Simpler Syntax

by Peter W. Culicover, Ray S. Jackendoff

This groundbreaking book offers a new and compelling perspective on the structure of human language. The fundamental issue it addresses is the proper balance between syntax and semantics, between structure and derivation, and between rule systems and lexicon. It argues that the balance struck by mainstream generative grammar is wrong. It puts forward a new basis for syntactic theory, drawing on a wide range of frameworks, and charts new directions for research.
In the past four decades, theories of syntactic structure have become more abstract, and syntactic derivations have become ever more complex. Peter Culicover and Ray Jackendoff trace this development through the history of contemporary syntactic theory, showing how much it has been driven by theory-internal rather than empirical considerations. They develop an alternative that is responsive to linguistic, cognitive, computational, and biological concerns.

Simpler Syntax is addressed to linguists of all persuasions. It will also be of central interest to those concerned with language in psychology, human biology, evolution, computational science, and artificial intelligence.

Genres: Linguistics

608 pages, Paperback

First published June 30, 2005

Reviews

[Gary Bruff · Rating 5 out of 5]

Simpler Syntax is the magnum opus of Peter Culicover and Ray Jackendoff (hereafter C&J), two highly respected leaders of generative grammar. The authors manage to present their heterodox views of syntax in a closely reasoned and non-polemical fashion, all the while suggesting that a post-Chomskyan coalition of non-derivational and constraint-based theories of syntax is both possible and desirable. Although the book is called Simpler Syntax, it is not an easy read. Not by a long shot. The reader should come prepared, at least to the data driven parts of this book, after having read some of the major work in both what C&J call Mainstream Generative Grammar (primarily Governemnt and Binding through Minimalist theory) as well as after exploring non-derivational theories of syntax(which could include Head-Driven Phrase Structure Grammar and Lexical-Functional Grammar among others). Although I have a degree in linguistics, I found several of the more technical arguments in Simpler Syntax to be beyond my abilities. The work does however provide a framework that is ‘simpler’ in comparison to many contemporary theories which, while striving to be minimalist, actually complicate matters unnecessarily. More on that below.

C&J begin their book with an overview on the generativists’ architecture of grammar. Apparently. there had been at one point a general consensus among syntacticians on how languages are ‘designed’. This consensus has faded considerably in light of the more recent Minimalist Program. Minimalism suffers from a lack of empirical coverage while at the same time being taxed with an extensive body of meta-theoretical restrictions. These restrictions seem to invite ‘mainstream’ linguists to delve deeper and deeper into less and less. Simpler Syntax dispenses with some of this meta-theoretical baggage, doing away with: hidden levels of syntax(covert LF movement, AGR and v shells); assumed uniformity of syntactic and semantic representation; obligatory, binary right-branching in phrase markers; uniform theta role alignment; complex and un-learnable derivations, where everything lexical moves into postulated functional heads; and an unrealistic absolute opposition between perfect syntax (Universal Grammar) and an idiosyncratic lexicon(all that is not universal). C&J propose alternatives to mainstream generative grammar’s more recent guiding principles, arguing that syntax divides up the work of bridging sound and meaning by interleaving a flatter (multiple-branching) tree with other (parallel or interface) representations. These other representations include a conceptual structure, a grammatical function tier, and an informational structure. In Simpler Syntax, nothing moves. Constructions are base generated, with relations such as binding or control being articulated through global constraints rather than through traces or other artifacts of movement-based accounts. In fact, there is no hard and fast demarcation between constructions and words. Both constructions and words bear meaning and are represented in the grammar as lying on a single continuum, with idioms and structures licensed by specific lexical items lying somewhere in middle of the spectrum.

The subsequent chapters are far more data based, and I found it tough to plow through many of the details which appear to be intended for experts in syntax. But I should say, having read Chomsky’s Minimalist Program a few months ago, I found Simpler Syntax to be a far superior, and far more explicit, treatment of binding and control. There are in C&J plenty of example sentences as well as a generous presentation of semantic and syntactic representations. In their discussion of the syntax-semantics interface, C&J present a fascinating and highly readable fragment of English structure, one that makes use of theories like Relational Grammar to account for ‘core’ syntactic phenomena like passivization and raising. But unfortunately for the patience of an amateur linguist (that being me), they then go on to explore more abstruse and marginal aspects of English syntax. Easiest of these to understand are their treatments of bare argument ellipsis and gapping, whereby information structure roles like ‘focus’ are employed to account for how both discourse and extra-linguistic context fill in missing structure in the course of everyday speech. So for example, “yeah, with Harriet,” is a complete sentence if it follows an expression like “Ozzie is flirting again.” C&J are at pains to demonstrate how conceptual structure, information structure, and the grammatical function tier (essentially semantic argument structure) provide a more compact and simpler account of how ellipses work. They show that there is no need to postulate underlying complete sentences (which then delete by rule the unexpressed parts of the sentence) in order to create or interpret a meaningful fragment of a sentence. C&J explain a wide range of phenomena--including ellipsis, gapping, sluicing, and sundry types of anaphora of verbs--with their idea of indirect licensing. Indirect licensing allows meaning to emerge through the synthesis of construction syntax and situational or semantic context. Wh-questions, right and left dislocations, and other ‘marked’ word orders are viewed as the product of informational structure mappings onto syntactic structure and not conceived of as derivations from an underlying unmarked order. C&J’s system therefore disallows covert-Wh movement (as has been postulated for Chinese) as well as intermediary traces in the syntax(there are no traces since nothing moves, cyclically or otherwise).

