Different Slants on Grammaticalization [e-book ed.] 9027252815, 9789027252814

Table of contents :
Different Slants on Grammaticalization
Editorial page
Title page
Copyright page
Table of contents

Introduction
1. Grammaticalization
2. Defining grammaticalization and grammar
3. Wide vs narrow views of grammaticalization
4. Recent issues in grammaticalization theory
5. The present volume
References

Section I Diachronic approaches
Chapter 1 From comparative standard marker to comparative adverb
1. Introduction
2. A short history of yori
3. Preceding studies on yori and language contact
3.1 Preceding studies on yori
3.1.1 Hida (1992, 2019)
3.1.2 Morioka (1999)
3.2 Language contact research on Japanese
4. Corpora
5. Survey results and discussion
5.1 Gradualness of grammaticalization
5.2 Collocational sequences that preserve the earlier ablative function of yori
5.2.1 Yori-collocations with particles
5.2.2 Yori co-occurring with sarani ‘moreover’
5.3 A collocational sequence that reflects the newer adverbial function of yori
5.4 Language contact and ‘extravagance’
5.5 Degrammaticalization and the emergence of the comparative adverb yori
6. Concluding remarks
Acknowledgments
Glossing conventions
Division of the history of Japanese in this study
References
Dictionaries
Corpora
Appendix. The major collocational patterns with yori (based on Shibasaki in press)
Chapter 2 From fear to reason
1. From fear to reason cross-linguistically
1.2 Grammaticalization as dependency vs expansion
1.3 Affectives in Mandarin
1.4 Apprehensives in Mandarin
2. Data retrieval
3. Analysis
4. Conclusions
References
Chapter 3 -maɾa in Mara
1. Introduction
2. On Mara
3. On -maɾa
4. The two auxiliary verb constructions with -maɾa
4.1 Expressing already
4.2 Expressing not yet
5. The grammaticalization of the -maɾa AVCs
5.1 ALREADY
5.1.1 Relaxation of selectional restrictions and expansion in collocational range
5.1.2 Loss of verbal properties and syntactic variability
5.1.3 Shift in anaphoric referencing
5.1.4 Loss of phonological substance
5.2 NOT YET
6. Old forms and cycles of change
6.1 The original formal means for expressing already/not yet
6.2 Diachronic cycle(s) of phasal polarity
6.3 The PhP cycle for -maɾa in Mara
7. Summary and conclusions
Acknowledgements
References

Section II Synchronic approaches
Chapter 4 Tracking Jespersen’s cycle in Veronese and Bresciano
1. Introduction
1.1 The expression of sentential negation
1.2 Grammaticalization of mica in standard and local varieties
1.2.1 Mica in Modern and Old Italian
1.2.2 Mica in Italo-Romance varieties
2. Methodology and research questions
3. Data analysis
Indicative
Protasis
Imperatives
Subjunctive subordinates
Discussion and conclusion
List of abbreviations
Note to the glossing system
References
Appendix
Chapter 5 Could be, might be, maybe
1. Introduction
2. Epistemic phrases and adverbials
3. Epistemic phrases/adverbials in an experimental study
3.1 Design and participants
3.2 Results
3.2.2 Comparing shadowed responses with input forms
3.2.3 Durations of shadowed items
4. Discussion
5. Conclusion
References
Chapter 6 The final-appendage construction in Japanese and Korean
1. Introduction
2. Data
3. Results
3.1 Japanese final-appendage structures and their Korean translations
3.2 Korean translations of Japanese utterances with the final-appendage construction
3.3 Distribution by grammatical function
4. Post-predicative differences in form and frequency
4.1 Adnominals
4.2 Nominals
4.3 Particle-less pronouns
4.4 Vocative phrases
4.5 Summary
5. Constructional entrenchment of final-appendage structures
6. Conclusion
Acknowledgements
Keys to abbreviations
References
Data sources
Chapter 7 New perspectives on phonological erosion as an aspect of grammaticalization
1. Introduction
2. The loss of phonological bulk
3. Phonological loss, compensation, and the efficiency constant
4. Accelerated and uncompensated loss as an aspect of right-sizing
5. Right-sizing extended to cases of grammaticalization
6. Right-sizing as a complex adaptive process with stochastic effect
7. Conclusions
References

Section III Interactive contexts
Chapter 8 On the development of discourse markers from elliptical structures
1. Introduction
2. Preliminaries
3. Data and EDM functions
3.1 DM functional domains
3.2 Interaction management functions of EDMs
3.2.1 Pause-filling
3.3 Information management
3.3.1 Disregard
3.3.2 Mirativity
3.3.3 Mitigation/deflection
3.3.4 Uncertainty
3.4 Interlocutor management
3.4.1 Feigned surprise
3.4.2 Challenge/protest
3.4.3 Discontent
3.4.4 Reproach
3.4.5 Sarcasm
3.4.6 Hesitance/reluctance
3.4.7 Upcoming disalignment
4. Historical development
4.1 Mwusun ‘what kind of’
4.2 Ilen ‘this kind of’
4.3 Celen ‘that kind of’
4.4 Weyn ‘what kind of’
4.5 Weynkel ‘what kind of a thing’
4.6 Mwel ‘what’
5. Discussion
5.1 Significance of elliptical structure
5.2 DM characteristics
5.2.1 Morphosyntax
5.2.2 Semantics
5.2.3 Phonology
5.2.4 Pragmatics
5.3 Source characteristics
6. Summary and conclusion
Funding
Abbreviations
Acknowledgements
References
Chapter 9 On the grammaticalization of ideophones
1. Introduction
1.1 On ideophones
1.2 Definition
1.3 Grammatical features
2. Ideophones as interactives
2.1 Interactives
2.2 Problems
3. Grammaticalization
3.1 Diagnostics
3.2 Ideophones in Siwu
3.3 Ideophones in Xhosa
3.3.1 The quotative construction
3.3.2 The modifier construction
3.4 Discussion
4. Conclusions
Abbreviations
References
Chapter 10 An emerging final particle
1. Introduction
2. Data
3. quoi in dictionaries
4. Hopper’s (1991) five principles of grammaticalization
4.1 Principle 1
4.2 Principle 2
4.3 Principle 3
4.4 Principle 4
4.5 Principle 5
5. Grammaticalization of the particle and beyond
6. Conclusion
References
Dictionaries
Corpus

Index

Citation preview

         

Different Slants on Grammaticalization Edited by Sylvie Hancil Vittorio Tantucci

   

Different Slants on Grammaticalization

Studies in Language Companion Series (SLCS) issn 0165-7763 This series has been established as a companion series to the periodical Studies in Language. For an overview of all books published in this series, please see benjamins.com/catalog/slcs

General Editor

Founding Editor

Elly van Gelderen

Werner Abraham

Arizona State University

University of Vienna / University of Munich

Editorial Board Alexandra D’Arcy

University of Victoria

Tine Breban

The University of Manchester

William A. Croft

University of New Mexico

Östen Dahl

Christian Lehmann University of Erfurt

Elisabeth Leiss

University of Munich

Marianne Mithun

University of California, Santa Barbara

Heiko Narrog

University of Stockholm

Tohuku University

Gerrit J. Dimmendaal

Johanna L. Wood

Marianne Hundt

Debra Ziegeler

University of Cologne University of Zurich

University of Aarhus University of Paris III

Ekkehard König

Free University of Berlin

Volume 232 Different Slants on Grammaticalization Edited by Sylvie Hancil and Vittorio Tantucci

Different Slants on Grammaticalization Edited by

Sylvie Hancil University of Rouen

Vittorio Tantucci Lancaster University

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

doi 10.1075/slcs.232 Cataloging-in-Publication Data available from Library of Congress: lccn 2023007770 (print) / 2023007771 (e-book) isbn 978 90 272 1375 4 (Hb) isbn 978 90 272 5281 4 (e-book)

© 2023 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Company · https://benjamins.com

Table of contents Introduction section i. Diachronic approaches chapter 1. From comparative standard marker to comparative adverb: On the contact-induced (de)grammaticalization of yori in modern through present-day Japanese Reijirou Shibasaki

1

20

chapter 2. From fear to reason: Grammaticalization as dependency vs expansion of the Mandarin apprehensive 怕 pà Vittorio Tantucci & Aiqing Wang

50

chapter 3. -maɾa in Mara: On the ongoing grammaticalization(s) of a Bantu ‘finish’ verb Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

74

section ii. Synchronic approaches chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano: Diatopic variation in sentential negation Marta Tagliani & Jelena Živojinović chapter 5. Could be, might be, maybe: Mechanisms of grammaticalization in synchronic use and perception David Lorenz chapter 6. The final-appendage construction in Japanese and Korean: To what extent is post-predicative position exploited in the two East Asian languages? Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization Charles Elerick

104

124

147

176

vi

Different Slants on Grammaticalization

section iii. Interactive contexts chapter 8. On the development of discourse markers from elliptical structures Seongha Rhee

198

chapter 9. On the grammaticalization of ideophones Alexander Andrason & Bernd Heine

237

chapter 10. An emerging final particle: The case of quoi ‘what’ in French Sylvie Hancil

263

Index

283

Introduction Vittorio Tantucci & Sylvie Hancil

University of Lancaster | University of Rouen

1.

Grammaticalization: Historical background

The term ‘grammaticalization’ was presumably first formalised in 1912 by the historical linguist Meillet. However, the grammatical evolution of languages was already amply discussed in the 19th century by linguists such as Bopp (1816; 1833), Wüllner (1831), or von der Gabelentz (1961[1891]). Grammaticalization arguably became a recognised sub-discipline of linguistics in the 1970’s, especially with the work of Givón, famously stating that language structure can be really understood once one gets to know how it evolved, which was embodied by the famous expression “Today’s morphology is yesterday’s syntax” (Givón 1971: 12, 1979). With a similar ethos, key books in the field of grammaticalization were produced by Lehmann (1995[1982]), Heine and Reh (1984), along with Traugott and Heine (1991a, 1991b). It was perhaps with the two textbooks by Heine et al. (1991) and Hopper and Traugott (1993) that grammaticalization was then widely recognised as an independent field of study within linguistics. These seminal works allowed the grammaticalization approaches to further expand into cross-linguistic and typological studies (Bybee et al. 1994; Ramat & Hopper 1998; Fischer et al. 2000). However, despite the growing interest towards a framework that could somewhat ‘predict’ how grammars evolve over time, poignant critiques of the theory were made in a special issue of Language Sciences (Campbell 2001; see also Newmeyer 1998). Others similarly cast doubt on the unidirectional hypothesis of grammatical change – i.e. the idea that the process of linguistic items becoming more grammatical is irreversible – starting with Ramat (1992) and culminating in Norde (2009) (see also Haspelmath 1999; Börjars & Vincent 2011 for critical discussions). The theory subsequently gathered new momentum with the conference series “New Reflections on Grammaticalization” (Wischer & Diewald 2002; Fischer et al. 2004; López-Couso & Seoane 2008; Seoane & López Couso 2008; Davidse et al. 2012; Smith et al. 2014); with the publication of key volumes such as Bisang et al. (2004), Stathi et al. (2010), Traugott & Trousdale (2010) and Van linden et al. (2010). Grammaticalization theories were also put into play in special issues in Language Sciences (Norde et al. 2013), Folia Linguistica (Von Mengden https://doi.org/10.1075/slcs.232.intro © 2023 John Benjamins Publishing Company

2

Vittorio Tantucci & Sylvie Hancil

& Simon 2014), again in Language Sciences (Breban & Kranich 2015), and in Heine & Kuteva’s book “World Lexicon of Grammaticalization” (2002), along with Narrog & Heine’s “The Handbook of Grammaticalization” (2011). After the turn of the century, grammaticalization also made its way into the generative paradigm (Roberts & Roussou 2003; van Gelderen 2004) and, preponderantly, in to Construction Grammar (i.a. Traugott & Trousdale 2013; Coussé et al. 2018; Pons & Loureda 2018; Sommerer & Smirnova 2020). Further advances into typological enquiry of language change have been made with the edited book “Grammaticalization from a Typological Perspective” (Narrog & Heine 2018) whereby areal overviews are provided about how grammaticalization occurs in different languages and language groups. New functional perspectives on data and by-products of grammaticalization phenomena have been flagged in edited books by Hancil & König (2014), Smith et al. (2015), van Olmen et al. (2016) and by Hancil et al. (2018). It is also important to note the recent and in many ways innovative textbook Grammaticalization by Narrog & Heine (2021: 1).

2.

Defining grammaticalization and grammar

As the name suggests, grammaticalization involves a process in which something becomes or is made grammatical (Lehmann 1995: 11). Another classic definition by Hopper and Traugott (2003: 18) treats grammaticalization as the “change whereby lexical items and constructions come in certain linguistic contexts to serve grammatical functions, and once grammaticalized, continue to develop new grammatical functions”. More recent is Boye and Harder’s definition (2012: 21) who suggest that “Grammaticalization is the diachronic change that gives rise to linguistic expressions that are by convention ancillary and as such discursively secondary”. Narrog and Heine (2021: 1) take a broader stance, as they view grammaticalization as “the development from lexical to grammatical forms and once the grammatical form has evolved, the development of further grammatical forms”. They also emphasise that “since the development of grammatical forms takes place in specific contexts and constructions, the study of grammaticalization is also concerned with constructions and context, including even larger discourse segments” (Narrog & Heine 2021: 1) But what is exactly grammatical and what is not? While an answer to this question may seem intuitive at first, in the last 30 years scholars in theoretical and historical linguistics found it increasingly challenging to identify grammar as a discrete category. For instance, a notoriously controversial phenomenon in the literature is the diachronic change and formation of discourse markers (DM), elsewhere defined as pragmatic markers (PM) (cf. Brinton 2017 for a detailed

Introduction

discussion). Do they undergo grammaticalization? Can they be considered as grammatical items? Is it necessary to adopt an alternative notion for such phenomena where lexical material develops into pragmatic markers, as captured by the concept of ‘pragmaticalization’ (Erman & Kotsinass 1993; Aijmer 1997: 2)? In this respect, a distinction is often made between so-called ‘sentence-grammatical phenomena’, and ‘discourse-pragmatic phenomena’, (Günthner & Mutz 2004). Pragmatic markers are often free from morpho-syntactic constraints such as obligatorification (when the choice is systematically constrained, use largely obligatory), paradigmaticization (small, tightly integrated paradigm), attrition (few semantic features; oligo- or monosegmental), condensation (the item modifies word or stem), coalescence (item is affix or even phonological feature of carrier), fixation (item occupies a fixed slot) (e.g. Lehmann 1995: 174). It is no secret that such view of grammar traditionally takes inflection as an important factor, as most of these tendencies typically move in the direction of morphosyntactic and phonological reduction. However, the boundary between grammar and discourse is not always easy to draw. If the think of sentence final particles (SFP) in a number of languages of the Southeast – such as Chinese or Thai – it is not straightforward whether these serve an exclusively grammatical vs discursive function. On the one hand, they underwent fixation as they are structurally bound to the end of the sentence, on the other, they are often not obligatory (Tantucci & Wang 2018, 2020b). The latter is a feature that is also shared by Chinese aspectual markers (cf. Wu 2005). SFP undoubtedly underwent paradigmaticalization, but being Chinese and Thai isolating languages, coalecence is less applicable than what we can see for inflectional languages such as Germanic or Romance. Have they undergone grammaticalization or pragmaticalization? Diewald (2011a) suggests that pragmaticalization can be treated as a subprocess of grammaticalization. Wischer (2000: 359) similarly proposed to tackle pragmaticalization as ‘grammaticalization on the text or discourse level’ and grammaticalization as ‘grammaticalization on the propositional level’. Arguably, in both processes linguistic items undergo recategorialization (see also Hopper on de-categorialization 1991: 22), as they change from a more open to a closer categorial system. Related to this issue, Nicolle (2011) suggests that grammaticalization primarily underpins the addition of procedural information (e.g. Traugott & Trousdale 2013) to the semantics of an expression. Linguistic expressions encode procedural meaning when they “contribute information about how to combine […] concepts into a conceptual representation” (Terkourafi 2011: 358–359), e.g. in indexical reference and information-structure marking (topic, definiteness, etc.), argument-structure marking (case), marking of temporal phase (aspect) or relationship to time of speaking (deictic tense). Accordingly, lexical items express

3

4

Vittorio Tantucci & Sylvie Hancil

conceptual information, while discourse markers, pronouns, tense, aspect, and modality markers contribute to procedural information. Such controversies also reflect a theoretical reassessment of the very notion of grammar in several paradigms over the last 30 years. Indeed, one of the most influential publications on the typology of grammatical change was the “Evolution of Grammar” by Bybee et al. (1994), which is centred on the way languages have a tendency towards phonological reduction and grammatical categorisation, which in their model is embodied by the progressive formation of ‘grams’, i.e. grammatical morphemes. This perspective does not differ from Lehmann’s strong emphasis on grammaticalization as a primarily structural mechanism that is necessarily geared towards inflectional and/or morphosyntactic reduction of linguistic material. An alternative view of grammatical knowledge and sedimentation in speakers’ mind was at the core of Construction Grammars (e.g. Langacker 1987; Fillmore 1988; Goldberg 1995, 2006; Traugott & Trousdale 2013), according to which lexicon and grammar are conceptually integrated, and therefore cannot be seen as discrete categories. This is a point that was also made in Hopper (1991: 19), when he suggested that “the only way to identify instances of grammaticization would be in relation to a prior definition of grammar; but there appear to be no clear ways in which the borders which separate grammatical from lexical and other phenomena can be meaningfully and consistently drawn”. In constructionist approaches, a grammar is stored in speaker’s memory as a large inventory of constructions, which constitute pairs of form and meaning of varying degrees of abstraction (which in the framework is referred to as schematicity) and productivity (the degree to which a construction is open to different/new lexical items). This means that the very schematic notion of subject-auxiliary inversion [SAI] can be seen as a construction, as much as the specific expression [kick the bucket] or the word [red]. What this entails for a theory of language change is that linguistic items that become more grammatical are primarily ones that are conceptualised as procedural elements that serve linguistic functions, e.g. such as marking the phasal realisation and/or the view point (i.e. aspect) of an event (as for the partly schematic progressive construction [BE V-ing], or the perfect one [HAVE, V-ed]), encoding definite reference (as for the specific definite article construction [the]) and so on. In this sense, the emphasis is not quite on the way linguistic material becomes phonologically reduced, but more distinctively on the way it serves general functions of language use, i.e. ones that allow speakers to ‘procedurally’ organise lexical and contentful material. In addition to procedurality, constructionist models also assume that grammatical change is gradual and incremental and leads to an increase in productivity and schematicity (e.g. Gisborne & Patten 2011; Traugott & Trousdale 2013).

Introduction

As pointed out in Heine (2011), a key aspect of this debate is whether maintaining that schematization in constructions and grammaticalization are two different types of change, or rather searching for an overarching theoretical framework that encompasses both (see also Nöel 2007). This latter approach often led to the inclusion of pragmatic markers and discursive functions as components of grammar, and therefore as items that undergo grammaticalization.

3.

Wide vs narrow views of grammaticalization

A distinction that ensued as a result of this debate was the one between a ‘narrow’ view in contrast with a ‘wide’ view of grammaticalization. The former is distinctively compatible with the diagnostics formulated in Lehmann (1982, 1995) but does not include DMs/PMs (Degand & Simon-Vandenbergen 2011: 290). The wide view takes into account DMs/PMs, as well as other forms of discourseoriented expressions (e.g. Traugott 1995; Brinton 2001; Barth-Weingarten & Couper-Kuhlen 2002; Mosegaard Hansen 1998; Lenker 2000; Traugott & Dasher 2002; Brinton & Traugott 2005: 136–140; Prévost 2011; Diewald 2011a, 2011b; van Bogaert 2011; Degand & Evers-Vermeul 2015: 67). The wide view rejects the early notion of pragmaticalization (cf. Traugott 1995, 2007) in favour of a distinction between two kinds of approaches, referred to as ‘GR’ (grammaticalization as reduction) and ‘GE’ (grammaticalization as expansion). In GR models, grammaticalization is viewed as reduction and increase in dependency. Examples of such approaches are Givón (1979), Heine & Reh (1984), Heine et al. (1991), Lehmann (1995, 2004), Bybee et al. (1991), Haspelmath (2004), and Boye & Harder (2012). While GR maintains that grammaticalization tends to be unidirectional for the most part (but see Norde 2009 for an in-depth account of degrammaticalization), GE may also involve ‘multidirectionality’ (Traugott & Trousdale 2013: 108). As remarked by Haselow (2014: 210), when understood as expansion and extra-sentential material, grammaticalization can be identified as involving (i) semantic change from conceptual/referential to procedural meanings and (ii) from truth-conditional to non-truth-conditional meanings, (iii) decategorialization (i.e. loss of morphosyntactic properties of the source forms) (iv) scope expansion from proposition-internal over proposition-external scope to scope over discourse, (v) a structural change from clause/sentence-internal to clause/sentence-external position, which often coincides with a movement to new structural positions, typically peripheral ones, and (vi) context extension or context generalization, i.e. use in an increasingly larger number of contexts (i.a. Tabor & Traugott 1998; Traugott & Dasher 2002; Heine 2002; Auer & Günthner 2005; Fanego 2010). Concerning this point, it is worth mentioning Heine’s (2011; 2018)

5

6

Vittorio Tantucci & Sylvie Hancil

clear stance against the wide view, as he provides an alternative explanation for the development of discourse markers. He argues that discourse markers involve extra-clausal material and are therefore to be treated as ‘theticals’ (see Dik 1997, as first discussed in Kaltenböck et al. 2011). Theticals can be the result of cooptation, which, in turn, hinges on some segment of linguistic discourse being transferred from one domain of discourse to another. Heine argues that the combination of coopotation + grammaticalization explains the special features of the development of discourse markers as components of the linguistic systems that should be kept separate from grammar. This is a model that is further discussed in the book “The rise of discourse markers” (Heine et al. 2021). In turn, a wide approach on the same issue is also elaborated in “Discourse Structuring Markers in English: A historical constructionalist perspective on pragmatics” by Elizabeth Traugott (2022).

4.

Recent issues in grammaticalization theory

Grammaticalization has been a key resource in the field of typology as it shed new light on cross-linguistic similarities among clines of change in different languages. Whilst being first centred on the diachronic study of languages with richly documented historical data, such as English, French, Spanish, etc. (but see LópezCouso & Seoane 2008), it was with the Oxford Handbook of Grammaticalization (Heine & Narrog 2001) that grammaticalization studies were widely extended to an increasingly diverse range of world’s languages (cf. Hancil et al. 2018). These and other studies put new emphasis not only on the effects of grammaticalization, but also on the motivations that trigger language change. A wide array of mechanisms have been theorised and redefined, such as reanalysis (Harris & Campbell 1995; Hopper & Traugott 2003), also refered to as neoanalysis (Croft 2000; Andersen 2001; Traugott & Trousdale 2013), routinization (Haiman 1994; Bybee 2010), entrenchment (e.g. De Smet 2017; Tantucci & Di Cristofaro 2020; Schmid 2020) analogy (Fischer 2008), ease of effort, economy, and perceptual factors (Fischer 2006), to name a few. An important pragmatic mechanism leading to grammatical change is famously argued to be extravagance, i.e. the desire to be noticed by others, with the corresponding effect of being over-informative in expressing grammatical relations (Keller 1994; Haspelmath 1999; Detges & Waltereit 2002). Accommodation and resonance are also treated as important pragmatic triggers of grammaticalization, as they involve changes that are determined by (dis-)alignment of speech to the other interlocutors (Beuls & Steels 2013; Schwenter & Waltereit 2010; Tantucci et al. 2018; Tantucci & Wang 2022). A key issue that arose with the new century was the intersection between a gradient synchronic system and grammaticalization (Aarts 2007; Traugott & Trousdale 2011), i.e. the degree to which it is possible to infer how neo/reanalysis

Introduction

may have been taking place in constructions that share some semantic and morphosyntactic analogies. Despite not strictly tackling grammaticalization per sé, synchronic approaches to change were also at the heart of sociolinguistic studies that relied on the so-called ‘apparent-time method’, taking speakers’ age cohorts as a proxy for linguistic changes across specific timespans (Sankoff & Blondeau 2007: 582; Tagliamonte & D’Arcy 2009: 61). Synchronic was also the innovative approach to grammaticalization developed by Croft (2007, 2010) who compared alternative verbalizations of the same experience of speakers in a controlled situation. Examples of innovation in lexical semantic change and grammaticalization were examined using parallel English narratives. The results challenged the traditional assumption that morphosyntactic innovation is rare and special mechanisms are needed for it to occur. Instead, the study suggested that grammaticalization may originate in the variation that is inherent in the verbalization of experience. New emphasis has also been given to real-time dynamics of the individual and the community in grammaticalization (e.g. Petré & Van de Velde 2018). Another fundamental aspect of grammaticalization which has become increasingly discussed in recent years is intersubjectivity, i.e. whether grammaticalization involves meanings that become increasingly ‘social’, that is ones that are increasingly centred on the addressee’s – or a specific/generic 3rd party’s – potential reactions to what is being said (Tantucci 2021). While several definitions of the notion of intersubjectivity have been proposed in the literature (e.g. Langacker 1991; Traugott & Dasher 2002; Traugott 2003; Verhagen 2005; Nuyts 2012; Tantucci 2015, 2017a), the most influential approach to intersubjectification as a diachronic phenomenon has probably been the one by Traugott (2013: 124) who sees it as “the development of meanings that encode speaker/writer’s attention to the cognitive stances and social identities of addressees” (Traugott 2003: 124). Potential intersections between intersubjectification, grammaticalization and unidirectional change have been at the core of dedicated volumes by Ghesquière et al. (2012) and Van Olmen et al. (2016). Another growing area of interest in broad approaches to grammaticalization hinges on semantic-pragmatic analogies between language change and language acquisition (e.g. Ellis & Larsen-Freeman 2006; Hengeveld 2011; Tantucci 2017b, 2020; Wei 2002). While phonological reduction and morphosyntactic erosion are controversial diagnostics for comparing diachronic and ontogenetic change (but see some notable exceptions in Diessel 2011), however remarkable similarities have been noted for unidirectional patterns of semantic-pragmatic reanalysis, as in both cases similar adaptive behaviours at play (Givón 2009). This has been tackled in particular with respect to joint attention (Ziegeler 1997; Diessel 2006) and semantic-pragmatic shifts from the awareness of one specific mind to the conceptualisation of social meanings, i.e. from immediate to extended intersubjectivity (Tantucci 2021b; Tantucci & Wang 2020; Tantucci & Di Cristofaro 2021). The

7

8

Vittorio Tantucci & Sylvie Hancil

research strands centred on intersubjectivity also led to a new dialogic turn in grammaticalization studies, whereby language change is addressed as a joint activity among interlocutors where structural and semantic-pragmatic reanalysis are triggered by both interlocutors in interaction (e.g. Haselow 2014, 2015; Haselow & Hancil 2018; Tantucci et al. 2018).

5.

The present volume

The present issue advances the theory and enriches the afore-mentioned debates in different ways. The notion of linguistic gradience is reflected in the synchronic approach of Marta Tagliani and Jelena Živojinović’s paper, which is centred on Jespersen’s cycle of grammaticalization of negative systems. Languages with exclusive preverbal negative markers (Stage I) often first get through a stage of discontinuous negation (Stage II), where the preverbal negative marker appears together with a postverbal negative element (cf. van der Auwera 2009 for a different account). Finally, the original preverbal marker disappears (Stage III), and the postverbal element conveys alone the negative semantic meaning to the sentence. This trend is confirmed by Tagliani and Živojinović with a diatopic-variational study on expressions of sentential negation in two Italo-Romance varieties, i.e., Bresciano and Veronese. In the Veronese peripheral data, the particle mia (local form of mica) functions as a reinforcing negative element, whereas central Bresciano shows a homogeneous pattern of post-verbal negation. Finally, in Lumezzanese the negative particle mia appears to be completely grammaticalized. Another synchronic approach to diachronic change is the one by David Lorenz, as he focuses on epistemic phrases such as (it) could/might be (that) in English as potential candidates for grammaticalizing into sentence adverb(ial)s. The aim is to check whether reduced forms (i.e. ones with it-omission) are preferred in potentially grammaticalizing contexts, e.g. when modifying a main clause ((it) could be this is correct). Indeed, corpus data from this study suggests that it-omissions are found in critical context across items (could be, might be) and register (spoken, informal writing). A ‘continuous shadowing’ experiment also partly confirms this, but at the same time shows some item-specific variation. What this indicates is that grammaticalizing contexts have an immediate effect on formal reduction and that language users have an active intuition for emerging or potential variational patterns. Vittorio Tantucci and Aiqing Wang provide the case study of the Mandarin apprehensive verb 怕 pà ‘to be afraid’, which developed a new epistemic meaning ‘to believe, to conclude that’. Apprehensives (e.g. Jing-Schmidt & Kapatsinski 2012) are known to shift diachronically from meanings of fear to ones of reason

Introduction

and have been observed in a number of languages, including the English construction I am afraid (Leech 1983; Palmer 2001). The paper provides corpus-data and a mixed effects logistic regression analysis showing that the transition from fear to reason in Chinese involves grammaticalization as increased dependency for constructions such as 不怕 bù pà ‘no fear’ > ‘although’ and 只怕 zhǐ pà ‘only fear’ > ’probably’. However, it also leads to grammaticalization as expansion, such is the case of 恐怕 kǒng pà ‘to be afraid’ > ‘to conclude that’. This supports the view that grammaticalization is not necessarily bound to increased reduction and dependency, but can be at work also as a mechanism of functional expansion and heterosemy (Lichtenberk 1991), even along the same cline of change, such is the one involving the same source (fear) and target (reason) domains of reanalysis. Reijirou Shibasaki’s paper also contributes to typological debates on grammaticalization as expansion (under the so-called broad view) as it discusses the change of the Japanese ablative marker yori, turning in to a comparative standard marker ‘than’ first and a comparative adverb ‘more’. Crucially, the progress of grammaticalization of yori appears as distinctively slow. Shibasaki suggests that this may be because functional expansion from a comparative standard marker to a comparative adverb is an unusual case in grammaticalization, and some collocational expressions may have partly impeded the development of yori becoming adverbial. Language contact with Western Languages is also a potential trigger of grammaticalization as intellectuals in the Meiji period dealt consciously with such differences in morphosyntactic structures between Japanese and western languages, giving rise to the newer comparative adverbial usage of yori, presumably facilitated not by the speakers’ un intended use of it but by their ‘intended’ learning of it. Another aspect of the study is that one collocational sequence yori issoo seems to have boosted the functional expansion of yori from a comparative standard (ablative) marker to a comparative adverb. This endorses the view that grammaticalization can be witnessed or facilitated in sequences of elements, but not in isolation (Traugott 2003; Himmelmann 2004). Lastly, Shibasaki points out that expansion as coexistence of ablative and adverbial usages at one synchronic stage is a good case of ‘layering’ (Hopper 1991), assisted by language contact. Mitsuko Narita Izutsu, Katsunobu Izutsu and Yong-Taek Kim provide a comparative analysis of parallel texts of Japanese novels and their Korean translations. They argue that the two languages, despite both being predicate final, differ in the exploitability of post-predicative positions. Korean’s clause-internal elements are less likely to occur after the predicate. The study yet shows a relatively higher proportion of vocative phrases in post-predicative position. In contrast, Japanese exhibits a relatively higher degree of constructional entrenchment of finalappendage structures. This entails that the post-predicative position in Korean is

9

10

Vittorio Tantucci & Sylvie Hancil

rather restrictive, i.e. comparatively more limited to clause-external elements than to clause-internal ones. Seongha Rhee’s paper similarly contributes to the broad/narrow debate as it is centred on a set of Korean DMs that arose from elliptical structures (EDMs), i.e ones characterised by the accusative case-marker or adnominalizer being still part of the construction. Such morphological markers would normally require the construct to occur in the theme-argument position or pre-nominal modifier position, EDMs have been through a process of expansion as they occur at non-licensed positions or even stand alone. They carry diverse functions across the interaction-, information and interlocutor-management domains, e.g. the speaker’s feigned surprise, discontent, challenge/protest, disregard, and so on. Rhee suggests that the development of these new affective and attitudinal functions are triggered by structural ellipsis, whereby the speaker is unable to complete the utterance as s/he is overwhelmed by the event or state of affairs that is being talked about. This change is accompanied with semantic generalization, whereby the scope is extended. This leads EDMs to move out of the licensed position and acquire new syntagmatic freedom. Interestingly, such DM function is not exclusively derivable from the source lexemes but also from structural changes (i.e. ellipsis). Something that sets EDMs apart from morphosyntactic, semantic, phonological, and pragmatic aspects typical of Discourse markers is also that they are often correlated with quite specific prosodies. Finally, Rhee identifies an important role of deixis and interrogatives in the source forms since they create indexicality and vagueness as important triggers of the constructs as EDMs. Alexander Andrason and Bernd Heine provide a compelling paper on ideophones, like English bang or thud. These are defined as interactive expressions used as vivid depictions of sensory imagery of states, events, objects, or qualities (cf. Dingemanse 2011, 2012, 2018; Dingemanse & Akita 2017; Andrason 2020, 2021). Ideophones are considered as a universal, typologically valid class of linguistic forms. While in European languages they may be limited, in others, such as Korean, Japanese and Basque, they may be as large as nouns and verbs (see Dingemanse 2018; Haiman 2018). Ideophones have been described as grammatical forms that are syntactically unattached and prosodically set off from surrounding text material. At the same time, they have also been treated as morphosyntactically integrated adverbials, verbals, or nominals in a number of languages. Andrason and Heine look at ideophones from the perspective of grammaticalization theory and provide a thought-provoking discussion about their ambivalent behaviour. An important contribution to dialogic and interactional approaches to grammaticalization is the one by Sylvie Hancil, as she discusses turn-taking as one of the key factors of innovative language use. In her paper, she looks at the rela-

Introduction

tionship between the emergence of grammatical patterns and intersubjectivity in sentence final position, in particular, she focuses on the change of final quoi in French. Final particles are a decisive category of linguistic expressions for the development of an interactional grammar. After reviewing the various meanings of final quoi in French dictionaries, she tests Hoppers’ (1991) five principles of grammaticalization, namely layering, divergence, specialization, persistence and decategorization and argues that the evolution of the particle cannot be restricted to a grammaticalization process. Rasmus Bernander, Antti Laine, Tim Roth and Lotta Aunio’s paper accounts for the grammaticalization of two auxiliary verb constructions (AVCs) based on the verb -maɾa ‘finish’ marking the two phasal polarity concepts translatable as already and not yet in the Western Serengeti language varieties of Mara Bantu. The two forms, despite being aspectually related seem occur in very different constructional frames and involve different historical backgrounds. The authors suggest that, although they share a joint paradigm or semantic space, their functions went through separate clines of change. These differences aside, both forms are analysed as being part of a type of recurrent linguistic cycle in Eastern Bantu. Verb forms that originally served to mark the concepts of with aspectual meanings similar to already and not yet developed other functional qualities, or even entered into oblivion. Finally, Charles Elerick’s study is focused on phonological erosion, a wellknown concomitant of grammaticalization. However, he suggests that accelerated reduction of phonological bulk is a more general diachronic phenomenon and can happen also in absence of grammaticalization per sé. Similarly, he notes that there are also instances of grammaticalization that do not involve significant phonological reduction. The latter is often offset by various mechanisms of rebulking. He argues that rightsizing is a key factor of rebulking phenomena. Right-sizing is an expression of Givón’s Quantity Principle (1990), which indicates that “less predictable information will be given more coding material” (Givón 1995: 49). Under the view of language as an adaptive and isostatic system, phonological loss interacts with mechanisms of compensation with or without grammaticalization. Elerick argues that phonological erosion that accompanies grammaticalization occurs as just one aspect of this diachronic process.

References Aarts, Bas. 2007. Syntactic Gradience: The Nature of Grammatical Indeterminacy. Oxford: OUP. Aijmer, Karin. 1997. I think – An English modal particle. In Modality in Germanic Languages: Historical and Comparative Perspectives, Toril Swan & Olaf Jansen-Westvik (eds), 1–47. Berlin: Mouton de Gruyter.

11

12

Vittorio Tantucci & Sylvie Hancil

Andersen, Henning. 2001. Actualization and the (uni)directionality of change. In Actualization: Linguistic Change in Progress [Current Issues in Linguistic Theory 219], Henning Andersen (ed.) 225–248. Amsterdam: John Benjamins. Auer, Peter & Günthner, Susanne. 2005. Die Entstehung von Diskursmarkern im Deutschen. Ein Fall von Grammatikalisierung? In Grammatikalisierung im Deutschen, Torsten Leuschner, Tanja Mortelmanns & Sarah De Groodt (eds), 335–362. Berlin: De Gruyter. Barth-Weingarten, Dagmar & Couper-Kuhlen, Elizabeth. 2002. On the final though: A case of grammaticalization? In Wischer & Diewald (eds), 345–361. Beuls, Katrien & Steels, Luc. 2013. Agent-based models of strategies for the emergence and evolution of grammatical agreement. PLoS ONE 8: e58960. Bisang, Walter, Himmelmann, Nikolaus P. & Wiemer, Björn (eds). 2009. What Makes Grammaticalization: A Look from its Fringes and its Components. Berlin: Mouton de Gruyter. Bopp, Franz. 1816. Über das Conjugationssystem der Sanskritsprache in Vergleichung mit jenem der griechischen, lateinischen, persischen und germanischen Sprache [Documenta Semiotica – Serie 1], reprint 1975. Frankfurt: Andreäsche Buchhandlung. Börjars, Kerstin & Vincent, Nigel. 2011. Grammaticalization and directionality. In Narrog & Heine (eds), 162–176. Boye, Kasper & Harder, Peter. 2012. A usage-based theory of grammatical status and grammaticalization. Language 88: 1–44. Breban, Tine & Kranich, Svenja (eds). 2015. What Happens after Grammaticalization? Secondary Grammaticalization and Other Late Stage Processes. Special issue of Language Sciences 47. Brinton, Laurel J. (2001). From matrix clause to pragmatic marker: The history of look-forms. Journal of Historical Pragmatics 2(2): 177–199. Brinton, Laurel J. 2017. The Evolution of Pragmatic Markers in English: Pathways of Change. Cambridge: CUP. Brinton, Laurel J. & Traugott, Elizabeth Closs. 2005. Lexicalization and Language Change. Cambridge: CUP. Bybee, Joan L. 2010. Language, Usage and Cognition. Cambridge: CUP. Bybee, Joan L., Perkins, Revere D. & Pagliuca, William. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago IL: The University of Chicago Press. Campbell, Lyle (ed.). 2001. Grammaticalization: A Critical Assessment. Special issue of Language Sciences 23(2–3). Coussé, Evie, Andersson, Peter & Olofsson, Joel (eds). 2018. Grammaticalization Meets Construction Grammar [Constructional Approaches to Grammar 21]. Amsterdam: John Benjamins. Croft, William. 2000. Explaining Language Change: An Evolutionary Approach. Harlow: Pearson Education. Croft, William. 2007. The origins of grammar in the verbalization of experience. Cognitive Linguistics 18(3): 339–382.

Introduction

Croft, William. 2010. The origins of grammaticalization in the verbalization of experience. Linguistics 48(1): 1–48. Davidse, Kristin, Breban, Tine, Brems, Lieselotte & Mortelmans, Tanja (eds). 2012. In Grammaticalization and Language Change: New Reflections [Studies in Language Companion Series 130]. Amsterdam: John Benjamins. Degand, Liesbeth & Evers-Vermeul, Jacqueline. 2015. Grammaticalization or pragmaticalization of discourse markers? More than a terminological issue. Journal of Historical Pragmatics 16(1): 59–85. Degand, Liesbeth & Simon-Vandenbergen, Anne-Marie. 2011. Introduction: Grammaticalization and (inter)subjectification of discourse markers. Linguistics 49(2): 287–294. De Smet, H. (2016). Entrenchment effects in language change. In Entrenchment and the psychology of language learning: How we reorganize and adapt linguistic knowledge (pp. 75–99). Detges, Ulrich & Waltereit, Richard. 2002. Grammaticalization vs. reanalysis: A semanticpragmatic account of functional change in grammar. Zeitschrift für Sprachwissenschaft 21: 151–95. Diessel, Holger. 2006. Demonstratives, joint attention, and the emergence of grammar. Cognitive Linguistics 17: 463-489. Diessel, Holger. 2011. Grammaticalization and language acquisition. In Narrog & Heine (eds), 130-141. Dik, Simon C. 1997. The Theory of Functional Grammar, Part 2: Complex and Derived Constructions [Functional Grammar Series 21]. Berlin: Mouton de Gruyter. Diewald, Gabriele. 2011a. Grammaticalization and pragmaticalization. In Narrog & Heine (eds), 450–461. Diewald, Gabriele. 2011b. Pragmaticalization (defined) as grammaticalization of discourse functions. Linguistics 49(2): 365–390. Ellis, Nick C. & Larsen-Freeman, Diane. 2006. Language emergence: Implications for applied linguistics – Introduction to the special issue. Applied linguistics 27(4): 558–589. Erman, Britt & Kotsinas, Ulla-Britt. 1993. Pragmaticalization: The case of ba and you know. Studier i Modern Sprakvetenskap 10: 76–92. Fanego, Teresa. 2010. Paths in the development of elaborative discourse markers: Evidence from Spanish. In Subjectification, Intersubjectification, and Grammaticalization, Kristin Davidse, Lieven Vandelanotte & Hubert Cuyckens (eds), 197–237. Berlin: Mouton de Gruyter. Fillmore, Carles J. 1988. The mechanisms of “construction grammar”. Annual Meeting of the Berkeley Linguistics Society 14: 35–55. Fischer, Olga. 2006. On the position of adjectives in Middle English. English Language and Linguistics 10(2): 253–288. Fischer, Olga. 2008. On analogy as the motivation for grammaticalization. Studies in Language 32(2): 336–382. Fischer, Olga, Rosenbach, Anette & Stein, Dieter (eds). 2000. Pathways of Change: Grammaticalization in English [Studies in Language Companion Series 53]. Amsterdam: John Benjamins.

13

14

Vittorio Tantucci & Sylvie Hancil

von der Gabelentz, Georg. 1891. Die Sprachwissenschaft. Ihre Aufgaben, Methoden und bisherigen Ergebnisse. Leipzig: Weigel Nachf. 2. überarbeitete Aufl.: Leipzig: Tauchnitz, 1901. van Gelderen, Elly. 2004. Grammaticalization as Economy [Linguistik Atuell/Linguistics Today 71]. Amsterdam: John Benjamins. Gisborne Nicholas & Patten, Amanda. 2011. Grammaticalization and construction grammar. In Narrog & Heine (eds), 92–104. Givón, Talmy. 1971. Historical syntax and synchronic morphology: An archaeologist’s field trip. Chicago Linguistic Society 7: 394–415. Givón, Talmy. 2009. The Genesis of Syntactic Complexity. Diachrony, Ontogeny, Neurocognition, Evolution. Amsterdam: John Benjamins. Goldberg, Adele E. 1995. Constructions: A Construction Grammar Approach to Argument Structure. Chicago IL: The University of Chicago Press. Goldberg, Adele E. 2006. Constructions at Work: The Nature of Generalization in Language. Oxford: OUP. Günthner, Susanne & Mutz, Katrin. 2004. Grammaticalization vs. pragmaticalization? The development of pragmatic markers in German and Italian. In Bisang, Himmelmann & Wiemer (eds), 77–107. Haiman, John. 1994. Ritualization and the development of language. Perspectives of Grammaticalization [Current Issues in Linguistic Theory 109], William Pagliuca (ed.) 3–28. Amsterdam: John Benjamins. Hancil, Sylvie & König, Ekkehard (eds). 2014. Grammaticalization. Theory and Data [Studies in Language Companion Series 162]. Amsterdam: John Benjamins. Hancil, Sylvie, Breban, Tine, & José Vicente Lozano (eds). 2018. New Trends in Grammaticalization and Language Change [Studies in Language Companion Series 202]. Amsterdam: John Benjamins. Harris, Alice C. & Campbell, Lyle. 1995. Historical Syntax in Cross-linguistic Perspective. Cambridge: CUP. Haselow, Alexander. 2014. Sequentiality in dialogue as a trigger for grammaticalization. In Hancil & König (eds), 203–204. Haselow, Alexander. 2015. Left vs. right periphery in grammaticalization. In Smith, Trousdale & Waltereit (eds), 157–186. Haselow, Alexander & Hancil, Sylvie. 2018. Rethinking language change from a dialogic perspective. Language Sciences 68: 1–5. Haspelmath, Martin. 1999. Why is grammaticalization irreversible? Linguistics 37: 1031–1068. Haspelmath, Martin. 2004. On directionality in language change with particular reference to grammaticalization. In Up and Down the Cline – The Nature of Grammaticalization [Typological Studies in Language 59], Olga Fischer, Muriel Norde & Harry Perridon (eds), 17–44. Amsterdam: John Benjamins. Hengeveld, Kees. 2011. The grammaticalization of tense and aspect. In Narrog & Heine (eds), 580–594. Heine, Bernd. 2002. On the role of context in grammaticalization. In Wischer & Diewald (ed.), 83–101.

Introduction

Heine, Bernd & Reh, Mechtild. 1984. Grammaticalization and Reanalysis in African Languages. Hamburg: Buske. Heine, Bernd & Kuteva, Tania. 2002. World Lexicon of Grammaticalization. Cambridge: CUP. Heine, Bernd & Reh, Mechthild. 1984. Grammaticalization and Reanalysis in African Languages. Hamburg: Buske. Heine, Bernd, Claudi, Ulrike & Hünnemeyer, Friederike. 1991. Grammaticalization: A Conceptual Framework. Chicago IL: The University of Chicago Press. Hopper, Paul J. 1991. On some principles of grammaticization. In Traugott & Heine (eds), Vol. 1, 17–35. Hopper, Paul J. & Traugott, Elizabeth Closs. 1993. Grammaticalization. Cambridge: CUP. Hopper, Paul J. & Traugott, Elizabeth Closs. 2003. Grammaticalization, 2nd edn. Cambridge: CUP. Keller, Rudi. 1994. On Language Change: The Invisible Hand in Language. London: Routledge. Lehmann, Christian. 1995[1982]. Thoughts on Grammaticalization. Munich: Lincom. First published in 1982 as Thoughts on Grammaticalization: A Programmatic Sketch [Arbeiten des Kölner Universalien-Projektes]. Cologne: University of Cologne. Lehmann, Christian. 2004. Theory and method in grammaticalization. Zeitschrift für Germanistische Linguistik 32(2): 152–187. Lenker, Ursula. 2000. Soplice and witodlice: Discourse markers in Old English. In Fischer, Rosenbach & Stein (eds), 229–249. López-Couso, María José & Seoane, Elena (eds). 2008. Rethinking Grammaticalization: New Perspectives [Typological Studies in Language 76]. Amsterdam: John Benjamins. Meillet, Antoine. 1912. L’évolution des formes grammaticales. Scientia (Rivista di Scienza) 12(26). Reprint: Meillet 1921:130–148, 6. von Mengden, Ferdinand & Simon, Horst (eds). 2014. What is it Then, This Grammaticalization? Special issue of Folia Linguistica 48(2). Mosegaard Hansen, Maj-Britt. 1998. The semantic status of discourse markers. Lingua 104(3–4): 235–260. Narrog, Heiko & Heine, Bernd. 2011. The Oxford Handbook of Grammaticalization. Oxford: OUP. Narrog, Heiko & Heine, Bernd (eds). 2018. Grammaticalization from a Typological Perspective. Oxford: OUP. Narrog, Heiko & Heine, Bernd. 2021. Grammaticalization. Oxford: OUP. Newmeyer, Frederick. 1998. Language Form and Language Function. Cambridge MA: The MIT Press. Nicolle, Steve. 2011. Pragmatic aspects of grammaticalization. In Narrog & Heine (eds), 401–412. Noël, Dirk. 2007. Diachronic construction grammar and grammaticalization theory. Functions of Language 14(2), 177–202. Norde, Muriel. 2009. Degrammaticalization. Oxford: OUP. Norde, Muriel, Beijering, Karen & Lenz, Alexandra (eds). 2013. Current Trends in Grammaticalization Research. Special issue of Language Sciences 36. Nuyts, Jan. 2012. Notions of (inter)subjectivity. English Text Construction 5(1): 53–76.

15

16

Vittorio Tantucci & Sylvie Hancil

Petré, Peter & Van de Velde, Freek. 2018. The real-time dynamics of the individual and the community in grammaticalization. Language 94(4), 867–901. Pons Bordería, Salvador & Loureda, Óscar (eds). (2018). Beyond Grammaticalization and Discourse Markers. Leiden: Brill. Prévost, Sophie. 2011. A propos from verbal complement to discourse marker: A case of grammaticalization? Linguistics 49(2): 391–413. Ramat, Paolo. 1992. Thoughts on degrammaticalization. Linguistics 30: 549–560. Ramat, Anna Giacalone & Hopper, Paul J. (eds). 1998. The Limits of Grammaticalization [Typological Studies in Language 37]. Amsterdam: John Benjamins. Ren, Wei. 2022. Effects of proficiency and gender on learners’ use of the pragmatic marker 吧 ba. Lingua 277, 103405. Roberts, Ian & Roussou, Anna. 2003. Syntactic Change. Cambridge: CUP. Sankoff, Gillian & Blondeau, Hélène. 2007. Language change across the lifespan: /r/ in Montreal French. Language 83: 560–588. Schwenter, Scott A. & Waltereit, Richard. 2010. Presupposition accommodation and language change. In Davidse, Vandelanotte & Cuyckens (eds), 75–102. Schmid, H. J. (2020). The dynamics of the linguistic system: Usage, conventionalization, and entrenchment. Oxford University Press. Seoane, Elena & María José López-Couso (eds). 2008. Theoretical and Empirical Issues in Grammaticalization [Typological Studies in Language 77]. Amsterdam: John Benjamins. Smith, Andrew D. M., Trousdale, Graeme & Waltereit, Richard (eds). 2015. New Directions in Grammaticalization Research [Studies in Language Companion Series 166]. Amsterdam: John Benjamins. Sommerer, Lotte & Smirnova, Elena (eds). 2020. Nodes and Networks in Diachronic Construction Grammar [Constructional Approaches to Grammar 27]. Amsterdam: John Benjamins. Stathi, Katerina, Gehweiler, Elke & König, Ekkehard (eds.). (2010). Grammaticalization: Current Views and Issues [Studies in Language Companion Series 119] Amsterdam: John Benjamins. Tabor, Whitney & Traugott, Elizabeth Closs. 1998. Structural scope expansion and grammaticalization. In Ramat & Hopper (eds), 229–272. Tagliamonte, Sali A. & D’Arcy, Alexandra. 2009. Peaks beyond phonology: Adolescence, incrementation, and language change. Language 85: 58–108. . Tantucci, V. (2015). Traversativity and grammaticalization: The aktionsart of 过 guo as a lexical source of evidentiality. Chinese Language and Discourse, 6(1), 57–100. Tantucci, Vittorio. 2017a. From immediate to extended intersubjectification: A gradient approach to intersubjective awareness and semasiological change. Language and Cognition 9(1): 88–120. Tantucci, Vittorio. 2017b. An evolutionary approach to semasiological change: Overt influence attempts through the development of the Mandarin 吧-ba particle. Journal of Pragmatics 120: 35–53. Tantucci, Vittorio. 2020. From co-actionality to extended intersubjectivity: Drawing on language change and ontogenetic development. Applied Linguistics 41(2): 185–214.

Introduction

Tantucci, Vittorio. 2021. Language and Social Minds: The Semantics and Pragmatics of Intersubjectivity. Cambridge: CUP. Tantucci, Vittorio, Culpeper, Jonathan & Di Cristofaro, Matteo. 2018. Dynamic resonance and social reciprocity in language change: The case of Good morrow. Language Sciences 68: 6–21. Tantucci, V., & Di Cristofaro, M. (2020). Entrenchment inhibition: Constructional change and repetitive behaviour can be in competition with large-scale “recompositional” creativity. Corpus Linguistics and Linguistic Theory, 16(3), 547–579. Tantucci, Vittorio & Di Cristofaro, Matteo. 2021. Pre-emptive interaction in language change and ontogeny: The case of [there is no NP]. Corpus Linguistics and Linguistic Theory, 17(3): 715–742. Tantucci, Vittorio & Wang, Aiqing. 2018. Illocutional concurrences: The case of evaluative speech acts and face-work in spoken Mandarin and American English. Journal of Pragmatics 138: 60–76. Tantucci, Vittorio & Wang, Aiqing. 2020a. From co-actions to intersubjectivity throughout Chinese ontogeny: A usage-based analysis of knowledge ascription and expected agreement. Journal of Pragmatics 167: 98–115. Tantucci, Vittorio & Wang, Aiqing. 2020b. Diachronic change of rapport orientation and sentence-periphery in Mandarin. Discourse Studies 22(2): 146–173. Tantucci, Vittorio & Wang, Aiqing. 2022. Resonance as an applied predictor of cross-cultural interaction: Constructional priming in Mandarin and American English interaction. Applied Linguistics, 43(1): 115–146. Terkourafi, Marina. 2011. The pragmatic variable: Toward a procedural interpretation. Language in Society 40(3): 343–372. Traugott, Elizabeth Closs. 1995. Subjectification in grammaticalization. In Subjectivity and Subjectivisation, Dieter Stein & Susan Wright (eds), 31–54. Cambridge: CUP. Traugott, Elizabeth Closs. 2003. From subjectification to intersubjectification. In Motives for Language Change, Raymond Hickey (ed.) 124–139. Cambridge: CUP. Traugott, Elizabeth Closs. 2007. Discussion article: Discourse markers, modal particles, and contrastive analysis, synchronic and diachronic. Catalan Journal of Linguistics 6: 139–157. Traugott, Elizabeth Closs. 2022. Discourse Structuring Markers in English: A Historical Constructionalist Perspective on Pragmatics [Constructional Approaches to Language 33]. Amsterdam: John Benjamins. Traugott, Elizabeth Closs & Dasher, Richard B. 2002. Regularity in Semantic Change. Cambridge: CUP. Traugott, Elizabeth Closs & Heine, Bernd (eds). 1991a. Approaches to Grammaticalization, Vol. 1: Theoretical and Methodological Issues [Typological Studies in Language 19:1]. Amsterdam: John Benjamins. Traugott, Elizabeth Closs & Heine, Bernd (eds). 1991b. Approaches to Grammaticalization, Vol. 2: Types of Grammatical Markers [Typological Studies in Language 19:2]. Amsterdam: John Benjamins. Traugott, Elizabeth Closs & Trousdale, Graeme (eds). (2010). Gradience, Gradualness and Grammaticalization [Typological Studies in Language 90]. Amsterdam: John Benjamins. Traugott, Elizabeth Closs & Trousdale, Graeme. 2013. Constructionalization and Constructional Change. Oxford: OUP.

17

18

Vittorio Tantucci & Sylvie Hancil

Van Bogaert, Julie. 2011. I think and other complement-taking mental predicates: A case of and for constructional grammaticalization. Linguistics 49(2): 295–332. Van linden, An, Verstraete, Jean-Christophe & Davidse, Kristin (eds). 2010. Formal Evidence in Grammaticalization Research [Typological Studies in Language 94]. Amsterdam: John Benjamins. Van Olmen, Daniel, Cuyckens, Hubert & Ghesquière, Lobke (eds). 2016. Aspects of Grammaticalization:(Inter)Subjectification and Directionality. Berlin. De Gruyter. Verhagen, Arie. 2005. Constructions of Intersubjectivity: Discourse, Syntax and Cognition. Oxford: OUP. Wischer, Ilse. 2000. Grammaticalization versus lexicalization: Methinks there is some confusion. In Fischer, Rosenbach & Stein (eds), 355–370. Amsterdam: John Benjamins. Wischer, Ilse & Diewald, Gabriele (eds). 2002. New Reflections on Grammaticalization [Typological Studies in Language 49]. Amsterdam: John Benjamins. Wu, Fu Xiang. 2005. “Hanyu tibiaoji “le” “zhe” weishenme bu neng qianzhixing shiyong (Why the aspectual markers “le” and “zhe” are not obligatory). Contemporary Linguistics 3. Wüllner, Franz. 1831. Über Ursprung und Urbedeutung der Sprachlichen Formen. Münster. Ziegeler, Debra. 1997. Retention in ontogenetic and diachronic grammaticalization. Cognitive Linguistics 8: 207–241.

section i

Diachronic approaches

chapter 1

From comparative standard marker to comparative adverb On the contact-induced (de)grammaticalization of yori in modern through present-day Japanese Reijirou Shibasaki Meiji University

This study investigates the grammaticalization of the Japanese ablative marker yori ‘from’ into a comparative standard marker ‘than’ and further into a comparative adverb yori ‘more.’ It is widely known that ablative markers can be grammaticalized into comparative standard markers. However, little is known about the direction of change from comparative marker to comparative adverb. This study takes a corpus-based approach to uncover the process of change. The ablative yori, which had been used since Old Japanese, began to take on the adverbial property ‘more’ in the early twentieth century. While some collocational sequences facilitate the grammaticalization of yori, the overall degree of change turns out to be very slow. Keywords: ablative markers, comparative standard markers, comparative adverbs, extravagance, language contact in Japanese, (de)grammaticalization, debonding, persistence

1.

Introduction

This study investigates the process of grammaticalization of the Japanese ablative marker yori ‘from’ into a comparative standard marker ‘than’ and the subsequent degrammaticalization into a comparative adverb yori ‘more.’ The following are examples from Present-Day Japanese (PDJ, hereafter). The corpora used for this study are introduced in Section 4. Note that examples from Showa Speech Corpus (SSC), Corpus of Spontaneous Japanese (CSJ) and Balanced Corpus of Contemporary Written Japanese (BCCWJ) are presented with its own sample ID number before the name of the corpus. From here onward elements in focus are underlined and yori is glossed according to the context in which it appears. https://doi.org/10.1075/slcs.232.01shi © 2023 John Benjamins Publishing Company

Chapter 1. From comparative standard marker to comparative adverb

(1) Ablative usage Kyoo-wa takoo-yori ki-te-morat.ta today-top another.school-from come-and-receive.perf Asura-kun-to katari-at-te-mitai to-omoi-masu. personal.name-Mr.-with talk-meet-and-want quot-think-pol ‘Today, (we) would like to talk with Asura (coming) from another school.’ (2002 Asurahiko no aoitori; PB29_00541; BCCWJ) (2) Comparative standard marker usage Kyoo-no o.tenki-wa kinoo-yori yoi-yoo.desu-ne. today-gen pol.weather-top yesterday-than good-appearance.cop (pol)-fp ‘(It) seems (that) Today’s weather is better than Yesterday’s.’ (2008 Yahoo Blog; OY13_06040; BCCWJ) (3) Comparative adverb usage Korera-no an-ga yori yoi mono to-naru-yoo, these-gen proposal-nom more good thing pt-become-purp mina-san-no iken-o boshuu.shi-masu. everyone-Mr./Ms.-gen opinion-acc application.do-cop.pol ‘(In order) to make these proposals better, (we) invite (public) opinions (about these).’ (2008 A PR magazine; OP75_00001; BCCWJ)

In (1), yori follows takoo ‘another school’ to indicate a source of information in a typical usage of an ablative marker; in this usage, yori often occurs with motion verbs such as iku ‘to go’ and kuru ‘to come.’ In (2), yori serves as a comparative standard marker with its preceding word kinoo ‘yesterday.’ In this usage, the speaker’s or writer’s evaluation is reflected in the predicate, as in yoi ‘be good.’ In (3), yori is used as a comparative adverb that modifies the following yoi ‘be good.’ Note that in (1) and (2), yori is attached as a case-marker onto the preceding words, takoo ‘another school’ and kinoo ‘yesterday,’ while in (3), yori is used as an independent word to modify the following adjective yoi ‘be good.’ That is, yori exhibits a layering of these three usages in PDJ. Further, both ablative and adverbial yori can be used together, as shown in (4). (4) Ryuudoo.hiritsu-yori yori kibishiku mita hojo.shihyoo fluidity.ratio-than more strict see.adn subsidiary.index to-i-eru-de-shoo. quot-say-can-cop-infr ‘(It) can be said that (it) is (a kind of ) stricter additional index than liquidity ratio.’ (2004 Hontoo ni dekiru shikin chootatsu; PB43_00708; BCCWJ)

The following example shows that the adverbial yori forms a compound with another adverbial issoo ‘further.’

21

22

Reijirou Shibasaki

(5) shienkin-yori yori issoo juuyoo-na-no-wa financial.backing-than more further important-cop.adn-nml-top senryaku.teki-shien-da. strategic-support-cop.npst ‘What is even more important than financial backing is (to provide) strategic support.’ (2008 Yahoo Blog; OY04_05267; BCCWJ)

Arguably, Hida and Asada (2018: 587–588) make an entry for yori-issoo ‘furthermore’ in their list of adverbial expressions in PDJ, which may lend support to the view that yori has been adverbialized enough in PDJ to be compounded with other elements (see Section 5.3 for related discussions).1 It is widely known that ablative markers can be grammaticalized into comparative standard markers (e.g., Stassen 1985: 39–45; Kuteva et al. 2019: 36–37); however, little is known about the direction of change from comparative standard marker to comparative adverb even in the framework of degrammaticalization. The whole issue is thus central to the current study. This paper is structured as follows. In Section 2, the history of yori is briefly explained. In Section 3, some preceding research closely related to the current study is introduced. Section 4 introduces the corpora used for this study. Section 5 provides the results of corpus-based surveys, each of which is discussed in terms of grammaticalization and degrammaticalization in connection with language contact. Section 6 is the conclusion of the study.

2.

A short history of yori

The element in focus here, i.e., yori, has been used since Old Japanese (OJ, hereafter). Nakata et al. (1983: 1720) summarize the functions of yori in OJ as follows. Translations are all by the current author, with the addition of possible corresponding expressions in English. Note that in as early as OJ, the ablative yori had already been polyfunctional in serving as a comparative standard marker as in (6e). See Stassen (1985) and Kuteva et al. (2019) mentioned above. (6) yori (ablative) a. a spatio-temporal starting point for activity, from b. a position along the way, by way of, through c. denoting an action, by means of, by d. denoting a range of action, from … to/until… e. denoting a comparative standard marker, than (Nakata et al. 1983: 1720) 1. In addition to these usages, yori can be used as an (almost) lexicalized expression yori-hokanai (from-other-not) ‘(there is) no choice (but to do something).’ See Hida & Asada (2018: 587) for other lexicalized or collocational expressions.

Chapter 1. From comparative standard marker to comparative adverb

Nakata et al. (1983: 1720–1721) go on to explain that in OJ, there were three other expressions synonymous with yori: yo, yuri, and yu. While potential derivational processes (e.g., from yo to yori and from yu to yuri) or possible directions of change (e.g., from yori to yo and from yuri to yu) are still open to debate, the fact that only yo and yori are found to have served as comparative standard markers (p.1721) is worth mentioning. However, even these two forms were extremely rare in OJ. On the other hand, the other ablative marker, kara ‘from,’ began to gain an increasing share of the usages in (6a) and (6b) from OJ onward, while the usage in (6c) has become fossilized in a limited set of expressions, such as kachi-yori ‘on foot (lit. ‘by foot’),’ according to Nihon Kokugo Daijiten (2006, Vol.13: 704; Nikkoku, hereafter). As a result, yori began to be used exclusively as a comparative standard marker from that time onward (ibid.).2 The comparative adverbial yori is considered to have appeared around the turn of the twentieth century. The earliest example cited in Nikkoku (2006) is shown in (7), which means that yori has developed its adverbial usage over the twentieth century. (7) Meiji-no shichoo-ni yori shokuchaku-su-beki-hazu-no Meiji.period-gen trend-to more direct.contact-do-should-supposed.to-gen chuutoo-shakai-o utsusu-nimo nao sukoburu sohon-de-aru. middle-class-acc copy-even.if still deeply crude-cop-exist ‘Even if (it) describes (people of ) the middle class who are supposed to be more trend-conscious during the Meiji period, (it) is still rough around the edges.’ (1899 Shakuhyoo Reigo; Nikkoku Vol.13, s.v. yori adv.)

Shibasaki (in press) examined five dictionaries of Japanese published from 1899 to 1912, and found that one dictionary, called Daijiten Ge (An unabridged dictionary, Vol.2, 1912), offers an explanation for the potential adverbial usage of yori in which yori is interpreted as corresponding to than or ‑er in English, originating in the translation of English into Japanese. Although the dictionary does not provide any crucial or actual example, it seems that in the early twentieth century, some speakers of Japanese might have realized the functional expansion of yori from a comparative standard marker (i.e., than) to a comparative adverb (i.e., more).3

2. Nikkoku (2006, Vol.13: 704) adds a note that among the four synonymous expressions, i.e., yori, yo, yuri, and yu, yori was by far the most frequently used form with a wider range of functions, as in (6). 3. Sansom (1928: 252–254) shows some examples with yori of those days; however, he makes no mention of the adverbial yori. Presumably, the adverbial yori was so infrequent in the early twentieth century that it might not have fallen under researchers’ notice.

23

24

Reijirou Shibasaki

3.

Preceding studies on yori and language contact

In this section, I refer mainly to the following two studies published in Japan, i.e., Hida (1992, 2019) and Morioka (1999), each of which is directly related to the present research (Section 3.1). After that, I briefly introduce some influential studies on language contact issues in a cross-linguistic context (Section 3.2).

3.1 Preceding studies on yori 3.1.1 Hida (1992, 2019) Hida (1992, 2019) examines a variety of English reference grammar books and their translated versions published in the early Meiji period and arrives at the conclusion that the adverbial yori was derived from the translation of English into Japanese. In what follows, I give an account of how Hida (1992) deciphers such a mysterious process of change from a comparative standard marker to a comparative adverb. The illustration in (8) is as faithful as possible to Hida’s (1992) illustration of the example. In (8), the line just above the original example stands for either glossing or grammatical relations in the original grammar books, while the line below the example gives a supplementary explanation, e.g., 人名 ‘personal name.’ The corresponding (9) is an English translation of the original (8). In (10), the English elements in focus, i.e., the comparative expressions, are relocated following a typical word order of Japanese. (8) ガ アル 長ケ高ク ガ アル ヨリ高ク 然シナガラ ガ アル 最モ高ク John is tall, James is taller, but George is the tallest. 人名 人名 人名 (Hida 1992:475) (9) nom be tall nom be yori.tall but nom be John is tall, James is taller, but George is personal.name personal.name personal.name (the).most.tall the tallest. ‘John is tall, James is taller, but George is the tallest.’ (10) (based on Hida 1992: 475)

Chapter 1. From comparative standard marker to comparative adverb

In the verbatim translation, yori seems to be used to indicate the comparative inflection of the adjective in (8), where yori does not carry the potential standard, John. In fact, relevant examples from English reference grammars cited in Morioka (1999) and Yagishita (2018) are similar in this respect. That is, yori is illustrated as a comparative adverb. Hida (2019: 191) thus argues that language learners in the Meiji period (e.g., college-bound youngsters and intellectuals) memorized these examples and their parallel translations faithfully and keenly, resulting in the emergence of the comparative adverb yori.4 Contact with western languages, especially English, is considered to have acted as a trigger for the reanalysis of the ablative comparative standard marker to the comparative adverb.5 Considering the fact that Japanese is a verb-final language, the reanalysis of X-yori (‘than X’) as yori Y (‘more Y (than X)’), as Hida (1992) argues, seems to be a man-made change caused by language contact. In more contemporary Japanese, yori can indisputably be used as a comparative adverb, but it seems to still be regarded as a comparative standard marker in typological studies. For example, Payne (1997: 89) refers to the following Japanese example in order to show the sequential relation of standard, marker and quality in a comparative construction. Note that one typographical error in the original has been corrected by the present author (i.e., meko > neko). (11) Japanese: standard-marker-quality Inu ga neko yori ookii. dog nom cat than big std mkr qual ‘The dog is bigger than the cat.’

(Payne 1997: 89)

4. Hida (2019: 191) notes that before the emergence of the comparative adverbial yori, Japanese had a bipartite system of positive and superlative forms. As one anonymous reviewer points out, there are some degree (but not comparative) adverbs such as sarani ‘additionally’, issoo ‘all the more’, nafo ‘still, more’, and harukani ‘(by) far’, some of which were used before the Meiji period. I am grateful to the reviewer for this point. Hida (2019) does not mention them; however, what he means there, presumably, is that those adverbs have not been used as comparative adverbs as shown in (11). In fact, the Corpus of Historical Japanese (CHJ) does not provide any crucial examples before the Meiji period in which those adverbs were used in the comparative construction. I will take this issue for my future study. 5. It should be mentioned that Sugimoto (1983: 232–242) notes the Dutch influence on the emergence of the comparative adverb yori in the Edo period, e.g., kleiner (klein-er [smallCOMP]) ‘smaller’ > yori chiisai (er-klein] ‘smaller.’ Worth noting here is that both Hida (1992, 2019) and Sugimoto (1983) attribute a significant part of the rise of the comparative yori to language contact, specifically translation of western languages, either English or Dutch, into Japanese. Also see Shibasaki (2021) on this issue.

25

26

Reijirou Shibasaki

In (11), yori is used as a comparative marker with its standard, neko ‘cat’: this yori is an ablative case-marker. The subject of the clause, inu ‘dog,’ is compared against the standard, neko ‘cat,’ in terms of the quality ookii ‘be big.’ On the other hand, if neko ‘cat’ is informationally active in the immediate discourse, it is omissible, as in (12): yori is automatically interpreted as a comparative adverb, because it is now bifunctional in the comparative construction. (12) Inu ga yori ookii. dog nom more big com.adv qual ‘The dog is bigger.’

3.1.2 Morioka (1999) Morioka (1999) pays attention to the following points: (1) a widely used English reader of those days, i.e., New National Fourth and Fifth Reader (1886–1907), and its Japanese translations (Reader translation); (2) wording in Japanese literature in both the Meiji and the Taisho periods (Japanese own wording), and (3) Japanese translation works from western languages (Literature translation). His observation is that yori served as a comparative adverb especially when translators added other adverbs such as issoo ‘further’ and sarani ‘moreover’ to lay emphasis on the comparative meaning of yori. In other words, these sequential forms are considered to be collocational, although he makes no reference to such technical terms. Here are some examples from Reader translation. (13) Reader translation a. Original: it was noisier than ever. (New National Fourth and Fifth Reader 1886–1907) b. Trans 1: sore-ga katsute-yori nao sawagashiku-arishi. that-nom formerly-than more be.noisy-exist.pst (Trans by Yutaka Shimada in the 1888 edition) c. Trans 2: sore-wa mae-yori-wa issoo sawagashiku-ariki. that-top once-than-top further be.noisy-exist.pst (Trans by Frederick W. Eastlake and Seiichiro Mayama in the 1902 edition) (Morioka 1999: 199)

In (13), the original English expression it was noisier than ever is translated into Japanese in two ways, as in (13b) and (13c). In both translations, yori remains the ablative, as connected by a dash to the preceding comparative standards, although it occurs with the following adverbs to emphasize the whole comparative meanings. Morioka (1999: 200) thus states that these formulaic sequences were created

Chapter 1. From comparative standard marker to comparative adverb

when intellectuals in these periods were sharply conscious of and struggled with translations of the comparative constructions in western languages (see also 6), gradually giving rise to the comparative adverb yori, as in (14). (14) Reader translation a. Original: The sun rose higher and higher,… (New National Fourth and Fifth Reader 1886–1907) b. Trans 1: Hi-ga masu-masu takaku nobori-shi… sun-nom more-more high(ly) rise-pst (Trans by Yutaka Shimada in the 1888 edition) c. Trans 2: Taiyoo-ga dan-dan takaku nobori-shi sun-nom step-step high(ly) rise-pst (Trans by Masao Wada in the 1900 edition) d. Trans 3: Taiyoo-wa yori takaku yori takaku nobori-ki sun-nom more high(ly) more high(ly) rise-pst (Trans by Frederick W. Eastlake and Seiichiro Mayama in the 1902 edition) (Morioka 1999: 200)

In (14), the comparative expression, i.e., higher and higher, is translated differently by using masu-masu ‘more and more,’ dan-dan ‘step by step’ and yori ‘more.’ It is true that these adverbs are subtly different in meaning from each other; however, these translations mirror the emergence of comparative adverbial forms including yori in these periods. It is important to note here that the comparative adverb yori was still in the early stages of its grammaticalization from the ablative case-marker (comparative standard marker) to the comparative adverb. Because of this, Morioka (1999) seems to include examples of both ablative and comparative adverbial usages, which succeeds in reflecting a layering of yori in these synchronic stages. Typically, linguistic elements are grammaticalized not without constraint but in particular sequential relationships in discourse (Traugott 2003: 624). In this respect, the sequential expressions with the comparative standard marker yori pointed out by Morioka (1999) are worthy of consideration. See Table 1. 6. It should be mentioned that Sugimoto (1983: 232–242) notes the Dutch influence on the emergence of the comparative adverb yori in the Edo period, e.g., kleiner (klein-er [smallCOMP]) ‘smaller’ > yori chiisai (er-klein] ‘smaller.’ Worth noting here is that both Hida (1992, 2019) and Sugimoto (1983) attribute a significant part of the rise of the comparative yori to language contact, specifically translation of western languages, either English or Dutch, into Japanese. Also see Shibasaki (2021) on this issue.

