Prepositional clauses in Spanish: A diachronic and comparative syntactic study 9781614510611

Table of contents :
Acknowledgments
List of tables
Abbreviations
1 Introduction
1 Topic of study
2 Sources of data
3 Data in historical studies
4 Organization of the book
2 Categories, syntax, and change
1 Introduction
2 Categories in prepositional clauses
2.1 The functional category Complementizer – COMP
2.2 The category Preposition
2.2.1 Prepositions: types and features
2.2.2 Licensing prepositional objects with functional prepositions
2.3 Prepositions and clauses
2.3.1 Prepositions and that-clauses
2.3.2 The hypothesis of the underlying preposition
2.3.3 Prepositional indirect interrogative clauses
2.3.4 Subordinating conjunctions as prepositional phrases
2.3.5 Prepositional complementizers
3 Syntactic framework
3.1 Case Theory
3.1.1 Case
3.1.2 Types of Case
3.1.3 Prepositions and Case
3.1.4 Critiques against abstract Case
3.1.5 Case and finite clauses
3.2 Determiners and clauses
3.3 Arguments vs. adjuncts
3.4 Historical syntax. Change and mechanisms
3.4.1 The nature of change
3.4.2 Mechanisms of change
3 Historical Spanish
1 Introduction
2 Prepositions and clauses in historical Spanish
2.1 Prepositions and nouns
2.1.1 Argumental and adjunct phrases
2.1.2 Variation in argumental prepositional selection
2.1.3 Conclusions
2.2 Prepositions and infinitives
2.2.1 Argumental and adjunct prepositional infinitives
2.2.2 Variation with prepositional infinitives
2.2.3 Summary and syntactic conclusions
2.2.4 The nominality of the infinitive
2.3 Prepositions and finite clauses
2.3.1 Previous studies based on syntactic analogy
2.3.2 Previous syntactic approaches: Pronouns, nominality and Case
2.3.3 Summary and conclusions of previous syntactic studies
2.3.4 Argumental prepositional que-clauses
2.3.5 Early examples of argumental que-clauses
2.3.6 Adverbial prepositional que-clauses (adjunct que-clauses)
2.3.7 Prepositional indirect interrogative finite clauses
2.3.8 Prepositional que-clauses and pronouns
2.3.9 Prepositional que-clauses and variation
2.3.10 Finite clauses and the article
3 Discussion and conclusions
4 Present-day Spanish
1 Introduction
2 Prepositions and (non-clausal) nouns
3 Prepositions and infinitives
4 Prepositional finite clauses
4.1 Prepositional que-clauses
4.2 Prepositional indirect interrogative finite clauses
4.3 Variation and optional prepositions
4.4 Case and the licensing of the finite clause
4.5 Clausal argumenthood and prepositionality
4.6 Finite clauses and the article
5 Discussion and conclusions
5 Crosslinguistic perspective: Romance languages I
1 Introduction
2 Latin and Proto-Romance
2.1 The Latin and Proto-Romance prepositional finite clause
2.2 The Latin and Proto-Romance prepositional infinitive
2.3 Conclusion
3 Portuguese
3.1 Historical Portuguese
3.1.1 Prepositions and nouns. Prepositions and infinitives
3.1.2 Prepositions and finite clauses
3.2 Present-day Portuguese
3.2.1 Prepositions and nouns. Prepositions and infinitives
3.2.2 Prepositional finite clauses
3.2.3 Nominality and the clause
3.3 Discussion and conclusions
6 Crosslinguistic perspective: Romance languages II
1 Introduction
2 French
2.1 Historical French
2.1.1 Prepositions and infinitives
2.1.2 Prepositional finite que-clauses. The ce element
2.2 Present-day French
2.2.1 Prepositions and infinitives
2.2.2 Prepositional finite clauses
2.3 Discussion and conclusions
3 Italian
3.1 Historical Italian
3.1.1 Prepositions and infinitives
3.1.2 Prepositional finite clauses
3.2 Present-day Italian
3.2.1 Prepositions and infinitives
3.2.2 Prepositional finite clauses
3.3 Discussion and conclusions
7 Conclusions
1 A crosslinguistic overview of prepositional finite clauses
2 Change in Spanish from a comparative perspective. Data and theoretical implications
2.1 The nominality of the clause and (re-)nominalization/substantivization
2.2 Case, pronominal alternatives, and the finite clause
2.3 E-language and syntactic change
References
Subject index

Citation preview

Manuel Delicado Cantero Prepositional Clauses in Spanish

Studies in Language Change

Edited by Cynthia Allen Harold Koch Malcolm Ross

Volume 12

Manuel Delicado Cantero

Prepositional Clauses in Spanish A Diachronic and Comparative Syntactic Study

ISBN 978-1-61451-061-1 e-ISBN 978-1-61451-062-8 ISSN 2163-0992 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. 6 2013 Walter de Gruyter, Inc., Boston/Berlin Typesetting: RoyalStandard, Hong Kong Printing and binding: Hubert & Co. GmbH & Co. KG, Göttingen ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

Acknowledgments First and foremost, I would like to thank the editors of this series, Cynthia Allen, Harold Koch and Malcolm Ross, for making it happen. I would like to acknowledge the financial support in the form of a Visiting Fellowship at the Australian National University’s Humanities Research Centre during semester 1 in 2012, which provided me with the necessary time to work on this book. This book represents a thoroughly revised, transformed, and updated version of my doctoral dissertation, defended at the Ohio State University. I want to express my deep gratitude to my advisor, Dieter Wanner, and the members of my dissertation committee, Brian Joseph, Javier Gutiérrez-Rexach and Janice Aski, for their constant help and support throughout the years and for their feedback on previous versions of this work. I have been very fortunate for the invaluable feedback received from friends and colleagues at different stages of this monograph. In the USA, in addition to the aforementioned, input from many friends and colleagues has found its way into this book, through comments or acting as informants (or both), including Patrícia Amaral, Álex Amaral and Flávia Cunha, Lorena Andueza, Carolina Castillo Trelles and Andy Lewis, Cécile d’Agaro, Melvin González Rivera, Edith Hernández, Ma Carmen Parafita, Sandro Sessarego, Vasilis Tsompanidis, and many others. I must also thank my friend and former colleague María Luisa Guardiola for being a role model. In Spain, my dear friends Alicia Mellado and Cecilia Sáez Colmenero have always been a source of inspiration and encouragement in this challenge. In Australia, thanks go to my colleagues in the School of Language Studies at ANU; among them, I owe particular thanks to Elisabeth Mayer, Catherine Travis, Cindy Allen, Jane Simpson, and Amalia Milman for being constant sources of support and learning. Above all, I am extremely grateful to William Steed for his comments and for being a praiseworthy compañero de viajes y aventuras and inspiring co-researcher since my arrival and to this day, despite the distance between Canberra and Townsville. Many colleagues have provided feedback in different forms at different conferences or other venues on the topic(s) discussed in this book, including Rachel Hendery, Leah Houle, Erich Round, Fred Weerman, John Whitman, among others, and many of those mentioned above, especially Dieter Wanner, Cindy Allen, and Catherine Travis. Last but not least, I am indebted to my family and friends in three continents, some already listed above, who have always supplied invaluable amounts of moral support to go on, regardless of the geographical distance. This book is especially dedicated to all and each of them.

Table of contents Acknowledgments List of tables xi Abbreviations xi

v

1 1 2 3 4

Introduction 1 Topic of study 1 Sources of data 3 Data in historical studies Organization of the book

2 1 2 2.1 2.2 2.2.1 2.2.2 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 3 3.1 3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 3.2 3.3 3.4 3.4.1 3.4.2

Categories, syntax, and change 10 Introduction 10 Categories in prepositional clauses 10 The functional category Complementizer – COMP 10 The category Preposition 16 Prepositions: types and features 16 Licensing prepositional objects with functional prepositions Prepositions and clauses 25 Prepositions and that-clauses 25 The hypothesis of the underlying preposition 29 Prepositional indirect interrogative clauses 32 Subordinating conjunctions as prepositional phrases 34 Prepositional complementizers 36 Syntactic framework 38 Case Theory 38 Case 38 Types of Case 40 Prepositions and Case 42 Critiques against abstract Case 45 Case and finite clauses 46 Determiners and clauses 48 Arguments vs. adjuncts 53 Historical syntax. Change and mechanisms 55 The nature of change 55 Mechanisms of change 59

5 5

22

viii

Table of contents

3 1 2 2.1 2.1.1 2.1.2 2.1.3 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.3 2.3.1 2.3.2

Historical Spanish 63 Introduction 63 Prepositions and clauses in historical Spanish 64 Prepositions and nouns 64 Argumental and adjunct phrases 64 Variation in argumental prepositional selection 65 Conclusions 67 Prepositions and infinitives 67 Argumental and adjunct prepositional infinitives 67 Variation with prepositional infinitives 69 Summary and syntactic conclusions 75 The nominality of the infinitive 76 Prepositions and finite clauses 80 Previous studies based on syntactic analogy 81 Previous syntactic approaches: Pronouns, nominality and Case 86 2.3.3 Summary and conclusions of previous syntactic studies 2.3.4 Argumental prepositional que-clauses 90 2.3.5 Early examples of argumental que-clauses 92 2.3.6 Adverbial prepositional que-clauses (adjunct que-clauses) 2.3.7 Prepositional indirect interrogative finite clauses 94 2.3.8 Prepositional que-clauses and pronouns 96 2.3.9 Prepositional que-clauses and variation 96 2.3.10 Finite clauses and the article 100 3 Discussion and conclusions 103 4 1 2 3 4 4.1 4.2 4.3 4.4 4.5 4.6 5

107 Present-day Spanish Introduction 107 Prepositions and (non-clausal) nouns 107 Prepositions and infinitives 112 Prepositional finite clauses 122 Prepositional que-clauses 122 Prepositional indirect interrogative finite clauses 124 Variation and optional prepositions 125 Case and the licensing of the finite clause 130 Clausal argumenthood and prepositionality 132 Finite clauses and the article 135 Discussion and conclusions 141

90

93

Table of contents

5 1 2 2.1 2.2 2.3 3 3.1 3.1.1 3.1.2 3.2 3.2.1 3.2.2 3.2.3 3.3

Crosslinguistic perspective: Romance languages I 157 Introduction 157 Latin and Proto-Romance 157 The Latin and Proto-Romance prepositional finite clause 157 The Latin and Proto-Romance prepositional infinitive 159 Conclusion 160 Portuguese 160 Historical Portuguese 160 Prepositions and nouns. Prepositions and infinitives 160 Prepositions and finite clauses 168 Present-day Portuguese 174 Prepositions and nouns. Prepositions and infinitives 174 Prepositional finite clauses 178 Nominality and the clause 183 Discussion and conclusions 193

6 1 2 2.1 2.1.1 2.1.2 2.2 2.2.1 2.2.2 2.3 3 3.1 3.1.1 3.1.2 3.2 3.2.1 3.2.2 3.3

Crosslinguistic perspective: Romance languages II Introduction 197 French 197 Historical French 197 Prepositions and infinitives 197 Prepositional finite que-clauses. The ce element Present-day French 215 Prepositions and infinitives 215 Prepositional finite clauses 220 Discussion and conclusions 231 Italian 234 Historical Italian 234 Prepositions and infinitives 234 Prepositional finite clauses 241 Present-day Italian 250 Prepositions and infinitives 251 Prepositional finite clauses 259 Discussion and conclusions 267

7 1 2

269 Conclusions A crosslinguistic overview of prepositional finite clauses 269 Change in Spanish from a comparative perspective. Data and theoretical implications 270

197

203

ix

x 2.1 2.2 2.3

Table of contents

The nominality of the clause and (re-)nominalization/ substantivization 271 Case, pronominal alternatives, and the finite clause E-language and syntactic change 278

References Subject index

292 311

274

List of tables Table 1. Prepositional clauses in present-day Romance languages Table 2. Prepositional clauses in historical Romance languages Table 3. Analogical syntactic change in Spanish

269 270 291

Abbreviations 1, 2, 3 A Acc AdvP Agr AgrP ANAPH AP ARTFL BADIP C COMP compl CORDE CP CRP CRPC D DAT DP E-language EPP F FinitenessP FNT ForceP FUT G GEN

first, second, third person adjective accusative adverb phrase agreement agreement phrase anaphoric adjective phrase American and French Research on the Treasury of the French Language BAnca Dati dell’Italiano Parlato complementizer complementizer complement Corpus Diacrónico del Español complementizer phrase Case Resistance Principle Corpus de Referência do Português Contemporâneo determiner Dative determiner phrase externalized language extended projection principle feature finiteness phrase Functional Nominalization Thesis force phrase future grammar Genitive

xii

Abbreviations

I-language INF INFL K KP MCVF N n nP NP OVI P PL Poss PossP PP Q QP RAE refl S SG Spec T θ TP V v VP vP Wh

internalized language infinitive inflection Kase Kase phrase Modéliser le changement: les voies du français corpus noun little noun little noun phrase noun phrase Opera del Vocabolario Italiano preposition plural possessor/possessive possessor phrase prepositional phrase quantifier quantifier phrase Real Academia Española reflexive 3rd person pronoun stage singular specifier tense theta tense phrase verb little verb verb phrase little verb phrase interrogative/relative

Chapter 1

Introduction 1 Topic of study This book addresses the syntactic description and evolution of Spanish argumental prepositional finite clauses introduced by que (‘that’). Particular attention is devoted to the examination of previous hypotheses and to the combination of the insight gained by traditional studies and the theoretical advances offered by current formal syntactic approaches to clausal complementation, always keeping the data at its center. Argumental finite clauses with que are situated within the general syntax of prepositional groups, namely prepositions with nouns, basic infinitival clauses, indirect interrogative finite clauses, and adjunct (adverbial) prepositional clauses, as appropriate. Argumental prepositional finite clauses are common in present-day Spanish. While grammatical in Portuguese as well, this configuration does not seem to be shared by other closely related languages such as French or Italian. Observe the following set of examples with argument clauses complementing the prepositional verb acordarse (to remember) and the corresponding equivalents in the aforementioned languages: (1)

a. Me acuerdo de que él era pequeño Me recall.1SG of that he was little

Spanish

b. Lembro-me de que ele era pequeno Recall.1SG-me of that he was little

Portuguese

c. Je me souviens (*de) qu’il était petit I me recall.1SG of that-he was little

French

d.

Italian

Mi ricordo (*di) che lui era piccolo Me recall.1SG of that he was little ‘I recall that he was little.’

While in Spanish and Portuguese the preposition de (‘of’) is permitted before a finite clause headed by the complementizer que (‘that’), in French and Italian the same syntactic configuration is ungrammatical. Furthermore, in older stages of Spanish such clauses are not generally attested (Tarr 1922; Serradilla 1995, 1996a, 1996b, 1997; Barra 2002, among others). Observe the missing preposition in the following 15th century sentence, featuring the verb membrarse (‘to remember’), a prepositional verb:

2 (2)

Introduction

no se miembra que fauoresció la toma de cantalapiedra not refl recalls that favored.3SG the seize of Cantalapiedra ‘And he does not remember that he favored the seize of Cantala-piedra.’ (Letras, Hernando del Pulgar, 15th c.)

Two comparative orientations that complement each other form the general outline of this research: 1. The first perspective is intralinguistic, and involves a diachronic description and discussion of relevant data regarding the evolution of Spanish argumental prepositional finite clauses, with attention to clausal argumenthood, clausal prepositional and non-prepositional licensing, and the role of syntactic analogy as the mechanism of change. Such historical overview serves as the basis for the description and analysis of current Spanish data. 2. The second perspective is crosslinguistic. It is centered around the description and discussion of the syntax of three other Romance languages – Portuguese, French, and Italian – both in their historical and current states. The purpose of this dimension of the study is to situate Spanish within a wider Romance context in order to compare grammatical/attested syntactic configurations and analyze and critique the theoretical repercussions of such comparisons, including to what extent apparent constraints against prepositional finite clauses are actually operative. Occasional data from Germanic languages – especially English and Swedish – are also brought into the discussion when relevant.

1.

2.

Two syntactic aspects are of particular importance: The syntax of the combination between prepositions and clauses. In order to do so, theoretical points raised include the nature of prepositions (functional and lexical), the categorial nature of clauses (Complementizer Phrases or CPs) and complementizers (Cs), and the role of syntactic licensing (Case Theory). Combinations of prepositions with nouns and infinitival clauses are also discussed. Given the nominal nature of prepositional objects, much attention is devoted to the issue of the nominality of the clause, especially the relation between nominality and the presence of a determiner with a clause, as described in very recent syntactic accounts of clausal nominalization (Panagiotidis and Grohmann 2009; and especially Kornfilt and Whitman 2011). Infinitival clauses are examined in greater detail in this matter.

Sources of data

3

2 Sources of data The evidence presented in this book comes from different sources. Certain longer examples have been shortened to highlight the relevant syntactic configurations, making sure the interpretation remains clear and unaltered. When the example or group of examples comes from the same written source and page, the reference is inserted immediately above the examples, at the end of the previous paragraph (3a). When the examples are from different written sources or from different pages in the same source, each example or group of examples from the same page contains a reference next to it or under it (3b), or next to/under the last example in groups of examples (see, for instance, 29a,b in chapter 2). Here is an illustration of this system: (3)

a. . . . Consider the following examples (x)

(Smith 2060: 4567):

a. Example ABC b. Example DEF

b. . . . Consider the following examples: (y)

a. Example GHI García (2070: 23) b. Example JKL Smith (2060: 4533)

The old data were compiled from a variety of scholarly publications (historical grammars, specialized articles, Ph.D. dissertations) and searches in corpora. The searches in these corpora were controlled for century but not for location, except for the case of historical Italian, where the examples are restricted to the Tuscany area. Here is the list: 1.

Historical Spanish: a. Davies, Mark, Corpus del Español b. Biblioteca Virtual Miguel de Cervantes: c. Real Academia Española, Corpus Diacrónico del Español (CORDE)

2.

Historical Portuguese: Davies, Mark and Michael Ferreira’s Corpus do Português:

3.

Historical French: Modéliser le changement: les voies du français corpus (MCVF ), developed as part of a Grands Travaux de recherche concertée (Conseil de recherches en

4

Introduction

sciences humaines du Canada) project directed by Dr France Martineau, from the Département de français at the Université d’Ottawa/University of Ottawa, Canada. 4. Historical Italian: Corpus OVI (Opera del Vocabolario Italiano) dell’Italiano antico, accessible through the website of the Istituto Opera del Vocabolario Italiano at the Consiglio Nazionale delle Ricerche and searchable with GattoWeb.

Several examples were obtained from having consulted OVI through the University of Chicago’s ARTFL.

As for the present-day languages, the data was collected from multiple scholarly publications (grammars, papers, book chapters, etc.). Informants were also consulted. In the case of Spanish, I added my own examples when required and consulted with other native speakers (Lorena Andueza, Carolina Castillo, Melvin González, and Edith Hernández), as properly mentioned. As for Portuguese, Patrícia Amaral (Portugal) and Álex Amaral and Flávia Cunha (Brazil) provided me with examples and grammaticality judgements, as indicated in the text. Cécile d’Agaro did the same for French, and Sandro Sessarego was my informant for Italian, once again as properly indicated in the text. Several online corpora were also used for Portuguese, French, and Italian: 1.

Portuguese Corpus de Referência do Português Contemporâneo (CRPC),Centro de Linguística at the Universidade de Lisboa, Portugal.

2.

French Corpus LEXIQUM, Secrétariat à la politique linguistique du gouvernement du Québec and Université de Montréal, Canada.

3.

Italian Corpus BAnca Dati dell’Italiano Parlato (BADIP), Karl-Franzens-Universität Graz, Austria.

Finally, several examples were taken from Google (indicated with URL and date of consultation), which has become a great source for colloquial data. The limited evidence from Germanic languages comes from grammars and other scholarly works.

Organization of the book

5

3 Data in historical studies Regarding historical data, it is customary to mention the limited availability of historical evidence as a factor which may question the value and weight of any conclusions. It is true that no historical study can have access to as much data as a present-day linguistic study, and that the recorded data may not be as pure (i.e. realistic, colloquial, non-formulaic, etc.) as desirable (Wanner 2000: 9–12; Fischer 2007: 12). Nevertheless, there is no other realistic solution than to make do with the resources at hand. Fortunately, in the case of Romance languages the evidence is sufficient for good results. Moreover, as expressions of the language of a speaker/writer or a group of them, historical texts are competent linguistic products as they are (Wanner 2000: 10). Due to the lack of native speakers and negative evidence, for obvious reasons, it is methodologically inadequate to speak of grammatical or ungrammatical constructions in the way those terms are used in studies of present-day languages with living native speakers and their intuitions. Instead, attested/ unattested is the actual category for old texts; the fact that some construction is not recorded does not imply that it was not in use at all. This does not mean that analyses of the recorded data may not lead to logical predictions in terms of grammaticality, but they always remain speculative. The goals of this book allow me to overlook any potential problems regarding the reliability of old texts as faithful reflections of the spoken language of a specific point in time. Given the importance of the 16th and 17th centuries in this book – for reasons that will be made clear later on – whether a particular early example dated in the 14th century actually reflects the language of that time or that of a previous state would not substantially alter the syntactic discussions in this book.

4 Organization of the book The book is organized in 7 chapters. The first chapter is this introduction. Chapter 2 introduces the necessary categories and the theoretical syntactic framework for the study. The first part introduces categories, starting with the concepts of complementizer (C) and the clause as complementizer phrase (CP), and their nominal distribution. The second important category is the preposition (P), divided into functional and lexical prepositions, following Rauh (1991, 1993, 2002), among many others. The concept of K(ase) is also introduced (Lamontagne and Travis 1987; Travis and Lamontagne 1992; Tremblay 1996). The next part

6

Introduction

combines both categories and revolves around prepositional clauses, with special attention to the existing constraints in languages such as English, which has molded to a degree the existing literature on the subject (most notably, Stowell 1981). Other important questions with direct consequences for the study include the phrasal nature of the traditional subordinating conjunctions (Pavón 1999, 2003, among many others), the special nature of prepositional complementizers (following Rizzi 1988), and the syntactic discussion of the cases in which an expected preposition fails to appear (Cano 1977–78, among others). The second part focuses on the theoretical framework. The first sections are devoted to Case Theory in a Chomskyan framework (Chomsky 1981, 1995), introducing the standard classification into morphological case and syntactic Case, and the subclassification between structural and inherent Case, the latter further differentiated from lexical Case (Woolford 2006). Of particular interest is the discussion of the literature on the finite clause as carrier of a Case feature (Stowell 1981; Bošković 1995; Lasnik, Uriagereka and Boeckx 2005, among many others). The following section concentrates on the issue of clausal nominality vis-à-vis clausal nominalization, and the role of the determiner, following recent works, especially Kornfilt and Whitman (2011). Infinitival clauses, as a point of comparison with finite clauses, are especially analyzed in this part. The next section deals with the problematic difference between arguments and adjuncts, where several criteria are examined. Recent works on the nominal nature of the Romance complementizers are particularly important (Manzini and Savoia 2005, 2011; Manzini 2010; Roussou 2010). The final section establishes the historical syntactic framework. It introduces the mechanisms of syntactic change (Harris and Campbell 1995; Wanner 2006) and comments upon the necessary link between more traditional approaches to change and formal Chomskyan approaches, especially by appealing to the role of E-language as locus of change and language creativity/innovations that will eventually provide the input for new I-languages (see Lightfoot 2006: 15). Chapter 3 is one of the core chapters of this book in that it documents, examines, and analyzes prepositional finite clauses in historical Spanish. After illustrating prepositional phrases involving regular nouns (regular DPs) and infinitival clauses in older Spanish, in order to show that prepositions could indeed take non-clausal and clausal complements, it centers on the main topic in this book: the evolution of argumental prepositional finite clauses in Spanish. It includes a summary of studies on the formation and syntax of prepositional finite clauses in general (both argumental and adjunct/adverbial), and, especially the formation of prepositional argumental finite clauses. Variation in the presence or absence of the expected preposition is documented. These latter studies are divided in two groups: those which rely on analogy for the emergence

Organization of the book

7

and expansion of prepositional clauses (Tarr 1922; Herman 1963; Serradilla 1995, 1996a, 1996b, 1997) and those which concentrate much more specifically on the syntactic structure inside those prepositional clauses without appealing to analogy (Moreno 1985–86, and especially Barra 2002). Abundant examples are presented and discussed in order to critique these approaches, in particular regarding clausal argumenthood, clausal nominality, and Case licensing. Special attention is devoted to the nominality and clausal nominalization of infinitival clauses and finite clauses, both introduced by que (‘that’) and indirect interrogative finite clauses. Chapter 4 presents and discusses present-day Spanish data, with a similar content layout as chapter 3. After examining prepositional phrases with nouns (regular DPs) and infinitival clauses, it focuses on prepositional finite clauses, including indirect interrogative finite clauses as well. Points of particular interest include current prepositional variation/optionality, nominality/clausal nominalization and its dependence on the determiner – with additional attention to colloquial Spanish – and the relation between prepositionality and argumenthood. The discussion in this section further calls into question the proposed hypotheses trying to explain the emergence of Spanish argument prepositional finite clauses and, especially, the role of Case (Barra 2002), with implications for Spanish and crosslinguistically. Chapter 5 consists of data from Latin and Portuguese. Latin (and ProtoRomance) is briefly documented to illustrate the existence of prepositional clauses before the actual Romance period. Portuguese receives detailed attention because an examination of this language, whose syntax was and is in so many ways comparable to that of Spanish, makes it possible to comment on what is similar and what is unexpectedly different from Spanish, in both historical and present-day terms. Prepositional optionality, clausal nominality and nominalization (including infinitival clauses; Raposo 1987a, 1987b) are also central parts of the description and analysis. Chapter 6 describes and analyzes historical and present-day French and Italian data. It follows a similar organizational pattern. Examining the syntactic possibilities of these languages at different points in history and in current times permits me to qualify the validity of the apparent constraint against argumental prepositional finite clauses typically mentioned in the literature on these languages (see, for instance, Elia, Martinelli, D’Agostino 1981 for Italian; Jones 1996 for French), and the additional expected theoretical consequences derived from such a constraint. It also discusses clausal argumenthood, clausal nominality and nominalization, the role of prepositional and non-prepositional clausal licensing and Case, and provides additional crosslinguistic material to compare Spanish with. French and Italian are particularly relevant for the

8

Introduction

role of (arguably) pronominal alternatives to direct prepositional selection of finite clauses (following especially the ideas and data presented in Rouquier 1990; Zaring 1992; Zaring and Hirschbühler 1997). Chapter 7 concludes the book. A first part focuses on the syntactic consequences for the description of prepositional finite clauses in Romance. Firstly, it provides a general overview of the existence of prepositional finite clauses, both argumental and adjunct ones, in the surveyed languages. In doing so, it visualizes the absence of a fully operative constraint against such configuration in those languages. It also summarizes the comments regarding the nominality of the finite clause in Romance and the syntactic role of the determiner in clausal nominalizations. It points out that there are two phenomena: finite clauses are distributionally nominal and essentially nominal thanks to their nominal complementizer, regardless of whether they further participate in clausal nominalization by combining with a determiner, thus creating a renominalization, or substantivization in fact (Yap, Grunow-Hårsta and Wrona 2011). That is to say, the nominality of the finite clause is shown to be independent from it projecting up to the DP level. Similar conclusions are drawn for infinitival clauses. Lastly, in showing that clauses can be licensed with or without the expected preposition (that is, positional licensing), it is concluded that Case must have been checked before and after the generalization of argumental prepositional finite clauses in Spanish (and Portuguese). Importantly, such conclusion means that French and Italian speakers could also license argumental finite clauses with or (frequently) without the expected preposition in argumental clauses. The second part of the conclusions lays out a solution to the emergence of argumental prepositional finite clauses in Spanish which combines both the traditional insight of the accounts based on analogy and the formal tools provided by very recent syntactic studies. The existence of sporadic early attestations of the relevant configuration before the 16th century shows that at least certain speakers could produce the relevant construction. Such sporadic output, other syntactic facilitating factors such as clausal argumenthood, and other already frequent models, as extensively reported in the literature, indicate that the necessary syntactic components which could favor the extension of the structure were in place. The main consequence for argumental clauses was a change in the materialization of inherent Case. As expected of analogy, other superficially – but only superficially – similar constructions were also affected. The possibility of conceptualizing analogy as an E-language mechanism foments a link between well-established (traditional) mechanisms of change and current formal syntactic analyses.

Organization of the book

9

The crosslinguistic and diachronic orientation of this study makes it possible to look at a group of closely related languages and unveil to what extent the explanations argued for one language in the literature are sustainable in another language or may lead to theoretical implications which are incompatible with the actual data.

Chapter 2

Categories, syntax, and change 1 Introduction The goal of this chapter is to introduce several theoretical tools that we will need for the discussion. There are two main groups. First, the syntax of finite clauses (COMP or C), prepositions (P), and the combination of both of them. The second has to do with the Theory of Case, in particular the nature of abstract (syntactic) Case, the role of Case licensing by prepositions, and Case and finite clauses.

2 Categories in prepositional clauses In current syntactic theory a finite clause is understood as a Complementizer Phrase, that is the projection of a head Complementizer. This section deals with the syntactic properties of complementizer phrases (CPs), their different types, in particular, declaratives and indirect interrogatives. In addition, two questions have a great impact on this book. On the one hand, I explore the nature of

as a subordinating conjunction vs. a prepositional phrase, and conclude that the latter is the most accurate analysis. On the other hand, I explore the categorial properties of the prepositions. The final section is devoted to the syntactic structure of those cases where there seems to be an invisible, underlying preposition. In all, this section aims to present and discuss the major points involved in the syntactic description of prepositional (finite) clauses.

2.1 The functional category Complementizer – COMP A complementizer (COMP or C) is a syntactic category corresponding with the traditional concept of subordinating conjunction. In keeping with the principle of endocentricity, complementizers are the heads of CPs, traditionally known as clauses, both finite and infinitival (see Rosenbaum 1967: 24; Bresnan 1970; Demonte 1977; Haegeman 1991: 111–112, among many others). A subordinate clause will be understood in these syntactic terms, a CP dependent on another category which either selects for it, if an argument, or rather is modified by the clause in the case of adjuncts (e.g., adverbial clauses).

Categories in prepositional clauses

11

There are several complementizers, introducing several types of clauses: content or declarative clauses, indirect interrogative (and indirect exclamative) clauses, and relative clauses. The discussion in this book will revolve around the first two, which can be grouped together under the general label of sentential complementation. Complementizers may introduce both finite and non-finite clauses. Those introducing finite clauses include: a. The complementizer par excellence is that (English), que (Spanish, French and Portuguese), che (Italian). Consider the following example: (1) María dijo [CP que iría a la fiesta] María said that would-go.3SG to the party ‘María said she would go to the party.’ b. Other complementizers introduce indirect interrogative clauses, for instance if/ whether (English), si (Spanish, French), se (Italian). Indirect interrogatives (and exclamatives), also CPs, may be introduced by wh-phrases located not in C, because they are not complementizers, but in Spec,CP after wh-movement (C would host an empty complementizer). Consider the following examples: (2)

a. María preguntó [CP si íbamos al cine] María asked if went.1PL to-the cinema ‘María asked whether we were going to the movies.’ b. María preguntó [CP dónde íbamos] María asked where went.1PL ‘María asked where we were going.’

The different complementizers and wh-phrases determine the type of clause (see Cheng 1991 on clause typing): a declarative clause in the case of that/que, a total indirect interrogative clause in the case of if/whether/si, and a partial indirect interrogative clause in the case of wh-phrases in Spec,CP (Haegeman 1991: 106; Fernández and Anula 1995: 314–321; Adger 2003: 292). Additionally, the complementizer that/que and wh-phrases may also introduce relative clauses, including free relative clauses. While declaratives and indirect interrogatives are both CPs, there are some further important categorial differences, in particular having to do with their nominality, which will be discussed extensively in this book. Sentential complementation can also be non-finite. Infinitival clauses are typically CPs or TPs (Adger 2003: 306, 313). Consider the following examples:

12 (3)

Categories, syntax, and change

a.

I tried [to call you]

b.

I wanted [to call you]

c.

Quería [llamarte] wanted.1SG call.INF-you ‘I wanted to call you.’

Certain infinitival clauses are introduced by prepositional-looking elements, the prepositional complementizers, usually including English for and French de/à (‘of/to’) and Italian di/a (‘of/to’) in certain cases. Gerunds in English may also be non-finite clauses, as in the well-known Acc/-ing and Poss/-ing constructions. With regards to their distribution and syntactic functions, finite and infinitival clauses can be arguments and adjuncts (Dixon 2006), open to crosslinguistic variation. Finite clauses can function as direct objects and subjects, including extraposition: (4) a.

María dijo [que iría a la fiesta] María said that would-go.3SG to the party

Direct object

‘María said she would go to the party.’ b.

María preguntó [si íbamos al cine] María asked if went.1PL to-the cinema

Direct object

‘María asked whether we were going to the movies.’ c.

[Que sea lunes] no tiene por qué ser malo That is Monday not has for what be bad

Subject

‘That it is Monday does not have to be a bad thing.’ d.

No tiene por qué ser malo [que sea lunes] Not has for what be bad that is Monday

Extraposition

‘It does not have to be a bad thing that it is Monday.’ e.

[Cuándo sea el examen] no me preocupa When is the exam not me worries

Subject

‘When the exam is going to be does not worry me.’ While Spanish allows clauses in all argumental contexts, other languages, notably English, are subject to certain restrictions. Consider the following set of examples:

Categories in prepositional clauses

(5)

13

a.

[That Jason arrived] infuriated Medea

b.

[Whether Agamemnon had triumphed] was unknown

c.

The Mayas already knew [that the world is round]

d.

I need to know [whether John wins or not]

e.

It didn’t matter (to Mary) [that John was an academic] Dixon (2006: 24)

f. *Is [that the world is round] obvious to you? Is [whether John wins or o not] of any great importance? h. *It depends on [that you come]

Adger (2003: 300) Kuno (1973: 363)

g.

i.

Kuno (1973: 370)

It depends on [whether you come or not]

English clauses may be subjects and direct objects, but those introduced by that – that-clauses – as opposed to those types introduced by an interrogative complementizer, are ungrammatical as subjects in inverted questions, which, along with additional evidence, has led linguists to argue that subject clauses are not really in subject position but rather are topicalized and co-referent with an empty pronoun in subject position (Koster 1978; Arlenga 2005).1 Most importantly for the goals of this book, that-clauses are ungrammatical as objects of prepositions. This matter will be discussed extensively throughout this book. Finite clauses are also found in adverbial clauses, which are syntactically adjuncts: (6)

a.

Before I went to the city, I called home

b.

I was reading when she arrived

c.

Jason became invisible, so that he could escape

d.

Llámame para que vayamos a comer Call-me for that go.1PL to eat.INF

Adger (2003: 329)

‘Call me so that we go eat.’

1 See also Adger (2003: 299–302); but see also Han (2005) for arguments against such analysis. Webelhuth (1992: 91) argues that CPs can be in subject position precisely because they are nominal. He further argues for a categorial differentiation between CP clauses (headed by that) and IP clauses (without that). Only the former can be equivalent to DPs, as they are headed by a category etymologically derived from a nominal, a demonstrative in the case of English that or German dass (Webelhuth 1992: 90).

14

Categories, syntax, and change

Before, when and so that introduce adverbial clauses in English. As adjuncts, they are not selected by any predicate and, therefore, they are optional. In principle, we may conclude that before, when and so that are complementizers as well, for they introduce subordinate clauses. Likewise, it is traditional to assert that Spanish adverbial clauses are introduced by subordinating conjunctions (hence, complementizers): porque (‘because’), para que (‘so that’), cuando (‘when’), antes (de) que (‘before’), etc. (see Kortmann 1998: 463) (but see section 2.2 in this chapter). Infinitives, like finite clauses, can also be subjects and complements, again with crosslinguistic constraints, and may appear in adverbial contexts. Here are some examples of Spanish: (7)

a. [Comer pan] es bueno eat.INF bread is good ‘To eat bread is good.’ b. Necesito [comer pan] need.1SG eat.INF bread ‘I need to eat bread.’ c.

Llámame para [ir a comer] Call-me for go.INF to eat.INF ‘Call me (in order) to go to eat.’

The previous examples show that clauses can occupy those positions typical of nouns, which explains why declarative and indirect interrogative clauses are normally labeled “noun clauses”. Not surprisingly, then, one of the recurrent topics in the sentential complementation literature has to do with the nominality or non-nominality of different types of clauses, which has consequently attracted a great deal of attention throughout the years. In several accounts, nominal projections are claimed to top otherwise “verbal” CPs, thus creating a DP-CP construction of sorts. For instance, such is the idea already in early works on complementizers and clausal complementation, as in Rosenbaum (1967) and Lees (1960), and later on, as in Davies and Dubinsky (1998). More recently, Han (2005) argues for a nominal shell on top of the CP (Han 2005: 99–100): (8) a.

[DP D [CP That he is a doctor]] is surprising

b.

I told him [DP [CP that she is gone]]

c.

I wonder about [DP [CP where I should go]]

Categories in prepositional clauses

15

Han supplies additional data from Korean, Greek and Spanish to support her analysis (see also Takahashi 2010). Note that this implies that clauses are not nominal themselves unless a D projection is added. Nominality also plays a great role in the differentiation between factive and non-factive clauses. Kiparsky and Kiparsky (1971) captured the difference between factive and non-factive complement clauses by positing an empty noun fact (sometimes overt, as in regret the fact that . . . or regret it that . . . , for instance) on top of the CP for the former type of clauses, thus effectively turning factive sentential predication into a DP for syntactic purposes. The typical description of the complementizer has assumed its “verbal” functional nature. As indicated above, this explains the need for some type of nominal projection on top of the otherwise “verbal” CP to turn it into a noun. However, very recent studies have argued that complementizers are themselves categorially nouns. Kayne (2008), Arsenijević (2009), and especially Manzini (2010), Manzini and Savoia (2005, 2011) for the Romance languages, and Roussou (2010) for Greek have recently challenged the assumed verbal nature of the complementizer by arguing that complementizer que/that/oti is actually a noun (or a relative pronoun), in a way undoing the traditional division between complementizer and relatives. Manzini and Savoia (2011) argue that (Romance) complementizers are nouns which select for an embedded proposition as their complement, obtaining categorial unity between complementizer and wh-phrases (relatives and wh-interrogatives), as follows (Manzini and Savoia 2011: 15): (9)

a. So che fai questo Know.1SG that do.2SG this ‘I know that you do this.’ b. Il lavoro che fai The work that do.2SG ‘The work that you do.’ c. Che fai? What do.2SG ‘What do you do?’

The difference between the complementizer and the relative or interrogative is not due to a categorial difference but rather has to do with interpretive reasons. If che binds a propositional variable, as in (9a) above, a complement clause obtains; if che binds and individual variable, as in (9b,c), a relative or interrogative reading obtains (adapted/modified from Manzini and Savoia 2011: 16):

16 (10)

Categories, syntax, and change

a0 . So [NP che [ CP [ fai questo]]] ! che x: x fai questo c0 .

[NP Che] [ CP [ fai ]] ! che x, fai x

In other terms, the classical CP structure is reorganized as consisting of a nominal element (now a noun phrase) plus a CP, which is the argument of the complementizer noun. Therefore, the complementizer is located outside of the embedded clause, outside of the CP, and it is the argument of the main verb, a preposition, etc.; it is not a functional category anymore. This new representation can be visually expressed in the following tree (simplified/modified from Manzini and Savoia 2011: 17): (11) So che fai questo

Roussou (2010) proposes a similar analysis for Greek declarative complementizers oti and pu and interrogative complementizer an (‘if’), and English that (see also Roberts and Roussou 2003: 110–121) and if, all categorially nominal. Note that this recent analysis automatically renders finite clauses necessarily nominal, even without a determiner (Roussou 2010: 587). This point will be of great importance later on in this book.

2.2 The category Preposition 2.2.1 Prepositions: types and features The second important category is the preposition (P) or, more generally, adposition. The differentiation between argumental prepositional finite clauses and adverbial/adjunct finite clauses depends in great measure on our understanding of the different types of prepositions and their corresponding properties. The Real Academia Española’s Nueva gramática de la lengua española, in its handbook version (RAE 2010: 557), defines prepositions as invariable words, usually atonic, which usually introduce a complement. Rauh (1991: 175) quotes Curme’s (1935: 87) description of the category preposition, representative of

Categories in prepositional clauses

17

English traditional grammar, as “a word that indicates a relation between the noun or pronoun it governs and another word, which may be a verb, and adjective, or another noun or pronoun”. In short, prepositions are traditionally described as invariable, short words that relate nominal categories. To this we can add that they belong to a closed class. In Chomskyan syntax, categories have been described in terms of positive and negative features: nouns are [+N] but [–V], and the opposite is said of verbs, which are [–N, +V]. Prepositions are described as the absence of features: [–N, –V] (Chomsky 1981). As such, they take complements of different types (12a–d) and may even have specifiers, i.e be modified by adverbs such as right (12e) (Rauh 1991: 181): (12) a.

Bill was here [PP before [NP Mary]]

b.

Bill was here [PP before]

c.

Bill was here [PP before [CP Mary came]]

d.

Bill came [PP from [PP behind the curtain]]

e.

Bill was here right before Mary

It would seem then that there is some type of conclusive definition of what a preposition is, but this is not the case. The literature generally agrees that they do not form a homogeneous class, at different levels.2 Firstly, prepositions usually require a (nominal) complement, but not all of them and not all the time. Transitive prepositions introduce a complement (12a), but prepositions can also be intransitive or be used intransitively if/when they do not have one (12b). Intransitive prepositions were traditionally categorized as adverbs. Secondly, prepositions can be simple words, such as on, of, with, or be more complex, including the transitive options of certain intransitive prepositions such as because of, outside of and complex (phrasal) prepositions such as in terms of, in view of, etc. (Hoffmann 2005: 23, Svenonius 2010: 130–131, 136). Thirdly, the boundary between prepositions and other categories is not always clear. As Asbury et al. (2008: 3–4) show, the limits between prepositions and adjectives (as in the case of near), prepositions and verbs (as in the case of regarding), prepositions and complementizers (see (10c) above) or even prepositions and nouns (as in the case of on top of ) is open to discussion. Distributional tests are key to discern between categories, without rejecting the possibility of 2 See also López (1970); Kurzon and Adler (2008: 3–4). See also Rauh (2010: 1–30, 389), who extensively criticizes the confusion between parts of speech and syntactic categories.

18

Categories, syntax, and change

multiple categorial membership (see Newmeyer 2000). Consequently, inventories of prepositions may differ in the literature; however, current syntactic studies would include deverbal prepositions such as considering, notwithstanding, during, according to, etc. and except, but, now, etc., not usually cited in traditional studies, on the basis that they do indeed pass several categorial distributional tests for prepositionality (Kortmann and König 1992: 672). Fourthly – and more importantly for our purposes – prepositions are normally classified in two groups: lexical prepositions and functional prepositions, somehow equivalent to the traditional classification between contentful and meaningless prepositions. While still open to debate,3 there is sufficient evidence in the literature to justify a syntactic differentiation along the lexical/functional divide, including some psycholinguistic evidence (Littlefield 2006, on acquisition). The lexical inventory consists of those prepositions which show case properties, strict subcategorization, argument structure, theta properties, inherent semantic features, and selectional requirements (Rauh 1993: 101, 121). Lexical prepositions select their complements. Consider, for instance, the following examples (Rauh 1993: 102): (13) a.

John stayed in the house

b.

John stayed in

c.

John went into the house

d. *John went into e.

I didn’t see him before the movie

f.

I didn’t see him before the movie began

g.

I didn’t see him before

The set above offers a rich sample of the properties and variety of lexical prepositions, which are all contentful: in and into express location, while before expresses time. They also vary in their selectional properties: while all three may select for a DP, only in and before may be intransitive. Notice that Rauh categorizes before as a prepositions even in (13f), where it introduces a clause and thus could be argued to be a complementizer. Consider the following extra examples regarding selectional restrictions of lexical Ps (Rauh 1993: 108): (14)

a.

Bill stayed above the creek

b. *Bill stayed above an hour 3 See Baker (2003: 303–325) and Botwinik-Rotem (2004: 13) for arguments that all prepositions are functional.

Categories in prepositional clauses

19

The ungrammaticality of (14b) proves that the lexical preposition above cannot combine freely with any type of DP but rather imposes semantic restrictions on its complements. Above requires a locative DP, not a temporal one. Lexical prepositions head their own projections: the prepositional phrase or PP. Rauh (1993) shows that PPs can be fronted (15a), can appear in cleft sentences (15b), can be coordinated (15c), can be substituted by proforms (15d), and can be modified by right (15e): (15)

a.

From the station came John

b.

It was before the movie that I saw John

c.

Before the movie began and until your arrival it was very boring Rauh (1993: 103)

d.

Bill refused before I did and John refused then too

e.

John stayed right in the house

Rauh (1993: 104)

Lexical prepositions also have case-assigning properties (for case, see section 3.1. in this chapter), as can be shown by the fact that they select for morphologically inflected pronouns (Rauh 1993: 107): (16) a. b.

Mary got the book from him Bill bought a book for her

As opposed to lexical prepositions, functional prepositions lack many of the previously noted properties. As other functional categories, functional prepositions lack selectional properties and do not restrict their complements semantically; this is done rather by the category selecting the preposition instead. In addition, they cannot be substituted for by a proform (17b), cannot be coordinated with other PPs (17d), and cannot be modified by right (17c) (Rauh 1993: 134; Horno 2002: 99–100, 173–190): (17)

a.

*Bill appealed to the station

Rauh (1993: 134)

b.

*Bill believes there

c.

*Bill believes right in science

Rauh (1993: 133)

d.

*Bill believes in science and during his life

Rauh (1993: 134)

The ungrammaticality of (17a) is due to the incompatibility of the selectional requirement of the verb appeal and its complement the station. The preposition

20

Categories, syntax, and change

to is not imposing any selectional requirements like lexical directional to would, which can freely select for locative complements independently of the verb. Notice the contrast in grammaticality between functional to (18a) and lexical to (18b) (Rauh 1993: 134): (18)

a. *Bill appealed to the station b. Bill went/sent a packet/walked/invited his friends to the station

In other words, functional prepositions do not θ-mark their complements. There is no thematic relation between the preposition and the complement. This can be seen again in the following examples (Rauh 2002: 17): (19) a. b.

The lawyer had no influence over his door John relied on his table

These examples are not syntactically ungrammatical, but they are semantically inadequate (Rauh 2002: 17). This awkwardness is due to the fact that it is not over or on which select for his door and his table respectively, but rather influence and relied. The same analysis applies to other functional Ps selected by verbs, nouns, or adjectives. Lack of θ-properties does not imply lack of Case-assigning properties. In fact, functional prepositions maintain their case-assigning properties, as can be seen in the morphological requirements of the pronoun: (20)

a.

Bill believes in her

Rauh (1993: 134)

b. *Bill believes in she In other words, lexical prepositions have a lexical-conceptual structure (LCS), while functional prepositions lack one (Horno 2002: 176). In this sense, functional prepositions would have no place in the LCS in an analysis along the lines of Culicover and Jackendoff (2005). Extra evidence of their different nature is that only lexical prepositions may in certain occasions be dropped in coordination (21a), while functional prepositions may not (21b) (Zaring 1991: 369; Demonte 1992: 426; Horno 2002: 175; but see section 2 in chapter 4 below for further discussion). Consider the following examples (Horno 2002: 175):

Categories in prepositional clauses

21

(21) a. La tesis versa sobre 4 el populismo y Ø el nacionalismo The thesis verses about the populism and the nationalism ‘The dissertation revolves around populism and nationalism.’ b. *La universidad prescindió de sus servicios y Ø su aportación valiosa The university dispensed of his services and his contribution valuable ‘The university dispensed with his work and valuable contribution.’ The deficiency of functional prepositions is supported in Svenonius (2007), who claims that all lexical (spatial) prepositions take one argument: Ground, but that functional prepositions (grammatical prepositions in his words) do not take a Ground argumental DP, since “the DP . . . is not originally a complement of the adposition, but is an argument of the verb, with the adposition being introduced separately” (Svenonius 2007: 88). While lexical prepositions are said to project their own PPs, Rauh (1993: 136) argues that functional prepositions do not project their own PP, as it is their complement which actually projects. Therefore, when the so-called complement of the functional preposition is a DP, the maximal projection would end up being a DP, not a PP. A similar perspective is entertained in Demonte (1992), who argues that some Spanish prepositional verbs do not have real Ps (see also Kempchinsky 1988: 204; Campos and Kempchinsky 1991: 175; see also Scorretti 1991: 158–161 for Italian). Botwinik-Rotem (2004: 41) claims that “[i]t is widely assumed that the internal argument of a PP-verb is not the PP, but rather the DP complement of the P”. Horno (2002: 215, fn 28) likewise argues that functional prepositions do not project a full PP, only up to P’. This selecting, thematic deficiency of functional prepositions is captured by the functional category K (ase), projecting a KP (Lamontagne and Travis 1987; Travis and Lamontagne 1992), as will be explained in section 3.1.3. in this chapter.5 4 Notice that sobre (‘about’), despite being selected by the predicate versar (and others such as hablar ‘to talk’) is considered a lexical preposition – and thus project a full PP – due to the fact that it seems to impose its own thematic requirements (see Rooryck 1996: 226, and the discussion in Neeleman and Weerman 2001: 130–132 for English about). This is not to say that the position is not argumental; rather, as happens with verbs such as guardar (‘to keep’), the complement of versar or hablar can be instantiated by using different PPs headed by different prepositions, among which sobre is but one option (see Demonte 1992: 418). 5 A similar idea lies behind Grimshaw (2000)’s Extended Projection, whereby all prepositions are the maximal projection of a nominal element. Grimshaw (2000: 119) claims that “CP and PP [are] the highest extended projections of the verbal system and the nominal system respectively, C standing in the same relationship to IP and VP as P does to DP and NP”. That is, “PPs are indeed a kind of nominal – the biggest kind there is” (Grimshaw 2000: 128).

22

Categories, syntax, and change

2.2.2 Licensing prepositional objects with functional prepositions For the reasons discussed in the preceding section, functional prepositions are said not to project a full-fledged PP. The immediate question is whether the role of these apparently negligible prepositions is such that they are syntactically invisible. This whole idea of prepositional verbs ( in particular) has been extensively advocated for by Cano (1981) for (historical) Spanish, arguing that constructions are just but another way of transitivity, even if this type includes a preposition. That is a more traditional way to capture the hypothesis that these verbs θ-mark the complement of the preposition but not the PP per se, and that the preposition does not really have much syntactic weight. In historical Spanish, there are verbs which alternate between a prepositional construction and a non-prepositional one. It is true that in some cases the presence or absence of the preposition produced different meaning. However, crucially, there are others where the direct object construction is very similar to the prepositional one (Cano 1977–78: 346): (22)

assi que el cauallero entendio en su muerte once that the knight found-out in his death ‘Once the knight found out about his death.’

(Crónica Alfonso XI)

For Cano (1977–78: 346), the value of the prepositional complement is entirely equivalent to that of a direct object. More evidence of what Cano calls prepositional transitivity is found in those verbs which could select for a variety of complements in the old language, including direct objects and several prepositional objects. Consider the following examples where such irregularity in the selectional properties of the verb pensar (‘to think’) is exemplified (Cano 1977–78: 349): (23)

a. Et quanto más pensé en la rreligión And as more thought.1SG in the religion ‘And the more I thought about religion.’ b. & pensó

(Calila e Dimna, 13th c.)

muy bien dél

& thought.3SG very well of-him ‘And he thought very well of him.’ (Conde Lucanor, Don Juan Manuel, 14th c.)

Categories in prepositional clauses

23

Cano (1977–78: 348) concludes that this variation proves that the prepositions involved are meaningless and thus essentially equivalent to a direct object. On the other hand, Rodríguez (2000: 238) argues that it is not possible to speak of prepositional transitivity given that it was common in Old Spanish for movement verbs to lack the current standard preposition. In these cases, however, given the intrinsically intransitive nature of the verbs, it is not possible to argue that, when missing the preposition, those same verbs were transitive. Neeleman (1997) and Neeleman and Weerman (2001) offer an approach to prepositional verbs which gives a specific formalization to Cano’s idea that prepositional verbs seem to behave as if the preposition were actually part of the verb and not a complete syntactic element of its own. As Neeleman (1997: 93) points out, “there does not seem to be a thematic relation between the PP and the verb, but rather between the verb-preposition combination and the DP contained in the PP”. Neeleman and Weerman (2001: 106) indicate that the preposition is indeed independent in the narrow syntax, thus the verb and the preposition do project their own phrases. Only at LF will the preposition incorporate in the verb, forming a complex predicate and θ-marking the complement DP together. Syntactically, the resulting tree would be as follows (Neeleman 1997: 104): (24)

The previous tree captures the idea that

θ-mark the complement DP without completely eliminating the presence of a PP, although it also implies that “[a]s opposed to other complements of V, however, no thematic relation licenses the presence of the prepositional phrase as a whole” (Neeleman 1997: 116). Idiomatic selection takes care of the particular one-to-one arbitrariness of preposition selection in each case (see also Hoekstra 1984: 112). Similarly, Horno (2002: 344, 358, 393, 460) concludes that some Spanish prepositional verbs such as contar con, hablar de, creer en or alegrarse de (‘to count on’, ‘to speak about’, ‘to believe in’, ‘to be happy about’) undergo reanalysis. Horno argues that the PPs involved in those PP-verbs are not truly arguments, but rather adjuncts, as they are not predicted in the corresponding conceptual structures of the selecting verbs. To be able to maintain and capture the obvious idiosyncratic selection typical of PP-verbs, she argues for syntactic reanalysis, whereby the group combines in a complex predicate. The preposition is

24

Categories, syntax, and change

also eliminated as a result. The resulting configuration arguably θ-marks the remaining DP (transformed now into an argument of ) and arguably does the same with Case (possibly lexical/inherent Case, as it would be linked to θ-marking). However, unlike Neeleman and Weerman’s reanalysis in LF, Horno argues for reanalysis in the syntax component and therefore faces certain problems (Baltin and Postal 1996), including the possibility of interpolation (25a) and the fact that the do not form a constituent while

do, as displacement shows (25b): (25)

a. Cuento con sin duda alguna la mejor ayuda Count.1SG with without doubt any the best help ‘Undoubtedly, I count on the best help possible.’ b. Con Pepe no puedo contar With Pepe not can.1SG count.INF ‘On Pepe I cannot count.’

A different approach is developed in Baker (2003). Baker regards all prepositions as functional categories and all PPs as strictly adjuncts, which obviously means that there is no such thing as prepositional arguments. However, his analysis must still account for the fact that many of the so-called prepositional verbs require the presence of their complement PPs. This can be seen in that a verb such as depend must co-occur with this complement introduced by on. Baker finds a way to overcome this descriptive problem by keeping the adjunct nature of all PPs but claiming that in these cases the PP is in reality co-referential with a nominal empty element which is the actual argument of the verb. Consider the following representation (Baker 2003: 321): (26) Chris [ VP [ depends < . . . θi > ei ] on checks ] The PP on checks is not the real argument of the verb, but, anyway, is said to license the nominal empty position complementing the verb. This hypothesis effectively advocates for a nominal object for these verbs, and places the PP out of sight. As Baker himself recognizes, however, this hypothesis still requires some extra work and further evidence for it to be complete. Adopting it would amount to eliminating other question that we have to answer: if all PPs are adjuncts no matter what, then it is useless to examine whether prepositional finite clauses are indeed argumental clauses because they would also be automatically adjuncts.

Categories in prepositional clauses

25

To sum up, there is a generally accepted analysis that the prepositions in prepositional verbs are somehow eliminated, either in overt syntax or later in the derivation. If functional Ps do not have a θ-grid, then they do not select any complements, that is, they act as bridges between the selecting category (verb or otherwise) and the object. The problems appear when trying to prove where and how functional prepositions become invisible for syntactic computation, as Neeleman and Weerman’s (2001) appeal to LF illustrates, and how to account for Case licensing.

2.3 Prepositions and clauses 2.3.1 Prepositions and that-clauses Linguists generally agree that English does not allow prepositional finite clauses introduced by that (that-clause for short). This is proven with the following examples, whose Spanish or Portuguese translations would all be grammatical: (27)

a.

*It depends on that you come

b. *I count on that she will come c.

*I am happy of that she got it

The same restriction applies to infinitives: (28)

a.

*I count on to go

b. *I am happy of to go c.

*We were speaking about to go

English finite clauses and infinitives contrast with DPs and gerunds. Emonds (1972: 31) indicates that “[a] clearcut indication that infinitives and sentences are not in the same category as gerunds and noun phrases is that the latter two but not the former two appear after the traditional class of prepositions (see also Quirk et al. 1985: 659, among many others). Consider the following set of examples illustrating such difference: (29)

a.

He insisted on an adjournment

b. He insisted that we adjourn c. *He insisted on that we adjourn

Huddleston and Pullum (2002: 1019)

26

Categories, syntax, and change

English may resort to intervening pronouns, but only in very limited cases: (30)

a.

I’ll see (to it) that it’s done on time

b. *It depends on it that you come c. *I am happy of it that she got it d. *We are interested in it that he studies here In conclusion, it would seem that there exists a constraint of the type *

in English. However, such constraint must be relaxed in its application, as there are examples directly challenging it. Firstly, Andersson (1974: 6), Keyser (1975: 28, fn 5) and other linguists have pointed out the existence of in that, except that, or but that, all adverbial

: (31) a. b.

In that there’s nothing left to say, I’m leaving

Napoli (1993: 111)

He said not a thing except that he was sorry Huddleston and Pullum (2002: 642)

Furthermore, Rauh (1993: 102), Napoli (1993: 111), and Huddleston and Pullum (2002: 600) find extra prepositional finite clauses in examples such as the following: (32)

a.

I wouldn’t cry just on account of [his mother yelled]

b. He left after [he saw her]

Napoli (1993: 111)

Huddleston and Pullum (2002: 600)

The same applies to before, since, and until: (33)

a.

Before [she left the house]

b. Since [she left the house] c.

Until [we came to the party]

In (32b) and (33) after, before, since, and until can arguably be categorized as prepositions or as complementizers. Larson (1990: 170) considers the structure of such prepositional clauses to be

despite the absence of the complementizer that, thus maintaining categorial integrity. On the other hand, Dubinsky and Williams (1995: 127) support the latter analysis, precisely because

Categories in prepositional clauses

27

they cannot introduce a that-clause.6 They contrast such constructions with additional evidence from certain dialects in South East USA where that may be optionally overt – thus a

analysis in their view – as (34) illustrates (Dubinsky and Williams 1995: 126): (34) a. b.

They never came to church without (that) they brought their Bibles Gene left despite (that) John said he wouldn’t

Granath and Wherrity (2005: 1) collected several additional examples of prepositional finite clauses, including the following with considering that (see also Granath 1997): (35) Considering that Pietro was not above twenty-three years of age when he undertook a load of heavy responsibilities, his conduct reveals him as one of the most intelligent as well as the most sympathetic of Byron’s entourage in Greece. (LOB G07 39) So far the attested examples are all adjunct clauses. However, the data prove that the constraint against prepositional finite clauses is not fully operative even for argumental clauses. Consider for instance the following example, found online, where resides in selects a finite clause: (36) Their plausibility resides in that they stand as innocuous, ordinary, Anybody’s Alternatives to what turned out, on some particular occasion, to be the actuality. (“At first I thought. A normalizing device for extraordinary events”, Gail Jeffferson, accessed in Googlebooks) (3-8-12) Granath and Seppänen (2004: 11) report more examples of argumental prepositional finite clauses, although not acceptable to all native speakers of English: (37)

a.

Their importance lies in that they took place at all . . . (The Guardian, 2003)

b. but all indications point to that discodermolide might be the next drug of choice . . . (CNN Science & Technology, June 26, 1997)

6 See also Webelhuth (1992: 92), who argues that before and after select for IPs. Van Gelderen (2009) argues that these prepositions were reanalyzed as complementizers.

28

Categories, syntax, and change

Despite the restrictions in English, other Germanic languages do pattern with Spanish in allowing propositional finite and infinitival clauses, both in argumental and adjunct positions. Such is the case of the Scandinavian languages. Observe for instance the following examples of Swedish, where the (lexical) preposition om (‘about’) selects for different categories: a DP, a finite clause and an infinitival clause (Teleman 2005: 1619): (38) a.

Han pratade om henne He talked about her ‘He talked about her.’

b.

Han pratade om [att han hade bott i Kina] He talked about that he had lived in China ‘He talked about the fact that he had lived in China.’

c.

Han pratade om [att fara till Kina] He talked about to go.INF to China ‘He talked about going to China.’

Prepositional finite clauses are also grammatical in adverbial clauses, as evidenced in the following examples: (39)

a. Vi åkte till Ipswich efter [att vi hade beslutat oss] We traveled to Ipswish after that we had decided us ‘We traveled to Ipswich after we had made our decision.’ Holmes and Hinchliffe (2003: 423) b. utan [att någon oroas av utvecklingen] without that anyone be-disturbed.GER of developments ‘Without anyone becoming uneasy about developments.’ Holmes and Hinchliffe (2003: 421) The corresponding infinitival clauses are also grammatical (McClean 1963: 212):

(40)

a. Han reste utan [att betala sin räkning] He left without to pay.INF his bill ‘He left without paying his bill.’ b. Efter [att ha tillbringat tolv åri Fjärran Östern] After to have.INF spent twelve years in-the-Far East ‘After having spent twelve years in the Far East.’

Categories in prepositional clauses

29

The situation in Norwegian is equally clear: “Norwegian prepositions can take clausal complements as objects” (Lødrup 2004: 69). Observe the following argumental and adjunct examples: (41) a. De pekte på [at det ikke var helt riktig] They pointed on that it not was quite right ‘They pointed out that it wasn’t quite right.’

Strandskogen (1986: 136)

b. Han insisterte på [at han jobber for en våpenhvile] He insisted on that he works for a cease-fire ‘He insisted that he works for a cease-fire.’

Lødrup (2004: 69)

c. Etter [at vi hadde spist], gikk vi en tur After that we had eaten, went we in walk ‘After eating, we went for a walk.’

Strandskogen (1986: 139)

d. De så på slåsskampen uten [at de løfte en finger] They saw on the-fight without that they lifted a finger ‘They watched the fight without lifting a finger.’ Strandskogen (1986: 142) As expected, prepositional infinitives are grammatical as well, both in argumental and adjunct positions: (42) a.

Per leste dikt uten [å le] Per read poems without to laugh.INF ‘Per read poems without laughing.’

b.

Lødrup (1991: 126)

Han gjemte seg for [ikke å bli funnet] He hid himself for not to be.INF found ‘He hid himself so as not to be found.’

Strandskogen (1986: 144)

Danish (Allan, Holmes, and Lundskær-Nielsen 2000: 113) and Icelandic (Thráinsson 1979: 25, 2007: 411) behave like Swedish and Norwegian, thus proving that the partial restrictions in English, including the nature of the clause, are not applicable crosslinguistically.

2.3.2 The hypothesis of the underlying preposition Rosenbaum (1967: 7, 81–92) argues that English clausal complements without prepositions are governed by a preposition in deep structure, which is deleted

30

Categories, syntax, and change

in surface structure (see also Lakoff 1968: 47). The existence of the preposition is supported by the fact that it resurfaces in the right contexts, for instance, whenever the clause becomes the subject of a passive (Rosenbaum 1967: 83): (43) That the plane flew at all was marveled at by them Consider also the results from pseudo-cleft, where the prepositional verb concur in maintains its prepositions (Granath and Wherrity 2005: 2): (44) What they concurred in was that there was strong scientific evidence that the risks of not immunising children were real French and Italian do not seem to allow prepositional sentential complementation either (see Elia, Martinelli, D’Agostino 1981 for Italian; Jones 1996 for French), but see chapter 5: (45)

a.

Sono sicuro (*di) che lei è felice am sure of that she is happy ‘I’m sure that she is happy.’

b.

Je suis sûre (*de) qu’il lira le journal I am sure of that-he will-read the journal ‘I’m sure that he will read the newspaper.’

It would seem that the syntactic analysis would imply that such clauses are direct objects. However, the Italian and French examples do not pronominalize with the expected direct object pronoun, as would be logical in direct complementation, but rather with the prepositional proforms en/y (French) and ne/ci (Italian), the typical substitutes for prepositional phrases (see Elia, Martinelli, D’Agostino 1981 for Italian; Jones 1996 for French): (46)

a.

Ne sono sicuro Of-it am sure ‘I’m sure of it.’

b.

J’en suis sûre I-of-it am sure ‘I’m sure of it.’

These examples reveal that apparently non-prepositional finite clauses are pronominalized as if they were introduced by a preposition, suggesting that they are in fact prepositional, even if there is no overt preposition.

Categories in prepositional clauses

31

Early transformational studies argued for the existence of a preposition in deep structure which would disappear on surface structure. Such is the analysis in Huot (1981: 65). Consider the ungrammaticality of the following sentences (Huot 1981: 65): (47) *Jean ne se souvient pas de que cette route était fermée Jean not refl recall not of that this road was closed pour travaux for works ‘Jean does not remember that this road was closed for construction.’ Huot assumes that there is a filter whereby the preposition is eliminated and the sentence becomes grammatical. In essence, the previous examples seem to suggest that prepositional verbs do indeed maintain their prepositions when selecting for finite clauses, with the only difference that the preposition is invisible, silent, empty, etc. An immediate consequence of such an approach is that examples which seem to not be prepositional finite clauses would turn out to be so. However, Huddleston and Pullum (2002) argue against such analysis. For these linguists, there are several problems with the previously mentioned examples. First, the examples of pseudo-cleft are too marginal in English and therefore questionable in their grammaticality. Furthermore, some pseudo-cleft sentences do not have a grammatical “reconstructed” prepositional form (Huddleston and Pullum 2002: 1020): (48)

a.

What we’re counting on is that they won’t all turn up

b. *We’re counting (on) that they won’t all turn up This second argument, though, is invalid, as it only remarks the initial ungrammaticality that we are to explain and is thus circular. Lastly, they point out several prepositional verbs which do not accept left dislocation (Huddleston and Pullum 2002: 1021): (49)

a. *That the report represents a serious indictment of the banks they concur in b. *That a peaceful resolution can be found we must all hope/pray for

In short, Huddleston and Pullum (2002: 1019) conclude that “[a]n oblique/object distinction cannot be justified by invoking latent prepositions”.

32

Categories, syntax, and change

A similar conclusion is reached for Norwegian and English in Lødrup (1991: 126–128, 2004: 84–88), again on the basis of the different behavior of prepositional and non-prepositional clauses in topicalization, for instance (Lødrup 2004: 61): (50)

a. *That it would rain, everybody hoped b. That the earth is round, everybody believed

Building on Darlymple and Lødrup (2000), Lødrup (2004) maintains a categorial/ functional differentiation “between clausal complements that are objects and clausal complements that are non-objects” (Lødrup 2004: 63), thus rejecting the non-existence of a potential underlying preposition in the former cases. While (50a) would be a non-object, (50b) would be an object, as it can be topicalized. Coincidentally, while Lødrup’s LFG-informed multiple categorization of CPs as either OBJ or COMP depending on their grammatical functions will not be followed in this book, his research shows that a language such as Norwegian, which allows optional preposition “deletion”, must have both types of clauses, and thus is a mixed language for clausal complementation in his terms (Lødrup 2004: 65). This is the same exact situation found in Spanish or Portuguese, as will be discussed in the following chapters. It seems, then, that there is no clear evidence in support of the existence of empty prepositions. What can actually be concluded, though, is that nothing in the previous discussion prevents the preposition-less clauses from being arguments in the appropriate selected context. 2.3.3 Prepositional indirect interrogative clauses Thus far we have seen that only certain prepositions accept finite clauses introduced by that in English. We could conclude that

is simply mostly ungrammatical. However, Huddleston and Pullum (2002: 1012) make it clear that “[t]raditional grammar accepts that prepositions can take finite interrogative clauses”. Therefore, there are additional prepositional finite clauses in English, as evidenced in this set of examples (Huddleston 1984: 340): (51)

a.

He was worrying about [who he should trust]

b. He raised the question of [why it had been concealed] Indirect interrogative clauses introduced by prepositions are also perfectly grammatical in Swedish (52a), Norwegian (52b), Danish (52c), and Icelandic (52d):

Categories in prepositional clauses

(52)

33

a. Jag tvivlar på [om han vill hjälpa oss] I doubt of whether he wants help.INF us ‘I doubt if he will want to help us.’ Holmes and Hinchliffe (2003: 325) b. Jeg lurer på [om han kommer] I wonder on whether he comes ‘I wonder if he is coming.’ c.

Strandskogen (1986: 137)

Jeg undrer mig over [hvem der kan have ringet] I wonder me over who he can have called ‘I wonder who he may have phoned.’

d.

Spore (1965: 203)

Jón var að hugsa um [hvort María mundi koma] Jon was to think.INF about whether Maria would come ‘John was thinking about whether Mary would come.’ Thráinsson (1979: 25)

In English, while prepositions are mostly ungrammatical with that-clauses, they remain optional in certain interrogative clauses: (53) We can’t agree (on) whether that is the best choice However, prepositional optionality does not apply across the board, as the following examples attest: (54)

a.

This is my account *(of ) who really won the war

b. This is my verdict *(of ) who is guilty c.

The dilemma *(of ) whether we should go or stay

Finally, interrogative infinitives can also be prepositional objects, again in opposition to non-interrogative infinitives (Huddleston and Pullum 2002: 641): (55) We can’t agree on [how much to charge] Indirect interrogatives seem to differ from that-clauses in English. Yim (1984) points out the paradox caused by the different behavior of interrogative clauses and that-clauses as objects of prepositions. Yim (1984: 195) argues for a syntactic difference between that-clauses and interrogative clauses, by assuming that indirect interrogatives fit well in the classical definition of a nominal category, namely being able to be the object of a preposition and receive Case. However,

34

Categories, syntax, and change

while appealing, Yim’s solution for English faces some serious crosslinguistic problems, starting with Spanish and Swedish, for instance, where both types of clauses can easily be prepositional. Furthermore, the fact that some that-clauses can indeed be introduced by prepositions, as shown in section 2.3.1 in this chapter, qualifies Yim’s conclusions with regards to the nominal nature of interrogative clauses vis-à-vis that-clauses in English (see also Roussou 2010). Additionally, the fact that that-clauses and indirect interrogative clauses do not share one syntactic environment is counterweighed by the ones they do. It is true that in English certain contexts discriminate between that-clauses and interrogatives clauses. Consider, for instance, the following examples: (56) a. *Is [that the world is round] obvious to you? Kuno (1973: 363) b. Is [whether John wins or o not] of any great importance? Kuno (1973: 370) But there are nevertheless syntactic contexts where both types of clauses show the same behavior, which consequently must be accounted for in similar ways. For example, both types can be the object of transitive verbs, as is well-known: (57)

a.

I know [that you can solve this problem]

b. I know [how you can solve this problem]

2.3.4 Subordinating conjunctions as prepositional phrases It was mentioned above (see section 2.3.1. in this chapter) that after, before, or until can be categorized as complementizers when they introduce a clause or, to simplify categorization, remain as prepositions which may take also a clause as their complement. Thus there is a theoretical debate on the limits and relationship between prepositions and the traditional subordinating conjunctions (complementizers). Emonds (1976, 1985) goes further claiming that all complementizers are indeed prepositions. However, as opposed to after or before, the nature of that or whether does not seem to fit in well with the current studies on the nature of prepositions. Similarly, Huddleston and Pullum (2002) include all subordinating conjunctions, except the complementizers that, whether, and if, into the group of prepositions. Traditional grammars of Spanish (and other Romance and Germanic languages) list porque (‘because’), para que (‘in order that’), hasta que (‘until’), etc. under the label subordinating conjunctions. In current syntactic terms,

Categories in prepositional clauses

35

porque, para que, hasta que and que (‘that’) or si (‘if’) would all be complementizers, located in C: (58) [ CP [ C que, porque, hasta que [ TP . . . ] ] ] Several linguists have challenged this analysis throughout the years. López (1970: 19, fn 23) comments that already Pottier and Coseriu had argued that so-called conjunctions such as para que are but prepositions introducing a finite clause. More recently, Pavón (1999, 2003) and Barra (2002) support the same analysis. Accordingly, para is always a preposition, be it followed by an infinitive, a finite clause or a noun, like the analysis for after, before or until mentioned above. Pavón (1999: 630–632) examines the two possible analyses of

in Spanish, finally rejecting the unitary analysis. In favor of a unitary analysis, Pavón highlights the following points: 1. The preposition and the complementizer seem to form a unit as they cannot be separated by the coordination of two clauses (59a) (see also Bosque 1998: 214). This is ungrammatical when the que-clause is selected by a verb (59b): (59)

a. *No salió porque quería quedarse en casa y Not went-out.3SG for-that wanted.3SG stay.INF in home and que no le apetecía ver a sus amigos that not him felt-like see.INF to his friends ‘He didn’t go out because he wanted to stay at home and because he didn’t feel like seeing his friends.’ Pavón (1999: 630) b. Creo que me voy a quedar en casa y que Think.1SG that me go.1SG to stay.INF in house and that voy a estudiar un poco más go.1SG to study.INF a little more ‘I think I’m going to stay home and I’m going to study a little more.’ Pavón (1999: 631)

Nevertheless, as Gaatone (1981: 206) points out, “insegmentabilité ne signifie pas inanalysabilité”, that is, units such as French au, aux, du or des, or Spanish del, al are not phonetically divisible but they contain two syntactic elements: the preposition and the article. Therefore, porque and the rest can be regarded exactly the same way.