Part 3 explores new answers to some rather worn out questions: those regarding binding and control. In both cases C&J dwell on both core and peripheral phenomena, and in both cases they develop arguments for the explanatory power of semantics-based representations. But I must admit that by the time I got to the “Interaction of the statue rule with binding” I had had just about enough. If we need to read page after page about why “Ringo fell on himself” can only mean that the actual Ringo fell on a statue of Ringo and not vice versa, then I feel we have truly moved beyond the ragged edge of language. The authors’ work on control was just as hair-splitting at times but less absurd in the contexts which their sentences construed. They make a thorough case for how semantic factors like voluntary control and unique control enable syntacticians to do away with PRO and (most instances of) t when discussing control predicates such as promise or persuade.

Part 4 looks at connections between clauses, an area that has not been beaten to death. C&J present many interesting cases (and some not so interesting) where, for example, the seemingly logically universal ‘and’ can have semantic, syntactic, and pragmatic properties which are unknown to or ignored by mainstream generativists. Moreover, clauses can be tied together in a number of ways, including intonation contour (“She listens to the Greatful Dead(rising), she gets depressed(falling)” means something different from “She gets depressed(rising), she listens to the Greatful Dead(falling)”). But alas, there is more binding theory here as well. The more I read about binding theory, the less I understand.

The final chapter of Simpler Syntax returns to the meta-theoretical discussion that the book started with. I found it refreshing to read that there are others who take umbrage with Chomsky’s recent obsession with language as a perfect system that floats through cognitive space like Galilleo’s spheres, removed from and entirely independent of any human concerns of communication or culture. I fully agree moreover with the authors’ advocacy for the scientific comparison of theories of grammar. While requiring linguists to be fluent in several theories, such a development would be a true advance over continued loyalty to an orthodoxy that has abandoned all realism and has remained transfixed by an ideal that might be eternally beyond reach. In my opinion, languages are each beautiful and in their own way perfect, but this does not need to entail that languages are all in some ‘deep’ sense absolutely identical. I disagree therefore with the mainstream’s apparent ‘underlyingly English hypothesis’. Some quotes from the concluding chapter of Simpler Syntax:

“The reason there are lots of languages rather than a single Universal Language is that a human language requires too much information to be carried on the genome.”

“We agree that a Grand Unified Theory would be an amazing scientific achievement. On the other hand, given that language is situated in the brain, itself a product of evolution, we find ourselves willing to back off from a Grand Unified Theory and to acknowledge that language is as riddled with eccentricities as every other part of the biological world, arising through contingencies of evolutionary history.”

As far as my criticism of Simpler Syntax goes, I have some good points and some bad points I would like to single out. On the good side, I am pleased that C&J have dared to incorporate pragmatic/informational and cognitive(conceptual-semantic) aspects to make their model of syntax more realistic, both psychologically and interpersonally. Simpler Syntax does not stick to one theory, nor does it focus on one module or one family of constructions. I commend the authors for their attempt at syntactic holism. I am also glad that the authors appear to invite readers to challenge widely accepted but somewhat arbitrary syntactic universals (to wit: binary branching, UTAH, light verb shells, universal SVO, Move, etc.). I feel that other established figures in linguistics should do more to keep syntacticians in honest conformity with the data, messy as that data may turn out to be. And although I didn’t find it fun to slog though, I acknowledge the authors’ erudition in digesting a vast body of work on binding and control in English.

On the down side, there were a few things which troubled me a little bit about Simpler Syntax. First, the book did not live up to its name in the sense that the phenomena covered here are actually extremely complicated in many cases. I feel the book could nevertheless provide a sound methodical machinery (less derivation, more representation) for approaching other languages, although it would not provide much that will count as Universal Grammar. But maybe it is time to give up on UG entirely. I found some of the authors’ grammaticality judgments odd, if not outright wrong relative to my blue color New Jersey dialect. Since some of their theoretical machinery is built from these judgments, I can’t help but wonder if it might not be impossible to get the whole thing, the whole syntactic kit and caboodle, in light of the increased variation encountered at English’s ragged edge. This holds true a fortiori with languages beyond English. As Huang sees evidence of covert wh-movement in Mandarin, do I believe this to be evidence of Minimalism’s Logical Form, or is Huang just trying to make Chomsky happy? Returning to Simpler Syntax: some of the sentence representations look suspiciously like notational alternatives to derivations rather than constraint-based theoretical innovations. Lines connecting grammatical functions (a hidden level, apparently) to surface constituents appear to be synonymous with lines that move lexical heads from theta to case positions in mainstream syntax. And at one point the authors admit that they need to employ a trace to make their non-derivational model hold up. Though commendably honest, this seems a bit inconsistent for a constraint-based theory that is otherwise very insistent that there is no movement.

But I don’t want to end on a sour note. The book is very good at what it does. If you are interested in alternatives to Gallilean linguistics (Chomsky post 1990), and don’t find yourself interested in the nuts and bolts of the Simpler Syntax theory, then the reader is encouraged to read part I and the last chapter. These chapters depict a new flavor of linguistics that is more attuned to the field’s rhetoric, more concerned with the aesthetics (or lack thereof) of syntactic representation and of theory creation, than most syntacticians would be inclined to accept. But Simpler Syntax does, I feel, come closer than most linguistics to something even more ineffable than good theory—the truth (whereof we can speak) about language.

[Thien]

Learning