27

28

Reijirou Shibasaki

Table 1. Collocational patterns with the comparative marker yori (based on Morioka 1999) Comp standard X plus comp marker yori

Particle

Adverb 1

X yori ‘than X’

Ø

nao ‘more’

Adverb 2

Adverb 3

issoo ‘further’

ito ‘very’

masu-masu ‘more and more’ yohodo ‘exceptionally’ issoo ‘further’ mata ‘besides’ wa (top)

issoo ‘further’ ooku ‘much’ sarani ‘moreover’

mo (emph)

hito-shio ‘even more’

nao ‘more’ hito-shio ‘even more’ ichi-bai ‘extra’ (lit. one-double) haruka-ni ‘far’ (ni = adverbial particle) sarani ‘moreover’ issoo-ni ‘further’ hanahadashiku ‘exceedingly’

It is surprising that there were/are such a large array of collocational sequences with yori, although some of these sequential forms are rarely observed in PDJ, e.g., X yori mo hito-shio (yasashii) ‘even much (kind)er than X’ and X yori mata issoo ito (kurusii) ‘even much further more (difficult) than X.’ On the other hand, the types of the comparative adverbial usages in Morioka (1999) are only those in (14). Presumably, as yori has been adverbialized, it has become semantically general enough to take the place of many of the adverbial expressions given in Table 1, which has not (just) been bleached in meaning but may have served to cover a wide range of meanings that used to be expressed by other sequential forms in Table 1, depending on the given context (see Bisang (2020) and Xing (2015, 2020) for the proposal of ‘semantic accretion’ in Mainland Asian languages). To sum up, Hida (1992, 2019), Morioka (1999) and the others mentioned elsewhere in this section all make a convincing case for understanding one type of contact-induced grammaticalization. Specifically, what they uncovered on the initial step in this particular grammaticalization deserves mention. Building on their

Chapter 1. From comparative standard marker to comparative adverb

pioneering works, this study takes a larger view from which to investigate how and to what extent yori has been grammaticalized over the twentieth century to the present, which I believe can provide a supplementary explanation to their research.

3.2 Language contact research on Japanese Language contact issues have drawn renewed attention from a wider range of researchers. For example, one can readily witness such huge thematic volumes on language contact as Adamou and Matras (2020), Grant (2020), and Hickey (2020), the last of which, by a curious coincidence, was revised and published in the same year. Further, some earlier works should also be addressed. Weinreich (1953: 42) introduces the idea of the Indo-Europeanization of Hebrew, showing the emerging preferred form of a new possessive (i.e., ha-bait šel-xa ‘the-house ofyou’) over the erstwhile possessive suffixation (i.e., bet-xa ‘house-your’). Muysken (2000: 139) reports that noun phrases are most likely to be borrowed through language contact, illustrating some Dutch elements found in Sranan, the Englishbased creole of Surinam on the northeastern coast of South America, while Clyne (2003: 76–79) classifies transferrable elements in language contact, e.g., lexical transference, phonological transference, prosodic transference, semantic transference, etc. between English and its immigrant languages. Despite the increasing trend in language contact research, reference to Japanese seems to be limited. For example, Thomason and Kaufman (1988: 32, 79–80) just touch on the adaptation of Chinese and English loanwords in Japanese, and Hickey (2010: 173–174) introduces a handful of loanwords, such as sweater in English to se(e)ta(a) in Japanese. Although Heine and Kuteva (2005) pay attention to either intensive or extensive contact between languages, Japanese is not included in their index. While there are other excellent studies, whether modest or large in scale, on language contact in Japan (e.g., Ishiwata (1986) and Loveday (1996) from a sociolinguistic perspective, Yanabu (2004) and the studies addressed in Section 3.1 on the translation effect of western languages on the creation of Modern Japanese expressions), any in-depth study of translation effects on the renewal of former linguistic elements from a (de)grammaticalization perspective has not been reported. This study thus takes a corpus-based quantitative approach to give a qualitative survey of the (de)grammaticalization of yori.

29

30

Reijirou Shibasaki

4.

Corpora

The corpora used for this study are as follows, summarized in (15). As shown in (7) in Section 2, Nikkoku (2006) provides the earliest example of the comparative adverbial yori from the very end of the nineteenth century. Therefore, I chose the Taiyo Corpus (Taiyo, hereafter) to examine the degree of gradualness of this particular functional expansion. In addition, I made the Aozora Corpus (Aozora, hereafter) because the 1930s–1940s are not sufficiently covered by existing corpora and these synchronic stages can be thought to provide examples for understanding some crucial directions of change.7 The Showa Speech Corpus (SSC, hereafter; see Maruyama 2020) and the Corpus of Spontaneous Japanese (CSJ, hereafter) are spoken-based corpora that cover the second half of the twentieth century. The Balanced Corpus of Contemporary Written Japanese (BCCWJ) is used to give supplemental surveys in Sections 5.2.2 and 5.3 as well as to collect relevant examples as shown in Section 1. (15) a. b. c. d. e.

5.

Taiyo Corpus (Taiyo, 1895, 1901, 1909, 1917, 1925) Aozora Corpus (Aozora, 1930s–1940s) Showa Speech Corpus (SSC, 1950s–1970s) Corpus of Spontaneous Japanese (CSJ, 2000s) Balanced Corpus of Contemporary Written Japanese (BCCWJ, 1976–2008)

Survey results and discussion

To begin with, I explain some coding properties in the use of the corpora introduced in Section 4. From all the examples of yori in Taiyo and CSJ, I randomly sampled 500 examples of each and manually removed irrelevant examples, such as tayori ‘news, hearing from someone,’ and niyori ‘by means of.’ The examples of yori in SSC turned out not to be evenly distributed across the periods; therefore, I divided the examples retrieved into those in the 1950s and those in the 1960s–1970s. CSJ is used to figure out the current changing situation of yori in the 2000s, which is expected to show the gradual and continual change in the function of yori in these spoken registers. For the sake of explanation, I will use the term grammaticalization until Section 5.4 instead of degrammaticalization, which is discussed in Section 5.5. 7. The Aozora corpus consists of approximately 772, 341 letters. The following are the texts that constitute the self-made corpus: Goshiki Onsen Sukī Nikki (1930), Joryūhaiku o Midokusu (1932), Hiruberuto Hōmonki (1932), Ginga tetsudō no Yoru (1934), Shumitoshiteno Dokusho (1935), Yukiguni (1937), Kaze Tachinu (1938), Robō no Ishi (1940), Sangetsuki (1941), Yakeato no Iesu (1946), Binbō Monogatari (1947), and Daigaku to Sono Sōchō (1948).

Chapter 1. From comparative standard marker to comparative adverb

5.1 Gradualness of grammaticalization Table 2 shows a broad overview of the functional expansion of yori from ablative case-marker to comparative adverb. Numbers represent raw frequencies in each stage, with the addition of percentages based on relative frequency. For example, 88 (96.7%) and 3 (3.3%) in the 1930s column mean that the ablative yori shares 96.7% of all the occurrences, while the comparative adverb shares 3.3%. Table 2. The pace of change in the grammaticalization of yori: An overview Taiyo 1895–1909 1917–1925

Aozora 1930s

1940s

SSC

CSJ

1950s

1960s–1970s

2000s

yori (ablative)

390 (100%)

105 (97.2%)

88 92 (96.7%) (96.8%)

181 (96.8%)

73 (89.0%)

316 (82.1%)

yori (adverb)

0 (0)

3 (2.8%)

3 (3.3%)

6 (3.2%)

9 (11.0%)

69 (17.9%)

3 (3.2%)

Overall, the speed of progress of this particular grammaticalization appears to be slow. While yori has been adverbialized almost steadily over the twentieth century, only about 18% of the total occurrences are used as adverbs even in the 2000s. A slow change in the expansion of yori can lend support to the well-known hypothesis characteristic of grammaticalization, i.e., gradualness. In the last hundred years or so, yori seems to have just started the functional expansion from the ablative to the comparative adverb. On the other hand, we may need to remember what was addressed in Section 3.1 as to the slow pace of change, namely, that this direction of change might not be a natural way of change. In fact, Stassen (1985), Heine (1997) and Kuteva et al. (2019) make no mention of the potential change from comparative standard marker to comparative adverb, which will be discussed in Section 5.5. On the other hand, if we narrow our focus to some particular sequential patterns with yori, the degree of variation may be different. In fact, Himmelmann (2004: 31) clearly mentions that where grammaticalization is witnessed it is “constructions and not isolated lexical items” (see Traugott (2003: 624) addressed in Section 3.1.2), which corresponds to the collocational sequences discussed in this study. Languages do not change in a similar manner at the same rate; however, if a given change can be quantified in a phased manner, the result would provide insights into other phenomena (see Section 5.2 below).8 8. The whole collocational sequences found in spoken Japanese are summarized in the Appendix, which is based on Shibasaki (in press). The observations discussed in Sections 5.2 and 5.3 are also built on it. Note that the two sequential forms discussed in Section 5.2 cannot

31

32

Reijirou Shibasaki

5.2 Collocational sequences that preserve the earlier ablative function of yori In Section 5.1, we have confirmed the gradual process of the grammaticalization of yori. The major sequential forms are summarized in the Appendix, but this does not necessarily provide such a detailed picture of the survey results as to allow us to fully understand what characterizes the progress of total grammaticalization. In this section, I thus illustrate the following two collocational sequences that are found to reflect ‘persistence’ (Hopper 1991) in grammaticalization. 5.2.1 Yori-collocations with particles Table 1 summarizes collocational patterns in which yori is used as its comparative (ablative case) marker, based on Morioka (1999), in which the three types of particles, i.e., Ø, wa, mo, are found to occur with the preceding yori-phrases to lay emphasis on the comparative meanings. What the survey results tell us about this construction is that once yori-expressions are accompanied by tangible particles, they always serve as a comparative standard marker, as in (16) and (17), i.e., ‘persistence’ (Hopper 1991). (16) …keishiki keishiki tte-iu-kedo-ne keishiki-yori-mo naiyoo-ja-nai-no…   form form quot-say-but-pt form-than-pt content-cop-neg-qp ‘… (people often) say that formality (is important), but what is said is (more important) than (the way people say), right?’ (1955 Kumiai Dankoo 2; C55_02_MT; SSC) (17) Ee saitekina parameetaa-yori-wa kanari ookii atai-ni um optimal parameter-than-top fairly big value-pt tot-te-ori-masu take-and-exist.hum-pol ‘Um, (we) set a much higher value (for this) than (for) optimization parameter.’ (2000s Male 30–34 yrs; A01M0264; CSJ)

On the other hand, yori-expressions with the intangible form, i.e., Ø, do not always serve as comparative standard markers; see (2) and (3) in Section 1. However, what is important is that the adverbialized yori does not occur with any particles, either tangible or intangible, according to the survey results. In a nutshell, the comparative adverb yori has been grammaticalized as a fixed form.

be considered as cases of lexicalization, because these are not completely unified lexical units that combine two words, while the sequential form yori issoo discussed in Section 5.3 can be a case of lexicalization because it is lemmatized at least in Hida and Asada (2018: 587–588).

Chapter 1. From comparative standard marker to comparative adverb

In addition, yori can be attached not only to phrases but to clauses, as in (18) and (19). In (18), the verb aruku ‘to walk’ takes the adnominal form that can be accompanied directly by case-markers such as yori, while in (19), the verb kodawaru ‘to cling to’ also takes the adnominal form but is nominalized by the following koto ‘thing’; the whole clause is used as the comparative standard. (18) Numa.zoi-o aruku-yori kanuu-o tsukat-ta-hoo-ga hayaku pond.bank-acc walk-than canoe-acc use-perf-choice-nom early mokutekichi-ni toochaku.suru-koto-ga-dekiru destination-to arrival.do-nml-nom-can ‘(We) can arrive at the destination earlier by using a canoe than by walking on the banks of a pond.’ (2000s Male 55–59 yrs; S07M1301; CSJ) (19) Ee jibun-no shuchoo-ni kodawaru-koto-yori-mo um self-gen assertion-pt cling-nml-than-pt gooi-no-keisei-o yuusen.suru-koto-ga ee nozomashii consensus.gen-formation-acc preference.do-nml-nom um desirable to-s.are-te-i.masu quot-do.pass-and-exist.pol ‘Um, it is desirable that (one should) take precedence of forming a consensus, um, over clinging to their claim.’ (2000s Male 25–29 yrs; A04M0730; CSJ)

Further, the predominant use of the ablative function appears to have created a variety of formulaic expressions, as partially listed in the Appendix, for example, nani-yori (mo) ‘above all (things),’ moto-yori ‘from the beginning, of course,’ dareyori ‘more than anyone,’ sore-yori ‘what’s more important, instead [lit. ‘from/than that’],’ to name just a few examples. The following is another example of a collocational and colloquial use of yori, which serves as a discourse marker, usually in sentence/utterance-initial position. (20) Maa jugyoo-ni shuuchuu.shi-te-i.mashi-ta-kara-ne. well class-pt concentration.do-and-exist.pol-pst-because-fp Tte-iu-yori-wa Eigo-to-puroguramingu-ga quot-say-than-top English-and-programming-nom idoo-na-nde-ne movement-cop.adn-because-fp ‘Well, (I) got focused in class. Rather, (I had to) move (to other classrooms) to take English and programming classes.’ (2000 Yahoo! Blog; OY08_00189; BCCWJ)

These survey results tell us that yori remains ablative-oriented in terms of frequency, as in Table 2, resulting in the production of ablative-based formulaic expressions of various types as shown in the Appendix. Considering the fact that

33

34

Reijirou Shibasaki

the grammaticalization of yori is still progressing slowly, it is far from impossible that yori will become more polyfunctional at least as long as it is ablative-oriented. In the next subsection, we will see another collocational sequence that retains the ablative function of yori. 5.2.2 Yori co-occurring with sarani ‘moreover’ Morioka (1999: 198) points out that the sequence of yori with the adverb sarani ‘moreover’ facilitates the progress of change in yori from the comparative standard (ablative) marker to the comparative adverb. Morioka (1999) makes a shrewd guess as to many aspects of contact-induced changes in Mod-J and carries out an intensive study of what gave rise to the adverbial function of yori. In what follows, I will thus investigate the degree of grammaticalization by comparing yori’s functions between 1895–1925 and 1976–2008. Table 4 summarizes the results of the surveys based on Taiyo (1895–1925) and BCCWJ (1976–2008). The number of examples retrieved from Taiyo is very small; therefore, all the examples are lumped together for the sake of convenience. Rel. freq. = relative frequency. Table 4. The degree of grammaticalization of yori in the sequence of yori sarani 1895–1925 1976–1989 1990–1999 2000–2003 2004–2008 yori (ABL) + sarani (ADV ) yori (ADV ) + sarani (ADV ) Total (rel. freq. %)

Total

38 (100%)

37 (97.4%)

62 (100%)

70 (100%)

83 (100%)

290 (99.7%)

---

1 (2.6%)

---

---

---

1 (0.3%)

38 (100%)

38 (100%)

62 (100%)

70 (100%)

83 (100%)

291 (100%)

It is obvious that in this particular sequence, i.e., yori sarani, yori has retained its ablative function over a century, contrary to Morioka’s (1999) expectation. His shrewd observation is no doubt perceptive; however, this survey result turns out to be inconsistent with his survey, showing a new and undeniable aspect of yori. Here are two examples from the ends of the nineteenth and twentieth centuries, respectively. In both examples, yori serves an ablative function. (21) yori (ablative) + sarani (adverb) ima-yori sarani okoshi-te shin-kyoku-o tsukur.i,… now-than moreover hang.tough-and new-song-acc make.and ‘(we) put in more effort than now on composing new songs, and …’ (1895 Kagura; Taiyo)

Chapter 1. From comparative standard marker to comparative adverb

(22) yori (ablative) + sarani (adverb) Beikoku-wa tousho-wa kokusai-kakaku-yori sarani yasuku U.S.A.-top at.first-top international-price-than moreover cheap hojokin.tsuki-de danpingu-yushutsu-o suru-keredomo,… subsidized-by dumping-export-acc do-but ‘At the beginning, the United States was active in subsidy and dumping export at a much lower price than at international prices, …’ (1987 Kokumin Shokkan o Teigensuru; LBb6_00010; BCCWJ)

On the other hand, the following is the only exceptional example that includes the adverbial yori in this particular sequence. (23) yori (adverb) + sarani (adverb) 2 Kaikyaku-zenkutsu < kiryoku-ga tsuku > Yori sarani Section 2 leg.straddled-forward.flexion stamina-nom attach more moreover kokyuu-o ookiku.shi, nou-e ketsueki-o sumuuzu.ni okuru tasuke-o-suru respiration big.do brain-to blood-acc smoothly send help-acc-do.adn undoo exercise ‘Section 2: Forward bending with legs straddled “Feel one’s spirit emerge.” (This is an) exercise that increases respiration much more and supplies blood to the brain.’ (1988 Konnani Yasete Iikashira; OB3X_00124; BCCWJ)

This example comes from a diet book. The first line is the section title, accompanied by the description of the effect of forward bending with legs straddled in the next lines. Here, yori is used in sentence-initial position without any preceding element; namely, the usage is adverbial. Of course, this usage is grammatically correct, but it is also true that as a native speaker of Japanese, I feel something is wrong with this sequence. One reason is that this sequential usage is quite exceptional, as shown in Table 4. What is important here, however, is that all the other cases of yori sarani illustrate the effect of the retention of yori’s ablative function. This survey result also lends support to the slow progression of the grammaticalization of yori.

35

36

Reijirou Shibasaki

5.3 A collocational sequence that reflects the newer adverbial function of yori We have thus far confirmed the gradual process of change in Section 5.1 and two cases that slow the progress of grammaticalization in Section 5.2. In this section, we turn to another collocational sequence, which is also pointed out by Morioka (1999), that is considered to have lubricated the grammaticalization of yori from the ablative marker to the comparative adverb, i.e., yori issoo. The survey results based on Taiyo and BCCWJ are shown in Table 5. The results in Table 5 stand in marked contrast to the results in Table 4. In the stage of 1895–1925, yori was ablative-oriented in the sequential use yori issoo in much the same way as yori was ablative-oriented in the sequence yori sarani in Table 4. However, from the stage of 1976–1989 onward, yori has become adverbialoriented in this collocational sequence, as predicted in Morioka (1999). Table 5. The degree of grammaticalization of yori in the sequence of yori issoo 1895–1925 1976–1989 1990–1999 2000–2003 2004–2008 yori (ABL) + issoo (ADV )

26 (96.3%)

yori (ADV ) + issoo (ADV )

1 (3.7%)

Total (rel. freq. %)

27 (100%)

6 (6.2%)

14 (9.3%)

17 (9.2%)

11 (5.1%)

Total 74 (11%)

91 (93.8%) 137 (90.7%) 168 (90.8%) 203 (94.9%) 600 (89%) 97 (100%)

151 (100%) 185 (100%) 214 (100%) 674 (100%)

Here are examples. In (24), yori is used as an ablative marker, while in (25), it is used as a comparative adverb. Albeit in the limited set of data used for this study, (25) is the earliest environment in which yori served as a comparative adverb in this sequential use. (24) yori (ablative) + issoo (adverb) Yo-wa jitsu.ni tooji-ni at.te ima-yori issoo I.formal-top by.my.faith those.days-pt exist-and now-than moreover chimitsu-naru giron-o kiki-shi… meticulousness-cop.adn discussion-acc hear-and ‘By my faith I listened to a more detailed discussion at that time than now and …’ (1985 Kokujiron to Ishokujuuron; Taiyo)

Chapter 1. From comparative standard marker to comparative adverb

(25) yori (adverb) + issoo (adverb) sore-wa bunka.teki shizen-o that-top cultural nature-acc sinryoku-to tairyoku-to-o genshi.teki spiritual.strength-and physical.strength-and-acc primitive toohi.suru-no-de-wa-naku.te, jinsei-no kyoosoo-ni escape.do-nml-cop-top-not.and life-gen competition-pt shizen-no soboku.na naka-kara nature-gen simple inside-from yori issoo taer-are-ru-dake-no more moreover bear-can-pres-only-gen hiki.dashi-tai-tame-ni ryokoo.suru-no-de-aru. pull.out-want-purp-pt trip.do-nml-cop-exist ‘That is, (I) make a trip not to escape from the natural way of things but to bring out both spiritual and physical strengths from the naive-ness of pristine nature, by which (I) can survive the battlefield of life much more (effectively).’ (1917 Shintoo Zasoo; Taiyo)

In terms of frequency, it is safe to say that yori has become adverbial-oriented in the second half of the twentieth century, and this survey result provides a piece of evidence for Morioka’s (1999) observational adequacy. Further, the sequenced adverbs yori issoo ‘much more’ can be regarded as a formulaic expression, as Hida and Asada (2018: 587–588) properly treat in their dictionary (see Section 1 for this discussion). The current study is part of the larger research project as noted in the acknowledgments. However, yori issoo is the only collocational sequence so far that facilitates the grammaticalization of yori in an obvious fashion from the ablative case-marker to the comparative adverb. It is true that there are other minor changes within the progress of yori’s grammaticalization, as shown in the Appendix. If we continue to enlarge the size of our data, I believe we will be able to find more interesting and crucial examples so that we can take this research project one step further.

5.4 Language contact and ‘extravagance’ In this subsection, I will briefly discuss the emergence of the adverbial yori from the viewpoint of ‘extravagance’ (Haspelmath 1999) as well as language contact (leading to contact-induced grammaticalization; Narrog & Heine 2021). The maxim of extravagance is explained as follows. (26) Extravagance: talk in such a way that you are noticed. (Haspelmath 1999: 1055)

37

38

Reijirou Shibasaki

This maxim is based on Keller’s (1994) invisible-hand theory on the assumption that “language change is an unintended byproduct of ordinary language use … viewed as an invisible-hand process, that is, a phenomenon that is the result of human actions” (Haspelmath 1999: 1054). On the other hand, the maxim of extravagance is unique in that speakers use innovative expressions to achieve social success, i.e., in order to be noticed and adopted by other speakers, by choosing “a new way of saying old things,” for example, by means of for with (Haspelmath 1999: 1057). Haspelmath’s maxim of extravagance may echo with Du Bois’s (2003) view of discourse and grammar as below. (27) Theoretical assumptions about discourse and grammar a. Speakers exploit available grammatical structures to realize their goals in speaking; b. The aggregate sum of what speakers do in discourse exhibits recurrent patterning beyond what is predicted by rules of grammar; c. Grammatical structure tends to evolve along lines laid down by discourse pattern: grammars code best what speakers do most. (Du Bois 2003: 49)

What is stated in (27a) and (27b) corresponds to “a new way of saying old things” (Haspelmath 1999: 1057) just mentioned above. The idea in (27c), which is also a synchronic view of ongoing change in conversation and interaction, can perfectly naturally give an account of the diachronic change of yori discussed above and summarized in the Appendix. As to the characteristics of contact-induced grammaticalization, Narrog and Heine (2021: 203–215) propose several diagnostics to see whether cases in focus are contact-induced or not. Once given phenomena can be regarded as potential cases of contact-induced grammaticalization, they further examine how such cases spread among a larger number of speakers. In what follows, we deepen our considerations of the emergence of the adverbial yori under the rubric of their diagnostics, i.e., ‘rare grammatical category’ and ‘demographic variables,’ because these are highly relevant to this particular phenomenon.9 9. Narrog and Heine’s (2021: 204–210) diagnostics for the evidence of contact-induced grammaticalization are (i) genetic patterning, (ii) rare grammatical category, and (iii) paired grammaticalization, while their diagnostics for the direction of transfer are (i) degree of grammaticalization, (ii) frequency of use, and (iii) demographic variables. They state that among the first set, “rare phenomena are particularly sensitive to contact-induced language change (p. 208)”; the other two are not related to the current case according to their definitions. Among the second set, the diagnostic of demographic variables is only applicable to the current case (see below in this subsection). Note that their diagnostics are defined around the premise of ‘neighboring languages’; however, as shown in this study, target languages seem not necessarily to exist side by side.

Chapter 1. From comparative standard marker to comparative adverb

The functional extension of the ablative to the free-standing adverb has not been addressed in studies on grammaticalization; therefore, this phenomenon fits the confines of rare grammatical category in Narrog and Heine’s diagnostic.10 The fact that this particular change has gone unnoticed even in studies on degrammaticalization (e.g., Joseph & Janda 1988; Norde 2009) can lend further support to the rarity of this change in grammar. Demographic variables mean that language users of a particular community begin to use a given expression in an innovative way, with the result that it becomes widespread in a larger variety of communities. As discussed in Section 3.1, intellectuals in the late Edo and Meiji periods felt and clarified distinctive differences in syntax and morphology between Japanese and western languages, and then made conscious efforts to acquire an innovative usage by renewing the ready-made element yori. In other words, intellectuals of those days as a specific population of language users appear to have started to choose a new way of using the existing yori in a different context (i.e., extravagance), under the influence of both intensive and extensive contact with western languages (i.e., contact-induced grammaticalization). To be precise, the change from the ablative to the comparative adverb is a case of degrammaticalization in a traditional view of language change, which will thus be discussed below.

10. It should be mentioned here that in Korean, the postposition ‑boda functionally extended from the erstwhile standard of comparison to the newer comparative adverb; Park (2018) may be related to the current issue. I am grateful to one anonymous reviewer for reminding me of this phoenomenon. In a similar vein, Park (in preparation) addresses the emergence of the Korean pragmatic marker mwullon 勿論 ‘of course, needless to say’ in the early twentieth century due to (written) language contact with Japanese that had already used the corresponding mochiron 勿論 ‘of course, needless to say.’ These cases could be another set of potential contactinduced (de)grammaticalization, specifically as a case of ‘paired grammaticalization’ as well as ‘rare grammatical category’ in 11. 11. Narrog and Heine’s (2021: 204–210) diagnostics for the evidence of contact-induced grammaticalization are (i) genetic patterning, (ii) rare grammatical category, and (iii) paired grammaticalization, while their diagnostics for the direction of transfer are (i) degree of grammaticalization, (ii) frequency of use, and (iii) demographic variables. They state that among the first set, “rare phenomena are particularly sensitive to contact-induced language change (p. 208)”; the other two are not related to the current case according to their definitions. Among the second set, the diagnostic of demographic variables is only applicable to the current case (see below in this subsection). Note that their diagnostics are defined around the premise of ‘neighboring languages’; however, as shown in this study, target languages seem not necessarily to exist side by side.

39

40

Reijirou Shibasaki

5.5 Degrammaticalization and the emergence of the comparative adverb yori In general, grammaticalization can be regarded as a cline along which a given lexical or content item becomes a grammatical one, further into a clitic and an affix, eventually leading to zero (Narrog & Heine 2021: 279). If a given item follows the reverse pathway of change, that would display a case of degrammaticalization, although it is also said that genuine cases of degrammaticalization are hard to come by (Narrog & Heine 2021: 280–286; also see Joseph & Janda 1988, Haspelmath 2004 and Norde 2009 for details). On the other hand, one type of degrammaticalization introduced by Norde (2009: 186–227), i.e., ‘debonding,’ can explain the change of the ablative yori to the comparative adverb yori. Debonding is defined as follows. (28) Debonding is a composite change whereby a bound morpheme in a specific linguistic context becomes a free morpheme . (Norde 2009: 186)

Norde (2009: 187) states that debonding has been the most frequently addressed topic in literatures on degrammaticalization, which includes the change of the Japanese clausal subordinator ‑ga ‘although’ to a discourse marker ga ‘however’ (Matsumoto 1988). The functional expansion of yori discussed elsewhere in the current paper can be another potential case of degrammaticalization.12 While yori qualifies as a case of degrammaticalization, we need to give some thought to the following point, i.e., gradualness. As shown in Table 2 (Section 5.1), the permeation of the newer comparative adverb function of yori turned out to be nearly steady but very slow; even in the 2000s, the ratio of the comparative function to the former ablative is 17.9%. As to the propagation of an innovative form, Denison (2003: 55) shows one idea that once it occurs, “the curve [S-curve; added by the present author] slowly climbs through single-figure percentages until the innovation is being used on maybe 20–25 per cent of the available occasions (20 per cent is the usual transition point chosen in such illustrations)” for a sharp rise in frequency.13 His idea is accountable for the slow progress of the comparative adverb yori, partly because it is a rare development from the erstwhile ablative from a cross-linguistic 12. In addition to the above discussions, notions such as ‘scope expansion’ (Tabor & Traugott 1998) and ‘structural scope’ (Lehmann 1995) are also related, however, both of which cannot be discussed due to a space constraint. I express my gratitude for this comment from one anonymous reviewer. 13. For a similar view on S-curve see Blythe and Croft (2012), while for a different view see Gries & Hilpert (2010). Note that even viewed from a sociolinguistic perspective (e.g., Labov 1994, Nevalainen and Raumolin-Brunberg 2017 and Nevalainen et al. 2018), our survey results on the gradualness of change (see Table 2) are consistent.

Chapter 1. From comparative standard marker to comparative adverb

perspective and partly because it is a potential case of degrammaticalization that is likely to have been caused by language contact, more specifically by a limited social group of people who received formal education (see Narrog & Heine 2021: 214; see Sections 3.1 and 5.4). As a consequence, it might have taken time for the newly emergent comparative adverb yori to be acknowledged.14 All in all, we are led to conclude that yori is a typical instance of grammaticalization, except for one advanced change from a comparative standard marker to a comparative adverb, which can be a good case of degrammaticalization as an additional case to what is discussed in Joseph & Janda (1988). Furthermore, the rise of the comparative adverb yori can be an example of ‘debonding’ in degrammaticalization (Norde 2009: 231), because the case-marking particle yori, either an affix or a clitic, developed into a free morpheme. The emergence of the free-standing adverb yori out of the erstwhile ablative is a complex phenomenon comprising grammaticalization and degrammaticalization, the latter of which is highly associated with language contact at least in this case. It is true that this survey result cannot go beyond the realm of one case study (see 15). However, our descriptive and theoretical analyses of the functional variation and change of yori will deserve attention as a point of departure for offering a good opportunity to extend our knowledge on the issues concerning (de)grammaticalization and language contact (see 16). 14. One anonymous reviewer pointed out that only the case of yori is not sufficiently enough to validate the emergence of the comparative adverb out of the ablative. I show my appreciation for the comment. As mentioned above, however, all the possible cases of degrammaticalization are far from impeccable (e.g., Joseph & Janda 1988; Narrog & Heine 2021: 279–286). Further, this particular case can be considered to be contact-induced (de)grammaticalization as discussed thus far; therefore, the process of change is very slow inescapably (see Table 2) and relevant data appears to be still very small. I will tackle this issue in my continued research. 15. One anonymous reviewer pointed out that only the case of yori is not sufficiently enough to validate the emergence of the comparative adverb out of the ablative. I show my appreciation for the comment. As mentioned above, however, all the possible cases of degrammaticalization are far from impeccable (e.g., Joseph & Janda 1988; Narrog & Heine 2021: 279–286). Further, this particular case can be considered to be contact-induced (de)grammaticalization as discussed thus far; therefore, the process of change is very slow inescapably (see Table 2) and relevant data appears to be still very small. I will tackle this issue in my continued research. 16. It should be mentioned here that in Korean, the postposition ‑boda functionally extended from the erstwhile standard of comparison to the newer comparative adverb; Park (2018) may be related to the current issue. I am grateful to one anonymous reviewer for reminding me of this phoenomenon. In a similar vein, Park (in preparation) addresses the emergence of the Korean pragmatic marker mwullon 勿論 ‘of course, needless to say’ in the early twentieth century due to (written) language contact with Japanese that had already used the corresponding mochiron 勿論 ‘of course, needless to say.’ These cases could be another set of potential contactinduced (de)grammaticalization, specifically as a case of ‘paired grammaticalization’ as well as ‘rare grammatical category’ in 17.

41

42

Reijirou Shibasaki

6.

Concluding remarks

In this study, I addressed some issues on the (de)grammaticalization of yori from a comparative standard marker ‘than’ to a comparative adverb ‘more’ and presented the following survey results. Firstly, the whole (de)grammaticalization process of yori turned out to be slow; the comparative adverbial usage of yori emerged around the turn of the twentieth century, although the older ablative and comparative standard usages were found as early as OJ (Sections 2 and 5.1). Secondly, such a slow progress of change may be due to the fact that the functional expansion from a comparative standard marker to a comparative adverb is a rare case of degrammaticalization (Section 5.5), and some collocational expressions have served to retain the older ablative case of yori (Section 5.2), i.e. ‘persistence’ (Hopper 1991). Thirdly, a typologically untypical change of yori, i.e., degrammaticalization, is considered to have been triggered by language contact with western languages (Section 3), through which intellectuals in the Meiji period dealt consciously with such differences in morphosyntactic structures between Japanese and western languages, giving rise to the newer comparative adverbial usage of yori, presumably facilitated not by the speakers’ unintended use of it but by their ‘intended’ learning of it (Section 5.4), i.e., ‘extravagance’ (Haspelmath 1999). Fourthly, while on the whole yori exhibits a slow process of degrammaticalization, one collocational sequence yori issoo is found to have boosted the functional expansion of yori from a comparative standard (ablative) marker to a comparative adverb (Section 5.3). This finding serves as a piece of evidence for the statement that (de)grammaticalization can be witnessed or facilitated in particular sequences of elements but not in isolation (Traugott 2003; Himmelmann 2004). Lastly, the coexistence of ablative and adverbial usages at one synchronic stage is a good case of ‘layering’ (Hopper 1991), and language contact of Japanese with

17. Narrog and Heine’s (2021: 204–210) diagnostics for the evidence of contact-induced grammaticalization are (i) genetic patterning, (ii) rare grammatical category, and (iii) paired grammaticalization, while their diagnostics for the direction of transfer are (i) degree of grammaticalization, (ii) frequency of use, and (iii) demographic variables. They state that among the first set, “rare phenomena are particularly sensitive to contact-induced language change (p. 208)”; the other two are not related to the current case according to their definitions. Among the second set, the diagnostic of demographic variables is only applicable to the current case (see below in this subsection). Note that their diagnostics are defined around the premise of ‘neighboring languages’; however, as shown in this study, target languages seem not necessarily to exist side by side.

Chapter 1. From comparative standard marker to comparative adverb

western languages brought about this precise change that otherwise would be impossible. Other issues connected with contact-induced (de)grammaticalization are addressed, albeit briefly, in translation studies, but those have yet to come under scrutiny in the same vein as tackled in this study (see Shibasaki in press for details). Therefore, approaches to (de)grammaticalization from the perspective of language contact are promising. It is hoped that this case study will become a springboard for future studies.

Acknowledgments This study would not have been possible without the aid of the following grants: Grant-in-Aid for Scientific Research (C) project (Pl: Reijirou Shibasaki, No. 19K00693) and Grant-in-Aid for Scientific Research (C) project (Pl: Reijirou Shibasaki, No. 22K00610) under the supervision of the Japan Society for the Promotion of Science. I express my sincere thanks to Martin Hilpert, Katsunobu Izutsu and Yongtaek Kim for their constructive feebacks. Further, I would like to extend my gratitude to Sue Browning for her editorial assistance to the final version.

Glossing conventions abl acc adn adv com cop fp emph gen hum infr mkr neg nml

ablative accusative adnominal adverb comparative copula final particle emphatic genitive humble inferential comparative marker negative nominalizer

nom npst pass perf pol pres pst purp pt qp qual quot std top

nominative non-past tense passive perfective polite present tense past tense purposive particle question particle = quality quality quotative comparative standard topic

Division of the history of Japanese in this study OJ MJ

Nara period: 710–784; Heian period: 794–1192 Middle Japanese: Kamakura period: 1192–1333; Muromachi period: 1336–1573; Azuchimomoyama period: 1573–1600

43

44

Reijirou Shibasaki

EMod-J Mod-J PDJ

Early Modern Japanese: Edo period: 1603–1867 Modern Japanese: Meiji period: 1868–1912; Taisho period: 1912–1926 Present-Day Japanese: Showa period: 1926–1989; Heisei period: 1989–2019; Reiwa period: 2019–to the present

Dictionaries Hida, Yoshifumi & Hideko Asada. 2018. Gendai fukushi yoohoo jiten (A Usage Dictionary of Contemporary Japanese Adverbs), revised edn. Tokyo: Tokyodo. Nakata, Norio, Toshimasa Wada & Yasuo Kitahawa (eds). 1983. Kogo daijiten (An Unabridged Dictionary of Old Japanese), compact edn. Tokyo: Shogakukan. Nikkoku=Nihon Kokugo Daijiten (The Dictionary of the Japanese Language), Vols. 1–13, 2nd edn, Yasuo Nikkoku (ed.), 2006. Tokyo: Shogakkan.

Corpora Aozora = Aozora Corpus (1930s-1940s), a self-made corpus (see Note 6) BCCWJ = Balanced Corpus of Contemporary Written Japanese (1976–2008) [last access: 26 August 2022] CHJ = The Corpus of Historical Japanese [last access: 26 August 2022] CSJ = Corpus of Spontaneous Japanese (2000s) [last access: 26 August 2022] SSC = Showa Speech Corpus (1950s-1970s) [last access: 26 August 2022] Taiyo = Taiyo Corpus, CD-ROM (1895, 1901, 1909, 1917, 1925). 2005. Tokyo: Hakubunkan Shinsha.

References Adamou, Evangelia & Matras, Yaron (eds). 2020. The Routledge Handbook of Language Contact. London: Routledge. Bisang, Walter. 2020. Grammaticalization in Chinese: A cross-linguistic perspective. In A Typological Approach to Grammaticalization and Lexicalization: East Meets West [Trends in Linguistics: Studies and Monographs 327], Janet Zhiqun Xing (ed.) 17–54. Berlin: De Gruyter Mouton. Blythe, Richard A. & Croft, William. 2012. S-curves and the mechanisms of propagation in language change. Language 88: 269–304. Clyne, Michael. 2003. Dynamics of Language Contact: English and Immigrant Language. Cambridge: CUP.

Chapter 1. From comparative standard marker to comparative adverb

Denison, David. 2003. Log(ist)ic and simplistic S-curves. In Motives for Language Changes, Raymond Hickey (ed.), 54–69. Cambridge: CUP. Du Bois, J. W. 2003. Discourse and grammar. In The New Psychology of Language, Vol. II, Michael Tomasello (ed.) 47–88. Hillsdale NJ: Lawrence Erlbaum Associates. Grant, Anthony P. (ed.). 2020. The Oxford Handbook of Language Contact. Oxford: OUP. Gries, Stephan T. & Hilpert, Martin. 2010. Modeling diachronic change in the third person singular: A multifactorial, verb- and author-specific exploratory approach. English Language and Linguistics 14(3): 293–320. Haspelmath, Martin. 1999. Why is grammaticalization irreversible? Linguistics 37(6): 1043–1068. Haspelmath, Martin. 2004. On directionality in language change with particular reference to grammaticalization. In Up and Down the Cline: The Nature of Grammaticalization [Typological Studies in Language 59], Olga Fischer, Muriel Norde & Harry Perridon (eds), 17–44. Amsterdam: John Benjamins. Heine, Bernd. 1997. Cognitive Foundations of Grammar. Oxford: OUP. Heine, Bernd & Kuteva, Tania. 2005. Language Contact and Grammatical Change. Cambridge: CUP. Hickey, Raymond (ed.). 2020[2010]. The Handbook of Language Contact, 2nd edn. Malden MA: Wiley-Blackwell. Hida, Yoshifumi. 1992. Tokyo Go Seiritsu no Kenkyuu (A Study on the Establishment of Tokyo Japanese). Tokyo: Tokyodoo. Hida, Yoshifumi. 2019. Meiji Umare no Nihongo (Japanese Words Created in the Meiji Period). Tokyo: Kadokawa. Himmelmann, Nikolaus P. 2004. Lexicalization and grammaticalization: Opposite or orthogonal? In What Makes Grammaticalization? A Look from Its Fringes and Its Components, Walter Bisang, Nikolaus P. Himmelmann & Björn Wiemer (eds), 21–42. Berlin: Mouton de Gruyter. Hopper, Paul J. 1991. On some principles of grammaticalization. In Approaches to Grammaticalization, Vol. I: Theoretical and Methodological Issues [Typological Studies in Language 19], Elizabeth C. Traugott & Bernd Heine (eds), 17–35. Amsterdam: John Benjamins. Ishiwata, Toshio. 1986. English borrowings in Japanese. In English in Contact with Other Languages. Studies in Honour of Broder Carstensen on the Occasion of his 60th Birthday, Wolfgang Viereck & Wolf-Dietrich Bald (eds), 457–471. Budapest: Akadémia Kiado. Joseph, Brian D. & Janda, Richard D. 1988. The how and why of diachronic morphologization and demorphologization. In Theoretical Morphology: Approaches in Modern Linguistics, Michael Hammond & Michael Noonan (eds), 193–210. San Diego CA: Academic Press. Keller, Rudi. 1994. On Language Change: The Invisible Hand in Language. London: Routledge. Kuteva, Tania, Heine, Bernd, Hong, Bo, Long, Haiping, Narrog, Heiko & Rhee, Seongha. 2019. World Lexicon of Grammaticalization, 2nd edn. Cambridge: CUP. Labov, William. 1994. Principles of Linguistic Change, Vol.1: Internal Factors. Oxford: Blackwell. Lehmann, Christian. 1995. Thoughts on Grammaticalization. Munich: Lincom.

45

46

Reijirou Shibasaki

Loveday, Leo J. 1996. Language Contact in Japan: A Socio-Linguistic History. Oxford: Clarendon Press. Maruyama, Takehiko. 2020. On the possibility of a diachronic speech corpus of Japanese. In Japanese Language from Empirical Perspective. Corpus-Based Studies and Studies on Discourse, Andrej Bekeš & Irena Srdanović (eds), 219–234. Ljubljana: Znanstvena založba FF. Matsumoto, Yo. 1988. From bound grammatical markers to free discourse markers: History of some Japanese connectives. In Proceedings of the Fourteenth Annual Meeting of the Berkeley Linguistics Society, Shelley Axmaker, Annie Jaisser & Helen Singmaster (eds), 340–351. Berkeley CA: BLS. Morioka, Kenji. 1999. Oobun Kundoku no Kenkyuu (A study of Western languages in their Japanese Pronunciation). Tokyo: Meiji-Shoin. Muysken, Pieter. 2000. Bilingual Speech: A Typology of Code-Switching. Cambridge: CUP. Narrog, Heiko & Heine, Bernd. 2021. Grammaticalization. Oxford: OUP. Nevalainen, Terttu & Raumolin-Brunberg, Helena. 2017. Historical Sociolinguistics, 2nd edn. London: Routledge. Nevalainen, Terttu, Palander-Collin, Minna & Säily, Tanja (eds). 2018. Patterns of Change in 18th-Century English: A Sociolinguistic Approach [Advances in Historical Sociolinguistics 8]. Amsterdam: John Benjamins. Norde, Muriel. 2009. Degrammaticalization. Oxford: OUP. Park, Jiyeon. In preparation. Semantic change of the Chinese loanword mwullon 勿論 in Korean. In The Emergence of Pragmatic Markers from Chinese Compounds in Chinese, Japanese and Korean: Perspectives from East Asian Languages and Beyond, Yukio Higashiizumi & R. Shibasaki (edn). Leiden: Brill. Park, Kang-Hun. 2018. Grammaticalization of Japanese postpositions: Focusing on yori cooccurring with negatives. Language and Linguistics 81: 51–73. Payne, Thomas E. 1997. Describing Morphosyntax: A Guide for Field Linguistics. Cambridge: CUP. Sansom, George B. 1928. An Historical Grammar of Japanese. Oxford: Clarendon Press. Shibasaki, Reijirou. 2021. Reanalysis and the emergence of adverbial connectors in the history of Japanese. In Studies at the Grammar-Discourse Interface: Discourse Markers and Discourse-Related Grammatical Phenomena [Studies in Language Series 219] Alexander Haselow & Sylvie Hancil (eds), 102–124. Amsterdam: John Benjamins. Shibasaki, Reijirou. In press. Gengo sesshoku to bunpooka nitsuite: Kingandai Nihongo no yori koobun o jirei toshite (Language contact and grammaticalization: A case of yori-constructions in Modern and Present-Day Japanese). In Nihongo to Kinronshogo niokeru Bumpooka to Bumpoohenka (Grammaticalization and Grammatical Changes in Japanese and Neighboring Languages), Hirofumi Aoki & Heiko Narrog (eds). Tokyo: Hituzi Syobo. Stassen, Leon. 1985. Comparison and Universal Grammar. Oxford: Blackwell. Sugimoto, Tsutomu. 1983. Nihon Hon’yakushi no Kenkyuu (A Study on the History of Translation in Japan). Tokyo: Yasaka-Shobo.

Chapter 1. From comparative standard marker to comparative adverb

Tabor, Whitney & Traugott, Elizabeth Closs. 1998. Structural scope expansion and grammaticalization. In The Limits of Grammaticalization [Typological Studies in Language 37], Anna Giacolone Ramat & Paul J. Hopper (eds), 229–272. Amsterdam: John Benjamins. Thomason, Sarah G. & Kaufman, Terrence. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley CA: University of California Press. Traugott, Elizabeth Closs. 2003. Constructions in grammaticalization. In The Handbook of Historical Linguistics, Brian D. Joseph & Richard D. Janda (eds), 624–647. Oxford: Blackwell. Weinreich, Ulrich. 1953. Languages in Contact. The Hague: Mouton de Gruyter. Xing, Janet Zhiqun. 2015. A comparative study of semantic change in grammaticalization and lexicalization in Chinese and Germanic languages. Studies in Language 39(3): 594–634. Xing, Janet Zhiqun. 2020. Introduction. In A Typological Approach to Grammaticalization and Lexicalization: East Meets West, Janet Zhiqun Xing (ed.) 1–13. Berlin: De Gruyter Mouton. Yagishita, T. 2018. Kindai Nihongo no Keisei to Oobun Chokuyakuteki Hyoogen (The Formation of Modern Japanese and Literal Translation-based Expressions). Tokyo: Bensei-Sha. Yanabu, Akira. 2004. Kidai Nihongo no Shisoo: Hon’yaku Seiritsu Jijoo (Thoughts of Modern Japanese: Circumstances on the Establishment of Translation). Tokyo: Hosei University Press.

Appendix. The major collocational patterns with yori (based on Shibasaki in press) Collocational pattern yori (adverb)

Subtotal (%/ the whole substage)

2000s

yori yoi/ii/etc. ‘be better’

5

-

2

yori ooi ‘be larger’/sukunai ‘be smaller’

-

4

3

yori seikakuna ‘precise’

-

-

3

yori issoo ‘much more’

-

-

3

others

1

5

58

3.2%

11.0%

17.9%

ima made yori ‘than ever before’

1

1

1

itsumo yori ‘than usual’

1

-

-

mae yori mo ‘than ever before’

-

1

1

others

-

1

41

1.1%

3.7%

11.2%

Subtotal (%/ the whole substage) Temporal expressions with yori (ablative)

1950s 1960–70s

47

48

Reijirou Shibasaki

Appendix. (continued) Collocational pattern General nouns with yori (ablative)

9

1

2

kijun yori ‘than/from the norm’

-

2

-

65

39

170

39.6%

51.2%

44.7%

first person + yori ‘than I’

10

2

-

second person + yori ‘than you’

3

1

-

reflexive + yori ‘than (one)self ”

1

-

3

Subtotal (%/ the whole substage)

Subtotal (%/ the whole substage) Interrogatives with yori (ablative)

7.5%

3.7%

0.8%

nani yori ‘above all (things)’

5

-

7

watashi nanka (‘what’) yori ‘than I’

2

-

3

dare yori mo ‘more than anyone’

-

-

1

Subtotal (%/ the whole substage) Demonstratives with yori (ablative)

Subtotal (%/ the whole substage)

2000s

Place name + yori ’from somewhere’

others Personal pronouns with yori (ablative)

1950s 1960–70s

3.7%

0%

2.9%

proximal + yori ‘than this/ these’

3

2

6

medial + yori ‘than that/ those’

11

5

7

distal + yori ‘than that/those’

2

4

1

sono tameni yori ‘than for that’

-

1

-

8.6%

14.6%

3.6%

Chapter 1. From comparative standard marker to comparative adverb

Appendix. (continued) Collocational pattern Adnominal forms of the verb with yori (ablative)

2000s

to-iu-yori-mo ‘rather’

13

3

18

tte-iu-yori-mo ‘rather’

7

5

5

omot-ta-yori (mo/ne) ‘than thought’

3

-

2

omot-te-iru-yori ‘than someone thinks’

-

-

1

28

3

24

27.3%

13.4%

12.9%

10

-

3

no-yori ‘than (e.g. what one ate)’

-

-

4

(to-itta) mono yori ‘than socalled …’

-

1

4

tte (yuu-koto) yori ‘rather’

1

1

1

to-iu-koto-yori ‘rather’

1

-

-

others Subtotal (%/ the whole substage) Nominalized expressions with yori (ablative)

1950s 1960–70s

(no) koto-yori ‘than (e.g. what is done)’

others

4

-

8

8.6%

2.4%

5.2%

moto-yori ‘from the beginning’

1

-

2

ijoo-yori ‘from the above’

-

-

1

Subtotal (%/ the whole substage)

0.5%

0%

0.8%

Total

187

82

385

Subtotal (%/ the whole substage) Others

49

chapter 2

From fear to reason Grammaticalization as dependency vs expansion of the Mandarin apprehensive 怕 pà Vittorio Tantucci & Aiqing Wang

Lancaster University | University of Liverpool

In this study we support the view that grammaticalization can unfold both in the form of increased dependency and increased expansion (cf. Traugott & Trousdale 2013), even when originating from the same original lexeme. We provide the case study of the Mandarin apprehensive verb 怕 pà ‘to be afraid’ which developed a new epistemic meaning ‘to believe, to conclude that’. The shift from fear to reason in so-called epistemic-apprehensives (Jing-Schmidt & Kapatsinski 2012) has been observed in a number of languages, including the English construction I am afraid (Leech 1983; Palmer 2001). What this study shows is that this transition in Chinese involves grammaticalization as increased dependency for constructions such as 不怕 bùpà ‘no fear’ > ‘although’ and 只怕 zhǐpà ‘only fear’ > ‘probably’. However, it also leads to grammaticalization as expansion, such is the case of 恐怕 kǒngpà ‘to be afraid’ > ‘to conclude that’. This supports and reinforces Traugott & Trousdale’s (2013) view that grammaticalization is not necessarily bound to increased reduction and dependency and can be at work also as a mechanism of functional expansion, even along the same cline of change, such is the present one involving the same source (fear) and target domains (reason) of reanalysis. Keywords: grammaticalization, apprehensives, Chinese, epistemic, evidentiality, expansion

1.

From fear to reason cross-linguistically

Literature in psychology often points to the important role of the communication of emotions in order to achieve interactional goals (e.g. Ekman 1984; Keltner & Gross 1999). Fear is no doubt one of the most basic emotions and is recognised to be communicated for different purposes in all languages (Ekman 1994; Wierzbicka

https://doi.org/10.1075/slcs.232.02tan © 2023 John Benjamins Publishing Company

Chapter 2. From fear to reason

1999). In this paper, we are interested in the semantic shift from expressive meanings of fear to epistemic ones hinging on conjecture, evaluation and/or certainty. This polysemy – surprisingly under-researched in the literature – has been attested typologically as the semantic category of the ‘apprehensionals’ (Dixon 1977, 1988), more commonly defined as ‘apprehensives’ (cf. Jing-Schmidt & Kapatsinski 2012; Yang & Fong 2015; Mitkovska et al. 2017; Boogaart 2020). Apprehensives have been observed in many languages, including Romance, Sino-Tibetan, Germanic, Slavic, Papuan (cf. Haiman 1980; Niculescu 2013; Mathiassen 1996) and involve the transition of fear expressions of affective stance towards a new epistemic reanalysis (e.g. Palmer 2001; Givón 2001; Yap et al. 2012; Yang & Yap 2015). An example of this is the ‑ada marker in To’aba’ita (an Austronesian language spoken in the Solomon Islands) which expresses epistemic reasoning, yet accompanied with a sense of dissatisfaction towards the likely outcome of the proposition (Gao 2003). Austin (1981: 225), with reference to Diyari, suggests that apprehensives indicate some situation which the speaker considers to be unpleasant and which should be avoided. The shift from the apprehensive to the epistemic meaning can be seen in the two usages of being afraid below, taken from the demographic sample of the British National Corpus (BNC1) below: (1) He had always been afraid of water.

BNC ACB 3408

(2) I am afraid that the Government have to take a rather broader view of what is in the interests of the economy. BNC C92 1955

In (1) to be afraid literally expresses the affective stance towards some event or state of affairs. On the other hand, the meaning of to be afraid in (2) is speech act oriented (e.g. Narrog 2012), as it is centred on the speaker’s reasoning process, that is, his/her evaluation about the likelihood of some state of affairs. The literal meaning of to be afraid is likely to be expressed via the [BE afraid of NP] construction, while the epistemic meaning of the lexeme tends to be structurally connected with the [BE afraid that CC2] construction. In the latter case, the element of fear is no longer in profile and a new proposition acts as a complementing clause (CC). The whole utterance is thus characterised by evaluative illocutionary force (Tantucci 2016a; Tantucci & Wang 2018, 2020a), viz. one that involves the speaker’s epistemic evaluation of about the posibility of some state of affairs as a result of a reasoning process. Leech (1983) discusses the pragmatics of epistemic-apprehensives – such as to be afraid in (2) – and argues that they convey the interpersonal function of

1. (10 April 2022). 2. Complementing clause.

51

52

Vittorio Tantucci & Aiqing Wang

preparing the addressee for some unwelcome yet likely impending event. Von Wright (1951) is the first to emphasise that FEAR predicates often merge more than one modal category, as they tend to express epistemic modality as well as attitudinal modality. Palmer (2001:13) notes that fear (and wish) expressions include a deontic component – hinging on the speaker’s attitude – as well as an epistemic one. The latter, accordingly, emphasises the speaker’s degree of certainty that a particular event will be realised. Givón (2001) considers that the deontic dimension of apprehensives is mainly at play with future events, whereas epistemic usages tend to index past events with some unresolved factual status. Lichtenberk (1995) talks of apprehensional-epistemics as predicates that convey the speaker’s uncertainty but also his/her evaluation of the proposition as being undesirable, as in the case of the dē in Standard Fijian. A key stage of the grammaticalization of apprehensives is when the expressive component of the expression is almost entirely dropped and the construction is used as politeness marker, functioning as a “neutral epistemic downtoner”. This is when the lexeme can appear at clause periphery with a new intersubjective function, no longer requiring a complement construction (e.g. That was overdone again, I am afraid. BNC HHV 9914). Another important diagnostic of change is temporal reference, as grammaticalised epistemic-apprehensives can be used with propositions of any temporal status: present, future or past (Lichtenberk 1995: 294–295). Consider the minimal pair below. In (3) I am afraid occurs in a biographical written monologue sampled from the BNC. The reason why I am concerned in (4) is not idiomatic is because the latter has not reached the same degree of epistemic reanalysis and therefore cannot refer to a past event via a complementising clause: (3) I am afraid that she had no success with me.

BNC B22 68

(4) * I am concerned that she had no success with me.

In the following section we will introduce the distinction between grammaticalization and dependency vs grammaticalization as expansion. While the past literature on grammaticalization (i.a Lehmann 1994; Bybee, Perkins & Pagliuca 1994) was almost entirely focused on the former, we will show that in Chinese the epistemic-apprehensive lexeme 怕 pà ‘to be afraid of ’ has been grammaticalising in both ways. This will suggest that grammaticalization as increased dependency and expansion are not necessarily incompatible, even as mechanisms originating from the same lexeme.

1.2 Grammaticalization as dependency vs expansion Grammaticalization has been traditionally defined as the “creation of grammatical categories” (Lehmann 2004: 183) and “the coming into being of grammatical mark-

Chapter 2. From fear to reason

ers such as case, tense, aspect, modality, mood, and connectives, etc” (Traugott & Trousdale 2013: 32). A key aspect of grammaticalization that has been traditionally overlooked is that it may be geared towards dependency or expansion. From the angle of “grammaticalization as reduction and increased dependency” as discussed in Traugott & Trousdale (2013: 32), – hereby, the GD model – grammaticalization involves increase in dependency and reduction of various aspects of the original expression (e.g. Givón 1979; Lehmann 1995; Haspelmath 2004). This means that a linguistic item becomes increasingly dependent on the constructional structure it is part of. Consider for instance the Latin periphrastic cantare habeo ‘sing:INF have:1sg’ whereby habeo turns into a future tense marker in French and Italian chanterai/canterò ‘sing:FUT:1sg’ (Fleischman 1982: 71). The grammatical change of habeo here is underpinned by a stronger morphosyntactic dependency on the [V:FUT] construction than it had on the original deontic [V habeo] construction. Typologically, this proved to be quite common, as it involves a cline of change starting from obligation, then shifting to intention and, finally, encoding future purpose meanings (cf. Bybee, Perkins, & Pagliuca 1994: 240). The same applies to the Old English numeral an ‘one’, which grammaticalised into the more schematic indefinite article a as a necessary component of of the [a(n) NP] construction. The GD approach is centred on reduction and increased dependency as signs of a new grammatical category that is being developed (Traugott & Trousdale 2013: 108). However, while there are domains – such as gender – where grammaticalization often involves increase in dependency, there are other domains – e.g. connectives or pragmatic markers – where grammaticalization may lead to a decrease in syntactic dependency. In a more recent tradition, grammaticalization is not only seen as morphosyntactic reduction and constructional dependency, but also has to do with the expansion of semantic-pragmatic, syntactic and collocational range (Himmelmann 2004), such as in the formation of pragmatic markers and the change in information structuring. This is what Traugott & Trousdale (2013: 33) identify as “grammaticalization as expansion”, henceforth, the GE model (Ibid.).3 Examples of GE are the notorious [WXDY ] construction What’s his fly doing in my soup? (cf. Fillmore & Kay 1999), which, in addition to its literal sense, acquired an expanded function expressing surprise accompanied by a hint of disapproval. The same applies to the [you don’t want X] construction in English, which diachronically acquired a new directive illocutionary force, thus functioning as a hortatory construct rather than an assertive statement (cf. Tantucci 2017a, 2021). This is one of the reasons why Himmelmann argues that gram3. See also Heine (2018) for a critical discussion on whether extra-clausal constituents should be characterised as grammaticalized.

53

54

Vittorio Tantucci & Aiqing Wang

maticalization crucially boils down to context-expansion (2004: 32). He notes that a sole focus on grammaticalising items alone is misleading, as constructions never grammaticalise out of context: “constructions (elements in context) and not individual items are the proper domain of grammaticization” (Himmelmann 2004: 31). In more traditional GD approaches, unidirectionality is morphemebound, to the point of having been considered irreversible (Haspelmath 1999). On the other hand, Traugott & Trousdale (2013) argue that in the case of GE, directionality is primarily centred on “expansion to more collocations and to more syntactic, semantic, and pragmatic options” [and] answers the question about how changes affect use in contexts and how contexts enable change (2013: 112). In recent years, research on language change has been increasingly concerned with the diachronic formation of conventional pairs of form and meaning, (as assumed in various versions of Construction Grammar, e.g. Goldberg 2006; Traugott & Trousdale 2013) rather than on the strict correspondence between morphosyntactic reduction and grammatical categorisation. Accordingly, Construction Grammar posits that the linguistic system is a network of constructions that may primarily vary in schematicity (abstraction), complexity (length) and procedurality (functional dimensions). Traugott & Trousdale (2013) account for the diachronic formation of grammar with a focus either on the creation or the changes in constructions, understood as conventional symbolic units (see e.g. Langacker 1987; Croft 2005). The key here is that constructional structures diachronically tend to become so idiosyncratic (Goldberg 1995) and/or so frequent (Goldberg 2006) that “the sign becomes entrenched as a form-meaning pairing in the mind of the language user” (Traugott & Trousdale 2013: 1). What this entails is a view of grammatical change that is no longer exclusively bound to morphosyntactic reduction, but rather involving any sort of change that leads to the re-organisation of a pair of form and meaning, i.e. a constructional unit. This includes pragmatic markers such as the adverbial actually or the more complex expression believe it or not (Tantucci 2021), the British English tag innit ‘isn’t it’, but also epistemic predicates acquiring a new performative speech-act orientation such as sentence peripheral usages of I insist or I think (Traugott 1989, 2016; Tantucci 2015a; Tantucci 2017b; Tantucci & Wang 2021, 2022a, 2022b). These are all part of the linguistic system as a result of grammatical change, and all part of speakers’ constructional knowledge. Grammatical change is thus at work for complex expressions such as [the ADJ‑er, the ADJ‑er] as much as for more atomic constructions such as the inflected English past tense [‑ed] or the definite article [the]. This also means that the grammatical change of a construction involves increased procedural meanings at the expenses of literal ones, such as for evidential usages of the pragmatic markers apparently or as it seems (Tantucci 2016b), both indicating that a piece of information is shared by other social members, rather than

Chapter 2. From fear to reason

literally expressing what is ‘visible at sight’ (Tantucci 2021). From a diachronic Construction Grammar perspective, the focus is no longer exclusively on the formation of morphemes or grams (cf. Bybee et al. 1994) as a by-product of phonetic reduction, but on how abstract form-meaning schemas themselves change as their members change (Traugott & Trousdale 2013: 95). The present study is centred on the Mandarin lexeme 怕 pà ‘to be afraid of ’. We provide three cases of constructional change involving either a reanalysis as dependency or expansion. Our goal is to show that both mechanisms are possible not just for separate lexemes of the linguistic system, but also for ones that are part of the same semantic cline of change, which in this case underpins fear as an affective source domain and reason as an epistemic target one. This tendency partly underpins heterosemy as it involves “etymologically related meanings [being] associated with different morphosyntactic categories” (Heine 2018: 27; but see also Lichtenberk 1991). What is distinctive of this case is that a process of differentiation is at play for the same lexeme both in the direction of grammaticalization as expansion and of grammaticalization as increased dependency.

1.3 Affectives in Mandarin: 恐怕 kǒngpà Most of the literature on the affective-epistemic shift in Chinese has been centred on 恐怕 kǒngpà, which is a verb constituted of two morphemes both expressing fear (Xinhua Dictionary 2004: 264, 372; Dictionary of Classical Chinese 2002: 886; Ōta 2003: 274–275). In Modern Mandarin, 恐怕 kǒngpà is rarely used to express the original meaning of fear and predominantly functions as a modal adverb of estimation or speculation (Modern Chinese Dictionary 2016: 748). During the 南宋 Nán Sòng Southern Song Dynasty (1127–1279), 恐怕 kǒngpà evolved from a main verb into an adverbial indicating worry or doubt, and during the 元 Yuán (1271–1368) and 明 Míng (1368–1644) Dynasties, it started to function as an epistemic adverbial expressing the foreseen outcome of some state of affairs (Liu 2011). An inferentialevidential reading is suggested by Sheng (2008) who notes that the use of 恐怕 kǒngpà is based on available information and some kind of cause-and-effect logic, therefore involving a high degree of certainty, rather than mere likelihood (Sheng 2008). However, some persistence (Hopper 1991) still remains, as the predictive function of 恐怕 kǒngpà in Modern Mandarin often intersects with the semantic component of being worried. It is mostly characterised by a negative prosody and is often used in unfavourable situations (Yang 2016). This view is also endorsed by Jing-Schmidt & Kapatsinski (2012) as they similarly suggest that 恐怕 kǒngpà tends to be used when the speaker is confident about his/her epistemic claim, yet wishes to be more circumspect (or even reluctant) when it comes to asserting his/her certi-

55

56

Vittorio Tantucci & Aiqing Wang

tude. Despite a clear tendency towards epistemicity, older expressive usages are yet still present in Modern Mandarin. Consider the pair of examples below: (5) 她说她恐怕过不了这个秋天了。 Tā shuō tā kǒngpà guò bù liǎo zhè gè qiūtiān le. She say she be-afraid live not le this cl4 autumn sfp5 ‘She said she was afraid that she wouldn’t survive this autumn.’ CCL6 / Fiction / 大陆作家 Dàlù zuòjiā ‘Mainland writers’ / 佳作 Jiāzuò ‘Outstanding works’ (6) 由于你不具备专业知识，恐怕你也看不出个所以然。 Yóuyú nǐ bú jùbèi zhuānyè zhīshi, kǒngpà nǐ yě kàn bù chū gè suǒyǐrán. Since you not have expertise knowledge it-is-probable you also see not out cl reason ‘Since you don’t have expertise knowledge, you probably can’t tell what happened.’ CCL / Spoken / 关于钱云会案的对话 Guānyú Qián Yúnhuì àn de duìhuà ‘A dialogue on Qian Yunhui’s case’

In (5) above, 恐怕 kǒngpà ‘to be afraid’ refers to the affective stance of the syntactic subject and literally conveys the emotional disclosure of her fear, namely of not being able to survive the Autumn. The illocutionary force of her reported speech is the one of an expressive speech act, that is, one that conveys the speaker’s feelings about themselves or the world (Searle 1976: 12). Expressives underpin a psychological condition, and thus not beliefs or intentions, which arise to given states of affairs (Norrick 1978: 279). This means that, in Searle’s terms, 恐怕 kǒngpà ‘to be afraid’ in (5) has no direction of fit. This is because such expressions indicate a mental or affective state, but do not convey the intention of changing the state of affairs of the world (world to word direction of fit ↑) that is external to the mind of the speaker. It similarly does not express the intention of expressing a proposition or a belief that would match the state of affairs of the external world (word to world direction of fit ↓) This is why in (5) the syntactic subject’s mental state could be indexed anaphorically via expressions overtly addressing the syntactic subject's feelings such as 她的心情就是这样 tā de xīnqíng jiùshì zhèyàng her feelings are/were just quite like that:

4. Classifier. 5. Sentence final particle. 6. Peking University diachronic corpus of Mandarin Chinese (CCL). (10 April 2022).

Chapter 2. From fear to reason

(5a) 她说她恐怕过不了这个秋天了。她的心情就是这样。 Tā shuō tā kǒngpà guò bù liǎo zhè gè qiūtiān le. Tāde xīnqíng jiù shì zhèyàng. She say she be.afraid live not pot7 this cl autumn sfp Her feeling exactly is this ‘She said she was afraid that she wouldn’t survive this autumn. Her feelings were just quite like that.’

Similarly, expressive usages of 恐怕 kǒngpà ‘to be afraid’, such as the one in (5), are compatible with a preceding expression 不恐怕 bù kǒngpà ‘not being afraid of ’ restricting the scope of the apprehension, as part of a contrastive construction [不 恐怕 bùkǒngpà NP, (只8 zhǐ) 恐怕 kǒngpà VP], literally meaning not being afraid of X, only being afraid of Y. Things are different in the case of 恐怕 kǒngpà in (6). There is now a semantic shift from the expression of emotions to epistemicity. In this case, 恐怕 kǒngpà indexes the speaking subject’s stance, rather than the syntactic subject’s one (cf. Benveniste 1971[1958]; Traugott 2003; Langacker 2008; Tantucci 2021). The illocutionary force of the utterance is thus evaluative and includes a word to world direction of fit (↓), as it now profiles the speaker’s epistemic reasoning about the likelihood of some state of affairs of the external world. This means that anaphoric reference to the speaking subject’s emotions is no longer a meaningful expression as it was in (5): (6a) 由于你不具备专业知识，恐怕你也看不出个所以然。*我的心情就是这 样。 Yóuyú nǐ bú jùbèi zhuānyè zhīshi, kǒngpà nǐ yě kàn bù chū gè suǒyǐrán. Wǒde xīnqíng jiù shì zhèyàng. Since you not have expertise knowledge it-is-probable you also see not out cl reason My feeling exactly is this ‘Since you don’t have expertise knowledge, you probably can’t tell what happened. *My feelings are just quite like that.’

The same applies for the ‘contrastive test’ mentioned above [不恐怕 bùkǒngpà NP, (只 zhǐ) 恐怕 kǒngpà Y ]. This is because 恐怕 kǒngpà here has already started to acquire a pragmatic marker function, expressing the speaker’s epistemic stance, rather than functioning as a full main verb expressing the emotions of a specific syntactic subject. The reason is that when it is used impersonally 恐怕 kǒngpà does not function as an imperative – as for full dynamic verbs such as 走 zǒu ‘go’ or stative ones as for 睡 shuì ‘sleep’ do – but necessarily conveys an evaluative illocutionary force. As we shall see in the case study in Section 3, this 7. Potential complement particle. 8. Only.

57

58

Vittorio Tantucci & Aiqing Wang

epistemic-apprehensive use of 恐怕 kǒngpà is prevalent in Mandarin, to the point of expressing the speaker’s stance even in cases when 恐怕 kǒngpà is immediately preceded by a personal pronoun other than 我 wǒ ‘I’. Consider Example (7) below: (7) 但你恐怕没有意识到这东西的破坏力。 Dàn nǐ kǒngpà méi yǒu yìshidào zhè dōngxī de pòhuài lì. But you be.afraid not have realise this stuff str9 destructive power ‘But I’m afraid you haven’t realised its destructive power.’ ccl / Spoken / 关于钱云会案的对话 Guānyú Qián Yúnhuì àn de duìhuà ‘A dialogue on Qian Yunhui’s case’

Example (7) clearly illustrates how 恐怕 kǒngpà in Modern Mandarin acquired a procedural function as a pragmatic marker, expressing the speaker’s stance as a default interpretation. In fact, despite occurring immediately before the personal pronoun 你 nǐ ‘you’, it refers to the speaker’s evaluation rather than indexing the syntactic subject’s emotions or beliefs. Such epistemic-apprehensive usage of 恐 怕 kǒngpà can arguably be considered as a case of grammatical expansion (GE), as the construction is now more grammatically independent, i.e. no longer constrained by reference to the syntactic subject of the clause. This is also true for the semantic interpretation of the lexeme, which is no longer bound to the expression of the inner feelings of the subject, despite retaining a negative prosody (Sinclair 1996), normally oriented towards an unpleasant state of affairs, generally on behalf of the addressee (cf. (7)). This is indeed what is argued both in JingSchmidt & Kapatsinski (2012) and Yang & Yap (2015), who suggest that the emerging meaning of the construction is generally one of high certainty, yet with an interpersonal function of hedging and sympathy towards the hearer, who is generally the bearer of the effects of the negative news. This GE stage is one of immediate intersubjectification (I-I) (Tantucci 2017a, 2017b, 2020, 2021; Tantucci & Wang 2020b), as the speaker markedly produces a surplus of meaning that preemptively addresses the hearer’s potential reactions to what is being said. This form of intersubjectivity is immediate (I-I) – rather than extended (E-I) – in that it is centred on a specific addressee who is present during the here-and-now of the interaction. At this stage, the epistemic reading is possible either with or without the occurrence of the first person pronoun. From (5) to (6–7) we can see a constructional reanalysis involving a shift from expressive to evaluative illocutionary force, formally corresponding to the transition from reference to the syntactic subject (morphosyntactically obligatory) to the speaking subject (no longer morphosyntactically obligatory). This corresponds to the transition between the possibility to have either a NP or a complementing clause (CC) as the object of 恐怕 kǒngpà 9. Structural particle.

Chapter 2. From fear to reason

(hence the ‘NP/CC’ notation in the slot in Example (8), meaning either noun phrase or proposition), to a new epistemic meaning, which is formally bound to a complementing clause (hence the CC notation). This is illustrated in (8) below: (8) [Synt-Subj 恐怕afraid np/cc] > [(Speak-Subj) 恐怕afraid/think cc]

A further stage of constructional expansion hinges on extended intersubjectivity (E-I). An utterance is of an E-I type when it bears a social meaning, involving the communicated awareness that a piece of information is relevant not only to a specific addressee, but also to other members of society (Tantucci 2015a, 2015b; Tantucci & Wang 2020b; Tantucci & Di Cristofaro 2021). The expressive component of 恐怕 kǒngpà of the E-I type is even less in profile whilst the communicative function the lexeme acquires an interpersonal evidential dimension (Tantucci 2013, 2021; Van Olmen 2019), although one may hope otherwise, one should concur that P. Different from evaluative usages (EvF(P)) of 恐怕 kǒngpà, the evidential inference of the construction depends on a presentative illocutionary force (PreF(P)) (Faller 2002; Tantucci 2016a, 2016b, 2021; Tantucci & Wang 2018, 2020a), which is in action when a piece of evidence from the external world is transmitted to the addressee/hearer based on some evidence, yet without the speaker’s subjective commitment to his/her own belief. This kind of invited inference (Traugoutt & Dasher 2002) is only possible in absence of the first person pronoun. A key aspect of this transition is that the construction is now no longer bound to dialogic interaction and the awareness of the potential emotions of a specific interlocutor. Rather, it can now occur in monologues and written registers, thus functioning independently as the apodosis of an assumed concessive move, although one may hope otherwise (Sweetser 1990: 69–73; Couper-Kuhlen & Thompson 2000; Squartini 2012): (9) 安徽省凤阳县的小岗村，现在恐怕没有谁不知道了。 ānhuī shěng fèngyáng xiàn de xiǎogǎng cūn, xiànzài kǒngpà méi yǒu shuí bù zhīdào le Anhui province Fengyang county str Xiaogang village now conclude not have who not know sfp10 ‘Now everybody will probably know of Xiaogang Village in Fengyang County, Anhui Province.’ ccl / Contemporary / Practical writing /中国农民调查 Zhōngguó nóngmín diàochá ‘Report on Chinese farmers’

Different from the literal affective usage in (5), 恐怕 kǒngpà in (9) above is entirely optional and has the pragmatic function of expressing the social awareness that a piece of information would be considered to be true by many other people in soci10. 了 le is a marker of current relevance to the present (a Sentence final particle SFP).

59

60

Vittorio Tantucci & Aiqing Wang

ety. In this case, the evidential reading does not depend on reference to a specific source of information, but rather on collective intentionality (cf. Tomasello 2019) and the idea that a statement is reliable – and pragmatically presented as external evidence – because it is expected to be found sensible by others: (10) [Synt-Subj 恐怕afraid np/cc] > [Speak-Subj 恐怕afraid/think cc] > [恐怕one-wouldconcurcc]

In (10), we reported the transition involving the grammatical expansion of 恐怕 kǒngpà originally occurring as an apprehensive, syntactic subject-bound predicate > to potentially expressing evaluative force > to a final stage of change also potentially expressing a social meaning grounded in interpersonal evidentiality (Tantucci 2013). Table 1. below summarises the three stages of change, their formal structure and the criteria to assess them. Table 1. Criteria for assessing the transition from apprehensive to epistemic usages of 恐 怕 kǒngpà Function

Formal structure

Apprehensive [Synt-Subj 恐怕 afraid NP/CC]

Criteria – Obligatory Syntactic subject. – The syntactic subject is flexible (e.g. 1st, 2nd, 3rd person). – Can be indexed by Subj 的心情就是这样 de xīnqíng jiùshì zhèyàng ‘that’s actually how Subj felt’. – Compatible with [not BE afraid of X, only BE afraid of Y ]. – Bound to present/future reference.

