Transitivity Alternations in Diachrony : Changes in Argument Structure and Voice Morphology [1 ed.] 9781443818100, 9781443817493

Citation preview

Transitivity Alternations in Diachrony Changes in Argument Structure and Voice Morphology

Transitivity Alternations in Diachrony: Changes in Argument Structure and Voice Morphology

By

Nikolaos Lavidas

Transitivity Alternations in Diachrony: Changes in Argument Structure and Voice Morphology, by Nikolaos Lavidas This book first published 2009 Cambridge Scholars Publishing 12 Back Chapman Street, Newcastle upon Tyne, NE6 2XX, UK

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

Copyright © 2009 by Nikolaos Lavidas All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-4438-1749-X, ISBN (13): 978-1-4438-1749-3

CONTENTS

List of Tables ......................................................................................................................................... vii Preface ................................................................................................................................................... ix List of Abbreviations and Acronyms ...................................................................................................... xi Chapter One ............................................................................................................................................ 1 Introduction 1.1 Aim and Objectives of the Research ........................................................................................... 1 1.2 Transitivity, Prototypical Transitivity, and Causativity ............................................................... 3 1.3 Transitivity Alternations.............................................................................................................. 9 1.4 The Study of Transitivity and the Diachronic Approach ........................................................... 14 1.4.1 The problems and the contribution of diachronic study ................................................... 14 1.4.2 Sources ............................................................................................................................. 15 1.5 Terminology Issues ................................................................................................................... 17 1.6 The Structure of the Monograph ............................................................................................... 20 Chapter Two ......................................................................................................................................... 25 Theoretical Framework 2.1 Transitivity, Syntax, and Lexicon-Syntax Interface .................................................................. 25 2.1.1 Transitivity and syntax ..................................................................................................... 26 2.1.2 Transitivity and lexicon-syntax interface ......................................................................... 36 2.1.3 The derivational basis issue: the relationship between the derivation of the transitive and intransitive types ........................................................................................................... 42 2.1.4 Morphology, argument structure, and transitivity ............................................................ 44 2.1.5 Summary .......................................................................................................................... 48 2.2 Language Change and Transitivity............................................................................................ 50 2.2.1 Generative Grammar and diachrony ................................................................................ 50 2.2.2 Language change, language acquisition, and reanalysis................................................... 52 2.2.3 Input, cues, and PLD ........................................................................................................ 53 2.2.4 Diachrony and transitivity in Greek: a review of the literature ........................................ 56 2.2.5 Summary .......................................................................................................................... 61 Chapter Three ....................................................................................................................................... 63 Diachronic Data and Analysis 3.1 From Homeric to Classical Greek ............................................................................................. 63 3.1.1 The overall picture ............................................................................................................ 63 3.1.2 Innovative causative use of intransitive verbs: from ablaut alternation to productive causativisation ..................................................................................................................... 65 3.1.3 Derivation and voice of the transitivity alternations: the contrast between the active causative and the non-active anticausative .......................................................................... 78 3.1.4 Summary .......................................................................................................................... 92 3.2 The Hellenistic-Roman Period .................................................................................................. 94 3.2.1 The general picture ........................................................................................................... 94 3.2.2 Innovative causative use of intransitive verbs: causativisation of intransitives vs. case changes in accusative verbs.................................................................................... 94 3.2.3 Transitivity alternation derivation and voice: changes towards marking of the anticausative with active voice ................................................................................ 106 3.2.4 Summary .........................................................................................................................118 3.3 Medieval Greek....................................................................................................................... 120 3.3.1 The general picture ......................................................................................................... 120 3.3.2 Innovative causative use of intransitive verbs: is there a limit to causativisations? ....... 121 3.3.3 Derivation of transitivity alternations and voice: absorption of the accusative case by the non-active voice and tendency towards active causatives and anticausatives ........ 130 3.3.4 From Medieval to Modern Greek ................................................................................... 152 3.3.5 Summary ........................................................................................................................ 161

Contents

3.4 Parallel Developments in the Diachrony of English and Romance ........................................ 162 3.4.1 Innovative causative use of intransitive verbs: direction of the changes ........................ 162 3.4.2 Derivation of transitivity alternations and voice: morphological instability in transitivity alternations and differentiation between anticausative and passive ............ 168 3.4.3 Summary ........................................................................................................................ 176 Chapter Four ....................................................................................................................................... 179 Diachronic Course and Derivation of Transitivity Alternations 4.1 Reanalysis, Cues, and Causativisation .................................................................................... 179 4.1.1 Reanalysis of intransitive as transitive ........................................................................... 179 4.1.2 Cues for reanalysis ......................................................................................................... 183 4.2 The Role of Voice in Syntactic Changes ................................................................................. 187 4.2.1 The course and the interpretation of changes in voice .................................................... 187 4.2.2 Voice head, anticausative, and passive ........................................................................... 191 4.2.3 The feature [non-act] ..................................................................................................... 200 4.3 Derivation of Transitivity Alternations: the Basic Causative Schema .................................... 204 4.3.1 Basic schema of alternating verbs: previous analyses .................................................... 204 4.3.2 The proposed analysis .................................................................................................... 206 4.4 Summary ................................................................................................................................. 213 Chapter Five ....................................................................................................................................... 217 Conclusions Bibliography ....................................................................................................................................... 223 Index ................................................................................................................................................... 257

LIST OF TABLES

1.1

Hopper & Thompson’s transitivity parameters............................................................................ 4

2.1

Language acquisition and change: I-language and E-language ................................................. 52

3.1

Middle and passive non-active forms in the future and past tense ............................................ 79

3.2

Syncretism of causative and anticausative types in English .................................................... 169

3.3

Morphological dissociation and periphrases ........................................................................... 174

3.4

Morphological dissociation and pronoun SE........................................................................... 174

3.5

Morphological dissociation between anticausatives and passives ........................................... 175

4.1

Formation of labile types in Greek .......................................................................................... 187

4.2

Overview of anticausatives and passives in the diachrony of Greek....................................... 198

PREFACE

This book is a revised version of my doctoral dissertation (“Changes in Verb Transitivity”, National and Kapodistrian University of Athens, 2006). My first thanks go to my dissertation committee: Dimitra Theophanopoulou-Kontou, Despoina Chila-Markopoulou, Amalia Moser. I would also like to thank the following people (in alphabetical order) for the comments, criticism, encouragement, and advice they provided at various stages in my research and preparation of this book: Artemis Alexiadou, Elena Anagnostopoulou, Georgios Babiniotis, Christoforos Charalambakis, Noam Chomsky, Burç ødem Dincel, Gaberell Drachman, Georgios Giannakis, Sabine Iatridou, Eleni Karantzola, Grammatiki Karla, Panagiotis Kontos, Carol Koulikourdi, Leonid Kulikov, Beth Levin, Angeliki Malikouti-Drachman, Amanda Millar, Elissavet Nouchoutidou, Nikolaos Pantelidis, Elizabeth Papadopoulou, Georgios Papanastasiou, Dimitra Papangeli, Anna Roussou, Nicolas Seferiades, Tasos Tsangalidis, Ianthi Maria Tsimpli, Christos Tzitzilis, Spyridoula Varlokosta. All errors are of course mine. Finally, I would like to thank all of the staff members of the Department of Theoretical and Applied Linguistics, School of English, Aristotle University of Thessaloniki for providing me with my “second home”.

Thessaloniki December 2009

LIST OF ABBREVIATIONS AND ACRONYMS

ABS ACC AGR ACT ANTIC AOR ASP CAUS Compl CP DAT DP EPP ERG FUT FUTPERF GEN GER IE IMP IMPERF IMPRFVE INF INTR IP MID NACT NACT1 NACT2 NOM NP O/OBJ OPT PART PASS PERF PERFVE PIE PL PLD PLUP POSS PP PRED PRES prt REFL S/SUBJ SG Spec TP TRANS UG V

absolutive case accusative case agreement active voice morphology anticausative morpheme aorist tense aspect causative morpheme Complement complementiser phrase dative case Determiner Phrase Extended Projection Principle ergative case future tense future perfect tense genitive case gerund Indo-European imperative mood imperfect tense imperfective aspect infinitive intransitive type Inflection Phrase middle voice non-active voice morphology type 1 of non-active voice morphology type 2 of non-active voice morphology nominative case Noun Phrase object optative mood participle passive voice (present) perfect tense perfective aspect Proto-Indo-European plural marker Primary Linguistic Data pluperfect tense possessive marker Prepositional Phrase Predication head present tense particle reflexive subject singular marker Specifier Tense Phrase transitive type/Transitivity head Universal Grammar verb

xii

VOC VoiceP vP VP WEAK

List of Abbreviations and Acronyms

vocative case Voice Phrase light verb Phrase Verb Phrase weak pronoun

CHAPTER ONE INTRODUCTION

1.1 Aim and Objectives of the Research This monograph examines the diachrony of verb transitivity (putting an emphasis on examples from diachrony of Greek verb transitivity) and, more particularly, the changes in causative verbs (verbs that denote change-of-state) and in transitivity alternations, defining the aspects of change in the lexicon that are constrained by Universal Grammar and those that are constrained by specific characteristics of language in every historical period. Furthermore, the aim is to show the mechanisms and processes of change in lexicon and in syntax that are concerned with the transitivity of verbs1. More specifically, I present the results of the research on diachronic changes in Greek in relation to transitivity alternations/alternating verbs: (1) English a. John opened the window b. The window opened Greek aǯ. ȅ īȚȐȞȞȘȢ ȐȞȠȚȟİ IJȠ ʌĮȡȐșȣȡȠ2 o Janis anikse to parathiro the.NOM Janis.NOM open.ACT.PAST.PERFVE.3SG the.ACC window.ACC ‘Janis opened the window’ bǯ. TȠ ʌĮȡȐșȣȡȠ ȐȞȠȚȟİ to parathiro anikse the.NOM window.NOM open.ACT.PAST.PERFVE3SG ‘The window opened’,

in comparison with non-alternating verbs: (a) transitive verbs (with no intransitive use): (2) English a. John cut the bread b. *The bread cut by itself Greek aǯ. O īȚȐȞȞȘȢ ȑțȠȥİ o Janis ekopse

1

2

IJȠ to

ȥȦȝȓ psomi

‘Transitivity’ is, in reality, a neutral term (which is why I will frequently call it as ‘(in)transitivity’) since it does not refer to the presence of a direct object but constitutes a characteristic of the verb that can receive a positive indicator/mark [+transitive] and means that the verb takes a direct object or a negative indicator/mark [-transitive] and demonstrates the obligatory absence of the direct object. Ǽxamples from Greek are quoted in both Greek and Roman scripts; the decision to transliterate the Greek can help linguists with no training in Greek to examine, check and cite the relevant examples given here (cf. Goldstein 2008). Examples will be given a broad phonological transcription; stress will be not marked except for the cases where the marking of stress is necessary for the argumentation (cf. Ralli 2009). As is well known, the phonological system of Greek has changed significantly during the Hellenistic-Roman period. For that reason and as details about the phonological system and its changes are not so relevant to this present study of syntactic changes, Ancient and Hellenistic-Roman Greek examples are transcribed according to the Classical Greek pronunciation (since not all of the changes were completed by the end of the Hellenistic period), whereas examples of the Medieval and Modern Greek period are given a Modern Greek transcription. For readability the vowels are presented by transliterating the Greek orthography, not in phonological transcription (cf. Kiparsky 2003): (i) for I write Ƞ: (ii) for , ou (iii) for , ei and (iv) for , e:. Phonologically corresponds to /o:/, corresponds both to /ou/ and to /oo/, both to /e:/ and /ei/, and to /İ:/. This ‘shortcut’ is not problematic in the present context because this present study of syntactic change does not depend on vowel quality.

2

Chapter One the.NOM Janis.NOM cut.ACT.PAST.PERFVE.3SG the.ACC ‘Janis cut the bread’ bǯ. *TȠ ȥȦȝȓ ȑțȠȥİ Įʌȩ *to psomi ekopse apo the.NOM bread.NOM cut.ACT.PAST.PERFVE.3SG by ‘*The bread cut by itself’,

bread.ACC ȝȩȞȠ IJȠȣ mono tu itself

(b) non-alternating unaccusative intransitive verbs: (3) English a. The tree fell b. *John fell the tree Greek aǯ. ȉȠ įȑȞIJȡȠ ȑʌİıİ to dhendro epese the.NOM tree.NOM fall.ACT.PAST.PERFVE.3SG ‘The tree fell’ bǯ. *ȅ īȚȐȞȞȘȢ ȑʌİıİ *o Janis epese the.NOM Janis.NOM fall.ACT.PAST.PERFVE.3SG ‘*Janis fell the tree’

IJȠ to the.ACC

įȑȞIJȡȠ dhendro tree.ACC

and (c) unergative intransitive verbs: (4) English a. Maria came b. *The bus came Maria back to the hotel Greek aǯ. Ǿ ȂĮȡȓĮ Ȓȡșİ i Maria irthe the.NOM Maria.NOM come.NACT.PAST.PERFVE.3SG ‘Maria came’ bǯ. *ȉȠ ȜİȦijȠȡİȓȠ Ȓȡșİ IJȘ ȂĮȡȓĮ ıIJȠ ȟİȞȠįȠȤİȓȠ *to leoforio irthe ti Maria sto ksenodhoxio the.NOM bus.NOM come.NACT.PAST.PERFVE.3SG the.ACC Maria.ACC to-the hotel ‘*The bus came Maria back to the hotel (‘the bus brought Maria back to the hotel’).

Emphasis is placed on transitivity alternations (ex. 1a, b, aǯ, bǯ) for three principal reasons: (a) the verb class (verbs denoting change-of-state) that participates in these alternations constitutes the central or prototypical class of transitive verbs (as I will argue in the second chapter/theoretical framework); (b) the diachronic behaviour of these verbs -in comparison with the diachronic behaviour of the exclusively transitive or exclusively intransitive verbs- can present the basic characteristics of the mechanism of change in verb transitivity; (c) the presence of a transitivity alternation even at a synchronic level comprises one of the basic means of change in verb valency since it concerns basic grammatical phenomena such as transitivisation and intransitivisation, as well as changes in voice morphology ((in)transitivity morphological markers). The research belongs to the field of diachronic studies and examines diachronic data through the prism of the Generative Grammar diachronic approach of Lightfoot (1979, 1999, 2006a, b). It contributes to the field of language change since it constitutes a systematic attempt at research on the diachrony of syntax and of the lexicon-syntax interface. Many studies have concerned themselves with the theoretical analysis of transitivity alternation in Modern Greek (Theophanopoulou-Kontou 1980, 1982, 1999b, 2004; Tsimpli 1989, 2006a; Alexiadou & Anagnostopoulou 1999a, b, 2004), with particularly significant results, but the issue of the diachronic development of transitivity has not been the subject of systematic research. The approach and basic position concerning the organisation of the lexicon and its interaction with syntax adopted by this monograph is that verb meaning has an effect on verb syntax (Levin & Rappaport Hovav 1995, 2005; Rappaport Hovav & Levin 1998). The verb has structured meaning that derives from the meaning of the event type. For verbs that participate in syntactic alternations, and particularly for verbs that participate in transitivity alternations, information on the event type can be

Introduction

3

marked by overt or zero morphemes3. This morphological process has syntactic consequences since it expresses the relationship that exists between the arguments of the verb. From the numerous general issues that can be raised concerning the argument structure, two principal issues are examined in this study: The first concerns the explanation of the limited possible verb structures that exist in the various periods not only of Greek but of all languages. This issue arises from the observation (Hale & Keyser 1993; Juarros-Daussà 2003) that a predicate (without the help of an additional lexical head) can only accept one (mono-transitive verbs) or at the most two (ditransitive verbs) arguments (with the exception of the external argument/subject4). The majority of theories on argument structure are based on ș-roles and/or primitives of the aktionsart. The aim of my analysis is to demonstrate the mechanisms that synchronously (in one period of the history of the Greek language) but also diachronically (in different periods in the history of the Greek language) limit the possibilities for the extension and modification of an argument structure and connect specific verb classes with specific structures. The second general issue concerns the correlation of the lexical-semantic structure, in other words, of the verb meaning as it is typically represented in modern theory, with syntactic structures. This problem is connected with the traditional problem of the correlation of (semantic) arguments (in other words the elements that semantically complete a verb) with syntactic representations, and it is concerned with the lexicon-syntax interface5. It is, of course, evident that an exhaustive examination of all the phenomena and questions connected with transitivity and all the verb classes of the periods covered by this study (5th century BC to the 17th/18th century AD) would lead to research that exceeds the size restrictions permissible for a monograph since a wide variation of syntactic behaviours and changes in the syntax of various verb classes can be observed. For this reason, the research into the diachronic theory of transitivity takes into consideration three specific questions: (a) What is the connection between transitive and intransitive types? Are transitive and intransitive types connected lexically with processes of (lexical) transitivisation (therefore, the intransitive type is the basic type [Dowty 1979; Pesetsky 1995]) or of (lexical) de-transitivisation (therefore, the transitive type is the basic type [Levin & Rappaport Hovav 1995; Chierchia 1989/2004; Reinhart 1997, 2000])? (b) What is the role of morphology? Is morphological marking connected with lexical derivation? Is there a link between the additional morphology and the derived form? (c) Why do languages and historical periods of the same language vary regarding which verbs participate or do not participate in transitivity alternations? What is the relationship between (a) intransitive alternating verbs (anticausatives), (b) intransitive passive verbs, and (c) transitive verbs (as a whole, and especially with causative transitive verbs)? (cf. the theoretical examination of (anti)causativisation by Alexiadou 2006a, c). Finally, a necessary supplement to a complete study of (in)transitivity would consist of the examination of changes in relation to the arguments and adjuncts that complete the transitive (causative) structures (agent, cause or instrument and patient) and the arguments that complete the intransitive structures (patient, cause or instrument and/or agent [in the case of intransitive passive structures]). Parallel to the aforementioned primary questions, an examination was also undertaken regarding the following issues that correlate with the previous central questions of diachronic research: the presence or loss of prefixes for innovative transitive or intransitive structures and the changes in the (direct) object case (genitive, dative morphological case, and prepositional phrases that replace or are replaced by the accusative morphological case). The tendency, as will be shown, is only towards the extension of the morphological accusative case, but a distinction between the different instances of the extension of the accusative case is vital.

1.2 Transitivity, Prototypical Transitivity, and Causativity In this section, I will try to present the reasons why I posit causative verbs as the focus of my research on transitivity, as well as the precise characteristics that distinguish them from the rest of the transitive verbs and, of course, from intransitive verbs. The theoretical analysis of the prototypical transitive 3 4

5

Cf. Drachman 2005, 2009. The external argument is ‘the argument of a predicate X which is not contained in the maximal projection of X; in general, this is the subject of a predicate’ (J. Kerstens, E. Ruys & J. Zwarts 1996-2001: Lexicon of Linguistics. Utrecht Institute of Linguistics OTS, Utrecht University – Williams 1980; Chomsky 1981, 1986). The correlation of the semantic structure with the syntactic structure is dealt with theoretically either by the same syntactic principles that also define the relations in the syntactic level or by the separate level of the lexicon-syntax interface, which is based on special rules of correlation, thematic hierarchies, and linking rules.

4

Chapter One

verbs, i.e. causative verbs, is based, on the one hand, on Tsunoda’s approach, which modified Hopper & Thompson’s classical study on transitivity, and, on the other hand, on the analyses of Levin & Rappaport Hovav. (a) Hypotheses on transitivity and unaccusativity Hopper & Thompson (1980, 1982) analysed transitivity as a continuum and attempted to show the parameters of the complete transitive sentence that are related to grammatical universals (which appear in the typology of languages) and to the use of language. Table 1.1 Hopper & Thompson’s transitivity parameters (1980: 252) High Transitivity

Low Transitivity

two or more participants (agent and object)

one participant

KINESIS

John hugged Maria. action.

Peter died. non-action

ASPECT

Kain killed Abel. telic

I like Helen. atelic

PUNCTUALITY

I ate it up. punctual

I’m eating it. non-punctual

VOLITIONALITY

Mary kicked the ball. volitional

Mary carried the book. non-volitional

AFFIRMATION

I wrote your name. I broke the glass. affirmative

I forgot your name The glass broke. negative

MODE

Peter smokes. realis

Peter does not smoke. irrealis

I picked up the phone. The agent is high in potency [+animate]

If I had picked up the phone. The agent is low in potency [-animate]

George startled me. John frightened me. The object is totally affected (total change of object)

The picture startled me. The picture frightened me. The object is not affected (partial change of object)

I drank the milk. High degree of individuation of object (proper nouns, concrete animates, singular number, countable, definite)6

I drank some milk. The object is not individuated

PARTICIPANTS

AGENCY

AFFECTEDNESS OF OBJECT

INDIVIDUATION OF OBJECT

I bumped into the table.

I bumped into Charles. In relation to the previously shown parameters, Hopper & Thompson formulated the Transitivity Hypothesis (1980: 255): ‘If two clauses (a) and (b) in a language differ in that (a) is higher in transitivity according to any of the features, then, if a concomitant grammatical or semantic difference appears elsewhere in clauses, that difference shows (a) to be higher in Transitivity’ 7. 6

7

It should be noted that the indirect animate object is considered an indication of greater transitivity than any direct object. Hagège (1982) also dealt with transitivity as a continuum. For Hagège, the contrast between ‘strong’ and

Introduction

5

The degree of transitivity (high-low) has morphosyntactic and semantic consequences: the morphological markedness of the object, according to the analysis by Hopper & Thompson, is attributed not to the distinction between subject and object but is the result of the degree of transitivity of the sentence as a whole, as it arises from the parameters of transitivity. Hopper & Thompson support their hypothesis by analysing, in a large number of languages, the case marking of the object, the case marking of the subject (which bears the semantic role of the agent or the patient), the incorporation of the object into the verb, the word order, the verb morphology, the anti-passive, and the reflexive structure. Furthermore, according to recent approaches, neither transitive verbs nor intransitive verbs make up a homogeneous group; the Unaccusativity Hypothesis (Permutter 1978; Burzio 1986) distinguishes two subclasses of intransitives, the unergative and the unaccusative intransitives, each one of which is related to a different syntactic structure. The unergative verbs (e.g., swim) have their sole argument in the position of the syntactic subject ([Spec, vP]) but have the position of the syntactic object empty ([Compl, VP]); the sole argument of the unaccusatives (e.g., appear) is generated in the [Compl, VP] position, but the subject position is empty. With the application of the rule ‘move Į’ (movement of NP) according to the older theoretical perspective or to receive/check the nominative case, the sole argument of the unaccusative verbs moves to the empty position of the subject. (5) a. unergative verb (swim, sing, dance): b. unaccusative verb (appear, intransitive break, intransitive open):

DP [vP V ] __ [vP V DP]

In thematic terms, the unergative verbs take agents as arguments and the unaccusative verbs take patients or themes as arguments. Burzio (1986) connected the ability of a verb to have an external argument with the ability to assign structural case (Burzio’s Generalisation): an unaccusative verb, precisely because it does not have an external argument, is unable to take an object in the accusative case, or, according to the Government and Binding theory, it does not assign structural case to its object8.

8

‘weak’ transitivity in different languages refers to: (1) the degree of volitionality for the accomplishment of an action, (2) the degree of the affectedness of the patient, (3) the degree of accomplishment of the action and (4) the degree of definiteness of the patient. The structure of intransitive predicates in Modern Greek and their relationship with the Unaccusativity Hypothesis have occupied many linguists (Markantonatou 1992; Kakouriotis 1993, 1994; Sioupi 1997; Embick 1998; Alexiadou & Anagnostopoulou 1999a, b; Theophanopoulou-Kontou 1999a, b, 2003a). Markantonatou (1992) and Alexiadou & Anagnostopoulou (1999a) attempted to define the criteria that distinguish the intransitive verbs of Modern Greek into unergative and unaccusative since the majority of the criteria used cross-linguistically cannot be applied to Modern Greek (all intransitive verbs in Modern Greek take the verb ‘have’ as auxiliary; Modern Greek does not have resultative and impersonal passive structures). An important criterion for Modern Greek verbs has been said to be the ability to form adjectival participles for unaccusative but not for unergative verbs: (1) a.

b.

ʌİıȝȑȞȠ ijȪȜȜȠ pezmeno filo fall.NACT.PART leaf ‘fallen leaf’ *IJȡİȖȝȑȞȠȢ ȐȞșȡȦʌȠȢ *treghmenos anthropos run.NACT.PART human ‘a person who has run a lot’

Furthermore, unaccusative verbs in the past tense, when the subject exhibits something countable, do not allow the presence of a progressive adverbial (and atelic interpretation); only when the subject exhibits quantity is the progressive adjunct (and atelic interpretation) allowed: (2) a. *ȅ ĬȘıȑĮȢ ȑijIJĮȞİ İʌȓ ȝȑȡİȢ ıIJȘȞ ǹșȒȞĮ *o Thiseas eftane epi meres stin Athina the.NOM Theseus.NOM arrive.ACT.PAST.IMPRVE.3SG for days in Athens ‘*Theseus was arriving for days in Athens’ b. ǹșȜȘIJȑȢ ȑijIJĮȞĮȞ İʌȓ ȝȑȡİȢ ıIJȘȞ ǹșȒȞĮ athlites eftanan epi meres stin Athina athletes.NOM arrive.ACT.PAST.IMPRFVE.3PL for days in Athens ‘Athletes were arriving for days in Athens’

Finally, Alexiadou & Anagnostopoulou (1999a) point out that unaccusative verbs of change-of-state in Modern

6

Chapter One

(b) Prototypical transitivity The transfer of emphasis from the transitivity continuum to prototypical transitivity was first undertaken by Cooreman, Fox & Givón (1984). For them, transitivity is concerned with three main characteristics of the sentence: (a) the agent: the prototypical transitive sentence has an explicitly denoted agent (or cause), who acts volitionally and has full control over the action; (b) the patient: the prototypical transitive sentence has an explicit patient who does not participate of his/her own volition in the action and does not control the action; and (c) the verb: a prototypical transitive sentence has a verb that shows the realis mood and perfective aspect. The sentences that bear the three aforementioned properties are prototypically transitive. The concept of prototypical transitivity has been accepted and used (sometimes with modifications) in many studies: Tsunoda (1985, 1994); DeLancey (1987); Croft (1990); Antonopoulou (1991); Kibrik (1993); Ono (1999); Lazard (2006); Næss (2007)9. Tsunoda & Lazard reject, however, the opinion that all the parameters of transitivity are connected equally among themselves: a sentence may exhibit low transitivity in relation to a parameter (for example, the definiteness of the object) and high transitivity in relation to another parameter (for example, the presence of the agent). Lazard considers the presence of two participants in a sentence as a fundamental and necessary condition to be able to talk about transitivity and not simply as one of the many determinants that indicate high transitivity. For Tsunoda, the close relationship of only some of the factors (the volitionality and agent) is also of particular importance, as is the existence of a hierarchy of the morphosyntactic marking (the affectedness of the object appears to constitute a more significant factor than the volitionality and the agent for the morphosyntactic marking of increased transitivity) and the scale of affectedness (a continuum of affectedness that has consequences for the morphosyntactic markedness of the object). Tsunoda, using as a basis the aforementioned correlations of high transitivity parameters, creates a hierarchy of verbs based on the probability that semantic classes of verbs have of being transitive crosslinguistically. The verbs that belong to the semantic classes that are higher in the hierarchy are more likely to be transitive across languages. The classes and their particular semantic characteristics are as follows: (i) Verbs with a direct effect on the argument of the patient and resultative meaning (causativity): ‘kill’, ‘break’, ‘bend’ (ii) Verbs with a direct effect on the argument of the patient but with non-resultative meaning: ‘hit’, ‘kick’, ‘shoot’ (iii) Verbs of perception; patient more attained: ‘see’, ‘hear’, ‘find’ (iv) Verbs of perception; patient less attained: ‘look’ (v) Verbs of pursuit: ‘search’, ‘wait’ (vi) Verbs of knowledge: ‘know’, ‘understand’, ‘remember’, ‘forget’ (vii) Verbs of feeling: Greek can form transitive structures, whereas the unergative verbs cannot: (3) a.

b.

9

ȉȠ ȕȐȗȠ to vazo the.NOM vase.NOM ‘The vase broke’ ȅ īȚȐȞȞȘȢ o Janis the.NOM Janis.NOM ‘Janis broke the vase’

ȑıʌĮıİ espase break.ACT.PAST.PERFVE.3SG ȑıʌĮıİ espase break.ACT.PAST.PERFVE.3SG

IJȠ to the.ACC

ȕȐȗȠ vazo vase.ACC

Antonopoulou recognises prototypical transitive verbs and non-prototypical transitive verbs; the core of transitivity is sentences that contain natural, punctual, significant, and total change of the state of the patient. According to Antonopoulou (cognitive approach of transitivity) transitivity concerns syntax, semantics, and pragmatics: a situation that is dealt with cognitively as transitive is comprised not only of two participants but also of their interaction and of the complete affectedness of one of the two. Transitivity is the superordinate category: the active transitive structure, the passive structure without agent (=anticausative structure), and the passive structure with agent make up the basic categories. The passive structure comprises the marked member of the category and has many characteristics of de-transitivisation in the form in which it is expressed. The focus on the patient is the main parameter of transitivitisation, whereas the defocalisation of the agent is considered the main parameter of de-transitivisation.

Introduction

7

‘love’, ‘like’, ‘want’, ‘need’, ‘fear’ (viii) Verbs of relationship: ‘have’, ‘possess’, ‘resemble’, ‘correspond’, ‘consist’ The aforementioned classes (semantically highly detailed) could, on the basis of their aktionsart and Levin’s approach to verb classes (1993, 2000, 2004), be limited to the following four classes10: (i) Resultative verbs with a direct effect on the patient argument (causative event structure): ‘kill’, break’, ‘bend’, ‘murder’ (ii) Non-resultative verbs with a direct effect on the argument of the patient (and surface contact) (simple event structure): ‘hit’, ‘kick’, ‘wipe’, ‘scratch’ (iii) Verbs of pursuit/ denotation of relation with the second argument (simple event structure): ‘search’, ‘wait’, ‘await’, ‘see’ (iv) Stative verbs of perception, knowledge, and feeling (simple event structure): ‘see’, ‘hear’, ‘know’, ‘understand’, ‘love’, ‘fear’ The resultative verbs with a direct effect on the argument of the patient (causative verbs) can be distinguished into smaller semantic classes (cf. Levin 1993; for Modern Greek data see Theophanopoulou-Kontou et al. 1998)11. (i) Verbs that mean ‘break’. Only the result (and the specific form of the result) is expressed; the manner of the action is not denoted: ıʌȐȦ spao ‘to break’, ȡĮȖȓȗȦ rajizo ‘to crack’, ıȣȞIJȡȓȕȦ sindrivo ‘to crash’, IJİȝĮȤȓȗȦ temaxizo ‘to cut-up/dismember’, șȡȣȝȝĮIJȓȗȦ thrimatizo ‘to shatter’, țȠȝȝĮIJȚȐȗȦ komatiazo ‘to cut-into-pieces’, įȚĮıʌȫ diaspo ‘to split’ . (ii) Verbs that mean ‘bend’. They have the same characteristics as the verbs of the first subclass; the only difference is the ability to be restored to their previous state (ȟİįȚʌȜȫȞȦ țȐIJȚ ksedhiplono kati ‘to unfold something’  *ȟİȡĮȖȓȗȦ țȐIJȚ kserajizo kati ‘to uncrack something’): ȜȣȖȓȗȦ lijizo ‘to bend’, IJıĮțȓȗȦ tsakizo ‘to crease’, ȗĮȡȫȞȦ zarono ‘to wrinkle’, IJıĮȜĮțȫȞȦ tsalakono ‘to crumple’, įȚʌȜȫȞȦ dhiplono ‘to fold’. (iii) Verbs that indicate ‘cooking’: ȕȡȐȗȦ vrazo ‘to boil’, ȗİıIJĮȓȞȦ zesteno ‘to heat’, țĮȓȦ keo ‘to burn’, ȝĮȖİȚȡİȪȦ majirevo ‘to cook’, ȥȒȞȦ psino ‘to roast’, IJȘȖĮȞȓȗȦ tighanizo ‘to fry’, țĮȕȠȣȡȞIJȓȗȦ kavurndizo ‘to sauté’, ȟİȡȠȥȒȞȦ kseropsino ‘to dry roast’, ıȚȖȠȥȒȞȦ sighopsino ‘to slow roast’, ıȚȖȠȕȡȐȗȦ sighovrazo ‘to simmer’. (iv) Verbs that indicate a change-of-state of a specific substance. These verbs have strict selectional restrictions12: 10

11

12

Differentiation of transitive verbs has also been supported in the Cognitive (Langacker 1991) and the Systematic-Functional approaches (Halliday 1980, 1985a, b): ergative vs. transitive verbs. Davidse (1991, 1992) and Lemmens (1998) consider as central the distinction between ergative (which correspond to what we call causative transitive verbs) and transitive structures (which correspond to what we call non-causative transitive verbs or accusative verbs). Davidse argues that in the ergative structures with two participants, the participant (‘medium’) who is affected by the event (the direct object) is presented as the most involved since it is both affected and participates in the process (ȅ ȆĮȞIJİȜȒȢ ȑȜȚȦıİ IJȠ IJȣȡȓ / o Pandelis eliose to tiri / the.NOM Pantelis.NOM melt.ACT.PAST.PERFVE.3SG the.ACC cheese.ACC / ‘Pantelis melted the cheese’; H DZȞȞĮ ȐȞȠȚȟİ IJȘȞ ʌȩȡIJĮ / i Ana anikse tin porta / the.NOM Anna.NOM open.ACT.PAST.PERFVE.3SG the.ACC door.ACC / ‘Anna opened the door’). The structural centrality of the aforementioned process is reflected by the fact that the participant (the core of the above process) can be isolated without the presence of the agent (TȠ IJȣȡȓ ȑȜȚȦıİ / to tiri eliose / the.NOM cheese.NOM melt.ACT.PAST.PERFVE.3SG / ‘The cheese melted’, Ǿ ʌȩȡIJĮ ȐȞȠȚȟİ / i porta anikse / the.NOM door.NOM open.ACT.PAST.PERFVE.3SG / ‘the door opened’). The ergative differs from the transitive structure which contains an ‘inactive’ affected participant. The main characteristic for transitive structures is the agent (‘actor’), who acts with volition. The fundamental structural characteristics for the transitive structure are, therefore, agent and process, which can be isolated in transitive sentences without an object (Ǿ ȐșȜȘıȘ ȕȠȘșȐİȚ / i athlisi voithai / the.NOM athletics.NOM help.ACT.PRES.3SG. / ‘Athletics helps’). The structure expands in cases that also contain the goal towards which the action of the agent is directed. The ‘goal’ in transitive examples does not participate in the process but is an ‘inactive’ component that is totally affected by the agent’s action. It is worth noting that Tzartzanos (1946) (in relation to voice and diatheses) has already distinguished between two sub-groups of transitive verbs, the simple transitives (affectedness of the object but without change-ofstate – ȤIJȣʌȫ xtipo ‘to hit’) and the causatives (change-of-state of the direct object – ıțȠIJȫȞȦ skotono ‘to kill’). The following are examples of selectional restrictions:

8

Chapter One

ıțȠȣȡȚȐȗȦ skuriazo ‘to rust’, ıĮʌȓȗȦ sapizo ‘to rot’, ȝȠȣȤȜȚȐȗȦ muxliazo ‘to moulder’, ȦȡȚȝȐȗȦ orimazo ‘to ripen’, ijȠȣıțȫȞȦ fuskono ‘to swell’13, İȜĮIJIJȫȞȦ elatono ‘to reduce’, ȝİIJȡȚȐȗȦ metriazo ‘to moderate’, ȝİȖĮȜȫȞȦ meghalono ‘to grow’, ĮȣȟȐȞȦ afksano ‘to increase’, ıȣȝʌȣțȞȫȞȦ simbiknono ‘to condense’, ıȣȝʌȚȑȗȦ simbiezo ‘to compress’, ȕİȜIJȚȫȞȦ veltiono ‘to improve’, ȤİȚȡȠIJİȡİȪȦ xiroterevo ‘to worsen’, țȠȞIJĮȓȞȦ konteno ‘to shorten’, ȝȚțȡĮȓȞȦ mikreno ‘to make smaller’, ȝİȚȫȞȦ miono ‘to diminish’, ıȣȡȡȚțȞȫȞȦ siriknono ‘to shrink’, įȚʌȜĮıȚȐȗȦ dhiplasiazo ‘to double’, ĮįȣȞĮIJȓȗȦ adhinatizo ‘to weaken’, ijĮȡįĮȓȞȦ fardheno ‘to widen’, ĮțȡȚȕĮȓȞȦ akriveno ‘to make more expensive’, įȚİȣȡȪȞȦ dievrino ‘to extend’, ıțȜȘȡĮȓȞȦ sklireno ‘to harden’, ȝĮȜĮțȫȞȦ malakono ‘to soften’, įȣȞĮȝȫȞȦ dhinamono ‘to strengthen’, ȝİȖİșȪȞȦ meghethino ‘to enlarge’, İʌȚIJĮȤȪȞȦ epitaxino ‘to quicken’, İʌȚȕȡĮįȪȞȦ epivradhino ‘to retard’ / ĮȜȜȐȗȦ alazo ‘to change’, ȝİIJĮȕȐȜȜȦ metavalo ‘to transform’, ĮȞĮIJȡȑʌȦ anatrepo ‘to reverse’ / ȟȣʌȞȫ ksipno ‘to waken’ / İȟȚıȠȡȡȠʌȫ eksisoropo ‘to equalize’, ıIJİȡİȫȞȦ stereono ‘to stabilize’, ʌĮȖȚȫȞȦ pajiono ‘to dazzle’ / șĮȝʌȫȞȦ thambono ‘to dim’, șȠȜȫȞȦ tholono ‘to cloud’, ȜĮȝʌȡȪȞȦ lambrino ‘to brighten’, ıțȠIJİȚȞȚȐȗȦ skotiniazo ‘to darken’, ijȦIJȓȗȦ fotizo ‘to illuminate’ / țȡȣȫȞȦ kriono ‘to chill’, ʌĮȖȫȞȦ paghono ‘to freeze’, șİȡȝĮȓȞȦ thermeno ‘to heat’, ȟİʌĮȖȫȞȦ ksepaghono ‘to defrost’ / țȜİȓȞȦ klino ‘to close’, ȕȠȣȜȫȞȦ vulono ‘to seal-up’, ĮȞȠȓȖȦ anigho ‘to open’, ȤĮȜĮȡȫȞȦ xalarono ‘to loosen’ / ȖİȝȓȗȦ jemizo ‘to fill’, ĮįİȚȐȗȦ adhjazo ‘to empty’ / ȜȚȫȞȦ liono ‘to melt’, ĮʌȠıȣȞșȑIJȦ aposintheto ‘to decompose’, įȚĮȜȪȦ dhialio ‘to dissolve’ / țĮșĮȡȓȗȦ katharizo ‘to clean’, ȕȡȦȝȓȗȦ vromizo ‘to dirty’ / ĮıʌȡȓȗȦ asprizo ‘to whiten’, ȝĮȣȡȓȗȦ mavrizo ‘to blacken’, țȚIJȡȚȞȓȗȦ kitrinizo ‘to yellow’, țȠțțȚȞȓȗȦ kokinizo ‘to redden’, ʌȡĮıȚȞȓȗȦ prasinizo ‘to make green’/ ĮȡȡȦıIJĮȓȞȦ arosteno ‘to sicken’, ȖȚĮIJȡİȪȦ jatrevo ‘to cure’ / ȕȠȣȜȚȐȗȦ vuliazo ‘to sink’. (v) Verbs that indicate movement/change of position: țȣȜȐȦ kilao ‘to roll’, ıIJȡȑijȦ strefo ‘to turn’, ʌİȡȚıIJȡȑijȦ peristrefo ‘to rotate’, ȝİIJĮțȚȞȫ metakino ‘to move’, ʌĮȡĮıȪȡȦ parasiro ‘to sweep-away’, ıIJȡȚijȠȖȣȡȓȗȦ strifojirizo ‘to spin’. (vi) Verbs of change of psychological state14. Pesetsky (1995) argues that the subject of these

Ǿ ȝȠȣıȚțȒ įȣȞȐȝȦıİ i musiki dhinamose the.NOM music.NOM intensify.ACT.PAST.PERFVE.3SG ‘The music intensified’ ??ȅ șȩȡȣȕȠȢ įȣȞȐȝȦıİ ??o thorivos dhinamose the.NOM noise.NOM intensify.ACT.PAST.PERFVE.3SG ‘The noise intensified’ ȅ īȚȐȞȞȘȢ įȣȞȐȝȦıİ IJȘ o Janis dhinamose ti the.NOM Janis.NOM intensify.ACT.PAST.PERFVE.3SG the.ACC ‘Janis turned up the music’ ???ȅ īȚȐȞȞȘȢ įȣȞȐȝȦıİ IJȘ o Janis dhinamose ti the.NOM Janis.NOM intensify.ACT.PAST.PERFVE.3SG the.ACC ‘Janis made his voice louder’

(1) a.

b.

c.

d.

13 14

ȝȠȣıȚțȒ musiki music.ACC ijȦȞȒ IJȠȣ foni tu voice.ACC his

The oblique lines indicate smaller verb classes showing change-of-state of specific substances. The psych-verbs are separated into three categories (Belletti & Rizzi 1998): (a) ‘admire’: experiencer [Verb theme] (b) ‘like’ : e[[Verb theme[Preposition-experiencer] [experiencer-dative] (c) ‘frighten’: e[[Verb theme]experiencer] The last category of psych-verbs is of great interest. For the theoretical analysis of these verbs as causatives, see Tsimpli (1989), Grimshaw (1991), Croft (1993), and Pesetsky (1995). For Modern Greek, cf. also Kakouriotis & Kitis (1999). These verbs: (i) Participate in transitivity alternations: TĮ ʌĮȚįȚȐ ĮȖȡȓİȥĮȞ ta pedhia aghriepsan the.NOM children.NOM aggravate.ACT.PAST.PERFVE.3PL ‘The children aggravated the dogs’ ȉĮ ıțȣȜȚȐ ĮȖȡȓİȥĮȞ ta skilia aghriepsan the.NOM dogs.NOM aggravated.ACT.PAST.PERFVE.3PL ‘The dogs became aggravated’ (*by the children)’

(1) a.

b.

IJĮ ıțȣȜȚȐ ta skilia the.ACC dogs.ACC (*Įʌȩ IJĮ ʌĮȚįȚȐ) (*apo ta pedhia) (*by the children)

(ii) Do not form a passive structure (2) a.

ȅ

ȀȫıIJĮȢ

șȪȝȦıİ

IJȘ

ijȓȜȘ

IJȠȣ

Introduction

9

verbs always bears the ș-role of cause: Șȡİȝȫ iremo ‘to calm’, ȘıȣȤȐȗȦ isixazo ‘to pacify’, ĮȞȘıȣȤȫ anisixo ‘to make anxious’, ĮȖȡȚİȪȦ aghrievo ‘to make fierce’, șȣȝȫȞȦ thimono ‘to anger’, ȞİȣȡȚȐȗȦ nevriazo ‘to annoy’, ıIJİȞȠȤȦȡȫ stenoxoro ‘to worry’, ʌİȚıȝȫȞȦ pizmono ‘to make stubborn’, IJȡİȜĮȓȞȦ treleno ‘to madden’, ȝʌİȡįİȪȦ berdhevo ‘to confuse’, ȤĮȡȠʌȠȚȫ xaropio ‘to gladden’, İȣȤĮȡȚıIJȫ efxaristo ‘to please’, ȚțĮȞȠʌȠȚȫ ikanopio ‘to satisfy’, įȚĮıțİįȐȗȦ dhiaskehazo ‘to entertain’, İȞșȠȣıȚȐȗȦ enthusiazo ‘to make enthusiastic’. (vii) Aspectual verbs: ĮȡȤȓȗȦ arxizo ‘to begin’, IJİȜİȚȫȞȦ teliono ‘to end’, ıȣȞİȤȓȗȦ sinexizo ‘to continue’, ıIJĮȝĮIJȫ stamato ‘to stop’, țȡĮIJȫ krato ‘to hold’, ʌȡȠȤȦȡȫ proxoro ‘to forward’, İʌĮȞĮȜĮȝȕȐȞȦ epanalamvano ‘to repeat’. To summarise, according to the aforementioned theoretical analyses of Hopper & Thompson (‘transitivity hypothesis’), Perlmutter & Postal, and Burzio (‘unaccusativity hypothesis’), Tsunoda and Levin, we come to the preliminary conclusion that transitivity is not a unified phenomenon: many classes of verbs can be identified according to their transitive or non-transitive behaviour. If we consider as transitive all verbs that take DPs in the default case of the object (accusative for Modern Greek), then we must distinguish (with mainly semantic criteria) the verbs that denote change-of-state (causative verbs – e.g., ‘murder’, ‘destroy’, ‘open’) as core and central (see section 2.1.1 for the corresponding syntactic criteria, as well as 2.1.2 for the correlation between syntactic and semantic structure). In the following section, we will see that causative verbs are also differentiated according to their participation in transitivity alternation (alternating vs. non-alternation causative verbs).

1.3 Transitivity Alternations Only one class of causative verbs, the alternating causative verbs, participate in transitivity alternations, i.e. they have the possibility for both transitive (causative) (agent/cause + Verb + patient) and intransitive (anticausative) use (patient + Verb + P(repositional)P(hrase)- cause/instrument): (6)

a. ȅ īȚȐȞȞȘȢ ȐȞȠȚȟİ IJȠ ʌĮȡȐșȣȡȠ o Janis anikse to parathiro the.NOM Janis.NOM open.ACT.PAST.PERFVE.3PL the.ACC window.ACC ‘Janis opened the window’ b. ȉȠ ʌĮȡȐșȣȡȠ ȐȞȠȚȟİ (Įʌȩ ȝȩȞȠ IJȠȣ) to parathiro anikse (apo mono tu)

o Kostas thimose ti fili tu the.NOM Kostas.NOM anger.ACT.PAST.PERFVE.3SG the.ACC friend.ACC his ‘Kostas angered his friend’ b. *Ǿ ijȓȜȘ IJȠȣ ȀȫıIJĮ șȣȝȫșȘțİ Įʌȩ IJȠȞ ȀȫıIJĮ *i fili tu Kosta thimothike apo ton Kosta the.NOM friend.NOM the.GEN Kostas.GEN anger.NACT.PAST.PERFVE.3SG by the Kostas ‘Kostas’s friend was angered by Kostas’

The periphrastic passive structures (İȓȞĮȚ + -ȝȑȞȠȢ / ine + -menos / ‘be + past participle’) are grammatical with these verbs. c.

Ǿ ijȓȜȘ IJȠȣ ȀȫıIJĮ i fili tu Kosta the.NOM friend.NOM the.GEN Kostas.GEN ‘Kostas’s friend is angered with Kostas’

İȓȞĮȚ ine is.3SG

șȣȝȦȝȑȞȘ thimomeni anger.PART

(iii) Form middle structures: ǹȣIJȐ IJĮ ȡȠȪȤĮ ıIJİȖȞȫȞȠȣȞ İȪțȠȜĮ afta ta ruxa steghnonun efkola these.NOM the.NOM clothes.NOM dry.ACT.PRES.3PL easily ‘These clothes dry easily’ b. Ǿ ǼȜʌȓįĮ șȣȝȫȞİȚ İȪțȠȜĮ i Elpidha thimoni efkola the.NOM Elpida.NOM anger.ACT.PRES.3SG easily ‘Elpida gets angry easily’ c. ǹȣIJȩ IJȠ ȕȚȕȜȓȠ įȚĮȕȐȗİIJĮȚ İȣȤȐȡȚıIJĮ afto to vivlio dhiavazete efxarista this.NOM the.NOM book.NOM read.NACT.PRES.3SG pleasantly ‘This book reads pleasantly’

(3) a.

ȝİ IJȠȞ ȀȫıIJĮ me ton Kosta with the Kostas

10

Chapter One the.NOM window.NOM open.ACT.PAST.PERFVE.3SG (by itself) ‘The window opened (by itself)’

Change-of-state, however, is also shown by causative verbs that do not alternate (non-alternating causative verbs), such as the verb ‘cut’ or the verb ‘murder’. In this case, the intransitive structures are ungrammatical15. These verbs (causatives that do not alternate) can participate in intransitive structures (in Modern Greek bearing non-active morphology) but only with a passive reading; anticausative interpretation (change-of-state without the presence of the agent) for these verbs is not possible. (7)

a. *ȉȠ ȥȦȝȓ țȩʌȘțİ (Įʌȩ ȝȩȞȠ IJȠȣ) *to psomi kopike (apo mono tu) the.NOM bread.NOM cut.NACT.PAST.PERFVE.3SG (by itself) ‘*The bread cut by itself’ b. ȉȠ ȥȦȝȓ țȩʌȘțİ Įʌȩ IJȠȞ īȚȐȞȞȘ to psomi kopike apo ton Jani the.NOM bread.NOM cut.NACT.PAST.PERFVE.3SG by the Janis ‘The bread was cut by Janis’

The characteristics of verbs that participate in causative and anticausative structure (alternating causatives) are as follows (Levin & Rappaport Hovav 1995: 89-119, 2005: 117-130): (a) They denote change-of-state: e.g., ıʌȐȦ spao ‘to break’, causative: [x CAUSES [y BECOMES predicate]] anticausative: BECOME [y BROKEN] (see also section 2.1.2.2) (b) They denote an event that can be imprinted semantically as evolving with or without external cause (causative-anticausative): alternating causatives: (8)

a. ȅ īȚȐȞȞȘȢ ȐȞȠȚȟİ IJȘȞ ʌȩȡIJĮ o Janis anikse tin porta the.ȃȅȂ Janis.NOM open.ACT.PAST.PERFVE.3SG the.ACC door.ACC ‘Janis opened the door’ (change-of-state; agent-Verb-patient) b. Ǿ ʌȩȡIJĮ ȐȞȠȚȟİ i porta anikse the.NOM door.NOM open.ACT.PAST.PERFVE.3SG ‘The door opened’ (unspecified for external cause)

cf. non-alternating causatives: (9)

a. ȅȚ ȞİȩȞȣȝijȠȚ ȑțȠȥĮȞ IJȘȞ IJȠȪȡIJĮ i neonimfi ekopsan tin turta the.NOM newly-weds.NOM cut.ACT.PAST.PERFVE.3PL the.ACC cake.ACC ‘The newly-weds cut the cake’ (change-of state; agent-Verb-patient) b. *Ǿ IJȠȪȡIJĮ ȑțȠȥİ / țȩʌȘțİ Įʌȩ ȝȩȞȘ IJȘȢ i turta ekopse / kopike apo moni tis the.NOM cake.NOM cut.ACT.PAST.PERFVE.3SG / cut.NACT.PAST.PERFVE.3SG by itself ‘*The cake (was) cut by itself’ (obligatory presence of external cause) c. Ǿ IJȠȪȡIJĮ țȩʌȘțİ Įʌȩ IJȠȣȢ ȞİȩȞȣȝijȠȣȢ i turta kopike apo tus neonimfus the.NOM cake.NOM cut.NACT.PAST.PERFVE.3SG by the newlywebs ‘The cake was cut by the newlywebs’ (passive type)

cf. transitive non-causatives: (10) a. ȅ īȚȐȞȞȘȢ ȑȖȡĮȥİ IJȠ țİȓȝİȞȠ o Janis eghrapse to kimeno the.NOM Janis.NOM write.ACT.PAST.PERFVE.3SG the.ACC text.ACC ‘Janis wrote the text’ (activity verb) b. *ȉȠ țİȓȝİȞȠ ȑȖȡĮȥİ / ȖȡȐijIJȘțİ *to kimeno eghrapse / ghraftike 15

Įʌȩ ȝȩȞȠ IJȠȣ apo mono tu

With the same meaning that the transitive type also has (‘to cut the bread’). I do not refer here, in other words, to structures such as ȉȠ ȖȐȜĮ ȑțȠȥİ / To ghala ekopse / the.NOM milk.NOM cut.ACT.PAST.PERFVE.3SG / ‘The milk curdled’; these structures do not correlate with the corresponding transitives (*DzțȠȥĮ IJȠ ȖȐȜĮ / *Ekopsa to ghala / cut.ACT.PAST.PERFVE.1SG the.ACC milk.ACC / ‘*I curdled the milk’).

Introduction

11

the.NOM text.NOM write.ACT.PAST.PERFVE.3SG/ write.NACT.PAST.PERFVE.3SG by itself

‘*The text wrote / was written by itself’ (obligatory presence of external cause) c. ȉȠ țİȓȝİȞȠ ȖȡȐijIJȘțİ to kimeno ghraftike the.NOM text.NOM write.NACT.PAST.PERFVE.3SG ‘The text was written’ (passive type)

(c) They allow agents and causes as external arguments in their transitive use: (11) a. ȅ īȚȐȞȞȘȢ ȐȞȠȚȟİ IJȘȞ ʌȩȡIJĮ o Janis anikse tin porta the.NOM Janis.NOM open.ACT.PAST.PERFVE.3SG the.ACC door.ACC ‘Janis opened the door’ b. ȅ ĮȑȡĮȢ ȐȞȠȚȟİ IJȘȞ ʌȩȡIJĮ o aeras anikse tin porta the.NOM wind.NOM open.ACT.PAST.PERFVE.3SG the.ACC door.ACC ‘The wind opened the door’ -cf. non-alternating causatives: c. *ȅ İȞșȠȣıȚĮıȝȩȢ ȑțȠȥİ IJȘȞ IJȠȪȡIJĮ *o enthusiasmos ekopse tin turta the.NOM enthusiasm.NOM cut.ACT.PAST.PERFVE.3SG the.ACC cake.ACC ‘*The enthusiasm cut the cake’

(d) They do not allow agents but allow causes in their anticausative use: (12) a. Ǿ i

ʌȩȡIJĮ porta

ȐȞȠȚȟİ anikse

*Įʌȩ IJȠȞ țȜİȚįĮȡȐ/ Įʌȩ IJȠȞ ĮȑȡĮ/ Įʌȩ IJȠ ȑȞIJȠȞȠ ıʌȡȫȟȚȝȠ *apo ton klidhara/ apo ton aera/ apo to endono sproksimo

the.NOM door.NOM open.ACT.PAST.PERFVE.3SG by the locksmith/ the wind/ the intense pushing

‘The door opened by *the locksmith / the wind / the intense pushing’ (anticausative type) b. Ǿ ʌȩȡIJĮ ĮȞȠȓȤIJȘțİ Įʌȩ IJȠȞ țȜİȚįĮȡȐ/ *IJȠȞ ĮȑȡĮ/ IJȘȞ țȜȦIJıȚȐ IJȠȣ īȚȐȞȞȘ i porta anixtike apo ton klidhara/ *ton aera/ tin klotsia tu Jani the.NOM door.NOM open.NACT.PAST.PERFVE.3SG by the locksmith/ the wind/ the kick of-the Janis

‘The door was opened by the locksmith / *the wind / Janis’s kick’ (passive type)

The differences between anticausative and passive types (important mainly for the non-active anticausatives of Greek, which bear the same morphology as the passives), can be summarised as follows: (i) In passive types, a by-PP that denotes agent can be present (and is always implied); in the anticausative types, the presence of a by-PP of the agent results in an ungrammatical sentence. (13) a. Ǿ ʌȩȡIJĮ ĮȞȠȓȤIJȘțİ Įʌȩ IJȠȞ țȜİȚįĮȡȐ i porta anixtike apo ton klidhara the.NOM door.NOM open.NACT.PAST.PERFVE.3SG by the locksmith ‘The door was opened by the locksmith’ (passive type) b. *Ǿ ʌȩȡIJĮ ȐȞȠȚȟİ Įʌȩ IJȠȞ țȜİȚįĮȡȐ16 *i porta anikse apo ton klidhara the.NOM door.NOM open.ACT.PAST.PERFVE.3SG by the locksmith ‘*The door opened by the locksmith’ (anticausative type)

(ii) In contrast with the passives, the anticausatives cannot allow agent-oriented adverbs and allow control with sentences of purpose. (14) a. Ǿ ʌȩȡIJĮ ĮȞȠȓȤIJȘțİ İʌȓIJȘįİȢ i porta anixtike epitidhes the.NOM door.NOM open.NACT.PAST.PERFVE.3SG on-purpose ‘The door was opened on purpose’ (passive type) b. *Ǿ ʌȩȡIJĮ ȐȞȠȚȟİ İʌȓIJȘįİȢ *i porta anikse epitidhes the.NOM door.NOM open.ACT.PAST.PERFVE.3SG on-purpose ‘*The door opened on purpose’ (anticausative type) 16

Despoina Chila-Markopoulou (personal communication) notes that structures such as this are beginning to be heard; she explains their acceptance (albeit to a limited degree) due to interpretation of the by-PP as cause and not as agent.

12

Chapter One c. Ǿ ʌȩȡIJĮ i porta

ĮȞȠȓȤIJȘțİ anixtike

ȖȚĮ ȞĮ ȝʌȠȣȞ ja na bun

the door.NOM open.NACT.PAST.PERFVE.3SG for to

ȠȚ ĮıIJȣȞȠȝȚțȠȓ i astinomiki

ıIJȠ ıʌȓIJȚ sto spiti

enter.3PL the policemen.NOM into-the house

‘The door was opened so for the policemen can enter into the house’ (passive type) d. *Ǿ ʌȩȡIJĮ ȐȞȠȚȟİ (Įʌȩ ȝȩȞȘ IJȘȢ) ȖȚĮ ȞĮ ȕȖİȚ Ș ȂĮȡȓĮ Įʌȩ IJȠ ıʌȓIJȚ *i porta anikse (apo moni tis) ja na vji i Maria apo to spiti the door.NOM open.ACT.PAST.PERFVE.3SG (by itself) for to get-out.3SG the Maria.NOM from the house

‘*The door opened (by itself) so for Maria can get out from the house’ (anticausative type)

(iii) Every transitive verb can be passivised, but only one sub-class of transitive verbs (“alternating causatives”) has anticausative types. (15) a. Ǿ ȂĮȡȓĮ ȐȞȠȚȟİ IJȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ i Maria anikse to xrimatokivotio the.NOM Maria.NOM open.ACT.PAST.PERFVE.3SG the.ACC safe.ACC ‘Maria opened the safe’ (causative type) b. ȉȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ ĮȞȠȓȤIJȘțİ Įʌȩ IJȘ ȂĮȡȓĮ to xrimatokivotio anixtike apo ti Maria the.NOM safe.NOM open.NACT.PAST.PERFVE.3SG by the Maria ‘The safe was opened by Maria’ (passive type) c. ȉȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ ȐȞȠȚȟİ (Įʌȩ ȝȩȞȠ IJȠȣ) to xrimatokivotio anikse (apo mono tu) the.NOM safe.NOM open.ACT.PAST.PERFVE.3SG (by itself) ‘The safe opened by itself’ (anticausative type) (16) a. Ǿ ȂĮȡȓĮ ȑțȠȥİ IJȠ ȥȦȝȓ i Maria ekopse to psomi the.NOM Maria.NOM cut.ACT.PAST.PERFVE.3SG the.ACC bread.ACC ‘Maria cut the bread’ (transitive, non-causative type) b. ȉȠ ȥȦȝȓ țȩʌȘțİ Įʌȩ IJȘ ȂĮȡȓĮ to psomi kopike apo ti Maria the.NOM bread.NOM cut.NACT.PAST.PERFVE.3SG by the Maria ‘The bread was cut by Maria’ (passive type) c. *ȉȠ ȥȦȝȓ ȑțȠȥİ/ țȩʌȘțİ Įʌȩ ȝȩȞȠ IJȠȣ *to psomi ekopse/ kopike apo mono tu the.NOM bread.NOM cut.ACT.PAST.PERFVE.3SG/ cut.NACT.PAST.PERFVE.3SG by itself ‘*The bread cut by itself’ (*anticausative type)

(iv) Only anticausative types allow the addition of the PP ‘by itself’ that denotes the absence of an external argument. (17) a. *ȉȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ ĮȞȠȓȤIJȘțİ Įʌȩ ȝȩȞȠ IJȠȣ *to xrimatokivotio anixtike apo mono tu the.NOM safe.NOM open.NACT.PAST.PERFVE.3SG by itself ‘*The safe was opened by itself’ (passive type) b. ȉȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ ȐȞȠȚȟİ Įʌȩ ȝȩȞȠ IJȠȣ to xrimatokivotio anikse apo mono tu the.NOM safe.NOM open.ACT.PAST.PERFVE.3SG by itself ‘The safe opened by itself’ (anticausative type)

(v) The anticausatives, but not the passives, can be accompanied by a DP-‘dative’ (in Modern Greek: in the genitive morphological case) which has, inter alia, the interpretation of the unintentional causer17. (18) ȉȠȣ īȚȐȞȞȘ IJȠȣ IJİȜİȓȦıİ IJȠ tu Jani tu teliose to the.GEN Janis.GEN he.GEN.WEAK finish.ACT.PAST.PERFVE.3SG the.NOM ‘Janis’s ink finished up (to him)’ (anticausative + DP-‘dative’)

ȝİȜȐȞȚ melani ink.NOM

(vi) In certain languages, there is also a morphological distinction between the passive and the anticausative: (19) Russian (Kulikov 1999a): Passive: periphrastic 17

Alexiadou, A. 2006c: “On (anti-)causative alternations”. Handout, EALing 2006,

Introduction

13

a. Palka byla slomana wood.NȅȂ is.PAST break.PAST.PART ‘The wood was broken by Peter’ Anticausative: suffix -sja18 b. Palka slomala’sǯ (*Petrom) wood.NOM break.ANTIC.PAST Peter.ERG ‘The wood broke’

Petrom Peter.ERG

The morphological derivation of the transitivity alternations is not expressed uniformly across languages, but neither is it expressed uniformly across all of the verbs of the same language. Languages vary with respect to the structure that constitutes the basis of morphological derivation. Haspelmath (1993) presents (on the basis of a typological research of transitivity alternations in twenty languages) three cross-linguistic patterns of morphological derivation, with the criterion of the presence of additional morphology in the derived type: (a) Causative pattern Turkish i. dĦú-mek ‘anticausative - fall’ fall-INF ii. dĦú-Ħr-mek ‘causative - throw’ fall-CAUS-INF19 (b) Anticausative pattern Russian i. otkryt’sja ‘anticausative open’ otkryt-ANTICAUS ii. otkryt ‘causative open’ (c) Without any direction in the morphological derivation x With the same type (“labile”) Modern Greek i. ıʌȐȦ spao ‘anticausative break’ ii. ıʌȐȦ spao ‘causative break’ 18

19

The suffix -sja with the imperfective past tense only forms an anticausative structure (obligatory absence of the agent). With the perfective past tense, it can also have a passive interpretation (the presence of the agent is possible). In Turkish, the causative morpheme can be connected to verb stems of different categories: (a) Causative morpheme + intransitive verb: (1) Ayúe bardag-ı yer-e Ayúe.NOM glass-ACC floor-DAT ‘Ayúe threw the glass on the floor’

dĦú-Ħr-dĦ fall-CAUS-PAST

(b) Causative morpheme + unergative verb (indirect cause): (2) Ayúe Zeynep’e acıklı bir hikaye anlat-arak Ayúe.NOM Zeynep.DAT sad a story telling-GER ‘Ayúe (in) telling a sad story to Zeynep made her cry’

onu she

a÷la-t-tı cry.CAUS.PAST

(c) Causative morpheme + transitive verb (the patient in the dative case functions as a second agent and strongly participates in the action): (3) Zeynep berber-e saç-ı-nı kes-tir-di Zeynep.NOM barber-DAT hair-POSS-ACC cut-CAUS-PAST ‘Zeynep got the barber to cut her hair’ (Göksel 1993)

(d) Causative morpheme + causative verb: (4) a.

b.

Mehmet Hasan’ı öl-dür-dü Mehmet.NOM Hasan.ACC die-CAUS-PAST ‘Mehmet killed Hasan’ Polis Mehmed-e Hasan’ı öl-dür-t-tü police.NOM Mehmet-DAT Hasan.ACC die-CAUS-CAUS-PAST ‘The police made Mehmet kill Hasan’

The first causative morpheme is considered as being added at the lexical domain, whereas the second causative morpheme is added at the syntactic level.

14

Chapter One

x With the addition of a morpheme to both verbs (‘equipollent’) Japanese i. tok-eru ‘anticausative melt’ ii. toku-asu ‘causative melt’ x With substitution (‘suppletive’) Finnish i. kuolla ‘anticausative die’ ii. tappaa ‘causative kill’20 To summarise, in the last two sections, I attempted to present arguments for the differentiation of transitive verbs into classes on which our research will be based. I endeavoured to show that transitive verbs do not comprise a unified class and that causative verbs (verbs denoting change-of-state) are the prototypical transitive verbs. The causative verbs are, moreover, divided into alternating causative verbs (verbs with transitive and intransitive (with the patient as subject) use) and non-alternating causative verbs. According to the above, a diachronic study of transitivity must take into consideration verb classes and in relation to these must search for systemisation in the changes. In the following section, I will try to present not only the problems that such a diachronic study encounters but also the likely positive aspects that arise from a diachronic study of transitivity.

1.4 The Study of Transitivity and the Diachronic Approach 1.4.1 The problems and the contribution of diachronic study The problems that face a diachronic study of transitivity are no different from those that every diachronic study faces. The following main difficulties are indicated (Roberts & Roussou 2003; Moser 2005; Faarlund 2006; Lightfoot 2006a, b): (a) A main methodological problem of diachronic linguistic studies is the limited data of what has survived or not survived by chance. The data that we have at our disposal are not, in all instances, spontaneous oral speech; they mainly represent written speech, and often they are not close to oral speech. Special care is needed so that the data that are finally selected for the analysis of the grammar of every period are as close as possible to the everyday speech of that period. Movements such as Atticism, which supported a return to an earlier form of the language, may cause additional difficulties for any attempt to examine everyday language21. (b) The analyses that adhere to the perspective of Generative Grammar are based on the grammaticality and ungrammaticality of structures, i.e. on the linguistic knowledge of the native speaker, on linguistic competence. Consequently, it is obvious that the methodology of diachronic research is necessarily differentiated from synchronic research since the conclusions drawn for the grammar of earlier historical periods are derived solely from observations in relation to the data of linguistic performance. More specifically, in diachronic studies, we are obviously unable to carry out experiments in order to enlarge the evidence that we will be using in the analysis, to minimise the possible errors that are contained in the data or to check the accuracy of our hypotheses. On the other

20

We can distinguish two further possible instances: (a) morphological marking of a different kind on causatives and anticausatives (Winnebago); (b) use of the same marker (ambiguous marker) for both the causatives and the anticausatives (Korean). ĺWinnebago (Guerssel et al. 1985): The causative type bears a prefix (gi-), whereas the anticausative type bears the morpheme -re-. (1) a.

b.

21

Kununga naanksik-ra gi-shishshannan Kununga the-wood gi-broke(causative) ‘Kununga broke the wood’ Naanksik-ra shiish-re-ena the-wood broke(anticausative)-re ‘The wood broke’

ĺKorean Some Korean unaccusative verbs derive corresponding causative types with additional causative morphology (class I), whereas some causative verbs derive corresponding unaccusative types with additional anticausative morphology (class II). The morphemes -(h)i- and the allomorphs -li-, -si-, and -ki- are ambiguous between the causative and the anticausative (Volpe 2005a, b). On the other hand, some negative remarks (for example, by the Atticist Phrynichus) about structures that were innovative for the period contribute to the study of language of the specific period and its differences from that of earlier periods.

Introduction

15

hand, when we study living languages, we acquire information from experiments regarding the acceptability of sentences, usually by eliciting judgements concerning the acceptability of sentences. The difficulties concerning the choice of data can, however, be reduced if we propose two hypotheses (Adams 1987; Santorini 1989): (i) the past is as the present and the general principles that derive from the study of living languages in the present, likewise, hold good for the earlier stages of language; and (ii) with regard to simple sentences, if a specific type of sentence does not appear in a representative corpus of data, then it is not grammatically possible in the language. This hypothesis is, of course, problematic since the lack of the presence of a structure in a corpus of data can always be the result of non-grammatical contextual factors or even coincidence. Despite that, for structurally simple sentences, it is extremely likely that the absence of the structure from a large corpus of data also means its absence from the structure of the language. The positive aspects of diachronic study that make linguists try particularly hard to deal with the aforementioned difficulties and to find the most suitable solutions to the problems already referred to, are, therefore, evident: (a) Diachronic research can offer us the appropriate examples in relation to a grammatical phenomenon from many historical periods and make possible the connection between a specific grammatical phenomenon and others that remain the same or change in different historical periods. The field of analysis is widening increasingly. Characteristic grammatical phenomena that are difficult to explain and interpret in only one historical period are found in different historical periods of one language or many languages. (b) The diachronic remarks give us the ability to strengthen the analysis of grammatical systems and to locate the role of universal principles and the differences in the parameters of language. Comparison between different grammatical systems and conclusions for language changes provide evidence for the areas of language that change, for the kinds and causes of change, and also for the characteristics of language that remain unchanged and constitute stable characteristics of human linguistic ability. The data concerning language change are of particular importance for approaches such as Generative Grammar since, as Kiparsky (1982a) first pointed out, they provide a gateway into the form of linguistic ability. Diachronic data contain types of information that are absent from synchronic data and make a vital contribution to the examination of the general principles of language. Through an understanding of the process of change, we can also understand the principles underlying the organisation of language. Given that the characteristics of a language in a specific historical period are the result of the interaction between the general principles of language and specific language changes, knowledge of diachronic processes is important in the exploration of the role that diachronic and general linguistic factors play, and, therefore, for the suitable formation of linguistic theory.

1.4.2 Sources The historical periods of the Greek language that are examined in this study are: Homeric, Classical, Hellenistic-Roman, Medieval, and Modern Greek (the latter only concerning the changes in relation to Medieval Greek). The reference to indicative examples from the Post-Byzantine/Early Modern Greek period (16th-17th and 18th centuries) aims at the strengthening of conclusions with regard to the tendencies that have been tracked. Further problems concerned with the setting of the boundaries of the historical periods and sub-periods are outside the goals of this monograph and do not concern me here. Special emphasis is given to Classical Greek. The study of Classical Greek is extensive for two reasons: (a) the system of transitivity (syntactic behaviour as well as the markers of transitivity) of Classical Greek has not been examined systematically to date and (b) we observe that Classical Greek (which constitutes the starting point of the present study on verb transitivity) differs considerably from Modern Greek with regard to verb transitivity. The general aim of this work and the domain of grammar that is analysed define the first restriction regarding the sources that are used for diachronic examination: the present study seeks to make known the similarities and differences in possible structures (transitive and intransitive structures) in the different periods of the history of the Greek language. Consequently, the comparative examination of every historical period in relation to the preceding period (as far as innovative structures or the loss of structures is concerned) constitutes the principal part of this study. The analysis begins with the Classical Greek period and uses Homeric Greek as a basis only for comparison and individual remarks on specific verbs. The data on which the analysis was based are derived from traditional Greek historical grammars, studies, and dictionaries (I refer here mainly to Jannaris 1897/1968, Psaltes 1913/1974, Mayser

16

Chapter One

1934/1970, Schwyzer & Debrunner 1950/2002, Chantraine 1953, Humbert & Kourmoulis 1957/2002, Browning 1983/1991 – Liddell, Scott, Jones & McKenzie 1940, Kriaras 1969/2009). In addition, systematic indexing was done of representative texts of the Classical Greek, Hellenistic, Medieval and Early Modern Greek periods. The data are divided into four large historical periods, which for the present analysis correspond to four distinctive grammatical systems: Classical Greek, HellenisticRoman Koine Greek, Medieval Greek, and Modern Greek. This present study in its entirety does not aim at an exhaustive judgement of data but is a study that is more oriented towards modern theory of language change and the testing of highly specific structures and aspects of the linguistic system in different historical periods. In any case, regarding the specific topic of transitivity, it can be observed that there has been, in the past, a total lack of examination of diachronic data and, consequently, this study constitutes the first attempt in this direction. (a) Classical Greek (5th – 3rd century BC) An attempt is made to locate evidence of everyday speech (in contradistinction to learned elements) in specific types of texts. The choice of texts is based on the histories of the Greek language of Babiniotis (1985/2002), Horrocks (1997), Christidis (2001) and the analyses mainly by Dover (1987), ChilaMarkopoulou (1990-91, 1999), Kopidakis (1999), Manolessou (2000, 2005), and Moser (1988, 2005). For example, the medical treatments of Hippocrates, written in a non-literary style, contain many elements of everyday speech (Manolessou 2005: 7-10). Additional limitations concerning the choice of data from Classical Greek comprise the avoidance of the effect of metre, archaisms, and particular register features; the emphasis is, nevertheless, on Attic Greek, which constitutes the dominant form of Classical Greek and the main source, through the Hellenistic and the Medieval periods, of Modern Greek (Lopez-Eire 1997; Horrocks 1997: 33-37). Beyond the grammars and the special studies on Classical Greek (Schwyzer 1939, 1959; Goodwin 1889/1965), texts representative of every text type were systematically indexed: Herodotus, Thucydides, Plato, Xenophon, Aristophanes, Demosthenes, Lysias, Euripides, Sophocles, Hippocrates, and Aristotle. (b) Hellenistic-Roman period (3rd century BC – 4th century AD) The New Testament, the non-literary papyri, and authors of written Koine such as Polybius and Epictetus comprise the most important sources of the colloquial language22: the subsequent development of the language shows that all of the particularities of the language of the New Testament are precursors of the changes that emerged in the following centuries; the non-literary papyri are invaluable since they contain elements of the colloquial language. There are, moreover, ample inscriptions in Hellenistic Greek, such as many informal letters and documents written on papyri (Moser 2005: 22-24). Apart from the evidence that was brought to notice from the studies of Mayser (1934), Mandilaras (1973), and Gignac (1976/1981), from the histories of the Greek language by Jannaris (1987), Horrocks (1997), and Christidis (2001), and from the monographs of Humbert (1930) and Janse (1993), collections of papyri, the Translation of the Septuagint, the New Testament, and the Acts and Lives of the Saints were also systematically indexed23. (c) Medieval Greek Systematic indexing of verbs of the semantic classes that we are concerned with was done in representative texts of this period (according to remarks made in the histories of Greek and by researchers of this period)24: (a) early and middle Medieval period (5th-11th century): the vernacular texts are few and the collection of data is particularly difficult. The texts that were chosen are: chronicles (Malalas: 6th century, 22

23

24

Despite the high degree of uniformity in written Hellenistic Koine, the differences between the authors are worth noting; the language of Polybius, for example, is more artificial than the language of Epictetus, which is much nearer to the everyday spoken language. For this purpose, electronic data were utilised: (a) TLG: Thesaurus Linguae Graecae, version E (CD-Rom) 2000 (and online version: ), University of California, Irvine, (b) PHI #7 Greek Documentary Texts (CD-Rom) (1991-1996), Packard Humanities Institute, Los Altos, California. I follow here the argumentation and the practice of David Holton and Geoffrey Horrocks: “Grammar of Medieval Greek Project”, : “... These dates [between 1100 and 1700] are chosen because texts in the vernacular become available in significant quantity only in the 12th century, and, although there is no obvious point at which to locate the end of the "medieval" period, by the 18th century important cultural and political changes are afoot. The period 1100-1700 constitutes a coherent whole in terms of the development of the Greek vernacular. ... In certain cases, early medieval texts (5th-11th century) will be taken into account, mainly to illuminate points of historical evolution or the earliest dating of phenomena”.

Introduction

17

Paschalion: 7th century, Georgios Monachos: 9th century, Theophanes the Confessor: 9th century), religious texts (John Moschos: 6th century, Romanos Melodos), and Protobulgarian inscriptions (Beševliev 1963). (b) late Medieval period (Beck 1971/1988; Vitti 1971/1978; Politis 1978; Browning 1983/1991; Tonnet 1993; Horrocks 1997; Adrados 1999; Kopidakis 1999; Eideneier 1999; Manolessou 2005): Digenis Akritas (the Grottaferrata and Escorial versions), the Ptochoprodromic poems (12th century), the poems of Spaneas, Michail Glykas, and Spanos, the Chronicle of Moreas (13th century), the metrical Byzantine romances of the 13th-15th centuries, the chronicles of the later period (Schreiner 1975), the Cypriot chronicle of Leontios Machairas, and the translation of the Assizes of Jerusalem and Cyprus; post-Byzantine period (Dimaras 1948; Politis 1978; Mastrodimitris 2006): anthologies of historical, philosophical, fictional texts, love poetry (Katalogia/ Erotopaignia, 15th century), Erofili, The Sacrifice of Abraham, Erotokritos (16th-17th centuries), The Chronicle of Rhodes of Georgilas, demotic songs from the collections of ǿ. Legrand (1880-1913, 1885) and N. Politis, and formal documents (Miklosich & Müller 1865). Useful data were also found in secondary sources, such as grammars, linguistic commentaries on specific texts and authors and studies of certain grammatical phenomena (Sofianos (16th century); Portius 1638; Psaltes 1913/1974; Browning 1983/1991). (d) Modern Greek Research on transitivity in Modern Greek is based on a personal corpus of data (written texts from the Internet: newspaper articles, taped sessions of the Greek Parliament, blogs), observations (examples of spontaneous oral speech from natural dialogues and from the radio and television), judgements of native speakers, personal linguistic intuition, and data from Modern Greek grammars and studies (articles and monographs).

1.5 Terminology Issues The terms that were and are used in relation to the various types of transitive and intransitive verbs and structures depend on the theoretical perspective to which the analyses adhere. The widest variation with regard to the use of terms concerns the verbs that can be used in transitive and intransitive structures, often without any change in morphology, and with the meaning of changeof-state. (20) a. Ǿ țĮȡȑțȜĮ i karekla the.NOM chair.NOM ‘The chair broke’ b. ȅ ȀȫıIJĮȢ o Kostas the.NOM Kostas.NOM ‘Kostas broke the chair’

ȑıʌĮıİ espase break.ACT.PAST.PERFVE.3SG ȑıʌĮıİ IJȘȞ espase tin break.ACT.PAST.PERFVE.3SG the.ACC

țĮȡȑțȜĮ karekla chair.ACC

Levin (1993) and Levin & Rappaport Hovav (1995) talk of causative alternation25. The terms causative and anticausative26 are used for the transitive and the intransitive members of the alternation respectively (e.g., Antonopoulou 1987), and the term middle intransitive only for the intransitive use (Theophanopoulou-Kontou et al. 1998). Clairis & Babiniotis (1999: 283-290) distinguish between two 25

26

Particularly frequent, especially in earlier times, is the term ergative or ergative pairs in literature. This is the term used by Halliday (1964) and Lyons (1968): ‘the term for the syntactic relation that exists between (1) the stone moved and (2) John moved the stone is “ergative”. The subject of the intransitive verb becomes the object of the corresponding transitive verb, and a new ergative subject is introduced as the “agent” (or “cause”) of the action referred to’ (Lyons 1968: 352). There have been serious objections raised, however, about the use of that term because it was initially linked to the typological categorisation of languages in nominative-accusative vs. ergative-absolutive. Dixon (1987) expresses his opposition to extending the meaning of the term ‘ergative’, believing that it could cause confusion. For the history of (anti)causative terms, see Haspelmath (1987: 8ff). The term anticausative was taken from Nedjalkov & Sil’nickij (1969) and was used mainly by Russian linguists and those with knowledge of the typological group of Leningrad/St. Petersburg. The term anticausative was also used by: Siewierska (1984), Haspelmath (1987, 1990), Comrie (1989), and Cennamo (1993). Other terms used for the anticausative are: decausative (Geniušiene 1987), deagentive (Lehmann 1992b), spontaneous (Kemmer 1993; Shibatani 1985), and verbal derivation (Lehmann 1972a).

18

Chapter One

categories of verbs (based on the non-change in voice morphology): (a) verbs that have as their basic use the transitive (e.g., Kapios aniji kati / someone.NOM open.ACT.PRES.3SG something.ACC / ‘Someone opens something’ – Kati aniji / something.NOM open.ACT.PRES.3SG / ‘Something opens’) that are called ergatives, and (b) verbs that have as their basic use intransitive (Kapios thimoni / someone.NOM anger.ACT.PRES.3SG / ‘Someone gets angry’ – Kati thimoni kapion / something.NOM anger.ACT.PRES.3SG somebody.ACC / ‘Something angers someone’) that are called causatives. The aforementioned distinction is supported by the frequency of the appearance of the transitive or the intransitive type and the ability to appear in more environments (for example, the intransitive use of causative verbs is deemed to be the most basic, as it is more frequent and is possible in more environments: (21) a. īȑȡĮıİ ʌȡȚȞ IJȘȞ ȫȡĮ IJȠȣ jerase prin tin ora tu get-old.ACT.PAST.PERFVE.3SG before the time his ‘(S/he) got old before his time’ b. ȅȚ ıțȠIJȠȪȡİȢ IJȠȞ ȖȑȡĮıĮȞ ʌȡȚȞ IJȘȞ ȫȡĮ IJȠȣ i skotures ton jerasan prin tin ora tu the.NOM worries.NOM he.ACC.WEAK get-old.ACT.PAST.PERFVE.3PL before the time his ‘Worries made him old before his time’

Following the Generative Grammar approach, I have decided on the term causative for the transitive type of the verb and the term anticausative for the intransitive type. I will call the transitive/intransitive verb alternation transitivity alternation, and the verbs that participate in alternation alternating causatives [in contrast with the causative (with change-of-state meaning) verbs that do not participate in alternation and which I will call non-alternating causatives]. The term anticausative is not used as such by Haspelmath (1993), where it only refers to intransitive verbs that are derived morphologically from transitives; it is used in the sense of the intransitive verbs that denote change-of-state, with the obligatory absence of the agent. CAUSATIVE TYPE causative meaning: ‘X makes Y (which pre-exists of the action of X) to change state (affected Y)’ AGENT/CAUSE/INSTRUMENT + Verb + PATIENT (=DIRECT EFFECT)

ANTICAUSATIVE TYPE anticausative meaning: ‘Y changes state by itself; external agent does not exist, only the cause or instrument may exist’ PATIENT + Verb + CAUSE/INSTRUMENT (by-PP)

Terms that derive from the theory of Generative Grammar are used for the most part without showing disregard for the terms and perspective of other approaches27. Transitive types 27

Cognitive and Systemic-Functional terminology (Lemmens 1998) is particularly detailed (possibly more than it needs to be in relation to grammatically encoded distinctions). Lemmens’s terminology is given below (1998: 46); the corresponding terms which are used in Generative Grammar are displayed in brackets. John killed Mary Jill died He ran a good race John kills Stale bread cuts easily Jill drowned Jean Jean drowned The system burst a pipe The window opened only with great difficulty

transitive effective transitive non-effective transitive pseudo-effective transitive objectless transitive middle ergative effective ergative non-effective ergative pseudo-effective ergative middle

(causative, non-alternating) (unergative, non-alternating) (transitive) (null object) (middle) (causative) (anticausative, unaccusative) (transitive) (middle)

Introduction

19

Causative (prototypical transitive) type: (22)

ȅ īȚȐȞȞȘȢ ȐȞȠȚȟİ IJȘȞ ʌȩȡIJĮ o Janis anikse tin porta the.NOM Janis.NOM open.ACT.PAST.PERFVE.3SG the.ACC door.ACC ‘Janis opened the door’

Accusative (transitive non-causative) type: (23)

ȅ īȚȐȞȞȘȢ ȐȖȖȚȟİ IJȘȞ o Janis angikse tin the.NOM Janis.NOM touch.ACT.PAST.PERFVE.3SG the.ACC ‘Janis touched the door’

ʌȩȡIJĮ porta door.ACC

Intransitive types Anticausative type: (24) a. Ǿ ʌȩȡIJĮ ȐȞȠȚȟİ ȟĮijȞȚțȐ i porta anikse ksafnika the.NOM door.NOM open.ACT.PAST.PERFVE.3SG suddenly ‘The door opened suddenly’ b. е̊ȱȱ ΔϱΏ΍Ζȱȱ πΕ΋ΐΝΌΉϟ΋ȱȱ ȱ he: polis ere:mo:theie: the.NOM city.NOM empty.PASS.AOR.OPT.3SG ‘if the city emptied’

(Modern Greek)

(Ancient Greek)ȱ

Passive type: (25)

H įȒȜȦıȘ ȣʌȠȖȡȐijIJȘțİ i dhilosi ipoghrafike the.NOM statement.NOM sign.NACT.PAST.PERFVE.3SG ‘The statement was signed by all the participants’

Įʌȩ ȩȜȠȣȢ IJȠȣȢ ıȣȝȝİIJȑȤȠȞIJİȢ apo olus tus simetexondes by all the participants

Reflexive type: (26)

ȅ īȚȐȞȞȘȢ ȟȣȡȓıIJȘțİ IJȠ ʌȡȦȓ o Janis ksiristike to proi the.NOM Janis.NOM shave.NACT.PAST.PERFVE.3SG the morning ‘Janis shaved in the morning’

Reciprocal type: (27)

ȅ īȚȐȞȞȘȢ țĮȚ Ș ȂĮȡȓĮ ijȚȜȒșȘțĮȞ o Janis ke i Maria filithikan the.NOM Janis.NOM and the.NOM Maria.NOM kiss.NACT.PAST.PERFVE.3PL ‘Janis and Maria kissed each other’

Middle type: (28)

ǹȣIJȩ IJȠ ȕȚȕȜȓȠ įȚĮȕȐȗİIJĮȚ ʌȠȜȪ İȣȤȐȡȚıIJĮ afto to vivlio dhiavazete poli efxarista this.NOM the.NOM book.NOM read.NACT.PRES.3SG very pleasantly ‘This book reads very pleasantly’

Many different verb types, whose only common characteristic is the presence of the morphological accusative case (see chapter 2), are unified under the term transitive non-causative verbs. For this reason, I propose and will use the term “accusative verbs” for transitive non-causative verbs. The difference between the causative and the accusative verbs is semantic and syntactic: (a) the direct object of the causatives always bears the thematic (ș-) role of the patient, whereas the object of the noncausative transitives (‘accusative verbs’) can be theme or goal, (b) the causative verbs denote changeof-state of the argument of the direct object, while the non-causative (‘accusative’) verbs do not denote change-of-state (but many other and varied meanings), and (c) the direct object of the causative type can be used as the subject of the intransitive-anticausative type, always without the presence of the agent (only a cause or an instrument can be present in the intransitive structure), with or without change of voice morphology, whereas the direct object of the non-causative (‘accusative’) verbs can be used as the subject of the intransitive-passive structure, always with the possible presence of the agent, and always with change of voice morphology. There are more types of intransitive structures, i.e. structures that do not allow the presence of a direct object. In reflexive, reciprocal and middle structures, the verbs, moreover, bear different verb voice morphology from that of the corresponding transitive structures (ȟȣȡȓȗȦ ksirizo ‘to shavetransitive’; ȟȣȡȓȗȠȝĮȚ ksirizome ‘to shave-reflexive’; ijȚȜȐȦ filao ‘to kiss-transitive’; ijȚȜȚȩȝĮıIJİ filiomaste ‘to kiss each other-reciprocal’; įȚĮȕȐȗȦ dhiavazo ‘to read-transitive’; įȚĮȕȐȗİIJĮȚ dhiavazete ‘to be read/reads-passive or middle’). In reflexive structures, the agent and the patient refer to the same (animate) person, and in the reciprocal structures, there are two agents (in the syntactic position of the subject), whereas the middle structures have a generic reading (Lekakou 2005).

20

Chapter One

The first two classes of phych-verbs that have been recognised from the relevant literature are dealt with as anticausative and causative types, respectively (Pesetsky 1995: 11-55). Class I: experiencer-Nom + theme/cause-Acc (anticausatives) (29) ȅ īȚȐȞȞȘȢ ijȠȕȐIJĮȚ o Janis fovate the.NOM Janis.NOM fear.NACT.PRES.3SG ‘Janis fears of ingratitude’

IJȘȞ tin the.ACC

ĮȤĮȡȚıIJȓĮ axaristia ingratitude.ACC

The transitive ijȠȕȓȗȦ fovizo ‘frighten’ [ĺ ‘make afraid of’] belongs to Class II-causatives. Class II: theme ĺ cause-Nom + experiencer/patient-Acc (causatives) (30) Ǿ ĮȞİȡȖȓĮ ĮȞȘıȣȤİȓ i anerghia anisixi the.NOM unemployment.NOM worry.ACT.PRES.3SG ‘Unemployment worries the citizens’

IJȠȣȢ tus the.ACC

ʌȠȜȓIJİȢ polites citizens.ACC

ȉhe intransitive ĮȞȘıȣȤȫ anisixo ‘to worry’ [ȠȚ ʌȠȜȓIJİȢ ĮȞȘıȣȤȠȪȞ i polites anisixun ‘the citizens worry’] belongs to Class I-anticausatives28. Regarding the terminology for the voice morphology that is connected with argument structure and transitivity, there are also many differentiations. As far as Modern Greek is concerned (which is the subject of many systematic synchronic analyses), there is general agreement about the voice that is connected to transitive structures (active voice), but the multiple functions of the verb endings that are connected to intransitive structures have resulted in uncertainty concerning the terminology in the literature: passive morpheme (Joseph & Philippaki-Warburton 1987; Holton et al. 1997; Tsimpli 1989), non-active morphology (Rivero 1990, 1992; Alexiadou & Anagnostopoulou 1999a, 1999b, 2004; Embick 1998, 2004; Tsimpli 2006a), middle-passive morphology (Theophanopoulou-Kontou 1980, 1982, 1983-4, 2004). I have chosen the terms non-active form and non-active morphology assuming that the essence of the present analysis is its distinction from the active form, which is used in transitive structures29, and that this term can cover all of the functions and syntactic behaviours and, mainly, all of the changes of this specific form in the diachrony of Greek.

1.6 The Structure of the Monograph In chapter 1 (Introduction), I described the subject and the goals of the study. This monograph investigates the diachrony of verb transitivity in Greek and, more specifically, the changes in causative verbs (transitive verbs that denote change-of-state) and in transitivity alternations. I follow, on the one hand, the approach of Tsunoda (1981, 1985, 1999), who modified the classical study of transitivity by Hopper & Thompson (1980) and, on the other hand, the analyses of Levin & Rappaport Hovav (1995, 1998, 2001, 2005) and distinguish the prototypical transitive verbs, i.e. the causative verbs, from the remaining transitives. Of particular interest to our research is the class of causative verbs that can participate in both transitive and intransitive structures (with the patient argument in the position of the subject), with or without change in voice morphology. Finally, in chapter 1, I analysed the difficulties and the benefits gained from the diachronic study of transitivity and concluded with the data sources and terminology issues. Chapter 2 lays out the theoretical framework that is adopted for the examination of transitivity. I maintain that the syntactic Minimalist approach (Chomsky 1995, 2000, 2005/2008) and the perspective of Levin & Rappaport Hovav (1995, 1998, 2005, 2006a) concerning semantic decomposition, the presence of verb classes, and correlation between lexicon and syntax (lexicon-syntax interface) can help in the interpretation of changes in transitivity. Emphasis is given to the central role of transitivity in syntactic analysis and the corresponding syntactic heads that are proposed to show the centrality of 28

Class III: theme-Nom + experiencer-Dat (1)

29

Ǿ ȩʌİȡĮ i opera the.NOM opera.NOM ‘Janis likes opera’

ĮȡȑıİȚ aresi like.ACT.PRES.3SG

ıIJȠȞ īȚȐȞȞȘ ston Jani to Janis

I will name ‘non-active form’ (without further distinctions) the verb form that differs from the active form of transitive structures. Non-active morphology is identified in Greek with the change of verb endings.

Introduction

21

transitivity in clause structure. The verbs that participate in transitivity alternations are analysed, and I attempt to expose the correlation between the syntactic structure and the semantic parts that arise from the decomposition of verb meaning. Different verb classes [causative alternating or non-alternating verbs, “accusative verbs” (=transitive non-causative verbs30), anticausative non-alternating, and unergative verbs] are derived from the interaction of the sub-predicates and the ș-roles. In the second part of the second chapter, the theoretical positions adopted in this monograph in relation to language change are developed. Lightfoot’s model (1979, 1999, 2006a) on diachronic change is outlined and the main areas of the approach to the topic under examination are analysed. Language change is correlated with language acquisition and is sought in the Internal language (I-language). The changes arise on the basis of linguistic data that children, when acquiring language, receive from their environment, and mainly from a reanalysis of those data to remove ambiguity successfully. The reanalysis of the structures is based on linguistic evidence that is to be found within the abstract syntactic structures and not in real sentences from data found in the environment. The third chapter is dedicated to the classification and analysis of diachronic data according to the theoretical principles of Generative Grammar. The analysis in each section is divided into two parts: (a) causativisations and (b) the role of voice morphology. In section 3.1, the transitivity in Ancient Greek is explored. In relation to the theoretical perspective, the data concerning Ancient Greek are classified according to the semantic verb classes. From the study of Classical Greek texts, it emerges that the process of causativisation was already highly systematic in the Classical period. Causativisation (emergence of a transitive type on the basis of an earlier non-alternating anticausative type) concerned all three classes (in relation to voice morphology) of anticausative verbs (types with active morphology, types with non-active morphology, and types with parallel use of active and non-active voice morphology) and a large number of verbs that denote change-of-state, chiefly from an external cause. The possible causes of change are sought in the presence of cognate direct objects with intransitive verbs and in the presence of prefixes. Finally, the different syntactic behaviour of prototypical transitive (causative) verbs in relation to the (‘accusative’) transitive verbs is presented. The causative nonalternating verbs have stable syntactic behaviour (they participate only in transitive structures and consistently require direct objects only in the accusative case), while the ‘accusative verbs’ exhibit instability in argument structure (change from ‘accusative verbs’ to intransitive verbs and vice versa) and in their case requirements. Regarding voice morphology, in Homeric Greek, the relationship between the causative and anticausative types is realised by vowel alternation (ablaut alternation), whereas the non-active voice is not obligatory for the passive structure; in Classical Greek, vowel alternation has been abandoned and the relationship between causative and anticausative type is marked through the contrast between active and non-active voice morphology. In section 3.2, the characteristics of the verb system and the changes during the Hellenistic-Roman period are analysed. The study follows the principles of the comparative linguistic approach (analysis of micro-variation) and inquires into the differences between the causative verbs of the period in relation to the corresponding causative verbs of Ancient Greek. The changes resulting in the presence of new transitivity alternations are examined. The picture presented by the process of causativisations remains steady. The causativisation of intransitive forms (active, non-active, and with parallel use of active and non-active voice) is especially productive. Causativisation is not systematic in the case of intransitives that denote change-of-state from an internal cause. Changes in the other direction (anticausative types from transitives) are not systematically observed. The situation is not the same with ‘accusative verbs’, where a higher degree of instability in relation to argument structure (‘accusative verb’ from intransitive verb and vice versa) and their case requirements can be observed (parallel use with the accusative case of other cases and PPs, spread of the accusative case to all verbs of this verb class). Finally, in relation to voice, the general tendency shows the extension of the active voice in the anticausative types and, consequently, of the marking with the same voice (active) of both the causative and anticausative types. This tendency is linked to the first appearances of the passive type (presence of the agent in the by-PP) of anticausative verbs and to a general rearrangement in the area of voice. In section 3.3, data from the Medieval and Early Modern Greek periods are presented. The classification of data follows the structure of the analysis of the Ancient Greek and Hellenistic-Roman periods. Causativisations and the factors that favoured them in this period are classified and studied. The distribution of the changes echoes our initial hypothesis that the changes in voice do not codify the changes and the direction of (anti)causativisation (=causative/anticausative extension). The changes in the structure of causative verbs can also be differentiated from the changes in the accusative verbs. The 30

I use the term “accusative verbs” for the non-causative transitive verbs on the basis of their semantic and syntactic behaviour and the diachronic tendency of their direct objects to change their morphological case into the accusative case.

22

Chapter One

changes in the accusative verbs are chiefly changes in morphological case, whereas the changes in causative verbs are changes in argument structure. With regard to the latter, I try to show that a dominant role is played by the presence of a clitic as the direct object of an earlier intransitive verb. The presence of a clitic comprises the first phase of the causativisation of intransitive verbs. On the other hand, the impossibility of causativising many intransitive verbs in Greek is connected with the possibility of the presence of DP in the accusative denoting cause (parallel use of accusative case and PP for the realisation of cause). Regarding voice, instability can be observed in this period, in the voice of both transitives and intransitives; this may be connected to the completion of changes for the absorption of the accusative case by the non-active form: the non-active voice is no longer connected to other arguments and their absorption (benefactive, goal, theme), but only with the accusative case, which the non-active voice blocks from being assigned (*non-active type + direct object in the accusative case). I argue that this change was instigated by the loss of the meaning of personal interest in the non-active transitives (“benefactive meaning”/middle voice reading) and by the spread of the accusative case as the default case morphology for direct objects. In turn, this change led to the extension of active anticausatives. Finally, the synchrony of the system of transitivity in Modern Greek is presented, but only as a basis for comparison with the corresponding system of Medieval Greek. In section 3.4, an analysis of diachronic data from the English language and some Romance languages is undertaken according to the principles that I used for Greek diachronic data. Likewise, I also distinguish voice morphology from the process of causativisations and from the changes in argument structure, and I present the areas of instability in the history of languages that express the passive and the anticausative types with different morphology. The cross-linguistic data are concerned mainly with English, for which there is a wealth of diachronic data in relation to the changes in argument structure, and with the Romance languages, because the information we have on the developments of voice in Latin and the Romance languages is detailed. I argue that the dual distinction of (a) causative and (b) ‘accusative verbs’ is particularly important and has an impact on verb syntactic changes: the changes in causative verbs are changes in argument structure, whereas the changes in ‘accusative verbs’ are mainly changes in the morphological case of their direct object. The dissociation of changes in argument structure from changes in voice morphology is also supported by diachronic development; the changes in voice morphology do not reflect the direction of changes in argument structure. In chapter 4, I move on to a general analysis and interpretation of the mechanisms of change in the area of transitivity and voice morphology, based on diachronic material presented in the previous chapter. In the first part of chapter 4, the causes of the changes in the argument structure are analysed. In all of the historical periods of Greek and English, the process of causativisation has been particularly productive. The change in the opposite direction (change to anticausative from non-alternating transitive) has not been at all productive. I cite significative examples that favoured reanalysis, and search for the abstract cues in the syntactic structure that led to the specific changes. The new structures are the result of the reanalysis of intransitive structures, mainly in sentences with the presence of a neuter DP that bears the same morphological ending in the nominative and accusative cases. I follow the approach of explanation through cues (cue-based approach; Lightfoot 1999, 2006a) and suggest that in structure reanalysis, the case markers comprise cues for language acquisition and consequently for language change. The instability in the voice morphology of anticausatives is not connected with the direction of changes in argument structure. The change that is observed in relation to voice morphology of causatives and anticausatives is the change from non-active anticausatives to active anticausatives. This change is mainly connected with the absorption of the accusative case by the non-active voice and the consequent presence of the Voice head [+agent] in the structure of passive types (in Ancient Greek, the presence of the construction “non-active verb + DP-accusative” was permitted). The change in voice morphology was dictated by a general linguistic principle, which requires dissociation between anticausatives and passives. The steady marking of passives with the non-active voice (the same as anticausatives) from the end of the classical Ancient Greek period and mainly from the HellenisticRoman period leads to changes in the voice morphology of anticausatives. This general principle is dictated by reasons of economy; according to those, the dissociation between the passive structure and the anticausative structure (passive: presence of Voice [+agent], non-active morphology; anticausative: absence of Voice, active morphology) is preferred to the presence (in both structures) of Voice with two sub-types (passive: Voice [+agent], non-active morphology; anticausative: Voice [-agent], non-active morphology). According to the findings of the present research and the new data that have come to light from diachronic study, I conclude by proposing the analysis of causative lexical semantic representation for all alternating verbs. In chapter 5, a synopsis of the results of the research is given, and I draw general theoretical conclusions concerned with the nature of transitivity and language change. The general issues of causation and graduality of change are raised: why language change appears as it does and why specific

Introduction

23

language change takes place instead of another possible change (e.g., causativisation instead of anticausativisation or the extension of active anticausatives instead of the extension of non-active anticausatives). If we assume that the linguistic system is internalised, then diachronic changes must also be associated with the internalised system. The spread of the changes (which take place over many generations; e.g., the causativisations for certain verbs remain for many centuries at the first stage when the verb can take as a complement only a clitic; or the extension of the active anticausative, which begins in the Early Medieval period and is not complete even today) has to be attributed to language use (External-Language). According to the aforementioned approach, we should distinguish the time during which the changes in the linguistic system occurred from the spread of changes in the entire linguistic community.

CHAPTER TWO THEORETICAL FRAMEWORK

2.1 Transitivity, Syntax, and Lexicon-Syntax Interface In this chapter, the general principles that govern the present research (a) for the analysis of every different synchrony (section 2.1) and (b) for the analysis of diachronic changes (section 2.2) are presented. I follow Chomsky’s approach (1995, 2000) and I assume that language is analysed as a result of the interference of the phonological and the lexical structure with the articulatory-perceptual and the conceptual-intentional systems, respectively. According to the Principles and Parameters Approach and the recent analyses of the Minimalist Programme, language embodies two main areas: the lexicon and the computational system. Grammar is a computational system which, based on choices from the lexicon, forms structural derivations which are interpreted by the articulatory-perceptual and conceptual-intentional systems. The lexicon determines the lexical items that participate in the derivation, whereas the computational system determines the derivation of the structure. The principal stage in all processes is the spell-out. After the spell-out, access to the lexicon is no longer possible (Chomsky 1995, 2001; cf., among many others, Theophanopoulou-Kontou 2002a; Haegeman 2006). The derivational and inflectional features of the lexical items belong to different parts of grammar: derivational morphology belongs to the lexicon, whereas inflectional morphology is made up of processes such as checking of Tense and Agreement features which occur in syntax. The lexical items also comprise the representation of the phonological form, the determination of the syntactic category and the semantic characteristics. There are two types of lexical items: lexical categories concerned with content words, such as Determiners/Nouns, Verbs, Adjectives, Adverbs, Prepositions, which are heads of Phrases (D(eterminer)P(hrase)/N(oun)P(hrase), V(erb)P(hrase), P(repositional)P(hrase), and grammatical categories (C(omplement)P(hrase), Agr(eement)(P)hrase, T(ense)P(hrase)) which play an important role in syntactic processes, such as Tense, Agreement, Feature (Case) checking and Specificity (Chomsky 1995). I assume that it is possible to predict to a high degree the syntactic representation of the arguments from verb meaning. I follow the analysis of Levin & Rappaport Hovav (1995, 2005, 2006a) which comprises a ‘bridging attempt’ between syntactic analysis and the lexical-semantic study of verb meaning. Consequently, this approach attempts to connect the pure syntactic analyses, which show generalisations concerned with argument structure without taking into consideration meaning, with the lexical-semantic analyses, which deal with typical semantic representations without raising syntactic issues that arise from these representations1. The recent approaches concerned with the lexicon-syntax interface agree that specific aspects of verb meaning are related to verb syntax and have syntactic consequences; for example, they define the means of mapping between the arguments and their syntactic positions. One example is the distinction between unaccusative and unergative intransitive verbs; this distinction is defined semantically but is represented syntactically (Levin & Rappaport Hovav 1995). In the present research, I assume that the correlation between the event type (semantic decomposition) of the predicate and the argument structure defines, to a large degree, the transitivity of the verb2. 1

2

The proposed analyses that are concerned with the correlation between semantics and syntax and which I attempt to connect by modifying the approach of Levin & Rappaport Hovav belong to two general analyses: (a) highly detailed theory of verb meaning (Dowty 1991; Jackendoff 1990; Pinker 1989) and (b) pure syntactic analysis of the lexicon (resulting in syntactic rules both for the lexicon and for the syntactic level – Emonds 1991; Hale & Keyser 1992, 1993, 2002). Moving in the opposite direction are approaches that accept the principal role of ș-roles and the hierarchy of șroles in syntactic representation (systematic correspondence of ș-roles and syntactic positions; Fillmore 1968; Perlmutter & Postal 1984; Baker 1988). For the problems in the adoption of a one-to-one correspondence of șroles and syntactic position, cf. Jackendoff (1987, 1990), Rappaport Hovav & Levin (1988), Pinker (1989), Dowty (1989), Emonds (1991), Levin & Rappaport Hovav (1994, 1995). The majority of modern studies argue that semantic features are not directly visible in syntax. Following Zubizarreta (1987), Rappaport & Levin (1988) and Grimshaw (1991), I assume that the positions in syntactic structure are not directly connected with

26

Chapter Two

To give an idea of the reasoning that follows, I assume that the Minimalist Approach for the syntactic level and for v as the bearer of causativity and agentivity, combined with a modified version of the Levin & Rappaport Hovav approach (1995, 1999, 2005) for the mapping between the semantic and the syntactic structure, can successfully create a balance between the needs of a system that is constrained by syntactic structures that govern the lexicon and syntax in the same way, and a detailed system of semantic (typical) representations. I assume that every verb is connected to two representations: the syntactic representation that encodes the syntactic features of a verb concerning the arguments it takes, and the lexical semantic representation (lexical conceptual structure; Hale & Keyser 1986, 1987, 2002; Jackendoff 1990), which encodes the aspects of verb meaning that are related to the syntax. I follow Pinker (1984, 1989) and assume that there is a sub-system of semantic structures that is part of the linguistic mechanism and is not found outside the linguistic mechanism, as suggested by Jackendoff; it is precisely for this reason that I assume that mapping and relations exist between the semantic (event) and the syntactic structures. The combination of lexical-semantic representation and syntactic representation that I follow was first ascertained by Rappaport & Levin (1988): it was argued that fundamental to the correlation between syntax and lexicon is the distinction between external and internal argument (Williams 1981). The external argument is expressed in syntax as external to vP in the [Spec, vP] position (Chomsky 1995, 2000). The internal argument is generated in the [Compl, VP] position. The aforementioned analyses and their synthesis make up the working hypothesis of this diachronic study: I will attempt to reveal the aspects of syntactic representation and lexicon-syntax interface that are concerned with and determine verb transitivity in each chronological period3.

2.1.1 Transitivity and syntax The goal of this section is the examination of the role of transitivity in the clause structure. I will present the syntactic representation of transitivity and the modern analyses that comprise the basis (and the working hypothesis) of the study of the various chronological periods of the Greek language. 2.1.1.1 Principles and concepts In the analysis of every chronological period and in the entire diachronic interpretation I adhere to the following general principles, which have their origin in the Government and Binding Approach and have been adopted (some with modifications, to which I will refer only when they are concerned with the present subject) by the Minimalist Programme: (a) Case theory It is generally accepted that every lexically expressed NP must be able to receive case. This principle is known as the Case Filter (in the Government and Binding approach): *NP if the NP is lexical and lacks a case feature The Case Filter functions in the surface structure of the sentence, after rules of movement (if any) have come first (Chomsky 1981: 49; cf., among many others, Haegeman 1991: 188-189; TheophanopoulouKontou 2002a: 104-122). Cases characterise different positions of the sentence; every NP receives the appropriate case marker according to the position it occupies in the sentence. Chomsky (1981: 170) has formulated the following general principles that define the case marker of an NP: (a) NP is nominative if governed by the Agr (b) NP is objective if governed by V with subcategorisation feature: +[-NP] (c) NP is oblique if governed by PP (d) NP is genitive in [NP-ȋ] (e) NP is inherently Case-marked as determined by the properties of its [-N] governor

3

ș-roles since the argument structure is pure syntactic representation. Reinhart, Embick, Papangeli, and Zombolou (for Modern Greek) have also concerned themselves systematically with argument structures within the Generative Grammar approach. As will be shown by the reasoning that follows, the analyses of the aforementioned linguists do not cover the issue of transitivity in their diachronic dimension. Moreover, for Reinhart and Papangeli, the ș-roles occupy a central position and not the event schemas, while Embick analyses argument structures based on the theory of Distributed Morphology, and Zombolou is concerned with semantic structures but not with syntactic ones.

Theoretical Framework

27

The most important point in case theory in the Government and Binding approach, which is also maintained in the Minimalist Programme, is not so much the interpretation of the empirical data of the morphological marking of case as the fact that through this principle, a number of mutually connected phenomena that are derived from case are examined together (Bobaljik & Wurmbrand 2008). It is, therefore, possible to make, through case, a unified interpretation for the promotion of the subject in passives (1a), anticausatives (1b), and dependent clauses (1c): (1)

a. ȉȠ țȜİȚįȓ ȕȡȑșȘțİ t to klidhi vrethike t the.NOM key.NOM find.NACT.PAST.PERFVE.3SG ‘The key was found by Janis’ b. ȉȠ ȜİȦijȠȡİȓȠ ȑijIJĮıİ to leoforio eftase the.NOM bus.NOM arrive.ACT.PAST.PERFVE.3SG ‘The bus arrived’ c. Ǿ ȂĮȡȓĮ ijĮȓȞİIJĮȚ ȞĮ ıȣȝʌĮșİȓ t i Maria fenete na simbathi t the.NOM Maria.NOM seem.3SG to like.3SG ‘Maria seems to like Janis’

(Įʌȩ IJȠȞ īȚȐȞȞȘ) (apo ton Jani) (by the Janis) t t IJȠȞ ton the.ACC

īȚȐȞȞȘ Jani John.ACC

The aim of the Minimalist Programme is the unification of nominative and accusative case assignment/checking: the hypothesis that supports something like this is the hypothesis that the assignment of all cases –the nominative of the subject is also included– involves a Spec-Head relationship and is determined by the locality constraint (Chomsky (2000) calls this relationship ‘agree’). This hypothesis is confirmed if we accept that all subjects move into the final position in which they appear from lower positions in the clause structure. An additional factor that strengthens the unified case assignment (or checking) is the view that external [of the VP] functional projections are responsible for the case of the objects (Chomsky 1991). (b) Accusative case and absorption I will adopt the position (Theophanopoulou-Kontou 2002a: 118; Catsimali 1990) that the accusative case of the direct object in Modern Greek is a structural case and not a lexical one. This means that it is determined by the position of the DP in relation to the Verb or the Preposition that governs it. Examples of ‘raising’ of the subject of the dependent clause argue in favour of the structural character of the accusative case: (2)

a. ȅ īȚȐȞȞȘȢ ʌİȡȚȝȑȞİȚ IJȘȞ ȐȞȠȚȟȘ ȞĮ ijIJȐıİȚ İʌȚIJȑȜȠȣȢ o Janis perimeni tin aniksi na ftasi epitelus the.NOM Janis.NOM wait.3SG the.ACC spring.ACC to arrive.3SG finally ‘Janis is waiting for the spring to finally arrive’ b. ǵȜȠȚ ȣʌȠȜȠȖȓȗȠȣȞ IJȘȞ ĮȞIJȓįȡĮıȘ IJȦȞ ijȠȚIJȘIJȫȞ ȞĮ İȓȞĮȚ ȝİȖȐȜȘ oli ipolojizun tin andidrasi ton fititon na ine meghali all.NOM expect.3PL the.ACC reaction.ACC the.GEN students.GEN to be big.NOM ‘Everybody expects that the students’ reaction will be strong’

The DP in the accusative belongs to the dependent clause: (3)

a. ȅ īȚȐȞȞȘȢ ʌİȡȚȝȑȞİȚ ȞĮ ijIJȐıİȚ Ș ȐȞȠȚȟȘ İʌȚIJȑȜȠȣȢ o Janis perimeni na ftasi i aniksi epitelus the.NOM Janis.NOM wait.3SG to arrive.3SG the.NOM spring.NOM finally ‘Janis is waiting for the spring to finally arrive’ b. ǵȜȠȚ ȣʌȠȜȠȖȓȗȠȣȞ ȞĮ İȓȞĮȚ ȝİȖȐȜȘ Ș ĮȞIJȓįȡĮıȘ IJȦȞ ijȠȚIJȘIJȫȞ oli ipolojizun na ine meghali i andidrasi ton fititon all.NOM expect.3PL to be big.NOM the.NOM reaction.NOM the.GENstudents.GEN ‘everybody expects the students’ reaction to be strong’

The relationship of unavailability of the presence of a DP in the accusative (absorption or blocking of the accusative case) with the non-active voice was recognised theoretically early on. The non-active type according to the Principles and Parameter approach absorbs (Chomsky 1982: 124; Baker 1988; Jaeggli 1986a; Roberts 1987a; Baker, Johnson & Roberts 1989; Afarli 1989):

28

Chapter Two

(a) the accusative case, (b) the external ș-role. The features of the accusative case and of the external ș-role are joined according to the hypothesis that the non-active ending is an argument to which the accusative case and the external ș-role are assigned. The non-active morpheme absorbs the accusative case and the external ș-role, and the DP-object cannot receive case in the [Spec, VP] position (Visibility Condition on Case assignment) and is raised to the [Spec, IP] (or [Spec, TP]) position. Since the ș-role of the external argument has been absorbed by the non-active ending, the subject of the corresponding transitive types (that bears the external șrole in the corresponding transitive structure) does not need to move to [Spec, IP] to receive case and, therefore, leaves empty the [Spec, ǿP] position allowing the direct object to move there. (4)

IP  DP Iǯ   D N I VP the book  VP PP   V P DP  by John V -en

Similarly, Jaeggli (1986a: 590) interprets the assignment of ș-roles as follows: (a) the non-active morpheme absorbs the external ș-role of the verb; (b) the non-active morpheme assigns the external ș-role to the ‘by’-phrase; (c) the ș-role which has been assigned to the PP percolates to the preposition ‘by’; the preposition ‘by’ assigns the external ș-role to the DP-complement of the Preposition. To the non-active ending of Modern Greek was given the role of the absorption of the accusative case by Tsimpli (1989) and Rivero (1990). Tsimpli (1989) argues that the suffix that denotes voice in Modern Greek is generated in Inflection, and that it absorbs the accusative case and the ș-role: (5) IP  I V  non-active AGR morphology Rivero (1990), following Pollock (1989b), assumes that the Voice is a functional head. The non-active verb ending absorbs the accusative case so that the empty category that remains from the incorporation of the object-anaphor (in cases of non-active reflexives such as ĮȣIJȠțĮIJĮıIJȡȐijȘțİ aftokatastrafike selfdestroy.NACT.PAST.PERFVE.3SG ‘s/he destroyed her/himself’) is not characterised as an empty category (pro). The non-active morphology acquires, in addition, the external ș-role of the verb, whereas the internal argument moves to the subject position4. (6) VoiceP  non-active VP morphology The main problem with the analyses that accept absorption of both the accusative case and the ș-

4

Papangeli (2004) contradistinguishes the absorption only of the accusative case by the non-active ending of Modern Greek from the absorption of dative and accusative by the clitic (for example, in Modern Italian). For Papangeli, the non-active morphology is not in a functional projection (v) but is a derivational (not inflectional) suffix in V, and prevents the verb feature of the accusative from being visible in syntax and from being checked by the DP.

Theoretical Framework

29

role of the external argument is that the agent-DP in the ‘by’-Phrase in the non-active passive structure receives the ș-role in a totally different way from the way in which the external ș-role is assigned to the same DP in the active transitive structure. For this reason, I will follow Collins’s (2005) analysis and hypothesise that the non-active type in Modern Greek blocks the presence of the accusative case but does not absorb the ș-role of the agent that can appear in the ‘by’-Phrase (for more on the adoption of Collins’s approach, see sections 2.1.2.2 and 4.2.2). 2.1.1.2 Extended VP (v and Voice) The majority of traditional definitions of transitivity emphasise the completion of a transitive verb with a direct object, in contrast with the absence of a direct object for the intransitive verb. It was considered that the fundamental morphological and syntactic criteria for the direct object were the different morphological markings of the subject and the object and passivisation (again, a link to voice is obvious here)5. The traditional groups of transitive and intransitive verbs were accepted by Transformational (and Generative) Grammar, which connected transitivity with transitive predicates that express relations between two arguments, in contrast with intransitive predicates that are concerned only with one argument. A result of the aforementioned was that transitivity was initially connected with passivisation and was dealt with as such over a long period by modern Linguistics. Transitivity was inseparably connected with the presence of a direct object, which can also comprise the subject of a corresponding passive structure (see the transformations that connect transitive and passive structure claimed by Zellig Harris (1956), and the common deep structure of transitive and passive structures by Noam Chomsky in Syntactic Structures (1957)). Chomsky (1965), in the same way, dealt with the relations between transitive-passive structure and transitivity-passivisation: transitivity remains totally connected to the presence of a direct object6, but the verbs now are marked in the lexicon as transitive or intransitive. In the Government and Binding approach, transitive and intransitive verbs have separate entries in the lexicon since they possess different properties of governing nouns: the transitive (in contrast to the intransitive) verbs assign case to their objects and ș-role to their subjects. Only over the last few years, in the analyses of Generative Grammar, has the need for a more highly detailed analytical examination of the phenomenon of transitivity been indicated through the promotion of the fundamental role that transitivity plays in grammar (Bowers 1997, 2002: transitivity comprises a separate functional category that cannot be expressed by other mechanisms of the Computational System). Many linguists have raised the issue of the presence of an additional head over the VP, which does not correlate with the inflectional features of the verb but with the arguments of the VP. This head introduces (or not) an external argument and subsequently defines the transitivity of the predicate and contains the relative agreement and case features. (a) v All the recent approaches of argument structure under the prism of the Minimalist Programme agree that a functional category between the functional category of Inflection and of the Verb exists with the following properties: (a) its Spec introduces the subject, (b) its head contains the agreement features with the object. The presence of a head that is connected with the arguments of the verb was proposed early on. Marantz in 1984 had already proposed that the external argument is not selected lexically by the verb head. Chomsky’s (1995) v, Kratzer’s (1996) Voice head, Borer’s (1996, 1998a, 1998b) ASP-or(iginator) are several examples of heads that are connected with the external argument, either with its introduction (Chomsky, Kratzer), or with the determination of the external argument as a pre-selected argument (Borer). In parallel, Johnson’s (1994) m heads, Chomsky’s (1995) AgrO, Borer’s (1996, 1998a, 1998b) ASP-e(vent) and Kratzer’s (2000) Telicity are heads that are connected to the internal 5

6

A counter-argument concerning these approaches is that the definition of passivisation has, many times, also been based on transitivity (a verb that has transitive characteristics can passivise), which thus leads to circularity (Siewierska 1984). Transitivity is also connected in Relational Grammar to the direct object, which, together with the subject, completes the transitive verb. A transitive verb is a verb that has both a subject and a direct object, whereas an intransitive verb is a verb that has neither the one nor the other complement (Perlmutter & Postal 1977, 1983; Perlmutter 1978). The subject and the direct object are defined in relation to word order, morphological marking, and the syntactic and semantic characteristics of the verb and the NPs.

30

Chapter Two

arguments of the verb. Chomsky (1993, 1995), following the split INFL hypothesis of Pollock (1989b), proposes the presence of the AgrS and AgrO heads higher than VP; these two heads are separated from the Tense head and play the role of checking the syntactic features of the subject and of the object, respectively. Later, Chomsky (1995), in relation to a study by Johnson (1991) on ditransitive verbs, proposed the insertion of the subject by a separate head, v, whose presence comprises a combination between the absence of the subject from the VP and its derivation lower than the position in which it appears in the spell-out. It is worth noting that the Agr Phrases are always present in the syntactic structure, whether their presence is essential for case or not, according to this approach7. (7) AgrSP  AgrS TP  T AgrOP  AgrO vP  DPsubj vǯ  v VP  V DPobj Bowers (2002) assumes that an optional functional category of Transitivity exists, which is subcategorised by the functional category of the Predication (as he calls v). The functional category of Transitivity may contain ij-features and assign accusative case, while it may also include the Extended Projection Principle (EPP) feature. The DPs that are marked with accusative case move to the [Spec, TransitivityP] position. In English, there is no overt realisation of Transitivity, whereas in Greek the morphological accusative case realises Transitivity. According to the aforementioned, the external argument of transitive verbs is found in the [Spec, PredicationP] position, which selects the Transitivity, whereas the external argument of unergative verbs is also found in the [Spec, PredicationP] position, which, however, selects a VP. The unaccusative verbs, similarly, do not select the Transitivity head. (8) a. transitives PredP DP

Predǯ Pred

TransP Trans

VP V

DP

b. unergatives PredP DP

Predǯ Pred

VP V

7

(PP)

In relation to the analysis of transitivity, it should be noted that the character of the subject as an argument in the above analyses is differentiated from its syntactic role, something, which is shown by the recognition of two different heads (AgrS and v) that are responsible for these two aspects of the external argument.

Theoretical Framework

31

c. unaccusatives PredP Pred

VP DP8

V

The functional category of Transitivity includes in transitive structures ij-features, whereas it includes in passive structures the non-active morpheme. In this second instance (passive structures), the object moves to the [Spec, PredicationP] position, and then (in English) to the [Spec, TP] position, where it receives nominative case. The presence of the Predication head helps in the analysis of structures with an expletive subject in the position of the subject and a DP in the accusative in specific languages (such as German) but is not dictated by a general theoretical need whereby its presence is completely necessary. I consider that the presence of the Predication head and the Transitivity head in the structures of verb classes is not justified by the syntactic behaviour of the verbs, and, therefore, it is uneconomical (particularly in the description of the Greek language), and I follow Theophanopoulou-Kontou (2004), who analyses verb structures according to the Minimalist Programme (Chomsky 1995) and the analysis of v. According to this analysis, the VP in transitive structures merges with a functional head (v; the basic feature of this abstract head is causativity) that inserts the subject in the [Spec, vP] position. I hypothesise that this functional head exists in the causative verbs (change-of-state meaning) but does not exist in the accusative verbs (transitive non-causative). In this way, I syntactically translate the distinction between the causative and accusative transitive verbs that was described in the first chapter (para. 1.2)9. (9) a. causatives vP external argument

vǯ v

VP V

DPobj

b. accusatives VP DPsubj

Vǯ V

DPobj

Since v is connected to causativity characteristics, I differ from Theophanopoulou-Kontou (2004) and hypothesise that v is present in anticausative structures (following Harley 1995, Collins 1997, Marantz 1997, Embick 1998, Arad 1999, Travis 1999, Alexiadou 2001, and Alexiadou & Anagnostopoulou 2004):

8

Transitivity for Bowers is an independent feature; it is not the result of a feature of the external argument. According to this analysis, the existence of impersonal verbs, that are purely transitive but they do not have an external argument, and whose existence demands the differentiation of transitivity from agentivity, may be predicted: (1) Es expletive subject ‘I feel cold’

9

friert mich freeze.3SG me.ACC

The above analysis also interprets why inversion structures and expletive structures with there in English are only grammatical with unaccusative verbs. This hypothesis is mainly supported by semantic characteristics that I cited in the previous chapter and counters the analyses of Chomsky (2000, 2005/2008) and Anagnostopoulou (1996, 2003). Chomsky argues that v is present in all transitive verbs, and in order to reverse this position, further (synchronic and diachronic) research on the ‘accusative verbs’ is required. Detailed research on the ‘accusative verb’ class (nonprototypical transitives) is not the object of this study and remains open to future research.

32

Chapter Two

(10) anticausatives vP vǯ v

VP V

DP

The above structure comprises the working hypothesis for this study. I hypothesise that the above structure holds if a close connection between the causatives and the anticausatives can be proven. v in anticausatives is [-agent], whereas v in causatives is [+agent]. I assume that through the lexical process, the insertion of the DP-agent is blocked in the above (anticausative) structure with the result that the [Spec, vP] position is not filled in the initial structure (for the lexical process, see 2.1.2). The single DP is generated in the [Compl, VP] position and moves to the [Spec, vP] position to receive nominative case10. (b) Voice The ‘little v approach’ is justified semantically by the analysis of Kratzer (1993), who argues that the distinguishing semantic features of the subject are derived from the insertion of a separate head (v) from the head that introduces the object. The head that is responsible for the internal argument is called Voice head by Kratzer and is higher than the vP. It has two abstract forms (in complementary distribution), one form that selects external arguments [+agent] and one form that does not select external arguments [-agent]. Voice differs from v because it is responsible for the accusative case of the object. (11) VoiceP

DP

Voiceǯ Voice [+accusative]

vP

v

VP

The non-active morpheme of the passive types that absorbs the accusative case is checked by the Voice (see Rivero’s analysis above). I follow the proposal of Collins (2005) in order to analyse the passive types11 and the correlation of the presence of non-active morphemes and the absence of the DPaccusative in Modern Greek. In accepting Collins’s analysis, I consolidate the assignment of the external argument by v in both the transitive and passive types. According to the Minimalist Programme, the assignment of all ș-roles is based on the configurational structure, as every syntactic position ([Spec, vP], [Compl, VP]) is taken as being connected to a specific ș-role. According to this view, it is essential that the ș-role of the external argument of the passive type is given in the same way as the ș-role of the external argument of the transitive type is given. (12) Passive type in Modern English – presence of an agent argument, Collins’s approach (2005)

10

11

I assume that the DP in this position can move to the subject position (because it is the single argument of the structure) where it checks the nominative case (I will show below that this movement is possibly marked with the presence of non-active morphology for those verbs in classical Ancient Greek, whereas it is not marked with non-active voice in Modern Greek). This movement is difficult to justify theoretically unless we accept a mechanism that rules out the checking of the case in the [Compl, VP] position if the nominative case feature in the [Spec, vP] position has not been checked first. I am not concerned with the derivation and licensing of the subject. For related analyses I refer to Alexiadou & Anagnostopoulou (1998), Spyropoulos & Philippaki (2001), Roussou & Tsimpli (2006). Passivisation belongs to the syntactic level; Horvath & Siloni (2005) present evidence that shows that passivisation is a syntactic process; syntactic processes do not change the ș-roles of the arguments.

Theoretical Framework

33

ex.: The book is written by John. TP DP

Tǯ

T subject movement in English V is

VP

VoiceP

ParticipleP Voiceǯ 

Participleǯ  Voice vP Participle VP by  DP vǯ V

external argument write the book John v ___________________________

phrase movement

The external argument derives in the [Spec, vP] position in the passive type, as in the transitive type. The features of the accusative case and the external ș-role are separate features, and they should be separated in the passive structures. In English, the morpheme (-EN) in passive structures does not absorb the external ș-role but blocks the assignment of the accusative case since there is only one more (apart from the by-Phrase) position for argument (which necessarily assigns nominative case)12. The preposition by in the passive structures in English does not form a PP but is the head of the VoiceP. The movement of the ParticipleP to the [Spec, VoiceP] is essential in English for the subject movement of the passive type to the [Spec, TP] position to be possible (subject position for English). Due to this movement, the subject (which is found in the ParticipleP; Collins names that type of movement “smuggling movement” of subject) can go beyond the DP (that is found in [Spec, vP] and blocks direct movement of the subject), as it moves together with the ParticipleP to the [Spec, VoiceP] position, and finally it can move to the [Spec, TP] position. The passive type in Greek is not formed periphrastically but with the addition of a non-active verb ending which is checked in the Voice head (cf. Rivero’s analysis above). Since Voice in Greek checks the verb voice morphology (verb ending -ȝĮȚ -me), then Collins’s schema must be modified for Greek; the preposition Įʌȩ (apo) ‘by’ is generated in a different position in Greek from what it is in English. Moreover, the preposition Įʌȩ (apo) ‘by’ does not have such restricted distributions in Greek as the preposition by in English: (a) by- only with passive types; its derivation in the Voice head is justified (cf. the leg of (*by) the table, proud of (*by) his friend, for (*by) John to be happy) (b) Įʌȩ (apo) ‘by’ – not only with passive types but also in structures in which the corresponding ones with by in English are ungrammatical, such as (i) with adverbs (ʌȐȞȦ Įʌȩ pano apo ‘above’), (ii) for the meaning of origin (IJȡĮʌȑȗȚ Įʌȩ ȟȪȜȠ / trapezi apo ksilo / ‘table from wood’; İȓȝĮȚ Įʌȩ / ime apo / ‘I am from’; ȑȡȤȠȝĮȚ Įʌȩ / erxome apo / ‘I come from’) or (iii) for the formation of genitive forms with diminutive nouns (Ș ȕȠȒșİȚĮ Įʌȩ IJȠ ʌĮȚįȐțȚ / i voithia apo to pedhaki / ‘the help from the little child’). I assume that in Greek, the preposition Įʌȩ / apo / ‘by’ and the DP that denotes the external argument comprise a constituent and are generated in the [Spec, vP] position, exactly as the external argument in transitive structures is generated in the [Spec, vP] position13. Furthermore, in Greek, head 12

13

As will be shown by the analysis of diachronic data, the non-active voice does not block the assignment of accusative case in Ancient and Hellenistic Greek but blocks the assigning of accusative case in Medieval and Modern Greek. That the presence of a PP in the [Spec, vP] position is possible can also be seen from the possibility of the expression of the external argument with case morphology (dative case) in passive structures in classical

34

Chapter Two

movement for checking of the non-active ending takes place and not phrase movement since it is not essential for there to be subject movement to the [Spec, TP] position, as in English14 (for more on this and the modification of Collins’s analysis that I attempt for Greek, see the full interpretation of the diachronic data in section 4.2). (13) Greek passive type – presence of the agent argument, Collins’s approach modification that I propose for languages with verb voice morphology Įʌȩ IJȠȞ īȚȐȞȞȘ ȖȡȐijIJȘțİ IJȠ ȕȚȕȜȓȠ apo ton Jani ghraftike to vivlio by the Janis write.NACT.PAST.PERFVE.3SG the.NOM book.NOM ‘the book was written by Janis’ TP T

VoiceP Voice non-active suffix

head movement

vP

PP (Įʌȩ/apo ‘by’-agent) Įʌȩ IJȠȞ īȚȐȞȞȘ v apo ton Jani by the Janis

vǯ VP V ȖȡȐijȦ ghrafo write

DP IJȠ ȕȚȕȜȓȠ to vivlio the book

The Voice head is considered, by certain linguists, also to be present in the structure of non-active anticausatives. I will deal with the anticausatives as a unified class and do not follow the tripartite distinction of anticausatives of Alexiadou & Anagnostopoulou (1999a, b, 2004), which is based on voice morphology in anticausatives in Modern Greek (active vs. non-active) (examples 14 vs. 15). I attempt to disassociate voice morphology from the direction of the anticausative derivation (for arguments, see 2.1.3), whereas I will also not consider that passives and anticausatives have the same structure (following the arguments of Theophanopoulou-Kontou 2004). Alexiadou & Anagnostopoulou (1999b, 2004) syntactically distinguish the anticausatives of Modern Greek according to the type of the PHRASE (XP) that follows the operator BECOME/RESULT. (14) a. Ǿ ǼȜʌȓįĮ ȟİʌȐȖȦıİ IJĮ ȥȐȡȚĮ i Elpidha ksepaghose ta psaria the.NOM Elpida.NOM defrost.ACT.PAST.PERFVE.3SG the.ACC fishes.ACC ‘Elpida defrosted the fish’ b. ȉĮ ȥȐȡȚĮ ȟİʌȐȖȦıĮȞ ta psaria ksepaghosan the.NOM fishes.NOM defrost.ACT.PAST.PERFVE.3PL ‘The fish defrosted’ (15) a. ȅ ȋĮȡȐȜĮȝʌȠȢ ȗȑıIJĮȞİ IJȠ ȖȐȜĮ o Xaralambos zestane to ghala the.NOM Haralambos.NOM warm.ACT.PAST.PERFVE.3SG the.ACC milk.ACC ‘Haralambos warmed up the milk’ b. ȉȠ ȖȐȜĮ ȗȑıIJĮȞİ/ ȗİıIJȐșȘțİ Įʌȩ ȝȩȞȠ IJȠȣ to ghala zestane/ zestathike apo mono tu the.NOM milk.NOM warm.ACT.PAST.PERRFVE.3SG/ warm.NACT.PAST.PERFVE.3SG by itself

14

Ancient Greek, as well as cross-linguistically (parallel with the PP). I assume that the availability of the presence of a PP in the [Spec, vP] position is licensed by the subcategorisation of the v of the passive structures. I will try to strengthen this view in para. 4.2. See the classic analysis of Alexiadou & Anagnostopoulou (1998) concerning the non-essentiality of the subject movement in Greek. It should be noted that in Greek, structures with a word order such as (1) are grammatical: (1) īȡȐijIJȘțİ Įʌȩ IJȠȞ īȚȐȞȞȘ ghraftike apo ton Jani write.NACT.PAST.PERFVE.3SG by the Janis ‘The book was written by Janis’

IJȠ ȕȚȕȜȓȠ to vivlio the.NOM book.NOM

Theoretical Framework

35

‘The milk warmed up by itself’ aǯ. Ǿ țȠȣȕȑȡIJĮ ȗȑıIJĮȞİ IJȘȞ ǼȚȡȒȞȘ i kuverta zestane tin Irini the.NOM blanket.NOM warm.ACT.PAST.PERFVE.3SG the.ACC Irini.ACC ‘The blanket warmed up Irini’ bǯ. Ǿ ǼȚȡȒȞȘ ȗİıIJȐșȘțİ/ *ȗȑıIJĮȞİ i Irini zestathike/ *zestane the.NOM Irini.NOM warm.NACT.PAST.PERFVE.3SG / warm.ACT.PAST.PERFVE.3SG ‘Irini warmed up’

Alexiadou & Anagnostopoulou argue that active anticausatives are derived from a structure with the operator BECOME (+ predicate: ȟİʌȐȖȦıİ ksepaghose ‘defrost.ACT.PAST.PERFVE.3SG’, + possession: ȗȑıIJĮȞİ zestane warm.ACT.PAST.PERFVE.3SG) (16a, 16b), whereas the non-active anticausatives of the type ȗİıIJȐșȘțİ zestathike warm.NACT.PAST.PERFVE.3SG from a structure with the operator RESULT, which includes VOICE (with these verbs, VOICE means [-agent, + Manner] (16c). The passives for Alexiadou & Anagnostopoulou are included in another schema with VOICE [+agent, + Manner] (17). (16) a.

vP

v BECOME

b.

AdjP

vP

BECOME

ȋP (= possessive structure) ȋǯ

DP

HAVE c.

vP

RESULT

VoiceP

Voice

(17)

DP

VP

VoiceP

Voice +Manner +agent

VP

According to my approach, the semantic features of anticausatives cross-linguistically, such as the absence of transitivity characteristics and the by-phrase that denotes the agent (cf. TheophanopoulouKontou 2003a, 2004) and the role of the voice morphology (which for the passives in Modern Greek is connected with the syntax (Voice) but for the anticausatives with the initial marking of the nonalternating intransitive and not with the transitivity alternation; see paras. 2.1.3 and 4.2), cannot support a parallel structure for the anticausatives and the passives. The accuracy of this approach needs, of course, support from diachronic data. To summarise, from the aforementioned arises the most important role of v and Voice for every

36

Chapter Two

analysis on transitivity15. (18) Causative Passive

v v v (by-PP) Voice

assigns the external ș-role checks the accusative case of the patient assigns the external ș-role absorbs the direct object in the accusative

v (i) bears causativity features: change-of-state and ș-role of patient (see also paras. 1.2 and 2.1.2.2), (ii) assigns the external ș-role to causative and passive structures, (iii) is distinguished between two types: the one type of v allows the presence of an agent, whereas the other type of v does not allow the presence of an agent (Harley 1995; Collins 1997; Marantz 1997; Embick 1998; Arad 1999; Travis 1999; Alexiadou 2001; Alexiadou & Anagnostopoulou 2004). Voice is connected to the accusative case of the direct object by blocking the presence of the accusative case of the direct object in passive types. The above analyses comprise the starting point of my diachronic analysis and I will be returning to these – and mainly to Collins’s analysis, which ǿ touched on only superficially here – when I discuss the distribution of diachronic data (chapter 3) and the overall interpretation (chapter 4).

2.1.2 Transitivity and lexicon-syntax interface 2.1.2.1 Semantic features, verb classes, and transitivity I assume that there is a lexicon-syntax interface in the sense of the hypothesis of the existence of verb classes that are determined semantic-syntactically (Levin 1993; Levin & Rappaport Hovav 1995, 2006a). I follow the Generative Grammar tradition and assume that the information that is fundamental to the syntactic and semantic use of a word can be isolated from any other information that is related to the word (mainly Hale & Keyser 1993, 2002; Levin & Rappaport Hovav 1995, 2005). Only information related to the linguistic system can be enlisted into a limited theory of lexical items. For example (Pesetsky 1995; Juarros-Daussà 2003), the meaning of ‘loudness of voice’ does not have grammatical consequences and is not related to the linguistic grammatical system if we take into consideration the cross-linguistic syntactic behaviour of the verbs ȥȚșȣȡȓȗȦ/psithirizo ‘whisper’ and ijȦȞȐȗȦ/fonazo ‘shout’. All speakers of Modern Greek recognise the common characteristics (‘denotation of way of speaking’) and the differences (‘denotation of loudness of voice’) of the verbs ‘whisper’ and ‘shout’ (for more on these verb classes, cf. Levin 1993: 240). These differences do not play any role in the syntactic environments in which these verbs appear in Modern Greek. Similarly, the denotation of the means of speaking in relation to the transmission of speech (without referring to the means) does not have syntactic consequences in Modern Greek. In contrast, in Modern English, the feature of the means of speech meaning determines the inability of their participation in the dative alternation, in contrast with, for example, verbs that denote ‘transfer of message’: (19) a. Ǿ DZȞȞĮ ȥȚșȪȡȚıİ/ ijȫȞĮȟİ IJȘȞ ĮʌȐȞIJȘıȘ ıIJȠ ıȣȝȝĮșȘIJȒ IJȘȢ i Ana psithirise/ fonakse tin apandisi sto simathiti tis the.NOM Anna.NOM whispered.3SG / shouted.3SG the.ACC answer.ACC to-the classmate her ‘Anna whispered / shouted the answer to her classmate’ b. Ǿ DZȞȞĮ ȥȚșȪȡȚıİ/ ijȫȞĮȟİ IJȠȣ ıȣȝȝĮșȘIJȒ IJȘȢ IJȘȞ ĮʌȐȞIJȘıȘ i Ana psithirise/ fonakse tu simathiti tis tin apandisi the.NOM Anna.NOM whispered.3SG / shouted.3SG the.GEN classmate.GEN her

(20) a.

b.

(21) a. b. 15

the.ACC answer.ACC

‘*Anna whispered / shouted her classmate the answer’ Ǿ DZȞȞĮ İȓʌİ IJȘȞ ĮʌȐȞIJȘıȘ ıIJȠ ıȣȝȝĮșȘIJȒ IJȘȢ i Ana ipe tin apandisi sto simathiti tis the.NOM Anna.NOM told.3SG the.ACC answer.ACC to-the classmate her ‘Anna told the answer to her classmate’ Ǿ DZȞȞĮ İȓʌİ IJȠȣ ıȣȝȝĮșȘIJȒ IJȘȢ IJȘȞ ĮʌȐȞIJȘıȘ i Ana ipe tu simathiti tis tin apandisi the.NOM Anna.NOM told.3SG the.GEN classmate.GEN her the.ACC answer.ACC ‘Anna told her classmate the answer’ John whispered/ shouted the answer to Mary *John whispered/ shouted Mary the answer

Chomsky (2000) argues that v* (v with full argument structure) is a head of a strong phase, whereas v in passives and unaccusatives is not a head of a strong phase because it does not have an external argument. According to Collins’s analysis, the passive v is v* (strong phrase head) since it has an external argument.

Theoretical Framework

37

(22) a. John told the answer to Mary b. John told Mary the answer

Consequently, the feature of the verb meaning of the ‘loudness of voice’ does not seem to influence verb syntactic behaviour at all, either in Modern Greek or in Modern English. In contrast, the constituent of ‘means’ in relation to the constituent of ‘message transfer’ can be an important linguistic component of meaning (as can be seen from the Modern English examples). I will argue that the participation of verbs in syntactic structures is also determined by their meanings (Levin 1993; Levin & Rappaport Hovav 1995, 2005). For example, native speakers of Modern Greek are aware that certain verbs participate in alternations concerning the expression of their arguments, and also that those alternations are often followed by change of meaning: locative alternation (23) a. Ǿ ȀĮIJİȡȓȞĮ ijȩȡIJȦıİ IJȠ ĮȣIJȠțȓȞȘIJȠ ȝİ IJȚȢ ȕĮȜȓIJıİȢ i Katerina fortose to aftokinito me tis valitses the.NOM Katerina.NOM loaded.3SG the.ACC car.ACC with the cases ‘Katerina loaded the car with the cases’ b. Ǿ ȀĮIJİȡȓȞĮ ijȩȡIJȦıİ IJȚȢ ȕĮȜȓIJıİȢ ıIJȠ ĮȣIJȠțȓȞȘIJȠ i Katerina fortose tis valitses sto aftokinito the.NOM Katerina.NOM loaded.3SG the.ACC cases.ACC to-the car ‘Katerina loaded the cases onto the car’ c. Ǿ DZȞȞĮ ıțȑʌĮıİ IJȠ țȡİȕȐIJȚ ȝİ IJȠ ʌȐʌȜȦȝĮ i Ana skepase to krevati me to paploma the.NOM Anna.NOM covered.3SG the.ACC bed.ACC with the quilt ‘Anna covered the bed with the quilt’ d. *Ǿ DZȞȞĮ ıțȑʌĮıİ IJȠ ʌȐʌȜȦȝĮ (ʌȐȞȦ) ıIJȠ țȡİȕȐIJȚ *i Ana skepase to paploma (pano) sto krevati the.NOM Anna.NOM covered.3SG the.ACC quilt.ACC (above) to-the bed ‘*Anna covered the quilt onto the bed’

The argument structure alternation that is concerned with transitivity is the transitivity alternation. Verbs such as ‘break’, ‘open’, and ‘melt’ show alternations in their transitivity (transitive and intransitive use), whereas verbs such as ‘cut’, ‘occur’, and ‘climb’ are stable regarding their (in)transitivity. The alternation is also known as causative-anticausative alternation because the transitive type of these verbs has causative meaning (cause-of-change meaning). The intransitive type refers to the result (inchoative) and comprises an unaccusative intransitive verb because its only argument has the ș-role of the patient. change-of-state meaning (24) a. ȅ ȀȫıIJĮȢ ȑıʌĮıİ IJȘ IJȗĮȝĮȡȓĮ o Kostas espase ti tzamaria the.NOM Kostas.NOM break.ACT.PAST.PERFVE.3SG the.ACC pane of glass.ACC ‘Kostas broke the pane of glass’ (agent – patient) a’. ȅ ĮȞİȝȠıIJȡȩȕȚȜȠȢ ȑıʌĮıİ IJȘ IJȗĮȝĮȡȓĮ o anemostrovilos espase ti tzamaria the.NOM tornado.NOM break.ACT.PAST.PERFVE.3SG the.ACC pane of glass.ACC ‘The tornado broke the pane of glass’ (cause – patient) b. Ǿ IJȗĮȝĮȡȓĮ ȑıʌĮıİ i tzamaria espase the.NOM pane of glass.NOM break.ACT.PAST.PERFVE.3SG ‘The pane of glass broke’ (patient) (25) a. Ǿ ǿȦȐȞȞĮ ȑțȠȥİ IJȠ ȥȦȝȓ i Ioana ekopse to psomi the.NOM Ioana.NOM cut.ACT.PAST.PERFVE.3SG the.ACC bread.ACC ‘Ioana cut the bread’ (agent – patient) b. *ȉȠ ȥȦȝȓ ȑțȠȥİ Įʌȩ ȝȩȞȠ IJȠȣ *to psomi ekopse apo mono tu the.NOM bread.NOM cut.ACT.PAST.PERFVE.3SG by itself ‘*The bread (was) cut by itself’ (patient) (26) a. ȅ ȆȑIJȡȠȢ ıțĮȡijȐȜȦıİ o Petros skarfalose the.NOM Petros.NOM climb.ACT.PAST.PERFVE.3SG ‘Petros climbed’ (agent) b. *Ǿ įȪȞĮȝȒ IJȠȣ ıțĮȡijȐȜȦıİ IJȠȞ īȚȐȞȞȘ *i dhinami tu skarfalose ton Jani

38

Chapter Two the.NOM strength.NOM his climb.ACT.PAST.PERFVE.3SG the.ACC Janis.ACC ‘*His strength climbed Janis’ - with the meaning: ‘it helped Janis to climb’ (cause – patient) (27) a. ȅ ȜĮȖȩȢ İȟĮijĮȞȓıIJȘțİ o laghos eksafanistike the.NOM rabbit.NOM disappear.NACT.PAST.PERFVE.3SG ‘The rabbit disappeared’ (patient) b. ȅ ȝȐȖȠȢ İȟĮijȐȞȚıİ IJȠȞ ȜĮȖȩ o maghos eksafanise ton lagho the.NOM magician.NOM disappear.ACT.PAST.PERFVE.3SG the.ACC rabbit.ACC ‘The magician made the rabbit disappear’ (agent – patient)

but in Modern English: (28) a. The rabbit disappeared b. *The magician disappeared the rabbit

The verbs that participate in the above argument structure alternation belong to semantically linked subclasses in relation to narrower semantic features (broad feature: ‘change-of-state’ – narrow features: ‘contact’, ‘movement’, ‘affectedness’, cf. Levin & Rappaport Hovav 1989, 1994, 1995, 2005). Consequently, the participation of a verb in a semantic group depends on (a) whether the specific verb satisfies the demands of the argument structure and (b) the narrower semantic constraints on the specific verb class. 2.1.2.2 Decomposition of meaning, event schemas, and causativity Having defined the above basic principles that I will follow for the syntactic analysis of transitivity, in this section, I attempt to determine the principles that govern the analysis of features of verbs that are transitive or intransitive; the aim, in other words, is to show why specific verbs have a specific syntactic representation (for example, why the verb ıʌȐȦ spao ‘to break’ has causative syntactic representation whereas the verb ȖİȜȐȦ jelao ‘to laugh’ has unergative syntactic representation). In relation to which verbs are causative, which intransitive anticausatives and which intransitive unergatives the same problem applies which also applies to and is concerned with syntactic knowledge, and which, consequently, can also extend to the domain of the lexicon (Levin & Rappaport Hovav 1995, 2005): the lexicon also appears to be governed by abstract principles, exactly as other domains of grammar are. The modern approaches of the lexicon-syntax interface agree that certain aspects of verb meaning are related to the syntactic behaviour of verbs and have syntactic consequences, determining, for example, how the mapping of arguments to syntactic positions occurs (‘the distinction between unergative and unaccusative intransitive verbs is not only semantically determined but also syntactically represented’ Levin & Rappaport Hovav 1995: 30). I will follow the ‘lexical decomposition’ approach, according to which lexical items are not atomic, holistic, unseparated units but can be decomposed into semantic parts (Gruber 1976; Dowty 1979; Jackendoff 1990; Hale & Keyser 1993). Adopting the above perspective, I hypothesise that the interaction between the event structure of a predicate and the argument structure determines verb transitivity. This position is supported by the analyses of Vendler (1967), Dowty (1979), and Pustejovsky (1991), among others, who assume that every verb has an event structure and that the event structure of every verb represents aspectual information. Vendler and Dowty organise the verbs into four main types: ACTIVITIES, ACCOMPLISHMENTS, STATES, and ACHIEVEMENTS. Pustejovsky distinguishes three event types: PROCESSES (the activities of Vendler and Dowty), STATES, and TRANSITIONS (the accomplishments and achievements of Vendler and Dowty). ȟȑȡȦ ksero ‘to know’, ĮȖĮʌȫ aghapo ‘to love’ . IJȡĮȖȠȣįȫ traghudho ‘to sing’, ȤȠȡİȪȦ xorevo ‘to dance’, ȥȐȤȞȦ psaxno ‘to search’. ACHIEVEMENTS ȜȚȫȞȦ liono ‘to melt’ (intr.), ʌİșĮȓȞȦ petheno ‘to die’. ACCOMPLISHMENTS ȤIJȓȗȦ xtizo ‘to build’, ȗȦȖȡĮijȓȗȦ ȝȓĮ İȚțȩȞĮ/zoghrafizo mia ikona/‘to paint a picture’, ıțȠIJȫȞȦ skotono ‘to kill’, ȜȚȫȞȦ liono ‘to melt’ (trans.).

(29) STATES ACTIVITIES

The event types are further separated into subevents: Dowty argues that the STATES are primitive predicates, and in parallel there are another four semantic primitive elements called connectives or operators: BE (state), BECOME (achievement), CAUSE (accomplishment), and DO (activity). All of the aspectual groups have a logical structure made up of one or more stative predicates that are connected with one of the logical connections. The accomplishments (‘to melt’, ‘to kill’, ‘to destroy’) show a complex event comprised of ACTIVITY and STATE-result. For example, the transitive verb ‘to melt’ has a complex semantic structure:

Theoretical Framework

39

[x CAUSES [y BECOME melted]] A predicate that denotes ACTIVITY, such as the verb ‘to laugh’, has the following structure: [x DO laughter] This type of semantic decomposition is called ‘thematic core’ by Pinker (1989) and ‘lexical semantic template’ by Levin & Rappaport Hovav (1995). As can be seen from the above, the distinction between complex and simple event schemas based on the presence or absence of the meaning of CAUSE (causativity) is especially important for the relationship of event schemas with argument structures and transitivity (Levin & Rappaport Hovav 1995, 2005). SIMPLE EVENT STRUCTURE (1) [x DOES ] (activity) (2) [x ] (state) (3) [BECOMES [x ]] (achievement) COMPLEX/CAUSATIVE EVENT STRUCTURE [[x ACTS ] CAUSES [BECOMES [y ]]] (accomplishment - causative)16 The simple event structure template is comprised of only one event (independent event schema), whereas the complex template is comprised of two events (or with another term: subevents). The meaning of causation of change-of-state is considered a broader term than the aktionsart of accomplishment and of telicity (McCawley 1976)17. Whether a verb has a complex structure or not, it is the presence of causativity that plays a role, not telicity)18. The majority of unergative intransitive verbs are ACTIVITIES, the anticausative intransitive verbs are ACHIEVEMENTS, and the alternating and non-alternating causative verbs are ACCOMPLISHMENTS, but there is no one-to-one correspondence between transitivity and event types since some transitive non-alternating verbs are activities (‘to search’), whereas some unergative verbs with the presence of a PP are accomplishments (see motion verbs in Modern English with PPs of direction: walked to the store)19. 16

17

18

19

Similarly, the transitive verb ‘to kill’ is considered as causative in contrast with the intransitive verb ‘to die’ and the ditransitive verb ‘to give’ as causative in contrast with the transitive ‘to have’. Therefore, it stands to reason that according to this theory, causativity and increased valency are connected. In contrast, Givón (1984, 1990a, b), Lazard (1984, 2002), Tsunoda (1994), and Cennamo (2000a, b) include completion/telicity of the event in the main parameters of causativity-transitivity equally (close to the factors of control and affectedness). The significant role that ‘cause of a change’ plays in the meaning of a verb and of a structure is also emphasised by Antonopoulou (1991). An attempt was made to express the approach of the verb meaning decomposition and verb classes based on the event type as purely syntactic in terms of pure syntactic derivation. The result of this attempt is the l(exical)-syntax (Travis 1991, 1994; ǹmberber 1993, 1996; Hale & Keyser 1993; Harley 1995; Baker 1997): event types and ș-roles are represented by X’ structures at the pre-syntactic domain. The event types are heads of projections, and the ș-roles are Specifiers (Hale & Keyser 1993; Travis 1994; Harley 1995; Amberber 1996). The rules of mapping are based on hierarchical relations. For example, if CAUSE is a sub-predicate, this will comprise the head of the highest VP in the lexical representation. The agent argument is found in the [Spec, CAUSE P] position (Baker 1997). The primitive semantic components of the event types are represented in X’ structures, and the event structure is represented in phrase structures that allow syntactic processes to take place at the lexical domain, below the VP. In the syntactic event schemas, the heads are v that encode the semantic operators BE, BECOME, CAUSE, DO, and STATE, whereas the complements contain the lexical information (Adj, N, P): VP2 = VoiceP(hrase) (Kratzer 1996) = EventP(hrase) (Harley 1995)  o ȐȞİȝȠȢ Vǯ o anemos  the wind CAUSE VP1  IJĮ įȑȞIJȡĮ Vǯ ta dhendra  the trees BE/BECOME țĮIJİıIJȡĮȝȝȑȞȠ katestrameno destroy.NACT.PAST.PERFVE.PART

40

Chapter Two

The verb classes (transitive, intransitive verbs that participate or do not participate in transitivity alternation) vary in relation to the number of subevents and ș-roles20. By modifying the analyses of Levin & Rappaport Hovav (1995) and Chierchia (1989/2004) and by putting emphasis on the Greek data and the central role that I argue is played by transitivity, we observe the following correspondences between verb syntax and verb meaning: (a) Causative non-alternating verbs (ıțȠIJȫȞȦ skotono ‘to kill’): Ætwo subevents (CAUSE: causing event, BECOME: state-result) Æagent [+specified] (b) Causative alternating verbs (ıʌȐȦ spao ‘to break’): Ætwo subevents (CAUSE: causing event, BECOME: state-result) Æagent [-specified] In intransitive structure: suppressed agent [-specified] the patient becomes the subject through DP movement (c) Intransitive non-alternating verbs = unaccusative non-alternating verbs (ȣʌȐȡȤȦ iparxo ‘to exist’): Ætwo subevents (for Chierchia) (CAUSE + BECOME) (but for Levin & Rappaport Ǿovav, only one subevent, BECOME) Æ[-agent] (d) Intransitive unergative verbs (țȠȜȣȝʌȐȦ kolimbao ‘to swim’): Æone subevent Æpresence of agent Æinternal cause of the change-of-state The syntactic variation is seen as the result of the extension of the basic event schema (Rappaport Hovav & Levin 1998, 2001; Levin & Rappaport Hovav 2005): Template augmentation “Event structure templates may be freely augmented up to other possible templates in the basic inventory of event structure templates” (Rappaport Hovav & Levin 1998: 111). The augmentation of an activity template, such as the verb IJȡȑȤȦ trexo ‘to run’, with the addition of the achievement template (connection with the CAUSE head) results in the event schema of accomplishment: (30) a. IJȡȑȤȦ trexo ‘to run’: [[x ACTS ]:

ȅ ȖȣȝȞĮıIJȒȢ IJȡȑȤİȚ o jimnastis trexi the.NOM trainer.ȃȅȂ run.ACT.PRES.3SG ‘the trainer runs’ b. IJȡȑȤȦ trexo ‘to run’: [[x ACTS ] CAUSES [BECOMES [y ]]]: ȅ ȖȣȝȞĮıIJȒȢ IJȡȑȤİȚ IJĮ ʌĮȚįȚȐ o jimnastis trexi ta pedhia the.NOM trainer.ȃȅȂ run.ACT.PRES.3SG the.ACCchildren.ACC

‘the trainer runs the children / forces the children to run’

Verbs such as ijIJȐȞȦ ftano ‘to arrive’, which are verbs with aktionsart and event template of

20

Through head-to-head movement, the predicate țĮIJİıIJȡĮȝȝȑȞȠ katestrameno destroy. NACT.PAST.PERFVE.PART ‘destroyed’ moves and is incorporated in the verb heads BE/BECOME in VP1 and CAUSE in VP2, deriving the transitive structure: ȅ ȐȞİȝȠȢ țĮIJȑıIJȡİȥİ IJĮ įȑȞIJȡĮ / o anemos katestrepse ta dhendra / the.NOM wind.NOM destroy.ACT.PAST.PERFVE.3SG the.ACC trees.ACC / ‘the wind destroyed the trees’. Disagreements have occurred concerning the relations of the various ș-roles and as far as the existence of a thematic hierarchy that determines the assignment of the ș-roles to the syntactic positions is concerned. The view that thematic hierarchy exists has already been expressed by Jackendoff (1990): (agent (experiencer(goal/source/location/(theme)))) Baker (1988: 46) assumes that there is mapping and correlation between the thematic hierarchy and the syntactic hierarchy; Baker’s Uniformity of Theta Assignment Hypothesis (1988: 46): ‘identical thematic relationships between items are represented by identical structural relationships between those items at the level of D-Structure’. Consequently, if a verb has an agent and theme in its thematic grid, then the agent will be linked to a higher syntactic position that will asymmetrically govern the theme-role.

