AUTHOR: Carnie, Andrew TITLE: Syntax SUBTITLE: A Generative Introduction, Second Edition SERIES: Introducing Linguistics PUBLISHER: Blackwell Publishing YEAR: 2006
Lara Reglero, Department of Modern Languages and Linguistics, Florida State University
This book is an introduction to syntactic theory which adopts a Principles and Parameters approach with an occasional eye on Minimalism. The book is intended for an upper division undergraduate introduction to syntax class but it could also be used in more advanced courses. The book is divided into five parts: Preliminaries (Chapters 1-5), The Base (Chapters 6-8), Movement (Chapters 9-12), Advanced Topics (Chapters 13-15) and Alternatives (Chapters 16-17). Being a second edition, the book has been considerably revised and contains quite a few differences with respect to the first edition. Just to mention a few examples, the chapters on X-bar theory have been expanded and now contain more trees and examples, the treatment of passives and locality conditions is different and now there is a new chapter on 'split VPs' and a new section on 'advanced topics'. In what follows I provide detailed summaries of the chapters and a critical evaluation of the book.
Chapter 1 ('Generative Grammar') introduces the theoretical framework for the book: Noam Chomky's Generative Grammar. Within this theory, the author adopts the Principles and Parameters approach and some recent developments from Minimalism. After an overview of some of the main fields in linguistics, the author focuses on the scientific study of syntax. For this, the author describes the main steps of any scientific method (observing data, making generalizations, developing hypotheses or rules and testing the hypotheses/rules against more data) and illustrates the method with anaphors in English. Possible sources of data include grammatical sentences from corpora and more importantly, ungrammatical sentences for which subconscious knowledge of language needs to be employed. The second part of the chapter attempts to uncover our subconscious knowledge of syntactic rules by adopting a perspective from language acquisition. The main claim is that many parts of language are innate, as proposed by Chomsky in his Universal Grammar (UG). Evidence for UG comes from different arguments: the logical problem of language acquisition, the underdetermination of the data, typological arguments (universals of language) and biological arguments (language is human-specific). In order to account for differences across languages, the concept of 'parameter setting' is introduced. The chapter concludes with an emphasis on the scientific method whose ultimate goal is to offer an explanatory adequate grammar.
In Chapter 2 ('Parts of Speech'), the author discusses the parts of speech (or syntactic category) words belong to: nouns (N), verbs (V), adjectives (Adj), adverbs (Adv) and prepositions (P). Given that semantic criteria are not completely reliable in determining parts of speech, the author suggests using other tests such as morphological distribution (derivational and inflectional suffixes) and syntactic distribution. A discussion of open vs. closed parts of speech follows, which is in turn followed by lexical vs. functional parts of speech. The author offers a thorough discussion of functional categories of English such as prepositions, determiners (D), conjunctions (Conj), complementizers (C), tense (T) and negation (Neg). The major parts of speech are divided into subtypes, also known as subcategories (marked through features). Nouns have subcategories such as plural vs. singular, count vs. mass, proper vs. common or pronoun vs. lexical noun. Verbs can be divided into two major subcategories: tense/finiteness and argument structure. As for argument structure, predicates can be classified as intransitive (one obligatory argument), transitive (two obligatory arguments) and ditransitive (three obligatory arguments).
Chapter 3 ('Constituency, Trees, and Rules') introduces the notions of '(hierarchical) structure' (as opposed to linear strings) and 'constituency'. These notions can be captured in tree structures (or bracketed diagrams). A sentence constituent (Tense Phrase (TP)) typically consists of a Noun Phrase (NP) and a Verb Phrase (VP). These phrases in turn contain other elements creating a hierarchical arrangement of the different elements in the structure. Trees are generated by means of phrase structure rules (PSRs) which adopt the following general format: XP -> XYZ (where XP stands for the label of the constituent, -> means 'consists of' and XYZ are the elements that make up the constituent, X being the 'head'). The author discusses the major phrases in detail (i.e. NP, VP, Adjective Phrase (AP), Adverb Phrase (AdvP), Prepositional Phrase (PP) and TP), provides examples, trees and the rules generating the tree structures. The chapter also contains a practical section on how to draw trees from two different perspectives (bottom-up and top-down) and a section on the relationship between syntactic trees and ambiguous sentences. At the end of the chapter, the author discusses constituency tests such as replacement, stand alone/sentence fragment, movement and coordination/conjunction. The chapter ends with an appendix on how to solve foreign language problems with and without word-by-word glosses.