Apprehensive/ [(Speak-Subj) 恐怕 – Optional Syntactic subject. Epistemic – If present, the syntactic subject is in the 1st person singular. afraid/think CC] – * Cannot be indexed by Subj 的心情就是这样 de xīnqíng jiùshì zhèyàng ‘that’s actually how Subj felt’. – * Incompatible with [not BE afraid of X, only BE afraid of Y ]. – Past, present, future reference. Evidential

[恐怕one-wouldconcur CC]

– Absence of Syntactic subject. – * Cannot be indexed by Subj 的心情就是这样 de xīnqíng jiùshì zhèyàng ‘that’s actually how Subj felt’. – * Incompatible with [not BE afraid of X, only BE afraid of Y ]. – Past, present, future reference.

1.4 Apprehensives in Mandarin: 只怕 zhǐpà and 不怕 bùpà While a specific study on the evidential reanalysis of 恐怕 kǒngpà is currently in preparation by the authors, in this paper we will primarily focus on the affec-

Chapter 2. From fear to reason

tive > epistemic shift and the tendency of the 怕 pà lexeme to move either in direction of grammatical dependency (GD) or grammatical expansion (GE). In this sense, the dependent variable of the analysis in Section 3 will involve whether a transition from the domain of ‘fear’ to the one of epistemicity has occurred or not. In order to do so, this study also includes two frequent constructions that are based on 怕 pà semantics, namely 不怕 bùpà and its reanalysis from ‘not being afraid’ to ‘despite/even though’, and 只怕 zhǐpà changing from the original meaning of ‘only being afraid of ’ to ‘very likely’. Consider examples (11–12) below: (11) 我不怕死。 Wǒ bù pà sǐ. I not be-afraid die ‘I’m not afraid of death.’ CCL / Qing Dynasty / Fiction / 小五义 Xiǎowǔyì ‘Five martyrs’ posterity’ (12) 不怕范围窄，越窄越便于聚精神。 Bùpà fànwéi zhǎi, yuè zhǎi yuè biànyú jùjīngshén. Not.fear scope narrow more narrow more easy concentrate ‘Even if the scope is narrow, that is not a problem, the narrower the scope, the easier to concentrate. CCL / Modern / News / 读者 Dúzhě ‘Readers’

In Example (11), 不怕 bùpà is used literally as the main verb of the sentence and indexes the emotions of an explicit syntactic subject. In (12) 不怕 bùpà is less compositional (cf. Tantucci & Di Cristofaro 2020), as it does not literally refer to the absence of fear of an experiencer, but rather underpins a reasoning process that is more bound to the epistemic domain than the expressive one (see Sweetser 1990: 69–73; Couper-Kuhlen & Thompson 2000; Squartini 2012 on the relationship between concessive moves and epistemicity). It is also more procedural as it encodes the specific function of introducing the protasis of a concessive construction [不怕 bùpà ‘despite’ X, Y ]. In the transition from (11) to (12) 不怕 bùpà clearly undergoes grammaticalization as increased dependency, as it shifts from acting as an independent predicate to a concessive operator. It can felicitously be replaced by 虽然 suīrán ‘although’ or 即使 jíshǐ ‘even if ’, which is not possible in Example (11). This concessive reanalysis is not dissimilar to what happened to a number of protasis operators in many European languages and beyond. Human negative reactions to events, such a ‘spite’, ‘contempt’, ‘regard’, ‘sorrow’, ‘regret’, ‘ingratitude’, ‘thoughtlessness’ and ‘displeasure’, often develop an extended intersubjective (E-I) concessive meaning, expressing an assumed reaction of a generic social persona to the content of the proposition. The concept of ‘fear’ fits nicely into this list, which can be exemplified by concessive markers like in spite of,

61

62

Vittorio Tantucci & Aiqing Wang

despite, regardless in English, a pesar de in Spanish, malgrado in Italian, ondanks and in weervil van in Dutch, huolimatta in Finnish and so on (König 2005: 821). Increased constructional dependency is also at play in the transition from Example (13) to (14) below of the lexeme 只怕 zhǐpà ‘to only be afraid of ’: (13) 我的哥, 快来关门罢! 我只怕老厌物又来缠。 Wǒde gē, kuài lái guān mén bà! Wǒ zhǐpà lǎo yànwù yòu lái chán. My older.brother quickly come close door SFP I probably old bastard again come pester ‘Dear, come in and close the door now, I am just afraid that old bastard may come and pester us.’ CNC / Yuan-Ming Dynasty /初刻拍案惊奇 Chūkè pāiàn jīngqí ‘The first shot of surprises’ (14) 你的脾气再大，只怕也大不过租界的法律。 nǐ de píqǐ zài dà, zhǐ pà yě dà bù guò zūjiè de fǎlǜ your temper increasingly big, only-fear big-not-pass concession law ‘No matter how determinate you are, it will not be enough to stop the concession laws.’ CNC / Fiction / 空谷兰 Kōng gǔ lán ‘Lonely Orchid’

In (14) 只怕 zhǐpà acquires an evaluative illocutionary force as a pre-verbal component in the apodosis of an incremental-hypothetical construction [再 zài ‘more’ X 只怕 zhǐpà ‘probably’ Y ], paraphrasable as more of X is likely to lead to Y. The lexeme here is also less compositional (cf. Tantucci & Di Cristofaro 2020), as the affective component of 只怕 zhǐpà is no longer at issue. This can be tested as the anaphoric expression *我的心情就是这样 wǒ de xīnqíng jiù shì zhèyàng ‘my feeling are just like that’ would be semantically incompatible with the speaker/writer’s evaluation.

2.

Data retrieval

We retrieved our data both from the diachronic and the modern section of the Chinese National Corpus (cncorpus.org11, hereforth CNC), which is a balanced corpus that comprises 7 milion and 19 milion characters respectively. Our study was centred on the fictional section of each diachronic sub-corpus of the CNC due to two reasons. First, only fictional texts are a balanced and constant source across the three periods of our enquiry: Yuan-Ming, Qing and Modern. Secondly, the fictional section has the advantage of providing a controlled environment that 11. Last accessed on 11/04/2022.

Chapter 2. From fear to reason

includes dialogic exchanges that could be representative of the formal and pragmatic features at stake in each sub-period (cf. Tantucci & Wang 2020a). Indeed, data from trials, plays, conversation in novels and letters is quite close to speech, and constitutes a key resource for diachronic investigation (cf. Culpeper & Kytö 2010; Tantucci 2017a, 2017b). Written language is widely acknowledged as a fundamental environment where to investigate language change (Traugott & Dasher 2002: 46) as ‘text provides a mode of speech’ (Olson 1994: xviii). Diachronies of words and constructions display well-attested reflexes in contemporary spoken data (Biber 1988) and changes of form and meaning that can be attested in diachronic re-analysis underpin sedimentation of spontaneous innovation (cf. Traugott & Trousdale 2013). The reason why we started our enquiry from the 元 Yuán Dynasty (1279–1368 AD) is because only few utterances of the three constructions were found during the 隋 Suí and 唐 Táng Dynasties and none prior to that, which would not lead to statistically significant results. Earliest attested apprehensive-epistemic usages of all the three constructions are indeed found during the 唐 Táng Dynasty.

3.

Analysis

Our multifactorial annotation included the source text, the lexeme (恐怕 kǒngpà, 只怕 zhǐpà or 不怕 bùpà), the Dynasties 元/明 Yuán/Míng (1279–1644), 清 Qīng (1644–1911) and Modern Period (1911-Present), whether the usage was epistemic and whether it occurred independently from a larger construction. One line of input of our analysis is given in Table 2 below: Table 2. Input for the diachronic multifactorial analysis of 怕 pà Source

Lexeme

Dynasty

Epistemic

Independent

Example

Chizi

Kongpa

Modern

Y

N

陈再道转[…]

We relied on the criteria for the assessment of the epistemic shift given in Sections 1.3 and 1.4. For the annotation of whether the lexeme would occur independently, we tested whether each entry (恐怕 kǒngpà, 只怕 zhǐpà or 不怕 bùpà) would be used as an internal constituent of a larger construction. An example of this is when 只怕 zhǐpà occurs in the apodosis as part of a larger hypothetical construction, e.g. [若是 ruòshì ‘supposing’ X 只怕 zhǐpà ‘probably’ Y ]:

63

64

Vittorio Tantucci & Aiqing Wang

(15) 只怕，你会失望的。 Zhǐpà, nǐ huì shīwàng de. be-afraid you will disappointed de ‘I am just afraid that you may be disappointed.’ cnc / Modern / 空谷兰 kōnggǔ lán ‘Lonely Orchid’ (16) 这下子若是皇帝一生气，只怕性命难保。 Zhè xiàzi ruòshì huánɡdì yī shēngqì, zhǐpà xìngmìng nánbǎo. This time supposing emperor get angry probably life difficult protect ‘If the emperor gets angry, it is unlikely to remain alive.’ CNC / Modern / 中国后妃列传 Zhōngguó hòufēi lièzhuàn ‘Legends of China’s royal concubines’

In (15) above, Sp/Wr12 expresses her concern about the Ad/H13’s reaction to some possible state of affairs. In this case, 只怕 zhǐpà occurs in isolation, independently from a larger morphosyntactic construction. In (16), on the other hand, 只怕 zhǐpà occurs in a specific slot, introducing an evaluative speech act in the apodosis of a hypothetical construction. In this case, the evaluation 只怕性命难保 zhǐpà xìngmìng nánbǎo ‘it is unlikely to remain alive’ is bound to the preceding structure 若是皇帝一生气 ruòshì huánɡdì yī shēngqì ‘if the emperor gets angry’. The same applies to cases where 不怕 bùpà occurs as an independent lexeme, as in (17) below, or when it functions as a specific constituent of a larger chunk, as in the case of the concessive construction [不怕 bùpà ‘even if ’ X, Y ] in (18): (17) 公子着急说∶“ 难道我不怕吗?” Gōngzǐ zháojí shuō∶ “Nándào wǒ bùpà ma?” Master anxious say surely I not-fear prt The master said anxiously: ‘Aren’t I afraid?’ cnc / Qing Dynasty / 侠女奇缘 Xiánǚ qíyuán ‘Legend of a marital woman’ (18) 老妈妈说: “不怕干部笑话，屋里当真连盐钱也拿不出了。” Lǎo māmā shuō: “Bùpà gànbù xiàohuà, wū lǐ dāngzhēn lián yán qián yě ná bù chū le.” Old woman say not.fear cadre laugh.at house inside really even salt money also take not out sfp ‘The old woman said: “I’m not afraid that you cadres will laugh at me – honestly, there’s not even money for salt in my home.” ’ CNC / Modern / 赤子 Chìzǐ ‘The sincere man’

To study the factors that were involved in the diachronic shift from the expressive to the epistemic domain of 怕 pà constructions we fitted a mixed effects multin12. Speaker/Writer. 13. Addressee/Hearer.

Chapter 2. From fear to reason

omial logistic regression (Baayen & Jaeger 2008). This model produces a binary outcome variable based on fixed factors and random factors. Random factors help to exclude variation that can depend on random or unpredictable circumstances or qualities. This allows the analyst to assess more reliably the remaining fixed factors, which are the ones that can be assumed to have direct effect on the prediction of the outcome variable. We thus fitted a model with: (i) epistemicity as the outcome variable; (ii) dynasty and structural (in)dependence as fixed effects; and (iii) the textual source of the 怕 pà usages as a random effect. The results of our model can be found in Table 3: Table 3. Mixed effects logistic regression of the epistemic shift of all 怕 pà affective constructions Random Effects Groups

Name

Variance

Std. Deviation

Source

(Intercept)

0.629

0.793

Estimate

Std. Error

Z value

Pr( > |t|)

(Intercept)

−3.141

0.494

−6.351

 2.14e-10***

PeriodQing

−0.249

0.559

−0.445

 0.656

PeriodModern

 3.198

0.526

 6.076

 1.23e-09***

IndependentN

−0.433

0.442

−0.98

 0.034

Wordzhipa

 3.579

0.404

 8.848

< 2e-16***

Wordkongpa

 2.066

0.375

 5.503

 3.73e-08***

DynastyQing:IndependentN

 1.1021

0.667

 1.652

 0.098.

DynastyModern:IndependentN

−0.841

0.641

−1.312

 0.189

Fixed Effects

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’

At the top of Table 3 are given the random effects of the model, which include the standard deviation, showing the variability from the predicted values with reference to the source of the text in which 恐怕 kǒngpà, 不怕 bùpà or 只怕 zhǐpà were used. The fixed effects can be found at the lower part of the table. Here, the Estimate column shows the coefficients of the slope for the fixed effects on whether 怕 pà is used epistemically or as a bare apprehensive. The predictors of the epistemic shift are time (i.e. the dynasty in which the construction is used) and whether the 怕 pà lexeme is used independently or as part of a larger construction. From the table we can see that, as a whole, there is a clear tendency of epistemic usages to increase, starting from the 元/明 Yuán/Míng Dynasties

65

66

Vittorio Tantucci & Aiqing Wang

(1279–1644) – which are located at the intercept of the model – up to the Modern period (1911–1949), (β = 3.198, Z = 6.076, p = 1.23e-09). We can also see that both 只怕 zhǐpà (β = 3.579, Z = 8.848, p < 2e-16) and 恐怕 kǒngpà (β = 2.066, Z = 5.503, p = 3.73e-08) tend to have a significant predominance of epistemic usages in contrast with 不怕 bùpà. This seems to indicate that the latter has a stronger persistence (Hopper 1991) of its original meaning, probably due to the negative polarity of the construction. This encodes the denial of a discourse-old or inferable proposition (cf. Prince 1992). That is, lexemes with negative polarity tend to index a propositional meaning rather than a performative one (the latter is at work when 不怕 bùpà is used as a grammaticalised concessive operator, meaning ‘even if ’ or ‘although’). Indeed, 不怕 bùpà in our data often literally negates that the speaker is or might be afraid of something. In Figure 1 are given the predicted probabilities for epistemicity as a whole over the three periods.

Figure 1. Predicted probabilities of epistemicity of 怕 pà lexemes from 1279 up to the present

In the plot, the x axis indicates time, while the y axis expresses the predicted probability that an epistemic reanalysis has taken place. Based on the plot, the first finding of this study is thus that the epistemic reanalysis of 怕 pà lexemes occurred most significantly over the 20th century, despite attested usages are present already during the 唐 Táng Dynasty. This may be connected with the so-called 五四运动 wŭsìyùndòng May Fourth Movement in China (Tantucci & Wang 2020a). Before the 1911, the official written language style in China was the so-called literary Chinese (文言文 wényánwén). Due to its official and prescriptive register, the written style of 文言文 wényánwén led to a mismatch

Chapter 2. From fear to reason

between the written language and the natural development of the spoken language (Pulleyblank 1995). Abrupt formal and pragmatic changes therefore occurred in the written language in between the end of the 19th and the beginning of the 20th century. The 五四运动 wŭsìyùndòng May the Fourth Movement was a political, cultural and anti-imperialist movement that affected the Chinese system of values, together with the language use per se. At this stage, the 文言文 wényánwén style was deemed by intellectuals as a ‘dead language’ and literature had now to be written in vernacular Chinese. There is little doubt that the abrupt change that we can appreciate in Figure 1 since 1911 was profoundly affected by this. At this point, in order to better assess grammatical dependency as a predictive component that may have affected the history of specific constructions, we fitted a new model only focusing on the two lexemes that in our data showed a distinctive tendency towards constructional and grammatical dependency, i.e. 只怕 zhǐpà and 不怕 bùpà. In Table 4 below are reported our results: Table 4. Mixed effects logistic regression of Grammatical reduction of 不怕 bùpà and 只 怕 zhǐpà Random Effects Groups

Name

Variance

Std. Deviation

Source

(Intercept)

0.644

 0.803

Estimate

Std. Error

Z value

Pr( > |t|)

(Intercept)

−0.861

0.566

−0.329

0.742

DynastyQing

−1.944

0.859

−2.262

0.024 *

DynastyModern

 0.455

0.669

 0.680

0.497

IndependentN

−1.284

0.485

−2646

0.008 **

DynastyQing:IndependentN

 3.22

0.787

 4.093

4.27e-05 ***

DynastyModern:IndependentN

 1.794

0.688

 2.607

0.009 **

Fixed Effects

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’

As we can see, in this case while the tendency towards epistemicity as such is no longer linear, as in the 清 Qīng Dynasty (1644–1911) we can even notice a decrease in epistemic usages (β = -1.944, Z = -2.262, p = 0.024) in contrast with the 元/明 Yuán/ Míng period (1279–1644). What is key now is that epistemicity tends to grow significantly in combination with structural dependency, which is both significantly at work during the 清 Qīng Dynasty and the Modern Period. This clearly indicates that the epistemic reanalysis of 只怕 zhǐpà and 不怕 bùpà involved a process of con-

67

68

Vittorio Tantucci & Aiqing Wang

structionalisation in which they started to acquire a procedural function as a specific component of a larger construction, while the opposite has been at play for 恐 怕 kǒngpà. In Figure 2 are reported the predicted probabilities of epistemic reanalysis 只怕 zhǐpà and 不怕 bùpà throughout the three periods whilst acting as dependent (blue dots) vs independent lexemes (red dot).

Figure 2. Predicted probabilities of epistemicity of 不怕 bùpà and 只怕 zhǐpà lexemes from 1279 up to 1949

From Figure 2 we can clearly see that the epistemic reanalysis is not linear when the lexemes are used as an independent construction (red dots). Quite differently, it grows incrementally through time in the form of grammatical dependence (GD), from 20% in the 元/明 Yuán/Míng period (1279–1644), to around 45% during the 清 Qīng Dynasty (1644–1911), to finally around 70% during the Modern period. This is an important finding because it confirms that grammaticalization is not necessarily bound to increased dependency and/or reduction, but can unfold in different forms even when originating from the same semantic source. This is the case of the Mandarin 怕 pà apprehensives, which all follow a clear trajectory towards the epistemic domain of reason out the expressive one of fear. However two apprehensive lexemes that include 怕 pà (不怕 bùpà and 只怕 zhǐpà)

Chapter 2. From fear to reason

went through a process of grammaticalization as increased dependency (DP), while one of them (恐怕 kǒngpà) underwent a process of grammatical expansion (GE). Importantly, this may not be the end of this story, as a new stage of increased dependency may arise (or have been arising) as a result of chunking (cf. Bybee 2010; Tantucci 2021), i.e. the automatisation of linguistic – as well as other kinds of – behaviours that are repeated in the same order. Repetition leads formal reduction and/or omission, so that shared meaning is dialogically accessed more quickly and produced more efficiently (Tantucci & Wang 2021, 2022a, 2022b, 2022c). In the present case, chunking may involve the conventionalised ellipsis of either the protasis or the apodosis of a concessive/hypothetical construction, leading to a new isolated use of 不怕 bùpà and 只怕 zhǐpà as independent epistemic operators. A specific corpus analysis centred on the Mandarin development from 1949 up to today may shed some important light on this hypothesis. All in all, the findings of this paper posit a challenge for an entirely reductionist view of grammatical change. While grammaticalization is necessarily bound to an increased procedural usage of language, this may yet underpin either the narrowing down of form and meaning or their expansion. More importantly, such processes may both occur at different stages of the grammatical change of similar constructions and even out of the same original lexeme, as the present study has shown.

4.

Conclusions

The present paper supports and reinforces the view that grammatical change is not necessarily bound to increased dependence and/or reduction of form and meaning. We focused on the change of the apprehensive lexeme 怕 pà as part of three different constructions, all undergoing a shift from a source domain of fear to a target domain of epistemicity. Despite a similar semantic trajectory, we showed that the grammatical change of 怕 pà – depending on the construct it is part of – has been both moved in the direction of grammatical dependency (GD), as well as in the direction of grammatical expansion (GE). These results aim to increase the awareness that constructional change and grammaticalization include a wider range of phenomena than it was originally assumed. They also suggest that grammatical change as dependency and expansion are not necessarily mutually exclusive, as they may be at work at different stages of the development of the same lexeme. This last point may be corroborated with evidence showing that either or both 不怕 bùpà and 只怕 zhǐpà may have gone through a new process of grammatical expansion and have become more syntactially autonomous during the 20th century.

69

70

Vittorio Tantucci & Aiqing Wang

References Austin, Peter. 1981. A Grammar of Diyari, South Australia [Cambridge Studies in Linguistics 32]. Cambridge: CUP. Benveniste, Emile. 1971[1958]. Subjectivity in language. In Problems in General Linguistics, 223–230. Trans. by Mary Elizabeth Meek. Coral Gables FL: University of Miami Press. Boogaart, Ronny. 2020. Expressives in argumentation: The case of apprehensive straks (‘shortly’) in Dutch. In From Argument Schemes to Argumentative Relations in the Wild. A Variety of Contributions to Argumentation Theory, Frans H. van Eemeren & Bart Garssen (eds), 185–204. Cham: Springer. Bybee, Joan L. 2010. Language, Usage and Cognition. Cambridge: CUP. Bybee, Joan L., Perkins, Revere D. & Pagliuca, William. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago IL: The University of Chicago Press. Couper-Kuhlen, Elizabeth & Thompson, Sandra A. 2000. Concessive patterns in conversation. In Cause, Condition, Concession, Contrast: Cognitive and Discourse Perspectives, Elizabeth Couper-Kuhlen & Bernd Kortmann (eds), 381–410. Berlin: Mouton de Gruyter. Dictionary of Classical Chinese. (2002). 古代汉语词典 Gudai Hanyu Cidian (Dictionary of Classical Chinese). Beijing: Commercial Press. Dixon, Robert M. W. 1977. A Grammar of Yidin [Cambridge Studies in Linguistics 19]. Cambridge: CUP. Dixon, Robert M. W. 1988. A Grammar of Boumaa Fijian. Chicago IL: The University of Chicago Press. Ekman, Paul. 1984. Expression and the nature of emotion. In Approaches to Emotion, Klaus R. Scherer & Paul Ekman (eds), 319–344. Hillsdale NJ: Lawrence Erlbaum Associates. Ekman, Paul. 1994. Strong evidence for universals in facial expressions: A reply to Russell’s mistaken critique. Psychological Bulletin 115(2): 268–287. Fillmore, Charles, Kay, Paul & O’Connor, Mary Catherine. 1999. Regularity and idiomaticity in grammatical constructions: The case of let alone. Language 64(3): 501–538. Fleischman, Suzanne. 1982. The Future in Thought and Language. Diachronic Evidence from Romance. Cambridge: CUP. Gao, Zeng Xia. (2003). 汉语担心一认识情态词 “怕” “看” “别” 的语法化 Hanyu danxin “pa” “kan” “bie” de yufahua [The grammaticalization of the Chinese epistemic-apprehensives “pa”, “kan” “bie”]. The Journal of the University of Chinese Academy of Social Sciences 1: 97–112. Givón, Talmy. 2001. Syntax: An Introduction. Amsterdam: John Benjamins. Goldberg, Adele. 1995. A Construction Grammar Approach to Argument Structure. Chicago IL: The University of Chicago Press. Goldberg, Adele. 2006. Constructions at Work: The Nature of Generalization in Language. Oxford: OUP. Haiman, John. 1980. Hua: A Papuan Language of the Eastern Highlands of New Guinea [Studies in Language Companion Series 5]. Amsterdam: John Benjamins. Haspelmath, Martin. 1999. Optimality and diachronic adaptation. Zeitschrift für Sprachwissenschaft 18(2): 180–205.

Chapter 2. From fear to reason

Haspelmath, Martin. 2004. On directionality in language change with particular reference to grammaticalization. In Up and Down the Cline – The Nature of Grammaticalization [Typological Studies in Language 59], Olga Fischer, Muriel Norde & Harry Perridon (eds), 17–44. Amsterdam: John Benjamins. Heine, Bernd. 2018. Are there two ways of approaching grammaticalization? In New Trends in Grammaticalization and Language Change [Studies in Language Companion Series 202], Sylvie Hancil, Tine Breban & José Vicente Lozano (eds), 23–54. Amsterdam: John Benjamins. Himmelmann, Nikolaus P. (2004). Lexicalization and grammaticization: Opposite or orthogonal? In What Makes Grammaticalization: A Look from its Components and its Fringes, Walter Bisang, Nikolaus P. Himmelmann & Björn Wiemer (eds), 21–42. Berlin: Mouton de Gruyter. Hopper, Paul. 1991. On some principles of grammaticalization. In Approaches to Grammaticalization, Vol.1: Theoretical and Methodological Issues [Typological Studies in Language 19:1], Elizabeth Closs Traugott & Bernd Heine (eds), 17–35. Amsterdam: John Benjamins. Jing-Schmidt, Zuo & Kapatsinski, Vsevolod. 2012. The apprehensive: Fear as endophoric evidence and its pragmatics in English, Mandarin, and Russian. Journal of Pragmatics 44(4): 346–373. König, Ekkehard. 2005. Concessive clauses. In The Encyclopedia of Language and Linguistics, Keith Brown (ed.), 820–824. Oxford: Elsevier. Langacker, Ronald W. 1987. Foundations of Cognitive Grammar: Theoretical Prerequisites. Sandford CA: Stanford University Press. Langacker, Ronald W. 2008. Cognitive Grammar: A Basic Introduction. Oxford: OUP. Lehmann, Christian. 1994. Predicates: Aspectual types. In The Encyclopedia of Language and Linguistics, Ronald E. Asher & James M. Y. Simpson, 3297–3302. Oxford: Pergamon. Lehmann, Christian. 1995. Thoughts on Grammaticalization. Munich: Lincom. Lehmann, Christian. 2004. Theory and method in grammaticalization. In Grammatikalisierung. Gabriele Diewald (ed.). Special issue of Zeitschrift für Germanistische Linguistik 32: 152–187. Lichtenberk, Frantisek. 1991. Semantic change and heterosemy in grammaticalization. Language 67(3): 475–509. Lichtenberk, Frank. 1995. Apprehensional epistemics. In Modality in Grammar and Discourse [Typological Studies in Language 32], Joan L. Bybee & Suzanne Fleischman (eds), 293–328. Amsterdam: John Benjamins. Liu, Xiaoying. 2011. 词义的历时演变与对外汉语词汇教学 – 以“恐怕”为例 Ciyi de lishi yanbian yu duiwai hanyu cihui jiaoxue (The diachronic evolution of lexical meaning and TCFL vocabulary teaching – Take ‘Kongpa’ as an example). Journal of Hebei University (Philosophy and Social Science) 3: 61–64. Mathiassen, Terje. 1996. A Short Grammar of Lithuanian. Bloomington OH: Slavica. Mitkovska, Liljana, Bužarovska, Eleni & Ivanova, Elena. 2017. Apprehensive-epistemic daconstructions in Balkan Slavic. Slověne = Словѣне. International Journal of Slavic Studies 6(2), 57–83. Modern Chinese Dictionary. 2016. 现代汉语词典 Xiandai hanyu cidian (Modern Chinese Dictionary). Beijing: Commercial Press.

71

72

Vittorio Tantucci & Aiqing Wang

Narrog, Heiko. 2012. Beyond intersubjectification: Textual uses of modality and mood in subordinate clauses as part of speech-act orientation. English Text Construction 5(1): 29–52. Niculescu, Dana. 2013. Romanian passive participles as complements of perception verbs. Revue Roumaine de Linguistique LVIII(1): 55–71. Norrick, Neal R. 1978. Expressive illocutionary acts. Journal of Pragmatics 2(3): 277–291. Ōta, Tatsuo. 2003. 中国语历史文法 Zhongguoyu lishi wenfa (Survey of the Historical Syntax of Chinese). Beijing: Peking University Press. Palmer, Frank R. 2001. Mood and Modality. Cambridge: CUP. Prince, Ellen F. 1992. The ZPG letter: Subjects, definiteness, and information-status. In Discourse Description: Diverse Analyses of a Fundraising Text [Pragmatics & Beyond New Series 16], Sandra A. Thompson & William Mann (eds), 295–325. Amsterdam: John Benjamins. Pulleyblank, Edwin G. 1995. Outline of Classical Chinese Grammar. Vancouver BC: University of British Columbia Press. Sweetser, Eve. 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: CUP. Searle, John R. 1976. A classification of illocutionary acts. Language in Society 5(1): 1–23. Sheng, Lichun. 2008. “大概”、“也许” 和 “恐怕” 的语义、语用分析 Dagai yexu he kongpa de yuyi yuyong fenxi [Semantic and pragmatic analysis of modal adverbs “dagai”, “yexu” and “kongpa”]. Chinese Language Learning 1: 45–51. Sinclair, John McH. 1996. The search for units of meaning. Textus 9: 75–106. Squartini, Mario. 2012. Evidentiality in interaction: The concessive use of the Italian future between grammar and discourse. Journal of Pragmatics 44(15): 2116–2128. Tantucci, Vittorio. 2013. Interpersonal evidentiality: The Mandarin V-过 guo construction and other evidential systems beyond the ‘source of information’. Journal of pragmatics 57: 210–230. Tantucci, Vittorio. 2015a. Epistemic inclination and factualization: A synchronic and diachronic study on the semantic gradience of factuality. Language and Cognition 7(3): 371–414. Tantucci, V. (2015b). Traversativity and grammaticalization: The aktionsart of 过 guo as a lexical source of evidentiality. Chinese Language and Discourse, 6(1), 57–100. Tantucci, Vittorio. 2016a. Toward a typology of constative speech acts: Actions beyond evidentiality, epistemic modality, and factuality. Journal Intercultural Pragmatics 13(2). Tantucci, Vittorio. 2016b. Textual factualization: The phenomenology of assertive reformulation and presupposition during a speech event. Journal of Pragmatics 101: 155–171. Tantucci, Vittorio. 2017a. From immediate to extended intersubjectification: A gradient approach to intersubjective awareness and semasiological change. Language and Cognition 9(1): 88–120. Tantucci, Vittorio. 2017b. An evolutionary approach to semasiological change: Overt influence attempts through the development of the Mandarin 吧-ba particle. Journal of Pragmatics 120: 35–53. Tantucci, Vittorio. 2021. Language and Social Minds: The Semantics and Pragmatics of Intersubjectivity. Cambridge: CUP.

Chapter 2. From fear to reason

Tantucci, V., & Di Cristofaro, M. (2020). Entrenchment inhibition: Constructional change and repetitive behaviour can be in competition with large-scale “recompositional” creativity. Corpus Linguistics and Linguistic Theory, 16(3), 547–579. Tantucci, Vittorio & Di Cristofaro, Matteo. 2021. Pre-emptive interaction in language change and ontogeny: The case of [there is no NP]. Corpus Linguistics and Linguistic Theory 17(3): 715–742. Tantucci, Vittorio & Wang, Aiqing. 2018. Illocutional concurrences: The case of evaluative speech acts and face-work in spoken Mandarin and American English. Journal of Pragmatics 138: 60–76. Tantucci, Vittorio & Wang, Aiqing. 2020a. Diachronic change of rapport orientation and sentence-periphery in Mandarin. Discourse Studies 22(2): 146–173. Tantucci, Vittorio & Wang, Aiqing. 2020b. From co-actions to intersubjectivity throughout Chinese ontogeny: A usage-based analysis of knowledge ascription and expected agreement. Journal of Pragmatics 167: 98–115. Tantucci, Vittorio & Wang, Aiqing. 2021. Resonance and engagement through (dis-)agreement: Evidence of persistent constructional priming from Mandarin naturalistic interaction. Journal of Pragmatics 175: 94–111. Tantucci, Vittorio & Wang, Aiqing. 2022a. Resonance as an applied predictor of cross-cultural interaction: Constructional priming in Mandarin and American English interaction. Applied Linguistics 43(1): 115–146. Tantucci, Vittorio & Wang, Aiqing. 2022b. Dialogic priming and dynamic resonance in Autism: Creativity competing with engagement in Chinese children with ASD. Journal of Autism and Developmental Disorders, 1–17. Tantucci, Vittorio & Wang, Aiqing. 2022c. Dynamic resonance and explicit dialogic engagement in Mandarin first language acquisition. Discourse Processes 59(7): 553–574. Traugott, Elizabeth Closs. 1989. On the rise of epistemic meanings in English: An example of subjectification in semantic change. Language 65(1): 31–55. Traugott, Elizabeth Closs. 2016. On the rise of types of clause-final pragmatic markers in English. Journal of Historical Pragmatics 17(1): 26–54. Traugott, Elizabeth Closs & Trousdale, Graeme. 2013. Constructionalization and Constructional Changes. Oxford: OUP. Van Olmen, Daniel. 2019. A diachronic corpus study of prenominal zo’n ‘so a’in Dutch: Pathways and (inter) subjectification. Functions of Language 26(2): 216–247. von Wright, Georg Henrik. 1951. Deontic logic. Mind 60(237): 1–15. Wierzbicka, Anna. 1999. Emotions across Languages and Cultures: Diversity and Universals. Cambridge: CUP. Xinhua Dictionary. 2004. 新华字典 Xinhua Zidian (Xinhua Dictionary). Beijing: Commercial Press. Yang, Bei. 2016. 汉语认识情态表达 “可能” “也许” “大概 ” “恐怕” 用法对比分析 Hanyu renshi qingtai biaoda keneng yexu dagai kongpa duibi fenxi (A corpus-based analysis of Chinese epistemic modals: Constrast among ‘Kěnéng’, ‘Yěxǔ’, ‘Dàgài’ and ‘Kǒngpà’). Journal of Guangdong University of Foreign Studies 27(6): 33–41. Yap, Foong Ha, Chor, Winnie & Wang, Jiao. 2012. On the development of epistemic ‘fear’ markers: An analysis of Mandar in kongpa and Cantonese taipaa. In Covert Patterns of Modality, Werner Abraham & Elisabeth Leiss (eds), 312–342. Newcastle upon Tyne: Cambridge Scholars.

73

chapter 3

-maɾa in Mara On the ongoing grammaticalization(s) of a Bantu ‘finish’ verb Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio University of Helsinki

This study accounts for the grammaticalization of the verb -maɾa ‘finish’ in the language varieties of the Western Serengeti branch of the Mara subgroup of Eastern Bantu. In these varieties, -maɾa is found in two different auxiliary verb constructions marking the two interrelated Phasal Polarity concepts of already and not yet. Whereas ‘finish’ in general and -maɾa cognates in particular are commonly grammaticalized into already in Bantu, the parallel development of a not yet ‑marker from the same source constitutes a much rarer development. Nonetheless, we propose that both constructions developed out of the same need to reinforce a reading of (non-)completion (and counter-factuality) and that they instantiate recurrent cyclic patterns of change significant for Eastern Bantu more broadly. Keywords: Bantu, Mara, auxiliary, phasal polarity, grammaticalization

1.

Introduction

The purpose of this paper is to offer a comprehensive account of two auxiliary verb constructions (AVCs) and the (ongoing) and interrelated processes of grammaticalization of which they are a part. These AVCs both involve the verb -maɾa ‘finish’ and occur in four Bantu language varieties spoken in the Mara region of northern Tanzania. In all of these language varieties – grouped together and referred to as Western Serengeti (WS) – the auxiliary verb -maɾa marks two versatile yet conceptually interrelated notions of phasal polarity (PhP), namely

https://doi.org/10.1075/slcs.232.03ber © 2023 John Benjamins Publishing Company

Chapter 3. -maɾa in Mara

already and not yet.1 Example (1) from Nata illustrates the use of the former and example (2) from Ishenyi the latter.2 (1) Nata aβá-án βá-nɔ m-bá-mare ɣu-sáɾɛk-a 2-children 2-dem foc-sm2-already.pfv inf-ruin-neut-fv ‘These children are already ruined.’ (2) Ishenyi ɣa-tá-ámaɾa ɣa-tɛɾɛk-u sm6-neg-not.yet sm6-bake-pass ‘They aren’t baked yet.’

The linguistic concept of phasal polarity (PhP) refers to “structured means of expressing polarity in a sequential perspective” (Van Baar 1997: 40; Kramer 2017). It encompasses a linguistic system consisting of four interrelated expressions denoting opposing phasal values and polarity, which in addition to already and not yet also include the expressions for still and no longer. In this systematization, already and not yet are interrelated in that the latter equals the externally negated variant of the former. In other words, “not yet P = (NOT (already P))”. In addition to the neutral notions of temporal-sequentially related phases, PhP expressions are often also associated with pragmatic and (inter-)subjective notions of counter-factuality or counter-expectation.3 This means that in the case of already, the event commenced earlier than expected, while its commencement is later than expected in the case of not yet. For example, in the utterance with already in (3), the speaker articulates his/her surprise over the fact that the time has elapsed faster than expected, while in the utterance with not yet in (4), s/he intends to counter the addressee’s apprehension of having arrived late at church.

1. These expressions have been assigned a plethora of different labels over the years. Most recent in this regard are the terms iamitive for already (cf. Dahl & Wälchli 2016; Dahl 2021) and nondum for not yet (Veselinova & Devos 2021). To avoid getting lost in terminology and partly overlapping and conflicting definitions, in this paper we will refer to these linguistic concepts merely with their English translations, including in the glossing of the examples. 2. The characteristics of the verb forms of which auxiliary -maɾa is a part and – in connection to this – the choice of interlinear segmentation of the examples, will be explained in later parts of this paper. 3. See, inter alia, Van der Auwera (1993: 620–622, 1998), Van Baar (1997: 27–35) and Kramer (2017) for elaborate discussions and illustrations of the contrast between pragmatically neutral and counter-factual uses of PhP markers; see also Schadeberg (1990) and Heine et al. (1991) for early accounts of this fact with specific reference to Bantu/African languages.

75

76

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

(3) Ikoma mbe wáɾjoβa e-eká-maɾ-a ko-hík-a t͡ʃa-sáá i-kɔ́mi oh God sm9-nucl-already-fv inf-reach-fv 10-time 10-ten ‘My God, it is already ten o’clock!’ (4) Nata u-tá-tiin-a tu-tá-ámaɾa tu-sɛɛm-a ɣu-sɛ́ ŋɡɛɾ-a sm2sg-neg-fear-fv sm1pl-neg-not.yet sm1pl-say-fv inf-praise-fv ‘Don’t worry, we have not said praises (i.e. started mass) yet.’

While the development of a ‘finish’ verb like -maɾa into a PhP marker already – and thence into interrelated concepts such as perfect, perfective aspect and past tense – is a well-known grammaticalization process along the verb-to-auxiliaryto-affix chain in Bantu (cf. Nurse 2008: 305; Nicolle 2012), as well as crosslinguistically (Bybee et al. 1994: 55–61, 70–73, 81–87; Kuteva et al. 2019: 173–180), there are far fewer accounts (if any?) of a parallel grammaticalization process where a ‘finish’ verb develops into marking not yet.4 In addition, although already and not yet are conceptually and paradigmatically related and are expressed with the same lexical verb as an auxiliary, they occur in very different constructional frames, which raises the question of the extent to which these constructions are historically related. In this paper, we argue that both auxiliary constructions with -maɾa were recruited to fill a functional gap brought about by the concomitant functional shifts / losses of the forms originally covering already and not yet. PhP markers are inherently focused, and the study suggests that the motivation for the recruitment of -maɾa into a PhP auxiliary was driven by the need to reinforce readings of (non-)completion (and counter-factuality). More specifically, we argue that the developments of both AVCs in WS are indicative of broader and recurrent patterns of diachronic cycles. From a methodological perspective, the paper aims, on the one hand, to showcase how reconstructional work established from grammaticalization studies may be employed on un(der)-documented language varieties such as the WS ones. On the other hand, it sets out to demonstrate how a grammaticalization framework may form a practical tool for the analysis and description of un(der)-documented 4. Here, it also suffices to mention that the verb -maɾa is not only employed as PhP marker in WS. It has also been recruited, in different constructional contexts, to express other functions too. Thus, it also occurs as an adversative coordinator ‘but’ and as a tail-head linkage marker. The not yet construction with -maɾa also has a dedicated function in subordinate contexts, marking successional ‘before’ clauses. To investigate these additional functionalities of -maɾa and their relationship to their usages as PhP markers is beyond the scope of this paper, but it would serve as an interesting topic for further research.

Chapter 3. -maɾa in Mara

varieties, where synchronic alternations and idiosyncrasies become more easily interpretable once they are seen in the light of the diachronic forces which have shaped them. In reaching these goals, the remainder of the paper will be organized in the following manner. In §2, we address the Mara language ecology and genealogy with a particular focus on WS, including a brief structural overview of these language varieties. In §3, we zoom in to the verb -maɾa and its semasiological background and in §4 we describe the two PhP constructions of which it forms a part. In §5, we account for semantic and formal indications of the grammaticalization of the two constructions, showing that they instantiate very different processes of change. In §6, we discuss the motivation and mechanism behind these grammaticalization processes and relate these to recurrent cycles of change significant in (Eastern) Bantu. A short summary and some final concluding remarks in §7 close this chapter.

2.

On Mara: Geography, genealogy and typology

As seen in Figure 1, among its many referential functions, the noun Mara may refer both to a geographically confined political region in northern Tanzania and a genealogical grouping of (Eastern) Bantu language varieties spoken both in Tanzania and across the border in Kenya. The Mara (formerly Musoma) region is situated in the northern parts of Tanzania, east of Lake Victoria and west of the Serengeti National Park, bordering Kenya to the north. The region is linguistically dense. Besides WS and several other Bantu language varieties, it also contains the Western Nilotic Luo and the Southern Nilotic Datooga languages. The impact of the national and socially dominant language Swahili (also a Bantu language) was noted for the area already by Heine (1976) and has since then had an increasing influence on the linguistic practices of the Mara communities. The younger generations, in particular, are all bilingual with Swahili and it would appear that Swahili is rapidly increasing in use in relation to the local minority language varieties. The Bantu family as a whole is a huge yet tightly knit language family – often compared to a large dialect continuum or spread zone (Güldemann 2018) – consisting of roughly 500 varieties spoken throughout a large part of Sub-Saharan Africa. Genealogically, Mara refers to a grouping within the East Nyanza branch of Great Lakes Bantu (Schoenbrun 1990; Nurse 1999; Nurse & Philippson 2003; Aunio et al. 2019), in turn a subpart of Eastern Bantu (Grollemund et al. 2015). In the traditional alpha-numeric referential system of Bantu languages, consisting of zones labelled through letters, further divided into groups marked by decimals

77

78

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

Figure 1. Map of the Mara languages and their neighbors

(and with individual languages indicated with basic numbers), the Mara group roughly corresponds to the JE40 languages of zone J (cf. Guthrie 1948, 1967–1971; Maho 2003, 2009).5 The Mara branch itself is split into North and South Mara. The Western Serengeti (WS) group – i.e. the focus of this study – forms a sub-branch of South Mara. The group consists of Ngoreme together with Ikoma, Ishenyi and Nata.6 5. These codes are primarily geographically based. Yet, the languages spoken around Lake Victoria, originally treated as belonging to the D and E-zones, have been re-classified on genealogical grounds into a new zone “J” (cf. Philippson & Grollemund 2019). As a consequence, these varieties are typically referred to with J as their first letter followed by their original classificatory letter, hence the doubled-lettered code “JE40” referring to the Mara group. 6. Ngoreme is coded with the ISO 639-3 code ngq / glottocode ngur1263, whereas the other varieties are all jointly coded with the ISO 639-3 code ntk and with the glottocode ikom1245 for

Chapter 3. -maɾa in Mara

The WS varieties are closely related, with an estimated lexical overlap of up to 85% between Ikoma, Nata and Ishenyi and 77% between Ngoreme and Ikoma (Roth 2018: 12; Hill et al. 2007). However, plenty of structural (micro-)variation between the varieties has been attested (see, e.g., Roth & Gibson 2019; Bernander & Laine 2020; Aunio 2017). Moreover, some features found in WS are quite striking from a Bantu comparative perspective, including ATR vowel harmony, final vowel deletion, a reduced tense system and the presence of uncanonical constituent orderings in certain pragmatic contexts. These exceptions aside, however, WS languages behave much like typical (Eastern) Bantu languages. They are agglutinative, with an elaborate noun class system and with morphologically intricate verb morphology. The noun class systems consist of up to 20 noun classes (with some minor variation in the inventory and productivity of certain classes), in control of an agreement system affecting the inflectional patterns of both nominal constituents and verbs. Apart from nominal indexation, the verbal word may contain a versatile set of both inflectional and derivational morphology. The verbal “morphotaxis” of WS, consisting of morphological slots dedicated to various functions, can be sketched as in Figure 2.7 foc8-(/ neg-) + sm- + tam-/neg- + tam- + om- + B(ase) + -tam + = loc Figure 2. The Bantu verbal template as it applies to WS

Note that vowel coalescence may occur across these morpheme borders in WS. All the WS varieties are also tonal, with lexical tones in the nominal system and grammatical tones for verbal inflection. Their tone systems are restricted and privative in character, with an opposition between H (high tones) – marked with an acute accent in this chapter – and Ø (toneless), left unmarked (see Aunio 2017 and further references therein for more on the prosodic characteristics of the WS varieties). Ikoma and Nata and isse1238 for Ishenyi. Within the traditional alpha-numeric coding system of Bantu languages Ngoreme is coded as JE401 whereas Ikoma, Ishenyi and Nata are all coded as JE45. 7. Abbreviations used in this article generally follows those outlined in the Leipzig Glossing Rules. The exceptions include the following: 1, 2, 3 etc. for noun classes; FV = Final Vowel; NEUT = Neuter; NUCL = Nucleative; OM = Object marker; SM = Subject Marker. ⁰ = Tentative reconstruction * = Reconstruction ** = Ungrammatical construction. 8. All WS varieties make use of what historically was most probably more of a bona fide focus prefix, formally realized as a homorganic nasal /N-/. Today, its function is more opaque. In short, it can be characterized as carrying the inverse function of a focus marker. That is, rather than marking predicate focus, its absence marks that the predicate verb is extra-focal. Nonetheless, it will be glossed as FOC in this chapter.

79

80

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

3.

On -maɾa: A lexical verb commonly involved in Bantu grammaticalization

The verb in focus in this paper, i.e. -maɾa, carries the lexical meaning ‘finish’ in all of WS, and is typically used as a transitive verb. When acting as a transitive verb, and in collocation with a second complement verb, -maɾa, still with the lexical meaning of ‘finish’, can be said to function as an “aspectualizer” (Freed 1979; Binnick 1991). This means that it highlights a distinct terminative phase of the event expressed by a dynamic complement verb, like ‘eat’ in (5).9 (5) Ishenyi n-a-marɛ ku-ɾááɣiɾ-a oβu-kíma foc-sm1-finish.pfv inf-eat-fv 14-ugali ‘He has finished eating ugali ( = stiff porridge).’

As shown by Roth (2018: 42, 69, 70, 104), -maɾa as an aspectualizer in WS may in fact be used with some punctive lexical verbs (i.e., verbs with a punctual “nucleus” or main phase), although they lack a terminative phase in their internal semantic structure. In these cases, -maɾa may contribute the “missing” terminative phase to the event described in the construction. A case in point is the inceptive punctive verb ‘arrive’, as seen in (6). (6) Ngoreme n-a-máre ko-hík-a foc-sm1-finish.pfv inf-arrive-fv ‘S/he has finished arriving.’

(Roth 2018: 69)

Otherwise – as expected from a ‘finish’ verb more generally (see Freed 1979: 135) and as also described for other Tanzanian Bantu languages such as Nyakyusa (Persohn 2017) and Nyamwezi (Kanijo 2019) – lexical -maɾa is banned from selecting punctive verbs, in particular those expressed with a non-volitional, nonagentive subject marker, such as ‘shoot’ and ‘see’ in (7) and (8) below. (7) Ishenyi ni-marɛ ku-ɾóɾ-a kilimant͡ʃáɾo sm1sg-finish.pfv inf-see-fv Kilimanjaro **‘I have finished seeing Mt. Kilimanjaro.’

9. In all the examples of this section, the lexical use of -maɾa is illustrated with the verb inflected in a perfect(ive) verb form whose formal and semantic peculiarities will be further addressed in §4.1.

Chapter 3. -maɾa in Mara

(8) Ishenyi na-a-marɛ ku-ɾás-u ɾi-sáási foc-sm1-finish.pfv inf-shoot-pass 5-bullet **‘She has finished being shot.’

The verb -maɾa is a reflex of the verb root *-màd- ‘finish’, reconstructible for ProtoBantu with cognates attested in most Bantu zones (see Bastin et al. 2002). As mentioned in §1, the recruitment of reflexes of *-màd- (along with other ‘finish’ verbs) into an auxiliary, and at a later stage into a prefix, marking already and interrelated concepts such as perfect, perfective aspect and past tense, constitutes a widespread, recurrent and thus well-known grammaticalization trajectory in Bantu. Ultimately, this situation can be seen to be tied in with the fact that while means of verbal inflection in Bantu languages have changed rapidly and frequently since Proto-Bantu, the patterns themselves and the functional categories expressed have remained relatively stable (Güldemann 1996, 1999; Nurse 2008). Similarly, patterns often re-occur across the vast Bantu family. Thus, it may not come as a surprise that verbs cognate with -maɾa are shown to have been recruited as auxiliaries in identical constructional frames to express similar functions in many sister languages to WS. Indeed, just through a quick glance at Tanzanian-Kenyan Bantu alone, we find a dispersed set of language varieties outside of the Mara group exhibiting grammaticalized variants of *-màd- at different levels of functional and formal maturation. This is illustrated in Table 1 below (see also Nurse 2008: 305).10 Following Robbeets & Cuyckens (2013), we may refer to these parallel grammaticalization processes as “parallel drift” or “Sapirian drift”, as they (have) occur(red) independently in several Bantu languages, although under the influence of their common origins and even with the same source material. For WS in particular, however, it cannot be ruled out that the grammaticalization of the AVC with -maɾa into already constitutes a shared innovation, a process that had at least partially started before proto-WS – or even some earlier ancestor group such as proto-South Mara – came to split up. This is because the use of -maɾa in a similar constructional frame expressing already is found in all of South Mara and also in some of the closer North Mara varieties, as indicated in Table 2. The AVC with -maɾa expressing not yet, on the other hand, seems to be a unique trait of WS. As this construction occurs in all four varieties while at the same time containing archaic morphology (see §5.2), it arguably forms a shared

10. In §6.3, Swahili will be discussed more extensively, as the language has not only developed an already-gone-perfect(ive) marker from * -màd-, but it has also developed yet another already marker from another ‘finish’ verb.

81

82

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

Table 1. Examples of grammaticalized markers originating from reflexes of *-màd- in Tanzanian and Kenyan Bantu Variety

Construction

Function

Reference

Ha

mar + pfv # INF

already

Harjula (2004: 109)

Manda

mal + pfv # INF

already

Bernander (2017: 207–217; 2021)

Nyamwezi

mal + pfv # INF

already

Jonsson (1954: 71)

Dawida-Kasigau

meri + pfv # INF

already

Nurse (1979: 254)

Luhya

mal + pfv # INF

already / perfect

Appleby (1961: 114)

Chaga (Vunjo)

maa-

perfective

Moshi (1994)

Chaga (Rombo)

me-

perfect

Nurse (1979: 254–255)

Swahili

me-

perfect(ive)

Ashton (1947)

Table 2. AVC with -maɾa as a marker of already in the Mara subgroup Subgroup S. Mara

N. Mara

Variety

Reflex of [*-màd- + pfv # INF] as already

Reference

WS

Ikoma, Nata, Ishenyi, Ngoreme

yes

[This chapter]

S.W. Mara

Ikizu

yes

Walker (2013: 179)

Zanaki

yes

Walker (2013: 177)

Kabwa

yes

Walker (2013: 170)

Simbiti

yes

Walker (2013: 173)

Kuria

no

Charwi (p.c.), Sillery (1936), Cammenga (2004)

Gusii

no

Cammenga (2002), Whiteley (1960)

innovation unique to WS, which makes it a defining feature of WS as a genealogical grouping. In sum, given the characterization above, it is perhaps not surprising that ‑maɾa in WS has also grammaticalized into marking already. At the same time, the corresponding development with *-màd-, or any other Bantu ‘finish’ verb for that matter, into a not yet marker as witnessed in WS is, as far as we know, previously unattested in Bantu (see Veselinova & Devos 2021).

Chapter 3. -maɾa in Mara

4.

The two auxiliary verb constructions with -maɾa

This section offers a further description of the constructional and functional characteristics of the two AVCs with -maɾa in WS, starting with already (§4.1) and continuing with not yet (§4.2). Anderson (2006, 2011: 2), largely influenced by Heine (1993), defines an auxiliary (verb) as “a verbal element on a diachronic form-function continuum standing between a fully lexical verb and a bound grammatical affix”. He goes on to define an auxiliary verb construction (AVC) as a mono-clausal complex construction consisting (minimally) of an auxiliary verb operating on a lexical predicate verb. Following these definitions, -maɾa in WS arguably acts as an auxiliary verb in two different AVCs, expressing two different functions. As further explained in this section, the two PhP constructions also relate to two different subtypes of AVCs, based on Anderson’s formal subdivision of auxiliary formations in terms of inflectional “headedness”. This formal difference further implies the instantiation of two different channels of formal grammaticalization along the verb-to-affix route.

4.1 Expressing already already is a positive PhP expression, conveying a retrospective focus on an earlier realization of an event that still holds. With regard to counter-exceptionality, already may also be used to convey that this realization commenced earlier than expected. When functioning as an already marker, -maɾa can be inflected in any finite perfect(ive) verb form, while it selects and operates on a second, infinitive verb which conveys the main part of the event. In examples (9) and (10), -maɾa acts as an already marker which occurs in the perfect(ive) verb form consisting of the suffix -iɾE. (9) Ngoreme a-mare ko-hík-a wa-íto hano sm1-already.pfv inf-arrive-fv 23-poss1pl here ‘She has already arrived here at our place.’ (10) Nata nááβa ní-maɾ-iɾɛ ku-β-á taɣitááɾi no, sm1sg-already-pfv inf-be(come)-fv 9.doctor ‘No, I am already a doctor ~ I have already become a doctor.’

83

84

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

Notice that while -maɾa retains its full form in (10), in (9) it has coalesced with the perfective suffix ‑iɾE. The curious result is that the two flaps of the verb root and the suffix have turned into an articulatorily more complex trill /r/. From a semantic point of view, it may furthermore be noticed that the temporal reading in these constructions may range from ‘have already’ to ‘be already’. Ultimately, this fact goes back to a common Bantu feature where change-of-state verbs like ‘be(come) a doctor’ in (10) above often gain a (present) resultative reading when inflected with perfect(ive) verb forms in Bantu. However, the reading may very well be rendered as ‘have become’, depending on clause-external issues of viewpoint or the general context in which the propositions are uttered.11 Except for perfective SM-B-iɾE, the meaning of already may also be expressed with -maɾa in a different “perfectoid” verb form, namely SM-Vká-B-a,12 as illustrated in (11). (11) Ikoma ne-eká-maɾ-a ko-hííʃi amá-nt͡ʃe sm1sg-nucl-already-fv inf-boil.caus 6-water ‘I have already boiled the water.’

Following Roth (2018), we will refer to this verb form as the “nucleative” (glossed as nucl). This verb form may convey a perfect meaning; this is arguably what it contributes to the proposition in (11), often with an immediate past (‘just’) flavor. However, it also conveys several other functions, including (present) progressivelike readings in addition to (firsthand/eyewitness) evidentiality, as further discussed in §6.1. In terms of the structural typology of auxiliaries, the AVC expressing already in WS thus belongs to the AUX[iliary]-headed type, i.e. with the auxiliary verb as both the syntactic and inflectional head and with the full verb carrying non-finite morphology (Anderson 2006, 2011). In this manner, the already AVC in WS arguably instantiates what Heine & Reh (1984: 114–116) refer to as the “nominal periphrasis” channel of formal grammaticalization along the verb-to-affix route.

11. See Botne (2022) for more on perfect(ive) marking in Bantu in general and Roth (2018: 54–70) for WS specifically. 12. The /V/ in this formula refers to an unspecified vowel, that is, a vowel which adapts its quality depending on the preceding vowel. The comparative data indicate that this vowel was originally an *a, as further discussed in §6.1. See Roth & Botne (forthcoming) on an elementbased analysis of the discourse functions associated with the formative a- in these and other verb constructions in Ishenyi.

Chapter 3. -maɾa in Mara

4.2 Expressing not yet not yet is a negative PhP expression that conveys the (continuing) non-realization of an event with a prospective focus on its future realization. In relation to pragmatic overtones of counter-expectation, not yet marks that the assumed beginning and duration of an event is delayed. As evident in examples (12) and (13), when expressing the notion of not yet, -maɾa is inflected very differently compared to when it expresses already. (12) Ngoreme tu-tá-ámaɾ-a tu-ɣɔ́ɾ-a e ŋɔ́mbɛ je-eɾe sm1pl-neg-not.yet sm1pl-buy-fv 9.cow 9-dem ‘We haven’t bought this cow yet.’ (13) Nata n-tjá-ámaɾa mɛɲ-a hánɔ to-ku-ɣí sm1sg-neg-not.yet sm1sg.know-fv where sm1pl-ipfv-go ‘I don’t know yet where we are going.’

To start with, not yet is a negative expression. Consequently, the construction includes negative morphology. The component of negation is expressed by the standard means in WS. Typically, that is through a negative prefix ta-, occurring in a position directly after the subject marker, as in the Ngoreme example in (12) above. However, there are many morpho-phonological quirks associated with the negative marker in WS. Thus, as illustrated in the Nata example in (13) above, the negative prefix and the 1sg subject prefix ne‑ have partly merged. Moreover, as seen in the Ikoma example in (14), when the SM is only a vowel the negative prefix may retract to a position before the SM to act as a coda on the initial syllable of the verbal word. (14) Ikoma naŋɡasi a-aɲí mu-máŋɡi tá-ámaɾa a-kwíɾ-u conc sm1-pres.cop 1-beautiful sm1.neg-not.yet sm1-marry-pass ‘Even though she is beautiful, she has not yet married.’

Finally, rather than the post-final prefix as in example (12) above, Ngoreme may instead optionally resort to a different pre-subject marker negative prefix ti‑ as in (15).13

13. See Laine et al. (forthcoming) for more information on the negation system of Ngoreme.

85

86

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

(15) Ngoreme ti-ná-ámaɾa máɲ-a ku-ɣámb-a ekí-íŋɡeɾesa neg-sm1sg-not.yet sm1sg.(come_to)know-fv inf-speak-fv 7-English ‘I don’t know (lit. haven’t come to know) (how) to speak English yet.’

The negative marking aside, -maɾa furthermore does not occur in any of the affirmative perfect(ive) verb forms found in WS (illustrated in §4.1). Instead, it is inflected in what must be commensurate with a verb form ⁰SM-a-B-a, an otherwise obsolete construction in WS (see §5.2 and §6.1 for further elaboration on this verb form). Finally, the second predicate verb is not coded as an infinitive. Instead it occurs in a type of finite yet dependent verb form (occasionally referred to as the “conjunctive” within Bantuistics; see Meeussen 1967: 113; Nurse 2008: 309), only used in complement clauses and other subordinate constructions in WS. Hence, in contrast to the auxiliary-headed nominal periphrasis of the AVC expressing already, the construction with -maɾa expressing not yet belongs to Anderson’s (2006, 2011) double-headed type and thus serves as an instantiation of Heine & Reh’s (1984: 116–120) serial periphrasis. Taking these facts together clearly shows that although the expression of already and not yet are closely interrelated on a semantic level, and in spite of the fact that they both contain the verb -maɾa, the two constructions are highly asymmetrical from a structural perspective and, as we shall see, also in terms of historical and ongoing developmental patterns.

5.

The grammaticalization of the -maɾa AVCs: Semantic and formal diagnostics

This and the next section focus on reconstructing the grammaticalization trajectories leading up to the two AVCs involving -maɾa in WS. In this endeavor, the methodology established within grammaticalization studies for the historical reconstruction of un(der-)documented languages is followed. As noted by Heine (2018) – and as demonstrated already in the semasiological account of -maɾa in §3 – this methodology mainly relies on diachronic reconstruction by means of the comparative method paired with internal reconstruction and typological generalizations. It may be argued that this approach works particularly well in the context of such a dense yet tightly-knit language family as Bantu, which can add a robust comparative backdrop to the reconstructional work on a linguistic expression’s history, development and spread through space and time, even when zooming in to a small selection of varieties.

Chapter 3. -maɾa in Mara

Following Norde & Beijering (2014), grammaticalization is here treated essentially as consisting of the composite changes of semantic reinterpretation and formal reanalysis. Indeed, there are both functional and formal traits which tease the two -maɾa constructions apart from their lexical sources. In this section, we will briefly zoom in to some of these traits as they are witnessed in the development first of already (§5.1) and then of not yet (§5.2). As will be seen, the formal differences noted in §4 above are reflected in the highly diversified historical backgrounds and developmental patterns of the two PhP constructions under consideration.

5.1 already The AVC expressing already in WS constitutes in many ways the more straightforward case of grammaticalization, as it adheres to a common grammaticalization route (see §2). It also exhibits typical diagnostics of (Bantu) AVC grammaticalization, indicative of the polysemic split from the source construction presented in §3. Below follow a few illustrations of various such functional and formal diagnostics of the grammaticalized status of the already AVC in WS. 5.1.1 Relaxation of selectional restrictions and expansion in collocational range In its grammatical(ized) role as an auxiliary expressing already, -maɾa shows a relaxation of its selectional restrictions indicative of host-class expansion (Himmelmann 2004), also known as an increase in type frequency (Traugott 2010). In practice, this means that -maɾa can be shown to be able to occur in a broader set of contexts and to collocate with an expanded set of complement verbs, as compared to its lexical etymon ‘finish’. Recall from §3 that -maɾa as a lexical aspectualizer is not totally banned from co-occurring with punctive complement verbs, as it may supply a “missing” coda phase on inceptive punctives like ‘arrive’. However, a combination of lexical ‘finish’ and punctive verbs with a nonvolitional agent like ‘see’ and ‘be shot’ do render a collocational clash. Yet, when -maɾa instead functions as a grammatical marker expressing already, this collocational restriction does not hold anymore, as illustrated in (16) and (17), being the re-iterated examples (7) and (8). (16) Ishenyi ni-marɛ ku-ɾóɾ-a kilimant͡ʃaɾo sm1sg-already.pfv inf-see-fv Kilimanjaro **‘I have finished seeing Mt. Kilimanjaro.’ BUT: ‘I have already seen Mt. Kilimanjaro.’

87

88

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

(17) Ishenyi na-a-marɛ ku-ɾás-u ɾi-sáási foc-sm1- already.pfv inf-shoot-pass 5-bullet **‘She has finished being shot.’ BUT: ‘She has already been shot.’

5.1.2 Loss of verbal properties and syntactic variability: The case of the locative enclitic An indication of stronger syntagmatic cohesion between the auxiliary -maɾa and the second verb and thus a loss of syntactic variability relative to the source construction comes from the position of the phoric locative enclitics. These may surface on lexical -maɾa when the verb takes a nominal argument (18) or an infinitival complement (19). (18) Ishenyi aβá-ána m-ba-marɛ́ = mu ɣa-ase 2-child foc-sm2-finish.pfv = loc18 6-all ‘The children have finished everything (there).’ (19) Ikoma ni-maɾ-iɾi = hó ɣo-t͡ʃíβ-a sm1-finish-pfv = loc16 inf-weed-fv ‘I have finished weeding there.’

However, as indicated in (20), when functioning as an auxiliary verb, a locative enclitic on -maɾa is unacceptable. It may only be attached to the second verb. (20) Ishenyi a. mo-mu-ɣɔ́ka ó-no amá-nt͡ʃe ŋ-ɡa-marɛ ku-hɛ́ tɛɾ-a = mu loc18-3-canal 3- dem 6-water foc-sm6-finish.pfv inf-pass-fv = loc18 ‘At this canal, the water has already passed through (there).’ b. ** mo-mu-ɣɔ́k-a ó-no amá-nt͡ʃe ŋ-ɡa-mare = mu ku-hɛ́ tɛɾ-a loc18-3-canal 3-dem 6-water foc-sm6-finish.pfv = loc18 inf-pass-fv 'At this canal, the water has already passed through (there).’

This loss of ability to carry a locative enclitic arguably shows a decrease in the verbal properties of the auxiliary -maɾa.

Chapter 3. -maɾa in Mara

5.1.3 Shift in anaphoric referencing To be sure, the grammaticalization of the AVC -maɾa + infinitive and the accompanying categorical reinterpretation affects not only -maɾa but also the second verb, as it is reinterpreted from a nominal-like verbal complement to being the (semantic) main verb. As a consequence, it may no longer be replaced with a pro-noun, as in the question to (21) in (22)a, but needs to be replaced with a pro-verb like ‘do’, as in (22)b (or with a construction where the full verb is explicitly expressed).14 (21) Nata n-á-marɛ ku-ɾááɣɛɾ-a foc-sm1-finish.pfv inf-eat-fv ‘She has finished eating.’ OR ‘She has already eaten.’ (22) Nata a. ne-ke á-marɛ foc-q sm1-finish.pfv ‘What has she finished? ** ‘What has she already (done)?’ b. ne-ke á-marɛ ɣu-kɔ́ɾ-a foc-q sm1-already.pfv inf-do-fv ‘What has s/he already done?’

5.1.4 Loss of phonological substance: The case of “detrillification” There are also several minor ongoing cases of segmental and suprasegmental shifts which are indicative of erosion and an increased fixation and bondedness between the auxiliary ‑mara and the second full verb. Here it suffices to mention one of the more striking ones, namely the case of “detrillification” in Ishenyi. Recall from §4.1 that the flap in -maɾa may coalesce with the flap of the perfective suffix -iɾE, the phonetic result being an articulatory more complex trill /r/. However, in Ishenyi, the already auxiliary may often be expressed with just a flap, whereas the lexical construction expressing ‘finish’ requires a trill (or the full formal realization /maɾ- iɾE/). As evidenced in (23), when realized as just SM‑maɾɛ, the lexical ‘finish’-reading of -maɾa is unacceptable.

14. See Bostoen et al. (2012) for a similar AVC diagnostic in the Bantu language Rundi; see also Divjak (2018) and Heine & Miyashita (2008) for (Indo-European) Polish and German, respectively.

89

90

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

(23) Ishenyi n-tu-maɾɛ ɣu-ɣámb-a na = ɣwe foc-sm1pl-finish.pfv inf-talk-fv com = 3sg ‘We have already spoken with her/him.’ **‘We have finished speaking with him.’

This fact is indicative of a loss of phonological substance and thus of a further structural simplification of the AVC. As there are no attestations of this feature in the other WS varieties besides Ishenyi, this would also constitute one of the few instances of (micro-)variation in the constructional realization of this AVC witnessed within WS.

5.2 not yet Although the grammaticalization process of the already construction forms a more canonical instance of auxiliation, the grammaticalized status of the not yet construction is actually more directly obtainable. This is based on the crucial factor (already hinted at in different parts of this paper) that the verb form ⁰SM-a-B-a in which -maɾa occurs is obsolete in contemporary WS. This means that this verb form, with or without negation, cannot be used with verb roots other than -maɾa. Hence, an intended construction like (24) is considered unacceptable. (24) Ngoreme **atááwansa aɾema ‘(intended:) S/he has not started cultivating.’

Not even the verb -maɾa can occur by itself in this verb form (whether with its lexical or grammatical reading), as shown in (25)a. Instead, it is dependent on a second verb on which it may only act as a grammatical operator conveying not yet, not the lexical ‘finish’, as illustrated in (25)b. (25) Ngoreme a. **atáámaɾa ‘(intended:) S/he has not finished (yet).’ b. atáámaɾa amáɾa ‘S/he has not finished (yet).’ **‘S/he has not finished finishing.’

There are also indications of further formal grammaticalization of the not yet AVC. In particular, there are instances of incremental attrition of the second verb subject prefix. This is most advanced in Ikoma, where there are attestations of the whole subject prefix of the second verb having been omitted altogether (although this still appears to be quite a rare phenomenon, not accepted by all speakers). This may be seen in the absence of the sm1pl morpheme on the second verb ‘see’ in example (26).

Chapter 3. -maɾa in Mara

(26) Ikoma to-ta-amáɾa mo-riɣi sm1pl-neg-not.yet om1-see ‘We have not seen her (yet).’

This formal development is in line with what one might expect from a doubleheaded serial periphrasis type of AVC such as the not yet-construction in WS, given the simple fact that double indexation of the subject is uneconomical once the construction has been grammaticalized (see Güldemann 2003).

6.

Old forms and cycles of change

This section sets out to reconstruct the grammaticalization pathway of the two AVCs with -maɾa in WS. Ultimately, we believe that the development of these constructions was driven by the need in these language varieties to reinforce the specific readings encapsulated in the linguistic expressions of already and not yet. More specifically, we believe, based on the lexical semantics of the source verb ‘finish’, that both AVCs were formed to highlight a reading of (non-)completion of a given event, i.e. the completion (already) or not (not yet) of the phase leading up to its realization (see Van Baar 1997: 86–87; see also Hopper’s 1991 notion of “persistence”). This development simultaneously strengthened the pragmatic readings of counter-factuality. Both already and not yet have been described as inherently focused expressions (König 1991: Chapter 7; Nicolle 2012; Güldemann 2008: 489–490; Veselinova & Devos 2021; Devos & Veselinova 2021), which at the same time entails that they are intrinsically prone to being neutralized and consequently in need of further reinforcement. Such a trait furthermore suggests the triggering of diachronic cycles, i.e. series of diachronic re-instantiations of the same patterns of development and loss within a single functional category. Indeed, as will be argued in the following, the development of both -maɾa AVCs can be analyzed as instantiating developmental patterns reminiscent of a diachronic cycle. In §6.2, we discuss the premises of such PhP cycles both for already and not yet from a general Bantu context. In §6.3, we zoom in to the particularities of the developments in WS and how they adhere to these sketched scenarios. First and foremost, however, the development of both AVCs with -maɾa in WS takes place in an environment where other formal means traditionally employed for expressing these concepts have come to drift away semantically, as is the case for already, or have become obsolete, as is the case for not yet. In relation to this fact, the grammaticalization of the -maɾa AVCs can be argued to

91

92

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

fill a functional void in these language varieties. We start, in §6.1, by looking at two simplex verb forms found in this area that may have served as the original markers of already and not yet and which concurrently form part of the morphological building blocks of the auxiliaries with -maɾa.

6.1 The original formal means for expressing already/not yet We may note, to begin with, that a formula *SM-a-B-a is generally assumed as a near past/perfect construction for either Proto-Bantu or at least early Bantu (Meeussen 1967: 109, 113; Nurse & Philippson 2006; Botne 2010, 2014; Dom et al. 2018), that is, a construction identical in shape to the negated verb form which -maɾa occurs in when expressing not yet. Next, Meeussen (1967: 109) posits a Proto-Bantu prefix *ka for the “limitative” slot (referring to a secondary prefix position occurring directly after the slot devoted to other TA prefixes) conveying both already and not yet. The combination of the aforementioned verb formula and the limitative prefix ka‑ would render the verb form ⁰SM-a-ka-B-a. Reflexes of this particular form with the meaning already/not yet, as well as more “neutral” perfect readings, often as a recent past perfect (‘just’), occur in many Great Lakes Bantu sister languages to WS, in particular in the Eastern Nyanza and Greater Luhya languages spoken on the eastern side of Lake Victoria; see, e.g., Mould (1981: 206–207), Nurse & Muzale (1999), Botne (2010). In WS, however, the cognate verb form SM-Vká-B-a has developed other peculiar functions. It may not convey the meaning of already alone and although it may at times acquire the function of a perfect (as was demonstrated in (11) in §4.1), it may also be found with a present progressive reading, not only with change-of-state but also with durative verbs, as evident in (27). What is more, as further seen in (28), SM-áka-B-a in WS has developed into a marker of (firsthand) evidentiality (cf. Roth 2018: 85–109; Crane et al. forthcoming). (27) Ikoma Wáka a-ká-ách-a! lion sm1-nucl-come-fv ‘Lion is coming!’ (28) Ikoma a-aɣá-ku sm1-nucl-die-fv ‘S/he has (just) died.’ {Context: An elderly person is very sick and near death; s/he is at home. Speaker is in the same room and can see the person OR in an adjacent room but can hear the person.}

Chapter 3. -maɾa in Mara

Similarly, the “bare” perfectoid verb form *SM-a-B-a is identical to the verb form with which -maɾa is conjugated when expressing not yet in WS. As was mentioned in §4.2, however, this verb form itself is obsolete and does not occur in any other contexts in WS, including with any verb other than -maɾa. Still, it appears in the comparative data. For example, the direct neighbor and close relative Kuria (North Mara) appears to freely make use of a negated variant of this verb form, as evident in (29). (29) Kuria (JE43; Cammenga 2004:288) bhá-tá-á-sǿm-a sm2-neg-a-read-fv ‘They have not (yet) read (up to now).’

Of particular interest here is the parentheses in the translation of the example, which suggest that the ‘yet’ reading is non-obligatory and thus that this construction may also convey a neutral plain negative perfect. That is exactly the kind of co-expression of not yet and plain negation which would trigger the emergence of an auxiliary construction specifying and reinforcing the reading of not yet, a fact to which we will turn to in the following section.

6.2 Diachronic cycle(s) of phasal polarity It has recently been suggested by Veselinova & Devos (2021) and Devos & Veselinova (2021) that the development of not yet markers in (Eastern) Bantu may be conveyed as instantiating a cyclical pattern of change. They sketch three developmental stages of such a scenario. In Stage 1, both not yet and plain (perfect/resultative) negation are expressed with the same simplex (morphological) verb form. To reinforce the specific not yet reading, the simplex verb form inflects on an auxiliary in a (foregrounding) auxiliary construction, in Stage 2. In Stage 3, this auxiliary construction, in turn, may fuse and undergo morphologization back into a synthetic simplex verb form, which may begin to neutralize into a regular negative (perfect/resultative) in need of further reinforcement, in which case the cycle may occur again. In their analysis of this type of developmental scheme, these authors focus on the comparatively more common use of “dummy” (lexically empty) verbs, e.g. copula verbs, as auxiliaries. With the use of -maɾa as a not yet auxiliary, WS has instead employed a source verb with lexical content which also contributes a semantic component to the construction. Still, we believe that the same type of motivation in terms of the need to reinforce the not yet reading was at play in WS too. Additionally, we believe that the semantically weighty origin of -maɾa brought with it an additional emphasis on the non-completion of the event.

