Like virtually all syntax textbooks written in the past 50 years or so, this one by Liliane Haegeman (hereafter H) begins with a discussion of Chomsky’s poverty of the stimulus argument, the notion that universal syntactic principles must be available to the child learning to speak. These universal principles, together with certain still rather mysterious language-specific parameters which license variation in or divergence from language’s otherwise inviolable formal template, equip the young learner with all the tools necessary for him or her to generate all of the grammatical sentences of a particular language, and only the grammatical sentences. H’s Introduction to Government and Binding Theory incorporates a voluminous syntactic literature in order to furnish the technical tools and metalinguistic vocabulary necessary for anyone to analyze syntax in a way that mirrors how children learn (activate?) their abstract and baroque rule systems. For Chomsky, the child applies universal principles and sets language-specific parameters en route to the achievement of a fully productive yet almost entirely subconscious knowledge of his or her grammar.
After this introduction, H’s book takes the reader through nearly 700 pages of formal syntactic representations and derivations of constructions, and not much else. The book proceeds from a mature Government and Binding theory up to the early days of the Minimalist Program. Even with a work of this length, H had to make significant choices over what to include and what to exclude from the mainstream generative canon. There are no alternative discussions of monostratal or non-derivational architectures of the human language faculty. Indeed, one of the criticisms I have against this book is that H abides a little too closely to the orthodoxy of Saint Noam. But Chomsky is still king of syntax for a reason. His well-reasoned arguments along with those of many of his followers are very often insightful, if not revelatory. Still, H’s presentation of the historical development and progressive evolution of the field of mainstream syntax implies that the schematic career of generative grammar and the publication history of Saint Noam are virtually isomorphic. Be that as it may, H generally makes wise and tasteful decisions about what to include in her book and what to pass over in silence.
But whatever your training in linguistics, you the reader will come away from this book with a renewed conviction that syntactic derivations entailing locality of government, cyclical transformations, and rigid constraints on representations at all levels have as their basis a kernel of psychological reality. The derivations of sentences presented by H are generally very satisfying as formal explanations of syntactic facts. And it helps that H provides a generous assortment of trees to go along with the less readable but far more numerous bracket notations.
The rationalist backbone of syntactic theory supports a logical system whereby the forms which linguistic data may take are constrained by meta-theoretical rules (binary branching, specifier-head agreement, strict locality, governed traces, etc.). Such meta-rules unfortunately tend to dictate whether a syntactic rule or representation is more natural or universal than another. In syntax, the theory regrettably derives the data at least as much as the data feeds back into the theory. In effect, the assumptions of Universal Grammar constitute a sort of bias about what is and is not a possible construction or grammar. What I am calling bias can go both ways. Some aspects of English syntax are now understood differently in light of research based on often radically ‘other’ languages. For example the split INFL hypothesis requires that agreement and tense be separate heads even though they are typically fused in English into a single head. I believe it is well worth our while to have our impressions of English colored by other languages, especially when the other languages are unambiguously distinct from English. Other languages can be inflectional or non-configurational, for example. In other cases, however, languages are seemingly forced to conform to universals that look a hell of a lot like English. For example, there are Chinese linguists who maintain that Chinese is exactly like English in moving question words to the beginning of the clause, except that in Chinese you don’t hear it. This ‘covert’ movement occurs (after the string has been sent to the phonology) in a quasi-mythical semantic interface called logical form. Another example of UG bias is deriving all basic word orders (somewhat spuriously) from a universally underlying subject-verb-object ordering.
I have a further criticism of this book which boils down to its overly modular approach to syntax. In the best of all possible worlds, H would have gone into the morpho-phonological and semantic-pragmatic aspects of using sentences in natural language. Instead, we are left with the onerous burden of explaining syntactic facts only in terms of other syntactic facts. While this promotes the academic autonomy of the field of linguistics, it amounts to bad science. Without anthropology, cognitive science, history, comparative/historical philology, sociology, and analytical semantics and pragmatics, linguistics would never have gotten off the ground. And I feel that the field of syntax now casts a much broader net than it did in 1994 when the second edition of Introduction to Government and Binding Theory was published. The final chapter of H’s book is on relativized minimality. Relativised minimality, which formalizes divergent patterns of local government within and between languages, has opened the door to other kinds of linguistic relativism. Maybe not Sapir-Whorf relativism, but linguists are certainly venturing further afield a la Boas. The idea that different languages are genuinely distinct is no longer considered the absurdity that it once was. And the left periphery (beginning of the sentence) these days hosts an array of positions for topicalized or focused elements projected from the functional categories in a particular language’s lexicon.
When writing over 25 years ago, H found it acceptable to look almost entirely at English, with some brief encounters with other, mostly European, languages. And even when looking just at English, H’s grammaticality judgments often did not match my own, especially with subjacency violations. I also have strong reservations against accepting the ontology of empty categories without qualification. H insists that empty categories are not really empty but are feature bundles copied and transmitted in local steps up the tree. But if you can’t hear the empty category... (examples are like PRO in ‘I tried PRO to leave’ where the subject of the infinitive is in an ungoverned and so silent position that receives its reference from the matrix subject ‘I’ which commands it; or take ‘Who(m) did you see t?’ where the person-object variable ‘whom’ is moved to the specifier of the +WH complementizer node that in turn holds the tense morpheme ‘did’; or take the passive ‘the ball was kicked’ which comes from ‘e was kicked the ball’, where ‘ball’ is theta-marked by the verb as theme before it moves to ‘e’ to receive structural case and fulfill the requirement that English sentences have surface subjects)... and if you can’t hear traces inherited from deep structure, and if you can’t hear logical form reconstruction with its silent, covert movement, then aren’t deep structure and logical form really like global and parallel representations of the lowly surface structure?
But I can’t blame a book for being a quarter century in theoretical arrears. As I noted, H’s book forms a thorough and still necessary framework for linguists who wish to get their feet wet whilst learning to demystify the arcane practice of syntactic analysis. I’ve read maybe half a dozen GB/PP/Minimalist textbooks, but for clarity of exposition, breadth of coverage, and succinctness of explanation, I would recommend this work over many others. H gives a clear sense of the wherefores and whys of syntax. And best of all, H’s writing lends an excitement and enthusiasm to what can be some often rather dry lines of inquiry. I see H’s textbook as occupying a crucial niche in time between the golden age of English-centric syntax and our current age with its massively increased appreciation of grammatical diversity, manifested by an openness to ‘new’ functional projections and ‘novel’ constructions. These projections, both from the lexicon and from the theory, demonstrate how different languages, while still being economical and constrained in their own ways, are nevertheless part of a much larger syntactic world that we are only just now beginning to explore.
I downloaded H’s book from the internet. There were a few pages missing, but otherwise the text was in good shape, no underlining or marginalia. Another freebie out there is Minimalist Syntax Revisited on Radford’s site. For a more modern and more expensive look at minimalism, see Carnie’s syntax textbook and workbook, which go into more recent topics like v shells and verb initial languages. For primary sources, I would check out Chomsky (especially Aspects of the Theory of Syntax (1965) and Remarks on Nominalization (1970)). I am currently working through the phenomenal Annotated Syntax Reader, 35 articles introduced and edited down by Kayne et al. H’s book might be regarded as a pre- or co-requisite for the annotated reader.