36

Categories, syntax, and change

2. A DP, adverb or pronoun cannot be conjoined with a que-clause if they are prepositional objects of a preposition; such ungrammaticality does not arise when they are direct objects of a verb (Pavón 1999: 631): (60)

a. *Te lo he traído para eso y que me des tu You it have.1SG brought for that and that me give.2SG your opinión opinion ‘I have brought it for that and to have your opinion.’ b. Dijo eso y que, por favor, no le molestaran más Said.3SG that and that, please, not him bother.3PL more ‘He said that and that he did not want to be bothered anymore.’

In favor of the PP analysis, Pavón (1999: 631) highlights that those same prepositions may introduce not only finite clauses, but also pronouns or other DPs, without any meaning change, which allows her to conclude that the PP analysis is the most realistic one (Pavón 1999: 631): (61) No fue por eso/que no quiso/ tener un compromiso Not was for that/that not wanted.3SG/ have.INF a commitment anterior previous ‘He didn’t go for that/because he didn’t want to/for having a previous appointment.’ Notice that this substitution test is exactly the same one adduced by Rauh (1993), Huddleston and Pullum (2002) and many others for English after, before or until. Additional support comes from studies of other languages, including Andersson (1975: 16) for the Scandinavian languages, Barreto (2002: 189) and Lemle (1984: 165) for Portuguese, and Le Goffic (1993: 436) and Gaatone (1981, 2004) for French, among others. It seems that the prepositional analysis is correct and I will adopt it in this book. 2.3.5 Prepositional complementizers There are specific functional prepositional-looking elements that introduce infinitival clauses, usually labeled prepositional complementizers. English for is said to be one of them. French à (‘to’) and de (‘of’) and Italian di (‘of’) and a

Categories in prepositional clauses

37

(‘to’) are also usually included in the same group. Consider the following Italian examples with di (Vincent 1997: 171): (62)

a. Paolo dice che Giorgio non può venire Paolo says that Giorgio not can come.INF ‘Paolo says that Giorgio cannot come.’ b. Paolo dice di non poter venire Paolo says of not can.INF come.INF ‘Paolo says that he (Paolo) cannot come.’

Vincent (1997) argues that di and a are complementizers in Italian because is equivalent to . Since che is a complementizer, so must be di and a. Leonard (1997: 137) argues that functional prepositions (his weak prepositions) are complementizers with Case-assigning properties (Leonard 1997: 178), effectively implying that prepositional complementizers have a mixed nature. Since Leonard examines only French and Italian, he concludes that prepositional complementizers occupy the C position because they are in complementary distribution with C que/che (Leonard 1997: 196). The Spanish data are not included. Rizzi (1992), in his well-known study about the fine structure of the left periphery and the split CP hypothesis, which breaks the CP into ForceP and FinitenessP, concludes that the complementizer che occupies the position of Force (the highest head), while its infinitival counterpart di occupies the Finite head (the lowest head). Di is then a complementizer, as (63) captures: (63) [ Force P [ Force que . . . [ Finiteness P [Finiteness di ] ] ] ] Other works approach the complementizer-like nature of these prepositions in a similar but not exactly identical way. Kayne (1999) speaks of complementizers that “look like” prepositions, such as Italian di, French de and English to. In his critique of Kayne’s paper, Borsley (2001) calls French à/de, Italian a/di and English to prepositional complementizers. On his part, Burchert (1993: 184) postulates, too, that French de is an infinitival complementizer in the following syntactic structure: (64)

Jean a promis [CP [C de [IP lire le livre]]] Jean has promised of read.INF the book ‘Jean has promised to read the book.’

38

Categories, syntax, and change

A partially similar alternative is Martineau and Motapanyane’s (2000) analysis of French, where prepositions like à or de in infinitival clauses are said to be categorially prepositions that occupy C. Rizzi (1988: 516) allows for a clear division between (functional) prepositions and (also functional) prepositional complementizers. Rizzi restricts the roster of prepositional complementizers by differentiating – initially for Italian, but applicable to other Romance languages – between those prepositions required by verbs with all kinds of complements (clause, nouns, etc.), such as the a in aspirare a (‘to aspire to’), from those prepositions – e.g., the prepositional complementizers di and a –which are only required with an infinitive, as in Credo di voler partire (‘I believe I want to leave’), but are not present with any other categories. These infinitival complements can be substituted by the pronoun lo (‘it’), evidence that they are not of true prepositional nature. In this way, it is possible to pin down prepositional complementizers to some specific uses of – especially – de, a in certain specific infinitival contexts under very precise circumstances. Should the preposition be retained with other categories, then it would not qualify as prepositional complementizer, but rather as a (lexical or functional) preposition. This is how I will describe prepositional complementizers in this book.

3 Syntactic framework 3.1 Case Theory The goal of this section is to introduce the concept and syntax of Case, with special attention to the difference between structural Case and inherent Case (and lexical Case), and to discuss the relationship between Case and prepositions, and Case and clauses.

3.1.1 Case Blake (2001: 1) defines case as “a system of marking dependent nouns for the type of relationship they bear to their heads”. This definition captures the essence of case as a relation between a selecting category and a nominal element (see also Butt 2006 for an extensive overview). Case in Chomskyan syntax serves to license nominal elements in the clause, as originally expressed in the – now superseded – Case filter, which established that every overt NP (= DP in current terms) must be assigned abstract Case

Syntactic framework

39

(Chomsky 1981: 49). Under this version, Case was assigned by certain categories – namely, Verb, Preposition, and Inflection (= Tense) – while others – namely, Noun and Adjective – could not and thus necessitated the assistance of additional prepositions. Case was assigned under government, for instance with verbs and prepositions selecting and governing their complements, or in a specifier-head configuration, as happens with Nominative Case (see Hornstein, Nunes and Grohmann 2005: 113, among many others). Consider the following examples: (65)

a. Juan compra el pan John buys the bread ‘John buys bread.’ b. Juan lo compra John it buys ‘John buys it.’

In (65a) Juan and el pan are nominals, and so they must be licensed in order to be visible to be part of the clause (the former Visibility requirement). While Juan receives its Nominative Case from inflection/tense, el pan is assigned Case by the verb compra. The same process takes place in the pronominal alternative in (65b). In (65b) the pronoun lo, given the context, is substituting for el pan in (65a). Notice, however, that in Spanish only personal pronouns such as lo show case morphology. As opposed to language such as Latin, German or Finnish, the overt, visual expression of case is not evident. Neither is it in English. However, Chomskyan linguistics (and other frameworks) assumes that the licensing which is overtly expressed by the morphological inflection on the pronoun in (65b) above is also present – albeit abstractly – in Juan in (65a, b) and el pan in (65a). Thus, abstract Case (capital C) plays the same role as the general property licensing nominals (DPs) in a clause as morphologically overt case (lowercase c) (Culicover 1997: 24, among many others). Consequently, morphological case is only a part of a wider notion of Case (see Haegeman 1991: 144; Culicover 1997: 24). Chomsky (1995: 110) indicates that “[i]n some languages (Sanskrit, Latin, Russian, . . .), Case is morphologically manifested, while in others, it has little (English, French, . . .) or no (Chinese, . . .) overt realization. In line with our general approach, we assume that Case is always present abstractly”. The notion of Case and its assignment/marking have been subject to thorough revision throughout the years, still maintaining its role as a nominal licensing property. In the Minimalist Program, much of the theoretical skeleton

40

Categories, syntax, and change

of Principles & Parameters is revised and reinterpreted, including Case. For one thing, the crucial relation of government was abandoned.7 In Minimalism Case is revisited in terms of feature checking. Case-checking categories such as light v or Tense are said to contain an uninterpretable Case feature which either checks and erases/values a matching Case feature present on whatever DP they select for, or establishes an Agree relation with it (Adger 2003: 217; Hornstein, Nunes and Grohmann 2005: 140). Therefore, assignment is transformed into feature matching/checking and feature valuing/erasing (Butt 2006: 75; Hornstein, Nunes and Grohmann 2005: 121).

3.1.2 Types of Case The linguistic literature mentions two main types of Case: structural Case and inherent Case. The former is independent of θ-assignment, while the latter is linked to θ-role assignment – discharged by the same category that θ-marks the complement. A third type – lexical Case – has been proposed to account for idiosyncratic Case (Woolford 2006). Structural Case is not strictly linked to a θ-role and is assigned to any DP realizing a particular syntactic configuration. Nominative and Accusative are generally assumed to be structural Cases. Consider the following pair of examples (Butt 2006: 57): (66)

a.

Julia melted the ice cream into mush

b. The ice cream melted into mush In both sentences the DP the ice cream receives the same θ-role: theme or patient. While the θ-role remains identical in both configurations, its abstract Case changes: while in (66a), the ice cream receives abstract Accusative Case, as is typical of direct objects in English, in (66b) this same DP receives abstract Nominative Case, typical of subjects in English (Butt 2006: 58). The dissociation between Case and θ-role marking essential to structural Case is clearly seen in the DP in subject position, which is θ-marked by light verb as its external argument (Hornstein, Nunes and Grohmann 2005: 102) but checks its Nominative Case feature with Tense. Likewise, a DP in object position

7 See Bobaljik and Wurmbrand (2008) and Lasnik (2008) for a historical overview of Case in Chomskyan generative grammar.

Syntactic framework

41

receives its θ-role from Verb, as its internal argument, but checks its Accusative Case feature with light verb. Inherent Case assignment can be regarded as the realization of a θ-role (Culicover 1997: 44), as it is linked to θ-role assignment (Blake 2001: 60; Woolford 2006: 123). Inherent Case is thus marked by the same category that θ-marks the complement, for instance, oblique Case with lexical prepositions. Haegeman (1991: 165–166) illustrates the classical description of inherent Case: (67)

a. Sie hilft ihm She helps him-DAT ‘She helps him.’

Haegeman (1991: 165)

b. Sie gedachte vergangener Freuden She remembered past-GEN joy-GEN ‘She remembered past moments of joy.’

Haegeman (1991: 166)

The kind of case – case, as it is morphologically overt – assigned to the pronoun ihm in (67a)– that is, Dative but not Genitive – is entirely dependent on the selecting category, here the verb helfen. Compare it to the verb gedanken in (67b), which, unlike helfen, selects for a Genitive object, not a Dative one. The thematic requirement of the German helfen specifies that it necessitates an argumental object in Dative case. Woolford (2006) reformulates nonstructural Case by differentiating between inherent Case and lexical Case. Lexical Case is typical of themes or internal arguments, idiosyncratically marked by the lexical categories Verb and Preposition. Inherent Case is reduced to being licensed by the functional category light v(erb) (Woolford 2006: 113). Apart from structural and inherent/lexical Case, usually related to argumental DPs, the literature offers additional types of case to account for specials circumstances, including non-argumental nominals. Schütze (2001) studies default Case, which he defines as the case forms “that are used to spell out nominal expressions (e.g., DPs) that are not associated with any case feature assigned or otherwise determined by syntactic mechanism” (Schütze 2001: 206), e.g. Accusative in English appositions, which are adjuncts. Another example is Larson’s (1985) account of adverbial nominals, which he argues license their Case on their own, i.e. thanks to the presence of a special feature [+F] in their own heads. Ernst (1996) argues for adjunct case for bare adjunct nominals. Butt (2006: 8–7), on her part, comments on for the need for adverbial case to account for non-argumental nominals.

42

Categories, syntax, and change

3.1.3 Prepositions and Case According to Woolford (2006), lexical prepositions check lexical Case. As a nonstructural Case, it involves both θ-marking and Case checking. Consider the following examples of prepositions: (68)

a.

He returned [PP from [DP his holidays]] |! +C, +θ

b.

The story [PP of [DP his holidays]]

(68a) shows the lexical preposition from selecting a complement DP. This DP receives its θ-role from the preposition and checks its oblique Case feature with the preposition as well. (68b), however, must be different because, as already discussed in section 2.2.1, such functional prepositions do not project a real PP and do not select for the DP they introduce. In other words, of in (68b) does not θ-mark the DP. The DP still licenses its oblique Case, as required, which is the function the literature has allocated to functional prepositions. In short, lexical prepositions check non-structural (lexical) Case straightforwardly while functional prepositions such as of in (68b) must do so in some other way. English of has been described as Case-marker which permits the licensing of the complement of nouns and adjectives (Chomsky 1981: 50); that is, functional prepositions are described as materializations or expressions of inherent Case (Eguren and Fernández 2004: 137, among many others). Napoli (1993: 193) explains this process with the example destruction of the city, arguing that “[t]he of that appears in [the destruction of the city], then, can be viewed as a P whose function is to assign Case to the following NP. Thus Ns and As take “prepositional Objects” rather than Direct NP Objects because Ns and As are not Case-assigners”. Going back to the German examples in (67) above, “P thus corresponds to inherent Cases in case-marking languages, for example, the accusative in the context of lehren or the genitive in the context of gedenken in German, which are also assigned lexically” (Rauh 1993: 136; see also Rauh 2002: 18; Stowell 1981: 126–127). Therefore, since nouns and adjectives may not check Case, the presence of the functional preposition is seen as the solution. It would seem, then, that the materialization of the preposition is a last resort operation to save an otherwise ungrammatical structure. Such is the classical solution presented in Chomsky (1981: 50–51). However, Chomsky (1995: 113) rejects such conclusion by noting that of cannot be inserted in any other “special” contexts to avoid ungrammaticality, as the following examples clearly show (Chomsky 1995: 113):

Syntactic framework

(69)

a.

*It seems of Susan to be here

b.

*I am proud of Bill to be here

c.

*Bill tried of Mary to be here

d.

*Bill believed sincerely of Sam

43

Chomsky (1995: 114) argues instead that this P is to be understood as the materialization of Case restricted to nominal ([+N]) complements. The link between functional preposition and Case finds its ultimate expression in works such as Lamontagne and Travis (1987: 177), Travis and Lamontagne (1992: 161) and Tremblay (1996), who reject the prepositional nature of functional prepositions and postulate that they belong to a new functional category called Kase (K). As Tremblay (1996: 86) puts it, “[d]ummy Case assigners do not belong to the category P, but rather to the functional category K (Kase). Dummy Case assigners not being inherently relational do not license arguments. This lexical difference expresses the fact that DCAs [Dummy Case assigners] do not link arguments”. This formalizes the basic idea in Rauh (1993), as the DP is the actual argument. The syntactic representations of KPs and PPs are as follows (Tremblay 1996: 87): (70)

Despite their similar layout, the essential difference resides in the fact that K lacks an argument structure and, as such, does not relate two DPs, which is exactly what a true preposition – a relator – does (Tremblay 1996: 87). Consequently, the specifier and complement positions of a preposition must always be occupied, assuming that the specifier may be filled with an empty PRO coindexed with a DP, as exemplified below (Tremblay 1996: 87): (71)

Mary put the book on the table

44

Categories, syntax, and change

KPs are consequently extended to other Case configurations, therefore becoming the syntactic category for all types of Case, including both structural Case, as in the object of the transitive verb appreciate in (72a), and inherent (or lexical) Case, as expressed by the functional of introducing the complement of the noun appreciation in (72b) (Travis and Lamontagne 1992: 166): (72)

a. Appreciate Mary

b. Appreciation of Mary

If inherent Case is expressed by functional prepositions and if prepositional verbs such as depend on, count on, acordarse de (‘to remember’), etc. involve functional prepositions,8 then those verbs involve materialization of inherent Case. Botwinik-Rotem (2004: 14, 52–56) argues that prepositional verbs are thematically underspecified, which in her approach translates into Accusative Case deficiency. Since the verb cannot check the Case feature of its nominal complement, the functional preposition – PCase in her terms9 – must do it. As for thematic relations, Botwinik-Rotem (2004: 52) claims that these verbs are thematically underspecified. On the other hand, Neeleman (1997: 93) and Neeleman and Weerman (2001: 104) (see section 2.2.2. in this chapter) argue

8 Assuming Case is still discharged under reanalysis, I assume that inherent Case would arguably also be checked by the resulting complex predicates in an analysis like Horno’s (2002). 9 Botwinik-Rotem (2004) considers all prepositions as functional categories. In her study, she argues for a tripartite subclassification: PRelation , PCase and PPredication. PRelation are lexical, contentful prepositions; PCase are Case-checking elements; PPredication introduce predicative NPs, as to in tough constructions.

Syntactic framework

45

that the verbs and the functional prepositions together (post LF-reanalysis) θmark the nominal argument of the verb.

3.1.4 Critiques against abstract Case Despite the role of Case in this book, I would like to indicate that the notion of Case as abstract Case has been subject to criticism by some linguists. The first critique questions the need for Case in grammar by remarking the fact that it is redundant and unrelated to morphological case. Marantz (2000: 11) argues that “the proper treatment of morphological case necessitates a complete break between abstract Case and morphological case . . . Giving content to the theory of morphological case allows for the elimination of abstract Case theory from the theory of syntax.” Likewise, McFadden (2004: 11) argues that “[t]rue case is a phenomenon of the post-Spell-out PF branch of the derivation, and in order to understand its real role in language, we must keep it separate from whatever handles DP-licensing within the pre-Spell-out narrow syntax” (see Legate 2008 for opposite views). Both Marantz (2000) and McFadden (2004) suggest that other syntactic/ semantic properties or features can account for the effects previously attributed to abstract Case. Marantz (2000: 11) finds that “[t]he mapping between semantic roles and argument positions, augmented by the subject requirement of the Extended Projection Principle, is sufficient to license NPs in argument positions.” The Extended Projection Principle feature or EPP renders the Case feature for the subject position redundant (Harley 1995; McFadden 2004: 7). McFadden and Marantz provide extra evidence to do without abstract Case (McFadden 2004: chapter 8). McFadden (2004: 277) concludes that what is needed is some type of semantic integration, since “[n]ominal phrases do not require abstract licensing beyond what is needed for integration into the semantic interpretation”. A different line of criticism comes from those linguists who regard abstract Case as merely another label for grammatical function (Alsina 2001), thus challenging the theoretical consistency of Case. Finally, Asbury (2008) claims that morphological case is reducible to P, D or φ-features. Thus, both Ps and morphological cases are one and the same category. Case features are reducible to other elements of the grammar. Consequently, Asbury (2008: 6) concludes that “case is an epiphenomenon”. Asbury (2008: 16) puts it, “both adposition phrases and noun phrases with morphological case are PPs”, so prepositions and Case/case would be the same category with different phonetic outputs. Furthermore, Asbury (2008: 178) asserts that

46

Categories, syntax, and change

“case is only present where it is literally visible: that is, where it is morphologically distinct”. To conclude, recent research has questioned the existence of abstract Case on the basis that it is redundant and that it is a separate phenomenon from morphological case, rejecting thus the main foundational reason for postulating its existence. 3.1.5 Case and finite clauses Do finite clauses carry and/or check Case? There are three fronts in this issue: those who argue against Case for finite clauses, those who propose the opposite, and those in the middle. The first group is represented by Stowell’s (1981) well-known Case Resistance Principle (CRP), which states that “Case may not be assigned to a category bearing a Case-assigning feature” (Stowell 1981: 146). Consistent with his theoretical framework, a clause contains a category which can assign Case (Inflection/ Tense), which immediately prevents the clause from receiving Case; thus, finite clauses cannot be assigned Case. Support for his hypothesis comes from English sentences like the following: (73)

a. *We were talking about that the Marines went to China b. *I consider that John came home to be fortunate

Stowell (1981: 149)

c. *Although that the house is empty may depress you. . .

Stowell (1981: 153)

d.

That Pauline moved to Kansas surprised me

Stowell (1981: 152)

e.

Paul already knows that Jim lives with his sister

Stowell (1981: 159)

The ungrammaticality of (73a, b, c) indicates that finite clauses are not like DPs, as they cannot be objects of prepositions or appear in other typically nominal configurations, which adds extra evidence in favor of his rejection of Case for clauses as only nominals are said to get Case. Sentences (73d, e) would seem to be counterarguments because they are nominal-like positions and should be getting Case. Stowell’s answer, in line with Emonds (1976), is to assume that those clauses are actually not in argumental positions but are rather topicalized. The argumental subject and object positions are in reality occupied by empty elements co-referential with the clause. This way, clauses are supposedly structurally displaced from Case positions. However, this account does not obtain the desired results, since θ-role and Case are assigned to the whole chain (Culicover 1997: 41, among others). Under the theoretical framework of the time, the finite clause and its co-referential pronoun still share Case and θ-role.

Syntactic framework

47

Picallo (2002: 141–142) argues that argumental clauses contain a [–Case] feature. Theoretically, clauses are specified for Case, but in the end they never possess any real Case. Picallo (2002: 142) adds that “[a]ny other morphological expression of Case is impossible because it manifests a form of the [+Case] specification, a property of nominals or nominal-like syntactic objects”, which in her view clauses are not. An identical Caseless assumption is adopted by Demonte and Fernández (2005). The opposite view is represented by works such as Plann (1986), who notes that Stowell’s CRP accounts for English but not for Spanish. The main reason is that in Spanish finite clauses can be selected for by prepositions, a Case-assigner. Likewise, Contreras (1985) criticizes Stowell’s CRP on the basis of Spanish and English data, especially,

, and concludes that clauses can receive Case. The position in the middle is represented by those linguists who find that finite clauses may receive Case only in some specific configurations. For instance, Napoli (1993: 193) notes “that the fact that phrases other than NPs (such as clauses) do not require Case does not prevent them from receiving Case if they should appear in a position to which Case is normally assigned”. Lasnik, Uriagereka and Boeckx (2005: 16) likewise argue that finite clauses may receive Case, for instance when in subject position: (74) a. *It seems that 2 + 2 = 4 to be true b. That 2 + 2 = 4 seems to be true The fact that (74b) is grammatical is proof that a finite clause can be in subject position, and, accordingly, receive Case. Notice that Stowell’s topicalization hypothesis is not advocated for. Another representative is Bošković (1995). Bošković (1995: 33) claims that “although clauses can appear in Caseless positions, they need Case when they function as subjects”, even when they undergo Topicalization (Bošković 1995: 38), contrary to Stowell (1981). Bayer, Bader, and Meng (2001: 470) point out that German dass-clauses have nominal properties without being exactly like DPs. These finite clauses can be in Nominative and Accusative positions and have those structural Cases checked. However, they cannot occupy a position where they would be assigned Dative. Thus structural Cases such as Nominative and Accusative may be realized by finite CPs, but not inherent Cases such as Dative. This section has exposed the continuing theoretical debate over how to better account for the relationship between (finite) clauses, Case, and nominality, all of which are of great relevance in this book.

48

Categories, syntax, and change

3.2 Determiners and clauses What does it mean for a clause to be nominal and nominalized? The literature on clausal nominalization is numerous. Well known works on the subject include Chomsky’s (1970) classical paper on nominalization, the comprehensive studies of Koptjevskaja-Tamm (1993) and Malchukov (2004); more languagespecifically, it is worth noting Abney’s (1987) study on English gerunds, a verbal/nominal category which has received extensive attention throughout the years, and Ross’s (1973) nominal squish scale. Other well regarded approaches include Bresnan’s (1997), who opts for head sharing in an LFG framework, or Malouf’s (2000), who speaks of mixed categories as well in an HPSG framework, among many others. The literature has traditionally acknowledged the categorial problem posed by these types of nominalizations, as they require accepting the existence of mixed categories (verbal and nominal), thus posing a theoretical problem for endocentricity. Clausal nominalization in Quechua (Lefebvre and Muysken 1988) and in Asian languages – particularly in Japanese, Chinese and Korean – has also received particular attention in the literature (see the many papers in Yap, Grunow-Hårsta and Wrona 2011); these languages are in general much richer in the morphological marking of nominalization than their Romance or Germanic counterparts, having developed, for instance, specific nominalizers, used for both clausal and non-clausal nominalizations. In section 2.1 in this chapter, I already showed that que-clauses appear in nominal contexts. Their nominality is ratified by their distribution since they must be licensed (and have their Case feature checked) as in those contexts typically reserved to nouns. Thus, to justify their distribution, it can be assumed that que-clauses must carry a [+nominal] feature, linked to a [+Case] feature. On the basis that some clauses may indeed combine with determiners in some languages, a second possible syntactic account for the nominality of the que-clause is to assume, as briefly mentioned in 2.1. above for some English cases, that an empty D (leading to a DP) projects in the relevant contexts. This way, all nominal-like elements would remain coherently DPs even when there is no visible determiner. The other side of the issue would be to wonder whether a determiner is always necessary for a clause to be nominal, and I will argue that the data actually suggest the contrary. Very recently, Kornfilt and Whitman (2011) offer a formal, typological analysis of clausal nominalization which maintains both endocentricity and categorial uniformity. Building on Borsley and Kornfilt (2000), Kornfilt and Whitman (2011: 1298) argue for the Functional Nominalization Thesis (FNT), which states the following:

Syntactic framework

49

(75) Nominal properties of a nominalization are contributed by a nominal functional projection. The nominalization has verbal properties below the nominal functional projection, nominal properties above it. While clausal nominalization remains still (unavoidably, as I see it) reminiscent of mixed categories, FNT maintains endocentricity by assuming homogeneous projections of categories forming a clausal nominalization. A nominal functional head will start the nominal part of the derivation. Kornfilt and Whitman (2011) argue for four different levels of clausal nominalization depending on when the nominal element projects in the tree: 1. CP nominalizations are high nominalizations in that the nominal projection is introduced on top of the complementizer phrase. Notice that the clause remains categorially verbal from the CP down. Spanish, Modern Greek and Polish, among others, display these nominalizations with finite clauses introduced by definite articles, “the sole “nominal” element in the projection of the embedded clause” (Kornfilt and Whitman 2011: 1299). Notice the following Spanish example (see the section on Spanish): (76) Lamento [DP el [CP que suban los precios]] Regret.1SG the that rise the prices ‘I regret that the prices are going up.’

The consequence is that que-clauses are only nominal when they turn into DPs, that is to say, only when an overt determiner is projected. This theoretical point will be amply discussed and critiqued later on. 2. vP nominalizations remain verbal up to vP. Such is the case of Dutch determiner nominal infinitives and Italian nominal infinitives. The latter, for instance, allow for articles, possessive determiners and adjectives, while at the same time the verb can still have direct objects. As a consequence, Kornfilt and Whitman claim that the resulting non-finite clause remains verbal in the bottom part (vP and below) and is nominal above the vP. A syntactic representation of the Italian nominal infinitive would be as follows (following Kornfilt and Whitman 2011: 1299; example from Zucchi 1993):

50 (77)

Categories, syntax, and change

Il suo continuo eseguire la canzone impeccabilimente The his continuous perform.INF the song impeccably ‘His continuous impeccable performance of the song.’

3. TP nominalizations are instantiated by Turkish nominalizations and English Poss/-ing gerunds, which do not allow adjectives or Nominative subjects. A syntactic representation of an English Poss/-ing gerund would be as follows (following Kornfilt and Whitman 2011: 1302): (78) Kim’s thankfully finishing the sonata

4. VP nominalizations are found in English complex event nominalization, Dutch nominal infinitives with determiners, and Irish verbal nouns, for instance. Typically verbal properties of the verb are not displayed (no direct object, optional argument realization), indicating that the nominal projection starts right on top of the verb. Consider the following syntactic representation of an English complex event nominalization (see the syntactic description in Kornfilt and Whitman 2011: 1302–1303, and the syntactic representation of event nominals in Rathert and Alexiadou 2010, among others). The KP of Anna Karenina must be located in the nominal part:

Syntactic framework

(79)

51

Ruth’s frequent reading of Anna Karenina

Panagiotidis and Grohmann (2009) offer a similar approach to clausal nominalization in terms of mixed projections and categorial switches. Clausal nominalizations consist of two categorially homogeneous parts (nominal and verbal) joined by a functional element named Categorial Switch. Switches are categorially mixed, having both nominal and verbal features – an interpretable nominal feature and an uninterpretable verbal feature – which act as the joining category that allows for the combination of a verbal category and a nominal one. Switches are the actual nominalizers. The authors illustrate their findings using English gerunds, which would be analyzed as in (80). (81) shows the location of switches for clausal and Poss/-ing gerunds in English (Panagiotidis and Grohmann 2009: 150): (80)

a.

[The finding of a treatment] overjoyed everyone Nominalization

b. [Their finding a treatment] overjoyed everyone c. (81)

Poss/-ing

[(Them) finding a treatment] overjoyed everyone Clausal

a. [DP D [SWITCH vP ]] Poss/-ing b. [SWITCH TP ] Clausal

Importantly, Panagiotidis and Grohmann note there is no need to posit an (empty) determiner for the clausal gerund, in spite of their being nevertheless nominal, as they can be objects of prepositions. Note too that all three previous gerunds share one more context: that of being in subject position.

52

Categories, syntax, and change

As in the case of Kornfilt and Whitman, Panagiotidis and Grohmann prove that there are syntactic limitations as to when nominalization (switches) can take place. In their case, they argue for the SWITCH Complement Hypothesis, whereby Prolific Domains – as described in Grohmann (2003), chunks of (argumentally, etc.) complete syntactic structure building a derivation, similar to Chomsky’s phases – delimit the size of nominalizations. In short, they argue for three types of possible clausal nominalizations (Panagiotidis and Grohmann 2009: 152): (82) 1. 2.

Switch at the Θ-domain (roughly, vP nominalization) Switch at the Φ-domain (TP nominalization)

3. Switch at the Ω-domain (CP nominalization) The Spanish example used above would fall into type 3, as does Modern Greek. The resulting syntactic representation would be as follows: (83)

Lamento [DP el [CP que suban los precios]] Regret.1SG the that rise the prices ‘I regret that the prices are going up.’

Kornfilt and Whitman (2011) criticize Panagiotidis and Grohmann’s stipulation of a new functional category which would only exist in order to produce clausal nominalizations. They also note that, while their study can account for four types, Panagiotidis and Grohmann’s lacks one context, that of VP nominalization, given that VP is not a prolific domain. Both accounts aim to maintain endocentricity as a property of syntactic structure, although in the end this process still requires mixing categories or mixing categorial features. Given their larger coverage of data and their more flexible derivational nominalization, I will adopt the typology in Kornfilt and Whitman (2011), as it will serve to illustrate, analyze, and compare this phenomenon in Spanish and other Romance languages in particular.

Syntactic framework

53

There is, however, one point which needs further examination. Kornfilt and Whitman’s (2011) explanation of CP nominalizations as verbal up to – and including – the CP level poses an interesting theoretical problem. Since queclauses without determiners are already distributionally nominal, their nominality needs to be captured in absence of a determiner as well. Is a clause or at least its complementizer nominal per se or does it require the presence of a determiner to become nominal? The comment by Panagiotidis and Grohmann (2009) on the nominal nature of clausal gerunds in English is of crucial importance, for it allows for the fact that clauses without determiners can be nominal (see also Roussou 2010: 587 for finite complementizers). Therefore, clausal nominality seems to be differentiable from clausal nominalization as defined above, as the former stems from distribution and does not require a determiner. In other words, already nominal clauses may combine with a determiner in participating in clausal nominalization, or, to be more exact, clausal substantivization (Yap, Grunow-Hårsta and Wrona 2011) or re-nominalization.

3.3 Arguments vs. adjuncts In this section I introduce the syntactic representation of arguments and adjuncts in current syntactic theory, with their corresponding implications. In addition, I enumerate some tests used in the literature on order to try to differentiate between argument and adjunct, with controversial results. Finally, I briefly mention problems some of tests face when confronted with data. Syntactically, the difference between complements (object arguments) and adjuncts can be easily visualized in a tree: while complements are syntactically sisters to the head they complement, adjuncts are not. This means that adjuncts are located further away from the head than complements (Radford 1988: 187; Carnie 2002: 117, among many others): (84)

The idea behind this representation is that complements are closer in that they are specifically required and selected by the heads, while adjuncts are more independent and not selected.

54

Categories, syntax, and change

In Minimalist terms, a complement phrase merges when selected in order to satisfy some features that must be checked and erased/valued, whereas adjuncts do not do have that effect. As Adger (2003: 112) indicates, adjuncts enter the derivation as an external accessory. In doing so, an adjunct, as opposed to a complement, “does not create a new [syntactic] object, it expands one of the old ones” (Adger 2003: 112). The crucial consequence is that adjuncts do not really exist for syntactic computations. Chomsky (1995: 325) suggests that adjunction “may not really belong to the system we are discussing here [Narrow Syntax or the syntactic component per se]”. In particular, Chomsky (2001: 17) assumes that adjuncts merge late, as opposed to complements. As a consequence, adjuncts don’t count: “if α is adjoined to β, the construction behaves as if α isn’t there apart from semantic interpretation” (Chomsky 2001: 17). The classical difference between arguments and adjuncts is the obligatoriness of the former and the optionality of the latter (Dowty 2003: 34). There are several other tests that have traditionally been applied to tell arguments and adjuncts apart (see Radford 1988: 189–192). The following are often cited examples of prepositional complements, where the preposition is particularly specified in each separate case (notice they are not interchangeable) (Radford 1988: 344): (85)

a.

I defer to your suggestion

b.

John waited for the taxi

c.

I can rely on you

d.

Nothing can detract from her merit

e.

He declaimed against Syntax

Several regularly prepositional verbs appear without their prepositions when selecting for finite clauses, which nevertheless are commonly described as complements: (86)

a. b.

I insist [that the Council’s decision be reconsidered] Radford (1988: 353) She inquired [who was coming]

c.

They wonder [when they’ll be allowed to leave]

d.

They debated [whether conclusive evidence had been presented] Radford (1988: 354)

Syntactic framework

55

However, there are still several widely recognized issues to be solved before a consistent differentiation between complements/arguments and adjuncts can be fully constructed. For instance, let us focus on the first test mentioned above: the optionality of the adjuncts. If arguments are always mandatory, this poses a problem in trying to account for common transitive verbs which can be used intransitively, such as comer (‘to eat’) or beber (‘to drink’) (see Vater 1978: 24 for similar issues in German). Several solutions have been proposed, but that does not eliminate the remaining basic theoretical problem.

3.4 Historical syntax. Change and mechanisms The fact that the properties of languages change across time is a profusely tested reality. As with many other aspects of linguistics, the very notion of change, that is, what constitutes change or when change happens is open to theoretical debate. Whether change is innovation (change at the level of the individual) or social change (change at the level of the society) has sparked much controversy in the literature. The role of the traditionally recognized mechanisms of change (reanalysis, analogy, and contact) in formal syntactic approaches is another open issue.

3.4.1 The nature of change Joseph and Janda (2003: 10–14) establish a tripartite division of the conceptualization of change: (87)

1.

Change as diachronic correspondences, that is, A became B is change.

2.

Change as innovations of the speaker, that is, any individual innovation – usually during language acquisition – constitutes change.

3. Change as social diffusion, that is, change as the complete dispersion in the whole speaking society. The first point, which is opposed to the other two, conceptualizes change as the correspondence between construction A at some point in history and its equivalent construction B at a later point in history. This definition of change wrongly pays no attention to intermediate steps in evolution. Joseph and Janda (2003: 13)

56

Categories, syntax, and change

remark that such is the view present in many grammaticalization studies, which focus only on the beginning and the end of change.10 The debate between change as innovation and change as social diffusion currently permeates the practice of Historical Linguistics. By viewing change as innovations of the individual speaker, generative historical linguists locate change in every readjustment of every speaker, in particular during L1 language acquisition, in accordance with Andersen’s (1973) seminal paper on language change and abduction. Hale (1998: 2) locates change (innovation) only in the acquisition process:11 (88)

A speaker with an established grammar (G1 ) will provide input for children acquiring the language. There are several moments in acquisition (S0 , etc.), all influenced by input from G1. These moments finish with the fixation of the child’s grammar: G2 . Change is then defined as the set of differences found between the input G1 and the resulting G2 . Therefore, all innovations are individual and synchronic by nature, hence abrupt (Roberts 2007: 295–297). The discontinuity of language change is crucial. The main idea in this generative approach is that change is located in the moment of acquisition by the child; it is abrupt and instantaneous, not gradual. This is the extrapolation to change of the core generative enterprise: to understand how and why children acquire a language. Innovation is seen as different from diffusion in the society. Relying on generative assumptions such as the autonomy of syntax, the importance of acquisition – as the central point of explanatory adequacy – and the well-known difference between I-language (competence or internalized language) and Elanguage (performance or externalized language), Hale (1998: 6) concludes that 10 See also Janda (2001), Joseph (2001), and Campbell (2001) for extensive criticism of so-called grammaticalization theory. 11 See also Lightfoot (2006: 89); Hale (2007: 12); and Roberts (2007: chapter 3).

Syntactic framework

57

“[d]iffusion through a population is simply not an I-language phenomenon” (see also Hale 2007: 35–47) and thus lies outside the scope of inquiry. Diffusion is just reduced to performance, to E-language (Lightfoot 2006: 89, 163). Change as social diffusion can be represented by Labov’s proposal that change has only come about once an innovation has been widely extended and accepted by a community of speakers; that is, a sociolinguistic view of change. Labov (1973: 195) states that he wants to show “how linguistic changes may be embedded in a social context, how they are evaluated, and how change may be activated at a particular time and place”. In a sociolinguistic model, “social variation does play a systematic role in linguistic change” (Labov 1973: 233). A determined change emerges first in a group and it spreads to other groups, in a slow process, until, if successful, it is fully integrated in the speech community of the language. Change is propagation, not simply (abrupt) innovation. It is necessary to highlight that generativists do not deny sociolinguistic factors and language use. Lightfoot (2006: 13–15) remarks that there is a difference between grammar change, that is, change in I-language, and language change, that is, change in E-language. While E-language is always in flux and in constant alteration, change in I-language is reduced to the process of language acquisition, and I-language is always systematic and variation-free. Variation exists in E-language. “Language change as a whole is a group phenomenon . . . And change at the level of languages often seems to take place gradually, spreading through the population from place to place and from group to group . . . In this perspective, the study of grammar change is fused with work on variation and the growth of grammar in children” (Lightfoot 2006: 13). Changes in E-language are very important as they “cause changes in Ilanguages” (Lightfoot 2006: 13) by altering the triggering experience of the new generation in such a way as to produce the setting of a different I-language. Moreover, in the generative framework there are other approaches that pay more attention to variation, in an intermediate position between strict generative positions and sociolinguistic claims (see also Fischer 2007 for extensive discussion and proposals on how to reconcile formal and functional approaches to morphosyntactic change). Kroch (2001: 699) asserts that “[t]he modern study of syntactic change . . . is also often couched in terms of learning; but . . . the study of diachrony adds complexities of its own”. Kroch (2001: 701) disagrees with other generative grammar approaches to change in several regards. Firstly, he considers that the standard idealization of instantaneous acquisition by an ideal learner is problematic in diachrony, given that constant perfect acquisition should lead to no

58

Categories, syntax, and change

change, contrary to facts. Furthermore he notes the theoretical problem posed by the fact that old texts do show variation (Kroch 2001: 721). The result is a framework that combines generative assumptions and variationist sociolinguistics. Kroch (2001: 720) indicates that “[g]iven the assumptions of generative grammar, variation in syntax which corresponds to opposed settings for basic syntactic parameters must reflect the co-presence in a speaker or speech community of mutually incompatible grammars”. Variation is hence captured by establishing separate, discrete competing grammars in a community or inside one individual (some type of diglossia). As a result, change is considered not as the abrupt moment of acquisition but as the gradual period of diffusion (competition between alternative grammars). The continuation of this idea is that those frequencies attested in old texts are reflecting competing grammars. After a period of competition, the innovative option may displace the old option (the popular S-shape curve), given that, as Pintzuk (2003: 525) states, “sociolinguistic, stylistic or psychological factors favor one option over the other.” The statistic study of these variants shows recurrent constants in the rate of change in the different contexts, what has been called the Constant Rate Effect (Kroch 1989). The competing grammars account has been criticized. Vincent (2000: 48) remarks that “[t]he problem with this approach is that every time a new set of competing forms are identified within a language, a new set of parallel grammars will have to be postulated”, which means that Kroch’s proposal has to face the problem of how many grammars an individual can have at the same time, a problem of overgeneration and lack of restriction – lack of economy (see Roberts 2007: 327–331). With the advent of Minimalism, all variation (parametric variation) is considered to be located in the features of the lexicon, particularly in the features of the functional categories, and so change comes down to change in features. For instance, Pintzuk, Tsoulas and Warner (2000: 12–13) note that the competing grammars approach could be simplified to competing features in a grammar. This would imply that it is not complete grammars which enter in competition but only lexical items inside one grammar. Roberts and Roussou (2003: 29) argue that variation among languages is due to differences in the way functional categories are phonetically realized. Formal approaches to historical research have traditionally faced the paradox of having to deduct change in I-language out of evidence from E-language, as there are obviously no informants available, only partial written records (if

Syntactic framework

59

at all), which may represent – to a better or worse degree – the real language spoken in that place at that time.12 Such I- vs. E-language problem is still present in the literature. Crisma and Longobardi (2009: 1–6) discuss the problem posed for formal theories when trying to capture I-language in diachrony, starting with the expected absence of grammaticality judgments, which can be solved by transforming grammatical/ ungrammatical into attested/unattested, the closest alternative available in historical studies. For them, E-language may be still useful if understood as “external but still individual, namely as the collection of utterances produced by a single speaker” (Crisma and Longobardi 2009: 2), but not if it is conceived as the social entity usually advocated for by the sociolinguists. On their part, Wanner (1991, 2006) and Matthews (2003) criticize, in one way or another, the secondary role attributed to E-language in Chomskyan historical syntax. In any case, all agree that E-language, in the form of the written records of the past, is the only source of information available for historical research (Miller and Wanner 2011: 119–120).

3.4.2 Mechanisms of change In their classic book on the subject, Harris and Campbell (1995) list three mechanisms of syntactic change, namely reanalysis, extension (analogy), and contact and borrowing. In this section I will concentrate on briefly introducing the first two mechanisms, which will be important for this book. Reanalysis is “a mechanism which changes the underlying structure of a syntactic pattern and which does not involve any modification of its surface manifestation”, understanding by underlying structure category membership, constituency, hierarchical configurations, cohesion, and grammatical relations (Harris and Campbell 1995: 50). Reanalysis directly depends on preexisting structural ambiguity (Harris and Campbell 1995: 72). For reanalysis to take place it is necessary to account for expressions creating or favoring the new ambiguous analysis. This is captured by Harris and Campbell (1995: 72–75) with their exploratory expressions, which are the direct result of the creativity of language. Such novel usages of language (originally for emphasis, clarity, production errors, etc.) will eventually incorporate into the language. As such, these novelties are initially part of E-language affecting new I-languages. 12 See Wanner (1991, 2006: 31–37) and Fischer (2007: 12–30) for discussion on the value of historical texts as valid representations of the language of the time and the need for philological tools in doing historical syntax.

60

Categories, syntax, and change

Harris and Campbell (1995: 77) further differentiate between reanalysis and its actualization, following Timberlake’s (1977) classic paper. Timberlake (1977: 168) defines actualization as “the elimination of rules or subrules in the norm that are evaluated as unmotivated with respect to the productive systemic principle established by the reanalysis”. Actualization is the accommodation of the newly reanalyzed structure to the rest of the syntactic contexts, shedding whatever special constrains or irregularities could be left; that is, actualization results in the regularization of the new structure in the grammar. The literature has pointed out the logical problem of the causation of reanalysis, the Regress Problem (Roberts 2007: 126). If new speakers/acquirers are to abduce a new analysis for a construction (Andersen 1973), it must mean that such secondary analysis could potentially be grammatically available already in the model grammar, which immediately poses the problem whether that older grammar already contained the innovative ambiguity and thus there is no actual reanalysis (see also De Smet 2009). It seems that there is no fool-proof conclusion to this paradox. Roberts (2007) points out that the solution lies in locating the new analysis outside of grammar proper, for instance contact; however, as he notes, not all changes are due to contact. Thus, “[t]hese subtle changes may be caused by some extrasyntactic, but still intralinguistic, factor such as phonological or morphological change” (Roberts 2007: 126), but such a hypothesis poses other problems and has to be severely downgraded when confronted with the actual data, as Allen (2002, 2010) clearly demonstrates. Out of the three mechanisms cited by Harris and Campbell (1995), only reanalysis has been regularly mentioned as a mechanismin formal syntactic studies (but see Kroch 2001 for contact). Reanalysis allows for the formalization of change in acquisition as the relation between the E-language of one generation and the I-language of the following one. In addition, reanalysis must be instantaneous by definition, and fits the view of change as innovation, including the hypothesis of parameter resetting (Roberts 2007: 122–141). For instance, Lightfoot’s (2006: 93) well-known example of the formation of English modals involves recategorization (from Verb to Modal). At times reanalysis seems to be disguised under a different label, grammaticalization. A clear example of this is Roberts (1993), who argues that grammaticalization is diachronic reanalysis, a particular mechanism where by a lexical element becomes functional by being based-generated in a functional domain, for instance the formation of Romance futures. Both Roberts (1993) and Roberts and Roussou (1999) follow the idea that grammaticalization involves the reanalysis of lexical material as functional material, producing structural simplification. Another example is Beths’s (1999) study of modal verb dare, which was

Syntactic framework

61

reanalyzed as Tense. This conception of reanalysis as (part of ) grammaticalization highlights the common use in the literature of the latter term as a supposed mechanism which turns out to be only a cover term grouping one or more of the traditionally described mechanisms of change (syntactic and otherwise), as extensively proved in Joseph (2001), Janda (2001), and Campbell (2001). (Analogical) extension is defined as the mirror process of reanalysis, that is, as “a mechanism which results in changes in the surface manifestation of a pattern and which does not involve immediate or intrinsic modification of underlying structure” (Harris and Campbell 1995: 51).13 As these linguists indicate, extension may eliminate exceptions and irregularities and it is systematic, in that it applies inside natural classes (for instance, subordinate clauses), although it may not affect every member of that class (Harris and Campbell 1995: 97, 101). The immediate consequence is variation in the language and gradualness in change (diffusion). Where the difference between actualization and analogy/extension lies has been open to debate. Harris and Campbell (1995: 80–81), however, maintain that actualization and extension are different, despite being coexistent in many cases. While some processes of analogy are more systematic – even rule-based – in the sense that they can be recognized as operating in well-defined linguistic paradigms or natural classes, the essential reliance of analogy on surface connections has made it largely incompatible with certain assumptions of Chomskyan generative grammar. In diachrony, it necessarily finds its place in E-language, given that extension needs to be gradual and cannot be reduced to innovation in acquisition. The effects of analogy as expressed in E-language will in the end serve as input for new children and will be incorporated as part of their Ilanguages (see Lightfoot 2006 for E-language changes as input for I-language). Despite this, there have been some attempts to formalize analogy, such as Kiparsky’s (2000) attempts to formalize analogy as optimization (i.e., simplification). It is worth noting that not all linguists see it possible to conceive of analogy as comprehensibly systematic. Joseph (1998) remarks that analogy is not exceptionless, a view also shared by Wanner (2006), who indicates that analogy is ubiquitous and unavoidable as a tendency. Wanner is skeptical about the systematicity of analogy in that there are no guarantees that similarity between forms is going to yield analogy, as there are also no guarantees that the analogical direction effected in one case is necessarily going to apply in similar cases; that is, “individually-based aleatory uncertainty in the direction of analogy” (Wanner 2006: 125; see also Anttila 2003). 13 See Hock (1991) for extensive documentation and discussion of the different types of changes that fall under the label analogy.

62

Categories, syntax, and change

In conclusion, reanalysis and analogy/extension are both syntactic mechanisms for change. While the former has been extensively formalized in Chomskyan historical syntax, the latter has not had the same predicament due to its dependence on surface connections, which, in any case, does not eliminate its importance, at the very least transformed into an E-language phenomenon. In the end, it is the data which must at the center of a historical study, beyond theory-internal issues.

Chapter 3

Historical Spanish 1 Introduction This chapter will be devoted to the syntax and history of Spanish prepositional finite clauses and related configurations. The study will develop in two main directions. Firstly, I will describe and analyze the syntax of prepositional phrases with (non-clausal) nouns and infinitives. Secondly, the theoretical tools presented in the previous chapter will be evaluated, with the purpose of clarifying to what extent they shed light on the history and properties of prepositional finite clauses. The study of the emergence of Spanish prepositional finite clauses has received some attention, in particular from two approaches. On the one hand, the traditional or classical approach to this linguistic problem of the origin of

has been that of analogy, that is, extension from other prepositional constructions in the language based on common surface properties. This will come as no surprise once the data are presented below. As will be seen, there were at least two prepositional syntactic constructions which would fit perfectly as the antecedent for the later prepositional finite clauses: both the prepositional infinitives and the PPs with nouns. On the other hand, more recent works have attempted to establish the internal properties which may be at stake in the change. This second group attempts to offer more specific information on the syntax of the configuration, opening at the same time important questions. Historical Spanish will be understood as covering the period prior to the extension of prepositional argumental finite clauses around the 16th/17th centuries, the key change around which this section will be revolving. Since argumental

became common in the 16th/17th century, my approach fortunately coincides with one of the standard periodization limits: that between Medieval Spanish – Middle Ages (up to the end of the 15th c.) and the beginning of Classical Spanish – Renaissance (in the 16th c.), although examples from the 16th/17th centuries may be included and discussed when necessary to understand the linguistic situation being examined. The data for this ample period are extracted from several books, scholarly articles, and other sources, as pertinently indicated; the rest of the examples come mainly from one online corpus, Davies’s Corpus del español, a free, solid corpus which allows for controlled searches, making it an indispensable tool for diachronic studies; RAE’s CORDE was occasionally used.

64

Historical Spanish

2 Prepositions and clauses in historical Spanish 2.1 Prepositions and nouns 2.1.1 Argumental and adjunct phrases This section focuses on prepositions and their (non-clausal) noun complements in historical Spanish, both pronouns and full noun phrases. The resulting PPs could be arguments of another category or adjuncts (adverbial phrases). Verbs, nouns and adjectives could select for apparent prepositional phrases with required prepositions: (1) a.

se podrie membrar de todas las cosas pasadas refl could.3SG recall.INF of all the things past ‘He could remember all the past things.’ (Estoria de España I, Alfonso X, 13th c.)

b.

saluo si se obliga a ello except if refl forces to it ‘Unless so required.’ (Leyes de estilo, c. 1310)

c.

seredes mas seguro de el will-be.2SG more sure of him ‘You will be surer about him.’ (De Doña Guillerma Desprats a Jaime II, 1318)

The same prepositions could combine with pronouns: (2)

Amembrando me de ti en mio lecho Remembering refl of you in my bed ‘Remembering you in my bed.’

(Biblia Latina, 13th c.)

These prepositions can be syntactically analyzed as functional prepositions, manifestations of K(ase) (inherent Case). Lexical prepositions are attested with nominal complements as well, both pronouns and full DPs, heading full PPs in adjunct positions: (3)

a. por mi e por todos los reyes for me and for all the kings ‘For me and for all the kings.’ (Documentos castellanos, Alfonso X, 13th c.)

Prepositions and clauses in historical Spanish

65

b. Dia es que vernan fasta ti desde assur fasta las Day is that come.3PL.FUT until you from Assur until the cibdades dela fortaleza cities of-the fortress ‘The day will arrive when they’ll come to you from Asur until the cities of the fortress.’ (Biblia romanceada judio cristiana, 14th c.)

2.1.2 Variation in argumental prepositional selection The previous section proves that prepositions could combine with nouns in historical Spanish. A different issue is whether those prepositions were rigorously fixed for every single selecting predicate. As Cano (1977–78: 336) clearly shows, the answer is negative, as one of the defining characteristics of the prepositional selection in medieval Spanish is actually its instability. There are two types of variation; firstly, when a single category may select for different prepositions without a (total) semantic difference and, secondly, when some governing categories may take either a prepositional object or a direct object. Observe the following examples of the first kind (Cano 1977–78: 349): (4) Pensar de = pensar en (lit. ‘to think of = to think in’) a.

Et quanto más pensé en la rreligión And as more thought.1SG in the religion ‘And the more I thought of religion.’ (Calila e Dimna, 13th c.)

b.

Mas de mi amor pensat But of my love think ‘But think of my love.’

(Razón de amor, 13th c.)

The second variable scenario is reflected in the following examples, where the presence or absence of a preposition does not seem to be relevant: (5)

a. Si uos un Dios creedes If you a God believe ‘If you believe in a god.’ (Disputa entre un cristiano y un judío, c 1220) b. creyme deste consejo & vete de aqui believe-me of-this advice & go-you of here ‘Believe my advice and leave.’ (Conquista de Ultramar, 13th c.) Cano (1977–78: 345)

66

Historical Spanish

c.

Creo en dios believe.1SG in god ‘I believe in god.’ (Libro de las tres creencias, Alfonso de Valladolid, 14th c.)

The variation between prepositional and non-prepositional selection is not restricted to verbal paradigms, since it is also found with nouns: (6)

a. si ellos parassen mientes la palabra que dice nuestro Señor If they stopped minds the word that says our lord ‘If they paid attention to the words of our Lord’. (Libro de los gatos, 15th c.)

Tarr (1922: 116)

b. Parad mientes en vos señor Stop minds in you sir ‘Pay attention to yourself, sir.’ (Crónica del Cid, 15th c.) Parar mientes requires a preposition but its absence is also attested. In addition, Tarr (122: 57–58, fn 2) adds that acordarse (‘to remember’) could be used with a direct object until the classical period. This same variation survives after the medieval period (Cano 1984: 238): (7)

a. Acuérdate lo que debes Remember-you the what must.2SG ‘Remember what you must.’ (Viaje, Lope, 17th c.) b. ¿Acordáisos de aquellos tiempos pasados . . . ? Remember-you of those times past ‘Do you remember those past times . . . ?’ (Lozana andaluza, Delicado, 16th c.)

The oscillation between the prepositional and the direct selection is still attested in the 16th century and beyond. The same can be said of non-prepositional complements of nouns, reanalyzed as direct objects, as the morphology of the pronouns suggests: (8) a.

como aquel que lo había gana Like that who it had cravings ‘Like that one who was longing for it.’ (Lazarillo, 16th c.) Cano (1984: 211)

Prepositions and clauses in historical Spanish

b.

67

Los avia gran piadad Them had.3SG great mercy ‘He had great mercy for them.’ (Vida, Santa Teresa, 16th c.) Cano (1984: 212)

2.1.3 Conclusions Functional and lexical prepositions could combine with nominal phrases, including full noun phrases and pronouns. There was some variation, and the same verb could frequently function with more than one required (lexical or functional) preposition. The lack of preposition shows that nominal phrases could – but crucially did not have to – be licensed via a preposition. They could be licensed positionally (Bosque and Gutiérrez 2009: 159). Therefore, prepositional verbs seem to be able to license their objects without the otherwise expected preposition. The same is possible with some nouns. Notice that this is not to say that nouns did generally not require prepositions to select for their objects in Spanish, only that, as happens with verbs, some nouns showed variation.

2.2 Prepositions and infinitives 2.2.1 Argumental and adjunct prepositional infinitives As happened with nouns, throughout the history we find different prepositions selecting for infinitives, both in argumental and adjunct positions. Let us start by examining some examples of prepositional argumental infinitives (Beardsley 1921: 245): (9)

a. Loth ovo miedo de seer en Segor Lot had fear of be.INF in Segor ‘Lot was scared of being in Segor.’ (La fazienda de Ultra Mar, Almerich, 13th c.) b. non dubdas en meterte a periglo de muerte not doubt.2SG in enter-you to danger of death ‘You do not hesitate in putting yourself in danger of death.’ (Primera Crónica General, Alfonso X, 13th c.)

68

Historical Spanish

Prepositional infinitives could also function as adjuncts in adverbial clauses. As expected, these are PPs introduced by lexical prepositions (Schulte 2003: 380): (10)

a. el rey caualgo con el a escorrirle fuera de la uilla the king rode with him to escort.INF-him out of the city ‘The king rode together with him to escort him out of town.’ (Primera Crónica General, Alfonso X, 13th c.) b. pora yrse pora alla for go.INF-refl toward there ‘In order to go there.’ (Primera Crónica General, Alfonso X, 13th c.)

Different types of prepositions – and/or prepositional complementizers, as appropriate – could combine with infinitives, as with nouns and pronouns. The parallelism is not perfect because prepositions such as desde (‘since, from’) and según (‘according to’) are never attested with infinitives, despite accepting other complements such as nouns (and even finite clauses in the present day) (Barra 2002: 241; Schulte 2004: 115, 2007: 518). Schulte (2003, 2004, 2007) shows that the range of prepositional infinitives was limited in the early texts. While a, de, en, and con are attested already in the 13th century in adverbial clauses, others will not combine with infinitives until later (Schulte 2004: 114, 2007: 522). In general, prepositional infinitives become more frequent around the 16th century (Schulte 2004: 106, 2007: 525), coinciding with an increase in the use of hypotactic finite and infinitival clauses (Schulte 2004: 12). A particular issue of interest has to do with the emergence of argumental prepositional infinitives.1 Schulte (2003, 2004, 2007) indicates that argumental infinitives (in Spanish and other Romance languages) result from the reanalysis of previously adjunct infinitives with clear adverbial interpretations, later favored by analogy (Schulte 2003: 377). Such could have been the case of, for instance, alegrarse de hacer algo (‘to be happy about doing something’), where a causal reading is still inferable today2 (Schulte 2004: 186–188). Other examples 1 As is well-known, with time certain PP-verbs reanalyze – and morphologize as units – as temporal-aspectual functional verbs (see Yllera 1980), projecting as Aspect or Tense phrases in a simple clause, similar to English modals. 2 This change helps understand why certain PPs even to this day retain certain levels of adverbial interpretation, as in the case of the causal connotations still present in the PP argument of alegrarse (‘to be happy’). It also explains why certain linguists, such as Horno (2002), argue that those PPs are not part of the conceptual structure of the selecting verbs but are adjuncts later reanalyzed as .

Prepositions and clauses in historical Spanish

69

are more dubious, however. For instance, acostumbrar a hacer algo (‘to be used to doing something’) does not lend itself to a quite straightforward adverbial interpretation of some sort, yet it is present in the old language.

2.2.2 Variation with prepositional infinitives The similarity between

and

remains when it comes to variation (Beardsley 1921; Cano 1977–78: 363; Schulte 2004: 190). Excellent evidence of the asystematicity of these constructions is the following example, where the same verb saber (‘to know’) combines with an infinitive both with a preposition and without one, all in one sentence (Cano 1977–78: 364): (11) omnes de corte que ssabían bien de trovar e cantar, men of court that knew well of compose.INF and sing.INF, e de joglares que ssopiesen bien tocar estrumentos and of minstrels that knew well play.INF instruments ‘Courtiers that knew how to compose and sing, and of minstrels that knew how to play instruments well.’ (Setenario, Alfonso X, 13th c.) In general, linguists note an “excess” of prepositions with infinitives in the old language (Herrero 2005: 89; Granvik 2012: 364–381, for ), some of which qualify as prepositional complementizers. In certain cases, one same verb may show different selectional properties, even in the same text. Take for instance the variability of temer (‘to fear’): (12) a.

Non los temades Not them fear ‘Don’t be afraid of them.’

b.

(Biblia ladinada I-i-3, c. 1400)

non temades seruir alos caldeos not fear serve.INF to-the Chaldeans ‘Don’t be afraid of serving the Chaldeans.’ (Biblia Escorial I.j.8, a 1300)

c.

non vos quebrantedes nin temades dellos not you break nor fear of-them ‘Do not break down or be afraid of them.’ (Biblia ladinada I-i-3, c. 1400)

70

Historical Spanish

d.

non temades de seruir alos Caldeos not fear of serve.INF to-the Chaldeans ‘Don’t be afraid of serving the Chaldeans.’ (Biblia ladinada I-i-3, c. 1400)

(12b) and (12d) contain the same sentence but from two different Bibles with two different syntactic representations, which clearly evidences the unpredictability of the presence or absence of the preposition. Beardsley (1921) and Cano (1977–78) offer comprehensive and data-rich descriptions of the use of prepositional infinitives in historical Spanish. The variation found in old texts can be classified in three groups, with individual categories – usually verbs – frequently qualifying for more than one type: (13) Types of variation in prepositional infinitives: a.

Optional use of preposition

b.

Use of more than one preposition

c.

Use of prepositional infinitives but direct selection of DPs

The first group consists of those predicates which may optionally take a preposition/prepositional complementizer when selecting for an infinitive (Beardsley 1921: 43–54, 73–84; Cano 1977–78: 367, 1984: 242). Consider the following examples with the verb usar (‘to usually do something’), which combines with no preposition in (14a), and with prepositions in (14b,c) (Cano 1977–78: 367): (14)

a. vsan caçar conellos liebres & perdices use.3PL hunt.INF with-them hares & partridges ‘They normally hunt hares and partridges with them.’ (Libro de la Caza, Don Juan Manuel, 14th c.) b. Los que agora vsan de caçar conellos The who now use of hunt.INF with-them ‘Those who now hunt with them.’ (Libro de la Caza, Don Juan Manuel, 14th c.) c.

falcones con que los omnes vsan acaçar falcons with which the men use to-hunt.INF ‘Falcons with which men normally hunt.’ (Libro de la Caza, Don Juan Manuel, 14th c.)

Prepositions and clauses in historical Spanish

71

This variation reaches past the Middle Ages (Cano 1984: 212): (15)

a. Nos amenazan quitarlos de los brazos de sus madres Us warn.3PL remove.INF-them of the arms of their mothers ‘They warn us with taking them off their mothers’ arms.’ (Guerra de Granada, Hurtado de Mendoza, 16th c.) b. Amenazava de venir sobre las tierras del Papa Warned.3SG of come.INF over the lands of-the Pope ‘He warned about taking over the Pope’s lands.’ (Diálogo de las cosas acaecidas en Roma, Alfonso de Valdés, 16th c.)

Prepositional optionality affects even some pronominal verbs, which are sporadically attested without their required preposition (Cano 1984: 246): (16) a.

no se espante V.S. el Marqués haber dado 500 ducados not refl scare you the marquis have.INF given 500 ducats ‘Sir, do not be horrified about the fact that the marquis has given 500 ducats.’ (Libro de la vida y costumbres, Enríquez de Guzmán, 16th c.)

b.

me maravillo hazer tal cosa me amaze.1SG do.INF such thing ‘I am amazed at doing such thing.’ (Teatro, Lope de Rueda, 16th c.)

c.

me acuerdo aver visto uno me recall.1SG have.INF seen one ‘I recall seeing one.’ (Diálogo de la lengua, Juan de Valdés, 16th c.)

The second type involves categories selecting for more than one preposition. As (14b,c) prove, categories in this group may also qualify for the kind of variation referred to in the first group. Beardsley (1921) demonstrates that membership in this group is rather common in historical Spanish (see also Cano 1977– 78: 341). A good example of this is començar (‘to start’), which usually takes de but may be found with a, even in the same sentence (Beardsley 1921: 111): (17)

los començaron a cometer et de los ferir them started.3PL to attack.INF and of them wound.INF ‘They started to attack them and to wound them.’ (Primera Crónica General, Alfonso X, 13th c.)

72

Historical Spanish

The third type of variation consists of those categories which may select prepositional infinitives and non-prepositional DP objects, typical prepositional complementizer cases, keeping in mind that, given the asystematicity of these configurations, non-prepositional infinitives may also be attested (see Cano 1984: 246–249): (18)

a. Nenguno non prueue de engañar a las gentes Nobody not try of deceive.INF to the people ‘No one try to deceive the people.’ (Fuero Juzgo, 13th c.) Cano (1977–78: 368) b. ¿Por que fazes de tardar anuestro companero? For what do.2SG of be-late.INF to-our mate? ‘Why are you making our mate be late?’

(Sendebar, 13th c.) Cano (1977–78: 369)

This syntactic property is maintained as well after the medieval period: (19) a.

te juro de casarme contigo you swear.1SG of marry.INF-me with-you ‘I swear I’ll marry you.’ (Comedia Seraphina, 16th c.) Cano (1984: 251)

b.

Yo había aceptado de me ir con ellos I had accepted of me go.INF with them ‘I had agreed to go with them.’ (Cartas, Cortés, 16th c.) Cano (1984: 252)

As indicated above, non-prepositional infinitives are not precluded with the same selecting verbs, as can be seen in this particular case with the same author in the same book (Cano 1984: 251): (20)

a.

que me consintáis de usar destos vocablos that me allow.2PL of use.INF of-these words ‘That you allow me to use these words.’ (Diálogo de la lengua, Juan de Valdés, 16th c.)

b.

porque me consintáis ser escasso en las demás for-that me allow.2PL be.INF short in the rest ‘So you allow me to be short with the rest.’ (Diálogo de la lengua, Juan de Valdés, 16th c.)

Prepositions and clauses in historical Spanish

73

It is important to note the existence of infinitives introduced by prepositional complementizers in the traditional subject position (Cano 1977–78: 338): (21) a.

Estonçe nos conuiene de fazer lo luego Then us is-better of do.INF it now ‘Therefore we’d better do it right away.’ (Poridat de poridades, 13th c.)

b.

Ffaçil es de lo dezir Y aun de fazer Easy is of it say.INF and even of do.INF ‘It is easy to say it and even easier to do it.’ (Bias contra Fortuna, Santillana, 15th c.)

Likewise, such infinitives in the traditional subject position are still attested after the medieval period (Cano 1984: 253): (22)

a. se le antojase de hacerse pastor refl him felt-like of do.INF-refl shepherd ‘He were to feel like becoming a shepherd.’ (Quijote, Cervantes, 17th c.) b. ni a vos se os irá de la memoria de hablar a nor to you refl you will-go.3SG of the memory of speak.INF to esa señora this lady ‘Nor will you forget about speaking with this lady.’ (Lozana andaluza, Delicado, 16th c.)

Infinitives of this group may refer back to a direct object pronoun (Beardsley 1921: 261–264), additional support for their complementizer nature. Consider the following examples: (23)

a. Yo non lo merezria de seer tan honrrada I not it would-deserve of be.INF so honored ‘I wouldn’t deserve to be so honored.’ (Vida de Sancta Oria, Berceo, 13th c.)

Beardsley (1921: 261)

b. qual serie lo meior: de yr a los moros o atender- los which would-be the best: of go.INF to the moors or wait-for them ‘Which one would be the best choice: to go to the moors or to wait for them.’ (Primera Crónica General, Alfonso X, 13th c.) Beardsley (1921: 262)

74

Historical Spanish

Unsurprisingly, bare infinitives may appear in the same extraposed configuration, as the following example proves (Beardsley 1921: 262): (24) lo que era peor de todo, uenir los dAffrica a Roma the what was worse of all, come.INF the of-Africa to Rome ‘What was worst of all, the Africans coming to Rome.’ (Primera Crónica General, Alfonso X, 13th c.) Infinitives introduced by a prepositional complementizer can themselves be substituted by a direct object pronoun, as expected, as the following 16th century fragment, extracted from the online Biblioteca Virtual Miguel de Cervantes, shows (see Cano 1984: 251): (25)

VALDÉS.- Esta cosa . . . , yo os prometo de hacerla. This thing I you promise of do.INF-it MARCIO.- . . . Una cosa me queda que demandar; prometedme One thing me rests that request.INF promise-me todos de no negármela, . . . all of not deny.INF-me-it VALDÉS.- Yo por mi parte lo prometo, . . . I for my part it promise PACHECO.- También yo lo prometo por la mía. Too I it promise for the mine ‘VALDÉS.- This thing . . . I promise you to do it MARCIO.-

. . . I have one more last request; promise me, all of you, not to deny it to me.

VALDÉS.-

For my part, I promise. . .

PACHECO.- For my part, I too promise. . .’ (Diálogo de la lengua, Juan de Valdés, 16th c.) The verb prometer (‘to promise’) takes a direct object lo (‘it’); this lo is substituting for de hacerla / de no negármela (lit. ‘of do.INF-it’, ‘of not deny.INF-me-it’). For Cano (1984: 248) these extra meaningless de with infinitives are one of the most typical alternations in Classical Spanish.

Prepositions and clauses in historical Spanish

75

2.2.3 Summary and syntactic conclusions The mixed situation of prepositional and non-prepositional infinitives raises the following theoretical questions. Infinitival adverbial clauses are very straightforward, as they are common PPs introduced by lexical prepositions. The situation with argumental infinitival clauses is more complicated. The fact that the infinitives could be substituted by direct object pronouns permits to conclude that the prepositions in examples (21–23, 25) are not actual prepositions, but rather prepositional complementizers. Additionally, the fact that even required prepositions, such as de with maravillarse (‘to marvel at’) and other verbs, could nevertheless be optionally left out suggests that their syntactic role was minimal, suggesting that they were functional prepositions or K(ase)s, while not fitting Rizzi’s classification for prepositional complementizer. Moreover, the negligible semantic import of most of these prepositions – apart from certain aspectual differences such as pensar (‘to think’) vs. pensar en (‘to think of’) (Serradilla 2001: 146; see also Demonte 1992 for present-day Spanish) – is further supported by those particular cases where more than one preposition could be selected by a particular category.3 The variation sampled in the previous section is the basis of Cano’s (1981) concept of prepositional transitivity. Whether it is transitivity or not, what is crucial is the fact that several verbs could variably license both a

(beyond just prepositional complementizers) or a bare infinitive as their complement suggests that both alternatives must have been equally licensed as internal arguments (see also López 1970: 177; Arjona 1978: 71–72, 89). According to the theory of Case presented in the previous chapter (section 3.1.), lexical prepositions would check lexical Case, while functional prepositions would be the manifestation of inherent Case (Kase). An important situation is created when trying to account for the non-prepositional objects of categories normally requiring a preposition. Let us focus on the following examples in particular: (26)

a. Me acuerdo ø aver visto una cosa (= 16c) b. Me maravillo ø hazer tal cosa (= 16b)

3 While functional prepositions are typically semantically empty, it is not a necessary condition; the fact that en can double up as aspectual value and add some meaning in pensar en does not deny its functional status, as aspectual markers, modals, auxiliary verbs, etc. are clearly semantically loaded functional categories (Cinque 2010: 11).

76

Historical Spanish

These must also be objects and argumental, so according to the theory Case must be checked. In principle, it should be some type of Accusative structural Case since they look like some type of direct object. However, acordarse and maravillarse are frequently prepositional and do not pronominalize with direct object pronouns. What is more, medieval Spanish had prepositional proforms equivalent to French and Italian y/ci, en/ne, namely hi/i/y and ende/end/ ent (Badia 1947), with which prepositional verbs such as maravillarse pronominalized: (27)

marabillóse ende mucho amazed.3G-refl of-it much ‘He was very amazed at it.’ (Libro de los estados, Don Juan Manuel, 14th c.)

A possible solution is self-licensed inherent Case, adduced in the literature to cover those constructions where there is no apparent Case-assigner, thus allowing the relevant nominal complements to be positionally self-licensed, as Bosque and Gutiérrez-Rexach (2009: 159) suggest, or in agreement with Ledgeway’s 2000 analysis of non-prepositional finite clauses in Neapolitan, or Barra’s 2002: 54–55 for Latin clauses.

2.2.4 The nominality of the infinitive Infinitives in historical Spanish display a typically nominal distribution, appearing as subject, object and prepositional object, as several of the examples in the previous sections have attested. The fact that they can be prepositional is generally seen as “a clear shift towards nominality, as the prototypical prepositional structure is undoubtedly [preposition + noun]” (Lapesa 2000: 547, 631–632; Schulte 2004: 98; Beardsley 1921; Tarr 1922; Raposo 1987b, probably the standard analysis regarding the nominality of Romance infinitives, via the morpheme –r). Consider the following example of coordination between PPs, one with a noun and the second one with an infinitival clause (Beardsley 1921: 2): (28)

sobrepuiaua [en piedad] et [en dar elmosnas] excelled.3SG in piety and in give.INF alms ‘He excelled at piety and at giving alms.’ (Primera Crónica General, Alfonso X, 13th c.)

Prepositions and clauses in historical Spanish

77

Examples where non-prepositional infinitives are coordinated with nouns also abound (Beardsley 1921: 2): (29)

a. demandaron [fabla] et [traer pleytesia con el rey] requested.3PL speech and bring.INF respect with the king ‘They requested information and to bring respect for the king.’ (Primera Crónica General, Alfonso X, 13th c.) b. prometiéndoles . . . [mucho bien] et [mucha merçed], et promising-them much good and much mercy and [meiorarles los fueros], et [baxarles los pechos] better.INF-them the privileges and lower.INF-them the pays ‘Promising them much good and mercy, and to better their privileges and to lower their payments.’ (Primera Crónica General, Alfonso X, 13th c.)

Having remarked the nominality of the infinitive in distributional terms, it is important to indicate that infinitives could additionally be introduced by articles and participate in clausal nominalization/substantivization (Lapesa 2000: 515– 668; Torres 2009). Such infinitives with article could also be introduced by a variety of prepositions (30a,b): (30)

a.

La hora del pedir The hour of-the request.INF ‘The time of requesting.’

(Apolonio, 13th c.)

