Usage-based Approaches to Japanese Grammar
Studies in Language Companion Series (SLCS) 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 http://benjamins.com/catalog/slcs
Editors Werner Abraham
University of Vienna / University of Munich
Elly van Gelderen
Arizona State University
Editorial Board Bernard Comrie
Max Planck Institute, Leipzig and University of California, Santa Barbara
William Croft
University of New Mexico
Östen Dahl
University of Stockholm
Gerrit J. Dimmendaal University of Cologne
Ekkehard König
Free University of Berlin
Christian Lehmann University of Erfurt
Marianne Mithun
University of California, Santa Barbara
Heiko Narrog
Tohuku University
Johanna L. Wood
University of Aarhus
Debra Ziegeler
University of Paris III
Volume 156 Usage-based Approaches to Japanese Grammar Towards the understanding of human language Edited by Kaori Kabata and Tsuyoshi Ono
Usage-based Approaches to Japanese Grammar Towards the understanding of human language Edited by
Kaori Kabata Tsuyoshi Ono University of Alberta
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.
Library of Congress Cataloging-in-Publication Data Usage-based Approaches to Japanese Grammar : Towards the understanding of human language / Edited by Kaori Kabata and Tsuyoshi Ono. p. cm. (Studies in Language Companion Series, issn 0165-7763 ; v. 156) Includes bibliographical references and index. 1. Japanese language--Usage. 2. Japanese language--Spoken Japanese. 3. Japanese language--Writing. I. Kabata, Kaori, editor of compilation. II. Ono, Tsuyoshi, editor of compilation. PL642.U83 2014 495.65--dc23 2014004751 isbn 978 90 272 5921 9 (Hb ; alk. paper) isbn 978 90 272 7031 3 (Eb)
© 2014 – 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 Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
Table of contents Acknowledgement List of contributors Introduction: Situating usage-based (Japanese) linguistics Tsuyoshi Ono and Ryoko Suzuki
vii ix 1
part 1. Cognition and language use Subordination and information status: A case of To and Koto complement clauses in Japanese Naomi H. McGloin
13
On state of mind and grammatical forms from functional perspectives: The case for Garu and Te-iru Yuki Johnson
37
Grammar of the internal expressive sentences in Japanese: Observations and explorations Shoichi Iwasaki
55
Subjectivity, intersubjectivity and Japanese grammar: A functional approach Rumiko Shinzato What typology reveals about modality in Japanese: A cross-linguistic perspective Kaoru Horie and Heiko Narrog
85
109
part 2. Frequency, interaction and language use If rendaku isn’t a rule, what in the world is it? Timothy J. Vance
137
Usage-based Approaches to Japanese Grammar
The semantic basis of grammatical development: Its implications for modularity, innateness, and the theory of grammar Yasuhiro Shirai
153
Interchangeability of so-called interchangeable particles: Corpus analysis of spatial markers, Ni and E Kaori Kabata
171
The re-examination of so-called ‘clefts’: A study of multiunit turns in Japanese talk-in-interaction Junko Mori
193
Activity, participation, and joint turn construction: A conversation analytic exploration of ‘grammar-in-action’ Makoto Hayashi
223
part 3. Language change and variation Context in constructions: Variation in Japanese non-subject honorifics Yoshiko Matsumoto
261
The use and interpretation of “regional” and “standard” variants in Japanese conversation Shigeko Okamoto
279
Index
305
Acknowledgement The preparation for the volume was generously supported by the Social Sciences and Humanities Research Council of Canada (#646-2003-1136), University of Alberta, the Faculty of Arts, the Department of East Asian Studies, and the Spoken Discourse Research Studio. We would like to extend our thanks to the symposium secretary Neill Walker and our many volunteers for their hard work, without which a successful symposium would simply not have been possible. We would also like to thank Diana Benschop, Robin Coogan, and Yumi Sasaki for their assistance in preparing the volume and Janice Brown and Sandra Thompson for their support in organizing the symposium.
List of contributors Makoto Hayashi University of Illinois at UrbanaChampaign, USA Kaoru Horie Nagoya University, Japan Shoichi Iwasaki University of Hawaii, Manoa/ University of California, Los Angeles, USA Yuki Johnson The Johns Hopkins University, USA Kaori Kabata University of Alberta, Canada Yoshiko Matsumoto Stanford University, USA Naomi Hanaoka McGloin University of Wisconsin-Madison, USA Junko Mori University of Wisconsin-Madison, USA
Heiko Narrog Tohoku University, Japan Shigeko Okamoto University of California, Santa Cruz, USA Tsuyoshi Ono University of Alberta, Canada Rumiko Shinzato Georgia Institute of Technology, USA Yasuhiro Shirai University of Pittsburgh, USA Ryoko Suzuki Keio University, Japan Timothy J. Vance National Institute for Japanese Language and Linguistics, Japan
introduction
Situating usage-based (Japanese) linguistics Tsuyoshi Ono and Ryoko Suzuki
1. Introduction Starting in the mid-70s, the field of linguistics saw an explosion of a new breed of research represented by the work of scholars including Bybee, Chafe, Comrie, Du Bois, Fillmore, Givón, Haiman, Hopper, Lakoff, Langacker, Li, Mithun, Talmy, and Thompson. This movement, though not organized as a research paradigm as a whole, shares a number of evolving theoretical assumptions, research goals, and focus of investigation as well as the types of data and methodologies employed. Its various specific subgroupings and instantiations have been (self-)identified by such designations and labels as Cognitive Grammar/Linguistics, Construction Grammar, (discourse-)functional linguistics, functionally-oriented linguistics, West Coast discourse-functional linguistics, and, more recently, usage-based linguistics, which we adopt for this volume. The present Japanese-focused volume contains articles written by the second and third generation of linguists belonging to this tradition. In this introduction, we will highlight a set of interrelated themes featured in this tradition to situate the articles in the volume. Our goal is to help the reader gain an understanding of the main findings and theoretical implications in our collective effort to better capture the nature of human language in general and the grammar of Japanese in particular. 2. Themes in usage-based linguistics 2.1
Universals and cross-linguistic orientation
As is found in some of the traditional schools of linguistics, usage-based linguistics has been very much concerned with universals. There are perhaps two distinctive characteristics for the study of universals in usage-based linguistics. (1) Since
Tsuyoshi Ono and Ryoko Suzuki
its inception in the 70s, usage-based linguists have had a strong cross-linguistic orientation; proposals for universals in usage-based linguistics were made based inductively on the examination of multiple languages (e.g., Keenan & Comrie 1977; Hopper & Thompson 1980). The motivation for the cross-linguistic orientation is obvious and inevitable in that one of the main goals of usage-based linguistics is to identify a set of universal principles of human language, and one has to look at more than one language to reach that goal. This characteristic originates in the work by Greenberg (e.g., 1963) and has resulted in the research tradition represented by such researchers as Givón, Li, Thompson, Bybee, Shibatani, Mithun, Comrie, Dryer, Croft, and Haspelmath. Proposals in other schools, in contrast, have strongly tended to be based on select languages, typically just English, though studies in recent years in the latter appear to involve more and more languages, perhaps due to influence from the cross-linguistic orientation of the former. You will find that papers by Horie and Narrog, Shinzato, and Shirai in this volume are particularly good representatives of this orientation. (2) Along with the cross- linguistic orientation to the search for universals, the suggested universals have been motivated by functional factors: (a) cognitive and discourse factors such as memory, attention, economy and iconicity (e.g., Chafe (1980), Givón (1983), Haiman (1985), and (b) affective, social, and interactional factors such as emotion, relationships among interactants, and (dis-)preferred social actions,(e.g., Ochs & Schieffelin 1989; Sacks, Schegloff & Jefferson 1974). Some of these functional motivations will be further discussed below. 2.2
External factors and interdisciplinary orientation
An important theme which characterizes usage-based linguistics is the recognition that language is not a static self-contained entity but is a living organism interfaced with, and thus shaped constantly by, a wide range of non-language factors including human cognition and interactional concerns. Cognition (often called simply semantics) Usage-based linguists argue that cognition plays an integral role in every facet of linguistic activities Langacker 1987. What speakers perceive and feel in context is tied to the linguistic structure representing the degree of beliefs and various perspectives taken by the speaker and others whose views are embedded in the utterance. Cognitive concerns are thus naturally reflected in the way language changes and is structured over time. Cognition forms a basis for a number of papers in the present volume including Horie and Narrog, Iwasaki, Johnson, McGloin, and Shinzato.
Introduction: Situating usage-based (Japanese) linguistics
Discourse/information structure/frequency Another factor which usage-based linguistics has been oriented to is what is called ‘information flow’ (Chafe 1994). For instance, the informational status of a referent, such as givenness, expectedness, familiarity and so on, is shown to correlate with several facets of grammar. One example is that the semantic and pragmatic status of noun phrases in a clause gives rise to crystalization of a particular argument structure: many languages grammaticize agentivity/topicality which leads to the same case-marking for subjects of intransitives and that of transitives, whereas many other languages code newness, with subjects of intransitives and objects of transitives using the same case-marking (See Du Bois 1985, 1987 for details). It is important to note that, in usage-based linguistics, the frequency of certain configurations in discourse is treated as an essential clue to understanding linguistic structure. The papers by Kabata and Shirai exhibit some of the characteristics of this thread. Interaction Investigation of language based on usage gains insights from careful observation of the ways in which interlocutors manage conversation moment-by- moment to achieve a wide range of communicative goals. In particular, turn-taking is one of the major topics of investigation among research on conversational interaction. Researchers discuss the nature of prosodic/grammatical units and other interactional factors, including gaze and gestures, interacting together to constitute a TCU (turn construction unit). This research thread has led to the formation of a new research paradigm called Interactional Linguistics, involving researchers such as Thompson, Couper-Kuhlen, Ford, Fox, Selting. The papers by Hayashi and Mori in the present volume are particularly good examples representing this thread. Interdisciplinary orientation As a natural consequence of having to deal with the above-mentioned factors, which are well studied in fields outside of linguistics, it is evident that usage-based linguistics has been an interdisciplinary endeavor. It necessarily closely intersects with psychology and cognitive science on the one hand, and sociology, anthropology and communication studies on the other. 2.3
Parting from intuition
Usage-based linguists have long expressed concerns regarding the use of speakers’ grammaticality judgment (the so-called native speakers’ intuition) of constructed examples as the primary data and methodology to study what is claimed to be core
Tsuyoshi Ono and Ryoko Suzuki
properties of language. For this reason, from early on, usage-based linguists distinguished themselves by taking into consideration semantics and pragmatics in analyzing linguistic structure even though they were then still examining constructed examples. In the mid-80s, going along with the idea that the study of the structure and use of language cannot be separated, usage-based linguists gradually shifted their attention to actual use: discourse. Initially this was done with more readily available types of data such as written texts and traditional folk tales and more easily collectable data such as elicited spoken narratives, but starting in the early/mid 90s, more and more researchers have been turning their attention to everyday talk, the primary form of language, on which both qualitative and quantitative analyses focusing on cognitive, discourse, affective, social and interactional factors are performed. This research has mostly been hypothesis-building in nature in that one observes data carefully and sorts out the patterns which emerge out of the data. All the contributions in the current volume address the problem of relying solely on intuition, though Iwasaki, Johnson, and McGloin particularly underscore the utility of the traditional methods. More recently, however, an increasing number of researchers have started employing rather sophisticated statistical methodologies on large scale corpora to test specific hypotheses and further engaging in experimental work trying to pin down specific factors in actual language use. Papers by Kabata and Vance are good examples of this newer trend. Unfortunately, progress along these lines is hampered by the general lack of large-scale corpora representing naturally occurring talk. That is, the large corpora currently available are severely limited in everyday talk, the primary form language, and instead made up mostly of various types of written sources and, in some rather limited cases, monologues and speeches as well as conversations created for research purposes, even though claims based on them are typically made as general properties of particular languages or human language in general. It should be pointed out that Japanese is no exception to this overwhelming pattern. Needless to say, constructed situations in experimental studies which are assumed to bring out human behavior in actual interactional contexts are, at the very least, suspect. This is perhaps the main reason why some usage-based linguists have tended to be skeptical about many of the findings of these recent statistical and experimental studies. It is our belief that in order to further advance the science of Japanese linguistics, the first step to take should be to try constructing a corpus representing various types of language use (determining what is a proper mix of various types of language use in itself is an important project of its own), especially representing naturally occurring talk. This would become a major milestone not only for the study of Japanese but also for human language in general since large-scale corpora of naturally occurring talk are still extremely limited.
Introduction: Situating usage-based (Japanese) linguistics
2.4
Non-discrete nature of linguistic categories
Another area where usage-based linguists distinguish themselves from those with more traditional views is the understanding and therefore treatment of linguistic categories. Traditionally, linguistic categories such as nouns vs. verbs have tended to be thought of as discrete entities. However, usage-based studies (Hopper & Thompson 1980, 1984; Givón 1983; Langacker 1987, 1991) have highlighted the non-discrete nature of linguistic categories where such categories as noun, verb, transitivity, and topic are better understood, not in terms of yes or no, but more or less, such as more/less nouny/verby, more/less (in)transitive, more/less topical, and so forth. This is, however, still far from being a mainstream view as observed in how these categories are discussed and treated by linguists with various theoretical orientations in the literature. In our view, studies dealing with the categories are minimally responsible for justifying their particular treatment rather than simply assuming that they are discrete. Again a number of papers in this volume, including Johnson, McGloin, and Okamoto, exhibit some characteristics of this thread. Perhaps not unrelated to the idea that linguistic categories are not discrete is the usage-based linguists’ view on the general structure of grammar. The other commonly held traditional view of grammar is that language consists of specific components such as syntax, phonetics, and lexicon, which might at first seem reasonable. Those who hold the view that language consists of components contend that it is simply a hypothesis while usage-based linguists say if it is a hypothesis, it needs to be tested. Again usage-based linguists commonly view that (1) boundaries among those components are much less discrete than has been assumed, (2) in fact, some of the standard separate components seem to be much better understood as having a different type of structure such as grammar and lexicon forming one single component (Langacker 1987) where the notion of grammar and lexicon is captured in terms of degree such as more or less grammatical/lexical, and (3) minimally one should not simply assume that language consists of components; it needs to be verified. Papers by Matsumoto, Okamoto, Shinzato, and Shirai in this volume give relevant discussion to this issue. Further, as we will suggest below, the nondiscreteness observed in these traditional components of grammar might simply be a consequence of the reality of actual human behavior: language change. 2.5
No division between synchrony and diachrony
One of the crucial questions that usage-based linguists would ask is why languages have become the way they are now. Paying special attention to recurrent patterns in discourse and where and how the patterns emerge, they view grammar as evolving through frequent use, thus constantly-changing, rather than static and fixed.
Tsuyoshi Ono and Ryoko Suzuki
In other words, rather than subscribing the commonly held distinction of synchrony and diachrony, the researchers take a non-discrete view in terms of the temporality of the data. Hopper (1987) coined the term ‘emergent grammar’ to indicate that there is nothing called a ‘grammar’ of language in reality, but what we see in everyday talk is the process of ‘grammaticization (grammaticalization)’ in which fluid uses become more fixed patterns through recurrence (See also work by Bybee 2010, Du Bois 1985, 1987 among others). A number of papers in this volume including Matsumoto, Shirai, and Shinzato address this issue. 3. Preview of the articles The reader will learn quickly that reflecting theoretical, methodological, and technological advancements made in and outside the field in recent years, the above articles all focus on specific phenomena employing various methodologies yet together manage to highlight general aspects of Japanese and human language. The first part of this book deals with papers that are concerned with the interaction between cognition and language use. The contribution by Naomi McGloin is perhaps most traditional in that she compares cases where the use of the complementizers no and koto are said to be interchangeable. Using constructed examples and internet data, McGoin, however, identifies some clear structural and informational differences associated with them where differences should be captured not in terms of ‘either or’ but in terms of degree. In particular, she shows that, unlike what has previously been suggested, no actually interjects a strong subjective stance by the speaker. Yuki Johnson fills the gap between the traditional understanding of grammar and actual usage by focusing on examples taken from conversation and e-mail as well as speaker judgment about them. Specifically, she focuses on garu and te-iru which have both been traditionally analyzed to describe a state of mind which belongs to someone other than the speaker. She finds that the use of these forms is closely tied with who the speaker considers her/his in- and out-group and contends that one not only needs the description of standard grammar but usage data to fully account for how they are used. The contribution by Shoichi Iwasaki focuses on a type of expression which has escaped the attention of researchers, such as a! itai! ‘Ouch!’. Iwasaki shows that these ‘(internal) expressive sentences’ have a distinct set of grammatical patterns which represent neurological experience of perception, emotion, and feeling which the speaker undergoes. Iwasaki’s study is based on constructed data, but he explores other ways in which one might study these expressions which are rather difficult to obtain examples of in actual language use.
Introduction: Situating usage-based (Japanese) linguistics
Rumiko Shinzato also looks into traditionally neglected non-propositional aspects of Japanese. Specifically, she goes over various areas of Japanese language intimately connected to the subjectivity and intersubjectivity distinction: the order of predicate elements, the mental (e.g., omou ‘think’) /speech act (e.g., iu ‘say’) verb difference, and soliloquy/dialogue distinction. She further highlights this connection reflected in diachronic change in terms of the unidirectionality of grammaticization. Kaoru Horie and Heiko Narrog also take a cross-linguistic approach to their study of modality in which they examine Japanese along with English, German, and Korean. By examining specific modal categories of these languages from a typological perspective, Horie and Narrog suggest that the grammatical encoding of modality varies cross linguistically much more than tense and aspect precisely because it is tied to the socio-cultural cognition and communicative practices of particular linguistic communities. Part 2 presents those papers that examined the influence of external factors, including frequency and interactions on language use. Timothy Vance tackles a well-known voicing phenomenon rendaku (also known as sequential voicing) which is, as he shows, fundamentally irregular. After reviewing some experimental studies which all result in partially regular patterns yet are not conclusive at all, Vance wonders in the end if the feeling of predictability based on analogy might result in the mistaken belief among speakers (perhaps linguists) that there is a relatively straightforward regular rule – what, as the current authors think, might be responsible for some (or many) of what we conceive of as grammatical rules. Yasuhiro Shirai’s study is cross-linguistic in nature by going over series of studies focusing on various grammatical phenomena in Japanese as well as English and Korean in order to establish the semantic basis of L1 grammatical development. He advocates a view of grammar acquisition which is guided by semantic prototypes formed from the input children are exposed to Shirai problematizes a number of theoretical assumptions and devices such as the discreetness of categories and the modularity of grammar often uncritically adopted in traditional approaches. The paper by Kaori Kabata highlights the discrepancy between what is represented in Japanese grammar and what empirical data reveals. Specifically, she examines data from various sources including native speakers’ judgments and corpora involving actual spoken and written language use and finds that the Japanese allative markers ni and e, which have been assumed be interchangeable exhibit clear skews bringing into questions where different types of data show different types of skews, how grammar should be understood and represented. Junko Mori adopts the Conversation Analysis (CA) framework to examine one particular type of the so-called cleft construction, X wa Y da, in which she
Tsuyoshi Ono and Ryoko Suzuki
finds the second Y da part is rarely produced in actual conversation. She instead shows that the first X wa part is tied to an extended unit of talk showing that the relevant connection might not be syntax, as has been assumed based on the analysis of constructed examples. Her findings were supported by CA based evidence, which most radically includes the analysis of bodily behavior demonstrating the intimate connection between language and bodily behavior. Perhaps in a very radical way, Makoto Hayashi directly examines the tight connection between grammar and interaction in conversational data. Specifically, Hayashi focuses on ‘joint turn construction’ – a practice whereby a participant in conversation completes a grammatical unit-in-progress initiated by another participant and shows that grammar and interaction continuously shape each other’s realization, again demonstrating the need of a dynamic view of grammar which was traditionally absent in the field of linguistics. The two papers in Part 3 deal with language change and variation. The article by Yoshiko Matsumoto analyzes data from actual speech and the internet. She reveals that non-subject honorific forms, which mark deference to the referent of non-subject arguments of the clause, are instead used to mark deference to the addressee (or the reader of the internet) demonstrating that grammatical device marking a relationship among the referents expressed in the utterance (semantics) has come to mark a relationship within the speech context (pragmatics). This new use seems to be rather widely observed and is in fact accepted by Japanese speakers, suggesting that they are not merely “errors” but another example of the change documented in the history of Japanese where referential honorifics changed into addressee honorifics. Shigeko Okamoto adopts conversation data in her study of regional dialects (Osaka and Yamaguchi) of Japanese with the background where switching of distinct dialects of Japanese has been commonly discussed based on informal observations and self-report survey data. She finds the actual use to be much more complex and radically different since, at various points in producing utterances, speakers variedly choose phonological, morphological, and lexical forms of the regional and Standard Japanese forms (i.e., variant choice). This results in (a) constant mixing of the two varieties within and across utterances and (b) different amounts of the two, depending on the speaker. She calls for the need to study individual human behaviors in depth and for a dynamic view of language to account for such a behavior. As a whole, these papers provide us with a rich array of data and methodologies which illuminate Japanese grammar in context. They reveal multiple new ways and layers of understanding of grammar as a usage-based phenomenon, naturally associated with plentiful of traces of ongoing change.
Introduction: Situating usage-based (Japanese) linguistics
References Bybee, Joan. 2010. Language, Usage and Cognition. Cambridge: CUP. DOI: 10.1017/ CBO9780511750526 Chafe, Wallace L. 1980. The Pear Stories: Cognitive, Cultural, and Linguistic Aspects of Narrative Production, Vol. 3. Norwood NJ: Ablex. Chafe, Wallace L. 1994. Discourse, consciousness, and time. Chicago IL: University of Chicago Press. Du Bois, John W. 1985. Competing motivations. In John Haiman (ed.), 343–65. Du Bois, John W. 1987. The discourse basis of ergativity. Language 63(4): 805–55. DOI: 10.2307/415719 Greenberg, Joseph H. 1963. Some universals of grammar with particular reference to the order of meaningful elements. In Universals of Language, Vol. 2, Joseph H. Greenberg (ed.), 73– 113. Stanford CA: Stanford University Press. Givón, T. 1983. Topic Continuity in Discourse: A Quantitative Cross-language Study [Typological Studies in Language 3]. Amsterdam: John Benjamins. DOI: 10.1075/tsl.3 Haiman, John (ed.). 1985. Iconicity in Syntax: Proceedings of a Symposium on Iconicity in Syntax, Stanford, June 24–6, 1983 [Typological Studies in Language 6]. Amsterdam: John Benjamins . Hopper, Paul J. 1987. Emergent grammar. BLS 13, 139–157. Hopper, Paul J. & Thompson, Sandra A. 1980. Transitivity in grammar and discourse. Language 56(2): 251–299. Hopper, Paul J. & Thompson, Sandra A. 1984. The discourse basis for lexical categories in Universal Grammar. Language 60(4): 703–752. Reprinted in Aarts, Bas, Denison, David, Keizer, Eveline & Popova, Gergana (eds). 2004. Fuzzy Grammar, 293–308. Oxford: OUP. Keenan, Edward L. & Comrie, Bernard. 1977. Noun phrase accessibility and universal grammar. Linguistic Inquiry 8(1): 63–99. Langacker, Ronald W. 1987. Foundations of Cognitive Grammar: Theoretical Prerequisites, Vol. 1. Stanford CA: Stanford University Press. Langacker, Ronald W. 1991. Foundations of Cognitive Grammar: Descriptive Application, Vol. 2. Stanford CA: Stanford University Press. Ochs, Elinor & Schieffelin, Bambi. 1989. Language has a heart. Text 9(1): 7–25. DOI: 10.1515/ text.1.1989.9.1.7 Sacks, Harvey, Schegloff, Emanuel A. & Jefferson, Gail. 1974. A simplest systematics for the organization of turn-taking for conversation. Language 50(4): 696–735. DOI: 10.2307/412243
part 1
Cognition and language use
Subordination and information status A case of To and Koto complement clauses in Japanese Naomi H. McGloin This paper examines one type of subordinate construction in Japanese, namely, complement clauses, in particular those where two different complementizers, to and koto, are interchangeable. Despite the general notion that main clauses code foreground information (or express a main assertion) while subordinate clauses code background information (or less profiled information), our examination shows that some subordinate clauses do code more important foreground information. Subordinate clauses are not by any means a unitary category; they display differing degrees of subordination. My goals, therefore, are to delineate differences between the two complementizer uses, and to show that clauses with to and koto present differing degrees of subordination, and that there is a correlation between degree of subordination and information status.
1. Introduction1 The status of subordinate clauses2 and subordination has been of interest to linguists for some time. Davidson (1979: 106–107), for example, offers the following criteria for subordinate constructions:
1. This paper was originally presented at the Symposium on Functional Approaches to Japanese Grammar, held at the University of Alberta, August 20–22, 2004. I am grateful to all the participants of the symposium for their comments. I am particularly indebted to Shigeko Okamoto for her comments on an earlier version of the paper, and editors of this volume, Tsuyoshi Ono and Kaori Kabata, for their detailed comments on the draft version of this paper. I am also grateful to the anonymous reviewer of this volume for insightful comments. 2. Matthiessen and Thompson 1988, for one, distinguishes two types of subordination – embedding (e.g., relative clauses) and clause combining (clauses connected by connectives, conditional/ temporal phrases, etc.). In this paper, I am using ‘subordination’ to cover both of these types.
Naomi H. McGloin
(1) Syntactic: a. non-finite verb b. complementizer, conjunction, relative pronoun (This means relative clauses, complement clauses and adverbial clauses are considered to be subordinate clauses.) c. special word order, impossibility of inversion (in English), presence of special mood d. ability to undergo movement as a constituent e. backwards pronominalization, deletion of subject by Equi, etc. f. proper inclusion within another clause Semantic: a. dependency of thought; greater ‘cohesion’ of constituents b. contents of the clauses are presupposed to be true, or not a separate assertion c. contents of the clause are in the scope of some higher predicate; negation, question, etc. d. clause plays a grammatical role in another clause; subject, object, etc. Pragmatic: a. clause expresses background information b. clause functions as a modifier of some other sentence, adjacent in discourse In addition, it has been claimed that subordinate clauses are more conservative and resist innovations. Matsuda (1998), for example, examines two competing potential suffixes (conservative –(tabe)rareru form and innovative –(tabe)reru form) and finds a statistically significant difference in their occurrences: the innovative form is more advanced in main clauses than in embedded clauses (p < .008). The relationship between the subordinate status of clauses and the information structure – how main and subordinate clauses code given and new information, or background and foreground information, has been the center of some controversy (Givón 1979; Fox & Thompson 1990; Tomlin 1985; Prideaux 1993). The conventional view is that subordinate clauses are presupposed, and hence should contain given information while new information is expected to reside in main clauses (Bever 1969, 1970; Givón 1979). Langacker (1991: 436) states, “A subordinate clause is ... one whose profile is overridden by that of the main clause.” Prideaux (1993: 57), although he actually argues against the following view, states: A further factor which might also be at work here is some version of the principle of iconicity (Haiman 1985; Givón 1989), according to which the more central, important, and new information could be placed iconically in the more salient
Subordination and information status
main clause, which of course can stand alone, while the less important given information finds itself relegated to an iconically less salient subordinate clause status. A similar possibility seems to hold for the proposed foreground-background distinction, where foreground information tends to be found in main clauses and background information in subordinate clauses.
Tomlin (1985) finds that the aforementioned correlation between information status and main-subordinate distinction holds, i.e., “independent clauses code foreground and pivotal information; dependent clauses code background information.” (p. 85) More work, however, has shown that this correlation does not always hold true (Schleppegrell 1992; Thompson 2002; Prideaux 1993). Prideaux (1993), for example, examines adverbial and relative clauses, both in oral and written narratives, and finds that there is a significant tendency for relative clauses to encode new information as well as a strong tendency for adverbial clauses to represent new information. The problem here, of course, is that notions of foreground/background information and new/old information do not always coincide. It is possible that clauses containing new information create a backgrounding effect in discourse. We will examine notions of foreground/background information later in the next section. Thompson (2002), moreover, claims that some complement clauses, which are grammatically subordinated, actually contain the main assertion of the utterance. Based on an examination of actual conversations, she argues that “in the majority of cases, the complement ‘overrides’ the ‘main clause’, and the ‘main clause’ is there to provide the speaker stance towards the assessments, claims, counterclaims, and proposals” (2002: 134). In other words, the complement clause represents foreground information. In the present paper, I will examine one type of subordinate construction in Japanese, i.e., complement clause. In particular, I will investigate cases where two different complementizers to and koto are interchangeable. My goals are two fold – (1) to delineate differences between the two complementizer uses, and (2) to show that clauses with to and koto present differing degrees of subordination, and that there is a correlation between degree of subordination and information status. In particular, I would argue that to complementizer, being less subordinate than koto, has an effect of foregrounding information presented in the complement clause. 2. Foreground vs. background information The notion of foreground vs. background information has been used in analyzing narrative organization (Longacre 1976; Hopper 1979). Hopper (1979: 213–214) defines “the parts of the narrative which relate events belonging to the skeletal
Naomi H. McGloin
structure of the discourse” as foreground, and the background as what “amplif[ies] or comment[s] on the events of the main narrative.” So, in (2), the gerund clause represents a background event, as opposed to the main clause, which represents a foreground event. We journeyed for several days, passing through a few villages. (Hopper 1979: 215) Applying these concepts beyond analysis of narrative structure, Tomlin (1985: 89) characterizes foreground information as “information which is more central or salient or important to the development of the discourse theme” and background information as “that which elaborates or develops foreground information.” In the present paper, I will propose that a certain construction has a ‘foregrounding effect’ in the sense that the speaker, by using this particular construction, presents a piece of information as particularly important to the main theme/ argument/claim or worthy of attention, thereby emphasizing or strongly asserting this information toward the addressee. On the other hand, a ‘backgrounding effect’ obtains when information is presented as expected or obvious or subordinate in importance. Observe the following sentences. (3) A: Hasami doko ni oita kana? scissors where at put-pst I wonder ‘Where did I put the scissors, I wonder.’ B: Soko ni aru janai.3 there at exist ‘It’s there, don’t you know?’ (4)
Dooshi toka demo, koko dattara ekaado o Verbs like even here cop-if picture card acc tsukatte yaru janai desu ka? use do Sooyuu no mo nai kara, ....4 such thing one also not exist because ‘In regards to verbs, here we practice using picture cards, right?, but since they don’t have such things, so ....’
In (3), janai has an expressive function, and B’s utterance ending with janai foregrounds the information. Here, B is pointing out the fact that the scissors are (right) there to A, who does not notice them. In (4), on the other hand, janai has a textual function. In (4), janai is used textually to provide an anaphoric reference 3.
This is taken from Ide (1984).
4. This example comes from a conversation between two native speakers of Japanese, which was recorded at an American university as part of a research project on Japanese discourse.
Subordination and information status
for the following pronominal expression (sooyuu no ‘such thing’). In the sense that the sentence ending with janai is there to introduce information for further comments, it can be said that this information is subordinate in importance to what follows, and hence is backgrounded.5 3. Degree of subordination It has been noted that various embedded clauses manifest different degrees of subordination to the main clauses (Kuno 1973; Minami 1974; Masuoka & Takubo 1992; Moriyama 2000). Moriyama classifies subordinate clauses into three types. Clauses which cannot have an independent subject, as in (5), are considered to have the greatest degree of subordination. (5) a. Taroo ga terebi o mi-nagara, benkyoo suru. nom tv acc see-while study do ‘Taroo studies while watching TV.’ b. *Taroo ga terebe o mi-nagara, Hanaoko ga benkyoo suru. ‘*Hanako studies during Taroo’s watching TV.’ Temporal and conditional clauses are considered to be less subordinate since they can have an independent subject different from the main clauses. However, they cannot embed modals (e.g., daroo ‘probably’) or the thematic particle wa, which express speaker subjectivity, as in (6b) and (6c). (6) a. Hanako ga ie ni kaeru to, Taroo ga nom home to return when nom sake o yooishita. sake acc prepared ‘When Hanako got home, Taroo served sake.’ b. *Hanako ga ie ni kaeru daroo to, Taroo ga sake o yooishita probably ‘*When Hanako probably went home, Taroo served sake.’ c. *Hanako wa ie ni kaeru to, Taroo ga sake o yooishita. top ‘As for Hanako, when she got home, Taroo served sake.’
5. McGloin (1999) argues that this textual function of janai (desu ka) is a relatively recent usage. See McGloin (1999) for the expressive and textual functions of janai (desu ka).
Naomi H. McGloin
Clauses such as -kara ‘because’ and -keredo ‘but’ are the least subordinate, and they share many syntactic characteristics with an independent clause. For instance, just like independent clauses, kara and keredo clauses not only can have an independent subject but also embed modals and the thematic particle wa, as in (7). (7) a. Heya wa shimatte iru kara, Taroo wa inai. room top close is because top exist-neg ‘The door to the room is shut, so Taroo is not in.’ b. Isogashii daroo kara, ato de denwa o suru. busy probably because later telephone acc do ‘Since you are probably busy, I will call you later.’ Now, what distinguishes even the least subordinate clauses such as -kara and -keredo from an independent clause is the fact that these clauses do not embed sentence final particles such as ne and yo, which often occur at the end of an independent sentence.6 (8) a. Kono hon wa omoshiroi yo. this book top interesting fp ‘This book is interesting, I tell you.’ b.
*Kono hon wa omoshiroi yo this book top interesting fp kara, yonde iru. because read be ‘*Since this book is interesting, I tell you, I am reading it.’
This can be summarized in Table I. Table I. More ← subordinate → less nagara ‘while’ tsutsu ‘while’ Can S1 and S2 have different subjects? Can S1 embed modals (e.g., daroo)? Can S1 embed a thematic particle wa? Do sentence final particles appear in S1?
NO NO NO NO
-to ‘when’ keredo ‘but’ tara, ba ‘if/when’ kara ‘because’ YES NO NO NO
YES YES YES NO
6. These clauses also do not embed imperatives, volitional forms, which occur in independent clauses.
Subordination and information status
4. Object complement clause in Japanese In the following, we will examine object complement construction in Japanese and show that a complement clause, which is syntactically subordinate to the main clause, does carry the main import of the sentence. It will be shown that there is a correlation between foregrounding effect and degree of subordination. 4.1
Overview
In Japanese, object complement clauses can be marked by to, koto, or no, as in (9). (9) a. John wa Mary ga nihongo o hanasu to itta. top nom Japanese acc speak cp say-pst ‘John said that Mary speaks Japanese.’ b. John wa Mary ga nihongo o hanasu koto o shitte iru. top nom Japanese acc speak cp acc know be. ‘John knows that Mary speaks Japanese.’ c. John wa Mary ga nihongo o hanasu no o kiita. top nom Japanese acc speak cp acc hear-pst ‘John heard Mary speaking Japanese.’ Kuno (1973: 213) observes: The koto and no clauses represent an action, state, or event that the speaker presupposes to be true, while the to clause represents an action, state, or event that does not have such a presupposition.
So, in (9a), the speaker does not presuppose the complement sentence to be true – i.e., Mary might or might not speak Japanese, while in (9b) and (9c), the speaker knows that it is true that Mary speaks Japanese. Hence, it is generally the case that to is associated with non-factivity, and koto and no with factivity. To marks an object complement clause of nonfactive predicates, such as verbs of saying, verbs of thinking and counter-factual verbs, while koto/no mark a complement clause of factive predicates. Furthermore, the difference between koto and no is said to be abstract vs. concrete. When the complement clause represents an abstract idea such as the complement of the verb ‘to learn,’ koto is used. On the other hand, for the complement clauses of verbs of sense perception, such as ‘to see’, ‘to hear,’ etc., which involve concrete and direct experiences, no is used.7
7. For more detailed discussions of factivity and complementation, see Kuno (1973), Suzuki (1997) and McCauley (1978).
Naomi H. McGloin
For the present paper, it is also important to note that some of the factive predicates – i.e., what are generally called ‘semi-factive’ predicates8 (e.g., ‘to know’, ‘to learn’, ‘to find out’, ‘to realize’) can mark their complement clauses either with to or koto, as in the following examples. (10) a. Tokyo wa bukka ga takai koto o shitte imasu. top cost nom expensive that acc know-be ‘I know that the cost of living is expensive in Tokyo.’ b. Tokyo wa bukka ga takai to shitte imasu. ‘I know that the cost of living is expensive in Tokyo. Since it is generally assumed that the difference between to and koto is the matter of factivity, it was somewhat of a mystery why the non-factive complementizer to sometimes co-occurs with factive predicates, namely those which are classified as semifactives. Kuno (1973: 217), for one, observed, “For some mysterious reason, siru ‘to get to know’ can be used with to in certain contexts.” (see Section 4.3 for more discussion on this.) In the following discussion, we will zero in on cases like (10). We will not examine the differences between koto and no. This is because the focus of this paper is to examine the relationship between degree of subordination and foregrounding effect and for this purpose we will be focusing on verbs which allow alternation between to and koto. 4.2
To vs. Koto: Syntactic differences and degree of subordination
To is generally said to mark a quotation, and as such it can introduce both direct and indirect quotations.9 In a case like (9a), however, it is considered to be a case 8. Semifactives are different from true factives in that “it is possible to construct sentences in which the truth of a semifactive complement cannot be inferred from the entire sentence.” (Hooper 1975: 114) So, if one says, “it is possible that I will find out that I have not told the truth,” the complement sentence is not presupposed to be true – even the speaker does not know whether he has told the truth or not. (cf. Karttunen 1971; Hooper 1975) 9. Direct quotations are generally distinguished from indirect quotations in Japanese by the use of a logophoric quotation marker. Moreover, if the to-marked sentence has final particles, polite desu/masu endings and deictic words such as ashita ‘tomorrow’, it can only be a direct quote. (i) is a case of a direct quote and (ii) an indirect quote. (i) Tanaka wa ‘Kyoo sensei ni aimashita yo’ to itta. top today teacher with meet-pst fp qt Say-pst ‘Tanaka said, “I met my teacher today.’’’ (ii) Tanaka wa sono hi sensei ni atta to itta. top that day teacher with meet-pst qt say-pst ‘Tanaka said he met his teacher that day.’ or more discussion on this, see Coulmas (1985), Maynard (1986), and Kamada (1988), among F others.
Subordination and information status
of indirect quotation and here to is considered to be an object complementizer in that it serves to link a complement clause to the main clause. In the following, we will examine how to and koto behave differently syntactically and demonstrate that to complement clauses show more syntactic independence than koto complement clauses. Applying three criteria for degree of subordination proposed by Moriyama (cf. Section 3), both to and koto clauses can have a different subject from that of the main clause. Second, to clauses embed certain types of modal expressions, while koto clauses do not. (11) a. Sekai wa heiwa ni naru daroo to omotta. world top peaceful become probably cp think-pst ‘I thought the world would probably become peaceful.’ b. Sekai ga heiwa ni naru daroo to wakatte ita. world nom peaceful become probably cp knew ‘I knew that the world would probably become peaceful.’ c. *Sekai ga heiwa ni naru daroo koto ga wakatta. world nom peaceful become probably cp nom know-pst ‘I knew that the world would probably become peaceful.’ (12) a. Nihon e ikoo to omotta Japan to go-volitional cp think-pst ‘I thought I would go to Japan.’ b. *Nihon e ikoo koto o kikaku shita. Japan to go-volitional cp acc plan do-pst ‘I planned that I would go to Japan.’ As for sentence-final particles, varying degrees of acceptability are observed. To complement clauses of non-factive verbs seem to accept sentence-final particles, as in (13a) and (13b). It is questionable in to complement clauses of semi-factive verbs, as in (13c). It is not acceptable in koto clauses of either semi-factive verbs or factive predicates as in (13d). (13) a. Benkyoo sureba dekiru yo to omou. study do-if can do fp cp think ‘I think (I) can do it if I study.’ b. Omoshiroi ne to omotta. interesting fp cp think-pst ‘I thought it was interesting.’
Naomi H. McGloin
c. *Omoshiroi yo to wakatte imasu. interesting fp cp know-be ‘I know it is interesting.’ d. *Omoshiroi yo koto ga wakatte imasu. So, according to Moriyama’s criteria, to clauses are actually less subordinate than koto clauses. There are, moreover, other syntactic differences between to complement clauses and koto clauses . For example, to complement clause can be pronominalized by soo, while koto clauses cannot, as in (14). (14) a. John wa nihongo ga yasashii to omotte top Japanese nom easy cp think iru. Watashi mo soo omou. be I also so think ‘John thinks Japanese is easy. I think so too.’ b. John wa amerika no da�gaku ga taihen top American lk college nom hard da to wakatte iru. cop cp know-be ‘John knows that American colleges are difficult.’
Watashi mo soo wakatte iru. I also so know-be ‘I also know so.’
c. John wa kuroo shita koto o wasurete iru. top hardship do-pst cp acc forget be *Watashi mo soo wasurete iru. I also so forget be ‘John forgot that (he) had a hard life. *I forgot so too.’ Soo is a sentential pronoun, and the fact that it can replace a to clause indicates that the to clause has a sentential status and thus is more independent than the koto clause.10 We can summarize these observations as follows: 10. It is interesting to note that the subject of the to complement clauses can be raised to the main object position, as in (iii), but not the subject of the koto clause (cf. Masuoka & Takubo 1992: 186; Kuno 1976). (iii) a. Nihonjin o kinben da to omoimasu ka? Japanese acc diligent cop cp think Q ‘Do you think Japanese are diligent people?’
Subordination and information status
Table II. More subordinate ← → less subordinate Complementizer Verb types Different subjects? Modal expressions? Soo pronoun? Sentence-final Particle
koto
to
True factives ‘forget’ Semifactives ‘know’
Semifactives ‘know’
Non-factives ‘think’
YES NO NO NO
YES YES YES NO
YES YES YES ?YES
To the extent that to complement clauses can embed modal expressions, they are more like independent sentences, and hence, we can say that to complement clauses show more independence as a clause than koto complement clauses. To clauses with semi-factive predicates, moreover, seem more subordinate than to clauses with non-factive predicates since the former does not allow sentence-final particles. Then, what is the difference between to clauses with non-factive predicates and independent sentences? To complement clauses of non-factive predicates are still subordinate in nature because (1) they cannot stand alone, (2) occurrence of sentence-final particles is not as free as in independent sentences, and (3) polite sentence endings, including polite request forms (e.g., -te kudasai) and V-mashoo ‘let’s V’ form, do not occur in to complement clauses. The use of these polite forms in to clauses generally turns the clauses into direct quotes. 4.3
To vs. Koto: Functional differences
In this section, we will try to delineate functional differences between complementizers to and koto. For this purpose, we will focus on those cases where alternation between to and koto is observed. In Section 4.1, we noted that the semifactive predicates (such as ‘to know’, ‘to learn’, ‘to find out’, etc.) take either to or koto as their complementizers. By examining these cases, we will propose that the functional difference between to and koto is not factivity, nor ‘distance’ as is proposed by Suzuki (1997), but rather that of foregrounding/backgrounding effects. b. *John o bengoshi de aru koto o wasurete ita. acc lawyer cp acc forget be-pst ‘I forgot that John is a lawyer.’ In (iiia), nihonjin is the object of the main verb omou, and at the same, it is the subject of the complement sentence kinben da.
Naomi H. McGloin
4.3.1 To as a marker of psychological distance Suzuki (1997) gives an interesting account for the co-occurrence of to and these semifactive predicates. Observe the following examples from Suzuki (1997: 298). (15)
Sore wa itsuka kaette kuru to ikura that top some day return come cp how much shitte ite-mo, hakkiri mimi ni suru made kesshite te know:is:even if clearly hear until never hand ni hairanai ‘jikkan’ datta. in enter:NEG real feeling was ‘Although I knew very well that he was coming home some day, I did not feel that would actually happen until I heard it with my own ears.’
(16)
Chishiki to shite wa ka de wa utsuranai tte11 knowledge as mosquitoes by top transmit:neg CP wakatte ite-mo kininarimashita. know:is:even if was concerned ‘Although I rationally knew that it was not communicable by mosquitoes, I was still worried (Suzuki 1997: 298).’
Using Frajzynger (1991)’s notions of de dicto (the domain of speech) and de re (the domain of reality), Suzuki argues that to, which also marks quotation, is characterized as a de dicto complementizer. Since what is quoted or reported is viewed by the speaker as belonging to someone else, she argues that, when the speaker uses to, s/he feels distanced from the information presented in the complement clause. The main function of to, she argues, “is to suggest the speaker’s psychological distance from the information expressed in the complement clause” (2001: 37). According to Suzuki, the speaker is psychologically distanced from the following kinds of information: (17) a. information which the speaker regards (or presents) as that belonging to others b. information which the speaker does not entirely consider to be true c. information which the speaker disapproves of in content d. information which is viewed by the speaker objectively In both (15) and (16), the complement clauses represent facts. However, a nonfactive to occurs. This is because the speaker is not entirely convinced of the reality 11. In (16), tte can be considered as a colloquial variant of to. The use of tte, however, is broader than to, and it is not certain if tte should indeed be analyzed simply as a variant of to. For discussions of tte, see Okamoto (1995) and R. Suzuki (1999), among others.
Subordination and information status
of the information contained in the complement clause. Sentence (16), for example, is said by a woman who went to Africa, where AIDS was an epidemic. While she intellectually knew that the AIDS virus is not spread by mosquitoes, emotionally she could not fully integrate this knowledge into her belief system, and in this sense, she feels distanced from this fact. This is an interesting account, and accounts for why to is used in the following example also. (18)
Mausu ga warui to wakatte itemo, (gachagachagacha) to yatchau. mouse nom bad cp know-be-even if (rattling noise) do Wakatte ru n da kedo.12 know:be nml cop but ‘Although I know that the (computer) mouse is bad, I still move it. I know (it’s bad), but..’
Before (18), the speaker said that she wished that somehow her computer get fixed miraculously. The speaker knows that the problem is a bad computer mouse, but she cannot resist moving it, just wishing it worked. So, we could say that here the speaker again has not integrated this fact into her belief system, or to is used here because the speaker disapproves of its content. 4.3.2 To as a device to foreground information When we look at actual discourse data,13 there are many examples of to where the speaker knows that the information represented in a complement clause is true and no psychological distance is detected on the part of the speaker. (19)
Hontoo wa ii ko da to wakatte ita node, chuui reality good child cop cp know-state-Past because caution shitara, sono tsugi kara wa chikoku mo nakunari, totemo do-when that next from to be late also cease very tsukiai yasuku narimashita yo. associate easy became fp (http://www.minaminippon.co.jp, 07/29/02) ‘I knew that (he) was actually a good kid, so I gave (him) a word of caution. Then, the following day, he stopped being late (for class) and (he) became easy to deal with.’
12. This example comes from the author’s field notes. 13. The majority of the data come from internet web sites. They were collected using the google search engine.
Naomi H. McGloin
(20) Kitakuni to onaji sharyoo to wakatte ita node joosha ga with same train(car) cp know be-pst because riding nom tanoshimi deshita. looking forward to was (http://homepage1.nifty.com) ‘I knew that it was the same train (car) as that of Kitakuni, so I was really looking forward to riding it.’ In (19) and (20), the complement clauses do not represent information that the speaker feels distanced from. On the contrary, the speaker seems quite convinced of the information presented in the complement clauses. In (19), the speaker is strongly convinced that the kid has a good heart (although others might not think so, as indicated by the use of hontoo wa), and that this is a positive characteristic, not something that the speaker disapproves of. Similarly, in (20), the speaker knew that the train s/he was taking was the same as another special train, and since s/he believes this to be true and is not wishing it otherwise, there is no reason for the speaker to feel distanced from the information. The actual data show that the cases where to is used with these semifactive predicates are when the speaker has some reason to ‘foreground’ or emphasize the content of the complement clauses. Observe the following example. (21) Isetan no mae de takushii ni noroo to shite, takushii noriba lk front at taxi in try to ride do taxi station ni iru kakari no hito ni hikoojoo made ikura? at is person in charge lk person to airport to how much to kiitara, 400 baatsu to iwarete, bikkuri. Saki ni cp ask:when cp say:passive surprised beforehand shitashirabe shite, investigation do → shinai kara kuukoo made wa 200 baatsu kurai to downtown from airport to top about cp shitte ita node meetaa de iku to ittara, takushii know:was because meter by go cp say:when taxi wa ok shite kurete, meetaa wa 180 baatsu deshita. top do give meter top cop:pst
‘I wanted to get a taxi in front of Isetan, so I asked the person in charge of the taxi how much it would cost to the airport. (He) said it would be 400 baht, and I was very surprised. I had done some homework before (the trip) and knew that it would be about 200 baht, so I asked the taxi driver if he would use the meter fare. He said OK, and the meter fare was 180 baht.’
Subordination and information status
In this passage, the speaker, who was traveling in Thailand, is talking about the taxi fare. The fact that it should be about 200 baht is what he/she had found out prior to the trip, and the speaker believes that this information is correct as opposed to what she was told by the taxi stand attendant. It does not seem that the speaker feels distanced from what he/she knew prior to the trip. There was, on the other hand, a discrepancy between what he/she had expected and what he/she was told, and since what he/she knew was the reason why he/she decided to go by meter rather than by a fixed fee, there is reason why the speaker might feel the need to assert/highlight (i.e., foreground) what he/she knew--i.e., the taxi fare should be about 200 baht. 4.3.3 To vs. Koto: Foregrounding/backgrounding effect My examination shows that to tends to be used with the verbs of knowing where there is some contrasting information or where the speaker has a need to impress a particular information on an addressee. Koto, on the other hand, has the effect of keeping an information in the background, as something expected, predictable, or taken for granted. What is involved here is not the semantic notion of assertion vs. presupposition, but rather a more pragmatic one, perhaps similar to what Lambrecht (1994: 66) calls ‘pragmatic accommodation’. Lambrecht talks about ‘accommodation for presupposition’ where the speaker/writer can create the presupposition in the hearer/reader’s mind by using the clause which requires the presupposition, and thereby “making it available as a background for the assertion in the following main clause.” (p. 69) In this sense, koto creates a backgrounding effect, while to creates a foregrounding effect. The following paragraph presents an interesting contrast between the use of to and koto. (22)
‘Conscientious railroad crossing’ (www2.justnet.ne.jp/~assoonas/OR33.HTML 5/6/02) Kono shigatsu (Heisei 13 nen) ni ten’nin ga atte, This April 13th year of Heisei in job transfer nom was tsuukin keiro ga kawatta no da ga, sono tochuu ni commuting route nom changed nml cop but that on the way at juutai ga hidoi fumikiri ga aru. traffic jam nom bad railroad crossing nom exist ‘This April, I was transferred to another office, and now I have to take a different route to my work. On the way, there is a railroad crossing with frequent traffic jams.’ (sentences omitted)
Naomi H. McGloin
ichijikan no uchi ippun teido wa ressha tsuuka one hour lk within one minute about top train passing no tame ni shadanki ga oriru no da for the purpose of crossing gate nom come down nml cop ga ato no 59fun wa mattaku ressha ga tooranai but rest lk 59 minutes top at all train nom pass-neg jikan na no da. time nml cop ‘The crossing gate comes down for one minute per hour so that the train can pass, but the rest of the hour – for 59 minutes, no train passes by.’ a →
Shikashi, mattaku ressha ga konai koto ga wakatte but (not) at all train nom come:Neg cp nom know itemo, dooro kootsuuhoo o mamoru to yuu koto ni be:even if traffic rule acc observe to naru to, fumikiri no chokuzen de kuruma wa become when crossing lk before at car top ichiji teishi o shinakereba naranai. momentary stop acc do:have to ‘But, even if we know that the train is not coming, we have to make a stop in order to observe the traffic rules.’ (sentences omitted) Sokode, fumikiri o tsuukasuru kuruma wa, ressha So, crossing acc pass through car top train ga konai nom come:neg
koto ga wakatte iru jikantai demo majimeni cp nom know be time even conscientiously ichijiteishi o suru node, onozuto kuruma no nagare stop acc do because naturally car lk flow ga waruku naru. nom bad become ‘Therefore, cars which pass through the crossing make a stop even during the time when we know that the train does not pass. Hence, naturally, the cars get backed up.’ (sentences omitted)
b →
c →
Ressha ga konai to wakatte iru jikan ni wa ichijiteishi train nom com-neg cp know be time at top stop o shinaide toorinukeru no ga gooriteki acc without doing pass through nml nom reasonable
Subordination and information status
d→
da to omou shi, soo sureba juutai ga okiru cop cp think and so do:if traffic jam nom happen koto ga nai. that nom not exist ‘I think it’s more reasonable to cross the railroad tracks without making a stop when we know that the train is not coming. If we do that, there won’t be any traffic jam.’ (omission) Ressha ga konai to wakatte iru jikan ni wa, train nom come:neg cp know be time at top jooji aoiro ni shite, kuruma ga teishi shinaide always green dat do car nom stop without doing toorinukerareru yoo ni shite oki, ressha ga tooru pass through:can dat do train nom pass toki dake kishingoo akashingoo ni sureba yoi no dearu. time only yellow light red light do:if good nml cop ‘When we know the trains are not coming, we should keep the light green all the time so that cars can pass through without making a stop. It should be fine to change the light to yellow or red only when the trains pass through.’
In this passage, the writer talks about the railroad crossing where cars have to come to a complete stop, thereby contributing to the backup of cars. The writer thinks it unreasonable that cars are required to come to a complete stop every time they pass through this crossing, since the trains pass by only once every hour. In this passage, the same phrase meaning ‘we know that the train does not pass by’ is used four times. In the first two instances (arrows a and b), koto is used as in ressha ga konai koto ga wakatte iru ‘(one) knows that the train is not coming’, while in the last two instances (c and d), to is used as in ressha ga konai to wakatte iru. It is true that grammatically both koto and to are acceptable in all four instances, but then why did the writer choose koto for the first two instances and to for the last two instances? This must not be simply accidental. I would argue that koto is used in the first two instances because the writer is simply stating the fact matter-of-factly, setting up a stage for what follows (i.e., backgrounding effect obtains). In the last two instances, on the other hand, the writer is trying to make the point that people should be allowed not to make a stop during the time period when people know that the train does not pass by. In other words, the information presented in the complement clause – i.e., that the train does not pass by – is crucial for the writer’s claim, and hence, the writer needs to impress this piece of information on the
Naomi H. McGloin
addressee (i.e., foregrounding effect obtains). I would argue that this subtle psychological attitude is reflected in the use of to and koto. Sunakawa (1988, 1989) argues that to, being a quotation marker, introduces the ‘dual places’ – the quoting itself and the quoted event. In case of verbs of judgment, what is quoted is the speaker’s own thought. Because of this ‘dual’ nature, to can present the event as a proposition which is experienced by the subject at the moment of ‘thought.’ Koto, on the other hand, does not have this ‘dual’ nature, and thus can only present an event “that the quoter has objectified and conceptualized.” (Maynard 1996: 208) I believe this is why to can infuse the speaker’s subjective stance, thereby emphatically present (i.e., foreground) an information. 5. Conclusion We started this investigation with the idea that main clauses code foreground information (or express a main assertion) while subordinate clauses code background information (or less profiled information). Our examination of some Japanese subordinate clauses has shown that this is not always the case – subordinate clauses do code more important foreground information. This supports recent findings which question the categorical distinction between main vs. subordinate clauses and, consequently, the idea that subordinate clauses are also subordinate or secondary in importance to the main clause (Haiman & Thompson 1984; Thompson & Mulac 1991; Englebretson 2003; Thompson 2002). An examination of subordinate clauses in Japanese, moreover, has shown that the subordinate clause is not by any means a unitary category, but that subordinate clauses display differing degrees of subordination. The correlation between independence of clauses and information status still seems to hold with respect to the degree of subordination – the more independent clauses are, the more foregrounding effect it projects. It was shown that, between the two complementizers in Japanese – to and koto, to complement clause shows less degree of subordination (or greater independence) and it has a foregrounding effect, while the koto clause with a greater degree of subordination shows a backgrounding effect. It has been pointed out that the meanings and functions of the complementizer to has a lot to do with the fact that it also serves as a quotation marker (Hayashi 1997; Maynard 1986; Suzuki 2001). In the past studies, however, this fact has been used to argue that to clauses introduce psychological distance because the quoted information is ascribed to someone other than the speaker. Suzuki (2001: 44), for example, states, “This grammatical non-incorporation reflects the psychological distance (i.e., psychological non-incorporation) that the speaker feels toward the information in the to-complement”. Along a similar line, Hayashi (1997: 580)
Subordination and information status
proposes that the core function of to is “invoking ‘dual voice’ in the utterance and shifting authority and responsibility for the proposition from the speaker who utters it”. These characterizations of to miss the point that a quoted utterance can also be the speaker’s own voice, for which the speaker takes full responsibility. Through examination of to and koto where they are interchangeable, I have argued that to has an opposite effect to ‘distancing’. By choosing to, the speaker can inject his/her strong subjective stance and this results in foregrounding information as central or important to the development of his/her point/theme. So, what does this study say about the nature of Japanese grammar and grammar in general? I would like to offer four observations. 1. This study offers an additional piece of evidence for the non-discrete nature of linguistic categories. It shows that not only the demarcation between main and subordinate clauses is non-discrete, subordinate clauses themselves are not unitary in nature. (cf. Okamoto & Ono 2008.) 2. Recent studies in discourse and conversation have clearly shown that there is a great deal of difference between our conception of grammar and the way we actually talk. The traditional concept of ‘sentence’, for example, has been questioned because speakers do not generally talk in full sentences consisting of a subject, an object and a predicate (Du Bois 1987; Ono & Thompson 1996; Thompson & Hopper 2001; Iwasaki & Ono 2002, just to mention a few). While I fully agree with the importance of looking at a particular construction in context, this primacy of spoken data (or conversational data) might be problematic. The complementizer to is a case in point. Although I knew that speakers use the complementizer to with semifactive verbs, I was not able to find any example in the many conversation data I had. I once heard an example of it, but this was not part of the recorded conversations. The google search, on the other hand, yielded a large amount of examples. So, the data used in this study were all from written discourse. The question, then, is: are there two grammars – grammar of written (Japanese) language and grammar of spoken (Japanese) language? Are the findings of the present study pertinent only to the grammar of written Japanese? 3. Related to the above question is the speakers’ intuition vs. what they actually do. The native speakers of the language can make reliable judgments about many constructions they rarely hear or utter. For example, the speaker probably rarely hears a sentence such as ‘Hanako ga yatta to hayagatenshita. (I misunderstood that Hanako did it)’, but a native speaker has no problem telling that, with this verb, to is fine but not koto. Native speaker’s intuition has to be part of our knowledge of grammar. As Newmeyer (2003: 692) puts it, “...there is a lot more to grammar than can be predicted from use in naturally occurring discourse.”
Naomi H. McGloin
4. Lastly, Japanese is the language where the speaker/writer’s subjective attitudes/ stance are grammaticalized in various forms, such as sentence-final particles, conjunctions, conditionals, etc. (Akatsuka 1985; Iwasaki 1993; McGloin 1977, among others.) A connective noni ‘although’ is more subjective while keredo ‘but’ is more objective, loosely speaking. The complementizer to, then, in a way, expresses the speaker/writer’s subjective stance, while koto indicates an objective stance. It seems to me that any grammar of Japanese would have to be able to account for the subjective/objective dichotomy, which seems to cut across a broad spectrum of phenomena in Japanese. List of abbreviations ACC CP COP DAT FP LK NEG
Accusative Complementizer Copula Dative Final Particle Linking Nominal Negative
NML NOM PST Q QT TOP
Nominalizer Nominative Past Question Quotation Topic marker
References Akatsuka, Noriko. 1985. Conditionals and the epistemic scale. Language 61(3): 625–639. DOI: 10.2307/414388 Bever, Thomas G. 1969. The comprehension and memory of sentences with temporal relations. Advances in Psycholinguistics, Giovanni B. Flores d’Arcais & Willem J.M. Levelt (eds), 285– 293. Amsterdam: North Holland. Bever, Thomas G. 1970. The cognitive basis for linguistic structures. Cognition and the Development of Language, John R. Hayes (ed.), 279–263. New York NY: John Wiley. Coulmas, Florian. 1985. Direct and indirect speech: General problems and problems of Japanese. Journal of Pragmatics 9: 41–63. DOI: 10.1016/0378-2166(85)90047-5 Davidson, Alice. 1979. Some mysteries of subordination. Studies in the Linguistic Science 9(1): 105–128. Du Bois, John. 1987. The discourse basis of ergativity. Language 63: 805–55. DOI: 10.2307/ 415719 Englebretson, Robert. 2003. Searching for Structure: The Problem of Complementation in Colloquial Indonesian Conversation [Studies in Discourse and Grammar 13]. Amsterdam: John Benjamins. DOI: 10.1075/sidag.13 Fox, Barbara & Thompson, Sandra A. 1990. A discourse explanation of the grammar of relative clauses. Language 66: 297–316. DOI: 10.2307/414888
Subordination and information status Frajzynger, Zygmunt. 1991. The de dicto domain in language. In Approaches to Grammaticalization, Vol. 1 [Typoloical Studies in Language 19(1)], Elizabeth Closs Traugott & Bernd Heine (eds), 219–251. Amsterdam: John Benjamins. Givón, T. 1979. On Understanding Grammar. New York NY: Academic Press. Givón, T. 1989. Mind, Code and Context: Essays in Pragmatics. Hillsdale NJ: Lawrence Erlbaum Associates. Haiman, John. 1985. Natural Syntax: Iconicity and Erosion. Cambridge: CUP. DOI: 10.1075/tsl.6 Haiman, John & Thompson, Sandra A. 1984. Subordination in Universal Grammar. BLS 10: 510–523. Hayashi, Makoto. 1997. An exploration of sentence-final uses of the quotative particle in Japanese spoken discourse. In Japanese/Korean Linguistics 6, Ho-Min Sohn & John Haig (eds), 565–581. Stanford CA: CSLI. Hooper, Joan B. 1975. On assertive predicates. In Syntax and Semantics 4, John Kimball (ed.), 91–124. New York NY: Academic Press. Hopper, Paul. 1979. Aspect and foregrounding in discourse. Syntax and Semantics 12, T. Givón (ed.), 213–241. New York NY: Academic Press. Ide, Sachiko, Ikuta, Shôko, Kawasaki, Akiko, Hori, Motoko & Haga, Hitomi. 1984. Syuhu no Issyuukan no Danwa Siryoo. Tokyo: Shuuei-shuppan. Iwasaki, Shoichi. 1993. Subjectivity in Grammar and Discourse. Theoretical Considerations and a Case Study of Japanese Spoken Discourse [Studies in Discourse and Grammar 2]. Amsterdam: John Benjamins. Iwasaki, Shoichi & Ono, Tsuyohi. 2002. ‘Sentence’ in spontaneous spoken Japanese discourse. In Complex Sentence in Grammar and Discourse: Essays in Honor of Sandra A. Thompson, Joan L. Bybee & Michael Noonan (eds), 175–202. Amsterdam: John Benjamins. Kamada, Osamu. 1988. Nihongo no dentatsu hyoogen (Reportive expressions in Japanese). Nihongogaku 7: 59–72. Karttunen, Lauri. 1971. Some observations on factivity. Papers in Linguistics 4: 55–69. DOI: 10.1080/08351817109370248 Kuno, Susumu. 1973. The Structure of the Japanese Language. Cambridge MA: The MIT Press. Kuno, Susumu. 1976. Subject raising. In Syntax and Semantics 5: Japanese Generative Grammar, Masayoshi Shibatani (ed.), 17–49. New York NY: Academic Press. Lambrecht, Knud. 1994. Information Structure and Sentence Form. Cambridge: CUP. DOI: 10.1017/CBO9780511620607 Langacker, Ronald. 1991. Foundations of Cognitive Grammar, Vol II: Descriptive Applications. Stanford CA: Stanford University Press. Longacre, Robert E. 1976. The discourse structure of the flood narrative. In Society of Biblical Literature: 1976 Seminar Papers, George W. MacRae (ed.), 235–262. Missoula MO: Scholars Press. Masuoka, Takashi & Takubo, Yukinori. 1992. Kiso Nihongo Bunpoo. Tokyo: Kurosio. Matsuda, Kenjirô. 1998. On the conservatism of embedded clauses. Theoretical and Applied linguistics at Kobe Shoin 1: 1–13. Matthiessen, Christian M.I.M. & Thompson, Sandra A. 1988. The structure of discourse and ‘subordination’. In Clause Combining in Grammar and Discourse [Typological Studies in Language 18], John Haiman & Sandra A. Thompson (eds), 275–329. Amsterdam: John Benjamins.
Naomi H. McGloin Maynard, Senko. 1986. The particle –o and content-oriented indirect speech in Japanese written discourse. In Direct and Indirect Speech, Florian Coulmas (ed.), 179–200. Berlin: Mouton de Gruyter. DOI: 10.1515/9783110871968.179 Maynard, Senko. 1996. Multivoicedness in speech and thought representation: The case of selfquotation in Japanese. Journal of Pragmatics 25: 207–226. DOI: 10.1016/03782166(94)00097-2 McCawley, Noriko Akatsuka. 1978. Another look at no, koto and to: Epistemology and complementizer choice in Japanese. In Problems in Japanese Syntax and Semantics, John Hinds & Irwin Howard (eds), 178–212. Tokyo: Kaitakusha. McGloin, Naomi H. 1977. The Speaker’s Attitude and the Conditionals to, tara and ba. Papers in Japanese Linguistics 5:181–191. McGloin, Naomi H. 1999. Pragmatic and discourse functions of the rhetorical negative question form, Zyanai desu ka. In Linguistics: In Search of the Human Mind, Masatake Muraki & Enoch Iwamoto (eds), 452–469. Tokyo: Kaitakusha Minami, Fujio. 1974. Gendai-Nihongo no Koozoo (Structure of Modern Japanese). Tokyo: Taishukan. Moriyama, Takuro. 2000. Koko kara hajimaru Nihongo bunpoo. Tokyo: Hitsuji-shobo. Newmeyer, Frederick. 2003. Grammar is grammar and usage is usage. Language 79(4): 682–707. DOI: 10.1353/lan.2003.0260 Okamoto, Shigeko. 1995. Pragmaticization of meaning in some sentence-final particles in Japanese. In Essays in Semantics and Pragmatics: In Honor of Charles J. Fillmore [Pragmatics & Beyond New Series 32], Masayoshi Shibatani & Sandra A. Thompson (eds), 219–246. Amsterdam: John Benjamins. Okamoto, Shigeko & Ono, Tsuyoshi. 2008. Quotative –tte in Japanese: Its multifaceted functions and degrees of ‘subordination’. In Crosslinguistic Studies of Clause Combining: The Multifunctionality of Conjunctions [Typological Studies in Language Series 80], Ritva Laury (ed.), 205–230. Amsterdam: John Benjamins. Ono, Tsuyoshi & Thompson, Sandra A. 1996. Interaction and syntax in the structure of conversational discourse: Collaboration, overlap, and syntactic dissociation. In Computational and Conversational Discourse Burning Issues—An Interdisciplinary Account, Eduard Hovy & David W. Scott (eds), 67–96. Berlin: Springer. DOI: 10.1007/978-3-662-03293-0_3 Prideaux, Gary D. 1993. Subordination and information distribution in oral and written narratives. Pragmatics & Cognition 1(1): 51–69. DOI: 10.1075/pc.1.1.05pri Schleppegrell, Mary. 1992. Subordination and linguistic complexity. Discourse Processes 15(2): 117–131. DOI: 10.1080/01638539209544804 Sunakawa, Yuriko. 1988. In’yoobun ni okeru ba no nijuusei ni tsuite (On the duality of ‘place’ in quotation). Nihongogaku 7: 14–29. Suzuki, Ryoko. 1999. Grammaticalization in Japanese: A Study of Pragmatic Particle-ization. PhD dissertation, University of California, Santa Barbara. Suzuki, Satoko. 1997. The relevance of factivity to complementizer choice in Japanese. Studies in Language 21(2): 287–311. DOI: 10.1075/sl.21.2.03suz Suzuki, Satoko. 2001. De dicto complementation in Japanese. In Complementation: Cognitive and Functional Perspectives [Converging Evidence in Language and Communication Research 1], Kaoru Horie (ed.), 33–57. Amsterdam: John Benjamins. Thompson, Sandra A. 2002. ‘Object complements’ and conversation: Towards a realistic account. Studies in Language 26(1): 125–163. DOI: 10.1075/sl.26.1.05tho
Subordination and information status Thompson, Sandra A. & Hopper, Paul. 2001. Transitivity, clause structure, and argument structure: Evidence from conversation. In Frequency and the Emergence of Linguistic Structure [Typological Studies in Language 45], Joan L. Bybee & Paul Hopper (eds), 27–60. Amsterdam: John Benjamins. Thompson, Sandra A. & Mulac, Anthony J. 1991. The discourse conditions for the use of complementizer that in conversational English. Journal of Pragmatics 15: 237–251. DOI: 10.1016/0378-2166(91)90012-M Tomlin, Russell. 1985. Foreground-background information and the syntax of subordination. Text 5(1–2): 85–122. DOI: 10.1515/text.1.1985.5.1-2.85
On state of mind and grammatical forms from functional perspectives The case for Garu and Te-iru Yuki Johnson The basic usages of the auxiliary verbs ~garu and ~(te)iru are generally understood as expressing desire and thoughts other than those of the speaker him/herself. A proposition accompanying~garu and ~(te)iru is often used with a modal, such as mitai, rashii, etc.; therefore, it belongs to the territory of the speaker domain. However, the proposition concerns another individual’s feelings that the speaker cannot perceive directly and is not an exclusive possession of the speaker. Using real-life usages that incorporate the notion of “territory of information” proposed by Kamio (1990) and the notion of “empathy,” by Kuno (1998), it is found that issues of hierarchy and empathy are deeply intertwined in their use in authentic communication and do not necessarily follow the description of standard grammar.
1. Introduction The current tendency in language research is toward a usage-based approach that emphasizes the importance of authentic utterances. Some practitioners of such an approach believe that grammar itself carries little relation to the proposition-like structures presented by formal linguists, who instead take the approach that sentences are structured according to propositions, verbs with predicates, and noun phrases with logical arguments. As described in Newmeyer (2003: 685), we rarely utter sentences comprising a subject, an object, and a verb, even though that is what grammars generate. Therefore, usage-based models argue that if real speech is not propositional, then neither should grammars be. In opposition to this premise, Newmeyer (2003) provides a variety of evidence to support the idea that mental grammar contributes to language use, but that usage does not necessarily reflect the grammar itself. While acknowledging a great disparity between sentences that grammars generate and a speaker’s actual
Yuki Johnson
utterances, Newmeyer defends the classical position with respect to the relationship between knowledge of language and use of language and challenges that ‘speakers mentally represent full grammatical structure, however fragmentally their utterances might be (2003: 682).’ As claimed by Chomsky (1965), our oral performance may not be entirely reliable in order to derive linguistic principles and parameters as it can be mistakeridden, depending on the environment the speaker is in as well as the physical condition of the speaker. Also, in oral performance in many languages of the world, word order may become random to a certain degree depending on the mood of the speaker and the content the speaker tries to convey. Japanese is no exception. A native speaker of Japanese may utter a sentence with a verb at an initial position when the speaker wants to quickly convey that s/he is going to Japan next year, saying iku no, nihon e, rainen, atashi.1 This type of utterance is often heard in oral production, though it does not coincide with the grammatical word order of Japanese as an SOV language or with the way people tend to express in writing. It seems hasty to conclude that grammars should not be propositional based exclusively on the instances of oral production, since our mental activity is not only represented in oral communication but also in writing, which seems to receive less attention in a usage-based approach regardless of the fact that both speaking and writing are productive skills. Unless a direct quote, a sentence such as iku no, nihon e, rainen, atashi is seldom produced; instead, atashi, rainen nihon e iku no may be considered an appropriate production in writing. Thus, it should be understood that usage in oral communication is ultimately derived from the same mental grammar as closely represented in written materials. This article considers the disparity seen in current linguistic research, with specific examination of spoken and written Japanese to uncover the nonsyntactic factors that affect usage, taking two grammatical patterns – garu and te-iru – as particular cases. The examination is an attempt to bring together the respective grammar and uses in a more comprehensive manner to bridge the gap between them. While the points made by Newmeyer (2003) are pertinent, a suggestion here is that grammar 1.
The gloss of this sentence is:
Iku no, nihon e, rainen atashi. Go-aux Japan loc next year I ‘I am going to Japan next year, I’m telling you.’
rammatical abbreviations used in this article are: ACC – accusative case; AUX – auxiliary G (modal auxiliaries and auxiliary verbs); CNT – contrastive marker; COMP – sentential complementizer; COP – copula; GEN – genitive case; LOC – locative case; NEG – negative morpheme; NOM – nominative case; PART – sentential particle; PASS – passive affix; PST – past tense affix; PERF – perfect affix; PROG – progressive affix (te-iru can be used to express either PERF or PROG); and TOP – topic marker.
On state of mind and grammatical forms from functional perspectives
and usage should not exhibit a complete disparity, but that grammar can be reexamined and reviewed based on usages in order to communicate sensibly and comprehensibly, since both are after all concerned with our mental activities. The data regarding the use of the auxiliary verbs, garu and te-iru were collected from various sources, including day-to-day oral communications and written documents (e-mail communication) gathered over the last two years. Also, each collected sentence was verified with native speakers of Japanese for its appropriateness to avoid single-handed intuitive judgment of the author. Whenever the use of garu and te-iru were acknowledged in oral communication or pseudo-oral communication2 in e-mail exchange (considering both formal and informal cases), uttered sentences were recorded with a note that explains the situations in which garu and te-iru were used. A total of 199 sentences were recorded and examined with regard to their use in oral communication, including 61 garu sentences and 41 tai sentences (without the use of garu) and 97 te-iru sentences. Each was compared with the formal descriptions of the use of garu and te-iru that appear in grammar reference books and language textbooks. Please refer to the data in the appendix appearing at the end of the chapter. 2. Grammar and usage of garu and te-iru 2.1
Garu
In most language texts and reference books,3 the auxiliary verb garu is explained as a constituent that is used when describing a third individual’s state of mind (specifically, desire) as opposed to the speaker’s own emotional state. Since one cannot make a direct observation or participate in someone else’s emotional state, garu is attached as a grammatical device to the root of an adjective, expressing meanings such as ‘(someone/something) has the appearance of ’ or ‘(someone/ something) appears to be.’ For this reason, some textbooks explain that garu may be replaced by the modal auxiliary (hereafter ‘modal’) soo ‘(someone/something) looks like.’ Also, garu alters the preceding predicate into a transitive verb: nomi-tagaru ‘(someone) wants to drink (something)’ and hoshi-garu ‘(someone) wants (something)’ for example, expresses someone’s future or habitual state of mind. In 2. This type of communication is herein provisionally termed ‘pseudo-oral,’ since in the case of casual communication via e-mail, language is often employed as if talking to a person directly, rather than in actual formal letter writing. 3. The books under examination are: Yookoso (Tohsaku, 1999), Genki (Nagano et al, 1999), Nakama (Makino, Hatasa & Hatasa, 1998), Kiso Nihongo 2 (Morita, 1980), Japanese: The Spoken Language (JSL) (Jorden with Noda, 1987), and The Dictionary of Japanese Grammar (Makino & Tsutsui, 1989).
Yuki Johnson
order to express a current situation, te-iru, which converts garu into a stative predicate, has to be attached. Please note the following examples. (1) a. Boku wa atarashii kuruma ga hoshii. I top new car nom want ‘I want a new car.’ b. Uti no mono ga anata ni aitaga-gatte-(i)ru-n-desu. my wife nom you dat meet-want-aux-perf-cop ‘My wife wants to meet you. (JSL: 135) (2) a. Watashi wa sabishii-desu. I top lonely-cop ‘I am lonely.’ b. Hayashi san wa sabishi-gatte-imasu. top lonely-aux-perf ‘Mr. Hayashi is lonely. (He appears to be lonely.)’ (3) a. Watashi wa tempura ga tabe-tai-desu. I top tempura nom eat-want-cop ‘I want to eat tempura.’ b. Tanaka san wa gohan o tabe-ta-gatte-imasu. top meal acc eat-want-aux-perf ‘Ms. Tanaka wants to eat meal. (Nakama: 409)’ (4) a. Boku wa rainen nihon e iki4-tai-desu. I top next year Japan loc go-want-cop ‘I want to go to Japan next year.’ ? b. Boku no tomodachi wa rainen nihon e iki-ta-gatte-iru. I gen friend top next year Japan loc go-want-aux-perf ‘My friend wants to go to Japan next year.’
c.
Boku no tomodachi wa ‘rainen ninon e iki-tai’. I gen friend top next year Japan loc go-want to itte-iru. comp say-prog ‘My friend says, ‘I want to go to Japan next year.’
Considering the subject in each of the above sentences (other than watashi and boku ‘I’), tsuma ‘wife,’ Hayashi-san ‘Mr. Hayashi,’ Tanaka-san ‘Mr. Tanaka,’ boku no 4. In fact, the morpheme boundary should be shown as ik-i-tai ‘want to go’ in which case a consonant verb is used to indicate the stem of the verb ik-. However, in order to avoid having many English glosses and due to limited space, the Japanese syllabary system is used to create a boundary. This is the reason for the desiderative form tai follows iki-tai and tabe-tai in this paper.
On state of mind and grammatical forms from functional perspectives
tomodachi ‘my friend’ are all other individuals; therefore, garu should be employed according to standard grammar.5 However, as can be detected in the data (in the appendix at the end of the article), it is noticeable that native speakers of Japanese seldom utter these sentences and that the definition ‘someone has the appearance of ’ or ‘someone appears to be’ does not fit particularly well for sentence (1b), while it suits sentence (2b) well and (3b) marginally. Sentence (4b) is somewhat awkward regardless of the reasonable-sounding English equivalent. It seems that the sentence is conveying that the speaker’s desire will be realized in the next year, rather than at the present time. This may be attributed to the use of the noun rainen ‘next year’ (that indicates a temporal distance from the current time) cooccurring with the current state of mind described by ta-gatte-iru. If the situation is described using sentence (4c), for example, the awkwardness is removed. Thus, we realize that for some cases garu may not be appropriate even for describing another individual’s desire. Let us observe some more example sentences.6 (5) Kodomo wa amai mono o tabe-ta-garu. kid top sweet thing acc eat-want-aux ‘Kids want to eat sweets. (Kids love to eat sweets.)’ (6)
Uchi no ryooshin wa otooto o daigaku ni my gen parents top brother acc university loc ire-ta-gatte-iru kedo, otooto wa amerika e enter-want-aux-perf but brother top America loc ryuugaku shi-ta-gatte-iru-n-da yo nee. study abroad-do-want-aux-perf-aux-cop part part ‘My parents want to let my younger brother go to college, but he wants to go to America to study, you know.’
(7) Cho san, Cheung san to kekkonshi-ta-gatte-(i)ru. with marry-want-aux-perf ‘Ms. Cho wants to get married to Mr. Cheung.’
(8) Suzuki buchoo, kaisha yame-ta-gatte-(i)ru. section chief company resign-want-aux-perf ‘Section Chief Suzuki wants to quit the company.’
5. Standard grammar here refers to explanations provided in elementary Japanese textbooks. They normally provide a brief explanation and do not include detailed information on exceptional cases and pragmatic uses, but the most basic uses of the grammar pattern. 6. Sentences (5) and (6) are actual utterances drawn from the original data, while (7) to (9) are modified by the author for demonstration purposes. Modal auxiliaries, such as mitai ‘looks like’ and rashii ‘seems like’ are removed from the original data for sentences (7) and (8), and a modal soo ‘looks like’ is replaced by (ta)-gatte-iru for sentence (9). The original of sentences (7)–(9) in the data are exhibited in sentences (10)–(12) that appear on the following page.
Yuki Johnson
(9) Sakai sensei wa nihon e kaeri-ta-gatte-iru. professor top Japan loc return-want-aux-perf ‘Professor Sakai wants to go back to Japan.’
Sentence (5) is acceptable as it describes a general tendency of kids’ tastes, not the state of mind of a person with whom the speaker is personally involved. Many people find sentence (6) also acceptable. The speaker is referring to his/her own family member’s desire, and the use of garu seems perfectly appropriate, as standard grammar suggests. A problem may arise with regard to sentences (7)–(9). Although they may be found in written materials, such as a report or journal where the author’s relationship with the person in question is neutral, in day-to-day conversation, they may not be uttered in this type of bare form.7 Even in casual conversation and even among young college students, speakers might avoid saying Sakai sensei, nihon e kaeri-ta-gatte-(i)ru ‘Professor Sakai wants to go back to Japan.’ Instead, as found in the data, a sentence with one of the modals is more commonly used. Observe sentences (10)–(12), drawn from the original data. (10) Cho san, Cheung san to kekkonshi-ta-gatte-(i)ru mitai yo. with marry-want-aux-perf aux part ‘It looks like Ms. Cho wants to get married to Mr. Cheung, you know?’ (11) Suzuki buchoo, kaisha yame-ta-gatte-(i)ru-rashii yo. section chief company resign-want-aux-perf-aux part ‘I hear that Section Chief Suzuki wants to quit the company, you know.’ (12) Sakai sensei, nihon e kaeri-tai-soo-da yo. Professor Japan loc return-want-aux-cop part ‘I hear that Professor Sakai wants to go back to Japan.’ Sentences (10)–(12), accompanying a modal such as mitai, soo, rashii, or ~n-desutte, are far more frequently uttered than sentences (7)–(9). In fact, sentence (12) is the case where garu does not accompany ta(i), but the use of the modal soo alone can convey the desire of the third person as a propositional content.8 These modals are used to refer to visual or sensory evidence and indicate how a speaker obtained the information (Johnson, 2004, 2008). Soo (also in the form of datte which is a colloquial equivalent of ~ to kii-ta ‘I heard’) and rashii are primarily used when the 7. ‘Bare form’ herein refers to sentences that do not accompany modal content, such as modal auxiliaries and sentence final particles. Also, in the collected data, the sentences originally all accompany some type of a modal, such as mitai ‘appears to be and’ ~desu-tte ‘I hear.’ Once the modal is removed, the sentences are all identified as awkward by native speakers of Japanese. 8. However, it is not that soo is replacing garu. Even if the sentence were to use garu, it may have been accompanied by some type of a modal.
