Date: Thu, 02 May 2002 02:22:27 +0000 From: Viatscheslav Iatsko Subject: Steedman (2002) The Syntactic Process
Steedman, Mark (2002) The Syntactic Process. MIT Press, x+330pp, paperback ISBN 0-262-19420-1, $19.95 (hardback ed., 2000)
Viatcheslav Iatsko, Department of English, Katanov State University of Khakasia.
[The hardback edition of this book was reviewed in http://linguistlist.org/issues/11/11-2540.html --Eds.]
Mark Steedman's new work can be classified as a monograph, in which this well known author attempts, on the one hand, to suggest a new syntactic conception based on the assumption of transparent relationship between syntax and semantics, and, on the other hand, to specify some theses of generative grammar using psychological and computational mechanisms. The book covers topics in formal linguistics, intonational phonology, and computational linguistics and can be of interest to post graduates and researchers engaged in various branches of syntax.
The monograph comprises a preface, three parts, 10 chapters, a "References" section, which is, in fact, an extensive bibliography on the problems of syntactic research, and an author-subject index.
In the preface the author emphasizes his intention to develop a simpler and more explanatory theory of such linguistic phenomena as coordination, extraction, and intonational phrasing paying particular attention to coordination and its interaction with other constructions in a number of languages. This theory Prof. Steedman calls the theory of Combinatory Categorial Grammar (CCG)
In the first (introductory) chapter the author substantiates the need to develop a theory of grammatical operations that allows a unified semantic representation to be built directly from surface forms. At the same time Prof. Steedman points out some faults of derivational conceptions, such as non-isomorphic correlation between the structure implicit in a derivation and the structure implicit in interpretation; inapplicability of derivational conceptions to languages with flexible order of the words; viewing surface structure as a level of representation. Chapter I is followed by the main body of the book, which comprises three parts. The first two parts of the book are concerned with language competence while the last one deals with performance mechanisms and computational issues.
Part I, entitled "Grammar and Information Structure", can be considered an introduction to CCG. It argues that the traditional notion of surface structure can be entirely replaced by a freer notion of surface constituency corresponding to information structure, and that this is the only derivational notion of syntactic structure that is linguistically necessary. Part I comprises chapters 2-5. Chapter 2 "Rules, constituents and fragments" provides some more detailed information for a radical rethinking of the nature of surface structure from coordination, parentheticalization, and intonation. The author describes consequences that follow from assuming a rule-to rule relation between syntax and semantics, such as 1) syntactic rules can only combine or yield constituents, 2) only grammatical entities that have interpretations are constituents, 3) syntactic and semantic rules should have the property of monotonicity. The rest of the chapter deals with crosslinguistic constraints manifested in coordination, parentheticals, and intonation structure.
The main drawback of this chapter is its constant promises to discuss some points later, which the author makes without indicating in what part of the book this discussion can be found. Such references without addresses occur almost on every page: "We will return to the nature of the mapping and its consequences below" (p.12), "...a point to which I return below" (p. 14,) "I will return to the question of the source of these asymmetries" (p.15), "Again I will return to this question" (p.16), "I will show later..." (p.17). In the next edition of the book the author should avoid such stylistic faults.
Chapter 3 "Intuitive basis of Combinatory Categorical Grammars" outlines CCG in terms of simple examples that motivate the individual rule types. It opens with the description of notation system that is needed to specify the kinds of things that a linguistic entity combines with and the kind of thing that results. The notation system adopted by the author is based on that one developed within the scope of Categorial Grammar, that is why the author describes its basic notions and principles outlining similarities between CCG, Lexical Functional Grammar and Government Binding Theory and pointing out differences between CCG, Tree Adjoining Grammar, Generalized Phrase Structure Grammar, and Head Driven Phrase Structure Grammar. He also discusses combinatory rules and combinators (the Bluebird, the Thrush, the Starling) that are needed to build up his theory.