93

94

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

Interestingly, it would appear to be just as possible to analyze the development of the positive already through the same analytical model of a recurrent diachronic cycle. A clear example comes from Swahili, where diachronic data also exist in support of this scenario; see Nurse & Hinnebusch (1993: 376), Nurse (2008: 60), Nicolle (2012), Furumoto (2019) and further references therein for reconstructional work on these constructions. Here, ‑mala, a cognate verb to -maɾa in WS, inflected in the cognate reflex of the perfect(ive) verb form -iɾE (obsolete in present-day Swahili) and also operating on an infinitive full verb, underwent grammaticalization instantiating the different stages of the cycle. As sketched in (30)a–c, ‑mala was recruited to reinforce the specific completive already reading (Stage 2), subsequently undergoing morphologization (Stage 3) into a prefix and neutralization into a perfect(ive) marker of its own, thus closing the cycle (back to Stage 1). (30) Swahili (G42) a. *tu-mal-ile ku-lima sm1pl-finish-pfv inf-cultivate-fv ‘We have finished cultivating / we have already cultivated.’ b. *tu-mele-ku-lim-a sm1pl-already.pfv-inf-cultivate-fv ‘We have (already) cultivated.’ c. tu-me-lim-a sm1pl-pfv-cultivate-fv ‘We have cultivated.’

At this stage, a new cycle kicked in. That is, in order to reinforce the reading of earliness and completion (and counter-factuality) once again, a new auxiliary verb emerged (Stage 2), namely -(kw)isha, another common Bantu verb meaning ‘finish’ (cf. Proto-Bantu *-jic-; Bastin et al. 2002). Inflected in the (now perfect(ive)) SM-me-B-a verb form, this already construction, in turn, exhibits various degrees of univerbation and formal condensation indicative of an inceptive Stage 3, as illustrated in (31). (31) Swahili (G42) a. ni-me-kwish-a ku-imb-a sm1sg -pfv-finish~already-fv inf-sing-fv ‘I have finished singing / I have already sung.’ b. ni-me-kwisha-imba ~ ni-me-sha-imba sm1sg-pfv-already-sing-fv ‘I have already sung.’

In sum, what is witnessed here is clearly a recurrent cycle where Auxiliary1 turns into Affix1, merely to be replaced by Auxiliary2 which, however, is inflected with Affix1 (i.e. the erstwhile Auxiliary1). Auxiliary2 subsequently also turns into an affix

Chapter 3. -maɾa in Mara

(Affix2). All stages involve material with the same source meanings that develop grammatical functions within the same “perfectoid pool”. In the following section, we will try to integrate these PhP cycles – i.e. the adapted already-cycle and the original not yet-cycle proposed by Veselinova & Devos (2021) and Devos & Veselinova (2021) – to see how they have played out for the two -maɾa AVCs in WS.

6.3 The PhP cycle for -maɾa in Mara Based on the discussion above we may hypothesize that at some putative historical Stage 1, already and not yet were expressed through some simplex morphological means, such as the forms presented in §6.1 above. It is also possible that already was expressed as an implicational reading of the perfect(ive) verb form with the suffix ‑iɾE (which is common in Bantu; see Botne forthcoming) Next, -maɾa was recruited and embedded in these simplex verb forms forming two auxiliary constructions to accentuate the reading of (non-)completiveness and counter-expectation. At some point in time, concurrent or sequential to this process, the verb form ⁰SM-a-B-a disappeared as a productive means of verbal conjugation, whereas the verb form SM-Vká-B-a became extensively associated with other functions. Both already and not yet would thus serve as prima facie instantiations of Stage 2 of the diachronic PhP cycles, as sketched in §6.2 above. At the same time, the signs of erosion witnessed for both constructions are arguably indicative of movements towards Stage 3. Similarly, there are functional indications that at least the not yet AVC with -maɾa is increasingly being neutralized into a broader and thus more neutral negative perfect. At least, it recurrently occurs as a marker of the perfect types associated with a speaker-oriented vantage point (see Botne forthcoming; see also Comrie 1976; Dahl & Wälchli 2016). A case in point is the negative experiential perfect construction in (32), which marks the situation referred as not having occurred even once (at some specific time in the past). Importantly, the second coordinated clause in this sentence cancels out any reading of the prospect of a future realization of the event depicted in the first clause, and thus a significant component of the meaning of not yet is lost. (32) Ngoreme ti-n-ámaɾa né-ɣi kéɲa té-ni-ku-ɣi he neg-sm1sg-pfr? sm1sg-go Kenya neg-sm1-ipfv-go emp ‘I have never (i.e. not once) been to Kenya and I will never go.’

Moreover, both AVCs may themselves be further reinforced, especially for expressions with a high pragmatic load of counter-factuality.

95

96

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

Thus, as (33) shows, in Ngoreme already may be reinforced with the adverb kaɾe (< *kàdé ‘olden times’; Guthrie 1967: 260). The cognate is also used for already in Kuria as well as outside of Mara, e.g. in other Kenyan-Tanzanian Bantu languages such as Digo (Nicolle 2013: 154), Lubukusu (Wasike & Diercks 2016), and perhaps also in Ndamba (Edelsten & Lijongwa 2010). (33) Ngoreme ɾóóbi a-máre ko-híka kaɾe βúnda Rhobi sm1-already.pfv inf-arrive already Bunda ‘Rhobi is already in Bunda.’

Similarly, there are instances in WS where the not yet reading of the construction with -maɾa itself is being reinforced. This reinforcement is executed through the use of another auxiliary -keeɾe marking still, thus being another PhP marker of WS.15 Example (34) is a case in point. (34) Ngoreme ɾóóbi a-kééɾe t-á-ámaɾa a-hík-a βúnda Rhobi sm1-still neg-sm1-prf? sm1-arrive-fv Bunda ‘Rhobi has not yet arrived in Bunda.’

Often, this doubled construction would seem to be used for marking counterexpectation. It also carries a more emphatic reading of disbelief and irritation similar to that of still not in English (see Van Baar 1997: 2), as evident in examples (35) and (36). (35) Ngoreme ɾóóbi kána o-kééɾe u-tá-ámaɾa o-swáɾ-a Rhobi oh sm2sg-still sm2sg-neg-prf? sm2sg-wear_clothes-fv ‘Oh Rhobi, are you STILL NOT dressed?!’ {I have told you three times already}

15. -keeɾe is either derived from the verb ikaɾa ‘sit, stay’ or a combination of the so called “persistive” (i.e. a still-marking) prefixal morpheme *kɪ́ -(reconstructible for Proto-Bantu; Nurse 2008: 246–249) and the copula verb -ɾe. Interestingly, -keeɾe + an infinitive verb has conventionalized into another not yet marker in WS, probably through the conventionalization of an inference invited from the inceptive status and thus non-factual reading of the infinitive verb, viz. ‘still to X’ > ‘not yet X’ (cf. Güldemann 1998; Nurse 2008: 147–148; Veselinova & Devos 2021). Since the auxiliary -maɾa cannot occur without a second predicate verb (see §5.2), -keeɾe alone may convey not yet in non-predicative elliptic contexts, such as in negative answers. See Laine et al. (forthcoming) for more information about this construction and its contrasts with the -maɾa construction with specific reference to Ngoreme.

Chapter 3. -maɾa in Mara

(36) Nata a-keeɾé ta-amáɾa a-ɣotooɾi kw-éɣi sm1-pers sm1.neg-prf? sm1-help inf-wash_up ‘Has he STILL NOT helped cleaning?’

7.

Summary and conclusions

This study has accounted for the grammaticalization of two auxiliary verb constructions involving the verb -maɾa ‘finish’ into markers of already and not yet in the Western Serengeti language varieties of Mara Bantu. As has been shown, there are both functional and formal indications of the grammatical(ized) status of both these AVCs, albeit in very different manners. Indeed, although the expressions of already and not yet form conceptually related markers of phasal polarity, and although they make use of the same lexical verb as auxiliary, they occur in very different constructional frames with different historical backgrounds. not yet is not merely the negated version of already but stands in a construction of its own. Thus, although they share a joint paradigm or semantic space, it is most reasonable to assume that the functions were developed separately, rather than from the same grammaticalization chain. With specific reference to the development of expressions of PhP, Van Baar (1997: 342) refers to this type of polygrammaticalization as the “plane model”, comparing it to an airplane that unloads its passengers at different stops where they continue to exist independently of each other, although they all originally flew in the same plane (i.e. they were instances of the same lexical source). These differences aside, it was argued that both forms can still be analyzed as being part of a type of recurrent linguistic cycle in Eastern Bantu, driven by the need to reinforce these inherently focused and thus expressive concepts. As argued, this occurred in an environment where the verb forms hypothesized to have originally served to mark the concepts of already and not yet had drifted away to develop other functional qualities, or even into oblivion.

Acknowledgements We wish to thank our Western Serengeti language consultants (and the Mara branch of SIL International). We gratefully acknowledge the financial support of the Kone Foundation for this research. The first author also wishes to thank the PK Hedborg Foundation and the Salén Foundation for making the finalization of this paper possible. An initial version of this study was presented at the 8th International Conference on Bantu languages in Essex, UK (held online) in June 2021; we want to thank the audience there for valuable questions and comments. We

97

98

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

also wish to thank the editor Sylvie Hancil and two anonymous reviewers for their thoughtful comments which helped improve the content of this paper. Thanks also to Mary Chambers for improving our English. This study has greatly benefited from the inspirational work of Maud Devos and from valuable discussions with her. All potential mistakes and shortcomings are our own.

References Anderson, Gregory D. S. 2006. Auxiliary Verb Constructions. Oxford: OUP. Anderson, Gregory D. S. 2011. Auxiliary verb constructions in the languages of Africa. Studies in African Linguistics 40(1–2): 1–409. Appleby, L. L. 1961. A First Luyia Grammar with Exercises, revised edn. Nairobi: East African Literature Bureau. Ashton, Ethel O. 1947. Swahili Grammar. London: Longmans, Green & Co. Aunio, Lotta. 2017. Syllable weight and tone in Mara Bantu languages. In Syllable Weight in African Languages [Current Issues in Linguistic Theory 338], Paul Newman (ed.) 191–214. Amsterdam: John Benjamins. Aunio, Lotta, Robinson, Holly, Roth, Tim, Stegen, Oliver & Walker, John B. 2019. The Mara Languages JE40. In The Bantu Languages, 2nd edn, Mark van de Velde, Koen Bostoen, Derek Nurse & Gérard Philippson (eds), 501–532. London: Routledge. Bastin, Yvonne, Coupez, André, Mumba, Évariste & Schadeberg, Thilo (eds). 2002. Bantu Lexical Reconstructions 3 / Reconstructions lexicales bantoues 3. Tervuren: Royal Museum for Central Africa. (22 December 2022). Bernander, Rasmus. 2017. Grammar and Grammaticalization in Manda: An Analysis of the wider TAM Domain in a Tanzanian Bantu Language. PhD dissertation, University of Gothenburg. Bernander, Rasmus. 2021. The phasal polarity marker (a)kona in Manda and its history. In The Expression of Phasal Polarity in Sub-Saharan African Languages, Raija L. Kramer (ed.) 41–72. Berlin: De Gruyter Mouton. Bernander, Rasmus & Laine, Antti. 2020. The formation of existential constructions in Western Serengeti – A micro-comparative exploration of variation and change. Africana Linguistica 26: 59–102. Binnick, Robert I. 1991. Time and the Verb: A Guide to Tense and Aspect. Oxford: OUP. Bostoen, Koen, Mberamihigo, Ferdinand & de Schryver, Gilles-Maurice. 2012. Grammaticalization and subjectification in the semantic domain of possibility in Kirundi (Bantu, JD62). Africana Linguistica 18: 5–40. Botne, Robert. 2010. Perfectives and perfects and pasts, oh my!: On the semantics of ‑ile in Bantu. Africana Linguistica 16: 31–63. Botne, Robert. 2014. Resultatives, remoteness, and innovation in Eastern and Southern Bantu T/A systems. Nordic Journal of African Studies 23(1): 16–30. Botne, Robert. Forthcoming. Tense and aspect. To appear in The Oxford Guide to the Bantu Languages, Lutz Marten, Nancy Kula, Ellen Hurst & Jochen Zeller (eds). Oxford: OUP.

Chapter 3. -maɾa in Mara

Bybee, Joan L., Perkins, Revere D. & Pagliuca, William. 1994. Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago IL: The University of Chicago Press. Cammenga, Jelle. 2002. Phonology and Morphology of Ekegusii: A Bantu Language of Kenya [East African Languages and Dialects 12]. Cologne: Rüdiger Köppe. Cammenga, Jelle. 2004. Igikuria Phonology and Morphology: A Bantu Language of South-West Kenya and North-West Tanzania. Cologne: Rüdiger Köppe. Comrie, Bernard. 1976. Aspect: An Introduction to the Study of Verbal Aspect and Related Problems. Cambridge: CUP. Crane, Thera Marie, Gunnink, Hilde, Kanijo, Ponsiano & Roth, Tim. Forthcoming. Aspect and evidentiality in four Bantu languages. To appear in Beyond Time, Astrid de Wit, Frank Brisard, Carol Madden-Lombardi, Michael Meeuwis & Adeline Patard (eds). Oxford: OUP. Dahl, Östen. 2021. “Universal” readings of perfects and iamitives in typological perspective. In The Perfect Volume: Papers on the Perfect [Studies in Language Companion Series 217], Kristin Melum Eide & Marc Fryd (eds), 43–64. Amsterdam: John Benjamins. Dahl, Östen & Wälchli, Bernard. 2016. Perfects and iamitives: Two gram types in one grammatical space. Letras de Hoje 51(3): 325–348. Devos, Maud & Veselinova, Ljuba. 2021. Some periphrastic ‘not yet’ constructions and cyclical change in Eastern Bantu languages. Presentation at the 8th International Conference on Bantu languages, Essex, 2–4 June. Divjak, Dagmar. 2018. Binding scale dynamics: Fact or fiction? In Aspects of Linguistic Variation [Trends in Linguistics. Studies and Monographs 324], Daniel Van Olmen, Tanja Mortelmans & Frank Brisard (eds), 9–42. Berlin: De Gruyter. Dom, Sebastian, de Schryver, Gilles-Maurice & Bostoen, Koen. 2018. The diachronic semantics of the dissociative past completive construction in the Kikongo language cluster (Bantu). Folia Linguistica Historica 39(2): 297–340. Edelsten, Peter & Lijongwa, Chiku. 2010. A Grammatical Sketch of Chindamba: A Bantu anguage (G52) of Tanzania [East African Languages and Dialects 21]. Cologne: Rüdiger Köppe. Freed, Alice F. 1979. The Semantics of English Aspectual Complementation. Dordrecht: Reidel. Furumoto, Makoto. 2019. *-Mala ‘finish’ derived perfect(ive) prefixes in Unguja dialects of Swahili. Swahili Forum 26: 63–88. Grollemund, Rebecca, Branford, Simon, Bostoen, Koen, Meade, Andrew, Venditti, Chris & Pagel, Mark. 2015. Bantu expansion shows that habitat alters the route and pace of human dispersals. Proceedings of the National Academy of Sciences of the United States of America 112(43): 13296–13301. Guthrie, Malcolm. 1948. The Classification of the Bantu Languages. London: OUP. Guthrie, Malcolm. 1967–1971. Comparative Bantu: An Introduction to Comparative Linguistics and Prehistory of the Bantu Languages, Vols. 1–4. Farnborough: Gregg International Publishers. Güldemann, Tom. 1996. Verbalmorphologie und Nebenprädikationen im Bantu: Eine Studie zur funktional motivierten Genese eines konjugationalen Subsystems. Bochum: Brockmeyer.

99

100

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

Güldemann, Tom. 1998. The relation between imperfective and simultaneous taxis in Bantu: Late stages of grammaticalization. In Afrikanische Sprachen im Brennpunkt der Forschung: Linguistische Beiträge zum 12. Afrikanistentag Berlin, 3.–6. Oktober 1996, Ines Fiedler, Catherine Griefenow-Mewis & Brigitte Reineke (eds), 157–177. Cologne: Rüdiger Köppe. Güldemann, Tom. 1999. The genesis of verbal negation in Bantu and its dependency on functional features of clause types. In Bantu Historical Linguistics, Jean-Mane Hombert & Larry M. Hyman (eds), 545–587. Stanford CA: CSLI. Güldemann, Tom. 2003. Grammaticalization. In The Bantu Languages, Derek Nurse & Gérard Philippson (eds), 182–194. London: Routledge. Güldemann, Tom. 2008. Quotative Indexes in African Languages: A Synchronic and Diachronic Survey. Berlin: de Gruyter. Güldemann, Tom. 2018. Areal linguistics beyond contact, and linguistic areas of Afrabia. In The Languages and Linguistics of Africa [The World of Linguistics 11], Tom Güldemann, (ed.) 448–545. Berlin: Mouton de Gruyter. Harjula, Lotta. 2004. The Ha Language of Tanzania: Grammar, Texts and Vocabulary [East African Languages and Dialects 13]. Cologne: Rüdiger Köppe. Heine, Bernd. 1976. Knowledge and use of second languages in Musoma Region: A quantitative study. Kiswahili 46: 49–59. Heine, Bernd. 1993. Auxiliaries: Cognitive Forces and Grammaticalization. Oxford: OUP. Heine, Bernd. 2018. Grammaticalization in Africa: Two contrasting hypotheses. In Grammaticalization from a Typological Perspective, Heiko Narrog & Bernd Heine (eds), 16–34. Oxford: OUP. Heine, Bernd & Miyashita, Hiroyuki. 2008. Accounting for a functional category: German drohen ‘to threaten’. Language Sciences 30: 53–101. Heine, Bernd & Reh, Mechthild. 1984. Grammaticalization and Reanalysis in African Languages. Hamburg: Buske. Heine, Bernd, Claudi, Ulrike & Hünnemeyer, Friederike. 1991. Grammaticalization: A Conceptual Framework. Chicago IL: The University of Chicago Press. Hill, Dustin, Lindfors, Anna-Lena, Nagler, Louise, Woodward, Mark & Yalonde, Richard. 2007. A Sociolinguistic Survey of the Bantu Languages in Mara Region, Tanzania. Dallas TX: SIL. Himmelmann, Nicolaus P. 2004. Lexicalization and grammaticalization: Opposite or orthogonal? In What Makes Grammaticalization? – A Look from its Fringes and its Components, Walter Bisang, Nikolaus P. Himmelmann & Björn Wiemer (eds), 21–42. Berlin: De Gruyter Mouton. Hopper, Paul J. 1991. On some principles of grammaticization. In Approaches to Grammaticalization, Vol. 1[Typological Studies in Language 19:1], Elizabeth Closs Traugott & Bernd Heine (eds), 17–35. Amsterdam: John Benjamins. Jonsson, Erland. 1954. Kinyang’wezi grammatik. Tabora, Tanganyika: Svenska Pingstmissionen. Kanijo, Ponsiano Sawaka. 2019. Aspectual Classes of Verbs in Nyamwezi. PhD dissertation, University of Gothenburg.

Chapter 3. -maɾa in Mara

Kramer, Raija. 2017. Position paper on phasal polarity expressions. Ms, University of Hamburg. (22 December 2022). Kuteva, Tania, Heine, Bernd, Hong, Bo, Long, Haiping Narrog, Heiko & Rhee, Seongha. 2019. World Lexicon of Grammaticalization, 2nd edn. Cambridge: CUP. König, Ekkehard. 1991. The Meaning of Focus Particles: A Comparative Perspective. London: Routledge. Laine, Antti, Bernander, Rasmus & Gibson, Hannah. Forthcoming. Negation in Ngoreme. To appear in Matti Miestamo & Ljuba Veselinova (eds), Negation in the Languages of the World. Berlin: Language Science Press. Maho, Jouni Filip. 2003. A classification of the Bantu languages – An update of Guthrie’s referential system. In The Bantu Languages, Derek Nurse & Gérard Philippson (eds), 639–651. London: Routledge. Maho, Jouni Filip. 2009. NUGL online: The online version of the new updated Guthrie List, a referential classification of the Bantu languages. Ms. (23 December 2022). Meeussen, Achiel Emiel. 1967. Bantu grammatical reconstructions. Africana Linguistica 3: 79–121. Moshi, Lioba. 1994. Time reference markers in KiVunjo-Chaga. Journal of African Languages and Linguistics 15: 127–160. Mould, Martin J. 1981. Greater Luyia. In Studies in the Classification of Eastern Bantu Languages, Thomas H. Hinnebusch, Derek Nurse & Martin Mould (eds), 181–236. Hamburg: Helmut Buske. Nicolle, Steve. 2012. Diachrony and grammaticalization. In The Oxford Handbook of Tense and Aspect, Robert Binnick (ed.) 370–397. Oxford: OUP. Nicolle, Steve. 2013. A Grammar of Digo: A Bantu Language of Kenya and Tanzania. Dallas TX: SIL International. Norde, Muriel & Beijering, Karin. 2014. Facing interfaces. A clustering approach to grammaticalization and related changes. Folia Linguistica 48(2): 385–424. Nurse, Derek. 1979. Classification of the Chaga Dialects: Language and History on Kilimanjaro, the Taita Hills, and the Pare Mountains. Hamburg: Buske. Nurse, Derek. 1999. Towards a historical classification of East African Bantu languages. In Bantu Historical Linguistics: Theoretical and Empirical Perspectives, Jean-Marie Hombert & Larry M. Hyman (eds), 1–41. Stanford CA: CSLI. Nurse, Derek. 2008. Tense and Aspect in Bantu. Oxford: OUP. Nurse, Derek & Hinnebusch, Thomas J. 1993. Swahili and Sabaki: A Linguistic History [University of California Publications in Linguistics 121]. Berkeley CA: University of California Press. Nurse, Derek & Muzale, Henry R. T. 1999. Tense and aspect in Great Lakes languages. In Bantu Historical Linguistics: Theoretical and Empirical Perspectives, Jean-Marie Hombert & Larry M. Hyman (eds), 517–544. Stanford CA: CSLI. Nurse, Derek & Philippson, Gérard. 2003. Towards a historical classification of the Bantu languages. In The Bantu Languages, Derek Nurse & Gérard Philippson (eds), 164–181. London: Routledge.

101

102

Rasmus Bernander, Antti Laine, Tim Roth & Lotta Aunio

Nurse, Derek & Philippson, Gérard. 2006. Common tense-aspect markers in Bantu. Journal of African Languages and Linguistics 27(2): 155–196. Persohn, Bastian. 2017. The Verb in Nyakyusa: A Focus on Tense, Aspect and Modality. [Contemporary African Linguistics 2]. Berlin: Language Science Press. Philippson, Gérard & Grollemund, Rebecca. 2019. Classifying Bantu languages. In The Bantu Languages, 2nd edn., Mark Van de Velde, Koen Bostoen, Derek Nurse & Gérard Philippson (eds), 335–354. London: Routledge. Roth, Tim. 2018. Aspect in Ikoma and Ngoreme: A Comparison and Analysis of Two Western Serengeti Bantu Languages. PhD dissertation, University of Helsinki. Roth, Tim & Gibson, Hannah. 2019. Differential diagnoses of language change in Ngoreme, a Bantu language of Tanzania. Africana Linguistica 25: 281–319. Roth, Tim & Botne, Robert. Forthcoming. Discourse-pragmatic functions of focus, tense, and aspect in Ishenyi narratives. To appear in Domains and Regions in Bantu Tense and Aspect, Robert Botne & Axel Fleisch (eds). Schadeberg, Thilo. 1990. Schon – noch – nicht – mehr: Das Unerwartete als grammatische Kategorie im Kiswahili. Frankfurter Afrikanistische Blätter 2: 1–15. Schoenbrun, David Lee. 1990. Early History in Eastern Africa’s Great Lakes Region: Linguistic, Ecological, and Archaeological Approaches, ca. 500 B.C. to ca. A.D. 1000. PhD dissertation, UCLA. Sillery, Anthony. 1936. Notes for a grammar of the Kuria language. Bantu Studies X: 9–30. Traugott, Elizabeth Closs. 2010. Grammaticalization. In Continuum Companion to Historical Linguistics, Silvia Luraghi & Vit Bubenik (eds), 269–283. London: Continuum. Van Baar, Tim. 1997. Phasal Polarity. Dordrecht: Foris. van der Auwera, Johan. 1993. ‘A lready’ and ‘still’: Beyond duality. Linguistics and Philosophy 16(6): 613–653. van der Auwera, Johan. 1998. Phasal adverbials in the languages of Europe. In Adverbial Constructions in the Languages of Europe, Johan van der Auwera & Dónall P. O. Baoill (eds), 25–145. Berlin: Mouton de Gruyter. Veselinova, Ljuba & Devos, Maud. 2021. NOT YET expressions as a lexico-grammatical category in Bantu languages. In The Expression of Phasal Polarity in African Languages, Raija L. Kramer (ed.) 445–498. Berlin: De Gruyter. Walker, John B. 2013. Comparative Tense and Aspect in the Mara Bantu Languages: Towards a Linguistic History. MA thesis, Trinity Western, Langley. Wasike, Aggrey & Diercks, Michael. 2016. Lubukusu adverbial syntax. Presentation at the 47th meeting of the Annual Conference on African Linguistics, University of California, Berkeley, 23–26 March. Whiteley, Wilfred Howell. 1960. The Tense System of Gusii [East African Linguistic Studies 4]. Kampala: East African Institute of Social Research.

section ii

Synchronic approaches

chapter 4

Tracking Jespersen’s cycle in Veronese and Bresciano Diatopic variation in sentential negation Marta Tagliani & Jelena Živojinović University of Verona | University of Graz

This article provides an insight in the expression of sentential negation in two Italo-Romance varieties, i.e., Bresciano and Veronese. The fieldwork data gathered in Lumezzane (BS), Brescia, San Giovanni Lupatoto (VR) and Nogara (VR) shows an intriguing range of variation. The Veronese peripheral data displays the particle mia (local form of mica) as a reinforcing negative element; central Bresciano shows a homogeneous pattern of postverbal negation, whereas in Lumezzanese the negative particle mia is completely grammaticalized. The general picture illustrates homogeneous negation patterns across the analysed structures, with all three stages of Jespersen’s cycle represented. Keywords: Jespersen's cycle, sentential negation, Italo-Romance, Veronese, Bresciano

1.

Introduction

In this article, we examine diatopic variation in the expression of sentential negation in two Italo-Romance varieties, Bresciano and Veronese by focussing on the use of the mica particle (in its local form mia) as a negative marker, which is attested in both substandard and regional Italian varieties. In the former, i.e., substandard Italian, mica has a presuppositional feature and is licensed under specific pragmatic conditions (cf. Cinque 1976). In the latter, i.e., Italo-Romance varieties, it instead seems to behave as a postverbal negation by gradually replacing the standard preverbal marker non across different syntactic configurations. Our aim is to provide an accurate description of the status of grammaticalization of mica within Jespersen’s cycle (1917) in these varieties by comparing the two central areas with two peripheral ones.

https://doi.org/10.1075/slcs.232.04tag © 2023 John Benjamins Publishing Company

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

In this introductory section, we first provide the relevant background information on the various syntactic strategies adopted in Modern Italian to express sentential negation (§1.1). Then, we briefly discuss the diachronic development of the item mica from Old to Modern Italian (§1.2.1). Finally, we present a short review of the literature on the use of mica negation in modern Italo-Romance varieties (§1.2.2). After this introduction, we describe the methods used for data collection (§2) and we present the linguistic data in the four geographical areas (§3). Eventually, we discuss their implications for the development of the negation system in Bresciano and Veronese varieties (§4).

1.1 The expression of sentential negation: An introduction Natural languages are characterized by an asymmetry between affirmative and negative sentences, with the latter being always marked in comparison to their affirmative counterpart in both language structure and use (Dahl 1979; Payne 1985; Horn 1989; Ladusaw 1996). In all word languages, the expression of sentential negation always requires the presence of an overt linguistic element (segmental or suprasegmental, such in the case of tone): in fact, differently from other grammatical functions (e.g., interrogative constructions), there is no language in which sentential negation can be expressed merely by means of a word order shift (Greenberg 1966; Horn 1989; Zeijlstra 2009). Nonetheless, natural languages display a lot of cross-linguistic variation in respect to how the negative meaning can be conveyed, with regard to the type, position and number of the negative markers employed. Zanuttini (2001) distinguished four main types of negative elements used to express sentential negation. First, some languages make use of special verbs to deny a sentence: it is the case of many Polynesian languages, in which negative verbs have scope over the entire clause. Instead, in languages such as Evenki (spoken in Siberia), the sentence’s negative meaning can be conveyed by means of a negative marker which has the properties of a finite auxiliary. A third type of negative markers are those anchored in the inflectional morphology of the verbs. For instance, Turkish (1) expresses sentential negation by means of a negative morpheme placed between the verbal stem and the temporal and personal inflectional affixes. (1) John elmalari ser-me-di John apples like.neg.pst.3sg ‘John doesn’t like apples’

The last class of strategies is represented by the use of negative particles to express sentential negation. These markers can be attached to the finite verb (e.g., in Czech) or not, as in the case of Modern Italian (Zanuttini 1997, 2001; Zeijlstra

105

106

Marta Tagliani & Jelena Živojinović

2007). In fact, Italian makes use of the negative particle non, a free-standing morphological word which occupies a fixed position to the immediate left of the finite verb: (2) Gianni non ha telefonato Gianni not aux call.pp ‘Gianni didn’t call’

In her influential works on negation in Romance varieties, Zanuttini (1997, 2001) developed a series of tests with the aim of determining the syntactic status of negative elements. Significantly, the analyses1 of different syntactic phenomena (e.g., clitic climbing), provided evidence that preverbal negative markers that attach to a finite verb such as the Italian non are syntactic heads having the entire VP as complement. Additional evidence comes from the blocking of verb movement attested in Paduan (Poletto 2000; Zanuttini 2001). In this northern Italian dialect, yes/no interrogative clauses require the C0 to be overtly filled. As a consequence, in interrogatives such (3a), the main verb moves from V0 to C0. However, the Head Movement Constraint (Travis 1984) prohibits this V-to-C movement in the presence of an overtly filled intervening head. Significantly, this is exactly what Zanuttini (2001) found for negative interrogatives (3b), confirming that the Paduan negative marker no is an intervening head which blocks the syntactic verb movement. (3) a.

Vien-lo? comes-he ‘Is he coming?’ b. *Vien-lo no? comes-he not? ‘Isn’t he coming?’

Further confirmation of the syntactic head status of preverbal negative particles has been provided by Merchant (2006), who developed the so-called “why not” test. Given that the “why not” construction is considered as a form of phrasal adjunction, Merchant (2006) predicted that such construction should not be allowed in those languages in which the negative markers are syntactic heads instead of phrasal elements. This is precisely what happens in many languages with preverbal 1. The same diagnostics have been used to investigate the syntactic status of those negative particles whose position remains unaffected by the surface position of the finite verb (e.g., nicht in German). Empirical evidence from verb movement analyses and the why not test in V2 languages provide support for the assumption that negative adverbs such as nicht behave instead as phrasal elements (i.e., XPs), thus occupying a different position in the clausal spine than preverbal negative markers.

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

negative markers, including Modern Italian (4), where the negative particle non cannot participate in the “why not” construction and the anaphoric form no must be used in its place. (4) *Perché non? why not ‘Why not?’

As said, Modern Italian makes use of the preverbal negative marker non to introduce the semantic negation. Nonetheless, sentential negation can be expressed through various other syntactic constructions. In (5), the semantic negative meaning is conveyed by the single neg-word2 nessuno placed preverbally, that behaves as a negative quantifier (e.g., the English nobody or no-one). No negative marker is required in such syntactic configuration. (5) Nessuno ha telefonato n-body has called ‘Nobody called’

The combination of the negative marker with a neg-word within the same sentence may yield different semantic interpretations. Specifically, when the preverbal non comes with a negative quantifier placed post-verbally as in (6), the sentence expresses one single semantic negation. The sentence in (6) is an instance of negative concord (NC): two or more negative elements that are able to individually express negation yield instead a single negative meaning when combined (Labov 1972). Being a non-strict NC language (Giannakidou 1997, 2000; Zeijlstra 2004), Modern Italian permits a negative concord reading only in syntactic configurations such as (6), that is, between a negative element in preverbal position (either a neg-word or a negative marker) and neg-words in postverbal position. (6) Gianni *(non) ha telefonato a nessuno Gianni not aux call.pp to n-body ‘Gianni didn’t call anybody’

NC constructions are however subject to syntactic locality constraints (Zeijlstra 2004): to provide a single semantic negative reading, all the negative elements must share the same scope domain. Instead, when the relationship is established between negative elements belonging to different clauses, e.g., the main clause and the subordinate one3, the interpretation is of double negation: along the lines of propositional logic, two negative elements cancel each other out yielding an affirmation. 2. In Laka’s (1990) terminology for negative indefinites in NC languages. 3. As long as the embedded clause is in the indicative mood.

107

108

Marta Tagliani & Jelena Živojinović

The NC reading of multiple negation constructions is also compromised when the neg-word combined with a negative marker is placed in preverbal position. In this case, a double negation (DN) reading of sentence (7) can be accessed modulo specific prosodic features.4 (7) Nessuno non ha telefonato n-body not aux call.pp dn reading: ‘Nobody didn’t call’ *nc reading: ‘Nobody called’

As we aim to provide the reader with an overview of how sentential negation is expressed in Modern Italian in the introductory section, we will not discuss the different proposals that have been made to account for the semantic phenomenon underlying negative concord (cf. Zeijlstra 2007, 2008 for a detailed discussion).

1.2 Grammaticalization of mica in standard and local varieties Jespersen's Cycle (1917) describes the diachronic development of negative particles and expressions across a variety of languages, providing evidence that negation systems with preverbal negative markers (Stage I) undergo the loss of this syntactic configuration. This linguistic process occurs gradually5: languages generally go through an intermediate stage of discontinuous negation (Stage II), in which the preverbal negative marker appears together with a postverbal negative element, which has mainly a supporting function as the original negative marker has become phonologically too weak to express negation by itself (cf. van der Auwera 2009 for a different account). In the final phase of the process, the original preverbal marker has disappeared (Stage III), and the postverbal element conveys the negative semantic meaning to the sentence. A well-known example of the cycle development is the French language (Zeijlstra 2004), where the lexical item pas (lit. ‘step’) has been reanalysed as a negative element in Modern French, occurring in a complex negative form together with the negative clitical element ne (9b). Evidence for a later stage is found in Colloquial French6 (9c), with the adverbial element pas being reinterpreted as a single negative marker.7 4. The primary stress of the sentence must be on the neg-word placed preverbally. 5. On the gradual character of diachronic change, cf. Carlier et al. (2012). 6. Note that, in Colloquial French, verbs such as s’inquiéter and s’occuper can even be used without any formal negation, but with negative meaning: T’occupe ! ‘don’t bother’. 7. Zeijlstra (2004: 52–57) provides a very detailed account on Jespersen’s cycle illustrating a total of 7 stages of diachronic development of sentential negation. Specifically, we find one negative marker preceding or following the finite verbs at Stage I. The marker becomes phono-

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

(9) a.

Je ne di I not say b. Je ne dis pas I not say pas c. Je dis pas I say pas ‘I do not say’

(Old French – Stage I)8 (Modern French – Stage II) (Colloquial French9 – Stage III)

In the next section, we describe the development of the item mica, a non-negative lexical item which is attested in both Modern Italian and local varieties in sentences conveying a negative meaning. As we will see throughout the discussion, the use of mica is not homogeneous across standard and non-standard varieties: while in the former it encloses a presuppositional feature, in the latter it seems to have undergone a grammaticalization path similar to that of the French pas, becoming a postverbal marker of sentential negation. 1.2.1 Mica in Modern and Old Italian Etymologically, mica indicated a quantity noun derived from the Latin mica(m)10 lit ‘crumb’, employed as a minimizer to express the smallest portion of an event (Garzonio 2019). (10) On stéde sciscerie miga de vin d’intrà one bushel of chickpeas and miga of wine of income ‘One bushel of chickpeas and a little of wine as income’ (Old Milanese, XVth century, Lancino Curti 6.12)11

logically too weak at Stage II. Subsequently, at Stage III, the phonologically weak marker is accompanied by an adverbial negative marker, resulting in a two-element sentential negation. The adverbial negative marker becomes obligatory, whereas the negative marker attached to the finite verb becomes optional at Stage IV. At Stage V, the negative marker preceding or following the finite verb is no longer available. Eventually, the negative adverb can maintain its position or appear in the form of a negative marker attached to the finite verb. Finally, at Stage VII, sentential negation is represented by one negative marker. 8. Example taken from Zeijlstra (2004: 54) 9. Also attested in reported conversations, popular literature, publicity, etc. 10. mica is still attested as a lexeme in some varieties, such as Catalan (cf. Solà Cortassa et al. 2008) and Norman French, e.g. Il ne mange miettece midi ‘He won’t have lunch today’ (cf. Grevisse & Goosse 2016). 11. Examples given in this section are taken from Garzonio (2019).

109

110

Marta Tagliani & Jelena Živojinović

The quantitative noun mica (or miga/mia in northern varieties) gradually underwent a process of grammaticalization12, becoming a postverbal marker of sentential negation, attested from the 19th century onwards (cf. Visconti 2009). Two competing forms are attested, mica (11) and né mica (12), the latter including the morpheme né attested also in Modern Italian neg-words (e.g., nessuno, ‘no one’) and probably derived from the Latin negative coordination particle nĕc. Hansen and Visconti (2009) argued that the complex form né mica is the older form. This was attested in earlier texts where it was sometimes also employed as a quantifier in similar fashion to other negated minimizer which have been grammaticalized as negative quantifiers (e.g., the Old Lombard negota ‘nothing’ from ne(c) gutta(m) lit. ‘not a drop’, cf. Middle Fr. Il ne parloit goutte ‘He didn’t speak at all’). The form né mica was then gradually substituted by the simpler form mica, although the two display a different syntactic distribution. When either form is in postverbal position, it always appears with a preverbal negative marker (11, 12). However, while the bare item mica can only surface post-verbally, the complex form né mica can also appear in preverbal position (13). (11) La grandezza delle magioni non cessa mica la febbre det largeness of + det houses not extinguishes mica det fever ‘The large size of a home does not extinguish the fever’ (Tesoro volgarizzato, 7.68) (12) Dio non ajuta nè-mica per preghiera se l’opera non seguita la God not help.3sg not + mica for prayer if det + deed not follow.pstp det pieganza. prayer ‘God doesn’t help due to a prayer if there is no deed following it’ (Trattato di virtù morali,17) (13) Federigo di Stuff ogià né-mica par che si celi… Frederick of Hohenstaufen already not + mica seem.3sg that refl hides.3sg ‘It already seems that Frederick of Hohenstaufen is not hiding …’ (Monte Andrea, Rime, tenz. 8, son. 1)

Furthermore, Garzonio (2019) reported instances of corrective constructions that display a complex form constituted by mica preceded by another negation, either né or non/no (14). The tendency to use this complex form is largely attested in both earlier and later texts. Remarkably, corrective focus constructions are con12. In this article, grammaticalization is intended in Meillet’s terms as a diachronic process by which an item moves from a lexical category into a more grammaticalized one, or from a grammatical one into an even more grammaticalized one (cf. Meillet 1912, but also Traugott & Heine 1991).

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

sidered by Visconti and Hansen (2009) to be at the origin of the presuppositional meaning that mica will eventually acquire in Modern Italian. (14) E qui s’intende di riso sfrenato e del continovato, and here impers + mean.3sg of laugh uncontrolled and of + det continued non miga della faccia rallegrare not mica of + det face amuse.inf ‘And here one means uncontrollable and persisting laugh, not just brightening.’ (Francesco da Barberino, Reggimento e costumi di donna, 1.6)

The distribution of mica in Modern Italian seems therefore mapped into that of the complex form né (or non) mica rather than that of the bare item mica in Old Italian. While an exhaustive explanation of the syntactic properties of the constructions with mica go beyond the aims of this paper13, it is worth underlying that the grammaticalization process of the item mica already began when there was an alternation between forms with and without the morpheme né, which has been lost during the diachronic development of the construction. This is a typical case of overlap, which is an indication of an ongoing grammaticalization process (cf. Lamiroy & De Mulder 2011: 302–318). In Modern Italian, mica has progressively undergone the loss of its original semantic component: nowadays, it only maintains its presuppositional feature (cf. Cinque 1976; Bernini & Ramat 1996; Zanuttini 1997; Pescarini 2009), supporting negation, as in the following example adapted from Tagliani & Rabanus (in press): (15) Speaker A: Metti il cappotto se esci. ‘Put on the coat if you go outside.’ Speaker B: Ma non fa mica freddo! but not make.3sg mica cold ‘But it is not cold!’

(Standard Italian)

In (15), mica appears as a discontinuous element of the negative syntactic construction, placed in postverbal position. In such pragmatic context, its occurrence is licensed by the presupposition that the weather is cold. 14 The preverbal position in Modern Italian may also be filled with the so-called negative adverb mica (cf. Manzini & Savoia 2002) as the only negative element of the clause, for instance:

13. The reader is referred to Garzonio (2019) for a comprehensive discussion. 14. It is necessary to note that certain examples with mica are not considered as Standard Italian and as such, may not be accepted by all speakers.

111

112

Marta Tagliani & Jelena Živojinović

(16) Mica ci vado. not there go.1sg ‘I’m not going there’

However, Garzonio (2019) does not consider the preverbal mica as a full negative marker replacing the existing particle non in the syntactic configuration, but rather as a focalized neg-word such as niente ‘nothing’ or nessuno ‘nobody’. The following examples illustrate how mica has the same distribution of other negwords in both negative (17) and interrogative (18) contexts: while in the former the preverbal marker non is needed to license its presence post-verbally, in the latter the same postverbal element is acceptable without any other negative element. (17) a.

Non mangia mica not eat.3sg mica ‘S/he is not eating’ b. Non mangia niente not eat.3sg nothing ‘S/he eats nothing’

(18) a.

Hai mica mangiato?15 aux mica eat.pp ‘Have you eaten?’ b. Hai mangiato niente? aux eat.pp nothing ‘Have you eaten anything?’

As previously mentioned, similarly to other neg-words, preverbal mica can be the only negative item in a sentence. However, from this position, it cannot license polarity adverbs such as mai ‘never’ or ancora ‘yet’.16 Further evidence that preverbal mica does not exhibit the same syntactic properties asthe standard marker non comes from corrective focus constructions, where a constituent can be negated by the bare mica only if it is in the left peripheral focus position (19b) or in elliptical fragments (Manzini & Savoia 2011). Instead, when the negated constituent does not move to a focus position, the use of preverbal mica in place of non (19a) is ungrammatical.

15. In polar questions, mica can occur combined with the preverbal negative marker such as Non hai mica mangiato? ‘Haven’t you eaten?’ (Pescarini & Penello 2012). 16. It can, however, appear in combination with a polarity adverb if in a postverbal position, such as Non ci sono mica mai andato ‘I have never gone there’.

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

(19) a.

(*)Nella salsa ho dimenticato (mica)/non il sale, ma il pepe. In + det sauce aux forget.pp mica not det salt but det pepper ‘I didn’t forget the salt in the sauce, but the pepper’ b. Mica il sale ho dimenticato. mica det salt aux forget.pp ‘It is not the salt that I have forgotten’

This evidence indicates that in Modern Italian mica is not inherently negative, but it rather displays the same licensing properties of other neg-words: namely, it can either appear under the scope of the standard negative marker non (i.e., in postverbal position) or as a stand-alone lexicalized element surfacing in a left peripheral focus position17 (Rizzi 1997, Benincà & Poletto 2004). According to Garzonio (2019), the development of the distribution of mica from Old to Modern Italian is intrinsically related to the loss of the morpheme né. In Old Italian, when mica is not in the scope of the preverbal negation non (e.g., in corrective constructions or in preverbal position), the negative additive Focus head must be lexicalized by means of the morpheme né in order to provide a non-veridical environment for the licensing of mica in the Existential Projection. As a result of a grammaticalization process (cf. Roberts & Roussou 2003), the lexical item mica has raised in the morpho-syntactic structure in Modern Italian, surfacing in the Focus position in the internal functional layer. This explains the need for a negative marker or for the particle né in such syntactic configurations. 1.2.2 Mica in Italo-Romance varieties After an overview of mica as sentential negation in Modern and Old Italian, we now focus on an outline of the existing literature on analogous constructions in modern Italo-Romance varieties. As pointed out in Penello & Pescarini (2008: 43), some Italo-Romance varieties display the postverbal mica particle obligatorily. This seems to be the case of some Veronese areas, such as Ca’ degli Oppi, whose speakers judged the sentences without mica as ungrammatical. For instance: (20) Ci no pol mia vegner? who not can mica come.inf ‘Who cannot come?’

17. This is a case of ‘syntactic parasitism’ (Garzonio & Poletto 2015), in which a lexical item carrying a given feature can surface in a structural position that normally encodes one of the features it is associated with (e.g., Focus in the case of mica).

113

114

Marta Tagliani & Jelena Živojinović

(21) No bizogna mia rivar tardi not need.3sg mica arrive.inf late ‘You don’t have to show up late’ (Illasi; ASIt Questionario n. 2 – Posizione: 27; cf. Dohi 2019: 54)

An analogous context may be found in some Emiliano microvarieties, displaying an obligatory postverbal marker which etymologically indicates a small quantity, such as mia, brisa, menga (cf. Benincà & Damonte 2009). (22) N va menga not go.3sg menga ‘It is not going/working’

(Modena)

(23) N va mia not go.3sg mica ‘It is not going/working’

(Carpi)

On the other hand, Penello and Pescarini (2008)18 claim that some Paduan microvarieties show an optional use of the postverbal mica, such as Carmignano in (24). (24) No so (mia) chi che laverà i piatti not know.1sg mica who that wash.fut.3sg det dishes ‘I don’t know who will wash the dishes’

In addition to examples exhibiting discontinuous negation, we find varieties with the local form of mica used as free-standing postverbal negator, such as for instance the Swiss Lombard dialects (Pescarini & Donzelli 2017) and some microvarieties of Bresciano (Tagliani & Rabanus in press). Such configurations are confirmed by Pescarini (2022), who argues that Italo-Romance shows an N1 type of negation with just a preverbal negative marker, a discontinuous N1 + N2 negation with a preverbal and a postverbal marker and N2 negation exhibiting only a postverbal marker. In his dataset comprising both Italo-Romance and GalloRomance varieties, N2 negation with mica is attested in Ticinese from Lugano, Comasco, Bergamasco, Cremonese, Grigionese, Novarese and in a very restricted set of examples of Mantovano and Milanese.

18. Also repoted by Dohi (2019: 54–58).

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

2.

Methodology and research questions

Our study is a follow-up to the empirical research conducted in Tagliani and Rabanus (in press) on two varieties of the Italo-Romance dialect known as “Bresciano”, aiming to broaden the investigation in a comparative perspective. For the purpose of the study, we compared these previous data with those from two varieties of “Veronese”, another Italo-Romance dialect (the western variety of Venetan) spoken in the northeast of Italy in the province of Verona. A translation task from Modern Italian into the speaker’s local variety was conducted in a central and a peripheral area (cf. Figure 1), specifically, in Brescia (central Bresciano) Lumezzane (peripheral Bresciano), Verona (central Veronese) and Nogara (peripheral Veronese).

Figure 1. Fieldwork areas

The study involves a total of 16 participants from the listed areas. The questionnaire (cf. Appendix) targets the following contexts for the investigation of negative markers: indicative sentences, protasis, imperatives and subjunctive subordinates. When possible, we compared the gathered data with examples extracted from the

115

116

Marta Tagliani & Jelena Živojinović

AIS Atlas19 in order to provide a diachronic perspective to the analysed data. Our attempt is therefore (1) to understand whether the Bresciano and Veronese areas are similar in terms of the expression of sentential negation, (2) to determine the degree of diatopic variation between the central and the peripheral area, (3) to examine Jespersen’s cycle (1917) within the analysed constructions.

3.

Data analysis

As mentioned in the previous section, we examine four different contexts of negation following Ashby (1981) and Pescarini and Donzelli (2017) who argue that certain constructions are more resistant to the loss of preverbal negation. Specifically, Ashby (1981) proposes the following hierarchy of syntactic contexts from the most to the least likely structures to undergo the loss of preverbal negation: indicative sentences > protasis > imperatives > subjunctive subordinate > negative concord. We will now examine one instance per context by comparing the AIS data with the gathered Veronese and Bresciano constructions.20

Indicative Ita. Questa donna non mi piace ‘I don’t like this woman’ (AIS 1678) (25) a.

La ma pias mia she to_me like.3sg mica b. Chesta fonna la ma pias mia this woman she to_me like.3sg mica

ais_lum21 new_lum, new_bs

19. The Linguistic and Ethnographic Atlas of Italy and Southern Switzerland (AIS) is a collection of maps of Italy and Southern Switzerland comprising lexemes, but also structures of a large set of microvarieties in these areas. The data collection was approximately carried out between 1919 and 1935. 20. For the detailed discussion of the data on the Bresciano varieties (also from a diastratic perspective) the reader is referred to Tagliani and Rabanus (in press). 21. AIS_LUM corresponds to Lumezzanese data extracted from the AIS Atlas. NEW_LUM indicates Lumezzanese fieldwork data, whereas NEW_BS stands for central Bresciano fieldwork data. Similarly, AIS_VER and AIS_CER stand for the Veronese and Cerea data extracted from the AIS Atlas, whereas NEW_VER and NEW_NO indicate central Veronese and Novara data. The discrepancy between the peripheral Veronese AIS and fieldwork data is due to the absence of data of the Nogara microvariety. As a consequence, we opted for the closest attested area in the atlas, which is Cerea.

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

(26) a.

Sta dona mi no la me piaze this woman to_me not her to_me like.3sg b. Sta dona no la me piaze this woman not her to_me like.3sg c. Sta dona no la me piaze mia this woman not her to_me like.3sg mica d. Sta dona no la me piase this woman not her to_me like.3sg

ais_ver ais_cer new_ver new_no

The comparison between (25a) and (26a–b) highlights a difference between the AIS Bresciano and Veronese data. On the one hand, the Lumezzanese AIS construction, only makes use of a postverbal marker which in Zeijlstra’s (2004) terms is a fully grammaticalized negative particle. On the other hand, the Veronese AIS structure shows a first stage no(n) preverbal negation. Fieldwork data collected in the area of Brescia illustrates a consistent use of the third stage negation, whereas the Veronese data shows diatopic variation. In particular, data collected in Nogara reveals the AIS pattern with just a preverbal or first stage negation, thus indicating no diachronic variation. Instead, the data for Central Veronese display discontinuous negation, placing itself at a later stage of the cycle than the peripheral variety spoken in Nogara.

Protasis Ita. Se non mangiamo ‘If we don’t eat’ (AIS 1278) (27) a.

Se non le manga if not her eat.3sg b. Se mangiom mia if eat.1pl mica

(28) a.

Se no magnemo if not eat.1pl b. Se non mangemo mia if not eat.1pl mica c. Se non magnemo if not eat.1pl

ais_lum new_lum, new_bs ais_cer new_ver new_no

Protasis shows very little variation when compared to indicative sentences in (25) and (26). Specifically, the Lumezzanese AIS example makes use of a preverbal negative marker, whereas the third stage negation is constant throughout the modern Bresciano data. The Veronese data representing protasis are analogous to that of the indicative: the AIS data reveal a preverbal negation strategy, which is still attested nowadays in Nogara. The new data gathered in the Veronese area

117

118

Marta Tagliani & Jelena Živojinović

exhibit diatopic variation also for the protasis context, with Central Veronese showing instances of discontinuous negation, typical of the second stage of the cycle.

Imperatives Ita. Non cadere! ‘Don’t fall!’ (AIS 1621) (29) a.

No burlà ado not fall.inf down b. Biösca mia fall.2sg.imp mica c. Burla mia zo fall.2sg.imp mica down

(30) a.

No sta mia cascar not stay.2sg.imp mica fall.inf b. Non cascar mia not fall.inf mica c. No cascar not fall.inf

ais_lum new_lum new_bs ais_cer new_ver new_no

As for imperative constructions, Bresciano presents a pattern similar to the previous syntactic contexts, with a first stage negation provided by the AIS data and a third stage negation in the modern central and peripheral data. Instead, the Veronese data provide a relatively different construction when it comes to the AIS peripheral example. Indeed, the imperative seems to be the only context which has discontinuous two stage negation, which is also attested in modern central Veronese (and in previous contexts for the same Veronese variety), but it is not found in the modern peripheral construction. Notwithstanding, the AIS_CER example in (30a) represents a different syntactic structure, with a presuppositional value that seems to be missing from the instances in 30 (b–c). As such, it does not seem to fall into the configuration in examination.

Subjunctive subordinates Ita. Che non lo troviate ‘That you don’t find it’ (AIS 1651) (31) a.

Che nol troaghef that not + it find.2pl.subj b. Che el truif mia that it find.2pl.subj mica

ais_lum new_lum, new_bs

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

(32) a.

Che no lo cate that not it find.2pl.subj b. Che non lo cati mia that not it find.2pl.subj mica c. Che no lo caté that not it find.2pl.subj

ais_cer new_ver new_no

Subjunctive subordinates follow the tendency identified in the previous contexts. In fact, the AIS_LUM data is at the first stage of Jespersen’s cycle, whereas the modern Bresciano data only show the third stage with no diatopic variation. As for Veronese, the peripheral AIS and modern data only use the preverbal no(n) negation, whereas the modern central Veronese displays consistent second stage negation with both preverbal and post-verbal markers.

Discussion and conclusion In light of the observations presented in the section on data analysis, let us now provide some further considerations on sentential negation in Bresciano and Veronese, along with a summary of the above-mentioned constructions. The Bresciano dialects spoken in both the central and the peripheral area display a homogeneous pattern of negative syntactic constructions, characterized by the presence of the post-verbal adverb mia as stand-alone negative element. A negation system of this kind indicates that the Bresciano varieties have already reached the third stage (Tagliani & Rabanus in press).As for the central area, the absence of AIS data prevents us from providing any indication on diachronic variation. As for the peripheral area, the AIS example in (25a) suggests that the postverbal negation in indicative contexts could have been productive in earlier stages of Lumezzanese as well. This seems to follow the hierarchy of syntactic contexts which are more likely to undergo the loss of the preverbal negative marker (cf. Ashby 1981). Regarding the new Veronese data, two main observations can be made. In a diachronic perspective, the comparison between the historical data collected at the AIS datapoint of Cerea and the new data gathered in Nogara exhibits no variation in the negative system. In fact, we report a homogeneous pattern of preverbal negation, typical of language varieties at the first stage of Jespersen’s cycle (1917). In this respect, we speculate that the instance of discontinuous negation reported in (30a) is not a relevant instantiation of the syntactic configuration under scrutiny as the particle mia seems to have a presuppositional value similar to that of Modern Italian (cf. Cinque 1976). In a diatopic perspective, we attest relevant and consistent differences between the central and the peripheral varieties

119

120

Marta Tagliani & Jelena Živojinović

of Veronese. The gathered data show in fact a discontinuous negation system for central Veronese, while the use of preverbal negative structures in the peripheral area. Significantly, this use in the two varieties is consistent across the four analyzed syntactic contexts. This allows us to locate the peripheral and the central Veronese varieties at stage one and two of Jespersen’s cycle (1917), respectively. Summing up, the four areas of interest present homogeneous negation patterns across the syntactic contexts under investigation. All three stages of Jespersen’s cycle (1917) are represented as follows: I. Stage (peripheral Veronese) II. Stage (central Veronese) III. Stage (central and peripheral Bresciano) Based on these observations, we conclude that the Bresciano and Veronese areas display different negation systems and as such, are not similar. Moreover, we report diatopic variation only within the Veronese area which is in earlier stages of Jespersen’s cycle than Bresciano, not featuring the particle mica as a fully grammaticalized and a purely functional element to the negation and being instead characterized by preverbal (peripheral Veronese) and discontinuous (central Veronese) negative structures. Finally, peripheral Veronese exhibits a crystallized negative system with respect to that attested at the time of the AIS data collection, showing no evidence of diachronic development towards later stages of the cycle (cf. peripheral Bresciano).22

List of abbreviations aux det fut imp inf neg

Auxiliary Determiners Future Imperative Infinitive Negation

pl pst pp refl sg subj

Plural Past Past Participle Reflexive Singular Subjunctive

22. In the discussion, we have not taken into consideration examples of negative concord structures, as the AIS Atlas lacks such configurations for the Veronese varieties. Although the pattern identified in our research for such varieties seem very consistent, further research is needed to establish whether negative concord structures may present remnants of earlier negation stages, as assumed in Ashby (1981) based on the data for Standard French.

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

Note to the glossing system Whenever the gloss for the temporal indication is not included, we mean the present tense.

References Ashby, William J. 1981. The loss of the negative particle ne in French: A syntactic change in progress. Language 57(3): 674–687. van der Auwera, Johan. 2009. The Jespersen cycle. In Cyclical Change [Linguistik Aktuell/Linguistics Today 146], Elly van Gelderen (ed.), 73–101. Amsterdam: John Benjamins. Benincà, Paola & Poletto, Cecilia. 2004. Topic, focus and V2: Defining the CP sub-layers. In The Structure of IP and CP. The Cartography of Syntactic Structures, Vol. II, Luigi Rizzi (ed.), 52–75. Oxford: OUP. Benincà, Paola & Damonte, Federico. 2009. Varianti sintattiche inter- e intra-individuali nelle grammatiche dialettali. In I parlanti e le loro storie. Competenze linguistiche, strategie comunicative, livelli di analisi. Atti del convegno Carini-Valderice, 23–25 ottobre 2008, Luisa Amenta & Giuseppe Paternostro (eds), 185–194. Palermo: Centro Studi Filogici e Linguistici Siciliana. Bernini, Giuliano & Ramat, Paolo. 1996. Negative Sentences in the Languages of Europe. A Typological Approach. Berlin: Mouton de Gruyter. Carlier, Anne, De Mulder, Walter & Lamiroy, Béatrice. 2012. The pace of grammaticalization in a typological perspective. Folia Linguistica 2: 1–15. Cinque, Guglielmo. 1976. Mica. Annali della Facoltà di Lettere e Filosofia dell’Università di Padova 1: 101–112. Dahl, Östen. 1979. Typology of Sentence Negation. Berlin: Walter de Gruyter. Dohi, Atsushi. 2019. La particella pa nelle varietà del ladino dolomitico con particolare attenzione al fassano. PhD dissertation, University of Trento. Garzonio, Jacopo & Poletto, Cecilia. 2015. On preverbal negation in Sicilian and syntactic parasitism. In Italo-Romance Morphosyntax, Diego Pescarini & Silvia Rossi (eds). A special issue of Isogloss. Open Journal of Romance Languages 2015: 132–149. Garzonio, Jacopo. 2019. Not even a crumb of negation: On mica in Old Italian. In Linguistic Variation: Structure and Interpretation, Ludovico Franco & Paolo Lorusso (eds), 263–282. Berlin: De Gruyter Mouton. Giannakidou, Anastasia. 1997. The Landscape of Polarity Items. PhD dissertation, Rijksuniversiteit Groningen. Giannakidou, Anastasia. 2000. Negative …Concord? Natural Language and Linguistic Theory 18: 475–523. Greenberg, Joseph H. 1966. Synchronic and diachronic universals in phonology. Language 42(2): 508–517. Grevisse, Maurice & Goosse, André. 2016. Le bon usage. Gembloux: Duculot. Horn, Lawrence R. 1989. A Natural History of Negation. Chicago IL: The Chicago University Press.

121

122

Marta Tagliani & Jelena Živojinović

Jespersen, Otto. 1917. Negation in English and Other Languages. Copenhagen: A.F. HØst. Labov, William. 1972. Negative attraction and negative concord in English grammar. Language 48(4): 773–818. Ladusaw, William. 1996. Negation and polarity items. In The Handbook of Contemporary Semantic Theory, Shalom Lappin (ed.), 311–331. Oxford: Blackwell. Laka, Miren Itziar. 1990. Negation in Syntax: On the Nature of Functional Categories and Projections. PhD dissertation, MIT. Lamiroy, Béatrice & De Mulder, Walter. 2011. Degrees of grammaticalization across languages. In The Oxford Handbook of Grammaticalization, Bernd Heine & Heiko Narrog (eds), 292–308. Oxford: OUP. Manzini, M. Rita & Savoia, Leonardo M. 2002. Negative adverbs are neither Adv nor Neg. In Proceedings of North East Linguistic Society 31, Masako Hirotani (ed.), 326–46. Amherst MA: GLSA. Manzini, M. Rita & Savoia, Leonardo M. 2011. Grammatical Categories: Variation in Romance Languages. Cambridge: CUP. Meillet, Antoine. 1912. L’évolution des formes grammaticales. Scientia 6(12): 384. Merchant, Jason. 2006. Why no(t)? Style 40(1–2): 20–23. A Festschrift for John Robert (Háj) Ross. Payne, John R. 1985. Negation. In Language Typology and Syntactic Description, Vol. I: Clause Structure, Timothy Shopen (ed.), 197–242. Cambridge: CUP. Penello, Nicoletta & Pescarini, Diego. 2008. La negazione: Variazione dialettale ed evoluzione diacronica. Quaderni di lavoroASIt 8: 42–56. Pescarini, Diego. 2009. ‘Presuppositional’ negation, modality and the {addressee}. Padua Working Papers 3: 22–27. Pescarini, Diego & Penello, Nicoletta. 2012. L’avverbio mica fra widening semantico e restrizioni sintattiche’. In Linguaggio e Cervello/Semantica. Paper presented at XLII Annual Meeting of the Società di Linguistica Italiana, Pisa, Italy, September 25–26, Pier Marco Bertinetto, Valentina Bambini & Irene Ricci (eds). Rome: Bulzoni. Pescarini, Diego & Donzelli, Giulia. 2017. La negazione nei dialetti della Svizzera italiana. Vox Romanica 76: 74–96. Pescarini, Diego. 2022. A quantitative approach to microvariation: Negative marking in central Romance. Languages 7: 87. Poletto, Cecilia. 2000. The Higher Functional Field. Evidence from Northern Italian Dialects. Oxford: OUP. Rizzi, Luigi. 1997. The fine structure of the left periphery. In Elements of Grammar [Kluwer Handbooks of Linguistics 1], Liliane Haegeman (ed.), 271–327. Dordrecht: Kluwer. Roberts, Ian & Roussou, Anna. 2003. Syntactic Change: A Minimalist Approach to Grammaticalization. Cambridge: CUP. Solà, Joan, Lloret, Maria Rosa, Mascaró, Joan, Pérez Saldanya, Manuel. 2008. Gramàtica del català contemporani. Barcelona: Editrial Empuries. Traugott, Elizabeth Closs & Heine, Bernd (eds). 1991. Approaches to Grammaticalization, Vol. 2: Types of Grammatical Markers [Typological Studies in Language 19:2]. Amsterdam: John Benjamins.

Chapter 4. Tracking Jespersen’s cycle in Veronese and Bresciano

Tagliani, Marta & Rabanus, Stefan. (In press). The cycle in language change: Insights from diachronic phonology and syntax of negation. Evolutionary Linguistic Theory 4(2), 191–228. Travis, Lisa DeMena. 1984. Parameters and Effects of Word Order Variation. PhD dissertation, MIT. Visconti, Jacqueline. 2009. From “textual” to “interpersonal”: On the diachrony of the Italian particle mica. Journal of Pragmatics, 41(5): 937–950. Visconti, Jacqueline & Hansen, Maj-Britt Mosegaard. 2009. On the diachrony of “reinforced” negation in French and Italian. In Grammaticalization and Pragmatics: Facts, Approaches, Theoretical Issues, Corinne Rossari, Claudia Ricci & Adriana Spiridon (eds), 137–71. Leiden: Brill. Zanuttini, Raffaella. 1997. Negation and Clausal Structure: A Comparative Study of Romance Languages. Oxford: OUP. Zanuttini, Raffaella. 2001. Sentential negation. In The Handbook of Contemporary Syntactic Theory, Mark Baltin & Chris Collins (eds). Hoboken NJ: Wiley & Sons. Zeijlstra, Hedde. 2004. Sentential Negation and Negative Concord. PhD dissertation, University of Amsterdam. Zeijlstra, Hedde. 2007. Negation in natural language: On the form and meaning of negative elements. Language and Linguistics Compass 1(5): 498–518. Zeijlstra, Hedde. 2009. On French negation. Annual Meeting of the Berkeley Linguistics Society 34(1): 447–458.

Appendix List of sentences used for the translation task from Modern Italian 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Questa donna non mi piace ‘I do not like this woman’ Stanotte non dormirò ‘Tonight I will not sleep’ Maria non vuole rimanere ‘Maria doesn’t want to stay’ Non capisco cosa vuoi dire ‘I don’t understand what you are trying to say’ Perché non vi sposate? ‘Why don’t you get married?’ Lui non sarebbe contento ‘He wouldn’t be happy’ Se non mangiamo perdiamo peso ‘If we don’t eat, we will lose weight’ Non ti muovere! ‘Don’t move!’ Non cadere! ‘Don’t fall!’ Stai attento che le galline non vadano in giardino ‘Be careful so that the chicken don’t go into the garden’ Mi meraviglio che non lo troviate ‘I am surprised that you can’t find it’ Non dormo mai il pomeriggio ‘I never sleep in the afternoon’ Lui non corre mai ‘He never runs’ Maria non ha mai fretta ‘Maria is never in a rush’ Non si muoveva più ‘He wasn’t moving any longer’ Sei venuta senza niente ‘You came empty handed’ Non è ancora matura ‘It is not ripe yet’

123

chapter 5

Could be, might be, maybe Mechanisms of grammaticalization in synchronic use and perception David Lorenz

Universität Rostock | Lunds Universitet

In grammaticalization, functional reanalysis and formal reduction are often regarded as elements of a unified diachronic process, though rooted in general communicative and cognitive preferences. The present study tests these claims in synchronic language use by investigating potential cases of grammaticalization. Epistemic phrases of the type (it) could/might be (that) in English are potential candidates for grammaticalizing into sentence adverb(ial)s. The question is whether shorter forms (here, it-omission) are preferred in potentially grammaticalizing contexts, e.g. modifying a main clause ((it) could be this is correct). I first summarize a corpus study, where overall higher rates of it-omission are found in critical context across items (could be, might be) and register (spoken, informal writing). A ‘continuous shadowing’ experiment partly confirms this finding but also shows that speakers/hearers are both more flexible and more conservative with could/ might be than with maybe / it may be that. The findings suggest that grammaticalizing contexts have an immediate effect on formal reduction even in the absence of change, and that language users have an active intuition for emerging variational patterns. Keywords: potential grammaticalization, synchronic usage, cognitive mechanisms of grammaticalization, epistemic phrases, adverbialization, morphological erosion

1.

Introduction

Grammaticalization as a type of language change is a multiplex phenomenon. It involves shifts on many levels (semantic and pragmatic, syntactic and morphological, phonetic and phonological) through different processes (e.g. divergence, specialization, cf. Hopper 1991), along several parameters (such as context extension,

https://doi.org/10.1075/slcs.232.05lor © 2023 John Benjamins Publishing Company

Chapter 5. Could be, might be, maybe

desemanticization, decategorialization, and erosion, cf. Narrog & Heine 2021, see also Lehmann 2015[1982]), and motivated by various mechanisms (such as reanalysis and analogy, Hopper & Traugott 2003, or automatization and categorialization, Bybee 2003; Lehmann 2017). A main tenet is that grammaticalization is a machinery in which these different parts operate together and interdependently. This leads to the question of how this machinery is built, or why “processes that need not be causally connected to one another are, nonetheless, so often observed to cluster” (Wiemer 2014: 426). As linguistic structures are created and replicated by speakers in interaction, “the conditioning factors for grammaticalization must in the first place be found at the speaker level” (Fischer 2007: 115–116). Thus, taking a cognitively oriented, usage-based approach, we can look for answers in investigating “fluent patterns of [synchronic] language use” (Hopper & Traugott 2003: 2), emerging discourse conventions (cf. Boye & Harder 2012), and the way humans process, store and replicate pieces of linguistic structure (cf. Bybee 2003; Lehmann 2017). Consequently, grammaticalization research should not only be concerned with long-term changes across generations of speakers, but also with spontaneous language use and perception on the individual level (cf. Fischer 2010; De Smet 2016; Petré & Van de Velde 2018). Gaeta (2004: 45) notes that “a lot of apparently messy structures are the result of very well motivated changes that, far from being whimsical, shed (at least some) light on the way language users perceive the world around them and on the language they use to express it”. Such statements are often made in a retrospective view of completed changes. The present study looks at it from the other end, focusing on potential grammaticalization that is not observably in progress. If general cognitive-communicative mechanisms and preferences are the source of gradual and complex change, they must be traceable even when change does not occur. The case study presented here investigates the epistemic phrases (it) could be (that) and (it) might be (that) in British English. They are seen as ‘candidates’ for grammaticalization towards sentence adverbs, which is a possible reading of might be in (1), but not of could be in (2) (examples from the Spoken BNC2014, Love et al. 2017). (1) should I do it for you? Might be there’s a knack to this (Spoken BNC2014 SUAB 95) (2) […] it could be that brainwaves can be transferred without physical contact (Spoken BNC2014 SM8P)

Based on a previous corpus study (Lorenz forthcoming), I will report on a ‘continuous shadowing’ experiment, which elicits spoken data and taps into the mental representations of language. The main focus is on the relation of the complmen-

125

126

David Lorenz

tizer that and the expletive pronoun it. On a grammaticalization pathway, a zero complementizer would put the epistemic phrase in the syntactic function of an adverbial (taking a bare clause complement, which can be analyzed as the main clause). As a consequence, we would expect morphological erosion in omitting the pronoun it (as seen in (1)) and univerbation of could/might + be (see Section 2 for detail). Thus, within the complex machinery of grammaticalization, we are dealing with the connection of (syntactic) function and (morphological) form. To analyze grammaticalization as a process, it can be necessary to dissect formal from functional changes (Detges & Waltereit 2002; Konvička 2019: 299). But the extent to which they are intertwined, especially in actual in-the-moment usage, is less often discussed. It is often assumed or implied that any change in form (bonding/univerbation, phonetic and morphological erosion) is a consequence that follows a change towards more grammatical meaning; but whether this happens immediately or only when the meaning change is in some way complete, or only after a frequency increase, is not always clear. For example, Traugott (2003: 645) offers the following definition of grammaticalization: “The process whereby lexical material in highly constrained pragmatic and morphosyntactic contexts is assigned grammatical function, and once grammatical, is assigned increasingly grammatical, operator-like function.” Apparently, “operator-like function” refers to scope increase as well as univerbation in her examples (indeed, instead of). Then, what are we to make of “once grammaticalized”? Does it apply to any usage event in which an item is “assigned grammatical function”, or does this function first need to become an established option, at least? As for erosion proper, Traugott views it as a property of “[l]ater grammaticalization” (2003: 644), and others associate it with ‘secondary grammaticalization’ (Breban 2014). Similarly, Heine (1993: 58) proposes that “cliticization and erosion normally start later than other shifts”, hence a temporal separation of steps, although he also sees desemanticization, decategorialization and erosion as “different components of one and the same mechanism” (Heine 2003: 578–579). On the other hand, for Lehmann (2015[1982]: 113) “phonological attrition and desemanticization go hand in hand” (see also Lehmann 2004), and Bybee et al.’s (1994: 20) ‘parallel reduction hypothesis’ proposes a “dynamic coevolution of meaning and form” rather than a clear-cut order of events. From a psychological perspective, Barth (2019: 1) finds that “in cases of grammaticalization, it is not an increase in frequency that results in reduction, but a decrease in informativity”, where ‘informativity’ appears to be largely co-extensive with ‘discourse prominence’ in the sense of Boye & Harder (2012). The present paper pursues the hypothesis that there is an immediate connection between function and form that is based in communication and cognition and need not be mediated by diachronic change or a frequency increase. An empirical investigation of epistemic phrases in

Chapter 5. Could be, might be, maybe

English is presented in Sections 2 and 3, and Section 4 provides a discussion of the results on the background given above.

2.

Epistemic phrases and adverbials

It is an attested pathway that sentence adverbs are the outcome of fusion and univerbation of larger, compositional structures. In particular, epistemic adverbs have emerged from “epistemic phrases” (Thompson & Mulac 1991) such as, in English, I think, it seems, it may be; at least in the last case, the process is completed as there is a fully conventional, monomorphemic adverb, maybe (see López-Couso & Méndez-Naya 2016 for its historical development). This adverbialization process is a type of grammaticalization in that it involves a shift towards discursive ancillariness through decategorialization, scope expansion, fusion, and often morphological erosion. What is a matrix verb phrase with propositional content in the source construction (think, seem, may be + complement clause) becomes part of an invariant form (i.e. the adverbial) that modifies a main clause and has scope over a proposition instead of being part of it (cf. Brinton & Traugott 2005: 136; Narrog & Heine 2021: 102). A recurrent type of phrases that can evolve into epistemic adverbs is exemplified by the following items: French peut-être (< îl peut être (que) ‘it can be (that)’), Estonian võib-olla (< (see) võib olla (et) ‘(it) may/can be (that)’) (Ramat & Ricca 1998); Swedish kanske (< (det) kan ske (att) ‘it can happen (that)’), Norwegian kanskje (< (det) kan skje (at) ‘it can happen (that)) (Beijering 2010, 2016; Norde et al. 2014), and English maybe (< it may be (that), López-Couso & Méndez-Naya 2016). In all of these, the source phrase involves a third person singular modal verb of ability/possibility and a main verb with existential meaning, glossed as ‘be’ or ‘happen’. The subject is an expletive impersonal pronoun which can be omitted and is lost in the process of adverbialization (morphological erosion). In addition to these “successful” cases, there are candidate constructions of the same type that have the potential for this adverbialization but are not, as it were, undergoing change along this path. While they may show tendencies towards occasional adverbial-like usage, there is no consistent actuation of change. Cases in point are German (es) kann sein / (es) mag sein. The subject pronoun es can be omitted and the phrase can occur in isolation (3a), but otherwise it does not show the profile of a sentence adverb like vielleicht (‘perhaps’). While the adverb modifies a main clause and is a constituent triggering V2 word order (3b), the phrase takes a subordinate dass-clause by default (3c), or may serve as a parenthetical element external to the main clause (3d).

127

128

David Lorenz

(3) a.

Kommen sie morgen? Kann sein. / Vielleicht. come-3pl they tomorrow can be / perhaps ‘Are they coming tomorrow? May-be/perhaps.’ b. *Kann sein / Vielleicht kommen sie morgen. *can be / perhaps come-3pl they tomorrow ‘Perhaps they are coming tomorrow.’ c. Kann sein / *Vielleicht, dass sie morgen kommen. can be / *perhaps that they come-3pl tomorrow ‘(it) may be that they are coming tomorrow.’ d. ?Kann sein / *Vielleicht sie kommen morgen. ?can be / *perhaps they come-3pl tomorrow ‘(it) may be they are coming tomorrow.’