On state of mind and grammatical forms from functional perspectives
information is obtained via an outside source, such as hearing, reading, and mitai, via the speaker’s visual impression, sensation, and/or supposition.9 The frequent occurrence of these modals demonstrates that the use of garu may be grammatically sufficient, but pragmatically insufficient to differentiate other individuals’ state of mind from the speaker’s own. However, as seen in sentence (6), when a sentence is concerned with a family member, a different concept may be employed. This observation suggests that the distinction between the in- and out-group of a speaker may also play a crucial role that is ultimately concerned with whether or not an emotional and spatio-temporal distance exists between the speaker and other individuals. Before further examining the characteristics of this issue, let us explore the behavior of another auxiliary verb – (te)-iru – which has a similar function to that of garu. 2.2
Te-iru
Most textbooks explain that the use of te-iru is required for verbs of mental activities, such as omou ‘to think,’ kangaeru ‘to think,’ and shinjiru ‘to believe,’ and when a speaker tries to signal that the thoughts belong to someone other than the speaker him/herself.10 The following are example sentences illustrating the use of omou.11 9. Several other modals exist, including hazu, ni chigainai, daroo, kamoshire-nai, n(o)-da and wake-da (Johnson, 2004). The first four function differently from yoo/mitai, soo, and rashii in that they express the degree of speaker conviction toward the truth or realization of the event described in the proposition. For example, while Nihon e kaeri-ta-gatte-iru soo-da ‘I hear that (s/ he) wants to go back to Japan’ means that the speaker tries to convey the proposition as true, hearsay information, Nihon e kaeri-ta-gatte-iru kamoshire-nai ‘It may be the case that (s/he) wants to go back to Japan’ conveys that the speaker is not sure about the truth of the proposition. These modals as well as n(o)-da and wake-da are not subject to examination in this paper. The difference between yoo and mitai is not discussed herein. However, mitai is considered an informal, colloquial equivalent of yoo (Makino & Tsutsui 1986, p.550), and the treatment of mitai here does not go beyond this observation. 10. It should be mentioned, however, that omou can accompany te-iru when the speaker’s thoughts are expressed – e.g., Boku wa jibun ga tadashii to omotte-iru ‘I’ve been thinking that I am right.’ The difference between omou and omotte-iru used for the thoughts of the speaker him/herself is that omou gives an impression that the speaker’s thoughts are grasped as a whole – equivalent to ‘I think,’ while omotte-iru is often used to suggest a time expanse between the time the speaker started to think and the time of speech – equivalent to ‘I’m thinking/I’ve been thinking that ~.’ Such observation is supported by the use of an adverb that indicates a duration of time, such as zutto ‘for a long time.’ Zutto is awkward in co-occurring with omou, while it is natural to co-occur with omotte-iru. 11. The function of omou ‘to think’ is two-fold, though sometimes difficult to identify a clear distinction. One function is to express a speaker’s opinion, indicating ‘this is what I think.’ In this
Yuki Johnson
(13) a.
Watashi wa korosa12-re-ta hito no otto ga I top kill-pass-pst person gen husband nom hannin-da to omoimasu. murderer-cop comp think ‘I think that the husband of the woman who was killed is the murderer.’
b.
Watashi wa korosa-re-ta hito no otto ga I top kill-pass-pst person gen husband nom hannin-da to omotte-imasu. murderer-cop comp think-prog ‘I’m thinking that the husband of the woman who was killed is the murderer.’
(14) *a. Otto wa hannin wa hoka ni iru to omou. my husband top murderer cnt elsewhere loc exist comp think ‘My husband thinks that the murderer is someone other than he.’ b.
Otto wa hannin wa hoka ni my husband top murderer cnt elsewhere loc iru to omotte-iru. exist comp think-perf ‘My husband thinks that the murderer is someone other than he.’
(15) *a. Sun sensei wa josei ga satsujin ni kakawatte-iru Professor top woman nom murder to relate-perf to omoimasu. comp think ‘Professor Sun thinks that a woman is involved in this murder.’ wa josei ga satsujin ni ??b. Sun sensei Professor top woman nom murder to kakawatte-iru to omotte-imasu. relate-perf comp think-perf ‘Professor Sun thinks that a woman is involved in this murder.’
case, omou is often added to soften the tone of the statement and avoid being too straightforward. The other is to express a speaker’s supposition, behaving exactly as a modal, such as kamoshirenai ‘may be’ and daroo ‘probably.’ Omou expresses a relatively high level of speaker conviction regarding the truth or realization of the proposition. Such use of omou is not a subject of examination in this paper. 12. Please refer to footnote #5 for the reason why the morpheme boundary is inserted this way. Technically, it should be koros-are-ta.
On state of mind and grammatical forms from functional perspectives
As shown in the example sentences, omou ‘to think’ is appropriate only when used to describe the speaker’s own state of mind. Omou can refer to the speaker’s state of mind at the time of speech, since the speaker can observe his/her inner feelings without a monitor or an acquisition device. Omotte-iru ‘thinking’ also refers to the speaker’s state of mind, although te-iru implies and lets one envision a time span – specifically, from the moment the speaker has decided who the criminal was up to the time of speech.13 Therefore, omotte-iru refers to the time expanse during which the speaker has been thinking. For other individuals’ state of mind, as seen in grammar explanation, only omotte-iru may be employed. This is due to the fact that the speaker cannot observe another individual’s mental activity internally at the time of speech, as others’ thoughts do not directly belong to the speaker’s mind. In order to learn and understand what is in another individual’s mind, it requires a learning process and time to acquire the thoughts, which in the end creates a spatio-temporal distance between the speaker and the other individual. Thus, te-iru itself involves duration of time to process the information, and only te-iru can be used to describe another individual’s state of mind. This observation is perfectly applicable in sentence (14b) when the speaker’s family member’s thought is in question. However, actual use of the language seems to be a bit different when concerned with individuals who have distance from the speaker in terms of social rankings. For example, sentence (15b) is a grammatical sentence with the appropriate use of te-iru. Nonetheless, most native speakers of Japanese claim that this kind of sentence sounds strange and is seldom uttered, unless as a reportive-style narrative. Sentence (15b) may sound like a public announcement. Or, in this case, if omotte-imasu ‘is thinking’ is used without accompanying any element, the speaker appears to be either very close to the individual or acting as a spokesman. In natural communication, the sentence does not end with te-iru but with some type of a modal. Please observe example sentences (16): (16) a. Sun sensei wa josei ga satsujin ni kakawatte-iru to professor top woman nom murder to related-perf comp omotte-(i)ru-n-desu-tte. think-perf-aux-cop-comp ‘I hear that Dr. Sun thinks that a woman is involved in this murder.’ b. Nara sensei wa otto ga yat-ta tte professor top husband nom murder-pst comp shinjite-(i)ru mitai. Believe-perf aux ‘It looks like Professor Nara believes that the husband is the murderer.’ 13. For further explanation of te-iru, please refer to Johnson (2004).
Yuki Johnson
c. Lee sensei wa otto wa mujitsu-da-tte professor top husband cnt innocent-cop-comp omotte-(i)ru-n-da-tte. think-perf-aux-cop-comp ‘I heard that Professor Lee thinks that the husband is innocent.’ A common feature seen in the attachment of a modal as a device is to exhibit a certain relationship between the speaker and the individual in question. By attaching a modal -tte (a semantic equivalent to hearsay soo), for example, the speaker makes it clear that s/he cannot directly perceive the other individual’s mind, but the information was somehow obtained through some unspecified method. Compared to this phenomenon, when the speaker represents his/her family member’s thoughts, the use of a modal may become somewhat loose. The speaker may convey her husband’s thoughts as if they belong to the speaker herself. This can be extended to people whom the speaker thinks very close – as family members – and the use of a modal becomes the personal preference. However, the point here is that the speaker can represent a family member’s thoughts without any device – simply omotte-iru is acceptable. 2.3
The commonality of garu and te-iru
We have seen a common characteristic in the use of garu and te-iru. That is, it is not necessarily that garu and te-iru ought to be used when desire and thoughts belong to someone other than the speaker him/herself. While grammatical, the use of garu and te-iru is not sufficient to describe the situation through the eyes of the speaker. Speakers tend to use extra devices to convey the information in actual communication, namely the use of a modal, such as yoo, soo, and rashii. Whether or not the device is needed seems to depend on the relationship between the speaker and the referent – i.e., whether the speaker considers the referent to belong to the speaker’s group, mentally, emotionally, and socially. 3. Theoretical justification Considering all aspects, the phenomenon mentioned in the previous section can be explained from the concept of ‘point-of-view’ which Kuno (1997) discusses under the framework of ‘empathy perspective.’ According to Kuno, empathy is ‘the speaker’s identification, which may vary in degree, with a person/thing that participates in the event or state that he describes in a sentence’ (1997: 206). A high degree of empathy signifies the speaker’s total identification with the person in
On state of mind and grammatical forms from functional perspectives
question, and a low degree of empathy signifies the lack of identification with the person in question. Although the concept of empathy discussed by Kuno covers a much broader range of phenomena, the concept of in- and out-group is in part integrated in the notion of ‘empathy.’ Individuals of the in-group are those to whom the speaker can emotionally relate and think that the speaker can represent their thoughts or feelings from his viewpoint and perspectives. Individuals of the out-group are those whom the speaker identifies at a distance, mentally, emotionally, and spatio-temporally. The speaker may not dare to represent their thoughts or feelings as if they belong to the speaker him/herself. Then, how can this phenomenon be justified in terms of linguistic theory? Part of a ‘Theory of Territory of Information’ proposed by Kamio (1990) helps to clarify this inquiry. Kamio’s (1990, 1994, 1997a, 1997b) theory of territory of information argues that language use depends on the territory to which information described by a sentence belongs. In this theory, whether or not a speaker/interlocutor possesses a given piece of information (i.e., whether s/he knows it or not) is clearly distinguished from whether a piece of information belongs to his/her territory. Kamio used the terms ‘direct form’ and ‘non-direct form’ for such distinction. He hypothesized that ‘the direct form is used when the speaker has adequate evidence for an assertion’ and ‘is appropriate when the information the form expresses is deep within the speaker’s territory of the information,’ whereas ‘the non-direct form is used when the speaker’s evidence is insufficient’ and ‘is appropriate when the information falls less deeply within the speaker’s territory or even outside it (1997: 146). In short, ‘direct form’ signals that the piece of information is dominated by the speaker, while the ‘non-direct form’ does not create such an implication. Please observe sentences (17) and (18) which represent the direct and non-direct form, respectively. (17) Toronto no natsu wa totemo atsui-desu. gen summer top very hot-cop ‘Summer in Toronto is very hot.’ (18) Toronto no natsu wa totemo atsui rashii-desu. gen summer top very hot aux-cop ‘I hear that summer in Toronto is very hot.’ In sentence (17), the speaker possesses the piece of information and can legitimately use the direct-form to represent the information regarding Toronto’s weather. If the speaker does not possess such information, the speaker cannot use the direct-form to convey the information, but must use a sentence ending with a form
Yuki Johnson
that signals that the information does not belong to his/her territory. This case is exemplified by sentence (18). This distinction also corresponds to propositional content and modal content. An exclusive piece of information that the speaker possesses is represented in a proposition while the fact that the speaker does not possess the information is represented by the use of a modal following a proposition. Kamio (1997a, 1997b) also states that since the direct form – a proposition – carries in part the notion that the information is dominated by the speaker, it generally creates a rather arrogant atmosphere which is not well accepted in society. Speakers try to adopt various devices to avoid creating such connotation by using modal auxiliaries, such as deshoo, mitai, yoo, rashii and sentence final particles. When examining garu and te-iru sentences in terms of the dichotomy of propositional content and modal content, we realize that both types of sentences are propositions, which means that, based on Kamio’s theory, the sentences describing another individual’s state of mind exclusively belong to the speaker. Please observe examples (19) and (20). (19) a. Otooto wa sony no furatto terebi o kai-ta-gatte-iru. my brother top sony gen flat tv acc buy-want-aux-perf ‘My brother wants to buy a Sony flat TV.’ b. Sensei wa sony no furatto terebi o kai-ta-gatte-iru. teacher top sony gen flat tv acc buy-want-aux-perf ‘Our teacher wants to buy a Sony flat TV.’ (20) a. Otooto wa terebi wa sonii ga ichiban ii to omotte-iru. my brother top tv cnt Sony nom best comp think-perf ‘My brother thinks that Sony is the best kind of TV.’ b. Sensei wa terebi wa sonii ga ichiban ii to omotte-iru. teacher top tv cnt Sony nom best comp think-perf ‘Our teacher thinks that Sony is the best kind of TV.’ Sentences in (19) and (20) are all propositions, and the information is supposed to be dominated by the speaker. However, the content is concerned with another individual’s state of mind. An apprehensive feeling regarding the use of propositions like (19b) and (20b) must be created by the conflict between the type of the sentence and the content of the proposition, and the speaker is urged to use a device to remove this conflict – namely, modal auxiliaries. On the other hand, the apprehensive feeling may not occur when the speaker refers to his/her own family members’ state of mind, which shows that interpersonal and intercommunity relationship plays a crucial role. Here, we see some commonality between the territory of information perspective and the empathy perspective demonstrating that the notion of ‘empathy’ and ‘in-group’ and
On state of mind and grammatical forms from functional perspectives
‘out-group’ are deeply related to each other and firmly rooted in the Japanese mind. Who belongs to the ‘in-’ or ‘out-’ depends on the speaker’s relationship with the individual, but social hierarchy also determines the language use. This statement in itself is nothing new. However, close examination of the use of garu and te-iru is another testimonial to the aptness of the acknowledged account. The following is a rough visualization of the examinations of garu and te-iru herein presented with some representative examples. Garu and te-iru are grammatical constituents that are used to describe a state of mind that belongs to an individual other than the speaker. For elementary to intermediate-level learners, this broad description should be sufficient. Nonetheless, actual use of these constituents is, in fact, more complicated, involving concepts that are crucial to the individuals who speak the language – namely the concepts of ‘in-group’ and ‘out-group’ intertwined with the concept of ‘empathy perspective.’ Individuals whom the speaker considers ‘in-group’ are those to whom the speaker is strongly attached in terms of empathy (an understanding and feeling of connectedness), while individuals whom the speaker considers ‘out-group’ are those with whom the speaker finds a lack of identification. The use of garu and te-iru alone is often seen to describe the state of mind belonging to individuals of the ‘in-group.’ On the other hand, for individuals of the ‘out-group,’ in order to avoid the seeming domination of their state of mind, additional devices, such as a modal, are frequently employed. Conceptually Out-group
Tabe-tai rashii
Tabe-tagatte -iru
Conceptually In-group ‘I’ Tabe-tai to omou ~Tabe-tai to omotte-iru
Tabe-tai to omotte-iru rashii
Figure 1.
Tabe-ta-gatte -iru soo-da
Tabe-tai to omotte-iru n-desu-tte
Degree of empathy High to low
Yuki Johnson
4. Concluding remarks This article examines the description of standard grammar and the actual use of the auxiliary verbs garu and te-iru, bringing together the respective grammar and usages in a more comprehensive manner to bridge the gap between them.14 Although the basic usages of garu and te-iru are to express desire and thoughts other than those of the speaker him/herself, it is found that issues of hierarchy and empathy are deeply intertwined in their use in authentic communication and do not necessarily follow the standard grammar. Even for a second or third person, they may not be used alone without accompanying a modal, such as mitai, soo, or rashii. Furthermore, garu and te-iru both fall in the domain of proposition structurally; therefore, they belong to the territory of the speaker domain, but not quite as an exclusive possession of the speaker due to the nature of the information – that is, another individuals’ feelings and thoughts which the speaker cannot perceive directly. Thus, starting from the basic description of garu and te-iru, this work extends the discussion to include real-life usages incorporating the notion of the territory of information proposed by Kamio (1990) so that their functions are more accurately described and understood. Observing the actual use of language is crucial to understanding how we think and process information. Just as crucial is to incorporate the actual use of language into existing grammars to reveal the appropriateness or insufficiency found therein. The statement made by Newmeyer (2003) that grammar is grammar and usage is usage is agreeable. However, having these two perspectives and familiarizing ourselves with the kind of discourse factors that interact with syntax, we can attain an important merit that is to protect us from making a narrow syntactic generalization. The results indicate that the formal description of these grammatical patterns does not precisely reflect their use, but provides merely a basic description of their functions. The results thus fill a gap between the formal description of grammar and the actual use of these grammatical patterns derived from actual use of the language.
14. The findings of the research done for these grammatical items are well incorporated in the Fundamentals of Japanese Grammar (Johnson, forthcoming).
On state of mind and grammatical forms from functional perspectives
Appendix Garu + Modal (a total of 61 sentences) Modal ϕ -ta-gatte-iru-n-desu-tte/dat-te -ta-gatte-iru soo -ta-gatte-iru mitai -ta-gatte-iru rashii -ta-gatte-iru tte itte-ta-yo -ta-gatte-iru tte kii-ta -ta-gatte-iru yoo Total
In-group
Out-group
Total
3 (60%) 3 (12%) 1 (7.6%) 4 (44.4%) 0 0 0 0 11 (18%)
2 (40%) 21 (87.5%) 12 (92.3%) 5 (55.5%) 3 (100%) 3 (100%) 3 (100%) 2 (100%) 50 (82%)
5 24 13 9 3 3 3 2 61
In-Group means that the speaker refers to a desire of a person who is considered to belong to the speaker’s group conceptually. Out-Group means the opposite of in-group. E.g.: Nihonjin tte gaikoku ni kite-mo, suguni nihonshoku tabe-ta-garu yo ne. ‘When Japanese come to a foreign country, they immediately want to eat Japanese food, don’t they?’
Tai + Modal (a total of 41 sentences) Cases where the use of garu is avoided and a modal may be added Modal
In-Group
Out-Group
Total
ϕ -tai-n-desu-tte/dat-te -tai soo -tai mitai -tai rashii Total
7 (100%) 0 2 (28.5%) 4 (30.7%) 0 13 (31.7%)
0 10 (100%) 5 (71.5%) 9 (69.3%) 4 (100%) 28 (68.3%)
7 10 7 13 4 41
E.g.: Ano hito wa ne, kirei na hito to kekkonshi-tai no. ‘My brother wants to get married to a pretty woman.’
Te-iru (a total of 97 sentences) Modal
In-group
Out-group
Total
ϕ -te-iru mitai -te-iru-n-desu-tte/da-tte -te-iru soo -te-iru rashii -te-iru-tte itte-ta-yo -te-iru yoo Total
4 (66.6%) 13 (35.1%) 1 (5.5%) 0 0 0 1 (25%) 19 (19.5%)
2 (33.4%) 24 (64.9%) 21 (95.5%) 16 (100%) 8 (100%) 4 (100%) 3 (75%) 78 (80.5%)
6 37 22 16 8 4 4 97
Yuki Johnson
Samples Samples 1, 2, and 3 are partial conversations/utterances recorded in casual daily conversation. [Sample 1] The following sample exhibits a case where both yame-ta-gatte-iru ‘wants to quit’ and yame-tai ‘wants to quit’ accompany a modal. The communication is exchanged between a 29 year-old male and a 32 year-old female who work at a Japanese motor company in Tokyo. A: Suzuki buchoo, kaisha, yame-ta-gatte-(i)ru-rashii yo. section chief company resign-want-aux-perf-aux part ‘I hear that Section Chief Suzuki wants to quit the company, you know.’ B: Un, atashi mo uwasa kii-ta. Yame-tai-mitai ne. Yes I too rumor hear-pst quit-want-aux part ‘Yes, I heard the rumor too. It looks like he wants to quit. I feel bad for him...’ [Sample 2] The following sample is one of the cases where a 35 year-old female in Tokyo talks about her own mother’s thoughts without using any modals. A:
Uchi no haha sa, Ken no koto seijitsu-da to my gen mother part gen about sincere-cop comp omotte-(i)ru kara, shinpai ira-nai yo. think-perf so worry need-neg part ‘Since my mother thinks that Ken is sincere, there is no need to worry.’
[Sample 3] The following sample is a case where a Japanese male college student in Toronto makes a declarative statement regarding a thought of his professor. It was uttered in a very casual situation where three other students were talking about the grade they received. A: Ano sensei, ore no koto zettai ni baka-da to omotte-(i)ru. that teacher I gen about definitely stupid-cop comp think-perf ‘That teacher definitely thinks that I am stupid.’
On state of mind and grammatical forms from functional perspectives
References Chomsky, Noam. 1965. Theories of Syntax. The Hague: Mouton. Johnson, Yuki. 2004. Modality and the Japanese Language. Ann Arbor MI: University of Michigan Center for Japanese Studies Publication Program. Johnson, Yuki. 2008. Fundamentals of Japanese Grammar. Hawaii HI: University of Hawaii Press. Jorden, Elinor Harz & Noda, Mari. 1987. Japanese: The Spoken Language, Part I. New Haven CT: Yale University Press. Kamio, Akio. 1990. Joohoo no Nawabari Riron. Tokyo: Taishukan. Kamio, Akio. 1994. The theory of territory of information: The case of Japanese. Journal of Pragmatics 21: 67–100. DOI: 10.1016/0378-2166(94)90047-7 Kamio, Akio. 1997a. Evidentiality and some discourse characteristics in Japanese. In Directions in Functional Linguistics [Studies in Language Companion Series 36], Akio Kamio (ed.), 145–171. Amsterdam: Johns Benjamins. DOI: 10.1075/slcs.36 Kamio, Akio. 1997b. Theory of Territory of Information [Pragmatics & Beyond New Series 48]. Amsterdam: Johns Benjamins. DOI: 10.1075/pbns.48 Kuno, Susumu. 1997. Functional Syntax. Chicago IL: Chicago University Press. Makino, Seichi, Hatasa, Yukiko Abe & Hatasa, Kazumi. 1998. Nakama I. New York NY: Houghton Mifflin. Makino, Seichi & Tsutsui, Michio. 1989. A Dictionary of Basic Japanese Grammar. Tokyo: Japan Times. Morita, Yoshiyuki. 1980. Kiso Nihongo 2. Tokyo: Taishukan. Nagai, Y., et al. 1999. Genki I. Tokyo: Japan Times. Newmeyer, Frederick. 2003. Grammar is grammar and usage is usage. Language 79: 682–707. DOI: 10.1353/lan.2003.0260 Tohsaku, Yasu-Hiko. 1999. Yookoso, 2nd edn. New York NY: McGraw-Hill College.
Grammar of the internal expressive sentences in Japanese Observations and explorations Shoichi Iwasaki ‘Internal expressive sentences’ such as a! itai ‘Oh, it hurts,’ or aa yokatta ‘Oh, I’m so happy!,’ reveal directly what the speaker perceives or senses internally, and contrast sharply with ‘descriptive sentences,’ such as taroo wa kuma o uchitometa ‘Taro shot the bear.’ This distinction has been noted in both Japanese and European languages but has not received much attention in linguistics in general. This paper explores this neglected type of sentence in terms of their morphosyntactic patterns (co-occurrence with outcry vocalization, such as a!, aa, waa, restricted predicate forms, the clipped adjective form, and inability to code an experiencer), semantic primitives labeled as Experience, Stimulus and Receptor, and extralinguistic factors based on neurological states of perception, emotion and feeling.
A group of Japanese sentences can be identified formally by their co-occurrence possibility with a short outcry vocalization such as a!, aa, or waa, as shown in (1). (1) a! itai! ‘Ouch!’ aa yokatta! ‘Oh, I’m so glad!’ waa ureshii! ‘Oh, I’m so happy!’ a! sakana ga shinderu! ‘Oh! A dead fish!’ (cf. * a! kono sakana ga shinderu! ‘Oh! This fish is dead!) (cf. * a! aru sakana wa sora o toberu! ‘Oh! Some fish can fly!’) These sentences refer to and reveal directly what the speaker perceives or senses internally, and contrast sharply with descriptive sentences, such as taroo wa kuma o uchitometa (‘Taro shot the bear’). In this paper, the type of sentences presented in (1) are referred to as ‘internal expressive sentences’ (or simply ‘expressive sentences’). Some types of internal expressive sentences have been identified as ‘Ausdruck’
Shoichi Iwasaki
(expressions) by Bühler 1933,1 ‘hyooshutu’ (revelations) by Sakuma (1967), and ‘kantan-bun’ (exclamatory sentences) by Onoe (2001: 168–178). These categories constitute a subset of what some other authors call ‘affective’ or ‘emotive’ expressions (e.g. Besnier 1990; Maynard 2000, 2002).2 Traditionally the internal expressive sentence constitutes part of the minor sentence type along with the ‘impricative (curse)’ and ‘optative’ (Sadock & Zwicky 1985: 162–5), and is contrasted with the major sentence type of the ‘declarative,’ ‘interrogative’ and ‘imperative’ sentences (Lyons 1977: 745). Linguists mainly concern themselves with the major sentence types, especially the proposition-making function of the declarative sentences, and have paid little attention to the other type of sentences including expressive sentences (see relevant comments, among others, in Mio, 2003: 66 [1948], Lyons 1982; Maynard 2000, 2002, 2007, and Linell 2005: 87–8). This paper, thus, tires to give due attention to this neglected sentence type, and attempts to provide a framework to analyze their specific grammatical patterns. It also proposes that a distinct set of grammatical patterns found for expressive sentences constitutes a separate system within a multi-layered grammatical organization of a human language. It should be noted, however, that much of the discussion presented here is of explorative nature based on my own observations. Nonetheless, I believe this is a necessary step towards future developments of the neglected area in linguistics. The following text is presented in six sections. Section One gives a brief summary of treatments of expressive sentences in the Japanese scholarship. Section Two first contrasts the expressive sentence with the more familiar type of descriptive sentence, then discusses distinct features of the former, and finally proposes three semantic primitives necessary for an analysis of expressive sentences. Section Three explores the possibility of analyzing expressive sentences in terms of nonlinguistic factors, namely various neurological processes, and delineates grammatical patterns associated with them. Section Four briefly discusses the descriptive sentence in relation to the expressive sentence, in order to understand the continuous nature between expressive and descriptive sentences. Section Five summarizes the paper. Section Six discusses a theoretical implication of the current study and different methodologies which may be used to analyze the expressive sentence.
1. See Bednarek (2006) and Innis (1982) for Bühler’s theory of linguistics known as the ‘Organon Model.’ 2. Maynard’s ‘emotives’ (2000, 2002) include a wide range of linguistic phenomena, such as linguistic devices that describe emotions, linguistic strategies that enact emotional attitude, and grammatical and rhetorical means which foreground emotive meanings. Other linguists who have recognized expressive functions in language include Bally (1965 [1925]), Jakobson (1960), Dong (1971), Lyons (1977: 50–56), Ochs & Schieffelin (1989), and Daneš (1994).
Grammar of the internal expressive sentences in Japanese
1. Japanese scholarship on the internal expressive sentence Traditional Japanese linguists have shown some interest in expressive and other related sentences. However, their analyses are not always without difficulties. One common difficulty is the strict formal means to identify these sentences. Yamada (1936: 935–6), for example, categorizes a sentence as the kandoo-kantai ‘impression-expressive style’ (one type of our ‘expressive sentence’) when it takes the form of vocative expression:
[Modifier Noun [hana]N [utsukushii]Mod beautiful flower ‘(What) a beautiful flower!’.3
(Vocative Particle)] ([yo])Prt Vocative Particle
This strict structural definition, however, assigns a different interpretation to two functionally similar sentences; the sentence; waa! kiree na hana! (wow – beautiful –flower) and waa! kono hana kiree! (wow – this flower – beautiful), though both have a similar emotive force, must be classified as completely different sentence types based upon constituent order, first being a sentence of kandoo-kantai ‘impression-expressive style’ (our expressive sentence) and the second being a sentence of juttai ‘subject-predicate sentence’ (our descriptive sentence) (see Morishige 1971: 25–7). Matsushita (1978 [1930]), in his philosophical discussion of sentence types, identifies the sentences of non-conceptual and conceptual subjective judgment as well as the sentence of deliberate judgment. Upon feeling a tremor, a speaker may issue the non-linguistic outcry, a! (a sign of non-conceptual subjective judgment) or describe it with a one-word expression, jishin! ‘Earthquake!’ (a sentence of conceptual subjective judgment). However, when jishin is followed by the copula, jishin da (It’s an earthquake!), the sentence is understood to be a sentence of description, implying deliberate judgment, rather than simply as expressive. This is because, according to Matsushita, it can be related to a more fully specified and logically constructed sentence, kore wa jishin da (this is an earthquake). However, jishin da may in fact be uttered as a response to one’s own perception, and in this case, it is a full, non-abbreviated sentence. This type of subjectless sentence is a hallmark of expressive sentences. 3. In addition to kandoo-kantai, Yamada identifies kiboo-kantai (the optative style), which takes the form of [NP] + [Optative Particle] with the meaning of ‘I wish for (Noun),’ e.g. [oizu shinazu no kusuri] + [moga] ‘How I wish for medicine of longevity and immortality!’ (Yamada 1954 [1936]: 949). The optative style can be found in Classical Japanese, but went out of use alongside the optative particles, such as moga(na). The optative style and optative particles are absent from Modern Japanese.
Shoichi Iwasaki
Mio (2003: 66 [1948]) improves on the form-based identification of sentence types by bringing in contextual information in the classification of sentence types. In his analysis, the same surface form may be analyzed differently, depending on its relation to the situation of its utterance. For example, in Mio’s view, jishin da (earthquake COP) (Earthquake!) could be an ‘underdeveloped sentence’ (mitenkaibun) in the case where a speaker utters it before analyzing the situation, as it depicts the speaker’s surprise at that moment. Alternatively, it may be a ‘partial sentence’ (bunsetu-bun) when it refers to a part of the complete idea for the situation, as in the utterance kore wa jishin da (this is an earthquake). This analysis can avoid the problem noted above with respect to Matsushita’s analysis. As Mio himself admits, however, the category of underdeveloped sentences is problematic as it includes both what Sakuma calls the ‘revelation type’ (hyooshutsu) which exposes the speaker’s inner emotional or cognitive state through such expressions as a! (oh!), ame da! (it’s raining!), and the ‘appeal type’ (uttae) which tries to get other’s attention. The latter category is exemplified by utterances such as hora! (look!) and kimi! (you!) (Sakuma 1952: 46–7)4. The analysis of expressive sentences presented in this paper expands upon Mio’s analysis, by identifying true expressive sentences on the basis of the nature of what is expressed (i.e. speaker’s internal, more specifically their neurological state) and how it is intended in the process of communication (i.e. not primarily intended for communication). 2. The nature of the internal expressive sentence This section outlines the characteristics of expressive sentences as opposed to description sentence and propose three unique semantic primitives that are relevant for analysis of expressive sentences. 2.1
Expressive and descriptive sentences
The ‘internal expressive’ sentence can be sharply contrasted with the ‘descriptive’ sentence. The distinction corresponds to that between mode pur (the intellectual mode) and mode vécu (the affective mode) of Bally (1965[1925]). As Maynard (2002: 26) explains, ‘in expressing the speaker’s emotional attitude, mode pur offers its description, producing a report of one’s inner sensations, as in I am getting mad. Mode vécu, on the other hand, enacts a live performance of the sensation, as in 4. These terms are Sakuma’s translation of Bühler’s Ausdruck and Appel (1934), see also Innis (1982).
Grammar of the internal expressive sentences in Japanese
Damn it! The descriptive sentence can be divided into ‘external descriptive’ and ‘internal descriptive’ sentences. The ‘external descriptive’ sentence is a familiar type of sentence which depicts a situation (an event, state, process, relation, or quality) that is actually or potentially observable, or objectively verifiable, in the external world (‘Mary stood up her date’, ‘The boy is growing up’, ‘John and I are friends’, ‘The window is broken’. ‘John is tall’). Consider the Japanese examples below: (2)5 External descriptive taroo ga kinoo te o kegashita ‘Taro hurt his hand yesterday..’ (name) nom yesterday hand acc hurt:past hanako wa me ga kiree da (name) top eye nom beautiful cop
‘Hanako has beautiful eyes.’
kono suupu wa karai this soup top spicy
‘This soup is spicy.
On the other hand, the ‘internal descriptive’ sentence describes situations that are internal to a person and are not directly observable or, in principle, not verifiable from outside. This includes mental processes, physical and emotional conditions, knowledge, and belief, among others (‘We expect John to win the race’, ‘Kim is angry at Jim’s sister’, ‘It strikes me that John is insincere’, ‘I had a bad headache yesterday’). See below for Japanese examples of this type of sentence: (3) Internal descriptive taroo ga hannin da to omou ‘(I) think Taro is the culprit.’ (name) nom culprit cop qt think:nonpast koko ga itai-n da yo here nom painful-se cop pp
‘It hurts here, you see.’
hara hetteta-n desu yo stomach decrease:past:asp-se cop pp
‘(I) was hungry, you know.’
suki yo like PP
‘(I) love (you), you know.’
In contrast, internal expressive sentences do not describe, but performs. English expressive sentences include; Oh, it hurts!, Boy, is he dumb!, and What a beautiful flower!. Notice these sentences exhibit an unconventional use of present tense form (hurts), word order, and verb-less structure. Japanese expressive sentences also show distinctive characteristics as discussed shortly, but first examine some examples. 5. The abbreviations used in this paper are: ACC = accusative marker, ASP = aspect, COP = copula, EXC = exclamation, NOM = nominative marker, NONPAST = non-past, PAST = past, PP = pragmatic particle, QT = quotative, SE = sentence extender, TOP = topic marker.
Shoichi Iwasaki
(4) Internal expressive (see also (1) above) a! karai! ‘Oh! it’s hot!’ exc spicy a! kono suupu karai! exc this soup spicy
‘Oh, this soup is so hot!’
a! sakana ga shinde-ru. exc fish nom die-asp:nonpast
‘Oh, a fish is dead!’
aa atama ga itai. exc head painful
‘Oh, I have a headache!’
waa. kireena hana. exc beautiful flower
‘Wow, what a beautiful flower!’
As I mentioned at the outset, the expressive sentence is defined by its co-occurring possibility with a short outcry vocalization such as a!, aa, or waa. In addition to this, there are other formal characteristics. First, internal expressive sentences cannot co-occur with the past tense. They are often appear in the present tense form (e.g., a karai! ‘Oh, hot!’ – expressing the current experience). When they appear with the past tense form of an adjective, it has to be interpreted as Perfect (e.g., aa karakatta ‘Oh, (that) was hot!’ – expressing the experience of the immediate lingering experience), not as Past (*a!/aa kinoo karakatta! ‘Oh, (that) was spicy yesterday’ is not possible.) This tense/aspect restriction is a consequence of expressive sentences being temporally deictic. Second, some of the internal expressive sentences show peculiar morphological contrast between speaker’s own and others’ experience; the form itai ‘painful’ can be used only for the speaker’s sensation (expressive sentences), but has to be modalized for someone else’s pain (descriptive sentence), e.g. ita-soo-da ‘appears to be painful’ and itagatte iru ‘acting as if painful.’ In other words, internal expressive sentences are completely egocentric expressions that reveal internal experiences of the speaker. Third, although internal expressive sentences always involve an experiencer, or a human who undergoes an internal situation, they never code it expressively (*a! boku ga itai! ‘Oh, I am in pain’); it can only be indexed indirectly through the act of exclamation. This is a crucial difference between the internal expressive sentence, e.g., a! itai! (Ouch! – lit. Painful!), and the internal descriptive sentence, boku mo atama ga itai (I also have an headache). Before leaving this section, it is pertinent to mention one other important characteristic of the expressive sentence; its lack of communication intent. The speaker of true internal expressive sentences such as itai! (‘Ouch!’) does not have an intention to communicate. This means that these sentences are sub-vocal, or ‘inner-speech’ (Vigotsky 1962) in nature. Once uttered, however, people around
Grammar of the internal expressive sentences in Japanese
may respond with, for example, daijoobu? (Are you OK?). Also a speaker may imitate an internal expressive sentence strategically to perform indirect speechacts; they may say urusai! (lit. It’s loud!) to make someone turn off the loud music. In fact, it is not always easy to separate a true internal expressive sentence and its strategic use. This is due to the more general problem of indirect speechacts, or a general lack of one-to-one correspondence between the form and the speech-act it performs (e.g., Lyons 1977: 745; Irvine 1982; Levinson 1983: 263–264; Palmer 1986: 23–25).6 Bednarek, in her critical review of Bühler’s Organon Model, notes, ‘it is not clear to what extent means of Ausdruck are employed intentionally or subconsciously and to what extent they are conventionalized’ (2006: 146). In this paper the perfomative use of internal expressive sentences will not be discussed further. The communicative function in language is, of course, crucial for understanding many aspects of language, but it is possible that it may be separated from other functions. Internal expressive sentences in their purest form represent a clear example of non-communication language. As will be discussed later, descriptive sentences can be also used with or without communicative intent. 2.2
Three semantic primitives of an expressive sentence
Descriptive sentences have been traditionally analyzed into a predicate and noun phrases with semantic roles such as ‘Experiencer’ ‘Agent,’ ‘Patient,’ ‘Theme,’ and ‘Instrument’. Internal expressive sentences cannot be meaningfully analyzed by these terms as many of them consist of one word. For an expressive sentence that consists of two terms, a noun and a verb/adjective, the traditional analysis can be performed, but will not capture the special nature of expressive sentences; the descriptive sentence neko ga sakana o tabeteiru ‘the cat is eating the fish’ and the expressive sentence a! neko ga sakana o tabeteru! ‘Oh, a cat is eating a fish!’ cannot be distinguished according to the traditional analysis. I propose in this paper, therefore, three entirely different semantic primitives; ‘Experience,’ ‘Stimulus,’ and ‘Receptor’. These semantic primitives have been extracted by observing what type of information appears when an expression contains only one word. In addition, they have been deduced by our general knowledge of neurological experiences. Sato (2004: 87), for example, describes how humans interpret external stimulus as follows:
6. Goffman (1983: 78–123) maintains that ‘response cries’ including ‘pain cries’ have communicative intent.
Shoichi Iwasaki
“Information contained in the stimulus from external world enters the nervous system through sensory organs such as eyes, ears, skin, tongue, and nose. The external information which goes through the sensory organs will be analyzed into temporal-spacio patterns in terms of impulse that the nervous system can use. Then the nervous system analyzes the pattern of impulse into numerous steps and processes them hierarchically in order to recreate the external image.” (translated and emphasis added by SI)
‘Experience’ is the recreated image, or the quality of the internal situation that the experiencer is currently undergoing, e.g. perception, emotion, or feeling. ‘Stimulus’ is something that causes a particular experience. ‘Receptor’ is a location through or at which an experience is registered. These three primitives appear differently in the linguistic expression depending on the exact type of internal situation that is being expressed. 2.3
Interim summary
The purpose of the preliminary sections above was to separate expressive sentences from descriptive sentences, and identify their features for further analyses. Expressive sentences in Japanese are formally characterized by possibility of cooccurrence with outcry vocalization, use of non-modal form, present tense form, Perfect aspect form, and inability to code an experiencer. Functionally, they are sentences of the ‘affective mode’ (mode vécu) and differ significantly from sentences of the ‘intellectual mode’ (mode pur). They reveal speaker’s internal experiences, such as perception, emotion, and feeling, at the time of the utterance. They are not intended for communication, at least in principle. I suggested three semantic primitives, Experience, Stimulus and Receptor, to analyze expressive sentences. My suggestion at this point is rather speculative, but it will be given some support in the subsequent discussion. Also, though we have discussed the notion of the internal situation that expressive sentences code, this notion is admittedly vague. Thus, in the next section we reassess the expressive sentence in terms of the notion of neurological conditions of the speaker that they refer to and the formal characteristics of utterances that express them. 3. Grammar of internal expressive sentences and the neurological processes In the previous section, some characteristics of internal expressive sentences were identified and their three semantic primitives proposed. This section delineates their grammatical system, or how the form is related to what it denotes. For this purpose, I propose to examine expressive sentences from a rather unconventional
Grammar of the internal expressive sentences in Japanese
perspective, i.e. distinct neurological states that humans experience such as reflex, perception, emotion and feeling. This proposal may sound ad hoc, but as will be seen, they will provide an important insight how expressive sentences are organized internally and with each other. It should be pointed out also that the discussion that follows crucially impinges upon the principal of iconicity (Haiman 1985), which assumes that there are direct correspondences between details of forms and the complexity of neurological process that they refer to; the simpler the form, the simpler the neurological process. Expressive sentences may take either one-term or two-term forms. As noted earlier in relation to Mio’s theory of expressive sentences, not all one-term expressions are internal expressive sentences. If a one-term expression such as jishin da (earthquake COP)7 is an elliptical sentence related to kore wa jishin da (This is an earthquake), it is a descriptive sentence resulting from the thought process of the speaker. However, if uttered as a reaction to a tremor a jishin da! (oh earthquake COP) ‘Oh, earthquake!’, it is an expressive sentence. The difference can be substantiated by an observation that the former can be put into the past tense sentence are wa jishin datta (‘That was an earthquake’), but the latter cannot (*a jishin datta!). While one-term expressions code one primitive in a single expression (karai! ‘Hot!’), two-term expressions code two primitives. In the expression a kimuchi karai! (‘Oh! (this) kimchi is spicy!), both Stimulus kimuchi (kimchi) and Experience karai (spicy) are expressed. In two-term expressions, the nominative case particle ga may be present. The use and non-use of the nominative particle will be taken up in Section 3.3.1 and 3.3.2 below. Before turning to the discussions of oneand two-term expressions, however, we must examine carefully what I have been calling the outcry vocalizations. I will show in the next section that these vocalizations can be termed ‘reflex expressions.’ 3.1
Reflex expression
As a point of departure for the discussion of reflex expressions, it is useful to refer to Nathan’s description of reflex as a neurological phenomenon. According to Nathan (2004), ‘(a) reflex occurs before an individual knows what struck him, what made him lift a foot or drop an object. It is biologically correct to be alarmed before one knows the reason.’ Reflex expressions are verbal emulation of an actual physical reflex, a verbal reaction without referring to the cause. Independently produced outcry expressions such as a!, aa, and waa, are a prima facie example
7. Both jishin! (‘Earthquake!’) and jishin da! (‘It’s an earthquake!’) are considered to be one-term expressions. One-term is, thus, defined either as one word or one word followed by the copula.
Shoichi Iwasaki
of verbal reflex (cf. Haiman’s symptomatic gesture 8). These are pre-linguistic expressions indexing only the existence of the experiencer/speaker, and fail to refer to any of the three primitives mentioned in Section 2.2. The reflex expressions often precede an expressive sentence, as shown in many examples to follow, such as a! itai! (‘Oh, it hurts!’) and ‘waa kiree!’ (‘Wow! Beautiful!’). There are short and long varieties of the reflex gestures a! and (w)aa 9, respectively. The former is pronounced with a glottal stop (shown here as an exclamation point) and indexes an immediate and acute experience, while the latter variety is pronounced with a long vowel without a glottal stop and indexes a slower and more prolonged process. To understand the difference, consider the following description of two neurologically different kinds of pain. When you hit your thumb with (a hammer), a sharp ‘first’ pain is felt immediately, (followed) later by a more prolonged aching, sometimes burning ‘second’ pain. (Basbaum & Jessell, 2000: 474)
The first pain may be expressed by the short reflex gesture and the second pain by the long variety. It is thus possible to utter a! (first pain) followed by aa (second pain) as in a! aa, but it is impossible to reverse the order, *aa. a!. 3.2
One-term expressions
One-term expressions code ‘perception,’ ‘feeling’ or ‘emotion.’ The grammar associated with one-term expressions can be described in terms of three features: (i) the possibility of clipped form, (ii) the possibility of co-occurring with a short outcry expression a!, and (iii) the possibility of co-occurring with a long outcry expression (w)aa.
8. Haiman offers the following observation: “A symptomatic gesture (let us say, a cry of pain like [aaaa]) accompanies a psychological state. That is, originally the gesture connotes the state. It becomes a signal which still connotes that state once it is recognized and responded to by some other animal. Finally, it becomes a sign (say, the English word ‘ouch’) which denotes the state only once it is emancipated both from the stimulus which produced it originally, and from the motivated state of which it served as a signal (1994: 15–16)” (Italics in original). Thus, only when a linguistic expression has been freed from the immediate stimulus can it serve as a descriptive sentence. 9. The two versions of long reflex gestures, aa and waa, are not functionally identical. The latter variety, waa, seems to appear when the speaker responds to an external Stimulus, and the former variety, aa, seems to appear when he/she responds to an internal(ized) Stimulus. Thus, as a response to the pain described in the quotation, waa is not appropriate. This matter could benefit from further investigation.
Grammar of the internal expressive sentences in Japanese
3.2.1 Perception expression In opposition to ‘reflex,’ the neurological process of ‘perception’ analyzes the input received at the receptor and transmitted to the central nervous system.10 ‘Perception expressions’ are a linguistic representation of such neurological analysis, and index either simple Experience, Stimulus, or marginally, Receptor. As Nathan notes regarding the neurological process of perception, ‘it is only after the immediate and automatic response that the cerebral cortex is involved and conscious perception begins’ (2004). In contrast to ‘reflex expressions,’ ‘perception expressions’ are genuine linguistic representations of internal states, referring to one of the three semantic primitives. Perception expressions that code Experience are mostly adjectives. Among several related forms, the most compact form is a clipped adjective sometimes fused with a short reflex expression. The clipped adjective is formed by deleting the final -i from an -i ending adjective, such as in ita-i > ita (painful).11 An expression with a clipped adjective may be fused with a reflex expression by deleting the reflex expression’s final glottal stop, e.g. a! ita! > aita!. They may be further reduced by deleting /i/ after /a/ to ata!. The final two vowels of an adjective, if they are /a/ and /i/, may also be fused as /e/ (ita-i > ite) particularly in male speech. The two most representative examples of perception expressions are ‘aita!’ (painful) and atsu!12 (hot). These clipped forms may be (partially) reduplicated; aitatata! (for a general pain), and achichichichi (for a heat pain). Table 1 below, shows this group of perception expressions. Though the numbers of adjectives that take the clipped form with a fused verbal reflex expressions is limited, the number of clipped form adjective may be on the rise. Those in Table 2 below may be used by some as perception expressions.13 10. In addition to the two types of motor control, reflex and reaction via perception, there is a third type, which is controlled by the brain stem that regulates posture and eye and head movements (Ghez 1991: 537–8). 11. The term ‘clipping’ usually refers to a morphological process which cuts off part of a word, e.g. ‘ad’ clipped from ‘advertisement’ in English. I am using this term in this paper to refer to the stem of the i-type adjective. My choice of this word is based on the observation that in a perception expression, the -i ending form (the full form – see below) is basic and the ‘clipped’ form is based on this form. This is evidenced as all adjectival perception expressions can be formed by the base form, but only selected ones can be ‘clipped.’ 12. Since the initial vowel in atsu- (hot) is /a/, the identity of /a/ in atsu! is ambiguous between the initial vowel of the original adjective and the reflex gesture a!. 13. See Sugiura (2005) for more information. Also, those forms in Tables 1 and 2 may have a long vowel instead of a glottal stop (e.g. amaa (sweet)). The distinction between ama! and amaa corresponds to the first and second types of pain just discussed.
Shoichi Iwasaki
Table 1. Perception Expressions of Experience [Clipped Adjective – Typical Cases] Experience
Clipped Form
English gloss14
Sensation
(a)ita!/(a)ite! /ata! atsu!/achi! (a)kayu!/(a)kai!
‘painful’ ‘hot (temperature)’ ‘itchy’
Table 2. Perception Expressions of Experience: [Clipped Adjective – Marginal Cases] Experience
Clipped Form
English gloss
Sensation
a!/(w)aa tsumeta!/ a!/(w)aa tsumete! a!/(w)aa samu!
‘cold’ (object temp.)
Gustatory
a!/(w)aa uma!/ume! a!/(w)aa mazu! a!/(w)aa kara!/ kare! a!/(w)aa ama! a!/(w)aa niga! a!/(w)aa suppa!/ suppe!
‘tasty (colloquial)’ ‘bad tasting’ ‘spicy’’ ‘sweet’ ‘bitter’ ‘sour’
Olfactory
a!/(w)aa kusa!/ a!/(w)aa kuse!
‘stinks/smelly’
Visual
a!/(w)aa kura!
‘dark’
Others
a!/(w)aa sugo! a!/(w)aa kowa!
‘great’ ‘scary’
‘cold’ (body temp.)
A much larger number of perception expressions appear with the full-form adjective. Some adjectives that appear in Tables 1 and 2 also appear in Table 3. All the expressions presented in Tables 2/3 can be preceded by a short or long reflex expression.15 Thus a! itai! refers to the immediate pain and aa itai to a prolonged after-the-effect pain. Perception expressions shown in Table 3 above are 14. I do not provide an approximate or exact translation of the expression in this and other tables in this paper; for aita! we only give the gloss ‘painful,’ not a functional equivalent ‘Ouch!’ or a more literal translation ‘Oh, it’s painful!’ as it is sometimes difficult to give precise equivalents in English. 15. Onoe (2001: 173) notes that adjectives of attribute, such as marui ((be) round), can also be used in an expressive sentence in a specific context, as in a! marui! (Oh, it’s round!). This is possible, for example, when a speaker expects the shape of an object to be square, but has just noticed it is round. This latter case belongs to the group of visual Experience in Table (3).
Grammar of the internal expressive sentences in Japanese
Table 3. Perception Expressions of Simple Experience: [Full Adjective] Experience
Full Form
English Gloss
Sensation
a!/aa itai! a!/aa atsui! a!/aa tsumetai! a!/aa samui! a!/aa kayui! a!/aa kaii!
‘painful’ ‘hot (temp.) ‘cold (object temp.) ‘cold (body temp.)’ ‘itchy’
Gustatory
a!/aa karai! a!/aa oishii! a!/aa umai! a!/aa mazui! a!/aa amai! a!/aa nigai! a!/aa suppai!
‘spicy’ ‘tasty’ ‘tasty (colloquial)’ ‘bad tasting’ ‘sweet’ ‘bitter’ ‘sour’
Olfactory
a!/aa kusai! a!/aa ii nioi!
‘stinks/smelly’ ‘fragrant’
Tactile
a!/aa zarazara! a!/aa katai!
‘rough surface’ ‘hard’
Auditory
a!/aa urusai!
‘noisy’
Visual
a!/aa mabushii! a!/aa akarui a!/aa kurai
‘dazzling’ ‘bright’ ‘dark’
Others
a!/aa kawaii! a!/aa kowai! a!/aa sugoi! a!/aa abunai! a!/aa atta!
‘cute’ ‘scary’ ‘great’ ‘dangerous’ ‘found it’ (=‘It’s here!’)
mostly adjectives, but also include other types of words such as an onomatopoeia (‘zarazara’ for roughness), a noun modified by an adjective ‘ii nioi’ (nice smell), or a verb for a change of state ‘atta!’ ((it’s) here < lit. existed). They all refer to a simple Experience. One-term perception expressions may also index a simple Stimulus in the form of [Noun (+Copula)], as shown in Table 4 below. Upon seeing a bug, for example, one may utter, a! mushi (da) or aa mushi (da)! (oh, a bug!).
Shoichi Iwasaki
Table 4. Perception Expressions of Simple Stimulus: [Noun (+Copula)] Stimulus
Noun (+Copula)
English gloss
a!/aa mushi (da)! a!/aa ame (da)! a!/aa jishin (da)!
‘bug’ ‘rain’ ‘earthquake’
Marginally, a perception expression may index the Receptor, such as ‘a! ashi!’ (oh! foot!) when something has happened to one’s foot.16 3.2.2 Emotion and feeling expressions ‘Emotion’ and ‘feeling’ are something that humans experience internally beside perception. Unlike ‘perception, ’ however, they do not allow the clipped form, e.g. a! *kuyashi! < kuyashii ‘regrettable.’ ‘Emotion’ and ‘feeling’ are similar to each other, but are distinguished psychologically and neurologically. Emotions are often associated with mental states that are impulsive and automatic. A working definition of emotion-proper provided by Damasio (2003: 53) includes: ‘(1) An emotion-proper, such as happiness, sadness, embarrassment, or sympathy, is a complex collection of chemical and neural responses forming a distinctive pattern. (2) The responses are produced by the normal brain when it detects an emotionally competent stimulus (an ECS), the object or event whose presence, actual or in mental recall, triggers the emotion. The responses are automatic.’17 In contrast, feelings are a complex resultant mental state given rise to by emotions and other factors, and almost always take time to register. A provisional definition of feeling given by Damasio (2003: 86) is ‘the perception of a certain state of the body along with the perception of a certain mode of thinking and of thoughts with certain themes’ The feeling is originated not only from the emotion-proper but also from any homeostatic reactions, e.g. appetite (Damasio 2003: 85). In short, while emotion is an automatic response, feeling is a perception originated by emotion and other reactions. The distinction is similar to that between first and second pains discussed earlier. Thus, I propose that emotion/feeling expressions can be distinguished by the type of outcry vocalization they co-occur. That is, if an expression is preceded by a short outcry vocalization, it is an emotion expression. If, on the other hand, it is preceded by a long counterpart, it is a feeling expression. It turns out that only a 16. I reported in Iwasaki (2006) that indicating the Receptor in a one-term expression is more natural in Korean than in Japanese. 17. This type of emotional state is often translated as joodoo (emotional movements) in Japanese.
Grammar of the internal expressive sentences in Japanese
Table 5. Experience of Emotion and Feeling18 Experience
Form
English gloss
Emotion (with a!)/ Feeling (with (w)aa)
a! /(w)aa kuyashii! a! /(w)aa yokatta! a! /(w)aa zannen! a!/(w)aa natsukashii a!/(w)aa ureshii a!/(w)aa tanoshii a!/(w)aa osoroshii a! /aa chikushoo!
‘regrettable’ ‘(was) good’ ‘regrettable’ ‘reminiscing’ ‘happy’ ‘fun’ ‘fearful’ ‘(Shit!)’
Feeling proper
aa kanashii aa tsurai
‘sad’ ‘hurt/painful’ (psychologically)
few are clearly feeling expressions, and many can refer to both emotion and feeling. Feeling expressions include aa kanashii ‘sad’ and aa tsurai ‘hurt/painful (psychologically). These cannot be preceded by a!.19 On the other hand, expressions like kuyashii ‘regrettable’ can be preceded by either variety. Like a! itai (first pain) and aa itai (second pain), a! kuyashii (emotion) and aa kuyashii (feeling) refer to the immediate reaction to a stimulus that is considered regrettable and to the processed reaction to it, respectively. Emotions are expressed by different types of words, but feelings are expressed only by adjectives. Both emotion and feeling expressions code Experience.20 3.2.3 An interim summary Table 6 below summarizes the discussion of reflex expressions and other one-term expressions that code perception, emotion and feeling, as defined by clipping possibility of adjectives, the co-occurrence possibilities with a!, and with (w)aa!.
18. Notice that most adjectives used for feelings and emotions have the ‘-shii’ ending. This is not accidental. Adjectives in Classical Japanese are classified according to their conjugational patterns into the -ku and -shiku conjugation adjectives, and in general they correspond to objective (e.g. taka-ku ‘high’) and subjective (e.g. kana-shiku ‘sad’) adjectives, respectively. The erstwhile -shiku ending adjectives have become the–shii ending adjectives in Modern Japanese. 19. Onoe notes that adjectives of feeling can be used as an emotion expression if the speaker registers it strongly at a particular moment; a! kanashii! (Oh, I’m so sad!) (2001: 173). 20. There may be some sentences of feeling that refer to Stimulus, such as a! kono yaroo! (Oh, Bastard! lit. oh, this guy!). This expression, however, may have become conventionalized, and refers to Experience. Similarly a! ckikusho! in Table (5) originally referred to ‘beasts’ (Stimulus), but has become completely conventionalized and code Experience.
Shoichi Iwasaki
Table 6. A Grid of Quality of Expression
Clipped Adjective Preceded by a! Preceded by (w)aa
Reflex
Perception
Emotion
Feeling
NA NA NA
± + +
– + +
– – +
This table reveals that the four types of neurological experiences have specific formal patterns. Reflex expressions are pre-linguistic expressions taking the form of either a! or (w)aa. Emotion and Feeling expressions do not take clipped adjectives, and both can co-occur with the long reflex expression. However, only emotion expression can take the short variety. Perception expressions may take the clipped form of adjective and can occur with either type of reflex expression. Thus, among the three linguistic expressions, the perception expression is least restricted, and can code a variety of semantic primitives. As shown below perception expressions can be further developed more freely in two-term expression compared to the emotion/feeling expressions. 3.3
Two-term expressions
Two-term expressions can be classified formally into two types. In one, the two terms are connected by the particle ga. The expression of this type indicates either a complex Stimulus or Stimulus-with-Experience. This type of expansion is available only for perception expressions. In the other, the two terms are juxtaposed with each other without the particle. The expression of this type successively presents a simple Stimulus/Receptor and Experience. 3.3.1 Complex perception: Expressive sentence with ga Two-term expressions may be expressions of complex perceptions, consisting of a noun and a verb/adjective connected by the particle ga of ‘neutral description’ (Kuno 1973). From the view of the information structure, these sentences are those with sentence focus, i.e. the whole of the information is new to the speaker. Semantically, these sentences do not express a proposition composed of a ‘subject’ and the ‘predicate.’ Those in (5) below identify the type of sensation (Experience) as e.g. itai ‘painful’ and the location of sensation (Receptor) as, e.g. onaka ‘stomach.’ Formally, the two primitives are connected with the particle ga.
Grammar of the internal expressive sentences in Japanese
(5)
Sensation Perception (Complex perception = Receptor + Experience) aa. onaka ga itai. exc stomach nom painful ‘Oh, stomach hurts!’
aa. me ga kayui. exc eye nom itchy ‘Oh, eyes are itchy!’ Other types of perception can be also expanded into a two-term expression with the particle ga. Note that the pain perceptions in (5) above are preceded by the reflex expression aa, indicating they represent the second phase of perception. In this case, prefacing with a! will lead to unacceptable expressions. Selection of a! or aa, however, fluctuates depending on the exact perception involved as shown in (6) below. In these examples, Stimulus (e.g. kyuuri ‘cucumber’) and Experience (e.g. oishii ‘delicious’) are coded in one expression. (6) (Complex perception = Stimulus + Experience) a. Gustatory Perception aa. kyuuri ga oishii! exc cucumber nom delicious ‘Oh, a cucumber is delicious!’ (Delicious cucumber!) b.
Olfactory Perception aa. happa ga ii nioi! exc leaf nom good smell ‘Oh, the leaves smell nice!’
c.
Auditory Perception aa. hae ga urusai! exc fly nom nosy ‘Oh, a fly is noisy!’
d.
Tactile Perception a!/aa. happa ga zarazara! exc leaf nom rough.surface ‘Oh, the leaves feel rough!’
e.
Visual Perception a!/aa. denki ga mabushii! exc light nom bright ‘Oh, the light is bright!’
Shoichi Iwasaki
Comparing the examples in (5) and (6) above, those in (5) are more readily accepted than those in (6)21. There are at least two potential reasons for this slight discrepancy. First, the pain expressions in (5) are far more common to appear in actual speech than other sensation expressions, and may have become formulaic. That is, those in (5) are familiar enough to be accepted without imagining an appropriate context, while those in (6) require proper contexts to be imagined. Second, in the pain expression, both Receptor and Experience exist within a speaker, making it easier for the two to be bridged, but in other perception expressions, Experience is internal but Stimulus external, making such cognitive operation more difficult. When the two primitives are both external, two-term expressions are also easy to construct. In Section 3.2.1 it was noted that a one-term perception expression such as a! mushi! is uttered upon registering a simple Stimulus. When the Stimulus to be referred to is more complex, grammar codes it as a two-term expression. Thus, upon seeing a bug (Stimulus) and simultaneously perceiving its condition (another Stimulus), a speaker may utter a! mushi ga shinderu (A dead bug! < lit. Oh! A bug is dead). Similarly, upon hearing a bug chirping, a speaker may utter a! mushi ga naiteru (A bug chirping! < lit. Oh! A bug is chirping). Some examples of visual perception and auditory perception are provided below in (7) and (8), respectively. Reflex markers can be either short or long.
(7) Visual perception (Complex perception = Complex Stimulus = Stimulus + Stimulus) a. a!/(w)aa mushi ga shinderu. exe bug nom dead ‘Oh, a bug is dead!’ b. a!/(w)aa okane ga ochiteru. exe money nom fallen ‘Oh, fallen money (on the ground)!’ c. a!/(w)aa inu ga hashitteru. exe dog nom run:asp:npst ‘Oh, a dog is running/passing by!’ d. a!/(w)aa hebi ga detekita. exe snake nom come.out:pst ‘Oh, a snake is coming out!’
21. According to a small sampling of opinions from eleven native Japanese speakers, both of the pain perception expressions in (5) are unanimously judged to be acceptable with the long outcry marker, aa, while judgments vary among other perception expressions in (6).
Grammar of the internal expressive sentences in Japanese
e. a!/(w)aa hoshi ga kiree. exe star nom beautiful ‘Wow, the stars are beautiful!’ f. a!/(w)aa kao ga akai. exe face nom red ‘Oh, (his/your etc.) face is red!’ (8) Auditory perception (Complex perception = Complex Stimulus) a. a!/(w)aa mushi ga naiteru. exe bug nom cry:asp:nps ‘Oh, (I hear that) a bug is chirping!’ b.
a!/(w)aa kaminari ga ochita. exe thunder nom fall:pasa ‘Oh, (I hear) thunder!’ (upon hearing a loud noise caused by thunder)
c.
a!/(w)aa mizu ga nagareteru. exe water nom flow:asp:nps ‘Oh, (I hear that) water is flowing!’ (upon hearing rather than seeing the flowing water)
According to Kuroda, the above examples are sentences of ‘thetic judgment,’ which represent ‘a simple recognition of the existence of an actual situation’ (Kuroda 1992: 22–23).22 In other words, when the speaker utters these sentences, he perceives the situation as a complex stimulus with no focus on either the entity or its temporary situation; no part of these sentences has been activated in the mind of the speaker, both are simultaneously activated on the spot. As said earlier, these sentences do not express a proposition, which must have a ‘subject’ and the ‘predicate’ independently. 3.3.2 Perception of ‘successive simple stimuli’: Expressive sentence without ga We noted earlier that in a sentence of thetic judgment, such as those involving [NP ga V/Adj.], neither the NP nor the V/Adj. receives saliency, or has been activated in the mind of the speaker independently. Based on this, it can be predicted that if the NP is assumed to be activated in the mind of a speaker (Chafe 1976, 1994) through modification with a demonstrative, the sentence will become ungrammatical. Compare the two sentences taken from (7) and presented again in (9) and their counterparts with demonstrative modifiers in (10) below. 22. Kuroda’s ‘thetic judgment’ sentences correspond to Kuno’s sentence of ‘neutral description’ (1973). We can treat all the sentences in (5) through (8) in either of these terms.
Shoichi Iwasaki
(9)
Perception expressions a!/(w)aa mushi ga shinderu. exe bug nom dead ‘Oh, a bug is dead!’
a!/(w)aa hoshi ga kiree! exc star nom beautiful ‘Wow, the stars are beautiful!’ (10)
Perception expressions with a demonstrative23 *a!/(w)aa kono mushi ga shinderu! exc this bug nom dead ‘This bug is dead!’
*a!/(w)aa ano hoshi ga kiree! exc that star nom beautiful ‘That star is beautiful!’ Those sentences in (10) become acceptable if a!/(w)aa is removed and ga interpreted as the particle of exhaustive listing with the meaning, such as kono mushi ga shinderu (It’s this bug that’s dead – not those others). This sentence, however, is no longer an expressive sentence, but a descriptive sentence.24 This will be discussed in Section 4 below. More significantly for the purpose of the present discussion, the sentences in (10) become perfectly acceptable as expressive sentences if the particle ga is removed, as shown in (11). (11) a!/(w)aa kono mushi __ shinderu! exc this bug _ dead ‘Oh, this bug is dead!’ a!/(w)aa. ano hoshi __ kiree! exc that star _ beautiful ‘Wow, that star is beautiful!’ This suggests that the sentences in (11) are not integrated thetic judgment sentences, but those that consist of two terms juxtaposed with one another, one referring to the identity of simple Stimulus (kono mushi (this bug) or ano hoshi (that 23. It is crucial that the modifier is a demonstrative. Since other types of modifiers like adjectives and genitival phrases do not specify the modified nouns as activated concepts, they will not create ungrammatical sentences with ga such as the following; a! akai mushi ga shinderu (Oh! a red bug is dead!), a! boku no mushi ga shinderu (Oh! my bug is dead!), and, a! Kimi-chan no mushi ga shinderu (Oh! Kimmy’s bug is dead!). 24. Note that the two functions of ga which Kuno identifies (1973), the neutral description and the exhaustive listing, are now defined by a distinction between two sentence types.
Grammar of the internal expressive sentences in Japanese
star)) and another referring to its state (shinderu (is dead) or kiree (is beautiful)). In other words, the sentences in (11) depict two simple stimuli in succession. It should be pointed out that emotion and feeling expressions cannot be turned into a thetic judgment sentence, but it can be turned into the type represented by (11); a! kono ongaku _ natsukashii! ‘Oh, this music conjures old memories!’) Finally, it should be noted that ga in all examples in (5) and (6) can be omitted. In some cases, acceptability will increase by removing the particle. According to the proposal above, there is a certain difference between the versions with and without ga. Compare (12–a) and (12–b). The former represents an instantaneous expression of a cucumber being delicious (‘Delicious cucumber!’), while the latter separately codes ‘the cucumber’ (external Stimulus) and ‘registering as delicious’ (internal Experience). In other words, (12–b) is synonymous with (12–c).25 (12) a. a!/aa kyuuri ga oishii! (cf. (6)) exc cucumber nom delicious ‘Oh, a cucumber is delicious!’ (Delicious cucumber!) b. a!/aa kyuuri __ oishii! exc cucumber delicious ‘Oh, (this) cucumber is delicious!’ c. a!/aa kono kyuuri __ oishii! exc this cucumber delicious ‘Oh, this cucumber is delicious!’ The observation that a ga marked noun phrase has not been activated in the minds of the speaker further explains the naturalness of (13–a, b, d) and the unnaturalness of (13–c) below (cf. Masunaga, 1988: 149–50). (13) a. b.
a! basu _ kita exc bus _ came ‘Oh, a bus is coming!’ a! basu ga kita exc bus _ came ‘Oh, a bus is coming!’
c. *a! kyuukyuusha _ kita exc ambulance _ came ‘Oh, an ambulance is coming!’ 25. The distinction found between (12–a) and (12–b) is not relevant when the noun is a body part (Receptor); a/aa onaka (ga) itai! ‘Oh, I have a stomachache!’ – see (5) above. This is most likely related to the fact that a body part can never be modified by a demonstrative such as kono ‘this.’ This matter requires a further investigation.
Shoichi Iwasaki
d. a! kyuukyuusha ga kita exc ambulance nom came ‘Oh, an ambulance is coming!’ Upon seeing an approaching bus, a person, if he/she has been waiting at a bus stop, can say (13–a) to him/herself without ga. If, however, a person accidentally saw an approaching bus, he/she would utter (13–b) in his/her mind. The reason why (13–c) sounds odd is because it is rare for people to wait for ambulances than buses (i.e. it is harder to imagine a proper context). If a person accidentally saw an approaching ambulances, he/she would utter (13–d). The point is that if the noun phrase represents a non-activated concept, ga marking is obligatory. In contrast, if it represents an activated concept, ga marking becomes problematic.26 3.3.3 Interim summary This section described two-term expressions. Such expressions describe visual, auditory, and other perceptions. The two terms in a two-term expression are connected with the particle ga or juxtaposed without it. The form, [NP-ga X], indicates the speaker’s perception as a whole ‘thetic judgment’, but the [NP _ X] indicates that the NP is independently identified or activated before a situation involving it is presented. 4. External descriptive sentences – revisited As we noted in the previous section, sentences in (11) and (12–b), such as a! kono mushi _ shinderu, are not thetic sentences of internal experience. However, they are not external descriptive sentences either. A sentence becomes fully developed as an external-descriptive sentence when the sentence is no longer temporally deictic (Section 2.1). This can be done by removing the outcry expression and inserting a particle, either ga or wa. (14)
External descriptive sentences kono mushi ga/wa shindeiru this bug nom/top die:asp:npst ‘This bug is dead.’
26. According to Matsunaga, it is possible to make a sentence like (13–c) acceptable by adding a sentence final particle such as yo or zo: a! kyuukyuusha _ kita yo as these particles move the focus from the NP to the verb (1988: 147). However, even without yo, it is possible to utter (13–c) in a context in which someone has called an ambulance, and the ambulance has just arrived. The function of yo here seems to create an understanding between the speaker and hearer that there is a shared understanding regarding the ambulance.
Grammar of the internal expressive sentences in Japanese
ano hoshi ga/wa kiree da that star nom/top beautiful cop ‘That star is beautiful.’ These sentences are familiar sentences to students of Japanese linguistics. The sentences with wa are sentences of ‘double judgment’ (Kuroda 1990, 1992), in which the speaker first singles out an entity and deliberately attributes a certain property to it, such as ‘Look at this bug, it is dead’. The sentences with ga, on the other hand, are sentences of ‘exhaustive listing’ (Kuno 1973), in which the speaker identifies an entity to be predicated, as in the sentence ‘It is this bug that is dead’. From the neurological point of view, both types represent the process of ‘cognition,’ which works to analyze a situation by sorting out a complex stimulus. There is another type of descriptive sentence, the generic sentence, as shown in (15). (15)
External descriptive sentences mushi wa chiisana ikimono da bug top small living.thing cop ‘The insect is a small living thing.’
hoshi wa yozora o terasu star top night.sky acc shine:npst ‘Stars shine across the night sky.’ When a speaker uses generic sentences, he/she leaves the deictic world of here and now, and enters the abstract, conceptual world. Before we leave this section, we should make a brief note on the communicative intent in language. One important feature of the internal-state expressions we have been discussing is their non-communicative nature (see 2.1). Descriptive sentences can also be used without any communicative intent. This happens when speakers use these sentences to organize thoughts for themselves. In other words, while communication is an important aspect of language, it is not a defining characteristic of language. 5. Summary Internal state expressions are linguistic representation of a speaker’s neurological experience of perception, emotion, and feeling that he/she is undergoing internally at the time of utterance, and are, in principle, self-directed speech and not intended for communication. Reflex expressions are pre-linguistic gestures denoted by a!, aa and waa in Japanese. Internal state expressions reveal one or two of the three primitives of Stimulus, Experience and Receptor. One-term perception
Shoichi Iwasaki
expressions in various forms with clipped or non-clipped adjectives and a limited number of other words refer to a simple Experience, Stimulus, or marginally, Receptor. One-term emotion and feeling expressions do not take clipped forms, but the former can co-occur with either a! or (w)aa, while the latter only with the long expression, (w)aa. Among the three types of internal states, only perception can be further elaborated into a two-term expression freely. Two-term expressions code complex perceptions. When the two terms are connected by the particle ga, the expression conveys the speaker’s thetic judgment applied to the whole utterance. When they are not connected with the particle, the expression is a successive production of two pieces of information, i.e. two different Stimuli a! mushi _ shinderu (‘Oh, the bug is dead!’), Receptor and Experience aa. onaka _ itai (‘Oh, my stomach hurts!’), and Stimulus and Experience a! kyuuri _ oishii (‘Oh, this cucumber is delicious!’). When the information is processed by a higher neurological function of cognition, the speaker can produce a sentence consisting of a noun followed by ga (‘exhaustive listing’) or wa (‘topic’) and a predicative element, as seen in the sentence kono mushi ga/wa shindeiru (This bug is dead). We hope to have shown that internal expressive sentences are best viewed as linguistic representations of neurological processes that humans undergo. This relationship will be eventually understood as part of the iconic principle pervasively found in language (Haiman 1985). The neurological event is a fundamental experience for humans, and thus it is expected to make a mark on all natural human languages. It is hoped that research such as this one will be conducted on diverse languages. 6. Discussion and conclusion In this concluding section, I will discuss a theoretical implication of the current research and methodological issues for studying expressive sentences. The expressive sentences in Japanese examined in this paper are largely ignored in linguistics. An obvious reason for such neglect is that these sentences do not convey a proposition, and thus are considered a sub-sentence in a field that is concerned with proposition making as a major endeavor of linguistic expressions. Even those who acknowledge the expressive function of language pay little attention to the grammatical system that underlies it. This paper thus tried to fill this gap by showing grammatical system which maps the form and the internal state that it reveals. Grammatical systems discovered include possibility of clipping of adjectives, co-occurrence patterns with different types of reflex expressions, expansion of one-term expression into two-term expression, the use and non-use of nominative particle ga and the use of demonstratives among others. These systems
Grammar of the internal expressive sentences in Japanese
are rather odd if seen from the confinement of traditional grammatical paradigm. However, native speakers know these patterns; e.g. they will never say *a! kono mushi ga shinderu (Oh! this bug is dead!) upon seeing a dead bug, but may say a! kono mushi __ shinderu (Oh! this bug is dead!) without the particle ga. Likewise, native speakers can distinguish the grammatical use of clipped forms for certain adjectives aita! (painful) and the ungrammatical use of such forms for other adjectives *a! ureshi! (happy). This means that the grammatical knowledge of native speakers includes recognition of these systems. The question remains however, of how to understand this system in relation to the more familiar systems that have been traditionally discussed in linguistics. For example, how should we understand the function of subject marking by the particle ga and its function in a thetic judgment sentence? Here I suggest that language consists of multiple dimensions defined by different functional domains, and each dimension may be organized grammatically differently with overlaps to various degrees. Besides the dimension of internal expressions, other dimensions of language that should be recognized include, conversational interaction, oral narratives, poetry, academic writing, sports casting and many more. These dimensions have been conventionally discussed in terms of speech genres, but we must now critically examine each genre’s internal grammar. Speakers acquire a different set of multiple grammars throughout their lives at different times, with different rates of learning and through different means. A mature speaker is thus a person who has integrated multiple grammars, and can use them separately or concurrently within a single discourse. The current study could be seen as an initial step towards understanding such grammatical knowledge and ability to use them in real discourse. Research into internal expressive sentences has been neglected both in formal and functional linguistics because of its methodological difficulty. Expressive sentences are not always available in the kind of discourse data linguists collect. Even in emotion-laden oral stories of personal tragedy, what one most often finds is a linguistic ‘description’ of reflex, perception, emotion, and feeling, or at best, one can find ‘acting’ of pain, hurt and joy, for example. These expressions are not, however, truly temporally deictic expressions. Furthermore, internal states such as pain, hurt, joy, and the like are more often than not communicated through extralinguistic features, such as ‘grunts’ and ‘noises’ (Campbell 2005). We can, however, consider a few methodologies that may be utilized. One possibility is to identify expressive sentences in a large spoken corpus. For example, Campbell used a corpus containing “almost five years of daily conversational speech” (2005: 116) to study ‘expressive speech’ (Campbell 2004, 2005). Though the chances of finding actual examples of expressive sentences will increase as the size of the corpus becomes larger, it may not be possible to able to get to all the
Shoichi Iwasaki
details of the phenomena, as expressive speech is by nature private, and may not be captured by the currently available methodologies. On the other hand, comic books contain many exclamations, curses, and other expressive language, and thus are promising sites in which expressive sentences can be studied. Goffman (1981: 114) advocates the use of cartoon and comic books to study ‘response cries’ as creators of such an art form must capture the ‘presumed inner state’ of cartoon figures.27 The following are a few relevant examples of expressive language that appear in Pocket Monster Special vol. 1 (2005, Hidenori Kusaka, Tokyo:Shogaku-kan):28 (16)
Pre-linguistic reflex expressions: a!? uwa! ku! a~a~ uwaa!? hi! aa!? uwaaaa!! waa!! achaa eee!? wa! uo!? oo! waaa!
Sensation perception expressions ite!/ tete!/atatatata/ (< itai ‘painful’) uatchi! (< atsui ‘hot’) Puzzlement he? e!? heee.
‘What?’
Curse kuso!/kusoo. kono yaroo!!
‘Shit!/Damn it!’ ‘Bastard!’
Discovery a! dokukurage a! kabigon a! booru miiikke
‘Oh! Dokkukurage! (a name of Pocket Monster) ‘Oh! Kabigon! (a name of Pocket Monster) ‘Oh! I found the ball!’
Combination u... nanda? kono kiri... kaaaa! kawaikunee yatsu
‘Hm? What? This fog...’ ‘Mmmmm. Not a cute one!’
27. See also Maynard (2002: 117–121) for a detailed discussion on the use of comics, fictions and television dramas as linguistic data. A caution is, however, necessary as the language in comics is a creation of the author, and may not correspond to the real use of language (see Kinsui 2003). 28. Diverse orthographic conventions are used for various effects in the original. Here we can convey only some of these.
Grammar of the internal expressive sentences in Japanese
Even this small sample gives both confirmation for some points made in the present paper and suggestions for further analysis. In this data are found pain expressions in the clipped form (ite!) and the fused form (uatchi! < u! atchi!) as well as a clipped form that lacks the initial vowel (tete < itete < itai itai; atatatata < a! itai itai itai). A slightly more complex sentence would include extra material; in a, booru miiikke, the verb miiikke is a clipped form from mitsuketa ‘found.’29 An expanded database of comics will most likely afford more varied forms used for internal expression. Another reasonable solution is an old technique of using introspection. There is always the danger of fabricating data to one’s own advantage in the introspection based research, but as has been noted, certain aspects of language such as ‘(m)eanings, mental imagery, emotions, and consciousness’ (Chafe 1994: 12) (and we could add other internal states – SI) can only be privately observable. For these aspects of language, introspection provides an important initial point of exploration. Of course, linguists who wish to study expressive sentences or similar phenomena such as inner speech (Vygotsky 1962: 44–8) need to struggle to pair up their theory with publicly observable linguistic behavior (Chafe 1994: 15), and accept criticisms on their judgment, and continue to revise their understanding of the phenomena. However, since it is unlikely that one methodology will reveal the entire range of phenomenon of internal expression, making use of a variety of methodologies is crucial. Only through such an open-minded approach will we advance our knowledge of human language. References Bally, Charles. 1965[1925]. Le langage et la vie. Genève: Droz. Basbaum, Allan I. & Jessell, Thomas M. 2000. The perception of pain. In Principles of Neural Science (4th edition), Eric R. Kandel, James Schwartz & Thomas M. Jessell (eds), 472–491. New York NY: McGraw-Hill. Bednarek, Monika. 2006. Bühler, Karl (1879–1963). In Encyclopedia of Language and Linguistics, 2nd edn, Keith Brown (ed.), 145–147. Oxford: Elsevier. Besnier, Nico. 1990. Language and affect. Annual Review of Anthropology 19: 419–451. DOI: 10.1146/annurev.an.19.100190.002223 29. Logically, booru (ball) is the direct object of mikke ((I) found) and should be able to take the accusative case particle o, but this is not an option here; *a! booru o mikke!. Though this is like a sentence that cannot take the nominative case particle (e.g. a! kono mushi _ shinderu (oh, this bug is dead!), the situation seems more complicated. In fact, there is no circumstance in which the accusative particle is allowed with this verb when the speaker makes an announcement about his/her discovery of an object whether the ball has been activated in the mind of the speaker. I will leave this matter for future research.
Shoichi Iwasaki Bühler, Karl. 1933. Ausdruckstheorie. Das system an der Geschichte aufgezeigt. Jena: Fischer. Campbell, Nick. 2004. Databases of expressive speech. Journal of Chinese Language and Computing 14(4): 295–304. Campbell, Nick. 2005. Getting to the heart of the matter: Speech as the expression of affect, rather than just text or language. Language Resources and Evaluation 39(1): 109–118. DOI: 10.1007/s10579-005-2699-y Chafe, Wallace L. 1976. Givenness, contrastiveness, definiteness, subjects, topics and point of view. In Subject and Topic, Charles N. Li (ed.). New York NY: Academic Press. Chafe, Wallace L. 1994. Discourse, Consciousness, and Time – The Flow and Displacement of Consciousness Experience in Speaking and Writing. Chicago IL: The University of Chicago Press. Daneš, František. 1994. Involvement with language and in language. Journal of Pragmatics 22: 251–64. DOI: 10.1016/0378-2166(94)90111-2 Damasio, Antonio. 2003. Looking for Spinoza: Joy, Sorrow, and the Feeling Brain. New York NY: Harcourt. Dong, Quang Phuc. 1971. English sentences without overt grammatical subject. In Studies out in Left Field: Defamatory Essays Presented to James D. McCawley on his 33rd or 34th Birthday, Arnold M. Zwicky A., Peter H. Salus, Robert I. Binick and Anthony L. Vanek (eds), 3–10. Edmonton: Linguistic Research Inc. Reprinted in 1992 by John Benjamins. Ghez, Claude. 1991. The control of movement (Chapter 35). In Eric R. Kandel, J. Schwartz and T. Jessell. Principles of Neural Science. (3rd edition), 533–547. New York: Elsevier. Goffman, Erving. 1983[1981]. Forms of Talk. Philadelphia PA: University of Pennsylvania Press. Haiman, John (ed.). 1985. Iconicity in Syntax [Typological Studies in Language 6]. Amsterdam: John Benjamins. DOI: 10.1075/tsl.6 Haiman, John (ed.). 1994. Ritualization and the development of language. In Perspectives of Grammaticalization [Current Issues in Linguistic Theory 109], William Pagliuca (ed.), 3–28. Amsterdam: John Benjamins. Innis, Robert E. 1982. Karl Bühler - Semiotic Foundations of Language Theory. New York NY: Plenum Press. DOI: 10.1007/978-1-4757-0923-0 Irvine, Judith. 1982. Language and affect: Some cross-cultural issues. In Georgetown University Roundtable on Languages and Linguistics, Heidi Byrnes (eds), 31–47. Washington DC: Georgetown University Press. Iwasaki, Shoichi. 2006. The structure of internal state expressions in Japanese and Korean. In Japanese/Korean Linguistics, Timothy J. Vance & Kimberly A. Jones (eds), 331–342. Stanford CA: CSLI. Jakobson, Roman. 1960. Linguistics and poetics. In Style in Language, Thomas A. Sebeok (ed.), 350–377. Cambridge: CUP. Kay, Paul & Fillmore, Charles. 1999. Grammatical constructions and linguistic generalization: The what’s X doing Y? construction. Language 75: 1–33 DOI: 10.2307/417472 Kinsui, Satoshi. 2003. Vaacharu nihongo: Yakuwarigo no nazo (Virtual Japanese: A mystery of role specified language). Tokyo: Iwanami Shoten. Kuno, Susumu. 1973. The structure of the Japanese language. Cambridge MA: The MIT Press. Kuroda, Sige-Yuki. 1990. The categorical and the thetic judgment reconsidered. In Mind, Meaning and Metaphysics. The Philosophy and Theory of Language of Anton Marty, Kevin Mulligan (ed.). Dordrecht: Kluwer. DOI: 10.1007/978-94-009-0505-4_6 Kuroda, Sige-Yuki. 1992. Japanese Syntax and Semantics: Collected Papers. Dordrecht: Kluwer. DOI: 10.1007/978-94-011-2789-9 Levinson, Stephen C. 1983. Pragmatics. Cambridge: CUP.