Lapesa (2000: 530)

b. sobre el partir de las tierras y el poner de los mojones about the divide.INF of the lands & the put.INF of the markers ‘About the division of the lands and the placement of the markers.’ (Romancero) Lapesa (2000: 560) c. el muncho comer sobejo the much eat.INF leftovers ‘Much eating leftovers.’ (Tratado médico, 15th c.) d.

el haber echado el Sol en aquella isla sus primeros rayos the have.INF thorwn the sun in that island its first rays ‘The Sun having laid its first rays on that island.’ (Comentarios reales, El Inca Garcilaso, 16th c.)

78

Historical Spanish

Determiners other than the article could introduce the infinitival clause: (31) a.

Sobreste crescer et minguar dell imperio About-this grow.INF and shrink.INF of-the empire ‘About the expansion and fall of the Empire.’ (Primera Crónica General, Alfonso X, 13th c.)

b.

Lapesa (2000: 560)

con aquel gozarla los ojos with that enjoy.INF-her the eyes ‘With the eyes enjoying her.’ (Los quince discursos, Morales, 16th c.) Keniston (1937: 545)

c.

crece mi firme seguiros grows my firm follow.INF-you ‘My firmly following you grows.’ (Cancionero, Juan del Encina, 15th c.)

Notice the following example, where an infinitive introduced by an article is coordinated with another infinitive without an article (Beardsley 1921: 12): (32)

fazie muchas epístolas del aorar de los ídolos et did.3SG many letters of-the honor.INF of the idols & de poner en los templos ell altar del uençimiento of put.INF in the temples the altar of-the victory ‘He wrote may letters about honoring the idols and about putting the altar of victory in the temples.’ (Primera Crónica General, Alfonso X, 13th c.)

Beardsley (1921: 2) notes that, despite being nominal-like, infinitives can still take objects, as (29) illustrates, and describes them as mixed categories, as has been traditional (see also Schulte 2004: 99–104). These differences can be easily captured in syntax by following Kornfilt and Whitman’s (2011) typological classification of determiners and clauses. In keeping with Kornfilt and Whitman (2011), infinitives can participate in different levels of nominalization/substantivization, depending on whether the infinitive would admit further nominal modification (such as an adjective, in the case of vP nominalizations) or not (TP and CP nominalizations). (30d) contains a clear postponed Nominative subject, therefore they must remain verbal high enough in the tree so as to allocate for a subject, which points to at least a TP nominalization. Given that there are no additional nominal modifiers forcing a lowering of the nominalization point and that, unlike the

Prepositions and clauses in historical Spanish

79

Turkish example Kornfilt and Whitman (2011) use to illustrate TP nominalizations, these infinitives must be verbal enough to be able to license a Nominative subject. While Mensching (2000) argues that postponed subjects remain in Spec,VP, Mensching (2000) and Sitaridou (2009) show that historical Spanish infinitives could have overt Nominative subjects in preverbal position, which could not be argued to remain in Spec,VP in any case. Mensching (2000: 106) argues that preverbal subjects are licensed in AgrP, that is, above TP, indirectly adding support for a level beyond TP (e.g., CP level) in infinitival clauses in historical Spanish. If we consider that nominalization is a dynamic derivational process and thus not a categorical process that must apply equally to every infinitive, (30d) could be regarded as a case of CP nominalization. In any case, in this particular cases the infinitives remain verbal high enough in the derivation so as to allow for internal arguments (direct objects, thus vP is present) and an external argument (the postponed subjects), as the trees in (33) visualize. Notice too that, while in (33a) (=31b) there is no evidence of aspectual/ modal or auxiliary verbs, (33b) (=30d) does include an auxiliary verb, which clearly justifies a nominalization point necessarily above TP: (33)

a. aquel gozarla los ojos

b. el haber echado el Sol en aquella isla sus primeros rayos

As expected, there are other nominalized infinitives which clearly differ from the previous ones and qualify as vP nominalizations. Such is the case of (30c) as there is a direct object licensed by vP and a quantifier modifying the verb: (34) el muncho comer sobejo

80

Historical Spanish

Finally, other infinitives – the more nominal-like – qualify as VP nominalizations, thus lacking many verbal properties of the verb such as internal arguments, etc., indicating that the nominal projection starts right on top of the verb. English complex event nominalization, such as Ruth’s frequent reading of Anna Karenina, are a clear example. The infinitive el aorar de los ídolos in (32) is another clear example. Its syntactic representation would be as follows: (35)

el aorar de los ídolos

Another case of this is the infinitive in (30a), el pedir. Beardsley (1921) lists many more infinitives of this type; they even admit plural inflection, as in aver (lit. ‘the having’, ‘possession’) and los averes (lit. ‘the havings’, ‘possessions’), or cantar (‘the singing’) and los cantares (‘the singings’) (Beardsley 1921: 6). These infinitives do not even project a thematic grid and thus have no arguments. In short, infinitival clauses were distributionally nominal in historical Spanish. Infinitives could additionally participate in nominalization/substantivizations at different levels, which also helps capture and formalize the traditional idea of their mixed nature. An interesting remaining issue is Lapesa’s (2000: 543) report that interrogative infinitives did not admit the article, in spite of the fact that they could be direct objects of a transitive verb, a nominal context, which further shows that nominality does not necessarily go hand in hand with the combination with articles.

2.3 Prepositions and finite clauses The previous sections have shown that there is no stability in the presence of the preposition, except in the adverbial cases, which seem unproblematic. Several

Prepositions and clauses in historical Spanish

81

examples revealed that noun phrases and infinitives could indeed be positionally licensed without the materialization of an inherent preposition. This section is devoted to exploring finite clauses (i.e., argumental que-clauses, indirect interrogatives, and adverbial clauses), with the purpose of investigating their combinatorial properties, their possible nominal nature, and how they compare to (non-clausal) nouns and infinitives. The notions of argumenthood vis-à-vis prepositionality and Case will fall under special scrutiny.

2.3.1 Previous studies based on syntactic analogy Herman (1963) is probably the principal traditional authority on the formation of so-called subordinating conjunctions in Romance – in current theoretical terms,

– and an advocate of the analogical account. His core hypothesis is based on the idea that the Romance languages conserved only a reduced number of conjunctions – complementizers – from Latin and consequently each language needed to create their own new sets following some specific analogical frames. This creative process is not only a Romance innovation; rather, it is already on-going in the Latin period, as Latin had already developed some finite subordination by using complementizers such as quod (originally a relative pronoun) or quia, a pattern later inherited by the Romance languages. Of all the Latin complementizers, only quod widens its use in Post-Classical Latin by becoming a multipurpose connector, a role later played by its Romance descendent que/che. The analogical frames leading to new complementizers in Post-Classical Latin and into the Romance languages are varied; three are of extreme interest for this research, especially the third type (Herman 1963: 74–101): (36)

a.

: causal eo quod (lit. ‘it that’)

b.

: causal pro eo quod (lit. ‘for it that’) c.

: pro quod (lit. ‘for that’)

As Herman (1963: 76) indicates, it is characteristic of the Latin syntax to find a quod clause referring back to a pronoun in the main clause, as in hoc dico quod (lit. ‘this I say, that. . .’). This syntactic phenomenon lies at the origin of the new conjunction eo quod (36a), initially a quod-clause referring back to an independent ablative pronoun, and from the Post-Classical period on, a fixed complementizer similar to quod (Herman 1963: 76).

82

Historical Spanish

The second creative process involves a preposition, with two possible paths: (36b) includes an intervening pronoun, while (36c) does not. Therefore, the configuration

is already present in Latin. This tendency to combine quod with another element increases, to the point that, as Herman (1963: 108) indicates, combinations of a first element (for instance, a preposition) plus quod are already in the rise within the Latin period, at the expense of quod functioning alone. The previous panorama will be echoed during the Proto-Romance period. In the period of the 6th–9th centuries, Herman notes the presence of que/quod in all the diverse linguistic offspring of Latin, including, of course, the prepositional groups. Consider the following Romance complementizers equivalent to English because, both with and without intervening pronouns (Herman 1963: 180–198): (37)

a. Spanish and Portuguese: a. por que (lit. ‘for that’) b. por esso que, por esto que (lit. ‘for that/this that’) b. French: a. por ço que, por ce que (lit. ‘for this that’) b. por que (lit. ‘for that’) c. Italian: a. perchè (lit. ‘for that’) b. per ciò che (lit. ‘for that that’)

Other such complementizers introduced by a preposition include Spanish posque (lit. ‘after that’), Italian poi che (lit. ‘after that’), French puisque (lit. ‘after that’), etc. According to Herman, all of these go back to Late Latin. Other conjunctions were created during the Romance period as innovations in each separate language, such as Spanish ata/fasta/fata que (lit. ‘until that’), de que (lit. ‘since that’); Italian acciò che (lit. ‘to this that’, so that); French à ce que (lit. ‘to this that’), de ce que (lit. ‘of this that’), etc. Herman’s work highlights the existence of several active analogical models producing new conjunctions from the Latin period on. Tarr (1922) studies the emergence of prepositional finite clauses in Spanish with particular attention to argumental clauses. Tarr also, assumes an analogical account and a regularization process, via the extension of the preposition used with nouns and infinitives to the finite clause (Tarr 1922: 253–254). Other structures present in the old language also serve as models for the new construction:

Prepositions and clauses in historical Spanish

83

1. Prepositional interrogative clauses and relative clauses (Serradilla 1995: 159): (38) se maravillaron mucho de [cuán pequeño hombre era Pedro refl marveled.3PL much of how little man was Pedro el Ermitaño] the Hermit ‘They were amazed at how little a man was Pedro the Hermit.’ (Gran Conquista de Ultramar, 13th c.) 2. The configuration

(Tarr 1922: 110): (39)

Los castellanos, pesándoles mucho de lo que la reyna The Castilians, aggrieving-them much of it that the queen donna Vrraca su sennora diera las fortalezas et los Mrs Urraca their lady gave the fortresses and the castiellos al rey de Aragon. . . castles to-the king of Aragón ‘The Castilians, being very aggrieved that queen doña Urraca their lady had given the fortresses and castles to the king of Aragón.’ (Primera Crónica General, Alfonso X, 13th c)

This pronominal construction is attested not only with functional prepositions such as de, a, en, but also in some prepositional finite clauses introduced by lexical prepositions, such as por aquello que, despues desto que, sin lo que, etc., as can be seen in the following example (Tarr 1922: 110–111): (40)

Et esa noble reina donna Berenguella muy alegre por aquello And that noble queen Mrs Berenguella very happy for that que su fijo el rey don Fernando auie conquerida la çipdat that her son the king Mr Fernando had conquered the city de Cordoua . . . fizo sus gracias a Dios of Córdoba made her thanks to God ‘And that noble queen Berenguella, very happy because her son Fernando had conquered the city of Córdoba, thanked God.’ (Primera Crónica General, Alfonso X, 13th c)

3. Prolepsis. In the proleptic construction the subject of the subordinate clause is anticipated in the main clause, where it is introduced by a preposition and

84

Historical Spanish

connected to the verb in the main clause, as can be seen in the next example (Tarr 1922: 120): (41)

me plaze de myo Çid que fizo tal ganancia me pleases of my Cid that did such victory ‘It pleases me that Myo Çid won.’ (Cid, 13th c.)

However, Tarr questions the actual relevance of the prolepsis, since it is only frequent with de, and it cannot account for the history of the majority of the prepositions. 4. The configuration

,4 where commo (‘how’) seems to work like que (Tarr 1922: 124): (42) membrandoseles de como son complidos cristianos remembering-refl-them of how are.3PL faithful Christians ‘Remembering that they are faithful Christians.’ (Siete Partidas, Alfonso X, 13th c.) Tarr (1922: 114–116) remarks that what is at stake is the nominal properties of the finite clause. The finite clause is latest to appear in nominal contexts where nouns and infinitives already could before. Bogard and Company (1989) study the introduction of the preposition de in the pattern , as evidenced in the following contrasting examples (Bogard and Company 1989: 258): (43) a.

Tengo miedo de que vengas have.1SG fear of that come.2SG ‘I am scared that you will come.’

b.

Tengo miedo ø que vengas have.1SG fear that come.2SG ‘I am scared that you will come.’

4 The medieval examples of de commo and en commo (‘of how’, ‘in how’) will not be studied here, since they seem to be fossilized forms (see Serradilla 1997: 247; and Herrero 2005: 143). In any case, they may be regarded as precursors of

, as Tarr (1922) and others have indicated. The use of commo as declarative complementizer – independently of the presence of not of a preposition – specialized for presupposed contexts, has been analyzed in Iglesias (2002).

Prepositions and clauses in historical Spanish

85

The historical appearance of de in such patterns is considered to be caused by analogy with the pattern , as in miedo de ir (‘fear of going’). This extension is supposed to eliminate a possible confusion between complement clauses and relative clauses (Bogard and Company 1989: 260). However, that hypothesis raises some questions. Serradilla (1995: 158) shows that the presence of the preposition does not always eliminate ambiguity between a complement and a relative clause, as the following example proves: (44)

Sepades que a nos es fecha rrelacion en que vosotros andays de Know that tous is done story in that you walk from lugar en lugar muchos tiempos e annos place in place many times and years ‘They inform us that you have been wandering around for a long time.’ (Documentos Villa III, 15th c.)

Here en que could be a prepositional relative (“letter in which. . .”) or the complement of the noun (“news that. . .”). In addition, if ambiguity were a problem in the past, Spanish should simply not have retained complement and relative clauses of the form in order to avoid further cases of ambiguity, but this is contrary to fact. Serradilla (1995) shares with Bogard and Company the idea that analogy is the key factor. The possible factors that may have favored the extension of the prepositional construction are listed below (see also Serradilla 1997: 236–260): 1. The existence of prepositional adverbial clauses, particularly those introduced by the same prepositions that introduce argumental clauses, as can be seen in the following example from the Corpus del Español: (45)

los procuradores del Rey fazen los yr alla a que the representatives of-the King make them go.INF there to that se fagan fijos dalgo refl make sons of-something ‘The King’s representatives make them go there so that they become noblemen.’ (Libro de los fueros de Castilla, Seudo-Fernando III, 13th c.)

2. Prepositional interrogative clauses, which seem to have entered the language before. However, this is also a late phenomenon, and it is not the case that prepositional indirect interrogatives were very common long before the emergence of argumental finite clauses with que.

86

Historical Spanish

3. Other configurations already listed by Tarr (1922), such as

,

, and the proleptic construction. The idea of analogy and the regularization process is reintroduced in Serradilla (1996b), but pointing out that the origin of the argumental prepositional finite clauses is to be found in the configuration

, which is also argued to be on the basis of the creation of Romance prepositional infinitives. The same hypothesis is exposed in Serradilla (1997: 177). The extension of the preposition from the nominal configuration

into the infinitive is considered to be easier due to the previous existence of prepositional infinitives in Late Latin. Serradilla (1997) adds the traditional idea that que became a default subordination marker, which explains why it is not necessary to use a preposition to select for it. The preposition appears when it is relevant semantically, that is, in adverbial clauses. She argues that only after que loses its status as subordination marker will other (functional) prepositions start combining with que-clauses.

2.3.2 Previous syntactic approaches: Pronouns, nominality and Case One of the analogical models proposed by Tarr (1922) is the configuration

, where the finite clause is said to be in apposition to a meaningless pronoun. This syntactic alternative becomes the locus of change in Moreno (1985–86). In this work, Moreno studies paratactic cataphora, namely the syntactic/discourse procedure which consists of a pronoun referring to a finite clause appearing later in the sentence. Examples are found in a variety of languages, including English and Latin: (46)

a.

They expect it of you that you cooperate

Moreno (1985–86: 178)

b.

Totum in eo est tectorium ut concinnum sit All in this is plasterwork that elegant is ‘All depends on the plasterwork being elegant.’ Moreno (1985–86: 175)

This is the configuration in old Spanish

, extendible as well to French de ce que and à ce que (Moreno 1985–86: 172): (47)

por tal fago aquesto que sirvan a so señor for so do.1SG this that serve.3PL to their lord ‘I do this so they serve their lord well.’

(Cid, 13th c.)

Prepositions and clauses in historical Spanish

87

For Moreno, the change from paratactic

into hypotactic

, thanks to the loss of the intervening pronoun, might explain the creation of prepositional finite clauses in Spanish. Schematically, Moreno (1985– 86: 174) proposes the following change: (48)

porque: por + esso, que > por + Ø, que > porque

Moreno suggests that the same type of change for argumental prepositional finite clauses. The Ø would represent the empty space left by the elimination of the pronoun (Moreno 1985–86: 174): (49)

Estaba contento de Ø que el trozo fuera tan abundante Was.3SG happy of that the piece was so abundant ‘He was happy that the piece was so large.’

Notice, however, that some syntactic issues require further elaboration for this hypothesis to be fully operative. The main one has to do with the fact that the second step, which assumes the pronoun-drop before the reanalysis, produces a syntactic “disturbance”, for it requires a main clause ending with a (transitive) preposition without complement, an analysis highly problematic for Spanish, where preposition-stranding is not allowed. In addition, the syntactic status of the que-clause would remain uncertain, in that, not being the complement of the preposition yet, it would still need to refer back but there would be nothing to refer to. Notice too that the supposition of a pro would be problematic, since, in pure syntactic terms, this pro would still be a pronoun, and what Moreno is arguing for is the elimination of the pronoun complement of the preposition. Another problem for Moreno’s account has to do with chronology. Porque (‘because’) or para que (‘so that’), for instance, are attested in early Spanish texts alongside their pronominal counterparts. Moreover, in the case of French a similar hypothesis had already been rejected by Herman (1963: 188) adducing that por ce que and por que are attested contemporarily (see section 2.1.2. in chapter 5). Barra (2002), a comprehensive and thorough work on the topic, the emergence of Spanish prepositional clauses is approached from a Chomskyan generative grammar perspective. One of the key theoretical features of this study is the rejection of analogy as a mechanism of syntactic change, as it is deemed vague and unsystematic.

88

Historical Spanish

Barra criticizes the previous analogical approaches on the basis that it leaves some forms unaffected. Barra rejects Herman’s (1963) analogical hypothesis for it predicts that any preposition can combine with que and produce a new subordinating conjunction, but such rule would overgenerate unattested forms. Furthermore, analogical extension from prepositional infinitives would not explain why some prepositions such as desde (‘from’) never combine with an infinitive. For Barra it is necessary to explain why the analogical process of Spanish and Portuguese did not apply in French or Italian, where prepositional finite clauses are more restricted. A final problem has to do with chronology, as Barra wonders why analogy did not apply before it actually did, as late as the 16th century for argumental clauses for instance (but see section 2.3.4 in this chapter). Barra’s study revolves around one very precise assumption about clausal syntax in medieval Spanish (Barra 2002: 71): (50)

[V i]

[ANAPH que i . . .]

[Main clause] [anaphoric adjunct clause] Medieval Spanish finite sentential syntax is reducible to the generalization in (50), whereby all clausal subordination with que-clauses is just apposition/ adjunction, not argumental complementation in any case, as a reflection of loose syntax (Barra 2002: 86, 111). Such general adjunction configuration would apply in all of the following examples: (51)

a. [coñoscoi ] [quei a ti lo deuo agradesçer] know.1SG that to you it must.1SG thank.INF ‘I know that I have to thank you for this.’

(Alexandre, 13th c.) Barra (2002: 74)

el alma b. [Conosçi [loi ] anoche . . .] [quei me sacarie knew.1SG-it last-night that me would-take.3SG the soul oy en aqueste dia] today in this day ‘I knew it last night that he would take my soul out on this day.’ (Alexandre, 13th c.) Barra (2002: 75) omne mortal] c. [e siempre te acuerdai ] [quei eres And always you remember that are.2SG man mortal ‘And always remember that you are mortal.’ (Rimado de Palacio, 14th–15th c.)

Barra (2002: 88)

Prepositions and clauses in historical Spanish

89

All examples in (51) would show the same syntactic pattern, a subordinated clause – syntactically an adjunct – referring back to some position in the main clause via a co-referent element (a pronoun (51a) or a noun, which would be the actual arguments in the main clause) or even without it (51a,c). Accordingly, in (51a) the verb does not assign Case to the clause. In (51b) Case is assigned to a pronoun, not to the clause. The situation is said to change once que-clauses become nominal by acquiring a [+N] feature in the 16th century. Once nominal, the finite clause may receive Case (Barra 2002: 186). The former adjunction patterns disappear and the clause integrates into the main clause as its argument/complement, either as direct object or as prepositional object (Barra 2002: 92, 173). The change brings about the possibility of (argument) prepositional finite clauses. As expected, two syntactic factors are especially relevant: 1. The features of the prepositions. Only some prepositions will be able to select for a finite clause as their complement (Barra 2002: 92 ff.) To meet the requirements, prepositions must be able to select for what Barra (2002: 197) calls 4-dimensional definite NPs, that is, definite noun phrases with temporal content, the same temporal content present in finite clauses (Barra 2002: 198). 2. The features of the finite clause. Finite clauses must become nominal to be prepositional objects. Barra’s study carries with it several important, far-reaching implications and problems. One of the implications is the fact that the appearance of the preposition is directly linked to the creation of argumental finite clauses in Spanish. The old adjunction-based system is substituted by proper argumental subordination only after the que-clause becomes nominal. A nominal que-clause can be not only the object of a preposition and check its Case feature, but also the direct object of a transitive verb. In addition, the fact that que-clauses could be introduced by lexical prepositions before the emergence of argumental prepositional finite clauses is a problem. Furthermore, old Spanish syntax has to account for other types of sentential complementation, such as indirect interrogatives, also finite clauses. Barra’s conclusions make clear crosslinguistic hypotheses, particularly important for sister languages such as French, Italian or Portuguese. If a language does not permit argumental prepositional finite clauses, it must mean that it does not have proper argumental subordination, as its clauses would not be nominal. A case in hand is French, which, if Barra’s hypothesis is valid, must lack finite sentential complementation, and such is the conclusion reached in Barra (2002: 399). This extension seems very risky, as Barra himself acknowledges.

90

Historical Spanish

2.3.3 Summary and conclusions of previous syntactic studies The analogical studies offer a sound hypothesis for the creation and replication of the prepositional pattern in Spanish and related languages, but have not focused so much on their syntactic explanations and inquiry into the formal structure of the language, e.g. a formal description of prepositional finite clauses. These are the questions that Barra and, in a lesser degree, Moreno focus on, but they include some succinct comparative evidence, which needs to be increased. It is imperative to examine what prepositional finite clauses are like in other related languages and locate the similarities and differences, which will help understand their syntactic nature. Moreover, the syntactic framework and related hypotheses presented before need to be tested for accuracy. In particular, several questions must be answered, especially the role of Case, especially in the creation of prepositional finite clauses, and whether a causality link between prepositionality, nominal features, and argumenthood of the finite clause can be actually established, as Barra’s study would imply.

2.3.4 Argumental prepositional que-clauses Linguists tend to draw a line about the 16th century for argumental prepositional finite clauses. Before that period, the preposition was (generally) not allowed, as the following 13th century examples illustrate, contrasting a prepositional infinitive with a non-prepositional finite clause dependent on the same noun (hueb/vos, ‘necessity’) and from the same text (Herrero 2005: 89): (52)

a. Nos huebos avemos en todo de ganar algo We needs have in all of win.INF something ‘We need to gain something in all.’

(Cid, 13th c.)

b. huevos avemos que nos dedes los marcos needs have.1PL that us give.2PL the marks ‘We need you to give us the marks.’

(Cid, 13th c.)

Notice also this early example where a non-prepositional clause depends on the noun aduenentiam (‘agreement’) (Bogard and Company 1989: 262):

Prepositions and clauses in historical Spanish

(53)

91

Ego Eolalia, facio aduenientiam cum filia mea Gometiza, I Eulalia, make agreement with daughter my Gometiza, quod ego do ei la medietat de les meas kasas del alraual that I give her the half of the my houses of-the hood ‘I Eulalia agree with my daughter Gometiza to give her half of the houses I own in the neighborhood.’ (DLE n 259, 12th c.)

Observe the following examples with PP-verbs (Barra 2002: 66): (54)

a. Demandole dixol que se marauellaua que Asked.3SG-him and told.3SG-him that refl marveled.3SG that con todos los otros tan mal se acordaua with all the others so bad refl agreed.3SG ‘He asked for him and told him that he was amazed that he had so bad relations with the rest.’ (Apolonio, 13th c.) b. tu me ayuda . . . que yo saque a Castylla del antygo dolor you me help that I take to Castile of-the ancient pain ‘Help me release Castile from the old pain.’ (Poema de Fernán González, 13th c.)

The same syntactic non-prepositional configuration applied to nouns and adjectives combining with que-clauses: (55)

a. e por esso e pauor que a esa quieras meior & for that have.1SG fear that to that love.2SG better ‘And because of that I am scared that you may love her better.’ (Razón de amor con los Denuestos del agua y el vino, 13th c.) González (1993: 77) b. Sennor, non so digno que tu entres so el mio techo Lord not am worthy that you enter under the my roof ‘Lord, I am not worthy that you enter my house.’ (Traducción del Evangelio de San Mateo, 13th c.) González (1993: 176)

The next task is to test the accuracy and explanatory power of the theoretical tools generally used, and explore the relationship between finite clauses and nouns and infinitives. Moreno (1985–86) established a diachronic correspondence

92

Historical Spanish

between early prepositional clauses with an intervening pronoun and posterior non-pronominal prepositional clauses. Barra (2002) argued for a change in the nominal features of the que-clause about the 16th century, and Case checking and argumenthood. However, both studies must face several challenges, both in terms of data (in Spanish and crosslinguistically) and theory-internal.

2.3.5 Early examples of argumental que-clauses Despite the seemingly consensus on 16th/17th centuries as the date for the emergence of argumental prepositional finite clauses, it is nonetheless true that many linguists have attested several early examples. Serradilla (1995: 149) concludes that the origins of argumental

are to be found already in the medieval period (see also Tarr 1922: 145, 254–256). Consider the following examples with nouns combining with prepositional finite clauses (Serradilla 1995: 150): (56)

a. con muy grand alegría de que auien conquistas todas with very big happiness of that had.3PL conquered all las tierras del mundo5 the lands of-the world ‘Very happy that they had conquered all the lands in the world.’ (Primera Crónica General, Alfonso X, 13th c.) b. por razón de que se fazien allí muchas serpientes for reason of that ref made.3PL there many snakes ‘Because many snakes proliferated there.’ (Primera Crónica General, Alfonso X, 13th c.)

As expected, early examples of prepositional finite clauses are attested with adjectives and verbs (Serradilla 1995: 153): (57)

a. penso en que fuesen fasta Hierusalen thought.3SG in that went.3PL until Jerusalem ‘He thought about them going to Jerusalem.’ (Gran Conquista de Ultramar, 13th c.)

5 Notice the causal connotation in this example.

Prepositions and clauses in historical Spanish

93

b. et non esperades a que vos afinque mas por ello and not wait to that you repeats more for it ‘And don’t wait for him to repeat it more.’ (Conde Lucanor, Don Juan Manuel, 14th c.) These examples document the sporadic existence of argumental prepositional finite clauses already in the medieval period.

2.3.6 Adverbial prepositional que-clauses (adjunct que-clauses) It would seem that combinations

were barred in historical Spanish, but, as Serradilla (1996b: 32) notes, non-selected prepositions (i.e., lexical prepositions introducing adverbial clauses) take que-clauses. Hence, adverbial clauses have made extensive use of the configuration

since the first attestations of the language without producing ungrammaticality. Observe the following examples: (58)

a. Por las antedichas razones et porque estas fiaduras en oluido For the foretold reasons and for-that these loans in forget non sean caydas not are fallen ‘For the above mentioned reasons and so that these loans do not fall into oblivion.’ (Fuero de Zorita de los Canes, 13th c.) Méndez (1995: 45) b. desde que salio dAffrica since that left.3SG of-Africa ‘Since he left Africa.’ (Primera Crónica General, Alfonso X, 13th c.) Méndez (1995: 131)

The immediate conclusion is that once again there was no constraint against prepositional finite clauses in old Spanish. Needless to say, the existence of these prepositional clauses has long been acknowledged in the literature. For instance, Herman (1963), who simply speaks of subordinating conjunctions and therefore does not contemplate the prepositional nature of these combinations.

94

Historical Spanish

2.3.7 Prepositional indirect interrogative finite clauses A second group of completive finite clause is the indirect interrogatives. These clauses are also finite CPs, therefore their historical behavior and syntactic properties must be examined to understand the features of finite sentential subordination. Indirect interrogative clauses appear without prepositions in historical Spanish, showing the same syntactic properties as que-clauses. The following examples include verbs, nouns, and adjectives: (59)

a. Marauillosse mucho que podrie seer Marvelled.3SG-refl much what could.3SG be.INF ‘He was very amazed at what it could be.’ (Primera Crónica General, Alfonso X, 13th c.)

Girón (1988: 164)

b. para mientes quan buena cosa es stop minds how good thing is ‘Consider what a good thing it is.’ (Castigos e documentos, Sancho IV, 13th c.)

Girón (1988: 163)

c. ruego te que pares mientes si fiz yo alguna cosa por ti beg.1SG you that stop minds if did I some thing for you ‘I beg you consider whether I did something for you.’ (Primera Crónica General, Alfonso X, 13th c.) Girón (1988: 163) d.

E por ser mas cierta si era assi And for be.INF more sure if was thus ‘And in order to be surer whether it was that way.’ (Gran Conquista de Ultramar, 13th c.)

As with que-clauses, there exist some early examples of prepositional indirect interrogatives before the 16th century (Serradilla 1995: 159; see example (38) above as well): (60)

a.

he dubdado sobre cuál haré antes have.1SG doubted about which will-do.1SG before ‘I have doubted which one I will do first.’ (Cárcel de amor, Diego de San Pedro, 15th c.)

Prepositions and clauses in historical Spanish

95

b. Fue mucho espantado de quan fuerte era Was very horrified of how strong was ‘He was horrified to see how strong he was.’ (Crónica Alfonso XI, 14th c.) As expected, examples are attested from the 16th century on as well, once again dependent on verbs, nouns, and adjectives: (61) a.

Algo estuvieron los españoles perplejos en si irían Some were the Spaniards perplexed in if would-go.3PL solos o no alone or not ‘The Spaniards were a little perplexed whether they would alone or not.’ (Segunda carta, Cortés, 16th c.)

b.

se me quito la duda de si era antojo refl me removed.3SG the doubt of whether was whim ‘I did not doubt any more whether it was a whim.’ (Vida, Santa Teresa, 16th c.) Keniston (1937: 641)

c.

yo me espanto de quan recontento bibo I me horrify of how happy live.1SG ‘I am horrified at how very happy I live.’ (Viaje del cielo, 17th c.)

To complete the picture, indirect interrogatives can be the complements of lexical prepositions introducing adverbial clauses: (62)

a. e fueron a Veralnorte, por si podrían tomar una and went.3PL to Veralnorte for if could.3PL take.INF one villa que llaman Oriola city that call.3PL Oriola ‘And they went to Veralnorte, just in case they could take a city called Oriola.’ (El Victorial, Díaz de Games, 15th c.) b. quiso hacer diligencia para si podría hallar wanted.3SG do.INF procedure for if could.3SG find.INF el cuerpo en el río the body in the river ‘He wanted to start procedures to find out whether the body was in the river.’ (Historia General del Perú, Murúa, 16th c.)

96

Historical Spanish

In conclusion, this second type of finite clause is attested combined with and without prepositions both in argumental clauses, and even with prepositions in adjunct (adverbial) clauses, in parallel with que-clauses.

2.3.8 Prepositional que-clauses and pronouns In historical Spanish prepositions could occasionally select for a pronoun to which the finite clause was arguably an adjunct, thus establishing a nonargument link between prepositions and finite clauses. Moreno (1985–86) considers it to be the origin of prepositional finite clauses. Observe the following example (Tarr 1922: 110; see (39) above): (63) pesándoles mucho de lo que la reyna . . . diera las fortalezas aggrieving-them much of it that the queen gave the fortresses ‘Being very aggrieved that queen had given the fortresses.’ (Primera Crónica General, Alfonso X, 13th c) In (63) pesando selects for de, which takes the pronoun lo as its complement. Arguably, the finite clause is an adjunct to the pronoun, “although [these constructions] do not have the full appositional force (particularly those with lo and aquello) of the separated forms de esto . . . que, etc. (cf. O. Fr. de ço . . . que) or the modern Spanish de esto: de que (de esto tengo miedo, de que, etc.)” (Tarr 1922: 109). Intervening pronouns are also attested in adverbial clauses (Moreno 1985–86), as the following example illustrates (Tarr 1922: 112): (64)

marauillauanse ende mucho por aquello que siempre le amazed.3PL-refl of-it much for that that always him uieran yazer encerrado en su alcaçar saw.3PL lie.INF locked in his fortress ‘They were amazed because they had always seen him shut up in his fortress.’ (Primera Crónica General, Alfonso X, 13th c.)

2.3.9 Prepositional que-clauses and variation The existence of argumental prepositional finite clauses did not eliminate the non-prepositional alternatives. As Cano (1985: 89–90) proves, the appearance of the prepositional alternative is not across the board and depends on each

Prepositions and clauses in historical Spanish

97

separate verb. As happened with nouns and infinitives, the available texts reflect ample optionality, which still exists in the current language6. Consider the following examples of the verb aguardar (‘to wait’), both with preposition (65a) and without it (65b). Both examples are the same author (Cano 1985: 83): (65)

a. ni él aguardó a que le respondiese nor he waited to that him answered ‘Nor did he wait to get an answer.’ (Novelas ejemplares, Cervantes, 17th c.) b. sin aguardar que Zoraida le respondiese without wait.INF that Zoraida him answered ‘Without waiting for Zoraida’s answer.’

(Quijote, Cervantes, 17th c.)

Notice that in both examples the mood in the subordinate clause remains subjunctive. The data around the 16th and 17th centuries show that both prepositional and non-prepositional alternates may anyway select for the same mood in the embedded clause. Consider for instance the following examples, where the verb quejarse (‘to complain’) takes indicative and subjunctive in the subordinate clause, regardless of the presence of a preposition and without meaning change or aspectual differences: (66)

Quejarse (de) que 1. Indicative in the embedded clause a. se quejaba que le habían quitado el cargo refl complained.3SG that him had.3PL stripped the position ‘He was complaining that they had been stripped of his position.’ ( Jornada de Omagua y Dorado, Francisco Vázquez, 16th c.) b. quejándose de que había hecho blasfemar su complaining-refl of that had.3SG made blaspheme.INF his nombre a los enemigos name to the enemies ‘Complaining that he had made the enemies blaspheme his name.’ (Predicación del Evangelio en las Indias, José de Acosta, 16th c.)

6 Not disregarding certain aspectual differences related to prepositions. See Cano (1985: 84).

98

Historical Spanish

2. Subjunctive in the embedded clause c. todos ellos se quejaban que con tales industriasel all they refl complained that with such tactics the gobernador y sus capitanes hubiesen querido entrevenir governor and his captains had wanted intervene.INF ‘All of them complained that in that way the governor and his captains had wanted to intervene.’ (Guerras civiles peruanas, Cieza de León, 16th c.) d.

quejándose de que el obispo se hubiese introducido en esta visita complaining-refl of that the bishop refl had introduced in this visit ‘Complaining that the bishop had joined this visit.’ (Política indiana, Solórzano Pereira, 17th c.)

Furthermore, certain verbs which are only attested selecting one mood in the embedded clause allow prepositional optionality. Again, the presence or absence of the preposition does not affect mood selection: (67)

Obligar (a) que (‘to force’) a. Y así el amor que a vuesa merced tengo, me obliga And so the love that to your mercy have.1SG, me forces a que diga algo to that say.1SG something ‘And therefore the love I have for you forces me to tell you something.’ (Vida del P. Baltasar Álvarez, Luis de la Puente, 16th c.) b. obligando que el tímido ganado atónito se esparza forcing that the shy cattle astonished refl scatters ‘Forcing the shy cattle to scatter in astonishment.’ (La gatomaquia, Lope, 16th c.)

While mood selection has evolved and become stricter in the current language, the previous examples illustrate that, at the time where argumental prepositional finite clauses are said to emerge, the presence or absence of those prepositions does not necessarily correlate with mood selection. If mood selection is linked to argumenthood, then we can at least suspect that the prepositional and non-prepositional examples may be argumental. A final point of interest is the presence of extra prepositions. On the one hand, certain verbs which were already peculiar in their prepositional selection

Prepositions and clauses in historical Spanish

99

of infinitives carry this peculiarity over to the finite clauses. Prometer could take a prepositional infinitive pronominalizable with lo, the typical test for prepositional complementizers. The same syntax is nevertheless attested with a finite clause. Compare the following examples: (68)

a.

Yo os prometo de hacerla I you promise of do.INF-it ‘I promise you to do it.’ (Diálogo de la lengua, Juan de Valdés, 16th c.) Cano (1984: 251)

b.

habéis me de prometer de que . . . no interromperéis have.2PL-me of promise.INF of that not will-interrupt.2PL el hilo de mi triste historia the thread of my sad story ‘You must promise me you won’t interrupt the thread of my story.’ (Quijote, Cervantes, 17th c.) Tarr (1922: 238)

Another similar example is temer (‘to fear’). Consequently, it comes as no surprise that temer accepts de que alongside bare que: (69)

aquellos que temían de que les succediesse alguna adversidad those that feared of that them happened some adversity ‘Those that were afraid of suffering some adversity.’ ( Jardín de flores curiosas, Antonio de Torquemada, 16th c.)

To be coherent with its infinitival syntax, prometer in (68b) and temer in (69) must combine with functional prepositions with a role copied from that of prepositional complementizers in their corresponding infinitival clauses (see Rizzi 1988), with the additional categorial problem stemming from the fact that they do no longer qualify as specialized infinitival complementizers. Nevertheless, that similarity is confirmed by examples where

is also substituted by a direct object pronoun: (70)

a. de que haya vencido a don Quijote de la Mancha, of that has beaten to don Quixote of the Mancha, póngolo muy en duda put.1SG it very in doubt ‘That he has beaten Don Quixote, I doubt it very much.’ (Quijote, Cervantes, 17th c.) Tarr (1922: 238)

100

Historical Spanish

b. De que seáis mujer, no me lo podéis negar Of that are.2PL woman, not me it can.2PL deny.INF ‘That you are a woman, you cannot deny it to me.’ (Las dos doncellas, Cervantes, 17th c.)

Tarr (1922: 242)

In these cases the que-clause cannot be introduced by a lexical, projecting preposition. Besides, following Barra’s (2002) explanation for related examples, it could be assumed that those examples of

are adjuncts coreferential with the pronoun lo, indicating that prepositional clauses are not necessarily arguments of the selecting predicate. 2.3.10 Finite clauses and the article Que-clauses, despite displaying a typically nominal distribution, are not attested with an article until the 16th/17th centuries. Lapesa (2000: 542–543) and Herrero (2005: 113–114) indicate that indirect interrogative clauses are first documented already in the second half of the 16th century. A search in the Corpus del Español retrieves a couple of examples with si-clauses as well. Consider the following examples: (71)

a. El cómo es esta que llaman unión, y lo que es, yo no The how is this that call.3PL union, and the what is, I not lo sé dar a entender it know.1SG give.INF to understand.INF ‘How this so-called union is and what it is, I cannot explain.’ (Vida, Santa Teresa, 16th c.) b. Y ya que avemos dicho el por qué es ‘monte’ Cristo And already that have.1PL said the for what is mount Cristo ‘And now that we have already said why it is ‘mount’ Cristo.’ (Nombres, Fray Luis de León, 16th c.) c. Trató con su dueña el cómo y cuándo sería Dealt.3SG with his woman the how and when would-be.3SG ‘He discussed with his wife how and when it would take place.’ (Alfarache, Mateo Alemán, 17th c.) Lapesa (2000: 543) d.

dejó indeciso el si el origen de os indios era de España left unsure the if the origin of the Indians was of Spain ‘He left unanswered whether the Indians were originally from Spain.’ (Tratado único y singular del origen de los indios occidentales, Diego Andrés Rocha, 17th c.)

Prepositions and clauses in historical Spanish

e.

101

no se observa el si era casado not refl observes the if was married ‘It is not noted whether he was married.’

(Quijote, Cervantes, 17th c.)

While several wh-phrases such as porqué (‘why’), cómo (‘how’) and cuándo (‘when’) themselves were and still are used as common nouns, as a sort of lower level clausal nominalization/substantivization, this option was never possible for complementizer que. In addition, porqué could already be used as a bare noun, without the article (again, nominality does not require a determiner; recall that indirect interrogative infinitival clauses did not take the article either despite their nominal distribution): (72)

a. ca aquel que tan sin por que oluido tanto bienfecho that that who so without for what forgot so-much well-done That that one who forgot so much benefit without any reason.’ (Crónica Troyana, 15th c.) b. mi viene vn cauallero que me quiere matar sin me comes a knight that me wants kill.INF without por-que E a me ansi llagado como vedes agora for-what and has me so wounded as see.2SG now ‘A knight came to me who wanted to kill me without reason and he has wounded me as you can see now.’ (Traduccción de Lanzarote del Lago, 15th c.)

Lapesa (2000) and Herrero (2005) document that the article with queclauses is attested almost a century later, becoming more frequent in the 18th century (Lapesa 2000: 544), even in uses unacceptable nowadays (Herrero 2005: 82, fn 10).7 Observe the following examples, the first of which dates back to the 16th century: (73)

a. Importúnele que no permita el que esté Vuestra Bother-him that not allows the that is your Merced en presencia de tal Magestad sin reverencia, temor, mercy in presence of such majesty without reverence, fear, y amor and love

7 For instance, with the verb saber (‘to know’), as in Infortunios de Alonso Ramírez, Carlos de Sigüenza y Góngora (1640).

102

Historical Spanish

‘Bother him so he does not allow that you are in presence of such majesty without reverence, fear, and love.’ (Epistolario espiritual, Juan de Ávila, 16th c.) Tarr (1922: 24) b. Si os da alivio el que yo me ausente If you gives relief the that I me be-absent ‘If it relieves you that I am away.’ (Antona García, Tirso, 17th c.) Herrero (2005: 82) c. El que sea esto así, yo lo sé; el por qué sea así, ignoro The that is this thus, I it know; the for what is thus, ignore.1SG ‘That this is thus, I know; why it is thus, I don’t know.’ (Persiles, Cervantes, 17th c.) Lapesa (2000: 544) These clauses with articles could be subjects and objects (Lapesa 2000: 544; Herrero 2005: 82), which points to their nominality, a CP-nominalization in terms of Kornfilt and Whitman (2011): (74) El que sea esto así

Interestingly, que-clauses with articles are not attested in the very nominal position of prepositional object, unlike indirect interrogative finite clauses, as the following example indicates (Lapesa 2000: 543): (75)

Tratóse del cómo y cuándo se avían de descubrir Dealt.3SG-refl of-the how and when refl had.3PL of discover unos a otros ones to others ‘It was discussed how and when they were supposed to disclose to each other.’ (Guerra de Granada, Hurtado de Mendoza, 16th c.)

The non-attestation of

does not fit well into the picture, as the que-clause is nominal and its combination with the article has been claimed as additional evidence of it (Barra 2002: 280–281). The same restriction is inoperative with indirect interrogative clauses.

Discussion and conclusions

103

3 Discussion and conclusions The previous panorama of historical Spanish syntax permits to reach several important conclusions and to evaluate the syntactic hypotheses presented, dealing with issues of prepositionality, argumenthood, Case checking, and nominality. First of all, it is possible to conclude that in historical Spanish prepositions could take nouns (non-clausal DPs), infinitives, and finite clauses – both adverbial clauses, common since the first attestations and argumental clauses, frequent from the 16th century on but with some early attestations – with variable degrees of optionality. Absence of the expected preposition is attested with all three groups, with higher levels in the case of clauses than with non-clausal nouns. For those speakers whose I-languages could produce non-prepositional variants where a preposition would have been otherwise required, the object category – clausal or not – unexpectedly non-prepositional, must have been equally licensed as argument, including Case checking. Secondly, the existence of the pattern

is clearly attested from the earliest texts. Thus, even if argumental prepositional clauses were not frequent – which is not to say that they did not exist – before the 16th century, the reason for their emergence and dispersion cannot be directly linked to some sort of constraint against combining prepositions and que-clauses. Obviously, this challenges Barra’s (2002) main hypothesis. As indicated above, Barra argues that finite clauses were not nominal categories before the 16th century, so adverbial clauses must be somehow exempt from Case for this hypothesis not to be immediately falsified. His solution resides in establishing a Case feature-related differentiation between the functional prepositions present in argumental prepositional finite clauses and the lexical prepositions introducing adverbial clauses. While the former do assign (check) Case, the latter do not. This special characteristic of lexical prepositions is further reliant on their adverbial origins (Barra 2002: 284–285). However, accepting that abstract Case exists and that it represents the covert counterpart of morphological case, the fact that these same prepositions are attested with case-inflected pronouns is problematic. Furthermore, lexical prepositions and functional prepositions alike combine with the same type of morphologically inflected pronouns, which again indicates that, according to the standard formulation of Case checking and nominal licensing, both must check oblique case. Consider the following examples with both lexical prepositions (76a) and functional prepositions (76b,c):

104

Historical Spanish

(76) a.

por mi e por todos los reyes del mio linage for me and for all the kings of-the my lineage ‘For me, and for all the kings of my lineage.’ (Documentos castellanos de Alfonso X – Castilla la Vieja, Alfonso X, 13th c.)

b.

Amembrando me de ti en mio lecho Remembering me of you in my bed ‘Remembering you in my bed.’ (Biblia Latina, 13th c.)

c.

Todos en ti creemos e a ti adoramos All in you believe.1PL and to you adore.1PL ‘We all believe in you and adore you.’ (Milagros de Nuestra Señora, Berceo, 13th c.)

Further support comes from Bassols (1992: 144), who points out that already in Latin there remain only minimal remnants of the adverbial origin of some of these prepositions. Evidence for this comes both from the fact that they can no longer be found used productively as adverbs Bassols (1992: 149), and from the fact that all of them do assign case. The existence of adverbial finite clauses and some early examples of argumental prepositional finite clauses points to the non-existence of any operative constraints barring

in old Spanish. The Case-based hypothesis also faces problems from the existence of the configuration

. In principle, such configuration could be considered as proof for the absence of Case-checking for argumental queclauses (see for instance Barra 2002). However, the same configuration is attested where Case-assignment is out of the question, i.e., adverbial clauses introduced by lexical prepositions, which means that ultimately its function cannot be to block

in order to prevent Case checking. The coexistence of porque and also questions such role. A third issue has to do with argumenthood of the que-clause in historical Spanish. Barra (2002) proposes that all que-clauses in Spanish were initially adjuncts, a conclusion directly related to the (apparent) absence of prepositional argumental finite clauses before the 16th century. The logical implication is that the presence of the preposition in argumental clauses indicates the creation of argumental clauses, as they must have turned nominal and thus can be properly licensed as such in other typically nominal positions, e.g., objects and subject. However, while it is normally agreed that the syntax displayed in the available

Discussion and conclusions

105

old texts is more colloquial and less strict than the preset standard,8 it is possible to evidence that at the very least certain verbs did indeed select for their argumental clausal objects. A clear test is that of extraction out of the embedded clause, a common test for argumenthood. Consider the following examples, all of which show wh-extraction from the embedded clause: (77)

a. & dixole ihesu xpisto que quieres que te faga And told.3SG-him Jesus Christ what want.2SG that you do.1SG ‘And Jesus Christ said to him: What do you want me to do to you?’ (Castigos, 13th c.) b. ¿Qué es lo que quieres que fagamos? What is the what want.2SG that do.1PL ‘What is what you want us to do?’ (Calila e Dimna, 13th c.) c.

¿quando quieres que rruegue por ti . . . ? when want.2SG that beg.1SG for you ‘When do you want me to pray for you?’ (Biblia Escorial I-j-4, c. 1400)

Extraction signals the argumenthood of the embedded clause in examples dated as early as the 13th century. Consequently, the que-clause must have the required feature loading to be properly licensed in nominal positions, including Casechecking. In other words, extraction provides additional evidence of the nominality of the clause long before the 16th century. This does not preclude the fact that other combinations with adverbial interpretation were really adjuncts. Those adjunct clauses could have nevertheless been licensed as nominals via adjunct Case or default Case. Continuing with argumenthood, another implication is that the presence of the preposition goes hand in hand with argumenthood. The previous paragraph already rejects such conclusion. One of the criteria applied to adjunct clauses is the lack of temporal/modal anchorage to the tense of the main clause, which is common in historical Spanish, full of violations of consecutio temporum. Barra (2002: 84) raises the issue of the presence or absence of the subjunctive mood, described as the selected mood (see Carrasco 1998, 2000; but see also the different views on subjunctive selection in Bosque 1990). However, as shown 8 But such is not the case in the colloquial language, which maintains many of those looser associations in subordination typically described for old Spanish, as in Dame el libro, que tengo que estudiar. (lit. ‘give me the book, that I have to study’, ‘give me the book because I have to study’).

106

Historical Spanish

in examples (66) and (67) above, the presence or absence of the preposition does not correlate one-to-one with mood selection in data from the 16th and 17th centuries. A similar scenario is found in the modern language. A final matter has to do with the combination with the article. It was shown that finite clauses are more restricted than their infinitival counterparts in only allowing the definite article el (‘the’). Crosslinguistic evidence will prove that the presence of the article is not necessary for nominality, since the que-clause is grammatical in nominal positions even without a determiner. Additionally, the absence of

in Spanish reveals the process of strict adjacency that seems to have operated in the past and still operates in the present in these combinations, both with functional and lexical prepositions, as will be discussed in the next chapter with present-day Spanish data. To sum up, Case is not the factor behind the emergence of prepositional finite clauses in Spanish. In the next chapters I will explore present-day Spanish and will as the implications discussed here have crosslinguistic consequences. Understanding how French, Italian, and Portuguese prepositional clauses work and have evolved, and contrasting them with Spanish, is vital and will provide additional important evidence as to the explanatory role of Case or nominality and argumenthood in the clausal domain.

Chapter 4

Present-day Spanish 1 Introduction This section moves us forward in time to present-day Spanish. The purpose is to examine the syntax of prepositional (finite) clauses, again in comparison with nouns and infinitives. The study of present-day Spanish concentrates on 20th– 21st century Spanish, including examples extracted from online searches in popular search engines such as Google.com which offers the researcher great amounts of colloquial Spanish from all over the world. Apart from my own native judgments, other native informants were consulted in certain cases, as indicated.

2 Prepositions and (non-clausal) nouns The syntax of prepositional phrases with non-clausal DPs as objects is much more homogeneous in present-day Spanish than in historical Spanish, probably in part due to standardization. While non-prepositional objects are not unknown in colloquial Spanish (1e), the following set of examples and grammaticality judgments illustrates the fact that variation is much less frequent nowadays: (1)

a. Cuento *(con) tu participación Count.1SG with your participation ‘I count on your participation.’ b. Me alegro *(de) tu participación Me be-happy.1SG of your participation ‘I am happy about your participation.’ c. Estoy cansado *(de) tu comportamiento Am tired of your behavior ‘I am tired of your behavior.’ d.

La propuesta *(de) reforma es necesaria The proposal of reform is necessary ‘The reform proposal is necessary.’

108

Present-day Spanish

e.

Me acuerdo (de) una cosa que me dijo Pepe Me recall.1SG of a thing that me said Peter ‘I recall one thing Peter told me.’

Finally, as expected, DPs can be the objects of lexical prepositions heading PPs functioning as modifiers (syntactic adjuncts): (2)

a. Lo hice por esa razón It did.1SG for that reason ‘I did it for that reason.’ b. Encontraron varios libros en el desván Found.3PL several books in the cellar ‘They found several books in the cellar.’

In the modern language some verbs may change their meaning depending on whether they take a preposition or not, partially different from the situation described for historical Spanish, where variation was much more prevalent. A well-known case is lexical aspect in pensar vs. pensar en (see Demonte 1992: 443–444): (3)

a. Pensó la respuesta Thought.3SG the answer ‘He came up with the answer.’ b. Pensó en la respuesta Thought.3SG in the answer ‘He thought of the answer.’

On the other hand, not all variation is aspectually motivated. For instance, Horno (2002: 94) wonders what the difference (if any) is between the following two sentences, where the presence or absence of the preposition really does not seem to alter the semantics: (4) a.

Rosa cuida de la casa Rosa cares of the house ‘Rosa takes care of the house.’

b.

Rosa cuida la casa Rosa cares the house ‘Rosa takes care of the house.’

Prepositions and (non-clausal) nouns

109

Other verbs with aspectual differences include soñar/soñar con (‘to dream’/ ‘to dream of’), hablar/hablar de (‘to speak’/‘to speak of’), and disfrutar/disfrutar de (‘to enjoy’/‘to rejoice in’). Importantly, Demonte (1992: 444) claims that the argument structure is the same in both cases. The complement checks Case as well. In agreement with the description of prepositions and Case presented before, we can assume that en is a functional preposition materializing inherent Case.1 Horno (2002) claims that PP-verbs such as hablar de, disfrutar de, pensar en, contar con actually involve lexical prepositions in adjunct position plus further reanalysis. However, Rauh (2002) gives evidence that the prepositions in PP-verbs of this kind do not impose their own thematic requirements – as opposed to the preposition sobre (‘about’), for instance – as the final interpretation depends on the verb (see also Barra 2002), sharply contrasting with their undoubtedly lexical locative counterparts. Besides, their double nature as aspectual markers argues in favor of their functional nature, as aspectual markers are indeed functional categories. Finally, extraction suggests that these PPs are not adjuncts, as opposed to truly adjuncts like the PP in (5c): (5)

a. ¿Quéi estás pensando en que hagamos ti hoy? What are.2SG thinking in that do.1PL today ‘What are you thinking we should do today?’ b. ¿Quéi cuentas con que hagamos ti mañana? What count.2SG with that do.1PL tomorrow ‘What do you count on us doing tomorrow?’ c.

*¿Quéi estás pensando hasta que compremos ti? What are.2SG thinking until that buy.1PL Lit. ‘What are you thinking until we buy?

Arguably, the hypothesis of syntactic reanalysis could also account for the extraction pattern, but it would also need to face the problems indicated in chapter 2, section 2.2.2 regarding interpolation/intervention and constituency (see Baltin and Postal 1996). The prepositions in contar con, hablar de, etc. offer the same results in the deletion under conjunction test (Zaring 1991: 369; Demonte 1992: 426; Horno 2002: 175; see section 2.2.1. in chapter 2):

1 Demonte (1992: 444) argues that it is materialization of structural Case by agreement.

110 (6)

Present-day Spanish

a. Cuento con Pedro y (con) sus hermanos Count.1SG with Pedro and with his brothers ‘I count on Peter and his brothers.’ b. Pienso en las vacaciones y (en) la vuelta al trabajo Think.1SG in the holidays and in the return to-the work ‘I think about my holidays and the return to work.’

This test reportedly tells lexical and functional prepositions apart in that only the former may be dropped. Recall the examples in (21) in chapter 2, section 2.2.1., repeated here for convenience (Horno 2002: 175): (7)

a. La tesis versa sobre el populismo y Ø el nacionalismo The thesis verses about the populism and the nationalism ‘The dissertation revolves around populism and nationalism.’ b. *La universidad prescindió de sus servicios y Ø su The university dispensed of his services and his aportación valiosa contribution valuable ‘The university dispensed with his work and valuable contribution.’

According to this test, the prepositions in (6) must be lexical, as they can be dropped. However, I would like to point out that the test actually renders wrong results. For one, I disagree with the grammaticality judgment in (7b), as I find the sentence completely grammatical. In addition, undoubtedly functional prepositions such as the so-called personal a in Spanish may also be dropped: (8) Vi a Pepe y (a) Juan Saw.1SG to Pepe and to John ‘I saw Pepe and John.’ Zaring (1991) already indicated that this test may not work as certain conjoined DPs may form a complex coordinated DP which is uniquely selected (and Casechecked) by one functional preposition, which easily accounts for the grammaticality of (8). However, other cases cannot be accounted this way. For instance, repeating the functional preposition en with PP-verb consistir en (‘to consist of’) unexpectedly produces a highly unacceptable sentence:

Prepositions and (non-clausal) nouns

(9)

111

*/??El programa consiste en 5 horas de clase los lunes y The program consists in 5 hours of class the Mondays and en 6 horas los martes in 6 hours the Tuesdays Lit. ‘The program consists of 5 class hours on Mondays and of 6 hours on Tuesdays.’

Furthermore, it also occurs that lexical prepositions cannot be dropped in certain circumstances, as Horno (2002: 175, fn 56) proves with the following example of directional a (‘to’): (10)

Primero iré a casa y luego *(a) la tienda First will-go.1SG to house and then to the store ‘First I will go home and then to the store.’

It seems that an excessively long first conjoined DP alters the reliability of the test and forces the appearance of lexical prepositions with the second DP: (11) a.

Hasta las 5 de la mañana el lunes y *(hasta) las 6, Until the 5 of the morning the Monday and until the 6 el martes the Tuesday ‘Until 5 AM on Mondays and until 6 AM on Tuesdays.’

b.

Lo hice inicialmente por mis hijos y *(por) mi mujer después It did.1SG initially for my kids and for my wife later ‘I did it initially for my kids and then for my wife.’

c.

Hablé por teléfono el lunes y *(por) internet el martes Spoke.1SG for phone the Monday and for internet the martes ‘I spoke on the phone on Monday and on the internet on Tuesday.’

d.

Estuve en la playa el lunes y *(en) la piscina el martes Were.1SG in the beach the Monday and in the pool the Tuesday ‘I was on the beach on Monday and in the swimming pool on Tuesday.’

Interestingly, the opposite does not necessarily apply, as the following example with functional prepositions demonstrates:

112

Present-day Spanish

(12) Consta de 5 capítulos en el primer libro y (de) 6, en el segundo Consists of 5 chapters in the first book and of 6 in the second ‘It consists of 5 chapters in the first book, and 6 in the second.’ It seems that the test points to the fact that lexical prepositions need to be present to check lexical Case, while functional prepositions, given the right configuration, may be absent even with additional intervening material (beyond Zaring’s 1991 suggestions), which suggests that the second DP is rendered Caseless or receives default Case. Notice that when morphological case is overt and therefore case is clearly checked, all prepositions must be repeated: (13) a.

Habla de ti y *(de) mí Speaks.3SG of you and of me ‘He is speaking of you and me.’

b.

Lo hice por ti y *(por) mí It did.1SG for you and for me ‘I did it for you and for me.’

In conclusion, the most important conclusion in this section is that both functional and lexical prepositions accept DPs as their objects. Interestingly, the question concerning how Case is checked has proven to be more cumbersome, as the few colloquial examples of preposition absence plus the intriguing results with coordination show that DPs may indeed be licensed without the mandatory presence of a functional preposition in certain configurations.

3 Prepositions and infinitives The situation is richer with infinitives. As in historical Spanish, the combination of prepositions and infinitives is common in the modern language. Prepositional complementizers are much less common. First, infinitival clauses can be complements of PP-verbs and other preposition-taking categories (Anula and Fernández 1997: 10): (14)

a. Es propenso a [cantar en público] Is prone to sing.INF in public ‘He is prone to singing in public.’

Prepositions and infinitives

113

b. Son expertos en arreglar averías Are.3PL experts in fix.INF breakdowns ‘They are experts in fixing breakdowns.’ Lexical prepositions heading adjunct PPs also accept an infinitival clause (Anula and Fernández 1997: 10): (15)

Compró lotería para ganar el premio Bought.3SG lotto for win.INF the prize ‘He bought lottery tickets to win the prize.’

Moreover, as happened in historical Spanish, infinitives may be introduced by de even when the verb does not select for that preposition with other complements, thus qualifying as cases with prepositional complementizers (Gómez 1999: 2128; see Real Academia Española 2010: 506 sanctioning these constructions as vulgar): (16) a.

Lo oí de entrar It heard.1SG of enter.INF ‘I heard him come in.’

b.

Me ha hecho de pensar Me has made of think.INF ‘He has made me think.’

As happens with non-clausal DPs, otherwise required prepositions can be absent in certain colloquial contexts in modern Spanish, as (17a,b) illustrate: (17)

a. Me alegro Ø poder ayudarte Me get-happy.1SG can.INF help.INF-you ‘I am happy to be able to help you.’ b. Me acuerdo Ø ir al cine del pueblo Me recall.1SG go.INF to-the cinema of-the town ‘I remember going to the town’s cinema.’ c. La idea *(de) estudiar inglés toda la tarde The idea of study.INF English all the afternoon ‘The idea of studying English all afternoon.’

114

Present-day Spanish

Therefore, in (17a,b) the infinitival clauses must still be licensed positionally without an overt functional preposition. Raposo (1987b) concludes that (Romance) infinitives are nominal categories since they must be introduced by a Case-marking category such as a preposition like other nouns: (18)

El miedo *(a) suspender el examen The fear to fail.INF the exam ‘The fear of failing the exam.’

Furthermore, infinitives can also be objects (19a) and subjects (19b), both typically nominal contexts: (19) a.

El niño teme suspender el examen The boy fears fail.INF the exam ‘The boy is afraid of failing the exam.’

b.

Suspender un examen es terrible Fail.INF an exam is terrible ‘To fail an exam is terrible.’

Thus, infinitival clauses display a typically nominal distribution in Spanish, which confirms their nominality. Additionally, certain infinitives may even combine with different determiners (Anula and Fernández 1997: 6): (20)

a.

Su caminar cansado His walk.INF tired ‘His tired walk.’

b.

Este tiritar de frío This shiver.INF of cold ‘Those shivers from the cold.’

c.

Ese telefonear de Pedro That phone.INF of Peter ‘That phoning of Peter’s.’

These same infinitives with determiners do not accept a more verbal modifier such as an adverb (Anula and Fernández 1997: 7):

Prepositions and infinitives

(21) a.

115

*Su caminar cansadamente His walk.INF tiredly Lit. ‘His walking tiredly.’

b.

*Este tiritar fríamente This shiver.INF coldly Lit. ‘This shivering coldly.’

The previous examples contrast with other kinds of infinitives which seem to combine both nominal and verbal properties (Anula and Fernández 1997: 8): (22)

a. El continuo dar dinero de María The continuous give.INF money of Mary ‘Mary’s continuous donation of money.’ b. El dar continuamente dinero María The give.INF continuously money Mary ‘Mary giving money all the time.’

As happened in historical Spanish, the infinitive in (22a) accepts the article, but nonetheless retains its argument structure, as the presence of a direct object evidences. What is more, despite the direct object, its subject is expressed via PP-modification. And (22b) is no different: an infinitive with article but modified by an adverb, with a direct object and a subject. The question of whether the infinitive is nominal or not, and also vis-à-vis the presence of a determiner, has received extensive attention in the literature, not only for Spanish (de Miguel 1995; Pérez 1998–99; Hernanz 1999; Ramírez 2003, among many others). For instance, based on the constraints such as those in (21) and others, Anula and Fernández (1997) conclude that any verbal-like category only nominalizes if and when they are introduced by a determiner. For that reason, the absence of the article indicates non-nominal (see also Plann 1981: 207). On the contrary, de Miguel (1996: 31) argues that “[t]he fact that a determiner heads the clause does not imply that the infinitival form has nominal status” (see also Hernanz 1999: 2205–2206), on the basis of examples of infinitives with articles that keep their verbal nature. However, the contexts where infinitives with determiners are found in present-day Spanish are most systematic than it might seem and certain properties may not co-exist, as Ramírez (2003: 121) correctly proves. Ramírez (2003) argues for a tripartite classification, purely nominal (DP-like for all purposes)

116

Present-day Spanish

(23a,b), sentential (CP-like for all purposes) (23c,d), and a third type – AgrP-like, which Ramírez (2003: 122) labels as mixed – which combines properties of both, such as determiners and adverbs or possessive determiner and direct object, etc. (23e,f). Consider also the unacceptability of (23h,i) (but see (23g)), where adjectives co-occur with an auxiliary verb (compound tense) and an aspectual functional verb (aspectual periphrasis): (23)

a. El suave murmurar The soft whisper.INF ‘The soft whispering.’ b. Su murmurar explica la reacción de María Her whisper.INF explains the reaction of Mary ‘Her whispering explains Mary’s reaction.’

Ramírez (2003: 118)

c. El escribir novelas ella explica su fama The write.INF novels she explains her fame ‘Her writing novels explains her fame.’ d.

El haber ella conducido ayer un camión explica The have.INF she driven yesterday a truck explains su alegría her happiness ‘Her having driven a truck yesterday explains her happiness.’ Ramírez (2003: 120)

e.

Aquel escribir novelas explica su fama That write.INF novels explains her fame ‘That writing of novels explains her fame.’

f. Su continuo conducir camiones imprudentemente Her continuous drive.INF trucks uncarefully ‘Her continuous reckless driving of trucks.’ g.

Su fácil poder recobrarse lo hace muy buen atleta His easy can.INF recover.INF-refl him makes very good athlete ‘His being able to recover so easily makes him a very good athlete.’ Ramírez (2003: 122)

h. ?Su constante tener que viajar en primera clase His constant have.INF that travel.INF in first class ‘His constantly having to travel in first class.’

Prepositions and infinitives

i.

117

??Ese constante haber estado en tensión es responsable That constant have.INF been in tension is responsible Lit. ‘That constant having been tense is responsible.’

This is exactly what is captured in Kornfilt and Whitman’s (2011) – and Panagiotidis and Grohmann’s (2009) – approach to clausal nominalization/ substantivization. Following Ramírez’s (2003) tripartite classification2 of infinitives with determiners, the different properties and modifiers/complements available can be formalized by positing different levels of nominalization/substantivization following Kornfilt and Whitman’s (2011) models, as follows (see also Alexiadou, Iordăchioaia and Schäfer 2011, who nevertheless argue for only two types of nominalized infinitives in Spanish): a. Example of VP nominalization, basically a noun from the morphological level; no direct object licensing (24) Su murmurar Her whisper.INF ‘Her whispering.’

2 Ramírez (2003) notes several interesting problematic examples, including the fact that nominalized infinitives in present-day Spanish may not express the object as a PP, unlike subjects (see Pérez 2002 for a solution in terms of Relativized Minimality). This applies to Italian as well (Zucchi 1993: 237). However, it is important to note that such syntactic configuration is attested in historical Spanish (Lapesa 2000: 532; Granvik 2012: 178). Since these problems have no implications on the syntax of prepositional (finite) clauses, they lie outside the scope of this book.

118

Present-day Spanish

b. Example of CP (or TP) nominalization (25)

El escribir novelas ella The write.INF novels she ‘Her writing novels.’

There are clear constraints attesting to the high level of nominalization, probably CP, but Alexiadou, Iordăchioaia and Schäfer (2011: 36) argue for TP nominalization. However, Kornfilt and Whitman (2011: 1301) indicate that one of the characteristics of Turkish TP nominalizations resides in the fact that subjects are marked as Genitives, not as nominatives, as is usual in verbal clauses in that language. Therefore, since the Spanish infinitival discussed here does contain a subject in (arguably) Nominative C/case, it seems to be nominalized above Turkish TP nominalizations. A possible argument for the TP nominalization option would be the fact that the subject is postposed, as opposed to many finite clauses, assuming licensing in situ, as in Mensching (2000). However, the subject must be expressed as is usual in tensed CPs, Nominative Case, and does not necessarily have to be postposed, as (23d) already illustrates (see also Fernández 1987 and Mensching 2000). Additionally, it is possible to have preverbal subjects, including subjects on top of aspectual projections and compound tenses, syntactically above TP, assuming a split TP, which Kornfilt and Whitman (2011) contemplate as an option: (26)

a. El yo haber acabado de pintar este cuadro es bueno The I have.INF finished of paint.INF this painting is good ‘For me to have finished this painting is good.’ b. El yo ir a hacer esto es un gran reto The I go.INF to do.INF this is a great challenge ‘For me to do this is a great challenge.’

Mensching (2000: 106) argues that preverbal subjects are licensed in AgrP, that is, above TP, indirectly adding support for a level beyond TP (e.g., CP level) in infinitival clauses, as in historical Spanish. Furthermore, the subject can be expressed neither as a possessive determiner (27a) nor via PP modification (27b),

Prepositions and infinitives

119

and no adjectival modification is allowed (30c), and it must contain at least a full Tense projection, as the compound tense form in (27d) shows. Infinitives may even project AspP with aspectual functional verbs, which merge above TP, as (27e) shows: (27)

a. *Su escribir novelas ella explica su fama Her write.INF novels she explains her fame Lit. ‘Her writing novels of hers explains her fame.’ Ramírez (2003: 119) b. *El escribir constantemente de ella explica su habilidad The write.INF constantly of her explains her skill Lit. ‘The constantly writing of hers explains her skill.’ Ramírez (2003: 120) c.

*El continuo escribir ella novelas explica su destreza The continuous write.INF she novels explains her skill Lit. ‘The continuous writing novels of hers explains her good skills.’ Ramírez (2003: 121)

d.

El haber escrito novelas ella explica su fama The have.INF written novels she explains her fame ‘Her having written novels explains her fame.’

e.

Ramírez (2003: 120)

El yo acabar de escribir este libro es maravilloso The I finish.INF of write.INF this book is great ‘For me to finish this book is great.’

In any case, all these possibilities are well accounted for assuming a dynamic nominalization process during the derivation which could start at different points. c. Finally, there are also examples of vP nominalizations, where the infinitival remains verbal up to vP. A typical example is Italian nominal infinitives (see Zucchi 1993), which allow for articles, possessive determiners and adjectives, both attesting a nominal projection and incompatible with a TP level nominalization, while at the same time the verb can still have direct objects, a verbal property. The non-finite clause remains verbal in the bottom part (vP and below) and is nominal above the vP:

120 (28)

Present-day Spanish

Su continuo conducir camiones imprudentemente Her continuous drive.INF trucks uncarefully ‘Her continuous reckless driving of trucks.’

The examples in (23g,h) pose an interesting problem as they show that infinitival nominalizations/substantivizations in Spanish can (and must) allow for adjectival modification in cases where the syntax seems to call for a nominalization point higher than vP, contra the suggestions in Kornfilt and Whitman (2011). In Spanish, this seems to apply only to some aspectual verbs (periphrastic expressions), always assuming that they must merge above vP (in TP or AgrP, possibly), as auxiliary verbs are unacceptable to ungrammatical with adjectival modification (see (23i) above; see also Hernanz 1999: 2358–2349). What they do maintain, as expected by Kornfilt and Whitman’s study, is the fact that the subject is still expressed as a determiner, typical of TP and vP nominalizations. This issue will reappear in Portuguese and Italian. What all these examples have in common is the fact that all of them end up becoming a DP. On the other hand, while any infinitive or infinitival clause can potentially be topped by a DP, it is not mandatory to project a DP to be nominal, as the distribution of infinitival clauses without determiners shows. Assuming that all infinitives in typically nominal positions are DPs leaves some important examples unexplained. While infinitives with overt determiners may be in subject (29a) and direct object positions (29b), the situation is unexpectedly different for prepositional objects. While not only VP but also the more verbal vP nominalizations require an overt D (29c,d), the opposite scenario arises with other types, as (29e,f ) illustrate (see also Pérez 1998–99: 153; Hernanz 1999: 2206). Indirect interrogative infinitives are attested with a determiner when introduced by a preposition (29g): (29)

a. (El) tener un hijo le cambió la vida The have.INF a kid him changed the life ‘To have a kid changed his life.’

Prepositions and infinitives

121

b. Lamento (el) tener que decir esto Regret.1SG the have.INF that say.INF this ‘I regret having to say this.’ c. Estoy cansado de*(l) continuo murmurar del mar Am tired of-the continuous whisper.INF of-the sea ‘I am tired of the constant whispering of the sea.’ d.

Lo vigilan por *(su) constante conducir camiones sin licencia Him watch.3PL for his constant drive.INF trucks without license ‘They watch him due to his constant driving without a license.’

e.

Está contento de(*l) tener un hijo Is happy of-the have.INF a kid ‘He is happy to have a kid.’

f. Luisa compró esto para (*el) poder ver películas Luisa bought this for the can.INF see.INF movies ‘Luisa bought this in order to be able to watch movies.’ g.

El conocimiento es el qué hacer y el por qué, la The knowledge is the what do.INF and the for what, the capacidad es el cómo hacerlo y por supuesto poder hacerlo ability is the how do.INF-it and for supposed can.INF do.INF-it ‘Knowledge means to know what to do and why; skill means to know how to do it and, of course, to be able to do it.’ (13-01-12)

Comparing (29c,d) and (29e,f ) would suggest a necessary return to the differentiation between nominal and verbal infinitives, but it was discarded above, plus notice that the determiner is even mandatory with vP nominalizations. The constraint is not against prepositional infinitivals either, but rather against an overt determiner or a determiner position specifically in such contexts. Fernández and de Dios (1991: 224) remark that infinitives cannot always be coordinated with DPs when they are objects of perception verbs such as ver (‘to see’) or oír (‘to hear’). However, determiner-less infinitives may be coordinated with plural bare nouns (30a), which are arguably NPs; as expected, these bare nouns can be objects of the same prepositions as well (30b):

122 (30)

Present-day Spanish

a.

Oyó gritar y golpes en la puerta Heard.3SG scream.INF and knocks on the door ‘He heard screaming and knocks on the door.’

b.