Chapter 4 ('Structural Relations') is a more technical chapter dealing with the geometry of trees. Syntactic trees consist of branches and nodes (root node, terminal nodes and non-terminal nodes). These elements enter into structural relations which include domination, precedence and c-command. Domination can be of two different types: exhaustive and immediate. 'Precedence' can also be immediate. The more complex notion of c-command is also defined, illustrated and divided into symmetric and asymmetric c-command with government being the local version of the term (phrase-government and head-government). Along with structural relations, there are grammatical relations, which can be also defined in structural terms. Examples include the notions of subject, direct object, object of a preposition and indirect object.
Chapter 5 ('Binding Theory') presents the basics of Binding Theory. For this, the terms R-expression, anaphor, and pronoun are illustrated and defined. In order to provide a more comprehensive account of Binding Theory, the author discusses important notions such as antecedent, indices, coindexing, coreference and binding (c-command and coindexing). Then, Principle A (responsible for the distribution of anaphors) is discussed and defined by appealing to a locality constraint: the binding domain. More specifically, an anaphor must be bound in its binding domain. In contrast, a pronoun must be free in its binding domain (Principle B). Finally, R-expressions must be free (Principle C).
Chapter 6 ('X-bar Theory') initially points to some inadequacies of the system, that is, it generates flat structures. By using the one-replacement test in the NP domain, the author shows that we need a more articulated tree with an intermediate level of structure (N' or N-bar) and a corresponding theory: X-bar theory. Evidence for the existence of V', Adj'/ Adv' and P' comes from do-so replacement, so-replacement and conjunction, respectively. In an attempt to simplify and generalize the list of PSRs, the author notes the following similarities: all phrases have obligatory heads, all non-head material is phrasal and optional and each major category contains three rules. In order to capture these similarities, any part of speech is replaced with the variable X. Then, the author introduces some new terminology: complements, adjuncts and specifiers. Complements and adjuncts are compared (especially in the NP and VP domains), differentiated and located in different positions in the tree. Specifiers (roughly, Ds) are also shown to be different from complements and adjuncts. Based on all this, three PSRs are introduced which make use of the variable X (i.e. the X-bar rules): the specifier rule (XP -> (YP) X'; where YP is a specifier), the adjunct rule (X' -> X' (ZP) or X' -> (ZP) X'; where ZP is an adjunct) and the complement rule (X' -> X (WP); where WP is a complement). To account for cross-linguistic variation, X-bar rules are made a bit more general by allowing specifiers and complements to appear on either side of the head. Given these options (i.e. parameters) the child needs to be exposed to the input and choose the right setting of the parameter. The chapter concludes with a step-by-step section on tree drawing by using X-bar theory.
Chapter 7 ('Extending X-bar Theory to Functional Categories') explores the specifier rule and the problem it raises for the theory: YP is phrasal but the D is not. In fact, Ds are heads and, as such, head their own projection: the Determiner Phrase (DP). Following Abney (1987), the proposal is that Ds are not within the NP. Instead, they take NPs as complements. The second goal of this chapter is to incorporate the Complementizer Phrase (CP) and TP rules (CP -> (C) TP and TP -> DP (T) VP), respectively) into X-bar theory. Under the new rules, the CP has C as its head and takes TP as its complement. All clauses are CPs including embedded and main clauses (the evidence for the later is taken from a cross-linguistics comparison of yes/no questions; Cs can be pronounced or null). The TP has T as its head, a subject DP as its specifier and VP as its complement. T can be an auxiliary or an inflectional ending on a verb. Since inflectional suffixes cannot be pronounced in isolation, the proposal is that the affix lowers to attach to the verb by means of an operation called T-affix lowering, whose motivation is morphophonological in nature.