In English, similar candidates are the epistemic phrases (it) could be and (it) might be. These are attested in various contexts relevant for grammaticalization. In a corpus study, Lorenz (forthcoming) investigates the complementation patterns of these items in spoken and informal written British English, with data from the British National Corpus (Spoken BNC1994 [BNC Consortium 2007] and Spoken BNC2014) and blog writing from the British section of the Corpus of Global Webbased Englishes (GloWbE, Davies 2013). Among others, epistemic (it) could/might be occur as matrix clause with a that-clause complement (4a–b), as well as with a bare clause complement (5a–b). (4) a.

you do know quite a lot so it could be that you might come across as slightly challenging (Spoken BNC2014 S8CW). b. […] so I got used to them early and rarely took them off. Might be that my eyes got even worse because of that […] (GloWbE GB Blogs)

(5) a.

Could be he was truly expecting a sugar skull but got a deviant alteration from a tattooist with a grudge against meatheads that kick 50 shades o shit outta their girls […] (GloWbE GB Blogs) b. […] the landlords in Saigon are if you like more commercial, more capitalist, it might be they were better organized than landlords in the north. (BNC1994 S_tutorial JJN)

In the latter context (5a–b), the epistemic items’ status is ambiguous between main clause (with the complementizer that omitted) and an adverbial modifying the following clause (compare maybe, perhaps). Thus, this is a ‘bridging context’ (Heine 2002) or ‘critical context’ (Diewald 2002, 2006) for grammaticalization, where the item can be structurally reanalyzed and, when it is interpreted as an adverbial, undergoes internal decategorialization and coalescence. Occurrence in a critical context as such is no proof of on-going grammaticalization, of course, especially since that-omission is common and possible irrespective of the matrix clause. But frequent use in these contexts can activate and reinforce the grammat-

Chapter 5. Could be, might be, maybe

icalized (adverbial) reading of the item. A shift from that-clause complementation to modifying a bare clause has been shown to be a crucial step in adverbialization (Thompson & Mulac 1991; López-Couso & Méndez-Naya 2016; Kaatari & Larsson 2019). Additional critical contexts found in the corpus study are the use as an isolated phrase (which is quite frequent in spoken language; ex. 6a-b) and as a parenthetical in non-initial positions (rare in both speech and writing, 7a-b). Other patterns were categorized as ‘pre-stage’ or ‘non-grammaticalizing’ contexts on the grammaticalization pathway. In ‘pre-stage’ contexts, the epistemic phrase cannot be immediately reanalyzed as a sentence adverb, although it serves a similar function, such as with that-clause complements (ex. 8, and 4a-b above) and with phrasal complements denoting a time, distance or quantity (ex. 9). Items in ‘non-grammaticalizing’ contexts contain an expletive subject (it or zero) but lack sentential scope or clear epistemic meaning (examples 10–11). (6) a.

maybe the government are polluting us and killing us off [S0008:] could be (Spoken BNC2014 S5PW) b. cos I’m out of the probationary period [S0592:] they can still fire you [S0597:] I think it might be (Spoken BNC2014 S6Q6)

(7) a.

[…] which puts you about forty […] fortyish could be becoming a dad? (Spoken BNC2014 S7JH) b. they are pretty cool (.) oh no maybe it’s not warm enough to sit outside it might be. (Spoken BNC2014 SMY5)

(8) It could be that Rock will lose to Punk, granting Punk his respect he already deserves. (GloWbE GB Blogs) (9) We know this is a massive success because 30 per cent of participants (it might be 40 – but since the numbers are made up anyway I can’t be bothered going back to check) don’t sign on again for three months. (GloWbE GB Blogs) (10) Meek, insecure, deferential – it could be suggested that Miliband needs some balls. (GloWbE GB Blogs) (11) but I think the assumption is with other phones just it could be me but I’ve never known anyone to have like a software problem with an Iphone (Spoken BNC2014 S4YQ)

Several of the above examples (4b, 5a, 6a, 7a) show omission of the expletive pronoun it. Pronoun omission can occur spontaneously in speech any time, but it can be linked to grammaticalization as morphological erosion, a reduction in form of the (potentially) grammaticalizing item.1 Lorenz (forthcoming) shows that the 1. Note that it-omission has been linked to more grammaticalized readings in other cases, such as evidential ∅ looks/seems like (Mélac 2022; see also López-Couso & Méndez-Naya 2014). For

129

130

David Lorenz

use in critical contexts clearly correlates with it-omission. Importantly, this means that a tendency towards erosion shows already with the ambiguous ‘critical contexts’, without any compelling signs of progressing grammaticalization. (Recall that any occurrence in a critical context does not imply that either the speaker or the hearer ‘reanalyze’ the item, in this case interpreting it as a sentence adverb rather than a phrase.) Figure 1 (from Lorenz forthcoming: Figure 3) shows the summary of it-omission rates by context type in speech (left panel) and blog writing (right panel). Table 1 shows the specific comparison of the rates of it-omission with that-clause complements (see Example 4a-b) and bare clause complements (critical context, 5a-b), which is the most important aspect to the present purpose.

Figure 1. (it) could/might be in the SpokenBNC corpora (spoken British English) and GloWbE-GB-B (blog writing): it-omission rates by grouped contexts Table 1. (it) could/might be in the SpokenBNC corpora (spoken British English) and GloWbE-GB-B (blog writing): it-omission rates with that-clause and bare clause complement complementation pattern

Spoken BNC 1994–2014

GloWbE-GB-B

could be

might be

could be

might be

that-clause

19% (12/63)

 2% (1/55)

 5% (16/323)

 2% (2/130)

bare clause

34% (13/38)

14% (3/22)

27% (13/49)

25% (4/12)

the present purpose, there is no need to assume that it alters the basic interpretation of epistemic possibility of (it) could/might be. Whether it does in nuances is an issue for further study.

Chapter 5. Could be, might be, maybe

If this finding can be generalized, it would mean that erosion/reduction and decategorialization are indeed intertwined. Reduction would be concomitant with reanalysis rather than following in its wake, and could not be put down to side effects such as increasing frequency (as has sometimes been suggested, e.g. Haspelmath 1999: 1062; Bybee 2003). Instead, there may be intuitive preferences in language use that match form to function immediately (and liken items by analogy or schema matching, e.g. could be to maybe; see Van Bogaert 2011 for elaboration). Such preferences would imply a tacit awareness of syntactic differences and their functionality in communication. If these preferences are at work any time, they might go a long way in explaining why grammaticalization seems to unify changes on several levels, and why it tends to follow recurrent, directional pathways (see Haspelmath 2008 for a similar argument from the perspective of linguistic universals). While more evidence is clearly needed to assess or refine this description, the idea that cognitive mechanisms provide the basis for grammaticalization has often been evoked in various ways (e.g. Bybee 2003; Hoefler & Smith 2009; Lehmann 2017). The results of the corpus study come with the caveat that they rest on a relatively small number of tokens in some of the relevant categories, and, especially in the spoken data set, there is a degree of uncertainty in the categorizations. These issues are inherent in the phenomenon under study. If adverbialized uses of could/ might be were highly frequent, we would simply consider them grammaticalized (like maybe); if spoken language followed rigid syntactic linearization, it would not be the arena of innovation and change. Experimental methods can be useful in validating specific corpus findings in a more controlled setting, and in adding the perspective of perception and processing (cf. Gilquin & Gries 2009; Lorenz & Tizón-Couto 2020). Grammaticalization phenomena have, to my knowledge, not often been researched with experimental methods. One exception is a self-paced reading study (maze task) by Hilpert and Correia Saavedra (2018) which in fact could not confirm (and thus cast into doubt) a hypothesized cognitive-communicative mechanism of grammaticalization, i.e. ‘asymmetric priming’ (Jäger & Rosenbach 2008). Another is Croft (2010), who could show the drive for morpho-syntactic innovation in a controlled elicited production setting. Christensen et al. (2021) use letter and text change detection tasks to show that discourse prominence can be measured and indeed captures the grammatical-vs-lexical distinction. These are important contributions which show that the diversity of experimental methods offers vast opportunities to gain insights into the cognitive underpinnings of grammaticalization and language change. The remainder of this article presents and discusses an experimental study that was designed to test the above corpus findings on perception and elicited production.

131

132

David Lorenz

3.

Epistemic phrases/adverbials in an experimental study

For the experimental study a ‘continuous shadowing’ task was created. In this task, participants hear and repeat utterances of pre-recorded narratives or dialogues (cf. Marslen-Wilson & Welsh 1978; Caines 2012). It serves as a window into the stored mental representations that the shadower may have of a given item. When an item in the input is distorted (e.g. “travedy” for tragedy in MarslenWilson & Welsh 1978) or when an expected element is missing (e.g. What ∅ I doing in Caines 2012), shadowers will, at a certain rate, restore it to the expected form that matches their mental representation (“tragedy”, What am I doing). In other words, such ‘fluent restorations’ are a sub-conscious reaction that can reveal the degree to which a given form or structure is intuitively accepted or rejected. In the same way, there may be ‘fluent reduction’ or ‘fluent deletion’ of elements in the input that are felt to be superfluous or out of place. For example, in a similar (non-continuous sentence shadowing) experiment by Lorenz (2012), participants often returned an input form “goinde” as gonna rather than going to, thus reducing rather than restoring material in their response. The target constructions of the present experiment were the epistemic phrases (it) could be (that), (it) might be (that) and (it) may be (that). They were presented in various forms with and without subject pronoun and complementizer: [+it +that] (e.g. it could be that), [-it +that] (∅ could be that), [+it -that] (it could be ∅), [‑it ‑that] (∅ could be ∅; this variant was left out with may be, because this form is a fully conventional adverb and not of interest here). Following the corpus study, it was assumed that the presence or absence of the complementizer that marks the difference between a ‘pre-stage’ and a ‘critical context’ for grammaticalization. The corpus finding that it-omission correlates with critical contexts leads to the following hypotheses for shadowing: i.

Fluent restoration of a missing pronoun it will be more prevalent when the complementizer that is present in the input: [-it +that] >> [+it +that] (e.g. ∅ could be that >> it could be that). ii. Fluent omission or reduction of it will be more prevalent when the complementizer that is absent from the input: [+it -that] >> [-it -that] (e.g. it could be ∅ >> ∅ could be ∅). iii. The input will be shadowed most faithfully when it conforms to either a full or an adverbial-like form: [+it +that] >> [+it +that]; [‑it ‑that] >> [‑it ‑that] (e.g. it could be that >> it could be that; ∅ could be ∅ >> ∅ could be ∅).

Chapter 5. Could be, might be, maybe

3.1 Design and participants The input material consisted of twenty-four short recorded dialogues (Examples 12–13), of which eleven included a target item (line B2) and thirteen served as distractors. Among the eleven target items, each participant would hear each variant once: four forms each of could be and might be ([+/-it +/-that]), and three of may be ([+it +that], [+it -that], [-it +that]). Every dialogue with a target item was created with two different variants; these were cross-spliced, so that the recordings were identical except for the form of the target item. Participants were assigned to one of two groups; each group would hear a given dialogue with one or the other form, e.g. Example (12) as given with [‑it + that] or with [‑it ‑that], and (13) with [+it -that] or [+it +that]. (12) A1: They’ve got a dog named Rupert. B1: Oh yes, the big white one. A2: He won’t let me pet him though. B2: Could be that he’s afraid of you. A3: Do I look scary to you? B3: It’s not the looks, it’s the smell. (13) A1: Jim’s band are playing another gig. B1: Do you know the date? A2: No, I saw it posted a while ago. B2: It might be we’ve missed it already. A3: That would be a shame, though. B3: We’ll have to find another excuse to go out.

The experiment was carried out online through Labvanced (Finger et al. 2017). The dialogues were presented in random order after a microphone test and a practice run. They were given auditorily only, with no written cues. Participants would hear the first line of a dialogue, repeat it, then hear the second line, and so on. They were instructed to “start repeating as soon as you can – even if the speaker hasn’t finished” and “speak clearly and in your own voice – don’t mimic the speaker’s accent or intonation”. Their responses were recorded.2

2. Time stamps for playback and recording were also registered, so that response latencies could be calculated. However, these turned out to be unreliable, probably due to slight distortions in the online recording process. When entire sentences are shadowed, such latencies are not necessarily a direct indicator of processing ease or difficulty with the target item; still, other studies have applied them in this way (Marslen-Wilson & Welsh 1978; Caines 2012).

133

134

David Lorenz

Sixty-one UK-based native speakers of English participated, who were recruited through the platform Prolific. Due to technical errors and the fragility of online recording and transmission, five data sets were not useable at all and a few others were incomplete. Still, the final yield was 565 analyzable tokens of target sentences (205 could, 207 might, 153 may).

3.2 Results While the main goal was to test for fluent restoration or omission on the word/ phrase level, the recordings provide a rich data base which can be explored on many levels. Therefore, the following is a mix of deductive hypothesis testing and inductive exploration. I used R (R Core Team 2019) for data analysis, Praat (Boersma & Weenink 2017) for acoustic measurements, and the R package rPraat (Bořil & Skarnitzl 2016) to combine them. 3.2.2 Comparing shadowed responses with input forms Out of the 565 relevant tokens, the target item was shadowed identically in 449 cases, leaving 116 (21%) non-identical responses. Assuming that entrenched and expectable forms will be repeated faithfully, non-identical responses are a first indicator that the input form was at odds with expectations. The distribution of non-identical responses across input forms is shown in Table 2, and tested with a simple logistic regression model identical ~ input_form (Figure 2). The reference level is it could be that, represented in the intercept, the coefficients indicate the probability of a non-identical response to a given input form relative to that reference level. The items that stand out for their high rates of non-identical responses are indeed those that do not match the proposed form-function correlation – mostly with the pronoun it retained in a ‘critical context’ with a bare clause complement ([+it -that]), and ∅ may be that ([-it +that]). The full phrase it may be that is the most faithfully shadowed item. These results support Hypothesis (iii) above. On the other hand, there are figures that run counter to it: the high rate of nonidentical responses with ∅ might be ∅, and especially the low rate with ∅ could be that (which was hypothesized to cause difficulties). Homing in on the non-identical responses, what forms are produced? Does the presence or absence of the complementizer that influence tendencies toward fluent restoration or omission, as hypothesized? Most non-identical responses do contain a fluent omission or restoration of it or that. In a few cases (n = 3) a different construction was returned – for example, one participant shadowed the line Could be that he’s afraid of you as Is it coz he’s afraid of you. These instances are not considered in the following analysis. Figure 3 shows for each relevant pair of

Chapter 5. Could be, might be, maybe

Table 2. Identical and non-identical shadowed responses by input form response form input form

identical

non-identical

it could be that

46

5 (10%)

∅ could be that

52

1 ( 2%)

it could be ∅

35

17 (33%)

∅ could be ∅

42

7 (14%)

it might be that

47

3 ( 6%)

∅ might be that

49

4 ( 8%)

it might be ∅

34

18 (35%)

∅ might be ∅

39

13 (25%)

it may be that

50

1 ( 2%)

∅ may be that

27

23 (46%)

it may be ∅

28

24 (46%)

Figure 2. Coefficients and confidence intervals (1 and 2 S.E.) of a logistic regression model identical ~ input_form (c = 0.78); reference levels identical = true, input_form = ‘it could be that’

135

136

David Lorenz

input items ([+/-that]) the rates of it-omission (left panel) and it-restoration (right panel) relative to identical responses. Another frequent observation is the fluent restoration of that, as shown in Figure 4.

Figure 3. Fluent it-omission and it-restoration by input form (p-values from Fisher’s exact test)

Figure 4. Fluent that-restoration and that-omission by input form (p-values from Fisher’s exact test)

Shadowers produce more fluent it-omission when the complementizer is zero. This effect is very clear in it may be ∅ >> ∅ may be ∅, a statistical trend with might be, and below significance with could be. Restoration of a missing pronoun is very rare and does not show any reliable trend. These results provide tentative support for Hypothesis (ii) above, but none for Hypothesis (i). In addition, that-restoration is surprisingly frequent (Figure 4, right panel); it seems to be weakly associated with the presence of pronoun it, but again this does not reach statistical significance and is most frequent with it may be ∅ >> it may be that. The input form ∅ may be that is the only one that is often returned with fluent omission of that. Overall, (it) may be (that) shows a clear trend to be shadowed either

Chapter 5. Could be, might be, maybe

as maybe or the full phrase it may be that. This is as expected given that maybe is a fully conventional adverb whose decategorialization is completed. With could be and might be, however, hearers/shadowers are much more flexible. The strongest trend is that of that-restoration, which reinforces the item’s syntactic status as a matrix clause rather than an adverbial. 3.2.3 Durations of shadowed items Rather than categorical presence or absence, we can look at the durations of items in the responses: are there patterns of gradient reduction?3 Under hypothesis (iii) above, we would especially expect shorter it when that is absent. The overall mean duration of it is 118msec (median = 112, s.d. = 42.4). Figure 5 shows the pronoun durations in the respective responses.4

Figure 5. Durations of pronoun it in different response items (y-axes are log scaled; boxes are median and 1st to 3rd quartile; whiskers extend to 1.5 * inter-quartile range; W and p-values from Wilcoxon rank sum tests)

A slight trend toward it-reduction with zero complementizer seems generally observable, but it is reliable only with could be. This is the reverse result of categorical it-omission (see Figure 3 above), where may be showed a very strong effect

3. Reported here are only the durations of pronoun it (where present) and the item could/might/may be. Durations of complementizer that were also considered but found to show no variation beyond chance. 4. These durations are to be viewed with some caution, as, firstly, they are not normalized against the speaker’s baseline speech rate; secondly, the picture would surely change if cases of fluent omission were included as having a duration of zero.

137

138

David Lorenz

and could be the weakest. A possible reason is that omission and duration are connected as categorical and gradient reduction: where there is a clear effect on omission, this offsets the effect on duration. With may be the distinction is strongly categorical. With could be, and to a small extent with might be, there is evidence also of a gradient reduction of the pronoun. All in all, the effects are tenuous (except for may be), but consistently in the direction predicted by Hypothesis (iii). Another interesting difference is seen in the durations of the X be items themselves. We might expect these to vary with the item’s potential adverbial status, that is, with the presence or absence of complementizer that, but this is not found. Instead, the durations vary with the presence or absence of pronoun it, as Figure 6 shows. They are generally shorter in variants without it (the two leftmost boxes in each graph), most clearly in the could be set (left panel), but also with might be (middle panel) and may be (right panel). This is surprising if we assume that it could be and ∅ could be carry the same epistemic information. However, it shows a correlation between it-omission and gradient reduction which suggests that shadowers tacitly interpret forms without the pronoun as less prominent in the utterance.

Figure 6. Durations of could/might/may be in different response items (y-axes are log scaled; boxes are median and 1st to 3rd quartile; whiskers extend to 1.5 * inter-quartile range; χ2 and p-values from Kruskal-Wallis rank sum tests)

Chapter 5. Could be, might be, maybe

4.

Discussion

With the experimental study presented above, and the corpus study that preceded it, I have tried to address issues of grammaticalization from a synchronic-cognitive perspective. The guiding idea is that if grammaticalization is a unified apparatus of interconnected mechanisms of change, and if these mechanisms are grounded in cognitive and communicative principles, then it is worthwhile to probe for these connections even without recourse to the diachronic dimension, in the use and perception of ‘candidate constructions’ which have an unfulfilled potential for grammaticalization or seem stuck at an incipient stage. This is one way of relating the community-wide language system (and its on-going history) to the language system as represented in the minds of contemporary individuals. We usually observe or interpret grammaticalization as a change in community grammar, and this is certainly adequate in a long-term perspective. But in its emergence through usage events, it is all down to the individual speakers and hearers – their speech behavior, their mental categories, and gradual changes therein (cf. Fischer 2010; De Smet 2016; Petré & Van de Velde 2018). We may therefore assume that the different speakers and writers in the corpora, and the sixty-one participants in the experiment, all differ, if slightly, in their mental representations of epistemic phrases and adverbials. What we see is that the usage of (it) could be and (it) might be as epistemic items is not bound by any strong conventions. Rather, speakers are flexible in their realization, omitting or retaining the pronoun it, or, in the shadowing task, reinserting the complementizer that. The fluent restoration of the complementizer is even, in a sense, conservative in a way that seems to run counter to grammaticalization. The difference becomes very clear in comparison with the forms with may, which show a strong pull towards either the adverb maybe or the phrase it may be that. The flexibility of could be and might be indicates that for most speakers they are not firmly entrenched as items with an adverbial function. On the other hand, it also shows their availability to be put to use in such functions. Taken together, the above findings on epistemic (it) could be (that) and (it) might be (that) show that things are very much in flux. A rather clear result of the corpus study is that it-omission correlates with bare clause complementation as a ‘critical context’ for grammaticalization. The continuous shadowing experiment provides some support for this, but it also exposes the contrast to the adverb maybe, which is quite clearly separated from the phrase it may be + that-clause. Beyond this main finding, there are several striking parallels between the corpus figures and fluent restoration or omission in the shadowing task. Firstly, the rates of it-omission are generally higher with could be than with might be in both

139

140

David Lorenz

studies. As could be is more frequent in critical contexts in the corpus data, this may affect its mental representation as an epistemic item and translate into an increased tendency to use it without the pronoun in other contexts as well.5 The difference in fluent it-omission in shadowing then results from differential expectations of the usage of could be and might be, which in turn affect their perception. Secondly, in the Spoken BNC corpora, that-clause complements are about twice as likely as bare clause complements overall, and still equally frequent when the pronoun it is absent. The shadowers in the experiment are on a par with these corpus data in their general preference for that-restoration, their acceptance of ∅ could/might be that, and their reluctance to restore the pronoun it. Again, this can be explained with the assumption that quantitative patterns in usage are reflected in the mental representation of a construction, and again this means that speakers/hearers orient themselves to experienced tendencies and do not categorically distinguish between matrix and adverbial-like could/might be. In addition, ∅ could/might be that is very similar to patterns of ‘semi-insubordination’, where a single matrix element takes a that-clause complement, e.g. strange/funny/odd that… (Kaltenböck 2021) or no wonder that… (Gentens et al. 2016). While ‘semiinsubordination’, as Kaltenböck (2021) argues, has specific functional properties (such as subjective stance and deictic anchoring) which do not apply to epistemic could/might be, it does provide a structural template to which the form ∅ could/ might be that can be matched.6 The durations of pronoun it (where present) and could/might/may be in the responses indicate that gradient phonetic reduction is at play, too. At least with it could be + bare clause, reduction of it is plausibly linked to the critical context for grammaticalization. On the other hand, reduction of could/might/may be seems to be independent of the complementation pattern and correlates instead with the

5. There is also a possible link to the general frequencies of could and might as modal auxiliaries: could is more frequent, with might on the decline (see Mair & Leech 2006: 326–327). However, might is more typically used with epistemic meaning (Collins 2007; Usonienė & Šoleniė 2012). 6. Interestingly, Kaltenböck (2021) reports that the elliptical forms ∅ strange/funny/odd favor ‘zero that-clause’ complements (e.g. Funny you should ask) compared to complete matrix clauses, which favor that-clause complements (e.g. It’s funny that you should ask). Thus, semiinsubordination shows the same correlation of subject ellipsis and that-omission as reported here for epistemic phrases. Gentens et al. (2016) propose that elliptical forms of the type ∅ no wonder that paved the way for the disjunct adverbial no wonder, starting in Late Middle English. This would mean that in this case subject omission promoted that-omission, i.e. the reverse order of what is hypothesized in the present study. This need not be a contradiction, as the cases, contexts and data are different enough. Still, Gentens et al. (2016: 137) also point out the association of shorter form and adverbial status.

Chapter 5. Could be, might be, maybe

absence of pronoun it. This can nonetheless be explained in terms of an immediate connection between form and status: When the phrase is bereft of its subject pronoun, it is more easily perceived as having low discourse prominence and hence realized more economically (cf. Boye & Harder 2012: 30).

5.

Conclusion

Overall, the variation in usage and perception of (it) could be and (it) might be is rather free, though not entirely chaotic. There are tendencies of adapting form to syntactic function, in categorical and gradient realizations. While we do not have comparable (spoken) data on the history of maybe (or the cognate epistemic adverbs in other languages, see Section 2), there is some evidence that it gradually conventionalized from similarly variable usage patterns (cf. López-Couso & Méndez-Naya 2016). Present-day maybe clearly is differentiated from the phrase it may be (that) in a way that could be and might be are not. A generalized proposal that can be made, following from the findings presented above, is as follows: Firstly, some items or constructions in a language can have a base potential for grammaticalization in that their meaning and structure allows them to be recruited to express more ‘procedural’ meanings (Traugott 2002) or grammatical functions, and be discursively backgrounded (Boye & Harder 2012). These items or constructions can easily occur in ‘critical contexts’ where a more grammaticalized interpretation is possible (cf. Diewald 2002). Such occurrences do not imply a speaker’s intention or a hearer’s reaction to ‘re-analyze’ the construction in the direction of grammaticalization; they are due to the flexibility and creativity of every-day language use (see also Croft 2010; Petré 2016). However, speakers and hearers do have an intuition for what is ‘more grammatical’, perhaps most plausibly in the sense of ‘discursively secondary’ (see Christensen et al. 2021). Secondly, then, the usage of a construction in a configuration that makes it appear ‘more grammatical’ (i.e., in a critical context) will have an immediate effect on its form and perception. It allows for more reduction, possibly because speakers tacitly recognize that the full form and structure is not needed to fulfil its function in the utterance (see also Barth 2019), or because of a formal analogy to an already grammaticalized item (as maybe in the case at hand; see also De Smet 2018: 87). Perception might be more conservative if hearers subconsciously reconstruct the full structure, but hearers might also expect (or readily accept) reduction when the context favors it (the experiment in Section 3 provides tentative evidence in both directions). This means, thirdly, that mental representations will be affected, not only in the increasing entrenchment of a reduced form when an item is grammaticalizing (which would correspond to

141

142

David Lorenz

the process of erosion), but even in the integration of low-frequency tendencies in potential or incipient grammaticalization. Finally, the findings suggest that erosion is not a side- or after-effect of grammaticalization but rather an intrinsic part of the machinery. To the extent that erosion is immediately associated with grammaticalizing usage contexts, grammaticalization also appears as a ‘pull factor’ in that items are more likely to get reduced just because they are (incidentally or deliberately) put on a grammaticalization pathway. This is only possible if language users’ mental representations include a notion of grammatical functions and of the morpho-phonological forms of the constructions that express them.

References Barth, Danielle. 2019. Effects of average and specific context probability on reduction of function words BE and HAVE. Linguistics Vanguard 5(1): article nr. 20180055. Beijering, Karin. 2010. The grammaticalization of Mainland Scandinavian MAYBE. Bergen Language and Linguistics Studies 1: 1–21. Beijering, Karin. 2016. Semi-insubordinate at-constructions in Norwegian: Formal, semantic and functional properties. Norsk Lingvistisk Tidsskrift 34: 161–182. BNC Consortium. 2007. The British National Corpus. Distributed by Bodleian Libraries, University of Oxford, on behalf of the BNC Consortium. Boersma, Paul & Weenink, David. 2017. Praat: Doing Phonetics by Computer [computer program]. Version 6.0.36. Amsterdam: University of Amsterdam. Bořil, Tomáš & Skarnitzl, Radek. 2016. Tools rPraat and mPraat. In Text, Speech, and Dialogue, Petr Sojka, Aleš Horák, Ivan Kopeček & Karel Pala (eds), 367–374. Cham: Springer. Boye, Kasper & Harder, Peter. 2012. A usage-based theory of grammatical status and grammaticalization. Language 88(1): 1–44. Breban, Tine. 2014. What is secondary grammaticalization? Trying to see the wood for the trees in a confusion of interpretations. Folia Linguistica 48(2): 469–502. Brinton, Laurel J. & Traugott, Elizabeth Closs. 2005. Lexicalization and Language Change. Cambridge: CUP. Bybee, Joan L. 2003. Cognitive processes in grammaticalization. In The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, Vol. 2, Michael Tomasello (ed.) 145–167. Mahwah NJ: Lawrence Erlbaum Associates. Bybee, Joan L., Perkins, Revere D. & Pagliuca, William. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago IL: The University of Chicago Press. Caines, Andrew. 2012. You talking to me? Corpus and experimental data on the zero auxiliary interrogative in British English. In Frequency Effects in Language Learning and Processing, Stefan T. Gries & Dagmar Divjak (eds), 177–205. Berlin: De Gruyter.

Chapter 5. Could be, might be, maybe

Christensen, Marie Herget, Kristensen, Line Burholt, Vinther, Nicoline Munck & Boye, Kasper. 2021. Grammar is background in sentence processing. Language and Cognition 13(1): 128–153. Collins, Peter. 2007. Can/could and may/might in British, American and Australian English: A corpus-based account. World Englishes 26(4): 474–491. Croft, William. 2010. The origins of grammaticalization in the verbalization of experience. Linguistics 48(1): 1–48. Davies, Mark. 2013. Corpus of Global Web-based English (GloWbE). (15 January 2021). De Smet, Hendrik. 2016. How gradual change progresses: The interaction between convention and innovation. Language Variation and Change 28: 83–102. De Smet, Hendrik. 2018. Entrenchment effects in language change. In Entrenchment and the Psychology of Language Learning: How We Reorganize and Adapt Linguistic Knowledge, Hans-Jörg Schmid (ed.) 75–100. Berlin: De Gruyter Mouton. Detges, Ulrich & Waltereit, Richard. 2002. Grammaticalization vs. reanalysis: A semanticpragmatic account of functional change in grammar. Zeitschrift für Sprachwissenschaft 21(2): 151–195. Diewald, Gabriele. 2002. A model for relevant types of contexts in grammaticalization. In New Reflections on Grammaticalization [Typological Studies in Language 49], Ilse Wischer & Gabriele Diewald (eds), 103–120. Amsterdam: John Benjamins. Diewald, Gabriele. 2006. Context types in grammaticalization as constructions. In Constructions, Special Vol. 1, Doris Schönefeld (ed.). Finger, Holger, Goeke, Caspar, Diekamp, Dorena, Standvoß, Kai & König, Peter. 2017. LabVanced: A unified JavaScript framework for online studies. International Conference on Computational Social Science (Cologne). Fischer, Olga. 2007. Morphosyntactic Change. Functional and Formal Perspectives. Oxford: OUP. Fischer, Olga. 2010. On problem areas in grammaticalization: Lehmann’s parameters and the issue of scope. In Formal Evidence in Grammaticalization Research [Typological Studies in Language 94], An Van linden, Jean-Christophe Verstraete & Kristin Davidse (eds), 17–42. Amsterdam: John Benjamins. Gaeta, Livio. 2004. Exploring grammaticalization from below. In What makes Grammaticalization? A Look from its Fringes and its Components, Walter Bisang, Nikolaus P. Himmelmann & Björn Wiemer (eds), 45–75. Berlin: Mouton de Gruyter. Gentens, Caroline, Kimps, Ditte, Davidse, Kristin, Jacobs, Gilles, Van linden, An & Brems, Lot. 2016. Mirativity and rhetorical structure: the development and prosody of disjunct and anaphoric adverbials with ‘no’ wonder. In Outside the Clause. Form and Function of Extra-clausal Constituents [Studies in Language Companion Series 178], Gunther Kaltenböck, Evelien Keizer & Arne Lohmann (eds), 125–156. Amsterdam: John Benjamins. Gilquin, Gaëtanelle & Gries, Stefan T. 2009. Corpora and experimental methods: A state-ofthe-art review. Corpus Linguistics and Linguistic Theory 5(1): 1–26. Haspelmath, Martin. 1999. Why is grammaticalization irreversible? Linguistics 37(6): 1043–1068.

143

144

David Lorenz

Haspelmath, Martin. 2008. Creating economical morphosyntactic patterns in language change. In Linguistic Universals and Language Change, Jeff Good (ed.) 185–214. Oxford: OUP. Heine, Bernd. 1993. Auxiliaries: Cognitive Forces and Grammaticalization. Oxford: OUP. Heine, Bernd. 2002. On the role of context in grammaticalization. In New Reflections on Grammaticalization [Typological Studies in Language 49], Ilse Wischer & Gabriele Diewald (eds), 83–101. Amsterdam: John Benjamins. Heine, Bernd. 2003. Grammaticalization. In The Handbook of Historical Linguistics, Brian D. Joseph & Richard D. Janda (eds), 575–601. Malden MA: Blackwell. Hilpert, Martin & Correia Saavedra, David. 2018. The unidirectionality of semantic changes in grammaticalization: An experimental approach to the asymmetric priming hypothesis. English Language and Linguistics 22(3): 357–380. Hoefler, Stefan H. & Smith, Andrew D. M. 2009. The pre-linguistic basis of grammaticalization: A unified approach to metaphor and reanalysis. Studies in Language 33(4): 886–909. Hopper, Paul. 1991. On some principles of grammaticization. In Approaches to Grammaticalization, Vol. 1: Theoretical and Methodological Issues [Typological Studies in Language 19:1], Elizabeth Closs Traugott & Bernd Heine (eds), 17–36. Amsterdam: John Benjamins. Hopper, Paul & Traugott, Elizabeth Closs. 2003. Grammaticalization, 2nd edn. Cambridge: CUP. Jäger, Gerhard & Rosenbach, Annette. 2008. Priming and unidirectional language change. Theoretical Linguistics 34(2): 85–113. Kaatari, Henrik & Larsson, Tove. 2019. Using the BNC and the Spoken BNC2014 to study the syntactic development of I think and I’m sure. English Studies 100(6): 710–727. Kaltenböck, Gunther. 2021. Funny you should say that: On the use of semi-insubordination in English. Constructions and Frames 13(1): 126–159. Konvička, Martin. 2019. Paradigms, host classes, and ancillariness. In Grammar – Discourse – Context: Grammar and Usage in Language Variation and Change, Kristin Bech & Ruth Möhlig-Falke (eds), 277–304. Berlin: De Gruyter. Lehmann, Christian. 2004. Theory and method in grammaticalization. Zeitschrift für Germanistische Linguistik 32(2): 152–187. Lehmann, Christian. 2015[1982]. Thoughts on Grammaticalization, 3rd edn. Berlin: Language Science Press. Lehmann, Christian. 2017. Grammaticalization and automation. JournaLIPP 5: 33–48. López-Couso, María José & Belén Méndez-Naya. 2014. From clause to pragmatic marker: A study of the development of like-parentheticals in American English. Journal of Historical Pragmatics 15(1): 66–91. López-Couso, María José & Méndez-Naya, Belén. 2016. From clause to adverb: On the history of maybe. In Outside the Clause. Form and Function of Extra-clausal Constituents [Studies in Language Companion Series 178], Gunther Kaltenböck, Evelien Keizer & Arne Lohmann (eds.) 157–176. Amsterdam: John Benjamins.

Chapter 5. Could be, might be, maybe

Lorenz, David. 2012. The perception of gonna and gotta – A study of emancipation in progress. In Proceedings of the 5th ISEL conference ExLing 2012, 27–29 August 2012, Athens, Greece, Antonis Botinis (ed.) 77–80. Athens: ISEL Editions. Lorenz, David. Forthcoming. Could be it’s grammaticalization: Usage patterns of the epistemic phrases (it) could/might be. To appear in Journal of English Linguistics. Lorenz, David & Tizón-Couto, David. 2020. Not just frequency, not just modality: Production and perception of English semi-modals. In Re-Assessing Modalising Expressions. Categories, Co-text, and Context [Studies in Language Companion Series 216], Pascal Hohaus & Rainer Schulze (eds), 79–107. Amsterdam: John Benjamins. Love, Robbie, Dembry, Claire, Hardie, Andrew, Brezina, Vaclav & McEnery, Tony. 2017. The Spoken BNC2014: designing and building a spoken corpus of everyday conversations. International Journal of Corpus Linguistics 22(3): 319–344. Mair, Christian & Leech, Geoffrey. 2006. Current changes in English syntax. In The Handbook of English Linguistics, Bas Aarts & April McMahon (eds), 318–342. Oxford: Blackwell. Marslen-Wilson, William D. & Welsh, Alan. 1978. Processing interactions and lexical access during word recognition in continuous speech. Cognitive Psychology 10(1): 29–63. Mélac, Eric. 2022. The grammaticalization of evidentiality in English. English Language & Linguistics 26(2): 331–359. Narrog, Heiko & Heine, Bernd. 2021. Grammaticalization. Oxford: OUP. Norde, Muriel, Rawoens, Gudrun & Beijering, Karin. 2014. Från matrissats till satsadverb? En diakron studie av adverbet kanske. (28 December 2022). Petré, Peter. 2016. Unidirectionality as a cycle of convention and innovation. Micro-changes in the grammaticalization of [BE going to INF]. Belgian Journal of Linguistics 30: 115–146. Petré, Peter & Van de Velde, Freek. 2018. The real-time dynamics of the individual and the community in grammaticalization. Language 94(4): 867–901. R Core Team. 2019. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. |(28 December 2022). Ramat, Paolo & Ricca, Davide. 1998. Sentence adverbs in the languages of Europe. In Adverbial Constructions in the Languages of Europe, Johan van der Auwera & Dónall P. O. Baoill (eds), 187–273. Berlin: Mouton de Gruyter. Thompson, Sandra A. & Mulac, Anthony. 1991. A quantitative perspective on the grammaticization of epistemic parentheticals in English. In Approaches to Grammaticalization, Vol. 2: Types of Grammatical Markers [Typological Studies in Language 19:2], Elizabeth Closs Traugott & Bernd Heine (eds), 313–330. Amsterdam: John Benjamins. Traugott, Elizabeth Closs. 2002. From etymology to historical pragmatics. In Studies in the History of the English Language: A Millennial Perspective, Donka Minkova & Robert Stockwell (eds), 19–49. Berlin: De Gruyter Mouton. Traugott, Elizabeth Closs. 2003. Constructions in grammaticalization. In The Handbook of Historical Linguistics, Brian D. Joseph & Richard D. Janda (ed.) 624–647. Oxford: Blackwell.

145

146

David Lorenz

Usonienė, Aurelia & Šolienė, Audrone. 2012. Choice of strategies in realizations of epistemic possibility in English and Lithuanian: A corpus-based study. In Corpus Studies in Contrastive Linguistics [Benjamins Current Topics 43], Stefania Marzo, Kris Heylen & Gert De Sutter (eds.) 141–166. Amsterdam: John Benjamins. Van Bogaert, Julie. 2011. I think and other complement-taking mental predicates: A case of and for constructional grammaticalization. Linguistics 49(2). 295–332. Wiemer, Björn. 2014. Quo vadis grammaticalization theory?, or: Why complex language change is like words. Folia Linguistica 48(2): 425–467.

chapter 6

The final-appendage construction in Japanese and Korean To what extent is post-predicative position exploited in the two East Asian languages? Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim Fuji Women’s University | Hokkaido University of Education | Georgia Institute of Technology

Japanese and Korean are both predicate-final (OV ) languages with relatively flexible constituent order. However, our analysis of parallel texts (Japanese novels and their Korean translations) demonstrates that the two languages differ in the exploitability of post-predicative position. Korean has a much lower frequency of final-appendage structures with clauseinternal elements (especially adnominals and subject nominals) less likely to occur in post-predicative position. On the other hand, Korean shows a relatively higher proportion of vocative phrases in post-predicative position despite a lower frequency of final-appendage structures. The results reveal that Japanese exhibits a relatively higher degree of constructional entrenchment of final-appendage structures. In contrast, post-predicative position in Korean is rather restrictive, more limited to clause-external elements than to clause-internal ones. Keywords: final-appendage construction, postposing, post-predicative position, Japanese, Korean

1.

Introduction

Japanese and Korean are generally regarded as typologically similar languages. They are agglutinative languages, in which grammatical relations are indicated by case and other postpositional particles.1 They both have SOV (predicate-final) 1. In this study, the term “grammatical relation” is used in a broader sense. It refers to “relationships within grammar, or at least within syntax and morphology” (Woolford 1999: 355). Topic phrases receive morphological marking in Korean and Japanese, and so do vocative phrases in https://doi.org/10.1075/slcs.232.06izu © 2023 John Benjamins Publishing Company

148

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

constituent orders, but the orders are relatively flexible. Because of the flexibility of constituent order, the two languages often allow some elements to be placed after predicates (i.e., in post-predicative position), especially in spoken discourse. Compare the Japanese (J) and Korean (K) sentences in (1) with those arranged in a different constituent order in the respective languages in (2).2 (1) J: […] sore-wa tyotto sagi-da-yo. that-top a.little fraud-cop-fp K: […] geugeo-n jom sagi-nde. that-top a.little fraud-(cop)fp    ‘[…] that’s a kind of fraud.’

(T. Kim 2011: 223)

(2) J: […] tyotto sagi-da-yo, sore-wa. a.little fraud-cop-fp that-top K: […] jom sagi-nde, geugeo-n. a.little fraud-(cop)fp that-top    ‘[…] (it)’s a kind of fraud, that is.’

(T. Kim 2011: 224)

In the canonical constituent orders of the two languages, subject nominals (sorewa or geugeo-n) are placed before the predicates (sagi-da-yo or sagi-nde) as in (1). However, these nominals can occupy post-predicative position as in (2). In this study, we refer to the structure where post-predicative position is occupied by some linguistic elements as “the final-appendage construction.” The schematic structure of the construction is formulated as [… predicate (,) X], where X represents a postposed element. Although such final-appendage construction is observed in the two languages, their realizations are not always identical. For example, compare an utterance from a Japanese novel with its Korean translation in (3). (3) J: Tanzyoobi purezento-daro, neetyan-no. birthday present-will.be big.sister.dim-gen    ‘(It)’s a birthday present, isn’t it, (my) big sister’s?’

(Tenki-no Ko p. 147)

Korean. Thus, topic and vocative particles are also treated as indicating grammatical relations here. 2. The romanization of Korean examples in our paper is largely based on Mungyobu hangeul romaja jeoggibeob 1959 nyeon pan ‘Ministry of Education Transliteration of Hangeul 1959.’ Examples cited from previous studies were modified in romanization and glossing for consistency. In some cases, the morphological analyses were also changed accordingly.

Chapter 6. The final-appendage construction in Japanese and Korean

K: Nuna saengil seonmul-i-ji? big.sister birthday present-cop-fp    ‘(It’s my) big sister’s birthday present, right?’ K’: ??Saengil seonmul-i-ji nuna-eui? birthday present-cop-fp big.sister-gen    ‘(It)’s a birthday present, isn’t it, my big sister’s?’

(Nalssi eui Ai p. 171)

(cf. Izutsu et al. 2021)

Japanese can place a nominal with a genitive case particle (neetyan-no ‘my big sister’s’) in post-predicative position as in (3J). In Korean, however, a genitive nominal is less acceptable in that position as in (3K’) than in the canonical prenominal position as in (3K).3 Previous studies have attempted to identify possible motivations for the postposing construction in Japanese and Korean. Although terminologies or their notional definitions may differ among researchers (e.g., “focus,” “importance,” “urgency/relevance”), most studies agree that postposing is motivated by a functional principle known as the “important information first (IIF)” principle (Simon 1989: 189). In languages with SOV constituent order, a speaker tends to encode “important or urgent information coming to the speaker’s mind first and thus being verbalized first” (Simon 1989: 189); hence less important information is produced later in the utterance. For example, when a speaker gives a warning to watch out for a car coming, (4a), which contains a postposed element, is more natural in a situation of imminent danger than (4b). The latter is more likely to be uttered, for example, by a mother giving a lesson about traffic safety to her kid. (4) a.

Abunai-yo, kuruma-ga kuru-kara. look.out-fp car-nom come-because ‘Look out, because a car is coming.’ b. Kuruma-ga kuru-kara abunai-yo. car-nom come-because look.out-fp ‘Look out, because a car is coming.’

(Simon 1989: 189)4

Similar functional motivations have been proposed in Korean (e.g., “the ‘urgency/ relevance of information’ factor” in T. Kim 2011) as well as in Japanese (Maynard 3. In Korean, the genitive case particle ‑eui is often unexpressed in spoken discourse as in (3K). If the adnominal phrase in (3K) comes after the predicate without the particle as in (i), the postposed element (nuna) sounds like a vocative expression rather than an adnominal phrase. (i) # Saengil seonmul-i-ji nuna? birthday present-cop-fp big.sister ‘(It)’s a birthday present, isn’t it, big sis?’ 4. The underline is in the original, indicating the position where the postposed expression is supposed to be in the canonical constituent order.

149

150

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

1989; Simon 1989; Takami 1995; Shimojo 2005). Taking into consideration the results of some previous studies (A. Kim 1988; Takami 1994), Fujii (1995: 195) finds it convincing that the “important information first” principle is a common motivation underlying the postposing constructions of SOV languages. If this functional motivation operates in both Japanese and Korean, why does it happen that postposing is not maintained in the translation of one language into the other, as in (3)? In general, translation preserves approximate equivalences in meaning, and the discourse context of a sentence in an original text is not normally changed in the translation. This would lead to the assumption that the functional motivation would likewise apply to both original sentences and their translations. An example like (3), however, suggests that it does not necessarily seem to be the case. The present study compares the utterances of the final-appendage construction in Japanese novels and their Korean translations, investigating how the two languages differ in the realization of the final-appendage construction. Our analysis of the parallel texts will demonstrate that the two languages differ in the exploitability of post-predicative position. Japanese exhibits a relatively higher degree of constructional entrenchment (Izutsu & Izutsu 2018) of final-appendage structures, considerably outnumbering Korean in their occurrences, especially in the occurrences of the structures where syntactically less flexible elements (clauseinternal arguments and adnominal phrases) occupy post-predicative position. The higher degree of entrenchment of the final-appendage construction is also evinced by the fact that Japanese shows a higher likelihood of pronouns appearing in post-predicative position without any indication of their grammatical functions within sentences.

2.

Data

We collected the utterances of the final-appendage construction in five Japanese novels and their corresponding Korean translations, as summarized in Table 1. The last two texts are the novelized editions of a TV drama and a movie, respectively. The square brackets following the titles of the Japanese novels and their Korean translations represent the abbreviations of the titles used in the subsequent discussion.

Chapter 6. The final-appendage construction in Japanese and Korean

Table 1. The Japanese novels and their Korean translations used in this study5 Japanese data (Original novel)

Author

Korean data (Translation)

Translator

Hyooten 1 [H] (1982 [1965])

Ayako Miura

Bingjeom [B] (2017 [1992])

Ho Choe

Shino Sakuragi

Gubichi neun Dal [G] (2015)

Yunog Yang

Kanako Nishi

Weontag [W] (2014)

Heuiseon Im

Yuji Sakamoto & Shinobu Momose

Choego eui Ihon 1 [C] (2018)

Jina Chu

Makoto Shinkai

Nalssi eui Ai [N] (2019)

Gyeongog Min

‘Freezing Point 1’ Dakoosuru Tsuki [D] (2016 [2013]) ‘Meadering Moon’ Entaku [E] (2013 [2011]) ‘The Round Table’ Saikoo-no Rikon 1 [S] (2013) ‘The Best Divorce 1’ Tenki-no Ko [T] (2019) ‘Weathering with You’

3.

Results

3.1 Japanese final-appendage structures and their Korean translations Final-appendage structures are classified into four types: simple postposing, multiple postposing, postposing with a preceding coreferential pronoun, and repetition. In this study, “postposing” does not presuppose any movement of a constituent from its original position in a sentence. The term “postposing” simply means the positioning of a linguistic element after a predicate. Although most previous studies confine their discussions to the postposing of clausal constituents (e.g., Simon 1989; Birner & Ward 1998; Shimojo 2005), our view of “postposing” includes the postposition of extra-clausal elements (e.g., vocative phrases) to the right of a predicate. The four types of final-appendage construction are exemplified in (5): 5. The years in the square brackets represent the publication years of the original novels, and the others indicate those of the paperback editions. The page numbers referred to in this paper are all taken from the latter publications.

151

152

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

(5) a.

Simple postposing Kuru-wayo, syuraba. come-fp battle.scene ‘Here comes a battle scene.’ b. Multiple postposing Nigedasi-tyat-ta-n-desu-yone, kare, keisatu-kara. escape-perf-pst-fn-cop-fp he police-from ‘He escaped from the police.’ c. Postposing with a preceding coreferential pronoun Sorei yaru-yo, taisyokukini. that give-fp severance.pay ‘(I)’ll give it (to you), (your) severance pay.’ d. Repetition Ee, onnanoko-ga hosii-wa, tiisana onnanoko-ga. yes girl-nom want-fp little girl-nom ‘Yes, (I) want a girl, a little girl.’

(S, p. 265)

(T, p. 233)

(T, p. 176)

(H, p.80)

“Simple postposing” is a structure in which a single linguistic element is placed after a predicate. In (5a), syuraba ‘battle scene,’ the grammatical subject of the sentence, follows the predicate kuru-wayo ‘come.’ Simon (1989: 6) states that “it is not unusual for two or three elements to appear here” (i.e., in post-predicative position). The construction in which post-predicative position is occupied by more than one element is called “multiple postposing” (Mihara 2015: 1). In (5b), two elements (the subject nominal kare ‘he’ and the locative adverbial keisatu-kara ‘from the police’) occur after the predicate with the final particle ‑yone. A postposed element is occasionally preceded by a pronoun coreferential with it, which is a phenomenon often referred to as right-dislocation. (5c) has a cataphoric pronoun sore ‘that,’ which is coreferential with the noun phrase taisyokukin ‘severance pay’ in post-predicative position. Finally, “repetition” is also a type of postposing in which a phrase (nearly) identical with the one preceding a predicate appears again in post-predicative position. In (5d), the speaker mentions onnanoko-ga ‘a girl’ in the canonical position of a grammatical object (immediately before a predicate) and then describes it again with a semantically more specific phrase tiisana onnanokoga ‘a little girl’ in post-predicative position. Table 2 summarizes the frequencies of utterances with the final-appendage construction in the Japanese data.6 Most of the final-appendage structures were

6. The figures in “multiple postposing” represent the number of the relevant final-appendage utterances, not the number of linguistic elements occurring in post-predicative position. Thus, an example like (5b) is counted as one token, although two elements are postposed.

Chapter 6. The final-appendage construction in Japanese and Korean

examples of simple postposing (93.2 percent). The remaining seven percent belonged to the other types of final-appendage construction. Table 2. Frequencies of utterances with the final-appendage construction in the Japanese data7 Simple postposing

Multiple postposing

Postposing with a preceding pronoun

Hyooten

69

0

0

1

70

Dakoosuru Tsuki

23

0

1

1

25

Entaku

42

2

0

0

44

Saikoo-no Rikon

80

4

3

3

90

Tenki-no Ko

60

3

1

1

65

274

9

5

6

294

(93.2%)

(3.1%)

(1.7%)

(2.0%)

Total

Repetition TOTAL

Table 3 compares the occurrences of utterances with the final-appendage construction in the Japanese texts (as shown in Table 2) and those in the Korean translations. As Table 3 shows, the Japanese texts contained 294 utterances of the final-appendage construction. In their Korean translations, however, there were only 90 utterances of similar final-appendage structures. That is, only 30.6 percent were expressed as final-appendage structures in Korean. The remaining 69.4 percent exhibited different syntactic realizations in the translations.

7. “Multiple postposing” can be regarded as a combination of two or more cases of postposing. In (i), for example, sore and onna-dake-no gengo occur after the interrogative predicate nani (an interrogative pronoun), representing cases of “simple postposing” and “postposing with a preceding pronoun,” respectively: (i) Nani1 sore, onna-dake-no gengo1? what that woman-only-gen language ‘What’s that, a language only for women?’

(S, p. 191)

The interrogative predicate nani is coreferential with the second postposed element (onna-dakeno gengo). In Table 2, these were counted as examples of “multiple postposing,” although other types of postposing (e.g., “simple postposing” and “postposing with a preceding pronoun”) are involved.

153

154

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

Table 3. Frequencies of utterances with the final-appendage construction in the Japanese novels and those of their Korean Translations Japanese original

Korean translation

Hyooten

70

17

Dakoosuru Tsuki

25

14

Entaku

44

25

Saikoo-no Rikon

90

19

Tenki-no Ko

65

15

Total

294

90

This result strikes us as interesting, because if the two languages had similar flexibility of constituent order, it would be reasonable to expect that Korean would likewise employ final-appendage structures in the translations. However, the result was contrary to this expectation; Korean appears to be more restrictive in the use of the final-appendage construction.

3.2 Korean translations of Japanese utterances with the final-appendage construction Except for the cases in which relevant utterances were not at all found in the Korean translations, there were four types of Korean renditions of Japanese finalappendage utterances: (i) Korean translations preserve final-appendage structures, (ii) postposed elements in Japanese utterances are realized in non-final position in Korean, (iii) postposed elements are translated as independent clauses, and (iv) postposed elements are not expressed at all. Examples (6)–(8) show that Korean encodes some linguistic elements in postpredicative position, as indicated in boldface. i.

Final position (6) J: Ee-naa doramusiki! be.good-fp drum.type    ‘(It)’s good, a front loader.’ K: Joh-gess-da, deureomsetaggi! be.good-will-fp drum.type.washing.machine    ‘(I) like (it), a front loader washing machine!’

(E, p. 64)

(W, p. 71)

Chapter 6. The final-appendage construction in Japanese and Korean

(7) J: Wasure-tei-ta-n-desu-mono, murai-san-nado. forget-perf-pst-fn-cop.pol-fp Murai-Mr.-and.the.like    ‘(I)’ve forgot about (someone) like Mr. Murai.’ K: Ijeo-beoryeo-sseo-yo, murai-ssi-gat-eun saram-eun. forget-perf-pst-fp Murai-Mr.-be.like-adn person-top    ‘(I)’ve forgot about a person like Mr. Murai.’ (8) J: Huzakeru-na, baka. kid.around-don’t idiot    ‘Don’t be silly, idiot.’ K: Usgiji-ma, i babo. make.laugh-don’t this idiot    ‘Don’t be silly, this idiot.’

(H, p. 29)

(B, p. 29)

(D, p. 22)

(G, p. 28)

In (6), post-predicative position is occupied by subject nominals without case particles in both Japanese and Korean. In (7), object nominals come after the predicates, accompanied by the particle of the general extender ‑nado and the topic marker ‑eun in the Japanese and Korean utterances, respectively. In (8), vocative phrases occupy utterance-final position with the demonstrative pronoun i ‘this’ additionally attached in the Korean translation. Examples (9)–(11) illustrate cases in which Japanese postposed elements are realized in non-final position in the Korean translations. In (9), the subject nominal (soree ‘that one’) occupies post-predicative position in Japanese, but the corresponding phrase occurs in the unmarked object position in the Korean translation. In (10), the adverbial phrase (sugoku ‘really’) appears in utterance-final position, although in unmarked constituent order it would precede the adjective (ii ‘good’) it modifies. On the other hand, the Korean translation preserves the unmarked order of modification with the adverbial (aju ‘very’) placed before the adjective (joha ‘good’) it modifies. In (11), the adverbial hitotu ‘one’ in Japanese occurs utterancefinally, but jom ‘a little,’ which roughly corresponds to hitotu, occupies initial position in Korean. ii. Non-final position (9) J: Ee-yan, soree. be.good-fp that    ‘(It)’s good, that (one).’ K: Aa, geugeo ha-gosipda. oh that do-want.to    ‘(I) want to do that.’

(E, p. 104)

(W, p. 111)

155

156

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

(10) J: Ii-n-zyanai-kana, sugoku. be.good-fn-isn’t:it-fp really    ‘(He) is good, really.’ K: Aju joha. very be.good    ‘(He) is very good.’ (11) J: Ikaga-desu, hitotu. how-cop.pol one    ‘Would (you) like one?’ K: Jom deureo bo-si-ji-yo. a.little eat look-hon-fp-fp    ‘Would (you) like some?’

(S, p. 241)

(C, p. 236)

(H, p. 357)

(B, p. 294)

Examples (12)–(14) represent cases in which the postposed elements of Japanese final-appendage utterances are translated as independent clauses (mostly independent utterances) in Korean. In (12), the subject nominal with a topic marker (monomorai-toiu yatu-wa ‘something called stye’) is postposed in the Japanese utterance, but in the Korean translation the corresponding part forms a separate utterance, as indicated by the period following the preceding utterance. In (13), the adverbial dokonimo ‘anywhere’ occurs after the predicate of the same sentence, but the Korean counterpart (eodiedo ‘anywhere’) stands as an independent utterance. In (14), the adverbial clause (ore yaru-nde ‘because I’ll do it myself ’) is postposed in Japanese, but the relevant part is expressed as a different utterance terminated by the final particle ‑yo. iii. Independent clauses (12) J: Yappari uturu-n-ya-na, monomorai-toiu yatu-wa. as.expected be.infectious-fn-cop-fp stye-known.as stuff-top    ‘A s expected, (it) is infectious, something called stye.’ (E, p. 37) K: Geureom jinjjaro olm-neun geo-nga bo-ne. Nun dareggi mal-i-ya. then truly move-adn thing-if see-fp eye stye word-cop-fp    ‘Then, (it) really seems infectious. I mean, eye stye.’ (W, p. 42) (13) J: Hina-san-ga inai-n-da, dokonimo! Hina-Ms.-nom be.neg-fn-cop anywhere    ‘(We) can’t find Hina anywhere!’

(T, p. 212)

Chapter 6. The final-appendage construction in Japanese and Korean

K: Hina-ssi-ga eobseojyeosseo. Eodiedo! Hina-Ms.-nom disappear.pst anywhere    ‘(We) find Hina is not (present). Anywhere!’ (14) J: Iya, daijoobu-desu, ore yaru-nde. no all.right-cop.pol I do-cp    ‘No, no problem, because I’ll do (it myself ).’ K: Ani, gwaenchanha-yo. Jigjeob ha-lge-yo. no don’t.care-fp directly do-will-fp    ‘No, no problem. (I) will do (it) on (my) own.’

(N, p. 242)

(S, p. 233)

(C, p. 229)

There were also cases where the postposed elements of Japanese utterances were not found in the Korean translations, as shown in (15)–(17). iv. Unexpressed (15) J: Nani-yo sore what-fp that    ‘What’s that?’ K: Museun mal-iya? what word-cop.fp    ‘What word is (that)?’ (16) J: Murai-mo warui otoko-zyanai-n-da, are-de. Murai-afp bad man-cop.top.neg-fn-cop that-with    ‘Murai isn’t a bad man, (though he may look) like that.’ K: Murai-do nabbeu-n nom-eun ani-lse. Murai-too be.bad-adn guy-top neg.cop-fp    ‘Murai isn’t such a bad guy.’ (17) J: Soremo Tanigawa-da-yo, Tanigawa. besides Tanigawa-cop-fp Tanigawa    ‘And what’s more, (her partner) is Tanigawa, Tanigawa.’ K: Gedaga sinlang-i Danigawa-rani! on.top.of.that bridegroom-nom Tanigawa-cop.quot.fp    ‘On top of it, the bridegroom is, they say, Tanigawa!’

(T, p. 186)

(N, p. 214)

(H, p. 105)

(B, p. 91)

(D, p. 106)

(G, p. 132)

In (15), the subject nominal corresponding to Japanese sore ‘that’ is missing in the Korean translation. Similarly in (16), the adverbial phrase are-de ‘(though he may look) like that’ is not translated in Korean, either. In (17), the Japanese example represents a case of repetition, where the complement nominal Tanigawa is rein-

157

158

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

troduced in utterance-final position, but no such repetition occurs in the Korean translation. Table 4 summarizes how the postposed elements of the Japanese finalappendage utterances (e.g., the boldfaced parts in (6J)–(17J) above) are realized in the Korean translations. The most frequent pattern of translation was not to retain the final-appendage construction in the Japanese originals but to express relevant parts in non-final position, which accounted for 36.3 percent. Only less than 30 percent of the Japanese data (29.7%) were translated similarly as final-appendage utterances in Korean. Some Japanese postposed elements were translated as independent clauses (23.1%) or not expressed at all (10.2%). Table 4. The Korean realizations of postposed elements in Japanese8 Final position

90

29.7%

Non-final position

110

36.3%

Independent clause

70

23.1%

Unexpressed

31

10.2%

Relevant part not found Total

2

0.7%

303

100%

If the two languages had similar flexibility of constituent order, utterances of the final-appendage construction would occur with similar frequency in the Korean translations. However, the final-appendage construction was only employed in less than 30 percent of the Korean data. This fact clearly shows that the two languages differ in the exploitation of post-predicative position.

3.3 Distribution by grammatical function To consider the differences in more detail, let us compare the postposed elements of the Japanese and Korean final-appendage utterances in terms of the grammatical functions those elements performed in the utterances. Table 5 shows the distribution of postposed elements by grammatical function. Compared with Korean, Japanese reveals a higher proportion of subject nominals, adverbials, and adnominals encoded in post-predicative position. About 27 percent of the Japanese utterances with the final-appendage construction contained subject nominals in post-predicative position in contrast to about 18 percent of their Korean counterparts. Also, the Japanese final-appendage utterances found 8. In the case of “multiple postposing,” two elements in utterance-final position were counted separately.

Chapter 6. The final-appendage construction in Japanese and Korean

Table 5. The distribution of postposed elements by grammatical function Japanese

Korean

Subject

83

27.4%

16

17.8%

Object

19

6.3%

6

6.7%

4

1.3%

2

2.2%

99

32.7%

22

24.4%

4

1.3%

0

0.0%

85

28.1%

42

46.7%

Interjection

5

1.7%

2

2.2%

Other

4

1.3%

0

0%

Total

303

100.0%

90

100.0%

Complement Adverbial Adnominal Vocative

four examples of adnominals in post-predicative position, but the Korean translations had no such examples. On the other hand, despite the smaller frequency of final-appendage utterances in Korean, vocative phrases (i.e., address phrases) accounted for about half of the postposed elements (46.7%), which is considerably higher in comparison to about 28 percent of their Japanese counterparts. In other words, the proportion of postposed elements serving as clausal constituents (subject, object, complement, adverbial, and adnominal) was higher in Japanese (69.0%) than in Korean (51.1%). The next section describes in more detail how such differences in the type of postposed element arise between the two languages.

4.

Post-predicative differences in form and frequency

4.1 Adnominals The post-predicative use of adnominal phrases was only observed in Japanese, as shown in Table 5. In (18J) and (19J), the genitive-marked adnominal phrases neetyan-no ‘(my) big sister’s’ and Keioosen-no ‘the Keio line’s’ modify the preceding nouns purezento ‘present’ and Hutyuu (a city’s name), respectively. Despite their functions as adnominal modifiers, the phrases occur after the predicates, not before the nouns they modify. On the other hand, the corresponding phrases in the Korean translations appear in non-final position, placed before the nouns modified, as in (18K) and (19K). (18) J: Tanzyoobi purezento-daro, neetyan-no. birthday present-will.be big.sister.dim-gen   ‘(It)’s a birthday present, isn’t it, (my) big sister’s?’

(T, p. 147)

159

160

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

K: Nuna saengil seonmul-i-ji? big.sister birthday present-cop-fp   ‘(It)’s (my) big sister’s birthday present, isn’t it?’

( = 3) (N, p. 171)

(19) J: De boku, Hutyuu-ni sun-deta-n-desu-yo, Keioosen-no. and I Fuchu-in live-prog.pst-fn-cop.pol-fp Keio.line-gen   ‘And I used to live in Fuchu, (a city) on the Keio line.’ (S, p. 36) K: Je-ga geo seon-e iss-neun huchu-e salass-geodeun-yo. I-nom Keio line-at be-adn Fuchu-at lived-fp-fp   ‘I used to live in Fuchu, which is on the Keio line.’ (C, p. 40)

4.2 Nominals The post-predicative occurrences of subject nominals also reveal a marked difference between the two languages. As seen in Table 5, the Japanese texts have a higher proportion of subject nominals in utterance-final position. One of the important contributors to this result is the “nani sore” construction, in which wh-predicates (or interrogative predicates) precede pronominal subjects, as in (20J) and (21J).9 The canonical constituent order of such sentences has a wh-predicate in final position as in (20’) and (21’). Final-appendage utterances such as (20J) and (21J) have non-predicate-final order with subject nominals (sore and anta) following the wh-predicates (nani and dare). In contrast, the Korean translations of the “nani sore” utterances still retain predicate-final constituent order, as in (20K) and (21K), respectively. [Non-predicate-final order: “nani sore” construction in Japanese] (20) J: Nani sore? what that   ‘What’s that?’ K: Geuge mweo-nya? that what-cop.fp   ‘What’s that?’ (21) J: Dare, anta? who you   ‘Who are you?’ K: Dangsin nugu-ya? you who-cop.fp   ‘Who are you?’

(S, p. 174)

(C, p. 171)

(T, p. 102)

(N, p. 118)

9. Pronominal subjects include simple pronouns (e.g., sore ‘that’) and nominals modified by demonstrative pronouns (e.g., sono kao ‘that face’) (see Table 6).

Chapter 6. The final-appendage construction in Japanese and Korean

[Canonical constituent order] (20’) J: Sore nani? that what   ‘What’s that?’ (21’) J: Anta dare? you who   ‘Who are you?’

Utterances with this kind of non-predicate-final order are referred to as “emotive type” (Ono & Suzuki 1992) or “emotionally charged utterances” (Ono 2006). It is argued that such non-canonical constituent order is “grammaticized to express emotion” (Ono 2006: 139) like surprise and disgust, characterized by the construction-specific intonation contour.10 In fact, “non-canonical order seems to be preferred over canonical order” (Ono & Suzuki 1992: 429) in such emotional utterances of Japanese. We found 20 examples of the “nani sore” construction in the Japanese data. As Table 6 shows, 19 examples of these are not translated into any forms of the finalappendage construction in Korean.11 In the Korean translations, the pronomi10. In spoken discourse, “nani sore” utterances are typically produced within a single intonation contour with no pause before the postposed elements (Ono & Suzuki 1992: 439; Ono 2006: 142). 11. The only exception is the following example: (i) J: Ha? Nani sore, onna-dake-no gengo? (S, p. 191, see also 12) huh what that woman-only-gen language   ‘Huh? What’s that, a language only for women?’ K: Eo, mweo-ya geuge yeoja-man-eui eoneo-ya? (C, p. 190) huh what-cop.fp that.nom woman-only-gen language-cop.fp   ‘Huh, what’s that, a language only for women?’ In both examples, post-predicative position is occupied by a pronominal subject (sore and geuge ‘that’), followed by the additional element specifying the referent of the interrogative pronoun (nani and mweo ‘what’). In the Korean translation, however, the nominative case particle ‑i is further added to the pronoun geugeos ‘that’ (geugeos + -i > geuge), although no such particle is used in the Japanese original. The case particle is in order in Korean but awkward in Japanese. 12. “Multiple postposing” can be regarded as a combination of two or more cases of postposing. In (i), for example, sore and onna-dake-no gengo occur after the interrogative predicate nani (an interrogative pronoun), representing cases of “simple postposing” and “postposing with a preceding pronoun,” respectively: (i) Nani1 sore, onna-dake-no gengo1? what that woman-only-gen language ‘What’s that, a language only for women?’

(S, p. 191)

The interrogative predicate nani is coreferential with the second postposed element (onna-dakeno gengo). In Table 2, these were counted as examples of “multiple postposing,” although other

161

162

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

nal subjects are placed in non-final position or not expressed at all. This fact shows that Korean is less tolerant of the placement of pronominal subjects after wh-predicates. Table 6. Examples of the “nani sore” construction and their Korean translations Japanese

Korean

Translation type

Itu-kara-desu, sore-wa. ‘How long has it been happening?’

(H) Eonje-puteo-i-nga-yo?

unexpressed

Nan-zyai, kore-wa! ‘What’s this?’

(H) Geureohdam eoddeohge doe-ngeoji!

non-final

Doosita-no, sono kao. ‘What happened to that face?’

(H) Eolgul-i wae geuraeyo?

non-final

Uwa, nan-ya koree. ‘Wow, what’s this?’

(E) Uwa! Ige mweo-ya?

non-final

Uwa, nan-ya aitu! ‘Wow, what is she doing?’

(E) Euag! Jyae mweohaneun-geo-ya?!

non-final

Nani are, atarasii katazukekata? ‘What’s that, a new way of cleaning up?’

(S) Jeoge da mweo-ya, saeroun jeongribeob-iya?

non-final

Nani kono kudaranai uta. ‘What is this stupid song?’

(S) I sisihan norae-neun mweo-ya.

non-final

Nani sono iikata! ‘How dare you say such a thing?’

(S) Museun mal-eul geureohge hae!

unexpressed

Nani sore? ‘What’s that?’

(S) Geuge mweo-nya?

non-final

E, nani sore? ‘Oh, what’s that?’

(S) Eo, geuge mweon soriya?

non-final

Nani sono kao. ‘What’s that face?’

(S) Geu pyojeong-eun mweo-ya.

non-final

Nani sore, omotya? ‘What’s that, a toy?’

(T) O! Ige mweo-rae? Jangnangam chong?

non-final

Nani-yo sore!? ‘What’s that?’

(T) Mweo-rago?!

unexpressed

Nani-yo sore, kasegeru-tte. ‘What do you mean by that, earning money?’

(T) Don-i doe-lgeo-ra-ni, geuge museun sori-ya?

non-final

······ E, nan-su-ka kore? ‘Huh, what’s this?’

(T) ······ Eo, geuge mweo-ya?

non-final

Dare, anta? ‘Who are you?’

(T) Dangsin nugu-ya?

non-final

types of postposing (e.g., “simple postposing” and “postposing with a preceding pronoun”) are involved.

Chapter 6. The final-appendage construction in Japanese and Korean

Table 6. (continued) Japanese

Korean

Translation type

Nani-yo sore. ‘What’s that?’

(T) Museun mal-iya?

unexpressed

Nan-da kore? ‘What’s this?’

(T) I seuwichi mweo-ji?

non-final

Nan-su-ka sore? ‘What’s that?’

(T) Geuge mweo-bnigga?

non-final

4.3 Particle-less pronouns As shown in Table 6, many Japanese examples of the “nani sore” construction (all except the first two examples) have pronominal subjects without particles (particle-less pronouns) in post-predicative position. That is, the utterance nani sore is much more common than nani sore-wa, where ‑wa is a topic particle. In Japanese, particle-less pronouns also occur in the post-predicative position of other types of the final-appendage construction. In (22J) and (23J), the pronouns kore ‘this’ and boku ‘I’ occupy post-predicative position without particles. (22) J: Iyaa nakanaka oisii-yo, kore. oh quite be.delicious-fp this   ‘Oh, (it) tastes quite good, this one.’ K: Aigo, igeos sangdanghi masiss-gumeon. oh this considerably taste.good-fp   ‘Oh, this tastes quite good.’ (23) J: Uun, sira-nai-yo boku. no know-neg-fp I   ‘No, I don’t know.’ K: Molla-yo. neg.know-fp   ‘(I) don’t know.’

(T, p. 20)

(N, p. 26)

(H, p. 91)

(B, p. 80)

Table 7 shows how Japanese particle-less pronouns in post-predicative position are translated into Korean.13 It represents the occurrences of postposed elements constituted of a pronoun alone. Of 37 cases, 29 examples (78.4%) do not preserve final-appendage structures. Most of them are either placed in non-final position as in (22K) or not expressed at all as in (23K) above. 13. The table includes examples where pronouns in post-predicative position are used as vocative phrases (see, for example, (26J)) below.

163

164

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

Table 7. The Korean counterparts of Japanese postposed particle-less pronouns Final position

Non-final position

Unexpressed

Independent clause

TOTAL

8

19

9

1

37

Interestingly, there were eight cases in which the Korean counterparts occupy utterance-final position, but they are translated differently from the Japanese originals. The Korean counterparts of Japanese particle-less pronouns have two main types of translation in post-predicative position: addition of particles and use of common nouns. In the first type, pronouns occupy post-predicative position in Korean, but particles are added to them to indicate the grammatical relations they have with the preceding elements. In examples (24)–(26), no particles are attached to utterance-final pronouns in Japanese, but in the Korean translations the case particle ‑ga or the topic particle ‑neun (phonologically contracted as ‑n) is added to the corresponding pronoun in final position. Such an additional marking suggests that Korean requires a more explicit marking of grammatical relations in post-predicative position.14 (24) J: Ukari-masu-ka, watasi? pass-pol-fp I   ‘Can I pass the exam?’ K: Habgyeogha-lgga-yo, je-ga? pass-will-fp I-nom   ‘Can I pass the exam?’ (25) J: Okori-masu-wa, watasi… get.angry-pol-fp I   ‘I’ll get angry.’ K: […] hwanae-lgeoye-yo, jeo-neun… get.angry-will-pol I-top   ‘I will get angry …’ (26) J: Dooiu sigoto-ga si-tai-n-da, omae? what.kind.of work-nom do-want.to-fn-cop you   ‘What kind of work do you want to do?’ K: Geunde eoddeon il-eul ha-gosipeo, neo-n?

(D, p. 33)

(G, p. 43)

(H, p. 10)

(B, p. 14)

(D, p. 32)

14. One of the reviewers observed that pronouns could appear in post-predicative position even without particles. However, it is important to note that this kind of observation still requires an explanation of why the Korean translators additionally used particles, which are not present in the Japanese originals.

Chapter 6. The final-appendage construction in Japanese and Korean

but what work-acc do-want.to you-top   ‘But what kind of work do you want to do?’

(G, p. 41)

The second way of translating Japanese particle-less pronouns into Korean is to employ common nouns instead of pronouns. Interestingly, these common nouns were all accompanied by demonstrative pronouns like geu ‘that’ or i ‘this’ in our Korean data. In examples (27)–(29), the Japanese original utterances contain the second and third person pronouns, temmee ‘you,’ kisamara ‘you bastards,’ and aitu ‘he,’ all of which are pejorative pronouns, normally used in informal speech. These pronouns are translated with the addition of demonstrative pronouns as i jasig-a ‘this bastard,’ i jasig-deul-a ‘these bastards,’ and geu saeggi ‘that bastard’ in the respective Korean translations.15 (27) J: […] tyotto mate-yo temee! a.little wait-fp you   ‘Wait a minute, you bastard!’ K: […] gidaryeo, i jasig-a! wait this bastard-voc   ‘Wait, you bastard!’ (28) J: Mate kisamara! wait you.pl   ‘Wait, you bastards!’ K: Geogi seo-rago, i jasig-deul-a! there stand-quot this bastard-pl-voc   ‘Stop (there), you bastards!’ (29) J: Iikagenni si-te hosii-noyo, aitu. moderately do-cp want-fp he   ‘(I) want (him) to stop annoying (me), that bastard.’ K: Jinjja eojiganhae-ya mal-i-ji geu saeggi. truth be.moderate-must work-cop-fp that bastard   ‘(He) must stop annoying (me), that bastard.’