Esto es perfecto para baños en el mar o para bañarse en casa This is perfect for baths in the sea or for bathe.refl in house ‘This is perfect for baths in the sea or bathing at home.’

Similar issues will arise with finite clauses, which are the main focus of this book, as will be discussed in the next section.

4 Prepositional finite clauses 4.1 Prepositional que-clauses In present-day Spanish prepositions may combine with a que-clause in argumental positions. Observe the following examples of verbs selecting for a prepositional clause: (31) a.

Me acuerdo de [que viniste a mi casa] Me recall.1SG of that came.2SG to my house ‘I remember that you came to my house.’

b.

Aspira a [que lo llamen de esa compañía] Aspires to that him call.3PL from that company ‘He aspires to be hired by that company.’

c.

Insiste en [que vengas] Insists in that come.2SG ‘He insists on your coming.’

d.

Cuento con [que vengas] Count.1SG with that come.2SG ‘I count on you coming.’

Nouns (32a,b) and adjectives (32c,d) can also combine with a prepositional que-clause:

Prepositional finite clauses

(32)

123

a. La idea de [que vengas] The idea of that come.2SG ‘The idea that you come.’ b. Su oposición a [que te dieran el premio] His opposition to that you gave.3PL the prize ‘His opposition to you being awarded the prize.’ c. Estoy seguro de [que Pepe vino] Am sure of that Pepe came ‘I am sure that Pepe came.’ d.

Es contraria a [que los medios actúen así] Is opposed to that the media act thus ‘She is against the media acting this way.’

As attested since the earliest texts in the medieval period, adverbial finite clauses also abound in present-day Spanish – syntactically adjuncts to the main clause – showing lexical prepositions taking a que-clause as their complement: (33)

a. Voy por [que quiero] Go. ISG for-that want. ISG ‘I go because I want to.’ b. Fui sin [que me llamaran] Went. 1SG without that me called.3PL ‘I went without having been called.’ However, certain prepositions do not accept finite clauses:

(34) a.

*ante que in-front-of that

b.

*contra que against that

c.

*bajo que under that

d.

*hacia que toward that

124

Present-day Spanish

Interestingly, these same prepositions do not accept infinitives either, pointing to an idiosyncratic semantic constraint – since syntactically there is no reason why, as those prepositions may take other DPs as complements – as highlighted in Barra (2002: 197), who shows that only 4-dimensional prepositions may select a clause.

4.2 Prepositional indirect interrogative finite clauses Indirect interrogative clauses – headed by the complementizer si or introduced by wh-phrases – can be prepositional objects, both in argumental and adjunct (adverbial) positions. Consider the following examples of prepositional verbs (35a), nouns (35b), and adjectives (35c): (35)

a. Platicaron sobre si le convenía viajar la Spoke.3PL about whether him suited.3SG travel.INF the semana entrante week entering ‘They spoke about whether he could travel the following week.’ Suñer (1999: 2151) b. Me carcome la duda de quién demonios se presentará Me preys the doubt of who demons refl will-come.3SG ‘It preys on my mind who the hell will show up.’ c. No estoy seguro (de) cómo y dónde le convendría Not am sure of how and where him would-suit.3SG establecer su negocio establish.INF his business ‘I am not sure how and where he should launch his business.’ Suñer (1999: 2153)

While not as frequently as que-clauses, indirect interrogative finite clauses may also appear as prepositional complements in adjunct (adverbial) positions: (36)

a. Por cómo habló, sé que no está contento For how spoke.3SG know.1SG that not is happy ‘Considering how he spoke, I know that he is not happy.’

Prepositional finite clauses

125

b. Te he comprado esto por si lo necesitaras en el viaje3 You have.1SG bought this for if it needed.2SG in the trip ‘I bought this for you just in case you need it in your trip.’ As can be seen, indirect interrogative finite clauses show a parallel behavior to que-clauses.

4.3 Variation and optional prepositions In present-day Spanish many argumental prepositional finite clauses still coexist with their corresponding non-prepositional counterparts, with variation in frequency of use and degree of acceptability among speakers, as expected (the well-known queísmo; see Rabanales 1977; Gómez 1999: 2133–2145, among others). The non-prepositional variants can be interpreted as the continuation of the original forms which were never completely substituted by the innovative ones (Tarr 1922: 65, fn 6). Functional prepositions de, a, and en may be left out, especially the first one: (37)

a. Me acuerdo (de) que viniste a mi casa Me recall.1SG of that came.2SG to my house ‘I recall that you came to my house.’ b. Estoy esperando (a) que salga el sol Am waiting to that rises the sun ‘I am waiting for the sunrise.’ c. Insiste (en) que vengas Insists in that come.2SG ‘He insists in your coming.’

3 Examples like this seem ambiguous between an indirect interrogative and a conditional interpretation. In principle, it is possible to find a context where the former reading is clearer, as follows: (1) Si lo necesitas o no, no lo sé, pero lo hice por eso If it need.2SG or not not it know.1SG but it did.1SG for that ‘Whether you need it or not, I don’t know, but I did it for that reason.’ Accepting that the si in indirect interrogative clauses and conditionals is the same category, as Spanish does not differentiate between if and whether, would eliminate the conflict as well.

126

Present-day Spanish

d.

Me amenaza que se irá de casa Me threatens that refl will-go.3SG of house ‘He threatens me with leaving home.’ (15-2-12)

e.

Todo depende que cumplan con los requisitos All depends that fulfill.3PL with the requirements ‘All depends on them fulfilling the requirements.’ (15-2-12)

f. En un principio me opuse que fuera In a principle me opposed.1SG that went.3SG ‘At first I opposed to her going.’ (15-2-12) g.

Su oferta estrella . . . consistía que los clientes de la operadora Its offer star consisted that the clients of the company podían hablar cuanto quisieran entre ellos could speak.INF as-much-as wanted among them ‘Its special offer consisted in unlimited calls for the phone company’s clients.’ (15-2-12)

Optionality, though, is not always possible when the semantics of the PPverb is not compositional anymore and the resulting PP-verb has become a unit, as is the case of contar con (see Horno 2002): (38) Cuento *(con) que vengas Count.1SG with that come.2SG ‘I count on your coming.’ Likewise, as in historical Spanish, the presence or absence of the preposition may translate in aspectual differences in certain verbs, such as, for instance, the aspectual difference between creer (‘to believe’) vs. creer en (‘to believe in’). With nouns, the non-prepositional alternatives are more easily attested:

Prepositional finite clauses

(39)

127

a. El hecho que llueva es bueno para el campo The fact that rains is good for the country ‘The fact that it is raining is good for the fields.’ b. Alegó que la negativa que los precandidatos únicos Claimed that the negative that the precandidates sole participen en debates viola el principio de equidad participate in debates violates the principle of equity ‘He claimed that the opposition to have lone pre-candidates participate in debates violates the equality principle.’ (15-2-12) c. Tenemos interés que vengan Have interest that come ‘We are interested in their coming.’

The expected preposition may also turn out to be absent when the selecting category is an adjective: (40)

a. Estoy seguro que vino Am sure that came ‘I am sure that he came.’ b. Independiente que venga otro pase Independent that comes other pass ‘Independent of whether another pass comes.’ (15-2-12) c.

Estoy interesado que me envies su catalogo Am interested that me send.2SG his catalogue ‘I am interested in you sending me his catalogue.’ (15-2-12)

The variation found with argumental que-clauses is also present when the object is an indirect interrogative finite clause: (41)

a. No estoy seguro quién estaba allí Not am sure who was there ‘I am not sure who was there.’

128

Present-day Spanish

b. El ministro dijo que todo depende si no hay impugnaciones The minister said that all depends if not there-are contests ‘The minister said that all depends on whether there are any contests.’ (15-2-12) c.

Es muy frecuente la pregunta cómo se usa el péndulo Is very frequent the question how refl uses the pendulum ‘The question how to use the pendulum is very frequent.’ (15-2-12)

There remain some combinations where the absence of the preposition causes ungrammaticality:4 (42) El hecho *(de) quién lo había roto era irrelevante The fact of who it had broken was irrelevant ‘The question of who had broken it was irrelevant’ In general, the behavior of que-clauses and indirect interrogative finite clauses is very similar when it comes to prepositional variation. In the case of adverbial clauses, it is crucial to point out that the colloquial language still retains several of the non-prepositional strategies for indicating cause or purpose, equivalent to those already attested in the medieval times, where subordination is usually considered to be rendered through pragmatics rather than proper syntactic embedding (Batllori, Sánchez, and Suñer 2001: 235–236; see Rabanales 1977: 564), just like in historical Spanish (Batllori, Sánchez, and Suñer 2001: 237): (43) a.

No voy a decirte nada, que te chivas Not go.1SG to tell.INF-you nothing that you tell ‘I’m not going to tell you anything because you’ll tell someone.’

b.

Grita fuerte, que se te oiga Scream strong that refl you hears ‘Scream loudly, so you’re heard.’

4 Interestingly, in English, where these prepositional constructions are much less frequent and where one would expect prepositions to never be mandatory, in reality this same context requires a preposition: (2) The question *(of ) who broke it is totally irrelevant now

Prepositional finite clauses

129

As a final brief note, it is interesting to note that the presence of the preposition – usually de – where it is not expected receives the name dequeísmo, somehow mimicking the typical context of the prepositional complementizer with infinitives in historical Spanish and colloquial Spanish, as indicated above. However, dequeísmo goes beyond those cases and is attested in many other configurations with que, including relative clauses, evidencing that its diffusion has been eminently analogical, surface-based (Gómez 1999; see also Demonte and Fernández Soriano 2005, who show that this de is not a Case preposition). Consider the following examples: (44)

a. Pienso de que tengo que ir Think.1SG of that have.1SG that go.INF ‘I think I have to go.’ b. Es probable de que llueva esta tarde Is probable of that rains this afternoon ‘It is probable that it will rain this afternoon.’

A related configuration, usually unnoticed in the literature (but see Rabanales 1977: 552; Gómez 1999: 2129), is a form of dequeísmo with indirect interrogative finite clauses, which seems to be the construction attested in the following examples, where the insertion of de is not normative: (45)

a.

Me pregunto de si no es ya la hora de empezar a Me wonder.1SG of if not is yet the hour of start.INF to hacer las cosas al revés do.INF the things at-the reverse ‘I wonder whether the time has not yet arrived to start doing things the other way around.’ (15-2-12)

b.

Le pregunte de cómo le podía hacer para formar Him asked.1SG of how it could.1SG make.INF for form.INF un sonido a sound ‘I asked him how I could manage to create sound.’ (15-2-12)

130

Present-day Spanish

4.4 Case and the licensing of the finite clause The previous examples show a high degree of prepositional variation in modern Spanish, with dialectal divergences as to the degree of acceptance. As prepositional objects, finite clauses, like infinitives and other DPs, must be licensed. In agreement with the description of abstract Case presented above, those clauses are to carry a Case feature which is checked by either a lexical preposition (lexical Case) or a functional preposition (materialization of inherent Case): (46)

a.

Me acuerdo [KP de [CP que viniste a mi casa]] Me recall.1SG of that came.2SG to my house ‘I recall that you came to my house.’

b.

Lo hice [PP para [CP que lo vieras]] It did.1SG for that it see.2SG ‘I made it so you could see it.’

Functional prepositions may be left out in the second conjunct in coordination, but those lexical prepositions introducing adverbial clauses may not (see Bosque 1998: 214): (47)

a.

La idea de que vengas y (de) que te guste The idea of that come.2SG and of that you likes ‘The idea of you coming and liking it.’

b.

Lo hice para que lo vieras y *(para) que lo compraras It did.1SG for that it see.2SG and for that it bought.2SG ‘I made it so you could see it and buy it.’

The results seem to be the exact opposite of the expectations, as it is the functional prepositions which can be absent, whereas the lexical preposition seems to be obligatory. Even in the case of grammaticalized PP-verbs such as contar con (‘to count on’) is it possible to find examples where the second conjunct is a CP introduced without a preposition: (48)

Ya que cuento con que sea el primero de muchos concursos Yet that count.1SG with that is the first of many contests y que cada vez los premios sean mejores and that each time the prizes are better ‘Since I count on it being the first of many contests and on the fact that the prizes will improve.’ (16-2-12)

Prepositional finite clauses

131

Notice that TPs can be coordinated without any problems: (49)

a.

La idea de que [TP vengas y te guste] The idea of that come.2SG and you likes ‘The idea of you coming and liking it.’

b.

Lo hice para que [TP lo vieras y lo compraras] It did.1SG for that it see.2SG and it bought.2SG ‘I made it so you could see it and buy it.’

In any case, I already discussed the unreliability of this test before. In (47a) we see that the second conjunct arguably receives Case either by forming a complex conjunct with the first conjunct, following Zaring’s (1991) idea, or by default Case as (inherent) Case may be positionally checked, if Case is actually necessary in all contexts, as (47a) may also be used simply as evidence for the optionality of Case-checking for finite clauses in Spanish. It may well be the case, though, that it is the first conjunct that matters and the only one which is properly licensed. As Zhang (2010: 51) shows with English examples, only the first conjunct (the external conjunct in her terms) must satisfy the requirements imposed by the selecting category. Evidence can be obtained from examples such as the one in (50), where it is interesting to note that it is the first (that is, external) conjunct which controls the optionality or not of the preposition. If the first conjunct is non-prepositional, the second must be so as well: (50)

La idea que vengas y (*de) que te quedes en casa The idea that come.2SG and of that you stay.2SG in home ‘The idea of you coming and staying home.’

(50) shows that the non-prepositional counterparts contain CPs that must be Case-licensed as well. The issue of Case-checking in non-prepositional alternates has very important theoretical consequences with historical implications regarding the causes behind the emergence of prepositional finite clauses (in Spanish and crosslinguistically) and its relation with Case. It has been proposed that the absence of preposition in these constructions equals adjunction in syntax, not sentential complementation (Barra 2002), which applies to both clausal subordination in Spanish before and after about the 16th century. Under this hypothesis, non-prepositional clauses remain syntactically adjuncts to PP-verbs, nouns or adjectives. However, as I proved for historical

132

Present-day Spanish

Spanish, syntactic adjunction blocks extraction out of the embedded clause, as the following example proves: (51)

*¿Qué te alegras porque comprara t Pepe? What you get-happy.2SG because bought Pepe Lit. ‘What are you happy because Pepe bought?’

Since adverbial are adjuncts, extraction is ungrammatical. Such is clearly not the case with selected clauses, as prepositional variation does not correlate with differences in argument extraction. Whether introduced by a preposition or not, the que-clauses in (52) must be argumental in all cases as extraction is grammatical, unlike in (51) (see Demonte and Fernández Soriano 2005: 1078): (52)

a. ¿Qué te acuerdas (de) que comprara t Pepe? What you recall.2SG of that bought Pepe ‘What do you recall Pepe buying?’ b. ¿Qué te alegras (de) que comprara t Pepe? What you get-happy.2SG of that bought Pepe ‘What are you happy Pepe bought?’

Furthermore, notice also the maintenance of equal mood selection in the subordinate clause in both prepositional and non-prepositional examples. In short, both prepositional and non-prepositional CPs must in principle be licensed in Case and be equally argumental.

4.5 Clausal argumenthood and prepositionality Another theoretical issue of relevance to the understanding of the syntax and history of prepositional finite clauses in Spanish has to do with argumenthood and prepositionality. Barra (2002) argues prepositional finite clauses emerge once que-clauses become nominal and may be objects of prepositions and check Case. Only once que-clauses became nominal did sentential complementation emerge in Spanish (and crosslinguistically). This is a very strong hypothesis which implies that there is a link between the existence of and argumenthood in a language. However, while many cases of de que, en que or a que are argumental – save those examples with clear adverbial readings in clear contexts

Prepositional finite clauses

133

where no preposition is required by any category, like old Spanish a que being a final ‘in order that’ – linguists have proved that not all of them are, despite involving a functional preposition and a clause. For instance, it is well known that Grimshaw (1990: 79) suggests that all prepositional finite clauses depending on nouns may actually be adjuncts, as they are not θ-marked by the noun or by the functional preposition. In the case of Spanish, Leonetti (1999b) proves that some prepositional finite clauses depending on nouns are indeed adjuncts, despite being prepositional. The first test has to do with the possibility of forcing an explicative apposition between commas, thus with a distinctive intonation: (53)

a. La solución de que no haya aparcamiento no puede ser The solution of that not is parking not can be.INF la grúa the truck ‘The solution to the lack of parking slots cannot be the tow-truck.’ Leonetti (1999b: 2090) b. *La The la the

solución, que no haya aparcamiento, no puede ser solution, that not is parking, not can be.INF grúa tow-truck

‘The solution, the lack of parking, cannot be the tow-truck.’ These que-clauses are arguments because they cannot be restated between commas. Compare (53) with the following group of examples of adjunct/appositive prepositional clauses, where the explicative apposition between commas is perfectly grammatical: (54)

a. La hipótesis de que el Caso no funciona es una posibilidad The hypothesis of that the Case not works is a possibility ‘The hypothesis that Case does not work is a possibility.’ b. La hipótesis, que el Caso no funciona, es una posibilidad The hypothesis, that the Case not works, is a possibility ‘The hypothesis, that Case does not work, is a possibility.’ c. El hecho de que vinieras fue bueno The fact of that came.2SG was good ‘The fact that you came was good.’

134

Present-day Spanish

d.

El hecho, que vinieras, fue bueno The fact that came.2SG was good ‘The fact, that you came, was good.’

The explicative apposition can be improved by adding an es decir (‘that is to say’) between the commas, an option which is not possible for the argumental example in (53). As Leonetti (1999b: 2091–92) shows, adjunct clauses may be restated as copulative sentences, which highlights the fact that the noun and the finite clause are at the same level (i.e., adjunction) and therefore there is no real subordination from one into the other. While (55a) is equivalent to (54a), (55b) does not correspond to (53a): (55)

a. La hipótesis es que el Caso no funciona The hypothesis is that the Case not works ‘The hypothesis is that Case does not work.’ b. #La solución es que no hay aparcamiento The solution is that not is parking ‘The solution is that there is no parking.’

Another test shows that only argumental clauses can combine with both definite and indefinite nouns (Leonetti 1999b: 2093): (56)

a. Una ventaja de que no llueva es que no hay ocasión de An advantage of that not rains is that not is occasion of perder el paraguas lose.INF the umbrella ‘An advantage of the fact that it is not raining is that there is no chance to lose your umbrella.’ b. *Un hecho de que sólo hayas tomado dos platos de A fact of that only have.2SG taken two dishes of tortellini indica que no te encuentras bien tortellini indicates that not you find well Lit. ‘A fact that you have only eaten two dishes of tortellini means that you’re not feeling well.’

The evidence presented thus far suffices to conclude that the presence of a (functional) preposition does not imply argumenthood of the finite clause. The

Prepositional finite clauses

135

same configuration may be an argument or an adjunct. After all, this is just logical in a language with other types of prepositional appositions, such as la ciudad de Toledo (‘Toledo city’), la calle de Alcalá (‘Alcalá Street’), or el mes de febrero (‘the month of February’) (see Leonetti 1999b: 2090, among many others). Consequently, both in current and historical terms, there is no one-to-one correlation between and argumenthood. Given that the tests show that an example such as el hecho de que . . . involves adjunction – with or without the preposition de – the fact that such constructions (and their argumental counterparts) were not (frequently) attested in the Middle Ages must have nothing to do with the creation of sentential complementation, as they remain adjuncts to this day. Therefore, the hypothesis that links the emergence of prepositional finite clauses to prepositional argumenthood is falsified.

4.6 Finite clauses and the article A finite clause can be optionally preceded by the definite article el in presentday Spanish – and only el, as opposed to nouns and some infinitival nominalizations (see section 3 in this chapter; see also Falk 1968; Plann 1981: 234) – seen as extra evidence in favor of its nominal nature (Picallo 2001, 2002; Barra 2002). As expected, they can be the objects (57a) and subjects (57b): (57)

a. Lamento (el) que no te prestaran ayuda Regret.1SG the that not you lent.3PL help ‘I regret that they didn’t give you a hand.’ b. (El) que te haya tocado la lotería es increíble The that you has taken the lottery is unbelievable ‘The fact that you have won the lottery is unbelievable.’

As indicated above, Kornfilt and Whitman (2011: 1299) consider these cases to be clear examples of CP nominalization, which I repeat here for convenience: (58)

Lamento [DP el [CP que no te prestaran ayuda]]

136

Present-day Spanish

CP nominalizations with finite clauses are grammatical in several other languages, including Polish, Modern Greek, Norwegian, Persian, Korean, Japanese, etc. Consider the following examples (Kornfilt and Whitman 2011: 1299): (59)

a. Jan oznajmil [to [ze Maria zmienia prace]] Jan announced that that Maria change job

Polish

‘Jan announced that Mary is changing her job.’ b. Dhen amfisvito [to [oti efighe]] Neg dispute.1SG the.ACC that left.3SG

Modern Greek

‘I do not dispute that he left.’ Roussou (1991: 78) provides additional examples of Modern Greek showing that finite clauses with articles5 may appear in other typically nominal positions, such as subject position and prepositional object: (60)

a.

[To [oti ehis filus]] simeni pola The that have.2SG friends mean.3SG much ‘The fact that you have friends means a lot.’

b. Apo [to [oti etreme]] From the that shook.3SG ‘From the fact that he was shaking. . .’ For Roussou (1991: 80), the presence of the article is ultimately forced upon Case requirements, as the article creates a nominal category, a DP, which can then be assigned Case, thus preventing the CP from receiving Case, in keeping with Stowell’s (1981) Case Resistance Principle. As has been discussed profusely before, this conclusion cannot be maintained in light of the Spanish data presented in this work. Panagiotidis and Grohmann (2009: 157) provide another example which illustrates the fact that indirect interrogative clauses may also be introduced by an article in Modern Greek: (61) Ghnorizo [to [pos agonizeste sklira]] know.1SG the how fight.2PL hard ‘I know how you struggle hard.’ 5 Some syntactic restrictions apply, e.g., direct objects and complementizer pu (‘that’) (Roussou 1991, 2010).

Prepositional finite clauses

137

However, the configuration is subject to two important constraints in standard Spanish. As has been long noticed in the literature (Demonte 1977; Leonetti 1999a; Picallo 2001, 2002; see Plann 1981 for the same idea for infinitives with articles), only factives are grammatical (in standard Spanish), as the following non-factive examples illustrate: (62)

a. Digo (*el) que tiene que venir Say.1SG the that have.3SG that come.INF ‘I say that he has to come.’ b. (*El) que tengo muchos libros es verdad The that have.1SG many books is truth ‘It is true that I have many books.’

In this sense, the nominalizations/substantivizations are more constrained now than in the 17th/18th centuries. More recently, Serrano (2009) has explored this constraint in detail and has concluded that it is not factivity but presupposition/ topicality which matters, as non-factive predicates such as evitar (‘to avoid’), fomentar (‘to promote’), justificar (‘to justify’), etc. may also take a finite clause with an article.6 These predicates also accept infinitives with articles. It is crosslinguistically important to highlight the fact that languages with a similar use of prepositional finite clauses as Spanish do not necessarily restrict their that-clauses with articles to just factive/presuppositional contexts. A perfect example of this is Swedish, which allows optional pronouns with factive (63a,b) and non-factive predicates (63c) (Brandtler 2010: 35): (63) a.

Jag beklagar (det) att jag sårade dig I regret it that I hurt you ‘I’m sorry that I hurt you.’

b.

Jag är förvånad över (det) att jag sårade dig I am surprised over it that I hurt you ‘I’m surprised that I hurt you.’

c.

Jag sa (det) att han förmodligen skulle förlora I said it that he probably would lose.INF

Non factive

‘I said to him that he would probably lose.’ 6 Picallo (2002: 119) convincingly proves that does not involve a covert noun fact. She also shows that el que-clauses are not syntactically equivalent to lo que-clauses (lit, ‘the.NEUTER that’, ‘what’).

138

Present-day Spanish

Such configuration resembles the optional use of it with certain predicates in English. Nevertheless, Swedish is not like English, where required prepositions cannot combine with a that-clause. The Swedish data pinpoint that pronominal alternatives (

) and direct prepositional combinations (

) are not incompatible – a problem for Moreno’s (1985–86) hypothesis –. Furthermore, this co-existence adds extra evidence in favor of the separation of Case and the existence of

, as it shows that the pronominal alternatives do not function as “buffer zones” preventing clauses from being selected by prepositions and, according to Case Theory, having their Case features checked. Moreover, while in Greek finite clauses may not be prepositional objects without the article (Roussou 1991; Vasilis Tsompanidis, pc), in standard Spanish it is actually the opposite (contra Plann 1981). No required preposition accepts the configuration as its complement: (64)

a.

Me alegro de(*l) que vinieras a mi casa Me get-happy.1SG of-the that came.2SG to my house ‘I am happy you came to my house.’

b.

Aspira a(*l) que lo llamen de esa compañía Aspires to-the that him call.3PL from that company ‘He aspires to get a call from that company.’

The constraint is equally operative with lexical prepositions in adverbial clauses, indicating that the restriction applies across the board regardless of the functional/lexical divide: (65)

Fui por (*el) que me apetecía Went.1SG for the that me felt-like ‘I went because I felt like it.’

As the examples in (29e,f ) show, similar constraints apply to CP/TP nominalization infinitives in standard Spanish (see also Hernanz 1999).7 7 Only the fossilized form al (lit. ‘to the’, ‘upon’) looks like

in examples such as the following: (3) Al comprar el libro At-the buy.INF the book ‘When buying the book.’ However, al has been analyzed as an idiosyncratic unanalyzable preposition (Schulte 2011; see also Torres 2009: 1728–1731) as its non-compositional semantics and particular syntactic behavior indicate.

Prepositional finite clauses

139

The ungrammaticality in (65) can be accounted for by remarking that they do not meet the factive/presupposition requirement. However, the same solution cannot apply in (64a), for instance, which is factive and yet ungrammatical in general. Interestingly, there seems to be some degree of variation in the modern language. An online search retrieves examples which prove that certain speakers may use

:8 (66)

a. no estoy seguro del que el programador funcione not am sure of-the that the programmer works ‘I am not sure that the programmer works.’

(21-2-12) b. aspira al que el establecimiento de la paz sea un aspires to-the that the establishment of the peace is a componente central component central ‘He aspires to make the establishment of peace a central component.’ (21-2-12) c. insisto en el que lleven a la familia insist.1SG in the that carry.3PL to the family ‘I insist that they carry their family.’ (21-2-12)

These speakers must have extended the grammaticality of to the context of prepositional object, possibly through E-language based analogy, and perhaps with the help of hypercorrection, to the point that they have erased the constraint operative for a majority of Spanish speakers.9 Furthermore, the examples in (66) are not semantically constrained as in standard Spanish, making it possible to find que-clauses with article even in non-factive/nonpresupposed contexts as the following, all ungrammatical for me: 8 Kany (1994: 440) reports some examples of in Chilean Spanish; some of them appear in what seem to be dequeísmo cases. 9 These kinds of innovations are expected to arise in a language that is spoken in so many countries. Another example is the el cual su N (lit. ‘the which his/her/their/its N’) construction studied in Delicado (2008) and Delicado (2012), which shows a true relative pronoun, el cual, used is the same contexts as the well-known quesuismo in possessive relative clauses.

140 (67)

Present-day Spanish

a. Esta bien documentado y es verdad el que esto no es posible Is well documented and is truth the that this not is possible ‘It is well documented and true that this is not possible.’ (21-2-12) b. Es cierto el que al observar la mucosa palpebral, Is true the that to-the observe.INF the mucous palpebral podemos determinar si la persona esta con anemia can.2PL determine.INF if the person is with anemia ‘It is true that by observing the eyelid mucous membrane we can determine if the person suffers from anemia.’ (21-2-12)

In any case, no examples were found with transitive verbs such as querer (‘to want’), with the grammaticalized PP-verb contar con (‘to count on’), or with lexical prepositions introducing adverbial clauses such as por or para (‘for’), which indicates that the extension is not across the board, as is typical with analogy. In any case, the reduced amount of examples available remarks that the constraint still applies for the majority of Spanish speakers. On the other hand, and again with varying degrees of acceptability for different speakers, certain indirect interrogatives are not subject to those two restrictions – especially those introduced by por qué, cómo and cuándo – as they are grammatical with an article and introduced by a preposition (68a–c), and are not semantically restricted to factive/presupposed contexts: (68)

a.

No me acuerdo de(l) por qué compré esto Not me recall.1SG of-the for what bought.1SG this ‘I don’t remember why I bought this.’

b.

ni yo me reconosco a veces por el como hablo nor I me recognize to times for the how speak.1SG ‘Sometimes I do not recognize myself in how I speak.’ (14-2-12)

c.

A la pregunta del si el BCE prevé reducir ese porcentaje To the question of-the if the CEB foresees reduce.INF that percentage ‘To the question whether the Central European Bank plans to reduce that percentage.’ (21-2-12)

Discussion and conclusions

d.

141

Nadie resuelve el cómo nos van a pagar las horas Nobody solves the how us go.3PL to pay.INF the hours ‘Nobody’s solving how they are going to pay for our work hours.’ (FORMESPA mailing list, 25-11-2011)

e.

Él siempre me pregunta el si su bronceado se está desvaneciendo He always me asks the if his tan refl is vanishing ‘He’s always asking me whether his tan is disappearing.’ (18-2-12)

Si-clauses with article, such as (68e), seem to be less frequent in the Spanishspeaking world than their counterparts with que, and are ungrammatical for many speakers, including myself. Those speakers for whom is ungrammatical have another context in common with Swedish speakers (Brandtler 2010: 35): (69)

Jag beklagar (*det) om jag sårade dig I regret it if I hurt you ‘I regret it if I hurt you.’

Brandtler (2010) argues that this ungrammaticality results from om-clauses not having a Spec,CP position for the pronoun. In the case of Spanish, it is possible to assume that the factive/presupposed semantic restriction applies here as well. As happens with que-clauses, it seems that certain speakers accept as a prepositional object, as illustrated in (68c). Again, this example is not grammatical for all speakers, including myself. The data presented show that Spanish finite clauses are distributionally nominal, like infinitival clauses. They may further combine with an article and produce a DP, subject to some semantic constraints and open to some variation.

5 Discussion and conclusions The data presented in this chapter support the conclusion that there is no restriction against prepositional clauses in present-day Spanish. Functional and lexical prepositions can take nouns, infinitives, and finite clauses. Despite standardization and the push from normative grammarians, variation persists, especially in the absence of finite clauses (queísmo), especially frequent in speakers from some Latin-American areas (see Rabanales 1977; Bentivoglio 1976; Gómez 1999, among many others).

142

Present-day Spanish

Furthermore, the data reinforce the conclusions reached in the chapter on historical Spanish regarding the fact that Case cannot be the factor responsible for the emergence of prepositional finite que-clauses. For instance, the existence of non-prepositional alternates with PP-verbs indicates that the argumental clause must be Case licensed like their prepositional counterparts. This conclusion has crosslinguistic consequences. The existence of (colloquial) examples with what must be classified as proper prepositional complementizers, and the fact that pragmatics persists as a possibility for non-prepositional adverbial clauses, are only two examples that prove that strategies found in the older texts were not completely eliminated and are part of the grammars of those individual speakers which can use them alongside the other alternatives, crucially highlighting that they are not incompatible in their I-languages. On the other hand, I showed that the presence of functional prepositions such as de introducing a clause does not necessarily turn it into an argument. This affects our understanding of the syntax of finite clauses in historical Spanish, since it shows that prepositionality does not create or prove argumenthood, contra the implications in Barra (2002). Adjunct prepositional finite clauses are not only introduced by lexical prepositions such as para (‘for’) or hasta (‘until’). Furthermore, the grammatical results of extraction with both prepositional and non-prepositional clauses shown above support the conclusion that both must be equally argumental. The data presented show that Spanish finite clauses are distributionally nominal, like infinitival clauses. They may further combine with an article and produce a DP, which brings us again to a very interesting categorial question. Finite clauses without an article appear in typically nominal contexts, and so do their counterparts with an article (despite certain constraints for queclauses). Kornfilt and Whitman (2011) refer to the former as CP nominalization, and indicate that the CP does not turn into a nominal unless the article is present (see also Plann 1981 for Spanish). Every nominalization requires some sort of functional nominal category, and the DP plays that role, the CP remaining fully verbal. Such use of the label nominalization does not capture the reality of the Spanish data (or that of Swedish, for instance). Spanish finite clauses need no article to be arguments, and as such, given that their distribution is typically nominal – with and without the determiner – it is necessary to maintain and represent their basic nominality as traditional grammars have usually asserted, and indicate that they may be further re-nominalized by adding an article and creating a DP, that is, they may be substantivized, in terms of Yap, GrunowHårsta and Wrona (2011: 14). Crucially, no determiner is necessary for clausal nominality, exactly as Panagiotidis and Grohmann (2009) claim for English

Discussion and conclusions

143

clausal gerunds (see also Roussou 2010: 587, discussed below). What seems necessary is to locate where clausal nominality is, that is, to find that functional nominal category Kornfilt and Whitman (2001) argue for in finite and infinitival clauses in Spanish (and other Romance languages, at least). There is recent evidence of the nominality of the finite clause in Romance. Roussou (2010), Manzini (2010) and Manzini and Savoia (2005, 2011), as briefly explained in chapter 2, section 2.1., have argued for the categorization of the complementizer que (and its relative and interrogative counterparts) as a noun, which takes a proposition as its complement. The resulting syntactic tree would be as follows (adapted/modified from Manzini and Savoia 2011: 15–17; see chapter 2, section 2.1, repeated here for convenience; see also Roussou 2010): (70)

So [NP che [ CP [ fai questo]]] Know.1SG that do.2SG this ‘I know that you do this.’

The complementizer is itself a nominal element (an N) located outside the CP which takes the CP as its argument. The complementizer, a noun, is the argument of the selecting category (the main verb, a preposition, etc.), not a functional verbal category. Roussou (2010) extends the complementizer-as-noun analysis to English that (see also Roberts and Roussou 2003: 110–121) and also to the Greek interrogative complementizer an (‘if’) and its English counterpart, if. Applied to Spanish, the resulting syntactic representations would be as follows: (71)

a. Lamento que sea difícil Regret.1SG that is difficult ‘I regret the fact that it is difficult.’

144

Present-day Spanish

b. Lamento el que sea difícil Regret.1SG the that is difficult ‘I regret the fact that it is difficult.’

Despite Manzini and Savoia’s (2005, 2011) and Manzini’s (2010) proposal that complementizer que (‘that’) and interrogative qué (‘what’) are the same word, there are important differences. Consider the following examples: (72)

a. Lamento (el) que sea difícil Regret.1SG the that is difficult ‘I regret the fact that it is difficult.’ b. *Lamento (el) que Regret.1SG the that Lit. ‘I regret the that.’ c.

*Lamento el que de la dificultad Regret.1SG the that of the difficulty Lit. ‘I regret the that of the difficulty.’

d.

*el que de esto the that of this Lit. ‘The that of this.’

e.

*el si de esto The if of this Lit. ‘The whether/if of this.’

f. el porqué The why Lit. ‘The why.’

Discussion and conclusions

g.

145

el porqué de esto the why of this Lit. ‘The why of this.’

Unlike porqué, que and si do not seem to behave as actual nouns, as they require only clausal complements. First, unlike porqué, que and si are unstressed, which may explain why they cannot be used in isolation. In addition, que and si are somehow constrained to participate only at higher level nominalizations, general for que, dialectally limited for si. The resulting configuration poses a further syntactic problem, as it consists of the nouns que or si taking a CP as their complements but without instantly blocking extraction out of the CP, contrary to usual theoretical assumptions about nominal islands in factives, for instance. Manzini (2010: 187) finds that the answer lies in the special nature of nominal complementizers, which, unlike regular NPs, allow for an additional escape hatch position in their left periphery that would permit extraction. Keeping the basic idea of the nominal nature – beyond distribution – of the complementizers in Romance, and in order to simplify the analysis and avoid the additional stipulations necessary to account for extraction, I believe that the evidence presented so far permits to combine the basic idea that que and si are essentially nominal – carrying a [+nominal] feature (in agreement with Barra 2002) and an associated Case feature – and the standard representation of que and si as complementizers inside the CP. The complementizers – whether que, si or Ø – would be the functional nominal categories, adapting Kornfilt and Whitman’s (2011) idea, that change the clausal derivation from verbal to nominal. The nominal category would not be located at the DP level, but rather, at a level equivalent to an NP. In cases of re-nominalization or substantivization, the determiner el creates a DP out of an otherwise already nominal projection; that is, the determiner is not mandatory for nominality (Roussou 2010: 587; in agreement with Holmberg and Platzack’s 1995: 19, fn 17 proposal for Scandinavian languages). The nominal complementizers ensure the proper licensing of their CPs (CP = NP) in nominal positions, maintaining phrasal and categorial coherence (given that C and CP are uniformly nominal) and endocentricity. The empty nominal complementizer can be argued to be present in indirect interrogative finite clauses. The resulting trees are as follows:

146 (73)

Present-day Spanish

a. Lamento que sea difícil (= 71)

b. Me pregunto por qué lo visité Me wonder.1SG why him visited.1SG ‘I wonder why I visited him.’

Whether introduced by an article or not, que-clauses produce the same semantic interpretations in terms of factivity/presupposition, showing that the overt determiner does not correlate with a semantic differentiation in terms of definite/indefinite complementizers as Roussou (2010: 592) establishes for Modern Greek. A similar functional nominal category must be present in infinitival clauses, in principle a nominal complementizer for CP infinitival clauses, assuming that all non-restructured infinitival clauses end up projecting at least a C. Raposo (1987b) argues for the morpheme –r as the nominalizer. Another option could be a nominal feature in T, as Raposo (1987a: 95) argues for inflected infinitives, despite creating a truly mixed category, or a null nominal affix, as Yoon (1996: 339) assumes for Spanish infinitival nominalization (her example is a vP nominalization, in Kornfilt and Whitman’s typology). Kayne (1999: 44) also argues for a nominal feature for infinitival clauses, which he describes as NPs rather than DPs as well, although with different conclusions regarding Case. A third possibility would be to assume an independent functional nominal projection on top of infinitival clauses which would carry the necessary nominal and Case features, especially for non-CP infinitival clauses.

Discussion and conclusions

147

Continuing with the analysis of indirect interrogative finite clauses, as indicated in the section on historical Spanish, a number of wh-phrases and clauses can be introduced by an article. Consider the following examples:10 (74) a. Me pregunto el porqué Me wonder.1SG the why ‘I wonder why.’ b. Me pregunto el por qué lo visité Me wonder.1SG the why him visited.1SG ‘I wonder why I visited him.’ c.

Me pregunto el porqué de mi visita Me wonder.1SG the why of my visit ‘I wonder about the reason of my visit.’

There is one important difference in the grammaticality of the absence of the article, as the following examples point out: 10 A possible analysis, extending Roussou (2010), Manzini (2010) and Manzini and Savoia (2005, 2011), which I will not pursue here in detail, would involve considering that certain Spanish indirect interrogative clauses are indeed headed by nominal categories. It can be concluded that porqué in both (1a) and (1b) is essentially the same noun, with the following syntactic representations: (4) a. Me pregunto el por qué lo visité Me wonder.1SG the why him visited.1SG ‘I wonder why I visited him.’

b. Me pregunto el porqué de mi visita Me wonder.1SG the why of my visit ‘I wonder about the reason for my visit.’

148 (75)

Present-day Spanish

a. Me pregunto el por qué lo visité Me wonder.1SG the why him visited.1SG ‘I wonder why I visited him.’ b. Me pregunto por qué lo visité Me wonder.1SG why him visited.1SG ‘I wonder why I visited him.’ c.

Me pregunto *(el) porqué de mi visita Me wonder.1SG the why of my visit ‘I wonder about the reason of my visit.’

d.

Me pregunto *(la) causa de tus dudas Me wonder.1SG the cause of your doubts ‘I wonder about the reasons for your doubts.’

This issue can be solved by appealing to the different levels of nominalizations/ substantivizations. (75a) is a high nominalization, like an el que-clause: (76) Me pregunto el por qué lo visité (= 74b)

On the other hand, (75c) is a basic conversion at the morphological level, equivalent to a VP nominalization for infinitives, with typically nominal PP modification.11 The determiner is mandatory like it is for other nouns such as causa in (75d) when there is modification. (75b) just displays a regular CP not involved in any nominalization/substantivization process, showing that indirect interrogative clauses do not need a determiner to be licensable in nominal positions. The nominality of a finite clause in Spanish does not require the overt presence of a determiner.

11 The same analysis in terms of lower nominalization vs. higher nominalization would apply to equivalent examples of porqué and other interrogatives in historical Spanish, especially once they are grammatical with an article.

Discussion and conclusions

149

This analysis can account better for all those other indirect interrogative finite clauses which are introduced by wh-phrases that cannot combine with an article but are nevertheless perfectly grammatical in nominal positions.12 The previous discussion has not addressed the existence of the constraint against

in standard Spanish, which is unexpected, given the nominality of que-clauses and the fact that ungrammaticality that does not arise with certain infinitival nominalizations or with indirect interrogatives (with different degrees depending on the speaker). This restriction cannot be caused by a potential incompatibility of the article with a tensed clause, as indirect interrogatives are grammatical. Moreover, is perfectly legitimate and licensable as an argument in subject and direct object positions. The fact that such constraint occurs even with factive verbs indicates that it is not due to semantic incompatibility (Serrano 2009). As mentioned before, Roussou (1991) argues that the overt presence of the article allows the CP to become a DP and receive Case from a preposition. However, it has extensively been discussed that clauses in Spanish need no article to be arguments. We can also discard the lack of nominal features, as it has already been proven that CPs are nominal in Spanish, so the cause must lie somewhere else. The traditional view on the formation of so-called subordinating conjunctions states that these elements are units, fixed complementizers (Herman 1963). Pavón (1999), Barra (2002) and others, however, have shown that the traditional view must be rejected and substituted by one that describes them as prepositions combined with a que-clause. Bosque (1998) argues that porque, nevertheless, seems to behave as if it were a unit, as if it had grammaticalized, as the following example illustrates: (77)

Porque lo buscas y *(por)que lo deseas13 For-that it seek.2SG and for-that it desire.2SG ‘Because you are looking for it and you want it.’

12 Consider the following examples: (5) a. Me pregunto (*el) con cuánto dinero fue Pepe allí Me wonder.1SG the with how-much money went Pepe there ‘I wonder with how much money Pepe went there.’ b. *Me pregunto el con cuánto dinero de Pepe Me wonder.1SG the with how-much money of Pepe Lit. ‘I wonder about the how much money of Pepe’s.’ 13 Si-clauses yield the same results: (6) Por si quieres venir o *(por) si quieres ir al cine For if want.2SG come.INF or for if want.2SG go.INF to-the cinema ‘Just in case you want to come or go to the cinema.’

150

Present-day Spanish

In view of this evidence, Bosque (1998: 214) argues for a dual analysis of porque (doble segmentación, ‘double segmentation’, in his own terms), reanalyzed as a complex unit in some cases and as

in others. This option is more flexible in the old language, where porque and que could be coordinated (Keniston 1937: 681), although I have already shown above that the behavior of the second conjunct can be independent from the requirements imposed by the selecting category on the first conjunct (Zhang 2010); furthermore, the repetition of porque is also attested in old texts. Notwithstanding the previous caveat, Bosque’s comments lead me to highlight the fact that, when a que-clause14 is the object of a preposition, the preposition, no matter what type – beyond porque or para que, as in Bosque (1998) – and the que-clause must observe strict adjacency (van Riemsdijk 1998), and accept no intervening materials, be it a determiner or other elements. What is more important, this restriction is not due to Case requirements, as Spanish does not impose strict adjacency for Case checking and, moreover, because intervention is perfectly grammatical with all prepositions when they combine with other DPs with article (78a). If strict adjacency is respected, intervention is once again permitted (78c). Tonic elements such as porqué or cómo do not require strict adjacency (78e–j), like the DP in (78a): (78) a.

Tengo que hablarte de sin duda alguna la Have.1SG that speak.INF-you of without doubt any the mejor opción para tu hogar best option for your home ‘I have to talk to you about the absolutely best option for your home.’

b.

*Tengo que hablarte de por supuesto que tienes Have.1SG that speak.INF-you of for granted that have.2SG que venir that come.INF ‘I have to talk to you about the fact that you must of course come.’

c.

Tengo que hablarte de que por supuesto tienes Have.1SG that speak.INF-you of that for granted have.2SG que venir that come.INF ‘I have to talk to you about the fact that you must of course come.’

14 Or a si-clause for some speakers. Obviously, the discussion of this constraint with siclauses is only relevant for those speakers who accept a si-clause with an article (in any case, not common or part of the standard language) except as a prepositional object. In any case, siclauses are not factive-presuppositional and thus are independently excluded from contexts where que-clauses are grammatical.

Discussion and conclusions

d.

151

*Me dio esto por, como era de esperar, si lo Me gave.3SG this for how was of expect.INF if it quería usar wanted.1SG use.INF Lit. ‘He gave this to me for, as expected, if I wanted to use it.’

e.

*/??A la pregunta de, cómo no, si lo había entendido To the question of how not if it had.1SG understood Lit. ‘To the question of, of course, whether I had understood it.’

f.

A la pregunta de, cómo no, por qué no fuimos To the question of how not why not went.1PL Lit. ‘To the question of, of course, why we didn’t go.’

g.

A la pregunta sobre, cómo no, el porqué de esta ausencia To the question about how not the why of this absence Lit. ‘To the question about, of course, the why of this absence.’

h.

A la pregunta sobre, cómo no, el por qué no fuimos To the question about how not the why not went.1PL Lit. ‘To the question of, of course, why we didn’t go.’

i.

Me preguntó por, claro está, cómo se resolvía eso Me asked.3SG for clear is how refl solved this Lit. ‘He asked me about, of course, how to solve that.’

j.

Me preguntó por, como te imaginas, el cómo se Me asked.3SG for as you imagine.2SG the how refl resolvía eso solved that ‘He asked me about, as you can imagine, how to solve this.’

The Romance complementizer que is indeed a clitic-like category (Wanner 1987: 47), an unstressed category (see Quilis 2012: 74, among many others), which supports the conclusion that the ungrammaticality of

in Spanish, even in factive/presupposed contexts where the article should in principle be allowed, is due to the cluster of the complementizer que (but also si for many speakers) and the preposition.15 Notice that the tonic noun porqué does not require strict adjacency (78f). 15 Arguably, the weak cluster

leans on a precedent or posterior stressed host, as evidenced by the fact that the cluster

may appear before or after intervening material.

152

Present-day Spanish

At a descriptive level, the effect described above prevents the D position from being overtly occupied or from being projected at all (see Delicado and González 2011 for an analysis in terms of incorporation):16 (79)

If maintaining a DP, C would have to moveto the empty D and from there form a weak cluster with the preposition. Since CPs can be arguments with or without the article, it would be descriptively accurate to claim that (Spanish) CPs can check Case without having to add a DP. If categories other than a DP can be – at least distributionally – nominal and appear in nominal positions (and check Case) without having to turn into DPs – for instance, PP arguments in Neeleman and Weerman (2001: 107–108)17 – then CPs could be among them. Actually, Dobrovie-Sorin, Bleam, and Espinal (2006: 60) argue that nominal categories other than DPs may appear in argumental positions. Additionally, arguing for a CP provides a clearer difference between strict adjacency of the C position versus non strict adjacency of DPs, including el porqué, etc. Furthermore, displacing the finite clause and separating it from the preposition does not allow the article to appear, which again suggests that there simply is no D projection at all: (80)

a. (*El) que estés aquí es de lo que me alegro The that are.2SG here is of the what me get-happy.1SG ‘The fact that you’re here is what I’m happy about.’ b. *El hecho de que estés aquí es de lo que me alegro The fact of that are.2SG here is of the what me get-happy.1SG ‘The fact that you’re here is what I’m happy about.’

16 An analysis in terms of incorporation (Delicado and González 2011) would automatically undermine the relevance of Case even further, as incorporated nouns do not carry Case features – arguably only DPs, especially the argumental ones, do – and would provide extra evidence against Barra (2002), as Case would simply not be involved at all in explaining the syntax – and consequently the emergence – of prepositional clauses in Spanish. 17 See also Culicover (1999: 96–101) for a more critical approach on categorization for nominal clauses and the issue of endocentricity.

Discussion and conclusions

c.

153

De que estés aquí es de lo que me alegro Of that are.2SG here is of the what me get-happy.1SG ‘The fact that you’re here is what I’m happy about.’

d.

Del hecho de que estés aquí es de lo que me alegro Of-the fact of that are.2SG here is of the what me get-happy.1SG ‘The fact that you’re here is what I’m happy about.’

A final piece of evidence comes from Leonetti (1999a: 826), who highlights that, despite its optionality, the presence of the determiner changes the syntax of the clause (whether finite or infinitival), as it creates an island for extraction:18 (81)

a. *¿Quién te sorprendió el que se retrasara una hora? Who you surpised.3SG the that relf delayed.3SG an hour? Lit. ‘Who did it surprise you the fact that he was an hour late?’ Leonetti (1999a: 826) b. ¿Quién te sorprendió que se retrasara una hora? Who you surpised.3SG that refl delayed.3SG an hour? Lit. ‘Who did it surprise you the fact that he was an hour late?’

In sum, the final syntactic tree is as follows, which shows that the unstressed cluster blocks the possibility of re-nominalization/substantivization; that is, no DP can be created: (82)

Case would still be checked, as expected. Coincidentally, this analysis provides further evidence for a differentiation between que and qué, as the latter can be introduced by an article even with por ( por el qué), while porque cannot.

18 This phenomenon in Spanish looks similar to the English that-trace phenomenon, where the literature has put forward arguments both for maintaining an empty C (later occupied by the verb or not) or for postulating that C simply does not project, exactly like D in our Spanish case, particularly if an empty D or a C-to-D movement are argued not to prevent extraction.

154

Present-day Spanish

This description of the facts partially rescues the traditional idea that subordinating complementizers somehow go together but without arguing for a reanalysis or morphologization of a PP into a complementizer and without losing the generalization that por, de and the rest are still prepositions in all contexts. The PP-analysis must still be maintained, as porque or de que (but also de si or por si, which are not traditionally listed as subordinating conjunctions) are perfectly analyzable in syntax, like other phonetic conglomerates such as al or del (Gaatone 1981: 206). The same analysis must apply to infinitival clauses in prepositional factive/ presupposed contexts in standard Spanish. Consider the examples in (83). Assuming a nominal complementizer positions for all non-restructured infinitival clauses (or other functional nominal category on top of the clause), (83b) reflects that the intervening expression claro está (‘clearly’, ‘of course’) must arguably be located after the complementizer and before TP. No D projects: (83)

a. de (*el) =C pagar el pan of the pay.INF the bread ‘Of paying for the bread.’ b. de (*el) =C, claro está, pagar el pan of the clear is pay.INF the bread Lit. ‘Of, of course, paying for the bread.’

Lower infinitival nominalizations/substantivizations are grammatical with an article in standard Spanish, as expected, and do not require strict adjacency, as any other DPs:

(84)

a. Hablamos de su continuo escribir novelas por la noche Talked.1PL of his continuous write.INF novels for the night ‘We talked about his continuous writing novels at night.’ b. Hablamos de, como te puedes imaginar, su continuo Talked.1PL of as you can.2SG imagine.INF his continuous escribir novelas por la noche write.INF novels for the night ‘We talked about, as you can imagine, his continuous writing novels at night.’

Discussion and conclusions

155

Such possibility is attested in old texts as well: (85)

en el hablar de aquellas buenas reglas in the speak.INF of those good rules Lit. ‘In the talking about those good rules.’ (Carta al inquisidor general don Diego de Arce y Reinoso, Palafox y Mendoza, 17th c.)

It is possible for certain speakers from all over the Spanish-speaking world to use the article with high infinitival nominalizations as prepositional objects, as the examplesin (86) illustrate. As expected, some speakers find all of them grammatical (Melvin González, pc, Puerto Rican Spanish), others find only some of them grammatical or acceptable (Carolina Castillo, Edith Hernández, pc, both Mexican Spanish), and some simply find all of them ungrammatical (my case, also Lorena Andueza, pc, both Spain Spanish): (86)

a.

el periodista recalcó que estaba “feliz” del ir the journalist remarked that was happy of-the go.INF al programa to-the program ‘The journalist indicated that he was happy to go to the program.’

(4-3-12)

b.

el hacer el equipaje no es un tema menor: no solo the do.INF the luggage not is a theme minor not only por el tener que “enfrentar” la maleta vacia for the have.INF that confront.INF the suitcase empty ‘Having to pack is not a minor issue, not only for having to confront the empty suitcase.’ (4-3-12)

c.

todo lo necesario para el tener el mejor equipo posible all the necessary for the have.INF the best team possible ‘Everything necessary to have the best possible team.’ (4-3-12)

156

Present-day Spanish

d.

En el sistema educativo de Tamaulipas existe personal que In the system educative of Tamaulipas exists staff that sin el tener el perfil docente, ejercen como maestros without the have.INF the profile teaching work as teachers ‘In the Tamaulipas education system there are staff who work as teachers without having teaching qualifications.’ (4-3-12)

What these examples reveal is that those speakers for whom they are grammatical are bypassing the strict adjacency active in standard Spanish. What is even more important, the contexts illustrated in (86) go beyond the semantic restriction for the presence of the article with factive/presupposed predicates, including adverbial clauses; rather, they pair with indirect interrogative clauses in allowing a much freer use of the article, an unsurprising result given the nominal value of the Spanish infinitive and the existence of different types of lower nominalizations (vP and VP nominalizations) which are perfectly grammatical in standard Spanish. In order to gain a wider and deeper understanding of the syntax of prepositional (finite) clauses beyond Spanish and to test and contrast the analysis presented in the literature for the emergence of prepositional finite clauses in Spanish with data from other related languages, the next chapters will be devoted to the description and discussion of historical and current Portuguese, French, and Italian data, with attention to Latin as well.

Chapter 5

Crosslinguistic perspective: Romance languages I 1 Introduction This chapter explores the syntax and history of prepositional finite clauses of the Romance language whose clausal syntax most closely resembles that of Spanish, Portuguese. Latin is also briefly analyzed. Non-clausal DPs and infinitives are also explored as appropriate. Being so closely related to Spanish, comparing the syntax and history of these languages offers a perfect scenario to examine and discover what is common, what is different – especially what is unexpectedly different – and what the theoretical implications are. In the following pages, I will show that there is no actual constraint against

in any of these languages, even if there are important differences. These differences inform the discussion and provide actual data for study and discussion in a crosslinguistic perspective. Variation/optionality is discussed too.

2 Latin and Proto-Romance 2.1 The Latin and Proto-Romance prepositional finite clause Latin had a rich finite and non-finite (e.g., bare infinitive, ACI infinitive, supine, gerund, and gerundive) clausal syntax, including argumental and adjunct clauses (see Vincent 1988: 65–73; Pinkster 1990, who offers a comprehensive study; or Horrocks 2011, among many others). Prepositional finite clauses are not just a purely Romance phenomenon either. Herman (1963: 74–86) documents the existence of an initially correlative eo quod (lit. ‘this that’) which eventually can be selected by several prepositions in Classical Latin, especially pro eo quod (lit. ‘for this that’). Consider the following examples: (1)

a. ab eo quod ignis propter splendorem fulget from this that fire for splendor shines ‘Because fire shines for splendor.’ (De lingua latina V, Varro) Herman (1963: 78)

158

Crosslinguistic perspective: Romance languages I

b.

iudicatus quidem est in eo, quod non credidit judged then is in this that not believed ‘He was judged because he didn’t believe.’ (Opera V, S. Eusebius Hieronymus)

c.

Herman (1963: 81)

pro eo quod pecunia ingenti me honorasti for it that money great me honored ‘Because you honored me with great wealth.’ (Historia Apollonii regis Tyri, 41)

Herman (1963: 84)

This syntactic construction where the pronoun is the real object of the preposition and the finite clause is an adjunct to the pronoun is also found in historical Spanish and other languages (for instance, modern Swedish). In addition to the correlative construction, Post-Classical and especially Late Latin already had a non-negligible number of truly prepositional adverbial finite clauses, all involving lexical prepositions selecting a quod-clause (see Väänänen 1995: 277). Unsurprisingly, those prepositions could combine with nouns as well. Observe the following examples: (2)

a. et multum bibet propter quod ardorem siccum pulmonis patitur and much drinks for that heat dry lung suffers ‘And he drinks a lot because he suffers from dry lung heat.’ (Mulomedicina Chironis, 4th c.) Herman (1963: 91) b. de quod Chlotarius regnum Burgundiae recipit from that Chlotar kingdom Burgundy received ‘Since Chlotar took the kingdom of Burgundy.’ (Fredegarii chronica, 7th c.)

Bassols (1992: 383)

c. post quod infantes non potuit habere after that kids not could have.INF ‘After she could have no kids.’ (Lex Salica XXIV, 6th c.)

Herman (1963: 94)

These sentences crucially illustrate that finite clauses could be prepositional objects already before the Romance period, even if this is just limited to adjunct clauses introduced by lexical prepositions. In his study of Spanish Late Latin texts (8th to 10th c., Proto-Romance period), Bastardas (1953: 93) finds that pro que (lit. ‘for that’, ‘because’) – which can be syntactically described as a PP – was already frequent. Consider the following example (Bastardas 1953: 93):

Latin and Proto-Romance

(3)

159

pro que erat in etate parbula for that was in age young ‘Because he was still too young.’ (Cartulario de San Vicente de Oviedo, 781-1200)

The existence of pro quod/pro que did not preclude the grammaticality of nonprepositional alternatives with final or causal interpretations (Bastardas 1953: 191), exactly as in historical and present-day Spanish. Bastardas (1953: 189) argues that pro quod/que is the result of the elimination of the pronoun eo in pro eo quod, as Moreno’s (1985–86) hypothesis for Spanish. While a feasible change in Latin, let us point out again that the same change has been rejected for historical French by Herman (1963) on the basis that both the pronominal and the non-pronominal alternatives are indeed contemporarily attested. We have already seen that porque is likewise attested from the earliest Spanish texts. What is more, Bastardas (1953: 189) highlights that the first examples of pro que date back to the 6th century and remarks that the combination

is typical of Late Latin. Both facts prove that the Romance period inherited a system where prepositional finite clauses were already grammatical.

2.2 The Latin and Proto-Romance prepositional infinitive In general, prepositional infinitives are considered a Romance creation. Despite the absence of a preposition, infinitives could still combine with nouns or adjectives, especially from the Post-Classical period on (Bassols 1992: 228–229). Furthermore, the infinitive could become a noun, particularly as VP nominalizations, according to the description in Bassols (1992: 229–230), as they could combine with different determiners and could be modified by adjectives. Bastardas (1953: 167), however, indicates that prepositional infinitives were already usual in popular speech as early as the 7th century. In particular, the (‘to + infinitive’) construction is already attested in the Latin period. Consider the following examples with different prepositions: (4) a.

tunc abiatis ad prendere de nos ipsa ereditate then have.2PL to take.INF of us this inheritance ‘Then you should take this inheritance from us.’ (Cartulario de San Vicente de Oviedo, 781-1200) Bastardas (1953: 162)

160

Crosslinguistic perspective: Romance languages I

b.

solatos populatos et alios pro populare lands populated and others for populate.INF ‘Inhabited lands and others yet to inhabit.’ (Becerro Gótico de Cardeña, 11th c.)

Bastardas (1953: 170)

2.3 Conclusion The purpose of this brief section has been to illustrate the grammaticality of both the

and the

configurations in the pre-Romance period, which explains why both are present, more or less profusely, in the Romance languages studied here, and which indicates that both kinds of clauses could, according to the theory, check Case already in Proto-Romance. This corresponds very well with the fact that (Classical and non-classical) Latin has finite clauses in nominal positions such as direct object where the clause must also check Case. Ernout and Thomas (1964) provide extensive examples of nominal finite clauses, both quod/que-clauses (and its variants with ut and other complementizers) and indirect interrogative clauses, also arguably nominal. Interestingly, Ernout and Thomas (1964: 295) remark that Latin quod used to introduce factive/presupposed clauses, thus yielding the same semantic effects as (standard) Spanish el que.

3 Portuguese 3.1 Historical Portuguese As in the case of Spanish, historical Portuguese is to be understood as a cover term for older periods of Portuguese. The examples range from as early as the 13th century up to the 17th century; the 16th and 17th centuries are also an important benchmark for this language as well. The examples were extracted from scholarly works, as indicated in each case, and from Davies and Ferreira’s Corpus do Português.

3.1.1 Prepositions and nouns. Prepositions and infinitives PP-verbs are attested since the earliest texts of historical Portuguese. As happened with Spanish, there was prepositional optionality. Mattos e Silva (1994: 83) indicates that certain PP-verbs could take a DP without a preposition or with more than one (see also Dias 1959: 11):

Portuguese

(5)

161

a. Et Deus cofonda a quantos creen en este adeujñador And god confuses to whoever believe in this fortune-teller ‘And may God confuse all those who believe in this fortune teller.’ (Cronica Troyana, 14th c.) b. nom creem outra vyda senom esta not believe.3PL other life if-not this ‘They do not believe in other life but this one.’ (Leal Conselheiro, 15th c.) c.

duvidarão de sua viagem doubted.3Pl of his trip ‘They questioned his trip.’ (Crónica do Conde D. Pedro de Meneses, 15th c.)

d.

de maneira que nom duvidaras em eles of manner that not will-doubt.2SG in them ‘So that you will not doubt about them.’ (Livro de vita Christi, 15th c.)

Another example of this variation is found with parar mentes (‘to pay attention’), which, exactly as in Spanish, is attested with and without the preposition de: (6)

a. parar mentes os meritos de cada huû stop.INF minds the merits of each one ‘To pay attention to the merits of each of them.’ (Euangelhos e epistolas con suas exposições en romãce, Garcia de Santa Maria, 15th c.) b. parar mentes no saluador posto na cruz stop.INF minds in-the savior put on-the cross ‘To pay attention to the savior on the cross.’ (Contemplacao de Sam Bernardo, 15th c.)

As for infinitives, PP-verbs, nouns and adjectives could combine with them with their required preposition: (7)

a. në obrigado a fazer në hûã cousa destas nor forced to do.INF not one thing of-these ‘Neither forced to do any of these things.’ (Terceyra Partida, Afonso X, 14th c.)

162

Crosslinguistic perspective: Romance languages I

b. que nõ fosse në hûû ousado de fazer força that not was not one daring of do.INF force ‘That no one dared do force.’ (Crónica Geral de Espanha de 1344, 14th c.) c.

diligencia em fazer serviços diligence in do.INF services ‘Diligence in providing services.’ (Livro de vita Christi, 15th c.)

Likewise, infinitives introduced by a lexical preposition could form adverbial clauses: (8) a.

o que fezesse esto por fazer algûû mal or that did.3SG this for do.INF some evil ‘Or that he would do this to cause pain.’ (Primeyra Partida, Afonso X, 14th c.)

b.

sem fazer hy muyta detëeça without do.INF there much stop ‘Without stopping there for long.’ (Crónica Geral de Espanha de 1344, 14th c.)

Mattos e Silva (1994: 67–68) and Dias (1959: 219–221) point out that prepositional optionality with infinitives is common in historical Portuguese, similar to the situation in Spanish. There is variation not only in the in the presence/ absence of the preposition or prepositional complementizer, as in the case of desejar and desejar a (‘to wish’), but also in the number used, as certain verbs could combine with infinitives with no preposition or with two, as is the case of começar/começar a/começar de (‘to start’) (see also Abraham 1938: 50–51; Russo 1942: 50–54; Dias 1959: 220; Said 1966: 339). Consider the examples in (9). (9a,b) show the verb dever (‘to have to’) with and without de; (9c,d,e) show variation with desejar: (9)

a. se acharem que com #V devo de lidar if tought.3PL that with five must.1SG of fight.INF ‘And I they thought that I must fight 5, I will.’ (Crónica Geral de Espanha de 1344, 14th c.) b. o que eu devo fazer the what I must do.INF ‘What I must do.’ (Crónica Geral de Espanha de 1344, 14th c.)

Portuguese

163

c. Et deseiaua de morrer And wished.3SG of die.INF ‘And he wished to die.’ d.

(Cronica Troyana, 14th c.)

deseio a fazer uoso prazer wish.1SG to do.INF your pleasure ‘I wish to please you.’ (Cronica Troyana, 14th c.)

e.

eu deseio ouujr algûa boa pallaura I wish hear.INF some good word ‘I wish to hear some good words.’ (Barlaam e Josephat, 14th c.)

Certain verbs, such as prometer (‘to promise’), may take a preposition when selecting an infinitival complement but yet show direct selection when the complement is a noun, as the following set of examples demonstrates, that is, typical of a prepositional complementizer: (10)

a. que ele prometia de pasar o mar com ele that he promised of pass.INF the sea with him ‘That he promised to pass the sea with him.’ (Cronica de Portugal, 15th c.) b. a outros prometia a fazer muytas mercees to others promised.3SG to do.INF many mercies ‘He promised to do many mercies.’ (Crónica Geral de Espanha de 1344, 14th c.) c.

soo lhe prometia alguum caminho de sua salvaçom only him promised.3SG some path of his salvation ‘He only promise dhim some path for his salvation.’ (Crónica de D. Duarte, Rui de Pina, 15th c.)

Furthermore, the same author may use the verb prometer with or without de in the same text: (11) a.

o qual . . . prometia entregar logo a el-rei de Portugal the who promised deliver.INF then to the-kind of Portugal ‘Who he promised to deliver then to the king of Portugal.’ (Historia da vida do Padre S. Francisco Xavier, Lucena, 16th c.)

164

Crosslinguistic perspective: Romance languages I

b.

ele prometia de permanecer só na confissão da fé he promised of stay.INF only in-the belief of-the faith católica Catholic ‘He promised to stay Catholic.’ (Historia da vida do Padre S. Francisco Xavier, Lucena, 16th c.)

The parallelism with Spanish is present even in allowing for non-prepositional infinitival selection with PP-verbs such as maravilhar-se (‘to marvel’), whose Spanish counterpart showed this very same property: (12) eu me maravilho haverdes por gram peguilho quatro forminhas I me marvel have.INF for great danger four frames ‘I’m amazed that you consider four shoe frames to be a great inconvenience.’ (Obra completa (A-M), Gil Vicente, 16th c.) Infinitival clauses introduced by a prepositional complementizer are attested not only as prepositional objects but also in subject position, once again as in Spanish (see Dias 1959: 219): (13) a.

A mim convem de hir a Castella To me is-convenient of go.INF to Castille ‘I should go to Castille.’ (Cronica de Dom Fernando, Fernão Lopes, 14th c.)

b.

mui grave seria de dizer quam forçosamente . . . very severe would-be.3SG of say.INF how forcibly ‘It would be very severe to say how forcibly. . .’ (Livro dos oficios, 15th c.)

There is one difference worth highlighting between Spanish and Portuguese infinitives, which is the existence of the inflected infinitive in the former. The inflected infinitive may be introduced by a preposition or prepositional complementizer and such configuration is attested already in the Middle Ages, as (14a) proves (see Brandão 1963: 441–443; Wireback 1994; and Martins 2001 for the emergence of the inflected infinitive out of Latin subjunctive forms). (14)

a. fariam bem de as leixarem would-do.3PL well of them leave.INF.3PL ‘They’d better leave them.’ (Livro dos oficios, 15th c.)

Portuguese

165

b. sem averem por ellas satisfaçom without have.INF.3PL for them satisfaction ‘Without getting any satisfaction from them.’ (Livro dos oficios, 15th c.) The following examples of inflected infinitives are also objects of lexical prepositions (forming causal adverbial infinitival clauses) and they show their own (preverbal) subjects as well, partially similar to examples attested in historical Spanish (see Mensching 2000: 27–28): (15)

por [hûûs seerem de hûa parte e outros doutra] for ones be.INF.3PL of one part and others of-other ‘Due to some being from one part and others, from another.’ (Leal Conselheiro, 15th c.)

As can be seen, infinitival clauses in historical Portuguese could appear in nominal positions. They could combine with determiners as well in clausal nominalization/substantivization, at different levels. Consider first the following examples of VP nominalizations – the most nominal – including sheer morphological conversion without any modifiers or arguments as PPs (16a,b), infinitives with subjects expressed as PPs (16c,d) or with direct objects expressed as PPs (16f ), and infinitives with the subject as a possessive determiner and modified by an adjective (16g): (16) a.

Acabado o jantar Finished the eat.INF ‘Once the food was finished.’ (Cronica de Dom Fernando, Fernão Lopes, 15th c.)

b.

quando os omes oem aquel cantar when the men hear that sing.INF ‘When the men hear that song.’ (Cronica Troyana, 14th c.)

c.

o pesar daquele que fez o pecado the hurt.INF of-that who did the sin ‘The pain of that who sinned.’ (Tratado de confissom, 14th c.)

d.

queo cresçer da fonte do Nylo that-the grow.INF of-the fountain of-the Nile ‘That the growing of the fountain of the Nile.’ (General estoria, 15th c.)

166

Crosslinguistic perspective: Romance languages I

e.

deste laurar de ferro of-this work.INF of iron ‘Of this working the iron.’ (Foraes, 15th c.)

f.

seu fazer perfeito his do.INF perfect ‘His perfect production.’ (Corte Enperial, 15th c.)

The examples illustrate the fact that different determiners, including the definite article and the demonstratives, are attested. The fact that these infinitives must take their complements via PPs results from the fact that their vP does not project. VP nominalizations may even be pluralized: (17)

a. cantaua taes cantares sang.3SG such sing.INF.PL ‘He sang such songs.’ (Barlaam e Josephat, 14th c.) b. ueería grandes pesares would-see.3SG great hurt.INF.PL ‘He would see great pain.’

(Cronica Troyana, 14th c.)

Given their DP status as their derivations include more nominal layers than verbal, it is not surprising to find that they are grammatical as objects of prepositions: (18)

no levar da perna in-the raise.INF of-the leg ‘In the act of raising the leg.’ (Livro da Ensinança de Bem Cavalgar Toda Sela, 15th c.)

It is also possible to find infinitival clauses where the nominalization point in the derivation is located higher than VP. Consider the following examples of nominalizations above the VP level, as evidenced by the fact that the vP can license direct objects (19a) but not overt subjects, still expressed as PPs (19a,b), and by the presence of adverbial modification (19b): (19) a.

o seu pedir tributo de vassalagem the his request.INF tribute of vassalage ‘His requesting vassalage.’ (Décadas da Asia, João de Barros, 16th c.)

Portuguese

b.

167

seu fazer e obrar ordenadamente his do.INF and act.INF orderly ‘His working in an orderly fashion.’ (Corte Enperial, 15th c.)

A comparison between (19b) and (16g) above illustrates the different level of nominalization/substantivization. According to Kornfilt and Whitman (2011), vP differ from TP nominalizations, which also license their subjects as Genitives, in allowing nominal-like modification such as an adjective on top of the vP projection. In the absence of specific examples with adjectives, the examples in (19) are nominalization at the very least at the vP level. Finally, high(er) nominalizations are also attested with inflected infinitives, showing full verbal properties, such as Nominative subject licensing, direct objects, full Tense projection, as evidenced by the presence of compound tenses: (20)

o terem-se empenhado tanto nas terras do inimigo the have.INF.3PL-refl worked much in-the lands of-the enemy ‘The fact of having put such effort in enemy lands.’ (Cartas, Vieira, 17th c.)

As discussed for Spanish, these examples illustrate CP nominalizations or TP nominalizations. Martins (2001, 2006) claims that historical Portuguese ECM, control and raising infinitives projected up to TP, as evidenced by the fact that clitics could climb out of the infinitival clause into the main clause much more frequently than in modern Portuguese, typical of reduced, restructured clauses. Since restructured infinitival clauses are reduced (Cinque 2006, among others), non-restructured infinitival clauses could nevertheless be CP arguments, as a generalization, and check Case. On the other hand, in the case of the prepositional infinitives discussed above, the presence of Nominative subjects differentiates them from the Turkish examples with deficient subject marking Tense presented by Kornfilt and Whitman (2011: 1301) as typical TP nominalizations. Furthermore, the nominalization in (20) applies at a higher point than those in (19), which, as I indicated, could be TP nominalizations according to the available evidence. In addition, assuming that each nominalization derivation is unique and that the same verbs may be nominalized at different points in the derivation, it is possible to argue for full CP status for (20), which is not part of a restructuring configuration and thus a potential (empty) C projection would not be problematic as it would not be blocking any extraction. For Mensching (2000), such infinitival clauses with subjects would be AgrP level infinitives, as AgrP is higher than TP in a split-TP approach to syntax, assuming that Nominative case is checked by Agreement. In

168

Crosslinguistic perspective: Romance languages I

the end, the essential point is that the upper part of the tree ends up as DP, and that the resulting phrase includes more verbal projections than in the case of vP and VP nominalizations. 3.1.2 Prepositions and finite clauses The history of Portuguese prepositional finite clauses resembles that of Spanish. In the older stages of the language, argumental finite clauses were not introduced by a preposition (see Mattos e Silva 1989: 743–744, 1994: 109–110): (21) a.

que Deus o quisesse ajudar que regesse ben that god him wanted help.INF that ruled.3SG well ‘That god wanted to help him rule well.’ (Crónica Geral de Espanha de 1344, 14th c.)

b.

nembrate que sem elles tu nom foras recall-you that without them you not were ‘Remember that without them you would not exist.’ (Livro da virtuosa bemfeitoria do infante Dom Pedro, 15th c.)

c.

maravilhousse que podya aquello seer marveled.3SG-refl what could that be.INF ‘He marveled at what that could be.’ (Crónica Geral de Espanha de 1344, 14th c.)

d.

eu som certo que sempre me consselharedes bë I am certain that always me will-advice well ‘I am sure that you will always give me good advice.’ (Crónica Geral de Espanha de 1344, 14th c.)