Theoretical Framework

41

achievement, have stable syntactic behaviour owing to their event template21. The template of these verbs cannot be ‘augmented’ and form another grammatical (not ungrammatical) event template. The same also happens with the accomplishments since the accomplishment template is the most complex event template possible22. Certain verbs appear in structures with more arguments than those that are related with the corresponding event template of their event structure (Grimshaw 1993; Rappaport Hovav & Levin 1998). This concerns verbs that denote activity (e.g., ȟİıțȠȞȓȗȦ kseskonizo ‘to dust’) and of which the event structure does not include a subevent that justifies the presence of an argument in the position of the syntactic object (the meaning of result and location towards which the movement takes place -and consequently an augmentation of the event structure- cannot be included). For example, if the surface on which the act that indicates a verb of activity takes place is also to be denoted, such as ȟİıțȠȞȓȗȦ kseskonizo ‘to dust’ (e.g., the desk), the surface must be inserted in the final event structure in another way since it does not correspond with the activity template. The additional argument that can be added derives only from the idiosyncratic meaning (root/constant) of the verb23 (it is required only by the specific meaning of the verb ȟİıțȠȞȓȗȦ kseskonizo ‘to dust’ which it distinguishes from the other verbs that belong to activities, such as IJȡȑȤȦ trexo ‘to run’), and not from the common meaning of activities. According to this remark, two kinds of arguments for each verb can be distinguished: the structural arguments that are required from both the event structure template and the idiosyncratic meaning and the pure idiosyncratic arguments that are required only by the idiosyncratic meaning of the verb, not by the event structure. The subjects of IJȡȑȤȦ trexo ‘to run’ and ȟİıțȠȞȓȗȦ kseskonizo ‘to dust’ realise structural arguments, whereas the object of ȟİıțȠȞȓȗȦ kseskonizo ‘to dust’ realises a pure idiosyncratic argument24. In Modern Greek, the remark by Chila-Markopoulou (1989) about the existence of various verb classes in relation to the object omission follows a similar line: (a) verbs + obligatory nominal complement25 (ǻȚĮȜȪȦ IJȘ ıȣȞİįȡȓĮ / dhialio ti sinedhria / destroy.ACT.PRES.1SG the.ACC meeting.ACC / ‘I destroy the meeting’ – *įȚĮȜȪȦ *dhialio destroy.ACT.PRES.1SG. ‘*I destroy’; țĮIJĮıțİȣȐȗȦ ȑʌȚʌȜĮ / kataskevazo epipla / fabricate.ACT.PRES.1SG furnitures.ACC ‘I fabricate furnitures’ – *țĮIJĮıțİȣȐȗȦ *kataskevazo / fabricate.ACT.PRES.1SG / ‘*I fabricate’) (this concerns 21

Dimitra Theophanopoulou-Kontou (personal communication) indicates the strong constraints that exist for these verbs when (in many cases with strong metaphorical meaning), on rare occasions, their event template ‘augments’: the presence of a PP and the meaning of the participation of the direct object in the action are necessary: (1) ȉȠȞ ȑijIJĮıĮ ton eftasa he.ACC.WEAK arrive.ACT.PAST.PERFVE.1SG ‘I made him reach the level I wanted’

22

23

24

25

ıIJȠ İʌȓʌİįȠ sto epipedho to-the level

ʌȠȣ pu that

ȒșİȜĮ ithela wanted.1SG

This approach has also been applied in other lexical categories. In the framework of Distributed Morphology (Halle & Marantz 1993; Marantz 1993, 1997; Harley 1998; Alexiadou 2001), the features of aktionsart and of meaning are assigned to abstract roots, which are not specified regarding their morphosyntactic category until they connect syntactically with the appropriate functional categories. For example, the root of ¥destroy and the root of transitive ¥grow are accomplishments and denote change-of-state, the root of ¥destroy denotes changeof-state from an external cause (presence of the agent), and the root of the intransitive ¥grow denotes internal change-of-state (absence of agent; its subject is the argument of a separate causative functional head). The idea that the idiosyncratic semantic information must be separated from the primitive predicates has been widely accepted, even if various names have been given to the two parts of meaning. For example, Levin & Rappaport Hovav name the idiosyncratic part of meaning ‘constant’ in their earlier analyses, whereas in their 1995 book, they name the idiosyncratic meaning ‘root’, thus following the example of Pesetsky (1995). We can, therefore, contradistinguish two classes of verbs with direct objects in relation to the character of the second argument: (a) Verbs with a complex event structure (and accomplishment aktionsart, that denote result and change-ofstate, e.g. ıʌȐȦ spao ‘to break’, ȟȘȡĮȓȞȦ ksireno ‘to dry’, ȜȚȫȞȦ liono ‘to melt’, ĮȞȠȓȖȦ anigho ‘to open’) that obligatorily take a direct object. (b) Verbs with a simple event structure (that do not denote change-of-state, e.g. ʌȓȞȦ pino ‘to drink’, ȖȡȐijȦ ghrafo ‘to write’, įȚĮȕȐȗȦ dhiavazo ‘to read’; the direct object is not affected) that can omit the direct object. Brisson (1994) also observes that an idiosyncratic argument may not be expressed if it has prototypical meaning for the verb (ȖȡȐijȦ ghrafo ‘to write’ + prototypical object (ȖȡȐȝȝĮ ghrama ‘letter’ or ȕȚȕȜȓȠ vivlio ‘book’), ȟİıțȠȞȓȗȦ kseskonizo ‘to dust’ + prototypical object/surface (ȑʌȚʌȜĮ epipla ‘furnitures’ or IJȡĮʌȑȗȚ trapezi ‘table’)). In contrast, verbs such as ıʌȐȦ spao ‘to break’ require an object in every environment (H MĮȡȓĮ ȑıʌĮıİ IJĮ ʌȚȐIJĮ / i Maria espase ta piata / the.NOM Maria.NOM break.ACT.PAST.PERFVE.3SG the.ACC dishes.ACC / ‘Maria broke the dishes’ – *Ǿ ȂĮȡȓĮ ȑıʌĮıİ / *i Maria espase / the.NOM Maria.NOM break.ACT.PAST.PERFVE.3SG / ‘*Maria broke’). Cf. Varlokosta 1994.

42

Chapter Two

complex event structures/causatives) and (b) verbs + optional nominal complement (įȚĮȕȐȗȦ dhiavazo ‘to read’; ȟİȞĮȖȫ ksenago ‘to guide’; ȠįȘȖȫ odhigo ‘to lead’; Ȧșȫ otho ‘to push’; ʌĮȡȘȖȠȡȫ parighoro ‘to console’; ijȦIJȠȖȡĮijȓȗȦ fotoghrafizo ‘to photograph’; țĮIJİȣșȪȞȦ katefthino ‘to direct’; İȞșĮȡȡȪȞȦ entharino ‘to encourage’; IJȡȫȦ troo ‘to eat’; ʌȜȑȞȦ pleno ‘to wash’). As far as the second verb class is concerned, the presence of semantic sub-groups is argued by Chila-Markopoulou: (i) verbs that have the same meaning with and without the object (ȖȡȐijȦ ghrafo ‘to write’; ȗȦȖȡĮijȓȗȦ zoghrafizo ‘to paint’; ĮʌȠijĮıȓȗȦ apofasizo ‘to decide’), (ii) verbs, that, when their object is omitted, imply one of the possible objects (ʌĮȓȗȦ pezo ‘to play’; ʌȜȑȞȦ pleno ‘to wash’), and (iii) verbs that have different meanings without an object (ijȠȣıțȫȞȦ IJȠ ȝʌĮȜȩȞȚ / fuskono to baloni / blow.ACT.PRES.1SG the.ACC balloon.ACC / ‘I blow the balloon up’ – ijȠȣıțȫȞȦ fuskono blow.ACT.PRES.1SG ‘I swelled’). Chila-Markopoulou maintains that the implied object can be represented by the expletive pro (arbitrary and generalised). Of special interest for the present study is the indication of the productivity of the structures with optional objects in Greek and crosslinguistically. Consequently, the verbs with two arguments but a simple event structure (ȟİıțȠȞȓȗȦ kseksonizo ‘to dust’) can (not obligatorily) take a direct object since the second argument is licensed only by the idiosyncratic meaning of the verb and, therefore, does not follow the general rules of mapping between event structure and syntax. In contrast, the verbs with complex, causative event structures (transitive ıʌȐȦ spao ‘to break’) obligatorily take a direct object since the direct object is licensed both by the event structure of the verbs and by their idiosyncratic meaning. According to the above, we conclude and posit as a working hypothesis for the analysis of diachronic data the following: (a) Meanings based on the event template (and mainly meanings concerned with the cause of the change-of-state) are connected to the mapping of lexical semantics and syntax. (b) It is not possible for all aspects of the argument realisation to result only from the event template structure. The realisation of the arguments appears to follow the event schema (the relations of importance between the subevents are also maintained in syntax), but the root (idiosyncratic) meaning also defines the argument realisation.

2.1.3 The derivational basis issue: the relationship between the derivation of the transitive and intransitive types Various approaches (theoretical and typological) that correlate the two members of the alternation and unify them into one lexical item have been developed for the relationship between the derivation of the transitive and of the intransitive members of the alternation. The question that arises concerning these analyses is which of the two -the transitive or the intransitive type- is the basic one. The noncorrespondence between the semantic and the morphological derivation has, as a result, theoretical disagreements concerning the basic and derived relationship between the transitive and intransitive types. There are four main proposals in relation to the question of whether the transitive or the intransitive type is the basic one and which type correspondingly is the derived type: (a) The morphological derivation reflects the semantic derivation. Haspelmath (1993: 105) considers that the verbs with different morphological patterns (causative or anticausative morphological pattern26) also have different semantic representation. According to this analysis, three classes of change-of-state verbs must be distinguished on the basis of morphology. Additional causative morphology ĺ meaning: [EVENT y ANTICAUS] Additional anticausative morphology ĺ meaning: [EVENT x CAUSE [y ANTICAUS]] Non-alternating causatives ĺ meaning: [EVENT x CAUSEh [y ǹȃȉǿCAUS]] It cannot be suppressed The analysis by Clairis & Babiniotis (1999) of Modern Greek comes close to this approach. The anticausative verbs are distinguished into two types: verbs with the intransitive verb as the basic type (for example: șȣȝȫȞȦ thimono ‘to become angry’ – they name them ‘causatives’) and verbs with the transitive verb as the basic type (for example: ĮȞȠȓȖȦ anigho ‘to open’ – they name them ‘ergatives’), depending on the frequency of use of the intransitive or the transitive type and the acceptance of the transitive and the intransitive structure by native speakers. (b) There is no derivational relationship between the causative and the anticausative type. The 26

In the sense of additional morphology in the causative or anticausative member of alternation, respectively.

Theoretical Framework

43

transitive and intransitive types (ıʌȐȦ spao ‘to break’) are connected semantically, but no type is derived from the other (Ȃarantz 1984: 181, 5.27). The verb ıʌȐȦ spao ‘to break’ has two sub-inputs in the lexicon, a transitive and an intransitive: their difference consists in the feature [+/- logical subject], which refers to whether the lexical input selects the logical subject (= agent) or not. According to Marantz, (i) both the transitive and the intransitive types are basic and non-derived and (ii) one of the two types is selected arbitrarily as the morphologically basic one. Alexiadou, Anagnostopoulou & Schäfer (2006) also maintain that there is no derivational relationship between the members of the transitivity alternation. Their approach is based on the distribution of the PP that denotes the external argument (agent, cause, instrument, and cause-event) in the passive and anticausative structures, and in the differences between the verbs that participate in the transitivity alternation in English, German, and Greek. The cross-linguistic differences that appear make up the evidence against the approach of lexical causativisation and anticausativisation. Alexiadou, Anagnostopoulou & Schäfer base their analysis on a syntactic decomposition of change-ofstate verbs in the components of VOICE and CAUSE. The cross-linguistic differences depend on the features of VOICE, the combination of VOICE with CAUSE and the various types of verb roots27. (c) The basic type is the intransitive (Lakoff 1968, 1970; Marchand 1969; Brousseau & Ritter 1991; Pesetsky 1995); the transitive type has an additional semantic constituent, that of CAUSE. The transitive type is derived from the intransitive since it is semantically more complex. The transitive/causative type contains the anticausative meaning, whereas the intransitive/anticausative type does not contain the causative meaning. According to this approach, the anticausative unaccusatives (that participate in the alternation) do not differ semantically from the remaining unaccusative verbs (that do not participate in the alternation). (d) The basic form is the transitive (Chierchia 1989/2004; Levin & Rappaport Hovav 1995, 2005); the intransitive type is derived from the transitive through the lexical derivational process of detransitivisation. Both the transitive and the intransitive types have complex semantic representation. [x CAUSE [y BECOME broken/open]] agent patient The verbs that participate in the alternation have unified lexical semantic representation that is connected with both their unaccusative intransitive and their causative use, and this lexical semantic representation is the causative (dyadic-complex event structure). In this analysis, I am going to support the presence of v similarly in the causative and anticausative types (see paras. 2.1.1.2 and 4.2.2). In intransitive structures, the argument that is subcategorised from the semantic constituent CAUSE (the agent) does not appear in the argument structure because in the mapping of the lexical semantic 27

In contrast with the cross-linguistic variation of intransitive structures, it is possible for the external argument in causative structures to be agent, instrument or cause in all of the languages under examination. According to Alexiadou, Anagnostopoulou & Schäfer, Voice is responsible for the insertion of the external argument and bears features that are related with the agent and the manner of expression of the verb action. VOICE [+agent] licenses the presence of agent, whereas VOICE [-agent] licenses the presence of CAUSE. In the anticausative structures, either VOICE is absent or there is VOICE [-agent] (an implied cause). Languages are divided into languages that allow the presence of VOICE [-agent] in passive structures, whereas VOICE is absent in anticausative structures, and languages that have an obligatory VOICE [+agent] head in passive structures, whereas in the anticausative structures they can either have no VOICE or have VOICE [-agent]. Modern Greek belongs to the second group of languages, and Modern English and German belong to the first. -Modern Greek: passive structures with PPs denoting the agent, and not the cause: (1) a.

b.

ȉĮ ȝĮȜȜȚȐ ıIJİȖȞȫșȘțĮȞ Įʌȩ IJȘȞ țȠȝȝȫIJȡȚĮ ta malia steghnothikan apo tin komotria the.NOM hair.NOM dry.NACT.PAST.PERFVE.3PL by the hairdresser ‘The hair was dried by the hairdresser’ *ȉĮ ȝĮȜȜȚȐ ıIJİȖȞȫșȘțĮȞ Įʌȩ IJȠȞ ȒȜȚȠ *ta malia steghnothikan apo ton ilio the.NOM hair.NOM dry.NACT.PAST.PERFVE.3PL from the sun ‘The hair dried from the sun’

-Modern Greek: anticausative structures with PPs denoting the cause (the presence of cause is not obligatory) (2) ȉȠ ȕȠȪIJȣȡȠ ȑȜȚȦıİ to vutiro eliose the.NOM butter.NOM melt.ACT.PAST.PERFVE.3SG ‘the butter melted from the heat’

(Įʌȩ IJȘ ȗȑıIJȘ) (apo ti zesti) (from the heat)

44

Chapter Two

representation to the argument structure, the position of CAUSE is bound. The binding of a position in the lexical semantic representation blocks the projection of this position in the argument structure. Cherchia (1989/2004) assumes that the participation of the unaccusative intransitive verbs (intransitive ıʌȐȦ spao ‘to break’) in the transitivity alternation shows that the unaccusative intransitive verbs are derived from ‘dyadic’-complex-causatives verbs, whereas the unergative intransitive verbs (that do not participate in the alternation – intransitive ȖİȜȐȦ jelao ‘to laugh’) are ‘monadic’- simple28. Levin & Rappaport Hovav (1995) agree that the unaccusative intransitive verbs that participate in the alternation (intransitive ıʌȐȦ spao ‘to break’) are dyadic-causatives, but they differ from Chierchia since they do not expand the aforementioned analysis to those unaccusative intransitive verbs that do not participate in the alternation (e.g., ȣʌȐȡȤȦ iparxo ‘to exist’). I will return analytically to the issue of whether or not there is a link between the causative and the anticausative structures when I present the diachronic data. According to the data (micro-variation of the Greek diachrony), I will attempt to strengthen my proposed analysis.

2.1.4 Morphology, argument structure, and transitivity In many languages, aspects of verb meaning and argument structure alternations are marked overtly by the derivational morphology of the verb. Even if the limits between derivational morphology and inflectional morphology are not always clear (Anderson 1985; Comrie 1985), derivational morphology is usually considered to be an element of the lexicon, whereas the inflectional morphology belongs to syntax (functional categories TP and AgrP; Chomsky 1995). The verb derivational morphemes are phonologically overt or zero and mark processes of changes in verb transitivity, rearrangement, addition or omission of verb arguments: locatives with prefixes in German (31) a. Hans pflanzt Bäume im Hans plants trees in-the b. Hans be-pflanzt den Garten Hans BE-plants the garden

Garten garden mit Bäumen with trees

omission of the object with the derivational marker (ending -sja) in Russian (32) a. Sobaka The-dog b. Sobaka The-dog

kusajet poþtal’ona bites the-postman kusajet-sja bites-SJA

valency augmentation with the use of the causative morpheme in Turkish (33) a. Hasan öldü Hasan.NOM die.PAST.3SG ‘Hasan died’ b. Ali Hasan-ı öl-dür-dü Ali.NOM Hasan-ACC die-CAUS-PAST.3SG ‘Ali killed Hasan’

The causative types in Modern Greek bear steadily active morphology, as do the majority of transitive verbs –with the exception of the deponents (which do not have any active morphology and are transitive non-active). Owing to the stable presence of the active voice morphology in causatives of Modern Greek, the morphology of causatives does not especially concern the analysis of Modern Greek29. Many recent analyses of the verb in Modern Greek within the Generative Grammar perspective give special emphasis to the morphological marking of the anticausative types (active or non-active depending on the verb) and base their subcategorisations and analyses of transitivity alternation on the morphology of the anticausatives30. Taking as basic criteria the presence/absence of the agent and cause, Theophanopoulou-Kontou (1999b, 2000a, 2004) differentiates the anticausatives of Modern Greek as follows: (a) [-agent, -cause]: ĮȞșȫ antho ‘to blossom’, ĮțȝȐȗȦ akmazo ‘to flourish’. 28 29 30

Reinhart (1991) agrees with Chierchia’s view. For the deponent verbs in Modern Greek, see Zombolou (1996, 1997). Of special interest is the fact that the anticausative verbs are dealt with by Tzartzanos (1946) as a class of passive verbs but with active morphology (Tzartzanos’s examples include the following: ĮȞȐȕȦ anavo ‘to light up-intransitive’, ıȕȒȞȦ zvino ‘to wipe out’, țȡȣȫȞȦ kriono ‘to get cold’, ʌĮȖȫȞȦ paghono ‘to freeze up’, ʌȡĮıȚȞȓȗȦ prasinizo ‘to become green’, ĮȡȡȦıIJĮȓȞȦ arosteno ‘to fall sick’, ȜȚȫȞȦ liono ‘to melt’). Their particular meaning, in other words the meaning of change-of-state, is not indicated by Tzartzanos.

Theoretical Framework

45

(b) [-agent, +cause]: ʌİșĮȓȞȦ petheno ‘to die’, ȝĮȡĮȓȞȠȝĮȚ marenome ‘to wither’, ȝȠȣȤȜȚȐȗȦ muxliazo ‘to mould’, țȠțțȚȞȓȗȦ kokinizo ‘to redden’, with gradation concerning the possibility of the metaphorical presence of agent: (34) ȝİ ʌȑșĮȞİȢ me pethanes me.ACC.WEAK die.ACT.PAST.PERFVE.2SG ‘You killed me’ (35) *ȝİ ȝȐȡĮȞİȢ *me maranes me.ACC.WEAK wither.ACT.PAST.PERFVE.2SG ‘*You withered me’

Theophanopoulou-Kontou does not differentiate syntactically the active (ȝȠȣȤȜȚȐȗȦ muxliazo ‘to mould’) and the non-active anticausatives (ȝĮȡĮȓȞȠȝĮȚ marenome ‘to wither’), but with semantic criteria observes that the non-active ending denotes the totally affected, possibly in analogy to the reflexives and chiefly when the subject is animate. (c) [+agent, +cause]: when the anticausatives are used with an agent, they are no longer anticausatives but passives (morphological syncretism), and the non-active ending is obligatory (țĮșĮȡȓıIJȘțİ Ș ıțȐȜĮ / katharistike i skala / clean.NACT.PAST.PERFVE.3SG the.NOM stairs.NOM / ‘the stairs were cleaned’; ĮȞȠȓȤIJȘțİ Ș ʌȩȡIJĮ / anixtike i porta / open.NACT.PAST.PERFVE.3SG the.NOM door.NOM / ‘the door was opened’). Embick (1998) considers that voice morphology in Modern Greek anticausatives does not correspond to any syntactic element but belongs to the morphological level and is assigned postsyntactically to a Verb, when this Verb (or the combination of v and V) is not locally adjacent to the external argument (unification of passive, reflexive, and anticausative verbs; cf. Marantz 1984; Chomsky 1995). Embick’s proposal for syntactic alternations is represented schematically below: V ĺ V – VOICE [mediopassive]/_DP non-external argument Tsimpli (2006a) argues that Voice is a feature of v only in cases of non-active morphology. The anticausative verbs that do not bear non-active morphology differ in relation to features of Transitivity/agent that are located in v, independently of Voice. The anticausatives with active morphology comprise unaccusative structures and the sole argument moves to the [Spec, TP] position. (36) a. ȉȠ ʌȠȣțȐȝȚıȠ ȜȑȡȦıİ to pukamiso lerose the.NOM shirt.NOM dirty.ACT.PAST.PERFVE.3SG ‘The shirt became dirty’ b. [TP IJȠ ʌȠȣțȐȝȚıȠ [vP ȜȑȡȦıİ [VP [DP IJȠ ʌȠȣțȐȝȚıȠ]]]] to pukamiso lerose to pukamiso the.NOM shirt.NOM dirty.ACT.PAST.PERFVE.3SG

ȉhe vP in anticausative structures does not have Transitivity or agent features and, as a result, the presence of an external ș-feature is not possible. Concerning non-active morphology, Tsimpli, following Kratzer (1996) and Alexiadou & Anagnostopoulou (2004), assumes that v/VOICE is included in the syntactic structure. The presence of non-active Voice has an effect on the verb transitivity. Voice locally attracts a ș-feature, which cannot be lexicalised as an independent argument (DP). A Logical Structure that correlates with the predicate arguments demands the interpretation of the șfeature (which the Voice has attracted) either as reflexive or non-reflexive (i.e. anticausative, passive, middle): [v/VoiceP DP [v/VoiceP v/ Voice [VP V ]]] ĺreflexive interpretation [TP DP [v/VoiceP v/Voice [VP V ]]] ĺanticausative, passive, middle interpretation In the first case, v has an agent feature, and this feature attracts the DP-subject to [Spec, vP]. In the second case, VOICE attracts the agent feature, i.e. the external argument. The distinguishing characteristic of Tsimpli’s analysis is that Voice attracts ș-roles (ș-attractor), and that the results in

46

Chapter Two

relation to Case are consequences of this function of Voice31. In the present analysis, I will separate voice morphology from the transitivity alternation derivation. My arguments for the dissociation of verb voice morphology (active or non-active) of anticausative types from the derivation, and the connection between causative and anticausative types are32: (a) The instability of the morphological marking of anticausative structures as it appears, for example, in the contemporary tendencies that can be observed in Modern Greek: (i) The observation of a tendency to extension of the non-active form in anticausative structures in Modern Greek (Theophanopoulou-Kontou 1999b): (37) a. ȁȑȡȦıİ lerose

Ș ȝʌȜȠȪȗĮ ȝȠȣ; but also ȜİȡȫșȘțİ i bluza mu lerothike

dirty.ACT.PAST.PERFVE.3SG the blouse.NOM my

dirty.NACT.PAST.PERFVE.3SG the blouse.NOM my

‘My blouse became dirty’ b. ǺȡȫȝȚıİ Ș ȝʌȜȠȪȗĮ ȝȠȣ; but also ȕȡȦȝȓıIJȘțİ vromise i bluza mu; vromistike soil.ACT.PAST.PERFVE.3SG the.NOM blouse.NOM my

Ș ȝʌȜȠȪȗĮ ȝȠȣ i bluza mu

Ș ȝʌȜȠȪȗĮ ȝȠȣ i bluza mu

soil.NACT.PAST.PERFVE.3SG the.NOM blouse.NOM my

‘My blouse became soiled’

(ii) A tendency (in the opposite direction) for substitution of many forms with non-active morphology with new intransitive active forms has been observed (Vassilaki 1988): (38) ʌȜȘșȪȞİIJĮȚ plithinete increase.NACT ȝİȚȫȞİIJĮȚ mionete reduce.NACT İȜĮIJIJȫȞİIJĮȚ elatonete diminish.NACT ȕİȜIJȚȫȞİIJĮȚ veltionete ameliorate.NACT İʌȚįİȚȞȫȞİIJĮȚ epidhinonete aggravate.NACT

– – – – –

ʌȜȘșĮȓȞİȚ plitheni increase.ACT ȝȚțȡĮȓȞİȚ mikreni shorten.ACT ȜȚȖȠıIJİȪİȚ lighostevi lessen.ACT țĮȜȣIJİȡİȪİȚ kaliterevi improve.ACT ȤİȚȡȠIJİȡİȪİȚ xiroterevi worsen.ACT

It should be noted that cross-linguistically, as well, there is disagreement between scholars of the French language, as to whether the use of the clitic se, or of the active endings in the anticausative structures, is undergoing development (cf., for example, Lagane 1967 and Rothemberg 1974). We find trends in both directions33: (39) a. Ce tableau a noirci the.NOM table.NOM has.3SG blackened ‘This table blackened’ b. Ce tableau s’ est noirci the.NOM table.NOM SE is.3SG blackened ‘This table blackened’ (rare; more frequent in the past) (Rothemberg 1974: 192) (40) a. La branche a cassé / s’ est cassée the.NOM branch.NOM has.3SG broken / SE is.3SG broken ‘The branch broke’ b. Le vase s’ est cassé / a cassé the.NOM vase.NOM SE is.3SG broken / has.3SG broken ‘ȉhe vase broke’ c. La crème a durci en se refroidissant the.NOM cream.NOM has.3SG hardened while SE cool.PART cǯ. Le suif s’ est durci en refroidissant 31

32

33

In contrast with Embick, who observes syncretism of all the functions (anticausative, reflexive, passive, middle) due to morphological underspecification, Tsimpli connects the morphological underspecification with the absorption of either the external argument (non-reflexive) or the internal argument (reflexivity). I think that Noam Chomsky’s view (personal communication) comes close to this approach: ‘In the hard sciences, these ubiquitous problems are avoided by keeping largely to the results of carefully constructed experiments, or else just putting anomalies to the side –not small ones either. In the study of language, what we observe are some extreme amalgam of historical accident and innumerable other factors. Particularly in morphology. That’s why, for example, Jespersen said that one can hope to discover a universal grammar for syntax, but not morphology’. In relation to the tendencies in Modern Greek and Modern French, as well as diachronic changes, I would mention Kemmer’s analysis (1993) (based on the approach of Haiman 1983) which distinguishes two contradictory motives in the field of change: first, the motive of economy, which states that the marking of semantic properties that are inherent in verb action must be avoided (resulting in unmarked intransitive verbs), and second, the motive of expressiveness, which contrasts with that of economy and is reflected in the extension of the non-active morphology and of the clitic.

Theoretical Framework

47

the.NOM lard.NOM SE is.3SG hardened while cool.PART ‘The cream / The lard hardened when cooled’ (Le Petit Robert 2000)

(b) The intense dialectical, diachronic, and cross-linguistic morphological variation concerning the marking of the anticausative (intransitive) member of the alternation: (i) Dialectical differences a. Anastasiadis (1976: 103) Cappadocian idiom spoken in Pharasa İȝȠȪ-ȝĮȚ emu-me fill.NACT ıİʌȠȪ-ȝĮȚ sepu-me rot.NACT ȜȚȠȪ-ȝĮȚ liu-me melt.NACT b. Drettas (2005) Pontic İȞȣȤIJȫ-șĮ enixto-tha be-overtaken-by-night.NACT İȕȡȐįȗ-ıIJĮ evradhz-sta be-overtaken-by-dark.NACT

Standard Modern Greek Ȗİȝȓȗ-Ȧ jemiz-o fill.ACT ıĮʌȓȗ-Ȧ sapiz-o rot.ACT ȜȚȫȞ-Ȧ lion-o melt.ACT Standard Modern Greek ȞȪȤIJȦ-ıĮ34 nixto-sa be-overtaken-by-night.ACT İȕȡȐįȚĮ-ıĮ35 evradhia-sa be-overtaken-by-dark.ACT

(ii) Differences in different periods of the Greek language. In relation to the Modern Greek intransitive psych-verb șȣȝȫȞȦ thimono get-angry.ACT, cf. the Medieval Greek intransitive psych-verb șȣȝȫȞȠȝĮȚ thimonome get-angry.NACT. (41)ȱ a. ϳȱȱ Ά΅Η΍ΏΉϿΖȱȱ ΌΙΐЏΑΉΘ΅΍ȱȱ ȱȱ ȱȱȱȱȱȱȱȱȱȱȱȱ o vasilefs thimonete the.NOM king.NOM get-angry.NACT.PRES.3SG ‘the king gets angry’ (Kallimachos and Chrysorroi, 2438; AD 14) ȱ b.ȱ Θϲȱȱ ΩΏΓ·ΓΑȱȱ БΖȱόΎΓΙΗΉΑȱΐΉ·ΣΏΝΖȱȱ πΌΙΐЏΌ΋ to aloghon os ikusen meghalos ethimothi the.NOM horse.NOM as heard.3SG a-lot get-angry.NACT.PAST.PERFVE.3SG ‘the horse got angry a lot as heard about these’ (An Entertaining Tale of Quadrupeds, 745; AD 14) ȱ c.ȱ ΘϱΘΉȱȱψȱȱ ΜΙΛχȱȱ πΌΙΐЏΌ΋ΎΉΑȱ tote i psixi ethimothiken then the.NOM soul.NOM get-angry.NACT.PAST.PERFVE.3SG ‘then the soul got angry’ (Leonardo Dellaporta, Poems, ǹ, 191; AD 14-15) ȱ d.ȱ Ύ΅ϠȱȱΔΣΕ΅ΙΘ΅ȱȱ πΌΙΐЏΌ΋ȱȱ Ύ΅ϠȱΉϢΖȱΚΙΏ΅ΎχΑȱȱ ΘϲΑȱȱ σΕΕ΍ΜΉΑ   ke parafta ethimothi ke is filakin ton eripsen  and immediately

ȱ

get-angry.NACT.PAST.PERFVE.3SG and in prison

he.ACC.WEAK threw.3SG

‘and he immediately got angry and threw him into prison’ (Leonardo Dellaporta, Poems, ǹ, 2265; AD 14-15) e.ȱ ΉΌΙΐЏΌ΋ΎΉΑȱȱ Γȱȱ Ά΅Η΍ΏΉϾΖȱȱ Γȱȱ ΐν·΅Ζȱȱ ȱ

ethimothiken o vasilefs o meghas get-angry.NACT.PAST.PERFVE.3PL the.NOM king.NOM the.NOM great.NOM ‘the king the great got angry’ (Apollonius of Tyre, The story of Apollonius (Kechagioglou), 204; AD 14) ȱ f.ȱ ϳȱΊ΅Ε΍ΗΘχΖȱ Φ·΅Α΅ΎΘκǰȱȱ ΌΙΐЏΑΉΘ΅΍ǰȱȱ ΐ΅ΑϟΊΉ΍ȱȱ o zaristis aghanakta, thimonete, manizi the dice-player.NOM become-irritated.ACT.PRES.3SG get-angry.NACT.PRES.3SG get-mad.ACT.PRES.3SG

‘the dice player became irritated, got angry and mad’ (Sachlikis, Paraenetic poem, 117; AD 15-16)

(iii) Cross-linguistic differences: Modern Greek: ȑıʌĮıİ espase / break.ACT.PAST.PERFVE.3SG German: zerbrechen / 34

35

*ıʌȐıIJȘțİ spastike break.NACT.PAST.PERFVE.3SG *sich zerbrechen

For some speakers the non-active form can be also grammatical: ȞȣȤIJȫșȘțĮ nixto-thika be-overtaken-bynight.NACT. For some speakers the non-active form can be also grammatical: ȕȡĮįȚȐıIJȘțĮ vradhia-stika be-overtaken-bydark.NACT.

48

Chapter Two Italian: French:

*rompere casser

/ /

si rompere se casser

(c) Moreover, the active or non-active morphology of anticausatives exerts no influence at all on the distribution of the PP in anticausative structures (Alexiadou, Anagnostopoulou & Schäfer 2006). The PPs, which appear with both active and non-active morphology, are: Įʌȩ ȝȩȞȠȢ IJȠȣ-ȝȩȞȘ IJȘȢ / apo monos tu/-oni tis / ‘by himself/herself’ (denoting cause or instrument, but not agent).

2.1.5 Summary I assume, according to the above, that three parameters must be taken into consideration in every analysis of transitivity and causativity: (a) the precise syntactic representation of the transitive and intransitive verbs, (b) the semantic decomposition that allows the mapping between semantic verb classes and syntactic structures, and (c) the voice morphology of the transitive and intransitive (anticausative, passive) types. It is possible to predict the syntactic representation of arguments to a large degree from the verb meaning. I follow the approach of Levin & Rappaport Hovav, which comprises a ‘bridging’ attempt between the syntactic analysis and the lexical-semantic study of verb meaning. Concerning syntactic representations, I follow the Minimalist Programme and the analyses that consider verb transitivity as important in recognising v as a syntactic head, a functional category between the categories of Tense and Verb with the following properties: (a) its Specifier inserts the subject and (b) its head contains the agreement features with the object. I hypothesise that this functional head exists in the causative verbs (meaning change-of-state) but does not exist in the accusative verbs (transitive non-causatives). In this way, we can syntactically translate the distinction between causative and accusative transitive verbs that we described in the first chapter (1.2 and 1.3). Since v is connected with causativity features, I hypothesise that it is also present in the anticausative structures (following Harley 1995; Collins 1997; Marantz 1997; Embick 1998; Arad 1999; Travis 1999; Alexiadou 2001; Alexiadou & Anagnostopoulou 2004). I adopt the syntactic distinction between causatives (v [+agent]), anticausatives (v [-agent]) and passives (presence of v and Voice) which is responsible for the absence of a direct object in the accusative case and for the Įʌȩ apo ‘by’-PP of the agent). The aforementioned structures comprise the working hypothesis for this monograph and hold good in cases where a close connection between the causatives and the anticausatives can be proven. v in anticausatives is [-agent], whereas v in the causatives is [+agent]; through a lexical process, the insertion of a DP-agent is blocked in anticausative structures with the result that the [Spec, vP] position is not realised in the initial structure. The head that is responsible for the insertion of the direct object is Voice, which is found higher than vP and VP. It also has two abstract types (in complementary distribution); the first type selects an external argument [+external argument] and the second type does not select an external argument [external argument]. The non-active morpheme of passive types that absorbs the accusative case is checked in the head of Voice. I follow Collins (2005) for the analysis of passive types and the presence of the non-active morpheme in parallel with the absence of DPs in the accusative in Modern Greek. In this way, we can unite the ș-role assignment (agent or non-agent; the common feature of prototypical transitives and passives is the possibility of the presence of an external argument (agentivity feature) as DP or PP) to the external argument in the [Spec, vP] position in both the transitive and the passive type. The external argument is generated in the [Spec, vP] position in the passive type, as happens exactly with the transitive type, whereas the non-active morpheme only blocks the accusative case assignment. v of passives has, of course, different features from v of transitives since it is subcategorised for the presence of the external argument-PP (agent or non-agent) in the Specifier of its Phrase; v of causatives is connected to causativity and the presence of DP-agent or non-agent as subject and DP-patient as object. I follow the presence of the lexicon-syntax interface, and the hypothesis of the existence of verb classes that are defined semantic-syntactically (Levin 1993; Levin & Rappaport Hovav 1995, 2006) and the representation of verb meaning through semantic decomposition (not through ș-role hierarchy; Levin 1993, 1999). The decomposition of the predicate comprises two basic types of semantic features: the primitive predicates and the roots. The semantic relations of the verbs with their arguments are defined in relation to the parts that arise from the semantic decomposition. The decomposition of the predicates happens in such a way that the verbs that belong to the same semantic class have common parts in their decomposition. The change-of-state verbs have as a common element the part of the structure that is comprised of the primitive semantic component BECOME and one root that names the specific situation that arises as a result. The roots in lexical semantic representation differentiate the members of the specific semantic verb classes.

Theoretical Framework

49

The majority of unergative intransitive verbs are ACTIVITIES, the anticausative intransitive verbs are ACHIEVEMENTS, and the alternating and non-alternating causative verbs are ACCOMPLISHMENTS, but there is no one-to-one correspondence between transitivity and event types since some transitive non-alternating verbs are activities and some unergatives with the presence of a PP are accomplishments. In any case, the different verb classes (transitive, intransitive verbs that do or do not participate in the transitivity alternation) vary in relation to the number of subevents and the ș-roles. By modifying the analyses of Levin & Rappaport Hovav (1995) and Chierchia (1989/2004) and putting an emphasis on the Greek diachronic data and on the central role that I assume is played by transitivity, I have ascertained the following correspondences between verb syntax and verb meaning: (a) causative non-alternating verbs (įȠȜȠijȠȞȫ dholofono ‘to murder’): two subevents, [+specified] agent; (b) causative alternating verbs (ıʌȐȦ spao ‘to break’): two subevents, [-specified] agent; in the intransitive structure suppression of the agent; (c) intransitive non-alternating verbs (ȣʌȐȡȤȦ iparxo ‘to exist’): two subevents; and (d) intransitive unergative verbs (țȠȜȣȝʌȐȦ kolimbao ‘to swim’): one subevent (presence of an agent, internal cause for the change-of-state). I separate voice morphology that marks causative and anticausative verbs from the syntactic representation and the semantic-syntactic mapping. The diachronic, dialectical, and cross-linguistic instability leads us to assume the dissociation of voice from the derivation of causative and anticausative types. The unstable morphological marking of anticausative structures is imprinted, for example, in the contemporary tendencies that can be observed in Modern Greek, where there is a tendency towards extension of the non-active ending in some anticausative types but also towards replacement of many forms with non-active morphology from new active anticausative forms. The active or non-active morphology of anticausatives does not, however, influence the PP distribution in Modern Greek: the PPs Įʌȩ ȝȩȞȠȢ IJȠȣ-ȝȩȞȘ IJȘȢ / apo monos tu-moni tis / ‘by himself / herself’ (denoting cause or instrument, but not agent) appear with both active and non-active anticausatives. All of the above clearly only comprises the framework of the analysis, the working hypothesis, and the basis on which this study is built. The perspective I described outlines the theoretical approaches and empirical analyses that I took into consideration during the study of the data. It leaves unanswered, however, the following questions, which will also comprise a goal in the present monograph: (a) What is the precise role of voice morphology in the diachrony (stable or dissimilar over chronological periods)? Does it stably encode lexical or syntactic information or not? (b) Do all semantic verb classes exhibit the same syntactic behaviour? What do the event templates and the wealth of diachronic data indicate for the derivation of the transitive from the intransitive type or vice versa? (c) Can the proposal (i) for syntactic differentiation of the passives and presence of the agent argument in the passive syntactic structures and (ii) for correlation of the causative and anticausative types be supported by diachronic data?

50

Chapter Two

2.2 Language Change and Transitivity The present section aims to present systematically the theoretical position adopted in this diachronic study and the perspective from which the diachronic development of the structures under examination is analysed. The issues that are touched on in this chapter are the Generative Grammar approach to language change and, specifically, Lightfoot’s theory (1979, 1999, 2006a) for the reanalysis and the connection between language acquisition and change through cues (see para. 2.2.3). The section finishes with the critical review of former diachronic studies related to issues that touch generally on Greek verb transitivity and lay the basis for every diachronic analysis of transitivity in the Greek language.

2.2.1 Generative Grammar and diachrony If we accept that the theory of grammar means the defining of the possible grammars of a natural language (Chomsky 1965, 1981, 2000), we put in place a higher constraint on possible diachronic changes: no grammar can be changed into a grammar that is not possible for a natural language. Following the approach developed within the framework of Generative Grammar theory, chiefly by Lightfoot (1979, 1991, 1999, 2006a, b), we hypothesise that the changes are accomplished when speakers (a generation of speakers) who acquire the language end up with a linguistic system that differs (in one or more aspects) from the grammatical system corresponding to the linguistic competence of the ideal speaker (I-language) of the previous generation, in other words, of the generation that provides input to speakers who acquire the grammatical system1. When the new generation of speakers replaces the older generation, then the change is carried over to the entire linguistic community. Consequently, the differences between chronological periods and the changes are not due to biological factors but to factors that are concerned with the input: if children hear input that is different to a certain degree, it is possible for them to form a different I-language system. Every child acquires an I-language, in other words a specific form of Greek, and not an External-language (E-language), namely Greek in its entirety. Changes in the E-language cause changes in the I-language (if people receive different input, they will acquire a new grammar), and vice versa. (a) Basic principles Concerning language change, I will follow the fundamental principles below (Lightfoot 1979; Aitchison 1991; MacMahon 1994; Theophanopoulou-Kontou 2000b; Roberts & Roussou 2003): (a) The spread of a change is a result of endogenous opacity or non-regularity of the system, and has, as a consequence, the simplification of grammar (limitation of the irregular/optional choices) or the creation of new patterns and contrasts (Lightfoot 1979). Lightfoot’s Transparency Principle requires the structures to be as simple as possible, and in this way, it refers to the properties of possible grammars2. The problematic relationship between the language system and its various realisations can lead to the reinterpretation and reanalysis of structures or to the gradual prevalence of other semantically or functionally corresponding structures. The emergence of change and reanalysis comprises a borderline point for grammar and indicates when the Transparency Principle3 intervenes. 1

2

3

I-language is the mental system that characterises the linguistic ability of humans and is represented in the brain of every person. It is the linguistic organ of humans, the system (Lightfoot 2006a: 7). The E-language is found outside in the real world, and it does not comprise a formed system. It constitutes the function of various grammars that people carry within them and can be influenced by frequency factors (for example, some people can use topicalisation more frequently than other people) or from non-systematic, ‘fixed’ expressions that do not correlate with other grammatical processes. The Transparency Principle has been referred to for other issues in the diachrony of the Greek language; cf. Theophanopoulou-Kontou 2000: “Locative adverbs and Case in Greek. A diachronic approach”, Glossologia 11-12, 1-40. The Transparency Principle was formulated within the Extended Standard Theory: ‘the structures must not be complex as far as it is possible, and the initial (deep) structures must be, as far as possible, as close to the surface structures’ (Lightfoot 1979: 121). The degree of exception can increase up to the point where it violates the Transparency Principle by exceeding the accepted level of complexity. Various problems have been noted concerning Lightfoot’s analysis, e.g. that he does not give a precise definition of transparency and opacity. For Harris & Campbell (1995), it is the ambiguity of the structures, not the opacity, that is the essential presupposition for reanalysis. Harris & Campbell (1995: 50) define reanalysis as the mechanism that alters the deep structure of a syntactic schema and does not include any modification of the surface form. The surface manifestations of change, such as a change in word order or morphological changes, occur after deep

Theoretical Framework

51

Language change is ‘therapeutic change’ and is essential as a ‘therapy’ of vagueness, whereas the only formal distinction between possible and non-possible change derives from the theory of grammar. My analysis differs from Lightfoot’s approach concerning the dynamic view of change. Lightfoot argues that the changes do not happen gradually but suddenly. For Lightfoot, the gradual pace of grammatical change can only be superficial; the spread of a change from area to area and from language group to language group -but not the grammatical change itself- can be gradual. Grammatical change tends to be irregular and ‘catastrophic’, and manifested by a group of phenomena that change simultaneously. I will follow as a working hypothesis the view that the diachronic approach presupposes a dynamic approach of language (Romaine 1981; Aitchison 1991; MacMahon 1994; Kakridi-Ferrari & Chila-Markopoulou 1996), together with a framework of analysis in which mainly internal factors play an essential role in change. Dynamic tendencies between the old and the new constitute the history of every language, and parallel forms shape the circumstances for subsequent changes (cf. Kroch 1989, 2000). (b) Diachronic study can describe language structures with the same methodology that modern analyses of Generative Grammar use (Kroch 1989, 2000; Haeberli 2002; Eythórsson 2003): grammar can be described as it was before and after a specific change. The comparison of different grammatical systems and the interpretation of their differences are the chief pursuit. Consequently, knowledge of two (or more) synchronic stages is needed. To find out what changed between chronological period a and chronological period b, we need to know what both of these periods were like as synchronic periods. A diachronic analysis can be divided into three parts: (i) the description of a series of synchronic periods; (ii) a comparison of a series of synchronic periods; and (iii) interpretation of the differences. Moreover, it is important to make a distinction between the two terms, grammar and language, since I am going to consider grammar as a system in the mind of humans (language competence, I-language), whereas language is the organised system that humans use in communication (language performance, E-language). (c) Following Clark & Roberts (1993: 202-209) and Roberts & Roussou (2003), I accept that the mechanism of acquisition prefers relatively simple representations. If the input is ambiguous, the child who acquires the language will choose the analysis of data that entails the simplest representation. For example, loss of movement in syntax leads to a reduction in complexity (Kayne 1994: the mechanism of movement is adjunction), in other words, to a simplest representation. If the child who acquires the language assumes that there is no movement, the preference of grammar for simplification is satisfied. The simplifications of grammar are always ‘local’ (they concern only a very specific and limited grammatical field), but they can lead to an increase in complexity elsewhere in the system. The simplification follows the preferable choice of the grammatical system which is for a one-to-one representation of features and lexical items to exist; if a feature has to be realised then the use of the same form of realisation for more than one feature is avoided. This approach explains why morphology is so important for syntactic change. The cause of a change must be sought in language indicators (triggers for Roberts & Roussou, cues for Lightfoot) that lead to the presence of a more complex system due to morphology. (b) Research areas: the diachronic study as comparative research There are two traditional study areas in diachronic Generative Grammar. The first area of study is the grammar of ancient languages (Classical Greek, Old English, Old High German). Studies in this area attempt to discover the principles that govern the internal grammar of speakers in the earlier phases of a language. It is a question, therefore, of synchronic analysis of earlier material. The aim of the research is not a change. The term 'diachronic' corresponds here to the diachronic character of the linguistic material that is being studied. The other main area of study is the comparison of earlier and later stages of the same language. Pintzuk (2003: 509), for example, defines diachronic syntax as the comparison of two stages of the same language and consequently as a branch of comparative syntax. According to this point of view, there is no great difference between the study of Classical Greek in relation to Modern Greek and of Modern Greek in relation to Korean. Only in the first case do we talk about changes, whereas in the second case, we do not. According to the latest approach, we can regard diachronic Generative Grammar as a branch of comparative syntax that is concerned with temporally correlated dialects (micro-variation). The basic difference between diachronic Generative Grammar and comparative studies is that the empirical data

reanalysis has happened. In relation to the question of why the structures do not return to their previous form in a given chronological period, the interpretation is given in relation to internal pressures for change (Keenan 1996: 3l; Longobardi 2001: 278); grammatical changes occur only when a reason exists that can cause grammatical change (such as change in the lexicon that leads to reanalysis).

52

Chapter Two

in diachronic study is E-language, whereas the goal of comparative study is I-language (Halle 1998). This fact is connected, of course, to a methodological problem: how can the E-language of texts (that have survived) be analysed in relation to the I-language (and knowledge) of speakers? The answer to such an important question can be supported by the admission that aspects of the data of the Elanguage can be interpreted as the output of the grammatical system in cases where we are able to set aside the effects of realisation parameters. The importance of tendencies (ongoing changes) and diachronic changes in the E-language that appear in the same chronological period are, in reality, far more widespread than they appeared above: the analysis of variation regarding changes in the E-language, and mainly the tendencies in the same chronological period, can reveal differences in the I-language. As long as a language changes, we can observe the course of the change, and this course supplies us with information about the nature and the organisation of the grammar (of a specific language or also of language in general), something that is not possible through a synchronic comparative study. This is the reason why recent studies in diachronic syntax place an emphasis not only on the sudden changes of parameters but also on synchronic variation and the consequences of gradual grammatical change (Kroch 1989, 2000). The detailed study of patterns of variation is one method for drawing conclusions in relation to the Ilanguage based on what is observed from the E-language. Moreover, some of the problems of diachronic study (limited data of everyday language, inability to examine the (un)grammaticality of certain structures) can be avoided if we examine the change as it develops (ongoing change) since, at least for one part of the transitory period of changes, there are native speakers (cf. Henry 1997).

2.2.2 Language change, language acquisition, and reanalysis The view that comparative grammar and diachronic grammar are identical fields is closely related to a dominant theory about change in Generative diachronic syntax, namely the approach concerning the link between change and acquisition that was first outlined by Lightfoot (1979). According to this approach, language changes because children, motivated by their exposure to different data, acquire a different grammar from that of their parents. The theory of language change is dependent on the theory of language acquisition that we accept. The idea of examining mainly language acquisition for the interpretation of language change was increasingly accepted after it was first overtly stated in David Lightfoot’s book, Principles of Diachronic Syntax (1979). Lightfoot opposes the ideas regarding the link between language change and the teleological development of language that were especially widespread in the 1970s. These theories presuppose that language change, even through many generations, follows a prescribed course. This idea, however, cannot under any circumstances be correct: every speaker refines his/her grammar anew; the child who acquires language and formulates his/her grammar has no knowledge concerning the history of the language and cannot follow a prescribed course. Lightfoot argues that a child who acquires language is genetically defined to construct a grammar of his/her language based solely on the language of his/her environment4. This view, which begins with Anderson (1973), is presented schematically in Table 2.1: Table 2.1 Language acquisition and change: I-language and E-language Grammar 1 - - - - - - - -

Grammar 2

I-language

Output 1 - - - - - - - - -

Output 2

E-language

If we replace Output 1 with the grammar of the parents (their E-language), what the diagram shows is that the child who acquires language constructs his/her own grammar (Grammar 2) based on Output 1. The child, of course, does not have access to the grammar (I-language) of his/her parents. There is only a relationship between Output 1 and Grammar 2, and there is no relationship between Grammar 1 and

4

Dimitra Theophanopoulou-Kontou (personal communication) contradistinguishes this view with the approach that the prescribed course may have a connection to the universality of the changes (if we accept that universality of changes exists). The difference in the approach followed in this monograph is that a prescribed course may exist but is only an epiphenomenon of the rules that a child follows when acquiring language and of the general rules that are valid for all languages. What such an approach excludes is the existence of rules that guide language development in one direction and not the other.

Theoretical Framework

53

Grammar 2. Grammatical changes occur when there is no concrete, explicit, visible relationship between the interpretation of the Output (E-language) and the Grammar (I-language). Grammatical phenomena cannot be acquired correctly if they are not reflected in language performance. It is evident that a grammatical feature that survives through many generations is clearly reflected in the Output (to be able to justify its acquisition by so many generations). If we observe change in one grammatical feature or in a structure, then something in the language performance of the previous generation must have changed and paved the way for reinterpretation of the feature or of the structure. For this reason, the questions that arise concerning language acquisition are particularly important for the understanding of the causes of language change: the direct cause of the change must be found in some alternation/change/innovation in the Primary Linguistic Data (PLD) that children who acquire language receive5. The significant data of linguistic experience are considered as PLD: the poverty-of-stimulus problem (for example, sentences (1a-b) are grammatical, but sentence (1c) is ungrammatical) is solved if we accept that children possess information that does not derive only from their linguistic experience (Lightfoot 2006a: 9). (1)

a. ȅ o the.NOM

īȚȐȞȞȘȢ Janis

ȤIJȪʌȘıİ xtipise

Janis.NOM

hit.ACT.PAST.PERFVE.3SG the.ACC hand.ACC

‘Janis hit his hand on the table’ b. ȅ īȚȐȞȞȘȢ ȤIJȪʌȘıİ o Janis xtipise the.NOM

Janis.NOM

IJȠ to

ȤȑȡȚ xeri ȤȑȡȚ xeri

IJȠȣ ıIJȠ tu sto

IJȠ to

ȤȑȡȚ xeri

hit.ACT.PAST.PERFVE.3SG the.ACC hand.ACC

IJȡĮʌȑȗȚ trapezi

his to-the table

IJȠȣ ʌȐȞȦ tu pano

hit.ACT.PAST.PERFVE.3SG the.ACC hand.ACC his

‘Janis hit his hand on the table’ c. *ȅ īȚȐȞȞȘȢ ȤIJȪʌȘıİ *o Janis xtipise the.NOM Janis.NOM

IJȠ to

on

to-the table

IJȠȣ IJȠ tu to his

ıIJȠ IJȡĮʌȑȗȚ sto trapezi IJȡĮʌȑȗȚ trapezi

the.ACC table.ACC

‘*Janis hit his hand the table’

Children possess the grammatical system of Universal Grammar and form their grammar (Ilanguage) by following PLD. Grammatical change is connected to the changes in the E-language since, based on the E-language, new I-languages arise according to the principles of Universal Grammar. The work of Diachronic Linguistics is to analyse and to compare (just like synchronic comparative linguistics) the different I-languages before and after the change and to connect the changes in the Ilanguage with the changes in the E-language. A structure or a category, however, can receive, during its diachronic course, marked features due to phonological changes, loss of morphology, or changes in word order. Children who acquire these structures are likely to analyse the structures that they observe in the speech of adults in a different way so that the structures correspond to a less complicated interpretation. A characteristic example are the verbs that were the predecessors of today’s English modal verbs: the creation of a class of modal verbs is owed to the concentration of particular distinguishing features (morphological, semantic, syntactic) in the predecessors of modal verbs in contrast to the rest of the verbs. The Transparency Principle was the reason that led to a change of category for these verbs and to their analysis in a different, new, and innovative way (reanalysis) when a high degree of obscurity was reached.

2.2.3 Input, cues, and PLD Children need linguistic experience, triggers, from the language of the preceding generation to be in a position to determine the features and the parameters of the language they acquire. Even if the number of features and parameters that need to be determined in a specific language is small and the choices for every feature and parameter are limited, children, according to this approach, must be able to consider all of the information in relation to all of the possible structures that correlate with every feature and parameter6. Consequently, it is obvious that the number of features to be acquired needs to be limited even further. It is precisely this role that is carried out by cues in Lightfoot’s theory: children recognise what to look for in the linguistic information that they receive, and they do not have to study every sentence individually to connect every sentence with every feature but have only to check the data that they receive for the designated structures or cues (Lightfoot 1999: 149, 259). 5

6

Of course, changes in PLD may be the result of other changes that have come before, such as loss or the weakening of overt morphology. Cf. Varlokosta 2005.

54

Chapter Two

Cues comprise the domains in which grammars are differentiated. Cues are found in the mental, abstract representations that are the result of comprehension and processing of the utterances of the children’s environment. When the child understands an utterance, s/he forms a mental, abstract representation of that utterance. A child who acquires language checks and processes the abstract representations that are derived from the input, looking for the cues. No cue is adopted in the grammar if it does not allow the syntactic analysis of the sentence to which it belongs and if it is not, in fact, required for the analysis of the particular sentence (Lightfoot 2006a: 85-86)7. The child who looks for the cues may or may not find them; the output of the grammar is entirely the result of the cues that the child finds; that is, the grammar is not formulated under any circumstances on the basis of the sentences that are heard. The utterances that children who acquire language hear comprise only the source for the cues (Lightfoot 2006a: 79). The total sum of cues that are found by the child acquiring a specific language can be considered similar to the lexicon. The child, in other words, recognises the cues in the same way that s/he recognises and acquires the lexical items of language (Lightfoot 2006a: 82). In what follows, I will present (as examples of analyses based on the cues) some elements from Lightfoot’s examination (2006b) of (a) the history of the modal auxiliary verbs in English (can, could, may, might etc.) and (b) the emergence of split genitives in Middle English. The creation of modal auxiliary verbs in English comprises an instance of change of a grammatical category; these specific verbs are initially generated in the VP, whereas after the change of the grammatical category, they are generated in the IP. The cause of this grammatical change can be looked for in earlier changes in the PLD: (a) the modal auxiliaries became separate morphological forms, the only members that survived from the class of preterite-present verbs and (b) the past tense form of those verbs overlapped in many instances phonologically with the subjunctive form; when the subjunctive form was lost, the past tense form survived, taking, however, the meaning of the subjunctive form instead of the past meaning. Universal Grammar provides a small number of grammatical categories, and lexical items are members of a grammatical category according to their morphological features and their distributional features. Morphological changes result in new PLD and can lead to a new formation of categories. The modal verbs of English, due to morphological changes, were enlisted at the beginning of the 16th century in a new grammatical category (verbs that are in the syntactic position of Inflection). Children explore their linguistic environment for cues of structure, such as IV or VP[V NP]. On the other hand, the morphological changes result in the deletion of the corresponding cue IV from the PLD, namely the deletion of the presence of the verb in the Inflection position (Lightfoot 1999, section 6.3). The presence in the Inflection position is possible for only one group of verbs, the modal verbs and the verb do. When the ability to move from the position of the Verb to the position of the Inflection is lost for all verbs, the position of Inflection is no longer a position to which verbs can be moved. The first morphological change, consequently, radically changes the expression of the cue IV, leading to the loss of the structure. An example of change in one part of the child’s linguistic experience, the morphology of cases, which also leads to other changes in the grammar, comprises the emergence of split genitives in Middle English (Lightfoot 1999, chapter 5). Lightfoot bases their diachronic analysis on the link between cases and ș-roles. ȉhe semantic relationship of the NP with its head is defined as a ș-role, whereas there is a one-to-one relationship, meaning that every NP has one case and one ș-role. In Old English, nouns can assign (genitive) case, towards both the left and the right. The result of this ability are split genitives: a phrase in the genitive case can be found both right and left of the governing head-noun. (2)

Inw3res broþur Ƞnd Healfdenes Inw3re.GEN brother and Healfdene.GEN ‘brother of Inw3re and of Healfdene’ (Anglo-Saxon Chronicle, 878; AD 11-12)

Moreover, in the instances of apposition, when two NPs appear side by side, split examples appear: the two words appear to the right and left of the head-noun. (3)

7

Ælfredes godsune cyninges Ælfred.GEN baptism king.GEN ‘the baptism of King Alfred’ (Anglo-Saxon Chronicle, 890; AD 11-12)

Clark (1992) and Gibson & Wexler (1994) have argued for a different perspective on the relation of language acquisition to the E-language.

Theoretical Framework

55

Middle English has examples of split genitives, but they are very different from those of Old English. For example, in Middle English, structures such as the following are possible: (4)

Ȁyng Priamus sone of Troy king Priam son of Troy ‘son of the king of Troy, Priam’ (Chaucer, Troilus and Cressida I, 2; AD 14)

According to Lightfoot, in these instances, the word found to the right of the head is not an apposition, nor is connected with coordination; moreover, it has no semantic relationship with the head (son). In Early Modern English, a new grammatical system is formed in which the features of the morphological case of the previous grammatical systems are absent. As a result of the loss of the case system, the genitive ending -es is reanalysed as clitic which marks the case not of the NPs but of the DPs. This may explain why groups of nouns in the genitive only begin to appear in this era. (5)

DP[

my uncle from Cornwall ]’s cat ‘the cat of my uncle from Cornwall’

Children, encountering a system that does not have morphological cases, reanalyse the old suffix of the genitive case as clitic. Structures such as Ælfredes godsune cyninges are reanalysed as Ælfred’s godsune king. (6)

DP[ DP [

Ælfred ] D’s NP[N godsune [ king ] ] ] Ælfred ’s baptism king ‘the baptism of king Ælfred’

Examples of this reanalysis from a text of the period can be found in (7), where the noun after the head is not marked for case: (7)

ȉhe Grekes hors Synoun ‘Sinon, the horse of the Greeks’ (Chaucer, Squire’s Tale, 209; AD 14)

Lightfoot argues that after the loss of the morphological case, the noun Sinon receives its case through its coindexing with the noun hors. The nouns are in apposition and for this reason bear the same ș-role; consequently, case marking for the noun Sinon is not required. On the other hand, the structure Ȁyng Priamus sone of Troy is not the same as the structure Ælfredes godsune cyninges, because the nouns Priamus and Troy are not in apposition, they are not coreferential, and consequently a means is needed for the assignment of a ș-role to Troy; this is precisely the function of the preposition of. The features of the new grammatical system can be interpreted if the children who acquire the language: (a) heard structures such as Ælfredes godsune cyninges, (b) did not have the morphological system of cases that their parents had, and (c) followed a grammatical system requiring all DPs to have a ș-role (Lightfoot 1999). Children explore their linguistic environment for morphological cases and, if these are found, the children use them as realisations of abstract cases. If the children who acquire the language do not find morphological cases, then a different grammar emerges in relation to the older period (here: Old English) that had morphological cases. The change in the grammar is a result of the restrictive theory of Universal Grammar requiring the presence of ș-roles and cases: the loss of the transparency of overt cases led to alternative ways of denoting ș-roles. The child searches for cues only in simple syntactic fields, in the primary acquisition domains (degree-0 learnability, Lightfoot 1991, 1994a, b). Consequently, children do not attempt to identify all of the input with specific features and criteria but search for specific abstract structures that are derived from the input, checking only simple structural fields. The output of children -in other words, the grammar they acquire- is the result of the cues that the children find8. An example of change that shows how Lightfoot’s approach functions in relation to the simple structural fields is the analysis of the change in the word order in English from object – verb to verb – object. The verb in Old English is generated in the final position, as in Modern Dutch and German. Even though the clearest cues for the determination of the parameter are found in the word order in the complement clauses, children (when 8

The fields of cues are not identified with the fields that change. According to this view, the changes in structurally complex fields are not ruled out; what is ruled out is the changing of grammatical features due to features and phenomena that are not concerned with simple structural fields. All of the changes begin from cues in structurally simple fields but may result in changes in other areas of the grammatical system as well (structurally simple or complex).

56

Chapter Two

they acquire language) do not have access to this linguistic information and must, therefore, define the parameter based only on the cues from the main clauses. The instances that could have facilitated the analysis of object – verb were (a) main clauses with the verb at the end, which were possible in Old English, and (b) the position of particles that remained in the position where they were generated (namely the final position of the clause) when the verb moved to the I(nflection) or to the C(omplementiser). With the passage of time, these cues of the original position of the verb (the verb after the object) decreased in frequency, until at the end of the period of Old English (12th century) their infrequency did not allow children to recognise that the language they were acquiring was a language with the verb in the final position. As a result, the structures were reanalysed as structures with the verb preceding the object. In the period of the reanalysis, in the complement clauses, the verb remained in the final position; if children had had access and checked (while acquiring) the complement clauses of their input, reanalysis would not have occurred. Consequently, the only way to form a different grammar in a child is if the child is exposed to different PLD to a high degree. Linguistic diachronic research aims to determine how grammars changed and how the corresponding linguistic experience of children changed before the changes took place so that the new grammars were the only possible results. This is the reason why research into diachronic changes through statistical analysis is not supported, according to this perspective. According to this perspective, the study of new grammars and grammatical change is linked to the study of language variation and acquisition.

2.2.4 Diachrony and transitivity in Greek: a review of the literature Language change in the area of transitivity in the Greek language has not been the object of in-depth enquiry. On the whole, the issue of diachronic changes in transitivity, and more specifically in transitivity alternations, has not been dealt with systematically in the literature. Earlier studies (generative or non-generative) contain only remarks that concern the diachrony of transitivity in Greek, but always in relation to other phenomena or only for one period (i.e. transitivity in an earlier period of Greek) or in relation to contemporary tendencies that can be observed in Modern Greek (ongoing changes). Even though the corresponding studies are, on the whole, limited, the following studies show the direction that needs to be taken in any study of transitivity in the Greek language. (a) (In)transitivisation, causativisation Concerning the derivation of transitive and intransitive verbs, there are only two studies from which some elements on the subject under examination can be deduced: the first study concerns the period of an earlier phase of Greek and the other contemporary tendencies of change that can be observed in Modern Greek. Of particular interest is the analysis by Ȁavoukopoulos (1988: mainly 288-315) for transitivity in the language of Homer; the framework for his study concerns cases and PPs in Homeric Greek. The analysis adheres to the functional theory of language as seen by Martinet’s functionalism. The verbs in Homer are classified in a transitivity continuum (more or less transitive) and are separated into verbs that obligatorily do not have an object, verbs for which it is possible to take an object, and verbs that must be connected to an object. With the help of statistical research, Ȁavoukopoulos indicates the high frequency of transitive verbs in Homeric data, mainly of transitive verbs that are obligatorily connected to a direct object (which is not omitted). The transitivity of verbs in Homeric Greek is linked to aktionsart and compounding. Concerning aktionsart, certain verbs are transitive in the present and the imperfect tense but intransitive (middle, reflexive, or passive) in the past, present perfect, and past perfect tenses: ΦΕΣΕ΍ΗΎΉȱ arariske (ACT.IMPERF.3SG - transitive), όΕ΅ΕΉȱ e:rare (ACT.AOR.3SG intransitive), ΩΕ΋Ε΅ȱ areDZ›Š (ACT.PERF.1SG - intransitive) ‘join, fit together - be joined closely together’; πΚϾΉ΍ȱ ephyei (ACT.PRES.3SG - transitive), σΚΙȱ ephu (ACT.AOR.3SG - intransitive), ΔΉΚϾΎ΅Η΍ȱ pephykasi (ACT.PERF.3PL - intransitive) ‘bring forth, produce, put forth - grow, wax, spring up or forth’. In contrast, a correlation between voice morphology (active or non-active) and (perfective or non-perfective) aspect is not observed. (8)

a. ΅ЁΘϲΖȱȱ Έȇȱȱ ΦΐΚϠȱȱ ΔϱΈΉΗΗ΍Αȱȱ οΓϧΖȱȱ ΦΕΣΕ΍ΗΎΉȱȱ ΔνΈ΍Ώ΅ autos d’ amphi podessin heois arariske pedila he.NOM prt around feet his join.ACT.IMPERF.3SG sandals.ACC ‘but he himself was fitting sandals about his feet’ (Homer, Odyssey, 14, 23; 8 BC) b. ϶ȱȱ Έχȱȱ Ύ΅Ϡȱȱ ΔκΗ΍Αȱ πΑϠȱΚΕΉΗϠΑȱȱ όΕ΅ΕΉΑȱȱ ϊΐ΍Α ho de: kai pasin eni phresin e:raren he:min

Theoretical Framework

57

the.NOM prt and all.DAT in minds join.ACT.AOR.3SG we.DAT ‘(our decision) that suited all of us just now in our minds’ (Homer, Odyssey, 4, 777; 8 BC)

As regards compounding, transitive verbs arise from an intransitive base with the addition of infixes (Ȃþnjğȁ
oideo: ‘to swell/become swollen’ > ȂþnjĝȀȁ
oidano: ‘to cause to swell’), but intransitive types are also created from a transitive base (ǴġȘȁ
le:tho:
‘to make one forget a thing’ > ǴǙȘĝȀȁ le:thano: ‘to be unseen/forget’). Consequently, the derivational morphemes are not exclusively connected to the process of derivation of the transitive or intransitive types (transitive and intransitive types derive with the same suffixes)9. The compound verbs are considered the result of transitivisation of intransitive simple verbs (ăȑȖDŽèDŽǬ
histamai ‘to stand’
ĺ
ñäģȑȖDŽèDŽǬ
epistamai ‘to understand/know’) or of intransitivisation of a transitive verb (ȅğȋȁ
phero: ‘bear/carry a load’ ĺ ñǰȅğȋȁ
ekphero: ‘run away/come to fulfilment’). (9)

a. ΗΓϠȱΈȇȱȱ soi d’

όΘΓ΍ȱΐξΑȱπ·Аȱȱ

Δ΅ΕΣȱΌȇȱ ϣΗΘ΅ΐ΅΍ȱȱ

ωΈξȱȱ

ΚΙΏΣΗΗΝȱȱȱ

e:toi men ego:

para th’ histamai

e:de

phylasso:

you.DAT now

I.NOM beside and stand.NACT.PRES.1SG myself.NOM watch-over.ACT.PRES.1SG

‘now beside you also I stand and ever watch over you’ȱ(Homer, Iliad, 5, 809; 8 BC)ȱ Έȇȱȱ πΔϟΗΘ΅ΘΓȱȱ σΕ·΅ȱȱ b. ΔΓΏΏΤȱȱ polla d’ epistato erga many.ACC prt know.MID.AOR.3SG works.ACC ‘he knew (to do) a lot of things’ (Homer, Iliad, 23, 705; 8 BC)

Regarding the derivational relationship between transitive and intransitive verbs, useful data can be deduced from the tendencies presented for Modern Greek: Theophanopoulou-Kontou (2002b) has indicated as a contemporary tendency in Modern Greek the production of several new transitive structures from pure intransitive verbs in analogy with existing transitivity alternations (verbs that form both a transitive and intransitive structure): (10) a. ȅȚ ȣʌȐȜȜȘȜȠȚ i ipalili

IJȠȣ ȣʌȠȣȡȖİȓȠȣ įȚȑȡȡİȣıĮȞ tu ipurjiu dhierefsan

ıIJȠȚȤİȓĮ stixia

ȖȚĮ IJȠ ȞȑȠ ȞȠȝȠıȤȑįȚȠ ja to neo nomosxedio

the employees.NOM the ministry.GEN leak.ACT.PAST.PERFVE.3PL elements.ACC about the new government-bill

‘The ministry employees leaked some elements about the new government bill’ b. To Ȟİȡȩ IJȘȢ ȕȡȠȤȒȢ ȐıIJȡĮȥİ IJȠ țĮȚȞȠȪȡȚȠ ȝȠȣ ĮȣIJȠțȓȞȘIJȠ to nero tis vroxis astrapse to kenurjo mu aftokinito the water.NOM the rain.GEN gleam.ACT.PAST.PERFVE.3SG the.ACC new.ACC my car.ACC

‘The rain water made my new car gleam’

Particularly productive are the causative structures of motion verbs (see Theophanopoulou-Kontou 2003a): (11) a. ȉıȠȪȜȘıĮȞ IJȠ ĮțȚȞȘIJȠʌȠȚȘȝȑȞȠ ĮȣIJȠțȓȞȘIJȠ tsulisan to akinitopiimeno aftokinito coast.ACT.PAST.PERFVE.3PL the.ACC stuck.ACC car.ACC ‘They let the car coast’ b. ȂȘȞ IJȠȞ ȖȣȡȓȗİȚȢ ʌȓıȦ ȩIJĮȞ ȑȤİȚ ȕȖİȚ Įʌȩ IJȠ ıʌȓIJȚ. ǼȓȞĮȚ ȖȡȠȣıȠȣȗȚȐ. min ton not

jirizis

piso otan exi

vji

apo to spiti. ine ghrusuzia.

he.ACC.WEAK turn.ACT.PRES.2SG back when has.3SG stepped-outside from the house is bad-luck

‘Don’t make him come back when he has stepped outside from the house. It is bad luck’

These structures are interpreted as tendencies of the system and aim at the promotion of the agent and the indirect cause. The tendencies are realised, according to Theophanopoulou-Kontou, in two directions: (a) promotion of the agent/cause in the position of the external argument of unaccusative verbs and (b) new categorisation of unergative verbs with the addition of an indirect cause (in a limited number of cases). The cases of the transitivisation of motion verbs are distinguished from the structures that denote one achievement in which the prototypical agent and the acting subject participate equally in the process10: 9

10

The derivation of verbs from nouns is another wide subject that is concerned with the diachrony of transitive and intransitive verbs. In Russian, for example, it has been observed that, frequently, a transitive verb is derived from a noun. The derivation of an intransitive verb from a noun is less frequent, and even less frequent is the derivation of a transitive from an intransitive verb (Zaliznjak 1977; Nichols 1993b). Cf. also the general remark by Mackridge (1985: 95-96) on flexibility in the transitivity of verbs in Modern Greek; according to Mackridge, many verbs known as intransitive are used by authors, and also in oral speech, as transitive as well.

58

Chapter Two

(12) dzșİȜĮ ȞĮ IJȘȞ ȤȠȡȑȥȦ țȐʌȠIJİ ithela na tin xorepso kapote wanted.ACT.PAST.PERFVE.1SG to she.ACC.WEAK dance.ACT.PERFVE.1SG once ‘I once wanted to dance with her’

(b) Transitivity and voice morphology Studies that are related to the diachrony of voice morphology, which expresses (in)transitivity in Greek, are few. Moreover, these observations are concerned with an old synchronic situation or contemporary tendencies of change which were ascertained in Modern Greek. The loss of middle voice/diathesis of transitive verbs (the benefactive meaning of middle voice ceases to be productive), as well as the tendency of denoting passive (not middle) structures by earlier intransitive non-active types, have been located in the period of transition from the Proto-IndoEuropean to Early Greek11. Parker (1976: 453) assumes that language change arises from the ‘misassignment of constituent structure to utterances’. According to Parker, all ancient languages in which transitive middle verbs took on passive interpretation had the word order of object – verb, the person and the number of subjects were denoted by the verb ending, and the subject was optional. Consequently since several nouns, particularly of neuter gender, had similar forms for the nominative and accusative case, certain sentences could have ambiguous analyses. In these cases, Parker assumes that middle transitive clauses such as (13a) could be structurally analysed as passive just like (13b) with the erroneous assignment of the structural constituents: (13) a. ̖ϲȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΩΗΘΙȱȱȱȱȱȱȱȱȱȱȱ ϷΏΏΙΘ΅΍ȱȱȱ










to asty ollytai the.ACC city.ACC destroy.NACT.PRES.3SG [patient] [non-passive] ‘(He) destroys with all his forces the city’ b. ̖ϲȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΩΗΘΙȱȱȱȱȱȱȱȱȱȱȱ ϷΏΏΙΘ΅΍ȱ to asty ollytai the.NOM city.NOM destroy.NACT.PRES.3SG [patient] [passive] ‘The city is destroyed’

In this way, Parker (1976: 459) interprets the presence of middle transitive verbs in Greek (similar to Sanskrit) which have both middle transitive and passive meaning: these ambiguous types represent the period of the language in which the change of non-productive denotation of personal interest (benefactive meaning) through non-active endings and the tendency towards the predominance of a passive interpretation of the non-active endings occurred. Lightfoot (1979: 245) attributes the aforementioned change to internal ‘vagueness/obscurity’, complexity, or non-regularity of the system. Lightfoot cites the example ȖĤ íȑȖȝ öèħȀDŽȖȂ 
to asty e:mynato / the.NOM city.NOM defend.MID.AOR.3SG
in which the non-active verb is the cause for two possible interpretations: (a)
‘(s/he) defends the city with all his/her strength’ (middle transitive interpretation) and (b) ‘the city is protected’ (passive interpretation). (14) a. ̖ϲȱȱ ΩΗΘΙȱȱȱȱȱȱȱȱ ωΐϾΑ΅ΘΓ to asty e:mynato the.ACC city.ACC defend.MID.AOR.3SG
[accusative] [transitive with middle interpretation] ‘(S/he) defended the city with all his/her strength’ (middle transitive interpretation) b. ̖ϲȱȱ ΩΗΘΙȱȱȱȱȱȱ ωΐϾΑ΅ΘΓȱȱȱȱȱȱȱȱȱȱȱȱ to asty e:mynato the.NOM city.NOM defend.MID.AOR.3SG
[nominative] [passive] ‘The city was protected’ (passive interpretation)

According to Lightfoot, it is the loss of middle interpretation for the öèħȀDŽȖȂ
e:mynato defended.MID.AOR.3SG
that
solves the opacity (a structure with two possible interpretations) since the structure is analysed now only as (b) is analysed. The loss of the middle voice for transitive verbs with the meaning of personal interest for the verb action (benefactive meaning) can be considered a ‘therapeutic’ response to a recently created opacity and, therefore, as a consequence of the 11

Cf. Giannakis 1997b.

Theoretical Framework

59

‘Transparency Principle’. In relation to voice morphology, Kavoukopoulos’s presentation (1988) on the morphology of verbs in Homeric Greek is analytical: the verbs in Homeric Greek are distinguished into: (1) deponents of active morphology (activa tantum) (õȋäȁ herpo: ‘to move slowly/walk’); (2) deponents of non-active (middle) morphology (media tantum) (ǰǗŝèDŽǬ
keimai
‘to be laid/lie/lie outstretched’); or (3) verbs with the ability to be used in both voices (ȑäğȋljȁ
sperkho
‘to set in rapid motion/be hasty of temper’). The transitive verbs that bear both active and non-active morphology, when they bear non-active morphology, denote in Homeric Greek: (a) coincidence of agent and patient; (b) cancellation of the direction of the action towards the object (the verb becomes intransitive); or (c) transitive diathesis (i) with objects that express personal relations to the subject (the action is directed towards the subject itself, a part of his/her body, personal tools, or related person, ǰǗȅDŽǴĠȀ ǰĥȇDŽȖȂ


kephale:n kopsato / head.ACC beat.MID.AOR.3SG), or (ii) when the action is of interest to the subject (benefactive meaning). It is worth noting the unavailability of the presence of a DP in the accusative case that denotes the animate object with non-active verbs. In addition, the difficulty of distinguishing between the passive and the transitive non-active (benefactive meaning/middle voice reading) structure in Homeric Greek has been indicated by Kavoukopoulos in cases where the passive structure is not accompanied by a PP denoting the agent (see Parker’s and Lightfoot’s remarks above). In relation to the statistical research of Kavoukopoulos, the highest percentage of verbs in Homer bear active morphology, the majority of active verbs are transitive (85.9%), and of the non-active verbs, a fairly high percentage are transitive, denoting personal interest for the verb action (benefactive meaning) (31.9%). Contemporary tendencies in voice morphology of transitive and intransitive verbs are analysed in the theoretical study of Theophanopoulou-Kontou (1999a, b, 2002b). The non-active form is expanded in anticausative verbs of Modern Greek. (15) ǺȡȫȝȚıİ vromise

Ș i

ȝʌȜȠȪȗĮ ȝȠȣ; but also ȕȡȦȝȓıIJȘțİ bluza mu vromistike

soil.ACT.PAST.PERFVE.3SG the.NOM blouse.NOM my

Ș i

ȝʌȜȠȪȗĮ ȝȠȣ bluza mu

soil.NACT.PAST.PERFVE.3SG the.NOM blouse.NOM my

‘my blouse soiled’

The non-active morphology in certain anticausative verbs in Modern Greek signals passive interpretation of the structure, more specifically, the presentation of the patient as directly affected, and manifests the tendency towards the use of these verbs as passive. At the same time, the reflexive verbs with stable non-active morphology (ȤIJİȞȓȗȠȝĮȚ xtenizome ‘to comb my hair’, ȝĮțȚȖȚȐȡȠȝĮȚ makijarome ‘to make up’) strengthen the presence of the non-active morphology in the anticausatives (Theophanopoulou-Kontou 1999a, b). On the other hand, Theophanopoulou-Kontou ascertains overgeneralisation of the non-active form in intransitive structures when the subject is [-inanimate] in the language of children (ȜİȚȫșȘțĮȞ ȠȚ ʌĮIJȐIJİȢ / liothikan i patates / melt.NACT.PAST.PERFVE.3PL the.NOM potatoes.NOM / ‘the potatoes melted’; ʌȠȞȑșȘțİ IJȠ ȤȑȡȚ ȝȠȣ / ponethike to xeri mu / sufferpain.NACT.PAST.PERFVE.3SG the.NOM hand.NOM my / ‘my hand suffered pain’; TheophanopoulouKontou 1983-4: 80), whose interpretation is based on the tendency to the morphological differentiation of verbs that denote intransitive structures in cases where a corresponding active transitive form is attested (transitive active -Ȧ/-o vs. intransitive non-active -ȝĮȚ/-me). (c) Cases In contrast, the subject of changes concerning the morphological case of the DP that accompany transitive verbs (which are only of peripheral concern to this present study, as a component to complete the overall picture of verb transitivity) has been well documented. All grammars and studies agree that the earlier and, at the same time, more stable, cases in Greek are those that denote the direct relationship with the word, namely the nominative and the accusative case. In contrast, the dative, principally a case of syncretisms, is the case that completely disappeared (following the course of the locative and the ablative) from Greek. Humbert (1930) analyses the replacement of the Ancient Greek dative by the genitive and the accusative (see also Mihevc-Gabrovec 1960: 20-24). The genitive (Jannaris 1897/1968: 333-340; Schwyzer & Debrunner 1950/2002: 112-172; Humbert & Kourmoulis 1957/2002: 256-272), moreover, is replaced in Modern Greek in many instances by PPs (mainly Įʌȩ apo ‘by’/ ȖȚĮ ja ‘from’ + accusative), but it is not particularly frequent in the plural. The genitives that specify a noun have been preserved, whereas the dividing and ablative uses of the genitive have been replaced (ñȑȘģȁ íȋȖȂȝ

esthio: artou / eat.ACT.PRES.1SG bread.GEN / ‘I eat from the bread’; ǗąȒ ȖŲȀ äDŽģnjȁȀ

heis to:n paido:n / one.NOM the.GEN children.GEN / ‘one from the children’; ȘDŽȝèĝȡȁ ȖȂŨ ǰĝǴǴȂȝȒ

thaumazo: tou kallous / admire.ACT.PRES.1SG the.GEN beauty.GEN / ‘I admire the

60

Chapter Two

beauty’). A reduction, however, can also be seen in the uses of the accusative case in Modern Greek in relation to the accusative case in Classical Greek (Jannaris 1897/1968: 327-330; Schwyzer & Debrunner 1950/2002: 85-112; Humbert & Kourmoulis 1957/2002: 243-256). In Classical Greek the accusative denotes direct object (ͪljǗǬȀ ȖĤȀ ăääȂȀ

ekhein ton hippon / have.ACT.PRES.INF the.ACC horse.ACC / ‘to have the horse’), cognate object (ȀǬǰņȀ ȀģǰǙȀ

nikan nike:n / win.ACT.PRES.INF win.ACC / ‘to win a victory’), direction (ÿǰȀǗŝȑȘDŽǬ íǴȑȂȒ

hikneisthai alsos / reach.NACT.PRES.INF sacred-grove.ACC / ‘to reach a sacred grove’), local or temporal dimension (éäğljǗǬȀ äȂǴǴȂĦȒ ȑȖDŽnjģȂȝȒ


apekhein pollous stadious / be-away.ACT.PRES.INF many.ACC stades.ACC / ‘to be far away’; ñLjģȁ ͪȖǙ ñȀǗȀġǰȂȀȖDŽ

ebio: ete: enene:konta / live.ACT.AOR.3SG years.ACC seventy.ACC / ‘s/he lived for seventy years’), or reference (njǬDŽȅğȋȁȀ ȖĠȀ ȅħȑǬȀ

diaphero:n te:n physin / be-different.ACT.PRES.PART the.ACC nature.ACC / ‘being different as for the nature’) (Humbert & ȀȠȣȡȝȠȪȜȘȢ 1957/2002)12. Of particular interest are the accusatives close to verbs that are normally intransitive for the denotation of direction13 (ăǰǗȖȂ ȂČȋDŽȀĥȀ

hiketo ouranon / arrive.MID.AOR.3SG sky.ACC / ‘s/he arrived in the skies’) and reference (íǴǠŲ
ȖȂĦȒ äĥnjDŽȒ

algo: tous podas / suffer-pain.ACT.PRES.1SG the.ACC feet.ACC / ‘I suffer pain in my feet’). Whereas the adverbial relations of location, time, and degree also continue to be marked by the accusative in Modern Greek, the direction is denoted by prepositions. Similarly, the accusative of reference was not preserved: already in the New Testament, the dative was strengthened at the expense of the accusative of reference, while in Modern Greek, the expressions that took the place of accusative of reference are marked completely differently; for example: (16) ȝȠȣ/ȝİ ʌȠȞȐİȚ IJȠ mu/me ponai to I.GEN/ACC suffer-pain.ACT.PRES.3SG the.NOM ‘I have a headache’

țİijȐȜȚ kefali head.NOM

ȝȠȣ mu my.GEN

For Catsimali (1990), the accusative case in Modern Greek, in contrast to Classical Greek, is a structural case, and shows the transition from a non-configurational to a partially configurational language14. Arguments for this analysis comprise the following: (i) in Modern Greek, in contrast to Classical Greek, the accusative is the obligatory (and unmarked) case of the direct object (for all verbs that assign the ș-role of theme or patient (affected objects), (ii) the prepositions in Modern Greek, but not in Classical Greek, are strictly adjacent to the DP that follows and is always found in the accusative case (the accusative in Modern Greek is the structural case which is given by the prepositions; in contrast, the cases that follow the prepositions in Classical Greek are specified by the lexicon according to their ș-roles), (iii) the dative was replaced by the accusative (or the genitive) for the expression of the experiencer with an impersonal verb (ȝİ ʌȠȞȐİȚ / me ponai / me.ACC suffer-pain.ACT.PRES.3SG)15, (iv) the case for the subject in Modern Greek is only nominative, in contrast to Classical Greek, where an infinitive could assign the accusative case to its subject in certain instances, and where the participles could have subjects in other cases besides the nominative case, and (v) the development of the auxilary verbs ȑȤȦ exo ‘have’16 and șȑȜȦ thelo ‘want’ ĺ șĮ tha ‘will’ has resulted in the separation of the Inflection from the Verb node.

12

13

14

15

16

Ǿumbert considers the accusative of cognate objects, of temporal dimension, of local dimension, and the absolutive accusative (participle of neuter gender denoting a general characteristic: òǰĥȀ
hekon ‘purposely’) as adverbial uses of the accusative case. Already in HerȠdotus, alternatively with preposition and with the suffix -įİ(Ȟ)/-de(n) (ɬȘġȀDŽnjǗ
Athe:na-de Athens-from ‘from Athens’). The cross-linguistic diachronic evidence for Catsimali originates from the analyses of the transition from Old to Modern English: (a) van Kemenade (1987) focused on two structural changes: (i) the change of the subject position (from [Spec, VP] to [Spec, IP]) and (ii) the change from object-verb to verb-object order; (b) Brody (1988) links the 'transitivisation' (nominative + verb + accusative) in Modern English of the impersonal structures (dative + verb + nominative) of Old English with the emergence of structural case in English from the 13th century onwards. The presence of the structural case results in 'transitivisation' of the aforementioned structures, the non-flexibility of word order, the adjacency of the structural arguments, and the loss of morphological marking. Brody, as well as Kemenade, links the changes with changes in Inflection (emergence in auxiliary verbs). In relation to the replacement of impersonal active verbs (for example, ñȀnjǬDŽȅğȋǗǬ èȂǬ

endiapherei moi / concern.ACT.PRES.3SG I.DAT / ‘I am interested in’) by personal non-active verbs (İȞįȚĮijȑȡȠȝĮȚ / endhiaferome / concern.NACT.PRES.1SG ‘I am interested in’). Cf. Aerts (1965), Moser (1988).

Theoretical Framework

61

2.2.5 Summary In this section, the fundamental principles for the analysis of diachronic change (which I have adopted in this present study) were presented. I follow Lightfoot’s approach and assume that the changes are accomplished when a generation of speakers who acquired language end up with a language system that differs in one or more aspects from the grammatical system that corresponds to the linguistic competence of the ideal speaker of the previous generation, i.e. the generation that provides input to the speakers who acquire the grammatical system. I adopt, consequently, the link between language change and acquisition. I follow these basic principles: (a) the productivity of a change is the result of internal irregularity of the system and results in the simplification of grammar or the creation of new patterns and oppositions. The Transparency Principle requires the structures to be as simple as possible and, in this way, it refers to the properties of possible grammars. The mechanism of acquisition prefers relatively simple representations: if the input is ambiguous, the child acquiring the language will select the analysis of data that involves the simplest representation; and (b) diachronic research is able to describe linguistic structures with the same methodology used by synchronic analyses in Generative Grammar. The main goal is the comparison of different systems and the interpretation of the differences. In addition, the analysis of the variation in the changes of the E-language, and mainly of the variation and the tendencies in one chronological period, can show the differences in the I-language. As long as language changes, we can follow the course taken by the changes, and this course provides us with information about the nature and the organisation of grammar. The principles of Universal Grammar and the determination of cues form the laws that lead the changes in grammar. Every grammatical change is interpreted if it is shown that: (a) the linguistic environment of children who acquire language changes and (b) the new phenomenon that appears is just like this because of certain general linguistic principles and new PLD. If we observe a change in one grammatical feature or structure, then something in the linguistic performance of the previous generation may have changed and paved the way for reinterpretation of the feature or structure. The direct cause of the change must be found in some form of alternation/change/innovation in the PLD that the children who acquire language receive. The number of features being acquired must, however, be further limited for language acquisition to be possible: this role is carried out by the cues of Lightfoot’s theory; children do not have to study every sentence separately and link every sentence with every feature but only have to check the data they receive for the designated structures or the cues. Children search for cues only in the simple syntactic domains; they do not attempt to identify all of the input with specific features and criteria but look for specific abstract structures by checking the simple structural domains only. The changes take place as a group of changes. The group of changes is the manifestation of only one choice in the formation of the specific grammar; the loss of morphological marking has consequences as to which lexical items are members of the different grammatical categories, and the changes in the members of the grammatical categories have consequences for the computational system (for example, for the possibility of Verb movement to I(nflection)). We can, therefore, hypothesise that there is a connection between the changes in morphology and in the syntactic changes (positions in the syntactic structure and movements). The new grammars that arise have, however, only local causes (and not causes that refer in total to the grammar of a specific language), and if there are no local changes, there are no new grammatical systems. According to this argument, the presence of a new Ilanguage in a child has to do with the primary conditions and details of a child’s linguistic experience. Finally, we have ascertained that language change in the area of Greek transitivity has not been the object of detailed research. Concerning the derivation of transitive and intransitive verbs, there are only two studies from which we are able to gather evidence for the subject under examination: the analysis of Kavoukopoulos (1988: mainly 288-315) for transitivity in the language of Homer, and the research of Theophanopoulou-Kontou (2002b) for synchronic tendencies in Modern Greek. There are also few studies related to the diachrony of voice morphology that expresses (in)transitivity in Greek. Parker (1976) and Lightfoot (1979) concern themselves with the loss of the meaning of personal interest for the verb action of the non-active voice (benefactive meaning), Kavoukopoulos (1988) with the voice morphology in Homeric Greek and Theophanopoulou-Kontou (1999a, b, 2002b) with the synchronic tendencies of change in the voice morphology of intransitive verbs in Modern Greek. The subject of change concerning the morphological case of DPs that accompany transitive verbs has been particularly systematically studied. Catsimali (1990) comes to the conclusion that the accusative case in Modern Greek, in contrast to Classical Greek, is a structural case. From the analyses of Greek diachrony we can conclude that there is a need for further research on: (a) the changes in relation to the processes of causativisation (and of the causes that favour causativisation); (b) the changes in the area

62

Chapter Two

of voice that are linked to transitivity, and (c) the unified or non-unified character of the spread of the use of the accusative case.

CHAPTER THREE DIACHRONIC DATA AND ANALYSIS

3.1 From Homeric to Classical Greek The distribution and analysis of the data that follow endeavour to give a general picture of the developments from the Ancient Greek period to the period of Modern Greek. Greek verbs are examined in relation to the changes in the argument structure and voice. This research attempts to add to the limited information that is afforded us by grammars on the diachrony of transitivity. The results of research in relation to the working hypotheses that ǿ put forward in the theoretical framework (chapter 2) are given for each period separately. A complete interpretation of the changes in the diachrony of the Greek language is referred to in chapter 4. An exploration of all possible changes and innovative verb structures (structures that were not grammatical in the previous period) does not lie within the scope of this monograph. Moreover, the presentation of all the changes would not have had a theoretical result if it were not accompanied by theoretical interpretation and analysis. For these reasons, the research is limited to the investigation of the following changes in the area of the lexicon: (a) the change from causative to anticausative verb (b) the change from anticausative to causative verb In parallel with that, I check the retention of the original type (causative and anticausative, respectively) and, of course, the instances of change of already alternating verbs (into non-alternating verbs): (a) an alternating verb changes into a non-alternating anticausative verb, (b) an alternating verb changes into a non-alternating causative verb. Emphasis in the research is given to the semantic group of the causative verbs, as has been defined by many studies in the area of Generative Grammar1. In the theoretical framework section, I presented the reasons for identifying prototypical transitivity with causative verbs and the central role played by causativity. The research aspires to present the causative verbs and their features and, of course, to reveal the accompanying differences in the verb system in different periods of the Greek language. The description of the data endeavours to reveal the way in which causative verbs emerged and the variation (morphological and syntactic) that occurs in every period, investigating the possible domains of change which are: the presence of new causative uses for verbs that were exclusively intransitive in the previous period (causativisation), the presence of new anticausative uses for verbs that were exclusively transitive (anticausativisation), the presence of new transitive (not causative) uses for intransitive verbs, and the presence of new intransitive (not anticausative) uses for verbs that were transitive. The main aim is to give a complete picture of (anti)causativity in every era and to analyse all aspects of the grammatical system that are linked to it: prefixes, cases, and voice morphology2. In parallel with that, the parts of the verb system that are connected to (anti)causativity but have not changed (remain stable from one period to the next and form the complete picture of every period) are described.

3.1.1 The overall picture It is generally accepted that when we talk about the Classical Greek language (5th-3rd century BC), we need to bear in mind the presence of many dialects, each of which had special significance in that period (Panayotou 1990, 1994; Luraghi, Pompei & Skopeteas 2005; Karali 2007). The two dominant 1

2

In the data used in this research, verbs and structures that were not expected in analyses and descriptions of the specific semantic group up to now have not been excluded but have arisen from the detailed research. The subject of the research is comprised only of the phenomenon of (anti)causativity, in other words, of only one and indeed a very specific part of the verb system. An endeavour has been made, however, to study this limited field in all of its possible manifestations and to reveal elements that arise from the study of the microvariation that appears between the periods of the Greek language.

64

Chapter Three

dialects were Attic and Ionic Greek, which particularly influenced the remaining dialects, as well as each other, creating a Koine dialect, the Attic-Ionic. Athens gained great political and cultural power as well as domination after the Persian Wars. Fear of the Persians caused many Greek cities to turn towards Athens, and this favoured the Attic dialect (Charalambakis 1980; Joseph 2001a). The general features of the grammatical system of the Classical Greek language that are concerned with transitivity are given as follows: Classical Greek generally permitted free word order: the subject – verb and verb – subject orders were equally possible. The object preceded or followed the verb (or the subject), whereas the clitics usually appeared in the second position (Aitchison 1979; Taylor 1994; Dik 1995; Dover 2001). The nominative case was the morphological case of the subjects of finite verbs, whereas the accusative was the usual case of the subjects of infinitives and the typical case for direct objects; therefore, the accusative case was a particularly frequent case. Some verbs idiosyncratically governed objects in other cases, of course. The dative case marked indirect objects, arguments that are interested in the action described by the verb, possession with the verb 'to be', agents with some passive verbs, or partial effect on the object (Joseph 2001a). As far as the aktionsart is concerned, according to Moser (2005), in Homeric Greek, every type of aktionsart was associated with certain forms of roots or with specific affixes, and quite a few verbs did not have a present or a past form. Moser also indicates the enormous morphological variation and the absence of consistency in the Homeric Greek voices: verbs that bear in the remaining tenses only middle voice have active present perfect, while verbs with active present bear middle morphology in other types. In the Classical period, on the other hand, we can talk about an aktionsart system since the three verb stems (present, past, and perfect) spread to all verbs with (a) vowel alternations/ablaut alternations of the root, (b) addition of affixes, and (c) types from different roots (verbs with similar meaning but with different aktionsart are used) (Moser 2005: 193-226). The opposition between the three stems was transformed into an aspectual opposition and did not refer at all to the distinction between dynamicity and stativity but only to the distinction between telicity and non-telicity. Certainly, the independence of the aktionsart (lexical aspect) system from the (grammatical) aspect had not been completed by the Classical period, as can be seen by the tendency to select for the past tense the type of (grammatical) aspect that was connected more with the aktionsart of the verb. Furthermore, in Classical Greek, expression only by verbs of roles and functions that were usually denoted with a Nominal Phrase and a copular verb becomes possible. In structures where the role is denoted by a verb, emphasis is given to the duration (Stassen 1997)3:

(1)

3

Similarly, in Latin, the expression of structures of copular verb and adjective by using only one verb is possible (Wierzbicka 1968: 376): (1) a.

b.

4 5

aǯȱ ΑΓΐΉϾΖȱȱ ΉϢΐϟȱȱ Ȧȱȱ Ά΅Η΍ΏΉϾΖȱȱ ΉϢΐϟŚȱȱȱ nomeus eimi basileus eimi shepherd.NOM be.ACT.PRES.1SG king.NOM be.ACT.PRES.1SG ‘I am a shepherd’ ‘I am a king’ bǯȱ ΑΓΐΉϾΝȱ ȱ Ȧȱȱ Ά΅Η΍ΏΉϾΝȱȱȱ ȱ nomeuo: basileuo: ȱ be-a-shepherd.ACT.PRES.1SG be-a-king.ACT.PRES.1SG ‘I am a shepherd’ ‘I am a king’ a. ̍ΓΏΓΆϲΑȱȱ Φ·νΏ΋Αȱȱ Θ΍ΑΤȱȱ ΎΉΕΣΘΝΑȱȱ ΑΓΐΉϾΉ΍ȱȱȱȱȱȱȱȱȱ kolobon agele:n tina kerato:n nomeuei devoid.ACC herd.ACC one.ACC horns.GEN be-shepherd.ACT.PRES.3SG ‘He is the shepherd of a herd devoid of horns5’ (Plato, Statesman, 265d, 4; 5-4 BC)

rosa rubra est rose.NOM red.NOM be.ACT.PRES.3SG ‘the rose is red’ rosa rubet rose.NOM be-red.ACT.PRES.3SG ‘the rose is red’

In these instances, the verb does not denote change-of-state, which will be denoted in later periods, but state. Other verbs with the meaning of state that could have been used instead of the structure copular verb + adjective: albeo ‘to be white’, areo ‘to be dry’, caleo ‘to be warm’, doleo ‘to be sad’, seneo ‘to be old’, timeo ‘to be afraid’, vigeo ‘to be strong’. For the transcriptions, cf. note 2 in para. 1.1. Most translations of the Ancient Greek examples are based (a) on Liddell, H. G., Scott, R. & Jones, H. S. 1996: A Greek-English Lexicon with a Revised Supplement. Oxford: Clarendon Press, (b) on Perseus Digital Library Project. Ed. G. R. Crane. Tufts University. , and (c) on Kahane, A. & Mueller, M.: The Chicago Homer. Northwestern University.

Diachronic Data and Analysis

65

b. Φ·΅ΌϲΖȱȱ ΑΓΐΓΌνΘ΋Ζȱȱ ΘΉȱ Ύ΅Ϡȱ ΑΓΐΉϿΖȱȱ Θϟȱȱ πΗΘ΍Α agathos nomothete:s te kai nomeus ti estin good.NOM lawgiver.NOM and and shepherd.NOM what.NOM is.ACT.PRES.3SG ‘What is to be both a good lawgiver and shepherd’ (Plato, Minos, 321c, 5; 5-4 BC)

In relation to the above (the ability of incorporating complex meanings into the verb) we need to point out that in Classical Greek, the incorporation of the direct object into the verb was possible. a. ΚνΕΝȱȱ ΈϱΕΙ phero: dory bear.ACT.PRES.1SG spear.ACC ‘I bear a spear’ ȱ b. ΈΓΕΙΚΓΕνΝȱȱ doryphoreo: spear-bear.ACT.PRES.1SG ‘I bear a spear’ (2)

̓ΓΏϧΘ΅΍ȱȱȱȱȱȱȱ ·ΤΕȱȱ ΈΓΕΙΚΓΕΓІΗ΍ȱȱȱȱȱȱȱȱȱȱ

ΐξΑȱȱȱ ΦΏΏφΏΓΙΖȱȱȱȱ

politai

men

gar

citizens.NOM prt

doryphorousi

ȱȱ

alle:lous

bear-spear.ACT.PRES.3PL prt

one-another.ACC

‘Citizens ward one another’ (Xenophon, Hiero, 4, 3, 3; 5-4 BC)

In the instances of verbs that normally take two objects, one of these objects can be incorporated into the verb, whereas the second object remains as the direct object in the accusative case (in some instances, the role of the incorporated object changes; the incorporated object denotes the means of the accomplishment of the action). ȱȱ

ȱ ȱ

ȱ

ȱ ȱ

ȱ

(3)

a.ȱȱ Έ΍ΚΕ΋Ώ΅ΘЗȱȱ ΘϲΑȱȱ diphre:lato: ton chariot-drive.ACT.PRES.1SG the.ACC ‘I drive a chariot in the skies’

ΓЁΕ΅ΑϱΑ

ouranon sky.ACC

b.ȱȱ ΆΓΙΎΓΏЗȱȱ ΔΣΌΓΖȱ boukolo: pathos tend-cattle.ACT.PRES.1SG passion.ACC ‘I serve the passion’ȱ c.ȱȱ ΌΙΕΗΓΚΓΕЗȱȱ thyrsophoro: bear-the-thyrsus.ACT.PRES.1SG ‘I lead/assemble the companies’

Ό΍ΣΗΓΙΖ

thiasous companies/troops/religious guilds.ACC

d.ȱȱ Έ΅ΗΐΓΏΓ·Зȱȱ dasmologo: subject-one-to-tribute.ACT.PRES.1SG ‘I subject the islanders to tribute’

ΘΓϿΖȱȱ

Α΋Η΍ЏΘ΅Ζ

tous the.ACC

ne:sio:tas islanders.ACC

e.ȱȱ ΛΉ΍ΕΓΘΓΑЗȱȱ kheirotono: stretch-out-the-hand.ACT.PRES.1SG ‘I elect the leaders’

strate:gous leaders.ACC

̕Ͽȱȱ

ΈȂǰȱȱ

Иȱȱ

ΘϲΑȱȱ

ΗΘΕ΅Θ΋·ΓϾΖȱ

΅ϢΔϿΑȱȱ

ΓЁΕ΅ΑϲΑȱȱ Έ΍ΚΕ΋Ώ΅ΘЗΑȱȱ

sy d’, ho: ton aipyn ouranon diphre:lato:n you.NOM prt the.VOC the.ACC steep.ACC sky.ACC chariot-drive.ACT.PRES.PART.NOM ‘And you, (Helios), whose chariot-wheels climb the steep sky’ (Sophocles, Ajax, 845; 5 BC)

3.1.2 Innovative causative use of intransitive verbs: from ablaut alternation to productive causativisation 3.1.2.1 Homeric Greek In Homeric Greek, the derivational relationship between the causative and anticausative type is associated with the ablaut alternation of the thematic vowel and with change of voice: As regards the glosses, they describe the words and the grammatical information that are essential for this present study, and not every grammatical information irrespectively of the goals of this research.

66

Chapter Three

ȱ

ΚΓΆЗȱphobo: ȱ

ȱ

ΚνΆΓΐ΅΍ phebomaiȱ flee-in-terror/flee-from.NACT

terrify/alarm/threaten.ACT

Lightfoot (1979) argues that these instances, in contrast to passive structures, are the result of a preposing rule6: 'X terrified Agamemnon' ĺ 'Agamemnon was afraid'. (4)

a. ϵΖȱΘΉȱȱ hos te

Ύ΅ϠȱΩΏΎ΍ΐΓΑȱȱ ΩΑΈΕ΅ȱȱ ΚΓΆΉϧȱȱ

Ύ΅Ϡȱȱ ΦΚΉϟΏΉΘΓȱȱ

ΑϟΎ΋Αȱȱȱȱ

kai alkimon

kai apheileto

nike:n

andra

phobei

who.NOM and warlike.ACC man.ACC terrify.ACT.PRES.3SG and take-away.3SG

victory.ACC

‘who terrifies even the warlike man, and takes away victory’ (Homer, Iliad, 16, 689; 8 BC) ЀΔȂȱд̄Ε·ΉϟΓ΍Η΍ȱȱ ΚνΆΓΑΘΓ7 b. ΦΏΏΤȱȱΎ΅Ϡȱȱ΅ЁΘΓϠȱȱ alla kai autoi hyp’ Argeioisi phebonto but and they.NOM by Argives flee-in-terror.NACT.IMPERF.3PL ‘but they themselves were running in fear from the Argives’ (Homer, Iliad, 11, 121; 8 BC)

The movement of the object of these verbs to the subject position is accompanied by alternation of the thematic vowel; in Homeric Greek, for certain verbs, the ș-role that is connected with the DP-object of the transitive type could appear in the DP-subject of the corresponding intransitive type, and this relationship was connected to the alternations in the stem vowel (Kurylowicz 1956: 87; 1964: 86-88). In relation to the passive structures, in the Homeric texts, we meet very few complete passive structures, namely instances of verbs with non-active morphology and the agent expressed with a preposition or with case marking (dative case). In the majority of the examples, we have intransitive verbs with active morphology and the presence of an agent, that is, no marking with non-active morphology: (5)ȱ πΑȱȱ ΎΓΑϟ΋Η΍ȱȱ ΔνΗΓ΍ΉΑȱȱ ЀΔȂȱΦΑΈΕΣΗ΍ȱȱ ΈΙΗΐΉΑνΉΗΗ΍Αȱȱȱȱȱȱȱȱ en konie:si pesoien hyp’ andrasi dysmeneessin in dust drop.ACT.AOR.OPT.3SG by men.DAT hostile.DAT ‘he shall drop in the dust under the hands of men who hate them’ (Homer, Iliad,, 6, 453; 8 BC)

Lightfoot concludes that in Homeric Greek, the PP of the agent is not derived through a postposing rule: the phrase of the agent is generated in the PP position [or with obligatory case (dative)] and is not generated in the subject position – for reasons of economy, we do not need to assume movement of the agent from the subject position to the PP position. Homeric Greek, of course, is a period during which the inherited relationship of the ablaut alternations and the movement of the DP for the anticausatives becomes more obscure in two ways: (a) the system of the ablaut alternations ceases to comprise a productive morphological process, and (b) the movement of the DP-object spreads in other verbs with non-active (middle) endings, in which there never existed thematic vowel ablaut alternation. That is, the DPs can appear as surface subject types of (middle) non-active forms and bearing the same ș-role as the direct object of the corresponding active form: õǴǰǗǬ ȖĤ ͪǠljȂȒ
/
helkei to egkhos / draw.ACT.PRES.3SG the.ACC sword.ACC (‘he draws the sword’) – ȖĤ ͪǠljȂȒ õǴǰǗȖDŽǬ / to egkhos helketai / the.NOM sword.NOM draw.NACT.PRES.3SG (‘the sword is drawn’). (6)

a. (̸ΎΘΓΕ΅) ϣΔΔΝΑȱȱ (Hektor) hippo:n

πΒΣΔΘΝΑȱ ΔΉΕϠȱȱȱΗϛΐȂȱ οΘΣΕΓ΍Γȱȱ

ΚϟΏΓ΍Γȱȱ

ρΏΎΉ΍ȱ

eksapto:n peri

philoio

helkei

se:m’ hetaroio

(Hektor.ACC) horses.GEN tie.PART around tomb

beloved.GEN companion.GEN drag.ACT.PRES.3SG

‘he ties him/Hektor to his horses and drags him around the tomb of his beloved companion’ (Homer, Iliad, 24, 51; 8 BC) b. ΔκΑȱ Έξȱ ΘȂȱ πΔ΍ΗΎϾΑ΍ΓΑȱ ΎΣΘΝȱ ρΏΎΉΘ΅΍ȱȱ ϷΗΗΉȱȱȱȱȱȱȱȱȱȱȱȱȱ Ύ΅ΏϾΔΘΝΑ

pan de t’ episkynion kato: helketai osse all prt and eyelids.ACC under tie.NACT.PRES.3SG eyes.ACC ‘his eyes are hooded under his sinking eyelids’ (Homer, Iliad, 17, 136; 8 BC)

kalypto:n cover.PART

At the same time, in many instances in Ancient Greek, the transitive structure with two arguments (agent/cause + patient – causative structure) comprises an innovative structure based on the earlier non6

7

The preposing rule has been considered transformational or lexical: for different opinions, the criteria, and the relevant bibliographical references, cf. Lightfoot (1979: 252). The PP denotes the cause (‘he was afraid because of the appearance of the Argives’) – the verb is a psych-verb; the anticausative psych-verbs cannot passivise or allow the presence of an agent in PP.

Diachronic Data and Analysis

67

active structure (anticausative structure): éǠĝǴǴȂèDŽǬ
agallomai ‘to take delight/pride in’ is earlier than éǠĝǴǴȁ
agallo: ‘to glorify/exalt/pay honour’, ȘğȋȂèDŽǬ
theromai ‘to become hot/warm’ is earlier than Șğȋȁ
thero: ‘to heat/make hot’, and èDŽģȀȂèDŽǬ
mainomai ‘to rage/be furious/be mad’ precedes èDŽģȀȁ
maino: ‘to madden’ (Lightfoot 1979: 250)8. 3.1.2.2 Classical Greek In Classical Greek, the use of intransitive verbs in causative structures with the meaning of ‘make someone to + ‘meaning of the intransitive verb’ ’ is already possible (not, however, especially systematic or productive)9. An initial evidence of the diachronic process of causativisation can come from an examination of the verb lists of the classic study by Jankuhn (1969). According to Jankuhn, the following verbs are only intransitive active verbs in Homer but appear as alternating (transitive and intransitive) in Classical Greek texts: ΦΏΓ·νΝȱalogeo: ‘to be unreasonable – madden’ȱȱ ·ΉΏΣΝȱgelao: ‘to laugh – deride’ȱ ·ΕΣΚΝȱgrapho: ‘to enroll oneself/be indicted – inscribe/propose’ȱȱ ΌΣΏΔΝȱthalpo: ‘to be warmed/be softened – warm/soften by heat’ ΌΉΕ΅ΔΉϾΝȱtherapeuo: ‘to be treated medically – do service/treat medically’ ΌΏΣΝȱthlao: ‘to crush/bruise’ȱȱ ΏΉ΍΅ϟΑΝȱleiaino: ‘to smooth/polish’ ΘΕЏ·Νȱtro:go: ‘to gnaw/be eaten – gnaw/nibble/eat’ȱ Λ΅ΏΉΔ΅ϟΑΝȱkhalepaino: ‘to be severe, irritated, violent – irritate/treat harshly’

ȱ Similarly, the following verbs are only intransitive non-active in Homer but participate in transitivity alternations in Classical Greek (as they form a corresponding active causative type): Φ·ΣΏΏΓΐ΅΍ȱagallomaiȱ‘to take delight/pride in’ ȱ ·ΉϾΓΐ΅΍ȱgeuomaiȱ‘to enjoy the taste of experience/have enjoyment of’ȱȱ πΕΙΌΕ΅ϟΑΓΐ΅΍ȱerythrainomaiȱ‘to become red/blush’ȱ ϊΈΓΐ΅΍ȱhe:domaiȱ‘to enjoy oneself/take one's pleasure/be glad’ȱ ΌνΕΓΐ΅΍ȱtheromai ‘to become warm/be burnt’ȱȱ ΐ΅ϟΑΓΐ΅΍ȱmainomai ‘to rage/be furious/be mad’ȱ ΐΉΏ΅ϟΑΓΐ΅΍ȱmelainomai ‘to grow black’ ΗφΔΓΐ΅΍ȱse:pomai ‘to rot/moulder’ȱ

During this period of Greek, we can distinguish three large categories of causativisation changes, namely the creation of an innovative causative type based on a non-alternating intransitive of the earlier period, i.e. Homeric: (a) active causative use on the basis of an active intransitive non-alternating type (the result is alternating active (of same form – ‘labile’) verb), (b) active causative use on the basis of an active and non-active (use of both the voices without semantic difference) intransitive nonalternating type (the result is alternating active causative – active/non-active anticausative verb), and (c) active causative use on the basis of a non-active non-alternating type (the result is an alternating active causative – non-active anticausative verb). It appears, therefore, that the voice of the initial anticausative corresponds to the voice of the initial non-alternating intransitive type. I refer to these three categories analytically below:

8

9

Lightfoot points out that the movement of the NP to the subject position did not happen with all (middle) nonactive verbs since in the reflexive structure ϳȱ ΗΘΕ΅Θ΍ЏΘ΋Ζȱ ΏΓϾΉΘ΅΍ȱ Ȧ ho stratio:te:s louetai / the.NOM soldier.NOM wash.NACT.PRES.3SG / ‘the soldier washes himself’, the subject also has the ș-role of the agent, whereas the object of the active type will always be the patient, something which shows that the NP ϳȱ ΗΘΕ΅Θ΍ЏΘ΋Ζ / ho stratio:te:s / the.NOM soldier.NOM has not been moved to a position higher than ΏΓϾΉΘ΅΍ȱ louetaiȱwash.NACT.PRES.3SG. Also see the corresponding analysis of Tsimpli (1989: 250 ff.) of reflexive verbs in Modern Greek: ‘the subject of the reflexive would not be a derived subject given that derived subjects always have a non-agent ș-role’. Indicative for the productivity of the causativisations (whether there is textual evidence or evidence from theoretical reconstructions of derivation) is Schwyzer’s observation (1939/1953: II 233 ff.): ‘seit dem Beginn der Überlieferung ist produktiv das Nebeneinander eines kausativen Aktivs und eines intransitives Mediums, wobei das Aktiv gewöhnlich jünger ist’ (‘from the beginning of the written tradition, the co-existence of a causative active and an intransitive middle type is productive, from which the active type is newer’).

68

Chapter Three

(a) Causative types from intransitive active The result of this kind of causativisation is that the same form of the verb has both transitive and intransitive use (use of both structures without morphological differentiation). This phenomenon is also well known cross-linguistically; verbs that are used with the same form in both the transitive and the intransitive structure are called labile (Larjavaara 2000: 216 ff.; Dixon 1996: 4)10. In the Germanic languages (and chiefly in English), a significant diachronic increase of the labile verbs can be observed. Kulikov (2003) (following, however, the opinions of Hirt (1937: 28) and Henry (1893: 121), who do not separate the labile causative/anticausative verbs from the verbs with optional presence of the object) maintains that more labile types existed in Proto-Indo-European and fewer in the younger languages.

ȱ

(7) ΔΏνΝȱpleo: ‘to sail/go by sea’ȱ i Homeric Greek intransitive active a.ȱ ΎΉϧΑΓΖȱȱ ЀΔνΕΌΙΐΓΖȱȱ ̇΍ϲΖȱȱ ΙϡϲΖȱȱ σΔΏΉΉΑȱȱ д̌Ώ΍ϱΌΉΑȱȱ keinos hyperthymos Dios hyios epleen Iliothen he.ȃȅȂ high-hearted.NOM Zeus.GEN son.NOM sail.ACT.IMPERF.3SG from-Ilion ‘he, the high-hearted son of Zeus was sailing from Ilion’ (Homer, Iliad, 14, 251; 8 BC) i Classical Greek intransitive active Ύ΅ΘΤȱȱ ΔΓΕΌΐϱΑȱ b. ΔΏΉϧȱȱ plei kata porthmon sail.ACT.PRES.3SG with stream ‘sail with the stream’ (Euripides, The Trojan Women, 102-104; 5 BC) transitive active πΔϠȱȱ ΔϱΑΘ΍΅ȱȱ ΎϾΐ΅Θ΅ȱȱ ΑΣ΍ΓΑȱȱ ϷΛ΋ΐ΅ȱȱȱ c. σΔΏΉΙΗ΅Αȱȱ epleusan epi pontia kymata naion okhe:ma sail.ACT.AOR.3PL over sea waves ship.ACC vehicle.ACC ‘they sailed the ships over the waves of the sea’ (Euripides, Iphigenia in Tauris, 409; 5 BC) (8) ·΋ΕΣΗΎΝ ge:rasko: ‘to become old/get older – raise my child’ i Homeric Greek intransitive active Γϡȱȱ Δ΅ΕΤȱ Δ΅ΘΕϠȱ ·νΕΓΑΘ΍ȱ Ύ΋ΕϾΗΗΝΑ a.ȱ ϵΖȱȱ hos hoi para patri geronti ke:rysso:n who.NOM him by

father aged

·φΕ΅ΗΎΉ

ge:raske

officiate-as-herald.PART grow-old.ACT.IMPERF.3SG

‘he was growing old in his herald's office by (Aineias') aged father’ (Homer, Iliad, 17, 325; 8 BC) i Classical Greek intransitive active πΑȱȱ ΘΓϾΘΓ΍Ζȱȱ ΌΉϱΖǰȱȱ ΓЁΈξȱȱ ·΋ΕΣΗΎΉ΍ b.ȱ ΐν·΅Ζȱȱ megas en toutois theos, oude ge:raskei mighty.NOM in them god.NOM not grow-old.ACT.PRES.3SG   ¶the god is mighty in them, and he does not grow old’ (Sophocles, Oedipus Tyrannus, 871; 5 BC) transitive active σΌΕΉΜ΅Αǰȱȱ ΓЁΈȂȱȱ π·φΕ΅Η΅Αȱȱ ΘΕΓΚϜȱȱȱȱ c.ȱ ΓЁȱȱ·ΣΕȱΐȂȱȱ ou gar m’ ethrepsan, oud’ ege:rasan trophe:i not as I.ACC reared.3PL nor grow-old.ACT.AOR.3PL nurture.DAT. ‘as they did not rear me, nor by their nurture did they bring me to old age’ (Aeschylus, Suppliant Women, 894; 6-5 BC) (9) Η΍ΝΔΣΝsio:pao: ‘to keep silence/be silent – keep secret/speak not of’ i Homeric Greek intransitive active Ά΅ΗϟΏΉ΍΅ǰȱȱ Η΍ΝΔφΗΉ΍΅Αȱȱ д̄Λ΅΍Γϟȱȱȱȱȱȱ a. ΉϢȱȱ ·ΣΕȱȱ ΘΓ΍ǰȱȱ ei gar toi, basileia, sio:pe:seian Akhaioi if only because queen.VOC keep-silence.ACT.AOR.OPT.3PL Achaeans.NOM ‘My queen, if only the Achaeans would be silent for you’ (Homer, Odyssey, 17, 513; 8 BC) 10

Other terms for the phenomenon: causative/inchoative (Haspelmath 1993), causative (Levin 1993, Levin & Rappaport 1994), ergative alternation/amphibious (Visser 1963-73: 99), ambitransitive (Dixon 1994: 18, 54, 217), causative-decausative (Dolinina 1989: 26), voice neutral (Tchekhoff 1980), S=O verbs (Dixon 1994), optionally transitive (Miller 1993), voiceless / non-diathetic (Schmalstieg 1982: 124). For English, see Jespersen 1927: 332-337 (move, turn, boil, improve); for crosslinguistic remarks, see Kitazume 1996, Tchekhoff 1980, Hewitt 1982, Šejxov 1987.

Diachronic Data and Analysis

69

i Classical Greek intransitive active Ύ΅Ϡȱȱ ΓЁΈξΑȱȱ ΪΑȱȱ ΉϥΔΓ΍ΐ΍ȱȱ b.ȱ Η΍ΝΔЗȱȱ sio:po: kai ouden an eipoimi keep-silence.ACT.PRES.1SG and nothing.ACC prt say.1SG ‘I keep silence and I will never say a word’ (Demosthenes, Ctesiphon, 268, 3; 4 BC) transitive active ̘ΝΎνΝΑȱȱ σΌΑΓΖȱȱ ΘΓΗΓІΘΓΑȱȱ ΦΑϙΕ΋ΐνΑΓΑȱȱ Η΍ΝΔЗȱȱȱȱȱ c.ȱ Ύ΅Ϡȱȱ Θϲȱȱ kai to Pho:keo:n ethnos tosouton ane:re:menon sio:po: and

the.ACC Phocians.GEN nation.ACC so-much

destroy.PART.ACC

ȱȱ

keep-silence.ACT.PRES.1SG

‘I do not reveal the destruction of the important nation of the Phocians’ (Demosthenes, Philippic, 3, 26, 6; 4 BC)

(b) Causative types from intransitive non-active Schwyzer & Debrunner (1950/2002: 274-277) observe for the following pairs of verbs that from the beginning of the written tradition, the coexistence of a transitive active verb and an intransitive nonactive verb is particularly productive. In these instances, the transitive type is newer. ϡΗΘ΋ΐ΍ȱ– ȱϣΗΘ΅ΐ΅΍ȱȱȱȱȱȱȱ ȱ

histe:mi – histamai ‘to make stand – stand’ Ά΅ϟΑΝȱ– Ά΅ϟΑΓΐ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

baino: – bainomai ‘to make go – walk’ ϷΏΏΙΐ΍ȱ– ϷΏΏΙΐ΅΍ȱ

ollymi – ollymai ‘to destroy – be destroyed/perish’ ϷΕΑΙΐ΍ȱ– ϷΕΑΙΐ΅΍ȱȱ

ornymi – ornymai ‘to make arise/stir – move/stir oneself’ Δ΅ϾΝȱ– Δ΅ϾΓΐ΅΍ȱ

pauo: – pauomai ‘to make end – cease/stop’ ΘΕνΚΝȱ– ΘΕνΚΓΐ΅΍ȱȱȱȱȱȱȱȱȱ

trepho: – trephomai ‘to feed – eat’ ΗΉϾΝȱ– ΗΉϾΓΐ΅΍ȱȱ

seuo: – seuomai ‘to set in motion – be put in quick motion’ ϣΊΝȱ– ρΊΓΐ΅΍ȱȱȱȱȱ

hizo: – hezomai ‘to make sit/seat/place – sit/sit down’ Κ΅ϟΑΝȱȮȱΚ΅ϟΑΓΐ΅΍ȱ

phaino: – phainomai ‘to cause to appear/reveal – appear/come to light’ Έ΍ΈΣΗΎΝȱ– Έ΍ΈΣΗΎΓΐ΅΍ȱ

didasko: – didaskomai ‘to teach – be taught/learn’ (10) ϷΏΏΙΐ΍ollymi ‘to destroy – be destroyed/perish’ intransitive non-active ȱΛΕϱΑУȱȱ ΔΓΏΏХȱȱ ΐΉΏΉΈ΅΍ΑϱΐΉΑΓ΍ȱȱȱȱȱȱȱȱ ȱ a.ȱ Ύ΅Ϡȱ΅ϢΉϠȱ ΘΉΏνΝΖȱȱ ϷΏΏΙΑΘ΅΍ȱȱ kai aiei teleo:s ollyntai khrono:i pollo:i meledainomenoi and always totally destroy.NACT.PRES.3PL time.DAT lot.DAT grow-black.PART.NOM.PL ‘and always they totally perish as they grow black after a lot of time’ (Hippocrates, On Glands, 14, 14; 5-4 BC) transitive active ΔϱΏΉ΍Ζȱȱ ϷΏΏΙΗ΍Αȱȱ ȱȱ b.ȱ ΅ЂΘ΋ȱȱ haute: poleis ollysin she.NOM cities.ACC destroy.ACT.PRES.3SG ‘she destroys cities’ (Sophocles, Antigone, 673; 5 BC)

For the psych-verb ΉЁΚΕ΅ϟΑΝȱŽž™‘›Š’—˜DZȱ‘to cheer/gladden’, only the non-active intransitive type is in evidence in Homer, whereas in Classical Greek the causative active use is possible in parallel with the anticausative non-active which is preserved.

70

Chapter Three

(11) ΉЁΚΕ΅ϟΑΝ euphraino: ‘to cheer/gladden – enjoy oneself/enjoy a pleasure’ i Homeric Greek intransitive non-active ρΎ΋ΏΓΑȱ a.ȱ Ύ΅Ϡȱȱ ΉЁΚΕ΅ϟΑΉΗΌ΅΍ȱȱ kai euphrainesthai heke:lon and gladden.NACT.PRES.INF at-ease ‘and to make merry at my ease’ (Homer, Odyssey, 2, 311; 8 BC) i Classical Greek intransitive non-active b. ϳȱȱ ΘϾΕ΅ΑΑΓΖȱȱ Έ΍Τȱ ΘΓІΘΓȱ Ύ΅Ϡȱȱ ΔΏΉϟΝȱ ΦΔȂȱ ΅ЁΘЗΑȱ ΉЁΚΕ΅ϟΑΉΘ΅΍




ho tyrannos dia touto kai pleio: ap’ auto:n euphrainetai the despot.NOM for that and more from them gladden.NACT.PRES.3SG ‘the despot gets more enjoyment out of them’ (Xenophon, Hiero, 4, 6, 2 – 5-4 B.C.) c.ȱ ψΐΉϧΖȱȱȱȱΘȂȱȱȱ΅ЇȱȱȱȱΓϡȱȱȱΦΎΓϾΓΑΘΉΖȱȱȱȱȱȱΐΣΏ΍ΗΘȂȱΪΑȱȱΓЂΘΝΖȱȱΉЁΚΕ΅΍ΑΓϟΐΉΌ΅ǰȱȱΓЁΛȱȱȱψΈΓϟΐΉΗΌ΅ȱ he:meis t’ au hoi akouontes malist’ an houto:s euphrainoimetha, oukh he:doimestha we.NOM and again the listen.PART.NOM most prt so

gladden.NACT.PRES.1PL not please.1PL

‘again, we listeners would thus be most comforted, not pleased’ (Plato, Protagoras, 337c, 1; 5-4 BC) transitive active ·νΑΝΑΘ΅΍ȱ d.ȱ ΦΏΏȂȱȱГΗΔΉΕȱȱ Γϡȱȱ ΦΌΏ΋Θ΅Ϡȱȱ ΓЁΛȱȱ ϵΘ΅Αȱȱ ϢΈ΍ΝΘЗΑȱȱ all’ ho:sper hoi athle:tai oukh hotan idio:to:n geno:ntai but

as

the

athletes.NOM not

when

amateurs.GEN

ΎΕΉϟΘΘΓΑΉΖǰȱȱ ΘΓІΘȂȱȱ

΅ЁΘΓϿΖȱȱ

ΉЁΚΕ΅ϟΑΉ΍

kreittones,

tout’

autous

euphrainei

superior.NOM

this.NOM they.ACC gladden.ACT.PRES.3SG

become.3PL

‘but as the athletes prove superior to amateurs, that does not gladden them at all’ (Xenophon, Hiero, 4, 6, 4; 5-4 BC)
e.ȱ ВΚΉΏΓϟ΋ȱȱ ΘΉȱȱ Ύ΅Ϡȱȱ ΉЁΚΕ΅ϟΑΓ΍ȱȱ ̇΍ϱΈΝΕΓΑ o:pheloie: te kai euphrainoi Diodo:ron help.3SG and and gladden.ACT.PRES.3SG Diodorus.ACC ‘he could help and please Diodorus’ (Xenophon, Memorabilia, 2, 10, 6; 5-4 BC)

As can be seen from the intransitive non-active example from Classical Greek (11b - Έ΍Τȱ ΘΓІΘΓȱ ΉЁΚΕ΅ϟΑΉΘ΅΍ȱȦȱdia touto euphrainetai / because-of that enjoy.NACT.PRES.3SG), the presence of cause is possible in anticausative structures; the cause is denoted, as in the specific example, with the PP njǬĝ
dia + Accusative, as well as with many other PPs: ñäģ epi + Dative, ñȀ en + Dative, éäĥ apo + Genitive or with a DP in the dative case. The verb ȝĮȡĮȓȞȠȝĮȚ marainomai ‘to wither’, exclusively intransitive during the Homeric period (with the possibility of the presence of the cause in a DP in the dative case or with PP), is causativised in the Classical period and also used as causative with active morphology: (12) ΐ΅Ε΅ϟΑΝȱmaraino: ‘to wither’ȱ i Homeric Greek intransitive non-active Δ΅ϾΗ΅ΘΓȱȱ Έξȱȱ ΚΏϱΒȱ a.ȱ ΔΙΕΎ΅Ϟχȱȱ πΐ΅Ε΅ϟΑΉΘΓǰȱȱ pyrkaie: emaraineto, pausato de phloks
fire.NOM wither.NACT.IMPERF.3SG stopped.3SG prt flame.NOM ‘the fire died down and the flames were over’ (Homer, Iliad, 23, 228; 8 BC)ȱ b. ΅ЁΘΤΕȱ πΔΉϠȱȱΎ΅ΘΤȱȱ ΔІΕȱȱ πΎΣ΋ȱȱ Ύ΅Ϡȱȱ ΚΏϲΒȱȱ πΐ΅ΕΣΑΌ΋ autar epei kata pyr ekae: kai phloks emaranthe: but when KATA- fire.NOM burned.3SG and flame.NOM wither.PASS.AOR.3SG ‘but when the fire had burned itself out, and the flames had died down’ (Homer, Iliad, 9, 212; 8 BC) i Classical Greek intransitive non-active Ύ΅Ϡȱȱ ΐ΅Ε΅ϟΑΉΘ΅΍ȱȱ ΑϱΗУ c.ȱ ΚΌϟΑΉ΍ȱȱ ·ΤΕȱȱ phthinei gar kai marainetai noso:i wane.3SG because and wither.NACT.PRES.3SG malady.DAT ‘for she is waning and wasting with her malady’ (Euripides, Alcestis, 203; 5 BC) d.ȱ ΦΏΏȂȱȱπΔϟΑΉΙΗΓΑǰȱȱ ̑΅ΑΌϟΔΔ΋ȉȱȱ ΎΦ·Аȱȱ Ύ΅ϠȱΗϿȱȱ ΐ΅Ε΅΍ΑϱΐΉΌ΅ȱȱȱȱȱȱ all’ epineuson, Ksanthippe:; kago: kai sy marainometha but nod-assent.IMP.2SG Xanthippe.VOC and-I.NOM and you.NOM wither.NACT.1PL ‘but, Xanthippe, nod assent; both I and you are losing vitality’ (Plato, Epigrams, 5, 80; 5-4 BC) transitive active

Diachronic Data and Analysis

71

e.ȱ ̓ΣΑΌȂȱȱ ϳȱȱ ΐν·΅Ζȱȱ ΛΕϱΑΓΖȱȱ ΐ΅Ε΅ϟΑΉ΍ȱȱȱȱȱȱȱȱȱ panth’ ho megas khronos marainei all.ACC the.NOM strong.NOM time.NOM wither.ACT.PRES.3SG ‘the strong time makes all things fade’ (Sophocles, Ajax, 713; 5 BC) ΘΉȱȱ ΑϱΗΓΑȱȱ ВΑϱΐ΅Η΅Ζǰȱȱ Ψȱȱ ΐ΅Ε΅ϟΑΉ΍ȱȱ ΐΉȱ f.ȱ ΌΉϱΗΙΘϱΑȱȱ theosyton te noson o:nomasas, ha marainei me heaven-sent.ACC and plague.ACC named.2SG which.NOM wither.ACT.PRES.3SG I.ACC ‘and have named the heaven-sent plague that wastes me’ (Aeschylus, Prometheus Bound, 597; 65 BC)

(c) Causative types from intransitive active / non-active Causative structures are also in evidence for intransitive verbs that can appear both in active and nonactive voice morphology. Causative use in this instance is innovative, whereas in Homeric Greek, only the intransitive structure was possible (with the same voice morphology that also bears in Classical Greek: active and non-active). The examples in (13) are of particular interest, as the causative structure is later abandoned and is ungrammatical in Modern Greek. It appears that verbs of change-of-state from internal cause that are causativised do not retain the causativisation in later periods. (13) ΏΣΐΔΝ lampo: ‘to shine – cause to shine’ i Homeric Greek intransitive active and non-active a.ȱ Λ΅ΏΎϲΖȱȱ σΏ΅ΐΔΉȱȱ ΔΉΕϠȱȱ ΗΘφΌΉΗΗ΍ȱȱ khalkos elampe peri ste:thessi bronze.NOM shine.ACT.IMPERF.3SG around chest ‘(such was the flare of the) bronze that was shining around his chest’ (Homer, Iliad, 22, 32; 8 BC) b.ȱȱ ΦΐΚϠȱȱ Έξȱȱ Λ΅ΏΎϲΖȱȱ πΏΣΐΔΉΘΓȱ amphi de khalkos elampeto about prt bronze.NOM shine.NACT.IMPERF.3SG ‘the bronze that closed about him was shining’ (Homer, Iliad, 22, 134; 8 BC) i Classical Greek intransitive active and non-active Έξȱȱ Ό΅ΑАΑȱȱ ΏΣΐΔΉ΍Ζȱ c. ΑІΑȱȱ nyn de thano:n lampeis now prt die.PART.NOM shine.ACT.PRES.2SG ‘now even dead you still shine’ (Plato, Epigrams, 7, 670; 5-4 BC) d.ȱ ΆΕΓΑΘϛΖǯǯǯȱȱ Ύ΅Ϡȱȱ πΎȱΘΓϾΘΓΙȱȱ ΏΣΐΔΉΗΌ΅΍ȱȱ ΔκΗ΅ȱǻΓϢΎϟ΅Ǽ bronte:s... kai ek toutou lampesthai pasa (oikia) clap-of-thunder.GEN and because-of that shine.NACT.PRES.INF whole.NOM (house) ‘a clap of thunder... the whole house is set ablaze because of that’ (Xenophon, Anabasis, 3, 1, 12; 5-4 BC) transitive active ΦΎΘ΅ϧΖȱȱ ΦΗΘνΕ΅ȱȱ ΏΣΐΜ΅Ζ e.ȱ πΐΆ΅ΏАΑȱȱ ̄Ϣ·΅ϟ΅΍ΖȱȱȱΘȂȱπΑΣΏΓ΍Ζȱȱ ΈϱΏ΍ΓΑȱȱ embalo:n Aigaiais t’ enalois dolion aktais astera lampsas cast.PART

Aegean.DAT and sea.DAT treacherous.ACC shores.DAT flame.ACC shine.ACT.AOR.PART.NOM

‘(an Achaean man) casting them the sea-shores of the Aegean, as he kindled a treacherous flame’ (Euripides, Helen, 1131; 5 BC)11

The causative types (old and innovative) bear stable active morphology, even if in Homeric and Classical Greek the formation of transitive types with non-active morphology was possible and with the meaning of ‘personal interest’ for the action (benefactive meaning) that the verb expresses (see the section on voice morphology below). 3.1.2.3 Factors and stages of causativisations The cognate objects appear to comprise an intermediate stage for the transitive use of intransitive verbs. Of particular frequency in Classical Greek is the syntax of intransitive verbs with a cognate noun in the dative case (cf. Andriotis 1937). The combination of intransitive verbs with a cognate object is possible independently of the morphology of the intransitive verb (active or non-active). -intransitive active and cognate object (14) a. ΑΓΗΉϧȱȱ 11

For more examples, cf. Lavidas 2006a.

ΐξΑȱȱ

ΑϱΗΓΑȱȱ

Φ·Εϟ΅Α

72

Chapter Three



nosei men noson agrian be-ill.ACT.PRES.3SG prt wound.ACC savage.ACC ‘he is afflicted with a savage wound’ (Sophocles, Philoctetes, 173; 5 BC) b. ΋ЁΘϾΛ΋Η΅Αȱȱ ΘΓІΘΓȱȱ Θϲȱȱ ΉЁΘϾΛ΋ΐ΅ȱȱȱ e:utykse:san touto to eutykhe:ma succeed.ACT.AOR.3PL. this.ACC the.ACC happiness.ACC ‘they have gained this success’ (Xenophon, Anabasis, 6, 3, 6; 5-4 BC)

ȱȱȱȱȱȱ

-intransitive non-active and cognate object (15) a. ΗΘΕΓΚΤΖȱȱ ΗΘΕνΚΉΗΌ΅΍ȱȱ  strophas strephesthai turnings.ACC twist.NACT.PRES.INF ‘to try every dodge’ (Plato, Republic, 405c; 5-4 BC) b. ΩΏΏ΋Αȱȱ ΏϾΔ΋Αȱȱ ΏΙΔϛΘ΅΍  alle:n lype:n lype:tai other.ACC sadness.ACC become-sad.NACT.PRES.3SG ‘he becomes sad again’ (Plato, Phaedo, 85a; 5-4 BC) ΔκΗ΅Αȱȱ πΔ΍ΐνΏΉ΍΅Α  c.ȱ πΔ΍ΐΉΏΓІΑΘ΅΍ȱȱ epimelountai pasan epimeleian care.NACT.PRES.3PL all.ACC care.ACC ‘they take the utmost care’ (Plato, Protagoras, 325c; 5-4 BC)

In contrast to the cognate objects, the prefixes are not connected with (anti)causativity in Classical Greek, or the derivation of innovative causative types and diachronic changes of verbs concerning transitivity from Homeric to Classical Greek; the presence of a prefix can result in either a transitive or an intransitive verb. Frequently, the presence of a prefix changes the ability of a verb to take a direct object. Transitive verbs in Classical Greek with the presence of a prefix12 are used as intransitive: transitives intransitives  Ω·Νȱago: ‘to lead’ ΦΑΣ·Νȱanago: ‘to return’  ΘΕνΔΝȱtrepo: ‘to direct’ πΔ΍ΘΕνΔΝȱepitrepo: ‘to turn’ (16) ̄Ε·ΉϧΓΖȱ ЖΑȱ ·ΤΕȱ΅ЁΘϲΖȱ д̄Ε·ΉϟΓΙΖȱ Ω·Νǯǯǯȱȱ ΘΓϾΗΈΉȱ Argeios o:n gar autos Argeious ago: ... tousde Argive.NOM be.PART as myself.NOM Argives.ACC take-away.ACT.PRES.1SG these.ACC ‘as I am an Argive myself, I take away ... these Argives’ (Euripides, Heracleidae, 139; 5 BC) (17)ȱ ΦΑΣ·Ή΍ȱȱ πΔϠȱȱ Θϲȱȱ Λ΅ΏΉΔЏΘ΅ΘΓΑȱ anagei epi to khalepo:taton return.ACT.PRES.3SG to the most-dangerous-pitch ‘and it returns to its most dangerous pitch’ (Hippocrates, Epidemics, 1, 3, 11; 5-4 BC)

ȱ Many intransitive verbs with a prefix13 (mainly withȱ njǬDŽ- dia-,ȱǰDŽȖDŽ- kata-,ȱéäȂ- apo-, and ñǰ- ek-) are, however, turned into transitives. Consequently, it is evident that the presence of a prefix cannot be directly connected with the transitive or the intransitive nature of the verb: Ύ΅Θ΅Α΅Ιΐ΅ΛЗȱkatanaumakho: ‘to conquer in a sea-fight/beat at sea’ȱ Ύ΅Θ΅ΔΓΏΉΐЗȱkatapolemȠ: ‘to reduce/make war against’ȱȱ ЀΚϟΗΘ΅ΐ΅΍ȱhyphistamai ‘to stand beneath/support’ ΐΉΘνΕΛΓΐ΅΍ȱmeterkhomai ‘to attack’ȱ ЀΔΉΕΆ΅ϟΑΝȱhyperbainȠ: ‘to mount/overstep’ȱ Δ΅Ε΅Ά΅ϟΑΝȱparabainȠ: ‘to go by the side of/overstep/transgress’ȱ ΔΕΓΗ·ΉΏЗȱprosgelȠ: ‘to smile at/greet’ȱ ȱ ЀΔΓΈϾΓΐ΅΍ȱhypodyomai ‘to go or get under or down into/put on’

12

13

Έ΍΅Ά΅ϟΑΝȱȱ

ΔΓΘ΅ΐϱΑȱ

diabaino: cross.ACT.PRES.1SG

potamon river.ACC

Έ΍νΕΛΓΐ΅΍ȱȱ

ΘχΑȱȱ

ΛЏΕ΅Αȱ

Regarding the binding of prefixes with the verb, which during the Classical period was very close (more so than in Early Greek), Moser (2005: 194) observes that tmesis has completely disappeared from the area of the verb, and Wackernagel refers to only one example of tmesis. Cf. Papanastasiou 2006.

Diachronic Data and Analysis

73

dierkhomai te:n kho:ran pass-across.NACT.PRES.1SG the.ACC country.ACC (18) a. Ύ΅Ϡȱȱ Ά΅ϟΑΉ΍ȱȱ Ύ΅Ϡȱȱ ΓϢΎΉϧ kai bainei kai oikei and go.ACT.PRES.3SG and stay.3SG ‘and he goes and stays...’ (Plato, Symposium, 195e, 3; 5-4 BC)ȱ b.ȱ Έ΍΅Ά΅ϟΑΉ΍ȱȱ ΘϲΑȱёΏΏφΗΔΓΑΘΓΑȱȱ ΗϿΑȱΘХȱΗΘΕ΅ΘΉϾΐ΅Θ΍ȱȱΉϢΖȱΘχΑȱ̈ЁΕЏΔ΋Αȱ diabainei ton Helle:sponton syn to:i strateumati eis te:n Euro:pe:n DIA-cross.ACT.PRES.3SG the.ACC Hellespont.ACC with the army to the Europe ‘he crosses the Hellespont with his army to Europe’ (Xenophon, Hellenica, 3, 2, 9; 5-4 BC) .

In some instances, the completion of the action denoted by the verb is expressed with prefixes; in other words, the role of the prefixes is concerned more with the aktionsart (lexical aspect) than with transitivity. ȱ ΦΔΓΐΣΛΓΐ΅ϟȱΘ΍ȱapomakhomai ti ‘to repel by fighting’ ΆΣΕΆ΅ΕΓΑȱȱ ΎΓ΍ΑϜȱȱ ΦΔΉΐ΅ΛνΗ΅ΑΘΓ

(19) ΘϲΑȱȱ





ton barbaron koine:i apemakhesanto




the.ACC barbarian.ACC together repel-by-fighting.MID.AOR.3PL
‘they had together driven the barbarian back’ (Xenophon, Hellenica, 6, 5, 34; 5-4 BC)

Prefixes function mainly, in fact, as markers of the initiation, termination, and duration of the action meaning14: (i) ΦΑ΅-/ana-: emphasis on the initiation of the action or on the attempt for its completion: (20)ȱ πΔΉϟȱȱ ΘΓ΍ȱȱ Ύ΅Ϡȱȱ ΎΕν΅ȱȱ ΦΑνΆΕ΅ΘΘΉΑȱȱ ϴΕΑϟΌΉ΍΅ epei toi kai krea anebratten ornitheia  since her and flesh.ACC stew.ACT.IMPERF.3SG fowl.ACC ‘since she was also stewing up some flesh of fowls’ (Aristophanes, Frogs, 509; 5-4 BC)

 (ii)ȱ ΦʌȠ-/apo-: expression of the completion of the verb action or of the failure of theaccomplishment of the verb action. At the same time, however, they mark the derivation of verbs (mainly transitive, causative or noncausative) from nouns, or add extra meaning (analogous to the meaning of the prefix when it is used as a preposition) to the verb base to which they are added: ΦΔΓΈΉϟΎΑΙΐ΍ȱapodeiknymi -‘to end up with a proof’ ΈΉϟΎΑΙΐ΍ȱdeiknymi -‘to express a specific case’ 

(21) ΦΔΓΈΉ΍ΎΑϾΑ΅΍ȱȱ ΘϟȱΔΓΘΉȱȱ ΑІΑȱ Ύ΅ΎЗΖȱȱ πΑȱΘ΅ϧΖȱΔϱΏΉΗ΍ȱȱȱȱΔΕΣΘΘΉΘ΅΍ȱȱ apodeiknynai ti pote nyn kako:s en tais polesi prattetai end-up-with-a-proof.ACT.PRES.INF what.ACC now badly in the cities manage.3SG ‘to point out what it is that is now badly managed in our cities’ (Plato, Republic, 473b; 5-4 BC)

Regarding the presence of suffixes15, I suppose that they also comprise markers of derivation of verbs from nouns, but, like the prefixes, in no instance are they identified with an exclusive mechanism of transitivisation or intransitivisation. With the addition of a suffix, both transitive and intransitive verbs are derived; consequently, they should be only [+Verb] markers (according to the terms of Hale & Keyser 2002) and not [+/-Transitive] markers. -ĮȦ/-ao:ȱ Θ΍ΐΣΝ timao: ‘to honour’; ΦΕ΍ΗΘΣΝȱaristao: ‘to take the ariston or midday meal’ (22)ȱ σ·Ν·νȱȱΚ΋ΐ΍ȱȱ ego:ge phe:mi

ΛΕϛΑ΅΍ȱȱ

ΔΣΑΘ΅ȱȱ

ΩΑΈΕ΅ȱȱ ΘϲΑȱȱ

ӥΕΝΘ΅ȱȱ

Θ΍ΐκΑǰȱ

khre:nai

panta

andra

Ero:ta

timan,

ton

I.NOM tell.1SG should.INF every.ACC man.ACC the.ACC Love.ACC honor.ACT.PRES.INF

Ύ΅Ϡȱȱ ΅ЁΘϲΖȱȱ

Θ΍ΐЗȱȱ

ΘΤȱȱ

πΕΝΘ΍ΎΤ

kai autos timo: ta ero:tika and myself.NOM honor.ACT.PRES.1SG the.ACC love-matters.ACC ‘I tell you now that every man should honor Love, and I myself do honor all love-matters’ (Plato, Symposium, 212b, 6; 5-4 BC) 14

15

Humbert & Kourmoulis (1957/2002) name these ‘empty’ prefixes due to the non-addition of another meaning to the verb meaning apart from the addition of ‘the completion of the verb action’ meaning. Humbert & Kourmoulis consider the role of prefixes to be fundamental since the increase in the use of the accusative case was greatly promoted by the increase in the use of prefixes of this type (‘empty’ prefixes) during the development of Greek. Cf. Papanastasiou 1998.