Chapter 8 ('Constraining X-bar Theory: The Lexicon') points to a problem with X-bar theory: it generates ungrammatical sentences. Since complements are optional in this theory, we can incorrectly generate transitive sentences without objects and intransitive sentences with direct objects. Whether a verb requires an object (or not) needs to be encoded in the lexicon along with any relevant selectional (semantic) restrictions a verb may have. Selectional restrictions are encoded in thematic relations (or theta roles) such as agent, experiencer, theme, goal, recipient, source, location, instrument and beneficiary. Theta roles can be used to represent the argument structure of a verb by means of a theta grid. In order to prevent X-bar Theory from overgenerating, the Theta Criterion is introduced (if a sentence violates the Theta Criterion, it will be marked as ungrammatical). In this chapter the author also considers verbs which take theta role-less expletive 'it' as subjects. This particular pronoun appears when there is no other subject in the sentence and its presence is forced by the Extended Projection Principle (EPP).
Chapter 9 ('Head-to-Head Movement') points to another problem raised by X-bar Theory: it undergenerates (it cannot generate some word orders). To solve this problem, we need transformations, that is, operations that take the output of X-bar rules (D(eep)-structure), change the structure, and give as their output a S(urface)-structure. In this chapter, the author examines one general transformation (head-to-head movement) and its different realizations (verb movement, T movement and do-support). For verb movement, the author compares French and English and notes that an adjunct can appear between the head and the complement in French (similar facts obtain with negation). The proposal is that V raises to T in French (verb movement) but T lowers to V in English (affix lowering). This point of cross-linguistic variation is captured by the verb movement parameter. To account for the Verb-Subject-Object (VSO) order in Irish, we need verb movement combined with the VP-internal subject hypothesis. To explain subject-auxiliary/verb inversion in yes/no questions, T -> C movement is postulated. Finally, there is another last resort transformation called do-support which applies in yes/no questions in English which lack an auxiliary. In these cases, a dummy verb 'do' is inserted in T to support an inflectional affix (the same facts obtain in negative sentences). The chapter concludes with a discussion on multiple auxiliaries and affix lowering in English and a brief appendix with tests for determining if a particular language has V -> T or affix lowering.
Chapter 10 ('DP Movement') explores another transformation which accounts for cases where DPs appear in unexpected positions according to theta theory: DP-movement. In this chapter, the author examines raising structures and concludes that DP-subjects receive an agent theta role in D-structure (specifier of VP) and subsequently undergo DP movement to fill the specifier position of the main clause TP. This transformation is motivated by the EPP requirement. Passive sentences can be analyzed in a similar way: the subject DP receives a theme theta role in D-Structure and then undergoes DP movement to the specifier of TP to satisfy the EPP. Based on some additional data from raising structures (expletive insertion does not satisfy the EPP), the author explains that the motivation for DP-movement is Case. Nominative case is assigned in the specifier of finite T and Accusative case is assigned as a sister to the verb. DPs without Case will be ruled out by the Case Filter. To conclude, the author examines raising structures with DPs in non-finite T contexts and the properties of the passive morpheme. The main conclusion is that DPs move from positions where they can't check Case to positions where they can.
In chapter 11 ('Wh-movement'), the author shows that wh-elements undergo movement from a Case position to the specifier of CP. This movement transformation is called wh-movement and is triggered by a [+WH] feature in C. The chapter provides detailed derivations for different types of wh-questions with all transformations indicated (wh-movement, T -> C, do-support, DP-movement). Wh-movement can also occur across clauses with wh-words stopping in each specifier of CP. In other words, wh-movement is successive cyclic. The author further shows that wh-movement is constrained, that is, there are certain contexts (called 'islands') out of which wh-movement is disallowed. Examples of island contexts include the Complex DP Island, the wh-island, subjects (the subject condition) and conjoined structures (the Coordinate Structure Constraint). Island effects (especially the wh-island) are explained under the Minimal Link Condition (MLC) which requires that movement be to the first potential position. The MLC can also account for locality effects with head movement and DP-movement in raising constructions. The author closes the chapter with a discussion on echo questions and shows that these questions do not involve movement and are not subject to the MLC.