The bare complementizer que could also introduce adjunct clauses with causal and other typically adverbial interpretations contemporarily expressed by prepositional que-clauses as well, as in Spanish (see Huber 1933: 287; Herman 1963: 142; Mattos e Silva 1994: 117; Barreto 2002: 166–167). Consider the following example (Huber 1933: 287): (22)

Rrogo-te que me digas cujo he este castello; que Beg.1SG-you that me tell.2SG whose is this castle that andei por muitas terras . . . e nunca vy tam fremoso walked.1SG for many lands and never saw.1SG so beautiful (Vida de Sancto Amaro, 14th c.)

Portuguese

169

Argumental prepositional que-clauses are usually claimed to appear to the 16th/17th centuries, chronologically similar to the increase in frequency in the Spanish case. Consider the following examples with required functional prepositions de, a, em, and com (see Dias 1959: 260–263, among many others): (23)

a. Deos não lhe manda nem o obriga a que seja virgin God not him orders nor him forces to that is virgin ‘God neither orders him nor forces him to remain a virgin.’ (Historia do Japan 1, Frois, 16th c.) b. pera que . . . nos ajudeys a que não a desmereçamos for that us help.2PL to that not it un-deserve.1PL ‘So that you help us be deserving of it.’ (Carta escrita en Cangoxima a 5 de novembro de 1549, Padre Mestre Francisco Xavier, 16th c.) c. que quanto o padre mais insistia em que lho contasse that as-much the father more insisted in that it told.3SG ‘The more the father insisted that he had told it.’ (Historia da vida do Padre S. Francisco Xavier, Lucena, 17th c.) d.

não tem conta com que os seos sejão poucos not has count with that the his are few ‘He does not care about the fact that his people are few.’ (Historia do Japam 2, Frois, 16th c.)

e.

me começava a alegrar de que fosseis vindo me started.1SG to be-happy.INF of that were.2PL come ‘I was starting to feel happy that you had come.’ (Cartas familiares, Francisco Manuel de Melo, 17th c.)

As expected, prepositional indirect interrogative finite clauses are also attested around the same period (Dias 1959: 264). I could find one example from the 15th century, (24b): (24) a.

não eram sabedores de quem os houvera not were.3PL aware of who them heard ‘They didn’t know who had heard them.’ (Décadas da Asia, João de Barros, 16th c.)

170

Crosslinguistic perspective: Romance languages I

b.

nom se nembre de quem foy not refl recall.3SG of who was ‘May he not remember who he was.’ (Leal Conselheiro, 15th c.)

As in Spanish, apart from the prepositional infinitive, there are other prepositional contexts which could as well have served as models in historical Portuguese,1 including adverbial prepositional finite clauses, earlier examples of

, and some

. As for

, Herman (1963) indicates that its use is sporadic, as in Spanish, and mentions that the most frequent type is por tal que (lit. ‘for such that’, ‘in order that’). However, the presence of this pronominal alternative coexists with

even in the same text, as the following example evidences: (25)

por tal que sejaaes saluos. por que assy he a verdadeira caridade for such that are.2PL safe for that thus is the true charity ‘So that you’re safe, because thus is true charity.’ (O compromisso da Confraria da Misericordia, Confraria da Santa Casa da Misercordia de Lisboa, 16th c.)

Adverbial prepositional que-clauses abound in historical Portuguese (Brandão 1963: 545, among many others), evidencing that que-clauses could be prepositional objects and thus qualified as nominal. Consider the following examples, dating up to the 16th century: (26)

a. & porque queremos senpre sseer a uosso sserviço & for-that want.1PL always be.INF to your service ‘And because I want to be always at your service.’ (Textos Notariais. Arquivo de Textos Notariais em Português Antigo, 13th c.) b. Fery o meu servo, porque elle aja medo Wounded.1SG the my servant, for-that he has fear ‘I hurt my servant so that he gets scared.’ (Fabulario Esopo, 15th c.)

Huber (1933: 308)

1 Castilho (2004) offers an interesting hypothesis for the emergence of de que combinations linked to reduplication of the proform ende and de-PPs. However, the phenomenon of the emergence of argumental prepositional finite clauses goes beyond the preposition de and affects, not one particular preposition, but rather a full syntactic configuration, and by extension any prepositions eligible in that context, including de but also a, en/em, con/com and others, which lie outside of the scope of Castilho’s hypothesis.

Portuguese

171

c. ata que traga aquela farinha do trijgo que leuou until that brings that flour of wheat that took.3SG ‘Until he brings back the wheat flour that he took.’ (Posturas do concelho de Lisboa, 14th c.) Moreover, Castilho (2004: 3) and especially Mollica (1995) document the existence of early examples; while some can be better interpreted as relative clauses, the remaining examples prove that, as in Spanish, this configuration was grammatical for certain speakers before the 16th/17th centuries. Additionally, a search in Davies and Ferreira’s online corpus of Portuguese retrieves a number of early examples of argumental prepositional finite clauses. As happens in Spanish – past and present – some of these de que-clauses add a causal connotation, but not all: (27)

a. në nëbro de que lle nõ correse sange nor recalled.3SG of that him not run blood ‘Nor did he remember that any blood was running in him.’ (Cronica Troyana, 14th c.) b. a muytos pesou de que el así fora acusado to many weighed of that he thus was accused ‘Many were sad that he had been accused thus.’ (Cronica Troyana, 14th c.) c.

per razõ de que os dictos herdamento e oliual téémos for reason of the the said possession and olives have.1PL ‘Because we own the aforementioned possessions and olive trees.’ (Textos Notariais, 13th c.)

Como (‘how’) functioning as que and the prepositional de como and em como are attested in historical Portuguese as well. Dias (1959: 263) points out that de como and em como are found in titles and with declarative verbs, among other contexts, as in Spanish (see also Huber 1933: 294–296; Herman 1963: 167; Mattos e Silva 1989: 735–736): (28)

a. nembrando-se em como Potem e Achilas matarom Pompeeo recalling-refl in how Pothinus and Achillas killed Pompey ‘Remembering that/how Pothinus and Achillas killed Pompey.’ (Vida e feitos de Júlio Cesar, 15th c.)

172

Crosslinguistic perspective: Romance languages I

b. que ouera nouas de como ele era finado that had.3SG news of how he was dead ‘That he had received news that he had died.’ (Crónica de D. Fernando, 15th c.) The non-prepositional alternative, illustrated in the examples in (29) below, has remained fully grammatical and frequent to this day, another feature that Portuguese shares with Spanish (see Dias 1959: 260–261; Brandão 1963: 545; Mollica 1995 for Portuguese queísmo). Consider the following non-prepositional examples dated in the 16th and 17th centuries: (29)

a. queixando-se que S. A. lhe proibisse não se meter complaining-refl that S.A. him forbade not refl enter.INF no negócio in-the business ‘Complaining that Your Majesty had barred him from joining the business.’ (Cartas, Vieira, 17th c.) b. se maravilhavam que fosse filha daquele pobre vilão refl marveled.3PL that was daughter of-that poor peasant ‘They marveled that she was indeed the daughter of that poor peasant.’ (Proveito, Fernandes Trancoso, 16th c.) c. ninguë vos obriga que deis, nem que queirais nobody you forces that give.2PL nor that want.2PL ‘Nobody forces you to either give or want to.’ (Contos & historias de proveito & exemplo, Fernandes Trancoso, 16th c.) d.

insistio que pelo menos quinze dias se auzentasse insisted that for-the less fifteen days refl was-absent ‘Insisted that he be absent for at least 15 days.’ (Historia do Japam I, Frois, 16th c.)

e.

que estivessem seguros que ainda não era chegada sua hora that were.3PL sure that still not was arrived their hour ‘That they were sure that it was not their time yet.’ (Historia do Japam 3, Frois, 16th c.)

Importantly, the non-attestation of (frequent) argumental prepositional queclauses in historical Portuguese does not correlate with lack of argumenthood in subordination, as Barra (2002) suggested for Spanish. The grammaticality of

Portuguese

173

extraction out of an embedded clause signals the existence of argumental queclauses: (30)

a.

quem direi que som . . . ? who will-say.1SG that am ‘Who will I say I am?’ (A Demanda do Santo Graal, 15th c.)

b.

O que Mandamos que assy sse cumpra The that order.1PL that thus refl fulfills ‘Which we order to be fulfilled this way.’ (Cortes portuguesas, 15th c.)

Finally, certain indirect interrogatives could participate in clausal nominalization/substantivization at different levels by taking an article, including porque (‘why’) and como (‘how’), also the most common ones in historical Spanish (see Dias 1959: 266). Consider the following examples of CP level nominalization/ substantivization: (31) a.

sem entender o porque o faz without understand.INF the why it does ‘Without understanding the reason why he does it.’ (Tratado sobre certas duuidas da nauegação, Pedro Nunes, 16th c.)

b.

Já sabeis o como e o porquê Yet know.2PL the how and the why ‘You already know how and why.’ (Cartas familiares, Francisco Manuel de Melo, 17th c.)

The same interrogatives are attested in lower level nominalizations, equivalent to infinitival VP nominalizations: (32)

a. Busquemos o porque de todalas discordias Seek.1PL the why of all-the disagreements ‘Let us look for the reasons for all the disagreements.’ (Diálogos, Amador Arrais, 16th c.) b. sem se saber o como without refl know.INF the how ‘Without knowing how.’ (Décadas da Asia, João de Barros, 16th c.)

174

Crosslinguistic perspective: Romance languages I

The existence of makes the fact that is unattested even stranger, especially for the hypothesis that the mark of sentential nominality is the presence of an article. Que-clauses in historical Portuguese appear in typically nominal positions and thus must be Case licensed as such, including by a lexical preposition in adverbial clauses, or positionally where the expected preposition is absent. The absence of the article does not alter the nominality of the clause.

3.2 Present-day Portuguese The examples for present-day Portuguese come from scholarly works, native speakers, and, in several instances, from the Corpus de Referência do Português Contemporâneo (CRPC) of the Centro de Linguística at the Universidade de Lisboa, as adequately indicated. 3.2.1 Prepositions and nouns. Prepositions and infinitives In present-day Portuguese, preposition-taking categories such as PP-verbs and nouns or adjectives may combine with a (non-clausal) DP: (33)

a. Eu me lembro bem da primeira noite I me recall well of-the first night ‘I remember the first night well.’ (29-2-12) b. eu ainda insisto em nós I still insist in us ‘I still insist on us.’ (29-2-12) c. a ideia de colapso da zona euro the idea of collapse of-the zone euro ‘The idea of the collapse of the Euro Zone.’ (29-2-12) d.

Inquérito obriga a novo estudo geológico Inquiry forces to new study geological ‘Inquiry forces a new geological study.’ (29-2-12)

Portuguese

175

Portuguese shows several relevant additional concomitances with Spanish. Firstly, the presence of the preposition may provoke aspectual differences, as is the case with pensar vs. pensar em (‘to think’ vs. ‘to think of’): (34) a.

*Pensou uma resposta durante cinco minutos Thought.3SG an answer for five minutes Lit. ‘He thought an answer for five minutes.’

b.

Pensou numa resposta durante cinco minutos Thought.3SG in-an answer for five minutes ‘He thought about an answer for five minutes.’

Secondly, certain required prepositions may still be absent in the modern language; interestingly, this option is attested with the verb lembrar-se (‘to remember’), which coincides with the behavior of Spanish acordarse (‘to remember’). (35a) illustrates the standard syntactic use with the preposition, while (35b) shows the non-prepositional version: (35)

a. Eu me lembro de uma coisa muito boa I me recall of a thing very good ‘I remember a very good thing.’ (29-2-12) b. nem me lembro a idade que tenho nor me recall.1SG the age that have.1SG ‘Nor do I recall my age.’ (CRPC, text J64298)

The same required prepositions admit (inflected and non-inflected) infinitives as well (see Duarte 2003: 622–623): (36)

a. Alguém já se lembrou de fazer isto? Someone yet refl recalled of do.INF this ‘Did someone remember to do this already?’ (29-2-12) b. Greve obrigou a suprimir 66 comboios Strike forced to eliminate.INF 66 trains ‘Strike forced to eliminate 66 trains.’ (29-2-12)

176

Crosslinguistic perspective: Romance languages I

c. a ideia de fazer um filme the idea of make.INF a film ‘The idea of filming a movie.’ (CRPC, text noCOD_1035004) Duarte (2003: 401–402) includes these required prepositions in the group of Case-markers (that is, KPs). On the other hand, they do not qualify as prepositional complementizers, for they continue to be (usually) necessary when selecting other categories such as nouns or pronouns (37a) and because they cannot be substituted by a direct object pronoun such as o (‘it’), but rather by a PP (37b) (Duarte 2003: 614): (37)

a. O João insistiu nisso The John insisted on-that ‘John insisted on that.’ b. *O João insistiu-o The John insisted-it Lit. ‘John insisted it.’

It is possible to find certain infinitival contexts where the expected required preposition is absent: (38) a.

Eu costumava adorar jogar esse jogo . . . , eu me lembro I used-to adore.INF play.INF that game I me recall fazer também a pizza do.INF also the pizza ‘I used to adore playing thta game. . . I also recall making pizza.’ (29-2-12)

b.

não me lembro vez alguma ter sofrido uma not me recall.1SG time some have.INF suffered an observação disciplinar observation disciplinary ‘I do not remember ever having received a disciplinary remark.’ (CRPC, text J150408)

Lexical prepositions may introduce infinitives and form adverbial infinitival clauses, as the following examples illustrate (Whitlam 2011: 136):

Portuguese

(39)

177

a. Você faz o favor de esperar até eu terminar de falar? You do the favor of wait.INF until I finish.INF of speak.INF ‘Would you please wait until I finish talking?’ b. Ele não consegue emagrecer por não fazer exercício He not gets lose-weight.INF for not do.INF exercise ‘He does not manage to lose weight as he does not exercise.’

Portuguese and Spanish even share the existence of a special prepositional complementizer for certain declarative verbs, although more restricted in Spanish. Consider the following examples (Duarte 2003: 621): (40)

a. Os pais disseram aos miúdos para vir(em) The parents told to-the kids for come.INF.PL para casa cedo for home early ‘The parents told the kind to return home early.’ b. Os jornalistas pediram ao chefe de redacção para mandar The journalists asked to-the boss of redaction for send.INF um repórter ao Médio Oriente a reporter to-the Middle East ‘The journalists asked their editor-in-chief to send a reporter to the Middle East.’

Duarte (2003) concludes that para is a prepositional complementizer because it is only present with infinitives (Duarte 2003: 622): (41)

a. *Os pais disseram aos miúdos para que viessem para The parents told to-the kids for that came.3PL for casa cedo home early ‘The parents told the kind to return home early.’ b. *Os pais pediram aos miúdos para isso The parents asked to-the kids for this ‘The parents asked this from the kids.’

The syntax of non-clausal DPs and infinitives is very similar to that of Spanish, even to the detail of allowing for grammatical licensing without the usually

178

Crosslinguistic perspective: Romance languages I

necessary preposition. With Raposo (1987b), the evidence presented in the previous pages indicates that Portuguese DPs and infinitives – save those involved in any type of clausal reduction via restructuring or other clausal reduction processes – check their Case features, either lexical Case with a lexical preposition or inherently with the materialization of a functional preposition or just positionally.

3.2.2 Prepositional finite clauses Portuguese mirrors Spanish in admitting both argumental and adjunct prepositional que-clauses: (42) a.

Os miúdos lembram-se de que os pais lhes The kids remember-refl of that the parents them prometeram uma bicicleta promised a bike ‘The kids remember that their parents had promised them a bike.’ Duarte (2003: 600)

b.

Eles autorizaram-nos a que consultásemos o manuscrito raro They authorized-us to that consulted.1PL the manuscript rare ‘They authorized us to consult the rare manuscript.’ Duarte (2003: 602)

c.

O João insistiu em que fôssemos à festa dele The John insisted in that went.1PL to party of-him ‘John insisted that we go to this his party.’

d.

Duarte (2003: 614)

Toda a equipa se mostrou receptiva a que os dois jovens All the team refl showed receptive to that the two youths fossem contratados were hired ‘The whole team was receptive to the idea of hiring both youngsters.’ Duarte (2003: 603)

e.

a ideia de que a Terra é redonda the idea of that the Earth is round ‘The idea that the Earth is round.’

Duarte (2003: 614)

Portuguese

179

Duarte (2003: 614) argues for a

analysis of the traditionally labeled subordinating conjunctions, in keeping with the analysis presented for Spanish. Evidence in favor comes from the fact that the same prepositions can introduce the other type of finite clause we have been discussing, the indirect interrogative finite clauses: (43) a.

não me lembro de quem era not me recall.1SG of who was ‘I do not remember who he was.’

b.

(CRPC, text A168486)

Não me lembro de como chegamos à fazenda Not me recall.1SG of how arrived.1PL to farm ‘I do not remember how we arrived at the farm.’ (CRPC, text noCOD_1000024)

c.

à pergunta de se a CAP está ou não à espera desta votação the question of if the CAP is or not to wait of-this voting ‘The question of whether CAP is waiting on this voting or not.’ (CRPC, text A116755)

Like Spanish, the optionality of the required preposition – that is, the continuation of the original form – is grammatical in Portuguese too, both in the case of que-clauses (traditionally called queísmo) and in the case of indirect interrogative clauses, as the following examples illustrate (Duarte 2003: 619; see also Mollica 1995: 17 for queísmo): (44)

a. o senhor secretário de Estado ainda nos convence que The mister secretary of State even us convinces that pagar impostos faz bem à saúde pay taxes does well to health ‘The secretary of State even convinces us that paying taxes is good for our health.’ (Sábado, 17/12/88) b. desejosos que o convívio resultasse num sucesso wishful that the meeting resulted in-a success ‘Wishful that the meeting would become a success.’ (Expresso, 4/11/89) c.

o facto que a Foz “cresceu junto a o rio . . .” the fact that the Foz grew next to the river ‘The fact that Foz grew up next to the river.’ (O Jornal Ilustrado, 23/8/91)

Duarte (2003: 619)

180

Crosslinguistic perspective: Romance languages I

d.

Já não me lembro quem foi que disse que a culpa do Yet not me recall.1SG who was that said that the guilt of-the atraso de Viana era dos Deputados delay of Viana was of-the representatives ‘I do not remember anymore who was it that said Viana’s backwardness was caused by its political representatives.’ (CRPC, text noCOD_1011032)

e.

não me lembro se era PPD not me recall.1SG if was PPD ‘I do not remember if it was PPD.’

(CRPC, text A163999)

f. quase nem me lembro para que era almost not me recall.1SG for what was ‘I almost do not remember what it was for.’ (CRPC, text A158562) As expected, lexical prepositions can take que-clauses and form adverbial finite clauses, as the following examples illustrate: (45)

a.

Fugiste para que ele não te visse Fled.2SG for that he not you saw ‘You ran away so he could not see you.’

b.

Brito (2003: 717)

sem que isso seja o seu objectivo principal without that that is the his objective principal ‘Without that being his main objective.’ (CRPC, text noCOD_1008951)

c.

até que isso aconteça until that this happens ‘Until that happens.’

(CRPC, text J47844)

Bechara (1999) and Brito (2003) analyze these combinations as PPs as well. By arguing for a PP analysis, the generalization is maintained that the affected prepositions remain categorially the same regardless of the complement they introduced, which can be a noun, an infinitive, or a finite clause, as in Spanish. Consider the following examples clearly showing that por (‘for’) remains a preposition in all contexts: (46)

a.

Houve seca em Portugal em 1981, porque não choveu Was drought in Portugal in 1981, for-that not rained ‘There was drought in Portugal in 1981 because it didn’t rain.’ Brito (2003: 711)

Portuguese

b.

181

Por não ter chovido em Portugal em 1981, houve seca For not have.INF rained in Portugal in 1981, was drought ‘Due to a lack of rain in Portugal in 1981, there was drought.’ Brito (2003: 712)

c.

Não choveu em Portugal em 1981, por isso houve seca Not rained in Portugal in 1981, for that was drought ‘It didn’t rain in Portugal in 1981, for that reason there was drought.’ Brito (2003: 713)

Adverbial clauses may also include an indirect interrogative finite clause, also introduced by lexical prepositions: (47)

a.

Podemos julgar o caráter de um homem por como ele Can.1PL judge.INF the character of a man for how he trata aqueles que não podem fazer nada por ele? treats those that not can do.INF nothing for him ‘Can we judge somebody’s character depending on how they treat those who have nothing to offer them?’ (1-3-12)

b.

Eu amei Star Wars . . . por como ele criou seu mundo I loved S W for how he created his world ‘I loved SW thanks to the way he created his world visually.’ (1-3-12)

Another point in common between Portuguese and Spanish is the presence of dequeísmo, that is, the presence of the preposition de where it is not required by the selecting category (usually verbs of communication) and thus does not appear with any of their other possible complements (nouns, infinitives, etc.) (Mollica 1995: 45): (48)

está escrito no final de que há muito escrúpulo is written in-the end of that has much qualm dessas mulheres of-those women ‘It is written in the end that he has much qualm of those women.’

182

Crosslinguistic perspective: Romance languages I

Where Spanish and (European) Portuguese diverge is in the fact that Portuguese allows particular uses of functional prepositions present with DPs or infinitives but absent with finite clauses. This is the case with verbs such as gostar (‘to like’), necessitar or precisar (‘to need’) (Duarte 2003: 636–637). Observe the asymmetry of the following Portuguese examples: (49)

a.

O João gosta da Maria The John likes of-the Mary ‘John likes Mary.’

b.

O João gosta de tocar flauta The John likes of play.INF flute ‘John likes playing the flute.’

c.

Duarte (2003: 637)

O João gosta que a Maria toque flauta The John likes that the Mary plays flute ‘John likes Mary to play the flute.’

Duarte (2003: 636)

Duarte (2003: 637) finds the explanation in Case Theory and asserts that verbs such as gostar are case-defective, which means that the Case-marker will emerge before Case-carrying categories such as nouns, pronouns, and infinitives. However, in keeping with the stipulations of Case Theory and seeing that que-clauses are indeed grammatical as prepositional objects in Portuguese, Duarte’s explanation cannot force the conclusion that the finite clause in Portuguese is not nominal, as such a hypothesis is automatically challenged by que-clauses as prepositional objects, a typically nominal environment. If finite clauses must check Case as any other categories found in nominal contexts, so must the Portuguese ones in such configurations, even when there is no preposition. These special verbs impose selectional constraints similar to the situation attested for all PP-verbs (and other categories) before

became established as a general grammatical (and, even more, normative) option, a change which in any case did not challenge the nominality of the finite clause in the medieval period. As happened in the old language, extraction proves that the non-prepositional sentential complement is indeed an argument and not an adjunct: (50)

O que gosta que lhe perguntem num primeiro encontro? The what like.3SG that you ask.3PL in-a first meeting ‘What do you like to be asked in a first meeting?’

(1-3-12)

Portuguese

183

On the other hand, unlike European Portuguese, is grammatical in Brazilian Portuguese (Álex Amaral, Flávia Cunha, pc; see also Mollica 1995: 54 for gostar de que and necessitar de que), which indicates, firstly, that for those speakers for whom this combination is indeed grammatical the verb gostar has regularized its overt selectional requirements and, secondly, that the finite clause may be overtly Case-marked: (51)

a. Não gosto de que ninguém me diga que isso . . . é ruim Not like.1SG of that nobody me says that that is mingy ‘I do not like to be told that that is mingy.’ (1-3-12) b. gosto de trabalhar e não gosto de que falem like.1SG of work.INF and not like.1SG of that speak.3PL mal do meu filho badly of-the my son ‘I like to work and I don’t like it when people speak ill of my son.’ (1-3-12)

Another difference lies in the existence of a prepositional environment where the optional preposition com is only present with que-clauses when selected by the verb fazer (‘to make’), which looks syntactically similar to cases of dequeísmo (Cyrino, Nunes and Pagotto 2009: 70): (52)

ele fazia com que nós lêssemos . . . os livros he did with that us read.1PL the books ‘He made us read the books.’

3.2.3 Nominality and the clause Both infinitival – inflected and non-inflected2 – (53a,b,c) and finite clauses (53e–h) are grammatical in typically nominal contexts in Portuguese other than the prepositional objects already presented above (see Duarte 2003: 606–616; Madeira 1994 for the inflected infinitive, among many others): 2 The nature of the syntactic constraints ruling where and why speakers use or may favor one type of infinitive has attracted much attention throughout the years. A detailed contribution to this matter lies outside the scope of this book.

184 (53)

Crosslinguistic perspective: Romance languages I

a. É difícil recebermos a informação a tempo Is difficult receive.INF.1PL the information to time ‘It is difficult for us to receive the information on time.’ b. O João lamenta os pedreiros não terem concluído The John regrets the workers not have.INF.3PL finished a obra the work

‘John regrets that the construction workers have not finished the construction.’ c. Os pedreiros lamentan não ter concluído a obra The workers regret not have.INF finished the work ‘The construction workers regret not having finished the construction.’ Duarte (2003: 621) d.

Que o atendimento público melhore é uma necessidade That the assistance public improves is a necessity ‘It is necessary to improve the public assistance.’

e.

Foi uma surpresa que o filme tivesse ganho o festival Was a surprise that the film had won the festival ‘It was surprising that the film had won the festival.’ Duarte (2003: 602)

f. O João sabe que estamos à espera dele The John knows that are.1PL to wait of-him ‘John knows that we are waiting for him.’ g.

Duarte (2003: 608)

O João perguntou se íamos à festa ou não The John asked if went.1PL to party or not ‘John asked whether we were going to the party or not.’ Duarte (2003: 612)

h.

perguntou onde estavam os telescópios asked.3SG where were the telescopes ‘He asked where the telescopes were.’

(CRPC, text noCOD_1059966)

All these examples support Raposo’s (1987b) well-known argument that Romance infinitives require Case and are nominal, like their finite counterparts do too.

Portuguese

185

Continuing with the discussion on the topic in historical Portuguese (see for instance Quicoli 1996; Martins 2001, 2006), the literature on modern Portuguese infinitives argues for TP infinitives and full-fledged CP infinitives, both of which appear – other restrictions notwithstanding – in nominal positions. While noninflected infinitival clauses are considered CPs across the board (Duarte 2003: 623) – except for cases of clausal reduction of some sort (ECM, restructuring, etc.) – inflected infinitives have been usually divided into TP infinitives and CP infinitives. Raposo (1987a) (see also Duarte 2003: 625) concludes that there exists a categorial difference between, on one side, inflected infinitives complements of factive verbs (54a,b), in subject position (54c,d) and adjunct prepositional infinitives (54d), and, on the other side, infinitives complements of declarative and epistemic predicates (54f,g). While the former can be TP or CP, the latter always project a full CP, like non-inflected infinitives. The primary evidence for this classification comes from subject-verb word order inside the infinitival clause in combination with assumptions about Case checking. Let us examine the following examples: (54)

a. Eu lamento os deputados terem trabalhado pouco I regret the representatives have.INF.3PL worked little ‘I regret that the representatives have worked little.’ b. Eu lamento terem os deputados trabalhado pouco I regret have.INF.3PL the representatives worked little ‘I regret that the representatives have worked little.’ Raposo (1987a: 87) c. Eles aprovarem a proposta será difícil They pass.INF.3PL the proposal will-be difficult ‘For them to pass the proposal will be difficult.’ d.

Será difícil eles aprovarem a proposta Will-be difficult they pass.INF.3PL the proposal ‘It will be difficult for them to pass the proposal.’ Raposo (1987a: 95)

e.

A Maria entrou em casa sem os meninos ouvirem The Mary entered in house without the kids hear.INF.3PL ‘Mary entered the house without the kid hearing her.’ Raposo (1987a: 97)

186

Crosslinguistic perspective: Romance languages I

f. *O Manel pensa os amigos terem levado o livro The Manel thinks the friends have.INF.3PL taken the book ‘Manel thinks his friends have taken the book.’ g.

O Manel pensa terem os amigos levado o livro The Manel thinks have.INF.3PL the friends taken the book ‘Manel thinks his friends have taken the book.’

Raposo (1987a: 98)

Whenever the subject is preverbal, Raposo (1987a) and Duarte (2003: 628) claim that the infinitives project only up to TP. Both assume a mechanism of Case transmission from the selecting category (verb or preposition) into the embedded Tense, which, as a result, can license its own subject’s Nominative Case (case with personal pronouns), a sort of indirect ECM (see also Martins 2001). Since preverbal subjects are also grammatical in argumental prepositional inflected infinitival clauses, Duarte (2003: 628) concludes that they too must be TPs. Whenever the subject either is (54b) or must only be postverbal (54f), Tense – usually an auxiliary verb (see Madeira 1994: 182) – moves to C and thus the infinitives are CPs (Duarte 2003: 629–630). However, the mechanism of Case transmission adduced to account for TP infinitives is rather questionable, given that it rests on theoretical assumptions already superseded, which requires rethinking subject licensing. In addition, Mensching (2000: 79–80) argues against this classification on the basis that the proposed mechanism seems to be transferring only Nominative case even when the selecting predicates are, for instance, transitive verbs, which would call for Accusative case. The same comment can be extended to prepositions. As already mentioned when discussing Spanish, Mensching (2000) claims that preverbal subjects in Portuguese are located in Spec,AgrP (Mensching 2000: 154), requiring that the relevant inflected infinitives project past TP, at least up to AgrP. Additionally, Mensching (2000: 89–91, 117–118) also refutes T-to-C movement (his Aux-to-Comp) for CP infinitives, as does Ambar (1994, 1999: 8, fn 13), who still argues in favor of a CP analysis for inflected infinitives selected by the group of (at least) declarative and epistemic verbs while positing preverbal subjects in Agr as Mensching does. On her part, while maintaining Raposo’s basic T-to-C assumption for inflected infinitives complement to declarative and epistemic verbs, Madeira (1994) independently shows that it is plausible, nevertheless, to argue for a CP analysis for all inflected infinitives, including those complement to factive predicates via V-raising, not to C, but to another functional head located between V and C. Recently, Santos et al. (2011) argue for a general CP analysis in European Portuguese.

Portuguese

187

Finally, from a comparative perspective, Spanish non-inflected infinitives do accept preverbal subjects under certain circumstances, which implies that preverbal subjects must be licensable regardless of the degree of inflection present in TP and ideally independently from the differentiation between non-inflected and inflected infinitives. Portuguese infinitives may combine with determiners and participate in nominalization/substantivization at different levels (Brito 2010). Firstly, there are many examples of VP nominalizations, characterized by their lack of proper argument structure, the possibility of pluralization, combination with articles and other determiners, etc. Consider the following example (Brito 2010: 1): (55)

O gritar constante das pessoas The scream.INF constant of-the persons ‘People’s constant screaming.’

When the nominalization point is high, we may obtain examples of CP and/ or TP-nominalization, linked to the discussion above regarding the TP or CP nature of infinitival clauses. Observe the following example by Brito (2010: 2), who considers it to be a CP nominalization/substantivization: (56)

O ter ela escrito esses poemas The have.INF she written those poems ‘Her having written those poems.’

Given that the auxiliary verb is preposed, this example would nevertheless qualify as CP nominalization for those of the TP/CP divide for inflected infinitives, as Raposo (1987a) and Duarte (2003). Consequently, the same example with a preverbal subject would need to be a TP nominalization for them (at least as high as AgrP for Mensching 2000). Furthermore, (57) could be recast as CP nominalizations by taking into consideration the evidence in Madeira (1994) or the categorial assumptions in Santos et al. (2011). In agreement with Ramírez’s (2003) tripartite classification for Spanish, Brito (2010: 2) establishes a further third type, which she considers to be a TP nominalization. Consider the following example (Brito 2010: 2, which she adapts from Ramírez 2003): (57)

O teu contínuo conduzir camiões imprudentemente The your continuous drive.INF trucks carelessly ‘Your continuously driving trucks carelessly.’

188

Crosslinguistic perspective: Romance languages I

Given the presence of patently nominal properties such as the possessive instead of a Nominative subject and adjectival modification, (57) cannot be display nominalization/substantivization at as high a point in the derivation as in (56). A TP level nominalization analysis for (57) remains problematic as, firstly, it contains no subject position – crucial for a TP nominalization in Kornfilt and Whitman’s approach – as it is expressed as a PP; and, secondly, typical nominal modification is allowed. Both are properties found in the Italian examples adduced by Kornfilt and Whitman (2011) as models of vP nominalizations. Brito (2010: 2) states that these infinitives may accept auxiliary verbs, arguably projected in the TP area. However, she finds that context grammatical but unacceptable (one question mark), which seems to suggest that the presence of the auxiliary verb is forcing the creation of a verbal syntactic projection too high. In Spanish such combinations also vary in grammaticality/acceptability (see (23g,h,i) in section 3 in chapter 4). Assuming a higher verbal projection than just vP to accommodate the auxiliary verb, possibly T as Brito (2010) suggests, would imply that TP level nominalizations can accept adjectival modification, contra Kornfilt and Whitman (2011). What is more, the same problem will reappear in Italian. Unlike Kornfilt and Whitman (2011), I have argued that the article itself is not the nominalizing category, as is particularly clear for que-clauses, whose distribution is nominal whether introduced by an article or not (see Roussou 2010: 587). In the discussion of Spanish I concluded that those clauses not introduced by an article are simply CPs, headed by a nominal complementizer, and that as CPs they must be able to have their Case checked. While proposing an empty D might have seemed a plausible answer, Portuguese turns out to provide additional important evidence of the contrary. Given that infinitives can participate in nominalization/substantivization at different levels, the immediate expectation would be to find Portuguese queclauses behaving like their infinitival counterparts (contra Madeira 1994: 192, fn 10), as happens in Spanish. However, such is not the case, for que-clauses may not be introduced by an article, that is, they cannot be re-nominalized or substantivized, which goes on to prove that a D (and DP) is not mandatory or even necessary for indicating nominality (see also Barra 2002): (58)

*O que ele fale português The that he speaks Portuguese ‘The fact that he speaks Portuguese.’ (Patrícia Amaral, pc)

The only way of rescuing this sentence is by including o facto que (‘the fact that’):

Portuguese

(59)

189

O facto de que ele fale português The fact of that he speaks Portuguese ‘The fact that he speaks Portuguese.’

The infinitives pattern with the que-clause in allowing o fa(c)to de que as well. Consider the following examples of infinitives with article and with o fa(c)to: (60)

a.

O (tu) criticares-me o tempo todo entristece-me The (you) criticize.INF.3SG-me the time all saddens-me ‘The fact that you criticize me all the time makes me sad.’ Duarte (2003: 630)

b. O facto de (tu) me criticares o tempo todo entristece-me The fact of (you) me criticize.INF.3SG the time all saddens-me ‘The fact that you criticize me all the time makes me sad.’ Duarte (2003: 631) The grammaticality of (60) makes the ungrammaticality of (58) even more unexpected and salient, particularly noticing that the very same semantic restriction operates in Portuguese whereby the article is only grammatical in factive/presupposed contexts (Duarte 2003: 630), with equal extraction differences, possible without the article and blocked with the article (Duarte 2003: 631), as in Spanish. Interestingly, as happened for some speakers of Spanish (but not in the standard language), (inflected) infinitive clauses may be prepositional objects when introduced by an article, as the examples below prove: (61) a.

não tem a ver com o não terem conquistado not has to see.INF with the not have.INF.3PL conquered medalhinhas medals ‘It has nothing to do with not having gotten any medals.’ (3-3-12)

b.

me preocupo imenso com o eles terem mudado me worry.1SG immense with the they have.INF.3PL changed de País of country ‘I am very worried about them having switched countries.’ (3-3-12)

190

Crosslinguistic perspective: Romance languages I

(61a) is grammatical in European Portuguese, while (61b) seems to be unacceptable for some speakers (Patrícia Amaral, pc); however, these same sentences are judged ungrammatical by Brazilian Portuguese speakers (Flávia Cunha and Álex Amaral, pc). The configuration

is in any case not grammatical across the board in European Portuguese either, as the examples in (62) illustrate (Patrícia Amaral, pc). Note that the ungrammaticality does not correlate with the lexical/functional divide, as both the adverbial clauses introduced by the lexical preposition por (‘for’) and the clauses introduced by the functional preposition de (‘of’) are judged ungrammatical. The Brazilian example (62g) is considered ungrammatical, while (62h) is judged as more acceptable. Interestingly, my Brazilian informants reject both (Flávia Cunha and Álex Amaral, pc):3 (62)

a. *Eu fiz isso por o eles serem meus amigos I did that for the they be.INF.3PL my friends ‘I did this because they are my friends.’ b. Eu fiz isso por eles serem meus amigos I did that for they be.INF.3PL my friends ‘I did this because they are my friends.’ c. *Eu fiz isso por o serem meus amigos I did that for the be.INF.3PL my friends ‘I did that because they are my friends.’ d.

Eu fiz isso por serem eles meus amigos I did that for be.INF.3PL they my friends ‘I did that because they are my friends.’

e.

*O facto do vir à festa The fact of-the come.INF to party ‘The fact of coming to the party.’

f. O facto de vir à festa The fact of come.INF to party ‘The fact of coming to the party.’ 3 Not to be confused with examples where o is a direct object clitic, as in the following example: (1) sem o eles saberem without it they know.INF.3PL ‘Without them knowing it.’ (4-3-12)

Portuguese

g.

191

caracterizado pelo facto do terem os garfos . . . characterized for-the fact of-the have.INF.3PL the forks Lit. ‘Characterized by the fact of the forks having . . .’

(3-3-12)

h. sabendo das coisas depois do elas terem ocorrido knowing of-the things after of-the they have.INF.3PL occurred ‘Finding out about things after they have happened.’ (3-3-12) Thus, while the grammaticality of the article with prepositional inflected infinitives is attested for certain speakers in certain cases, it seems to be reduced or infrequent at the very least and subject to dialectal variation. The elimination of the article restores grammaticality (which is what we find in standard Spanish as well, with factive/presupposition restrictions). It could be the case that articles and finite clauses are incompatible in Portuguese for independent reasons, but this hypothesis is immediately falsified once indirect interrogative finite clauses are brought into the picture. Consider the following examples of indirect interrogatives with article (CP nominalizations/ substantivizations): (63) a.

é difícil esquecer o como e o quando apareceu is difficult forget.INF the how and the when appeared a primeira menstruação the first menstruation ‘It is difficult to forget how and when the first menstruation happened.’ (CRPC, text R1007)

b.

o quanto a nossa sociedade enferma em termos de the how-much the our society sickens in terms of valores humanos values human ‘How much our society becomes sick in terms of human values.’ (CRPC, text J164378)

c.

o porque o governo da Aliança Democrática optou the why the government of-the Alliance Democratic opted por um modelo diferente for a model different ‘Why the government of the AD opted for a different model.’ (CRPC, text A160444)

192

Crosslinguistic perspective: Romance languages I

d.

me pergunto o porque de algumas coisas terem acontecido me asked.1SG the why of some things have.INF.3PL occurred (3-3-12)

e.

Quanto à pergunta de se defendemos ou não uma As-for the question of if defend.1PL or not an economia social do mercado economy social of-the market ‘Regarding the question of whether we defend or not a social market economy.’ (CRPC, text A5811)

Indirect interrogative finite clauses may be prepositional objects with an article, with varying degrees of grammaticality and/or acceptability for different speakers, as in Spanish. The example (64d), with

, is ungrammatical in European Portuguese; with no article

is perfect (64e) (Patrícia Amaral, pc). My Brazilian informants vary in their judgments although they find the presence of the article either totally ungrammatical in (64a–d) (Flávia Cunha, pc) or oscillating from somewhat acceptable (but better without the article) in (64a,c) to plainly ungrammatical in (64b,d) (Álex Amaral, pc). Both agree that the absence of the article is grammatical (64e): (64)

a.

a pergunta sobre o porque o poder se tornou tão maléfico the question about the why the power refl turned so evil ‘The question about the reasons why power became so evil.’

(5-3-12)

b.

cuidar do porque escolheu formato angulos aqui care.INF of-the why chose.3SG format angle here ‘To take care of the reason why he chose an angle format here.’ (CRPC, text L0439P0011Xstar)

c.

as verdadeiras razões do porquê se julga existir the true reasons of-the why refl judge.3SG exist.INF em Angola um serviço de telefonia móvel celular deficiente in Angola a service of telephone mobile cellular deficient ‘The true reasons as to why the Angolan mobile phone service is deemed to be deficient.’ (CRPC, text J146385)

Portuguese

d.

193

respondendo a pergunta, do se eh possivel criar answering the question of-the if is possible create.INF uma obra prima sem sofrimento a work prime without suffering ‘Answering the question whether it is possible to create a mater piece without suffering.’

(5-3-12)

e.

por que persiste a pergunta de se a Rússia pertence à Europa? why persists the question of if the Russia belongs to Europe ‘Why does the question whether Russia is in Europe still persist? (5-3-12)

All these examples supply evidence against any special constraint against for certain speakers. Additionally, they also provide comparative support for the fact that substantivized indirect interrogative finite clauses are not subject to the factive/presupposed semantic constraint and can participate in different types of levels of nominalization. Porque and como can be nominalized at a lower level, as in Spanish, as the following examples illustrate: (65)

a. lhe é impossível explicar o como e o porquê da sua him is impossible explain.INF the how and the why of-the his opção de vida option of life ‘It is impossible for him to explain the manners and reasons of his life style.’ (CRPC, text J53577) b. Me pergunto o porque disso Me ask.1SG the why of-that ‘I wonder about the reasons for that.’ (3-3-12)

3.3 Discussion and conclusions Historical Portuguese and historical Spanish show many concomitances when it comes to the emergence of prepositional argumental finite clauses. In both cases, the available documentation shows that the non-prepositional variant

194

Crosslinguistic perspective: Romance languages I

was the most frequent option, but not the only one, as the attestation of some few early examples indicates. In both languages, it is possible to find infinitives and non-clausal DPs licensed without the presence of the otherwise required preposition, which suggests that arguments could indeed be licensed positionally, and such licensing mechanisms should also be available for nonprepositional que-clauses. Furthermore, the early attestation of prepositional adverbial clauses highlights that que-clauses were nominal then as they are nowadays, given that they had to check their lexical Case, according to the premises of current Case Theory. Present-day Portuguese and present-day Spanish, in all their vast dialectal richness, present many common options in the syntax of prepositional clauses. In both languages, lexical and functional prepositions can take finite clauses, and thus check their Case features as nominal projections. In addition, both permit the absence of the preposition in certain circumstances – the continuation of the original form – usually in colloquial speech. Extraction tests indicate that, as in Spanish, those non-prepositional varieties are not syntactically adjuncts, contra the expectations arising from Barra’s (2002) work. The preference toward using inflected infinitives instead of queclauses makes the Portuguese data somewhat different from that of Spanish, but, in any case, extraction does prove that the non-prepositional variants are arguments as extraction is possible: (66)

a. Diz-nos o que te lembras ter-te marcado mais Tell-us the that you recall.2SG have.INF-you marked more a nível emotivo to level emotive ‘Tell us what you remember having marked you the most at an emotional level.’

(3-3-12) b. O que me lembro que era . . . foi o sentimento muito The that me recall.1SG that was was the feeling very profundo da desigualdade deep of-the inequality ‘What I recall it being was the very deep feeling of inequality.’ (CRPC, text L0275P0011Xstar) c. a festa, à qual o João insistiu que fôssemos the party the which the John insisted that went.1PL ‘The party which John insisted we went to.’ (Patrícia Amaral, pc)

Portuguese

195

The fact that both languages share other special uses of prepositions not found in several other Romance languages, such as the phenomenon of dequeísmo, highlights the existence of comparable (probably analogical) processes operative for their speakers beyond the common existence of prepositional finite clauses in both argumental and adjunct/adverbial contexts, which in itself sets them apart from non-Iberian Romance languages (to some extent). The same conclusions reached in Spanish can be applied to Portuguese regarding the nature of nominal complementizers – whether que, si or empty for indirect interrogative finite clauses – with additional support from Manzini and Savoia (2005, 2011), Manzini (2010), and Roussou (2010). Infinitival clauses, being equally distributionally nominal throughout the history of Portuguese and currently, require a similar nominal category or feature for proper licensing and Case checking. In the case of clearly CP infinitives – whether inflected or noninflected – we can extend the suggestions raised for Spanish and assume an (empty) nominal complementizer, representing the functional nominal category Kornfilt and Whitman (2011) mention, without having to recur to additional DP layers with an empty determiner. Following Madeira (1994) and Santos et al. (2011), all infinitival clauses could be homogenously treated as CPs. For an analysis along the lines of the differentiation between TP and CP inflected infinitival clauses in Raposo (1987a) and Duarte (2003), Raposo (1987a: 95) himself argues that Inflection (=Tense) is essentially nominal, which can be reinterpreted as carrying the necessary [+nominal] feature. The result is, however, a truly mixed category, violating categorial coherence, as pursued in Kornfilt and Whitman (2011). Yoon’s (1996) null nominal affix for infinitival clauses would also be extended to Portuguese. A final possibility would be to argue for a functional nominal category topping TP infinitival clauses, a hypothesis coherent with the fact that an nP projection in VP and vP nominalizations had to be already added to account for nominal modifiers. In any case, the main conclusion is that it is agreed that Portuguese infinitival clauses – equivalent to NPs – were and are nominal and must check Case. If the derivation adds a determiner in order to re-nominalize/substantivize, then a DP tops the clause. There are also particularly relevant differences with theoretical implications. First of all, as has just been pointed out, Portuguese speakers can use an inflected infinitive which appears in many contexts where Spanish speakers would require a finite que-clause. Secondly, the ungrammaticality of articles with que-clauses in Portuguese comes as a big surprise, particularly compared to the grammatical possibilities with inflected infinitives and indirect interrogative finite clauses. What both languages prove is that clausal nominality is independent from clausal nominalization/substantivization. The inflected infinitival seems to occupy the position

196

Crosslinguistic perspective: Romance languages I

reserved for the que-clause in Spanish combining with determiners. Portuguese indirect interrogative clauses mirror their Spanish counterparts as well. Thirdly, as expected of different languages, there exist some differences, such as the ungrammaticality of gostar de que in European Portuguese, or the ungrammaticality of

in many varieties of Spanish. However, even in these cases there are concomitances, as there exist speakers of Portuguese, especially in Brazil, who have extended the preposition to finite clausal complements and find gostar de que perfectly grammatical. Likewise, there exist speakers of Spanish (in Mexico or Puerto Rico, for instance) for whom prepositional infinitives with an article are an option. In conclusion, the history and present state of the Portuguese language provides extensive crosslinguistic support in favor of the hypothesis that the emergence of prepositional argumental finite clauses is not the result of changes in the nominality of the que-clause (contra Barra 2002).

Chapter 6

Crosslinguistic perspective: Romance languages II 1 Introduction This chapter explores the syntax and history of prepositional finite clauses of French and Italian, with attention to the past and the present-day. Non-clausal DPs and infinitives are also explored as appropriate. Comparing the syntax and history of these languages offers a perfect scenario to examine and discover what is common, what is different – especially what is unexpectedly different – and what the theoretical implications are.

2 French 2.1 Historical French Historical French is used here as a general term for older French. The data analyzed in this section range from the 12th century up to the 17th century. The examples were obtained from scholarly works, indicated in each case. The rest of the historical examples were extracted from the Modéliser le changement: les voies du français corpus (MCVF), an electronic corpus developed as part of the GTRC project directed by Dr France Martineau, from the Département de français at the Université d’Ottawa/University of Ottawa, Canada.

2.1.1 Prepositions and infinitives As in the other Romance languages examined so far, infinitives could be introduced by a preposition in historical French (see Moignet 1973: 297–301, 304, 313; Pearce 1990: 226–238; van Reenen and Schøsler 1993; Buridant 2000: 394, 467; Martineau 2000, 2002). Many of these cases, in particular in this period, qualify as infinitives introduced by prepositional complementizers, that is to say, they are CPs, not PPs. Other cases, especially those introduced by clearly PP-verbs taking de, must be categorized as PPs, such as parler de (‘to speak of’), as in (1d,e), where the (true) preposition de is maintained both with an infinitival complement and a (non-clausal) DP complement (see Pearce 1990:

198

Crosslinguistic perspective: Romance languages II

234; see also Moignet 1973: 303–304). Restructuring, especially frequent with , was common before the 17th century (see Pearce 1990; Martineau and Motapanyane 2000; Martineau 2002). Consider the following examples: (1) a.

D’euls disarmer ne s’oublia Of-them disarm.INF not refl-forgot.3SG ‘He did not forget to disarm them.’ (Conquête de Constantinople, Robert de Claris, 13th c.)

b.

se je tendoie a avoir la if I tried to have.INF it ‘If I tried to have it.’

c.

(Queste del Saint Graal, 13th c.) Pearce (1990: 237)

Mult se fait fiers de ses armes porter Much refl makes proud of his arms carry.INF ‘He becomes very proud to bear his arms.’ (Roland, 11th c.) Pearce (1990: 236)

d.

sans plus jamais parler de tuer without more never speak.INF of kill.INF ‘Without ever again speakimg about killing.’ (Cent Nouvelles Nouvelles, 15th c.)

e.

nous retournerons a parler de la vieille we will-return to talk.INF of the old ‘We will talk again about the old lady.’ (Cent Nouvelles Nouvelles, 15th c.)

Prepositional optionality is attested as well (see Haase 1965: 281–294, 327– 338; Ménard 1988: 164–165; Moignet 1973: 297–299; Martineau 2000, 2002), with different degrees of optionality and variation in different periods (Martineau 2002: 75) and with a noted increase of the use of de over time (Pearce 1990: 238; Buridant 2000: 393, 400, 467–468, 473–474), syntactically linked to the loss of restructuring (Martineau 2000: 177; Martineau and Motapanyane 2000: 226; Martineau 2002: 80–81). Observe the following examples with verbs allowing multiple selection (be it truly prepositional or involving prepositional complementizers), like their equivalents in Spanish (van Reenen and Schøsler 1993: 528):

French

(2)

199

a. pensent Ø vivre touz iours think.3PL live.INF all days ‘They think they are going to live forever.’ (Le sermon, Guischart de Beaulieu, 12th c.) b. que il n’i pensent pas a atteindre that they not-there think not to reach.INF ‘That they do not think to arrive there.’ (Miracles de Notre Dame de Chartres, Jean le Marchant, 13th c.) c. pensa tost de moi sivre thought.3SG soon of me follow.INF ‘He thought to follow me soon.’ (Roman de la rose, Guillaume de Lorris, 13th c.)

These examples illustrate that historical French shows equivalent patterns of variability as those attested in Spanish. Darmesteter (1931: 140, 4th book; see also Haase 1965) indicates that this variability persists as late as into the 17th century, where the otherwise expected preposition or prepositional complementizer may be absent (Darmesteter 1931: 140, 4th book): (3)

a. Pour quoy est necessaire se pourveoir d’amys For what is necessary refl obtain.INF of-friends ‘For which it is necessary to obtain friends.’ (Mémoires, Commines, 15th c.) b. Le plus aspre . . . mestier . . . c’est faire dignement le roy The more rough work is do.INF decently the king ‘The rougher job is to be a decent king.’ (Essais, Montaigne, 16th c.)

Prepositional optionality is likewise attested with certain non-clausal DPs, as illustrated in Buridant (2000: 394–395) (see also Darmesteter 1931: 307, 4th book). Additionally, Siminenko (2010) studies the progressive elimination of the non-prepositional genitive, maintained as late as the 15th century, and its substitution for the prepositional one. In other words, the licensing of nouns and infinitival clauses must have included a direct, non-prepositional option in cases where such option alternates with a more frequent prepositional one, again as in Spanish, which provides a background syntactic configuration extendable to finite clauses.

200

Crosslinguistic perspective: Romance languages II

Another point in common with Spanish is the existence of infinitives in subject position introduced by prepositional complementizers (Darmesteter 1931: 304, 4th book; see also Haase 1965: 282–283 for examples from the 17th century) (Darmesteter 1931: 304, 4th book): (4) a.

D’amer est mervilleuse cose Of-love.INF is marvelous thing ‘To love is a wonderful thing.’ (Amadas et Ydoine, 13th c.)

b.

folie est de trespasser le vostre comandement madness is of trespass.INF the your order ‘It is crazy to violate your order.’

(Queste del Saint Graal, 13th c.)

Furthermore, historical French also presents infinitival clauses with overt subjects: (5)

a. Quel joie a cascun Est d’habiter frères en un What joy to each is of-live.INF brothers in one ‘What joy it is for everybody that brothers live in unity.’ (Le voyage de St. Brandan, 12th c.) Mensching (2000: 120) b. jusques à en estre souvent l’âme désemparée du corps until to of-it be.INF often the-soul abandoned of-the body ‘Until the soul was frequently abandoned by the body.’ (Le quart livre des faicts et dicts héroiques du bon Pantagruel, Rabelais, 16th c.) Mensching (2000: 122)

As expected, lexical prepositions could take infinitival clauses and form adverbial clauses: (6)

a. A lui tranchier mar querras fer To it cut.INF tort will-want.2SG iron ‘In order to cut it you will need iron.’

(Roman d’Eneas, 12th c.) Ménard (1988: 166)

b. Après mangier sont alé reposer After eat.INF are gone rest.INF ‘After eating, they went to rest.’

(Le Couronnement de Louis, 12th c.)

French

201

c. Ne veintroiz mie par foïr Not will-win.2PL nothing for flee.INF ‘You will not win by fleeing.’ (Roman d’Eneas, 12th c.) Ménard (1988: 167) Infinitival clauses in historical French can appear in typically nominal positions, such as prepositional object – both with functional and lexical prepositions, as already shown – subject and (direct) object positions: (7)

a. Tencier, fet il, est laide chose Argue.INF did he is hateful thing ‘Arguing, said he, is a bad thing.’ (Roman de Thèbes, 12th c.) b. Ja vos cuidions nos trover a Kamaalot Yet you think us find.INF in Camelot ‘You already think about meeting us in Camelot.’ (Queste del Saint Graal, 13th c.) c.

Moignet (1973: 196)

envoierent messages a Cesar por proier le de pes sent.3PL messages to Caesar for beg.INF him of peace ‘They sent messages to Caesar to beg for peace.’ (Fet des Romains, 13th c.)

Buridant (2008: 43)

Thus, historical French is similar to historical Spanish in having (distributionally) nominal infinitival clauses, arguably CPs (AgrPs in Mensching’s 2000 terms), with the exception of restructured infinitives or otherwise clausal reduction cases, which in any case are reported to be rather frequent, as happened with Portuguese, for instance, and which are licensed necessarily differently (see Cinque 2006). Their distributional nominality can be captured, as indicated for Spanish and Portuguese, by arguing for a functional nominal category carrying the [+nominal] feature necessary to properly license the relevant infinitival clauses in nominal positions and carrying the associated Case feature. The possibilities indicated for Spanish and Portuguese are logically available for French as well. In addition to their being distributionally nominal, infinitives could be introduced by a determiner in historical French and participate in nominalization/ substantivization, as Buridant (2008) extensively evidences. Consider the following examples:

202

Crosslinguistic perspective: Romance languages II

(8) a.

pour les dous boires que je leur donnai el desert for the two drinks.INF that I them gave the dessert ‘And for the two drinks that I gave them for dessert.’ (Le Livre du Graal, 13th c.) Buridant (2008: 46)

b.

doubtant le retourner de son adversaire doubting the return.INF of his enemy ‘Afraid of the return of his enemy.’ (Cent Nouvelles Nouvelles, 15th c.) Buridant (2008: 179)

Buridant (2008) clearly shows that in historical French infinitival nominalizations/substantivizations are attested, a configuration which is less frequent in the modern language (see also Ménard 1988: 159). The examples in (8) above all illustrate what Buridant labels as lexicalized nominalized infinitives (Buridant 2008: 44). These infinitives admit all kinds of essentially nominal properties, such as a variety of determiners, pluralizations, adjectival modification, and even case morphology (Buridant 2008: 44–82). Following Kornfilt and Whitman (2011), these examples qualify as VP nominalizations. Buridant (2008) establishes a second group of infinitival nominalizations labeled non-lexicalized nominalized infinitives (Buridant 2008: 82). It is in this group where we find examples of nominalizations at least up to the vP level – which can be introduced by prepositions –. The examples in (9a,b) show that nominalized/substantivized infinitives that can be modified by an adverbial or even license a direct object. Assuming a PRO subject, they could even be regarded as high as CP or at least TP nominalizations;1 should similar examples express the subject in the form of a possessive, then they would illustrate vP level nominalizations more clearly. (9c), with its subject in the form of a possessive and adverbial modification of the event expressed by the infinitive, qualifies as a vP nominalization:2 1 Sleeman (2010: 156) argues that Old French nominalized infinitives could project up to AspP, thus at least above TP. 2 Sleeman (2010: 156–157) highlights the existence of some nominalizations which seem to be a mixed case in a dual division nominal vs. verbal infinitives: (1) Li dormirs atempreement The sleep.INF.NOM moderately ‘Sleeping moderately.’ These nominalizations maintain the case morphology typical of nouns in the period but allow for direct objects and adverbs, not expected in nouns. Kornfilt and Whitman’s (2011) classification and description of nominalizations helps understand where to locate these examples; the point of nominalization must be high enough to allow for a vP, which lies in between the

French

(9)

203

a. Il n’i a que dou belement Aler ariere It not-there has that of-the voluntarily go.INF behind ‘We can simply go back.’ (Guillaume de Dole, 13th c.) b. Si n’i ot que del avaler Le pont Thus not-there remains that of-the descend.INF the bridge ‘The only missing thing is to lower the bridge.’ (Yvain, Chrétien de Troyes, 12th c.) c.

Buridant (2008: 85)

pour vostre chevauchier a tous les jours for your horse ride.INF to all the days ‘For your horse riding everyday.’

(Ménagier de Paris, 14th c.) Buridant (2008: 197)

In short, the data indicate that historical French was similar to Spanish in allowing (distributionally) nominal and nominalized/substantivized infinitival clauses, and prepositional infinitival clauses introduced by functional prepositions required by the selecting category (argumental clauses), by prepositional complementizers (and thus CPs, not actual PPs) – including subject clauses introduced by de – and by lexical prepositions heading adverbial clauses (adjunct clauses). In all typically nominal contexts, the infinitival clause must have checked Case, be it inherent, lexical or structural Case.

2.1.2 Prepositional finite que-clauses. The ce element This section will focus particularly on historical French sentential complementation and adjunction involving combinations of prepositions and que-clauses. The use of alternatives with ce (‘this’) in prepositional que-clauses is also of particular interest. Let us first review certain relevant characteristics also shared with historical Spanish. Consider the following examples:

more nominal VP and the more verbal TP/CP nominalization. Therefore, these special infinitives can be vP nominalizations which maintain a typical feature of Old French, as a result of the derivation or rather by analogy with VP nominalizations, assuming that these examples were exceptional. The maintenance of this special nominal feature is not that unexpected once we take into consideration that vP nominalizations manifest other nominal characteristics, such as adjectival modification, presence of a variety of determiners, etc.

204 (10)

Crosslinguistic perspective: Romance languages II

a. Nuncent al pedre que nel pourent truver Announce.3PL to-the father that not-him could.3PL find.INF ‘They announce to the father that they could not find him.’ (Vie de Saint Alexis, 11th c.) b. Je sai bien ou tu vas et que tu demandes I know well where you go and what you ask ‘I know very well where you’re going, what you’re asking.’ (Sept sages de Rome, 13th c.) Buridant (2000: 693) c.

Fuiant s’en vint, qu’il n’i pout mes ester Fleeing refl-of-there came, that-he not-there could more stay.INF ‘He fled because he could not resist any longer.’

d.

(Roland, 11th c.)

Prist l’olifan, que reproce n’en ait Took.3SG the-orphan that blame not-of-it has.3SG ‘He took the orphan to avoid all blame.’ (Roland, 11th c.) Buridant (2000: 565)

e.

m’en sui aperceüe Que j’en ait été deceüe me-of-it am aware that I-of-it have been deceived ‘But I have found out today that I have been deceived.’ (Vergy, 13th c.) Buridant (2000: 574–575)

f. il a grant poor de sa desloiauté qu’ele ne li nuise He has great fear of his disloyalty that-he not him damages ‘He has great fear that his disloyalty would bring him problems.’ (Mort le roi Artu, 13th c.) Buridant (2000: 576) g.

Ke vols que jo te face? What want.2SG that I you do.1SG ‘What do you want me to do to you?’ (Quatre livres des rois, 12th c.) Rouquier (2003: 346)

h.

fault. . . parler comme le roy estoit allé en Bourbonnoys must talk.INF how the king was gone in B. ‘It is necessary to mention a little that the king had gone to B.’ (Mémoires, Philippe de Commynes, 15th c.)

(10a) serves as good example of a que-clause in a typically nominal position (the direct object), thus supporting the nominality of the clause (ce que-clauses were also possible, see Rouquier 1990). The same nominal distribution applies in

French

205

the case of indirect interrogative clauses (10b). (10c,d) show that the que-clause was polyfunctional in historical French too; particularly, these examples pinpoint the role of pragmatics, and the fact that adverbial clauses could be nonprepositional (see Jensen 1974; Bertin 1997: 41–43; Buridant 2000: 565). (10e) shows an example of a non-prepositional que-clause nevertheless substituted by the PP-proform en. (10f ) illustrates the phenomenon of prolepsis with the preposition de taking the DP sa desloiauté rather than the que-clause. (10g) shows that, despite the role of pragmatics in subordination, as in Spanish, argumental clauses are attested as early as the 12th century, as extraction out of the embedded clause proves. (10h) illustrates the use of comme (‘how’) in declarative sentences. All these examples and constructions are attested in historical Spanish. The nominal complementizer as the necessary functional nominal category to account for proper nominal licensing of finite clauses argued for Spanish can be extended to French as well. Let us move on to the use of de ce que and à ce que, as these will be more important in the discussion of argumental clauses in present-day French. Herman (1963: 223) notes that the form de ce que (lit. ‘of this that’) is already attested in the oldest French texts, especially frequent from the 12th century on, precisely in those contexts where de is selected by a verb or combines with any other category. Observe the following examples: (11) a.

Guenes est fels d’iço qu’il le traït Guenes is furieux of-this that-he him betrayed ‘Guenes is furious that he betrayed him.’

b.

(Roland, 11th c.)

Achimelech s’esmerveillad de çó que David vint si sultifs Achimelech refl-marvelled of this that David came so alone ‘Achimelech marveled that David came on its own.’ (Li quatre li bre des reis, 12th c.) Herman (1963: 223)

c.

la novele . . . de ce que ele amenoit le lyeon et le chevalier the news of this that she brought the lion and the knight ‘The news that she was bringing the lion and the knight.’ (Yvain, Chrétien de Troyes, 12th c.)

d.

dolenz de ce que ge ne puis veoir Lancelot hurt of this that I not can see.INF Lancelot ‘Hurt that I cannot see Lancelot.’

(Mort le roi Artu, 13th c.)

The example in (12) illustrates the fact that de ce que could be separated and form a correlation:

206

Crosslinguistic perspective: Romance languages II

(12) De ceo li semblot grant merveille k’il la veeit blanche Of this him seemed great marvel that-he her saw white ‘He was amazed at seeing her white.’ (Lais, Marie de France, 12th c.) Jensen (1974: 48) Jensen (1974: 48, 1990: 503) states that de ce que was common with verbs of emotion, adding a strong causal connotation (see also Darmesteter 1931: 192, 4th book; Haase 1965: 275–276), as has also been discussed in Spanish and Portuguese. Consider the following example (Jensen 1990: 503): (13) mut li aveit pesé de ceo qu’il l’ot si cungeé much him had weighed of this that-he her-had thus left ‘He had suffered much because he had left her this way.’ (Chievrefoil, Marie de France, 12th c.) Notice, however, that this strong connotation is possible with some examples in (11) but not in all of them, for instance, (11c). Furthermore, the causal interpretation is possible in historical Spanish (Herrero 2005: 92) and even in present-day Spanish, as can be seen in Me alegro de que vengas (‘I am happy that you are coming’) being easily reinterpreted as Me alegro porque vienes (‘I am happy because you’re coming’) despite clear syntactic differences between the two. A similar ambiguity may be obtained with (certain) contexts of à ce que, as in historical and present-day Spanish a que-clause. The same predicates that combine with de ce que are attested being complemented directly by a que-clause while maintaining the causal connotation, again as in Spanish: (14)

a. je sui dolenz que tu ies si maubailliz I am hurt that you are thus damaged ‘I am hurt that you are so damaged.’

(Queste del Saint Graal, 13th c.)

b. la novele Que venue est . . . La plus bele feme the news that come is the most pretty woman ‘The news that the prettiest woman has come.’ (L’Escoufle, Jean Renart, 13th c.) Indirect interrogative finite clauses can be objects of PP-verbs such as se mervellier (‘to marvel’) without the preposition (see Kunstmann 1990: 114, who finds no prepositional indirect interrogative finite clauses in his Old French corpus):

French

(15)

207

se mervellierent que li empereur voloit faire refl marveled.3PL what the emperor wanted do.INF ‘They were amazed about what the emperor wanted to do.’ (La Conqueste de Constantinople, Robert de Clari, 13th c.)

À ce que (lit. ‘to this that’) is first documented in the 12th century as a temporal “conjunction”, according to Herman (1963: 220). Observe the following examples. (16a) is interpreted as temporal by Herman (1963), whereas (16b) indicates purpose for Jensen (1974: 37, 83–84): (16) a.

Á çó qu’il al pruveire parlad To this that-he to-the priest spoke ‘Once he spoke with the priest.’ (Li quatre libre des reis, 12th c.) Herman (1963: 220)

b.

Labourez A ce que tout soit nettement Ordonné Work to this that all is cleanly organized ‘Work so that everything is neatly organized.’ (Estoire de Griseldis, 14th c.)

Jensen (1974: 37)

A search in the MCVF corpus retrieved certain examples that provide a plausible complement/argumental interpretation of à ce que (as happens with de ce que): (17)

a. por savoir se vos vos porriez acorder a ce que il for know.INF refl you you could agree.INF to this that he fust li mieldres chevaliers del monde was the best knight of-the world ‘In order to know if you could agree that he was the best knight in the world.’ (Mort le roi Artu, 13th c.) b. il furent consillié a ce que il se renderoient a lor They were advised to this that they relf surrendered to their dame la roine lady the queen ‘They were advised to surrender to their lady the queen.’ (Chroniques, Jean Froissart, 14th c.)

208

Crosslinguistic perspective: Romance languages II

c. on peut s’opposer à ce, que l’alteration et corruption one can refl-oppose.INF to this that the-alteration and corruption . . . ne nous esloigne trop de nos . . . principes not us move-away too-much of our principles ‘We can fight against that alteration and corruption moving us away from our principles.’ (Essais, Montaigne, 16th c.) Note that the preposition may be also absent in the same contexts: (18)

a. s’acorderent que le roy alast assieger Alixandre refl-agreed.3PL that the king went capture.INF Alexander ‘They agreed that the king should go capture Alexander.’ (Mémoires ou Vie de Saint Louis, Jean de Joinville, 13th c.) b. qui s’oposoit que sa fille vingt en ce peys who refl-opposed that his daughter came in this country ‘Who was against his daughter coming to this country.’ (Annales de l’Hôtel-Dieu de Montréal, Marie Morin, 17th c.)

The traditional analysis of ce (‘this’) indicates that it is a pronoun to which the clause is adjoined since ce could be the direct complement of prepositions and verbs at least until the 16th century. Furthermore, syntactically ce que could be separated and form a correlation, ce anticipating and referring to the clause. Correlations of this type are attested not only in prepositional combinations such as de ce que or por ce que; Rouquier (1990) cites examples in subject and object positions. Consider the following example (Rouquier 1990: 48): (19) Et quant ce virent les puceles que lor petite dame vient And when this heard the girls that their little lady comes ‘And when the girls heard that their lady was coming.’ (Perceval, Chrétien de Troyes, 12th c.) With time, ce ceases to be attested in object positions with verbs and remains only with prepositions (Darmesteter 1931: 72–73, 4th book). However, it is precisely the differences noted between the separated ce . . . que in a correlation/extraposition and the adjacent ce que which has led linguists such as Rouquier (1990) herself and Zaring and Hirschbühler (1997) to claim that adjacent ce que does not involve a truly pronominal ce. One important piece of evidence comes from the ability of ce to refer to clauses introduced by wh-phrases other than the complementizer que. In (20a) we find ce . . . que

French

209

where (20b) reads ce . . . quant (‘when’) for the same section of the text, each reading coming from different editions of different manuscripts of Perceval (Rouquier 1990: 61): (20)

a.

Ce mout li desabeli que il einsi l’avoit perdue This much him disliked that he thus her-had lost ‘He hated the fact that he had lost her this way.’ (Perceval, Chrétien de Troyes, 12th c.)

b.

Ce mout li desabeli quant il einsi l’avoit perdue This much him disliked when he thus her-had lost ‘He hated the fact that he had lost her this way.’ (Perceval, Chrétien de Troyes, 12th c.)

While adjacent ce que is attested, the expected ce quant is not (Rouquier 1990: 61). Rouquier (1990) considers adjacent ce que to be a conjunction, that is to say, a (complex) complementizer. Following Rouquier’s (1990) lead, Zaring and Hirschbühler (1997) argue that the ce in adjacent ce que is actually a clausal determiner; as such, ce que would qualify as clausal nominalization/substantivization (from early Old French to the present day; see Zaring 1992), as follows (Zaring and Hirschbühler 1997: 371): (21)

[PP por [DP ce [CP que . . .]]]

The data clearly support the conclusion that in no period of the history of the French language can we find attestations of argumental de que or à que with no ce, unlike in Spanish and Portuguese. However, while essentially true, this assertion still requires examining two aspects to reach a more thorough understanding of the syntax of prepositional clauses in French and especially how their evolution and syntactic characteristics compare to Spanish. The first aspect has to do with the actual role of the ce strategy, particularly vis-à-vis Case licensing for clauses (and nominality). In principle, it can be argued that the role of ce is to avoid illegitimate Case assignment to a – nonnominal – finite clause; the pronoun ce, being a nominal element, would check Case instead. Such would be in agreement with hypotheses such as Stowell’s (1981) Case Resistance Principle, and, in the realm of Romance syntax, the hypothesis laid out by Barra (2002) for Spanish. However, as in Spanish, the syntax of adverbial clauses holds the key. Like Spanish and Portuguese, adverbial prepositional finite clauses are attested in historical French (see Buridant 2000: 567). First of all, if the historical

210

Crosslinguistic perspective: Romance languages II

French que-clause had not been a nominal category capable of checking Case, we would be left without a way to account for the syntax of adverbial prepositional finite clauses where lexical prepositions check (lexical) Case, as illustrated by the following examples: (22)

a. Ensemble furent jusque a Deu s’en ralerent Together were.3PL until-that to God refl-there returned ‘They were together until they returned to God.’ (Saint Alexis, 11th c.) b. La noit demurent tresque vint al jur cler The night rest until-that came to-the day clear ‘They remained the night until the clear day.’ (Roland, 11th c.) c. Alés vos ent avant que pis ne vos aveigne Go you from-here before that worse not you comes ‘Go away before something worse comes to you.’ (Roman de Troie, Benoît de Sainte-Maure, 12th c.) d.

Jensen (1990: 481)

des que ele me voit venir since that she me sees come.INF ‘Once she sees me coming.’

(Yvain, Chrétien de Troyes, 13th c.)