Chapter 4 "Explaining constraints on natural grammar" defines the space of possible CCGs more exactly, and summarizes the ways in which some basic constraints on bounded and unbounded constructions, including some apparently "nonconstituent" coordinations, emerge as inevitable consequences of the theory. It also briefly considers the way in which scope ambiguities for natural language quantifiers can be handled within a strictly monotonic grammatical framework. The author characterizes combinatory rules, described earlier, by the principles of adjacency, consistency, and inheritance and analyzes asymmetries associated with subject extraction and with heavy NP shift constructions. Chapter 5 "Structure and intonation" concludes part I by showing that the level of intonation structure can be directly subsumed under surface syntax as it is viewed by the author. It also shows that the level of information structure is captured in the semantic interpretations that the grammar compositionally assigns to the non standard constituents of surface structure/intonation structure. This chapter covers a wide variety of tunes and informational constituents that provide convincing evidence that the same constraints apply in syntactic and prosodic domains, the latter being reflected in information structure of sentences comprising theme and rheme as its main components. The idea to distinguish between background and focus as components of theme and rheme (pp.106-109) opens up further possibilities of addressing a range of questions in semantics and syntax.
Part II entitled "Coordination and word order" comprises chapters 6-7. It continues the development of the grammatical theory and consists of two connected case studies. The first, in chapter 6 "Cross-serial dependencies in Dutch", concerns cross-serial multiple dependencies characteristic of the verb raising construction in Dutch. This phenomenon is examined in interaction with the earlier account of extraction and coordination. This chapter has some interesting generalizations concerning Germanic main-clause order. Chapter 7 "Gapping and the order of constituents" looks at gapping in English and Dutch, applying a new categorial theory of medial gapping and verb-initial coordination to cover a wide range of constructions and the full range of language universals.
Part III "Computation and performance" comprises chapters 7-10. It turns to issues of computation and human grammatical performance. Chapter 8 "Combinators and grammars" examines questions of expressive and automata-theoretic power and the general nature of the parsing problem for CCGs. Chapter 9 "Processing in context" discusses a specific architecture for a parser, including the role of semantic interpretation. It is at this point that the author substantiates his earlier thesis of a direct relation between grammar and processor. Chapter 10 "The syntactic interface" summarizes the architecture of the theory as a whole, its role in acquisition and performance. And its relation to other theories of grammar.
Generally, Prof. Steedman's monograph is based on two sound ideas 1) to introduce the notion of information structure that can be directly derived from surface structure, 2) to apply at once syntactic, semantic,and phonological analyses so as to create a "combinatory" (in Steedman's terms) theory. Both ideas are in no way new. The idea to derive deep structures directly from surface structures was suggested by me in Iatsko (1998a, 1998b). The idea of a comprehensive, integrational approach to the analysis of linguistic units was formulated and realized in lexicography by Y. Apresian (Apresian, 1986). Later I realized the same idea in the conception of integrational discourse analysis (Iatsko 1998c, Iatsko 2002). Prof. Steedman's contribution belongs to the domain of syntax; it is another manifestation of a revolutionary situation in contemporary linguistics characterized by the competition between a number of derivational and non-derivational conceptions.
The specific feature of this monograph is profound syntactic analysis; it contains much material that can be widely used in further research so as to establish a new paradigm in the field.
REFERENCES Apresian Y. D. (1986). Integrational description of language and a defining dictionary. In: Voprosy yazykoznania. No 2. P.57-70. (In Russian).
Iatsko V. A. (1998a).Reasoning as a type of scientific discourse. Abakan: Katanov State University of Kahakasia Press. (In Russian)
Iatsko V. A. (1998b). Textual deep structure . In: Text, speech, dialogue. Proceedings of the first workshop. Brno: Masaryk University Press.
Iatsko V. A. (1998c) Deep structure of proposition and deep structure of discourse. In: Linguistics in Potsdam. No 4. P. 72-89.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER V. Iatsko is a professor in the Department of English and Head of the Computational Linguistics Laboratory at the Katanov State University of Khakasia located in Abakan, Russia. His research interests include text summarization, text grammar, TEFL, contrastive analysis of English and Russian syntax.