(T, p. 82)

(N, p. 83)

(T, p. 225)

(N, p. 256)

(D, p. 108)

(G, p. 135)

15. One of the reviewers said that the usage of third person pronouns is restrictive in spoken Korean regardless of the position in which they occur. This may in part be related to the usage of common nouns in post-predicative position as seen in (29). However, the limited usage of third person pronouns cannot explain why second person pronouns are also replaced with common nouns as in (27) and (28) above. This again suggests that particle-less pronouns are less favored in post-predicative position in Korean than in Japanese.

165

166

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

Our data shows that except for one example of demonstrative adverb, all the other seven cases undergo either of the two changes in linguistic form when finalappendage structures are retained in Korean (see 16 for the remaining example, which represents the addition of a case particle to a pronoun in post-predicative position).18 This suggests that Korean is less likely to use pronouns alone in postpredicative position.

16. The only exception is the following example: (i) J: Ha? Nani sore, onna-dake-no gengo? (S, p. 191, see also 17) huh what that woman-only-gen language   ‘Huh? What’s that, a language only for women?’ K: Eo, mweo-ya geuge yeoja-man-eui eoneo-ya? (C, p. 190) huh what-cop.fp that.nom woman-only-gen language-cop.fp   ‘Huh, what’s that, a language only for women?’ In both examples, post-predicative position is occupied by a pronominal subject (sore and geuge ‘that’), followed by the additional element specifying the referent of the interrogative pronoun (nani and mweo ‘what’). In the Korean translation, however, the nominative case particle ‑i is further added to the pronoun geugeos ‘that’ (geugeos + -i > geuge), although no such particle is used in the Japanese original. The case particle is in order in Korean but awkward in Japanese. 17. “Multiple postposing” can be regarded as a combination of two or more cases of postposing. In (i), for example, sore and onna-dake-no gengo occur after the interrogative predicate nani (an interrogative pronoun), representing cases of “simple postposing” and “postposing with a preceding pronoun,” respectively: (i) Nani1 sore, onna-dake-no gengo1? what that woman-only-gen language ‘What’s that, a language only for women?’

(S, p. 191)

The interrogative predicate nani is coreferential with the second postposed element (onna-dakeno gengo). In Table 2, these were counted as examples of “multiple postposing,” although other types of postposing (e.g., “simple postposing” and “postposing with a preceding pronoun”) are involved. 18. The only exception is the example in which both Japanese and Korean utterances contain locative pronouns in utterance-final position as in (i): (i) J: Suwat-te suwat-te, koko. sit-cp sit-cp here   ‘Sit down, sit down, here.’ K: Anja boa, anja boa, yeogi. sit see sit see here.’   ‘Sit down, sit down, here.’

(S, p. 142) (C, p. 143)

The fact that the pronoun yoegi used as a place adverbial ‘here’ appears in post-predicative position suggests that the Korean final-appendage construction is more open to expressions which are relatively free in position, as with the case of vocative phrases that will be discussed in Section 4.4.

Chapter 6. The final-appendage construction in Japanese and Korean

4.4 Vocative phrases The Korean translations showed a much lower frequency of final-appendage structures. However, they contained a relatively higher proportion of vocative phrases (i.e., address phrases). As seen in Table 5, about half of the postposed elements (46.7%) in Korean were vocative phrases, as exemplified in (30) and (31), where vocative phrases appear in utterance-final position, exactly like the Japanese counterparts. (30) J: Urusai boke! be.noisy idiot   ‘Shut up idiot!’ K: Siggeureoweo, babo-ya! be.noisy idiot-voc   ‘Shut up, idiot!’

(E, p. 79)

(W, p. 85)

(31) J: Konna kyakutanka-de nani-o yare-tte iu-n-da, such average.customer.spend‑at what-acc do.imp-quot say-fn-cop bakayaroo. idiot   ‘What are (they) telling (us) to do with this average customer spend,(D, p. 9) idiot?’ K: I danga-ro daechae mweol mandeuleo‑nae‑neunya‑go i this unit.price‑at on.earth what.acc make-put.out-q-quot this babo-ya! idiot-voc   ‘What the hell are (they) telling (us we should) make with this aver- (G, p.12) age customer spend, this idiot?’

Again in Korean, vocative phrases in post-predicative position are often accompanied by additional markings, for example, the vocative particle ‑(y)a as in (30) and/or the demonstrative pronoun i as in (31) (see also (8), (27) and (28) above), although no such markings are used in the Japanese original utterances.

4.5 Summary: Post-predicative position in Korean To summarize the observations thus far, a comparison of the final-appendage utterances in the Japanese novels and their Korean translations reveals that Korean showed a much lower frequency of final-appendage structures (only about 30 percent of the Japanese final-appendage utterances). Also, in the Korean translations, adnominals and subject nominals (especially particle-less pronouns as used in the

167

168

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

“nani sore” construction) occurred less frequently in post-predicative position. For postposed particle-less pronouns, many of the Korean translations did not retain the encoding pattern of Japanese original utterances. Even when such pronouns are used in the post-predicative position of the Korean utterances, they are either additionally followed by particles or replaced with common nouns. On the other hand, despite a lower frequency of final-appendage utterances, the Korean translations represented a relatively higher proportion of vocative phrases in post-predicative position, which are often accompanied by the additional markings of vocative particles and/or demonstrative pronouns. In other words, Korean postposed elements are, if anything, more morphophonologically and/or syntactically marked than Japanese, requiring a more explicit grammatical relation marking or a more contentful or emphatic rendition of their corresponding Japanese elements.19

5.

Constructional entrenchment of final-appendage structures

The difference between Japanese and Korean in the use of the final-appendage construction can be considered in terms of the notion of “constructional entrenchment,” that is, the state and process of a construction being “entrenched (or stored as a unit)” (Traugott & Trousdale 2013: 48) in the minds of speakers. In our previous study (Izutsu & Izutsu 2018), we discussed the constructional entrenchment of final-tagged structures in Irish and American English. Comparing two English corpora (SPICE-Ireland and the Santa Barbara Corpus of Spoken American English), the study revealed that both type and token frequencies of final tags (i.e., pragmatic markers in utterance-final position) were higher in Irish English than in American English.20 The results led us to conclude that “Irish English displays a greater degree of the entrenchment of the final-tag construction [[ANCHi FTj] ↔ [S conclude verbalization of propositioni with attitudej]] than American English” (Izutsu & Izutsu 2018: 414). Our comparison of the Japanese texts and their Korean translations points to a similar difference in the constructional entrenchment of final-appendage structures.21 We have seen that Japanese exhibited a much greater frequency of the 19. The present results will need to be replicated for verification by using the reverse set of data (Korean novels and their Japanese translations). See Izutsu, Izutsu & Kim (2021) for a qualitative analysis of examples including this reverse data set. 20. SPICE-Ireland (Kirk et al. 2011a) represents the Systems of Pragmatic Annotation in the Spoken Component of ICE-Ireland (Kirk et al. 2011b). 21. We employ the term “constructional entrenchment” rather than “contructionalization” (more precisely, “grammatical constructionalization”), because our studies do not illustrate how

Chapter 6. The final-appendage construction in Japanese and Korean

elements occupying post-predicative position and allowed a higher frequency of clause-internal elements (e.g., subject nominals and adnominals) to be placed in that position. On the other hand, Korean showed a high frequency of vocative phrases in post-predicative position, while it was more restrictive on the placement of clause-internal elements in the position. Furthermore, as seen in the “nani sore” construction, pronominals occupied post-predicative position, often without particles in Japanese. In Korean, however, pronominals were less likely to occur in post-predicative position, and when they occurred, they tended to be accompanied by particles or replaced with common nouns. These differences reveal that the two languages markedly differ in the exploitability of postpredicative position. Japanese allows a more flexibility and productivity of postpredicative position than Korean, thereby suggesting a relatively higher degree of constructional entrenchment of final-appendage structures. This contrast in constructional entrenchment can explain the difference in the development of utterance-final connectives, or more precisely that of the retrospective use of final connectives, between Japanese and Korean. As discussed in our previous study (Izutsu & Izutsu 2018), Irish English, which shows a higher degree of constructional entrenchment of final-tagged structures, has the finalparticle use of but as in (32) like Australian English (Mulder & Thompson 2008; Mulder, Thompson & Penry Williams 2009), and it also has the final particle so as in (33). However, American English does not have such final-particle uses of these connectives.22 Final particle but (the retrospective use of final but) (32) a. < # > Och that’s easy to do but isn’t it (ICE-Ireland S1A-042) b. B: < # > The perm makes me look like Shirley Temple A: < # > Yeah she’s young and cute but < # > What what else do you want (ICE-Ireland S1A-059, also in Kirk & Kallen 2010: 203)

a new form-meaning pairing has come into being (see also Izutsu & Izutsu 2018). We describe how the final-appendage and final-tagged constructions, both of which are already present as symbolic constructions, are exploitable in different languages or varieties. 22. The term “final particle” means the retrospective use of connectives, which involves a connective (e.g., but) retrospectively combining two preceding clauses (X, Y ) as in X, Y but. It does not include the prospective or “final hanging” use (Mulder & Thompson 2008; Mulder, Thompson & Penry Williams 2009), which results from suppressing the information to be expressed in the second clause, as in X but (…). Izutsu & Izutsu (2014) distinguish the two uses of final connectives into the “backshift” and “truncation” types, respectively.

169

170

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

Final particle so (the retrospective use of final so) (33) a. B: < # > and there was a blue light flashing in the window and everything < # > So I said there was < { > < [ > probably a fire < /[ > C: < # > < [ > It was you < /[ > < /{ > opened the curtains so B: < # > Yeah

(ICE-Ireland S1A-050) b. < # > And < , > what kind of music do you like so < , > < { > < [ > if < /[ > you don’t like the music they play in the clubs (ICE-Ireland S1A-057)

A similar contrast is observed between Japanese and Korean. Although the present data did not have examples of final connectives, it is known that the two languages allow postposing adversative connectives (demo ‘but’ and guende ‘but’) in utterance-final position, as given in (34). (34) J: Mondai-wa sinkoku-yo demo. problem-top serious-fp but K: Munje-ga simgaghae geunde. problem-nom be.serious but   ‘The problem is serious, though.’

(Sohn & Kim 2014: 242)

However, like Irish English, such use of final connectives seems to be more pervasive in Japanese. Apart from adversative connectives, the causal connective dakara ‘so’ can also be used in utterance-final position in Japanese as in (35J), but the corresponding Korean connective geureonigga ‘so’ is less acceptable in that position as in (35K). (35) J: Yohoo-de-wa kyoo hura-nai-tte itten-no dakara. forecast-at-top today rain-neg-quot be:saying-fp so (Izutsu & Izutsu 2013: 223) ?? K: Ilgiyebo-eseo‑neun bi-ga an onda-go haess-da-nigga-yo, weather:forecast‑at-top rain-nom neg come-quot said-quot-because-fp geureonigga. so   ‘(I)’m telling (you) that the weatherman says it won’t rain today, I mean!’

Similarly, the Japanese (colloquial) connective datte, a causal connective of rebuttal implication, can also be used in utterance-final position as in (36J), but the Korean connective of a similar meaning geuchiman ‘but’ is less likely to occur in the position as in (36K). The single question mark (?) indicates that geuchiman is unnatural when it forms a single intonation contour with the preceding clause, while the Japanese datte is perfectly natural even when produced in a single contour with no pause preceding it.

Chapter 6. The final-appendage construction in Japanese and Korean

(36) J: Kakkoii-n-da-mon, datte. be.cool-fn-cop-fp so/but K: ?Jal saengyeoss-geodeun-yo, geuchiman. be.well appear.pst-because-fp but   ‘(It)’s cool, though.’

(School of Lock, 3/6/2016)

Such a contrast in the acceptability of utterance-final connectives can be viewed as a reflection of the difference in the entrenchment of the final-appendage construction between the two languages, just like the one found between Irish and American English. The comparison of Japanese texts with their Korean translations shows that the use of the final-appendage construction is more versatile and prevalent in Japanese, which is further endorsed by the fact that Japanese allows a wider variety of connectives to be used in post-predicative position than Korean. These observations suggest that the Japanese final-appendage construction has become a more schematic category accommodating a wider variety of expressions (both in type and token) in post-predicative position.

6.

Conclusion

Japanese and Korean are known for relatively flexible constituent order, and similar functional motivations, for example, the “important information first” principle (Simon 1989) and the “urgency/relevance of information” factor (T. Kim 2011), have been proposed for the postposing construction of the two languages. However, our contrastive analysis of Japanese texts and their Korean translations reveals that the two languages substantially differ in the exploitability of postpredicative position. Just as Irish and American English exhibit different degrees of entrenchment of the final-tag construction, Japanese and Korean likewise differ in the degree of constructional entrenchment of final-appendage structures. The fact that the Korean translations manifested a much lower frequency of utterances with such structures suggests that Korean shows a lower degree of entrenchment of the final-appendage construction. This observation is also supported by the fact that despite the lower frequency of final-appendage utterances, post-predicative position in Korean is more open to elements with high positional flexibility (e.g., vocative phrases) than to clause-internal elements (e.g., adnominals, subjective nominals). These positionally flexible elements can appear in post-predicative position without much help from the final-appendage construction. We consider that such a difference in the degree of entrenchment of the finalappendage construction is primarily attributed to a higher degree of “grammatical relation sensitivity” of Korean (Izutsu, Izutsu, & Kim 2021). In post-predicative

171

172

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

position, which cannot encode a grammatical relation by means of constituent order, Korean more strongly requires a morphological marker (case, topic, or vocative particles) indicating a grammatical relation, as seen in Sections 4.2 to 4.4. This greater sensitivity to grammatical relations is in accordance with E. Kim’s observation that Korean more prefers an explicit encoding of grammatical subjects than Japanese and such subjects are more likely to occupy sentence-initial position in Korean (E. Kim 2013). In other words, the more a language is sensitive to indicating a grammatical relation either with morphological marking or constituent order, or both, the more restrained the speakers are in the placement of clausal constituents outside of the canonical constituent order; that is, the final-appendage construction is less entrenched in the minds of the speakers. Although postposing may be driven by similar functional motivations in Japanese and Korean, to what extent the final-appendage construction is entrenched in the minds of speakers of the two languages significantly differs depending on the grammatical relation sensitivity of the languages.

Acknowledgements We are deeply indebted to our reviewers for their helpful and constructive comments and suggestions. Our thanks also go to Sylvie Hancil for planning the Gramm3 conference, which had to be canceled due to the COVID-19 pandemic. All remaining errors and shortcomings are of course ours.

Keys to abbreviations acc adn afp cop cp dim fn fp gen hon imp

accusative adnominal additive focus particle copula connective particle diminutive formal noun final particle genitive honorific imperative

neg nom perf pl pol prog pst top q quot voc

negation nominative perfect plural polite progressive past tense topic question quotative vocative

Chapter 6. The final-appendage construction in Japanese and Korean

References Birner, Betty J. & Ward, Gregory. 1998. Information Status and Noncanonical Word Order in English [Studies in Language Companion Series 40]. Amsterdam: John Benjamins. Fujii, Seiko. 1995. Nihongo-no gojun-no gyakuten-nitsuite: Kaiwa-no naka-no nagare-o chuushinni (Word order reversal in Japanese: With special reference to a flow of conversation). In Nichieigo-no Uhooidookoobun: Sono Koozoo-to Kinoo (Rightward Movement in Japanese and English: Their Structure and Function), Ken-ichi Takami (ed.), 167–198. Tokyo: Hitsuzi. Izutsu, Katsonobu & Izutsu, Mitsuko Narita. 2013. From discourse markers to modal/final particles: What the position reveals about the continuum. In Discourse Markers and Modal Particles: Categorization and Description [Pragmatics & Beyond New Series 234], Liesbeth Degand, Bert Cornillie & Paola Pietrandrea (eds), 217–235. Amsterdam: John Benjamins. Izutsu, Katsunobu, Izutsu, Mitsuko Narita & Kim, Yong-Taek. 2021. Grammatical relation sensitivity: Some different conceptions of pre/post-predicative structues. Japanese/Korean Linguistics 28. (29 December 2022). Izutsu, Mitsuko Narita & Izutsu, Katsonobu. 2014. Truncation and backshift: Two pathways to sentence-final coordinating conjunctions. Journal of Historical Pragmatics 15(1): 62–92. Izutsu, Mitsuko Narita & Izutsu, Katsonobu. 2018. Cross-varietal diversity in constructional entrenchment: The final-tag construction in Irish and American English. In New Trends in Grammaticalization and Language Change [Studies in Language Companion Series 202], Sylvie Hancil, Tine Breban & José Vicente Lozano (eds), 381–430. Amsterdam: John Benjamins. Kim, Alan Hyun-Oak. 1988. Preverbal focusing and type XXIII languages. In Studies in Syntactic Typology [Typological Studies in Language 17], Michael Hammond, Edith A. Moravcsik & Jessica Wirth (eds), 147–169. Amsterdam: John Benjamins. Kim, Eunae. 2013. Nihongo-to kankokugo-niokeru shugo-no arawarekata-nitsuite (On the occurrences of subjects in Japanese and Korean). Journal of the Faculty of Integrated Human Studies and Social Sciences, Fukuoka Prefectural University 22(2): 55–62. Kim, Taeho. 2011. An empirical study of postposing constructions in Korean. Linguistic Research 28(1): 223–238. Kirk, John M. & Kallen, Jeffrey L. 2010. How Scottish is Irish Standard English? In Northern Lights, Northern Words. Selected Papers from the FRLSU Conference, Kirkwall 2009, Robert McColl Millar (ed.), 178–213. Aberdeen: Forum for Research on the Languages of Scotland and Ireland. (29 December 2022). Kirk, John M., Kallen, Jeffrey L., Lowry, Orla, Rooney, Anne & Mannion, Margaret. 2011a. The SPICE-Ireland Corpus: Systems of Pragmatic Annotation for the Spoken Component of ICE-Ireland. Version 1.2.2. Belfast: Queen’s University Belfast and Dublin: Trinity College Dublin. Kirk, John M., Kallen, Jeffrey L., Lowry, Orla, Rooney, Anne & Mannion, Margaret. 2011b. International Corpus of English: Ireland Component. The ICE-Ireland Corpus. Version 1.2.2. Belfast: Queen’s University Belfast and Dublin: Trinity College Dublin.

173

174

Mitsuko Narita Izutsu, Katsunobu Izutsu & Yong-Taek Kim

Maynard, Seiko K. 1989. Japanese Conversation: Self-contextualization through Structure and Interactional Management. Norwood NJ: Ablex. Mihara, Ken-ichi. 2015. Chizu seisaku keikaku-niokeru nihongo-no koochibun (A cartographic approach to right dislocation in Japan). Studies in Japanese Language and Culture 25: 1–11. Mulder, Jean & Thompson, Sandra A. 2008. The grammaticization of but as a final particle in English conversation. In Crosslinguistic Studies of Clause Combining: The Multifunctionality of Conjunctions [Typological Studies in Language 80], Ritva Laury (ed.), 179–204. Amsterdam: John Benjamins. Mulder, Jean, Thompson, Sandra A. & Penry Williams, Cara. 2009. Final but in Australian English conversation. In Comparative Studies in Australian and New Zealand English: Grammar and Beyond [Varieties of English Around the World G39], Pam Peters, Peter Collins & Adam Smith (eds), 339–359. Amsterdam: John Benjamins. Ono, Tsuyoshi. 2006. An emotively motivated post-predicate constituent order in a ‘strict predicate final’ language: Emotion and grammar meet in Japanese everyday talk. In Emotive Communication in Japanese, Satoko Suzuki (ed.), 139–153. Amsterdam: John Benjamins. Ono, Tsuyoshi & Suzuki, Ryoko. 1992. Word order variability in Japanese conversation: Motivations and grammaticization. Text 12(3): 429–445. School of Lock. 3 June, 2016. (19 February, 2020). Shimojo, Mitsuaki. 2005. Argument Encoding in Japanese Conversation. New York NY: Palgrave Macmillan. Simon, Mutsuko Endo. 1989. An Analysis of the Postposing Construction in Japanese. PhD dissertation, University of Michigan. Sohn, Sung-Ock S. & Kim, Stephanie Hyeri. 2014. The interplay of discourse and prosody at the left and right periphery in Korean: An analysis of kuntey ‘but.’ In Discourse Functions at the Left and Right Periphery: Crosslinguistic Investigations of Language Use and Language Change [Studies in Pragmatics 12], Kate Beeching & Ulrich Detges (eds), 221–249. Leiden: Brill. Takami, Ken-ich. 1994. Nichiego taishookenkyuu (5): Nichieigo-no koochibun-to joohookoozoo (A contrastive study of Japanese and English: Postposing and information structure in Japanese and English). Eigo Seinen (The Rising Generation) 140(5): 230–232. Takami, Ken-ichi. 1995. Nichieigo-no koochibun-to joohookoozoo (Postposing and information structure in Japanese and English). In Nichieigo-no Uhooidookoobun (Rightward Movement in Japanese and English: Their Structure and Function), Ken-ichi Takami (ed.), 149–165. Tokyo: Hitsuzi. Traugott, Elizabeth Closs & Trousdale, Graeme. 2013. Constructionalization and Constructional Changes. Oxford: OUP. Woolford, Ellen. 1999. Grammatical relations. In The MIT Encyclopedia of the Cognitive Sciences, Robert A. Wilson & Frank C. Keil (eds), 355–357. Cambridge MA: The MIT Press.

Chapter 6. The final-appendage construction in Japanese and Korean

Data sources Miura, Ayako. 1982 [1965]. Hyooten 1. Tokyo: Kadokawa. (Bingjeom, trans. by Ho Choe. Seoul: Hong Shin Publishing, 2017 [1992]). Nishi, Kanako. 2013 [2011]. Entaku. Tokyo: Bungeishunju. (Weontag, trans. by Heuiseon Im. Bucheon: Bookstory Publishing, 2014). Sakamoto, Yuji. & Momose, Shinobu. 2013. Saikoo-no Rikon 1. Tokyo: Fusosha. (Choego eui Ihon 1, trans. by Jina. Chu. Seoul: Bagha, 2018). Sakuragi, Shino. 2016 [2013]. Dakosuru Tsuki. Tokyo: Futabasha. (Gubichi neun Dal, trans. by Yunog Yang. Seoul: Hyeondae Munhag, 2015). Shinkai, Makoto. 2019. Tenki no Ko. Tokyo: Kadokawa, 2019. (Nalssi eui Ai, trans. by Gyeongog. Min. Seoul: Daeweonssiai, 2019).

175

chapter 7

New perspectives on phonological erosion as an aspect of grammaticalization Charles Elerick

The University of Texas at El Paso

Phonological erosion as a concomitant of grammaticalization has been a prevailing assumption for four decades. Phonological erosion is, in fact, a general diachronic process with reduction occurring at the rate of about 15–20% per millennium. Reduction related to grammaticalization is often faster than the nominal rate and this has supported the standard theory. In a language, which is an adaptive and isostatic system, phonological loss, at nominal or accelerated rates, whether associated with grammaticalization or otherwise, interacts with mechanisms of compensation. Such interaction, a complex process, supports lexical right-sizing, an expression of the Quantity Principle. Phonological loss that accompanies grammaticalization, often significant and often uncompensated, is one aspect of this general diachronic process.

1.

Introduction

Antoine Meillet (1921) published his broadly conceived exploration of grammaticalization, including discussion that foreshadowed later studies, over a century ago. He in fact invented the term in its French form. In the decades to follow, grammaticalization received some continuing notice and was cited as an explanation for change in individual forms. Lehmann (1985) energized its study when he systematically engaged lexical and morpho-syntactic aspects of grammaticalization and demonstrated its importance and explanatory potential. His work, which incorporated important additional insights regarding language change, will be revisited later in this paper. Heine and Reh (1984: 67) codified the essentials that characterize this complex process, including the bleaching of referential semantic content, an increase in frequency, a restriction in context of occurrence, and the frequent erosion of phonological bulk. Such developments, including erosion, are widely evidenced in the Latin > French continuation. We see, for example, cependant [spɑ̃

https://doi.org/10.1075/slcs.232.07ele © 2023 John Benjamins Publishing Company

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

dɑ̃] ‘however’ < Latin ecce pendente ‘with that hanging’. The four essentials, taken together, have served as a point of departure for myriad subsequent studies. In a paper that reviewed the place of grammaticalization within the larger framework of historical linguistics, Joseph (2015) offered significant observations regarding apparent cases of grammaticalization for which competing, if not compelling, alternative analyses are available. He recognized the importance of grammaticalization as a diachronic process but cautioned against over-enthusiasm for any single linguistic concept or explanation. His work serves as a reminder that the reexamination of any linguistic precept and informing data by looking to larger contexts is always in order. The enshrined observation by Heine and Reh (1984) that phonological reduction is a typical concomitant of grammaticalization is the issue that will be reexamined in this study. The essential validity of the claim that grammaticalization tends to promote phonological erosion has been affirmed by subsequent studies, for example the extending work of Heine, Claudi & Hünnemeyer (1991). In a corrective vein, Campbell (2001) pointed out that phonological erosion, whether associated with grammaticalization or not, was an inherent dimension of language change. Schiering (2010) noted that phonological erosion is not to be expected to consistently accompany grammaticalization. This paper will offer an exploration of how the phonological reduction that may be associated with grammaticalization is, in fact, part of a much larger regime of diachronic change. In the last four decades that have seen much investigation of grammaticalization as an aspect of historical linguistics, questions other than phonological reduction have received the most attention. In their status review of the role and function of grammaticalization Traugott and Hopper (1993) did not substantively address the issue of phonological reduction per se. They did note, correctly and significantly, the preference for unmarked or reduced phones in forms that result from grammaticalization. Their observation of the infrequent occurrence of high profile, expressive phones, for example strident palatals, is a significant observation. The treatment of phonetic erosion in corresponding material in Traugott and Hopper’s 2nd edition (2003), ten years on, did not significantly extend the earlier version. Boye and Harder (2012), in a prominently placed paper that purported to reformulate how grammaticalization should be understood, offer a few random observations on phonological reduction. They observed, once again, that grammaticalization typically involves the loss of phonological bulk and they mentioned the standard assumption that increased frequency of occurrence promotes reduction. They noted that “the low discourse prominence of grammatical expressions motivates their low phonological prominence”, a point that had been made by Traugott and Hopper. In sum, one can only note that while repeatedly referenced for almost four decades, with some pertinent discussion, phonological

177

178

Charles Elerick

reduction as an aspect of grammaticalization has received more notice than elucidation. It has not been the subject of systematic investigation. The present study, as indicated, will show that the accelerated reduction of phonological bulk that is often associated specifically with grammaticalization must be understood within the framework of the loss of phonological bulk as a general diachronic phenomenon. It is also the case that accelerated loss of phonological bulk can occur in forms that have not evolved as instances of grammaticalization. In addition, there are instances of grammaticalization that do not involve significant phonological reduction, and these can also be explained as a reflection of that same general framework of understanding. Phonological reduction at either an accelerated or nominal rate or the absence of significant erosion, contrary to the standard assumption, is in fact embedded in the fundamental nature of language. Language is to be understood as a complex and ever-evolving adaptive system (Beckner 2009), or perhaps better, an isostatic, i.e. self-rebalancing, system and, recalling the memorable treatise of Hockett (1968), one that is stable but not well-defined and subject to “wear and tear.” As an orientation to the overarching diachronic process, extended to implications for grammaticalization, we should note these prime assumptions, outlined here and elaborated in the ensuing discussion. – – – – –

2.

Phonological erosion is a ubiquitous and ongoing process. Lost phonological bulk is offset by various mechanisms of rebulking. The loss and compensation cycle maintains, stochastically, a balance between informatic quantity and complexity of sign, is a key premise of information theory. The compensation for low (lost) bulk has a “right-sizing” effect but under certain conditions right-sizing may involve accelerated debulking. The loss of phonological bulk associated with the grammaticalization proceeds as one special aspect of this more general diachronic principle and associated process.

The loss of phonological bulk: The preliminary benchmark

Phonological erosion and the resulting diachronic loss of phonological bulk is readily observable when we have data from successive stages of a language. We can assemble an extended and representative sample of a natural lexical class and compare earlier forms with their reflexes. A particularly fruitful source of data is the Latin > Romance continuation. Note that in this paper, most of the data are drawn from the Latin and the Romance progression, a goldmine for the study of diachrony.

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

The working assumption is that Latin > Romance data reflects decay and compensation at or near some universal rate but this may be confirmed or modified on the basis of studies undertaken in other language families. Also, the Latin > Romance data offers a view of about two millennia but we should assume that decay and compensation took place before and will continue and therefore the treatment in the present paper should not be taken as representing a closed end outcome. A review of food and foodstuff terms in Latin that have French reflexes provides an informal but instructive picture of loss of phonological bulk. Note avēna- > avoine ‘oats’; aliu- > ail ‘garlic’; lactūca- > laitue ‘lettuce’; vīnu- > vin ‘wine’; lacte- > lait ‘milk’; pane- > pain ‘bread’; melle- > miel ‘honey’; butyrū- > beurre ‘butter’. A more quantifiable result emerges when we consider, with attention to phonetic detail, a set of Latin terms for common (non-exotic) animals, domestic and non-domestic, all of which have reflexes in Modern French and which evidence descent by popular transmission. The data in Table 1 show a significant loss of phonological bulk. The 68 phones that comprise the Latin etyma are represented by 40 continuing phones in the French reflexes, representing a loss at the rate of about 20% per millennium. Table 1. Phonological erosion in Latin > French animal terms Latin etymon (generalized stem)

Phones

Mod. Fr. reflex

Phones

Loss(gain)

musca- ‘fly’

5

mouche [muʃ ]

3

2

pūlice- ‘flea’

6

puce [pys]

3

3

apicula- ‘bee’

7

abeille [abɛj]

4

3

formīca- ‘ant’

7

fourmi [furmi]

5

2

ciconia- ‘stork’

7

cigogne [sigɔɲ]

5

2

aquila- ‘eagle’

6

aigle [ɛgl]

3

3

caballu- ‘horse’

7

cheval [ʃɛval]

5

2

lupu- ‘wolf ’

4

loup [lu]

2

2

cane- ‘dog’

4

chien [ʃjɛ̃ ]

3

2(1)

cattu- ‘cat’

5

chat [ʃa]

2

3

capra- ‘she-goat’

5

chèvre [ʃɛvr]

4

1

vacca- ‘cow’

5

vache [vaʃ ]

3

2

As phonetic erosion and the loss of phonological bulk proceeds, at the rate of about 20% or more per millennium, as suggested by the data in Table 1, a language does not waste away. A study of the Latin > Spanish evolution of anatomical terms. Elerick (2016) addresses this issue. That investigation of phonetic erosion

179

180

Charles Elerick

also shows a rate of loss in the 20% per millennium range. In addition, it addresses larger dimensions of the evolving lexicon and reveals various mechanisms that effect the compensation for low bulk or continuing erosion, often with the previously mentioned right-sizing effect. Some illustrative data involving a natural lexical set taken from the study are displayed in Table 2. Most significantly the study illustrates how phonological bulk is maintained, stochastically, in a lexical sector. Table 2. Loss, compensation, and right-sizing in Latin > Spanish anatomical terms Classical Vulgar/Late Sp. Segments lost term Latin etymon reflex by erosion of Mod. Spanish term

Gained by (a) phonological processes, (b) morphological augment, (c) lexical replacement

Segments in Span. vs. corresponding Classical term

digitu‘finger’

digitu-

dedo

2

pede‘foot’

pede-

pie

2

cubitu‘elbow’

cubitu-

codo

2

nare‘nose’

*narīce-

nariz

1

b) 2

5 vs. 4

aure‘ear’

auricula-

oreja

3

b) 4

5 vs. 4

gena‘cheek’

maxilla-

mejilla 2

a) 1 c) 2

6 vs. 4

genu‘knee’

rotella-

rodilla 1

c) 3

6 vs. 4

14 gained

33 vs. 32

13 of 44 ± 17% per millennium

3.

4 vs. 6 a) 1

3 vs. 4 4 vs. 6

Phonological loss, compensation, and the efficiency constant

The principled Latin > Spanish data and analysis illustrate loss and compensation at the lexical level and offer a clear illustration of a linguistic principle of prime importance. This principle entails the balance, indeed tension, between complexity of expression and informatic content, referenced above as a core concept of information theory. The pioneering work was published by Shannon and Weaver (1949) and we also have the ground breaking lexico-statistic work of Zipf (1965), which originated in the same period. Their mutually reinforcing conclusions laid

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

the groundwork for the expectation of token-content balance. Their findings were elaborated with reference to language that they dealt with as a synchronic and static system and with no cross-language data. The heuristic value, and in fact the explanatory utility, of the insight laid out decades ago has been supported by such related studies as Bennet (1987), Frank & Jaeger (2008), Piantadosi, Tily & Gibson (2011), and Coupe et al. (2019). Givón (1991) offered a convincing articulation of the Quantity Principle, the essential proportionality between content and bulk that is consistent across languages. By extension, and important for the present paper, the quantity principle can be assumed to hold between any language and an evolved successor. Givón additionally pointed out that the quantity principle represents an instance of non-arbitrariness in language. Linguists have long noted the arbitrary relationship between form and sign to the point that they reflexively think of language as an arbitrary system. The non-arbitrariness that is an entailed assumption of the quantity principle is likewise implied by the findings presented in this paper. The Latin > Spanish anatomical terms data show that the principles of information theory are relevant for the study of language as an evolving system that is governed by a diachronic corollary to the informatic balance that has been reliably postulated for any synchronic state. A logical extension of the principle of proportionality of information and its representation as sign would be an efficiency constant, the diachronic expression of the quantity principle. The persistent and adaptive balance is observable in all levels of language organization. As the evidence for this principle of language is explored, we should keep in mind that the phonological reduction that may attend grammaticalization is embedded in this larger and fundamental property of language as an isostatic, i.e. always rebalancing, evolutionary system. The stochastic balance that prevails through the erosion and compensation cycle as it shapes and reshapes the lexicon, i.e. the efficiency constant, always a stochastic effect that shows variation from token to token, is evidenced by successor articulations at higher levels of organization. A set of equivalent expressions from Latin and two Romance daughters, Spanish and French, demonstrate the utility, and probable validity, of the working hypothesis. Note that since in Latin vowel quantity is distinctive, there is a moreic dimension, which is not factored into these comparisons of bulk expressed as segment counts. Also, the Latin laryngeal glide [h] could arguably be excluded from a segment count. (1) a.

Nōn sumus oblītī. (14 segments) ‘We have not forgotten.’ b. No hemos olvidado. (14) c. Nous n’avons pas oublié. [nu na vɔ̃ pa u bli je]. (14)

181

182

Charles Elerick

(2) a.

Quō vadis? (8) ‘where are you going?’ b. ¿Adónde vas? (9) c. Où est-ce que tu vas? [u ɛ skə ty va] (9)

(3) a.

Dīxit sē legisse lībrōs quōs mīsissēs. (33) ‘He said that he had read the books that you sent’. b. Dijo que había leído los libros que (tu) mandaste. (34(36)) c. Il a dit qu’il avait lu les livres que vous avez envoyés. [i la di ki la ve ly le li vrə kə vu za ve ɑ̃ wva je]. (34)

(4) a.

Hodiē mortuus est eius pater. (24) ‘His father died today.’ b. Hoy falleció su padre. (16) c. Son père est décédé aujourd’hui. [sɔ̃ pɛ re de se de o ӡur dɥi] (19)

(5) a.

Cūr subrīdēs? (11) ‘Why are you smiling’? b. ¿Por qué sonríes? (12) c. Pourquoi souris-tu? [pur kwa su ri ty] (12)

(6) a.

Nōn essēmus pōma. (14) ‘We do not eat apples b. No comemos manzanas (17) c. Nous ne mangeons pas de pommes. [nu nə mɑ̃ ӡɔ̃ pa də pɔm]. (15)

The global segment count for the six corresponding clauses shows an essential equivalence in the three languages, 104 segments for Latin, 102 for Spanish, and 103 for French and therefore these data stand as a significant demonstration of the adaptive persistence of the information-sign balance, i.e. the efficiency constant. In the foregoing data we see many instances of phonological reduction in lexical material, instances of right-sizing by lexical replacement, and on a larger scale, compensation and right-sizing by various morphosyntactic innovations. Instances of phonological loss at the lexical level are Spanish dijo < dixit (-2); que < quōs (-2); vas < vadis (-2); hoy < hodiē (-3) and French père < pater (-2); livres < librōs (-2); vas < vadis (-3) ; est < est (-2). We see compensation through lexical replacement in late Latin of edere by comedere and mandūcāre, specifically the reflexes of these bulkier verbs in the inflected forms comemos < comedimus and mangeons < mandūcāmus, both of which, not unexpectedly, show subsequent and ongoing phonological reduction. We also see lexical substitution with bulking effect in the replacement of pomum by manantia > Spanish manzana. Highly significant is the rebulking effect of the elaboration within French of aujourd’hui, which replaced, while incorporating, the eroded hui < Latin hodiē ‘today’. Also

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

note the replacement of the Latin cūr ‘why’, a form of low bulk which was the reduced successor of early Latin kwor, by bulkier innovated material in French, pourquoi, and in Spanish, por qué. Regarding the replacement in the Latin period of shorter forms by more robust successors, with consequences for Romance, see the extended discussion with ample documentation of Lüdtke (1974) and note in addition the observation by Bourciez (1946: 56) of the general tendency to replace shorter words with “une forme plus résistante”, ‘a stronger form’. There are additional and highly significant diachronic developments that serve the isostatic effect, that is, the efficiency constant, that are morphosyntactic. We can take another look at the French nous ne mangeons pas de pommes, 6c) above. We have noted the loss of bulk through phonological decay. However, the compensatory bulk that supports the efficiency constant results from syntactic innovations, i.e., subject-obligatory nous, the presence of the innovated discontinuous negative grammaticalized particle pas, and the obligatory partitive de. We should note that these multiple bulk-adding features may be understood at least in part as compensation for a somewhat higher phonological decay rate in French. The higher decay rate was occasioned, according to Wartburg (1934: 56) by Frankish-Gallo-Romance bilingualism, in particular the persisting effect of the Germanic strong stress accent on the phonology of nascent French. At any rate, there was operating a combination of phonological, lexical, and syntactic variables, and possibly other factors, that had far-reaching effect. And this illustration of the efficiency constant as a diachronic principle establishes its usefulness as a working hypothesis that motivates further exploration of right-sizing and specifically how this concept might explain phonological outcomes that attend grammaticalization.

4.

Accelerated and uncompensated loss as an aspect of right-sizing

The maintenance of stochastically constant efficiency and more generally the quantity principle entails, as illustrated, the right-sizing by compensation for low or lost phonological bulk. But there is a mirror-image process that also results in rightsizing. The other aspect of right-sizing involves the reduction of forms which for some reason have bulk beyond what is commensurate with their informatic value. This often involves the diachronic development of forms that originated as multimorphemic innovations, as innovations by composition. Consider in this light the familiar case of the origin and development of French même ‘self-same’ and its Spanish cognate mismo, used to indicate mild emphasis. The Romance forms have their origin in the Latin *metipsimus, an extravagantly synthesized emphatic form,

183

184

Charles Elerick

some uses of which foreshadow those of its reflexes in the Romance daughters. The loss of bulk, to wit, 10 > 3 in French and 10 > 5 in Spanish, shows the accelerated erosion that can result in such cases. Terms for numbers above 10 were innovated in pre-historic Latin. Those for 11–15 are of special interest here. The attested Latin forms ūndecim ‘eleven’, duodecim ‘twelve’, tredecim, ‘thirteen’, quattuordecim ‘fourteen’, and quīndecim ‘fifteen’ are each composed of two morphemes, the first meaning ‘one’, ‘two’ etc., and the second meaning ‘ten’. The reflexes of these forms, as a set, in the Romance daughters show the effect of accelerated reduction. Illustrative of Romance outcomes are the French and (Spanish) reflexes onze (once), douze (doce), treize (trece), quatorze (catorce). The 45 segments that constitute the Latin set are reduced to 20 in French and to 25 in Spanish. Thus, the evolution of this limited sector of the lexicon illustrates the loss of phonological bulk with patent rightsizing effect, specifically the reduction of morpho-phonological elaborations that exceeded their semantic yield. A notable instance of rapid phonological erosion is seen in the innovation in early modern Spanish of a new 2nd Pers. Sg pronoun usted ‘you (formal). It had its origin in the NP vuestra merced, ‘your mercy’, referring to a real or fictitious attribute of a person of higher rank and it became a form of address as a personal pronoun, supporting a plural extension, ustedes. The 13 segments in vuestra merced quickly, in a matter of decades, were reduced to just five in what Lapesa (1981: 392) included in an otherwise unspecified set of “casos especiales de desgaste” ‘special cases of erosion’, mentioning only high frequency as a probable factor. An illustrative example from the referential lexicon is Latin petroselīnum ‘parsley’, with its French reflex persil. The Latin form was borrowed from Greek, a noun + noun compound meaning ‘rock celery’. The five-syllable Latin form represented a phonological elaboration that put it out of range of a reasonable bulkto-content ratio and it is a good example of right-sizing by accelerated reduction. This case, like the other material just discussed, illustrates the motivated acceleration of phonological loss with right-sizing effect as a general diachronic outcome. We see in these examples a loss of bulk that is of a magnitude that is sometimes associated with grammaticalization in cases that do not, except for the marginal case of usted, involve that diachronic process. We see loss that is not compensated, as it is in the lexicon at large when loss results in underweight forms.

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

5.

Right-sizing extended to cases of grammaticalization

As discussed in the preceding section, right-sizing is a complex global diachronic process that is (1) responsive to the quantity principle and (2) effected by various means, sometimes by which bulk is added, or sometimes through the reduction of bulk by phonological erosion. Phonological reduction taken as a frequent outcome that accompanies or follows from grammaticalization is part of this diachronic process and should not be understood as phenomenon-specific. The discovery that the frequent and not-insignificant link between grammaticalization and phonological reduction is subordinate to the larger framework sheds new light on that aspect of language change. The concept of right-sizing turns out to be useful in explaining various phonological outcomes that are associated with grammaticalization. Here is preview is what that exploration entails. – – – –

Right-sizing can involve the selective reduction of phonological bulk within forms. The loss of bulk in grammaticalized forms may be minimal if such forms represent a good quantity fit, e.g., an innovated preposition that is already monosyllabic, without reduction, or no reduction beyond a nominal rate. The loss of bulk may proceed as an effect of grammaticalization in conjunction with other loss, general erosion, with the two being difficult to sort out. Grammatical forms can themselves, on occasion, be right-sized by adding bulk.

The selective reduction of the phonological contour of forms as they are grammaticalized or become part of an innovative construction has been long documented. These future forms in French (e.g. tiendrai ‘I will have’) and Spanish (e.g., tendré ‘I will have’) were new in Romance, built with the Latin elements tenēre ‘to have’ plus habeō ‘I have’. In what had been innovated as a periphrastic future construction, the forms of habēre, were rapidly grammaticalized. They assumed a deontic force which was then quickly bleached to a simple future. Thus, inflected future forms emerged, with a loss of about 40% of the initial phonological bulk. (Note that in both the Spanish and French forms the non-etymological /d/ is inserted, an example of compensation by phonological addition and the French form shows an added segment by diphthongization in the stem.) We know that the loss of phonological bulk was fast. We see the attested right-sized future forms, salvarai ‘I will come to the aid of ’ and non prindrai ‘I will not make (enter into)’ in the oath that Louis swore at Strasbourg in 842, a premier example of Romance, by then established as a language distinct from Latin. The preserved record in a MS. dating from about 1000 is presumably reflective of the original. The forms in question attest to rapid phonological reduction, significantly exceeding the 15–20% per millennium

185

186

Charles Elerick

as demonstrated by decay in the general vocabulary earlier in the paper. Also significant is the selective reduction of material originating in the habēre paradigm while the lexically significant bases lose little, and even show innovated segments. In a study of the origin of the related Conditional paradigm in Spanish, Bybee (2010) notes the disproportionate reduction of the material proceeding from habēre and flags that development as a not-unexpected concomitant of grammaticalization. It is instructive to look again at the history of French low double-digit terms, onze ‘eleven’ douze ‘twelve’ etc. which show accelerated reduction. When we compare the accelerated reduction in the French future forms and these number terms, we see a similar outcome. The accelerated reduction eliminates phonological material in a way that is highly selective. Analogously to the verb forms, the number terms lose dispensable phones but not the denominating material in the first syllable or the word-final signature material [dz > z]. We see right-sizing, and right-configuring as comparable diachronic outcomes associated both with forms that do and with forms that do not involve grammaticalization. The loss of bulk in grammaticalized forms may be minimal under some circumstances. Keeping in mind the importance of right-sizing as an influencing factor, a form that is consistent with the quantity principle with little or no reduction may maintain its bulk. Also, a continuing identity with a form with lexical meaning can retard reduction. An example is French malgré ‘despite’. This adversative marker is the grammaticalized reflex of Latin malum grātum ‘bad feeling’. It shows significant phonological erosion. But when we look at the French reflexes of the individual constituent words, mal ‘bad’ and gré ‘feeling’ which retain their lexical status we find the exact same extent of phonological erosion. Therefore, the erosion evident in malgré cannot be attributed to grammaticalization. Similarly, Latin salvum ‘safe’ has a lexical reflex sauf ‘safe’ and a grammaticalized reflex sauf ‘save for; except’, both showing the same phonological reduction. The conclusion is clear. Phonological erosion at the nominal rate of 15–20% results here in grammaticalized forms that are of a right, or reasonable, size, i.e., not multi-syllabic, with respect to their grammatical semantic function. A comparison with a very different reduction outcome is instructive. Fr. il i a ‘there is, there are’ is the reflex of Latin illud ibi habet ‘it has there’. The difference in the amount of erosion seen in this material vs. that seen in the case of malgré is a function of how much reduction was needed for right-sizing, a lot in the one case, but relatively little in the latter. The question of what constitutes “right-sizing” ultimately requires additional research and cross-linguistic comparison. But, for example, we can be reasonably sure that if we look at scores of languages representing many families, prepositions expressing a comitative function, ‘with’, will consist of one or two syllables in virtually all of them.

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

A complex and highly interesting case of grammaticalization with no significant loss of bulk is seen in the succession of Latin middle voice verbs to their successors, which cannot be understood as direct reflexes, in the Romance daughters. In Latin, there are many middle voice verbs that show the same morphology that is used to form the Passive of active verbs, e.g., vereor ‘I fear’, cf. amor ‘I am loved’. In the Romance daughters the middle voice employs the morpho-syntax of reflexive constructions. Some good examples of the Latin-Romance correspondences are seen in Table 3. The morphological material that marks the verb as middle is underlined. Table 3. Latin and Romance middle voice correspondences Latin

Italian

French

Spanish

expergiscitur ‘he wakes up’

si sveglia

Il se réveille

se despierta

reminīscitur ‘he remembers’

si ricordi

Il se souvient

se acuerda

The replacement of the middle voice morpho-syntactic regime of Latin by the clitic, pseudo-reflexive construction in the Romance daughters occurred as a scantily attested macro-evolutionary and axial development. Meyer-Lübke (1926:317f ) recognized the importance of this development and included it in his discussion of “problemas”, by which he meant “challenges”, in the history of Romance. It has not been convincingly explained though see relevant material in Miller (2010). But the evolution of the middle voice can be engaged as a non-obvious case of grammaticalization that is highly significant in itself and which also provides a special illustration of the principle of right-sizing. Data are available that foreshadow what happened widely and with consequence presumably in late Latin and manifestly in Romance. At the beginning of the 1st cent. CE The Roman poet Ovid wrote: (7) a.

…laxantur corpora rūgīs … (Ars Amatoria 3.73) ‘…bodies sag with wrinkles …’

The unaccusative expresses as a middle voice verb with passive morphology. A century later the Roman poet Juvenal wrote: (8) a.

…trēs rūgae subeant et sē cutis arida laxet … (Sat. VI. 143) ‘[just let] three wrinkles show up and [let] her dry skin start to sag …’

In (8) the middle voice of the same verb, laxō, laxāre, expresses as a pseudoreflexive. These contrasting cases forecast a defining development in Romance. What are the implications for grammaticalization, right-sizing, and the efficiency constant? Taking the last question first, we notice that a pseudo-reflexive replacement involving sē for a ‑ur middle voice construction is bulk-neutral. We

187

188

Charles Elerick

are never surprised when loss and compensation balance each other. Next, can we understand the extension of sē, and other reflexive forms to a new function in middle voice constructions as grammaticalization? Yes, we can as the continuing discussion will show. And it is an interesting case of movement on the grammaticalization cline, discussed early by Hopper and Traugott (1993). Reflexive pronouns in Latin, including the special instance in the verse from Juvenal, are clause constitutes, as evidenced by the occurrence of sē on the periphery of the clause, at distance from the governing verb. The syntax of the reflexes of such pronouns in Romance is highly restrictive, consistent with their grammaticalization and incorporation into the Romance clitic regime. Semantic bleaching proceeds with the syntactic change, probably as an effect of the syntactic change itself. Specifically, true reflexives have a referent, with a Patient Ɵ-role, and they are under identity with the subject of the verb they are in construction with. Such a construction is like any other clause that has a transitive verb with a Direct Object. The pseudo-reflexives that occur in middle voice constructions in Romance are under identity and in agreement with the subject of the verb but they express no Ɵ-role and in addition have only an incidental and redundant referent. We are looking at an especially interesting case of language change, a mode of grammaticalization, that simply entails the disengagement of a Ɵ-role assignment. This has the effect of stranding the continued pronominal form, with the observed semantic and morpho-syntactic consequences. The Roman poet Martial supplies a late 1st cent. example of the Latin predecessor morpho-syntax. (9) a.

…praeceps sanguineā dum sē rotat ursus harenā. (Epig. I.XI. 1) ‘…while the frenzied bear turns around in the bloody sand’.

In this clause, sē is on the cusp between a Patient reflexive and a non-Patient form occurring in a middle voice construction. It is significant that there are equivalent middle voice verbs in at least two Romance daughters, Italian si gira, ‘he turns around’ and Spanish se da vuelta ‘he turns around’ As to the right-sizing issue, here we have another example of a diachronic size check that results in no loss in Italian or Spanish and in French there is little except for the loss of a vowel in liaison, mirroring what happens with true reflexives, as well as the phonological reduction in 1st and 2nd pers. plural forms, which reduction also tracks that seen in their corresponding forms when used as true reflexives. Given the phonological stability of the key forms, it has been easy to overlook the true extent of the morpho-syntactic evolution, including grammaticalization, that is entailed in this aspect of the history of Romance. Finally, as noted in the case of malgré and sauf above, we must assume that the morphosyntax of true reflexives and the clitic regimes that emerge in the various daughters support the formal characteristics of middle voice constructions.

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

The third issue in this continuing exploration of right-sizing and grammaticalization involves the interaction of loss that is plausibly connected with grammaticalization as de-semanticization and loss that proceeds as a reflection of the general tendency for phonological decay. The subsequent development of the reconstructed pre-Germanic forms for the number words eleven and twelve, following Voyles (1987), illustrates how general phonological reduction synergizes with other developments. It is not necessary to mention that the data displayed in Table 4 is also interesting in relation to the origin and evolution of corresponding forms in Romance as discussed earlier. Note that the innovated forms reflect a counting practice that devised a way to move past ten fingers. Table 4. The origin and evolution of West Germanic terms for 11 and 12 *oin-lip-ants one-leave-ing

*oin-lip one-leave

*duό-lip two-leave

*duό-lip two-leave

OE endlefen

OHG einlif

OE twelf

OHG zwelif

ModE eleven

NHG elf

ModE twelve

NHG zwölf.

This data shows how general decay and right-sizing by reduction may further interact with reduction that tends to be promoted by the loss of the referential semantics of constituent lexical material, as often seen in grammaticalization. This reinforces the claim of this paper that all reduction should be investigated with the knowledge that there are complex interactive factors in the loss of phonological bulk. The fourth effect of right-sizing and the operation of the quantity principle is the rebulking by augmentation of a grammatical form that has lost bulk to the extent that it is underweight and has lost viability. This is akin to yet different from the replacement of underweight non-lexical forms such as the pale Latin demonstratives is, ea, id ‘this’ by more robust forms as discussed by Palmer (1954) and subsequently by Wanner (1987). For background, we see augmentation frequently in lexical forms as evidenced above in the data on Latin > Spanish anatomical terms. An additional example is Latin semita- ‘path’ > Sp. senda which itself has been displaced in contemporary Spanish by the morphologically rebulked sendero. In grammatical forms, such augmentation is less common but the fact that it does happen demonstrates the principled nature of right-sizing. An illustration of reduction and subsequent right-sizing of a non-lexical form by augmentation retraces the history of Fr. avec ‘with’. The cycle starts with Latin apud ‘at, near, in the company of ’, which was continued in French and Catalan, but in no other Romance daughter. In French, it displaced the Latin preposition cum ‘with’ to indicate both accompaniment and instrument while also retaining the earlier meaning ‘at the home of ’ or ‘in the company of ”. The attestation of the evolution of apud and evolving forms in Gallo-Romance and Old French is extensive enough to support

189

190

Charles Elerick

a confident account of eventual outcomes while not supporting an account that would satisfy Neo-grammarian expectations for control of phonological specifics. Returning to the Oath of Strasbourg (842), we see, in the MS from about 1000, the effect of generic phonological reduction, apud > ab. (10) a.

…et ab Ludher nul plaid nunquam prindrai… ‘… and with Lothair never will I enter into any arrangement …’

The phonological reflexes of apud in OF also include such orthographic representations as od, ad, ot, or o and these forms, used with the meanings ‘with’ (instrumental), ‘at, in the presence of ’ and ‘with’ (comitative), are variously attested throughout the OF period as for example in the 9th cent. text that recalls the martyrdom of St. Eulalia. (11) a.

…ad une spede… ‘…with a sword …’,

In two early OF texts of the life of St. Leger we see variants attesting to the survival of reduced forms with the meaning ‘in the presence of ’. (12) a. … il nos aiud ob ciel senior… b. … il nos auit ot cel seignor… ‘…he helps us (intercedes for us) with the heavenly Lord …’,

And in the mid-12th century text of the Chason de Roland we see (13) a.

Ensembl’ od lui Rollant e Oliver. ‘Together with him (were) Roland and Oliver.’

But in the 12th century, in the Chanson Toile, we see (14) a.

A trente dames que avuec moi menrai. ‘And thirty dames whom I shall bring with me’.

Subsequently, the innovated avec succeeds as the sole form that expresses instrumental and comitative ‘with’. The augmented, right-sized avec shows, according to the standard theory, the addition of material that continues Latin hoc ‘this’, that expands ab ~ av, the remnant reflex of apud, which we have seen attested. See Voretzsch (1918: 91) for discussion. Quite interestingly, a parallel grammatical form, the OF reflex of Latin sine ‘without’, had lost viability through erosion and was right-sized, not by augment, but through the replacement by a bulkier alternative. The successor form, sans, is the grammaticalized and reduced (right-sized) form of the OF reflex of Latin absēns ‘away from’. And at about the same time that avec emerged, the phonologically weak o, ob, etc. used in the sense of “accompaniment” yielded to an innovation that follows the common pattern of grammaticalization. The OF noun (en) chies(e) < Lat.

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

(in) casā ‘house’ was appropriated as a bulkier substitute, and with rapid desemanticization, was recast as a preposition, yielding the Mod. Fr. chez. ‘at the house of ’. The loss of phonological bulk we see in (en) chies(e) > chez resulted in several segments resolving as just two, and we are reminded of an essential question addressed in this paper. How much of the loss reflects generic phonological decay and how much should be attributed to reduction and right-sizing that should be attributed to grammaticalization? This is one more example of reduction associated with grammaticalization that cannot be treated as isolated from the larger erosion regime. In the various cases of phonological right-sizing associated with grammaticalization or grammatical forms, we have seen cases of accelerated reduction, selective reduction, reduction that is simply consistent with the otherwise expected decay rate, and rebulking, either by morphological augmentation or replacement by more robust lexical material. The standard assumption regarding phonological reduction as a typical and expected concomitant of grammaticalization must be rephrased as follows. Grammaticalization often occasions phonological erosion but such reduction is a function of the larger imperative for right sizing. Rightsizing, i.e., the diachronic response to the quantity principle, can result in either reduction, which may vary as to its extent, or augmentation, which occurs when motivated by shape or circumstance.

6.

Right-sizing as a complex adaptive process with stochastic effect

Phonological decay, right-sizing in its various aspects, as explored in this paper and including both loss and gain of bulk, and the imperative of the efficiency constant produce a diachronic succession of principled linguistic states which are macro-expressions of the quantity principle. If it is true that any language, as it changes, produces such a non-arbitrary and foreordained result, at a stochastic level, how can we understand and explain this macro-diachronic outcome as an unconscious, adaptive process that is non-teleological in its effectuation? A short and preliminary answer is that language change moves on the current of the infinitesimal numbers of speech transactions of its speakers. A good place to start a broader exploration is the remark by Lehmann in his 1985 paper, cited above, that “language activity is creative” attributed by him as causal to the process of grammaticalization. He goes so far as to say that speakers “want to change.” Lehmann emphasized the role of speakers, and by implication, speech acts in linguistic innovation and change. Bernd, Claudi and Hünnemeyer (1991) have engaged and advanced the discussion regarding the fundamentally creative nature of grammaticalization and this dimension of language change. It

191

192

Charles Elerick

is a complex process by which new ways of using lexical, syntactic, and semantic resources to express complex propositions and relationships are devised, i.e., created. It is a collective process that involves the unconscious cognitive activity of individual speakers, that is exploratory, experimental, and adaptive, and which produces accumulating micro-changes. The literature of grammaticalization is replete with examples of such linguistic creativity. In the case of adjustments in phonological bulk that may accompany grammaticalization, we would not want to say that there is “creativity” involved. It proceeds rather as a matter of experimentation and the progressive result of “successful” experiments. What does this mean? Remembering that change in phonological bulk and right-sizing is a general process which happens to frequently occur as a concomitant of grammaticalization, the experimentation that takes place can involve phonological erosion, morphological enlargement or truncation, or lexical substitution. In any such case, essentially what happens is a speaker experiments, unconsciously, with a variant and such an articulated experimental instance may turn out to be viable, be imitated by other speakers, and thus propagate. There are cases of right-sizing by reduction that are transparent and that concretely illustrate the process. This happens when a new term of a certain bulk quickly increases in frequency and is replaced almost immediately and completely by a clipped form. A example from 20th century English is the replacement of condominium by condo. Examples of this type of transparent right-sizing by lexical clipping abound. In the larger scheme of erosion and right-sizing, every single case of loss, or occasional addition, of segmental material or the morphological augment seen in data presented above also involves experimental articulations that succeeded because they were imitated by other speakers and very importantly, were incorporated into the grammars of new speakers. Experimentation, imitation, and propagation is a familiar formula. It emphasizes the role of speech and speakers. It recalls the groundbreaking work of Labov (1972) who, importantly, identified speaker prestige as a factor in accelerating the spread of an innovation and whose concepts are aptly enlisted to explain the macrodynamic that drives right-sizing. Gillièron (1918) had prosecuted these essential insights in his careful exposition of creative linguistic evolution as a counter-point to the then-prevailing mechanistic concepts of the Neo-grammarians, with great explanatory value. What induces speakers to engage in speech behavior that ultimately leads to alterations of a linguistic system? The loss of phonological bulk through erosion is essentially the result of the operation of the law of least effort and recalling information theory, the urge to make good use of demand on the channel. Erosion has its origin in speech acts in less formal registers and unguarded articulation.

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

Such erosion resulting in reductive restructuring is not a controversial concept. And the impetus for such reduction and critically, the tolerance of a more radical reduction of sign is enhanced when, as in grammaticalization, higher predictability lowers the semasiological value and therefore the need for, or in fact tolerance of, tediously excessive bulk. The results are easy to find and understand when we have historical data to inspect. With respect to compensation, what induces speakers to experiment transactionally in ways that might, and we know do, generate the rebulking that supports the efficiency constant? We also have a lot of data that show rebulking outcomes but it is much harder to explain the rebulking process, with all its complexity, than it is to explain phonological reduction. Grice (1975), looking at the linguistic significance of what happens at the level of the speech act, points out that speakers observe a Cooperative Principle, one aspect of which is to consciously or consciously make an effort to make their message recoverable and intelligible. Grice furthermore posits a principle of Right Quantity as an aspect of his Cooperative Principle and this entails speakers’ accommodation of what they perceive of their interlocutors’ need for informatic specificity and bulk. Their perceived need to provide viable forms undoubtedly contributes to the propagation of more robust forms. So, a speech situation may involve a speaker’s motivated choice of a more robust form over a form of less bulk. For example, if a Latin speaker, referring to their grandfather used a form of avus, the ablative sg. say [áwō], the acoustic prominence and therefore viability, would be quite low. But if that speaker made recourse to the available alternative *aviolō, ‘dear grandfather’, a diminutive of affection, intelligibility would have been enhanced. In addition to the issue of acoustic viability, the larger form with its additional bulk, adding one or two syllables and 20 to 40 centi-seconds, provides the listener a little additional processing time. We know that the alternate form in fact represented a compelling preference because its reflex is seen in the Spanish abuelo ‘grandfather, a connotatively neutral form, and the shorter form went extinct. Examples of such lexical rebulking outcomes also abound and many instances have been cited in the foregoing discussion. What else induces speakers to use speech acts to unconsciously instigate changes that may result in the addition of phonological bulk? There are in fact, many additional motivations to experiment with bulk-adding speech. Such cases seem to reflect a certain preference for the bulkier of two options for expressing essentially the same content. Simply put, humans like to make noise. They like to command the speech channel for a few extra seconds. They like to sound like they have more to say than they in fact do. Most often these urges are met by simply adding more filler speech in a conversational intervention, but the urge also expresses as the choice of a longer word over a shorter one. Compensation by morpho-syntactic innovation responds to these same motives and inclinations.

193

194

Charles Elerick

An example is the innovation in French of est-ce que ‘is it that’ preceding what is stated or asked in a subordinate clause. It is regularly used as semantically empty phonologically bulk and with no suggestion of introducing a clarification. Some articulatory extensions may be employed in the effect of not appearing to be too monosyllabic and therefore not communicatively forthcoming. A good example in contemporary English is the rapid increase in shortened clausal responses to questions, “I do” or “I have not” in preference to “yes” or “no.” Given these various motivations, if a speaker of English says utilize in preference to use or a French speaker says visualizer in preference to voir, we cannot be sure which urge or what combination of urges to choose bulk over efficiency is in play. A not-unexpected response in some quarters to these examples will be to dismiss them as random, anecdotal, and even trivial. The fact is, isostatic evolution goes forward solely on the basis of random and “trivial” occurrences. Multiplied by orders of magnitude, they result in the rebulking that sustains the efficiency constant. In the same way, the individual, often stepwise, reductions of form that represent random wear and tear or respond to the quantity principle, including those associated with grammaticalization, might also appear trivial. Speakers of a language as agents of adaptive linguistic evolution unconsciously effect the changes that power principled linguistic evolution and such evolution is principled even as the process itself appears to be chaotic. Finally, we must note that consistent with the nature of a complex adaptive system, the loss of bulk, aggregated, and the aggregated compensation proceed as two distinct, untidy processes that produce results that intersect in a stochastic universe and maintain an evolving and principled system.

7.

Conclusions

The clear implication of the data presented in this paper is that in any instance of grammaticalization that has been or might be investigated, the question of accompanying change in form, typically reduction, should be considered within the framework of the general process of phonological erosion, the global principles of diachronic right-sizing and the quantity principle with its diachronic analogue, the efficiency constant. At the same time, the extensive illustration of how such a complex adaptive process apparently works, based on data drawn principally from the Latin > Romance continuation, must be considered provisional while motivating further investigation. The initial demonstration provided in the present paper suggests that effort invested in testing its assumptions and precepts using data from other languages and language families will yield significant results.

Chapter 7. New perspectives on phonological erosion as an aspect of grammaticalization

The present study should also serve to reinforce the understanding that speakers of a language, as unconscious participants in a complex process, bring about language change. The representation of a historical antecedent being replaced by its reflex, x > y, is convenient and efficient, but is a distorting abbreviation of the complex process that can only go forward and be meaningfully understood as one effected by those speaker-agents. On a more general level, the demonstration that language change, as systemic change, is messy even as it is principled represents a break from the stance that has dominated linguistics for over a century. Saussure insisted that linguistics was about describing languages as stable systems, at a fixed point in time, as a welldefined complex of formal and psychological interrelationships. In other words, he was both a structuralist and a positivist. Structuralist precepts and practitioners, Bloomfield and others, defined linguistics until after mid-century when generative modeling, in fact a sophisticated extension of structuralism, emerged with great vigor. The present paper makes a case for the view that linguistics, in the first instance, is about understanding how languages are built to respond to wear and tear, and other challenges, and to replace, on a continuing basis, one system, that is sufficiently stable but open, with a successor.

References Beckner, Clay, Blythe, Richard, Bybee, Joan L. Christiansen, Morton, H., Croft, William, Ellis, Nick C., Holland, John, Ke, Jinyun, Larsen-Freeman, Diane & Schoenemann, Tom. 2009. Language is a complex adaptive system. Position paper. Language Learning 59(Supplement 1): 1–27. Bennett, Charles H. 1987. Information, dissipation, and the definition of organization. In Emerging Syntheses in Science, David Pines (ed.) 215–233. Reading: Addison-Wesley. Bourciez, Édouard. 1946. Éléments de linguistique Romane, 4th edn. Paris: Librairie C. Klincksieck. Boye, Kasper & Harder, Peter. 2012. A usage-based theory of grammatical status and grammaticalization. Language 88(1):1–44. Bybee, Joan L. 2010. Language, Usage, and Cognition. Cambridge: CUP. Campbell, Lyle. 2001. What’s wrong with grammaticalization? Language Sciences 23: 113–161. Elerick, Charles. 2016. Lexical right-sizing in the evolution of Spanish: Understanding language as a complex adaptive system. Bulletin of Hispanic Studies 96(3): 183–204. Frank, Austin F. & Jaeger, Florian T. 2008. Speaking rationally: Uniform information density as an optimal strategy for language production. Proceedings of the Annual Meeting of the Cognitive Science Society 30. (30 December 2022). Gillièron, Jules. 1918. Généalogie des mots qui désignent l’abeille, d’apres l’Atlas linguistique de la France. Paris: Champion.

195

196

Charles Elerick

Givón, Talmy. 1991. Isomorphism in the grammatical code: Cognitive and biological considerations. Studies in Language 15(1): 85–114. Grice, H. Paul. 1975. Logic and conversation. In Syntax and Semantics, Vol. 3: Speech Acts, Peter Cole & Jerry L. Morgan (eds), 41–58. New York NY: Academic Press. Heine, Bernd, Claudi, Ulrike & Hünnemeyer, Friederike. 1991. Grammaticalization. Chicago IL: The University of Chicago Press Heine, Bernd & Reh, Mechthild. 1984. Grammaticalization and Reanalysis in African Languages. Hamburg: Helmut Buske. Hockett, Charles F. 1968. The State of the Art. The Hague: Mouton. Hopper, Paul F. & Traugott, Elizabeth Closs. 1993[2003]. Grammaticalization. Cambridge: CUP. Joseph, Brian. 2015. Multiple sources and multiple causes multiply explored. In On Multiple Source Constructions in Language Change [Benjamins Current Topics 79], Hendrik De Smet, Lobke Ghesquiére & Freek Van de Velde (eds), 205–221. Amsterdam: John Benjamins. Labov, William. 1972. Sociolinguistic Patterns. Philadelphia PA: University of Pennsylvania Press. Lapesa, Rafael. 1981. Historia de la lengua española, 9th edn. Madrid: Gredos. Lehmann, Christian. 1985. Grammaticalization: Synchronic variation and diachronic change. Lingua e Stile 20: 303–318. Lüdtke, Helmut. 1974. Historia del léxico románico. Madrid: Gredos. Meyer-Lübke, Wilhelm. 1926. Introducción a la lingüística románica. Madrid: Revista de filología española. Miller, Gary D. 2010. Language Change and Linguistic Theory, Vol. II: Morphological, Syntactic, and Typological Change. Oxford: OUP. Palmer, Leonard Robert. 1954. The Latin Language. London: Faber and Faber. Piantadosi, Steven T., Tily, Harry & Gibson, Edward. 2011. Word lengths are optimized for efficient communication. PNAS 108(9): 3526–3529. Schiering, René. 2010. Reconsidering erosion in grammaticalization. In Grammaticalization. Current Views and Issues [Studies in Language Companion Series 119], Katerina Stathi, Elke Gehweiler & Ekkehard König (eds), 73–100. Amsterdam: John Benjamins. Shannon, Claude & Weaver, Warren. 1949. The Mathematical Theory of Communication. Urbana IL: University of Illinois Press. Voretzsch, Karl. 1918. Altfranzösische Sprache, 5th edn. Halle: Max Niemeyer. Voyles, Joseph. 1987. The cardinal numerals in pre-and Proto-Germanic. Journal of English and Germanic Philology 86: 487–499. Wanner, Dieter. 1987. The Development of Romance Clitic Pronouns. Berlin: Mouton de Gruyter. von Wartburg, Walter. 1934. Évolutión et structure de la langue française. Berlin: Teubner. Zipf, George K. 1965. The Psycho-Biology of Language. Cambridge MA: The MIT Press.

section iii

Interactive contexts

chapter 8

On the development of discourse markers from elliptical structures Seongha Rhee

Faculty of Liberal Arts, Mahidol University | Hankuk University of Foreign Studies

Korean has a group of DMs that developed from sentential fragments through ellipsis (EDMs), i.e., those that were morpho-syntactically incapable of standing alone. They carry diverse functions across discourse domains. In particular, they tend to mark the speaker’s affect, and this development is largely due to their lexical and structural characteristics. A diachronic analysis shows that their development is closely tied to semantic generalization, scope extension, and movement to non-licensed positions. The functions of EDMs are often correlated with their characteristic prosody rather than with their forms. All EDMs involve either deictic expressions or interrogatives which are related to indexicality and vagueness that play a significant role in the development of EDM functions. Keywords: discourse marker, ellipsis, prosody, periphery, stance, indexicality, vagueness

1.

Introduction

Discourse markers (DMs) develop from various source lexemes and constructions, and carry diverse and often multiple functions, largely depending on the contexts in which they occur (Fischer 2006 and works therein; Koo 2012, 2018; Heine et al. 2021: 64; Rhee & Koo 2021: 77) as well as the prosodic features that accompany them (Hancil 2013 and works in Hancil & Hirst 2013). It is notable that many DMs in common use are structurally incomplete, e.g., I know, You know, I see, Look, Sort of, Kind of I mean, I think, I guess, Let’s say, It’s like, Like, etc. in English; wokan ‘I think’, woxiang ‘I think’, etc. in Chinese, among others, which clearly show structural incompleteness from the absence of phrasal or clausal complements or of external arguments like the sentential subject. When linguistic forms are recruited as DMs, the structural characteristics of https://doi.org/10.1075/slcs.232.08rhe © 2023 John Benjamins Publishing Company

Chapter 8. On the development of discourse markers from elliptical structures

their sources may disappear, as shown in Aijmer (1996: 19), who observed with reference to the English openers well and now, “They can be analysed neither as elliptical structures nor as special parts of speech.” Korean has one group of DMs that developed from sentential fragments through ellipsis, i.e., those that were morpho-syntactically incapable of standing alone. This class of DMs will be referred to as elliptical discourse markers (EDMs henceforth). Even though EDMs present interesting aspects in terms of their structure and function, they have not yet received much attention. This paper intends to fill the research gap by (i) describing the functions of EDMs, (ii) analyzing their developmental paths, and (iii) discussing the theoretical importance of certain aspects of their development. This paper is organized in the following manner. Section 2 presents the scope of research and briefly introduces previous research on some EDMs. A few preliminary examples of EDMs are also provided. Section 3 illustrates major DM functions exemplified by EDMs. Section 4 presents the developmental paths of each EDM to identify the semantic and structural characteristics that contributed to the development of EDMs. Section 5 discusses some theoretical issues, such as significance of elliptical structures, DM characteristics, and source characteristics. Section 6 summarizes the findings and concludes the paper. The data for historical development has been taken from the 21st Century Sejong Corpus, a 200-million-word corpus, developed by the Korean Ministry of Culture and Tourism and the National Institute of the Korean Language. The corpus encompasses data from the 15th to the early 20th century (1446–1912). The contemporary data is taken from the Drama-Cinema Corpus, a 24-million-word corpus consisting of 7,424 drama, sitcom, and movie scenarios, compiled by Min Li while at Seoul National University (currently at Tsinghua University). Scenarios are dated from 1992 through 2015. The concordance program used in the data search is UNICONC, developed by Jinho Park.1

2.

Preliminaries: EDMs and previous studies

Korean is an agglutinating and polysynthetic language with a large number of grammatical suffixes for nouns and verbs. In a sentence, these lexical forms occur with morphosyntactic markers, which signal, for example, the grammatical relations for nouns, or tense, aspect, modality, and mood for verbs. Grammatical markers are necessarily licensed by the semantic and syntactic properties of the 1. Special thanks go to the developers of these corpora and the search program for their generosity in granting their use for academic research.