The immediate conclusion is that, firstly, it is necessary to discard the hypothesis that que-clauses in historical French were not nominal – as they are attested as prepositional objects – and secondly, there cannot have existed a real constraint against prepositional finite clauses and thus against Case licensing for clauses. If ce checked Case, so did the que-clauses attested in the same configuration. What is more, the examples of (22) become secondary evidence in the critique of the supposed role of the intervening ce as a Case-blocking device. As Rouquier (1990: 55) and others (see Grevisse 1980: 1249; Marchello-Nizia 1997: 370–376) have extensively evidenced, the pronominal strategy was grammatical in adverbial clauses throughout the medieval period, even for prepositions that could take a que-clause directly. Observe the following lists: (23)

a. Presence and absence of ce attested: après ce que/après que (lit. ‘after this that’, ‘after that’) avant ce que/avant que (lit. ‘before this that’, ‘before that’) por ce que/por que (lit. ‘for this that’, ‘for that’)

French

211

par ce que/par que (lit. ‘for this that’, ‘for that’) dès ce que/dès que (lit. ‘since this that’, ‘since that’) tres ce que/tres que (lit. ‘until this that’, ‘until that’) pendant ce que/pendant que (lit. ‘during this that’, ‘during that’) selon ce que/selon que (lit. ‘according this that’, ‘according that’) sans ce que/sans que (lit. ‘without this that’, ‘without that’) puis ce que/puis que (lit. ‘since this that’, ‘since that’) devant ce que/devant que (Jensen 1990: 481) (lit. ‘before this that’, ‘before that’) sauf ce que/sauf que (Moignet 1973: 251) (lit. ‘save this that’, ‘save that’) excepté ce que/excepté que (Moignet 1973: 251) (lit. ‘except this that’, ‘except that’) ja soit ce que/ja soit que (Marchello-Nizia 1997: 375) (lit. ‘already be this that’, ‘already be that’) b. Only direct selection attested: avers que (lit. ‘against that’) envers que (lit. ‘towards that’) entre que (lit. ‘between that’) jusque que (lit. ‘until that’) c. Obligatory presence of ce: à ce que (lit. ‘to this that’) de ce que (lit. ‘of this that’) en ce que (lit. ‘in this that’) endroit ce que (lit. ‘towards this that’) o ce que (lit. ‘with this that’) parmi ce que (lit. ‘among this that’) sur ce que (lit. ‘about this that’) Consider the examples with ce in (24). Note that some of these prepositions are attested in the same period in direct selection (see (22)):

212

Crosslinguistic perspective: Romance languages II

(24) a.

por ce que l’uevre est fause et mauvese for this that the-work is false and bad ‘Because the work is false and bad.’ (Establissement des mestiers de Paris, Étienne Boileau, Estienne Boilique, 13th c.)

b.

devant ce que li Graax venist before this that the Graax came ‘Before Graax came.’

c.

(Queste del Saint Graal, 13th c.)

Aprés ce que le roy fu couronné After this that the king was crowned ‘After the king was crowned.’ (Mémoires ou Vie de Saint Louis, Jean de Joinville, 13th c.)

d.

dès ce que leur entrée fut faicte à Rouen Since this that their entry was done to Rouen ‘Once they had entered Rouen.’ (Mémoires, Philippe de Commynes, 15th c.)

Furthermore, a preposition such as sans (‘without’), which takes the que-clause directly nowadays, is still attested with an intervening pronoun as late as the 17th century, well beyond the Old French period (Grevisse 1980: 1250): (25)

sans ce que je l’ai trouvé without this that I him-have found ‘Without the fact that I have found him.’ (Vaugelas, 17th c.)

Darmesteter (1930: 180, 2nd book) – like Moreno (1985–86) – established a diachronic correspondence between the pronominal alternatives and the nonpronominal ones in some cases, while in general accepting that the reasons behind the elimination or maintenance of ce seem rather arbitrary. Furthermore, it is possible to attest both the absence and the presence of the ce element in the period, which suggests that for certain speakers both option had to be equally grammatical contemporarily: (26)

a. Après que la bonne femme eut dit la confession generale After that the good woman had said the confession general ‘After the good laday had confessed.’ (Cent Nouvelles Nouvelles, 15th c.)

French

213

b. Après ce que en la chambre furent descenduz After this that in the room were descended ‘After having descended to the room.’ (Cent Nouvelles Nouvelles, 15th c.) More importantly, what is clear in the syntax of

is that Case was not the factor; the pronoun ce may not be described as a last resort strategy to avoid Case on finite clauses. The same applies to Zaring and Hirschbühler (1997) analysis of ce as clausal determiner. Zaring and Hirschbühler (1997: 379, fn 13), citing Stowell’s (1981) Case Resistance Principle, assume that ce prevents the clause from illegitimately getting Case. However, such assertion must be discarded, both on intralinguistic grounds, as discussed above, and also for crosslinguistic reasons, as Spanish, Portuguese, and Swedish amply illustrate. The second context of special interest is the actual attestation of a grammatical alternative to de que in historical French. Tobler (1905: 207) mentions the existence of a de quoi (lit. ‘of which’) which is syntactically equivalent to Spanish de que. Consider the following examples: (27)

a. non sans regret de quoy . . . nous n’avons eu des myettes not without regret of that we not-have had some crumbs du pain of-the bread ‘Not without regretting that we have not had breadcrumbs.’ (Marguerite d’Alençon u. G. Briçonnet, 16th c.) Tobler (1905: 207–208) b. L’autre se plaint dequoy ses bras tant beaux The-other refl complains of-that his arms so beautiful devienent longs rameaux become large branches ‘The other complains that his beautiful arms are becoming large branches.’ (Metamorphose, Marot, 16th c.) Glauning (1873: 15)

This is the closest French has ever come to Spanish de que, that is, a preposition selecting for a finite CP without ce and therefore with direct selection, with identical causal connotations as de ce que permits (see also Jensen 1974: 48, who adds similar examples but with dont). Quoi is the tonic relative/interrogative pronoun found in, for instance, pourquoi (lit. ‘for what’, ‘why’). Nevertheless, Tobler (1905: 208) argues that this quoi is identical to the complementizer que (see also Meyer-Lübke 1900: 655–656, who adds that de quoi is not frequent until the 15th century).

214

Crosslinguistic perspective: Romance languages II

The complementizer nature of this quoi finds support in other uses which show the interchangeability between que and quoi; for instance, por quoi alongside por que, and par que alongside par quoi, where the atonic que takes the place of quoi (Tobler 1905: 208; see also Moignet 1973: 162, 248, 321; Kunstmann 1990: 334; Schlieben-Lange 1992: 348). Consider the following example of quoi in place of que (Tobler 1905: 208): (28)

Coument porra de ceste estoire ouvrer, Par quoi la puist How will-can.3SG of this story work.INF for what it can.3SG seur verité fonder on truth found.INF ‘How he will be able to work with story in fashion to found it on truth.’ (Enfances Ogier, Adenes Le Roi, 13th c.)

In addition, Tobler (1905: 209, fn 2) notes that the opposite is also attested, that is to say, the atonic que in the place of the expected tonic quoi (Tobler 1905: 209, fn 2): (29)

a. A faire chose de conrie To do.INF thing of which-one-laughs ‘To do something we can laugh at.’ (Fabliaux et contes. Dou pouvre mercier, 12/13th c.) b. Ce fu cele. . . De c’om porroit plus grant bien dire. . . It was this of who-one could more great well say.INF ‘It was her of whom one could say. . .’

( Joufroi de Poitiers, 13th c.)

In conclusion, historical French shows certain similarities with historical Spanish in terms of prepositional optionality and variation. A number of patterns are found in both languages, with more or less diachronic consistency. In sum, we can conclude that infinitival clauses and que-clauses were nominal and could be arguments and adjuncts. Indirect interrogative clauses could also be direct objects and non-prepositional objects of PP-verbs, as shown above.3 3 I have not found evidence of prepositional indirect interrogative clauses in the literature or in my searches in the MCVF corpus; likewise, I have only found one possible case of nominalized/ substantivized indirect interrogative finite clause in the corpus, despite them being grammatical in present-day French (as in le pourquoi, lit. ‘the why‘). Here is the example: (1) s’il en avoit le pourquoy comme j’ay if-he of-it had the why as I-have ‘If he had the reasons as I do.’ (Cent Nouvelles Nouvelles, 15th c.)

French

215

Que-clauses could, as in Spanish, combine directly with verbs and other selecting categories, or via a

construction. As indicated above, as in Spanish, adverbial connotations can be obtained in several cases with de ce que and à ce que clauses which in other respects look like arguments. A comparison between French and Spanish becomes particularly relevant when the higher frequency of the intervening ce in French and its theoretical repercussions, especially for Case Theory, are taken into consideration. The optionality of ce with many lexical prepositions reveals that its presence is not due to incompatibility between Case and que-clauses, for, in that case, direct selection should have been totally ungrammatical.

2.2 Present-day French The examples for present-day French come from scholarly works, native speakers, online searches, and, in several instances, from the corpus LEXIQUM, a joint project of the Secrétariat à la politique linguistique du gouvernement du Québec and the Université de Montréal in Canada, as indicated in each case.

2.2.1 Prepositions and infinitives The combinatorial options in present-day French are much more rigid than they were in the past and the nominal distribution of the infinitive is likewise more constrained (see Kayne 1999: 49). As indicated in the previous section, several patterns formerly with à (‘to’) switched with time to de (‘of’), which became the most frequent prepositional complementizer. Nevertheless, there are some cases where optionality is still grammatical (Grevisse 1980: 864–891; Jones 1996: 60), such as nier de/Ø (‘to deny’), aimer de/à/Ø (‘to love’), or souhaiter de/Ø (‘to wish’) (Grevisse 1980: 873): (30)

a.

Il souhaitait de voir écraser cette voiture He desired of see.INF crash.INF this car ‘He desired to see this car crash.’ (La Revanche, André Thérive)

b.

Je ne souhaite connaître que vous I not desire know.INF that you ‘I only wish to know you.’ (Le Retour de l’enfant prodigue, André Gide)

216

Crosslinguistic perspective: Romance languages II

In certain cases, the choice of preposition is linked to a difference in meaning (see Grevisse 1980: 884–891) or to different constructions, such as active vs. passive (see Jones 1996: 60). French and Spanish even share a verb with prepositional optionality. Consider the following example of se souvenir (‘to remember’), which usually takes a de but can be used without it (Grevisse 1980: 874): (31) Je me souviens avoir protesté I me recall have.INF protested ‘I recall having protested.’ (A. Billy, in Figaro litterarie, 12/11/1960) An important restriction for prepositional infinitives in present-day French and a clear difference with Spanish is that, out of all the selected or required prepositions, only de and à can combine with an infinitive (Kempchinsky 1992: 35). Thus, the only grammatical options are represented in the following examples (Hawkins and Towell 2001: 273): (32)

a. Marie refuse de sortir Mary refuses of leave.INF ‘Mary refuses to go out.’ b. C’est utile à savoir This-is useful to know.INF ‘It is useful to know.’

Other possible required prepositions, while grammatical and necessary with DPs, are barred. Compter sur (‘to count on’) just eliminates the preposition (33a), whereas consister en (‘to consist of’) switches from en (‘in’) to à (33c) (Kempchinsky 1992: 35): (33)

a. Je compte (*sur) avoir l’aide de mes amis I count on have.INF the-help of my friends ‘I count on having my friends’ help.’ b. *L’activité consiste en parler avec les visiteurs The-activity consists in speak.INF with the visitors ‘The activity consists in speaking with the visitors.’ c. L’activité consiste à parler avec les visiteurs The-activity consists to speak.INF with the visitors ‘The activity consists in speaking with the visitors.’

French

217

The restriction does not apply to adverbial clauses, as lexical prepositions can take an infinitival clause as complement: (34) a.

Jules a mangé avant de se coucher Jules has eaten before of refl go-to-bed.INF ‘Jules ate before going to bed.’

b.

Jones (1996: 385)

J’y suis allé pour parler avec lui I-there am gone for speak.INF with him ‘I went there in order to speak with him.’

c.

après avoir acheté une glace after have.INF bought an ice-cream ‘After having bought an ice-cream.’

Hawkins and Towell (2001: 305)

Infinitival clauses may appear in subject and object positions, typically nominal contexts, as well: (35)

a. [CP Aller en vacances] est possible Go.INF in holidays is possible ‘To go on holidays is possible.’ b. Je veux [CP aller à Townsville] I want go.INF to Townsville ‘I want to go to Townsville.’

The previous examples point out that infinitives are distributionally nominal categories (nominal clauses) in those contexts and thus must check Case with lexical prepositions, as required of prepositional objects. When de and à introduce an infinitival clause in French, in many cases they are prepositional complementizers rather than as actual prepositions (Kayne 1975, 1981, 1999; Cinque 1990: 34–37; Jones 1996: 59). Consider the following examples: (36)

a. Jean se réjouit de partir bientôt à l’étranger Jean refl rejoices of leave.INF soon to the-overseas ‘Jean is pleased to be leaving overseas soon.’ b. Jean se réjouit que cette affaire soit terminée Jean refl rejoices that this matter is finished ‘Jean is pleased that this matter is over.’

Huot (1981: 48)

218

Crosslinguistic perspective: Romance languages II

c. Jean se réjouit de ce voyage Jean refl rejoices of this trip ‘Jean is pleased with this trip.’ d.

Huot (1981: 49)

Jean m’a dit de partir Jean me-has said of leave.INF ‘Jean has told me about leaving.’

e.

Jean m’a dit qu’il est parti Jean me-has said that-he is left ‘Jean has told me that he is gone.’

f. Jean m’a dit ça Jean me-has said this ‘Jean has told me that.’ Following Rizzi’s classification for prepositional complementizers, only (36d) shows a prepositional complementizer in the infinitival clause, as it is barred from other contexts (36e,f ). On the contrary, de in (36a) does not qualify as a prepositional complementizer as it is maintained with other categories. (36b) will be discussed in the next section (see Huot 1981 for an analysis in terms of prepositional deletion; see chapter 2, section 2.3.2.). While the status of de as a prepositional complementizer in the right contexts is widely accepted, there is not such agreement as for the status of à as prepositional complementizer. Huot (1981) argues that it is a regular preposition, on the basis that most contexts are selected by PP-verbs and other categories which require either à or another preposition with other categories (nouns or que-clauses) (Huot 1981: 85). Consider the following set of examples: (37)

a. Pierre aspire à revenir dans sa ville natale Pierre aspires to return.INF in his city natal ‘Pierre aspires to return to his hometown.’ b. Pierre aspire à la retraite Pierre aspires to the retirement ‘Pierre aspires to retirement.’

Huot (1981: 85)

c. Pierre est expert à organizer des séances récréatives Pierre is expert to organize.INF some sessions recreative ‘Pierre is expert in organizing recreation sessions.’

French

d.

219

Pierre est expert en mécanique Pierre is expert in mechanics ‘Pierre is an expert in mechanics.’

Huot (1981: 87)

While à in (37a) clearly does not qualify as prepositional complementizer given that it is necessary with other complements, Huot (1981: 87) extends the same conclusion to (37d) because the same verbs need a preposition with nouns. The issue here is that the otherwise selected preposition, as shown above in (33), must be changed to à. What the actual form of the preposition turns out to be, in Huot’s terms, does not alter the fact that that that syntactic position must remain truly prepositional. Other factors must be independently responsible for the mandatory modification of the preposition. Furthermore, Huot (1981) shows that à cannot be substituted by the direct object pronoun le, unlike what happens with prepositional complementizer de (Huot 1981: 454): (38) *Pierre l’aime, à faire de grandes promenades à bicyclette Pierre it-loves to do.INF of big walks to bike ‘Pierre loves it, going for long bike rides.’ Martineau (1990: 151), citing Rochette (1988), shows, however, that this class of verbs cannot substitute their infinitival complements by a direct object pronoun whether introduced by à or de, which weakens this test. Nevertheless, the prepositional nature of à finds some additional support in Canac Marquis’s (1996) study of tough constructions, who claims that à in such sentences as ce livre est difficile à lire (‘this book is easy to read’) is a preposition, not a complementizer (Canac Marquis 1996: 40). In any case, if à is found not to be specialized to introduce only infinitives but rather either reappears with other categories or combines with a PP-verb which – for reasons beyond the scope of this chapter – has to swap prepositions, then it is fair to simply conclude that it does not meet the requirements for the (strict) definition of prepositional complementizer in Romance used here. On the other hand, what it does retain is its functional nature when introducing an argumental PP. In the end, this categorial issue is limited to infinitival clauses, the classical realm of prepositional complementizers in Romance, as their finite counterparts – de ce que and à ce que – cannot be syntactically analyzed but as involving prepositions, whether functional if argumental or lexical even, if used in adverbial causal or final clauses.

220

Crosslinguistic perspective: Romance languages II

An important difference between French and Spanish infinitives is the lack of a productive nominalization/substantivization mechanism in the former. Unlike several periods of historical French, present-day French only present a number of fixed VP nominalizations and a certain revitalization of the VP level nominalization pattern in certain areas such as sports, philosophy, politics, etc. (Buridant 2008: 281–352). Any higher level nominalizations/substantivizations or productive derivational nominalizations/substantivizations are ungrammatical. Consider the following example (Barra 2002: 279): (39)

(*Le) boire de la bière le matin empêche de travailler The drink.INF of the beer the morning prevents of work.INF ‘Drinking beer in the morning prevents from working.’

The article is ungrammatical even when the infinitive shows identifiable nominal properties such as PP-modification. However, as previously discussed in the sections on Spanish and Portuguese, the implicational link “nominality, ergo presence of the article” cannot be maintained. The presence of the article only indicates a DP projection, while infinitival clauses can be described, as que-clauses, as nominals (essentially equivalent to an NP) as discussed for Spanish, Portuguese, and historical French. A hypothetical lack of nominality of infinitival clauses in French cannot be total in pure theoretical terms, as its grammaticality in subject, direct object and prepositional object positions would be left unanswered and unaccounted for, even if it they do not participate in all possible nominal contexts as freely as in Spanish (see, for instance, Kayne 1999).

2.2.2 Prepositional finite clauses Argumental prepositional finite clauses with direct selection of the que-clause as in Spanish are disallowed in French; ce is mandatory (see Sandfeld 1965, among many others). As with infinitives, it is mostly de and à which are found in these finite constructions with a ce que-clause, although some examples of en and sur are also reported in the literature (see Sandfeld 1965: 44; Grevisse 1980: 1281). Observe the following examples: (40)

a. Jean abuse de ce que son frère est plus faible Jean abuses of this that his brother is more weak ‘Jean takes advantage of his brother being weaker.’

French

221

b. Jean se refuse à ce que sa présense à cette réunion Jean refl refuses to this that his presence to this reunion soit annoncée is announced ‘Jean refuses that his presence in the meeting be announced.’ Huot (1981: 159) c.

Je me plains de ce que tu sois en retard I me complain of this that you are in late ‘I complain about your being late.’

d.

Rowlett (2007: 146)

La difficulté consiste en ce que nous avons peu de temps The difficulty consists in this that we have little of time ‘The difficulty consists in the fact that we have little time.’ Le Goffic (1993: 343)

Similarly to Spanish, certain predicates – but interestingly not all of them – may combine directly with a que-clause without

(Zaring 1992: 58–60). Observe the differences in the following set of examples. The use of à ce in (41a) is considered grammatical in colloquial French by some native speakers (Cécile d’Agaro, pc): (41)

a. Je m’habitue *(à ce) qu’elle dorme pendant la journée I me-accustom to this that-she sleeps during the day ‘I am getting used to her sleeping during the day.’ b. Paul est attentif *(à ce) que ses enfants fassent leurs Paul is attentive to this that his children do their devoirs le soir homework the night ‘Paul pays attention that his kids do their homework at night.’ c.

Les étudiants se plaignent toujours (de ce) qu’ils ne The students refl complain always of this that-they not peuvent pas nous comprendre can not us understand.INF ‘The student always complain that they cannot understand us.’ Zaring (1992: 58)

222

Crosslinguistic perspective: Romance languages II

d.

Je me souviens *(de) son adresse/ *(d’)avoir dit cela/ I me recall of his address of-have.INF said that (*de ce) que Marc est parti assez tôt of this that Marc is left rather early ‘I remember his address/having said that/that Marc left rather early.’ Zaring (1992: 59)

As Zaring (1992: 58) points out, the majority of these examples that allow optionality involve de; à tends to be obligatory. Importantly, the fact that certain PP-expressions do nevertheless reject

actually shows that French shares with Spanish the configuration where bare que-clauses can be licensed as arguments. As for the syntactic representation of

, the pronominal nature of ce was already rejected already in the historical period. For Bouchard and Hirschbuhler (1987: 50) and Rouquier (1990: 53), ce que was reanalyzed as a complementizer and therefore a unit, as illustrated in the following analysis: (42) Elle se plaint [KP de [CP ce que [TP tu n’étudies pas]]] She refl complains of this that you not-study not ‘She complains about you not studying.’ On the other hand, as mentioned before (see Zaring and Hirschbühler 1997), Zaring (1992) argues that ce is actually a clausal determiner, not a pronoun (see also Pollock 1992: 453; Rowlett 2007: 146). Evidence in favor of this comes from the fact that ce does not qualify independently as a pronoun in present-day French. For instance, it cannot be the object of the preposition; instead, cela or ça (‘this’) must be used (Zaring 1992: 56): (43) *Je m’habitue à ce I me-accustom to this ‘I am getting used to this.’ Furthermore, not being a pronoun, ce does not totally block extraction, adding support to the argumental status of selected de ce que and à ce que and rejecting the hypothesis whereby the que-clause is adjoined to ce: (44)

a. Qu’est-ce que tu t’habitues à ce qu’elle fasse toute seule? What-is-this that you you-accustom to this that-she does all alone ‘What are you getting used to her doing all alone?’

French

223

b. ?Avec quel copain souffre-t-il (de ce) que Chantal sorte? With which friend suffers-t-he of this that Chantal leaves Lit.‘With which friend does he suffer from Chantal’s going out?’ Zaring (1992: 65) c.

Qui tiens-tu à ce que Jean rencontre? Who insist-you to this that Jean meets Who do you insist Jean should meet?’ Bouchard and Hirschbuhler (1987: 50)

Extraction is not grammatical across the board (Zaring 1992: 64–66, 1993), and but, in any case, Zaring (1992, 1993) provides an independent account for such constraints based on tense constraints, while preserving argumenthood. Additional evidence against a pronominal analysis of ce comes from the fact that that

does not form a constituent independent from the finite clause, as is evidenced by the fact that it cannot be substituted by the corresponding prepositional proform y or en while at the same time leaving the finite clause in its place (Zaring 1992: 67): (45)

a.

*J’y veillerai qu’il se couche de bonne heure I-to-it will-watch that-he refl goes-to-bed of good hour ‘I will see to it that he goes to bed early.’

b.

J’y veillerai I-to-it will-watch ‘I will see to it.’

A final piece of evidence comes from the fact that ce is regularly used in French as a demonstrative determiner, as in ce livre (‘this book’) (Zaring 1992: 68). The resulting tree is as follows (Zaring 1992: 69): (46)

J’ai profité de [DP ce [CP qu’il a fait beau]] I-have benefitted of this that-it has made beautiful ‘I took advantage that it was a nice day.’

As is, Zaring’s analysis leaves certain issues unexplained. One important issue revolves around the fact that in present-day French, unlike historical French, the syntactic productivity of this ce is null; it is restricted to de ce que and à ce que (and en ce que and sur ce que for some speakers) and does not

224

Crosslinguistic perspective: Romance languages II

work with any other categories (Zaring 1992: 56; Authier and Reed 2010: 12), including infinitives and indirect interrogative clauses: (47)

a.

*Je m’habitue à ce dormir pendant la journée I me-accustom to this sleep.INF during the day ‘I am getting used to it sleeping during the day.’

b.

Zaring (1992: 56)

*Paul ne s’est jamais soucié de ce s’il restait de Paul not relf-is never worried of this if-it remained of l’argent à la fin du mois the-money to the end of-the month ‘Paul never cared about whether or not there was any money left at the end of the month.’ Authier and Reed (2010: 17)

This lack of productivity points out that de ce que and à ce que have become fixed. This idea is captured by Authier and Reed (2010), who conclude that de ce que and à ce que are complementizers.4 However, assuming that they are complementizers means that their projection is not prepositional. I contend that such a hypothesis is too powerful as it is easily falsified by the substitution test in (45). The preposition in de ce que and à ce que “belongs” to the select4 Authier and Reed (2010) point out that for certain predicates the use of que vs. de ce que produces certain meaning differences. Huot (1981) already notes that certain uses of de ce que are different from que. It is expected that two different ways of expressing the same predication might become specialized in the right contexts, but, importantly, this seems not to be applicable across the board. It is also true that in some cases de ce que combines with indicative whereas the equivalent bare que-clause requires the subjunctive, but again not all the time (for instance, se désoler; see below). On the other hand, Authier and Reed (2010) cite Bat-Zeev Shyldkrot (2008), who argues that the prepositional variety adds a causal connotation. This argument is much more problematic. First, the causal connotation has been present since the first uses of de ce que. Secondly, such connotation is not limited to French, as it applies to Spanish as well; it is expected from the very semantics of certain verbs, especially emotive verbs, that their theme can be interpreted as the cause of the emotion (Authier and Reed 2010 do note that this connotation could not account for all those predicates which take de ce que). It is very clear, though, that this de que argument is nevertheless very different from a porque adjunct, as the former allows extraction, for instance, while the former does not. The same applies in French, as Zaring (1992) proves. Bat-Zeev Shyldkrot (2008: 119) actually illustrates this difference with the verb se désoler, which happens to select the indicative mood both with de ce que and bare que. Bat-Zeev Shyldkrot’s argument for French sounds similar to the hypothesis laid out for certain Spanish de que clauses which claims that de is a marker of evidentiality (iconicity) (Schwenter 1999; Demonte and Fernández Soriano 2005), a hypothesis that has been adequately falsified (Gutiérrez-Rexach and Silva-Villar 2007). What is more, Grevisse (1980: 1347) indicates that the causal connotation can be obtained with both bare que and de ce que variants, exactly as in Spanish, which cancels this special use of de ce que.

French

225

ing category, which is very clear in the case of PP-verbs. Instead, I argue that the answer resides in combining Zaring’s (1992) and Authier and Reed’s (2010) accounts. One of the properties Zaring (1992) herself already highlights is that ce must always be adjacent to the finite clause and allows no interpolation, thus questioning the internal divisibility of ce que (Zaring 1992: 56): (48)

*Je m’habitue à ce évidemment qu’elle dorme pendant la journée I me-accustom to this evidently that-she sleeps during the day ‘I am getting used to it evidently that she sleeps during the day.’

This indivisibility is compatible with Authier and Reed’s (2010) account. At the same time, in descriptive terms, it is exactly the same strict adjacency found in their Spanish counterparts, in that case without a determiner. Thus, in order to maintain the prepositionality of the constructions – functional prepositions, in reality – which in any case cannot freely combine with any predicate (that is, à ce que does not combine with se soucier, for instance, and de ce que does not combine with vouloir), I assume that the syntactic analysis is that of a (functional) preposition selecting for a , following Zaring’s account, but clearly idiosyncratically reduced to only a limited – arbitrary – number of prepositions which have continued to include ce throughout the years, particularly considering that ce in adverbial clauses was optional during periods of historical French. That is to say, lack of productivity signals that speakers have isolated those particular combinations – probably through surface (analogical) chopping of a construction. Taking into consideration Gaatone’s (1981) view that unsegmentability does not equal unanalizability, it is possible to maintain that, as in Spanish

weak clusters, the internal structure of the construction can still be described as

. Again, Spanish al or del or French au or des are phonetic units which nevertheless permit internal syntactic segmentation. Zaring’s analysis brings Case Theory back into the picture. She claims that the determiner ce serves to nominalize the finite clause and thus prevents it from illicitly getting Case, in keeping with Stowell’s Case Resistance Principle (Zaring 1992: 71). I have already commented on and subsequently discarded this hypothesis5 when discussing the role of ce in adverbial clauses in historical French. It is undeniable that French is really very similar to English in its restric5 Burchert (1994) discusses this configuration as well and, noting Spanish

, argues that there is an empty noun which is the head of n NP complement of the determiner ce. However, Picallo (2002: 119) convincingly proves that such a hypothesis is inadequate for Spanish, which can be extended to French.

226

Crosslinguistic perspective: Romance languages II

tion on the nominal contexts where finite clauses may appear (see Huot 1981: 100–102), contexts where the Spanish que-clause is perfectly grammatical. Accordingly, several linguists conclude that que-clauses are not nominal in French (Rowlett 2007: 146–147). This generalization is nevertheless too strong. On the one hand, as in the case of adverbial clauses in historical French, the role of ce as a block for Case checking is inadequate. For instance, note that ce cannot be included to rescue the ungrammaticality of the same contexts Huot (1981) lists as evidence for the non-nominal nature of French que-clauses. The same extends to indirect interrogative clauses (but see below). This again proves that ce is not a last resort element against Case. Furthermore, ce is not required with infinitival or finite clauses when they appear in typically nominal positions where they are to check (see Bošković 1995 for similar arguments for English): (49)

a.

Demandez-lui donc s’il ne connaît pas une auberge Ask-him then if-he not knew not a hostel ‘Ask him if he did not know a hostel.’

b.

Je ne sais vraiment pas quel charme il peut trouver I not know truly not what charm he can find.INF à cette région at this region ‘I do not know what charm he can find in this region.’ Huot (1981: 130)

c.

Je pense qu’il dit la verité I think that-he says the truth ‘I think he is telling the truth.’

Additional evidence for Case checking comes from adverbial clauses, which demonstrate that there exists no restriction against direct prepositional selection of a que-clause. Observe the following examples: (50)

a. pour que ce voyage se passe sans ennui for that this trip refl passes without ennui ‘So that this trip is not boring.’ b. Pierre est parti Pierre is left demander son request.INF his

Huot (1981: 59)

sans que personne ait songé à lui without that person has thought to him adresse address

‘Pierre left without no one thinking of asking him for his address.’ Huot (1981: 493)

French

227

c. Dès que Pierre sera là, nous commencerons les travaux Since that Pierre will-be there we will-start the works ‘Once Pierre arrives there, we will start the work.’ d.

Selon que son travail sera fini ou non According-to that his job will-be ended or not ‘Depending on whether his work is done or not, Jean will join us on August 15.’ Huot (1981: 82)

If understood as subordinating conjunctions, they would be categorized as units, as complementizers, but then we would lose the same generalization as in Spanish or Portuguese, these prepositions can select other nominal categories beyond que-clauses. This idea was already captured by Gaatone (1981: 204), who explains the syntactic structure of avant que by indicating that the preposition introduces and subordinates a clause nominalized by que. This same PP structure applies to après que, dès que, pour que, outre que, pendant que, au lieu que, depuis que, sans que (Gaatone 1981: 206; see also Jones 1996: 382).6 Nonprepositional adverbial clauses are also grammatical in French (Rowlett 2007: 148): (51)

a. Viens ici que je te voie mieux Come here that I you see better ‘Come here so I can see you better.’ b. Il est pas venu qu’il est malade He is not come that-he is sick ‘He didn’t come because he is sick.’

6 Rowlett (2007: 147) argues for a PP analysis as well, and proposes that the syntactic analysis could be either

, with direct selection, or rather

, as he argues that French que-clauses are not nominal. This solution is theoretically sound, of course, but it adds a level of empty categories which seems unnecessary, particularly given the evolution of French. Ce was possible with a wide array of que-clauses in the past and that optionality (in adverbial clauses, subject clauses, and even direct object clauses) was lost in favor of direct selection. Ce que is only possible with de ce que and à ce que (with some cases of en ce que and sur ce que), which is exactly why these combinations are special in French and merit being considered as fossilized. Non-relative ce que is not productive in present-day French. Furthermore, in the section on Spanish I showed that it was unnecessary to argue for an empty determiner either for que-clauses when the article is not overt. I think Leonetti’s (1999a) comments regarding the differences in extraction between que and el que are sufficient to argue against empty Ds for all que-clauses in Spanish, and thus by extension, to French. To this we could add Manzini and Savoia’s (2005, 2011), Manzini’s (2010), and Roussou’s (2010) claims with regard to the nominality of que/that.

228

Crosslinguistic perspective: Romance languages II

Thus, que-clauses and indirect interrogative finite clauses are nominal,7 as they contain a nominal complementizer, as argued for Spanish and Portuguese as well, in keeping with recent research on the subject (Manzini and Savoia 2005, 2007, 2011; Manzini 2010; Roussou 2010), not only for Romance, but also for English that (Kayne 2008; Roussou 2010), subject to similar constraints as French que. As Manzini (2010: 186) claims for Italian, other interpretive factors must be barring the que-clause from certain nominal contexts, not Case. The last section is devoted to the syntax of indirect interrogative finite clauses. In principle, no selected preposition seems to be able to introduce an indirect interrogative finite clause (Huot 1981: 56):8 (52)

a. *Je ne me souviens pas de si le magasin est ouvert I not me recall not of if the store is open ‘I don’t remember whether the store opens in August.’ b. *Je ne me souviens pas d’à qui j’ai prêté ce livre I not me recall not of-to who I-have lent this book ‘I do not remember who I lent this book to.’ Grammaticality is restored once the preposition is removed (Huot 1981: 152):

(53)

Je ne me souviens pas à qui j’ai prêté ce livre I not me recall not to who I-have lent this book ‘I do not remember who I lent this book to.’

The ungrammaticality of (52) cannot account for the fact that these same clauses cannot freely combine with nouns or adjectives for all speakers, with or without the mediation of a preposition (Huot 1981: 128–131). However, (54b) is indeed grammatical in colloquial French for some speakers (Cécile d’Agaro, pc):

7 See Kayne (1999: 49) for a similar idea for infinitives. 8 Indirect questions introduced by seem to be grammatical as prepositional objects (Jones 1996: 494–495; Nakamura 2008: 2580) (Jones 1996: 495, fn 7): (1) Je ne suis pas sûr de ce que Pierre fera I not am not sure of this that Pierre will-do ‘I am not sure what Pierre will do.’ In this case, though, these clauses are free relative clauses in reality (Jones 1996: 495, fn 7; Rowlett 2007: 233).

French

(54)

229

a. *Es-tu sûr (de) si Jean a reçu ces papiers? Are-you sure of if Jean has received these papers? ‘Are you sure whether Jean has received these papers.’ Huot (1981: 131) b. *Jean était. . . incertain (de) quand il deviendrait titulaire Jean was unsure of when he would-become tenured ‘Jean was always unsure when he would get tenure.’ Huot (1981: 153)

What is more, a further look at colloquial French offers a different picture. First of all, several linguists report a number of colloquial examples of

as far back as the 80s: (55)

a. Je ne me souviens pas de qui a dit cela I not me recall not of who has said that ‘I do not remember who said that.’

Huot (1981: 57)

b. On peut s’interroger sur quelles étaient les circonstances We can refl-wonder.INF about which were the circumstances ‘We can wonder which the circumstances were.’ (Radio Informations, 08/2000) Blanche-Benveniste (2001: 91) Huot (1981) herself also reports several counterexamples to the second constraint, beyond (54b) according to my native informant: (56)

a. Soyez attentif à de quoi Pierre parlera Be attentive to of what Pierre will-speak ‘Pay attention to what Pierre will speak about.’ b. Soyez attentif de quoi Pierre parlera Be attentive of what Pierre will-speak ‘Pay attention to what Pierre will speak about.’

Huot (1981: 491)

c. As-tu fait attention s’il restait assez d’essence? Have-you made attention if-it remained enough of-gasoline ‘Have you paid attention to whether there was enough gas left?’ Huot (1981: 489) The ungrammaticality of

in presentday French must be heavily qualified, as evidenced by the large amount of examples found online (see also Nakamura 2008):

230 (57)

Crosslinguistic perspective: Romance languages II

a. Je ne me souviens pas de si ça a déjà été dit I not me recall not of if that has yet been said ‘I don’t remember if that has already been said.’ (15-3-12) b. jolicoeur a commenté sur Pourquoi François Bayrou peut gagner J. has commented on why F. B. can win.INF ‘J. talked about why F.B. can win.’ (21-3-12)

Additionally, a search in the online corpus Lexiqum, a joint project of the Secrétariat à la politique linguistique du gouvernement du Québec and the Université de Montréal in Canada, retrieves abundant examples of the relevant configuration from political speeches in Canadian French, which leaves absolutely no doubt as to the grammatical use of prepositional indirect interrogative finite clauses even in more formal registers: (58)

un constat de comment on traite notre jeunesse a report of how one treats our youth ‘A report on how we treat our youth.’ (Lexiqum, Assemblée nationale du Québec)

What is more, several wh-phrases and their corresponding indirect interrogative finite clauses do participate in nominalization/substantivization by combining with articles, as le pourquoi (lit. ‘the why’) and le comment (lit. ‘the how’), which are found in the Lexiqum corpus. In this respect, French is again like Spanish and Portuguese: (59)

a. les intentions ou le pourquoi de certaines modifications the intentions or the why of certain modifications ‘The purpose or the reasons behind certain modifications.’ (Lexiqum, Assemblée nationale du Québec) b. le pourquoi je suis content d’intervenir the why I am happy of-intervene ‘The reason why I am happy to speak.’ (Lexiqum, Commissions parlementaires, Québec)

French

231

c. dans le comment on devrait nous-même proposer les choses in the how one should ourselves propose.INF the things ‘In how we should propose things ourselves.’ (Lexiqum, Assemblée nationale du Québec) To sum up, while que-clauses are still restricted, as their English counterparts, it can be proven that they must be nominal, being headed by a nominal complementizer, because they are licensed in typically nominal positions. As a result, on theoretical grounds, the role of ce as a clausal determiner cannot be that of blocking Case on the que-clause. The resulting syntactic analysis shows that

is the French version of Spanish el que-clauses, without the semantic constraints operative in (standard) Spanish. The many examples found online support the idea that indirect interrogative finite clauses in French are nominal as well, and may participate in substantivizations, as in the case of Portuguese, where que-clauses are equally ungrammatical introduced by articles.

2.3 Discussion and conclusions The data presented in the previous sections allow concluding that speakers of French accept – and accepted in the past – certain degrees of prepositional – and prepositional complementizer – optionality and have used

in adverbial contexts. Adding the evidence from colloquial and even more formal French, it is fair to conclude as well that in present-day French, in particular, there is no full constraint against prepositional finite clauses. There are some specific points that deserve particular attention when French and Spanish are compared regarding the emergence of prepositional finite clauses in the latter, especially taking into consideration the hypothesis laid out for Spanish in Barra (2002). Firstly, according to the analysis presented here, both French and Spanish have argumental prepositional finite clauses. In particular, the role of ce as a pronoun to which the que-clause would adjoin has been discarded. Furthermore, both for historical periods of French and for present-day French, the role of ce as a deterrent of sentential Case checking has been rejected; rather, following Zaring’s (1992) insight, and partially in keeping with Authier and Reed’s (2010) paper, I argue that French

are fixed, not productive and yet syntactically analyzable combinations of prepositions and the equivalent to Spanish el que-clauses. What is more, while shortly lived, certain speakers of Classical French developed a new kind of argument prepositional finite clause with de quoi, not attested today. For those speakers’ I-languages, direct selection

232

Crosslinguistic perspective: Romance languages II

of a quoi(=que)-clause by a functional preposition was grammatical alongside the common de ce que, etc. Secondly, the presence of the clausal determiner, while not productive today, was not limited to argumental clauses as it was available for adverbial clauses – adjuncts – in alternation with direct selection. The fact that direct selection was grammatical for the same speakers who could insert ce indicates that at the very least those lexical prepositions must have checked Case on the (nominal) que-clause exactly as they must have done with the DP projection created by ce. In other words, the use of ce was independent from the argument/ adjunct divide, as the use of de que is in present-day Spanish (Leonetti 1999b). Thirdly, the fact that several predicates such as se souvenir (‘to remember’) do not take a prepositional finite clause despite requiring the preposition for other categories suggests that que-clauses can be licensed positionally, as in Spanish. Non-prepositionality does not equal adjunction, as extraction shows that, with or without , these clauses are indeed argumental. However, again extraction proves otherwise (Huot 1981: 68): (60)

a.

Quel est ce livre que Jean ne se souvenait pas que son Which is this book that Jean not refl recalled not that his père avait déjà acheté? father had yet bought ‘Which is the book that Jean did not remember that his father had already bought?’

b. Quel est ce prix que Jean se réjouit que son frère Which is this prize that Jean refl rejoices that his brother ait gagné? has gained ‘Which is the prize Jean is very pleased his brother has won?’ The fact that French does have argumental clauses is crucial for the discussion of Spanish. Barra (2002: 399) argues that French must lack argumental finite clauses; since Barra argues that French does not have argumental prepositional finite clauses, que-clauses must have never been nominal and thus must have never become argumental in any position. However, the evidence clearly shows otherwise. Fourthly, French also shares with Spanish the incompleteness of prepositional selectional patterns. For instance, depuis (‘since’) accepts both a nonclausal DP (61a) and a que-clause (61b), but not an infinitive (61c) (Huot 1981: 80):

French

(61) a.

233

Il n’était pas revenu depuis la fin de la guerre He not-was not returned since the end of the war ‘He had not returned since the end of the war.’

b.

Il n’était pas revenu depuis que son travail l’avait appelé He not-was not returned since that his job him-had called à Paris to Paris ‘He had not retuned since his job had taken him to Paris.’

c.

*Il n’était pas revenu depuis avoir déménagé He not-was not returned since have.INF move-out ‘He had not returned since moving out.’

Contre (‘against’) exhibits even more limited selectional properties, in that it only takes non-clausal DPs; (62c) illustrates that ce cannot be added to save a derivation (Huot 1981: 94): (62)

a. Pierre s’est toujours élevé contre ses mesures Pierre refl-is always risen against his measures ‘Pierre has always protested against his measures.’ b. *Pierre s’est toujours élevé contre que de telles mesures Pierre refl-is always risen against that of such measures soient soumises à un vote are subjected to a vote ‘Pierre has always protested against having such measures put to the vote.’ c. *Pierre s’est toujours élevé contre ce que de telles Pierre refl-is always risen against this that of such mesures soient soumises à un vote measures are subjected to a vote ‘Pierre has always protested against having such measures put to the vote.’

Finally, French provides an extra look into the imperfect relationship between nominality, prepositions, and presence of articles in substantivization. While infinitival clauses, que-clauses and indirect interrogative finite clauses

234

Crosslinguistic perspective: Romance languages II

can all be prepositional objects to some degree, only VP level infinitival nominalizations/substantivizations are grammatical in present-day French and only indirect interrogative finite clauses can be introduced by a determiner, *le que (lit. ‘the that’) being ruled out, as in Portuguese, and ce que being nowadays severely constrained as prepositional object. In short, while French imposes and has historically imposed more constraints on the syntax of que-clauses than Spanish or Portuguese, they are and were nevertheless grammatical in typically nominal positions – as is and was the case with infinitival clauses – and thus must check Case according to the general expectations of Case Theory. Argumental que-clauses are grammatical in present-day French and are attested in historical French. French had and currently has prepositional finite que-clauses as well, involving a (fixed, nonproductive) clausal determiner.

3 Italian 3.1 Historical Italian As with the other languages analyzed in this book, by historical Italian is intended as a cover term for older stages of the language, with special attention to the Tuscany area. The examples are dated in the 13th and 14th centuries, with one example from the 19th century. The data are extracted from scholarly works and from the online corpus Opera del Vocabolario Italiano (OVI), accessible through the website of the Istituto Opera del Vocabolario Italiano at the Consiglio Nazionale delle Ricerche and also through the ARTFL project. 3.1.1 Prepositions and infinitives The infinitive in historical Italian shares many properties with that of other Romance languages. First, it may be in subject or object positions, both in bare form or introduced by a prepositional complementizer, especially di (‘of’) (Rohlfs 1969; Egerland 2010, among many others). These possibilities are attested in other Romance languages, as evidenced in previous sections: (63) a.

Ne le cose aperte voler argomentare è simigliante In the things open want.INF argue.INF is similar matezza come alluminare la chiarità del sole con. . . foolishness like light.INF the clarity of-the sun with ‘To want to argue about evident things is as stupid as to light up the clarity of the sun with. . .’ (Fiori e vita di filosafi, 13th c.)

Italian

b.

235

Per cessare fame e sete non è bisogno tentare il mare For cease.INF hunger and thirst not is necessary try.INF the sea ‘In order to satisfy hunger and thirst, it is not necessary to try to sail.’ (Fiori e vita di filosafi, 13th c.) Egerland (2010: 826)

c.

nelle quali non bisogna di contare il nome del parlieri in-the which not needs of tell.INF the name of-the speaker ‘In which it is not necessary to tell the name of the speaker.’ Egerland (2010: 825)

d.

e mellio desiderava passare per forza che per amore and better desired pass.INF for force that for love ‘And he preferred to pass by force than by love.’ (Fatti di Cesare, 13th c.)

e.

tanto desiderava di combattare con Pompeio so desired of fight.INF with Pompeius ‘He desired so much to fight Pompeius.’ (Fatti di Cesare, 13th c.)

Interestingly, Rohlfs (1969: 98) points out that some of these verbs could actually take di with nouns as well (see Egerland 2010: 106). Therefore, at least for certain speakers, verbs such as domandare, sperare, or temere were indeed prepositional and did not restrict the use of di to that of a prepositional complementizer. Consider the following set of contrasting examples with the verb temere (‘to fear’) taken from the same texts, actually showing the same behavior of temer in historical Spanish: (64)

a.

il giudicie che teme di far vendetta the judge that fears of do.INF revenge ‘The judge that fears to take revenge.’ (Trattati morali di Albertano da Brescia, Andrea da Grosseto, 13th c.)

b.

teme essere ben fedele fears be.INF well faithful ‘He fears to be faithful.’ (Trattati morali di Albertano da Brescia, Andrea da Grosseto, 13th c.)

c.

Et qua(n)do alcuno luogo teme di periculo And when some place fears of danger ‘And when he fears danger somewhere.’ (Trattati di Albertano da Brescia volgarizzati, 13th c.)

236

Crosslinguistic perspective: Romance languages II

d.

no(n) temere le paraule dela tua moglie airata not fear.INF the words of-the your wife angry ‘Do not fear your angry wife’s words.’ (Trattati di Albertano da Brescia volgarizzati, 13th c.)

(Other) PP-verbs, nouns, and adjectives may take an infinitival clause as their complement (see Rohlfs 1969: 93–98): (65)

a. certo di conquistare questo regno sure of conquer.INF this kingdom ‘Sure about conquering this kingdom.’ (Libro de’ vizî e delle virtudi, Giamboni, 13th c.)

Egerland (2010: 861)

b. Usura è uno studioso desiderio d’avere alcuna cosa Usury is a studied wish of-have.INF some thing oltre la sorte beyond the loan ‘Usury is the conscious wish to have more than just a loan.’ (Libro de’ vizî e delle virtudi, Giamboni, 13th c.) Egerland (2010: 867) As the other old Romance languages, infinitives with overt nominative subjects – especially in postverbal position – are also attested, again in more contexts than present-day Italian (Mensching 2000: 107–112, 130–135; see also Egerland 2010: 873): (66)

a. È una grande fatica a conservare l’uomo la pecunia Is a great effort to keep.INF the-man the money ‘It is a great effort for a man to keep his money.’ (Prediche inedite, Giordano da Pisa, 14th c.) Mensching (2000: 108) b. da questo essere abbandonati gli infermi da’ vicini from this be.INF abandoned the sick from neighbors ‘From this fact, namely that the ill were abandoned by their neighbors.’ (Decameron, Boccaccio, 14th c.) Mensching (2000: 109)

Lexical prepositions could take infinitives and form adverbial infinitive clauses (see Rohlfs 1969: 93–102; Egerland 2010: 870–873): (67)

a. per fare la pace co li romani for do.INF the peace with the Romans ‘In order to make peace with the Romans.’ (Storie de Troia e de Roma, 13th c.)

Italian

237

b. senza fare alcuna dimora without do.INF any delay ‘Without any delays.’

(Fatti di Cesare, 13th c.)

Furthermore, unlike present-day Italian (Egerland 2010: 870), the old language allowed for the prepositions in (‘in’) and con (‘with’) to introduce infinitives directly; is still attested as late as the 19th century: (68)

a.

le loro case afforzorono con asserragliare le vie con legname the their houses reforced with close.INF the ways with wood ‘They reinforced their houses by blocking the paths with wood.’ (Cronica delle cose occorrenti ne’ tempi suoi, Compagni, 14th c.)

b.

cioè in dare speranza that-is in give.INF hope ‘That is, in giving hope.’ (De Amore di Andrea Cappellano volgarizzato, 14th c.) Egerland (2010: 870)

c.

La guerra finí con riconoscere tutti il nuovo duca The war ended with recognize.INF all the new duke ‘The war ended by all accepting the new duke.’

(Manzoni, 19th c.) Rohlfs (1969: 101)

Fluctuation in the presence or absence of prepositions – or prepositional complementizers, in certain cases – is attested in historical Italian, in similar circumstances to those attested for other Romance languages. Firstly, as Rohlfs (1969: 83, 98) and Egerland (2010: 819) remark, an infinitive introduced by di (‘of’) could alternate with a bare infinitive. Observe the following group of examples: (69)

a. Neuno è sì vecchio che non si creda di potere No-one is this old that not refl believes of can.INF vivere un anno live.INF a year ‘No one is so old not to believe themselves to be able to live one more year.’ (Fiori di filosafi e d’altri savi e d’imperadori, 13th c.) b. non credo potere avere il regno di Cielo not believe can.INF have.INF the kingdom of heaven ‘And I do not believe to be able to have the kingdom of heaven.’ (Il Libro de’ vizî e delle virtudi, Giamboni, 13th c.)

238

Crosslinguistic perspective: Romance languages II

A second type of variation is attested with categories that may combine with either a (‘to’) or di (Egerland 2010: 863): (70)

a. è obligato naturalmente di dire verità is forced naturally of say.INF truth ‘He is naturally forced to tell the truth.’ (Libro de’ vizî e delle virtudi, Giamboni, 13th c.) b. è obbligato a rendere guiderdone is forced to offer.INF recompense ‘He is forced to provide a recompense.’ (Tesoro di Brunetto volgarizzato da Bono Giamboni, 13th c.)

As mentioned above, variation in the presence or absence of preposition is also attested with nouns (non-clausal DPs), showing that nouns could be licensed by certain predicates with and without a preposition, as in historical Spanish. Consider the following additional examples (Egerland 2010: 92): (71)

a. mi misi a cercare di questa donna me put to seek.INF of this woman ‘I started to look for this woman.’ (Vita nuova, Dante, 14th c.) b. molte volte l’andai cercando many times him-walked.1SG seeking ‘I was looking for him many times.’ (Vita nuova, Dante, 14th c.) c. Neun uomo può giudicare de le cose che debbono avenire No-one man can judge.INF of the things that must come.INF ‘No man can judge what is to come.’ (Libro de’ vizî e delle virtudi, Giamboni, 13th c.) d.

pregandoli che giudicassero la mia visione begging-them that judged the my vision ‘Begging them to judge my vision.’

(Vita nuova, Dante, 14th c.)

Historical Italian infinitival clauses, introduced or not by prepositional complementizers, can be assumed to be nominal CPs (AgrP for Mensching 2000; see the discussion on this issue for Spanish, Portuguese, and French), save for the usual cases of restructuring and other clausal reduction (see Egerland 2010: 830–835). In sum, as categories appearing in typically nominal contexts, historical Italian infinitives must therefore check Case.

Italian

239

Infinitives could be introduced by determiners and thus participate in nominalizations/substantivizations in historical Italian (Rohlfs 1969: 80–82; Vanvolsem 1983), as all the other languages examined in this book. Consider the following examples of VP nominalizations, as they have no arguments (see the subject expressed as PP in (72c)), may be modified by adjectives (72c–e), and can be pluralized (72e) (see especially Vanvolsem 1983: 136–149). They accept determiners other than the definite article (72d) and even no determiner at all (72e), which are bare NPs (see Vanvolsem 1983: 113): (72)

a. Ne l’andare dee l’uomo essere savio di non andare In the-walk.INF must the-man be.INF wise of not walk.INF roppo piano too slow ‘The man must be wise when walking and avoid walking too slowly.’ (Fiori e vita di filosafi, 13th c.) b. col guatare uccide with-the look.INF kills ‘It kills with its look.’ (Quaresimale fiorentino, Giordano da Pisa, 14th c.) Egerland (2010: 871) c.

lo suo dolcissimo salutare the his very-sweet greet.INF ‘His very sweet greeting.’ (Vita nuova, Dante, 14th c.)

d.

per quello dormire for that sleep.INF ‘For that sleeping.’ (Quaresimale fiorentino, Giordano da Pisa, 14th c.) Egerland (2010: 875)

e.

di belli donari of beautiful give.INF.PL ‘Of good gifts.’ (Novellino, 13th c.)

Egerland (2010: 876)

Nominalizations/substantivizations higher than the VP level are also attested (Vanvolsem 1983: 87–104, 149–165; Egerland 2010: 877–879). Consider the following examples:

240 (73)

Crosslinguistic perspective: Romance languages II

a. e’n sul far frutto and-in its-the do.INF fruit ‘And in its producing fruit.’ (Rime, Petrarca, 14th c.) Vanvolsem (1983: 87) b. non ci torrà lo scender questa roccia not us will-prevent.3SG the descend.INF this rock ‘It will not prevent us from descending this rock.’ (Inferno, Dante, 14th c.) Egerland (2010: 878) c. l’onestamente andare the-honestly walk.INF ‘Walking honestly.’ d.

(Decameron, Boccaccio, 14th c.) Vanvolsem (1983: 154)

L’usare la dimistichezza d’uno uomo una donna The-use.INF the confidence of-a man a woman ‘For a woman to show the confidence of a man.’ (Decameron, Boccaccio, 14th c.) Vanvolsem (1983: 156)

e.

l’aver vinta una colomba the-have.INF beaten a dove ‘The fact of having defeated a dove.’ (Decameron, Boccaccio, 14th c.) Vanvolsem (1983: 152)

f. Dell’aver io parlato ben mi son pentuto Of-the have.INF I spoken well me am regretted ‘I am sorry to have spoken well.’ (Ammaestramenti degli antichi latini e toscani, Bartolomeo da San Concordio, 14th c.) Egerland (2010: 877) (73a) is clearly a vP nominalization since the verb has its own direct object but the subject is expressed via a possessive determiner. (73b) likewise shows an infinitive with a direct object, thus qualifying at least as a vP nominalization. Since the subject is not overt, it is possible to assume that it must be PRO and thus it could be assumed that the level should be higher than just vP. The same reasoning applies in (73c), which contains an infinitive modified by an adverb and arguably with a covert PRO subject. (73d) contains an overt nominative subject, which is postverbal, also typical of historical Spanish. (73e) contains a compound tense. (73f) shows a nominalized infinitive with auxiliary movement, an overt nominative subject, and adverbial modification. The same categorial

Italian

241

discussion presented for Spanish can be extended to these Italian examples, as (73d,e,f) point to at least TP level nominalizations if not higher, especially due to the presence and licensing of the subject, which, in agreement with Kornfilt and Whitman (2011), would signal a level above TP. For Mensching (2000: 108), postverbal subjects in historical Italian can be licensed in situ, that is, inside TP. The presence of auxiliary verbs in both (73e) and (73f) seems to require a nominalization point higher than just TP. The fact that the auxiliary (and only the auxiliary) is fronted provides additional evidence for the higher level of nominalization/substantivization. If Aux-to-Comp has applied (see Mensching 2000: 109–111 for discussion), then it would be a genuine case of CP nominalization. In any case, there is no doubt that (73e,f,g) qualify as high nominalizations/ substantivizations exactly as in Spanish.

3.1.2 Prepositional finite clauses The situation in historical Italian is similar to that of Spanish, Portuguese or French, in that che-clauses and indirect interrogative finite clauses selected by verbs, nouns or adjectives otherwise requiring prepositions were generally introduced directly without prepositions (Salvi 2010: 946), as the following examples illustrate: (74) a. Jo so che tu ti meravegli, che io non cognosandoti, I know that you you marvel, that I not knowing-you, ti ho saludado you have.1SG greeted ‘I know you are amazed that, even though I do not know you, I have greeted you.’ (Storia di Apollonio di Tiro, 14th c.) Frenguelli (2002: 19) b. Non ti maravigliare se li uoimini vanno a Dio Not you marvel if the men go to god ‘Do not be surprised if men go to god.’ (Fiori e vita di filosafi, 13th c.) Barbera (2010: 1011) c.

mandârno il bando che tutti li poveri andassero alla riva sent.3PL the message that all the poor went to-the shore ‘And they sent the message that all the poor people should go to the shore.’ (Novellino, 13th c.) Frenguelli (2002: 19)

242

Crosslinguistic perspective: Romance languages II

d.

avendo novelle che’l suo figliuolo era coronato re di Buemia having news that-the his son was crowned king of B. ‘Getting the news that his son was cronwed as king of Bohemia.’ (Cronica delle cose occorrenti ne’ tempi suoi, Compagni, 14th c.) Frenguelli (2002: 20)

e.

Ciascuno è certo che la natura umana è perfettissima Everyone is certain that the nature human is most-perfect di tutte l’altre nature of all the-other natures ‘Everybody is sure that the human nature os the most perfect of all.’ (Convivio, Dante, 14th c.) Frenguelli (2002: 20)

f. non è certo a che la cosa debbia pervenire not is sure to what the thing must arrive.INF ‘It is not sure where the thing must get to.’ (Rettorica, Latini, 13th c.) Munaro (2010: 1152) g.

Ben può Hector essere sicuro che, s’io lo troverò, Well can Hector be.INF sure that, if-I him will-find, elli ucciderà me o io lui he will-kill me or I him ‘Hector can be completely sure that, if I find him, he will kill me or I will kill him.’ (La storia di Troia, Binduccio dello Scelto, 14th c.)

The OVI corpus does not retrieve any examples of prepositional indirect interrogative finite clauses either. Even though the general construction is non-prepositional, there is some evidence of the preposition a taking a che-clause already in the 14th century: (75)

a. fino a che la loro virtude non agguaglia alla fama until to that the their virtue not equals to-the fame ‘Until their virtue equals fame.’ (Ammaestramenti degli antichi latini e toscani, Bartolomeo da San Concordio, 14th c.) b. tutte attendendo miravano a che Lia o a parlare o a all waiting looked to that Lia or to speak.INF or to partirsi si disponesse leave.INF-refl refl prepared ‘All waiting were looking forward to Lia’s getting ready to speak or leave.’ (Comedia delle ninfe fiorentine, Boccaccio, 14th c.)

Italian

c.

243

aspettando a che riuscir volesse waiting to that go-out.INF wanted.3SG ‘Hoping that he would want to go out.’ (Decameron, Boccaccio, 14th c.)

Those prepositional contexts are attested without a. Notice the prolepsis in (76b): (76) a.

in fino che egli non fosse risuscitato in until that he not were resuscitated ‘Until he resuscitated.’ (Esposizione del Simbolo degli Apostoli, Domenico Cavalca, 14th c.)

b.

aspettando a Guanto, ch’el venisse waiting to Guanto that-he came ‘Waiting for Guanto to come.’ (Centiloquio, A. Pucci, 14th c.)

In addition to this, the existence of adverbial clauses introduced by a lexical preposition once again evidences the fact that che-clauses could be prepositional objects (see Medici 1978: 338–344; Vegnaduzzo 2010: 806–809; Zennaro 2010: 953; among many others) and check Case accordingly. Consider the following examples: (77)

a. perché l’altre parole siano meglio intese For-that the-other words are better understood ‘So that the other words can be better understood.’ (Rettorica, Latini, 13th c.) Vegnaduzzo (2010: 807) b. dacché ebbero inteso quel che le Virtú voleano from-that had.3PL understood what that the Virtues wanted ‘Once they understood what the Virtues wanted.’ (Libro de’ vizi e delle virtudi, Giamboni, 13th c.) Barbera (2010: 991) c.

prima che partisse quella before that left that ‘Before that one would leave.’ (Convivio, Dante, 14th c.) Fiumara (2006: 176)

d.

Sanza che’n noi trovasse trieva o patti without that-in us found.3SG truce or pacts ‘Without getting any truce or pact from us.’ (Il Fiore e il Detto d’Amore, 14th c.)

244

Crosslinguistic perspective: Romance languages II

e.

purché mal non n’avenga for-that bad not of-it-comes ‘So that no harm is caused.’

(Tesoretto, Latini, 13th c.) Vegnaduzzo (2010: 809)

The fact that che-clauses could check Case is supported by their attestation in other typically nominal argumental positions, including subject and direct object positions (Salvi 2010). The same applies to indirect interrogative finite clauses (78c): (78) a.

è ben vero che’l regno di Cielo . . . non si può conquistare is well true that-the kingdom of heaven not refl can conquer.INF ‘It is clear that the kingdom of heaven cannot be conquered.’ (Libro de’ vizi e delle virtudi, Giamboni, 13th c.) Salvi (2010: 942)

b.

Disse ancora beato Agostino che dDio è fine di tutte le cose Said also beatus A. that god is end of all the things ‘Augustus said too that god is the end of all things.’ (Cronica fiorentina, 13th c.) Salvi (2010: 945)

c.

la madre il domandò . . . che nel consiglio fosse fatto the mother him asked what in-the council was done ‘His mother asked what had been done in the council.’ (Fiori e vita di filosafi, 13th c.) Munaro (2010: 1152)

A further piece of evidence of the argumenthood of the che-clause is several contexts comes from the grammaticality of extraction out of an embedded clause (Munaro 2010: 1157): (79)

a. Che vuoli tu ch’io ti doni . . . ? What want you that-I you give ‘What do you want me to give you?’

(Novellino, 13th c.)

b. come credi che’l fuoco de l’amor divino ch’è di how think.2SG that-the fire of the-love divine that-is of virtù maravigliosa lavori nell’anima? virtue marvelous works in-the-soul ‘How do you think that the fire of divine love, which is of marvelous virtue, works in the soul? (Libro de’ vizi e delle virtudi, Giamboni, 13th c.)

Italian

245

The existence of argumental clauses in early Italian does not imply the absence of a multifunctional, pragmatically interpreted che – exactly as in Spanish, Portuguese and French – which in certain contexts appears to act as a non-prepositional alternate of adverbial clauses introduced by lexical prepositions (Tekavčić 1980: 465–466; Salvi 2010: 769–771; Barbera 2010: 983–987). Consider the following examples (Tekavčić 1980: 465): (80)

a. Noi eravam partiti già da ello, ch’io vidi . . . We were left yet from it, that-I saw ‘We had already left, when I saw. . .’ (Inferno, Dante, 14th c.) b. Andiam che la via lunga ne sospigne Go that the way long of-it impels ‘Let us go, because the long way impels us to.’ (Inferno, Dante, 14th c.)

Going back to prepositional combinations, several prepositions admitted a construction with an intervening pronoun – similar to French but not analyzable as a clausal determiner – to form adverbial clauses (Medici 1978: 338), even coexisting with direct selection as is the case of perché vs. perciocché (Rohlfs 1969: 179). Observe the following examples: (81)

a. per ciò che ella amava ben lui for this that she loved well him ‘Because she loved him very much.’ (Decameron, Boccaccio, 14th c.) Dardano (1995: 42) b. in ciò che Tulio l’ appella antichissimo sì dimostra che non in this that Tulio it-calls very-old refl shows that not sia da credere is from believe.INF ‘In the fact that Tulio calls it very old it is proven that it is not believable.’ (Rettorica, Latini, 13th c.)

Among them, a ciò che is usually cited as a final or causal complementizer (Rohlfs 1969: 182; Tekavčić 1980: 472; Pantiglioni 2010: 1088). Tekavčić (1980: 472) argues that the che-clause is adjoined (in apposition) to the pronoun ciò:

246

Crosslinguistic perspective: Romance languages II

(82) canto sì amorosamente a ciò che sia gaudente lo meo sing.1SG thus lovely to this that is happy the my corag[g]io di bona speranza heart of good hope ‘I sing this lovely so that my heart be happy with good hope.’ (A pena pare ch’io, Messer Jacopo Mostacci, 13th c.) A ciò che also combines with categories that usually require a. Observe the examples in (83). The purposive connotation, as has been discussed for other clauses in other Romance languages – several à ce que-clauses in French or a que-clauses in Spanish, for instance – is not that obvious: (83)

a. Conviene a ciò che l’uomo abbia dal cominciamento Is-convenient to this that the-man has from-the beginning buoni costumi good customs ‘It is advisable that men have good habits from the beginning.’ (Tesoro di Brunetto Latini volgarizzato da Bono Giamboni, 13th c.) b. intende a ciò, che si dica, che questa . . . hae in sè virtude intends to that that refl says that this has in it virtue ‘He wants them to say that this one has virtue inside.’ (L’Ottimo Commento della Commedia, t. II Purgatorio, 14th c.)

The argumental interpretation is further supported by the fact that those predicates can be attested directly taking a che-clause, pointing to a syntactic analysis equivalent to that in (74) above: (84)

a. conviene che matematica sia partita in quattro scienze is-convenient that math is divided in four sciences ‘For that reason, it is necessary to divide math in four sciences.’ (Rettorica, Latini, 13th c.) b. intende che si facciano tra loro le paci intends that refl do.3PL among them the peaces ‘He wants them to make peace.’ (Orosio, Giamboni, 13th c.)

A search in the OVI corpus retrieves examples of a similar construction with di, namely di ciò che (lit. ‘of this that’):

Italian

(85)

247

a. I Savj si maravigliarono di ciò che lo vidono sì abbaìto The Wise refl marveled of that that him saw thus sad ‘The Sages were amazed to see him so depressed.’ (Libro dei Sette Savj di Roma, 13th c.) b. Josafat avea grande allegrezza di ciò che ’l Signore . . . avea Josafat had great happiness of this that the lord had difesa la veritade defended the truth ‘Josafat was very happy that the lord had defended the truth.’ (Leggenda Aurea, 14th c.) c. molto tristi di ciò che non trovavano nulla very sad of that that not found nothing ‘Very sad not to find anything.’ (Leggenda Aurea, 14th c.) d.

dolente era di ciò, che li due tribuni li avevano tolta la hurt was of that that the two courts him had taken the gloria del rifiutare glory of-the refuse.INF ‘He was hurt that the two courts had denied him the glory of a rebuttal.’ (Fatti di Cesare, 13th c.)

e.

quello era malvagio segno, di ciò che la bestia si fuggì that was evil sign of that that the beast refl fled del suo sacrificio of-the its sacrifice ‘That was a bad sign that the beast had escaped its sacrifice.’ (Fatti di Cesare, 13th c.)

f. ornamento di sentenze è una dignitade la quale proviene embellishment of sentences is a principle the which comes di ciò che in una diceria si giugne una sentenza . . . of this that in a speech refl adds a sentence ‘Sentence embellishment is a principle which comes from the fact that in a speech a sentence is added. . .’ (Rettorica, Latini, 13th c.) As happens with equivalent combinations in Spanish de que or French de ce que, di ciò che may add a causal connotation, clearly seen in the following example:

248 (86)

Crosslinguistic perspective: Romance languages II

È Giraldo degno di pena di ciò che commise furto? Is Giraldo deserving of penalty of this that committed theft ‘Does Giraldo deserve a penalty for stealing?’ (Rettorica, Latini, 13th c.)

However, firstly, some examples do not allow the causal connotation (see (85e, f ), for instance); secondly, as in the case of aciocché, several predicates without the preposition still maintain the causal connotation, which indicates that the causal connotation may not derive from di ciò che being a specialized “causal subordinating conjunction” (a causal complementizer). Rather, the analysis can be the same as for the examples in (74), especially considering the examples with me/aravigliarsi in (74a,b). Consider also the following additional examples: (87)

a. fue troppo dolente che T. iera tornato in Cornovaglia was too hurt that T. was returned in Corwall ‘He was very hurt that T. had retuened to Conrwall.’ (Tristano Riccardiano, 13th c.) b. avendo allegrezza che tanto bene volesse Stazio al suo having happiness that so-much well wanted Statius to-the his maestro Virgilio teacher Virgil ‘Being very happy that Statius loved his teacher Virgil so much.’ (Commento al Purgatorio, Francesco da Buti, 14th c.)

The internal syntactic structure of the group

has traditionally been described as that of a pronoun to which a che-clause adjoins (for instance, Tekavčić 1980: 472). In historical Italian, ciò is indeed attested in other pronominal contexts, including prepositional object (without a che-clause) and antecedent of a relative clause, where quello che (lit. ‘that which’) is also attested (see Benincà and Cinque 2010: 500–501): (88)

a. pentuto di ciò, fu liberato per l’orazione di costui sorry of that, was freed for the-prayer of that ‘Sorry about that, he was released thanks to his prayer.’ (Leggenda Aurea, 14th c.) b. ciò ch’io dico this that-I say ‘What I say.’ (Libro di buoni costumi, Paolo da Certaldo, 14th c.)

Italian

249

Notice the fact that di ciò in (88a), is analyzable as the complement of pentuto, not as a causal adjunct; the same pentuto combines with a clause in (85c) above, where it has however been analyzed as a causal adjunct by some linguists. The fact that the PP is a complement in (88a) supports the argumental analysis of the clause in (85c). Ciò also appears in cataphoric uses referring to a finite clause – che-clauses (89a), adverbial perché (89b) or even indirect interrogative clauses introduced by se (89c) – a construction likewise possible in other Romance languages: (89)

a.

per ciò è appellato controversia che diversi diversamente for this is called controversy that different differently sentono di quel detto o di quel fatto feel of that say or of that fact ‘It is called controversy because different people feel differently about a say or a fact.’ (Rettorica, Latini, 13th c.)

b.

per ciò s’era cieco perché non potea guardare le for this refl-was blinded because not could.3SG look.IN F the femine sanza carnale desiderio di peccare women without carnal desire of sin.INF ‘He had blinded himself because he could not look at women without feeling the carnal urge to sin.’ (Fiori di filosafi, 13th c.)

c.

di ciò era questione, se doveano essere puniti o no of this was question if must.3PL be.INF punished or not ‘There was the question whether they had to be punished or not.’ (Rettorica, Latini, 13th c.)

Unlike the evolution of French

, in present-day Italian speakers do not use any type of di ciò che or aciocché as French speakers do with de ce que and à ce que, as will be discussed in the next section. The pronominal nature of ciò seems to be acceptable in older stages of Italian, but it has nevertheless been reduced across time as ce did in French. Although ciò has never become a determiner in Italian, unlike ce in French, its pronominality has declined, being completely substituted by quello (‘this’) in relative clauses, for instance, and not being possible as prepositional object any longer either (see Vanelli 2010: 356–357). In sum, the data presented above confirm that prepositions could take finite clauses as their complements in historical Italian, including several cases of (argumental) a che. Furthermore, both che-clauses and indirect interrogative

250

Crosslinguistic perspective: Romance languages II

finite clauses are found in typically nominal contexts, as is the case in all the other Romance languages studied in this book. Extraction also confirms that che-clauses could be argumental already in early Italian. While the number of prepositional finite clauses is reduced, it has been shown that several predicates could indeed take both a prepositional and non-prepositional che-clause and thus license it with or without the preposition. The fact that Italian never truly developed a prepositional finite clause system as in Spanish and Portuguese, however, cannot be due to any lack of argumental clauses or nominality of the finite clause in early Italian, contra the assumptions in Barra (2002). The same nominal complementizer proposed for Spanish, Portuguese, and French can be justified for Italian. The same theoretical conclusion reached for finite clauses, namely the nominal complementizer for che-clauses and indirect interrogative finite clauses, and the nominality of the infinitival clause (and its licensing) are applicable in historical Italian. The data show that infinitival clauses could additionally participate in nominalization/ substantivization at different levels. Several indirect interrogative finite clauses, but not che-clauses, could be introduced by an article (Renzi 2010: 311). Observe the following examples extracted from the OVI corpus: (90)

a.

ascolta il perché hear the why ‘Hear the reason why.’ (Filocolo, Boccaccio, 14th c.)

b. poi che vi piace di sapere il perché. . . io sia venuto for that you pleases of know.INF the why I am come ‘Since you want to know why I have come.’ (Filocolo, Boccaccio, 14th c.) The role of

is much more reduced in Italian than in French but as in the latter the presence of ciò cannot be to block Case checking by the finite clause. Firstly, there are many adverbial clauses which would violate such constraint. Secondly, (at least) two of those pronominal combinations alternated with non-pronominal ones, namely aciocché and a che and, especially, percioché and perché. Italian shares several contexts such as prolepsis and cataphoric reference with Spanish as well.

3.2 Present-day Italian The data are extracted from scholarly works, from native informants, and online searches, as indicated. One example was extracted from BAnca Dati dell’Italiano

Italian

251

Parlato (BADIP), a free access corpus of the Karl-Franzens-Universität Graz in Austria.

3.2.1 Prepositions and infinitives Let us start by indicating that the Italian infinitive is grammatical in such typically nominal positions as subject and direct object, where by assumption they check (structural) Case. It can also be the complement of a noun or an adjective (see especially Skytte 1983: 322–349 for adjectives and Skytte 1983: 350–407 for nouns): (91) a.

Andare in vacanza potrebbe servire anche a te Go.INF in vacation could serve.INF also to you ‘Going on vacation would be useful for you too.’

b.

Piero vorrebbe andare in vacanza Piero would-want go.INF in vacation ‘Piero would want to go on vacation.’

c.

Skytte and Salvi (1991: 483)

La sua volontà di partire non poteva essere contrastata The his will of leave.INF not could be.INF contrasted ‘His intention of leaving could not be confirmed.’ Skytte and Salvi (1991: 545)

d.

Era ansioso di incontrarla Was anxious of meet.INF-her ‘He was anxious to meet her.’

Skytte and Salvi (1991: 551)

As for prepositional infinitives, Rizzi’s (1988) distinction between real PPs and prepositional complementizers in Italian. The former are depicted in the (92); the functional preposition is obligatory whether the object is an infinitive or other category (Acquaviva 1991: 637): (92) a.

Mi meraviglio di vederti così abbattuto Me marvel.1SG of see.INF-you thus depressed ‘I am amazed to see you this depressed.’

b.

Mi meraviglio del tuo comportamento Me marvel.1SG of-the your behavior ‘I am amazed at your behavior.’

252

Crosslinguistic perspective: Romance languages II

When the preposition is only used with infinitives and, in addition, the infinitival clause can be substituted with a regular direct object pronoun (93c, d), it qualifies as a prepositional complementizer (see Rizzi 1988; Skytte and Salvi 1991: 523, among many others): (93) a.