Grammar of the internal expressive sentences in Japanese Linell, Per. 2005. The Written Language Bias in Linguistics: Its Nature, Origins and Transformations. London/New York: Routledge. Lyons, John. 1977. Semantics, Vol. 1. Cambridge: CUP. Lyons, John. 1982. Deixis and subjectivity: Loquor ergo sum? In Speech, Place and Action: Studies in Deixis and Related Topics, Robert J. Jarvella & Wolfgang Klein (eds), 101–24. New York NY: Wiley. Masunaga, Kiyoko. 1988. Case deletion and discourse context. In Papers from the Second International Workshop on Japanese Syntax, William J. Poser (ed.), 145–156. Stanford CA: CSLI. Matsushita, Daisaburo. 1978[1930]. Kaisen hyoojun nihon bunpoo (teisei saihan) (Revised Standard Japanese Grammar (corrected and re-published)). Tokyo: Bensei-sha. Maynard, Senko K. 2000. Jooi no gengo-gaku: ‘Ba-kooshoo-ron’ to nihongo hyoogen no patosu. Tokyo: Kuroshio shuppan. Maynard, Senko K. 2002. Linguistic Emotivity. Centrality of Place, the Topic-comment Dynamic, and an Ideology of Pathos in Japanese Discourse [Pragmatics & Beyond New Series 97]. Amsterdam: John Benjamins. DOI: 10.1075/pbns.97 Maynard, Senko K. 2007. Linguistic Creativity in Japanese discourse: Exploring the Multiplicity of Self, Perspectives, and Voice [Pragmatics & Beyond New Series 159]. Amsterdam: John Benjamins. DOI: 10.1075/pbns.159 Mio, Isago. 2003[1948]. Kokugohoo bunshoo ron (A theory of Japanese sentences). Tokyo: Hitsuji-shoboo. Morishige, Satoshi. 1971. Nihonbunpoo no shomondai (Issues in Japanese grammar). Tokyo: Kasama shoin. Nathan, Peter W. 2004. Human nervous system. In Encyclopædia Britannica. (11 August, 2004). Ochs, Elinor & Schieffelin, Bambi. 1989. Language has a heart. TEXT 9(1): 7–25. DOI: 10.1515/ text.1.1989.9.1.7 Onoe, Keisuke. 2001. Bunpoo to imi I (Grammar and meaning I). Tokyo: Kuroshio-shuppan. Palmer, Frank R. 1986. Mood and Modality. Cambridge: CUP. Sadock, Jerrold M. & Zwicky, Arnold M. Speech-act distinctions in syntax. In Language Typology and Syntactic Description: Clause Structure, Timothy Shopen (ed.). Cambridge: CUP. Sakuma, Kanae. 1952. Gendai nihongohoo no kenkyuu. Tokyo: Koosei-sha Koosei-kaku. Sakuma, Kanae. 1967. Nihon-teki hyoogen no gengo kagaku (Language science of Japanese expression). Tokyo: Koosei-sha Koosei-kaku. Sato, Masahiko. 2004. Noo-shinkei to koo doo (The brain-nervous system and behavior). Tokyo: Iwanami Shoten. Sugiura, Katsumi. 2005. ‘Samu!’ ‘uma!’ nado ni mirareru bunpooka ni kansuru koosatsu: Koobun bunpoo no shiten. The Japanese Cognitive Linguistics Association Conference Handbook, 153–156. Vygotsky, Lev S. 1962[1934]. Thought and Language, trans. by Eugenia Hanfmann & Gertrude Vakar. Cambridge MA & New York NY: The MIT Press & John Wiley & Sons. Yamada, Yoshio. 1954[1936]. Nihon bunpōgaku gairon (Introduction to Japanese grammar). Tokyo: Hobunkan. DOI: 10.1037/11193-000
Subjectivity, intersubjectivity and Japanese grammar A functional approach Rumiko Shinzato This survey study intends to show the benefits of adopting the functional linguistic perspectives of emphasizing the non-referential and emotive/ subjective aspects of language, which are largely neglected in mainstream formalist linguistics. Specifically, this paper will first point out the fact that both Japanese linguist Haga and French linguist Benveniste independently proposed very similar dichotomies, juttei vs. dentatsu (Haga 1954) and subjectivity vs. intersubjectivity (Benveniste 1971[1958]) for their analyses of the nonpropositional parts of language. Subsequently, drawing upon published studies, this paper will point out the relevance of these two dichotomies to syntactic organization (predicate order in particular), soliloquy and dialogue (or mental and speech act) dyad, and diachronic change (namely, the unidirectionality in grammaticalization from subjectification > intersubjectification (Traugott 2003).
1. Introduction1 It is probably fair to say that mainstream formalist linguistics has built on analyses of discrete and close-ended data, out of discoursal context, and at the exclusion of the speaker (cf. Givón 1982; Hopper 1997). It has treated language as an abstract, static, mechanical, and non-humanistic object (Finegan 1995). Actual data, which faithfully represent the reality of language in everyday life, but do not conform to the formalist models, were ignored and conveniently replaced by introspection (Laury & Ono 2005). Naturally, in this school, referential, rational and objective aspects of linguistic expressions were of the utmost concern, and non-referential, emotive, and subjective aspects were largely neglected (Lyons 1982, 1995; Mushin 2001; Maynard 2002; Suzuki 2006). Thus, the downplaying or dismissal of the 1. I would like to thank Kaori Kabata and Tsuyoshi Ono for their valuable comments and advice on an earlier version of this paper. I am also grateful to a reviewer of this paper for her/ his comments. Any shortcomings remain my own.
Rumiko Shinzato
speaker has been commonplace in the generative tradition (Mushin 2001: 3), let alone other interactants, conversational context, and subtle cues of participants’ visual and bodily conduct, which have been vigorously investigated in Conversational Analysis (CA) (Selting & Couper-Kuhlen 2001; Prevignano & Thibault 2003; Hayashi 2003). In stark contrast, a branch of traditional Japanese linguistics (kokugogaku) has always been concerned with the subjective, emotive and interactive parts of language, which have been integral to the Japanese language (Tokieda 1941; Watanabe 1953; Minami 1974). Testimonial to this is a long tradition of Japanese linguistic studies (Watanabe 1953, inter alia) on chinjutsu (roughly subjectivity + intersubjectivity2 in the sense of Benveniste). Closely related to chinjutsu is the notion of ima/koko/watashi ‘now/here/I’, which has been widely recognized and integrated in grammatical descriptions of Japanese (Nakau 1994; Watanabe 1995; Onoe 2001 [1998]).3 Not surprisingly, this notion relates to ground in cognitive linguistics (Langacker 1991), deixis (Bühler 1934; Lyons 1977; Duchan et al. 1995), and subjectivity (Benveniste 1971[1958]) in Western functional linguistics. The importance of non-referential and emotive/subjective aspects of the Japanese language has been also emphasized in Western functional linguistics. A 2. An anonymous reviewer offers a different interpretation of Benveniste’s notion of “intersubjectivity.” S/He claims that “Benveniste’s intersubjectivity is really a close relative of ‘objectivity’ (shared knowledge etc.), as the term is used in philosophy, rather than of ‘subjectivity’ in the Traugottian (or Haga) sense.” Here, I still follow Traugott’s interpretation of Benveniste (Trangott & Dosher 2002: 20): “Benveniste saw the SP/W-AD/R dyad as the condition or ground for linguistic communication, and characterized this relationship as one of ‘intersubjectivity’ – in communication each participant is a speaking subject who is aware of the other participant as speaking subject.” (SP/W-AD/R = speaker/writer–addressee/reader) 3. The speaker’s now relates to Akatsuka’s (1979: 10–11) notion of immediate experiencer. Akatsuka claims that the speaker cannot express disparagement of himself at the speaker’s now, when he is an immediate experiencer as in her example (i), but he could negatively evaluate himself when he is not an immediate experiencer as in her example (ii). (i) *watashi wa orokanimo yuubinya ga tegami o kaihuu-shite iru to omou. (present) (ii) watashi wa orokanimo yuubinya ga tegami o kaihuu-shite iru to omotte ita. (past) I stupidly mailman letters opening is that think ‘I stupidly {(i) think/(ii) thought} the mailman {(i) is/(ii) was} opening our mail.’ e speaker’s now is also analogous to Nakau’s (1979: 235–238) notion of the speaker’s instantaTh neous present. The simple present tense omou refers to the speaker’s instantaneous present, while the progressive, omotte-iru does not. The negation of the former is impossible as in (iii), but the latter can be negated, as in (iv) (Both are Nakau’s examples with my gloss and translation). (iii) * watashi wa Ann wo shoojikida to omou no dewa-nai. (iv) watashi wa Ann wo shoojikida to omotte-iru no dewa-nai. I top obj honest comp think that is not ‘It’s not that I (iii) think/(iv) am thinking that Ann is honest.’
Subjectivity, intersubjectivity and Japanese grammar
common and pervasive belief among Japanese researchers in this camp is that Japanese is so imbued with subjectivity or intersubjectivity that it is impossible to talk about Japanese grammar (conditionals, complementizers, sentence-final forms, temporals, giving-receiving verbs, conjunctions, etc.) without reference to this key notion (see Akatsuka 1979, 1985; Iwasaki 1993; Kamio 1997; Kuno 1987; Kuroda 1973; also Maynard 1993). For instance, Maynard (1993: 4) comments that it is impossible to speak Japanese without expressing “one’s personal attitude toward the content of information and toward the addressee”. Following the spirit of these previous studies both in Japan and the US, I will show the benefits of adopting functional linguistic perspectives. Specifically, this paper will first point out the fact that both Japanese linguist Haga and French linguist Benveniste independently proposed very similar dichotomies, juttei vs. dentatsu (Haga 1954) and subjectivity vs. intersubjectivity (Benveniste 1971 [1958]) for their analyses of the non-propositional parts of language. Subsequently, it will point out the relevance of these two dichotomies to syntactic organization (predicate order in particular), soliloquy and dialogue (or mental and speech act) dyad, and diachronic change (namely, the unidirectionality in grammaticalization, in the sense of Traugott 2003). Readers should be warned that this paper is meant to be a survey study in nature, not meant to be an original case study. 2. Subjectivity vs. intersubjectivity Traditionally, the non-referential part of the sentence has been referred to as modality (mood) vis-à-vis proposition, the referential part of the sentence. Modality is an extremely broad and general concept that includes such concepts as deontic/ epistemic distinction, epistemic qualification, evidentiality, mood, perspective, and temporality (Narrog 2005). Though these sub-concepts are useful and enlightening in their own light, this paper capitalizes on subjectivity and intersubjectivity. This is because it is through this concept that the intertwinedness of structure, semantics/pragmatics, and diachrony is clearly exhibited in Japanese. Subjectivity and intersubjectivity are defined as follows: Subjectivity denotes the speaker’s attitude towards the proposition, while intersubjectivity exerts the illocutionary force directed towards the addressee.4 In Japanese linguistics, Haga 4. The following definitions on subjectivity may serve as a useful reference for (inter)subjectivity: Lyons (1982: 102) – “the term ‘subjectivity’ refers to the way in which natural languages, in their structure and their normal manner of operation, provide for the locutionary agent’s expression of himself and his own attitudes and beliefs...”
Rumiko Shinzato
Table 1. Comparison of Haga’s and Benveniste’s concepts5 Haga (1954) Benveniste (1971[1958])
juttei ‘judgment’ = the speaker’s attitude toward the proposition subjectivity = the expression of “the attitude of the speaker with respect to the statement he is making”
dentatsu ‘communication’ = communication of proposition and propositional attitude to the addressee intersubjectivity5 = “which alone makes linguistic communication possible”
(1954) first proposed this division in chinjutsu (roughly ‘modality’). The Japanese terms he created are juttei for subjectivity, and dentatsu for intersubjectivity. In Western linguistics, Benveniste (1971[1958]) recognized this distinction independently (see Table 1 for their correspondence). From their parallel characterizations as seen in Table 1, it becomes evident that the key words for subjectivity and intersubjectivity are “speaker’s attitude” and “communication” respectively. What is more, it also becomes apparent that the existence of the addressee in context is essential for the latter, though it is not the case for the former.6 As will be discussed in Section 2.2, the presence or nonpresence of the addressee is critical for the mental vs. speech act verb distinction as well as the soliloquy vs. dialogue dyad. Having seen two extremely similar dichotomies arising from the opposite ends of the world, one may ask if there is any relationship between subjectivity and intersubjectivity and if there is one, what the nature of the relationship is. Traugott and Dasher’s (2002: 22) view resonates through this paper: “Subjectivity is a prerequisite to intersubjectivity, inasmuch as SP/W’s attitude toward AD/R’s is a function of the perspective of SP/W (RS: SP = speaker, W = writer, AD = addressee, R = reader)”. The following two sections will demonstrate how this implicational Finegan (1995: 1) – “...expression of self and the representation of a speaker’s (or, more generally, a locutionary agent’s) points of view in discourse – what has been called a speaker’s imprint...the intersection of language structure and language use in the expression of self.” 5. As an anonymous reviewer points out, this should be understood as a function of “intersubjectivity” rather than the definition of it. 6. Verhagen (2005: 7–8) include two conceptualizers, the speaker and the hearer, in his formulation of the construal configuration, and claims as far as to say “...even in the absence of an actual addressee, a speaker (for example, one making a note in a personal diary) is committed to the assumption that her utterance is in principle interpretable by someone else sharing the knowledge of certain conventions. The idea that some utterance could in principle only be interpretable for a single individual makes the idea that it is an instance of language void.” However, he (ibid: 18) does recognize the case where only the speaker is profiled, such as the context when the speaker utters a non-interactional sign of disgust or frustration (Yech, Damn).
Subjectivity, intersubjectivity and Japanese grammar
relationship is substantiated in Japanese syntax, soliloquy/dialogue dyad, and diachronic change. 2.1
Predicate order
It is well recognized in traditional Japanese linguistics that Japanese predicates, that is, verbs, auxiliaries, and sentence-final particles, are connected in a fixed order, which is not reversible. As seen in example (1), adopted from Haga (1954), a verb is first followed by the sentence-final particle ka, expressing the speaker’s doubt, and it is then followed by another sentence-final particle, ne, which functions to form the addressee-directed question. Needless to say, the former corresponds to subjectivity and the latter to intersubjectivity as defined in this paper. (1) [[[shibai ga hajimaru] ka] ne7] play sbj start Doubt/Wonder Question [[[jojutsu] =‘proposition’ I with I only with I + II with I + II + III
+ juttei] =‘subjectivity’
+ dentatsu] =‘intersubjectivity’
II III ‘The play will start.’ ‘I wonder if the play will start.’ ‘(I wonder, therefore I ask) Is the play starting?’
Since the reverse order, ne ka, is not possible, it follows that the speaker first expresses his doubt about the proposition (i.e., the play is starting), and then directs the question to the addressee (i.e., illocutionary act of questioning). It is evident that Japanese predicative order proceeds from proposition to subjectivity and finally to intersubjectivity. Or, in Haga’s terms, it goes from jojitsu, to juttei, and then to dentatsu. The strict unidirectionality observed here is consistent with the implicational relationship suggested by Traugott and Dasher (2002) as mentioned above. Haga’s view has been further explored in Minami (1974), which analyzes Japanese sentences to be organized in a layered structure with four distinct levels in it. In Table 2, elements in bold represent ones unique to that level and not shared by the lower level, but included in the next higher level.8 7. Haga (1954) treats ne and naa to be different functionally. The former is characterized as dentatsu ‘intersubjective’, while the latter juttei ‘subjective’. An utterance can end with the subjective elements alone. 8. The VP in English in the early days of generative grammar is formulated as (a). In this PS rule, the verb phrase might have been singing has a deep structure like (b), from which a surface structure (c) is produced after affix hopping is applied (see Binnick 1991).
Rumiko Shinzato
Table 2. Minami’s sentence production levels Level
Example
A
(Nimotsu ga) Yokohama ni tsuku (luggage sbj) at arrive ‘(The luggage) will arrive at Yokohama.’ Kinou nimotsu ga Yokohama ni tsui-ta yesterday arrive-past ‘The luggage arrived at Yokohama yesterday.’ Nimotsu wa tabun kinoo Yokohama ni tsuita daroo top probably inf ‘Probably, the luggage arrived at Yokohama yesterday.’ Soodana, nimotsu wa tabun kinoo Yokohama ni tsuita daroo yo Well I-tell-you ‘Well, I tell you, probably, the luggage arrived at Yokohama yesterday.’
B
C
D
Note here that Minami augmented Haga’s analysis by recognizing a similar hierarchical order in nominal elements as well. According to Minami, levels from A to D represent an increasing degree of sentence-hood (bunrashisa). That is, level D represents a syntactically more complex, and semantically richer sentence than level A. Note also that Minami’s levels C and D correspond to Haga’s juttei and dentatsu, or subjectivity and intersubjectivity respectively. The layered structure advanced in traditional Japanese linguistics finds its counterpart in Role and Reference Grammar (Foley & Van Valin 1984), Hengeveld’s (2005 [1989]) layered model, and Rijkhoff ’s (2002) model of noun phrase and verb phrase correspondence. In Hengeveld’s layered model (2005 [1989]: 5–6), four different operators are nestled in a structure as in Figure 1.
a. VP → Tense (Modal) (have + en) (be +ing) V b. VP → Past may have en be ing sing c. VP → may + Past have be + en sing + ing
is elegant and simplistic treatment of the English VP with enormous generative power Th appeared extremely attractive and convincing to students. Attractive and workable though it may be at a purely syntactic and mechanical level, placing the precedence of tense over modal is counterintuitive on semantic grounds. As Nakau (1979: 225–229) rightfully argues (see footnote 2), modality (i.e., subjectivity as used in this paper) is only concerned with speaker’s instantaneous present, thus it is outside the scope of tense. The Japanese predicative order corroborates this. Tense belongs to Minami’s B level, but modality is an element of C level. The evidential/epistemic mood ⊃ tense in scope is claimed in Cinque’s (1999) cross-linguistic study of adverbs as well. In addition, Bybee & Paglinca’s (1985: 33–35, 196–200) research finds that of fifty languages, all but one conform to the same scope relation (mood ⊃ tense).
Subjectivity, intersubjectivity and Japanese grammar (E1: [π4 ILL (S) (A) (π3 X1: [proposition] (X1))] (E1) (π2 e1: [π1 Predβ (X1) (X2)...(Xn)] (e1)) π1: Predicate operators π2: Predication operators
π3: Proposition operators π4: Illocution operators
Figure 1. Hengeveld’s model (2005 [1989]: 5)
The definitions of the operators from the innermost π1 to the outermost π4 are given in (2): (2) Predicate operators (π1) capture the grammatical means which specify additional properties of the set of SoAs (RS: state of affairs) by a bare predication. Predication operators (π2) capture the grammatical means which locate the SoAs designated by a predication in a real or imaginary world and thus restrict the set of potential referents of the predication to the external situation(s) the speaker has in mind. Proposition operators (π3) capture the grammatical means through which the speaker specifies his attitude towards the (truth of the) proposition he puts forward for consideration. Illocution operators (π4) capture the grammatical means through which the speaker modifies the force of the basic illocution of a linguistic expression so as to make it fit his communicative strategy.
It is interesting to see that his proposition and illocution operators parallel Haga’s juttei and dentatsu and Benveniste’s subjectivity and intersubjectivity dyads. Just as Haga’s juttei and Benveniste’s subjectivity are concerned with the speaker’s attitude towards the proposition (cf. Table 1), so are Hengeveld’s proposition operators. Similarly, as Haga’s dentatsu and Benveniste’s intersubjectivity put the communicative aspect of language use in focus, so do Hengeveld’s illocution operators. In addition, the hierarchical relationship held between them (Illocution operators ⊃ Proposition operators) is also consistent with that of intersubjectivity (dentatsu) ⊃ subjectivity (juttei). In Minami’s model, the nominal elements unique to each level have close semantic relationships with their corresponding predicate elements in the same level (cf. Table 2). For instance, at Level B, kinoo ‘yesterday’ is in a natural liaison with the past tense marker, -ta. Similarly, at Level C, the modal adverb, tabun ‘probably’ readily connects to the inferential auxiliary, daroo, and the same is true for the semantic tie between the discourse marker, soodana ‘well’ and the interpersonal particle, yo. We also find in Rijkhoff ’s (2002) model (cf. the top half of
Rumiko Shinzato Location
π2b
σ2b
Quantiy
π2a π1
Quality
σ2a σ1
Time
tense
semalfactive iterative, etc. aspect
demonstr. number pronoun numeral
verbal aspect nominal aspect
VERB
adverbs/adverbials of: manner speed, etc. frequency
NOUN
adjective
Space ω1 ω2a ω2b
Quality Quantity Location
lexical numeral
time place Rel. cl possessor NP, etc.
τ1 τ2a τ2b
Figure 2. Symmetry in the underlying structure of the clause and the NP Rijkhoff (2002: 224)
Figure 2)9 similar bi-directional extensions, as in π1 (=qualifying clause operator) corresponding to σ1 (=qualifying clause sattellite). Rijkhoff (2002: 216) states that “in the layered model elements of the linguistic expression that belong together semantically also occur together in the underlying structure of that linguistic expression.” Thus, he contends that verbal aspects π1 and adverbs of manner and speed σ1 come closest to the semantic nucleus, the verb. In addition, Rijkhoff also recognizes a similar bidirectional extension in NP (the bottom half of Figure 2). He sees that as the verb phrase expresses a verbal aspect (Aktionsart= ‘mode of action’), a noun phrase represents a nominal aspect (Seinsart= ‘mode of being’) and both are semantically parallel. The symmetry between NPs and clauses proposed by Rijkhoff neatly illustrates that the orders in NP and the clause are not only fixed, but also organized by the same principles. He names three such principles: principle of domain integrity, principle of head proximity, and principle of scope. What underlies these, according to Rijkhoff (2002: 253), is a more general 9. In Rijkhoff ’s later model (2005: 87), two more layers are added: ‘Kind’ in the inner layer of ‘Scope’ and ‘Discourse-Referential in the outer layer of ‘Location’.
Subjectivity, intersubjectivity and Japanese grammar
iconic principle that states “what belongs together semantically is also placed together syntactically”.10 The quality of a noun is more or less its inherent or characteristic properties which relate only to a nominal head, while quantity and location are non-inherent and external features of a referent (ibid: 220). Another significant contribution in this connection is offered by Nuyts who argues that her hierarchical relationship below (2001: 347)11 stems from differences in conceptualization and cognition. (3)
evidentiality ⊃ epistemic modality ⊃ deontic modality ⊃ time ⊃ quantificational aspect ⊃ qualificational aspect
In the above cline from the bottom end to the top end, Nuyts sees the gradual broadening of scope to correlate with widening scope of the state of affairs involved, as in the relation between participants (lower level) ⊂ state of affairs as a whole (medium level) ⊂ external situation (higher level). Cognitively, Nuyts interprets the cline (from bottom to top) to be representing (ibid: 355) “a decreasing role of direct perception of the state of affairs, and an increasing role for interpretation and creative involvement on the part of the speaker.” In other words, the lower ends on the ladder indicate what the speaker directly perceived, while the higher ends exhibit abstract deductive reasoning from perceptions of other states of affairs. The increasing involvement of the speaker towards the higher end is extremely reminiscent of the Japanese predicate order, especially the hierarchical relationship between proposition ⊂ subjectivity, as Nuyts’ quantificational/qualificational aspect is considered to be in the domain of proposition, while epistemic modality clearly belongs to the subjective domain involving the speaker. As noted with examples (6), Nuyts’ projection of the hierarchical relationship between mental and speech act verbs is also consistent with subjectivity ⊂ intersubjectivity. Summing up this section, it was demonstrated that both Western and Japanese linguistics recognized the concepts of subjectivity and intersubjectivity and 10. Similarly, Narrog (2002: 233) also views the correlation between structural and semantic hierarchy in Japanese to be grounded in Givόn’s (1995: 51) proximity principle: “Entities that are closer together functionally, conceptually, or cognitively will be placed together at the code level, i.e. temporally or spatially.” 11. The hierarchical relationship presented here is the same in content, but slightly different in appearance from its original. Her horizontal layout is replaced by a vertical one, and her symbol ‘>’ is replaced by ‘⊃’ to avoid confusion since the ‘>’ is used to indicate the directionality of a diachronic change in this paper.
Rumiko Shinzato
consistently represented their scope relation as intersubjectivity ⊃ subjectivity in a hierarchical structure. Accounts for such fixed orders based on iconicity and cognition were also provided. The next section explores the relevance of the subjectivity vs. intersubjectivity dyad to semantic/pragmatic dichotomies.12 2.2
Mental vs. speech act verb dichotomy
2.2.1 Implicational relationship between mental and speech act verbs The implicational relationship held between subjectivity and intersubjectivity also has a bearing on the mental vs. speech act verb relationship. From the subjectivity and intersubjectivity dyad as described in Table 1, it is a very small leap to associate it with the mental and speech act verb opposition, or “I think X” and “I say X” contrast. It is also natural to wonder if the same heirarchical relationship (intersubjectivity ⊃ subjectivity) holds true for the mental and speech act verb relationship. Indeed, such an implicational relationship is observed in these two distinct types of verbs. Both Leech (1983) and Nakau (1994) claim that what speech act verbs refer to meta-implicate what is described by mental verbs. For instance, Nakau (1994: 85) offers the following implicational relationship as in (4). (4)
Speech act verbs a. I say/state/assert/claim/tell you (that) b. I ask (you) inquire/question (wh-) c. I promise (to do) d. I order (you to do) e. I confirm (that) f. I deny (that)
Mental verbs ⊃ I believe (that) ⊃ I wonder (wh-) ⊃ I intend (to do) ⊃ I want (you to do) ⊃ I know (that) ⊃ I doubt (that)
Nakau explains that in order to declare Gus’ guilt in (5b), the speaker has to first believe in that proposition. Likewise he maintains that one asks a question about something, because one wonders about it (cf. the example in [1] where ka ‘wonder’ is followed by ne ‘ask’)13. In other words, what is embodied in mental verbs is a prerequisite for the acts denoted by their corresponding speech acts to take place. (5) a. I believe that Gus is guilty. b. I assert that Gus is guilty.
12. The structural hierarchy and “scope increase” in grammaticalization are discussed in more detail in Shinzato (2007). 13. For a more elaborate study of mental and speech act verbs such as their structural parallel, semantic proximity and implicational relationship, please refer to Shinzato (2004).
Subjectivity, intersubjectivity and Japanese grammar
Undoubtedly, the embeddability of mental verbs in speech act verbs, but not vice versa, is also observed in English, thus furthering the point above. Nuyts (2001: 318) observes that on semantic grounds, (6a) is plausible, but (6b) is not: (6) a. I say that I think that they have run out of fuel. b. ?I think that I say that they have run out of fuel. 2.2.2 Mental vs. speech act verb distinction to soliloquy vs. dialogue distinction Another interesting point about the mental and speech act verb opposition is Nakau’s remark that speech act verbs always assume the existence of the underlying second person ‘you’. This is because it is meaningless to engage in questioning, ordering, or promising if there is no one to whom these speech acts are directed. In principle, speech acts are communicative acts, and thus necessitate the existence of the second person. The absence vs. presence of the underlying you recalls the two distinct modes of speech, namely, soliloquy and dialogue, since soliloquy requires no audience, whereas dialogue does. What this leads to is the pairing of mental verbs to soliloquy and speech act verbs to dialogue. In fact, whether a proposition can be embedded in ~to omou ‘I think that ~’ serves as a diagnostic test for whether or not it can be uttered in soliloquy (see Nitta 1991, Moriyama 1997). Two volitional expressions in Japanese show an interesting contrast with regard to the embeddability in omou ‘I think’. The plain verb, expressing volition, can be embedded in omou as in (7), but when the plain verb is annexed by tsumori da ‘intend to~’, it can not be embedded in omou ‘I think’, but can be in iu ‘I say’ as shown in (8). What can be embedded in omou is apparently something which goes through the speaker’s mind privately, and thus it can be uttered in soliloquy. In contrast, what is embedded in iu ‘I say’ is what is to be uttered to someone in dialogue. (7) Ore wa itsuka yatsu o nagutte-yaru to omotta. I top someday him obj hit-give comp thought ‘I thought I’d hit him some day.’ (8) Boku wa ganbaru tsumorida to (*omotta/itta). I top try hard intend to comp thought/said ‘I (thought/said) that I intend to try hard.’ Japanese lexical sensitivity to the soliloquy/dialogue opposition is also seen in the two distinct question particles in Old Japanese (OJ). The two OJ question particles, namely ka and ya, are a case in point. It is generally agreed that ka forms a selfinquiry, doubt, or wonder, while ya makes an other-inquiry (see Ono, 1993; Sakakura 1993; Serafim & Shinzato 2000; Shinzato & Serafim 2003, 2013). Ayuhisho, a grammar book written in 1778 (quoted in Ono 1993) differentiates the two as omohu ‘wonder/doubt’ and tohu ‘ask’. Similarly, Sakakura (1993) notes that ka
Rumiko Shinzato
frequently cooccurs with omou as in ~ka to omou ‘I wonder if ’, whereas ~ya is coupled with tohu ‘I ask’ as in ~ya to tohu ‘I ask if ’. Observe examples (9) and (10) below: (9)
Wa ga seko ni mata ha aha-ji ka to omohe-ba my love dat again top see-neg kp comp think-if ka kesa no wakare no sube nakari-tsuru kp this morning ’s parting sbj helpless-perf ‘I wonder if it is because I think I won’t see my love again that the parting this morning was (so) helplessly (sad).’ (Manyooshuu 540)
(10) Hatsusegaha haya-mi hayase wo musubi-agete fast-since water obj scoop aka-zu-ya imo to tohi-shi kimi ha mo satisfy-neg-kp my love comp ask-past you top emph ‘Since Hatsuse rapids were fast, you scooped the water (for me with your hands and asked’, “My love, you haven’t had enough (water), have you? (Have it to your heart’s content)”’ (Manyooshuu 1452) In Okinawan, a sister language to OJ, sentence-internal ga and sentence-final i (the cognates of the OJ kakari particles, ka and ya respectively) unexceptionally form self-directed vs. other-directed questions.14 Consequently, ga can be uttered in soliloquy, but i has to be directed to an interlocutor in a dialogue. The mental vs. speech act verb contrast of want and order is also lexicalized in OJ as the irrealis + na vs. irrealis + ne. The former expresses the speaker’s wish/ hortation, while latter indicates his request toward the addressee (Tanabe 1953; Morishige 1971; Yamaguchi 1985). (11) Yamatachibana wo tsutoni tsumi-ko-na. mountain orange obj gift pick-come-na ‘I wish to go pick mountain oranges for a gift (to someone).’ (Manyooshuu 4471) (12) Hitome mi ni ko-ne. just once see to come-ne ‘Just once, come see it.’
(Manyooshuu 4077)
These lexicalized forms did not survive into Middle Japanese (9C~). But their cognates in Okinawan, that is, the irrealis form + na and the irrealis form + i, are still extant, and express the same contrast of the speaker’s intention vs. the request 14. Together with a wh-word, the sentence-final ga forms a wh-question directed to the addressee, complementing i, which only forms a yes/no question. This ga is differentiated from the sentence-internal KP ga. For more detailed discussion, see Shinzato and Serafim (2000).
Subjectivity, intersubjectivity and Japanese grammar
for the addressee. That is, the mental vs. speech act verb contrast of want and order is readily seen in Okinawan cognates of OJ irrealis + na vs. irrealis + ne. Needless to say, the Okinawan irrealis + i cannot be uttered in soliloquy, but only in dialogue. In Modern Japanese, the sentence-final particle ne is noted for its function to build collaboration, cooperation, and confirmation with the addressee (Saji 1957, Cook 1989, inter alia). In this context, it is illuminating to observe two experimental studies done on this particle. Tsuchii and Omori (2000) explored the frequency of the use of ne in relation to aizuchi, chiming-in, or back-channeling. In Japanese discourse, it is commonplace for the listener to signal his/her attentiveness and involvement (i.e., ‘I am listening’) by nodding his/her head frequently (cf. Mizutani 1981). In Tsuchii and Omori’s study, two control groups were formed unbeknownst to the participants (students). When students presented their research to their teacher, the teacher was instructed to limit chiming-in gestures in the first group, while this restriction was lifted for the second group. Tsuchii and Omori’s study found that the students in the first group (without chiming-in) used ne at a noticeably higher frequency rate than the students in the second group. They interpret this difference as follows: when the expected signal is not sent by the listener (the teacher), the speaker senses some oddity about his interaction, and thus is compelled to use ne more often to remedy the awkward situation and establish a healthy speaker-listener interaction. Watamaki (1997) analyzed tape-recorded conversation data between a care-taker and two children: a child with mental retardation, and a child with autism. The finding of this study is that the child with mental retardation used ne frequently just like a normal child, but the autistic child did not use ne at all. This difference was only observable for ne, and not for the other particles. Watamaki interprets this as the result of impairment of interpersonal skills in the autistic children. What these two studies illustrate is the highly interpersonal and intersubjective nature of ne, in accordance with Haga’s characterization of its function as dentatsu ‘communication’. What emerges from this subsection is a three-way relationship as summarized in Table 3, in which the subjectivity/intersubjectivity opposition relates to the mental/speech act verb dichotomy, by way of which the subjectivity/intersubjectivity dyad further parallels the soliloquy/dialogue distinction. Table 3. Three-way correspondence subjectivity intersubjectivity
mental verbs speech act verbs
soliloquy dialogue
Rumiko Shinzato
Though the mental vs. speech act verb distinction, or the soliloquy vs. dialogue dyad, could hardly become objects of serious research in formalist linguistics, in cognitive linguistics, the subjectivity vs. intersubjectivity dyad made an inroad into research. According to Evans (2004: 34), what he calls subjective information (information of internal states) is encoded in body format representation, while what he calls intersubjective information (visual-spatial information from the external world) is encoded in 3D format. The former feeds into the latter, which is further elaborated as a conceptual system, then as a linguistic system. This also shows the far-reaching nature of the notions of subjectivity and intersubjectivity. 2.3
Unidirectionality in grammaticalization
The previous sections discussed the heirarchical and implicational relationship between subjectivity and intersubjectivity in a synchronic frame. This section points out its additional relevance to diachronic change, especially to the unidirectionality in grammaticalization. Not surprisingly, the subjectivity/intersubjectivity dichotomy finds its counterpart in grammaticalization theory as subjectification/i ntersubjectification, which Traugott (2003: 128) defines as follows: (13) “...while subjectification is a mechanism whereby meanings become more deeply centered on the speaker, intersubjectification is a mechanism whereby meanings become more centered on the addressee....The hypothesis is that, for any lexeme L, intersubjectification is historically later than and arises out of subjectification.” A perfect example of the shift from subjectification to intersubjectification comes from Japanese as in the development of mental verbs into speech act verbs. Traugott and Dasher (1987: 570) note the following development (MV and SAV stand for mental verb and speech act verb respectively). (14) kotowaru: attested as MV meaning <> from the 8th C, and a SAV with the meanings <> from the late 12th C, <> from the late 13th C. The present meaning <> is attested from the mid 19th C. mitomeru: mi-miru <> + tomeru <> attested as a MV meaning <> from the 13th C, and as a SAV from the 17th C. A similar path is also observed in the English verb, find, as below: (15) find: appears from OE on as a MV; does not appear as SAV until 1400, and then only in legal contexts (e.g. find guilty means <>)
Subjectivity, intersubjectivity and Japanese grammar
Referring to such a unidirectional developmental path, they (ibid: 570) state: (16) “Being in a certain state of mind is a prerequisite for a speech act... claim involves the speaker’s belief in the proposition, order involves the speaker’s desire for the addressee to do whatever is named in the proposition...” Note that their remark is strikingly similar to Nakau’s implicational relationship (cf. example 4), and in line with the Japanese predicative order (cf. example 1). In this junction, it should be mentioned that Romaine and Lange (1991: 265) follows Traugott and Dasher (1987) in hypothesizing that the expression, be like first developed as a marker of thought as in (17), and then as a marker of speech, as in (18). (17) My mother said, “David where are you,” and he just came right out. I was like, I thought I lost him. I really thought I lost him. (ibid: 265) (18) I was like, “Mom?” She was like “What!” (ibid: 253) Romaine and Lange (ibid: 243) found that in their data, like tends to be used for self-representation, say for the speech of others, and go for both. Regarding this statistical difference, they state, “Insofar as only the speaker can have access to his/ her own thoughts, and like (rather than say or go) is more likely to be used for the representation of thought, this trend is not surprising.” The higher rate of the occurrence of be like with the 1st person subject is also noted in Tagliamonte and Hudson (1999). After all, we know what others are thinking from what they say, not by our omniscient power. Another relevant mental vs. speech act contrast can be seen in the developmental history of I see – you see. According to OED, I see as in the sense of ‘understand’, that is, the next stage of development from its concrete visual perception sense, was attested around 15C as in (19), while you see as an interpersonal discourse marker appeared around 17C as in (20). This again conforms to the directionality of subjectification > intersubjectification. (19) Now I see and vnderstande that myn old synne hyndereth me and shameth me. (1470–85 MALORY Arthur XIII. xix. 639) (20) Because, you see, the present Government has 1,900,000 l. (1657 CROMWELL Sp. 21 Apr. in Carlyle Lett. & Sp.) Some Japanese modal auxiliaries such as rashii ‘it appears’, yooda ‘it seems’ and sooda ‘I hear’ are polysemous between evidential meaning and mitigating meaning. In Nitta’s (1992) example (21), yooda (yoo desu) has an evidential function, indicating the proposition it appends to was yielded through the speaker’s inferential process based on the available information (in this case what is expressed in the previous sentence). In contrast, the same auxiliary in (22), does not express
Rumiko Shinzato
any evidential meaning, but rather it makes the tone softer and indirect. Thus, it concerns more with politeness than evidentiality. Needless to say, the former aligns with mental verbs, while the latter does so with speech act verbs. (21)
Me-o-korasu to rei-no-futari ga butchouzura-de dete kita. gaze when those two sbj serious-look came out Doomo senka wa sappari datta yooda Somehow results top not great was seem ‘When I was gazing, those two came out with a serious-look (on their faces). It seems that the results were not that great.’
(22)
(After checking his watch to make sure the time has already passed) Jikoku ni natta yoodesu. Honjitsu no kaigi wa kore de time has come seem today ’s meeting top this with ohiraki-ni shi tai to omoimasu. ajourn make would like to comp think ‘It seems the time is up. I would like to adjourn today’s meeting now.’
Nitta (1992: 7) asserts that what he calls ‘communication modality’ as in (22) was developed later than ‘judgement modality’ as in (21). Though his analysis is synchronic in essence, nonetheless, here again, the same directionality from mental to speech act verbs, or subjectification → intersubjectification is confirmed.15 Though not reflecting the contrast between mental and speech act verbs per se, Japanese examples such as (23) are also consistent with the unidirectionality (subjectification > intersubjectification) under discussion. As illustrated below, this unidirectionality is not limited to a particular part of speech, but applies to items across the board: (a) verb; (b) formal noun; (c) clausal connective; (d) quotative conditional; and (e) sentence-final particle. In the following examples, the earlier stages depict the happenings in the speaker’s inner world (affect, inference, judgement, exclamation), while the later stages exemplify other-oriented speech acts (camaraderie, invited hearer’s inference, vocative, summon). The earlier stages correspond to subjectification, as opposed to the later ones which exhibit intersubjectification: (23) a. te shimau (affective marker > social dialect/camaraderie) (Strauss & Sohn 1998)
15. In the case of yooda/yoodesu, the subjective/intersubjective distinction does not seem to be ‘dichotomous’, but rather ‘a matter of degree’. In fact, Traugott (2003: 134) also states, ‘...although it may in some instances be difficult to determine whether a new meaning is strictly subjective before it becomes intersubjective, nevertheless, nonsubjective > intersubjective > subjective is hypothesized not to be likely.’
Subjectivity, intersubjectivity and Japanese grammar
b. wake (speaker’s inference > invited hearer’s inference) (Suzuki 1998); mono (subjective judgement > amae, dependency on the addressee (Fujii 2000) c. demo (subjective adversative connective > discourse marker of claiming floor, and changing topic) (Onodera 2004) d. ttara (metalinguistic and subjective judgement > vocative) (Shinzato 2007) e. na (exclamation > summon) (Onodera 2004) In summing up this section, it discussed the diachronic counterpart of the synchronic hierarchical relation between subjectivity and intersubjectivity. Drawing on existing studies, it demonstrated that the diachronic shift from mental verbs > speech act verbs, and also the shift from mental acts > speech acts are prevalent in Japanese across grammatical categories. 3. On the fundamentality of subjectivity and intersubjectivity The above discussions on the relevance of the notions of subjectivity/intersubjectivity to Japanese syntactic organization, the mental/speech act verb difference, soliloquy/dialogue distinction, and furthermore, the unidirectionality in grammaticalization lead to the close mesh of synchrony, diachrony and modes of speech. As is well known, in some formalist frameworks, morphology is seen as autonomous, modular, and divorced from diachronic and comparative analyses (Aronoff 1994). In the functional frameworks, morphology is viewed to be intertwined with diachronic, psychological as well as communicative processes (Bybee et al. 1994; Axelrod 1999). As has been substantiated on neurological grounds (see Ariel 1998: 15), it is probably true that some level of modularity or autonomy in morphology exists. However, faced with the rigid predicate order in Japanese as well as crosslinguistically and its interdependence with other seemingly unrelated aspects of language, diachrony and speech modes, it is difficult, if not impossible, to ignore such interdependence of morphology and other components of language.16 It 16. Ariel (1998: 15) states that Fodor’s (1983) input-system and central system distinction may have a neurologically based parallel: “Left brain-damaged patients have selective semantic, syntactic, and morphological deficits. Right brain-damaged subjects are impaired in drawing inferences based on contextual assumptions, in extracting the text macrostructure, the discourse theme, the main point, coherent, connections among textual propositions, a distinction between important and trivial elements, humorous points, inferences about the emotional state of others, etc.” Additionally, Cain (2002) supports the modularity of Fodor’s input system based on pathological data. However, for a critique of Fodor’s theory, see Prinz (2006).
Rumiko Shinzato
Table 4. Related notions and studies Notion
Applicable to:
Example
‘ego vs. non-ego’ (McCawley 1978)
complementizer choice
‘waga-koto ‘my affairs’ vs. hito-goto’ ‘other people’s affairs’ (Watanabe 1991) ‘experiencing self vs. observing self ’ (Shinzato 2003) uchi ‘inside’ vs. soto ‘outside’ (Quinn 1994)
adverb choice
no = S’s own mental state; internalized knowledge koto = others’ mental state; not yet internalized knowledge zuibun ‘very’ cannot modify waga-koto, but taihen ‘very’ can.
simple vs. stativized verb form transitivity, aspect, modality, nominalizers, locative particles
simple tense = experiencing self; stativized verb = observing self ni ‘in/at’ + shite mo =even given [its] being X to ‘with/alongside’ + shite mo =even supposing [it] to be X
seems that strict adherence to an autonomy thesis would deprive one of seeing the richness and intriguing facts of language. Likewise, the soliloquy and dialogue distinction has never been a concern in formal analyses. The sensitivity of Japanese syntax to such a distinction assures us of the indispensability of such a notion in grammatical description. If the soliloquy and dialogue distinction is viewed as the difference between inside vs. outside the speaker’s world, then this distinction is found relevant to such well-known notions as self vs. others, or uchi ‘inside’ vs. soto ‘outside’ as seen in some representative studies below. As they permeate Japanese grammar so deeply, it also reminds us that these exist for a reason. Table 4 summarizes sample works which use corollaries of the uchi vs. soto notion and their applicability in Japanese grammar. McCawley (1978: 189–190) observes that the complementizer no is more natural when the speaker’s own mental state is at issue as in (24a), but the other factive complementizer koto becomes more natural when other’s mental state is at issue as in (24b). (24) a.
otoosan ga omae no koto o konnani shinpai shite Father your matter this much worried yatte-iru {no/ ?koto} ga mada omae ni wa wakara-n no ka? for you yet yo to realize-not Q ‘Can’t you see that I, your father, am this worried about you?!’
b. minasan gata ga omae no koto o annani shinpai shite everybody your matter that much worried
Subjectivity, intersubjectivity and Japanese grammar
kudasatte-iru {no / koto}17 ga mada omae ni wa for you yet you to wakara-n no ka? realize-not Q ‘Can’t you understand yet that other people are concerned that much about you?!’
She also notes that if a piece of information such as the earth is round which the speaker acquired from an external source is internalized, no is possible, but if it is not yet internalized into the speaker’s knowledge base as in Today, I learned that the earth is round, but I just can’t believe it, then koto is possible. Watanabe (1991) observes that degree adverbs can be classified depending on their affinity with my/ others’ affairs. For instance, zuibun ‘very’ cannot modify my affairs such as *watashi wa zuibun ureshii ‘I am very happy’, but it can modify a similar emotional state of other people as in Tanaka-san wa zuibun ureshisooda ‘It seems Tanaka is very happy.’ In contrast, taihen ‘very’, a synonym of zuibun can modify both my and others’ affairs. In a similar vein, Shinzato (2003) argues that simple tense forms are used to express an event the speaker is experiencing, while its stativised version is used for the event the speaker has observed. For instance, if the speaker himself is attacked, the simple present tense as in nani o suru nda [what-OBJ-do-it is that] ‘what are you doing (to me)?’ is used; in contrast, when he catches someone’s unexpected behavior, the stativised form as in nani o shite-iru nda [what-OBJ-do-be-it is that] ‘what are you doing (there)?’ is more likely to be used. Here again, Japanese grammar’s sensitivity to self vs. others is evident. This territorial distinction is also seen in Quinn’s (1994) account where the semantic contrast of uchi ‘inside’ vs. soto ‘outside’ seen in two locative particles, ni ‘in/at’ and to ‘with/along side’ is carried out in several constructions of which they are a part. To quote just one example, X ni shitemo ‘even given [its] being X’ presents presupposed, and fully internalized knowledge of the speaker, while it is not the case with X to shite mo ‘even supposing [it] to be X’. Seeing the correlation between synchrony, diachrony and speech modes, and realizing what underlies such intertwinedness are the notions of subjectivity and intersubjectivity, it would be impossible to deny the fundamentality of these notions in Japanese grammar, and perhaps in the grammar of any language. 17. One may ask as to the difference between no and koto here. Though it is not addressed specifically with this particular example, Akatsuka (ibid: 185) explains a similar case with the verb matsu ‘wait’. She states that with no, the state of affairs in the complement is taken for granted, thus, the speaker would be disappointed if it does not materialize. In contrast, with koto, the same verb matsu is more synonymous with kitaisuru ‘expect’. That is, Akatsuka (ibid: 181) sees the choice between no and koto to be the difference in the degree of the speaker’s commitment to the truth of the proposition: no indicates the speaker’s complete endorsement, while koto does not show such a strong endorsement.
Rumiko Shinzato
4. Conclusion Finegan (1995) deplores the fact that the term subjectivity did not make an entry of its own in International Encyclopedia of Linguistics, with only peripheral mention of it under literary pragmatics. As discussed in this paper, faced with such far-reaching applicability of this notion, it is clear that an attempt to construct grammar, especially Japanese grammar, with the incorporation of subjectivity and intersubjectivity broadens our understanding of the working of the Japanese language. In this sense, humanistic, or better yet, real(istic) linguistics is undoubtedly in order.18 Real(istic) linguistics sees “language not strictly as form nor as the expression of propositional thought, language not as autonomous structure not as representing logical propositions, but language as an expression – an incarnation, even – of perceiving, feeling, speaking subjects (ibid: 2)”. In this context, Beneveniste’s words carry special weight: ‘If LANGUAGE is, as they say, the instrument of communication, to what does it owe this property?’ Abbreviations COM COMP COND DAT EMPH INF INT
comitative complementizer conditional dative emphasis inference intention
KP NEG OBJ PERF SBJ SP TOP
kakari particle negative object perfect subject sentence final particle topic
References Akatsuka, Noriko. 1978. Another look at no, koto, and to: Epistemology and complementizer choice in Japanese. In Problems in Japanese Syntax and Semantics, John Hinds & Irwin Howard (eds), 178–212. Tokyo: Kaitakusha. Akatsuka, Noriko. 1979. Epistemology, Japanese syntax, and linguistic theory. Papers in Japanese Linguistics 6: 7–28. Akatsuka, Noriko. 1985. Conditionals and epistemic scale. Language 61: 625–639. DOI: 10.2307/414388
18. I owe this term to Tsuyoshi Ono.
Subjectivity, intersubjectivity and Japanese grammar Ariel, Mira. 1998. Mapping so-called ‘pragmatic’ phenomena according to ‘linguistic-extralinguistic’ distinction. In Functionalism and Formalism in Linguistics, Vol. II: Case Studies [Studies in Language Companion Series 42] Michael Darnell, Edith A. Moravcsik, Michael Noonan, Frederick J. Newmeyer & Kathleen Wheatley (eds), 11–38. Amsterdam: John Benjamins. Aronoff, Mark. 1994. Morphology by Itself. Cambridge MA: The MIT Press. Axelrod, Melissa. 1999. Lexis, grammar and grammatical change. In Functionalism and Formalism in Linguistics, Vol. II: Case Studies [Studies in Language Companion Series 42] Michael Darnell, Edith A. Moravcsik, Michael Noonan, Frederick J. Newmeyer & Kathleen Wheatley (eds), 39–58. Amsterdam: John Benjamins. Benveniste, Emile. 1958 [Tr. 1971]. Problems in General Linguistics, trans. by Mary Elizabeth Meek. Coral Gables FL: University of Miami Press Binnick, Robert I. 1991. Time and the Verb: A Guide to Tense and Aspect. Oxford: OUP. Bühler, Karl. 1934 [Tr. 1990]. Theory of Language: The Representational Function of Language, trans. by Donald Fraser Goodwin. Amsterdam: John Benjamins. Bybee, Joan L. & Pagliuca, William. 1985. Cross-linguistic comparison and the development of grammatical meaning. In Historical Semantics: Historical Word-Formation, Jacek Fisiak (ed.), 59–83. Berlin: Mouton de Gruyter. Bybee, Joan L., Perkins, Revere & Pagliuca, William. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago IL: The University of Chicago Press. Cain, Mark J. 2002. Fodor: Language, Mind and Philosophy [Key Contemporary Thinkers]. Oxford: Blackwell. Cinque, Guglielmo. 1999. Adverbs and Functional Heads: A Cross-linguistic Perspective [Oxford Studies in Comparative Syntax]. Oxford: OUP. Cook, Haruko Minegishi. 1989. The sentence-final particle ne as a tool for cooperation in Japanese conversation. Japanese/Korean Linguistics 1: 29–44. Duchan, Judith F., Bruder, Gail A. & Hewitt, Lynn E. 1995. Deixis in Narrative: A Cognitive Science Approach. Hillsdale NJ: Lawrence Erlbaum Associates. Evans, Vyvyan. 2004. The Structure of Time: Language, Meaning and Temporal Cognition [Human Cognitive Processing 12]. Amsterdam: John Benjamins. DOI: 10.1075/hcp.12 Finegan, Edward. 1995. Subjectivity and subjectivisation: An introduction. In Subjectivity and Subjectivisation: Linguistic perspectives, Stein Dieter & Susan Wright (eds), 1–15. Cambridge: CUP. DOI: 10.1017/CBO9780511554469.001 Fodor, Jerry. 1983. The Modularity of Mind. Cambridge MA: The MIT Press. Foley, William & Van Valin Jr., Robert D. 1984. Functional Syntax and Universal Grammar. Cambridge: CUP. Fujii, Seiko. 2000. Incipient decategorization of MONO and grammaticalization of speaker attitude in Japanese discourse. In Pragmatic Markers and Propositional Attitude [Pragmatics & Beyond New Series 79], Gisle Andersen & Thorstein Fretheim (eds), 85–118. Amsterdam: John Benjamins. Givόn, Talmy. 1982. Logic vs. pragmatics, with human language as the referee: Toward an empirically viable epistemology. Journal of Pragmatics 6: 81–113. DOI: 10.1016/03782166(82)90026-1 Givόn, Talmy. 1995. Functionalism and Grammar. Amsterdam: John Benjamins. DOI: 10.1075/ z.74 Haga, Yasushi. 1954. Chinjutsu to wa nanimono? Kokugo Kokubun 23: 47–61.
Rumiko Shinzato Hayashi, Makoto. 2003. Joint Utterance Construction in Japanese Conversation [Studies in Discourse and Grammar 12]. Amsterdam: John Benjamins. DOI: 10.1075/sidag.12 Hengeveld, Kees. 2005 [1989]. Layers and operators in functional grammar. In Crucial Readings in Functional Grammar, Matthew P. Anstey & J. Lachlan Mackenzie (eds), 1–40. Berlin: Mouton de Gruyter. DOI: 10.1515/9783110901191.1 Hopper, Paul J. 1997. Dispersed verbal predicates in vernacular written narrative. In Directions in Functional Linguistics [Studies in Language Companion Series 36], Akio Kamio (ed.), 1–18. Amsterdam: John Benjamins. Iwasaki, Shoichi. 1993. Subjectivity in Grammar and Discourse: Theoretical Considerations and a Case Study of Japanese Spoken Discourse [Studies in Discourse and Grammar 12]. Amsterdam: John Benjamins. Kamio, Akio. 1997. Territory of Information. Amsterdam: John Benjamins. DOI: 10.1075/ pbns.48 Kuno, Susumu. 1987. Functional Syntax: Anaphora, Discourse, Empathy. Chicago IL: University of Chicago Press. Kuroda, Sige-Yuki. 1973. Where epistemology, style and grammar meet: A case study from Japanese. In A Festschrift for Morris Halle, Stephen R. Anderson & Paul Kiparsky (eds), 377–391. New York NY: Rinehart and Winston. Langacker, Ronald. 1991. Foundations of Cognitive Grammar, Vol. II. Stanford CA: Stanford University Press. Laury, Ritva & Ono, Tsuyoshi. 2005. Data is data and model is model: You don’t discard the data that doesn’t fit your model! Language 81(1): 218–225. DOI: 10.1353/lan.2005.0026 Leech, Geoffrey N, 1983. Principles of Pragmatics. London: Longman. Lyons, John, 1977. Semantics, Vol. 2. Cambridge: CUP. Lyons, John. 1982. Deixis and subjectivity: Loquor, ergo sum? In Speech, Place, and Action, Robert J. Jarvella & Wolfgang Klein (eds), 101–124. New York NY: John Wiley & Sons. Lyons, John. 1995. Linguistics Semantics: An Introduction. Cambridge: CUP. DOI: 10.1017/ CBO9780511810213 McCawley, Noriko A. 1978. Epistemology and Japanese syntax: Complementizer choice. CLS 14: 272–284. Maynard, Senko K. 1993. Discourse: Modality: Subjectivity, Emotion and Voice in Japanese Language [Pragmatics & Beyond New Series 24]. Amsterdam: John Benjamins. DOI: 10.1075/ pbns.24 Maynard, Senko K. 2002. Linguistic Emotivity: Centrality of Place, the Topic-Comment Dynamic, and an Ideology of Pathos in Japanese Discourse [Pragmatics & Beyond New Series 97], Amsterdam: John Benjamins. DOI: 10.1075/pbns.97 Minami, Fujio. 1974. Gendai Nihongo no Koozoo Tokyo: Taishukan Shoten. Mizutani, Osamu. 1981. Japanese: The Spoken Language in Japanese Life. Tokyo: Japan Times. Morishige Satoshi. 1971. Nihon Bumpoo no Sho-mondai. Tokyo: Kasama Shoin. Moriyama, Takuro. 1997. ‘Hitorigoto’ o megutte. Nihongo Bumpoo: Taikei to Hoohoo, Yoshiaki Kawabata & Yoshio Nitta (eds), 173–188. Tokyo: Hituzi Shobo. Mushin, Ilana. 2001. Evidentiality and Epistemological Stance Narrative Retelling [Pragmatics & Beyond New Series 97]. Amsterdam: John Benjamins. DOI: 10.1075/pbns.87 Nakau, Minoru. 1994. Principles of Cognitive Semantics. Tokyo: Taishukan Shoten. Nakau, Minoru. 1979. Modaritii to meidai. Eigo to Nihongo to – Hayashi Eiichi Kyooju Kanreki Kinen Ronbunshuu, Hayashi Eiichi Kyooju Kanreki Kinen Ronbunshuu Kankoo Iinkai (eds), 223–250. Tokyo: Kuroshio Shuppan.
Subjectivity, intersubjectivity and Japanese grammar Narrog, Heiko. 2002. Imironteki kategorii toshite no modaritii. Cognitive Linguistics II: Categorization, Toshio Ohori (ed.), 217–251. Tokyo: University of Tokyo Press. Narrog, Heiko. 2005. On defining modality again. Language Sciences 27(2): 165–192. DOI: 10.1016/j.langsci.2003.11.007 Nitta, Yoshio. 1991. Ishi no hyoogen to kikite sonzai. Kokugogaku 165: 1–13. Nitta, Yoshio. 1992. Handan kara hatsuwa dentatsu e – Denbun enkyoku hyoogen o chuushin ni -. Nihongo Kyoiku 77: 1–13. Nuyts, Jan. 2001. Epistemic Modality, Language, and Conceptualization [Human Cognitive Processing 5]. Amsterdam: John Benjamins. DOI: 10.1075/hcp.5 Ono, Susumu. 1993. Kakari Musubi no Kenkyuu. Tokyo: Iwanami Shoten. Onodera, Noriko O. 2004. Japanese Discourse Markers. Synchronic and Diachronic Discourse Markers [Pragmatics & Beyond New Series 132]. Amsterdam: John Benjamins. DOI: 10.1075/pbns.132 Onoe, Keisuke. 2001[1990]. Bunpoo-ron – Chinjutsuron no tanjoo to shuuen. Bunpoo to Imi, 278–300. Tokyo: Kurohio Shuppan. Onoe, Keisuke. 2001[1998]. Ichigo-bun no yoohoo: ‘Ima/koko’ o hanarenai bun no kentoo no tame ni. Bunpoo to Imi, 217–229. Tokyo: Kurohio Shuppan. Prevignano, Carlo L. & Thibault, Paul J. 2003. Discussing Conversation Analysis: The Work of Emanuel A. Schegloff. Amsterdam: John Benjamins. DOI: 10.1075/z.118 Prinz, Jessee J. 2006. Is the mind really modular? In Contemporary Debates in Cognitive Science, Robert J. Stainton (ed.), 22–36. Oxford: Blackwell. Quinn, Charles J. 1994. UCHI/SOTO: Tip of a semiotic iceberg? ‘Inside’ and ‘outside’ knowledge in the grammar of Japanese. In Situated Meaning: Inside and Outside in Japanese Self, Society, and Language, Jane Bachnik & Charles Quinn Jr. (eds), 247–294. Princeton NJ: Princeton University Press. Rijkhoff, Jan. 2002. The Noun Phrase. Oxford: OUP. DOI: 10.1093/acprof:oso/9780198237822.0 01.0001 Rijkhoff, Jan. 2005. Layers, levels and contexts in Functional Discourse Grammar. In The Noun Phrases in Functional Discourse Grammar, Daniel García Velasco & Jan Rijkhoff (eds), 63– 115. Berlin: Mouton de Gruyter. Romaine, Suzanne & Lange, Deborah. 1991. The use of like as a marker of reported speech and thought: A case of grammaticalization in progress. American Speech 66: 227–279. DOI: 10.2307/455799 Saji, Keizo. 1957. Shuujoshi no kinoo. Kokugo Kokubun 26: 23–31. Sakakura, Atsuyoshi. 1993. Nihongo Hyoogen no Nagare. [Iwanami Seminaa Bukkusu 45]. Tokyo: Iwanami Shoten. Selting, Margret & Couper-Kuhlen, Elizabeth. 2001. Studies in Interactional Linguistics [Studies in Discourse and Grammar 10]. Amsterdam: John Benjamins. DOI: 10.1075/sidag.10 Serafim, Leon A. & Shinzato, Rumiko. 2000. Reconstructing the Proto-Japonic kakari musubi, *…ka…-(a)m-wo. Gengo Kenkyu 118: 81–118. Shinzato, Rumiko. 2003. Experiencing self vs. observing self: The semantics of stative extensions in Japanese. Language Sciences 25: 211–238. DOI: 10.1016/S0388-0001(01)00025-0 Shinzato, Rumiko. 2004. Some observations concerning mental and speech act verbs. Journal of Pragmatics 36(5): 883–897. DOI: 10.1016/S0378-2166(03)00002-X Shinzato, Rumiko. 2007. (Inter)subjectification, Japanese syntax and scope increase. Journal of Historical Pragmatics 8(2): 171–206. DOI: 10.1075/jhp.8.2.03shi
Rumiko Shinzato Shinzato, Rumiko & Serafim, Leon A. 2003. Kakari musubi in comparative perspective: Old Japanese ka/ya and Okinawan -ga/-i. Japanese/Korean Linguistics 11: 189–202. Shinzato, Rumiko & Serafim, Leon A. 2013. Synchrony and Diachrony of Okinawan Kakarimusubi in Comparative Perspective with Premodern Japanese. Leiden: Brill. Strauss, Susan & Sohn, Sung-ock. 1998. Grammaticalization, aspect, and emotion: The case of Japanese -te shimau and Korean -ia/e pelita. Japanese/Korean Linguistics 8: 217–230. Suzuki, Ryoko. 1998. From a lexical noun to an utterance-final pragmatic particles: Wake. In Studies in Japanese Grammaticalization, Toshio Ohori (ed.), 67–92. Tokyo: Kuroshio Publishers. Suzuki, Satoko (ed.). 2006. Emotive Communication in Japanese [Pragmatics & Beyond New Series 151]. Amsterdam: John Benjamins. DOI: 10.1075/pbns.151 Tagliamonte, Sali & Hudson, Rachel. 1999. Be like et al. beyond America: The quotative system in British and Canadian youth. Journal of Sociolinguistics 3(2): 147–172. DOI: 10.1111/14679481.00070 Tanabe, Masao. 1953. Iwayuru atsurae no joshi ‘namu’ nit suite. In Kindaichi Hakase Koki Kinen Gengo Minzoku Ronsoo, Masayo Tanabe (ed.) 477–504. Tokyo: Sanseido. Tokieda, Motoki. 1941. Kokugogaku Genron. Tokyo: Iwanami Shoten. Traugott, Elizabeth Closs. 2003. From subjectification to intersubjectification. In Motives for Language Change, Raymond Hickey (ed.), 124–139. Cambridge: CUP. Traugott, Elizabeth Closs & Dasher, Richard. 1987. On the historical relation between mental and speech act verbs in English and Japanese. In Papers from the Seventh International Conference on Historical Linguistics [Current Issues in Linguistic Theory 48], Anna Giacalone Ramat, Onofrio Carruba & Giuliano Bernini (eds), 561–573. Amsterdam: John Benjamins. Traugott, Elizabeth Closs & Dasher, Richard. 2002. Regularity in Semantic Change. Cambridge: CUP. Tsuchii, Shoichi & Omori, Akira. 2000. Aizuchi o toosei-shita komyunikeeshon ni okeru joshi ne no hindo no henka. Ninchi Kagaku 7(1): 107–111. Verhagen, Arie. 2005. Constructions of Intersubjectivity. Oxford: OUP. Watamaki, Toru. 1997. Lack of the particle-ne in conversation by a child with autism: A case study. Hattatsu ShogaiKenkyu 19: 146–157. Watanabe, Minoru. 1953. Jojutsu to chinjutsu - Jutsugo bunsetsu no koozoo. Kokugogaku 13–14: 20–34. Watanabe, Minoru. 1991. ‘Wagakoto-hitogoto’ no kanten to bunpooron. Kokugogaku 165: 1–14. Watanabe, Minoru. 1995. Tokoro to toki no shitei ni kakawaru go no ikutsuka – Imiron-teki ni. Kokugogaku 181: 18–19. Yamaguchi, Yoshinori. 1985. Kodai Nihongo Bumpoo no Seiritsu no Kenkyuu. Tokyo: Yuseido.
What typology reveals about modality in Japanese A cross-linguistic perspective* Kaoru Horie and Heiko Narrog This paper presents a functional-typological analysis of three linguistic manifestations of modality-related phenomena in Japanese. When compared with English, German, and Korean, Japanese is characterized by a modal system that encodes event, epistemic, and evidential modalities, and a relatively impoverished grammatical mood, as well as a rich discourse system of sentencefinal particles that can be grouped into speaker-oriented and hearer-oriented particles. The modality system in Japanese demonstrates a relatively high degree of elaboration in formal coding of evidential and discourse modalities. The cross-linguistic differences in the degrees of elaboration among different subcategories of modality as presented in this study require an explanation beyond the confines of grammar to find a link between grammar and other cognitive and communicative systems.
1. Introduction Tense, aspect and modality, collectively referred to as TAM, are grammatical categories frequently occurring together close to the verb stem and affecting the meaning of the co-occurring verb to varying degrees. There is a fundamental difference between tense and aspect, on one hand, and modality, on the other: “Tense, rather obviously, is concerned with the time of the event, while aspect is concerned with the nature of the event, particularly in terms of its ‘internal temporal constituency’ (Comrie 1976: 3). (...) Modality differs from tense and aspect in that it * This research was supported in part by grants from the Tohoku University 21st Century Program in Humanities . Special thanks are due to Yoshi Ono and Kaori Kabata for organizing the Functional Approaches to Japanese Grammar conference (U. of Alberta, August 2004) and offering extensive constructive criticism and helpful comments. Thanks also go to Andrew Barke, Robin Coogan, Nathan Hamlitsch, Ahran Kim, Sujin oh and Ryan Spring, for their editorial assistance.
Kaoru Horie and Heiko Narrog
does not refer directly to any characteristic of the event, but simply to the status of the proposition.” Modality is a semantic category “concerned with the status of the proposition that describes the event”, the term “proposition” covering “events, actions, situations, states, etc.” (Palmer 2001: 1). The semantic category of modality is manifested grammatically in two major ways: (i) “modal systems” most typically illustrated by modal verbs, and (ii) “mood” most typically illustrated by indicative vs. subjunctive moods (Palmer 2001). Bybee’s seminal work on verbal affix ordering (Bybee 1985) presents a similar observation that, among the three categories, aspect and tense are more relevant to (or more directly affect) the meaning of the verb than mood, the latter broadly corresponding to (ii) above. Crucially, this relative ordering in terms of semantic relevance is reflected in the ordering of aspect and tense affixes, which tend to be placed more closely to the verb stem than mood affixes. The fundamental difference between modality and the other two verbal categories is also reflected in the degree of ubiquity of their grammaticalization pathways, as pointed out by Horie (1997). Horie (1997), a critique of Bybee, Perkins, and Pagliuca (1994) noted that grammaticalization pathways proposed for Tense and Aspect grammatical morphemes (henceforth ‘grams’; see Bybee 1985) are more readily applicable to Japanese (e.g., “be/have > Resultative > Perfective/Simple Past”) than those proposed for Modality grams (e.g. “Agent-oriented Modalities > Epistemic Modalities”). The greater cross-linguistic variability of modality grams can arguably be attributed to the fact that they index the speaker’s opinion/ attitude toward the proposition, which can be expressed in more variable ways cross-linguistically than temporal specifications of the event. This observation is in fact echoed by Palmer (2001: 2): “In all typological studies there is considerable variation in the ways in which languages deal with grammatical categories, and there is probably more variation with modality than with other categories” (emphasis added). Modality is thus coded in a more varied way than are other verbal categories like tense, aspect, and voice, as the following remark elucidates: “Languages have at least the following means at their disposal to express modality: (a) modal verbs, (b)verbs denoting (various modes of) knowledge and belief, (c) modal adverbs, (d) modal particles, (e) evidentials, (f) grammatical mood.” (Kiefer 1999: 223). The greater cross-linguistic variability of modality grams is arguably responsible for the paucity of typologically oriented studies of modality in Japanese. As a result, despite intensive scholarly efforts engendered on the minute (and meticulous) analyses of Japanese modality grams by Japanese linguists for the past twenty or so years, it remains to be seen precisely what distinguishes Japanese modality systems from their counterparts in other languages and what is shared between them.
What typology reveals about modality in Japanese
By contrasting the mood and modality system of Japanese with those of other languages in Sections 2 and 3, which differ from Japanese typologically to varying degrees, we address the following questions in Section 4: (i) what is unique to the Japanese mood and modality system relative to other languages, and (ii) what motivates the grammar of a language to have the shape it does. Section 5 presents our conclusion. 2. Revisiting the semantic and formal categories of modality in Japanese: A cross-linguistic assessment This section first presents an overview of semantic categories of modality and their formal coding in linguistic typology, and then examines the language-particular observations of Japanese modality, presented in Japanese linguistics from a typological perspective. 2.1 Modality and its formal coding in linguistic typology Considering the greater cross-linguistic variability of modality relative to other grammatical categories (e.g., tense) pointed out in Section 1, it is natural that languages differ rather considerably in terms of the types of modal meaning a language preferentially elects to encode grammatically. As for the types of modal meaning encoded in typologically diverse languages, Palmer (2001), in a most widely cited typological study of mood and modality, first draws attention to the relevance of the following semantic distinction:
(1) realis and irrealis
Regarding this distinction, Palmer (2001. 1) cites the definition of Mithun (1999: 173): The realis portrays situations as actualized, as having occurred, or actually occurring, knowable through direct perception. The irrealis portrays situations as purely within the realm of thought, knowable only through imagination.
This distinction, though not directly relevant to Japanese, is known to have grammaticalized in many European languages typically by means of a contrast between ‘indicative’ and ‘subjunctive’ mood, as in the following pair of English examples: (2) a. They insisted that John was there. (indicative: “realis”) b. They insisted that John be there. (subjunctive: “irrealis”)
Kaoru Horie and Heiko Narrog
Another cross-linguistically relevant semantic distinction noted by Palmer is the following one:
(3) propositional modality event modality
This distinction partially overlaps with the traditional distinction between ‘epistemic’ and ‘deontic’ modalities, though the former is more inclusive than the latter. The distinction between propositional modality and epistemic modality is illustrated by the following pair of examples (ibid: 7, partially modified). (4) a. Kate may be at home now. ~ It is possible (possibly the case) that Kate is at home now. (propositional modality) b. Kate may come in now. ~ It is permitted for Kate to come in now. (event modality) (4a) describes “the speaker’s judgment of the proposition that Kate is at home”. (4b), in contrast, depicts “the speaker’s attitude towards a potential future event, that of Kate coming in” (ibid: 7–8). Both propositional modality and event modality are further divided into subcategories, but we will be concerned only with the following subcategories of propositional modality in this paper: (5) epistemic modality evidential modality Epistemic modality encodes speakers’ “judgments about the factual status of the proposition”, whereas evidential modality expresses “the evidence they have for its factual status” (ibid: 8). These two types of modality are respectively illustrated by the following pair of Japanese examples: (6) a. [Asu yuki-ga huru ] kamo sirenai. tomorrow snow-nom fall it may be ‘It may snow tomorrow’. (epistemic modality)
b. [Asu yuki-ga huru] soo da. tomorrow snow-nom they say ‘I hear it will snow tomorrow’. (evidential modality) In this paper, we will examine the grammatical means of encoding the following three types of modality: (7) a. event modality b. epistemic modality c. evidential modality
What typology reveals about modality in Japanese
Let us now move on to the formal/coding dimension of modal meaning from a typological perspective. Palmer notes that there are basically two ways in which the semantic category of modality is grammatically encoded, as in (8): (8) a. modal system b. mood ‘Modal system’ and ‘mood’ are formally coded by one of the following three grammatical markers according to Palmer (p.19), who notes that modal verbs seem to be confined to modal systems: (9) a. modal verb b. inflection c. individual suffixes, clitics, and particles Though there is not a direct correspondence between modal system/mood in (8) and their formal manifestations (9), there is a strong tendency for ‘modal system’ (8a) and ‘mood’ (8b) to be coded by ‘modal verb’ (9a) and ‘inflection’ (9b) respectively. For instance, modal system (8a) is illustrated by modal auxiliaries (‘modal verb’) in English (4) and Japanese (6), while mood (8b) is illustrated by indicative and subjunctive moods in English (2). ‘Individual suffixes, clitics, and particles’ (9c) can encode various modal meanings including discursive meaning and manifest ‘discourse system’, which we will introduce in this section. The relative degrees of prominence between the two formal means can vary between languages as observed by Palmer: Both may occur within a single language, e.g., in German, which has a modal system of modal verbs and mood (indicative and subjunctive), and in Central Pomo (...). In most languages, however, only one of these devices seems to occur, or at least, one is much more salient than the other. (p.4, emphasis added)
As we will see closely in Section 3, Japanese belongs to a group of languages where mood is virtually absent. 2.2 Modality and its formal coding in Japanese Having reviewed Palmer’s typological survey relating to the semantic and formal/ coding dimensions of modality, we are now ready to examine the language-particular notions of modality in Japanese linguistics from a typological perspective. In Japanese linguistics, it has been customary to divide sentence structure into two major semantic components, i.e. proposition (variously referred to as ‘meidai’ (proposition), ‘genpyoo zitai’ (the state of affairs expressed) etc.) and modality
Kaoru Horie and Heiko Narrog
(variously referred to as ‘muudo’ (mood), ‘genpyoo taido’ (the attitude with which to express the state of affairs) etc.), as schematized and explained in (10): (10) [Proposition] [Modality] In modern Japanese, most researchers have come to an agreement that Japanese sentences can be divided into a propositional content and modal content (...) Propositional content expresses an objective statement while modal content expresses a speaker’s subjective judgment or attitude toward the proposition. (Johnson 2003, emphasis added)
A cursory illustration of the distinction between ‘proposition’ and ‘modality’ is provided by example (11) from Nitta (2000: 81), where ‘P’ and ‘M’ respectively refer to the layers of proposition and modality. (11) [M Nee tabun [P kono ame toobun yama-nai] daroo ne] ip probably this rain for a while stop-neg it will be that ip “You see, probably it won’t stop raining for a while, will it?” (Nitta 2000: 81) Some Japanese linguists, including Yoshio Nitta and Takashi Masuoka (Nitta 1991, 2000; Masuoka 1991, 2000) propose further dividing the layer of modality into two subcategories, i.e. ‘modality oriented toward proposition’ (“meidai meate no modaritii”) and ‘modality of discourse and communication” (“hatuwa, dentatu no modaritii”) (see Onoe 1996 for a critique of this kind of extended view of modality). In a similar vein, Maynard (1993: 38–39) introduces the notion of “discourse modality”, which refers to “information that does not or only minimally conveys objective propositional message. (...) Discourse Modality operates to define and to foreground certain ways of interpreting the propositional content in discourse”. It is fair to say that Maynard’s ‘discourse modality’ is parallel to Nitta and Masuoka’s ‘modality of discourse and communication’. We will employ the term ‘discourse modality (DM)’ to distinguish it from ‘modality’, or ‘modality oriented toward proposition’. This subdivision can be illustrated as in (11’): (11’) [DM Nee [M tabun [P kono ame toobun yama-nai] daroo] ne] “You see, probably it won’t stop raining for a while, will it?” The differing degrees of relevance of modality categories to the propositional content of the sentence are iconically reflected in their differential positions, i.e. ‘modality’ being situated closer to Proposition than ‘discourse modality’. In Japanese Linguistics, two distinct types of modal meaning have been identified and respectively referred to as ‘modality (oriented toward proposition)’ and ‘discourse modality’. These descriptive labels seem to make sense when we consider Japanese alone (see (11’)). However, when we contrast modality categories in
What typology reveals about modality in Japanese
Japanese with those in other languages, it is necessary to examine how such labels are compatible with the typologically established terminology such as that employed in Palmer (1986, 2001) and Bybee et al. (1994). It is also very important to examine what kind of grammatical means a language has available to encode and allocate modal meaning in its sentence structure. In Japanese linguistics, much attention has been paid to the semantic aspects of modality (see Narrog 2002a, 2005, 2009a, b), but relatively little attention seems to have been paid to its formal/ coding dimension. The notion of ‘modality’ (modality oriented toward proposition) in Japanese linguistics doesn’t translate straightforwardly into the typologically-oriented terminology introduced in Section 2.1, but seems to encompass both propositional (epistemic and evidential) and event modalities. In terms of formal coding, ‘modality’ in Japanese is manifested as a modal system of modal verbs (auxiliaries) as we will see in Section 3. The notion of ‘modality’ in Japanese is thus not wholly incompatible with the typologically oriented notions of modality. The notion of ‘discourse modality’, in contrast, appears to be less compatible with the typologically established terminology of modality. This arguably reflects the fact that, unlike mood and prototypical modal systems, grammaticalized coding of a speaker’s discursive stance relative to the discursive content and to her/his addressee, or discourse modality, is less prominent cross-linguistically. Palmer (1986: 58) notes the existence of a ‘discourse system’ as a (rather minor) subcategory of modal systems and offers the following observation:1 Modals have an important part to play in discourse, as the participants express their opinions and attitudes, and, in general, interact with one another. It is not wholly surprising, therefore, that there are systems which are more directly concerned with discourse relations.
Though not cross-linguistically prominent, the existence of a ‘discourse system’ in Japanese is undisputable, as we will see in detail in Section 3. We therefore suggest that when we contrast Japanese with other languages, ‘discourse system’ be included among the inventory of grammatical means to encode modality, on par with modal system and mood as in (8’), instead of being included within modal 1. A typical instance of discourse system cited by Palmer (1986: 61) is the set of sentence-final particles in Mandarin Chinese (based on Li & Thompson 1981: 238ff) illustrated in (a): (a)
le ne ba ou a/ya ma
currently relevant state response to question solicit agreement friendly warning reduce forcefulness question
Kaoru Horie and Heiko Narrog
systems as it is under Palmer’s classification, when we contrast Japanese with other languages. (8’) a. modal system b. mood c. discourse system 3. Modality and its formal coding in Japanese, Korean, English, and German This section provides a comparison of modality grams in Japanese and its counterparts in Korean, English, and German. As with other grammatical phenomena (e.g., tense, aspect, voice), modality in Japanese has been contrasted primarily with its English counterpart (e.g., Sawada 1995) without paying due attention to the typological variability of the grammatical category in question. We believe that by contrasting modality in Japanese with its counterparts in languages of differing typological profiles, such as Korean, English, and German, we can have a more realistic and ‘relativized’ picture of modality in Japanese. The four languages selected represent two sets of East Asian and European languages and they each are known to exhibit intriguing morpho-syntactic and semantic differences (English-German and Japanese-Korean). English and German have been contrasted extensively by Hawkins (1986) from the perspective of Comparative Typology, a theoretical framework designed to reveal cross- linguistic variation in form-meaning mapping based on various lexical and morpho-syntactic contrasts. Similar contrasts in form-meaning mapping have been pointed out with another pair of languages, Japanese and Korean, in a series of works by the first author of this paper (Horie 1998a, b, 2001, 2002a, b, 2003a, b, Horie & Sassa 2000). In the domain of mood and modality, Horie has made attempts to present a comparative study of Japanese and Korean (Horie & Taira 2002, Horie 2003b), but it has become increasingly obvious that a further crosslinguistic comparison beyond these two East Asian languages is needed in order to provide a better-balanced relativized picture of modality in Japanese. Building on the previous Japanese-Korean contrastive studies of modality (Horie & Taira 2002, Horie 2003b), we will extend our inquiry into English and German in the remainder of this section. 3.1
Modal systems in Japanese, Korean, English, and German
Japanese has a rather elaborate modal system of suffixes (e.g. strong inferential daroo ‘(I) predict’), so-called ‘formal nouns’ (grammaticalized nouns) plus copulas
What typology reveals about modality in Japanese
(e.g. evidential yoo-da (appearance-be) ‘it appears’), and other syntactic constructions (e.g. weak possibility kamo sirenai ‘it may be that’) that encode event, epistemic, and evidential modalities: (10) Asu-wa hayaku kaet-te mo ii desu yo. tomorrow-top early go home-may cop:pol ip ‘You may go home early tomorrow’. (event modality) (11) a. [Asu yuki-ga huru] kamo sirenai. tomorrow snow-nom fall it may be ‘It may snow tomorrow’. (epistemic modality) b. [Asu yuki-ga huru] soo da. they say ‘I hear it will snow tomorrow’. (evidential modality) We can see that the polysemy in terms of event-epistemic modality (e.g., two primary senses of may in English), which is prevalent in European languages, is absent in Japanese. It must be noted, however, that the event-epistemic polysemy was observable in Classical Japanese, as was the case with mu (encoding a speaker’s volition and prediction/conjecture) and besi (encoding both event and epistemic modality senses of should) (cf. Narrog 2002b). A similar allocation of modality meanings by means of suffixes, formal nouns with copulas, and other syntactic constructions, is observable in Korean (Korean examples are from Horie 2003b unless otherwise noted): (12) a. Yeki-se tampay-lul phiwe-to toy-pnikka? this place-loc cigarette-acc smoke-even good-pol:q ‘May I smoke here?’ (event modality) b. Ce-nun naynyen-ey hankwuk-ey kal ci to molu-pnita. I-top next year-loc Korea-to go:adn:fut may-pol:ind ‘I may go to Korea next year’. (epistemic modality) c. Nwu-ka wa-ss na pota. someone-nom come-past seem:ind ‘I think/It seems that someone is here’.(evidential modality) (Martin 1992: 705) It is important to note, however, that Korean behaves differently from Japanese in terms of the grammatical coding of modal meanings. First, unlike Japanese, Korean has a verbal suffix that encodes both event and epistemic modalities, i.e. -keyss-:
Kaoru Horie and Heiko Narrog
(13) a.