199

200

Seongha Rhee

sentence in which they occur. For instance, if a noun occurs with an accusative case marker ‑((l)u)l, the noun phrase should be an argument of a transitive verb, and if a verb occurs with an adnominalizer ‑n/nun/(u)l, the form should be a modifier of a noun phrase that follows it (note that Korean is head-final). Therefore, if such forms occur without a transitive verb (as in (1)) or a noun phrase (as in (2)), they are syntactically defective, i.e., they are fragments (note: utterances that do not require morphemic gloss are given in English translation within square brackets): (1) (At a small food-stand in a local market, A asks her husband B to buy many different kinds of dishes from C and they begin to eat them. A eats so much that B is a bit worried.) B: [Hey.. you shouldn’t complain later that you have a stomachache. You eat that much and you don’t get a stomachache, then it means your stomach is not normal.] C: aikwu imsin-i-kwuman mwel… cal mek-e kulay-ya nwun intj pregnacy-be-mir dm much eat-end do.so-if eye ccakccaki an naw-a… unmatched.pair not come.out-end   ‘Well, she is pregnant, DM ( = no doubt)! Eat well. Only that way you will have a normal baby (lit., no baby with differently-sized eyes will come out).’ (2003 Drama Nolan sonswuken, Episode #121) (2) (A and B are art high school students. B turns around a corner and bumps into A. A’s palette falls onto B’s white slippers and stains them.) A: ilen!! coysongha-mta. takk-atuli-lkey-yo. dm be.sorry-dec clean-ben-prom-pol   ‘DM ( = O, no!) I’m sorry. I will wash (them) for you!’ B: (frowns looking down his slippers stained with colors.) (2006, Drama Kwung, Episode #1)

In (1), the EDM mwel, developed from mwe-l [what-acc], does not carry its original meaning ‘what’ nor does it function as a verbal theme argument. Instead, it is used to express C’s conviction about her statement (i.e. abnormal appetite is a sign of pregnancy) and disregard of any potential disagreement (“what else can you say?”). The preceding utterance is a mirative-marked sentence, literally ‘(it) is pregnancy – surprising!’ (note that argument omission is free and even preferred in speech in Korean as long as it does not hinder correct interpretation). The utterance following mwel is an imperative sentence, meaning ‘Eat well so you

Chapter 8. On the development of discourse markers from elliptical structures

will have a normal baby.’2 Therefore, mwel ‘what’, despite its accusative marker ‑l, is a stand-alone fragment with neither a governing verb nor a specified sentential subject, functioning as an EDM of emphasis. Similarly, (2) involves ilen, developed from ileh-n [be.like.this-adn], with the literal meaning ‘this kind of ’, which would normally be followed by a noun, by virtue of containing the adnominalizer ‑n. However, as is evident from the excerpt, ilen occurs as a stand-alone form with no noun following it. Even semantically, the original meaning ‘this kind of ’ does not fit in the context. Instead, it is used to express A’s surprise about her mistake and embarrassment for the damage it caused to B. The mirative function of the EDM ilen resembles that of an interjection. Korean has about 200 DMs, many of which are syntactically defective, largely because of the syntactic idiosyncrasy of relatively free argument omission in the language. However, the degrees of perceived ellipsis are variable and the present paper focuses on the DMs with the highest level of perceived ellipsis, thus unambiguously ellipsis-based, i.e., those marked with the adnominalizer ‑n or the accusative case-marker ‑l. Therefore, the bare interrogative pronoun mwe ‘what’ is not included in the EDMs, whereas the accusative marked counterpart mwel ‘what-acc’ is. There are about six EDMs in this category, as shown in (3) and (4), with their source constructions (note that weynkel in (4b) involves both an adnominalizer and accusative case-marker; see Section 4 for a more detailed account of the etymology): (3)

ADN-marked EDMs

(4)

ACC-marked EDMs

a. ilen b. celen c. mwusun d. weyn

< < < mues > mwues > mwe]. Mwel is marked with the accusative case-marker ‑l, and thus is expected to occur in the object position licensed by a transitive verb. It is a reduced form of mwuesul, which is still in use in PDK, especially in formal registers. In historical data, we see many instances of the accusative-marked predecessor forms in the 17th century, e.g., muses-ul kitAlo-li-o [what-acc wait-fut-q] ‘what should (I) wait for?’ (1617, Tongkwuksinsok Yel 6:71b), muses-ul wihAye [what-acc for] ‘for what; why’ (1677, Pakthongsa enhay II: 52a), mwues-Al po-si-liiska [what-acc see-hon-fut-q] ‘what would you see?’, (1618–1667(?), Sekwungilki 5a), etc. It is interesting to note that mwuesul (and its variants) tends to occur increasingly in argumentative contexts from the 18th century, as shown in the following:

Chapter 8. On the development of discourse markers from elliptical structures

(30) (A shopper at an archery shop complains about the quality of bows with an intention to buy a bow at a lower price, and the shop-keeper argues back) [You nonsensically complained that this bow is of poor quality, and] cyelen hwal-ul sto mwues-ul saonap-ta hA-nA-nta that.kind.of bow-acc again what-acc be.poor-comp say-pres-q ‘what are you complaining about that kind of (premium quality) bow as being poor in quality?’ (1790, Mongenokeltay 6:13b)

This argumentative usage becomes popular in colloquial genres in the reduced form mwel, often collocating with the verb ha- ‘do’, i.e., mwel ha- ‘do what?’, to mean ‘what is the point of doing it?’ In this usage, mwel ‘what’ tends to become superfluous (note that the theme argument role of mwuesul in the example is weak as well). This weakening process, in sinsosel novels, further led the form to have a wider scope, resembling an adverb, i.e., from the verbal argument to the sentential adverb, carrying the ‘why’ meaning, as in (31a), and further in challenging contexts, as in (31b), carrying the EDM function: (31) a.

(A visits B with a large package of fresh persimmons, and B gladly divides them into small bundles so they can be equitably distributed among family members. Looking at B, A says:) [Don’t bother.] ku-skacis-kes-ul mwue-l kulihA-o that-pej-thing-acc what-acc do.so-q ‘Why do you do so with such petty things?’ [Just keep them all together and you help yourself.] (1913/1914, Kumkwukhwa 017) b. (The unemployed husband asks for lunch complaining that he is hungry. The wife is annoyed.) mwue-l enusAy sicyanghA-ta-n mal-i-yo what-acc already be.hungry-comp-adn word-be-end ‘What (are you talking about)? How are you saying that you are hungry already?’ [You’d better get used to going hungry.] (1915, Kumkangmwun 38)

In (31a), the speaker refers to the expensive fruit with a pejorative form as a politeness strategy as it is what she herself brought. It is noteworthy that the theme of the main verb kuliha- ‘do so’ is kuskaciskes ‘that insignificant thing’ (note that it is marked by the accusative ‑ul) and thus mwue ‘what’, though marked by the accusative ‑l, is not the true theme argument. In strictly formal terms, these two accusative-marked nominals may be interpreted as appositives. But a more intuitively appealing interpretation is that the function of mwuel here is ‘why’ rather than ‘what’. In (31b), the annoyed speaker challenges her husband’s asking for food. It is noteworthy that in the sentence there is no transitive verb to take mwuel as its theme argument. Thus, mwuel is syntactically superfluous and its meaning

221

222

Seongha Rhee

is closer to ‘what are you talking about?’ or ‘why are you saying that?’ The syntactic independence and the metatextual meaning that occurred around the turn of the 20th century identify mwuel (and its further reduced variant mwel) as an EDM of challenge. The series of semantic and syntactic changes described above suggests that the function of marking challenge/protest of the EDM mwel developed from its scope extension from an argument role to a non-argument role and its concomitant semantic shift from ‘what’ to ‘why’.

5.

Discussion

In the preceding exposition we have looked at the historical patterns of the lexemes that eventually developed into EDMs. Based on the description, we will discuss in this section a few select issues that bear significance in the study of DM development.

5.1 Significance of elliptical structure We restricted our attention to the DMs that developed from elliptical structures. A number of grammaticalization studies in Korean, e.g., Sohn 1995; Park & Sohn 2002; Jung 2001; Rhee 2012, 2020a; Koo & Rhee 2001; Koo 2005, among others) have noted that many grammatical markers, notably sentence-enders, arose from the process known as ‘insubordination’ (Evans 2007, 2009; Evans & Watanabe 2016). Rhee (2020a) notes that from the inventory of 381 sentence-enders in contemporary Korean as listed in Kim (2001: 147–151) a vast number of them are the product of insubordination. This process, according to Rhee (2020a), is the strategic use of ellipsis, whereby the speaker invites the addressee to pragmatic inferences. Strategic ellipsis gives the benefit to the speaker, e.g., economy of verbalization, no burden of commitment to the elided detail, and enriched interpretation from the addressee (Rhee 2020a: 22). For instance, many DMs of agreement, especially those used for back-channeling, historically originated from discourse segments in elliptical structures. What has been elided is strategically withheld by the speaker to indicate that it is so obvious that it does not need to be explicitly expressed, i.e., the high degree of its being obvious warrants its ellipsis and at the same time serves as an endorsement of the truthfulness or firmness of what the previous speaker has just said, thus, agreement (Rhee 2015: 20). A similar observation can be made with respect to EDMs. Let us take the Example (9), repeated here as (32) for convenience:

Chapter 8. On the development of discourse markers from elliptical structures

(32) ( = (7)) (A, who fell in love with B, calls B on the phone, wishing to see her,) B: [Excuse me? You mean me? Why? No, I’m not particularly busy, but … Where are you? Near my home? Why?] A: [I want to go out on a date with you.] B: ney? teyithu-yo? ani weyn… pardon date-pol.end dm dm   ‘Pardon me? You mean, a date? DM ( = wait a minute) DM ( = I’m surprised) …’ A: [Didn’t I ask you yesterday?] (2007, Sitcom Kechimepsi haikhik Episode #144)

Speaker B says weyn (lit. ‘what kind of ’) and does not continue the utterance to complete it but elides what is expected to follow with some elongated decay (note the elliptical notation). In this situation, the addressee will have the following line of pragmatic inferencing: (i) the speaker ended her utterance with a modifier, so I can wait for her to complete the utterance; (ii) the intonation contour and the decay show that the utterance will not continue; (iii) now the discourse reached the TRP; (iv) from the general cooccurrence pattern and the context, the noun most likely to be modified by weyn but not explicitly expressed may be teyithu ‘date’; (v) the speaker must have reasons not to complete the seemingly incomplete sentence; (vi) she must have been unable to utter the word and others following it because she was surprised very much; and (vii) weyn is the marker of surprise, rather than a modifier ‘what kind of ’, which is not followed by a modified noun after all. This line of inferencing can easily lead to the functional reinterpretation of weyn from an interrogative modifier to a mirative EDM. This pattern may be applicable to most EDMs with minor modification in detail. The significance of the role of ellipsis in the development of EDMs is that not only the lexical sources but also the structural patterns play an important role. For instance, as discussed above, signaling surprise, one of the most prominent EDM functions, is derived from structural ellipsis, which leads the addressee to infer that the speaker is unable to complete the utterance by being overwhelmed by the event or state of affairs. This pragmatic interpretation is often strengthened by their other occurrence patterns such as the repetition of the EDMs (e.g., ilen, ilen! celen celen! etc., cf. (5), (10), (16), (17) above), stammering (e.g., mwu..mwusun, i..ilen, etc., cf. (12) above), frequent cooccurrence with an interjection (e.g., aikwu, eykwu, etc., cf. (1) above), use of exclamation marks in case of written texts (cf. (2), (15) above), among others.

223

224

Seongha Rhee

5.2 DM characteristics A number of defining characteristics of DMs have been proposed by many DM researchers. However, most researchers also note that the properties are not definitional but are only prototypical (Heine 2013: 1209). Commonly cited properties include: syntactic independence, formulaicity, prosodic separation, non-truthconditional, non-restrictive and procedural meaning, non-compositionality, formal shortness, omissibility, indexicality, multifunctionality, positional variability, etc. (Brinton 1996; Heine 2013: 1209; Heine et al. 2021: 10–11, among others). Based on these, we will review the properties of EDMs, in terms of morphosyntax, semantics, phonology, and pragmatics. 5.2.1 Morphosyntax In morphosyntactic terms, DMs tend to be syntactically independent, omissible, formulaic, and positionally variable (Hansen 1997; Tabor & Traugott 1998; Waltereit 2002; Barth-Weingarten & Couper-Kuhlen 2002; Heine 2013; Haselow 2014; Traugott 2014; Beeching & Detges 2014; Brinton 2017; Rhee & Koo 2021, among others). EDMs exhibit all such properties. Their syntactic freedom is particularly notable because they still bear the markers of syntactic dependence, i.e., the accusative case-marker or the adnominalizer. Syntactic independence of EDMs has been acquired largely through two processes: the segment’s movement and ellipsis. In the case of accusative-marked noun phrases, the source form moves out of the argument position into a nonargument position either by being extrapolated to the periphery or by occurring with a better qualified argument, whereby it is made superfluous or functionally inconspicuous. In the case of the latter, for example, in some expressions like mwe-l mekulke-l sa-ss-ta [something-acc food-acc buy-pst-dec] ‘(I) bought some food’, there are two potential theme arguments ‘something’ and ‘food’, both marked by the accusative case (note that in structural terms these two nominals may be appositives), but ‘food’ is more specific and informative than ‘something’. This can push mwe-l out of the argument structure, which, in turn, gives it syntactic freedom. Similarly, in the case of the adnominalizer-marked source forms, they may occur in a position where the modifier-modified relation is not prominent, such as when the two are separated by a clausal, or otherwise complex, modifier. For instance, in example (20c) above, the modifier ilen ‘this kind of ’ and the modified noun nom ‘fellow’ are separated by a complex modifier of nom ‘fellow’, i.e., ‘who will be bound by the police rope and take a nap at a police station’. Because of the intervening lengthy modifier, the modifier function of ilen ‘this kind of ’ becomes vague.

Chapter 8. On the development of discourse markers from elliptical structures

The other process whereby the EDM source forms gained independence is the use of explicit ellipsis, i.e., not completing the syntactically incomplete utterance by supplying the qualified verb (in the case of accusative-marked forms) or the modified noun (in the case of adnominalizer-marked forms). In addition to noncontinuance, the ellipsis can also be marked by a closing prosody (as can be indicated by an exclamation mark in case of written texts), as discussed in 5.1 above. EDMs are also omissible without loss in meaning at any significant level. This is closely related to the syntactic freedom EDMs acquired. Through the acquisition of syntactic freedom, EDMs have lost their functional significance at the argument-structure level (see 5.2.2 for more discussion). EDMs are highly formulaic, i.e., they are invariable and do not allow internal morphological modification (Heine et al. 2021: 11). This property is robust with EDMs because EDMs are relatively less complex in their formal makeup. They are highly unitized (i.e., ‘univerbated’ Lehmann 2015[1982]) and do not allow any internal modification. This is clear from the fact that accusative case-markers and adnominalizers are not periphrastic but morphological markers, and morphological markers are directly suffixed to their hosts. EDMs are also positionally variable. As a result of the syntactic freedom they acquired, they do not occur in the traditionally licensed positions. Not all EDMs freely occur at any position but their positions are closely tied to their functions. For instance, interjection-like EDMs such as ilen, celen, and weynkel tend to occur at the left-periphery, at utterance-initial positions. The EDMs that display the greatest level of positional freedom are those used for pause-filling, e.g., ilen and mwusun. Such positional freedom is a natural consequence of the fact that lexical search may be necessitated at any point in an utterance, even though they tend to occur at major phrasal junctures. Similarly, EDMs of the reluctance/hesitancemarking function, similar to those of pause-filling, also tend to occur at any point in an utterance because EDMs of such a function tend to occupy the position immediately before the expression the speaker has reservation to verbalize. 5.2.2 Semantics DMs carry non-restrictive and procedural meaning (Hansen 1997; Fraser 1999; Schourup 1999; Wilson 2011; Haselow 2015; Heine et al. 2021, among others). This semantic aspect is definitional of DMs because they are given such a functional status by virtue of their discourse-organizing functions rather than their contribution to the propositional meanings. The functions of DMs in the interaction-, topic-, information-, and interlocutor-management domains are all beyond the propositional levels, i.e., the discourse level. Their advancement into the discourse level is through the loss of the original lexical content meanings in specific contexts, i.e., desemanticization (Heine et al. 2021: 81).

225

226

Seongha Rhee

In the same way, EDMs carry non-restrictive and procedural meanings. By moving out of the argument structure their capability as a clause- or phrase-level modifier has significantly diminished and instead they have acquired the procedural meanings, such as “I’m not yet finished” (pause-filling, 3.2.1), “This is surprising” (mirativity, 3.3.1), “Aren’t you surprised?” (feigned surprise, 3.4.1), etc. Also notable in EDM semantics is that they are multifunctional, a property often cited by DM researchers (Maschler 1994; Jucker 2002; Aijmer et al. 2006; Koo & Rhee 2018; Rhee & Koo 2021). Further, Heine et al. (2021: 244) note that in many cases of DMs it is hard, if not impossible, to determine the ‘core’ ‘basic’ function. Evidently, this difficulty is from the interaction of diverse factors participating in the formation of DM meanings. As is clearly shown in Table 1 in 3.1, EDMs are also highly multifunctional. As noted in Rhee (2020b) and Rhee & Koo (2021), there are multiple factors that determine the DM function, and the same is true with EDMs. Among the key determinants are pragmatic inferences (see 5.1 above), source semantics (see 5.3 below), and prosody (see 5.2.3 below). Through intricate interaction of these determinants, there arise diverse functions that are closely related with each other. In terms of cross-domain multifunctionality, EDMs cut across discourse domains such as interaction, information, and interlocutor management. For instance, as shown in Table 1 in 3.1, multiple forms carry the functions of pause-filling (interaction management); emphasis, disregard, mirativity, etc. (information management); and feigned surprise, challenge/ protest, discontent, sarcasm, reluctance, hesitance, and upcoming disalignment (interlocutor management). 5.2.3 Phonology In terms of phonology, DMs tend to exhibit prosodic separation and shortness (Zwicky 1985; Schiffrin 1987; Bolinger 1989; Brinton 1996, 2017; Hansen 1997; Ajmer 2002; Dehé & Kavalova 2006; Maschler 2009; Song 2013; Hancil 2013; Hancil & Hirst 2013 and works therein; Heine 2013, 2018; and Sohn & Kim 2014; Sohn 2016; Haselow 2017, among many others; see, however, Heine et al. 2021: 51–52 for vulnerability of this criterion). For Korean DMs, elaborate analyses of prosody, by means of spectrogram features have been provided in a number of works, mostly targeting a single DM (notably, Song 2013, 2014, 2015; Song & Shin 2014; Sohn & Kim 2014; Lee & Sohn forthcoming). In the present research of macroscopic nature, however, such an in-depth analysis is beyond the immediate scope of interest due to multiplicity of forms and space limitation. For this reason, there are studies that made use of more simplified prosodic analyses based on impressionistically identified features. For instance, in the analyses of kule-DMs (i.e., those based on kule- ‘be so’), Rhee (2020b) and Rhee and Koo (2021), making use of such simplified means, note that prosodic features, e.g., intonation contour, duration, subsequent pause, etc. play

Chapter 8. On the development of discourse markers from elliptical structures

an important role in determining DM functions. Following this line of research, the characteristic prosodies of EDMs are identified as shown in Table 2. Table 2. Characteristic prosody of EDMs Domain

Function

Interaction

Pause-Filling

Information

Interlocutor

EDM ilen

celen

long level

mwusun weyn

mwel

weynkel

long level

Negation

short rising

short rising

short rising

Emphasis

short falling

short falling

short rising

Disregard

short rising

short rising

short rising

Mirativity

short falling

short falling

short rising

Mitigation/ Deflection

short rising

Uncertainty

long level

Feigned Surprise

short falling

Challenge/ Protest

short rising

Discontent

long rising

Reproach Sarcasm

short falling

short rising long falling

short rising

short rising

short rising

short rising

short rising

short rising

long rising

long rising

long rising

long rising

short rising

long rising

short rising

long rising

long rising

long rising

long rising

long rising

Reluctance/ Hesitance

long level

long level

Upcoming Disalignment

long rising

long rising

long rising

The prosodic characteristics of EDMs as shown in Table 2 present a few noteworthy generalizations. First of all, the same EDM may have variable prosodic realizations depending on its functions. For instance, the EDM ilen may have

227

228

Seongha Rhee

long-level, short-falling, and long-rising prosody, variable with the functions it carries. Second, certain functions may have signature prosody across all EDMs. For instance, pause-filling EDMs are invariably long-level; emphasis, disregard and challenge/protest EDMs, short-rising; discontent, sarcasm, and upcoming disalignment EDMs, all long-rising, and reluctance/hesitance EDMs, long-level, etc. Third, there is no EDM that has a uniform prosody across functions. This suggests that prosody is more closely related to functions than to individual forms. Other characteristic prosody patterns observable from the analysis, with respect to pause, are as follows: (33) a.

With a preceding/following pause: uncertainty, mitigation, discontent, reproach, reluctance b. Without a preceding/following pause: negation, emphasis, disregard, mirative, feigned surprise, challenge/protest, sarcasm

The prosodic patterns suggest that, in general, the EDMs with functions associated with indeterminate attitude tend to occur with a pause, whereas those associated with determinate attitude tend to occur without a pause. This is in line with the general tendency in language use involving affective and attitudinal stance, not restricted to EDMs, a finding similar to that in Rhee (2020b). 5.2.4 Pragmatics In an exposition of DMs across languages, Heine et al. (2021: 244–245) note that many DM sources are deictic in nature (e.g., temporal adverbs, imperatives involving perception verb), and that in their development into DMs they shift from temporal deictics to discourse deictics (Levinson 2006) or indexicals (Aijmer 2002; Aijmer et al. 2006; Furkó 2014). EDMs bear special significance in this respect. EDMs involve, in addition to the structural sources of involving an accusative case-marker and an adnominalizer, two major functional sources, i.e., deictic expressions (e.g., ile ‘like this’ and cele ‘like that’) and interrogatives (e.g., mwusu ‘what’, wey ‘why’, and mwe ‘what’). Deictic expressions are thus named because of the crucial role of context in determining the reference (cf. the Greek etymology of deixis ‘display, demonstration’). Thus, the EDMs ilen and celen are inherently indexical, whose correct interpretation lies solely in the context. Similarly, interrogatives are closely related to deixis by virtue of their reference being tied to temporal, personal, spatial, or other contexts with one common semantic property of espistemic modality, i.e., [unidentified]. These functional sources create indexicality and vagueness that are of highly pragmatic relevance. For further discussion of these source characteristics see 5.3 below.

Chapter 8. On the development of discourse markers from elliptical structures

5.3 Source characteristics As briefly discussed in 5.2.4, EDMs have two prominent sources, i.e., indexicality and vagueness. Indexicality is directly related to deixis and vagueness to ‘sortal’ deixis (see below) and interrogatives. These conceptual categories play an important role in the development of EDMs and thus merit our discussion. Korean has a tripartite spatial system, i.e., speaker-proximal (i ‘this’), speakermedial (ku ‘that’), and speaker-distal (ce ‘that’, in addition, the referent must be visible either physically or mentally). All these deictic expressions, either alone or combined with other forms, developed into DMs (Park 2006; Kim & Lee 2007; Lee 2010; Kim 2012; Shin 2012; Koo & Rhee 2018; Rhee & Koo 2021, among others). Deictic expressions involved in EDM are the proximal i and the distal ce, a pair with the greatest conceptual distance, as in ilen and celen. These two opposite deictic expressions create an interesting functional contrast. For instance, when the proximal-based EDM ilen is used, as in (5) and (10) above, the main objects in the discourse, e.g., ‘the water-bike’ in (5) and ‘hairpins, stockings, etc.’ in (10), are close to the speaker in physical terms, but, more importantly, the speaker places them close to himself or herself in the mental space. For this reason, the speaker is more deeply involved in the scene. Therefore, the speaker of (6), the inventor of the water-bike that is irreparably broken, signals stronger astonishment, despair, helplessness, etc. that is felt within him, and the speaker of (10), the interior designer intending to renovate the room, signals his disappointment and urgency of change felt within him. The EDM ilen signals this psychological involvement in addition to its respective function of pause-filling and mirativity-marking. This engaging effect, however, is not present with the distal-based EDM celen. For instance, the usage of the EDM celen is illustrated in Example (17), in which the speaker is uttering the example sentence while watching TV news. In the news, a packaged food producer’s contaminated food is being revealed and the speaker uses the EDM celen as part of condemning the producer. This distalbased EDM signals, in addition to reproach, the speaker’s mental distance from the one being condemned. For this reason, in condemning contexts, the EDM celen is far more frequently used than the proximal-based EDM ilen. The second conceptual category is interrogatives. Interrogatives inherently involve indeterminacy and thus interrogatives and indefinites, both pronominal and adverbial, are sometimes conflated (e.g. ‘who-someone’, ‘what-something’, etc.), a point indicated in 4.1 above. Thus, mwusun is ambiguous between ‘what kind of ’ and ‘some kind of ’; weyn between ‘for what reason’ and ‘for some unknown reason’; mwel between ‘what’ and ‘something’; and weynkel ‘what kind of thing’ and ‘something of unknown origin’. Furthermore, deictic expressions involved in the development of EDMs are ile ‘be like this’ and cele ‘be like that’

229

230

Seongha Rhee

(note that ‑le is an adverbializer), which are sortal, different from the more individuated ‘this’ and ‘that’. Therefore, ile and cele (the bases of the EDMs ilen and celen) refer to a category of similar members, and this characteristic creates vagueness, as is clearly felt from their meanings: ‘this kind of ’ and ‘that kind of ’. The effect of vagueness in EDMs is that it leads to creation of the vaguenessrelated concepts, e.g., uncertainty, reluctance, pause-filling, mitigation/deflection, etc. In (19) above, for instance, the speaker uses the EDM mwusun, which is syntactically extraneous because of a better-qualified modifier talun ‘other’ for the noun being modified il ‘development’. The speaker, while in lexical search for an appropriate word, fills the spot with a vague term (analogous to the English kind of). Since all EDM sources involve vagueness, the function of this concept bears special significance in the development of EDMs.

6.

Summary and conclusion

A subset of Korean DMs arose from elliptical structures, hence EDMs. This paper analyzed six of them that still carry the accusative case-marker or adnominalizer. Despite the presence of these morphological markers, which require them to occur in the theme-argument position or pre-nominal modifier position, EDMs occur at non-licensed positions or even stand alone. They carry diverse functions across the interaction-, information-, and interlocutor-management domains. Notably, many of them mark the speaker’s (feigned) surprise, discontent, challenge/protest, disregard, etc. The development of these affective and attitudinal functions is evidently due to the structural ellipsis, i.e. the speaker is unable to complete the utterance by being overwhelmed by the event or state of affairs. A diachronic analysis of these EDMs shows a number of interesting aspects of DM development. For instance, the source forms undergo semantic generalization, whereby the scope is extended, and the forms move out of the licensed position, whereby they acquire syntagmatic freedom. It is also notable that a DM function is not only derivable from the source lexemes but also from the structures themselves (i.e., ellipsis). EDMs carry most DM characteristics in morphosyntactic, semantic, phonological, and pragmatic aspects. Especially, their functions are often correlated with their characteristic prosodies. Also significant is the role of deixis and interrogatives in the source forms since they create indexicality and vagueness that play a significant role in the development of EDM functions.

Chapter 8. On the development of discourse markers from elliptical structures

Funding This research was supported by the research fund of Faculty of Liberal Arts, Mahidol University, Thailand and the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2021S1A5A2A01060515)

Abbreviations acc adn ben comp dec dm edm end excl gen hon hum imp inst intj

accusative adnominal benefactive complementizer declarative discourse marker ellipsis-based discourse marker sentence-ender exclamative genitive honoriric humilific imperative instrumental interjection

lmt mir nom nomz pej perf pol pres proh prom pst purp q rept top

limitative mirative nominative nominalizer pejorative perfective polite present prohibitive promissive past purposive question reportative topic

Acknowledgements Special thanks go to Bernd Heine and Alexander Andrason for their kind comments and suggestions on an earlier version of this paper. Also thanks go to Julia Pozy for making suggestions for improvement in style and content.

References Aijmer, Karin. 1996. Conversational Routines in English: Convention and Creativity [Studies in Language and Linguistics]. London: Longman. Aijmer, Karin. 2002. English Discourse Particles: Evidence from a Corpus [Studies in Corpus Linguistics 10]. Amsterdam: John Benjamins. Aijmer, Karin, Foolen, Ad & Simon-Vandenbergen, Anne-Marie. 2006. Pragmatic markers in translation: A methodological proposal. In Approaches to Discourse Particles, Kerstin Fischer (ed.), 101–114. Amsterdam: Elsevier.

231

232

Seongha Rhee

Barth-Weingarten, Dagmar & Couper-Kuhlen, Elizabeth. 2002. On the development of final though: A case of grammaticalization? In New Reflections on Grammaticalization [Typological Studies in Language 49], Ilse Wischer & Gabriele Diewald (eds), 345–361. Amsterdam: John Benjamins. Beeching, Kate & Detges, Ulrich (eds). 2014. Discourse Functions at the Left and Right Periphery: Crosslinguistic Investigations of Language Use and Language Change. Leiden: Brill. Bolinger, Dwight. 1989. Intonation and its Uses: Melody in Grammar and Discourse. London: Edward Arnold. Brinton, Laurel J. 1996. Pragmatic Markers in English: Grammaticalization and Discourse Functions. Berlin: Mouton de Gruyter. Brinton, Laurel J. 2017. The Evolution of Pragmatic Markers in English: Pathways of Change. Cambridge: CUP. Dehé, Nicole & Kavalova, Yordanka. 2006. The syntax, pragmatics, and prosody of parenthetical what. English Language and Linguistics 10(2): 289–320. Evans, Nicholas. 2007. Insubordination and its uses. In Finiteness: Theoretical and Empirical Foundations, Irina Nikolaeva (ed.), 366–431. Oxford: OUP. Evans, Nicholas. 2009. Insubordination and the grammaticalization of interactive presuppositions. Paper presented at Methodologies in Determining Morphosyntactic Change Conference, Museum of Ethnography, Osaka, March. Evans, Nicholas & Watanabe, Honoré (eds). 2016. Insubordination [Typological Studies in Language 115]. Amsterdam: John Benjamins. Fischer, Kerstin. 2006. Toward an understanding of the spectrum of approaches to discourse particles: Introduction to the volume. In Approaches to Discourse Particles, Kerstin Fischer (ed.), 1–20. Amsterdam: Elsevier. Fraser, Bruce. 1999. What are discourse markers? Journal of Pragmatics 31: 931–952. Furkó, Bálint Péter. 2014. Cooptation over grammaticalization. Argumentum 10: 289–300. Hancil, Sylvie. 2013. Introduction. In Hancil & Hirst (eds), 1–31. Hancil, Sylvie & Hirst, Daniel (eds). 2013. Prosody and Iconicity [Iconicity in Language and Literature 13]. Amsterdam: John Benjamins. Hansen, Maj-Britt Mosegaard. 1997. Alors and donc in spoken French: A reanalysis. Journal of Pragmatics 28(2): 153–187. Haselow, Alexander. 2014. Sequentiality in dialogue as a trigger for grammaticalization. In Grammaticalization – Theory and Data [Studies in Language Companion Series 162], Sylvie Hancil & Ekkehard König (eds), 203–234, Amsterdam: John Benjamins. Haselow, Alexander. 2015. Left vs. right periphery in grammaticalization: The case of anyway. In New Directions in Grammaticalization Research [Studies in Language Companion Series 166], Andrew D. M. Smith, Graeme Trousdale, & Richard Waltereit (eds), 157–186. Amsterdam: John Benjamins. Haselow, Alexander. 2017. Spontaneous Spoken English: An Integrated Approach to the Emergent Grammar of Speech. Cambridge: CUP. Haspelmath, Martin. 2011. The gradual coalescence into ‘words’ in grammaticalization. In The Oxford Handbook of Grammaticalization, Heiko Narrog & Bernd Heine (eds), 342–355.Oxford: OUP.

Chapter 8. On the development of discourse markers from elliptical structures

Heine, Bernd. 2013. On discourse markers: Grammaticalization, pragmaticalization, or something else? Linguistics 51(6): 1205–1247. Heine, Bernd. 2018. Are there two different ways of approaching grammaticalization? In New Trends on Grammaticalization and Language Change [Studies in Language Companion Series 202], Sylvie Hancil, Tine Breban & José Vicente Lozano (eds), 23–54. Amsterdam: John Benjamins. Heine, Bernd, Kaltenböck, Gunther, Kuteva, Tania & Long, Haiping. 2021. The Rise of Discourse Markers. Cambridge: CUP. Jeon, Yeong-Ok. 2002. A study on discourse markers in Korean. Journal of Speech Communication 4: 113–145. Jucker, Andreas H. 2002. Discourse markers in Early Modern English. In Alternative Histories of English, Richard Watts & Peter Trudgill (eds), 210–230. London: Routledge. Jung, Yonhee. 2001. Grammaticalization of Korean Clause Connectives. PhD dissertation, Hankuk University of Foreign Studies, Seoul. Kim, In-gyu. 2012. A Study on Usages of Korean Reference Items: Focused on the Data of Oral Conversation in Gyeongbuk Dialect. MA thesis, Kyungpook National University, Daegu, Korea. Kim, Mary Shin, Rhee, Seongha & Smith, Hye Young. 2021. The Korean discourse particle kulssey across discreet positions and contexts in talk-in-interaction. Journal of Pragmatics 182: 16–41. Kim, Myung-Hee. 2005. The development of Korean interrogatives as discourse markers. Discourse and Cognition 12(2): 41–63. Kim, Myung-Hee. 2006. Discourse functions of interrogative mwusun in Korean. Discourse and Cognition 13(2): 21–42. Kim, Myung-Hee & Lee, Jeonghwa. 2007. The role of subjectivity and intersubjectivity in the grammaticalization of icey in Korean. Discourse and Cognition 14(2): 27–49. Kim, Sunyoung. 2016. Discourse marker functions of museun in spoken discourse. The Language and Culture 12(3): 83–98. Kim, Tae Yeop. 2001. Kwuke Congkyelemiuy Mwunpep (Grammar of Korean sentence-final markers). Seoul: Kwukhakcalyowen. Kim, Young-Ran. 2003. Lexical Information and Teaching Methods of Korean Interrogatives for Korean Language Education. PhD dissertation, Sangmyung University, Seoul. Kim, Xianghua. 2002. A Study of Discourse Markers in Korean. PhD dissertation, Keimyung University, Daegu, Korea. Ko, Yong-Kun. 1991. Phyocwun Cwungseykwuke Mwunpeplon (Middle Korean grammar). Seoul: Tower Press. Koo, Hyun Jung. 2005. Conditional markers in speech context: From conditional to politeness. Discourse and Cognition 12(2): 1–22. Koo, Hyun Jung. 2012. On grammaticalization of ‑kinun construction in Korean. Discourse and Cognition 19(3): 1–28. Koo, Hyun Jung. 2018. Hankwuke tamhwaphyoci yenkwuuy tonghyangkwa cenmang (Research tends and prospect of Korean discourse markers). In Hankwuke Yenkwuuy Saylowun Hulum (New research trends of Korean), Yong-Kwon Shin, (ed.), 221–255. Seoul: Pagijong Press.

233

234

Seongha Rhee

Koo, Hyun Jung & Rhee, Seongha. 2001. Grammaticalization of sentential end marker from a conditional marker. Discourse and Cognition 8(1): 1–19. Koo, Hyun Jung & Rhee, Seongha. 2018. On the emergence of polyfunctionality of discourse markers: The case of kulay ‘it is so’ in Korean. Paper presented at the 36th International Conference of the Spanish Society for Applied Linguistics (AESLA-2018), Universidad de Cádiz, Cádiz, Spain, 19–21 April. Lee, Heejoo & Sohn, Sung-Ock S. Forthcoming. From interrogative to disaffiliative stance marker: An analysis of mwusun ‘what’ in Korean conversation. University of California, Los Angeles. Lee, Hye Seung. 2020. Korean interrogative mwusun ‘what kind of ’ as a discourse marker. Japanese/Korean Linguistics 26: 323–333. Lee, Hyo Sang. 1991. Tense, Aspect, and Modality: A Discourse-pragmatic Analysis of Verbal Affixes in Korean from a Typological Perspective. PhD dissertation, University of California, Los Angeles. Lee, Hyo Sang. 2015. Modality. In The Handbook of Korean Linguistics, Lucien Brown & Jaehoon Yeon (eds), 249–268. Oxford: Wiley Blackwell. Lee, Kigap. 2010. Dialectal variation of Korean discourse markers, geunyang, geujeo, and geuman. Journal of Korean Dialectology 11: 5–35. Lee, Kwang-Ho. 1994. Cwungseykwuke Mwunpeplon (Middle Korean grammar). Seoul: Hakyeonsa. Lee, Kwang-Ho. 2009. Meaning property and form change of ‘musu’ and ‘musuk/musum/musus.’ Korean Language and Literature 151: 35–57. Lehmann, Christian. 2015[1982]. Thoughts on Grammaticalization, 3rd edn. Berlin: Language Science Press. Levinson, Stephen C. 2006. Deixis. In The Handbook of Pragmatics, Laurence R. Horn & Gregory L. Ward (eds), 87–120. London: Blackwell. Maschler, Yael. 1994. Metalanguaging and discourse markers in bilingual conversation. Language in Society 23: 325–366. Maschler, Yael. 2009. Metalanguage in Interaction: Hebrew Discourse Markers [Pragmatics & Beyond New Series 181]. Amsterdam: John Benjamins. Onodera, Noriko O. 2004. Japanese Discourse Markers: Synchronic and Diachronic Discourse Analysis [Pragmatics & Beyond New Series 132]. Amsterdam: John Benjamins. Park, Keun-young. 2006. Grammaticalization of ‘I, geu, jeo’. The Language and Culture 2(1): 65–93. Park, Mee-Jeong & Sohn, Sung-Ock S. 2002. Discourse, grammaticalization, and intonation: An analysis of ‑ketun in Korean. Japanese/Korean Linguistics 10: 306–319. Park, Sang-Sook. 2020. Discourse Functions and Prosody in the Expression of Interrogative. PhD dissertation, The Catholic University of Korea, Seoul. Rhee, Seongha. 2012. Context-induced reinterpretation and (inter)subjectification: The case of grammaticalization of sentence-final particles. Language Sciences 34: 284–300. Rhee, Seongha. 2015. On the emergence of Korean markers of agreement. Journal of Pragmatics 83: 10–26. Rhee, Seongha. 2018. Grammaticalization of the plural marker in Korean: From object to text to stance. Journal of Language Sciences 25(4): 221–249.

Chapter 8. On the development of discourse markers from elliptical structures

Rhee, Seongha. 2020a. The power of the unspoken: Ellipsis and grammaticalization in Korean and beyond. Studies in Linguistics 56: 1–36. Rhee, Seongha. 2020b. On determinants of discourse marker functions: Grammaticalization and discourse-analytic perspectives. Linguistic Research 37(2): 289–325. Rhee, Seongha & Koo, Hyun Jung. 2017. Audience-blind sentence-enders in Korean: A discourse-pragmatic perspective. Journal of Pragmatics 120: 101–121. Rhee, Seongha & Koo, Hyun Jung. 2020. From quotation to surprise: The case in Korean. Journal of Pragmatics 155: 83–100. Rhee, Seongha & Koo, Hyun Jung. 2021. On divergent paths and functions of ‘background’-based discourse markers in Korean. In Studies at the Grammar-Discourse Interface [Studies in Language Companion Series 219], Alexander Haselow & Sylvie Hancil (eds), 77–100. Amsterdam: John Benjamins. Ryu, Chang-Don. 1980. Ehwisa Yenkwu (A study of lexical history). Seoul: Iwoo Publishing. Schiffrin, Deborah. 1987. Discourse Markers. Cambridge: CUP. Schourup, Lawrence. 1999. Discourse markers. Lingua 107: 227–265. Shin, Eun Kyung. 2012. An Analysis of the Function ‘irae, geurae, jeorae’ in Spoken Korean. MA thesis, Konkuk University, Seoul. Sohn, Sung-Ock. 1995. On the development of sentence-final particles in Korean. Japanese/Korean Linguistics 5: 219–234. Sohn, Sung-Ock. 2016. Development of the discourse marker kulentey (‘but’ ‘by the way’) in Korean: A diachronic and synchronic approach. Journal of Historical Pragmatics 17(2): 231–254. Sohn, Sung-Ock & Kim, Stephanie Hyeri. 2014. The interplay of discourse and prosody at the left and right periphery in Korean: An analysis of kuntey‘but’. In Discourse Functions at the Left and Right Periphery: Crosslinguistic Investigations of Language Use and Language Change, Kate Beeching & Ulrich Detges (eds), 221–249. Leiden: Brill. Song, Inseong. 2013. Functions and prosodic features of discourse marker mweo. Korean Linguistics 58: 83–106. Song, Inseong. 2014. Prosodic features and discourse functions of discourse marker mak. Korean Linguistics 65: 211–236. Song, Inseong. 2015. Functions and Prosodic Features of Korean Discourse Markers. PhD disseration, Korea University, Seoul. Song, Inseong & Shin, Jiyoung. 2014. The realization of functions and morphological, prosodic features of discourse marker jom. Korean Linguistics 62: 315–339. Tabor, Whitney & Traugott, Elizabeth C. 1998. Structural scope expansion and grammaticalization. In The Limits of Grammaticalization [Typological Studies in Language 37], Anna Giacalone Ramat & Paul J. Hopper (eds), 229–272. Amsterdam: John Benjamins. Traugott, Elizabeth Closs. 2014. Intersubjectification and clause periphery. In Intersubjectivity and Intersubjectification in Grammar and Discourse [Benjamins Current Topics 65], Lieslotte Brems, Lobke Ghesquière & Freek Van de Velde (eds), 7–27. Amsterdam: John Benjamins. Waltereit, Richard. 2002. Imperatives, interruption in conversation and the rise of discourse markers: A study of Italian guarda. Linguistics 40: 987–1010.

235

236

Seongha Rhee

Wilson, Deirdre. 2011. The conceptual-procedural distinction: Past, present and future. In Procedural Meaning: Problems and Perspectives, Victoria Escandell-Vidal, Manuel Leonetti & Aoife Ahern (eds), 3–31. Bingley: Emerald. Zwickey, Arnold M. 1985. Clitics and particles. Language 61(2): 283–305.

chapter 9

On the grammaticalization of ideophones Alexander Andrason1, 2 & Bernd Heine3 1

Living Tongues Institute for Endangered Languages | 2 University of Cape Town | 3 University of Cologne

Ideophones, like English bang or thud, are interactive expressions used as vivid depictions of sensory imagery of states, events, objects, or qualities (cf. Dingemanse 2011, 2012, 2018; Dingemanse & Akita 2017; Andrason 2020, 2021). They are claimed to represent a universal class of linguistic forms – that is, any given language can be expected to have a set of them. That set may be highly limited, as is the case in many European languages, but it may as well be almost as large as that of lexical categories like nouns and verbs. There are thousands of ideophones in the spoken usage of languages like Korean, Japanese and Basque, which have 4500 or more of them (see Dingemanse 2018; Haiman 2018). Ideophones exhibit an ambivalent structural behavior. On the one hand, they have been described as grammatical forms that are syntactically unattached and prosodically set off from surrounding text material. On the other hand, they have also been described as morphosyntactically integrated adverbials, adjectivals, verbals, or nominals in a number of languages. Building on some earlier work (especially Dwyer & Moshi 2003), the goal of the present paper is to look at ideophones from the perspective of grammaticalization theory with a view to accounting for this ambivalent behavior. It is argued in the paper that we are dealing here with a process that differs from ‘canonical’ grammaticalization in that the end-product of the process is a lexical rather than a grammatical form. Keywords: context extension, decategorialization, desemanticization, discourse construction, erosion, grammaticalization, ideophone, Siwu, Xhosa

https://doi.org/10.1075/slcs.232.09and © 2023 John Benjamins Publishing Company

238

Alexander Andrason & Bernd Heine

1.

Introduction

The main goal of the present paper is to apply grammaticalization theory to the study of ideophones (Narrog and Heine 2021).1 Following some other authors studying ideophones (Ibarretxe-Antuñano 2017; Andrason 2020; Andrason, Phiri & Fehn forthcoming) as well as other kinds of interactives (Nübling 2004; StangeHundsdörfer 2016; Andrason & Dlali 2020; Heine 2023), the present paper is based on a prototype approach to categorization (cf. Taylor 1995; Evans Green 2006; Janda 2015). Accordingly, we view the category of ideophones as a radial network of members that vary in their compliance with the prototype – the ideal exemplar constructed by linguists – and, through a family-resemblance relationship, emanate from the categorial center to peripheries. Members that fully comply with the prototype are canonical and, in the model, populate the center of the category; the members that comply with the prototype minimally are non-canonical and occupy the margins of the category where they, to some extent, overlap with the peripheral members of other categories. Between these two extremes, there is a cloud of semi-canonical members that exhibit a gradual decrease in their compliance with the prototype and, correspondingly, a more remote placement with respect to the categorial nucleus.

1.1 On ideophones Ideophones are also known by a number of other names, such as ‘expressives’ or ‘mimetics’, for a list of terms that are employed for ideophones, see Dingemanse (2018: 2, Table 1). They are poorly represented in most European languages (Diffloth 1972: 440; Basque being a remarkable exception; see Ibarretxe-Antuñano 2017), English examples include forms like bang, boing, boom, plop, pow, ta-dah, thud, vroom, and zoom. In some languages of Africa and eastern Asia, by contrast, they belong to the most numerous form classes. There are thousands of ideophones in the spoken usage of a number of languages: Korean is said to have more than 5000 of them, Japanese 4500, Gbeya over 3000, and Zulu 3000 (Dingemanse 2018: 15, 2019: 26; Koo & Rhee 2018). To the extent that ideophones can be expected to be found in any language of the world, there might be justification to call them a ‘universal category’ – to our knowledge, no language has so far been identified that would lack ideophones. But considering that up to now only a small part of the world’s languages has been described appropriately, no claim to this effect is made here.

1. We wish to express our gratitude to Seongha Rhee and an anonymous reviewer for highly valuable comments on an earlier version of this paper.

Chapter 9. On the grammaticalization of ideophones

The grammatical status of ideophones has been a challenge to researchers. They have been classified as adverbs, adjectives, pseudo-adjectives, onomatopoeic adjectives, verbs, derived verbs, predicatives, or else as something unique, differing from such word classes. Dwyer and Moshi (2003: 174, (2)) described ideophones in terms of six properties, while Andrason (2021: 124–31) proposes six (yet non-exhaustive) phonological, five morphological, and five syntactic properties characterizing this type. Ideophones have been covered in great detail and in all their various aspects of behavior in the work of Mark Dingemanse (see References) and the reader is referred to that work for more information (see also Voeltz & Kilian-Hatz 2001; Haiman 2018). This paper is restricted to features of ideophones that immediately relate to their grammaticalization. With reference to their relative degree of grammaticalization, the status of ideophones differs considerably from one language to another.

1.2 Definition Working on Bantu languages in southern Africa (see, e.g., Voeltz & Kilian-Hatz 2001), Doke (1935: 118) proposed the by now classic definition of ideophones: A vivid representation of an idea in sound. A word, often onomatopoeic, which describes a predicate, qualificative or adverb in respect to manner, colour, sound, smell, action, state or intensity. (Doke 1935: 118)

More recent work was shaped most of all by that of Dingemanse (e.g., 2011, 2012, 2019). Describing ideophones as “conventionalised words that combine iconic and arbitrary form-meaning mappings”, he defines them in the following way: “[…] a canonical ideophone is a member of an open lexical class of marked words that depict sensory imagery” (Dingemanse 2019: 13, 18). The definition proposed in (1) is based on that by Dingemanse (2019) but describes ideophones not as words or a lexical class but rather as a type of interactives, for reasons to be given in Section 2. (1) Definition of ideophones (Heine 2023, Section 3.5.2) An ideophone is an interactive used for a vivid depiction of sensory imagery of a state, event, object, or quality.

The notion ‘sensory imagery’ covers both perceptions of the external world and inner sensations, it includes in much the same way the sensory domains of sight, touch, hearing, kinaesthesia, taste, color, texture, and inner physiological conditions (Dingemanse 2018: 3, 2019: 15). Concerning the notion ‘depiction’, see Clark (2016; 2019; see also Dingemanse 2013, 2015; Heine 2023, Section 2.1.3). The phrase ‘a state, event, object, or quality’ in (1) refers to the fact that ideophones can depict a range of quite different concepts. With regard to the four

239

240

Alexander Andrason & Bernd Heine

aspectual (Aktionsart) categories distinguished by Vendler (1967), ideophones can depict the aspectual contours of accomplishments, achievements, activities, or states. English ideophones such as tick-tock and vroom, for instance, are suggestive of the aspectual contours of activities whereas ideophones like bang and thud are suggestive of achievements (Kita 1997: 404; Heine 2023, Section 3.5.3.2). Ideophones are frequently treated in tandem with onomatopoeias (or ‘onomatopoetics’, or ‘onomatopes’), that is, forms characterized by a sound symbolic relationship between the depicting and the depicted item, or syntactically isolated forms that “phonetically imitate noises” (Meinard 2015: 151). In many treatments it remains largely unclear what exactly the status of the two is relative to one another. The stance adopted here is that expressed by Andrason, Phiri and Fehn (forthcoming) thus: Onomatopoeias occupy the first (lowest) position on the implicational hierarchy of ideophones that leads from ideophones denoting auditory sensations (onomatopoeias) to ideophones denoting visual sensations, and, further, to ideophones denoting psychological and cognitive states. (Andrason, Phiri and Fehn forthcoming: 4)

With this stance, which is based on Dingemanse (2012: 663), we follow a number of students of ideophones in assuming that onomatopoeias are located within the semantic space covered by ideophones, and that onomatopoeias form a “subset of ideophones imitative of sound” (Dingemanse 2018: 19) – hence: “All onomatopes are ideophones, but not all ideophones are onomatopes” (Haiman 2018: 82).

1.3 Grammatical features With regard to the grammatical status of ideophones, it is in particular two features that need to be mentioned. On the one hand, ideophones have been described as having an argument structure and, on the other hand, they tend to be associated with three kinds of constructions. While ideophones can occur on their own, forming distinct utterances, Dingemanse (2011: 145) makes the important observation that they need to be “supported by other talk”, even if that “other talk” may be implied rather than being formally encoded. In the framework of Heine (2023, Section 2.2), that “other talk” is described as an argument of ideophones, called the T (theme) argument.2 Thus, in Example (2), use of the ideophone vroom depends in its interpretation on what it refers to, which in this example is the preceding text piece The car went instantiating its T argument – 2. An anonymous reviewer rightly observes that this usage of ‘theme’ as a discourse argument differs from that found in other theoretical frameworks as well as in work on information structure.

Chapter 9. On the grammaticalization of ideophones

that is, without reference to some kind of T argument stated or implied, use of ideophones would normally not be meaningful. (2) English The car went vroom.

(Dingemanse 2021)3

Since T need not be overtly expressed, ideophones are not uncommonly used as stand-alones, forming utterances of their own, with T being implied rather than encoded, being recoverable from the preceding context, as in the English example in (3). (3) English Ta-dah!

In a number of languages, the three main discourse constructions listed in Table 1 tend to be distinguished in the way ideophones are deployed in discourse. Consider the English examples in (4) illustrating the constructions. Following Heine (2023, Section 3.5.3.2) the constructions are referred to as the free construction, as in (4a), the quotative construction, cf. (4b), and the modifier construction, cf. (4c). (4) English a. Thud! b. And the vase went thud! c. The vase fell to the ground – thud!

In the free construction, the ideophone constitutes an independent utterance of its own, that is, its T argument is not realized overtly, even if the ideophone may be preceded and/or followed by other discourse material. In the quotative construction, the ideophone is presented as an obligatory discourse complement of its T argument.4 The latter has the structure of a quotative clause, where the quotative marker frequently takes on the form of a verb for ‘say’, ‘do’, ‘go’ or of other verbs whose main or only function it is to introduce direct speech. English speakers frequently use be like, or go as a quotative marker instead, as in (4b), where the ideophone is presented like a direct speech utterance introduced by went, and the latter can be interpreted simultaneously as the main verb and as a quotative marker introducing the ideophone. Being a discourse complement, the ideophone cannot be omitted in this type of construction. In the modifier construction, as in (4c), the ideophone is presented as a kind of afterthought or appositional parenthetical, that is, it is an omissible text seg-

3. Ideophones are generally printed in bold in this paper. 4. Importantly, discourse complements are not syntactic complements, they have semanticpragmatic scope over text pieces larger than a sentence.

241

242

Alexander Andrason & Bernd Heine

ment having the appearance of a discourse modifier and being semantically compatible with its T argument. Table 1. The main discourse constructions used for presenting ideophones (Heine 2023, Table 3.9) Construction

Discourse status of ideophone

Marking

Placement

Free construction

Independent utterance

No marking

No constraints

Quotative construction

Discourse complement

Verb for ‘say’, ‘do’, ‘go’, etc., and/or a quotative marker

In the slot reserved for direct speech

Modifier construction

Appositional parenthetical

No marking

After the text piece they refer to

2.

Ideophones as interactives

Most analyses of ideophones converge on treating them as a category of their own, to be distinguished from other kinds of grammatical or lexical categories, forming a separate ‘word’ class and/or being only marginally attached to grammar, if at all. According to Heine (2023) in contrast, ideophones are not part of sentence grammar but rather constitute one of the ten types of expressions within interactive grammar, jointly with attention signals, directives, discourse markers, evaluatives, interjections, response elicitors, response signals, social formulae, and vocatives. Ideophones thus are classified as interactive forms or, in short, as interactives.

2.1 Interactives Interactives are prefabricated routine forms and include what Ferguson (1981) describes as social formulas and Coulmas (1981: 2–3) as conversational routines, that is, “highly conventionalized prepackaged expressions whose occurrence is tied to more or less standardized communication situations”. They are ‘interactive’ in the sense that they are grounded in the world of social interaction. One noteworthy characteristic of the category of interactives is that, at least in some of their usages, they can form utterances of their own, consider the English examples in Table 2. But, in addition, they share a set of properties that define interactives as a category and distinguish them from elements of sentence grammar, such as nouns, pronouns, verbs, adverbs, etc. A definition of interactives is

Chapter 9. On the grammaticalization of ideophones

presented in (5). Note that the definition makes no mention of the functions of interactives, for which see Heine (2023, Chapter 3). Table 2. The main types of interactives Type

English example

Attention signal

Hey!

Directive

Shh!

Discourse marker

Indeed.

Evaluative

Great!

Ideophone

Thud!

Interjection

Ouch!

Response elicitor

Right?

Response signal

No.

Social formula

Bye!

Vocative

Mom!

(5) An interactive is an invariable deictic form that is in some way set off from the surrounding text semantically, syntactically and prosodically and can neither be negated nor questioned.

For cross-linguistic evidence on which the properties in (5) are based, see Heine (2023, Section 1.3). As was observed in Section 1, categorization in this paper rests on prototype theory. With regard to the definition in (5) this means that prototypical instances of interactives conform to all the properties while less prototypical ones may lack one or the other of the properties. The violation of each feature itself may be scalar. For instance, prosodic detachment of ideophones from the surrounding text may be more or less radical, and lack of compliance with morphological invariability may be gradient (Andrason 2020: 148). Ideophones are for the most part in accordance with the definitional properties, as is demonstrated in Heine (2023, 3.5.7).5 Thus, the English examples ta-dah, thud and vroom in (2) to (4) are invariable and typically separated from the surrounding discourse syntactically and prosodically, and they cannot normally be negated or questioned. Problems exist, however, with the modifier construction of ideophones, where the ideophone is not semantically set off but rather accepts only semantically compatible T arguments – in other words, where the meaning 5. An anonymous reviewer rightly suggests that ideophones might well be interpreted “as primitive (or pre-) lexical items which are not yet part of the sentence grammar.”

243

244

Alexander Andrason & Bernd Heine

of the ideophone constrains the nature of the preceding text piece serving as its T argument. Thus, the following utterance does not seem to be well-formed because the T argument The pillow slipped to the floor and the ideophone pow are not semantically compatible with one another. (6) English ?The pillow slipped to the floor, pow!

(Dwyer & Moshi 2003: 176, (6))

A second problem exists with the deictic status of ideophones in (5) in that it has been argued that ideophones are not or not clearly indexicals (Dingemanse 2011:155). There is, however, also an alternative view on this issue (Haiman 2018: 136), and the issue is in need of further research (see Heine 2023, 3.5.7, (e)). In spite of such issues, ideophones conform on the whole to the grammatical structure of interactives, and in cases where they do not, this requires an explanation.

2.2 Problems When asserting in Section 2.1 that ideophones conform on the whole to the grammatical structure of interactives then this applies to what one might wish to call prototypical instances of ideophones but not really to many other of their uses. Ideophones have been classified as adverbs, adjectives, pseudo adjectives, onomatopoeic adjectives, verbs, derived verbs, predicatives, qualitatives, etc. (Dwyer and Moshi 2003). For example, in the Sepik language Iatmul of Papua New Guinea, ideophones are treated as a subclass of adverbs (Jendraschek 2012: 312–317) while in the Somali language of northeastern Africa they are said to form a subclass of nouns (Dingemanse & Akita 2017: 506). Rather than representing a clearly delineable category, ideophones in many languages exhibit a high amount of structural variation, behaving in some of their uses as interactives but as words of sentence grammar in other uses. In fact, not uncommonly, ideophones do not conform to the definitional properties of interactives in (5) of Section 2.1. Take the case of the ideophone kpokporo ‘be hard’ of the Siwu language of southeastern Ghana in (7), which clearly is at variance with those properties: It is not invariable, taking prefixes, it is not syntactically detached nor prosodically set off from the surrounding text, it can be negated (cf. the negation marker ‑i-), etc. (7) Siwu (Kwa, Niger-Congo; Dingemanse 2017a: 371, (12)) […] ì-i-kpokporo. it-neg-ide.be:hard ‘It isn’t kpokporo [hard].’

Chapter 9. On the grammaticalization of ideophones

A look at the use patterns of ideophones in languages across the world suggests that, while they do show the properties in (5) in some of their uses, overall they may have the appearance of a structural motley collection. In this respect they differ from other types of interactives, none of which exhibits such a variable behavior. The question then is what accounts for this behavior. The answer proffered in this paper is that, unlike other types of interactives, ideophones have a strong propensity to undergo grammaticalization. This is not a new hypothesis, it has also been expressed in some form or other in a number of other studies, especially by Dwyer and Moshi (2003). What is new, however, is that we propose a framework that accounts for all the main features of this variable behavior. To this end, a catalog of four parameters is presented in Section 3.1, and these parameters are applied to two specific languages, namely Siwu in Section 3.2 and Xhosa in Section 3.3.

3.

Grammaticalization

In their study of the development of Japanese ideophones, Dingemanse (2017b: 197) and Dingemanse and Akita (2017) conclude that ideophones show “an inverse relation between prosodic foregrounding and morphosyntactic integration: Ideophones that are more deeply integrated in the structure of the clause lose their prosodic foregrounding.” As is argued in the present paper, this observation is hardly surprising when looking at the development of ideophones from the perspective of grammaticalization theory.

3.1 Diagnostics In order to find ways of describing processes of grammaticalization, a wide range of criteria have been proposed (see Narrog & Heine 2021: 44–65). In this paper we are restricted to the parameters proposed by Heine and Kuteva (2007: 133–46; see also Kuteva et al. 2019). These parameters rest on a cross-linguistic survey and capture salient features of forms undergoing grammatical change, based on the assumption that grammaticalization affects all major components shaping linguistic expressions. The four parameters distinguished are presented in (8). (8) Parameters of grammaticalization (Heine & Kuteva 2007: 33–46) a. Context extension: The rise of new meanings when linguistic expressions are extended to new contexts, leading to context-induced reinterpretation.6 b. Desemanticization (ʽsemantic bleachingʼ): Loss or generalization of meaning content or functions in such contexts.

245

246

Alexander Andrason & Bernd Heine

c.

Decategorialization: Loss of morphosyntactic properties characteristic of lexical or other less grammaticalized forms.7 d. Erosion (ʽphonetic reductionʼ): Loss of phonetic substance, including prosodic features.

These parameters make it possible to look at each of the main components of grammar independently and in their own right, where (8a) involves the pragmatic, (8b) the semantic, (8c) the morphosyntactic, and (8d) the phonetic component. And the ordering of the four parameters reflects the diachronic sequence in which they typically apply: Grammaticalization tends to start out with context extension, which triggers desemanticization, subsequently decategorialization, and finally erosion. Erosion thus is the last parameter involved, and in many cases of grammaticalization it is not, or not yet, a relevant parameter. With reference to ideophones, Dingemanse and Akita (2017: 506) propose the following pre-conditions for grammatical integration in Japanese: (i) linear position (peripheral items are less integrated), (ii) syntactic optionality (optional items are less integrated), and (iii) embedding in morphosyntactic structure (less deeply embedded items are less integrated). These are in fact relevant criteria for identifying instances of grammaticalization. All three criteria are restricted to the parameter of decategorialization: Ideophones, being typically placed ‘optionally’ at the periphery of an utterance (cf. (i, ii)), lose their status as a syntactically unattached category and are gradually integrated and embedded in the structure of a sentence (cf. (iii)).8 In addition, these authors also propose another criterion to define grammatical integration, namely what they call ‘expressiveness’, arguing that the more expressive an ideophone is, the less it is integrated in the morphosyntactic structure of the sentence (Dingemanse and Akita 2017: 506). Expressiveness, which is one of the manifestations of what Heine (2023, Section 4.3) calls ‘expressive reinforcement’, is described there with reference to the three features in (9). All features ultimately concern the ability to either add or modify phonetic substance and/or substance used for the expression of ideophones. Accordingly, loss of these features, and hence, increased grammatical integration, falls ultimately within the parameter of erosion in (8). 6. The parameter of context extension covers essentially the definition of grammaticalization as proposed by Hopper and Traugott (2003: 18). 7. Note that the term ʽdecategorializationʼ differs from that employed by Hopper (1991: 22) in that it is restricted to morphosyntactic properties, that is, unlike that of Hopper it does not include meaning or function. 8. However, ‘optional placement’ does not apply to the quotative construction, where use of ideophones is obligatory (Section 1.3).

Chapter 9. On the grammaticalization of ideophones

To conclude, while avoiding the term ‘grammaticalization’, Dingemanse and Akita (2017: 507–11) use two of the four parameters distinguished in (8), namely decategorialization and erosion, to describe what they call ‘deideophonization’ and what is called here the grammaticalization of ideophones. (9) Features of expressiveness (Dingemanse & Akita (2017: 507–11) a. Intonational foregrounding (e.g., pitch), b. Phonational foregrounding (e.g., breathy voice, creaky voice, whisper), c. Expressive morphology (e.g., reduplication, lengthening)

Applying the parameters in (8) to ideophones suggests that the structure of ideophones can be described along a scale extending from non-grammaticalized forms at the initial stage to fully grammaticalized forms at the final stage of grammaticalization. Table 3 lists the grammatical features characterizing these two stages. Table 3. The initial and final stages to be expected in the grammaticalization of ideophones (Heine 2023, Table 4.9) Initial stage

Final stage

General

The ideophone forms a distinct unit of discourse

The ideophone is integrated within word classes of sentence grammar

Function

Expressive

Non-expressive

Meaning

Does not change the clausal meaning

Contributes to the meaning of a clause

Syntax

Unattached

Constituent of a clause

Morphology

Invariable

May take inflectional and/or derivational morphology

Prosody

Prosodically distinct, forms an intonation unit of its own

Integrated in the intonation contour of the clause

Placement

Free, can form an utterance of its own

Restricted to slots provided by its morphosyntactic status as a noun, verb, adjective, or adverb

Discourse functions

Resistance to being negated or interrogated

Can freely be negated and interrogated

There are a few languages displaying simultaneously the initial and the final stages of evolution. Yucatec Mayan of Mexico, as sketched by Dingemanse and Akita (2017: 506), seems to be one of them, showing the two endpoints in the development of ideophones: At one end, ideophones are characterized by syntactic independence, expressive morphology, and marked prosody – all features suggestive of the initial stage. At the other end, ideophones seem to lack these properties, being morphosyntactically integrated, that is, strongly grammaticalized. They are

247

248

Alexander Andrason & Bernd Heine

“instantiated as verbs” showing aspectual marking – features to be expected in the final stage of development where ideophones are fully integrated into the structure of a sentence. However, in most languages the situation is more complex. Rather than distinguishing neatly between an initial and a final stage, one is more likely to find ranges of intermediate usages of ideophones, being located somewhere along the scale between the two stages. In the next sections, such a situation is looked at in more detail.

3.2 Ideophones in Siwu The Siwu language of southeastern Ghana belongs to the Kwa branch of the NigerCongo languages of Africa. In his description of ideophones in Siwu, Dingemanse distinguishes five morphosyntactic constructions (Dingemanse 2011: 145–8, 2017a: 366–70, 2017b). These constructions and their frequency in discourse are listed in Table 4. Ignoring the category ‘Others’, which is not further described,9 the five constructions are illustrated in (10) (ideophones are printed in bold). Note that these are morphosyntactic constructions of sentence grammar rather than discourse constructions like the ones in Table 1. Table 4. Ideophone constructions in Siwu (Dingemanse 2011: 145–8, 2017a: 366–70, 2017b) Construction (Dingemanse 2017a)

Number of tokens

Percentage

Holophrase

27

12 %

Complement

56

26 %

Adverbial

101

46 %

Adjectival

13

6%

Predicative

11

5%

Others

11

5%

219

100 %

Total

9. The category ‘Others’ “lumps together a small number of cases of ideophones occurring in syntactic environments that do not slot neatly into the five types described above” (Dingemanse 2017a: 370).

Chapter 9. On the grammaticalization of ideophones

(10) Constructions of ideophones in Siwu (Kwa, Niger-Congo; Dingemanse 2017a: 366–370) a. The holophrase construction10 Alɛ Kàntɔ kùgɔ ɔ̀-sɛ ɔ̀-bara ũ ara lo. ↑Tsintsintsintsintsin!↑ like Kàntɔ how 3.sg-hab 3.sg-do his things fp neatly.ide.emph ‘Just like Kàntɔ, the way he does his stuff. Tsintsintsintsintsin! [neatly]’ b. The complement construction kà ì-bara gelegele~gelegele ing nc-do ide.shiny~emph ‘It’ll be gelegele-gelegele [shiny].’ (Lit. ‘It does gelegele-gelegele‘) c. The adverbial construction i-tsi Si i-fudza-ɔ ↑fututu~tutututu↑ nc-head if s.nc-be.white-sg.o ide.pure.white~emph ‘That your head may become white fututu-tutututu [pure white].’ d. The adjectival construction bo-nà ɔ̀-turi gbógbóró-à kere. 1.pl-get ɔ-person ide.tough-adj just ‘We’ve got a sturdy person here.’ e. The predicate construction ( = (7))11 […] ì-i-kpokporo. it-neg-ide.be:hard ‘It isn’t kpokporo [hard].’

It would seem that the five morphosyntactic constructions can be related to the discourse constructions proposed in Table 1 of Section 1.3. As we saw there, ideophones are presented as independent utterances (e.g., Thud!) in the free construction. This applies to the holophrase construction in (10a): Tsintsintsintsintsin is a full-fledged ideophone corresponding to our definition of interactives, functioning as a “microscopic sentence” (Dingemanse 2017a: 367–368; see below). At the same time, (10a) can equally well be interpreted as an instance of the modifier construction. In the quotative construction of Section 1.3, ideophones are presented as discourse complements of quotative clauses introduced by verbs like ‘say’, ‘do’, ‘go’, or a quotative marker (e.g., And the vase went, thud!). We argue that this also applies to Dingemanse’s complement construction illustrated in (10b). In this language, verbs introducing the ideophone are se ‘be’, ba ‘have’, bara ‘do’ or nyɔ ‘look’. In Example (10b) it is the verb bara ‘do’ which functions like a quotative index 10. Here and elsewhere, arrows (↑) mark the beginning and the end of prosodic foregrounding, such as markedly high pitch. 11. Whether gbogboro in (10d) and ‑kpokporo in (10e) are related forms does not become entirely clear. Note that the former is an adjective and the latter appears to be a verb of state.

249

250

Alexander Andrason & Bernd Heine

(cf. ‘It does gelegele-gelegele’). We will, however, not deal with this construction any further here because it does not become entirely clear from the data available whether gelegele-gelegele is really a full-fledged ideophone unaffected by grammaticalization. This is different in the case of the remaining examples (10c–e). What characterizes these three examples is, first, that they can be interpreted as instances of the modifier construction. As we saw in Section 1.3, in this construction the ideophone is added to the preceding text as a kind of appositional parenthetical forming an independent utterance (e.g., The vase fell to the ground – thud!). In fact, the free constructions in Example (10) can alternatively be interpreted also as a modifier construction where the ideophone tsintsintsintsintsin is added to the preceding text piece as an independent utterance as a kind of appositional parenthetical. The examples in (10c–e), in contrast, are analyzed best as grammaticalized versions of the modifier construction. Thus, ignoring (10b), we hypothesize that the examples in (10) instantiate a scale of grammaticalization. In this scale, each of the constructions distinguished by Dingemanse (2017a) represents a different stage of development. The scale can be described in terms of a four-stage scenario of the following kind:12 Stage I. The free construction This is the stage where there is a full-fledged ideophone, as tsintsintsintsintsin is in (10a). It corresponds to our definition of interactives in (5): The ideophone functions as an independent unit, not being part of the preceding utterance, syntactically, semantically and prosodically complete, forming an intonation unit of its own. Stage I ideophones are not only “devoid of syntactic integration” but also “feature expressive morphology and prosodic foregrounding” (Dingemanse 2017a: 373). Stage II. The adverbial construction This can be called a stage of early grammaticalization. In accordance with the grammaticalization parameter of context extension in (8a), the use of the ideophone, like fututu-tutututu in (10c), has been extended to function as a modifier of the predicate, thereby losing its status as a syntactically free form. Desemanticization had the effect that semantically as well, it lost its independent status, now used to provide a depictive rendering of the scene described by its T argument in the preceding predicate phrase – that is, its meaning is dependent on that of the predicate. Via decategorialization, the ideophone lost its morphosyntactically independent status by gradually turning from a discourse modifier into an adverbial modifier of the main predicate. However, 12. As was just noted, we are ignoring here the construction in (10b).

Chapter 9. On the grammaticalization of ideophones

grammaticalization has not been highly advanced; the ideophone is still part of what Dingemanse (2017a: 370) calls a ‘free construction’. For example, the ideophone has retained a large amount of expressive and syntactic freedom. Stage III. The adjectival construction The ideophone undergoes further decategorialization by losing further independence, shifting from what Dingemanse (2017a: 370) calls ‘free construction’ to ‘bound construction’: It may now bear inflectional morphology and undergo regular processes of tonal change. The ideophone gbogboro ‘tough’ has given rise to the adjective gbogboro-à ‘tough’, illustrated in (10d), it functions very much like an “ordinary derived adjective” directly modifying a noun. The adjectival suffix ‑à, used to derive adjectives from stative verbs, thus has been extended to the ideophone. This is suggestive of the grammaticalization parameter of context extension whereby the use of the ideophone was extended to the domain of stative verbs, thereby sharing morphological features with the latter. Furthermore, grammaticalization affected the tonal structure of the form via the parameter of erosion: The ideophone gbogboro ‘tough’ lost its independent tone form: It was adapted to its tonal environment, in the present example rising to extra-high tone following the ‘raising noun’ ɔ̀-turi ‘person’ in (10d). Finally, in accordance with the parameters of desemanticization and erosion, the ideophone also lost its ability to use expressive morphology and prosodic foregrounding, respectively. But grammaticalization does not seem to have been concluded in that the ideophone has not been fully integrated in its new morphological class. For example, the ideophone uses the adjectival suffix ‑à often but not always, which may be taken to signal that morphological integration is not (yet) complete – the suffix is still optional, it has not yet become an obligatory part of the ideophone. Stage IV. The predicate construction Having already turned into a bound construction at Stage III, at Stage IV the ideophone has been grammaticalized further, turning largely into a unit of sentence grammar, which is that of a verb in the case of kpokporo in (10e). It underwent decategorialization to the extent that it lost most of the morphosyntactic independence characteristic of ideophones: Like normal verbs, it now takes verbal derivations and inflections for subject agreement, aspect, and negation, as kpokporo does in (10e). And via the parameter of erosion, it has lost much or all of its prosodic distinctiveness. On the other hand, decategorialization also had the effect that the erstwhile ideophone no longer takes a

251

252

Alexander Andrason & Bernd Heine

T argument, that is, it no longer needs a preceding text piece as an argument, now acting as a predicate of sentence grammar. Importantly, a Stage IV ideophone shares with Stage III ideophones that it lacks a number of other salient properties of Stage I ideophones. In particular, it lost the ability to foreground information prosodically and to present new information and, even more importantly, it is no longer immune to being negated and to being questioned. Note that “Siwu ideophones are not negated, nor are they used in question formation” (Dingemanse 2017a: 371) – that is, Stage I ideophones of Siwu conform to the definitional properties of interactives in (5) of Section 2.1. These properties are lost in Stage IV ideophones. In short, Stage IV ideophones are fairly deeply integrated in the structure of the clause. Nevertheless, even strongly grammaticalized elements are likely to retain some features of their earlier category status, bearing witness to their origin. To conclude, the findings presented in this section suggest that the analysis proposed by Dingemanse (2017a) can be re-analyzed in terms of a scale of grammaticalization of one particular discourse construction, viz. the modifier construction. These findings are summarized in Table 5. Table 5. Discourse constructions of Siwu ideophones Morphosyntactic construction of Dingemanse (2017a)

Discourse construction

Grammaticalization

Example

Holophrase

Free construction, or modifier construction

Stage I (nongrammaticalized)

(10a)

Complement

Quotative construction

?

(10b)

Adverbial

Modifier construction

Stage II

(10c)

Adjectival

Modifier construction

Stage III

(10d)

Predicative

Modifier construction

Stage IV

(10e)

3.3 Ideophones in Xhosa In their grammaticalization behavior, ideophones in Siwu are far from presenting an isolated case. The present section looks at another African language, namely the Bantu language Xhosa of South Africa. To this end, two of the discourse constructions proposed in Table 1 are looked at in more detail, namely the quotative and the modifier constructions.