Gianni ha promesso di aiutarci Gianni has promised of help.INF-us ‘Gianni has promised to help us.’

b.

Acquaviva (1991: 635)

Gianni l’ha promesso Gianni it-has promised ‘Gianni has promised that.’

The other possible prepositional complementizer, a, seems to be subject to constraints similar to those mentioned by Huot (1981) for French. In any case, the answer lies in keeping a strict definition of prepositional complementizer as the prepositional-looking element which is used to introduce only infinitives. In those cases where a reappears with other categories or combines with a PP-verb which must change prepositions, as happens in French, it does not qualify as prepositional complementizer by definition. Certain predicates allow variation with either no meaning change (94a,b,c) or with some degree of it (94d,e) (see Skytte and Salvi 1991: 532–534): (94)

a.

Non mi fido a / di lasciarlo solo in casa Not me trust.1SG to of leave.INF-him alone in house ‘I don’t feel comfortable leaving him alone at home.’

b.

Ho sbagliato a prendere le tue misure Have.1SG mistaken to take.INF the your measures ‘I have made a mistake in taking your measures.’

c.

Ho sbagliato nel prendere le tue misure Have.1SG mistaken in-the take.INF the your measures ‘I have made a mistake in taking your measures.’ Skytte and Salvi (1991: 532)

d.

Non pensa ancora a sposarsi Not thinks still to marry.INF-refl ‘He does not plan to get married.’

Italian

e.

253

Penso di partire domani Think.1SG of leave.INF tomorrow ‘I think I’ll leave tomorrow.’

Skytte and Salvi (1991: 533)

Lexical prepositions forming an adverbial infinitive clause are grammatical: (95)

a. Vado a Milano per trovare un mio amico Go.1SG to Milan for find.INF a my friend ‘I am going to Milan to meet a friend of mine.’ Acquaviva (1991: 652) b. La riaccompagnò a casa senza dirle una parola Her accompanied.3SG to house without say.INF-her a word ‘He accompanied her back home without saying a word.’ Skytte and Salvi (1991: 483)

A characteristic that Italian shares with French, but not with Spanish or Portuguese, is the fact that argumental infinitives may only be introduced by a reduced number of functional prepositions. Only di, a, and some limited examples with da (‘from’) in fixed phrases such as (‘to have/give to do’) are allowed, as the following examples illustrate (see Skytte 1983: 487–536; Cinque 1990): (96)

a. *Contavo su essere onesto Counted.1SG on be.INF honest ‘I counted on being honest.’

Cinque (1990: 35)

b. si accontentò di guardarlo senza dire nulla refl satisfied of look.INF-him without say.INF nothing ‘He was satisfied with looking at him without saying anything.’ Maiden and Robustelli (2000: 369) c. L’hanno costretto a lasciare la scuola Him-have.3PL forced to leave.INF the school ‘The forced him to leave the school.’ d.

Skytte and Salvi (1991: 499)

Si dimenticò che aveva da finire la spesa Refl forgot.3SG that had from finish.INF the shopping ‘He forgot that he had to finish doing the shopping.’ Skytte and Salvi (1991: 532)

254

Crosslinguistic perspective: Romance languages II

As in French, some prepositional verbs change their required prepositions when they take an infinitive. Observe the following contrasting examples (Acquaviva 1991: 649): (97)

a. Conto di / *su essere a casa per la sette Count.1SG of on be.INF at house for the seven ‘I count on being at home by seven.’ b. Conto *dei / sui ritardi dei treni Count.1SG of-the / on-the delays of-the trains ‘I take the train delays for granted.’

Both examples include the same verb, but, while su may take non-clausal nouns, di is necessary for an infinitival clause. Notice that di is not acceptable for nouns. One of the options available in Italian consists of adding an article, exactly the opposite to Spanish. Consider the following examples, all of which would be ungrammatical without the determiner: (98)

a.

Non per questo cessava dal pensare a lui Not for this ceased.3SG of-the think.INF to him ‘Not for this did he stop thinking about him.’

b.

insistendo nel notare che il modo di essere. . . insisting in-the note.INF that the way of be.INF ‘Insisting on noting that the way of being.’ (Etica e politica, Croce) Skytte and Salvi (1991: 567)

c.

Finirai col farlo arrabbiare Will-end.2SG with-the do.INF-him anger.INF ‘You will end up making him angry.’

Skytte and Salvi (1991: 534)

The facts in (98) show that there is a syntactic difference between the presence and the absence of the determiner, that is, between projecting a DP or not, exactly what I have discussed for Spanish, independently of the nominal properties of the infinitive per se (Kayne 2000: 284) and despite it not being grammatical in all nominal contexts (Kayne 2000: 288). However, the problem must actually lie on the preposition and the construction rather than on the infinitival clause per se, given that infinitives with articles

Italian

255

are not always grammatical in all the logically possible and expected cases (see Salvi 1985: 264–266), such as the following ones: (99)

a. *Mi pento dell’aver parlato Me regret.1SG of-the-have.INF spoken ‘I regret having spoken.’

Skytte and Salvi (1991: 569)

b. *Mi dispiace dell’averlo offeso Me dislikes of-the-have.INF-him offended ‘I am sorry to have offended him.’

Salvi (1985: 264)

Salvi (1985: 264) indicates that the article is never possible in control configurations. However, note that in (99b), for instance, di is not even a true preposition but a prepositional complementizer and thus the article is likely to be ungrammatical simply on configurational grounds, as it would need to be located under the CP projection, not outside of it, as expected. What is more, a limited number of lexical prepositions, namely in, con and certain uses of da, require an article (Salvi 1985: 263), despite most lexical prepositions being generally grammatical without the article. Recall that the infinitival in (98b) above is argumental and still requires the article, contra Salvi’s (1985) hypothesis:9 (100)

a. Nel guardarlo, Anna trasse. . . un sospiro profondo In-the look.INF-him, Anna brought a sigh deep ‘While looking at him, Anna gave a deep sigh.’ (Menzogna e sortilegio, Morante) b. Col fargli continui dispetti With-the do.INF-him continuous snubs ‘By snubbing him all the time.’

Skytte and Salvi (1991: 568)

c. *Lo aiutò con inviargli denaro Him helped.3SG with send.INF-him money ‘He helped him by sending him money.’

Salvi (1985: 264)

9 Salvi’s (1985) comment is important in that it highlights once again the syntactic differentiation between the presence of the article (a DP) and the absence of it, as I did for Spanish, thus reinforcing the idea that the latter configuration does not project an empty D. In any case, the hypothesis about control works for several example sin Italian, but it does not seem to account for Spanish, as the introduction of an overt subject, and thus the elimination of PRO control, does not correlate with a grammatical use of the article with infinitives, which remains equally ungrammatical (in standard Spanish).

256

Crosslinguistic perspective: Romance languages II

Present-day Italian is rich in infinitival nominalizations/substantivi-zations (Vanvolsem 1983; Salvi 1985; Zucchi 1993), sharing many of the properties with Spanish and Portuguese. Following Kornfilt and Whitman’s (2011) model, first of all, Italian speakers can use VP nominalizations like the following: (101)

a. Giovanni udì il mormorare sommesso Giovanni heard the whisper.INF weak ‘Giovanni heard the weak whisper.’

Zucchi (1993: 219)

b. Lo scrivere interminabile di Sibila The write.INF neverending of Sibila ‘Sibila’s neverending writing.’

Zucchi (1993: 225)

The infinitives mormorare and scrivere are modified by adjectives and, crucially, have no arguments; in (101b) the subject is expressed by a PP. As Zucchi (1993: 224) remarks, (101a) is in principle ambiguous between a VP level and a vP level nominalization/substantivization, as the verb could as well take a direct object. Since there is no evidence of an argumental structure in the actual example, it is sufficient to claim that it is a nominalization at least at the VP level. There are clearer cases of vP nominalizations/substantivizations, including the following ones (Zucchi 1993: 228): (102)

a. Il suo mormorare sommessamente The his whisper.INF weakly ‘His weakly whispering.’ b. Il suo mormorare parole dolci The his whisper.INF words sweet ‘His whispering sweet words.’ c.

Quel suo mormorare parole dolci That his whisper.INF words sweet ‘That whispering sweet words of his.’

d.

Il suo continuo essere a corto di denaro The his continuous be.ING at short of money ‘His continuously being short of money.’

In (102b,c,d) the verbs must remain verbal up to the vP as witnessed by the presence of direct objects. In (102a) the adverb indicates that the nominalization

Italian

257

point must be higher than VP. In all these cases, the external argument is not expressed, being conveyed by means of a possessive determiner instead. There is a third type which seems to project up to TP while retaining the vP nominalization property of accepting adjectival modification. Consider the following examples: (103)

a. Il suo dover prender moglie The his must.INF catch.INF wife ‘His having to get a wife.’

Zucchi (1993: 234)

b. Il suo continuo dover cambiar casa The his continuous must.INF catch.INF wife ‘His continuously having to switch houses.’ c.

Il semplice essere statto interpellato The simple be.ING been addressed ‘The simple having been addressed.’

Zucchi (1993: 235)

For Zucchi (1993: 235), these are VP-infinitival NPs. However, the nominalization point must be higher than just VP, as evidenced, for instance, by the presence of an argumental grid. Given the grammaticality of modals and auxiliary verbs, the level of nominalization/substantivization must be high enough to include their projections. (103a,b) are also grammatical in Spanish, while the construction with a compound tense and an adjective (103c) seems highly unacceptable to ungrammatical in Spanish, and somewhat unacceptable in Portuguese (Brito 2010: 2). A last group of nominalizations/substantivizations seem to involve cases of Aux-to-Comp (Mensching 2000: 138), which point to their being CP level nominalizations (sentence-infinitival NPs for Zucchi 1993: 231): (104)

a. L’avere egli scritto quella lettera The-have.INF he written that letter ‘His having written that letter.’

Zucchi (1993: 220)

b. *Il suo avere egli scritto quella lettera The his have.INF he written that letter Lit. ‘His having he written that letter.’ c. *Il continuo essere egli a corto di denaro The continuous be.INF he at short of money Lit. ‘The continuous being he short of money.’

Zucchi (1993: 228)

258

Crosslinguistic perspective: Romance languages II

As expected of these high nominalizations, no adjectives or possessive determiners are allowed. Infinitives with articles can appear in a number of contexts, especially the ones that contain a larger nominal projection (VP, vP, and possibly TP with adjectival modification; see Salvi 1985: 266). They may especially appear in subject position; the object position is more restricted, as VP nominalizations are grammatical but CP/TP nominalizations are not (105e), which is syntactically interesting as it is unexpected of a DP; it also contrasts with the wider use of the infinitive without the article in Italian and with the grammatical Spanish counterpart in (105f). Consider the following examples: (105)

a.

Il vederti qui non mi sorprende The see.INF-you here not me surprises ‘Seeing you here does not surprise me.’

b.

Vederti qui non mi sorprende See.INF-you here not me surprises ‘Seeing you here does not surprise me.’ Skytte and Salvi (1991: 566)

c.

Gianni udì il mormorare sommesso del mare Gianni heard the whisper.INF weak of-the sea ‘Gianni heard the soft whispering of the sea.’

d.

Zucchi (1993: 253)

Non sopporto quell suo essere così smemorato Not stand.1SG that his be.INF thus forgetful ‘I can’t stand his being constantly forgetful.’

e.

*Non sopporto l’essere Giorgio così smemorato Not stand.1SG the-be.INF Giorgio thus forgetful ‘I can’t stand Georgio’s being constantly forgetful.’ Salvi (1985: 266)

f.

No pude soportar el haber tenido un retraso Not could.1SG stand the have.INF had a delay ‘I couldn’t stand having had a delay.’

Higher level nominalizations are generally limited to factive contexts (Salvi 1985: 265; Zucchi 1993: 248; but see Salvi 1985: 266 for additional constraints). Consider the following ungrammatical example of a CP nominalization (Skytte and Salvi 1991: 566):

Italian

(106)

259

*Era possibile l’essere egli rientrato in patria Was possible the-be.INF he returned in homeland Lit. ‘It was possible having he returned home.’

This context is also ungrammatical in Spanish (see Plann 1981: 209). However, as has been discussed above, such semantic constraints are not in place for prepositions since con and in, exhibit idiosyncratic behavior in requiring the article; other prepositions either accept the presence or absence of the article (tra, for instance); finally, many simply exhibit strict adjacency and admit no article, as in (standard) Spanish.

3.2.2 Prepositional finite clauses Italian groups with French in not allowing direct selection of the finite clause in argumental contexts. Consider the following examples with prepositional verbs: (107)

a.

Max si dimentica (*di) che ha promesso un viaggio a Pio Max refl forgets of that has promised a trip to Pio ‘Max forgets that he promised Pio a trip.’ Elia, Martinelli, D’Agostino (1981: 349)

b.

Luca si lamenta (*di) che mangia poco con Maria Luca refl regrets of that eats little with Maria ‘Luca regrets eating seldom with Maria.’ Elia, Martinelli, D’Agostino (1981: 375)

The same constraints apply to finite clauses with nouns and adjectives: (108)

a. Ciò dipende dal fatto (*di) che tu sei venuto That depends from-the fact of that you are come ‘That depends on the fact that you have come.’ Elia, Martinelli, D’Agostino (1981: 284) b. Sei certo (*di) che la cintura sia allacciata? Are sure of that the belt is fastened ‘Are you sure the belt is fastened?’ Maiden and Robustelli (2000: 324)

260

Crosslinguistic perspective: Romance languages II

Other PP-verbs which may take che-clauses directly include assicurarsi (‘to make sure’), dimenticarsi (‘to forget’), meravigliarsi (‘to marvel’), ricordarsi (‘to remember’), etc. (see Acquaviva 1991: 647). Extraction shows that the checlause can be argumental: (109)

a. ti dici nella mente che cosa ti ricordi che sta you say.2SG in-the mind what thing you recall.2SG that is dicendo il libro in quella pagina saying the book in that page ‘You say to yourself what you remember the book is saying on that page.’ (7-6-12) b. il quale insistono che sia in corso the which insist.3PL that is in run ‘Which they insist should be going on.’ (10-4-12)

Despite the examples in (107) and (108), it is clear that che-clauses are not barred from carrying and checking Case, as the adverbial clauses in (109) evidence. As a matter of fact, this point is highlighted in Manzini (2010: 186) as evidence of the nominal status of the Italian complementizer che. Observe the following examples: (110)

a. Perché possa comprare For-that can.3SG buy.INF ‘So that he can buy.’ b. Perché ha mentito For-that has lied ‘Because he has lied.’

Tekavčić (1980: 463)

c. Ciò è avvenuto senza che io potessi far nulla This is occurred without that I could do.INF nothing ‘This happened without my being able to avoid it.’ d.

Rizzi (1988: 510)

Ha fatto di tutto salvo che non ha voluto perdere l’onore Has done of all save that not has wanted lose.INF the-honor ‘He has done everything except that he did not want to lose his honor.’ Tekavčić (1980: 459)

Italian

e.

261

Dopo che Serena conobbe Mario After that Serena met Mario ‘After Maria met Mario.’

Giusti (1991: 726)

At the same time, as happens in Spanish, several potentially grammatical combinations are simply not used (Elia, Martinelli, D’Agostino 1981: 285–286): (111)

*in che /*su che/ *dentro che/ *sopra che/ *con che/ *contro che in that on that inside that over that with that against that

Moreover, for certain speakers of Italian the selected preposition a can take a che-clause resulting in an argumental prepositional finite clause as illustrated in the following examples (see Rizzi 1988: 509; Scorretti 1991: 160–161; among many others): (112)

a.

Ada bada (a) che tutto riesca Ada cares to that all succeeds ‘Ada takes care that everything turns out well.’ Elia, Martinelli, D’Agostino (1981: 295)

b.

Tengo (a) che sia avvertita Have.1SG to that is warned ‘I am anxious that she be informed.’

c.

Fa attenzione (a) che gli studenti siano preparati Make attention to that the students are prepared ‘Pay attention that the students are ready.’

Ledgeway (2000: 82)

The same possibility is available for nouns and adjectives: (113)

a.

Marì ha interesse (a) che Diego la perdoni Mari has interest to that Diego her forgives ‘Mari is interested in Diego’s forgiveness.’ Elia, Martinelli, D’Agostino (1981: 343)

b.

Sono contrario (a) che tu parta subito Am opposed to that you leave immediately ‘I am opposed to your leaving right away.’

Cinque (1990: 36)

262

Crosslinguistic perspective: Romance languages II

There is a very clear difference between the behavior of di and that of a (Scorretti 1991: 161), since di che is never grammatical. Acquaviva (1991) tries to capture the difference between a and di by arguing that the P a actually introduces a final adverbial clause, that is, an adjunct. In Italian a may introduce adverbial final clauses, thus showing different syntactic and semantic properties. Consider the following examples where a equals per (Acquaviva 1991: 652): (114)

a. Vado a Milano a trovare un mio amico Go.1SG to Milan to find.INF a my friend ‘I am going to Milan to meet a friend of mine.’ b. Vado a Milano per trovare un mio amico Go.1SG to Milan for find.INF a my friend ‘I am going to Milan to meet a friend of mine.’

However, there are many problems with that hypothesis. First, as Bertuccelli Papi (1991: 824) shows, can be argumental in certain cases, as it complements predicates which require a goal as their argument. Furthermore, as discussed for Spanish, Portuguese, and French, the adverbial interpretation is simply not obtainable in all cases; for instance, in the following case (Rizzi 1988: 519): (115) Sono favorevole (a) che tu parta Am favorable to that you leave ‘I am favorable to your leaving.’ What is more, not all the predicates that can take a che may substitute it by an actual final adverbial clause. Fare attenzione can (180), but, to my knowledge, aspirare, for instance, cannot (see Bertuccelli Papi 1991: 824 for similar evidence with infinitives): (116)

occorre fare la massima attenzione acciochè le feci occurs do.INF the maximum attention to-this-that the feces degli animali malati non contaminino i mangimi of-the animals sick not pollute the fodder ‘In general, it is necessary to pay maximum attention to avoid the feces of the sick animals polluting the fodder.’

Italian

263

In the end, it is important to remark that, unlike adjuncts, argumental a checlauses cannot freely combine with any predicate to express purpose, unlike acciocché, as all the cases of (112) and (113) paradigmatically require a with other categories; what is more, as Scorretti (1991: 160) points out, not all PPverbs which take a are actually used with a che by the speakers either.10 The situation with indirect interrogative finite clauses is similar to that of French. Se-clauses (whether/if-clauses) cannot be prepositional objects (Elia, Martinelli, D’Agostino 1981: 286; Fava 1991: 676), but they may be licensed without a preposition, as che-clauses can. Consider the following examples: (117) a.

Dubitava se gli fosse permesso Doubted.3SG if him were allowed ‘He doubted whether it was allowed to him.’

b.

Fava (1991: 676)

Lo tormentava il dubbio se gli convenisse . . . farlo Him tormented the doubt if him should do.INF-it ‘He was tormented by the doubt whether he should do it.’ Fava (1991: 715)

10 In general, the pattern of extraction with a che-clauses – but not with other types, as (109) proves – seems to be different from other Romance languages. Consider the following examples, which show that the interrogative strategy does not involve actual extraction out of the embedded clause, since the preposition is also “fronted” (probably, merged in situ) with the wh-phrase (Sandro Sessarego, pc): (1) a. A cosa sei favorevole che faccia? To thing are.2SG favorable that do.1SG ‘What are you favorable to me doing?’ b. A che cosa aspirate che faccia io da solo? To what thing aspire.2PL that do I from alone ‘What do you expect me to do alone?’ In any case, at the very least, given that there is no resumptive pronoun in the embedded clause, the displaced object of the embedded verbs (faccia in both cases) must be still syntactically accounted for, be with operator movement or otherwise, suggesting that the embedded clauses are not barriers, that is, they must be arguments. Moreover, while a che-clauses are usually described as disallowing extraction (see Cinque 1990: 37), it seems that under certain circumstances extraction is grammatical even with a: (2) Ci sono argomenti da approvare che tutti hanno interesse There are arguments from approve.INF that all have interest a che siano condotti a buon esito? to that are driven to good success ‘Are there arguments pending approval which all are interested in seeing successfully solved?’ (4-4-12)

264

Crosslinguistic perspective: Romance languages II

c.

dipende se gli da la merendina depends if him gives the snack ‘It depends on whether he gives him a snack.’ (BADIP, M E 6 273 A)

Despite the fact that no preposition can take an indirect interrogative with se, these clauses are pronominalized with the corresponding proform ne, as expected of a di-PP (Sandro Sessarego, pc): (118)

Se è già partito, non ne sono certo If is yet left not of-it am certain ‘Whether he has already left, I am not certain of it.’

As per their argumental nature despite the absence of the preposition, extraction is once again grammatical, which serves as confirmation: (119)

a. Non ricercare nulla che dubiti se sia giusto o ingiusto Not pursue.INF nothing that doubt.2SG if is fair or unfair ‘Don’t pursue anything you doubt whether is fair or unfair.’ (4-4-12) b. per il cielo che dubiti se sia reale o no for the sky that doubt.2SG if is real or not ‘For the sky which you doubt whether is real or not.’ (4-4-12)

In addition, se-clauses can appear in subject and direct object position, typically nominal positions where they must check Case: (120)

a. non e’ importante se il bambino non e’ ancora “pulito” not is important if the kid not is still clean ‘It is not important whether the kid is not yet clean.’ (4-4-12) b. Saprò se hai investito in azioni Will-know.1SG if have.2SG invested in stock ‘I will know if you have invested in stock.’

Fava (1991: 703)

Italian

265

On the other hand, indirect interrogative finite clauses introduced by whphrases are perfectly grammatical as prepositional objects (Fava 1991: 676), as the following examples prove: (121)

a.

Dubitava su che cosa gli fosse veramente permesso Doubted.3SG on what thing him was truly allowed ‘He doubted about what was truly allowed to him.’

b.

Fava (1991: 676)

dipende da perché non è venuto depends from why not is come ‘It depends on the reason why he came.’ (4-4-12)

c.

La questione di chi sia l’autore dell’Iliade The question of who is the-author of-the-Iliad ‘The question of who the author of the Iliad is.’ (4-4-12)

Contrary to the cases with a che, the preposition is considered to be mandatory in standard Italian (Fava 1991: 700), although a search online suggests that it may also be left out in equivalent contexts: (122) a.

Avevo dubitato *(di) che cosa gli sarebbe potuto accadere Have.1SG doubted of what thing him would-be could happen.INF ‘I had doubted about what could have happened to him.’ Fava (1991: 700)

b.

Kylar non dubitava cosa avrebbe fatto Logan Kylar not doubted thing would-have done Logan ‘Kylar had no doubts about what Logan would have done.’ (Oltre le tenebre. L’angelo della notte, Brent Weeks, accessed in Googlebooks) (4-4-12)

The final issue revolves around the combination with the article. While Spanish allows the article with que-clauses, Italian does not: (123) *Il che venga alla festa non significa niente The that comes to-the party not means nothing ‘His coming to the party does not mean anything.’

266

Crosslinguistic perspective: Romance languages II

As in Portuguese, the absence of the article with a che-clause cannot be interpreted as a sign of non-nominality.11 Italian che-clauses can be prepositional objects and appear in subject and direct object positions, as in Spanish or Portuguese, for instance: (124)

a. Vuoi che ti porti in braccio? Want.2SG that you carry.1SG in arm ‘Do you want me to carry you in my arms?’ Maiden and Robustelli (2000: 315) b. Che Luigi non sappia l’italiano è assurdo That Luigi not knows the-Italian is absurd ‘For Luigi not to speak Italian is absurd.’ Maiden and Robustelli (2000: 319)

Several indirect interrogative finite clauses can participate in nominalizations/substantivizations with articles, again exactly as in Spanish, Portuguese or French: (125)

a. Informati del come e il perché si è fatto Informed of-the how and the why refl is done ‘Informed about how and why it was done.’

Lapesa (2000: 551)

b. dipende dal perché depends from-the why ‘It depends on the reason.’ (4-4-12) c. Dipende dal perche’ del bacio Depends from-the why of-the kiss ‘It depends on the reason for the kiss.’ (4-4-12) d.

tutto dipende dal perchè si sia alzato all depends from-the why relf is rised ‘All depends on why he got up.’ (4-4-12)

11 The same must apply to se-clauses. Remember that in standard Spanish se-clauses cannot appear with an article despite being perfect prepositional objects. Any constraints must be due to factors other than nominality and Case (see Roussou 2010 for English if ).

Italian

267

3.3 Discussion and conclusions The data presented in the previous sections indicate that speakers of historical (Tuscany) Italian behaved similarly to the speakers of historical Spanish in limiting the number of argumental prepositional finite clauses; however, there existed no restrictions on Case checking. The early examples of a che-clause attested in historical Italian already, in addition to the grammaticality of prepositional che-clauses in adverbial clauses, permit to conclude that the operating constraint did not block Case checking with finite clauses. The fact that che-clauses could appear in other typically nominal positions further supports their nominal nature (see Castelli 1988: 333; Manzini 2010; Roussou 2010). Prepositional infinitival clauses and infinitival clauses with prepositional complementizers were also grammatical, with optionality/variation, and in more contexts than in the present-day language, especially given the current idiosyncratic behavior of in and con, which could nevertheless take an infinitive without the article in the old stages of the language. Verbs requiring a preposition other than a or di swap their prepositions, as happens in French. Once again, despite these constraints, infinitival clauses can be used as grammatical subjects and direct objects, thus pointing to their nominal nature. Italian is rich in infinitives with articles, displaying all levels of nominalization/substantivization. As in historical Spanish and especially French, Italian texts show the existence of pronominal alternatives, namely acciocché, di ciò che, perciò che or in ciò che. While a che could alternate with a ciò che in historical Italian but not in present-day Italian, di ciò che has simply disappeared. As discussed for French, the last resort value of the pronominal alternative is questioned because both a and per are attested with direct selection. The grammaticality of a che-clauses in cultured registers of Italian indicates that for certain speakers there continues to be no (full) constraint against argumental prepositional finite clauses, even if only for one preposition. In any case, the fact that the presence of a is optional – and the only option for those who do not accept a che-clauses and for all other selected prepositions, including di – shows that che-clauses can be licensed directly as arguments as in Spanish. Needless to say, the pattern is completed by the several lexical prepositions which take che-clauses in adverbial clauses, where lexical Case is thus checked, which Manzini (2010) in particular interprets as clear evidence for the nominal nature of the complementizer che in Italian. Importantly, not all lexical prepositions can take a che-clause, exactly as is the case for Spanish. Indirect interrogative finite wh-phrases can be prepositional, although seclauses cannot, as in French; likewise, some wh-phrases participate in further nominalizations, such as perché, as in Spanish, Portuguese and French. The

268

Crosslinguistic perspective: Romance languages II

che-clause cannot combine with a determiner as it does with some indirect interrogative finite clauses, but again this is shared by French and, more importantly, by Portuguese, whose rich argumental prepositional que-clauses serve to pinpoint that the absence of the article does not mean non-nominality, rather only the absence of a DP layer. In short, despite Italian speakers imposing certain constraints on the grammaticality of prepositional che-clauses, the configuration is attested in historical texts of Italian and is part of the grammar of a number of speakers. Additionally, Italian speakers could and can license che-clauses without the otherwise required preposition.

Chapter 7

Conclusions 1 A crosslinguistic overview of prepositional finite clauses The findings of this investigation concerning the presence or absence of prepositional clauses in Romance are summarized in Tables 1 and 2: Table key: Z Y/ Z Y

extensive grammatical use limited grammatical use ungrammatical or unattested

Argumental

Adjunct

P + que/checlause

P + indirect interrogative finite clause

P + infinitive

P + que/ checlause

P + infinitive

Spanish

Z

Z

Z

Z

Z

Portuguese

Z

Z

Z

Z

Z

French

Y de ce que à ce que

Y/Z Colloquial examples

Y/Z de, à

Z

Z

Italian

Y/Z a che

Y/Z Ungrammatical with se; grammatical with wh-phrases

Y/Z di, a, da Article required

Z

Z

Table 1: Prepositional clauses in present-day Romance languages

270

Conclusions

Argumental

Adjunct

P + que/checlause

P + indirect interrogative finite clause

P + infinitive

P + finite que/checlause

P + infinitive

Historical Spanish

Y/Z Some early examples; ~16th on

Y/Z Some early examples; ~16th on

Z

Z

Z

Historical Portuguese

Y/Z Some early examples; ~16th on

Y/Z ~17th on

Z

Z

Z

Historical French

Y/Z de ce que à ce que de quoi

Y

Y/Z de, à

Z ce / Ø

Z

Historical Italian

Y/Z a che

Y

Z

Z

Z

Table 2: Prepositional clauses in historical Romance languages

All of these languages allow adverbial prepositional finite clauses; any restrictions typically affect argumental clauses.

2 Change in Spanish from a comparative perspective. Data and theoretical implications After examining the syntax and history of prepositional clauses (both finite and infinitival) in several related languages, it is time to focus on Spanish and reach several conclusions regarding the case of Spanish prepositional finite clauses, especially the crosslinguistic implications on the validity of certain assumptions from previous works. In chapter 3 I summarized the three extant general approaches to the emergence of prepositional argumental finite clauses: 1. Analogy-based analysis; patterns extend from

to

with other additional patterns 2. Pronoun-drop analysis; the elimination of the intervening pronoun creates

Change in Spanish from a comparative perspective

3.

271

Case and nominal features; clauses are not nominal until the 16th century, when they can then receive Case and become prepositional objects.

In order to evaluate these approaches and find the mechanisms of change that may explain the creation and extension of prepositional argumental finite clauses, it is necessary to summarize and critique the following theoretical points: 1. Nominality; nominality and substantivization, the article 2. Pronominal alternatives 3. Prepositionality and argumenthood 4. Case and the finite clause

2.1 The nominality of the clause and (re-)nominalization/ substantivization The nominal nature of the finite clause is essential in this research for clear theoretical implications, especially, the role of Case Theory. It also affects the nature of – crosslinguistically valid – categorization in linguistics. Clauses may be nominal at two different levels, with syntactic and crosslinguistic repercussions. Firstly, clauses are distributionally grammatical in typically nominal contexts such as subject, direct object and prepositional objects. Therefore, que/ che-clauses are nominal in present-day Romance, as they were too in the past, despite additional existing constraints. Whether introduced by a (functional) preposition or licensed positionally, que/che-clauses (and their indirect interrogative counterparts) are nominal this way. In order to account for this, in keeping with recent research on the subject (Manzini and Savoia 2005, 2011; Manzini 2010; and Roussou 2010, in particular), I have argued for a [+nominal] feature categorially expressed in a nominal complementizer (que/che, empty or si/se; see Manzini 2010: 195 for Italian, extensible to other Romance languages), thus materializing the functional nominal category Kornfilt and Whitman (2011) require for clausal nominalizations. This nominal complementizer is – and has always been – essentially nominal, as the evidence presented for the historical periods suggests. Special constraints, not related to nominality, must be in place for idiosyncratic behavior. Given their also nominal distribution, infinitival clauses can be argued to be licensable in nominal positions. This can be achieved via a nominal complementizer heading their CPs, assuming an extensive analysis of infinitival clauses as CPs, save for restructured infinitives. The latter are nevertheless licensed in some

272

Conclusions

other way as part of the restructured monoclausal configuration, via assuming a mixed category for infinitival clauses and maintaining Raposo’s (1987a) analysis of a [+nominal] non-finite INFL or Raposo’s (1987b) nominalizer –r, via adding a null nominal affix (Yoon 1996: 339), or via a special nominal category topping the infinitival clauses materializing the functional nominal category. Given that the focus of this book is finite clauses, I will leave these options open for further research. In all these cases, the resulting projections are finite or infinitival CPs (or infinitival TPs/AgrPs) which are essentially equivalent to NPs, rather than DPs, perfectly grammatical as arguments (Dobrovie-Sorin, Bleam, and Espinal 2006: 60). Additionally, both finite and infinitival clauses may participate in a second process, labeled clausal nominalization, by co-appearing with a determiner and thus projecting a DP, again with interesting crosslinguistic results and implications. This is a second way in which clauses are nominal, namely by becoming DPs, exactly like regular NPs may also be bare or project a DP. The analysis following Kornfilt and Whitman (2011) has proven satisfactory, as it allows accounting for the different types of categorial nominalizations in Romance by hypothesizing categorially coherent projections in one same derivation and thus preventing exocentricity. Such a classification accounts for the nominal and verbal features especially visible in many infinitival constructions and solve the apparent categorial aporia created by the unclear properties of the infinitive. I have shown that VP and vP nominalizations create a highly nominal category out of an otherwise verbal (radical) category. While vP nominalizations seem to require a determiner, certain infinitival VP nominalizations can appear without a determiner (for instance, as a bare plural noun, as in Italian, see example (72e) in chapter 6, section 3.1.1.) again suggesting that these nominalizations do not need to project a DP to be nominal. The fact that vP nominalizations require a determiner may be due to the fact that modified nouns tend to require a determiner as well in Romance. On the other hand, higher types of nominalizations require more attention for several reasons. First of all, given the already nominal que/che-clause, adding a determiner and projecting a DP – a CP level nominalization for Kornfilt and Whitman (2011) – needs to be conceptualized as re-nominalization or substantivization (Yap, Grunow-Hårsta, Wrona 2011). While it would have seemed logical to ascertain the existence of an empty determiner position and regularize the categorial nature of CP as DP across the board, I showed that this is not the case. Evidence for this comes from the unexpected ungrammaticality of sentences such as the following:

Change in Spanish from a comparative perspective

273

(1) *Me acuerdo del que viniste a mi casa Me recall.1SG of-the that came.2SG to my house ‘I remember that you came to my house.’ While prepositions take nominal complements, it is surprising that

in Spanish is generally ungrammatical (although open to some dialectal variation). Strict adjacency allows describing such configuration and permits to do away with an empty determiner and reinforces the idea that que/che-clauses appear in nominal contexts even without the determiner. Additional evidence comes from Portuguese, a language which, despite sharing many properties with Spanish (in the present and in the past), allows no . What is more, there are clear syntactic consequences linked to the presence or the absence of the determiner projection in terms of extraction barriers, as Leonetti’s (1999a) proves. The co-appearance of the article with a que-clause in general Spanish is also semantically constrained. The important conclusion here is that whether a language can have more or less constructions with determiners and clauses – if any – is not directly related to the nominality of its finite and non-finite CPs (or TPs/AgrPs), which can nevertheless appear in typically nominal contexts without the determiner, contra the assumption in Kornfilt and Whitman (2011) whereby the determiner is the functional nominal category turning a verbal CP into a noun in the case of CP nominalizations/substantivizations. As previously discussed, Panagiotidis and Grohmann (2009: 150, fn 5) claim that English clausal –ing clauses require no empty determiner layer in order to be nominal(ized), an idea similarly expressed in Roussou (2010: 587) for finite clauses. The examination of indirect interrogative finite clauses reveals additional interesting differences as well. In general, indirect interrogative finite clauses appear in typically nominal contexts, regardless of the wh-phrases in Spec,CP. On the other hand, several of those wh-phrases can further combine with determiners. An examination of the resulting contexts reveals that in this case too there are different levels, as is the case with infinitives, from lower nominalizations of the PP por qué (lit. ‘for what’, ‘why’) into the noun porqué, to CP level nominalizations/substantivizations. Consider the following examples: (2)

a. sin porqué without why ‘With no reason.’ b. sin razón without reason ‘With no reason.’

274

Conclusions

c. sin porqués without whys ‘With no reasons.’ d.

sin razones without reasons ‘With no reasons.’

The fact that porqué can be used as noun without a determiner, like razón does, is additional evidence that creating a DP is not a conditio sine qua non for nominality. Furthermore, while porqué can be a lower level nominalization and behave syntactically as a real noun, the same does not extend to que, which seems to be constrained to participating only in CP level nominalizations/ substantivizations, possibly for its clitic/atonic/weak nature. Another difference between que/che-clauses and indirect interrogative finite clauses is located in the fact that the latter are grammatical with determiners not only in Spanish, but also in Portuguese, French, and Italian, open to dialectal variation and certain operating constraints, especially for si/se-clauses. Let’s remember for instance that in Spanish is not grammatical for all speakers (me inclusive). Furthermore, indirect interrogatives in Italian can be accompanied by an article, a configuration that is nevertheless ungrammatical in English or Swedish. The case of Swedish particularly relevant in terms of crosslinguistic theoretical assumptions and hypothetical implications/expectations in terms of nominality of CPs as Swedish nevertheless shares with Spanish the extensive grammaticality of prepositional argumental finite clauses (att-clauses in Swedish). Going back to Spanish prepositional finite clauses, the main argument in Barra (2002), the hypothesis that the absence of prepositional argumental queclauses in pre-16th century Spanish texts is due to the lack of nominality of the que-clause, is therefore too strong and cannot be maintained.

2.2 Case, pronominal alternatives, and the finite clause The data presented in this book support the conclusion that clauses, in being nominal, can carry and check Case features, extensively for lexical Case with lexical prepositions, or inherent Case, either positionally (see Bosque and Gutiérrez-Rexach 2009: 159; through the same mechanism that licenses default Case or adjunct Case, for instance) or, subject to other restrictions, through the materialization of a functional preposition or K (Lamontagne and Travis 1987; Travis and Lamontagne 1992; Tremblay 1996):

Change in Spanish from a comparative perspective

(3)

275

a. Functional preposition materializes Case

b. Case checked positionally

An alternative licensing option for (4b) is to assume the underlying preposition hypothesis (see chapter 2, section 2.2.2.), although I have already discussed the additional theoretical problems it raises. Furthermore, it seems simpler to argue, in line with the results for CP vs DP projections, that there is a representational syntactic difference when the preposition is overt (KP projects) whereas the absence of the preposition indicates absence of KP, particularly when it is possible to assume just positional Case licensing by assuming a Case feature in CP (possibly available already from the Numeration/Lexical Array). In any case, such covert preposition alternative is easily expressed following Travis and Lamontagne’s (1992: 166) analysis, since they argue for a Kase projection for Accusative Case, for instance, which is non-prepositional. The same mechanism could be applied here: (4) Empty, underlying preposition, empty Kase

Prepositional optionality was shown to be grammatical in all periods of Spanish, especially with que-clauses. A similar degree of optionality is found in Portuguese, while Italian and French show optionality but clearly tend to favor

276

Conclusions

the absence of the preposition in argumental clauses, without forgetting the existence of a che in present-day Italian, de quoi in Classical French, de ce que and à ce que in French, etc. The reality of non-prepositional licensing finds additional support in a Germanic language whose prepositional clauses are similar to Spanish: Swedish. As Ralph (1975: 670) claims, for many PP-verbs and adjectives, such as tänka på (‘to think of’) or lycklig över (‘happy about’), the preposition is optional when taking an att-clause (see also Holmes and Hinchliffe 2003: 325). As in Spanish, optionality does not apply across the board; for instance, the preposition is mandatory with (non-clausal) nouns (Ralph 1975: 676–677). Likewise, indirect interrogative finite clauses in English may be introduced by a preposition which nevertheless can be left out in many contexts as well. In sum, whether DP or CP (even TP/AgrP in the case of some infinitival clauses), Romance clauses must be checking Case, at the very least when in nominal contexts. The same must have applied to historical periods of the languages, where I showed that clauses could appear in subject and direct object positions. It has also been shown that in older stages of Spanish, Portuguese, Italian, and French (even in Late Latin) clauses could be prepositional objects, e.g., que/che-clauses introduced by prepositions in adjunct position (adverbial clauses). As a consequence, any constraints, namely in argumental prepositional finite clauses, must not stem from nominality-dependent factors preventing Case checking. Another aspect that has been linked to Case is the presence of the pronominal alternative, as opposed to direct prepositional selection. As discussed above, the classical example is French de ce que or à ce que, which have been interpreted in the literature, as reviewed above, as que-clauses adjoined to the pronoun ce. However, it was shown above that such analysis is untenable, for several reasons. First of all, Zaring (1992) proves that ce stopped being analyzable as a pronoun already in old French and was reanalyzed as a determiner; certainly, in present-day French ce is not used as a pronoun any longer. What is more important, textual evidence shows that ce was also used in adverbial clauses in alternation with direct selection; thus, speakers were not using it as a last resort to prevent the CP from checking Case. Furthermore, even clearer cases of pronominal intervention, as in historical Italian and Spanish, have been shown to have coexisted with direct selection (e.g., a ciò che vis-à-vis a che; por aquello que vis-à-vis porque). Additional evidence comes from Swedish again (Andersson 1973: 3). Observe the following examples: (5)

a. Efter det att han avslutat sin examen fick han ett jobb After that that he finished his exam got he a job ‘After finishing his exam, he got a job.’

Andersson (1975: 51)

Change in Spanish from a comparative perspective

277

b. Vi akte till Ipswich efter att vi hade beslutat oss We traveled until I. after that we had decided us ‘We traveled to Ipswich after we had made our decision.’ Holmes and Hinchliffe (2003: 423) c.

Hans enda anspråk på ryktbarhet bestod däri, att His only claim to fame consisted therein, that han en gang hade räddat ett drunknande barn he one time had saved a drowning child His only claim to fame consisted in his once having saved a drowning child.’ McClean (1963: 213)

d.

Hans enda anspråk på ryktbarhet bestod i att han. . . His only claim to fame consisted in, that he ‘His only claim to fame consisted in the fact that . . .’ Erich Round (pc)

Both lexical and functional prepositions, both argumental and adjunct clauses allow direct selection or pronominal intervention in Swedish, which provides additional evidence outside the Romance realm that the presence of the pronoun is not intended as a mechanism to prevent Case on a clause. A third important issue revolves around the link between prepositionality and argumenthood. Barra (2002) argues that finite clauses became arguments once they could be prepositional. This implies a link between argumenthood and the presence of prepositions with clauses. Such a hypothesis has immediate crosslinguistic implications as well. This matter can be approached from two perspectives. Firstly, Leonetti (1999b) proves – for present-day Spanish but applicable to other stages and other (Romance) languages – that prepositionality does not necessarily correlate with argumenthood. For instance, from the moment that the prepositional finite que-clause introduced by functional de which is modifying a noun such as hecho (‘fact’) is determined to be an adjunct, not a complement, the attestations of those particular prepositional clauses in Spanish texts of the 16th century are to be analyzed as equally adjunct despite participating in the general increment of frequency in prepositional finite clauses. That is to say, not all the changes involving prepositions and finite clauses are a sign of argumenthood. The immediate consequence is that the hypothesis that stems from Barra’s (2002) analysis is weakened. Secondly, the grammaticality of extraction out of the embedded clause highlights that sentential complementation was a fact already before the 16th century; that is to say, que/che-clauses were argumental. While it is typically

278

Conclusions

remarked that syntax was “looser” in the older stages of the language, as the multifunctionality of que-clauses indeed reveals, it is nevertheless also true that pragmatics-dependent subordination continues to be perfectly grammatical in present-day Spanish and other languages, as evidenced above (remember also Batllori, Sánchez, and Suñer 2001). The link between nominality, Case, and argumenthood extends beyond Spanish. From a crosslinguistic perspective, Barra (2002: 399) suggests that the purported absence of argumental prepositional finite clauses in French implies that that language lacks argumental clauses altogether. Leaving aside Zaring’s (1992) conclusions about French de ce que and à ce que and subsequent discussion establishing that such combinations can be indeed argumental, it is impossible to maintain that such is the case even if French speakers are more limited in the types of required prepositions that may take a finite clause. On the one hand, French is as rich as Spanish is in terms of adverbial

, where again the clause must be the complement of the lexical preposition. Furthermore, extraction out of the embedded clause was attested in historical French and continues to be widely available in present-day French, thus indicating that sentential complementation is part of the grammar. The French que-clause is nominal as it is in Spanish and Portuguese, any remaining constraints deriving from some other factors. The conclusions reached for French are valid for Italian as well. To sum up, an analysis of prepositional argumental finite clauses based on the lack of Case and nominality fails to account for all relevant contexts where (finite) clauses are grammatical. Case Theory as presented in this book is expected to be open to further revisions with time, even to the point of rejecting Case as a linguistic feature with a prime role in licensing. Such is the view purported by McFadden (2004), for instance, as briefly explained in chapter 2, section 3.1.4., who argued for an EPP feature instead of abstract Case. Whatever the theoretical construct, what will always remain is the data presented above showing combinations between prepositions and clauses and clauses in other typically nominal contexts.

2.3 E-language and syntactic change Once the role of Case and nominality has been removed as the factors behind the emergence of prepositional argumental finite clauses, the examination of the data reveals that the change is a question of the surface. Inasmuch as the role of (functional) prepositions (Ks) is semantically and syntactically reduced, beyond creating a KP not necessary when the preposition is not present, there

Change in Spanish from a comparative perspective

279

seems to be no significant syntactic or semantic change beyond the presence of the functional preposition/materialization of K (Case) between the state of affairs before the rise in frequency of argument prepositional finite clauses, when queclauses are licensed positionally, and after. What mechanism(s) lie(s) behind the rise in frequency? Where is change located? In order to solve these questions, it is important to step back and look at the data from a general perspective to see what we can learn from the chronology of the configuration in Spanish. The first point that needs to be highlighted is the fact that the syntactic configuration is already present before the 16th century. Consider the following early examples (see especially Serradilla 1996a): (6)

A que-clause a. el mj pacto que yo. . . tengo a que non oujesse fijo que. . . the my pact that I have to that not had son that ‘The deal I have regarding the fact that he would not have a son who. . .’ (Biblia romanceada judio cristiana, 14th c.) b. los otros seran embiados a que complan de derecho the others will-be sent to that fulfill.3PL of right ‘The rest will be sent to fulfill their duties.’ (Leyes del estilo, 14th c.) c. por que vlixes ayudara a que elena fuese dada a menalao for that Ulysses helped to that Helen was given to M. ‘So that Ulysses would assist in giving Helen to Menelaus.’ (Sumas de la historia troyana, Leomarte, 14th c.)

(7)

De que-clause a. dixo Turín al rrey que bien cierto fuesse de que eso said Turin to-the king that well true was of that that poco que él sabía que gelo diría little that he knew that him-it would-say ‘Turin told the King that he should stay reassured that he would tell him whatever he knew.’ (Libro de los Estados, Don Juan Manuel, 14th c.) Serradilla (1996a: 14) b. quel gradeciesen de que no los mandaua matar that-him thanked of that not them ordered kill.INF ‘That they should thanbk him for not ordering their assassination.’ (Primera Crónica General, Alfonso X, 13th c.) Serradilla (1996a: 21)

280

Conclusions

c.

diciendo las fablas de que los el uenciera saying the stories of that them he beat ‘Telling the stories about his having beaten them.’ (Primera Crónica General, Alfonso X, 13th c.) Serradilla (1996a: 96)

(8) En que-clause penso en que fuessen hasta Hierusalem thought.3SG in that went.3PL until Jerusalem ‘He thought about them going to Jerusalem.’

(Libro de los azores, 13th c.)

What the previous examples indicate is that the syntax of prepositional finite clauses was already present and available for (certain) speakers before the frequency increase documented in 16th century texts. Let us first concentrate on the combination a que-clause. As indicated above, the combination a que and its equivalents in other Romance languages have been considered as adverbial clauses by some linguists. The main reasons to support such a hypothesis are their final or purposive connotations and the existence of proper final or purposive a que, which is not selected by any predicate. There are indeed examples in historical Spanish where a que introduces a proper final adverbial clause, as the preposition is not part of the thematic grid of any predicate (Sánchez 1999: 274): (9)

mandoles que fuesen adelante a que le atajasen las tierras sent.3SG-them tha went.3PL ahead to that him explored the lands ‘He sent them ahead to explore the land.’ (Gran Cronica de Alfonso XI, 14th c.)

Likewise, examples (6b,c) above allow a final connotation. From a crosslinguistic point of view, equivalent examples of both types are attested in early French and Italian. (10)

Clearly final adverbial clauses a. Labourez A ce que tout soit nettement Ordonné Work to this that all is neatly organized ‘Work so that everything is neatly organized.’ (Estoire de Griseldis, 14th c.)

Jensen (1974: 37)

b. Addunque odi la dottrina, ad ciò che tu abbia scienzia Thus heard.1SG the doctrine to this that you have science ‘Thus I listened to the doctrine, so that you have science.’ (Dei trattati morali, Andrea da Grosseto, 13th c.)

Change in Spanish from a comparative perspective

281

(11) A selected by a predicate a. il furent consillié a ce que il se renderoient a lor They were advised to this that they relf surrendered to their dame la roine lady the queen ‘They were advised to surrender to their lady the queen.’ (Chroniques, Jean Froissart, 14th c.) b.

se questa scienza intende. . . a ciò che noi siamo buoni if this science intends to this that we are good ‘If this science has the intention of making us good.’ (Reggimento de’ principi di Egidio Romano, 13th c.)

The ambiguity extracted from the previous examples invites the conclusion that the change here is due to a reanalysis of adjuncts into arguments at some point in early historical Romance languages. Such is the hypothesis in Schulte (2003, 2004, 2007), who argues that Romance argumental infinitives result from former adjunct infinitives reanalyzed as complements, later favored by analogy. Certainly reanalysis alone cannot account for all infinitival clauses. While the semantic connotation was and is present in, for instance, alegrarse de hacer algo (‘to be happy about doing something’), it is not in acostumbrar a hacer algo (‘to be used to doing something’), yet it is present in the old language. In any case, a previous reanalysis of PP-verbs via infinitival clauses would necessarily mean that those PP-verbs must have already been syntactically ready to take a finite clause with their required preposition; that is, once reanalyzed, there would be no need for a second reanalysis targeting que-clauses. The reanalysis of a que-clause is not sufficient to account for all the contexts in Spanish. First of all, the use of a que-clause as a truly final adverbial clause was not as common in Spanish as it was – and still is – in Italian or French (see Sánchez 1999: 318; Herrero 2005: 382–384). That is to say, the configuration a que-clause is not attested following the necessary verbs and other categories to support reanalyses for all of them (not even in French or Italian; see Herman 1963: 220 for Italian). Secondly, many of the predicates in (6) are also attested with a non-prepositional que-clause while still maintaining the final connotation, which shows that the apparent semantic/adverbial connotation is not located (only) in the actual preposition. Enviar (‘to send’), for instance, could take an infinitive without the preposition as well:

282

Conclusions

(12) a.

fue emplazado que viniesse a oyr el Riepto was ordered that came.3SG to hear.INF the duel ‘He was ordered to come hear the duel.’ (Ordenamiento de Alcalá, Alfonso XI, 14th c.)

b.

embiol que viesse aquel destruymjento sent-him that saw.3SG that destruction ‘He sent him to see that destruction.’ (General Estoria V, Alfonso X, 13th c.)

c.

ayudándole que faga sienpre tales cosas helping-him that does.3SG always such things ‘Helping him do those things forever.’ (Siete partidas, Alfonso X, 13th c.)

Thirdly, not all combinations permit a final connotation. This is exactly what Sánchez (1999: 318) notes for ayudar a que (‘to help that. . .’), which is attested early. The same applies to other combinations attested later on, such as aspirar a que (‘to aspire that. . .’) or acostumbrarse a que (‘to get used to the fact that. . .’). A reanalysis does not seem likely for these combinations. Fourthly, the same connotation which can be obtained in the previous medieval examples is maintained in present-day Spanish. PP-verbs such as ayudar or enviar include the a-complement in their theta-grids as ditransitive verbs. Reanalysis does not necessarily eliminate ambiguity (for instance, going to in English), but the current argumental examples permit extrapolating this analysis to similar contexts in the old language. In sum, the previous comments suggest that the examples in (6) can be also analyzed as argumental. Since there is no reason to support the claim that the early examples in (6) were not already argumental, the reanalysis must have taken place before. In any case, the syntax of the que-clause was already favorable in the medieval period, as extraction proves that sentential complementation was already in place before the 16th century, and the que-clause was perfectly nominal. The fact that a que/che-clauses, even those with the same final connotation, do permit extraction in the present-day languages serves as indirect support that integration took place. Chronologically, a que-clauses with PP-verbs increase in frequency from the 16th century on, which suggests an introduction of the preposition a in a variety of contexts thanks to pre-existing prepositional models and to the availability of argumental que-clauses. A search in the Corpus del Español reveals that aguardar a que (‘to wait until. . .’), obligar a que (‘to force something/someone

Change in Spanish from a comparative perspective

283

to. . .’), aspirar a que (‘to aspire for someone to. . .’), mirar a que (‘to watch out that. . .’), enseñar a que (‘to teach someone to. . .’), consentir a que (‘to agree that. . .’), etc. are first attested in the 16th century, once again indicating that, as of the available evidence, for these verbs there are no previous contexts where a reanalysis could have in principle been possible. Furthermore, the absence of the prepositions is also attested, and even in the same text (13b): (13) a.

Aviso y torno a avisar que nadie desmaye Advise.1SG and turn.1SG to advise.INF that nobody faints ni dexe de ser en la corte bueno nor stops of be.INF in the court good ‘I urgently advice everybody not to weaken or stop being good in the palace.’ (Menosprecio de corte y alabanza de aldea, Antonio de Guevara, 16th c.)

b.

no aspiro a que ella induzga a maravilla, sino que a not aspire.1SG to that she induces to marvel but that to mi calor se proporcione me heat refl provides ‘I do not expect it to be marvelous, but rather to provide my heat.’ (Rimas, Leonardo de Argensola, 16th c.)

c.

nos obliga que les guardemos respeto como a Santos us forces that them keep.1PL respect like to saints ‘It forces us to show them respect like saints.’ (Diálogos teológicos, Tirso de Molina, 17th c.)

In addition, reanalysis is out of the question for verbs and other categories introduced after the 16th century, as is the case for propender/propenso (17th century for propenso, see Corominas 1998: 449), which nevertheless enter the language copying the already existing model of a for all their complements, including the que-clause. Notice too that in present-day Spanish it is grammatical for certain speakers not to use the preposition: (14)

a. propenso a que en el Reino floreciesen las Artes prone to that in the kingdom blossomed the arts ‘Prone to a blossoming of the arts in the kingdom.’ (Cartas eruditas y curiosas, Feijoo, 18th c.)

284

Conclusions

b. He dicho que la legislación debe propender a que Have.1SG said that the legislation must tend.INF to that cada labrador no labre sino la tierra que. . . each farmworker not works but the land that. . . ‘I have said that the legislation must tend to reflect the principle that each farmworker should only work the land that. . .’ (Informes al Consejo sobre la Ley Agraria, 18th c.) c.

¿Quién es propenso que le de obesidad . . . Who is prone that him gives obesity ‘Who is prone to be obese?’

(21-4-12)

As for de que, the reanalysis option was already contemplated by Tarr (1922: 239, 259, fn 4) but it seems more doubtful. First of all, unlike French, the common adverbial use of de que was temporal (‘since’, ‘the moment that’; see Herrero 2005: 228–231) rather than causal: (15)

fuesse de que sopo lo que auja acaeçido left.3SG-refl of that knew the that had happened ‘He left once he knew what had happened.’ (Exemplario por ABC, 15th c.)

Causal interpretations are indeed possible for examples (7a,b) above (see in addition Herrero 2005: 92). Consider also the following additional example (Serradilla 1996a: 88): (16) erró mucho de que non metió sus fijas consigo mistook.3SG much of that not took his daughters with-him ‘He made a big mistake by not taking his daughters with him.’ (Conquista de Ultramar, 13th c.) In contrast, the causal connotation is difficult to obtain in (7c). Furthermore, as with a que-clause, the adverbial connotation continues to be operative in present-day Spanish, but crucially not for all types either, as no causal connotation is possible for acordarse de que (‘to remember that. . .’) or el hecho de que (‘the fact that. . .’), where the de que-clause is actually a prepositional adjunct (see chapter 4, section 4.5.). It is the same issue raised by Bat-Zeev Shyldkrot’s (2008) analysis of French de ce que as a causal complementizer; however, the syntax of proper causal adverbial clauses, e.g. porque clauses, is different from

Change in Spanish from a comparative perspective

285

de que-clauses, as the former, when argumental, allow extraction despite potential causal connotations, whereas the former does not: (17)

a. ¿Qué te alegras de que hiciera Pepe? What you get-happy.2SG of that did Pepe ‘What are you happy Pepe did?’ b. *¿Qué te alegras porque hiciera Pepe? What you get-happy.2SG because did Pepe ‘What are you happy because Pepe did?’

Chronologically, de que-clauses are profusely attested from the 16th century; the first attestations are not only restricted to clearly isolatable context(s) – that is, those with possible causal connotations – which could have undergone adjunct-to-argument reanalysis. Actually, the preposition is present in contexts where no causal connotation is feasible (18a,b), yet it is already attested in the 16th century: (18)

a. la tristeza es causa de que se dexe de hacer the sadness is cause of that refl stops of do.INF ‘Sadness leads to quit doing things.’ (Luz del alma cristiana, Felipe de Meneses, 16th c.) b. la idea de que el buen tiempo se afianzaría the idea of that the good weather refl would-establish ‘The idea that the good weather would establish.’ (Epistolario, Ginés de Sepúlveda, 16th c.) c.

muy maravillado de que aquel cavallero quisiesse tomar very amazed of that that knight wanted take.INF contienda con él solo fight with he only ‘Very amazed at the fact that that knight wanted to statr a fight with him alone.’ (Espejo de príncipes y cavalleros, Ortúñez de Calahorra, 16th c.)

Moreover, the causal connotation is also obtainable even when the preposition de is absent, as the following set of examples demonstrates:

286

Conclusions

(19) a.

muy maravillado que criatura tuviesse tan soberano poder very marveled that creature had so sovereign power ‘Very amazed that someone could have such powers.’ (El Crotalón, Cristóbal de Villalón, 16th c.)

b.

que estuviese alegre que él era su padre that were.3SG happy that he was his father ‘That he were happy that he was his father.’ (Suma y narración de los incas, Betanzos, Juan de, 16th c.)

The reanalysis option is even more difficult to support for en que-clauses. En que-clauses could be used to introduce temporal clauses, and not very frequently (Herrero 2005: 231–232). Moreover, it is implausible that the combination insistir en que (‘to insist that. . .’) would be the result of a reanalysis from an adjunct temporal clause into an argumental clause; yet such configuration is attested once again in 16th century texts (20a). The preposition may still be absent (20b,d): (20)

a.

Vuelvo a insistir en que no he abandonado Return.1SG to insist.INF in that not have.1SG quit ‘I insist again that I have not quit’ (Epistolario, Ginés de Sepúlveda, 16th c.)

b.

insistían que hiciesen rostro contra ellos insisted.3PL that did.3PL face against them ‘They insisted in facing them.’ (Crónica del Perú, Cieza de León, 16th c.)

c.

sin reparar en que fuesen espurios o ilegítimos without notice.INF in that were.3PL spurious or illegitimate ‘Without noticing whether they were spurious or illegitimate.’ (Política indiana, Solórzano Pereira, 17th c.)

d.

Sin reparar que eran espurios o ilegítimos Without notice.INF that were.3PL spurious or illegitimate ‘Without noticing that they were spurious or illegitimate.’ (Política indiana, Solórzano Pereira, 17th c.)

e.

sin reparar que fuessen de los bastos y grosseros without notice.INF that were.3PL of the rude and coarse ‘Without noticing whether they were rude or coarse.’ (Historia de la conquista y población de la provincia de Venezuela, Oviedo y Baños, 18th c.)

Change in Spanish from a comparative perspective

287

Finally, regarding con que-clauses, a search in the Corpus del Español shows examples of argumental con que-clauses in the 16th century. Consider the following examples of contentarse con (‘to be satisfied with’), also with prepositional optionality: (21) a.

no se contentaba con que el día antes se hubiese not refl contented.3SG with that the day before refl had despedido dél bid-farewell of-him ‘He was not satisfied with the fact that the day before he had bid farewell to him.’ (Crónica de la Nueva España, Cervantes de Salazar, 16th c.)

b.

aún no se contentava que lo tuviessen por señor still not refl contented.3SG that him had.3PL for lord ‘He was still not satisfied with their having him as their lord.’ (Espejo de príncipes y cavalleros, Ortúñez de Calahorra, 16th c.)

Con que used in adverbial contexts is attested in consecutive clauses in the 16th century (Herrero 2005: 535–536), particularly frequent from the 17th century on. It is likewise difficult to see the possible reanalysis in this case, both on semantic and chronological grounds. A reanalysis is also out of the question with contar con (‘to count on’), as it is a late formation, itself a grammaticalized, semantically non-compositional PP-verb: (22)

a. contando con que el aumento de riqueza . . . proporcionará counting with that the increase of wealth will-provide la construcción de canales the construction of channels ‘Counting on the fact that the increase in wealth will make the consturction of channels financially viable.’ (Elementos de economía política con aplicación particular a España, Marqués de Valle Santoro, 17th c.) b. es forzoso contar con sus pasiones is forceful count.INF with his passions ‘It is imperative to count on his passions.’ (Elementos de economía política con aplicación particular a España, Marqués de Valle Santoro, 17th c.)

288

Conclusions

c. contaba con dispersar todo el ejército counted.3SG with disperse.INF all the military ‘He counted on dispersing the whole military.’ (Correspondencia, Jovellanos, 18th c.) The general overview of the data leads to two important conclusions. Firstly, one of the possible mechanisms of syntactic change cannot account for all the changes with all the relevant prepositional configurations, if any. Reanalysis is not the mechanism that can explain the innovative prepositional que-clauses. It may have been the mechanism that created some PP-verbs in the past (see Schulte 2004 for infinitives), but not out of the configuration

. Secondly, it is clear that the relevant syntax was already available for certain speakers as there are several early examples of prepositional combinations beyond adverbial clauses (here contact via translations may have been a factor as well), sentential complementation was in place, and the que-clause was nominal and could be positionally licensed as argument, which nevertheless remained a possibility after the change. There must have been a different mechanism whereby prepositional verbs and sentential complement (and adjunct) taking nouns and adjectives must have transposed the preposition used with other categories, which was more common and was also increasing in the case of infinitives (Schulte 2004), resulting in a change in the materialization of inherent Case with finite clauses. The innovation is not located in the underlying syntactic analysis but rather on the surface, via regularization of existing patterns. Let us observe the following two sets of examples: (23)

a. el que teme ser odioso, no sabe gobernar the who fears be.INF hateful not knows rule.INF ‘He who fears to be hateful does not know how to rule.’ (Crónica del Reino de Chile, Mariño de Lobera, 16th c.) b. teme que le dé con ellos fears that him give.1SG with them ‘He fears that I will hit him with them’ (La madre de la mejor, Lope de Vega, 16th c.) c. teme de recaer fears of relapse.INF ‘He is afraid of relapsing.’ (Coloquios espirituales y sacramentales, González de Eslava, 16th c.)

Change in Spanish from a comparative perspective

d.

289

temiendo de que no le matasen fearing of that not him killed ‘Fearing that they would kill him.’ (Crónica del Perú, Cieza de León, 16th c.)

(24) a.

de que haya vencido a don Quijote de la Mancha, póngolo of that has beaten to don Quixote of the Mancha, put.1SG-it muy en duda very in doub ‘That he has beaten Don Quixote, I doubt it very much.’ (Quijote, Cervantes, 17th c.) Tarr (1922: 238)

b.

De que seáis mujer, no me lo podéis negar Of that are.2PL woman, not me it can.2PL deny.INF ‘That you are a woman, you cannot deny it to me.’ (Las dos doncellas, Cervantes, 17th c.)

Tarr (1922: 242)

Both sets show striking properties unusual in present-day Spanish. In the first set, the verb temer (‘to fear’), attested with prepositional optionality with infinitives throughout the Middle Ages, and which also accepts a que-clause, at a certain point extends the preposition to the que-clause, and creates a regularized prepositional pattern ungrammatical in present-day Spanish, both with que-clause and with infinitives. There is no semantic difference between (23b) and (23d). The second set is even stranger as it shows the extension of a de which would otherwise be expected to appear only with infinitives as prepositional complementizer (see Serradilla 1995: 152, fn 12 for examples of this de que in subject position). The resulting configuration is ungrammatical in present-day Spanish as well. These sets show two different generalizations of de, whether a real functional preposition or just a prepositional complementizer with a theoretically complicated twist; that is to say, based surface-dependent generalizations. Analogy is such surface-dependent mechanism of change (see Harris and Campbell 1995; Joseph 1998; Wanner 2006). Not being a mere diachronic correspondence, analogy provides the channel whereby speakers multi-copied the prepositional frame and spread it to more contexts from other previously prepositional constructions, such as

or

, in agreement with previous works on the subject (see chapter 3, section 2.3.1.). Barra’s (2002) critical view of analogical change in syntax is understandable, given that the seeming asystematicity of analogy is difficult to reconcile with the concept of I-language. However, it is possible to reconcile analogy with formal syntax by relocating this mechanism in E-language. Such move

290

Conclusions

joins the list of recent studies such as Roberts and Roussou (2003), van Gelderen (2004), Batllori et al. (2005), and Roberts (2007) in accommodating approaches and mechanisms traditionally reserved for non-formal frameworks. The same attempt is found in Adger and Smith (2005), who formalize variation. Fischer (2007) argues in favor of a combination of both formal and functional perspectives on change. In the case of analogy, Kiparsky (2000) has already dealt with the formalization of morphological analogy from a non-Chomskyan perspective. Analogical change does not necessarily apply across the board and speakers may skip certain possible candidates in a group (Wanner 2006), as evidenced by the fact that certain expected combinations are nevertheless ungrammatical. A good example is se souvenir (‘to remember’) in French, or even some gaps in adverbial clauses (see Herman 1963 and Barra’s 2002 constraint of 4-dimensional prepositions). As an E-language phenomenon, such exceptional gaps in analogy are a logical possibility, not a theoretical problem. Additionally, in certain cases other factors must be in place, such as semantic restrictions (selection restrictions), in agreement with the general provisos established by Manzini (2010), as in the case of hablar de que (lit., ‘to speak of that. . .’) vs *hablar sobre que (lit., ‘to speak about that. . .’), which crucially cannot be related to Case or nominality at all, as both take indirect interrogative finite clauses, infinitives, and non-clausal DPs. The sporadic – but nevertheless grammatical – examples attested before the 16th century evidence that some speakers could produce the target configuration. Such examples, probably the exploratory expressions Harris and Campbell (1995: 72–75) talk about – with additional help from the extensive existence of

early in the language in adverbial clauses (see chapter 3, section 2.3.1.) – established the syntactic foundations which helped the spread and regularization caused by multiple analogical changes, in line with the traditional view of this change (Tarr 1922; Herman 1963; Serradilla 1995, 1996a, 1996b, 1997). The increase in prepositional infinitives documented in Schulte (2004) may also help explain the timing of this change. From such E-language output, new learners of the language must have incorporated the new prepositional configurations in their I-languages, to the point that such changes became to be firmly established for many speakers (see the discussion of the role of E-language for I-language change in Lightfoot 2006), without forgetting, nevertheless, the maintenance of the older non-prepositional alternative, which is especially common for several native speakers of Spanish in American countries. Table 3 summarizes the proposed analogical syntactic change and other relevant syntactic factors:

Change in Spanish from a comparative perspective

Early attestations Syntax is ready Early examples of a que, de que, en que + nominal que-clause + clausal argumenthood + adverbial prepositional que-clauses

291

Extension/syntactic analogy Implemented from 15/16th centuries on )

Difference in materialization of (inherent) Case Overt linker Facilitated by pre-existing syntax

Table 3: Analogical syntactic change in Spanish

Semantically intact, the new prepositional configurations did affect the phonological output of the derivation; in a majority of cases, it affected the materialization of inherent Case, that is, it resulted in the overt syntactic projection of KP: (25) Visualization of change in the materialization of inherent Case

In other cases, the preposition does not qualify as KP, showing that it is indeed surface that counts here. For those PP-verbs selecting a lexical preposition, as might be the case in preocuparse por (‘to take care of’), the introduction of prepositions is configurationally accounted for already on independent grounds, as (non-selected) lexical prepositions were extensively taking que-clauses already in adverbial clauses anyway. In any case, a search in the Corpus del Español shows that preocuparse por is first attested with a que-clause in the 20th century.

References Abney, Steven. 1987. The English Noun Phrase in its sentential aspect. Ph.D. diss., Massachusetts Institute of Technology. Abraham, Richard D. 1938. A Portuguese Version of the Life of Barlaam and Josaphat. Paleographical Edition and Linguistic Study. Philadelphia: University of Pennsylvania. Acquaviva, Paolo. 1991. Frasi argomentali: Completive e soggettive. In Grande grammatica italiana di consultazione. II. I sintagmi verbale, aggetivale, avverbiale. La subordinazione, Lorenzo Renzi and Giampaolo Salvi (eds.), 633–674. Bologna: Il Mulino. Adger, David. 2003. Core Syntax. A Minimalist Approach. Oxford: Oxford University Press. Adger, David and Jennifer Smith. 2005. Variation and the Minimalist Programme. In Syntax and variation: Reconciling the biological and the social, Leonie Cornips and Karen P. Corrigan (eds.), 149–178. Amsterdam/Philadelphia: Benjamins. Alexiadou, Artemis, Gianina Iordăchioaia, and Florian Schäfer. 2011. Scaling the variation in Romance and Germanic Nominalizations. In The Noun Phrase in Romance and Germanic. Structure, Variation and Change, Petra Sleeman and Harry Perridon (eds.), 25–40. Amsterdam/Philadelphia: Benjamins. Allan, Robin, Philip Holmes, and Tom Lundskær-Nielsen. 2000. Danish: An essential grammar. London/New York: Routledge. Allen, Cynthia L. 2002. Case and Middle English genitive noun phrases. In Syntactic Effects of Morphological Change, David W. Lightfoot (ed.), 57–80. Oxford: Oxford University Press. Allen, Cynthia L. 2010. Substantival adjectives in the history of English and the nature of syntactic change. In Grammatical Change: Theory and Description, Rachel Hendery and Jennifer Hendriks (eds.), 9–25. Canberra: Pacific Linguistics. Alrenga, Peter. 2005. A sentential subject asymmetry in English and its implications for complement selection. Syntax 8 (3): 175–207. Alsina, Alex. 2001. Is Case another name for grammatical function? Evidence from object asymmetries. In Objects and other subjects: Grammatical functions, functional categories, and configurationality, William D. Davies and Stanley Dubinsky (eds.), 77–102. Dordrecht: Kluwer Academic Publishers. Ambar, Maria Manuela. 1994. “Aux-to COMP” and lexical restrictions on verb movement. In Paths Towards Universal Grammar: Studies in Honor of Richard S. Kayne, Guglielmo Cinque, Jan Koster, Jean-Yves Pollock, Luigi Rizzi and Raffaella Zanuttini (eds.), 1–23. Georgetown: Georgetown University Press. Ambar, Maria Manuela. 1999. Infinitives vs. participles. In Semantic Issues in Romance Synta, Esthela Treviño and José Lema (eds.), 1–20. Amsterdam/Philadelphia: Benjamins. Andersen, Henning. 1973. Abductive and deductive change. Language 49 (4): 567–595. Andersson, Lars-Gunnar. 1973. Some subordinating constructions in Swedish. Gothenburg Papers in Theoretical Linguistics 17: 1–22. Andersson, Lars-Gunnar. 1975. Form and function of subordinate clauses. Ph.D. diss., University of Göteborg. Andersson, Anders-Borje. 1974. The NP status of that-clauses and infinitives. Gothenburg Papers in Theoretical Linguistics 24: 1–16. Anttila, Raimo. 2003. Analogy: The warp and woof of cognition. In The Handbook of Historical Linguistics, Brian D. Joseph and Richard D. Janda (eds.), 425–440. Malden: Blackwell. Anula Rebollo, Alberto and Marina Fernández Lagunilla. 1997. Infinitivos nominales y verbales en español. Espéculo 5: 1–11.