74

Chapter Three

-İȦ/- eo: ȱ ΦΕ΍ΌΐνΝ arithmeo: ‘to number/count/reckon up’; ΉЁΘΙΛνΝ eutykheo: ‘to be prosperous/fortunate’; ΉЁΈ΅΍ΐΓΑνΝ eudaimoneo:’ to be prosperous, well off/truly happy; ΚΓΆνΝȱ phobeo: ‘to terrify/alarm/threaten; ϴΛνΝȱokheo:ȱ‘to endure/suffer/continue’ (23) ΛΕφΐ΅Η΍Αȱȱ ·ΤΕȱȱ ΉЁΘΙΛЗȱȱ ȱ ȱ khre:masin gar eutykho: goods.DAT prt be-happy.ACT.PRES.1SG ‘I am happy with these goods’ȱ(Euripides, Fragments, 143, 1; 5 BC)

-ȠȦ/-oo: ȱ ΐ΍ΗΌϱΝ misthoo: ‘to let out for hire/farm out/offer to let’; Ί΋ΐ΍ϱΝȱze:mioo: ‘to cause loss/do damage to/penalize’; ΈΓΙΏϱΝ douloo: ‘to enslave’; πΏΉΙΌΉΕϱΝeleutheroo: ‘to set free’ (24)ȱ πΔΉϟȱȱ ΗΉȱȱ ΘΓІΈȂȱȱ πΏΉΙΌΉΕЗȱȱ ΚϱΑΓΙȱ epei se toud’ eleuthero: phonou as you.ACC that.GEN set-free.ACT.PRES.1SG murder.GEN ‘for of this murder I acquit you’ (Euripides, Hippolytus, 1449; 5 BC)

-ĮȗȦ/-azo: ȱ ΥΕΔΣΊΝ harpazo: ‘to snatch away/carry off’; Έ΍ΎΣΊΝ dikazo: ‘to judge/ sit in’ (25)ȱ ΥΕΔΣΊΉ΍ȱȱ ΐνΗΓΑȱȱ harpazei meson seize.ACT.PRES.3SG waist.ACC ‘he seizes me around the waist and 17, 3; 4 BC)

Ύ΅Ϡȱȱ πЏΌΉ΍ȱȱ

ΐΉȱȱ

ΉϢΖȱΘΤΖȱΏ΍ΌΓΘΓΐϟ΅Ζ

kai eo:thei me eis tas lithotomias and throw.3SG I.ACC to the quarries throws me into the quarries’ (Demosthenes, Against Nicostratus,

-ȚȗȦ-izo:ȱ πΏΔϟΊΝ elpizo: ‘to hope for/look for/expect’; ΑΓΐϟΊΝȱnomizo:ȱ‘to use customarily/own/acknowledge’ (26)ȱ ΓЁΈвȱȱ ЀΐΉϧΖ πΏΔϟΊΉΘΉȱȱ ΈνΒΉΗΌ΅΍ȱȱ ψΐκΖȱ oud’ hymeis elpizete deksesthai he:mas not you.NOM expect.ACT.PRES.2PL await.INF we.ACC ‘you do not yourselves expect them to await our attack’ (Xenophon, Anabasis, 6, 5, 17; 5-4 BC)

-ĮȚȞȦ/-aino:ȱ ΉЁΚΕ΅ϟΑΝ euphraino: ‘to cheer/gladden’; ΔΓ΍ΐ΅ϟΑΝ poimaino: ‘to herd/tend’; Η΋ΐ΅ϟΑΝȱse:maino:ȱ‘to show by a sign/indicate/point out’ (27) д̄ΏΏΤȱΎ΅ϠȱΔΏΣΑΓ΍ȱȱ alla kai planoi but

ȱΎ΅Ϡȱȱ ΦΑ΅ΆΣΗ΍ΉΖȱȱ

Λ΅ΏΉΔ΅Ϡȱȱ

Θ΅ЁΘΤȱȱ

Η΋ΐ΅ϟΑΓΙΗ΍Αȱȱȱ

kai anabasies

khalepai

tauta

se:mainousin

and leading-astray.NOM and mounting.NOM difficulties.NOM these.NOM indicate.ACT.PRES.3PL

‘but both leadings astray and mountings mean difficulties’ (Hippocrates, De diaeta, i-iv, 93, 23; 5-4 BC)

-ĮȚȡȦ/-airo:ȱ Ύ΅Ό΅ϟΕΝȱkathairo:ȱ‘to cleanse/clear’ ȱ (28)ȱ ΛΕϱΑΓΖȱȱ Ύ΅Ό΅ϟΕΉ΍ȱ ΔΣΑΘ΅ȱȱ ·΋ΕΣΗΎΝΑȱȱ ϳΐΓІȱ khronos kathaire panta ge:rasko:n homou time.NOM clear.ACT.PRES.3SG everything.ACC age.PART.NOM together ‘Time purges all things, aging with them’ (Aeschylus, Eumenides, 286; 6-5 BC)

-ıțȦ/-sko: Έ΍ΈΣΗΎΝ didasko: ‘to instruct a person/teach a thing’; Δ΍ΔϟΗΎΝȱpipisko:ȱ‘to give/make drink’ (29)ȱ Ύ΅Ϡȱ οΘνΕΝΑȱȱ ΎΝΏϾΉ΍ȱȱ ΛΕ΋ΗϟΐΝΑȱȱ Ύ΅Ϡȱȱ Έ΍ΈΣΗΎΉ΍ȱȱ Ύ΅ΎΣ kai hetero:n ko:lyei khre:simo:n kai didaskei kaka and others.GEN hinder.3SG useful.GEN and teach.ACT.PRES.3SG bad.ACC ‘they keep one from useful occupations and teach what is bad’ (Xenophon, On Hunting, 13, 3, 1; 5-4 BC)

The majority of causative verbs are formed from nouns or adjectives with the addition of the suffixes ȩȦ, -ĮȓȞȦ, -ȪȞȦ, -ȓȗȦ / -oo: -aino: -yno: -izo: Anticausative verbs are formed with the addition of the suffix -ıțȦ/-sko:

Diachronic Data and Analysis

75

ΈΓΙΏϱΝ douloo: ‘to enslave’ Έ΋ΏϱΝ de:loo: ‘to make visible/show’ ΛΕΙΗϱΝ khrysoo: ‘to make golden/gild’ Ί΋ΐ΍ϱΝ ze:mioo: ‘to cause loss/do damage to/penalise’ ΌΉΕΐ΅ϟΑΝ thermaino: ‘to warm/heat’ ΎΓ΍Ώ΅ϟΑΝ koilaino: ‘to hollow/scoop out’ Η΋ΐ΅ϟΑΝ se:maino: ‘to show by a sign/indicate’

ȱ Ά΅ΕϾΑΝ baryno: ‘to weigh down/oppress/depress’ ψΈϾΑΝ he:dyno: ‘to make pleasant/gratify’ ϴΒϾΑΝȱoksyno: ‘to sharpen/point’

(30)ȱ ΘϜȱȱ te:i

ΏνΒΉ΍ȱȱ

Θ΅ϾΘϙȱȱΔΓΏΏΤȱΊ΋ΐ΍ΓІΑΘ΅ϟȱȱ

ΘΉȱΎ΅Ϡȱȱ Ί΋ΐ΍ΓІΗ΍Αȱȱ

οΘνΕΓΙΖȱȱ

leksei

taute:i polla

te kai

heterous

ze:miountai

ze:miousin

the.DAT. word.DAT this.DAT much loss.NACT.PRES.3PL and and loss.ACT.PRES.3PL

others.ACC

‘by this saying they often both lose and make others lose’ ǻPlato, Laws, 916e, 3; 5-4 BC) (31)ȱ ΘΓІΘΓȱȱ ΘΤȱȱ ЀΔΉΕΗ΅ΕΎνΓΑΘ΅ȱȱ Ύ΅Ό΅ϟΕΉ΍ȱȱ Ύ΅Ϡȱȱ ΎΓ΍Ώ΅ϟΑΉ΍ touto ta hypersarkeonta kathairei kai koilainei that.NOM the.ACC fleshy.ACC clean.ACT.PRES.3SG and hollow.ACT.PRES.3SG ‘that cleans and hollows the fleshies’ (Hippocrates, De ulceribus, 14, 6; 5-4 BC) (32)ȱ ΓЁȱȱ ΐϱΑΓΑȱȱ ΗϧΘΓΑȱȱ ΦΏΏΤȱȱ Ύ΅Ϡȱȱ ΔΓΘϲΑȱȱ ψΈϾΑΉ΍

ou monon siton alla kai poton he:dynei not only food.ACC but and drink.ACC make-pleasant.ACT.PRES.3SG ‘it makes not only food but also drink pleasant’ (Xenophon, Symposium, 4, 8, 7; 5-4 BC) ·΋ΕΣΗΎΝ ge:rasko: ‘to grow old’ ψΆΣΗΎΝ he:basko: ‘to come to puberty/reach-show the outward signs of manhood’ ·ΉΑΉ΍ΣΗΎΝȱgeneiasko: ‘to begin to get a beard’ ȱ (33)ȱ Ύ΅Ϡȱȱ ΑІΑȱȱ ΩΔ΅΍Ζȱȱ ΩΘΉΎΑΓΖȱȱ ·΋ΕΣΗΎΝȱȱ

ΈΙΗΘ΅ΑΓΘΣΘΝΖȱȱ

kai nyn apais ateknos ge:rasko: dystanotato:s and now childless.NOM barren.NOM grow-old.ACT.PRES.1SG piteously ‘and now I am childless and I grow old piteously’ (Euripides, Suppliants, 967; 5 BC)

Furthermore, the derivation of causative verbs from corresponding intransitive or transitive noncausatives is possible with the presence of a suffix: -ȚȗȦ/-izo: Υ·ΑϟΊΝȱǻΥ·ΑνΝǼȱȦ hagnizo: (hagneo:) ‘to cleanse away/purify’ ΗΝΚΕΓΑϟΊΝȱǻΗΝΚΕΓΑЗǼ / so:phronizo: (so:phrono:) ‘to recall a person to his senses/chasten’ ΔΏΓΙΘϟΊΝȱǻΔΏΓΙΘνΝǼȱȦȱploutizo: (plouteo:) ‘to make wealthy/enrich’ (34) a.ȱ sophron-o: ĺ intransitive ΘΤȱȱ ΘЗΑȱȱ ΩΏΏΝΑȱȱ ΔΕΣΘΘΓΑΘΉΖȱȱ ΗΝΚΕΓΑΓІΗ΍Α
ȱ ȱ Γϡȱȱ hoi ta to:n allo:n prattontes so:phronousin those.NOM the.ACC the.GEN others.GEN do.PART.NOM chasten.ACT.PRES.3PL ‘those who do others' business are chastened’ (Plato, Charmides, 163a, 11; 5-4 BC) b. sophron-iz-o: ĺ causative Ύ΅Ϡȱȱ ψȱȱ

ΘΓ΍΅ϾΘ΋ȱȱ ϏΘΘ΅ȱȱ

ΗΝΚΕΓΑϟΊΉ΍Αȱȱ

ϡΎ΅Αχȱȱ

ΉϨΑ΅΍ȱȱ ΦΑΌΕЏΔΓΙΖ




kai

he:

toiaute:

so:phronizein

hikane:

einai

and

the.NOM such.NOM

he:tta

anthro:pous

defeat.NOM chasten.ACT.PRES.INF adequate.NOM be.INF men.ACC

‘such a defeat is adequate to make men be chastened’ (Xenophon, Cyropaedia, 3, 1, 20; 5-4 BC) -ȣȞȦ/-yno: ȱ ΦΏ·ϾΑΝȱǻΦΏ·ЗǼ / algyno: (algo:) ‘to pain/grieve/distress’ ȱ Ό΅ΕΕϾΑΝȱǻΌ΅ΕΕЗǼȱȦ tharryno: (tharro:) ‘to encourage/cheer’ (35) a. alg-o: ĺ intransitive Ώϱ·ΝΑȱȱ ΦΏ·Зȱȱ ΎΏϾΝΑ ȱ ȱ ΘЗΑȱȱ to:n logo:n algo: klyo:n the.GEN words.GEN pain.ACT.PRES.1SG hear.PART.NOM ‘I pain when I hear these plans’ (Sophocles, Philoctetes, 86; 5 BC) b. alg-yn-o: ĺ causative ·Ȃȱȱ ΓЁΈνΑǰȱ ΦΏ·ΙΑЗȱȱ Ένȱȱ ȱ ȱ Ύ΅ϟΘΓ΍ȱȱ ΔΕΓΎϱΜΝȱȱ kaitoi prokopso: g’ ouden, algyno: de yet accomplish.1SG prt nothing.ACC grieve.ACT.FUT.1SG prt

ΗΉȱ

se you.ACC

76

Chapter Three ‘and yet I accomplish nothing by this, and merely cause you grief’ (Euripides, Hippolytus, 1297; 5 BC)

3.1.2.4 Changes in accusative verbs: sub-categories I should point out that as regards verb transitivity not all the changes observed are of the same form. Under the term 'transitivity' or 'changes in verb transitivity', many grammatical phenomena have been listed from time to time, without any differentiation, in traditional grammars. According to the theoretical approach I am following, which I presented in the corresponding section, causative verbs comprise a central group of transitive verbs and need to be distinguished from the remaining transitive verbs. Transitivity has been defined as the presence of a direct object in the accusative case. The changes in relation to the presence of a direct object and its morphological case are not always concerned with the emergence of causative verbs, which have the specific meaning of change-of-state and always require a patient argument as the direct object. Changes in relation to the presence or absence of arguments that bear other ș-roles (cause, agent, etc.) and accompany verbs with other meanings (not change-of-state), or in relation with the morphological case of arguments with other șroles apart from that of the patient, are concerned with non-causative verbs (or accusative verbs, in the sense that the only thing they have in common is that they are able to take objects in the accusative case). The possible changes in non-causative verbs are: (a) from transitive to intransitive (anticausative [with a patient argument as subject] or non-anticausative) and (b) from transitive to causative. Noncausative transitive verbs can be the result of changes of: (a) intransitive verbs (which take a nonpatient direct object and do not denote the change-of-state of a patient) and (b) causative verbs (which do not take patient arguments but other arguments as direct objects). The morphological case also plays an important role in the shaping of the system of transitive (non-causative) verbs; the accusative case can be the result of a change in the syntax of the verb (which in earlier periods took DPs in the genitive or dative case as direct objects). Finally, it is possible for there to be probable alternations in structures (namely, one verb can participate in more than one structure). The above comprise the description of all possible changes in the area of transitivity; the data of Classical Greek, however, limit these possible changes in the lexicon of the verbs. For causative verbs, we observed that the new causative verbs are the result of the causativisation of exclusively intransitive verbs (anticausatives) of Homeric Greek. Also of significance is the retention of the earlier intransitive use. Earlier research on non-causative (accusative) verbs is also scarce16. A complete examination of accusative verbs is not the goal of our monograph, but the following elements of syntactic behaviour and the changes in the accusative verbs are also concerned generally with the issue of causativity (prototypical transitivity). Of particular interest is that the accusative verbs of Classical Greek do not exclusively take a DP in the accusative case as their object, as do the causatives of Classical Greek, but show evidence of alternations between the accusative and other cases or also with PPs. With psych-verbs, active and non-active, according to the data from Classical Greek, the argument of theme/cause could be denoted with a DP in the accusative case or alternatively with a PP (ɚƧˁepi + dative, ƛƠƓ dia + accusative), with a DP in the genitive case (usually animate), or with a DP in the dative case (usually inanimate): active morphology Ό΅ΙΐΣΊΝ thaumazo: ‘to wonder/marvel/honour’ ΉЁΈ΅΍ΐΓΑϟΊΝ eudaimonizo: ‘to call-account happy’ ΐ΅Ύ΅ΕϟΊΝ makarizo: ‘to bless-deem-pronounce happy/congratulate’ ΓϢΎΘϟΕΝ oiktiro: ‘to pity/have pity upon/be sorry that’ Ί΋ΏЗ ze:lo: ‘to vie with/be jealous of/envy’ Λ΅ϟΕΝ khairo: ‘to rejoice/be glad/take pleasure in a thing’ ·΋ΌνΝ ge:theo: ‘to rejoice’ ΦΏ·νΝ algeo: ‘to feel bodily pain/suffer/be ill’ ΔΉΑΌЗ pentho: ‘to bewail/lament/mourn’ Έ΅ΎΕϾΝ dakryo: ‘to weep/shed tears’ 16

See Theophanopoulou-Kontou (2000b: 13-14): the genitive case in its diachronic course bore various ș-roles. For the gradual decline of the genitive, cf. Jannaris (1897), Chatzidakis (1915), Schwyzer & Debrunner (1950), Humbert & Kourmoulis (1957/2002), Tsamadou (1984), Nakas (1987), Campos (1991), Nikiforidou (1991), and Setatos (1995): (a) already there was a tendency towards a decline of the genitive case in earlier times and (b) PPs replaced case types (genitive, ablative, instrumental) in the gradual decline of cases (Jannaris 1897: 337, and for PPs as abstract cases, cf. Dimitriadis 1994a, b; Fykias 1995).

Diachronic Data and Analysis

77

ΌΕ΋ΑЗ thre:no: ‘to sing a dirge/wail’ ΚΌΓΑЗȱphthonoDZ ‘to bear malice/grudge/be envious’ȱ

 non-active morphology Ω·΅ΐ΅΍ȱagamai ‘to wonder/admire a person or thing’ ϴΕ·ϟΊΓΐ΅΍ orgizomai ‘to grow angry’ ϊΈΓΐ΅΍ he:domai ‘to enjoy oneself/take one's pleasure’ ΐνΐΚΓΐ΅΍ memphomai ‘to blame/censure’ ΚΓΆΣΐ΅΍ phobamai ‘to be seized with fear/be affrighted’ ΅ϢΗΛϾΑΓΐ΅΍ aiskhynomai ‘to be ashamed/feel shame’ ΅ϢΈΓІΐ΅΍ aidoumai ‘to be ashamed/stand in awe of/fear’ ΩΛΌΓΐ΅΍ akhthomai ‘to be loaded/grieved’ πΎΔΏφΘΘΓΐ΅΍ ekple:ttomai ‘to be struck with terror or amasement’ ϴΈϾΕΓΐ΅΍ȱodyromaiȱ‘to lament/bewail/mourn’ (36) a. ΅ϢΗΛϾΑΓΐ΅΍ȱȱ Έξȱȱ ΘΤΖȱȱ πΐ΅ΙΘΓІȱȱ ΗΙΐΚΓΕΣΖ aiskhynomai de tas emautou symphoras be-ashamed.NACT.PRES.1SG prt the.ACC of-myself.GEN misfortunes.ACC ‘I am ashamed of my misfortunes’ (Aristophanes, Plutus, 774; 5-4 BC) πΔϠȱ ΘϜȱȱ πΐ΅ΙΘΓІȱȱ Κ΅ΙΏϱΘ΋Θ΍ b.ȱ ΓЂΘΝΖȱȱ ΅ϢΗΛϾΑΓΐ΅΍ȱȱ houto:s aiskhynomai epi te:i emautou phaulote:ti so be-ashamed.NACT.PRES.1SG of the my ineptitude ‘so ashamed I am of my own ineptitude’ (Plato, Theages, 130c, 6; 5-4 BC) ΘΓ΍ȱȱ Θ΅ϧΖȱȱ ΔΕϱΘΉΕΓΑȱȱ Υΐ΅ΕΘϟ΅΍Ζȱ c. ΅ϢΗΛϾΑΓΐ΅΍ȱȱ aiskhynomai toi tais proteron hamartiais be-ashamed.NACT.PRES.1SG verily the.DAT past errors.DAT ‘I am verily ashamed at my past errors’ (Aristophanes, Knights, 1355; 5-4 BC)

Motion verbs17 behave syntactically like the accusative verbs (with optional accusative case) since they can take an argument that denotes the goal of the movement in the accusative case. The object apparently comprises an idiosyncratic (not event) argument that arises from the specific meaning of every verb individually, not from the event structure of the verb group (since it concerns all the verbs). (37)ȱ Ύ΅Ϡȱȱ ΘΓϾΘΓΙΖȱȱ ΐξΑȱȱ Γϡȱȱ ̕ΙΕ΅ΎϱΗ΍Γ΍ȱȱ ΉЁΌϿΖȱȱ Ύ΅ΘΤȱȱΘΣΛΓΖȱȱȱ ȱΚΌΣΑΓΙΗ΍Α
kai toutous men hoi Syrakosioi euthus kata takhos phthanousin and these.ACC prt the.NOM Syracusans.NOM immediately in haste reach.ACT.PRES.3PL ‘these the Syracusans managed immediately to snatch up in haste’ (Thucydides, The Peloponnesian War, 6, 101, 6; 5 BC)

The argument denoting the experiencer can, moreover, be in the accusative or dative case: ΦΕνΗΎΉ΍areskei +DP-accusative / dative

‘it pleases (us/them)’ (38) a. ΗΓϠȱȱ ΈȂȱȱ ΓЁΎȱȱ ΦΕνΗΎΉ΍ȱȱ ΘΦ·ΣΌ΅ȱ soi d’ ouk areskei tagatha you.DAT prt not please.ACT.PRES.3SG the-good.ACC ‘but good treatment does not please you’ (Euripides, Medea, 621; 5 BC) b. ΉϢȱȱ ΓЇΑȱȱ Θϟȱȱ ΗΉȱȱ ΘΓϾΘΝΑȱȱ ΦΕνΗΎΉ΍ȱ ei oun ti se touto:n areskei if well something.NOM you.ACC these.GEN please.ACT.PRES.3SG ‘if any of this pleases you’ (Thucydides, The Peloponnesian War, 1, 128, 7; 5 BC)

Arguments denoting the theme appear as the direct object not only in the accusative but also in the dative case, and with certain verbs also in a PP (without semantic difference). πΑΓΛΏЗenokhlo: + DP-accusative / dativeȱ‘to bother/trouble/annoy’ ΦΈ΍ΎЗadiko: + DP-accusative / ǗþȒ eis-PP / äȋĥȒ pros-PP ‘to do wrong/do wrong in the eye of the law’ ΏΙΐ΅ϟΑΓΐ΅΍lymainomai + DP-accusative / ǗþȒ eisPP/ äȋĥȒ pros-PP ‘to cleanse from dirt’ ΏΝΆκΐ΅΍lo:bamai + DP-accusative / ǗþȒ eis-PP/äȋĥȒ pros-PP ‘to outrage/maltreat’

(39) a. πΑΓΛΏΉϧΖȱȱ enokhleis bother.ACT.PRES.2SG

17

Cf. Charalambakis 1976.

ΐΉȱȱȱ

me

I.ACC

78

Chapter Three ‘you bother me’ȱ(Plato, Alcibiades 1, 104d, 2; 5-4 BC)
ωΑЏΛΏΉ΍ȱȱ ΑІΑȱȱ ψΐϧΑȱȱ ϳȱȱ ̘ϟΏ΍ΔΔΓΖ
b. ΓЁΎȱȱ ΪΑȱȱ ouk an e:no:khlei nyn he:min ho Philippos not prt bother.ACT.IMPERF.3SG now we.DAT the.NOM Philip.NOM ‘Philip would not (have survived to) trouble us today’ (Demosthenes, Olynthiac 3, 5, 9; 4 BC)

Every accusative (non-causative transitive) verb takes a direct object in the genitive case if the action of the verb affects the object only partially (partitive) (Wackernagel 1928: 212-213). (40) a. ΘΓІȱȱ tou

ΐΓϾΑΓΙȱȱ ΔϟΑΉ΍ȱȱ

Ά΅Η΍ΏΉϿΖȱȱ Ύ΅Ϡȱȱ ΩΏΏΓΙȱȱ

ΓЁΈΉΑϲΖȱȱ

ΔΓΘ΅ΐΓІ



mounou pinei

basileus

oudenos

potamou

kai

the.GEN only.GEN drink.ACT.PRES.3SG king.NOM and

allou

other.GEN nobody.GEN river.GEN

‘the only river from which the king will drink’ (Herodotus, The Histories, 1, 188, 6; 5 BC) b. ΔΓΏϿΑȱȱ ΔϟΑΉ΍Αȱ ΓϨΑΓΑ  polyn pinein oinon much drink.ACT.PRES.INF drink.ACC ‘to drink much wine’ (Plato, Symposium, 176c; 5-4 BC) (41) a. ϵΘ΍ȱȱ ϳȱΐξΑȱȱ ΊЗΑΘ΅Ζǰȱȱ ϳȱΈξȱȱ ΑΉΎΕΓϿΖȱȱ πΗΌϟΉ΍ȱȱ hoti ho men zo:ntas, ho de nekrous esthiei that the one.NOM alives.ACC the other.NOM deads.ACC eat.ACT.PRES.3SG ‘that the one eats alives, while the other deads’ (Demosthenes, Fragments, 13, 63, 2; 4 BC) ΪΑȱȱ Ύ΅Ϡȱȱ ВΐЗΑȱȱ πΗΌϟΉ΍Αȱȱ ΅ЁΘЗΑ b.ȱ ψΈνΝΖȱȱ he:deo:s an kai o:mo:n esthiein auto:n pleasantly prt and raw.GEN eat.ACT.PRES.INF they.GEN ‘that he pleasantly eats them raw’ (Xenophon, Hellenica, 3, 3, 6; 5-4 BC)

Similarly, verbs that denote feelings take a complement in the accusative or in the genitive case, according to how precise or complete the perception of the event is from the agent’s point of view. It is possible for some of these verbs to take an active participle as a complement. (42) a. ϔΗΌΓΑΘΓȱȱ ΘχΑȱȱ ΆΓφΌΉ΍΅Α e:isthonto te:n boe:theian perceive.NACT.IMPERF.3PL the.ACC help.ACC ‘(the Peloponnesians) became aware of the coming help’ (Thucydides, The Peloponnesian War, 2, 94, 3; 5 B.C)  b. ϔΗΌΉΘΓȱȱ ΘΓІȱȱ ·νΏΝΘΓΖ e:istheto tou gelo:tos perceive.NACT.AOR.3SG the.GEN laughter.GEN ‘(Philip) heard the laughter’ (Xenophon, Symposium, 1, 16; 5-4 BC) (43) a. πΐΉΐΑφΐ΋Αȱȱ ΘχΑȱȱ ̽Η΍ϱΈΓΙȱȱ ·ΉΑΉ΅ΏΓ·ϟ΅Αȱȱȱ ememne:me:n te:n He:siodou genealogian remember.NACT.PLUP.1SG the.ACC Hesiod.GEN genealogy.ACC ‘I remembered the genealogy of Hesiod’ (Plato, Cratylus, 396c; 5-4 BC) ΘЗΑȱȱ Δ΅Ώ΅΍ЗΑȱȱ ΔΕΣΒΉΝΑȱȱ ΐνΐΑ΋ΑΘ΅΍ b. ψΈνΝΖȱȱ he:deo:s to:n palaio:n prakseo:n memne:ntai pleasantly the.GEN past.GEN deeds.GEN remember.NACT.PERF.3PL ‘with joy they recall their past deeds’ (Xenophon, Memorabilia, 2, 1, 33; 5-4 BC)

3.1.3 Derivation and voice of the transitivity alternations: the contrast between the active causative and the non-active anticausative The issue that concerns us in this section is the derivation of causative and anticausative verbs in Classical Greek (basic and derived type if we accept that they both arise from a common lexical input). Following the contemporary theoretical approach of (anti)causativity, two topics must be examined: (a) the kind of arguments that complement the verbs (agent, cause, instrument -comparing causative and anticausative with passive types) and (b) the voice morphology of causative and anticausative types since all of the approaches up to now attribute a role in the determination and marking of the derivation direction to voice morphology (the causative is derived from the anticausative type or the anticausative from the causative type; of great importance also is the relationship of voice morphology in causatives and anticausatives in relation to other types, mainly transitive non-causative (accusative) and passive).

Diachronic Data and Analysis

79

3.1.3.1 Voice in Ancient Greek: an overview The verb systems of Greek and Sanskrit indicate that the Proto-Indo-European system of voice was comprised of two voices18 that expressed the opposition between active and middle meaning (Delbrück 1897: 412-439; Brugmann 1897-1916/1930: 678-711; Wackernagel 1950: 119-144; Schwyzer & Debrunner 1950: 222-242; Szemerényi 1996: 255-257). Classical Greek had a system of distinction between three voices: active, middle, passive, but middle and passive morphology are distinguished only in the future and the past tense. Very rapidly, however, (possibly during the Classical Greek period, certainly by the end of Classical Greek) middle and passive endings are used alternately without any difference. The sole non-active form in the remaining tenses is formed based on the morphological alternation of the ending (denoting person, as well as number and voice): -Ȧ/-o: vs. -ȠȝĮȚ/-omai. The category of voice (active, passive, middle) and the categories of tense (present, past, future) and aspect (imperfective and perfective) concern all verbs, whether finite (with denotation also of modality and person and number in agreement with the subject), or non-finite (namely, without denotation of person, number, modality). All combinations of categories are not realised morphologically; for example, no different types of middle (ĺ non-active I) and passive (ĺ non-active II) voice morphology in the present, the imperfect, the present perfect, and the past perfect tenses exist; there are also no other types of modality for the imperfect apart from the indicative (cf. Joseph 2001a; Papanastasiou & Petrounias 2007; Pantelides 2008). Table 3.1 Middle and passive non-active forms in the future and past tense Different middle and passive non-active forms exist only in the future and in the past tenses: FUTURE PAST non-active middle form Ÿ non-active ǿ -ıȠȝĮȚ -somai -ıĮȝȘȞ -same:n non-active passive form Ÿ non-active ǿǿ

-șȘ-ıȠȝĮȚ -the:-somai

-(ș)Ș-Ȟ -(th)e:-n -șȘ-ıĮȝȘȞ -the:-same:n

In relation to voice in Classical Greek: (a) The middle type is not identified with specific structures (reflexive or anticausative), nor is the passive type identified with the passive structure in Classical Greek. I suppose, therefore, that it is concerned with, at least from the Classical era onwards, two different morphological types of the same non-active category, which are used alternately. For this reason I prefer and propose the terms nonactive form I and non-active form II, and I do not use the traditional terms middle and passive form. Already in Classical Greek, the traditionally named middle type (Ÿ non-active I) was used in passive structures (with the presence or absence of an agent-PP)19. 18

19

Haacke (1852), from the typical opposition of only active endings with the middle endings, has deduced the presence of only two voices and has attempted the unification of the functions of middle voice under the meaning of the patient. For Kowaleck (1887), even if the development of passive meaning from the original middle voice ‘obscures’ to a certain degree the functions of the original middle voice, the middle voice must have denoted (a) movement that is autonomously accomplished by the subject and (b) active participation of the subject in the action. Grosse (1889, 1891) has associated the middle voice with the denotation of emphasis. The gradual development of the passive structure was commented on for the first time by Wistrand (1941), who indicated that the passive structure was by no means systematically present in Homeric Greek, in contrast to Herodotus, where it already appears to be a part of the grammatical system. Wistrand (actually, in complete agreement with the modern theory) uses as a criterion for the passive structure the possibility of the presence of an agent argument. Moreover, in relation to earlier analyses of verb voice in Homeric and Classical Greek, Zsilka (1964, 1965) has observed that both the middle and passive types can have either intransitive or passive use, without consequent identification of a type with only one function. Jankuhn (1969) has argued that the exploration of the original meaning of Greek middle voice is impossible; the only certainty is that the passive meaning made up one of the meanings of middle voice. Of particular importance is his attention to the diachronic differences between Homeric and Classical Greek lists of verbs from Homeric epics and Classical Greek prose writing; many transitive verbs of Classical Greek were also transitive in Homeric Greek but they did not have passive structure; conversely, for many middle intransitives in Homeric Greek there was no active transitive (causative) type, as there was in Classical Greek. This strengthens the position that the direction of diachronic changes cannot have been the same for passive and causative structures: from what were originally transitive verbs, it appears that secondary reflexive and passive types originate, whereas causative verbs appear to be coined from what were originally intransitive anticausative types. Chatzidakis (1934: 23ff.): ‘ȞȦȡȓȢ ıȘȝİȚȫșȘțĮȞ ȤȡȒıİȚȢ ʌȠȣ ʌĮȡĮȕȓĮȗĮȞ IJȘ įȠȝȚțȒ įȚĮijȠȡȠʌȠȓȘıȘ ȝȑıȘȢ țĮȚ

80

Chapter Three

(44) πΎΉϟΑΓ΍ȱȱ ΔΓΏ΍ΓΕΎφΗΓ΍ΑΘΓȱȱ ЀΔϲȱȱ ΘЗΑȱȱ ΦΑΘ΍ΔΣΏΝΑȱȱȱȱȱ ekeinoi poliorke:sointo hypo to:n antipalo:n these.NOM besiege.MID.FUT.OPT.3PL by the adversaries ‘they would be besieged by their adversaries’ (Xenophon, Hellenica, 7, 5, 18; 5-4 BC)

Furthermore, from classical Ancient Greek, the traditionally named passive (Ÿ non-active II) type is used productively in intransitive non-passive structures (with psych-verbs, in anticausative structure): (45) Ύ΅ϟȱȱ ΐΉȱȱ πΤΑȱȱ πΒΉΏν·ΛϙΖǰȱȱ ΓЁΎȱȱ ΦΛΌΉΗΌφΗΓΐ΅΍ȱȱ  kai me an ekselegkhe:is, ouk akhthesthe:somai and I.ACC if refute.2SG not grieve.PASS.FUT.1SG ‘and if you refute me, I shall not be grieved’ (Plato, Gorgias, 506c, 1; 5-4 BC)

(b) From a purely morphological perspective, we can distinguish in the non-active form II of the future and the past tenses an additional non-active morpheme, the morpheme -șȘ- -the:-: (46) ȱ ΏΙȬΌφȬΗΓΐ΅΍ȱȱ ly-the:-somai unbind/release-the:-somai

Ȧȱ ȱπȬΏϾȬΌ΋ȬΑŘŖ / e-ly-the:-n unbind/release-the:-n

In the future tense, the morpheme -șȘ- -the:- and -ȠȝĮȚ -omai express the non-active, whereas in the past tense, only the morpheme -șȘ- -the:- expresses the non-active. As regards the suffix -v -n of the past tense (ñȟǗäǴĝǠǙȀ
ekseplage:n ‘I was struck with terror or amasement’), initially it must have denoted the ending of the past of the active verbs of the conjugation -ȝȚ -mi, and it was also later used for reflexive and anticausative (but not passive) structures of verbs with active morphology21. -Ȟ -n 1. active suffix (for a specific conjugation (verbs in -ȝȚ -mi)) ĺ 2. only intransitive (reflexive and anticausative) suffix ĺ 3. non-active (reflexive, anticausative and passive) suffix

20

21

ʌĮșȘIJȚțȒȢ ijȦȞȒȢ ıIJȠȞ ǹȩȡȚıIJȠ țĮȚ IJȠȞ ȂȑȜȜȠȞIJĮ’ (‘from early on uses were noted that violated the structural differentiation of middle and passive voice in the past and in the future tenses’). Humbert & Kourmoulis (1957/2002) argue that the formation of the passive [Ÿnon-active II] past tense with -șȘȞ the:n has been completed in Homer, that the passive [Ÿnon-active II] future tense with -șȘıȠȝĮȚ -the:somai, unknown in Homer and Herodotus, is not in evidence before Aeschylus, and that the passive future tense was formed on the basis of the past tense. Cf. the view of Andersen (1993a) that only the derivational morpheme -șȘ- -the:- is connected with nontransitive structures. In contrast, the verb suffixes (agreement suffixes with the subject) are not directly connected with transitivity. See also Ȁühner & Gerth (1898), who refer to the intransitive (reflexive and anticausative) use of the ending -(Ș)Ȟ -(e:)n of the past tense of many transitive verbs in relation to the active past tense formed with the consonant -ı- -s- before the ending, which is transitive: ΈϾΝȱdyo:ȱ ϣΗΘ΋ΐ΍ histe:mi ΚϾΝphyo: 

active past σΈΙΗ΅ edysa ‘I hid something’ σΗΘ΋Η΅ este:sa ‘I placed something’ σΚΙΗ΅ ephysa ‘I gave birth to something’

past with -Ȟ -n σΈΙΑedyn ‘I hid’ σΗΘ΋Αeste:n ‘I stood’ σΚΙΑephyn ‘I came into being’

In many instances, the contrast between the (traditionally named ensigmos ‘formed with -s-’) active past tense and the (traditionally named rizikos ‘formed with the root’) -(Ș)Ȟ -(e:)n past tense (σΗΘ΋Η΅ este:sa ‘I placed something’ / σΗΘ΋Α este:n ‘I stood’) is related to the opposition of (causative) transitive and non-transitive (anticausative) type (Brugmann 1897-1916/1930: 73-79): σΚΙΗ΅ephysa ‘I gave birth to something’ (1) Δ΅ϧΈ΅Ζȱȱ ΓЃΖȱȱ σΚΙΗ΅ȱȱ ΎΦΒΉΌΕΉΜΣΐ΋Αȱ paidas hous ephysa kaksethrepsame:n children.ACC which.ACC grow.ACT.AOR.1SG and-raised.1SG ‘the children I begot and raised’ (Euripides, Medea, 1349) σΚΙΑ ephyn ‘I came into being’ (2) ϣΎΉ΍ȱȱ ΔΣΏ΍Αȱȱ ϵΌΉΑΔΉΕȱȱ ΦΕΛΤȱȱ ΘЗΑȱȱ Ύ΅ΎЗΑȱȱ ΥΐϠΑȱȱ σΚΙȱ hikei palin hothenper arkha to:n kako:n hamin ephy ah again from-where beginning.NOM the.GEN bad.GEN we.DAT come-into-being.ACT.AOR.3SG ‘Ah! here again the beginning of our troubles came into being’ (Aristophanes, Acharnians, 821; 5-4 BC)

Diachronic Data and Analysis

81

πΒνΔΏ΋Β΅ekseple:ksa (transitive)

/ πΒΉΔΏΣ·΋Αekseplage:n (intransitive) ‘I struck with terror or amasement’ ‘I was struck with terror or amasement’  ΦΑνΘΕΉΜ΅anetrepsa (transitive) / ΦΑΉΘΕΣΔ΋Αȱanetrape:n (intransitive) ‘I overturned, upset’ ‘I was upset, disheartened’ ȱ σΚ΋Α΅ephe:na (transitive) / πΚΣΑ΋Αephane:n (intransitive) ‘I brought to light, caused to appear’ ‘I came to light, appeared’ (47) a. πΐξȱ ·ΤΕȱȱ πΎΉϟΑ΋ȱȱ ψȱȱ ΦΕΛχȱȱ ΓЁΎȱ πΒνΔΏ΋ΒΉΑ eme gar ekeine: he: arkhe: ouk ekseple:ksen I.ACC because that.NOM the.NOM govenrment.NOM not frighten.ACT.AOR.3SG ‘for that government did not frighten me’ (Plato, Apology, 32d, 4; 5-4 BC) b. Ύ΅Ϡȱȱ π·Аȱȱ ΦΎΓϾΗ΅Ζȱȱ πΒΉΔΏΣ·΋Αȱȱ ȱ kai ego: akousas ekseplage:n and I.NOM hear.PART.NOM frighten.PASS.AOR.1SG. ‘and I, when I heard him, was frightened’ (Plato, Republic, 336d, 5; 5-4 BC)

The suffix -v -n is used mainly in reflexive and anticausative (but not passive) structures in Homer; of the 22 instances recorded by Delbrück (1897), only 2 have a purely passive interpretation (πΔΏφ·΋Α eple:ge:n ‘I was struck, smitten’ / πΘϾΔ΋Αetype:n ‘I was beaten, I struck myself’). (48) ρΏΎΉ΅ȱ ΔΣΑΘ΅ȱȱ ΐνΐΙΎΉΑȱȱ ϵΗΗȂȱȱ πΘϾΔ΋ helkea panta memyken hoss’ etype: wounds.NOM all.NOM heal.ACT.PERF.3SG which.NOM strike.PASS.AOR.3SG ‘all his wounds have been closed up where he was struck’ (Homer, Iliad, 24, 420-421; 8 BC)

All of the remaining examples are not passive; many, in fact, denote change-of-state (Δ΅·ϛΑ΅΍ page:nai ‘to become solid/stiffen’, Ε΅·ϛΑ΅΍ȱrage:nai ‘to break’, Θΐ΅·ϛΑ΅΍tmage:nai ‘to be divided’). Consequently, there appears to have been a connection between the earlier ending -v -n and anticausativity. (49) Ύ΅Ϡȱȱ ΑΣΔ΅΍ǰȱȱ kai napai, and

ΓЁΕ΅ΑϱΌΉΑȱȱ ΈȂȱΩΕȂȱȱ

ЀΔΉΕΕΣ·΋ȱȱ

ΩΗΔΉΘΓΖȱȱ

΅ϢΌφΕ

ouranothen

d’ ar’

hyperrage:

aspetos

aithe:r

prt as

break.PASS.AOR.3SG

endless.NOM

bright-air.NOM

ravines.NOM from-sky

‘and the ravines, as endless bright air spills from the heavens’ (Homer, Iliad, 16, 300; 8 BC)

The newer verbs with the addition of the morpheme -șȘ- -the:- are, in Homer, mainly reflexives, anticausatives, and, in fewer cases, passives; Grosse (1889, 1891) refers to only 30 with a purely acquire.PASS.AOR.INF, ΈΓΌΉϟ΋ dotheie: passive interpretation (ΎΘ΅ΌϛΑ΅΍ ktathe:nai give.PASS.AOR.OPT.3SG) from the 129 examples that he notes (intransitive: Φ·ΉΕΌϛΑ΅΍ agerthe:nai gather.PASS.AOR.INF, ·ΙΐΑΝΌϛΑ΅΍ gymno:the:nai strip.PASS.AOR.INF, ΛΓΏΝΌϛΑ΅΍ kholo:the:nai anger.PASS.AOR.INF)22. passive structure (50) ΛΕφΐ΅ΘȂȱȱ ΦΔ΅΍ΘϟΊΓΑΘΉΖǰȱȱ ρΝΖȱȱ ΎȂȱ ΦΔϲȱȱ ΔΣΑΘ΅ȱȱ ΈΓΌΉϟ΋ khre:mat’ apaitizontes, heo:s k’ apo panta dotheie: possessions.ACC demand.PART.NOM until and APO all.NOM give.PASS.AOR.3SG ‘demanding our possessions until all were given back’ (Homer, Odyssey, 2, 78; 8 BC)

intransitive structure (51) ̍΅Ϡȱȱ ΘϱΘΉȱȱ Έχȱȱ ΔΉΕϠȱȱ ΎϛΕ΍ȱȱ ̓ΓΗΉ΍ΈΣΝΑȱȱ πΛΓΏЏΌ΋ȱȱ ȱ kai tote de: peri ke:ri Poseidao:n ekholo:the: and then prt about heart Poseidon.NOM anger.PASS.AOR.3SG ‘and then Poseidon was angered about the heart’ (Homer, Iliad, 13, 206; 8 BC)23

22

23

See Andersen (1993a, b, 1994), according to whom Classical Greek did not possess passive voice, only active and middle; the middle type of Classical Greek was able to receive many possible interpretations depending on the context; anticausative, reflexive, reciprocal, and passive are among the interpretations. The result is the final predominance of the passive (Ÿnon-active II) type, namely of the additional suffix (πΚΓΆφΌ΋Αȱ ephobe:the:n terrify.PASS.AOR.1SG,ȱ πΔΓΕΉϾΌ΋Αȱ eporeuthe:n drive.PASS.AOR.1SG,ȱ πΎΓ΍ΐφΌ΋Αȱ ekoime:the:n sleep.PASS.AOR.1SG) at the expense of the middle (Ÿnon-active I).

82

Chapter Three

3.1.3.2 Distribution of active and non-active voice Referring to the distribution of active and non-active morphology in Classical Greek, we ascertained that non-active morphology is generally considered to be the marked instance in relation to active voice (Ȁühner & Gerth [1898-1904]/1963: 89; Smyth 1920: 389-398; Ruijgh 1975: 359; Bakker 1994: 24)24. (a) Non-active in transitive structures (benefactive meaning) The non-active voice morphology is frequently used in transitive structures and adds the meaning that the result of the verb action concerns the subject; in other words, it leads to absorption of the benefactive (indirect reflexivity in traditional terminology): ΘϟΌ΋ΐ΍ȱȱ

ΑϱΐΓΑ 

tithe:mi place.ACT.PRES.1SG ‘I make a law’

nomon law.ACC 

ΘϟΌΉΐ΅΍ȱȱ

ΑϱΐΓΑȱȱȱȱ

tithemai nomon place.NACT.PRES.1SG law.ACC ‘I pass a law in my own interests’25 ȱ Έ΍ΈΣΗΎΝȱȱ

Θ΍ΑΣȱȱ

didasko: tina teach.ACT.PRES.1SG somebody.ACC ‘I give lessons to someone’ Έ΍ΈΣΗΎΓΐ΅ϟȱȱ

Θ΍Α΅ȱȱȱȱȱȱȱȱȱȱ

didaskomai tina teach.NACT.PRES.1SG somebody.ACC ‘I take care that someone who I have an interest in follows lessons, frequently a father for his son’ ȱ Δ΅Ε΅ΗΎΉΙΣΊΝȱȱ

Α΅ІΑȱȱȱȱȱȱȱȱȱȱ

paraskeuazo: prepare.ACT.PRES.1SG ‘I prepare a ship’

naun ship.ACC ȱ

Δ΅Ε΅ΗΎΉΙΣΊΓΐ΅΍ȱȱ

Α΅ІΑȱȱȱȱȱ

paraskeuazomai naun prepare.NACT.PRES.1SG ship.ACC ‘I prepare a ship for myself’ ȱȱȱ

24

25

For Duhoux (1988), non-active morphology denotes a marked relation between verb and arguments; in all instances, the central meaning of middle voice (= non-active morphology) is the presence of a verb action as incorporated in the personal sphere of the argument, which is presented as the agent. Duhoux’s analysis is very close to what I maintain since, according to Duhoux, middle voice denotes, among other things, action that is realised solely by the subject (anticausative meaning), whereas the active voice denotes the participation not only of the subject in the realisation of the action (causative meaning): non-active voice morphology – active voice morphology ·ΉϾΓΐ΅΍ȱgeuomai ‘to dine’ ͒ ·ΉϾΝȱgeuo: ‘to make someone dine’ σΏΔΓΐ΅΍ȱelpomai‘to hope’͒σΏΔΝȱelpo:‘to make someone hope’  ΎΓ΍ΐΣΓΐ΅΍ȱkoimaomai‘to sleep’͒ΎΓ΍ΐΣΝȱkoimao:‘to make someone sleep’ Ώ΅ΑΌΣΑΓΐ΅΍ȱlanthanomai‘to lose’͒Ώ΅ΑΌΣΑΝȱlanthano:‘to make someone lose’ ΐ΍ΐΑφΗΎΓΐ΅΍ȱmimne:skomai‘to remember’͒ΐ΍ΐΑφΗΎΝȱmimne:sko:‘to make someone remember’ ǻΦΔǼϱΏΏΙΐ΅΍ȱ(ap)ollymai‘to destroy myself’͒ǻΦΔǼϱΏΏΙΐ΍ȱ(ap)ollymiȱ‘to make someone destroy himself/herself’ ϴΕ·ϟΊΓΐ΅΍ȱorgizomai‘to become enraged’͒ϴΕ·ϟΊΝȱorgizo: ‘to make someone enraged’ ϷΕΑΙΐ΅΍ȱornymai‘to stand up’͒ϷΕΑΙΐ΍ȱornymi‘to make someone stand up’ ϴΕΛνΓΐ΅΍ȱorkheomai‘to dance’͒ϴΕΛνΝȱorkheo:‘to make someone dance’ Δ΅ϾΓΐ΅΍ȱpauomai‘to stop’ ͒Δ΅ϾΝȱpauo:‘to make someone stop’ ΔΉϟΌΓΐ΅΍ȱpeithomai‘to be persuaded’͒ΔΉϟΌΝȱ™Ž’‘˜DZ‘to make someone to be persuaded’ ǻπΔ΍ǼΜ΋ΚϟΊΓΐ΅΍ȱ(epi)pse:phizomai ‘to vote’͒ǻπΔ΍ǼΜ΋ΚϟΊΝȱ(epi)pse:phizo:‘to make someone vote’

See the observations of Indian grammarians (Wackernagel, Vorlesungen I, p. 124): active voice parasmaipadam: ‘verb for the other participant’, middle voice in transitive structure atmanepadam: ‘verb for the participant himself, the result of verb action concerns the subject itself’. Based on the Indian grammarians, Bechert & Gonda (1960) define as the initial function of middle voice the denotation of an action that is related or affects a person or an object. Furthermore, Benvenist (1950) calls middle voice ‘internal diathesis’ in contrast to active voice which is ‘external diathesis’.

Diachronic Data and Analysis

83

ΉϢΕφΑ΋Αȱȱȱȱȱȱȱȱȱȱ poio: eire:ne:n make.ACT.PRES.1SG peace.ACC ‘I bring peace to the warring sides’ ΔΓ΍Зȱȱ

ΔΓ΍ΓІΐ΅΍ȱȱ

ΉϢΕφΑ΋Αȱ

poioumai eire:ne:n make.NACT.PRES.1SG peace.ACC ‘I make peace with my enemy’ (52) Φ·ΕΣΚΓΙΖȱȱ agraphous unwritten.ACC ȱ ΘΓϾΖȱȱ ·вȱȱ ȱ tous g’ the.ACC prt ȱ ȱ σΛΓ΍Ζȱȱ ΪΑȱȱΓЇΑȱȱ ekhois an oun have.2SG prt so

Θ΍Α΅Ζȱȱ ΓϨΗΌ΅ǰȱȱ σΚ΋ǰȱȱ Иȱȱ ͒ΔΔϟ΅ǰȱȱ ΑϱΐΓΙΖDzȱȱ tinas oistha, ephe:, o: Hippia, nomous? ȱ some.ACC know.2SG said.3SG the.VOC Ippias.VOC laws.ACC

Ένȱȱ

de prt

πΑȱΔΣΗϙǰȱ σΚ΋ǰȱ

ΛЏΕθȱȱȱ

Ύ΅ΘΤȱȱ Θ΅ЁΘΤȱȱ

ΑΓΐ΍ΊΓΐνΑΓΙΖǯ

en pase:i, ephe:, in all said.3SG

kho:rai country

kata for

nomizomenous. observe.PART.ACC

tauta same

ΉϢΔΉϧΑǰȱȱ σΚ΋ǰȱȱ

ϵΘ΍ȱȱ Γϡȱȱ

ΩΑΌΕΝΔΓ΍ȱȱ ΅ЁΘΓϿΖȱȱ σΌΉΑΘΓDzȱȱ

eipein,

ephe:,

hoti hoi

anthropoid

say.INF

said.3SG that

the.NOM men.NOM

ethento? ȱ

autous

those.ACC make.MID.AOR.3PLȱ

ȱ ǯǯǯȱΘϟΑ΅Ζȱȱ

ΓЇΑǰȱȱ σΚ΋ǰȱȱ

... tinas oun, some.ACC prt ȱ

ȱ ȱ

ΑΓΐϟΊΉ΍Ζȱ ΘΉΌΉ΍ΎνΑ΅΍ȱȱ

ephe:, said.3SG

ΘΓϿΖȱȱ

ΑϱΐΓΙΖȱȱ

π·Аȱȱ

ΐνΑǰȱȱ σΚ΋ǰȱȱ

ΌΉΓϿΖȱȱ

ΓϨΐ΅΍ȱȱ

ego: I.NOM

men, prt

theous gods.ACC

oimai tous nomous toutous think.1SG the.ACC laws.ACC these.ACC

ephe:, said.3SG

ΘΓϾΘΓΙΖDzȱȱ

nomizeis tetheikenai tous nomous toutous? think.2SG make.ACT.PERF.INF the.ACC laws.ACC these.ACC

ΘΓϧΖȱȱ

ΦΑΌΕЏΔΓ΍Ζȱȱ

ΌΉϧΑ΅΍ȱ

tois the.DAT

anthropois men.DAT

theinai make.ACT.AOR.INF

ΘΓϿΖȱȱ

ΑϱΐΓΙΖȱȱ

ΘΓϾΘΓΙΖȱȱ

“Do you know what is meant by ‘unwritten laws’, Hippias?” “Yes, those that are uniformly observed in every country.” “Could you say that men made them?”... “Then by whom have these laws been made, do you suppose?” “I think that the gods made these laws for men” (Xenophon, Memorabilia, 4, 4, 19; 5-4 BC) (53) a. non-active transitive ΦΔϲȱȱ

ϴΏΙΕνΝΑȱȱ

ΔΓ΍ΉІΑΘ΅΍ȱȱ

Η΍Θϟ΅ȱȱ

ǽΓϡȱ̄Ϣ·ϾΔΘ΍Γ΍Ǿ

apo olyreo:n poieuntai sitia [hoi Aigyptioi] from wheat make.NACT.PRES.3PL food.ACC [the Egyptians.NOM] ‘they make food from wheat (to eat for themselves)’ (Herodotus, The Histories, 2, 36; 5 BC) b. active transitive Γϡȱȱ

ΦΗΘΓϠȱȱ

ΩΏΉΙΕΣȱȱ

ΘΉȱȱΎ΅Ϡȱȱ ΩΏΚ΍Θ΅ȱȱ

πΔΓϟΉΙΑ

hoi astoi aleura te kai alphita epoieun the.NOM townsmen.NOM wheat.ACC and and barley.ACC make.ACT.IMPERF.3PL ‘the inhabitants of the cities prepare a meal with flour from wheat and flour from barley (for the Persian king)’ (Herodotus, The Histories, 7, 119; 5 BC) (54) non-active transitive ̇΋ΐΓΛΣΕ΋Ζǯǯǯȱȱ

ΓЁΎȱȱ

ΦΔΓΎνΎΕΙΔΘ΅΍ȱȱ

ΘχΑȱȱ

ΓЁΗϟ΅Α

De:mokhare:s... ouk apokekryptai te:n ousian Demochares.NOM not conceal.NACT.PERF.3SG the.ACC property.ACC  ‘Demochares... has not concealed his property’ (Demosthenes, Against Aphobus 2, 28, 3; 4 BC)

Frequently, the non-active verb denoting personal interest (benefactive meaning) or relationship with the subject is accompanied by a reflexive or a possessive pronoun: (55) a.ȱ ΓϡȱΈȂȱȱ ȱ ο΅ΙΘϲΑȱȱ πΔ΍ΗΚΣΒ΅ΗΌ΅΍ȱȱȱ  hoi d’ heauton episphaksasthai the.NOM others himself.ACC draw-dagger.MID.AOR.INF ‘while others [say that] he drew his dagger’ (Xenophon, Anabasis, 1, 8, 29; 5-4 BC) b.ȱ πΔΉΈΉϟΒ΅ΑΘΓȱȱ ΘΤΖȱȱ ΅ЀΘЗΑȱȱ ΦΕΉΘΣΖ epedeiksanto tas hauto:n aretas

84

Chapter Three display.MID.AOR.3PL the.ACC themselves.GEN valors.ACC ‘they displayed their valor’ (Isocrates, 4 (Panegyricus), 85; 5-4 BC)

At the same time, there is evidence of instances where the personal relationship is denoted by the presence of the possessive pronoun close to an active type (in the following example, the structure “active type + possessive pronoun + object” is contradistinguished by the structure “non-active type + object”). (56) Γϡȱȱ ΩΑΈΕΉΖȱȱ Γϡȱ Φ·΅ΌΓϠȱȱ ΘΤȱȱ hoi andres hoi agathoi ta

ΐξΑȱȱΩΏΏ΅ȱȱΘΓϿΖȱȱȱȱȱ΅ЀΘЗΑȱȱȱ

men alla

the men.NOM the good.NOM the.ACC prt

ȱ

ȱ

ΘΤȱȱ

ΐξΑȱȱ

ΩΏΏ΅ȱȱ

tous

other

ȱΈ΍ΈΣΗΎΓΙΗ΍ǯǯǯȱȱ

ȱȱЀΉϧΖȱ

hauto:n

hyeis

didaskousi...

the.ACC themselves.GEN sons.ACC teach.ACT.PRES.3PL

ΩΕ΅ȱȱ

ΘΓϿΖȱȱ

ЀΉϧΖȱȱ

Έ΍ΈΣΗΎΓΑΘ΅΍ȱ

ta men alla ara tous hyeis didaskontai the.ACC prt other.ACC therefore the.ACC sons.ACC teach.NACT.PRES.3PL ‘good men have their sons instructed in everything else (but this)... therefore they have their sons instruct in everything else’ (Plato, Protagoras, 325b, 3; 5-4 BC)

The reflexive non-active verbs are able to take a direct object in the accusative case in classical Ancient Greek (ΏΓϾ˜ΐ΅΍ȱΘϲȱΗЗΐ΅ /ȱlouomai to so:ma / wash.NACT.PRES.1SG the.ACC body.ACC / ‘I wash my body’), if the object-goal is directly related to the subject (something that is not possible in Modern Greek, where the non-active reflexive verbs are obligatorily intransitive)26. a. Ancient Greek ȱȱ ΏΓϾΓΐ΅΍ȱȱ ΘΤΖȱȱ louomai tas wash.NACT.PRES.INF the.ACC ‘s/he washes her/his hands’ ȱȱ ȱȱ ȱ ȱ ȱ ȱ

ΛΉϧΕ΅Ζȱȱȱ

ȱ

kheiras hands.ACC

ΑϟΊΓΐ΅΍ȱΘΤΖȱΛΉϧΕ΅ΖȱȦȱnizomaiȱtas kheirasȱȦȱ‘to wash my hands’ ΎϱΔΘΓΐ΅΍ȱΘχΑȱΎΉΚ΅ΏφΑȱȦȱkoptomai te:n kephale:n / ‘I beat my head’ π·Ύ΅ΏϾΔΘΓΐ΅΍ȱegkalyptomai ‘to veil/wrap up’ ΔΉΕ΍ΆΣΏΏΓΐ΅΍ȱperiballomai ‘to throw round’ ΔΉΕ΍ΘϟΌΉΐ΅΍ȱperitithemai ‘to place-put round’

b. Modern Greek *ʌȜȑȞȠȝĮȚ *plenome wash.NACT.PRES.1SG

IJĮ ȤȑȡȚĮ ta xerja the.ACC hands.ACC

Έχȱȱ ΏΓϾΓΑΘ΅΍ȱȱ ЂΈ΅Θ΍ȱȱ Θϲȱȱ Δ΅ΕΣΔ΅ΑȱȱΘϲȱȱΗЗΐ΅ȱ (57) a.ȱ ΓЁȱȱ·ΤΕȱ ou gar de: louontai hydati to parapan to so:ma not because prt wash.NACT.PRES.3PL water.DAT the.ACC altogether the body ‘for they never wash their bodies with water’ (Herodotus, The Histories, 4, 75, 7; 5 BC) b.ȱ ϳȱΐξΑȱȱ ΦΗΔϟΈ΅ȱȱ ΦΑ΅΍ΕΉϧΘ΅΍ǰȱȱ ϳȱΈξȱȱ ΌЏΕ΅Ύ΅ȱȱ πΑΈϾΉΘ΅΍ǰȱ ho men aspida anaireitai, ho de tho:raka endyetai, the one.NOM shield.ACC raise.NACT.PRES.3SG the other.NOM

ȱ

ȱ

ȱ

corslet.ACC wear.NACT.PRES.3SG

ϳȱΈξȱȱ

ΎΑ΋ΐϧΈ΅Ζȱȱ

ύȱȱ ΎΕΣΑΓΖȱȱ

ύȱȱ ΊΝΗΘϛΕ΅ȱȱ

ΔΉΕ΍ΘϟΌΉΘ΅΍

ho de

kne:midas

e: kranos

e: zo:ste:ra

peritithetai

the other.NOM

leggings.ACC or

helmet.ACC or warrior's-belt.ACC

put.NACT.PRES.3SG

‘the one is raising his shield, the other is wearing his corslet, the other is putting on his leggings or his helmet or his belt’ (Aristotle, On the cosmos, 399b; 4 BC)

The non-active voice is also used in transitive structures with deponent verbs (for an analysis of the deponent verbs in diachrony, cf. Lavidas & Papangeli 2006). (58) π·Аȱȱ ̍ΏΉ΍Αϟ΅Αȱȱ ego: Kleinian I.NOM Cleinias.ACC

ϊΈ΍ΓΑȱȱ

ΐξΑȱȱ

ΌΉЗΐ΅΍ȱȱ

ύȱȱ

ΘΫΏΏ΅ȱȱ

ΔΣΑΘ΅

he:dion

men

theo:mai

e:

talla

panta

more-pleasantly

prt

gaze.NACT.PRES.1SG than

the-other.ACC all

‘I would rather gaze at Cleinias than at all the other (beautiful objects in the world)’ (Xenophon, 26

According to the above, we would be able to hypothesise that non-active morphology in classical Ancient Greek denoted [+subject affectedness] (meaning that also exists in the three structures: passive, anticausative, and reflexive), whereas Hellenistic-Roman Koine should comprise the intermediate stage towards the denotation of [-transitive] by non-active morphology (Lavidas 2002, 2003b, 2005).

Diachronic Data and Analysis

85

Symposium, 4, 12; 5-4 BC)

(b) Active voice in passive structures In Homeric Greek (as we have also seen from the beginning of the section), the presence of the agent (in PP or dative case) was particularly systematic with active verbs and the patient argument as subject. This concerns typical instances of passive structure but with active morphology. Jankuhn (1969) observes the phenomenon but fails to give a systematic interpretation: Ά΅ΕϾΌΝȱbarytho: ‘to ach’(Homer, Iliad, 16, 519; 8 BC)ȱ Ύ΅ΘφΕ΍ΔΉȱkate:ripe ‘to crumble’(Homer, Iliad, 5, 92; 8 BC)ȱ ΔΣΗΛΝȱpaskho: ‘to suffer’ (Homer, Odyssey, 2, 134; 8 BC)ȱ ΚΉϾ·Νȱpheugo: ‘to flee/ run away’ (Homer, Iliad, 21, 553, Homer, Iliad, 18, 149; 8 BC)ȱ

The same phenomenon is also encountered in Classical Greek, where in many instances there is evidence of structures of intransitive active verbs with a PP denoting the agent (ĺ passive structure) or the cause (ĺ passive or anticausative structure) (cf. George 2005). ΔϟΔΘΝȱpipto: ‘to fall down/cast oneself down/fall violently upon’ ǻΦΔΓǼΌΑφ΍ΗΎΝȱ(apo)thne:isko: ‘to die’ πΎΔϟΔΘΝ ekpipto: ‘to fall out/fall down/be dislocated/die’ ΔΣΗΛΝ paskho: ‘to suffer’ ΚΉϾ·Ν pheugo: ‘to flee/avoid/escape’ȱ ΘΉΏΉΙΘЗȱteleuto: ‘to accomplish/die’

Consequently, in Classical Greek, a passive structure with active morphology and the presence of an agent is not ungrammatical. ȉhis is in contrast to Modern Greek, where the denotation of the agent in anticausative structures is ungrammatical (*ʌȑșĮȞİ Įʌȩ IJȠȞ ȀȫıIJĮ / *pethane apo ton Kosta / die.ACT.PAST-PERFVE.3SG by the Kostas / ‘*s/he died by Kostas – she was killed by Kostas’). Examples from the different categories are given below: passive structure with active morphology and PP-agent (59) a. ΦΔνΌ΅ΑΉȱȱ ЀΔϲȱȱ ̝Ό΋ΑΣΈΉΝȱȱ ΦΑΈΕϲΖȱȱ ̖Ε΋Λ΍ΑϟΓΙȱȱȱ apethane hypo Athe:nadeo: andros Tre:khiniou die.ACT.AOR.3SG by Athenades man Trachinian ‘he was killed by Athenades, a Trachinian’ (Herodotus, The Histories, 7, 213, 2-3; 5 BC) b.ȱ ΅ЁΘΓϟȱ ·Ήȱ ΦΔνΌΑϙΗΎΓΑȱ ЀΔϲȱȱ ϡΔΔνΝΑȱȱȱȱȱ ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ autoi ge apethne:iskon hypo hippeo:n they.NOM prt die.ACT.IMPERF.3PL by cavalry ‘they were killed by any of the enemy's cavalry’ (Xenophon, Cyropaedia, 7, 1, 48; 5-4 BC) c.ȱ ΓϡȱΈȂȱȱ ΩΏΏΓ΍ȱȱ ΚΉϾ·ΓΑΘΉΖȱ σΔ΍ΔΘΓΑǰȱȱ ΔΓΏΏΓϠȱȱ ΐξΑȱȱ ЀΔϲȱϡΔΔνΝΑ hoi d’ alloi pheugontes epipton, polloi men hypo hippeo:n ȱ the rest.NOM flee.PART.NOM die.ACT.IMPERF.3PL many.NOM prt by horsemen ‘and the rest were cut down as they fled, many by the horsemen’ (Xenophon, Hellenica, 7, 1, 31; 5-4 BC) ΐξΑȱȱ ΘΉΏΉΙΘλȱȱȱȱȱȱȱȱ ΈϱΏУȱȱ ЀΔϲȱȱ ̕Δ΅Ε·΅ΔΉϟΌΉΓΖȱ d.ȱ ̝Ε΍΅ΔΉϟΌ΋Ζȱȱ Ariapeithe:s men teleutai dolo:i hypo Spargapeitheos Ariapithes.NOM prt die.ACT.PRES.3SG treachery.DAT by Spargapithes ‘Ariapithes was treacherously killed by Spargapithes’ (Herodotus, The Histories, 4, 78, 8; 5 BC) e.ȱ ΉϢΖȱΘχΑȱΉϡΕΎΘχΑȱȱ πΗΔϟΔΘΉ΍ȱ ЀΔϲȱȱ ΘЗΑȱȱ πΚϱΕΝΑȱȱȱȱ eis te:n heirkte:n espiptei hypo to:n ephoro:n in the prison throw.ACT.PRES.3SG by the Ephors ‘he is thrown into prison by the Ephors’ (Thucydides, The Peloponnesian War, 1, 131, 2; 5 BC)

 intransitive structure with active morphology and PP-cause (60) a.ȱ ϵΘ΍ȱȱ ЀΔϲȱȱ ΘЗΑȱ ̐ΙΐΚЗΑǯǯǯȱ Η΅ΚЗΖȱȱ πΑΌΓΙΗ΍ΣΗΝȱȱ ȱ hoti hypo to:n Nympho:n... sapho:s enthousiaso: that by the nymphs surely inspire.ACT.FUT.1SG ‘I shall surely be possessed of the nymphs’ (Plato, Phaedrus, 241e; 5-4 BC) b.ȱ ΘΕΙΚλΖȱȱ ЀΔϲȱȱ ΔΏΓϾΘΓΙȱȱ ΘϛΖȱȱ ȱ ΗΓΚϟ΅Ζȱȱȱ tryphais hypo ploutou te:s sophias be-indolent.ACT.PRES.2SG by wealth the.GEN wisdom.GEN  ‘you are indolent on account of your wealth of wisdom’ (Plato, Euthyphro, 12a; 5-4 BC)  c.ȱ ΐΉΌϾΉ΍Αȱȱ ЀΔϲȱȱ ΘΓІȱȱ ΓϥΑΓΙȱȱ ΦΚ΍ΒϱΐΉΌ΅ȱȱ methyein hypo tou oinoi aphiksometha

86

Chapter Three be-drunken.ACT.PRES.INF by the wine arrive.NACT.FUT.1PL ‘be driven on by the wine to a state of intoxication’ (Xenophon, Symposium, 2, 26; 5-4 BC)

According to the above, it emerges that in many instances passive structure (with the presence of the agent in PP) without a non-active verb ending is possible. (61)ȱ ω··νΏΌ΋ȱ e:ggelthe:

ΘΓϧΖȱȱ

ΘЗΑȱȱ

̕ΙΕ΅ΎΓΗϟΝΑȱΗΘΕ΅Θ΋·ΓϧΖ ΓϥΎΓΌΉΑȱϵΘ΍ȱ

tois

to:n

Syrakosio:n

ΚΉϾ·Γ΍ΉΑȱȱЀΔϲȱΘΓІȱΈφΐΓΙȱȱȱȱ

strate:gois oikothen hoti pheugoien hupo tou de:mouȱ

announced.3SG the.DAT the.GEN Syracusan.GEN generals.DAT from-home

that

expel.ACT.PRES.3PL by people

‘word came from home to the Syracusan generals that they had been banished by the citizens’ (Xenophon, Hellenica, 1, 1, 27; 5-4 BC)

It should be noted that the verb ijİȪȖȦ pheugo: in Classical Greek has the meaning of ‘I am exiled/I run away, ‘they expel me/I undergo persecution’ and is used as both a transitive and intransitive verb: (62)ȱ ΚΉϾ·Ή΍Ζǰȱȱ σΚ΋ǰȱȱ Иȱȱ ̕ЏΎΕ΅ΘΉΖǰȱȱ ϳȱȱ ̇΍ΓΑΙΗϱΈΝΕΓΖ pheugeis, ephe:, o: So:krates, ho Dionysodo:ros expel.ACT.PRES.2SG said.3SG the.VOC Socrates.VOC the.NOM Dionysodorus.NOM ‘you are running away, Socrates, said Dionysodorus’ (Plato, Euthydemus, 297b, 4; 5-4 BC)

(c) Non-active voice in passive structures In Classical Greek, we also find, however, the first signs of productive use of non-active morphology in passive structures. The agent in these structures is referred to only exceptionally (Ernout 1908-9: 329 ff.; Schwyzer 1943), despite its presence with active intransitive structures as well. The necessity or not of the presence of the agent (in a PP or DP in the dative/ablative case) for the analysis of a structure as passive in Classical Greek has been a subject of particular discussion in the literature (Wackernagel 1904; Kuryłowicz 1935, 1964; Meillet 1903/1964; Schmidt 1963a, b; Jankuhn 1969: 22-27; George 2005). In instances where the PP appears, the choice of a specific agent- or cause-PP depends (cf. George 2005): (a) on the semantic group to which the verb belongs: verbs that mean ‘to give’ and ‘to send’ (ditransitive verbs) are accompanied by the PPs πΎ ek, Δ΅ΕΣ para + Genitive, verbs that mean ‘to think’ are accompanied by the PPs ΔΕϱΖ pros, Δ΅ΕΣ para + Dative; and (b) on the finiteness of the verb; the participles frequently take different PPs from finite verbs. For Luraghi (2000: 279-285; 2003: 128-129) prepositions πΎȱ ek and ΔΕϱΖȱ pros accompany verbs with a low degree of transitivity (according to Hopper & Thompson’s (1980) scale). In contrast, verbs with a high degree of transitivity (transitivity is related to the presence of the meaning of the prototypical agent in these structures) take the preposition ЀΔϱȱhypo. x

ЀΔϱhypo+ genitive : usually animate (but also +cause) πΘ΍ΐκΘΓȱȱ (63) Ήϥȱȱ Θ΍Ζȱȱ

ЀΔϲȱȱ ΘΓІȱȱ ΈφΐΓΙ


ei tis etimato hypo tou de:mou if someone.NOM honour.NACT.IMPERF.3SG by the people   ‘if someone was honoured by the commons’ (Xenophon, Hellenica, 2, 3, 15; 5-4 BC)

x

x

x

(64) ωΑΣ·Ύ΅Ηΐ΅΍ȱȱ ЀΔϲȱȱ ΘЗΑȱȱ ·Ή·ΉΑ΋ΐνΑΝΑ e:nagkasmai hypo to:n gegene:meno:n complel.NACT.PERF.1SG by the.GEN occur.GEN.PART ‘I have now been compelled by what has occurred’ (Lysias, 12 (Against Eratosthenes), 3; 5-4 BC)  πΒȱeks ǻπΎȱek)+ genitive : usually cause  ΗΓІȱȱ Ά΍ΣΊΓΑΘ΅΍ȱȱ ΘΣΈΉ   (65) πΎȱȱ ek sou biazontai tade from you press.NACT.PRES.3PL these.NOM ‘they are hard pressed by you’ (Sophocles, Antigone, 1073; 5 BC) ΔΕϱΖpros+ genitive : usually animate (but also +cause) (66) πΌ΋ΏϾΑΌ΋Αȱȱ ΗΘϱΐ΅ȱȱ ΔΕϲΖȱΘϛΗΈΉȱ ΘϛΖȱȱ ·ΙΑ΅΍ΎϱΖ ethe:lynthe:n stoma pros te:sde te:s gynaikos emasculate.PASS.AOR.3SG mouth.NOM from that the woman ‘(I felt) my mouth emasculated by this woman's words’ (Sophocles, Ajax, 651; 5 BC) Δ΅ΕΣpara+ genitive : animate (67) ΘΤȱȱ ΈЗΕ΅ȱȱ ΔνΐΔΉΘ΅΍ȱȱ Δ΅ΕΤȱȱ ΘΓІȱ Ά΅Η΍ΏΉϾΓΑΘΓΖ ta do:ra pempetai para tou basileuontos the.NOM gifts.NOM send.NACT.PRES.3SG by the king ‘gifts are sent by the (successive) kings (of Persia)’ (Herodotus, The Histories, 7, 106; 5 BC)

Diachronic Data and Analysis

x

x

x

87

(68) Δ΅ΕΤȱȱ ΔΣΑΘΝΑȱȱϳΐΓΏΓ·ΉϧΘ΅΍ȱȱ  para panto:n homologeitai from all agree.NACT.PRES.3SG ‘as is agreed from all’ (Xenophon, Anabasis, 1, 9, 1; 5-4 BC) Έ΍Τdia+ genitive : instrument  (69) Έ΍Τȱȱ ΘΓІΘΝΑȱȱ ΧΔ΅ΑΘ΅ȱȱ ΔΕΣΘΘΉΘ΅΍ȱȱȱȱȱȱ ȱȱȱȱȱȱȱȱȱȱ dia touto:n hapanta prattetai from they.GEN all.NOM manage.NACT.PRES.3SG ‘everything is managed by them’ (Demosthenes, Olynthiac 3, 31, 3; 4 BC) ΦΔϱapo+ genitive : animate (agent)  Έξȱȱ ΓЁΈξΑȱȱ ΦΔȂȱȱ ΅ЁΘЗΑȱȱ σΕ·ΓΑȱȱ ΦΒ΍ϱΏΓ·ΓΑ (70) πΔΕΣΛΌ΋ȱȱ eprakhthe: de ouden ap’ auto:n ergon aksiologon do.PASS.AOR.3SG prt nothing.NOM by they.GEN deed.NOM great.NOM ‘nothing great was achieved by them’ (Thucydides, The Peloponnesian War, 1, 17, 1; 5 BC) dative: (usually with verbal adjectives or perfect) animate (agent)  (71) ГΖȱȱ ΐΓ΍ȱȱ ΔΕϱΘΉΕΓΑȱȱ ΈΉΈφΏΝΘ΅΍ȱȱȱȱȱ ȱȱȱȱȱȱȱȱȱȱ ho:s moi proteron dede:lo:tai as I.DAT before indicate.NACT.PERF.3SG ‘as it has been indicated by me before’ (Herodotus, The Histories, 1, 18, 10; 5 BC) (72) ΔνΔΕ΅ΎΘ΅΍ȱȱ ΘΓϧΖȱȱ ΩΏΏΓ΍Ζȱȱȱȱ ȱ ȱȱȱȱȱȱȱȱȱȱ pepraktai tois allois do.NACT.PERF.3SG the.DAT others.DAT ‘(critics of) what was conducted by others’ (Demosthenes, Olynthiac 2, 27, 6; 4 BC) (73) ΔΓΏΏ΅Ϡȱȱ ΌΉΕ΅ΔΉϧ΅΍ȱȱ Ύ΅Ϡȱȱ Δ΅ΑΘΓΈ΅Δ΅ϠȱΘΓϧΖȱȱ Ϣ΅ΘΕΓϧΖȱȱ ΉЂΕ΋ΑΘ΅΍ȱ pollai therapeiai kai pantodapai tois iatrois heure:ntai many treatments.NOM and all-kind.NOM the physicians.DAT discover.NACT.PERF.3PL ‘many treatments of all kinds have been discovered by physicians’ (Isocrates, On the peace, 39; 5-4 BC)

According to what we have seen above, in classical Ancient Greek, the presence of the following arguments in the position of the subject of non-active verbs (and the assignment of nominative case) is possible: (a) benefactive; the person who benefits from the action or the interested person: (74)ȱ БΖȱȱ πΐ΅ΙΘХȱȱ ΌΕνΐΐ΅ȱȱ ΌΕΉΜ΅ϟΐ΋Αȱȱ π·Џ ho:s emauto:i thremma threpsaime:n ego: as myself.DAT nursling.ACC rear.MID.AOR.1SG I.NOM ‘I reared him myself as my own foster-son’ (Sophocles, Oedipus Tyrannus, 1143; 5 BC)

(b) theme/patient: (75)ȱ Θ΍Ζȱȱ πΘ΍ΐκΘΓȱȱ ЀΔϲȱȱ ΘΓІȱȱ ΈφΐΓΙȱȱ tis etimato hypo tou de:mou someone.NOM honour.NACT.IMPERF.3SG by the people ‘s/he was honoured by the commons’ (Xenophon, Hellenica, 2, 3, 15; 5-4 BC)

(c) The presence in the subject position of the argument that denotes the goal (in the genitive, dative, or second accusative case) with ditransitive verbs is also possible in Classical Greek, whereas the argument of the theme/patient remains in the accusative case since the non-active type is able to assign accusative case to the argument of the theme/patient. Consequently, I cite examples of passivisation in Classical Greek: (i) With (mono)transitive verbs: the argument that receives the dative or genitive case in the transitive structure is now in the subject position of the non-active verb. πΔ΍ΆΓΙΏΉϾΝȱepibouleuo: + DP-DAT ‘to plot/contrive against’ πΔ΍ΆΓΙΏΉϾΓΐ΅΍ȱepibouleuomai + by-PPȱ‘be the object of plots’

ȱ

ΚΌΓΑЗȱ™hthonȠ: + DP-DAT ‘to bear grudge/be envious-jealous’ ΚΌΓΑΓІΐ΅΍ȱphthonoumai + by-PP ‘to be envied/begrudged’ Δ΍ΗΘΉϾΝȱpisteuo: + DP-DAT ‘to trust/put faith in/rely on a person’ Δ΍ΗΘΉϾΓΐ΅΍ȱpisteuomai + by-PP ‘to be entrusted with a thing’ȱ

ȱ ΦΔ΍ΗΘЗȱapisto: + DP-DATȱ‘to disbelieve/distrust’

ȱ

ȱ

ȱ

88

Chapter Three

ΦΔ΍ΗΘΓІΐ΅΍ȱapistoumai + by-PPȱ‘to be distrusted’

ȱ ΔΓΏΉΐЗȱpolemo: + DP-DAT ‘to make hostile/make an enemy of’ ΔΓΏΉΐΓІΐ΅΍ȱpolemoumai + by-PP ‘to be made an enemy of’ ȱ

(76) ȱ

ЀΔдȱȱ ̝Ό΋Α΅ϟΝΑȱȱ πΔ΍ΆΓΙΏΉΙϱΐΉΌ΅ȱ he:meis hyp’ Athe:naio:n epibouleuometha  we.NOM by Athenians plot.NACT.PRES.1PL ‘we are the objects of the designs of the Athenians’ (Thucydides, The Peloponnesian War, 1, 82; 5 BC)

ψΐΉϧΖȱȱ

ϴΏ΍·ΝΕЗȱoligo:ro: + DP-GEN ‘to esteem lightly/make small account of’ ȱ ϴΏ΍·ΝΕΓІΐ΅΍ȱoligo:roumai + by-PP ‘to be neglected’ ΩΕΛΝȱarkho: + DP-GEN ‘to make a beginning of/lead’ȱ ΩΕΛΓΐ΅΍ȱarkhomai + by-PP ‘to be ruled/governed’

(77) ȱ Θ΅ІΘ΅ȱȱ ·ΤΕȱȱ ΔΕЗΘΓΑȱ ВΏ΍·ΝΕφΌ΋ȱȱ ΐΉΘΤȱ ΘχΑȱȱ ΉϢΕφΑ΋Α tauta gar pro:ton o:ligo:re:the: meta te:n eire:ne:n these.NOM because first disregard.PASS.AOR.3SG after the peace ‘for these were the places that were disregarded immediately after the peace’ (Demosthenes, Phillipic 4, 8; 4 BC)

(ii) With ditransitive verbs: as can be seen from the following examples, the assignment of accusative case from a non-active verb is possible in Classical Greek: (78) a. πΎȱȱ ΘϛΖȱȱ ΘΓϾΘΓΙȱȱ ΛΉ΍ΕϲΖȱ πΘ΅ΐ΍ΉΙϱΐΉΌ΅ȱȱ ΘΓϿΖȱȱ ΑϱΐΓΙΖȱ ek te:s toutou kheiros etamieuometha tous nomous  from the he.GEN hand dispense.NACT.IMPERF.1PL the.ACC laws.ACC ‘the laws were dispensed to us from his hands’ (Lysias, Against Nicomachus, 30, 3; 5-4 BC)  ΘχΑȱȱ ΦΕΛφΑ b. ЀΔϲȱȱ ΘΓІȱȱ ̏΍Ώ΋ΗϟΓΙȱȱ ̝Ε΍ΗΘΓ·ϱΕΉΝȱȱ ΦΔΉΗΘνΕ΋ΘΓȱȱ hypo tou Mile:siou Aristogoreo: apestere:to te:n arkhe:n by the Milesian Aristagoras deprive.NACT.PLUP.3SG the.ACC rule.ACC ‘he was deprived of his rule by Aristagoras of Miletus’ (Herodotus, The Histories, 6, 13; 5 BC)  c. ΩΏΏΓȱȱ Θ΍ȱȱ ΐΉϧΊΓΑȱȱ ΉЁΌϿΖȱȱ πΔ΍Θ΅ΛΌφΗΉΗΌΉȱ allo ti meizon euthys epitakhthe:sesthe other.ACC something.ACC more.ACC immediately dictate.PASS.FUT.2PL  ‘you will be immediately dictated some more oppressive condition’ (Thucydides, The Peloponnesian War, 1, 140; 5 BC)

ȱ In Classical Greek, in cases of non-active ditransitive verbs, the argument denoting the (animate) person was marked with nominative case (as the subject of the non-active verb), whereas the other argument remained in the accusative case in the passive structure as well: ΦΚ΅΍ΕЗȱaphairo: ‘to take away/exclude/separate’ Θ΅ΐ΍ΉϾΝ tamieuo: ‘to deal out/dispense/manage’

+1. DP-ACC +1. DP-ACC

+2. DP-ACC / GEN +2. DP-GEN / DAT

ϷΜ΍Αȱȱ ΦΚ΅΍ΕΉϧΘ΅΍   (79) ΘχΑȱȱ te:n opsin aphaireitai the.ACC eye.ACC take-away.NACT.PRES.3SG ‘(but if any attempts to gaze recklessly upon him), he blinds their eyes / they are blinded (their eyes)’ (Xenophon, Memorabilia, 4, 3 ,14; 5-4 BC)

It is also possible with non-active ditransitive verbs for the DP in the accusative case (of the corresponding structure with the active type) to change into nominative and for the DP in the dative case to remain in the dative: (80)ȱ ΘΓϧΗ΍ȱȱ πΔΉΘνΘΕ΅ΔΘΓȱȱ ψȱȱ ΚΙΏ΅Ύφ toisi epetetrapto he: phylake: they.DAT charge.NACT.PLUP.3SG the.NOM guard.NOM ‘to them (Ionians) the guard had been charged’ (Herodotus, The Histories, 7, 10, 34; 5 BC)

Diachronic Data and Analysis

89

3.1.3.3 Causatives and anticausatives: voice morphology The distribution of voice morphology in causative and anticausative types appears to be highly systematic. Causative verbs in classical Ancient Greek are consistently active, whereas anticausatives are denoted with non-active voice; there are very few instances of anticausatives that can bear both (without a difference in meaning) non-active and active voice morphology. (a) Consistent active causative – non-active anticausative Causative verbs bear consistently active morphology; the use of non-active endings for the denotation of “personal interest of the subject for the verb action” (benefactive meaning) with causatives types is not possible. Anticausative verbs (even those verbs denoting natural, innate affectedness of the subject) are also marked consistently with a non-active ending (the active types in brackets are for the causative member of the alternation): ΚϾΓΐ΅΍ȱǻΚϾΝǼȱȮȱphyomai (phyo:) ‘to grow/wax’ ΘΕνΚΓΐ΅΍ȱǻΘΕνΚΝǼȱȮȱtrephomai (trepho:) ‘to grow/increase’ ΗφΔΓΐ΅΍ȱǻΗφΔΝǼȱȮȱse:pomai (se:po:) ‘to rot/moulder’ ΘφΎΓΐ΅΍ȱǻΘφΎΉΝǼȱȮȱte:komai (te:ko:) ‘to melt/be dissolved/melt away’ Ϲφ·ΑΙΐ΅΍ȱǻϹφ·ΑΙΐ΍ǼȱȮȱrhe:gnymai (rhe:gnymi) ‘to break/break asunder/burst’

(81) a. Θϲȱȱ ΗЗΐ΅ǯǯǯȱȱ ΓЁΎȱȱ πΐ΅Ε΅ϟΑΉΘΓȱȱ to so:ma... ouk emaraineto the.NOM body.NOM not wither.NACT.IMPERF.3SG ‘the body (meanwhile)... did not waste away’ (Thucydides, The Peloponnesian War, 2, 49; 5 BC) ǻΑϱΗΓΖǼȱȱ ΐ΅Ε΅ϟΑΉ΍ȱȱ ΐΉ b. Ψȱȱ ha (nosos) marainei me which.NOM (illness) wither.ACT.PRES.3SG I.ACC ‘(and have named the heaven-sent plague) that wastes (and stings) me (with its maddening goad)’ (Aeschylus, Prometheus Bound, 597; 6-5 BC) (82) a. ǻΏϟΐΑ΋Ǽȱȱ ΉϨΘȇȱȱ πΒ΋ΕΣΑΌ΋ȱȱ ΩΑȱȱȱȱȱȱȱȱȱȱȱ (limne:) eit’ ekse:ranthe: an lake.NOM afterwards dry.PASS.AOR.3SG prt ‘(the lake) would become dry afterwards’ (Aristotle, Meteorology 353a; 4 BC) ΗȂȱȱ ϳȱȱ ̅ΣΎΛ΍ΓΖ b. Β΋Ε΅ΑΉϧȱȱȱ kse:ranei s’ ho Bakkhios dry.ACT.FUT.3SG you.ACC the.NOM Dionysus.NOM ‘Dionysus will parch you up’ (Euripides, Cyclops, 575; 5 BC)

With the aforementioned verbs, the presence of a DP denoting the cause (or the agent in structures with passive interpretation) is very frequent. (83) ГΗΘΉȱȱ ΘΤȱȱ ho:ste ta so

ΚΙΘΤȱȱ

ΎϟΑΈΙΑΓΖȱȱ ЀΔϲȱȱ ΐξΑȱȱ ΘΓІȱȱ

ЂΈ΅ΘΓΖȱȱ

ΗφΔΉΗΌ΅΍

phyta

kindynos

hypo men tou

hydatos

se:pesthai

by

water.GEN

rot.NACT.PRES.INF

the.ACC plants.ACC risk.NOM

prt

the

‘so the plants would run the risk of damping off through too much water’ (Xenophon, Economics, 19, 11, 7; 5-4 BC)

The corresponding transitive (active) verbs denote also the cause or the agent that affects the nature of the patient argument: (84) ΅ϡȱ ψΗΙΛϟ΅΍ȱ hai he:sykhiai

ΗφΔΓΙΗ΍ȱȱ

Ύ΅ϠȱȱΦΔΓΏΏϾ΅Η΍ȱȱ

ΘΤȱȱ

ΘϛΖȱȱ

ΔϱΏΉΝΖȱΔΕΣ·ΐ΅Θ΅

se:pousi

kai apollyasi

ta

te:s

poleo:s

pragmata

the stillness.NOM rot.ACT.PRES.3PL and destroy.ACT.PRES.3PL the.ACCthe.GEN city.GEN situation.ACC

‘stillness causes decay and destruction of the city’ (Plato, Theaetetus, 153c, 7; 5-4 BC)

The instances of verbs with alternate (without any difference in meaning) use of the non-active and active anticausative type are limited (see the following examples in 85-88). These instances reinforce a different way of dealing with the voice morphology of anticausative verbs from the voice morphology of the other structures. Possibly, this alternation of active and non-active form reflects a corresponding morphological alternation of the intransitive type before it was causativised. A second possibility is that the active form arose as a result of a morphological ongoing change. The active form appears in those few verbs in free distribution with the non-active form, to prevail in the next period. I will return to this topic when examining data from the Hellenistic period. A basic conclusion, in any case, is that the presence of the non-active form is consistent in the anticausative structures in Classical Greek, either as

90

Chapter Three

the only grammatical voice morphology in the majority of instances, or in free distribution with the active (as a result of an ongoing change) in very few instances. The arguments for the systematic presence of non-active voice in anticausative verbs of Classical Greek is strengthened by the following examples from the two semantic groups of verbs that have been described for Modern Greek, where they bear consistent active voice, by Alexiadou & Anagnostopoulou (2004). The anticausative verbs of both semantic groups bear non-active voice in Classical Greek, in contrast to the corresponding Modern Greek verbs. Concerning the first group (structure: BECOME + Predicate; ĮıʌȡȓȗȦ asprizo ‘to whiten’, cf. para. 2.1.4), the anticausative verb type in classical Ancient Greek consistently has a non-active ending: (85)ȱ πΎΎ΅Ό΅ϟΕΝ ekkathairo: ‘to cleanse/clear out’ a.ȱ intransitive πΑȱΘΓϾΘΓ΍ΖȱȱΘΓϧΖȱȱȱȱȱΐ΅Όφΐ΅Η΍ΑȱȱοΎΣΗΘΓΙȱϷΕ·΅ΑϱΑȱȱȱȱȱΘ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱΜΙΛϛΖȱȱȱȱȱȱȱπΎΎ΅Ό΅ϟΕΉΘ΅ϟȱȱ

en toutois

tois

mathe:masin hekastou organon

ti

psykhe:s ekkathairetai

in these.DAT the.DAT studies.DAT every.GEN instrument.NOM some.NOM soul.GEN cleanse.NACT.PRES.3SG

‘there is in every soul an instrument (of knowledge) that purifies (and kindles afresh) by such studies (when it has been destroyed and blinded by our ordinary pursuits)’ (Plato, Republic, 527d, 8; 5-4 BC) b. transitive ΦΔΓΔΕϟΗ΅Ζȱȱ

ǽρΎ΅ΗΘΓΖǾȱȱΔκΑȱȱȱȱȱΘϲȱȱ σΑΉΕΌΉȱȱȱΘЗΑȱ

ϴΚΕϾΝΑȱȱȱȱȱȱȱȱπΎΎ΅Ό΅ϟΕΉ΍

apoprisas

[hekastos] pan

ophryo:n

saw-off.PART.NOM [each]

to enerthe to:n

all.ACC the beneath

ekkathairei

the.GEN eyebrows.GEN cleanse.ACT.PRES.3SG

‘each saws off all the part beneath the eyebrows, and cleans the rest’ (Herodotus, The Histories, 4, 65, 4; 5 BC) (86)ȱ ΎΏνΝȦȱΎΏΉϟΝ kleo:/kleio: ‘close’ a.
intransitive
БΖȱȱ ψȱȱ ΌϾΕ΅ȱȱ ΎΉΎΏΉϟΗΉΘ΅΍ȱ
ho:s he: thyra kekleisetai as the.NOM door.NOM close.NACT.FUTPERF.3SG ‘the door will then close’ (Aristophanes, Lysistrata, 1071; 5-4 BC) b. transitive ȱ

ȱ

ΎΏΉϠΖȱϳΐΝΑϾΐΝΖȱȱȱȱȱϊȱȱȱΘΉȱȱȱЀΔϲȱȱΘϲΑȱ΅ЁΛνΑ΅ȱȱΘЗΑȱΊФΝΑȱȱΎ΅ϠȱϗȱȱȱȱΘΤΖȱΌϾΕ΅ΖȱȱΎΏΉϟΓΙΗ΍Α

kleis homo:nymo:s he: te hypo ton aukhena to:n zo:io:n kai he:i tas thyras kleiousin kleis equivocally

which and under the neck

the livings and which the doors.ACC close.ACT.PRES.3PL

‘the equivocal use of the word kleis (key) which denotes both the bone at the base of the neck and the instrument with which we lock our doors’ (Aristotle, Nicomachean Ethics, 1129a, 31; 4 BC) (87) ΘφΎΝ te:ko: ‘to melt’ intransitive ϳȱȱ ΦΗΌΉΑνΝΑȱЀΔϲȱ ϴΈΙΑνΝΑȱ ϢΗΛΙΕЗΑȱȱ a. ΘφΎΉΘ΅΍ȱȱ te:ketai ho astheneo:n hypo odyneo:n iskhyro:n melt.NACT.PRES.3SG the.NOM sick.NOM by pains.GEN strong.GEN ‘the sick is wasting away by the strong pains’ (Hippocrates, Epidemics, i-iii, 1, 19, 20; 5-4 BC) b. ΘφΎΉΘ΅΍ȱȱ ·ΤΕȱ πΑȱ ΅ЁΘϜȱȱ Θϲȱȱ ȱ ΚΏν·ΐ΅ ȱ te:ketai gar en aute:i to phlegma melt.NACT.PRES.3SG as in her the.NOM phlegm.NOM ‘as the phlegm (one of the four humours in the body) is melting in her body’ (Hippocrates, Epidemics, i-iii, 2, 1, 1; 5-4 BC) transitive ΔІΕȱȱȱȱȱȱ ȱȱΘφΎΉ΍ȱȱ ΘϲΑȱȱ ΎΕϾΗΘ΅ΏΏΓΑȱȱ c.ȱ ГΗΔΉΕȱȱΎ΅ϠȱΘϲȱȱ ho:sper kaiȱ to ȱpyr ȱȱte:kei ton  krystallon  as ȱ and the.ȃȅȂ ȱ fire.NOM ȱmelt.ACT.PRES.3SG the.ACC ice.ACC ‘as the fire melts the ice’ (Aristotle, Problems, 873a, 28; 4 BC) (88) ΐΉ΍З meio: lessen, diminish a. intransitive ȱ

Ύ΅ΎЗΖȱΈξȱȱ

ΘΓϾΘΝΑȱȱ

ΔΕ΅ΘΘΓΐνΑΝΑȱȱ Γϡȱȱ

kako:s de

touto:n

prattomeno:n

badly prt

ȱthese.GEN do.PART.GEN

ΓϨΎΓ΍ȱȱ

ΐΉ΍ΓІΑΘ΅΍ȱ

hoi ȱ

oikoi

meiountai 

the.NOM

goods.NOM

diminish.NACT.PRES.3PL

‘if they act incompetently, the goods are diminished’ (Xenophon, Economics, 3, 15; 5-4 BC) b. transitive ΐΉ΍Γϧȱȱ

Ύ΅ΌȇȱϵΗΓΑȱ ΪΑȱ

ΈϾΑ΋Θ΅΍ȱ

Θϲȱȱ

·Ή·ΉΑ΋ΐνΑΓΑȱ

meioi

kath’ oson an

dyne:tai

to

gegene:menon

Diachronic Data and Analysis

91

diminish.ACT.PRES.3SG as much prt can.3SG the.ACC occurrence.ACC ‘he belittles the occurrence as much as possible’ (Xenophon, Hiero, 2, 17, 5; 5-4 BC)

Concerning the verbs of the second group (ȗĮȡȫȞȦ zarono ‘to wrinkle’, ȗİıIJĮȓȞȦ zesteno ‘to warm’, ıțȓȗȦ skizo ‘to tear’, ıțȠȡʌȓȗȦ skorpizo ‘to scatter’, ȖțȡİȝȓȗȦ gkremizo ‘to demolish/collapse’, įȚĮȜȪȦ dialio ‘to dissolve’; with active morphology in Modern Greek when their subject is inanimate and they denote a partial change), in classical Ancient Greek, as the verbs in the first group, they are marked with non-active morphology: (89) Έ΍΅ΏϾΝ dialyo: ‘to dissolve’ a. intransitive πΒȱȱ ЙΑȱȱ

ΗϾ·ΎΉ΍Θ΅΍ȱȱ

Ύ΅ϠȱΉϢΖȱΨȱȱ

Έ΍΅ΏϾΉΘ΅΍

eks ho:n sygkeitai kai eis ha dialyetai from which.GEN compose.NACT.PRES.3SG and in which.ACC dissolve.NACT.PRES.3SG ‘from which it composes and into which it dissolves’ (Aristotle, On coming to be and passing away, 1, 8, 12; 4 BC) b. transitive ΑІΑȱȱ ΐξΑȱȱ

ΘϲΑȱȱ

ΗϾΏΏΓ·ΓΑȱȱ

ΘϱΑΈΉȱȱ

Έ΍ΣΏΙΗΓΑ

nyn men ton syllogon tonde dialyson now prt the.ACC assembly.ACC that.ACC dissolve.ACT.IMP.2SG ‘for now dismiss this assembly’ (Herodotus, The Histories, 7, 10, 41; 5 BC) (90)ȱπΕ΋ΐϱΝȱere:moo: ‘to abandon/desert/desolate’ a. intransitive ΉϢȱȱ ψȱȱ




ΔϱΏ΍Ζȱȱ

πΕ΋ΐΝΌΉϟ΋

ei he: polis ere:mo:theie: if the.NOM city.NOM desolate.PASS.AOR.3SG ‘if the city becomes desolate’ (Thucydides, The Peloponnesian War, 1, 10, 2; 5 BC) b.
transitive
ΘχΑȱȱ ΛЏΕ΅Α πΕ΋ΐΓІΑ te:n kho:ran ere:moun the.ACC land.ACC desolate.ACT.PRES.INF ‘to make the land (Attica) empty’ (Andocides, On the Peace, 21, 8; 5-4 BC)

(91) ΗΛϟΊΝ skhizo: ‘to tear’ a.
intransitive ΗΛϟΊΓΑΘ΅΍ȱȱ ΔΣΑΘϙȱȱ Ύ΅ΘΤȱΘχΑȱЀΗΘνΕ΅Αȱ

ǻ΅ϡȱΚΏνΆΉΖǼȱȱ (hai phlebes) skhizontai pante:i kata te:n hysteran (the.NOM veins.NOM) tear.NACT.PRES.3PL on-every-side in the ovary ‘(the veins) tear on every side in the ovary’ (Aristotle, Generation of Animals, 745b, 34; 4 BC) b. transitive ȱ

ϵΔΝΖȱȱ ΈΕІΑȱȱ ЀΏΓΘϱΐΓ΍ǰȱȱ

ΗΛϟΊΓΙΗ΍ȱȱ

ΎΣΕ΅ȱȱ

ΚΓΑϟУȱȱ

ΔΉΏνΎΉ΍ȱ

hopo:s dryn hylotomoi, skhizousi kara phonio:i pelekei as oak.ACC woodmen.NOM tear.ACT.PRES.3PL head.ACC murderous.DAT axe.DAT ‘just as woodmen chop an oak, (Aegisthus) split his head with murderous axe’ (Sophocles, Electra, 99; 5 BC)

ȱ

ȱ

(92)ȱΌΉΕΐ΅ϟΑΝ thermaino: ‘to warm’ a. intransitiveȱ ΉϢΖȱȱ ȱ Θϲȱȱ Έξȱȱ ΔΉΕ΍ΝΗΌξΑȱȱ to de perio:sthen eis the prt propel-round.NOM into ‘(and the air) which is thus propelled Timaeus, 79e, 3; 5-4 BC) b. transitive ЂΈΝΕȱȱ

Θϲȱȱ ΔІΕȱȱ πΐΔϧΔΘΓΑȱȱ

ΌΉΕΐ΅ϟΑΉΘ΅΍

to pyr empipton thermainetai the fire stream.PART.NOM warm.NACT.PRES.3SG round becomes warmed by streaming into the fire’ (Plato,

πΑȱȱ ΛΙΘΕ΍ΈϟУȱȱ ΌΉΕΐ΅ϟΑΉΘΉ

hydo:r en khytridio:i thermainete water.ACC in little-pot warm.ACT.IMP.2PL ‘make warm water in a little pot’ (Aristophanes, Acharnians, 1175; 5-4 BC) (93) ΅ЁΒΣΑΝȱȮȱ΅ЄΒΝȱauksano: - aukso: ‘to increase’ ȱ intransitive ΩΑΌΕΝΔΓΖȱȱ a. ГΗΘΉȱȱ ΦΔνΐ΅ΌΓΑǯǯǯȱȱȱ Ύ΅Ϡȱȱ Έ΍Τȱȱ Θϟȱȱ ho:ste apemathon... kai dia ti anthro:pos so forgot.1SG and for what man.NOM

΅ЁΒΣΑΉΘ΅΍

auksanetai grow.NACT.PRES.3SG

92

Chapter Three ‘I forgot... even about the reason why people are growing old’ (Plato, Phaedo, 96c, 7; 5-4 BC) b.ȱ ΘΓϾΘΓ΍ΖȱΈχȱȱΘΕνΚΉΘ΅ϟȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΉȱΎ΅Ϡȱ΅ЄΒΉΘ΅΍ȱȱȱȱȱȱȱȱȱΐΣΏ΍ΗΘΣȱ·ΉȱΘϲȱȱȱΘϛΖȱȱΜΙΛϛΖȱȱȱȱȱȱȱΔΘνΕΝΐ΅ toutois de: trephetai te kai auksetai malista ge to te:s phsykhe:s ptero:ma these.DAT nourish.NACT.PRES.3SG and grow.NACT.PRES.3SG then the.NOM the soul.GEN wings.NOM

ȱ

‘by these then the wings of the soul are nourished and grow’ (Plato, Phaedrus, 246e, 3; 5-4 BC) transitive ΗϿΑȱȱ ΘϜȱȱπΐϜȱȱ ϹЏΐϙȱȱ c. Ύ΅Ϡȱȱ ΘχΑȱȱ πΐχΑȱȱ ΛЏΕ΅Αȱȱ ΅ЄΒΉ΍Ζȱȱ kai te:n eme:n kho:ran aukseis syn te:i eme:i ȱrho:me:i and the my land.ACC grow.ACT.PRES.2SG with the my ȱ strength ‘and with my own strength you increase my realm’ (Xenophon, Cyropaedia, 5, 5, 33; 5-4 BC)

It is worth noting the very frequent passive use (+presence of agent) of the above verbs in the data that I have examined; the interpretation of the structures as anticausative or passive is difficult in many cases (ambiguity). (b) Parallel presence of non-active and active anticausative type without semantic difference On the other hand, there are instances of parallel non-active and active morphology for anticausatives. In these instances, the anticausative bears active or non-active morphology (without difference in meaning), whereas the causative bears consistently active morphology. I cite indicative instances: (94)ȱΐΉΏ΅ϟΑΝȱmelaino: ‘˜ȱblacken, make black’ȱ intransitive non-active ΐξΑȱ ΔΕΉΗΆΙΘνΕΝΑȱȱ ΅ϡȱȱ ΏΉΙΎ΅Ϡȱȱ ΘΕϟΛΉΖȱȱπΐΉΏ΅ϟΑΓΑΘΓ a.ȱ Ύ΅Ϡȱ ΘЗΑȱȱ kai to:n men presbytero:n hai leukai trikhes emelainonto and the.GEN prt old-men.GEN the.NOM white hair blacken.NACT.IMPERF.3PL ‘the white hair of the old men grew dark’ (Plato, Statesman, 270e, 7; 5-4 BC) intransitive active ΘϛΖȱΗ΅ΕΎϲΖǯǯǯȱȱ ΐΉΏ΅ϟΑΉ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΐξΑȱЀΔϲȱΔ΅Ώ΅΍κΖȱΗΙ·Ύ΅ϾΗΉΝΖȱ b.ȱ Δ΅Ώ΅΍ϱΘ΅ΘΓΑȱϸΑȱȱ palaiotaton on te:s sarkos… melainei men hypo palaias sygkauseo:s oldest part.NOM the flesh.GEN blacken.ACT.PRES.3SG prt by ȱȱȱold ȱ ȱcombustion ‘all the oldest part of the flesh... blackens by the continued combustion’ (Plato, Timaeus, 83a, 7; 5-4 BC) transitive active c. ΘχΑȱȱ Έξȱȱ ΗΣΕΎ΅ȱȱ ΐΉΏ΅ϟΑΉ΍Ъȱ te:n de sarka melainei? the.ACC prt flesh.ACC blacken.ACT.PRES.3SG ‘it blackens the flesh’ (Aristotle, Problems, 966b, 2; 4 BC) (95)ȱΘΉϟΑΝ teino: ‘to stretch by force/pull tight’ intransitive non-active ΘΓІȱȱ ΎΓϟΏΓΙȱȱ ΘΓІȱȱ ΐνΗΓΙȱȱ ΘΉϟΑΉΘ΅΍ȱȱ ΔΣΏ΍Αȱ ΚΏνΜȱȱȱȱȱȱ a. Έ΍Τȱȱ dia tou koilou tou mesou teinetai palin phleps through the hollow the middle stretch.NACT.PRES.3SG again vein.NOM ‘the vein is stretching again through the middle hollow’ (Aristotle, The History of Animals, 513b, 4; 4 BC) intransitive active ·ΤΕȱȱ πΖȱȱ ΘΓІΘΓΑȱǻΑΝΘ΍΅ϧΓΑȱΐΙΉΏϱΑǼȱπΎȱȱ Δ΅ΑΘϲΖȱȱ ΘΓІȱΗЏΐ΅ΘΓΖȱ b. ΘΉϟΑΓΙΗ΍ȱȱ teinousi gar es touton (no:tiaion myelon) ȱȱek pantos tou ȱso:matos stretch.ACT.PRES.3PL prt

in this (southern marrow)

ȱȱfrom everywhere.GEN the body.GEN

‘they are stretching in this southern marrow starting from everywhere in the body’ (Hippocrates, On Generation, 1, 13; 5-4 BC) transitive active ΘЗΑȱȱ Ώϱ·ΝΑȱȱ c. ΑІΑȱȱ ·ΓІΑȱȱ ΗΙΛΑΓϿΖȱȱ ΘΉϟΑΝȱȱ nyn goun sykhnous teino: to:n logo:n now prt long.ACC stretch.ACT.PRES.1SG the.GEN speeches.GEN ‘now, at any rate, I am rather extending my speeches’ (Plato, Gorgias, 519e, 1 ; 5-4 BC)

3.1.4 Summary From the study of texts from the Classical Greek period, we see that the process of causativisation already was systematic in the Classical Greek period and concerned with a large number of verbs denoting change-of-state mainly from an external cause. I distinguish three large categories of causativisation changes (three large types of creation of an innovative/new causative type based on a non-alternating intransitive verb of the previous period, i.e. the Homeric Greek period): (a) active

Diachronic Data and Analysis

93

causative use on the basis of an older active intransitive non-alternating type (the result is an alternating active (‘labile’) verb); (b) active causative use on the basis of an older active and non-active (use of both voices without semantic difference) intransitive non-alternating type (the result is an alternating active causative – active/non-active anticausative verb); and (c) active causative use on the basis of an older non-active intransitive non-alternating type (the result is alternating active causative – non-active anticausative verb). It appears, consequently, that the voice morphology of the anticausative type corresponds to the voice morphology of the initial non-alternating intransitive type. The presence of a cognate object with intransitive verbs can be considered as an intermediary stage, a stage favouring the innovative causative use of intransitive verbs. The combination of intransitive verbs with cognate objects is possible independently of the morphology of the intransitive verb (active or non-active). We cannot argue that derivational prefixes or suffixes favoured the causativisations in the same way since not only instances of causativisation of intransitives with the use of a prefix or suffix (the majority of causative verbs are formed from nouns or adjectives with the addition of suffixes -ȩ-(Ȧ) -Ƞ-(Ƞ:) / -ĮȓȞ(Ȧ) -ain-(o:) / -ȪȞ-(Ȧ) -un-(o:)/ -ȓȗ-(Ȧ) -iz-(o:)) are observed but also instances of the use of derivational prefixes and suffixes with a transitive base, resulting in an intransitive type. The finding that the anticausativisation process (new anticausative uses for older non-alternating causative types) is not systematic and productive (as the causativisation process is) is of particular importance for the analysis. The causative non-alternating verbs have a stable syntactic behaviour both in relation to the structures in which they participate (only transitives) and in relation to the accusative case, which they exclusively and consistently require. In contrast, the accusative verbs (transitive noncausative verbs) exhibit instability regarding argument structure (intransitive emerges from older accusative verbs and vice versa) and their case requirements. The accusative verbs of Classical Greek do not require exclusively a DP in the accusative case, as do the causatives of Classical Greek, but alternations between the accusative and other cases and/or PPs are in evidence. Concerning the arguments of the structures that are examined here, the presence of an agent, cause or instrument in the subject position is possible with causative structures, whereas with anticausative structures, the presence of an agent is excluded in Classical Greek. An agent, cause or instrument can be denoted with a by-Phrase in passive structures. In Classical Greek, the following arguments can be in the subject position (bearing nominative case) of non-active types: (a) benefactive, (b) theme/patient, and (c) goal (in the genitive, dative, or second accusative case) with ditransitive verbs (whereas the theme/patient argument remains in the accusative case since the non-active type was able to assign the accusative case to the theme/patient argument). The observations on the voice morphology can be summarised as follows: generally, in the system of voice of Ancient Greek, the non-active is the marked voice, denoting personal interest with transitive types (benefactive meaning), whereas in the future and past tenses, it is distinguished in two morphological forms (non-active I and non-active II) without, however, any difference in the structures where these forms are used (both forms are used alike in the same structures). The structure “non-active verb + object in the accusative” cannot be ruled out, as it appears not only from the systematic presence of transitive structures with non-active voice (with benefactive meaning) but also from the passivisation of ditransitive verbs where the DP in the accusative does not change case (resulting in passive structures with non-active verbs and the second object in the accusative case). In Homeric Greek, the relationship between causative-anticausative was denoted with an ablaut alternation, whereas the non-active voice was not essential for the passive structure (where an agent is present). In Classical Greek, the ablaut alternation has been lost, and the relationship causativeanticausative continues to be marked by the opposition of active and non-active voice, whereas the passive type is marked with active voice but has also begun to be marked very productively with the non-active voice. The voice distribution in causative and anticausative types is very systematic. Causatives are consistently active (the use of non-active endings to denote personal interest (benefactive meaning) with causative types is not possible), whereas the anticausatives are expressed with non-active voice. There are very few instances of anticausatives with parallel use (without difference in meaning) of both non-active and active voice. These instances strengthen the different way of dealing with the anticausative voice morphology from the way of dealing with the voice morphology of the other intransitive structures. The alternation of the active and non-active intransitive type reflects the corresponding morphological alternation of the intransitive type before it was causativised (when it was non-alternating intransitive) or occurs as the result of an ongoing morphological change towards the expression of the anticausative type with active voice; that is, an active form appears for those few verbs (and is used in free distribution with the older non-active form), to prevail in the next period. The basic conclusion, however, is that the presence of the non-active form is stable in the anticausative structures in Classical Greek, either as the only voice morphology in the majority of

94

Chapter Three

instances or in free distribution with an active form in very few instances.

3.2 The Hellenistic-Roman period 3.2.1 The general picture The chronological boundaries of the Hellenistic period are unclear (Moser 2005: 227-229). The military campaigns of Alexander the Great (336-323 BC) are considered the starting point of this period, which saw as an aftermath the generalisation of the use of the Greek language of that era as Koine. During this period, Greek became a second language for the people of the regions conquered by Alexander the Great (Asia Minor, Egypt, Syria, Lebanon, Mesopotamia, and Persia). There is more disagreement, however, concerning the end point of this period; for the majority of studies, the final date is the end of the 5th century AD (Browning 1983/1991; Babiniotis 1985/2002; Horrocks 1997; Kopidakis 1999; Eideneier 1999; Adrados 1999), whereas Tonnet (1993) places the end of the Hellenistic period in the 7th century. Hellenistic Koine is believed to have evolved from the Attic dialect of the Classical period (Chatzidakis 1975/1892; Robertson 1934). The influence of the remaining dialects on the Attic dialect during its development into Hellenistic Koine has not made any change to the dominance of the Attic dialect27. Hellenistic Koine slowly dominated oral language, but written language was divided since the writers who addressed their work to a wide public wrote in Hellenistic Koine (lyric poetry, laudations). Certain distinguished historiographers, orators and philosȠphers, however, sought a return to the Attic Classical Greek dialect (the Atticist movement), and some genres of poetry were already written in traditional literary classical dialects as well. Hellenistic Greek became, however, the language of sacred Christian texts: the 39 books of the Old Testament were translated from Judaic and Aramaic into Greek in Alexandria (translation of the Septuagint) to meet the needs of the Greek-speaking Judeans of the region, and the 27 books of the New Testament were written directly in Hellenistic Koine (cf. Beck 1971/1988; Browning 1983/1991; Karali 1990, 1992; Glykatzi-Ahrweiler 1999). In the Hellenistic and Roman period (3rd BC - 4th AD), significant changes are observed at all levels of the linguistic system, i.e. phonological, morphological, semantic, and syntactic (Charalambakis 1981, 1993, 1999; Papanastasiou 2007). An examination of all of these changes would be excessive and outside the objectives of this study (for details of the changes, cf. Horrocks 1997 and Christidis 2007, among others). The majority of changes are concerned with the phonological domain: the tone of Classical Greek was replaced by dynamic stress resulting in loss of prosody; diphthongs were monophthongised (/ei/ ĺ /i/, /ai/ ĺ /e/) and double consonants were pronounced as one consonant. In morphology, the tendency that prevails is simplification, mainly through analogy: irregular comparatives are replaced, the variation in the formation of verb types is limited, and dual number is lost. In syntax, the dative case (and in many instances the genitive case also) is replaced by PPs, the accusative replaces the genitive and the dative case, and the optative mood gradually gives its position to structures with a verb and a complement clause (tendency towards analytic structures).

3.2.2 Innovative causative use of intransitive verbs: causativisation of intransitives vs. case changes in accusative verbs 3.2.2.1 Causativisations During this period in the area of transitivity alternations new causative uses appear for exclusively intransitive verbs of Classical Greek. The original intransitive verbs bear active or non-active morphology and retain the voice morphology that they bore as intransitive after their causativisation as well. In their innovative causative use, they always bear active morphology (in other words, similarly to anticausatives in the case of original active intransitives (‘labile forms’) and dissimilarly to anticausatives in the case of original non-active intransitives). Consequently, the presence of the external argument (agent/cause) and the active morphology of transitive structures are more closely connected than the presence of the direct object and active morphology. It appears that active voice is not related to the presence of the accusative case but to the presence of the agent as subject (in the nominative case) of the structure. Similarly, we have seen (section 3.1.3) voice morphology being connected with the external argument in cases of middle transitive verbs (with the meaning of personal interest of the external argument-subject/benefactive meaning). The causativisations of active 27

There is no unanimous agreement concerning the origin of the Hellenistic Koine dialect. Some grammarians and linguists have argued that it arose from the four basic dialects, others that it was itself a fifth dialect.

Diachronic Data and Analysis

95

intransitive verbs form a passage for the creation of transitivity alternations without any change of voice (same voice for both the causative and the anticausative type, i.e. ‘labile forms’), i.e. the kind of transitivity alternations that we meet in Modern Greek (causative and anticausative members of the alternation with active morphology) but not in classical Ancient Greek (active causative vs. non-active anticausative). I distinguish the following three cases of causativisation: (a) Causative types from intransitive active Already in the Ptolemaic era (cf. Mayser 1934/1970), the first examples of innovative causative types are apparent. In the following examples, it appears that the structures with an infinitive and a DPsubject in the accusative case (clitic or not) could be reanalysed as transitive structures. (1)ȱ ΦΏΓ·νΝȱalogeo: ‘to be-unreasonable’ ȱ ΆΓΙΏϱΐΉΑϱΖȱȱ ΐΉȱȱ ΦΏΓ·ϛΗ΅΍  ȱ boulomenos me aloge:sai wish.PART.NOM me.ACC be-unreasonable/crazy.ACT.AOR.INF ‘wishing to drive me crazy’ (ȉebtunis Papyri I, 138; UPZ I (Urkunden der Ptolemäerzeit (ältere Funde), 648, 18; 3-2 BC) (2)ȱ Έ΍΅ΘΕϟΆΝdiatribo: ‘to delay’  ΉϢΖȱȱ Θϲȱȱ ΐχȱȱ

Έ΍΅ΘΕϧΜ΅΍ȱȱ

΅ЁΘϱΑȱȱȱ

eis to me: diatripsai auton to the not delay.ACT.AOR.INF he.ACC ‘for not to delay him’ (BGU (Aegyptische Urkunden aus den Staatlichen Museen zu Berlin, Griechische Urkunde) 1208, 21, 28a)

In the examples in (3), the patient arguments (which are in the subject position in the corresponding intransitive structures) are in the accusative case, whereas the DPs in the genitive denote the cause of the emotion expressed by the verbs. (3) 

ΉЁΈΓΎνΝ eudokeo: ‘to please’ a. ΋ЁΈϱΎ΋Η΅Ζȱȱ

ΐΉȱȱ ΘϛΖȱȱ Θ΍ΐϛΖ e:udoke:sas me te:s time:s please.ACT.AOR.2SG I.ACC the.GEN honour.GEN ‘you pleased me with an honour’ (Lond. (Greek Papyri in the British Museum. London) I, 3 (p. 46))  b. ΦΔ΋ΙΈϱΎ΋ΗΣΖȱȱ ΐΉȱȱ ΘϛΖȱȱ Θ΍ΐϛΖȱȱ    ape:udoke:sas me te:s time:s    disregard.ACT.AOR.2SG I.ACC the.GEN honour.GEN ‘you did not please me with an honour’ (Leid. (Papyri Graeci Musei Antiquarii Lugduni-Batavi) P. 3) 

Both in the translation of the Septuagint (Old Testament) and in the New Testament, in many instances, causative interpretation is given to verbs that were solely intransitive in Classical Greek (cf. also Moulton & Turner 1963): Translation of the Septuagint; 2-1 BC a. Ά΅Η΍ΏΉϾΝ basileuo: ‘to cause someone to rule’ b. πΒ΅ΐ΅ΕΘΣΑΝ eksamartano: ‘to cause someone to make a mistake’ New Testament; AD 1  a. ΦΑ΅ΘνΏΏΝ anatello: ‘to make someone stand up’  b. ΦΑ΅Κ΅ϟΑΝȱanaphaino: ‘to make someone appear’  c. ΐ΅Ό΋ΘΉϾΝ mathe:teuo: ‘to make someone a pupil elsewhere’  d. Ύ΅Θ΅ΎΏ΋ΕΓΑΓΐЗ katakle:ronomo: ‘to cause someone to inherit.ACT’  e. ΌΕ΍΅ΐΆΉϾΝ thriambeuo: ‘to lead to victory’ f. ΉϢΕ΋ΑΉϾΝ eire:neuo: ‘to cause the combatants to make peace’

The following transitivity alternations (intransitive type in active voice / transitive type in active voice) are innovative structures of the Hellenistic-Roman period (in Classical Greek, the verbs were intransitive non-alternating). (4) Φ·ΝΑ΍Зȱago:nio:ȱ‘to be distressed’ intransitive active ·ΤΕȱȱ a.ȱ Φ·ΝΑ΍Зȱȱ ago:nio: gar be-distressed.ACT.PRES.1SG because

ȱ Ύ΅ЁΘϱΖ

kautos and-myself.NOM

96

Chapter Three ‘as I am myself distressed’ (Menander, Misumenus, 272; 4-3 BC) transitive active ΘΤΖȱȱ ΔΉΊ΍ΎΤΖȱ ΈΙΑΣΐΉ΍Ζ b.ȱ Φ·ΝΑ΍ЗΗ΅΍ȱȱ ago:nio:sai tas pezikas dynameis be-distressed.ACT.PRES.PART.NOM the.ACC land forces.ACC ‘as they made the land forces distressed’ (Polybius, Histories, 1, 20, 6; 3-2 BC) i in the previous period - Classical Greek: exclusively intransitive active c.ȱ ̍΅ϟȱΐΓ΍ȱπΈϱΎΉ΍ȱȱȱȱȱȱȱȱȱȱϳȱȱȱ̓ΕΝΘ΅·ϱΕ΅Ζȱȱ όΈ΋ȱΘΉΘΕ΅ΛϾΑΌ΅΍ȱȱȱȱȱȱΘΉȱΎ΅ϠȱΦ·ΝΑ΍κΑȱ kai moi edokei ho Pro:tagoras e:de: tetrakhynthai te kai ago:nian and I.DAT seemed.3SG the Protagoras.NOM already provoke.NACT.INF and be-distressed.ACT.PRES.INF ȱ

‘Protagoras seemed to me to be in a thoroughly provoked and distressed state’ (Plato, Protagoras, 333e, 3; 5-4 BC) (5)ȱ πΒ΅ΐ΅ΕΘΣΑΝȱeksamartano: ‘to fail of one's purpose/go wrong’ intransitive active a.ȱ ЀΔΓΐΑ΋ΗΘνΓΑȱȱ Έξȱȱ Ύ΅Ϡȱȱ Δ΅ΘνΕΝΑȱȱ Φ·΅ΌЗΑȱ πΑϟΓΙΖǰȱȱ ϵΘ΅Αȱȱ πΒ΅ΐ΅ΕΘΣΑΝΗ΍Α

hypomne:steon de kai patero:n agatho:n enious, hotan eksamartano:sin remind

prt and

fathers.GEN good.GEN few.ACC when go-wrong.ACT.PRES.3PL

‘one must also remind the fathers of some good men, when they are doing wrong’ (Plutarch, Precepts of Statecraft, 810b, 2; AD 1-2)

transitive active b. Ύ΅Ϡȱȱ πΒφΐ΅ΕΘΉΖȱȱ kai ekse:martes and

tou parorgisai

ΐΉȱ meȱ

the.GEN provoke-to-anger.INF

I.ACC

ΘϲΑȱΏ΅ϱΑȱΐΓΙȱȱΘϲΑȱ͑ΗΕ΅ϛΏȱȱΘΓІȱȱΔ΅ΕΓΕ·ϟΗ΅΍ȱȱ

ton laon mou ton Israe:l

go-wrong.ACT.AOR.2SG the people.ACC my the Israel

‘and you made my people, the Israel, do wrong and provoke me to anger’ (Septuagint, Kings III, 16, 2, 3; 2-1 BC) c.ȱ Θ΅ІΘȇȱȱ ΦΎΕ΅Ηϟθȱȱ Ύ΅Θ΅Ά΍΅ΊϱΐΉΑΓ΍ȱȱ ΘϲΑȱΏΓ·΍ΗΐϲΑȱȱ πΒ΅ΐ΅ΕΘΣΑΓΙΗ΍Αȱ taut’ akrasiai katabiazomenoi ton logismon eksamartanousin the.NOM bad-mixture.DAT constrain.NACT.PART.NOM the reasoning-power.ACC go-wrong.ACT.PRES.3PL

‘these make reasoning power go awry as they are constrained by the bad mix’ (Plutarch, Table Talk, 705c, 3; AD 1-2)ȱ i in the previous period - Classical Greek: intransitive active28 ΦΏΏȇȱȱ Φΐ΅Όϟθȱȱ ΘϜȱȱ πΐϜȱȱ ȱȱȱȱȱȱȱȱ d.ȱ ϵΘ΍ȱȱ ΓЁΛȱȱ οΎАΑȱȱ πΒ΅ΐ΅ΕΘΣΑΝȱȱ hoti oukh heko:n eksamartano: all’ amathiai te:i eme:i that not purposely go-wrong.ACT.PRES.1SG but ignorance.DAT the.DAT my.DAT ‘that I do not do wrong on purpose (it is not an intentional error), but due to my ignorance’ (Plato, Gorgias, 488a, 3; 5-4 BC)

(b) Causative types from intransitive non-active At the same time, new active causative uses for earlier non-active anticausative verbs appear. In this case, the innovative causative type is accompanied by a change in morphology, which means that an active causative type becomes possible in parallel for old non-active anticausative verbs. For example: intransitive ûnjȂèDŽǬ he:domai ‘to enjoy oneself/take one's-pleasureĺ innovative transitive use ûnjȁ he:do:  intransitive ÷ȖȖŲèDŽǬ he:tto:mai ‘to be less than another/inferior to’ ĺ innovative transitive use ÷ȖȖĝȁ he:ttao:  intransitive èDŽģȀȂèDŽǬ mainomai ‘to rage/be furious’ĺ innovative transitive use ñǰ èDŽģȀȁ (ek)maino: intransitive ñȀȖȋğäȂèDŽǬ entrepomai ‘to hesitate/feel misgiving/feel shame’ĺ innovative transitive use ñȀȖȋğäȁ entrepo:




(6) ϊΈΝ he:do: ‘to please/delight/enjoy’ intransitive non-active ϳȱȱ ΐνΌΙΗΓΖȱȱ πΐΚΓΕΓϾΐΉΑΓΖȱȱ ΓϥΑΓΙȱ a.ȱ ΓϩΓΑȱȱ·ΓІΑȱϊΈΉΘ΅΍ȱȱ hoion goun he:detai ho methysos emphoroumenos oinou so then enjoy.NACT.PRES.3SG the drunk.NOM fill.NACT.PART.NOM wine.GEN ‘so then the drunk enjoys as he is filled with wine’ (Sextus Empiricus, The Outlines of Pyrrhonism, 3, 195, 5; AD 2-3)

transitive active b.ȱ БΖȱȱȱΘϲȱΗ΍ΘϟΓΑȱȱ ȱύȱȱ Θϲȱȱ ΔΓΘϲΑȱȱ ho:s to sition e: to poton as

ϊΈΉ΍ȱΐξΑȱȱ

ΘϲΑȱȱ Κ΅·ϱΑΘ΅ȱȱ

ύȱȱ Δ΍ϱΑΘ΅ȱȱȱȱ

he:dei men

ton

e: pionta

the food.NOM or the drink.NOM enjoy.ACT.PRES.3SG the

phagonta

eat.PART.ACC or drink.PART.ACC

‘as the food or the drink pleases the person who is eating or drinking’ (Sextus Empiricus, Against the

28

Rarely transitive but with cognate object (in Herodotus and in Sophocles).

Diachronic Data and Analysis

97

Mathematicians, 7, 368, 3; AD 2-3)

i in the previous period - Classical Greek:
exclusively intransitive non-active c.ȱ ΩΏΏΓΖȱȱΘ΍Ζȱȱ ȱȱȱȱύȱϣΔΔУȱȱ Φ·΅ΌХȱȱ ύȱΎΙΑϠȱȱȱȱ ύȱϷΕΑ΍Ό΍ȱȱȱȱ ϊΈΉΘ΅΍ǰȱȱ allos tis e: hippo:i agatho:i e: kyni e: ornithi he:detai, other

somebody.NOM or horse.DAT good

ΓЂΘΝȱȱ

houto:

or dog.DAT or bird.DAT enjoy.NACT.PRES.3SG so ȱ

Ύ΅Ϡȱȱ σΘ΍ȱȱ ΐκΏΏΓΑȱȱ ϊΈΓΐ΅΍ȱȱ

ΚϟΏΓ΍Ζȱȱ

Φ·΅ΌΓϧΖ

kai eti mallon he:domai philois agathois and even more enjoy.NACT.PRES.1SG friends.DAT good.DAT ‘others have a fancy for a good horse or dog or bird: my fancy, stronger even than theirs, is for good friends’ (Xenophon, Memorabilia, 1, 6, 14; 5-4 BC) (7) ΏΙ·ϟΊΝ lygizo: ‘to bend/twist as one does a withe’ intransitive non-active ӥΕΝΘΓΖȱȱ ЀΔȇȱȱ ΦΕ·΅ΏνΝȱȱ πΏΙ·ϟΛΌ΋ΖЪ a.ȱ ΓЁΎȱȱ ΅ЁΘϲΖȱȱ ouk autos Ero:tos hyp’ argaleo: elygikhthe:s? not yourself.NOM Love.GEN by troublesome.GEN bend.PASS.AOR.PERF.2SG ‘you did not bend by the troublesome Love?’ (Theocritus, Idylls, 98; 4-3 BC) transitive active Θ΍ȱȱ ΘЗΑȱȱ ΐΉΏЗΑȱȱ b.ȱ ΐ΋Έξȱȱ ΐχΑȱȱ ΦΑ΅·Ύ΅ΊΓΐνΑΓΙΖȱ ΏΙ·ϟΊΉ΍Αȱȱ me:de me:n anagkazomenous lygizein ti to:n melo:n neither not force.PART.ACC bend.ACT.PRES.INF one.ACC the body-parts.GEN ‘neither were they forced to bend one from their parts’ (Aelian, On the characteristics of animals, 2, 11, 46; AD 2-3) i in the previous period - Classical Greek:
exclusively intransitive non-active c.ȱ ΔΓΈϱΖȱȱ Α΍Αȱȱ ΩΕΌΕΓΑȱȱ ϗȱȱ ΏΙ·ϟΊΉΘ΅΍ȱȱ podos nin arthron he:i lygizetai foot.GEN him ankle.NOM which.DAT bend.NACT.PRES.3SG ‘(he caught him) by the foot where the ankle bends’ (Sophocles, Trachiniae, 779; 5 BC)

(c) Causative types from intransitive active / non-active In the case of the verb ȜİȣțĮȓȞȦ leukaino: ‘to whiten’, we have the causativisation of an intransitive active / non-active verb. The new causative type bears active morphology, whereas the anticausative type bears active (in some sentences) and non-active morphology (in other sentences) without any semantic difference (exactly like the original intransitive type). (8) ΏΉΙΎ΅ϟΑΝ leukaino: ‘to whiten’ intransitive non-active ·ΤΕȱȱ ΔΓΏΏΤȱȱ ΐνΏ΅Α΅ǰȱȱ Ψȱȱ ΏΉΙΎ΅ϟΑΓΑΘ΅΍ȱ a. ΉϢΗϠȱȱ eisi gar polla melana, ha leukainontai be.3PL prt many.NOM black.NOM which.NOM whiten.NACT.PRES.3PL ‘as there are many blacks which (can) whiten’ (Ammonius, in Porphyrii isagogen sive quinque voces, 114, 3; AD 5) intransitive active ΅ϡȱȱ ΘΕϟΛ΅Ζȱȱ ΘϛΖȱȱ ΎΉΚ΅ΏϛΖȱȱ ΅ЁΘΓІ b. ΏΉΙΎ΅ϟΑΓΙΑȱȱ leukainoun hai trikhas te:s kephale:s autou ȱȱȱȱȱȱȱȱ whiten.ACT.PRES.3PL the.NOM hair the.GEN head.GEN his ‘the hair of his head is getting white’ (Apocrypha, Revelation, versio tertia, 322, 4; AD 2) transitive active ΐξΑȱȱ ΘΤȱȱ πΗΌφΐ΅Θ΅ȱȱ c. ΏΉΙΎ΅ϟΑΉ΍ȱȱ leukainei men ta esthe:mata whiten.ACT.PRES.3SG prt the.ACC clothes.ACC ȱ ‘that whitens the clothes’ (Sextus Empiricus, Against the Mathematicians, 9, 247; AD 2-3) i in the previous period - Classical Greek
intransitive active Έȇȱȱ πΑΘ΅ІΌ΅ȱȱ ΉϢΎϱΘΝΖȱȱ ΘΓІȱȱ ΅ϣΐ΅ΘΓΖȱȱ d. ΦΌΕΓ΍ΊΓΐνΑΓΙȱȱ athroizomenou d’ entautha eikoto:s tou haimatos collect.NACT.PART.GEN prt here similarly the.GEN blood.GEN ȱ

ΘΓІΘΓȱȱ

ΐξΑȱȱ πΕΙΌΕ΅ϟΑΉ΍ǰȱȱ

Ψȱȱ

Έξȱȱ ΦΔΓΏΉϟΔΉ΍ǰȱ

ΏΉΙΎ΅ϟΑΉ΍ȱȱ



touto 

men  erythrainei,

ha

de

leukainei 

this.NOM prt 

redden.ACT.PRES.3SG which.NOM prt

apoleipei,

lose.ACT.PRES.3SG whiten.ACT.PRES.3SG

‘as the blood seems to collect here, this is getting red; when it is losing blood, it is getting whiter’ (Aristotle, Problems, 890a, 9; 4 BC) intransitive non-active

98

Chapter Three e. ЀΔϲȱȱ Έξȱȱ ΔΙΕϲΖȱȱ Ύ΅Ϡȱȱ ΛΕϱΑΓΙȱȱ Δ΅ΛϾΑΉΘ΅΍ȱȱ Ύ΅Ϡȱȱ ΏΉΙΎ΅ϟΑΉΘ΅΍ȱȱȱȱ  hypo de pyros kai khronou pakhynetai kai leukainetai by prt fire and time thicken.3SG and whiten.NACT.PRES.3SG ‘it is getting thicker and whiter from fire and time’ (Aristotle, Meteorology, 383b, 29; 4 BC) 

The causativisations of intransitive verbs are subject to restrictions; in the Hellenistic and Roman period, causative types were not formed for all intransitive verbs. Also in the Hellenistic-Roman period, verbs denoting change-of-state from an internal cause remain intransitive non-alternating verbs. The verb ȤĮȓȡȦ khairo: ‘to rejoice/be glad’ remains intransitive non-alternating: (9)

Λ΅ϟΕΝȱkhairo: ‘to rejoice/be glad’ Λ΅ϟΕΉ΍Ζȱȱ Θ΍Αϟȱ a. πΑȱȱ ΅ЁΘΓϧΖȱȱ ΓϩΖȱȱ en autois hois khaireis tini in these.DAT which.DAT rejoice.ACT.PRES.2SG something.DAT ‘at the times when you are delighted with a thing’ (Epictetus, Discourses, 3, 24, 88, 1; AD 1-2) Ύ΅ϠȱȱΦ·΅ΏΏ΍κΗΌΉǰȱȱȱȱȱȱȱȱȱϵΘ΍ȱȱϳȱȱΐ΍ΗΌϲΖȱȱȱȱЀΐЗΑȱȱȱΔΓΏϿΖȱȱπΑȱΘΓϧΖȱΓЁΕ΅ΑΓϧΖ b.ȱ Λ΅ϟΕΉΘΉȱȱ khairete kai agalliasthe, hoti ho misthos hymo:n polys en tois ouranois rejoice.ACT.IMP.2PL and be-glad.NACT.IMP.2PL that the reward.NOM you.GEN much in the skies

i

ȱ

‘rejoice, and be glad, for great is your reward in heaven’ (New Testament, Matthew, 5, 12, 1; AD 1)

similar in the previous period (Classical Greek): Λ΅ϟΕΉ΍ȱȱ ρΏΎΝΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΘϲȱπΔϟΗΘ΅ΗΌ΅΍ȱȱȱΎ΅Ϡȱȱΐ΅ΑΌΣΑΉ΍Αȱ c.ȱ ϵΔϙȱȱȱȱȱΘ΍Ζȱ hope:i tis khairei helko:n to epistasthai kai manthanein if someone.NOM rejoice.ACT.PRES.3SG drag.PART.NOM the.ACC know.INF and learn.INF ‘if anyone is delighted to drag the expressions “know” and “learn” (one way and another)’ (Plato, Theaetetus, 199a, 5; 5-4 BC) ΘЗΑȱȱ ΅ЁΘЗΑȱȱ ·΍·ΑΓΐνΑΝΑȱȱ ΘΉȱΎ΅Ϡȱȱ ΦΔΓΏΏΙΐνΑΝΑȱ d.ȱ ΔΣΑΘΉΖȱȱ Γϡȱȱ ΔΓΏϧΘ΅΍ȱȱ pantes hoi politai to:n auto:n gignomeno:n te kai apollymeno:n all the citizens.NOM the.GEN same.GEN births.GEN and deaths.GENȱ Ύ΅Ϡȱȱ ΏΙΔЗΑΘ΅΍Ъȱȱ ȱ ȱ ȱ Δ΅Ε΅ΔΏ΋ΗϟΝΖȱȱ Λ΅ϟΕΝΗ΍ȱȱ paraple:sio:s khairo:si kai lypo:ntai?ȱȱȱȱȱȱȱȱȱ ȱ alike rejoice.ACT.PRES.3PL and grieve.ACT.PRES.3PL ‘all the citizens rejoice and grieve alike at the same births and deaths?’ (Plato, Republic, 462b, 6; 5-4 BC)

The verb éȑȘǗȀŲ
astheno: ‘to be weak/feeble/sick’ remains intransitive non-alternating, and of particular interest is the case of the verb ȜȐȝʌȦ lambo: ‘to shine’, which was causativised in the Classical Greek period but is intransitive non-alternating in the Hellenistic-Roman period. (10) ΦΗΌΉΑЗ astheno: ‘to be weak/feeble/sickly’ a.ȱ БΖȱȱ ΓЇΑȱȱ όΎΓΙΗΉΑȱϵΘ΍ȱΦΗΌΉΑΉϧǰȱȱ ho:s oun e:kousen hoti asthenei,

ȱΘϱΘΉȱΐξΑȱȱσΐΉ΍ΑΉΑȱȱπΑȱСȱȱȱȱώΑȱΘϱΔУȱȱΈϾΓȱψΐνΕ΅Ζ

tote men emeinen en ho:i e:n topo:i dyo he:meras

when therefore heard.3SG that be-sick.ACT.PRES.3SG then prt stayed.3SG in which was.3SG place two days

‘when therefore he heard that he was sick, he stayed two days in the place where he was’ (New Testament, John, 11, 6, 2; AD 1) ·ΤΕȱȱ ΦΗΌΉΑЗ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ b.ȱ ΐΉΘΤȱΘϲΑȱȱ ΌΉΕ΍ΗΐϲǽΑȱπΕ·ΓǾȱ ǽΏǾ΅ΆφΗΓΐ΅ǽ΍Ǿǰȱȱ ΩΕΘ΍ȱȱ meta ton therismo[n ergo] [l]abe:soma[i], arti gar astheno:i after the harvest work.ACC take.MID.FUT.1SG immediately prt be-sick.ACT.PRES.1SG ‘I will take over the work after the harvest; as I am sick now’ (BGU (Aegyptische Urkunden aus den Staatlichen Museen zu Berlin, Griechische Urkunde) 2, 594 v 6) c.ȱ ϵΘ΍ȱȱ ωΗΌνΑ΋Ύ΅ȱȱ ΔΣΏ΅΍ hoti e:sthene:ka palai that be-sick.ACT.PERF.1SG long-ago ‘that I became sick long ago’ (Tebtunis Papyri 2, 414 r 9) (11) ΏΣΐΔΝȱlambo: ‘to shine’ a. д̈Ύȱȱ ΗΎϱΘΓΙΖȱΚЗΖȱȱ ek skotous pho:s

ΏΣΐΜΉ΍ǰȱȱ

ȱ϶ΖȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱσΏ΅ΐΜΉΑȱȱȱȱȱȱȱȱȱȱȱȱȱπΑȱΘ΅ϧΖȱΎ΅ΕΈϟ΅΍ΖȱψΐЗΑȱ

lampsei,

hos

elampsen

en tais kardiais he:mo:n

from darkness light.NOM shine.ACT.FUT.3SG which.NOM shine.ACT.AOR.3SG

in the hearts our

‘light will shine out of darkness, which has shone in our hearts’ (New Testament, 2 Corinthians, 4, 6, 2; AD 1) Θϲȱȱ ΔΕϱΗΝΔΓΑȱȱ ΅ЁΘΓІȱȱ БΖȱȱ b.ȱ Ύ΅Ϡȱȱ σΏ΅ΐΜΉΑȱȱ kai elampsen to proso:pon autou ho:s and shine.ACT.AOR.3SG the.NOM face.NOM his as

ϳȱȱ

ϊΏ΍ΓΖ

ho the

e:lios sun

Diachronic Data and Analysis

99

‘and his face shone like the sun’ (New Testament, Matthew, 17, 2, 2; AD 1)29

3.2.2.2 Factors and stages of causativisations During the Hellenistic period, as in the Ancient Greek period, the fact that intransitive verbs (mainly but not only psych-verbs) are connected with cognate objects (thereby gaining a kind of transitive nature) can be considered an intermediary stage for the transitivisation of intransitive verbs. I assume that this phenomenon strengthens the tendency towards the transitivisation of certain intransitive verbs. Usually, the cognate object is accompanied by an adjective or by a relative pronoun (Andriotis 1937): (12) a. πΚΓΆφΌ΋Η΅Αȱȱ ΚϱΆΓΑȱȱ ΐν·΅Α ephobe:the:san phobon megan ȱ ȱ fear.PASS.AOR.3PL fear.ACC great.ACC ‘they were greatly afraid (they feared a great fear)’ (New Testament, Mark, 4, 41, 1; AD 1) Λ΅ΕΤΑȱȱ ΐΉ·ΣΏ΋Αȱ ȱ b. πΛΣΕ΋Η΅Αȱȱ ekhare:san kharan megale:n rejoice.NACT.ǹȅȇ.3PL joy.ACC great.ACC ‘they rejoiced with great joy’ (New Testament, Matthew, 2, 10, 1; AD 1) c. πΎ΅Ιΐ΅ΘϟΗΌ΋Η΅Αǯǯǯȱȱ Ύ΅Іΐ΅ ekaumatisthe:san... kauma scorch.PASS.AOR.3PL heat.ACC ‘(people) were scorched with heat’ (New Testament, Revelation, 16, 9; AD 1) (13) a. Θϲȱȱ ΆΣΔΘ΍Ηΐ΅ȱȱ ϶ȱȱ π·Аȱȱ Ά΅ΔΘϟΊΓΐ΅΍ȱ to baptisma ho ego: baptizomai the.ACC baptism.ACC which.ACC I.NOM baptize.NACT.PRES.1SG ‘the baptism that I am baptized with’ (New Testament, Mark, 10, 38, 3; AD 1) b. ψȱȱ Φ·ΣΔ΋ȱȱ ϋΑȱȱ ω·ΣΔ΋ΗΣΖȱȱ ΐΉȱ ȱ e: agape: he:n e:gape:sas me the.NOM love.NOM which.ACC love.ACT.AOR.2SG I.ACC ‘the love with which you loved me’ (New Testament, John, 17, 26, 3; AD 1)

As also appears from the examples, the verbs that take a cognate object can be of active or non-active morphology. The derivation of compound transitive verbs seems to be also in the same direction. In postClassical Greek, compound transitive verbs appear with a noun as their first compound element (ex. 14: ȖȘ ge: ‘earth/land’) and with a direct object in the accusative case (as in Classical Greek). (14) a. ΘχΑȱȱ ·ϛΑȱȱ ·ΉΝΕ·ΓІΐΉΑ te:n ge:n geo:rgoumen ȱȱȱȱȱ ȱ the.ACC land.ACC cultivate.ACT.PRES.1PL ‘we cultivate the land’ (Zenon Papyri, Catalogue général des antiquités égyptiennes du Musée du Caire 59130, 2, 256a) ·ΉΝΕ·З b.ȱ ΦΐΔΉΏЏΑ΅Ζȱȱ ύȱȱ Δ΅Ε΅ΈΉϟΗΓΙΖȱȱ ampelo:nas e: paradeisous geo:rgo: ȱȱȱ vineyards.ACC or paradises.ACC cultivate.ACT.PRES.1SG ‘I cultivate vineyards or paradises’ (Revenue Laws of Ptolemy Philadelphus, 36, 15)

Prefixes induce a change in the nature of verb transitivity during this period (as in the Ancient Greek period). The addition of a prefix to a verb could, however, result in either a transitive or intransitive type. Consequently, prefixes cannot be considered indicators of transitivisation or intransitivisation in Hellenistic-Roman Koine since they lead to the production of either transitive or intransitive verbs and not consistently to production of a specific type (of specific (in)transitivity) of verb. Voice, again, does not play a role in cases of transitivisation (active or non-active verb may comprise the base for the addition of a prefix). In the case of a causative type with a prefix, however, again only an active verb can comprise the base of the derivation (causative type: prefix + active form). (a) Change into transitive (accusative or causative) verb Ύ΅Θ΅ȱȬΆ΅ϟΑΝ baino: ‘to attain’ ȱ ȱ ȱ ȬΆΕ΅ΆΉϾΝ -brabeuo: ‘to give judgement against/rule over’

29

For more examples, cf. Lavidas 2006a.

100

Chapter Three

ȱȱȱȱȱȱȱȱȱ ȬΔΓΑЗ -pono: ‘to subdue’ ȱȱ ȬΗΓΚϟΊΓΐ΅΍ -sophizomai ‘to outwit by sophisms or fallacies’ Έ΍΅ȱȱȬΆ΅ϟΑΝȱdia -baino: ‘to pass over’ ȱȱ ȬνΕΛΓΐ΅΍ -erkhomai ‘to go through/pass through’ ȱȱ ȬΔΏνΝȱ-pleo: ‘to sail through a strait or gap/sail across’ ȱȱ ȬΔΓΕΉϾΓΐ΅΍ȱ-poreuomai ‘to pass across/go through’ Δ΅Ε΅ȱȬνΕΛΓΐ΅΍para -erkhomai ‘to go by, beside, or past/pass by’ (New Testament, Mark, 6, 48; AD 1) ΔΉΕ΍ȱȬνΕΛΓΐ΅΍ peri -erkhomai ‘to go round/go about’ ȱȱ ȬϟΗΘ΋ΐ΍ -iste:mi ‘to place round’ ȱȱ ȬΘΕνΛΝ-trekho: ‘to run round (searching)’ (New Testament, Mark, 6, 55; AD 1) ȱȱ ȬΣ·Ν -ago: to lead/draw/bring round’ ΔΕΓȱȬΣ·Νpro -ago: ‘to lead forward or onward/escort on their way’ (New Testament, Matthew, 2, 9; AD 1) ЀΔΉΕȱȱȬνΛΝȱhyper -ekho: ‘to hold over’ ЀΔΓȱȱȬΘΕνΛΝ hypo -trekho: ‘to run in under/run in between/intercept’ (15) a.ȱ ̏νΑΝΑȱȱ ΔΉΕ΍ΔΓΕΉϾΉΘ΅΍ȱȱ ΘχΑȱȱ ̖Ε΍ΎΓΐϟ΅Α Meno:n periporeuetai te:n Trikomian Meno.NOM go-round.NACT.PRES.3SG the.ACC Trikomia.ACC ‘Meno goes round Trikomia’ (Urkunden der Ptolemäerzeit (ältere Funde) 76, 2) b.ȱ πΔΉ΍ΕκΘΓȱȱ Έ΍΅Ά΅ϟΑΉ΍Αȱȱ ΘϲΑȱȱ ΔΓΘ΅ΐϱΑȱ epeirato diabainein ton potamon try.NACT.IMPERF.3SG pass-over.INF the.ACC river.ACC ‘he was trying to pass over the river’ (Diodorus Siculus, Library, 22, 13, 4; 1 BC)ȱ ΓЇΑȱȱ πΎΉϧΑΓ΍ȱȱ Ύ΅Θ΅ΗΓΚϟΊΓΑΘ΅΍Ъ
c.ȱ Θϟȱȱ ti oun ekeinoi katasophizontai? what.ACC prt these.NOM outwit-by-sophisms-or-fallacies.NACT.PRES.3PL ‘Hence what may they outwit by sophisms?’ (Aelian, On the characteristics of animals, 12, 42, 12; AD 2-3) d.ȱ ωΕЏΘ΅ȱȱΔϱΘΉΕΓΑȱϴΕΌΓϧΖȱΘΓϧΖȱȱΈϱΕ΅Η΍ΑȱȱύȱΎΉΎΏ΍ΐνΑΓ΍ΖȱΈ΍΅ΔΓΕΉϾ΋Θ΅΍ȱȱȱȱȱȱȱȱȱȱΘχΑȱΛЏΕ΅Αȱ΅ЁΘЗΑȱ e:ro:ta poteron orthois tois dorasin e: keklimenois diaporeue:tai te:n kho:ran auto:n ask.3SG whether

upright the spears.DAT or bowed pass-through.NACT.PRES.3SG the land.ACC theirȱ

‘he was asking if he should pass through their land holding the spears upright or bowed’ (Plutarch, Lysander, 22, 2, 3; AD 1-2) ȱȱȱȱȱȱȱȱ e.ȱ Ύ΅ϠȱȱΔ΅ΕνΕΛΉΗΌΉȱȱ ΘχΑȱȱΎΕϟΗ΍Αȱȱ ȱΎ΅ϠȱΘχΑȱȱΦ·ΣΔ΋ΑȱȱΘΓІȱΌΉΓІ kai parerkhesthe te:n krisin kai te:n agape:n tou theou ȱ ȱ and pass/disregard.NACT.PRES.2PL the judgement.ACC and the love.ACC the god.GEN ‘and you disregard the judgement and the love of God’ (New Testament, Luke, 11, 42, 4; AD 1) f.ȱ ΔΕΓΣ·Ή΍ȱȱ ΘΉȱȱ ΅ЁΘΤΖȱȱ Ύ΅Ϡȱȱ πΒΣ·Ή΍ȱȱ ψȱȱ ΦΌΙΐϟ΅ȱ proagei te autas kai eksagei he: athymia lead-forward.ACT.PRES.3SG and

these.ACC and

release.ACT.PRES.3SG

the

lack-of-spirit.NOM

‘lack of spirit leads forward and releases them’ (Aelian, On the characteristics of animals, 1, 15, 9; AD 2-3)

In any case, the formation of verbs denoting change-of-state with a prefix, from the Hellenistic Koine period and after, is frequent; despite that, we cannot consider there to be a direct connection between the prefixes and causativisation since this is not the sole function of prefixes (see point (b) [change into intransitive] below: the result of the addition of the prefix is an intransitive type). ΦΔϱȬȱapo- ȱ ΦΔΓ·΅΍Зȱapogaio: ‘to form a stone’ ȱ (16) ΦΔΓ·΅΍ЏΗϙΖȱȱ

ΘχΑȱȱ Δ΅ΛΉϧ΅Αȱȱ ЂΏ΋Αȱ apogaio:se:is te:n pakheian hyle:n form-a-stone.ACT.FUT.2SG the.ACC thick.ACC material.ACC ‘you will form a stone from the thick material’ (Pseudo-Galen, De affectuum renibus insidentium dignotione et curatione liber, 19, 672, 10; post AD 2)

(b) Change into intransitive (not only anticausative) verb a. Ω·Ν ago: ‘to lead someone’ ЀΔΣ·Ν hypago: ‘to go’ ȱ ΔΉΕ΍Σ·Ν periago: ‘to pass by’ ȱ πΔ΅ΑΣ·Ν epanago: ‘to return’ ΔΕΓΣ·Ν proago: ‘to go in front’

Diachronic Data and Analysis

101

b. ΗΘΕνΚΝ strepho: ‘to lead someone in the direction of’ ȱ ЀΔΓΗΘΕνΚΝ hypostrepho: ‘to return’ ȱ πΔ΍ΗΘΕνΚΝ epistrepho: ‘to return’ ΦΑ΅ΗΘΕνΚΝ anastrepho:ȱ‘to come back/return’ ΦΔΓΗΘΕνΚΝ apostrepho:ȱ‘to move off’ (17) a. БΖȱȱ ·ΤΕȱȱ ЀΔΣ·Ή΍Ζȱȱ ΐΉΘΤȱȱ ΘΓІȱΦΑΘ΍ΈϟΎΓΙȱΗΓΙȱȱ πΔȇȱȱ ΩΕΛΓΑΘ΅ ho:s gar hypageis meta tou antidikou sou ep’ arkhonta ȱȱ when because go.ACT.PRES.2SG with the adversary your to magistrate ‘for when you are going with your adversary before the magistrate’ (New Testament, Luke, 12, 58; AD 1) πΖȱȱ Θϲȱȱ ΅ЄΏ΍ΓΑȱȱ Θϲȱȱ ϥΈ΍ΓΑ b.ȱ Ύ΅Ϡȱȱ ЀΔΓΗΘΕνΚΉ΍ȱȱ kai hypostrephei es to aulion to idion and return.ACT.PRES.3SG to the cottage the his-own ‘and he returns to his own cottage’ (Aelian, On the characteristics of animals, 12, 7, 18; AD 2-3) c.ȱ ̕ΐ΍ΎΕϟΑ΋Ζȱȱ ΦΑ΅ΗΘΕνΚΉ΍ȱȱ πΒȱȱ ΩΗΘΉΝΖ Smikrine:s anastrephei eks asteo:s Smikrinis.NOM return.ACT.PRES.3SG from city ‘Smikrinis returns from the city’ (Menander, The Litigants (Epitrepontes), 577; 4-3 BC)

3.2.2.3 Changes in accusative verbs: sub-categories In traditional grammars, the innovative accusative (transitive non-causatives) verb types are dealt with in exactly the same way as the innovative causative types (Mayser 1934/1970: 89-105; Jannaris 1897/1968: 356-365; Schwyzer & Debrunner 1950/2002: 279-302; Humbert & Kourmoulis 1957/2002: 94-104). We can see, even so, from the examination of data that the presence of a verb with a complement in the accusative case does not constitute in all instances a similar phenomenon. The following instances do not constitute causativisation of intransitive verbs: (a) Intransitive verbs denoting profession changed into accusative verbs with specific denotation of the profession as the direct object (specification of verb meaning). This change is concerned with verbs belonging to the semantic group of activities; the DP-accusative with these verbs constitutes an argument that derives only from the idiosyncratic verb meaning, not from the meaning of the event schema of the verb (Grimshaw 1991, 1993; Rappaport Hovav & Levin 1998): active and non-active morphology ·Ε΅ΐΐ΅ΘΉϾΝ grammateuo: ‘to be secretary/hold his office’ ΎΝΐ΅ΕΛνΝ ko:markheo: ‘to be the head man of a village’ȱ ΚΙΏ΅Ύ΍ΘΉϾΝ phylakiteuo: ‘to serve as police official’ ΛΓΕ΋·З khore:go: ‘to lead a chorus/defray the cost of bringing out a chorus at the public festivals’ȱ ΘΕ΍΋Ε΅ΕΛЗ trie:rarkho: ‘to be trierarch’ ΦΑΘ΍·ΕΣΚΓΐ΅΍ȱantigraphomai ‘to be a checking or copying clerk, a public officer’ ȱ

(18) a.ȱȱ transitive ϳȱȱ ΌΉϲΖȱȱ ψΐϧΑȱȱ οΓΕΘΤΖȱȱ ΛΓΕ΋·Ήϧȱ ȱ ho theos he:min heortas khore:gei the.NOM god.NOM we.DAT feast.ACC defray-the-cost.ACT.PRES.3SG ‘the God grants us feasts’ (Plutarch, On tranquillity of mind, 477e, 1; AD 1-2) i in the previous period - Classical Greek:
b.ȱ intransitive ΓЁΎȱΦΔΓΎνΎΕΙΔΘ΅΍ȱΘχΑȱΓЁΗϟ΅ΑǰȱȱȱΦΏΏΤȱΛΓΕ΋·ΉϧȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ϠȱΘΕ΍΋Ε΅ΕΛΉϧȱ

ouk apokekryptai

te:n ousian, ȱȱalla

khore:gei

ȱȱȱȱȱkai trie:rarkhei

not concealed.3SG ȱȱȱȱȱthe property.ACC but defray-the-cost.ACT.PRES.3SG and act-as-trierarch.ACT.PRES.3SG

‘(Demochares) has not concealed his property, but acts as choregus and as trierarch’ (Demosthenes, Against Aphobus 2, 3, 3; 4 BC)

Regarding the non-active intransitives that are used as accusative (transitive non-causatives) verbs, the innovative transitive use may also be accompanied by voice change (non-activesĺ actives), or a DP may simply be added without voice change (non-active intransitive verbs ĺ non-active accusative verbs + DP; active intransitive verbs ĺ active accusative verbs + DP). As regards this characteristic, the addition of DP-accusative differs from causativisation (the new causatives bear only active morphology).

