The study also highlights the constructs of current linguistic theory, arguing for distinctive features and the notion 'onset' and against some of the claims of Optimality Theory and Usage-based accounts.
The importance of Henk Zeevat's new monograph cannot be overstated. [...] I recommend it to anyone who combines interests in language, logic, and computation [...]. David Beaver, University of Texas at Austin
Date: Fri, 30 Dec 2005 07:51:23 -0800 From: John Fry <email@example.com> Subject: Argument Encoding in Japanese Conversation
AUTHOR: Shimojo, Mitsuaki TITLE: Argument Encoding in Japanese Conversation PUBLISHER: Palgrave Macmillan YEAR: 2005
John Fry, SRI International
Linguists with even the most cursory acquaintance with the Japanese language are likely aware that it encodes grammatical and pragmatic features using postpositional particles, for example '-wa' (topic particle) and '-ga' (subject particle). A vast literature in Japanese linguistics attempts to explain the nuances conveyed by these particles, often by contrasting minimal pairs like 'hi-ga noboru' ('the sun rises') and 'hi-wa noboru' ('the sun, it rises'). The question is often put this way: when does a speaker choose '-ga' instead of '-wa', and vice-versa?
A welcome and insightful addition to this literature is Mitsuaki Shimojo's _Argument Encoding in Japanese Conversation_. Relying on quantitative and qualitative analyses of spontaneous Japanese conversations, Shimojo develops an account of how Japanese speakers encode subject and object arguments using a variety of grammatical strategies, among them wa- and ga-marking. Shimojo explains the speaker's choice of encoding strategies in terms of discourse salience and ''mental processing instructions'' in the tradition of Talmy Givon (1983, 1993).
The book consists of three introductory chapters, followed by five chapters that detail Shimojo's analyses and results.
Chapter 1, the introduction, delineates the scope of the investigation. Shimojo restricts his attention to the core argument roles of subject and direct object; peripheral argument roles such as locatives and indirect objects are ignored. Next, Shimojo identifies six ''encoding types'' to be investigated; that is, six ways that the subject and object arguments can be manifested in actual Japanese utterances. The six encoding types are: (1) ga-marking; (2) o-marking; (3) wa-marking; (4) zero anaphor (i.e., argument ellipsis); (5) zero particle (i.e., particle ellipsis); and (6) post-predicative encoding (postposing).
Chapter 2 is a comprehensive yet engaging review of the previous literature on the six encoding types. Most earlier studies (e.g. Fry 2003) focused on the grammatical or discourse properties of just one encoding type, such as wa-marking or ellipsis, in isolation. By examining all six encoding types within the same conversational corpus, Shimojo hopes to develop a more coherent, integrated account than the ones he reviews in this chapter.
Chapter 3 describes the conversational Japanese data on which the book's analyses are based. Following the methods of Maynard (1989), Shimojo recorded, videotaped, and transcribed the spontaneous conversations of eight pairs of native Japanese speakers. The resulting four hours of conversational data yielded a total of 7909 ''clausal units''. This chapter also describes how subjects and objects were identified within the conversational transcripts. One question this chapter does not answer is why Shimojo went to the trouble of creating and transcribing a brand new set of Japanese conversations in support of his research. Were no existing conversational data appropriate (for example, the transcribed Japanese conversations available from the Linguistic Data Consortium)? Also unclear is why the conversations needed to be videotaped.
Chapters 4 and 5 are concerned with discourse salience. Following Givon (1983), Shimojo adopts two quantitative measures of saliency: Referential Distance (the backwards distance to the most recent coreferential expression, measured in clausal units) and Referential Persistence (encompassing frequency of reference and uninterrupted persistence, again measured in clausal units). These chapters, stuffed full of tables and statistical results, are a bit tedious to read, and perhaps could have been shortened. The upshot is that the Referential Distance measurements show that zero anaphors tend to encode highly salient entities, while ga- and o-marking are associated with newer or less salient information. In terms of Referential Persistence, wa- and o-marking is associated with highly persistent entities, while zero particles and the post-predicative construction are both associated with less persistent information, which suggests that these two encoding types have a defocusing function.
Chapter 6, entitled ''The Six Argument Encoding Types as a System,'' presents a single unified account of argument encoding in Japanese. The most striking feature of Shimojo's system is how he organizes the six encoding types into contrasting pairs. A typical linguistic analysis of Japanese will contrast wa-marking with ga-marking. Shimojo, however, proposes that wa-marking is better contrasted with particle ellipsis, and that ga-marking and o-marking should be contrasted with argument ellipsis. His reasoning runs roughly as follows. First, ga- marking, o-marking, and argument ellipsis all contribute to 'cataphoric focusing' (i.e., maintenance of salience), but they differ in that ellipsis is applied to anaphorically salient (e.g., old) information, while ga- marking and o-marking are applied to nonsalient (e.g., new) information. On the other hand, wa-marking and particle ellipsis can be contrasted in terms of how they are used to specify referents. Particle ellipsis is used for 'absolute' (i.e., non-contrastive) specification of an entity, whereas wa-marking encodes contrastiveness (and thereby helps to maintain salience). Finally, the post-predicative construction stands apart from the other five encoding types; its function, which is to defocus unimportant information, does not contrast directly with any of the others. Shimojo offers dozens of example utterances from his corpus in order to illustrate and support his thesis (which is one reason why this chapter weighs in at almost 100 pages).
The last section of Chapter 6 recasts Shimojo's system into a discourse processing account, whereby each encoding type represents a specific set of ''mental processing instructions for the hearer''. For example, a zero anaphor instructs the hearer to search for a coreferential link and to continue the activation of the referent, whereas a post-predicative construction invites the hearer to deactivate the referent. At a meager seven pages, this is the only section of the book that I wished had been longer. How literally are we to interpret these ''mental processing instructions''? What kind of mental discourse model is presupposed here? Is there a relevant psycholinguistic literature? This section deserves its own chapter, along with expanded background material and discussion.
Chapter 7 elaborates further on the Japanese post-predicative construction. Accounts of this construction traditionally have been production-based (speaker-oriented), whereas Shimojo's account in Chapter 6 is comprehension-based (hearer-oriented). Shimojo defends his approach by identifying flaws in the production-based accounts and showing that his conversational data are better explained in comprehension-based terms.
Final conclusions are offered Chapter 8, where Shimojo further develops and refines his thesis that the argument encoding system serves to guide the hearer's processing of utterances in spontaneous conversation. Shimojo emphasizes that the speaker's choice of encoding type is made not in isolation, but rather in relation with other encoding types within the system.
In sum, Shimojo's book is an impressive example of quantitative and qualitative linguistic analysis of spontaneous conversation. The overall thesis, that six different argument encoding types form a system that guides comprehension, is original, insightful, and empirically defensible. While perhaps overly long, the book is engagingly written, with no apparent typos. Researchers in Japanese linguistics, especially those interested in discourse analysis and processing, will want to read it.
Fry, John (2003) Ellipsis and wa-marking in Japanese conversation. New York: Routledge.
Givon, Talmy (1983) Topic Continuity in Discourse: a quantitative cross-language study. Amsterdam: John Benjamins.
Givon, Talmy (1993) ''Coherence in Text, Coherence in Mind.'' Pragmatics & Cognition, 1:171-227.
Maynard, Senko K. (1989) Japanese conversation: self- contextualization through structure and interaction management. Norwood, NJ: Ablex.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
John Fry is a Research Linguist at SRI International in Menlo Park, CA, USA.