References

293

Arsenijević, Boban. 2009. Clausal complementation as relativization. Lingua 119 (1): 39–50. Asbury, Anna. 2008. The morphosyntax of Case and adpositions. Ph.D. diss., Utrecht University (LOT). Asbury, Anna, Jakub Dotlačil, Berit Gehrke, and Rick Nouwen (eds.). 2008. Syntax and Semantics of Spatial P, Amsterdam/Philadelphia: Benjamins. Authier, J. Marc and Lisa A. Reed. 2010. On the categorial status of French à/de ce que, University of Pennsylvania Working Papers in Linguistics 16 (1): 10–20. http://repository.upenn. edu/pwpl/vol16/iss1/3 Badia Margarit, Antonio. 1947. Los Complementos Pronominalo-adverbiales Derivados de ibi e inde en la Península Ibérica. Madrid: Consejo Superior de Investigaciones Científicas. Baker, Mark. 2003. Lexical Categories. Verbs, Nouns, and Adjectives. Cambridge: Cambridge University Press. Baltin, Mark and Paul M. Postal. 1996. More on reanalysis hypotheses. Linguistic Inquiry 27 (1): 127–145. Barbera, Manuel. 2010. 2. I costrutti casuali. In Grammatica dell’Italiano Antico. Vol. II, Giampaolo Salvi and Lorenzo Renzi (eds.), 973–1014. Bologna: Il Mulino. Barra Jover, Mario. 2002. Propiedades Léxicas y Evolución Sintáctica. El Desarrollo de los Mecanismos de Subordinación en Español. La Coruña: Toxosoutos. Barreto, Therezinha Maria Mello. 2002. Observações sobre as conjunções no século XVI. In O Português Quinhentista. Estudos Lingüísticos, Rosa Virgínia Mattos e Silva and Américo Venãncio Lopes Machado Filho (eds.), 161–193. Salvador: Editora da Universidade Federal de Bahia/Universidade Estadual de Feira de Santana (EDUFBA/UEFS). Bastardas Parera, Juan. 1953. Particularidades Sintácticas del Latín Medieval (Cartularios Españoles de los Siglos VIII al XI). Barcelona: Escuela de Filología. Bassols de Climent, Mariano. 1992. Sintaxis Latina. 10th ed. Madrid: Consejo Superior de Investigaciones Científicas. Batllori, Montse, Carlos Sánchez, and Avelina Suñer. 2001. La evolución del sistema de nexos adverbiales del español: un proceso cíclico entre pragmática y sintaxis (el caso de que). In Lengua, Discurso, Texto. I Simposio Internacional de Análisis del Discurso. Vol. 1, José Jesús de Bustos Tovar (ed.), 227–240. Madrid: Visor. Batllori, Montse, Maria-Lluïsa Hernanz, Carme Picallo, and Francesc Roca. 2005. Grammaticalization and Parametric Variation. Oxford: Oxford University Press. Bat-Zeev Shyldkrot, Hava. 2008. Complétives introduites par Prep que P vs Complétives introduites par Prep ce que P, Langue française 157: 106–122. Bayer, Josef, Markus Bader and Michael Meng. 2001. Morphological underspecification meets oblique case: Syntactic and processing effects in German. Lingua 111 (4–7): 465–514. Beardsley, Wilfred A. 1921. Infinitive Constructions in Old Spanish. New York: Columbia University Press. Bechara, Evanildo. 1999. Moderna Gramática Portuguesa. Rio de Janeiro: Lucerna. Benincà, Paola and Guglielmo Cinque. 2010. La frase relativa. In Grammatica dell’Italiano Antico. Vol. I, Giampaolo Salvi and Lorenzo Renzi (eds.), 469–507. Bologna: Il Mulino. Bentivoglio, Paola. 1976. Queísmo y dequeísmo en el habla culta de Caracas. In 1975 Linguistics Colloquium on Hispanic Linguistic, Frances M. Aid, Melvyn C C Resnick and Bohdan Saciuk (eds.), 1–18. Washington, D.C.: Georgetown University Press. Bertin, Annie. 1997. L’Expression de la Cause en Ancien Français. Genève : Droz. Bertuccelli Papi, Marcella. 1991. 2.5. Le frasi finali. In Grande Grammatica Italiana di Consultazione. I. La Frase. I Sintagmi Nominale e Preposizionale, Lorenzo Renzi and Giampaolo Salvi (eds.), 818–825. Bologna: Il Mulino.

294

References

Beths, Frank. 1999. The history of dare and the status of unidirectionality. Linguistics 37 (6): 1069–1110. Blake, Barry J. 2001. Case. Cambridge: Cambridge University Press. Blanche-Benveniste, Claire. 2001. Prépositions à éclipses. Travaux de linguistique 42–43: 83–95. Bobaljik, Jonathan David and Susi Wurmbrand. 2008. Case in GB/Minimalism. In The Oxford Handbook of Case, Andrej Malchukov and Andrew Spencer (eds.), 44–58. Oxford: Oxford University Press. Bogard, Sergio and Concepción Company. 1989. Estructura y evolución de las oraciones completivas de sustantivo en el español. Romance Philology 43: 258–273. Borsley, Robert D. 2001. What do ‘prepositional complementizers’ do? Probus 13 (2): 155–171. Borsley, Robert D. and Jaklin Kornfilt. 2000. Mixed extended projections. In The Nature and Function of Syntactic Categories, Robert D. Borsley (ed.), 101–131. New York/San Diego: Academic Press. Bošković, Željko. 1995. Case properties of clauses and the Greed Principle. Studia Linguistica 49 (1): 32–53. Bosque, Ignacio (ed.). 1990. Indicativo y Subjuntivo. Madrid: Taurus. Bosque, Ignacio (ed.). 1998. Las Categorías Gramaticales. Madrid: Síntesis. Bosque, Ignacio and Javier Gutiérrez-Rexach. 2009. Fundamentos de Sintaxis Formal. Madrid: Akal. Botwinik-Rotem, Irena. 2004. The category P. Features, projections, interpretation. Ph.D. diss., Tel-Aviv University. Bouchard, Denis and Paul Hirschbuhler. 1987. French quoi and its clitic allomorph que. In Studies in Romance Languages, Carol Neidle and Rafael Núñez Cedeño (eds.), 39–60. Dordrecht: Foris. Brandão, Cláudio. 1963. Syntaxe Clássica Portuguêsa. Belo Horizonte: Universidade de Minas Gerais. Brandtler, Johan. 2010. The evaluability hypothesis. The syntax and semantics of polarity item licensing in Swedish. Ph.D. diss., Lund University. Bresnan, Joan W. 1970. On complementizers: Toward a syntactic theory of complement types. Foundations of Language 6: 297–321. Bresnan, Joan W. 1997. Mixed Categories as Head Sharing Constructions. In Proceedings of the LFG97 Conference, Miriam Butt and Tracy Holloway King (eds.), 1–17, Stanford: Center for the Study of Language and Information (CLSI). cslipublications.stanford.edu/LFG/2/ bresnan-lfg97.pdf Brito, Ana Maria. 2003. Subordinação adverbial. In Gramática da Língua Portuguesa. 7th ed. Maria Helena Mira Mateus, Ana Maria Brito, Inês Duarte, Isabel Hub Faria, Sónia Frota, Gabriela Matos, Fátima Oliveira, Marina Vigário, and Alina Villalva (eds.), 695–728. Lisboa: Caminho. Brito, Ana Maria. 2010. Tense in two types of nominalization of the infinitive in Portuguese. Abstract for Workshop on Tense and Aspect 2010 FCSH-UNL/CLUNL. http://www.fcsh.unl. pt/linguistica/wta2010/abstracts/Tense_in_two_types_of_nominalization_of_the_ infinitive_in_Portuguese.pdf Burchert, Frank. 1993. On the French complementizers que, dont and de. Probus 5 (3): 179–191. Burchert, Frank. 1994. Embedded French complement clauses and CP-recursion. Linguistics in Postdam 1: 13–23. Buridant, Claude. 2000. Grammaire Nouvelle de l’Ancien Français. Paris: Sedes.

References

295

Buridant, Claude. 2008. La Substantivation de l’Infinitif en Français : Etude Historique, Paris: Champion. Butt, Miriam. 2006. Theories of Case. Cambridge: Cambridge University Press. Campbell, Lyle. 2001. What’s wrong with grammaticalization? Language Sciences 23 (2–3): 113–161. Campos, Héctor and Paula Kempchinsky. 1991. Case absorption, theta structure and pronominal verbs. In New Analysis in Romance Linguistics, Douglas Kibee and Dieter Wanner (eds.), 171–185. Amsterdam/Philadelphia: Benjamins. Canac Marquis, Réjean. 1996. The distribution of à and de in tough constructions in French. In Grammatical Theory and Romance Languages, Karen Zagona (ed.), 35–46. Amsterdam/ Philadelphia: Benjamins. Cano Aguilar, Rafael. 1977–78. Cambios en la construcción de los verbos en castellano medieval. Archivum 27–28: 335–379. Cano Aguilar, Rafael. 1981. Estructuras Sintácticas Transitivas en el Español Actual. Madrid: Gredos. Cano Aguilar, Rafael. 1984. Cambios de construcción verbal en español clásico. Boletín de la RAE 64: 203–255. Cano Aguilar, Rafael. 1985. Sobre el régimen de las oraciones completivas en español clásico. In Philologica Hispaniensa in Honorem Manuel Alvar. II. Lingüística, Julio FernándezSevilla, Humberto López Morales, José Andrés de Molina, Antonio Quilis, and Gregorio Salvador (eds.), 81–93.Madrid: Gredos. Carnie, Andrew. 2002. Syntax. A Generative Introduction. Malden: Blackwell. Carrasco Gutiérrez, Ángeles. 1998. La correlación de tiempos en español. Ph.D. diss., Universidad Complutense de Madrid. Carrasco Gutiérrez, Ángeles. 2000. La Concordancia de Tiempos. Madrid: Arco/Libros. Castelli, Margherita. 1988. La nominalizzazione. In Grande Grammatica Italiana di Consultazione. I. La Frase. I Sintagmi Nominale e Preposizionale, Lorenzo Renzi and Giampaolo Salvi (eds.), 333–355. Bologna: Il Mulino. Castilho, Célia Maria Moraes de. 2004. Primeras histórias sobre a diacronia do dequeísmo: o clítico locativo en e o dequeísmo das orações relativas no PM. VI Seminário do Projeto para a História do Português Brasileiro. http://www.mundoalfal.org/de-que-ismo.htm. Also available in Para a História do Português Brasileiro, T. Lobo, I. Ribeiro, Z. Carneiro, and N. Almeida (eds.), 183–222. Salvador: Editora da Universidade Federal da Bahia, 2006. Cheng, Lisa Lai Shen. 1991. On the typology of wh-questions, Ph.D. diss., Massachusetts Institute of Technology. Chomsky, Noam. 1970. Remarks on nominalization. In Readings in English Transformational Grammar, Roderick A. Jacobs and Peter S. Rosenbaum (eds.), 184–221. Georgetown University Press: Washington, D.C. Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris. Chomsky, Noam. 1995. The Minimalist Program. Cambridge: The MIT Press. Chomsky, Noam. 2001. Derivation by phase. In Ken Hale: A Life in Language, Michael Kenstowicz (ed.), 1–52. Cambridge: The MIT Press. Cinque, Guglielmo. 1990. Types of Ā-dependencies. Cambridge: The MIT-Press. Cinque, Guglielmo. 2006. Restructuring and Functional Heads. The Cartography of Syntactic Structures. Volume 4, Oxford: Oxford University Press.

296

References

Cinque, Guglielmo. 2010. Mapping spatial PPs: An introduction. In Mapping Spatial PPs. The Cartography of Syntactic Structures, Volume 6, Guglielmo Cinque and Luigi Rizzi (eds.), 3–25. Oxford: Oxford University Press. Contreras, Heles. 1985. Clausal case-marking and the CRP. In Selected papers from the XIIIth Linguistic Symposium on Romance, Larry D. King and Catherine A. Maley (eds.), 19–30. Amsterdam/Philadelphia: Benjamins. Corominas, Joan. 1998. Breve Diccionario Etimológico de la Lengua Castellana, 3rd ed., Madrid: Gredos. Crisma, Paola and Giuseppe Longobardi. 2009. Change relatedness and inertia in historical syntax, In Historical Syntax and Linguistic Theory, Paola Crisma and Giuseppe Longobardi (eds.), 1–13. Oxford: Oxford University Press. Culicover, Peter W. 1997. Principles and Parameters. An Introduction to Syntactic Theory. Oxford: Oxford University Press. Culicover, Peter W. 1999. Syntactic Nuts. Hard Cases, Syntactic Theory, and Language Acquisition. Oxford: Oxford University Press. Culicover, Peter W. and Ray Jackendoff. 2005. Simpler Syntax. Oxford: Oxford University Press. Curme, George Oliver. 1935. A Grammar of the English Language. Vol. 2: Parts of Speech and Accidence, Boston: D. C. Heath & Comp. Cyrino, Sonia, Jairo Nunes, and Emilio Pagotto. 2009. Complementação. In Gramática do Português Culto Falado no Brasil. Volume 3. A Construção da Sentença, Ataliba T. de Castilho (general coord.), Mary Aizawa Kato and Milton do Nascimento (orgs.), 43–96, Campinas: Editora da Unicamp. Darmesteter, Arsène. 1930. Cours de Grammaire Historique de la Langue Française. Deuxième Partie: Morphologie. Paris: Librairie Delagrave. Darmesteter, Arsène. 1931. Cours de Grammaire Historique de la Langue Française. Quatrième Partie: Syntaxe. Paris: Librairie Delagrave. Dardano, Maurizio. 1995. Note sulla prosa antica. In La Sintassi dell’Italiano Letterario, Maurizio Dardano and Pietro Trifone (eds.), 15–50. Roma: Bulzoni. Darlymple, Mary and Helge Lødrup. 2000. The grammatical functions of complement clauses. In Proceedings of the LFG00 Conference, Miriam Butt and Tracy Holloway King (eds.), 1–18. Stanford: Center for the Study of Language and Information (CLSI). csli-publications. stanford.edu/LFG/5/lfg00dalrympl-lodrup.pdf Davies, William D. and Stanley Dubinsky. 1998. Sentential subjects as complex NPs: New reasons for an old account of subjacency. In CLS 34. Part 1: Papers from the Main Session, M. Catherine Gruber, Derrick Higgins, Kenneth S. Olson, and Tamra Wysocki (eds.), 83–94. Chicago: Chicago Linguistic Society. Delicado Cantero, Manuel. 2008. Innovation in relativization: the new Spanish possessive relative el cual su N. Kansas Working Papers in Linguistics 30: 51–65. Delicado Cantero, Manuel. 2012. Relative and possessive features in colloquial Spanish. In Current Formal Aspects of Spanish Syntax and Semantics, Melvin González-Rivera and Sandro Sessarego (eds.), 78–94, Newcastle upon Tyne: Cambridge Scholars Publishing. Delicado Cantero, Manuel and Melvin González Rivera. 2011. Restrictions on nominalization in Spanish finite clauses. Paper presented at the 2011 Australian Linguistics Society Annual Conference, ANU, Canberra. Demonte, Violeta. 1977. La Subordinación Sustantiva. Madrid: Cátedra.

References

297

Demonte, Violeta. 1992. Linking and case: The case of prepositional verbs. In Theoretical Analyses in Romance Linguistics, Christiane Laefer and Terrell A. Morgan (eds.), 415–456. Amsterdam/Philadelphia: Benjamins. Demonte, Violeta and Olga Fernández Soriano. 2005. Features in comp and syntactic variation: the case of ‘(de)queísmo’ in Spanish. Lingua 115 (8): 1063–1082. Dias, Augusto Epiphanio da Silva. 1959. Syntaxe Historica Portuguesa. Lisboa: Livraria Clássica Editora. Dixon, Robert M. W. 2006. Complement clauses and complementation strategies in typological perspective. In Complementation: a Cross-linguistic Typology. Robert M. W. Dixon and Alexandra Y. Aikhenvald (eds.), 1–48. Oxford: Oxford University Press. Dobrovie-Sorin, Carmen, Tonia Bleam, and M. Teresa Espinal. 2006. Bare nouns, number and types of incorporation. In Non-definiteness and Plurality, Svetlana Vogeleer and Liliane Tasmowski (eds.), 51–79. Amsterdam/Philadelphia: Benjamins. Dowty, David. 2003. The dual analysis of adjuncts/complements in Categorial Grammar. In Modifying Adjuncts, Ewald Lang, Claudia Maienborn, and Cathrine Fabricius-Hansen (eds.), 33–66. New York: Mouton de Gruyter. Duarte, Inês. 2003. Subordinação completiva – as orações completivas. In Gramática da Língua Portuguesa. 7th ed. Maria Helena Mira Mateus, Ana Maria Brito, Inês Duarte, Isabel Hub Faria, Sónia Frota, Gabriela Matos, Fátima Oliveira, Marina Vigário, and Alina Villalva (eds.), 593–651. Lisboa: Caminho. Dubinsky, Stanley and Kemp Williams. 1995. Recategorization of prepositions as complementizers: The case of temporal prepositions in English. Linguistic Inquiry 26 (1): 125–137. Egerland, Verner. 2010. 1, 2.1–2.3, 2.5, 3–6. Frasi subordinate all’infinito. In Grammatica dell’Italiano Antico. Vol. II, Giampaolo Salvi and Lorenzo Renzi (eds.), 817–35, 856–879. Bologna: Il Mulino. Eguren, Luis and Olga Fernández Soriano. 2004. Introducción a una Sintaxis Minimista. Madrid: Gredos. Elia, Annibale, Maurizio Martinelli, and Emilio D’Agostino. 1981. Lessico e Strutture Sintattiche: Introduzione alla Sintassi del Verbo Italiano. Napoli: Liguori. Emonds, Joseph E. 1972. A reformulation of certain syntactic transformations. In Goals of Linguistic Theory, S. Peters (ed.), 21–62. Englewood Cliffs: Prentice-Hall. Emonds, Joseph E. 1976. A Transformational Approach to English Syntax. New York: Academic Press. Emonds, Joseph E. 1985. A Unified Theory of Syntactic Categories. Dordrecht: Foris. Ernout, Alfred and François Thomas. 1964. Syntaxe Latine. 2d ed. Paris: Klincksieck. Ernst, Thomas. 1996. On adjunct Case in Chinese. Cahiers de linguistique – Asie Orientale 25 (2): 167–197. Falk, Julia S. 1968. Nominalizations in Spanish. Ph.D. diss., University of Washington. Fava, Elisabetta. 1991. 1.3. Le frasi interrogative indirette. In Grande Grammatica Italiana di Consultazione. II. I Sintagmi Verbale, Aggetivale, Avverbiale. La Subordinazione, Lorenzo Renzi and Giampaolo Salvi (eds.), 675–720. Bologna: Il Mulino. Fernández Lagunilla, Marina. 1987. Los infinitivos con sujetos léxicos en español. In Sintaxis de las lenguas románicas, Violeta Demonte and Marina Fernández Lagunilla (eds.), 125–147. Madrid: Ediciones El Arquero. Fernández Lagunilla, Marina and Alberto Anula Rebollo. 1995. Sintaxis y Cognición. Introducción al Conocimiento, el Procesamiento y los Déficits Sintácticos. Madrid: Síntesis.

298

References

Fernández Lagunilla, Marina and Francisco Javier de Dios López. 1991. Dos análisis gramaticales de ciertas construcciones completivas de infinitivo en español: a propósito de los verbos causativos y de percepción. Revista Española de Lingüística 21 (2): 217–232. Fischer, Olga. 2007. Morphosyntactic Change. Functional and Formal Perspectives. Oxford: Oxford University Press. Fiumara, Francesco. 2006. La sintassi del Convivio. Tesi di laurea specialistisca. Università di Pisa. Frenguelli, Gianluca. 2002. Tre Studi di Sintassi Antica e Rinascimentale. Roma: Aracne. Gaatone, David. 1981. Conjonctions et locutions conjonctives en français. Folia Linguistica 14: 195–211. Gaatone, David. 2004. Les prépositions forment-elles une classe? In Lexique, Syntaxe et LexiqueGrammaire. Syntax, Lexis & Lexicon-Grammar. Papers in Honour of Maurice Gross, Christian Leclère, Eric Laporte, Mireille Piot, and Max Silberztein (eds.), 211–221. Amsterdam/ Philadelphia: Benjamins. van Gelderen, Elly. 2004. Grammaticalization as Economy. Amsterdam/Philadelphia: Benjamins. van Gelderen, Elly. 2009. Renewal in the left periphery: economy and the complementiser layer. Transactions of the Philological Society 107 (2): 131–195. Girón Alconchel, José Luis. 1988. Las Oraciones Interrogativas Indirectas en Español Medieval. Madrid: Gredos. Giusti, Giuliana. 1991. 2. Frasi avverbiali: Temporali, causali e consecutive. In Grande Grammatica Italiana di Consultazione. II. I Sintagmi Verbale, Aggetivale, Avverbiale. La Subordinazione, Lorenzo Renzi and Giampaolo Salvi (eds.), 720–738. Bologna: Il Mulino. Glauning, Friedrich. 1873. Syntaktische Studien zu Marot, ein Beitrag zur Geschichte der Französischen Syntax. Nördlingen: C. H. Beck’sche Buchhandlung. Gómez Torrego, Leonardo. 1999. La variación en las subordinadas sustantivas: dequeísmo y queísmo. In Gramática Descriptiva de la Lengua Española, Ignacio Bosque and Violeta Demonte (eds.), 2105–2148. Madrid: Espasa-Real Academia Española. González Ollé, Fernando. 1993. Lengua y Literatura Españolas Medievales. Textos y Glosario. Madrid: Arco/Libros. Granath, Solveig. 1997. Verb Complementation in English. Omission of Prepositions before THAT-clauses and TO-infinitives. Göteborg: Acta Universitatis Gothoburgensis. Granath, Solveig and Aimo Seppänen. 2004. Prepositions and that-clauses in English. Proceedings from NAES 2004, Ninth Nordic Conference for English Studies. Granath, Solveig and Michael P. Wherrity. 2005. Prepositions with that-clause complements in tagged corpora, with a special focus on in that. Proceedings from The Corpus Linguistics Conference Series. Grammar 1 (1): 1–9. Granvik, Anton. 2012. De de. Estudio histórico-comparativo de los usos y la semántica de la preposición de en español, Ph.D. diss., University of Helsinki. Grevisse, Maurice. 1980. Le Bon Usage. Grammaire Française avec des Remarques sur la Langue Française d’aujourd’hui. Paris/Gembloux: Duculot. Grimshaw, Jane. 1990. Argument Structure. Cambridge: The MIT Press. Grimshaw, Jane. 2000. Locality and extended projection. In Lexical Specification and Insertion, Peter Coopmans, Martin B.H. Everaert, and Jane Grimshaw (eds.), 115–133. Amsterdam/ Philadelphia: Benjamins.

References

299

Gutiérrez-Rexach, Javier and Luis Silva-Villar. 2007. Predicative complementation: A new approach to dequeísmo. Ms Ohio State University/Mesa State College (Colorado Mesa University). Haase, Alfons. 1965. Syntaxe Française du XVII e Siècle. 5th ed. Paris: Delagrave. Haegeman, Liliane. 1991. Introduction to Government and Binding Theory. Oxford: Blackwell. Hale, Mark. 1998. Diachronic syntax. Syntax 1: 1–18. Hale, Mark. 2007. Historical Linguistics. Theory and Method. Malden: Blackwell. Han, Hye Jin. 2005. A nominal-shell for CPs: Beyond subject CPs. In Proceedings from the Annual Meeting of the Chicago Linguistic Society 41. 1. The Main Session, Rodney L. Edwards, Patrick J. Midtlyng, Colin L. Sprague, and Kjersti G. Stensrud (eds.), 95–109. Chicago: Chicago Linguistic Society. Harley, Heidi. 1995. Abstracting away from abstract case. In Proceedings of NELS 25, part I, J. Beckman (ed.), 207–221. Amherst: Graduate Linguistics Students’ Association, University of Massachusetts. Harris, Alice C. and Lyle Campbell. 1995. Historical Syntax in Cross-linguistic Perspective. Cambridge: Cambridge University Press. Hawkins, Roger and Richard Towell. 2001. French Grammar and Usage. New York: McGraw-Hill. Herman, József. 1963. La Formation du Système Roman des Conjonctions de Subordination. Berlin: Akademie-Verlag. Hernanz, M. Lluïsa. 1999. El infinitivo. In Gramática Descriptiva de la Lengua Española, Ignacio Bosque and Violeta Demonte (eds.), 2197–2356. Madrid: Espasa-Real Academia Española. Herrero Ruiz de Loizaga, F. Javier. 2005. Sintaxis Histórica de la Oración Compuesta en Español. Madrid: Gredos. Hock, Hans Henrich. 1991. Principles of Historical Linguistics. 2d ed. Berlin/New York: Mouton de Gruyter. Hoekstra, Teun. 1984. Government and the distribution of sentential complements in Dutch. In Sentential Complementation. Proceedings of the International Conference held at UFSAL, Brussels. June, 1983, Wim de Geest and Yvan Putseys (Eds.), 105–116. Dordrecht: Foris. Hoffmann, Sebastian. 2005. Grammaticalization and English Complex Prepositions. A Corpusbased Study. New York: Routledge. Holmberg, Anders and Christer Platzack. 1995. The Role of Inflection in Scandinavian Syntax. New York: Oxford. University Press. Holmes, Philip and Ian Hinchliffe. 2003. Swedish: a Comprehensive Grammar. London/New York: Routledge. Horno Chéliz, Ma Carmen. 2002. Lo que la Preposición Esconde. Estudio sobre la Argumentalidad Preposicional. Zaragoza: Prensas Universitarias de Zaragoza. Hornstein, Norbert, Jairo Nunes, and Kleanthes K. Grohmann. 2005. Understanding Minimalism. Cambridge: Cambridge University Press. Horrocks, Geoffrey. 2011. Latin syntax. In A Companion to the Latin Language, James Clackson (ed.), 1–22. Oxford: Wiley-Blackwell. Blackwell Reference Online. Huber, Joseph. 1933. Gramática do Português Antigo. Lisboa: Fundação Calouste Gulbenkian. Huddleston, Rodney. 1984. Introduction to the Grammar of English. Cambridge: Cambridge University Press. Huddleston, Rodney and Geoffrey K. Pullum. 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Huot, Hélène. 1981. Constructions Infinitives du Français. Le Subordonnant de. Genève: Droz.

300

References

Iglesias Recuero, Silvia. 2002. Sobre cómo enunciativo en la lengua medieval y clásica. In Actas del V Congreso Internacional de Historia de la Lengua Española, María Teresa Echenique Elizondo and Juan P. Sánchez Méndez (eds.), 675–690. Madrid: Gredos. Janda, Richard D. 2001. Beyond “pathways” and “unidirectionality”: on the discontinuity of language transmission and the counterability of grammaticalization. Language Sciences 23 (2–3): 265–340. Jensen, Frede. 1974. The Syntax of the Old French Subjunctive. The Hague/Paris: Mouton. Jensen, Frede. 1990. Old French and Comparative Gallo-Romance Syntax. Tübingen: Niemeyer. Jones, Michael Allan. 1996. Foundations of French Syntax. Cambridge: Cambridge University Press. Joseph, Brian D. 1998. Diachronic Morphology. In The Handbook of Morphology, Andrew Spencer and Arnold M. Zwicky (eds.), 351–373. Malden: Blackwell. Joseph, Brian D. 2001. Is there such thing as “grammaticalization”? Language Sciences 23 (2–3): 163–186. Joseph, Brian D. and Richard D. Janda. 2003. On language, change, and language change – or, of history, linguistics, and historical linguistics. In The Handbook of Historical Linguistics, Brian D. Joseph and Richard D. Janda (eds.), 3–180. Malden: Blackwell. Kany, Charles E. 1994. Sintaxis Hispanoamericana. Madrid: Gredos. Kayne, Richard. 1975. French Syntax. Cambridge: The MIT Press. Kayne, Richard. 1981. On certain differences between French and English. Linguistic Inquiry 12 (3): 349–371. Kayne, Richard. 1999. Prepositional complementizers as attractors. Probus 11: 39–73. Kayne, Richard. 2000. Parameters and Universals. Oxford: Oxford University Press. Kayne, Richard. 2008. Why isn’t this a complementizer?, Ms NYU. To appear in Functional Structure from Top to Toe: A Festschrift for Tarald Taraldsen, Peter Svenonius (ed.), New York: Oxford University Press. http://as.nyu.edu/docs/IO/2652/Kayne0608This.pdf Kempchinsky, Paula. 1988. On inherent Case-marking. In Proceedings of the Seventh West Coast Conference on Formal Linguistics. Hagit Borer (ed.), 203–216. Stanford: Stanford Linguistics Association. Kempchinsky, Paula. 1992. Clausal complements and Case Theory in Romance. Probus 4: 17–51. Keniston, Hayward. 1937. The Syntax of Castilian Prose. The Sixteenth Century. Chicago: The University of Chicago Press. Keyser, Samuel Jay. 1975. A partial history of the relative clause in English. Papers in the history and structure of English. University of Massachusetts Occasional Papers in Linguistics 1: 1–33. Kiparsky, Paul. 2000. Analogy as optimization: ‘exceptions’ to Sievers’ Law in Gothic. In Analogy, Levelling, Markedness. Principles of Change in Phonology and Morphology. Aditi Lahiri (ed.), 15–46. Berlin/New York: Mouton de Gruyter. Kiparsky, Paul and Carol Kiparsky. 1971. Reprint. Fact. In Semantics: An Interdisciplinary Reader in Philosophy, Linguistics and Psychology, Danny D. Steinberg and Leon A. Jakobovits, L. A. (eds.), 345–369. Cambridge: Cambridge University Press. Original edition, in Progress in Linguistics, The Hague: Mouton, 1970. Koptjevskaja-Tamm, Maria. 1993. Nominalizations. London/New York: Routledge. Kornfilt, Jaklin and John Whitman. 2011. Afterword: Nominalizations in syntactic theory. Lingua 121 (7): 1297–1313. Kortmann, Bernd. 1998. Adverbial subordinators in the languages of Europe. In Adverbial Constructions in the Languages of Europe, Johan van der Auwera in collaboration with Dónall P. Ó. Baoill (eds.), 457–561. Berlin/New York: Mouton de Gruyter.

References

301

Kortmann, Bernd and Ekkehard König. 1992. Categorial reanalysis: the case of deverbal prepositions. Linguistics 30: 671–697. Koster, Jan. 1978. Why subject sentences don’t exist. In Recent Transformational Studies in European Languages, Samuel J. Keyser (ed.), 53–64. Cambridge: The MIT Press. Kroch, Anthony. 1989. Reflexes of grammar in patterns of language change. Language Variation and Change 1: 199–244. Kroch, Anthony. 2001. Syntactic change. In TheHandbook of Contemporary Syntactic Theory, Mark Baltin and Chris Collins (eds.), 699–729. Malden: Blackwell. Kuno, Susumu. 1973. Constraints on internal clauses and sentential subjects. Linguistic Inquiry 4 (3): 363–385. Kunstmann, Pierre. 1990. Le Relatif-interrogatif en Ancien Français. Genève: Droz. Kurzon, Dennis and Silvia Adler (eds.). 2008. Adpositions. Pragmatic, Semantic and Syntactic Perspectives. Amsterdam/Philadelphia: Benjamins. Labov, William. 1973. The social setting of linguistic change. In Current Trends in Lingusitics. Volume II. Diachronic, Areal, and Typological Linguistics, Thomas A. Sebeok (ed.), 213–251. The Hague/Paris: Mouton. Lakoff, Robin T. 1968. Abstract Syntax and Latin Complementation. Cambridge: The MIT Press. Lamontagne, Greg and Lisa Travis. 1987. The syntax of adjacency. WCCFL 6: The Proceedings of the Sixth West Coast Conference on Formal Linguistics: 173–186. Lapesa, Rafael. 2000. Estudios de Morfosintaxis Histórica del Español. Madrid: Gredos. Larson, Richard. 1985. Bare NP-Adverbs. Linguistic Inquiry 16: 595–621. Larson, Richard. 1990. Extraction and multiple selection in PP. The Linguistic Review 7: 169–182. Lasnik, Howard. 2008. On the development of Case Theory: Triumphs and challenges. In Foundational Issues in Linguistic Theory. Essays in Honor of Jean-Roger Vergnaud, Robert Freidin, Carlos P. Otero, and Maria Luisa Zubizarreta (eds.), 17–41. Cambridge: The MIT Press. Lasnik, Howard, and Juan Uriagereka, with Cedric Boeckx. 2005. A Course in Minimalist Syntax. Foundations and Prospects. Malden: Blackwell. Le Goffic, Pierre. 1993. Grammaire de la Phrase Française. Paris: Hachette Supérieur. Ledgeway, Adam. 2000. A Comparative Syntax of the Dialects of Southern Italy: a Minimalist Approach. Oxford: The Philological Society-Blackwell. Lees, Robert B. 1960. The Grammar of English Nominalizations. The Hague: Mouton. Lefebvre, Claire and Pieter Muysken. 1988. Mixed Categories: Nominalizations in Quechua. Dordrecht: Kluwer Academic Publishers. Legate, Julie Anne. 2008. Morphological and abstract Case. Linguistic Inquiry 39 (1): 55–101. Lemle, Miriam. 1984. Análise Sintática (Teoria Geral e Descrição do Portugués). São Paulo: Editora Ática. Leonard, William Patrick. 1997. The syntax of weak prepositions: Locality constraints and lexical dependencies in English and Romance. Ph.D. diss., University of California-Los Angeles. Leonetti, Manuel. 1999a. El artículo. In Gramática Descriptiva de la Lengua Española, Ignacio Bosque and Violeta Demonte (eds.), 787–890. Madrid: Espasa-Real Academia Española. Leonetti, Manuel. 1999b. La subordinación sustantiva: las subordinadas enunciativas en los complementos nominales. In Gramática Descriptiva de la Lengua Española, Ignacio Bosque and Violeta Demonte (eds.), 2083–2104. Madrid: Espasa-Real Academia Española. Lightfoot, David. 2006. How New Languages Emerge. Cambridge: Cambridge University Press. Littlefield, Heather A. 2006. Syntax and acquisition in the prepositional domain: Evidence from English for fine-grained syntactic categories. Ph.D. diss., Boston University.

302

References

Lødrup, Av Helge. 1991. Clausal complements in English and Norwegian. Norsk LingvistiskTidsskrift 9: 105–136. Lødrup, Av Helge. 2004. Clausal complementation in Norwegian. Nordic Journal of Linguistics 27 (1): 61–95. López, María Luisa. 1970. Problemas y Métodos en el Análisis de Preposiciones. Madrid: Gredos. Madeira, Ana Maria. 1994. On the Portuguese inflected infinitive. UCL Working Papers in Linguistics 6: 179–203. Maiden, Martin and Cecilia Robustelli. 2000. A Reference Grammar of Modern Italian. New York: McGraw-Hill. Malchukov, Andrej L. 2004. Nominalization/Verbalization: Constraining a Typology of Transcategorial Operations. Munich: LINCOM Europa. Malouf, Robert. 2000. Verbal gerunds as mixed categories in Head-driven Phrase Structure Grammar. In Syntax and Semantics. Volume 32. The Nature and Function of Syntactic Categories, Robert D. Borsley (ed.), 133–166. New York: Academic Press. Manzini, M. Rita. 2010. The structure and interpretation of (Romance) complementizers. In The Complementizer Phase, E. Phoevos Panagiotidis (ed.), 167–199. Oxford: Oxford University Press. Manzini, M. Rita and Leonardo M. Savoia. 2005. I Dialetti Italiani e Romanci. Morfosintassi Generativa. Alessandria: Edizioni dell’Orso. Manzini, M. Rita and Leonardo M. Savoia. 2011. Grammatical Categories. Variation in Romance Languages. Cambridge: Cambridge University Press. Marantz, Alec. 2000. Case and licensing. In Arguments and Case, Eric Reuland (ed.), 11–30. Amsterdam/Philadelphia: Benjamins. Marchello-Nizia, Christiane. 1997. La Langue Française aux XIV e et XV e Siècles. Paris: Nathan. Martineau, France. 1990. La montée du clitique en moyen français: une étude de la syntaxe des. constructions infinitives, Ph.D. dissertation, University of Ottawa. Martineau, France. 2000. Les compléments infinitifs de l’ancien français au français moderne: concurrence entre les prépositions a, de et l’absence de préposition. In Le Moyen Français: le Traitement du Texte, Edition, Apparat critique, Glossaire, Traitement électronique. Actes du IX e Colloque International sur le Moyen Français, Claude Buridant (ed.), 165–183. Strasbourg: Presses Universitaires de Strasbourg. Martineau, France. 2002. Grammaticalisation et changement de paramètre: la préposition à, élément introducteur d’un complément infinitif. Scolia 15: 71–85. Martineau, France and Virginia Motapanyane. 2000. Infinitive subordinators and verb restructuring in French. In New Approaches to Old Problems, Steven N. Dworkin and Dieter Wanner (eds.), 217–231. Amsterdam/Philadelphia: Benjamins. Martins, Ana Maria. 2001. On the origin of the Portuguese inflected infinitive: A new perspective on an enduring debate. In Historical Linguistics 1999: Selected Papers from the 14th Conference on Historical Linguistics, Laurel J. Brinton (ed.), 207–222. Amsterdam/ Philadelphia: Benjamins. Martins, Ana Maria. 2006. Aspects of infinitival constructions in the history of Portuguese. In Historical Romance Linguistics: Retrospective and Perspectives, Randall S. Gess and Deborah Arteaga (eds.), 327–355. Amsterdam/Philadelphia: Benjamins. Matthews, Peter. 2003. On change in ‘E-language’. In Motives for Language Change, Raymond Hickey (ed.), 7–17. Cambridge: Cambridge University Press. Mattos e Silva, Rosa Virgínia. 1989. Estruturas Trecenstistas. Elementos para uma Gramática do Português Arcaico. Lisboa: Impresa Nacional-Casa da moeda.

References

303

Mattos e Silva, Rosa Virgínia. 1994. O Português Arcaico. Morfologia e Sintaxe. São Paulo: Contexto. McClean, R. J. 1963. Teach Yourself Swedish. A Grammar of the Modern Language. London: The English Universities Press. McFadden, Thomas. 2004. The position of morphological case in the derivation: A study of the syntax-morphology interface. Ph.D. diss., University of Pennsylvania. Medici, Mario. 1978. Congiunzioni. In Enciclopedia dantesca. Roma: Istituto della enciclopedia italiana fondata da Giovanni Treccani. 334–344. Ménard, Philippe. 1988. Syntaxe de l’Ancien Français. 4th ed. Paris: Editions Bière. Méndez García de Paredes, Elena. 1995. Las Oraciones Temporales en Castellano Medieval. Sevilla: Universidad de Sevilla. Mensching, Guido. 2000. Infinitive Constructions with Specified Subjects. A Syntactic Analysis of the Romance Languages. New York: Oxford University Press. Meyer-Lübke, Wilhelm. 1900. Grammaire des Langues Romanes. Paris: H. Welter. de Miguel, Elena. 1996. Nominal infinitives in Spanish: An aspectual constraint. Canadian Journal of Linguistics 41 (1): 29–52. Miller, D. Gary and Dieter Wanner. 2011. Theoretical trends in historical syntax. Diachronica 28 (1): 119–131. Moignet, Gérard. 1973. Grammaire de l’Ancien Français. Morphologie-Syntaxe. Paris: Éditions Klincksieck. Mollica, Maria Cecilia. 1995. (De) Que Falamos? Rio de Janeiro: Edições Tempo brasileiro Ltda. Moreno Cabrera, Juan Carlos. 1985–86. Tipología de la catáfora paratáctica: entre la sintaxis del discurso y la sintaxis de la oración. Estudios de Lingüística Universidad de Alicante (ELUA) 3: 165–192. Munaro, Nicola. 2010. La frase interrogativa. In Grammatica dell’Italiano Antico. Vol. II, Giampaolo Salvi and Lorenzo Renzi (eds.), 1147–1185. Bologna: Il Mulino. Nakamura, Takuya. 2008. Sur les interrogatives indirectes construites avec sur: leur distribution avec verbes à deux compléments. In Congrès Mondial de Linguistique Française – CMLF’08, J. Durand, B. Habert, and B. Laks (éds.), 2577–2586. http://www.linguistiquefrancaise.org http://dx.doi.org/10.1051/cmlf08232 Napoli, Donna Jo. 1993. Syntax. Theory and Problems. New York: Oxford University Press. Neeleman, Ad. 1997. PP-complements. Natural Language and Linguistic Theory 15: 89–137. Neeleman, Ad and Fred Weerman. 2001. Flexible Syntax. A Theory of Case and Arguments. Dordrecht: Kluwer Academic Publishers. Newmeyer, Frederick J. 2000. The discrete nature of syntactic categories: Against a prototypebased account. In Syntax and Semantics. Volume 32. The Nature and Function of Syntactic Categories, Robert D. Borsley (ed.), 221–250. New York: Academic Press. Panagiotidis, E. Phoevos and Kleanthes K. Grohmann. 2009. Mixed projections: Categorial switches and prolific domains. Linguistic Analysis 35 (1–4): 141–161. Pantiglioni, Massimo. 2010. 5. Le frasi finali. In Grammatica dell’Italiano Antico. Vol. II, Giampaolo Salvi and Lorenzo Renzi (eds.), 1086–1094. Bologna: Il Mulino. Pavón Lucero, Ma Victoria. 1999. Clases de partículas: preposición, conjunción y adverbio. In Gramática Descriptiva de la Lengua Española, Ignacio Bosque and Violeta Demonte (eds.), 565–655. Madrid: Espasa-Real Academia Española. Pavón Lucero, Ma Victoria. 2003. Sintaxis de las Partículas. Madrid: Visor. Pearce, Elizabeth. 1990. Parameters in Old French Syntax: Infinitival Complements. Dordrecht: Kluwer Academic Publishers.

304

References

Pérez Vázquez, Yuya. 1998–99. Status categoriale dell’infinito in italiano e in spagnolo. Quaderni del Laboratorio di Linguistica 12: 149–170. http://linguistica.sns.it/QLL/QLL98/ YPV.StatusInfItSp.pdf Picallo, M. Carme. 2001. Nominalized clauses, clausal arguments and agreement. Catalan Working Papers in Linguistics 9: 69–84. Picallo, M. Carme. 2002. Abstract agreement and clausal arguments. Syntax 5 (2): 116–147. Pinkster, Harm. 1990. Latin Syntax and Semantics. London/New York: Routledge. Pintzuk, Susan. 2003. Variationist approaches to syntactic change. In The Handbook of Historical Linguistics, Brian D. Joseph and Richard D. Janda (eds.), 509–528. Malden: Blackwell. Pintzuk, Susan, George Tsoulas, and Anthony Warner. 2000. Syntactic change: Theory and method. In Diachronic Syntax. Models and Mechanisms, Susan Pintzuk, George Tsoulas, and Anthony Warner (eds.), 1–22. Oxford: Oxford University Press. Plann, Susan. 1981. The two el + infinitive constructions in Spanish. Linguistic Analysis 7 (3): 203–240. Plann, Susan. 1986. On Case-marking clauses in Spanish: Evidence against the Case Resistance Principle. Linguistic Inquiry 17 (2): 336–345. Pollock, Jean-Yves. 1992. Opérateurs nuls, dont, questions indirectes et théorie de la quantification. In Hommages à Nicolas Ruwet: De la Musique à la Linguistique, Liliane Tasmowski and Anne Zribi-Hertz (eds.), 440–463. Ghent: Communication and Cognition. Quicoli, A. Carlos. 1996. Inflection and parametric variation: Portuguese vs. Spanish. In Current Issues in Comparative Grammar. Robert Freidin (ed.), 46–100. Dordrecht: Kluwer Academic Publishers. Quilis, Antonio. 2012. Principios de Fonología y Fonética Españolas Madrid: Arco/Libros. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985. A Comprehensive Grammar of the English Language. New York: Longman. Rabanales, Ambrosio. 1977. Queísmo y dequeísmo en el español de Chile. In Estudios sobre el Español Hablado en las Principales Ciudades de América, Juan M. Lope Blach (ed.), 541–569. México: Universidad Nacional Autónoma de México. Radford, Andrew. 1988. Transformational Grammar. A First Course. Cambridge: Cambridge University Press. Ralph, Bo. 1975. On the nature of preposition deletion in Swedish. In The Nordic Languages and Modern Linguistics 2, Karl-Hampus Dahlstedt (ed.), 666–684. Stockholm: Almqvist & Wiksell International. Ramírez, Carlos Julio. 2003. The Spanish nominalized infinitives: A proposal for a classification. Toronto Working Papers in Linguistics 21: 117–133. Raposo, Eduardo. 1987a. Case Theory and the Infl-to-Comp: The inflected infinitive in European Portuguese. Linguistic Inquiry 18 (1): 85–109. Raposo, Eduardo. 1987b. Romance infinitival clauses and Case theory. In Studies in Romance languages, Carol Neidle and Rafael A. Núñez-Cedeño (eds.), 237–249. Dordrecht: Foris. Rathert, Monika and Artemis Alexiadou. 2010. Introduction. In The Semantics of Nominalizations Across Languages and Frameworks, Monika Rathert and Artemis Alexiadou (eds.), 1–8. Berlin/New York: De Gruyter Mouton. Rauh, Gisa. 1991. Prepositional forms in the lexicon: problems and suggestions. In Approaches to Prepositions, Gisa Rauh (ed.), 169–223. Tübingen: Gunter Narr Verlag. Rauh, Gisa. 1993. On the grammar of lexical and non-lexical prepositions in English. In The Semantics of Prepositions. From Mental Processing to Natural Language Processing, Cornelia Zelinsky-Wibbelt (ed.), 99–150. Berlin/New York: Mouton de Gruyter.

References

305

Rauh, Gisa. 2002. Prepositions, features, and projections. In Perspectives on Prepositions, Heinrich Bellermann (ed.), 3–23. Tübingen: Niemeyer. Rauh, Gisa. 2010. Syntactic Categories. Their Identification and Description in Linguistic Theories. Oxford: Oxford University Press. Real Academia Española. 2010. Nueva Gramática de la Lengua Española. Manual. México D.F.: Real Academia Española, Asociación de Academias de la Lengua Española, Editorial Planeta Mexicana (Espasa). van Reenen, Pieter and Lene Schøsler. 1993. Les indices d’infinitif complément d’objet en ancien français. In Actas do XIX Congreso Internacional de Lingüística e Filoloxía Románicas. Vol V, Ramón Lorenzo Vázquez (ed.), 523–545. La Coruña: Fundación Pedro Barrié de la Maza. Renzi, Lorenzo. 2010. L’articolo. In Grammatica dell’Italiano Antico. Vol. I, Giampaolo Salvi and Lorenzo Renzi (eds.), 297–347. Bologna: Il Mulino. van Riemsdijk, Henk. 1998. Head Movement and Adjacency. Natural Language and Linguistic Theory 16: 633–678. Rizzi, Luigi. 1988. Il sintagma preposizionale. In Grande Grammatica Italiana di Consultazione. I. La Frase. I Sintagmi Nominale e Preposizionale, Lorenzo Renzi (ed.), 507–531. Bologna: Il Mulino. Rizzi, Luigi. 1992. The fine structure of the left periphery. In Elements of Grammar. Handbook in Generative Syntax, Liliane Haegeman (ed.), 281–337. Dordrecht: Kluwer Academic Publishers. Roberts, Ian. 1993. A formal account of grammaticalisation in the history of Romance futures. Folia Linguistica Historica 13 (1–2): 219–258. Roberts, Ian. 2007. Diachronic Syntax. Oxford: Oxford University Press. Roberts, Ian and Anna Roussou. 1999. A formal approach to “grammaticalization”. Linguistics 37 (6): 1011–1041. Roberts, Ian and Anna Roussou. 2003. Syntactic Change. A Minimalist Approach to Grammaticalization. Cambridge: Cambridge University Press. Rochette, Anne. 1988. Semantic and syntactic aspects of Romance sentential complementation, Ph.D. diss., Massachusetts Institute of Technology. Rodríguez Gallardo, Ángel. 2000. Complementos preposicionales argumentales en español clásico. Verba 27: 223–248. Rohlfs, Gerhard. 1969. Grammatica storica della Lingua Italiana e dei suoi Dialetti. Sintassi e Formaziones delle Parole. Torino: Einaudi. Rooryck, Johan. 1996. Prepositions and minimalist Case-marking. In Studies in Comparative Germanic Syntax. Vol. II, Höskuldur Thráinsson, Samuel Epstein, and Steve Peter (eds.), 226–256. Dordrecht: Kluwer Academic Publishers. Rosenbaum, Peter S. 1967. The Grammar of English Predicate Complement Constructions. Cambridge: The MIT Press. Ross, John R. 1973. Nouniness. In Three Dimensions of Linguistic Theory, Osamu. Fujimura (ed.), 123–257. Tokyo: The TEC Corporation. Rouquier, Magali. 1990. Le terme ‘ce que’ régissant une complétive en ancien français. Revue romane 25: 47–72. Rouquier, Magali. 2003. La séquence est-ce dans les interrogatives en qui/que en ancien et en moyen français. French Language Studies 13: 339–362. Roussou, Anna. 1991. Nominalized clauses in the syntax of Modern Greek. UCL Working Papers in Linguistics 3: 77–100. Roussou, Anna. 2010. Selecting complementizers. Lingua 120 (3): 582–603.

306

References

Rowlett, Paul. 2007. The Syntax of French. Cambridge: Cambridge University Press. Russo, Harold J. 1942. Morphology and syntax of the Leal conselheiro. Ph.D. diss., University of Pennsylvania. Said Ali, Manuel. 1966. Gramática Histórica da Língua Portuguesa. São Paulo: Edições Melhoramentos. Salvi, Giampaolo. 1985. L’infinito con l’articolo. In Sintassi e Morfologia della Lingua Italiana d’Uso. Teorie e Applicazioni Descrittive. Atti del XVII Congresso Internazionale di Studi, Annalisa Franchi de Bellis and Leonardo M. Savoia (eds.), 243–268. Roma: Bulzoni. Salvi, Giampaolo. 2010. Frasi subordinate argomentali. In Grammatica dell’Italiano Antico. Vol. II, Giampaolo Salvi and Lorenzo Renzi (eds.), 939–951. Bologna: Il Mulino. Sánchez Jiménez, Santiago U. 1999. La expresión lingüística de la finalidad en textos históricocronísticos medievales, Ph.D. diss., Universidad Complutense de Madrid. Sandfeld, Kristian. 1965. Syntaxe du français contemporain. Les propositions subordonnées. Genève : Droz. Santos, Ana Lúcia, Inês Duarte, Acrisio Pires, and Jason Rothman. 2011. Early inflected infinitives and late V-to-C movement. In Proceedings of the 35th Boston University Conference on Language Development, Nick Danis, Kate Mesh, and Hyunsuk Sung (eds.), 540–552. Somerville: Cascadilla Press. Schlieben-Lange, Brigitte. 1992. The history of subordinating conjunctions in some Romance languages. In Internal and External Factors in Syntactic Change, Marinel Gerritsel and Dieter Stein (eds.), 341–354. Berlin/New York: Mouton de Gruyter. Schulte, Kim. 2003. Pragmatic relevance as cause for syntactic change: The emergence of prepositional complementizers in Romance. In Historical Linguistics 2001, Barry Blake and Kate Burridge (eds.), 377–389. Amsterdam/Philadelphia: Benjamins. Schulte, Kim. 2004. Pragmatic causation in the rise of the Romance prepositional infinitive. A statistically-based study with special reference to Spanish, Portuguese and Romanian. Ph.D. diss., University of Cambridge. Schulte, Kim. 2007. What causes adverbial infinitives to spread? Evidence from Romance. Language Sciences 29: 512–537. Schulte, Kim. 2011. On the position of overt subjects in infinitival clauses in Spanish and Portuguese: Pragmatic, semantic and frequency-based motivations. Paper presented at the International Conference on Historical Linguistics 2011, Osaka, Japan. Schütze, Carson T. 2001. On the nature of default Case. Syntax 4 (3): 205–238. Schwenter, Scott A. 1999. Evidentiality in Spanish morphosyntax: A reanalysis of dequeísmo. In Estudios de variación sintáctica, María José Serrano (ed.), 65–87. Madrid: Editorial Iberoamericana. Scorretti, Mauro. 1991. Complementizers in Italian and Romance. Ph.D. diss., University of Amsterdam. Serradilla Castaño, Ana. 1995. Sobre las primeras apariciones de construcciones preposicionales ante que completivo en español medieval. Factores determinantes. EPOS XI: 147–163. Serradilla Castaño, Ana. 1996a. Diccionario Sintáctico del Español Medieval. Verbos de Entendimiento y Lengua. Madrid: Gredos. Serradilla Castaño, Ana. 1996b. Las formas analíticas de “entendimiento y lengua” ante que completivo en español medieval. Breves notas acerca de su evolución en la época clásica. Revista de Filología Románica 13: 13–38. Serradilla Castaño, Ana. 1997. El Régimen de los Verbos de Entendimiento y Lengua en Español Medieval. Madrid: Ediciones de la Universidad Autónoma de Madrid.

References

307

Serradilla Castaño, Ana. 2001. Frecuencia de preposiciones en el régimen de construcciones transitivas: hacia una periodización lingüística de la Edad Media. Cahiers de linguistique et de civilisation hispaniques médiévales 24: 143–165. Serrano, Silvia. 2009. Sustantivas precedidas por el en posición de OD. Talk given at the Seminario de Lingüística Teórica y Ciencia Cognitiva, CSIC, Madrid. http://www.lineas. cchs.csic.es/lycc/sites/lineas.cchs.csic.es.lycc/files/serrano.pdf Siminenko, Alexandra. 2010. Disappearance of Old French juxtaposition genitive and case: a corpus study. In Actes du Congrès de l’ACL 2010/2010 CLA Conference Proceedings, Melinda Heijl (ed.), 1–15. http://homes.chass.utoronto.ca/~cla-acl/actes2010/CLA2010_ Simonenko.pdf Sitaridou, Ioanna. 2009. On the emergence of personal infinitives in the history of Spanish. Diachronica 26 (1): 36–64. Skytte, Gunver. 1983. La Sintassi dell’Infinito in Italiano Moderno. Revue romane, numéro supplémentaire 27. Copenhague: Munksgaards Forlag. Skytte, Gunver and Giampaolo Salvi. 1991. Frasi subordinate all’infinito. In Grande Grammatica Italiana di Consultazione. II. I Sintagmi Verbale, Aggetivale, Avverbiale. La Subordinazione, Lorenzo Renzi and Giampaolo Salvi (eds.), 497–569. Bologna: Il Mulino. Sleeman, Petra. 2010. The nominalized infinitive in French: structure and change. Linguística. Revista de Estudos Linguísticos da Universidade do Porto 5: 145–173. De Smet, Henk. 2009. Analysing reanalysis. Lingua 119 (11): 1728–1755. Spore, Palle. 1965. La Langue Danoise. Phonétique et Grammaire Contemporaines. Copenhague: Akademisk Forlag. Stowell, Timothy Angus. 1981. Origins of phrase structure. Ph.D. diss., Massachusetts Institute of Technology. Strandskogen, Åse-Berit and Rolf Strandskogen. 1986. Norwegian: An Essential Grammar. London/New York: Routledge. Suñer, Margarita. 1999. La subordinación sustantiva: la interrogación indirecta. In Gramática Descriptiva de la Lengua Española, Ignacio Bosque and Violeta Demonte (eds.), 2149– 2196. Madrid: Espasa-Real Academia Española. Svenonius, Peter. 2007. Adpositions, particles, and the arguments they introduce. In Argument Structure, Eric Reuland, Tanmoy Bhattacharya, and Giorgos Spathas (eds.), 63–104. Amsterdam/Philadelphia: Benjamins. Svenonius, Peter. 2010. Spatial P in English. In Mapping Spatial PPs. The Cartography of Syntactic Structures, Volume 6, Guglielmo Cinque and Luigi Rizzi (eds.), 127–160. Oxford: Oxford University Press. Takahashi, Shoichi. 2010. The hidden side of clausal complements. Natural Language and Linguistic Theory 28 (2): 343–380. Tarr, Frederick Courtney. 1922. Prepositional complementary clauses in Spanish with special reference to the works of Pérez Galdós. Revue Hispanique 56: 1–264. Tekavčić, Pavao. 1980. Grammatica storica dell’italiano. II. Morfosintassi. Bologna: Il Mulino. Thráinsson, Höskuldur. 1979. On complementation in Icelandic. Ph.D. diss., Harvard University. Timberlake, Alan. 1977. Reanalysis and actualization in syntactic change. In Mechanisms of Syntactic Change, Charles Li (ed.), 141–77. Austin: University of Texas. Tobler, Adolf. 1905. Mélanges de grammaire française. Paris: Alphonse Picard et fils. Torres Cacoullos, Rena. 2009. Las nominalizaciones de infinitivo. In Sintaxis Histórica del Español, tomo II: La Frase Nominal, Concepción Company Company (ed.), 1673–1738. México: Fondo de Cultura Económica, Universidad Nacional Autónoma de México.

308

References

Travis, Lisa and Greg Lamontagne. 1992. The Case Filter and licensing of empty K. Canadian Journal of Linguistics 37 (2): 157–174. Tremblay, Mireille. 1996. Lexical and non-lexical prepositions in French. In Configurations. Essays on Structure and Interpretation, Anna-Maria di Sciullo (ed.), 79–98. Somerville: Cascadilla Press. Väänänen, Veikko. 1995. Introducción al Latín Vulgar. 3rd ed. Madrid: Gredos. Vanelli, Laura. 2010. I demostrativi. In Grammatica dell’Italiano Antico. Vol. I, Giampaolo Salvi and Lorenzo Renzi (eds.), 349–357. Bologna: Il Mulino. Vanvolsem, Serge. 1983. L’Infinitivo Sostantivatoin Italiano. Firenze: Accademia della Crusca. Vater, Heinz. 1978. On the possibility of distinguishing between complements and adjuncts. In Valence, Semantic Case, and Grammatical Relations, Werner Abraham (ed.), 21–45. Amsterdam/Philadelphia: Benjamins. Vegnaduzzo, Stefano. 2010. Frasi subordinate al congiuntivo. In Grammatica dell’Italiano Antico. Vol. II, Giampaolo Salvi and Lorenzo Renzi (eds.), 791–816. Bologna: Il Mulino. Vincent, Nigel. 1988. Latin. In The Romance Languages, Martin Harris and Nigel Vincent (eds.), 26–78. New York: Oxford University Press. Vincent, Nigel. 1997. Complementation. In The Dialects of Italy, Martin Maiden and Mair Parry (eds.), 171–178. London/New York: Routledge. Vincent, Nigel. 2000. Competition and correspondence in syntactic change: null arguments in Latin and Romance. In Diachronic Syntax: Models and Mechanisms, Susan Pintzuk, George Tsoulas, and Anthony Warner (eds.), 25–50. Oxford: Oxford University Press. Wanner, Dieter. 1987. The Development of Romance Clitic Pronouns. From Latin to Old Romance. Berlin: Mouton de Gruyter. Wanner, Dieter. 1991. Historical syntax and Old Spanish text files. In Linguistic Studies in Medieval Spanish, Ray Harris-Northall and Thomas D. Cravens (eds.), 165–190. Madison: The Hispanic Seminary of Medieval Studies. Wanner, Dieter. 2000. Beyond parameters. In New Approaches to Old Problems, Steven N. Dworkin and Dieter Wanner (eds.), 3–32. Amsterdam/Philadelphia: Benjamins. Wanner, Dieter. 2006. The Power of Analogy. An Essay on Historical Linguistics. Berlin: Mouton de Gruyter. Webelhuth, Gert. 2002. Principles and Parameters of Syntactic Saturation. New York: Oxford University Press. Whitlam, John. 2011. Modern Brazilian Portuguese Grammar. A Practical Guide. London/New York: Routledge. Wireback, Kenneth J. 1994. The origin of the Portuguese inflected infinitive. Hispania 77 (3): 544–554. Woolford, Ellen. 2006. Lexical Case, Inherent Case, and argument structure. Linguistic Inquiry 37 (1): 111–130. Yap, Foong Ha, Karen Grunow-Hårsta, and Janick Wrona. 2011. Introduction: Nominalization strategies in Asian languages. In Nominalization in Asian Languages. Diachronic and Typological Perspectives, Foong Ha Yap, Karen Grunow-Hårsta, and Janick Wrona (eds.), 1–58. Amsterdam/Philadelphia: Benjamins. Yllera, Alicia. 1980. Sintaxis Histórica del Verbo Español: Las Perífrasis Medievales. Zaragoza: Universidad de Zaragoza. Yim, Young-Jae. 1984. Case-tropism: The nature of phrasal and clausal case. Ph.D. diss., University of Washington. Yoon, James Hye Suk. 1996. Nominal gerund phrases in English as phrasal zero derivations. Linguistics 34 (2): 329–356.

References

309

Zaring, Laurie. 1991. On prepositions and Case-marking in French. Canadian Journal of Linguistics 36 (4): 363–377. Zaring, Laurie. 1992. French ce as a clausal determiner. Probus 4 (1): 53–80. Zaring, Laurie. 1993. On a type of argument-island in French. Natural Language and Linguistics Theory 11: 121–174. Zaring, Laurie and Paul Hirschbühler. 1997. Qu’est-ce que ce que? The diachronic evolution of a French complementizer. In Parameters of Morphosyntactic Change, Ans van Kemenade and Nigel Vincent (eds.), 353–379. Cambridge: Cambridge University Press. Zennaro, Luigi. 2010. 1. Frasi temporali. In Grammatica dell’Italiano Antico. Vol. II, Giampaolo Salvi and Lorenzo Renzi (eds.), 953–973. Bologna: Il Mulino. Zhang, Niina Ning. 2010. Coordination in Syntax. Cambridge: Cambridge University Press. Zucchi, Alessandro. 1993. The Language of Propositions and Events: Issues in the Syntax and the Semantics of Nominalization. Dordrecht: Kluwer Academic Publisher.

On-line corpora Biblioteca Virtual Miguel de Cervantes. www.cervantesvirtual.com Davies, Mark. Corpus del Español, www.corpusdelespanol.org Davies, Mark and Michael Ferreira. Corpus do Português, www.corpusdoportugues.org Real Academia Española. Corpus Diacrónico del Español (CORDE), www.rae.es Corpus de Referência do Português Contemporâneo (CRPC), Centro de Linguística at the Universidade de Lisboa, Portugal. Modéliser le changement: les voies du français corpus (MCVF ), developed as part of a Grands Travaux de recherche concertée (Conseil de recherches en sciences humaines du Canada) project directed by Dr France Martineau, Département de français, Université d’Ottawa/ University of Ottawa, Canada. Corpus OVI (Opera del Vocabolario Italiano) dell’Italiano antico, accessible through the website of the Istituto Opera del Vocabolario Italiano at the Consiglio Nazionale delle Ricerche and searchable with GattoWeb. . University of Chicago: Opera del Vocabolario Italiano (ARTFL). BAnca Dati dell’Italiano Parlato BADIP Karl-Franzens-Universität Graz (Austria) badip.uni-graz.at Corpus LEXIQUM, Secrétariat à la politique linguistique du gouvernement du Québec and Université de Montréal, Canada.

Subject index adjunct 23–24, 53–55, 64, 67–68, 108–109, 113 – see also adjunct finite clauses adjunct finite clauses 10, 13–14, 16, 27–29, 276–277, 281, 288 – che-clauses in historical Italian 249, 262 – overview 269–270 – que-clauses in historical Spanish 88–89, 93, 95–96, 100, 104–105, 277, 285–286 – que-clauses in historical Portuguese 168, 170 – que-clauses in historical French 203, 214, 232 – que-clauses in present-day Spanish 123– 124, 131–135, 142, 195, 224 fn 4, 277, 284 – que-clauses in present-day Portuguese 178, 195 – quod-clauses in Latin 158 – see also adverbial finite clauses adverbial finite clauses 10, 13–14, 16, 26, 28, 276, 280 – che-clauses in historical Italian 243, 245, 249–250, 260, 267, 280–281 – che-clauses in present-day Italian 262, 267, 281 – que-clauses in historical French 203, 205, 209–210, 276, 280–281 – que-clauses in historical Portuguese 162, 170, 174 – que-clauses in historical Spanish 68–69, 85–86, 93, 96, 103–105, 280–281, 284, 287, 290–291 – que-clauses in present-day French 225– 227, 231–232, 278, 281 – que-clauses in present-day Spanish 123, 128, 130, 132, 138, 140, 142, 156, 195, 284 – que-clauses in present-day Portuguese 180, 195 – quod-clauses in Latin 158 – see also adjunct analogy – criticism of syntactic analogy 87–88, 289

– and evolution of Portuguese clauses 195 – and evolution of Romance infinitives 68, 281 – and evolution of Spanish clauses 63, 68, 81–86, 88, 90, 129, 139–140, 270, 291 – extension 59, 61–63, 82, 85–86, 88–89, 140, 271, 289, 291 – as mechanism of change 55, 61–62, 289– 290 apposition 41, 86, 88, 96, 133–135, 245 – see also pronoun intervening in prepositional selection argument 10, 12, 16, 21, 29, 32, 40–41, 43, 45–47, 50, 53–55, 68 fn 2, 80, 105, 109, 115, 132, 152, 272, 277–278 – see also argumental finite clauses argumental finite clauses – che-clauses in historical Italian 241–244, 246–248, 250 – che-clauses in present-day Italian 259– 263, 267 – early attestations of prepositional clauses 92–94, 104, 171, 194, 270, 279–280, 288, 290–291 – overview 269–270 – que-clauses in historical French 204–209, 213–215 – que-clauses in historical Portuguese 168– 169, 172–173, 193–194, 196 – que-clauses in historical Spanish 85–93, 96–98, 103–106, 142, 270–271, 274, 279, 282–288, 291 – que-clauses in present-day French 220– 226, 231–232 – que-clauses in present-day Portuguese 178–180, 182–183, 194 – que-clauses in present-day Spanish 12, 109, 122–128, 132–135, 143, 149, 152– 153 – test argumental vs adjunct clause 133– 135, 232, 277 Case – abstract/syntactic 38–40, 103, 130 – adjunct 41, 105, 274

312

Subject index

– – – –

adverbial 41 criticism of abstract Case 45–46, 278 default 41, 105, 112, 131, 274 inherent 24, 38, 40, 41–44, 44 fn 8, 47, 64, 75–76, 109, 130–131, 178, 203, 274, 288, 291 – Kase 43, 75, 275, 291 – lexical 38, 40–42, 44, 75, 112, 130, 178, 194, 210, 267, 274 – materialization of (inherent) Case 42–44, 81, 130, 178, 274, 279, 288, 291 – morphological 39, 41, 45–46, 103, 112 – positional licensing of 67, 76, 81, 114, 131, 174, 178, 194, 232, 271, 274–275, 279, 288 – structural 38, 40, 44, 47, 76, 109 fn 1, 203, 251 Case features in clauses 46–48, 89, 130, 138, 145–146, 178, 194, 201, 210, 260, 274, 275–276 – Case Resistance Principle 46–47, 136, 209, 213, 225 – Case without expected selected preposition see Case positional licensing of Chinese 48 complement clause 14–15, 30, 32, 34, 38, 54, 68, 75, 85, 87–89, 95, 112, 123–124, 131–132, 135, 138, 143, 145, 163, 180, 182, 185–186, 196–197, 203, 206–208, 217, 219, 236, 249, 251, 262, 277–278, 281–283, 288 – see also argumental finite clause complementizer – as functional category 10–16 – as nominal 15–16, 143–147 fn 10, 154, 188, 195, 205, 227 fn 6, 228, 231, 250, 260, 266, 267, 271 – prepositional complementizer 36–38, 68– 70, 72–75, 99, 112–113, 129, 142, 162– 164, 176–177, 197–200, 203, 215, 217– 219, 231, 234–238, 251–252, 255, 267, 289 Danish 29, 32 dequeísmo, 129, 139 fn 8, 181, 183, 195 determiner see nominalization clausal Dutch 49–50

E-language change in 56–62, 139, 289–290 endocentricity 10, 48–49, 52, 145, 152 fn 17 English 11–14, 16–17, 25–29, 31–34, 36–37, 39–42, 46–48, 50–51, 53, 60, 68 fn 1, 80, 82, 86, 128 fn 4, 131, 138, 142–143, 153 fn 18, 225, 228, 231, 266 fn 11, 273– 274, 276, 282 exocentricity 272 factive 15, 137, 139–141, 145–146, 149–150 fn 14, 151, 154, 156, 160, 185–186, 189, 191, 193, 258 Finnish 39 French à ce que, 82, 86, 205–208, 211, 215, 219, 220–225, 227 fn 6, 246, 249, 269– 270, 276, 278, 280–281 French de ce que, 82, 86, 205–208, 211, 213, 215, 219, 220–225, 227 fn 6, 232, 247, 249, 269–270, 276, 278, 284 French de quoi, 213, 231, 270, 276 Functional Nominalization Theory see nominalization clausal German 13 fn 1, 39, 41–42, 47, 55 Germanic 28, 34, 48, 276 grammar change see I-language change in Greek (Modern) 15–16, 49, 52, 136, 138, 143, 146 Icelandic 29, 32 I-language change in 56–61, 289–290 indirect interrogative clause 11, 14, 32–34, 47 – in historical French 204–207, 214 – in historical Italian 241–242, 244, 249– 250 – in historical Portuguese 169–170, 173– 174 – in historical Spanish 81, 85, 89, 94–96, 100–102 – in Latin 160 – overview 269–270 – in present-day French 224, 226, 228–231, 233–234 – in present-day Italian 263–268 – in present-day Portuguese 179–181, 191– 193, 195–196

Subject index

– in present-day Spanish 120–121, 124–125, 127–129, 136, 140–141, 144–149, 151, 156 – see also argumental finite clauses infinitival clause 10, 11–12, 14 – adverbial 75, 176, 190, 200, 217, 219, 236, 253 – inflected infinitive in Portuguese 146, 164–165, 167, 175, 185–187, 189–191, 194, 195 – and prepositions 25, 28–29, 33, 35, 113, 67–74, 75, 82, 85–86, 88, 90, 99, 103, 112–114, 120–121, 124, 130, 138, 138 fn 7, 141, 154–156, 159–165, 170, 175–178, 180, 182–183, 197–201, 215– 220, 232–234–240, 251–255, 262, 267, 269–270, 289–291 – see also complementizer prepositional complementizer; indirect interrogative clause infinitive see infinitival clause Italian acciocché, 82, 245–246, 262–263, 267 Italian a che, 242–243, 246, 248–250, 261– 263 fn 10, 266–267, 269–270, 276 Japanese 48, 136 Kase see Case Korean 15, 48, 136 language change see E-language change in mood selection 97–98, 105, 106, 132, 224 fn 4 nominal complementizer see complementizer as nominal nominal feature in clauses 48, 89–90, 92, 145–146, 149, 195, 201, 203 fn 2, 271– 272 – see also complementizer as nominal nominalization clausal – constraints in Spanish 149–156, 273 – Functional Nominalization Thesis 48–49 – in historical French 201–203, 209, 213, 276

313

– in historical Italian 239–241, 250 – in historical Portuguese 165–168, 173–174 – in historical Spanish 76–80, 100–102, 106 – in Latin 159 – in present-day French 220, 222–225, 227 fn 6, 230–234 – in present-day Italian 254–259, 265–268 – in present-day Portuguese 187–193, 195– 196 – in present-day Spanish 114–121, 135–156 – substantivization/re-nominalization 53, 142, 145, 153, 188, 193, 202–203, 214 fn 3, 234, 239, 241, 250, 256, 257, 266–267, 271–274 – Switch Categorial 51–52 – syntactic representation of 16, 48–53, 272–274 Norwegian 29, 32, 136 optionality prepositional 33, 71, 97, 103, 127, 131, 160, 162, 179, 198–199, 214– 216, 222, 231, 267, 275–276, 287, 289 – see also queísmo Persian 136 Polish 136 Portuguese inflected infinitive see infinitival clause inflected infinitive in Portuguese preposition – functional 18–25, 37–38, 42–45, 64, 67, 75, 83, 86, 99, 103, 109–112, 114, 125, 130, 133–134, 142, 169, 178, 182, 190, 194, 203, 225, 232, 251, 253, 271, 274– 275, 277–279, 289 – lexical 18–21, 28, 38, 41–42, 44 fn 9, 64, 67–68, 75, 83, 89, 93, 95, 100, 103– 104, 106, 108–113, 123, 130, 138, 140– 142, 158, 162, 165, 174, 176, 178, 180– 181, 190, 194, 200–201, 203, 210, 215, 217, 219, 232, 236, 243, 245, 253, 255, 267, 274, 277–278, 291 – test lexical vs. functional 19–21, 109– 112 – underlying 29–32, 275 prepositional complementizer see complementizer presuppositional see factive

314

Subject index

prolepsis 83–84, 86, 205, 243, 250 pronoun intervening in prepositional selection – in historical French 203, 205, 208–213, 276 – in historical Italian 249–250 – in historical Portuguese – in historical Spanish 96, 100 – in Latin 157–159 – in present-day French 225–226 – pronoun-drop analysis 86–87, 91–92, 159, 212, 270 – in Swedish 277 – see also French à ce que; French de ce que; Italian acciocché; Italian a che; prolepsis; Spanish a que; Spanish de que; Spanish en que

Spanish a que, 85, 93, 97–98, 122–123, 125, 132–133, 279–284, 289, 291 Spanish de que, 84, 92, 97–100, 125, 130– 135, 150, 152–154, 279–280, 284–285, 289–291 – see also dequeísmo Spanish en que, 85, 92, 109, 125, 132, 280, 286, 291 strict adjacency 106, 150–156, 225, 259, 273 subject position clause in 12–13, 34, 47, 51, 73, 136, 164, 185, 200, 217, 258, 289 subordinating conjunction 10, 14, 34, 81, 88, 93, 149, 154, 179, 227, 248 see also complementizer Swedish 28–29, 32, 34, 137–138, 141–142, 158, 213, 274, 276–277 Switch Complement Hypothesis see nominalization clausal Switch Categorial

queísmo, 125, 141, 172, 179 reanalysis – as mechanism of change 59–62 – and evolution of Romance infinitives 68, 281 – and evolution of Spanish clauses 87, 154, 281–288 – prepositional verbs 23–24, 45, 109

θ-role marking 20, 22–24, 40–42, 45, 133 Turkish 50, 79, 118, 167 variation prepositional 23, 65–67, 69–75, 96–100, 107–109, 125–130, 132, 141, 161–164, 198, 214, 238, 252 – see also optionality prepositional