102

Chapter Three

(b) Greater specialisation of verb meaning, but not, however, causative interpretation, is given to intransitive verbs in the New Testament (Mayser 1934/1970) with the addition of a direct object. We can distinguish the following cases: (i) verbs denoting an action that becomes specified with the presence of a direct object. active πΑΉΕ·Зȱenergo: ‘to be in action or activity/operate’ (Mayser 1934/ 1970: II 1, 87ff., II 2, 310ff.) ϡΉΕΓΙΕ·Зȱhierourgo: ‘to sacrifice/minister the gosper’ ЀΆΕϟΊΝȱhybrizo:ȱ‘to run riot (in the use of superior strength or power)’ ȱ πΑΉΈΕΉϾΝȱenedreuo: ‘to lie in wait for/lay snares for’ȱ ΔΏΉΓΑΉΎΘЗȱpleonekto: ‘to have or claim more than one's due/be greedy/have some advantage’ȱ ΉЁΗΉΆЗȱeusebo: ‘to live or act piously or reverently’ ȱ ΌΕ΍΅ΐΆΉϾΝȱthriambeuo: ‘to triumph’ȱ ΐνΑΝȱmeno:ȱ‘to stay/wait’ ȱ Ύ΅ΕΘΉΕЗȱkartero:ȱ‘to be steadfast, patient’ȱ ϴΑΉ΍ΕЏΘΘΝȱoneiro:tto: ‘to dream’ [intransitive in Classical Greek, transitive with DP in accusative case in

Polybius: the change does not lead to causative interpretation (‘to cause someone to dream’), but to denotation of a specific object which is implied (Polybius, Histories, 5,108,5; 17,15,13; 3-2 BC)]ȱ non-active πΐΔΓΕΉϾΓΐ΅΍ȱemporeuomai ‘to travel for business/make gain of/cheat’ ȱ

(19) a. Γϡȱȱ ·ΤΕȱȱ ̍κΕΉΖȱȱ ΎϾΑ΅Ζȱȱ ϡΉΕΓΙΕ·ΓІΗ΍ hoi gar Kares kynas hierourgousi the.NOM prt Carians.NOM dogs.ACC sacrifice.ACT.PRES.3PL ‘as the Carians sacrifice dogs’ (Plutarch, Alexandrian Proverbs, 1, 73, 2; AD 1-2) i in the previous period - Classical Greek:
intransitive b. ΗΛΉΈϲΑȱȱ ·ΤΕȱȱ ΓЁΈξȱΉϩΖȱȱȱȱȱȱȱȱȱȱȱ ΘЗΑȱȱ Κ΅ϾΏΝΑȱȱȱȱȱ ΔΕϲΖȱ ΦΏφΌΉ΍΅Αȱȱ ϡΉΕΓΙΕ·Ήϧȱȱ skhedon gar oude heis to:n phaulo:n pros ale:theian hierourgei near prt not one.NOM the mean.GEN to truth sacrifice.ACT.PRES.3SG ‘nearly none of the mean men truly performs sacred rites’ (Aristotle, Fragments, 1, 16, 102; 4 BC) (20) a. ГΗΘΉȱȱȱΎ΅ΘΤȱȱȱ ΘΓϿΖȱȱ ЂΔΑΓΙΖȱȱȱΘϲΑȱȱ ̘ϟΏ΍ΔΔΓΑȱȱ Θ΅ІΘȇȱȱ ϴΑΉ΍ΕЏΘΘΉ΍Αȱ
ho:ste kata tous hypnous ton Philippon taut’ oneiro:ttein that during the sleeps the.ACC Philip.ACC these.ACC dream.ACT.PRES.INF ‘that Philip even dreamed of them in his sleep’ (Polybius, Histories, 5, 108, 5; 3-2 BC) i in the previous period - Classical Greek:
intransitive b. ·ΉΝΐΉΘΕϟ΅ΖȱΘΉȱΎ΅ϠȱΘΤΖȱȱȱΘ΅ϾΘϙȱȱȱοΔΓΐνΑ΅ΖǰȱȱȱȱϳΕЗΐΉΑȱȱȱБΖȱȱȱϴΑΉ΍ΕЏΘΘΓΙΗ΍ȱΐξΑȱȱΔΉΕϠȱΘϲȱȱϷΑȱ geo:metrias te kai tas taute:i hepomenas, horo:men ho:s oneiro:ttousi men peri to on geometry.GEN and

the.ACC that.DAT following.ACC see.1PL as

dream.ACT.PRES.3PL about the being

‘geometry, and the studies that accompany it, are, as we see, dreaming about being’ (Plato, Republic, 533b; 5-4 BC)

(ii) Psych-verbs; here the direct object denotes the cause of the specific emotional state.ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ active Ό΅ΕΕЗȱtharro:ȱ‘to be of good courage/take courage’ȱ Ό΅ΙΐΣΊΝȱthaumazo:ȱ‘to wonder/marvel’ȱ πΏΉЗȱeleo:ȱ‘to have pity on/show mercy to/feel pity’ ȱ ΓϢΎΘϟΕΝȱoiktiro:ȱ‘to pity, pity for or because of a thing’ȱ ΎΏ΅ϟΝȱklaio:ȱ‘to weep for/lament/cry for’ ȱ ΔΉΑΌЗȱpentho:ȱ‘to bewail/lament’ȱ ΉЁΈΓΎЗȱeudoko:ȱ‘to be content with/find pleasure in a person or thing’ȱ ΉЁΛ΅Ε΍ΗΘЗȱ(ΘΓІΘΓ) eukharisto: (touto) ‘to be glad for that’

(21) a. ΫΕ΅ȱȱ ΐχȱȱΘΤȱȱ ΐ΍ΎΕϱΘ΅Θ΅ȱȱ ΘЗΑȱȱ ̕Θ΅ΘΉϟΕ΅ΖȱΎΏ΅ϟΝȱȱ Ύ΅ΎЗΑǯǯǯDz ara me: ta mikrotata to:n Stateiras klaio: kako:n...? therefore not the.ACC smallest.ACC the.GEN Statira cry.ACT.PRES.1SG bad.GEN ‘therefore, I do not cry even about the smallest disasters in Statira...?’ (Plutarch, Alexander, 30, 8, 5; AD 1-2) i in the previous period - Classical Greek:
intransitive

Diachronic Data and Analysis

103

b. ΎΏ΅ϟΉ΍ȱȱ ·ȇȱȱ ΩΎΓ΍Θ΍Αȱȱ πΑȱȱ ΛΉΕΓϧΑȱȱ ΚϟΏ΋Αȱȱ σΛΝΑ klaiei g’ akoitin en kheroin phile:n ekho:n cry.ACT.PRES.3SG prt wife.ACC in hands beloved.ACC have.PART.NOM ‘yes, (he) weeps, holding his beloved wife in his arms’ (Euripides, Alcestis, 201; 5 BC) non-active Ύ΅ΙΛκΐ΅΍ kaukhamai ‘to boast/vaunt oneself’ ΎϱΔΘΓΐ΅΍ȱkoptomai ‘to beat or strike oneself/beat one's breast or head through grief30’ȱ ΦΔΓΕΓІΐ΅΍ȱaporoumai ‘to look away from’ ȱ ϡΏΣΗΎΓΐ΅΍ȱhilaskomai ‘to be merciful, gracious’ȱ ΅ϢΈΓІΐ΅΍ȱaidoumai ‘to be ashamed’ȱ Ύ΅Θ΅ΔΏφΗΗΓΐ΅΍ȱkataple:ssomai ‘to be panic-stricken, astounded’ ȱ

(22) a.ȱ Θϲȱȱ ·ΤΕȱȱ ϴΒϾΕΕΓΔΓΑȱȱ ΘϛΖȱȱ ΘϾΛ΋Ζȱȱ ϡΏΣΗΎΓΐ΅΍ to gar oksyrropon te:s tykhe:s hilaskomai ȱ

the.ACC prt unstable.ACC the.GEN luck.GEN be-gracious.NACT.PRES.1SG ‘I am appeased with the instability of luck’ (Heliodorus, Ethiopian Romance, 10, 2, 1; AD 3) i in the previous period - Classical Greek:
intransitive ·ΙΑ΅΍Ύ΋ϟΓ΍Η΍ȱȱ ΎΉΕΘϱΐΓ΍Η΍ȱȱ ϡΏΣΗΎΓΑΘΓȱ b.ȱ ΛΓΕΓϧΗ΍ȱȱ khoroisi gynaike:ioisi kertomoisi hilaskonto choruses.DAT female.DAT satirical.DAT be-gracious.NACT.IMPERF.3PL ‘they were gracious with the female choruses in the satirical mode’ (Herodotus, The Histories, 5, 83, 3; 5 BC)

(c) A third mode of movement from the class of intransitive verbs to the class of accusative verbs constitutes the replacement of the PP by DP in the accusative case. (23) a. Ύ΅Θ΅·ΝΑϟΗΘ΅ΗΌ΅΍ȱȱ ΘΓϿΖȱȱ ̝Λ΅΍ΓϾΖ katago:nistasthai tous Akhaious subdue.NACT.AOR.INF the.ACC Achaeans.ACC ‘to subdue the Achaeans’ (Polybius, Histories, 2, 45, 4; 3-2 BC) b.ȱ Ύ΅Θ΅·ΝΑϟΊΉΘ΅΍ȱȱ Θϲȱȱ ΜΉІΈΓΖȱȱȱ katago:nizetai to pseudos overpower.NACT.PRES.3SG the.ACC falsehood.ACC ‘(she) overpowers falsehood’ (Polybius, Histories, 13, 5, 6; 3-2 BC)

Even if an innovative anticausative use of originally non-alternating causative verbs is not in evidence, the innovative intransitive use of originally exclusively accusative verbs can be observed. It appears, consequently, that the movement from the [+transitive] (= [+accusative case]) to [-transitive (= [accusative case] cannot be completely ruled out, but the productive and systematic change from causative non-alternating to causative alternating verb (with innovative availability of anticausative structure) is ruled out. The verbs πΎΘ΍ΑΣΗΗΝȱŽ”’—Šœœ˜DZȱ‘to turn away somebody’, πΎΔΏνΎΝȱŽ”™•Ž”˜DZ‘to prepare’ are also used from this period (chiefly from the Roman period) in intransitive structures (and they continue to be used in the accusative-transitive structures), with the meanings ‘to jolt’ and ‘to prepare oneself’ respectively (Kapsomenakis 1938 and Mayser 1934/1970: II 1, 82). In the majority of examples of new intransitive use of accusative verbs, however, we see omission of the implied object. Many accusative verbs in the New Testament are used without a direct object (the direct object arises easily from the context): ΐχȱȱ ΦΔΓΗΘΉΕφΗϙΖ

me: apostere:se:is not defraud.ACT.AOR.2SG (New Testament, Mark, 10, 19; AD 1) ΦΈ΍ΎΉϧΘΉȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ adikeite do-wrong.ACT.PRES.2PL (New Testament, 1 Corinthians, 6; AD 1) ϴΚΉϟΏΉ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ opheilei owe.ACT.PRES.3SG (New Testament, Matthew, 23, (16, 18); AD 1)

30

Cf. Giannakis 1999.

104

Chapter Three ǻŘŚǼȱ a. ΦΏΏΤȱȱЀΐΉϧΖȱȱȱȱ ΦΈ΍ΎΉϧΘΉȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ȱΎ΅ϠȱΦΔΓΗΘΉΕΉϧΘΉǰȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ϠȱΘΓІΘΓȱΦΈΉΏΚΓϾΖȱ alla hymeis adikeite ȱkai apostereite, kai touto ȱadelphous but you.NOM do-wrong.ACT.PRES.2PL and defraud.ACT.PRES.2PL and that brothers.ACC ‘but you yourselves do wrong, and defraud, and that against your brothers’ (New Testament, 1 Corinthians, 6, 8; AD 1) i in the previous period - Classical Greek:
transitive ΦΈ΍ΎЗ

b. ϵΘ΍ȱȱ ΓЁΈνΑ΅ȱȱ




hoti oudena adiko:

that nobody.ACC do-wrong.ACT.PRES.1SG ‘that I wrong no man’ (Xenophon, Economics, 11, 22, 6; 5-4 BC)

In contrast to the causative verbs that take a direct object consistently in the accusative case, the non-causative verbs exhibit parallel use of other cases or PPs and accusative, so the accusative case finally spreads to all sub-classes of non-causative verbs. From the Ptolemaic era onwards, the use of the accusative case increases, in parallel with the cases –genitive and dative– that were assigned, in earlier times, by the non-causative verbs to their direct object. This alternation of cases in Hellenistic-Roman Koine constitutes a significant difference between the non-causative verbs and the causative verbs, which only take a DP in the accusative case as a complement: ΦΎΓϾΝȱȱ

ΚΝΑφΑȱȦȱȱ

ΚΝΑϛΖȱ

akouo: phone:n / phone:s hear.ACT.PRES.1SG voice.ACC / voice.GENȱȱ ȱ ·ΉϾΓΐ΅΍ȱȱ

ΗΙΐΔϱΗ΍ΓΑȱȦȱȱ

ΗΙΐΔΓΗϟΓΙ

geuomai symposion / symposiou enjoy.NACT.PRES.1SG symposium.ACC / symposium.GEN πΔ΍Ώ΅ΑΌΣΑΓΐ΅΍ȱȱ

ΐνΘΕ΅ȱȦȱȱ

ΐνΘΕΝΑ

epilanthanomai forget.NACT.PRES.1SG

metra / rules.ACC /

metro:n rules.GEN

πΑΓΛΏνΝ+ DP-accusative (or -dative)

enokhleo: annoy.ACT.PRES.1SG Δ΍ΗΘΉϾΝ+ DP-accusative (or -dative)

pisteuo: believe/trust.ACT.PRES.1SG (25) a. ΐΓ΍ȱȱ πΑΓΛΏΉϧΖȱ moi enokhleis  I.DAT annoy.ACT.PRES.2SG ‘you annoy me’ (Heliodorus, Ethiopian Romance, 10, 9, 5; AD 3) b. Θϟȱ ΐȂȱ πΑΓΛΏΉϧΖǰȱ ΘΣΏ΅ΑDz ti m’ enokhleis, talan? why I.ACC annoy.ACT.PRES.2SG misery.VOC ‘you, misery, why do you annoy me?’ (Menander, Misumenus, 189; 4-3 BC)

The alternation in the case requirements of non-causative verbs (genitive/dative/PP) strengthened the process of the replacement of the dative and the genitive cases by the accusative case. Below, ǿ cite verbs that could alternatively take complements of different types: πΏΉΙΌΉΕЗȱeleuthero set-free.ACT ΗФΊΝȱso:izo save.ACT ΩΕΛΓΐ΅΍ȱarkhomai be-ruled/governed.NACT ȱ ΐΣΛΓΐ΅΍ȱmakhomai fight.NACTȱ ΦΎΓΏΓΙΌЗȱakoloutho: follow.ACT ρΔΓΐ΅΍ȱhepomaiȱfollow.NACT Λ΅ΏΉΔ΅ϟΑΝȱkhalepaino be-angry-with.ACTȱ Θ΍ΐΝΕΓІΐ΅΍ȱtimo:roumaiȱexact-vengeance-from.NACT

(26) a. ΘХȱ to:i the.DAT

ΎΏνΔΘϙȱ

ΓЁȱ

klepte:i ou thief.DAT not

+ DP-GEN / by, from-PP + GEN + DP-GEN / by, from-PP + GEN ƸȱDP-GEN / by, from-PP + GEN ƸȱDP-DAT / against-PP + ACC + DP-DAT / with-PP + DATȱ + DP-DAT / with-PP + DAT ƸȱDP-DAT / at-PP + DAT / for-PP + DAT + DP-GEN / about-PP + GEN

Λ΅ΏΉΔ΅ϟΑΉ΍Ζȱ

khalepaineis be-angry.ACT.PRES.2SG

Diachronic Data and Analysis

105

‘you will not be angry with the thief’ (Epictetus, Discourses, 1, 18, 11; AD 1-2) Έ΍Τȱ ·νΕΓΑΘ΅ȱ ΔΓΑ΋ΕϲΑȱ ΉЁΑΓІΛΓΑȱ ȱ b. ΉϢȱ Λ΅ΏΉΔ΅ϟΑΉ΍Ζȱ ei khalepaineis dia geronta pone:ron eunoukhon  if be-angry.ACT.PRES.2SG because old-man.ACC dishonest.ACC castrated.ACC ‘if you are angry with an old dishonest and castrated man’ (Plutarch, Artaxerxes, 17, 8, 4; AD 1-2) Θ΅ϧΖȱ ΎΓ΍Α΅ϧΖȱ ΘЗΑȱ ΦΑΌΕЏΔΝΑȱ ΔΕΓΏφΜΉΗ΍Αȱ

(27) a. ΓЁΈξȱ ΐ΅ΛϱΐΉΌ΅ȱ oude makhometha

tais

koinais

to:n

anthro:po:n prole:psesin

neither fight.NACT.PRES.1PL the.DAT common.DAT the.GEN people.GEN preconception.DAT

‘we do not fight against the common preconceptions of the people’ (Sextus Empiricus, Against the Mathematicians, 8, 157, 2; AD 2-3) b. ΔΏ΅ΑЏΐΉΌ΅ȱΎ΅Ϡȱ ΔΕϲΖȱΦΏΏφΏΓΙΖȱ ΐ΅ΛϱΐΉΌ΅Dzȱ  plano:metha kai pros alle:lous makhometha? differ.1PL and to each-other.ACC fight.ACT.PRES.1PL ‘(here, in things) about which all men differ and dispute?’ (Epictetus, Discourses, 2, 7, 8; AD 1-2)

Spread of the accusative instead of the dative and the genitive case can also be observed in the texts of Polybius (Krebs 1882; de Foucault 1972: 103-105): (28) a.ȱ ΘΤΖȱȱ ̕ΙΕ΅ΎΓϾΗ΅Ζȱȱ πΔΓΏνΐΓΙΑ tas Syrakousas epolemoun  the.ACC Syracuse.ACC fight.ACT.IMPERF.3PL ‘they were fighting against Syracuse’ (Polybius, Histories, 2, 42, 3; 3-2 BC) ΘΓϿΖȱȱ ͦ΋·ϟΑΓΙΖ b.ȱȱ Δ΅Ε΅ΗΔΓΑΈΉϧΑȱȱ paraspondein tous Rhe:ginous break-faith.ACT.PRES.INF the.ACC Rhegians.ACC ‘to break faith with the people of Rhegium’ (Polybius, Histories, 1, 7, 8; 2, 46, 3; 3-2 BC)

The replacement of the dative by the accusative case is very frequent in the texts of the New Testament for objects of verbs meaning ‘disapproval, hate, promise’ (Humbert 1930; Mihevc-Gabrovec 1960: 2024; Jannaris 1897: 341-347): ΉЁ΅··ΉΏϟΊΓΐ΅΍ euaggelizomai ‘to bring good news’ ΚΌΓΑνΝ phthoneo: ‘to be envious’

(29) a. ϢΈΓϿȱ ·ΤΕȱ ΉЁ΅··ΉΏϟΊΓΐ΅΍ȱ ЀΐϧΑȱ Λ΅ΕΤΑȱ ΐΉ·ΣΏ΋Α   idou gar euaggelizomai  hymin kharan megale:n behold prt bring-good-news.NACT.PRES.1SG you.DAT joy.ACC great.ACC ‘for behold, I bring you good news of great joy’ (New Testament, Luke, 2, 10, 3; AD 1) b. ΓЁȱ ΚΌΓΑΓІΐΉΑȱ ΓЇΑȱ ψΐΉϧΖȱ ΓЁΈνΑ΅ȱ ȱ   ou phthonoumen oun he:meis oudena  not be-envious.ACT.PRES.1PL therefore we.NOM nobody.ACC ‘therefore, we are not envious of anybody’ (Palladius, de gentibus Indiae et Bragmanibus, 2, 57, 15; AD 4-5)

In Koine of the Roman period, the accusative case is used as the default case, not only for direct objects but also instead of oblique cases. The accusative is also used as the default case assigned by prepositions. Moreover, the adverbial use of the accusative is extended, especially in the denotation of location, time, and cause. Finally, the distinction between the location meaning of the dative and the ‘movement towards a direction’ meaning denoted by the accusative case leads gradually to its abandonment in favour of the accusative. (30) a.ȱ Θ΅ІΘ΅ȱȱ Έξȱȱ ΔΣΑΘ΅ȱȱ ΗΙΑΉΑϛȱȱ ΉϢΖȱȱ ΘϲΑȱȱ Λ΍ΘЗΑ΅ tauta de panta synene: eis ton khito:na these.NOM prt all.NOM unite.3SG to the.ACC chiton.ACC ‘all these unite into the chiton’ (Oxyrhynchus Papyrus, VI, 929; AD 2) ΉϢΖȱȱ ΘϲΑȱΦ·ΕϲΑȱ ΐχȱ πΔ΍ΗΘΕΉΜΣΘΝȱȱ ΉϢΖȱȱ ΘΤȱȱ ϴΔϟΗΝ b.ȱ ϳȱȱ ho eis ton agron me: epistrepsato: eis ta opiso: the.NOM to the field not return.ACT.AOR.IMP.3SG to the.ACC back ‘let him who is in the field not return back’ (New Testament, Mark, 13, 16; AD 1)

The frequent replacement of the accusative of reference by the dative case constitutes a change in the opposite direction. (31)ȱ Ύ΅Ό΅Ε˜ϟȱȱ katharoi

ΘϜȱȱ

Ύ΅ΕΈϟθ

te:i

kardiai



106

Chapter Three pure.NOM the.DAT heart.DAT ‘(blessed are) the pure in heart’ (New Testament, Matthew, 5, 8; AD 1)

The type of dative that survived the longest was the personal dative (Browning 1983). The accusative, and later the genitive, of personal pronouns replaced the dative case from the 1st century BC in nonliterary papyri, in phrases such as: (32) a. ·ΕΣΚΓΐ΅ϟȱȱ ΗΉ (1 BC), graphomai se write.NACT.PRES.1SG you.ACC

c. ΦΔΓΗΘΉΏЗȱȱ ΗΉ (1 BC) apostelo: se send-off.ACT.FUT.1SG you.ACC

ȱ

d. Φ·ϱΕ΅ΗϱΑȱȱ ȱ agorason buy.ACT.AOR.IMP.2SG

b.ȱȱ ΉϥΕ΋ΎΣȱȱ eire:ka tell.ACT.PERF.1SG

ΗΓΙ (AD 4), sou you.GEN

ΐΓΙ (AD 2) mou I.GEN

3.2.3 Transitivity alternation derivation and voice: changes towards marking of the anticausative with active voice The aim of this section is the examination of voice of transitive and intransitive types during the Hellenistic-Roman period. The analysis is two-pronged: (a) examination of the arguments accompanying transitive and intransitive types so that possible differences between the types appear, and chiefly so we can analyse the differences between the PPs that accompany the non-active passive types and the active and non-active anticausative types (always in relation to the question of the derivation direction); and (b) distribution of the active and the non-active voice and analysis of the role of voice in relation to (in)transitivity. 3.2.3.1 Anticausatives, passives, and argument realisation In relation to the expression of arguments with anticausative and passive types I have ascertained the following: (i) In Hellenistic Koine, ЀΔϱ hypo ‘by’ was the preposition that was common to all verbs for denotation of the agent (the preposition Δ΅ΕΣ para ‘beside/from’ only accompanied verbs such as ʌȑȝʌȦ pempo: ‘send’ (ditransitive verbs); in other words, it was used mainly for the goal indirect object) (George 2005)31. (ii) In the era of the translation of the Septuagint, ЀΔϱ hypo ‘by’ was the preposition common to all verbs for the expression of the agent. Moreover, the preposition ΦΔϱ apo ‘by’ appears with verbs that mean sending, mainly denoting the source. Specifically, the preposition ΦΔϱ apo ‘by’ seems to begin with DPs that are closer to the location and the cause meaning (cf., in Modern Greek, the use of the preposition ΦΔϱ apo ‘by’ also for locative meanings). (iii) In the New Testament, ЀΔϱ hypo ‘by’ was the common preposition for the expression of the agent. The preposition ΦΔϱ apo ‘by’ is used when the agent is inanimate or with compound verbs with the prefix ΦΔϱ- apo-. (33) a. ΦΏΏΤȱΎ΅ΌАΖȱ ΈΉΈΓΎ΍ΐΣΗΐΉΌ΅ȱ ЀΔϲȱȱΘΓІȱΌΉΓІȱΔ΍ΗΘΉΙΌϛΑ΅΍ȱ Θϲȱ ΉЁ΅··νΏ΍ΓΑȱ   alla ȱkatho:s dedokimasmetha hypo tou theou pisteuthe:nai to euaggelion  but ȱas ȱapprove.NACT.PERF.1PL ȱ by the God entrust.INF ȱ the gospel.ACC  ‘but even as we have been approved by God to be entrusted with the gospel’ȱǻNew Testament, 2 Thessalonians, 4, 2; AD 1) ΦΔϲȱ ΦΗΌΉΑΉϟ΅Ζȱ b. πΈΙΑ΅ΐЏΌ΋Η΅Αȱ edynamo:the:san apo astheneias become-strong.PASS.AOR.3PL from weakness ‘from weakness they were made strong’ (New Testament, Hebrews, 11, 34, 3; AD 1)

In Polybius, the use of the dative case is also frequent (according to the classical Ancient Greek model) with past tense, present perfect tense or verbal adjective: (34) ЀΔξΕȱȱ hyper 31

ЙΑȱȱ

ψΐϧΑȱȱ

ΉϥΕ΋Θ΅΍

ho:n

he:min

eire:tai

In contrast, in Classical Greek, agent and cause are expressed consistently with the preposition ЀΔϱȱhypo ‘by’ (PP) or with DP in the dative.

Diachronic Data and Analysis

107

of which.GEN we.DAT speak.NACT.PERF.3SG ‘of which we have already had occasion to speak’ (Polybius, Histories, 3, 47, 4; 3-2 BC)

Polybius also uses the preposition Ѐʌϱ hypo ‘by’ (even close to a present perfect type): (35) ΅ЁΒϱΐΉΑΓΖȱȱ ЀΔϲȱΘΓІȱȱ ΔΏφΌΓΙΖȱȱ ΘЗΑȱȱ ȱȱȱΦΑ΅Θ΋ΎΓΐνΑΝΑȱΛ΍ϱΑΝΑ auksomenos hypo tou ple:thous to:n anate:komeno:n khiono:n swell.NACT.PRES.PART.NOM by the.GEN quantity.GEN the.GEN melting.GEN snows.GEN ‘because it is then swollen by the melting snows’ (Polybius, Histories, 2, 16, 9; 3-2 BC)

3.2.3.2 Distribution of active and non-active voice (a) Passive structures In contrast to Classical Greek, in which it was possible for there to be passive structures with (expression of the agent and) active voice, in the Hellenistic-Roman period, only anticausatives (and not passives) that take PPs denoting cause (Δ΅ΕΤȱparaȱȦȱЀΔϱȱhypo) can bear active voice. (36)ȱ ΐχȱȱЀΔϲȱȱȱ·ΙΑ΅΍ΎϲΖȱύȱΔ΅΍ΈϲΖȱόȱȱΘ΍ΑΓΖȱȱΔ΅Ε΅ΔΏ΋ΗϟΓΙȱΦΑΌΕЏΔΓΙȱΓϡȱȱΦΐΉϟΑΓΑΉΖȱȱΦΔΓΌΑϚΗΎΝΗ΍Α me: hypo gynaikos e: paidos e: tinos paraple:siou anthro:pou hoi ameinones apothne:isko:sin not by

woman

or child

or somebody such-like

human the.NOM better.NOM die.ACT.PRES.3PL

‘so that the better sort do not die because of a woman or child or a such like human being’ (Plutarch, Apophthegmata Laconica, 228d, 12; AD 1-2)

As in Classical Greek, in the post-classical period, it is possible to have in the subject position of nonactive types arguments that are in the genitive or dative case in the corresponding transitive structure. At the same time, non-active types can take DPs in the genitive, dative, or accusative (in the case of ditransitive verbs) as complements. Mayser 1934/1970, Ptolemaic era: ȬΦΑΘ΍Ώ΅ΐΆΣΑΓΐ΅΍
antilambanomai ‘to hold back/interrupt’ (Rhein. 17, 17 (110-99) / 18, 35 / 19, 2 (108a) / Fay. 12, 34 / Lond. II 401(p.13) 30 (116 – 11a)) ȬΎ΅Θ΅·ΉΏЗΐ΅΍ȱkatagelo:maiȱ‘to be derided’ (PSI VI 577, 26 (248a) ȬΦΕΎ˜Іΐ΅΍ȱarkoumai ‘to be satisfied with’ (Par. 38, 11 (160a) / BGU VI 1247, 13 (149a) / Tor. I 2, 18 (117a) (37)ȱ πΎΉϧΑΓ΍ȱ ΈȂȱ ΦΕΎΓІΑΘ΅΍ȱ ΘΓϧΖȱ ο΅ΙΘЗΑȱ  ekeinoi d’ arkountai tois heauto:n  they.NOM prt satisfy.NACT.PRES.3PL the.DAT their.GEN ‘but they are satisfied with their own’ (Epictetus, Discourses, 4, 6, 38; AD 1-2)

In New Testament, with non-active ditransitive verbs, (a) DPs in the genitive/dative case change into nominative, and (b) the retention of the accusative case of the second object is possible: with dative case in transitive structure: ȱ π·Ύ΅ΏΓІΐ΅΍ȱǻπ·Ύ΅ΏЗǼȱegkaloumai (egkalo:) ‘to bring a charge/accusation’ ΐ΅ΕΘΙΕΓІΐ΅΍ȱmartyroumai ‘to ascribe to somebody’ȱ ΛΕ΋ΐ΅ΘϟΊΓΐ΅΍ȱkhre:matizomai ‘to negotiate/transact business to one's own profit/make-money’ȱ Έ΍΅ΎΓΑΓІΐ΅΍ȱdiakonoumai ‘to minister to one's own needs/serve oneself’ȱ (38) ̓ΉΕϠȱ ΔΣΑΘΝΑȱ ЙΑȱ π·Ύ΅ΏΓІΐ΅΍ȱ ЀΔϲȱ ͑ΓΙΈ΅ϟΝΑȱ peri panto:n ho:n egkaloumai hypo Ioudaio:n about all.GEN which.GEN accuse.NACT.PRES.1SG by Jews.GEN ‘concerning all the things whereof I am accused by the Jews’ (New Testament, Acts, 26, 2, 2; AD 1) with genitive case in transitive structure: Ύ΅Θ΋·ΓΕΓІΐ΅΍ kate:goroumai ‘to accuse somebody of something’  Ύ΅Θ΅·΍ΑЏΗΎΓΐ΅΍katagino:skomai ‘to know/ recognize’ ΔΏ΋ΕΓІΐ΅΍ple:roumai ‘to fulfil’ 

(39) a. Θϟȱ Ύ΅Θ΋·ΓΕΉϧΘ΅΍ȱ ЀΔϲȱ ΘЗΑȱ ͑ΓΙΈ΅ϟΝΑȱ ti kate:goreitai hypo to:n Ioudaio:n what.ACC accuse.NACT.PRES.3SG by the Jews ‘about why he was accused by the Jews’ (New Testament, Acts, 22, 30, 2; AD 1)

108

Chapter Three b. ΔκΖȱΈȂȱȱȱϳȱȱȱȱΐΉΗΣΊΝΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘϱΔΓΖȱȱȱȱȱȱȱȱЀΔϲȱȱȱΘϛΖȱȱΔ΅Ώ΍ΕΕΓϟ΅ΖȱΦΚΕΓІȱΘΉȱȱΔΏ΋ΕΓІΘ΅΍ȱ pas d’ ho mesazo:n topos hypo te:s palirroias aphrou te ple:routai all prt the be-in-the-middle.NOM place.NOM by

the backwater

ȱȱfoam ȱȱand fill.NACT.PRES.3SG

‘all the place which was in between fills with backwater foam’ (Diodorus Siculus, Library, 1, 32, 9; 1 BC) ditransitive verbs: the structure non-active type + DP in the accusative is possible ΔΑΉІΐ΅ȱȱ πΔΓΘϟΗΌ΋ΐΉΑ (40) a. ςΑȱȱ hen pneuma epotisthe:men   one.ACC Spirit.ACC drink.PASS.AOR.1PL ‘and were all given to drink into one Spirit’ (New Testament, 1 Corinthians, 12, 13; AD 1)  b. πΑΉΈ΍ΈϾΗΎΉΘΓȱȱ ΔΓΕΚϾΕ΅Αȱ enedidysketo porphyran ȱȱ clothe.NACT.IMPERF.3SG purple.ACC ‘he was clothed in purple’ (New Testament, Luke, 16, 19; AD 1)

As for non-active types that, as transitives, take two accusatives, the DP denoting the theme remains in the accusative case in the structure with the non-active type as well. (41) a.ȱ ΦΔ΅΍Θ΋ΌφΗΓΑΘ΅΍ȱȱ Θϲȱȱ πΔ΍·νΑ΋ΐ΅ apaite:the:sontai to epigene:ma   demand.PASS.FUT.3PL the.ACC profit.ACC ‘they will demand for the profit’ (Tebtunis Papyri 61b 359 = 72, 355) b.ȱ Θϲȱȱ ϡΐΣΘ΍ϱΑȱȱ ΐΓΙȱȱ ϶ȱȱ ΔΉΕ΍ΉΆΉΆΏφΐ΋Α to himation mou ho periebeble:me:n the.ACC clothes.ACC my which.ACC wear.NACT.PLUP.1SG ‘the clothes which I have worn’ (Magdalen Papyrus 6, 6, 221a)



Consequently, the assignment of accusative case to the DP that accompanies verbs in non-active voice is possible in the following instances: (i) The theme argument is marked with accusative case in the passive structure of ditransitive verbs. (ii) The object is marked with accusative case in the passive structure of ditransitive verbs (the noun which is marked with the dative case in the transitive structure can in the New Testament also become the subject of the passive type). (b) Loss of the benefactive meaning From the Hellenistic period, the marking of the benefactive (‘personal interest’) meaning by non-active morphology begins to be abandoned. The active and non-active transitive type are used, the one next to the other, without an essential difference in meaning, even in the same sentence. The feature [+benefactive meaning] (subject participation or interest of the subject for the verb action) that was attributed by the non-active voice in Classical Greek in these cases was lost: (42) a. ΐΣΛ΅΍Ε΅Αȱȱ ΗΔ΅ΗΣΐΉΑΓΖȱȱ ȮȱσΗΔ΅Η΅Αȱȱ ΐ΅Λ΅ϟΕ΅Ζ makhairan spasamenos – espasan makhairas large-knife.ACC break.NACT.AOR.PART – break.ACT.AOR.3PL large-knife.ACC (without any meaning difference)32 (Ptolemaic papyri, Ȃayser 1929: 111) ΘϲΐȱΐϾΕ΍ΓΑȱȱ ȮȱΘ΍ΐЏΐΉΌ΅ȱȱ ΘϲΑȱΐϾΕ΍ΓΑ b. Θ΍ΐЗΐΉΑȱȱ timo:men tom myrion – timo:metha ton myrion estimate.ACT.PRES.1PL the numberless.ACC – estimate.NACT.PRES.1PL the numberless.ACC (in both cases it means ‘calculate/estimate’) (Ȃayser 1929: 112)

In many instances, active voice instead of non-active was used, but also vice versa. These changes comprise evidence of changes in the voice system Ƞf the Hellenistic period, resulting in instability in the voice system (as we can see from the tendencies observed in the ongoing changes). The active transitive type (instead of the classical Ancient Greek non-active transitive) is used in many instances where the subject participates in the verb action: (43) a. Μ΋ΚϟΊΉ΍ȱȱ 32

ΘχΑȱȱ

Έ΅ΔΣΑ΋Α

Similar instances of active and non-active voice alternation without semantic difference are to be found in the following: Par. 23, 7 (165a)/ 12, 15 (157a)/ Teb. I 138 (IIa): non-active, Par. 40, 40 (156a): active.

Diachronic Data and Analysis

109

pse:phizei te:n dapane:n   count.ACT.PRES.3SG the.ACC cost.ACC ‘he counts the cost’ (New Testament, Luke, 14, 28; AD 1) with the meaning of the Classical Greek transitive ȥȘijȓȗȠȝĮȚ pse:phizomai

Grammarians (mainly Atticists) comment on the choice of voice and often give arbitrary interpretations about the meaning of the active and the non-active form: (44) a.ȱ Θϱȱȱȱȱȱȱȱȱȱȱ·ΉȱȱȱȱΐχΑȱȱȱȱȱΎ΅ΌνΊΉΗΌ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱΘΓІȱȱȱȱȱȱȱȱȱΎ΅ΌϟΊΉ΍ΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ϠȱȱΘϲȱȱΎΣΌ΍ΗΓΑȱ to ge me:n kathezesthai tou kathizein kai to kathison the.ACC prt

not

sit.NACT.PRES.INF the.GEN sit.ACT.PRES.INF and the sit.ACT.AOR.IMP.2SG

ΘΓІȱȱȱȱȱȱȱȱȱȱΎΣΌ΋ΗΓȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΫΕȂȱȱȱȱȱΓϨΗΌȂȱȱȱȱȱȱȱȱȱȱȱϵΘ΍ȱȱȱȱȱΈ΍ΉΑφΑΓΛΉDz

tou kathe:so ar’ oisth’ hoti diene:nokhe? the.GEN sit.NACT.PERF.IMP.2SG prt know.2SG that differ.PERF.3SG ‘you do not know that ‘kathezesthai’ differs from ‘kathizein’ and ‘kathison’ from ‘kathe:so’?’ (Lucianus, Soloecista, 11; AD 2) b. Ύ΅ϠȱΘϲȱȱȱȱȱȱȱȱȱΎ΅Θ΅ΈΓΙΏΓІΑȱȱȱȱȱȱȱȱȱȱΗϿȱȱȱȱȱȱȱȱȱȱΐξΑȱȱȱϥΗΝΖȱȱȱȱΘ΅ЁΘϲΑȱȱȱȱȱȱȱȱΘХȱȱȱȱȱȱȱȱΎ΅Θ΅ΈΓΙΏΓІΗΌ΅΍ȱ kai to katadouloun sy men iso:s tauton to:i katadoulousthai ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ and the.ACC enslave.ACT.PRES.INF you.NOM prt maybe the-same.ACC the.DAT enslave.NACT.PRES.INF ΑΉΑϱΐ΍Ύ΅Ζǰȱȱȱȱȱȱȱȱȱȱπ·АȱȱȱȱȱȱΈξȱȱȱȱȱȱΓϨΈ΅ȱȱȱȱȱȱȱȱȱȱȱΈ΍΅ΚΓΕΤΑȱȱȱȱȱȱȱȱȱΓЁΎȱȱϴΏϟ·΋ΑȱȱȱȱȱȱȱσΛΓΑ ȱ

nenomikas, ego: de oida diaphoran ouk olige:n ekhon think.PERF.2SG I.NOM prt know.1SG difference.ACC not little.ACC have.PART ‘and you have thought that ‘katadouloun’ maybe is the same as ‘katadoulousthai’; but I know that there is a large difference between them’ (Lucianus, Soloecista, 12; AD 2) c.ȱ ΅ϢΘЗȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅Ϡȱȱ΅ϢΘΓІΐ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΈ΍΅ΚξΕΉ΍ȉȱΘϲȱΐξΑȱ·ΤΕȱ΅ϢΘЗȱȱȱȱȱȱȱȱȱȱȱȱȱȱπΔϠȱȱΘΓІȱΧΔ΅ΒȱȱΘ΍ȱȱ aito: kai aitoumai diapherei; to men gar aito: epi tou hapaks ti ask-for.ACT.PRES.1SGand ask-for.NACT.PRES.1SG differ.3SG the first ask-for.ACT.PRES.1SG for the once something

ȱ

ȱ

Ώ΅ΆΉϧΑȱȱȱȱȱȱȱȱΎ΅ϠȱΐχȱȱΦΔΓΈΓІΑ΅΍ǰȱȱȱȱȱȱΘϲȱȱΈȂȱȱȱȱȱ΅ϢΘΓІΐ΅΍ȱȱȱȱȱȱȱȱȱȱȱπΔϠȱΘΓІȱΛΕφΗ΅ΗΌ΅΍ȱȱΉϢΖȱΦΔϱΈΓΗ΍Α

labein

kai me: apodounai,

to d’

aitoumai

epi tou khre:sasthai eis apodosin

take.ACT.PRES.INF and not return.ACT.AOR.INF the other ask-for.NACT.PRES.1SG for the use.INF to return.ACC

‘‘aito:’ and ‘aitoumai’ are different; as the first one means that I ask for something in order to take it once and not to give it back, the other I ask for something to use it and return it’ (Ammonius, 7; AD 5) d.ȱ ΔΏΉϧΗΘ΅ȱȱȱȱȱΈξȱȱȱȱȱΘЗΑȱȱȱȱȱȱȱȱȱȱΔ΅Ό΋Θ΍ΎЗΑȱȱȱΎΓ΍ΑЗΖȱȱȱȱȱΏ΅ΐΆΣΑΉΘ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱπΑνΕ·Ή΍΅Αȱ pleista de to:n pathe:tiko:n koino:s lambanetai energeian many.NOM prt the.GEN passive.GEN similarly take.NACT.PRES.3SG action.ACC Ύ΅Ϡȱȱ ΔΣΌΓΖȱȱ

ȱ

ȱ

Η΋ΐ΅ϟΑΓΑΘ΅ȱȱ

σΌΉ΍ȱ

̝ΘΘ΍ΎХȱȱȱȱȱȱȱȱ

ȱ

kai pathos se:mainontaȱ ethei Attiko:i and state.ACC mean.PART.ACC habit.DAT Attic.DAT ‘many from the passives (non-active voice morphology) are considered to mean both action and state according to the Attic practice’ (Bachmann, Anecd. ii 300, 8)

In the following extract, Phrynichus (second half of the 2nd century AD), in his common way, disapproves of the use of a verb that does not accord with that which can be witnessed in classical texts: (45) ΦΔΉΏΉϾΗΓΐ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱΔ΅ΑΘΣΔ΅Η΍ȱΚΙΏΣΘΘΓΙǰȱȱȱΓЄΘΉȱȱȱ·ΣΕȱȱΓϡȱȱȱȱȱȱȱȱȱΈϱΎ΍ΐΓ΍ȱȱȱȱȱȱȱȱȱϹφΘΓΕΉΖǰȱȱ apeleusomai pantapasi phylattou, oute gar hoi dokimoi rhe:tores, depart.MID.FUT.1SG always avoid.IMP.2SG neither as the.NOM notable.NOM rhetorician.NOM ΓЄΘΉȱȱȱψȱȱȱȱȱȱȱȱȱȱȱΦΕΛ΅ϟ΅ȱȱȱȱȱȱȱΎΝΐУΈϟ΅ǰȱȱȱȱȱȱȱΓЄΘΉȱȱȱ̓ΏΣΘΝΑȱȱΎνΛΕ΋Θ΅΍ȱȱȱȱȱȱȱȱΘϜȱȱȱȱȱȱȱȱȱȱȱȱΚΝΑϜǰ

oute

he:

arkhaia

ko:mo:idia,

oute Plato:n

kekhre:tai

te:i

pho:ne:i,

neither the.NOM ancient.NOM comedy.NOM neither Plato.NOM use.PERF.3SG the.DAT voice.DATȱȱȱȱȱȱȱ

ȱ ΦΑΘȂȱȱȱȱΈȂȱȱȱ΅ЁΘΓІȱȱȱȱΘХȱȱȱȱȱȱȱΩΔΉ΍ΐ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΛΕЗȱȱȱȱȱȱȱȱȱȱȱΎ΅ϟȱȱΘΓϧΖȱȱȱȱȱϳΐΓΉ΍ΈνΗ΍ΑȱȱȱБΗ΅ϾΘΝΖǯ

ant’

d’ autou

to:i

apeimi

khro:

kai tois

homoeidesin ho:sauto:s.

instead prt this.GEN the.DAT depart.ACT.FUT.1SG use.IMP.2SG and the.DAT similar.DAT

in-like-manner

‘avoid always ‘apeleusomai’, as neither the notable rhetoricians nor the ancient comedy nor Plato have used this voice; instead of that use ‘apeimi’ and the similar verbs to that alike’ (Phrynichus, Ǽcloga, 24; AD 2)ȱȱȱ

The reason for the loss of the distinction and the special meaning of the non-active transitive structure can be sought in the following: (a) Both voices (active and non-active) could be used in transitive structures with agent as subject, with the result in many instances of them being in alternation already in Classical Greek without any particular obvious difference in meaning.

110

Chapter Three (46) ΘΓϾΘΝΑȱȱ touto:n these.GEN ΅ϡȱȱ

Α΅ΙΘ΍ΎϲΑȱȱ

Δ΅ΕΉϟΛΓΑΘΓȱȱ

nautikon ships.ACC

pareikhonto Korinthioi, furnish.NACT.IMPERF.3PL Corinthians.NOM

ΈȂȱȱ

ΩΏΏ΅΍ȱȱ ΔϱΏΉ΍Ζȱȱ

̍ΓΕϟΑΌ΍Γ΍ǰ

ΔΉΊϱΑȱȱ

Δ΅ΕΉϧΛΓΑǯȱȱ

hai d’ allai poleis pezon pareikhon. the.NOM prt others cities.NOM infantry.ACC furnish.ACT.IMPERF.3PL ‘of these, the Corinthians furnished ships, the other cities furnished infantry’ (Thucydides, The Peloponnesian War, 2, 9, 3; 5 BC)

(b) The different meanings of the non-active (middle) voice were difficult to unify in only one meaning, particularly as the non-active (middle) voice also marked the passivisation (the transparent meaning of passivisation competed with the other meanings). Lexical substitution (in certain tenses) (cf. 47a) and the deponent verbs with non-active voice and with direct object in the accusative (when, in fact, synonyms with active voice appeared in parallel to them; Lavidas & Papangeli 2006) rendered the voice system more opaque (cf. 47b). < (47) a.ȱ ΔϟΓΐ΅΍ piomai < drink.NACT.PRES.1SG – b. ΆΓϾΏΓΐ΅΍ boulomai – will/wish.NACT.PRES.1SG

ʌϟȞȦ pino: drink.ACT.PRES.1SG πΌξΏΝȱȱȱ

ethelo: will/wish.ACT.PRES.1SG

Of interest in this period is the denotation of the person who is interested in the verb action with a pronoun in the dative and not with a non-active transitive verb: (48)ȱ

ΘχΑȱȱ Ϲ΍ΑΓΔϾΏ΋Αȱȱ οΘΓϟΐΝΖȱȱ ΦΑΓϟ·Ή΍Αȱȱ ΅ЁΘХȱ te:n rhinopyle:n hetoimo:s anoigein auto:i the.ACC side-gate.ACC immediately open.ACT.PRES.INF he.DAT ‘to open immediately the side gate for him’ (Polybius, Histories, 8, 25, 8; 3-2 BC)

The use of clitics for the denotation of the benefactive, instead of the use of non-active transitive structures, can be connected to reanalysis, in this period, of the clitics in the genitive case, from pronouns governed by a noun (possessives) to pronouns that are governed by a verb (Horrocks 1997: 59-60). (49)

Ώ΅ΐΆΣΑΉ΍Ζȱȱ

ΐΓΙȱȱ

lambaneis mou take.ACT.PRES.2SG I.GEN ‘you receive letters from me’

Θ΅ȱȱ

·ΕΣΐΐ΅Θ΅ȱȱ

ta the.ACC

grammata letters.ACC

The presence of the pronoun in the second position (resulting in adjacency with the verb) can lead to an interpretation of the pronoun analogous with that of the ethic dative which also appeared in the same position. During the post-classical period (the era of the gradual loss of the middle transitive structure (benefactive meaning)), the combination of non-active voice and personal pronouns for the denotation of the benefactive is highly systematic: (50)

ΦΑ΅ΌЏΐΉΌ΅ȱȱ

ο΅ΙΘΓϿΖȱȱ

ΘХȱȱ

͑΋ΗΓІ

anatho:metha heautous to:i Ie:souȱȱȱȱ entrust.MID.AOR.1PL ourselves.ACC the.DAT Jesus.DAT ‘we entrust ourselves to Jesus’ (Martyr. Matthaei, 174, 11)

The fact that the marking of the benefactive (of the subject who participates in the verb action; or, according to the traditional terminology, of the indirect reflexivity) by the non-active morphology was becoming restricted (and the replacement of this structure by the structure “active verb + personal pronoun that denotes the benefactive”) results in the more frequent marking of the (direct) reflexivity with a reflexive pronoun: (51) a. ο΅ΙΘϲΑȱȱ heauton myself.ACC

πΔ΍Έ΍ΈϱΑ΅΍

epididonai give-freely.ACT.PRES.INF

Diachronic Data and Analysis

111

‘to give myself freely’ (Ȃayser 1929: 104) b. πΘ΅ΔΉϟΑΝΗΉΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱο΅ΙΘϲΑȱȱȱȱȱȱȱȱȱȱ·ΉΑϱΐΉΑΓΖȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱЀΔφΎΓΓΖȱȱȱȱȱȱȱȱΐνΛΕ΍ȱȱΌ΅ΑΣΘΓΙ etapeino:sen heauton genomenos hype:koos mekhri thanatou make-lowly.ACT.AOR.3SG himself.ACC become.PART.NOM subject.NOM until death ‘he made himself lowly as he became a subject until his death’ (New Testament, Philippians, 2, 73; AD 1)

The use of reflexive pronouns with non-active verbs should, however, be considered evidence of the confusion in the use of voice morphology. The reflexive pronouns are already being used with nonactive verbs with reflexive interpretation in Classical Greek, but not systematically and productively: (52)ȱ Γϡȱȱ ΐνΑȱȱ Κ΅Η΍ȱȱ hoi men phasi the.NOM one.NOM report.3PL πΔ΍ΗΚΣΒ΅΍ȱȱ

Ά΅Η΍Ών΅ȱȱ

ΎΉΏΉІΗ΅΍ȱȱ

Θ΍Α΅ȱȱ

basilea king.ACC

keleusai order.INF

tina ȱ someone.ACC

ΈȂȱȱ

ο΅ΙΘϲΑȱȱ

΅ЁΘϲΑȱȱ̍ϾΕУǰȱȱ

Γϡȱȱ

πΔ΍ΗΚΣΒ΅ΗΌ΅΍ȱ

episphaksai auton Kyro:i, hoi d’ heauton episphaksasthai slay.ACT.AOR.INF he.ACC Cyrus.DAT the.NOM other.NOM himself.ACC slay.MID.AOR.INF ‘and one report is that the king ordered someone to slay him upon the body of Cyrus, while others say that he slew himself with his own hand’ (Xenophon, Anabasis, 1, 8, 29; 5-4 BC)

In Hellenistic Koine, the reflexive pronoun is used very frequently and fairly systematically with reflexive non-active verbs, which constitutes evidence that the non-active ending did not also selfevidently denote reflexivity: (53) a. ΏφΜΓΐ΅΍ȱȱ πΐ΅ΙΘХȱȱ ЀΐκΖȱȱ Ώ΅ϲΑȱȱ πΐΓϟ  le:psomai emauto:i hymas laon emoi  take.MID.FUT.1SG myself.DAT you.ACC people.ACC I.DAT ‘I will take you with me as you are my own people’ (Septuagint, Exodus, 6, 7; 2-1 BC) b.ȱ ЀΔΓΏ΍ΔϱΐΉΑΓΖȱȱ Η΅ЁΘХ hypolipomenos sauto:i leave.MID.AOR.NOM yourself.DAT ‘as you were left to yourself’ (Ȃayser 1929: 104)

(c) Active versus non-active future At the same time, non-active (mainly future tense) types of verbs that formed the present tense with active endings are replaced, in most instances, by active forms: ΦΎΓϾΗΝakouso: hear.ACT.FUT [instead of ΦțȠϾıȠȝĮȚ akousomai hear.NACT.FUT] Υΐ΅ΕΘφΗΝ hamarte:so: fail.ACT.FUT (New Testament, Matthew, 18, 21; AD 1) ΦΔ΅ΑΘφΗΝ apante:so: meet.ACT.FUT (New Testament, Mark, 14, 13; AD 1)ȱ ΥΕΔΣΗΝ harpaso: snatch-away.ACT.FUT (New Testament, John, 10, 28; AD 1) ΆΏνΜΝblepso: see.ACT.FUT (New Testament, Matthew, 13, 14; AD 1) Έ΍ЏΒΝdio:kso:pursue.ACT.FUT (New Testament, Matthew, 23, 34; AD 1) πΐΔ΅ϟΒΝempaikso: mock.ACT.FUT (New Testament, Mark, 10, 34; AD 1) ϹΉϾΗΝrheuso:flow.ACT.FUT (New Testament, John, 7, 38; AD 1)

(54) a. Ύ΅Ϡȱ ΦΔ΅ΑΘφΗΉ΍ȱ ЀΐϧΑȱ ΩΑΌΕΝΔΓΖȱΎΉΕΣΐ΍ΓΑȱ ЂΈ΅ΘΓΖȱ Ά΅ΗΘΣΊΝΑȱ   kai ȱ apante:sei hymin anthro:pos keramion hydatos bastazo:n and ȱȱmeet.ACT.FUT.3SG you.DAT man.NOM pitcher.ACC water.GEN carry.PART.NOM ‘and (there) a man carrying a pitcher of water will meet you’ (New Testament, Mark, 14, 13, 2; AD 1) b. ΦΎΓϾΗΉΘΉȱ Ύ΅ϠȱΓЁȱΐχȱ ΗΙΑϛΘΉǰȱ Ύ΅ϠȱȱȱΆΏνΔΓΑΘΉΖȱΆΏνΜΉΘΉȱȱȱȱȱȱȱȱΎ΅ϠȱΓЁȱΐχȱϥΈ΋ΘΉȱ akousete kai ou me: syne:te, kai blepontes blepsete ȱȱȱȱkai ou me: ide:te hear.ACT.FUT.2PL and no not understand.2PL and see.PART see.ACT.FUT.2PL and no not see.2PL

‘you will hear, and will in no way understand. In seeing, you will see, and will in no way perceive’ (New Testament, Acts, 28, 26; AD 1)

Atticism, of course, reintroduces the Classical Attic non-active future to such an extent that even unattested non-active types are preferred in a frenzy of imitation instead of the actives since Atticists considered the use of non-actives instead of actives as an element of the Attic dialect. The replacement of non-active with active future types in the Roman papyri is not as frequent as it is in the New Testament and Hellenistic Koine but is more frequent than in the Ptolemaic papyri (there is only one instance there: πΎΚΉϾΒΉ΍Αȱekpheukseinȱ‘to flee out or away/escape’ –Ȃayser 1929/1970: i2, 2, 130). In the Roman papyri, we come across the following (among other) active future types

112

Chapter Three

(Gignac 1981): (55) a. ρΝΖȱȱ ΦΎΓϾΗΝȱȱȱȱȱȱȱ heo:s akouso: until hear.ACT.FUT.1SG (Pȅxy. (The Oxyrhynchus Papyri) 294.15-16; 22 AD) b. ЀΔ΅ΎΓϾΗΉ΍Αȱȱȱȱȱȱȱȱȱȱȱ hypakousein   hearken/obey.ACT.FUT.INF (BGU (Aegyptische Urkunden aus den Staatlichen Museen zu Berlin, Griechische Urkunde) 747; 139 AD / Pȅxy. (The Oxyrhynchus Papyri) 1667.11) c. ΦΔ΅ΑΘφΗΓȱȱȱȱȱȱȱȱȱȱȱ    apante:so meet.ACT.FUT.1SG ǻPGen. (Les Papyrus de Genève) 56; 346 AD)  d. ΦΑ΅ΔΏΉϾΗΝȱȱȱȱȱȱȱȱȱ   anapleuso:  sail-upwards.ACT.FUT.1SG ǻBGU (Aegyptische Urkunden aus den Staatlichen Museen zu Berlin, Griechische Urkunde) 601.17; AD 2 / SB (Sammelbuch griechischer Urkunden aus Aegypten) 4317.8; circa 200 AD)

Concerning the non-active I (middle) and non-active II (passive) endings, which are only distinguished in the past and the future, free alternation is observed during the Hellenistic era, something which had already begun during the Classical era. The new element for this period is the extension of non-active II (passive) endings. The following deponent and intransitive verbs begin to form non-active II/passive future and past instead of non-active I/middle in the Roman papyri (Chatzidakis 1892/1975: 193-200)33: (56) a. ϐΗΌΣΑΌ΋Α e:isthanthe:n feel.PASS.AOR.1SG (PȂich. (Michigan Papyri) 486.7; AD 2 b. πΏΙΔφΌ΋elype:the: be-grieved.PASS.AOR.3SG (PMich. (Michigan Papyri) 497.15; AD 2  c. ΐΉΏ΋ΌϜΖmele:the:is care.PASS.AOR.2SG (PMich. (Michigan Papyri) 466.35; AD 2

3.2.3.3 Change of non-active into active anticausative The majority of anticausatives in the Hellenistic-Roman period are active. The active voice is extended to anticausatives that participate in transitivity alternations and is not connected with the production of new alternations. At the same time, as an indication of change in progress, in this period, anticausative types with active (corresponding to the new type) and non-active voice morphology (corresponding to the type of the previous period) without difference in meaning are in evidence. (a) Expansion of the active anticausative type Particularly frequent in this era is the change in voice morphology of anticausative verbs that, from being non-active, develop into active. This is a change in voice morphology, i.e. in the marking of the anticausative verbs, even though grammarians (Kühner & Gerth [1898-1904]/1963; Krüger, Blaß & Debrunner 1961; Brugmann & Thumb 1913) consider that it was concerned with the formation of a new intransitive from a transitive verb, not taking into consideration that there already existed in previous periods a corresponding anticausative verb with non-active morphology, and that the intransitive use of these verbs is not innovative34. 33

The non-active I/middle past is also used with other verbs (for example, with ΦΕΑνΓΐ΅΍ ‘to deny’); something that again indicates that there is no clear separation in the field of use of these two endings:

ȱ

34

(1)ΉϢȱΦΕΑ΋ΗϱΐΉΌ΅ǰȱΎΦΎΉϧΑΓΖȱΦΕΑφΗΉΘ΅΍ȱψΐκΖȱ ei arne:sometha, ȱkakeinos ȱarne:setai ȱȱȱhe:mas if deny.NACT.FUT.1PL and-he.NOM deny.NACT.FUT.3SG we.ACC ‘if we deny him, he also will deny us’ (New Testament, 2 Timothy, 2, 12; AD 1) 



Similarly, the observations and the following examples from Moulton & Turner (1963) do not so much concern changes in causative and anticausative verbs as intransitive verbs with an agent-DP as subject and active morphology (unergative intransitives that change from non-active into active). Again, it is not about a shift of transitive into intransitive verbs but about a clear change in voice morphology since the corresponding intransitive verbs were in evidence (with non-active morphology) in earlier periods. The majority of these verbs denote movement; there are also, however, verbs of change-of-state (anticausatives) among the verbs referred to: Δ΅ΕΣ·Νȱparago:ȱ‘to pass by’ (New Testament, Matthew 9, 27; 20, 30, New Testament, Mark 15, 21; AD 1)ȱ ΔΕΓΣ·Νproago:‘to lead the way/go before’ (New Testament, John 2, 9; AD 1)) ΔΉΕ΍Σ·Νperiago: ‘to come round’ (New Testament, Matthew 4, 23; ǹ 13, 11; AD 1)

Diachronic Data and Analysis

113

Mayser (1934/1970) refers to the following verbs from the Ptolemaic period as instances of innovative intransitive use of earlier transitive verbs. He talks, however, about verbs that in earlier times already had intransitive use but with non-active voice. This erroneous ascertainment of Mayser (and other grammarians for other verbs) about the innovative intransitive structure is a result of the phonological merger concerning the voice of intransitive (earlier non-active) and transitive structure from the moment that the intransitive structure is also marked with active voice. ΦΔ΅ΏΏΣΗΗΝȱapallasso: ‘to get off free/escape’ȱ

ȱ

ȱ ȱ



ȱ ΦΏϾΔΝΖȱȱ

ΦΔ΅ΏΏΣΗΗΉ΍Ζȱȱȱȱȱȱ

alypo:s without-pain

apallasseis get-off-free.ACT.PRES.2SG (The Petrie Papyri II 2 (3) 1)

 Έ΍΅ΆΣΏΏΝȱdiaballo:ȱ‘to pass over/cross’

ȱ ȱȱ

ȱ ωΑ΅·ΎΣΗΌ΅΍ȱȱ

Έ΍΅Ά΅ΏΉϧΑ

e:nagkasthai force.NACT.PERF.INF

diabalein pass-over.INF (Tebtunis Papyri I, 23, 3)

ΦΏΏΣΗΗΝȱallasso:ȱ‘to change’ πΔ΍ΆΣΏΏΝȱepiballo: ‘to go straight towards/lie upon’ πΎΘΓΔϟΊΝȱektopizo:ȱ‘to take oneself from a place/go abroad’ȱ ΦΑ΅ΊΉΙ·ΑϾΝȱanazeugnyo:ȱ‘to yoke or harness again/withdraw’ȱ Ύ΍ΑЗȱkino:ȱ‘to move forward’ ȱ

Quite a few anticausative structures from the Hellenistic Koine begin to be denoted with active morphology in contrast to Classical Greek. The transitivity alternation (causative and anticausative type) already existed in Classical Greek, with the difference that the anticausative type was non-active (agent/cause + ΦΑΓϟ·Ν anoigo: open.ACT.PRES.1SG + patient, patient + ΦΑΓϟ·ΉΘ΅΍ anoigetai open.NACT.PRES.3SG).The change concerns only the morphology of the anticausative (patient + ΦȞȠȓȖİȚ anoigei open.ACT.PRES.3SG) without other syntactic changes in the existing alternation. The result is the use of the same voice for both the transitive and the intransitive type (labile transitivity alternation). (57) active transitive ΅ΑΙ·Ν΍ȱȱ ΘΓϿΖȱȱ ϴΚΌ΅ΏΐΓϾΖȱȱ ΐΓΙ a. πΒ΅ϟǽΚΑ΋ΖǾȱȱȱȱΦΑΓϟ·Νȱȱ eksai[phne:s] anoigo: anygo:i tous ophthalmous mouȱȱ suddenly open.ACT.PRES.1SG open.ACT.PRES.1SG the.ACC eyes.ACC my ‘and suddenly I open my eyes’ (UPZ (Urkunden der Ptolemäerzeit (ältere Funde)) 1, 78 rp 1) active intransitive b. ΓЂΘΝΖȱȱ ΓЁΎȱȱ ΦΑΓϟ·Ή΍ȱȱ Θϲȱȱ ΗΘϱΐ΅ȱȱ ΅ЁΘΓІ outo:s ouk anoigei to stoma autou so not open.ACT.PRES.3SG the.NOM mouth.NOM his ‘so his mouth does not open’ (New Testament, Acts, 8, 32, 6; AD 1)

In the following example containing an orthographic alternation a significant factor is revealed that contributed to the change: the homophony of the endings -ΉΘ΅΍ /-etai/ (3rd person, singular number, non-active voice) and -ΉΘΉ/-ete/ (2nd person, plural number, active voice). The phonological merger of these two endings is the result of the phonological change that happened in the period of HellenisticRoman Koine and resulted in the pronunciation of the diphthong /ai/ as /e/. This homophony (when there was no overt expression of the subject) may have led, in instances such as that given in the example, to the change of a non-active into an active anticausative, and in other instances (not with this specific verb, as it was already an alternating verb) to causativisation of an intransitive non-alternating verb.

πΔ΅ΑΣ·Νepanago: ‘to withdraw/retreat’ (New Testament, Matthew 21, 18; AD 1) πΔ΍ΆΣΏΏΝ epiballo: ‘to lie upon/be put upon’ (New Testament, Luke 15, 21; AD 1) ΦΔνΛΝapekho: ‘to be away or far from’ (New Testament, Luke 15, 20; AD 1) ΎΏϟΑΝklino: ‘to incline towards/come to end’ (New Testament, Luke 9, 12; 24, 29; AD 1) ΅ЁΒΣΑΝȱauksano:ȱ‘to grow/increase’ȱ ΚϾΝȱphyo:ȱ‘to grow up/appear’

114

Chapter Three (58)ȱ ΎΕΓϾΉ΍ȱȱ ΌϾΕ΅Αȱȱ Ύ΅Ϡȱȱ ΦΑΓϟ·ΉΘ΅΍ȱȱ ΅ΑΙ·ΉΘΉȱȱ krouei thuran kai anoigetai anugete knock.ACT.PRES.3SG door.ACC and open.NACT.PRES.3SG open.ACT.PRES.2PL ‘he knocks at the door and it opens’ (UPZ (Urkunden der Ptolemäerzeit (ältere Funde)) 1, 79 rp, 7)

Whereas active morphology for intransitive structures has already begun to be used, at the same time the non-active form also continues to be used. It is mainly used, of course, for passive structures with presence of an agent or with an implied agent, and not with a cause. (59) a.ȱ ΎΕΓϾΉΘΉǰȱȱ Ύ΅Ϡȱȱ ΦΑΓ΍·φΗΉΘ΅΍ȱȱ ЀΐϧΑȱȱ ȱȱȱȱȱȱ krouete, kai anoige:setai hymin knock.IMP.2PL and open.MID.FUT.3SG you.DAT ‘knock and it will be opened for you’ (New Testament, Matthew, 7, 7, 1; AD 1) b.ȱ ΉЁΌνΝΖȱȱ ΦΑΓϟ·ΉΗΌ΅΍ȱȱ ΘχΑȱȱ Ϲ΍ΑΓΔϾΏ΋Αȱȱ ΅ЁΘХȱ euthe:o:s anoigesthai te:n rhinopyle:n auto:i immediately open.NACT.PRES.INF the.ACC side-gate.ACC he.DAT ‘so that the side gate will be opened for him’ (Polybius, Histories, 8, 25, 10; 3-2 BC)

ȱ The verb ĮȣȟȐȞȦ auksano: ‘to grow’ had both active transitive and non-active intransitive types during the Classical Greek period. In the Hellenistic period, the anticausative structure bears active voice morphology. (60) active transitive ·ΤΕȱ ΘϲΑȱ ΚΌϱΑΓΑȱ a. ΅ЄΒΉ΍ȱ   auksei gar ton phthonon grow.ACT.PRES.3SG prt the.ACC envy.ACC ‘that grows the envy’ (Plutarch, On the Control of Anger, 462c, 1; AD 1-2) active intransitive b. Ύ΅Θ΅ΐΣΌΉΘΉȱȱ ΘΤȱȱ ΎΕϟΑ΅ȱȱ ΘΓІȱȱ Φ·ΕΓІȱȱ ΔЗΖȱȱ΅ЁΒΣΑΓΙΗ΍Αȱȱ katamathete ta krina tou agrou po:s auksanousin consider.IMP.2PL the.NOM lilies.NOM the.GEN field.GEN how grow.ACT.PRES.3PL ‘consider the lilies of the field, how they grow’ (New Testament, Matthew, 6, 28, 2; AD 1) c.ȱ ΅ЁΒΣΑΉΘΉȱȱȱȱȱȱȱȱȱȱΈξȱπΑȱΛΣΕ΍Θ΍ȱΎ΅Ϡȱ·ΑЏΗΉ΍ȱΘΓІȱΎΙΕϟΓΙȱψΐЗΑȱȱȱȱΎ΅ϠȱΗΝΘϛΕΓΖȱд̌΋ΗΓІȱ̙Ε΍ΗΘΓІ auksanete de en khariti kai gno:sei tou kyriou he:mo:n kai so:te:ros Ie:sou Khristou grow.ACT.IMP.2PL prt in grace and knowledge the Lord.GEN our

and Savior

Jesus.GEN Christ.GEN

‘grow in the grace and knowledge of our Lord and Savior Jesus Christ’ (New Testament, 2 Peter, 3, 18, 1; AD 1)

The following example is indicative of the structures that can lead to a change in the intransitive type voice morphology. The accusative case of the subject of the infinitives is identified with the accusative case of the direct object of finite verb types. Consequently, in the cases of transitive infinitive types, the DP in the accusative case, which is the complement of the infinitive, can be analysed by native speakers as the object of a transitive active verb or as the subject of an intransitive active verb35. With these specific anticausative verbs, the second analysis (as subject of intransitive type) results in the innovative active voice morphology of the intransitive structures. (61)ȱ πΎΉϧΑΓΑȱȱ ΈΉϧȱȱ ΅ЁΒΣΑΉ΍Αǰȱȱ πΐξȱȱ Έξȱȱ πΏ΅ΘΘΓІΗΌ΅΍ȱȱȱȱȱ ekeinon dei auksanein, eme de elattousthai he.ACC must.3SG increase.ACT.PRES.INF I.ACC prt decrease.NACT.PRES.INF ‘He must increase, but I must decrease’ (New Testament, John, 3, 29-30; AD 1)

As can be seen from the following examples, the non-active forms had not been entirely abandoned. Active and non-active forms were used at the same time with the same structures (in the anticausative structures, the presence of only the cause and not the agent was possible) until the final dominance of only one of the two types. (62) non-active intransitive (as in Classical Greek) 35

Clearly this is concerned with an instance of opacity (a structure that can receive two interpretations). The issue is theoretically very important. Its interpretation (and the very important question of why changes of this type appeared beginning in the Hellenistic-Roman and not during the Ancient Greek period) comprises the object of chapter 4, which attempts to explain the overall grammatical change in all periods of Greek, as well as cross-linguistically.

Diachronic Data and Analysis

115

a.ȱ ΅ЄΒ΅ΗΌ΅΍ ΘχΑȱȱ ΎЏΐ΋Α auksasthai te:n ko:me:n grow.MID.AOR.INF the.ACC village.ACC ‘so that the village grows’ (BGU (Aegyptische Urkunden aus den Staatlichen Museen zu Berlin, Griechische Urkunde) 2, 389, 10) b.ȱ πΈϱΎΓΙΑȱȱ ΉϢȱȱ Ύ΅Ϡȱȱ ΔΉΑ΍ΛΕ΅ϧΖȱ πΏΔϟΗ΍Αȱȱ ΋ЁΒΣΑΓΑΘΓ edokoun ei ȱ kai penikhrais elpisin e:uksanonto ȱȱ ȱ supposed.3PL if ȱ and poor.DAT hopes.DAT grow.NACT.IMPERF.3PL ‘they supposed that they were growing even with poor hopes’ (P. Cair. Masp (Papyrus grecs d'époque byzantine, Catalogue général des antiquités égyptiennes du Musée du Caire) 3, 67295, pg 3, 15)

The verb éǴǴĝȑȑȁ allasso: ‘to change’ is especially characteristic of the confusion that was caused and the misinterpretations that exist in traditional grammatical descriptions of these verbs. The active éǴǴĝȑȑȁ allasso: ‘to change’ was considered by traditional historical grammarians (Jannaris 1897/1968; Schwyzer & Debrunner 1950/2002; Chantraine 1953; Humbert & Kourmoulis 1957/2002) as a new intransitive use of a transitive verb. A much closer look at the data shows that it is a case of change in voice morphology: the specific verb also existed earlier as an intransitive non-active, whereas from the Hellenistic era, it also consistently appears as an active intransitive. Initially, the intransitive non-active form (the earlier form) and the intransitive active form (the newer form) were to be found in free alternation, without aspectual or other semantic differences, but as synchronic variations (tendencies of change)36. (63) ΦΏΏΣΗΗΝ allasso: ‘to change’ intransitive active Θϲȱȱ ΗΛϛΐ΅ȱȱ ΦΏΏΣΗΗΉ΍ȱ a.ȱ Ύ΅Ϡȱȱ ΆΏνΔΉȱȱ kai blepe to skhe:ma allassei and see.IMP.2SG the.NOM form.NOM change.ACT.PRES.3SG ‘and see that the form is changing’ (Dorotheus, Greek fragments, 421, 2; 1 BC-AD 1) intransitive non-active ΐξΑȱȱ ΘχΑȱȱ ΔΓΕΚΙΕκΑȱ bǯȱ ωΏΏΣΛΌ΅΍ȱȱ e:llakhthai men te:n porphyran change.NACT.PERF.INF prt the.ACC cloths-of-purple.ACC ‘that the cloths of purple were changed’ (BGU (Aegyptische Urkunden aus den Staatlichen Museen zu Berlin, Griechische Urkunde) 4, 1141, 41) transitive active Θϲȱ Δ΅΍ΈΉϾΉΗΌ΅΍ǰȱȱΓЁΛȱ ϣΑȂȱ ΦΏΏΣΒΝΐΉΑȱ ΘΤΖȱ ЀΔΓΌνΗΉ΍Ζ c. πΔϠȱ epi to paideuesthai, oukh hin’ allakso:men tas hypotheseis upon the educate.INF not to change.ACT.AOR.1PL ȱ the.ACC constitutions.ACC ‘(we should enter) upon a course of education, not in order to change the constitution of things’ (Epictetus, Discourses, 1, 12, 17; AD 1-2) d.ȱ ϳΔϱΘΉΕΓΑȱȱȱȱȱȱȱȱȱȱȱȱȱΈξȱȱȱȱȱ΅ЁΘЗΑȱȱȱȱȱ ΓЁΎȱ σΒΉΗΘ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΦΏΏΣΒ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΘϲΑȱȱȱȱȱ ρΘΉΕΓΑ opoteron de auto:n ouk eksesti allaksai ton heteron which-of-two.NOM prt

these.GEN not

be-possible.3SG. change.ACT.AOR.INF the.ACC other.ACC

‘it is not possible of either of these two to change the other’ (P.Cair.Isid.(The Archive of Aurelius Isidorus in the Egyptian Museum, Cairo, and the University of Michigan), 80 rp 12)

It is worth noting the observation of Jannaris (1897/1968) that the transitive and intransitive uses of verbs with active morphology were much more frequent in the New Testament than they were in Classical Greek. He refers to the following verbs: ΗΘΕνΚΝ strepho: ‘to turn about, aside’ȱ ΦΔΓΕΕϟΔΘΝ aporripto: ‘to throw away/put away/throw oneself down’ πΑ΍ΗΛϾΝ eniskhyo: ‘to strengthen/confirm/prevail in or among’ ΦΚ΅ΑϟΊΝ aphanizo: ‘to hide/destroy/disappear’ȱ Έ΍ΓΕϾΗΗΝ diorysso: ‘to dig through’ ΎΕΓϾΝ krouo: ‘to strike’ Φ·ΓΕΣΊΝ agorazo: ‘to buy in the market’ȱ ΦΑ΅ΘνΏΏΝ anatello: ‘to make rise up/rise/appear’ ΅ЁΒΣΑΝ auksano: ‘to grow’ 36

Semantic differentiation is not observed in anticausative active and non-active structures. Certainly, the nonactive structures with passive meaning (implied or overt presence of agent, i.e. passive structures) differ from the active structures with anticausative meaning.

116

Chapter Three

πΔ΍ΗΘΕνΚΝȱȱepistrepho: ‘to return’

ȱ Of course, Jannaris does not differentiate here between transitive verbs with object omission and intransitive verbs that change into alternating verbs. The complete picture of the era, however, is that of the change from non-active anticausative verbs into active. I assume that this change is connected with and possibly due to two factors: (i) The possibility of passivisation for certain of the anticausative verbs creates the need for differentiation of the anticausative structure (the agent argument cannot be present) from the passive structure (possibility of the presence of the agent argument). The difference between the two structures is expressed through the morphological contrast of the active voice for the anticausatives and the nonactive voice for the passives (active voice morphology (anticausatives) vs. non-active voice morphology (passives)). The need for this distinction makes a strong appearance in this period because the passive structures are consistently marked from this period and afterwards with non-active voice and not with active voice as in Classical Greek (Lightfoot 1979), and because (from this period and after) the same preposition is being used (ЀΔϱȱhypo ‘by’) both for the agent (only in passive structures) and for the cause (in anticausative structures). (ii) The general changes in the area of voice (mainly the loss of the non-active transitive structure for the denotation of “personal interest for the verb action”/benefactive meaning) and the impossibility of semantic linking of non-active voice and transitive structures (only the deponent transitive verbs can, moreover, bear non-active morphology) resulted in instability in the area of voice. (b) Continuation of the presence of non-active anticausatives On the other hand, quite a few transitivity alternations continue (in spite of the general extension of active anticausatives) to mark with non-active voice the anticausative type (the same as in classical Ancient Greek). We have observed that in classical Ancient Greek, the change of a state as a result of a natural process is denoted with a non-active form (ΑϱΗΓΖȱ ΐ΅Ε΅ϟΑΉ΍ȱ ΐΉȱ Ȧȱ nosos marainei me / illness.NOM wither.ACT.PRES.3SG I.ACC / ‘an illness withers me’ -ȱ ΑϱΗУȱΘϲȱΗЗΐ΅ȱΓЁΎȱπΐ΅Ε΅ϟΑΉΘΓȱȦ noso:i to so:ma ouk emaraineto / illness.DAT the.NOM body.ACC not wither.NACT.IMPERF.3SG / ‘the body did not wither from the illness’). According to the data of Hellenistic-Roman Koine for the verbs of natural process, we usually meet the structure “DP-nominative + active type + DP-accusative” (ȑȝȑȖğǴǴȁ systello: ‘to draw-together’, njǬDŽȑȖğǴǴȁ diastello: ‘to put asunder/expand’, ñȟDŽȖèģȡȁ eksatmizo: ‘to smoke/ steam’, èDŽȋDŽģȀȁ maraino: ‘to wither’ + DP-Accusative): (64)

Ύ΅Ό΅ϟΕΓΙΗ΍ȱΎ΅ϠȱȱΗΙΗΘνΏΏΓΙΗ΍ȱ

ȱȱȱȱȱȱΘχΑȱȱȱȱȱȱȱȱȱȱЂΏ΋ΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΉϢΖȱ΅ЁΘϲȱΘϲȱȱπΈЏΈ΍ΐΓΑȱ

kathairousi kai systellousi te:n hyle:n eis auto to edo:dimon clean.3PL and draw-together.ACT.PRES.3PL the.ACC material.ACC to that the eatable ‘(friction and digestion) clean and draw the material together, so that it can be eatable’ (Plutarch, Table Talk, 693e, 4; AD 1-2)

Concerning the non-active form, the absence of certainty as to whether it is a reflexive, passive, or anticausative structure (with an illness or a function of the body as cause) depends on the verb meaning: (65) a.ȱ ΅ϡȱȱΐξΑȱȱȱΦΕΘ΋Εϟ΅΍ǰȱȱȱȱΚ΋ΗϟΑǰȱȱȱȱΗΙΗΘνΏΏΓΑΘ΅ϟȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΉȱΎ΅ϠȱȱȱΈ΍΅ΗΘνΏΏΓΑΘ΅΍ȱȱ hai men arte:riai, phe:sin, systellontai te kai diastellontai ȱ the prt

arteries.NOM say.3PL draw-together.NACT.PRES.3PL and

put-asunder.NACT.PRES.3PL

‘it is said that the arteries both draw together and put asunder’ (ǹnonymous Londiniensis, medical, 29, 4; AD 1) b. ΦΑΌΉϧȱȱ ·ΤΕȱψȱȱ ȱȱΜΙΛχȱȱȱ ȱȱΔΕϲΖȱȱȱȱȱπΔ΍ΗΘφΐ΋ΑǰȱȱϳΔϱΘΉȱȱ΅ϡȱ anthei gar he: psykhe: pros episte:me:n, hopote hai ȱ blossom.ACT.PRES.3SG as the.NOM soul.NOM towards knowledge when the.NOMȱȱ

ȱ ȱ

ΘΓІȱȱ

ȱ

ΗЏΐ΅ΘΓΖȱȱ ΦΎΐ΅Ϡȱȱ

ΛΕϱΑΓΙȱȱ ΐφΎΉ΍ȱȱȱ

ΐ΅Ε΅ϟΑΓΑΘ΅΍ǯ

tou so:matos akmai khronou me:kei marainontai. the.GEN body.GEN vigour.NOM time.GEN lenght.DAT wither.NACT.PRES.3PL ‘as the soul blossoms to knowledge whenever the body’s vigour withers at length’ (Philo Judaeus, On Dreams, 1, 11, 2; 1 BC-AD 1) c. ΐν·΍ΗΘΓΖȱ ȱΈȂȱ megistos d’ biggest.NOM prt

ЖΑȱ

ΘЗΑȱ

o:n to:n be.PART.NOM the.GEN

ЀΔΓΎΣΘΝȱ ΘϱΔΓΙΖȱ

ΗΙΗΘνΏΏΉΘ΅΍ȱ

hypokato: topous

systelletai

ΥΔΣΑΘΝΑȱ ΔΓΘ΅ΐЗΑdzȱ Ύ΅ΘΤȱΈξȱȱΘΓϿΖȱ

hapanto:n potamo:n… kata de tous all rivers.GEN in prt the

Diachronic Data and Analysis

117

under places.ACC draw-together.NACT.PRES.3SG ‘it is the biggest of all the rivers... it is drawing together under the earth’ (Diodorus Siculus, Library, 1, 32, 4; 1 BC)

Example (65b) is indicative of the preference during the Hellenistic-Roman Koine period of morphological marking ΦΑΌЗ antho: ‘to blossom’ with an active form, whereas in the same sentence, ΐ΅Ε΅ϟΑΓΐ΅΍ marainomai ‘to wither’ is marked with a non-active form. An important difference between the two verbs is the absence of transitive ΦΑΌЗ antho: ‘to blossom’ (which could be passivised), in contrast to the presence of the transitive ΐ΅Ε΅ϟΑΝ maraino: ‘to wither’ 37. It should be noted that the non-active morphology of verbs such as ΐ΅Ε΅ϟΑΓΐ΅΍ marainomai ‘to wither’, according to the diachronic study, has to be relative to the generalisation of the non-active form in verbs characterised as middle reflexive or pseudo-reflexive and in which we can observe a high level of affectedness of the subject-patient, a characteristic of the reflexives38. The diachronic study indicates that the verbs of the ΐ΅Ε΅ϟΑΓΐ΅΍ marainomai ‘to wither’ group also showed a high level of affectedness and, in many examples, showed reflexive interpretation. The verb ΐ΅Ε΅ϟΑΓΐ΅΍ marainomai ‘to wither’ (and the remaining verbs of the same category) was considered by the native speakers as ‘middle reflexive/pseudo-reflexive’, whereas the non-active morphology prevailed due to the high level of affectedness arising from the lexical meaning of the verb. In relation to the above, it is important to stress that although the data indicate “passive” interpretation of the anticausatives in certain instances, regarding the presence of the non-active morphology in anticausatives, the role of reflexivity may be basic39.

ȱ

(66) transitive active Έξȱ ΎΝΚχȱ Ύ΅Ϡȱ ΆϟΓΖȱ οΈΕ΅ϧΓΖȱ πΔϠȱΗΛΓΏϛΖȱ ΦΔΓΎΉϟΐΉΑΓΖȱ a. ψΗΙΛϟ΅ȱ   he:sykhia de ko:phe: kai bios hedraios epi skhole:s apokeimenos  stillness.NOM prt noiseless.NOM and life.NOMsteady.NOM on leisure lay.PART.NOM ΓЁȱ ΐϱΑΓΑȱ ΗЏΐ΅Θ΅ȱ ΦΏΏΤȱ Ύ΅Ϡȱ ΜΙΛΤΖȱ ΐ΅Ε΅ϟΑΉ΍ȱ  ou monon so:mata alla kai psykhas marainei not only bodies.ACC but and souls.ACC wither.ACT.PRES.3SG ‘noiseless stillness and steady life with leisures wither not only the bodies but also the souls’ (Plutarch, Is the saying "live in obscurity" right?, 1129d, 5; AD 1-2) intransitive non-active ȱ·ΤΕȱΌΣΏΏΉ΍ȱ ȱΎ΅ϠȱΔΣΑΘ΅ȱ ΐ΅Ε΅ϟΑΉΘ΅΍ȱȱȱ b.ȱ ΔΣΑΘ΅ȱ panta gar thallei kai panta marainetai everything.NOM as bloom.ACT.PRES.3SG and everything.NOM wither.NACT.PRES.3SG ‘as everything blooms and withers’ (The Life of Aesop, Vita Westermanniana, 110, 6; AD 2)ȱ c. ΐ΅Ε΅ϟΑΉΘ΅΍ȱ ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΈξȱȱ΅ЁΘ΅ϧΖȱȱȱȱΘΤȱΔΘΉΕΤȱȱȱȱȱȱȱȱΎ΅ϠȱΩΆΕΝΘΓ΍ȱȱȱȱȱȱȱȱȱȱ·ϟΑΓΑΘ΅΍ȱȱȱΘΓϧΖȱȱϢΛΌϾΗ΍ȱȱ marainetai de autais ta ptera kai abro:toi ginontai tois ikhthysi wither.NACT.PRES.3SG prt these.DAT the wings.NOM andȱuneatable.NOM become.3PL the

fishes.DAT

‘their wings wither and become uneatable to fishes’ (Aelian, On the characteristics of animals, 15, 1, 29; AD 2-3)ȱ

The verbs ΏΙΔЏ lypo: ‘to cause grief’ȱŠ—ȱ ΌΉΕΐ΅ϟΑΝȱthermaino: ‘to warm’had transitive active and intransitive non-active use during the Classical Greek period. Similarly, active transitive, non-active intransitive use, is available during the Hellenistic period. (67) ΏΙΔЏ lypo: ‘to grieve’ intransitive non-active a.
н̄ΕΣ·ΉȱϳȱȱȱΚΌΓΑЗΑȱȱȱȱȱȱȱȱΛ΅ϟΕΉ΍ЪȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȮ̒ЁΈ΅ΐЗΖǰȱȱȱΦΏΏΤȱȱȱΐκΏΏΓΑȱΏΙΔΉϧΘ΅΍ ara ge ho phthono:n khaire? –oudamo:s, alla mallon lypeitai 37

řŞ

39

Indeed, both verbs denote a biological process. According to the observations of Levin & Rappaport Hovav (1995) for other languages, ‘blossom’ could not be considered anticausative as long as the criterion for (anti)causativity is the presence of active transitive and intransitive type. For Theophanopoulou-Kontou (1999a, b, 2004), ‘blossom’ is found at the end of the continuum with the possibility of presence of only metaphorical cause (?Ƞ ȑȡȦIJĮȢ IJȘȞ ȐȞșȘıİ / ?o erotas tin anthise / the.NOM love.NOM she.ACC.WEAK blossom.ACT.PAST-PERFVE.3SG / ‘love made her blossom’). Cf. Modern Greek ıțȠȡʌȓȗȠȝĮȚ skorpizome ‘to scatter’, ıȣȖțİȞIJȡȫȞȠȝĮȚ singendronome ‘to collect’, İȟĮijĮȞȓȗȠȝĮȚ eksafanizome ‘to disappear’ (Theophanopoulou-Kontou 2004). The factors for the spread of the non-active voice are related to the fact that the non-active voice becomes transparent for functions such as the passive and the reflexive (presence of agent), resulting in differentiation from the anticausative interpretation (absence of agent).

118

Chapter Three prt

the envious.NOM rejoice.ACT.PRES.3SG by-no-means but

rather grieve.NACT.PRES.3SG.

‘Does an envious man rejoice? - By no means; he rather grieves’ (Epictetus, Discourses, 2, 12, 7; AD 1-2) b. ΘχΑȱȱ πΔ΍ΗΘΓΏχΑȱȱ Ύ΅Ϡȱȱ ΦΑν·ΑΝΑȱȱ Ύ΅Ϡȱȱ πΏΙΔφΌ΋Αȱȱ ΉΏ΋Δ΋Ό΋Α te:n epistole:n kai anegno:n kai elype:the:n ele:pe:the:n ȱ the.ACC letter.ACC and read.1SG and grieve.NACT.AOR.1SG grieve.NACT.AOR.1SG ‘he read the letter and grieved’ (BGU (Aegyptische Urkunden aus den Staatlichen Museen zu Berlin, Griechische Urkunde) 4, 1079 rp r 9) transitive active σΘ΍ȱȱ ΈϾΑΝΑΘ΅΍ȱȱ ΏΙΔΉϧΑȱȱ ΐ΋ΈνΑ΅ȱ c.ȱ ΐ΋Έξȱȱ ΆΓΙΏ΋ΌνΑΘΉΖȱȱ ȱȱȱȱȱ ȱȱȱȱ me:de boule:thentes eti dyno:ntai lypein me:dena not wish.PART.NOM more can.3PL grieve.ACT.PRES.INF nobody.ACC ‘he would be rendered incapable of causing grief to anyone’ ǻPolybius, Histories, 15, 31, 5-6; 3-2 BC)ȱ ΘΓϾΘΝ΍ȱȱ ΐξȱȱ ΏΙΔφΗΉ΍Ζȱȱ ΏΓ΍Δ΋Η΍Ζȱȱ ΔΣΑΘΝΖȱ d.ȱ ΓЁΈξΑȱȱ ouden touto:i me lype:seis loipe:sis panto:s nothing that.DAT I.ACC grieve.ACT.FUT.2SG grieve.ACT.FUT.2SG in-all-ways ‘you will not grieve me with that at all’ (BGU (Aegyptische Urkunden aus den Staatlichen Museen zu Berlin, Griechische Urkunde) 1, 248 r 29) Cf. the presence of the cognate object ΓЇΗ΅ȱȱ ΏΙΔΉϧȱȱ ΐΉȱȱ ΏϾΔ΋Αȱȱ ΦΑϟ΅ΘΓΑȱȱ ȱȱȱ e.ȱ ȁ̄ΏΏȂ΅ЂΘ΋ȱȱ ΘΓ΍΅ϾΘ΋ȱΘ΍Ζȱȱ ȱȱȱȱȱȱȱȱ ȱ all’ haute: toiaute: tis ousa lypei me lype:n aniaton but

she.NOM such

somebody be.PART.NOM grieve.ACT.PRES.3SG I.ACC grief.ACC incurable

‘but she –as is– causes me incurable grief’ (Heliodorus, Ethiopian Romance, 2, 33, 4; AD 3) (68) ΌΉΕΐ΅ϟΑΝ thermaino: ‘to warm’ intransitive non-active Ύ΅Ϡȱȱ ΛΓΕΘΣΊΉΗΌΉȱ a.ȱ е̗ΔΣ·ΉΘΉȱȱ πΑȱΉϢΕφΑϙǰȱȱ ΌΉΕΐ΅ϟΑΉΗΌΉȱȱ hypagete en eire:ne:i, thermainesthe kai khortazesthe go.IMP.2PL in peace warm.NACT.PRES.2PL and sate.NACT.PRES.2PL ‘go in peace; warm and sate yourselves’ (New Testament, James, 2, 16, 2; AD 1) b.ȱ ΦΑΌΕ΅Ύ΍ΤΑȱȱ ΔΉΔΓ΍΋ΎϱΘΉΖǰȱȱ ϵΘ΍ȱȱΜІΛΓΖȱȱȱȱώΑǰȱȱȱȱΎ΅ϠȱȱȱπΌΉΕΐ΅ϟΑΓΑΘΓ anthrakian pepoie:kotes, hoti psykhos e:n, kai ethermainonto fire-of-coals.ACC make.PART.NOM that cold was and warm.NACT.IMPERF.3PL ‘having made a fire of coals, for it was cold, and they were warming themselves’ (New Testament, John, 18, 18, 2; AD 1)ȱ transitive active c. ΗϿȱȱȱϊΏ΍ΓΖȱȱȱΉϨȉȱȱȱȱȱȱȱȱȱΈϾΑ΅Η΅΍ǯǯǯȱȱΎ΅ϠȱȱΘΤȱΗЏΐ΅Θ΅ȱΘЗΑȱȱΦΑΌΕЏΔΝΑȱΌΉΕΐ΅ϟΑΉ΍ΑȱȱȱȱȱȱΗΙΐΐνΘΕΝΖȱ sy he:lios ei; dynasai... kai ta so:mata to:n anthro:po:n thermainein symmetro:s you.NOM sun be.2SG can.2SG and the bodies.ACC the.GEN men.GEN warm.ACT.PRES.INF properly ȱ

‘you are the sun; you can... warm the bodies of men properly’ (Epictetus, Discourses, 3, 22, 5; AD 1-2)

3.2.4 Summary The picture of the causativisation processes remains consistent during the Hellenistic-Roman period. The causativisations of intransitive types are, on the whole, especially productive except in the instances of intransitives denoting change-of-state from internal cause. The initial intransitive nonalternating verbs are of active or non-active morphology and preserve the voice morphology that they bore as intransitives in their anticausative structure and after their causativisation. In the innovative transitive structure, they always bear, however, active morphology. The presence of an external argument (denoting the agent or the cause) and the active morphology of the transitive structures are more closely connected than the presence of the direct object and the active morphology. It appears that active voice is not related to the presence of the accusative case but to the presence of the agent in the subject position (in the nominative case). In any case, the causativisations of intransitive verbs of active morphology open up a gateway for the formation of transitivity alternations without voice change. The reanalysis of intransitive types as transitives is favoured by the structures with infinitive and a DP (or clitic) as subject in the accusative. Moreover, in this period, the new causative types were also strengthened by the possibility of the presence of intransitive types with cognate objects. In contrast, the prefixes and the suffixes encode differences in aktionsart and do not contribute to the presence of new causative types. Changes in the opposite direction (anticausativisations, i.e. new anticausative types from transitives) are not systematically observed. In contrast to the general picture of changes in the verb system of the Hellenistic-Roman period, the causative verbs have stable syntactic behaviour.

Diachronic Data and Analysis

119

The situation is not the same with accusative verbs, where a greater degree of instability in relation to the argument structure (particularly free shift from intransitives to accusatives and vice versa) and their case requirements (parallel use with the accusative of other cases and PPs; extension of the accusative case in the DPs-direct objects of all verbs belonging to this group) is observed. Even if innovative anticausative use of initially non-alternating causative verbs is unattested, an innovative intransitive (not anticausative) use of initially exclusively accusative verbs is observed. The shift, therefore, from [+transitive] (= [+accusative case]) to [-transitive] (= [-accusative case]) cannot be totally ruled out, whereas the systematic shift from causative non-alternating to causative alternating (with the creation of anticausative type) can be ruled out. From an examination of the data, it is seen that the innovative syntactic behaviour of an old intransitive verb with a complement in the accusative case does not, in all instances, comprise a similar phenomenon: (a) intransitive verbs denoting profession were turned into accusative verbs with specific denotation of the profession as the direct object. This concerns verbs that belong to the semantic group of activities; the DP-accusative with these verbs comprises an argument deriving only from the idiosyncratic meaning, not also from the meaning of the event schema of the verb. (b) Similarly, in the New Testament, the verb meaning of intransitive verbs that denoted action or were psych-verbs is afforded greater specialisation (indication of the specific action or the cause of the psychological condition) with the addition of the direct object but not, however, causative interpretation as well. (c) The replacement of the PP by DP in the accusative comprises a third way of shifting from the intransitive verb group to the accusative verb group. Concerning voice morphology, the passive structures are, moreover, consistently marked with nonactive voice in the Hellenistic-Roman period. In contrast to the Classical Greek period, during which passive structures with (parallel expression of the agent and) active voice were possible, in the Hellenistic-Roman period, only anticausative (and not passive) types accompanied by a cause-PP (Δ΅ΕΤȱpara ‘from’ /ȱЀΔϱȱhypo ‘by’) bear active voice. Even if the non-active voice begins to lose the “personal interest” meaning (benefactive meaning) in transitive structures, the structure “non-active verb + DP-accusative” continues to be possible, as well as the passivisation of ditransitive verbs with the retention of the accusative case of the second object. The reasons for the abandonment of the distinction and the separate meaning of the non-active transitive structure can be sought in the following: (a) both voices (active and non-active) could be used in transitive structures with the agent as subject, with the result that, in many instances, they are in alternation already in Classical Greek without any particularly obvious semantic difference; (b) the quite different meanings that the non-active (middle) voice could receive were difficult to unify in one and only one meaning, especially from the moment when the transparent meaning of passivisation competed with the other meanings; and (c) the use of clitics40 for the denotation of the benefactive, instead of the use of non-active transitive structures, is frequent; the use of clitics is connected to the reanalysis, in this period, of the clitics in the genitive case from pronouns governed by a noun (possessive) to pronouns governed by a verb. Concerning the transitivity alternations, we have ascertained a general tendency to the spread of active voice with the anticausative verbs and encoding with the same voice (active) of both the causative and the anticausative types. The active voice spreads to anticausatives that participate in transitivity alternations and is not connected with the derivation of new alternations from initial nonalternating transitive or intransitive verbs. It is concerned, in other words, with a change in the voice morphology of anticausative verbs, even if grammarians have argued that it is concerned with the creation of a new intransitive verb from a transitive verb, not taking into consideration that there already existed in earlier periods a corresponding anticausative verb with non-active morphology (and that consequently the intransitive use of these verbs is not innovative). In parallel, as an indication of change in progress, in this period, anticausative types that can bear active (corresponding to the new type) and non-active voice morphology (corresponding to the type of the earlier period) without difference in meaning are in evidence. This tendency for the spread of active anticausatives is linked with: (a) the first appearances of passive types of anticausative verbs (the patient argument in the subject position and presence of the agent in the by-Phrase) and with a general reordering in the area of voice; (b) the loss of the non-active transitive structure for the denotation of “personal interest” (benefactive meaning) and the following impossibility of a semantic connection between non-active voice and transitive structures; (c) the phonological merger of the non-active ending -ΉΘ΅΍ /-etai/ (3rd person, singular number, non-active voice) with the active ending ȬΉΘΉ/-ete/ (2nd person, plural number, active voice). This homophony – when there was no overt expression of the subject– led through reanalysis to a change from non-active 40

Cf. Drachman 1997.

120

Chapter Three

anticausative into active; and (d) the coincidence of the accusative case of the subject of the infinitives with the accusative case of the direct object of the finite types; in instances of transitive infinitive types, the DP in the accusative case can be analysed by native speakers as the object of a transitive active verb or as the subject of an intransitive active verb, while the second analysis (subject of intransitive) results in the innovative active voice morphology of the intransitive structures.

3.3 Medieval Greek 3.3.1 The general picture Sources for the oral language of the Early Medieval period (5th-10th century AD), in comparison with the sources for the Hellenistic-Roman and the Late Medieval periods, are very few and the texts of everyday register are very limited. Papyri are only preserved from the beginning of this period (until the 8th century), but they are rare. For the Middle and mainly the Late Medieval Greek period (11th-15th century AD), the sources are chiefly extracts of dialogues, expressions of reactions at the Hippodrome, books on the interpretation of dreams, practical medicine, wishes, curses, and quite a few chronicles (Glykatzi-Ahrweiler 1999). In the period of the Comnenus Dynasty, novels became well-known and loved. The language of this particular type of text is mixed, with learned elements and a strong influence from colloquial speech. In the 12th century, the Ptochoprodromika poems, possibly written by Theodoros Prodromos, resemble everyday speech. Evidence about the language of the 14th century is to be found in novels: the Chronicle of Moreas, and the texts of Dellaporta and Falieros. For the 15th century, the official and non-literary (notarial archives) documents from Crete, the Ionian islands, and the Cyclades are of great help (Manolessou 2003; Karantzola, Tiktopoulou & Frantzi 2006). Due to the limited number of non-literary texts in everyday language from the period between the 11th and 15th centuries, our data also cover texts of the 16th and 17th centuries. After the fall of the Byzantine Empire in 1453, the stabilisation of the form of the Modern Greek dialects and idioms begins (Tzitzilis 2000). Cretan literature begins with didactic texts in the 15th century and flourishes in the 16th and 17th centuries with works that cover all theatrical types; pastoral novels, tragedies, comedies, religious dramas (Horrocks 1997; Simeonidis 1999). In the second half of the 15th century, the beginning of ‘demotic’ Modern Greek poetry is recorded with the appearance of didactic and religious texts in simple vernacular41 language (and the introduction of sermons in vernacular language in the tradition of the Greek church)42. The general changes that appear to go hand in hand with the changes in the area of transitivity are concerned chiefly with the nouns and the morphological case of nouns. Kapsomenakis (1938) has observed that the marking of subjects of finite verbs with accusative case is frequent in papyri in the Byzantine period. (1) ΦΏΏΤȱȱ Ύ΅Ϡȱȱ ΘΓϿΖȱȱ ̓νΕΗΓΙΖȱȱ ώΏΌΉΑȱȱ πΑȱȱ ̖φΑ΍ȱ ala ke tus Persus ilthen en Tini ȱȱȱȱ but and the.ACC Persians.ACC come.ACT.PAST.PERFVE.3SG in Tini ‘but Persians also came to Tini’ (P. Russ-Georg (Papyri russischer und georgischer Sammlungen) IV (Die Kome-Aphrodito Papyri der Sammlung Lichaþov) Anhang S.100 (619/29p), 3-4)

i ΘΓϿΖȱ ̓νΕΗΓΙΖȱ Ȧȱ tus Persus / ‘the Persians’: the subject of the finite verb ώΏΌΉΑ ilthen ‘come’in accusative case DPs in the accusative as subjects of finite verbs appear also elsewhere in papyri of the Byzantine era: (2)

41 42

a.ȱ ΦΏΏΓΘΕϟ΅Αȱ ·ΙΑ΅ϧΎ΅Αȱȱ πΎΏ΋ΕΓΑϱΐ΋ΗΉΑȱȱ ΅ЁΘϱΑ  alotrian jinekan eklironomisen afton foreign.ACC woman.ACC inherit.ACT.PAST.PERFVE.3SG he.ACC ‘a foreign woman has inherited him’ ((P. Oxy. (The Oxyrhynchus Papyri)VII 1067, 7/9, IIIp)

Cf. Pantelides 1999, 2003. I will follow David Holton and Geoffrey Horrocks (“Grammar of Medieval Greek Project”, ): “... These dates [between 1100 and 1700] are chosen because texts in the vernacular become available in significant quantity only in the 12th century, and, although there is no obvious point at which to locate the end of the "medieval" period, by the 18th century important cultural and political changes are afoot. The period 1100-1700 constitutes a coherent whole in terms of the development of the Greek vernacular. ... In certain cases, early medieval texts (5th-11th century) will be taken into account, mainly to illuminate points of historical evolution or the earliest dating of phenomena”.

Diachronic Data and Analysis b. ЙΑȱȱȱȱȱȱȱȱȱȱȱΘΤȱȱ ΐνΘΕ΅ȱȱ ȱΎ΅ϠȱΘΤΖȱ·Ή΍ΘΑϟ΅Ζȱȱ on ta metra ke tas jitnias

121

Έ΍ΤȱȱΘЗΑȱΔΕΓΎΘ΋Θ΍ΎЗΑȱȱ

ΘνΘ΅ΎΘ΅΍ȱ

dhia ton proktitikon

tetakte

which.GEN the meters.NOM and the adjoining-areas.ACC from the previous-owners.GEN define.NACT.PERF.3SG

‘the meters and the adjoining areas which have been defined by the previous owners’ (CPR (Corpus Papyrorum Raineri) 4 (=Chrest II 159), 9/10, 51/3p) c. Θ΅ϾΘ΋Αȱȱ ΘχΑȱȱ πΔ΍ΗΘΓΏχΑȱȱ π·ΕΣΚ΋ȱȱ πΑȱȱ ̋ΐΓϾΉ΍ taftin tin epistolin eghrafi en Thmui that.ACC the.ACC letter.ACC write.NACT.PAST.PERFVE.3SG in Thmui ‘that letter was written in Thmui’ (P. Par. (Paris Papyri in Notices et Extraits XVlll, ii, 18, 12, 3 (Iip)) d. πΗΚΕ΅·ϟΗΌ΋ȱȱ ΘχΑȱȱ ΎνΏΏ΅Αȱȱ ΅ЁΘΓІȱȱȱȱȱ esfrajisthi tin kelan aftu seal.NACT.PAST.PERFVE.3SG the.ACC cell.ACC his ‘his cell was sealed’ (P. Oxy. (The Oxyrhynchus Papyri)VII 1067, 12/3)

In the above examples, we can observe that in many instances the DPs in the accusative are subjects of non-active finite verbs43. Also indicative of the instability in the nominal morphological system is the presence of the opposite phenomenon, namely the use of nominative instead of accusative case for the objects of finite verbs. Again, examples of non-active finite verbs with their object in the nominative case are very frequent: (3)

ȱ

a. ΐχȱȱ ΏΙΔ΋ΌϜΖȱȱ ϳȱȱ ΌΣΑ΅ΘΓΖ mi lipithis o thanatos   not grieve.NACT.PERFVE.2SG the.NOM death.NOM ‘do not grieve because of the death’ (Lefebvre's corpus of Greek Christian inscriptions from Egypt, 482, 1/ 4) ΚϱΕΓΖȱȱ ΦΔΓΈЏΗΝȱȱ ΗΓ΍ b. ϳȱȱ o foros apodhoso si the.NOM tribute.NOM give.ACT.PERFVE.1SG you.DAT ‘the tribute I am going to give you’ (P. ȅxy. (The Oxyrhynchus Papyri) VIII 1126, 13/4 (Vp))

These two tendencies clearly helped and strengthened the reanalysis of causative and anticausative verbs. These tendencies are also connected with the loss of the infinitives that take subjects in the accusative case (Joseph 1978/1990; for the parallel use of infinitive and finite types in translations of earlier texts during this period, cf. Karla & Lavidas 2003). Linked to this phenomenon is the coincidence of the nominative and accusative case for the nouns of the first declension and the impossibility of distinguishing between the two morphological cases (see Lehmann 2005 for corresponding phenomena of confusion between the nominative and accusative case in Latin).

3.3.2. Innovative causative use of intransitive verbs: is there a limit to causativisations? 3.3.2.1 Causativisations The process of the production of innovative causative types continues to be particularly productive. The causative type is created based on an earlier non-alternating intransitive verb of active, non-active or both active and non-active voice. (a) Causative types from intransitive active According to Kapsomenakis (1938), in (4), we have one of the first examples of transitive use of the verb ĮʌȠșȞȒıțȦ apothnisko/ʌİșĮȓȞȦ petheno ‘to die’ in Byzantine papyri. Kapsomenakis observes that the first use of the verb as transitive is not literal (it does not talk about the end of biological life) but refers to the presence of circumstances that severely afflict and torture the speaker. (4)

ΐχȱȱ ·νΑ΋Θ΅΍ȱȱ

ΦΐνΏΉ΍΅Αǰȱȱ

ϣΑ΅ȱȱ ΐχȱȱ ΦΔΓΌ΅ΑΓІΗ΍ȱȱ

ΐΉ

mi jenite amelian, ina mi apothanusi me not occur.NACT.PRES.3SG negligence.NOM to not die.ACT.PERFVE.3PL I.ACC ‘so that negligence does not occur, for that they will not make me die’ (P. Iand. (Papyri Iandanae II 23, 14 (VI/VII p))

43

This phenomenon can be interpreted as: (a) ability of non-active verbs to assign accusative case or (b) the sole argument is not moved to the position in which nominative case is given.

122

Chapter Three

According to the above, Chatzidakis (1927) erroneously places the first causative use of the verb ĮʌȠșȞȒıțȦ apothnisko/ʌİșĮȓȞȦ petheno ‘to die’ in Modern Greek. The verb ĮʌȠșȞȒıțȦ apothnisko/ʌİșĮȓȞȦ petheno ‘to die’ consistently intransitive during the Hellenistic era and has intransitive and transitive use throughout the Medieval period up to the Early Modern Greek period (17th century)44. (5)

intransitive active ΔΉΌ΅ϟΑΓΙΗ΍Αȱȱ ΅ЁΘΓϟȱȱ  a.ȱ ΈξΑȱȱ dhen pethenusin afti not die.ACT.PRES.3PL they.NOM ‘they do not die’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 263, 20; AD 17) transitive active b.ȱȱ ΩΑΉΐΓΖǰȱȱΉϨΔΉΑȱȱ ΅ЁΘφǰȱȱ ·ϟΑΓΐ΅΍ǰȱȱ Ύ΅ϠȱσΘΊ΍ȱȱ ΐΔ΅ϟΑΝȱȱ ΉϢΖȱΘχΑȱΐϾΘ΍Αȱ anemos, ipen afti, jinome, ke etzi beno is tin mitin ȱȱȱȱȱȱȱȱ ȱ wind said.SG she.NOM become.1SG and so enter.1SG in the noseȱ ȱ ȱ ȱ ȱ ΅ЁΘΓΑЗΑȱȱ ϳΔΓІȱȱ Όξȱȱ ΑΤȱȱ ΘΓϿΖȱȱȱ Δ΅΍Ό΅ϟΑΝȱȱ aftonon opu the na tus petheno these.GEN who want.3PL to they.ACC.WEAK die.ACT.IMPRFVE.1SG ‘I become as the wind, she said, and like that I enter the nose of those whom I want to make die’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 163, 15-16; AD 17) exclusively intransitive in the previous period (Hellenistic-Roman): Έξȱ Ύ΅Ϡȱ ̝ΑΑϟΆ΅Ζȱ ϳȱ ΗΘΕ΅Θ΋·ϱΖȱ c. ΦΔνΌ΅ΑΉȱ apethane de kai Annibas ho strate:gos die.ACT.AOR.3SG prt and Anivas.NOM the.NOM commander.NOM ‘also Anivas, the commander, died’ (Diodorus Siculus, Library, 13, 86, 3; 1 BC)

Regarding the verb ĮȡȡȦıIJĮȓȞȦ arosteno ‘to fall ill’ (which is found in the list given by Chatzidakis (1927) with the causativisations that he assumes to be innovative for Modern Greek), again the data from the Medieval and Early Modern Greek period show that it was causativised before the Modern Greek period (like the verb ʌİșĮȓȞȦ petheno ‘to die’). In both examples below, the object is a clitic. (6)

intransitive active ΆΓΗΎϲΖȱȱ ΦΕΕЏΗΘ΋ΗΉ a. Ύ΅Ϡȱȱ ϳȱȱ ke o voskos arostise and the.NOM shepherd.NOM fall-ill.ACT.PAST-PERFVE.3SG ‘and the shepherd fell ill’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 269, 45; AD 17) transitive active b.

Ύ΍ȱȱȱΓȱȱ ΏΓ·΍ΗΐϱΖǰȱȱΓΔΓϾȱȱΆ΅Ώ΅ǰȱȱ ·΋ȱ·΍΅ϟΑΉ΍ȱȱ

ki o ȱ lojismos,

opu

vala,

ji jieni

ȱȱ·΋ȱ΅ΕΕΝΗΘΉϟȱ

ȱji arosti

ȱΐΉ

me

and the thought.NOM which ȱput.1SG or ȱcure.ACT.PRES.3SG or fall-ill.ACT.PRES.3SG I.ACC.WEAK

‘and the thought I had either cures or makes me ill’ (Erotokritos, I, 2094; AD 17) i in the previous period - Hellenistic-Roman period: only intransitive Ύ΅ϠdzȱωΕΕЏΗΘ΋ΗΉΑ  c. Ύ΅Ϡȱ σΔΉΗΉΑȱ ͞ΛΓΊϟ΅Ζȱ kai epesen Okhozias kai... e:rro:ste:sen and fell.3SG Ohozias.NOM and fall-ill.ACT.PAST.PERFVE.3SG ‘and Ohozias fell and became ill’ (Septuagint, Kings IV, 1, 2, 3; 2-1 BC ) often with cognate object d. ̍΅ϠȱωΕΕЏΗΘ΋ΗΉȱ ȱȱȱȱȱȱΘϲȱȱȱΔ΅΍ΈΣΕ΍ΓΑȱȱȱ΅ЁΘΓІȱΦΕΕΝΗΘϟ΅ΑȱΎΕ΅Θ΅΍ΤΑȱȱȱΗΚϱΈΕ΅ȱ Ȁai e:rro:ste:se to paidarion autou ȱarro:stian ȱȱȱkrataian sphodra and fall-ill.ACT.PAST.PERFVE.3SG the little-boy.NOM his

illness.ACC ȱȱȱserious.ACC severe.ACC

‘and his little boy fell ill with a very serious illness’ (Septuagint, Kings III, 12, 24g, 2; 2-1 BC) e. ̏ΉΘΤȱ·ΤΕȱΘΕϟ΅ȱσΘ΋ȱωΕΕЏΗΘ΋Η΅ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΦΕΕΝΗΘϟ΅ΑȱΘχΑȱΦΔϲȱΗΔΏ΋ΑϲΖȱΎ΅ϠȱΗΘΓΐΣΛΓΙȱȱ meta ȱȱgar ȱtria ete: ȱe:rro:ste:sa ȱ arro:stianȱȱȱȱ te:n apo sple:nosȱȱȱkai ȱstomakhou after because three years fall-ill.ACT.PAST.PERFVE.1SG illness.ACC the.ACC from spleen and stomach

‘after three years I fell ill with an illness of the spleen and stomach’ (Palladius, Lausiac history, 35, 11, 11; AD 4-5)

Psaltes (1913/1974) in his study on Byzantine chronicles refers to the following instances of causativisation of intransitive active verbs: ȬΈΙΑ΅ΐϱΝȱdhinamoo ‘to strengthen’ (Doukas, 299, 22; AD 15) 44

On the characteristics of lability in Early Modern Greek, cf. Karantzola & Lavidas (2009-to apper).

Diachronic Data and Analysis

123

ȬπΏ΅ΘΘΓΑνΝelatoneo ‘to lessen’ (Paschal Chronicle, 37, 2; AD 7) ȬπΕΉ΍ΔϟΊΝeripizo ‘to ruin’ (Leo Grammaticus, 69, 1; AD 10)  ȬΔ΅ΎΘΉϾΝȱpaktevo ‘to make somebody come to terms’ (Paschal Chronicle, 726, 7; AD 7)

It is worth noting that quite a few of the causativisations were not preserved and were not grammatical in the following period, i.e. Modern Greek. This indicates that, while the direction regarding the addition of an external argument to the structures of anticausative verbs is consistent, the presence of the external agent does not exhibit the same consistency for all anticausative verbs. The verbs that do not accept consistently an external agent in the subject position usually denote change-of-state from an internal cause. Causativisation that was not preserved is the causativisation of the intransitive verbs ʌȜȠȣIJȓȗȦ plutizo ‘to grow rich’, ĮıșİȞȫ astheno ‘to become ill’, ıȚȖȫ sigho ‘to keep silence’ and țȚȞįȣȞİȪȦ kindhinevo: ‘to be in danger’. (7) ΔΏΓΙΘϾΑΝȱǻΔΏΓΙΘϟΑΝȱplutinoȱΔΏΓΙΘϟΊΝ plutizo) ‘to grow rich’ intransitive active45 ΑνΓΖǯǯǯȱȱ Έ΍Τȱȱ ΑΤȱȱ ΔΏΓΙΘϾΑϙȱ a. Ύ΅Ϡȱȱ ϳȱȱ ke o neos... dhia na plutini ȱ ȱ ȱ and the.NOM young.NOM for to grow-rich.ACT.PERFVE.3SG ‘and the young... for to grow rich’ (Leonardo Dellaporta, Poems, ǹ, 1991; AD 14-15) transitive active ΑΤȱȱ Ηξȱȱ ΔΏΓΙΘϟΗΝȱȱ ΗϾΑΘΓΐ΅ȱȱȱ b. ΪΑȱȱσΑȂȱȱΎ΅Ϡȱȱ ΌνΏϙΖȱȱ an en’ ke thelis na se plutiso sindoma if is and want.2SG to you.ACC.WEAK grow-rich.ACT.PERFVE.1SG soon ‘and you would like it so that I will soon make you rich’ (Leonardo Dellaporta, Poems, ǹ, 1985; AD 14-15) i in the previous period - Hellenistic-Roman period: intransitive active and non-active πΔΏΓϾΘ΋ΗΉΑȱ Θ΅ΛνΝΖȱ ΈϟΎ΅΍ΓΖȱ ЕΑȱ ȱ c. ΓЁΈΉϠΖȱ oudeis eploute:sen takheo:s dikaios o:n nobody.NOM grow-rich.ACT.AOR.3SG fast fair.NOM be.ACT.PRES.PART.NOM ‘nobody grew rich fast if s/he was fair’ (Menander, Fragments, 294, 1; 4-3 BC) d. ϵΘ΍ȱȱπΑȱΔ΅ΑΘϠȱȱȱȱȱȱȱπΔΏΓΙΘϟΗΌ΋ΘΉȱ ȱπΑȱ΅ЁΘХǰȱπΑȱΔ΅ΑΘϠȱΏϱ·УȱΎ΅ϠȱΔΣΗϙȱ·ΑЏΗΉ΍ȱ hoti en panti eploutisthe:te en auto:i, en panti logo:i kai pase:i gno:sei that in everything grow-rich.NACT.AOR.2PL in him in all speech and all knowledge ‘that in everything you grew rich in him, in all speech and all knowledge’ (New Testament, 1 Corinthians, 1, 5, 2; AD 1)

It is worth noting that in the following period (Modern Greek), the structure ijIJȫȤȣȞİ IJȚȢ ȠȚțȠȖȑȞİȚİȢ / ftoxine tis ikojenies / make-poor.ACT.PAST.PERFVE.3SG the.ACC families.ACC / ‘it made the families poor’ is grammatical, but the structure *ʌȜȠȪIJȚıİ IJȚȢ ȠȚțȠȖȑȞİȚİȢ / plutise tis ikojenies growrich.ACT.PAST.PERFVE.3SG the.ACC families.ACC / ‘it made the families rich’ is ungrammatical; in Medieval Greek, both causative structures were possible: (8)

a. Ύ΅ϠȱΦΑ΅ΌΙΐκΘ΅΍ȱȱ ΘξΖȱȱ ΆΓΏξΖȱȱȱ ϳΔΓϿȱȱȱȱȱȱȱȱȱȱ ΘϲΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ πΔΘΝΛΣΑ΅Α ke anathimate tes voles opu ton eptoxanan ȱ and remember.3SG the

dice.ACC which.NOM he.ACC.WEAK make-poor.ACT.PAST.PERFVE.3PL

‘and he remembers the dice which made him poor’ ǻSachlikis, Paraenetic poem, 169; AD 15-16) b. ΪΑȱσΑȂȱȱȱΎ΅Ϡȱȱ ΌνΏϙΖȱȱȱȱȱ ΑΤȱ Ηξȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΔΏΓΙΘϟΗΝȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΗϾΑΘΓΐ΅ȱȱȱ an en’ ke thelis na se plutiso sindoma if is and want.2SG to you.ACC.WEAK grow-rich.ACT.PERFVE.1SG soon ‘and you would like it so that I will soon make you rich’ (Leonardo Dellaporta, Poems, ǹ, 1985; AD 14-15) (9) ΦΗΌΉΑЗ astheno ‘to become ill’ intransitive active ·ϱΑ΅ΘΣȱȱ ΐΓΙȱȱ ωΗΌνΑ΋Η΅Αȱȱ ΦΔϲȱΘΤΖȱȱ ΐΉΘ΅ΑΓϟ΅Ζ a. ΘΤȱ ta ghonata mu isthenisan apo tas metanias  the.NOM knees.NOM my become-ill.ACT.PAST.PERFVE.3PL by the repentances ‘my knees were affected (became ill) by my acts of repentance’ (Ptochoprodromos, ǿV, 628; AD 11-12) 45

When nothing is mentioned, the examples come from Late Medieval texts (cf. note 40 in this chapter ).

124

Chapter Three b. ΦΔξȱ ΘχΑȱȱ ΔΏϛΒ΍Αȱȱ ΦΗΘνΑ΋ΗΉΑȱȱ Ύ΅Ϡȱȱ πΔνΌ΅ΑΉΑȱ ape tin pliksin astenisen ke epethanen from the boredom become-ill.ACT.PAST-PERFVE.3SG and died.3SG ‘he became ill and died from boredom’ (Leontios Machairas, Chronicle, 3, 463; AD 15) transitive active ΘΓΑǰȱȱ Έ΍ΤȱΑΤȱΈΙΑ΅ΐЏΗϙȱΘϲΑȱπΛΌΕϱΑȱΘΓΙȱ c. Έ΍ϱΘ΍ȱȱ ǻϳǼȱ̍ϾΕ΍ΓΖȱȱ ΦΗΘΉΑΉϧȱȱ dhioti (o) Kirios asteni ton, dhia na dhinamosi ton exthron tu because (the) Lord.NOM become-ill.ACT.PRES.3SG he.ACC.WEAK for to strengthen.3SG the enemy.ACC his

‘because Lord makes him ill, to strengthen his enemy’ (Leontios Machairas, Chronicle, 3, 421; AD 15) i in the previous period - Hellenistic-Roman period: intransitive active d.ȱȱ ΓЇΑȱȱȱȱȱȱȱόΎΓΙΗΉΑȱȱϵΘ΍ȱȱΦΗΌΉΑΉϧǰȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘϱΘΉȱΐξΑȱȱȱσΐΉ΍ΑΉΑȱπΑȱСȱȱȱȱȱώΑȱȱȱΘϱΔУȱΈϾΓȱψΐνΕ΅Ζȱ oun e:kousen hoti asthenei, ȱ tote men emeinen en ho:i e:n topo:i dyo he:meras therefore heard.3SG. that become-ill.ACT.PRES.3SG then prt stayed.3SG in which was.3SG place two days

‘when therefore he heard that he was sick, he stayed two days in the place where he was’ (New Testament, John, 11, 6, 3; AD 1) (10) Η΍·З sigho ‘to keep silence’ intransitive active a.ȱ Ύ΅ϠȱΗϟ·΋ΗΉȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ϠȱȱΚϾΏ΅ΒΉǰȱȱȱȱȱȱȱΘϟΔΓΘΉȱȱȱȱȱȱȱȱȱΐχȱȱȱΐΓІȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΉϥΔϙΖȱȱ ke sijise ke filakse, tipote mi mu ipis and keep-silence.ACT.IMP.2SG and be-careful.2SG nothing.ACC not I.GEN.WEAK say. ACT.2SG.

‘and keep silence and be careful and do not say anything to me’ (The War of Troy, 6610; AD 14) transitive active b. ·ΓΕ·ϲΑȱȱȱȱȱΌξΏΉ΍ȱȱȱȱȱΘχΑȱȱȱȱȱȱȱȱȱΌΏϧΜ΍ΑȱȱȱȱȱȱΘ΋ΖȱȱȱΗ΍·φΗΉ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ϠȱȱȱΘϲΑȱȱȱȱȱȱȱΔϱΑΓΑ ghorghon theli tin thlipsin tis sijisi ke ton ponon fast

want.3SG the.ACC sorrow.ACC her

keep-silence.ACT.PERFVE.3SG and the.ACC pain.ACC

‘she wants to suppress her sorrow and her pain’ (The War of Troy, 5759; AD 14) i in the previous period - Hellenistic-Roman period: intransitive active and non-active Έξȱ ΔκΑȱ Θϲȱ ΔΏϛΌΓΖ c. ̳Ηϟ·΋ΗΉΑȱ Ǽsige:sen de pan to ple:thos keep-silence.ACT.AOR.3SG prt all.NOM the.NOM multitude.NOM ‘all the multitude kept silence (and they listened to Barnabas)’ (New Testament, Acts, 15, 12, 1; AD 1)

The verb ĮȡȖȫ argho ‘to be late’ is causativised in Medieval Greek; the causative structure, however, is limited only to sentences with clitics as direct objects in Modern Greek. Of interest is the opposite course of the synonymous verb țĮșȣıIJİȡȫ kathistero ‘to be late/delay’ of which there is no evidence of systematic transitive use in Medieval Greek (in the era, namely, in which the verb ĮȡȖȫ argho is causative) but which is causativised during the Modern Greek period. (11) ΅Ε·Џ argho ‘to be late’ intransitive active ϴΏϟ·ΓΑȱȱ ϳȱȱ ̖ΕЏϞΏΓΖ a.ȱ ΪΑȱȱ ΩΕ·΋ΗΉΑȱȱ an arjisen olighon o Troilos ȱȱȱȱ if be-late.ACT.PAST.PERFVE.3SG little the.NOM Troilos.NOM ‘if Troilos was a little late’ (The War of Troy, 6071-2; AD 14) transitive active b. ΔΓΏΏΣȱȱΆ΅ΕϿΑȱȱΘΓϿΖȱȱΚ΅ϟΑΉ΍ǰȱȱΈ΍΅ΘϠȱΘϱΗ΅ȱΘΓϿΖȱȱȱȱȱȱȱȱΩΕ·΋ΗΉΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱϳȱΔϱΏΉΐΓΖȱȱ pola varin tus feni, dhiati tosa tus arjisen o polemos much burdensome them seem.3SG as so they.ACC.WEAK be-late.ACT.PAST.PERFVE.3SG the war.NOM

‘it seemed very burdensome (to the Greek nobles), as the war (of Troy) made them delay’ (The War of Troy, 5963-4; AD 14) i in the previous period - Hellenistic-Roman period: intransitive active  ΐξΑȱ ȱȱ·ΤΕȱΦΕ·ΓІΗ΍ǰȱ ȱΐχȱ ΆΓΙΏϱΐΉΑ΅΍ȱ ΔϱΕΕΝȱ ΔΏ΅ΑκΗΌ΅΍ȱ c. ΅ϡȱ   hai men gar argousi, me: boulomenai porro: planasthai the.NOM one.NOM as be-late.ACT.PRES.3PL not will.PART.NOM far wander.INF ‘as these delay, not willing to wander far away’ (Plutarch, Causes of Natural Phenomena, 917c, 2; AD 1-2) (12) Ύ΅ΌΙΗΘΉΕЗ kathistero ‘to be late/delay’ intransitive active

Diachronic Data and Analysis

125

a. ΘΓϿΖȱȱ ΐξΑȱȱ ΔΕΉΗΆΉΙΘΤΖȱȱ πΒȱ΅ЁΘϛΖȱȱ ΔνΐΔΉ΍Αȱȱ Έ΍΅Η΅ΚφΗΓΑΘ΅Ζǰȱȱ tus men prezveftas eks aftis pembin dhiasafisondes, the.ACC prt commissioners.ACC from her send.INF make-clear.PART.ACCȱ ϣΑ΅ȱȱ ΐχȱȱ Ύ΅ΌΙΗΘΉΕЗΗ΍ȱȱ ΘЗΑȱȱ Ύ΅΍ΕЗΑ ina mi kathisterosi ton keron to not be-late.ACT.IMPRFVE.3PL the.GEN circumstances.GEN ‘to send the commissioners from her to make everything clear, so that they will not delay because of the circumstances’ ǻConstantine VII Porphyrogenitus, de legationibus, 311, 12; AD 10)ȱ i in the previous period - Hellenistic-Roman period: intransitive active ЀΔΣΘΓΙΖȱ ΦΑ΅Ύ΅ΏΉΗΣΐΉΑΓ΍ȱȱΔ΅ΕЏΕΐΝΑȱȱ σΛΉΗΌ΅΍ȱȱ ΘΓІȱ Ύ΅΍ΕΓІȱ b. ΘΓϿΖȱ  tous hypatous anakalesamenoi ȱ paro:rmo:n ȱ ekhesthai tou kairou the.ACC consuls.ACC call.PART ȱ act.PART have.INF the.GEN time.GEN   Ύ΅Ϡȱ ΐχȱ Ύ΅ΌΙΗΘΉΕΉϧΑ  kai me:  kathysterein and not be-late.ACT.PRES.INF ‘the Senate called upon the Consuls to act at once and see that they did not delay’ (Polybius, Histories, 27, 6, 4; 3-2 BC)

Of particular interest is the rare and non-systematic use of ijİȪȖȦ fevgho ‘to flee/go away/leave’ in transitive structures (Apostolopoulos 1984). The meaning of ijİȪȖȦ fevgho ‘to flee/go away/leave’ in these instances corresponds to the meaning of the verb ĮʌȠijİȪȖȦ apofevgho ‘to avoid’ (with the prefix Įʌȩ- apo-). It is not concerned, therefore, with causative use but with transitive use in which the object has the ș-role of the theme (change-of-state or total affectedness of the object is not denoted). This transitive use has been abandoned in Modern Greek, where this meaning is consistently denoted by the verb ĮʌȠijİȪȖȦ apofevgho ‘to avoid’ but is preserved in the dialect of Ȃodern Cypriot (Lavidas & Therapontos 2006): (13) a. Early Medieval Greek ȱ



ȱ

ΚϾ·ΝΐΉΑȱȱ

ΘΤΖȱȱ

ΎΓΗΐ΍ΎΤΖȱȱ

ϳΐ΍Ώϟ΅Ζ

fighomen tas kozmikas omilias go-away.ACT.PERFVE.1PL the.ACC earthly.ACC discussions.ACC ‘to avoid the earthly discussions’ (John Moschos, Leimonarion (The Spiritual Meadow), CLIX, A, 12; AD 6-7) Late Medieval Greek b. ΚϾ·Νȱȱ ΘΓϿΖȱȱ ϷΚΉ΍Ζ   figho tus ofis go-away.ACT.PERFVE.1SG the.ACC snakes.ACC ‘to avoid the snakes’ (Kallimachos and Chrysorroi, 228; AD 14) c. ΚϾ·Νȱȱ ΘϲΑȱȱ ΔϱΑΓΑ  figho ton ponon go-away.ACT.PERFVE.1SG the.ACC pain.ACC ‘to avoid the pain’ (Kallimachos and Chrysorroi, 712; AD 14) d. ΚΉϾ·ΓΙΗ΍Αȱȱ ΘϲΑȱȱ ΎϱΔΓΑ fevghusin ton kopon go-away.ACT.PRES.3PL the.ACC effort.ACC ‘they avoid the effort’ (Leonardo Dellaporta, Poems, ǹ, 137; AD 14-15) e. ϳΔΓϿȱȱȱȱȱȱȱΉϢΖȱȱΘϲΑȱΎϱΗΐΓΑȱȱȱΔΉΕΔ΅ΘΓІΑȱΑΤȱȱΚΉϾ·ΓΙΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘνΘΓ΍΅ȱȱȱȱȱȱȱΔΣΌ΋ȱ opu is ton kozmon perpatun na fevghun tetia pathi wherever in the earth walk.3PL to go-away.ACT.IMPFVE.3PL such.ACC passions.ACC ‘wherever in earth they walk, let them avoid such passions’ (Leonardo Dellaporta, Poems, ǹ, 1537; AD 14-15)

 The following alternations in the expression of the verb complement show precisely that it does not concern a causative verb but a transitive verb with the meaning ‘to avoid’. (14) a. ΦΔϲȱȱ ΘχΑȱȱ ΛЏΕ΅ȱȱ ΚΉϾ·Ή΍ apo tin xora fevji from the country go-away.ACT.PRES.3SG ‘he leaves the country’ (Sachlikis, Paraenetic poem, 98; AD 15-16) ΦΔϲȱȱ ΘΤȱ ΈЗΕ΅ȱ ȱȱ b. ΑΤȱȱ ΚΉϾ·ϙȱȱ na fevji apo ta dhora to go-away.ACT.IMPFVE.3SG from the gifts ‘to flee away from the gifts’ (Leonardo Dellaporta, Poems, ǹ, 1098; AD 14-15)

126

Chapter Three c. ΘξΖȱȱ ψΈΓΑξΖǯǯǯȱȱ ΔЗΖȱȱ ΑΤȱΘξΖȱȱ πΒΝΚϾ·΋Dzȱȱȱȱ tes idhones... pos na tes eksofiji? ȱ the.ACC pleasures.ACC how to them.ACC.WEAK EKSO-go-awayǯǯǯř ȱ ‘pleasures, how to flee away from them?’ (Leonardo Dellaporta, Poems, ǹ, 1504; AD 14-15) ΌΣΑ΅ΘΓΑȱȱ ΅ΔϱΚΉΙ·ΉΑȱȱ ΘΓΙȱΗЏΐ΅ΘΓΖȱȱȱΉΎΉϟΑ΋ȱȱȱ d. ΘΓΑȱȱ ton thanaton apofevjen tu somatos ekini the.ACC death.ACC APO-go-away.ACT.PAST.IMPRFVE.3SG the body.GEN she.NOM ‘she was avoiding the death of the body’ (Apollonius of Tyre, The story of Apollonius (Kechagioglou), 542; AD 14)

(b) Causative types from intransitive non-active Psaltes (1913/1974) notes the following causativisations (innovative active causative type based on an earlier non-active anticausative type) in the Byzantine Chronicles: i Ύ΅ΌνΊΓΐ΅΍ȱkathezome sit.NACT ĺȱtransitiveȱΎ΅ΌνΊΝȱkathezo sit.ACT (Doukas, 166, 19; AD 15) Early Medieval Greek i ΎΓΙΕΣΊΝȱkurazo tire.ACTȱȱȱ ȱ ΉΎΓϾΕ΅ΗΉȱȱ ΔΣΑΘ΅Ζ ekurase pandas tire.ACT.PAST.PERFVE.3SG all.ACC (Theophanes Confessor, Chronicle, 448, 9; AD 8-9)46 i ȱ΅·΅Α΅ΎΘΓϾΐ΅΍aghanaktume exasperate.NACT ĺ transitive ΅·΅Α΅ΎΘЏ aghanakto exasperate.ACT (Theophanes Confessor, Chronicle, 186, 16; AD 8-9)

It seems, however, that in the majority of cases causativisation occurs on the basis of an active intransitive verb and that in a few cases it begins from a non-active intransitive. Evidently, this specific type of change correlates with the way the structure is reanalysed; a structure with an active intransitive verb is more easily reanalysed as a transitive structure (it needs only the analysis of the DP-subject as DP-object) than a structure with a non-active verb; in the second case, therefore, the anticausative needs initially to be analysed by the speakers as a passive structure for the presence of the agent to be considered possible (see section 4.1). 3.3.2.2 Factors and stages of causativisation The role of prefixes is also of particular interest in Medieval Greek. The addition of a prefix results, in many cases, in the availability of transitive syntactic behaviour of a verb that is intransitive without the prefix. In many instances, causative verbs are formed with the addition of a prefix in the Medieval Greek period. patient arguments can also be combined with prefixed causatives, many times, as a clitics47 (see chapter 4). (15) a. Early Medieval Greek ϳȱΈξȱȱȱ̇΍ϱΗΎΓΕΓΖȱȱȱȱΔΕΉΗΆΉϾΗ΅ΖȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ΘΉΔΕΣΙΑΉΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ΅ЁΘϱΑ

 o dhe Dhioskoros 



the

prezvefsas

kateprainen

afton 





Dioskorus.NOM act-as-ambassador.PART.NOM KATA-assuage.ACT.PAST.PERFVE.3SG he.ACC.WEAK 

 ‘Dioskorus acted as ambassador and assuaged him’ (Theophanes Confessor, Chronicle, 252B, 15; AD 8-9)   Late Medieval Greek b. ΔνΑΘΉȱ ΘχΑȱȱ πΔ΅ΕνΘΕΉΛΓΑȱȱ ΦΔȂȱЙΈΉȱȱ Ύ΅ϠȱΦΔȂȱȱπΎΉϧΌΉΑ pende tin eparetrexon ap’ odhe ke ap’ ekithen five she.ACC.WEAK PARA- run.ACT.PAST.IMPFVE.3PL from here and from there ‘five men were chasing her here and there’ (Digenis Akritas, Ǽscorial, 1497; AD 12) c. π·Аȱȱ ΑΤȱ·νΑΝȱȱ ȱΘϾΕ΅ΑΑΓΖȱȱ ȱΑΤȱΘϲΑȱȱ Ύ΅Θ΅Δ΍ΎΕΣΑΝȱȱȱȱ ȱ egho na jeno tiranos na ton katapikrano I.NOM to become.1SG tyrant.NOM to he.ACC.WEAK KATA-grieve.ACT.PERFVE.1SG ‘to become tyrant and cause him grief’ (Livistros and Rodamni, 274; AD 14)

The connection of the prefixes with causativisation cannot, however, be direct since causative uses are 46

47

Chatzidakis includes ȟİțȠȣȡȐȗȠȝĮȚ ksekurazome rest.NACT ĺ ȟİțȠȣȡȐȗȦ ksekurazo rest.ACT in the Modern Greek causativisations; it appears, however, that the causativisation of țȠȣȡȐȗȦ kurazo tire.ACT (as well as ȟİțȠȣȡȐȗȦ ksekurazo rest.ACT) is a Medieval causativisation. Cf. the presence of clitic in the first instances of causativisation (ȝĮȢ IJȡȑȤİȚ / mas trexi / we.ACC.WEAK run.ACT.PRES.3SG).

Diachronic Data and Analysis

127

witnessed with the same verbs without the prefix: (16) a. ΧΔ΅Βȱȱ ψΐκΖȱȱ Δ΍ΎΕ΅ϟΑΉ΍ȱȱȱȱȱ  apaks imas pikreni once we.ACC grieve.ACT.PRES.3SG ‘once only he grieves us’ (Michael Glykas, 100; AD 12) b. Ύ΅Ϡȱȱ ΘϠȱȱ Δ΍ΎΕ΅ϟΑΉ΍Ζȱȱ ΘχΑȱȱ πΐχΑȱȱ Ύ΅ΕΈϟ΅Αȱȱȱ ke ti pikrenis tin emin kardhian and why grieve.ACT.PRES.2SG the.ACC my heart.ACC ‘and why you grieve my heart’ (Florios and Platziaflora, 1653; AD 15)

The fact that prefixes and causativisation are not directly connected (the presence of a prefix does not always mean causative interpretation and structure) is also indicated by the presence of intransitive verbs with prefixes. The addition of the prefix can result in either a transitive or intransitive verb: (17) a. Ύ΅Ϡȱȱ Θϲȱȱ ΗΔ΅ΌϟΑȱȱ ΘΓΙȱȱ πΒνΔΉΗΉΑȱ ke to spathin tu eksepesen and the.NOM sword.NOM he.GEN EK- fall.ACT.PAST.PERFVE.3SG ‘and his sword fell’ (Digenis Akritas, Ǽscorial, 1272; AD 12) ȱȱȱȱȱȱ b. ΛΌξΖȱȱ πΔ΅Ε΅ΆΕ΅Έ΍ΣΗΘ΋ΐ΅Αȱȱ Ύ΅Ϡȱȱȱ πΐΉϟΑ΅ΐΉΑȱΦΔЗΈΉ xthes eparavradhiastiman ke eminamen apodhe yesterday PARA- be-overtaken-by-night.NACT.PAST.PERFVE.1PL and stayed.1PL here ‘we were overtaken by night and stayed here’ (Digenis Akritas, Ǽscorial, 338; AD 12)

The use of cognate objects (namely of DPs that have the same root as the verb) for intransitive verbs is more closely connected with the causativisations in Medieval Greek (cf. relative observations for Classical Ancient and Hellenistic-Koine, paras. 3.1.2.3 and 3.2.2.2). It possibly concerns the first stage of the causativisation of intransitive verbs. Andriotis (1937) notes the use of cognate objects with intransitive verbs; usually the cognate object is modified by an adjective and can precede or follow the verb: (18) non-active morphology ΓЁΈξΑȱȱ ΎΓ΍ΐΓІΐ΅΍ a. ЂΔΑΓΑȱȱ ipnon udhen kimume sleep.ACC not sleep.NACT.PRES.1SG ‘I cannot sleep’ (Digenis Akritas, Ǽscorial, 843; AD 12) Θ΅ϾΘ΅Ζȱȱ πΑΘ΅ϾΌ΅ȱΘΤΖȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎΓΏΣΊΓΐ΅΍ b. ΘΤΖȱΘΕΉϧΖȱΎΓΏΣΗΉ΍Ζȱȱ tas tris kolasis taftas entaftha tas kolazome the three punishments.ACC these.ACC here they.ACC.WEAK be-punished.NACT.PRES.1SG ‘I am punished here with three punishments’ (Ptochoprodromos, IV, 149; AD 11-12) c.ȱȱ Ύ΅Ϡȱȱ Ύ΅ΘȂȱȱ ΅ЁΘϛΖȱȱ ΌΙΐЏΑΉΘ΅΍ȱȱ ΌΙΐϱΑȱȱ πΒ΅·Ε΍ΗΐνΑΓΑ ke kat’ aftis thimonete thimon eksaghrizmenon and against she.GEN become-angry.NACT.PRES.3SG anger.ACC furious.ACC ‘and he becomes furiously angry against her’ (Livistros and Rodamni, 3333; AD 14) d.ȱ ΦΔϲȱȱ πΑΘΕΓΔϛΖȱȱ ΘΓΙȱ ΐ΅ϟΑΉΘ΅΍ȱȱ ΐ΅Αϟ΅Αȱȱ ΦΏΏΤȱȱ ΐΉ·ΣΏ΋Αȱ apo endropis tu menete manian ala meghalin from shyness his rage.NACT.PRES.1SG rage.ACC but intense.ACC ‘from shyness his rage becomes intense fury’ (Livistros and Rodamni, 2072; AD 14) e.ȱ ΏΙΔΉϧΘ΅΍ȱȱ ΏϾΔ΋Αȱȱ ΦΐΉΘΕΓΑȱȱȱ ȱ lipite lipin ametron feel-sorrow.NACT.PRES.3SG sorrow.ACC excessive.ACC ‘and he feels excessive sorrow’ (The Story of Achilles, 602; AD 15) active morphology ΐ΅Ε΅ΐϱΑ  f. Ύ΅Ϡȱȱ ΑΤȱȱΐ΅Ε΅ϟΑϙȱȱ ke na mareni maramon and to wither.ACT.PERFVE.3SG withering.ACC ‘to wither’ (Sachlikis, Paraenetic poem, 1502; AD 15-16) g. Ύ΅Ϡȱ ΑΤȱ ΐ΅Ε΅ϟΑϙȱ ΐ΅Ε΅ΐϲΑȱ ϳΕ΍ΗΘ΍ΎϲΑ ke na mareni maramon oristikon and to wither.ACT.PERFVE.3SG withering.ACC permanent.GEN ‘to wither totally’ (Florios and Platziaflora, 1515; AD 15) with a compound object h. ΐξȱ ϵΏ΋Αȱȱ ΘχΑȱΈ΍΅ΎΕΣΘ΋Η΍ΑȱϵΔΓΙȱȱ ΎΕ΅ΘΉϧȱȱ ȱ ψȱȱ ̏ΓΌЏΑ΋ me olin tin dhiakratisin opu krati i Mothoni

128

Chapter Three with all the rule.ACC which rule.ACT.PRES.3SG the.NOM Mothoni.NOM ‘with all the property which Mothoni rules’ (Chronicle of Moreas, ȇ, 2855; AD 14) in a passive structure i. ΎΏЗΗΐ΅ȱȱ ΘϛΖȱȱ ΘϾΛ΋Ζȱȱ πΎΏЏΗΌ΋ȱȱ ΐΓΙȱ   klosma tis tixis eklosthi mu   spin.NOM the.GEN fate.GEN spin.NACT.PAST.PERFVE.3SG I.GEN.WEAK ‘the turns of fate have for me been spun’ (Kallimachos and Chrysorroi, 468; AD 14)48 

3.3.2.3 Limitations on causativisations The causativisation of intransitive verbs is not without restrictions and is not completely free for all intransitive verbs, and especially for the verbs denoting change from internal cause49: (19) ΏΣΐΔΝ lambo ‘to shine’50 intransitive active Θϲȱȱ ΔΕϱΗΝΔϱΑȱȱΘΓΙ a.ȱ Ύ΅Ϡȱȱ БΖȱ ȱϊΏ΍ΓΖȱȱσΏ΅ΐΔΉΑȱȱ   ke os ilios elamben to prosopon tu  and as sun shine.ACT.PAST.IMPRFVE.3SG the.NOM face.NOM his ‘and his face was shining like the sun’ (Digenis Akritas, Ǽscorial, 612; AD 12) b.ȱ ϢΈΓϿȱ Ύ΅΍ΕϲΖȱȱ ·ΏΙΎϾΘ΅ΘΓΖȱȱ ΘХȱ ̍΅ΏΏ΍ΐΣΛУȱȱ ΏΣΐΔΉ΍ȱȱȱ idhu keros ghlikitatos to Kalimaxo lambi here occasion.NOM sweetest.NOM the.DAT Kalimaxos.DAT shine.ACT.PRES.3SG ‘the sweetest occasion shines for Kalimachos’ (Kallimachos and Chrysorroi, 1848; AD 14)

Concerning the verb ĮȞșȫ antho ‘to blossom’, Psaltes (1913/1974) and Apostolopoulos (1984) indicate only one example of causative use (which was later abandoned): (20)ȱ Θϲȱȱ ΕϱΈΓΑȱȱ ΦΑΌΉϧȱȱ ΘχΑȱȱ ΛΣΕ΍Α   to rodhon anthi tin xarin   the.NOM rose.NOM blossom.ACT.PRES.3SG the.ACC grace.ACC  ‘the rose makes its grace blossom’ (Kallimachos and Chrysorroi, 1593; AD 14)



The interpretation of the example, of course, can also be as follows: ‘it blossoms because of its grace’ (Apostolopoulos’s translation is: ‘la rose fleurit la grâce’). (21) intransitive active (particularly systematically) ΘΓІȱ ΔϱΌΓΙȱ Θϲȱ ΎΏ΅ΈϠΑȱ ȱ a. ΦΑΌΉϧȱȱ anthi tu pothu to kladhin blossom.ACT.PRES.3SG the.GEN passion.GEN the.NOM branch.NOM ‘the branch blossoms because of the passion’ (Livistros and Rodamni, 3633; AD 14)ȱ b. ΦΚΓІȱȱ ǻΘϲȱΈνΑΈΕΓΑǼȱȱ ΦΑΌφΗϙȱȱ afu (to dhendron) anthisi after (the tree) blossom.ACT.PERFVE.3SG ‘since the tree blossoms’ (Agapios Landos, Geoponikon, 28, 3-4; AD 17)ȱ

The non-availability of causativisation also appears to be due to the presence of DP in the accusative for the denotation of the cause (not of the patient) with many of the intransitive verbs of internal change that are not causativised (cf. 4.1.1). Consequently, it does not appear that the main element for the presence of transitivity alternation is the presence of an external cause of the change-of-state (we have located some verbs of change-of-state from internal cause that are causativised) but the absence of DP in the accusative (which usually denotes the cause). (22) a. intransitive verb + DP-cause in accusative case Θ΋Αȱ ȱ ȱ π·νΏ΅ΗνΑȱȱ ejelasen tin laugh.ACT.PAST.PERFVE.3SG she.ACC.WEAK ‘he laughed at her’ (Sachlikis, Paraenetic poem, 72-73; AD 15-16) 48 49

50

For more examples, cf. Lavidas 2006a. Beth Levin (personal communication) points out that every development in transitivity alternations and every process of causativisation has to be analysed in relation to the semantic characteristic of “change from internal or external cause”. Intransitive active and non-active in Classical Greek ĺ intransitive active in Hellenistic and Medieval.

Diachronic Data and Analysis ȱȱ

129

b. only the intransitive structure is possible ·ΉΏκȱȱ

ψȱȱ

Ύ΅ΕΈ΍Σȱȱ

Θ΋Ζȱ

jela i kardhia tis laugh.ACT.PRES.3SG the.NOM heart.NOM her ‘her heart laughs’ (Sachlikis, Paraenetic poem, 369; AD 15-16)

We have ascertained, consequently, that in Greek, all types of accusative (transitive non-causative) verbs are not causativised, in contrast to languages that have causative morphemes at their disposal and can also causativise accusative verbs. For example, the verb ĮȞĮȝȑȞȦ anameno ‘to wait’ is only transitive and is not causativised (it does not take a DP-patient in the accusative case as the direct object) with the meaning ‘make someone wait’ (in Levin’s terms: the extension of the event schema of the specific verb is not possible): (23) Άϟ·Ώ΅Ζȱ vighlas

τΗΘ΋ΗΉΑȱ

ΔΓΏΏΤΖȱȱȱ Ύ΅Ϡȱ ΦΑ΅ΐνΑ΅ΗϟΑȱȱ

ΘΓΑȱȱȱȱȱ

estisen

polas

anamenasin

ton

wait.ACT.PAST.IMPRFVE.3PL

he.ACC.WEAK

ke

guards.ACC put.PAST.PERFVE.3SG many.ACC and

‘he put many guards and they were waiting for him’ (Digenis Akritas, Ǽscorial, 1029; AD 12)

3.3.2.4 Changes in accusative verbs: sub-categories As regards the changes concerning the accusative verbs of Medieval Greek, the rules of case assignment gradually change following the wider tendency of assignment of accusative case by all the verbs, causative and non-causative. In the post-Classical period (Catsimali 1990), uncertainty concerning the use of the genitive and the accusative had already emerged. Browning (1983/1991) argues that: (a) from the early 5th century, in the texts of lives of saints, uncertainty concerning the use of the dative is observed; (b) 5th-7th century: ‘question’ verbs assign accusative case to their object, whereas for many of the dative meanings, the genitive case is used; (c) 7th-9th century: in the Chronicle of Theophanes, the verb DŽþȖğȁ eteo ‘to ask for’ takes objects in the dative case; (d) 9th-11th century: in Constantine VII Porphyrogenitus, the verbs ǰȋDŽȖğȁ krateo ‘to govern’ and ǰȝȋǬǗħȁ kirievo ‘to dominate’ take objects in the accusative case. Similarly, Psaltes (1913/1974) observes that the following verbs were required to take DPs in the dative or genitive case in Classical and post-Classical Greek but were required to take DPs in the accusative case (the default case for direct objects) in the Byzantine chronicles: dative ĺ accusative ΆΓ΋ΌЗ voitho ‘to help’ ΔΓΏΉΐЗ polemo ‘to fight’ ΛΕЗΐ΅΍ȱxrome ‘to use’ genitive ĺ accusative ΦΔΓΏ΅ΑΌΣΑΓΐ΅΍ȱapolanthanome ‘to forget’ȱ ΚΉϟΈΓΐ΅΍ fidhome ‘to spare’ȱ ΛΝΕϟΊΓΐ΅΍ xorizome ‘to separate’ȱ ȱ ȱ ΛΝΕϟΊΓΐ΅΍ȱȱ ΘΓϿΖȱȱ ΦΈΉΏΚΓϾΖȱȱȱȱ ȱ xorizome tus adhelfus separate.NACT.PRES.1SG the.ACC brothers.ACC ‘I part from my brothers’ (Kallimachos and Chrysorroi, 2561; AD 14) ȱ

ΘχΑȱȱ

ΗΙΐΆΓΙΏχΑȱȱ

ΦΔΉΏ΅Όϱΐ΋Αȱȱȱȱ

ȱ

tin simvulin apelathomin the.ACC advice.ACC forget.NACT.PAST.PERFVE.1SG ‘I forgot the advice’ (Kallimachos and Chrysorroi, 109; AD 14))

In Digenis Akritas (12th century), an almost total absence of the dative case is ascertained. The accusative and the genitive case are interchangeable in indirect objects (both in DPs and clitics): (24) a.ȱ ΘΓϿΖȱȱ Φ·ΓϾΕΓΙΖȱȱ ΘΓΙȱ σΏΉ·ΉΑ    tus aghurus tu elejen    the.ACC young.ACC his say.ACT.PAST.IMPRFVE.3SG ‘he was saying to the young’ (Digenis Akritas, Ǽscorial, 1, 499; AD 12) b.ȱ ΘЗΑȱȱ ΦΈΉΏΚЗΑȱȱ Θ΋Ζȱȱ σΏΉ·ΉΑȱȱȱȱȱ ȱ ton adhelfon tis elejen



130

Chapter Three the.GEN brothers.GEN her say.ACT.PAST.IMPRFVE.3SG ‘she was saying to her brothers’ (Digenis Akritas, Ǽscorial, 1, 390; AD 12)

3.3.3 Derivation of transitivity alternations and voice: absorption of the accusative case by the non-active voice and tendency towards active causatives and anticausatives 3.3.3.1 Voice in Medieval Greek: an overview The changes in the area of voice, as shown by their first emergence during the Hellenistic-Roman period, continued into the Medieval Greek period. Non-active voice appears with intransitive verbs for which we would expect active voice, but the active also continues to replace the non-active voice in instances of earlier transitive non-active verbs that denote interest of the subject for the verb action (benefactive meaning). Wolf (1911-1912) indicates the use of non-active forms instead of active for intransitive verbs in Malalas and Theophanes: ΚΌΣΑΓΐ΅΍fthanome arrive.NACT (Theophanes Confessor, Chronicle, 378, 14; AD 8-9)

Apostolopoulos (1984) interprets the presence of non-active voice morphology in the novel Kallimachos and Chrysorroi as denotation of the verb action with particular emphasis. ΐΓ΍ΕΓΏΓ·ΉϧΘ΅΍ȱmirolojite lament.NACT.PRES.3SG(Kallimachos and Chrysorroi, 2360; AD 14) ΒνΑΓΑΘ΅΍ȱksenonde be-removed.NACT.PRES.3PL (Kallimachos and Chrysorroi, 267; AD 14)ȱ

At the same time, however, in the same text similar structures with active voice morphology and use of a clitic which expresses the person who experiences the emotion are in evidence: (25) a. ΌΕ΋ΑφΗΝǰȱȱ ΎΏ΅ϾΗΓΐ΅ϟȱȱ ΗΉȱȱ thriniso, klafsome se  mourn.ACT.FUT.1SG cry.NACT.FUT.1SG you.ACC.WEAK ‘I will mourn and cry for you’ (Kallimachos and Chrysorroi, 1142; AD 14) b. Ύ΅Ϡȱ ΎΏ΅ϾΗΝȱȱ ΐΓΙȱȱ ΘχΑȱȱ ΘϾΛ΋Αȱȱȱȱȱȱȱȱȱȱȱ ȱȱȱȱȱȱȱȱȱȱȱ ke klafso mu tin tixin and cry.ACT.FUT.1SG I.GEN.WEAK the.ACC fate.ACC ‘and I will cry for my fate’ (Kallimachos and Chrysorroi, 706; AD 14)

On the other hand, in the Medieval Greek documents of South Italy and Sicily (Minas 1994), which were written between the 10th and 15th centuries, non-active verb types instead of active for transitive verbs are in evidence51: (26) a. Έ΍ΈϱΐΉΌ΅ȱȱ Ύ΅Ϡȱȱ ΘχΑȱȱ ΎΓΙΏΘΓϾΕ΅Αȱȱ dhidhometha ke tin kulturan give.NACT.PRES.1PL and the.ACC culture.ACC ‘we also give them the culture’ (ȉ, 117, 1121)52 ΧΔΉΕȱȱ Ύ΅ΘΉ·Ε΅ΜΣΐ΋Α b. ΘΤȱȱ ta aper kateghrapsamin the.ACC which.ACC describe.NACT.PAST.PERFVE.1SG ‘these which I described’ (GM, 60, 1123)53 Θϲȱȱ Θϟΐ΋ΐ΅ȱȱ ȱ c. πΏ΅ΆϱΐΉΌ΅ȱȱ elavometha to timima take.NACT.PAST.PERFVE.1PL the.ACC payment.ACC ‘we took the payment’ (ȉ, 205, 1155)

Minas contradistinguishes the aforementioned tendency with the use of active instead of non-active 51

52

53

Minas indicates the instability in voice morphology which is observed during this period: (a) strong presence of non-active endings in the position of active, a phenomenon which appears to spread widely in the postClassical and Medieval Greek periods, (b) many non-active verbs (already from the period of Hellenistic Koine) were turned into active. T ÆTrinchera Francisco, Syllabus graecarum membranarum, Sala Bolognese: Arnado Forni 1978 (first edition: Napoli 1865). GM Æ André Guillou, Les actes grecs de S. Maria di Messina. Enquête sur les populations grecques d’ Italie du Sud et de Sicile (XIe – XIVe s.), Palermo: Istituto siciliano di studi bizantini e neoellenici, 1963.

