SIL Electronic Book Reviews 2003-010
URL: http://www.sil.org/silebr/silebr2003-010
Anaphora: A cross-linguistic study
Oxford Studies in Typology and Linguistic Theory. Oxford and New York: Oxford University Press, 2000. Pp. 396. hardback $119.00, paper $35.00. ISBN 0198235291 (hardback), 0198235283 (paper).
Reviewed by Steve Nicolle
BTL (East Africa) and SIL International
This is an impressive book which I recommend to all linguists; it will challenge your theoretical assumptions and expand your typological horizons. Huang draws on data from around 550 genetically and structurally diverse languages, but this book is far more than just a typological overview of anaphoric systems (which the introductory chapter provides). The broad thesis of the book is as follows: Purely syntactic and/or semantic analyses of (NP and VP) anaphora are theoretically and empirically inadequate; chapters 2 and 3 are devoted to demonstrating this. Instead, syntactic and semantic analyses need to be supplemented with, and in some languages replaced by, a pragmatic account. Crucially, Huang claims that languages vary along a continuum between pragmatically-oriented and syntactically-oriented languages. Chapter 4 aims to demonstrate that of the current pragmatic theories, Huang’s neo-Gricean approach can best account for the relevant data. This pragmatic account of anaphora is then applied to the analysis of switch reference and discourse anaphora (chapter 5).
In chapter 2, Huang first discusses Chomsky’s binding conditions and shows that these do not hold in various languages; for example, long-distance reflexivisation violates binding condition A: an anaphor is bound in a local domain:
| (1) |
Tuki, Biloa 1991:850
|
Huang then illustrates problems with generative accounts of PRO, Control theory (which is the attempt within the Principles-and-Parameters framework to explain how PRO is assigned an antecedent), and null subjects and null objects. All of the analyses formulated within P&P, Minimalism, and Optimality theory are, he argues, inadequate, since different languages will require different licensing and identifying strategies, some of which are nonsyntactic. Chapter 2 concludes with a discussion of long-distance reflexivization (see example 1) in which Huang lists universal tendencies in addition to an implicational scale for long-distance reflexive complement types and some language-specific properties (pp. 90–100). Again he argues that generative analyses are inadequate to account for the data (pp. 101–130).
Chapter three, which is somewhat inaccurately entitled “Semantic Approaches to Anaphora,” begins by discussing four analyses of VP anaphora, three syntactic and one semantic (pp. 131–156). Analyses of VP anaphora have to account for the possible interpretations of sentences such as (2) in which either Susan adores Mary’s piano teacher or Susan adores Susan’s piano teacher.
| (2) | Mary adores her piano teacher, and Susan does too. |
Huang then returns to the topic of binding and control, demonstrating that various semantic analyses are inadequate.
Huang then describes two related phenomena: logophoricity and long-distance reflexivization. Logophoricity is:
…the phenomenon whereby the “perspective” of an internal protagonist of a sentence or discourse, as opposed to that of the current, external speaker, is being reported by some morphological and/or syntactic means.
Using data from various African languages (pp. 176–189), Huang proposes a number of implicational hierarchies for logophoric pronouns, their NP antecedents, and the predicates or complementizers which license them. He then demonstrates that long-distance reflexives in Chinese, Japanese, and Korean conform to these same hierarchies (pp. 190–199). A unified account of logophoric pronouns in African languages and long-distance reflexives in East Asian languages is therefore desirable, and Huang concludes chapter 3 with a formal analysis of these two anaphoric linking devices within the framework of Discourse Representation Theory (DRT).
Chapter 2, dealing with the inadequacies of syntactic accounts of NP anaphora is very detailed, but still accessible to non-syntax specialists. The argumentation is clear and the exemplification is excellent. Chapter 3 is less coherent and little is gained by introducing a DRT analysis as this framework is not used elsewhere in the book.
The pivot of Huang’s thesis is chapter 4, “Pragmatic Accounts of Anaphora.” According to Huang’s neo-Gricean pragmatic account, binding condition B (a pronominal is free in a local domain) results from a generalized conversational implicature derived from the fact that a reflexive form was not used. If a language has no reflexive form, co-reference is the default implicature when a regular pronoun is used.
This account is for the most part convincing, and the claim that certain types of anaphoric usage depend on the generation of implicatures has the benefit of accounting for apparent violations of putative syntactic constraints (since an implicature is by definition cancellable). However, it fails, in my opinion, to deal satisfactorily with languages like Korean in which either a reflexive or a pronoun can be used for co-reference. Huang (p. 225) notes that the use of the reflexive “indicates some sort of unexpectedness,” but as to why a pronoun can be co-referential when the language has reflexive forms, he simply stipulates that “the grammar allows the unmarked pronoun to be used to encode co-reference.”
A more fundamental problem exists, I believe, with the neo-Gricean approach to implicature. In this framework, meaning can be broadly divided into “what is said” and what is implicated. It seems counter-intuitive to say that reference assignment is the result of identifying an implicature, particularly when, in the Gricean tradition, ”what is said” is defined as the semantically determined aspects of meaning plus disambiguation and reference assignment. Alternative pragmatic approaches which allow pragmatic aspects of what is said should, in theory, be able to provide a more coherent account. One such theory, which Huang discusses, is relevance theory (Sperber and Wilson 1995). Unfortunately, I doubt whether relevance theory as it currently stands is capable of providing an empirically falsifiable account of anaphora resolution (although Matsui 2000 provides a start in her account of bridging reference). This is a challenge that needs to be met.
Chapter 4 concludes with a comparison of syntactically oriented languages (e.g. English) and pragmatically oriented languages (e.g. Chinese, Japanese, Korean). This typology is discussed further in chapter 5, which deals with switch-reference systems and discourse anaphora.
To conclude, Huang’s account may not be perfect, but his command of the topic is impressive and he provides compelling evidence for the proposed typological distinction between syntactic and pragmatic languages. This book sets the standard against which any future account of anaphora, syntactic or pragmatic, will be measured.
References
Biloa, Edmond. 1991. Anaphora and binding. Linguistics 29:845-859.
Matsui, Tomoko. 2000. Bridging and Relevance. Amsterdam/Philadelphia: John Benjamins.
Sperber, Dan and Dierdre Wilson. 1995. Relevance: Communication and cognition, 2nd edition. Oxford: Basil Blackwell.