Na ton an pat-keyss-tako yayki-n ha-n I money neg receive-keyss-quot story-top say-adn:past cek eps-e. occasion not exist-ind ‘I have never said that I wouldn’t receive money’. (event modality)
(b) Nakksi-yo? Eme, nemwu caymiiss-keyss-ta... fishing-pol ah very interesting-keyss-ind ‘(Are you) fishing? Ah, it looks very interesting’. (epistemic modality) Secondly, Korean has a suffix -te- (realized alternately as -ti- or -t-) that encodes “retrospective” evidential modality, or “the speaker’s past perception, observation, or experience in declaratives and the hearer’s in interrogatives” (Sohn 1994: 47), as in (14), where -tela encodes the speaker’s recollection of a past experience: (14)
Pangkum cen-ey kasstaw-ass-nuntey uysik-i just now before-in go and visit-past-conj consciousness-nom eps-usi-tela. not exist-hon-retro:ind ‘I visited (the patient) just a while ago, but (the patient) was unconscious’.
This kind of ‘retrospective’ modality is not grammatically coded in Japanese modal systems, again pointing to a difference in the organization of the modal systems between Korean and Japanese. It should be noted, however, that a similar modal meaning was expressible in Old Japanese by means of a modal suffix keri. English and German behave similarly in terms of the shared coding of event and epistemic modalities, a commonly observed phenomenon within and beyond European languages: (15) a. You must leave now. (event modality) b. He must be tired. (epistemic modality) (16) a. Dieses Problem muss gelöst werden this problem mod solved become ‘This problem must be solved’. (event modality) b. Hans muss jetzt schön in Edmonton angekommen sein. Hans mod now already in Edmonton arrived be ‘Hans must have already arrived in Edmonton by now’. (epistemic) There is a crucial difference between English and German modal systems in terms of evidential modality. Unlike English, German has a set of modal auxiliaries that can encode evidential meaning, as in (17):
What typology reveals about modality in Japanese
(17) Hans soll jetzt schon in Edmonton sein. Hans mod now already in Edmonton be. ‘(According to some third person) Hans is already in Edmonton now’. (evidential modality) A contrastive summary of modal systems in the four languages is presented in Table 1. From Table 1, we can see that the modal system in Japanese, not surprisingly, is more similar to that in Korean than to the two other languages in terms of its morpho-syntactic inventories. Specifically, both Japanese and Korean modal systems are composed of periphrastic constructions and more synthetic suffixes, while English and German modal systems are almost exclusively composed of auxiliaries. However, when we take a closer look at the composite semantic categories of modality in the four languages, we get a more relativistic, continuous picture of cross-linguistic variation, as illustrated in Table 2: Table 1. Modal systems in Japanese, Korean, German, and English Japanese
Korean
German
English
Suffixes (-da roo, beki-da, -soo-da, -mitai-da, -rasii, -tai, -(r) (ar)eru, -uru, etc.) Formal noun (yoo-da, mono-da, etc.) Syntactic construction (-ka-mo sire nai, nakereba naranai, -temo ii, -te hosii, etc.)
Suffixes (-te-,-ti-, -tey- etc. (retro-spective), -keyss- (speaker’s intent, conjecture), -ess- (irrealis: conditonal clause), etc.) Formal noun (-kes kathta, -swu issta etc.) Syntactic construction (-ci moluta, -to toyta, ...) Verbal suffixes and periphrasis systematically present; synthetic nature of suffix more manifested than in Japanese (e.g., –keyss-)
Modal auxiliaries (müssen, dürfen, können, wollen, sollen, möchten)
Modal auxiliaries (must, can, may, should, will, shall, ought to, etc.)
Semi-modal auxiliaries (brauchen, scheinen) Auxiliaries (ist zu, haben zu)
Semi-modal auxiliaries (seem to, want to, etc.) Auxiliaries (is to, has to)
Modal auxiliaries and semi-modals systematically present
Modal auxiliaries and semi-modals systematically present
Verbal suffixes and periphrasis systematically present
Modal systems elaborately developed and playing more prominent roles than grammatical mood in all four languages.
Kaoru Horie and Heiko Narrog
Table 2. Semantic categories of modality and their formal manifestations contrasted Japanese
Korean
nakereba naranai, -te-ya hata/toyta, mo ii,-beki-da, -(y) -to toyta/cohta, oo, -tai, etc. -keyss, -ko siphta, etc. Epistemic -daroo, -keyss Modality -ka-mo sirenai, -nun/n/l -ni tigai nai, -hazu ci moluta da etc. -nun/n/l tus hata adverbs adverbs Evidential -yoo-da, -te-, -tey -mitai-da, rasii, -na pota, -nun/n/l -soo-da moyang-ita Event Modality
German
English
müssen, können, must, can, should, dürfen, brauchen, wol-may, ought to, shall, len, möchten, sollen, will, have to haben zu müssen, können, must, can, should, dürfen, mögen may, ought to
conditional adverbs adverbs sollen, wollen, Virtually (möchten) absent
Evidential modality is highly grammaticalized in Japanese and Korean, while it is absent in English, with German situated in between. German and English systematically exhibit polysemy in event and epistemic modalities, while Korean does so only sparingly and Japanese virtually lacks such polysemy.
Table 2 demonstrates that both German and English exhibit event-epistemic modal polysemy, i.e., a single auxiliary encoding both modalities (e.g., English must, German müssen), a phenomena virtually absent in Japanese and Korean except for Korean -keyss that encodes both modalities. It must be noted that Old Japanese exhibited such polysemy abundantly, e.g., mu, besi (cf. Horie 1997, Narrog 2002b). Another interesting observation is that, as far as evidential modality is concerned, English is virtually lacking in this category in striking contrast to Japanese and Korean. The latter two languages have a rich system of evidentials, with some interesting cross-linguistic differences observed such as the absence versus presence of the ‘retrospective’ evidential modality encoding. German is an interesting case in point in that it has hearsay evidential auxiliaries like sollen unlike English. 3.2 Mood in Japanese, Korean, German, and English Mood, or verbal inflection whose primary function is to encode the distinction between realis and irrealis, is not a prominent grammatical feature in Japanese. ‘Unmarked’ conclusive (sentence-final) verb forms in Japanese, which primarily encode non-past or past tense, do not overtly encode ‘indicative mood’, as in (18): (18) a. Gohan-o mainiti taberu. rice-acc every day eat:nonpast ‘I eat rice every day’.
What typology reveals about modality in Japanese
b. Gohan-o mainiti tabeta. eat:past ‘I ate rice every day’. Overt marking of modal meaning by inflectional forms is thus limited to the conclusive verb forms marking ‘propositive’ and ‘imperative’ mood, as in (19): (19) a. Gohan-o tabe-yoo. eat-propos ‘Let’s eat’. b. Gohan-o tabe-ro. eat-imp ‘Eat!’ Similarly to Japanese, Korean has overt conclusive verbal forms encoding ‘propositive’ and ‘imperative’ mood, as in (20): (20) a. Ka-ca. go-propos ‘Let’s go’. (plain style) b. Ka-kela (plain style, used by elderly (Ahran Kim, p.c.)) ‘Go !’ Unlike Japanese, however, Korean has conclusive verbal forms that encode ‘indicative mood’ and ‘interrogative’ mood, as in (21): (21) a. Cikum pakkey pi-ka manhi o-n-ta. now outside rain-nom much come-pres-ind ‘It is raining heavily outside.’ b. Cikum pakkey pi-ka manhi o-nunya? now outside rain-nom much come-Q ‘Is it raining heavily now outside?’ Furthermore, Korean has an adnominal verbal ending –(u)l that encodes ‘irrealis’ (future/probability), as in (22): (22) ka-l salam go-adn:fut person ‘the person who will/is supposed to go’ In German, compared to English, mood is still alive and present, relatively speaking. It has two types of subjunctive mood, i.e. Subjunctive I and Subjunctive II, to mark subordinate clause, irrealis, and quotation:
Kaoru Horie and Heiko Narrog
(23) a. Sie sagte, sie habe jetzt genug von allem. she said she has.subji now enough of all ‘She said that she had enough of all that’. (subjunctive I marking indirect speech) b. Mit etwas mehr Glück hätte ich hier gewonnen. with a little more luck have.subjii I here won ‘With a little more luck I would have won here’. (subjunctive II as irrealis) In English, in contrast, mood is clearly dying out. Subjunctive verb forms have survived only as the ‘marked’ forms of were in conditional clauses and as citation forms in complement clauses to a limited number of directive verbs like suggest, as in (24): (24) a. If you were here, we would be very happy. b. The director suggested that all the staff members be punctual. A contrastive summary of mood in the four languages is presented in Table 3. Table 3. Mood (Verbal inflection) Japanese
Korean
German
English
Mood encoded by conclusive form V + (r)u (unmarked)
Mood encoded by conclusive form V + ta, -e, -ney, etc. (indicative) V + ca, -sey, etc. (propositive) V + nya, -ni, etc. (interrogative) V + (e)la, -key, etc. (imperative)
Indicative (unmarked)
Indicative (unmarked)
Mood encoded by attributive form: -(u)l (irrealis)
+e/ würden + V
V + ta (unmarked) V + (y)oo (propositive) V + e/ro (imperative)
Imperative (unmarked) Indicative (unmarked) Subjunctive: Subjunctive I: V+e Subjunctive II: V (Prt/umlaut)
Function: Marking of subordinate clauses, irrealis, quotation Verbal inflection Verbal inflection cod-Indo-European mood partially coding ing modal systems partially modal meaning, but not meaning more preserved represent- ing gramovertly than matical mood Japanese as a whole
(Subjunctive: citation form in a limited number of complement clauses to directive verb, suppletive form were in conditional clauses
Indo-European mood systems virtually disappearing
What typology reveals about modality in Japanese
Table 3 indicates that grammatical mood is still alive and well in German, maintaining an indicative-subjunctive distinction, while it has virtually ceased to function in English barring a very limited number of syntactic environments (e.g. complements to a limited number of directive verbs like advise). More controversial are the statuses of mood in Japanese and Korean, i.e. whether it is proper to consider verbal suffixes encoding modal meanings in these languages as representing ‘grammatical mood’ in the same way that verbal inflections do in European languages. As we adopt a typological approach in this paper that treats the function of a linguistic form as constant and its formal manifestation as variable in this paper, we will consider verbal suffixes in Japanese and Korean, which encode ‘(ir)realis’ meaning to represent ‘grammatical mood’. Japanese has a very limited number of verbal affixes with modal meaning, so it is safe to say that mood plays a very limited role, if any, in Japanese. Though it is a matter of degree, Korean appears to manifest grammatical mood more overtly than Japanese does in that it has verbal suffixes positively marking indicative mood, e.g. V-ta. By comparing Tables 1 and 3, we can conclude that, for English, Japanese, and Korean, modal systems play more prominent roles than grammatical mood. In German, however, the division of labor between modal systems and mood is better balanced than it is for the other three languages. 3.3
Discourse systems in Japanese, Korean, English, and German
Discourse systems, i.e. modals indexing discourse relations among participants typically coded by modal particles, is not a major topic of discussion in Palmer (1986) or its second edition (2001). However, discourse systems merit our attention as they exhibit high systematicity in Japanese, leading to the postulation of “Discourse Modality” (or Modality of Discourse and Communication) (cf. 11’) in Japanese linguistics. As discussed by Horie and Taira (2002), Japanese and Korean differ, rather strikingly, in terms of the differential degrees of elaboration of their discourse systems. Japanese has a highly elaborate system of sentence-final particles that are divided into speaker-oriented particles or addressee-oriented particles: (25) a. (speaker-oriented) yo, ze, zo b. (addressee-oriented) ne, na (26) a. Ame-ga huri hazime-ta yo. rain-nom fall begin-past sfp ‘It has started raining, as I can assure you (e.g. the speaker telling his interlocutor about the weather on his end’.
Kaoru Horie and Heiko Narrog
b. Ame-ga huri hazime-ta ne. rain-nom fall begin-past sfp ‘It has started raining, as we can see (e.g. the speaker telling his interlocutor about the rain they are watching together)’. Similar to Japanese, Korean has a set of sentence-final suffixes, as in (27): (27) -ci, -kwun(a), -ney etc. (28) a. Na-nun mos ka-ci-yo. I-top cannot go-sup-pol ‘I am unable to go, I suppose’.
(Sohn 1994: 353)
b. Nwun-i o-nun-kwun. snow-nom come-pres-app ‘It is snowing!’
(Sohn 1994: 354)
Unlike Japanese sentence-final particles, however, Korean sentence-final particles do not constitute a uniform set with similar discursive functions such as (28a, b), and some of them, e.g.-ci and -kwun(a), respectively encoding a speaker’s “supposition” (Sohn 1994) and “newly made inference” (Lee 1991, Choi 1995), are more properly regarded as markers of evidential modality, as suggested by Horie and Taira (2002). German is again an interesting case in point. Unlike English, which virtually lacks this grammatical category, German has a set of so-called ‘modal particles’ which apparently serve some discursive functions (29): (29) doch (‘after all, though, just, truly, surely...’) schon (‘never fear, no doubt, surely, as a matter of course’) denn (‘evidently, as is well known, as I learn...’) etc. (Palmer 2001: 60) These modal particles can relate the content of speech to the immediate discourse context, as in (30a), or serve some interactive (addressee-oriented) function, as in (30b), where doch serves to mitigate the speaker’s utterance: (30) a. Ich hab’s doch immer gewusst. I have.it mod always known ‘I have always known that’. (and now it turns out to be true/but you wouldn’t listen to me etc.) b. Halt doch deine Klappe! keep.shut mod your mouth ‘Keep your mouth shut’. A contrastive summary of discourse systems is presented in Table 4:
What typology reveals about modality in Japanese
Table 4. Discourse systems in Japanese, Korean, German, and English Japanese
Korean
German
English
Sentence-final particles (yo, ne, zo, sa, na etc.) Divided into Speakeroriented or Heareroriented particles
Sentence-final suffixes (-ci, -ney, -kwuna etc.) Speaker-oriented meaning or evidentiality coded
Modal particles (doch, Virtually absent aber, auch, bloß, denn etc.) Serve to fit the content Virtually absent of an utterance to the content of speech
Japanese sentence-final particles can be divided into Speaker-oriented group (yo, ze, zo, sa etc.) and Hearer-oriented group (ne, na etc.). Korean and German counterparts apparently lack the latter category. English lacks this modal category altogether, primarily resorting to intonation and lexical means.
Japanese has a set of modal particles whose functions are oriented to either of two major discourse participants, i.e. speaker (represented by particle yo) and hearer (represented by ne). Korean and German, though possessing the category of modal particles category, do not appear to exhibit similar orientation to discourse participants unlike their Japanese counterparts. Korean and German modal particles are apparently more oriented toward the content of the utterance than to the discourse participants (see Horie & Taira 2001 for a detailed contrast between Japanese and Korean discourse systems). This modal category is entirely lacking in English2, which primarily resorts to intonation and lexical means to index discourse relations between speaker and hearer. 4. Why does Japanese have the distribution of modality categories it has? A communicative-discursive perspective Our contrastive study in Section 3 doesn’t necessarily present a neat demarcation of modality categories along geographical or genetic groupings. What emerges from our contrastive study is a more relativistic, continuous spectrum of modal categories distributed across the four languages. The four languages contrasted are shown to manifest slightly differing distributions of modal categories. In terms of the grammatical coding of modal meanings, the four languages exhibit the differing degrees of elaboration, as shown in (31). Bold and ‘( )’ indicate the greater and lesser degrees of elaboration respectively, while its absence indicates the virtual absence of grammatical coding: (31) a. modal system: b. mood: c. discourse system:
[J, K, G, E] [(J), K, G, (E)] [J, (K), G]
2. See Abraham 2009 for a contrast between German and English.
Kaoru Horie and Heiko Narrog
In terms of the types of modal meanings grammatically coded, the four languages exhibit the following cross-linguistic variation as in (32). Please note that the symbols (a)–(c) in (31) and (32) indicate the correspondence between the types of grammatical marking and the types of modal meaning coded therein: (32)
a. a’. b. c.
event modality, epistemic modality: evidential modality: realis/irrealis: discourse modality:
[J, K, G, E] [J, K, G] [(J), K, G, (E)] [J, (K), G]
How can we interpret the cross-linguistic distribution of formal and semantic categories of modality presented in (31) and (32)? As stated in the preceding sections, we adopt a functional typological approach to grammar. This approach pays close attention to its communicative and discursive foundations in line with the research frameworks of “Discourse and Grammar” by Du Bois (1987, 2003) and “Typological Discourse Analysis” by Myhill (1992). Specifically, we see grammar of a particular language not as a highly abstract and static competence-driven system of knowledge but as an emergent, dynamic set of communicative-discursive practices preferred in that linguistic community, some of which are highly grammaticalized and other parts of which are subject to varying degrees of conventionalization (see also Hopper 1998, Horie 2004, Narrog & Horie 2005). In what follows, we present a relativistic view of Japanese grammar, in this case modality, and the communicative-discursive motivations for its language-particular distribution, in comparison to the grammars of the other three languages. Among the four languages, English exhibits the least degree of elaboration in terms of grammatical coding of modal meanings in general, whereas German is shown to give overt encoding to all four types of modal meaning (32). This contrast ties in with the observation presented in Hawkins’ comparative typological study (1986) of English and German in terms of form-meaning correspondence. It was suggested there that English consistently tends toward greater surface structure ambiguity than German. In other words, German favors form-meaning transparency to a greater extent than English. This contrast in terms of form-meaning mapping holds in the domain of mood and modality. German thus elects to code all four types of modal meaning, whereas English does so rather sparingly. It has turned out that Japanese and Korean are situated in between the two poles in terms of the degree of explicit grammatical coding of modal meanings, as shown in (32). This suggests that Japanese and Korean prioritize different types of modal meaning among the four types, as discussed in Horie and Taira (2002). Japanese manifests itself as a language that has a modal system that encodes event, epistemic, and evidential modalities, a relatively impoverished grammatical
What typology reveals about modality in Japanese
mood, and a rich discourse system of sentence-final particles that can be grouped into speaker-oriented and hearer-oriented ones. Relative to the other languages contrasted, the modality system in Japanese can be characterized by the relatively high degree of elaboration in formal coding of evidential and discourse modalities. As for the prominence of evidential modality in Japanese, it was suggested in Horie (2000) that Japanese prioritizes the semantic distinction between directly perceived phenomena (typically visually perceived phenomena) and indirectly perceived (e.g., inferred) phenomena, leading to an elaborate system of complementation and a lexicon sensitive to this distinction (e.g., no, tokoro vs. koto, mono vs. koto) (see Makino (1983) for a similar observation). It is not unreasonable to assume that this socio-cultural preference, which may be deeply rooted in the Japanese culture from an ancient period of its history, is responsible for the development of grammaticalized evidential modality, already present in the Old Japanese auxiliaries such as rasi, kemu, and the distinction between past auxiliaries ki (directly experienced past) and keri (indirectly inferred past). The relatively impoverished mood coding in Japanese presents a rather striking contrast with Korean, which has an elaborate system of marking modal meanings by verbal inflectional suffixes, some of which are sensitive to the coding of realis and irrealis. As suggested by Horie (2000), Korean appears to prioritize the semantic distinction between realis and irrealis over that between ‘directly perceived phenomena’ and ‘indirectly perceived phenomena’, which is prioritized in Japanese. The greater/lesser degrees of elaboration of mood-marking inflectional suffixes between Korean and Japanese are an indication of this cross-linguistic difference. This difference is also arguably reflected in the differential semantic foundations of complementizer choice between Korean and Japanese, i.e., (u)m (realis) vs. ki (irrealis) in Korean, and no/tokoro (directly perceived phenomena) vs. koto (indirectly perceived phenomena). What then is the motivating factor behind the high elaboration of the discourse system in Japanese? In Horie (2002a, 2003a), based on morpho-syntactic contrasts between Japanese and Korean, it was suggested, by drawing explanatory insights from Hawkins’s (1986) contrastive study of English and German, that the differing degrees of tightness in form-meaning mapping between the two languages may be a reflection of different priorities in underlying cognitive- functional principles, form-meaning isomorphism (transparency) and economy. Specifically, it was argued that the greater emphasis on form-meaning mapping isomorphism in Korean and that on economy in Japanese are correlated with differing socio-cultural communicative norms between the two linguistic communities, i.e. prioritization of getting one’s message across without being misunderstood (Korean) versus that of contextual disambiguation of what is implicated
Kaoru Horie and Heiko Narrog
(Japanese). Horie and Taira (2002) posited that the tendency toward contextdependent ambiguity resolution leads to the development of grammatical resources such as sentence-final particles in Japanese that closely monitor the flow of information between discourse participants. This arguably led to the prominence of formal coding of discourse modality in Japanese. In contrast, the tendency toward clarity/explicitness of coding was argued to be responsible for the development of mood-marking suffixes in Korean, which are more closely related to the realis/irrealis status of the propositional content. This contrast was illustrated in Horie and Taira (2002: 188, partially modified) as in (33) (see example (11’) for the notation): (33) [Discourse structure[Sentence structure: [proposition] [M] [DM]]] K>J J>K We can see from (33) that Japanese and Korean manifest differential preferences in terms of the types of modal meaning they elect to encode grammatically. Through our extended contrastive study incorporating English and German, we observed that languages differ, rather considerably, in terms of the types of modal meaning grammatically encoded, the formal means employed to encode modal meaning, and, more importantly, in terms of the degree of elaboration in grammatical encoding. German was shown to pay due overall attention to the coding of different types of modal meaning, while English was shown to be rather selective, in line with the overall contrast between the two languages in terms of the ‘tightness’ of form-meaning mapping suggested by Hawkins (1986). Japanese and Korean are situated in between the two poles, and their differential choices are arguably influenced by the different socio-cultural communicative norms between the two linguistic communities. What we have offered as motivating factors for the particular makeup of the Japanese Modality system need further verification by related disciplines like cultural anthropology and comparative cultural studies. However, compared to such grammatical categories as tense and aspect, it does not seem unreasonable to assume that modality, which exhibits greater cross-linguistic variability, is more intimately affected and shaped by the particular socio-cultural communicative priorities of the linguistic community. 5. Conclusion and implications for grammar From a functional-typological perspective, this study presents a critical assessment of definitions of the grammatical category of modality based on the speaker’s
What typology reveals about modality in Japanese
subjective attitude/opinion, widely circulating within the Japanese linguistic community.3 Typological studies on modality, notably Palmer (2001), are shown to be highly suggestive in that they provide us with cross-linguistically applicable terminology. Our contrastive study of modality grams in Japanese, Korean, English, and German confirms the importance of relativizing the modality system in Japanese by employing typologically established terminology and contrasting it with its counterparts in other languages. The preceding analyses have the following implications for the study of Grammar:
(I) Grammar as a reflection of socio-cultural cognition and communicative practices.
Grammatical categories such as modality are useful abstractions employed in linguistic analysis. However, abstractions can prevent us from recognizing the fact that grammar is employed by real people living in a particular linguistic community and exercising specific socio-cultural practices. Recent years have witnessed a convergence of typological research interest on (socio-cultural) cognitive and communicative foundations of grammar, as represented by the emerging research disciplines such as Cognitive Typology (Kemmer 2003; Horie 2002b; Horie & Pardeshi 2009; see also Heine 1997), Ethnosyntax (Enfield 2002; see also Enfield & Levinson 2006), Discourse and Grammar (Du Bois 2003; Ariel 2009), and Grammar-in-Interaction (Ford, Fox & Thompson 2003). These approaches do not take grammar to be an autonomous system. Instead, grammar is viewed as an open system closely interacting with other cognitive and socio-cultural systems like perception, inference, discourse organization, communicative practices, and culture. The cross-linguistic differences in the degree of elaboration among different subcategories of modality presented in this study require an explanation, which 3. Reflecting continuing interest in modality phenomena, a collected volume on foundational issues and recent advances in the study of modality was published (Frawley 2006). de Haan (2006), included in Frawley (2006), offers a useful survey of expressions of modality from a typological perspective. In regards to modality phenomena in Japanese, a collected volume with the title Japanese Modality (Pizziconi & Kizu 2009), and a monograph entitled Modality in Japanese (Narrog 2009a: See also Narrog 2009b) were published recently. The former is based on a selection of papers presented at the international conference on Revisiting Japanese Modality held at SOAS of the University of London in 2006. Moriya and Horie (2009), included in Pizziconi and Kizu (2009), offers a contrastive and historical linguistic analysis of the Japanese modal system based particularly on a comparison with its Korean counterpart. Sawada (2012, in press) are a collection of the most recent papers dealing with theoretical and analytical issues in the study of Japanese modality (see Horie, in press, and Narrog, in press, included in Sawada, in press).
Kaoru Horie and Heiko Narrog
entails going beyond the confines of grammar and finding a link between grammar and other cognitive and communicative systems. (II) Grammar as the interface of indigenous descriptive tradition and typological analysis. Grammatical categories like modality are not just abstractions; they have concrete manifestations in a particular language. As such individual languages have indigenous descriptive grammatical traditions, which may be very elaborate with various language-particular ramifications. The traditional descriptive approaches to modality in Japanese outlined in Section 2.2 are one such example. We should not dismiss such indigenous descriptive grammatical traditions altogether as they can offer insight into the language-particular classification and organization of grammatical categories (e.g., part-of-speech system). At the same time, we should not be misled by the idiosyncrasies involved in indigenous descriptive terminology. ‘Modality’ in Japanese is a prime example of a grammatical category at the intersection of rich indigenous descriptive traditions and typological analysis, with not much interaction between them until now. A typological approach such as the one exercised in this paper can thus serve as a reality check for language-particular grammatical analyses and terminology (see also de Haan 2006). Abbreviations ACC ADN APP CONJ COP FUT IMP IND IP LOC MOD NOM
Accusative Adnominal Apperceptive Conjunctive Copula Future Imperative Indicative Interactional Particle Locative Modal Nominative
NEG POL PAST PRES PROPOS Q QUOT RETRO SUBJ SUP TOP
Negative Politeness Past Present Propositive Question Quotative Retrospective Subjunctive Suppositive Topic
What typology reveals about modality in Japanese
References Abraham, Werner. 2009. Die Urmasse von Modalität und ihre Ausgliederung Modalität anhand von Modalverben, Modalpartikeln und Modus. In Modalität und Evidentialität. Epistemik und Evidentialität bei Modalverb, Adverb, Modalpartikel und Modus, Werner Abraham & Elisabeth Leiss (eds), 251–302. Tübingen: Stauffenburg. Ariel, Mira. 2009. Pragmatics and Grammar. Cambridge: CUP. Bybee, Joan L. 1985. Morphology [Typological Studies in Language 9]. Amsterdam: John Benjamins. DOI: 10.1075/tsl.9 Bybee, Joan, Perkins, Revere & Pagliuca, William 1994. The Evolution of Grammar. Tense, Aspect, and Modality in the Languages of the World. Chicago IL: The University of Chicago Press. Choi, Soonja. 1995. The development of epistemic sentence-ending modal forms and functions in Korean. In Modality in Grammar and Discourse [Typological Studies in Language 32], Joan Bybee & Suzanne Fleischman (eds), 165–204. Amsterdam: John Benjamins. Comrie, Bernard. 1976. Aspect. Cambridge: CUP. de Haan, Ferdinand. 2006. Typological approaches to modality. In The Expression of Modality, William Frawley (ed.), 27–69. Berlin: Mouton de Gruyter. Du Bois, John W. 1987. The discourse basis of ergativity. Language 63: 805–855. DOI: 10.2307/415719 Du Bois, John W. 2003. Discourse and grammar. In The New Psychology of Language, Vol. 2, Michael Tomasello (ed.), 47–87. Mahwah NJ: Lawrence Erlbaum Associates. Enfield, Nicholas J. (ed.) 2002. Ethnosyntax: Explorations in Culture and Grammar. Oxford: OUP. Enfield, Nicholas J. & Levinson, Stephen C. (eds). 2006. Roots of Human Sociality. Oxford: Berg. Ford, Cecilia, Fox, Barbara A. & Thompson, Sandra A. 2003. Social interaction and grammar. In The New Psychology of Language, Vol. 2, Michael Tomasello (ed.), 119–143. Mahwah NJ: Lawrence Erlbaum Associates. Frawley, William (ed.) 2006. The Expression of Modality. Berlin: Mouton de Gruyter. DOI: 10.1515/9783110197570 Hawkins, John A. 1986. A Comparative Typology of English and German. London: Croom Helm. Heine, Bernd. 1997. Cognitive Foundations of Grammar. Oxford: OUP. Hopper, Paul J. 1998. Emergent grammar. In The New Psychology of Language, Michael Tomasello (ed.), 155–175. Mahwah NJ: Lawrence Erlbaum Associates. Horie, Kaoru. 1997. Form-meaning interaction in diachrony: A case study from Japanese. English Linguistics 14: 428–449. Horie, Kaoru. 1998a. Functional duality of case-marking particles in Japanese and its implications for grammaticalization: A contrastive study with Korean. In Japanese/Korean Linguistics 8: 147–159. Horie, Kaoru. 1998b. On the polyfunctionality of the Japanese particle No: From the perspectives of ontology and grammaticalization. In Studies in Japanese Grammaticalization: Cognitive and Discourse Perspectives, Toshio Ohori (ed.), 169–192. Tokyo: Kurosio. Horie, Kaoru. 2000. Complementation in Japanese and Korean: A contrastive and cognitive linguistic approach. In Complementation: Cognitive and Functional Perspectives [Converging Evidence in Language and Communication Research 1], Kaoru Horie (ed.), 11–31. Amsterdam: John Benjamins. DOI: 10.1075/celcr.1
Kaoru Horie and Heiko Narrog Horie, Kaoru. 2001. Kootyakugo ni okeru bunpooka no tokutyoo ni kansuru ninti gengogakuteki koosatu (A cognitive linguistic analysis of the characteristics of grammaticalization in agglutinative languages). In Ninti gengogaku ronkoo (Papers in cognitive linguistics), Masaaki Yamanashi et al. (eds), 89–131. Tokyo: Hituzi. Horie, Kaoru. 2002a. A comparative typological account of Japanese and Korean morpho- syntactic contrasts. Eonehag 32 (The Linguistic Society of Korea): 9–32. Horie, Kaoru. 2002b. Verbal nouns in Japanese and Korean: Cognitive typological implications. In Culture, Interaction, Language, Kuniyoshi Kataoka & Sachiko Ide (eds), 77–101. Tokyo: Hituzi. Horie, Kaoru. 2003a. What cognitive linguistics can reveal about complementation in non-IE languages: Case studies from Japanese and Korean. In Cognitive Linguistics and Non-Indo European Languages, Eugene H. Casad & Gary B. Palmer (eds), 363–388. Berlin: Mouton de Gruyter. DOI: 10.1515/9783110197150.9.363 Horie, Kaoru. 2003b. Differential manifestations of “modality” between Japanese and Korean A typological perspective. In Empirical and Theoretical Investigations into Language, Shuji Chiba et al. (eds), 205–216. Tokyo: Kaitakusya. Horie, Kaoru. 2004. Danwa to ninti (Discourse and cognition). In Ninti bunpoo ron II (Essays in cognitive linguistics II), Yoshihisa Nakamura (ed.), 247–278. Tokyo: Taishukan. Horie, Kaoru. In press. Modaritii no ruikeiron (A typology of modal expressions). In Modaritii: Riron to hoohoo (Modality: Theory and method), Harumi Sawada (ed.). Tokyo: Hituzi. Horie, Kaoru & Sassa, Yuko. 2000. From place to space to discourse: A contrastive linguistic analysis of Tokoro and Tey. In Japanese/ Korean Linguistics 9, Mineharu Nakayama & Charles J. Quinn (eds), 181–194. Stanford CA: CSLI. Horie, Kaoru & Taira, Kaori. 2002. Where Korean and Japanese differ: Modality vs. discourse modality. In Japanese/ Korean Linguistics 10, Noriko Akatsuka & Susan Strauss (eds), 178– 191. Stanford CA: CSLI. Horie, Kaoru & Pardeshi, Prashant. 2009. Gengo no taiporozii: Ninti ruikei ron no apurooti (A typology of languages: Cognitive typological approach). Tokyo: Kenkyusya. Johnson, Yuki. 2003. Modality and the Japanese Language. Ann Arbor MI: The University of Michigan Press. Kemmer, Suzanne. 2003. Human cognition and the elaboration of events: Some universal conceptual categories. In The New Psychology of Language, Vol. 2. Michael Tomasello (ed.), 89–118. Mahwah NJ: Lawrence Erlbaum Associates. Kiefer, Ferenc. 1999. Modality. In Concise Encyclopedia of Grammatical Categories, Keith Brown & Jim Miller (eds), 223–229. Oxford: Elsevier. Lee, Hyosang. 1991. Tense, Aspect, and Modality. A Discourse-Pragmatic Analysis of Verbal Affixes in Korean: From a Typological Perspective. PhD dissertation, UCLA. Li, Charles & Thompson, Sandra A. 1981. Mandarin Chinese. Berkeley CA: The University of California Press. Maikino, Seiichi. 1983. How sensitive is the Japanese language to directly Perceivable phenomena? In Issues in Syntax and Semantics. Festschrift for Masatake Muraki, Kazuko Inoue (ed.), 127–144. Tokyo: Sansyuusya. Martin, Samuel E. 1992. A Reference Grammar of Korean. Tokyo: Tuttle. Masuoka, Takashi. 1991. Modaritii no bunpoo (Grammar of modality). Tokyo: Kurosio. Masuoka, Takashi. 2000. Meidai to modaritii no kyookai o motomete (In search for the boundary between proposition and modality). In Nihongo bunpoo no syosoo (Aspects of Japanese grammar), Takashi Masuoka (ed.), 87–98. Tokyo: Kurosio.
What typology reveals about modality in Japanese Maynard, Senko K. 1993. Discourse Modality. Subjectivity, Emotion and Voice in the Japanese Language [Pragmatics & Beyond New Series 24]. Amsterdam: John Benjamins. DOI: 10.1075/pbns.24 Myhill, John. 1992. Typological Discourse Analysis. Oxford: Blackwell. Mithun, Marianne. 1999. The Languages of Native North America. Cambridge: CUP. Moriya, Testuharu & Horie, Kaoru. 2009. What is and what is not language-specific about the Japanese modal system? A comparative and historical perspective. In Japanese Modality: Exploring its Scope and Interpretations, Barbara Pizziconi & Mika Kizu (eds), 87–114. Houndmillls: Palgrave Macmillan. Narrog, Heiko. 2002a. Imiron kategorii to site no modaritii (Modality as a semantic category). In Ninti Gengogaku 2: Kategoriika (Cognitive linguistics 2: Categorization), Toshio Ohori (ed.), 217–251. Tokyo: University of Tokyo Press. Narrog, Heiko. 2002b. Polysemy and indeterminacy in modal markers. The case of Japanese Beshi. Journal of East Asian Linguistics 11: 123–167. DOI: 10.1023/A:1014948516744 Narrog, Heiko. 2005. On defining modality again. Language Sciences 27: 165–192. DOI: 10.1016/j.langsci.2003.11.007 Narrog, Heiko. 2009a. Modality in Japanese: The Layered Structure of the Clause and Hierarchies of Functional Categories [Studies in Language Companion Series 109]. Amsterdam: John Benjamins. DOI: 10.1075/slcs.109 Narrog, Heiko. 2009b. Modality, modariti and predication – The story of modality in Japan. In Japanese Modality. Exploring Its Scope and Interpretation, Barbara Pizziconi & Mika Kizu (eds), 9–35. Houndmills: Palgrave Macmillan. Narrog, Heiko. In press. Modaritii no teigi o megutte (Concerning the definition of modality). In Modaritii: Riron to hoohoo (Modality: Theory and method), Harumi Sawada (ed.). Tokyo: Hituzi. Narrog, Heiko & Horie, Kaoru. 2005. Hanasi kotoba ni okeru kanoo hyoogen (Potential expressions in spoken Japanese). In Gengogaku to nihongo kyooiku IV, Masao Minami (ed.), 99– 110. Tokyo: Kurosio. Nitta, Yoshio. 1991. Nihongo no modaritii to ninsyoo (Modality and person in Japanese). Tokyo: Hituzi. Nitta, Yoshio. 2000. Ninsiki no modaritii to sono syuuhen (Epistemic modality and its surroundings). In Modaritii (Modality). Nihongo no modaritii to ninsyoo (Modality and person in Japanese), Takuroo Moriyama et al. (eds), 81–159. Tokyo: Iwanami Syoten. Onoe, Keisuke. 1996. Bun o doo mita ka: Zyutugoron no gakusiteki tenkai (How did they view a sentence? A historical development of the analysis of predication in Japanese). Nihongogaku (Journal of Japanese Linguistics) 15(8): 4–12. Palmer, Frank R. 1986. Mood and Modality. Cambridge: CUP. Palmer, Frank R. 2001. Mood and Modality, 2nd edn. Cambridge: CUP. DOI: 10.1017/ CBO9781139167178 Pizziconi, Barbara, & Kizu, Mika (eds). 2009. Japanese Modality: Exploring its Scope and Interpretations. Houndmillls: Palgrave Macmillan. Sawada, Harumi. 1995. Studies in English and Japanese Auxiliaries. Tokyo: Hituzi. Sawada Harumi (ed.). 2012. Modaritii: Zirei kenkyuu (Modality: Case studies). Tokyo: Hituzi. Sawada, Harumi (ed.). In press. Modaritii: Riron to Hoohoo (Modality: Theory and method). Tokyo: Hituzi. Sohn, Homin. 1994. Korean. London: Croom Helm.
part 2
Frequency, interaction and language use
If rendaku isn’t a rule, what in the world is it? Timothy J. Vance The morphophonemic voicing phenomenon in Japanese known as rendaku is highly irregular, but several factors are believed to make rendaku more or less likely. This paper reviews some experiments intended to test the psychological reality of three such factors: Lyman’s Law, the semantic relationship between the two elements in noun + verb compound nouns, and salient semantic or phonological resemblances between novel compounds and existing compounds. The evidence suggests that each of these factors has at least a detectable effect on responses in experimental situations. Any realistic overall account of rendaku will have to incorporate a significant degree of intractable irregularity, but it will also have to be consistent with the intuition of naïve native speakers that rendaku is predictable.
1. Introduction Many Japanese morphemes exhibit a well-known voicing phenomenon called rendaku 連濁.1 Such a morpheme has one allomorph beginning with a voiceless obstruent and another allomorph beginning with a voiced obstruent. An example is a morpheme meaning ‘bird’: it appears with voiceless initial /t/ in /tori/ 鳥 ‘bird’ and in /tori + kago/ 鳥籠 ‘birdcage’ (cf. /kago/ ‘cage’), and it appears with voiced initial /d/ in /hači + dori/ 蜂鳥 ‘hummingbird’ (cf. /hači/ ‘bee’). When a morpheme shows this kind of alternation, the allomorph that begins with a voiced obstruent can appear only non-word-initially, and it is customary to say that rendaku ‘occurs’ in a word that contains this allomorph. Thus, in the examples involving /tori/~/dori/, rendaku occurs in /hači + dori/. Also, it is convenient to refer to an allomorph that begins with a voiced obstruent as the ‘voiced allomorph’ or the ‘rendaku allomorph’ of the relevant morpheme. In the case of /tori/~/dori/
1. Martin (1952: 48) offers ‘sequential voicing alternation’ as an English translation of the Japanese technical term rendaku. The phenomenon is now widely known among linguists, and many recent publications in English refer to it as (Japanese) rendaku.
Timothy J. Vance
‘bird’, the voiced allomorph is /dori/. The examples in (1) show all the pairs of alternating phonemes that fall under the rendaku label.2 (1)
a. b. c. d. e. f. g. h.
/f/~/b/ /h/~/b/ /t/~/d/ /k/~/g/ /c/~/z/ /s/~/z/ /č/~// /š/~//
/fune/ 船 ‘boat’ /hako/ 箱 ‘box’ /tama/ 玉 ‘ball’ /kami/ 紙 ‘paper’ /cuka/ 塚 ‘mound’ /sora/ 空 ‘sky’ /čikara/ 力 ‘strength’ /širuši/ 印 ‘symbol’
/kawa + bune/ 川船 ‘river boat’ /haši + bako/ 箸箱 ‘chopstick case’ /me + dama/ 目玉 ‘eyeball’ /kabe + gami/ 壁紙 ‘wallpaper’ /ari + zuka/ 蟻塚 ‘anthill’ /hoši + zora/ 星空 ‘starry sky’ /soko + ikara/底力‘latent strength’ /ya + iruši/ 矢印 ‘arrow symbol’
Notice that /b/ alternates with /f/ (as in /fune/~/bune/) and with /h/ (as in /hako/~ /bako/), not with /p/. The /f/~/b/ and /h/~/b/ alternations reflect the that /f/ and /h/ in native Japanese words are both descended from a single phoneme that was once pronounced [p].3 Notice also that /z/ alternates both with /c/ (as in /cuka/~/zuka/) and with /s/ (as in /sora/~/zora/), and that // alternates both with /č/ (as in /čikara/~/ikara/) and with /š/ (as in /širuši/~/iruši/). These pairings reflect the mergers of voiced fricatives and affricates in Tokyo Japanese, that is, the loss of earlier phonemic distinctions between [z] and [dz] and between [ʑ] and [ɟʑ]. Because of all these changes, the difference between the paired voiced and voiceless obstruents in (1) is often more than just the presence or absence of voicing. It is not immediately obvious that all the alternating phoneme pairs in (1) should be treated as instances of a single phenomenon. The same sort of question arises in connection with the three parallel voiceless/voiced fricative alternations in English nouns: /f/~/v/ (as in /wʊlf/ wolf versus /wʊlv + z/ wolves), /θ/~/ð/ (as in /bæθ/ bath versus /bæð + z/ baths), and /s/~/z/ (as in /haus/ house versus /hauz + əz/ houses). These three English pairings are all phonetically parallel, but many noun morphemes end in /f/ or /θ/ both in the singular and in the plural (e.g., reef /rif/ versus reefs /rif + s/, myth /mɪθ/ versus myths /mɪθ + s/), house is the only morpheme that shows the /s/~/z/ alternation, and no morpheme shows a parallel /ʃ/~/ʒ/ alternation. It is far from certain that ordinary native speakers of English intuitively recognize the three fricative alternations as instances of a single more abstract phenomenon. When it comes to rendaku, however, there is no real doubt that native speakers of Japanese see all the alternations in (1) as instances of a single more general phenomenon, in spite of the phonetic complications noted above. One likely reason is 2. The phonemic transcriptions of modern Japanese in this paper follow the analysis I adopted in Vance (1987: 9–47). I assume a uniform phonemic inventory for all vocabulary strata. 3. For details on the changes that earlier [p] has undergone, see Kiyose 1985; Hamano 2000; and Unger 2004.
If rendaku isn’t a rule, what in the world is it?
that the Japanese rendaku alternations are much more widespread than the English fricative alternations. The Japanese alternations appear in a very large number of morphemes, while the English alternations are confined to a small set of noun morphemes. At the same time, almost any preceding compound element or prefix provides an environment for the voiced allomorph of an alternating Japanese morpheme. In the English case, the plural morpheme is the only environment for the allomorphs ending with a voiced fricative.4 The Japanese writing system provides what is probably an even more powerful reason for native speakers to see the rendaku alternations as a unitary phenomenon: modern kana spelling represents all the alternations in exactly parallel fashion. The kana voicing diacritic called dakuten 濁点 represents more than just the addition of voicing in some cases, and the relationships between kana symbols with and without the voicing diacritic mirror the alternations shown in (1) above. For example, the diacritic is added to the symbols for /ta/ た, /sa/ さ, /ka/ か, and /ha/ は to write the syllables /da/ だ, /za/ ざ, /ga/ が, and /ba/ ば. Because of the two mergers of voiced fricatives and affricates, each of the syllables /zu/, /i/, /a/, /o/, and /u/ has two possible spellings. In most cases, the diacritic is added to /su/ す, /ši/ し, /ša/ しゃ, /šo/ しょ, and /šu/ しゅ to write /zu/ as ず, /i/ as じ, /a/ as じゃ, /o/ as じ ょ, and /u/ as じゅ.5 But when a morpheme has a voiceless allomorph that begins with one of /cu/ つ, /či/ ち, /ča/ ちゃ, /čo/ ちょ, and /ču/ ちゅ, the practice is to write its voiced allomorph by simply adding the diacritic and writing /zu/ as づ, /i/ as ぢ, /a/ as ぢゃ, /o/ as ぢょ, and /u/ as ぢゅ. As a result, in terms of kana spelling, rendaku is just the addition of the dakuten diacritic, as in ありづか for /ari + zuka/ 蟻塚 ‘anthill’ (compare つか for /cuka/ 塚 ‘mound’) and そこぢから for / soko + ikara/ 底力 ‘latent strength’ (compare ちから for /čikara/ 力 ‘strength’). 2. Fundamental irregularity According to Okumura (1955), it is extremely difficult to specify when rendaku occurs, although there are certain tendencies. This candid assessment could 4. In a few cases, an alternating English noun morpheme is also related to a verb with a stemfinal voiced fricative. For example, the verb house is pronounced /hauz/, just like the allomorph of the noun morpheme that appears in the plural. There are also some examples of a verb stem ending in /ð/ that has a different vowel than the related noun. One of these is bathe /beð/, with /e/ instead of the /æ/ of bath /bæθ/. Given the differences in meaning, the noun and verb in such a pair presumably cannot be analyzed as just allomorphs of the same morpheme. 5. The situation was quite different before the postwar orthographic reforms of 1946. See Seeley (1991: 104–125, 153–154) for discussion of the differences between prewar and postwar kana spellings.
Timothy J. Vance
perhaps be improved slightly by replacing the words ‘extremely difficult’ with the word ‘impossible’; certainly nothing in the half century of subsequent research on rendaku suggests that Okumura was overly pessimistic. The rendaku alternations are fundamentally irregular in two ways. First, there are morphemes that never exhibit rendaku, even though no putative constraint would be violated.6 The second elements in the three compounds in (2) are morphemes of this type. (2) /suna + kemuri/ 砂煙 ‘clouds of sand’ (cf. /kemuri/ ‘smoke’) /asa + cuyu/ 朝露 ‘morning dew’ (cf. /cuyu/ ‘dew’) /kucu + himo/ 靴紐 ‘shoelace’ (/himo/ ‘string’) The important point here is not the absence of rendaku in the three compounds in (2) but the fact that the three morphemes meaning ‘smoke’, ‘dew’, and ‘string’ simply do not have voiced allomorphs. The other kind of irregularity is that many morphemes sometimes exhibit rendaku and sometimes do not, even when no putative constraint is relevant. The examples in (3) illustrate. (3) /ki/ 木 ‘wood’
/cumi + ki/ 積み木 ‘(toy) wooden blocks’ /yose + gi/ 寄せ木 ‘wooden mosaic’
/tama/ 玉 ‘ball’
/mizu + tama/ 水玉 ‘water droplet’ /yu + dama/ 湯玉 ‘bubbles in boiling water’
/šima/ 島 ‘island’
/uki + šima/ 浮き島 ‘floating island’ /hanare + ima/ 離れ島 ‘solitary island’
/hi/ 日 ‘sun’
/asa + hi/ 朝日 ‘morning sun’ /niši + bi/ 西日 ‘westering sun’
In addition, there are individual words, such as those in (4), that can be pronounced either with or without rendaku.7 6. For an introduction to the putative constraints on rendaku, see Vance (1987: 136–146). 7. Of course, an individual speaker might prefer one alternative to the other for each word. In the case of /oku + fukai/~/oku + bukai/, Matsumura (1988) lists both pronunciations, but Shinmura (1998) lists only /oku + bukai/, and NHK Hōsō Bunka Kenkyūjo 1998 lists only /oku + fukai/. As for /waru + kuči/~/waru + guči/ and /ki + kaeru/~/ki + gaeru/, all three of these dictionaries list both alternatives. The rendaku in /ki + gaeru/ seems to be a recent development in modern ‘standard’ Japanese. Jorden and Chaplin (1962: 45) gives /ki + kaeru/, without rendaku, as the only pronunciation for this lexical item, but Jorden and Noda (1988: 301) gives /ki + kaeru/ and /ki + gaeru/ as alternative pronunciations, and the great majority of modern Tokyo speakers seem to use only /ki + gaeru/. Although /oku + fukai/~/oku + bukai/ and /ki + kaeru/~/ki + gaeru/ are inflectional forms, the morpheme boundaries between stem and inflectional suffix are not relevant here and are not marked. In the case of verbs, there is serious
If rendaku isn’t a rule, what in the world is it?
(4) /waru + kuči/~/waru + guči/ 悪口 ‘bad mouthing’ /ki + kaeru/~/ki + gaeru/ 着替える ‘change clothes’ /oku + fukai/~/oku + bukai/ 奥深い ‘deeply recessed’ On the other hand, rendaku is pervasive in the existing vocabulary, and it often occurs in newly coined words. As a comparison, consider once again the singular/ plural voicing alternations in English nouns (as in /wʊlf/ wolf versus /wʊlv + z/ wolves, /bæθ/ bath versus /bæð + z/ baths, and /haus/ house versus /hauz + əz/ houses). These English alternations are all losing ground; nouns that alternate for older speakers often do not alternate for younger speakers, and the alternations are never extended to new nouns. Also, English orthography provides no hint that the alternations are parallel. 3. Lyman’s Law The most important of the putative constraints on rendaku is known as Lyman’s Law, which says that when the second element in a combination already contains a voiced obstruent, rendaku does not occur.8 The examples in (5) illustrate. (5) a. /tani + gawa/ 谷川 ‘valley river’ cf. /kawa/ ‘river’ b. /tani + kaze/ 谷風 ‘valley wind’ cf. /kaze/ ‘wind’ Since /kawa/ does not contain a voiced obstruent, rendaku in /tani + gawa/ (5a) does not violate Lyman’s Law. In contrast, since /kaze/ does contain a voiced obstruent (namely, /z/), if the form in (5b) were */tani + gaze/, with rendaku, it would violate Lyman’s Law. The modern Japanese vocabulary contains only a very small number of exceptions to Lyman’s Law, the best known of which is probably /nawa + bašigo/ 縄梯子 ‘rope ladder’ (cf. /hašigo/ ‘ladder’). A study reported in Vance (1980) asked native speakers of Japanese to consider compounds consisting of a real first element and a made-up second element. Each made-up element and each compound was provided with a definition, and the task was to decide whether each compound sounded better with rendaku or
doubt about whether the division of inflectional forms into a stem followed by an inflectional suffix is a realistic way of representing the knowledge of native speakers. See Vance (1991) and especially Klafehn (2003). 8. Lyman’s Law is named after Benjamin Smith Lyman, who was the first non-Japanese to write about it (see Lyman 1894). According to Miyake (1932: 136), the Japanese scholar Motoori Norinaga (1730–1801) stated categorically that if the second element of a compound contains a voiced obstruent, its initial consonant does not voice.
Timothy J. Vance
without. In some cases, rendaku would violate Lyman’s Law, and in other cases it would not. Two of the test items appear in (6). (6) a. (川での競争) kawa de no kyōsō 川＋きだけ 川＋ぎだけ
‘competition on a river’ /kawa + kidake/ /kawa + gidake/
b. (川の状態) kawa no jōtai 川＋たつか 川＋だつか
‘condition of a river’ /kawa + tacuka/ /kawa + dacuka/
Notice that one of the two choices in (6a) (/kawa + gidake/) violates Lyman’s Law, while the other choice (/kawa + kidake/) does not. Neither of the two choices in (6b) violates Lyman’s Law. The graph in Figure 1 shows the percentage of rendaku responses by each subject to items with no voiced obstruent in the second element and to items with a voiced obstruent anywhere in the second element. Rendaku responses to items of the second type are like /kawa + gidake/ in (6a), i.e., they are violations of Lyman’s Law. The graph shows a wide range of individual differences, but every subject except S12 gave some responses that violated Lyman’s Law. On the other hand, for every subject, the proportion of rendaku responses to items without a voiced
Rendaku responseses
80
60 no Lyman’s Law violation Lyman’s Law violation
40
20
7
4 14 5
8 10 1 13 3 Subject
6
2
9 11 12
Figure 1. Rendaku responses to compounds consisting of a real first element and a madeup second element (Vance 1980)
If rendaku isn’t a rule, what in the world is it?
obstruent in the second element was higher than the proportion of rendaku responses to items with a voiced obstruent in the second element. The difference was statistically significant for every subject except S1.9 What, then, is the ontological status of Lyman’s Law? Does it make sense to say that there is a constraint on something (i.e., rendaku) that is not itself a rule?10 And why does Lyman’s Law persist if most people are willing to violate it in an experimental situation? 4. Direct object + verb stem This section considers two-element compound nouns consisting of a noun element followed by the so-called stem form of a transitive verb. The abbreviation N + V = N denotes a compound of this type. If the noun element is semantically the direct object of the verb element (DO + V = N), rendaku is supposedly inhibited (Okumura 1955; Sakurai 1966: 41; Yamaguchi 2011). The examples in (7) follow this putative pattern. (7) a. DO + V = N: /ai + cuke/ 味付け ‘flavoring’ cf. /ai o cukeru/ 味を付ける ‘add flavor’ b. nonDO + V = N: /kugi + zuke/ 釘付け ‘attaching with nails’ cf. /kugi de cukeru/ 釘で付ける ‘attach with nails’ As Kindaichi (1976) demonstrates with long lists of examples, however, resistance to rendaku in DO + V = N compounds is just a tendency in the existing vocabulary. The first three examples in (8) all have the form DO + V = N, but rendaku occurs. In (8d), on the other hand, the noun element is not the direct object of the verb element (nonDO + V = N), but rendaku does not occur. (8) a. /kuruma + dome/ 車止め ‘wheel block’ cf. /kuruma o tomeru/ 車を止める ‘stop wheels’ b. /kui + biki/ 籤引き ‘drawing lots’ cf. /kui o hiku/ 籤を引く ‘draw lots’ 9. The statistical test used was chi-square. The number of responses with and without rendaku was treated as one dimension and the items in which rendaku was or was not a violation of Lyman’s Law was treated as the other dimension of a 2×2 contingency table. Ihara and Murata (2006) used essentially the same method with two much larger groups of subjects, and they report comparable results. For a more recent experimental study of Lyman’s Law, see Kawahara (2012). 10. I thank George Bedell for putting the question this way in the discussion after a talk I gave at International Christian University in Tokyo in 2002.
Timothy J. Vance
c. /hotaru + gari/ 蛍狩り ‘firefly hunting’ cf. /hotaru o karu/ 蛍を狩る ‘hunt fireflies’ d. /kata + kake/ 肩掛け ‘shawl’ cf. /kata ni kakeru/ 肩に掛ける ‘put on the shoulders’ Needless to say, the presence or absence of rendaku is not an issue if the related verb does not begin with a voiceless obstruent or if it contains a non-initial voiced obstruent. For example, in the case of /muši + yoke/ 虫除け ‘insect repellent’ (cf. the verb /yokeru/ ‘avoid’), there simply is no possibility of rendaku, and in the case of /ne + sage/ 値下げ ‘price reduction’ (cf. the verb /sageru/ ‘lower’), rendaku would violate Lyman’s Law. Setting such cases aside, nonDO + V = N compounds without rendaku like (8d) are quite rare. Most nonDO + V = N compounds follow the proposed pattern and, like /kugi + zuke/ in (7b), have rendaku. In contrast, DO + V = N compounds with rendaku like (8a–c) are quite common.11 Nakamura and Vance 2002 reports an experimental study of the proposed pattern. As a preliminary step, a representative sample of the existing vocabulary was carefully constructed.12 This sample contains 403 N + V = N items susceptible to rendaku, that is, items for which the related verb begins with a voiceless obstruent 11. Sugioka (2005: 217–218) argues that there is a productive rule for creating new DO + V = N compounds, whereas new nonDO + V = N compounds result from what she characterizes as a kind of analogy. She also claims that rendaku always occurs when it is possible in newly coined nonDO + V = N compounds but seldom occurs in newly coined DO + V = N compounds. To be more precise, Sugioka says that the contrast is between initial noun elements that are adjuncts and those that are arguments, and arguments presumably include subjects as well as direct objects. Kindaichi (1976: 12) suggests that Subject + V = N compounds resists rendaku regardless of whether the verb element is transitive or intransitive, but in earlier work Sugioka (1986: 108, n. 24) disagrees and says that Subject + V = N compounds do not seems to resist rendaku. 12. A revised version of the vocabulary sample described in this paragraph is available to any interested reader as a pdf file . A detailed explanation of how the sample was constructed is included in Nakamura and Vance 2002, but the main points can be summarized as follows. We started with a large reverse dictionary organized by part of speech (Kazama 1979) and made a list of all the simplex verbs beginning with a voiceless obstruent and not containing a medial voiced obstruent. We then restricted the list to verbs that are reasonably common in modern Japanese by eliminating all those that do not appear as a headword in a medium-sized Japanese-English dictionary (Hasegawa et al. 1986). For each remaining verb, we found every N + V = N compound listed in a small reverse dictionary (Kitahara 1990) but discarded any compound that does not appear as a headword in the Japanese-English dictionary. We then tried to classify each remaining compound as DO + V = N or nonDO + V = N and eliminated those for which the classification was problematic. Finally, Nakamura (a native speaker of Japanese) eliminated compounds that were not in her active vocabulary or for which her own pronunciation differed from the pronunciation given in the Japanese-English dictionary. The version of the sample now available online incorporates a few revisions and corrections done in 2012 by a second native speaker of Japanese (Akiko Takemura).
If rendaku isn’t a rule, what in the world is it?
+DO
−DO
+R
143
145
−R
104
11
Figure 2. Representative sample of N + V = N compounds in existing vocabulary
and does not contain non-initial voiced obstruent. The properties of interest are distributed as in Figure 2, where + DO means a DO + V = N compound, −DO means a nonDO + V = N compound, +R means that rendaku occurs, and −R means that rendaku does not occur. It is clear that rendaku occurs in a much higher proportion of nonDO + V = N (−DO) compounds (145/156 = 93%) than DO + V = N ( + DO) compounds (143/247 = 58%). On the other hand, rendaku is hardly exceptional in +DO compounds, since it occurs in more than half (58%) of the relevant items. In an earlier experimental study, Kozman (1998) asked subjects to respond to ten novel N + V = N compounds, each plausibly ambiguous between a +DO meaning and a −DO meaning, in which the nominal element is understood as instrumental. One of Kozman’s items had the alternative pronunciations /aši + kaki/ (without rendaku) and /aši + gaki/ (with rendaku). Twenty subjects heard the −R pronunciation and twenty subjects heard the +R pronunciation. The task was to choose one of two written definitions for each item. In this case, the definitions were as in (9). (9) a. +DO: E no kurasu de seito ga ashi o kaku. ‘The students draw a foot in drawing class.’ b. −DO: Ude no fujiyū na hito ga ashi de kaku. ‘The person with disabled arms draws with his foot.’ There were 200 responses to +R items and 200 responses to −R items (10 items × 20 subjects for each). The percentage of +DO responses was 37% for +R and 45.5% for −R items, and while the difference (74/200 versus 91/200) is in the predicted direction, it is not statistically significant. Kozman’s results thus provide no support for the idea that native speakers of Japanese internalize the putative regularity even as a tendency. The Nakamura and Vance 2002 experiment involved a production task. The expectation was that the results would corroborate Kozman’s findings, but the actual results were not quite in line with this expectation. Twenty-one subjects participated, and the task was to pronounce N + V = N compounds in response to spoken prompts that were mostly of the form NOUN (PARTICLE) VERB. There
Timothy J. Vance
were eight verbs and four prompts for each verb, two +DO prompts and two −DO prompts. For each verb, one +DO prompt and one −DO prompt were each intended to elicit an existing compound. The other +DO prompt and the other −DO prompt were each intended to elicit a novel compound. The prompts for two existing compounds are shown in (10). (10) a. PROMPT: /mono o hosu/ 物を干す ‘dry things’ (+DO) EXPECTED RESPONSE: /mono + hoši/ ‘drying rack’ 物干 (existing; −R) b. PROMPT: /kage de hosu/ 陰で干す ‘dry in the shade’ (−DO) EXPECTED RESPONSE: /kage + boši/ ‘drying in the shade’ 陰干し (existing; +R) All the existing compounds that the prompts were intended to elicit conform to the putative generalization: the eight +DO items lack rendaku, as in (10a), and the eight −DO items have rendaku, as in (10b). The prompts for two novel compounds are shown in (11). (11) a. PROMPT: /kucu o hosu/ 靴を干す ‘dry shoes’ (+DO) EXPECTED RESPONSE: /kucu + hoši/ or /kucu + boši/ 靴干し ‘shoe drying’ (novel) b. PROMPT: /yoru hosu/ 夜干す ‘dry at night’ (−DO) EXPECTED RESPONSE: /yoru + hoši/ or /yoru + boši/ 夜干し ‘drying at night’ (novel) The subjects were divided into two groups. Those in one group first heard all the prompts for existing compounds and then heard all the prompts for novel compounds. The subjects in the other group first heard all the prompts for novel compounds and then heard all the prompts for existing compounds. In this experimental design, the dependent variable is the number of rendaku responses, and there are three factors: subject group (existing items first vs. novel items first), item set (existing items vs. novel items), and noun role (direct object vs. other role). In the ANOVA results, all three two-way interactions were significant, and so was the main effect of noun role (+DO vs. −DO).13 The interactions will be interpreted one by one in the following paragraphs. 13. The three-way interaction was not significant. The main effect of subject group (novel items first vs. existing items first) was significant by items (F[1,28] = 33.20; p<.001) but not by subjects (F[1,19] = 3.08; p>.09). The main effect of item set (novel words vs. existing words) was significant by subjects (F[1,19] = 6.49; p<.02) but not by items (F[1,28] = 2.47; p>.12). The main effect of noun role (direct object vs. other) was significant both by items (F[1,28] = 181.22; p<.0001) and by subjects (F[1,19] = 316.31; p<.0001).
If rendaku isn’t a rule, what in the world is it? % rendaku responses
80
Existing
60
Novel
Existing
Novel
40 20 Existing first
Novel first
Figure 3. Subject-group × item-set interaction
As the graph in Figure 3 shows, the subject-group×item-set interaction reflects the fact that the subjects who responded to existing items first gave more rendaku responses to the existing compounds than to the novel compounds, whereas the subjects who responded to the novel items first gave more rendaku responses to the novel compounds than to the existing compounds.14 This subject-group×itemset interaction was surprising, since there is no reason to expect the order of testitem presentation to have any effect on whether or not an existing word would be produced with rendaku. The graph in Figure 4 shows that both groups of subjects gave more rendaku responses to nonDO + V = N compounds than to DO + V = N compounds. The subject-group×noun-role interaction reflects that fact that the gap was wider for the subjects who responded to the existing items first than for the subjects who responded to the novel items first.15
% Rendaku responses responses
−DO 80
−DO
60 40 +DO
+DO
20
Existing first
Novel first
Figure 4. Subject-group×noun-role interaction 14. The subject-group×item-set interaction was significant both by items (F[1,28] = 36.14; p < .0001) and by subjects (F[1,19] = 21.86; p < .0001). 15. The subject-group×noun-role interaction was significant both by items (F[1,28] = 19.46; p < .0001) and by subjects (F[1,19] = 9.23; p < .007).
% rendaku responses responses
Timothy J. Vance −DO 80
−DO
60 40
+DO
20
+DO
Existing set
Novel set
Figure 5. Item-set×noun-role interaction
This interaction just means that the main effect of noun role (i.e., the preference for rendaku in −DO items) was stronger for the subjects who responded to the existing items first than for the subjects who responded to the novel items first. Like the graph in Figure 4, the graph in Figure 5 shows that the subjects gave more rendaku responses to nonDO + V = N compounds than to DO + V = N compounds. The item-set×noun-role interaction reflects that fact that the gap was wider for the novel items than for the existing items.16 This interaction means that the responses to novel items were more in line with the putative generalization about N + V = N compounds than the responses to existing items – a completely unexpected result. Notice that the subjects produced many ‘incorrect’ responses to prompts that were intended to elicit existing N + V = N words. Despite the fact that these existing words are all DO + V = N compounds without rendaku or nonDO + V = N compounds with rendaku, the subjects gave 36% rendaku responses for the DO + V = N items and 22% no rendaku responses for the nonDO + V = N items. A small amount of random error is inevitable in an experimental situation, but these percentages of ‘incorrect’ responses are much too high to ignore. It seems likely that the experimental situation was stressful enough to interfere with lexical retrieval, and if so, the subjects sometimes had to coin a word rather than retrieve a known item in response to an existing-set prompt.17 In any case, in contrast to the earlier 16. The item-set×noun-role interaction was significant both by items (F[1,28] = 15.00; p<.001) and by subjects (F[1,19] = 39.93; p<.0001). 17. Kaori Kabata suggests that the low frequency of some of the existing test items might also be part of the explanation for the high error rate. As a crude way of correcting for differences in salience, Nakamura and Vance (2002) considered using frequency to weight each existing N + V = N compound in the representative sample of the existing vocabulary in Figure 2, but only
If rendaku isn’t a rule, what in the world is it?
study by Kozman (1998), noun role (direct object vs. other) emerged in this study as a significant factor. 5. Analogy Ohno (2000: 161) argues that a kind of analogy is the basic mechanism for extending rendaku to new vocabulary items: ‘Native speakers refer to existing compounds and refer to a semantically and/or phonetically parallel form when they determine the rendaku/non-rendaku form of a novel compound.’ For example, Ohno says that /kami/~/gami/ 髪 ‘hair’ always appears as /gami/ when it is the final element in a compound, except in the word /kuro + kami/ 黒髪 ‘black hair’. When Ohno presented his subjects with 白髪 (‘white’+‘hair’), 27 of 31 chose /širo + kami/ over /širo + gami/. In contrast, Ohno says that /či/~/i/ 血 ‘blood’ never appears as /i/ when it is the final element in a compound, except in the word /hana + i/ 鼻血 ‘nosebleed’. When Ohno presented his subjects with 耳血 (‘ear’+‘blood’), 37 of 43 chose /mimi + i/ over /mimi + či/. Ohno assumes that the novel compounds meaning ‘white hair’ and ‘ear bleeding’ strongly bias native speakers of Japanese toward accessing the existing items /kuro + kami/ and /hana + i/. This is what he means by a ‘semantically parallel’ form. As a result, Ohno argues, most native speakers choose the form for each novel compound that we would not expect if the choice depended simply on the proportion of existing items with rendaku. As an example of the influence of ‘phonetically parallel’ forms, Ohno compares the existing compounds /waka + kusa/ 若草 ‘young grass’ (without rendaku) and /i + gusa/ 藺草 ‘rush’ (with rendaku). A majority (25/35) of Ohno’s subjects preferred /aka + kusa/ (without rendaku) over /aka + gusa/ for 赤草 (‘red’ + ‘grass’), but a majority (25/35) also preferred /ki + gusa/ (with rendaku) over /ki + kusa/ for 黄草 (‘yellow’ + ‘grass’). Ohno attributes the difference to the influence of phonologically similar /waka + kusa/ and /i + gusa/. Ohno (2000: 162) goes on to suggest that ‘If there is no possible reference to existing rendaku forms, the item does not undergo rendaku in a novel compound.’ For example, Ohno says that /taka/ 鷹 ‘hawk’ is rendaku-immune (i.e., does not
12 of the 403 compounds are listed in the one readily available source of frequency data (Kokuristu Kokugo Kenkyūjo 1962). In any case, since the database for this source was magazines published in 1956, the reported frequencies are not likely to correspond very well to salience for the subjects in Nakamura and Vance experiment, who were born around 1980. It would be interesting to be look at the errors in existing items in the light of frequency information from a more recent and larger database.
Timothy J. Vance
alternate with /daka/).18 When Ohno presented his subjects with 大鷹 (‘large’ + ‘hawk’) and 赤鷹 (‘red’ + ‘hawk’), 41 of 43 chose /oo + taka/ over /oo + daka/ and /aka + taka/ over /aka + daka/. Ohno also says that /surume/ 鯣 ‘dried squid’ does not occur as the final element in any existing compound, and when he presented his subjects with 大鯣 (‘large’ + ‘dried squid’) and 赤鯣 (‘red’ + ‘dried squid’), 41 of 43 chose /oo + surume/ over /oo + zurume/ and /aka + surume/ over /aka + zurume/. As Ohno (2000: 162) points out, this kind of analogy accounts for the persistence of Lyman’s Law, even if this ‘constraint’ is just a historical residue. On the other hand, it is not clear why speakers would extend rendaku to elements they have never heard, such as the made-up elements in the Vance 1980 experiment that was summarized above. 6. An illusion of regularity? This paper has reviewed the results of three kinds of psycholinguistic experiments, all designed to explore some aspect of the familiar yet puzzling phenomenon known as rendaku. Even on the generous assumption that these experiments were well enough designed to avoid hopeless confounding by uncontrolled extraneous variables, none of them gives us more than a tantalizing glimpse at whatever the underlying ‘big picture’ might be. It is relatively easy to makes sense of each set of results on its own, but reconciling them all with each other and with still other sets of results is a challenge for any theory of ‘grammar’ that aspires to encompass the (morpho-) phonological aspects of language in a psychologically plausible way. If we take this challenge seriously, even the (sometimes vague) feelings of native speakers are part of what ultimately needs to be explained. The analogical decisions explored by Ohno (2000) seem to turn on perceived similarity, and it may be that such decisions are close enough to certainty in any particular case for any particular individual speaker that rendaku in general feels predictable. This kind of predictability is not exactly the same thing as regularity, but it is not hard to see how a person could translate a feeling of predictability into a (clearly mistaken) belief that there somehow must be a relatively straightforward rule that could be discovered and even taught explicitly to non-natives.
18. Although no compounds ending in /daka/ meaning ‘hawk’ are common enough to be included in smaller dictionaries, a comprehensive dictionary such as Kōjien (Shinmura 1998) includes several headwords that fit this description. There are also species names, such as /aka + hara + daka/ 赤腹鷹 ‘red-bellied hawk’, that do not appear in ordinary dictionaries but do appear in field guides for bird watchers.
If rendaku isn’t a rule, what in the world is it?
References Hamano, Shoko. 2000. Voicing of obstruents in Old Japanese: Evidence from the sound-symbolic stratum. Journal of East Asian Linguistics 9: 207–225. DOI: 10.1023/A:1008367619295 Hasegawa, Kiyoshi, et al. (eds). 1986. Ōbunsha waei chūjiten. Tokyo: Ōbunsha. Ihara, Mutsuko & Murata, Tadao. 2006. Some experiments on sequential voicing. On’in Kenkyū/ Phonological Studies 9: 17–24. Jorden, Eleanor & Ito Chaplin, Hamako. 1962. Beginning Japanese, Part II. New Haven CT: Yale University Press. Jorden, Eleanor & Noda, Mari. 1988. Japanese: The Spoken Language, Part 2. New Haven CT: Yale University Press. Kawahara, Shigeto. 2012. Lyman’s law is active in loanwords and nonce words: Evidence from naturalness judgment studies. Lingua 122: 1193–1206. DOI: 10.1016/j.lingua.2012.05.008 Kazama, Rikizō (ed.). 1979. Tsuzuriji gyakujun hairetsu gokōsei ni yoru Daigenkai bunrui goi. Tokyo: Fuzanbō. Kindaichi, Haruhiko. 1976. Rendaku no kai. Sophia Linguistica 2: 1–22. Kitahara Yasuo (ed.). 1990. Nihongo gyakubiki jiten. Tokyo: Taishūkan. Kiyose, Gisaburō N. 1985. Heianchō hagyō-shiin p-onron. Onsei no Kenkyū 21: 73–87. Klafehn, Terry. 2003. Emergent Properties of Japanese Verb Inflection. PhD dissertation, University of Hawai‘i. Kokuritsu Kokugo Kenkyūjo. 1962. Gendai zasshi kyūjusshu no yōgo yōji, Vol. 1: Sōki oyobi goihyō. Tokyo: Shūei Shuppan. Kozman, Tam. 1998. The psychological status of syntactic constraints on Rendaku. Japanese/Korean Linguistics 8, David Silva (ed.), 107–120. Stanford CA: CSLI. Lyman, Benjamin Smith. 1894. The Change from Surd to Sonant in Japanese Compounds. Philadelphia PA: Oriental Club of Philadelphia. Martin, Samuel E. 1952. Morphophonemics of Standard Colloquial Japanese. Supplement to Language [Language Dissertation No. 47]. Matsumura, Akira (ed.). 1988. Daijirin. Tokyo: Sanseidō. Miyake, Takeo. 1932. Dakuon-kō. Onsei no Kenkyū 5: 135–90. Nakamura, Kumiko & Vance, Timothy J. 2002. Rendaku in noun + verb-stem compounds: A production task. Paper presented at LP’2002, Urayasu, Chiba. NHK Hōsō Bunka Kenkyūjo (ed.) 1998. Nihongo hatsuon akusento jiten, new edn. Tokyo: Nippon Hōsō Shuppan Kyōkai. Ohno, Kazutoshi. 2000. The lexical nature of Rendaku in Japanese. Japanese/Korean Linguistics 9, Mineharu Nakayama & Charles J. Quinn Jr. (eds), 151–164. Stanford CA: CSLI. Okumura, Mitsuo. 1955. Rendaku. Kokugogaku jiten, Kokugo Gakkai (ed.), 961–962. Tokyo: Tōkyōdō. Sakurai, Shigeharu. 1966. Kyōtsūgo no hatsuon de chūi subeki kotogara. Nihongo hatsuon akusento jiten, Nihon Hōsō Kyōkai (ed.), 31–43. Tokyo: Nihon Hōsō Shuppan Kyōkai. Seeley, Christopher. 1991. A History of Writing in Japan. Leiden: E.J. Brill. Shinmura, Izuru (ed.). 1998. Kōjien, 5th edn. Tokyo: Iwanami. Sugioka, Yoko. 1986. Interaction of Derivational Morphology and Syntax. New York NY: Garland. (University of Chicago PhD dissertation, 1984).
Timothy J. Vance Sugioka, Yoko. 2005. Multiple mechanisms underlying morphological productivity. Polymorphous Linguistics: Jim McCawley’s Legacy, Salikoko S. Mufwene, Elaine J. Francis & Rebecca S. Wheeler (eds), 204–223. Cambridge MA: The MIT Press. Unger, J. Marshall. 2004. Alternations of m and b in Early Middle Japanese: The deeper significance of the sound-symbolic stratum. Japanese Language and Literature 38: 323–337. DOI: 10.2307/4141293 Vance, Timothy J. 1980. The psychological status of a constraint on Japanese consonant alternation. Linguistics 18: 245–267. DOI: 10.1515/ling.1980.18.3-4.245 Vance, Timothy J. 1987. An Introduction to Japanese Phonology. Albany NY: SUNY Press. Vance, Timothy J. 1991. A new experimental study of Japanese verb morphology. Journal of Japanese Linguistics 13: 145–166. Yamaguchi, Kyoko. 2011. Accentedness and Rendaku in Japanese deverbal compounds. Gengo Kenkyu 140: 117–134.
The semantic basis of grammatical development Its implications for modularity, innateness, and the theory of grammar Yasuhiro Shirai This chapter reviews semantic bias observed in the acquisition of grammatical categories, and examines its theoretical implications. It first reviews the acquisition of tense-aspect morphology, in particular of English, and proposes a model of grammatical development called the “input-based prototype formation”, taking examples from cross-linguistic data. I will then discuss semantic bias observed in the acquisition of pronominal case in English, conditionals in Japanese and Korean, causative morphology in Japanese, nominative case omission in Japanese and Korean, and relative clause construction in Japanese. Based on the review, I will argue against modularity and the innateness hypothesis, which constitutes the fundamental assumptions espoused by generative grammar, and argue for further analysis of semantic bias in frequency distribution in the input language that children are exposed to.
1. Introduction In a study of English acquisition, I have investigated the effect of input on the acquisition of tense-aspect morphology, and proposed an explanation for the acquisition patterns of past tense and progressive aspect markers based on input frequency (Shirai 1991, 1994; Shirai & Andersen 1995). Specifically, it has been observed that in the acquisition of various languages (e.g., French, Italian, Japanese, Greek, Turkish, English) children start to use the past tense form with telic verbs (achievements/accomplishments) such as break, fall, and the progressive form with dynamic atelic verbs (activities) such as play, watch. In the study of English acquisition, I also found that about 60% of progressive morphology and of past tense morphology in child-directed speech was used with activity verbs and achievement verbs respectively. Importantly, this association between tense-aspect
Yasuhiro Shirai
morphology and verb type was found to be even stronger (almost 100%) in early child language. I have attributed this to children’s prototype formation of semantic representation for these grammatical markers. This paper discusses the importance of this model, which I call “input-based prototype formation” for the development of closed-class items, taking examples from crosslilnguistic data. I will discuss the acquisition of tense-aspect, pronominal case in English (Budwig 1996), conditionals in Japanese and Korean (Akatsuka & Clancy 1993), causative morphology in Japanese (Shirai, Miyata, Naka & Sakazaki 2000, 2001), nominative case omission in Japanese and Korean (Miyamoto, Wexler, Aikawa & Miyagawa 1998; Lee & Wexler 2001), and relative clause construction in Japanese (Ozeki & Shirai 2010). I will then discuss the implications of this model for the general theory of grammatical development. 2. Tense-aspect Bickerton (1981, 1984, see also 1999) proposed the Language Bioprogram Hypothesis, based primarily on his research on the genesis of creole languages. Observing that even genetically unrelated creole languages share common linguistic features in domains as varied as the causative construction, nominal specificity, and tenseaspect marking, Bickerton hypothesized that these commonalties must come from a bioprogram–i.e., knowledge that is innately specified for all human children. Bickerton’s argument is that the commonalties among genetically unrelated creole languages can be explained only if we postulate that children modify the pidgin input they are exposed to in ways that are compatible with the bioprogram. 2.1
Lexical aspect
Before I discuss Bickerton’s proposal, let me briefly define lexical aspectual classes that are to be a basis for understanding the main point of the paper. Inherent aspect (otherwise known as situation aspect or aktionsart) refers to the temporal semantic characteristics of the verb and elements associated with it (arguments and some adjuncts) (Smith 1997). Vendler’s (1957) four classes are the most well-known:
State: love, exist, know, believe, think that... Activity: run, walk, swim, think about... Accomplishment: make a chair, walk to the store Achievement: reach the summit, die, win the race, notice
State terms refer to stative situations that do not change unless some other force changes them. States are stative, while the other three classes are dynamic. Among
The semantic basis of grammatical development
the dynamic situations, activity terms involve duration, but they do not have an inherent endpoint. That is, one can stop the action at any time point, but the fact remains that the action has been done. Activity is therefore atelic (non-telic), not involving an inherent endpoint. Accomplishment terms, on the other hand, are telic, involving a necessary endpoint. One cannot truthfully say she made a chair if she stops in the middle of making a chair. Finally, achievement terms refer to situations that are linguistically conceived as punctual or instantaneous. Achievements, therefore, are [+punctual] and [+telic].1 These four classes can be schematically represented as follows (Andersen 1990):
State ______________ Activity ~~~~~~~ Accomplishment ~~~~~~~x Achievement x
love, contain, know run, walk, swim paint a picture, build a house fall, drop, win the race
In this schematization, a long solid line is used to represent states, because states have no apparent beginning point or endpoint and endure indefinitely unless some external force changes them. The wavy lines for activities and accomplishments indicate the dynamic duration of an action, while x for accomplishments and achievements represents a punctual point of change of state, signaling telicity (i.e., natural endpoint). 2.2
Language bioprogram hypothesis: PNPD and SPD
Bickerton (1981) postulated two types of aspectual distinctions that human children are innately endowed with: the PNPD (the Punctual-Non-Punctual Distinction) and the SPD (the State-Process Distinction), both of which he claimed are grammatically marked in all “true” creole languages as defined by him. He reviewed studies in the acquisition of French (Bronckart & Sinclair 1973) and Italian (Antinucci & Miller 1976) in support of the PNPD, and studies in English (Brown 1973; Kuczaj 1978) and Turkish (Slobin & Aksu 1980) in support of the SPD. Regarding the PNPD, it was observed that children acquiring French and Italian tend to attach past marking to ‘punctual’ (i.e., telic) verbs, but not to ‘nonpunctual’ (i.e., atelic) verbs (Bronckart & Sinclair 1973; Antinucci & Miller 1976).2 Based on this observation, Bickerton claimed that children are marking ‘punctuality’ 1. Smith (1997) proposes a fifth category semelfactive, which is punctual but atelic, such as jump, kick, knock. Although this revision of the Vendlerean verb classification has important consequences for the acquisition of tense-aspect (Shirai & Andersen 1995), this paper uses the four-way classification to simplify the discussion. 2. The terminology used by Bickerton (1981) is slightly different from those used by aspectologists. His ‘punctual’ roughly corresponds to ‘telic’, and ‘nonpunctual’ to ‘atelic’ as defined above.