Diachronic Data and Analysis

131

forms for transitive verbs again in Medieval Greek documents of South Italy and Sicily (examples 27ae). Evidently, it is concerned with the completion of the change from non-active transitive verbs (absorption of the benefactive by the non-active ending) into active transitive verbs, which had begun in the Hellenistic-Roman era. This change must also be connected with the non-availability of the assignment of accusative case by non-active morphology, which emerges in the Medieval Greek period. Psaltes (1913/1974) also locates relative changes in the Byzantine chronicles (cf. examples in 28). (27) non-active ĺ active ΘϲΑȱȱ ΦΈΉΏΚϱΑȱȱ ΐΓΙ a. ϣΑ΅ȱȱ πΔ΍ΐΉΏφΗϙȱȱ ina epimelisi ton adhelfon mu to take-care.ACT.PERFVE.3SG the.ACC brother.ACC my ‘to take care of my brother’ (ȉ, 46, 1050) b. ΔΕϲȱ ΔΣΑΘΝΑȱȱ ΘЗΑȱȱ ΩΏΏΝΑȱȱ ΉϢΖȱȱ Θ΅ІΘ΅ȱȱ πΔΕΓΌϾΐ΋Η΅ pro pandon ton alon is tafta eprothimisa than all the other in these volunteer.ACT.PAST.PERFVE.1SG ‘I volunteered for these more than anything else’ (C, 4, 1097)54 c. ΘΤȱȱ ΛΝΕΣΚ΍΅ȱȱ ΘΤȱȱ ΧΔΉΕȱȱ ΛΘЗΐΉΑȱȱ ȱ ȱ ta xorafia ta aper xtomen the.ACC fields.ACC the.ACC which.ACC acquire.ACT.PRES.1PL ‘the fields which we acquired’ (C, 295, 1139) d. ϵΗΘ΍Ζȱȱ ΩΑȱȱ πΔ΅Α΅ΗΘϜȱȱ Ύ΅Ϡȱȱ ΆΓϾΏϙȱȱ ȱ ostis an epanasti ke vuli whoever.NOM prt rebel.3SG and want.ACT.PRES.3SG ‘whoever may rebel and want’ (ȉ, 105, 1116) e. ΔΣΑΘΉΖȱȱ ϳΐΓΌΙΐ΅ΈϲΑȱȱ ΗΎνΜΓΙΗ΍Αȱȱ ȱ ȱ pandes omothimadhon skepsusin everybody.NOM with-one-accord think.ACT.FUT.3PL ‘everybody will think about that with one accord’ (Rȕǯ, 110, 1185)55 ΏΙΐ΅ϟΑΝȱȱ ȱ ȱ (28) ΏΙΐ΅ϟΑΓΐ΅΍ȱȱȱĺȱȱȱ limenome limeno cleanse-from-dirt.NACT cleanse-from-dirt.ACT (Doukas, 136, 6; AD 15) ΦΗΔΣΊΓΐ΅΍ȱȱȱĺȱȱȱ ΦΗΔΣΊΝȱ   ȱ aspazome aspazo kiss.NACT kiss.ACT (Doukas, 306,22; AD 15)

From the aforementioned, it appears that the changes in the area of voice cannot be confined to only one period of Greek. The changes in voice of certain verbs had already begun during the Classical Greek period, other changes in the voice of verbs began in the Hellenistic period, but the final outcome of changes are to be found in Medieval Greek56. 3.3.3.2 Passive structures and absorption of the accusative case The predominance of the accusative as the case of the direct object is connected with the absence of structures with non-active verbs and direct object in the accusative (presence in the subject position of an argument which is found in the genitive or dative case in the corresponding transitive structure, and retention of the accusative case of an argument that is in the accusative case and in the corresponding transitive structure). There are, consequently, two phenomena that emerge together: the accusative replaces the genitive and the dative case, while theme/goal arguments (which are now in the accusative case in the corresponding transitive structures) cannot be in the subject position of the non-active type (the theme and goal arguments, usually marked with genitive or dative case, can be found in the subject position of non-active types in the Classical and post-Classical period). Ditransitive verbs with DPs in the genitive/dative and in the accusative case require from the Medieval Greek (Byzantine chronicles) two accusatives57: 54

55

56

57

C Æ Cusa Salvatore, I diplomi greci ed arabi di Sicilia, Köln – Wien: Böhlau 1982 (first edition: Palermo 1868, 1882). Rȕ Æ Robinson Gertrude, History and Cartulary of the Greek Monastery of S’ Elias and S’ Anastasios of Carbone, II. Cartulary, 1-2, Roma: Pont. Institutum Orientalium Studiorum, 1929. Cf. the use of non-active transitive types by Anna Komnene (AD 11-12). The morphological rules which Anna Komnene follows are, in total, classical, except for the ‘atticist’ use of the non-active instead of the active voice with transitive verbs. Similarly, in the Timarion (probably a work of Nikolaos Kallikles, imitation of Lucian) there are many non-active types that were active even in Classical Greek. Also, in Byzantine novels (Apostolopoulos 1984), the verbs which only have active morphology can be either

132

Chapter Three

(29) ȱ

ΘΓϿΖȱȱ ΈϾΓȱȱ ΈϟΈΉ΍ȱȱ ϳȱȱ Ά΅Η΍ΏΉϿΖȱȱ ΘΩΏΓ·΅ȱȱȱȱ tus dhio dhidhi o vasilefs talogha the.ACC two give.ACT.PRES.3SG the.NOM king.NOM the-horses.ACC ‘the king gives the horses to these two’ (Kallimachos and Chrysorroi, 898; AD 14)

ȱ

Regarding the PPs which introduce arguments in structures with non-active verb types, the situation generally is as follows: 4th-7th century: in the papyri (4th century), the preposition ЀΔϱipo ‘by/from’ denotes the agent; in the 6th and 7th centuries, the use of the preposition Δ΅ΕΣ para ‘from/by’ is very frequent; in Malalas and Moschos (6th century), the common preposition for the agent is ЀΔϱ ipo, but both Δ΅ΕΣ para and ΦΔϱ apo ‘by/from’ are found as prepositions of the agent argument. 8th-11th century: both the prepositions Δ΅ΕΣ para and ΦΔϱ apo are introduced by the copyists in the manuscripts of the New Testament in this period (Δ΅ΕΣ para: vernacular of 6th-12th century), 12th century: in the Escorial manuscript of Digenis Akritas, the preposition Δ΅ΕΣ para is used in passive structures, whereas in the Grotteferrata manuscript of Digenis Akritas, the prepositions ЀΔϱipo and Δ΅ΕΣ para are used. In the Ptochoprodromika, the use of a PP with non-active verbs is rare: the preposition Δ΅ΕΣ para is used twice, and the preposition πΎ ek ‘by/from’ is used once. Consequently, it appears that the preposition Δ΅ΕΣ para + genitive succeeded the preposition ЀΔϱ ipo for the denotation of the agent argument, and the preposition ΦΔϱapo was not the main preposition for the agent until the 12th century. Evidence for the availability of the multiple expression of the agent can be found in Old and Middle English and in Old Spanish, as well (Old English: Mitchell 1985: 334348, Middle English: Mustanoja 1969: 442, Old Spanish: Penny 1991: 103). 3.3.3.3 Voice in causative and anticausative verbs 3.3.3.3.1 The relationship of voice to changes in argument structures The absence of a connection between voice in the anticausatives and the processes of (anti)causativisation can be shown by the presence of verbs (mainly psych-verbs) that bear both voices alternatively during the Medieval Greek period. As in the case of verbs with prefix, where the presence of the prefix cannot be connected with (anti)causativisation, so, too, can voice and (anti)causativisation be considered unconnected. The intransitive psych-verbs șĮȣȝȐȗȦ thavmazo ‘to admire’, IJȠȜȝȫ tolmo ‘to dare’ and ȤĮȓȡȦ xero ‘to rejoice’ appear alternatively with active or non-active voice even in the same text. The DP in the accusative case denotes the cause (the cause is denoted in the accusative case or with PP İȟĮȚIJȓĮȢ eksetias because-of / ȝİ me with etc.). (30) a. Early Medieval Greek ̒Ёȱ ΘΓΏΐЗΐΉΑȱ ΘΤȱ ΔΕϲΖȱЀΐκΖȱ Ύ΅Θ΅ΚΌΣΗ΅΍  U tolmomen ta pros imas katafthase not dare.ACT.PRES.1PL the.ACC.PL to you anticipate.ACT.PERFVE.INF ‘we do not dare to anticipate yours’ (Ioannes Malalas, Chronographia, 18, 68, 89; AD 5-6) Late Medieval Greek b. ΐ΋ΈξΑȱȱ ΘΓΏΐφΗΉΘ΅΍ȱȱ Ύ΅ΑΉϠΖ transitive or intransitive. It appears, consequently, that in this period also there is no connection between transitivity and voice: Verbs with stable active morphology which are transitive in the novel “Kallimachos and Chrysorroi” (AD 14): ·νΐΝȱjemo fill.ACT (325), ЀΔΓΐνΑΝȱipomeno endure.ACT (1364). Verbs with stable active morphology which are intransitive in “Kallimachos and Chrysorroi”: Ά΅ΈϟΊΝȱvadhizo walk.ACT (2562), ΚΌΣΑΝȱ fthano arrive.ACT (85), ΘΕνΛΝȱ trexo run.ACT (640), ΚΉϾ·Νȱ fevgho leave.ACT (113). Similarly, verbs with non-active morphology can be both transitive (deponents) and intransitive: Verbs with stable non-active morphology which are transitive in “Kallimachos and Chrysorroi”: ΐ΋Λ΅ΑЗΐ΅΍ mixanome prepare.NACT (1710), ΗΔΏ΅ΛΑϟΊΓΐ΅΍ȱ splaxnizome have-pity.NACT (543), ΈνΛΓΐ΅΍ȱ dhexome accept.NACT (764), ΚΌν··Γΐ΅΍ȱfthengome speak.NACT (2450), Έ΍΋·ΓІΐ΅΍ȱdhiighume narrate.NACT (2022). Verbs with stable non-active morphology which are intransitive in “Kallimachos and Chrysorroi”: ΔΓΕΉϾΓΐ΅΍ȱ porevome march.NACT (60), Ύ΅ΘΓΙΑΉϾΓΐ΅΍ȱ katunevome camp.NACT (1228), σΕΛΓΐ΅΍ȱ erxome come.NACT (491), ϴΕΛΓІΐ΅΍ orxume dance.NACT (1320).

Diachronic Data and Analysis

133

midhen tolmisete kanis nothing.ACC dare.NACT.FUT.3SG nobody.NOM ‘nobody should dare anything’ (An Entertaining Tale of Quadrupeds, 702; AD 14) c. Ό΅ΙΐΣΊΓΑΘ΅΍ȱȱ ΘΤΖȱȱ ΔΕΣΒΉ΍Ζȱȱ ΘΓІȱȱ ΑΉΝΘνΕΓΙȱȱȱȱȱ thavmazonde tas praksis tu neoteru admire.NACT.PRES.3PL the.ACC actions.ACC the.GEN younger.GEN ‘they admire the actions of the younger man’ (Digenis Akritas, Ǽscorial, 779; AD 12) d. Γϡȱȱ ΦΔΉΏΣΘΉΖȱ Ό΅ΙΐΣΊΓΙΗ΍ȱȱ ΘχΑȱȱ ωΈΓΑχΑȱ ΘϛΖȱȱ ΎϱΕ΋Ζ i apelates thavmazusi tin idhonin tis koris the.NOM klephts.NOM admire.NACT.PRES.3PL the.ACC beauty.ACC the.GEN young-girl.GEN

‘the klephts admire the beauty of the young girl’ (Digenis Akritas, Ǽscorial, 1151; AD 12)

The psych-verb ȤĮȓȡȦ xero ‘to rejoice/enjoy’ is in evidence as a verb which can bear both non-active and active voice morphology in Medieval Greek. In (31c and d), it bears non-active morphology, as does the synonymous verb İȣijȡĮȓȞȠȝĮȚ efrenome ‘to rejoice’. In (31a), it also bears non-active morphology as does the synonymous verb ĮȖȐȜȜȦ aghalo ‘to take delight/enjoy’, which is found in the same extract. This example is in contrast to (31b), which is from the same author, and in which the verb ȤĮȓȡȦ xero ‘to rejoice/enjoy’ is in active voice, whereas the synonymous verb which is used here (IJȑȡʌȠȝĮȚ terpome ‘to delight/enjoy’) is in non-active voice. (31) Early Medieval Greek a.ȱ Λ΅ϟΕΉΘ΅΍ȱȱ xerete

ϳȱȱ ΎϱΗΐΓΖȉȱȱ Φ·ΣΏΏΉΘ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱψȱΗϾ·ΎΏ΋ΘΓΖȱΎ΅ϠȱϵΏΓΑȱΘϲȱΔ΅ΏΣΘ΍Αǯȱ

o

kozmos;

aghalete

i singlitos

ke olon to palatin.

rejoice.NACT.PRES.3SG the people.NOM be-glad.NACT.PRES.3SG the senate.NOM and all the palace.NOM

‘the people rejoice, the senate and everybody in the palace is glad’ (Constantine VII Porphyrogenitus, De ceremoniis, 2, 103, 23; AD 10) b. ̙΅ϟΕΉ΍ȱȱ ϳȱȱ ΎϱΗΐΓΖȱ ϳΕЗΑȱȱ ΗΉȱȱ ΅ЁΘΓΎΕΣΘΓΕ΅ȱȱ xeri o kozmos oron se aftokratoraȱ rejoice.ACT.PRES.3SG the people.NOM see.PART.NOM you.ACC.WEAK emperorȱ ΈΉΗΔϱΘ΋Αǰȱȱ Ύ΅Ϡȱȱ ψȱȱ ΔϱΏ΍Ζȱȱ ΗΓΙȱȱ ΘνΕΔΉΘ΅΍ȱȱ dhespotin, ke i polis su terpete ȱ sovereign and the city.NOM your be-glad.NACT.PRES.3SG ȱ ‘the people rejoice seeing you sovereign emperor, and your city is glad’ (Constantine VII Porphyrogenitus, De ceremoniis, 2, 91, 29; AD 10) Late Medieval Greek c. Λ΅ϟΕΉΘ΅΍ȱȱ ΉϢΖȱȱ ΘχΑȱȱ ·Ε΅ΚφΑȱȱ ΘΓΙ xerete is tin ghrafin tu rejoice.NACT.PRES.3SG with the letter his ‘he rejoices with his letter’ (Livistros and Rodamni, 1718; AD 14) d.ȱ Ύ΅΍ȱȱ Λ΅ϟΕΓΑΘ΅΍ȱȱ Ύ΅΍ȱȱ ΉΙΚΕ΅ϟΑΓΑΘ΅΍ǰȱȱ ΐ΍ΎΕΓϟȱȱ ΘΉȱΎ΅΍ȱΐΉ·ΣΏΓ΍ ke xeronde ke efrenonde, mikri te ke meghali and rejoice.NACT.PRES.3PL and enjoy.NACT.PRES.3PL young.NOM and old.NOM ‘they rejoice and enjoy, both yound and old’ (Apollonius of Tyre, The story of Apollonius (Kechagioglou), 791; AD 14)

Parallel use of active and non-active voice is also observed with the anticausative verbs ȤĮȜȐȦ xalao ‘to ruin’, ĮȣȟȐȞȦ afksano ‘to grow’, and ȗİıIJĮȓȞȦ zesteno ‘to warm’. Active and non-active forms are possible in semantically equivalent anticausative structures, while very close in the same text, verb synonyms or verbs belonging to exactly the same semantic group are in evidence, for which, however, a different voice has been chosen. (32) a.ȱ Ύ΅Ϡȱȱ ΘϱΘΉȱȱϳȱȱ ̐ΉΔΘϱΏΉΐΓΖǰȱȱ ΘϛΖȱȱ ͦϱΈΓΙȱȱ Ά΅Η΍Ών΅Ζǰȱ ke tote o Neoptolemos, tis Rodhu vasileas, and then the.NOM Neptolemos.NOM the.GEN Rhodes.GEN king.NOM ώΏΌΉȱȱ

Ύ΅Ϡȱȱ πΛ΅ΏΣΗΌ΋ΎΉȱȱ

Ύ΅ΘΉΔΣΑΝȱΘЗΑȱȱ ̖ΕЏΝΑȱȱȱȱȱȱ

ilthe ke exalasthike katepano ton Troon came.3SG and ruin.NACT.PAST.PERFVE.3SG at the Troyans ‘and then Neptolemos, the king of Rhodes, came and destroyed himself fighting against the Troyans’ (The War of Troy, 7516-7; AD 14) b.ȱ ΗΔφΘ΍΅ȱȱ ΐΓΙǯǯǯȱȱ ΘΤȱ ϳΔΓϟ΅ȱȱ ΧΚ΍Ύ΅ȱȱ Ύ΅Ϡȱȱ πΛ΅ΏΣΗ΅Αȱ spitia mu... ta opia afika ke exalasan houses.ACC my the.ACC which.ACC left.1SG and ruin.ACT.PAST.PERFVE.3PL ‘my houses ... which I left and they became ruins’ (Michail Maras, Katasticho, 149, ǹǯ, ǿȗȞ 3; AD 16)

134

Chapter Three

In the following examples, the anticausative verb ĮȣȟȐȞȦ afksano ‘to grow’ appears in one instance with active morphology next to an antonym which bears non-active voice (increase – decrease), and in another instance with non-active morphology and again next to a verb of the same semantic group, synonymous but with active voice. (33) a. ΗϾΑȱȱ ΗΓ΍ȱȱ ΔΣΑΘ΅ȱȱ ΅ЁΒφΗΉ΍ȱȱ Ύ΅Ϡȱ ΐΉ΍ΝΌφΗΉΘ΅΍   sin si panda afksisi ke miothisete with you all.NOM grow.ACT.FUT.3SG and decrease.NACT.FUT.3SG ‘all grows and decreases with you’ (Papyri Graecae Magicae, II, ȋǿǿǿ, 519-20) b.ȱ Ύ΅ϠȱϳȱΔϱΏΉΐΓΖȱ΋ЁΒΣΑΉΘΓȱȱ Ύ΅ϠȱπΔΏφΌΙΑΉΑȱȱ ψȱΐΣΛ΋   ke o polemos ifksaneto ke eplithinen i maxi  and the war.NOM grow.NACT.PAST.IMPRFVE.3SG and multiply.ACT.PAST.IMPRFVE.3SG the battle.NOM

‘and the war became bigger and the battles multiplied’ (Florios and Platziaflora, 687; AD 15) 

From the above, it is shown that the role of voice for intransitive and mainly for intransitive anticausative verbs (namely, for verbs that are alternating, with causative and anticausative type) is peripheral in Medieval Greek. The voice of anticausatives cannot be connected with processes of transitivisation, causativisation or anticausativisation, and seems to be concerned with the marking of anticausative structures independently of the presence and the role of the causative structure. Consequently, voice cannot be considered to reflect the basic character of the causative or anticausative structure in relation to the other existing structure or to mark the course of derivation (from transitive to intransitive or vice versa). Only the stable presence of active voice in the causative verbs of Greek constitutes causativisation evidence since the non-active verbs that obtain a novel causative use change to active verbs for the causative structures, and they do not participate in causative structures with non-active voice in any case (even in the period when the non-active transitive structure was used to express the benefactive meaning). 3.3.3.3.2 Causatives, anticausatives: distribution of active and non-active voice Three groups of verbs in Medieval Greek in relation to causative and anticausative voice can be distinguished: verbs with active causative and non-active anticausative voice morphology, verbs with active causative voice and availability of parallel use of non-active and active voice concerning the anticausative structures, and verbs with active causative and anticausative voice (resulting in exactly the same type of verb for transitive and intransitive). The verbs of the three groups that are further presented do not exhibit any change in relation to the previous (Hellenistic-Roman) period concerning the marking of voice. (a) Active causative – non-active anticausative The first group of verbs bear active endings in causative structures and non-active endings in anticausative structures. It is concerned, in other words, with verbs which continue with regard to the expression of (anti)causativity the situation of Classical Greek distinguishing the causatives from the anticausatives in terms of voice morphology. The anticausative verbs are marked with the voice that is also used for passive, reflexive and reciprocal structures. Concerning the specific verb meanings, the following should be noted: The verbs ȝĮȡĮȓȞȦ mareno ‘to wither’, ȟȘȡĮȓȞȦ ksireno ‘to dry’, țĮșĮȓȡȦ kathero ‘to clean’ and șİȡȝĮȓȞȦ thermeno ‘to warm’ retain the causative and anticausative structure without change in relation to Classical Greek. Again, the presence of a clitic in the causative structure is characteristic. (34) ΐ΅Ε΅ϟΑΝ mareno ‘to wither’ intransitive non-active a.ȱ ΅ЁΘχȱȱ Ύ΅ϠȱΤΑȱπΐ΅Ε΅ϟΑΉΘΓΑǰȱȱ   afti ke an emareneton,

σΏ΅ΐΔΉΑȱȱ

ȱБΖȱϳȱȱόΏ΍ΓΖ

elamben

os o ilios

she.NOM even if wither.NACT.PAST.IMPRFVE.3SG shine.ACT.PAST.IMPRFVE.3SG like the sun

‘even if she was withering, she was shining like the sun’ (Digenis Akritas, Ǽscorial, 183; AD 12)



b.ȱ ǻψȱΜΙΛφǼȱȱ Δ΅ΕΤȱΎ΅΍ΕϲΑȱπ·φΕ΅ΗΉΑǰȱȱ πΜϾ·΋Αǰȱȱ πΐ΅ΕΣΑΌ΋Αǯȱ (i psixi) para keron ejirasen, epsijin, emaranthin. (the soul) with time got-old.3SG got-cold.3SG wither.NACT.PAST.PERFVE.3SG ‘(the soul) got old, cold and withered with time’ (Michael Glykas, 203; AD 12) c.ȱ ǻΎΣΏΏΓΖǼȱΔЗΖȱπΐ΅ΕΣΑΌ΋ȱȱ ΔΕϲȱΎ΅΍ΕΓІȱ ΦΔϲȱΎΏ΅ΙΌΐΓІȱΎ΅ϠȱΔϱΑΓΙ  (kalos) pos emaranthi pro keru apo klafthmu ke ponu  

Diachronic Data and Analysis

135

(beauty) how wither.NACT.PAST.PERFVE.3SG before time from cry and pain ‘how the beauty withered before time from cries and pain’ (Kallimachos and Chrysorroi, 1591; AD 14) d.ȱ π··ϟΒϙȱȱ Θϲȱȱ ΗϟΈ΋ΕΓΑȱȱ ΉϢΖȱȱ ΘχΑȱȱ ϹϟΊ΅Αǰȱȱ ΐ΅Ε΅ϟΑΓΑΘ΅΍ȱȱ engiksi to sidhiron is tin rizan, marenonde touch.3SG the.NOM metal.NOM to the root wither.NACT.PRES.3PL ‘it touches the root with metal, and they wither’ (Agapios Landos, Geoponikon, 10, 6-7; AD 17) transitive active ΐ΅ΖȱΘΓϾΘ΋Αȱȱ ΘχΑȱȱ ΗΘΕΣΘ΅Αȱȱ ϵΏ΋Αȱ ȱ e.ȱȱ Ύ΅Ϡȱȱ Ύ΅ϾΗ΋ȱȱ Ύ΅Ϡȱΐ΅ΕΣΑ΋ȱȱ ke kafsi ke marani mas tutin tin stratian olin and burn.3SG and wither.ACT.PERFVE.3SG our that.ACC the.ACC army.ACC all.ACC ‘to burn and make all our army wither’ (Digenis Akritas, Ǽscorial, 887; AD 12) f.ȱȱ Θϲȱ ΩΑΌΓΖȱȱ ΘΓІȱȱ ΔΕΓΗЏΔΓΙȱȱ ΗΓΙȱȱ πΐΣΕ΅ΑΉΑȱȱ ψȱΌΏϟΜ΍Ζ    to anthos tu prosopu su emaranen i thlipsis  the flower.ACC the face.GEN your wither.ACT.PAST.PERFVE.3SG the sadness.NOM ‘sadness made the flower of your face wither’ (Digenis Akritas, Ǽscorial, 192; AD 12) g.ȱȱ ϳΔΓϿȱȱ Ηξȱȱ πΐΣΕ΅Α΅Αȱȱ ΓϡȱΩΑΈΕΉΖȱȱ ΓϡȱΒνΑΓ΍ opu se emaranan i andhres i kseni that you.ACC wither.ACT.PAST-PERFVE.3PL the men.NOM the foreign.NOM ‘that the foreign men made you wither’ (Spanos, ǹ, 233; ǹ, 276; B, 99; AD 14/15) Θϲȱȱ ΎΣΏΏΓΖȱ ΗΓΙȱ ΐ΅Ε΅ϟΑΓΙΑȱ  h.ȱȱ Γϡȱ ·ΤΕȱ ΗΙΛΑΓϟȱȱȱΗΓΙȱϴΈΙΕΐΓϠȱȱ   i ghar sixni su odhirmi to kalos su marenun ȱ the as frequent your wailings.NOM the.ACC beauty.ACC your wither.ACT.PRES.3PL ‘as your frequent wailing makes your beauty wither’ (Florios and Platziaflora, 96; AD 15) i.ȱȱ Ύ΅Ϡȱȱ ΐ΅Ε΅ϟΑΉ΍ȱȱ ΘΓΑȱȱ ψȱȱ ΔΉϧΑ΅ȱȱ ΘΓΙȱȱȱ ȱ ke mareni ton i pina tu and wither.ACT.PRES.3SG he.ACC.WEAK the hunger.NOM his ‘and his hunger makes him wither’ (Leonardo Dellaporta, Poems, ǹ, 2359; AD 14-15)

The verb is also causative with the addition of a prefix (usually țĮIJĮ- kata-), which has the meaning of “complete change-of-state” (the aktionsart is modified by the prefix). The prefix adds, consequently, an extra meaning to the denotation of change-of-state but is not completely necessary for causative use. The verb can also be causative without the prefix (cf. 35e) as there are, moreover, examples of intransitive structures with (cf. 35f) or without a prefix. (35) a. ΑΤȱȱΎϱΔΘ΋ȱȱ ΘχΑȱΎ΅ΕΈϟ΅ȱȱ ΐΓΙǰȱΑΤȱȱ ΘχΑȱȱ Ύ΅Θ΅ΐ΅Ε΅ϟΑ΋ na kopti tin kardhia mu, na tin katamareni to cut.3SG the heart.ACC my to she.ACC.WEAK KATA-wither.ACT.IMPRFVE.3SG ‘to break my heart, to make it wither’ (Livistros and Rodamni, 3090; AD 14) b. ǻΎΏ΅ΌΐϲΖȱΎ΅ϠȱΌΕϛΑΓΖǼȱȱ πΒΉΐΣΕ΅ΑΉΑȱȱ ΔκΗ΅Αȱȱ ΦΑΌΕЗΔΓΙȱȱ ΚϾΗ΍Αȱȱ (klathmos ke thrinos) eksemaranen pasan anthropu fisin (cry and lamentatio) EK-wither.ACT.PAST.PERFVE.3SG every.ACC human.GEN nature.ACC ‘(cries and lamentation) made every human’s nature wither’ (Kallimachos and Chrysorroi, 1599; AD 14) c. Γϡȱȱ ΔϱΑΓ΍ȱȱ ΘχΑȱȱ Ύ΅ΕΈϟ΅Αȱȱ ΘΓΙȱȱ πΎ΅ΘΉΐΣΕ΅ΑΣΑȱȱ ȱȱΘ΋Α i poni tin kardhia tu ekatemaranan tin the pains.NOM the heart.ACC his KATA-wither.ACT.PAST.PERFVE.3PL she.ACC.WEAK ‘the pains made his heart wither’ (Florios and Platziaflora, 296; AD 15) d. ǻϳȱΔΓΘ΅ΐϲΖȱϳȱΔϾΕ΍ΑΓΖǼȱΜϾΒϙȱΎ΅ϠȱΎ΅Θ΅ΐ΅ΕΣΑϙȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΤΖȱοΆΈΓΐφΑΘ΅ȱΈϾΓȱϊΐ΍ΗΙȱΚΏνΆ΅Ζȱȱ (o potamos o pirinos) psiksi ke katamarani tas evdhominda dhio imisi flevas (the river the of-fire) freeze.3SG and KATA-wither.ACT.PERFVE.3SG the seventy two half veins.ACC

‘(the river of fire) will freeze and make the seventy two and a half veins wither’ (Spanos, B, 184; AD 14/15) causative without prefix e. ǻϳȱΔΓΘ΅ΐϲΖȱϳȱΔϾΕ΍ΑΓΖǼȱȱȱΜϾΒϙǰȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ϾΗϙǰȱȱȱȱȱΐ΅ΕΣΑϙȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΤΖȱοΆΈΓΐφΑΘ΅ǯǯǯ (o potamos o pirinos) psiksi, kafsi, marani tas evdhominda... (the river the of-fire) freeze.3SG burn.3SG wither.ACT.PERFVE.3SG the.ACC seventy ‘(the river of fire) will freeze, burn, make the seventy... wither’ (Spanos, ǻ, 1780; AD 14/15) anticausative with prefixȱ Έ΍Τȱȱ ΔϱΌΓΑȱ ΘϛΖȱȱ ̔ΓΈΣΐΑ΋Ζȱ f. Θϲȱȱ ΔЗΖȱȱ Ύ΅Θ΅ΐ΅Ε΅ϟΑΉΘ΅΍ȱ to pos katamarenete dhia pothon tis Rodhamnis the how KATA-wither.NACT.PRES.3SG from passion the Rodamni.GEN ‘how he withers from his passion for Rodamni’ (Livistros and Rodamni, 3184; AD 14)

The verb ȟȘȡĮȓȞȦ ksireno ‘to dry’ is usually intransitive in Medieval Greek, and the causative use of

136

Chapter Three

the verb is not particularly systematic. Of interest is the use of the adverb ʌĮȡĮȤȡȒȝĮ paraxrima ‘immediately’ in some of the following examples (that denote a spontaneous event), and whose presence can comprise a criterion for the spontaneous change-of-state in the case of the anticausative. (36) Β΋Ε΅ϟΑΝȱksireno ‘to dry’58 intransitive non-active Early Medieval Greek a. Ύ΅ϠȱΔ΅Ε΅ΛΕϛΐ΅ȱπΒ΋ΕΣΑΌ΋ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱψȱȱΘΓ΍΅ϾΘ΋ȱΎ΅ϠȱΘ΋Ώ΍Ύ΅ϾΘ΋ȱȱΌΉΓΈЏΕ΋ΘΓΖȱȱΔ΋·φȱȱȱ ke paraxrima eksiranthi i tiafti ke tilikafti theodhoritos piji and immediately dry.NACT.PAST.PERFVE.3SG the that

and of-such-age given-by-God spring.NOM

‘and that spring of such age, given by God, immediately dried’ (John Moschos, Leimonarion (The Spiritual Meadow), LXXX, D, 11; AD 6-7) ψȱȱ ΎΓΏΙΐΆφΌΕ΅ȱ  b. πΘϱΏΐ΋ΗΉΑȱΉϢΔΉϧΑǯǯǯȱȱȱΎ΅ϠȱΉЁΌνΝΖȱȱȱȱȱȱȱȱȱ πΒ΋ΕΣΑΌ΋ȱȱ   etolmisen ipin... ke eftheos eksiranthi i kolimvithra   dared.3SG speak.INF and immediately dry.NACT.PAST.PERFVE.3SG the baptistery.NOM ‘he dared to speak... and the baptistery immediately dried’ (Theophanes Confessor, Chronicle, 234Ǻ, 19; AD 8-9)  Late Medieval Greek c. ϵΘ΅Αȱȱ Β΋Ε΅ΑΌΓІΑȱȱ ΅ЁΘΤȱȱ ΘΤȱȱ ΚϾΏΏ΅ otan ksiranthun afta ta fila when dry.NACT.PERFVE.3PL these.NOM the.NOM leaves.NOM ‘when these leaves dry’ (Agapios Landos, Geoponikon, 6, 5; AD 17) transitive active πΒφΕ΅ΑΉΑȱȱ ǻψȱΎ΅ІΗ΍ΖȱΘΓІȱ̽ΏϟΓΙǼȱȱ d. ΘϲΑȱȱ ǻΎ΅ΕΔϲΑȱΘΓІȱΌνΕΓΙΖǼȱȱ ton (karpon tu therus) eksiranen (i kafsis tu Iliu) he.ACC (fruit.ACC the.GEN summer.GEN) dry.ACT.PAST.PERFVE.3SG (the heat.NOM the sun.GEN)

‘(sun’s heat) dried it (the summer fruit)’ (Agapios Landos, Geoponikon, 11; AD 17) ΦΏ΅Θ΍ΗΐνΑΓΑȱΎΕν΅Ζǯǯǯȱȱ Β΋Ε΅ϟΑΉ΍ȱȱ ΘχΑȱȱ ΗΣΕΎ΅ȱȱ e. Θϲȱȱ to alatizmenon kreas... ksireni tin sarka the.NOM salted.NOM meat.NOM dry.ACT.PRES.3SG the.ACC flesh.ACC ‘the salted meat... dries the flesh’ (Agapios Landos, Dietetics, 15, 26; AD 17) ȱ

The verb Ȝȣʌȫ lipo ‘to sadden’ participates in the transitivity alternation in Medieval Greek; it is worth noting the limitation of its causative use in the following period (i.e. Modern Greek) (it is mainly used with a clitic in the direct object position). (37) ΏΙΔЏ lipo ‘to sadden’ intransitive non-active ψȱȱ ΜΙΛφȱȱ ΐΓΙȱȱ a. Ύ΅Ϡȱȱ ΏΙΔ΋Όϛȱȱ   ke lipithi i psixi mu     and sadden.NACT.PERFVE.3SG the.NOM soul.NOM my ‘and my soul became sad’ (Digenis Akritas, Ǽscorial, 492; AD 12) b. Ύ΅Ϡȱȱ ΦΔϲȱΘΓІǰȱȱ π·ΑЏΕ΍ΗΉǰȱȱ ΘϠȱȱ σΑ΅΍ȱȱ Θϲȱȱ ΏΙΔΓІΐ΅΍ȱȱȱȱ ȱȱ ke apo tu, eghnorise, ti ene to lipume and from that learned.3SG what be.3SG the.NOM sadden.NACT.PRES.1SG ‘and from that, he learned what is to be sad’ (Livistros and Rodamni, 112; AD 14) transitive active c. Early Medieval Greek ΘΓϿΖȱȱ

ΘϛΖȱȱ

ΗΙ·ΎΏϛΘΓΙȱȱ

ΐΉ·ΣΏΝΖȱȱ

πΏϾΔΉ΍

tus the.ACC

tis the.GEN

singlitu senate.GEN

meghalos considerably

elipi sadden.ACT.PAST.IMPRFVE.3SG



‘he considerably saddened the men of the senate’ (Theophanes Confessor, Chronicle, 239Ǻ, 7; AD 8-9)

d. Late Medieval Greek Ύ΅Ϡȱȱ

ΔΣΏ΍Αȱȱ

ΘΓϿΖȱȱ

πΏϾΔ΋ΗΉΖȱ

ke palin tus elipises and again they.ACC.WEAK sadden.ACT.PAST.PERFVE.2SG ‘and you saddened them again’ (Livistros and Rodamni, 550; AD 14) 58

Cf. the connection between transitivity alternation, voice and the participle -ȝİȞȠȢ -menos (see also point (d) below): (1)

·ΉΑνΗΌΝȱȱ

Έχȱȱȱȱȱȱȱȱȱȱȱ ψȱȱȱȱȱȱȱȱȱȱȱȱȱ

ΛΉϟΕȱȱȱȱȱȱȱȱ

ΗΓΙȱ

jenestho dhi i xir su be.IMPER.3SG therefore the.NOM hand.NOM your ‘therefore let your hand dry’ (Spanos, ǻ, 1476; AD 14/15)

Β΋Ε΅ΐνΑ΋

ksirameni dry.NACT.PAST.PERFVE.PART.NOM

Diachronic Data and Analysis

137

In the anticausative non-active structures, the cause can be denoted with a DP in the accusative case: (38) Ύ΅΍ȱȱ ΘϟΖȱȱ Α΅ȱȱ ΐ΋Αȱȱ ΘΓΑȱȱ ΏΙΔ΋ΌΉϟȱȱȱȱȱȱ ke tis na min ton lipithi and who.NOM to not he.ACC.WEAK sadden.NACT.PERFVE.3SG ‘and who is not going to be grieved because of him’ (Apollonius of Tyre, Rimada V, 321; AD 14)

The causative structure Ȝȣʌȫ lipo ‘to cause grief’ + DP in the accusative (patient) tends to be abandoned precisely when themes and causes with psych-verbs begin to be consistently denoted with accusative case. In the following examples, I cite verbs of the Medieval period (mainly psych-verbs) which are accompanied by a theme or a cause in the accusative case and which are not causativised: (39) ȱ ȱ

(40)

(41) ȱ

Early Medieval Greekȱ ΎΏ΅ϟΉ΍Ζȱȱ

ΘΤΖȱȱ

Υΐ΅ΕΘϟ΅Ζȱȱ ΗΓΙ

kleis tas amartias su cry.ACT.PRES.2SG the.ACC sins.ACC your ‘you are crying for your sins’ (John Moschos, Leimonarion (The Spiritual Meadow), LXXI, C, 5; AD 6-7) Late Medieval Greek Ύ΅Ϡȱȱ Γϡȱȱ ΦΔΉΏΣΘΉΖȱȱ ΑΤȱȱ ΘϲΑȱȱ ΘΕνΐΓΙΗ΍Αȱȱȱȱȱ  ke i apelates na ton tremusin and the.NOM klephts.NOM to he.ACC.WEAK tremble.ACT.IMPRFVE.3PL ‘and the klephts to tremble because of him’ (Digenis Akritas, Ǽscorial, 784; AD 12) Φ·΅Α΅ΎΘΉϧȱȱ ΘΓϿΖȱȱ ΔϱΑΓΙΖȱȱȱȱȱȱ aghanakti tus ponus exasperate.ACT.PRES.3SG the.ACC pains.ACC ‘he was exasperated by his pains’ (Kallimachos and Chrysorroi, 1709; AD 14)

(42) a.ȱ ΐχȱȱΚΓΆ΋ΌϛΖȱȱ ΘϲΑȱȱ ΌΣΑ΅ΘΓΑȱȱ ȱΔ΅ΕΤȱΐ΋ΘΕϲΖȱȱ Ύ΅ΘΣΕ΅Α mi fovithis ton thanaton para mitros kataran not be-afraid.NACT.PERFVE.2SG the.ACC death.ACC from mother.GEN curse.ACC ‘do not be afraid of death from a mother's curse’ (Digenis Akritas, Ǽscorial, 2; AD 12) b. ΑΤȱ ΐχȱȱΚΓΆκΘ΅΍ȱȱ ΎϟΑΈΙΑΓΑȱȱ ΉϢΖȱȱΘΓϿΖȱȱΆ΅ΕνΓΙΖȱȱ ΔΓΏνΐΓΙΖ   na mi fovate kindhinon is tus vareus polemus to not be-afraid.NACT.PRES.3SG danger.ACC in the heavy fighting ‘do not be afraid of danger in heavy fighting’ (Digenis Akritas, Ǽscorial, 1785; AD 12)

The cause can be denoted with a DP in the accusative case or alternatively with a PP: (43) a.ȱȱ ΐχΑȱ Θϲȱȱ ΌΏ΍ΆϛΖȱ ·ΤΕȱ Ών·Νȱȱ ΗΉȱ min to thlivis ghar legho se not it.ACC.WEAK be-distressed.NACT.PERFVE.2SG because say.1SG you.ACC ‘do not be distressed by that, I say to you’ (Digenis Akritas, Ǽscorial, 372; AD 12) ȱȱȱȱȱȱ b. ΐχȱ ΌΏϟΆΉΗ΅΍ǰȱȱ ΑΉЏΘΉΕΉǰȱȱ σΑΉΎΉΑȱȱ ΘϛΖȱȱ ΐ΋ΘΕϱΖȱȱ ΗΓΙ mi thlivese, neotere, eneken tis mitros su not be-distressed.NACT.IMPFVE.2SG youngster.VOC because the mother your ‘do not be distressed, youngster, because of your mother’ (Digenis Akritas, Ǽscorial, 424; AD 12) (44) a. ΘϲΑȱȱ ̏΅Ε·΅ΕϟΘ΋Αȱȱ ΌΏϟΆΓΑΘ΅΍ȱ     ton Margharitin thlivonde the.ACC Margharitis.ACC be-distressed.NACT.PRES.3PL ‘they are distressed for Margaritis’ (The War of Troy, 6865; AD 14) ȱȱȱȱȱȱ b. ϳȱȱ ̝Λ΍ΏΏΉϿΖȱȱ πΌΏϟΆ΋ΎΉȱȱ ȱΈ΍ΤȱȱȱȱΘϲΑȱȱπΒΣΈΉΏΚΓΑȱΘΓΙ o Axilefs ethlivike dhia ton eksadhelfon tu the.NOM Achilles.NOM be-distressed.NACT.PAST.PERFVE.3SG because the cousin his ‘Achilles was distressed because of his cousin’ (The War of Troy, 6093; AD 14)

Due to the non-availability of causativisation, the causative meaning is denoted with a periphrastic structure or with a different verb (the formation with the suffixes -ȚȗȦ -izo, -Ȧ(ȣ)ȞȦ -o(i)no is very productive; cf. the related observations for Modern Greek, in Giannakidou & Merchant 1999, and in Alexiadou & Anagnostopoulou 2004): (45) a. ΘϲΑȱȱ ton

ΔΕϟ·Ύ΍Δ΅ȱȱ

πΚΓΆνΕ΍ΊΉΖ

pringipa

efoverizes

138

Chapter Three the.ACC prince.ACC threaten.ACT.PAST.IMPRFVE.2SG ‘you were threatening the prince’ (Chronicle of Moreas, 5133; AD 14 ) b. ϵΔΓΙȱȱ ΑΤȱȱΗΓІȱȱ ΘϲΑȱȱ Λ΅ΕϾΑϙȱȱ ϳȱȱȱ ΌΉϱΖȱ opu na su ton xarini o theos that to you.GEN.WEAK he.ACC.WEAK gladden.ACT.PERFVE.3SG the god.NOM ‘that God will gladden him for you’ (Leontios Machairas, Chronicle, 3, 401; AD 15)ȱ c. ϳΔΓϟ΅ȱȱ ΘϲΑȱȱ ΔΓ΍φΗϙȱȱ ΑΤȱȱ Λ΅ΕϜ na xari opia ton piisi whoever.NOM he.ACC.WEAK make.3SG to be-glad.ACT.PERFVE.3SG ‘whoever will make him glad’ (Florios and Platziaflora, 797; AD 15)

Similarly, in Medieval Greek the verbs IJıĮțȓȗȦ tsakizo ‘to crush’ and ȝĮȗȫȞȦ mazono ‘to gather’ form active causative and non-active anticausative structures. (46) IJıĮțȓȗȦ tsakizo ‘to crush’ intransitive non-active Γϡȱȱ ̕΅Ε΅ΎφΑΓ΍ȱȱȱ a. Ύ΅Ϡȱȱ πΘΊ΅ΎΎϟΗΘ΋Η΅Αȱ ke etzakistisan i Sarakini and crush.NACT.PAST.PERFVE.3PL the.NOM Saracen.NOM ‘and the Saracen were crushed’ (Leontios Machairas, Chronicle, 3, 285; AD 15) b. Έ΍ϱΘ΍ȱπΘΊ΅ΎϟΗΌ΋ȱȱ ρΑ΅ȱΐΉ·ΣΏΓΑȱΘЗΑȱȱ̆ΣΏΏΝΑȱȱȱȱȱȱȱΎ΅ΕΣΆ΍ȱȱȱȱΉϢΖȱΌΣΏ΅ΗΗ΅Α  dhioti etzakisthi ena meghalon ton Ghalon karavi is thalasan as crush.NACT.PAST.PERFVE.3SG one big.NOM the.GEN French.GEN ship.NOM in sea ‘as a big ship of the French destroyed at sea’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 356, 7-8; AD 17) transitive active ΘΓϿΖȱȱ ̕΅Ε΅Ύ΋ΑΓϾΖȱȱȱ c. Ύ΅Ϡȱ πΘΊΣΎΎ΍Η΅Αȱȱ ke etzakisan tus Sarakinus and crush.ACT.PAST.PERFVE.3PL the.ACC Saracen.ACC ‘and they destroyed the Saracen’ (Leontios Machairas, Chronicle, 3, 286; AD 15) (47) ȝĮȗȫȞȦ mazono ‘to gather’ intransitive non-active ϵΏΏ΅ȱ ΘΓІΘ΅ȱȱ ǻΈ΋ΑνΕ΍΅ǼȱΉϢΖȱȱΘΤȱȱΛνΕ΍΅ȱ ΘΝΖȱ a. Ύ΅Ϡȱ ΑΤȱ ΐ΅ΊΓΎΘΓІΑȱȱ ke na mazoktun ola tuta (dhineria) is ta herja tus and to gather.NACT.PERFVE.3PL all.NOM these.NOM (dinars) in the hands their ‘so that all these (dinars) are gathered in their hands’ (M. Grigoropoulos, notary of Chandaka, 52, 120; AD 16) b. ψȱȱ Μφȱȱ ΐΓΙȱΗΙΐΐ΅ΊЏΑΉΘ΅΍ȱȱ Ύ΍ȱȱ ώΕΌ΅Αȱȱ ΘΤȱЂΗΘΉΕ΅ȱȱ ΐΓΙȱȱ i psi mu simazonete ki irthan ta istera mu the soul.NOM my SYN-gather.NACT.PRES.3SG and came.3SG the last-years.NOM my ‘my soul gathers itself and the last years of my life have come’ (The Sacrifice of Abraham, 192; AD 16) οΒ΅ΎΓΗϟ΅΍ΖȱΛ΍Ώ΍ΣΈΉΖȱȱΩΑΌΕΝΔΓ΍ȱȱ ΘΓІȱȱΔΓΏνΐΓΙ c. Ύ΅Ϡȱȱ Έ΍Ήΐ΅ΊЏΛΌ΋Η΅Αȱȱ   ke dhiemazoxthisan eksakosies xiliadhes anthropi tu polemu  and DIA-gather.NACT.PAST.PERFVE.3PL six-hundred thousand persons the war.GEN ‘six-hundred thousand persons of war gathered’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 350, 30-31; AD 17) transitive active ΐν·΅ȱȱΗΘΕΣΘΉΙΐ΅ȱȱΦΔϲȱΘΓϿΖȱ̖΅ΘΣΕΓΙΖȱ d. ǻϳȱΛΣΑ΋ΖȱΘЗΑȱ̖΅ΘΣΕΝΑǼȱȱ πΐΣΊΝΒΉȱȱ (o xanis ton Tataron) emazokse megha stratevma apo tus Tatarus (the leader.NOM the Tatars.GEN) gather.ACT.PAST.PERFVE.3SG large troop.ACC from the Tatars

‘(the leader of the Tatars) gathered a large troop of Tatars’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 347, 1; AD 17)

With regard to the presence of the causative ijȜȑȖȦ flego ‘to inflame’, there are many disagreements in the dictionaries (cf. also Zombolou 2004). Medieval Greek data show evidence of its anticausative nonactive use as well as the causative active use of this specific verb. The verb is also used as causative with a prefix. (48) ΚΏν·Ν flegho ‘to inflame/burn’ intransitive non-active a. Early Medieval Greek πΔΉ΍Έχȱȱ ΓЇΑȱȱ

πΎȱȱ

ΘΓІȱȱ

ΔΙΕΉΘΓІȱ πΚΏν·ΉΘΓ

epidi

ek

tu

piretu

un

eflejeto

Diachronic Data and Analysis

139

as prt from the fever inflame.NACT.PAST.IMPRFVE.3SG ‘as he was burning from the fever’ (John Moschos, Leimonarion (The Spiritual Meadow), CLXXXIV, C, 2; AD 6-7) Late Medieval Greek b. ΗΙΛΑΣΎ΍Ζȱȱ ΌνΏΉ΍ȱȱ ΚΏν·ΉΗΌ΅΍ȱȱ ΉϢΖȱȱ ΔϱΌΓΑȱȱ ΘϛΖȱȱ ϴΈϾΑ΋Ζȱȱȱȱȱȱȱȱ sixnakis theli flejesthe is pothon tis odhinis very-often want.3SG inflame.NACT.PRES.INF in passion the grief.GEN ‘he very often wants to burn in a passion of grief’ (Florios and Platziaflora, 209; AD 15) c. ΑΤȱȱ ΚΏν·ΉΗ΅΍ǰȱȱ ΑΤȱ Ύ΅ϟΉΗ΅΍ȱȱȱ na flejese, na keese to inflame.NACT.IMPRFVE.2SG to burn.NACT.IMPERFVE.2SG ‘to inflame, to burn’ (Leonardo Dellaporta, ǹ, 99; AD 14-15) transitive active σΚΏΉΒΉΑȱȱ ΑνΎΙΑȱȱȱȱȱȱȱ ȱȱ d. ΓЁȱȱ ·ΉΑǽνΘ΋ΖȱȏȱȏȱȏǾΓΖȱȱ u jen[etis _ _ _]os efleksen nekin not ancestor.NOM inflame.ACT.PAST.PERFVE.3SG corpse.ACC ‘he is not an ancestor... burned the corpse’ (Funerary Inscript., 1037, 4) σΚΏΉΒΉȱȱ ψȱ ȱΚΏϱ·΅ȱȱ ΘϛΖȱȱ Φ·ΣΔ΋Ζȱȱ e. ΘϲΑȱȱ ton eflekse i flogha tis aghapis he.ACC.WEAK inflame.PAST.PERFVE.3SG the flame.NOM the love.GEN ‘the flame of love set him on fire’ (Livistros and Rodamni, 2932; AD 14) use of the prefix kata-; transitive f. Θϲȱȱ ΔІΕȱȱ Φΐ΋Λ΅ΑΉϧȱȱ Ύ΅Ϡȱȱ ΓЁΈξΑȱȱ ΘϲΑȱȱ Ύ΅Θ΅ΚΏν·Ή΍ȱȱȱȱȱ ȱ to pir amixani ke udhen ton katafleji the fire.NOM stop.3SG and not he.ACC.WEAK KATA-inflame.ACT.PRES.3SG ‘the fire stops and does not burn him’ (Livistros and Rodamni, 314; AD 14)

(b) Parallel presence of non-active and active anticausative type without difference in meaning The second group of verbs is made up of verbs that bear active voice morphology in the causative type and can bear both non-active and active voice morphology in the anticausative type. The active anticausative type is, in most instances, the result of change (during the Hellenistic or the Medieval periods) in the voice morphology of a verb which already participated in transitivity alternation. The trend is for replacing the non-active form with an active form and for the same causative and anticausative voice morphology. The change is gradual and concerns all the Medieval period, as the two types are found in competition until one prevails over the other. The verbs IJİȜİȚȫȞȦ teliono ‘to finish’, ıȤȓȗȦ sxizo ‘to tear’, ıțȠȡʌȓȗȦ skorpizo ‘to scatter’, ȕȣșȓȗȦ vithizo ‘to sink’ belong to this group of verbs. (49) ΘΉΏΉ΍ЏΑΝȱteliono ‘to finish’ intransitive non-active a. Early Medieval Greek Ύ΅Ϡȱȱ ΓЂΘΝΖȱȱ πΘΉΏΉ΍ЏΌ΋

ke utos and so

eteliothi finish.NACT.PAST.PERFVE.3SG

‘and it finished like that’ (John Moschos, Leimonarion (The Spiritual Meadow), IV, B, 6; AD 6-7)

Late Medieval Greek ΦΎϱΐ΋ȱȱ ϳȱȱ ΔϱΏΉΐΓΖȱȱ ΈξΑȱȱ πΘΉΏΉ΍ЏΌ΋ȱȱȱȱȱȱ ȱ b. ϷΘ΅Αȱ ȱ ȱ otan akomi o polemos dhen eteliothi when yet the.NOM war.NOM not finish.NACT.PAST.PERFVE.3SG ‘when the war was not yet finished’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 343, 2829; AD 17) intransitive active ψȱȱ ΏΉ΍ΘΓΙΕ·ϟ΅ȱ c. Ύ΅Ϡȱȱ ϵΑΘ΅ΑȱȱπΘΉΏΉϟΝΗΉΑȱȱ ke ondan eteliosen i liturjia and when finish.ACT.PAST.PERFVE.3SG the.NOM mass.NOM ‘and when the mass finished’ (Leontios Machairas, Chronicle, 3, 416; AD 15) d. πΘΉΏΉϟΝΗ΅Αȱȱ ΘΤȱȱ σΘ΋ȱȱ ΐΓΙȱ eteliosan ta eti mu finish.ACT.PAST.PERFVE.3PL the.NOM years.NOM my ‘my years finished’ (The Sacrifice of Abraham, 194; AD 16) e. ϳΔΓІȱȱ πΘΉΏΉϟΝΗΉΑȱȱ ψȱȱ ΐΣΛ΋ȱȱȱȱ opu eteliosen i maxi that finish.ACT.PAST.PERFVE.3SG the.NOM battle.NOM ‘that the battle finished’ (Agapios Landos, Geoponikon, 42, 17; AD 17)

140

Chapter Three transitive active ΔЗΖȱȱΑΤȱ ΘχΑȱȱ ȱΘΉΏΉ΍ЏΗΝȱȱ f. Ύ΅Ϡȱȱ ΔЗΖȱȱ ΑΤȱ·ΕΣΜΝȱȱΘχΑȱΦΕΛφΑǰȱȱ ke pos na ghrapso tin arxin, pos na tin telioso and how to write.1SG the beginning.ACC how to it.ACC.WEAK finish.ACT.PERFVE.1SG ‘and how to write the beginning, how to finish it’ (Imberios and Margarona, 1; AD 15-16) g. ΦΏΏΤȱΌνΏΝȱȱȱΓϡȱȱȱϹ΋ΌνΑΘΉΖȱȱȱȱȱȱȱȱȱȱȱȱπΔϟΘΕΓΔΓ΍ǯǯǯȱȱȱȱȱȱΘΉΏΉ΍ЗΗ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΧΔ΅ΑΘ΅ȱȱ ala thelo i rithendes epitropi... teliose apanda but want.1SG the assigned.NOM wardens.NOM finish.ACT.PERFVE.INF everything.ACC ‘but I want that the assigned wardens... finish everything’ (M. Grigoropoulos, notary of Chandaka, 48, 59; AD 16) h. ΑΤȱȱ ΘΉΏ΍ЏΗΝȱȱ Θ΍Αȱȱ Δ΅ΕΓϾΗ΅Αȱȱ ΘΓΙȱ Ύ΅Ϡȱȱ ΉϢΗΘΉΕΓΑȱȱ Έ΍ΦΌϟΎ΍Αȱȱ na telioso tin parusan tu ke isteron dhiathikin to finish.ACT.PERFVE.1SG the present.ACC his and later.ACC will.ACC ‘that I finish his present and later will’ (Manolis Varouchas, Notarial acts, 66, 4-5; early AD 17 (1597-1613))

The verb IJİȜİȚȫȞȦ teliono ‘to finish’ frequently forms causative structures with the addition of the prefix ȟİ- kse-. Again, however, the prefix denotes greater emphasis on the completion of the action (change-of-state) than on verb causativity since the prefix can be added to form an anticausative verb, as well. (50) prefix and causative structure a. ΒΉΘνΏΉ΍ΝΗΉȱǻπΗϾǼȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΔΕΓΌΙΐΉΕΤȱȱȱΘΤȱȱȱȱȱȱȱȱȱȱȱȱΗϷΛΝȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΐ΍Ώ΋ΐνΑ΅ȱ kseteliose (esi) prothimera ta soxo milimena KSE-finish.ACT.IMP.2SG willingly the.ACC you-have.1SG say.PART.ACC ‘finish willingly what I have said to you’ (The Sacrifice of Abraham, 20; AD 16) prefix and anticausative structure ΑΤȱȱ ΒΉΘΉΏΉ΍ΝΌΓІΑȱȱ πΘΓϾΘ΋ȱ ΘχΑȱ ψΐνΕ΅ b. ΗΔΓϾΈ΅ΒΉȱȱ spudakse na kseteliothun etuti tin imera make-haste.IMP.2SG to KSE-finish.NACT.PERFVE.3PL that the day ‘make haste so that they will finish today’ (The Sacrifice of Abraham, 785; AD 16)

Particularly frequent is the addition of a prefix to the verb ıȤȓȗȦ sxizo ‘to tear’. The addition of the prefix, most of the time, results in an anticausative structure. (51) ΗΛϟΊΝȱsxizo ‘to tear’ use of the prefix dia-; intransitive non-active Early Medieval Greek Έ΍΅ΗΛϟΊΉΗΌ΅΍ a. ΎΦΑΘΉІΌΉΑȱȱ πΈϱΎΓΙΑȱȱ ΅ϡȱȱ е̔Νΐ΅ϞΎ΅Ϡȱȱ ΗΔΓΑΈ΅Ϡȱȱ kandefthen edhokun e Romaike sponde dhiasxizesthe thence thought.3PL the Roman treaty.NOM DIA-tear.NACT.PRES.INF ‘and thence they thought that the Roman treaty was broken’ (Constantine VII Porphyrogenitus, de legationibus, 223, 9; AD 10) without prefix; intransitive non-active Early Medieval Greek b. πΒ΅ϟΚΑ΋Ζȱȱ πΗΛϟΗΌ΋ȱȱ ψȱȱ ·ϛȱȱ eksefnis esxisthi i ji suddenly tear.NACT.PAST.PERFVE.3SG the ground.NOM ‘the ground rent apart suddenly’ (Ioannes Malalas, Chronographia, 18, 19, 55; AD 5-6) Late Medieval Greek c. ΑΤȱȱ ΗΛϟΊΉΘ΅΍ȱȱ πΎȱȱ ΘΓϿΖȱȱ ΔϱΑΓΙΖ    na sxizete ek tus ponus   to tear.NACT.IMPRFVE.3SG from the pains ‘to be torn apart by the pains’ (Livistros and Rodamni, 46 ; AD 14) intransitive active Late Medieval Greek d. ΔЗΖȱȱ ΓЁΎȱȱ σΗΛ΍ΗΉΑǰȱȱ ΔЗΖȱȱ ΓЁΎȱπΕΕΣ·΋Αȱȱ ȱȱȱȱȱΘϱΘΉȱΘΓІȱ pos uk esxisen, pos uk erajin tote tu how not tear.ACT.PAST.PERFVE.3SG how not break.NACT.PAST.PERFVE.3SG then the ΈΕΣΎΓΑΘΓΖȱȱ ψȱȱ ΈΙΗЏΈ΋Ζȱȱ ΎΓ΍Ώϟ΅ȱȱ ȱ ȱ ȱ ȱ Δ΅ΑΘΓΚΣ·ΓΙȱȱ pandofaghu dhrakondos i dhisodhis kilia all-devouring.GEN dragon.GEN the ill-smelling abdomen.NOM ‘how the ill smelling abdomen of the devouring dragon did not tear apart and break’ (Kallimachos and Chrysorroi, 686-7; AD 14) transitive active

Diachronic Data and Analysis

141

e. Early Medieval Greek σΗΛ΍ΗΉȱȱ

ΘΤȱȱ ϡΐΣΘ΍΅ȱȱ

Ύ΅Ϡȱȱ

ΗΣΎΎΓΑȱȱ

ΔΉΕ΍ΉΆΣΏΉΘΓ

esxise ta imatia ke sakon perievaleto tear.ACT.PAST.PERFVE.3SG the clothes.ACC and coarse-cloth.ACC put-round.3SG ‘he tore his clothes and tied a coarse cloth around himself’ (Ioannes Malalas, Chronographia, 5, 41, 75; AD 5-6)

The verbs ıțȠȡʌȓȗȦ skorpizo ‘to scatter’ and ȕȣșȓȗȦ vithizo ‘to sink’ retain in the Medieval Greek period their causative active and anticausative non-active and active morphology. It is concerned consequently in these verbs with a change in voice morphology of the anticausative that began in the Hellenistic period and continued up to the Medieval period. Moreover, with the verb ıțȠȡʌȓȗȦ skorpizo ‘to scatter’, the presence of a prefix is very frequent. (52) ΗΎΓΕΔϟΊΝȱskorpizo ‘to scatter’ intransitive non-active Early Medieval Greek a. ǻ̕΅ΐ΅ΕΉϧΘ΅΍ǼȱΗΙΑΣ·ΓΑΘ΅΍ǰȱȱȱȱȱȱȱ ΓЁΛȱȱГΕΐ΋ΗΉȱȱȱΎ΅ΘЮȱ΅ЁΘЗΑȱΎ΅ϠȱȱȱπΗΎΓΕΔϟΊΓΑΘΓȱȱ ȱȱȱȱȱ ȱ (Samarite) sinaghonde, ux ormise kat’ afton ke eskorpizondo (Samaritans) gather.NACT.PRES.3PL not throw.3SG against them and scatter.NACT.PAST.IMPRFVE.3PL

‘Samaritans were gathering, he did not throw himself against them and they were scattering’ (Ioannes Malalas, Chronographia, 18, 35, 83-84; AD 5-6) Late Medieval Greek ΔΏϛΌΓΖȱȱ ΗΎΓΕΔ΍ΗΌϜȱȱ b. БΗΤΑȱȱ Θϲȱȱ osan to plithos skorpisthi when the.NOM crowd.NOM scatter.NACT.PAST.PERFVE.3SG ‘when the crowd scatters’ (Kallimachos and Chrysorroi, 1937; AD 14) c. Ύ΅Ϡȱȱ ΘΤȱȱ ΑΉΕΤȱȱ ΗΎΓΕΔϟΊΓΙΑΘ΅΍ȱȱ ȱ ke ta nera skorpizunde and the.NOM water.NOM scatter.NACT.PRES.3PL ‘and the water is scattering’ (Florios and Platziaflora, 1201; AD 15) intransitive active d. ΘΓȱȱ ΔΏΓϟΓΑȱȱ ȱΈ΍ΉΗΎϱΕΔ΍ΗΉΑǰȱȱ ΉΔΑϟ·΋Η΅Αȱ Γ΍ȱȱ ΔΣΑΘΉΖȱȱȱ ȱȱȱȱȱ ȱ to plion dhieskorpisen, epnijisan i pandes the.NOM ship.NOM DIA-scatter.ACT.PAST.PERFVE.3SG drown.3PL the.NOM all.NOM ‘the ship tore apart, all were drowned’ (Apollonius of Tyre, The story of Apollonius, (Kechagioglou), 137; AD 14) e. ΔϟΉȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱρΑ΅ȱȱΔΓΘϛΕ΍ȱΎΕ΅ΗϠȱȱΩΈΓΏΓΑȱȱȱȱΑΤȱΗΎΓΕΔϟΗϙȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱψȱȱȱΎ΅ΎΓΗϾΑ΋ȱȱȱȱȱȱȱȱȱȱΘΓΙΖȱȱȱ ȱȱȱȱȱȱ pie ena potiri krasi adholon na skorpisi i kakosini tus drink.IMP.2SG one glass wine.ACC truly.ACC to scatter.ACT.PERFVE.3SG the malevolence.NOM their

‘drink a glass of wine without guile, so that their malevolence ceases’ (Agapios Landos, Dietetics, 62, 12; AD 17) transitive active ΘΓΙΖȱȱ ϵΏΓΙΖȱȱȱȱ f. πΗΎϱΕΔ΍ΗΣȱȱ eskorpisa tus olus scatter.ACT.PAST.PERFVE.1SG they.ACC.WEAK all.ACC ‘I made them all scatter’ (Digenis Akritas, Ǽscorial, 1714; AD 12) g. πΗΎϱΕΔ΍ΗΉȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΤȱȱΔΏΓϾΘ΋ȱȱȱȱȱȱȱΘ΋ΖȱΎ΅ϠȱΘЗΑȱΔΘΝΛЗΑȱȱΘΤȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΈϟΈΉ΍ȱȱ eskorpise ta pluti tis ke ton ptoxon ta dhidhi scatter.ACT.PAST.PERFVE.3SG the wealth.ACC her and the poor.GEN they.ACC.WEAK give.3SG

‘she scattered her wealth and gave them to the poor people’ (Leonardo Dellaporta, Poems, ǹ, 1899; AD 14-15) (53) ΆΙΌϟΊΝ vithizo to sink’ intransitive non-active Early Medieval Greek Έξȱȱ ΏΓ΍ΔΓϿΖȱȱ ЀΔϲȱȱ ΔΏφΌΓΙΖȱȱ a. ΘΓϿΖȱȱ tus dhe lipus ipo plithus the.ACC prt rest.ACC under much

ЀΈΣΘΝΑȱȱ ΆΙΌ΍ΗΌϛΑ΅΍

idhaton water

vithisthine sink.NACT.PERFVE.INF

‘that the rest of them sunk under much water’ (Ioannes Malalas, Chronographia, 1, 3, 63-64; AD 5-6)

intransitive active Late Medieval Greek b. πΚ΅ϟΑΉΘϱΑȱȱ ΗΓΙȱȱ ϵΘ΍ȱȱ ψȱȱ ·ϛȱȱ ЀΔΣ·Ή΍ȱΑΤȱȱ ΆΙΌϟΗϙȱȱ efeneton su oti i ji ipaji na vithisi seemed.3SG you.GEN that the ground.NOM go.3SG to sink.ACT.PERFVE.3SG ‘it seemed to you that the ground was going to sink’ (The War of Troy, 7022; AD 14)

142

Chapter Three transitive active c. БΗΤΑȱΑΤȱόΌΉΏΉȱȱȱȱȱȱȱȱȱȱǻϳȱΗΉ΍ΗΐϱΖǼȱȱȱȱȱȱȱΑΤȱȱΘχΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱǻ·ϛΑǼȱȱ osan na ithele (o sizmos) na tin (jin) as

to wanted.3SG (the earthquake) to she.ACC.WEAK (ground)

ΆΙΌϟΗϙȱȱ

vithisi sink.ACT.PERFVE.3SG

‘as (the earthquake) wanted to sink it (the ground)’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) I, 335; AD 17)

The Medieval verb ĮȜȜȐııȦ alaso ‘to change’ combines two main meanings: the meaning of a change of clothes (which is a stable and very frequent collocation and meaning for this verb, resulting in the frequent absence of the word ‘clothes’ as easily implied direct object) and the meaning of the change of the previous situation in which the subject or the object was found. The meaning ‘to change clothes’ is clearly not causative and does not show the syntactic behaviour of causative verbs. Only the second meaning is causative and retains the active causative structure and the parallel use of non-active and active anticausative structure in Medieval Greek. (54) ΦΏΏΣΗΗΝȱalaso ‘to change’ active transitive; ‘they change clothes’ presence of the direct object Early Medieval Greek ΘχΑȱȱ ϢΈϟ΅Αȱȱ ΗΘΓΏφΑ ȱ a. Ύ΅Ϡȱȱ ρΎ΅ΗΘΓΖȱȱ ΅ЁΘЗΑȱȱ ΦΏΏΣΗΗΉ΍ȱȱ ke ekastos afton alasi tin idhian stolin and each.NOM they.GEN change.ACT.PRES.3SG the his-own outfit.ACC ‘and each of them changes his own outfit’ (Constantine VII Porphyrogenitus, De ceremoniis, 2, 105, 7; AD 10) active transitive; ‘they change (clothes)’ omission of the direct object Early Medieval Greek b. ̙Εχȱȱ ΉϢΈνΑ΅΍ȱȱ ȱΔЗΖȱȱΦΏΏΣΗΗΓΙΗ΍ΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΓϡȱȱΈΉΗΔϱΘ΅΍ȱȱȱπΑȱΘ΅ϧΖȱοΓΕΘ΅ϧΖȱΎ΅ϠȱΔΕΓΉΏΉϾΗΉΗ΍ȱ ȱ ȱ xri idhene pos alasusin i despote en tes eortes ke proelefsesi must know.INF how change.ACT.PRES.3PL the bishops.NOM at the feast and processions ‘you must know how the bishops change (their clothes) at feasts and processions’ (Constantine VII Porphyrogenitus, De ceremoniis, 1, 175, 2; AD 10) Late Medieval Greek Ύ΅Ϡȱȱ ΦΏΏΣΗΗΉ΍ c. Ύ΅Ϡȱȱ πΏΓϾΗΘ΋ȱȱ ke elusti ke alasi and shampooed.3SG and change.ACT.PRES.3SG ‘and he shampooed and changed (his clothes)’ (ȉhe Song of Armuris, 191; AD 11) d. ЂΔ΅·ΉȱȱȱȱȱȱȱȱȱȱΉϢΖȱȱΘχΑȱΘνΑΘ΅ȱΐΓΙȱȱΎ΅ϠȱȱΚνΕΉȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΐΓΙȱȱȱȱȱȱȱΑȂȱȱȱΦΏΏΣΒΝ  ipaje is tin tenda mu ke fere mu n’ alakso go.IMP.2SG to the tent my and bring.IMP.2SG I.GEN to change.ACT.PERFVE.1SG ‘go to my tent and bring me clothes to change into’ (Digenis Akritas, Ǽscorial, 1194; AD 12) Ύ΅΍ȱȱ ΏΓΙΗΌϛΖȱȱȱȱȱȱȱȱȱȱΎ΅ϠȱȱΦΏΏΣΒ΋Ζ  e. Κ΅·ΉϧΑȱ Ύ΅Ϡȱȱ Δ΍ΉϧΑȱȱ ȱȱσΛΉ΍Ζȱȱ fajin ke piin exis ke lusthis ke alaksis eat.INF and drink.INF have.2SG and shampoo.2SG and change.ACT.PERFVE.2SG ‘you have to eat and drink and shampoo and change (clothes)’ (Digenis Akritas, Ǽscorial, 1779; AD 12)

The availability of the omission of the object with the meaning ‘to change clothes’ facilitates the morphological change (into active) of the anticausative ĮȜȜȐȗȦ alazo ‘to change the previous state’, which was marked with non-active voice morphology in Classical Greek. This change began with the parallel use of active voice morphology in Hellenistic Greek and continues in Medieval Greek. The intransitive examples do not comprise passive structures with availability of presence of an agent, as in Modern Greek. In Medieval Greek, structures such as Ș ȩȥȘ ȐȜȜĮȟİ / i opsi alakse / the.NOM appearance.NOM change.ACT.PAST.PERFVE.3SG / ‘his appearance changed’ (spontaneous, without the intervention of external factor) are witnessed both with non-active and with active voice morphology. Consequently, these parallel active and non-active intransitive structures, which are seen in the Medieval Greek period, only echo a tendency of change in voice morphology towards active anticausatives (which has not been completed; that is why the two types appear at the same time), and not non-active forms which are used in other (passive) structures (as in Modern Greek). (55) intransitive non-active ΆϟΓΖȱ ΓЈΘΓΖȱ ΩΗΘ΅ΘΓΖȱ ΦΏΏΣΗΗΉΘ΅΍ȱ Ύ΅ΌȂȱȱ ГΕ΅Αȱȱ a. ϳȱ o vios utos astatos alasete kath’ oran the life.NOM that.NOM fickle.NOM change.NACT.PRES.3SG every hour

Diachronic Data and Analysis



143

‘this fickle life is changing every hour’ (Michael Glykas, 386; AD 12) ΌνΏΉ΍ȱ ΦΏΏ΅ΛΌϛȱ ϳȱ ΑΓІΖȱ Θ΋Ζȱ b. ΓЂΘΝΖȱ ΘϛΖȱȱ utos tis theli alaxthi o nus tis so she.GEN.WEAK want.3SG change.NACT.PERFVE.3SG the mind.NOM her ‘so her mind is going to change’ (The War of Troy, 5959; AD 14) ψȱȱ ΐΓΕΚφȱȱ ΘΓΙȱ c. πΎȱȱ ΘϛΖȱȱΎ΅ΎΓΔ΅ΌΉϟ΅Ζȱȱ ΘΓΙȱȱ ωΏΏΣ·΋Αȱ ek tis kakopathias tu ilajin i morfi tu from the sufferings his change.NACT.PAST.PERFVE.3SG the figure.NOM his ‘his figure changed from his sufferings’ (Velthandros and Chrysandza, 1266; AD 14) intransitive active d. ΦΔϲȱȱΘϲΑȱΚϱΆΓΑȱǀΗȂǁȱϵΏΉΖȱΘΓΙΖȱΦΏΏΣΗΗ΅ΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΓΙΖȱȱ΅ϡȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΛΕϱΉΖȱȱȱȱ apo ton fovon oles tus alasan tus i xroes from the fear all their change.ACT.PAST.PERFVE.3PL their the.NOM skins.NOM ‘all their skins changed from fear’ (The War of Troy, 5999; AD 14) e. ΑȂȱȱ ΦΏΏΣΒΉ΍ȱȱ ψȱ ϷΕΉΒφȱȱ ΗΓΙȱȱȱ n’ alaksi i oreksi su to change.ACT.PERFVE.3SG the.NOM wish.NOM your ‘that your wish is going to change’ (The Sacrifice of Abraham, 808; AD 16) f. ·ΑΝΕϟΊΓΙΗ΍Αȱȱ ϵΘ΅Αȱȱ ΌνΏϙȱȱ ΑΤȱ ΦΏΏΣΒϙȱȱ ϳȱȱ Ύ΅΍ΕϱΖȱȱ ghnorizusin otan theli na alaksi o keros know.3PL when want.3SG to change.ACT.PERFVE.3SG the weather.NOM ‘they know when the weather is going to change’ (Agapios Landos, Tois Enteuksomenois, į, 16; AD 17) transitive active ΓЁΎȱȱ όΏΏ΅Β΅Ζȱȱ ΔΓΘνȱ g. ΌϾΕ΅Αȱȱ thiran uk ilaksas pote door.ACC not change.ACT.PAST.PERFVE.2SG never ‘you did not ever change a door’ (Ptochoprodromos, ǿ, 84; AD 11-12) h. Θϲȱȱ Δ΋·΅ΈϱΗΛΓ΍ΑΓΑȱȱ πΎϱΔ΋ȱȱȱȱΎ΅ϠȱȱȱΪΖȱȱ Θϲȱȱ ΦΏΏΣΒΓΙΑȱȱȱȱ  to pighadosxinon ekopi ke as to alaksun the rope-of-well.NOM cut.3SG and let it.ACC.WEAK change.ACT.PERFVE.3PL ‘the rope of the well was cut and they should change it’ (Ptochoprodromos, ǿ, 55; AD 11-12)ȱ

The verb ĮȣȟȐȞȦ afksano ‘to grow’ also belongs to this group of verbs. Regarding what Vassilaki (1988) observes about the ongoing morphological change of that verb during the Modern Greek period, this ongoing change appears to reflect something which started much earlier. The verb was active causative and non-active anticausative in Classical Greek; in the Hellenistic Koine active morphology begins to be witnessed (in parallel to non-active) for the anticausative. Moreover, it is frequently used with the meaning ‘to grow old’, a meaning that is not denoted by the corresponding Modern Greek verb. The situation in the Medieval Greek period is similar to the situation of the Hellenistic-Roman period. (56) ΅ЁΒΣΑΝ afksano ‘to grow’ intransitive non-active Early Medieval Greek ΋ЁΒΣΑΓΑΘΓȱȱ Ύ΅ΘΤȱȱ ·νΑΓΖȱ΅ЁΘЗΑ a. Ύ΅Ϡȱȱ ΔΣΑΘ΅ȱΘΤȱȱ ϷΑΘ΅ǯǯǯȱ ke panda ta onda... ifksanondo kata jenos afton and all the beings.NOM grow.NACT.PAST.IMPRFVE.3PL according race their ‘and all the beings... were multiplying according to their race’ (Ioannes Malalas, Chronographia, 1, 4, 83-85; AD 5-6) Late Medieval Greek Ύ΅Ϡȱ πΑ΋Ώ΍Ύ΍ЏΌ΋Αȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ b. ǻϳȱΆ΅Η΍ΏΉϿΖǼȱ Ύ΅Ϡȱȱ ΋ЁΒφΑΌ΋Αȱȱ   (o vasilefs) ke ifksinthin ke enilikiothin (the king) and grow.NACT.PAST.PERFVE.3SG and came-of-age.3SG ‘(the king) grew and came of age’ (Spanos, ǹ, 174; Ǻ, 7-8; AD 14/15) c. Ύ΅Ϡȱ ϳȱ ΔϱΏΉΐΓΖȱȱ ΋ЁΒΣΑΉΘΓȱ   ke o polemos ifksaneto and the war.NOM grow.NACT.PAST.IMPRFVE.3SG ‘the war grew bigger’ (Florios and Platziaflora, 687; AD 15) intransitive active Early Medieval Greek d. Ύ΅Ϡȱ ГΗΔΉΕȱΆΕνΚΓΖȱ ΅ЁΒΣΑΝȱȱȱȱȱ ke osper vrefos afksano and like baby grow.ACT.PRES.1SG ‘and I grow like a baby’ (Romanus Melodus, Cantica, 11, 14, 2; AD 6)

144

Chapter Three Late Medieval Greek ΘϱΘΉΖȱȱ ϳȱ ͑ΝΗφΚȱ e. ΅ЄΒ΅ΑΉȱ afksane totes o Iosif grow.ACT.PAST.IMPRFVE.3SG then the Joseph.NOM ‘then Joseph was growing older’ (Leonardo Dellaporta, Poems, ǹ, 488; AD 14-15) transitive active f. ǻΗЗΐ΅Ǽȱȱ ΘχΑȱȱ ΛΣΕ΍Αȱ Ύ΅Ϡȱ ΘχΑȱ ψΈΓΑχΑǯǯǯȱȱ ΅ЁΒΣΑΉ΍ȱ (soma) tin xarin ke tin idhonin... afksani (body) the grace.ACC and the pleasure.ACC grow.ACT.PRES.3SG ‘(the body) is growing its grace and pleasure’ (Kallimachos and Chrysorroi, 796; AD 14) g. ǻϳȱЃΔΑΓΖǼȱ ΅ЁΒΣΑΉ΍ȱȱ ΘχΑȱȱ σΐΚΙΘΓΑȱ ΌνΕΐ΋Αȱȱȱȱȱȱȱȱȱȱȱȱ (o ipnos) afksani tin emfiton thermin (the sleep) grow.ACT.PRES.3SG the innate.ACC warmth.ACC ‘(the sleep) is increasing the innate warmth’ (Agapios Landos, Dietetics, 4, 3-4; AD 17) transitive; with a prefix ΘΓІȱȱ ΔΙΕϲΖȱ ΘχΑȱ ΚΏϱ·΅Αȱ πΔ΅ΙΒΣΑΉ΍ h. ΘΤȱ ΒϾΏ΅ȱ ta ksila tu piros tin floghan epafksani the wood.NOM the fire.GEN the flame.ACC EPI-grow.ACT.PRES.3SG ‘as the wood increases the flame of the fire’ (Spaneas, 131; AD 12)

Of similar meaning and with exactly the same syntactic behaviour is the verb ʌȜȘșĮȓȞȦ plitheno ‘to increase/grow’. The anticausative ʌȜȘșĮȓȞȦ plitheno ‘to increase/grow’ can be non-active and active, and there is also causative active structure. (57) intransitive non-active ψȱȱ ΈϾΑ΅ΐϟΖȱȱ ΘΝΑ a. Έ΍Τȱ ΑΤȱ ΐχΑȱ ΔΏ΋ΌΙΑΌϜȱȱ dhia na min plithinthi i dhinamis ton for to not increase.NACT.PERFVE.3SG the strength.NOM their ‘so that their strength do not increase’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 346, 27; AD 17) intransitive active b. ΦΗΘΕ΅ΔΘ΅ϠȱȱȱȱȱȱȱΎ΅ϠȱȱΆΕΓΑΘ΅ϠȱȱȱȱȱȱȱȱȱȱΉϢΖȱΘϲȱȱΎΉΚΣΏ΍ΑȱȱΘΓΙȱπΔΏφΌΙΑ΅Αȱȱȱȱȱȱ astrapte ke vronde is to kefalin tu eplithinan lightning.NOM and thunder.NOM to the head his increase.ACT.PAST.PERFVE.3PL ‘lightning and thunder increased over his head’ (Spanos, ǻ, 929; AD 14/15) c. ΑΤȱ ΔΏ΋ΌϾΑϙȱȱ ϳȱȱ ΎϱΗΐΓΖ na plithini o kozmos to increase.ACT.PERFVE.3SG the people.NOM ‘so that the people increase’ (Leonardo Dellaporta, Poems, ǹ, 1540; AD 14-15) ΉϢΖȱ ΘχΑȱȱ ̍΍Θ΅Ϭ΅Α d. ǻψȱΛΕ΍ΗΘ΍΅Α΍ΎχȱΔϟΗΘ΍ΖǼȱ πΔΏφΌΙΑΉΑȱȱ (i xristianiki pistis) eplithinen is tin Kitaian (the Christian faith) increase.ACT.PAST.PERFVE.3SG in the Kitaia ‘(the Christian faith) was growing in Kitaia’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 350, 25; AD 17) transitive active e. Ύ΅΍ȱ ΅ΙΘΉϟΑΓ΍ȱȱȱȱȱȱ ΔΣΑΘ΅ȱ ΗΘνΎΓΙΑΘ΅΍ȱ Θ΅ȱ Ώϱ·΍΅ȱȱȱȱȱȱȱȱȱ Α΅ȱ ΔΏ΋Ό΅ϟΑΓΙ   ke aftini panda stekunde ta loja na plithenu   and those.NOM always stand.3PL the words.ACC to increase.ACT.PRES.3PL ‘and those always try to increase the words’ (Apollonius of Tyre, Rimada ǹ, Ǽ, 620; AD 14)

The verbs ȜȣȖȓȗȦ lijizo ‘to bend’, ĮȞȠȓȖȦ anigho ‘to open’, țȜİȓȞȦ klino ‘to close’, țĮȓȦ keo ‘to burn’ and ıȕȒȞȦ zvino ‘to extinguish’ are witnessed as active causatives and non-active or active anticausatives during the Medieval Greek period. In the case where active voice morphology is used in the anticausative, causative and anticausative type bear exactly the same voice morphology, whereas in the case of non-active anticausatives, the relationship between the causative and the anticausative resembles the relationship between the transitive and the passive or the reflexive. The presence of the two types (non-active and active), however, reflect ongoing changes in the voice of the anticausative type, as they are given shape in the Hellenistic period and continue in the Medieval Greek period). (58) ΅ΑΓϟ·Ν anigho ‘to open’ intransitive non-active Early Medieval Greek a. όΈ΋ȱ ΐΓ΍ȱ д̌ΓΕΈΣΑ΋Ζȱȱ idhi mi Iordhanis

ΦΑΓϟ·ΉΘ΅΍ȱȱȱȱ

anijete

Diachronic Data and Analysis

145

already I.DAT Jordan.NOM open.NACT.PRES.3SG ‘Jordan already opened for me’ (Romanus Melodus, Cantica, 16, 3; AD 6) b. ΦΑΓϟ·ΓΑΘ΅΍ȱ Έξȱ ΅ϡȱȱ ΘΕΉϧΖȱ ΌϾΕ΅΍ȱȱ anighonde dhe e tris thire open.NACT.PRES.3PL prt the.NOM three doors.NOM ‘the three doors open’ (Constantine VII Porphyrogenitus, De ceremoniis, 405, 8; AD 10) Late Medieval Greek intransitive active c. όΑΓ΍ΒΉΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱψȱȱȱȱȱΎ΅ΕΈϟ΅ȱȱȱȱȱȱȱΐΓΙǰȱȱȱπΛΣΕ΋ΑȱȱȱȱȱȱȱȱȱȱψȱȱȱȱΜΙΛφȱȱȱȱȱȱȱȱȱΐΓΙ iniksen i kardhia mu, exarin i psixi mu open.ACT.PAST.PERFVE.3SG the heart.NOM my rejoiced.3SG the soul.NOM my ‘my heart opened, my soul rejoiced’ (Livistros and Rodamni, 664; AD 14) d. ΘΕνΛΉ΍ǰȱȱȱ ΘΤΖȱΔϱΕΘ΅Ζȱȱȱȱȱ Ών·Ή΍ǰȱȱȱȱȱ ΐϱΑ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱ ΘΝΑȱȱȱȱȱȱȱȱȱȱȱ ΦΑΓϟ·ΓΙΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ Δ΅Ε΅ΙΘϟΎ΅ trexi, tas portas leji, mone ton anighun paraftika run.3SG the doors.ACC say.3SG alone.NOM they.GEN open.ACT.PRES.3PL immediately ‘he runs, says to the doors, they immediately open alone’ (Kallimachos and Chrysorroi, 1280; AD 14) transitive active ΌϾΕ΅Αȱȱ ΐΓΙȱ e. ΪΑȱȱΓЁΎȱȱ ΦΑΓϟΒϙΖȱ   an uk aniksis thiran mu     if not open.ACT.PERFVE.2SG door.ACC my ‘if you do not open my door’ (Ptochoprodromos, ǿǿ, 98; AD 11-12)  f. ΘχΑȱ ΚΙΏ΅ΎχΑȱȱ ΐξȱȱ ΦΑΓϟ·ΓΙΗ΍Α    tin filakin me anighusin the jail.ACC I.ACC.WEAK open.ACT.PRES.3PL ‘they open the prison for me’ (Livistros and Rodamni, 3225; AD 14) g. Ύ΍ȱ ΩΑΓ΍ΒΉȱȱ ΘΗϠȱȱ Φ·ΎΣΏΉΖ  ki anikse tsi angales and open.ACT.PAST.PERFVE.3SG the.ACC arms.ACC ‘and he opened his arms’ (The Sacrifice of Abraham, 1108; AD 16)

The following examples can also be analysed as instances of passivisation of the causative verb with availability of presence of an agent: (59) 

ΔΓΘξȱȱ ΐχȱ

ΦΑΓ΍ΛΌЗΗ΍Αȱȱ

ǻΓϡȱΔϱΕΘΉΖǼȱȱȱȱ 

pote mi anixthosin (i portes)  never not open.NACT.PERFVE.3PL (the doors) ‘so that (the doors) will be never opened’ (The War of Troy, 7158; AD 14) (60) ΎΏΉϟΑΝȱklino ‘to close’ intransitive non-active ΘϛΖȱȱȱ ΐΉΘ΅ΑΓϟ΅Ζȱȱ ψȱȱ ΌϾΕ΅ȱȱȱȱȱ a. ΘЏΕ΅ȱȱ πΎΏΉϟΗΌ΋ΎΉΑȱȱ tora eklisthiken tis metanias i thira now close.NACT.PAST.PERFVE.3SG the.GEN regret.GEN the door.NOM ‘and now the door of regret closed’ (Leonardo Dellaporta, Poems, Ǻ, 111-112; AD 14-15) ΎΏΉ΍ΗΌЗΗ΍Αȱȱ ΅ϡȱȱ ΌϾΕ΅΍ȱȱ b. ΔΕϠΑȱȱ prin klisthosin e thire before close.NACT.PERFVE.3PL the.NOM doors.NOM ‘before the doors close’ (Leonardo Dellaporta, Poems, Ǻ, 153; AD 14-15) intransitive active Θϲȱȱ ΘνΕΐΉΑΓΑȱȱ ΘϛΖȱ ΘΕνΆ΅Ζȱ c. ȱ ΘχΑȱȱ οΗΔνΕ΅Αȱ ϳΔΓϿȱȱ σΎΏΉ΍ΗΉȱȱ tin esperan opu eklise to termenon tis trevas the evening when close.ACT.PAST.PERFVE.3SG the deadline.NOM the truce.GEN ‘the evening when the truce deadline ceased’ (The War of Troy, 6572-3; AD 14) d. Θϲȱ ΘνΕΐΉΑΓΑȱ πΔ΍ΌΙΐΓІΑȱ ΑΤȱȱΎΏΉϟΗϙȱȱ ΘϛΖȱȱ Φ·ΣΔ΋ΖȱǻΉϢΕφΑ΋ΖǼȱȱȱȱ to termenon epithimun na klisi tis aghapis (irinis) the time.NOM wish.3PL to close.ACT.PERFVE.3SG the love.GEN (peace.GEN) ‘they wish that the time of love (of peace) will come to an end’ (The War of Troy, 7388; AD 14) e. ϵΘ΍ȱϳȱȱȱΎ΅΍ΕϲΖȱȱȱȱȱȱȱȱȱȱΗΙΑνΎΏΉ΍ΗΉΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ϠȱȱΓЁȱȱΗΙΑΘΉΏΉϧȱȱȱȱΘϲΑȱȱΗΔϱΕΓΑȱ oti i keros sineklisen ke u sindeli ton sporon that the weather.NOM SIN-close.ACT.PAST.PERFVE.3SG and not conduce.3SG the seed.ACC

‘that the weather closed in and was not conducive to seeds’ (Livistros and Rodamni, 947; AD 14) cf. transitive active ΗΚ΍ΎΘχΑȱȱ ΘχΑȱȱ ΌϾΕ΅Αȱȱȱȱ ȱ ȱ f. ΎΏΉϟΉ΍ȱȱ klii sfiktin tin thiran

ȱ

146

Chapter Three close.ACT.PRES.3SG tight.ACC the.ACC door.ACC ‘he closes the door tight’ (Ptochoprodromos, ǿ, 26; AD 11-12)

The non-active intransitive țȜİȓȞȠȝĮȚ ‘to close’ is a passive type (since passive interpretation and overt or implied presence of the agent is possible) in the examples (61a and b): (61) Early Medieval Greek Έξȱȱ Ύ΅Ϡȱȱ ΘΤȱȱ πΕ·΅ΗΘφΕ΍΅ȱ a. πΎΏΉϟΗΌ΋Η΅Αȱȱ eklisthisan dhe ke ta erghastiria ȱȱȱ close.NACT.PAST.PERFVE.3PL prt and the workshops.NOM ‘the workshops were also closed’ (Theophanes Confessor, Chronicle, 199P, 25; AD 8-9) Late Medieval Greek b. ΦΔϲȱΘΓІȱΑІΑȱȱȱΓϡȱȱȱΔϱΕΘΉΖȱȱȱȱȱȱȱΘ΋ΖȱȱΘϛΖȱȱȱȱȱȱȱȱȱ̖ΕΓϟ΅ΖȱȱȱȱȱȱȱȱΑΤȱȱȱΎΏΉ΍ΗΌΓІΗ΍ apo tu nin i portes tis tis Trias na klisthusi from the now the doors.NOM her the.GEN Troy.GEN to close.NACT.PERFVE.3PL ‘that from now on the doors of Troy shall be closed’ (The War of Troy, 7157; AD 14) (62) Ύ΅ϟΝ keo ‘to burn’ intransitive non-active Early Medieval Greek a. Ύ΅ϠȱȱπΎ΅ϾΌ΋ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΔΓΏϿȱȱȱȱȱȱȱȱȱȱȱΐνΕΓΖȱȱȱȱȱȱȱȱΘϛΖȱȱȱȱȱȱȱȱȱȱΔϱΏΉΝΖȱȱ  ke ekafthi poli meros tis poleos  and burn.NACT.PAST.PERFVE.3SG much.NOM part.NOM the.GEN city.GEN ‘and much of the city was burned down’ (Theophanes Confessor, Chronicle, 278Ǻ, 27; AD 8-9) Late Medieval Greek b. πΎ΅ϾΌ΋Η΅ΑȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΤȱȱΗΔΏΣΛΑ΅ȱȱȱȱȱȱȱΘΓΙǰȱπΛΣΌ΋ΑȱψȱȱȱȱΎ΅ΕΈϟ΅ȱȱȱȱȱȱȱȱΘΓΙ    ekafthisan ta splaxna tu, exathin i kardhia tu burn.NACT.PAST.PERFVE.3PL the bowels.NOM his lost.3SG the heart.NOM his ‘his bowels were burned, his heart was lost’ (Digenis Akritas, Ǽscorial, 300; AD 12) c. πΈΤȱȱΑΤȱȱΐΔϛΎ΅ȱȱȱȱΗΘχȱȱȱȱȱȱȱΚΝΘ΍ΤȱȱΑΤȱȱΎ΅ϟ·ΓΙΐΓΙȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΐξȱȱΗνΑ΅ edha na bika sti fotia na keghumu me sena here to enter.1SG into-the fire to burn.NACT.PAST.IMPRFVE.1SG with you ‘to get in the fire here, to become burned for you’ (The Sacrifice of Abraham, 868; AD 16)

The intransitive active structures were promoted by the possibility of omission of the implied object in transitive structures (with a generalised meaning). The example ΑΤȱΎ΅ϟΓΙΑȱπΎȱΘχΑȱΚΏϱ·΅Α /ȱna keun ek tin floghan / to burn.ACT.PRES.3PL from the flame can be also interpreted as: ‘so that they burn everybody who touches them’ and the example ΈξΑȱΔΕνΔΉ΍ȱΑΤȱΘΤȱΘΕЏ·ΉΘΉȱϵΘ΅ΑȱΎ΅ϟΓΙΗ΍Α / dhen prepi na ta trojete otan keusin / not must to them eat.2PL when burn.ACT.PRES.3PL as ‘you should not eat them, when they burn everybody who eats them’. (63) intransitive active ΎΓΎΎ΍ΑϲΆ΅Κ΅ȱȱ ΑΤȱȱ Ύ΅ϟΓΙΑȱȱ πΎȱȱ a. Ύ΅Ϡȱȱ ΛΉϟΏ΋ȱȱ ke xili kokinovafa na keun ek and lips.NOM red-coloured.NOM to burn.ACT.PRES.3PL from

ΘχΑȱȱ

ΚΏϱ·΅Αȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

tin the

floghan flame

‘and the red coloured lips to become burned from the flame’ (Florios and Platziaflora, 813; AD 15)



ȱ ȱ ȱ ȱ ȱ ȱ ȱ

b. ǻΘΤȱΚ΅·΍ΣǼȱ ΈξΑȱȱ ΔΕνΔΉ΍ȱȱ ΑΤȱΘΤȱ ΘΕЏ·ΉΘΉȱ ȱϵΘ΅Αȱ Ύ΅ϟΓΙΗ΍Αȱȱȱ   (ta faja) dhen prepi na ta trojete otan keusin (the foods) not must to them eat.2PL when burn.ACT.PRES.3PL ‘you should not eat it (food) when it burns’ (Agapios Landos, Dietetics, 5, 131; AD 17) transitive active Early Medieval Greek c.ȱ ̳ΤΑȱȱ Ύ΅ϾΗϙȱȱ ΐΉȱȱ Θϲȱȱ ʌІȡ ȱ ȱ ean kafsi me to pir ȱ if burn.ACT.PERFVE.3SG I.ACC.WEAK the fire.NOM ȱ ‘if the fire burns me’ (Ioannes Malalas, Chronographia, 18, 16; AD 5-6) Late Medieval Greek d. ȱ ΘΓϿΖȱ σΎ΅ΜΉΑȱ ϳȱ ΔϱΌΓΖ ȱ tus ekapsen o pothos ȱ they.ACC.WEAK burn.ACT.PAST.PERFVE.3SG the passion.NOM ȱ ‘the love grieved them, the passion burned them’ (Livistros and Rodamni, 3732; AD 14)

The verb ıȕȒȞȦ zvino ‘to extinguish’ is referred to in the list by Chatzidakis (1927) as a product of causativisation which occurred in Modern Greek. From Medieval Greek data it appears, however, that

Diachronic Data and Analysis

147

the active causative verb was already in use in the Medieval Greek period, and at the same time as the non-active and active anticausative. (64) ΗΆφΑΝ zvino ‘to extinguish’ intransitive active a. ΐΤȱȂΑ΅ΑȱȱΎΉΕϠΑȱȱȱȱȱȱȱȱȱȱȱΦΚΘΓϾΐΉΑΓȱπΎΕΣΘΓΙΑȱȱΎ΍ȱȱόΗΆ΋ΗνȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΐΓΙȱȱȱ  ma ’nan kerin aftumeno ekratun ki izvise mu but one candle.NOM lighted.NOM hold.1SG

and extinguish.ACT.PAST.PERFVE.3SG I.GEN.WEAK

‘but I was holding a lighted candle and it was extinguished by my bad luck’ (The Sacrifice of Abraham, 408; AD 16) b. Ύ΅Ϡȱȱ ΌνΏΉ΍ȱȱ ΗΆϾΗϙȱȱ Θϲȱȱ Ύ΅ΎϲΑȱ ȱȱȱ ke theli zvisi to kakon and want.3SG extinguish.ACT.PERFVE.3SG the evil.NOM ‘and the evil is going to be extinguished’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 272, 128; AD 17) c. Έ΍Τȱ ΑΤȱ ΐχȱ ΗΆϾΗϙȱ ψȱȱ ΚΝΘϟ΅ȱ dhia na mi zvisi i fotia for to not extinguish.ACT.PERFVE.3SG the fire.NOM ‘so that the fire will not be extinguished’ (Agapios Landos, Geoponikon, 54, 18; AD 17) intransitive non-active πΎ΅ϧΘΓΑȱȱ ΑϾΎΘ΅Αȱȱ ψΐνΕ΅Αȱȱ ΔΣΑΘΓΘΉǰȱȱ d. ϳΔΓϿȱ opu eketon niktan imeran pandote, which.NOM burn.NACT.PAST.IMPERFVE.3SG night day always ΔΓΘνȱȱ

ΘΓΙȱȱ

ΓЁΎȱ

πΗΆΉΑΑνΘΓΑȱȱ

ȱ

pote tu uk ezvenetonȱ never it.GEN.WEAK not extinguish.NACT.PAST.IMPRFVE.3SG ‘that was always burning night and day, and was never extinguished’ (The War of Troy, 6393; AD 14)ȱ transitive active Ύ΍ȱȱȱȱΩΑΉΐΓΖȱȱȱȱȱȱΗΓІȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘϲȱȱȱ ȱȱȱȱȱȱȱΗΆφΗΉ΍ȱȱ e.ȱ ΎΉΕϠȱȱ keri ki anemos su to zvisi candle.ACC and wind.NOM you.GEN.WEAK it.ACC.WEAK extinguish.ACT.PERFVE.3SG ‘it is a candle and the wind will extinguish it’ (The Sacrifice of Abraham, 908; AD 16) ΑΤȱ ΘχΑȱȱ πΗΆφΗ΋Ζȱȱȱȱȱ f.ȱ ΅ΙΘχΑȱ ΘχΑȱ ΚΏϱ·΅Αȱȱ aftin tin floghan na tin ezvisis this the flame.ACC to she.ACC.WEAK extinguish.ACT.PERFVE.2SG ‘you should extinguish that flame’ (Leonardo Dellaporta, Poems, ǹ, 1663; AD 14-15) g.ȱ ǻψȱ΅Ε·ϟ΅Ǽȱȱ ΗΆϾΑΉ΍ȱȱ ΘχΑȱȱ σΐΚΙΘ΋Αȱȱ ΌνΕΐ΋Αȱ (i arghia) zvini tin emfitin thermin (the indolence) extinguish.ACT.PRES.3SG the innate warmness.ACC ‘(the indolence) extinguishes the innate warmness’ (Agapios Landos, Dietetics, 3, 23; AD 17)

(c) Active causative and anticausative The third group of verbs comprises verbs which bear the same voice morphology for both the causative and anticausative type. This is concerned, in many instances, with verbs which were active intransitive and were causativised in the Hellenistic Koine, but also with much earlier verbs which, in all probability, first appeared in the Medieval period with active morphology for both of the structures. The verbs ȕȡȐȗȦ vrazo ‘to boil’ and ȜȚȫȞȦ liono ‘to melt’ participate with active voice morphology in both causative and anticausative structure. For these verbs, there can be no certainty about whether they comprise the result of innovative causativisation or a change in voice morphology of the anticausative type, as data for these verbs in vernacular texts dating from before Medieval Greek are scarce. (65) ΆΕΣΊΝȱvrazo ‘to boil’ intransitive active ΑΤȱȱ ΐχȱȱ ΆΕΣΊϙ  a. ΔΏχΑȱȱ ΆΏνΔΉȱȱ plin vlepe na mi vrazi but watch.2SG to not boil.ACT.PRES.3SG ‘but watch so that it will not boil’ (Ptochoprodromos, ǿǿǿ, 128; AD 11-12) b. σΆΕ΅ΗΉΑȱȱ ψȱȱ Ύ΅ΕΈϟ΅ȱȱ ΐΓΙǰȱȱΔ΅ΕνΎΉ΍ȱ ΓЁΈξΑȱȱ Ά΅ΗΘΣΊΝȱȱȱ evrasen i kardhia mu, pareki udhen vastazo boil.ACT.PAST.PERFVE.3SG the heart.NOM my more not bear.1SG ‘my heart boiled, I cannot bear more’ (Michael Glykas, 386; AD 12) ΘΓȱȱ ΅ϟΐ΅ȱȱ νΆΕ΅ΗΉΑȱ c. ΉΙΌϾΖȱȱ

148

Chapter Three efthis to ema evrasen immediately the blood.NOM boil.ACT.PAST.PERFVE.3SG ‘the blood boiled immediately (his blood ran again/he became alive)’ (Apollonius of Tyre, The story of Apollonius (Kechagioglou), 137; AD 14) d. ǻΎΕ΍ΌΣΕ΍΅ǰȱϷΗΔΕ΍΅Ǽȱȱ ΆΕΣΊΓΙΑȱȱ ϴ·Ώ΍·ΝΕϱΘΉΕ΅ȱȱȱȱȱ (kritharia, ospria) vrazun oghlighorotera (barley, pulse) boil.ACT.PRES.3PL faster ‘(barley and pulse) boil faster’ (Agapios Landos, Geoponikon, 3, 6; AD 17) transitive active ΘξΖȱȱ ΈΓΑΘΣΎΕΉΖȱȱ e. Ύ΅Ϡȱȱ σΆΕ΅Ί΅Αȱȱ ke evrazan tes dhondakres and boil.ACT.PAST.IMPRFVE.3PL the.ACC pincers.ACC ‘and they were boiling the pincers’ (Leontios Machairas, Chronicle, 3, 440; AD 15) f. Θϲȱ ϳΔΓϧΓΑȱȱ ǻΐνΏ΍Ǽȱȱ ΆΕΣΗΉȱȱ ΔΕЗΘ΅ȱȱ ΐΓΑ΅ΛϱΑȱȱ to opion (meli) vrase prota monaxon the which.ACC (honey) boil.ACT.IMPER.PERFVE2SG first alone.ACC ‘boil it (the honey) first alone’ (Agapios Landos, Geoponikon, 50, 33; AD 17)ȱ ΆΕΣΗ΋Ζȱȱ ΦΑΘΣΐ΅ȱ ΐξȱ ΘΤȱȱ ϷΗΔΕ΍΅ g. ΑΤȱ ΘΤȱȱ na ta vrasis andama me ta ospria to they.ACC.WEAK boil.PRFVE.2SG together with the pulse ‘to boil them together with pulse’ (Agapios Landos, Dietetics, 32, 14; AD 17) h. Ύ΅ϠȱȱΦΑ΅ΎΣΘΝΑνȱȱȱȱȱΘΓȱȱȱȱȱȱȱȱΐΉȱȱȱρΑ΅ȱΒϾΏΓΑȱȱȱȱȱϵΗ΋ΑȱȱȱȱȱГΕ΅ΑȱȱΘϲȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΆΕΣΊΉ΍Ζȱȱȱ ke anakatone to me ena ksilon osin oran to vrazis and stir.IMPER.2SG it.ACC with one piece-wood as-much hour it.ACC.WEAK boil.ACT.PRES.2SG

‘and stir it with a piece of wood as you boil it’ (Agapios Landos, Dietetics, 83, 9-10; AD 17) (66) Ώ΍ЏΑΝȱliono ‘to melt’ intransitive active a. ΐξȱ ΑφΗΘΉ΍ΉΖȱΎ΅Ϡȱȱ ΐξȱȱΈν΋ΗΉΖǰȱΎ΍ȱȱ όΏ΍ΝΗΉΑȱȱ ψȱȱȱȱ Ύ΅ΘΣΕ΅ȱ me nisties ke me dheises, ki iliosen i katara with fasts and with prayers and melt.ACT.PAST.PERFVE.3SG the curse.NOM ‘the curse melted away with fasts and prayers’ (The Sacrifice of Abraham, 1142; AD 16) b. ΘΤȱΔΉΕ΅ΗΐνΑ΅ȱȱȱπΈ΍ΣΆ΋Η΅ΑȱΎ΅ϠȱȱΘΤȱȱ·Ε΅ΐΐνΑ΅ȱȱπΏ΍ЗΗ΅ǰȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱπΔΣΜ΅Η΍ȱȱΘΤȱȱΎΏ΅φΐ΅Θ΅ȱ ta perazmena edhiavisan ke ta ghramena eliosa, epapsasi ta klaimata the past-times.NOM went-by.3PL and the written.NOM melt.ACT.PASTPERFVE.3PL stopped.3PL the cries.NOM

‘past times went by and the written melted away, the cries stopped’ (The Sacrifice of Abraham, 1127; AD 16) transitive active ȱȱȱΎ΅ϠȱΕπΆΓΎΣΕΝȱȱΎ΅ϠȱȱΘχΑȱΎΓΐΉΗ΍ϲΑȱȱȱȱȱΈ΍ΤȱȱȱΘΓϾΘΓΙȱΘΓІȱΑΓΘ΅ΕϟΓΙȱ c. ΦΎϱΐ΋ǰȱȱȱȱȱΏ΍ЏΑΝȱȱ akomi, liono ke revokaro ke tin komesion dhia tutu tu notariu furthermore melt.ACT.PRES.1SG and revoke.1SG and the command.ACC through that this notary

‘furthermore, I rescind and revoke the command through a notary’ (Michail Maras, Katasticho, 149, ǹǯ, ȚșȞ 10; AD 16)

Finally, the verbs IJȡȠȝȐȗȦ tromazo ‘to scare’ and ĮȡȤȓȗȦ arxizo ‘to begin’ are active causatives and anticausatives in Medieval Greek. Chatzidakis (1927) includes them in the causativisations of Modern Greek, but Medieval Greek data contradict this and show that the same type (active) was in use in both the causative and the anticausative structure in the Medieval Greek. (67) ΘΕΓΐΣΊΝ tromazo ‘to scare’ intransitive active a. Ύ΅Ϡȱ ψȱ ΜΙΛφȱ ΐΓΙȱȱ πΘΕϱΐ΅ΒΉΑȱ ke i psixi mu etromaksen and the soul.NOM my scare.ACT.PAST.PERFVE.3SG ‘and my soul was scared’ (Digenis Akritas, Ǽscorial, 1113; AD 12) b. ΗφΐΉΕΓΑȱ πΗΆνΗΘ΋Αȱ ϳȱ ΓΙΕ΅ΑϱΖǰȱȱ πΘΕϱΐ΅ΒΉΑȱȱ ϳȱȱ ΎϱΗΐΓΖȱȱ simeron ezvestin o uranos, etromaksen o kozmos today erased.3SG the.NOM sky.NOM scare.ACT.PAST.PERFVE.3SG the.NOM people.NOM ‘today the sky was erased, the people scared’(Leonardo Dellaporta, Poems, ī, 349; AD 14-15) transitive active πΘΕΓΐΣΒ΅Η΍Αȱȱ ΘϲΑȱȱ ΐν·΅Αȱȱ ͑ΝΣΑΑ΋Α c. ΈΕΣΎΓΑΘΉΖǯǯǯȱȱ dhrakondes... etromaksasin ton meghan Ioanin dragons.NOM scare.ACT.PAST.PERFVE.3PL the great Ioanis.ACC ‘dragons... scared the great Ioannis’ (Leonardo Dellaporta, Poems, ǹ, 1414; AD 14-15)

Diachronic Data and Analysis

149

(68) ΅ΕΛϟΊΝȱarxizo to start’ intransitive active ϳȱȱ ΏΓ·΍ΗΐϱΖȱȱ ȱ ȱ a. ΘϱΘΉȱȱ όΕΛ΍ΗΉΑȱȱ tote irxisen o loghizmos then begin.ACT.PAST.PERFVE.3SG the reasoning.NOM ‘then the reasoning began’ (Leonardo Dellaporta, Poems, ǹ, 69; AD 14-15) transitive active Θ΋Αȱ ΅Ό΍ΆΓΏφȱ Γȱ ΒνΑΓΖ b. ΣΕΛ΍ΗΉΑȱȱ arxisen tin athivoli o ksenos begin.ACT.PAST.PERFVE.3SG the discussion.ACC the foreigner.NOM ‘the foreign man began the discussion’ (Apollonius of Tyre, Rimada, ǹ, 843; AD 14)

Moreover, the instances of Medieval Greek verbs, of which the corresponding earlier verbs were quite different, are significant, as they show a general tendency for active causative and anticausative, in other words, for the use of exactly the same type in the causative and anticausative structure. (69) Δ΅·ЏΑΝ paghono ‘to freeze’ intransitive active Δ΅·ЏΗϙȱȱ ϳȱȱ ΔΓΘ΅ΐϱΖ a. ϵΘ΅Αȱ otan paghosi o potamos when freeze.ACT.PERFVE.3SG the river.NOM ‘when the river freezes’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 347, 2; AD 17) transitive active Early Medieval Greek ΘχΑȱȱ ΌΣΏ΅ΗΗ΅Αȱ ȱ ȱ  b. ГΗΘΉȱ Δ΅·ЗΗ΅΍ȱȱ   oste paghose tin thalasan    so freeze.ACT.PERFVE.INF the sea.ACC ‘to freeze the sea’ (Theophanes Confessor, Chronicle, 297, 11; AD 8-9) 

(d) Change in voice (active from non-active anticausative) Regarding the following verbs change in voice morphology of the anticausative verb can be ascertained. It is not concerned here, consequently, with innovative causativisation or with innovative anticausativisation, but with change in the voice of verbs that can have causative and anticausative use. The changes have to do with anticausative structure and are in the direction of the expression of the anticausative with active voice. The marking of the anticausative type with active voice in parallel with the earlier non-active voice is frequent. It can be concluded, therefore, that the changes in the area of voice are not concerned solely with the Hellenistic-Roman period. The verb ʌȜĮIJȪȞȦ platino ‘to widen’ begins to mark the anticausative structure with active morphology, in other words, to use exactly the same type for both the causative and the anticausative structure, during the Medieval Greek period. (70) intransitive active a. ώΏΌΉΑȱ ψȱ ΑϾΒ... ȱ πΔΏΣΘΙΑΉΑȱȱ Θϲȱȱ ΗΎϱΘΓΖȱȱ ilthen i niks... eplatinen to skotos came.3SG the night.NOM widen.ACT.PAST.PERFVE.3SG the darkness.NOM ‘the night came... the darkness widened’ (Kallimachos and Chrysorroi, 2229; AD 14) b. πΔΏΣΘΙΑΉΑȱȱ ϳȱϊΏ΍ΓΖǰȱȱ σΏ΅ΐΜΉΑȱȱ Δ΅ΑΘ΅ΛϱΌΉΑȱȱȱ ȱȱȱȱ   eplatinen o ilios, elampsen pandaxothen widen.ACT.PAST.PERFVE.3SG the sun.NOM shone.3SG everywhere ‘the sun spread, shone everywhere’ (Florios and Platziaflora, 635; AD 15) intransitive non-active Θϲȱȱ ΗΘϱΐ΅ȱȱ ΘΓΙȱȱ БΖȱϳȱΚΓϾΕΑΓΖȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ c. πΔΏ΅ΘϾΑΌ΋ȱȱ eplatinthi to stoma tu os o furnos widen.NACT.PAST.PERFVE.3SG the mouth.NOM his as the oven ‘his mouth widened like an oven’ ǻSpanos, ǹ, 76; AD 14/15) d. Ύ΅ϠȱΘΤȱȱȱΈΉ΍ΑΤȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ Δ΅ΕνΈΕ΅ΐΓΑȱȱ Ύ΅ϠȱȱψȱȱȱΛ΅ΕΤȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ πΔΏ΅ΘϾΑΌ΋ȱȱȱȱȱȱȱȱȱ ke ta dhina paredhramon ke i xara eplatinthi and the misfortunes.NOM passed.3PL

and the happiness.NOM widen.NACT.PAST.PERFVE.3SG

‘and the misfortunes passed and happiness spread’ (Velthandros and Chrysandza, 1303; AD 14) transitive active e. ΘΓϿΖȱȱ Ώϱ·ΓΙΖȱȱ ΐΓΙȱȱ ΔΏ΅ΘϾΑΝ   tus loghus mu platino the words.ACC my widen.ACT.PERFVE.1SG ‘I widen my words’ (Livistros and Rodamni, 3055; AD 14)

150

Chapter Three

The verbs ʌĮȪȦ pavo ‘to end’, İȣijȡĮȓȞȦ efreno ‘to gladden’ and ȡȒȖȞȣȝȚ righnimi ‘to break/shatter’ have an active causative type in Medieval Greek and the availability of free alternation between the non-active (earlier) and active (later) anticausative type. (71) Δ΅ϾΝȱpavo ‘to end’ intransitive active a. ΑΤȱȱ Δ΅ϾΗΓΙΑȱȱ ΅ϡȱȱ ϴΈϾΑ΅΍ȱȱ ΐΓΙȱȱȱȱȱ na pafsun e odhine mu to end.ACT.PERFVE.3PL the.NOM dolors.NOM my ‘so that my dolors end’ (Michael Glykas, 163; AD 12) b. Ύ΍ȱȱ ΪΖȱȱ ΔΣΜΉ΍ȱ ΅ЁΘ΅χȱȱ ψȱȱ ΘΕΓΐΣΕ΅ȱȱ ki as papsi aftai i tromara and let end.ACT.PERFVE.3SG this the terror.NOM ‘let this terror end’ (The Sacrifice of Abraham, 150; AD 16) intransitive non-active Ά΅Η΍ΏΉϾΓΑΘΓΖȱȱ ϳȱ ΔϱΏΉΐΓΖȱ πΔ΅ϾΌ΋ c. ψΏϟΓΙȱȱ iliu vasilevondos o polemos epafthi sun.GEN sink.PART.GEN the war.NOM end.NACT.PAST.PERFVE.3SG ‘when the sun sank, the war ended’ (An Entertaining Tale of Quadrupeds, 1075; AD 14) transitive active Δ΅ϾΗϙȱȱ ΘχΑȱȱ Θ΅Ε΅ΛφΑȱȱ d. Έ΍Τȱȱ ΑΤȱ dhia na pafsi tin taraxin for to end.ACT.PERFVE.3SG the disorder.ACC ‘to end the disorder’ (Leontios Machairas, Chronicle, 3, 314; AD 15) (72) ΉЁΚΕ΅ϟΑΝȱefrenoȱ‘to gladden’ȱ intransitive non-active Early Medieval Greek οΓΕΘΣΊΓΑΘΉΖȱȱ ΉЁΚΕ΅΍ΑϱΐΉΌ΅ȱȱ a.ȱȱ ΦΉϠȱȱ ai eortazondes efrenometha always celebrate.PART.NOM gladden.NACT.PRES.1PL ‘we always celebrate and gladden’ (Romanus Melodus, Cantica, 41, 4; AD 6) Late Medieval Greek b. πΑΉ·΅ΏΏ΍ΣΗΌ΋Αȱȱ ψȱȱ Ύ΅ΕΈϟ΅ȱȱ ΘΓΙȱȱ Ύ΅Ϡȱȱ eneghaliasthin i kardhia tu ke rejoice-exceedingly.NACT.PASTPERFVE.3SG the heart.NOM his and ΉЁΚΕΣΑΌ΋Αȱȱ

ψȱȱ

ϷΜ΍Ζȱȱ

ΘΓΙȱȱ

efranthin i opsis tuȱ gladden.NACT.PAST.PERFVE.3SG the appearance.NOM his ‘his heart rejoiced exceedingly and his appearance gladdened’ (Spanos, ǹ, 229; AD 14/15) intransitive active ψȱȱ Ύ΅ΕΈ΍Σȱȱ ΐΓΙȱȱȱ  c. ΋ЄΚΕ΅ΑΉȱȱ   ifrane i kardhia mu  gladden.ACT.PAST.PERFVE.3SG the heart.NOM my ‘my heart gladdened’ (Ptochoprodromos, ǿ, 266; AD 11-12)  transitive active Early Medieval Greek ΗΓΚϲΖȱȱ ΉЁΚΕ΅ϟΑΉ΍ȱ Δ΅ΘνΕ΅ȱȱȱȱȱȱȱȱȱ d. ΙϡϲΖȱȱ ios sofos efreni patera son.NOM wise.NOM gladden.ACT.PRES.3SG father.ACC ‘a wise son makes his father gladden’ (Constantine VII Porphyrogenitus, De administrando imperio, p, 2; AD 10) (73) Εφ·ΑΙΐ΍ȬΕ΅ǻ·ǼϟΊΝ righnimi-ra(j)izo ‘to break/shatter’ȱ intransitive non-active Early Medieval Greek ΑІΑȱ ϳȱȱ ΔΉΑΌφΕ΋Ζȱ Λ΍ΘЏΑȱ a. е̔φ·ΑΙΘ΅΍ȱȱ Righnite nin o penthiris xiton shatter.NACT.PRES.3SG now the mourning.NOM chiton.NOM ‘the mourning chiton is rending now’ (Romanus Melodus, Cantica, 17, 12, 1; AD 6) Late Medieval Greek b. ψȱȱ Ύ΅ΕΈ΍Σȱȱ ΐȂȱȱ πΕΕΣ·΋ȱȱȱ ȱ i kardhia m’ eraji the heart.NOM my shatter.NACT.PAST.PERFVE.3SG

Diachronic Data and Analysis

151

‘my heart broke’ (The Sacrifice of Abraham, 789; AD 16) intransitive active ΗΓΙȱȱ Ε΅ϬΗΓΙΗ΍ȱȱȱȱ  c. Ύ΅Ϡȱȱ ΘΤȱȱ ΔΏΉΙΕΣȱȱ ke ta plevra su raisusi and the sides.NOM your shatter.PERFVE.3PL ‘and your sides will break’ (Spanos, ǻ, 309; AD 14/15) d. Ύ΅ϠȱΦΚȂȱȱΘχΑȱΌΏϧΜ΍ΑȱȱȱΘχΑȱȱȱΔΓΏΏχΑȱΎȂȱȱȱȱΓϡȱȱȱ ΔνΘΕΉΖȱȱȱȱȱȱȱȱȱȱπΕΕΣ·΍Η΅Α   ke af’ tin thlipsin tin polin k’ i petres erajisan    and from the sorrow the much and the stones.NOM shatter.ACT.PAST.PERFVE.3PL ‘even the stones shattered from so much sorrow’ (Velthandros and Chrysandza, 1279; AD 14) e. Ύ΅Ϡȱ πΕΣ·΍ΗΉΑȱ ϳȱ ΔϾΕ·ΓΖȱ ΘΓІȱ ΐΓΑ΅ΗΘ΋Ε΍ΓІȱȱȱ ke erajisen o pirghos tu monastiriu and break.ACT.PAST.PERFVE.3SG the tower.NOM the monastery.GEN ‘and the monastery tower shattered’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) I, 337; AD 17)

That the presence of both active and non-active intransitive types is a case of voice alternation due to synchronic variation which leads to change is also shown by the use of non-active endings with the earlier stem of the verb (ȡȘȖȞ- righn-), and of active endings with the newer stem of the verb (ȡĮȖragh-), and in both instances with exactly the same subject choices and the same meaning (ȡȒȖȞȣIJĮȚ Ș țĮȡįȚȐ / righnite i kardhia / break.NACT.PRES.3SG the.NOM heart.NOM / ‘the heart breaks’ – ȡĮȖȓȗİȚ Ș țĮȡįȚȐ / rajizi i kardhia / break.ACT.PRES.3SG the.NOM heart.NOM / ‘the heart breaks’). It is worth noting that the same phenomenon is also observed with the verb ıĮʌȓȗȦ sapizo rot – ıȒʌȦ sipo rot. The earlier stem of the verb (ıȘʌ- sip-) is combined with non-active voice, while at the same time the newer stem of the verb (ıĮʌȚȗ- sapiz-) bears active voice morphology. (74) earlier stem ıȘʌ- sip- + non-active voice ΉϢΖȱȱ ΎΕ΍ΌΣΕ΍ǰȱȱ ΈξΑȱȱ ΗφΔΓΑΘ΅΍ a. πΤΑȱȱ Έξȱȱ ΔΣΏ΍Αȱȱ ΛЏΗϙΖȱȱ ΘΤȱΎϾΘΕ΅ȱȱ  ean dhe palin xosis ta kitra is krithari, dhen siponde   if prt again bury.2SG the citrons.ACC to barley not rot.NACT.PRES.3PL  ‘if you bury the citrons in barley, they do not rot’ (Agapios Landos, Geoponikon, 20, 24; AD 17) b. ϵΗ΅ȱΒϾΏ΅ȱȱȱȱȱȱȱȱȱȱȱΎϱΜΉ΍Ζȱȱȱȱ΅ЁΘϲΑȱΘϲΑȱΐϛΑ΅ǯǯǯȱȱΔΓΘνȱȱΘΓΙΖȱȱ ȱȱȱȱȱȱȱȱȱȱȱȱȱȱΈξΑȱȱȱȱȱΗφΔΓΑΘ΅΍ȱȱ osa ksila kopsis afton ton mina... pote tus dhen siponde as wood.ACC cut.2SG this the month never it.ACC.WEAK not rot.NACT.PRES.3PL ‘as much wood as you cut this month... it will never rot’ (Agapios Landos, Geoponikon, 32, 7; AD 17) newer stem ıĮʌȚȗ- sapiz- + active voice ΈξΑȱ Η΅ΔϟΊΓΙΑȱȱ ϴ·Ώϟ·ΝΕ΅ȱȱȱȱȱȱȱȱ c. Ύ΅Ϡȱȱ ΩΏΏ΅ȱ ΗΘ΅ΚϾΏ΍΅ǯǯǯȱ ϳΔΓІȱȱȱ ke ala stafilia... opu dhen sapizun oghlighora and other grapes.NOM which.NOM not rot.ACT.PRES.3PL fast ‘and other grapes... which do not rot fast’ (Agapios Landos, Geoponikon, 41, 2; AD 17) ȱ d. ΌνΏΉ΍Ζȱȱ ΑΤȱΚΙΏΣΒϙΖȱȱ ΘΤȱȱΦ·ΎΓϾΕ΍΅ȱȱȱȱȱȱȱȱȱȱȱȱΔΓΏϿΑȱȱΎ΅΍ΕϲΑȱΑΤȱȱȱΐχȱȱ Η΅ΔφΗΝΗ΍ȱ thelis na filaksis ta anguria polin keron na mi sapisosi want.2SG to conserve.2SG the cucumbers.ACC long time to not rot.ACT.PERFVE.3PL ‘you want to conserve the cucumbers for a long time so that they will not rot’ (Agapios Landos, Dietetics, 54, 27; AD 17)

The above show that the participles ending in -ȝİȞȠȢ -menos of those verbs that are in evidence from the Medieval period arise from the parallel use of non-active voice for those verbs (for analysis of this issue, cf. Moser 1988, 2005; Manolessou 2005). (75) a. ΋ЈΕ΅ȱȱ ΘχΑȱȱ ivra tin found.1SG the

πΎΎΏ΋Ηϟ΅Αȱ

ϵΏ΋Αȱ

eklisian church.ACC

olin rajizmenin whole.ACC shatter.NACT.PAST.PERFVE.PART.ACC.SG

Ϲ΅·΍ΗΐνΑ΋Α

‘I found the whole church shattered’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) I, 332; AD 17)

b. ΑΤȱ ΐχΑȱ σΛΓΙΑȱȱ Ϲϱ·΅΍Ζȱȱ Η΅Δ΋ΐνΑ΅΍Ζ na min exun rojes sapimenes to not have.3PL berries.ACC rot.NACT.PAST.PERFVE.PART.ACC.PL ‘so that they do not have rot berries’ (Agapios Landos, Geoponikon, 41, 15; AD 17)

A prefix is also added to the verb ȡĮȖȓȗȦ rajizo ‘to crack’ in many instances to form either a causative or anticausative structure: (76) prefix in the causative type

152

Chapter Three Early Medieval Greek a.ȱ Ά΅ΗΎ΅Αϟ΅ǯǯǯȱȱ Έ΍΅ΕΕ΋·ΑϾΑ΅΍ȱȱ vaskania... dhiarighnine evil-eye.NOM

DIA-shatter.ACT.PRES.INF

ΘΓϿΖȱȱ

ΘϛΖȱȱ

Κ΍Ώϟ΅Ζȱȱ

ΌΉΗΐΓϾΖȱȱ

tus

tis

filias

thezmus

the.ACC

the.GEN friendship.GEN institutions.ACC

‘the evil eye... that rends the institution of friendship apart’ (Constantine VII Porphyrogenitus, de legationibus, 193, 26; AD 10) prefix in the anticausative type Early Medieval Greek b.ȱȱ Ύ΅ΌΤǯǯǯȱȱȱȱȱȱȱȱΔΕΓΗΎΙΑΓІΐΉΑǰȱΎΪΑȱ ΓϡȱȱȱȱȱȱȱȱȱȱȱΛΕ΍ΗΘΓΐΣΛΓ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱ Έ΍΅ΕΕφ·ΑΙΑΘ΅΍ katha... proskinumen, kan i xristomaxi dhiarighninde which.ACC worship.1PL even-if the.NOM Jesus-opponents.NOM DIA-shatter.NACT.PRES.3PL ‘which... we worship, even if the opponents of Jesus are breaking themselves apart’ (George the Monk, Chronicle, 773, 17; AD 9)

3.3.4 From Medieval to Modern Greek 3.3.4.1 Causativisations The aim of this section is not the analysis of the synchrony of Modern Greek -which, in any case, has been examined systematically by Generative Grammar and other approaches. I attempt here to complete the presentation of the diachronic course of transitivity alternations by giving some evidence of the ongoing changes which are observed in Modern Greek (based on the comparison with the previous language system of Medieval Greek). The causativisations which appear in Modern Greek are, in literature, a well-known subject; they were treated by Chatzidakis some eighty years ago and Theophanopoulou-Kontou more recently. Chatzidakis (1927) and Theophanopoulou-Kontou (2003a) refer to instances of causativisation in Modern Greek, in other words to intransitive verbs which begin to be used, for the first time, as transitives with a patient argument in the object position. In the analyses, it is noted that causativisations in Modern Greek are frequently dealt with by native speakers and grammarians as a wrong choice and with marginal usage. A look at diachronic data indicates analogous changes in other periods of Greek as well, and indeed indicates how essential for the language system is the mechanism of the introduction of an external argument to the intransitive/anticausative structure. In chapter 4, I will return to this issue and attempt to interpret this frequently appearing mechanism of causativisation59. Regarding the semantic group of verbs which are causativised, the greater part of the list of verbs given by Chatzidakis are psych-verbs, whereas Theophanopoulou-Kontou mainly examines motion verbs. Of particular interest is the fact that new causative types in Greek (and cross-linguistically) are in evidence only for certain motion verbs (IJȡȑȤȦ trexo ‘to run’, ʌİȡʌĮIJȫ perpato ‘to walk’ – Ƞ įȚİȣșȣȞIJȒȢ ȝİ IJȡȑȤİȚ ĮıIJĮȝȐIJȘIJĮ / o dhiefthindis me trexi astamatita / the.NOM director.NOM I.ACC.WEAK run.ACT.PRES.3SG continuously / ‘the director makes me run continuously’) whereas they are absent from others (țȠȜȣȝʌȫ kolimbo ‘to swim’, ıțĮȡijĮȜȫȞȦ skarfalono ‘to climb’). Theophanopoulou-Kontou (2003a) argues that the motion verbs in Modern Greek in the framework of unaccusativity theory can be separated into unaccusative and unergative with a basic criterion being the presence/absence of transitivity alternations, and that transitive types, where they are encountered, consist of structures analogous with those that are in evidence with unaccusative verbs that denote change-of-state or are considered idiosyncratic. The causative use of motion verbs constitutes, according to Theophanopoulou-Kontou, a widening of their semantic structure with the addition of an abstract causative verb. The widening of the semantic structure of the verb results in the assignment of the agent/external cause ș-role and in the widening of the syntactic structure with the necessary presence of the functional head (v) specified as causative. 59

For the transitive and intransitive use of verbs in Modern Greek, cf. also Mackridge (1990: 154-155), and Bousboukis (1995), who also refers to the metaphor in poetic speech: transitive use of otherwise intransitive verbs ijİȪȖȦ IJȠȣ ĮȜȩȖȠȣ IJȘȞ ȠȡȝȒ țĮȚ IJȠȣ ıʌĮșȚȠȪ IJȠȞ IJȡȩȝȠ fevgho tu aloghu tin ormi ke tu spathiu ton tromo leave.ACT.PRES.1SG the horse.GEN the impetus.ACC and the sword.GEN the terror.ACC ‘I flee away from the impetus of the horse and the terror of the sword’ (Solomos, Free Besieged, The Temptation; AD 19) b. ȖȚĮ ȞĮ ıİ țȠȚȝȘșȫ ʌĮȡȐȞȠȝĮ ja na se kimitho paranoma for to you.ACC.WEAK sleep.NACT.PERFVE.1SG illegally ‘to sleep with you illegally’ (Elytis, The Light Tree and the Fourteenth Beauty; AD 20)

(1) a.

Diachronic Data and Analysis

153

(a) Causative from active intransitive type The overwhelming majority of verbs which are causativised in Modern Greek are initially intransitive with active morphology. This is to be expected since active morphology (when it is the exclusive voice morphology or when it appears in free distribution with non-active voice morphology in an intransitive verb) favours the reanalysis process of the intransitive verb. Chatzidakis (1927) has described as innovative (for the language of his time) the causative uses of the following (among other) verbs, which were active intransitives: (77) Chatzidakis (1927): ȘıȣȤȐȗȦ isixazo calm.ACT ȘıȪȤĮıĮ IJȠ isixasa to calm.ACT.PAST.PERFVE.1SG the.ACC ‘I calmed the child’

ʌĮȚįȓ pedhi child.ACC

IJȡȠȝȐȗȦ tromazo scare.ACT ȝ’ İIJȡȩȝĮȟİȞ m’ etromaksen I.ACC.WEAK scare.ACT.PAST.PERFVE.3SG ‘he scared me’ șȣȝȫȞȦ thimono anger.ACT IJȠȞ İșȪȝȦıİȢ ʌȐȜȚ ton ethimoses pali he.ACC.WEAK anger.ACT.PAST.PERFVE.2SG again ‘you made him angry again’ ʌİșĮȓȞȦ petheno die.ACT IJȠȞ ʌȑșĮȞİ ıIJȠ ȟȪȜȠ ton pethane sto ksilo he.ACC.WEAK die.ACT.PAST.PERFVE.3SG to beating ‘he made him die by beating him’ ȖȣȡȓȗȦ ʌȓıȦ / jirizo piso / return.ACT IJȠȞ ȖȪȡȚıĮ ton jirisa he.ACC.WEAK return.ACT.PAST.PERFVE.1SG ‘I made him return’

ʌȓıȦ piso back

ȟİIJȡȣʌȫȞȦ ksetripono flush-out.ACT IJȠȞ İȟİIJȡȪʌȦıĮȞ ton eksetriposan he.ACC.WEAK flush-out.ACT.PAST.PERFVE.3PL ‘they flushed him out’ ȖİȡȞȐȦ jernao get-old.ACT ȝİ İȖȑȡĮıĮȞ IJĮ me ejerasan ta I.ACC.WEAK get-old.ACT.PAST.PERFVE.3PL the ‘the torments made me get old’

ȕȐıĮȞĮ vasana torments.NOM

ȝİȖĮȜȫȞȦ meghalono grow.ACT İȖȫ ı’ İȝİȖȐȜȦıĮ ego s’ emeghalosa I.NOM you.ACC.WEAK grow.ACT.PAST.PERFVE.1SG ‘I grew you up’ ĮȞȐȕȦ anavo light.ACT ȐȞĮȥİ IJȠ anapse to light.ACT.PAST.PERFVE.3SG the.ACC ‘he lit the oil lamp’ ıȕȒȞȦ zvino extinguish.ACT IJȠȞ ȑıȕȘıĮȞ ton ezvisan

ȜȣȤȞȐȡȚ lixnari oil-lamp.ACC

IJĮ ta

ȑȟȠįĮ eksodha

154

Chapter Three

he.ACC.WEAK extinguish.ACT.PAST.PERFVE.3PL the ‘expenses wiped him out’

expenses.NOM

ȖȚȐȞȦ jano cure.ACT İȖȫ șĮ ıİ ȖȚȐȞȦ egho tha se jano I.NOM will you.ACC.WEAK cure.ACT.PRES.1SG ‘I will cure you’ ȖȜȣțĮȓȞȦ ghlikeno sweeten.ACT İȖȫ șĮ IJȠȞ egho tha ton I.NOM will he.ACC.WEAK ‘I will sweeten him’

ȖȜȣțȐȞȦ ghlikano sweeten.ACT.PERFVE.1SG

ʌİȚșĮȡȤȫ pitharxo obey.ACT ĮȣIJȩȢ ʌİȚșĮȡȤİȓ aftos pitharxi he.NOM obey.ACT.PRES.3SG ‘he makes the horse obey’

IJȠȞ ton the

ȓʌʌȠ ipo horse.ACC

Theophanopoulou-Kontou (2003a) indicates the productivity of Modern Greek causativisations with meanings as ‘to make’, ‘to force/oblige, ‘to accompany’. Among the examples referred to, the following causative verbs are the result of the causativisation of intransitive active verbs: (78) Theophanopoulou-Kontou (2003a): IJȘȞ ʌİȡʌȐIJȘıİ ıİ ȩȜĮ IJĮ ȝȑȡȘ ʌȠȣ İȓȤĮȞ IJȩıȠ ĮȖĮʌȒıİȚ tin perpatise se ola ta meri pu ixan toso aghapisi she.ACC.WEAK walk.ACT.PAST.PERFVE.3SG to all the places which had.3PL so loved.3PL ‘he walked her / accompanied her in all places which they had so much loved’ Ƞ ȈȘȝȓIJȘȢ o Simitis the Simitis.NOM ‘Simitis makes Laliotis run’

IJȡȑȤİȚ IJȠȞ trexi ton run.ACT.PRES.3SG the

ȁĮȜȚȫIJȘ Lalioti Laliotis.ACC

įȚĮȡȡȑȦ ʌȜȘȡȠijȠȡȓİȢ dhiareo plirofories leak.ACT.PRES.1SG information.ACC ‘I reveal some information’ țĮȜʌȐȗȦ kalpazo ‘to gallop’ ıİȡȖȚĮȞȓȗȦ serjanizo ‘to perambulate’ IJĮȟȚįİȪȦ taksidhevo ‘to travel’ țȣțȜȠijȠȡȫ kikloforo ‘to go round’ ĮȣIJȠțIJȠȞȫ aftoktono ‘to commit suicide’

It should be noted that neither are causativisations in Modern Greek connected with changes in voice morphology. The intransitive verbs which are causativised are in active voice and the introduction of the external argument is without consequences for the voice of the causative and anticausative type. If we compare the behaviour of verbs which are causativised in Modern Greek with their corresponding behaviour in Medieval Greek, we can observe the following: The verb ȦȡȚȝȐȗȦ orimazo ‘to mature’, which is causativised in Modern Greek, is consistently intransitive in Medieval Greek. (79) innovative causative use a. IJȠȞ ȦȡȓȝĮıĮȞ ȠȚ ton orimasan i he.ACC.WEAK mature.ACT.PAST.PERFVE.3PL the ‘the difficulties made him mature’ b. ȠȚ įȣıțȠȜȓİȢ ȦȡȓȝĮıĮȞ i dhiskolies orimasan the difficulties.NOM mature.ACT.PAST.PERFVE.3PL ‘the difficulties made Janis mature’ i in the previous period - Medieval Greek: intransitive active

įȣıțȠȜȓİȢ dhiskolies difficulties.NOM IJȠȞ ton the

īȚȐȞȞȘ Jani Janis.ACC

Diachronic Data and Analysis

155

c. ΦΚΓІȱȱ БΕ΍ΐΣΗΓΙΑȱȱ ΘΤȱȱ ΗΘ΅ΚϾΏ΍΅ȱȱȱȱȱ afu orimasun ta stafilia after mature.ACT.PERFVE.3PL the.NOM grapes.NOM ‘after the grapes mature’ (Agapios Landos, Geoponikon, 13, 10 – AD 17) d. ǻΩ·Ε΍΅ȱΐΓІΗΎΏ΅ǼǯǯǯȱΩΚΉΖȱΘ΅ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΑΤȱБΕ΍ΐΣΒΓΙΑǯǯǯȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΑΤȱȱȱΗΘ΅Κ΍Έ΍ΣΗΓΙΑȱ (aghria muskla)... afes ta na orimaksun... na stafidhiasun (wild mosses) let they.ACC.WEAK to mature.ACT.PERFVE.3PL to dry-up.3PL ‘let them (wild mosses) ripe and dry up’ (Agapios Landos, Geoponikon, 43, 2-3; AD 17)

The verb ȖİȡȞȐȦ jernao ‘to get old’ is causativised in Modern Greek in relation to its consistently intransitive use during the Medieval Greek period. It is concerned, however, with an instance of verb re-causativisation since the verb was initially used in causative structures in Classical Greek. Its causative use was further abandoned, to be re-used with an external argument in Modern Greek. Notably, the verb is included in the causativisations which Chatzidakis mentions in the language of his time. The process of causativisation appears not to have been completed since the structure with clitic as direct object is the only one that is grammatical. (80) innovative causative use a. IJȠȞ ȖȑȡĮıĮȞ ȠȚ țĮțȠȣȤȓİȢ ton jerasan i kakuxies he.ACC.WEAK get-old.ACT.PAST.PERFVE.3PL the hardships.NOM ‘the hardships aged him’ b. ?ȠȚ IJĮȜĮȚʌȦȡȓİȢ ȖȑȡĮıĮȞ IJȠȞ Įįİȡijȩ ?i talepories jerasan ton adherfo the discomforts.NOM get-old.ACT.PAST.PERFVE.3PL the brother.ACC ‘the discomforts aged my brother’ i in the previous period - Medieval Greek: intransitive active Δ΅΍ΈϟΑǰȱȱ π·φΕ΅Η΅ c.ȱ όΐΓΙΑȱȱ imun pedhin, ejirasa was.1SG child.NOM get-old.ACT.PAST.PERFVE.1SG ‘I was a child, I got old’ (Michael Glykas, 1; AD 12) d.ȱ ϵΘ΅Αȱȱ ·΋ΕΣΗϙΖȱȱ Ύ΅Ϡȱȱ πΗϾȱȱȱ ȱ otan jirasis ke esi when get-old.ACT.PERFVE.2SG and you.NOM ‘when you also get old’ (An Entertaining Tale of Quadrupeds, 784; AD 14)

ȝȠȣ mu my

The psych-verbs ȘıȣȤȐȗȦ isixazo ‘to rest/quieten’ and Șȡİȝȫ iremo ‘to calm’ were intransitive nonalternating in Medieval Greek but take a DP (denoting change-of-state) as direct object in Modern Greek. As the causative type of the verb ȘıȣȤȐȗȦ isixazo ‘to rest/quieten’, the verb țĮșȘıȣȤȐȗȦ kathisixazo ‘to appease/put sb.’s mind at rest’ (with the addition of the prefix KATA-) is also used60. (81) innovative causative use a. ʌȡȠıʌȐșȘıİ ȞĮ ȝİ ȘıȣȤȐıİȚ prospathise na me isixasi tried.3SG to I.ACC.WEAK quieten.ACT.PERFVE.3SG ‘s/he tried to quieten me’ b. țĮșȘıȪȤĮıİ IJȘ ȂĮȡȓĮ țĮȚ ȩȜĮ kathisixase ti Maria ke ola KATA-quieten.IMP.2SG the Maria.ACC and everything.NOM ‘appease Maria and everything will be fine’ i in the previous period - Medieval Greek: intransitive active

60

șĮ ʌȐȞİ țĮȜȐ tha pane kala will go.3PL fine

As in all instances of transitivity alternations the selectional restrictions play an important role: (1) ??țĮșȘıȪȤĮıİ (İıȪ) IJȠ ??kathisixase (esi) to KATA-quieten.ACT.PERFVE.IMP.2SG (you) ‘appease the baby’ (2) *IJȠ ȝȦȡȩ *to moro the baby.NOM ‘the baby appeased’

ȝȦȡȩ moro the.ACC

țĮșȘıȪȤĮıİ kathisixase KATA-quieten.ACT.PAST.PERFVE.3SG

baby.ACC

156

Chapter Three



Early Medieval Greek c. Ύ΅Ϡȱȱ ЀΔΣ·Νȱȱ ΉϢΖȱΘχΑȱσΕ΋ΐΓΑǰȱȱ Ύ΅Ϡȱȱ ψΗΙΛΣΊΝ ke ipagho is tin erimon, ke isixazo and go.1SG to the desert and quieten.ACT.PRES.1SG ‘and then I go to the desert and become calm’ (John Moschos, Leimonarion (The Spiritual Meadow), ȋȋ, B, 4; AD 6-7) d. Ύ΅Ϡȱȱ ωΗϾΛ΅ΗΉΑȱȱ ψȱȱ ΔϱΏ΍Ζȱȱȱȱȱ ȱ   ke isixasen i polis and quieten.ACT.PAST.PERFVE.3SG the city.NOM ‘and the city quieted’ (Theophanes Confessor, Chronicle, 286Ǻ, 1; AD 8-9) 

(82) innovative causative use a. Ș ijȦȞȒ IJȘȢ ȝȘIJȑȡĮȢ ȘȡȑȝȘıİ IJȠ ʌĮȚįȓ i foni tis miteras iremise to pedhi the voice.NOM the mother.GEN calm.ACT.PAST.PERFVE.3SG the child.ACC ‘the voice of the mother calmed the child’ b. ȝİ ȘȡȑȝȘıĮȞ IJĮ ȜȩȖȚĮ IJȘȢ me iremisan ta loja tis I.ACC.WEAK calm.ACT.PAST.PERFVE.3PL the words.NOM her ‘her words calmed me’ i in the previous period - Medieval Greek: intransitive active ωΎϱΐΔ΋ΗΉΖǰȱȱ σΔΉΗΉΖȱȱȱȱ c. ωΕνΐ΋ΗΉΖǰȱȱ iremises, ikombises, epeses calm.ACT.PAST.PERFVE.2SG lay.ACT.PAST.PERFVE.2SG fall.ACT.PAST.PERFVE.2SG ‘you calmed down, laid down, fell down’ (Michael Glykas, 165; AD 12)

It is also worth noting the causativisation of the intransitive verbs ȜȒȖȦ ligho ‘to end/expire and ȜȚȖȠıIJİȪȦ lighostevo ‘to reduce’. The verbs were consistently intransitive during the Medieval Greek period but begin to be used as causatives in Modern Greek. There is not, however, complete agreement among native speakers over the newer causative structure. (83) innovative causative use a. Ƞ įȚĮȚIJȘIJȒȢ ȑȜȘȟİ IJȠȞ ʌȠįȠıijĮȚȡȚțȩ o dhietitis elikse ton podhosferiko the referee.NOM end.ACT.PAST.PERFVE.3SG the football ‘the referee ended the football match’ (example from television) b. Ș ĮȞIJȓįȡĮıȒ IJȠȣ ȜȚȖȩıIJİȥİ IJȚȢ ʌȚșĮȞȩIJȘIJİȢ i andidhrasi tu lighostepse tis pithanotites

ĮȖȫȞĮ aghona match.ACC ıȣȞİȞȞȩȘıȘȢ sinenoisis

the reaction.NOM his reduce.ACT.PAST.PERFVE.3SG the possibilities.ACC understanding.GEN

‘his reaction reduced the possibilities for understanding’ i in the previous period - Medieval Greek: intransitive active c. Early Medieval Greek ȱȱȱȱȱȱ ȱ ȱȱȱȱȱȱȱȱȱȱȱ

ǻϳȱΔϱΏΉΐΓΖǼȱȱ ΉϢΖȱȱ ΘΓІΘΓȱȱ Θϲȱ

ΘνΏΓΖȱȱ ΘΓІȱȱ

ΔΉΑΘΉΎ΅΍ΈΉΎΣΘΓΙȱȱ σΘΓΙΖȱȱ

(o polemos) (the war)

telos end

pendekedhekatu fifteen

ΘϛΖȱȱ

is in

tuto this

to the

̝Α΅ΗΘ΅ΗϟΓΙȱȱ

tu the.GEN

Ά΅Η΍ΏΉϟ΅Ζȱȱ σΏ΋ΒΉΑȱȱȱ ȱ

etus year.GEN 



tis Anastasiu vasilias eliksen  the.GEN Anastasios.GEN reign.GEN end.ACT.PAST.PERFVE.3SG ‘(the war) ended at the end of the fifteenth year of Anastasios’s reign’ (Theophanes Confessor, Chronicle, 229Ǻ, 6-7; AD 8-9)  d. Late Medieval Greek ΔΣΏ΍Αȱ

ВΏ΍·ϱΗΘΉΙΗΉΑȱȱ

ǻψȱȱȱΛΕ΍ΗΘ΍΅Α΍ΎχȱȱΔϟΗΘ΍ΖǼȱȱȱ

palin olighostefsen (i xristianiki pistis) again reduce.ACT.PAST.PERFVE.3SG (the Christian faith) ‘(the Christian faith) again dwindled’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 350, 25; AD 17)

(b) Causative types from intransitive non-active The examples of non-active anticausative verbs which are causativised during the Modern Greek period are obviously fewer. Chatzidakis includes in his list the causativisation of the non-active verb ȟİțȠȣȡȐȗȠȝĮȚ ksekurazome ‘to repose/rest’, whereas Theophanopoulou-Kontou cites the causativisation of the verb İȟİȜȓııȠȝĮȚ ekselisome ‘to evolve’, which is not yet completely accepted by all native speakers, and the causativisation of the intransitive verb ĮʌȠıȪȡȠȝĮȚ aposirome ‘to withdraw’. In the

Diachronic Data and Analysis

157

case of causativisation of a non-active intransitive verb, change of voice occurs at the same time since the causative types can only bear active voice (anticausative non-active Æ causative active). (84) ȟİțȠȣȡȐȗȠȝĮȚ ksekurazome repose.NACT ĺ Ƞ ĬİȩȢ ȞĮ IJȠȞ ȟİțȠȣȡȐıİȚ o Theos na ton ksekurasi the God.NOM to he.ACC.WEAK repose.ACT.PERFVE.3SG ‘may God make him rest’ ĺ

İȟİȜȓııȠȝĮȚ ekselisome develop.NACT șĮ İȟİȜȓȟȠȣȞ IJȠ İțʌĮȚįİȣIJȚțȩ ıȪıIJȘȝĮ tha ekseliksun to ekpedheftiko sistima will develop.ACT.PERFVE.3PL the educational.ACC system.ACC ‘they will develop the educational system’ ĮʌȠıȪȡȠȝĮȚ ĺ aposirome withdraw.NACT

ĮʌȠıȪȡȦ aposiro withdraw.ACT

The new causative type of certain verbs comprises, in reality, the reappearance of an earlier causative use. The active causative type of the verb ĮȞĮʌIJȪııȦ anaptiso ‘to develop’ is for Modern Greek a new and very disputed type. The meaning of the Modern Greek ĮȞĮʌIJȪııȦ anaptiso ‘to develop’ is clearly different from the meaning of the Ancient Greek causative verb ĮȞĮʌIJȪııȦ anaptysso:. In Classical Greek, the verb ĮȞĮʌIJȪııȦ anaptysso: as intransitive has the meaning of the intransitive ‘to open’. The transitive ĮȞĮʌIJȪııȦ anaptysso: has the meaning of the causative ‘to open’. (85) Ancient Greek intransitive non-active a. ΔΘνΕΙ·΅Ζȱȱ πΑȱȱ ΘΓϧΖȱȱ·΍··ΏϾΐΓ΍Ζȱȱ ΔΕΓΗΌΉΘνΓΑȱȱϵΔΝΖǰȱȱϵΘ΅Αȱȱ ΐξΑȱ Έ΍΅ϟΕ΋Θ΅΍ǰȱ pterugas en tois gigglumois prostheteon hopo:s, hotan men diaire:tai, flaps at the joints detachable that when prt elevate.NACT.PRES.3SG ΦΑ΅ΔΘϾΗΗΝΑΘ΅΍ǰȱȱ ϵΘ΅Αȱȱ Έξȱȱ Ύ΅Θ΅ϟΕ΋Θ΅΍ǰȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ πΔ΍ΎΏΉϟΝΑΘ΅΍ȱ ȱȱȱȱȱȱȱȱȱȱȱȱȱ ϳΐΓϟΝΖȱȱ homoio:s anaptusso:ntai, hotan de kataire:tai, epikleio:ntai correspondingly open.NACT.PRES.3PL

when

prt lower.NACT.PRES.3SG close.NACT.PRES.3PL

‘there should be detachable flaps at the joints, in order that, when the arm is elevated, they may open correspondingly, and may close when it is lowered’ (Xenophon, On the Art of Horsemanship, 12, 6; 5-4 BC) transitive active ȱȱȱȱȱȱȱȱ ΛΕϱΑΓΖȱȱ ΗΝΘ΋Εϟ΅Ζȱȱ ·ΤΕȱȱ ΚΣΕΐ΅Ύ΅ȱ b. ΦΑ΅ΔΘϾΗΗΉ΍ȱȱ ȱȱȱȱȱ ȱ anaptyssei khronos so:te:rias gar pharmaka open.ACT.PRES.3SG time.NOM salvation.GEN as drugs.ACC ‘as the time makes the drugs of salvation unfold’ (Sophocles, Fragments, 301, 2; 5 BC)

In the data of Hellenistic and Roman Koine, only intransitive examples can be found, whereas in Medieval Greek, it is rarely used, and only in transitive structures. (86) Hellenistic-Roman Koine intransitive non-active Ύ΅ΌϱΈУȱȱ ·ϟΑΉΗΌ΅΍ǯǯǯ a. БΖȱ ΘϲΑȱȱ ΓЁΈϲΑȱ ΌϾΕ΅Αȱ ΩΏΏ΋Αȱȱ ΘϜȱȱ hos ton oudon thyran alle:n te:i kathodo:i ginesthai... as the entrance door other the.DAT descent.DAT become.INF Ύ΅Ϡȱȱ ΦΑΉΔΘϾΗǹȘȱȱ ȱ ȱ kai aneptysseto and open.NACT.IMPERF.3SG ‘as the entrance became another door to the descent... and opened’ (Heliodorus, Ethiopian Romance, 1, 29, 1, 4; AD 3) i in the previous period - Medieval Greek: transitive active Early Medieval Greek Χΐ΅ȱȱ ΘХȱȱ ΦΑ΅ΔΘϾΒ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘϲȱȱΎΉΚΣΏ΅΍ΓΑȱΘΓІΘΓ b. Ύ΅Ϡȱȱ ΉЈΕΓΑȱȱ ke evron ama to anaptikse to kefaleon tuto and

find.IMP.2SG immediately the.DAT open.ACT.PAST.PERFVE.INF the book.ACC this

‘and find it the moment that you open that book...’ (John Moschos, Leimonarion (The Spiritual Meadow), LV, B, 8; AD 6-7)

158

Chapter Three

The ongoing causativisation of the verb constitutes an indication of the reintroduction of the external argument for a verb which in earlier language systems could have had an external argument (cf. Lemmens 1999). The second particular instance which I would like to draw attention to is the verb İȟİȜȓııȦ ekseliso ‘to evolve’. The causative type İȟİȜȓııȦ ekseliso ‘to evolve’ is not accepted by all speakers of Modern Greek. It is concerned with a newer and disputed active causative type. Again, semantic change in relation to the Ancient Greek ĮȞĮʌIJȪııȦ anaptysso: / anaptiso, which meant ‘unfold’, is observed. (87) Ancient Greek intransitive non-active a. πΒΉΏϟΘΘΉΘ΅΍ȱȱȱȱȱȱȱȱȱσΎ΅ΗΘΓΖȱȱϳȱȱȱΗΘϟΛΓΖǰȱȱϥΑ΅ȱȱΓϡȱȱȱΎΕΣΘ΍ΗΘΓ΍ȱπΑ΅ΑΘϟΓ΍ȱΦΉϠȱΘΓϧΖȱΔΓΏΉΐϟΓ΍ΖȱȱИΗ΍Αȱ   ekselittetai hekastos ho stikhos, hina hoi kratistoi enantioi aei tois polemiois o:sin  evolve.NACT.PRES.3SG each the rank.NOM to the best opposite.NOM always the enemies.DAT be.3PL

‘each rank countermarches, in order that the best men may always be face to face with the enemy’ (Xenophon, Constitution of the Lacedaimonians, 11, 8; 5-4 BC) transitive active ΘχΑȱ ΚΣΏ΅··΅ȱ b. πΒΉΏϟΘΘΓΙΗ΍ȱ   ekselittousi te:n phalagga evolve.ACT.PRES.3PL the army.ACC ‘they make the army countermarch by ranks’ (Xenophon, Constitution of the Lacedaimonians, 11, 9, 6; 5-4 BC)

The verb is intransitive for Modern Greek, but a tendency to its causativisation (with the meaning ‘to ameliorate’, ‘to modernise’) is observed (Theophanopoulou 2003a): (88) innovative causative use a. ?İȟİȜȓııȠȣȞ IJȠ İțȜȠȖȚțȩ ıȪıIJȘȝĮ ?ekselisun to eklojiko sistima evolve.ACT.PRES.3PL the electoral.ACC system.ACC ‘they modernise the electoral system’ b. ?İʌȚșȣȝȠȪȞ ȞĮ İȟİȜȓȟȠȣȞ IJȘ įȚȠȓțȘıȘ IJȘȢ ?epithimun na ekseliksun ti dhiikisi tis wish.3PL to evolve.ACT.PERFVE.3PL the administration.ACC the ‘they wish to modernise the administration of Greece’

ǼȜȜȐįĮȢ Eladhas Greece.GEN

The opposite direction of change, namely an innovative anticausative use (with the meaning of changeof-state/anticausative meaning) of an earlier transitive verb does not comprise a productive phenomenon for the ongoing changes in Modern Greek. In contrast, different types of semantic change in transitive verbs that begin to be used also in intransitive constructions are observed. The instances of intransitive types which derive from transitive verbs are not instances of anticausativised verbs (with the meaning of change-of-state) but instances of detransitivised verbs (the non-active morphology in the detransitivised verbs marks the absorption of the accusative case). For example, in the case of the verbs of the group of ıȣıIJȑȜȜȠȝĮȚ sistelome ‘to contract/become smaller’, the ascertainment of change in meaning is particularly important in relation to the connection or non-connection between the earlier transitive structure and the innovative intransitive structure (Zombolou 2004)61. The transitive verb ıȣıIJȑȜȜȦ systello: of Classical Greek had a very different meaning (‘to take down the sails of a ship’) from the intransitive verb ıȣıIJȑȜȜȠȝĮȚ sistelome ‘to contract/become smaller’ of Modern Greek. Consequently, the intransitive verb of Modern Greek cannot be connected to the Ancient Greek transitive ıȣıIJȑȜȜȦ systello: and considered an instance of secondary intransitive (anticausative) use of a transitive verb. 3.3.4.2 Changes in accusative verbs: sub-categories The accusative (transitive non-causative) verbs of Modern Greek show great (in relation to causative verbs) flexibility and alternations in expression of the object, as do the accusative verbs of all periods of the Greek language. An example of this flexibility is constituted by the (non-causative) accusative motion verbs, which are verbs of activity, and their direct object does not comprise a part of the event 61

Furthermore, the archaic Modern Greek verb is replaced by another verb which expresses a similar meaning, the verb ȝȚțȡĮȓȞȦ mikreno ‘to minimize’ with active morphology. The same happens with the Modern Greek verbs įȚĮıIJȑȜȜȦ dhiastelo ‘to expand’ (in Medieval Greek its meaning was ‘to open’), ıİȓȦ sio ‘to rock’, İțIJȡȑʌȦ ektrepo ‘to divert’ which are replaced by the active verbs ȝİȖĮȜȫȞȦ meghalono ‘to expand/ grow’, IJĮȡĮțȠȣȞȫ tarakuno ‘to rock/ shake’, ʌĮȡĮıȪȡȦ parasiro ‘to divert/ drift’ (Zombolou 2004).