Chapter 12 ('A Unified Theory of Movement') explores the possibility of unifying all movement rules into a single rule: Move. Since all movement is motivated by locality, it is proposed that movement is constrained by Full Interpretation. Feature checking needs to occur in a local configuration such as specifier-head, head-complement or head-head configuration. In order to capture cross-linguistic variation, the system needs to be modified. There is only D-structure (and no S-structure), a branching point of SPELLOUT and two interface levels, phonetic form (PF) and logical form (LF). Overt movement takes place between D-structure and SPELLOUT and covert movement between SPELLOUT and LF. Differences between English and Chinese wh-movement are explained under the wh-parameter (English is overt and Chinese is covert). The Verb movement parameter captures the differences between English and French when it comes to verb movement (French is overt and English is covert). The existence of covert movement is independently supported by MLC effects in wh-in-situ languages and quantifier scope in English.
Chapter 13 ('Expanded VPs') examines the problems raised by ditransitive structures (presence of two complements, c-command, adjacency). A possible solution involves a modification of the syntactic structure so that DP1 c-commands DP2. For this, an extra projection is needed (labeled vP), which corresponds to a light verb (CAUSE). Case assignment to DP1 is also discussed. Based on evidence from Irish, the author explains that DP1 receives accusative case in a new functional category: AgrO (Object Agreement). This proposal works for ditransitive verbs which take a PP or a CP as their final argument. Verbs that take two DP complements are more complex. It is proposed that the direct object is introduced by the light verb CAUSE and moves to Spec AgrOP. In the same fashion, the indirect object is introduced by another light verb (LOCATE or POSSESS) and moves to a different functional category for case reasons: AGRIOP.
In chapter 14 ('Raising, Control, and Empty Categories') the author discusses raising (subject-to-subject and subject-to-object) and control (subject and object). Main predicates in raising constructions do not assign an external theta role. The subject of the embedded clause does not have Case so it raises to the empty subject position of the main clause to satisfy Case and the EPP requirement. In contrast, main predicates in control constructions assign an external theta role and do not involve raising. The external theta role of the embedded predicate is assigned to a null pronoun without Case under the name of PRO. The author offers different tests to distinguish raising from control: theta grids of matrix predicates, behavior of idioms and the extraposition construction. A discussion of subject-to-object raising follows. In this case, the subject of the embedded sentence raises to the object position of the matrix verb to receive accusative Case. In object control examples, there is a Caseless PRO and no DP movement for Case. Next, the author considers the nature of PRO and concludes that it is not subject to binding theory but to a different module of the grammar called control theory. The chapter closes with a discussion of 'pro' and the null subject parameter.
Chapter 15 ('Advanced Topics in Binding Theory') starts by summarizing chapter 5 ('Binding Theory'). Then, the author explains that binding principles hold at LF. For this, we need to assume the Copy Theory (Chomsky, 1993) and propose that at least one copy of the chain needs to be subject to the binding principles. After examining numerous data, the author refines the notion of 'binding domain' and reformulates the binding principles as follows. According to Principle A, ''one copy of an anaphor in a chain must be bound within the smallest CP or DP containing it and a potential antecedent'' (p. 429). According to Principle B, ''a pronoun must be free within the smallest CP or DP containing it but not containing a potential antecedent. If no such category is found, the pronoun must be free within the root CP'' (p. 431).
Chapters 16 ('Lexical-Functional Grammar') and 17 (Head-Driven Phrase Structure Grammar') offer a brief overview of two alternative theories of generative grammar. For Lexical-Functional Grammar (LFG), the author discusses the mechanisms of c-structure, functions, the lexicon, F-structure and explains how this theory can account for phenomena such as head-to-head movement/head mobility, passives, raising and control, and wh-movement. For Head-Driven Phrase Structure Grammar (HPSG), the author discusses the notion of features and points to one important difference between HPSG and LFG: HPSG is compositional. The chapter also includes a discussion of rules (the Head Complement Rule and the Head Modifier Rule), binding and long distance dependencies.