Yasuhiro Shirai
rather than pastness when they use past tense markers. Regarding the SPD, Bickerton claimed that, because children rarely (if ever) make the error of incorrectly attaching progressive marking to stative verbs, they must already know the distinction between state and non-state (i.e., process). Bickerton also noted that Turkish children at the early stages use the indirect past marker -mIs and the direct past marker -dI to differentiate between static and dynamic events, violating the norms of adult grammar. This, Bickerton argued, provides further evidence in support of his hypothesis. Bickerton (1981: 163) thus suggested that the acquisitional patterns of tenseaspect morphology in these studies can only be explained by assuming an innate bioprogram. He admits that a bioprogram specification may not appear in child language in an explicit form, but argues that if there is some indication of a bioprogram at work, and if the phenomenon cannot be explained by other theories of acquisition, this constitutes a significant support for the bioprogram hypothesis. 2.3
Input distribution as explanation
In a series of studies, I have proposed an alternative account for the acquisitional patterns of tense-aspect morphology which Bickerton had earlier attributed to the bioprogram. The proposal is that the pattern of acquisition is primarily driven by the distribution of the tense-aspect morphology in the input, and by children’s initial process of category formation. Bickerton (1981) evidently assumes that the distribution in the input has no significant consequences in the acquisitional patterns of tense-aspect morphology. In reviewing the acquisition of Italian (Antinucci & Miller 1976), which shows that perfective past forms (referred to as ‘participials’) are only attached to ‘punctual’ (i.e., telic) verbs, and imperfective past forms only to activity verbs, Bickerton (1981: 174) suggested that in child Italian perfective past and imperfective past are in complementary distribution, the first being used for punctual verbs, the second for nonpunctual ones. He states: Note that this does not reflect anything in Italian grammar; all Italian verbs, whether punctual or nonpunctual, activity or change-of-state verbs, have both perfective and imperfective past tenses.
But while it may be possible to attach perfective or imperfective past forms to verbs of any verb type, not all verb classes are equally likely to take perfective and imperfective past forms in real-life language use. There are natural relationships between perfective past and telic verbs (Bickerton’s punctual verbs) on the one hand, and between imperfective past and atelic verbs (Bickerton’s nonpunctual
The semantic basis of grammatical development
verbs), on the other. Leone (1990), cited in Andersen (1993), quantitatively establishes this association: in an adult-adult native speech sample (an interview), perfective past forms are predominantly attached to telic verbs. For example, based on token count, 97% of the perfective past forms were attached to telic verbs (achievements and accomplishments), while 72% of imperfective past forms are attached to atelic verbs (activities and states). If such distributional bias is present in the input, it is not surprising that children follow this trend in acquiring tense-aspect morphology. In my previous research (Shirai 1991, 1994; Shirai & Andersen 1995) I tested this hypothesis by investigating the use of verbal morphology by three children acquiring English and by their mothers in the CHILDES database. I found that children’s early past marking is predominantly restricted to achievement verbs. This supports one of the predictions of Bickerton’s bioprogram hypothesis. At the earliest, emergent stage of past tense verbs, most of the past inflections are attached to achievement verbs (i.e., [+punctual], [+telic]). In the acquisition of English verbs, children use the bare stem as the all-purpose verbal form at the earliest stage (Brown 1973). Then gradually they begin to attach verbal morphology on selective verb forms. What happens here is underextension of past morphology; children rarely overextend past tense morphology to inappropriate contexts (Brown 1973; Kuczaj 1976). Then children gradually extend the application of past tense morphology to contexts that deviate from this initial verb class, i.e., they attach past morphology to atelic (Bickerton’s non-punctual) verbs, thus approximating the adult norm. At the same time, however, I also found that the caretakers’ past marking is likewise skewed in a direction congruent with the children’s nearly absolute trend: approximately 60% of past-tense markers in the caretakers’ input occurred with achievement verbs. It may have been that this high correlation of punctuality and telicity with past morphology resulted in the near absolute tendency observed in the children’s output. Let us now turn to the acquisition of progressive marking, which concerns both the PNPD and the SPD. First, Bickerton (1981) claims that children acquiring English first mark non-punctuality (i.e., imperfective) by means of the progressive -ing forms, and then later mark punctuality (i.e., telicity) by means of irregular past tense forms.3 Shirai and Andersen (1995) show that the earliest use of progressive marking by the three children is predominantly on activity verbs, which is in line with Bickerton’s claim that children acquiring English mark nonpunctuality 3. Bickerton (1981) made this prediction based on the observation (e.g., Brown 1973) that in the acquisition of English, irregular past attains acquisition criterion (i.e., 90% supplied in obligatory contexts) after progressive -ing, but before regular past tense -ed.
Yasuhiro Shirai
(i.e., lack of telicity) by -ing. This again, however, can be attributed to input: Approximately 60% of the mothers’ uses of progressive are with activity verbs. With regard to the use of progressive marking on stative verbs, Bickerton’s prediction is that children will not make the error of incorrectly attaching progressive marking to stative verbs because they know in advance the distinction between states and non-states (i.e., what Bickerton calls processes). Shirai (1994), however, shows that children do incorrectly attach progressive marking to stative verbs. It was found that one of the three children studied (Naomi) used progressive marking with state verbs, and some of the uses were obviously ungrammatical (e.g., *seeing light, *needing). Furthermore, the pattern of error is guided by parental input: the mother of this child was the only one who used progressive marking with stative verbs, which suggests that the pattern of error is guided by parental input. One additional piece of evidence for Bickerton’s SPD is the acquisition of Turkish. As noted briefly above, there are two past tense markers in Turkish, one for direct experiences (-dI), and the other for indirect experiences (-mIs). According to Bickerton (1981), who cites Slobin and Aksu (1980), Turkish children at early stages use -mIs and -dI to differentiate between static and dynamic events because the distinction is innately programmed. He argues that evidentiality is not part of the bioprogram, and therefore Turkish children misinterpret this direct vs. indirect past as the grammatical marking that makes the state-process distinction. Aksu-Koç (1988: 56), however, suggests an alternative, input-based account for this phenomenon, although she does not mention Bickerton’s SPD. She states: The alternative route whereby the -mIs particle acquires the function of indicating stativity in children’s speech can be hypothesized on the basis of a significant feature of baby-talk in Turkish. Although it is not an independently established fact, it is a strongly shared observation that adults talk to infants and young children in the evidential with -mIs. That is they are likely to comment on existing states or on resultant states that have come about in the child’s presence as well as otherwise, with the -mIs particle, seemingly violating their own rules.
If adults use -mIs for states, disregarding their own norm, it is not surprising that children mistake -mIs for a marker of stativity – even without any bioprogram. In sum, the evidence from language acquisition that Bickerton discussed to support his bioprogram hypothesis can be accounted for by input-based learning. The strong initial correlation between punctual/telic verbs and past marking, and that between progressive/imperfective marking and atelic verbs, which Bickerton interpreted as children’s bioprogram-based marking of the punctual-nonpunctual distinction, can instead be attributed to the skewed distribution in the input. The lack of overextension of stative verbs, which supports Bickerton’s state-process hypothesis, was not observed in Shirai (1994), and incorrect use of stative progressive
The semantic basis of grammatical development
was attributed to input. Next I will discuss a possible mechanism behind such a learning pattern. 3. Distributional learning and prototype-based initial representations Although the importance of input and particular organizations of target linguistic structures for acquisition studies has been stressed by many researchers (e.g., Bowerman 1985; Ochs 1985; Clancy 1989; Lieven 1994), the actual analysis of how input contributes to the initial representation of linguistic forms has not been systematically attempted until recently. Budwig (1996: 143) notes that although language acquisition research has shown that children link the use of particular linguistic forms with particular meanings/functions in ways not necessarily identical with the adult model, little investigation has been undertaken regarding the basis for such linkage. Budwig’s studies are particularly relevant to the present discussion in that she started out from the position that children will associate particular linguistic forms with some conceptual units that are cognitively salient (Budwig 1986, 1989). It was indeed found that three of the six children she studied associated the first person nominative pronoun I with low agentivity and control, while associating the genitive my and accusative me with high agentivity and control. More specifically, these children used my/me in situations where they act as prototypical agent, often by attempting to use language to bring about change (i.e., to control the situation around the child), as in Me open that.4 On the other hand, I was often used with stative verbs, which are low in agentivity and control, in that stative verbs normally do not refer to agentive actions. Budwig (1989) suggested that these children are marking agentivity and control, which are closely related to the linguistic notion of transitivity (Hopper & Thompson 1980), by using distinct forms, disregarding the adult model (see also Clahsen 1986, in which German children mark low transitivity by the -t inflection on the verb). This is in line with Slobin’s (1985) Basic Child Grammar, which suggests that children use specific linguistic forms to mark a scene involving prototypical agents. Budwig (1996), however, investigated the input to these three children, and found that 59% of caretakers’ uses of I involved a state verb, such as I like..., I think... Also, 61% of their uses of I are associated with non-control acts. This means that there was a high correlation of stativity, which is low in agentivity and control,
4. The issue of why children make such errors is still controversial (e.g., Schütze 1999, 2001, Rispoli 2002, 2005).
Yasuhiro Shirai
with the caretakers’ use of I. Thus, Budwig suggests that the input distribution may be the source of children’s acquisitional pattern.5 The convergence of figures between Shirai’s and Budwig’s studies – 60% – is suggestive. Shirai’s study also suggests the possibility that when 60% of a particular linguistic form is associated with one particular meaning/function, then children may make a strong connection between the two, as if the association was absolute; about 60% of past tense morphology and progressive morphology was used with achievement ([+punctual, +telic]) and with activity ([+dynamic, -telic]), respectively, in the mothers’ speech, and children initially restricted these morphological markings to those types of verbs almost exclusively. In the case of Italian, the distributional bias was more dramatic; perfective past is used with telic verbs more than 95% in adult speech, which surely would contribute to the initial restriction of perfective past to telic verbs by children (Antinucci & Miller 1976), which Bickerton attributed to a bioprogram. I argue that the mechanism behind this input-driven early restriction is that of distributional analysis and prototype formation (see Shirai & Andersen 1995; Andersen & Shirai 1996). Children are not just passively imitating what caretakers say. They actively reorganize their linguistic representations based on the distributional information in the input, and create the initial prototype. In the case of tense-aspect morphology, children appear to be guided by lexical aspectual features such as telicity and punctuality in creating the initial prototype, while in the case of Budwig’s pronominal case acquisition, the three children are guided by the notion of agentivity and control. When the initial prototype does not coincide with the adult model, it results in underextension (past, -ing, and I), and/or overextension (My/me). If we assume this type of learning mechanism on the part of children, we do not need to posit an innate bioprogram to explain the pattern of past tense acquisition in Bronckart and Sinclair (1973) and Antinucci and Miller (1976), which Bickerton used as supporting evidence for his bioprogram hypothesis. Furthermore, this input-based prototype formation has been tested by a simulation utilizing a connectionist network. Li and Shirai (2000, Ch.7) report a simulation of the acquisition of English as a first language by a self-organizing network, which showed a pattern of acquisition very similar to that observed in human children. This is important in that the network did not use “back-propagation” in which the network is ‘taught’ how its output differs from the target. Thus, by using unsupervised learning for this simulation, Li and Shirai bypassed the no-negative evidence 5. Slobin himself has also changed his position regarding Basic Child Grammar, and now doubts that there are any special notions that are universally grammaticizable (e.g., Slobin 1997).
The semantic basis of grammatical development
problem. The fact that this input-based prototype formation model can be implemented on a connectionist network without negative evidence suggests that it is a realistic and viable model of grammatical development. Now, let us turn to other areas of grammatical development to see the applicability of this model. There are four other linguistic domains I would like to discuss where this prototype hypothesis works well. 4. Other evidence for input-driven acquisition 4.1
Causative morpheme -sase in Japanese
Shirai, Miyata, Naka and Sakazaki (2000, 2001) found that children’s early use of the Japanese causative -sase is mainly restricted to ‘indirect causation’ such as permissive and assistive causation, although in the adult norm, it can be used for direct causation of the manipulative/directive type. This initial restriction gradually is weakened as children grow older. When we looked at the input, it was found that the causative morphology in the caretaker’s speech was mostly used for permissive and assistive causation. 4.2
Conditionals in Japanese and Korean
Akatsuka and Clancy (1993) report that Japanese and Korean children, when they start to use conditionals, begin with what they called D-conditionals to express order, prohibition and permission, such as Tabecha dame ‘(lit.)It’s no good if (you) eat’, and that the development of regular conditionals such as “If I eat this, I’ll have a stomachache” is slow. They attribute this to the high frequency of D-conditionals in the input. They further observed that one Korean child who started to use both types of conditionals at the same time had parents who used both at the same ratio. 4.3
Nominative case-marker drop in Japanese and Korean
Another interesting area where the effect of input is observed is the acquisition of nominative case markers in Japanese and Korean. Miyamoto, Wexler, Aikawa, and Miyagawa (1999) investigated the deletion of case markers in child Japanese, and found that children tend to omit nominative case markers more often for the subject NPs of unaccusative verbs than otherwise. Lee and Wexler (2001) analyzed the data from Korean child language more systematically, and also found that the subject NPs of unaccusative verbs have higher rate of nominative case omission. What is interesting is that they also found that adults’ speech addressed to children is characterized
Yasuhiro Shirai
by a higher rate of nominative case omission than adult-directed speech. 6 Thus, it appears that instead of the deficient grammar (i.e., the lack of A-chain formation) as these researchers suggested, children’s pattern of nominative-case marker omission seems to be driven by input-based form-meaning association. 4.4
Relative clause constructions in Japanese
Finally, Ozeki and Shirai (2010) found that the head nouns of early relative clauses that Japanese children use are mostly generic nouns and pro-forms that refer to inanimate entities, such as mono ‘thing’, tokoro ‘place’, -no ‘one’, (72% for 2-yearolds) and the modifying clauses involve mostly predicates referring to stative/generic attributes of the head noun modified as in yakyuu suru mono ‘the thing with which (people) play baseball’ (73% for 2 year-olds). These percentages closely reflect maternal speech addressed to children (54% and 73%, respectively, for all ages combined), whereas distribution in adult-adult discourse is quite different (28% and 49%, respectively). All these studies indicate that children’s early knowledge of grammatical structures is very different from that of adults. In particular children are semantically restricted compared to adults, and it appears that the formation of such restricted semantic prototypes is heavily influenced by the input they are exposed to. 5. A methodological objection to the prototype model One common objection to this type of prototype model is that it may reflect, not the child’s competence, but just a discourse context in which the child’s spontaneous conversation is recorded. In other words, it is just the demand of the discourse context in which children are engaged in conversation that determined the semantic bias in children’s use of grammatical forms, not because their semantic representation is distinct from that of adults. Weist (1989) makes such an argument to account for skewed distribution in the use of tense-aspect morphology in Polish. This is a valid criticism, and the real test of whether the skewed distribution is purely based on discourse factors must come from experimental studies. Regarding tense-aspect, comprehension studies (e.g., Li & Bowerman 1998 for Chinese, and Stoll 1998 for Russian) show that children’s competence is in fact limited, and that children have higher comprehension scores in prototypical combinations rather than non-prototypical combinations. In the domain of causative morphology in Japanese, Morikawa and Ono’s (1997) production experiment shows that 6. Mays and Ono (1993) also report that the pattern of use and omission of nominative marker -ga by a boy acquiring Japanese was strikingly similar to that observed in the input.
The semantic basis of grammatical development
children, and especially younger children, have more difficulty using -sase for direct causation than indirect causation. These studies suggest that children’s restricted production patterns also reflect their restricted semantic representation. 6. Input vs. innateness in language acquisition The observations discussed in this paper considerably weaken the innateness hypothesis espoused by theories of generative grammar. As is well known, Chomsky’s theory of Universal Grammar presupposes innate linguistic principles to solve the logical problem of language acquisition. That is, children can uniformly acquire a highly complex grammatical system of language without negative evidence. Language acquisition is possible only if children are endowed with innate linguistic knowledge, the theory goes. Much of generative research on language acquisition, therefore, investigates linguistic items that are relevant to this innateness hypothesis. In this paper, I discuss the proposal by Stephen Crain, who is one of the leading researchers in generative acquisition research (see also Crain & Wexler 1999). Crain and Thornton (1998) specifically argue that innate principles (a) emerge early (b) are universal, and (c) appear without decisive evidence from the environment. In chapter five of their book, Crain and Thornton (henceforth C & T) argue against the models of language acquisition that primarily rely on input (what they call the ‘input-matching model’). As examples of this model, they discuss Bates and MacWhinney’s (1989) Competition Model and Hirsh-Pasek and Golinkoff ’s (1996) Coalition Model. Tomasello’s usage-based model (Tomasello 2003) as well as the prototype model proposed in this paper is also considered to be a variant of the input-matching model. C & T argue that the input-matching model is inadequate because it cannot explain some facts of language acquisition. In particular, they claim that “the most compelling evidence of innateness probably comes from the observation that children sometimes make nonadult grammatical hypotheses” (C & T 1998: 37). As an example, they discuss “the medial wh-” – an unnecessary repetition of a wh- phrase within a sentence. For example:
*What do you think what pigs eat?
They argue that the input-matching model would be hard-pressed to provide an account of this phenomenon. However, this is not necessarily problematic for the input-matching model. This can be treated as an example of incorrect segmentation (or chunking) of linguistic units, of a sort that is often observed during language acquisition. Presumably, the children first uttered the string [what do you think], which they
Yasuhiro Shirai
encounter frequently, and then added [what pigs eat] to convey their intended meaning, i.e., to ask what pigs eat. What C &T are missing is the fact that children do not merely imitate what they hear, but actively reorganize linguistic data addressed to them, creating their own syntactic and semantic generalizations. If we accept this simple fact, children’s non-adult hypotheses are not problematic for the input-matching model at all. C & T’s interpretation of such non-adult hypotheses is that they are instantiations of UG. The medial wh-, for example, can be found in some dialects of German, and thus is part of the inventory of UG, they argue. But such an argument begs the question of which types of children’s errors are UG-driven, and which are not, because we do not have an exhaustive inventory of UG yet. Should we consider that children’s non-adult hypotheses concerning Japanese relative clauses (discussed in 4.4.) part of UG? Is there such a language where relative clauses are restricted to stative/generic predicates, as has been found in child Japanese (Ozeki & Shirai 2010)? Should we consider that the restriction to stative/ generic predicates in early L1 relative clauses is part of UG? C & T would also have difficulty explaining individual differences among children. They state: Clearly... different children encounter different input. This observation leads us to expect different children to adopt different hypotheses assuming that they are basing their hypothesis on experience. On the other hand, if all children in a linguistic community adopt the same linguistic hypotheses, then we can infer that language development is guided by innate principles, and not by learning-theoretic mechanisms (C & T 1998: 37).
Recall that three children in Shirai (1994) and four children in Akatsuka and Clancy (1993) showed individual differences in their acquisition of stative progressives and conditionals (respectively) as a consequence of different input distribution they received. This further indicates that the development of grammatical functors is guided by learning-theoretic mechanisms, not by innate principles. 7. Prototypes vs. modularity Newmeyer (1983: 13–14) specifically argued against the theory of the acquisition of syntax grounded in ‘semantic prototypes’ to support the autonomy of formal grammar. For the formal grammatical theory, syntactic categories are classical categories – past tense is past tense, and progressive aspect is progressive aspect, without internal structure; thus there is no good member or bad member of a particular linguistic category. The prototype theory of linguistic representation squarely goes against this assumption, and is at odds with the autonomy of syntactic
The semantic basis of grammatical development
structure from semantics. Therefore, the notion that children are guided in their learning of syntactic constructions by the semantically prototypical members of the class is problematic for the formal grammatical theory. Newymeyer based his arguments against the prototype theory on Maratsos and Chakley (1980), an influential paper which advocated the acquisition of syntax based on distributional information. Maratsos and Chakley’s argument against the prototype model is mainly based on the observation that it is extremely rare that children make categorical errors such as *He is nicing to them (instead of he is being nice to them) even though the prototype theory predicts that somewhat verblike adjectives (e.g., nice) will be miscategorized as verbs. This prediction by Maratsos and Chakley overlooks the fact that the prototype model of acquisition is essentially an underextension model, i.e. the child’s initial semantic representation is “restricted” to prototypical ones only, and will be gradually expanded to approximate the adult norm, and that therefore, overuse is rare. During the process of acquisition, sometimes semantic overextensions do occur, but these are also driven by semantic prototypes, as exemplified in the study of tense-aspect acquisition by Shirai (1993, 1994). Therefore, the empirical basis of the prototype formation model is solid, and it casts doubt on the modular theory of grammatical development where children are assumed to acquire syntactic categories as classical categories without regard to their internal structure. 8. What is the nature of the grammar that children acquire? The proponents of the innateness hypothesis generally assume a modular theory of grammar, in which the distinction between lexical and grammatical categories is of paramount importance. Pinker (1999), for example, proposes the words-andrules theory of language, in which words are manipulated by rules, the former being learned through associative memory while the latter are manipulated by productive computational mechanisms. This approach, although elegant in some ways, cannot deal with the continuous nature of lexical and grammatical categories observed in language. As grammaticization research shows, it is extremely difficult to draw a line between lexical and grammatical categories, since it is a universal of language change that lexical items grammaticize into functional elements and that therefore any language exhibits what Bybee (1985) calls the lexical-derivational-inflectional continuum. Furthermore, the lexical-grammatical dichotomy approach cannot deal with somewhat unpredictable constraints on grammatical operations observed in language (Bolinger & Sears 1981; Pawley & Syder 1983). By making a clear dichotomy
Yasuhiro Shirai
between words and rules, we lose sight of important facts of language, which consists of numerous cases of semi-productive phenomena (see also Jackendoff & Pinker (2005). Following usage-based, functional-typological-cognitive approaches to grammar (e.g., Langacker 2000, Goldberg 1995, Croft 2001, Bybee & McClleland 2005), I assume that the grammar that children need to acquire consists of large numbers of networks of constructions with different degrees of productivity. At the highest end of productivity is what has been considered to be “rules” while at the lowest end are words, and in between there are numerous semi-productive constructions such as idioms, formulas, and derivational morphemes (see Tomasello 2003 for an elaborated scenario of how such grammar is acquired by children through inputdriven pattern finding). 9. Conclusion In the field of first language acquisition, the role of input in grammatical development has generally been considered to be unimportant (but see Bowerman 1985). There appear to be two main reasons for this. First, early correlational research such as Brown (1973) and Newport, Gleitman and Gleitman (1977) showed that the acquisition of grammatical morphemes and constructions did not correlate with the frequency of these forms in caretaker speech. Second, influential theories of language acquisition did not place importance on input. For example, Chomsky’s Universal Grammar emphasized the grammatical principles that children are innately endowed with, and treated input only as trigger of Universal Grammar, whereas Piaget’s constructivism and Slobin’s Basic Child Grammar stressed the importance of the conceptual basis that children already have on which to map grammatical/linguistic structures. However, research on the effect of input is not very useful if we only investigate correlations between input frequency and acquisition of various grammatical structures, as was done by Brown (1973) and Newport et al. (1977); rather, we should look at the effect of input within a specific linguistic domain, as Valian (1999) pointed out. To understand the role of input in grammatical development, we need more focused research of the type discussed in this paper, which investigates the relationship between input and output in particular linguistic domains.7 That will shed new light to the mechanism of language acquisition, which in turn will help us understand the nature of grammar. 7. Recently, researchers who subscribe to the usage-based theory of language acquisition (Tomasello 2003) have embarked on just such research, looking into the relationship between input data and children’s language within particular linguistic domains (Cameron-Faulkner, Lieven & Tomasello 2003 for various constructions such as wh-questions, copulas and transitives, Deissel & Tomasello 2000 for relative clauses, Rowland, Pine, Lieven & Theakston 2003 for whquestions, Theakston, Lieven, Pine & Rowland 2004 for verbs and their argument structure).
The semantic basis of grammatical development
References Akatsuka, Noriko & Clancy, Patricia M. 1993. Conditionality and deontic modality in Japanese and Korean: Evidence from emergence of conditionals. Japanese/ Korean Linguistics, Vol. 2, Patricia M. Clancy (ed.), 177–192. Stanford CA: CSLI. Aksu-Koç, Ayhan. 1988. The Acquisition of Aspect and Modality: The Case of Past Reference in Turkish. Cambridge: CUP. DOI: 10.1017/CBO9780511554353 Andersen, Roger W. 1990. Seminar on the acquisition of tense and aspect. Unpublished lecture, University of California at Los Angeles. Andersen, Roger W. 1993. Four operating principles and input distribution as explanations for underdeveloped and mature morphological systems. In Progression and Regression in Language, Kenneth Hyltenstam & Åke Viborg (eds), 309–339. Cambridge: CUP. Andersen, Roger W. & Shirai, Yasuhiro. 1996. Primacy of aspect in first and second language acquisition: The pidgin/creole connection. In Handbook of Second Language Acquisition, William C. Ritchie & Tej K. Bhatia (eds). San Diego CA: Academic Press. Antinucci, Francesco & Miller, Ruth. 1976. How children talk about what happened. Journal of Child Language 3: 169–189. DOI: 10.1017/S0305000900001434 Bates, Elizabeth & MacWhinney, Brian. 1989. Functionalism and the competition model. In The Cross-linguistic Study of Sentence Processing, Brian MacWhinney & Elizabeth Bates (eds). Hillsdale NJ: Lawrence Erlbaum Associates. Bickerton, Derek. 1981. Roots of Language. Ann Arbor MI: Karoma. Bickerton, Derek. 1984. The language bioprogram hypothesis. The Behavioral and Brain Sciences 7: 173–188. DOI: 10.1017/S0140525X00044149 Bickerton, Derek. 1999. Creole languages, the Language Bioprogram Hypothesis, and language acquisition. In Handbook of Child Language Acquisition, William C. Ritchie & Tej K. Bhatia (eds), 195–220. San Diego CA: Academic Press. Bolinger, Dwight & Sears, Donald A. 1981. Aspects of Language, 3rd edn. New York NY: Harcourt, Brace Jovanovich. Bowerman, Melissa. 1985. What shapes children’s grammars? In The Crosslinguistic Study of Language Acquisition, Vol. 2: Theoretical Issues, Dan Isaac Slobin (ed.), 1257–1319. Hillsdale NJ: Lawrence Erlbaum Associates. Bronckart, J. Paul & Sinclair, Hermina. 1973. Time, tense, and aspect. Cognition 2: 107–130. DOI: 10.1016/0010-0277(72)90032-7 Brown, Roger. 1973. A First Language: The Early Stages. Cambridge MA: Harvard University Press. DOI: 10.4159/harvard.9780674732469 Budwig, Nancy. 1986. Agentivity and Control in Early Child Language. PhD dissertation, University of California at Berkeley. Budwig, Nancy. 1989. The linguistic marking of agentivity and control in child language. Journal of Child Language 16: 263–284. DOI: 10.1017/S0305000900010412 Budwig, Nancy. 1996. What influences children’s patterning of forms and functions in early child language? In Social Interaction, Social Cognition and Language: Essays in Honor of Susan Ervin-Tripp, Dan Isaac Slobin, Julie Gerhardt, Amy Kyratzis & Jiansheng Guo (eds, 143–156). Mahwah NJ: Lawrence Erlbaum Associates. Bybee, Joan L. 1985. Morphology [Typological Studies in Language 9]. Amsterdam: John Benjamins. DOI: 10.1075/tsl.9
Yasuhiro Shirai Bybee, Joan L. & McClelland, James L. 2005. Alternatives to the combinatorial paradigm of linguistic theory based on domain general principles of human cognition. The Linguistic Review 22: 381–410. DOI: 10.1515/tlir.2005.22.2-4.381 Cameron-Faulkner, Thea, Lieven, Elena & Tomasello, Michael. 2003. A construction based analysis of child directed speech. Cognitive Science 27: 843–873. DOI: 10.1207/s15516709 cog2706_2 Clahsen, Harald. 1986. Verb inflections in German child language: Acquisition of agreement markings and the functions they encode. Linguistics 24: 79–121. DOI: 10.1515/ling.1986. 24.1.79 Clancy, Patricia. 1989. Form and function in the acquisition of Korean wh-questions. Journal of Child Language 16: 323–347. DOI: 10.1017/S0305000900010448 Crain, Stephen & Thornton, Rosalind. 1998. Investigations in Universal Grammar: A Guide to Experiments in the Acquisition of Syntax and Semantics. Cambridge MA: The MIT Press. Crain, Stephen & Wexler, Kenneth. 1999. Methodology in the study of language acquisition: A modular approach. In Handbook of Language Acquisition, William C. Ritchie & Tej K. Bhatia (eds), 387–425. New York NY: Academic Press. Croft, William. 2001. Radical Construction Grammar. Oxford: OUP. DOI: 10.1093/acprof:oso/9 780198299554.001.0001 Diessel, Holger & Tomasello, Michael. 2000. The development of relative clauses in spontaneous child speech. Cognitive Linguistics 11: 131–151. Goldberg, Adele. 1995. Constructions. Chicago IL: University of Chicago Press. Hirsh-Pasek, Kathy & Golinkoff, Roberta Michnick. 1996. The Origins of Grammar: Evidence from Early Language Comprehension. Cambridge MA: The MIT Press. Hopper, Paul J. & Thompson, Sandra A. 1980. Transitivity in grammar and discourse. Language 56: 251–299. Jackendoff, Ray & Pinker, Steven. 2005. The nature of the language faculty and its implications for evolution of language: Reply to Fitch, Hauser & Chomsky. Cognition 97, 211–225. Kuczaj, Stan A. 1976. -ing, -s, and -ed: A Study of the Acquisition of Certain Verb Inflections. PhD dissertation, University of Minnesota. Kuczaj, Stan A. 1978. Why do children fail to overgeneralize the progressive inflection? Journal of Child Language 5: 167–171. DOI: 10.1017/S0305000900002026 Langacker, Ronald. 2000. A dynamic usage-based model. In Usage-Based Models of Language, Michael Barlow & Suzanne Kemmer (eds), 1–63. Stanford CA: CSLI. Lee, Hyeonjin & Wexler, Kenneth. 2001. Nominative case omission and unaccusatives in Korean acquisition. In Selected Papers from the Twelfth International Conference on Korean Linguistics, Hee-Don Ahn & Nam-Kil Kim (eds), 263–279. Seoul: Kyungjin Munhwasa. Leone, P. 1990. Tense and aspect in Italian: A bias in the distribution of verbal inherent semantics and verbal morphology. Unpublished paper, Applied Linguistics, University of California, Los Angeles. Li, Ping & Bowerman, Melissa. 1998. The acquisition of lexical and grammatical aspect in Chinese. First Language 18: 311–350. DOI: 10.1177/014272379801805404 Li, Ping & Shirai, Yasuhiro. 2000. The Acquisition of Lexical and Grammatical Aspect. Berlin: Mouton de Gruyter. DOI: 10.1515/9783110800715 Lieven, Elena V. M. 1994. Crosslinguistic and crosscultural aspects of language addressed to children. In Input and Interaction in Language Acquisition, Clare Gallaway & Brian J. Richards (eds), 56–73. Cambridge: CUP. DOI: 10.1017/CBO9780511620690.005
The semantic basis of grammatical development Maratsos, Michael P. & Chalkley, Mary Anne. 1980. The internal language of children’s syntax: The ontogenesis and representation of syntactic categories. In Children’s Language, Vol. 2, Keith E. Nelson (ed.), 127–214. New York NY: Gardner Press Mayes, Patricia & Ono, Tsuyoshi. 1993. The acquisition of the Japanese subject marker ga and its theoretical implications. In Japanese/Korean Linguistics, Vol. 2, Patricia M. Clancy (ed.), 239–247 Stanford CA: CSLI. Miyamoto, Edson T., Wexler, Ken, Aikawa, Takako & Miyagawa, Shigeru. 1999. Case-dropping and unaccusatives in Japanese acquisition. In Proceedings of the 23rd annual Boston University Conference on Language Development, Annabel Greenhill, Heather Littlefield & Cheryl Tano (eds), 443–452. Somerville MA: Cascadilla Press. Morikawa, H. & Ono, S. 1997. Children’s understanding of causatives in Japanese: A cross- sectional study. Paper presented at the SRCD Biennial Meeting, Washington DC, April 2. Newmeyer, Frederick. 1983. Grammatical Theory: Its Limits and Possibilities. Chicago IL: University of Chicago Press. Newport, Elissa, Gleitman, Henry & Gleitman, Lila. 1977. “Mother, I’d rather do it myself ”: Some effects and non-effects of maternal speech style. In Talking to children: Language Input and Acquisition, Catherine E. Snow & Charles A. Ferguson (eds), 109–149. Cambridge: CUP. Ochs, Elinor. 1985. Variation and error: A sociolinguistic approach to language acquisition. In The Crosslinguistic Study of Language Acquisition, Vol. 1: The Data. Daniel Isaac Slobin (ed.), 783–838. Hillsdale NJ: Lawrence Erlbaum Associates. Ozeki, Hiromi & Shirai, Yasuhiro. 2010. Semantic bias in the acquisition of relative clauses in Japanese. Journal of Child Language 37: 197–215. DOI: 10.1017/S0305000909009489 Pinker, Stephen. 1999. Words and Rules. New York NY: Harper Collins. Pawley, Andrew & Syder, Frances. 1983. Two puzzles for linguistic theory: Native-like selection and native-like fluency. In Language and communication, Jack C. Richards & Richard Schmidt (eds), 191–226. London: Longman. Rispoli, Matthew. 2002. Theory and methods in the study of the development of case and agreement: A response to Schutze. Journal of Child Language 29: 151–159. DOI: 10.1017/ S0305000901004962 Rispoli, Matthew. 2005. When children reach beyond their grasp: Why some children make pronoun case errors and others don’t. Journal of Child Language 32: 93–116. DOI: 10.1017/ S0305000904006658 Rowland, Caroline F., Pine, Julian M., Lieven, Elena V. M. & Theakston, Anna L. 2003. Determinants of acquisition order in wh-questions: Re-evaluating the role of caregiver speech. Journal of Child Language 30: 609–635. DOI: 10.1017/S0305000903005695 Schütze, Carson T. 1999. Different rates of pronoun case error: Comments on Rispoli (1998). Journal of Child Language 26: 749–755. DOI: 10.1017/S0305000999003992 Schütze, Carson T. 2001. Productive inventory and case/agreement contingencies: A methodological note on Rispoli (1999). Journal of Child Language 28: 507–515. DOI: 10.1017/ S0305000901004731 Shirai, Yasuhiro. 1991. Primacy of Aspect in Language Acquisition: Simplified Input and Prototype. PhD dissertation, University of California, Los Angeles. Shirai, Yasuhiro. 1993. Inherent aspect and the acquisition of tense/aspect morphology in Japanese. In Argument Structure: Its Syntax and Acquisition, Heizo Nakajima & Yukio Otsu (ed.), 67–82. Tokyo: Kaitakusha.
Yasuhiro Shirai Shirai, Yasuhiro. 1994. On the overgeneralization of progressive marking on stative verbs: Bioprogram or input? First Language 14: 67–82. DOI: 10.1177/014272379401404005 Shirai, Yasuhiro & Andersen, Roger W. 1995. The acquisition of tense/aspect morphology: A prototype account. Language 71: 743–62. DOI: 10.2307/415743 Shirai, Yasuhiro, Miyata, Susanne, Naka, Norio, & Sakazaki, Yoshiko. 2000. The acquisition of causative morphology: Why does it correlate with the imperative? In Proceedings of the Thirtieth Annual Child Language Research Forum, Eve V. Clark (ed.), 87–94. Stanford CA: CSLI. Shirai, Yasuhiro, Miyata, Susanne, Naka, Norio, & Sakazaki, Yoshiko. 2001. The acquisition of causative morphology in Japanese: A prototype account. In Issues in East Asian Language Acquisition, Mineharu Nakayama (ed.), 183–203. Tokyo: Kurosio. Slobin, Dan Isaac. 1985. Crosslinguistic evidence for the Language-Making Capacity. In The crosslinguistic study of language acquisition. Vol. 2: Theoretical issues, Dan Isaac Slobin (ed.). Hillsdale NJ: Lawrence Erlbaum Associates. Slobin, Dan Isaac. 1997. The origins of grammaticizable notions: Beyond the individual mind. In The Crosslinguistic Study of Language Acquisition, Vol. 5: Expanding the Contexts, Dan Isaac Slobin (ed.), 265–323. Mahwah NJ: Lawrence Erlbaum Associates. Slobin, Dan Isaac & Aksu, Ayha A. 1980. Acquisition of Turkish. Ms. Cited in Bickerton 1981. Smith, Carlota S. 1997. The Parameters of Aspect, 2nd edn. Dordrecht: Kluwer. DOI: 10.1007/97894-011-5606-6 Stoll, Sabine. 1998. The role of Aktionsart in the acquisition of Russian aspect. First Language 18: 351–377. DOI: 10.1177/014272379801805405 Theakston, Anna L., Lieven, Elena V. M., Pine, Julian M. & Rowland, Caroline F. 2004. Semantic generality, input frequency and the acquisition of syntax. Journal of Child Language 31: 61–99. DOI: 10.1017/S0305000903005956 Tomasello, Michael. 2003. Constructing a Language: A Usage-based Theory of Language Acquisition. Cambridge MA: Harvard University Press. Valian, Viginia. 1999. Input and language acquisition. In Handbook of Child Language Acquisition, William C. Ritchie & Tej K. Bhatia (eds), 497–530. New York NY: Academic Press. Vendler, Zeno 1957. Verbs and times. The Philosophical Review 66: 143–160. DOI: 10.2307/ 2182371 Weist, Richard M. 1989. Aspects of the Roots of Language: Commentary on Cziko. First Language 9: 45–49. DOI: 10.1177/014272378900902504
Interchangeability of so-called interchangeable particles Corpus analysis of spatial markers, Ni and E Kaori Kabata Three different sets of empirical data are compared to examine the usage distributions of ni and e, two particles that have been traditionally considered to be interchangeable in marking allative relations. The data from the speaker’s judgment test are more consistent with the semantic analyses of the two particles, exhibiting interaction between semantic contexts and particle preference. However, the usage distributions in the spoken and written corpora exhibited a strong bias for ni regardless of the semantic environment. The discrepancy between what is generally considered to be ‘acceptable’ in Japanese grammar and the actual usage data stresses the impact of the empirical data in the further understanding of lexical representation.
1. Introduction Recent developments in corpus linguistics have demonstrated the magnitude of impact of usage data on linguistics research and consequently the understanding of grammar. In particular, in the area of cognitive linguistics, a usage-based approach to language, where grammatical rules and patterns are claimed to be abstracted from actual patterns of usage (e.g., Langacker 1987, 2000), corpus data present powerful empirical evidence for the important roles of frequency and conventionalization (e.g., Barlow & Kemmer (eds) 2000). Usage data also play an essential role in studies of grammaticalization and acquisition processes, which are considered closely relevant to the synchronic usage patterns of language, as demonstrated by Newman and Rice (2004). Moreover, corpus data, as a collection of actual uses, may challenge some of the previous claims about mental representations of lexical information, which have largely been based solely on a linguist’s native-speaker intuitions. An examination of natural conversational data in Japanese, for example, has revealed that sentential elements are often expressed
Kaori Kabata
after predicates, rather than before (Ono & Suzuki 1992) and that ga, which is generally considered a subject marker or a nominative case marker, is most often used to convey pragmatic functions, signaling new and/or unpredictable information (Ono, Thompson & Suzuki 2000). Similarly, Yamazaki (2002) has shown that omae, a word usually translated as the second-person pronoun ‘you’, is often used at the sentence-final position to convey speakers’ emotive factors. Despite the general consensus as to the importance of empirical data, including corpus data, in cognitive linguistics research, however, researchers have yet to resolve methodological issues involved in such approaches. As Sandra and Rice (1995) argue, the results of an experimental study are inevitably subject to taskeffects: While an off-line sorting task would tend to emphasize the differences among meanings, a similarity judgment task would more likely emphasize their resemblance. Furthermore, an on-line task, in which measurements are taken while participants are actually processing the meanings, would have a different focus from that of off-line tasks, which are generally believed to be a more indirect measure. Similarly, in corpus-based studies, various aspects of the corpus, including the modes of texts and participants, would inevitably influence the implications to be drawn from the analyses, as demonstrated by Masuda (2002). The primary goal of this paper is to show how different empirical data may lead to different conclusions for linguistic categorization of elements under investigation. The focus of this paper is the usage distributions of ni and e, which have been traditionally considered to be interchangeable in marking allative relations (e.g., Martin 2004). Three different sets of empirical data compared are; (i) data from a native speaker judgment test, (ii) data from a spoken corpus called CALLHOME, and (iii) data from a written corpus based on novels and essays. The results did not support the interchangeablility the two so-called interchangeable particles. While the data from the speaker’s judgment test were more consistent with the semantic analyses of the two particles, exhibiting interaction between semantic contexts and particle preference, the usage distributions in the spoken and written corpora exhibited a strong bias for ni regardless of the semantic environment. Moreover, metaphorical use of e in the form of e-to was frequently found only in the written texts, although e, when used by itself, was limited to the spatial expressions. By unveiling the discrepancy between what is generally considered to be ‘acceptable’ in Japanese grammar and the actual usage patterns found in corpus data, this study stresses the impact of the empirical data in the further understanding of lexical representation.
Interchangeability of so-called interchangeable particles
2. Semantic characterization of Ni vs. E Ni and e, which are both used to indicate a direction or a goal, have not attracted as much, if any, attention in Japanese linguistics, as have the locative marking ni and de (Kumashiro 1994; Masuda 2002), the causative marking ni and o (Ikegami 1986, 1987), or the subject marking ni and ga (Kabata 1999; Shibatani 1978). This lack of attention by linguists is probably due to the fact that they are simply thought to be interchangeable (e.g., Martin 2004). Kaiser, Ichikawa, Kobayashi, and Yamamoto, for example, state: With V of motion, e can be used instead of ni to mark a core case (dative of direction). Whereas ni indicates the goal of a motion, e is said to be concerned more with the direction towards the goal, but in practice the two are often interchangeable (2001: 134).
Such semantic overlap between a direction marker and a destination marker is actually far from surprising, since the two concepts are often subsumed under the term allative, which is defined by Svorou (1994) as following: If the lm [landmark]...is treated as the destination, then the tr [trajector] is moving in an allative motion in the direction of the lm. In this case, the implication is that the tr has started the movement with the intention of reaching the lm, and other things being equal, it will reach it (1994: 27) [emphasis added].
Many languages allow one grammatical morpheme to describe both relations, as shown in (1), in Maori. These examples are from Bauer (1993):1 (1) a. Ka hoki eeraa ki Rurunui t/a return those Rurunui ‘They returned to Rurunui.’
[goal]
b. E whakahorohoro ana au ki te kaainga t/a hurry t/a Isg the home ‘I’m hurrying home.’
[direction]
Despite the semantic overlap of the two particles in modern Japanese as discussed in the previous literature, they are regarded as having developed with two 1.
Below is the list of abbreviations used in this paper: acc accusative nom nominative all allative past past tense conn connective qt quotative cop copula t/a tense/aspect marker gen genitive top topic marker loc locative
Kaori Kabata
distinct functions. Although no direct record is available to indicate the origin of ni, because most of its uses are found in the earliest written records of Japanese, Kabata (2000) argues, based on cross-linguistics and circumstantial evidence, that one of its most basic senses is that of describing the spatial goal. The particle e is generally believed to have evolved out of the noun he ‘area, vicinity’, around 7th or 8th century, taking on the sense describing a direction of motion (e.g., Martin 2004; Yamaguchi & Akimoto 2001). Around the middle of Heian Era (794 – 1191) e started to indicate the endpoint of motion, and thus to overlap semantically with ni. Figure 1 provides a schematic illustration of the two distinct but related senses of the two markers, based on the previous literature. The basic sense of ni is the one illustrated in Figure 1b. The focus is on the endpoint of a motion, best described by goal-oriented verbs like tsuku ‘arrive’, hairu ‘enter’, and tsukeru ‘attach’. The basic sense of e, on the other hand, is to mark the direction of a motion whose focus is on the source or the path, as with the verbs like deru ‘exit’, susumu ‘proceed’, and mukau ‘head (for).’ Verbs like iku ‘go’ and kuru ‘go’ do not have strong focus on either the endpoint or the source/path, and are equally compatible with either type of marker. While e is generally considered as being limited to the marking of spatial relations, ni is associated with a wide array of functions marking both spatial and nonspatial relations, as shown in (2a) and (3a). However, when used in a prenominal phrase with the genitive marker no, e marks both spatial and non-spatial relationships, as shown in (2b) and (3b):
TR
LM
a. ALLATIVE
TR
LM
b. GOAL marker NI (focus on the endpoint)
Figure 1. Ambiguity of ALLATIVE marker
TR
LM
c. DIRECTION marker E (focus on the source/path)
Interchangeability of so-called interchangeable particles
(2) a. nihon ni/e shuppatsu-suru Japan all departure-do ‘(I am going to) depart for Japan.’ b. nihon *ni/e no shuppatsu Japan all gen departure ‘a departure for Japan.’ (3) a. kodomo ga oya ni/*e hanpatsu-suru child nom parents all rebel-do ‘A child rebels against his parents.’ b. kodomo no oya *ni/e no hanpatsu child gen parents all gen rebel ‘A child’s rebel against his parents’ Thus, previous literature indicates that the semantic distinction these two directional markers had originally has faded over time and e has extended its usage into the semantic field originally occupied by ni. However, those claims have been solely based on the researchers’ native intuition or at best unsystematic observation. As Kabata (in preparation) has found, the patterns of child acquisition of these two markers differ, with the directional marker ni emerging much earlier and being much more frequent than e. Assuming children’s language acquisition is much influenced by the adults’ input (e.g., Hallan 2001), one would expect to find some differences between the two markers in their usage patterns by native speakers. But how do they differ, if they do, and how can we explain such differences? Three sets of empirical data are compared to address this question. 3. Ni and E in the speaker judgment data 3.1
Procedure
20 native speakers of Japanese, including 4 male and 16 female, participated in the experiment.2 The participants were all studying at the University of Alberta, Canada, at the time of the experiment. The participants had been living outside of
2. This test was conducted as part of a comparative linguistics study in which I compared the semantic distributions of Japanese particles and those of Korean counterparts. The results were presented jointly with Jeong-Hwa Lee at the 6th CSDL conference, Rice University, Houston, TX, October 12–14, 2002.
Kaori Kabata
Japan for periods of time ranging from 5 months to 10 years, with the average being 2 years and 1 month. They all spoke Japanese on a daily basis.3 The stimuli for the study were 50 sentences (shown in Appendix with English translation), which were differentiated in terms of several semantic factors, including the verb types (goal-oriented, path-oriented, or neutral), the tense and aspect of the predicate (present perfect, past, or future), and the size of the location or the container (small, middle, or large). Five sets of stimuli were prepared by arranging the 50 sentences in five different random orders. Participants were asked to read the instruction carefully and provide personal information before starting the experiment. The participants’ task was to rate the acceptability of the two allative particles, e and ni, in each stimulus sentence. They were instructed to base their judgment on their native intuition and to proceed at their own pace. The acceptability scale ranged from ‘1’ to ‘5’, with the rating ‘1’ and ‘5’ indicate that participants judged ni and e as being exclusively acceptable respectively. The rating ‘3’ indicates that either particle is equally acceptable, and ‘2’ or ‘4’ that one was more preferred to the other, although both are acceptable. The obtained data were subject to further statistics analyses. 3.2
Results
The bar graph in Figure 2 illustrates the overall particle distribution of the 50 stimulus sentences presented in original Japanese in Appendix. A higher or lower score on the rating scale indicates stronger preference for e or ni by the participants. As can be seen in the graph, ni was generally preferred over e: The distribution of ratings ranged rather narrowly from 1.25 to 3.47, with the average score at 2.33.4 ANOVA tests revealed that the type of Japanese locomotive verbs – goaloriented verbs (e.g., tsuku ‘arrive’), source-oriented verbs (e.g., shuppatsusuru ‘depart’), or neutral type (e.g., iku ‘go’) – were found to interact with the choice of 3. The dialectal variation and deviation from so-called standard Japanese were pointed out during the discussion at the symposium. The author is aware that there are some dialectal variations in the usages of ni and e, as listed in Tokugawa (1989). Martin (2004: 46) also mentions that speakers in downtown Tokyo tend to use e in place of ni. However, in this paper any dialectal usage of ni and e, if any, were treated as individual differences in the data. This decision was made for the following reasons: (i) Variations in language use may exhibit dialectal differences, generational changes, etc., which are impossible to distinguish unless detailed studies are conducted, which is beyond the scope of the present study. (ii) The sentences presented in the speaker judgment task were written in what is generally considered to be standard Japanese, with which all the participants were familiar through formal educations they had received in Japan. 4. The Korean data were rather contrastive to the Japanese data, with the scores more widely distributed ranging from 1 (only ey is acceptable) to 5 (only (u)lo is acceptable). The average score for the Korean data was 2.80.
Interchangeability of so-called interchangeable particles ni 1
e 2
3
4
5
1 Taro went to America last week 2 Taro has gone to America 3 Taro will go to America 4 Taro left for America last week 5 Taro has left for America 6 Taro is going to leave for America tomorrow 7 Taro arrived in America last week 8 Taro has arrived in America 9 Taro will arrive in America tomorrow 10 Taro went to the airport last week 11 Taro has gone to the airport 12 Taro will go to the airport 13 Taro left for the airport last week 14 Taro has left for the airport 15 Taro is going to leave for the airport tomorrow 16 Taro arrived in the airport last week 17 Taro has arrived in the airport 18 Taro will arrive in the airport tomorrow 19 Taro went to school last week 20 Taro left for school tomorrow last week 21 Taro arrived in school last week 22 Taro went to work last week 23 Taro left for work tomorrow last week 24 Taro arrived in work last week 25 Taro put the car inside the warehouse 25 Taro is going to the car inside the warehouse 27 the car moved inside warehouse 28 the car is going to move inside warehouse 29 Taro took the car outside the warehouse 30 Taro is going to take the car outside the 31 the car moved outside warehouse 32 the car is going to move outside warehouse 33 Taro poured tea inside/into the cup 34 Taro is going to pour tea inside/into the cup 33 the coin fell inside/into cup 36 the coin is going to fall inside/into cup 37 Taro spill tea onto the table 38 Taro is going to spill tea onto the table 39 tea spilled onto the table 40 tea will spill outside of the cup 41 Nile river pours into the Mediterranean sea 42 Taro turned his chin upward 43 Taro turned his chin downward 44 Taro turned his chin leftward 45 California is in contact with Pacific ocean 46 my cheek came in contact with the floor 47 Taro’s hand came in contact with my face 48 Taro’s plam came in contact with my nose 49 Nevada is contiguous to Pacific ocean 50 Taro’s hand almost touched my face
Figure 2. Allative particle distribution in Japanese based on the 50 stimulus sentences
Kaori Kabata
particle at a significant level (F = 49.91, p< .01). The mean rating score was 1.64 for sentences with tsuku ‘arrive’ (S7-S9 and S16-S18), indicating a strong preference for ni. With iku ‘go’ and shuppatsusuru ‘depart,’ the preference for e was stronger, although even for the latter the preference was far from being exclusive. The mean rating scores were 2.9 for sentences with iku ‘go’ (S1-S3, and S10-S12) and 3.2 for those with shuppatsusuru ‘depart’ (S4-S6, and S13-S15) respectively. The size of the goal location or the tense or aspect did not show a significant interaction with the particle choice with any of these locomotive verbs. In sentences describing movement into a container – the container size used in the stimuli varied from small (e.g., a cup), to middle-size (a garage), to extra large (the Mediterranean Sea) – the size of the goal location was found to influence particle preference (F = 18.11, p< .01). The acceptability ratings for these sentences were 1.7 with small containers like a cup (S33-S36), 2.4 with middle-seized containers like a garage (S25–28), 2.6 with large containers like the Mediterranean Sea (S41), indicating that speakers’ perceived acceptability for e is higher with a larger container. The data also indicated that there were rather distinct patterns of particle preference with regard to the type of event, namely ‘turning’ (S42-S44) vs. ‘contact’ (S45–S50). In describing the former, described by mukeru ‘turn’ in the stimulus sentences, there was a strong preference for the particle e: The mean scores for S42-S44 was 2.32. In contrast, ni was judged as almost exclusively acceptable in the sentences with fureru ‘contact’. For S45–S50, the acceptability rating scores were 1.5. Neither the direction of the ‘turning’ movement (i.e., upward, sideways, etc.) nor the size of the location for ‘contact’ showed interaction with the particle preference. To summarize, the results of the speaker judgment test were consistent with the semantic analyses of the two particles in general. While ni is generally more preferred as a spatial goal marker by native speaker participants, the preference between the two spatial markers interact with various semantic factors including the type of locomotive verbs and the container size. These semantic factors are interpreted as setting the described event as either more goal-oriented (and more compatible with ni) or more source/path-oriented (more compatible with e). 4. Ni and E in spoken data 4.1
Methodology
The purpose of this corpus analysis was to examine the usage distributions of ni and e and determine whether they reflect the semantic characteristic of each
Interchangeability of so-called interchangeable particles
particle as discussed in the previous literature. Spoken speech data were collected from a corpus of telephone conversations from CALLHOME Japanese Speech, collected and released in 1997 by the Linguistic Data Consortium.5 The corpus consisted of 120 files of recorded phone conversations between the participants, who were native speakers residing in the United States, and their family members or close friends. The length of transcribed conversations, which were contiguous segments taken from approximately 30-minute-long phone calls, varied from 5 minutes to 10 minutes, totaling 1,100 minutes for the entire corpus. All utterances containing the particle e were extracted and analyzed. In the case of ni, since it is associated with variety of usages, only those that contain the allative usage describing movement in the spatial domain were subject to the analysis. Sentences that contain a verb that describe movement in the spatial domain but are used metaphorically (e.g., used with a non-spatial noun or noun phrase, such as ninenme ‘the second year’ as in ninenme ni hairu ‘enter the second year’) were also included in the data. The types of verbs and nouns that were used with the two particles were examined. Verbs were analyzed in term of the goal- or source-orientedness: while verbs like tsuku ‘arrive’, hairu ‘enter’, and oku ‘put/place’ are, in a neutral context, characterized as profiling the endpoint of the motion (i.e., the location towards which the object is moving), and are categorized as goal-oriented. Verbs like deru ‘exit’, shuppatsusuru ‘depart’, on the other hand, profile the location away from which the object is moving, and are categorized as source-oriented. Verbs that are not strongly associated with either the goal or the source of a movement were treated as neutral. 4.2
Results
There were only 78 instances of e, while 639 out of the total 3898 instances of ni were found to be used as a spatial goal marker, suggesting the dominance of ni and the limited use of e in describing spatial goal. Out of the total 78 instances of e, 15 instances appeared either in verbless utterances, as part of a fixed phrase tsugi-kara-tsugi-e ‘one after another’, or in a prenominal phrase ...e-no discussed above in (2) and (3). The remaining 63 instances of e were used with 22 different verbs, as shown in Table 1.6 Iku ‘go’, which is considered to be neutral in terms of goal- or source-orientedness, appeared with e 29 times, or 46% of all the instances. Kuru ‘come’, kaettekuru ‘come back’, which are 5.
Information about LDC corpora is available at .
6. The lists in the tables are for verbs and noun phrases that occurred with each particle in more than 2% of the total instances.
Kaori Kabata
Table 1. List of frequently co-occurring verbs with e and ni in the spoken corpus (Verbs that were appear on the both lists are indicated in bold) E
Ni
Token frequency Type frequency iku ‘go’ kuru ‘come’ kaetteiku ‘return’ deru ‘exit’ yoru ‘drop by’ okuru ‘send’ tsureteiku ‘take’
63 22 29 4 4 4 3 2 2
15 others verbs
15
Token frequency Type frequency iku ‘go’ kuru ‘come’ hairu ‘enter’ ireru ‘put in’ deru ‘exit’ denwasuru ‘phone’ kaeru ‘return’ okuru ‘send’ noru ‘get into’ tsuku ‘arrive’ kaku ‘write’
639 62 212 69 57 37 28 28 22 18 14 13 13
51 other verbs
156
also interpreted as a neutral type, were used with e 4 times each. Despite the assumed semantic compatibility, source-oriented verbs, such as deru ‘exit’, were found rather infrequently in only 6 instances. Goal-oriented verbs were also used very infrequently, with verbs like tsuku ‘arrive’ and oku ‘put down’ appearing in only 7 instances altogether. Sixty-two different verbs were found with ni, resulting in a token vs. type ratio at as high as 10: 1, far higher than that of e, suggesting that some of the verbs were used highly frequently with ni. In fact, ni was used with iku ‘go’ 212 times (33%), followed by kuru ‘come’ at 69 instances (11%). Goal-oriented verbs hairu ‘enter’ and ireru ‘put in’ produced 57 and 37 instances respectively, together constituting 15% of all the instances of ni, indicating its compatibility with sentences that focus on the endpoint of the motion. Other goal-oriented verbs found relatively frequently with ni included noru ‘get in/on’ and tsuku ‘arrive’. However, ni was also used with source-oriented verbs, such as deru ‘exit’, dasu ‘put/take out’ and dekakeru ‘go out’, altogether in 34 instances (5%), far more frequently than e, which appeared with them only in 6 instances. The particle e was found to be fairly limited in terms of the range of collocating NPs, as well as verbs. Only 27 different nouns were found in all 76 cases: Two instances where e was used in an idiomatic expression or in a prenominal expression were excluded from the analyses. The most frequent was the demonstrative locative
Interchangeability of so-called interchangeable particles
NP soko ‘there’ at 13 times (17%), followed by the interrogative locative NP doko ‘where’ at 12 times (16%). The combination of interrogative and demonstrative locative nouns provided 33 instances, or 43% of all instances of e. Moreover, of the 12 instances of doko ‘where’, 7 instances appeared with the verb iku ‘go’, forming a string doko e iku. E was also used with generic or unspecified locative nouns such as tokoro ‘place’, ... no hoo ‘in the direction of ...’, and mukoo ‘over there’. Two hundred and sixty different nouns were used in the 640 instances of ni. The most frequently used nouns included place NPs like nihon ‘Japan’ (34 instances) and amerika ‘America’ (22 instances), whose high frequency was not surprising since the phone calls were made between the States and Japan. Definite place nouns like uchi ‘home’ (33 instances), and gakko ‘school’ (20 instances) were also frequent in the data. Locative demonstrative NPs such as koko ‘here’, soko ‘there’, sotchi ‘there’, kochira or kotchi ‘here’ appeared 71 times (11%). Kochira and kotchi ‘here’ together comprised 28 instances, 21 of which were used with the verb kuru ‘come’ as in kotchi ni kuru ‘come this way’. This frequent use of ni with kuru was somewhat contrastive to the use of e with the verb, which appeared only 4 times, and only once with kotchi ‘here’ as in kotchi e kuru. Ni was used with various place nouns of various sizes, including countries (74 instances or 12%), states and towns (60 instances or 9%), and buildings and rooms (127 instances or 20%), as well as other concrete nouns like fuutoo ‘envelope’, saifu ‘wallet’, and reezooko ‘refrigerator. Ni was also used with various relational nouns such as ... no ue ‘top’, shita ‘beneath’, naka ’inside’, ushiro ‘back’, for a total of 33 instances, with 8 different nouns. Unspecified locative nouns like tokoro ‘place’ and ... no hoo ‘in the direction of ’ also appeared 15 and 14 times respectively. Metaphorical expressions containing temporal nouns like ni-nen-me ‘the second year’, and abstract nouns like jootai ‘situation’ and nakama ‘circle of people’ were also found, though with a limited range of verbs, such as hairu ‘enter’ and ochiiru ‘fall into’. In sum, an examination of e and ni in spoken data has indicated that, contrary to the expectation based on the semantic analysis, ni is much more frequently used as a goal marker regardless of the locative sizes or the verb types. Even souceoriented verbs like deru ‘exit’ was a lot more frequent with ni at 28 instances as opposed to 4 instances with e. Usage distribution of e was fairly restricted both in terms of the collocating verbs and nouns: the verb iku ‘go’ comprised over a third of the all the instances of e, and the interrogative locative NP doko ‘where’ and locative demonstrative NPs including soko ‘there’ kotchi ‘here‘ together counted over 40%. While ni also exhibited some of its frequent fixed patterns, such as ...ni iku, and kotchi ni kuru, it was used with a much more variety of verbs and NPs, suggesting its dominance as an allative marker.
Kaori Kabata
Table 2. List of frequently co-occurring NPs with e and ni in the spoken corpus E
Ni
Token frequency Type frequency soko ‘there’ doko ‘where’ tokoro ‘place’ kotchi ‘this way/here’ ...no hoo ‘the direction of ’’ mukoo ‘over there’ asoko ‘over there’ ..no soto ‘outside [of]’ amerika ‘America’ namikawa (place name)
76 27 13 12 7 6 4 3 2 2 2 2
17 other spatial nouns
23
Token frequency Type frequency uchi ‘home’ nihon ‘Japan’ kochira/kotchi ‘here’ amerika ‘America’ gakkoo ‘school’ tokoro ‘place’ soko ‘there’ ..no hoo ‘the direction of ’
640 260 34 34 28 22 20 19 18 13
227 other spatial nouns 14 non-spatial nouns
439 13
5. Ni and E in written text data 5.1
Methodology
Table 3 lists the ten novels and essays used for the written corpus. All the texts, taken on line from Aozora Bunko at http://www.aozora.gr.jp/, were written within the last two decades, and were thus expected to represent current usage of the two particles. Each instance of the two particles ni and e were extracted and analyzed in the same manner as with the CALLHOME data. 5.2
Results
The written corpus contained 565 instances of e, as opposed to 3013 instances of ni as a goal-marker, demonstrating that ni is more prevalent than e in the written texts as in the spoken texts, but that the preference did not appear as strong. The ratio of overall frequencies between e and ni was roughly 1 to 5 in the written corpus, as opposed to 1 to 8 in the spoken corpus. However, a closer examination of the data revealed that e is rather limited in its usage distribution, despite its overall frequency. The total 565 instances included 191 instances (34%) where e preceded no in the prenominal expression ...e-no, which appeared only once in the spoken data. There were also 11 instances of fixed expressions such as tsugi-kara-tsugi-e-to ‘one after another’ and doko-e-yara ‘somewhere (out there)’. The remaining 363 instances were subject to further analyses.
Interchangeability of so-called interchangeable particles
Table 3. List of novels and essays used for the written corpus Author
year
Title [English translation]
Akino, T. Hamano, S. Ichikawa, Y. Kareha. Koba, T. Koizumi, H. Kuwahara, I. Sano, R. Takahashi, Y.
1993 1998 2000 2001 1999 1999 1998 1993 1997
Yamagata, H.
1999
Rokku Nanajuunendai [Rock in the ‘70’s] Shintoshi Ongaku Nooto [A Music Note in New Towns] Hookago no Rock’n Roll [Rock’n Roll after School] Great Gatsby (Translation of The Great Gatsby by F.S. Fitzgerald) Sophia Mimizu Monogatari 2 [An Earthworm Story 2] Ningen no Kihon [The Foundation of Human Beings] Warera Rifutaa [We Weightlifters] Ongaku no Han-hooron Josetsu [Introduction to Anti-Music Methodology] Garan to Bazaaru (Translation of The Cathedral and Bazaar by E. Raymond)
Table 4. List of frequently co-occurring verbs with e and ni in the written corpus E
Ni
Token frequency Type frequency iku ‘go’ mukau ‘head (for) susumu ‘proceed’ hairu ‘enter’ modoru ‘return’ aruku ‘walk’ yattekuru ‘come’
341 165 34 19 10 9 8 8 8
141 other spatial VPs 17 non-spatial VPs
208 37
Token frequency Type frequency hairu ‘enter’ mukau ‘head (for)’ iku ‘go’ deru ‘exit’ kuru ‘come’ tatsu ‘stand’
3013 710 136 104 101 91 68 65
704 other spatial VPs
2448
Among the 363 instances of e were 22 verbless sentences. As shown in Table 4, the remaining 341 instances of e were found with 165 different verbs, resulting in a token vs. type ratio of 2 to 1, indicating a more even usage distribution of e with verbs than in the spoken data, where the token vs. type ratio was 3 to 1. The verb iku ‘go’ was again the most frequent in the written data, at 34 instances (10%, as opposed to 46% in the spoken data). Path-focused verbs like mukau ‘head (for)’ and susumu ‘proceed’ were also frequent at 19 and 10 instances respectively. In contrast to the spoken data, e was used with a variety of goal-focused, neutral,
Kaori Kabata
Table 5. List of frequently co-occurring NPs with e and ni in the written corpus E
Ni
Token frequency Type frequency ..(no) hoo ‘the direction of ..(no) naka ‘inside [of]’ gakkoo ‘school’ soko ‘there’ tokoro ‘place’ doko ‘where’
363 226 19 15 13 6 5 5
172 other concrete NPs 4 temporal NPs 44 abstract NPs
247 5 48
Token frequency Type frequency ..(no) naka ‘inside [of]’ ..(no) ue ‘on top [of]’
3013 1354 99 66
1088 other concrete NPs 37 temporal NPs 227 abstract NPs
2457 70 321
and source/path-focused verbs in the written data. E was used with 35 goalfocused verbs including hairu ‘enter’ (9 instances) and modoru ‘return’ (8 instances). E was also used with verbs that describe non-spatial events such as changes in the cognitive domain. They include kaeru ‘change’ (3 instances), henkasuru ‘change’ (3 instances), seechoosuru ‘grow’ (3 instances), and tenkansuru ‘convert’ (2 instances). (4) presents some of the examples: (4) a. ... sore ga subarashii mono e to seechooshi-ta. that nom wonderful thing all qt grow-past ‘(and) it grew into something wonderful.’ b.
Fankii wa futatabi wadai wo ta e tenji-yoo to ki wo Fanky top again topic acc else all change-will qt be tsukau. anxious. ‘Fanky is anxious to try to change the topic again.’
Turning to the distribution of NPs, 363 instances of e appeared with 226 different NPs. The most frequent was (... no) hoo ‘the direction [of ...]’ (19 times), followed by (... no) naka ‘inside’ (15 times) and gakkoo ‘school’ (13 times). Demonstrative NPs, such as soko ‘there’, koko ‘here’, and kochira ‘here/this way’, and the interrogative locative NP, doko ‘where’, which together occupied 43% of the total instances of e in the spoken data, appeared only 14 times (less than 4%) in the written data. Moreover, unlike the spoken data, the written data contained 48 instances of e found with 44 abstract nouns, many of which were used with verbs describing
Interchangeability of so-called interchangeable particles
spatial motions, such as oikomu ‘chase’, michibiku ‘lead’, and tsuresaru ‘take (somebody) away’. Some of the examples are shown in (5): (5) a. ...kyokugenjookyoo e jibun o oikomi... ultimate state all self acc chase ‘... chase himself into the ultimate state, and...” b. mata tsugi no mokuhyoo e mukat-te-iku. again next gen goal all head-conn-go ‘(then) he moves on again to head for the next goal.’ However, a closer examination revealed many of the instances of e used with goalfocused verbs, and most of those with non-spatial verbs, appeared before the particle to as in ...e-to V, as illustrated in (4a) above. In fact, 142 instances of e were followed by to in the written data, comprising 42% of all the instances where e was used with a verb. Among the 142 instances of e-to were 24 where e was used with a goal-oriented verb, including 4 with hairu ‘enter’ and all the 5 instances with tsunagaru ‘connect.’ It was also found that 29 out of the 37 instances (78%) of nonspatial verbs, and 38 of 48 instances (78%) of abstract nouns appeared with e-to. Table 6 illustrates the distribution of VPs and NPs with e after all the instances with e-no, e-to and verbless sentences are excluded. While the lists of the frequently co-occurring verbs or nouns are not very different from those in Tables 4 and 5, the number of non-spatial VPs and NPs are radically lower, indicating that the function of e, unless used with to, is mostly limited to the spatial expressions. Only Table 6. Distribution of VPs and NPs used with e with e-no, and e-to excluded VPs
NPs
Token frequency Type frequency iku ‘go’ susumu ‘proceed’ mukau ‘head (for) modoru ‘return’ yattekuru ‘come’ hairu ‘enter’ utsusu ‘move’ okurikomu ‘send in’ deru ‘exit’
201 106 33 8 8 7 7 5 4 4 4
90 other spatial VPs 7 non-spatial VPs
118 7
Token frequency Type frequency ..(no) hoo ‘the direction [of]’ gakkoo ‘school’ ..(no) naka ‘inside [of]’ sentoo ‘public bath’ dokoka ‘somewhere’ doko ‘where’ soko ‘there’ ...(no) tokoro ‘place’
202 127 14 12 9 6 4 4 4 4
114 other concrete NPs 5 abstract NPs
150 5
Kaori Kabata
11 instances were found where e was used to describe an abstract event, as in mokuhyoo e mukau ‘head for an aim’ and kyokugenjookyoo e oikomu ‘chase (somebody) into an extreme situation.’ Only 1 instance was found where e was used with a non-spatial VP and an abstract NP. Compared to e, ni was used more frequently and in a wider range of collocation environment in the written texts as well, exhibiting again its dominant status in the semantic expression of spatial goals. While ni also appeared with other particles following it, such as mo ‘also’ (as in ni-mo), wa (a topic maker), made ‘until’, and demo ‘as well’, in 103 instances, unlike the occurrences of e, none of the uses of ni in these cases were seen to deviate from the function of marking a spatial goal: the other particle was used to add some discourse-level information and the sentences would be just as acceptable if ni were used by itself. The allative use of ni was found in 3013 instances with 710 different verbs, resulting in a token vs. type ratio 4 to 1, as opposed to 2 to 1 for e, indicating that some verbs were closely associated with ni. All types of verbs, including goal-oriented and source- or path- oriented verbs, as well as those that are neutral, were found at high frequency. As shown in Table 4, among the most frequently used verbs in the data were hairu ‘enter’, a goal-oriented verb, mukau ‘head (for)’ and deru ‘exit’, both path/source-oriented verbs. Verbs that are neutral in terms of goal- or sourceorientedness, such as iku ‘go’ and kuru ‘come’, also appeared frequently. Ni was used with 1354 different NPs, with very few of them appearing more frequently than 2% or even 1% of the total instances, as shown in Table 5. Nonspatial nouns were used with spatial verbs to describe abstract events in 391 instances (13%). They included temporal expressions like mirai ‘future’ as in mirai ni (...wo) hakobu ‘carry ... to the future’, and abstract nouns like kokoro ‘mind’, ketsuron ‘conclusion’ used in expressions like kokoro ni todoku ‘reach one’s mind’ or ketsuron ni tassuru ‘reach a conclusion’.7 6. Discussion Findings from the present study indicate discrepancy among what is generally believed as their semantic property based on semantic analyses, native speakers’ perception of semantic distribution of e and ni, and their actual use. Compared to the native speaker judgment data, ni’s dominance over e as a spatial goal marker was much more robust in the spoken and written corpus data, with the frequency 7. Of course, there are many instances where ni is used to describe non-spatial, abstract events. In this particular study, however, only the instances were ni appears with spatial verbs were subject to the analysis.
Interchangeability of so-called interchangeable particles
ratio of instances of e to ni at roughly 1 to 8 and 1 to 5 respectively. While native speaker ratings in the judgment test indicated a higher acceptability of e in sentences with a focus on the source, e was not found any more frequently than ni with source-/path-focused verbs in neither the spoken nor the written corpus. On the other hand, there was a strong preference for ni in contexts with a focus on the endpoint of a movement, both in the experimental data and in the corpus data. Although noun type exhibited interaction with particle choice with verbs with a meaning related to ‘enter’ in the native speaker judgment test, the corpus data indicated that ni is preferred in such contexts regardless the size of the location. The data also indicated that there are certain characteristics in the patterns of particle uses in spoken and written texts. The spoken text included certain groups of expressions that appeared repeatedly. For example, e appeared with locative demonstrative or interrogative nouns soko ‘there’ and doko ‘where’ nearly half the time in the spoken corpus, but only 5% in the written corpus. Similarly, iku ‘go’ was the most frequently used verb both with e and ni in the spoken texts, at 46% and 33% of all the instances respectively, but it appeared only about 10% with e and less than 5% with ni in the written texts. The use of e in the written texts exhibited different usage patterns from those in the spoken texts. The prenominal expression e-no was particularly frequent in the written text, appearing 34% of all the instances of e, as opposed to one single occurrence in the spoken text. The uses of e in describing abstract events were found only with the written text, and most of them appeared with e-to, which appeared 142 times in total. It seems that e-to functions as a complex particle, like e-no, and marks both spatial and non-spatial goal, as does e-no. E-to was used to describe both spatial events, shown in (6a) and abstract events with which neither e nor to can be acceptable without each other, as shown in (6b). The examples are taken from the written corpus: (6) a.
boku wa shibafu o tsukkit-te, jibun no uchi {e-to/e} I top lawn acc go through-conn self gen home isoi-da. hurry-past ‘I hurried home across the lawn.’
b.
hito wa kooyuu no naka de fukusuu no kachikan person top friendship gen inside loc multiple gen value acc o motta ningen {e-to / *e / *to} seechooshi-te-iku no da. have human grow-conn-go nom cop ‘People grow up to become a human being with multiple value systems being (involved) in friendship.’
Kaori Kabata
The function of e-to seems to have largely been neglected in Japanese linguistics. Martin (2004: 1007–8) is among the few who discuss the e-to expression, but limits himself to saying that e-to emphasizes the goal sense. Martin fails to recognize the fact that in today’s Japanese e-to can mark non-spatial relations that e alone would not: While e was used to describe non-spatial goals in some instances, the extent of such uses was still fairly limited to fixed expressions. How e-to has developed and is going to develop is an interesting issue, which I hope will be further investigated in a future study. Further examination of usage patterns of e and ni revealed some semantic characteristics of certain verbs. In the spoken texts, iku ‘go’ was the most frequently used verb both with e and ni (46% and 33% of the total instances of each particle), whereas kuru ‘come’ was used almost exclusively with ni. Kuru appeared with ni in 69 instances (11% of all the instances of ni), but with e in only 4 instances (6% of all the instances of e). Moreover, 22 of the instances where kuru occurred with ni made use of the expression koko/kotchi ni kuru ‘come here/this’, while iku took a wider range of locative nouns as its landmark. Although the two deictic verbs are commonly treated as if they differ only in terms of the direction of movement, they seem to also differ in the degree of specificity of the landmark, likely with kuru indicating more focus on the goal. A similar usage pattern was observed for the verbs ireru ‘put in’ vs. dasu ‘take out’ and hairu ‘enter’ vs. deru ‘exit’. While the source-focused verbs dasu and deru were used both with e and ni, the goal-focused verbs ireru and hairu were used almost exclusively with ni. Despite the common assumption of the interchangeability of e and ni, the results suggested that native speakers are not using these two particles randomly, as commonly suggested in dictionaries and grammar books. Rather, ni has become dominant in the semantic field of spatial goal marking to the point where e is used in a very limited contexts, both in terms of the overall frequency of occurrence and in terms of the types of nouns and verbs it takes. This is rather contradictory to the previous literature, which implies that e has extended its function to take on the goalmarking function previously associated with ni. At the same time, e seems to have acquired a grammatical function to mark abstract relations when used with no in the adnominal marker e-no or with to in e-to. The high frequency of e-no and e-to in written texts further suggests the discrepancy between spoken and written speech. These findings are, however, only tentative due to limitations of the study. A future experimental study will have to be based upon a larger corpus and be tested with more participants. Individual differences should also be taken into consideration. The findings also have to be tested against different types of corpus data. While representing spontaneous unprepared speech, telephone conversations may deal with different topics from face-to-face conversations, which might have influenced the data. More different types of written texts also need to be analyzed.
Interchangeability of so-called interchangeable particles
7. Implication to grammar Over the last two decades, cognitive and functional linguists have devoted their research to demonstrate that language “must be described as an integral facet of overall psychological organization”, and that “well-formedness judgments are often matters of degree, and reflect the subtle interplay of semantic and contextual factors” (Langacker 1988: 4–5). Moreover, such a view of grammar diverges from that generally held by formal linguists who maintain that language is a self- contained module and that grammar is an algorithmic system that generates only grammatical sentences. Based on three different types of empirical data, the results of this study have provided a supporting piece of evidence for the cognitive linguistics view of the grammar, by showing that the speakers’ use of the two goalmarking particles, which have been traditionally treated as if they were semantically poor and were mainly associated with grammatical functions, indicate the interaction between the their conceptualization of the situation and the semantics of the particles. The two particles are not used randomly, nor are as interchangeable as suggested by some of the previous studies. The results of the present study have, however, suggested we need to further establish the empirical methods to approach the issues dealing with grammar. The discrepancy between the speakers’ judgment data and the frequency data from the spoken and written corpora leaves us wonder what is indicated by ‘acceptability’ or ‘interchangeability’. Assuming that grammar is at least in part acquired through entrenchment and conventionalization (e.g., Tomasello 2000), frequency data from the actual usage should be expected to play some role in determining the nature of lexical items or constructions, as argued by Newman and Rice (2004). But what role they should play is yet to be established. The results of this study also provided yet another piece of evidence for the discrepancies between the language uses between spoken and written texts. While the cognitive linguistics view to grammar assumes the variation among speakers, how can it accommodate the variation within speakers? Previous studies have already demonstrated the two texts differ in a variety of grammatical features, including word order, choice of lexical and grammatical morphemes, and syntactic complexity (Quaglio & Biber 2006). The present study has shown that the usage patterns of grammatical morphemes, namely particles, may also differ. Does such difference, then, indicate that there are two sets of grammar, one for spoken and the other for written? There is also the issue of individual differences: What is acceptable to one speaker may not be as acceptable to another. It may be due to dialectal variation, or to generational variation. On the other hand, it may simply reflect differences in the conceptualization of events. Language is dynamic, reflecting the flexibility and elasticity of speaker’s minds. Further development of corpus
Kaori Kabata
linguistics will help us better understand the nature of grammar and its role in capturing the dynamicity of language. References Barlow, Michael & Kemmer, Suzanne. 2000. Usage Based Models of Language. Stanford CA: CSLI. Bauer, Winifred. 1993. Maori. London: Routledge. DOI: 10.4324/9780203403723 Hallan, Naomi. 2001. Path to prepositions?: A corpus-based study of the acquisition of a lexicogrammatical category. In Frequency and the Emergence of Linguistic Structure [Typological Studies in Language 45], Joan L. Bybee & Paul Hopper (eds), 91–120. Amsterdam: John Benjamins. Ikegami, Yoshihiko. 1986: The goal as agent. Sophia Linguistica 20–21:7–14. Ikegami, Yoshihiko. 1987. ‘Source’ vs. ‘goal’: A case of linguistic dissymmetry. In Concepts of Case, René Dirven & Günter Radden (eds), 122–146. Tübingen: Gunter Narr. Kabata, Kaori. 1999. On the agent/experiencer continuum: A cognitive approach to the ga/ni alternation in Japanese. In The Twenty-Fifth LACUS Forum, Shin-Ja Hwan & Arle R. Lommel (eds), 375–386. Chapel Hill NC: LACUS. Kabata, Kaori. 2000. Japanese Ni: A Cognitive Analysis of a Lexically Complex Particle. PhD dissertation, University of Alberta. Kabata, Kaori. In prep. A corpus-based study of child acquisition of Japanese particles. Kaiser, Stefan, Ichikawa, Yasuko, Kobayashi, Noriko & Yamamoto, Hilofumi. 2001. Japanese: A comprehensive Grammar. London: Routledge. Kumashiro, Toshiyuki. 1994. On the conceptual definitions of adpositions and case markers: A case for the conceptual basis of syntax. In Papers from the Thirtieth Regional Meeting of the Chicago Linguistic Society, Katharine Beals et al. (eds), 236–250. Chicago IL: CLS. Langacker, Ronald W. 1987. Foundations of Cognitive Grammar, Vol. 1: Theoretical Prerequisites. Stanford: Stanford University Press. Langacker, Ronald W. 1988. An overview of cognitive grammar. In Topics in Cognitive Linguistics [Current Issues in Linguistic Theory 50], Brygida Rudzka-Ostyn (ed.), 3–48. Amsterdam: John Benjamins. Langacker, Ronald W. 2000. A dynamic usage-based model. In Barlow & Kemmer (eds), 1–64. Martin, Samuel Elmo. 2004. A Reference Grammar of Japanese. Honolulu HI: University of Hawaii Press. Masuda, Kyoko. 2002. A Cognitive Approach to Japanese Locative Postpositions Ni and De: A Case Study of Spoken and Written Discourse. PhD dissertation, University of Arizona. Newman, John & Rice, Sally. 2004. Patterns of usage for English sit, stand, and lie: A cognitivelyinspired exploration in corpus linguistics. Cognitive Linguistics 15(3): 352–396. DOI: 10.1515/cogl.2004.013 Ono, Tsuyoshi & Suzuki, Ryoko. 1992. Word order variability in Japanese conversation: motivations and grammaticization. Text 12(3): 429–445. DOI: 10.1515/text.1.1992.12.3.429 Ono, Tsuyoshi, Thompson, Sandra A. & Suzuki, Ryoko. 2000. The pragmatic nature of the socalled subject marker ga in Japanese: Evidence from conversation. Discourse Studies 2(1): 55–84. DOI: 10.1177/1461445600002001003
Interchangeability of so-called interchangeable particles Quaglio, Paulo & Biber, Douglas. 2006. The grammar of conversation. In The Handbook of English Linguistics, Bas Arts & April McMahon (eds). Oxford: Blackwell. DOI: 10.1002/9780470753002.ch29 Sandra, Dominiek & Rice, Sally. 1995. Network analyses of prepositional meaning: Mirroring whose mind-the linguist’s or the language user’s? Cognitive Linguistics 6: 89–130. DOI: 10.1515/cogl.1995.6.1.89 Shibatani, Masayoshi. 1978. Nihongo no bunseki (An analysis of Japanese language). Tokyo: Taishukan. Svorou, Soteria. 1994. The Grammar of Space [Typological Studies in Language 25]. Amsterdam: John Benjamins. DOI: 10.1075/tsl.25 Tokugawa, Munemasa. 1989. Nihon hogen daijiten (A comprehensive dictionary of Japanese dialects). Tokyo: Shogakukan. Tomasello, Michael. 2000. The item-based nature of children’s early syntactic development. Trends in Cognitive Sciences 4(4): 156–163. DOI: 10.1016/S1364-6613(00)01462-5 Yamaguchi, Akido & Akimoto, Morihide. 2001. Nihongo bunpoo daijiten (A comprehensive dictionary of Japanese grammar]. Tokyo: Meiji Shoin. Yamazaki, M. 2002. From PERSON marker to AFFECT marker: The grammaticization of the Japanese 2nd person singular pronoun omae in conflict talk. Paper presented at the 6th Conference on Conceptual Structure, Discourse, and Language, October 11–14, 2002, Rice University, Houston TX.