Diachronic Data and Analysis

159

structure but its presence is caused only by the idiosyncratic meaning of each verb. It is owing to this idiosyncratic connection of the second argument to the syntactic behaviour of the verb and to its relation with only one part of verb meaning (that of the meaning of the particular verb and not of its event structure and of the meaning of the semantic group to which the verb belongs) that the instability concerning the presence of the second argument and its means of expression is to be expected. Basea-Bezantakou (1992) distinguishes between three groups of motion verbs. Alternation between morphological cases and PPs is observed with all three verb groups, something which is not possible with causative verbs (they constantly take only DPs in the accusative case as direct objects). Causative verbs (already causative in preceding periods or causativised during the Modern Greek period) consistently have the patient argument (the argument which undergoes the change-of-state) in the accusative case. (89) a. ȅ īȚȐȞȞȘȢ ȑıʌĮıİ IJȠ ʌĮȡȐșȣȡȠ o Janis espase to parathiro the Janis.NOM break.ACT.PAST.PERFVE.3SG the window.ACC ‘Janis broke the window’ b. *ȅ īȚȐȞȞȘȢ ȑıʌĮıİ IJȠȣ ʌĮȡȐșȣȡȠȣ *o Janis espase tu parathiru the Janis.NOM break.ACT.PAST.PERFVE.3SG the window.GEN ‘Janis broke the window / Janis broke a part of the window’ c. *ȅ īȚȐȞȞȘȢ ȑıʌĮıİ ıIJȠ ʌĮȡȐșȣȡȠ *o Janis espase sto parathiro the Janis.NOM break.ACT.PAST.PERFVE.3SG to-the window ‘Janis broke the window / Janis broke a part of the window’

The motion verbs of Basea-Bezantakou’s first group, verbs denoting the direction of the movement towards a particular point (End-point) can have an object expressed as DP in the accusative case but also as PP. This syntactic behaviour holds good independently of whether the second argument is inanimate or animate. It is clearly concerned with verbs with a low degree of transitivity (see chapter 2: contrast between prototypical transitive verbs and accusative verbs). For example, the verb ʌȜȘıȚȐȗȦ plisiazo ‘to approach’ has a very low degree of transitivity since the DP is not affected by the action of the verb, as it happens with verbs that denote the result of an action (causative verbs). (90) a. ʌȜȘıȚȐȗȦ IJȠ ʌĮȡȐșȣȡȠ plisiazo to parathiro approach.ACT.PRES.1SG the.ACC window.ACC ‘I approach the window’ b. ʌȜȘıȚȐȗȦ ıIJȠ ʌĮȡȐșȣȡȠ plisiazo sto parathiro approach.ACT.PRES.1SG to-the window ‘I approach close to the window’ c. IJȠȞ ʌȜȘıȚȐȗȦ / ʌȜȘıȚȐȗȦ IJȠȞ ton plisiazo / plisiazo ton he.ACC.WEAK approach.ACT.PRES.1SG / approach.ACT.PRES.1SG the ‘I approach him / I approach Iason’ d. ʌȜȘıȚȐȗȦ ıIJȠȞ ǿȐıȠȞĮ plisiazo ston Iasona approach.ACT.PRES.1SG to-the Iason ‘I approach close to Iason’

ǿȐıȠȞĮ Iasona Iason.ACC

Basea-Bezantakou analyses the accusative case as syntactic expression with a PP in which the preposition and the article have been omitted62: 62

Additionally, the End-point can be replaced with the purpose of the action. The purpose of the movement in a broad sense is the destination of the movement, so, therefore, is semantically the End-point of the movement. (1) a.

b.

c.

ʌȐȦ ıʌȓIJȚ pao spiti go.ACT.PRES.1SG home.ACC ‘I go home’ ʌȐȦ ıIJȠ ıʌȓIJȚ ȞĮ ijȐȦ pao sto spiti na fao go.ACT.PRES.1SG to-the home to eat.1SG ‘I go home to eat something’ ʌȐȦ ȖȚĮ ȞĮ ijȐȦ

160

Chapter Three

(91) a. ȑȡȤȠȝĮȚ ıʌȓIJȚ erxome spiti come.NACT.PRES.1SG home.ACC ‘I come home’ b. ȑȡȤȠȝĮȚ ıIJȠ ıʌȓIJȚ erxome sto spiti come.NACT.PRES.1SG to-the home ‘I come home’

In the second group of motion verbs are included verbs with a complement which refers to the distance travelled. The distance travelled can be expressed with PP or with DP in the accusative case. (92)

(93)

ʌİIJȐ Įʌȩ IJȠ ȑȞĮ įȑȞIJȡȠ ıIJȠ ȐȜȜȠ peta apo to ena dhendro sto alo fly.ACT.PRES.3SG from the one tree to-the other ‘It flies from one tree to the other’ įȚĮȕĮȓȞȦ IJȘ ȜİȦijȩȡȠ dhiaveno ti leoforo cross.ACT.PRES.1SG the.ACC avenue.ACC ‘I cross the avenue’

The meaning of the motion verbs of the third group does not clearly define a specific local denotation. Again, alternation of the means of denotation of the second argument is possible: (94) a. țĮIJİȕĮȓȞȦ kateveno go-down.ACT b. țĮIJİȕĮȓȞȦ IJȘ ıțȐȜĮ kateveno ti skala go-down.ACT.PRES.1SG the stairs.ACC ‘I go down the stairs’ c. țĮIJİȕĮȓȞȦ ıIJȘ ıțȐȜĮ kateveno sti skala go-down.ACT.PRES.1SG in-the stairs ‘I go down to the stairs’ d. țĮIJİȕĮȓȞȦ Įʌȩ IJȘ ıțȐȜĮ kateveno apo ti skala go-down.ACT.PRES.1SG from the stairs ‘I go down using the stairs’ e. țĮIJİȕĮȓȞȦ ȝİ IJȘ ıțȐȜĮ / ȝİ IJȠ ĮıĮȞıȑȡ kateveno me ti skala / me to asanser go-down.ACT.PRES.1SG with the stairs / with the elevator ‘I go down using the stairs / the elevator’ f. țĮIJİȕĮȓȞȦ ıIJȘȞ ʌȩȡIJĮ kateveno stin porta go-down.ACT.PRES.1SG to-the door ‘I go down to the door’ g. țĮIJİȕĮȓȞȦ ıIJȘȞ ǹșȒȞĮ kateveno stin Athina go-down.ACT.PRES.1SG to-the Athens ‘I go down to Athens’ h. țĮIJİȕĮȓȞȦ ȞĮ ıİ įȦ kateveno na se dho go-down.ACT.PRES.1SG to you.ACC.WEAK see.1SG ‘I go down to see you’ i. țĮIJİȕĮȓȞȦ țȐIJȦ ȞĮ ıȠȣ įȫıȦ

d.

pao ja go.ACT.PRES.1SG for ‘I go to eat something’ ʌȐȦ ȖȚĮ pao ja go.ACT.PRES.1SG for ‘I go for lunch’

na to

fao eat.1SG

ijĮȖȘIJȩ fajito food

IJȠ įȑȝĮ

Diachronic Data and Analysis

161

kateveno kato na su dhoso to dhema go-down.ACT.PRES.1SG down tȠ you.ACC.WEAK give.1SG the parcel.ACC ‘I go down to give you the parcel’

3.3.5 Summary In the analysis of the Medieval and Early Modern Greek periods I separated (as I did also for the remaining periods) the changes in argument structure (causative, anticausative types) from the changes in the voice system (in contrast to the descriptions of historical grammars that analyse structures and voice as being completely interdependent, without clearly distinguishing -in the majority of caseswhether a change in the morphological encoding of voice is involved or an innovative type which did not exist in the preceding period). This means of the distribution of changes reflects our working hypothesis, that the changes in voice do not encode the changes and the direction of derivation in (anti)causativity. As with the Ancient Greek and Hellenistic-Roman period, I differentiate between the changes in the structure of the causatives (core transitive verbs) and the changes in the accusative verbs (non-causative transitive verbs). The changes in the accusative verbs are mainly changes in case, whereas the changes in causative verbs are changes in argument structure (new causative types with presence of an agent/cause in the subject position based on an earlier anticausative verb). Concerning the changes in argument structure, the leading role is played by the presence of the clitic as direct object of an earlier intransitive verb. The clitic comprises the first phase of the causativisation of intransitive verbs. It is a necessary phase of the causativisation (so that the presence of a DP in the accusative case as object of a former intransitive verb would be possible from a period and then), but the completion of causativisation in all instances is not obligatory; it is possible for a causativisation (which has not been completed and with which only the presence of a clitic [not DP] was possible as direct object) to be abandoned in a later period. The instability in the system of nominal morphology, as it appears through the use of nominative instead of accusative case for the objects of finite verbs in this period (evidently, this instability is connected to the loss of infinitives which accept subjects in the accusative case), clearly helped and also strengthened the reanalysis of anticausative verbs. In the majority of cases, causativisation is based on an active intransitive verb, but in a few cases it begins on the basis of a non-active intransitive verb. Clearly, this specific type of change is related to the means of reanalysis of the structure: an intransitive structure with an active verb is reanalysed more easily into a transitive structure (the analysis of the DP-subject as DP-object is enough) than a structure with a non-active verb; in the latter case, the intransitive (anticausative) verb must initially be analysed by the native speakers as a passive structure for the presence of an agent to be considered possible (intransitive/anticausative Æ interpretation as passive Æ hypothesis of presence of transitive type). The verbs that do not retain causativisation, and which, in other words, do not constantly have an external argument, are verbs denoting change-of-state from internal change (e.g., ȣȖȚĮȓȞȦ ijieno ‘to be/become healthy’, ĮıșİȞȫ astheno ‘to become ill’, ʌȜȠȣIJȓȗȦ plutizo ‘to grow rich’). Consequently, even if causativisation is a process which applies to all periods, it is not a process without limitations and restrictions. The non-availability of the causativisation of intransitive verbs is also linked to the availability of presence of a DP in the accusative case denoting the cause with these intransitive verbs (parallel use of DP in the accusative case or PP for the denotation of the cause). Therefore, we can conclude that intransitive verbs with a cause expressed as DP in the accusative case and accusative verbs with a theme in the direct object position are not causativised systematically (in contrast to languages with causative morphemes, in which the causativisation of accusative verbs is frequent). The changes in the accusative verbs are mainly changes in the case and not in the transitivity system. In contrast to causative verbs, which consistently assign accusative case, the accusative verbs can take DPs in the accusative, genitive/dative case or PPs for the accusative case finally to prevail.ȱ Regarding voice, in this period instability in the voice of both transitive and intransitive verbs is observed. This instability is connected to the completion of an important change in the voice system. In Medieval Greek, the change for the absorption of the accusative case is completed: non-active voice is no longer connected to the absorption of other arguments (benefactive, goal, theme) but only of the accusative case; non-active voice prevents accusative case from being assigned (blocking/absorption: *non-active verb type + DP-accusative as direct object). I consider that this change was instigated by the loss of the benefactive meaning (which was denoted by the non-active transitives) and by the extension of the accusative case as the case of direct objects. In its turn, this change (diachronic chain of changes) led to the strengthening of active anticausatives (active from non-active anticausatives). The absence of connection between the anticausative voice and the processes of (anti)causativisation is also evident from the presence of alternative active and non-active anticausative types of the same verb

162

Chapter Three

without a difference in meaning during Medieval Greek. I interpreted this parallel presence of the two types (active and non-active anticausatives) as an ongoing change (the use of both active and nonactive endings is in evidence until the active endings dominate). I followed the same approach for the changes identified as being ongoing changes in Modern Greek and examined them in relation to the preceding period: they are distinguished as changes in the verb argument structure (causativisations) and changes in voice morphology (continuation of the spread of active anticausatives). Similarly, in Modern Greek as well, the changes in transitivity alternations (changes in argument structure) are differentiated from the changes in the accusative verbs (mainly changes in case morphology). What remains is (a) to examine if the specific means of distribution and analysis of the data only concern the diachrony of Greek or can also be applied to data from other languages (section 3.4), and (b) to end up with a complete interpretation of the diachronic processes and mechanisms based on the partial conclusions (chapter 4).

3.4 Parallel Developments in the Diachrony of English and Romance In this section, I will attempt, using the same criteria of distribution and analysis [(i) changes in the argument structure/causativisations; (ii) voice instability], to examine the development of the same categories in the diachronic data of other languages. The aim is also to test classification and analysis of the changes that I undertook for Greek using cross-linguistic data. This endeavour comprises a first attempt at analysis of several cross-linguistic data, according to the criteria that I followed in previous sections for the Greek language. This distribution and analysis of the data have not been undertaken for cross-linguistic data in past literature (that is precisely the merit of this present attempt). Diachronic examples (taken from original texts, earlier articles63, histories and historical grammars) and the course of the changes can be incorporated in a dual classification: (a) causative verbs (and changes in argument structure) are distinguished from the accusative verbs (and changes in case), (b) changes in voice (instability of intransitive voice) are distinguished from the direction of the changes in argument structure (causativisations). This present section gives us, in parallel, essential elements (mainly about periphrastic passive types and pronouns) which enable a more complete interpretation of the changes (chapter 4). In contrast to the numerous, detailed synchronic analyses of argument structure and transitivity (which also represent different perspectives), diachronic analyses are not as systematic. In most cases, diachronic observations in the literature concerning these issues are incorporated into traditional grammatical descriptions, or only synchronic variations and tendencies that lead to changes or are a result of changes are referred to. The majority of data comes from studies which do not belong to the Generative Grammar perspective, and since I endeavour a data analysis according to the approach I presented in the previous sections, I cite some elements from the interpretations that were undertaken in the literature only additionally and only in relation to the main questions I raised. I assume that the conclusions which arise from cross-linguistic remarks contribute to a more complete interpretation of the changes in transitivity. In this present section, I outline the analysis of the different sub-types of the phenomena, and mainly how these different sub-types are connected; the overall interpretation is presented in the following section (chapter 4).

3.4.1 Innovative causative use of intransitive verbs: direction of the changes The data of this section mainly come from the English language. According to the criteria I posed for the Greek language, the changes in verb syntax are separated into changes in argument structure (new causative type from an earlier anticausative type) and changes in case (extension of the accusative/structural case). The course of changes in argument structure is in one direction: causativisation (with introduction of an agent argument) of earlier anticausative non-alternating types. I assume that these criteria for diachronic changes in transitivity apply cross-linguistically. (a) Causativisations In comparison with the few studies that exist for other languages, we have a very clear picture of the history of verb transitivity in English. According to Visser’s (1963-73) historical grammar, already in Old English, there are a small group of verbs that participate in the transitivity alternation and that can, in other words, be used in both transitive and intransitive structures, and with the same type, without 63

Only parenthetically and where it is necessary for a wider picture of the changes do ǿ comment on the interpretations that have been given in the literature.

Diachronic Data and Analysis

163

change in voice morphology (Visser (1963-73: para. 1320) calls them amphibious verbs). (1) ‘break’ intransitive a. Hit is on us eallum swutol & gesene þaet we aer þysan oftor it is to us all clear and seen that we.NOM before that oftener braecan þonne we bettan broke.3PL than we.NOM amended.3PL ‘it is clear and well seen in all of us that we have previously more often transgressed than we have amended’ (Wulfstan, Sermo ad Anglos, 42, 3; AD 11) transitive b. He braec þone bordweall he.NOM broke.3SG that.ACC shield.ACC ‘he broke that shield’ (Battle of Malden, 277; AD 10)64

The verbs in (2) have the same behaviour with brecan ‘to break’ and are included in the examples given by Visser. Lumsden (1987) analyses the possibility of participation in both structures as a result of the lexical process of the introduction of a cause in the lexical representation of the intransitive verb. (2)

acweccan ‘to shake’ byrnan ‘to burn’ fleon ‘to flee’ openian ‘to open’ wlitigan ‘to get more beautiful’

Visser (1963-73: para. 132, 134) talks about a “wholesale process of transitivation” as a result of changes; nearly all of the verbs that expressed movement and sound production and were intransitive in Old English are causative in Modern English: –climb, bound, burst, creep, dive, flow, glide, run, spring –bellow, crow, groan, grunt, laugh, lisp, neigh, stammer, weep, whisper, whistle

The processes of causativisation apply to all periods of English and not just one specific period: the verb cook (which was derived from the noun cook) is used from the 17th century as transitive and intransitive, but originally it was only intransitive. Rissanen (1999: 256) refers to several instances of initially intransitive motion verbs which are able to take an object in the imperative mood in the 16th and 17th century. Earlier intransitive verbs are also used in a causative structure in Modern English (with the meaning of change-of-state) (Rissanen 1999: 253-257). (3)

(4)

(5)

Meet me tomorrow. Ile flie my Hawke with yours ‘Meet me tomorrow. I will fly my hawk with yours’ ([HC] Heywood Woman killed with kindness, 93; AD 17) After swim him and apply bathes ‘ǹfter you make him swim and give him a bath’ (Thomas de Grey, The Compleat Horseman – AD 17 [OED s.v. swim v. 13]) They likewise grow some rice and tobacco (John Campbell, A Political Survey of Great Britain – AD 18 [OED s.v. grow v. 14])

In contrast to Generative Grammar which dealt with the diachrony of transitivity mainly within the framework of theoretical synchronic studies, greater emphasis was given to the diachronic aspect of transitivity by several studies under the perspective of Cognitive (and Functional-Systemic) Linguistics (mainly by Davidse (1991) and Lemmens (1998)). In these studies, abundant data are cited: I assume that these data can also be interpreted as data that strengthen the argument concerning the monodirectionality of changes in argument structure (causatives from earlier non-alternating anticausatives) and the non-connection of voice with the direction of the derivation of new types. On the other hand, I do not follow the specific analyses of the changes according to this approach, which in the majority of cases stress the relationship between very subtle semantic changes and changes in the real world with the grammatical changes under examination (for my own overall analysis, cf. sections 2.1, 2.2 and 4). Davidse (1991) has observed the very frequent semantic change of intransitive verbs denoting autonomous process in verbs denoting semi-autonomous process and which can be used also in 64

For more examples, cf. Lavidas 2006a.

164

Chapter Three

transitive structures with the patient argument as the object65; this is, of course, a result of the introduction of an external cause66. (6)

a. The motor is running b. She is running the motor

Geyskens (1997) and Davidse & Geyskens (1998) present data on the causativisation of verbs of manner of movement in English with the introduction of an animate agent (run, jump, march), but also verbs that do not denote movement (such as bleed, work), based on a large corpus of data. (7)

a. What a nice gentleman you are, to walk Mary home b. The captain marched them off to the main camp (from Davidse and Geyskens 1998: 169)

Davidse & Geyskens characterise the above examples as ergative causativisations of intransitives and they separate them, in contrast to Levin & Rappaport Hovav (1995: 115) and Pinker67, from the transitive causativisations (ex. 8): (8)

a. He beeped the horn b. They may fly the women and children out on Thursday

Of the 7158 examples with verbs of manner of movement in Geyskens’s (1997)68 data, only in 115 do we have ergative causativisation of the intransitive, and there is no example of regular causativisation of the intransitives leap, swim and drive. It was ascertained, consequently, that ergative causativisation of intransitives is very rare in relation to the typical intransitive structure of these verbs. The percentage of transitive use of these verbs is as follows: fly 0.2%, jump 0.3%, run 0.5%, race 1.9%, gallop 2.5 %, canter 2.6%, walk 2.6%, trot 4.2%, march 6.6% Ÿ it is obvious that the ergative causativisation of intransitives is a marked structure69.

Lemmens (1998: 86-97) analyses the processes of verb changes as processes of extension of verb meaning. For example, the verb starve, which initially only meant ‘to die’ (absence of external cause), approached the meaning of the verbs hunger and thirst, gradually acquiring the meaning ‘to die from hunger/thirst’ (namely, presence of specific cause). The lexical changes are accompanied by changes in argument structure: starve begins to be used in structures with two participants. Initially, the (ergative) structures with two participants appear to be an extension of the intransitive prototypical pattern (such as today’s: I walk the dog ‘I take the dog for a walk’, I jump the horse ‘I make the horse that I ride jump’), but in the end they result in the change of the verb into a regular (ergative) transitive verb (with emphasis on the patient who participates in the process, and with the availability of alternation with the transitive structure with the patient in the subject position). The same did not happen with the verbs hunger70 and thirst, which were turned into verbs denoting intense desire and remain consistently intransitive. Lemmens does not follow Davidse (1991: 115), who argues that the transitive use 65

66

67

68

69

70

Lemmens calls these structures ergative structures. I do not follow this specific terminology. For details relating to Lemmens’s analysis, cf. notes 10 and 27 in chapter 1. Halliday (1985b: 146) ascertained the spread of that structure (‘ergative schema’) in English over the last five hundred years. Pinker (1989: 226) describes the two verb groups of unergative intransitives which can be causativised in Modern English, as follows: (i) verbs that denote the manner of volitional movement (walk, gallop, trot, race, run, jump, jog), (ii) motion verbs denoting the means of transport (drive, fly, cycle, ferry, boat, sail). Geyskens (1997), in a study based on a corpus, ascertains the presence of six types of ergative causativisations of intransitives in relation to three factors: (a) the co-extension of instigation and induced action, (b) the natural causativisation (natural contact), (c) the character of power asymmentry. Clearly, Geyskens’s subcategorisation is highly detailed and may include characteristics that are are not encoded linguistically (e.g., the natural contact). In relation to the obligatory PP denoting direction with the above verbs, it is argued that only for the first group is the presence of the PP of direction obligatory: I ran her to the doctor), whereas for the other groups the PP of direction is not required. The verb hunger in a certain period was used in transitive (ergative) structures, but later a tendency prevailed for a return to its initial syntactic behaviour (denotation of the intransitive process of an intense desire for something), possibly because the meanings of starve and hunger were very close resulting in semantic change for only one of the two verbs.

Diachronic Data and Analysis

165

(ergativisation) of intransitive verbs is limited in cases of intransitives denoting ‘control’ of verb activity by the subject. As counter-examples to the position of Davidse, Lemmens refers to the following sentences: (9)

a. The nurse burped the baby (from Smith 1978) b. The doctor bled the patient (from Levin & Rappaport Hovav 1995)

Changes have also been observed in relation to transitive verbs in English. Transitive verbs that gave emphasis to the subject and to the complete affectedness of the object give (after the semantic change) emphasis to the object, which is affected and at the same time participates in the process. This semantic change also results in syntactic change since the object can be transformed into subject in the intransitive structure of the same verb (an instance also of ‘ergativisation’). This happened, for example, with verbs of the group suffocate (Lemmens 1998: 93-97). From the 16th century onwards, even the initially exclusively transitive verbs (strangle, throttle, stifle, smother) began to be used in intransitive structures [(i) subject + verb + patient and (ii) patient + verb]. Prior to, or at the same time, with the consistent appearance of the aforementioned verbs in intransitive structures also, their meaning was modified: the specification of the cause is not so exact since other types of causes are added which can result in suffocation, and a degree of overlap of the meaning between the different verbs with similar meaning is observed, that is, the area of use of each verb is not absolute. (10) He could not lie down… he should strangle (the subject which participates in the process is also the patient)

The basic area of disagreement between Lemmens and Davidse is the inclusion of lexical meaning and the meaning of the structure in the examination of diachronic changes. For Lemmens, it is impossible to separate the lexical from the structural meaning and to say whether the change in verb meaning came before the participation of the specific verb in various structures that carry different meaning. The semantic changes are the result of a strong interaction between the lexical and the structural meaning, and for this reason, it is impossible to ascertain in all the cases and absolutely which change caused the other; in other words, if the extension of meaning caused the participation of the verb in other structures, or whether the participation in other structures resulted in the differentiation of meaning. In contrast, however, to the verb starve (for which the new schema [new alternating verb from an intransitive verb] was established), concerning the verbs of the group ‘suffocate’ (intransitive from transitive) a new change71 is observed, and tendency for a return to their initial state (the same can also be ascertained for hunger) and for participation only in transitive structures (and not in intransitive structures). The causes for the changes in verb meaning and also for their participation in intransitive structures (ergative patterns) are being sought, according to this approach, again in experience, in the real world: for example, the verb murder was never used as intransitive (in an ergative pattern), because the subject of this verb could never be considered (in sentences such as John murdered), based on experience from the real world, both agent and affected entity. The problem with theories which interpret language change strictly in relation to the real world is that they have difficulties in analysing language acquisition (how a child learns in such a short space of time all these linguistic structures if they are so connected [according to these approaches] with the real world). Concerning the reasons which I assume contributed to the emergence of causative use based on earlier intransitive use, the presence of the structure be + past participle should be noted (Mustanoja 1960: 429, Visser 1963-73: para. 134ff.). In earlier periods of the English language, the auxiliary verb ‘be’ (and not the auxiliary ‘have’) was used with past participles of intransitive verbs for the denotation of the state-result (it was crumpled : it has crumpled). The coincidence of this construction with the passive construction results in the agent also being implied in the first structures (denoting state-result

71

Further to these obvious directions (ergative from alternating and transitive from ergative), Lemmens also points out the existence of more complex cases of change; for example, the verb abort was an intransitive verb which initially acquired transitive use also (it was ergativised) (‘the embryo is affected and participates in the verb action’) and further to that it lost its intransitive use and became absolutely transitive (‘the women bear the role of the ǹgent’), but only in its literal uses (since its metaphorical uses continue to follow the earlier ergative pattern (both transitive and intransitive use is possible)). Its use as exclusively transitive (‘transitivisation’ according to Lemmens) is still to be found in progress and is gradually extended to the metaphorical uses as well. The course of all these changes is attributed by Lemmens to the real world, even to ethical values which change: for example, its second change (which now is exclusively transitive) in Modern English is connected to the recognition or non-recognition of the right of a woman to decide on the birth of the child she is carrying.

166

Chapter Three

with intransitive verbs)72. This development (of passive interpretation) leads to the interpretation of sentences such as the one in example (11) (the potatoes are grown), either as a present perfect structure of an intransitive-anticausative verb (the potatoes have grown), or as a passive structure which arises from a corresponding transitive structure (the presence of the agent is possible)73. (11) the potatoes are grown (cf. also the structure “hee was walkt” in (12))

In other cases, the passive structures for anticausative verbs comprise evidence for the availability of a corresponding transitive structure. The availability of the presence of passive (+agent) in parallel with the anticausative intransitive type (-agent) may lead to the reinterpretation of the intransitive verb as a verb which can participate in a transitive structure since the majority of transitive verbs also form passive structures. In this case, consequently, the connection transitive Æ passive (when there is a transitive type of the verb, then a passive type also exists) is considered by speakers to hold good in reverse also, namely passive Æ transitive (when a passive type of the verb exists, then a transitive type also exists). (12) Hee (=the old man) was walkt into the parke He was walked into the park ‘he was walked into the park / they walked him (the old man) into the park’ ([Helsinki Corpus] Robert Armin, A Nest of Ninnies, 43; AD 17)

The remaining directions of changes which are in evidence (anticausative from alternating verb) can be interpreted as abandonment of causativisation in the later period, but in any instance a pure anticausativization (anticausative from causative verb) is not observed. In Old English, psych-verbs also participated in both causative and anticausative structures, in contrast to Modern English (Fischer & van der Leek 1987). For example, the verb lystan with the experiencer in the dative case and the cause in the nominative case meant ‘to cause desire’, while with the experiencer in the nominative case 72

Of particular interest is the fact that in Old English the passive type (auxiliary ‘be’ + uninflected passive participle) was initially formed with inflected passive participle: (1) On þære ilcan tide wurdon twegen æþelingas afliemde of Sciþþian on that same hour were two nobles exiled from Scythia ‘at the same time the two nobles were exiled from Scythia’ (King Alfred, Orosius’ History, 1 10.29.14; AD 9)

According to the meaning of the passive structure, two auxiliaries were used in Old English for the formation of the periphrastic passive type: weorþan for verbs of activity and change-of-state, beon/wesan for states/results. (2) Ĉær wearÿ Alexander Þurhscoten mid anre flan there was Alexander pierced with one arrow ‘then Alexander was pierced by an arrow’ (King Alfred, Orosius’ History, 3, 9.134.22; AD 9) (3) wæs sum broðor syndriglice mid godcundre gife gemæred & geweorðad was a brother especially with divine gift celebrated and honoured ‘...one brother was especially proclaimed and honoured for having a divine gift’ (Bede 4 25.342.3; AD 9-10)

73

The semantic distinction between the two auxiliaries gradually became less clear, and the auxiliary (weorþan) was lost, to be finally replaced in Early Modern English (Moessner 1994) by the verb get. Lumsden connects the changes from intransitive to causative with two other parallel changes, the change in the verbs which took indirect objects as complements in earlier periods and which now take direct objects, and the change in adjuncts, which in Old English were never expressed with a DP in the accusative case (but always with a PP), but in Modern English are also expressed with structural/accusative case. Lumsden argues that in Old English the assignment of the non-inherent (accusative) case should have been marked morphologically on the verb, and, in contrast, in Modern English the assignment of the non-inherent (structural now, and without any case marking of the noun) case does not need to be marked morphologically on the verb. The prefixes (ge-, be-) and the change of the vowel (vowel mutation/umlaut) comprised the necessary verb characteristics for the marking of the non-inherent case (accusative) assignment, and the intransitive verbs remained unmarked. The prefixes and the change of the vowel became superfluous features when (from the 12th century) the situation with the default characterisation of the object changed (according to Lumsden’s terms: default direct object in Old English = [+inherent]; default direct object in Modern English = [-inherent], or, in other words, in Modern English the assignment of case is structural). In the grammar of Modern English special marking on the verb is not needed for it to assign the non-marked (structural) accusative case.

Diachronic Data and Analysis

167

it meant ‘to desire’: NOM-cause NOM-experiencer

V V

DAT-experiencer

causative structure anticausative structure

In the causative and anticausative structure, the verb lician could also be used in the same way; conversely, in Modern English the corresponding verb is only anticausative with the theme argument in the accusative case (like); the corresponding causative verb with the experiencer in the accusative case is the verb please. (13) a. an is selost þæt þu Gode licie one is best that you.NOM God.DAT please.2SG ‘the one thing that is the best to do is to please God’ ((Bl. Hom.) The Blickling Homilies 6; AD 10) b. Ƞn ÿec ic wel licade in whom I.NOM well was-pleased.1SG ‘I was very pleased with that’74 (Lindisfarne Gospels, Mark 1.11; AD 10)

Cases of loss of causative use we have encountered also in Greek, the only difference being the change of the anticausative impersonal (the experiencer in the dative case) to anticausative personal verb (the experiencer in the nominative case) in English; this change is concerned with the case marking of the sole argument of the anticausative verb and is not witnessed in the history of the Greek language (Lavidas & Papangeli 2005). To summarise, the changes, also in English, are as follows: (a) alternating from anticausative non-alternating (namely, causativisation), (b) anticausative from alternating (namely, abandonment of causativisation). The following changes are not productive: (c) alternating from causative, (d) causative from alternating. (b) Changes in accusative verbs: sub-categories Moreover, other verbs appear to take a direct object in Modern English in contrast to Old English, apart from the cases of causativisations. This development is concerned with changes of intransitive into accusative verbs and is the result of: (a) mainly, the loss of the preposition with verbs denoting movement; cf. in Modern English the verbs enter, cross; (b) replacement of the genitive by the accusative case for the cause (with intransitive verbs). The genitive is replaced by the accusative case mainly during the Middle English period. In Early Modern English, verbs such as fear or like are connected with a DP in the genitive or PP denoting the cause (Visser 1963-73: para. 370-393), and gradually take DPs in the accusative case as well. (14) a. Bona weorces gefeah destroyer.NOM work.GEN rejoiced ‘the destroyer rejoiced at the work’ (Descent into Hell (Ex. Bk., Krapp) 88; AD 10) b. Alas, why, fearing of times tiranie (Shakespeare, Sonnet, 52; AD 16-17)

Of interest is the change in English of transitives into intransitive reflexives: the direct object of the verb, the affected object, when it is co-referred with the subject, is incorporated in the verb, resulting in the same verb being ambiguous and able to participate in both a transitive and a reflexive structure. wash ‘I wash something/someone’ – wash ‘I wash myself’ The emergence of intransitive reflexive types is not limited to one only period of English: Old English Middle English Modern English

74

baðian washe(n) clothe(n) shave

‘wash (i) accusative verb (ii) reflexive verb’ ‘wash (i) accusative verb (ii) reflexive verb’, ‘dress (i) accusative verb (ii) reflexive verb’ ‘shave (i) accusative verb (ii) reflexive verb’ (Oxford English Dictionary (OED) 2009)

Cf. cases of causative-anticausative psych-verbs that lose their causative use in a later period (similarly to the diachrony of the Greek language).

168

Chapter Three

The observations of Kulikov’s study are in the same direction (1999a, b, 2003) for Vedic. The verb púúyati was both intransitive (‘to prosper’) and transitive (‘to make prosper’). The intransitive was the first use, and the transitive use arose from reanalysis of the intransitive structure. This reanalysis was based on: (a) structures with cognate NP in the accusative (“content accusative”), namely with the abstract noun which derives from the root puú with the meaning ‘prosperity’, (b) structures with accusative case which refers to one aspect, parameter of the verb meaning (accusative of reference). Kulikov’s hypothesis for the development of types of the same form (labile types) is as follows: (i) NP-nominative + V‘prospers’ (ii) NP-nominative + NP-cognate object-accusative + V‘prospers’ (iii) NPi-nominative + NP(part of NPi)-accusative + V‘X prospers regarding hisi financial condition’ (iv) NPi + NP(part of NPj)-accusative + V‘makes prosper’ The change from (iii) to (iv) is called by Kulikov ‘alienation of parameter’: bhúma vȓĞvam puúyati ‘earth prospers in everything [which is on it]’ ĺ transitive interpretation: ‘earth makes everything [which is on it] prosper’. Kulikov only refers to the specific verb but hypothesises that the reanalysis of intransitive verb with accusative of reference (denoting a parameter of the verb action) into transitive/causative verb, and, consequently, into patterns of verbs of same form in intransitive and transitive (labile verbs), may also be extended to other labile verbs.

3.4.2 Derivation of transitivity alternations and voice: morphological instability in transitivity alternations and differentiation between anticausative and passive The changes in voice are not connected to the changes in argument structure. I have attempted to examine this with data from the diachrony of the Greek language, and I will endeavour in this section to strengthen it with some more cross-linguistic data. The voice system is characterised by diachronic instability, which I attribute to variation and changes which finally result (in every case) in the morphological differentiation between anticausative and passive. 3.4.2.1 English The changes in the morphological encoding of voice in verbs participating in transitivity alternations in the history of the English language can be distinguished into two sub-classes: (a) Morphological change of the causative type and syncretism of causative and anticausative type Quite a number of cases of apparent causativisation or anticausativisation in reality comprise morphophonological changes in Old English which led to the merger of different earlier (causative and anticausative) verb types. The morpho-phonological changes led to the creation of verbs that do not have any morphological or phonological difference in their causative and anticausative use (labile types). This does not mean that they comprise innovative appearances of causative or anticausative structures, but that due to morphological changes there is a phonological merger of the causative with the anticausative type. In the majority of cases, it is the anticausative type which changes and becomes identical with the causative (Ȃöhlig & Klages 2002). Even so, the fact that the specific types (which participate in the transitivity alternation) begin from one period and after to show phonological merger, and that the same form is used for both of the structures, is not without significance for transitivity alternation. The overall picture of the phenomenon indicates that this merger of the causative and the anticausative type either derives from syntactic change (causativisation of the intransitive type) or from morpho-phonological change (of the intransitive type in analogy to the transitive type). According to this second shift, the transitive type is transformed, loses its additional characteristics which distinguished it, and takes exactly the same form as that of the intransitive type75. This precise tendency for syncretism I translate syntactically as presence of v in both the causative and anticausative type (see paras. 2.1.1.2 and 4.3). 75

In many traditional grammatical descriptions of Greek, there is talk of innovative intransitive use of the earlier transitive type in cases in which, in reality, we have change of the intransitive type (which pre-existed and was not created as an innovative type) resulting in its merger with the transitive type.

Diachronic Data and Analysis

169

More specifically, Old English retained the distinction it had inherited from Proto-Germanic between a primary strong stative or inchoative type and a secondary weak causative type: sincan ‘to sink’ meltan ‘to melt’ brinnan/ beornan ‘to burn’

– – –

sencean ‘to cause to sink’ mieltan ‘to cause to melt’ bærnan ‘to cause something to burn

The phonological differentiation in the alternation meltan (intransitive) – mieltan (transitive) had probably already been neutralised in Old English, whereas the first evidence of the phonological merger of the two types as meltan – mieltan, brinnan/beornan – bærnan comes from Middle English texts (the transitive type changes ĺ the intransitive type finally prevails also for the transitive: meltanmelt) (Visser 1963-73: 131). At the same time, yet another morphological change led to the coincidence of transitive and intransitive types with the prevalence of the intransitive type for both structures: the abandonment of the use of the prefixes ge- and be- with transitive verbs during the Medieval period (Krahe & Meid 1969: 37). For Old English, the preverbal ge- functioned as a causativisation morpheme since its addition to an intransitive verb resulted in the creation of a causative verb. feallan ‘to come down’ hwîtian ‘to be/become white’ minsian ‘to diminish/get smaller’

– gefeallan ‘to bring sth. down’ – gehwîtian ‘to make sth. white’ – geminsian ‘to make sth. smaller’ (Visser 1963-1973: paragr. 134, Kastovsky 1992: 380)

From the following table of causative verbs with the prefix ge- it seems that there are no factors obstructing intransitive (unaccusative) verbs from becoming causative; that which probably hinders several verbs (denoting change-of-state from an internal cause) from being transitive seems only to be language conservatism (tendency towards a return to the previous state) which slows down language development (cf. Eythórsson 2003, for subject case morphology in the history of the Icelandic language). Just as in modern languages, in Old English the area of the transmission of light includes a great number of anticausative predicates in contrast to the causative predicates, which are few in number (Díaz Vera 2000). scinan ‘to emit light/be shining’ bierhtan ‘to shine brightly’ beorhtian ‘to become brightly’ glówan ‘to shine with sudden bright light’ twinclian ‘to shine with a winking light’ scymrian ‘to shine with a trembling light’ lihtan ‘to begin to shine’

gescinan ‘to make sth. emit light/be shining’ gebierhtan ‘to make sth. shine brightly’ gebeorhtian ‘to make sth. become brightly’ -

Similarly, with the prefix be- intransitive verbs turned into causatives: dyrnan ‘to hide-intr.’ flow ‘to roll’

bedyrnan ‘to conceal’ beflowan ‘to pour’

The verbs with these prefixes were lost at the end of Old English and during the Early Middle English period (for Visser the reasons are phonological). The loss of these prefixes resulted in the creation of the same types for both the transitive and the intransitive structure. (b) Morphological change of the anticausative type and syncretism of causative and anticausative type The anticausative type changed and formed exactly as the causative in the period subsequent to Old English. The following table shows such cases: Table 3.2 Syncretism of causative and anticausatives type in English Middle English drye(n) ‘to dry up’ (< Old English drygan : adjective dry), circa 1200 Early Middle English fill (< Old English fyll), beginning 17th century Middle English close(n) (< Old English clysan / Middle English closen : close < Old Frisian clos – adjective), circa 14th century Modern English clean (from adjective clean), 18th century (Ȁrahe & Meid 1969: 37; OED 2009)

170

Chapter Three

The retention of the morpho-phonological differentiation is, however, also possible, as would appear in the cases of licgean – lecgean, feallan – fellan (lie – lay, fall – fell). On the other hand, openian ‘to open’ is already used without morphological change for both the causative and the anticausative structure already in Old English. It is very difficult (as all the studies of Old English agree) to determine whether the anticausative or causative use was the first since the reconstruction of Proto-Germanic does not give trustworthy evidence. The cognate earlier verbs of the Germanic family also exhibit identical types for the transitive and intransitive structure: Icelandic: Old Saxon: Old High German:

opna ‘to make sth. be open / become open’ opanon ‘to make sth. be open / become open’ offanõn ‘to make sth. be open, be known / become open’ (Heidermanns 1993: 640)

Furthermore, during the Middle English period the general tendency is for the reflexive pronoun which accompanies intransitive verbs to be abandoned: only some intransitive active verbs appear at the same time and with the reflexive pronoun (Mustanoja 1960: 431): (15) a. he rideth hym ful right he rode.3SG himself.ACC full right ‘then sharply he did veer off to the right’ (Chaucer, The Canterbury Tales: The Knight's Tale (CT I), 1691 [1: 1693]; AD 14) b. No neer Atthenes wolde he go ne ride ‘No nearer Athens would he walk or ride’ (Chaucer, The Canterbury Tales: The Knight's Tale (CT I), 968 [1: 970]; AD 14)

The same alternation (active verb with or without reflexive pronoun) is also observed in Early Modern English: (16) a. ǿ prepared my self to be redye I prepared myself to be ready ([Helsinki Corpus] An Elizabethan in 1582: the diary of Richard Madox, 84; AD 16) b. So the Frenchmen prepared to interrupt his arrival ([Helsinki Corpus] The trial of Sir Nicholas Throckmorton, 66 Cii; AD 16)

Many intransitive verbs which were formed with a reflexive pronoun had already lost the reflexive pronoun by the 16th-17th century. In Early Modern English, the verbs cure, shape, sell, hide begin to be used without a reflexive pronoun (Hermodsson 1952: 65; Visser 1963-1973: 145). The result of this change is the merger of the causative and the anticausative type. The gradual decrease of the use of reflexive pronouns is considered to have helped generally in the formation of intransitive verbs without other morphological marking (Visser 1963-1973 and OED 2009): (17) One desperate greefe cures with an others languish ‘A desperate grief is cured by another grief’ (Shakespeare, Romeo and Juliet, I, ii; AD 16-17)

Just as the denotation of intransitive verbs with the structure “active verb + reflexive pronoun” was abandoned, so the reflexive pronoun took over the role of the expression of the emotional involvement of the subject and of the emphasis: (18) The Bees high them home as fast as they can ‘The bees hasten home as fast as they can’ (Warder, True Amazons, 124 [OED s.v. ‘hie’ v. 3]; AD 18)

3.4.2.2 Latin – Romance languages (a) Reflexive pronouns (Latin) In Archaic Latin (as in Classical Greek), the majority of anticausative verbs were expressed mainly with a non-active (passive) type (suffix -r)76, whereas alternation between the non-active morphology 76

The non-active suffix -r of Latin is usually dealt with in literature either as a morpheme which initially had an impersonal interpretation (Ernout 1908-09: 273-279; Lindsay 1894: para. 21, 1897: para. 65), or as a nonactive suffix (Ǻassols de Climent 1948: para. 5; Hofmann & Szantyr 1965: para. 62, note a) which first had a

Diachronic Data and Analysis

171

and the analytical structure “active verb + se” or “active verb + sibi” is in evidence (Cennamo 1998a, b; Reichenkron 1933; Hatcher 1942; Kemmer 1993). The expression of the anticausative through the type active verb + se especially spreads in Late Latin (Meyer-Lübke 1899: III, para. 382-385; Reichenkron 1933; Orbàn 1974; Cennamo 1993, 1998a, b; Michaelis 1998). (19) donec se vulnus limpidet then SE wound.NOM clean.3SG ‘and then the wound cleanses itself’ (Mulomedicina Chironis, 86; AD 4) (20) servare se possunt save.INF SE can.3PL ‘they can save themselves’ (Palladius, De Agri, 3, 25, 18; AD 4)

During the Empire era, the use of the pronoun se for the marking of the anticausative verbs increased and was also extended to verbs which bore the non-active suffix -r in both the Archaic and the Classical Latin periods. (21) minuente se morbo memory SE weakened ‘memory weakens’ (Plinius, Naturalis Historiae, 23, 50; AD 1)

Consequently, anticausatives were marked more and more with a pronoun in periods when the nonactive verb ending consistently denoted passive structures. At the end of the 4th century, active morphology (-o) begins to be used more and more (alternatively to a pronoun) with anticausative verbs denoting a spontaneous change-of-state (muto ‘to change’, scindo ‘to break’) but also states which often include the presence of an external cause (sano ‘to cure’, vexo ‘to pressure’) (Hofmann & Szantyr 1965; Feltenius 1977). In the same period, the type se + active verb also denoted passive structures since its use was also extended to verb meanings which implied an agent, resulting in the se also becoming the marker of passive in Late Latin (from approximately the 4th century AD) (Cennamo 1998a, b, 1999)77. The marking of the anticausative with active morphology (the same as for the causative) and not with a pronoun (or non-active morphology) correlates with the hypothesis of the dissociation of the morphology of the anticausative from the morphology of the passive. When the passive begins to be denoted with a pronoun, the use of the pronoun for the anticausatives is significantly restricted and the anticausatives are denoted with active endings like the causatives. Similarly, just as, the expression of the passive with a periphrastic type is extended (esse/fio/venire + past participle; Ǽrnout 1909; Reichenkron 1933; Winters 1984; Michaelis 1998), so too is the use of the pronoun se in the anticausatives extended again (see the Old Italian data above). The generalisation which arises from the above is as follows: in the case where (i) the anticausative marker also expresses the passive type, the new anticausative types tend to be expressed with a different marker from that of the passives; (ii) the passive marker also expresses the anticausative type, the new passive types tend to be expressed with a different marker from that of the anticausatives. (b) Passive periphrases (Latin – Old Italian) The pronoun se, which is, at the beginning of Old Italian, a marker of the anticausative structures (after

77

passive interpretation (Flobert 1975; Baldi 1977). 3rd-4th century AD: se also with intransitive verbs which mean mental process, 5th-6th century AD: se also with transitive verbs which mean mental process, 8th-9th century AD: se with intransitive verbs which mean speech action. The use of se with transitive verbs reflects a restructuring process in the system of voice (it possibly occurred in the 4th-5th century AD), when the passive endings were being replaced (increasingly) in everyday speech by the structure se (sibi) + active verb in all the verb groups (Norberg 1943: 151-174; Cennamo 1998a, b). In Late and Medieval Latin the coincidence of se and sibi becomes more systematic, until se replaces sibi in its environments (this also comprises yet another aspect of the same phenomenon, of the gradual loss of the passive endings). According to Cennamo, se and sibi mark the unergative and the unaccusative types respectively, in the 4th and 5th century AD. Also of interest is the coincidence of verb groups which take the pronoun sibi and appear with DP in the accusative case in the subject position. Both of these changes (extension of sibi (instead of the passive endings) and extension of accusative case (instead of nominative)) are dealt with by Cennamo as changes in the direction of active syntax.

172

Chapter Three

the extension of the periphrases for the passive structures), is again extended during Old Italian to verbs which did not allow an anticausative interpretation (passive structures with presence of agent). Therefore, a new ambiguity arises between the anticausative and passive type (Michaelis 1998). (22) passive (with presence of agent): perché la nostra città si regge da Ghibellini then the our ciy SE governed by Ghibellinis ‘then our city was governed by the Ghibellinis’ (Dino Compagni, II, 3, III; AD 13-14; Kontzi 1958: 70) (the agent here is a concrete group of people)

In contrast to Old Italian, in Modern Italian the structure *se + PP-agent is ungrammatical. This contrast between Old and Modern Italian indicates that a new change preceded in the direction of the obligatory absence of agent in the structures with se (obligatory anticausative interpretation)78. Sansò (2005/2009) analyses the aforementioned change79 as a new change of se from passive (+agent) into anticausative (obligatory absence of the agent; the PP denotes only the cause) marker. The passives with se and presence of agent are much more frequent in Old Italian than in the following periods of the language, whereas in Modern Italian they have been lost. Consequently, the course of se in Italian is: anticausative and passive Æ anticausative80. More specifically: in the first vernacular texts of Old Italian of the 13th century, the passives with the pronoun (SE) and presence of the agent are very widespread; the change of the reflexive into a passive marker is at an advanced stage in the first texts of Old Italian81. In the next period, from 1350 until 1450, the availability of the presence of an agent in structures with a pronoun still exists but the number of passives with pronoun and presence of an agent decreases significantly. The number of structures with a specific agent is reduced even more in the following period, from 1450 until the 17th century. In the 17th and 18th century only specific types of agent appear and are, indeed, very rare in structures with a pronoun, whereas the presence of a generic agent is still possible: (23) a. la monarchia di Persia si rovesciò da Alessandro the monarchy of Persia was abolished by Alexander ‘The Persian monarchy was overthrown by Alexander’ (Vico, 33; AD 18) b. per la festa della lunazione di febbraio si fanno per tutti alcune stiacciate until the celebrations of moon of February SE make from everybody some pies ‘until the February moon celebrations some pies were made by everybody’ (Lorenzo Magalotti, Relazione della China, 112; AD 17)

Finally, in the 19th century only one limited group of generic agents is possible: da tutti ‘from everyone’, da molti ‘from many’, dagli uomini ‘from human species’82. 78

Se + PP-agent is ungrammatical in Modern Italian, French, Portuguese, Catalan and grammatical but only with a collective subject in Spanish. For the extension of the pronoun SE in passive and impersonal structures (as a form of grammaticalisation), cf. Kemmer 1993; Cennamo 1993; Wehr 1995; Parry 1998; Sansò 2005/2009. They all recognise the direction: reflexive > middle/anticausative > passive/impersonal. Michaelis comments on the return to the situation of Old Italian as blocking of the process of grammaticalisation through regularisation; these processes follow for a second time the course of grammaticalisation but in the opposite direction. 79 Sansò tries to test unidirectionality in the emergence of new interpretations (which he calls ‘grammaticalisation’) of the marker SE: reflexive > middle > (object drop) > anticausative > impersonal/passive > impersonal of verbs which are able not to denote the object > impersonal of intransitive verbs (Cennamo 1993: 34-35; Geniušiene 1987; Kemmer 1993: chapter 5; Michaelis 1998: 86; Haspelmath 2003: schema 17; for Greek and French: Lavidas 2004). Haspelmath (2004: 33) describes phenomena such as the above as ‘retraction’. Retraction appears when in the chain of grammaticalisation a newer member becomes opaque (for example, the impersonal use of man was extended in Middle English but was abandoned later). 80 Conversely, in Greek the non-active with anticausative interpretation recedes; see below and other differences between the diachronic course of the pronoun se and the non-active endings of Greek. 81 che si dovesse bandire la nostra processione la primaia domenica di ciascheuno mese per Angnello banditore ‘that our parade, the first Sunday of every month, is announced publicly by Agnello, the town crier of our city’ (Libro degli ordinamenti della Compagnia di Santa Maria del Carmine, §26, 60.12; AD 13) 82 A parallel development is observed in other Romance languages as well: the structures with pronoun and agent were especially widespread in the early stages but tended to be abandoned during the history of the languages: Piedmontese dialect (passive structures with pronoun and agent until the 18th century; Parry 1998: 97, 105106), Spanish (passive structures with pronoun and agent first appear in the 15th century, but tend to disappear in Modern Spanish; Monge 1955). Sansó interprets the phenomenon according to the approach of prototypicality; in Old Italian, the overlap of the periphrastic passive structure and the structure with the

Diachronic Data and Analysis

173

The development of the markers of (anti)causativity, and especially of passive periphrases and the pronoun se in Italian, has been examined under the prism of the grammaticalisation approach. The study and the theory of grammaticalisation are outside the scope of this monograph, but the specific phenomena of changes which are observed in relation to the morphological markers and the analytical denotation of the passive, the reflexive and the anticausative type concern every study of the diachrony of (anti)causativity. The development of periphrases in Old Italian which gradually undertake the expression of the passive, leaving the pronoun to denote only the anticausative, can be given schematically as follows (Michaelis 1998): verb + adjective ĺ anticausative interpretation and absence of agent ĺ passive interpretation and presence of agent In what follows, I outline the development of two periphrases (fieri, venire) which initially had anticausative interpretation, then anticausative and passive interpretation, finally to lose the anticausative interpretation, which, moreover, is only denoted with the pronoun se and not with passive periphrases. (a) auxiliary verb fieri ‘exist’, ‘become + adjective’ (factive meaning) ĺ (with perfect participle and change-of-state verbs) anticausative types (24) ficum contundito usque dum minutum fiat fig.ACC cut.2SG until small become.3SG ‘Cut the fig until it is reduced to small pieces’ (Chiron 890; AD 4)

The connection between the above structure and transitive clauses with the corresponding lexical verb, here with the verb minuere, results in the periphrasis receiving anticausative interpretation. The presence of a corresponding transitive structure leads to the anticausative reanalysis of the periphrasis: (25) Claudia ficum minuit Claudia.NOM fig.ACC shrink.3SG ‘Claudia makes the fig shrink’ (the corresponding transitive structure)

The anticausative interpretation usually arises when the auxiliary fieri is connected to the perfect participle. The anticausative interpretation has the function of the intermediate type for transition to the passive interpretation. (26) scalda etqualiter, ut tota sca(l)data fiat heat evenly that all heated becomes ‘heat evenly, so that it becomes in its entirety heated’ (Compositiones Lucenses, L26; AD 7-8; Svennung 1941)

Of course, in the era when the passive interpretation of the periphrasis appears (with availability of the presence of an agent), the earlier anticausative interpretation (without presence of agent) also remains available. The distinction of the anticausative and passive type usually happens from the context. The fact that the former interpretation also remains active, even after the appearance of the innovative type, is connected to the stages of grammaticalisation where the earlier and new type co-appear (Heine 1993: 48ff.). An example of an unambiguous passive type from the era of passive reanalysis of the periphrasis is the following: (27) sed per sacerdotes fiant gubernatas but by priests become.3PL govern.PART.FUT.ACC ‘but they will be governed by the priests’ (Cap. Gen. 783; AD 8; Muller 1924:79, Reichenkron 1993: 43)

(b) verb venire motion verb Æ + adjective (factive meaning) Æ anticausative Æ passive

pronoun was greater, whereas in Modern Italian, the complete differentiation of the two structures has been formed and the non-prototypical cases which are covered by both structures are few.

174

Chapter Three

This concerns a change from lexical to grammatical meaning (grammaticalisation, cf. the classic studies of Lehmann (1982, 1985), and for Greek diachrony, cf. Moser (1988)). Regarding this periphrasis, the coexistence of the earlier with the newer structure had already been observed in both the pre-classical and classical Latin. In that period, both the factive and the anticausative newer interpretation for the periphrasis were possible. anticausative (with perfect participle): (28) confusa venit vox confused.NOM come.3SG voice.NOM ‘the voice is confused’ (Lucretius, De Rerum Natura, 4, 560; 1 BC; Lukrez 4, 562)

In vulgar Latin, there are still examples of ambiguity of the verb venire as (i) a lexical motion-verb, (ii) a factive type, and (iii) an anticausative type. There are, however, also examples which are unambiguously anticausative: (29) sed magis ut deveniant quasi vaporati but more how become as vaporise ‘but many times they vaporise in a way’ (Anthimus, 10, 21; AD 6)

From the 4th century, the verb venire can have passive interpretation (venire + participle of verbs which take two arguments). The full picture of the aforementioned changes is that: (a) the verbs venire and fieri develop gradually from lexical to auxiliary verbs of passive structure; (b) at the same time as the extension of periphrases for the passive types, the type with se instead of the periphrasis also begins to be extended for the anticausative structure. The general tendency, therefore, which we ascertain in relation to the anticausative and the passive type is an attempt at the dissociation of the anticausative type so that the anticausative is expressed with a different morphology from the passive; the anticausative bears active endings when the passive begins to be denoted with the pronoun se or again is expressed with the pronoun se when the passive is consistently denoted with periphrases. The tendency for morphological dissociation between the anticausative and the passive appears in the tables below: Table 3.3 Morphological dissociation and periphrases periphrases (fieri, venire + adjective) interpretation stage a -lexical interpretation of ĺ verb + adjective

stage b -anticausative interpretation -also lexical interpretation of verb + adjective

ĺ

stage c -passive interpretation -also lexical interpretation of verb + adjective -no anticausative interpretation

Table 3.4 Morphological dissociation and pronoun SE pronoun (SE) interpretation stage a stage b stage c stage d -reflexive ĺ -anticausative interpretation ĺ -passive interpretation ĺ -anticausative -also reflexive -(also anticausative interpr.) -also reflexive -also reflexive -no passive interpretation

Diachronic Data and Analysis

175

Table 3.5 Morphological dissociation between anticausatives and passives the general course of the changes 1. a. ANTICAUSATIVE periphrasis

b. ANTICAUSATIVE pronoun se

Æ 2.

Æ 3.

PASSIVE with the same periphrasis

ANTICAUSATIVE it changes to a structure with pronoun (se), so it differentiates from the passive

PASSIVE pronoun se

PASSIVE it ceases to be marked by a pronoun and differentiates from the anticausative

Moreover, Portuguese diachrony (Naro 1968) strengthens the position that this type of changes has, as its final result, the morphological dissociation of the anticausative and passive type: the pronoun (se) is extended to the anticausatives, when their use for passive types ceases to be productive and the passives are expressed periphrastically. In Old Portuguese, the following structure was possible: pronoun se + verb + PP-agent (passive structure with se) The structures with se were not possible with anticausative types in Old Portuguese, but are available from 1500 onwards: (30) Ƞnde se trata de pecado e viçios where SE sin and violence were cured ‘where the sin and the violence were cured’ (Ropica 8; AD 16)

In Modern Portuguese, the presence of PP-agent with the pronoun se (only anticausative structures) is not available, and the passive structure is formed only as a periphrastic structure with the auxiliary ser83: Portuguese SE: passive ĺ passive + anticausative interpretation ĺ anticausative interpretation; the passive is expressed periphrastically The same course is observed in French diachrony84 (Heidinger & Schäfer 2006): (a) the PP par-agent with the pronoun (SE) is ungrammatical in Modern French (Ruwet 1972; Melis 1990a, b; Lamiroy 1993; Lagae 2002); (b) the PP par-agent with the pronoun (SE) was grammatical in previous periods of the language (Stéfanini 1962; Ǻrunot 1965, 1966; Ruwet 1972; Fellbaum & Zribi-Hertz 1989; Lamiroy 1993)85. 83

84

85

In Russian as well, the denotation of anticausative and passive with the same morpheme was abandoned; therefore, it is concerned with a general phenomenon. An important distinction exists between languages with passive periphrasis and languages without passive periphrasis. Cf. Heidinger & Schäfer (2006): (a) the PP par-cause was not lost during the history of French; (b) the PP parcause was already more frequent than the par-agent from the first period (it is logical because initially the structure was anticausative, not passive; it later also received passive interpretation and in the end only anticausative); (c) the PP par-cause became more frequent in the third period and is still present in Modern French, but with lesser frequency. For French as well, Lagane (1967) ascertains the following two tendencies from the classical period until today: (i) Tendency for transitivisation of intransitive verbs, of which Claude Favre de Vaugelas (17th century) had expressed his objection: “Sortez ce cheval, pour dire faites sortir ce cheval, ou tirez ce cheval est très mal dit, encore que cette façon de parler se soit rendue fort commune à la court et par toutes les provinces”. (ii) Tendency to limit the pronoun se, and expression of the intransitive with active morphology, as can be observed in everyday oral speech. Even if specific diachronic studies are absent, the group of labile verbs in French appears to have been

176

Chapter Three

French SE: anticausative + passive (availability of cause + agent) ĺ ĺ only anticausative (only cause, not agent); the passive is expressed only with periphrasis The instability in the area of voice is not only concerned with European languages. Instability is also observed in Vedic. Kulikov (1999a, 1999b, 2003) argues that the schema of labile verbs has a secondary character. Specifically, this schema arises from a series of factors: (a) from the many functions of the middle endings (anticausative, passive, reflexive intransitives as well as middle transitives [benefactive interpretation]), (b) from the merger of several middle participles which are formed from passive and non-passive types, (c) from the syntactic reanalysis of intransitive structures with an accusative of reference or cognate NP in the accusative as transitive/causative, (d) from the use of the perfect type in both transitive and intransitive structures (already from ProtoIndo-European)86. In Middle and Late Vedic, the formation of labile verbs is getting restricted. In the Atharvaveda text, very few labile types are in evidence. The loss of labile types correlates with two morphological changes: (a) with the extension of two morphemes of valency change: the new causative morpheme -Ȑya(Jamison 1983) and the passive morpheme -yȐ- (Kulikov 2001a) comprise overt markers of changes of transitivity. (b) with the de-grammaticalisation87 of middle voice (Kulikov 2003, 2006): many functions of the PIE middle voice are transferred to other markers (for example, reflexivity is frequently denoted by the pronoun tanú- (which had the original meaning ‘body’)88.

3.4.3 Summary The main endeavour of this section was to test the criteria of distribution and analysis proposed in the previous sections on data from more languages. I argue that a dual distinction of (a) causative and (b) accusative verbs is especially important and reflects on syntactic verb changes: the changes in causative verbs are changes in argument structure, whereas the changes in the accusative verbs are mainly changes in case. I argue, furthermore, that the distinction between changes in argument structure and changes in voice is supported by data from diachronic development; the changes in voice do not encode the direction of the changes in argument structure (causativisations). The cross-linguistic data presented here are mainly concerned with English, for which we also have rich sources of diachronic data chiefly in relation to the changes in argument structure, and the Romance languages in relation to Latin, as the information that we have for the developments in voice is particularly detailed. According to the criteria that I posited for Greek, the changes in the verb transitivity are distinguished into changes in argument structure (causative from anticausative type) and changes in case morphology (extension of accusative/structural case). The course of changes in argument structure is in one direction: causativisation with the introduction of the agent argument for earlier anticausative non-alternating types. We have ascertained that these criteria for the diachronic change in transitivity also apply cross-linguistically. The process of causativisation concerns all periods of English and is mainly favoured by the presence of passive interpretation (+agent) of anticausative verbs and by the presence of the structure “auxiliary verb ‘be’ + passive participle” both for transitive verbs (passive type) and for intransitive verbs (type that denotes the result of the action). Regarding the remaining directions of change which are witnessed (anticausative from alternating verb), we have ascertained

86 87 88

extended. Vaugelas mentions: “...il faut remarquer que de toutes les erreurs qui se peuvent introduire dans la langue, il n’ y en a point de plus aisée à établir que de faire un verbe actif d’ un verbe neutre, parce que cet usage est commode en ce qu’il abrège l’expression, et ainsi il est incontinent suivi et embrassé de ceux qui se contentent d’ être entendus sans se soucier d’ autre chose; on a bien plus tôt dit sortez ce cheval, ou entrez ce cheval que faites entre ce cheval”. It is precisely this “d’ être entendus” that leads to the preservation of a language tradition, slowing the development to a certain extent. The co-appearance of two morphological types (in our case) or structures clearly presents in general a transition to a new situation. Cf. Giannakis 1992, 1997a. With the meaning of the loss of interpretations which existed in earlier periods. Kulikov argues that, for Vedic, labile types are not a primary grammatical phenomenon, as they are for example for Modern English, but of secondary character.

Diachronic Data and Analysis

177

that they relate to cases of abandonment of causativisation in a later period, and in no case is pure anticausativisation observed (anticausative from causative verb). Apart from the causativisation process, transformation of intransitive into accusative verbs resulted in the presence of a direct object for verbs in Modern English that did not take a direct object in Old English. This development is the result of: (a) the change from PP to DP of the complements of motion verbs; (b) replacement of the genitive by the accusative case for the DP denoting the cause. As regards voice, it appears also cross-linguistically that voice is not connected to the processes of (anti)causativisation. The result, indeed, of changes in the morphology of causative and anticausative types is, in many cases, the same as the result of syntactic causativisation: syncretism of the causative and the anticausative type. In English, the phonological merger of the two types arose from morphophonological change of the transitive (loss of prefix or change of the root vowel) or of the intransitive type (loss of the reflexive pronoun). The changes in voice in Latin and in the Romance languages are particularly important for an overall interpretation and the general linguistic principle which I maintain governs the changes in transitivity (differentiation of anticausative and passive type) and which I will endeavour to formulate in detail in the next chapter (chapter 4). ǿ interpret the instability in voice which is observed in Latin and the Romance languages as follows: passive and anticausative type change so that they are encoded morphologically in a different way; in cases where the passive type was encoded (due to other changes) in the same way as the anticausative type (e.g., with the same periphrasis), the anticausative changed (e.g., a structure with a pronoun is used instead of the periphrasis). The aforementioned observation is based on the change: (a) of the expression of the anticausative with a clitic (SE) to its expression with an active form, when the use of the clitic (SE) for the passive type became productive in Late Latin; (b) of the expression of the anticausative with a periphrasis to its expression with a clitic (SE), when the marking with a periphrasis of the passive type in Italian also became productive. In Late Latin (approximately from the 4th century AD), the earlier anticausative type “se + active verb” also denoted passive structures, when its use further extended to verb meanings which implied agent; this resulted in the expression of the anticausative with an active form (without pronoun). When, however, the passive was expressed systematically with periphrastic type, the anticausative was again marked with the pronoun se. On the other hand, at the beginning of Old Italian, the pronoun se is a marker of the anticausative but is extended again during Old Italian to verbs which formed passive structures with presence of the agent. The ambiguity between anticausative and passive leads to a new reanalysis which results in a new change, again in the direction of the obligatory absence of the agent from the structures with se (exclusively anticausative) appearing in Modern Italian. The passive is denoted with periphrases; these periphrases also follow the same course (at a certain period denoting both the passive and the anticausative) to end in the exclusive expression of the passive. The development of periphrases in Old Italian, which gradually take on the expression of the passive, allowing the pronoun only to denote the anticausative, is as follows: “verb + adjective” ĺ anticausative interpretation and absence of agent ĺ passive interpretation and presence of agent. The same observation holds true both for French and for Portuguese: the anticausative is denoted with pronoun in the period when the passive ceases to be grammatical with pronoun and begins to be denoted with periphrasis. All the above strengthen the hypothesis of the dissociation of the anticausative from the passive.

CHAPTER FOUR DIACHRONIC COURSE AND DERIVATION OF TRANSITIVITY ALTERNATIONS

In this chapter, I attempt to give an overall interpretation of the diachrony of transitivity with reference to data and the changes during all periods of the Greek language, but also according to cross-linguistic diachronic remarks deriving mainly from the history of the English language. In the previous sections, we observed the distribution of argument structures and voice morphology and endeavoured to find connections between changes and factors which led to these changes. In relation to transitivity alternations, the direction of changes was stable (in all periods): production of innovative causative types from active or non-active anticausatives. Regarding the voice system, the passives are constantly non-active from the Hellenistic-Roman period, and, in parallel with that (and mainly from the Medieval period), the anticausatives change and begin to bear active voice morphology. In this section I will attempt to interpret the changes overall and with the help of rich sources of diachronic data to form an adequate interpretation of the structure of causative and anticausative types. In the two first sections I examine and analyse the following according to the theoretical model which I put forward in the second chapter (Lightfoot’s approach, extended VP and lexicon-syntax interface): (a) the ambiguity and structural reanalysis that appear to play a significant role in these changes (diachronic interpretation of causativisations); (b) the relationship of voice morphology with the direction of the derivation of causatives and anticausative types. The chapter ends with some arguments from the diachronic study which strengthen the hypothesis of the causative template.

4.1 Reanalysis, Cues, and Causativisation 4.1.1 Reanalysis of intransitive as transitive Ambiguity and reanalysis appear to comprise the factors that instigate changes in the area of (anti)causativisation and play a deciding role in defining the way in which these changes appear. Commenting on the innovative passive meaning of middle transitive verbs, Parker (1976) observes that the structural analysis of the passive sentence can be assigned to a sentence with middle transitive meaning, in cases where (i) the word order is object – verb, (ii) there is ambiguity concerning the case of the object, and (iii) the subject is not expressed overtly: (1)

ΘϲȱΩΗΘΙȱΦΐϾΑΉΘ΅΍

to astu amunetai a. Θϲȱȱȱ ΩΗΘΙȱȱȱȱȱȱȱȱȱ ΦΐϾΑΉΘ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ǻΓЇΘΓΖǼ to astu amunetai (outos) the.ACC city.ACC defend.NACT.PRES.3SG[middle interpretation] (he-implied subject) ‘(he) is defending the city with all his strenghth’ (middle transitive structure) b. Θϲȱȱȱȱ ΩΗΘΙȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΦΐϾΑΉΘ΅΍ to astu amunetai the.NOM city.NOM defend.NACT.PRES.3SG[passive interpretation] ‘the city is protected’ (passive structure)

From the moment that the structural analysis of the sentence as passive (resulting in its ambiguity as middle transitive and as passive) became also possible, the middle transitive structure was abandoned as a productive process and the middle transitive type was turned into a transitive type without any semantic difference from the active transitive type, to be abandoned gradually (the active (and not the middle) type is used after this change systematically and productively in transitive structures). The above interpretation is supported by the general position of Parker: (a) language change is the result of hypotheses which the native speaker forms by constructing his/her grammar, (b) more than one hypothesis can arise in relation to an area of data resulting in the creation of erroneous

180

Chapter Four

correspondences between structural constituents and sentences, (c) a deviant hypothesis can be formed if (i) it does not violate Universal Grammar, (ii) it does not create semantically insoluble problems. Ligthfoot (1979) describes in more typical terms these changes as follows: the two opaque structures (middle transitive and passive) were reanalysed leading to the domination of the passive structure according to the Transparency Principle. He hypothesises that two changes occurred during the transition from Proto-Indo-European to Early Greek. The consequence of these changes was opacity which caused in its turn the loss of the transitive middle diathesis. Referring to Vennemann (1974a, b, c) and Friedrich (1976), Lightfoot hypothesises that Proto-Indo-European had as underlying structure the structure Subject – Verb – Object; change in the basic word order before the early texts of Greek led to the structure Subject – Object – Verb (the analyses/structures (a) and (b) are possible). With the underlying structure Subject – Verb – Object a clause such as: Θϲȱ ΩΗΘΙȱ ΦΐϾΑΉΘ΅΍ / to asty amynetai could be analysed in only one way (2b). Only after the new change to Subject – Object – Verb were both structural analyses/interpretations available (analyses/structures (a) and (b) are available). This opacity/surface ambiguity (which arose from the aforementioned change in word order) ceased to exist with the gradual loss of middle transitive meaning and finally the clause ΘϲȱΩΗΘΙȱΦΐϾΑΉΘ΅΍ / to asty amynetai could only be analysed as (2b). (2) a.

VP

b.

VP

 DP ȱ



V

DP

V

ΘϲȱΩΗΘΙȱȱȱȱȱȱȱȱ

ΦΐϾΑΉΘ΅΍ȱ

ΘϲȱΩΗΘΙȱȱȱȱȱȱȱȱȱȱ

ΦΐϾΑΉΘ΅΍ȱ

to asty the city [accusative]

amynetai defend.NACT.PRES.3SG [transitive with middle interpretation]

to asty amynetai the city defend.NACT.PRES.3SG [nominative] [passive]

According to the remarks above and to Lightfoot, two stages for the loss of middle transitive diathesis are to be observed: (i) non-active transitive verbs (additional benefactive meaning [‘interest of the subject for the action that the verb describes’]) ceased to be distinguished semantically from the corresponding active transitive verbs; (ii) further to that, passive (non-active II) endings prevailed in the past and future tenses in place of the middle (non-active I) endings in cases where middle (non-active I) endings had passive interpretation. I assume that we have a similar reanalysis of argument structure in cases of causativisation in the diachrony of Greek, and chiefly in sentences with the presence of a neuter DP since neuter DPs bear the same morphological ending for the nominative and accusative case. In (3) the declension of a neuter gender noun in Classical Greek is shown (Jannaris 1897/1968: 111-120): (3)

example (neuter DP): ΘϲȱΈЗΕȬΓΑ / to do:r-on / ‘gift, present, gift of honour’ Singular Plural ΘϲȱΈЗΕȬΓΑȱ ΘΤȱΈЗΕȬ΅ȱ Nominative

ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

Accusative
ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

Vocative
ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

Genitive ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

Dative
ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

˜ȱ˜DZ›Ȭ˜—ȱȱ ΘϲȱΈЗΕȬΓΑȱ ˜ȱ˜DZ›Ȭ˜—ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΈЗΕȬΓΑȱ ˜DZ›Ȭ˜—ȱȱȱȱȱȱȱȱȱȱȱȱȱ ΘΓІȱΈЏΕȬΓΙȱ ˜žȱ˜DZ›Ȭ˜žȱȱȱȱȱȱȱȱȱȱȱȱȱ ΘЗ΍ȱΈЏΕȬΝ΍ȱȱ ˜DZ’ȱ˜DZ›Ȭ˜DZ’ȱȱȱȱȱȱȱȱȱȱ

Šȱ˜DZ›ȬŠȱȱȱȱȱȱȱȱȱ ΘΤȱΈЗΕȬ΅ȱ Šȱ˜DZ›ȬŠȱȱȱȱȱȱȱȱȱȱ ΈЗΕȬ΅ȱ ˜DZ›ȬŠȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ ΘЗΑȱΈЏΕȬΝΑȱ ˜DZ—ȱ˜DZ›Ȭ˜DZ—ȱȱȱȱȱȱȱȱȱȱȱȱȱ ΘΓϧΖȱΈЏΕȬΓ΍Ζȱ ˜’œȱ˜DZ›Ȭ˜’œȱȱȱȱȱȱȱȱ

ȱ ȱ ȱ ȱ

Intransitive active verbs with a neuter DP in the subject position can be reanalysed as causatives: (4)

ψΗϾΛ΅ΗΉΑȱȱȱȱ ΧΔ΅Αȱȱ Θϲȱȱ ΔΏϛΌΓΖ

he:sykhasen hapan to ple:thos ΧΔ΅Αȱȱ a. ψΗϾΛ΅ΗΉΑȱȱ he:sykhasen hapan calm.ACT.PAST.PERFVE.3SG (intransitive) all.NOM ‘all the people calmed down’

Θϲȱȱ

ΔΏϛΌΓΖ

to ple:thos the.NOM people.NOM

ĺ reanalysis ĺ b. ψΗϾΛ΅ΗΉΑȱȱ

ΧΔ΅Αȱȱ

Θϲȱ

ΔΏϛΌΓΖȱȱȱ

ǻ˜ЈΘΓΖȱȦȱ΅ЂΘ΋Ǽ

Diachronic Course and Derivation of Transitivity Alternations he:sykhasen

hapan

to

calm.ACT.PAST.PERFVE.3SG (causative) all.ACC

ple:thos

181 (houtos / haute:)

the.ACC people.ACC

(s/he – implied subject)

‘(s/he) calmed all the people’

In (5-14) I repeat from sections 3.1, 3.2 and 3.3 causative structures which are innovative in Classical Greek, Hellenistic-Roman Koine and Medieval Greek (the numbers in the second set of parentheses refer to the example numbers in the corresponding section): Classical (5) (7c) σΔΏΉΙΗ΅Αȱȱ πΔϠȱȱ ΔϱΑΘ΍΅ȱȱ ΎϾΐ΅Θ΅ȱȱ ΑΣ΍ΓΑȱȱ ϷΛ΋ΐ΅ȱȱȱ epleusan epi pontia kymata naion okhe:ma sail.ACT.AOR.3PL over sea waves ship.ACC vehicle.ACC ‘they sailed the ships over the waves of the sea’ (Euripides, Iphigenia in Tauris, 409; 5 BC) (6) (11d) ВΚΉΏΓϟ΋ȱȱ ΘΉȱȱ Ύ΅Ϡȱȱ ΉЁΚΕ΅ϟΑΓ΍ȱȱ ̇΍ϱΈΝΕΓΑ o:pheloie: te kai euphrainoi Diodo:ron help.3SG and and gladden.ACT.PRES.3SG Diodorus.ACC ‘he could help and please Diodorus’ (Xenophon, Memorabilia, 2, 10, 6; 5-4 BC) Hellenistic-Roman (7) (6b) БΖȱȱȱȱΘϲȱΗ΍ΘϟΓΑȱȱȱȱȱȱύȱȱΘϲȱΔΓΘϲΑȱȱȱȱȱȱȱȱϊΈΉ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΐξΑȱȱΘϲΑȱΚ΅·ϱΑΘ΅ȱȱȱȱȱȱȱȱύȱȱΔ΍ϱΑΘ΅ȱȱȱ ho:s to sition e: to poton he:dei men ton phagonta e: pionta as

the food.NOM or the drink.NOM enjoy.ACT.PRES.3SG prt.

the eat.PART.ACC or drink.PART.ACC

‘as the food or the drink pleases the person who is eating or drinking’ (Sextus Empiricus, Against the Mathematicians, 7, 368, 3; AD 2-3) (8) (7b) ΐ΋ΈξȱȱȱȱΐχΑȱȱ ΦΑ΅·Ύ΅ΊΓΐνΑΓΙΖȱȱȱȱȱȱȱȱΏΙ·ϟΊΉ΍Αȱȱ Θ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱ ΘЗΑȱ ΐΉΏЗΑȱ

me:de me:n anagkazomenous lygizein ti to:n melo:n neither not force.PART.ACC bend.ACT.PRES.INF one.ACC the.GEN body-parts.GEN ‘neither were they forced to bend one from their parts’ (Aelian, On the characteristics of animals, 2, 11, 46; AD 2-3) (9) (8c) ΏΉΙΎ΅ϟΑΉ΍ȱȱ ΐξΑȱȱ ΘΤȱȱ πΗΌφΐ΅Θ΅ȱ leukainei men ta esthe:mata whiten.ACT.PRES.3SG prt the.ACC clothes.ACC ȱ ‘that whitens the clothes’ (Sextus Empiricus, Against the Mathematicians, 9, 247; AD 2-3) Medieval (and Early Modern) Greek (10) (9c) Έ΍ϱΘ΍ȱȱ ǻϳǼȱȱ̍ϾΕ΍ΓΖȱȱ ȱΦΗΘΉΑΉϧȱȱ ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΓΑǰȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΈ΍ΤȱȱȱΑΤȱΈΙΑ΅ΐЏΗϙȱȱΘϲΑȱȱπΛΌΕϱΑȱȱȱȱΘΓΙȱ dhioti (o) Kirios asteni ton, dhia na dhinamosi ton exthron tu because (the) Lord.NOM become-ill.ACT.PRES.3SG he.ACC.WEAK for to strengthen.3SG

the enemy.ACC his

‘because Lord makes him ill, to strengthen his enemy’ (Leontios Machairas, Chronicle, 3, 421; AD 15) (11) (10b) ·ΓΕ·ϲΑȱȱȱȱΌξΏΉ΍ȱȱȱȱȱȱΘχΑȱȱȱȱȱȱȱȱΌΏϧΜ΍ΑȱȱȱȱȱȱΘ΋ΖȱȱȱΗ΍·φΗΉ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΎ΅ϠȱȱΘϲΑȱȱȱȱȱȱȱȱΔϱΑΓΑ ghorghon theli tin thlipsin tis sijisi ke ton ponon fast

(12) (6b)

want.3SG the.ACC sorrow.ACC her

keep-silence.ACT.PERFVE.3SG and the.ACC pain.ACC

‘she wants to suppress her sorrow and her pain’ (The War of Troy, 5759; AD 14) țȚ Ƞ ȜȠȖȚıȝȩȢ, ȠʌȠȪ ȕĮȜĮ, ȖȘ ȖȚĮȓȞİȚ ȖȘ ĮȡȡȦıIJİȓ ki o lojismos, opu vala, ji jieni ji arosti and the thought.NOM which put.1SG or cure.ACT.PRES.3SG or

ȝİ me

fall-ill.ACT.PRES.3SG I.ACC.WEAK

‘and the thought I had either cures or makes me ill’ (Erotokritos, I, 2094; AD 17) (13) (7b)

ΪΑȱȱσΑȂȱȱȱΎ΅ϠȱȱȱΌνΏϙΖȱȱȱȱȱȱȱȱΑΤȱȱȱΗξȱȱȱȱȱ

ΔΏΓΙΘϟΗΝȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

ΗϾΑΘΓΐ΅ȱȱȱȱ

an en’ ke thelis na se plutiso sindoma if is and want.2SG to you.ACC.WEAK grow-rich.ACT.PERFVE.1SG soon ‘and you would like it so that I will soon make you rich’ (Leonardo Dellaporta, Poems, ǹ, 1985; AD 14-15) (14) (5b)

ΩΑΉΐΓΖǰȱȱΉϨΔΉΑȱȱȱȱȱȱ΅ЁΘφǰȱȱȱȱȱȱȱȱȱȱ·ϟΑΓΐ΅΍ǰȱȱȱȱȱȱȱȱȱΎ΅ϠȱȱσΘΊ΍ȱȱȱȱȱΐΔ΅ϟΑΝȱȱȱȱȱΉϢΖȱȱΘχΑȱΐϾΘ΍Αȱ

anemos, ipen afti, jinome, ke etzi wind said.SG she.NOM become.1SG and so ΅ЁΘΓΑЗΑȱȱ

ϳΔΓІȱȱ Όξȱȱ

ΑΤȱȱ ΘΓϿΖȱȱ

beno is tin mitin enter.1SG in the noseȱ

ȱ

Δ΅΍Ό΅ϟΑΝȱȱ

aftonon opu the na tus petheno these.GEN who want.3PL to they.ACC.WEAK die.ACT.IMPRFVE.1SG ‘I become as the wind, she said, and like that I enter the nose of those whom I want to make die’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 163, 15-16; AD 17)

The innovative causative uses are promoted in all periods of Greek from the following characteristics: (a) the availability of subject drop (null subject) in Greek, (b) the surface word order in Greek which was never fixed, but always free, (c) neuter DPs which are ambiguous and can denote nominative or accusative case, especially when they are in a postverbal position, (d) the characteristics, the position and the use of clitics (which create an unbreakable union with the

182

Chapter Four

verb) since the presence of a clitic as direct object appears to be obligatory in the first phase of the change from intransitive verbs to causatives [(IJȠȞ ʌȡȦșȣʌȠȣȡȖȩ) *(IJȠȞ) ʌĮȡĮȓIJȘıĮȞ / (ton prothipurgho) *(ton) paretisan / (the prime minister) *(he.ACC.WEAK) resign.ACT.PAST.PERFVE.3PL / ‘they made the prime minister resign’ – Roussou & Tsimpli 2007], (e) the earlier structures of psych-verbs with the very frequent presence of a clitic for the experiencer: ȝȠȣ ĮȡȑıİȚ / mu aresi / I.GEN.WEAK like.ACT.PRES.3SG / ‘I like it’ (mainly in relation to the use of clitics with psych-verbs, such as:ȱ ĮȞȘıȣȤȫ anisixo ‘to worry’ – ȝİ ĮȞȘıȣȤİȓ me anisixi I.ACC.WEAK worry.ACT.PRES.3SG ‘it worries me’), (f) the many alternative ways of denotation of the cause with psych-verbs which are available in the history of Greek: PP, DP-dative, DP-accusative (ĮȞȘıȣȤȫ anisixo ‘to worry’ / ȞIJȡȑʌȠȝĮȚ drepome ‘to be ashamed’ + DP-DAT, PP (įȚĮ dhia / ȖȚĮ ja / İȟĮȚIJȓĮȢ eksetias ‘because of’), DP-accusative), (g) the large amount of functional vocabulary1 in Greek. Alexiadou (2006a) indicates that Greek is a language with productive causativisations since it has a relatively large amount of functional vocabulary at its disposal and a relatively small list of roots: in Greek, in contrast to languages such as English, different meanings arise from the combination of functional features with a small number of roots2. In languages such as English that have a large number of roots and a small number of functional categories (functional vocabulary), causativisation is expressed more with periphrastic causative structures (make +DP-Accusative +verb), and roots denoting internal cause of the change-of-state can be causativised with periphrastic causative structures (make Peter laugh). In English, transitivity alternations are possible only with verbs of change-of-state, whereas in other languages transitivity alternations are possible with more verbs, such as (i) with verbs of spontaneous movement, verbs of appearance and existence, or intransitive unergatives (15), and (ii) with verbs that are causatives but non-alternating in English: destroy – țĮIJĮıIJȡȑijȦ katastrefo ‘to destroy’3. (15) Japanese: a. Ȁotozuke-a kie-ta message-NOM disappear-PAST ‘the message disappeared’ b. Dareka-ga kotozuke-o keshi-ta somebody-NOM message-ACC disappear-CAUS.PAST ‘somebody made the message disappear’

Consequently, whereas in English causativisation is productive with periphrastic causative structures (cf. verbs of change-of-state from internal cause; Harley (2005) analyses the lexical verb ‘cause’ as an overt spell-out of Voice4), in Greek there are two possibilities: (a) periphrasis5, (b) clitic (Roussou & 1

2

3

4

The cross-linguistic differences of alternating verbs are interpreted based on the possible combinations of functional items and roots: languages with productive causativisations have a relatively large functional vocabulary but a relatively small list of roots. The different meanings arise from the combination of functional items with roots (Reinhart 2000; Arad 2002); for example, Japanese: keshi-ta: disappears-causative.past English has a relatively large list of roots and a small functional vocabulary. One root can be connected with different parts of the structure and can express all the parts of the structure. Old English had more functional elements; there were prefixes such as ge- which allowed the formation of causative types for verbs which are not available in Modern English: limpan ‘it happens’ – gelimpan ‘to make somebody exist’. After the loss of the functional vocabulary (prefixes and endings) during Middle and Early Modern English, productive causativisation of intransitives begins. In Greek, both of the processes exist: both verb endings and intense causativisation. One hypothesis could be that the causativisations spread when the formation of the unmarked causative type (with the change: non-active ĺ active voice morphology) begins to be productive in Greek; as essentially it is then that the functional vocabulary needed for the differentiation between anticausative and causative ceases to exist (active verb endings for both types). Other differences of English: (a) bare roots can be used either as nouns or as verbs (conflation); (b) the English roots modify v and form resultative structures (in contrast to Greek; cf. Giannakidou & Merchant (1999) and Horrocks & Stavrou (2003)). The roots correspond to the idiosyncratic part of the semantic decomposition of Levin & Rappaport Hovav (Levin 1993; Levin & Rappaport Hovav 1995, 2006; see also para. 2.1.2). Possible interpretation: (i) there is always a causative corresponding type; but it is possible for a verb to have been “frozen” as one type in the lexicon of a language (for example, intransitive Modern English arrive; cf. Reinhart 2000); (ii) the derivational patterns are correlated cross-linguistically with more than one possible classifications; in other words, the corresponding verbs do not mean the same in all languages. Cf. Wright (2002), McKoon & MacFarland (2000), for the relationship between verbs denoting change-of-state from internal cause and causative verbs: (1) a.

Early summer heat blossomed fruit trees across the valley. (LN 1999)

Diachronic Course and Derivation of Transitivity Alternations

183

Tsimpli (2007) analyse the clitic as spell-out of Transitivity/Voice (Ǿ ȖȣȝȞĮıIJȚțȒ IJȘȞ ĮįȣȞȐIJȚıİ IJȘ ȂĮȡȓĮ / i jimnastiki tin adhinatise ti Maria / the.NOM exercise.NOM she.ACC.WEAK makethin.ACT.PAST.PERFVE.3SG the.ACC Maria.ACC / ‘exercise made Maria thin’)).

4.1.2 Cues for reanalysis I interpret the diachronic changes of causativisations according to the cue-based approach of Lightfoot (1999, 2006a, b) assuming that case marking comprises a cue of language acquisition (Eythȩrsson 2000, 2002). According to the cue-based approach, the grammatical output of each new language period is a result of cues that children find when they acquire their language. A cue can be considered a type of structure or a feature of grammar which comes from the input. Children search their linguistic environment for the designated structures or cues. The cues that appear only in specific grammars comprise the parameters (which are defined by the Universal Grammar). The cues are found in mental representations that are the result of the comprehension and the processing of the utterances. When a child understands an utterance, even if only partially, s/he forms a type of mental representation of the utterance. The child searches for these cues in those representations which come from the input (Lightfoot 1999: 149). The cues are components of the I-language as they appear in the mental representations which come from the processing of utterances. They are accepted in the children’s grammar if they are essential and allow for the analysis of sentences which are found in the data of their linguistic experience (Dresher 1999; Lightfoot 2006a: 85-86). Following Eythȩrsson’s approach (for the diachronic analysis of Icelandic6) I argue that the case which is assigned to the DP in one specific structural position –for example, in the [Compl, VP] position– can function as a cue; the different cases mark the different structures in which case is assigned. As referred to above, the neuter DPs in Greek bear the same suffix in the nominative and the accusative case, and, therefore, the neuter DPs in nominative with intransitive verbs can be analysed as DPs in the accusative. When the child acquiring the language searches for cues, s/he may hypothesise, based on the features which s/he meets, that if an unaccusative verb has a patient argument marked

b. c. d.

5

6

Salt air and other pollutants can decay prints. (LN 1982) Raindrops selectively erode clay particles. (BNC B1E) The onset of temperatures of 100 degrees or more, on top of the drought, has withered crops. (NYT 1986)

For verbs denoting change-of-state from internal cause that are causativised (1) the presence of an agent is not possible, (2) the cause appears to be connected inherently with the process of change-of-state. The result in such cases is indirect causativity; the schema of direct causativity is less productive. Cf. the view of Shibatani & Pardeshi (2001) (with indirect causativity the separation of the event-cause and the event-result is possible) and Levin & Rappaport Hovav (1999) and Ramchand (2006a, b) (there is temporal independence of the events). Eythórsson endeavours to explain the “dative sickness” (extension of dative case) and “nominative sickness” (extension of nominative case) in Icelandic according to Lightfoot’s cue-based approach. The dative case replaces the accusative case in Icelandic in the subject of verbs which assign the ș-role of the experiencer. (1) Mig > mer I.ACC I.NOM ‘I need a knife’

vantar need

hníf knife.ACC

The second change concerns the tendency of impersonal structures to become personal, when either the DP in the accusative or the DP in the dative changes into nominative case and agrees with the verb. (2) Bátinn > báturinn rak að landi boat.ACC ship.NOM drifted.3SG to land ‘boat drifted to the shore’ (from Eythórsson 2002)

The DPs which change in these cases bear the grammatical role of the subject (according to the criteria for subjects) in Icelandic (in contrast to corresponding structures in German, for example). The changes for marking of the subject with DP in the nominative case, independent of the meaning of the verb, appear in the entire history of Icelandic; on the other hand, the changes for marking of the subject with dative which concern the semantic group of verbs with subject with the ș-role of experiencer begin from the middle of the 19th century. When a child searches for cues, s/he will hypothesise from the cues that come from the general pattern of the language that the argument in the [Spec, TP] position receives structural case and that this case must be nominative. This may result in a grammar with subjects only in the nominative case. Studies on child language strengthen this hypothesis (children who acquire Icelandic up until the age of 2;5 use nominative case where adults use an oblique subject).

184

Chapter Four

with accusative case, then it is possible also for an external argument (cause or agent) to be introduced in the subject position in the structure of the specific verb. In (16) and (17), examples from Medieval Greek of structures which are ambiguous and consequently comprise structures which can be reanalysed are cited: (16)

(17)

ψΗϾΛ΅ΗΉΑȱȱ ΧΔ΅Αȱȱ Θϲȱȱ ΔΏϛΌΓΖȱȱ ke isixasen apan to plithos and calm.ACT.PAST.PERFVE.3SG all.NOM/ACC the.NOM/ACC people.NOM/ACC ‘and he calmed all the people / all the people calmed down’ (Ioannes Malalas, Chronographia, 16, 19, 35; AD 5-6) ΎȂȱȱ σΔΉΗΉΑȱȱ ȂΖȱȱ ΘΤȱȱ ΔΓΈΣΕ΍΅ȱȱ ΘΓΙǰȱȱ Θϲȱȱ ΎΏΣΐΐ΅Αȱȱ ΦΕΛ΍ΑϟΊΉ΍ k’ epesen ’s ta podharia tu, to klaman arxinizi Ύ΅Ϡȱȱ

and

fell.3SG

to the

feet

his

the.NOM/ACC cry.NOM/ACC begin.ACT.PRES.3SG

‘and she fell at his feet, she began the crying / the crying began’ (ǺGV I, 327, 198; AD 17)

The reanalysis of intransitive structures as causative was strengthened by two more factors: (a) the use of the accusative case both for the object of finite types and for the subject of infinitives7; (b) the possibility of the formation of (passive) participles with the suffix -ȝİȞȠȢ -menos not only for transitive but also for intransitive verbs. An infinitive with a DP in the accusative case may lead to ambiguity and: (i) in the case of active intransitive verb, to causative innovative analysis (infinitive + accusative: the infinitive does not denote agreement and therefore the accusative can be analysed either as subject or as object of the infinitive) (see 18a); (ii) in the case of active transitive verb, to its analysis as active anticausative, if the DP was ambiguous between the accusative and the nominative, and an active causative and non-active anticausative type already existed (DP-accusative as object + active causative infinitive: the DP can be analysed erroneously as subject (as the nominative and accusative cases coincide phonologically) and the verb erroneously as active anticausative verb) (see 18b). (18) reanalysis with result: an innovative causative type a. ΈΓϿΖȱȱȱȱȱȱȱȱȱȱȱȱȱ΅ЁΘХȱȱπΒΓΙΗϟ΅ΑǯǯǯȱȱΘΓІȱΉϢΕ΋ΑΉІΗ΅΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱΘΤΖȱπΎΎΏ΋Ηϟ΅ΖȱΔΣΗ΅ΖȱΘΤΖȱ̄Ϣ·ϾΔΘΓΙȱ dhus afto eksusian... tu irinefse tas eklisias pasas tas Ejiptu give.IMP.2SG him



power

the.GEN pacify.ACT.PERFVE.INF the churches.ACC all

the Egypt.GEN

‘give him the power... so that he can pacify all churches in Egypt’ (Theophanes Confessor, Chronicle, 373B, 5-6; AD 8-9) reanalysis with result: an active (instead of an earlier non-active) anticausative type ΅ЄΒΉ΍Αȱȱ b. ΗΉΏφΑ΋Αȱȱ selinin afksin moon.ACC increase.ACT.PRES.INF   ‘so that the moon increases’ (Papyri Graecae Magicae, I, II, ȋǿǿǿ, 64-65)

The formation of the “passive participle” (ending -ȝİȞȠȢ -menos) is connected with the availability of passivisation of a transitive (causative) verb and therefore with the presence of an active transitive type. In the Medieval Greek period (12th-15th century), extension of the formation of the “passive participle” is observed as well as with verbs which bear active morphology and have stative, anticausative or unergative interpretation (something that is also retained in Modern Greek, cf. Tzartzanos 1946/1991: 330-331; Moser 1988: 145-152; ǹlexiadou & Anagnostopoulou 1999a: 25). Manolessou (2005) provides the following examples (12th-15th century): (19) a. ȖȠȞĮIJȓȗȦ ĺ ghonatizo kneel.ACT b. țȚIJȡȚȞȓȗȦ ĺ kitrinizo yellow.ACT c. ʌİȚȞȐȦ ĺ pinao hunger.ACT d. IJȡȫȦ ĺ troo eat.ACT e. ȞȣıIJȐȗȦ ĺ 7

Cf. Karali 1994.

ȖȠȞĮIJȚıȝȑȞȠȢ ghonatizmenos kneel-MENOS țȚIJȡȚȞȚıȝȑȞȠȢ kitrinizmenos yellow-MENOS ʌİȚȞĮıȝȑȞȠȢ pinazmenos hunger-MENOS ijĮȖȦȝȑȞȠȢ faghomenos eat-MENOS ȞȣıIJĮȖȝȑȞȠȢ

Diachronic Course and Derivation of Transitivity Alternations

185

nistazo nistaghmenos feel-drowsy.ACT feel-drowsy-MENOS

This concerns a Medieval Greek characteristic which is not observed in Classical Greek, except in a minority of cases (Chatzidakis 1924, 1927)8. In the 16th century, only the gerund [with the suffix -ȠȞIJĮȢ -ondas] is found in the active voice, whereas in the non-active voice only the present perfect participle has remained (cf. Karla 2002, for the changes in the Ancient Greek participles in the versions of the “Life of Aesop”); only a very small percentage of the participles of the original post-Classical text remain in the later translation; the greater part is transformed into sentences with a finite verb)9. The fact that passive participles began to be formed for both unaccusative and unergative verbs which bear active voice morphology constitutes an indication of changes in the voice system of Greek during the Medieval period (Manolessou 2005). Moreover, both forms of the endings of the participle change: the active ending for the participles (-ȠȞIJĮȢ -ondas) ceases to exist as a participle ending, and the passive ending for the participles is also limited to nominal use (cf. Lavidas 2003a, for a diachronic examination of the structure İȓȝĮȚ + -ȝİȞȠȢ / ime + -menos / ‘to be + V-menos’; Laskaratou & Philippaki10 1984, Setatos 1985, Sklavounou 2000, Anagnostopoulou 2001a, for the similarities between passive participles and adjectives). (20) a. Ύ΅Ϡȱȱ ΘΤȱȱ ΑΉΚΕΤȱȱ ΗΓΙȱȱ Η΅ΔΓІΑ ke ta nefra su sapun and the.NOM kidneys.NOM your rot.ACT.PRES.3PL ‘and your kidneys are rotting’ (Spanos, ǻ, 311; AD 14/15) b. ΔΓΏΏΣΎ΍ΖȱȱΈξȱȱ Ύ΅Ϡȱ ΩΑȱ ΆΕΝΐϜȱȱ Ύ΅Ϡȱȱ σΑ΍ȱȱ Η΅Δ΋ΐνΑΓΑȱȱ ǻΘϲȱΦΏΉϾΕ΍Ǽ polakis dhe ke an vromi ke eni sapimenon (to alevri) many-times prt and if stink.3SG and is.3SG rot.NACT.PART.NOM (the flour) ‘it (the flour) stinks and is rotten many times’ (An entertaining tale of quadrupeds, 441; AD 14) (21) ϳȱȱ ·νΕΓΑΘ΅Ζȱȱ ΘΕΉΛΣΐΉΑΓΖȱȱ ΘχΑȱȱ ΌϾΕ΅Αȱȱ ΘΓΙȱȱ ΦΑΓϟ·Ή΍ o jerondas trexamenos tin thiran tu aniji the old-man.NOM run.NACT.PART.NOM the door.ACC his open.3SG ‘the old man runs and opens his door’ (ǺGV I, 326, 196; AD 17)

In the Medieval Greek period, as well as in all periods, we can find examples of possible reanalysis resulting not only in an innovative active causative type, in cases of verbs which were exclusively intransitive non-alternating, but also in changes in voice morphology (active instead of non-active anticausative type), in cases of verbs which already had active causative and non-active anticausative type. The reanalysis is particularly favoured in cases of verbs in third person with a neuter DP as their argument; the (neuter) DP can be analysed both as DP in the nominative case that has the role of the subject of the verb in third person and as DP in the accusative case that has the role of the object of the verb in third person (for which a subject with the meaning ‘s/he’ is implied). The reanalysis of the intransitive type can lead to a causative type (earlier intransitive ĺ causative interpretation is possible as well) (22) Medieval Greek Ύ΍ȱ

πΈΤȱ

ΔЗΖȱȱ ΉϨΑ΅΍ȱȱ ΐΔΓΕΉΘϲȱ Θϲȱȱ

ΘΣΗΗ΍ΐΓȱȱ ΑΤȱȱ

Ώ΍ЏΗΉ΍Dzȱȱ

ki edha pos ine boreto to tasimo na liosi? and here how is possible the vow.ACC to melt.ACT.PERFVE.3SG ‘and here how is it possible to make the vow melt away?’ (The Sacrifice of Abraham, 672; AD 16)

8

9

10

The same phenomenon can be observed generally in Balkan languages: non-active participles (Lindstedt 2002; Tzitzilis & Symeonidis 2000). Dimitra Theophanopoulou-Kontou (personal communication) also indicates the reverse tendency in Modern Greek (non-active deponent verb ĺ active gerund): İțȝİIJĮȜȜİȪȠȝĮȚ ekmetalevome exploit.NACT ĺ İțȝİIJĮȜȜİȪȠȞIJĮȢ ekmetalevondas exploit.ACT.GER, įȑȤȠȝĮȚ dhexome accept.NACT ĺ įȑȤȠȞIJĮȢ dhexondas accept.ACT.GER. Cf. Haspelmath (1994): (i) passive participles usually begin as resultatives (they describe a situation which arises from a previous event) ĺ this also explains the formation of these participles in many languages for both causative and unaccusative verbs; (ii) in certain languages, the resultative participles develop into pure passive participles (orientated towards the object, and not just towards the patient argument); (iii) these participles finally also lose their past tense characteristic. Moser (2005) observes that Greek types in -ȝİȞȠȢ -menos retain the resultative nature, something which also explains their use in the so-called present perfect II of active voice. Laskaratou & Philippaki place the derivation of the middle passive participle in the lexicon and not in the transformational domain.

186

Chapter Four

The reanalysis can lead to an active anticausative type (change in voice morphology) (23) Hellenistic-Roman period ΓЁΎȱ ΦΑΓϟ·Ή΍ȱȱȱ

Θϲȱ

ΗΘϱΐ΅ȱȱȱȱ

΅ЁΘΓІȱ

ouk anoigei to stoma autou not open.ACT.PRES.3SG the mouth.NOM his ‘his mouth does not open’ (New Testament, Acts, 8, 32; AD 1)

In (24-29) I repeat from sections 3.2 and 3.3 examples with active (after a change) anticausatives from the Hellenistic-Roman Koine and the Medieval Greek period (the numbers in the second parenthesis refer to the example numbers in the corresponding section): Hellenistic-Roman period (24) (60c) ΅ЁΒΣΑΉΘΉȱȱȱȱȱȱȱȱȱȱȱΈξȱȱπΑȱȱΛΣΕ΍Θ΍ȱȱΎ΅Ϡȱ·ΑЏΗΉ΍ȱȱΘΓІȱȱΎΙΕϟΓΙȱȱψΐЗΑȱȱȱȱΎ΅ϠȱȱΗΝΘϛΕΓΖȱд̌΋ΗΓІȱ̙Ε΍ΗΘΓІ auksanete de en khariti kai gno:sei tou kyriou he:mo:n kai so:te:ros Ie:sou Khristou grow.ACT.IMP.2PL prt in grace

and knowledge the Lord.GEN our

and Savior

Jesus.GEN Christ.GEN

‘grow in the grace and knowledge of our Lord and Savior Jesus Christ’ (New Testament, 2 Peter, 3, 18, 1; AD 1) Θϲȱ ΗΛϛΐ΅ȱȱ ΦΏΏΣΗΗΉ΍ȱ (25) (63a) Ύ΅Ϡȱȱ ΆΏνΔΉȱȱ kai blepe to skhe:ma allassei and see.IMP.2SG the.NOM form.NOM change.ACT.PRES.3SG ‘and see that the form is changing’ (Dorotheus, Greek fragments, 421, 2; 1 BC-AD 1) Medieval (and Early) Modern Greek (26) (67a) Ύ΅Ϡȱ ψȱ ΜΙΛφȱ ΐΓΙȱȱ πΘΕϱΐ΅ΒΉΑȱ ke i psixi mu etromaksen and the soul.NOM my scare.ACT.PAST.PERFVE.3SG ‘and my soul was scared’ (Digenis Akritas, Ǽscorial, 1113; AD 12) (27) (67b) ΗφΐΉΕΓΑȱ πΗΆνΗΘ΋Αȱ ϳȱȱȱȱ ȱȱΓΙΕ΅ΑϱΖǰȱȱπΘΕϱΐ΅ΒΉΑȱȱ ϳȱȱ ΎϱΗΐΓΖȱȱ simeron ezvestin o uranos, etromaksen o kozmos today erased.3SG the.NOM sky.NOM scare.ACT.PAST.PERFVE.3SG the.NOM people.NOM ‘today the sky was erased, the people scared’(Leonardo Dellaporta, Poems, ī, 349; AD 14-15) ψȱȱ Ύ΅ΕΈ΍Σȱȱ ΐΓΙȱȱȱ  (28) (72c) ΋ЄΚΕ΅ΑΉȱȱ ifrane i kardhia mu  gladden.ACT.PAST.PERFVE.3SG the heart.NOM my ‘my heart gladdened’ (Ptochoprodromos, ǿ, 266; AD 11-12)  (29) (70a) ώΏΌΉΑȱ ψȱ ΑϾΒ... ȱ πΔΏΣΘΙΑΉΑȱȱ Θϲȱȱ ΗΎϱΘΓΖȱȱ ilthen i niks... eplatinen to skotos came.3SG the night.NOM widen.ACT.PAST.PERFVE.3SG the darkness.NOM ‘the night came... the darkness widened’ (Kallimachos and Chrysorroi, 2229; AD 14)

The question that arises is why these changes which appear so natural did not prevail completely, resulting, for example, in the causativisation of all the intransitives (denoting change-of-state) or in the active voice morphology for all the anticausatives. The incomplete extension of causativisations and active anticausatives could have been attributed to the conservatism of the Greek language11 which is a homogeneous language in a large degree with a strong tradition of linguistic purity (cf. the observations of Theophanopoulou-Kontou on language “errors” and non-acceptance of certain types by the community of native speakers). Moreover, intransitive verbs with anticausative meaning of change-of-state without an external agent cannot be causativised. For example, the verbs ȝȠȣȤȜȚȐȗȦ muxliazo ‘to mould’ and ĮȞșȓȗȦ anthizo ‘to blossom’ / ĮȞșȫ antho ‘to blossom’ are not causativised in any period of the Greek diachrony. I assume that the unavailability of lexical causativisation is linked to the unavailability of a relation between the meaning of those verbs and the meaning of the direct cause of the change-of-state since in these specific verbs only indirect causativisation is possible, as the periphrastic causative structures, in which they participate, show (Bittner 1999; Piñon 2001): (30) Ǿ ijȡȠȞIJȓįĮ 11

IJȠȣ țȘʌȠȣȡȠȪ

ȑțĮȞİ

IJȠȞ țȒʌȠ

ȞĮ

ĮȞșȓıİȚ

Eythórsson argues that the incomplete extension of the nominative case could have been attributed to the conservatism of Icelandic, a language without dialectical differences, spoken in an isolated country without language contact but with extension of school education in all social classes. On the other hand, the availability of the dative case for the subject-experiencers also comprises a cue for children who overgeneralise the dative with arguments denoting the experiencer resulting in the loss of subjects which bear accusative case and denote the experiencer. The result of such a change is thematic case for the DP-subjects. Despite the attempts of the purists in Iceland, the dative case is extended and may be considered as the norm for the subject-experiencers for many speakers.

Diachronic Course and Derivation of Transitivity Alternations i frondidha tu kipuru ekane ton the care.NOM the gardener.GEN made.3SG the ‘the care of the gardener made the garden blossom’ (31) The heat caused the flowers to wilt (32) Ǿ ȖȣȝȞĮıIJȚțȒ ȑțĮȞİ IJȘ ȂĮȡȓĮ i jimnastiki ekane ti Maria the exercise.NOM made.3SG the Maria.ACC ‘the exercise made Maria lose weight’

187

kipo na anthisi garden.ACC to blossom.ACT.PERFVE.3SG ȞĮ na to

ĮįȣȞĮIJȓıİȚ adhinatisi lose-weight.ACT.PERFVE.3SG

The intransitive verbs of change-of-state from internal cause are not alternating, even though they have causative meaning in periphrastic causative structures denoting indirect cause of the change-of-state. The meaning of change-of-state from internal cause does not rule out all types of causativisation but only direct causativisation.

4.2 The Role of Voice in Syntactic Changes 4.2.1 The course and the interpretation of changes in voice The picture of the diachronic development of transitivity alternations is completed with the ascertainment of changes in voice morphology of certain anticausative verbs, without further syntactic changes in the causative and anticausative use of those verbs. These changes are linked to the instability in the voice of the anticausative types and not to the production of new types and the direction of the production; new transitivity alternations are always the result of the reanalysis of an intransitive verb as transitive and are not related to the changes in the voice system. The differences in voice can be found in the following areas: (a) Ancient Greek: (i) non-active type + DP-accusative as direct object; (ii) active passives; (iii) non-active anticausatives. I assume as a working hypothesis for this present section that the non-absorption of the accusative case by the non-active voice is correlated with the presence of active passives and that the non-active anticausatives are connected with the presence of active passives; (b) from the Early Medieval Greek period: (i) *non-active type + DPaccusative case as direct object (the non-active suffix blocks/absorbs the accusative case); (ii) nonactive passives; (iii) active anticausatives. The absorption of the accusative case by the non-active type is correlated with the productivity, from the Hellenistic-Roman but mainly the Early Medieval period, of non-active passives, whereas the extension of active anticausatives is connected with the absence of active passives. Non-active anticausative verbs change into actives (ĮȞȠȓȖİIJĮȚ anoigetai open.NACT ĺ ĮȞȠȓȖİȚ anoigei open.ACT) from the Hellenistic-Roman period but mainly from the Medieval Greek period. Two factors which concern the specific period of change need to be taken into consideration for the interpretation of those changes (and the instability in voice morphology). First, the availability of passive interpretation, namely the presence of an agent for certain of the anticausative verbs (interpretation: ‘the door was opened by the locksmith’), created the need for the anticausatives to be differentiated from the passive structures (absence-presence of agent) through the active voice morphology of intransitive anticausatives. Second, more general changes are observed in the area of voice morphology during the Hellenistic but mainly the Medieval period: the presence of non-active voice morphology results only in the absence of a direct object in the accusative case, namely in the absorption of the accusative case; the indirect object (which is in the genitive or dative case in the corresponding transitive structure) is not absorbed (the passivisation of the goal argument is no longer possible). Table 4.1 Formation of labile types in Greek Classical Greek: ΦΑΓϟ·Ν anoigo: ΦΑΓϟ·Γΐ΅΍ anoigomai

Hellenistic and Medieval Greek: ΦΑΓϟ·Ν anoigo: / anigho ΦΑΓϟ·Ν anoigo: / anigho

causative ‘open’ anticausative ‘open’

active non-active

causative ‘open’ anticausative ‘open’

active active

Concerning the absorption of the accusative case, which I consider to be a very important characteristic for the entirety of the changes in voice morphology, the following need to be noted about

188

Chapter Four

the differences between Ancient Greek (non-absorption of accusative case) and Medieval Greek and the reverse direction taken by the same phenomenon as observed in the history of the French language. (a) Absorption of the accusative case In Ancient Greek, theme and goal arguments could both be in the subject position of a non-active type; the absorption of the genitive and dative case was available, even if the alternative syntax of ditransitive verbs with both DPs in the accusative case was not available12: ¾ active Verb + DP-ACC + DP-DAT ¾ (a) non-active Verb + DP-DAT + ЀΔϱ hypo by-PP (b) non-active Verb + DP-ACC + ЀΔϲ hypo by-PP  ĺȱȱ ΩΏΏΓȱȱ Θ΍ȱȱ ΐΉϧΊΓΑȱȱ πΔ΍ΘΣΛΌ΋ΗΌΉȱȱ ǽЀΐΉϧΖǾ allo ti meizon epitakhthe:sthe [hymeis]   other.ACC something.ACC greater.ACC impose.PASS.FUT.2PL [you]  ‘you will be imposed some greater demand’ (Thucydides, The Peloponnesian War, 1, 140, 5; 5 BC) the corresponding transitive structure: ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ 

πΔ΍ΘΣΒΓΙΗ΍ȱȱ ΩΏΏΓȱȱ Θ΍ȱȱ ȱΐΉϧΊΓΑȱȱ ЀΐϧΑȱȱ epitaksousi allo ti meizon hymin impose.ACT.FUT.3PL other.ACC something.ACC greater.ACC you.DAT ‘(they) will impose on you some greater demand’

The situation in Classical Greek resembles the situation in Modern English: both the argument in the accusative and the argument in the dative of the corresponding transitive type can be subjects of the passive type. (33) a. John gave the book to Mary b. John gave Mary the book a’ The book was given to Mary b’ Mary was given the book

(Acc-Passive) (Dat-Passive)

The passive of a dative structure (Dat-Passive) is relatively recent in the history of the English language. In Modern English, the cases of passivisation of objects of prepositions are rare (He ruled (over) mighty and diverse nations ĺ The three kingdoms were ruled over by a single sovereign). In Classical Greek, these cases of passivisation are more frequent: a large number of verbs which took an object in the dative or genitive case in transitive structures either as an option or obligatorily were able to have these DPs in the subject position in their passive structures. a.

b.

πΔ΍ΆΓΙΏΉϾΝȱȱ

΅ЁΘХ

epibouleuo: plot-against.ACT.PRES.1SG

auto:i he.DAT

Ύ΅Θ΅ΚΕΓΑЗȱȱ

΅ЁΘΓІȱ

kataphrono: autou think-slightly-of.ACT.PRES.1SG he.GEN (34) classical Ancient Greek ϵΗΓ΍ȱȱȱȱ

ГΗΔΉΕȱȱ Ύ΅Ϡȱȱ

ψΐΉϧΖȱȱ

ЀΔȂȱȱ

̝Ό΋Α΅ϟΝΑȱȱ

πΔ΍ΆΓΙΏΉΙϱΐΉΌ΅ȱȱȱȱȱ

hosoi ho:sper kai he:meis hyp’ Athe:naio:n epibouleuometha who.NOM as and we.NOM by Athenians plot.NACT.PRES.1PL ‘all who like us are the objects of the designs of the Athenians’ (Thucydides, The Peloponnesian War, 1, 82, 1; 5 BC)

Therefore, in Classical Greek non-active voice morphology did not block the assignment of accusative case (Lightfoot (1979) (under the prism of Transformational Grammar) characterises the passivisation in Ancient Greek as a lexical rule). In (35-41) I repeat (from sections 3.1 and 3.2) examples with a non-active verb type and a DP in the accusative case from the Classical Greek and the Hellenistic-Roman period (the numbers in the second 12

That is a possible structure for ditransitives in Modern English (Adams 1971; Feldman 1978; Rappaport Hovav & Levin 2008). For dative alternation and passivisation, cf. Larson (1988a, b); Baker (1985); Anagnostopoulou (2003b). In Classical Greek also, the passivisation of the goal is grammatical only when the goal is animate. Adams (1971) argues that this concerns a combination of goal and experiencer which is passivised.

Diachronic Course and Derivation of Transitivity Alternations

189

set of parentheses refer to the example numbers in the corresponding section): Classical Greek (35) (78a) πΎȱȱȱȱ ΘϛΖȱȱ ΘΓϾΘΓΙȱȱ ΛΉ΍ΕϲΖȱ πΘ΅ΐ΍ΉΙϱΐΉΌ΅ȱȱ ΘΓϿΖȱȱ ΑϱΐΓΙΖȱ ek te:s toutou kheiros etamieuometha tous nomous  from the he.GEN hand dispense.NACT.IMPERF.1PL the.ACC laws.ACC ‘the laws were dispensed to us from his hands’ (Lysias, Against Nicomachus, 30, 3; 5-4 BC)  (36) (78b) ЀΔϲȱȱ ΘΓІȱȱ ̏΍Ώ΋ΗϟΓΙȱȱ ̝Ε΍ΗΘΓ·ϱΕΉΝȱȱȱ ΦΔΉΗΘνΕ΋ΘΓȱȱ ΘχΑȱȱ ΦΕΛφΑ hypo tou Mile:siou Aristogoreo: apestere:to te:n arkhe:n by the Milesian Aristagoras deprive.NACT.PLUP.3SG the.ACC rule.ACC ‘he was deprived of his rule by Aristagoras of Miletus’ (Herodotus, The Histories, 6, 13; 5 BC) (37) (78c) ΩΏΏΓȱȱ Θ΍ȱȱ ΐΉϧΊΓΑȱȱ ΉЁΌϿΖȱȱ πΔ΍Θ΅ΛΌφΗΉΗΌΉ allo ti meizon euthys epitakhthe:sesthe other.ACC something.ACC more.ACC immediately dictate.PASS.FUT.2PL  ‘you will be immediately dictated some more oppressive condition’ (Thucydides, The Peloponnesian War, 1, 140; 5 BC) Hellenistic-Roman period (38) (40a) ςΑȱȱ ΔΑΉІΐ΅ȱȱ πΔΓΘϟΗΌ΋ΐΉΑ hen pneuma epotisthe:men   one.ACC Spirit.ACC drink.PASS.AOR.1PL ‘and were all given to drink into one Spirit’ (New Testament, 1 Corinthians, 12, 13; AD 1) (39) (40b) πΑΉΈ΍ΈϾΗΎΉΘΓȱȱ ΔΓΕΚϾΕ΅Αȱ enedidysketo porphyran ȱ ȱ clothe.NACT.IMPERF.3SG purple.ACC ‘he was clothed in purple’ (New Testament, Luke, 16, 19; AD 1) (40) (41a) ΦΔ΅΍Θ΋ΌφΗΓΑΘ΅΍ȱȱ Θϲȱȱ πΔ΍·νΑ΋ΐ΅ apaite:the:sontai to epigene:ma    demand.PASS.FUT.3PL the.ACC profit.ACC ‘they will demand for the profit’ (Tebtunis Papyri 61b 359 = 72, 355) (41) (41b) Θϲȱȱ ϡΐΣΘ΍ϱΑȱȱ ΐΓΙȱȱ ϶ȱȱ ΔΉΕ΍ΉΆΉΆΏφΐ΋Α to himation mou ho periebeble:me:n the.ACC clothes.ACC my which.ACC wear.NACT.PLUP.1SG ‘the clothes which I have worn’ (Magdalen Papyrus 6, 6, 221a)

From the Medieval Greek period, however, the co-presence of a non-active type and a DP in the accusative case as direct object is not possible. The course of these structures in the diachrony of the French language is exactly the reverse. I will now consider the reverse direction as observed in French to reach general conclusions concerning the changes in voice morphology and the presence of DP in the accusative as direct object of non-active types. (b) The reverse direction (French) Of particular interest is the reverse course (in relation to the non-absorption of the accusative in Ancient Greek and the absorption of the accusative in Medieval Greek) which is ascertained in the diachrony of the French language (Miller 2001): in Modern French, structures with clitic (se) and a DP in the accusative that denotes the part of the body of the subject are completely grammatical (Miller refers to these structures by the fairly traditional term indirect reflexives)13. SE + direct object in the accusative case (42) a. Il se lave he SE washes ‘he washes his hands’ b. Il se rogne he SE shaves ‘he shaves his beard’ c. Il s’ essuya he SE wiped ‘he wiped his eyes’

les the.ACC

mains hands.ACC

la the.ACC

barbe beard.ACC

les the.ACC

yeux eyes.ACC

In Old French (and before 1150) structures such as the above were ungrammatical. Verbs with clitic se were unable to take a DP in the accusative as direct object. Structures which expressed meanings

13

Only with objects which are not separated from the subject (inalienable) (*ils se lavent la chemise) and only if the object is influenced by verb action (*il se tourne la tête).

190

Chapter Four

corresponding to the meanings of Modern French with clitic and a DP in the accusative were as follows: (43) a. Il lui lave he he.DAT washes ‘he washes his hands’ b. Il lave ses he washe his ‘he washes his hands’

les the.ACC

mains hands.ǹCC

mains hands.ACC

Miller (2001) assumes that the changes from Old to Modern French allowed the incorporation (Baker 1988, 1995) of se into the verb, and that the incorporation of se into the verb results in the nonabsorption of the accusative case by the se. The ethical dative and the accusative of reference (in which case, se does not absorb the dative case) comprised the cues for the innovation of the use of se + DP in the accusative case (indirect reflexivity). In cases of ethical dative and accusative of reference the clitic did not absorb the case which is assigned by the verb, something which aided the necessary structural requirement for the presence of the structure “se + accusative”, which according to Miller is incorporation14. The structures in Modern French are analysed as follows: se is incorporated into the verb root (for example, lav- ‘to wash’), and this union is moved to the functional category of Voice15. (44) VoiceP  Subject ili Voiceǯ   Voice VP   V DP lavsei  In Greek, the direction of the changes was the reverse: (a) Ancient Greek: non-active type + Accusative ĺ (b) Medieval Greek: *non-active type + Accusative. The cues for the non-availability of the presence of DPs-accusative as direct objects of non-active types need to be sought in the extension of the accusative case as case of the objects (instead of the genitive and the dative; Catsimali 1990) and in the loss of the benefactive meaning which non-active voice bore in Greek. The benefactive meaning resulted in the productive presence of structures “active type + accusative” (absence of personal interest for the verb action) vs. “non-active type + accusative” (presence of personal interest for the verb action). When the benefactive meaning was lost, the structure “active type + accusative” became more systematic and was no longer in opposition to the structure “non-active + accusative” with a different 14

15

Miller assumes that the following five cues led to the new situation (se + accusative): (1) environments where se did not need case -raising structures: plus se fait fiers ‘s/he became more vivid’ (Chanson de Roland – R 111,1) -intransitives of change-of-state: Carles se pasmet ‘Carles died’ (R 2891) (2) datives: si li briset les os ‘s/he broke his bones’ (R 1200) (3) structures with two objects, the second of which receives case through coindexing: si la baisa le pié ‘he kissed her on her foot’ (A 443) (4) ethical dative: Rolland nel se doüst penser ‘Rolland should not have thought that’ (R 355) (5) every clitic appears close to the verb to which it refers: d’ euls desarmer ne s’oublia ‘to disarm them, he did not forget’ (Clavi 2265) In (1): se without case, in (2) and (4): se in free distribution with dative pronoun, in (5): incorporation of se into the verb. The structures where se did not take structural case strengthened the presence of the necessary structural conditions for the incorporation of the clitic se into the verb: the adjunction of a head X to a structurally governing ȋȠ-level category is termed incorporation (Baker 1988, 1995). Since, in general, the incorporated elements do not need case, the incorporated se allows the presence of an additional object in the accusative (the structural (accusative) case of the verb has not been assigned). After the incorporation of se, the clitic se appears exactly before the inflected verb also in structures with auxiliary verbs and periphrastic causative verbs (faire): (1) il se peut defendre ĺ il peut se defender (2) il se fait laver aux enfants ĺ il fait se laver les enfants

Diachronic Course and Derivation of Transitivity Alternations

191

meaning. Consequently, non-active voice was correlated with the absence of the accusative. I assume that this change is also linked to the presence of active anticausatives during the same period. In the following section I concentrate on the interpretation of the extension and on the structure of active anticausatives.

4.2.2 Voice head, anticausative, and passive I assume that the non-absorption of the accusative in Ancient Greek as opposed to the absorption of accusative in Medieval Greek is linked to the presence of active passives and non-active anticausatives in Ancient Greek as opposed to the presence of non-active passives and active anticausatives in Medieval Greek. What I propose for the interpretation of these structures is the presence of Voice [agent] for the anticausatives and absence of Voice for the passives of Ancient Greek, in contrast to the absence of Voice for the anticausatives and presence of Voice [+agent] for the passives of Medieval Greek. According to this approach, the presence of the Voice head is connected to the non-active morphology of Greek (cf. Rivero 1990 and section 2.1.1) and with the introduction or the nonavailability (absorption) of introduction of the internal argument in the accusative case (cf. Collins 2005, para. 2.1.1.2, and below in this section). The diachronic data from English (and Latin) (section 3.4), in which it appears that the morphological marking of the anticausative structure changes when it becomes similar to the morphological marking of the passive structure, add another argument for the non-correlation between the anticausative and the passive structure. I assume that the non-connection of the same morphological marker at the same time with both the passive and the anticausative types holds as a general linguistic principle. This principle is applied to the diachrony of Greek as follows: (a) Ancient Greek: when non-active voice is not directly connected to the syntax of passive types [passive types frequently bear active voice (with presence of agent)] non-active voice productively marks the anticausatives (stable presence of non-active anticausatives)16, (b) Medieval Greek: when non-active voice marks in a stable manner passive types (in reality, beginning from Hellenistic Koine, but mainly from Early Medieval Greek), then the marking of anticausatives with active voice is productive17. (a) Derivation of passive types Concerning the structure of passive types I follow and modify the approach of Collins (2005), who argues that the external argument is present and is generated also in the passive structure in the same structural position in which it is generated in the transitive structure (cf. also 2.1.1.2, point b). According to prevailing approaches, the non-active ending absorbs not only the accusative case but also the external ș-role of the verb18. A problem with these approaches is the fact that the agent (the external argument) is generated in the subject position ([Spec, vP]) in transitive structures, whereas it comprises the complement of the preposition ‘by’ in passive structures. The aforementioned problem can be dealt with if we follow Collins’s analysis, which suggests the same position for the external argument in both the passive and transitive structures. Collins, based on the analysis of the relationship between transitives and passives by Chomsky (1957: 42-43, 78-81) and translating it according to the Minimalist Programme, proposes an analysis according to which the external argument is generated in the passive structure in the same way that it is also generated in the transitive structure. Consequently, the external argument with the passive type is generated in the [Spec, vP] position in exactly the same position as with the transitive type (cf. also Goodall 1997; Watanabe 1993: 337; Mahajan 1994: 297), whereas the non-active voice morphology does not absorb the external ș-role but only the accusative case19. According to the rules of subcategorisation of v, v accepts the preposition 16

17

18

19

A parallel cross-linguistic argument: when passive types begin to be marked with periphrases, the pronoun becomes the stable expression of anticausatives (see para. 3.4.2). A parallel cross-linguistic argument: when the pronoun is also used for passive types, active voice morphology begins to prevail in the anticausatives (see para. 3.4.2). I do not follow the analysis of the absorption of the external ș-role by the non-active voice which is proposed by Embick; Embick with the [non-act(ive)] feature in v unites all the types (passives, anticausatives) which are considered not to have an external argument. The passive, however, has the external argument of the agent in the Įʌȩ/apo ‘by’-Phrase. On the other hand, I follow Embick concerning the possibility of the presence of the [non-act] feature either in v (from the Hellenistic and mainly in the Medieval Greek period and afterwards) or in Root for the deponents; I assume that this holds not only for the deponents but also for the non-active anticausatives – for details, cf. the following paragraph. I do not follow Embick in the connection of [non-act] in v with the absence of the external argument (cf. Collins 2005). According to the diachronic data, non-active voice does not absorb and does not block the assignment of the

192

Chapter Four

‘by’ in its Spec in passive but not in transitive structures. ǿ maintain that the difference between Greek and English passive types is generally concerned with the differences of the passive structure in languages with verb voice morphology (Greek) and languages with periphrastic passive type (auxiliary verb + participle, English). While, according to Collins (cf. 2.1.1.2 point b), the preposition ‘by’ in passive structures of languages with periphrastic passive types does not form a PP but is the head of the VoiceP20, I assume that in languages where voice morphology is marked as a suffix the preposition ‘by’ does not derive in the Voice head but the Voice head is responsible for the non-active morpheme (cf. the classic analyses of Rivero and Tsimpli). In Greek, the head movement (v ĺ Voice) is responsible for the non-active morphology21. The preposition Įʌȩ apo ‘by’ together with the DP comprise a constituent for Greek and are generated (as PP) in the [Spec, vP] position. The presence of the PP in the [Spec, vP] position is justified if we accept Chomsky’s basic principle (1957), as translated by Collins (2005) in terms of the Minimalist Programme, for the presence of the external argument in the same position in both transitive and passive structures; there is no reason to deny the possibility of the expression of the external argument as PP if we consider the possibility of expression of the external argument in the dative case (namely, with morphological case, in parallel to the possibility of expression as PP) in classical Ancient Greek, Hellenistic Koine and many other languages. Moreover, even Collins initially proposes for English the solution of the presence of the PP-external argument in the [Spec, vP] position, which he rejects for reasons of economy (that have to do with the features of v in the lexicon), confronting, however, a serious problem in the analysis of Kiswahili which has verb voice morphology (he finally declares that he is unable in the present stage to interpret the passive types of this language). v (but not V) is optionally subcategorised for the presence of by-PP in the Spec position (Watanabe 1993: 337) only in the passive and not in the transitive structures. Consequently, v has the optional subcategorisation frame [_ by Prep]. One solution for Greek is to accept that the subcategorisation frame [_ by Prep] is an optional characteristic which is added when the numeration has been completed (for example, when v has been chosen by the lexicon; cf. Chomsky 1995: 236 for the optional characteristics). The presence of the PP in the [Spec, vP] position is optional, in exactly the same way that the presence of subject of transitive types is optional in languages with rich verb morphology, such as Greek22. Therefore, as can be seen from the structures below, the correspondence between the transitive and the passive structure is complete for the Greek language. Passive type; presence of the agent argument, the external argument (agent) as PP in [Spec, vP] (45) IJȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ ĮȞȠȓȤIJȘțİ to xrimatokivotio anixtike the.NOM safe.NOM open.NACT.PAST.PERFVE.3SG ‘the safe was opened by the robbers’

20

21

22

Įʌȩ apo by

IJȠȣȢ tus the

ȜȘıIJȑȢ listes robbers

accusative case in Classical and Hellenistic Greek, but absorbs the accusative case in Medieval and Modern Greek. In English, the Voice head (‘by’) checks the accusative case of the DP which is in the [Spec, vP] position (where the traditional complement of the preposition ‘by’ is found) in the same manner in which ‘for’ in English checks the case of the DP in structures such as [CP For John to win will be nice]. The fact that in Greek the case assignment in structures such as the above is not the same also supports my proposal for modification of the structure according to the Greek data. Moreover, I assume that the structure I propose for Greek and its differentiation from the corresponding English structure is associated with the availability of null subject in Greek in contrast to English. In contrast to English and languages with periphrastic passive types, where phrase movement and not head movement takes place. Cf. also Collins’s remarks in the same paper on the differences between English and Kiswahili. The following sentences comprise evidence that the external argument is structurally present in the passive structures even when it is not expressed (cf. structural presence of non-overt subjects in pro-drop languages). The non-overt external argument can bind a reflexive (1), it can licence the presence of a depictive secondary predicate (2); it behaves, in other words, exactly like the overt external argument (3). (1) a. Such privileges should be kept to oneself (Baker, Johnson & Roberts 1989: 228) b. Damaging testimony is always given about oneself in secret trials (Roberts 1987a) (2) a. The book was written drunk b. At the commune, breakfast is usually eaten nude c. This song must not be sung drunk (Baker 1988: 318) (3) a. Breakfast is eaten nude by the campers b. Breaksfast is eaten by the campers nude

Diachronic Course and Derivation of Transitivity Alternations

193

TP  T VoiceP  Voice vP non-active  suffix PP vǯ (Įʌȩ apo by)  external v VP argument  position V DP  Įʌȩ-IJȠȣȢ-ȜȘıIJȑȢ ĮȞȠȓȤIJȘțİ IJȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ apo-tus-listes anixtike to xrimatokivotio by-the-robbers open.NACT.PAST.PERFVE.3SG the.NOM safe.NOM Passive type; the external argument (agent) in [Spec, vP] position is not phonologically expressed (46) IJȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ to xrimatokivotio the.NOM safe.NOM ‘the safe was opened’

ĮȞȠȓȤIJȘțİ anixtike open.NACT.PAST.PERFVE.3SG

TP  T VoiceP    Voice vP non-active  suffix pro vǯ external  arg. position v VP  V DP ĮȞȠȓȤIJȘțİ IJȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ anixtike to xrimatokivotio open.NACT.PAST.PERFVE.3SG the.NOM safe.NOM Causative type; the external argument (agent) in the [Spec, vP] position is not phonologically expressed (47) ȐȞȠȚȟĮȞ aniksan open.ACT.PAST.PERFVE.3PL ‘(they) opened the safe’

IJȠ to the.ACC

ȤȡȘȝĮIJȠțȚȕȫIJȚȠ xrimatokivotio safe.ACC

TP  T vP    pro vǯ  v VP   V DP ȐȞȠȚȟĮȞ IJȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ aniksan to xrimatokivotio open.ACT.PAST.PERFVE.3PL the.ACC safe.ACC

194

Chapter Four

Passives in Medieval and Modern Greek bear productively non-active endings. Non-active voice consistently absorbs/blocks the accusative case in Medieval and Modern Greek since its presence makes the introduction of the DP in the accusative case impossible; the Voice head is responsible for the presence of non-active voice in passive types in Medieval Greek. At the same time the presence of the preposition Įʌȩ apo ‘by’ is available, which is licensed by v (v is characterised as [+agent]). I therefore propose the following structure –with presence of the Voice[+agent] head– for the passives of Medieval and Modern Greek: (48)

Ș ıȣȞȑȜİȣıȘ įȚĮȜȪșȘțİ Įʌȩ IJȠȣȢ įȚĮijȦȞȠȪȞIJİȢ i sinelefsi dhialithike apo tus dhiafonundes the.NOM assembly.NOM dissolve.NACT.PAST.PERFVE.3SG by the dissidents ‘the assembly was dissolved by the dissidents’

TP   T VoiceP   Voice vP [+agent ]   non-active PP (Įʌȩ apo) suffix (external argument)

vǯ  Įʌȩ IJȠȣȢ įȚĮijȦȞȠȪȞIJİȢ v VP apo tus diafonundes [+agent]  by the dissidents V DP įȚĮȜȪșȘțİ Ș ıȣȞȑȜİȣıȘ dhialithike i sinelefsi dissolve.NACT.PAST.PERFVE.3SG the.NOM assembly.NOM

In contrast, the passives (cf. tree-diagram 49) in Ancient Greek do not have a Voice head in their structure since in many cases they bear active voice (although agent arguments can be present). The passive type bears active endings whereas the PP denoting the agent is again generated (despite the absence of the Voice head, which I associated only with the absorption of the accusative case and not with the preposition ‘by’ as does Collins for English) in the [Spec, vP] position (as is demanded by the specific verb and licensed by the subcategorisation of the passive type). From the Hellenistic-Roman period onwards, the non-active passive is the stable situation since the Voice[+agent] head is further included in the structure of the passives. The absence of Voice from the passives of Ancient Greek and the presence of Voice in the passives of Medieval Greek is linked to the absorption of the accusative by the non-active types which begins from the Hellenistic-Roman period and is stabilised during the Medieval Greek period. As seen from the above, the loss of the benefactive meaning, which non-active endings in transitive structures bore, and the extension of the accusative as the structural case of the direct object of the active transitive verbs (with the sole exception of the very few deponent verbs, which do not, however, possess both active and non-active forms) leads to the direct connection of the non-active ending with the Voice head (which, as I suggested previously, is connected with the absence of the accusative case and (in Greek) with the non-active voice). (49)

Γϡȱȱ

ΈȂȱΩΏΏΓ΍ȱȱ

σΔ΍ΔΘΓΑȱȱ

hoi d’ alloi epipton the.NOM rest.NOM cut-down.ACT.IMPERF.3PL ‘the rest were cut down by the horsemen’

23

ȱ

Γϡȱȱȱȱȱȱȱȱȱȱȱȱȱ ΈвȱΩΏΏΓ΍ȱȱȱ ΚΉϾ·ΓΑΘΉΖȱȱȱȱȱȱȱȱ σΔ΍ΔΘΓΑǰȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ

ЀΔϲȱ

ϡΔΔνΝΑ23

hypo by

hippeo:n horsemen

ΔΓΏΏΓϠȱȱȱȱȱȱȱ ΐξΑȱȱȱȱȱȱ ЀΔϲȱȱȱȱϡΔΔνΝΑǰȱȱ

oi d’ alloi pheugontes epipton, polloi men hupo hippeo:n, the.NOM rest.NOM flee.PART.NOM cut-down.ACT.IMPERF.3PL many.NOM prt by horsemen ȱ ΔΓΏΏΓϠȱȱ Έξȱȱ ЀΔϲȱȱ ΘЗΑȱȱ ̍ΉΏΘЗΑ polloi de hupo to:n Kelto:n many.NOM prt by the Celts ‘and the rest were cut down as they fled, many by the horsemen and many by the Celts’ (Xenophon, Hellenica, 7, 1, 31; 5-4 BC)

Diachronic Course and Derivation of Transitivity Alternations

195

TP  T vP  PP vǯ (Įʌȩ apo-agent)  ЀΔϲȱϡΔΔνΝΑ v VP hypo hippeo:n [+agent]  by the horsemen V DP σΔ΍ΔΘΓΑȱȱȱȱ ΓϡȱΩΏΏΓ΍ȱ epipton hoi alloi cut-down.ACT.IMPERF.3PL the.NOM rest.NOM In (50-54) I repeat from section 3.1 passive structures with active type from Classical Greek (the numbers in the second set of parentheses refer to the example numbers in the corresponding section):

ȱ

(50) (59a) ΦΔνΌ΅ΑΉȱȱ ЀΔϲȱȱ ̝Ό΋ΑΣΈΉΝȱȱ ΦΑΈΕϲΖȱȱ ̖Ε΋Λ΍ΑϟΓΙȱȱȱ apethane hypo Athe:nadeo: andros Tre:khiniou die.ACT.AOR.3SG by Athenades man Trachinian ‘he was killed by Athenades, a Trachinian’ (Herodotus, The Histories, 7, 213, 2-3; 5 BC) (51) (59b) ΅ЁΘΓϟȱ ·Ήȱ ΦΔνΌΑϙΗΎΓΑȱ ЀΔϲȱȱ ϡΔΔνΝΑȱȱȱȱȱ ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ autoi ge apethne:iskon hypo hippeo:n they.NOM prt die.ACT.IMPERF.3PL by cavalry ‘they were killed by any of the enemy's cavalry’ (Xenophon, Cyropaedia, 7, 1, 48; 5-4 BC) (52) (59c) Γϡȱȱ ΈȂȱΩΏΏΓ΍ȱȱ ΚΉϾ·ΓΑΘΉΖȱ σΔ΍ΔΘΓΑǰȱȱ ΔΓΏΏΓϠȱȱ ΐξΑȱȱ ЀΔϲȱϡΔΔνΝΑ hoi d’ alloi pheugontes epipton, polloi men hypo hippeo:n ȱ the rest.NOM flee.PART.NOM die.ACT.IMPERF.3PL many.NOM prt by horsemen ‘and the rest were cut down as they fled, many by the horsemen’ (Xenophon, Hellenica, 7, 1, 31; 54 BC) (53) (59d) ̝Ε΍΅ΔΉϟΌ΋Ζȱȱ ΐξΑȱȱ ΘΉΏΉΙΘλȱȱȱȱȱȱȱȱ ΈϱΏУȱȱ ЀΔϲȱȱ ̕Δ΅Ε·΅ΔΉϟΌΉΓΖȱ Ariapeithe:s men teleutai dolo:i hypo Spargapeitheos Ariapithes.NOM prt die.ACT.PRES.3SG treachery.DAT by Spargapithes ‘Ariapithes was treacherously killed by Spargapithes’ (Herodotus, The Histories, 4, 78, 8; 5 BC) (54) (59e) ΉϢΖȱΘχΑȱȱΉϡΕΎΘχΑȱȱ πΗΔϟΔΘΉ΍ȱ ЀΔϲȱȱ ΘЗΑȱȱ πΚϱΕΝΑȱȱȱ eis te:n heirkte:n espiptei hypo to:n ephoro:n in the prison throw.ACT.PRES.3SG by the Ephors ‘he is thrown into prison by the Ephors’ (Thucydides, The Peloponnesian War, 1, 131, 2; 5 BC)

(b) Derivation of anticausatives We have ascertained the following in relation to the voice morphology of anticausative alternating types in Greek: (a) two sub-cases are found: active anticausatives (beginning from the Hellenistic period); non-active anticausatives (Classical Greek). (b) voice morphology is not correlated with the derivation of the transitive from the intransitive type or vice versa, or with the distinction of various groups of anticausatives in the same period. The structure that I propose for the anticausatives of Ancient Greek (cf. section 3.1.3) is the structure in (55). The anticausatives bear in Ancient Greek constantly non-active voice. I assume that this is due to the presence higher than v of the functional category Voice[-agent]. Voice has the feature [-agent] since only the presence of Įʌȩ apo ‘by’-Phrase denoting cause or instrument, but not agent, is possible24. I assume that v is essential in all the anticausatives (cf. section 2.1.1 and below 4.3) since it is connected with the meaning of causativity (meaning of change-of-state) which is a basic element of the group of anticausatives. I do not connect v exclusively with the presence of agent (I hypothesise, following Alexiadou & Anagnostopoulou (2004), that there are two types of v: v [+agent] or [-agent]), but with the nominative case of the subject. The fact that v is present in both the causative and anticausative structures in all periods needs to be justified. Its presence in the structure of all the causatives and anticausatives is owing to the connection between causatives and anticausatives in a common causative lexical schema, which I will argue in detail in section 4.3. For the time being, I 24

For the features [+agent], [-agent], cf. Alexiadou & Anagnostopoulou (1999a, 2004); Alexiadou, Anagnostopoulou & Schäfer (2006).

196

Chapter Four

hypothesise that the presence of v is essential for the nominative case of both the subject of causatives (v[+agent]) and that of anticausatives (v[-agent]). In addition, Voice is responsible for the non-active morphology in the anticausative structures. The non-active morpheme also blocks the assignment of accusative case in anticausatives of Ancient Greek. (55)

πΒ΋ΕΣΑΌ΋ȱȱ

ΔΉΈϟΓΑȱȱ ekse:ranthe: pedion dry.PASS.AOR.3SG field.NOM ‘the field dried by drought’

πΒȱȱ

ΦΑΓΐΆΕϟ΅Ζ

eks from

anombrias drought

TP  T VoiceP  Voice vP [-agent]  non-active suffix vǯ  v VP  [-agent]   VP PP  P-cause V DP πΒȱΦΑΓΐΆΕϟ΅Ζ ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱ πΒ΋ΕΣΑΌ΋ȱȱȱȱ ʌİįȓȠȞ eks anombrias ekse:ranthe: pedion from drought dry.PASS.AOR.3SG field.NOM In (56-59) I repeat examples from section 3.1 with non-active anticausatives from Classical Greek the numbers in the second set of parentheses refer to the example numbers in the corresponding section): (56) (81a) Θϲȱȱ ΗЗΐ΅ǯǯǯȱȱ ΓЁΎȱȱ πΐ΅Ε΅ϟΑΉΘΓȱȱ to so:ma... ouk emaraineto the.NOM body.NOM not wither.NACT.IMPERF.3SG ‘the body (meanwhile)... did not waste away’ (Thucydides, The Peloponnesian War, 2, 49; 5 BC) (57) (85a) πΑȱΘΓϾΘΓ΍ΖȱȱȱΘΓϧΖȱȱȱȱȱȱΐ΅Όφΐ΅Η΍ΑȱȱȱοΎΣΗΘΓΙȱϷΕ·΅ΑϱΑȱȱȱȱΘ΍ȱȱȱȱȱȱȱȱȱȱȱȱȱȱΜΙΛϛΖȱȱȱȱȱȱπΎΎ΅Ό΅ϟΕΉΘ΅ϟȱ en toutois tois mathe:masin hekastou organon ti psykhe:s ekkathairetai in these.DAT the.DAT studies.DAT every.GEN instrument.NOM some.NOM soul.GEN cleanse.NACT.PRES.3SG

‘there is in every soul an instrument (of knowledge) that purifies (and kindles afresh) by such studies (when it has been destroyed and blinded by our ordinary pursuits)’ (Plato, Republic, 527d, 8; 5-4 BC) (58) (104b) ΓЁ Έ΍Ύ΅ΗΘφΕ΍Σȱȱ ΘΉȱΎ΅Ϡȱȱ Ϣ΅ΘΕΉϧ΅ȱȱ ΔΓΏΏΤȱȱ ΦΑΓϟ·ΉΘ΅΍ ou dikaste:ria te kai iatreia polla anoigetai not courts.NOM and dispensaries.NOM many.NOM open.NACT.PRES.3SG ‘not many courts of law and dispensaries opened’ (Plato, Republic, 405a, 4; 5-4 BC) (59) (87b) ΘφΎΉΘ΅΍ȱ ·ΤΕȱ πΑȱ΅ЁΘϜȱ Θϲȱ ΚΏν·ΐ΅ȱ te:ketai gar en aute:i to phlegma melt.NACT.PRES.3SG as in her the.NOM phlegm.NOM ‘as the phlegm is melting in her body’ (Hippocrates, Epidemics, i-iii, 2, 1, 1; 5-4 BC)

In parallel with the absorption of the accusative by the non-active voice and the predominance of non-active passives, a tendency for predominance of active anticausatives (as early as the HellenisticRoman period) appears. The structure (in 60) which I propose for the anticausatives of Medieval and Modern Greek (when the extension is completed) does not include Voice. The absence of Voice results in active voice stably for the anticausatives of Medieval and Modern Greek (the anticausatives bear systematically active voice). v is also connected here with causativisation (change-of-state meaning, presence of the ș-role of the patient), which arises from the complex event schema of causative and anticausative types25 (cf. sections 2.1.1, 2.1.2, and 4.3), and is specified as [-agent]. Consequently, in 25

v is present in causatives and anticausatives. It is connected with the presence of the complex subevent for these verbs. That causativity means a complex event schema has been shown in the chapter where I discussed prototypical transitives (causatives) and accusative verbs (para. 2.1.3). The fact that the same event schema

Diachronic Course and Derivation of Transitivity Alternations

197

both causative and anticausative structures and in all periods, v is connected with the causativity feature (with availability of the presence of the agent in causative structures and non-agent in anticausative structures26) and only indirectly with active voice (when Voice is absent), whereas Voice (absent in anticausative structures of Medieval Greek) is responsible for the non-active voice and the absorption of accusative case. (60) πΘΕϱΐ΅ΒΉΑȱȱ ϳȱȱ ȱ etromaksen ho ȱ scare.ACT.PAST.PERFVE.3SG the.NOM ȱ ‘people were scared because of that’

ΎϱΗΐΓΖȱȱ

Έ΍Τȱȱ

ΘΓІΘΓ

kosmos people.NOM

dia from

touto that

TP  T vP    vǯ    v VP [-agent]  VP PP  P-cause V DPȱȱȱȱȱȱȱȱΈ΍ΤȱΘΓІΘΓȱȦ dia touto / from that ȱ πΘΕϱΐ΅ΒΉΑȱȱȱȱ

etromaksen scare.ACT.PAST.PERFVE.3SG

ϳȱΎϱΗΐΓΖȱȱȱȱȱȱ

ho kosmos

ȱ

the.NOM people.NOM

In (61-64) I repeat examples from section 3.3 with active anticausatives from the Medieval Greek period (the numbers in the second set of parentheses refer to the example numbers in the corresponding section): Medieval Greek ψȱ ΜΙΛφȱ ΐΓΙȱȱ πΘΕϱΐ΅ΒΉΑȱ (61) (67a) Ύ΅Ϡȱ ke i psixi mu etromaksen and the soul.NOM my scare.ACT.PAST.PERFVE.3SG ‘and my soul was scared’ (Digenis Akritas, Ǽscorial, 1113; AD 12) (62) (66a) ΐξȱ ΑφΗΘΉ΍ΉΖȱ Ύ΅Ϡȱȱ ΐξȱΈν΋ΗΉΖǰȱȱȱ Ύ΍ȱȱ όΏ΍ΝΗΉΑȱȱ ψȱȱȱȱ Ύ΅ΘΣΕ΅ȱ me nisties ke me dheises, ki iliosen i katara with fasts and with prayers and melt.ACT.PAST.PERFVE.3SG the curse.NOM ‘the curse melted away with fasts and prayers’ (The Sacrifice of Abraham, 1142; AD 16) όΕΛ΍ΗΉΑȱȱ ϳȱȱ ΏΓ·΍ΗΐϱΖȱȱ ȱ ȱ (63) (68a) ΘϱΘΉȱȱ tote irxisen o loghizmos then begin.ACT.PAST.PERFVE.3SG the reasoning.NOM ‘then the reasoning began’ (Leonardo Dellaporta, Poems, ǹ, 69; AD 14-15) (64) (69a) ϵΘ΅Αȱ Δ΅·ЏΗϙȱȱ ϳȱȱ ΔΓΘ΅ΐϱΖ otan paghosi o potamos when freeze.ACT.PERFVE.3SG the river.NOM ‘when the river freezes’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 347, 2; AD 17)

ȉhe question which arises is why the above change happened in the area of voice morphology, in other words why active anticausatives were extended and spread. The answer to this question was supported theoretically in the structural differentiation between anticausatives and passives, which is linked to the need also for morphological differentiation between the anticausative and the passive. The need for differentiation between the voice of the anticausatives and the voice of the passives has to be sought in the principle of economy which governs language change (cf. Roberts & Roussou (2003), who talk about economy in relation to syntactic movements; cf. section 2.2.1): for reasons of economy the presence of two types of Voice [+agent] and [-agent] in the same period is avoided; the presence of two types of Voice occurs from the Hellenistic period onwards (the passive type is constantly non-

26

also applies to anticausatives remains to be proved (see 4.3). Non-agent = cause or instrument.

198

Chapter Four

active, resulting in passive structure with Voice [+agent] at the same time as the structure of the anticausative includes Voice [-agent]). Table 4.2 Overview of anticausatives and passives in the diachrony of Greek Ancient Greek

(already from the end of Ancient Greek) Hellenistic-Roman period

Hellenistic and mainly Medieval Greek

anticausative Voice [-agent]

anticausative Voice [-agent]

anticausative absence of Voice27

passive absence of Voice

passive Voice [+agent]

passive Voice [+agent]

From the above, the relationship between causatives and anticausatives is indirectly evident (which I syntactically represented with the presence of v). Although the tendency is for the dissociation of anticausative and passive (through the loss of one of the two types of Voice), causative and anticausative are connected since they bear two types of the same head (v, +/-agent). The relationship between causative and anticausative types and the exact manner of suppression of the one argument in anticausatives (while there are two positions in the structure) will be analysed in section 4.3.2. Concerning the distinction between two types of anticausatives (anticausatives of Ancient Greek with Voice head and anticausatives of Medieval Greek without Voice head), my approach is very close to that of Alexiadou, Anagnostopoulou & Schäfer (2006), who, on the basis of voice morphology, distinguish in the same synchrony two groups of anticausatives with different structure: (a) (group I) anticausatives which have no additional morphology (they bear active voice morphology) in relation to the corresponding causatives; I assume that the anticausatives of Medieval and Modern Greek belong to this group28: [v [Root]] (b) (group II) anticausatives which have non-active morphology (namely, additional morphology from the causative type); I assume that the anticausatives of classical Ancient Greek belong to group II: [Voice (-external argument -agent) [v [Root]]] To the first group belong verbs of change-of-state from internal cause (ȝİȖĮȜȫȞȦ meghalono ‘to grow’, ĮȞșȓȗȦ anthizo ‘to blossom’) or from unspecified cause, whereas verbs of change-of-state from external cause belong to the second group. According to this approach, the role of voice is as follows: non-active voice marks the demotion of the external cause in the verbs of the second group, which are considered to be derived from a basic transitive type. The verbs of the first group do not include Voice (=absence of external cause), and consequently are not marked with non-active morphology29. 27

28

29

I am going to deal with the anticausatives of Medieval and Modern Greek, which follow the anticausatives of the preceding period and bear non-active voice (as exceptions to the general rule), in the next paragraph. Cross-linguistically, verbs which in certain languages do not possess additional morphology in the anticausative but have additional morphology in the causative also belong to this group: Armenian Group I: Group II:

Causative V+suffix V

cor-ats-nel batsel

Anticausative V V+suffix

coranal bats-v-el

‘to dry’ ‘to open’

Japanese Group I: Group II:

Causative V+suffix V

kawak-as-u war-u

Anticausative V V+suffix

kawak-u war-er-u

‘to dry’ ‘to break’

Korean Group I: Group II:

Causative V+suffix V

mal-li-ta tat-ta

Anticausative V V+suffix

malu-ta tat-hi-ta

‘to dry’ ‘to close’

Cf. the parametrisation of this analysis by Schäfer (2006, 2008): when there is voice morphology, every language individually can use a marked schema to denote the absence of external cause (absence of external cause even when the root demands an external cause); in Modern English, in which no voice morphology exists the situation is different: there is expletive Voice which does not have a semantic content; sich of Modern German is an expletive external argument in the [Spec, VoiceP] position. Its presence is idiosyncratic and depends on how unlikely the spontaneous occurrence of an event is (the only role of the expletive Voice is to allow the absence of the external argument (and that only in languages which have voice morphology -such as Modern German)). Consequently, Schäfer adds a third group of anticausatives: Meaning Syntax Spell-out active: [agent [Voice{D, agent} [V [Root]]]] (active)

Diachronic Course and Derivation of Transitivity Alternations

199

My differentiation in relation to the aforementioned analysis lies in the fact that I do not assume that both structures of anticausatives function in parallel in the same synchrony. Additionally, these two groups of anticausatives cannot cross-linguistically and diachronically be separated in the same synchrony on the basis of the above semantic differentiation: verbs of the second group (external cause) frequently change and do not bear special additional morphology (they become active), not for semantic reasons but from the moment that the system of voices changes gradually and marks the anticausative type with active voice (the non-active marks the passive type). At the same time the diachronic course of the anticausatives does not appear to indicate connection between voice and direction of change in the argument structures: namely, I did not ascertain changes in voice that are correlated with the introduction or suppression of the external cause (or internal cause); the changes were in the same direction of the production of an innovative causative type from an earlier nonalternating intransitive of active or non-active voice. Consequently, we cannot distinguish two different groups of anticausatives30 according to voice morphology and meaning, but I argue that in different synchronies the structure of anticausatives is different and, therefore, the manner of morphological marking is different. Specifically: i In Classical Greek, the verb is marked concerning voice morphology (it bears non-active morphology) but that does not mean identification between Voice and absence of the external argument (cf. above the approach of Collins). The structure of all the anticausatives is: [Voice [v [ V]]]31 i From the Hellenistic Koine and mainly from the Medieval Greek period and later (the change is particularly gradual, and the resistance to ‘regularisation’ and return to the previous situation particularly strong), non-active morphology systematically marks the passive types (the presence of accusative case with non-active type is not possible), and the structure of all anticausative verbs is the following: [v [V]] This change is owing to the presence of Voice[+agent] in the passive structure from the HellenisticRoman Koine period and onwards. I assume that for reasons of economy the anticausatives gradually lose the Voice[-agent] from their structure. The presence of two sub-types of Voice (Voice[+agent] for passives and Voice[-agent] for anticausatives) is uneconomical and leads to the loss of Voice for anticausatives and to their marking with active endings. To summarise, morphological instability and variation32 are not a result of the process of the

30

31

32

passive: [Voice{agent} [V [Root]]] (passive) anticausative-I: [V [Root]] (without marking) anticausative-II: [Voice{Ø} [V [Root]]] (passive, clitic-si) anticausative-III: [Expl. [Voice{D, Ø } [V [Root]]]] (sich) A problem for the “morphocentric approaches” of denotation of the derivation direction is how to explain examples such as the following from Japanese: koware-ru/kow-as-u (‘to break’). Both the intransitive and the transitive are morphologically derived, the intransitive with addition of the derivational passive morpheme (r)are, the transitive with the addition of the morpheme -(s)as-, an allomorph of the derivational causative morpheme -(s)ase (Jacobsen 1992). There is no simple type which is not morphologically derived from the root kow-. As an alternative interpretation the following may be proposed: absence of Voice for anticausatives as well (non-active anticausatives) (cf. Theophanopoulou-Kontou 2004), whereas non-active voice in classical Ancient Greek would mark the movement of internal arguments in the subject position, and would mark all the unaccusative verbs which demand movement of the internal argument from the [Compl, VP] position to the subject position. In relation to the role of voice in transitivity alternations, studies of language acquisition also show that children who acquire languages with special voice morphology have particular problems with voice but not with transitive uses. Morikawa (1996) observes that children do not make errors with lexical causatives, but make errors with morphological causatives (periphrastic structures of causatives) where they omit causative morphemes: (1) a. b.