Chapter 9. On the grammaticalization of ideophones

3.3.1 The quotative construction All ideophones in Xhosa may be introduced by quotative or performative markers, that is, verba dicendi and verba facendi (-thi and tsho, both meaning ‘say/ do’) – in accordance with the quotative construction in Table 1 of Section 1.3.13 Xhosa distinguishes the three variants in (11) to instantiate the quotative construction, as illustrated there. In (11a), the ideophone mhu-u-u is ‘performed’ as if it were imitating the actual sound made by a cow. In accordance with the initial stage of Table 3, the ideophone is subject to expressive manipulation (see (9)) and is syntactically unattached and prosodically set off by a pause. (11b) illustrates what we suggest is a variant of the quotative construction, differing from the latter in lacking a quotative index. In the third variant of the quotative construction, the ideophone precedes rather than follows its quotative clause, hence appearing utterance-initial, cf. (11c). (11) Three quotative constructions of ideophones in Xhosa (Bantu, Niger-Congo; Alexander Andrason, own data) a. Inkomo i-th-e , mhu-u-u mhu-u-u mhu-u-u nc.cow sa-say/do-perf [pause] ide.mooo ide.mooo ide.mooo ‘The cow [said] “mooo!”’ b. Inkomo , mhu-u-u! mhu-u-u! nc.cow [pause] ide.mooo ide.mooo ‘The cow [said] “mooo!”’ c. Ngxingxilili , uloliwe ide.stop [pause] nc.train ‘The train has stopped’

All three constructions in (11) conform to our definition of ideophones in (1) and to the set of features in Table 3 characterizing the initial stage in the grammaticalization of ideophones – that is, non-grammaticalized ideophones. But the quotative construction of Xhosa has also been grammaticalized and, as is to be expected in grammaticalization, the non-grammaticalized and the grammaticalized constructions co-exist side by side in the present usage of the language. The following reconstruction is confined to the constructions in (11a) and (11b), thus ignoring the construction in (11c), for which there are no appropriate data. It is hypothesized that (12a) is a grammaticalized form of (11a), and (12b) is a grammaticalized form of (11b). What the process from the structures in (11) to the ones in (12) must have involved is the following changes: First, the ideophones lost their status as independent utterances, turning into predicates of the 13. The performative meaning of thi/tsho is concrete, unlike English do, which may also be used as a desemanticized auxiliary.

253

254

Alexander Andrason & Bernd Heine

respective clauses (decategorialization), second, they also lost their prosodic distinctiveness, that is, the pause feature separating them from the preceding clause (erosion), third, they also lost their ability to mark expressiveness (see (9) above), such as being lengthened, or being accompanied by gestures. Finally, grammaticalization was not restricted to the ideophones; rather it affected the construction as a whole, leading from a complex construction [quotative clause – ideophone] to a mono-clausal construction [subject – predicate]. It is especially the erstwhile quotative verb ‑thi or ‑tsho (cf. i-th-e in (11a)) which is affected by grammaticalization. It lost part of its lexical meaning in favor of its new function of introducing an ideophone-derived predicate (desemanticization). And the original agglutinative structure w-a-thi (NC-PAST-say/do) is fused into the more synthetic form wee, that is, it underwent decategorialization (loss of boundaries separating the three morphemes) and erosion (loss of consonant and merger of a-i to ee). (12) Verbal predicate derived from ideophone in Xhosa (Bantu, Niger-Congo; Alexander Andrason, own data) a. Inkomo i-th-e mhuu nc.cow sa-say/do-perf ide.mooo ‘The cow [said] “mooo!”’ or ‘The cow mooed’ b. Inkomo mhuu nc.cow ide.mooo ‘The cow [said] “mooo!”’ or ‘The cow mooed’

But grammaticalization is far from complete. While being verb-like predicates in (12), the ideophones retained their invariable format, that is, they did not (yet) acquire salient features of Xhosa verbs. Most of all, they can neither take inflections and derivations nor can they take a negation marker or be interrogated. In sum, grammaticalization of Xhosa ideophones in the quotative construction would seem to have led to a change in the direction form ideophone to (deficient) verb, but it is far from having reached the final stage in Table 3. 3.3.2 The modifier construction As we saw in Table 1, the ideophone is placed in the modifier construction after the text piece it refers to as a kind of appositional parenthetical, set off syntactically and prosodically, serving as a modifier of that text piece. The ideophone dyumpu, expressing the idea of splashing in (13), illustrates this construction. The ideophone exhibits marked prosody, being separated by a pause and realized with unusual loudness and articulatory energy, often attesting to a high extent of expressiveness (e.g., the formative vowels are lengthened, and the form tends to be repeated), and

Chapter 9. On the grammaticalization of ideophones

it is accompanied by distinctive gesturing. The ideophone in (13) thus corresponds to the definition in (1) and exhibits the features of the initial stage in Table 3. (13) The modifier construction of ideophones in Xhosa (Bantu, Niger-Congo; Alexander Andrason, own data) Ndi-w-e e-manzi-ni , dyumpu! 1sg.sa-fall-perf loc-nc.water-loc [pause] ide.splash ‘I have fallen in the water, splash!’

Superficially, the construction illustrated in (14) looks much the same as that in (13); as a matter of fact, however, it is quite different. (14) presents the grammaticalized ideophone construction, differing from (13) most of all in the following: Rather than forming an utterance of its own, dyumpu in (14) has the appearance of an adverb, rendered in the translation of (14) by ‘in a splashing manner’. And in fact, this change can be hypothesized to be the result of grammaticalization processes. First, dyumpu lost its syntactic independence (decategorialization), being integrated in the structure of its host clause. Second, it underwent erosion: The prosodic boundary separating it from the host clause is lost, and dyumpu also lost much of its prosodic and phonation-related expressiveness (see (9)). (14) Adverbial modifier derived from ideophone in Xhosa (Bantu, Niger-Congo; Alexander Andrason, own data) Ndi-w-e dyumpu e-manzi-ni 1sg.sa-fall-perf ide.splash loc-nc.water-loc ‘I have fallen in a splashing manner in the water’

While having changed from independent ideophone towards adverb, the shift from interactive to category of sentence grammar is far from having being concluded, in that dyumpu differs from canonical adverbs in not taking the overt adverbial prefix ka-, thus, *kadyumpu would be ungrammatical – dyumpu, as well as other ideophones having made use of the modifier construction, are barred from using the prefix. To conclude, like ideophones grammaticalizing via the quotative construction (Section 3.3.1), those using the modifier construction moved towards a category of sentence grammar but did not reach the endpoint of the process – the former have not become full-fledged members of the class of verbs and the latter and the latter are not full-fledged adverbs. Thus, neither of the two source constructions has given rise to the structure described in Table 3 as the final stage of ‘deideophonization’.

255

256

Alexander Andrason & Bernd Heine

3.4 Discussion The pervasive grammaticalization of ideophones towards verbs, adverbs, adjectives, etc. seems to stem from the fact that virtually any ideophone may, from the beginning of its grammatical life, be recruited for this purpose. The analysis presented in Sections 3.2 and 3.3 shows that not much is gained by reducing grammaticalization to its initial and its final stage, as we did in Table 3: Much of the behavior of ideophones, captured by Dingemanse (2017a) with reference to the adverbial, the adjectival and the predicate constructions, concern usages of ideophones that are intermediate between the initial and the final stages – be that in Siwu or in Xhosa. Grammaticalization is a gradual process and the four-stage scenario proposed for Siwu in Section 3.2, and also the differential behavior of ideophones in Xhosa captures but a few structures along the scale or chain of grammaticalization. Note further that with the rise of a new stage, the earlier stage does not disappear but rather is likely to continue to be used. Thus, we saw that in the Siwu language all four stages distinguished co-exist side by side in present-day language use in the form of separate constructions, and exactly the same applies to the two constructions in Xhosa. The observations made here confirm findings presented earlier by Dwyer and Moshi (2003): Their distinction between a ‘primary’ and a ‘grammaticalized’ ideophone roughly corresponds to that between Stage I and all the other stages, respectively. Since grammaticalization takes place in specific contexts one will expect that a given ideophone is less strongly grammaticalized in some contexts but more strongly grammaticalized in other contexts. Thus, rather than with a simple distinction between a primary (non-grammaticalized) usage and a grammaticalized usage of the expression concerned, we are confronted with a complex situation where a range of different, contextually defined usages are found in the language concerned, bearing witness to the various effects of one general process of grammaticalization. And these observations also conform to the hypothesis proposed by Dwyer and Moshi (2003) according to which the development leads from an expressive to an analytic dimension of meaning: The more ideophones are grammaticalized, the more they lose their expressive force and their status as ‘expressives’ (Kaplan 1999). A slightly different interpretation is proposed by Dingemanse (2011: 150) in what he refers to as a process of deideophonization. Rather than grammaticalization, he draws on the terms ‘taming’ and the lexicalization of ideophones:14 14. Concerning ‘lexicalization’ as a diachronic process, see Brinton & Traugott (2005). This issue is taken up at the end of this section.

Chapter 9. On the grammaticalization of ideophones

This implies a lexicalisation scenario. Ideophones, normally free, can be tamed. They can be pressed into service as ordinary words. The more frequently an ideophone is used, the easier it will be to recruit it as an ordinary word. (Dingemanse 2011: 150)

The term ‘grammaticalization’ is also avoided by Haiman (2018: 104), who uses the term ‘domestication’ of ideophones instead: “I will identify as the result of ‘domestication’, the process whereby ideophones become assimilated to prosaic language”, where ‘prosaic language’ roughly corresponds to the notion of sentence grammar used here. Haiman (2018: 268) also characterizes the grammaticalization process as “a whirlwind tour of a handful of languages with ideophones in different stages of assimilation or domestication, from ‘rebel’ to ‘tame’ status”. A number of examples of ‘domestication’ are discussed in Haiman (2018: 268–81). Terms such as ‘taming’ and ‘domestification’ are intuitively appealing to refer to specific manifestations of change in ideophones. They are, however, not adopted here for the following reasons. First, the terms raise the question of how they are defined relative to other kinds of grammatical change. Second, they imply that there are linguistic units that can be classified as ‘untamed’ or ‘undomesticated’ and it does not become entirely clear how such units can be described. And finally, and most importantly, we argue that the changes to be observed in ideophones can be accounted for cross-linguistically in terms of grammaticalization in much the same way as changes in other kinds of linguistic forms. Thus, we are dealing with a more general process – one that has been observed to be at work across grammatical categories and languages, as has been demonstrated in Kuteva et al. (2019).15 This fact does not seem to be captured by terms such as ‘taming’ and ‘domestification’. But the development of ideophones is of a special kind. Grammaticalization is usually defined as leading from lexical to grammatical and from grammatical to even more grammatical forms (e.g., Narrog & Heine 2021). In the case of ideophones, however, the development leads from syntactically and prosodically detached forms to lexical forms rather than to grammatical forms. It can therefore be argued that rather than grammaticalization, the process is one of lexicalization.16 The reasons for insisting that we are nevertheless dealing here with the former are of two kinds. For one thing, the process examined, leading from grammatical independence to grammatical integration within sentence grammar, is in accordance with parameters of grammaticalization rather than of lexicalization, as we saw in Section 3.1. And for another, this process is of a different kind than processes that are traditionally captured by the label ‘lexicalization’ (see 15. In Kuteva et al. (2019), 544 pathways of grammaticalization in more than one thousand languages are documented. 16. We are grateful to Seongha Rhee (personal communication) for this suggestion.

257

258

Alexander Andrason & Bernd Heine

Brinton and Traugott 2005). Ignoring lexicalization as a synchronic phenomenon, diachronic lexicalization is in most cases of two kinds: It involves either fusion, that is, loss of boundaries between separate morphological elements leading to fusion and new lexical items. Alternatively, it leads from grammatical to lexical elements – that is, it constitutes a reversal of grammaticalization and is therefore widely known as ‘degrammaticalization’.17 Neither of the two processes is involved here and we therefore follow Dwyer & Moshi (2003) in treating the development from ideophone to verb as a special line of grammaticalization. This special line, leading from grammatical independence to grammatical integration within sentence grammar, has so far not been documented (but see Heine 2023, Section 4.2). It has in particular the following effect: Whereas in ‘canonical’ grammaticalization, words like nouns and verbs tend to lose their inflectional and derivational morphology on the way to becoming grammatical markers, the opposite can happen in ideophones. As we saw in Section 3.2, for example, ideophones are invariable expressions, taking no affixes. But once they are grammaticalized to verbs they are likely to acquire the morphological attributes of verbs, such as taking inflectional and derivational affixes.18

4.

Conclusions

The main point made in this paper is nicely summarized by Dingemanse (2017b: 198) in the following way: The tell-tale signs of depiction that occur when ideophones are morphosyntactically independent all disappear when ideophones lose their freedom and are assimilated to become more like normal words. (Dingemanse 2017b: 198)

The paper is based on a prototype framework, its goals were, first, to show that grammaticalization theory provides a tool for constructing a diachronic profile of ideophones and, second, to understand why ideophones behave structurally and typologically the way they do. With regard to the first goal, attention was drawn in Section 3 to the observation made by Dingemanse (2017b: 197) and Dingemanse and Akita (2017) according to 17. An example of fusion would be English hand in cap > handicap, while the latter includes ‘clipping’, e.g., when an affix turns into a lexical item, cf. English ex- > ex, as in My ex takes care of the children. Clipping though is not included by Brinton and Traugott (2005: 98) in lexicalization. 18. An anonymous reviewer of the paper makes the thought-provoking suggestion that the grammaticalization of ideophones might be interpreted as one of ‘predicatization.’

Chapter 9. On the grammaticalization of ideophones

which ideophones show “an inverse relation between prosodic foregrounding and morphosyntactic integration: Ideophones that are more deeply integrated in the structure of the clause lose their prosodic foregrounding”: Why do expressiveness and morphosyntactic freedom go together so naturally in ideophones? Conversely, why do ideophones tend to lose their expressiveness when they are more deeply integrated in the utterance? (Dingemanse & Akita 2017: 526)

As we saw in Section 3, there is an explanation for this fact: It is a predictable effect of grammaticalization and can be accounted for in terms of the parameters presented in (8) of Section 3.1. Thus, in accordance with the grammaticalization parameter of decategorialization, an ideophone undergoing grammaticalization is likely to lose much of its morphosyntactic freedom, that is, its detached syntactic status (cf. Section 1.1) and to be morphosyntactically integrated in the structure of a sentence. And in accordance with the parameter of erosion, that ideophone can also be expected to lose part or all of its expressiveness, such as prosodic and phonational foregrounding (see (9)). As was observed in Section 3.4, however, the development from ideophone to words of sentence grammar is of a different kind than that of grammaticalization as it has been described so far, resulting in the rise of new lexical rather than grammatical forms. With regard to the second goal, the findings presented account for the fact that ideophones exhibit a highly complex structural outfit. On the one hand, they may show most or all of the properties defining them as interactives, cf. (5) of Section 2.1. On the one other hand, they may exhibit many of the structural properties commonly found in words of sentence grammar such as verbs, adjectives and adverbs. From the perspective of grammatical typology, these observations suggest that ideophones are a highly unusual grammatical category. For one thing, their membership includes, on the one hand, free, syntactically and prosodically detached forms belonging to interactive grammar. On the other hand, it also includes forms that may be largely or entirely integrated in the structure of sentence grammar, being part of the syntactic hierarchy of a clause. And for another, providing a grammatical description of ideophones in terms of a fixed set of properties that applies across languages is a task that is close to impossible, for the following reason: Languages differ greatly from one another in the way and the extent to which they have grammaticalized their ideophones. This means, on the one hand, that much of what characterizes the structural behavior of ideophones crosslinguistically is located simultaneously in two different domains of grammar, namely sentence grammar and interactive grammar (Heine 2023). On the other

259

260

Alexander Andrason & Bernd Heine

hand, it also means that there are presumably no two languages exhibiting exactly the same grammatical structure in all their ideophones. To be sure, this problem does not exist if comparative typology is restricted to non-grammaticalized ideophones, that is, uses of ideophones conforming to the grammatical features of what was described in Table 3 as the initial stage of ideophone development. This is the procedure adopted in most studies of ideophones, and the procedure has in fact made it possible to identify ideophones as a distinct type of interactive grammar. But an important task for future research is it to come to grips with the overall grammatical structure of ideophones, that is, with the use of ideophones across linguistic discourse. The present paper could achieve hardly more other than pointing out problems that are associated with such a task.

Abbreviations 1, 2, 3 1a adj emph fp hab ide ing loc

first, second, third person noun class 1a adjective emphasizer final particle habitual ideophone ingressive locative

nc neg o past perf pl s sa sg

noun class marker negation marker object marker past perfect plural subject marker subject agreement singular

References Andrason, Alexander. 2020. Ideophones as linguistic “rebels” – The extra-systematicity of ideophones in Xhosa. Part 1. Asian and African Studies 29(2): 119–165. Andrason, Alexander. 2021. Ideophones as linguistic “rebels” – The extra-systematicity of ideophones in Xhosa. Part 2. Asian and African Studies 30(1): 1–30. Andrason, Alexander & Dlali, Mawande. 2020. The (crucial yet neglected) category of interjections in Xhosa. STUF – Language Typology and Universals 73(2): 159–217. Andrason, Alexander, Phiri, Admire & Fehn, Anne-Maria. M. forthcoming. Onomatopoeias in Tjwao. Canadian Journal of Linguistics. Brinton, Laurel J. & Traugott, Elizabeth Closs. 2005. Lexicalization and Language Change. [Research Surveys in Linguistics]. Cambridge: CUP. Clark, Herbert H. 2016. Depicting as a method of communication. Psychological Review 123(3): 324–347.

Chapter 9. On the grammaticalization of ideophones

Clark, Herbert H. 2019. Depicting in communication. In Human Language: From Genes to Behavior, Peter Hagoort (ed.). Cambridge MA: The MIT Press. Coulmas, Florian (ed.). 1981. Explorations in Standardized Communication Situations and Prepatterned Speech, Vol. 2: Conversational Routine [Janua Linguarum. Series Maior, 96]. Berlin: De Gruyter Mouton. Diffloth, Gérard. 1972. Notes on expressive meaning. Papers of the Chicago Linguistic Society 8: 440–447. Dingemanse, Mark. 2011. The Meaning and Use of Ideophones in Siwu. PhD dissertation, Radboud University, Nijmegen. Dingemanse, Mark. 2012. Advances in the cross-linguistic study of ideophones. Language and Linguistics Compass 6(10): 654–672. Dingemanse, Mark. 2013. Ideophones and gesture in everyday speech. Gesture 13(2): 143–165. Dingemanse, Mark. 2015. Ideophones and reduplication: Depiction, description, and the interpretation of repeated talk in discourse. Studies in Language 39(4): 946–970. Dingemanse, Mark. 2017a. Expressiveness and system integration: On the typology of ideophones, with special reference to Siwu. STUF – Language Typology and Universals 70(2): 363–384. Dingemanse, Mark. 2017b. On the margins of language: Ideophones, interjections and dependencies in linguistic theory. In Dependencies in Language, Nicholas J. Enfield (ed.), 195–202. Berlin: Language Science Press. Dingemanse, Mark. 2018. Redrawing the margins of language: Lessons from research on ideophones. Glossa: A Journal of General Linguistics 3(1): 1–30. Dingemanse, Mark. 2019. ‘Ideophone’ as a comparative concept. In Ideophones, Mimetics, and Expressives [Iconicity in Language and Literature 16], Kimi Akita & Prashant Pardeshi (eds.), 13–33. Amsterdam: Benjamins. Dingemanse, Mark. 2021. Ideophones. In The Oxford Handbook of Word Classes, Eva van Lier (ed.). Oxford: OUP. Dingemanse, Mark & Akita, Kimi. 2017. An inverse relation between expressiveness and grammatical integration: On the morphosyntactic typology of ideophones, with special reference to Japanese. Journal of Linguistics 53(3): 501–532. Doke, Clement Martyn. 1935. Bantu Linguistic Terminology. London: Longman. Dwyer, David & Moshi, Lioba. 2003. Primary and grammaticalized ideophones. In Linguistic Typology and Representation of African Languages, John M. Mugane (ed.), 173–185. Trenton NJ: Africa World Press. Evans, Vyvyan & Green, Melanie. 2006. Cognitive Linguistics: An Introduction. Edinburgh: EUP. Ferguson, Charles A. 1976. The structure and use of politeness formulas. Language in Society 5: 137–151. Ferguson, Charles A. 1981. The structure and use of politeness formulas. (Reprint of Ferguson 1976). In Explorations in Standardized Communication Situations and Prepatterned Speech, Vol. 2: Conversational Routine [Janua Linguarum. Series Maior 96], Florian Coulmas (ed.), 21–36. Berlin: De Gruyter Mouton. Haiman, John. 2018. Ideophones and the Evolution of Language. Cambridge: CUP. Heine, B. 2023. The Grammar of Interactives. Oxford: OUP.

261

262

Alexander Andrason & Bernd Heine

Heine, B. & Kuteva, T. 2007. The Genesis of Grammar: A Reconstruction [Studies in the Evolution of Language 9]. Oxford: OUP. Hopper, Paul J. 1991. On some principles of grammaticization. In Approaches to Grammaticalization, Vol. 1: Theoretical and Methodological Issues [Typological Studies in Language 19: 1], Elizabeth Closs Traugott & Bernd Heine (eds), 17–35. Amsterdam: John Benjamins. Hopper, Paul J. & Elizabeth Closs Traugott. 2003. Grammaticalization. Cambridge: CUP. Ibarretxe-Antuñano, Iraide. 2017. Basque ideophones from a typological perspective. Canadian Journal of Linguistics 62(2): 196–220. Janda, Laura A. 2015. Cognitive linguistics in the year 2015. Cognitive Semantics 1: 131–154. Jendraschek, Gerd. 2012. A Grammar of Iatmul. Habilitationsschrift, Universität Regensburg. Kaplan, David. 1999. The meaning of ouch and oops: Explorations in the theory of meaning as use, transcribed by Elizabeth Coppock. Howison Lecture in Philosophy, delivered at the University of California at Berkeley. Kita, Sotaro. 1997. Two-dimensional semantic analysis of Japanese mimetics. Linguistics 35: 379–415. Koo, Hyun Jung & Rhee, Seongha. 2018a. Ideophones and attenuatives in Korean. Paper presented at the 51st Annual Meeting, Societas Linguistica Europaea, Tallinn University, Estonia, 29 August–1 September 1. Kuteva, Tania, Heine, Bernd, Hong, Bo, Long, Haiping, Narrog, Heiko & Rhee, Seongha. 2019. World Lexicon of Grammaticalization, 2nd edn. Cambridge: CUP. Meinard, Maruszka Eve Marie. 2015. Distinguishing onomatopoeias from interjections. Journal of Pragmatics 76: 150–168. Narrog, Heiko & Heine, Bernd. 2021. Grammaticalization. Oxford: OUP. Nübling, Damaris. 2004. Die prototypische Interjektion: Ein Definitionsvorschlag. Zeitschrift für Semiotik 26(1–2): 11–46. Stange-Hundsdörfer, Ulrike. 2016. Emotive Interjections in British English. A Corpus-Based Study on Variation in Acquisition, Function and Usage [Studies in Corpus Linguistics 75]. Amsterdam: John Benjamins. Taylor, John R. 1995. Linguistic Categorization: Prototypes in Linguistic Theory. Oxford: OUP. Vendler, Zeno. 1967. Linguistics and Philosophy. Ithaca CA: Cornell University Press. Voeltz, F. K. Erhard & Kilian-Hatz, Christa (eds). 2001. Ideophones [Typological Studies in Language 44]. Amsterdam: John Benjamins.

chapter 10

An emerging final particle The case of quoi ‘what’ in French Sylvie Hancil

University of Rouen

Turn-taking is one of the contexts leading to innovative language use. In this article, I will explore the relationship between the emergence of grammatical patterns and intersubjectivity. Final particles belong to the category of linguistic expressions that generate the production of interactional grammar. After reviewing the various meanings of final quoi in French dictionaries, I will test the five principles of grammaticalization, namely layering, divergence, specialization, persistence and decategorization and I will try to show that the evolution of the particle cannot be restricted to the grammaticalization process. Keywords: final particle, quoi, intersubjectivity, grammaticalization, cooptation

1.

Introduction

Recent research in language change has acknowledged the importance of identifying pragmatic contexts which allow for new usages of a form (Traugott & Dasher 2002; Traugott 2010). Turn-taking (Detges & Waltereit 2011) is one of the contexts leading to innovative language use. In this article, I will explore the relationship between the emergence of grammatical patterns and intersubjectivity, which is defined as the encoding of “the Speaker’s attention to the cognitive stances and social identities of the Addressee” (Traugott 2012: 9), relating to aspects such as politeness, and turn taking as social norms in a more general perspective. From a specific cognitive perspective, intersubjectivity refers to the “mutual management of cognitive states” (Verhagen 2005: 3), that is, to the engagement of speakers in cognitive coordination with others. The notion of ‘grammar’ is understood here in a broader way (Traugott 2003: 626; Heine, Kaltenböck, Kuteva & Long 2013: 156) as a knowledge system that allows speakers to produce text, that is, a piece of meaningful language that serves a particular purpose in an individual context. https://doi.org/10.1075/slcs.232.10han © 2023 John Benjamins Publishing Company

264

Sylvie Hancil

Intersubjectivity is closely related to questions of politeness (face-concerns), syntactic position (the peripheries are more closely related to intersubjective meanings than clause-medial positions, see Beeching & Detges 2014), and social accommodation (convergence on each other’s usage patterns as a symbol of psychological approximation, which may trigger overuse and grammaticalization of particular patterns, Du Bois & Giora 2014), all of which may trigger the rise of innovative form-meaning pairings. These aspects have been largely neglected in grammaticalization theory, partly because language has long been seen predominantly from a structural perspective as a set of abstract rules and fixed structures. However, since language is a ressource for social interaction, the emergence of grammatical patterns often lies in language used in interaction and can thus not be studied independently of interactive aspects. Final particles are part of these linguistic expressions that participate in the production of interactional grammar. The purpose of this paper is to examine the use of the final particle quoi ‘what’ in French and evaluates its degree of grammaticalization in relation to Hopper’s (1991) five principles of grammaticalization. The paper is organized as follows. After reviewing the various meanings of quoi in French dictionaries, I will test the five principles of grammaticalization, namely layering, divergence, specialization, persistence and decategorization, and I will show that the evolution of the particle cannot be restricted to the grammaticalization process.

2.

Data

Most of the occurrences are identified in the French Web Corpus 2017. This French corpus comprises texts collected from the Internet. The corpus belongs to the TenTen corpus family which is a set of the web corpora built using the same method with a target size of 10-plus billion words. Data was crawled by the SpiderLing web spider in April 2012 and consists of almost 10 billion words. The corpus contains many varieties of the French language – European, Canadian and African French. Here it was chosen to focus on the French variety from France. There are 2,925,924,634 tokens, thereby almost 3 billion tokens. There are 2,023 occurrences of final quoi in this sub-corpus, hence 0.69 per million words.

3.

quoi in dictionaries

According to the Dictionnaire historique de la langue française, the use of quoi in initial position to express astonishment or indignation is ancient (around 1180),

Chapter 10. An emerging final particle

whereas its use in final position after a word, summarizing an enumeration, an idea, is quite recent (early 19th century), and regarded as colloquial. The Grand Robert indicates that the use of quoi in the sense of pardon ? is considered as rather impolite. Nevertheless, Goose (1993: 1073) notices that quoi with an indication of astonishment is used even in the formal style. Chanet (2001: 56) remarks that the particle quoi is not popular and that its role is to be an attempt to build up an intersubjective space by resorting to an inferential activity of the co-speaker. In the seventeenth- and eighteenth-century dictionaries, quoi, written quoy, is identified as an adverb or interjection. Thus, in Furetière (1690), the demonstrative quoy is also an adverb marking admiration and interrogation, as in “Quoy! / He quoy ! Quoy donc!” ‘Quoi! Eh quoy! Quoi, then!’ ; in Féraud (1787: 334), quoi is also an interjection employed in exclamations, as in “Quoi ! toute la nature ne frémit-elle pas! ” ‘Quoi! Does nature not thrill!’. In the nineteenth century, Littré (1876) explains the word in terms of interjection expressing indignation, and for Pierre Larousse (1889, Dictionnaire complet), the interjection quoi indicates astonishment, as in “Quoi ! vous partez?” ‘Quoi! You are leaving?’. It is worthwhile noting the change taking place in dictionaries of the second half of the twentieth century, in which quoi is systematically labelled as “pronoun”, in the section “interrogative” and “exclamative”. Nevertheless, some dictionaries sometimes mention its status of an interjection (Le Petit Robert 1967). Indeed, in the 1960s-1970s, lexicographers gradually extended the definition of uses of quoi. The values identified are more diverse and are classified as being colloquial (Petit Robert 1967; TLFi 1990), or elliptical uses of the spoken language (TFLi 1990). As a discourse particle, quoi does not have a very well defined status inside the parts of speech. The entry dates in dictionaries of this use of quoi indicate the beginning of the development of its use in contemporary French, in literary French, but also in spoken language, namely around the 1950s-1960s. The dictionary definitions show an evolution in the meanings and extend from the strictly interjective meaning, marking surprise or indignation at the end of an enumeration or and explanation, to a more enunciative and more dialogal value (see TLFi, to refute a possible objection, to establish a connivence with the co-speaker, relying on the common ground shared by the interlocutors.) Let us now test the five principles of grammaticalization (Hopper 1991).

4.

Hopper’s (1991) five principles of grammaticalization

The five principles of grammaticalization (Hopper 1991) will show that five semantic-pragmatic values can be identified: indefiniteness, summative, evaluative, especially intensifying and phatic.

265

266

Sylvie Hancil

4.1 Principle 1: Layering The principle of layering illustrates the fact that an emerging form does not automatically replace an existing expression for the grammatical phenomenon examined, thereby explaining that old and new forms co-exist, as illustrated in the Examples (1)–(3) where three markers can coexist: (1) bon c’est pas si grave (hein / n’est-ce pas / quoi) (ccc) ‘Well, it’s not so bad, hein / n’est-ce pas / quoi’ (2) oui : / et il fait beau aujourd’hui : (hein / n’est-ce pas / quoi) /// ah c’est agréable ce temps-là (ccc) ‘Yes, and it’s sunny today, hein / n’est-ce pas / quoi. Ah, this weather is nice.’ (3) fâcheux temps (n’est-ce pas / hein / quoi) (prf ) ‘Difficult times, n’est-ce pas / hein / quoi’

4.2 Principle 2: Divergence The principle of divergence shows that existing forms take on new meanings in specific contexts. A large variety of meanings is exemplified in the following examples, extending from resumptive expressions, via modalising expressions, to interactive and phatic expressions. First, the discourse particle quoi can occur with an explanation marked with slight impatience or with a word that sums up an idea or an enumeration (Grand Robert, TLFi), as in (4) : (4) Un peuple de candidats à la bourgeoisie, un peuple d’aspirants à la bedaine. Les pantoufles, quoi! (Larbaud). ‘A people of bourgeois candidates, a people of potbellied aspirants Slippers, quoi !’

Beeching (2004: 71) also notices that the particle quoi usually follows a reformulation or a final term of the summary of the preceding discourse, which corresponds to the process of production and planification explained in Chanet (2001 : 70) : (5) […] il y avait des sacs d’olives + pas des sacs + des cartons quoi + des + des cagettes + d’olives […] (Chanet 2001 : 70) ‘there were bags of olives + not bags + cardboard boxes quoi + some + some crates’

Chapter 10. An emerging final particle

Chanet (2001) proposes the following représentation: (6) il y avait des sacs d’olives pas des sacs des cartons quoi des des cagettes + d’olives (ibid.) ‘There were bags of olives not bags cardboard boxes quoi some some crates + of olives.’

The author also says that quoi is used in places where it is difficult to say if it is a reformulation or an enumeration. It is also employed to end an enumeration, whose last term has a pragmatically resumptive meaning (Chanet 2001: 72), as in (7)–(8) : (7) […] apparemment les gens sont sont sympas euh sincères euh humains quoi et pas du tout […] ‘Apparently people are are nice erm sincere, er human quoi, and not at all.’ (8) voilà serrurier menuisier euh + + enfin tous les métiers d’atelier quoi + (ibid.) ‘Right, locksmith, carpenter erm + + right all the workshop’s jobs, quoi.’

According to this linguist, quoi indicates all the other terms that could have been referred to instead in the list can be included in this class. quoi opens up a paradigm or a field of possible referents and incites the co-speaker to mobiliz their/his or her knowledge of the background to set up a schematisation. Lefeuvre (2006 : 113) also underlines the presence of a paradigm : La recherche de la bonne expression impliquée par quoi se traduit par la présence d’un paradigme qui comporte plusieurs segments correspondant à autant de formulations possibles. ‘The search for the right expression implicated by quoi is translated into the presence of a paradigm which comprises several segments corresponding to as many formulations as possible.’ F. Lefeuvre (2006: 113)

By citing the following Example (9) : (9) […] fils d’immigré, flic de gauche résolument, sensible, déchiré, rebelle. Attachant, quoi ! Izzo est mort et ne peut plus protester. ‘A son’s immigrant, resolutely a left-wing cop, sensitive, dishevelled, rebellious. Endearing, quoi ! Izzo is dead and cannot protest anymore.’

267

268

Sylvie Hancil

Lefeuvre (2006) proposes the following paraphrase quoi’s value of approximation as: « Attachant, c’est quelque chose comme cela » ‘Endearing, something like that.’

or as: « Attachant, c’est finalement ce que je pourrais dire de mieux de cet individu ; cela s’approche le mieux, mais pas complètement » (ibid.). ‘Endearing, that’s ultimately the best I could say about this individual ; it’s close the best, but not quite.’

Lefeuvre (2006) also mentions that quoi synthesizes the preceding formulations (ibid.). My analysis shows there is an opening of paradigms; however, the above paraphrase implies that quoi closes rather than opens a paradigm. Contrary to Chanet (2001), I think that the speaker introduces a paradigm and then ends this formulation with quoi. Second, according to Chanet (2001: 74), quoi signals that the speaker evaluates the representation that his speech is constructing and that he is wondering whether the information that he is offering is sufficient to allow the co-speaker to reconstruct this representation. Chanet cites operations of resumption accompanying quoi that ends an enumeration with a specific lexicon such as sorte de, genre de, espèce de, type de, style de ‘kind of, sort of ’, along with the operation of referencing made by the use of a blurred/vague categorizer (truc, chose, ‘thing’), with the use of a comparison such as comme ça ‘like that’, as in (10–(11) : (10) il faut pas il faut pas euh + mettre un une sorte de + + de jugement de valeur quoi euh dire […] ‘You should not you should not erm + give a sort of judgement of value quoi erm say …’ (11) […] c’est les boules Quiès euh des trucs comme ça quoi il y a rien d’autre ‘It’s the earplugs erm things like that quoi there’s nothing else.’

Lefeuvre (2006: 115) also notices that quoi does not allow a definitive validation by itself. Namely, quoi shows that the assertion cannot exist by itself and introduces some doubt, as shown in the following example where a phrase is repeated as the speaker does not find any better formulation : (12) il est engourdi il est engourdi quoi ‘He is numb he is numb quoi’

Chapter 10. An emerging final particle

Beeching (2004: 196) cites the following example to illustrate the value of the inadequation of the expression : (13) ménager des moments où on peut se se détendre un peu et faire autre chose, ne pas avoir que des contraintes dans dans sa vie, quoi, hein ? ‘A llowing time/Creating time when you can relax a bit and do something else, not having constraints in your life, quoi, huh ?’

I agree that quoi indicates that the speaker is making an evaluation (Chanet 2001: 78), which Lefeuvre calls the speaker’s reflective process about what he has just said (Lefeuvre 2006: 112). Nevertheless, it does not seem right to say that the speaker doubts adequacy of the information. This doubt can be implied, so it seems, by the expression quoi, but by using quoi, the speaker gives up their search for the right words and forces the co-speaker to accept what he has just said despite its possible incompleteness. The speaker may or may not be satisfied with his expression, but he puts an end to the quest with quoi : Third, quoi, when it is used at the end of an utterance, can also be employed to establish a bond with the co-speaker about the identification of what it is about (TLFi), as shown in (14): (14) Tout ce qu’ils possédaient, leur campagne, les charrettes, […] un chien avec sa chaîne, tout quoi (Céline, Voyage, 1932, p. 17). ‘Everything they possessed, their countryside/land, the carts, … a dog on a leash, everything quoi ‘

With regard to this interactive aspect, Chanet (2001: 74) notices that quoi signals that the speaker invites the co-speaker to resort to stereotypical knowledge (which is not provided by the discourse) to reconstruct this representation. It concerns examples of comparisons, most of the time using comme ‘like’, as in (15): (15) comme ceux que l’on achète dans le commerce quoi non ‘like those that we buy in shops quoi, right ?

Chanet (2001: 75) also mentions that by using quoi, the speaker invites the cospeaker to share the subjectivity of his or her positioning and that quoi conveys an implicit discourse of the type : « ce que je viens de dire, tu aurais sans doute pu le dire aussi » (id.: 76) ‘What I just said, you could have said it, too’ : 1) une « opération de quantification » marquée par un quantificateur « universel » (tous les), tout le temps, toujours, jamais (f ), ou 2) une « opération de graduation d’une échelle » portée par les termes tels que complètement, carrément, assez, trop, un peu (g) ou par des termes axiologiques, tels que

269

270

Sylvie Hancil

somptueux, fabuleux, atroces, ainsi que quoi 3) avec un vraiment évaluatif ou 4) avec des modalisateurs épistémiques (je pense (que), je trouve (que), peutêtre (h), des commentaires métadiscursifs tels disons (que)) (id. : 74–5). (f ) […] parce que la langue euh tu l’utilises tous les jours quoi + et euh […] (id. : 74) (g) […] c’est c’est trop risqué quoi c’est c’est vraiment risqué (id. : 75) (h) […] et puis d’autres ben euh moi j’aime pas les prononcer puis je trouve ça ça te rabaisse à quelque part quoi je sais pas + + (ibid.) Chanet (2001 : 76) ‘(1) an operation of quantification marked by a “universal quantifier” (everyone of them), all the time, always, never (coll.), or (2) an “operation of graduation of scale” indicated by the terms such as completely, really, enough, too much, a little (3) or by axiological terms, such as sumptuous, fabulous, terrible, alongside quoi with a real evaluative or 4) with epistemic modalizers (I think that, I believe that, perhaps (h), metadiscursive comments such as let’s say that (id. 74–5). (f ) […] because the language uhm you use it daily quoi (g) […] it’s it’s too risky quoi it’s it’s really risky (id. 75) (h) […] and then others well uhm I don’t like to pronounce them and I find that it it degrades you to something quoi I don’t know

I can add the case of exaggerated affirmation introduced by Beeching (2004: 196), as in (16) : (16) […] quand on voit les car-ferries qui passent c’est c’est superbe quoi, on adore ça, nous.[rires] ‘[…] when we see the ferries passing it’s it’s superb quoi, we love that, we. (laughter)’

According to Beeching (2004), the use of a universal quantifier favours the occurrence of quoi, since the speaker means that he cognitively scans the totality of a class of objects and that he cannot go any further in this scanning ; it is an idea of extremity that allows quoi to appear. Lefeuvre (2006: 112) denies Chanet’s position by saying that the role in relation to the co-speaker (invitation to sharing) identified by Chanet does not seem obvious, even in situations of dialogue. For instance, she notices that the example does not show that quoi is the vector of knowledge shared with the co-speaker and that it is non ‘no’ that makes the link with the co-speaker. We agree with

Chapter 10. An emerging final particle

Lefeuvre in so far as it is rather the expression cited by comme and the use of on which could indicate the stereotypical knowledge. But I agree with Chanet’s remark (2001: 79), that quoi is fundamentally « attentive » to the co-speaker and that it can also indicate the desire to see his or her own words echoing a possible discourse from the co-speaker. However, it seems to us necessary to distinguish being attentive to the co-speaker and inviting to share his or her attitude on what he or she has just said. By using quoi, does the speaker give the co-speaker the choice not to share his opinion ? Regarding this point of view, Morel and Danon-Boileau (1998: 102) notice that quoi means to the co-speaker that you offer your position to yourself and that it is not subject to discussion, namely the use of quoi leads to the speaker’s validation of the referential content which has been asserted and rejects any coenunciative sharing. The authors propose the paraphrase « je me donne comme étant seul à penser ce que je pense » ‘I am the only one to think what I think’ or « je clôture en disant que je donne un contenu qualitatif qui m’est personnel, dont je suis le seul à pouvoir énoncer les propriétés distinctives » ‘I end the speech by saying that I give a qualitative content that is mine and that I am the only one to be able to pronounce the distinctive properties’. Fourth, Chanet (2001: 77) describes the particle quoi as intervening in an argumentative strategy, most of the time a concessive one, as in (17): (17) moi je pense qu‟il existe euh un seul français mais qui peut avoir euh différents aspects quoi + enfin ça reste du français + […] ‘I think uhm one single French exists but it can have different aspects quoi + well it remains French.

This example can be compared to the Example (18) of “contradictions” cited by Beeching (2004: 196) (18) Je suis de nationalité française mais je suis très contente d’être bretonne, je suis fière d’être bretonne quand même, quoi ‘My nationality is French but I am very happy to be a Briton, I am proud to be a Briton anyway, quoi’

Fifth, Wilmet (2003: 286) contends that quoi is used as a landmark of spoken language, which is in accordance with Andrews’s discussion (1989: 196). The latter calls it a “terminator”, with no semantic or affective value and whose only function in speech is to signal the end. For Auchlin, it is even difficult to attribute to quoi, on the structural plane, other properties than that of punctuating the discourse by signaling through its position at end of a unit (E. Roulet et al. 1985: 102). Delierre (1997: 25) also considers the particle as a marker of oral punctuation.

271

272

Sylvie Hancil

Kerbrat-Orecchioni (2005: 33) classifies this use of quoi as punctuating with the help of alors ‘then’ and bon ‘well’ in the analysis of a plenary conference.

4.3 Principle 3: Specialization The principle of specialization leads to the singling out of one or a few forms for grammaticalization. When grammaticalization of the new form is complete, there should be no alternative choice for the marker examined: no interchangeable term should be available. Nevertheless, the principle is not fully supported by the data since the expression systematically has a synonym in final position. The synonym for an indefinite meaning is the expression en quelque sorte ‘sort of ’, as in (19) : (19) Dans le Positif de juin, je vois que les films étrangers sont cités sous leur titre français, jamais originaux. Est-ce grave? Suis-je un peu snob? < /s > < s > THE LINE-UP ça sonne quand même un peu mieux que LA RONDE DU CRIME, titre qui fait penser à 50 films différents! < /s > < s > D’autre part, dans la rubrique DVD, à propos de ce film de Siegel d’ailleurs excellent, Fabien Baumann désigne le personnage joué par Robert Keith (sans citer l’acteur luimême) par une expression vulgaire que je n’aime pas ne rapporte pas ici, et qui ostracise deux catégories de population. Tt d’un coup, je me suis retrouvé dans un article un peu “mode” d’un magazine un peu … je sais pas … un peu “magazine”, quoi. Est-ce moi qui débloque? (sacd.fr) ‘In the June issue of The Positive, I see that the foreign movies are listed under their French title, never the original. Is it bad ? Am I a bit snobbish ? The Lineup rather sounds a bit better than The Circle of the crime, a title that makes you think of 50 different movies ! Besides, in the DVD section, in this movie by Siegel which is by the way excellent, Fabien Baumann denotes the character played by Robert Keith (without quoting the actor himself ) with a vulgar expression that I don’t like and won’t repeat here, and which ostracizes two categories of the population. All of sudden, I find myself in an article that is a bit « fashionable » in a magazine that is bit … I don’t know …a bit “magazine” quoi. Maybe I’m exaggerating ?’

At the end of a list, the particle quoi can be paraphrased by enfin ‘in sum’, as in (20): (20) Sortie de l’Amphi, des gens qui passent à côté de vous sans un bonjour ni merde (alors qu’aujourd’hui, vraiment, un merde m’aurait fait plus plaisir), rentrée à l’appart’ avec un mal de tête naissant, la flemme qui m’a écrasée toute la journée. Un peu honte quand même de dire que je suis arrivée hier et que je n’ai toujours pas fait mon lit, ni la vaisselle, et je crois bien que quand j’ouvrirai les yeux demain, je serais carrément malade. J’ai faim et pas faim en même

Chapter 10. An emerging final particle

temps, et même la clope que je viens de griller à cet instant ne m’a pas motivée plus que ça. < /s > < s > J’aurais du bosser mes cours, lire, enfin, faire quelque chose de “bien”, quoi. Ou alors quelque chose tout court. (cowblog.fr) ‘Outside the amphitheater, people pass by you without saying hello or shit (whereas today, really, saying shit would have pleased me), back to the flat with a burgeoning headache, the laziness that crushed me all day long. A bit ashamed anyway to say that I arrived yesterday and still haven’t made my bed, or done the the dishes, and I do believe that when I open my eyes tomorrow, I will be quite unwell. I’m hungry and not hungry at the same time, and even the cigarette that I just smoked hasn’t given me any motivation. I should have studied, read, well, done something “well”, quoi. Or at least done something.’

or can be synonymous of bref ‘well, briefly’, as in (21) : (21) Alerte au Parlement! Il y a fort à parier que la présence des députés était étonnamment élevée en ce jour de séance. La raison de cette assiduité peu coutumière? La visite dans l’hémicycle de Pamela Anderson. L’ancienne pin-up aux formes parfaites, cheveux au vent, qui courait sur la plage, cintrée dans son maillot rouge, pour aller sauver au péril de sa vie les baigneurs imprudents. Bref, la star d’ “Alerte à Malibu”, quoi. Mais que vient faire à Paris, l’actrice canadienne de 49 ans aux traits figés par le Botox? Réanimer des élus privés d’air ? Leur proposer une formation de secouriste ? Rien de tout cela. (femmeactuelle.fr) ‘A lert/Newsflash: Parliament ! It is strongly believed that the presence of MPs was astonishingly high on that day of the session. What is the reason for such unfamiliar attendance? Pamela Anderson’s visit. The former pin-up of perfect porportions, hair windswept, running along the beach, clad in her red bathing suit, risking her life to save struggling swimmers. In a word, the Bay watch star, quoi. But what on earth is the 49-year-old Canadian actress with her Botox-frozen face doing in Parliament ? Resucitating MPs? Hosting a first aid course ? None of the above.’

or can be expressed by donc ‘consequently’, as in (22)-25): (22) X : OK. Donc ça, tu l’as lancé quand tu étais encore salarié ? H : Oui, quand j’étais encore salarié, oui. Là, je m’étais même renseigné puisque j’avais dû monter une auto-entreprise et puis savoir si c’était compatible avec mon boulot salarié, quoi. H : Et donc, voilà en fait, j’avais lancé ça … bêtement ! Et puis, ça a marché et je me suis dit : “Effectivement ! Je suis capable de générer de la valeur. Donc, allez, sautons dans le bain, quoi. Et donc, j’ai quitté mon boulot, là où je travaillais, dans l’enceinte de Grande Distribution, qui était un très bon boulot, qui était bien payé et tout, pas de soucis !. (lemarketsamurai.fr) ‘X : OK. So you launched the programme when you were still working? H : Yes, I was still working, yes. I even had inquired since I had set up my own

273

274

Sylvie Hancil

business and then I had to know whether it was compatible with my paid work. H : And then, here we are in fact, I’d launched that …like a fool ! And then, it worked and I told myself « « Wow! I am capable of generating some value. So, let’s go, let’s jump in, quoi. And then I left my job, where I worked in retail, which was a very good job, well paid, no worries !’ (23) bon c’est pas si grave (hein / n’est-ce pas / quoi) (ccc) ‘well, it is not so bad (hein / n’est-ce pas / quoi).’ (24) oui : / et il fait beau aujourd’hui : (hein / n’est-ce pas / quoi) /// ah c’est agréable ce temps-là (ccc) ‘well : and it’s sunny : : (hein / n’est-ce pas / quoi). Ah, this weather’s nice.’ (25) fâcheux temps (n’est-ce pas / hein / quoi) (prf ) ‘Difficult times, (n’est-ce pas / hein / quoi)’.

The particle quoi can be used as a modalizer meaning, more specially an intensifier, vraiment ‘really’, as in (26): (26) Amour. 49. Pourrais-tu aimer quelqu’un de physiquement .pas avantagé? Je sais pas. 50. Es-tu deja tombé amoureux d’un(e) inconnu(e) ? Oui, mais c’était pas vraiment “amoureux”, quoi. 1. Nom : Et non, vous saurez pas, lol 2. Prénom : Coline 3. Sexe : Féminin 4. Age : 13 ans et euhattendez je compte1 mois et1 semaine ! (codelyoko.fr) “Love 49 : Could you love someone who is physically challenged ? I don’t know. Have you fallen in love with a stranger ? Yes, but I wasn’t really “in love”, quoi. 1 : Name : And no, you won’t know, lol 2. First name : Coline 3. Sex : Female 4. Age : 13 and erm, hang on, one month and one week’.

When the particle quoi has lost its semantic value, it can be synonymous with the question tag n’est-ce pas or the particle hein ‘huh’ to keep the conversation rolling, as illustrated in (27)–(30): (27) On paye, et on file se planquer sur la terrasse, non sans avoir emporté son disque de commande, celui qui se mettra à clignoter quand votre repas aura été préparé et qu’il faudra aller le chercher. Ca se passe comme ça. On déjeune ensuite, entre l’ombre et le soleil, soudain isolé de la ville et de sa pagaille. Bientôt, on viendra boire des coups ou un café filtre à 1€ à la buvette, entre une poule et un plan de tomates. A l’étage tous les jeudis soirs, on peut également faire ses courses en direct avec des petits producteurs, grâce à la Ruche qui dit Oui, hébergée dans ce lieu définitivement tourné vers le collaboratif et l’écoresponsabilité. “Old is Gold”, quoi. Ce lieu est une blague! Ambiance sympa mais côté valeurs… (timeout.fr) ‘You pay and go hide on the terrace, but not without the disk, the one that will flash when your meal is ready and you have to go fetch it. This is how it goes.

Chapter 10. An emerging final particle

You then have lunch, between shadow and sunshine, suddenly isolated from the town and its mess. Soon, you will come to drink a few beers or 1€ filter coffee at the snack bar between a hen and a tomato patch. Upstairs, every Thursday night, you can also do your shopping, buying direct from small producers, thanks to the Hive that says Yes, hosted by this place firmly anchored in collaboration and eco-responsability. “Old is Gold”, quoi. This place is a joke ! Great atmosphere but a bit granola.’ (28) bon c’est pas si grave (hein / n’est-ce pas / quoi) (ccc) ‘Well, it is not so bad, (hein / n’est-ce pas / quoi).’ (29) oui : / et il fait beau aujourd’hui : (hein / n’est-ce pas / quoi) /// ah c’est agréable ce temps-là (ccc) ‘Yes, and it is sunny today, (hein / n’est-ce pas / quoi). Ah, it is nice that weather.’ (30) fâcheux temps (n’est-ce pas / hein / quoi). prf ) ‘A h, miserable weather,’ (n’est-ce pas / hein / quoi’.’

4.4 Principle 4: Persistence The principle of persistence shows that vestiges of the earlier meaning of contrast are still retained in the emerging final marker: the indefinite meaning persists, as exemplified below, where the indefinite value of the particle is reinforced by markers of hesitation such as Je sais pas ‘I don’t know’ and Comment vous dire ? ‘How can I tell you’?, as in (31): (31) A l’époque (j’étais jeune),je trouvais que c’était un drôle de choix que de faire élever ses enfants par sa mère. Je m’aperçois que le choix, elle ne l’avais peutêtre pas. Que de choses doivent changer, dans ce pays qu’est la France, qui a tendance à oublier ceux de ses habitants qui se trouvent loin de ses yeux. Akynou, si toutes vos illustrations d’images sont aussi réussies que l’histoire que je viens de lire, eh bien vous devriez, à mon avis, commencer sérieusement à vous chercher un éditeur. D’autant qu’en commençant à lire, j’avais un peu peur du “misérabilisme”. Comment vous dire? le syndrome “Cosette”, quoi. Ici, cet écueil (d’autant plus traître que l’émotion est toujours facile, l’effet à portée de main, dès qu’on met en scène un enfant), est évité grâce en partie au personnage de la “grande soeur” et sa réconfortante révolte (akynou.fr) ‘Back then (I was young) I thought it was a strange choice to have children raised by their mother. I realize that this choice, may not have been a choice at

275

276

Sylvie Hancil

all. That things must change, in this country that is France, with its tendency to forget about people who are out of sight. Akynou, if all your illustrations are as successful as the story that I’ve just read, well, you really should, in my opinion, start seriously looking for a publisher. When I started reading, I was a bit afraid of ‘miserabilism’. What do I mean? the “Cosette” syndrome, quoi. Here, this pitfall (all the more treacherous as the emotion is always easy, the effect is within easy reach as soon as you describe a child), is avoided thanks partly to the ‘big sister’ character and her comforting revolt.’

When the phrasing raises problems, the speaker regularly uses the particle quoi to accompany the linguistic production. This difficulty that the speaker has in expressing themself is explicitly underlined in the context by the use of j’veux dire ‘I mean’ or repetitions, as in (32): (32) et puis après ils étaient très étonnés parce que quand ils voulaient entrer en prépa on les prenait après les gens qui venaient du public ben + euh ça paraît normal quoi j’veux dire euh euh les lycées du (…) je pense que à l’intérieur de Paris globalement euh (..) les les écoles sont sont d’un bon niveau voilà. (CFPP2000, Yvette Audin, 7) ‘And then after they were quite astonished because when they wanted to start preparatory training for the Grande Ecoles, they took them after the people from the public school system quoi I mean erm erm the high schools of the (…) I think that inside Paris globally erm (…) the the schools are are at a good level, that’s it.’

There are repetitions, false starts, markers of hesitations such as euh ‘er’, pauses and repetition of grammatical markers, as shown in (33)–(34) : (33) spk2 : j’aime beaucoup le quartier euh: Saint-Michel euh:::: le côté SaintMichel boulevard Saint-Germain euh après on va dire que j’m’y connais pas trop trop en quartier dans les vers les autres côtés quoi spk1 : d’accord et puis pour vivre euh (CFPP2000, Lucie Da Sylva 7) ‘spk2 ; I like the neighbourhood erm of Saint Michel erm … the side Saint Michel boulevard Saint Germain erm after they will say that I do not really know the neighbourhood towards on the other sides quoi’ (34) spk2 : et ça m’a fait un ef- un effet bizarre hein c’était bizarre c’était un peu l’Amérique + vous savez c’était amus – on s’rendait pas compte du tout de: de c’que ça allait entraîner d’modifications mais là aussi c’étaient des débuts d’modifications considérables et c’était am- amusant quoi spk1: mais alors + vous en dites quoi de ces modifications vous m’avez parlé du quartier des petites boutiques (CFPP2000, Pauline Debordes, 7) ‘spk2 : And it made me an eff- weird effort huh it was weird it was a bit America + you know it was funn- we didn’t realize at all that it was going to bring

Chapter 10. An emerging final particle

about modifications but, it was also the beginnings of immense modifications and it was fun-funny quoi.’

4.5 Principle 5: Decategorization The principle of de-categorization argues that the new form shifts away from its original syntactic category as an indefinite pronoun, which is stuck to the predicate. According to Lefeuvre, Morel and Teston-Bonnard (2011), the particle quoi is, in the majority of cases, post-rhematic. A typical example is (35): (35) Il faut que je travaille tellement sur moi.. Que j’arrête d’avoir cette peur débile et ces a prioris à la con. < /s > < /p > < p > < s > Personne ne croit à mon challenge. Personne. < /s > < s > Ils sont tous là à sourire, désabusés. A ricaner ou à lever les yeux au ciel. Ils n’y croient pas. < /s > < s > Je me lèverai pour la mer, nom de Dieu. Ils ne me croient pas, n’y croient pas. < /s > < s > Rien, que dalle, nada. J’ai réalisé aussi qu’autour de moi, il n’y avait que le vide. Si peu de gens pour combler l’espace. Un cercle désespérément ridicule en numéros. Je veux devenir superficielle. Je voudrais avoir une vie «normale», quoi. Oublier un instant à quel point je peux être différente. (cowblog.fr) ‘I have to work so much on myself. I have to stop having this stupid fear and these stupid thoughts. Nobody believes my challenge. Nobody. There they all are, smiling, disillusioned. Sneering or rolling their eyes. They don’t believe me. I will stand up for the sea, for God’s sake. They don’t believe me, they don’t believe it. Nothing at all, nada. I have also realized that around me was just an void. So few people to fill in the gap. A circle desperately ridiculous in numbers. I want to become superficial. I wish I could have a “normal” life, quoi. Forget for a moment/second how different I am.’

Lefeuvre, Morel and Teston-Bonnard (2011) have examined three corpora and the majority of cases are post-rhematic. The first subcorpus is from RFC (Recherche sur le Français Contemporain) ; it lasts around 3 hours and has 188 occurrences, including 23 examples of pronouns (20%), and 95 examples of quoi as a discourse particle (80%). Out of these 95 occurrences of quoi, 74 stand at the end of the rheme (77.9%), including 23 at the end of a turn unit (31% of th total of ends of a rheme). The second corpus is CLAPI (Corpus de Langues Parlées en Interaction) and it has 135 hours of data. It is a multimedia databank of recorded corpora in real life in various contexts : professional interactions, institutional or private conversations, commercial situations, didactic situations, medical interactions, etc. Out of the 35 hours open to the public, it gathers 1041 occurrences of quoi, including 335 pronouns, and 706 discourse particles. Out of the 706 examples, 576 are at the end of a rheme (81%), with half of them at the end of a turn unit (40% of

277

278

Sylvie Hancil

the total of ends of rheme), and 130 (19%) have other distributions, at the middle of a rheme, for instance. In the third corpus called Corpus de Français Parlé Parisien (CFPP2000) (Branca et al.), there are 460 occurren,ces of quoi, including 245 concern the discourse particle (53%). Out of these 245 examples, quoi is present at the end of the rheme in 173 occurrences (70%). Sometimes the particle quoi occurs between the rheme and the post-rheme, as in (36) : (36) à une culture/ mais si tu veux euh . euh bon je travaille sur l’espace/ et tu sais avec quel sé/sérieux mais alors la notion d’enracinement euh me me met hors de moi/parce que je trouve que: c’est une c’est bien c’ est bien européen quoi ça tu vois qu’ils n’ont l’impression de vivre que quand ils sont propriétaires (bruit de camion) euh euh …quand ils sont propriétaires d’un lieu/ quand ils peuvent défendre un territoire/ enfin je trouve ça très animal\ (CLAPI Corpus Enquête de sociologie urbaine) ‘to a culture/but if you want erm erm well, I work on the space and you know how serious I am but then the notion of entrenchment erm drives me me mad because I find it is a it is quite European quoi all that, you see they feel they are living only when they are owners (sound of a truck) erm erm when they are owners of a place, when they can defend a territory, well, I find this quite animal-like.’

or it can be used in a cleft, as in (37) : (37) sinon je pense que [c’est euh vers vingt-cinq vingt-six ans quoi] qu’on atteint vraiment la plénitude mais euh quand on voit euh David Douillet par exemple qui est l’exemple suprême du judoka et qui a trente ans et qui s’apprête à faire les euh + Jeux Olympiques de Sydney quoi + (CRFP . PRI -BOR 1) ‘Otherwise I think that it is erm towards age 25 26 quoi that we really reach fullness but erm when you see erm David Douillet for instance who is the ultimate example of the judoka and who is 30 and is ready to erm go to the Olympics in Sydney, quoi.’

It can be employed inside the rheme, too, within a nominal syntagm, as in (38): (38) c’est /à, a-/ il y a un petit peu de tout il y a de la technique il y a du euh de la compétition + de la pédagogie et puis euh surtout le le brevet de secourisme quoi qui est important aussi (CRFP.PRI-BOR-1) ‘It is at, there is a litle bit of all that is technical, there is erm some competition + pedagogy and also erm above all, first-aid training quoi that is also important.’

or between the verb and the direct object, as in (39):

Chapter 10. An emerging final particle

(39) et en fait je trouve euh que euh si tu veux d’un côté je préfère quoi euh ne pas avoir d’encadrement tout ça parce que je me sens su-plus libre de mes mouvements quoi (ex. cited from Chanet 2001) ‘and in fact I find erm that erm if you want in a way I prefer quoi erm not having a support all that because I feel mo-more free in my movements quoi.

When the particle is post-rhematic, it can be preceded by a pause or not, namely it can be attached to the preceding main clause or not. So its syntactic nature can be an IP-adjunct or an orphan (Haegeman 1991). It is worth noting that the various types of semantic values encountered are not dependent on the syntactic nature of the constructions examined, whatever their syntactic category may be.

5.

Grammaticalization of the particle and beyond

The following path of grammaticalization can be reconstructed for final quoi : (40) INDEF pronoun > INDEF REL > INTENSIFIER > FILLER

According to the previous sections, final quoi has acquired metacommunicative meanings which are not included in the utterance meaning and are best explained in terms of cooptation. There is an increase in scope, so the semantic-pragmatic meanning expressing intensification belongs to cooptation. The particle ends its race as a filler, so grammaticalization explains the phenomenon. When the particle is syntactically independent form the host clause, it can be said to be related to cooptation ; when it is attached to the host clause, the process that best illustrates this is grammaticalization. So it can be argued that the evolution of final quoi is a combination of grammaticalization and cooptation. The cooptation process, as is elaborated by Heine (2013) is particularly useful : (a) to overcome constraints imposed by linearization in structuring texts, (b) to package together larger segments of discourse, (c) to place a text in a wider perspective, e.g. by providing explanations, comments, and supplementary information, (d) to describe the inner state of the speaker, and/or (e) to involve the hearer in the discourse (Heine 2013: 1221)

6.

Conclusion

To conclude, final quoi is not fully grammaticalized since the principle of specialization is not met by the data. Besides, the grammaticalization path shows that both grammaticalization and cooptation are intertwined in the process at various stages. Finally, the findings offer another example that classifies grammar as an

279

280

Sylvie Hancil

emergent, activity-related phenomenon; this research enriches the debate on the way usage affects the structure of a language. The grammatical structure is subject to variation, and therefore to change. The studies show that there exist at least certain types of changes that cannot be explained only by system-internal factors, but by factors taking into account the interaction between speaker and co-speaker.

References Andrews, Barry. 1989. Terminating devices in spoken French. International Review of Applied Linguistics in Language Teaching 27(3): 193–216. Beeching, Kate. 2002. Gender, Politeness and Pragmatic Particles in French [Pragmatics & Beyond New Series 104]. Amsterdam: John Benjamins. Beeching, Kate. 2004. Pragmatic particles – Polite but powerless? Tone-group terminal hein and quoi in contemporary spoken French. Multilingua 23: 61–84. Beeching, Kate & Detges, Ulrich. 2014. Discourse Functions at the Left and Right Periphery: Crosslinguistic Investigations of Language Use and Language [Studies in Pragmatics 12]. Leiden: Brill. Chanet, Catherine. 2001. 1700 occurrences de la particule quoi en français parlé contemporain: Approche de la ‘distribution’ et des fonctions en discours. Marges Linguistiques 2: 56–80. Delierre, Charles. 1997. Tu fais attention à c’que tu dis, quoi ou première approche de quoi. Travaux de l’Institut de Phonétique de Strasbourg 27 : 25–61. Detges, Ulrich & Waltereit, Richard. 2011. Turn-taking as a trigger for language-change. In Rahmen des Sprechens. Beiträge zu Valenztheorie, Varietätenlinguistik, Kreolistik, Kognitiver und Historischer Semantik, Sarah Dessì Schmid, Ulrich Detges, Paul Gévaudan, Wiltrud Mihatsch & Richard Waltereit (eds), 175–189. Tübingen: Narr. Du Bois, John W. & Giora, Rachel. 2014. From cognitive-functional linguistics to dialogic syntax. Cognitive Linguistics 25 (3), 351–357. Goose, André. 1993. Le Bon Usage, 13th edn. Paris: Duculot. Haegeman, Liliane. 1991. Parenthetical adverbials: The radical orphan approach. In Aspects of Modern English Linguistics: Papers Presented to Masatomo Ukaji on his 60th Birthday, Shuki Chiba, Akira Ogawa, Yasuaki Fuiwara, Norio Nawada, Osamu Koma & Takao Yagi (eds), 232–254. Tokyo: Kaitakusha. Heine, Bernd, Kaltenböck, Gunther, Kuteva, Tania & Long, Haiping. 2013. An outline of discourse grammar. In Functional Approaches to Language, Shannon Bischoff & Carmen Jany (eds), 155–206. Berlin: De Gruyter Mouton. Hopper, Paul J. 1991. On some principles of grammaticalization. In Approaches to Grammaticalization, Vol.1: Theoretical and Methodological Issues [Typological Studies in Language 19:1], Elizabeth Closs Traugott & Bernd Heine (eds), 17–35. Amsterdam: John Benjamins. Kerbrat-Orecchioni, Catherine. 2005. Le discours en interaction. Paris: Armand Colin Lefeuvre, Florence. 2006. Quoi de neuf sur quoi? Étude morphosyntaxique du mot quoi. Rennes: PUR.

Chapter 10. An emerging final particle

Lefeuvre, Florence, Morel, Mary-Annick & Teston-Bonnard, Sandra 2011. Valeurs prototypiques de quoi à travers ses usages en français oral. Bulletin de la Société Néophilologique (Neuphilologische Mitteilungen) 112(1): 37–59. Morel, Mary-Annick & Danon-Boileau, Laurent. 1998. Grammaire de l’intonation. Paris: Ophrys. Roulet, Eddy, Auchlin, Antoine, Moeschler, Jacques & Rubatte, Christian. 1985. L’articulation du discours en français contemporain. Berne: Peter Lang. Traugott, Elizabeth Closs. 2003. Constructions in grammaticalization. In The Handbook of Historical Linguistics, Brian Joseph & Richard Janda (eds), 624–643. Oxford: Blackwell. Traugott, Elizabeth Closs. 2010. Grammaticalization. In Continuum Companion to Historical Linguistics, Silvia Luraghi & Vit Bubenik (eds), 269–283. London: Continuum. Traugott, Elizabeth Closs. 2012. Intersubjectification and clause periphery. English Text Construction 5: 7–28. Traugott, Elizabeth Closs & Dasher, Richard B. 2002. Regularity on Semantic Change. Cambridge: CUP. Verhagen, Arie. 2005. Constructions of Intersubjectivity. Discourse, Syntax and Cognition. Oxford: OUP.

Dictionaries Furetière, Antoine 1690. Dictionnaire universel. Le Grand Robert de la langue française 2013. Le Petit Robert 1967. paris: Société du Nouveau Littré, le Robert. Littré, Emile 1876. Dictionnaire de la langue française. Société du Littré. Larousse, Pierre 1889. Dictionnaire complet. Paris: Larousse Trésor de la langue française informatisé, (TLFi). 1990

Corpus Corpus de français parlé parisien (CFPP) French Web corpus 2017

281

Index A ablative marker 9, 20, 22, 23, 36 accusative case marker 200 adnominalizer 10, 200, 201, 213, 217, 224, 225, 228, 230 adverbialization 124, 127, 129 apprehensive verb 8, 50 auxiliary 4, 11, 74, 75, 76, 81, 83, 84, 87, 88, 89, 93, 94, 95, 96, 97, 98, 105, 120, 253 B bilingualism 183 Bantu 11, 74, 239, 252, 253, 254, 255, 261 Bresciano 8, 104, 105, 114, 115, 116, 117, 118, 119, 120 C case particles 155 comparative adverb 9, 20, 21, 22, 23, 25, 26, 27, 28, 30, 31, 32, 33, 34, 36, 37, 39, 40, 41, 42 comparative standard marker 9, 20, 21, 22, 23, 24, 25, 27, 31, 32, 41, 42 complex adaptive system 194, 195 construction 2, 4, 5, 9, 10, 12, 13, 15, 16, 18, 25, 26, 27, 31, 32, 46, 50, 51, 52, 53, 54, 71, 72, 74, 76, 77, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 105, 106, 107, 108, 110, 111, 113, 116, 117, 118, 119, 127, 132, 134, 139, 140, 141, 142, 145, 147, 148, 149, 150, 151, 152, 153, 154, 158, 160, 161, 162, 163, 166, 168, 169, 171, 173, 174, 185, 187, 188, 186, 198, 201, 209, 213, 233, 240, 241, 242, 243, 246, 248, 249, 250, 251, 252, 253, 254, 255, 256, 279, 281 constructional entrenchment 9, 147, 150, 168, 169, 171, 173

critical context 8, 124, 128, 129, 130, 132, 139, 140, 141 D debonding 20, 40 degrammaticalization 5, 15, 16, 20, 22, 30, 39, 40, 41, 42, 46, 58 deixis 10, 228, 229, 230, 234 E ellipsis 10, 69, 140, 198, 199, 201, 217 epistemic phrase 8, 124, 125, 126, 127, 129, 132, 139, 140, 145 experimental study 131, 132 F final-appendage construction 147, 148, 150, 152, 153, 154, 158, 163, 166, 168, 171, 172 G grammatical relation sensitivity 171, 172 decategorization, 11, 263, 264, 277 diatopic variation 104, 116, 117, 119, 120 divergence 11, 124, 219, 263, 264, 266 epistemic 8, 50, 51, 52, 54, 55, 56, 57, 58, 59, 60, 61, 63, 64, 65, 66, 68, 69, 70, 71, 72, 73, 124, 125, 126, 127, 128, 129, 130, 132, 138, 139, 140, 141, 145, 146, 270 expansion 5, 9, 10, 23, 30, 31, 40, 42, 47, 50, 52, 53, 54, 55, 58, 59, 60, 61, 69, 87, 99, 127, 235 expressive 51, 56, 57, 58, 59, 61, 64, 68, 70, 72, 97, 177, 238, 246, 247, 250, 251, 253, 254, 255, 256, 259, 261

extravagance 6, 20, 37, 38, 39, 42 final particle 43, 59, 152, 156, 169, 170, 172, 174, 263, 264 French 11, 53, 108, 109, 123, 127, 176, 179, 182, 183, 184, 185, 186, 187, 188, 189, 194, 217, 263, 264, 271, 280, 281 H history of French 186 I ideophone 10, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 260 261 indexicality 10, 198, 224, 228, 229, 230 inference 59, 96, 222, 226 information theory 181 interaction 8, 10, 17, 38, 50, 59, 72, 73, 125, 143, 145, 176, 189, 203, 204, 225, 226, 227, 230, 233, 234, 242, 264, 277, 280 interactive 10, 189, 232, 237, 239, 242, 243, 244, 249, 250, 255, 259, 260, 261, 264, 266, 269 Italo-Romance 8, 104, 105, 113, 114, 115, 121 J Japanese 9, 10, 20, 22, 23, 24, 25, 26, 29, 30, 39, 40, 41, 42, 43, 44, 45, 46, 47 K Korean 9, 10, 39, 41, 46, 147, 148, 149, 150, 151, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173 Jespersen’s cycle 8, 104, 108, 116, 119, 120

284

Different Slants on Grammaticalization

L language contact 9, 20, 22, 24, 25, 27, 29, 37, 39, 41, 42, 43, 44, 45, 46, 47 layering 9, 11, 21, 27, 42, 263, 264, 266 Latin 53, 109, 110, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 193, 194 lexical right-sizing 195 M Mandarin 8, 16, 17, 50, 55, 56, 58, 60, 68, 69, 71, 72, 73 Mara 11, 74, 77, 78, 79, 81, 83, 89; 93, 95, 96, 97, 98, 100, 102 middle voice 187, 188 mimetic 238, 261 modifier 10, 159, 200, 213, 214, 215, 216, 217, 218, 223, 224, 226, 230, 241, 242, 243, 250, 252, 254, 255 morphological erosion 124, 126, 127, 129 N “nani sore” construction 160, 162, 163, 168, 169 negation 8, 85, 90, 93, 100, 104, 105, 106, 107, 108, 109, 110, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122

P particles 3, 11, 17, 32, 101, 105, 106, 108, 147, 148, 155, 163, 164, 168, 169, 172, 173, 231, 232, 234, 236 particle-less pronouns 163, 164, 165, 167, 168 perception 72, 124, 125, 131, 139, 140, 141, 145, 239 phasal polarity 11, 72, 74, 75, 93, 97, 98, 101 postposing 147, 149, 150, 151, 152, 153, 158, 161, 162, 166, 170, 171, 172, 173, 174 post-predicative position 147, 148, 150, 152, 155, 158, 161, 163, 164, 165, 166, 167, 168, 169, 171 persistence 11, 20, 32, 42, 55, 66, 91, 182, 263, 264, 275 phonological erosion 11, 176, 177, 178, 179, 184, 186, 191 potential grammaticalization 124, 125 prosody 55, 58, 143, 174, 198, 202, 225, 226, 227, 228, 232, 234, 247, 254 Q quantity principle 11, 176, 181, 183, 185, 186, 189, 191, 194 quotative 42, 100, 172, 241, 242, 246, 249, 252, 253, 254, 255

S sentential negation 8, 104, 105, 107, 108, 109, 110, 113, 116, 119, 123 specialization 11, 263, 264, 272, 279 Siwu 237, 244, 245, 248, 249, 252, 256, 261 speech 6, 17, 20, 30, 44, 46, 51, 56, 63, 72, 73, 129, 130, 137, 139, 142, 145, 165, 191, 192, 193, 196, 199, 200, 219, 232, 233, 241, 242, 261, 265, 268, 271 spoken data 63, 125, 131 synchronic 6, 7, 8, 9, 14, 17, 27, 38, 42, 72, 77, 100, 121, 124, 181, 196, 202, 234, 235, 258 V vagueness 10, 198, 228, 229, 230 Veronese 104, 113, 115, 116, 117, 118, 119, 120 vocative particles 148, 168, 172 X Xhosa 237, 245, 252, 253, 254, 255, 256, 260

This volume on grammaticalization focuses on new theoretical and methodological challenges underpinning language change. It provides new approaches and insights deepening our understanding of the cognitive, pragmatic, and socio-cultural mechanisms that trigger the formation and the change of grammars. In this volume, grammaticalization is dealt with diachronically, synchronically and as a by-product of dialogic interaction. Another key feature of this book is language diversity; as it includes studies on language families ranging from Niger-Congo, Koreanic, Japonic, Sino-Tibetan to Germanic and Romance. The novel aspects of grammaticalization addressed are new slants on the fundamental debate about grammaticalization as expansion vs reduction; the grammatical formation of ideophones; the semantic domain of fear as a source and a trigger of grammatical change, and many other aspects of semantic and morphosyntactic development.

     

   