Appendix List of stimulus sentences in the native speaker judgment test. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
太郎は先週アメリカ__行った 太郎はもうアメリカ__行ってしまった 太郎は明日アメリカ__行く 太郎は先週アメリカ__出発した 太郎はもうアメリカ__出発してしまっ た 太郎は明日アメリカ__出発する 太郎は先週アメリカ__着いた 太郎はもうアメリカ__着いた 太郎は明日アメリカ__着く 太郎は先週空港__行った 太郎はもう空港__行ってしまった 太郎は明日空港__行く 太郎は先週空港__出発した 太郎はもう空港__出発してしまった 太郎は明日空港__出発する
Taro went to America last week. Taro has gone to America. Taro will go to America. Taro left for America last week. Taro has left for America. Taro is going to leave for America tomorrow. Taro arrived in America last week. Taro has arrived in America. Taro will arrive in America tomorrow. Taro went to the airport last week. Taro has gone to the airport. Taro will go to the airport. Taro left for the airport last week. Taro has left for the airport. Taro is going to leave for the airport tomorrow.
Kaori Kabata 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50.
Taro arrived in the airport last week. Taro has arrived in the airport. Taro will arrive in the airport tomorrow. Taro went to school last week. Taro left for school tomorrow last week. Taro arrived in school last week. Taro went to work last week. Taro left for work tomorrow last week. Taro arrived in work last week. Taro put the car inside the warehouse. Taro is going to the car inside the warehouse. The car moved inside warehouse. The car is going to move inside warehouse. Taro took the car outside the warehouse. Taro is going to take the car outside the warehouse. 車は倉庫の外__出た The car moved outside warehouse. 車は倉庫の外__出るらしい The car is going to move outside warehouse. 太郎はカップ__紅茶を注いだ Taro poured tea inside/into the cup. 太郎はカップ__紅茶を注ぐだろう Taro is going to pour tea inside/into the cup. コインがカップの中__落ちた The coin fell inside/into cup. コインがカップの中__落ちそうだ The coin is going to fall inside/into cup. 太郎はテーブル__紅茶をこぼしてしま Taro spill tea [out of the cup and] onto the った table. 太郎はテーブル__紅茶をこぼしそうだ Taro is going to spill tea onto the table. 紅茶がテーブル__こぼれた Tea spilled onto the table.. 紅茶がテーブル__こぼれそうだ Tea will spill onto the table.. ナイル川は地中海__注いでいる Nile River pours into the Mediterranean Sea. 太郎はあごを上__向けた Taro turned his chin upward. 太郎はあごを下__向けた Taro turned his chin downward. 太郎は頭を左__向けた Taro turned his head leftward. ナイル川は地中海__注いでいる California is in contact with Pacific Ocean 私の頬が床__触れた My cheek came in contact with the floor. 太郎の手が私の顔__触れた Taro’s hand came in contact with my face. 太郎の手が私の鼻__触れた Taro’s palm came in contact with my nose. ネバダ州は大平洋__近接している Nevada is contiguous to Pacific Ocean. 太郎の手が私の顔__触れそうだ Taro’s hand almost touched my face. 太郎は先週空港__着いた 太郎はもう空港__着いた 太郎は明日空港__着く 太郎は先週学校__行った 太郎は今朝学校__出かけた 太郎は今朝学校__着いた 太郎は先週仕事場__行った 太郎は今朝仕事場__出かけた 太郎は今朝仕事場__着いた 太郎は車を倉庫の中__入れた 太郎は車を倉庫の中__入れるつもりだ 車は倉庫の中__入った 車は倉庫の中__入るらしい 太郎は車を倉庫の外__出した 太郎は車を倉庫の外__出すつもりだ
The re-examination of so-called ‘clefts’ A study of multiunit turns in Japanese talk-in-interaction Junko Mori Using conversation analysis as a central tool for analysis, this chapter examines how conversational participants utilize a cleft construction, which begins with -no wa topic clause, as a resource to construct and process their turns at talk. Unlike the well-formed structure of this construction described in reference grammar, it is rare for the speakers in talk-in-interaction to end the unit initiated with the -no wa topic clause with a clause marked by a nominalizer. The current analysis demonstrates how the the -no wa topic clause serves as a prospective indexical that pre-announces the nature of the forthcoming extended talk and provides the recipients an interpretive framework, and thereby contributes to the investigation of the mutual relationship between interaction and grammar.
1. Introduction The aim of this chapter is to demonstrate how the methodological framework of conversation analysis (CA) (cf. Sacks, Schegloff & Jefferson 1974) conceptualizes ‘grammar’ in its studies of talk-in-interaction. More precisely, the chapter explores how grammar serves as a resource for conversational participants to construct and process their turns at talk and thereby organize their participation in interaction. To exemplify this perspective, this study examines one type of cleft constructions in Japanese as it is used in talk-in-interaction. The following sentence introduced by Iwasaki (2002: 207) represents the target construction: (1)
Iwasaki (2002: 207) [daredemo shitte-iru] no wa anyone know nml top [ningen wa itsuka wa shinu] to yuu koto da. man top someday top die qt say nml cop ‘What everyone knows is the fact that people eventually die.’
Junko Mori
As shown in (1), this construction, which can be considered comparable to the English WH-cleft construction, essentially takes the basic topic-comment structure, X wa Y da: a complement marked by the nominalizer no and the topic marker wa is followed by a complement marked by a nominalizer such as koto and the copula da, which constitutes the comment component. Indeed, Harada (2000: 439 [originally published in 1972]) asserts that the predicate component in this construction must consist of a noun phrase and a copula. As we examine spontaneous talk-in-interaction, however, we find many cases in which the -no wa topic clause is not followed by a corresponding comment component marked by a nominalizer, even though what follows the topic clause develops into a full clause or longer. To be more exact, 27 cases of such utterances are found in the current database of approximately five hours of conversations and interviews. In fact, the ‘well-formed’ structure such as shown in (1) never appears in this database.1 The following excerpts exemplify the kinds of cases found in the data (the transcripts consist of original Japanese utterances, interlinear glosses, and approximate English translations, which do not necessarily reflect the structural features of original Japanese utterances. The transcription conventions and abbreviations are provided in the appendix): (2) [Homestay 30105] ((A graduate research assistant is interviewing Aoyama, a middle-aged woman, who has hosted a number of international students.)) Aoyama:
... = tada watashi ga ichiban yana no wa, .hh but I S most unpleasant nml top ukeireru toki wa kitai to:: .hh °yaa:° accept time top expectation and ano atarashii kata to no deai de ne? uhm new person with lk encounter cop fp ‘... but what I don’t like most is, when I take them in, because of expectations and excitement for meeting a new person,’
1. The database for this study consists of six face-to-face conversations and four interviews that were videotaped for the project titled ‘Learning Through Listening Towards Advanced Japanese Proficiency’ conducted under the Center for Advanced Language Proficiency Education and Research (CALPER) . The participants in these interactions were provided with topics such as cultural differences, intercultural communications, gender roles, educational systems, and so on, which are frequently covered in intermediate and advanced Japanese language textbooks. The participants include both males and females, whose ages range from their early twenties through their sixties at the time of recording.
The re-examination of so-called ‘clefts’
IR:
[ha:i yes ‘yes.’
Aoyama: [sugoku tanoshii desu kedo, .hhh really fun cop but kaeru toki ga iya desu ne yappari. = return time S unplesant cop fp expectedly ‘It is really fun, but .hhh I don’t like the time when they go home, as you might expect.’ IR: = aa:::: oh ‘oh:::’ (3) [Kokusaika 3645] ((Three graduate students in the twenties, Aya, Nori, and Shin, are talking about their international experiences.)) Aya:
watashi ga omotta no WA, nihon GA, I S thought nml top Japan S
amerikakabureshiteru no to issho DE::, Americanized nml with same cop
kankoku mo::, (0.5) >nihonkabureshiteru mitaina< Korea also Japanized like
tokoro [o kanjita [tokoro ga aru no ne, aspect O felt aspect S exist nml fp
‘What I thought was, like in the same way in which Japan has been Americanized, Korea, too (0.5) has been Japanized, there are things that I felt that way.’ Shin: Nori:
[aa::::: oh ‘Oh::’
[
[aa:: soo na n da. oh so cop nml cop ‘Oh::, is that so.’
(4) [Kyoiku 3510] ((A graduate research assistant is interviewing a male middle school principal on various issues concerning the on-going educational reform.)) Principa1: ... tada watashi jishin ga omotteiru no wa ne, but I self S think nml top fp
Junko Mori
senseitachi ni mo iu n desu ga, .hhh teachers dp also say nml cop but
ichijikanjuu o hyooka no tame ni one-hour-entire O assessment lk purpose for
jugyoo o suru n ja nai yo °to.° class O do nmk cop:top neg fp qt
‘... but what I myself think is, I tell other teachers this, too, but, .hhh it is not that we design the entire class hour just for the purpose of assessment.’ IR:
°un° uh-huh ‘uh huh.’
(0.9)
Principal: kodomotachi ni rikaisaseru tame ni ri- (0.5) children dp understand:cau purpose for
jugyoo suru n de. class do nml cop
‘We do it for the purpose of having kids understand the materials.’ IR:
°ha:::° yeah ‘Yeah.’
As shown above, in spontaneous talk-in-interaction, we can observe units initiated with the -no wa clause ending not with a nominalizer, but with the reappearance of the verb or nominal adjective used in the topic clause (as shown in (2)), or with the occurrence of a verb or nominal adjective different from, but similar to, the one used in the topic clause (as shown in (3)). Other cases include those that are marked by the quotative particle to (as shown in (4)), or those that are not accompanied by any clear grammatical marking that could indicate the end of the corresponding comment component (as examined later in this chapter). From the perspective of prescriptive grammar, these utterances may be considered illformed performance errors, but the fact that such a practice often occurs in interactions involving different speakers tells us something about the nature of grammar-in-interaction. Further, in some of these cases, several other units of talk are produced between the -no wa topic clause and the clause ending with a verb or nominal adjective that corresponds to the topic clause, seemingly marking a possible completion of the extended talk, as we will see in the analysis section. Namely, the -no wa topic clause appears to initiate or ‘project’ what is called a ‘multiunit
The re-examination of so-called ‘clefts’
turn’ (e.g., Schegloff 1982, 1996), serving as a ‘prospective indexical’ (Goodwin 1996) of the subsequent talk of the current speaker and establishing an interpretative framework for the recipients of the talk.2 This study investigates the interrelationship between this ‘grammatical practice’ (cf. Lerner & Takagi 1999; Hayashi 2001, 2004a) and the sequential organization of the talk and actions accomplished. In particular, it investigates how the target construction affects the initiation and realization of a multiunit turn, paying close attention to both the current speaker’s and the recipients’ treatments of the ongoing development of talk. The following Section 2 briefly reviews previous studies on cleft constructions and clarifies the analytical approach taken by the current study. Section 3 summarizes the findings of conversation analytic inquiries concerning how the turn-taking operation may shift from the participants’ exchange of single-unit turns to one party’s initiation of a multiunit turn. Subsequently, Section 4 provides close analyses of two excerpts of multiparty interaction that involve the target construction in order to demonstrate how the trajectory presented by the -no wa topic clause, along with co- occurring non-verbal behaviors, influence the ongoing development of talk-in-interaction and the shift in the participation framework. Finally, the concluding section discusses the implication of this observation to the conceptualization of grammar. 2. Discourse functional studies of cleft constructions Over the last thirty years, a number of studies have examined the use of cleft constructions in discourse (e.g., Gundel 1977; Prince 1978; Kamio 1990; Collins 1991; Geluykens 1991; Kim 1995; Weinert & Miller 1996; Jucker 1997; Davidse 2000; Lambrecht 2001; Herriman 2003, 2008; Hopper 2004, 2007 among others).3 Concepts such as given/known versus new (e.g., Prince 1978; Kamio 1990), foreground versus background (Jucker 1997) or focus /emphasis (e.g. Gundel 1977; Geluykens 1991) have been introduced to discuss how such constructions may contribute to the management of information status and the construction of discourse. Other studies such as Weinert and Miller’s (1996) suggest that WH-clefts often introduce topics or mark an important starting point for the following discourse. More recently, Herriman (2003) discusses how WH-clefts can be used as a rhetorical device to gain more argumentative force and argues that basic WH-clefts often involve a contrastive development of what has previously been expressed in the text. What is 2. See Hayashi (2004b) for the analysis of another example of a prospective indexical in Japanese, i.e., the distal demonstrative pronoun are. 3. The fact that the majority of studies reviewed in this section are based on English data is simply because, to my knowledge, there are few studies of this Japanese construction that have taken discourse-functional approaches.
Junko Mori
common among these studies is their concentration on written corpus (some of them rely on invented examples) or spoken narratives. Even when the databases that are used include conversational data, little attention has been paid to the interactional dimension of conversation or the recipients’ reaction to the unfolding talk. Indeed, Geluykens (1991: 354) explicitly acknowledges this shortcoming. Among previous studies on cleft constructions, the approaches taken by Kim (1995) and Hopper (2004) are probably the closest to that taken in the current study because they consider the kinds of service WH-clefts can provide for interactional enterprises in spontaneous English conversations. The findings reported by them parallel what was observed in the Japanese data examined in this study. For instance, Kim (1995: 251) reports that verbs marking a speaker-internal state such as realize, want, feel, think, know, hesitate, enjoy, see, object to, bother, etc., and verbs marking a metalinguistic dimension such as mean, say, etc. constitute the majority of WH-clefts found in his English conversational data and that many of them are used with the first person subject. This tendency is also observed in the data analyzed in this study as exemplified in Excerpts (2), (3), and (4). Kim further notes that his data include many cases of WH-clefts containing a full clause in the focus position that have no THAT (and sometime not even a BE copula). According to Kim, in such cases, the initial WH clause is often followed by background information rather than the focused utterance. The following excerpt introduced by Kim exemplifies such a case: (5)
Kim (1995: 273) 143 D: Y- yeah well it’s in these big 144 (1.0) 145 D: [va:ses, [you know. [Huge. 146 W: [Vats 147 H: [Yeah. [Yeah. 148 D: But what I didn’t realize at the time was I had always 149 been thinking well a:ll anything alcoholic has been (.) 150 distilled and is oka:y. 151 (0.2) 152 D: But this isn’t. 153 (0.3) 154 D: It’s just made from (.) [I mean t’ey jus155 J: [Oh yea:h, 156 D: It’s just ferme:nted. It’s not disti::lled.
In this case, the focused utterance of the WH-cleft initiated in line 148 is found in line 152 (‘this isn’t’). What follows right after the WH clause (‘I had always been... and is oka:y’) provides ‘background information that serves as a preliminary to the
The re-examination of so-called ‘clefts’
main point that is projected to be forthcoming in the subsequent talk’ (p. 273). Hopper (2004) also demonstrates that WH-clefts, or the bi-clausal construction with a wh-part and a focus part consisting of a noun phrase, are seldom fully exemplified in spontaneous talk and that what follows the WH- part is a stretch of discourse of indeterminate length, including a noun phrase as one of many possibilities. These observations appear, to a certain extent, to be comparable to what is found in the Japanese data examined here. That is, the majority of units initiated with the -no wa clause contain no nominalizer or copula marking the end of the comment component, and in some cases, multiple units of talk are produced between the -no wa clause and what appears to be its corresponding comment component. The close observation of conversational data has led Kim as well as Hopper to propose an alternative view of the so-called WH-cleft to the conventional view of grammar, which was primarily based on written discourse and invented examples. Kim (1995) states that ‘various functions of WH-clefts may be better understood in terms of the speaker’s adding the initial WH-clause before the presentation of a major point in the focused utterance rather than in terms of a syntactic operation in which an item is picked up and extraposed to be given focus’ (p. 252). Similarly, Hopper (2004) suggests that the WH-clause in these cases should be regarded as an ‘emergent discourse particle.’ He further states that the primary roles of the WH-cleft are ‘to defer an upcoming segment of discourse’ and ‘to provide a cognitive space for the speaker to plan the next part of the utterance, or a social space to stave off interruption.’ The current study pursues this alternative understanding of the workings of so-called clefts proposed by Kim and Hopper, by examining the seemingly comparable construction in Japanese initiated by the -no wa clause. The description of the WH-cleft’s functions offered by Hopper, especially its social dimension, I believe, can be further enriched by incorporating CA perspectives and the findings regarding multiunit turns. This study thus aims to illustrate the interactive processes that the speakers and the recipients engage in for the achievement of extended talk, shedding light on how the target construction affects the recipients’ conduct as well. More specifically, the analysis uncovers how the recipients anticipate a forthcoming completion of the turn, or in other words, an occasion when their full-fledged response to the current speaker’s talk becomes relevant. 3. Single unit turns versus multiunit turns This section offers a brief overview of CA’s account for the organization for turntaking, which provides a critical analytical framework for the current study. Of particular importance to the current study is the concept of a multiunit turn, which
Junko Mori
is considered something to be achieved in ordinary conversation, where the completion of every turn-constructional unit (TCU) can be susceptible to a transfer of speakership.4 As described by Sacks, et. al. (1974) and further discussed in subsequent studies including Schegloff (1996), Ford and Thompson (1996), Ford, Fox, and Thompson (1996), Selting (2000), and Ford (2004), among others, the TCU is not a kind of unit that is defined a priori by researchers, but something that is realized by the conversational participants in talk-in-interaction with reference to multiple resources, including syntax, prosody, sequential organization, and bodily conduct. Recent studies of Japanese conversations by Mori (1999), Tanaka (1999, 2000), Hayashi (2003a), Iwasaki (2008, 2009), and Nakamura (2009), among others, also attest that the basic rules described by Sacks, et al. (1974) apply to Japanese, while its word order and other grammatical features may change the temporal development of a TCU and the recipients’ ability to project its possible completion. To crudely summarize, the participants in Japanese conversations are found to demonstrate their orientation towards the end of a predicate, following intonation, and some type of action completion, and most notably the combination of all of the above, in recognizing a possible completion of a TCU. While Sacks, et al. (1974) describe such fundamental mechanisms for turntaking in which each party is essentially allotted one TCU at a time, they also acknowledge that there are occasions when one party accomplishes talk longer than 4. Sacks, Schegloff, and Jefferson (1974) describe a basic set of rules for turn-taking operation as follows: (i) For any turn, at the initial transition-relevance place of an initial turn-constructional unit: (a) If the turn-so-far is so constructed as to involve the use of a ‘current speaker selects next’ technique, then the party so selected has the right and is obliged to take next turn to speak; no others have such rights or obligations, and transfer occurs at that place. (b) If the turn-so-far is so constructed as not to involve the use of a ‘current speaker selects next’ technique, then self-selection for next speakership may, but need not, be instituted; first starter acquires rights to a turn, and transfer occurs at that place. (c) If the turn-so-far is so constructed as not to involve the use of a ‘current speaker selects next’ technique, then current speaker may, but need not continue, unless another selfselects. (ii) If, at the initial transition-relevance place of an initial turn-constructional unit, neither (i)a nor (i)b has operated, and, following the provision of (i)c, current speaker has continued, then the rule-set a-c is re-applied at the next transition-relevance place and recursively at each next transition-relevance place, until transfer is effected. (p. 704) These rules are not prescriptive rules that constrain conversational participants’ behaviors, but rather intersubjective norms towards which the conversational participants are found to demonstrate their orientation. Thus, they do not, for instance, exclude the possible occurrences of overlapping talk or silence between turns, but rather they offer explanations for how such overlaps or silence may occur and how they can be interpreted or treated by the participants themselves.
The re-examination of so-called ‘clefts’
one TCU, or what is called a multiunit turn. Multiunit turns are different from cases in which no one self-selects at the end of a TCU and thus the current speaker ends up continuing his or her talk. Rather, the establishment of a multiunit turn involves the speaker’s pre-announcing, in some fashion, that his or her upcoming talk will require more than one TCU, and securing co-participants’ continued recipiency of the projected talk. One of the typical devices for accomplishing multiunit turns described in CA studies is what is called ‘story preface’ (e.g., Sacks 1974, 1992; Jefferson 1978). Namely, when one is about to tell a story in conversational interaction, the person often starts with a story preface such as ‘I heard the most wonderful thing yesterday,’ or ‘Something awful happened to me today,’ which serves to preannounce the fact that he or she is going to tell the co-participants what ‘the wonderful thing’ or ‘something awful’ was and that this telling can take up more than a single TCU. The story preface gives the co-participants a chance to align themselves as the story recipients and helps them understand the nature of the telling (i.e., a ‘wonderful’ or ‘awful’ story) and process the subsequent talk accordingly; they wait for the moment when the story reaches the ending that describes why it was wonderful or awful and provide an appropriate response. Another device that enables a participant to deliver a multiunit turn is what Schegloff (1980) calls a ‘preliminary to preliminaries,’ or ‘pre-pre’ in short. A typical pre-pre, for instance, is a question such as ‘Can I ask you a question?’ What happens following this question and the recipient’s compliance, however, is not always the intended question that the person has just received permission to ask. Instead, some preliminaries such as an explanation of the relevance of the question or information necessary to establish the content of the question can occur before the question as shown in (6): (6)
[#12 BC, Beige, 18–19] 11 B: I like tuh ask you something 12 A: Shoot. 13 B: Y’know I’ad my license suspended 14 fuh six munts, 15 A: Uh huh 16 B: Y’know for a reason which, 17 I rathuh not, mention tuh you, in 18 othuh words, -- a serious reason, 19 en I like tuh know if I w’d 20 talk tuh my senator, or– 21 somebuddy, could they help me 22 get it back,
(Schegloff 1980: 105) ((pre-pre)) ((preliminaries))
((question))
Junko Mori
In such a case, the co-participants who complied to be a responder to the question process the preliminaries delivered in a multiunit turn, assuming that these are relevant to the projected, forthcoming question. In addition to these characterizations of entire turns that can project forthcoming multiunit turns, features that accompany these turns have been discussed by Schegloff (1980, 1996) who notes ‘dummy terms’ such as that or something used in these turns, and by Goodwin (1996) who names such items as ‘problem’ or ‘terrible’ or ‘wonderful’ that characterize the upcoming telling as prospective indexicals. Goodwin (1996: 384) describes how such a prospective indexical enables the recipients to engage in the analysis of a current speaker’s talk, which is ‘a dynamic process that extends through time as subsequent talk and the interpretative framework provided by the prospective indexical mutually elaborate each other.’ The findings of these CA studies that describe devices and mechanisms observed in the establishment of multiunit turns help us consider one way in which the -no wa clause is used in Japanese talk-in-interaction. Namely, it appears that the -no wa clause can serve as a prospective indexical forecasting an upcoming multiunit turn. As noted in the previous section, there are remarkable similarities between what Kim and Hopper report regarding WH-clefts in English conversations and what the current study has found regarding the Japanese so-called clefts initiated with the -no wa topic clause. However, there are also crucial differences, in particular, concerning to what extent such marked word orders change the temporal realization of talk-in-progress. Namely, while the English WH-cleft does not drastically change the placement of a verb from its non-cleft counterpart as illustrated in (7a) and (7b), the cleft version and the non-cleft version of comparable Japanese utterances differ significantly in their placement of the verb or nominal adjective as shown in (8a) and (8b). (7) a. What I thought is [ ] b. I thought that [ ] (8) a. watashi ga omotta no wa [ ] I S thought nml top (8) b. [ ] to omotta. QT thought This difference brings about significant consequences in the temporal development of talk, which is experienced by the participants of conversation. That is, the canonical word order in Japanese positions verbs or nominal adjectives indicating the speaker’s internal state, which often occur in such a construction used in talkin-interaction, at the very end of the utterance, or after revealing the contents of
The re-examination of so-called ‘clefts’
the speaker’s thoughts or feelings (see Lerner and Takagi (1999), Tanaka (1999, 2000) and Nakamura (2009)) regarding some consequences of this word order upon the turn-taking operations in Japanese). On the other hand, by utilizing the -no wa topic clause, the speaker can preannounce the nature of the following talk, and as a result, the recipients can process the following talk accordingly. And yet, the frequent reproduction of the same, or a very similar, verb or nominal adjective used in the topic complement at the completion of projected talk (as exemplified in (2) and (3)) may also indicate Japanese speakers’ orientation towards the canonical, predicate-final word order in regard to the negotiation of turn-taking and participation. The following analysis section will introduce two excerpts from multi-party interactions where the target grammatical practice occurs. The purpose of this analytical section is not necessarily to provide generalizable claims that apply to all the cases in which the target construction is observed, but to emphasize the importance of documenting how such a construction is used as a grammatical practice in response to a specific interactional need in a particular sequential context, and how it is treated by the participants themselves. While it is not my intention to disregard the necessity of investigating the aggregate functions of this construction, I believe that close analysis of both the speakers’ and the recipients’ conduct in the unfolding talk will illustrate what ‘grammar’ may mean to the language users themselves. 4. Analysis of so-called ‘clefts’ in multiparty talk-in-interaction This section introduces two excerpts from multi-party interactions to demonstrate how the findings of CA studies discussed in the previous section can be applied to the study of the target grammatical practice. The first excerpt involves a case in which a unit of talk initiated with the -no wa clause appears to come to a possible completion when the speaker produces a predicate that essentially reiterates the one included in the topic clause. The second excerpt, on the other hand, involves a case in which a unit of talk initiated with the -no wa clause does not include any explicit linguistic marking of possible completion. The analysis of these cases addresses not only how the target grammatical practice affects the organization of participation, but also how multimodal practices, including gaze shifts, body movements, gestures, etc., may influence the ways in which the participants develop and process the talk-in-interaction and actions-in-progress. These other practices may enhance the projection activated by grammar in some cases, or they may serve to cancel it out in other cases.
Junko Mori
4.1
A case of a multiunit turn that ends with a predicate corresponding to the topic clause
The first case shown in (9) is extracted from a multiparty interaction among Mari, Taka, and Eri, who had just returned to Japan after studying abroad in the U.S. for one year at the time of the data collection. By using the -no wa topic clause, Eri characterizes the upcoming talk as a description of something that surprised her at the beginning of her stay in the US. Following the topic clause, Eri first refers to the fact that libraries in the US are open until late at night, ending this utterance with the tag-like expression jana::i, which, taken out of the context, could be viewed as marking a possible completion of the current turn and yielding the next turn to another speaker. However, the other participants do not start taking a full turn at that moment. Indeed, Eri quickly takes a short in-breath and continues her talk with the connective nanka (‘like’). This next unit initiated with dakara ends with the predicate, sugoina naa::toka omotta: (‘thought impressive’), which reiterates the content of the -no wa topic clause, sugoi bikkurishita (‘really surprised’), or at least expresses the same sort of assessment. Mari’s full-fledged response to Eri’s talk starts only after Eri produces this reiterated assessment and the subsequent token u::n yielding speakership (cf. Iwasaki, 1997). After the presentation of the transcript of this segment, we will consider what kinds of resources contribute to Eri and her co-participants’ accomplishment of this multiunit turn. (9) JA10915 15 Eri:
[.hhhhhhhhh=
16
=soo i- suggoi saisho:: suggoi bikkurishita no wa:::, .hhh so real first real surprised nml top
17
nan daro, yonaka no juuniji toka niji toka what cop midnight lk 12-oclock or 2-oclock like madeuntil
18 toshokan oo- (0.3) oopu[nshiteru jana::i. .hhh= library open tag ‘Right i- what really surprised me first is:::, .hhh what can I say, like until 12 o’clock midnight or 2 o’clock, libraries are o- (0.3) open, right?’ 19 Mari:
[u::::n uh-huh ‘uh huh.’
20 Eri: =nanka sore made minna sa::: ttoka itte isshookenmei like that until all mim like say hard
The re-examination of so-called ‘clefts’
21 benkyooshiteru tte yuu no wa sugoi studying qt say nml top impressive naa::to[ka omotta:.= fp like thought ‘Like everyone is studying really hard until then, and I thought that was really impressive.’ 22 Mari:
[u::n. uh-huh ‘uh huh.’
23 Eri: =u::n. uh-huh ‘uh huh.’ 24 Mari: shisetsu ga totonotteru yo ne[:. jujitsushiteru yo ne[: facility S established fp fp substantial fp fp ‘Their facilities are really nice, aren’t they? Very substantial.’ 25 Eri: 26
[u::n uh-huh ‘uh huh.’
[u::n. uh-huh ‘uh huh.’
u::n. uh-huh ‘uh huh.’
The accomplishment of this multiunit turn involves several factors including 1) grammar as a resource for designing and processing the turn, 2) co-occurring non-verbal behaviors which provide another frame of reference for processing the on-going development of talk, and 3) the sequential context in which this particular multiunit turn occurred. In the following discussion, we will examine the excerpt above, considering these three factors. First, as mentioned earlier, the -no wa topic clause appears to serve as a prospective indexical that provides the recipients with an interpretive framework, with which to process the on-going development of talk. Namely, the recipients in this case process Eri’s talk by considering what could have made her so surprised in the U.S., and the completion of the talk that uncovers the description of something worthy of surprise becomes the moment when their response to this telling becomes relevant. By processing this on-going development of talk in a linear fashion, one may consider it to be possible that Eri was surprised to know how late American libraries are open. In fact, Mari produces a minimal token at the moment when this information becomes available, namely when the talk reaches the halfway point in the vocalization of the verb oopunshiteru (‘open’) and the remaining
Junko Mori
component becomes rather predictable. However, the clause-final expression for this unit does not necessarily correspond to how this talk was initiated. That is, although the tag-like expression jana::i (line 18) might be considered the end of a turn in other contexts, it is not a suitable ending for the unit initiated with the -no wa clause. By this I do not mean that the recipients considered that such a unit should end with a nominalizer as described in prescriptive grammar. Rather, the frequent pattern of a predicate identical or similar to the one in the topic clause, which occurs at a possible completion of the unit (as exemplified in (2) and (3)), seems to establish such an expectation against which the recipients process the on-going talk. The next unit indeed ends with a predicate that resonates with the characterization provided in the -no wa topic clause. It seems that what surprised Eri, or what she found impressive, is the fact that students were studying so hard until late at night. Eri’s addition of the token u::n (line 23) further confirms that she is ready to realign herself as a recipient of the next speaker’s talk (cf. Hayashi & Yoon, 2009). The construction of this multiunit turn is also supported by Eri’s non-verbal behaviors including gaze shifts, body movements, and mimes. First, in overlap with the end of the prior turn by Taka, Eri produces a rather lengthy in-breath in line 15, indicating that she is ready to take the next turn (cf. Jefferson 1984; Linton & Lerner 2004). At the same time Eri looks away from the current speaker, Taka, and looks up in the air and then down at the table. Such a gaze direction that is not directed to either one of the co-participants, along with the self-addressed question nan daro (‘what can I say’) (line 17), indicates that Eri is engaging in some kind of solitary word search (cf. Goodwin & Goodwin 1986; Hayashi 2003b); in this case, she appears to be trying to search for a way to construct her upcoming talk on what really surprised her. Around the time when she produces the nominal linking particle no in line 17, Eri finally starts directing her gaze towards Mari. Then, as Mari produces the reactive token u:::n (line 19) immediately after Eri’s production of the initial component of the predicate oopun (‘open’), Eri shifts her gaze back to the mid air between Mari and Taka while uttering the tag-like expression jana::i (line 18). By directing her gaze at Mari when referring to the library’s closing hours, Eri seems to make sure that the information is heard by Mari. On the other hand, by not gazing at any recipient when producing the tag question, Eri seems to indicate that the tag question should not be considered the end of the turn yielding speakership. During the production of the next unit starting with nanka (‘like’), Eri leans forward and mimes the ways in which these students are studying hard while producing the mimetic expression sa::: ttoka itte. Eri then brings her gaze up and looks towards Mari during the entire talk in line 21, selecting Mari to be the next speaker. As such, Eri’s gaze shifts and body orientations appear to enhance the initiation, continuation, and termination of this multiunit turn.
The re-examination of so-called ‘clefts’
Finally, it is also important to consider in what kind of interactional context Eri launches this multiunit turn and what kinds of actions she achieves by delivering this turn in this particular format. Excerpt (10) shows the segment preceding Eri’s initiation of her multiunit turn. This segment starts after a short silence that emerged at a possible completion of the prior sequence, in which they were talking about the American students’ grade consciousness. As evident from the transcript, it is actually Mari who initiates the topic of the students’ studious attitude by asking the others, who attended different American universities during their study-abroad experience, whether or not they also noticed students studying diligently in the cafeteria. Eri responds to this question by providing a description that directly contrasts with the one offered by Mari, i.e., that there were people chatting in the cafeteria rather than studying (lines 4 and 6), although she utters this partial disagreement with the use of the connective particle -kedo (line 4) (cf. Mori 1999) and laughter to downgrade its potentially negative force. Taka, on the other hand, shifts the topical focus from a cafeteria to a library while maintaining the theme of Mari’s utterance (line 5), but he ceases this turn during the overlapping talk produced by Eri. Thus, we see that it is in fact Taka who brought up the topic of a library in this conversation. Reviewing the prior sequence, it becomes clear that Eri’s multiunit turn incorporated these previous contributions made by the other participants. (10) JA10915 1 Mari: soo:ne, (0.3) kafeteria de s- gakusee- (0.3) minna soo fp cafeteria in students all 2 benkyoo [shitenakatta? studying:neg:pst ‘Right. (0.3) In a cafeteria, students (0.3) were all studying, weren’t they? 3
Taka:
[hh hh hh
4 Eri: ts .hhhh u::n [>nanka< shabetteru hito mo ita uh-huh well chatting people also existed kedo ne,= but fp ‘ts .hhh uh hu:h well there were some chatting, though.’ 5 Taka:
[toshokan library ‘Library’
6 Eri: =suggoi ikioi de(h). [hehehehehehe .hhhhh impressive speed cop ‘at an impressive speed. hehehehehehe .hhhh
Junko Mori
7 Mari:
[iru ne, iru ne.= exist fp exist fp ‘There are, there are.’
8 Mari: =[un u:n uh-huh ‘an uh huh’ 9 Eri: [u::::n uh-huh ‘uh huh’ 10 Mari: [°demo::° but ‘but’ 11 Taka: [toshokan toka mo hon[toni:: library like also really ‘like the library, too, rea::ly’ 12 Eri:
[TOSHOkan wa- (0.3) soo soo library top so so soo.= so ‘As for the library- (0.3) right right right.’
13 Taka: =hito wa benkyooshiteiru, hh. nihon de wa ne- (0.5) people top studying Japan in top 14 toshokan de wa gakusee wa neteiru n [da library in top students top sleeping nml cop kedomo .hhh but ‘People are studying (0.6) whereas in Japan sl- (0.5) in the library, students are sleeping.’ In line 11, Taka re-initiates his prior utterance that was terminated in overlap. This time, Taka continues his talk despite the fact that Eri initiates a turn reacting to the introduction of the term toshokan (‘library’) in overlap with his talk (line 12). Taka’s utterance that refers to his observation of American library scenes in fact supports Mari’s original utterance that describes the studious attitude of American students; while students are often sleeping in the library in Japan, people are studying in the library in the US. It is in this context that Eri deploys the target construction.
The re-examination of so-called ‘clefts’
Triggered by Taka’s reference to American students’ studying in a library, Eri introduces a circumstantial factor that influences the difference in students’ study habits between the US and Japan. But this is done not as a single turn conveying the new piece of information, but as a unit of talk embedded within the multiunit turn, which indexes her contribution as being in agreement with the others. Namely, by carefully crafting this multiunit turn, Eri makes her contribution not as an abrupt topic shift emphasizing the availability of facility, but as a continuous contribution to both of the co-participants’ descriptions of studious behaviors of American students. In this section, we examined a case in which the multiunit turn ends with a predicate that corresponds to the content of a pre-announcement effected by the -no wa topic clause. The examination of the next case, in which the producer of the -no wa clause does not produce any explicit linguistic marking that corresponds to the initial component of multiunit talk demonstrates how such a marking may affect the development a multiunit turn. 4.2
A case of a multiunit turn that does not end with a clear linguistic marking
Excerpt (11) demonstrates another case in which the -no wa topic clause is used to initiate a multiunit turn. The participants in this interaction, three college students, are also talking about their previous year’s study-abroad experiences. Chika studied in China, whereas Emi and Kana studied in the U.S. (11) JA20615 1 Chika: maa kuni no jijoo [mo aru kedo ne::. well country lk circumstance also exist but fp ‘Well it also depends on the situation of each country, though.’ 2
Emi:
[.hhhhhhhhhhhhhhh=
3 =[are da yo ne tabun sa:::, kankyoo da yo ne:::, that cop fp fp maybe fp environment cop fp fp ‘You know, maybe it is the environment.’ 4 Kana: [hee:::: hn ‘hm::::n.’ 5 Emi:
watashi amerika de omotta no wa sa:::, (0.4) I America in thought nml top fp
6 kyanpasu ni sunderu jan min[na ga::, .hh campus in living tag everyone S ‘What I thought in America wa:::s, (0,4) everyone’s living on campus, right? .hh’
Junko Mori
[uu:::[::n uh-huh ‘uh hu::::n’
7 Kana:
[u::n uh-huh ‘uh huh.’
8 Chika: 9
Emi:
de ryoo ni sa::: minna isshoni sundete:::, .hhh and dorm in fp everyone together living
10
koo:: benkyoosu- gakkoo ni sunderu wake da kara sa::, like study school in living nml cop because fp
11
maa gakkoo o shu toshite benkyoo no ba ni well school O main as study lk place in
12 [sunderu wake da kara:,= living nml cop because ‘And in dorms everyone is living together, and .hhh like study- because people are living in school, well in a place of study having a school as the center, and so.’ 13 Chika: [un un un °un° uh-huh uh-huh uh-huh uh-huh ‘uh huh uh huh uh huh uh huh’ 14 Kana: =u::[:n uh-huh ‘uh huh’ 15 Emi:
[koo- .hh (0.2) yappa- (0.3) >nan te no< jibun no like expectedly what qt fp self lk
16 seekatsu no chuushin ga gakkoo chuushin?= life lk center S school center ‘like- .hh (0.2) you know (0.2) what can I say the center of your own life is school?’ 17 Chika: =un un un.= uh-huh uh-huh uh-huh ‘uh huh uh huh uh huh’ 18 Emi: =benkyoo chuushin ni mawatteru:. study center in revolving ‘your life is revolving around study’ 19 Kana: sore wa [aru yo ne[::. that top exist fp fp ‘That’s true, isn’t it.’
The re-examination of so-called ‘clefts’
20 Chika:
[haa::::::: hmm ‘hm::::::m
The segment shown in (11) starts at the time when Chika makes a concluding remark on her prior telling concerning the relationship between China’s one-child policy and the tendency for parents to direct a significant amount of attention to their only child and to urge the child to study diligently. For the transition from Chika’s talk about China to her own about the US, Emi reiterates the basic idea introduced by Chika that it is up to each country’s circumstances whether or not students in that country study diligently. In doing so, she changes the term jijoo (‘circumstance’) used by Chika (line 1) to kankyoo (‘environment’) (line 3), but both utterances basically propose that it is the various surrounding factors that affect students’ motivations and behaviors. Right after this transitional statement, which also seems to forecast that she is going to talk about something concerning the impact of kankyoo (‘environment’), Emi utters the -no wa topic clause to characterize her forthcoming talk as being concerned with her impressions of the U.S. (line 5). First, it is noticeable that Emi in (11) produces a clause marked by the tag-like expression jan right after the -no wa topic clause (line 6), in a way very similar to what we observed in Eri’s multiunit turn in (9), although Emi in this case adds the post-predicate element minna ga:: (‘everyone’) with no break in intonation and no beat of silence (cf. Couper-Kuhlen and Ono, 2007; Ono and Suzuki, 1992). As in the previous case, the recipients only produce minimal tokens at this point and do not take a full-fledged turn, and thereby indicate their understanding that more talk by Emi is on the way. In the meantime, Emi, after taking a short in-breath, continues her talk, elaborating on the typical living arrangements for American university students. This next unit of talk does not reach a possible completion for some time, either in terms of its syntax or content. Namely the -te form (line 9) and the connective particle kara (‘because’) (lines 10 and 12) project more talk to follow. Further, Emi’s talk up to line 12 describes the living arrangements that she observed, which does not correspond to the pre-announced nature of talk, namely what she ‘thought about.’ The recipients indeed indicate their understanding of such a continuous development by just providing the token u:n and its variants rather than full turns. The earliest moment when the projected multiunit turn can be considered to have reached a possible completion is at the end of line 16, when Emi produces what can be considered the main clause following the -kara clause, jibun no seekatsu no chuushin ga gakkoo chuushin? (‘the center of your life is school’). This clause can also be considered as introducing the content of what she thought when she was in the U.S. However, the end of this unit is not marked by a nominalizer or the re-occurrence
Junko Mori
of a predicate identical or similar to the one in the topic clause. As such, the recipients, Chika and Kana, only produce the token un and/or head nods. Further, Emi’s non-verbal behaviors also insinuate that this is not a completion of the on-going action. During the production of her talk in lines 15, 16, and 18, Emi makes notable gestures as illustrated in the following transcripts and in Figures 1 to 3. Starting immediately before the onset of the utterance in line 15, Emi raises her right arm and holds it up, keeping her right hand raised at about the height of her head (Figure 1). This serves as a preparatory stage for the forthcoming circling gestures that appear to correspond to the idea expressed in her talk that students’ lives revolve around school and study (cf. Streeck 1988, 1993; McNeil 1992, Goodwin 2000a, 2000b). By the time Emi’s talk reaches the end of line 16, she has produced the circling gesture (Figure 2) twice, but her hand comes back to the position indicated in Figure 1, which appears to indicate that this series of gestures and accompanying talk, and most importantly, the action accomplished by these multimodal means, is not yet complete.
Figure 1
Figure 2
Figure 3
(12) JA20615 Keeping the right hand Circling the right at the height of her eyes hand and arm (Figure 1) 1st time 15 Emi: [koo- .hh (0.2) yappa- (0.2) >nan te no< jibun no like expectedly what qt fp self lk
The re-examination of so-called ‘clefts’
Circling the right hand and arm 2nd time (Figure 2)
16 seekatsu no chuushin ga gakkoo chuushin?= life lk center S school center ‘like- .hh (0.2) you know (0.2) what can I say the center of your own life is school?’ 17 Chika: =un un un.= uh-huh uh-huh uh-huh ‘uh huh uh huh uh huh’
Circling the right hand and arm 3nd time
Bringing the hand back to the home position (Figure 3)
18 Emi: =benkyoo chuushin ni mawatteru. study center in revolving ‘your life is revolving around study’ In fact, Emi reiterates her point in line 18, this time by completing the unit with the verb mawatteru (‘revolving’) and the falling contour. In addition, her circling gesture, which is produced for the third time during the first half of the utterance in line 18, also comes to a completion when she brings her hand back to the home position shown in Figure 3 (Sacks & Schegloff 2002) in the second half of the utterance in line 18. Kana indeed treats this point as a place where it is relevant to produce more than just a minimal token. Her turn in line 19, sore wa aru yo ne:: (‘That’s true, isn’t it’) treats Emi’s preceding talk to be not merely a description but an expression of her interpretation of the situation. However, this response offered by Kana is still relatively short and is produced while Kana looks down at the table or at her hands on her knees rather than at Emi, who has just expressed her viewpoint and to whom Kana seems to be expressing this agreement. Emi, who shifts her gaze from Chika to Kana at the end of the utterance in line 18, observes this, and continues to extend her multiunit turn by providing more detailed description of what she observed in the U.S., which made her think that American students’ lives revolves around the school (line 21) as shown below. (13) JA20615 21 Emi: [( 22
nanka) sa: .hh kekkyoku kae- heya ni kae like fp after-all room to retrun
ttemo: ruumumeeto hon yondetari sa:, n- nanka even-if roommate book reading:or fp like
Junko Mori
23
benkyooshiteru jan. .hh jugyoo no aida ni:: sa studying tag classes lk between in fp
24
heya ni chotto sa:: (0.7) kaeru toki mo::, maa .hhh room to a-little fp return time also well
25
ruumumeeto ga itari inakattari da kedo:::, roommate S exist:or exist:neg:or cop but
26
de toshokan ni mo::, and library in also
‘like .hh after all re- even when you return to your room, your roommate may be reading, or like studying, you know? .hh when you go back to your room between classes for a short time, too, well you may or may not find your roommate, but, in libraries, too,’ ((18 lines of Emi’s extended talk and Kana’s insertion of short turns that collaboratively tell their observation of campus life in the U.S. and Chika’s reactive tokens and repair initiators are omitted.)) 45 Emi: de::: sooyuu no de:: koo (0.4) u::n (0.6) toshokan and such one in like uh-huh library 46
itte: tte [katachi::: (.) de benkyooshita kara= go qt style in studied because
‘and in such a way like (0.4) uh huh (0.6) we went to libraries, and in that fashion we studied, and so’ 47 Kana:
[u:::n uh-huh ‘uh huh’
48 Emi: =yonaka toka made. .hh mo nanka (0.8) su- (0.8) midnight like until really like 49
tomodachi ga::, ruumumeeto ga tomodachi, = tomodachi ga:: friend S roommate S friend friend S
50 ma
yoowa benkyoo nakama mitaina kanji well in-short study companion like feeling na no ne. cop nml fp
51 [.hhh dakara:. so
The re-examination of so-called ‘clefts’
‘until like midnight. .hh really like (0.8) s- (0.8) your friend, your roommate is your friend, = your friend, well in short, is your study companion, something like that. .hhh so::’ 52 Kana: [u:::::n uh-huh ‘uh huh’ 53
(1.0)
54 Emi: u:::n sore ga chigatta yo ne::.= uh-huh that S differed fp fp 55 =nanka hitori janai kara:: like alone cop:top:neg because 56
benkyoo[shiteru °(no ga)°. studying nml S
‘uh huh that was very different. Cause like you are not alone, when studying.’ 57 Chika:
[aa::::: nakama ga iru kanji?= oh companion S exist feeling ‘oh:::: you feel like you have companions?’
58 Kana: =°u::n° ato ke- u:n kekkoo:: nanka- (0.4) ko uh-huh also uh-huh rather like daigaku de university at 59
kore o yaritai tte yuu node:::, kiteru ko ga >sugoi< this O do:want qt say because coming child S really
60 ooi ki ga shita... many feeling S had ‘uh huh also ke- uh huh rather like (0.4) thi- I felt that there are really a lot of them who come to the university because they want to do this.’ During Emi’s turn lasting from line 5 in Excerpt (11) until line 51 in Excerpt (13), Kana and Chika’s participations are limited to continuers, Kana’s short turns that contribute to the description of campus life in the U.S. and Chika’s reactive tokens and repair initiators. Indeed, Emi’s expansion of her telling includes more of a description of what she observed, or the environment in which American students are situated, rather than the expression of her thought, which she pre-announced to do in the -no wa topic clause.
Junko Mori
It is noticeable that there is one second of silence in line 53, when Emi ceases to continue her extended talk, but the other participants do not initiate their responses to her telling either. After this silence, Emi delivers the upshot of her lengthy description, i.e., that the environment in the U.S. which makes these students think that they are not studying alone is what is different from the environment surrounding college students in Japan. This upshot can be considered the content of what she thought concerning college students in the U.S., which was projected by the -no wa topic clause and the transitional remark from Chika’s prior telling to her own telling of the U.S. that Emi made in line 3 in (11). Finally, after this conclusive remark, which corresponds to the ways in which this extended talk was initiated, Kana launches her first full-fledged turn and shifts the topic. While the two focal cases examined in this section share certain features, including that the -no wa topic clause is used to initiate and project a forthcoming extended talk and that units of talk such as those marked by the tag-like expression janai or jan can be embedded in the projected turn space, they also exhibit differences. Most notably, the completion of Emi’s multiunit turn in (11) exhibits more ambiguity than that of Eri’s talk in (9). This ambiguity seems to have been caused in part by the fact that Emi does not produce a predicate that reiterates what was pre-announced in the -no wa topic clause and thereby explicitly marks the end of this multiunit turn. Another possibility that seems worth pursuing by collecting more cases is the relationship between the nature of the verbs used in the -no wa clause and their indexicality. That is, compared to the verb omotta (‘thought’) used in this case, bikkurishita (‘surprized’) used in the previous case or other types of emotive expressions appear to be more semantically loaded and therefore are able to offer clearer clues for the recipients to determine when the climax of a telling is delivered. 6. Conclusion This chapter examined how one type of so-called cleft construction is used in talkin-interaction. First, it demonstrated how rare it is for the speakers in talk-in-interaction to end the unit initiated with the -no wa topic clause with a clause marked by a nominalizer. Instead, the speakers often mark the end of the unit initiated with the -no wa topic clause by reiterating the same or a similar predicate used in the topic clause. Further, the chapter investigated how such a construction may influence the ways in which the conversational participants organize their interaction. That is, the -no wa topic clause appears to serve as a prospective indexical that pre-announces the nature of the forthcoming extended talk and provides the recipients an interpretive framework. As a result, the speakers may produce
The re-examination of so-called ‘clefts’
utterances marked by the tag-like expression jan(ai), which offer preliminary background information necessary for delivering what is pre-announced by the topic clause. The recipients, then, perceive this tag-like expression which occurs within a projected multiunit turn to be a place where their production of minimal tokens is relevant rather than full-fledged responses. Finally, the chapter demonstrated how important it is, when studying grammar-in-interaction, to pay attention to the concurring non-verbal behaviors that provide another frame of reference for designing and processing one’s turn. What is described in this study supports at least one of the arguments made by scholars who have established a research tradition that investigates the mutual relationship between interaction and grammar (e.g., Ochs et al. 1996; Selting & Couper-Kuhlen 2001; Ford et al. 2002). Namely, the data presented here clearly demonstrate how grammar serves as a resource for organizing talk-in-interaction. While the current study may not be able to present sufficient hard evidence for the other argument, i.e., grammar is a consequence of the necessities of social interaction, or the emergent grammar perspective (cf. Hopper 1988), it can at least support the idea put forth by Iwasaki and Ono (2002: 176), who argue that ‘it is possible that the spoken and written modes of language, while sharing a common core, may also differ in certain aspects of their grammars,’ citing Kokuritsu Kokugo Kenkyuujo (1963) and Miller (1995). How can we, then, account for the difference between grammars for spoken and written modes of a language, as well as their common core? With regard to the target construction examined in this chapter, there must be a certain shared sense of expected trajectory of talk for the conversational participants to project and anticipate further development of talk. But at the same time, this intersubjective understanding might be quite different than, or at least not identical to, what is described as the prescriptive ‘well-formedness’ of linguistic structures, which has been developed based on written data and introspection. Hopper (2004) indeed comments on the relationship between the two modes of grammars as it follows: ‘Most studies of the pseudocleft in spoken discourse have tacitly assumed that the spoken version is a reduced or “degenerate” form of the fuller construction found in writing and other planned discourse... I suggest that the reverse is the case: that prefabricated fragments of discourse are acquired for their usefulness in managing effective discourse, and that planned and written modes have normativized these bits and pieces into longer and more rule-governed syntactic constructions.’ (p. 173)
I hope the current chapter has succeeded in exhibiting how the speakers tend to treat the -no wa clause as a prefabricated fragment that projects an imminent multiunit talk, while a planned and written mode of Japanese discourse, as well as the prescriptive grammar primarily based on such data, tend to introduce the complete structure
Junko Mori
of cleft constructions as a norm. CA’s rigorous attention to social actions accomplished through talk-in-interaction should continue to aid the exploration of grammar as emergent practices that are acquired, employed, and modified by its users. Acknowledgement I would like to thank Kasumi Kato and students in my Japanese language courses for calling my attention to this phenomenon. Research for this article was funded in part by a grant from the United States Department of Education (CFDA 84.229, P229A020010-03). However, the content does not necessarily represent the policy of the Department of Education, and one should not assume endorsement by the Federal Government. Appendix: Transcription conventions and abbreviations Transcript symbols [ The point where overlapping talk and/or gesture starts (0.0) length of silence in tenths of a second (.) micro-pause less than 2/10 of a second underlining relatively high pitch CAPS relatively high volume :: lengthened syllable - cut-off; self-interruption = ‘latched’ utterances ?/./, rising/falling/continuing intonation respectively ! animated tone, not necessarily an exclamation ( ) unintelligible stretch (word) transcriber’s unsure hearings (( )) transcriber’s descriptions of events, including nonvocal conduct hh audible outbreath .hh audible inbreath (hh) laughter within a word > < increase in tempo, as in a rush-through ° ° a passage of talk quieter than the surrounding talk
The re-examination of so-called ‘clefts’
Abbreviations used in the interlinear gloss CAU causative suffix COP various forms of copula verb be DP dative particle FP final particle LK nominal linking particle NEG negative NML nominalizer
O PSS QT Q S TOP
object particle passive suffix quotative particle question particle subject particle topic particle
References Collins, Peter C. 1991. Cleft and Pseudo-Cleft Constructions in English. London: Routledge. DOI: 10.4324/9780203202463 Couper-Kuhlen, Elizabeth & Ono, Tsuyoshi. 2007. ‘Incrementing’ in conversation: A comparison of practices in English, German and Japanese. Pragmatics 17(4): 513–552. Davidse, Kristin. 2000. A constructional approach to clefts. Linguistics 38: 1101–1131. DOI: 10.1515/ling.2000.022 Ford, Cecilia E. 2004. Contingency and units in interaction. Discourse Studies 6(1): 27–52. DOI: 10.1177/1461445604039438 Ford, Cecilia E. & Thompson, Sandra A. 1996. Interactional units in conversation: Syntactic, intonational, and pragmatic resources for the management of turns. In Ochs, Schegloff & Thompson (eds), 134–84. Ford, Cecilia E., Fox, Barbara A. & Thompson, Sandra A. 1996. Practices in the contruction of turns: The ‘TCU’ revisited. Pragmatics 6: 427–454. Ford, Cecilia E., Fox, Barbara A. & Thompson, Sandra A. (eds). 2002. The Language of Turn and Sequence. Cambridge: CUP. Geluykens, Ronald. 1991. Discourse functions of it-clefts in English conversation. Communication and Cognition 24(3–4): 343–358. Goodwin, Charles. 1996. Transparent vision. In Ochs, Schegloff & Thompson (eds), 370–404. Goodwin, Charles. 2000a. Gesture, aphasia, and interaction. In Language and Gesture, David McNeil (ed.), 84–98. Cambridge: CUP. DOI: 10.1017/CBO9780511620850.006 Goodwin, Charles. 2000b. Action and embodiment within situated human interaction. Journal of Pragmatics 32: 1489–1522. DOI: 10.1016/S0378-2166(99)00096-X Goodwin, Marjorie Harness & Goodwin, Charles. 1986. Gesture and coparticipation in the activity of searching for a word. Semiotica 62: 51–75. DOI: 10.1515/semi.1986.62.1-2.29 Gundel, Jeanette K. 1977. Where do cleft sentences come from? Language 53: 543–559. DOI: 10.2307/413176 Harada, S.I. 2000 [1972]. Hobun ryooshiki no ‘no’. In Syntax and Meaning: S. I. Harada Collected Works in Linguistics, 439–447. Tokyo: Taishukan Publishing. Hayashi, Makoto. 2001. Postposition-initiated utterances in Japanese conversation: An interactional account of a grammatical practice. In Selting & Couper-Kuhlen (eds), 317–344. Hayashi, Makoto. 2003a. Joint Utterance Construction in Japanese Conversation [Studies in Discourse and Grammar 12], Amsterdam: John Benjamins. DOI: 10.1075/sidag.12
Junko Mori Hayashi, M. 2003b. Language and the body as resources for collaborative action: A study of word searches in Japanese conversation. Research on Language and Social Interaction 36 (2): 109–141. DOI: 10.1207/S15327973RLSI3602_2 Hayashi, Makoto. 2004a. Discourse within a sentence: An exploration of postpositions in Japanese as an interactional resource. Language in Society 33: 343–376. DOI: 10.1017/ S0047404504043027 Hayashi, Makoto. 2004b. Projection and grammar: Notes on the ‘action-projecting’ use of the distal demonstrative are in Japanese. Journal of Pragmatics 36: 1337–1374. DOI: 10.1016/j. pragma.2004.05.006 Hayashi, Makoto & Yoon, Kyung-Eun. 2009. Negotiating boundaries in talk. In Conversation Analaysis: Comparative Perspectives, Jack Sidnell (ed.), 248–276. Cambridge: CUP. DOI: 10.1017/CBO9780511635670.009 Herriman, Jennifer. 2003. Negotiating identity: The interpersonal functions of wh-clefts in English. Functions of Language 10(1): 1–30. DOI: 10.1075/fol.10.1.02her Herriman, Jennifer. 2008. The interpersonal function of clefts in English and Swedish. Languages in Contrast 8(2): 143–160. DOI: 10.1075/lic.8.2.02her Hopper, Paul. 1988. Emergent grammar and the a priori grammar postulate. In Linguistics in Context, Deborah Tannen (ed.), 117–134. Norwood NJ: Ablex. Hopper, Paul. 2004. The openness of grammatical constructions. Proceedings from the Annual Meeting of the Chicago Linguistics Society 40(2): 153–175. Hopper, Paul. 2007. Linguistics and micro-rhetoric: A twenty-first century encounter. Journal of English Linguistics 35(3): 236–252. DOI: 10.1177/0075424207305307 Iwasaki, Shimako. 2008. Collaborative Construction of Talk in Japanese Conversation. PhD dissertation, University of California, Los Angeles. Iwasaki, Shimako. 2009. Initiating interactive turn spaces in Japanese conversation: Local projection and collaborative action. Discourse Processes 46: 226–246. DOI: 10.1080/016385 30902728918 Iwasaki, Shoichi. 1997. The Northridge earthquake conversations: The floor structure and the ‘loop’ sequence in Japanese conversation. Journal of Pragmatics 28: 661–93. DOI: 10.1016/ S0378-2166(97)00070-2 Iwasaki, Shoichi. 2002. Japanese [London Oriental and African Language Library 5]. Amsterdam: John Benjamins. Iwasaki, Shoichi & Ono, Tsuyoshi. 1999. ‘Bun’ saikoo: Kaiwa ni okeru ‘bun’ no tokuchoo to nihongo kyooiku e no teian. In Gengogaku to Nihongo Kyooiku: Jitsuyooteki Gengo Kenkyuu no Koochiku o Mezashite, Yukiko Sasaki Alam (ed.), 129–144. Tokyo: Kuroshio. Iwasaki, Shoichi & Ono, Tsuyoshi. 2002. ‘Sentence’ in spontaneous spoken Japanese discourse. In Complex Sentences in Grammar and Discourse: Essays in Honor of Sandra A. Thompson, Joan L. Bybee & Michael Noonan (eds), 175–202. Amsterdam: John Benjamins. Jefferson, Gail. 1978. Sequential aspects of storytelling in conversation. In Studies in the Organization of Conversational Interaction, Jim Schenkein (ed.), 219–48. New York NY: Academic Press. Jefferson, Gail. 1984. Notes on some orderliness of overlap onset. In Discourse Analysis and Natural Rhetorics, Valentina D’Urso & Paolo Leonardi (eds), 11–38. Padova: CLEUP Editore. Jucker, Andreas H. 1997. The relevance of cleft constructions. Multilingua 16 (2–3): 187–198. DOI: 10.1515/mult.1997.16.2-3.187
The re-examination of so-called ‘clefts’ Kamio, Akio. 1990. Cleft sentences and the territory of information. In Interdisciplinary Approaches to Language: Essays in Honor of S. Y. Kuroda, Carol J. Georgopoulos (ed.) 353–371. Dordrecht: Kluwer. Kim, Kyu-hyun. 1995. Wh-clefts and left-dislocation in English conversation: Cases of topicalization. In Word Order in Discourse [Typological Studies in Language 30], Pamela A. Downing & Michael Noonan (eds), 247–296. Amsterdam: John Benjamins. Kokuritsu Kokugo Kenkyuujo (National Language Research Institute). 1963. Hanashi Kotoba no Bunkei, 2: Dokuwa shiryoo ni yoru kenkyuu. Tokyo: Shuuei Shuppan. Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39: 456– 516. DOI: 10.1515/ling.2001.021 Lerner, Gene H. & Takagi, Tomoyo. 1999. On the place of linguistic resources in the organization of talk-in-interaction: A co-investigation of English and Japanese grammatical practice. Journal of Pragmatics 31(1): 49–75. DOI: 10.1016/S0378-2166(98)00051-4 Linton, Larry D. & Lerner, Gene H. 2004. Before the beginning: Breath taking in conversation. Ms, University of California, Santa Barbara. McNeill, David. 1992. Hand and Mind: What Gestures Reveal about Thought. Chicago IL: University of Chicago Press. Miller, Jim. 1995. Does spoken language have sentences? In Grammar and Meaning: Essays in Honour of Sir John Lyons, Frank R. Palmer (ed.), 116–135. Cambridge: CUP. DOI: 10.1017/ CBO9780511620638.007 Mori, Junko. 1999. Negotiating Agreement and Disagreement in Japanese: Connective Expressions and Turn Construction [Studies in Discourse and Grammar 8]. Amsterdam: John Benjamins. DOI: 10.1075/sidag.8 Nakamura, Kanae. 2009. Interactive Negotiation of Perspectives in Japanese: Predicate-final Structure as a Resource to Organic Interaction. PhD dissertation, University of WisconsinMadison. Ochs, Elinor, Schegloff, Emanuel A. & Thompson, Sandra A. (eds). 1996. Interaction and Grammar. Cambridge: CUP. DOI: 10.1017/CBO9780511620874 Ono, Tsuyoshi & Suzuki, Ryoko. 1992. Word order variability in Japanese conversation: Motivation and grammaticization. Text 12: 429–445. DOI: 10.1515/text.1.1992.12.3.429 Prince, Ellen. 1978. A comparison of wh-clefts and it-clefts in discourse. Language 54(4): 883– 906. DOI: 10.2307/413238 Sacks, Harvey. 1974. An analysis of the course of a joke’s telling in conversation. In Explorations in the ethnography of speaking, Richard Bauman & Joel Sherzer (eds), 337–53. Cambridge: CUP. Sacks, Harvey. 1992. Lectures on Conversation. Oxford: Backwell. Sacks, Harvey & Schegloff, Emanuel A. 2002. Home position. Gesture 2(2): 133–146. DOI: 10.1075/gest.2.2.02sac Sacks, Harvey, Schegloff, Emanuel A. & Jefferson, Gail. 1974. A simplest systematics for the organiztion of turn-taking for conversation. Language 50: 696–735. DOI: 10.2307/412243 Schegloff, Emanuel A. 1980. Preliminaries to preliminaries: ‘Can I ask you a question?’ Sociological Inquiry 50(3–4): 104–52. DOI: 10.1111/j.1475-682X.1980.tb00018.x Schegloff, Emanuel A. 1982. Discourse as an interactional achievement: some uses of ‘uh huh’ and other things that come between sentences. In Georgetown University Roundtable on Language and Linguistics, Deborah Tannen (ed.), 71–93. Washington DC: Georgetown University Press.
Junko Mori Schegloff, Emanuel A. 1996. Turn organization: One intersection of grammar and interaction. In Ochs, Schegloff & Thompson (eds), 52–133. DOI: 10.1017/CBO9780511620874.002 Selting, Margaret. 2000. The contruction of units in conversational talk. Language in Society 29: 477–517. DOI: 10.1017/S0047404500004012 Selting, Margaret & Couper-Kuhlen, Elizabeth (eds). 2001. Studies in Interactional Linguistics [Studies in Discourse and Grammar 10]. Amsterdam: John Benjamins. DOI: 10.1075/ sidag.10 Streeck, Jürgen. 1988. The significance of gesture: How it is established. IPrA Papers in Pragmatics 2: 60–83. Streeck, Jürgen. 1993. Gesture as communication, I: Its coordination with gaze and speech. Communication Monographs 60: 275–299. DOI: 10.1080/03637759309376314 Tanaka, Hiroko. 1999. Turn-taking in Japanese Conversation: A Study in Grammar and Interaction [Pragmatics & Beyond New Serie 56]. Amsterdam: John Benjamins. Tanaka, Hiroko. 2000. Turn projection in Japanese talk-in-interaction. Research on Language and Social Interaction 33(1): 1–38. DOI: 10.1207/S15327973RLSI3301_1 Weinert, Regina & Miller, Jim. 1996. Cleft constructions in spoken language. Journal of Pragmatics 25: 173–206. DOI: 10.1016/0378-2166(94)00079-4
Activity, participation, and joint turn construction A conversation analytic exploration of ‘grammar-in-action’* Makoto Hayashi “Once we register that language figures in the actual, practical activities of the lives of people and societies, and that how the language is configured is more than incidentally related to its involvement in those activities, it is readily apparent that, at the very least, attention must be paid to what the relationship is between activity, action and the orderly deployment of language called grammar.” (Schegloff, Ochs & Thompson 1996: 21)
This chapter pursues an action-oriented view of language as a vehicle for participation in everyday communicative activities. To illustrate how local regularities in the deployment of language as forms of participation are shaped by the organization of communicative activities, the chapter examines ‘joint turn construction’ – a practice whereby a participant in conversation completes a grammatical unitin-progress initiated by another participant. The analysis shows, on the one hand, how grammar shapes relevant ways in which participants ‘package’ their action in completing another speaker’s turn-at-talk. It also demonstrates, on the other hand, how grammatical structure emerges as the outcome of situated actions of the participants.
* This chapter is based on a larger study on joint utterance construction in Japanese conversation (Hayashi 2003). I am grateful to John Benjamins for permission to reprint in this article some of the materials published in Chapter 3 of that earlier manuscript. I wish to thank the following people for their helpful comments at various stages of the development of my analyses presented in this article: Barbara Fox, Charles Goodwin, Kaori Kabata, Curtis LeBaron, Junko Mori, Tsuyoshi Ono and Hiroko Tanaka. Any remaining shortcomings are my responsibility.
Makoto Hayashi
1. Introduction The goal of this chapter is to explore the ramification of an interaction-oriented view of language for the study of grammar. The fundamental question that motivates this exploration is the issue of how we should conceptualize the relationship between grammar (or ‘structure’, more broadly) and situated human action. The relationship between structure and action, or the tension between them, has been much debated in various fields in the past several decades. In the framework of classical structuralism, structure has been conceptualized as externally and internally constraining rules that regulate situated human conduct, existing prior to, and independently of, local contexts of action. Such a conceptualization of a priori, autonomous structure exerting a dominant influence over situated action has been repeatedly questioned and challenged in a number of disciplines in the human and social sciences (Bourdieu 1977; Giddens 1979, 1984; Vygotsky 1978; Garfinkel 1967; Zimmerman & Boden 1991; Lave & Wenger 1991; Hopper 1987, 1988, 1998; Linell 1982, 2005; among others). One crucial problem with the structuralist view, it has been pointed out, is its failure to capture the relationship between structure and action in dynamic terms. This failure stems from the static dichotomies presupposed in the structuralist thinking, such as system vs. use, langue vs. parole, competence vs. performance, macro-system vs. micro-genesis, etc., where the former in each pair is viewed as an invariant, self-contained structure that is clearly delineated from the latter, while the latter is seen as playing no meaningful role in the constitution of the former. In reaction to this static view of the structure-action relationship, an alternative has been proposed that regards as central the essentially reflexive and dialectical relationship between structure and action (see the works cited above). In this alternative view, structure does not simply underlie action as its source and precondition, but it is also the outcome of emergent regularities in situated actions in local contexts of interaction. Structure and action are thus viewed as mutually shaping one another, with one feeding into the very constitution of the other. The present study pursues this alternative conceptualization and explores the reflexive relationship between grammar and action. The point of departure for this exploration is a shift away from the static view of language as close-knit systems of abstract forms, and a move toward an action-oriented view that sees language as a vehicle for participation in everyday communicative activities. To illustrate how local regularities in the deployment of language as forms of participation are shaped by the organization of communicative activities, I focus my analysis on ‘joint turn construction’ – a practice whereby a participant in conversation completes a grammatical unit-in-progress initiated by another participant. The following fragment provides an example of the practice in question (see the appendix for
Activity, participation, and joint turn construction
the glossary of the transcript symbols and abbreviations used in the excerpts cited in this paper): (1) [RKK 24] 1 Harumi: demo::: (.) °onna no hito de irezumi but female lk person cp tattoo no hito tte::° lk person qt “Bu:::t (.) °women with tattoos ((on their bodies))...” 2
(1.2)
3→Seiji:
mita koto nai.
saw event not.exist “((you)) have never seen”
Joint turn construction offers a fruitful site for exploring the interpenetration of grammar and action because we can directly observe how the details of the organization of activities-in-progress bear on the ‘grammatical meshing’ of separate speakers’ utterances within a single turn at talk. This chapter thus explores the workings of ‘grammar-in-action’ observed in the process of negotiating and achieving joint courses of action through co-constructing turns at talk in a variety of activity contexts. The database for the present study consists of 11 hours of naturally-occurring conversations among native speakers of Japanese (10 hours of videotaped face-toface conversations and one hour of audiotaped telephone conversations). The conversations took place among family members, friends, and acquaintances in causal settings. Some of the participants in the conversations are speakers of Tokyo Japanese, while others are speakers of the Kansai variety. 2. Activity and participation as analytic concepts The present study takes the notions of ‘activity’ and ‘participation’ as central for the analysis of grammar-in-action. Underlying this analytic focus on activity and participation is the idea that language is always situated in actual context of use and that its deployment constitutes social action. The view of language as social action has a rather long history and it has been pursued in a variety of disciplines, including anthropology (e.g., Malinowski [1935] 1978; Hymes 1972; Gumperz 1982), philosophy (e.g., Wittgenstein 1953; Austin 1962; Searle 1969), psychology (e.g. Vygotsky 1978), sociology (e.g., Goffman 1981a), and linguistics (e.g., Labov & Fanshel 1977; Levinson 1979). It is important to note here, however, that using the
Makoto Hayashi
notions of activity and participation as the frame of reference for analyzing language does not simply entail identifying individual utterances with certain kinds of action, such as a ‘promise,’ an ‘apology,’ etc. Rather, it gives us a way of thinking about larger frameworks within which interdependent actions operate coherently. In this way of proceeding with an analysis of language, one does not start with individual utterances, as most linguists and speech act theorists do, but with the larger courses of events – i.e., activities – that interacting participants engage in and build together by producing utterances and other conduct in concert with one another. Individual pieces of linguistic (and other) conduct, then, are seen as forms of participation in, and integral parts of, such activities. It is crucial to note, moreover, that the notion of activity should not be understood as referring to some static contextual frame within which participants are made to behave in certain ways. Rather, activity should be conceptualized as a temporally-unfolding, dynamic process that is organized by both the speaker and recipients on a moment-by-moment basis. To illustrate what this means, let us examine the following excerpt discussed by Goodwin and Goodwin (1987), in which a temporally-unfolding assessment activity is co-constructed by two participants who achieve intricate coordination of their conduct by reference to one another’s with split-second precision. (2) [Goodwin and Goodwin 1992a: 78; 1992b: 168] 1 Nancy: Jeff made an asparagus pie ((lowers upper ((nod with trunck)) eyebrow flash)) 2 it was s::so[: goo:d. [ 3 Tasha:
[I love it. °Yeah I love that. ((nods)) ((starts to withdraw gaze))
In this exchange, the details of the unfolding course of Nancy’s conduct in line 2 (e.g., the emerging syntactic structure of the form [it] + [copula] + [adverbial intensifier] + [assessment adjective], the prosodic emphasis on the intensifier so, the gestural marking of emphasis, etc.) progressively provide the recipient with resources to make inferences about what activity the speaker is engaged in at the moment. This temporally-unfolding activity-context allows Tasha to project with some precision what is going to happen next in the ongoing course of Nancy’s assessment. Tasha, then, utilizes this projection as a resource to organize her own conduct in such a way as to initiate an assessment of her own simultaneously with
Activity, participation, and joint turn construction
Nancy’s deployment of the assessment adjective good in line 2. By doing so, she accomplishes precisely-synchronized participation in the ongoing assessment. Central to understanding how a situated activity such as this is organized as a temporally-unfolding, interactively-constituted phenomenon is the notion of ‘participation’ (cf. Philips 1972; Goffman 1981b; C. Goodwin 1981; M. Goodwin 1990). When an activity is initiated, the participants align themselves toward the event-in-progress as well as toward one another in specific ways. This situated configuration of alignment proposes a particular framework for relevant participation by reference to which the participants organize their conduct and produce activity-relevant actions during the course of the activity. For instance, when a participant initiates a telling of a story, this emerging activity invokes a participation framework in which one can relevantly align oneself as a story recipient by producing uh-huh’s while the telling is in progress, as a ‘co-teller’ by joining in the ongoing telling, or as a ‘heckler’ by producing conduct that derails the intended course of the telling (cf. Sacks 1974; C. Goodwin 1984; Lerner 1992). It also projects what will constitute a relevant next action to be produced by co-participants, such as producing laughter at the climax of a story characterized as ‘funny.’ In the case of an assessment activity, it invokes a participation framework that proposes displays of agreement or disagreement to be relevant next actions by recipients (cf. Pomerantz 1978, 1984a). Hence, we can see that Tasha in fragment (2) above organizes her conduct by reference to such a participation framework invoked by the emerging course of Nancy’s assessment and accomplishes activity-relevant participation – a display of strong agreement – by producing a matching assessment of her own even before hearing the core of her interlocutor’s assessment (i.e., the adjective good in line 2). Note here that the details of Tasha’s conduct, including the precise placement of her utterance with regard to the emerging course of Nancy’s utterance and the accompanying head nods, are shaped by her orientation to accomplishing a particular kind of participation made relevant by the ongoing activity. Thus, analyzing utterances and other behaviors as forms of participation in situated activities sheds light on the ways in which interactional dynamics shape the details of how participants’ vocal and non-vocal conduct (including the grammatical structuring of utterances) figures in interaction. Recognition of the relevance of activity-contexts and participation frameworks to the details of how the grammatical structuring of utterances is configured is quite useful for the present investigation of grammar-in-action. In what follows, I examine instances of joint turn construction with a focus on the activities that the participants engage in and the participation frameworks such activities invoke. I also investigate how the current non-speakers make use of joint turn construction to accomplish delicately maneuvered participation in the activity-in-progress.
Makoto Hayashi
Through this examination, I demonstrate how the organization of activity and the grammatical organization of utterances mutually shape one another. 3. Differentiated participation in situated activities through joint turn construction In this section, I examine instances of joint turn construction in Japanese observed in a number of different activity-contexts. For the sake of convenience, in what follows I use the term ‘first speaker’ to refer to the participant who initiates a turn to be completed by another, and ‘second speaker’ for the participant who completes the first speaker’s turn-in-progress. The utterance that completes the first speaker’s utterance-in-progress is referred to as a ‘completing utterance.’ 3.1
Interactive achievement of shared perspectives
One common type of action that joint turn construction is employed to accomplish is demonstrating affiliation with another participant when such a display of affiliation is made relevant by the activity-in-progress (Lerner 1987; Antaki et al. 1996; Hayashi & Mori 1998). As discussed above, when a participant presents his/her stance or perspective toward some object or event, e.g., through an assessment, such a stance- or perspective-display regularly invokes a participation framework which makes it relevant for recipients to agree or disagree with the prior speaker (For English see Pomerantz 1984a; Goodwin & Goodwin 1987. For Japanese see Mori 1999). Among the variety of ways of ‘doing agreeing’ (e.g., producing an agreement token such as ‘yeah’ and the like), saying a version of ‘what the first speaker is going to say’ before he/she says it is a way for the second speaker to demonstrate (as opposed to merely claim) that he/she shares the same stance or perspective as the first speaker. In other words, joint turn construction can be used by participants to show each other that their minds are together on the issue being discussed. A close examination of the processes of achieving shared perspectives through joint turn construction reveals that, in Japanese, the second speaker regularly deploys certain types of grammatical endings in his/her completing utterance that are devoted to the work of ‘perspective sharing.’ Japanese is a ‘predicate-final’ language in which a predicate (typically a verb, but sometimes a predicate adjective or a predicate nominal) regularly occurs at the end of a clausal/sentential utterance. Therefore, joint turn construction is often realized by supplying a predicate that comes at the end of the ongoing utterance initiated by the first speaker. In addition, a clause/ sentence-final predicate in Japanese can be followed by so-called ‘utterance-final elements’ (Tanaka 1999), such as auxiliary verbs and sentence-final particles, which
Activity, participation, and joint turn construction
serve to indicate; i) the speaker’s affect, perspective, epistemic stance, etc., toward the proposition expressed, and ii) the speaker’s interpersonal sensitivity towards the addressee in the speech situation (Watanabe 1953; Haga 1954; Uyeno 1971; Cook 1990, 1992; Iwasaki 1993a, 1993b; Maynard 1993; Kamio 1994, 1997). This structural possibility of deploying epistemic and interpersonal stance markers after clause/sentence-final predicates allows Japanese speakers who perform joint turn construction to manage finely-tuned grammatical displays of particular ‘footing’ (Goffman 1981b; Antaki et al 1996) on which they complete another speaker’s utterance-in-progress. In other words, unlike joint turn construction in English, the negotiation of whose ‘voice’ the second speaker is animating through joint turn construction and what kind of participatory stance he/she is taking can be indicated by the grammatical form of the completing utterance in Japanese.1 In the instances examined for the present study, joint turn construction used for achieving ‘stance/perspective sharing’ regularly involves the delivery of the following types of clusters of utterance-final elements after a predicate: a. [predicate] + yo ne (or yo na) b. [predicate] + mon ne (or mon na) c. [predicate] + [tag-question-like element, such as jan(ai)] What is common in these grammatical endings is that they contain the elements of both the speaker’s assertion and a solicitation of acknowledgment from the recipient. For instance, in (a) and (b), the final particle yo and the nominalizer mon both indicate a sense of insistence in asserting a claim, while the final particle ne or na seeks agreement or confirmation from the recipient. In (c), assertion and agreement/confirmation-seeking are conflated in the tag-question-like expression jan(ai). The juxtaposition of these two elements, i.e., assertion and agreement/ confirmation-seeking, is important for accomplishing shared perspectives through joint turn construction for the following reasons. First, by invoking the sense of an assertion of a claim, the second speaker indicates that what he/she is voicing in the completing utterance is not merely a guess of what the first speaker would say to complete his/her utterance, but also the second speaker’s own assertion of a stance toward the issue under discussion. In other words, with yo or mon, the second speaker claims his/her ‘entitlement’ (Sacks 1992) to the content of the jointly produced utterance. On the other hand, by soliciting confirmation and agreement from the first speaker, the second speaker proposes his/her utterance as a collaborative (rather than an individual) assertion of a shared perspective, for which the 1. See also Morita (2002), who made similar observations regarding the relationship between the grammatical endings of completing utterances and the speaker’s ‘authority’ toward his/her interlocutors.
Makoto Hayashi
first speaker’s validation is indispensable. To put it in a slightly different way, by deploying such grammatical endings as yo ne, mon ne, and jan(ai), the second speaker embeds the ‘voices’ of both participants, i.e., the first and second speakers, within the completing utterance. And when the first speaker in the subsequent turn confirms the second speaker’s rendition of the ‘dual-voiced’ completion, a shared perspective is sequentially and interactively accomplished. Let us examine some instances. Fragment (3) presents an excerpt from a discussion about ‘tooth-brushing’ among four participants. Prior to the beginning of this excerpt, Muneo and Yurie stated that they used to have quite a few cavities because they did not like brushing their teeth. In response to this, the other two, Shoko and Kanji, state that they love brushing their teeth so much that their dentists told them that they did it too much – the enamel on their teeth was getting scoured off. Examine the following fragment to see how Shoko and Kanji interactively accomplish ‘alliance formation’ – an alliance of ‘lovers of tooth-brushing.’ (3) [TYC 31] 1 Shoko: ... .hh ha- haguki to ha no kokorahen ga gum and tooth lk around.here sp “... .hh The area around here between the gum and the teeth...” 2
be.scoured.off etc. “... gets scoured off, or something?”
3 Yurie:
“Oh:::::.”
4 Kanji:
kezuraresugi toka:?
a::::::. u:::::n na- (.) soo- tabun soo na n deshoo
yeah
ne::.=
so
probably so
fp “Yeah::: (.) ((I)) guess that’s probably what ((it)) is.”