Taro-ga naku Taro cries Hanako-ga Taro-o Hanako-NOM Taro-ACC ‘Hanako makes Taro cry’

nak-ase-ru make-cry

Again, although it appears that children have acquired in the early stages the meaning of causativity, they learn to mark causativity morphologically later. Moreover, Pye (1991, 1994) ascertains in K’iche Maya the contrast between the exact acquisition of the syntactic behaviour of the verbs and the problems of acquisition of morphological marking of transitivity: errors in transitive use of exclusively intransitive verbs are not observed, but quite a few cases of omission of causative morphemes are ascertained, or overgeneralisation of the causative morpheme where it is not needed. In the acquisition of Inuktikut, the errors are also unidirectional (Allen 1996). All the errors concern the transitive use of intransitive verbs, and all could have been avoided if the morpheme for the periphrastic causative verbs had been added. Incorrect causativisations

200

Chapter Four

creation of transitivity alternations (causativisation or anticausativisation from basic anticausative or causative corresponding type) but are concerned with the morphological marking of anticausative verbs which already participated in transitive and intransitive structures. According to a general linguistic principle which I hypothesise is in force, the presence in every synchrony of only one type of Voice (which is responsible for non-active voice in Greek) is preferred to the presence of two types of Voice ([+agent]: passive and [-agent]: anticausative). The separation begins from the Hellenistic-Roman period and is strengthened during the Medieval Greek period. According to the overall picture that arises, in the same period, it is possible to have: (a) presence of Voice [+agent] in the structure of passives (presence of non-active voice) but not in the structure of anticausatives (absence of non-active voice) (for example, Medieval Greek); (b) presence of Voice [-agent] in the structure of anticausatives (presence of non-active voice) but not in the structure of passives (absence of non-active voice) (for example, Ancient Greek). The diachronic changes in the area of voice are due to a general linguistic principle which dictates the non-availability of the presence in the same period of two types of Voice ([+agent] and [-agent]) for reasons of economy (during the Hellenistic-Roman period, Voice (nonactive endings and absence of accusative case) was present in both passives and anticausatives).

4.2.3 The feature [non-act] Anticausatives in Greek and mainly in the Medieval Greek period change into actives and non-actives. I attributed this change to the loss of Voice[-agent] from their structure, which is due to the emergence of Voice[+agent] in the passives of that period (when non-active voice was furthermore stable for the passives after the loss of the benefactive meaning). Concerning the anticausatives of Medieval and Modern Greek, which retain non-active voice, I will follow and modify Embick’s (2000) approach for the deponent transitive verbs of Latin. Embick, examining the presence of analytical and synthetic types of the present perfect in Latin, argues that the emergence of analytical and synthetic types is defined from the presence of the feature [pass(ive)] in certain verbs. The feature [pass] (a) is systematically linked with passivisation, and (b) is included inherently in certain verb roots33 for reasons that have nothing to do with passive syntax. For Embick, many of the deponent transitive verbs, which are a sub-group of transitive verbs, should be dealt with as verbs which are inherently specified with this feature which is responsible for their morphological declension. Some verbs have this feature (to be consistent with the terms that are used

and errors in voice morphology are also found in the data of Turkish child language which were examined by Aksu-Koç & Slobin (1985: 848): -causativisation of intransitive types: (2) *Su-nu kalk this-ACC arise.IMP.2SG ‘raise this’

-overgeneralisation of the causative morpheme, even with verbs which are lexical causative: (3) *kes-tir-di-m cut-CAUS-PAST-1SG ‘I cut it’

Errors in transitivity and voice morphology in Spanish child language are reported by Lȩpez Ornat (1994). The transitivity errors are very few: -only two examples of transitive use of intransitive non-alternating verbs: (4) *te voy a you will-1SG to ‘I am going to kill you’

morir die

(use of morir ‘to die’ instead of the lexical suppletive matar ‘to kill’)

33

Even though children use the clitic se from an early stage, the incorrect omission of se with the verbs caer ‘to fall’ and romper ‘to break’ is observed, and its use with intransitive verbs which in adult language are not formed with se: comer ‘to eat’ and Ilorar ‘to cry’. Embick’s roots correspond to the idiosyncratic part (the part that differentiates the specific verbs from the verb group to which they belong) of the semantic decomposition of Levin. Moreover, as Embick himself states, the hypothesis of the roots begins from the verb groups and their analysis by Levin (1993): these abstract roots belong to specific semantic groups of Levin’s type of groups (1993). ¥DESTROY denotes change-of-state from external cause and implies an agent, whereas ¥GROW denotes change-of-state from internal cause and does not imply the presence of an agent. The term ‘root’ was first proposed by Pesetsky and was later also adopted by Levin herself (who began this discussion of idiosyncratic features but did not initially name them ‘roots’).

Diachronic Course and Derivation of Transitivity Alternations

201

throughout this study34, I will name this feature [non-act] for reasons that have nothing to do with passive syntax or relevant verb meanings which refer to the entire verb group. (a) Anticausatives with the [non-act] feature In endeavouring to analyse the diachronic development of the voice of transitivity alternations and the role of voice in the derivation of transitivity alternations, I assume that the same also holds true for the group of anticausatives which did not change into actives (and remained non-actives until today): the anticausatives which did not change into actives inherently have (in the verb root) the [non-act] feature. The feature is stable diachronically as it belongs and corresponds to the root (to the idiosyncratic part) of the verb and not to the common part of the verb with the remaining verbs of the same group (event schema of change-of-state verbs; cf. section 2.1.2). If the verbs of this group (non-active anticausatives of Medieval and Modern Greek -which are found to be in contrast to the general tendency from the Hellenistic, and mainly from the Medieval Greek period, for active anticausatives) appear with nonactive endings for systematic syntactic reasons, we would expect them to form a unified group, with particular semantic and syntactic behaviour, but this is not the case. Embick’s arguments for the presence of the [non-act] feature in the deponents of Latin are as follows: (i) The transitive ([non-act]) deponent verbs have the same syntax as the other non-deponent transitive verbs: Subject-Nominative + Verb + Object-Accusative; (ii) There is no additional or special meaning in the deponent verbs that does not exist in the nondeponent verbs. Deponent and non-deponent verbs can be connected in paratactic constructions (cf., for example, the verbs consector ‘I chase’ and aggredior ‘I attack’ which are found in Latin texts in paratactic constructions with corresponding active transitives); (iii) The deponents resemble psych-verbs with derived subject and passive morphology (see Belletti & Rizzi 1988); we would expect them to have the same meaning and syntax as the psych-verbs, but this is not the case; (iv) Some deponent verbs can be passivised (e.g., hortor ‘I encourage, I incite’); (v) The nominalisation with agentive meaning of deponent verbs (like the nominalisation of transitive verbs) is possible. The same arguments can also apply to the non-active anticausatives, from the moment that their presence is not productive from Hellenistic-Roman Koine, and mainly from the Medieval Greek period and afterwards: (a) the anticausative non-active ([non-act]) verbs behave syntactically like the remaining anticausatives: Subject-Nominative (patient) + Verb, (b) there is no reason in relation to their lexical meaning for these verbs to be separated (in a systematic manner and in all cases) from the remaining active anticausatives (cf. the diachronic parallel use of anticausative verbs with active and non-active voice morphology), (c) some of the anticausatives can also have passive syntax; they can, in other words, be passivised (IJȠ ȤȡȘȝĮIJȠțȚȕȫIJȚȠ ĮȞȠȓȤIJȘțİ Įʌȩ IJȠȞ ȜȘıIJȒ / to xrimatokivotio anixtike apo ton listi / the.NOM safe.NOM open.NACT.PAST.3SG by the burglar / ‘the safe was opened by the burglar’), (d) and the non-active anticausatives are nominalised in the same way as the active anticausatives: (65)

țĮȜȣIJİȡİȪȦ kaliterevo improve.ACT ȕİȜIJȚȫȞȠȝĮȚ veltionome ameliorate.NACT

ĺ ĺ

țĮȜȣIJȑȡİȣıȘ kaliterefsi ‘improvement’ ȕİȜIJȓȦıȘ veltiosi ‘amelioration’

Consequently, the deponent verbs and the non-active anticausatives of Medieval and Modern Greek bear the [non-act] feature, which is simply a feature of their verb root and does not correlate systematically with passive syntax. It is similar to the characteristics of conjugation that are basic to the correct morphological formation. The difference between the characteristics of the conjugation and the [non-act] feature is that the [non-act] feature is (a) sometimes inherently assigned to the verb and (b) other times is systematically correlated with passive syntax (and Voice). (b) The two types of non-active voice morphology From the above it appears that the area of voice is characterised by relations in only one direction: from the Medieval Greek period, the presence of the agent by-phrase entails the presence of the [non-act] feature, in other words, non-active voice morphology, but the opposite does not always hold true: the presence of the [non-act] feature does not entail presence of a by-Phrase denoting an agent, as there are deponent transitives and non-active anticausative verbs. Consequently, I assume that there are two non34

For the reasons behind the choice of these specific terms, cf. 1.5.

202

Chapter Four

active morphemes: (a) a non-active morpheme which is derived in syntax, in the functional category of Voice, and (b) a non-active morpheme which is lexical. Travis (2000), examining languages with causative morphemes, reaches a similar conclusion: she finds two causative morphemes; the one functions in the syntax and the other belongs to the lexicon and is concerned with specific verbs (the causative morphemes of the lexicon and the syntax in certain languages are also distinguished phonologically). From the above it would appear, of course, that my analysis differs from Embick’s (1998) analysis, who deals with the anticausatives of Modern Greek in the same way as with the passives, not giving much weight to the active anticausatives (for a critical review, cf. Theophanopoulou-Kontou (2004) and ǹlexiadou & ǹnagnostopoulou (2004)). For Embick, passives and anticausatives share the common property of absence of external argument; their difference is that the structure of passives has an agentive v, whereas the structure of anticausatives has a non-agentive v. This analysis does not hold true either for Medieval and Modern Greek (passives: non-actives, anticausatives: change towards marking with active voice) or for Classical Greek (passives: frequently with active morphology, anticausatives: systematically non-actives). According to my approach the [non-act] feature has syntactic consequences and is interpretable when it appears in Voice, whereas it is not interpretable when it appears arbitrarily with deponent verbs and with non-active anticausative verbs35. When the [non-act] feature is generated in Voice, it affects Merge, and no external argument can be generated; on the other hand, the verb root includes the [nonact] feature if the verb root is deponent transitive (according to Embick) or anticausative non-active (according to my approach). According to this analysis, the presence of [non-act] in the verb root is not concerned with and does not affect syntactic behaviour. v can have an external argument even if the [non-act] appears in the root36. In (66-73) I repeat from section 3.3 examples with non-active anticausatives (verbs with the [nonact] feature in their root) from Medieval Greek (the numbers in the second set of parentheses refer to the example numbers in the corresponding section): Medieval (and Early Modern) Greek (66) (34a) ΅ЁΘχȱȱ Ύ΅ϠȱΤΑȱȱ πΐ΅Ε΅ϟΑΉΘΓΑǰȱȱ afti ke an emareneton, she.NOM even if

σΏ΅ΐΔΉΑȱȱ

ȱБΖȱϳȱόΏ΍ΓΖ

elamben

os o ilios

wither.NACT.PAST.IMPRFVE.3SG shine.ACT.PAST.IMPRFVE.3SG like the sun

‘even if she was withering, she was shining like the sun’ (Digenis Akritas, Ǽscorial, 183;AD 12) πΐ΅ΕΣΑΌ΋Αǯȱ (67) (34b) ǻψȱΜΙΛφǼȱȱ Δ΅ΕΤȱΎ΅΍ΕϲΑȱȱ π·φΕ΅ΗΉΑǰȱȱ πΜϾ·΋Αǰȱȱ (i psixi) para keron ejirasen, epsijin, emaranthin. (the soul) with time got-old.3SG got-cold.3SG wither.NACT.PAST.PERFVE.3SG ‘(the soul) got old, cold and withered with time’ (Michael Glykas, 203; AD 12) (68) (34c) ǻΎΣΏΏΓΖǼȱȱ ΔЗΖȱȱ πΐ΅ΕΣΑΌ΋ȱȱ ΔΕϲȱȱΎ΅΍ΕΓІȱ ȱΦΔϲȱΎΏ΅ΙΌΐΓІȱΎ΅ϠȱΔϱΑΓΙ (kalos) pos emaranthi pro keru apo klafthmu ke ponu (beauty) how wither.NACT.PAST.PERFVE.3SG before time from cry and pain ‘how the beauty withered before time from cries and pain’ (Kallimachos and Chrysorroi, 1591; AD 14) (69) (36a) Ύ΅ϠȱΔ΅Ε΅ΛΕϛΐ΅ȱȱȱȱπΒ΋ΕΣΑΌ΋ȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱȱψȱȱΘΓ΍΅ϾΘ΋ȱȱΎ΅ϠȱȱΘ΋Ώ΍Ύ΅ϾΘ΋ȱȱΌΉΓΈЏΕ΋ΘΓΖȱȱȱΔ΋·φȱȱ ke paraxrima eksiranthi i tiafti ke tilikafti theodhoritos piji and immediately

dry.NACT.PAST.PERFVE.3SG the that

and of-such-age given-by-God spring.NOM

‘and that spring of such age, given by God, immediately dried’ (John Moschos, Leimonarion (The Spiritual Meadow), LXXX, D, 11; AD 6-7) ΅ЁΘΤȱȱ ΘΤȱȱ ΚϾΏΏ΅ (70) (36c) ϵΘ΅Αȱȱ Β΋Ε΅ΑΌΓІΑȱȱ otan ksiranthun afta ta fila when dry.NACT.PERFVE.3PL these.NOM the.NOM leaves.NOM ‘when these leaves dry’ (Agapios Landos, Geoponikon, 6, 5; AD 17) ΐΓΙȱȱ ΗΙΐΐ΅ΊЏΑΉΘ΅΍ȱȱ Ύ΍ȱȱ ώΕΌ΅Αȱȱ ΘΤȱȱЂΗΘΉΕ΅ȱȱ ΐΓΙȱ (71) (47b) ψȱȱ Μφȱȱ i psi mu simazonete ki irthan ta istera mu the soul.NOM my gather.NACT.PRES.3SG and came.3SG the last-years.NOM my 35

36

In relation to the synthetic perfects of deponents in Latin, Embick deals with the [non-act] feature as a syntactic characteristic within an approach which accepts roots in syntax. The presence of the [non-act] feature in the root, and consequently in the passivisation process since the root is moved to v, allows Embick to interpret the blocking of the movement of Aspect to Tense with deponents in the derivation of synthetic perfects (something which results in the derivation of only analytical perfects in the case of deponents). An extension to this approach in the area of nouns was attempted by Alexiadou (2001), who suggested that the different forms of nouns reflect differences in the number and type of functional projections in which the abstract roots are introduced.

Diachronic Course and Derivation of Transitivity Alternations

203

‘my soul gathers itself and the last years of my life have come’ (The Sacrifice of Abraham, 192; AD 16) (72) (47a) Ύ΅Ϡȱ ΑΤȱ ΐ΅ΊΓΎΘΓІΑȱȱ ϵΏΏ΅ȱ ΘΓІΘ΅ȱȱ ǻΈ΋ΑνΕ΍΅ǼȱΉϢΖȱȱΘΤȱȱΛνΕ΍΅ȱ ΘΝΖȱ ke na mazoktun ola tuta (dhineria) is ta herja tus and to gather.NACT.PERFVE.3PL all.NOM these.NOM (dinars) in the hands their ‘so that all these (dinars) are gathered in their hands’ (M. Grigoropoulos, notary of Chandaka, 52, 120; AD 16) (73) (47c)ȱȱ Ύ΅Ϡȱȱ Έ΍Ήΐ΅ΊЏΛΌ΋Η΅Αȱȱ οΒ΅ΎΓΗϟ΅΍Ζȱȱ Λ΍Ώ΍ΣΈΉΖȱȱ ΩΑΌΕΝΔΓ΍ȱ ΘΓІȱȱȱ ΔΓΏνΐΓΙ ke dhiemazoxthisan eksakosies xiliadhes anthropi tu polemu  and gather.NACT.PAST.PERFVE.3PL six-hundred thousand persons the war.GEN ‘six-hundred thousand persons of wargathered’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 350, 30-31; AD 17)

Embick characterises the system of Latin as unique in terms of the presence of the [non-act] feature and of the deponents that bear inherently (in certain roots) a characteristic which is connected systematically with syntax. Embick’s analysis is radical because it overturns a model of syntax and morphology interaction which disallows roots to be connected inherently with syntactic features. Embick’s proposal for Latin deponents, which comprise a case where the features which are connected with the roots are not only morphological, appears from the above to allow for the analysis of other (besides deponents) cross-linguistic phenomena37. For Greek diachrony, we are able to hypothesise, moreover, a link between the change of transitive middle verbs (denoting the benefactive meaning) of classical Ancient Greek into deponent transitives of Modern Greek (without semantic contribution of the non-active ending; the non-active feature is in the verb root, as in Latin deponents) and the change of non-active anticausatives of classical Ancient Greek (the majority of anticausatives bore non-active voice) only into a small group of non-active anticausatives of Modern Greek (the non-active feature is in the verb root in contrast to the majority of active anticausatives). The overall picture in relation to the voice of the anticausative alternating types and the [non-act] feature of the root is as follows: (a) The non-active anticausatives of Medieval Greek are marked with the [non-act] feature in their root. The feature in the root corresponds every time to the morphological marking of the initially exclusively (non-alternating) intransitive type which, further, is causativised and becomes an alternating (transitive and intransitive) type. (b) Following this approach, Voice comprises in many cases a feature of the verb root (the verbs with [non-act] feature in the root comprise the corresponding intransitive category of transitive deponents), and as a feature of the verb root it corresponds to the initial marking of voice of the nonalternating intransitive which is later causativised. The non-active anticausatives of Medieval and Modern Greek have the [non-act] feature in their root (as do the deponent transitives), but do not have a Voice head (as do the passive types). Consequently, concerning the distinction between the two groups of anticausatives which are recognised by Alexiadou, Anagnostopoulou & Schäfer (2006), my approach is differentiated, as, according to the analysis of Alexiadou, Anagnostopoulou & Schäfer, Voice is present in both the nonactive anticausatives and the passives, with the result that the differences between anticausatives and passives are not taken into consideration (cf. the counter-arguments of Theophanopoulou-Kontou 2004)38. In contrast, according to my analysis, the presence of non-active voice morphology is owing to a different reason in every case: (a) in classical Ancient Greek, it is due to the presence of the Voice head in anticausatives (but not in passives, which are not systematically non-active but can also be active); (b) in Medieval Greek, it is mainly due to the presence of the [non-act] feature in the root of some anticausatives, but also to the presence of Voice only in passives (with stable presence of an external argument (overtly expressed or not) according to the modification of Collins’s approach).

37

38

Cf. Baker (2003: 280): ‘there is a residue of morphology that seems to have nothing to do with syntax’. Cf. also Embick’s analysis (1996) for the alternation of an overt and a null causative morpheme in Hupa: the causative morpheme is dealt with as the head of a verb projection, but the variation which appears is compared with the variation of past morphemes (-0, -d, -t) in English. Embick claims that it should be characterised as a property of the roots of these specific verbs whether the causative head is realised as ł (overt causative morpheme) or as -0 in the transitive type. This alternation of allomorphs, however, only appears in transitivity alternations; in other cases, there is no alternation between -0 and -ł. Moreover, the Salish languages comprise a counter-argument; they generally allow for verbs denoting changeof-state from an external cause to be expressed with the first structure of anticausatives, without a Voice head.

204

Chapter Four

4.3 Derivation of Transitivity Alternations: the Basic Causative Schema 4.3.1 Basic schema of alternating verbs: previous analyses A vital question for every study of transitivity is the derivation of causative and anticausative type and the relationship (or the absence of relationship) between the two types in the lexicon; an answer, in other words, is sought in relation to (a) if the verbs begin in the lexicon as causatives or anticausatives, or some as causatives and others as anticausatives; and (b) if causatives and anticausatives are connected in any way in the lexicon. Diachronic data and the analysis that I undertook are directly linked to the theoretical issue of the basic type (transitive or intransitive from which the other is derived). In analyses of Modern Greek until today, the direction of derivation has been linked to voice morphology. For Theophanopoulou-Kontou (1999b, 2004), anticausative morphology is connected with the structure of anticausatives which is a structure of VP (and not vP)39. In the case of the active anticausatives (țȠțțȚȞȓȗȦ kokinizo ‘to redden’, ĮȡȡȦıIJĮȓȞȦ arosteno ‘to fall sick’), the intransitive structure is the basic since it is accepted by all native speakers (in contrast to the transitive structure which is accepted by only some native speakers). The presence of the causative member of the alternation is the result of transitivisation with the addition of an external agent or cause/instrument in the external argument position (Theophanopoulou-Kontou 2001: 55-56, 2002b). In the case of nonactive anticausative verbs, voice morphology is connected with the characteristics (a) of the totally affected subject (IJȠ IJȡĮʌİȗȠȝȐȞIJȘȜȠ ȜİȡȫșȘțİ / to trapezomandilo lerothike / the.NOM tablecloth.NOM dirty.NACT.PAST.PERFVE.3SG / ‘the tablecloth got dirty’ - ȝʌȜȑȤIJȘțİ IJȠ ʌȩįȚ ȝȠȣ ıIJȠ ȤĮȜȓ / blextike to podhi mu sto xali / tangle.NACT.PAST.PERFVE.3SG the.NOM foot.NOM my tothe carpet / ‘my foot was tangled in the carpet’), (b) of the subject involvement in the verb action/state (ıIJİȞȠȤȦȡȚȑȝĮȚ stenoxorieme ‘to be distressed’, ıȣȖțȚȞȠȪȝĮȚ singinume ‘to be moved’). Consequently, for Theophanopoulou-Kontou, non-active morphology in anticausatives does not affect syntax but is connected with the predicate meaning and with purely lexical factors. For ǹlexiadou & ǹnagnostopoulou (1999, 2004), the morphological variation in Modern Greek is correlated with the structural heterogeneity of anticausative verbs: the anticausatives do not have a unified structure: (a) BECOME + predicate (deadjectival40) [țȠțțȚȞȓȗȦ kokinizo ‘to redden’, ĮįİȚȐȗȦ adjazo ‘to empty’, ĮıʌȡȓȗȦ asprizo ‘to whiten’]; (b) RESULT + Voice (Voice = agent/+manner), [IJıĮȜĮțȫȞȠȝĮȚ tsalakonome ‘to wrinkle’, ȜİȡȫȞȠȝĮȚ leronome ‘to dirty’]; (c) BECOME + possessive structure [IJıĮȜĮțȫȞȦ tsalakono ‘to wrinkle’, ȜİȡȫȞȦ lerono ‘to dirty’]. In the case of non-active anticausatives, the voice morphology marks the process of de-transitivisation (the same also applies to reflexive and passive verbs); in the case of active anticausatives, the transitive type is the result of transitivisation. No structure, however, whether the anticausative or the causative, is derived from the other (there is no directionality in transitivity alternations). ǹlexiadou, Anagnostopoulou & Schäfer (2006) base their analysis on the syntactic decomposition of change-of-state verbs in the components of VOICE and CAUSE: the structure of causatives, anticausatives and passives is [VOICE [CAUSE [ROOT]]]. The differences, however, between the passives and the anticausatives (cf. Theophanopoulou-Kontou 2004) and the role of voice morphology (in passives: absorption/blocking of accusative case, in anticausatives: relation of voice morphology to the initial morphological marking of the intransitive/anticausative verb) cannot support the analysis of a parallel structure for passives and anticausatives. The majority of theoretical proposals (a particularly high number; every attempt at analysis of the phenomenon also endeavours to solve this problem) that have been posited for the derivation and the relation between causative and anticausative types can be classified into three large groups: (a) According to the first group of approaches, the lexical semantic representations of anticausative and causative types differ from language to language and from predicate to predicate, so that they reflect the primary (non-derived with additional (causative or non-causative) morpheme) morphological marking (cf. mainly Haspelmath (1993) and, for Modern Greek, Clairis & Babiniotis (1999)). The central idea behind this is that the morphological direction of derivation (causative and anticausative morphology), in a language and cross-linguistically, is indicative of how causative and anticausative change-of-state predicates are connected. According to this approach, a verb such as ‘to freeze’ has anticausative semantic representation in Arabic (where the causative structure is 39

40

The difference between anticausatives and passives (absence vs. presence of an agent) has to do with the absence/presence of a functional head for the agent (v), and, therefore, with the absence/presence of the external ș-role (agent) which is introduced into the structure by the v head. Cf. Stavrou 1996, 1999.

Diachronic Course and Derivation of Transitivity Alternations

205

morphologically marked) but causative semantic representation in Swahili (where the anticausative structure is morphologically marked)41. The problem, indeed, with approaches such as this is that having as their aim the preservation of the correspondence between voice morphology and lexical semantic representation, they deal with the semantic structure of synonymous verbs differently in different languages. This results in theoretical problems which are concerned with language acquisition and the hypothesis of the uniformity of the lexicon (lexicon uniformity principle; Reinhart 1997), which requires a simpler and more regular theory of lexicon. Additionally, according to what I have presented in the previous paragraph, we cannot connect morphological derivation with the derivation of the causative or anticausative type from an anticausative or causative type respectively. (b) The second group of analyses supports the presence of a common anticausative lexical semantic representation for both causative and anticausative verbs42 (Lakoff 1970; McCawley 1979; Dowty 1979; Williams 1981; Pesetsky 1995). This approach is based particularly on the structures where the causative verb is morphologically marked (but faces problems in the interpretation of structures with morphologically marked anticausative verbs). Arguments against the underlying anticausative structure are as follows: (i) no interpretation is given of why verbs such as ‘to cut’ do not have an anticausative type but have only a passive type, whereas verbs such as ‘to open’ have both anticausative and passive type; (ii) for certain verbs, there are selectional restrictions on intransitive uses, but these restrictions do not apply to transitive uses. (74)

a. ȀĮșȐȡȚıİ IJȘȞ ĮȣȜȒ Katharise tin avli clean.ACT.PAST.PERFVE.3SG the.ACC yard.ACC ‘s/he cleaned the yard’ b. ???Ǿ ĮȣȜȒ țĮșȐȡȚıİ ???I avli katharise the.NOM yard.NOM clean.ACT.PAST.PERFVE.3SG ‘the yard was cleaned’

(c) The third solution is to accept a causative representation for both causative and anticausative structures. The lexical process of the semantic anticausativisation suppresses/saturates/blocks the expression of the initial (argument which causes the change or) event. For example, both the causative 41

42

Nuñon’s semantic analysis is also close to this approach: both causative and anticausative verbs are derived from a third source. The idea is again to analyse the structures according to morphology but not in absolute relation to the surface morphology. Anticausative morphology is connected to the derivation of the intransitive type from an alternating verb stem. The difference between languages consequently lies in the explicit marking of the derivation of the causative and the anticausative by the alternating verb stem. The restriction on the derivation of causative verbs is the availability of presence of agent for the types of events which are denoted by causative verbs, whereas the restriction on anticausative verbs is the non-denotation of change-of-state necessarily arising from the agent for the types of events which are denoted by anticausative verbs. This theory interprets the absence of causative structures for anticausative non-alternating verbs and anticausative structures for causative non-alternating verbs. In the case of ıʌȐȦ IJȠȞ ȩȡțȠ ıȚȦʌȒȢ / spao ton orko siopis / break.ACT.PRES.1SG the.ACC oath.ACC silence.GEN / ‘I break the oath of silence’ – *Ƞ ȩȡțȠȢ ıȚȦʌȒȢ ȑıʌĮıİ / *o orkos siopis espase / the.NOM oath.NOM silence.GEN break.ACT.PAST.PERFVE.3SG / ‘the oath of silence broke’, Nuñon recognises two meanings: according to the first meaning of the breaking of specific objects, an agent is not required; according to the second meaning of the abstract breaking, an agent is required (obligatory agent ĺ only causative (non-alternating) type). Cf. also Parson’s approach (1990: chapter 6), according to which both the causative and the anticausative verbs are derived from related stative adjectives. Vcausative Adjstat. % ' Vanticausative The only difference from Nuñon’s model is the different interpretation of cross-linguistic morphological data. According to this analysis, (anti)causative morphology has a double role: (a) the change of category (adjective ĺ verb); (b) the morphological representation of causativity (CAUSE + BECOME) or anticausativity (BECOME). Moreover, according to this analysis, the change-of-state predicate should also be present both in the derivation of the causative type from a stative adjective and in the derivation of the anticausative type from a stative adjective. It should be noted that the common representation according to this approach concerns all the verbs which participate in the alternation and not only one group of verbs (approaches which separate the verbs which participate in transitivity alternations and their derivation are part of the analyses in (a) above).

206

Chapter Four

țȜİȓȞȦ klino ‘to close’ and the anticausative țȜİȓȞȦ klino ‘to close’ have the semantic representation of a dyadic causative verb (Levin & Rappaport Hovav 1995: 108): Causative țȜİȓȞȦ klino close.ACT Lexical semantic representation [[x does something] causes [y becomes CLOSED]] Mapping rules Ļ Ļ Argument structure x

Anticausative țȜİȓȞȦ klino close.ACT Lexical semantic representation [[x does something] causes [y becomes CLOSED]] Mapping rules p p Argument structure ‡

The causative analysis must concern all the unaccusative verbs, as has been maintained by Chierchia (1989/2004) and Pustejovsky (1995): an unaccusative verb, for example, ȑȡȤȠȝĮȚ erxome ‘to come’, which does not have causative use, is correlated with a causative verb meaning, such as the meaning of ijȑȡȞȦ ferno ‘to bring’, but this causative meaning is either not lexicalised or is lexicalised by a verb which is not morphologically correlated with the intransitive use. According to this approach, the fact that verbs from all the semantic sub-groups of unaccusatives are causativised in the diachrony of the Greek language, or the fact that in the dialect St’át’imcets43 all the unaccusatives are alternating in contrast to Modern English and to the diachrony of the English language, are owing to lexical idiosyncratic reasons and differences between the specific languages and periods44. Concerning the non-alternating causatives, Reinhart (1997) analyses the causative verbs that do not have an unaccusative type in relation to a restriction in the reduction of ș-roles: a ș-role which is characterised as [+mental state] cannot be removed from the verb structure, and the prediction is exact since the study of diachrony shows that we do not have anticausativisations for causative nonalternating verbs (therefore, the process of reduction of an element [+mental state] is impossible).

4.3.2 The proposed analysis In this present section I present my basic arguments in favour of the connection between causative and anticausative types through a common lexical semantic representation. Arguments in favour of basic causative and against anticausative lexical semantic representation comprise the following (Chierchia 1989/ 2004; Levin & Rappaport Hovav 1995; Pustejovsky 1995; Reinhart 1997)45: 43

44

45

Predicates such as ‘punch’, ‘whip’, ‘build’ (non-alternating causatives denoting actions which cannot take place without the intervention of an external agent, in languages such as Greek, English or Italian) have unaccusative types in St’át’imcets (a Salish dialect; Southwest British Columbia); Davis & Demirdache (1995). St’át’imcets is theoretically very important as all the verb roots in this dialect are unaccusative; the unergative and transitive verbs are always morphologically derived with the addition of an (in)transitivised suffix. Levin & Rappaport Hovav do not agree with this analysis: the unaccusative verbs which are not alternating mean appearance (“appearance verbs”); the absence of meaning of external cause in these verbs interprets the cross-linguistic absence of transitivity alternation for the verbs denoting appearance. This prediction, however, does not apply to St’át’imcets, or to the Greek data since the verbs ‘appear’ and ‘disappear’ form transitivity alternations with the meaning ‘x made y appear/become visible’, ‘x made y disappear’. I do not agree with the counter-arguments to this approach posited by Alexiadou, Anagnostopoulou & Schäfer (2006) and Alexiadou (2006a): (a) presence of causative morphology. According, however, to the analysis I am following for voice (cf. the previous paragraph), the presence of voice morphology cannot result in either an argument or counter-argument for any approach. (b) the unaccusatives denoting change-of-state which are not alternating (ijșİȓȡȠȝĮȚ fthirome ‘to decay’, ĮȞșȫ antho ‘to blossom’, İȟİȜȓııȠȝĮȚ ekselisome ‘to evolve’) cannot be interpreted as the result of a lexical anticausativisation. They describe a change-of-state which arises from an internal cause. Verbs of this semantic group, however, are alternating in other languages (Salish; cf. Davis & Demirdache 1995) and other periods of the same language. (c) the difference between anticausatives and passives in Greek, English and German cannot be interpreted on the basis of the presence in the passives and the absence in the anticausatives of an argument: the passive type of Modern Greek does not permit the presence of a PP denoting cause, but only agent or instrument, in contrast to the periphrastic passive types of English and German which permit the presence of an agent, instrument or cause-PP. The anticausative types of Modern Greek can accept cause or instrument-PP, in contrast to the anticausatives of English and German which only accept cause. The following examples from Modern Greek are passive, however, and the presence of agent or cause is possible:

Diachronic Course and Derivation of Transitivity Alternations

207

(a) The anticausative verbs have more selectional restrictions than causative verbs: ȅ īȚȐȞȞȘȢ țĮșȐȡȚıİ IJȠ o Janis katharise to the.NOM Janis.NOM clean.ACT.PAST.PERFVE.3SG the.ACC ‘Janis cleaned the room’ anticausative: ???ȉȠ įȦȝȐIJȚȠ țĮșȐȡȚıİ46 ???to dhomatio katharise the.NOM room.NOM clean.ACT.PAST.PERFVE.3SG ‘The room was cleaned’

(75) causative:

įȦȝȐIJȚȠ dhomatio room.ACC

Furthermore, the causative non-alternating verbs (such as ‘cut’), which do not form an anticausative type, allow as their subject only agents or agents and instruments, but do not allow the presence of

(1) a.

Ǿ ʌȩȜȘ țĮIJĮıIJȡȐijȘțİ Įʌȩ i poli katastrafike apo the.NOM city.NOM destroy.NACT.PAST.PERFVE.3SG by ‘the city was destroyed by the earthquake’

IJȠȞ ton the

ıİȚıȝȩ sizmo earthquake

b.

Ǿ ʌȩȜȘ țĮIJĮıIJȡȐijȘțİ Įʌȩ i poli katastrafike apo the.NOM city.NOM destroy.NACT.PAST.PERFVE.3SG by ‘the city was destroyed by the soldiers’

IJȠȣȢ ıIJȡĮIJȚȫIJİȢ tus stratiotes the soldiers

(d) cross-linguistic differences concerning verbs which cross-linguistically are alternating. The core of verbs which are alternating are cross-linguistically stable; there is, however, cross-linguistic variation: English: nonalternating causatives destroy, kill – Greek: alternating causatives țĮIJĮıIJȡȑijȦ katastrefo ‘to destroy’, ıțȠIJȫȞȦ skotono ‘to kill’. Perhaps, however, the above examples from Modern Greek can be analysed as passives, as in Modern Greek the passive type is also formed with a non-active ending (as are some anticausatives) and not with periphrasis as in English. (e) moreover, it generally appears that Modern Greek has fewer selectional restrictions for intransitives than English or German. Causativisation appears to be more productive in Greek. The following example is grammatical in Modern Greek, but not in English and German: (2) a.

b.

ȉȠ ʌĮȖțȩıȝȚȠ ȡİțȩȡ ȑıʌĮıİ to pangozmio rekor espase the world record.NOM break.ACT.PAST.PERFVE.3SG ‘the world record was broken’ ȉȠ ıȣȝȕȩȜĮȚȠ ȑıʌĮıİ to simvoleo espase the contract.NOM break.ACT.PAST.PERFVE.3SG ‘the contract broke’

On the other hand, it has to do with particular lexical, idiosyncratic and specific restrictions for every language, if we look at cases such as: (3) ‘break his heart’: a.

but:

46

b.

*ȑıʌĮıİ *espase break.ACT.PAST.PERFVE.3SG ‘(she) broke his heart’ ȡȐȖȚıİ rajise crack.ACT.PAST.PERFVE.3SG ‘(she) cracked his heart’

IJȘȞ tin the

țĮȡįȚȐ kardhia heart.ACC

IJȠȣ tu his

IJȘȞ tin the

țĮȡįȚȐ IJȠȣ kardhia tu heart.ACC his

Wasow (1977: 333) argues that transitivity alternations are defined by lexical idiosyncratic characteristics of verbs, and supports his opinion with examples such as (1) – (4) from English. In these examples, however, the main difference that can be observed is the presence of more restrictions for anticausative than for causative structures: (1) a.

John dropped the rope John lowered the rope b. The rope dropped *The rope lowered (3) a. John shattered the light bulb John demolished the light bulb b. The light bulb shattered *The light bulb demolished

(2) a.

John darkened his hair John tinted his hair b. His hair darkened *His hair tinted (4) a. We moved the boxes We transported the boxes b. The boxes moved *The boxes transported

208

Chapter Four

cause (Smith 1970; Levin & Rappaport Hovav 1995; Reinhart 2000, 2002). On the contrary, the alternating causatives (such as ‘break’) can have as subject agents, instruments or causes. This difference may be interpreted with the approach of the underlying causative structure: a basic causative verb can leave the external argument unspecified if its thematic nature is unspecified (agent or instrument or cause); in contrast, when the verb is specified concerning the nature of the external argument, the external argument position cannot be removed (or bound with a lexical process). The same argument also applies to diachronic data: the intransitive structures are more restricted in relation to the causative structures in all periods. This fact is connected with the much freer causativisations of initial intransitives and with the much more restricted and non-systematic innovative intransitive uses of initial transitive verbs47. (b) The reflexive adjuncts which are connected to a common causative argument can be present even in anticausative structures. (76)

Ș ʌȩȡIJĮ ȑțȜİȚıİ Įʌȩ ȝȩȞȘ IJȘȢ i porta eklise apo moni tis the.NOM door.NOM close.ACT.PAST.PERFVE.3SG by itself ‘the door closed by itself’

In all periods of the history of the Greek language, the presence of reflexive adjuncts (DP in the dative or PP) with anticausative types is possible. (c) The cause, which refers to the initial (which causes the change) change-of-state subevent, can appear in anticausative structures comprising evidence of the presence of this common subevent in the anticausative structure, exactly as it is present in the causative structure (Pustejovsky 1995). (77)

Ș ʌȩȡIJĮ ȐȞȠȚȟİ Įʌȩ IJȠȞ i porta anikse apo ton the.NOM door.NOM open.ACT.PAST.PERFVE.3SG by the ‘the door opened by the wind’

ĮȑȡĮ aera wind

The reference and access to the initial subevent (which causes the change-of-state) is possible in the event structure of unaccusatives, something which renders the introduction of cause possible. In contrast, the reference and access to the agent of the event (Pustejovsky 1995) is not possible. The phrase Įʌȩ apo ‘by’+agent and the sentences denoting the purpose of the action (ȖȚĮ ȞĮ ... ja na ‘for to’ ...) are ungrammatical as they do not refer to the initial event but to the agent of the event. The agent is not part of the lexical representation of the predicate and no chain of correlation with the agent can be achieved. The Greek anticausative verbs, active and non-active, accept the presence of a cause in all periods, in Classical Greek (in which they bear constantly non-active voice), in the Hellenistic and Medieval Greek periods (when the spread of active voice begins) and in Modern Greek. Consequently, the presence of the cause is stable and independent of other factors in anticausative structures. (d) the aktionsart of both causatives and anticausatives is the same: generally, the causative verbs are telic (they include a process which leads to a change); similarly, the anticausatives are also telic (they show a simple change-of-state). Changes in the aktionsart of anticausative and causative types in the different diachronic periods are not observed. The anticausatives and causatives belong consistently to the event schema of accomplishments, whereas the diachronic derivation of the causative from the anticausative type (with addition of an external argument) is stable. (e) Reinhart (2000, 2002) theoretically differentiates between anticausativisations in lexicon, the result of which are intransitive types (as the lexical semantic representation is in every case causative), and reflexivisations which take place in syntax, of which result are reflexive types (from lexical transitive types)48. The present diachronic study shows the systematic presence of innovative causative types in the lexicon in every period. Further study is needed for innovative reflexive uses: my prediction is that the direction of the reflexivisations in the diachronic lexicon will be the opposite; innovative uses will be concerned with new intransitive (reflexive) types from initially exclusively transitive verbs. 47

48

ǿ exclude, of course, the hypothesis that the derivation may be different in every period, not only for theoretical reasons, but because in all periods the same direction of change (causativisations) is observed. According to this approach (Reinhart 2000 and Papangeli 2004 for Modern Greek), the role of the presence of the [+cause] is particularly important: the [+cause] verbs permit (i) the absorption of the external argument (expletivisation), which results in an unaccusative entry in the lexicon, and (ii) the self-passivisation (absorption of the internal argument) which is a syntactic process.

Diachronic Course and Derivation of Transitivity Alternations

209

(f) The purely diachronic argument: transitivity alternations are unstable. The productivity and the direction of transitivity alternations strengthen the approach of the causative lexical semantic representation. According to this approach, if an anticausative verb appears as a lexical item in one language, then its innovative causative use is also very possible. This is also shown by contemporary tendencies which are observed cross-linguistically, as can also be seen in the following examples: (78) English

anticausative: The roof deteriorated causative: %The pine needles deteriorated the roof (Levin & Rappaport Hovav 1995) anticausative: I pomodori sono cresciuti the tomatoes have.INTR grow.PART ‘the tomatoes have grown’ causative: %I figli, Gianni li ha cresciuti bene his-own son Gianni him has.TRANS grow.PART well ‘his son, Gianni has grown him well’ (Chierchia 1989/ 2004)

Italian

This is also shown by the Greek diachronic data. In (79-83) I repeat from sections 3.1, 3.2, and 3.3 cases of instability in argument structure concerning all periods of Greek (the numbers in the second set of parentheses refer to the example numbers in the corresponding section): Ancient Greek (79) (9) Η΍ΝΔΣΝsio:pao: ‘to keep silence/be silent – keep secret/speak not of’ i Homeric Greek intransitive active Ά΅ΗϟΏΉ΍΅ǰȱȱ Η΍ΝΔφΗΉ΍΅Αȱȱ д̄Λ΅΍Γϟȱȱȱȱȱȱ a. ΉϢȱȱ ·ΣΕȱȱ ΘΓ΍ǰȱȱ ei gar toi, basileia, sio:pe:seian Akhaioi if only because queen.VOC keep-silence.ACT.AOR.OPT.3PL Achaeans.NOM ‘My queen, if only the Achaeans would be silent for you’ (Homer, Odyssey, 17, 513; 8 BC) i Classical Greek intransitive active Ύ΅Ϡȱȱ ΓЁΈξΑȱȱ ΪΑȱȱ ΉϥΔΓ΍ΐ΍ȱȱ b.ȱ Η΍ΝΔЗȱȱ sio:po: kai ouden an eipoimi keep-silence.ACT.PRES.1SG and nothing.ACC prt say.1SG ‘I keep silence and I will never say a word’ (Demosthenes, Ctesiphon, 268, 3; 4 BC) transitive active ̘ΝΎνΝΑȱȱ σΌΑΓΖȱȱ ΘΓΗΓІΘΓΑȱȱ ΦΑϙΕ΋ΐνΑΓΑȱȱ Η΍ΝΔЗȱȱȱȱȱ ȱȱ c.ȱ Ύ΅Ϡȱȱ Θϲȱȱ kai to Pho:keo:n ethnos tosouton ane:re:menon sio:po: and

the.ACC Phocians.GEN nation.ACC so-much

destroy.PART.ACC

keep-silence.ACT.PRES.1SG

‘I do not reveal the destruction of the important nation of the Phocians’ (Demosthenes, Philippic, 3, 26, 6; 4 BC) (80) (12) ΐ΅Ε΅ϟΑΝȱmaraino: ‘to wither’ȱ i Homeric Greek intransitive non-active Δ΅ϾΗ΅ΘΓȱȱ Έξȱȱ ΚΏϱΒȱ a.ȱ ΔΙΕΎ΅Ϟχȱȱ πΐ΅Ε΅ϟΑΉΘΓǰȱȱ pyrkaie: emaraineto, pausato de phloks
fire.NOM wither.NACT.IMPERF.3SG stopped.3SG prt flame.NOM ‘the fire died down and the flames were over’ (Homer, Iliad, 23, 228; 8 BC)ȱ i Classical Greek intransitive non-active Ύ΅Ϡȱȱ ΐ΅Ε΅ϟΑΉΘ΅΍ȱȱ ΑϱΗУ b.ȱ ΚΌϟΑΉ΍ȱȱ ·ΤΕȱȱ phthinei gar kai marainetai noso:i wane.3SG because and wither.NACT.PRES.3SG malady.DAT ‘for she is waning and wasting with her malady’ (Euripides, Alcestis, 203; 5 BC) transitive active ϳȱȱ ΐν·΅Ζȱȱ ΛΕϱΑΓΖȱȱ ΐ΅Ε΅ϟΑΉ΍ȱȱȱȱȱȱȱȱȱ c.ȱ ̓ΣΑΌȂȱȱ Panth’ ho megas khronos marainei all.ACC the.NOM strong.NOM time.NOM wither.ACT.PRES.3SG ‘the strong time makes all things fade’ (Sophocles, Ajax, 713; 5 BC) Hellenistic-Roman (81) (6) ϊΈΝ he:do: ‘to please/delight/enjoy’ intransitive non-active ϳȱȱ ΐνΌΙΗΓΖȱȱ a.ȱ ΓϩΓΑȱȱ·ΓІΑȱϊΈΉΘ΅΍ȱȱ hoion goun he:detai ho methysos so then enjoy.NACT.PRES.3SG the drunk.NOM

πΐΚΓΕΓϾΐΉΑΓΖȱȱ

ΓϥΑΓΙȱ

emphoroumenos oinou fill.NACT.PART.NOM wine.GEN

210

Chapter Four ‘so then the drunk enjoys as he is filled with wine’ (Sextus Empiricus, The Outlines of Pyrrhonism, 3, 195, 5; AD 2-3)

transitive active b.ȱ БΖȱȱȱΘϲȱΗ΍ΘϟΓΑȱȱ ho:s to sition as

ȱύȱȱ Θϲȱȱ ΔΓΘϲΑȱȱ

ϊΈΉ΍ȱΐξΑȱȱ

ΘϲΑȱȱ Κ΅·ϱΑΘ΅ȱȱ

ύȱȱ Δ΍ϱΑΘ΅ȱȱȱȱ

e: to poton

he:dei men

ton

e: pionta

phagonta

the food.NOM or the drink.NOM enjoy.ACT.PRES.3SG the eat.PART.ACC

or drink.PART.ACC

‘as the food or the drink pleases the person who is eating or drinking’ (Sextus Empiricus, Against the Mathematicians, 7, 368, 3; AD 2-3)

i in the previous period - Classical Greek:
exclusively intransitive non-active Θ΍Ζȱȱ ȱȱȱȱύȱϣΔΔУȱȱ Φ·΅ΌХȱȱ ύȱΎΙΑϠȱȱȱȱ ύȱϷΕΑ΍Ό΍ȱȱȱȱ ϊΈΉΘ΅΍ǰȱȱ c.ȱ ΩΏΏΓΖȱȱ allos tis e: hippo:i agatho:i e: kyni e: ornithi he:detai, other

somebody.NOM or horse.DAT good

Ύ΅Ϡȱȱ σΘ΍ȱȱ ΐκΏΏΓΑȱȱ ϊΈΓΐ΅΍ȱȱ

ΓЂΘΝȱȱ

houto:

or dog.DAT or bird.DAT enjoy.NACT.PRES.3SG so ȱ

ΚϟΏΓ΍Ζȱȱ

Φ·΅ΌΓϧΖ

kai eti mallon he:domai philois agathois and even more enjoy.NACT.PRES.1SG friends.DAT good.DAT ‘others have a fancy for a good horse or dog or bird: my fancy, stronger even than theirs, is for good friends’ (Xenophon, Memorabilia, 1, 6, 14; 5-4 BC) (82) (7) ΏΙ·ϟΊΝ lygizo: ‘to bend/twist as one does a withe’ intransitive non-active ӥΕΝΘΓΖȱȱ ЀΔȇȱȱ ΦΕ·΅ΏνΝȱȱ πΏΙ·ϟΛΌ΋ΖЪ a.ȱ ΓЁΎȱȱ ΅ЁΘϲΖȱȱ ouk autos Ero:tos hyp’ argaleo: elygikhthe:s ? not yourself.NOM Love.GEN by troublesome.GEN bend.PASS.AOR.PERF.2SG ‘you did not bend by the troublesome Love?’ (Theocritus, Idylls, 98; 4-3 BC) transitive active Θ΍ȱȱ ΘЗΑȱȱ ΐΉΏЗΑȱȱ b.ȱ ΐ΋Έξȱȱ ΐχΑȱȱ ΦΑ΅·Ύ΅ΊΓΐνΑΓΙΖȱ ΏΙ·ϟΊΉ΍Αȱȱ me:de me:n anagkazomenous lygizein ti to:n melo:n neither not force.PART.ACC bend.ACT.PRES.INF one.ACC the.GEN body-parts.GEN ‘neither were they forced to bend one from their parts’ (Aelian, On the characteristics of animals, 2, 11, 46; AD 2-3) i in the previous period - Classical Greek:
exclusively intransitive non-active Α΍Αȱȱ ΩΕΌΕΓΑȱȱ ϗȱȱ ΏΙ·ϟΊΉΘ΅΍ȱȱ c.ȱ ΔΓΈϱΖȱȱ podos nin arthron he:i lygizetai foot.GEN him ankle.NOM which.DAT bend.NACT.PRES.3SG ‘(he caught him) by the foot where the ankle bends’ (Sophocles, Trachiniae, 779; 5 BC) Medieval (and Early) Modern Greek (83) (5) ʌİșĮȓȞȦ petheno ‘to die’ intransitive active ΔΉΌ΅ϟΑΓΙΗ΍Αȱȱ ΅ЁΘΓϟȱȱ  a.ȱ ΈξΑȱȱ dhen pethenusin afti not die.ACT.PRES.3PL they.NOM ‘they do not die’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 263, 20; AD 17) transitive active ·ϟΑΓΐ΅΍ǰȱȱ Ύ΅ϠȱσΘΊ΍ȱȱ ΐΔ΅ϟΑΝȱȱ ΉϢΖȱΘχΑȱΐϾΘ΍Αȱ b.ȱȱ ΩΑΉΐΓΖǰȱȱΉϨΔΉΑȱȱ ΅ЁΘφǰȱȱ anemos, ipen afti, jinome, ke etzi beno is tin mitin ȱȱȱȱȱȱȱȱ ȱ wind said.SG she.NOM become.1SG and so enter.1SG in the noseȱ ȱ ΑΤȱȱ ΘΓϿΖȱȱ Δ΅΍Ό΅ϟΑΝȱȱ ȱ ȱ ȱ ΅ЁΘΓΑЗΑȱȱ ϳΔΓІȱȱ Όξȱȱ aftonon opu the na tus petheno these.GEN who want.3PL to they.ACC.WEAK die.ACT.IMPRFVE.1SG ‘I become as the wind, she said, and like that I enter the nose of those whom I want to make die’ (ǺGV (Legrand, Bibliothèque grecque vulgaire) Iǿǿ, 163, 15-16; AD 17) i exclusively intransitive in in the previous period (Hellenistic-Roman): Έξȱ Ύ΅Ϡȱ ̝ΑΑϟΆ΅Ζȱ ϳȱ ΗΘΕ΅Θ΋·ϱΖ c. ΦΔνΌ΅ΑΉȱ apethane de kai Annibas ho strate:gos die.ACT.AOR.3SG prt and Anivas.NOM the.NOM commander.NOM ‘also Anivas, the commander, died’ (Diodorus Siculus, Library, 13, 86, 3; 1 BC)

In the opposite case, if we accept a common anticausative representation for both the transitive and intransitive types (the anticausatives to be classified in the lexicon as verbs from which the external argument is absent), there would be no reason for those verbs to be transformed much more easily into causatives than the unergative intransitives (ijIJȐȞȦ ftano ‘to arrive’, ıțĮȡijĮȜȫȞȦ skarfalono ‘to climb’, țȠȜȣȝʌȫ kolimbo ‘to swim’; cf. Chierchia 1989/2004: 41).

Diachronic Course and Derivation of Transitivity Alternations

211

Furthermore, it is not possible for an innovative anticausative use to arise in cases where only a causative structure is available. In English, causative predicates such as compose (in contrast to decompose), murder and sterilise which do not have anticausative use are stably causative (Davis 2000). The same applies to Modern Greek, where the corresponding verbs with anticausative interpretation and obligatory absence of agent are ungrammatical: (84) a. ???ıȣȞIJȓșİIJĮȚ Įʌȩ ȝȩȞȠ IJȠȣ ???sindithete apo mono tu compose.NACT.PRES.3SG by itself ‘*it composes by itself’ b. *įȠȜȠijȠȞȒșȘțİ Įʌȩ ȝȩȞȠȢ IJȠȣ *dholofonithike apo monos tu murder.NACT.PAST.PERFVE.3SG by himself

‘*he was murdered by himself’ c. *ĮʌȠȜȣȝȐȞșȘțİ Įʌȩ ȝȩȞȠ IJȠȣ *apolimanthike apo mono tu

(ıȣȞİIJȑșȘ (sinetethi

Įʌȩ IJȠȣȢ İʌȚıIJȒȝȠȞİȢ) apo tus epistimones)

compose.NACT.PAST.PERFVE.3SG by the scientists

‘it was composed by the scientists’ (įȠȜȠijȠȞȒșȘțİ Įʌȩ IJȠȞ țĮIJȠȤȚțȩ ıIJȡĮIJȩ) (dholofonithike apo ton katoxiko strato) murder.NACT.PAST.PERFVE.3SG by the occupation army

‘he was murdered by the occupation army’ (ĮʌȠȜȣȝȐȞșȘțİ Įʌȩ IJȠȣȢ ȖȚĮIJȡȠȪȢ) (apolimanthike apo tus jatrus)

disinfect.NACT.PAST.PERFVE.3SG by itself disinfect.NACT.PAST.PERFVE.3SG by the doctors

‘*it disinfected by itself’

‘it was disinfected by the doctors’

The Greek diachronic data (productive causativisations but not anticausativisations) strengthen such an analysis since no systematic innovation of anticausative structures occurs in cases where only the causative structure was lexicalised. The causative verbs of Classical Greek which do not participate in anticausative structures (non-alternating causative verbs) do not change, they have stable syntactic behaviour, i.e. they are constantly non-alternating causatives. In addition, the dissociation of voice morphology from the derivation of innovative transitivity alternations (as we observed in the diachronic data) and the fact that in the synchrony of Classical Greek the anticausative members of the alternation (even the verbs corresponding to the active anticausatives of Modern Greek) bore stable non-active morphology49 (85a, b, c, d) appear to exclude (i) the relationship of voice morphology with the lexical derivation of causatives and anticausatives, as well as the counter-argument for the anticausative lexical semantic representation only for anticausatives with non-active voice, and (ii) the analysis of anticausatives as a non-unified group of verbs, whose differences in voice morphology are owing to differences in aktionsart. (85)

a. ψȱȱ Έξȱȱ ΌΣΏ΅ΗΗ΅ȱȱ Δφ·ΑΙΘ΅΍ȱȱ he: de thalassa pe:gnytai the.NOM prt sea.NOM freeze.NACT.PRES.3SG ‘the sea freezes’ (Herodotus, The Histories, 4,28; 5 BC) bǯȱ ΓЁ Έ΍Ύ΅ΗΘφΕ΍Σȱȱ ΘΉȱΎ΅Ϡȱȱ Ϣ΅ΘΕΉϧ΅ȱȱ ΔΓΏΏΤȱȱ ΦΑΓϟ·ΉΘ΅΍ ou dikaste:ria te kai iatreia polla anoigetai not courts.NOM and dispensaries.NOM many.NOM open.NACT.PRES.3SG ‘not many courts of law and dispensaries opened’ (Plato, Republic, 405a, 4; 5-4 BC) cǯȱ ǻΘϲȱΦ··ΉϧΓΑǼȱȱ ΎΏΉϟΉΘ΅΍ȱȱ Έξȱȱ πΖȱȱ ΘχΑȱȱ țĮȡįϟȘȞ (to aggeion) kleietai de es te:n kardie:n the vessel.NOM close.NACT.PRES.3SG prt in the heart ‘the vessel closes in the heart’ (Hippocrates, On the Heart, 12, 4; 5-4 BC) dǯȱ ΘΤȱȱ Η΍Θϟ΅ǯǯǯȱȱ ΌΉΕΐ΅΍ΑϱΐΉΑ΅ȱȱ ȟȘȡĮϟȞİIJĮȚ ta sitia... thermainomena kse:rainetai the.NOM grain.NOM warm.PART.NOM dry.NACT.PRES.3SG ‘the grain, if warmed, dries’ (Hippocrates, On Affections, 59, 4; 5-4 BC)

In relation to uni-directionality of the changes we have observed50, evidence also comes from studies on language acquisition of verbs (cf. the relationship between language change and language acquisition according to the theory of Lightfoot, section 2.2)51. Children’s errors concerned with 49

50

51

Both the anticausative verbs denoting change-of-state as a result of a natural, biological or spontaneous process (the majority of which are active in Modern Greek; Vassilaki 1988) and the anticausative verbs which derive from adjectives (and in Modern Greek correspond to verbs which are active) bear non-active voice morphology in Classical Greek. The productive change that we ascertained is as follows: alternating verb type (anticausative + causative use) from anticausative (non-alternating) type. In contrast, non-productive change is the following: alternating verb type (causative + anticausative use) from causative (non-alternating) type. Nouchoutidou (2003) and Tsimpli (2006a, b) refer to errors concerned with voice morphology of passives,

212

Chapter Four

causative uses of intransitive verbs are systematic and productive. The first study on the acquisition of transitivity alternations was conducted by Bowerman (1974, 1982). Bowerman ascertains the systematic creation of innovative causative types by children who acquire their native language. The new causative types are created on the basis of intransitive verbs such as: unaccusatives which are not alternating (86)

a. *I’m gonna disappear it under the washrag b. *Water bloomed these flowers

unaccusatives which are not alternating as they have a lexical suppletive causative: (87)

a. *I come it closer (ĺ bring) b. *Go me to the bathroom (ĺ take) c. *I’m gonna fall this on her (ĺ drop)

unergatives (which do not, of course, have causative use): (88)

a. *I’m singing it (ĺ with the meaning ‘I make the musical box sing/work’) b. *You cried her c. *I wanta swim that

Lord (1979), who observes the transitive use of 80 intransitive verbs, agrees with the remarks of Bowerman. Lord characterises a few examples as erroneous intransitive use of transitive nonalternating verbs. If Lord is correct, then we have to do with errors during language acquisition in both directions. Careful observation, however, of the few examples of ‘erroneous or innovative’ intransitive use in the data given by Lord shows that it is usually a case of unsuccessful formation of the passive type (as an agent is implied or present). Only in one case do we have anticausative use of a causative (non-alternating) type: (89)

*Do you think it will fix

In other cases, a PP-agent can be added (therefore, they are passive and not anticausative structures) and they would be entirely grammatical if the passive type had been formed correctly: (90)

a. *It won’t lose (= it won’t get lost) b. *Pull! (= get moved, be pulled – the child pulls the father by the hand, his father does not move and the child tells him that he has to move, to be pulled in the direction that the child wants him to head). c. *And corn doesn’t crunch, it eats d. *They don’t seem to see. Where are they? e. *They attract by the peanuts

The above remarks about the very productive causativisations of unaccusative verbs can be connected with the remark of Montrul (1997: 118) that the causativisation of an unaccusative is more natural than that of an unergative verb, because the causativisation of the unaccusative means completion of the VP2 (according to Montrul’s proposal of analysis) with the introduction of the CAUSE subevent and the agent argument52. In contrast, in the case of unergatives, the agent argument is found in the Spec of the subevent of CAUSE (or [Spec, vP]), and, consequently, causativisation of unergative means demotion of the existing agent in [Spec, VP1] (or simply VP) and addition of a second agent. Montrul53 assumes that the default transitive template (Ÿ corresponds to a basic causative schema or causative lexical semantic structure), based on which transitive and intransitive structures are derived, leads both L1 and L2 acquisition (Hypothesis of the Default Transitive Template)54.

52

53

54

reflexives and anticausatives in L1 acquisition of Greek. Other interpretations of errors in language acquisition assign the preference for transitive uses, even in cases of non-alternating intransitives, to the very important role which is played by agentivity in grammar (Finer & Roeper 1989; Budwig 1990; DeLancey 1990; Verrips 1993). Montrul’s study is of particular importance for the area of the L2 acquisition of the lexicon. Montrul does not observe influence of L1 in the acquisition of the causative argument structure (same transitivity errors are observed (mainly erroneous causativisations) independently from L1); in contrast, L1 plays a very important role in the acquisition of derivational morphology which expresses (anti)causativity. For Montrul, the semantic decomposition is based on syntactic principles and not on mapping in the interface.

Diachronic Course and Derivation of Transitivity Alternations

213

(91) VP2  agent Vǯ  CAUSE VP1  patient Vǯ  BECOME predicate If the alternating verbs have the causative as basic structure, then the anticausative type is more complex and structurally marked than the causative type since it comprises an additional step in derivation. The assignment of the default transitive template to one non-alternating transitive verb, the suppression of the agent and the binding of the CAUSE with the patient (the process, namely, for the derivation of intransitive from transitive types in erroneous de-transitivisations, which are not at all frequent in L1 acquisition and are shown to be not at all productive as diachronic changes) is a much more uneconomical process than that of the simple correlation of a non-alternating unaccusative with the causative template and the completion of the [Spec, CAUSE P] position with an agent argument (the process, namely, for the derivation of a transitive type from a non-alternating intransitive in erroneous transitivisations). The (causative) lexical semantic representation of alternating verbs explains for Montrul why the errors in one direction are in any case many more than in the other direction: the causative errors are syntactically and semantically less marked than the anticausative errors. The above arguments strengthen our hypothesis for a basic causative template (or a causative lexical semantic structure) and are correlated with our proposal for the presence of the functional head of v in the syntactic representation of both causatives and the anticausatives (cf. chapter 2 and 4.2). (92)

vP  +/-agent vǯ  v VP [+/-agent]  V patient x [[x does something] causes

y [y becomes STATE]]

In the causative types, both arguments appear: agent (or cause or instrument) in the [Spec, vP] position and patient in [Compl, VP]. In the anticausative types, the agent is lexically bound (according to the analysis of Levin & Rappaport Hovav 2005): anticausative type Lexical semantic representation [[x does something] causes [y becomes CLOSED]] Mapping rules Ļ Ļ Argument structure ‡

The result of the lexical binding of the agent is that the patient argument is the subject of the intransitive structure (after movement to the subject position, [Spec, vP])55.

4.4 Summary The goal of this present section was the overall interpretation of changes in transitivity and voice morphology which are concerned with the processes of (anti)causativisation. The first part of the interpretation sought the causes of changes in argument structures. In all 55

I do not concern myself at all in this monograph with the syntactic position of the subject. For Greek, I refer to Alexiadou & Anagnostopoulou (1998), Spyropoulos & Philippaki (2001), Roussou & Tsimpli (2006).

214

Chapter Four

periods of Greek, as well as cross-linguistically, the causativisation process (new causative use (with the meaning of change-of-state) of an earlier intransitive non-alternating verb) is particularly productive. The change in the opposite direction (anticausative from non-alternating causative) is not at all productive. The new structures are the result of reanalysis of intransitive structures. The neuter DPs which bear the same endings for nominative and accusative, the availability of subject drop (pro-drop), the free word order, the characteristics and the position of the clitics (which create an unbreakable unit with the verb) and the small number of verb roots (i.e. idiosyncratic semantic components which arise from the verb decomposition) in Greek favoured the reanalysis of intransitive types in all periods. In all periods, we find examples of possible reanalysis resulting in not only an innovative active causative type (in the case of verbs which were exclusively non-alternating intransitive) but also in change in voice morphology and in active anticausative types (in the case of verbs which already had active causative and non-active anticausative type). I follow the cue-based approach (Lightfoot 1999, 2006a, b) and assume that in reanalysis of the argument structures the case markers comprise cues for language acquisition and change. The case which is assigned to the DP in a specific structural position –for example, in the [Compl, VP] position– can function as a cue, the different morphological case types marking the different structures in which case is assigned. The neuter DPs in the nominative case can be reanalysed as DPs in the accusative since the neuter DPs bear the same ending for nominative and accusative. The reanalysis is favoured particularly in the cases of third person verbs with neuter DPs; these DPs can be analysed as nominatives and subjects of the third person verbs or as accusatives and objects of third person verbs (for which verbs a subject with the meaning ‘he/she’ is implied). The child who acquires language searches for cues and hypothesises (according to the characteristics which s/he comes across) that if an anticausative verb has a patient argument marked with accusative case, then it is possible in the structure of the specific verb for an external argument to be added in the subject position. Apart from the same morphology of the neuter DPs for the nominative and the accusative, two more elements contribute to this direction of changes: (a) the use of the accusative case for the object of finite verbs as well as for the subject of infinitives, and (b) the formation of participles in -ȝİȞȠȢ -menos both for transitive and intransitive verbs. Concerning voice morphology, the instability in voice of anticausatives is not connected with the direction of changes in the argument structure. The change that is observed in relation to the voice of causatives and anticausatives from the Hellenistic-Roman period and more intensely from the Medieval Greek period is the change of non-active anticausatives into active. This change is not connected with changes in the argument structures but with more general changes in the structure of the specific verbs and mainly with the absorption of the accusative case by non-active voice and the following presence of the Voice [+agent] head in the structure of passive types (in Ancient Greek, the structure non-active type + DP-accusative was grammatical). The differences in voice can be summarised as follows: (a) Ancient Greek: (i) the structure “non-active type + DP-accusative as direct object” is grammatical, (ii) passives can bear active voice morphology, (iii) anticausatives have non-active endings; (b) from Early Medieval Greek and afterwards: (i) the structure “non-active + DP-accusative as direct object” is ungrammatical (the non-active blocks/absorbs the accusative case), (ii) the passives bear non-active voice morphology, (iii) most of the anticausatives have active endings. In relation to the above observations, I assume that the anticausatives of Medieval Greek, which are systematically active, do not have a Voice head. The anticausatives of Medieval Greek (and Modern Greek) that bear non-active voice, however, have the lexical feature [non-act] in their root. The passives of Medieval Greek (and of Modern Greek) are non-active and have a Voice[+agent] head which is responsible for the non-presence of a DP in the accusative. The anticausatives of Ancient Greek are non-active and have a Voice[-agent] head. I distinguish two non-active anticausative types: non-active types which are marked in the lexicon (in the root; Medieval Greek anticausatives) and nonactive types which are marked in syntax (due to the presence of Voice; Ancient Greek anticausatives). The change in voice morphology is dictated by the principle of economy, owing to which I assume a general linguistic principle is applied which requires the separation of the anticausatives and passives. The stable marking of passives with non-active voice (the same as with anticausatives) from the end of the classical Ancient Greek period, and mainly from the Hellenistic-Roman period, leads to changes in the voice of anticausatives. The reason is that, according to this principle, language selects the separation (passive: Voice [+agent], non-active voice morphology; anticausative: absence of Voice, active voice morphology) from the presence of Voice with two sub-types (passive: Voice [+agent], nonactive voice morphology; anticausative: Voice [-agent], non-active voice morphology). Concerning the anticausatives of Medieval and Modern Greek which retain non-active voice I follow and modify Embick’s approach for the deponent transitive verbs of Latin: the anticausatives which did not change into actives inherently have (in the verb root) the [non-act] feature. This feature is stable diachronically because it belongs to the root of the verb and not to the common part of the verb with the remaining

Diachronic Course and Derivation of Transitivity Alternations

215

verbs of the same semantic group. If the verbs of this group appeared with non-active endings for systematic syntactic reasons, they would have formed a uniform group with particular semantic and syntactic behaviour, but this does not occur. From the above, the need for a connection between causatives and anticausatives (which I represented syntactically with the presence of v) is evident. Whereas the tendency is for dissociation of anticausatives and passives (through the loss of one of the two types of Voice), causatives and anticausatives are connected since they bear two types of the same head (v, +/-agent). I argue that the two factors which I analysed (causativisations and instability in voice) strengthen the idea of a basic causative schema both for the causative and for the anticausative type (both types are derived from a structure with v, which is responsible for the active voice morphology in the causative type in all periods and in the anticausative type in Medieval Greek). The fact that both types (causative and anticausative) are derived lexically from a common causative verb is also apparent from the strong tendency of anticausatives (intransitive verbs of change-of-state) to causativise in the diachrony of many languages, and from the absence of anticausative sub-groups based on voice in all periods. The instability of transitivity alternations comprises a clear diachronic argument for the above view. The productivity and the direction of the transitivity alternations strengthen the approach of the basic causative lexical semantic representation, according to which if an anticausative type appears as a lexical item in one language, then its innovative causative use is also entirely possible; but it is not similarly possible for innovative anticausative uses to arise in cases where only a causative structure is possible. The diachronic data from the history of the Greek language (productive causativisations and not anticausativisations) provide arguments in favour of such an approach since we do not observe systematic innovation of new anticausative structures in cases where only a causative structure was lexicalised. The causative verbs of Classical Greek which do not participate in anticausative structures (they are non-alternating causatives) do not change in the following periods but have a stable syntactic behaviour, namely they are stable non-alternating causatives.

CHAPTER FIVE CONCLUSIONS

The goal of this monograph was the exploration of the diachrony of transitivity and, more specifically, of causative (prototypical transitive) verbs. The theoretical model adopted is in accordance with the later developments in the theory of Generative Grammar and specifically in the area of mapping lexical semantic information with syntactic structures, from which arises the hypothesis of verb groups. The lexicon is governed by rules and the lexical items are structured as complex units comprising semantic parts and not as unified units. Following the underlying rules of the lexicon, native speakers in every period discover the meaning of each verb and especially the aspect of verb meaning which has syntactic consequences. In the diachronic approach to the specific phenomenon I took into consideration principles from Generative diachronic syntax. One main difference between diachronic syntax and other types of comparative syntax is that the empirical data of diachronic syntax come from the E-language. The methodological problem, of course, is how an interpretation can be made of the E-language of diachronic texts in relation to the I-language of the speakers. Aspects of the E-language may be interpreted as the output of underlying grammar in cases where we are really able to leave aside effects of performance parameters. Additionally, a group of inter-related changes, which have been ascertained in diachronic texts, can reflect a change in the grammatical system. Furthermore, the significance of the changes in the E-language, which occur in the same period, is greater than it appears to be from the above for two reasons: (a) a difference in the grammatical system may justify a group of surface changes between different periods; (b) the analysis of variation in the E-language may uncover differences in the grammatical system. As a language changes, we are able to follow the course of the changes, and this course of the changes finally gives us information on the nature and the organisation of the language (the specific one and languages in general), which is not available in synchronic comparative research. Consequently, a detailed study of the patterns of variation is one method of drawing conclusions for the grammatical system based on data on the E-language. Grammar may change in the space of two generations. The change occurs when the choices of the speakers fall together in one grammatical system which differs at least in one characteristic/parameter from the system of the speakers of the previous generation. The interpretation of language change is, therefore, connected with the interpretation of language acquisition. Primary Language Data guide the child to acquire the language since they function as a source of the essential cues for the development of the language. Language change is the result of a different function and interpretation of precisely these cues1. According to the theoretical framework I followed, I have attempted to interpret the diachronic differences and changes in transitivity. Whereas the lexico-syntactic templates of causative and anticausative types are considered to be common cross-linguistically and across different periods of the same language, the languages and the diachronic periods of the languages differ concerning the morphological marking of the transitivity alternations of the verbs denoting change-of-state. Consequently, the main questions were: (a) what are the mechanisms of change of transitivity alternations, and are they the same in all periods? and (b) what role does the specific manner in which transitivity alternations are morphologically marked play in diachronic changes? The approach I followed can interpret the types of changes in transitivity which we observed in Greek diachrony, as well as cross-linguistically. Not all verb groups are created in the same manner. The different verb groups are the result of the interaction between the event types and the ș-roles. The type of subevents, the number and type of arguments and more specific semantic information (type of CAUSE and type of agent) define which verbs participate in transitivity alternations and which do not. The diachronic innovative causativisations (of non-alternating intransitive verbs) are a result of the knowledge (native speakers have) of the common meaning of intransitive verbs which belong to the specific verb group (change-of-state) and their mapping with the basic causative template (which is common to both the causative and the anticausative type). This basic (causative) template would appear to lead to diachronic productive causativisation of earlier non-alternating intransitive verbs. 1

Cf. also Lavidas 2006a, b.

218

Chapter Five

The diachronic data of Greek were classified and analysed, on the one hand, in relation to the aforementioned questions and, on the other hand, in relation to possible changes (according to the theory), the aim being to explore whether there are limits to the possible changes. The possible and ascertained changes are as follows: (a) alternating verb from anticausative intransitive; (b) anticausative intransitive verb from alternating. The following changes are possible, but they have not been ascertained in the data of Greek diachrony: (c) alternating verb from causative; (d) causative verb from alternating. The changes in the morphological encoding of transitive and intransitive verbs were also taken into consideration. The following changes were ascertained: (a) in the voice morphology of intransitives which participate in the transitivity alternation; (b) in the voice morphology of transitive and intransitive verbs which do not participate in the transitivity alternation. Moreover, the following possible but unascertained changes were examined: (c) in the voice morphology of initially exclusive intransitives which change into alternating; (d) in the voice morphology of transitives which participated in the transitivity alternation. The results of this present study indicated the following: (a) The process of causativisation is particularly productive in all periods. The causativisations are the result of reanalysis of an intransitive type as transitive based on the cue of case. The ambiguity of the accusative and nominative case comprises the reason for the reanalysis. The changes in causative verbs (changes in argument structure) are differentiated from the changes in the accusative (nonprototypical transitive) verbs (mainly changes in morphological case). (b) There is no connection between voice morphology and the direction of change in transitivity. The basic principle which leads the changes in the voice system demands separation of the anticausatives from the passives: in the diachrony of Greek, this appears with the presence of Voice[agent] in the structure of the anticausatives of Ancient Greek, whereas Voice was absent from the structure of the passives, and from the absence of Voice from the anticausatives of Medieval Greek, whereas Voice[+agent] was stably present in the passives. This separation of the anticausatives from the passives is supported by a principle of economy, according to which I assume that the presence of a head (e.g., Voice) with two different types in the same period (e.g., Voice[-agent], Voice[+agent]) is uneconomical. A critical point in the development of changes in voice morphology was comprised by the characteristic of the absorption/blocking of the accusative case by non-active types. (c) The changes are chain-like: the extension of the accusative case and the loss of the benefactive meaning, which the non-active transitive types bore, are connected with the absorption/blocking of the presence of only the accusative case by the non-active voice (the other cases are not blocked); the absorption only of the accusative case by the non-active voice is connected with the stable presence of Voice in the passives, and the presence of Voice in the passives leads to the extension of the active voice in the anticausatives. I distinguish the differentiation of anticausatives and passives from the link between anticausatives and causatives: a diachronic argument in favour of the connection between the anticausatives and the causatives in a basic lexical causative template is the direction of changes in the argument structures, which leads consistently to the creation of a new causative from an intransitive type, as well as instability in the area of voice (cf. paras. 3.1.2, 3.2.2, 3.3.2). Linguists who are concerned with diachronic Generative Grammar agree that the process of language change is complex since many factors play a role in the transition from one structure aǯ to an innovative structure bǯ. One of the goals of the theory of language change is to find the aspects of changes which are determined by universal mechanisms and the aspects which are determined by the characteristics of every language. Even if, as Hale (1996) states, it appears to be very difficult, although not impossible, to achieve the above purpose, the similarities between the processes of change, which we observed cross-linguistically at the level of argument structure, led us to the conclusion that diachronic changes in verb items in the lexicon follow a basic lexical semantic structure (causative template), whereas the changes in derivational voice morphology are not correlated with the diachronic derivation of the intransitive from the transitive type, but with the structure of the causatives, the anticausatives and the passives, and the relation between morphological marking of anticausatives with the marking of passives. This conclusion is not in agreement with the approach of the direct connection of voice morphology with the derivation and the direction of derivation of transitive and intransitive types proposed earlier in the literature. The diachronic data revealed that in relation to argument structure all periods would appear to

Conclusions

219

follow the same linguistic principles. Causativisations were observed in all phases of Greek diachrony. The changes are characterised by uni-directionality since productive anticausativisations are not seen to result in an alternating or non-alternating innovative intransitive type from an earlier causative type. The changes concerning morphological encoding are independent of the derivation of new transitivity alternations. This can be seen from the fact that the changes in voice morphology are only concerned with intransitive verbs which were already participating in the alternation; namely, the change in voice of the intransitive verb which is causativised is not productive and the intransitive verb which becomes alternating retains its original morphological marking. The rich diachronic data also reveal changes which occur in the structure of the anticausative (presence of Voice[-agent] in Classical Greek; absence of the Voice[-agent] from the Medieval Greek onwards), and the distinct tendency towards the differentiation between the structure of the anticausative from the structure of the passive. The tendency towards the differentiation of the anticausative from the passive can be expressed as follows: (a) Ancient Greek: (i) non-active anticausatives, (ii) passives bearing active voice, and (iii) possibility of presence of DP in the accusative as direct object of a non-active type; (b) Medieval Greek: (i) active anticausatives; (ii) passives only in non-active voice, (iii) and blocking/absorption of the accusative case by the non-active type. In addition, this present analysis offers empiric arguments in favour of the approach of the interpretation of language change based on the principles of transparency-reanalysis and cues (Lightfoot 1999, 2006a, b). The diachronic data of Greek show that the case marking of the sole argument of intransitive verbs is directly connected with the reanalysis of several intransitive verbs as transitive. The ambiguity of the nominative and the accusative case of the neuter DPs favours reanalysis. I assume that the accusative case in the DP denoting the patient argument functions as a cue for the possibility of presence of an agent in the position of the external argument of the specific verb. The reanalysis of the intransitive type is also connected with other characteristics of the language: the availability of null subject (pro-drop), the position and the characteristics of the clitics, and the number of roots in a language strengthen the reanalysis of an intransitive as transitive. Limitations in the reanalysis are (im)posed by the meaning of the change from an internal cause (intransitive verbs, that denote change from an internal cause and are reanalysed as alternating, very frequently lose their new causative use in the following period) and from the availability of denotation with accusative case (alternatively of genitive/dative case or PP) of the cause of the event which the intransitive verb denotes. This present analysis strengthens the theoretical position of Lightfoot according to which language change is expected when there is no obvious connection between the interpretation of the output (the Elanguage) and the internalised grammar (I-language). The grammatical system cannot be acquired correctly if it is not reflected precisely (clearly and without ambiguity) in the language realisation, and, consequently, the grammatical characteristics which survive for many generations are clearly reflected in the PLD. If the grammatical characteristics change in any period, something in the language realisation of the previous generation must have changed and led the following generation to a new interpretation. I assume that changes that lead to variation in the PLD play an important role in the search for an interpretation of the changes under investigation; the language of a specific community can gradually become different from the language which earlier comprised the trigger for language acquisition. These changes in the E-language differentiate the input of the children who acquire the language and comprise the reason for a different determination of the parameters. Every change in the PLD can, of course, also be the result of a previous change. The general question which arises from the above is whether gradual changes in the E-language can be considered to guide changes in the I-language or if, in certain cases, the opposite occurs, namely if the changes in language realisation should be considered as changes in the I-language2. The essence of this question is whether we can accept the existence of changes in the E-language without there being changes in the I-language. The opinion accepted by the majority of linguists is that the changes in the E-language reflect changes in the I-language; this, of course, brings us to the question of what in reality the E-language is, and whether it really exists. What we observe is that the presence of the concept of the E-language is essential since the E-language contains the linguistic cues (in our phenomenon, the case markings) which lead the acquisition of the I-language and finally also the changes. This approach leads to further questions such as: (a) whether a universal group of cues exists (just as there is a universal group of categories) for which the child searches in his/her input, or whether the cues are 2

Chomsky (2000c: 32) notes in this regard: ‘changes in usage … may in fact be marginal changes of I-language, or changes in belief systems, here construed as (narrowly described) C[omputational]-R[epresentational] systems of the mind, which enrich the perspectives and standpoints for thought, interpretation, language use and other actions’.

220

Chapter Five

characteristics of the data of the input, so that which comprises a cue in one language is not a cue in another language for a similar grammatical phenomenon (e.g., DP denoting the patient in the accusative case); (b) what the connection is between the cues and linguistic maturation (if we accept linguistic maturation as part of our interpretation for language acquisition); (c) what principles influence that which can comprise a possible, relative or real cue; and (d) how complex the cues can be. In relation to the above, therefore, the issue of the causation of the changes is important, namely, the reason why language change appears when it appears, and why this specific language change takes place instead of another possible change (e.g., causativisation instead of anticausativisation or spread of active anticausatives instead of spread of non-active anticausatives). It is well known that a child does not inherit from his/her parents the parameters which are concerned with the specific language that s/he acquires, and that s/he does not have direct access to the grammar of the older generation. In contrast, the child develops his/her grammar based on the grammar of the previous generation, namely the output s/he receives. Syntactic change comprises grammatical reanalysis of the data by the generation which acquires the language. For the approach I followed, the presence of a structured Universal Grammar and the connection between language acquisition and change are essential. Language acquisition is the point where grammatical change occurs. Two different grammars may have a similar underlying form but a different triggering and different mechanisms (e.g., syntactic movements) to produce the structures which appear to be the same at the spell-out. The Transparency Principle, however, limits the form of the grammar; if the grammar becomes structurally very complex, then it is no longer able to be acquired. One solution in relation to the exact form of the relation between acquisition and change is that children have a tendency to analyse language in the simplest way possible (e.g., to believe that the Voice head only has the characteristic [+agent] or [-agent] and not alternatively both of them). Many linguists have maintained that economy and simplification are in essence the moving forces behind language change. I also connected linguistics economy and simplification with the change which is concerned with the separation of anticausatives and passives: the loss of one type of Voice (Voice[-agent]) results in the simplification of the system which had two types of Voice with the characteristics [+agent] and [-agent]. Linguists who supported the principle of simplification and economy, however, always had to deal with the fact that languages never end up as particularly simple systems. Although it is certain that there is no general and complete trend of simplification, it is possible to maintain that the changes are triggered by structurally local simplifications which may create complexity at other points in the system. A classic example is phonological change, which simplifies morphological endings but leads to more complex syntax (cf. the change in the verb suffixes resulting in the phonological merger of the 2nd person plural active of present tense -ete (-İIJİ) and the 3rd person singular non-active of present tense -etai (-İIJĮȚ)). We do, of course, need to be careful in the use of the terms simple and easy. The opinion I put forward that every generation attempts to find the simplest solution for the derivation of similar structures with the structures of the previous generation is very different from claiming that every generation attempts to simplify the grammar. Consequently, I reject the opinion that every generation attempts to simplify the E-language and I assume that every generation attempts to simplify the process by which the data are processed. The main reason for this is that, if we claim that something changed in the grammar because it was very complex or difficult, then we must indeed search for an answer not only as to why this complexity and difficulty arose but also how it survived for many centuries until it changed. Although I do not accept the hypothesis that there are grammatical phenomena that are so complex that there is a connection between the tendency towards simplification of the grammar and language change, I do not reject the terms simple and complex in their entirety. If the grammatical phenomena differ in regard to complexity, an examination of grammatical change and the complexity of the phenomena would indeed make for a very interesting subject of research. For example, complex phenomena (such as the two sub-categories of v or Voice) may be the result of interaction between many factors concerned with the specific language, so that they are useful as cues of language acquisition and, therefore, of stability throughout the centuries. The above are connected to a fundamental issue that concerns the approach I followed for language change: how precise is language acquisition in the case of a monolingual child? If it is precise to a high degree and if all the choices and processes by the child (reinterpretation and redetermination of structures) that I proposed above are not a reality, then language change does not arise from the linguistic domain which changes in each case. For example, phonology and morphology may be considered as external pressures on syntax. On the other hand, if acquisition is not perfect, then there is room for change which is provoked by internal to the specific domain causes, and syntactic change can arise from reanalysis of syntactic structures. Directly related to all of the above is the graduality of the changes. More specifically,

Conclusions

221

causativisation for certain verbs remains for many centuries at the first stage of the combination of the verb with clitics only, and the spread of the active anticausatives began with the Early Medieval period and is not complete even today. I translated the latter fact as a characteristic [non-act] in the root of the anticausatives which persist in not being turned into actives. If we consider the language system to be internalised, then not only must diachronic changes be connected to the internalised system but also the extension of changes (which take place through many generations) must be attributed to something different, to language use (E-language). According to the above argument, we have to disassociate the time during which the changes in the language system occurred from the extension of the changes in all of the language community. This means that the graduality of language change may be located in two areas: (a) the changes that begin as tendencies in the E-language are gradually extended in the linguistic community; (b) these changes in their turn cause changes in the I-language, which also pass through their extension stage in the linguistic community. The monograph endeavoured to trace evidence of reanalysis, change and the gradual extension of changes in the area of verb transitivity. The central role played by transitivity in the verb system and its connection to other characteristics of the grammatical system were made apparent. Concerning this connection of transitivity to other grammatical characteristics, issues which require further study are as follows: (a) the relationship between the diachrony of transitivity alternations with the diachrony of other alternations, such as ditransitive verbs (įȓȞȦ IJȠȣ īȚȐȞȞȘ IJȠ ȕȚȕȜȓȠ / dhino tu Jani to vivlio / give.ACT.PRES.1SG the.GEN Janis.GEN the.ACC book.ACC / ‘I give the book to John’ – įȓȞȦ ıIJȠȞ īȚȐȞȞȘ IJȠ ȕȚȕȜȓȠ / dhino sto Jani to vivlio / give.ACT.PRES.1SG to-the Janis the.ACC book.ACC / ‘I give John the book’) and locative alternation (ijȠȡIJȫȞȦ IJȠ ʌȜȠȓȠ ȝİ İȝʌȠȡİȪȝĮIJĮ / fortono to plio me emborevmata / load.ACT.PRES.1SG the.ACC ship.ACC with cargo / ‘I load the ship with cargo’ – ijȠȡIJȫȞȦ ıIJȠ ʌȜȠȓȠ İȝʌȠȡİȪȝĮIJĮ / fortono sto plio emborevmata / load.ACT.PRES.1SG on-the ship cargo.ACC / ‘I load cargo onto the ship’); (b) the development of other intransitive types, such as reflexives, reciprocals and middle structures; of particular theoretical interest is whether the development of these types corresponds to the development of anticausative types; (c) the reasons why the clitics comprise the first phase of causativisation of intransitive verbs; moreover, the exploration of the development of the relationship between clitics and transitivity is extremely useful; (d) the relationship between Tense and Aspect with transitivity (tenses such as the present perfect have been connected with the intransitive type in Early Greek), and the precise study of all prefixes and suffixes (e.g., -ȚȗȦ -izo, -ȦȞȦ -ono etc.) which have to do with transitivity; (e) the structure “İȓȞĮȚ + -ȝİȞȠ / ine + meno / be + PART” [İȓȞĮȚ ijIJȚĮȖȝȑȞȠ Įʌȩ / ine ftiaghmeno apo / be.3SG make.PART by / ‘it is made from’] and its relationship with the diachrony of passives, anticausatives and transitivity alternations; (f) of particular interest is the analysis of the causes of the appearance of suppletion for change-of-state verbs (e.g., ȡȓȤȞȦ rixno ‘to throw’ – ʌȑijIJȦ pefto ‘to fall’) and the development and restrictions of the periphrastic expressions of the causative meaning (țȐȞȦ IJȠȞ īȚȐȞȞȘ ȞĮ ȑȡșİȚ / kano ton Jani na erthi / make.ACT.PRES.1SG the.ACC Janis.ACC to come.NACT.PRES.3SG / ‘I make Janis come here’, ȕȐȗȦ vazo ‘to put/force’, ĮijȒȞȦ afino ‘to let’); (g) the optional omission of the object (direct or indirect) and the changes within regard to the verbs which optionally omit their objects, as well as the position, the omission and the relationship of the subject to the changes in transitivity; (h) a further examination of the diachrony of more languages would be useful in relation to the non-connection of the causativisation processes with the voice morphology, and in relation to the principle of the dissociation of the anticausatives from the passives. The analysis of diachronic changes in languages which have causative morphemes would also be useful; it needs to be ascertained whether the changes are in the same direction (productive causativisations) and whether the role of voice morphology corresponds to the role of voice morphology in languages with anticausative morphemes, as in Greek; (i) I assume that the presence of other general principles that govern language change, such as the one I proposed here (pressure for the dissociation of the anticausatives from the passives), is highly likely; a further examination of the restrictions on language change and the possible grammatical systems is essential.

BIBLIOGRAPHY

Abraham, W. 1997. “Kausativierung und Dekausativierung: zu Fragen der verbparadigmatischen Markierung in der Germania”. In T. Birkmann, H. Klingenberg, D. Nübling & E. RonnebergerSibold (eds.), Vergleichende germanische Philologie und Skandinavistik. Festschrift für Otmar Werner. Tübingen: Niemeyer, 13-28. Ackema, P. & Schoorlemmer, M. 1995. “Middles and nonmovement”. Linguistic Inquiry 26: 173-197. Adams, D. Q. 1971: “Passives and problems in Classical Greek and English”. Working Papers in Linguistics, Ohio State University 10: 1-7. Adams, M. 1987. “From Old French to the theory of pro-drop”. Natural Language and Linguistic Theory 5: 1-32. Adrados, F. 1999. Historia de la lengua griega de los origenes a nuestros dias. Madrid: Gredos. Aerts, W. J. 1965. Periphrastica. An Investigation into the Use of einai and echein as Auxiliaries or Pseudo-Auxiliaries in Greek from Homer up to the Present Day. Amsterdam: A. M. Hakkert. Åfarli, T. 1989. “Passive in Norwegian and English”. Linguistic Inquiry 20: 101-108. Agapitos, P. A. 2006. ǹijȒȖȘıȚȢ ȁȚȕȓıIJȡȠȣ țĮȚ ȇȠįȐȝȞȘȢ. Athens: ȂǿǼȉ. Aitchison, J. 1979. “The order of word order change”. Transactions of the Philological Society 197679: 43-65. —. 1991. Language Change: Progress or Decay? 2nd edition. New York: Cambridge University Press. Aksu-Koç, A. & Slobin, D. 1985. “The acquisition of Turkish.” In D. I. Slobin (ed.), The Crosslinguistic Study of Language Acquisition. Volume I: The Data. New Jersey: Hillsdale Erlbaum, 839-878. Aland, K. et al. (eds.) 19912. Novum Testamentum Graece et Latine. Institutum Studiorum Textus Novi Testamenti Monasteriensi (Westphaliae), Stuttgart: Deutsche Bibelgesellschaft. Alexiadou, A. 2001. Functional Structure in Nominals. Nominalization and Ergativity. Amsterdam/ Philadelphia: J. Benjamins. —. 2006a. “On the morpho-syntax of anticausative verbs”. Workshop on Syntax, Lexicon and Event Structure, Jerusalem, July 2006. —. 2006b. “Uniform and non-uniform aspects of pro-drop languages”. In P. Ackema, P. Brandt, M. Schoorlemmer & F. Weerman (eds.), Arguments and Agreement. Oxford: Oxford University Press, 127-158. —. 2006c. “On (anti-)causative alternations”. Handout, EALing 2006, [accessed on 1 December 2009]. Alexiadou, A. & Anagnostopoulou, E. 1998. “Parametrizing Agr: word order, verb movement, and EPP checking.” Natural Language and Linguistic Theory 16: 491-539. —. 1999a. “Tests for unaccusativity in a language without tests for unaccusativity”. In A. Moser (ed.), Proceedings of the 3rd International Conference on Greek Linguistics. Athens: Ellinika Grammata, 23-31. —. 1999b. “Non active morphology and the direction of transitivity alternations”. NELS 29: 27-40. —. 2000. “Greek syntax: a principles and parameters perspective”. Journal of Greek Linguistics 1: 171221. —. 2004. “Voice morphology in the causative-inchoative alternation: evidence for a non-unified structural analysis of unaccusatives”. In ǹ. ǹlexiadou, E. Anagnostopoulou & M. Everaert (eds.), The Unaccusativity Puzzle: Explorations of the Syntax-Lexicon Interface. Oxford: Oxford University Press, 114-136. —. 2006. “From hierarchies to features: person splits and direct inverse alternations”. In C. Boeckx (ed.), Agreement Systems. Amsterdam/ Philadelphia: J. Benjamins, 41-62. Alexiadou, A., Anagnostopoulou, E. & Schäfer, F. 2006. The properties of anticausatives crosslinguistically. In M. Frascarelli (ed.), Phases of Interpretation. Berlin/ New York: Mouton de Gruyter, 187-211. Alexiou, S. 1985. ǺĮıȓȜİȚȠȢ ǻȚȖİȞȒȢ ǹțȡȓIJȘȢ (țĮIJȐ IJȠ ȤİȚȡȩȖȡĮijȠ IJȠȣ ǼıțȠȡȚȐȜ) țĮȚ IJȠ DZıȝĮ IJȠȣ ǹȡȝȠȪȡȘ. Athens: Ermis. —. 1986. ǺȚIJıȑȞIJȗȠȣ ȀȠȡȞȐȡȠȣ ǼȡȦIJȩțȡȚIJȠȢ. Athens: Ermis. Alexiou, S. & Aposkitis, M. 1988. ǼȡȦijȓȜȘ. ȉȡĮȖȦįȓĮ īİȦȡȖȓȠȣ ȋȠȡIJȐIJıȘ. Athens: Stigmi. Allen, S. E. M. 1996. Aspects of Argument Structure Acquisition in Inuktitut. Amsterdam/ Philadelphia:

224

Transitivity Alternations in Diachrony

J. Benjamins. Alsina, A. 1992. “On the argument structure of causatives”. Linguistic Inquiry 23(4): 517-555. Amberber, M. 1996. “The transitivity of verbs of 'saying' revisited”. In J. Camacho, L. Choueiri, & M. Watanabe (eds.), The Proceedings of the Fourteenth West Coast Conference on Formal Linguistics. Stanford, CA: CSLI Publications, 1-15. —. 1997. Transitivity Alternations, Even-Types and Light Verbs. PhD Dissertation, McGill University. Anagnostopoulou, E. 1994. Clitic Dependencies in Modern Greek. PhD Dissertation, University of Salzburg. —. 1996. “Object Experiencers in Greek”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 16: 254-265. —. 2001a. “On the distinction between verbal and adjectival passives; evidence from Greek”. Workshop on Participles, Tübingen, April 17-18, 2000. —. 2001b. “Two classes of double object verbs: the role of zero-morphology”. In M. van Oostendorp & E. Anagnostopoulou (eds.), Progress in Grammar. Articles at the 20th Anniversary of the Comparison of Grammatical Models Group in Tilburg. Dutch Royal Academy Publications: Roquade. http://www.roquade.nl/meertens/progressingrammar/toc.html. —. 2003a. The Syntax of Ditransitives: Evidence from Clitics. Berlin/ NewYork: Mouton de Gruyter. —. 2003b. “Participles and voice”. In A. Alexiadou, M. Rathert & A. von Stechow (eds.), Perfect Explorations. Berlin/ New York: Mouton de Gruyter, 1-36. Anagnostou, S. 1903. ȁİıȕȚĮțȐ. Athens: Konstantinidou. Anastasiadis, V. K. 1975. “ǿıIJȠȡȓĮ țĮȚ ȖȜȫııĮ IJȘȢ ȀĮʌʌĮįȠțȓĮȢ țĮȚ IJȠ ȚįȓȦȝĮ IJȦȞ ĭĮȡȐıȦȞ”. ȂȚțȡĮıȚĮIJȚțȐ ȋȡȠȞȚțȐ 16: 150-184. —. 1976. Ǿ ıȪȞIJĮȟȘ ıIJȠ ĭĮȡĮıȚȫIJȚțȠ ȚįȓȦȝĮ IJȘȢ ȀĮʌʌĮįȠțȓĮȢ. PhD Dissertation, University of Ioannina. Anastasiadi-Symeonidi, A. 1986. “Ǿ ijȪıȘ țĮȚ Ș ʌĮȡĮȖȦȖȚțȩIJȘIJĮ IJȠȣ ıȤȘȝĮIJȚıIJȚțȠȪ ıIJȠȚȤİȓȠȣ -ʌȠȚȫ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 7: 49-70. —. 2002. ǹȞIJȓıIJȡȠijȠ ȜİȟȚțȩ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ. Thessaloniki: Institute for Modern Greek Studies (Manolis Triantafyllidis Foundation). Anastasiadi-Symeonidi, A. & Chila-Markopoulou, D. 2003. “ȈȣȖȤȡȠȞȚțȑȢ țĮȚ įȚĮȤȡȠȞȚțȑȢ IJȐıİȚȢ ıIJȠ ȖȑȞȠȢ IJȘȢ İȜȜȘȞȚțȒȢ: ȝȚĮ șİȦȡȘIJȚțȒ ʌȡȩIJĮıȘ”. In A. Anastasiadi-Symeonidi, A. Ralli & D. ChilaMarkopoulou (eds.), ȉȠ ȖȑȞȠȢ. Athens: Patakis, 13-56. Andersen, H. 1973. “Abductive and deductive change”. Language 49: 765-793. Andersen, P. K. 1993a. “Zur Diathese”. Historische Sprachforschungen 106: 177-231. —. 1993b. “Eine alternative Sprachtypologie für das Reflexiv”. Folia Linguistica 27: 107-146. —. 1994. “Remarks on Dionysius Thrax’s concept of diathesis”. Historiographia Linguistica 21: 1-37. Anderson, S. R. 1971. “On the role of deep structure in semantic interpretation”. Foundations of Language 7(3): 387-396. —. 1985. Phonology in the Twentieth Century: Theories of Rules and Theories of Representations. Chicago: University of Chicago Press. Anderson, S. R. & Lightfoot, D. W. 2002. The Language Organ: Linguistics as Cognitive Physiology. Cambridge: Cambridge University Press. —. 2006. “Biology and language: a response to Everett (2005)”. Journal of Linguistics 42(2): 377-383. Andrews, A. D. 1985. “The major functions of the noun phrase”. In T. Shopen (ed.), Language Typology and Syntactic Description. I. Clause Structure. Cambridge: Cambridge University Press, 62-154. Andriotis, N. P. 1933. “Ȇİȡȓ IJȘȢ ĮȡȤȒȢ IJȦȞ ȕȠȡİȓȦȞ ȖȜȦııȚțȫȞ ȚįȚȦȝȐIJȦȞ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ”. ǼʌİIJȘȡȓȢ ǼIJĮȚȡİȓĮȢ ǺȣȗĮȞIJȚȞȫȞ ȈʌȠȣįȫȞ 10: 340-352. —. 1937. “ȉȠ ıȪıIJȠȚȤȠȞ ĮȞIJȚțİȓȝİȞȠȞ İȚȢ IJȘȞ ȝİıĮȚȦȞȚțȒȞ țĮȚ ȞȑĮȞ İȜȜȘȞȚțȒȞ”. ǹșȘȞȐ 47: 181-202. —. 1953. “Ǿ ȤȡȒıȘ ĮȝİIJȐȕĮIJȦȞ ȡȘȝȐIJȦȞ ȦȢ ȝİIJĮȕĮIJȚțȫȞ ıIJȘ ȝİıĮȚȦȞȚțȒ țĮȚ ȞȑĮ İȜȜȘȞȚțȒ”. ȆȡȠıijȠȡȐ İȚȢ ȈIJȓȜʌȦȞĮ Ȇ. ȀȣȡȚĮțȓįȘȞ İʌȓ IJȘ İȚțȠıȚʌİȞIJĮİIJȘȡȓįȚ IJȘȢ țĮșȘȖİıȓĮȢ ĮȣIJȠȪ (1926-1951) [ǼȜȜȘȞȚțȐ, ʌĮȡȐȡIJȘȝĮ 4]. Thessaloniki, 51-62. Antonopoulou, E. 1987. Prototype Theory and the Meaning of Verbs, with Special Reference to Modern Greek Verbs of Motion. PhD Dissertation, University of London. —. 1991. Agent-defocusing Mechanisms in Spoken English: a Cognitive Explanation of Impersonalization. Parousia 16. National and Kapodistrian University of Athens. Anttila, A. 1997. Variation in Finnish Phonology and Morphology. PhD Dissertation, Stanford University. Apostolopoulos, P. 1984. La langue du roman byzantin “Callimaque et Chrysorrhoé”. Athens. Arad, M. 1998. VP-Structure and the Syntax-Lexicon Interface. PhD Dissertation, University College London.

Bibliography

225

—. 2002. “Universal features and language-particular morphemes”. In A. Alexiadou (ed.), Theoretical Approaches to Universals. Amsterdam/ Philadelphia: J. Benjamins, 15-40. Arlotto, A. 1972. Introduction to Historical Linguistics. Boston: Houghton-Mifflin. Arndt, W. F. & Wilbur Gingrich, F. 1957. A Greek-English Lexicon of the New Testament and Other Early Christian Literature. Chicago: University of Chicago Press. Aronoff, M. 1994. Morphology by itself. Stems and Inflectional Classes. Linguistic Inquiry Monographs. Cambridge, MA: MIT Press. Asenova, P. 2000. “Quelques observations sur l’emploi de la diathèse non-active dans les langues balkaniques”. In Ch. Tzitzilis & Ch. Symeonidis (eds.), ǺĮȜțĮȞȚțȒ īȜȦııȠȜȠȖȓĮ. ȈȣȖȤȡȠȞȓĮ țĮȚ įȚĮȤȡȠȞȓĮ. Thessaloniki, 27-39. Babiniotis, G. 1975. ȃİȦIJȑȡĮ īȜȦııȠȜȠȖȓĮ. Athens. —. 1980. ĬİȦȡȘIJȚțȒ īȜȦııȠȜȠȖȓĮ. ǼȚıĮȖȦȖȒ ıIJȘ ıȪȖȤȡȠȞȘ īȜȦııȠȜȠȖȓĮ. Athens. —. 1985/2002. ȈȣȞȠʌIJȚțȒ ȚıIJȠȡȓĮ IJȘȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ, ȝİ İȚıĮȖȦȖȒ ıIJȘȞ ǿıIJȠȡȚțȠıȣȖțȡȚIJȚțȒ īȜȦııȠȜȠȖȓĮ. Athens. —. 1985. ǿıIJȠȡȚțȒ ȖȡĮȝȝĮIJȚțȒ IJȘȢ ĮȡȤĮȓĮȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. I: ĭȦȞȠȜȠȖȓĮ. Athens. —. 1986. ǼȚıĮȖȦȖȒ ıIJȘ ȈȘȝĮıȚȠȜȠȖȓĮ. Athens. Baker, M. 1985. “The mirror principle and morphosyntactic explanation”. Linguistic Inquiry 16: 373415. —. 1988. Incorporation: a Theory of Grammatical Function Changing. Chicago: University of Chicago Press. —. 1995. “Lexical and nonlexical noun incorporation”. In U. Egli, P. E. Pause, C. Schwarze, A. von Stechow & G. Wienold (eds.), Lexical Knowledge in the Organization of Language. Amsterdam/ Philadelphia: J. Benjamins, 3-33. —. 1997. “Thematic roles and syntactic structure”. In L. Haegeman (ed.), Elements of Grammar, Dordrecht: Kluwer, 73-137. —. 2003. Lexical Categories: Verbs, Nouns, and Adjectives. Cambridge: Cambridge University Press. Baker, M., Johnson, K. & Roberts, I. 1989. “Passive arguments raised”. Linguistic Inquiry 20(2): 219252. Bakker, E. J. 1994. “Voice, aspect and aktionsart. Middle and passive in Ancient Greek”. In B. Fox & P. J. Hopper (eds.), Voice: Form and Function. Amsterdam/ Philadelphia: J. Benjamins, 23-47. Bakker, W. F. & van Gemert, A. F. 1987. ȂĮȞȩȜȘȢ ǺĮȡȠȪȤĮȢ: ȞȠIJĮȡȚĮțȑȢ ʌȡȐȟİȚȢ. ȂȠȞĮıIJȘȡȐțȚ ǹȡĮțȜȓȠȣ (1597-1613). Rethymne. —. 1988. ǿıIJȠȡȓĮ IJȠȣ ǺİȜȚıĮȡȓȠȣ. Athens: ȂǿǼȉ. —. 1996. Ǿ șȣıȓĮ IJȠȣ ǹȕȡĮȐȝ. Iraklion: University of Crete Press. Baldi, P. 1977. “Remarks on the Latin R-form verbs”. Zeitschrift für vergleichende Sprachforschung 90: 222-257. Bartelink, G. J. M. 1971. Callinicos. Vie d’Hypatios. Paris: Cerf. —. 1974. Palladio. La Storia Lausiaca. Verona: Fondazione Lorenzo Valla. —. 1994. Athanase d’Alexandrie: Vie d’Antoine. Paris: Cerf. Basea-Bezantakou, C. 1992. ȈȘȝĮıȚȠȜȠȖȚțȒ șİȫȡȘıȘ IJȦȞ ȡȘȝȐIJȦȞ țȚȞȒıİȦȢ IJȘȢ ȞȑĮȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. Athens: Kardamitsas. Bassols de Climent, M. 1948. Sintaxis histórica de la lengua latina. II.1. Las formas personales del verbo. Barcelona: Escuela de Filología. Bauer, B. 2000. Archaic Syntax in Indo-European. The Spread of Transitivity in Latin and French. Berlin/ New York: Mouton de Gruyter. Beaton, R. 1996. The Medieval Greek Romance. 2nd edition. London: Routledge. Beaucamp, J. (ed.) 2004. Recherches sur la chronique de Jean Malalas, I. Actes du Colloque La Chronique de Jean Malalas: genèse et transmission. Paris: Association des Amis du Centre d’Histoire et Civilisation de Byzance. Bechert, J. 1964. Die Diathesen von “idein” und “horan” bei Homer. München. Beck, H.-G. 1971/1988. Geschichte der byzantinischen Volksliteratur. Munich / ǿıIJȠȡȓĮ IJȘȢ ȕȣȗĮȞIJȚȞȒȢ įȘȝȫįȠȣȢ ȜȠȖȠIJİȤȞȓĮȢ. Athens: MIET. Beekes, R. S. P. 1995. Comparative Indo-European Linguistics: an Introduction. Amsterdam/ Philadelphia: J. Benjamins. Belletti, A. & Rizzi, L. 1987. “The case of unaccusatives”. Linguistic Inquiry 6(3): 291-352. —. 1988. “Psych-verbs and theta-theory”. Natural Language and Linguistic Theory 6: 291-352. Benedicto, E. (ed.) 2000. The UMass Indigenous Languages Volume. UMass Occasional Papers in Linguistics 20. Amherst: University of Massachusetts. Bentley, D. & Eythórsson, T. 2004. “Auxiliary selection and the semantics of unaccusativity”. Lingua 114(4): 447-471.

226

Transitivity Alternations in Diachrony

Benua, L. 2000. “Yup’ik antipassive and the AspP hypothesis”. In E. Benedicto (ed.), The UMass Indigenous Languages Volume. UMass Occasional Papers in Linguistics 20. Amherst, MA: UMass, GLSA, 107-138. Benua, L. & Borer, H. 1996. The Passive/Anti-Passive Alternation. Paper presented at GLOW 1996. Benveniste, E. 1950. “La phrase nominale”. Bulletin de la Société de Linguistique de Paris 46: 19-36. Bernd, H. 1993a. “Bekommen ohne etwas zu bekommen: zur Grammatikalisierung des Dativpassivs”. Sprache und Literatur in Wissenschaft und Unterricht 71: 26-33. Bernd, H. 1993b. Auxiliaries. Cognitive Forces and Grammaticalization. Oxford: Oxford University Press. Beševliev, V. 1963. Die protobulgarischen Inschriften. Berliner byzantinistische Arbeiten Bd 23. Berlin: Akademie Verlag. BGU = Aegyptische Urkunden aus den Königlichen (Staatlichen) Museen zu Berlin, Griechische Urkunden. Berlin. Vol. 1-18 (1895-2000). BGV Æ Legrand, É. Bhatt, R. & Embick, D. 2004. “Causative derivations in Hindi”. Ms. University of Massachusetts at Amherst & University of Pennsylvania. Bittner, M. 1999. “Concealed causatives”. Natural Language Semantics 7: 1-78. Bittner, M. & Hale, K. 1996. “The structural determination of case and agreement”. Linguistic Inquiry 27(1): 1-68. Blass, F., Debrunner, A. & Funk, R. W. 1961. A Greek Grammar of the New Testament and Other Early Christian Literature. Cambridge: Cambridge University Press. Blass, F., Debrunner, A. & Rehkopf, F. 1984. Grammatik des neutestamentlichen Griechisch. Göttingen: Vandenhoeck und Ruprecht. Bobaljik, J. & Wurmbrand, S. 2008. “Case in GB/ Minimalism”. In ǹ. Malchukov & A. Spencer (eds.), Handbook of Case. Oxford: Oxford University Press, 44-58. Bonnet, M. 1898. Acta Apostolorum Apocrypha. Vol. 2.1. Leipzig: Mendelssohn. Borer, H. 1994. “The projection of arguments”. In E. Benedicto & J. Runner (eds.), Functional Projections. UMass Occasional Papers in Linguistics 17. Amherst, MA: UMass, GLSA, 19-47. —. 1996. “The construct in review”. In J. Lecarme, J. Lowenstamm & U. Shlonsky (eds.), Studies in Afroasiatic Grammar. Holland Academic Graphics, 30-61. —. 1998a. “Morphology and syntax”. In A. Spencer & A. Zwicky (eds.), The Handbook of Morphology. Oxford: Blackwell, 151-190. —. 1998b. “Passive without theta grids”. In S. G. Lapointe, D. K. Brentari & P. M. Farrell (eds.), Morphological Interfaces and its Relations to Phonology and Syntax. Stanford, CA: CSLI Publications, 60-99. Bousboukis, A. D. 1994-5. “ĭȦȞȑȢ țĮȚ įȚĮșȑıİȚȢ ȡȘȝȐIJȦȞ ıIJȘ ȃȑĮ ǼȜȜȘȞȚțȒ țĮȚ IJȘȞ ǿIJĮȜȚțȒ”. ȁİȟȚțȠȖȡĮijȚțȩȞ ǻİȜIJȓȠȞ 19: 103-137. Bowerman, M. 1974. “Learning the structure of causative verbs: a study in the relationship of cognitive, semantic, and syntactic development”. Papers and Report on Child Language Development 8: 142-178. Stanford, CA: Stanford University. —. 1982. “Evaluating competing linguistic models with language acquisition data: implications of developmental errors with causative verbs”. Quaderni di Semantica 3: 5-66. Bowers, J. 1991. “The syntax and semantics of nominals”. In S. Moore & A. Wyner (eds.), Proceedings from SALT I, Cornell Working Papers in Linguistics 10: 1-30. —. 1993. “The syntax of predication”. Linguistic Inquiry 24: 591-656. —. 1997. “A binary analysis of resultatives”. Proceedings of the 1997 Linguistics Society Conference, Texas Linguistic Forum 38: 43-58. —. 2002: “Transitivity”. Linguistic Inquiry 33(2): 183-224. Braine, M. D. S., Brody, R. E., Fish, S. M., Weisberger, M. J., & Blum, M. 1990. “Can children use a verb without exposure to its argument structure?” Journal of Child Language 17: 313-342. Brekke, M. 1988. “The Experiencer constraint”. Linguistic Inquiry 19(2): 169-180. Brisson, C. 1994. “The licensing of unexpressed objects in English verbs”. Chicago Linguistic Society 30: 90-102. Brody, M. 1989. “Old English impersonals and the theory of grammar”. UCL Working Papers in Linguistics 1: 262-294. London: University College London. Brousseau, A.-M. & Ritter, E. 1991. “A non-unified analysis of agentive verbs”. In D. Bates (ed.), The Proceedings of the Tenth West Coast Conference on Formal Linguistics. Stanford, CA: CSLI Publications, 53-65. Browning, R. 1978. “The language of Byzantine literature”. In S. J. Vryonis (ed.), The Past in Medieval and Modern Greek Culture. Malibu: Undena, 104-133.

Bibliography

227

—. 1983. Medieval and Modern Greek. Cambridge: Cambridge University Press. Brugmann, K. 1897-1916/1930. Grundriss der vergleichende Grammatik der indogermanischen Sprache. 2nd edition. Berlin: Mouton de Gruyter. Brugmann, K. & Thumb, A. 1913. Griechische Grammatik. 4th edition. München: von Müller. Brunot, F. 1953/1965. La pensée et la langue. Paris: Masson. Brunot, F. 1966. Histoire de la langue française des origines à nos jours. ȉome 3. La formation de la langue classique 1600 – 1660. Paris: Colin. Bubenik, V. 1989. Hellenistic and Roman Greece as a Sociolinguistic Area. Current Issues in Linguistic Theory 57. Amsterdam/ Philadelphia: J. Benjamins. —. 1993. “Dialect contact and koineization: the case of Hellenistic Greek”. International Journal of the Sociology of Language 99: 9-23. —. 1996. “Toward a theory of syntactic change”. Diachronica 13(1):155-170. Buck, C. D. 1955. The Greek Dialects. 2nd edition. Chicago: University of Chicago Press. —. 1910. Introduction to the Study of the Greek Dialects. Boston: Ginn. Budwig, N. 1990. “The linguistic marking of non-prototypical agency: an exploration into children’s use of passives”. Linguistics 28: 1221-1252. Burzio, L. 1986. Italian Syntax: a Government-Binding Approach. Dordrecht: Reidel. Bussmann, H. 1983. Lexikon der Sprachwissenschaft. Stuttgart: Kröner. Butt, M. & Geuder, W. (eds.) 1998. The Projection of Arguments: Lexical and Compositional Factors. Stanford, CA: CSLI Publications. Bybee, J. L. 1985. Morphology: a Study of the Relation between Meaning and Form. Typological Studies in Language 9. Amsterdam/ Philadelphia: J. Benjamins. —. 1988. “Semantic substance vs. contrast in the development of grammatical meaning”. Proceedings of the 14th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 247-264. Bybee, J. L. & Pagliuca, W. 1985. “Cross-linguistic comparison and the development of grammatical meaning”. In J. Fisiak (ed.), Historical Semantics, Historical Word-Formation. Trends in Linguistics Studies and Monographs 29. Berlin/New York: Mouton de Gruyter, 59-83. Bybee, J. L., Perkins, R. & Pagliuca, W. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Campos, H. 1991a. “Preposition stranding in Spanish”. Linguistic Inquiry 22(4): 741-750. —. 1991b. “Silent objects and subjects in Spanish”. In H. Campos & F. Martínez-Gil (eds.), Current Studies in Spanish Linguistics. Washington, D.C.: Georgetown University Press, 117-141. Campos, H., Charnitski, M. & Woodard, W. 1990. “Silent prepositional phrases in Spanish”. The Georgetown Journal of Languages and Linguistics 1(1): 13-24. —. 1991. “Silent prepositional phrases and Pro-PP in Spanish”. Probus 3(1): 1-21. Canac Marquis, R. 2000/2003. “Asymmetry, syntactic objects and the mirror generalization”. Ms. Simon Fraser University / Published in A. M. DiSciullo (ed.), Asymmetry in Grammar. Amsterdam/ Philadelphia: J. Benjamins, 65-94. Carrier, J. & Randall, J. H. 1992. “The argument structure and syntactic structure of resultatives”. Linguistic Inquiry 23(2): 173-234. Carrier-Duncan, J. 1985. “Linking of thematic roles in derivational word formation”. Linguistic Inquiry 16: 1-34. Catsimali, G. 1990. Case in Modern Greek. Implications for Clause Structure. PhD Dissertation, University of Reading. Cennamo, M. 1993. The Reanalysis of Reflexives: a Diachronic Perspective. Napoli: Liguori. —. 1998a. “The loss of the voice dimension between Late Latin and Early Romance”. In M. S. Schmid, J. R. Austin & D. Stein (eds.), Historical Linguistics 1997. Amsterdam/ Philadelphia: J. Benjamins, 77-100. —. 1998b. “Costruzioni passive ed impersonali in Veneziano e in Napoletano antico”. In A. Englebert et al. (eds), Actes du XXIIe Congrès International de Linguistique et Philologie Romanes, vol. 2: Les Nouvelles Ambitions de la Linguistique Diacronique. Tübingen: Niemeyer, 91-103. —. 1999. “Late Latin pleonastic reflexives and the unaccusative hypothesis”. Transactions of the Philological Society 97: 103-150. —. 2000a. “(In)transitivity and object marking: some current issues”. In G. Fiorentino (ed.), Romance Objects. Berlin/ New York: Mouton de Gruyter, 49-104. —. 2000b. “Patterns of ‘active’ syntax in Late Latin pleonastic reflexives”. In J. C. Smith & D. Bentley (eds.), Historical Linguistics 1995. Amsterdam/ Philadelphia: J. Benjamins, 35-56. Chantraine, P. 1953. Grammaire homérique. II: Syntaxe. Paris: Klincksieck. —. 1970-1974. Dictionnaire étymologique de la langue grecque. Paris: Klincksieck.

228

Transitivity Alternations in Diachrony

Charalambakis, Ch. 1976. “Phthano”: ein Beitrag zur griechischen Wortforschung. Heidelberg. —. 1980. ȀİijȐȜĮȚĮ Įʌȩ IJȘȞ ǿȞįȠİȣȡȦʌĮȧțȒ īȜȦııȠȜȠȖȓĮ țĮȚ IJȘȞ ȚıIJȠȡȓĮ IJȘȢ ĮȡȤĮȓĮȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. Rethymne. —. 1981. ǿıIJȠȡȓĮ IJȘȢ ȝİIJĮțȜĮıȚțȒȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ: Ș EȜȜȘȞȚıIJȚțȒ KȠȚȞȒ. Rethymne. —. 1993. “Zum Sprachverfall in der griechischen Antike”. In G. W. Most, H. Petersmann & A. M. Ritter (eds.), Festschrift für Albecht Dihle zum 70. Geburtstag. Göttingen, 36-45. —. 1999. “ȈȣȞIJĮțIJȚțȑȢ țĮȚȞȠIJȠȝȓİȢ IJȘȢ ȀȠȚȞȒȢ”. In M. Z. Kopidakis (ed.), ǿıIJȠȡȓĮ IJȘȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. Athens: E.L.I.A., 100-101. Chatzidakis, G. N. 1907. ȂİıĮȚȦȞȚțȐ țĮȚ ȃȑĮ ǼȜȜȘȞȚțȐ. Vol. 2. Athens: Marasli. —. 1915/1930/1992. ǹțĮįȘȝİȚțȐ ĮȞĮȖȞȫıȝĮIJĮ. Athens: Vas. G. Vassiliou. —. 1927. “Ȇİȡȓ ȝİIJĮȕȚȕĮıIJȚțȒȢ įȚĮșȑıİȦȢ”. Proceedings of the Academy of Athens 1927: 16-21. —. 1934. īȜȦııȠȜȠȖȚțĮȓ ȑȡİȣȞĮȚ, Vol. 1. Athens. —. 1977. īȜȦııȠȜȠȖȚțĮȓ ȑȡİȣȞĮȚ, Vol. 2. ȀȑȞIJȡȠȞ ȈȣȞIJȐȟİȦȢ IJȠȣ ǿıIJȠȡȚțȠȪ ȁİȟȚțȠȪ, ȁİȟȚțȠȖȡĮijȚțȩȞ ǻİȜIJȓȠȞ, ȆĮȡȐȡIJȘȝĮ 2. Athens: Academy of Athens. Chierchia, G. 1982. “Nominalization and Montague Grammar”. Linguistics and Philosophy 5: 303-354. —. 1985. “Formal semantics and the grammar of predication”. Linguistic Inquiry 16: 417-443. —. 1989/2004. “A semantics for unaccusatives and its syntactic consequences”. Ms., Cornell University, Ithaca, New York. / Published in A. Alexiadou, E. Anagnostopoulou & M. Everaert (eds.), The Unaccusativity Puzzle: Explorations of the Syntax-Lexicon Interface. Oxford: Oxford University Press, 22-59. Chila-Markopoulou, D. 1989. “‘EȜȜİȚʌIJȚțȐ’ ijĮȚȞȩȝİȞĮ ıIJȘ ȃȑĮ ǼȜȜȘȞȚțȒ. Ǿ ʌİȡȓʌIJȦıȘ IJȠȣ ĮȞIJȚțİȚȝȑȞȠȣ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 9: 389-406. —. 1990-91. “ȆȡȠȕȜȒȝĮIJĮ įȚĮȤȡȠȞȚțȒȢ ıȪȞIJĮȟȘȢ: ȠȚ İȜİȪșİȡİȢ ĮȞĮijȠȡȚțȑȢ ʌȡȠIJȐıİȚȢ ıIJĮ ȝİıĮȚȦȞȚțȐ țĮȚ ȞȑĮ İȜȜȘȞȚțȐ”. Glossologia 9-10: 13-42. —. 1994. “Problems in diachronic syntax: free relatives in Medieval and Modern Greek”. In I. Philippaki-Warburton, K. Nicolaidis & M. Sifianou (eds.), Themes in Greek Linguistics. Papers from the First International Conference on Greek Linguistics, Reading, September 1993. Amsterdam/ Philadelphia: J. Benjamins, 413-419. —. 1999. “ȉȠ 'ĮȩȡȚıIJȠ ȐȡșȡȠ' IJȘȢ İȜȜȘȞȚțȒȢ: įȚĮȤȡȠȞȚțȒ ʌȡȠıȑȖȖȚıȘ”. In A. Moser (ed.), Proceedings of the 3rd International Conference on Greek Linguistics. Athens: Ellinika Grammata, 537-545. —. 2000. “The indefinite article in Greek: a diachronic approach”. Glossologia 11-12: 111-130. —. 2001. “'ĬȑȜȦ ȞĮ... ĮȜȜȐ įİȞ șĮ...'. ĭĮȚȞȩȝİȞĮ ȖȡĮȝȝĮIJȚțȠʌȠȓȘıȘȢ țĮȚ IJȡȠʌȚțȒȢ ʌȠȜȣİțȝİIJȐȜȜİȣıȘȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 21: 822-833. Chomsky, N. 1957. Syntactic Structures. The Hague: Mouton de Gruyter. —. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. —. 1981. Lectures on Government and Binding. Dordrecht: Foris. —. 1986. Barriers. Cambridge, MA: MIT Press. —. 1993. “A minimalist program for linguistic theory”. In K. Hale & S. J. Keyser (eds.), The View from Building 20. Essays in Linguistics in Honor of Sylvain Bromberger. Cambridge, MA: MIT Press, 149. —. 1995. The Minimalist Program. Cambridge, MA: MIT Press. —. 1999/2001. “Derivation by phase”. MIT Occasional Papers in Linguistics 18 / Published also in M. Kenstowicz (ed.), Ken Hale: a Life in Language. Cambridge, MA: MIT Press, 1-52. —. 1999. “Language design”. In S. Griffiths (ed.), Predictions: 30 Great Minds on the Future. Oxford: Oxford University Press, 30-32. —. 2000. “Minimalist inquiries: the framework”. In R. Martin, D. Michaels & J. Uriagereka (eds.), Step by Step: Essays in Minimalist Syntax in Honor of Howard Lasnik. Cambridge, MA: MIT Press, 89-155. —. 2005. “Three factors in language design”. Linguistic Inquiry 36: 1-22. —. 2005/2008. “On phases”. Ms, MIT. / Published in C. Otero et al. (eds.), Foundational Issues in Linguistic Theory. Cambridge, MA: MIT Press, 133-166. —. 2006. Language and Mind. Cambridge: Cambridge University Press. —. 2006/2007. “Approaching UG from below”. Ms., MIT. / Published in U. Sauerland & H. M. Gärtner (eds.), Interfaces + Recursion = Language?. Berlin/ New York: Mouton de Gruyter, 1-29. Christidis, A.-Ph. 2005. ǿıIJȠȡȓĮ IJȘȢ ĮȡȤĮȓĮȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. ǹȡȤĮȚȠȖȜȦııȓĮ țĮȚ ĮȡȤĮȚȠȖȞȦıȓĮ ıIJȘ ȂȑıȘ ǼțʌĮȓįİȣıȘ 1. Institute for Modern Greek Studies (Manolis Triantafyllidis Foundation). —. (ed.) 2007. A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press. Cinque, G. 1988. “On si constructions and the theory of Arb”. Linguistic Inquiry 19: 521-581. Clairis, Ch. & Babiniotis, G. 1999 (in collaboration with: E. Bakakou, A. Moser & S. Skopeteas).

Bibliography

229

īȡĮȝȝĮIJȚțȒ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ. 2. ȉȠ ȡȒȝĮ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ: Ș ȠȡȖȐȞȦıȘ IJȠȣ ȝȘȞȪȝĮIJȠȢ. Athens: Ellinika Grammata. —. 2005 (in collaboration with: E. Bakakou, A. Moser & S. Skopeteas). īȡĮȝȝĮIJȚțȒ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ. ǻȠȝȠȜİȚIJȠȣȡȖȚțȒ – ǼʌȚțȠȚȞȦȞȚĮțȒ. Athens: Ellinika Grammata. Clark, E. V. 1992. “Conventionality and contrast: pragmatic principles with lexical consequences”. In E. F. Kittay & A. Lehrer (eds.), Frames, Fields, and Contrasts: New Essays in Semantic and Lexical Organization. Hillsdale, NJ: Lawrence Erlbaum Associates, 171-188. Clark, R. & Roberts, I. 1993. “A computational model of language learnability and language change”. Linguistic Inquiry 24: 299-345. Clogg, R. 1986. A Short History of Modern Greece. Cambridge: Cambridge University Press. Collins C. T. 1997. “Argument sharing in serial verb constructions”. Linguistic Inquiry 28: 461-497. —. 2005. “A smuggling approach to the passive in English”. Syntax 8(2): 81-120. Comrie, B. 1985. Tense. Cambridge: Cambridge University Press. —. 1989. Language Universals and Linguistic Typology: Syntax and Morphology. Oxford: Blackwell. Comrie, B. & Polinsky, M. (eds.) 1993. Causatives and Transitivity. Amsterdam/ Philadelphia: J. Benjamins. Cooper, G. L. 1998. Attic Greek Prose Syntax. Ann Arbor: University of Michigan Press. Cooreman, A., Fox, B. & Givón, T. 1984. “The discourse definition of ergativity”. Studies in Language 8: 1-34. Croft, W. 1990. Typology and Universals. Cambridge: Cambridge University Press. —. 1991. Syntactic Categories and Grammatical Relations. Chicago: University of Chicago Press. —. 1993. “The role of domains in the interpretation of metaphors and metonymies”. Cognitive Linguistics 4: 335-370. Cruse, D. A. 1973. “Some thoughts on agentivity”. Journal of Linguistics 9: 11-23. Curme, G. O. 1947. English Grammar. New York: Barnes & Noble Books. Cuxac, C. 1987. “Transitivité en LSF, structures de l’iconicité”. In D. François-Geiger (ed.), La transitivité et ses corrélats. Travaux/ Centre de Linguistique 1. Paris: Université René Descartes, 15-50. Davidse, Ȁ. 1991. Categories of Experiential Grammar. PhD Dissertation, University of Leuven. —. 1992. “Transitivity/ergativity: the Janus-headed grammar of actions and events”. In M. Davies & L. Ravelli (eds.), Advances in Systemic Linguistics: Recent Theory and Practice. London: Pinter, 105135. Davidse, K. & Geyskens, S. 1998. “Have you walked the dog yet? The ergative causativization of intransitives”. WORD 49(2): 155-180. Davis, H. 2000. “Salish evidence on the causative-inchoative alternation”. In W. U. Dressler, O. E. Pfeiffer, M. Pöchtrager & J. R. Rennison (eds.), Morphological Analysis in Comparison. Amsterdam/ Philadelphia: J. Benjamins, 25-60. Davis, H. & Demirdache, H. 1995. Agents and Events. Paper presented at the GLOW Colloquium, University of Tromsø, Norway. Dawkins, R. M. 1916. Modern Greek in Asia Minor. Cambridge: Cambridge University Press. —. 1932. Leontios Makhairas. Recital concerning the Sweet Land of Cyprus entitled “Chronicle”. Vol. I. Oxford: Clarendon Press. De Boor, C. 1883/1963. Theophanis Chronographia. Hildesheim: Georg Olms. DeLancey, S. 1984a. “Notes on agentivity and causation”. Studies in Language 8(2): 181-213. —. 1984b. “Transitivity and ergative case in Lhasa Tibetan”. Proceedings of the 10th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 131-140. —. 1987. “Transitivity in grammar and cognition”. In R. Tomlin (ed.), Discourse Relations and Cognitive Units. Amsterdam/ Philadelphia: J. Benjamins, 53-68. —. 1990. “Cross-linguistic evidence for the structure of the Agent prototype”. Papers and Reports on Child Language Development 29: 141-147. Department of Linguistics, Stanford. —. 1991. “Event construal and case role assignment”. Proceedings of the 17th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 338-353. Delbrück, B. 1888/1968. Altindische Syntax. Syntaktische Forschungen 5. Darmstadt: Wissenschaftliche Buchgesellschaft. —. 1897. Vergleichende Syntax der indogermanischen Sprachen. Zweiter Teil. Strassburg: Trübner. Demonte, V. 1986. “Predication and passive”. In I. Bordelois, H. Contreras & K. Zagona (eds.), Generative Studies in Spanish Syntax. Dordrecht: Foris, 51-66. —. 1988. “Remarks on secondary predicates: c-command, extraction and reanalysis”. The Linguistic Review 6: 1-39. —. 1991a. “Linking and case: the case of prepositional verbs”. In T. Morgan & C. Laeufer (eds.),

230

Transitivity Alternations in Diachrony

Theoretical Analyses in Romance Linguistics. Amsterdam/ Philadelphia: J. Benjamins, 413-450. —. 1991b. Detrás de la palabra. Estudios de gramática del espanol. Madrid: Alianza Universidad. —. (ed.) 1994. Gramática del español. México: El Colegio de México. Denis, A.-M. 1970. Fragmenta pseudepigraphorum quae supersunt Graeca. Leiden: Brill. Déroche, V. 1994. “L’Apologie contre les Juifs de Léontios de Néapolis”. Travaux et Mémoires 12: 45104. Díaz Vera, J. 2000. “The development of causation in Old English and its interaction with lexical and syntactic diachronic processes”. Cuadernos de investigación filológica XXVI: 17-38. Dieterich, K. 1898/1970. Untersuchungen zur Geschichte der Griechischen Sprache von der Hellenistischen Zeit bis zum 10. Jahrhundert N. Chr. Hildesheim: Georg Olms. Dik, H. 1995. Word Order in Ancient Greek: a Pragmatic Account of Word Order Variation in Herodotus. Amsterdam studies in classical philology 5. Amsterdam: J. C. Gieben. Dimaras, K. Th. 1948. ǿıIJȠȡȓĮ IJȘȢ ȞİȠİȜȜȘȞȚțȒȢ ȜȠȖȠIJİȤȞȓĮȢ. Athens: Ikaros. Dimitriadis, ǹ. 1994a. “Clitics and island-insensitive object drop”. Studies in the Linguistic Sciences 24(1-2): 153-169. —. 1994b. “Clitics and object drop in Modern Greek”. In C. Giordano & D. Ardron (eds.), Proceedings of the 6th Student Conference in Linguistics (SCIL-6) (MIT Working Papers in Linguistics 23). University of Rochester, New York, 95-115. Dindorf, L. 1831. Ioannis Malalae Chronographia. Bonn. —. 1832. Chronicon paschale. Bonn. Dixon, R. 1987. “Studies in ergativity: introduction”. Lingua 71: 1-16. —. 1994. Ergativity. Cambridge: Cambridge University Press. —. 1996. Australian Languages. Cambridge UK: Cambridge Language Surveys. Dobrovie-Sorin, C. 1998. “Impersonal se constructions in Romance and the passivization of unergatives”. Linguistic Inquiry 29(3): 399-437. Dolinina, I. B. 1989. “Teoreticheskie aspekty glagol'noj mnozhestvennosti”. Tipologiya iterativnyh konstrukcij. Leningrad. Doris, K. 2007. “Ǿ ıȤȑıȘ IJȠȣ İȜȜȘȞȚțȠȪ ȖȜȦııȚțȠȪ ȚįȚȫȝĮIJȠȢ IJȘȢ ȋȚȝȐȡĮȢ ȝİ IJĮ ȐȜȜĮ ȞİȠİȜȜȘȞȚțȐ ȚįȚȫȝĮIJĮ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 17: 198-209. Doron, E. 2003. “Agency and voice: the semantics of the Semitic templates”. Natural Language Semantics 11(1): 1-67. Dover, K. J. 2000. Greek Word Order. Bristol: Bristol Classical Press. —. 1987. Greek and the Greeks. Oxford: Blackwell. Dowty, D. R. 1979. Word Meaning and Montague Grammar. Dordrecht: Reidel. —. 1989/1991. “Thematic proto-roles and argument selection”. Ms. / Language 67(3): 547-619. Drachman, G. 1997. “Some properties of clitics (with special reference to Modern Greek)”. In A. Alexiadou & T. A. Hall (eds.), Studies on Universal Grammar and Typological Variation. Amsterdam/ Philadelphia: J. Benjamins, 219-248. —. 2005. “A note on ‘shared’ allomorphs”. Journal of Greek Linguistics 6: 5-38. —. 2009. “Meaning-variation and change in morphology”. In N. Lavidas, E. Nouchoutidou & M. Sionti (eds.), New Perspectives in Greek Linguistics. Newcastle: Cambridge Scholars Publishing, 134. Drakakis, M. G. 2004. ȂȚȤĮȒȜ ȂĮȡȐȢ, ȞȠIJȐȡȚȠȢ ȋȐȞįĮțĮ. ȀĮIJȐıIJȚȤȠ 149, ȉȩȝȠȢ ǹǯ [16/1 – 30/3 1549]. Iraklion: Vikelea Dimotiki Vivliothiki. Dresher, E. 1999. “Charting the learning path: cues to parameter setting”. Linguistic Inquiry 30: 27-67. Drettas, G. 2005. “Un retour aux diathèses: en réexaminant la voix moyenne”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 25: 185-196. Drody, M. & Manzini, M. 1988. “On implicit arguments”. In R. Kempson (ed.), Mental Representations. Cambridge University Press, 105-130. Duhoux, Y. 1998. “Grec écrit et grec parlé. Une étude contrastive des particules aux Ve-IVe siècles”. In A. Rijksbaron (ed.), New Approaches to Greek Particles. Amsterdam: J. C. Gieben, 15-48. —. 1988. “Le système verbal grec: l’état mycénien”. In A. Rijksbaron, H. A. Mulder & G. C. Wakker (eds.), In the Footsteps of Raphael Kühner. Amsterdam: J. C. Gieben, 121-134. —. 1992/2000. Le verbe grec ancien: éléments de morphologie et de syntaxe historiques. Louvain-laNeuve, Belgium: Peeters. —. 1997. “Quelques idées reçues, et néanmoins fausses, sur les particules grecques”. L’Antiquité classique 66: 281-288. Egea, J. M. 1996. La Crónica de Morea. Madrid: Consejo superior de Investigaciones Científicas. Eideneier, H. 1977. Spanos: eine byzantinische Satire in der Form einer Parodie. Berlin/ New York: Mouton de Gruyter.

Bibliography

231

—. 1991. Ptochoprodromos. Neograeca Medii Aevi V. Köln: Romiosini. —. 1999. Von Rhapsodie zu Rap. Aspekte der griechischen Sprachgeschichte von Homer bis heute. Tübingen: Narr. Elliott, W. N. 1986. “On the derivation of en-clitics”. Syntax and Semantics 19: 97-121. Embick, D. 1996. “Causativization in Hupa”. Proceedings of the 22th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 83-94. —. 1998. “Voice systems and the syntax-morphology interface”. MIT Working Papers in Linguistics 32: 41-72. —. 2000. “Features, syntax, and categories in the Latin Perfect”. Linguistic Inquiry 31(2): 185-230. —. 2003. “Locality, listedness and morphological identity”. Studia Linguistica 57: 143-169. —. 2004. “Unaccusative syntax and verbal alternations”. In A. Alexiadou, E. Anagnostopoulou & M. Everaert (eds.), The Unaccusativity Puzzle: Explorations of the Syntax-Lexicon Interface. Oxford: Oxford University Press, 137-158. Emonds, J. E. 1991. “Subcategorization and syntax-based theta-role assignment”. Natural Language and Linguistic Theory 9: 369-429. —. 2000. Lexicon and Grammar: the English Syntacticon. Berlin/ New York: Mouton de Gruyter. Engelhardt, M. 2000. “The projection of argument-taking nominals”. Natural Language and Linguistic Theory 18(1): 41-88. Epstein, S. 2000. Essays in Syntactic Theory. London/ New York: Routledge. Epstein, S., Groat, E., Kawashima, R. & Kitahara, H. 1998. A Derivational Approach to Syntactic Relations. New York: Oxford University Press. Ernout, A. 1908-9. “Recherches sur l’emploi du passif latin”. MSL 15: 273-333. —. 1909. “De l’emploi du passif dans la Mulomedichina Chironis”. Philologie et Linguistique: Mélanges offerts à Louis Havet. Paris: Hachette, 131-150. Everaert, M. 1986. The Syntax of Reflexivization. Dordrecht: Foris. Eythórsson, T. 2000. “Dative vs. nominative: changes in quirky subjects in Icelandic”. In D. Nelson & P. Foulkes (eds.), Leeds Working Papers in Linguistics and Phonetics 8: 27-44. —. 2002. “Changes in subject case-marking in Icelandic”. In D. Lightfoot (ed.), Syntactic Effects of Morphological Change. Oxford: Oxford University Press, 196-212. —. 2003. “The change that never happened: the story of oblique subjects”. Journal of Linguistics 39(3): 439-472. Eythórsson, T. & Barðdal, J. 2005. “Oblique subjects: a common Germanic inheritance”. Language 81(4): 824-881. EĬǼī/HNC = Institute for Language and Speech Processing, Hellenic National Corpus. Faarlund, J. T. 1990. Syntactic Change. Towards a Theory of Historical Syntax. Berlin/ New York: Mouton de Gruyter. —. 2001. Grammatical Relations in Change. Amsterdam/ Philadelphia: J. Benjamins. —. 2004. The Syntax of Old Norse. Oxford: Oxford University Press. Feldman, H. 1978. “Passivizing on datives in Greek”. Linguistic Inquiry 9: 499-502. Fellbaum, C. & Zribi-Hertz, A. 1989. “La construction moyenne en français et en anglais: étude de syntaxe et de sémantique comparées”. Recherches Linguistiques de Vincennes 18: 19-57. Feltenius, L. 1977. Intransitivizations in Latin. Uppsala: Almqvist & Wiksell. Fillmore, C. J. 1968. “The case for case”. In E. Bach & R. T. Harms (eds.), Universals in Linguistic Theory. New York: Holt, Rinehart & Winston, 1-88. —. 1970. “The grammar of hitting and breaking”. In R. A. Jacobs & P. S. Rosenbaum (eds.), Readings in English Transformational Grammar. Waltham, MA: Ginn, 120-133. Finer, D. L. & Roeper, T. 1989. “From cognition to thematic roles: the projection principle as an acquisition mechanism”. In R. J. Matthews & W. Demopoulos (eds.), Learnability and Linguistic Theory. Dordrecht: Kluwer, 177-210. Fischer, O. & van der Leek, F. 1987. “A ‘case’ for the Old English impersonal”. In W. Koopman, F. van der Leek, O. Fischer & R. Eaton (eds.), Explanation and Linguistic Change. Amsterdam/ Philadelphia: J. Benjamins, 79-120. Flobert, P. 1975. Les verbes déponents Latins des Origines à Charlemagne. Paris: Les Belles Lettres. Foley, W. A. & Van Valin, R. D. 1984. Functional Syntax and Universal Grammar. Cambridge: Cambridge University Press. Folli, R. & Harley, H. 2005. “Consuming results in Italian and English: flavors of v”. In P. Kempchinsky & S. Slabakova (eds.), Aspectual Inquiries. Dordrecht: Springer, 95-120. Foucault, J.-A. de. 1972. Recherches sur la langue et le style de Polybe. Paris: Les Belles Lettres. Friedrich, P. 1975. Proto-Indo-European Syntax. The Order of Meaningful Elements. Journal of ǿndoeuropean Studies Monograph N.1.

232

Transitivity Alternations in Diachrony

Frisk, H. 1960-70. Griechisches etymologisches Wörterbuch. Heidelberg: C. Winter. Fu, J. Q. 1994. On Deriving Chinese Derived Nominals. Evidence for V-to-N Raising. PhD Dissertation, University of Massachusetts. Fukui, N. & Speas, M. 1986. “Specifiers and projection”. MIT Working Papers in Linguistics 8: 128172. Fykias, I. 1995. “ȅȚ ʌȡȠșİIJȚțȑȢ ijȡȐıİȚȢ ıIJȘ ȞȑĮ İȜȜȘȞȚțȒ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 15: 352-359. —. 1996. “ȅȚ ʌȡȠșİIJȚțȑȢ ijȡȐıİȚȢ țĮȚ IJȠ ıȪıIJȘȝĮ IJȦȞ ʌIJȫıİȦȞ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 16: 210-218. Gaedicke, C. 1880. Der Accusativ im Veda. Breslau: Koebner. Gavarró, A. 1992. “Empty objects in Catalan”. In A. Branchadell, N. Martí, B. Palmada, J. Quer, F. Roca, J. Solà & E. Vallduví (eds.), Catalan Working Papers in Linguistics 1992. Vol. 2. Universitat Autónoma de Barcelona, 145-161. Gehring, A. 1891. Index homericus. Leipzig: Teubner. Geldart, E. M. 1870. The Modern Greek Language in its Relation to Ancient Greek. Clarendon Press Series. Oxford: Clarendon Press. Gelzer, H. 1893. Leontios von Neapolis, Leben des heiligen Iohannes des Barmherzigen, Erzbischofs von Alexandrein. Freiburg: Akademische Verlagsbuchhandlung von J. C. B. Mohr. Geniušiene, E. 1987. The Typology of Reflexives. Berlin/ New York: Mouton de Gruyter. George, C. H. 2005. Expressions of Agency in Ancient Greek. Cambridge: Cambridge University Press. Geyskens, S. 1997. The Ergativization of Intransitives: a Lexico-Grammatical Description. M.A. Thesis, Dept. of Linguistics, K.U.Leuven. Giannakidou, ǹ. & Merchant, J. 1999. “Why Giannis can’t scrub his plate clean: on the absence of resultative secondary predication in Greek”. In A. Moser (ed.), Proceedings of the 3rd International Conference on Greek Linguistics. Athens: Ellinika Grammata, 93-103. Giannakis, G. K. 1992. “Reduplication as a morphological marker in the Indo-European languages”. WORD 43(2): 161-196. —. 1997a. Studies in the Syntax and Semantics of the Reduplicated Presents of Homeric Greek and Indo-European. Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft. —. 1997b. “Ǿ ʌȡȠȧıIJȠȡȓĮ IJȠȣ ȡȘȝĮIJȚțȠȪ ıȣıIJȒȝĮIJȠȢ IJȘȢ ĮȡȤĮȓĮȢ İȜȜȘȞȚțȒȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 17: 28-41. —. 1999. “ȆĮȡĮIJȘȡȒıİȚȢ ʌȐȞȦ ıIJȚȢ ȑȞȞȠȚİȢ țȩʌIJȦ țĮȚ įȚĮȞȑȝȦ, įȚĮȚȡȫ IJȘȢ ĮȡȤĮȓĮȢ İȜȜȘȞȚțȒȢ țĮȚ IJȘȢ ȚȞįȠİȣȡȦʌĮȧțȒȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 19: 81-94. Gibson, E. & Wexler, K. 1994. “Triggers”. Linguistic Inquiry 25(3): 407-454. Gignac, F. T. 1976/1981. A Grammar of the Greek Papyri of the Roman and Byzantine Periods. Vol. ǿ, Phonology, vol. II, Morphology. Milano: Cisalpino-Goliardica. Giorgi, A. & Longobardi, G. 1991. The Syntax of Noun Phrases. Configuration, Parameters and Empty Categories. Cambridge: Cambridge University Press. Givón, T. 1984/1990. Syntax: a Functional-Typological Introduction. Amsterdam/ Philadelphia: J. Benjamins. —. 1990a. “Reflexives, complementation and clause integration”. In C. Feagan & P. Muhlhausler (eds.), Papers in Honor of C. J. Bailey. Amsterdam/ Philadelphia: J. Benjamins. —. 1990b. “Tale of two passives: internal reconstruction”. In C. Duncan-Rose & T. Venneman (eds.), Papers Presented to Robert P. Stockwell on the Occasion of his 60th Birthday. London: CroomHelm. —. 2001. Syntax. An Introduction. Amsterdam/ Philadelphia: J. Benjamins. Glykatzi-Ahrweiler, Ǿ. 1999. “ǺȣȗȐȞIJȚȠ: İȚıĮȖȦȖȒ”. In ǿıIJȠȡȓĮ IJȘȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. Athens: Ǽ.L.I.A., 126-135. Godard, D. 1992. “Extraction out of NP in French”. Natural Language and Linguistic Theory 10: 233277. Göksel, A. 1993. Levels of Representation and Argument Structure in Turkish. PhD Dissertation, University of London. Gołąb, Z., Heinz, A. & PolaĔski, K. 1968. Słownik terminologii jĊzykoznawczej. Warszawa: PWN. Goldberg, A. 1995. Constructions: a Construction Grammar Approach to Argument Structure. Chicago: University of Chicago Press. Goldstein, D. M. 2008. “Review of ‘Dik H. 2007. Word Order in Greek Tragic Dialogue’. Oxford: Oxford University Press”. Bryn Mawr Classical Review 2008.06.08. Gonda, J. 1979. The medium in the Rgveda. Leiden: Brill. Goodall, G. 1997. “ș-alignment and the by-phrase”. Chicago Linguistic Society 33(1) [The Main Session]: 129-139.

Bibliography

233

Goodspeed, E. J. 1915. Die ältesten Apologeten. Göttingen: Vandenhoeck & Ruprecht. Goodwin, W. W. 1889/1965. Syntax of the Moods and Tenses of the Greek Verb. New York. —. 1897. Greek Grammar. 2nd edition. New York: Ginn & Company. Grimshaw, J. 1979. “Complement selection and the lexicon”. Linguistic Inquiry 10: 279-326. —. 1981. “Form, function, and the language acquisition device”. In C. L. Baker & J. McCarthy (eds.), The Logical Problem of Language Acquisition. Cambridge, MA: MIT Press, 165-182. —. 1991. Argument Structure. LI Monographs. Cambridge, MA: MIT Press. —. 1993. “Minimal projection, heads, and optimality”. Ms. Rutgers University. http://roa.rutgers.edu/ files/68-0000/roa-68-grimshaw-3.pdf. —. 1993/1997. “Projection, heads, and optimality”. Linguistic Inquiry 28: 373-422. Grimshaw, J. & Vikner, S. 1993. “Obligatory adjuncts and the structure of events”. In E. Reuland & W. Abraham (eds.), Knowledge and Language II: Lexical and Conceptual Structure. Dordrecht: Kluwer, 143-155. Gropen, J., Blaskovich, J. & DeDe, G. 1996. “Come it closer: causitive errors in child speech”. In A. Stringfellow, D. Cahana-Amitay, E. Hughes & A. Zukowski (eds.), BUCLD 20: Proceedings of the 20th Annual Boston University Conference on Language Development. Somerville, MA: Cascadilla Press, 272-283. Gruber, J. 1965. Studies in Lexical Relations. PhD Dissertation, MIT. —. 1976. Lexical Structures in Syntax and Semantics. Amsterdam: North Holland. Guerssel, M., Hale, K., Laughren, M., Levin, B. & White Eagle, J. 1985. “A cross-linguistic study of transitivity alternations”. Papers from the Parasession on Causatives and Agentivity. Chicago Linguistic Society: 48-63. Gvozdanoviü, J. 1996. “Tracing the origin of a change”. In J. Gvozdanoviü (ed.), Language Change and Functional Explanations. Berlin/ New York: Mouton de Gruyter, 71-87. Haacke, A. 1852. Beitrage zu einer Neugestaltung der griechischen Grammatik. Nordhausen: A. Büchting. Haeberli, E. 2002. “Inflectional morphology and the loss of verb-second in English”. In D. Lightfoot (ed.), Syntactic Effects of Morphological Change. Oxford: Oxford University Press, 88-106. Haegeman, L. 1991. Introduction to Government and Binding. Oxford: Blackwell. Haegeman, L. 2006. Thinking Syntactically: a Guide to Argumentation and Analysis. Oxford: Blackwell. Hagège, C. 1982. La structure des langues. Paris: PUF. Haiman, J. 1983. “Iconic and economic motivation”. Language 59: 781-819. Hale, K. 1996. “UG and the roots of linguistic diversity”. In J. D. Bobaljik, R. Pensalfini & L. Storto (eds.), Papers on Language Endangerment and the Maintenance of Linguistic Diversity. MIT Working Papers in Linguistics 28: 137-161. Hale, K. & Keyser, S. J. 1986. Some Transitivity Alternations in English. Lexicon Project Working Papers 7. Cambridge, MA: MIT Working Papers in Linguistics. —. 1992. “The syntactic character of thematic structure”. In I. M. Roca (ed.), Thematic Structure: its Role in Grammar. Berlin: Foris, 107-143. —. 1993. “On argument structure and the lexical expression of syntactic relations”. In K. Hale & S. J. Keyser (eds.), The View from Building 20. Essays in Linguistics in Honor of Sylvain Bromberger. Cambridge, MA: MIT Press, 53-109. —. 1995/1997. “The limits of argument structure”. Ms., MIT / Published in A. Mendikoetxea & M. Uribe Extebarria (eds.), Theoretical Issues at the Morphology-Syntax Interface. Bizkaia: UPV, 203230. —. 1996. “On the double-object construction”. Ms., MIT. —. 2002. Prolegomenon to a Theory of Argument Structure. Cambridge, MA: MIT Press. Hale, K. & Salamanca, D. 2000. “Theoretical and universal implications of certain verbal entries in dictionaries of the Misumalpan languages”. In W. Frawley et al. (eds.), Making Dictionaries. Berkeley, CA: University of California Press, 25-59. Halle, M. 1998. “The stress of English words: 1968-1998”. Linguistic Inquiry 29(4): 539-568. Halle, M. & Marantz, A. 1993. “Distributed Morphology and the pieces of inflection”. In K. Hale & J. Keyser (eds.), The View from Building 20. Essays in Linguistics in Honor of Sylvain Bromberger. Cambridge, Ȃǹ: MIT Press, 111-176. Halliday, M. A. K. 1964. “Syntax and the consumer”. Monograph Series on Languages and Linguistics 17: 11-24. Washington, D.C.: Georgetown University. —. 1980. “Functions of language”. Sophia Linguistica 6: 31-42. —. 1985a. “Systemic background”. In J. D. Benson & W. S. Greaves (eds.), Systemic Perspectives on Discourse. Norwood, N.J.: Ablex, 1-15.

