Syntactic Theory: A Formal Introduction

by Ivan A. Sag, Thomas Wasow

Syntactic A Formal Introduction is unlike any other introductory textbook on the market. Targeting students with strong formal/mathematical skills, but assuming no particular previous background, this book focuses on the development of precisely formulated grammars whose empirical predictions can be directly tested. The book begins with the inadequacy of context-free phrase structure grammars, motivating the introduction of feature structures, types and type constraints as ways of expressing linguistic generalizations. Step by step, students are led to discover a grammar that covers the core areas of English syntax that have been central to syntactic theory in the last quarter century, complementation, control, 'raising constructions', passives, the auxiliary system, and the analysis of long distance dependency constructions. Special attention is given to the treatment of dialect variation, especially with respect to African American Vernacular English, which has been of considerable interest with regard to the educational practice of American school systems.

Genres: Linguistics, Textbooks, Nonfiction

440 pages, Hardcover

First published April 1, 2003

Reviews

[Ushan · Rating 5 out of 5]

This is a textbook of syntax used at Stanford University and at the University of Washington, where the authors teach; it is based on a flavor of generative grammar called Head-Driven Phrase Structure Grammar. Ever since the ancient Greeks, words in a language have been assigned to syntactic categories: nouns, verbs, adjectives and so on; phrases also belong to categories: "my hovercraft" is a noun phrase, "full of eels" is an adjectival phrase, "is full of eels" is a verb phrase, and "my hovercraft is full of eels" is a sentence. This grammar realizes that this is not enough: categories must have internal structure. The types are Boolean (a verb is either auxiliary or not), enumerated (a personal pronoun is first-, second- or third-person), an ordered list (in "Bill Gates gave the library a million bucks", "gave" has a list of two objects, "the library" and "a million bucks"), or a set of key-value pairs where the values are either primitive types or other sets of key-value pairs. Sometimes two elements are required to be the same in Prologesque unification: in French the gender and number of the noun and its modifier adjectives must agree. The grammar also allows transformations that transform valid phrases into other valid phrases; "Bill Gates gave the library a million bucks" becomes "Bill Gates gave a million bucks to the library" becomes "A million bucks were given to the library by Bill Gates". A topic that was new to me is raising and control: English has subject-raising verbs ("Bill Gates seems rich" = "It seems that Bill Gates is rich"), object-raising verbs ("I expect Bill Gates to be rich" = "I expect that Bill Gates is rich"), subject-control verbs ("I want to be rich" = "I want it to be so that I am rich") and object-control verbs ("I persuaded Bill Gates to found a company" = "I persuaded Bill Gates to make it so that Bill Gates founds a company"); their internal structure is such that they transform differently from other verbs. This book gradually builds up a grammar of a significant subset of English to be ever more complicated; it touches on such topics as the expletive it and there ("It is raining"), gaps ("Bill is easy to work with" = "It is easy to work with Bill") and even verbless sentences in African American Vernacular English ("[I voted for Barack Obama because] he black"). There are exercises that mention other languages, including an Australian Aboriginal one and a Salvadorean Indian one, but the emphasis is on English.

[Kirk Lowery · Rating 5 out of 5]

It's the bible of linguistics for computational linguists.

[Othman · Rating 3 out of 5]

the book is not bad, but unless you're a computational linguist or someone whose life has no value whatsoever, why the hell would you wanna waste your time on HPSG? I mean,,, C'mon

[Robert Muller · Rating 5 out of 5]

If you want to be up-to-date on modern linguistics theory, or if you want to explore a modern analysis of the syntax of the English language, this is the book. It brings together a lot of work done in linguistics from the 1980s on. It is also the basis for a good deal of work in natural language processing in computer science and machine learning. As a computer scientist I find it interesting that they cite very little literature on the development of the type model they use (essentially the object oriented model with abstraction and inheritance emphasized), which goes back all the way to Sowa in the 1970s (and obviously has roots going all the way back to Aristotle). I would not use this book to understand the philosophy or history of the approach, it's more of a manual on how to use it. As such, it is clear, concise, and well written, with excellent exercises and summaries. Don't forget to read Chapter 16, the last chapter, which points toward future developments. Spoiler alert: they are basically unifying syntactic theory (Chomsky et al.) with Semiotics (Saussure , Eco, et al.) in a theory of signs. And, as a computer scientist again, the impact of this on the design of the syntactic type structure is huge, making it much more simple and straightforward and an excellent basis for NLP.

[Justin · Rating 5 out of 5]

This is the textbook used to teach syntax courses in head-driven phrase structure grammar (HPSG). It provides a solid introduction to the topic and has a lengthy appendix that provides a useful reference.

Since there isn't much of a plot summary I can give, I'll just let this review "divulge" a little joke on the cover. On the bottom right portion of the cover, there's an image of four people. Although it's hard to make out, this image is supposed to show the three authors (Sag, Wasow, and Bender) being introduced to each other by the publisher. They're all dressed in formal clothes; hence, the subtitle of the book, "A Formal Introduction". This is a little bit clearer in the first edition of the book, which had this image larger and didn't include Bender as the book was just authored by Sag and Wasow back then. Apparently, the position of the people in this image was supposed to reflect a syntax tree.

[Kira]

I own this now. Yaay!