5 Shoko: watashi mo so[o iwa]reru. I also so be.told “((They)) say that to me, too.” [ ] 6
Muneo:
7
cp n cp
[((sniff))]
(0.5)
8 Muneo: [( [
)] ]
Activity, participation, and joint turn construction
9 Kanji:
[tada nanka moo] tsurutsuru shite nai to: but like emp smooth is not if But, like, if ((the teeth)) aren’t smooth,”
da kara::] 10 (0.3) i[ya don’t.like cp because “(0.3) ((I)) would hate ((it)), so...” [ ]
11 Shoko: 12 Kanji:
] ne::. [soo desu yo so cp fp fp “((That))’s right, isn’t it?”
yappari tetteeteki ni:: (0.3)
as.expected thoroughly “... you know, thoroughly:: (0.3)”
13→Shoko: migakimasu yo [ne::.] brush fp fp “((we)) brush ((our teeth)), don’t ((we))?” ] [ 14 Kanji:
[°u:::]::n.° “°Yeah:::::°”
In lines 1–2, Shoko asks Kanji for confirmation about what his dentist told him about brushing too much. Kanji offers confirmation in line 4 and Shoko seconds it by saying, watashi mo soo iwareru (‘((They)) say that to me, too’), in line 5. Kanji then states that he would not like it if his teeth were not smooth (lines 9–10) and Shoko agrees rather emphatically with soo desu yo ne:: (‘((That))’s right, isn’t ((it))?’) at line 11. In these exchanges between Shoko and Kanji, we already observe moves of affiliation (e.g., lines 5 and 11) that display their alignment as ‘lovers of tooth-brushing.’ It is in this activity-context that joint turn construction is employed to collaboratively achieve a shared perspective. At line 13, Shoko supplies a predicate (migakimasu ‘to brush’) that completes Kanji’s utterance-in-progress in line 12. Note here that Shoko produces the combination of the final particles yo + ne after the predicate. As described above, by deploying these particles, Shoko presents the predicate migakimasu (‘to brush’) as her own assertion of a stance, on the one hand, and proposes it as a shared perspective that requires Kanji’s confirmation, on the other. The sense that the yo ne ending conveys is roughly expressed in the translation with the pronoun ‘we’ (which is not present in the original Japanese) rather than ‘you,’ and the tag question, ‘don’t we?’ In line 14 Kanji confirms Shoko’s completion and through this sequence of actions by the two participants, a shared perspective is interactively and collaboratively achieved.
Makoto Hayashi
Fragment (4) provides another instance of stance/perspective sharing through joint turn construction. This fragment is taken from a conversation among four middle-aged women in which the participants talk about various issues and concerns related to middle-aged women. The following portion of the conversation presents a part of storytelling by Emiko, who describes her experience at a boutique to the other participants. Prior to the beginning of the fragment, Emiko mentioned that a sales woman recommended a dress with a belt around the waist, which, in Emiko’s opinion, does not look good on a middle-aged woman’s ‘lumpy’ body because the belt would only make visible the layers of body fat. In lines 1–2, 4, 6–7, Emiko states that such a dress would look good on someone who is tall and thin. In the course of this assessment utterance by Emiko, one of the recipients, Yaeko, accomplishes activity-relevant participation – display of agreement and alignment with Emiko as another middle-aged woman – by co-constructing a shared stance toward the event under discussion through joint turn construction (see lines 11–12, 14, and 16). (4) [OBS 6] 1 Emiko: are wa ne:, hoso:kutte NE:, (0.3) that tp fp thin:and fp 2
ano uwazee ga atte:,= uhm height sp exist:and “If ((you)) were thin and (0.3) uhm, tall, and”
3 Yaeko: =so[o da yo ne:, so cp fp fp “That’s right.”
[
4 Emiko: [zentai ga:,= whole sp “((your)) whole body...” 5 Yaeko: =soo soo. so so “Right, right.” 6 Emiko: u- ano: hosoi kara↑, .hh beruto shitemo uhm thin because belt do:even.if 7 su[teki na n] da kedo:= nice cp n cp although “... were thin, then ((you)) would look nice in that ((sort of dress)), even if ((you)) wore a belt, bu:t,”
Activity, participation, and joint turn construction
[
]
8 Yaeko: [u::n u:n.] “Uh huh, uh huh.” 9 Hanae: =u:n. “Uh huh.” 10 Emiko: ((TSK))hn. “((TSK)) hn” 11→Yaeko: obasan no ne, middle-aged.woman lk fp 12→
bo[kon ga ne, mata yokee [bokon mim(lumpy) mim sp fp also extra “middle-aged women will only get...”
]
[
]
[
13 Emiko: [obasan no [bokon bokon shita no ga sa:,] middle-aged lk mim mim is n sp fp woman “middle-aged women’s lumpy ((body)) will...” 14→Yaeko: =bokon b(h)ok(h)on te [hitotsu huechau mim mim qt one increase dake= only “... another layer on ((their)) lumpy ((bodies))...”
[
15 Emiko: [so. “Right.” 16→Yaeko: =[da mon ne(hh). cp thing fp “... won’t ((they))?”
[
17 Hanae: =[(ikani)mo ne::: heh heh heh indeed fp “Indeed. heh heh heh” During the course of the first part of Emiko’s assessment utterance (lines 1–2, 4, and 6–7), we already observe affiliative moves by Yaeko, who produces a number of agreement/acknowledgment tokens (lines 3, 5, and 8). In lines 11–12, 14, and 16, Yaeko upgrades the display of her agreement/affiliation by providing a completion
Makoto Hayashi
to Emiko’s utterance-in-progress.2 In this completing utterance, Yaeko offers her understanding of what would happen if a middle-aged woman were to wear a dress with a belt around the waist. Note that, by deploying the combination of mon + ne at the end of her utterance (line 16), Yaeko presents her utterance not only as her understanding of what Emiko was going to say, but also as an assertion of her independent understanding of the situation, i.e., an independent understanding as another middle-aged woman who faces similar issues and concerns as Emiko and who shares the same perspective toward the kind of event that Emiko is commenting on. Thus, through Yaeko’s joint turn construction with these specific utterancefinal elements, as well as Emiko’s acknowledgment (line 15), the two participants manage an interactive accomplishment of alliance as middle-aged women who share the same concerns about their bodies and clothes. In this subsection, I examined two instances of joint turn construction performed in the context in which the first speaker engages in the activity of presenting his/her stance or perspective on some object or event. Such an activity-context invokes a participation framework which makes it relevant for recipients to agree or disagree with the speaker. Within such a framework for relevant participation, a recipient can employ joint turn construction as a device to interactively achieve ‘shared perspective’ with the speaker. Specifically, I showed that the second speaker’s participatory stance toward achieving shared perspective with the first speaker shapes the grammatical endings of his/her completing utterance in a particular way, i.e., the deployment of utterance-final elements that serve to present the second speaker’s completing utterance as voicing the two participants’ perspectives embedded in it. And when the first speaker confirms the completing utterance, a shared perspective is sequentially and interactively accomplished. In the next subsection, I examine instances in which the second speakers employ joint turn construction to display their empathetic understanding in the context in which the first speaker engages in the telling of his/her personal (past) experience. 2. Here, Yaeko accomplishes joint turn construction by providing the second component of a ‘compound turn constructional unit’ (Lerner 1991) in the form of [X-kedo] + [Y] (‘[Although X] + [Y]’). One reader pointed out that kedo has come to be used as a kind of final particle in conversational Japanese and that the presence of [X-kedo] does not always project a subsequent production of the [Y] component. That being true, what is important in this instance is the fact that the compound structure provides a resource for Yaeko to achieve joint turn construction irrespective of whether Emiko intended to terminate her turn after [X-kedo] or continue on to produce the [Y] component. In other words, as will be seen in some other instances below as well, quite independent of the first speaker’s intention, a co-participant can act on the publicly available structure of the talk and build his/her participation in a manner that is relevant to the ongoing activity.
Activity, participation, and joint turn construction
3.2
Differentiated displays of empathetic understanding of another’s experience
Another type of action for which joint turn construction is commonly employed is displaying an empathetic understanding of another participant’s experience. The activity-context in which this type of joint turn construction is observed is different from that of stance/perspective sharing described above. In the cases of stance/ perspective sharing, the first speaker presents (or is about to present) his/her perspective or stance toward the kind of event that, in principle, is available for an independent evaluation by the second speaker. In other words, the two participants have basically equal access to the event toward which a shared stance is negotiated through joint turn construction. In the cases of displaying an empathetic understanding of another’s experience, on the other hand, the first speaker engages in the activity of telling about his/her personal (past) experience, which in principle is only available to the person who experiences (or experienced) it. What the second speaker negotiates in this environment, then, is a display of a ‘vicarious’ understanding of someone else’s specific experience which is essentially unavailable to him/her. Let us examine an instance. Fragment (5) (presented above in Fragment (1)) is taken from a larger sequence of talk in which the participants discuss the kinds of people that they see in a sentoo (‘public bathhouse’) in Japan. Prior to the beginning of the fragment, the two male participants (Seiji and Akira) stated that they often see people with tattoos on their back – typically associated with yakuza (‘Japanese mafia’) members. In lines 1–2, Harumi, a female participant, starts to talk about her experience regarding women with tattoos on their bodies in public bathhouses. Examine Seiji’s joint turn construction at line 4. (5) [RKK 24] 1 Harumi: demo::: (.) °onna no hito de irezumi no but female lk person cp tattoo lk 2
person qt “Bu:::t (.) °women with tattoos ((on their bodies))...”
3
(1.2)
4→Seiji:
saw event not.exist “((you)) have never seen”
hito tte::°
mita koto nai.
5 Harumi: u:::n. “Right.”
Makoto Hayashi
Harumi’s utterance in lines 1–2 is an initiation of a telling of her personal experience. Among the relevant actions that the recipients can do as a response to such a telling is displaying their understanding of what is being told and Seiji accomplishes this by employing joint turn construction in line 4. Note here that the personal experience being told is in principle inaccessible to Seiji, especially given that the experience being discussed pertains to the female section of the public bathhouse which is off limits for men. Thus, what Seiji engages in by supplying a predicate that completes Harumi’s utterance-in-progress is offering his candidate, empathetic understanding of Harumi’s personal experience, rather than sharing a perspective toward it. Harumi then confirms his candidate understanding in the subsequent turn. The difference between a display of empathetic understanding of another’s experience and stance/perspective sharing manifests itself also in the grammatical endings employed in the completing utterance. As discussed in the previous subsection, elements occurring after the predicate in Japanese often indicate the speaker’s attitude or stance toward the content of the preceding statement and/or toward the interlocutor(s). And we have seen that the instances of stance/perspective sharing exhibit particular types of grammatical endings occurring after the predicate in the completing utterance. In the instances of displaying empathetic understanding, on the other hand, such stance-displaying grammatical endings are typically lacking. Particularly, the sense of asserting his/her own point of view, which would be conveyed by yo and mon in the instances of stance/perspective sharing discussed above, is consistently absent in instances of empathetic understanding. Thus, in fragment (5), Seiji’s completing utterance is delivered with no utterance-final elements that display the second speaker’s stance. Especially since it lacks utterance-final elements indicating insistence or an assertion of one’s own entitlement to the content of the utterance (e.g., yo and mon), the second speaker’s completing utterance is produced and treated simply as a rendition of what would have been said by the first speaker, which in turn makes relevant the first speaker’s confirmation or disconfirmation of that rendition. Thus, in contrast to the ‘dualvoiced’ character of the completing utterance in stance/perspective sharing which embeds the voices of both participants in it, the second speaker who engages in a display of empathetic understanding simply ‘animates’ (in Goffman’s (1981b) sense) the first speaker’s voice in his/her completing utterance and subjects it to confirmation or disconfirmation by the first speaker in the next turn. The next fragment presents a case in which a co-participant engages in transforming a telling of an essentially unavailable, ‘particular’ personal experience of another participant into a sharing of an accessible, ‘generic’ experience via joint turn construction. In this instance, the second speaker manages a finely-tuned display of her understanding of selective aspects of the first speaker’s experience
Activity, participation, and joint turn construction
– aspects that transcend the particularity of individual experiences and that are generic enough to be accessible to others. In fragment (6), Hayao engages in the activity of telling his recipient, Izumi, about what he experienced at their mutual female friend’s wedding. Note that Izumi did not attend the wedding and has never seen the brother of the mutual friend’s that Hayao talks about in his telling. Therefore, the content of Hayao’s telling – i.e., the striking resemblance that he observed between their mutual female friend and her younger brother at the wedding – is only accessible to Izumi through Hayao’s talk. (6) [TI 14] 1 Hayao: honma sono KUse made sokkuri tte yuu really uhm habit even exactly.same qt say “Really, uhm, the fact that even ((his)) little habits are...” no [wa yappari ne: nakanaka ano::: .hhh] 2 N tp you.know fp rather uhm “...exactly the same ((as his sister’s)) is, you know, like, uhhhm, .hhh”
[
]
3 Izumi:
[he:::::::::::::::::::::::::::::::::.] “Oh wo:::::::::::::::::::::::::::::::::::::::::::::::::w”
4 Hayao: mitete ne::, be.seeing fp “...while seeing ((it))” 5 Izumi: u:::n.hh= “Uh huh. hh” 6 Hayao: =m(h)oo: chotto:= emp a.little “...l(h)ike” 7→Izumi:
=waratte shim(h)a(h)u hh hh .hh laugh can’t.help “((you)) can’t help l(h)augh(h)ing hh hh .hh”
In lines 1–2, 4, and 6, Hayao reports that even the little habits are almost identical between the two siblings. He bases his telling on his actual experience of having observed the resemblance at the wedding (mitete ne::, ‘while seeing ((it))’ in line 4). Izumi then delivers a predicate at line 7 that completes Hayao’s utterance-in-progress, thereby displaying her empathetic understanding of the situation being told about. Recall that Izumi does not have the kind of direct experience that Hayao possesses to draw on when producing this completing utterance. Nonetheless, she
Makoto Hayashi
accomplishes a delicately maneuvered display of her understanding of a selective aspect of the event being recounted. Note that Izumi uses the non-past predicate form, waratte shimau (‘can’t help laughing’), rather than the past-tense form, waratte shimatta (‘couldn’t help laughing’), in completing Hayao’s utterance-inprogress. By using the non-past form, she constructs her completing utterance as an understanding of a generic experience, i.e., how anyone (including herself) would react if he/she were in the same situation. If Izumi had deployed the pasttense form of the predicate, waratte shimatta (‘couldn’t help laughing’), at line 7, it would be heard as an offer of her candidate understanding of Hayao’s past experience, as in “you, Hayao, couldn’t help laughing on that particular occasion in the past when you saw the resemblance.” Thus, while Izumi has only limited access to the particular experience that Hayao had, she nonetheless exhibits access to a selective aspect of the experience, i.e., a generic aspect that anyone would share if he/ she were in the same situation.3,4 This instance documents a process whereby a world of ‘shared experience’ – experience that transcends the particularity of private experiences and is ‘identical for all practical purposes’ (Schutz 1962; Garfinkel 1967; Heritage 1984) among different individuals – is brought into being through a finely-differentiated display of empathetic understanding via the grammatical design in the completing utterance.5 In this subsection, I examined how the second speakers, in the course of negotiating joint participation in the first speaker’s telling of his/her personal experience, accomplish finely-differentiated displays of their empathetic understanding of the first speaker’s personal experience. We observed that the second speakers’ orientation to the inaccessibility of the first speaker’s personal experience shapes the grammatical design of their completing utterances, especially in terms of the absence of utterance-final elements that convey the sense of claiming 3. In other words, the ‘you’ in the translation of line 7 does not (solely) refer to the interlocutor, Hayao, but to any individual who would find him/herself in such a situation, i.e., the generic ‘you.’ Note that, in the Japanese original, the subject of the predicate does not have to be, and is not, expressed, and therefore, the transformation from the personal to the generic is achieved solely through the tense form of the predicate. 4. One reader pointed out that, at least on the written transcript, it is possible to interpret Izumi’s utterance in line 7 as her own reaction to Hayao’s telling, i.e., “I (=Izumi) can’t help laughing!” Although I lack technical vocabulary to defend my interpretation, the prosody with which line 7 is said makes it sound like a proffer of her candidate understanding, rather than an expression of her reaction to the ongoing telling. 5. Goodwin and Goodwin (1987) discuss similar cases from English conversations in which two participants who engage in a concurrent assessment activity manage fine displays of differential access to the event being assessed. There as well, different tense forms are mobilized to accomplish finely-tuned displays of differential access.
Activity, participation, and joint turn construction
one’s entitlement to the content of the jointly produced utterance. We also observed that co-participants can mobilize grammatical and other resources to accomplish the transformation of a telling of a private personal experience into a sharing of the understanding of generic aspects of the experience. The delicately-engineered interactive process observed above might be the participants’ enactment of what Schutz (1962) would call ‘reciprocity of perspectives;’ that is, despite their non-identical experiences and despite their lack of access to the full particularity of one another’s experiences, participants still proceed on the assumption that their experiences are ‘identical for all practical purposes’ and can achieve shared understanding that transcends individuals’ private experiential worlds. Next, I examine instances in which a ‘teaming-up’ of two participants is accomplished through joint turn construction. 3.3
Assisted explaining
So far I have examined cases in which the second speaker who produces the completing utterance is not only an addressed recipient of the first speaker’s ongoing utterance but also directs the completing utterance to the first speaker. The configuration of this type of joint turn construction can be represented by the following diagram (See Figure 1). In this subsection, I discuss instances in which the second speaker who produces the completing utterance is a non-addressed participant, and the completing utterance is directed to a third party who is the addressed recipient of the first speaker’s ongoing utterance.6 Schematically (See Figure 2). ongoing utterance A
B completing utterance
A = first speaker B = second speaker [addressed recipient of A’s talk]
Figure 1.
6. It is Gene Lerner’s work (Lerner 1987, 1996; Lerner & Takagi 1999) on joint turn construction in English that first brought my attention to the difference in the ‘directionality’ of joint turn construction in my Japanese cases.
Makoto Hayashi ongoing utterance C
A
completing utterance
B A = first speaker B = second speaker [non-addressed participant] C = third party [addressed recipient of A’s and B’s talk]
Figure 2.
As seen in the diagram above, this type of joint turn construction occurs in multiparty (i.e., more-than-two-party) conversation. And it is used to form a local alignment of two (or more) participants as an ‘interactional team’ vis-à-vis a third party. One common activity-context in which such ‘teaming-up’ is observed is one in which some sort of ‘explaining’ takes place. On such occasions, the second speaker uses joint turn construction to accomplish participation in the first speaker’s explaining as a ‘co-explainer,’ addressing the completing utterance to a thirdparty explainee. I adopt the term ‘assisted explaining’ from Lerner and Takagi (1999) to refer to this use of joint turn construction. Let us examine some instances of interactional teaming-up in assisted explaining. Fragment (7) is taken from a conversation in which two married couples are talking. In this segment, Yurie and her husband Kanji are explaining to the other couple their experience of having had trouble renting a wedding dress of the right size for Yurie. (7) [TYC 44] 1 Kanji: 2
saizu: ga: ne::, (.)
size sp fp “The ((right)) size, ((you)) know, (.)” ano yappa
chicchai kara[::]
uhm as.expected small because “uhm, ((you)) see, since ((she))’s small,”
[ ]
Activity, participation, and joint turn construction
3 Shoko:
[u:]:::::n. “Uh huh”
4→Yurie:
[(na)i: n desu yo::::::.] not.exist N cp fp “isn’t there.”
[
5 Kanji:
[sentakushi ga::, kagira]retete. options sp be.limited:and “((our)) options were limited, and,”
]
In this instance, we observe the co-participant’s finely-tuned ‘grammatical work’ for accomplishing the interactional task of assisted explaining. In line 1, Kanji initiates an utterance about the size of wedding dresses by producing a nominal (saizu ‘size’) marked with the subject particle ga. After a micro-pause, he constructs a kara-clause (roughly equivalent to a ‘because’-clause in English), in which he presents Yurie’s small stature as a reason for something. After Shoko’s acknowledgment token in line 3, then, Yurie starts up and participates in the talk-in-progress as a ‘co-explainer’ of the situation under discussion through joint turn construction. She provides a predicate for the subject nominal that Kanji constructed in line 1 (saizu: ga: ne:: + (na)i: n desu yo::::::. ‘There’s no ((right)) size’) and brings Kanji’s utterance-in-progress to completion. It is interesting to compare Yurie’s completion in line 4 and Kanji’s own completion in line 5 in terms of how these two participants contextualize the grammatical trajectory of the utterance-so-far. As discussed above, Yurie builds her utterance in line 4 as a continuation and completion of Kanji’s utterance in line 1. Through this completion, her utterance in effect retroactively contextualizes the kara-clause in line 2 as a clause-internal insertion. Figure 3 represents this operation schematically. On the other hand, Kanji’s utterance in line 5 produced in overlap with Yurie’s completing utterance contextualizes the utterance-so-far slightly differently. That is, Kanji’s own completion of his utterance-in-progress only provides a continuance of what he stated in line 2 (ano yappa chicchai kara:: [line 2] + sentakushi ga::, kagiraretete [line 5] ‘Uhm, ((you)) see, since ((she))’s small, ((our)) options were limited, and’), and is structurally disconnected from the utterance in line 1. Thus, Kanji’s completion in line 5 retroactively displays that he has abandoned the subject nominal that he constructed in line 1 (saizu: ga: ne:: ‘The ((right)) size, ((you)) know,’). See Figure 4.
Makoto Hayashi Line 1:
Line 4:
saizu: ga: ne::, The ((right)) size, ((you)) know,”
(na)i: n desu yo::::::. “isn't there.”
SUBJECT
PREDICATE
Line 2: ano yappa chicchai kara:: “uhm, ((you)) see, since ((she))’s small,”
Figure 3.
Line 1:
Line 2:
saizu: ga: ne::, “The ((right)) size, ((you)) know,”
ano yappa chicchai kara:: “uhm, “((you)) see, since ((she))’s small,”
retroactively displays that line 1has been abandoned Line 5: Sentakushi ga::, kagiraretete. “((our)) options were limited, and,”
Figure 4.
In this instance, then, quite independent of the original speaker’s intention (i.e., the abandonment of the subject nominal in line 1), a co-participant acts on the publicly available structure of the talk that has already been produced and builds her participation in the ongoing activity as a co-explainer based on that structure. This instance can thus be seen as suggesting that language in interaction does not (simply) reside in the speaker’s mind, but exists between participants, so to speak,
Activity, participation, and joint turn construction
as a publicly available, social structure that mediates multiple participants’ action in concert with one another. Before moving on to discuss another subtype of joint turn construction used for assisted explaining, a note on the grammatical ending of the completing utterance observed in fragment (7) is in order. Here again, we observe that the second speaker’s participatory stance toward the activity-in-progress shapes the grammatical endings of his/her completing utterance in a particular way. Notice that in fragments (7), the second speaker deploys the particle yo at the end of her completing utterance (line 4). As stated above with regard to the sequence of yo and ne (Section 2.1), the particle yo indicates the sense of insistence in making one’s own assertion and of claiming his/her entitlement to the content of the utterance that it marks. It was argued there that yo occurring in the second speaker’s completing utterance serves to present his/her own voice in the completing utterance. Note then that, unlike the cases of stance/perspective sharing in which yo co-occurs with another utterance-final element, ne, which serves to solicit confirmation or agreement from the addressee, the completing utterance in fragment (7) only has the particle that presents the speaker’s voice and lacks any utterance-final elements that seeks the addressee’s confirmation. This presents a markedly different situation from the cases discussed above, especially those that involve interactive achievement of shared perspective or shared experience, in which a confirmation of the second speaker’s completing utterance by the first speaker is solicited as an indispensable part of the constitution of the action being performed through joint turn construction. It is important to recall here that, in the cases of assisted explaining, the second speaker participates in the ongoing activity of explaining as a ‘co-explainer,’ i.e., someone who has virtually equal access and entitlement to what is being explained with the first speaker. Also important is the fact that the second speaker addresses his/her completing utterance not to the first speaker whose utterance is being completed, but to a third party who is assumed not to have prior access to the event being explained. Within this participation framework, then, the second speaker is entitled to voice a completing utterance as her own assertion (i.e., with yo), while she does not need to solicit confirmation from the addressee (i.e., with ne). Thus, here again, unlike assisted explaining through joint turn construction in English (cf. Lerner & Takagi; 1999), we can see how, in Japanese, the participation framework in which completing utterances are produced for achieving assisted explaining has a direct consequence for the grammatical composition of those completing utterances. Let us now move on to discuss a slightly different type of joint turn construction used for assisted explaining. In the previous example, the co-members of the interactional team formed through joint turn construction have virtually equal access to the event that they explain. In the next example, we observe that the first
Makoto Hayashi
and second speakers have different levels of access/entitlement to the event being explained. More specifically, it will be seen that the first speaker (A) voices information/experience/etc. that ‘belongs to’ a co-present participant (B), and directs the utterance to a third party (C). Then, the co-present participant (B) whose information/experience/etc. is being discussed joins the first speaker’s (A) explaining by means of joint turn construction. It may be useful here to invoke the notions of ‘AB-events’ and ‘B-events’ proposed by Labov and Fanshel (1977) to describe the difference between the two types of assisted explaining being discussed. Labov and Fanshel (1977) set up a taxonomy of events that are talked about in interaction in terms of the distribution of knowledge among the participants. Thus, an ‘A-event’ is an event or piece of information to which the speaker (A) has more access than a co-participant (B) – typically the recipient. A ‘B-event’ is an event or piece information to which a co-participant (B) – again, typically the recipient – has more access than the speaker (A) does. An ‘ABevent’ is an event or piece of information to which both the speaker (A) and a coparticipant (B) have equal access. Accordingly we can describe the two types of assisted explaining as follows: In the first type of assisted explaining observed in fragment (7), the first speaker (A) explains an ‘AB-event’ to a third party (C) and the co-participant (B) who has equal access to the event joins the first speaker in the explaining. In the second type being discussed here, on the other hand, the first speaker (A) explains a ‘B-event’ to a third party (C) and the co-participant (B) who has more access to the event joins the first speaker in the explaining. Let us examine an example. In fragment (8), Sanae talks about the co-present participant Ryoko’s past experience of having done some volunteer work for the homeless, but she directs this talk to a third party, Tomoe (not shown in the transcript). In line 5, Ryoko, whose experience is being discussed, comes in with an utterance that completes Sanae’s ongoing utterance and thereby accomplishes her participation in the talk as a ‘co-explainer.’ (8) [HR 6] 1 Sanae:
soo ryoko chan nanka ippai sonna n
so
Ryoko tl like a.lot such N shite:,=ano:: do:and uhm “Right, Ryoko does that kind of thing a lot,=uhhm”
.hhh (0.3) kama:- kamagasaki no:, 2 Kamagasaki lk “.hhh (0.3) in Kama:- Kamagasaki,”
3 Ryoko: u:n. “Uh huh”
Activity, participation, and joint turn construction
4 Sanae:
takidashi toka mo[:
food.drive etc. also “... a food drive, also,”
[
5→Ryoko:
[u::n. [ikkai itta:. once went “Uh huh. ((I)) went to, once.”
[
6 Sanae:
[itta n ya tte. went N cp qt “((she)) went to, ((I)) heard.”
Labov and Fanshel (1977: 100) point out that when speaker A makes a statement about a ‘B-event,’ it is hearable as an invitation for co-participant B to produce an utterance of confirmation.7 To put it in another way, the voicing of words that pertains to a ‘B-event’ by speaker A invokes a participation framework that makes relevant for the ‘author/owner’ of the statement (B) to confirm the words produced by the ‘animator’ (A). In the instance above, Sanae’s voicing of a ‘B-event’ that pertains to Ryoko’s past experience makes it relevant for Ryoko as the author/ owner of the statement to provide confirmation of the content of the ongoing telling by the animator, Sanae. Within this framework of relevant participation, Ryoko provides confirmation first through an acknowledgment token (u::n) and then by supplying a completion for Sanae’s utterance-in-progress. Note here that Ryoko could have confirmed Sanae’s telling by producing an acknowledgment token after Sanae completes her telling. Instead, by providing a completion to Sanae’s utterance and addressing it to the third-party addressee, Ryoko displays her participatory stance as a ‘co-explainer’ who takes an active part in constructing the telling toward the third party. Thus, in this instance, we see that joint turn construction is used as a device for the ‘author/owner’ of an experience being discussed to join in the explaining acitivty-in-progress by forming a local alignment of an ‘interactional team’ with another participation vis-à-vis the third-party explainee.8 7. A similar observation is made by Pomerantz (1980), who shows that participants methodically use their ‘limited access’ to information as a device to elicit that information from the party who has more access to it. 8. Note that the difference in the participatory status between the first and the second speakers manifests itself in the grammatical forms of the completions that they produce in lines 5 and 6. Thus, while Ryoko produces the simple past-tense form of a verb (itta ‘went’), Sanae displays her participatory role as the ‘animator’ by adding the quotative particle tte at the end of her completion in line 6 and thereby marking her utterance as hearsay.
Makoto Hayashi
In this subsection, I described two types of assisted explaining achieved through joint turn construction that occur in different participation frameworks. The difference pertains to the distribution of knowledge between co-explainers about the matter being discussed. The first instance showed a case in which the co-members of the explaining party participate in the activity of explaining as those who have virtually equal access to the event being relayed to a third party. In the second instance, the first speaker who has less access to the event being explained engages in a ‘B-event’ telling and thereby invites a co-participant who has more access to the event to participate in the ongoing activity as the author/owner of the telling. In both types, the practice of joint turn construction allows for the interactive construction of a local alliance between two participants vis-à-vis a third party. In the next, final subsection, I examine instances in which joint turn construction is used as a way to convert a dispreferred action-in-progress into a preferred action. 3.4
Converting a dispreferred action to a preferred action
Conversation analysts have demonstrated (e.g., Pomerantz 1978, 1984a, 1984b; Davidson 1984; Atkinson & Drew 1979; Schegloff, Jefferson & Sacks 1977; Sacks 1987[1973]; Jefferson 1987; Mori 1999, among others) that “there is a ‘bias’ intrinsic to many aspects of the organization of talk which is generally favourable to the maintenance of bonds of solidarity between actors and which promotes the avoidance of conflict” (Heritage 1984: 265) and that the participants have available a variety of systematic practices by which they can organize their conduct to contribute to the maintenance of social solidarity. Thus, when an action that could potentially cause a conflict among the participants (i.e., a so-called dispreferred action) is imminent in a given interactional context, its producer and/or recipient(s) can mobilize those practices to prevent it from being (fully) actualized. In this subsection, I examine some ways in which the recipient of an imminent dispreferred action mobilizes joint turn construction to preempt the production of the dispreferred action by the prior speaker. Based on data from English conversation, Lerner (1987, 1996) reports that, in the course of a speaker’s delivery of an utterance that is recognizably embodying an imminent dispreferred action (e.g., disagreement, other-correction, etc.), the addressee can make a mid-turn entry and co-opt the completion of the first speaker’s utterance-in-progress, thereby preempting the emerging dispreferred action in mid course. Through this co-optation, Lerner demonstrates, the second speaker can strategically convert an imminent dispreferred action into a collaboratively achieved, preferred action in the same domain of activity (e.g., from disagreement
Activity, participation, and joint turn construction
to agreement, from other-correction to self-correction, etc.). A similar practice is observed in Japanese conversation as well. The following fragment shows an instance of the conversion of other-correction into self-correction. Schegloff, Jefferson and Sacks (1977) demonstrate a preference for self-correction over other-correction in the organization of repair in English conversation. Though there is no systematic study that has explicated the preference organization in the operation of repair in Japanese, some preliminary inspection of data from Japanese conversation suggests that Schegloff et al.’s (1977) description of the preference for self-correction over other-correction applies to Japanese as well. In fragment (9), an emerging utterance that is recognizable as an imminent other-correction (i.e., a dispreferred action) is converted into a self-correction (i.e., a preferred action) through the practice of joint turn construction. At the beginning of this fragment, the participants are discussing the origins of place names in the Plains in the United States. In lines 1–3, Kooji asks Hideo, who has lived in the United States, for confirmation of his understanding that the word ‘Oklahoma’ came from a Native American language. Instead of directly confirming Kooji’s understanding, Hideo suggests that many place names in the Plains also came from Native American languages (lines 4, 6–7, and 9). He then goes on to state that the Plains is an area which many Native Americans have inhabited for a long time (line 9, 11, and 13). While he was moving on to mention something about Nebraska (line 13), Kooji comes in (line 14) and produces an understanding check as to whether the situation is much like the situation in Hokkaido in Japan, which has been inhabited for a long time by an indigenous people called Ainu and which has many place names that came from words in the Ainu language. Examine the following fragment and see how the interaction transpires after Kooji’s understanding check in line 14. (9) [FM 23] 1 Kooji:
nan okurahoma tte: (3.0) nanka are ja nakatta what Oklahoma qt like that cp was.not “Doesn’t Oklahoma mean, like, that...”
2
uhm Indian lk what-you-call-it like
imi
meaning lk “... uhm Isn’t ((it)) a word that means, like, Indian’s ...”
kono INdian no nanchara
mitaina
no
3 (0.5) kotoba chig(h)atta kke? word was.not Q “...what-you-call-it?”
Makoto Hayashi
4 Hideo: kororado mo: maa: Colorado also well “Colorado is also, well,” 5 Kooji:
aa soo na n?
oh so cp N “Oh is ((that)) right?”
6 Hideo: n chuu ka m- mma ne- ano::: zenbu:: (0.5) qt:say Q well uhm all “((I)) mean, well, uhm, all:: (0.5)” 7
zenbu tte koto nai kedo kekkoo=
all qt thing not but “not all, but, pretty much...”
pretty.much
soo na n ka= 8 Kooji: =kono hen wa kekkoo this area tp pretty.much so cp N Q “Is ((it)) that ((the situation)) is pretty much like that around this area [i.e., the Plains]?”
9 Hideo: =ooi yo. sono:: (0.6) motomoto: many fp uhm originally “Quite a bit. Uh::m, (0.6) ((It))’s originally...” 10 Kooji:
oo hoo hoo.
“Oh, uh huh, uh huh.”
11 Hideo: indian no: Indian lk “... a place where Indians ...” 12 Kooji:
[un.] “Uh huh”
[
]
13 Hideo: [oru] toko yashi neburasuka mo nanka exist place cp:and Nebraska also like sono:= uhm “... live, and Nebraska is also, like, uhm ...”
14 Kooji:
=hokkaidoo mitaina mon ka. Hokkaido like thing Q “Is ((it that the situation is)) something like Hokkaido?”
15
(1.7)
Activity, participation, and joint turn construction
16→Hideo: nmaa sono hokkaidoo no:= well uhm Hokkaido lk “Well, uhm, Hokkaido’s...” 17→Kooji:
=chimee na:.= place.name fp “place names, right?”
18 Hideo: =soo [soo soo soo.] “Right, right, right, right.”
[
19 Kooji:
[(
] )] toka: etc. )”
“Like (
20 Hideo: nn soo soo [soo soo.] “Yeah, right right right right.”
[
]
21 Kooji: [nantoka ] betsu toka. etc. “Like, such-and-such betsu.” Notice that Hideo does not immediately respond to Kooji’s understanding check in line 14 as to whether the situation in the Plains in the United States is much like the situation in Hokkaido in Japan. In fact, there is quite a long silence in line 15 following Kooji’s understanding check. As shown by the work on the preference organization in English and Japanese (Pomerantz 1984a; Mori 1999), such a delay in response that has been made relevant by the prior utterance strongly projects an upcoming dispreferred action. In addition, when Hideo initiates an utterance in line 16, he prefaces it with tokens like nmaa (‘well’) and sono (‘uhm’), which are commonly observed in utterances embodying dispreferred actions. Given these ‘harbingers’ of a dispreferred action, Hideo appears to be initiating some action other than a direct confirmation of Kooji’s understanding proffered in line 14. I argue that his utterance hokkaidoo no: (the word for Hokkaido followed by the ‘linker’ no, which, just like the English genitive -’s, projects another noun to be produced after it) can be heard as an initiation of other-correction seeking to specify what Kooji means by hokkaidoo mitaina mon (‘situation like Hokkaido’) in his understanding check. I say this because Kooji’s utterance in line 14 is placed after Hideo’s statement about the population of Native Americans in the Plains (lines 9, 11, and 13), and this sequential location makes it equivocal whether the ‘situation like Hokkaido’ refers to the population of the native people (i.e., Ainu) in Hokkaido, or the place names in Hokkaido that came from the Ainu language (or both).
Makoto Hayashi
It is after the stressed and stretched no in hokkaidoo no:, then, that Kooji makes a mid-turn entry and supplies the noun chimee (‘place name’), which completes the projected structure of the form [Noun no + Noun], i.e., [hokkaidoo no + chimee] (‘Hokkaido’s + place names’). By co-opting the completion of Hideo’s utterancein-progress and supplying the word that specifies what is referred to by hokkaidoo mitaina mon (‘situation like Hokkaido’) himself, Kooji not only preempts the emerging other-correction being produced, but also converts it into a self-correction of his own prior utterance in line 14. Note also that Kooji presents the completion with the utterance-final particle na, which is often used to seek confirmation/ acknowledgment from the recipient. In the following turn, Hideo provides confirmation tokens (soo soo soo soo ‘right right right right’). Through this interactive process, the conversion of an imminent dispreferred action (other-correction) into a preferred action (self-correction) is collaboratively achieved. In the next fragment, in the face of the first speaker’s utterance-in-progress that projects an incipient disagreement, the second speaker employs joint turn construction and anticipatorily voices the projected disagreement himself. By doing so, he provides the first speaker with a next-turn slot for an agreeing response to the proffered completion, so as to convert an imminent disagreement to a collaboratively achieved agreement. Unlike the previous fragment, however, the first speaker in this fragment disregards the second speaker’s attempt to preempt a disagreement by using the practice of ‘delayed completion’ (Lerner 1989) and completes his disagreeing utterance himself. As a result, a preferred action, i.e., agreement, is not collaboratively achieved. The participants shown in this fragment are two male graduate students in economics. In this segment, they are discussing their research methods, particularly the use of statistics. The focus of their discussion is on how many samples they would need to obtain reliable results for their research. Examine how the interaction transpires, with a special focus on Akira’s joint turn construction in line 9 and Seiji’s delayed completion in lines 10–11. (10) [RKK 19] 1 Akira:
=s:soko de::: (0.7) kyuujuu shiraberu there 90 examine
hitsuyoo nai yo na: 2 necessity not.exist fp fp “There’s no need to examine 90 out of them ((=100)), right?”
3
(2.5)
4 Akira:
“sa-”
sa-
Activity, participation, and joint turn construction
5
(0.7)
6 Seiji:
well but not.know fp “We::::::::::::ll .hh but ((you)) never know”
7 Akira:
=[bi- ] “bi-”
u:::::::::::::n .hh demo wakannai yo=
[
]
8 Seiji: =[daka]ra:: i- (2.0) doo yuu:: (0.7) so what.kind.of “So:: i- (2.0) what kind of (0.7)” 9→Akira:
test oh result sp come.about Q “test, oh I mean, result will come about.”
10→Seiji:
↑kentee o:: suru ka tte yuu what.kind.of test O do Q qt say
11→
koto ni yoru yo na::= thing on depend fp fp “((It)) will depend on what kind of test ((you)) do.”
12 Akira:
=u::n. “Yeah.”
kentee a:: kekka ga deru
ka.
doo yuu
In lines 1–2, Akira produces an agreement-seeking utterance about how many samples they would need to obtain reliable results, thus making it relevant for Seiji to offer a response in the next turn. However, Seiji delays his response considerably, first by not initiating talk immediately (lines 3–5), and then by producing a prolonged preface, u:::::::::::::n (‘we:::::::::::::ll’). He then produces a disagreementimplicating connective, demo (‘but’), followed by a qualifier, wakannai yo (‘((you)) never know’), which constitutes a mild disagreement. Akira appears to have heard wakannai yo as a complete response from Seiji to his earlier agreement-seeking utterance and thus starts to talk on completion of wakannai yo (line 7). On the other hand, Seiji continues on to produce a further utterance (line 8), projecting, with dakara (‘so’), some elaboration on his previous remark wakannai yo. Akira withdraws from the simultaneous talk immediately, thereby letting Seiji continue. However, when Seiji pauses after producing doo yuu (‘what kind of ’), Akira comes in and provides an utterance that completes Seiji’s utterance-in-progress. In his completing utterance, Akira voices the projected elaboration of Seiji’s disagreement on his behalf, thereby attempting to preempt the disagreement import of the utterance. What Akira constructs using Seiji’s doo yuu (‘what kind of ’) here is the
Makoto Hayashi
clause doo yuu kekka ga deru ka (‘what kind of result will come about’),9 which is designed to be grammatically connected to wakannai yo (‘((you)) never know’) in Seiji’s earlier utterance. To be more specific, the co-constructed clause is built as a postposed embedded clause for wakannai yo, which serves as the matrix clause, as shown below:
Matrix clause [wakannai yo] + (‘((you)) never know’)
Embedded clause [doo yuu kekka ga deru ka] (‘what kind of result will come about’)
We call this embedded clause ‘postposed,’ because, unlike in English, the canonical clause order in a complex sentence in Japanese is [embedded clause] + [matrix clause], as in doo yuu kekka ga deru ka + wakannai yo (‘((You)) never know what kind of result will come about’). Thus, by co-constructing a postposed embedded clause for Seiji’s earlier utterance and providing him with a next-turn slot for an agreeing response to the completing utterance, Akira preempts the emerging disagreement and attempts a collaborative achievement of agreement. However, unlike the first speaker in fragment (9), Seiji does not produce an agreeing/confirmatory response after the completing utterance proffered by Akira. Instead, he employs ‘delayed completion’ (Lerner; 1989), i.e., a practice that speakers use to finish a discontinued utterance after an intervening utterance by another participant. Lerner (1989) notes that delayed completion is used as a device to disregard another participant’s intervening utterance and cancel its sequential import. In the present case, while Akira’s joint turn construction in line 9 makes relevant Seiji’s next-turn receipt – and next-turn agreement, in particular – for a collaborative achievement of the conversion of disagreement into agreement, Seiji’s delayed completion cancels the sequential relevance of his receipt and reinstitutes a disagreement import of his earlier utterance. And by doing so, he turns the table, as it were, and provides Akira with a next-turn receipt slot (line 12). Note also that Seiji engages in skillful ‘grammatical work’ to accomplish the interactional work of disregarding Akira’s joint turn construction. As described above, Akira’s joint turn construction contextualizes Seiji’s doo yuu in line 8 as the beginning of the postposed embedded clause for wakannai yo in line 6. In his delayed completion in lines 10–11, however, Seiji recontextualizes doo yuu as the beginning of a new grammatical unit that is independent of wakannai yo, thereby simultaneously recontextualizing wakannai yo as being already complete in line 6, as seen in the following: Line 6
demo wakannai yo (‘But ((you)) never know.’)
9. Kentee (‘test’) at the beginning of Akira’s utterance in line 9 is replaced by kekka (‘result’).
Activity, participation, and joint turn construction
Lines 10–11
doo yuu ↑kentee o:: suru ka tte yuu koto ni yoru yo na:: (‘((It)) will depend on what kind of test ((you)) do.’)
Thus, Seiji accomplishes the disregarding of Akira’s joint turn construction not simply by delivering another rendition of completion, but by proposing an alternative grammatical structure for the resulting utterance.10 In this instance, then, we observed that, while the second speaker uses joint turn construction to attempt a conversion of an emerging disagreement into a collaboratively achieved agreement, the first speaker employs delayed completion to decline such collaboration and reinstitute the disagreement import of his original utterance. In this subsection, I examined instances in which, in the face of an emerging dispreferred action, the recipient employs joint turn construction to accomplish a distinct type of participation, i.e., converting an imminent dispreferred action into a preferred action in the same domain of activity. It was noted that such a conversion is achieved not unilaterally but collaboratively; thus, when the first speaker does not collaborate, as in fragment (10), an attempt to convert a dispreferred action into a preferred action through joint turn construction can fail. 4. Concluding remarks The goal of this chapter has been to explore the reflexive relationship between grammar and social action. Taking as a point of departure the perspective that language is essentially a vehicle for participation in everyday communicative activities, I examined cases of joint turn construction observed in Japanese conversations in order to investigate how grammar and action mutually shape one other in situated language use in real-life settings. I showed, on the one hand, how grammar shapes relevant ways in which participants ‘package’ their action. As a speaker’s utterance unfolds, its emerging grammatical structure allows co-participants to recognize what kind of action the speaker is executing, how it is going to develop, and what range of responsive actions will be relevant next. Co-participants utilize this ‘projection’ of the unfolding course of the speaker’s action as a resource to shape their own grammatical conduct (i.e., producing an utterance that is fitted to the grammatical trajectory of the current speaker’s ongoing utterance) in order to accomplish activity-relevant participation. The foregoing analysis also demonstrated, on the other hand, how grammar emerges as the outcome of action and 10. It is also of interest to note that Seiji incorporates into his delayed completion the word kentee (‘test’), which was discarded by Akira in his co-participant completion in favor of kekka (‘result’). Thus, Seiji’s delayed completion makes visible his rejection of Akira’s co-participant completion not only at the grammatical level, but also at the lexical/semantic level.
Makoto Hayashi
participation. We observed some specific ways in which the speaker’s finely-differentiated participatory stances toward the ongoing activity shapes the details of the grammatical structuring of the jointly produced utterance. Thus, the examination of joint turn construction presented in this chapter provided a conspicuous documentation of how grammar is a resource for, and an outcome of, an interactional achievement of concerted action by multiple participants. Or, we might go one step further and argue that “[g]rammar is not only a resource for interaction and not only an outcome of interaction, it is part of the essence of interaction itself ” (Schegloff, Ochs & Thompson 1996: 38). I hope that the present chapter has provided some testimony for this statement. On a final note, I would like to address the issue of what kind of ‘grammar’ we are talking about here. Some readers might object to my statement in the preceding paragraph by saying that it only applies to grammar as instantiated in particular utterance forms, but not to grammar as generalized knowledge underlying diverse linguistic activities across situations. I acknowledge that grammar involves situation-transcending aspects of linguistic routines that exist as tacit knowledge allowing speakers and hearers to engage in communicative activities across many different contexts (Linell 2005) and I concede that the present study does not (directly) deal with that aspect of grammar. I would argue, however, that such an objection is based on the view that grammar as generalized knowledge is clearly delineated and completely severed from grammar as instantiated in situated local contexts. The point being made here is that situation-transcending, generalized knowledge can only come about as a result of accumulation and sedimentation of our everyday experience of language use in situated communicative activities. If this is true, then the only way we can gain analytic access to how grammar as situation-transcending knowledge is configured and reconfigured is to pay serious attention to grammar as instantiated on situated local occasions of talk-in-interaction. 5. Appendix 5.1
Transcript symbols
[ ] (0.0) (.) underlining CAPS ::
The point where overlapping talk starts The point where overlapping talk ends length of silence in tenths of a second micro-pause relatively high pitch relatively high volume lengthened syllable
Activity, participation, and joint turn construction
- = ?/./, ! ( ) (word) (( )) hh .hh (hh) > < ° ° bolding 5.2 CP EMP FP LK MIM N 5.3
glottal stop self-editing marker “latched” utterances rising/falling/continuing intonation respectively animated tone, not necessarily an exclamation unintelligible stretch transcriber’s unsure hearings transcriber’s descriptions of events -- e.g., ((sniff)) audible outbreath audible inbreath laughter within a word increase in tempo, as in a rush-through a passage of talk quieter than the surrounding talk section of note, highlighted by transcriber
Abbreviations used in the interlinear gloss various forms of copula verb be emphasis marker final particle linking particle mimetics nominalizer
O Q QT SP TL TP
object particle question particle quotative particle subject particle title marker topic particle
Double parentheses in the translation lines
Elements in double parentheses in the translation lines indicate those elements that are not expressed in the Japanese original but are supplied by the author for the reader’s ease of understanding. References Antaki, Charles, Díaz, Félix & Collins, Alan F. 1996. Keeping your footing: Conversational completion in three-part sequences. Journal of Pragmatics 25: 151–171. DOI: 10.1016/03782166(94)00081-6 Atkinson, J. Maxwell & Drew, Paul. 1979. Order in Court. London: Macmillan. Austin, John L. 1962. How to Do Things with Words. Oxford: OUP. Bourdieu, Pierre. 1977. Outline of a Theory of Practice. Cambridge: CUP. DOI: 10.1017/ CBO9780511812507 Cook, Haruko Minegishi. 1990. An indexical account of the Japanese sentence-final particle no. Discourse Processes 13: 410–439. DOI: 10.1080/01638539009544768
Makoto Hayashi Cook, Haruko Minegishi. 1992. Meanings of non-referential indexes: A case study of the Japanese sentence-final particle ne. Text 12: 587–539. DOI: 10.1515/text.1.1992.12.4.507 Davidson, Judy. 1984. Subsequent versions of invitations, offers, requests, and proposals dealing with potential or actual rejection. In Structures of Social Action: Studies in Conversation Analysis, J. Maxwell Atkinson & John Heritage (eds), 102–128. Cambridge: CUP. Garfinkel, Harold. 1967. Studies in Ethnomethodology. Englewood Cliffs NJ: Prentice-Hall. Giddens, Anthony. 1979. Central Problems in Social Theory: Action, Structure, and Contradiction in Social Analysis. Berkeley CA: University of California Press. Giddens, Anthony. 1984. The Constitution of Society: Outline of the Theory of Structuration. Berkeley CA: University of California Press. Goffman, Erving. 1981a. Forms of Talk. Philadelphia PA: University of Pennsylvania Press. Goffman, Erving. 1981b. Footing. In Forms of Talk, Erving Goffman, 124–159. Philadelphia PA: University of Pennsylvania Press. Goodwin, Charles. 1981. Conversational Organization: Interaction between Speakers and Hearers. New York NY: Academic Press. Goodwin, Charles. 1984. Notes on story structure and the organization of participation. In Structures of Social Action: Studies in Conversation Analysis, J. Maxwell Atkinson & John Heritage (eds), 225–246. Cambridge: CUP. Goodwin, Charles & Goodwin, Marjorie Harness. 1987. Concurrent operations on talk: Notes on the interactive organization of assessments. IPrA Papers in Pragmatics 1: 1–54. Goodwin, Charles & Goodwin, Marjorie Harness. 1992a. Context, activity and participation. In The Contextualization of Language [Pragmatics & Beyond New Series 22], Peter Auer & Aldo Di Luzio (eds.), 77–99. Amsterdam: John Benjamins. Goodwin, Charles & Goodwin, Marjorie Harness. 1992b. Assessments and the construction of context. In Rethinking Context: Language as an Interactive Phenomenon, Alessandro Duranti & Charles Goodwin (eds), 151–189. Cambridge: CUP. Goodwin, Marjorie Harness. 1990. He-Said-She-Said: Talk as Social Organization among Black Children. Bloomington IN: Indiana University Press. Gumperz, John J. 1982. Discourse Strategies. Cambridge: CUP. DOI: 10.1017/ CBO9780511611834 Haga, Yasushi. 1954. ‘Chinjutsu’ to wa nanimono? (What is ‘chinjutsu’?) Kokugo-kokubun 23(4): 241–255. Hayashi, Makoto. 2003. Joint Utterance Construction in Japanese Conversation [Studies in Discourse and Grammar 12]. Amsterdam: John Benjamins. DOI: 10.1075/sidag.12 Hayashi, Makoto & Mori, Junko. 1998. Co-construction in Japanese revisited: We do ‘finish each other’s sentences.’ In Japanese/Korean Linguistics, Vol. 7, Noriko Akatsuka, Hajime Hoji, Shoichi Iwasaki & Susan Strauss (eds), 77–93. Stanford CA: CSLI. Heritage, John. 1984. Garfinkel and Ethnomethodology. Cambridge: Polity. Hopper, Paul J. 1988. Emergent grammar and the a priori grammar postulate. In Linguistics in Context: Connecting Observations and Understanding, Deborah Tannen (ed.), 117–134. Norwood NJ: Ablex. Hopper, Paul J. 1987. Emergent grammar. Berkeley Linguistics Society 13: 139–157. Hopper, Paul J. 1998. Emergent grammar. In The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, Michael Tomasello (ed.), 155–175. Mahwah NJ: Lawrence Erlbaum Associates.
Activity, participation, and joint turn construction Hymes, Dell. 1972. Models of the interaction of language and social life. In Directions in Sociolinguistics: The Ethnography of Communication, John J. Gumperz & Dell Hymes (eds), 35– 71. New York NY: Holt, Reinhart & Winston. Iwasaki, Shoichi. 1993a. Subjectivity in Grammar and Discourse: Theoretical Considerations and a Case Study of Japanese Spoken Discourse [Studies in Discourse and Grammar 2]. Amsterdam: John Benjamins. Iwasaki, Shoichi. 1993b. The structure of the intonation unit in Japanese. In Japanese/Korean Linguistics, Vol. 3, Soonja Choi (ed.), 39–53. Stanford CA: CSLI. Jefferson, Gail. 1987. On exposed and embedded correction in conversation. In Talk and Social Organisation, Graham Button & John R. E. Lee, 86–99. Clevedon: Multilingual Matters. Kamio, Akio. 1994. The theory of territory of information: The case of Japanese. Journal of Pragmatics 21: 67–100. DOI: 10.1016/0378-2166(94)90047-7 Kamio, Akio. 1997. Territory of Information [Pragmatics & Beyond New Series 48]. Amsterdam: John Benjamins. DOI: 10.1075/pbns.48 Labov, William & Fanshell, David. 1977. Therapeutic Discourse: Psychotherapy as Conversation. New York NY: Academic Press. Lave, Jean & Wenger, Etienne. 1991. Situated Learning: Legitimate Peripheral Participation. Cambridge: CUP. DOI: 10.1017/CBO9780511815355 Lerner, Gene H. 1987. Collaborative Turn Sequences: Sentence Construction and Social Action. PhD dissertation, University of California, Irvine. Lerner, Gene H. 1989. Notes on overlap management in conversation: The case of delayed completion. Western Journal of Speech Communication 53: 167–177. DOI: 10.1080/105703 18909374298 Lerner, Gene H. 1991. On the syntax of sentences in progress. Language In Society 20, 441–458. Lerner, Gene H. 1992. Assisted storytelling: Deploying shared knowledge as a practical matter. Qualitative Sociology 15: 247–271. DOI: 10.1007/BF00990328 Lerner, Gene H. 1996. Finding ‘face’ in the preference structures of talk-in-interaction. Social Psychology Quarterly 59: 303–321. DOI: 10.2307/2787073 Lerner, Gene H. & Takagi, Tomoyo. 1999. On the place of linguistic resources in the organization of talk-in-interaction: A co-investigation of English and Japanese grammatical practices. Journal of Pragmatics 31: 49–75. DOI: 10.1016/S0378-2166(98)00051-4 Levinson, Stephen C. 1979. Activity types and language. Linguistics 17: 365–399. DOI: 10.1515/ ling.1979.17.5-6.365 Linell, Per. 1982. The Written Language Bias in Linguistics. Linköping: University of Linköping, Department of Communication. Linell, Per. 2005. The Written Language Bias in Linguistics: Its Nature, Origins, and Transformations. London: Routledge. DOI: 10.4324/9780203342763 Malinowski, Bronislaw. [1935]1978. Coral Gardens and Their Magic, 2 Vols. London: Allan & Urwin. Maynard, Senko K. 1993. Discourse Modality: Subjectivity, Emotion and Voice in the Japanese Language [Pragmatics & Beyond New Series 24]. Amsterdam: John Benjamins. DOI: 10.1075/pbns.24 Mori, Junko. 1999. Negotiating Agreement and Disagreement in Japanese: Connective Expressions and Turn Construction [Studies in Discourse and Grammar 8]. Amsterdam: John Benjamins. DOI: 10.1075/sidag.8
Makoto Hayashi Morita, Emi. 2002. Stance marking in the collaborative completion of sentences: Final particles as epistemic markers in Japanese. In Japanese/Korean Linguistics,Vol.10, Noriko Akatsuka, Susan Strauss & Bernard Comrie. Stanford CA: CSLI. Phillips, Susan U. 1972. Participant structures and communicative competence: Warm Springs children in community and classroom. In Functions of Language in the Classroom, Courtney B. Cazden, Vera P. John and Dell Hymes (eds), 370–394. New York NY: Columbia Teachers Press. Pomerantz, Anita. 1978. Compliment responses: Notes on the co-operation of multiple constraints. In Studies in the Organization of Conversational Interaction, Jim Schenkein (ed.), 79–112. New York NY: Academic Press. Pomerantz, Anita. 1980. Telling my side: ‘Limited access’ as a ‘fishing’ device. Sociological Inquiry 50: 186–198. DOI: 10.1111/j.1475-682X.1980.tb00020.x Pomerantz, Anita. 1984a. Agreeing and disagreeing with assessments: Some features of preferred/dispreferred turn shapes. In Structures of Social Action: Studies in Conversation Analysis, J. Maxell Atkinson & John Heritage (eds), 57–101. Cambridge: CUP. Pomerantz, Anita. 1984b. Pursuing a response. In Structures of Social Action: Studies in Conversation Analysis, J. Maxwell Atkinson & John Heritage (eds), 152–163. Cambridge: CUP. Sacks, Harvey. 1974. An analysis of the course of a joke’s telling in conversation. In Explorations in the Ethnography of Speaking, Richard Bauman & Joel Sherzer (eds), 337–353. Cambridge: CUP. Sacks, Harvey. [1973]1987. On the preferences for agreement and contiguity in sequences in conversation. In Talk and Social Organisation, Graham Button & John R.E. Lee, 54–69. Clevedon: Multilingual Matters. Sacks, Harvey. 1992. Lectures on Conversation, 2 Vols., Gail Jefferson (ed.), with introductions by Emanuel A. Schegloff. Oxford: Blackwell. Schelgoff, Emanuel A., Ochs, Elinor & Thompson, Sandra A. 1996. Introduction. In Interaction and Grammar, Elinor Ochs, Emanuel A. Schegloff & Sandra A. Thompson (eds.), 1–51. Cambridge: CUP. Schegloff, Emanuel A., Jefferson, Gail & Sacks, Harvey. 1977. The preference for self-correction in the organization of repair in conversation. Language 53: 361–382. Schutz, Alfred. 1962. Collected Papers, Vol. 1. The Hague: Martinus Nijhoff. Searle, John. 1969. Speech Acts: An Essay in the Philosophy of Language. Cambridge: CUP. DOI: 10.1017/CBO9781139173438 Tanaka, Hiroko. 1999. Turn-taking in Japanese Conversation: A Study in Grammar and Interaction. Amsterdam/Philadelphia: John Benjamins. Uyeno, Tazuko. 1971. A Study of Japanese Modality: A Performative Analysis of Sentence Particles. PhD dissertation. University of Michigan. Vygotsky, Lev S. 1978. Mind in Society: The Development of Higher Psychological Processes. Cambridge MA: Harvard University Press. Watanabe, Minoru. 1953. Jojutsu to chinjutsu - jutsugo-bunsetsu no koozoo (Describing and stating: The structure of predicative clauses). Kokugogaku 13–14. Wittgenstein, Ludwig. 1953. Philosophical Investigations. Oxford: Blackwell. Zimmerman, Don H. & Boden, Deirdre. 1991. Structure-in-action: An introduction. In Talk and Social Structure: Studies in Ethnomethodology and Conversation Analysis, Deirdre Boden & Don Zimmerman (eds), 3–20. Berkeley CA: University of California Press.
part 3
Language change and variation
Context in constructions Variation in Japanese non-subject honorifics* Yoshiko Matsumoto This paper illustrates the importance of context for grammar by examining discourse data of the nonsubject honorific construction o – Verb (stem) – suru. Building upon previous studies that argued that non-subject honorifics are subject to a pragmatic condition of benefit transfer between the subject and nonsubject referents, I argue that the targeted referents of nonsubject honorification are reorganized to the two participants of the discourse: the speaker and the addressee. The variations in the o-Verb (stem) – suru form are explained from the constructional and frame semantics perspectives as a process of cognitive and intersubjective (e.g., Traugott and Dasher 2002) reorganization of grammatical constructions motivated by contextual conditions of use and the speaker’s intention with respect to the addressee.
1. Introduction The integration of context in grammatical description is crucial to linguistic theories that consider language as part of human behavior. However, the same linguistic phenomena may be viewed quite differently depending on the goals and interests of researchers, and what is considered crucial support for a certain * I am grateful to those who gave me comments at various occasions when parts or a version of this paper was presented, especially the participants of the Symposium on Functional Approaches to Japanese Grammar: Toward the Understanding of Human Languages, at the University of Alberta in 2004 where implications of the symposium papers on the claims made by Newmeyer (2003) were discussed. Thanks are also due to the audience at the International Cognitive Linguistics Conference (2001), where an earlier analysis of the uses of the o-V-suru construction, later published in Matsumoto (2008), was first presented. I am also indebted to Elizabeth Traugott for her invaluable discussion. Matsumoto (2008) focuses on cognitive reorganization from the point of view of intersubjectivity, which can be described within the framework of Construction Grammar. The basic analysis of the construction is presented here as an example in support of the theoretical approach of the present volume. My sincere gratitude goes to Kaori Kabata and Tsuyoshi Ono, the organizers of the symposium and the editors of this volume.
Yoshiko Matsumoto
theoretical view may not be considered important for others. For example, Newmeyer’s paper “Grammar is grammar and usage is usage”(2003) is illustrative of this theoretical variability by bringing our attention back to a longstanding split among linguists on the question regarding whether or not the speaker’s knowledge of grammar is fundamentally separate from one’s knowledge of the real world and of uses of linguistic constructions in context. While this position is not shared in all strands of formal or generative theories, research of the same theoretical persuasion as Newmeyer’s would maintain the autonomy of syntax or grammar, in that the grammar is never fundamentally affected by the contexts in which the language is used. In such a view pragmatic and real-world knowledge of the interlocutors would not be central to grammar even where such knowledge is recognized to be crucial to the construal of linguistic constructions, to the process of grammaticalization and intersubjectivification as well as to constructions that are sensitive to information status. As the examination of various constructions such as in Laury and Ono (2005) shows, the dismissive position regarding usage and context would exclude from grammar phenomena that have traditionally been areas of substantial interest within linguistics. Constructions in Japanese in which pragmatic and world knowledge play a crucial role in grammatical description are well exemplified by those that have spatial, temporal or social deictic expressions (including honorifics), constructions expressing sensations, and also by such seemingly less subjective constructions as the noun-modifying construction. Japanese noun-modifying constructions, i.e., complex noun phrases including what are traditionally categorized as relative clause constructions, present a good example of such phenomena (Matsumoto 1997a, inter alia). In these constructions, there is no explicit marking of the grammatical relationship between the head noun and the predicate, yet semantic and pragmatic compatibility of the constituents determine the meaning and the acceptability of the specific construction. Thus, for example, the noun-modifying construction [[toire ni ikenai]komaasyaru] ‘[[bathroom GOAL go.cannot] commercial]; commercials (because of which) (x) cannot go to the bathroom’ in the attested utterance konogoro [[toire ni ikenai]komaasyaru] ga ookute komaru “(I am) in trouble because there are many commercials (because of wanting to watch which) (I) cannot go to the bathroom” is a grammatical Japanese phrase that is readily construable to competent Japanese speakers who are cognizant of life in Japan. There are various factors that support this construability. There is shared real-world knowledge of what people usually do during TV commercials, e.g. going to the bathroom. There is semantic and pragmatic knowledge that the negative state of affairs expressed in the subordinate clause (i.e. the inability to go to the bathroom) would be compatible with an idea that there was a condition that caused such state. One can also conjecture that the increased number of interesting
Context in constructions
commercials could cause inconvenience to the TV viewers who use commercial breaks as bathroom breaks. The construal of this noun phrase is not based on the syntactically most likely analysis, in which the head noun, “commercial”, is understood to be the subject referent of the one-place predicate, i.e. what cannot go to the bathroom. Unlike Chomsky’s well-known semantically anomalous but syntactically well-formed sentence, Colorless green idea sleeps furiously, the syntax of Japanese noun-modifying constructions does not force a reading which may then be judged as semantically or pragmatically anomalous reading unless its usage is contextually supported. In other words, unlike what is arguably the case in English, pragmatics, just as much as syntax, is at the basis of Japanese grammar, and a purely syntactic analysis devoid of pragmatic and contextual considerations does not give an accurate acceptability judgment of such constructions. The Japanese non-subject honorific construction in the form o-Verb(stem)suru, presents another instance in which context plays a crucial role in providing a systematic understanding of its structure, as well as of the variation and change in use. If the notion of the grammar of a language is to be confined exclusively to the abstraction of general ways in which words are strung together, conditions that motivate innovative (i.e., normatively unacceptable) usages may well be outside of the scope of investigation, as suggested by Newmeyer that “grammars are not fragile, fluid and temporary” as he suspects that “we could carry on a conversation with Shakespeare, who lived four hundred years ago. And the problems we would have with him would more likely be lexical and low-level phonological rather than syntactic” (Newmeyer 2003: 698). The changes that I will examine here would involve ungrammaticality based not on the wrong “shape” of words and sentences nor on any anomaly at the lexical or phonological level, but based on choices of subject and nonsubject referents. The phenomenon might be considered as a collection of performance errors, but as I will discuss below, they can be explained systematically if we take the context into serious consideration. It may not be completely clear, however, what contextual factors are relevant and how they can be represented in grammatical constructions. In this paper, I will focus on uses of the Japanese non-subject honorific construction in the form o-Verb(stem)-suru to illustrate that the two types of “frames” proposed by Fillmore in his theory of Frame Semantics (e.g., 1975, 1982) are crucial in incorporating relevant contextual information into a grammatical construction. The two “frames” are: (1) what may be called a “cognitive frame”, which is evoked by lexical meanings and which contains cognitively profiled roles/elements; for example, a larger context evoked by verbs such as buy, sell, and cost, i.e., the “commercial event” frame, which includes roles such as seller, buyer, goods, money, and (2) an “interactional frame”, which represents the conceptualization of the discourse situation between the speaker and the addressee: from knowledge of deictic
Yoshiko Matsumoto
categories to knowledge of discourse genres, such as a folktale, and an obituary (see also Tannen 1979; Matsumoto 1997b). Fillmore’s notion of frame has various aspects and has been further developed, but the following quote provides a general sense of the concept. By the term “frame” I have in mind any system of concepts related in such a way that to understand any one of them you have to understand the whole structure in which it fits; when one of the things in such a structure is introduced into a text, or into a conversation, all of the other are automatically made available. I intend the word “frame” as used here to be a general cover term for the set of concepts variously known, in the literature on natural language understanding, as “schema”, “script”, “scenario”, “ideational scaffolding”, “cognitive model”, or “folk theory”. (1982: 111)
In this approach to semantics, the first type of frame, the “cognitive frame”, has received more focus in research, but I will show that the two types of frames are both crucial part of grammar. In the construction of the honorific prefix o, a verb stem and the light verb suru, meaning ‘do’, the verb form conventionally indicates that a non-subject referent, such as the direct object referent, is “honorified”, or is the target of the speaker’s respect, while the subject referent is abased. Recently, however, numerous uses of the o-Verb-suru form that deviate from this conventional description have been attested. In those uses, the target of respect is not denoted by a non-subject and the referent of the subject is not abased. The target of respect is usually the addressee, who may be referred to by the subject. For such uses, it is not at all accurate to refer to the o-Verb(stem)-suru form as a “non-subject” honorific. Crucial to discovering a system in these seemingly deviant variations is to incorporate the context into the analytical framework of the construction. I argue that when we incorporate speech contexts into our analysis, the variations in the o-Verb (stem) – suru form can be systematically explained from the constructional and frame semantics perspectives, rather than being merely exceptional or irregular instances of the construction. In a previous study of non-subject honorifics (Matsumoto 1997b), I showed that such forms are normally subject to a pragmatic condition of benefit transfer between the two main participants of the event, namely, the subject and nonsubject referents expressed or implied in the propositional content of the construction. I have also noted that speakers tend to present their actions as benefiting the addressee (see also Tsujimura 1992). Building upon these findings, I have described how the targeted referents of nonsubject honorification are reorganized from the two relevant participants in the proposition to the two participants of the discourse; namely, the speaker and the addressee (Matsumoto 2008). This suggests that the nonsubject honorific construction in Japanese is undergoing a systematic (and to a certain extent predictable) change and is acquiring an addressee honorific use in addition to being a propositional referent honorific.
Context in constructions
In terms of frames, the combination of “cognitive frames” and “interactional frames” plays a crucial role. Examining attested discourse instances of the o-Verb (stem) – suru form, I will illustrate that the recent variations are reflections of intersubjective (e.g., Traugott & Dasher 2002) reorganization of the cognitive frame of the grammatical construction motivated by elements of the “interactional frame”, i.e., speech context and the speaker’s intention with respect to the addressee. The advantage of using both types of frames to describe the o-Verb-suru construction is the capability of illustrating the relation between the contextual participants and the event participants, and how these two interact. Notions available in Frame Semantics and Construction Grammar well accommodate a systematic account of contextually dependent constructions, such as honorifics, including their variations and change. In the following, I will discuss the uses of o-Verb(stem)-suru form of nonsubject honorifics more in detail, following the lines of arguments presented in Matsumoto (2008). 2. Background of non-subject honorifics Honorific forms of Japanese predicates, after Harada’s generative analysis (1975), have generally been classified into propositional honorifics and performative honorifics, otherwise known as referent honorifics and addressee honorifics, respectively.1 The use of referent honorifics is licensed by the speaker’s relation to the referents of the subject and/or non-subject of the sentence, or the relation between the referents of the subject and some non-subject of the sentence; while the use of addressee honorifics depends on the context of speech, notably on the speaker’s relation to the addressee. Referent honorifics are further subcategorized into subject honorifics and nonsubject honorifics depending on whether the triggering referent is expressed as the subject or a nonsubject of the sentence.2 These structural 1. There have also been numerous studies treating the system and the history of Japanese honorifics in the frameworks of traditional Japanese grammar and ethnomethodology (e.g., Yamada 1924; Tokieda 1941; Tsujimura 1967, 1968; Hayashi & Minami 1973–74; Kokuritsu Kokugo Kenkyujo (The National Language Research Institute) 1957, to name a few). 2. Harada classified honorific forms into subject honorifics (SH), object honorifics (OH) and performative honorifics (PH). I adopt the terminology “nonsubject honorifics”, following Kuno (1987), instead of Harada’s “object honorifics”, since the target NP of such construction need not be limited to a direct or indirect object. (For criticisms for Harada’s account, see also, e.g., Kikuchi 1980; Kuno 1983; Martin 1975; McCawley 1993.) Kuno, however, doesn’t explain which referent among the many possible semantic and grammatical roles can license a specific nonsubject honorific construction.
Yoshiko Matsumoto
classifications are not free of problems, but they highlight the two major types of honorific constructions (referent honorifics and addressee honorifics), which can be used independently or in combination, i.e. a predicate may exhibit referent honorifics only, addressee honorifics only, neither, or both. The focus of this paper is one of the referent honorific forms, the nonsubject honorific construction, which is in the general form o-Verb (stem)-suru (‘Honorific Prefix – Verb (stem) -do’). This is one of the forms for which variations and changes in use have been most commonly noted in newspaper articles and in popular books on the “correct” use of honorifics. In the conventional use of this construction, for example in (1), the teacher, the object referent of the honorific verb o-tasuke-sita ‘helped (NSH: Non-Subject Honorific), is elevated while the subject referent, Abe-san, is humbled. (2) is a similar non-honorific sentence, given as a contrast. In (2), neither the subject nor object referent is elevated or downgraded. (1) Abe-san ga sensei o o-tasuke-sita. Mr. Abe nom teacher acc hp-help-did ‘Mr. Abe helped (Non-Subject Honorific: NSH) the teacher. (2) Abe-san ga Oda-san o tasuketa. Mr. Abe nom Mr. Oda acc helped ‘Mr. Abe helped Mr. Oda. Examples that do not fit into the conventional uses are given in (3) and (4). (3) was announced by a pilot on an international flight (Matsumoto 1997b). (3)
ato go-zikan hodo de hizuke-henkoo-sen o more 5-hours about in international dateline acc o-mukae-itasi-masu. HP-meet-do.HUM-AH ‘(We) will meet (NSH) the international dateline in about 5 hours.’
The part in bold face, o-mukae-itasi- is an example of a nonsubject honorific form in which -itasi- is the stem of itasu, a humble form of the verb suru ‘do’, and the underlined masu is an addressee honorific. From the conventional description of nonsubject honorifics, the sentence would be anomalous because the object of mukae- ‘meet/reach’ -- i.e., the‘international dateline’ -- can hardly be the target of honorification. The only apparent potential target of honorific use in this discourse context is the addressees, i.e. the passengers on the airplane, (for whom the addressee honorific masu is used as shown in the example,) but they are not a participant in the described event. Another, perhaps more striking example is given in (4).
Context in constructions
(4)
[a delivery service clerk to a customer] ..soo site-itadaku to o-mati-suru o-zikan mo arimasen kara... so do-receive.hum if hp-wait-do hp-time also not.exist.ah so ‘...if you could do so, there would be no time for you (= you don’t need) to wait (NSH) (for us), so...’
This example was uttered by a delivery/pick-up service clerk on the phone to his customer who wanted her out-going package to be picked within a short window of time. The service clerk was recommending that the customer should pack the items using her own box and fill in forms before the clerk arrived, so that she would not have to wait for him to make a package. In this specific context, the utterance is interpreted in the way in which the subject referent of the o-Verb(stem)suru construction o-mati-suru is the customer, whereas the nonsubject referent is the speaker. This use is deviant from the normative perspective since the speaker, the referent of a nonsubject, is raised, while the subject referent, the customer, is downgraded. Variants such as (3) and (4), which are attested in spoken and written discourse, as we see in more examples below, suggest that the nonsubject honorific construction in Japanese is undergoing a change in usage and is gradually becoming an addressee/performative honorific. In a frame of a polite discourse, the targeted referents of the nonsubject honorific construction o-Verb(stem)-suru are re-organized from referents of the subject and the nonsubject in the described event to the two participants of the discourse, the speaker and the addressee. This type of mapping between the domain of the sentence describing an event and the speech context has also provided a systematic explanation to other linguistic phenomena such as “speech act causality” argued by Sweetser (1990), and speech act qualification in English and Japanese (Lakoff 1980; Matsumoto 1985, 2001), and indeed, a similar drift in usage from referent to addressee honorifics has been observed in other Japanese honorific constructions, such as -masu. (Dasher 1995; Traugott & Dasher 2002). In the following, I will illustrate three main uses of the o-Verb(stem)-suru construction and demonstrate the importance of context in grammar with reference to ideas in Frame Semantics and Construction Grammar. 3. O-Verb(stem)-suru construction: Use 1 – Nonsubject honorifics In a reexamination of the o-Verb(stem)-suru construction elsewhere (Matsumoto 1997b), I found several noteworthy properties. One is that, in the conventional use of the o-Verb(stem)-suru form, there is an implied relationship of benefit between the two (typically, human) participants, in which the exalted nonsubject referent is
Yoshiko Matsumoto
either the beneficiary or the source of benefit, depending on the predicate and on the context. The determination of which non-subject referent may license nonsubject honorifics depends, therefore, not only on the speaker’s decision to elevate one of the nonsubject referents but also on the consonance of the action or state that is described in the sentence with a pragmatic notion of benefactivity involving such nonsubject referent. The importance of benefactive relationship between the subject referent and the honorific target of the nonsubject referent is illustrated by the contrasts in the (a) and (b) sentences of examples (5) and (6). (5) a. Abe-san ga sensei o o-tasuke-sita. Mr. Abe nom teacher acc hp-help-did ‘Mr Abe helped (Non-Subject Honorific: NSH) the teacher. b.# Abe-san ga sensei o o-korosi-sita. Mr. Abe nom teacher acc hp-kill-did ‘Mr Abe killed (Non-Subject Honorific: NSH) the teacher. (6) a. Abe-san ga sensei kara hon o o-kari-sita. Mr. Abe nom teacher dat book acc hp-borrow-did ‘Mr Abe borrowed (Non-Subject Honorific: NSH) a book from the teacher. b.# Abe-san ga sensei kara hon o o-nusumi-sita. Mr. Abe nom teacher dat book acc hp-steal-did ‘Mr Abe stole (Non-Subject Honorific: NSH) a book from the teacher. The first of each pair, i.e., (5a) and (6a), contains a verb that is semantically benefactive, whereas verbs in the second of the pair, i.e., (5b) and (6b), denote counterbenefactive actions. Example (5b) and (6b) with counter-benefactive expressions are normally not acceptable without a special context. This unacceptability or the rarity of use is also borne out in an Internet search, which did not yield such an instance, although this does not mean that sentences with o-korosi-suru are absolutely impossible. A similar Internet search readily finds examples with verbs of benefactive actions such as (7) and (8).3 (7) [excerpts from a monthly magazine Seiron] Bannen Sonbun o o-tasuke-siteita koto kara mo, later years Sun Wen acc hp-help-did.prog fact from also
3. In contrast, there are many sentences with o-tasuke-suru and o-kari-suru. The web search of o-tasuke-simasu (HP-help-do.AH), for example, yielded 3,330 cases.
Context in constructions
tyuugoku to no kankei o omonjiteita koto ga China with gen relationship acc was valuing fact nom wakarimasu understand ‘The fact that [my grandfather] was valuing the relationship with China could also be seen from the fact that (he) was helping (NSH) Sun Wei in his later years.’
(8)
[web page description] HP yoo sozai o o-kari-siteimasu. home page use materials acc hp-borrow-do.prog.ah ‘(We) have borrowed (NSH) materials for home pages.’
In examples (1), (3)–(5) and (7) the target of honorification, a nonsubject, is a core argument of the main verb expressed as the direct object. It is worth noting, however, that the nonsubject target can be a non-argument as illustrated in (6a) and (8), and perhaps more clearly in (9), which is an example from a web page dedicated, according to the writer, to beginners of aromatherapy. The target of the honorification in (9) is an implicit nonargument referent, namely, the neophyte aromatherapists. (9) [web page for aromatherapy novices] ... aroma terapii no kihontekina koto o aromatherapy gen basic thing acc o-kaki-sinakereba toyuu simeikan ni karare, ... hp-write-do.neg.cond comp mission.sense by be.driven ‘being driven by the sense of mission that (I) should write (NSH) the basics of aromatherapy...’ These examples show that, regardless of the grammatical role that the target nonsubject takes, an elevated referent is understood to be in a benefit transfer relation with the subject referent in the o-Verb(stem)-suru construction. 4. O-V(stem)-suru construction: Direction toward performative honorifics 4.1
Use 2: Subject referent ≈ Speaker; non-subject referent ≈ addressee
In (9), the aromatherapy novices, who are the target of honorification, are also the readers (addressees) of the web page, and the recipients of the benefit of the writing. In other words, the humbled subject referent is the writer and the target of the benefit transfer and of honorification is the addressee. This is well-motivated since
Yoshiko Matsumoto
the most important and inescapable relation in a speech context is that of the speaker with the addressee. There are numerous instances in which this condition holds. (10) and (11) are attested examples in which the referent of the indirect object is the addressee and the subject referent is the speaker. Researchers such as Dasher (1995) consider the examples of this kind to be prototypical cases of the nonsubject honorific expressions. (10) anoo, tyotto o-tazune-simasu ga... um a little hp-ask-do.ah but ‘(lit.) um, (I) have a question to ask (NSH) (you), ...’ [at a department store] (11) zyaa atode ronbun o-okuri-simasu then later academic article hp-send-do.ah ‘(I) will send (NSH) (you) (my) article later, then.’ [after a colloquium] Example (12) is cited by Tsujimura (1992: 478) as an utterance by a cooking instructor in a TV program. The semantic goal of pouring is the dish that the instructor (i.e., the speaker) is cooking, which would not be a usual trigger for honorifics. Tsujimura explains that the speaker used the form to imply the pouring of soy sauce is done for the benefit of the addressees, the TV viewers. It is doubtful whether the naturalness of speech would be maintained if the viewers were expressed as a nonsubject in the sentence (i.e., to say “I’m pouring in a little soy sauce for you” in this context is unnatural), but such intention of benefit is implicit in the nature of the program in which the cooking demonstration is done for the benefit of the viewers and the speech context is normally intended to be refined and polite. Similarly in (13), an example from a web page, the prospective vacationers at the National Park Resort Villages will benefit from the link and are the target of the honorification, although they are the nonargument referents, not explicitly mentioned in the sentence. (12)
[cooking instructor in a TV cooking program; cited in Tsuji-mura 1992] o -syooyu o syoosyoo o-ire-itasimasu. hp-soy sauce acc a little hp-pour in-do.hum.ah ‘(I) pour in (NSH) a little soy sauce (to something that is being cooked).’
(13) [description of a web link to the Association of National Park Resort Villages] Nihon no kyuuka o o-tasuke-simasu. Japan gen vacation acc hp-help-do.ah ‘(The Association of National Park Resort Villages) will help (NSH) Japanese vacation.’
Context in constructions
(14)
[pilot on an international flight] (same as (3)) ato go-zikan hodo de hizuke-henkoo-sen o more 5-hours about in international dateline acc o-mukae-itasimasu. hp-meet -do.hum.ah ‘(We) will reach (NSH) the international dateline in about 5 hours..’