234

Transitivity Alternations in Diachrony

—. 1985b. An Introduction to Functional Grammar. London: Edward Arnold. Harley, H. 1995. “Abstracting away from abstract case”. NELS 25: 207-221. —. 1998. “You’re having me on: aspects of have”. In J. Guéron & A. Zribi-Hertz (eds.), La grammaire de la possession. Paris: Université Paris X - Nanterre, 195-226. —. 1999. “Denominal verbs and aktionsart”. Papers from the UPenn/MIT Roundtable on the Lexicon. MIT Working Papers in Linguistics 35: 73-85. —. 2005. “How do verbs get their names? Denominal verbs, manner incorporation and the ontology of verb roots in English”. In N. Erteschik-Shir & T. Rapoport (eds.), The Syntax of Aspect. Oxford: Oxford University Press, 42-64. Harris, A. C. & Campbell, L. 1995. Historical Syntax in Cross-Linguistic Perspective. Cambridge Studies in Linguistics 74. Cambridge: Cambridge University Press. Harris, Z. 1956/1970. Introduction to Transformations. Transformations and Discourse Analysis Papers, No. 2. Philadelphia: University of Pennsylvania. [1970, in Papers in Structural and Transformational Linguistics. Dordrecht/ Holland: D. Reidel.] Haspelmath, M. 1987. “Transitivity alternations of the anticausative type”. Arbeitspapiere, N.F., Nr. 4. Cologne: Institut für Sprachwissenschaft der Universität zu Köln. —. 1990. “The grammaticization of passive morphology”. Studies in Language 14: 25-71. —. 1993. “More on the typology of inchoative/causative verb alternations”. In B. Comrie & M. Polinsky (eds.), Causatives and Transitivity. Amsterdam/ Philadelphia: J. Benjamins, 87-120. —. 1994. “Passive participles across languages”. In B. Fox & P. J. Hopper (eds.), Voice: Form and Function. Amsterdam/ Philadelphia: J. Benjamins, 151-177. —. 2003. “The geometry of grammatical meaning: semantic maps and cross-linguistic comparison.” In M. Tomasello (ed.), The New Psychology of Language. Vol. 2. Mahwah, NJ: Lawrence Erlbaum, 211-242. —. 2004. “On directionality in language change with particular reference to grammaticalization”. In O. Fischer, M. Norde & H. Perridon (eds.), Up and Down the Cline: the Nature of Grammaticalization. Amsterdam/ Philadelphia: J. Benjamins, 17-44. Hatcher, A. G. 1942. Reflexive Verbs: Latin, Old French, Modern French. Paris: Les Belles Lettres. Haugen, E. 1982. Scandinavian Language Structures: a Comparative Historical Survey. Tübingen: Niemeyer. Hawkins, J. A. 1983. Word Order Universals. New York: Academic Press. Hay, J., Kennedy, C. & Levin, B. 1999. “Scalar structure underlies telicity in 'Degree Achievements'”, Proceedings of SALT 9: 127-144. Hazout, I. 2000. “Adjectival genitive constructions in modern Hebrew: a case study in coanalysis”. The Linguistic Review 17: 29-52. Heidermanns, F. 1993. Etymologisches Wörterbuch der germanischen Primäradjektive. Berlin/ New York: Mouton de Gruyter. Heidinger, Ȃ. & Schäfer F. 2006. “On the French reflexive passive and anticausative: a diachronic view from the par-phrase”. Paper presented at the conference Diachro3, Paris, 20-22 September 2006. Heimerdinger, J. 1996. “Word order in Koine Greek. Using a text-critical approach to study word order patterns in the Greek text of Acts”. Filologia Neotestamentaria 18: 139-180. Heine, B. 1993. Auxiliaries: Cognitive Forces and Grammaticalization. Oxford: Oxford University Press. Henry, A. 1994. “Viewing change in progress: the loss of V2 in Hiberno-English imperatives”. In A. van Kemenade & N. Vincent (eds), Parameters of Morphosyntactic Change. Cambridge: Cambridge University Press, 273-296. Hermodsson, L. 1952. Reflexive und intransitive Verba im älteren Westgermanischen. Uppsala: Almquist & Wiksells. Hesse, R. 1980. Syntax of the Modern Greek Verbal System. Opuscula Byzantina et Neograeca 2. Copenhagen: Museum Tusculanum Press. Hewitt, B. G. 1981. “Caucasian languages”. In B. Comrie (ed.), The Languages of the Soviet Union. Cambridge: Cambridge University Press, 196-237. Higginbotham, J. 1986. “Linguistic theory and Davidson’s program in semantics”. In E. LePore (ed.), Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson. Oxford: Blackwell, 29-48. Hirt, H. 1937. Indogermanische Grammatik, VII: Syntax II. Heidelberg: C. Winter. Hochberg, J. 1986. “Children’s judgements of transitivity errors”. Journal of Child Language 13(2): 317-334. Hock, H. H. 1986. Principles of Historical Linguistics. Berlin/ New York: Mouton de Gruyter.

Bibliography

235

Hoekstra, T. 1987. Transitivity: Grammatical Relations in Government-Binding Theory. Dordrecht: Foris. —. 1988. “Small clause results”. Lingua 74: 101-139. —. 1992. “Aspect and theta theory”. In I. M. Roca (ed.), Thematic Structure: its Role in Grammar. Berlin: Foris, 145-174. Hofmann, J. B. & Szantyr, A. 1965. Lateinische Syntax und Stilistik. München: C. H. Beck. Holm, G. 1952. Om s-passivum i svenskan. Lund: Gleerup. Holton, D. 1974. ǻȚȒȖȘıȚȢ IJȠІ ̝ȜİȟȐȞįȡȠȣ. The Tale of Alexander. The Rhymed Version. Thessaloniki. Holton, D., Mackridge, P. & Philippaki-Warburton, I. 1997. Greek: a Comprehensive Grammar of the Modern Language. London: Routledge. Hopper, P. & Thompson, S. 1980. “Transitivity in grammar and discourse”. Language 56(2): 251-299. —. (eds.) 1982. Syntax and Semantics 15. Studies in Transitivity. New York/ London: Academic Press. Hopper, P. J. & Traugott, E. C. 1993. Grammaticalization. Cambridge: Cambridge University Press. Horrocks, G. 1991. “ȅȚ țȜȚIJȚțȑȢ ĮȞIJȦȞȣȝȓİȢ ıIJȘȞ ȚıIJȠȡȓĮ IJȘȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 11: 39-51. —. 1997. Greek: a History of the Language and its Speakers. London: Longman. —. 2007. “Syntax: from Classical Greek to Koine”. In A.-Ph. Christidis (ed.), A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press, 618-631. Horrocks, G. & Stavrou, M. 2003. “Actions and their results in Greek and English: the complementarity of morphologically encoded aspect and syntactic resultative predication”. Journal of Semantics 20: 297-327. Horvath, J. & Siloni, T. 2005. “Active lexicon: adjectival passives”. Ms., Tel Aviv University. Presentation at the Semitic Syntax Workshop, GLOW 28, University of Geneva, March 29-30. Householder, F., Kazazis, K. & Koutsoudas, A. 1964. Reference Grammar of Literary Dhimotiki. Bloomington: Indiana University / The Hague: Mouton. Hult, K. 1990. Syntactic Variation in Greek of the 5th Century AD. Studia Graeca et Latina Gothoburgensia LII. Gothenburg: Acta Universitatis Gothoburgensis. Humbert, J. 1930. La disparition du datif en grec du Ier au Xe siecle. Paris. —. 1945. Syntaxe Grecque. Collection de Philologie Classique II. Paris: Klincksieck. Humbert, J. & Kourmoulis, G. 1957/2002. ȈȣȞIJĮțIJȚțȩȞ IJȘȢ ǹȡȤĮȓĮȢ ǼȜȜȘȞȚțȒȢ īȜȫııȘȢ. Athens. Hunger, H. 1981. Anonyme Metaphrase zu Anna Komnena, Alexias XI- XIII. Ein Beitrag zur Erschließung der byzantinischen Umgangssprache. Wien: Verlag der österreichischen Akademie der Wissenschaften. Hunger, H. & K. Vogel 1963. Ein byzantinisches Rechenbuch des 15. Jahrhunderts. Wien: Österreichische Akademie der Wissenschaften. Hunt, A. S. & Edgar, C. C. (eds.) 1932. Select Papyri I: Private Affairs. Cambridge, Mǹ: Loeb Classical Texts. Iand Æ Kapsomenakis, S. G. Ingham, R. 1995. “Acquiring a semantic constraint on argument structure: the role of linguistic input”. Reading Working Papers in Linguistics 2: 23-43. Jackendoff, R. 1972. Semantic Interpretation in Generative Grammar. Cambridge, MA: MIT Press. —. 1977. X' Syntax: a Study of Phrase Structure. Linguistic Inquiry Monograph 2. Cambridge, MA: MIT Press. —. 1983. Semantics and Cognition. Cambridge, MA: MIT Press. —. 1987. “The status of thematic relation in linguistic theory”. Linguistic Inquiry 18(3): 369-411. —. 1990. Semantic Structure. Cambridge, MA: MIT Press. Jacobsen, W. 1992. The Transitive Structure of Events in Japanese. Tokyo: Kurioso. Jaeggli, O. 1986a. “Passive”. Linguistic Inquiry 17: 587-622. —. 1986b. “Three issues in the theory of clitics: case, doubled NPs, and extraction”. Syntax and Semantics 19: 15-42. James, M. R. 1982. The ȉestament of Abraham. Cambridge: Cambridge University Press. Jamison, S. 1983. Function and form in the -áya-formations of the Rig Veda and Atharva Veda. Göttingen: Vandenhoeck & Ruprecht. Janda, R. & Joseph, B. 2003. “On language, change, and language change – or, Of history, linguistics, and historical linguistics’. In B. Joseph & R. Janda (eds.), The Handbook of Historical Linguistics. Oxford: Blackwell, 3-180. Jankuhn, H. 1969. Die passive Bedeutung medialer Formen untersucht an der Sprache Homers. Göttingen: Vandenhoeck & Ruprecht. Jannaris, A. 1897/1968. An Historical Greek Grammar, chiefly of the Attic dialect, as written and spoken from Clasical Antiquity down to the Present Time, founded upon the Ancient Texts,

236

Transitivity Alternations in Diachrony

Inscriptions, Papyri, and Present Popular Greek. London: MacMillan. (Reprint 1987, Hildesheim: Georg Olms). Janse, M. 1993. “La position des pronoms personnels enclitiques en grec néo-testamentaire à la lumière des dialectes néo-helléniques”. In C. Brixhe (ed.), La koiné grecque antique: I. Une langue introuvable?. Nancy: Presses Universitaires de Nancy, 83-121. —. 1998. “Cappadocian clitics and the syntax-morphology interface”. In B. D. Joseph, G. C. Horrocks & I. Philippaki-Warburton (eds.), Themes in Greek Linguistics II. Amsterdam/ Philadelphia: J. Benjamins, 257-281. Jasanoff, J. H. 1978. Stative and Middle in Indo-European. Innsbrucker Beiträge zur Sprachwissenschaft 23. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Jeffreys, E. 1998. Digenis Akritis. The Grottaferatta and Escorial Versions. Cambridge Medieval Classics 7. Cambridge: Cambridge University Press. Jeffreys, E., Croke, B. & Scott, R. 1990. Studies in John Malalas. Sydney: Australian Association for Byzantine Studies. Jelinek, E. 1998. “Voice and transitivity as functional projections in Yaqui”. In M. Butt & W. Geuder (eds.), The Projection of Arguments: Lexical and Compositional Factors. Stanford, CA: CSLI Publications, 195-225. Jespersen, O. 1927/1954. A Modern English Grammar on Historical Principles, Volume III (Syntax). London: Allen & Unwin. Johnson, K. 1991. “Object positions”. Natural Language and Linguistic Theory 9: 577-636. —. 1994. “Bridging the gap”. Ms., University of Massachusetts, Amherst. Joseph, B. 1978/1990. Morphology and Universals in Syntactic Change: Evidence from Medieval and Modern Greek. PhD Dissertation, Harvard University / New York: Garland Publishers. —. 1992. “Diachronic explanation: putting speakers back into the picture”. In G. D. Davis & G. K. Iverson (eds.), Explanation in Historical Linguistics. Current Issues in Linguistic Theory 84. Amsterdam/ Philadelphia: J. Benjamins, 123-144. —. 2001a. “Ancient Greek”. In J. Garry, C. Rubino, A. Faber & R. French (eds.), Facts about the World’s Major Languages. An Encyclopedia of the World’s Major Languages, Past and Present. New York: The H. W. Wilson Company, 256-262. —. 2001b. “Is there such a thing as ‘Grammaticalization’?” Language Sciences 23 [special issue: Grammaticalization: a Critical Assessment, edited by L. Campbell]: 163-186. —. 2002. “Morphologization from syntax”. In B. Joseph & R. Janda (eds.), Handbook of Historical Linguistics. Oxford: Blackwell, 472-492. Joseph, B. & Philippaki-Warburton, I. 1987. Modern Greek. Croom Helm Descriptive Grammar Series. London: Croom Helm. Juarros-Daussà, E. 2000. The Syntactic Operator se in Spanish. UMass Amherst. —. 2003. Structure and the Lexicon-Syntax Interface. PhD Dissertation, University of Massachusetts, Amherst. Kahane, A. & Mueller, M.: The Chicago Homer. Northwestern University. 1 December 2009.

Kaklamanis, S. & Lambakis, S. 2003. ȂĮȞȠȣȒȜ īȡȘȖȠȡȩʌȠȣȜȠȢ, ȞȠIJȐȡȚȠȢ ȋȐȞįĮțĮ (1506-1532). ǻȚĮșȒțİȢ, ǹʌȠȖȡĮijȑȢ – ǼțIJȚȝȒıİȚȢ. Iraklion: Vikelea Dimotiki Vivliothiki. Kakouriotis, T. 1993. Argument Realization. Thessaloniki. —. 1994. “Theta roles and middle constructions”. Proceedings of the 8th International Symposium of English and Greek, 279-304. Thessaloniki. Kakouriotis, A. & Kitis, E. 1999. “The case of 'fovame'[fear] and other psychological verbs”. In A. Moser (ed.), Proceedings of the 3rd International Conference on Greek Linguistics. Athens: Ellinika Grammata, 131-140. Kakridi-Ferrari, Ȃ. & Chila-Markopoulou, D. 1996. “Ǿ ȖȜȦııȚțȒ ʌȠȚțȚȜȓĮ țĮȚ Ș įȚįĮıțĮȜȓĮ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ ȦȢ ȟȑȞȘȢ ȖȜȫııĮȢ”. In Ǿ ȃȑĮ ǼȜȜȘȞȚțȒ ȦȢ ȟȑȞȘ ȖȜȫııĮ – ȆȡȠȕȜȒȝĮIJĮ įȚįĮıțĮȜȓĮȢ. Athens: Goulandri-Chorn Foundation, 17-51. Kalitsounakis, I. E. 1928. “ǹȞȑțįȠIJĮ ȀȡȘIJȚțȐ ıȣȝȕȩȜĮȚĮ İț IJȘȢ ǼȞİIJȠțȡĮIJȓĮȢ”. Proceedings of the Academy of Athens 3: 483-519. Kallulli, D. 2006. “Unaccusatives with dative Causers and Experiencers: a unified account”. In D. Hole, A. Meinunger & W. Abraham (eds.), Datives and Similar Cases. Amsterdam/ Philadelphia: J. Benjamins, 271-300. Kalonaros, P. P. 1940. ȉȠ ȋȡȠȞȚțȩȞ IJȠȣ ȂȠȡȑȦȢ. Athens. Kapsomenakis, S. G. 1938. Voruntersuchungen zu einer Grammatik der Papyri der nachchristlichen Zeit. Münchener Beiträge zur Papyrusforschung und antiken Rechtsgeschichte. Munich: Beck. —. 1953. “Beiträge zur Historischen Grammatik der Griechischen Dialekte Unteritaliens”.

Bibliography

237

Byzantinische Zeitschrift 46: 320-348. Karali, M. 1990. “ȈİȚȡȐ IJȦȞ ȩȡȦȞ, İʌȚIJĮııȩȝİȞĮ ıIJȠȚȤİȓĮ țĮȚ İʌĮijȒ įȚĮȜȑțIJȦȞ: İʌȓįȡĮıȘ IJȘȢ İȜȜȘȞȚıIJȚțȒȢ țȠȚȞȒȢ ıIJȘ įİȜijȚțȒ įȚȐȜİțIJȠ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 10: 25-45. —. 1992. “ĭȩȡȝȠȣȜİȢ țĮȚ įȚȐIJĮȟȘ IJȦȞ ȩȡȦȞ IJȘȢ ʌȡȩIJĮıȘȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 13: 69-90. —. 1993. “Infinitival constructions and case assignment: a case of a consecutive formula in the delphic corpus”. In I. Philippaki-Warburton, K. Nicolaidis & M. Sifianou (eds.), Themes in Greek Linguistics. Papers from the First International Conference on Greek Linguistics, Reading, September 1993. Amsterdam/ Philadelphia: J. Benjamins, 405-412. —. 2007. “The use of dialects in literature”. In A.-Ph. Christidis (ed.), A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press, 974-985. Karantzola, E. 2007. “From humanism to Enlightenment: the teaching of Ancient Greek and its grammar”. In A.-Ph. Christidis (ed.), A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press, 1241-1249. Karantzola, E. & Lavidas, N. 2009-to appear. “Lability, case, and split ergativity in diachrony: evidence from Early Modern Greek”. In L. Kulikov & N. Lavidas (eds), Special issue on lability. Karantzola, E., Tiktopoulou K. & Frantzi, K. T. 2006. “ȉĮ ȑȖȖȡĮijĮ ȦȢ ʌȘȖȑȢ ȖȚĮ IJȘȞ ʌȡȫȚȝȘ ȞİȠİȜȜȘȞȚțȒ”. International Congress in Neograeca Medii Aevi VI, September 29 - October 2, Ioannina, Greece. Karla, G. A. 2001. Vita Aesopi: Überlieferung, Sprach und Edition einer frübyzantinischen Fassung des Äsopsromans. Wiesbaden: Reichert. Karla, G. A. & Lavidas, N. 2003. “ȉȠ ĮʌĮȡȑȝijĮIJȠ ıIJȠȞ ȐȟȠȞĮ IJȘȢ įȚĮȤȡȠȞȓĮȢ IJȘȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ”. Proceedings of the 6th International Conference on Greek Linguistics. http://www.philology.uoc.gr/ conferences/6thICGL/ Kastovsky, D. 1992. “Word-Formation”. In R. M. Hogg (ed.), The Cambridge History of the English Language. Volume 1: The Beginnings to 1066. Cambridge: Cambridge University Press, 355-400. Katsogiannou, M. 1995. “ȉȠ ȡȒȝĮ ıIJȠ İȜȜȘȞȚțȩ ȚįȓȦȝĮ IJȘȢ ȀȐIJȦ ǿIJĮȜȓĮȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 15: 542-553. —. 1997. “ȉĮ İȜȜȘȞȚțȐ IJȘȢ ȀȐIJȦ ǿIJĮȜȓĮȢ: ȝȠȡijȠȜȠȖȓĮ IJȠȣ ȠȞȩȝĮIJȠȢ țĮȚ İȟȑȜȚȟȘ IJȠȣ țȜȚIJȚțȠȪ ıȣıIJȒȝĮIJȠȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 17: 328-341. Katsouros, A. 1955. “ȃĮȟȚĮțȐ įȚțĮȚȠʌȡĮțIJȚțȐ ȑȖȖȡĮijĮ IJȠȣ 16Ƞȣ ĮȚȫȞȠȢ”. ǼʌİIJȘȡȓȢ ȂİıĮȚȦȞȚțȠȪ ǹȡȤİȓȠȣ 5: 47-91. Kavoukopoulos, F. 1988. Les expansions casuelles et prépositionnelles du prédicat. Essai de syntaxe homerique. PhD Dissertation, Université Paris V. Kayne, R. 1984. Connectedness and Binary Branching. Dordrecht: Foris. —. 1991. “Romance clitics, verb movement and PRO”. Linguistic Inquiry 22(4): 647-686. —. 1994. The Antisymmetry of Syntax. Cambridge, MA: MIT Press. Kechagioglou, G. 2004. ǹʌȠȜȜȫȞȚȠȢ IJȘȢ ȉȪȡȠȣ. ȊıIJİȡȠȝİıĮȚȦȞȚțȑȢ țĮȚ ȞİȩIJİȡİȢ İȜȜȘȞȚțȑȢ ȝȠȡijȑȢ. Thessaloniki: Institute for Modern Greek Studies (Manolis Triantafyllidis Foundation). Kechagioglou, G. (ed.) 1997. Ǿ ʌĮȜĮȚȩIJİȡȘ ʌİȗȠȖȡĮijȓĮ ȝĮȢ. [vol. B1, 15th cent.-1830]. Athens: Sokolis. Keenan, E. L. 1996. “The semantics of determiners”. In S. Lappin (ed.), The Handbook of Contemporary Semantic Theory. Oxford: Blackwell, 41-63. Kemenade, ǹ. van. 1987. Syntactic Case and Morphological Case in the History of English. Dordrecht: Foris. Kemmer, S. 1993. The Middle Voice. Amsterdam/ Philadelphia: J. Benjamins. Kempinsky, P. 1991. “On a characterization of a class of ditransitive verbs in Spanish”. In H. Campos & F. Martínez-Gil (eds.), Current Studies in Spanish Linguistics. Washington: Georgetown University Press, 201-240. Kerstens, J., Ruys, E. & Zwarts, J. 1996-2001. Lexicon of Linguistics. Utrecht Institute of Linguistics OTS, Utrecht University. Keyser, S. J. & Roeper, T. 1984. “On the middle and ergative constructions in English”. Linguistic Inquiry 15: 381-416. —. 1992. “Re: the abstract clitic hypothesis”. Linguistic Inquiry 23(1): 89-125. Kibrik, A. A. 1993. “Transitivity increase in Athabaskan languages”. In B. Comrie & M. Polinsky (eds.), Causatives and Transitivity. Amsterdam/ Philadelphia: J. Benjamins, 47-67. Kiparsky, P. 1979. “Analogical change as a problem for linguistic theory”. In B. Kachru (ed.), Linguistics in the Seventies: Directions and Progress. Studies in the Linguistic Sciences 8(2), 72-96.

238

Transitivity Alternations in Diachrony

—. 1982a. “Word-formation and the lexicon”. In F. Ingemann (ed.), Proceedings of the 1982 MidAmerica Lingistics Conference. Kansas: Lawrence, 3-29. —. 1982b. Explanation in Phonology. Dordrecht: Foris. —. 2003. “Accent, syllable structure, and morphology in Ancient Greek”. In E. Mela-Athanasopoulou (ed.), Selected Papers from the 15th International Symposium on Theoretical and Applied Linguistics. Thessaloniki, 81-106. Kitazume, S. 1996. “Middles in English”. WORD 47: 161-183. Klaiman, M. H. 1991. Grammatical Voice. Cambridge: Cambridge University Press. Kontosopoulos, N. 1994. ǻȚȐȜİțIJȠȚ țĮȚ ȚįȚȫȝĮIJĮ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ. Athens. Kontzi, R. 1958. Der Ausdruck der Passividee im älteren Italienischen. Tübingen: Niemeyer. Koontz-Garboden, A. & Levin, B. 2005. “The morphological typology of change-of-state event encoding”. In G. Booij, E. Guevara, A. Ralli, S. Sgroi & S. Scalise (eds.), On-line Proceedings of the Fourth Mediterranean Morphology Meeting (MMM4). Universita degli Studi di Bologna, Bologna, Catania, 21-23 September 2003: 185-194.http://mmm.lingue.unibo.it/mmm-proc/MMM4/ 185-194-KoontzGarboden-Levin-MMM4.pdf Koontz-Garboden, A. 2008. “Monotonicity at the lexical semantics-morphosyntax interface”. In E. Elfner & M. Walkow (eds.), Proceedings of the 37th Annual Meeting of the Northeast Linguistic Society. Amherst, MA: UMass, GLSA. Kopidakis, M. Z. 1999. ǿıIJȠȡȓĮ IJȘȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. Athens: E.L.I.A. Kordoni, V. 2001. Psych Verb Constructions in Modern Greek: a Semantic Analysis in the Hierarchical Lexicon. PhD Dissertation, University of Essex. KostȠula, D. 1991. ǹȖĮʌȓȠȣ ȁȐȞįȠȣ īİȦʌȠȞȚțȩȞ. Volos: Tinos Publications. Kotzia, E. & Efthymiou, E. (eds.) 2001. īİȞİIJȚțȒ ȖȡĮȝȝĮIJȚțȒ țĮȚ ıȣȖțȡȚIJȚțȒ ĮȞȐȜȣıȘ. Athens: Kastaniotis. Kowaleck, H. 1887. Über Passiv und Medium vornehmich im Sprachgebrauche des Homer. Wedel: Danzig. Krahe, H. & Meid, W. 1969. Germanische Sprachwissenschaft. Bd. 1: Einleitung und Lautlehre. Bd. 2: Formenlehre. Bd. 3: Wortbildungslehre. Berlin/ New York: Mouton de Gruyter. Kratzer, A. 1988. “Stage-level and individual-level predicates”. In M. Krifka (ed.), Genericity in Natural Language. Proceedings of the 1988 Tübingen Conference, SNS-Bericht: 88-42, 247-284. —. 1993. “The event argument and the semantics of voice”. Ms., University of Massachusetts, Amherst. —. 1995. “Stage-level and individual-level predicates”. In G. Carslon & F. J. Pelletier (eds.), The Generic Book. Chicago: Chicago University Press, 125-175. —. 1996. “Severing the external argument from its verb”. In J. Rooryck & L. Zaring (eds.), Phrase Structure and the Lexicon. Dordrecht: Kluwer, 109-139. —. 1999. The Event Argument and the Semantics of Verbs. University of Massachusetts at Amherst. —. 2000. “Building statives”. In L. J. Conathan et al. (eds.), Proceedings of the 26th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 385-399. —. 2003. “Building resultatives”. In C. Maienborn & A. Wöllstein- Leisten (eds.), Event Arguments in Syntax, Semantics, and Discourse. Tübingen: Niemeyer, 177-212. Krebs, F. 1882. “Die Präpositionen bei Polybius”. In M. Schanz (ed.), Beitrage zur historischen Syntax der griechischen Sprache. Heft I. Würzburg: A. Stuber, 3-4. Kriaras, E. 1955. ǺȣȗĮȞIJȚȞȐ ȚʌʌȠIJȚțȐ ȝȣșȚıIJȠȡȒȝĮIJĮ. Athens: Aetos. —. 1957. “īȜȦııȚțȐ ijĮȚȞȩȝİȞĮ ıİ įȘȝȫįȘ ȝİıĮȚȦȞȚțȐ țİȓȝİȞĮ”. ǼȜȜȘȞȚțȐ 15: 191-202. —. 1969-2006. ȁİȟȚțȩ IJȘȢ ȝİıĮȚȦȞȚțȒȢ İȜȜȘȞȚțȒȢ įȘȝȫįȠȣȢ ȖȡĮȝȝĮIJİȓĮȢ 1100-1669. Vol. 1-16. Thessaloniki: Centre for the Greek Language. Krifka, M. 1989. “Nominal reference, temporal constitution and quantification in event semantics”. In R. Bartsch, J. van Benthem & P. van Emde Boas (eds.), Semantics and Contextual Expression. Dordrecht: Foris, 75-115. —. 1992. “Thematic relations as links between nominal reference and temporal constitution”. In I. A. Sag & A. Szabolcsi (eds.), Lexical Matters. Stanford, CA: CSLI Publications, 29-54. —. 1998. “The origins of telicity”. In S. Rothstein (ed.), Events and Grammar. Dordrecht: Kluwer, 197-235. Kroch, A. 1989. “Reflexes of grammar in patterns of language change”. Language Variation and Change 1: 199-244. —. 2000. “Syntactic change”. In M. Baltin & C. Collins (eds.), The Handbook of Contemporary Syntactic Theory. Oxford: Blackwell, 629-739. Kroch, A. & Taylor, ǹ. 1997. “Verb movement in Old and Middle English: dialect variation and language contact”. In A. van Kemenade & N. Vincent (eds.), Parameters of Morphosyntactic

Bibliography

239

Change. Cambridge: Cambridge University Press, 297-325. —. 2001. “Verb-object in early Middle English”. In S. Pintzuk, G. Tsoulas & A. Warner (eds.), Diachronic Syntax: Models and Mechanisms. Oxford: Oxford University Press, 132-163. Krüger, K. W. 1998. Attic Greek Prose Syntax (Edited by: G. L. Cooper). Ann Arbor: University of Michigan Press. Krumbacher, K. 1897. Die Geschichte der Byzantinischen Literatur. 2nd edition. Munich: Beck. Kühner, R. & Gerth, B. [1898-1904]/1963. Ausführliche Grammatik der Griechischen Sprache. Zweiter Teil: Satzlehre. Munich: Max Hueber. Kulikov, L. I. 1998. “Causative constructions in Tuvinian: towards a typology of transitivity”. In L. Johanson (ed.), The Mainz Meeting. Proceedings of the Seventh International Conference on Turkish Linguistics, August 1994. Wiesbaden: Harrassowitz, 258-264. —. 1999a. “May he prosper in offspring and wealth: a few jubilee remarks on the typology of labile verbs and Sanskrit púsyati 'prospers; makes prosper'”. In E. V. Rakhilina & Y. G. Testelets (eds.), Tipologija i teorija jazyka: Ot opisanija k ob"jasneniju. K 60-letiju A.E. Kibrika [Typology and Linguistic Theory: from Description to Explanation. For the 60th birthday of Aleksandr E. Kibrik]. Moscow: Jazyki russkoj kul'tury, 224-244. —. 1999b. “Split causativity: remarks on correlations between transitivity, aspect, and tense”. In W. Abraham & L. I. Kulikov (eds.), Tense-Aspect, Transitivity and Causativity. Essays in Honour of Vladimir Nedjalkov. Amsterdam/ Philadelphia: J. Benjamins, 21-42. —. 2001a. “Between passive and reflexive: the Vedic presents with the suffix -ya-”. In G. van der Meer & A. G. B. ter Meulen (eds.), Making Sense: from Lexeme to Discourse. In honor of Werner Abraham at the Occasion of his Retirement. Groningen: Center for Language and Cognition, 13-20. —. 2001b. “Causatives”. In M. Haspelmath, E. König, W. Oesterreicher & W. Raible (eds.), Language Typology and Language Universals. An International Handbook. Vol 2. Berlin/ New York: Mouton de Gruyter, 886-898. —. 2003. “The labile syntactic type in a diachronic perspective: the case of Vedic”. SKY Journal of Linguistics (Journal of the Linguistic Association of Finland) 16: 93-112. —. 2006. “Passive and middle in Indo-European: reconstructing the early Vedic passive paradigm”. In W. Abraham & L. Leisiö (eds.), Passivization and Typology: Form and Function. Proceedings of the Workshop on Passive at the 20th Scandinavian Conference of Linguistics. Amsterdam/ Philadelphia: J. Benjamins, 62-81. Kuryłowicz, J. 1935. Etudes indoeuropéennes I. Krakow. —. 1956. L’apophonie en indo-européen. Krakow. —. 1964. The Inflectional Categories of Indo-European. Heidelberg: C. Winter. —. 1965. “The evolution of grammatical categories”. Diogenes 51: 55-71. Kyriazidis, N. I. (ed.) 1992. ȉĮ İȜȜȘȞȚțȐ IJȠȣ ȂĮțȡȣȖȚȐȞȞȘ ȝİ IJȠȞ ȣʌȠȜȠȖȚıIJȒ. Athens: Papazisis. La Fauci, N. 1988. Objects and Subjects in the Formation of Romance Morphosyntax. Indiana University Linguistics Club Publications. Lagae, V. 2002. “Le passif pronominal: une forme complémentaire du passif périphrastique?”. In V. Lagae, A. Carlier & C. Benninger (eds.), Cahiers Chronos 10, Temps et aspect: de la grammaire au lexique. Amsterdam: Rodopi, 133-149. Lagane, R. 1967. “Les verbes symétriques: économie morphosyntaxique et différenciation sémantique”. Cahiers de Lexicologie 10: 21-30. Lakoff, G. 1968. Ordering of Transformational Rules. Bloomington: Indiana University Linguistics Club. —. 1970. Irregularity in Syntax. New York, NY: Holt, Rinehart & Winston. —. 1972. “The global nature of the nuclear stress rule”. Language 48: 285-303. —. 1977. “Linguistic gestalts”. Chicago Linguistic Society 13: 236-287. Lallot, J. (ed.) 1997. Apollonius Dyskole: De la constuction (syntaxe). Paris: Librairie Philosophique J. Vrin. Lambert, J. A. 1935. Le roman de Libistros et Rhodamné. Amsterdam: Van de Noord. Lambros, S. 1908. “ΕΉϧΖȱ πΔ΍ΗΘΓΏ΅Ϡȱ ΘΓІȱ Ύ΅ΕΈ΍Α΅ΏϟΓΙȱ ̅΋ΗΗ΅ΕϟΝΑΓΖȱ πΑȱ ΘϜȱ Έ΋ΐЏΈϙȱ ·ΏЏΗΗϙ”. ̐νΓΖȱ̴ΏΏ΋ΑΓΐΑφΐΝΑȱ5: 20-39. Lamiroy, B. 1993. “Pourquoi il y a deux passifs”. Langages 109: 53-72. Lampe, G. W. H. 1961. A Patristic Greek Lexicon. Oxford: Oxford University Press. Langacker, R. 1977. “Syntactic reanalysis”. In C. N. Li (ed.), Mechanisms of Syntactic Change. Austin: University of Texas Press, 57-139. —. 1991. Concept, Image, and Symbol: the Cognitive Basis of Grammar. 1: Cognitive Linguistics Research. Berlin/ New York: Mouton de Gruyter. Larjavaara, M. 2000. Présence ou absence de l’objet. Limites du possible en français contemporain.

240

Transitivity Alternations in Diachrony

PhD Dissertation, Helsinki: Academia Scientiarum Fennica. Larson, R. K. 1988a. “Implicit arguments in Situation Semantics”. Linguistics and Philosophy 11: 169201. —. 1988b. “On the double object construction”. Linguistic Inquiry 19: 335-391. —. 1988c. “Scope and comparatives”. Linguistics and Philosophy 11: 1-26. —. 2003. “Time and event measure”. In J. Hawthorne & D. Zimmerman (eds.), Philosophical Perspectives 17, Language and Philosophical Linguistics. Oxford: Oxford University Press, 247258. Lascaratou, C. 1984. The Passive Voice in Modern Greek. PhD Dissertation, University of Reading. Lascaratou, C. & Philippaki-Warburton, I. 1984. “Lexical versus transformational passives in Modern Greek”. Glossologia 2-3: 99-109. Lavidas, N. 2002. ǻȠȝȑȢ ȝİ ʌȡȠȠʌIJȚțȒ IJȠȣ ǻȑțIJȘ țĮȚ Ș țĮIJĮȞȠȝȒ -Ȧ/ -ȝĮȚ: įȚĮȤȡȠȞȚțȒ ʌȡȠıȑȖȖȚıȘ. MA Thesis. Department of Linguistics. National and Kapodistrian University of Athens. —. 2003a. “ǻȠȝȑȢ ‘İȓȝĮȚ + ȝİIJȠȤȒ’ / ‘être + ȝİIJȠȤȒ’ țĮȚ ȖȡĮȝȝĮIJȚțȠʌȠȓȘıȘ”. In ȃ. ȁĮȕȓįĮȢ et al. (eds.), APC-1. Athens: National and Kapodistrian University of Athens, 85-95. —. 2003b. “MİIJĮȕȚȕĮıIJȚțȑȢ İȞĮȜȜĮȖȑȢ IJȘȢ ǼȜȜȘȞȚțȒȢ țĮȚ ȝİIJĮȕȠȜȑȢ ıIJȘ ȝİIJĮȕĮIJȚțȩIJȘIJĮ”. Proceedings of the 6th International Conference on Greek Linguistics. http://www.philology.uoc.gr/ conferences/6thICGL/ —. 2004. “ȂİIJĮȕȚȕĮıIJȚțȑȢ İȞĮȜȜĮȖȑȢ: įȚĮȤȡȠȞȚțȑȢ țĮȚ ıȣȖȤȡȠȞȚțȑȢ IJȐıİȚȢ ıIJȘȞ ǼȜȜȘȞȚțȒ țĮȚ IJȘ īĮȜȜȚțȒ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ EȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 24: 369-381. —. 2005. “The diachrony of the Greek anticausative morpholȠgy”. In A. Alexiadou (ed.), Studies in the Morpho-syntax of Greek. Newcastle: Cambridge Scholars Publishing, 106-136. —. 2006a. Changes in Verb Transitivity. PhD Dissertation. Department of Linguistics. National and Kapodistrian University of Athens. —. 2006b. “The development of the Greek labile verbs”. In S. Kittilä & L. Kulikov (eds.), Diachronic Typology of Voice and Valency-Changing Categories. Amsterdam/ Philadelphia: J. Benjamins. Lavidas, N. & Papangeli, D. 2005. “Impersonals in the diachrony of Greek”. Workshop on Voice Morphology. Proceedings of the 7th International Conference on Greek Linguistics. University of York. —. 2006. “Deponency in the diachrony of Greek”. In M. Baerman, G. G. Corbett, D. Brown & A. Hippisley (eds.), Deponency and Morphological Mismatches. Oxford: Oxford University Press, 97126. Lavidas, N. & Therapontos, C. 2006. “ȂİIJĮȕĮIJȚțȑȢ İȞĮȜȜĮȖȑȢ ıIJȘȞ ȀȣʌȡȚĮțȒ”. 5th International Conference on Greek dialects. Athens, May 2006. Lazard, G. 1984. “Actance variations and categories of the object”. In F. Plank (ed.), Objects. Towards a theory of Grammatical Relations. London: Academic Press, 269-292. —. 2002. “Transitivity revisited as an example of a more strict approach in typological research”. Folia Linguistica 36(3-4): 141-190. Lebaux, D. 2000. Language Acquisition and the Form of the Grammar. Amsterdam/ Philadelphia: J. Benjamins. Legrand, É. 1880-1913. Bibliothèque Grecque Vulgaire. Paris: Maisonneuve. —. 1885. Bibliographie Hellénique (XVe-XVIe siècles). Vol. 2. Paris. Lehmann, C. 1982/1995. Thoughts on Grammaticalization. LINCOM Studies in Theoretical Linguistics 01. Munich: LINCOM Europa. —. 1985. “Grammaticalization: synchronic variation and diachronic change”. Lingua E Stile 20: 303318. —. 2005. “Latin causativization in typological perspective”. 13ème Colloque International de Linguistique Latine, Bruxelles, 4-8 Avril 2005, http://www.christianlehmann.eu/publ/ latin_causativization.pdf Lehmann, W. 1972a. “Contemporary linguistics and Indo-European studies”. Publications of the Modern Language Association of America 87: 976-993. —. 1972b. “Germanic”. In W. K. Wimsatt (ed.), Versification: Major Language Types. New York: New York University Press, 122-135. —. 1992a. “Comparative linguistics”. In E. C. Polomé & W. Winter (eds.), Reconstructing Languages and Cultures. Berlin/ New York: Mouton de Gruyter, 5-19. —. 1992b. “SWÆS. From the middle to pronominal reflexive markers”. In C. Blank et al. (eds.), Language and Civilization. A Concerted Profusion of Essays and Studies in Honour of Otto Hietsch. Frankfurt: Peter Lang, 139-146. —. 1992c. Historical Linguistics. 3rd edition. London: Routledge. Lekakou, M. 2005. In the Middle, Somewhat Elevated. The Semantics of Middles and its

Bibliography

241

Crosslinguistic Realization. PhD Dissertation, University of London. Lemmens, M. 1998. Lexical Perspectives on Transitivity and Ergativity. Causative Constructions in English. Amsterdam/ Philadelphia: J. Benjamins. —. 1999. “Diachronic perspectives on lexical and constructional interdependency in English”. http://www.univ-lille3.fr/Recherche/silex/lemmens. Levin, B. 1993. English Verb Classes and Alternations: a Preliminary Investigation. Chicago: University of Chicago Press. —. 1999. “Objecthood: an event structure perspective”. Proceedings of the 35th Annual Meeting of the Chicago Linguistic Society - Part 1: Papers from the Main Session: 223-247. —. 2000. “Aspect, lexical semantic representation, and argument expression”. Proceedings of the 26th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 413429. —. 2004. “Verbs and constructions: where next?”. Presentation at the Western Conference on Linguistics, University of Southern California, Los Angeles, CA, 12-14 November 2004. —. 2006. “First objects and datives: two of a kind?”. Parasession on Theoretical Approaches to Argument Structure, 32nd Annual Meeting of the Berkeley Linguistics Society. University of California, Berkeley, CA, 10-12 February 2006. Levin, B. & Rappaport Hovav, M. 1988. “Non-event -er nominals: a probe into argument structure”. Linguistics 26(6): 1067-1083. —. 1989. “An approach to unaccusative mismatches”. NELS 19: 314-328. —. 1992. “The lexical semantics of verbs of motion: the perspective from unaccusativity”. In I. M. Roca (ed.), Thematic Structure: its Role in Grammar. Berlin: Foris, 247-269. —. 1994. “A preliminary analysis of causative verbs in English”. Lingua 92: 35-77. —. 1995. Unaccusativity: at the Syntax-Lexical Semantics Interface. Linguistic Inquiry Monographs 26. Cambridge, Mǹ: MIT Press. —. 1999. “Two structures for compositionally derived events”. Proceedings of SALT 9: 199-223. —. 2005. Argument Realisation. Cambridge: Cambridge University Press. —. 2006a. “Constraints on the complexity of verb meaning and VP structure”. In H.-M. Gärtner, R. Eckardt, R. Musan, & B. Stiebels (eds.), Between 40 and 60 Puzzles for Krifka. http://www.zas. gwz-berlin.de/40-60-puzzles-for-krifka —. 2006b. “Dative verbs: a crosslinguistic perspective”. Presentation at the Lexis and Grammar Conference, Palermo, Italy, September 6-9, 2006. Levin, L. 1986. Operations on Lexical Forms: Unaccusative Rules in Germanic Languages. PhD Dissertation, MIT. Li, F. ȋ. 1991. “An examination of causative morphology from a cross-linguistic and diachronic perspective”. In L. M. Dobrin et al. (eds.), Papers from the 27th Regional Meeting of the Chicago Linguistic Society - Part 1: The General Session: 344-359. Liddell, H. G., Scott, R. & Jones, H. S. 1996. A Greek-English Lexicon with a Revised Supplement. Oxford: Clarendon Press. Lightfoot, D. W. 1979. Principles of Diachronic Syntax. Cambridge: Cambridge University Press. —. 1991. How to Set Parameters: Arguments from Language Change. Cambridge, MA: MIT Press. —. 1993a. “Degree-0 learnability”. University of Maryland Working Papers in Linguistics 1: 22-38. —. 1993b. “Why UG needs a learning theory: triggering verb movement”. In C. Jones (ed.), Historical Linguistics: Problems and Perspectives. London: Longman, 190-214. —. 1994a. “Degree-0 learnability”. In B. C. Lust, G. Hermon & J. Kornfilt (eds.), Syntactic Theory and First Language Acquisition. Vol. 2. Binding, Dependencies, and Learnability. Hillsdale, NJ: Erlbaum, 453-471. —. 1994b. “Shifting triggers and diachronic reanalyses”. University of Maryland Working Papers in Linguistics 2: 110-135. —. 1995. “Why UG needs a learning theory: triggering verb movement”. In A. Battye & I. G. Roberts (eds.), Clause Structure and Language Change. New York, NY: Oxford University Press, 31-52. —. 1997. “Catastrophic change and learning theory”. Lingua 100(1-4): 171-192. —. 1999. The Development of Language: Acquisition, Change, and Evolution. Oxford: Blackwell. —. 2002a. “Myths and the prehistory of grammars”. Journal of Linguistics 38(1): 113-136. —. 2002b. “More myths”. Journal of Linguistics 38(3): 619-626. —. 2004. “Abstraction and performance: commentary on Fischer”. Studies in Language 28(3): 741744. —. 2006a. How New Languages Emerge. Cambridge: Cambridge University Press. —. 2006b. “Cueing a new grammar”. In A. van Kemenade & B. Los (eds.), The Handbook of the History of English. Oxford: Blackwell, 24-44.

242

Transitivity Alternations in Diachrony

Lindsay, W. M. 1894. The Latin Language. Oxford: Clarendon Press. Lindstedt, J. 2002. “Is there a Balkan verb system?”. Balkanistica 15: 323-336. Ljungvik, H. 1932. Beitrage zur Syntax der Spatgriechischen Volkssprache. Uppsala: Almqvist &Wiksells. Longobardi, G. 2001. “The structure of DPs: some principles, parameters and problems”. In M. Baltin & C. Collins (eds.), Handbook of Contemporary Syntactic Theory. Oxford: Blackwell, 562-603. López-Eire, A. 1993. “De l’attique à la koiné”. In C. Brixhe (ed.), La koiné grecque antique: une langue introuvable? Nancy: Presses Universitaires de Nancy, 41-57. —. 1997. Retırica clĮsica y teorȞa literaria moderna. Madrid: Arco Libros. López-Ornat, S. 1994. La adquisición de la lengua española. Madrid: Siglo XXI. Lord, C. 1979. “Don’t you fall me down: children’s generalizations regarding cause and transitivity”. Papers and Reports on Child Language Development 17. Stanford, CA: Stanford University Department of Linguistics, 81-89. Lumsden, J. 1987. Syntactic Features: Parametric Variation in the History of English. PhD Dissertation, MIT. Luraghi, S. 2000. “Spatial metaphors and agenthood in Ancient Greek”. In C. Zinko & M. Offisch (eds.), 125 Jahre Indogermanistik in Graz. Graz: Leykam, 283-298. —. 2003. On the Meaning of Prepositions and Cases. The Expression of Semantic Roles in Ancient Greek. Amsterdam/ Philadelphia: J. Benjamins. Luraghi, S., Pompei, A. & Skopeteas, S. 2005. Ancient Greek. Munich: LINCOM Europa. Lyons, J. 1968. Introduction to Theoretical Linguistics. Cambridge: Cambridge University Press. —. 1977. Semantics. Cambridge: Cambridge University Press. Macfarland, T. 1995. Cognate Objects and the Argument/Adjunct Distinction in English. PhD Dissertation, Northwestern University. Mackridge, P. 1985. The Modern Greek Language: a Descriptive Analysis of Standard Modern Greek. Oxford: Oxford University Press. —. 1990. “Katharevousa (c.1800–1974): an obituary for an official language”. In Ȃ. Sarafis & M. Eve (eds.), Background to Contemporary Greece. London: Merlin Press, 25-51. —. 1993. “An editorial problem in medieval Greek: the position of the object clitic pronoun in the Escorial Digenes Akrites.” In N. M. Panayiotakis (ed.), AȡȤȑȢ IJȘȢ ȃǼ ȁȠȖȠIJİȤȞȓĮȢ/ Origini della letteratura neogreca. Venice, 325-342. MacMahon, A. 1994. Understanding Language Change. Cambridge: Cambridge University Press. Mahajan, A. 1994. “Active passives”. In R. Aronovich, W. Byrne, S. Preuss & M. Senturia (eds.), Proceedings of the Thirteenth West Coast Conference on Formal Linguistics. Stanford CA: CSLI Publications, 286-301. Mallen, E. 1990. “Clitic movement inside noun phrases”. Studia Linguistica 44: 1-29. Mandilaras, V. 1973. The Verb in the Greek Non-Literary Papyri. Athens. —. 19934. Ǿ įȠȝȒ IJȘȢ ĮȡȤĮȓĮȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. Athens: Kardamitsas. Manney, L. 2000. Middle Voice in Modern Greek: Meaning and Function of an Inflectional Category. Amsterdam/ Philadelphia: J. Benjamins. Manolessou, I. 1999. “ǹʌȩ IJȘȞ ȚıIJȠȡȓĮ IJȠȣ ȐȡșȡȠȣ”. In A. Moser (ed.), Proceedings of the 3rd International Conference on Greek Linguistics. Athens: Ellinika Grammata, 492-499. —. 2000. Greek Noun Phrase Structure: a Study in Syntactic Evolution. PhD Dissertation, University of Cambridge. —. 2001. “The evolution of the demonstrative system in Greek”. Journal of Greek Linguistics 2: 119148. —. 2003. “Ǿ ĮȟȓĮ IJȦȞ ȝȘ ȜȠȖȠIJİȤȞȚțȫȞ ʌȘȖȫȞ ȖȚĮ IJȘȞ ȑȡİȣȞĮ IJȘȢ ȂİıĮȚȦȞȚțȒȢ ǼȜȜȘȞȚțȒȢ”. ȁİȟȚțȠȖȡĮijȚțȩȞ ǻİȜIJȓȠȞ 24: 61-88. —. 2005. “From participles to gerunds”. In M. Stavrou & A. Terzi (eds.), Advances in Greek Generative Syntax: in Honor of Dimitra Theophanopoulou-Kontou. Amsterdam/ Philadelphia: J. Benjamins, 241-283. Manousakas, M. I. 1961. “ǺİȞİIJȚțȐ ȑȖȖȡĮijĮ ĮȞĮijİȡȩȝİȞĮ İȚȢ IJȘȞ İțțȜȘıȚĮıIJȚțȒȞ ȚıIJȠȡȓĮȞ IJȘȢ ȀȡȒIJȘȢ IJȠȣ 14Ƞȣ-16Ƞȣ ĮȚ. (ȆȡȦIJȠʌĮʌȐįİȢ țĮȚ ʌȡȦIJȠȥȐȜIJĮȚ ȋȐȞįĮțȠȢ)”. ǻİȜIJȓȠȞ IJȘȢ ǿıIJȠȡȚțȒȢ țĮȚ ǼșȞȠȜȠȖȚțȒȢ ǼIJĮȚȡİȓĮȢ IJȘȢ ǼȜȜȐįȠȢ 15: 149-233. —. 1964. “ǼȜȜȘȞȚțȐ ȞȠIJĮȡȚĮțȐ ȑȖȖȡĮijĮ Įʌȩ IJĮ “Atti antichi” IJȠȣ ĮȡȤİȓȠȣ IJȠȣ ǻȠȪțĮ IJȘȢ ȀȡȒIJȘȢ”. ĬȘıĮȣȡȓıȝĮIJĮ 3: 73-102. —. 1995. ȁİȠȞȐȡįȠȣ ȃIJİȜȜĮʌȩȡIJĮ ȆȠȚȒȝĮIJĮ (1403 / 1411). Athens: Academy of Athens. Manzini, M. R., Roussou, A. & Savoia, L. M. 2009-to appear. “The morphosyntax of non-active voice in Greek and Albanian”. Proceedings of Mediterranean Syntax Meeting 2. Leiden: Brill. Marantz, A. 1982. “Re reduplication”. Linguistic Inquiry 13: 435-482.

Bibliography

243

—. 1984. On the Nature of Grammatical Relations. Cambridge, MA: MIT Press. —. 1993. “Implications of asymmetries in double object constructions”. In S. A. Mchombo (ed.), Theoretical Aspects of Bantu Grammar. Stanford: CSLI Publications, 113-150. —. 1997. “No escape from syntax: don’t try morphological analysis in the privacy of your own lexicon”. In A. Dimitriadis, L. Siegel, C. Surek-Clark & A. Williams (eds.), University of Pennsylvania Working Papers in Linguistics, vol. 4.2, Proceedings of the 21st Annual Penn Linguistics Colloquium. Philadelphia, 201-225. —. 2005. “Objects out of the lexicon: objects as event”. Presentation at the University of Vienna, June 11, 2005. Maratsos, M., Fox, D., Becker, J. & Chalkey, M. A. 1985. “Semantic restrictions on children’s passives”. Cognition 19: 167-191. Maratsos, M., Gudeman, R., Gerard-Ngo, P., & DeHart, G. 1987. “A study in novel word learning: the productivity of the causative”. In B. MacWhinney (ed.), Mechanisms of Language Acquisition. Hillsdale, NJ: Lawrence Erlbaum, 89-113. Marchand, H. 1964a. “A set of criteria for the establishing of derivational relationship between words unmarked by derivational morphemes”. Indogermanische Forschungen 69: 10-19. —. 1964b. “Die Ableitung desubstantivischer Verben mit Nullmorphem im Englischen, Französischen und Deutschen”. Die Neueren Sprachen 63: 105-118. —. 1969. The Categories and Types of Present-Day English Word-Formation. München: Beck. Markantonatou, S. 1992. “ȇȘȝĮIJȚțȐ ȠȣıȚĮıIJȚțȐ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 13: 235-254. —. 1993. The Syntax of Modern Greek NPs with a Deverbal Head. PhD Dissertation, University of Essex. Martinet, A. 1979. Grammaire fonctionnelle du Francais. Paris: CREDIF. Mastrodimitris, P. D. 2006. ǼȚıĮȖȦȖȒ ıIJȘ ȞİȠİȜȜȘȞȚțȒ ijȚȜȠȜȠȖȓĮ. 6th edition. Athens: Domos. Mavromanolaki, G. 2002. “ȇȒȝĮIJĮ ĮȜȜĮȖȒȢ țĮIJȐıIJĮıȘȢ țĮȚ țȡȚIJȒȡȚĮ ĮȞĮȚIJȚĮIJȚțȩIJȘIJĮȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 22: 429-439. Mayser, E. 1929/1970. Grammatik der Griechischen Papyri aus der Ptolemaerzeit. Berlin: Mouton de Gruyter. —. 1934/1970. Grammatik der Griechischen Papyri aus der Ptolemäerzeit. Band II 1 Satzlehre. Berlin: Mouton de Gruyter. McCawley, J. D. 1968/1973. “Lexical insertion in a transformational grammar without deep structure”. Chicago Linguistic Society 4: 71-80. / Published also in J. D. McCawley 1973, 154-166. —. 1971/1973. “Prelexical syntax”. Report of the 22nd Annual Roundtable Meeting on Linguistics and Language Studies. Washington D.C.: Georgetown University Press, 19-33. / Published also in J. D. McCawley 1973, 343-356. —. 1973: Grammar and Meaning. Tokyo: Taishukan. —. 1976/1979. “Remarks on what can cause what”. In M. Shibatani (ed.), The Grammar of Causative Constructions. Syntax and Semantics 6. New York: Academic Press, 117-129. Published also in J. McCawley (ed.), Adverbs, Vowels, and Other Objects of Wonder. Chicago: University of Chicago Press, 101-112. McKoon, G. & Macfarland, T. 2000. “Externally and internally caused change-of-state verbs”. Language 76: 833-858. McNulty, E. 1988. The Syntax of Adjunct Predicates. PhD Dissertation, University of Connecticut. Megas, G. A. 1975. īİȦȡȖȓȠȣ ȋȠȪȝȞȠȣ, Ǿ țȠıȝȠȖȑȞȞȘıȚȢ, ĮȞȑțįȠIJȠȞ ıIJȚȤȠȪȡȖȘȝĮ IJȠȣ Țİǯ ĮȚȫȞȠȢ. ɤȝȝİIJȡȠȢ ʌĮȡȐijȡĮıȚȢ IJ̎Ȣ īİȞȑıİȦȢ țĮˀ ɢȟȩįȠȣ IJ̎Ȣ ȆĮȜĮȚ˾Ȣ ǻȚĮșȒțȘȢ. Athens: Academy of Athens. Meier-Brugger, M. 1992. Griechische Sprachwissenschaft. Vol. 1. Berlin/New York: Mouton de Gruyter. Meillet, A. 1903/1964. Introduction à l’étude comparative des langues indo-européennes. [1937: Paris: Hachette /1964: University of Alabama Press.] Melis, L. 1990a. “Pronominal verbs in Old and Modern French, or How prototypes can be restructured on the basis of permanent meaning effects”. Belgian Journal of Linguistics 5: 87-108. —. 1990b. La voie pronominale. La systématique des tours pronominaux en français moderne. Louvain-la-Neuve/ Paris: Duculot. Méndez Dosuna, J. 2000. “Review of Horrocks 1997”. Journal of Greek Linguistics 1: 274-295. Mereu, L. (ed.) 1999. Boundaries of Morphology and Syntax. Amsterdam/ Philadelphia: J. Benjamins. Mertzios, K. D. 1961-2. “ȈIJĮȤȣȠȜȠȖȒȝĮIJĮ Įʌȩ IJĮ țĮIJȐıIJȚȤĮ IJȠȣ ȞȠIJĮȡȓȠȣ ȀȡȒIJȘȢ ȂȚȤĮȒȜ ȂĮȡȐ (1538-1578)”. ȀȡȘIJȚțȐ ȋȡȠȞȚțȐ 15-16: 228-308. Meyer-Lübke, W. 1899. Grammatik der romanischen Sprachen. Leipzig: Reisland.

244

Transitivity Alternations in Diachrony

Michaelis, S. 1998. “Antikausativ als Brücke zum Passiv: fieri, venire und se im Vulgärlateinischen und Altitalienischen”. In W. Dahmen et al. (eds.), Neuere Beschreibungsmethoden der Syntax romanischer Sprachen. Tübingen: Narr, 69-98. Migne, J.-P. 1857-66. Patrologiae cursus completus (series Graeca) 87: 2852-3112. Paris: Garnier. Mihevc-Gabrovec, E. 1960. Études sur la syntaxe de Ioannes Moschos. Razprave Univerze v Ljubljani. Miklosich, F. & Müller, J. 1860-90. Acta et diplomata graeca medii aevi sacra et profana. Vienna: C. Gerold. —. 1865. Acta et diplomata graeca medii aevi, sacra et profana. Vol. III. Vienna: C. Gerold. Miller, D. G. 1993. Complex Verb Formation. Amsterdam/ Philadelphia: J. Benjamins. Miller, D. G. 2001. “Innovation of the indirect reflexive in Old French”. In L. J. Brinton (ed.), Historical Linguistics 1999, Selected Papers from the 14th International Conference on Historical Linguistics, Vancouver, 9-13 August 1999. Amsterdam/ Philadelphia: J. Benjamins, 223-239. Minas, K. 1994. Ǿ ȖȜȫııĮ IJȦȞ įȘȝȠıȚİȣȝȑȞȦȞ ȝİıĮȚȦȞȚțȫȞ İȜȜȘȞȚțȫȞ İȖȖȡȐijȦȞ IJȘȢ ȀȐIJȦ ǿIJĮȜȓĮȢ țĮȚ IJȘȢ ȈȚțİȜȓĮȢ. Athens. Mirambel, A. 1959. La langue grecque moderne, description et analyse. Paris: Klincksieck. Mitchell, B. 1985. Old English Syntax. Oxford: Clarendon Press. Mitsakis, K. 1967. The Language of Romanos the Melodist. Byzantinisches Archiv Heft 11. Munich: Beck. Mittwoch, A. 1991. “In defence of Vendler’s achievements”. Belgium Journal of Linguistics 6: 71-86. —. 1998. “Cognate objects as reflections of davidsonian event arguments”. In S. Rothstein (ed.), Events and Grammar. Dordrect: Kluwer, 309-332. Moessner, L. 1994. “Early Modern English passive constructions”. In D. Kastovsky (ed.), Studies in Early Modern English. Berlin/New York: Mouton de Gruyter, 217-232. Möhlig, R. & Klages, M. 2002. “Detransitivization in the history of English from a semantic perspective”. In T. Fanego & M. J. López-Couso (eds.), English Historical Syntax and Morphology. Selected Papers from 11 ICEHL, Santiago de Compostela, 7-11 September 2000. Amsterdam/ Philadelphia: J. Benjamins, 231-254. Monge, F. 1955. “Las frases pronominales de sentido impersonal en español”. Archivo de Filología Aragonesa 7: 1-102. Montrul, S. 1997. Transitivity Alternations in Second Language Acquisition: a Crosslinguistic Study of English, Spanish and Turkish. PhD Dissertation, McGill University. Moorhouse, A. C. 1982. The Syntax of Sophocles. Leiden: E. J. Brill. Moravcsik, E. A. 1978. “On the distribution of ergative and accusative patterns”. Lingua 45: 233-279. Moravcsik, G. & Jenkins, R. J. H. 1967. Constantine Porphyrogenitus. De administrando imperio. 2nd edition . Washington, D.C.: Dumbarton Oaks. Morikawa, M. 1996. “Passives in Japanese”. Journal of Sugiyama Jogakuen University. Bulletin of Universities and Institutes 24. Moser, A. 1988. The History of Perfect Periphrases in Greek. PhD Dissertation, University of Cambridge. —. 1993. “īȡĮȝȝĮIJȚțȠʌȠȓȘıȘ țĮȚ ȕȠȘșȘIJȚțȐ ȡȒȝĮIJĮ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 14: 161-175. —. 1994a. “The interaction of lexical and grammatical aspect in Modern Greek”. In I. PhilippakiWarburton, K. Nikolaidis & M. Sifianou (eds.), Themes in Greek Linguistics. Papers from the First International Conference on Greek Linguistics, Reading, September 1993. Amsterdam/ Philadelphia: J. Benjamins, 137-144. —. 1994b. ȆȠȚȩȞ țĮȚ ĮʌȩȥİȚȢ IJȠȣ ȡȒȝĮIJȠȢ. Parousia 30. National and Kapodistrian University of Athens. —. 1996. “ȉȠ ʌȠȚȩȞ İȞİȡȖİȓĮȢ Ȓ ȐʌȠȥȘ IJȠȣ ȡȒȝĮIJȠȢ IJȘȢ ǼȜȜȘȞȚțȒȢ”. In Ǿ ȃȑĮ ǼȜȜȘȞȚțȒ ȦȢ ȟȑȞȘ ȖȜȫııĮ. ȆȡȠȕȜȒȝĮIJĮ įȚįĮıțĮȜȓĮȢ. Athens: Goulandri-Chorn Foundation, 77-96. —. 2005. DZʌȠȥȘ țĮȚ ȤȡȩȞȠȢ ıIJȘȞ ȚıIJȠȡȓĮ IJȘȢ ǼȜȜȘȞȚțȒȢ. Parousia. National and Kapodistrian University of Athens. Moser, A. & Chila-Markopoulou, D. 2001a. “ȉİȜȚțȩIJȘIJĮ țĮȚ ĮȞĮijȠȡȚțȩIJȘIJĮ ıIJȘ ȡȘȝĮIJȚțȒ ijȡȐıȘ IJȘȢ ȃǼ: ʌȠȚȩȞ İȞİȡȖİȓĮȢ țĮȚ ȐȡșȡȠ”. In G. Aggouraki et al. (eds.), Proceedings of the 4th International Conference on Greek Linguistics. Thessaloniki: University Studio Press, 138-145. —. 2001b. “īȜȦııȚțȒ İȟȑȜȚȟȘ țĮȚ ȖȜȦııȚțȒ ʌȡĮȖȝĮIJȚțȩIJȘIJĮ”. [Review of: G. Horrocks. 1997: Greek: a History of the Language and its Speakers. London: Longman.] Glossologia 13: 145-173. Motsios, G. 1990. ǼȜȜȘȞȚțȒ ʌİȗȠȖȡĮijȓĮ 1600-1821. Athens: Grigoris. Moulton, J. H. 1908. A Grammar of NT Greek. Edinburgh: Clark. Moulton, J. H. & Turner, ȃ. 1963. A Grammar of New Testament Greek. 3. Syntax. Edinburgh: T. & T. Clark.

Bibliography

245

Mustanoja, T. F. 1960. A Middle English Syntax. Part 1. Parts of Speech. Helsinki: Société Néophilologique. Næss, Å. 2007. Prototypical Transitivity. Amsterdam/ Philadelphia: J. Benjamins. Nakas, Th. 1987. ȉĮ İʌȚȡȡȘȝĮIJȚțȐ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ. ȆȡȠȕȜȒȝĮIJĮ ȣʌȠțĮIJȘȖȠȡȚȠʌȠȓȘıȘȢ. PhD Dissertation, National and Kapodistrian University of Athens. Naro, A. 1968. History of Portuguese Passives and Impersonals. PhD Dissertation, MIT. Nedjalkov, V. P. & Sil'nickij, G. G. 1969. “Tipologija kauzativnych konstrukcij. Tipologija morfologiþeskogo i leksiþeskogo kauzativov”. In A. A. Cholodoviþ (ed.), Tipologija Ȁauzativnych Ȁonstrukcij. Morfologiþeskij Ȁauzativ. Leningrad: Nauka, 5-50. —. 1988. “Resultative, passive, and perfect in German”. In V. P. Nedjalkov (ed.), Typology of Resultative Constructions. Amsterdam/ Philadelphia: J. Benjamins, 411-433. Nedyalkov, I. V. 1991. “Recessive-accessive polysemy of verbal affixes”. Languages of the World 1: 431. Neu, E. 1968a. Das hethitische Mediopassiv und seine indogermanischen Grundlagen. Wiesbaden: Harrassowitz. —. 1968b. Interpretation der hethitischen mediopassiven Verbalformen. Wiesbaden: Harrassowitz. Newton, B. 1972. “The dialect geography of Modern Greek passive inflections”. Glotta 50: 262-289. Nichols, J. 1993a. “Ergativity and linguistic geography”. Australian Journal of Linguistics 13: 39-89. —. 1993b. “Transitive and causative in the Slavic lexicon: evidence from Russian”. In B. Comrie et al. (eds.), Causatives and Transitivity. Amsterdam/ Philadelphia: J. Benjamins, 69-86. Niehoff-Panagiotidis, J. 1994. Koine und Diglossie. Mediterranean Language and Culture Monograph Series 10. Wiesbaden: Harrassowitz. Nikiforidou, K. 1991. “The meanings of the genitive: a case study in semantic structure and semantic change”. Cognitive Linguistics 2: 149-205. —. 2007. “Language change”. In A.-Ph. Christidis (ed.), A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press, 124-131. Norberg, D. 1943. Syntaktische Forschungen auf dem Gebiete des Spätlateins und des frühen Mittellateins. Uppsala. —. 1944. Beiträge zur spätlateinischen Syntax. Uppsala. Nouchoutidou, E. 2003. “ȆĮȡĮIJȘȡȒıİȚȢ ıIJȘȞ țĮIJȐțIJȘıȘ IJȦȞ ȝȑıȦȞ-ʌĮșȘIJȚțȫȞ țĮȚ IJȦȞ ĮȣIJȠʌĮșȫȞ ȡȘȝȐIJȦȞ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ”. In N. Lavidas et al. (eds.), APC-1. Athens: National and Kapodistrian University of Athens, 31-42. Nuñes, M. L. 1993. “Argument linking in English derived nominals”. In R. D. Van Valin (ed.), Advances in Role and Reference Grammar. Amsterdam/ Philadelphia: J. Benjamins, 375-432. Oertel, H. 1926. The Syntax of Cases in the Narrative and Descriptive Prose of the BrƗhmaas. Heidelberg. Oesterreicher, W. 1992. “SE im Spanischen. Pseudoreflexivität, Diathese und Prototypikalität von semantischen Rollen”. Romanistisches Jahrbuch 43: 237-260. Olsen, M. B. 1996. “Telicity and English verb classes and alternations: an overview”. Computer Science Technical Report Series. CS-TR-3607, UMIACS-TR-96-15. College Park: University of Maryland. —. 1997. A Semantic and Pragmatic Model of Lexical and Grammatical Aspect. New York: Garland. Olsson, B. 1925. Papyrusbriefe aus der frühesten Römerzeit. Uppsala. Ono, K. 1999. “Intransitive - transitive phrase pairs in Japanese”. Theoretical Linguistics 25: 15-29. Orban, A. P. 1974. “Die Entwicklung des Mediums vom Lateinischen zu den romanischen Sprachen mit besonderer Rucksicht auf die Theorien der lateinischen Grammatiker”. Zeitschrift fur Vergleichende Sprachforschung 88: 235-45. Oxford English Dictionary (OED) 2009, 4th edition. P.Cair.Zen. = Zenon Papyri, Catalogue général des antiquités égyptiennes du Musée du Caire. Ed.: C. C. Edgar. Cairo. Vol. 1-5 (1925-1940). P.Oxy. = The Oxyrhynchus Papyri. London. Vol. 1-68 (1898-2003). P.Zen.Pestm. = Greek and Demotic Texts from the Zenon Archive. Ed.: P. W. Pestman. Leiden: 1980. Pagkalos, G. 1955. Ȇİȡȓ IJȠȣ ȖȜȦııȚțȠȪ ȚįȚȫȝĮIJȠȢ IJȘȢ ȀȡȒIJȘȢ. Athens. Palm, J. 1955. Über Sprache und Stil des Diodoros von Sizilien. Lund: Gleerup. Palmer, L. R. 1963. The Interpretation of Mycenaean Greek Texts. Oxford: Clarendon Press. —. 1980. The Greek Language. London: Faber & Faber. Panagiotakis, N. 1990. ĭȡĮȖțȓıțȠȢ ȁİȠȞIJĮȡȓIJȘȢ. ȀȡȘIJȚțȩȢ ȝȠȣıȚțȠıȣȞșȑIJȘȢ IJȠȣ įȑțĮIJȠȣ ȑțIJȠȣ ĮȚȫȞĮ. ȂĮȡIJȣȡȓİȢ ȖȚĮ IJȘ ȗȦȒ țĮȚ IJȠ ȑȡȖȠ IJȠȣ. Venice. Panayotou, A. 1990. La langue des inscriptions de Macédoine (IVe s. a. C.-VIIe s. p. C.). Phonétique, phonologie et morphologie. Nancy.

246

Transitivity Alternations in Diachrony

—. 1994. “Ancient Greek dialects in Northern Greece”. In I. Philippaki-Warburton, K. Nicolaidis & M. Sifianou (eds.), Themes in Greek Linguistics. Papers from the First International Conference on Greek Linguistics, Reading, September 1993. Amsterdam/ Philadelphia: J. Benjamins, 421-426. Pantelides, N. 1999. “ȂȚĮ ʌȡȫIJȘ ȖȜȦııȠȖİȦȖȡĮijȚțȒ ʌȡȠıȑȖȖȚıȘ ıIJĮ IJȑȡȝĮIJĮ IJȦȞ ȝİıȠʌĮșȘIJȚțȫȞ İȞİıIJȫIJĮ țĮȚ ʌĮȡĮIJĮIJȚțȠȪ IJȦȞ «ȕĮȡȣIJȩȞȦȞ» ȡȘȝȐIJȦȞ ıIJȘȞ ȆİȜȠʌȩȞȞȘıȠ”. ȁİȟȚțȠȖȡĮijȚțȩȞ ǻİȜIJȓȠȞ 22: 133-203. —. 2003. “The active imperfect of the verbs of the 2nd conjugation in the Peloponnesian varieties of Modern Greek”. Journal of Greek Linguistics 4: 3-43. —. 2008. “Ǿ 'İȞȠʌȠȓȘıȘ IJȠȣ ʌĮȡȦȤȘȝȑȞȠȣ': ȚıIJȠȡȚțȒ ʌȡȠıȑȖȖȚıȘ ȝİ ȕȐıȘ IJȘ ȝĮȡIJȣȡȓĮ IJȦȞ ȞİȠİȜȜȘȞȚțȫȞ įȚĮȜȑțIJȦȞ”. In A. Moser, E. Bakakou-Orfanou, Ch. Charalambakis & D. ChilaMarkopoulou (eds.), īȜȫııȘȢ ȋȐȡȚȞ. Athens: Ellinika Grammata. Papadimitriou, S. D. 1896. Stefan Sakhlikis i ego stikhotvorenie ‘AijȒȖȘıȚȢ ȆĮȡȐȟİȞȠȢ’. Odessa. Papadopoulos, Th. 1977. ȃȚțȠȜȐȠȣ ȈȠijȚĮȞȠȪ īȡĮȝȝĮIJȚțȒ IJȘȢ țȠȚȞȒȢ IJȦȞ ǼȜȜȒȞȦȞ īȜȫııĮȢ. Athens: Kedros. Papadopoulos-Kerameus, A. 1909. Varia graeca sacra. St. Petersburg: Kirschbaum. Papanastasiou, G. C. 1998. “ȇȘȝĮIJȚțȐ țȜȚIJȚțȐ İʌȚșȒȝĮIJĮ IJȘȢ ȚȞįȠİȣȡȦʌĮȧțȒȢ țĮȚ IJȘȢ ĮȡȤĮȓĮȢ İȜȜȘȞȚțȒȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 18: 401-413. —. 2006. “Ǿ ʌȡȩșİıȘ İȞ ıIJȘ ıȪȞșİıȘ: ǼȜȜȘȞȚțȒ, ȁĮIJȚȞȚțȒ țĮȚ ȉİȣIJȠȞȚțȑȢ ȖȜȫııİȢ”. Proceedings of the 6th International Conference on Greek Linguistics. http://www.philology.uoc.gr/conferences/ 6thICGL/ —. 2007. “Morphology: from Classical Greek to Koine”. In A.-Ph. Christidis (ed.), A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press, 610-617. Papanastasiou, G. C. & Petrounias, E. B. 2007. “The morphology of Classical Greek”. In A.-Ph. Christidis (ed.), A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press, 571-589. Papangeli, D. 2004. The Morphosyntax of Argument Realization: Greek Argument Structure and the Lexicon-Syntax Interface. PhD Dissertation, University of Utrecht. Papathomopoulos, M. & Jeffreys, E. M. 1996. ȅ ȆȩȜİȝȠȢ IJȘȢ ȉȡȦȐįȠȢ. Athens: ȂǿǼȉ. Parker, F. 1976. “Language change and the passive voice”. Language 52: 449-460. Parry, M. 1998. “The reinterpretation of the reflexive in Piedmontese: impersonal se constructions”. Transactions of the Philological Society 96: 63-116. Parsons, T. 1990. Events in the Semantics of English. Cambridge, MA: MIT Press. Partee, Ǻ. 1965. Subject and Object in Modern English. PhD Dissertation, MIT. —. 1989. “Binding implicit variables in quantified contexts”. In C. Wilshire, R. Graczyk & B. Music (eds.), Papers from the 25th Regional Meeting of the Chicago Linguistics Society - Part I: The General Session: 342-365. Penny, R. 1991/2002. A History of the Spanish Language. Cambridge/ New York: Cambridge University Press. Perlmutter, D. 1978. “Impersonal passives and the Unaccusative Hypothesis”. Proceedings of the 4th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 157189. Perlmutter, D. & Postal, P. 1977. “Toward a universal characterization of passivization”. Proceedings of the 4th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 394-417. —. 1983. “Some proposed laws of basic clause structure”. In D. Perlmutter (ed.), Studies in Relational Grammar 1. Chicago: University of Chicago Press, 81-128. —. 1984. “Impersonal passives and some relational laws”. In D. Perlmutter & C. Rosen (eds.), Studies in Relational Grammar 2. Chicago: University of Chicago Press, 126-170. Perseus Digital Library Project. Ed. G. R. Crane. Tufts University. 1 December 2009. . Pesetsky, D. 1995. Zero Syntax. Experiencers and Cascades. Cambridge, MA: MIT Press. PHI #7 Greek Documentary Texts (Cd-Rom) (1991-1996). Packard Humanities Institute, Los Altos, California. PHib.: Grenfell, B. P. et al. 1906-1955. The Hibeh Papyri. London: Egypt Exploration Fund. Philippaki-Warburton, I. 1992a. ǼȚıĮȖȦȖȒ ıIJȘ ĬİȦȡȘIJȚțȒ īȜȦııȠȜȠȖȓĮ. Athens: Nefeli. —. 1992b. “On mood and complementizers in Modern Greek”. Reading University Working Papers in Linguistics 1: 5-41. —. 2007. “The syntax of Classical Greek”. In A.-Ph. Christidis (ed.), A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press, 752-757.

Bibliography

247

Philippides, D. M. 1986. Ǿ șȣıȓĮ IJȠȣ ǹȕȡĮȐȝ ıIJȠȞ ȘȜİțIJȡȠȞȚțȩ ȣʌȠȜȠȖȚıIJȒ. Athens: Ermis. Philippides, D. M. & Holton, D. 1996. ȉȠȣ țȪțȜȠȣ IJĮ ȖȣȡȓıȝĮIJĮ. ȅ ǼȡȦIJȩțȡȚIJȠȢ ıİ ȘȜİțIJȡȠȞȚțȒ ĮȞȐȜȣıȘ. Athens: Ermis. Pieris, Ȃ. & Nikolaou-Konnari, A. 2003. ȁİȠȞIJȓȠȣ ȂĮȤĮȚȡȐ ȋȡȠȞȚțȩ IJȘȢ ȀȪʌȡȠȣ. ȆĮȡȐȜȜȘȜȘ įȚʌȜȦȝĮIJȚțȒ ȑțįȠıȘ IJȦȞ ȤİȚȡȠȖȡȐijȦȞ. Nicosia: Center for Scientific Research. Pinker, S. 1984. Language Learnability and Language Development. Cambridge, MA: Harvard University Press. —. 1989. Learnability and Cognition: the Acquisition of Argument Structure. Cambridge, MA: MIT Press. Pinkster, H. 1985. “Latin cases and valency grammar - Some problems”. In C. Touratier (ed.), Syntaxe et Latin. Aix-en-Provence: Université de Provence, 163-189. Piñón, C. 2001. “A finer look at the causative-inchoative alternation”. In R. Hastings, B. Jackson & Z. Zvolenszky (eds.), Proceedings of Semantics and Linguistic Theory 11. Ithaca, N.Y.: CLC Publications. Pintzuk, S. 1999. Phrase Structures in Competition: Variation and Change in Old English Word Order. New York, NY: Garland. —. 2003. “Variationist approaches to syntactic change”. In B. Joseph & R. Janda (eds.), The Handbook of Historical Linguistics. Oxford: Blackwell, 509-528. Plank, F. 1985. “The extended accusative / restricted nominative in perspective”. In F. Plank (ed.), Relational Typology. Berlin/ New York: Mouton de Gruyter, 269-310. Plank, F. 1995. “Syntactic change: ergativity”. In J. Jacobs et al. (eds.), Syntax: an International Handbook of Contemporary Research. Vol. 2. Berlin/ New York: Mouton de Gruyter, 1184-1199. Pokorny, J. 1959. Indogermanisches etymologisches Wörterbuch. Bern: Francke. Politis, L. 1964. īİȦȡȖȓȠȣ ȋȠȡIJȐIJıȘ ȀĮIJȗȠȪȡȝʌȠȢ, țȦȝȦįȓĮ. Iraklion. Politis, N. G. 1978. ǼțȜȠȖĮȓ Įʌȩ IJĮ IJȡĮȖȠȪįȚĮ IJȠȣ İȜȜȘȞȚțȠȪ ȜĮȠȪ. Athens: Vagionakis. Pollock, J.-Y. 1989a. “Opacity, genitive subjects and extraction from NP in English and French”. Probus 1(2): 151-162. —. 1989b. “Verb movement, UG and the structure of IP”. Linguistic Inquiry 3: 21-58. Portius, S. 1638/1889. Grammatica Linguae Graecae Vulgaris. Paris: F. Vieweg. Premper, W. 1987. Kausativierung im Arabischen (Ein Beitrag zur sprächlichen Dimension der Partizipation). Arbeiten des Kölner Univesalien-Projekts [AKUP] 66. Köln: Institut für Sprachwissenschaft. Psaltes, S. 1913/1974. Grammatik der byzantinischen Chroniken. Göttingen: Vandenhoeck & Ruprecht. PTeb.: Grenfell, B. P. et al. 1902-1976. The Tebtunis Papyri. London/ NY: H. Frowde, Oxford University Press. Pustejovsky, J. 1991. “The syntax of event structure”. Cognition 41: 47-81. —. 1995. The Generative Lexicon. Cambridge, MA: MIT Press. Pye, C. 1991. “Learning to constrain verb transititvity”. In F. Ingemann (ed.), 1990 Mid-America Linguistics Conference Papers. Department of Linguistics, University of Kansas, 294-304. —. 1993. “A cross-linguistic approach to the causative alternation”. In Y. Levy (ed.), Other Children, Other Languages. Hillsdale, NJ: Erlbaum, 243-263. Pylkkänen, L. 2002. Introducing Arguments. PhD Dissertation, MIT. Rahlfs, A. 1935. Septuaginta. Stuttgart: Württembergische Bibelanstalt. Ralli, A. 1984. “ȂȠȡijȠȜȠȖȓĮ IJȠȣ İȜȜȘȞȚțȠȪ ȡȒȝĮIJȠȢ țĮȚ șİȦȡȓĮ IJȠȣ ȁİȟȚțȠȪ: ȝİȡȚțȑȢ ʌȡȠțĮIJĮȡțIJȚțȑȢ ʌĮȡĮIJȘȡȒıİȚȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 5: 61-80. —. 1986. “ȀȜȓıȘ țĮȚ ʌĮȡĮȖȦȖȒ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 7: 2949. —. 1988. Eléments de morphologie du grec moderne: la structure du verbe. PhD Dissertation, University of Montreal. —. 1996. “ȈȤȑıİȚȢ ıȪȞșİıȘȢ țĮȚ ʌȡȠıijȣȝĮIJȠʌȠȓȘıȘȢ: Ș ʌİȡȓʌIJȦıȘ IJȦȞ șİȝĮIJȚțȫȞ ȡȩȜȦȞ ıIJĮ ȡȘȝĮIJȚțȐ ıȪȞșİIJĮ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 17: 136-148. —. 2005. ȂȠȡijȠȜȠȖȓĮ. Athens: Patakis. —. 2009. “Morphology meets dialectology: insights from Modern Greek dialects”. Morphology 19 (1), 87-105. Ramchand, G. 1997. Aspect and Predication. Oxford: Clarendon Press. —. 2003. “First phase syntax”. Ms., University of Oxford. —. 2006a. “Direct and indirect causation in Hindi”. Paper presented at the Workshop on Clitics, Intonation and Causatives, University of Konstanz, 19 January 2007. —. 2006b. “Verb meaning and the lexicon”. Ms., University of Tromsø. —. 2008. Verb Meaning and the Lexicon. A First Phase Syntax. Cambridge: Cambridge University

248

Transitivity Alternations in Diachrony

Press. Rappaport Hovav, M. & Levin, B. 1988. “What to do with Theta-roles”. In W. Wilkins (ed.), Syntax and Semantics 21: Thematic Relations. New York: Academic Press, 7-36. —. 1992. “-er nominals: implications for a theory of argument structure”. In T. Stowell & E. Wehrli (eds.), Syntax and Semantics 26: Syntax and the Lexicon. New York: Academic Press, 127-153. —. 1998. “Building verb meanings”. In Ȃ. Ǻutt & W. Gender (eds.), The Projection of Arguments. Lexical and Compositional Factors. Stanford: CSLI Publications, 97-134. —. 2000. “Classifying single argument verbs”. In P. Coopmans, M. Everaert & J. Grimshaw (eds.), Lexical Specification and Insertion. Amsterdam/ Philadelphia: J. Benjamins, 269-304. —. 2001. “An event structure account of English resultatives”. Language 77: 766-797. —. 2007. “Deconstructing thematic hierarchies”. In A. Zaenen, J. Simpson, T. H. King, J. Grimshaw, J. Maling & C. Manning (eds.), Architectures, Rules, and Preferences: Variations on Themes by Joan W. Bresnan. Stanford, CA: CSLI Publications, 385-402. —. 2008. “The English dative alternation: the case for verb sensitivity”. Journal of Linguistics 44: 129167. Rappaport, M. 1989. “Unaccusativity and verbs of motion”. 2nd Symposium of English and Greek: Description and/ or Comparison of the two Languages. Thessaloniki: Aristotle University of Thessaloniki, 113-140. Reichenkron, G. 1933. “Passivum, Medium und Reflexivum in den Romanischen Sprachen”. Berliner Beiträge zur Romanischen Philologie 3: 1-69. Reinhart, T. 1991. “Lexical properties of ergativity”. Lecture presented at the Conference on Lexical Structure, Utrecht. http://www.let.uu.nl/~Tanya.Reinhart/personal/Papers/pdf/1991-lecture-utrecht. pdf —. 1997. “Syntactic effects of lexical operations: reflexive and unaccusatives”. OTS Working Papers in Linguistics, University of Utrecht, Utrecht, The Netherlands. —. 2000. “The theta system: syntactic realization of verbal concepts”. OTS Working Papers in Linguistics, University of Utrecht, Utrecht, The Netherlands. —. 2002. “The Theta System–An Overview”. Theoretical Linguistics 28(3): 229-290. Reinhart, T. & Siloni, ȉ. 2005. “The lexicon-syntax parameter: reflexivization and other arity operations”. Linguistic Inquiry 36(3): 389-436. Reis, M. 1973. “Is there a rule of subject-to-object raising in German?”. Chicago Linguistic Society 9: 519-529. Rijksbaron, A. 2002. ȉhe Syntax and Semantics of the Verb in Classical Greek. Amsterdam: J. C. Gieben. Rissanen, M. 1999. “Syntax”. In R. Lass (ed.), The Cambridge History of the English Langauge, III: 1476-1776. Cambridge: Cambridge University Press, 187-331. Ritter, E. & Rosen, S. T. 1998. “Delimiting events in syntax”. In M. Butt & W. Geuder (eds.), The Projection of Arguments: Lexical and Compositional Factors. Stanford, CA: CSLI Publications, 135-164. Rivero, M.-L. 1990. “The location of nonactive voice in Albanian and Modern Greek”. Linguistic Inquiry 21: 135-146. —. 1992. “Adverb incorporation and the syntax of adverbs in Modern Greek”. Linguistics and Philosophy 15: 289-331. Rix, H. 1976. Historische Grammatik des Griechischen. Darmstadt: Wissenschaftliche Buchgesellschaft. Roberts, I. 1987a. “On valency affecting rules”. In P. Cordin (ed.), Ipotesi e Applicazioni di Teoria Linguistica dal XIII Incontro di Grammatica Generativa. Dipartimento di Storia della Civilta Europea, Facolta di Lettere e Filosofia, Trento, 150-195. —. 1987b. The Representation of Implicit and Dethematized Subjects. Dordrecht: Foris. —. 1993. Verbs and Diachronic Syntax: a Comparative History of English and French. Dordrecht: Kluwer. Roberts, I. & Roussou, A. 2003. Syntactic Change. Minimalist Approach to Grammaticalization. Cambridge: Cambridge University Press. Robertson, A. T. 1934 [1919/1923]. A Grammar of the Greek New Testament in the Light of Historical Research. 4th edition. Nashville: Broadman. Robinson, J. A. 1891. The Passion of S. Perperua. Cambridge: Cambridge University Press. Roeper, T. 1987. “Implicit arguments and the head-complement relation”. Linguistic Inquiry 18(2): 267-310. —. 2000. “Inherent binding and the syntax-lexicon interface: distinguishing DP, NP, and N”. In P. Coopmans, M. Everaert & J. Grimshaw (eds.), Lexical Specification and Insertion. Amsterdam/

Bibliography

249

Philadelphia: J. Benjamins, 305-328. Roeper, T. & Siegel M. 1978. “A lexical transformation for verbal compounds”. Linguistic Inquiry 9(2): 199-260. Rohlfs, G. 1924: Griechen und Romanen in Unteritalien. Geneva: Leo S. Olschki. —. 1962: Neue Beiträge zur Kenntnis der unteritalienischen Gräzität. Munich: Bayerische Akademie der Wissenschaften. Rohrbacher, B. 1994. The Germanic Languages and the Full Paradigm: a Theory of V-to-I Raising. PhD Dissertation, University of Massachusetts, Amherst. Romaine, S. 1981. “The transparency principle: what it is and why it doesn’t work”. Lingua 55: 93116. Rooryc, J. & Zaring, L. 1996. Phrase Structure and the Lexicon. Dordrecht: Kluwer. Rosen, C. 1984. “The interface between semantic roles and initial grammatical relations”. In D. Perlmutter & C. Rosen (eds.), Studies in Relational Grammar. Chicago: University of Chicago Press, 38-77. Rothemberg, M. 1974. Les verbes à la fois transitifs et intransitifs en français contemporain. The Hague/ Paris: Mouton de Gruyter. Rothstein, S. 1983. The Syntactic Forms of Predication. PhD Dissertation, MIT. —. 1999. “Fine-grained structure in the eventuality domain: the semantics of predicative adjective phrases and be”. Natural Language Semantics 7: 347-420. Roussou, A. 2009. “Voice morphology and ergativity in Modern Greek”. Proceedings of the 8th International Conference on Greek Linguistics. http://www.linguist-uoi.gr/cd_web/docs/english /031_roussouICGL8_OK.pdf Roussou, A. & Tsimpli, I.-M. 2006. “On Greek VSO again”. Journal of Linguistics 42: 317-354. —. 2007. “Clitics and transitivity”. In A. Alexiadou (ed.), Studies in the Morpho-syntax of Greek. Newcastle: Cambridge Scholars Publishing, 138-174. Rouveret, A. 1978. “Result clauses and conditions on rules”. In S. J. Keyser (ed.), Recent Transformational Studies in European Languages. LI Monographs. Cambridge, MA: MIT Press, 159-187. Rozwadowska, B. 1988. “Thematic restrictions on derived nominals”. In W. Wilkings (ed.), Thematic Relations. Syntax and Semantics vol. 21. New York: Academic Press, 147-165. Ruijgh, C. L. 1975. “Review of C. G. Gual: El sistema diatetico en el verbo griego”. Lingua 36: 359370. Ruwet, N. 1972. Théorie syntaxique et syntaxe du français. Paris: Éditions du Seuil. Rydbeck, L. 1967. Fachprosa, Vermeintliche Volkssprache und Neues Testament. Acta Universitatis Upsaliensis. Uppsala: University of Uppsala. Rydén, L. 1963. Das Leben des heiligen Narren Symeon von Leontius von Neapolis. Stockholm: Almqvist & Wiksell. Sansò, ǹ. 2005/2009. “Grammaticalization paths or prototype effects? A history of the agentive reflexive passive in Italian”. Ms / Published in T. Mortelmans (ed.), Prototypes and Grammaticalization: Grammaticalization as Prototype?, Language Sciences (special issue). Santorini, B. 1989. The Generalization of the Verb Second Constraint in the History of Yiddish. PhD Dissertation, University of Pennsylvania. Sathas, K. 1877. ȂİıĮȚȦȞȚțȒ ǺȚȕȜȚȠșȒțȘ, Vol. VI: ǹıȓȗĮȚ IJȠȣ ǺĮıȚȜİȓȠȣ IJȦȞ ǿİȡȠıȠȜȪȝȦȞ țĮȚ IJȘȢ ȀȪʌȡȠȣ. Paris: Maisonneuve. Sawai, N. K. 1996. “An event structures analysis of 'ify/ize'”. Ms, University of Calgary. Calgary, Alberta. SB: Kießling, E. (ed.) 1963. Sammelbuch griechischer Urkunden aus Ägypten. 6. Band, Nr. 8964-9641. Wiesbaden: O. Harrassowitz. Schäfer, F. 2006. “Middles as voiced anticausatives”. Paper presented at Student Session of the “DGfS & GLOW Summer School”, University of Stuttgart. —. 2008. The Syntax of (Anti-)Causatives: External Arguments in the Change-of-State Contexts. Amsterdam/Philadelphia: John Benjamins. Schenk von Stauffenberg, A. 1931. Die römische Kaisergeschichte bei Malalas. Stuttgart: W. Kohlhammer Verlag. Schenkl, H. 1916. Epicteti dissertationes ab Arriano digestae. Leipzig: Teubner. Schmalstieg, W. 1982. “The shift of intransitive to transitive passive in the Lithuanian and IndoEuropean verb.” Baltistica 18: 119-134. Schmidt, K. H. 1963a. “Präteritum und Medio-Passiv”. Sprache 9: 14-20. —. 1963b. “Zum altirischen Passiv”. IF 68: 257-275. Schreiner, P. 1975. Die byzantinischen Kleinchroniken. Wien: Verlag der österreichischen Akademie

250

Transitivity Alternations in Diachrony

der Wissenschaften. Schwyzer, E. 1939/1953. Griechische Grammatik auf der Grundlage von Karl Brugmanns griechischer Grammatik. Vol. 1-3. Munich: C. H. Beck. —. 1943. Zum persönlichen Agens beim Passiv, besonders im Griechische. Abhandlungen der Preußischen Akademie der Wissenschaften. Philos.-histor. Klasse, Jahrgang 1942: Nr 10. Berlin: Akademie der Wissenschaften. —. 1953. Griechische Grammatik. Vol. 1, Lautlehre. Wortbildung, Flexion. Handbuch der Altertumswissenschaft 2.1.1. München: Beck. —. 1959. Griechische Grammatik. 3rd edition. München: Beck. Schwyzer, E. & Debrunner, A. 1950/2002. Griechische Grammatik. München: C. H. Beck. —. 1966. Griechische Grammatik. Band II, Syntax und Syntaktische Stilistik. Munich. SEG: Supplementum Epigraphicum Graecum. Selkirk, E. 1982. The Syntax of Words. LI Monographs. Cambridge, MA: MIT Press. Setatos, M. 1985. “ȉȡȠʌȚțȩIJȘIJİȢ IJȠȣ ȡȒȝĮIJȠȢ ıIJȘȞ țȠȚȞȒ ȞİȠİȜȜȘȞȚțȒ (ȀȃǼ)”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 6: 175-182. —. 1995. “ȅȚ ʌȡȠșȑıİȚȢ IJȘȢ țȠȚȞȒȢ ȞİȠİȜȜȘȞȚțȒȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 15: 33-58. —. 2007. “Semantic change”. In A.-Ph. Christidis (ed.), A History of Ancient Greek: from the Beginnings to Late Antiquity. Cambridge: Cambridge University Press, 667-676. Shibatani, M. 1985. “Passives and related constructions: a Prototype analysis”. Language 61: 821-848. Shibatani, M. & Pardeshi P. 2001. “The causative continuum”. In M. Shibatani (ed.), The Grammar of Causation and Interpersonal Manipulation. Amsterdam/ Philadelphia: J. Benjamins, 85-126. Siewierska, A. 1984. The Passive: a Comparative Linguistic Analysis. London: Croom Helm. Siloni, T. 1997. “Event nominals and the construct state”. In L. Haegeman (ed.), The New Comparative Syntax. London: Longman, 165-188. Sioupi, A. 1997. “Ǿ ʌĮȡĮȖȦȖȒ IJȦȞ įȠȝȫȞ ȝȑıȘȢ įȚȐșİıȘȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 17: 216-228. —. 2001. “The distribution of object bare singulars”. In G. Aggouraki et al. (eds.), Proceedings of the 4th International Conference on Greek Linguistics. Thessaloniki: University Studio Press, 292-299. Sklavounou, E. 2000. “ȉĮ İʌȓșİIJĮ IJȘȢ ȞȑĮȢ İȜȜȘȞȚțȒȢ ıİ -ȝȑȞȠȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 20. 499-510. Skopeteas, S. & Verhoeven, E. 2003. “ȈȣȞIJĮțIJȚțȒ țȣȡȚĮȡȤȓĮ IJȦȞ ȑȝȥȣȤȦȞ ıIJȘȞ ȚıIJȠȡȓĮ IJȘȢ ǼȜȜȘȞȚțȒȢ”. Proceedings of the 6th International Conference on Greek Linguistics. http://www. philology.uoc.gr/conferences/6thICGL/ Slabakova, R. 1997. “Bulgarian preverbs: aspect in phrase structure”. Linguistics 35: 673-704. Smirniotopoulos, J. C. 1991/1992. Lexical Passives in Modern Greek. PhD Dissertation, The Ohio State University / Outstanding Dissertations in Linguistics series, New York: Garland Press. Smith, C. S. 1970. “Jespersen’s ‘move’ and ‘change’ class and causative verbs”. In M. Ali Jazayery, E. Polomé & W. Winter (eds.), Linguistic and Literary Studies in Honor of Archibald A. Hill. Vol. 2: Descriptive Linguistics. The Hague: Mouton de Gruyter, 101-109. Smith, M. D. 1978. “The acquisition of word meaning: an introduction”. Child Development 49: 950952. Smith, O. L. 1999. The Byzantine Achilleid: the Naples version. Wien: Verlag der österreichischen Akademie der Wissenschaften. Smyth, H. W. 1920/1956/1974. Greek Grammar. Cambridge, MA: Harvard University Press. Snyder, W. 1998. “On the aspectual properties of English derived nominals”. In U. Sauerland & O. Percus (eds.), MIT Working Papers in Linguistics 25: 125-139. Sofianos, N. 1977 [1544]. īȡĮȝȝĮIJȚțȒ IJȘȢ țȠȚȞȒȢ IJȦȞ ǼȜȜȒȞȦȞ ȖȜȫııȘȢ. Athens: Kedros. Song, J. J. 1990. “On the rise of causative affixes: a universal-typological perspective”. Lingua 82: 151-200. —. 1996. Causatives and Causation: a Universal Typological Perspective. London: Longman. Sophocles, E. A. 1870/1914. Greek Lexicon of the Roman and Byzantine Periods. Cambridge, MA: Harvard University Press. Speas, M. 1990. Phrase Structure in Natural Language. Dordrecht: Kluwer. —. 1994. “Null arguments in a theory of economy of projection”. In E. Benedicto & J. Runner (eds.), Functional Projections. UMass Occasional Papers in Linguistics 17. Amherst, MA: UMass, GLSA, 179-208. —. 1994/2005. “Economy, agreement and the representation of null arguments”. http://www.people.umass.edu/pspeas/prodrop.pdf / In P. Ackema, P. Brandt, M. Schoorlemmer & F. Weermann (eds), Agreement and Argument Structure. Oxford: Oxford University Press, 35-75.

Bibliography

251

Sportiche, D. 1989. “Le mouvement syntaxique: constraintes et paramètres”. Langages 95: 35-80. —. 1996. “Clitic constructions”. In J. Rooryck & L. Zarrig (eds.), Phrase Structure and the Lexicon. Dordrecht: Kluwer, 223-277. —. 1998. Partitions and Atoms of Clause Structure. London: Routledge. Spyropoulos, V. & Philippaki-Warburton, I. 2001. “'Subject' and EPP in Greek”. Journal of Greek Linguistics 2: 149-186. Stamatakos, ǿ. 1949/ 1990. ǿıIJȠȡȚțȒ ȖȡĮȝȝĮIJȚțȒ ǹȡȤĮȓĮȢ ǼȜȜȘȞȚțȒȢ. Athens. Stassen, L. 1997. Intransitive Predication. Oxford: Clarendon Press. Stavrou, M. 199. “Adjectives in Modern Greek: an instance of predication, or an old issue revisited”. Journal of Linguistics 32: 76-112. —. 1999. “The position and serialization of APs in the DP”. In A. Alexiadou, G. Horrocks & M. Stavrou (eds.), Studies in Greek Syntax. Dordrecht: Kluwer, 201-226. Stechow, A. von 1995. “Lexical decomposition in syntax”. In U. Egli, P. Pause, C. Schwartze, A. von Stechow & G. Wienold (eds.), Lexical Knowledge in the Organization of Language. Amsterdam/ Philadelphia: J. Benjamins, 81-117. Stefanini, J. 1962. La voix pronominale en ancien et en moyen français. Publication des Annales de la Faculté des Lettres 31. Aix-en-Provence: Ophrys. Steinbach, M. 2002. Middle Voice: a Comparative Study in the Syntax-Semantics Interface. Amsterdam/ Philadelphia: J. Benjamins. Stempel, R. 1996. Die Diathese im Indogermanischen. Formen und Funktionen des Mediums und ihre sprachhistorischen Grundlagen. Innsbrucker Beiträge zur Sprachwissenschaft. Vorträge und kleinere Schriften 67. Innsbruck: Institit für Sprachwissenschaft. Svennung, J. 1941. Compositiones Lucenses. Studien zum Inhalt, zur Textkritik und Sprache. Uppsala. Sybesma, R. 1992. Causatives and Accomplishments: the Case of Chinese ba. Dordrecht: Holland Institute of Generative Linguistics. Symeonidis, Ch. P. 1999. “MİIJȐ IJȘȞ DZȜȦıȘ. ǼȚıĮȖȦȖȒ”. ǿıIJȠȡȓĮ IJȘȢ İȜȜȘȞȚțȒȢ ȖȜȫııĮȢ. Athens: E.L.I.A. Szabolcsi, A. 1994. “The noun phrase”. Syntax and Semantics 27: 179-274. Szemerényi, O. 1960. Studies in the Indo-European System of Numerals. Heidelberg: C. Winter. —. 1996. Introduction to Indo-European Linguistics. Oxford: Oxford University Press. Tabachovitz, D. 1946. “Phénomènes linguistiques du vieux grec dans le grec de la basse époque”. Museum Helveticum 3: 144-179. Talmy, L. 1976. “Semantic causative types”. In M. Shibatani (ed.), Syntax and Semantics 6: The Grammar of Causative Constructions. New York: Academic Press., 43-116. —. 1985. “Lexicalization patterns: semantic structure in lexical forms”. In T. Shopen (ed.), Language Typology and Syntactic Description. Vol. 3. Cambridge: Cambridge University Press, 57-149. —. 1988. “The relation of grammar to cognition”. In B. Rudzka-Ostyn (ed.), Topics in Cognitive Linguistics. Amsterdam/ Philadelphia: J. Benjamins, 165-205. Taplin, O. 1977. The Stagecraft of Aeschylus. Oxford: Clarendon Press. Taylor, A. 1994. “The change from SOV to SVO in Ancient Greek”. Language Variation and Change 6(1): 1-37. Tchekhoff, C. 1980. “The organization of a voice-neutral verb: an example in Avar”. International Review of Slavic Linguistics 5: 219-230. Tenny, C. L. 1987. Grammaticalizing Aspect and Affectedness. PhD Dissertation, MIT. —. 1992. “The aspectual interface hypothesis”. In I. A. Sag & A. Szabolcsi (eds.), Lexical Matters. Stanford, CA: CSLI Publications, 1-27. —. 1994. Aspectual Roles and the Syntax-Semantics Interface. Dordrecht: Kluwer. —. 1995a. “How motion verbs are special. The interaction of linguistic and pragmatic information in aspectual verb meanings”. Pragmatics and Cognition 3: 31-73. —. 1995b. “Modularity in thematic versus aspectual licensing. Paths and moved objects in motion verbs”. Canadian Journal of Linguistics 40: 201-234. Terzi, A. 1992. PRO in Finite Clauses. A Study of the Inflectional Heads of the Balkan Languages. PhD Dissertation, CUNY Graduate Center. —. 1996. “The linear correspondence axiom and the adjunction site of clitics”. In A. M. DiSciullo (ed.), Configurations: Essays on Structure and Interpretation. Somerville: Cascadilla Press, 185199. —. 1999. “Cypriot Greek clitics and their positioning restrictions”. In A. Alexiadou, G. Horrocks, M. Stavrou (eds.), Studies in Greek Syntax. Dordrecht: Kluwer, 227-240. Thavoris, A. I. 1971. Ǿ ȖȜȫııĮ ȝĮȢ ıIJĮ ȤȡȩȞȚĮ IJȘȢ ȉȠȣȡțȠțȡĮIJȓĮȢ. Ioannina: University of Ioannina.

252

Transitivity Alternations in Diachrony

Theophanopoulou-Kontou, D. 1973. “ȂİIJĮıȤȘȝĮIJȚıIJȚțȒ ĮȞȐȜȣıȚȢ IJȘȢ ĮʌȡȠıȫʌȠȣ țĮȚ ʌȡȠıȦʌȚțȒȢ ʌĮșȘIJȚțȒȢ ıȣȞIJȐȟİȦȢ ȜİțIJȚțȫȞ țĮȚ įȠȟĮıIJȚțȫȞ ȡȘȝȐIJȦȞ İȞ IJȘ ĮȡȤĮȓĮ İȜȜȘȞȚțȒ”. ǹșȘȞȐ 73-74: 625-660. —. 1980. “ȂȑıĮ ĮȣIJȠʌĮșȒ ȡȒȝĮIJĮ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ”. A Festschrift for ī. ǿ. ȀȠȣȡȝȠȪȜȘȢ, 1-21. —. 1982. “ȉĮ ȝȑıĮ ȡȒȝĮIJĮ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ: ȝİȡȚțȑȢ ʌȡȠțĮIJĮȡțIJȚțȑȢ ʌĮȡĮIJȘȡȒıİȚȢ ıIJȠ ıȪıIJȘȝĮ IJȦȞ įȚĮșȑıİȦȞ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 2: 51-78. —. 1983-4. “Patient vs non patient orientation of the action and the voice distinction in Modern Greek”. Glossologia 2-3: 75-90. —. 1986-87. “ȀİȞȑȢ țĮIJȘȖȠȡȓİȢ țĮȚ țȜȚIJȚțȐ ıIJȘ ȃȑĮ ǼȜȜȘȞȚțȒ: Ș ʌİȡȓʌIJȦıȘ IJȠȣ ȐȝİıȠȣ ĮȞIJȚțİȚȝȑȞȠȣ”. Glossologia 5-6: 41-68. —. 1989. ȂİIJĮıȤȘȝĮIJȚıIJȚțȒ ıȪȞIJĮȟȘ: Įʌȩ IJȘȞ șİȦȡȓĮ ıIJȘȞ ʌȡȐȟȘ. Athens: Kardamitsas. —. 1994. “Transformational grammar and Modern Greek syntax: an overview and some problematic cases”. In I. Philippaki-Warburton, K. Nikolaidis & M. Sifianou (eds.), Themes in Greek Linguistics. Papers from the First International Conference on Greek Linguistics, Reading, September 1993. Amsterdam/ Philadelphia: J. Benjamins, 11-20. —. 1999a. “Ǿ ĮȞĮIJȠȝȓĮ IJȠȣ ȜȐșȠȣȢ”. Proceedings of the 'Conference on the Greek Language 19761996'. Athens, 253-259. —. 1999b. “ȉĮ ȝȑıĮ ĮȝİIJȐȕĮIJĮ ȡȒȝĮIJĮ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ țĮȚ Ș ȝİıȠʌĮșȘIJȚțȒ țĮIJȐȜȘȟȘ”. In A. Moser (ed.), Proceedings of the 3rd International Conference on Greek Linguistics. Athens: Ellinika Grammata, 114-122. —. 2000a. “ǻȠȝȑȢ IJȘȢ ȃǼ ȝİ ʌȡȠȠʌIJȚțȒ IJȠȣ įȑțIJȘ țĮȚ Ș țĮIJĮȞȠȝȒ -Ȧ/ -ȝĮȚ: IJĮ ĮȞIJȚȝİIJĮȕȚȕĮıIJȚțȐ țĮȚ IJĮ ʌĮșȘIJȚțȐ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 20: 146-157. —. 2000b. “ȉȠʌȚțȐ İʌȚȡȡȒȝĮIJĮ țĮȚ 'ʌIJȫıȘ' ıIJȘȞ ǼȜȜȘȞȚțȒ: įȚĮȤȡȠȞȚțȒ ʌȡȠıȑȖȖȚıȘ”. Glossologia1112: 1-40. —. 2001. “ȁȐșȘ ıIJȘ ȤȡȒıȘ IJȘȢ ȖȜȫııĮȢ: ĮȜȒșİȚĮ țĮȚ ȝȪșȠȢ”. In G. Charis (ed.), ǻȑțĮ ȝȪșȠȚ ȖȚĮ IJȘȞ İȜȜȘȞȚțȒ ȖȜȫııĮ. Athens: Patakis, 53-61. —. 2002a. īİȞİIJȚțȒ ıȪȞIJĮȟȘ. ȉȠ ʌȡȩIJȣʌȠ IJȘȢ ȀȣȕȑȡȞȘıȘȢ țĮȚ ǹȞĮijȠȡȚțȒȢ ǻȑıȝİȣıȘȢ. Athens: Kardamitsas. —. 2002b. “ȂİIJĮȕȚȕĮıIJȚțȑȢ įȠȝȑȢ ıIJȘ ȃȑĮ ǼȜȜȘȞȚțȒ. DzȞĮ ʌĮȡĮȖȦȖȚțȩ ıȤȒȝĮ”. In Ch. Clairis (ed.), Recherches en linguistique greque. Paris: L’Harmattan, 243-246. —. 2003a. “ȉĮ ȡȒȝĮIJĮ țȓȞȘıȘȢ IJȘȢ ȃȑĮȢ ǼȜȜȘȞȚțȒȢ țĮȚ Ș ȝİIJĮȕȚȕĮıIJȚțȒ IJȠȣȢ ȤȡȒıȘ”. In D. Theophanopoulou-Kontou et al. (eds.), ȈȪȖȤȡȠȞİȢ IJȐıİȚȢ ıIJȘȞ ǼȜȜȘȞȚțȒ īȜȦııȠȜȠȖȓĮ: ȝİȜȑIJİȢ ĮijȚİȡȦȝȑȞİȢ ıIJȘȞ ǼȚȡȒȞȘ ĭȚȜȚʌʌȐțȘ- Warburton. Athens: Patakis, 237-255. —. 2003b. “ǼȟȑȜȚȟȘ țĮȚ ȝİIJĮȕȠȜȒ ıIJȘ ȖȜȫııĮ: șİȦȡȘIJȚțȒ ʌȡȠıȑȖȖȚıȘ-ʌȡȠȕȜȒȝĮIJĮ”. ȂȑȞIJȠȡĮȢ (special issue) 4: 62-73. —. 2004. “The structure of VP and the mediopassive morphology. The passives and anticausatives in Modern Greek”. Parousia 15-16: 173-206. Theophanopoulou-Kontou, D. et al. 1998. ȃİȠİȜȜȘȞȚțȒ ıȪȞIJĮȟȘ: șİȦȡȓĮ-ĮıțȒıİȚȢ. Athens: National and Kapodistrian University of Athens. Thumb, A. & Scherer, A. 1959. Handbuch der griechischen Dialekte. 2nd edition. Heidelberg: C. Winter. Thurn, I. 2000. Ioannis Malalae chronographia. Berlin/ New York: Mouton de Gruyter. Timberlake, A. 1977. “Reanalysis and actualization of syntactic change”. In C. N. Li (ed.), Mechanisms of Syntactic Change. Austin/ London: University of Texas Press. Tischendorf, C. 1866: Apocalypses Apocryphae. Leipzig: Mendelssohn. TLG: Thesaurus Linguae Graecae. University of California, Irvine. 1 December 2009. . Tonnet, H. 1993. Histoire du grec moderne. Paris: L’Asiathèque. Torrego, E. 1999. “Nominative subjects and pro-drop INFL”. Syntax 1(2): 206-219. Traugott, E. 1992. “Syntax”. In R. M. Hogg (ed.), The Cambridge History of the English Language, Vol. 1. Cambridge: Cambridge University Press. Travis, L. 1991. “Derived objects, inner aspect, and the structure of VP”. Paper presented at NELS 22. —. 1992. “Inner aspect and the structure of VP”. Cahiers de Linguistique de l’UQAM 1: 132-146. —. 1994. “Event phrase and a theory of functional categories”. Proceedings of the 1994 Annual Conference of the Canadian Linguistics Society, 559-570. —. 2000. “Event structure in syntax”. In C. Tenny and J. Pustejovsky (eds.), Events as Grammatical Objects: the Converging Perspectives of Lexical Semantics and Syntax. Stanford, CA: CSLI Publications, 145-185. Tsamadou-Jacoberger, I. 1998. Le nom en grec moderne: marqueurs et opérations de détermination.

Bibliography

253

Paris: L’Harmattan. Tsangalidis, A. 1999. Will and tha: a Comparative Study of the Category Future. PhD Dissertation, Aristotle University of Thessaloniki. Tsavati, I. 1987. ȅ ȆȠȣȜȠȜȩȖȠȢ. Athens: ȂǿǼȉ. Tsimpli, I.-M. 1989. “On the properties of the passive affix in Ȃodern Greek”. UCL Working papers in Linguistics 1: 235-260. —. 1990. “The clause structure and word-order in Modern Greek”. UCL Working Papers in Linguistics 2: 226-255. —. 1995. “Focusing in Modern Greek”. In E. Kiss (ed.), Discourse Configurational Languages. Oxford: Oxford University Press, 176-204. —. 2005. “Peripheral positions in early Greek”. In M. Stavrou & A. Terzi (eds.), Advances in Greek Generative Syntax: in Honor of Dimitra Theophanopoulou-Kontou. Amsterdam/ Philadelphia: J. Benjamins, 178-216. —. 2006a. “The acquisition of voice and transitivity alternations in Greek as a native and second language”. In S. Unshworth, T. Parodi, A. Sorace & M. Young-Scholten (eds.), Paths of Development in L1 and L2 Acquisition. Amsterdam/ Philadelphia: J. Benjamins, 15-55. —. 2006b. “Ǿ ĭȦȞȒ ıIJȘȞ ǼȜȜȘȞȚțȒ: ʌİȡȚȖȡĮijȒ IJȠȣ ıȣıIJȒȝĮIJȠȢ țĮȚ ȝİȜȑIJȘ IJȘȢ ĮȞȐʌIJȣȟȒȢ IJȠȣ ıIJȘȞ İȜȜȘȞȚțȒ ȦȢ ȝȘIJȡȚțȒ țĮȚ įİȪIJİȡȘ/ ȟȑȞȘ ȖȜȫııĮ”. In S. Moschonas (ed.), H ıȪȞIJĮȟȘ ıIJȘ ȝȐșȘıȘ țĮȚ ıIJȘ įȚįĮıțĮȜȓĮ IJȘȢ İȜȜȘȞȚțȒȢ ȦȢ ȟȑȞȘȢ ȖȜȫııĮȢ. Athens: Patakis, 62-116. Tsimpli, I.-M. & Roussou, A. 1996. “Negation and polarity items in Greek”. The Linguistic Review 13: 49-81. Tsiouni, V. 1972. ȆĮȚįȚȩijȡĮıIJȠȢ įȚȒȖȘıȚȢ IJȦȞ ȗȫȦȞ IJȦȞ IJİIJȡĮʌȩįȦȞ. München: Institut für Byzantinistik und Neugriechische Philologie der Universität. Tsolakis, E. T. 1959. “ȂȚȤĮȒȜ īȜȣțȐ ıIJȓȤȠȚ”. ǼʌȚıIJȘȝȠȞȚțȒ ǼʌİIJȘȡȓȢ ĭȚȜȠıȠijȚțȒȢ ȈȤȠȜȒȢ ǹȡȚıIJȠIJİȜİȓȠȣ ȆĮȞİʌȚıIJȘȝȓȠȣ ĬİııĮȜȠȞȓțȘȢ, ȆĮȡȐȡIJȘȝĮ 3, 3-22. Tsopanakis, A. G. 1994. ȃİȠİȜȜȘȞȚțȒ īȡĮȝȝĮIJȚțȒ. Thessaloniki: Kyriakidis. Tsunoda, T. 1985. “Remarks on transitivity”. Journal of Linguistics 21: 385-396. —. 1994. “Transitivity”. In R. E. Asher (ed.), The Encyclopedia of Language and Linguistics. Oxford: Pergamon Press, Vol. 9: 4670-77. Turner, N. 1963. A Grammar of NT Greek. Vol. III, Syntax. Edinburgh: T. & T. Clark. Tzartzanos, A. [1946, 1963] 1991. ȃİȠİȜȜȘȞȚțȒ ıȪȞIJĮȟȚȢ (IJȘȢ țȠȚȞȒȢ įȘȝȠIJȚțȒȢ). Thessaloniki: Kyriakidis. Tzitzilis, Ch. 2000. “ȃİȠİȜȜȘȞȚțȑȢ įȚȐȜİțIJȠȚ țĮȚ ȞİȠİȜȜȘȞȚțȒ įȚĮȜİțIJȠȜȠȖȓĮ”. Ǿ ǼȜȜȘȞȚțȒ īȜȫııĮ țĮȚ ȠȚ ǻȚȐȜİțIJȠȓ IJȘȢ. Athens: Ministry of National Education and Religious Affairs, Center for the Greek Language, 15-22. Tzitzilis, Ch. & Symeonidis, Ch. (eds.) 2000. ǺĮȜțĮȞȚțȒ īȜȦııȠȜȠȖȓĮ: ıȣȖȤȡȠȞȓĮ țĮȚ įȚĮȤȡȠȞȓĮ. Thessaloniki: University of Thessaloniki. Valetas, G. 1947. ̝ΑΌΓΏΓ·ϟ΅ȱΘϛΖȱΈ΋ΐΓΘ΍ΎϛΖȱΔΉΊΓ·Ε΅Κϟ΅Ζǯȱ̝ΔϲȱΘϲȱ̏΅Λ΅΍ΕκȱБΖȱΘϲȱ̕ΓΏΝΐϱ (13401827). Athens: Petros Ranos. Valin, R. D. Van & LaPolla, R. J. 1997. Syntax: Structure, Meaning and Function. Cambridge: Cambridge University Press. Valin, R. D. Van & Wilkins, D. P. 1996. “The case for 'Effector': case roles, agents, and agency revisited”. In M. Shibatani & S. A. Thompson (eds.), Grammatical Constructions. Oxford: Clarendon Press, 289-322. Varlokosta, S. 1994. Issues on Modern Greek Sentential Complementation. PhD Dissertation, University of Maryland. —. 2005. “Eventivity, modality and temporal reference in early child Greek”. In M. Stavrou & A. Terzi (eds.), Advances in Greek Generative Syntax: in Honor of Dimitra Theophanopoulou-Kontou. Amsterdam/ Philadelphia: J. Benjamins, 217-240. Vassilaki, S. 1986. La constitution des relations réfléchies et le passif: Etude des verbes en -ȝĮȚ du grec moderne. PhD Dissertation, Université Paris VII. —. 1988. “Ǿ ȝȠȡijȠȜȠȖȓĮ IJȘȢ ʌĮșȘIJȚțȒȢ ijȦȞȒȢ ıIJĮ ȃȑĮ ǼȜȜȘȞȚțȐ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 8: 185-204. —. 1999. “ȀȡȚIJȒȡȚĮ ȖȚĮ IJȘȞ ȠȡȚȠșȑIJȘıȘ IJȘȢ ȝȑıȘȢ ijȦȞȒȢ ıIJĮ ȃȑĮ ǼȜȜȘȞȚțȐ”. In A. Moser (ed.), Proceedings of the 3rd International Conference on Greek Linguistics. Athens: Ellinika Grammata, 52-60. Vendler, Z. 1957. “Verbs and times”. Philosophical Review 56: 143-160. —. 1967. “Verbs and times”. In Z. Vendler (ed.), Linguistics in Philosophy. Ithaca: Cornell University Press, 97-121.

254

Transitivity Alternations in Diachrony

Vennemann, T. 1974a. “Language type and word order”. Acta Universitatis Carolinae, Philologica 5, Linguistica Generalica I: 219-229. —. 1974b. “Theoretical word order studies: results and problems”. Papiere zur Linguistik 7: 5-25. —. 1974c. “Topics, subjects, and word order: from SXV to SVX via TVX”. In J. Anderson & C. Jones (eds.), Historical Linguistics: Proceedings of the First International Congress of Historical Linguistics, Edinburgh, September 1973. Vol. II. Amsterdam: North-Holland, 339-376. Verkuyl, H. J. 1989. “Aspectual classes and aspectual composition”. Linguistics and Philosophy 12: 3994. —. 1993. A Theory of Aspectuality. Cambridge: Cambridge University Press. Verrips, M. 1993. “Implicit agents in passives and anticausatives of Dutch children”. In F. Drijkoningen & K. Hengeveld (eds.), Linguistics in the Netherlands 1993. Amsterdam/ Philadelphia: J. Benjamins, 165-176. Vihman, V.-A. 2002. “Middle voice in Estonian”. Transactions of the Philological Society 100: 131160. Visser, F. 1963-1973. A Historical Syntax of the English Language. Leiden: E. J. Brill. Visvizis, I. 1951. “ȃĮȟȚĮțȐ ȞȠIJĮȡȚĮțȐ ȑȖȖȡĮijĮ IJȦȞ IJİȜİȣIJĮȓȦȞ ȤȡȩȞȦȞ IJȠȣ ǻȠȣțȐIJȠȣ IJȠȣ ǹȚȖĮȓȠȣ (1538-1577)”. ǼʌİIJȘȡȓȢ IJȠȣ ǹȡȤİȓȠȣ ǿıIJȠȡȓĮȢ ǼȜȜȘȞȚțȠȪ ǻȚțĮȓȠȣ 4: 1-167. Vitti, M. 1960. “Il poema parenetico di Sachlikis”. ȀȡȘIJȚțȐ ȋȡȠȞȚțȐ 14: 173-200. —. 1971/1978. Storia della letteratura neogreca. Torino. / ǿıIJȠȡȓĮ IJȘȢ ȞİȠİȜȜȘȞȚțȒȢ ȜȠȖȠIJİȤȞȓĮȢ (ȝİIJȐijȡĮıȘ: Ȃ. ǽȠȡȝʌȐ). Athens: Odysseas. Volpe, M. 2005a. Japanese Morphology and its Theoretical Consequences: Derivational Morphology in Distributed Morphology. PhD Dissertation, SUNY Stony Brook. —. 2005b. “The morpho-syntax of expressive derivations in Japanese: subject honorification and object honorification in Distributed Morphology”. http://ling.auf.net/lingBuzz/000087. Vouaux, L. 1922. Les actes de Pierre. Paris: Letouzey & Ané. Wackernagel, J. 1904/1953. “Studien zum griechischen Perfectum”. Programm zur akademischen Preisverteilung, 3-32 / Published also in Kleine Schriften. Göttingen: Vandenhoek and Ruprecht, 1000-1021. —. 1924. Vorlesungen über Syntax II. Basel: E. Birkhauser. —. 1926-1928: Vorlesungen über Syntax. 2nd edition. Basel: Birkhäuser. [1926. I. Erste Reihe. 1928. II. Zweite Reihe.] —. 1950. Vorlesungen über Syntax mit besonderer Berücksichtigung von Griechisch, Lateinisch und Deutsch. Basel: Birkhäuser. Wagner, W. 1870/1970. Medieval Greek Texts: being a Collection of the Earliest Compositions in Vulgar Greek, prior to the year 1500. Amsterdam: Gruner. —. 1874. Carmina Graeca Medii Aevi. Leipzig: Teubner. Wasow, T. 1977. “Transformations and the lexicon”. In P. W. Culicover, T. Wasow & A. Akmajian (eds.), Formal Syntax. New York: Academic Press, 327-360. Wassiliewsky, B. & Jernestedt, V. 1965. Cecaumeni Strategikon. Amsterdam: Hakkert. Watanabe, A. 1993. Agr-Based Case Theory and its Interaction with the A-Bar System. PhD Dissertation, MIT. Wehr, B. 1995. SE-Diathese im Italienischen. Tübingen: Narr. Weierholt, K. 1963. Studien im Sprachgebrauch des Malalas. Symbolae Osloenses suppl. 18. Oslo. Wierzbicka, A. 1968. “On the semantics of the verbal aspect in Polish”. To Honor Roman Jakobson. Essays on the occasion of his seventieth birthday. The Hague: Mouton, 2231-2249. Williams, E. 1980. “Predication”. Linguistic Inquiry 11(1): 203-238. —. 1981. “Argument structure and morphology”. The Linguistic Review 1(1): 81-114. —. 1983. “Against small clauses”. Linguistic Inquiry 14(2): 287-308. —. 1984. “Grammatical relations”. Linguistic Inquiry 15(4): 639-673. —. 1987. “English as an ergative language: the theta structure of derived nouns”. Chicago Linguistic Society 23: 366-375. Williams, R. S. 1994. “A statistical analysis of English double object alternation”. Issues in Applied Linguistics 5: 37-58. Winters, M. E. 1984. “Steps toward the Romance passive inferrable from the Itinerarium Egeriae”. Romance Philology 37(4): 445-454. Wistrand, E. 194l. Über das Passivum. Göteborg. Wolf, K. 1911-1912. Studien zur Sprache des Malalas, I. Formenlehre, II. Syntax. München: Progr. Wonder, J. P. 1990. “Noun incorporation in Spanish”. Hispanic Linguistics 4(1): 149-177. Wright, S. 2002. “Transitivity and change-of-state verbs”. Proceedings of the 28th Annual Meeting of the Berkeley Linguistics Society. Berkeley: Berkeley Linguistics Society, 339-350.

Bibliography

255

Wright, S. & Levin, B. 2000. “Unspecified object contexts with activity and change-of-state verbs”. Paper presented at the 74th Annual LSA Meeting, Chicago, IL, January 6-10, 2000. Young, R. W. & Morgan, W. 1980. The Navajo Language. A Grammar and Colloquial Dictionary. Albuquerque: University of New Mexico Press. Zagona, K. 1998. Verb Phrase Syntax: a Parametric Study of English and Spanish. Dordrecht: Kluwer. Zaliznjak, A. A. 1977. Grammatiþeskij slovar’ russkogo jazyka. Moscow. Zambelios, S. 1986 [1852]. DZıȝĮIJĮ įȘȝȠIJȚțȐ IJȘȢ ǼȜȜȐįȠȢ, İțįȠșȑȞIJĮ ȝİIJȐ ȝİȜȑIJȘȢ ȚıIJȠȡȚțȒȢ ʌİȡȓ ȝİıĮȚȦȞȚțȠȪ ǼȜȜȘȞȚıȝȠȪ. Athens: D. N. Karavias. Zombolou, K. 1996. “ȉĮ ĮʌȠșİIJȚțȐ ȡȒȝĮIJĮ IJȘȢ ȞȑĮȢ İȜȜȘȞȚțȒȢ: ıȤȘȝĮIJȚıȝȩȢ ʌĮșȘIJȚțȒȢ ijȦȞȒȢ țĮȚ įȚȐșİıȘȢ, șİȝĮIJȚțȠȓ ȡȩȜȠȚ țĮȚ ȖȞȦıIJȚțȒ ıȘȝĮıȚȠȜȠȖȚțȒ ĮȞȐȜȣıȘ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 16: 160-171. —. 1997. “ǹȞĮȗȘIJȫȞIJĮȢ IJȠ įȡȐıIJȘ: ĮʌȠșİIJȚțȐ ȡȒȝĮIJĮ ȑȞĮȞIJȚ ȝİıȠʌĮșȘIJȚțȒȢ įȚȐșİıȘȢ. ȂȚĮ ȐȜȜȘ ʌȡȠıȑȖȖȚıȘ IJȠȣ İȜȜȘȞȚțȠȪ ȡȘȝĮIJȚțȠȪ ıȣıIJȒȝĮIJȠȢ”. ȂİȜȑIJİȢ ȖȚĮ IJȘȞ ǼȜȜȘȞȚțȒ īȜȫııĮ/Studies in Greek Linguistics 17: 229-242. —. 2004. Verbal Alternations in Greek: a Semantic Analysis. PhD Dissertation, University of Reading. Zsilka, J. 1966. “Probleme des Aorists bei Homer”. Acta Antiqua Academiae Scientiarum Hungaricae 14: 33-59. Zubizarreta, M. L. 1987. Levels of Representation in the Lexicon and in the Syntax. Dordrecht: Foris.

INDEX accomplishment, 5, 38-41, 65, 73 accusative case, 3, 5, 19, 21, 22, 27, 28, 30, 32, 33, 36, 48, 58, 59-61, 64, 65, 73, 76, 77, 84, 87, 88, 93-95, 99, 101-105, 107, 108, 114, 118-121, 128-132, 136, 137, 158-161, 166-168, 171, 176, 180, 181, 183-194, 196, 197, 199, 200, 204, 214, 218-220 accusative verb, 7, 19, 21, 31, 48, 76, 77, 93, 94, 101, 103, 118, 119, 129, 158, 159, 161, 162, 167, 176, 196 achievement, 38-41, 57 active, 20-22, 44, 82, 86, 89, 90, 93, 94, 95, 106-108, 110, 112-114, 116-119, 130, 131, 133, 134, 139, 142, 144, 147, 149, 151, 154, 156, 161, 179, 182, 185-187, 190, 191, 193, 194, 196, 198-200, 202, 208, 214, 215, 218, 219 activity, 11, 38-41, 102, 158, 164, 166 Adams, D. Q., 15, 188 adjective, 64, 99, 106, 127, 169, 173, 174, 177, 205 adjunct, 5 adverb, 135 affectedness, 5-7, 38, 39, 84, 89, 117, 125, 165 agent, 3-7, 9-11, 13, 17-19, 21, 22, 29, 32, 3441, 43-45, 48, 49, 57, 59, 66, 67, 76, 78, 79, 82, 85-87, 89, 92-94, 106, 107, 109, 112119, 122, 126, 132, 142, 145, 152, 161, 162, 164-166, 171-173, 175-177, 183, 184, 186, 187, 191-201, 204-208, 210, 212-214, 217-220 agreement, 25, 29, 48, 79, 80, 184, 218 aktionsart, 3, 7, 39, 41, 56, 64, 73, 118, 135, 208, 211 Alexiadou, A., 2, 3, 5, 6, 12, 20, 31, 32, 34-36, 41, 43, 45, 48, 90, 137, 182, 195, 198, 202, 203, 206, 213 ambiguity, 21, 50, 92, 172, 174, 177, 179, 180, 184, 218, 219 Anagnostopoulou, E., 2, 5, 6, 20, 31, 32, 3436, 43, 45, 48, 90, 137, 184, 185, 188, 195, 198, 203, 204, 206, 213 analogy, 45, 57, 94, 168 Andersen, P. K, 80, 81 animacy, 4, 20, 45, 59, 76, 86, 87, 88, 159, 164, 188 anticausative, 6, 9-14, 17-23, 31, 32, 34, 37, 39, 42-49, 59, 63, 65-68, 70, 76, 78-82, 84, 85, 88-90, 92-94, 96, 97, 100, 103, 106, 112-119, 121, 122, 126, 132-136, 138-144, 146-149, 151, 152, 154, 156, 158, 161, 162, 166-179, 182, 184-187, 191, 195-214, 217219, 221 anticausativisation, 23, 43, 63, 93, 134, 149, 168, 176, 200, 205, 206, 220 Antonopoulou, E., 6, 18, 39 Arabic, 204

Arad, M., 31, 36, 48, 182 argument structure, 3, 20-22, 25, 26, 29, 36-39, 44, 63, 93, 118, 132, 161-164, 168, 176, 179, 180, 199, 209, 212-214, 218, 219 Armenian, 198 aspect, 6, 56, 64, 73, 79, 163, 168, 171, 217 atelicity/atelic, 4, 5 auxiliary, 5, 54, 60, 165, 166, 173-176, 190, 192 Babiniotis, G., 16, 18, 42, 94, 204 Baker, M., 25, 27, 39, 40, 188, 190, 192, 203 Bakker, E. J., 82, Basea-Bezantakou, C., 159 benefactive, 22, 58, 59, 61, 71, 82, 83, 87, 89, 93, 94, 108, 110, 116, 119, 130, 134, 161, 176, 180, 190, 194, 200, 203, 218 blocking, 27, 36, 161, 172, 202, 204, 218, 219 Borer, H., 29 Brugmann, K., 79, 80, 112 Burzio’s generalisation, 5, 9 Campos, H., 76 Case, 25-28, 46, 50 Catalan, 172 causative, 1, 3, 7, 9, 10, 12-14, 17-22, 31, 3639, 41-44, 46, 48, 49, 57, 63, 65, 67, 68, 70-76, 78-80, 82, 88, 89, 92-99, 101-104, 112, 113, 118, 119, 121, 123-129, 132, 134149, 151-159, 161-163, 165-171, 176, 179, 181, 182, 184-187, 190, 195, 196, 198-201, 203-214, 217-219, 221 causativisation, 3, 21-23, 43, 56, 62, 63, 65, 67, 68, 71, 76, 92-94, 97, 100, 101, 113, 118, 122, 123, 126-128, 132, 134, 137, 146, 147, 149, 152, 154-158, 161-164, 166-169, 176, 179, 180, 182, 186, 187, 196, 199, 200, 212, 213, 218, 220, 221 cause, 3, 6, 9-11, 13, 14, 17-21, 37-45, 48-51, 53, 54, 57, 58, 61, 66, 67, 69-71, 74-76, 78, 85, 86, 89, 92-95, 98, 102, 105-107, 113, 114, 116-119, 123, 126, 128, 132, 136, 137, 152, 161, 163-167, 169, 171, 172, 175, 176, 182-184, 186, 187, 195-200, 203, 204, 206208, 213, 219, 221 Cennamo, M., 17, 39, 171, 172 change-of-state, 1, 2, 6-10, 14, 17-21, 31, 3644, 48, 49, 64, 71, 76, 81, 92, 98, 100, 112, 118, 123, 125, 128, 135, 140, 152, 155, 158, 159, 161, 163, 166, 169, 171, 173, 182, 183, 186, 187, 190, 195, 196, 198, 200, 201, 203,-206, 208, 211, 213, 215, 217, 221 Charalambakis, Ch., 64, 77, 94 Chatzidakis, G. N., 76, 79, 94, 112, 121, 122, 126, 146, 148, 152, 153, 155, 156, 185 Chierchia, G., 3, 40, 43, 44, 49, 206, 209, 210

258

Transitivity Alternations in Diachrony

Chila-Markopoulou, D., 11, 16, 41, 42, 51 Chomsky, N., 3, 20, 25- 27, 29-31, 36, 44-46, 50, 191, 192, 219 Christidis, A.-Ph., 16, 94 clause structure, 21, 26, 27 clitic, 22, 23, 28, 46, 55, 95, 118, 122, 126, 130, 134, 136, 155, 161, 177, 182, 189, 190, 198, 200 Collins, C. T., 29, 31-34, 36, 48, 191, 192, 194, 199, 203 complement/complementation, 23, 28, 29, 41, 55, 78, 94, 101, 104, 114, 119, 125, 160, 191, 192, 248 complexity, 50, 51, 58, 220 compounding, 56, 57 conceptual structure, 26 conservatism (of the learner), 169, 186 cues, 22, 50, 51, 53-55, 61, 183, 190, 214, 217, 219, 220 dative alternation, 36, 188 dative case, 13, 33, 64, 66, 70, 71, 76, 77, 85, 88, 94, 105-108, 129, 131, 161, 166, 167, 183, 186-188, 190, 192, 219 Davidse, Ȁ., 7, 163-165 Davis, H., 206, 210 decomposition, event, 20, 25, 38, 39, 43, 48, 182, 200, 204, 212, 213 degree-0, 55 detransitivisation, 43 diathesis, 58, 59, 82, 180 directionality, 163, 204, 211, 219 Distributed Morphology, 26, 41 ditransitive, 3, 30, 39, 86-88, 93, 106-108, 119, 188, 221 domain, 13, 15, 38, 39, 94, 185, 220 Drachman, G., 3, 119 drift, 158 Dutch, 55 economy, 22, 46, 66, 192, 197, 199, 200, 214, 218, 220 E-language, 50-54, 61, 217, 219, 220, 221 Embick, D., 5, 20, 26, 31, 36, 45, 46, 48, 191, 200-203, 214 English Early Middle, 169 Early Modern, 55, 166, 167, 170, 182 Middle, 54, 55, 132, 167, 169, 170, 172 Modern, 32, 36-39, 43, 60, 163-167, 169, 176, 182, 188, 198, 206 Old, 51, 54, 55, 60, 132, 162, 163, 165-170, 176, 182 ergative case, 7, 17, 18, 68, 163-165 Event, 40 experiencer, 8, 20, 40, 60, 77, 166, 167, 182, 183, 186, 188 Extended Projection Principle (EPP), 30 external argument, 3, 5, 11, 12, 26, 28-33, 36, 43, 45, 46, 48, 57, 94, 118, 152, 154, 155,

157, 161, 184, 191-194, 198, 199, 202-204, 207, 208, 210, 214, 219 Eythórsson, T., 51, 169, 183, 186 Faarlund, J. T., 14 feature, 26, 28, 30-32, 36-38, 43, 45, 48, 51, 53, 61, 108, 183, 191, 195, 196, 200-203, 214 ij-features, 30, 31 French Modern, 46, 175, 189, 190 Old, 189 Fykias, I., 76 Generative Grammar, 2, 14, 15, 18, 21, 26, 29, 36, 44, 50, 51, 61, 63, 152, 162, 163, 217, 218 genitive case, 3, 12, 26, 33, 54, 55, 59, 60, 76, 78, 86, 87, 93-95, 104, 105, 107, 110, 119, 129, 131, 132, 161, 167, 176, 187, 188, 190, 219 George, C. H., 85, 86, 106 German, 31, 43, 44, 48, 51, 55, 170, 183, 198, 206, 207 Germanic/Proto-Germanic, 68, 168, 170 gerund, 185 Giannakis, G. K., 58, 103, 176 goal, 7, 19, 22, 26, 40, 49, 52, 61, 76, 77, 84, 87, 93, 106, 131, 161, 187, 188, 213, 217 Government and Binding Theory, 5, 26, 27, 29 grammaticalisation, 172, 173, 176 Greek classical Ancient Greek, 25, 35, 37, 89, 93, 94, 97, 99-101, 105, 118, 120, 128, 208, 212, 219, 220 Early Medieval, 125, 126, 132, 133, 136146, 149-152, 155-157, 187, 191, 214 Hellenistic-Roman, 1, 15, 16, 21, 22, 84, 93, 95, 98, 99, 104, 106, 107, 112-114, 116, 118-120, 122-124, 130, 134, 143, 149, 157, 161, 179, 181, 186-189, 194, 196, 198-201, 209, 210, 214 Homeric, 15, 21, 56, 59, 61, 63-68, 70-72, 76, 79, 85, 92, 93, 209 Late Medieval, 120, 125, 126, 132, 133, 136-144, 146, 150, 156 Medieval, 15, 16, 22, 47, 120, 123, 124, 126, 127, 129-136, 138, 141-144, 146149, 152, 154-158, 161, 181, 184-191, 194, 197-203, 208, 214, 215, 218, 219 Modern, 1, 2, 5-10, 12, 14-17, 19-21, 2629, 32-37, 41-49, 51, 56-61, 63, 67, 71, 84, 85, 90, 91, 94, 106, 117, 120-126, 136, 137, 142, 143, 146, 148, 152, 154159, 161, 184-186, 191, 193, 196, 198, 200-204, 206-208, 210, 211, 214 Grimshaw, J., 8, 26, 41, 101 Haegeman, L., 25, 26 Hale, K., 3, 25, 26, 36, 38, 39, 73, 218

Index

Harley, H., 31, 36, 39, 41, 48, 182 Haspelmath, M., 13, 17, 18, 42, 68, 172, 185, 204 Hopper, P., 4, 5, 9, 20, 86 Horrocks, G., 16, 17, 94, 110, 120, 182 Humbert, J., 16, 59, 73, 76, 80, 101, 105, 115 Hupa, 203 Icelandic, 169, 170, 183, 186 I-language, 21, 50-53, 61, 183, 217, 219, 221 imperative, 163 impersonal, 5, 31, 60, 167, 170, 172, 183 inchoative, 37, 68, 168 incorporation, 5, 28, 65, 190 infinitive, 60, 95, 114, 118, 119, 121, 184 input, 43, 50, 51, 54, 55, 61, 78, 183, 219, 220 instrument, 3, 9, 18, 19, 43, 48, 49, 78, 87, 90, 93, 195-197, 204, 206, 207, 213 internal argument, 26, 28, 30, 32, 46, 191, 199, 208 intransitive, 1-5, 9, 10, 13-15, 17-22, 25, 29, 35, 37-44, 46-49, 56-61, 63, 65-73, 75, 76, 79-81, 84-86, 89-103, 106, 112-115, 117119, 121-128, 130-132, 134-136, 138-152, 154-158, 161-172, 175-177, 179, 180, 182187, 195, 199, 200, 203-206, 208-213, 215, 217, 218, 219, 221 Italian Modern, 28, 172, 177 Old, 171-173, 177 Jackendoff, R., 25, 26, 38, 40 Jannaris, A., 16, 59, 76, 101, 105, 115, 180 Japanese, 14, 182, 198, 199 Joseph, B., 20, 64, 79, 121, 143 Kakouriotis, T., 5, 8 Karali, M., 64, 94, 184 Karantzola, E., 120, 121 Kavoukopoulos, F., 59, 61 Kemmer, S., 17, 46, 171, 172 Keyser, S. J., 3, 25, 26, 36, 38, 39, 73 Kiparsky, P., 1, 15 Korean, 14, 51, 198 Kratzer, A., 29, 32, 39, 45 Kroch, A., 51, 52 Kulikov, L. I., 13, 68, 167, 168, 176 language acquisition, 21, 22, 50, 52-54, 61, 165, 183, 199, 205, 211, 212, 214, 217, 219, 220 Latin Archaic, 170 classical, 22, 64, 121, 170, 171, 174, 176, 177, 191, 200-203, 214 Medieval, 171 Lavidas, N., 5, 6, 71, 84, 98, 110, 121, 125, 127, 163, 167, 172, 185, 217 learnability, 55 Levin, B., 2- 4, 7, 9, 10, 17, 20, 25, 26, 36- 41,

259

43, 44, 48, 49, 68, 101, 117, 128, 129, 164, 165, 182, 183, 188, 200, 205-207, 209, 213 lexical binding, 44, 72, 213 lexicon, 1-3, 20, 25, 26, 29, 36, 38, 43, 44, 48, 51, 60, 63, 64, 76, 179, 182, 185, 192, 203, 205, 208, 210, 212, 214, 217, 218 Lightfoot, D. W., 2, 14, 21, 22, 50-55, 58, 59, 61, 66, 67, 116, 179, 180, 183, 188, 211, 214, 219 linking rules, 3 little/light v, 6, 8, 12, 26, 28-26, 39, 43, 45, 48, 60, 80, 81, 98, 152, 163, 168, 170, 182, 191, 193-199, 201, 202, 204, 213, 214, 220 location, 40, 41, 60, 105, 106 locative, 37, 59, 106, 221 mapping, 25, 26, 38-40, 42, 44, 48, 49, 212, 217 markedness, 5, 6 Merge, 202 merger, 113, 119, 168-170, 176, 177, 220 Michaelis, S., 171-173 middle, 9, 17-20, 22, 45, 46, 56, 58, 59, 64, 66, 67, 79-82, 92, 94, 107, 110, 112, 117, 119, 172, 176, 179, 180, 183, 185, 203, 221 voice, 22, 58, 59, 64, 79, 82, 176 Minimalist Programme, 25-27, 29, 31, 32, 48, 191, 192 morphology, 3, 5, 10, 11, 13, 14, 17, 19-22, 25, 28, 32-34, 42, 44-46, 48, 49, 51, 53, 54, 59, 61, 64, 66, 70, 71, 76-80, 82, 84-86, 89, 9194, 96, 97, 99, 101, 108-110, 112-119, 127, 130-133, 139, 141-144, 147, 149, 151, 152, 158, 161, 162, 169-171, 174-176, 179, 182, 184-188, 191, 195, 198, 199, 201-206, 211, 212, 214, 215, 218, 220, 221 Montrul, S., 212, 213 Moser, A., 14, 16, 60, 64, 72, 94, 151, 174, 184, 185 motion (verbs of), 39, 57, 59, 69, 152, 158160, 163, 164, 173, 174, 176 movement, 5, 8, 26, 32-34, 38, 40, 41, 51, 61, 66, 67, 77, 79, 94, 103, 105, 112, 159, 163, 164, 167, 182, 192, 199, 202, 213 Nikiforidou, K., 76 nominative case, 5, 17, 22, 26, 27, 31-33, 5860, 64, 87, 88, 93, 94, 107, 116, 118, 121, 161, 166-168, 171, 180, 181, 183-186, 195, 213, 214, 218, 219 non-active, 21, 22, 27, 32, 33, 61, 71, 81, 82, 84, 89, 93, 106-110, 112, 113, 116, 117, 119, 121, 129, 130, 132-134, 142, 149, 151, 152, 161, 185, 187, 188, 190, 191, 194-201, 203, 208, 211, 214, 218, 219 object, 1, 3-7, 9, 17-19, 22, 28-32, 36, 41, 42, 44, 48, 55, 56, 58-61, 64-68, 71, 76-78, 82, 84, 87, 93, 102, 103, 107, 108, 114, 115, 119, 121, 122, 125-127, 129, 131, 142, 146,

260

Transitivity Alternations in Diachrony

152, 158, 159, 161, 163, 165-167, 172, 176, 179, 184, 185, 187-190, 214, 221 cognate, 60, 71, 72, 93, 96, 99, 118, 122, 127, 168 direct, 1, 4, 7, 19, 21, 27-29, 36, 41, 42, 48, 56, 60, 64-66, 72, 76-78, 84, 94, 99, 101105, 110, 114, 118, 119, 124, 129, 131, 136, 142, 155, 158, 159, 161, 166, 167, 176, 182, 187, 189, 190, 194, 214, 219 indirect, 64, 106, 129, 166, 187 oblique, 8, 26, 105, 183 opacity, 50, 58, 114, 180 Panayotou, A., 64 Pantelides, N., 79, 120 Papanastasiou, G. C., 72, 73, 79, 94 Papangeli, D., 26, 28, 84, 110, 167, 208 parameter, 6, 53, 55, 168, 217 participle, 9, 60, 78, 135, 165, 171, 173, 174, 176, 184, 185, 192 partitive, 78 passive, 3, 5, 6, 9-13, 19-22, 29, 31-34, 36, 4346, 48, 49, 56, 58, 59, 64, 66, 78-81, 84-86, 88, 89, 92, 93, 106-109, 112, 113, 115-117, 119, 126, 127, 132, 134, 142, 144, 145, 161, 162, 165, 166, 168, 170-177, 179, 180, 184, 185, 187, 188, 191, 192, 194, 195, 197-201, 203-207, 212, 214, 219 passivisation, 29, 32, 87, 93, 110, 116, 119, 145, 184, 187, 188, 200, 202, 208 patient, 3, 5-7, 9, 10, 13, 14, 19, 20, 36-38, 40, 43, 48, 58-60, 67, 76, 79, 85, 87, 89, 93, 95, 102, 113, 117, 119, 126, 128, 129, 137, 152, 159, 163-165, 184, 185, 196, 201, 212-214, 219, 220 periphrasis, 173-177, 182, 207 Perlmutter, D., 9, 25, 29 Pesetsky, D., 3, 8, 20, 36, 41, 43, 200, 205 phase, 22, 36, 56, 161, 182, 221 Philippaki-Warburton, I., 20 phonology, 220 Pinker, S., 25, 26, 39, 164 Portuguese, 172, 175, 177 Postal, P., 9, 25, 29 predicate, 3, 10, 25, 29, 35, 38-40, 45, 48, 192, 204, 205, 208, 212 prefix, 14, 72, 73, 93, 99, 100, 106, 125-127, 132, 135, 138-141, 144, 151, 152, 155, 169, 177 preposition/preposition phrase, 28, 33, 55, 60, 66, 73, 86, 106, 116, 132, 159, 167, 191, 192, 194 preterite-present, 54 primary linguistic data (PLD), 53, 54, 56, 61, 219 Principles and Parameters, 25 pro-drop, 192, 213, 219 pronoun, 83, 84, 99, 110, 111, 170-177, 190, 191 Proto-Indo-European, 58, 68, 79, 176, 180

psych-verb, 8, 47, 66, 70, 76, 80, 99, 119, 132, 133, 137, 152, 155, 166, 167, 182, 201 raising, 25, 27, 84, 190 Ramchand, G., 183 Rappaport Hovav, M., 2-4, 10, 17, 20, 25, 26, 36-41, 43, 44, 48, 49, 101, 117, 164, 165, 182, 183, 188, 205-207, 209, 213 reanalysis, 21, 22, 50, 52, 53, 55, 56, 110, 118, 119, 121, 152, 161, 168, 173, 176, 177, 179, 180, 183-187, 213, 218-221 reconstruction, 170 reduction, 51, 60, 206 reflexive, 5, 19, 45, 46, 56, 59, 67, 79, 80, 81, 83, 84, 110, 111, 116, 117, 134, 144, 167, 170, 172-174, 176, 177, 192, 204, 208 Reinhart, T., 3, 26, 44, 182, 205-208 resultative, 5-7, 182, 185 Roberts, I., 14, 27, 50, 51, 192, 197 Roeper, T., 212 Romance, 22, 103, 104, 118, 157, 162, 170, 172, 176, 177 root, 41, 42, 48, 64, 80, 127, 134, 168, 177, 182, 190, 198-203, 214, 221 Rothemberg, M., 46 Roussou, A., 14, 32, 50, 51, 182, 197, 213 Russian, 13, 17, 44, 57, 175 Salish, 203, 206 Sansò, ǹ., 172 selectional restrictions, 8, 155, 205, 207 semantics, 6, 42 Sioupi, A., 5 source, 16, 40, 54, 106, 204, 217 Spanish Modern, 172 Old, 132 spell-out, 182 spread, 21, 23, 50, 51, 62, 64, 117, 119, 130, 149, 162, 164, 182, 197, 208, 220, 221 stability, 220 Stavrou, M., 182, 204 subcategorisation, 26, 34, 164, 191, 194 subevent, 40, 41, 49, 196, 208, 212 subject, 2, 3, 5, 8, 14, 17, 19, 20, 26-34, 40, 41, 43, 45, 48, 56-61, 63-67, 76, 79, 80, 82-89, 91, 93-95, 98, 107-110, 112-114, 117-120, 123, 126, 130, 131, 142, 151, 152, 161, 164, 165, 167, 169, 170-172, 179-181, 183186, 188, 189, 191, 192, 195, 199, 201, 204, 207, 213, 214, 219-221 subjunctive, 54 suffix, 13, 28, 34, 55, 60, 73, 75, 80, 81, 93, 170, 171, 183-185, 187, 192-194, 196, 198, 206 syncretism, 45, 46, 168, 169, 177 telicity/telic, 4, 208 template, 39-42, 179, 212, 213, 217, 218 tense, 5, 13, 54, 56, 64, 79, 80, 106, 111, 185,

Index

220 thematic (ș-) role, 3, 9, 21, 25, 26, 28, 29, 32, 33, 36, 37, 39, 40, 45, 48, 49, 54, 55, 60, 66, 67, 76, 125, 152, 183, 191, 196, 204, 206, 217 theme, 8, 19, 20, 22, 40, 60, 76, 77, 87, 93, 108, 125, 131, 137, 161, 167, 188 Theophanopoulou-Kontou, D., 2, 5, 7, 18, 20, 25-27, 31, 34, 35, 41, 44-46, 50, 52, 57, 59, 61, 76, 117, 152, 154, 156, 185, 186, 199, 202-204 Thompson, S., 4, 5, 9, 20, 86 transition, 58, 60, 173, 175, 180, 218 transitive, 1-7, 9-15, 17-22, 28-33, 37-44, 48, 49, 56-59, 61, 63, 66-73, 75, 76, 78-84, 86, 87, 89-97, 99, 101-104, 106-110, 112-119, 121-128, 130-132, 134, 136, 138-150, 152, 157-159, 161-171, 173, 176, 177, 179, 180, 184, 187, 188, 191, 194, 195, 198-206, 208-210, 212-214, 217-219 transitivisation, 2, 3, 6, 56, 57, 60, 73, 99, 134, 165, 175, 204 Transparency Principle, 50, 53, 59, 61, 180, 220 trigger, 219 Tsimpli, I.-M., 2, 8, 20, 28, 32, 45, 46, 67, 182, 183, 192, 211, 213 Turkish, 13, 44, 199 Tzartzanos, A., 7, 44, 184 Tzitzilis, Ch., 120, 185

261

underspecification, 46 unergative, 2, 5, 6, 13, 18, 21, 25, 30, 38-40, 44, 49, 57, 112, 152, 164, 171, 184, 185, 206, 210, 212 unidirectionality, 172 Universal Grammar (UG), 1, 53-55, 61, 180, 183, 220 Varlokosta, S., 41, 53 Vassilaki, S., 46, 143, 211 Vedic, 167, 176 Visser, F., 68, 162, 163, 165, 167, 169, 170 voice, 2, 7, 8, 18-22, 28, 29, 33-37, 45, 46, 48, 49, 56, 58, 59, 61, 63, 66-68, 71, 78-82, 85, 88-90, 93-95, 101, 104, 106-116, 118, 119, 129-135, 139, 141, 142, 144, 147, 149, 151, 152, 154, 156, 161-163, 168, 171, 174, 176, 177, 179, 185-187, 189, 191, 192, 195-201, 203-206, 211, 213, 214, 218, 219, 221 morphology, 2, 18-22, 33-35, 45, 46, 48, 49, 56, 58, 59, 61, 63, 71, 78, 79, 88, 89, 93, 94, 111, 112, 114-116, 118, 119, 130, 134, 139, 141, 142, 144, 147, 149, 152, 154, 162, 179, 185-187, 189, 192, 195, 197-200, 204-206, 211, 213, 214, 218, 219, 221 Williams, E., 3, 26, 205 Wright, S., 182 Zombolou, K., 26, 44, 138, 158