The sentence in (14), which was given as (3) earlier, was an announcement made by the captain of an international flight. Under the analysis of the conventional use of the o-Verb(stem)-suru construction given in the last section, it should be unacceptable, since the only explicit nonsubject referent, the international dateline, is an unlikely target of benefit transfer and therefore of honorification. The only potential target of the honorification in the given context is the passengers, but their involvement to the described event, i.e. meeting the international dateline, is minimal and therefore the benefit of such event to them is less evident than in the other two examples. For this reason, the sentence probably strikes normative users of the construction as even odder than the other two examples. However, it is clear in these examples that the speaker is being humbled and the addressee (or the reader) is elevated through the use of the o-Verb(stem)-suru construction in the speech context, where the speaker wishes to be polite and to imply some sort of benefit relation with the addressee. This observation leads us to the third and currently controversial use of the construction. 4.2
Use 3: Addressee honorific use (Subject Referent ≈ Addressee)
In the controversial, but increasingly attested, uses, the subject referent is the addressee. That is, the addressee, the intended target of honorification appears in the subject position, and the term nonsubject honorification is no longer descriptive. Examples such as (15) – (20) could easily be regarded simply as deviations from the normative point of view. On the other hand, similar usage was also observed by Hudson (1999), who pointed out the fact that o-Verb(stem)-suru forms are sometimes used as hyper-polite forms, and by the report of the Council of the National Language (Bunka Cho (Bureau of Culture)1996), which noted that more than 40% of respondents in a survey judged examples such as these to be acceptable. This tendency is also supported by relatively frequent use (about 40 to over 200 cases depending on the expression) of such variations found on a Internet search in 2003. (15)
[receptionist to a guest] Kotira de o-mati-site-kudasai. here loc hp-wait-do-give.imperative.ah ‘Please wait (NSH) here.’
Yoshiko Matsumoto
(16)
[attendant at a fast food store to a customer; cited in Asahi Newspaper] O-motikaeri-simasu ka hp-take out-do.ah question particle ‘(lit.) Are (you) taking it out (NSH)?/Is this for takeout?’
(17)
[delivery service clerk to a customer] (=(4)) ..soo site-itadaku to o-mati-suru o-zikan mo so do-receive.hum if hp-wait-do hp-time also arimasen kara... not.exist.ah so ‘...if you could do so, you would not have time to wait (NSH) around, so...’
(18)
[waitress to diners at an upscale Japanese restaurant] Nori o o-nose-site mesiagatte kudasai. dried.sea.weed acc hp-place.on.top-do eat.sh give.me.ah ‘Please place (NSH) the dried seaweed on top before you eat (SH) (it).’
(19) [college student to an elderly person, prefacing her question about the interviewee’s personal background] ..eto, o-kotae-dekiru han’i de kamawanai n desu ga...4 um hp-answer-do extent fine. with nmlz is.ah but ‘.. um, it’s OK if you could answer (NSH) to the extent that you feel comfortable...’ (20) [graduate student to a professor, an invited guest speaker at a graduate colloquium] OHP o o-tukai-suru n desu ne ohp acc hp-use-can.do nmlz is.ah sentence fin. particle ‘(You) are going to use (NSH) an OHP, aren’t you?’ It is interesting to note that these examples were often attested in the context of the general relation of service provider to client, as in the cases in (15) – (18). For instance, in (18), the waitress is explaining to guests at a Japanese restaurant that the seaweed that was brought on a separate plate was to be put on the dish before eating. It is clear from the context that the target of honorification is not the object referent, the seaweed. In all examples (15) – (20), it still remains true that the oVerb(stem)-suru construction is used in the context in which the speaker intends to be humble and polite to the addressee and the speaker and addressee are engaging in or implying a benefit transfer relation at the occasion of communicating the
4. The word dekiru ‘can do’ is the potential form of suru ‘do.’
Context in constructions
described event. In this sense, the pragmatic properties of the o-Verb(stem)-suru construction are still maintained. 5. Frame semantic representations of the three uses of the O-Verb(stem)-Suru construction The differences among the three uses of o-Verb(stem)-suru construction are schematically given in the following diagrams. The inner box represents a cognitive frame, the event described by the construction focusing on the two participants who are crucial in the construction, while the outer box represents the frame of the speech context, an interactional frame. Specifically, it represents the politeness context with the two main participating parties, i.e. the speaker and the addressee. The first diagram illustrates the conventional use (Use 1) of o-Verb(stem)-suru construction as a nonsubject referent honorific, and the second one shows the middle stage (Use 2), so-to-speak, at which the subject referent of the described event is identified with the speaker and the addressee is identified with the explicit or implied target of benefit transfer in the event. Use 2 can be seen as a special case of Use 1. In the third diagram (Use 3), however, the identification between the speaker and the subject referent, and between the nonsubject referent and the addressee is not maintained, and the benefit transfer relation is relevant only between the speaker and the addressee, the two prominent participants of the speech context rather than of the described event. Therefore, this diagram represents what is in effect a performative or addressee honorific construction. Although these three uses are concurrently found at present, the third use has begun to be noticed most recently among the three, and is treated in how-to books as one of the common mistakes of honorific uses. Use 1
Speaker
Subj Ref
Non-Subj Ref
Event Benefit transfer Politeness speech context
Addressee
Yoshiko Matsumoto
Use 2
Speaker
Subj Ref Event
Non-Subj Ref
Addressee
Benefit transfer
Politeness speech context
Use 3
Speaker
Subj Ref
Non-Subj Ref
Addressee
Event
Politeness speech context
Benefit transfer
The currently observed tendency is that the profiled referents of the o-Verb(stem)suru construction are shifting from the subject and nonsubject referents of the sentence describing the event to the two prominent participants of the discourse. The direction of diachronic change in honorifics from the referent to addressee honorifics has been pointed out by researchers including Tsujimura (1992, etc.), Dasher (1995), and Traugott and Dasher (2002). For example, the addressee honorific form -masu originated in a referent honorific. Interestingly, in his analysis of honorific grammaticalization, Dasher (1995) stated that the o-Verb(stem)-suru form does not participate in this direction of change. Although the use of oVerb(stem)-suru is not identical to the use of -masu in that there should be a non-subject referent in the described event and a benefit transfer relation between the speaker and the addressee, the data and analysis that I provided in this paper show that this construction also participates in the general direction of development and change manifested in the previous findings especially when analyzed from the point of view of intersubjectification. These observations therefore illustrate that usage and its context are indeed fundamental to the construction of grammar.
Context in constructions
6. Conclusion The diagrams of the three uses of the o-Verb(stem)-suru construction are identified with representations of the “cognitive frame” and “interactional frame” of the utterances (e.g., Fillmore 1982) in the analysis above. The specific interactional frame illustrated as the outer box in the diagram in the o-Verb(stem)-suru construction is a very general one-- the politeness speech context in which the speaker and the addressee are situated. The inner box in the diagram illustrates the cognitive frame, which represents the event described by the utterance. The advantage of using both types of frames to describe the o-Verb(stem)-suru construction is the capability of illustrating the relation between the contextual participants and the event participants, and how these two interact, which is schematically indicated by connecting lines between them. It is also noteworthy that concepts in Frame Semantics and Construction Grammar well accommodate even those constructions that are contextually sensitive, such as honorifics constructions, and allow us to describe their variations and change. The findings in this light strongly suggest that the usage and the context of linguistic constructions can crucially affect the grammatical conditions of the constructions. I also touched upon the significance of a pragmatically supported network of knowledge in determining the grammaticality of constructions such as complex noun phrases in Japanese. Pragmatic factors, usage, and context are not peripheral or accidental, but are critically woven into the way in which a grammar of human language operates. Without such pragmatic and contextual knowledge, linguistic facts to which systematic explanations can be given would be very limited: the grammaticality of noun phrases such as [[toire ni ikenai] komaasyaru] (mentioned earlier) cannot be explained. The same is true of other similar examples that defy regular syntactic explanations. Further, the seemingly ungrammatical honorific constructions which are the focus of this paper, in fact follow the general tendency of variations and direction of change, would have to remain as unexplained deviations. What these considerations demonstrate is that a human language is not a collection of codes that exist void of human experience and speech contexts, and that a grammar of a human language will be limited and impoverished if it does not reflect such property.
Yoshiko Matsumoto
7. List of abbreviations ACC: AH: COMP: COND: DAT: GEN: HP:
Accusative Addressee Honorific Complement Conditional Dative Genitive Honorific Prefix
HUM: NEG: NMLZ: NOM: NSH: PROG: SH:
Humble Form Negavie Nominalizer Nominative Nonsubject Honorific Progressive Subject Honorific
References Bunka cho (Bureau of Culture). 1996. Kokugo Shingi-kai Hookokusho (Report of Council of the National Language) 20. Tokyo: Ministry of Finance. Dasher, Richard Byrd. 1995. The Grammaticalization in the System of Japanese Predicate Honorifics. PhD dissertation, Stanford University. Fillmore, Charles. 1975. An alternative to checklist theories of meaning. In Proceedings of the First Annual Meeting of the Berkeley Linguistics Society, 123–131. Berkey CA: BLS. Fillmore, Charles. 1982. Frame semantics. In Linguistics in the Morning Calm, Linguistic Society of Korea (ed.), 111–138. Seoul: Hanshin. Harada, S-I. 1976. Honorifics. In Syntax and Semantics, 5: Japanese Generative Grammar, Masayoshi Shibatani (ed.), 499–561. New York NY: Academic Press. Hayashi, Shiro & Fujio Minami. 1973–4. Keigo Kooza (Lectures on Honorifics). Tokyo: Meiji Shoin. Hudson, M. 1999. Teinei-hyoosiki tosite no kenzyoo-doosi (Japanese humbling verbs as politeness markers). In Gengogaku to nihongo kyooiku (Linguistics and Japanese language education), Yukiko Sasaki Alam (ed.), 259–274. Tokyo: Kurosio Syuppan. Kikuchi, Yasuto. 1980. Jooge-taiguu-hyoogen no kijutu (Description of expressions for the treatment of status distinctions). Kokugogaku (Japanese linguistics): 39 - 54. Kokuritsu Kokugo Kenkyujo (The National Language Research Institute), 1957. Keigo to keigoishiki (Honorifics and the attitudes toward honorifics). Tokyo: Shueido Shuppan. Kuno, Susumu. 1983. Sin Nihon-bunpoo Kenkyuu (New studies in Japanese grammar). Tokyo: Taishukan. Kuno, Susumu. 1987. Honorific marking in Japanese and the word formation hypothesis of causatives and passives. Studies in Language 11(1): 99–128. DOI: 10.1075/sl.11.1.05kun Lakoff, Robin T. 1980. How to look as if you aren’t doing anything with words: Speech act qualification. Versus Quaderni di Studi Semiotici Milano 26/27, 29–47. Laury, Ritva & Ono, Tsuyoshi. 2005. Data is data and model is model: You don’t discard the data that doesn’t fit your model! Language 81: 218–225. DOI: 10.1353/lan.2005.0026 Martin, Samuel Elmo. 1975. A Reference Grammar of Japanese. New Haven CT: Yale University Press. Matsumoto, Yoshiko. 1985. A sort of speech act qualification in Japanese: chotto. Journal of Asian Culture 9: 143–159.
Context in constructions Matsumoto, Yoshiko. 1997a. Noun-Modifying Constructions in Japanese: A Frame Semantic Approach [Studies in Language Companion Series 35]. Amsterdam: John Benjamins. DOI: 10.1075/slcs.35 Matsumoto, Yoshiko. 1997b. The rise and fall of Japanese nonsubject honorifics: The case of ‘oVerb-suru’. Journal of Pragmatics 28: 719–740. DOI: 10.1016/S0378-2166(97)00072-6 Matsumoto, Yoshiko. 2001. Tyotto: Speech act qualification in Japanese revisited. Japanese Language and Literature 35(1): 1–16. DOI: 10.2307/489703 Matsumoto, Yoshiko. 2008. Variations in Japanese honorification. Deviations or a change in the making? In Constructional Re-organization [Constructional Approaches to Language Series 5], Jaakko Leino (ed.), 89–104. Amsterdam: John Benjamins. McCawley, James D. 1993. A Japanese and Ainu linguistic feast (Review of Shibatani 1990). Journal of Linguistics 29: 469–484. DOI: 10.1017/S0022226700000402 Newmeyer, Frederick J. 2003. Grammar is grammar and usage is usage. Language 79, 682–707. Sweetser, Eve. 1990. From etymology to pragmatics. Cambridge: CUP. DOI: 10.1017/CBO9780 511620904 Tannen, Deborah. 1979. What’s in the frame? Surface evidence for underlying expectations. In New Directions in Discourse Processes, Roy O. Freedle (ed.), 137–181. Norwood NJ: Ablex. Tokieda, Motoki. 1941. Kokugogaku genron (The theory of Japanese linguistics). Tokyo: Iwanami Syoten. Traugott, Elisabeth Closs & Dasher, Richard Byrd. 2002. Regularity and Semantic Change. Cambridge: CUP. Tsujimura, Toshiki. 1967. Gendai no keigo (Modern honorifics). Tokyo: Kyobunsha. Tsujimura, Toshiki. 1968. Keigo no sitekikenkyuu (Historical studies on honorifics). Tokyo: Tokyo-do Shuppan. Tsujimura, Toshiki. 1992. Keigo ronkoo (Essays on honorifics). Tokyo: Meiji Shoin. Yamada, Yoshio. 1931[1924]. Keigohoo no kenkyuu (The studies of honorification). Tokyo: Hobunkan.
The use and interpretation of “regional” and “standard” variants in Japanese conversation Shigeko Okamoto It is often said that language standardization has been steadily advancing in modern Japan and that speakers in regional Japan are now bi-dialectal and code-switch between “Standard” and “regional” Japanese. The notion of codeswitching, however, assumes the existence of varieties, or well-defined linguistic systems, that are distinct from each other. In this study, I examine the use of “Standard Japanese” and “regional dialects” and argue that it is much more complex and dynamic than what can be possibly accounted for in terms of the notion of code-switching involving two distinct varieties. I explore an alternative account employing the notion of variant choice and characterize the social meanings of “Standard” and “regional” variants as context-dependent and as multiple and ambiguous.
1. Introduction It is often noted that due to a variety of factors associated with the nation’s modernization since the Meiji Restoration in 1867--e.g. the government language policy, which promoted “Standard Japanese” (“SJ”, hereafter)1 and the development of mass communication and transportation system--language standardization has been steadily advancing in modern Japan, reducing regional differences (e.g., Kobayashi et al. 1996; Yoneda 1997; Chapter 10 in Kunihiro et al. 1999; Sanada 2002; Kobayashi & Shinozaki 2003). At the same time, however, it has also been pointed out that “regional varieties” of Japanese have not lost their vitality, and that many speakers are now bi-dialectal and code-switch between the “regional” and “standard varieties” according to the situation (e.g., Inoue 1988; Shibata 1988; Miyake 1995; Long 1996; Sanada 1996; Carroll 2001). For example, Inoue (1988: 1. In this paper I use double quotes around the terms that refer to particular varieties of Japanese, such as “Standard Japanese” and “Osaka dialect,” because these categories are not set in stone and do not constitute linguistic systems that are distinct from each other, as will be discussed in the ensuing sections.
Shigeko Okamoto
19–20) states that while standardization is proceeding, dialectal forms are still used vigorously (in informal situations), showing “a kind of bi-dialectalism.” Long (1996: 122) notes that whereas the typical trend in “Eastern-Japanese dialect changes” has been “a shift from the local dialect to the standard,” “the trend in Western-Japanese dialects has been towards bi- (or multi-) dialectalism” with “situational code-switching.” Sanada (1996, 2000) uses the term neo-hoogen, or neodialect, to refer to a koodo ‘code’, or a linguistic system, that lies in between hoogen ‘dialect’ and hyoojun-go ‘standard language’. Speakers are then said to code-switch between the three “varieties” depending on the situation (Sanada 1996: 7; cf. Ball 2004; Hosotani 2004). These previous studies provide us with some general ideas about the use of “regional dialects” and “Standard Japanese.” However, they are usually based on informal observations or self-report survey data, which tend to elicit answers according to the respondents’ normative expectations about language use. It is thus unclear exactly how and why individual speakers use their “regional dialect” and “Standard Japanese” in actual conversations. In particular, terms such as “bi-dialectalism” and “situational code-switching” imply that there are distinct varieties, or dialects, of Japanese and that speakers use one variety at a time depending on the situation. But can we in fact identify such distinct varieties, or codes, as welldefined linguistic systems? If we can, how? If not, why? There have been numerous studies of code-switching behavior, but most of them involve two (or more) different languages (e.g., most of the chapters in Heller 1988; Milroy & Muysken 1995; Auer 1998, 2000; Li Wei 2000; Gardner Chloros 2009), and not different dialects of the same language. Some “exceptions” include Blom & Gumperz (1972) on Norwegian and Giacalone Ramat (1995) on Italian, both of which treat the two varieties/dialects in question as distinct codes, or linguistic systems (but cf. Auer 1976 cited by Giacalone Ramat 1995, Maehlum 1996). In this study, I analyze the use of “regional” and “Standard” Japanese in conversations and argue that it is much more complex and dynamic than what can be possibly accounted for in terms of the notion of code-switching, involving two distinct varieties.2 I then attempt to explore an alternative account employing the notion of variant choice. My analysis demonstrates that the social meanings of “Standard” and “regional” variants vary according to the context (i.e., are indexical) and are thus multiple and ambiguous and that the speaker uses these variants as resources for constructing a desired social context. My approach is in line with recent research on sociolinguistic variation that emphasizes the need to consider fluid indexical meanings, viewing language as a social practice taking place in specific social, cultural, and historical contexts (e.g., Agha 2003; Eckert 2008, 2012; Blommaert 2010; Johnstone 2. See Section 3.3 for further discussion of the notion of code-switching.
The use and interpretation of “regional” and “standard” variants
2010). Based on the findings of this study I will also consider their implications for the notion of Japanese grammar and that of dialect as a linguistic category. 2. Method The conversations examined in this study come from two sets of data collected in two locales in Western Japan: one involving speakers in Osaka City, one of the major cities in Japan, and the other involving speakers in Shuho-cho, a small rural town in Yamaguchi prefecture. The Osaka data consist of 7 conversations in which one speaker, Teru,3 talks with different persons, as shown in Table 1. Teru was born and grew up in Osaka; her interlocutors are all from the Kansai region, which includes Osaka. The Yamaguchi data consist of 9 conversations. One speaker, Sachiko, is in 6 conversations, and another speaker, Yuri, is in 4 conversations; each of them talks with a different person in each conversation, as shown in Table 2. (In one of the conversations, Sachiko and Yuri are the participants.) Sachiko was born in Fukuoka, Kyushu, but has lived in Shuho-cho and its vicinity most of the time since childhood; Yuri was born and lived in Shuho-cho throughout her life. The interlocutors of Sachiko and Yuri are also from Shuho-cho or its vicinity. The researcher was not present in any of the conversations. The three pivotal speakers, Teru, Sachiko, and Yuri, were asked to record conversations in which they talk with people in different kinds of relationships, such as a family member, close friend, superior, or acquaintance. That is, these three pivotal speakers chose their interlocutors on their own. The topics of conversations were self-selected.4 Table 1. Conversations from the Osaka data set T-1: T-2: T-3: T-4: T-5: T-6: T-7: 3.
Teru (T, female, age 28), talking with her close friend (A, female, age 27) over the phone her husband (B, male, age 25) at home her co-worker (C, female, age 26) during a break from work her supervisor (D, female, age 28) and other co-workers (C and E, female, age 26 and F, male, age 27) at a staff meeting her supervisor (D) and co-workers (C and E) during a break her mother-in-law (G, age late 40s) over the phone a customer (H, female) at the department store where she works The names of all participants are pseudonyms.
4. For the Yamaguchi data the first 8 minutes of each conversation was examined. However, for the Osaka data it was not possible to have the same length of each conversation, because some of the conversations were very short; the lengths of the seven conversations vary from 3 to 8 minutes.
Shigeko Okamoto
Table 2. Conversations from the Yamaguchi data set Sachiko (S, female, age 58), talking with S-1 her mother (I, female, age 78) at S’s home S-2 her older sister (J, female, age 60) at J’s home S-3 her daughter (K, female, age 23) at their home S-4 her close friend (L, female, age 59) at L’s home S-5* her daughter’s friend (Y, female, age 25) at Y’s home S-6 her daughter’s former high school teacher (M, female, age 69) at M’s home Yuri (Y, female, age 25), talking with Y-1 her mother (N, female, age 44) at their home Y-2 her father (O, male, age 58) at their home Y-3 her close friend (P, female, age 23) at P’s home Y-45 her friend’s mother (S, female, age 58) at Y’s home
In analyzing the data, I use the term variants (of a variable) to refer to individual linguistic forms, or different manifestations of a particular linguistic item (e.g. the particles na and ne are variants, see Table 3), and the term code to refer to a linguistic system, or variety, consisting of a set of linguistic (phonological, morpho-syntactic, lexical, etc.) properties. Further, as noted earlier, I use terms such as “Osaka Table 3. Examples of variants relevant to the Osaka data set Phonological variants a. The use of a contour tone within a vowel in one- or two-mora words (e.g. hi ‘fire’, having a L to H6 contour; me in ame ‘rain’, having a H to L contour) in “OD,” but not in “SJ.” b. Devoicing of i and u in a voiceless environment (e.g. i in kita ‘north’) occurs much less frequently in “OD” than in “SJ” (Inoue 1989). c. Pitch-accent patterns differ considerably, as illustrated below: “OD” “SJ”
kaze ga HHH LHH
kawa ga HLL LHL
kasa ga LLH HLL
ame ga LHL HLL
N.B. kaze ‘wind’, kawa ‘river’, kasa ‘umbrella’, ame ‘rain’, ga, a subject marker (Shibatani 1990)
5.
In Table 2, S-5 and Y-4 are the same conversation.
6. L stands for a low pitch, and H a high pitch. 7. Abbreviations used for word classes are as follows: A, adjective; Adv, adverb; AN, adjectival noun; N, noun; and V, verb.
The use and interpretation of “regional” and “standard” variants
Morphological variants Interactional particles “OD” A/AN/N/V de7 A/V n A/V ne (n) A/AN/N/V yan na(a)
“SJ” A/AN/N/V yo A/V no A/V no(yo)/n da A/AN/N/V janai ne(e)
English gloss I tell you you know you see right?, isn’t it? right?, isn’t it
Inflectional endings “OD” Adv-oo/uu AN/N-ya A/AN/N/V-yaro(o) V-haru V-hen/n V-n V-oote/oota V-toru
“SJ” Adv-ku N/AN-da A/AN/N/V-daro(o) V-reru/o-V-ni naru V-nai V-ru V-tte/tta V-teru
English gloss adverbial ending copulative auxiliary isnt’ it ...?, probably subject-referent honorific negation (not V) V infinitive V gerundive/past V progressive/stative
“SJ” S tte
English gloss quotative
Conjunctive particles “OD” S te Lexical variants Verbs/Adjectives/Adjectival Nouns/Adverbs “OD” akan donai ee erai honma sha::nai shindo(i) yooke
“SJ” ikenai doo/donna ni ii taihen na hontoo shikataganai tsukareta takusan
English gloss not good how good, fine terrible really cannot be helped tired many/much
“SJ” (sore) ja(a)/nara sore nara/son nara (sore)de dakara dattte/demo/dakedo
English gloss then then and then, so therefore but
Conjunctions “OD” hona honnara (ho)nde soya kara soya kedo
Shigeko Okamoto
dialect” (“OD” hereafter), “Yamaguchi dialect” (“YD” hereafter), and “regional dialects” in double quotes, because they are broad cultural categories and cannot be regarded as well-defined, static linguistic categories that are distinct from each other (see Sections 3 for further discussion). Note also that many expressions do not have “regional” and “Standard” variants--i.e., the same expressions are used in different “varieties.” In analyzing the data, I focus only on the forms that I, the analyst, identified as “regional” and “Standard” variants in consultation with native speakers of “OD” and “YD,” Nakagawa (1982) and Yamamoto (1982).8 Some readers, however, may not entirely agree with the identifications used in this study, which is possible due to the dynamic nature of the indexical values associated with these variant forms (see Section 3.4 for further discussion). For the Osaka data, I examine phonological, morphological, and lexical variables.9 For the Yamaguchi data, I examine only morphological and lexical variables10 because phonological features, including the pitch-accent patterns, are mostly the same as those in “SJ,” as noted in Nakayama (1982) and Shibatani (1990). Tables 3 and 4 show examples of the variants relevant to the two data sets. 3. Results and discussion 3.1
Inter- and intraspeaker variation
Table 5 shows Teru’s use of SJ variants in total as well as in each of the three categories--phonology, morphology, and lexicon. Table 6 shows Sachiko’s and Yuri’s use of SJ variants in total as well as in each of the two categories--morphology and lexicon. As we can see in these Tables, there is wide interspeaker variation. Overall, Teru used “SJ” variant forms 38% of the time, Sachiko 16% of the time, and Yuri 48% of the time. In particular, Table 6 shows a clear difference between the two speakers, Sachiko and Yuri, even though they have lived in the same town most of their life and spoke in similar situations for the present data. For example, when 8. For “OD” I (a native speaker of “Kansai dialect,” which includes “OD”) consulted one other native speaker for “OD”; for “YD” I consulted three native speakers for “YD.” 9. Some of the morphological and lexical variables may appear to be phonological variables. However, I treat a variable as morphological if it involves a grammatical function (e.g. verb conjugations) or pragmatic function (e.g. interactional particles). I treat a variable as lexical if it involves only a particular lexical item or a particular group of lexical items. 10. I examined only those variables that occurred frequently and were identified as “YD” and “SJ” variants by the native-speaker consultants, as they could not give “SJ” counterparts for some of the “regional” forms.
The use and interpretation of “regional” and “standard” variants
Table 4. Examples of variants relevant to the Yamaguchi data set Morphological variables Inflectional endings “YD” Adv-oo/uu AN/N-ya/ja A/AN/N/V-yaro(o) A/AN/N/V–jaroo V-choru V-n V-oote/oota
“SJ” Adv-ku AN/N-da A/AN/N/V-daro(o) A/AN/N/V-daro(o) V-teru V-nai V-tte/tta
English gloss adverbial ending copulative auxiliary isnt’ it ...?, probably isnt’ it ...?, probably V progressive/stative negation (not V) V gerundive/past
Lexical variables Verbs/Adjectives/Adjectival Nouns/Adverbs “YD” ee donee sonee
“SJ” ii do, donna so, sonna
English gloss good, fine how such
“YD”
“SJ”
English gloss
hetara hede, hete honde, hode (he)jakedo (he)yakara
soshitara sorede sorede, dakara dakedo, demo dakara
then and, so so, therefore but so, therefore
Conjunctions
the interlocutor is the speaker’s mother, Sachiko used “SJ” variants only 5% of the time, while Yuri used them 40% of the time; when the interlocutor is a close friend, Sachiko used “SJ” variants only 5% of the time, while Yuri used them 49% of the time; and when the interlocutor is the speaker’s superior (in terms of age, status, etc.), Sachiko used “SJ” variants 62% of the time, while Yuri used them 85% of the time. A number of factors, including the generational difference,11 may contribute to these interspeaker differences, which require more investigation. But the important point here is the fact that interspeaker variation can be quite large. These Tables also show that there is strikingly large intraspeaker variation. Although there are individual differences, all three speakers altered the use of 11. Although there are individual differences, younger generations tend to use more “Standard” variants than older generations.
Shigeko Okamoto
Table 5. Use of “SJ” variants by Teru Morphology
Lexicon
Total12
almost none
“SJ”/Total # of tokens (%) 16/123 (13)
“SJ”/Total # of tokens (%) 4/40 (10)
“SJ”/Total # of tokens (%) 20/163 (12)
almost none
36/183 (20)
16/49 (33)
52/232 (22)
almost none
25/77 (33)
3/8 (38)
28/85 (33)
almost none
25/57 (44)
19/25 (76)
44/82 (54)
almost none
75/126 (60)
29/48 (60)
104/174 (60)
almost none
5/7 (71)
6/7 (86)
11/14 (79)
some
35/36 (97)
9/10 (90)
44/46 (96)
217/609 (36)
86/187 (46)
303/796 (38)
Conversation # Phonology (interlocutor(s)) “SJ” T-1 (close friend) T-2 (husband) T-3 (colleague) T-4 (supervisor & colleagues) T-5 (supervisor & colleagues) T-6 (mother-in-law) T-7 (customer) Total
variants considerably depending on the situation/interlocutor. For example, the use of “SJ” morphological and lexical variants ranged from 12% to 96% for Teru, from 5% to 62% for Sachiko, and from 27% to 85% for Yuri. Further, the ratios of “regional” and “SJ” variants change gradually from conversation to conversation. 3.2
Mixing “regional” and “Standard” variants
Tables 5 and 6 show that all three speakers mix “regional” and “Standard” variants in a quite complex way. Although the proportions differ, they all mixed the two kinds of variants in all the conversations examined here. Some of the mixings occur across sentences, while others occur within the same sentence or even within the same phrase (see below for examples). Tables 5 and 6 also show that the speakers mix variants within a particular category of variables (e.g. morphological variables). That is, all three speakers mixed “regional” and “Standard” morphological
12. “Total” combines the frequencies of morphological and lexical variables.
The use and interpretation of “regional” and “standard” variants
Table 6. Use of “SJ” variants by Sachiko and Yuri Conversation # (interlocutor)
S-1 (mother) S-2 (older sister) S-3 (daughter) S-4 (close friend) S-5 (daughter’s friend) S-6 (daughter’s teacher) Total Y-1 (mother) Y-2 (father) Y-3 (close friend) Y-4 (friend’s mother) Total
Morphology
Lexicon
Total
“SJ” /Total # of tokens (%) 0/13 (0)
“SJ”/Total # of tokens (%) 1/8 (13)
“SJ “/ Total # of tokens (%) 1/21 (5)
0/40 (0)
6/27 (22)
6/67 (9)
2/35 (6)
4/12 (33)
6/47 (13)
0/41 (0)
4/41 (10)
4/82 (5)
7/30 (23)
8/12 (67)
15/42 (36)
6/14 (43)
7/7 (100)
13/21 (62)
15/173 (9)
30/107 (28)
45/280 (16)
9/27 (33)
3/3 (100)
12/30 (40)
11/52 (21)
5/8 (63)
16/60 (27)
10/35 (29)
20/26 (77)
30/61 (49)
20/25 (80)
15/16 (94)
35/41 (85)
50/139 (36)
43/53 (81)
93/192 (48)
variants, includeing the variants of the same variable) in every conversation but in differing proportions; they also mixed “regional” and “Standard” lexical variants (including the variants of the same variable) in almost all conversations. For the phonological variables, however, the speaker (Teru) used “OD” variants most of the time and “SJ” variants only occasionally. Additionally, variant mixings occur across different categories of variables (phonological, morphological, lexical). For example, Table 5 shows that Teru does not utilize the phonological, morphological, and lexical variables to the same extent. She hardly used “SJ” phonological forms, but used “SJ” morphological and
Shigeko Okamoto
lexical forms much more frequently.13 In other words, when she used “SJ” morphological or lexical forms, they were usually pronounced with the “OD” phonological patterns (e.g. using the “SJ” lexical variant ii ‘good’ instead of its “OD” counterpart ee, but pronouncing ii with the “OD” pitch pattern LH rather than with the “SJ” pattern HL).14 Speakers also mixed variants across morphological and lexical variables, using, for example, an “SJ” lexical form and a “regional” morphological form in the same sentence (see below for examples). The kinds of variant mixings observed in the data are summarized in Table 7. Another point to note here is that not all variables in the same category are used in the same way. For example, for some morphological variables, “SJ” variants were used only infrequently, while for others, “regional” variants were used only infrequently; and still for others, both variants were used relatively frequently, as shown in Tables 8 and 9.15 Although more research is required, this difference in the use of variants among different variables may suggest that the process of standardization does not affect all variables to the same extent (at a given point in time). Table 7. Ways of mixing “regional” and “Standard” variants a. Variant mixings in the same conversation – Intersentential mixings – Intrasentential mixings b. Variant mixings within the same category of variables – Using both “regional” and “Standard” morphological variants in the same conversation – Using both “regional” and “Standard” lexical variants in the same conversation – Using “regional” (“OD”) phonological variants primarily, and “Standard” variants only occasionally in the same conversation c. Variant mixings across different categories of variables – Using “OD” phonological variants along with “SJ” lexical and/or morphological variants in the same conversation – Using “SJ” lexical (or morphological) variants and “regional” morphological (or lexical) variants in the same conversation
13. Although more research needs to be done, this asymmetry in the use of phonological and morphological/lexical features may suggest that it is more difficult to acquire/change phonological features than morphological and lexical features. 14. This kind of mixing was in general not seen in the Yamaguchi data since the “YD” phonological patterns are mostly the same as the “SJ” patterns. 15. Tables 8 and 9 show the three speakers’ use of “regional” and “Standard” variants for the individual morphological variables that had more than 10 tokens in all the conversations each speaker had.
The use and interpretation of “regional” and “standard” variants
Table 8. Teru’s use of individual morphological variables in Conversations T1-T7 “OD”form
“SJ” form
S-te V-oote/oota ‘ Adv-oo/uu V-toru/totta na(a) V-hen A/AN/N/V de A/V ne (n) AN/N-ya/yat(ta) A/AN/N/V-yaro(o) A/AN/N/V-yan
S-tte V-tte/-tta Adv-ku V-teru/teta ne(e) V-nai A/AN/N/V yo A/V no(yo)/n da AN/N-da/dat(ta) A/AN/N/V-daro(o) A/AN/N/V-janai
# “SJ”of tokens (%) 31 (91%) 12 (86) 10 (77) 40 (71) 64 (50) 7 (35) 6 (26) 6 (15) 4 (4) 1 (3) 0 (0)
Table 9. Sachiko’s and Yuri’s use of individual morphological variables in Conversations S1-S6 and Y1-Y4 “YD” form
“SJ” form
# of “SJ” tokens (%)
Sachiko V-oote/oota V-chor(u) V-n AN/N-ya/ja
V-tte/tta V-ter(u) V-nai AN/N-da
10 (45) 1 (3) 2 (2) 0 (0)
Yuri V-oote/oota Adv-oo/uu V-n V-chor(u) AN/N-ya/ja
V-tte/tta Adv-ku V-nai V-ter(u) AN/N-da
30 (88) 8 (67) 3 (19) 4 (15) 5 (10)
Let us look at some examples.16 [In the following examples “regional” variants are boldfaced and “Standard” variants are underlined. Phonological forms (e.g. pitchaccent patterns) are not shown.] 16. ? . , :: (.) (n)
The following transcription conventions are used in this study. rising intonation = xxx = “xxx” overlapping with “yyy” falling intonation = yyy = slight fall indicating continuity = = xxx latching lengthened segment [xxx] nonlexical phenomena (e.g. [laugh]) pause of less than 1 second (xxx) unintelligible speech pause of n seconds
Shigeko Okamoto
(1) [from Conversation T-5: Teru (T), talking with her supervisor (D) and other coworkers (C and E) during a break at work] 1 T: ... dakara koo nagaku kuroku ke ga koo atte, ‘ ... so he has this long, black hair,’ 2
C:
[laugh]
3 T:
hige ga koo tsunagatteru kara:: ( . ) ne, ‘and his mustache is connected like this.’
4 D:
= =soo yuu hito inakatta. ‘That sort of person wasn’t there.’
5 T:
= =inakatta desu::? ‘There wasn’t?’
6 D:
uun ‘No.’
7 T:
orahen katta n ka na:: ‘I wonder if he wasn’t there.’
8 D:
=uun= ‘No.’
9 T:
=chau= hito otta n kamo shiren. ‘A different person may have been there.’
10
mayuge ga iyoo ni futoine =yan,= ‘His eyebrows are unusually thick.’
11 C:
= [laugh]=
12 T:
tsunagatteru tte kanji. ‘It’s like they are connected to each other.’
13 C:
= =[laugh] wairudo:: ‘He’s wild.’
14 D:
[laugh] wairudo-kee ‘He’s a wild-type’
15 T: . . .
demo chotto otaku ya ne. ‘But he’s a little nerd, isn’t he?’
16 T:
a:: oishi soo. ‘oh, that sounds good.’
17
ii na:: ‘That’s nice.’
The use and interpretation of “regional” and “standard” variants
Example (1) is from the Osaka data set. In lines 1–14, Teru mixed “OD” and “SJ” morphological and lexical variants intersententially. That is, in lines 1, 3, 5, and 12, she used only “SJ” variants, and in lines 7, 9, and 10 she used only “OD” variants. In lines 15 and 17, on the other hand, Teru mixes “OD” and “SJ” variants intrasententially. In 15 demo is an “SJ” lexical variant, ya an “OD” morphological variant, and ne an “SJ” morphological variant; in 17 ii is an “SJ” lexical form, and na:: an “OD” morphological form. Phonologically, Teru used “OD” variants in all her utterances except for the vowel e in ke ‘hair’ in line 1, for which she did not use the “OD” form with a contour tone (H to L). In other words, even if Teru used “SJ” morphological and lexical forms, they were pronounced with “OD” forms. For example, in line 5, she used the “SJ” lexical variant i ‘to be’ and the “SJ” morphological variant -nakat(ta) ‘negative’ but pronounced the whole sentence using the “OD” phonological forms, including the pitch-accent patterns (HHLLL for inakatta instead of the “SJ” LHLLL) and the voiced long vowel u:: in desu:: ‘copulative auxiliary’. Note also that she rephrases inakatta in line 5 as orahenkatta in line 7; that is, she not only mixes “SJ” and “OD” forms in the same category of variables (i.e. morphological and lexical), but also for the same variables i ‘to be’ and -nai ‘nagative’ in the same conversation.
(2) [from Conversation S-5/Y-4; Sachiko (S), talking with Yuri (Y) at Yuri’s house] 1 S: Y-san nara koso:: (2) kyooryoku-shite moratte honto kanshashichoru yo? ‘I’m grateful that you (Y) are helping me (as I thought you would).’ 2 Y:
ie ie ie. ‘No, no, no.’
3 K-san mo sugoi benkyoo gambatchotte mitai ya kara ma watashi ga kore gurai dattara watashi mo honto o-tetsudai dekiru n de, ‘K seems to be studying very hard, and I can help her with (small) things like this, so ...’ 4 S:
=ureshii= ‘I’m happy.’
5 Y:
=itsu demo= itte kudasai. ‘Please ask me any time.’
6 S:
ureshii, honto. ‘I’m happy, really.’
7
iya sasuga dookyuusee ja ne:: ‘Well, you are worthy of (her) classmate, aren’t you?’
Shigeko Okamoto
8 Y:
ie ie ie honto. ‘No, no, no, really.’
9 K-san mo hoikuen no koro kara zu::tto issho da shi, ma, kookoo wa chigatta kedo, moo kochira daigaku sunde kara mo, iroiro ohanashi-shitari, tegami no yaritori toka mo shiteru kara, ‘I’ve been with K since kindergarten, although our high schools were different, and after I was done with the college, we’ve been doing things like chatting and exchanging letters, so ...’ . . . 10 S: sore hodo shimpai wa sen kedo ne::, yappari nanto naku shimbun toka de ne:: iroiro ano kiji o miru to ne::, ano ‘I’m not worried that much, (but) as expected, when I see various things in newspapers, and the like, uh ...’ 11 Y:
u::n onna no ko da shi. ‘and she is a girl, so ...’
12 S:
[laugh] chotto po- porotto omou koto ga aru kedo ne::, ‘[laugh] There are times I think of her a little bit, but,’
13 ee ka:: shitai koto o sashichoru n yake:: to =omotte= ne::, ‘(then) I think it’s probably OK, because I’m letting her do what she wants to do,’ 14 Y:
=uun,= sugoi =na::= ‘Yeah, that’s great.’
15 S:
=soo=omotchoru i ne::. ‘I think so.’
16 Y:
sugoi wakari areru okaasan de ii na:: tto omotte, ‘I thought it’s nice that you are a very understanding mother,’
Example (2) is a conversation between Sachiko and Yuri from the Yamaguchi data set. Let us look at Sachiko’s speech first. In lines 1 and 10 Sachiko mixed “YD” and “SJ” morphological variants intra-sententially. In 7 she used one “YD” morphological variant. In lines 13 and 15, she mixes “YD” and “SJ” forms intra-sententially as well as across different categories of variables. That is, in line 13 the “SJ” morphological variant omotte ‘think’ is used along with the “YD” lexical variants ee ‘it’s OK’ and sashi ‘let do’ and the “YD” morphological variants -choru ‘(V) progressive/stative’, ya ‘copulative auxiliary, and -ke:: ‘because’. In line 15, the “SJ”
The use and interpretation of “regional” and “standard” variants
lexical variant soo ‘so’ and the “SJ” morphological variant omot ‘think’ is used along with the “YD” morphological variants choru ‘(V) progressive/stative’ and i (for emphasis). Compared to Sachiko, Yuri used fewer “YD” forms (see also Figure 2 and Table 6). In line 3, Yuri mixed “YD” and “SJ” morphological variants intra-sententially. Note also that for the variable da (a copulative auxiliary), Yuri used both “YD” variant ya and “SJ” variant da within the same sentence. In the other turns (lines 5, 9, 11, and 16) in this extract, Yuri used only “SJ” variants (“SJ” morphological forms in lines 5, 9, and 11, and “SJ” lexical and morphological forms in line 16). These analyses demonstrate that the use of “regional” and “Standard” variants is highly complex and varies widely among speakers and across situations. In what follows I consider how best we can account for these linguistic practices. 3.3
Code-switching or variant-choice?
As mentioned earlier, the use of “regional” and “Standard Japanese” has been described in terms of the notion of code-switching. This treatment implies that one can identify dialects, or codes, as distinct linguistic systems. However, the data examined in this study indicate that it is difficult to identify such distinct linguistic categories. I argue then that the kind of data examined here can be better accounted for in terms of the notion of variant choice (for each individual variable), rather than that of code-switching. Before we discuss this issue further, let us first consider the notion of code-switching. Gumperz (1982: 66) defines code-switching as “meaningful juxtaposition of what speakers must consciously or subconsciously process as strings formed according to the internal rules of two distinct grammatical systems” (cf. Brown 2003) As noted earlier, in research on code-switching “two distinct grammatical systems” have mainly concerned different languages rather than different dialects of the same language. Example (3) illustrates code-switching between two different languages, Swahili and English.
(3) [Setting: A conductor on a Nairobi bus has just asked a passenger where he is going in order to determine the fare (in Swahili). Myers-Scotton 2000: 151–152, quoted from Scotton and Ury 1977. “P” stands for passenger, and “C” conductor.] 1 P (Swahili): Nataka kwenda posta. ‘I want to go to the post office.’ 2 C (Swahili): Kutoka hapa mpaka posta nauli ni senti hamsini. ‘From here to the post office, the fare is 50 cents.’ (Passenger gives conductor a shilling from which there should be 50 cents in change.)
Shigeko Okamoto
3 C (Swahili): Ngojea change yako. ‘Wait for your change.’ (Passenger says nothing until a few minutes have passed and the bus nears the post office where the passenger will get off.) 4 P (Swahili):
Nataka change yangu. ‘I want my change.’
5 C (Swahili):
Change utapata, Bwana. ‘You’ll get your change, mister.’
6
P (English):
I am nearing my destination.
7
C (English): Do you think I could run away with your change?
According to Myers-Scotton, in line 6 the passenger switches from Swahili to English, which is what Myers-Scotton considers a marked language in the given situation, in order to “encode authority and educational status,” although “the conductor counters by matching the passenger’s marked choice, showing that he too can compete in any power game (involving here ability to speak English)” (Myers-Scotton 2000: 151). As this example illustrates, code-switching is used to convey a particular social or pragmatic meaning or meanings by language alternation. The meanings signaled by code-switching are diverse, including solidarity, distance, authority, irony, a particular speech act (e.g., request, refusal), emphasis, sequential contrast (e.g., side- remark), and topic change. (see Heller 1988; Auer 1998; 2000; Gardner-Chloros et al. 2000; Li 2000; Myers-Scotton 2000; Gardner-Chloros 2009 among others). The Japanese data examined earlier seem quite different from the kind of codeswitching involving two languages as illustrated in Example (3).17 In the Japanese case, there are many forms that are “shared” by “SJ” and a “regional variety” (“OD” or “YD”), forms that are not variables. Syntactic structures in the two kinds of varieties are largely the same. Further, as can be seen in Examples (1) and (2), only some of the morphological and lexical items are variables, having “regional” and “Standard” variants. Phonologically, “OD” forms are quite different from “SJ” forms, but “YD” forms are quite similar, if not exactly the same. These facts make it difficult to distinguish between dialects as distinct varieties. Let us consider, for example, lines 5 and 7 in Example (1), reproduced here as (1a). (1) a. [from Conversation T-5] 5 T: =inakatta desu::? ‘There wasn’t?’ 17. Note, however, that even when two “languages” are involved, there are many cases that cannot be fully accounted for in terms of the notion of language as a discrete linguistic system (e.g. Brown 2003; Pennycook 2003; Blommaert 2010).
The use and interpretation of “regional” and “standard” variants
6 D: uun ‘No.’ 7 T: orahen katta n ka na:: ‘I wonder if he wasn’t there.’ Can we say, for example, that Teru switches to “OD” when she says orahen in line 7? But that would mean that the utterance in line 5 is “SJ,” when in fact its status is ambiguous. Phonologically, the whole utterance in line 5 is made with “OD” forms; lexically and morphologically, i and na are “SJ” variants, but what about katta desu? Is it an “SJ” or “OD” form? The syntactic structure of this utterance is not a variable, either. How do we treat those forms that are not variables? Similar questions are applicable to the utterance in line 7. The English-Swahili code-switching example discussed above concerns switching codes within a conversation. But the term is sometimes used in the sense of code choice, or situational code-switching (Blom & Gumperz 1972). In this case, a particular code, or language, is chosen in a particular situation (e.g. school, home). Here, too, one must assume the existence of distinct codes, or grammatical systems. We saw earlier that Teru, Sachiko, and Yuri all changed the amount of “SJ” (or “regional”) forms used in each conversation examined. But in all conversations they used both “regional” and “standard” variants. In other words, the use of variants is a matter of degree, which makes it difficult to draw non-arbitrary lines between “dialects.” If we regard any admixture of “regional” and “Standard” forms as constituting a hybrid dialect, or neo-hoogen ‘neo-dialect’ (Sanada 1996, 2000), the speech of Teru, Sachiko, and Yuri in every conversation examined in this study should be regarded as a hybrid dialect. But such a treatment cannot capture the considerable intraspeaker variation across different conversations, since such variation does not seem to be random but functionally motivated (see Section 3.4 for further discussion). It may be possible for a speaker to identify an individual linguistic form as a “regional” or “Standard” variant to a certain extent, although this identification itself is not always determinate, as indicated by the use of double quotes around expressions like “OD” or “SJ” variants (see Section 3.4 for further discussion). The data examined in this study suggest that one cannot identify a series of utterances in a conversation as representing a particular dialect, or grammatical system, that is distinct from other dialects. In sum, my analyses show that the notion of variant-choice is more appropriate than that of code-switching to account for the kind of linguistic behavior seen in the present data. The speaker chooses one of the variants for each variable at each relevant point and then produces utterances that have a greater or lesser amount of “Standard” (or “regional”) variants. I argue that variant-choice can be a resource for the speaker’s style management, which I will discuss in Section 3.4.
Shigeko Okamoto
3.4
Variant-choice as a resource for style management
The fact that the speakers gradually altered the proportion of “regional” and “Standard” variants in different conversations suggests that the choice of variants is not random, but rather functionally motivated. The examination of how variants are distributed in different contexts indicates that it is closely related to the speaker’s affective stance, such as formal or polite attitude, toward various features of the context. (I use the notion of formality in the sense of one’s restrained or spontaneous/relaxed attitude toward a given situation (Okamoto 1999, 2008b), and the notion of politeness in the sense of one’s deferential attitude toward another individual (Agha 1993).) The assessment of what stance to take may be made based on the consideration of a variety of aspects of the context, such as the relationship between the speaker and the interlocutor (e.g., relative status, degree of intimacy), domain (e.g., work, home), genre (e.g., making a report at a business meeting, small talk, soliloquy), speech act type (e.g., request, apology), and topic. And it may affect the choice of various linguistic expressions, including honorifics18 and (non-)standard forms.19 In the current data, our three speakers used “regional” variants most frequently with the interlocutors that are equal or lower-status persons and/or familiar persons. The use of “regional” variants decreased when they spoke with their superiors and/or persons they don’t know well. Further, even in the same conversation, variants were used differentially. For example, in a staff meeting in Conversation T-4, “SJ” variants were used more when the speaker (Teru) was presenting a report to the whole group than when she was participating in a discussion with other participants. Also, when a speaker (Teru) was engaging in a dialogue with a superior in Conversation T-5, “SJ” variants were used more frequently than when she made soliloquy-like utterances in the same conversation, as seen in lines 5 and 7 in Example (1). These differential uses thus seem to be related to the speaker’s stance toward each context, that is, his/her assessment of what is the appropriate degree of formality and/or politeness he/she should display in a given situation. As is well known, “SJ” was created at the beginning of modern Japan based on the variety presumably used by educated Tokyoites (e.g., Yasuda 1999), or by residents in the Yamanote area in Tokyo. That is, originally, the variety now known as “SJ” was one of the regional (as well as class-based) dialects. However, I argue that when it was selected as “SJ,” that is, when it was “recontexualized” (Blommaert 18. See Okamoto 1999 and 2008a for a discussion of formality and honorifics. 19. I assume that a particular linguistic form (e.g., an honorific) does not directly index a particular contextual aspect (e.g., degree of intimacy or status difference between participants); rather, such an aspect is indirectly indexed through the pragmatic meaning (e.g., formality) associated with each linguistic form. See Okamoto 2004 for further discussion of this issue.
The use and interpretation of “regional” and “standard” variants
2005) in the national context and juxtaposed and compared with other regional dialects, it was reinterpreted as an index of social meanings, such as correct, superior, formal, and polite, in contrast to the social meanings assigned to non-Standard Japanese, that is, other regional dialects (see Agha 2003: 251 for a discussion of the parallel reinterpretation of Received Pronunciation in English; see also Milroy 2001: 535 for a discussion of the notion of standard language culture that tends to assign certain values to standard varieties). Accordingly, it started to be used more frequently in contexts thought to require the speaker to take a formal and/or polite stance. If this is the case, does that mean that “regional” and “Standard” variants index the speaker’s affective stance of formality and/or politeness rather than his/ her regionality? I argue that the indexical meanings of these variants are ambiguous between formality/politeness and regionality (cf. Ball 2004). In fact, these variants used in specific contexts may be further interpreted as indexes of other meanings. For example, “SJ” variants interpreted as the speaker’s formal and/or polite stance may be further linked to a particular relationship (e.g., status difference) or a particular genre (e.g. public speech). Or they may be further accorded affective meanings such as beautiful, refined, and feminine. Moreover, depending on the context, “positive” meanings such as formal and polite may be reinterpreted negatively as aloof, cold, and bland, as, for instance, when a speaker of Standard Japanese does not attempt to assimilate his/her speech to the local dialect after moving to regional Japan (e.g., Sato 1996; also cf. Okamoto 2011, in which she discusses similar variability of indexical meanings in regard to honorifics). That is, the indexical meanings of these variants are variable, multiple, and ambiguous. Eckert (2008: 463), drawing on Silverstein’s (2003) notion of idexical order, argues that “[p]articipation in discourse involves a continual interpretation of forms in context, an in-the-moment assigning of indexical values to linguistic forms. A form with an indexical value, what Silverstein calls an nth order usage, is always available for reinterpretation – for the acquisition of an n + 1st value.” The present analysis also indicates that the meanings of the two kinds of variants are indeterminate in that not everyone may share the same interpretations for all the variables. As mentioned earlier, I, the analyst, classified these variants in consultation with native-speakers of “OD” and “YD,” Nakagawa (1982), and Yamamoto (1982) in order to have a certain reference point for my data analysis.20 But I have been using terms such as “Standard” and “regional” in double quotes throughout this paper, because these classifications are not absolute. The data examined here suggest that standardization is definitely affecting speakers in 20. As I discuss here, not everyone may classify variants in exactly the same way, but even if one uses different classifications, one is likely to find diversity in the use of these variants, although further study is called for regarding this issue.
Shigeko Okamoto
regional Japan, and that there are wide interspeaker (or intergenerational) variation in the degree of standardization. Accordingly, the meanings of a particular variable may be indeterminate or ambiguous due to individual differences. Take, for instance, the interactional particle ne/na. If the speaker primarily uses the “OD” variant na and uses the “SJ” variant ne only occasionally, he/she may perceive ne more strongly as an “SJ” variant, or “foreign,” expression as well as an formal and/ or polite expression. If, on the other hand, the speaker uses ne rather than na almost always, he/she may not perceive ne as an “SJ” variant, or “foreign,” form, nor as a formal and/or polite expression.21 Furthermore, as noted earlier, the degree of standardization may vary considerably depending on the variable. If a particular variable (e.g., the particles ne and na) becomes completely standardized (for a particular speaker), then there will be no variant-choice that can be used as a resource for style-management concerning formality or politeness. This suggests that the indexical meanings of variants of each variable (e.g., na/ne) are inseparable from the context of use, and that they are not fixed but relative to specific contexts. Eckert (2008: 453) argues that: “the meanings of variables are not precise or fixed but rather constitute a field of potential meanings – an indexical field, or constellation of ideologically related meanings, any one of which can be activated in the situated use of the variable.” Such a dynamic view of indexical meanings is also required in accounting for the interpretation of Japanese variables in question. As we saw above, the speakers mixed two kinds of variants in intriguing ways-inter- and intrasententially, within the same kind of variables (e.g., morphological variables), and across different kinds of variables (e.g., phonological and morphological variables). It seems that mixing the two kinds of variants in an array of ways enables the speaker to express different stances by indexing different degrees of regionality, formality and/or politeness (as well some other context-related meanings), thereby fine-tuning the overall effect. Johnstone (2010: 389), in her discussion of dialects (in English and other languages) in contact, argues for viewing “regional speech not just as an automatic consequence of where a person was born or raised but as a resource for social action,” and that work on ‘style’ (e.g., Eckert and Rickford 2001; Coupland 2007) “has showed how social identities can be evoked or created through the use of particular linguistic forms and has suggested that, at least for some people and in some ways, regional forms could serve such purposes.” The variant choice we saw in this study offers an additional example of such use of linguistic forms as a social action, in which variants serve as resources for style management in creating a desired context, including identities and relationships. 21. Note also that variables such as da, janai, sa (in “SJ”) tend to be used only in informal situations even in “SJ.” The use of such “SJ” variants by speakers of “regional dialects” may not be associated with formality.
The use and interpretation of “regional” and “standard” variants
4. Conclusion and implications for grammar The foregoing analyses demonstrate that it is too simplistic to characterize the kind of linguistic behavior observed in this study as code-switching between two or three distinct varieties of Japanese. Notions such as hoogen ‘dialect’ and neo-hoogen presuppose that there exist well-defined speech varieties, or linguistic systems “out there,” that can be clearly distinguished from each other. These terms may be useful, and perhaps even necessary, to refer to broad regional differences. But at the same time, the present data suggest that these notions are reifications of discrete categories vis-à-vis the synchronic and diachronic continuum of differences. Such a static view of language cannot fully capture the complex and dynamic process involved in the use and interpretation of variant choice, as observed in this study. Lastly, I would like to discuss briefly some broader theoretical issues. Although more research involving more speakers in many different regions needs to be done, the findings of this study raise a number of questions regarding the notion of Japanese grammar as well as that of dialect as a linguistic category. Here I note three interrelated issues to consider: 1. Linguistic variation and the grammar of a language: When we say “Japanese grammar,” what should be considered as constituting the Japanese language? How can the vast regional and individual variation be dealt with? If the grammar of Japanese (or any other language) reflects the shared knowledge among the native speakers in a particular speech community, how can we conceive of “native speakers” or “a speech community”? To what extent is the knowledge shared? The kind of variation observed in this study is not simply free variation, but rather concerns evolving structure. How can the diverse knowledge individual speakers have be represented as the grammar of a particular language? 2. Non-discreteness of linguistic categories: Just as the notion of grammar of a particular language requires reconsideration, the present study also raises a question of whether one can talk about the grammar of a particular “dialect” (e.g. such as “SJ,” “OD”, and “YD”). It might be possible to talk about the grammar of a dialect if one were to refer to some idealized or vaguely characterized linguistic systems as dialects. The notion of code-switching assumes the existence of such linguistic systems. However, the analysis in this study suggests that such a view is not congruent with what individual speakers like Teru, Sachiko, and Yuri know and do in practice, and that dialectal differences are on a continuum of synchronic and diachronic variation. 3. Ambiguity and fluidity of indexical meanings: As discussed above, the indexical meanings of variants examined in this study suggest that the relation of form and meaning is not fixed, but is shaped through use embedded in a
Shigeko Okamoto
historical context in which standardization of Japanese has been advancing for a variety of reasons, as noted in Section 1.22 That is, the relation of form, meaning, and context of use are in a dialectic relationship. How can the grammar of a particular language deal with this kind of dynamic process? The three questions raised here are in line with the theoretical orientation in recent research that seeks a more dynamic approach to sociolinguistic variation, viewing language as a social practice rather than as an abstract linguistic system. In discussing language use and globalization, Blommaert (2010: 18), for example, claims that “[e]stablished notions such as ‘language’, ‘culture’, or ‘place’ are not useful in an analysis of objects that are necessarily mixed, hybrid, local as well as delocalized (or delocalizable), dynamic and unstable.” Similarly, Johnstone (2010: 392) argues that while people “can be said to speak ‘the same’ language or variety, or to ‘share’ ways of talking,” we must assume that “such sharing is never complete,” and that “it must be renegotiated in every interaction.” She contends that “languages and dialects, like localities, are ‘imagined’,” and culturally constructed. In sum, the observations made in this study suggest a need for a view of language as a dynamic process that constantly renews linguistic structure, rather than as a static autonomous entity, or to use Milroy’s (2001) term, “a stable synchronic finite-state idealization.” In order to account for variation and change in linguistic structure, we need to examine closely individual speakers’ patterns of linguistic practice, which may be much fuzzier and much more unstable than we may expect them to be. 5. Acknowledgment I would like to thank those who participated in the data collection, in particular, Chika Kobayashi, Kimiko Izumi, and Takae Izumi. This study is an extension of Okamoto (2008a, b). I’m grateful to Kaori Kabata and Tsuyoshi Ono for their valuable comments on the earlier version of this paper, which was presented at the Symposium on Functional Approaches to Japanese Grammar (2004). I also wish to thank the other participants of the symposium as well as the anonymous reviewer and the following people for their helpful comments and discussion: Hiromi Ano, Reiko Ano, Ritva Laury, Karen Mistry, and Raymond Weitzman.
22. See Yasuda (1999), Komori (2000), and Carroll (2001) for discussions of the historical background of the standardization of Japanese.
The use and interpretation of “regional” and “standard” variants
References Agha, Asif. 1993. Grammatical and indexical convention in honorific discourse. Journal of Linguistic Anthropology 3(2): 131–163. DOI: 10.1525/jlin.1993.3.2.131 Agha, Asif. 2003. Social life of cultural value. Language and Communication 23: 231–273. DOI: 10.1016/S0271-5309(03)00012-0 Auer, Peter. 1976. Konversationelle standard dialect-kontinua (code shifting). Deutsche Sprache 2: 97–124. Auer, Peter. (ed.). 1998. Code-Switching in Conversation: Language, Interaction, and Identity. London: Routledge. Auer, Peter. 2000. A conversation analytic approach to code-switching and transfer. In The Bilingual Reader, Li Wei (ed.), 166–187. London: Routledge. Ball, Christopher. 2004. Repertoires of registers: Dialect in Japanese discourse. Language and Communication 24: 355–380. DOI: 10.1016/j.langcom.2004.01.004 Blom, Jan Petter & Gumperz, John J. 1972. Social meaning in linguistic structure: Code-switching in Norway. In Directions in Sociolinguistics: The Ethnography of Communication, John J. Gumperz & Dell Hymes (eds), 407–434. New York NY: Holt, Rinehart and Winston. Blommaert, Jan. 2005. Discourse: A Critical Introduction. Cambridge: CUP. DOI: 10.1017/ CBO9780511610295 Blommaert, Jan. 2010. The Sociolinguistics of Globalization. Cambridge: CUP. DOI: 10.1017/ CBO9780511845307 Brown, Becky. 2003. Code-convergent borrowing in Louisiana French. Journal of Sociolinguistics 7(1): 3–23. DOI: 10.1111/1467-9481.00208 Carroll, Tessa. 2001. Language Planning and Language Change in Japan. Richmond: Curzon Press. Coupland, Nicholas. 2007. Style: Language Variation and Identity. Cambridge: CUP. DOI: 10.1017/CBO9780511755064 Eckert, Penelope. 2008. Variation and the indexical field. Journal of Sociolinguistics 12(4): 453– 476. DOI: 10.1111/j.1467-9841.2008.00374.x Eckert, Penelope. 2012. Three waves of variation study: The emergence of meaning in the study of sociolinguistic variation. Annual Review of Anthropology 41: 87–100. Eckert, Penelope & Rickford, John. 2001. Style and Sociolinguistic Variation. Cambridge: CUP. Gardner-Chloros, Penelope. 2009. Code-switching. Cambridge: CUP. Gardner-Chloros, Penelope, Charles, Reeva & Cheshire, Jenny. 2000. Parallel patterns?: A comparison of monolingual speech and bilingual codeswitching discourse. Journal of Pragmatics 32: 1305–1341. DOI: 10.1016/S0378-2166(99)00120-4 Giacalone Ramat, Anna. 1995. Code-switching in the context of dialect/standard language relations. In Milroy & Muysken (eds), 45–67. Gumperz, John J. 1982. The sociolinguistics of interpersonal communication. In Discourse Strategies, John J. Gumperz (ed.), 9–37. Cambridge: CUP. DOI: 10.1017/CBO9780511 611834.004 Heller, Monica (ed.). 1988. Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin: Mouton. DOI: 10.1515/9783110849615 Hosotani, F. 2004. Osaka-shi hoogen washa no sutairu kirikae [Style switching of speakers of Osaka City dialect]. Handai Shakai Gengogaku Keyuu Nooto (Osaka University Sociolinguistics Research Notes) 4: 42–59.
Shigeko Okamoto Inoue, Fumio. 1988. Dialect image and new dialect forms. Tokyo University of Foreign Languages, Area and Culture Studies 38: 13–22. Inoue, Fumio. 1989. Kotoba-zukai Shinfuukee (New Landscape of Language Use). Tokyo: Kaiyama Shoten. Johnstone, Barbara. 2010. Indexing the local. In The Handbook of Language and Globalization, Nikolas Coupland (ed.), 386–405. Oxford: Blackwell. DOI: 10.1002/9781444324068.ch17 Kobayashi, Takashi & Shinozaki, Kooichi (eds). 2003. Gaidobukku: Hoogen Kenkyuu (A Guidebook: Dialect Research). Tokyo: Hitsuji-shobo. Kobayashi, Takashi., Shinozaki, Kooichi & Onishi, Takeharu (eds). 1996. Hoogen no Genzai (The Present State of Dialects). Tokyo: Meiji-shoin. Komori, Yōichi. 2000. Nihon-go no Kindai (The Japanese Language in Modern Times), 2nd edn. Tokyo: Iwanami Shoten. Kunihiro, Tetsuya, Inoue, Fumio & Long, Daniel (eds). 1999. Takesi Sibata: Sociolinguistics in Japanese Contexts. Berlin: Mouton. DOI: 10.1515/9783110821307 Long, Daniel, 1996. Quasi-standard as a linguistic concept. American Speech 71: 118–135. DOI: 10.2307/455480 Maehlum, Brit. 1996. Codeswitching in Hemnesberget: Myth or reality? Journal of Pragmatics 25: 749–761. DOI: 10.1016/0378-2166(95)00027-5 Milroy, James. 2001. Language ideologies and the consequences of standardization. Journal of Sociolinguistics 5(4): 530–555. DOI: 10.1111/1467-9481.00163 Milroy, L. & Muysken, P. (eds). 1995. One Speaker, Two Languages: Cross-Disciplinary Perspectives on Code-Switching. Cambridge: CUP. DOI: 10.1017/CBO9780511620867 Miyake, Yoshimi. 1995. A dialect in the face of the standard: A Japanese case study. In Proceedings of the 21st Annual Meeting of the Berkeley Linguistics Society, 217–225. Berkeley CA: BLS. Myers-Scotton, Carol. 2000. Code-switching as indexical of social negotiations. In The Bilingual Reader, Li Wei (ed.), 137–165. London: Routledge. Nakagawa, K. 1982. Yamaguchi-ken no hoogen (The dialect of Yamaguchi prefecture). In Kooza Hoogengaku: Chuugoku, Shikoku-chihoo no Hoogen (Seminar in Dialectology: Dialects of Chuugoku and Shikoku Areas), Kiichi Iitoyo, Sukezumi Hino & Ryōichi Satō (eds), 142– 173. Tokyo: Kokushokankokai. Okamoto, Shigeko. 1999. Situated politeness: Manipulating honorific and nonhonorific expressions in Japanese conversations. Pragmatics 8(2): 51–74. Okamoto, Shigeko. 2004. Ideology in linguistic practice and analysis: Gender and politeness in Japanese revisited. In Japanese Language, Gender and Ideology: Cultural Models and Real People, Shigeko Okamoto & Janet S. Shibamoto Smith (eds), 113–130. Oxford: OUP. Okamoto, Shigeko. 2008a. Speech style and the use of regional (Yamaguchi) and standard Japanese in conversations. In Japanese Style Shift [Pragmatics & Beyond New Series 180], Kimberly Jones & Tsuyoshi Ono (eds), 229–249. Amsterdam: John Benjamins. Okamoto, Shigeko. 2008b. The use of “regional” and “standard” Japanese in conversations: A case study from Osaka. In Japanese Applied Linguistics: Discourse and Social Perspectives, Junko Mori & Amy Snyder Ohta (eds), 132–159. New York NY: Continuum. Okamoto, Shigeko. 2010. Politeness in East Asia. In Handbooks of Pragmatics, Vol. 6: Interpersonal Pragmatics, Miriam A. Locher & Sage Lambert Graham (eds), Berlin: Mouton de Gruyter. Okamoto, Shigeko. 2011. The use and interpretation of addressee honorifics and plain forms in Japanese: Diversity, multiplicity, and ambiguity. Journal of Pragmatics 43: 3673–3688.
The use and interpretation of “regional” and “standard” variants Pennycook, Alistair. 2003. Global Englishes, rip slyme, and performativity. Journal of Sociolinguistics 7(4): 513–533. DOI: 10.1111/j.1467-9841.2003.00240.x Sanada, Shinji. 1996. Chiiki-go no Dainamizumu: Kansai-hen (Dynamism of Regional Languages: The case of Kansai). Tokyo: Ofu. Sanada, Shinji. 2000. Henyoo-suru Osaka-kotoba (Changing Osaka Language). Gengo (Language) 29(1): 49–53. Sanada, Shinji. 2002. Hoogen no Nihon Chizu (A Dialect Map of Japan). Tokyo: Koodansha. Sato, Kazuyuki. 1996. Hoogen-shuryuu-shakai: Kyoosei toshite no Hoogen to Hyoojungo: Tohokuhen (Dialect-dominant Society: Coexisting Dialectal and Standard Varieties: The Case of Tohoku). Tokyo: Ofu. Shibata, Takeshi. 1988. Ikite iru Nihon-go: Hoogen tansaku (Living Japanese: Exploring dialects). Tokyo: Kodansha. Shibata, Takeshi. 1999[1965]. The rise and fall of dialects. In Takeshi Shibata: Sociolinguistics in Japanese Contexts, Tetsuya Kunihiro, Fumio Inoue & Daniel Long (eds), 183–206. Berlin: Mouton. (Translated from the Japanese original published in 1965). Shibatani, Masayoshi. 1990. The Languages of Japan. Cambridge: CUP. Silverstein, Michael. 2003. Indexical order and the dialectics of sociolinguistic life. Language and Communication 23: 193–229. DOI: 10.1016/S0271-5309(03)00013-2 Wei, Li. 2000. The Bilingualism Reader. London: Routledge. DOI: 10.4324/9780203461341 Yamamoto, Toshiharu. 1982. Osaka-fu no Hoogen. Kooza Hoogengaku: Kinki-chihoo no Hoogen (Seminar in Dialectology: Dialects of the Kinki Area), Kiichi Iitoyo, Sukezumi Hino & Ryōichi Satō (eds), 196–228. Tokyo: Kokushokankokai. Yasuda, Toshiaki. 1999. Kokugo to Hoogen no Aida: Gengo-koochiku no Seeji-gaku (Between the National Language and Dialects: Political Science of Language Construction). Kyoto: Jinbun Shoin. Yoneda, Masato. 1997. Survey of standardization in Tsuruoka, Japan: Comparison of results from three surveys conducted at 20-year intervals. Nihongo Kagaku (Japanese Linguistics) 2: 24–39.
Index A accomplishment 154, 155, 204, 205, 234 achievement 153–155, 157, 160, 199, 221, 228, 243, 252, 254 activity 38, 45, 153–158, 160, 219, 223, 225–228, 231, 232, 234, 235, 237, 238, 240, 242, 243, 246, 253, 254, 256, 257 addressee 8, 16, 27, 30, 86–89, 96–99, 101, 115, 123, 124, 229, 243, 245, 246, 261, 263–267, 269–276, 302 adjectives 65–67, 69, 70, 74, 78, 79, 165, 202, 283, 285 affective 2, 4, 56, 58, 62, 100, 296, 297 – stance 296, 297 affiliation 228, 231, 233 agent 61, 87, 88, 110, 159, 190 agentivity 3, 159, 160, 167 agreement 114, 115, 168, 169, 209, 213, 221, 227–229, 232, 233, 243, 247, 250–253, 257, 258 aktionsart 92, 154, 170 alignment 227, 231, 232, 240, 245 allative 7, 171–174, 176, 177, 179, 181, 186 allomorph 137–139 alternation 20, 23, 137, 138, 152, 190, 294 analogy 7, 144, 149, 150 animator 245 aspect 7, 33, 59, 60, 62, 77, 91–93, 102, 105, 108–110, 116, 128, 131, 132, 150, 153–157, 160, 162, 164, 165, 167–170, 173, 176, 178, 195, 238, 254, 296 inherent – 154, 169 lexical – 154 assisted explaining 239–241, 243, 244, 246 atelic 153, 155–158 Ausdruck 55, 58, 61
author 25, 39, 41, 42, 80, 116, 176, 183, 245, 246, 255 auxiliary verbs 37–39, 50, 228 B background information 13–16, 30, 35, 198, 217 backgrounding effect 15, 16, 27, 29, 30 back-propagation 160 bare form 42 Basic Child Grammar 159, 160, 166 benefactive actions 268 benefit transfer 261, 264, 269, 271–274 bunsetu-bun 58 C causative 153, 154, 161, 162, 170, 173, 219 chinjutsu 86, 88, 105, 108, 256, 258 clipped form 64–66, 68, 70, 81 Coalition Model 163 code 3, 13–15, 30, 33, 55, 60, 62–65, 69, 70, 72, 75, 78, 93, 126, 275, 279, 280, 282, 293–295, 299, 301, 302 – switching 279, 280, 293–295, 299, 301, 302 cognitive frame 263–265, 273, 275 communicative intent 61, 77 Competition Model 163, 167 complementizer 13–15, 20, 21, 23, 24, 30–32, 34, 35, 38, 102, 104, 106, 127 compound 137, 139, 141, 143–146, 148–150, 234 conditional 13, 17, 100, 104, 120, 122, 276 connectionist 160, 161
construction – Grammar 1, 168, 261, 265, 267, 275 grammatical – 82, 220, 261, 263 noun-modifying– 262 o-V-suru – 261 contextual information 58, 263 contextual participants 265, 275 control 65, 82, 97, 159, 160, 167 conversation analysis 7, 107, 193, 256, 258 corpus 4, 79, 171, 172, 178–180, 182–184, 186–190, 198 creole 154, 155, 167 D degrees of subordination 13, 15, 17, 30 deliberate judgment 57 dentatsu 33, 85, 87–91, 97, 107 descriptive sentences 55, 56, 58, 61, 62, 76, 77 destination 173, 294 deviations 271, 275, 277 dialect 8, 100, 164, 191, 279–282, 284, 293–300, 299, 301–303 regional – 8, 279, 280, 284, 297, 298 neo- 280, 295 dialogue 7, 85, 87–89, 95–98, 101, 102, 296 direct causation 161, 163 direction 145, 157, 173, 174, 178, 181, 182, 184, 185, 188, 206, 269, 274, 275 direct object 81, 143, 146, 149, 264, 269 disagreement 207, 221, 227, 246, 250–253, 257 discourse system 109, 113, 115, 116, 123–125, 127 dispreferred action 246, 247, 249, 250, 253
Usage-based Approaches to Japanese Grammar distributional bias 157, 160 distributional learning 159 double judgment 77 dynamic 8, 83, 106, 126, 153–156, 158, 160, 168, 189, 190, 202, 224, 226, 279, 280, 284, 298–300 E emotion 2, 6, 55, 62–64, 68–70, 75, 77–79, 106, 108, 133, 257 empathetic understanding 234–238 entitlement 229, 236, 239, 243, 244 event A- 244 AB- 244 B- 244–246 – modality 112, 117, 118, 126 – participants 265, 275 evidential 90, 99, 100, 109, 110, 112, 115, 117–120, 124, 126, 127, 158 – modality 112, 117–120, 124, 126, 127 exclamatory sentences 56 exhaustive listing 74, 77, 78 experience 6, 55, 60–72, 75–78, 82, 118, 164, 207, 232, 234–240, 243–245, 254, 275 experiencer 55, 60–62, 64, 86, 190 expressions deictic – 79, 262 egocentric – 60 emotive – 216 one-term – 63, 64, 69 pre-linguistic – 64, 70 reflex – 63–65, 69, 70, 77, 78, 80 two-term – 63, 70, 72, 76, 78 expressive sentences 6, 55–63, 74, 78–81 external descriptive 59, 76, 77 F footing 229, 255, 256 foreground information 13–16, 25, 30 foregrounding effect 16, 19, 20, 27, 30 formality 296–298 frames 263–265, 275
frame semantics 261, 263–265, 267, 275, 276 French 85, 87, 153, 155, 301 frequency 3, 7, 35, 97, 135, 148, 149, 153, 161, 166, 170, 171, 180–186, 188–190 G gesture 64, 65, 212, 213, 218, 219, 221, 222 glottal stop 64, 65, 255 grammaticalization 6, 7, 33, 34, 82, 85, 87, 94, 98, 101, 105, 107, 108, 110, 131, 132, 165, 171, 190, 191, 221, 262, 274, 276 grammaticization (see grammaticalization) Greek 153 H honorifics 8, 261, 262, 264–270, 274–277, 296, 297, 302 hyooshutsu 56, 58 I iconicity 2, 9, 14, 33, 63, 82, 94 identities 298 imperfective 156–158 impression-expressive style 57 indexical – field 298, 301 – meanings 280, 297–299 – order 303 indirect causation 161, 163 indirect speech-acts 61 inflection 113, 120, 122, 159, 168 inflectional suffixes 127 innateness 153, 163, 165 input 7, 65, 101, 153, 154, 156–164, 166–170, 175 – based prototype formation 153, 154, 160, 161 – frequency 153, 166, 170 – matching model 163, 164 interactional frame 263, 265, 273, 275 interaction and grammar 131, 193, 217, 221, 258 internal expressive sentences 55, 59–63, 78, 79 interspeaker variation 284, 285 intersubjectification 85, 98–100, 108, 274
intersubjective 89, 97, 98, 100, 200, 217, 261, 265 intraspeaker variation 284, 285, 295 introspection 81, 85, 217 irrealis 96, 97, 111, 119–122, 126–128 irregular past 157 Italian 153, 155, 156, 160, 168, 280 J judgment test 171–173, 178, 187 juttai 57 juttei 85, 87–91 K kandoo-kantai 57 kantan-bun 56 Korean 7, 33, 68, 82, 105, 108, 109, 116–129, 131–133, 151, 153, 154, 161, 167–169, 175, 176, 256–258 L Language Bioprogram Hypothesis 154, 155, 167 language policy 279 layered structure 89, 90, 133 lexical-derivational-inflectional continuum 165 Lyman’s Law 137, 141–144, 150, 151 M medial wh- 163, 164 mental verbs 94, 95, 97, 98, 100, 101 mitai 37, 41–43, 45, 48, 50–52, 119, 120, 291 mitenkai-bun 58 modal – auxiliary 39 – content 42, 48, 114 modality epistemic – 93, 107, 110, 112, 117, 118, 120, 126, 133 deontic – 93, 167 mode pur 58, 62 mode vécu 58, 62 mood 14, 38, 83, 87, 90, 109–111, 113–116, 119–123, 125–128, 133 grammatical – 109, 110, 119, 122, 123, 126
Index indicative – 120, 121, 123 subjunctive – 121 morpheme boundary 40, 44 morphology 101, 105, 131, 151–154, 156, 157, 160–162, 167–170, 284, 286, 287 multi-layered grammatical organization 56 multimodal 203, 212 multiunit turn 196, 197, 199, 201, 202, 204–207, 209, 211, 213, 216, 217 N native speakers 3, 7, 16, 31, 39, 41, 42, 45, 79, 137–139, 141, 145, 149, 150, 175, 179, 186, 188, 225, 284, 299 n-desu-tte 42, 45, 51 negative evidence 160, 161, 163 neurological states 55, 63 nominative case 38, 63, 81, 153, 154, 161, 162, 168, 172 O object complement clauses 19 other-correction 246, 247, 249, 250 outcry 55, 57, 60, 62–64, 68, 72, 76 overextension 158, 160 P partial sentence 58 patient 61, 118 perception 6, 19, 55, 57, 62–74, 76–81, 93, 99, 111, 118, 129, 186 perfect 38, 60, 62, 98, 104, 176 perfective 110, 156, 157, 160 performative 258, 265, 267, 269, 273 pidgin 154, 167 PNPD 155, 157 Polish 162 politeness 100, 130, 273, 275, 276, 296–298, 302 – speech context 275 pragmatic 3, 14, 27, 34, 41, 59, 94, 105, 108, 132, 172, 190, 219, 261–264, 268, 273, 275, 284, 294, 296 – accommodation 27
preferred action 246, 247, 250, 253 preliminary to preliminaries 201 progressive 38, 86, 153, 156–158, 160, 164, 168, 170, 276, 283, 285, 292, 293 projection 93, 203, 220, 222, 226, 253 pronominal case 153, 154, 160 proposition 30, 31, 37, 43, 44, 48, 50, 56, 70, 73, 78, 87–89, 91, 93–95, 99, 103, 110, 112–115, 128, 132, 229, 264 propositional 7, 37, 38, 42, 48, 85, 87, 88, 104, 105, 112, 114, 115, 128, 264, 265 – content 42, 48, 114, 128, 264 – modality 112 proposition-making function 56 prospective indexical 193, 197, 202, 205, 216 prototype 153, 154, 159–165, 169, 170 pseudo-oral communication 39 psychological distance 24, 25, 30 punctual 122, 155–158, 160 R rashii 37, 41–43, 46–48, 50–52, 99 realis 111, 120, 123, 126–128 real-world knowledge 262 Received Pronunciation 297 receptor 55, 61, 62, 65, 68, 70–72, 75, 77, 78 recontexualized 296 referent 3, 8, 46, 93, 263–274, 283 regionality 297, 298 regular past 157 relative clause 153, 154, 162, 262 revelation type 58 S self-correction 247, 250, 258 self-organizing network 160 semelfactive 155 sentence-final particles 21, 23, 32, 34, 89, 109, 115, 123–125, 127, 128, 228 sentence focus 70
soliloquy 7, 85, 87–89, 95–98, 101, 102, 296 soo 22, 23, 29, 39, 41–43, 46, 50, 51, 60, 112, 117, 119, 120, 195, 204, 207, 208, 230–232, 244, 248–250, 267, 272, 290, 292, 293 SPD 155–158 speech act verbs 93–95, 97, 98, 100, 101, 107, 108 speech community 299 speech contexts 264, 275 stance/perspective sharing 229, 232, 235, 236, 243 standardization 279, 280, 288, 297, 298, 300, 302, 303 Standard Japanese 8, 83, 176, 279, 280, 293, 297, 302 state 6, 19, 25, 37, 39, 41–43, 45, 46, 48, 49, 58, 59, 64, 67, 68, 75, 77, 78, 80, 82, 91, 93, 94, 99, 101–103, 113–115, 154–156, 158, 159, 164, 173, 185, 198, 202, 230, 247, 262, 268, 300, 302 stative 40, 107, 154, 156, 158, 159, 162, 164, 170, 283, 285, 292, 293 stimulus 55, 61–65, 67–75, 77, 78, 176–178, 191 style 45, 57, 82, 106, 121, 169, 214, 295, 296, 298, 301, 302 style-management 298 subject-predicate sentence 57 subordinate clauses 13–15, 17, 18, 30, 31, 122 T telic 153, 155–158, 160 tense 7, 38, 59, 60, 62, 63, 86, 90, 91, 102, 103, 105, 109–111, 116, 120, 128, 131, 132, 153–158, 160, 162, 164, 165, 167–170, 173, 176, 178, 238, 245 past – 38, 60, 63, 91, 120, 153, 156–158, 160, 164, 173 present – 59, 60, 62, 86, 103 thetic judgment 73–76, 78, 79, 82 transitivity 5, 9, 35, 102, 159, 168 Turkish 153, 155, 156, 158, 167, 170 typology 83, 109, 111, 116, 129, 131, 132 comparative – 116, 131 linguistic – 111
Usage-based Approaches to Japanese Grammar U unaccusative 161 underdeveloped sentence 58 underextension 157, 160, 165 Universal Grammar 9, 33, 105, 163, 164, 166, 168
usage 1–6, 8, 9, 17, 34, 37–39, 50, 53, 163, 166, 168, 170–172, 175, 176, 178, 179, 181–183, 187–190, 262, 263, 267, 271, 274, 275, 277, 297 – based approach 37, 38, 171 – based model 163, 168, 190
V variant choice 8, 279, 280, 293, 295, 296, 298, 299 variant mixings 287, 288 verbal inflection 120, 122 verbal reflex 64, 65 voicing 7, 137–139, 141, 151, 229, 234, 245