This textbook is an excellent introduction to syntactic theory. The author manages to present all the complexities of the Principles and Parameters framework in a student-friendly way. The book is well-written and the discussion and argumentation are very clear and surprisingly accessible to the reader. Each chapter contains several theoretical points, followed by an extremely useful section on ''ideas, rules, and constraints introduced in this chapter'', which provides definitions/short summaries of the main ideas discussed in each chapter. This section is in turn followed by a ''further reading'' section which contains a list of useful references for the reader. Each chapter also contains a problem section which is divided into two parts. The first part corresponds to the ''general problem sets'' and provides numerous exercises for regular students. The second part is the ''challenge problem sets'' and would be more beneficial for more advanced/honors students. Each exercise is labeled according to content and it also indicates the type of skill involved (creative and critical thinking, application of skills, data analysis) and the level of difficulty (basic, intermediate, challenge, extra challenge). Without a doubt, all this information would be very valuable for both instructors and students. Furthermore, the author indicates in the main text when students should be ready to solve the relevant exercise(s). The book also has a website with further resources.
Throughout the book, the author manages to combine theoretical explanations with numerous examples and trees to illustrate and support the theoretical claims under discussion. The author uses extremely good visuals and analogies to facilitate the understanding of complex notions. PS trees are explained in depth and complete derivations are provided for more complex examples. The author also added very useful appendixes on tree drawing, data analysis etc. which make the apparently daunting task of syntactic analysis much more accessible to the reader. Additionally, the author included a number of gray textboxes with more advanced information, clarifications and historical facts, all of which will be of great interest for students.
In terms of topic choice, the author is influenced by previous textbooks such as Cowper (1992), Haegeman (1994) or Radford (1988) but adds two new topics not covered in introductory manuals: LFG and HPSG. This gives students the chance to compare and evaluate different frameworks. The book also has a few chapters with more advanced topics which are more open-ended in nature and show that syntax is an open field where more research needs to be done. One of the big strengths of this book is the emphasis on cross-linguistic data and analysis. All chapters contain examples in English and numerous examples in other languages. The problem sets are often based on languages other than English. This clearly shows to the reader how important it is to conduct research in all possible languages and come up with theories which can account for generalizations across different languages.
Even though the book is very well organized, some of the terms and discussion could have been introduced in more detail or connected more clearly. For example, the author's first illustration of hierarchical structure is unexpectedly too advanced. It could have been more useful to discuss hierarchy in a simpler domain such as the NP and then introduce the more complex TP level (a comparison with a flat structure would have been useful as well). The author could have also considered discussing hierarchy without assigning specific labels (NP, VP etc.) to the different nodes in the tree. This part of the discussion could have been postponed until the discussion of phrases. In general, the connection between phrases, the tree structure(s) and the notion of constituency is not crystal clear. All these crucial notions are discussed in the book but instructors might want to reorganize part of the material and show more precisely how all these notions are related. Many of the trees included in the first chapters of the book contain flat structures (or trees with hierarchical and flat structure combined). This can be confusing to the reader given that the author has argued for hierarchical structures. It would have been useful to include a note explaining that all the flat structures would be revised in later chapters. As for constituency tests, the author could have provided ungrammatical counterparts for all tests. This would illustrate better what constitutes a phrase/constituent and what doesn't. The concept of 'head' could be made a bit more explicit (its obligatory nature). It might have helped introduce in more detail the notion of head in earlier chapters along with the concepts of 'complement', 'adjunct' and 'specifier'.
The author explains island effects under the MLC. The problem is that the account can only be applied to wh-islands. It would have been useful to point to a possible direction to account for other cases too or even mention how previous mechanisms (bounding nodes) explained island effects. In his discussion of echo questions, the author could have limited the discussion to cases with only one wh-phrase in situ. Multiple question contexts can be more confusing given that they can also have a true question interpretation (in fact, the distinction between echo question and true question could have been introduced at the beginning of the discussion). Other points to consider are the following: the division of binding theory into two separate chapters could be reconsidered (but instructors can easily explain the two chapters simultaneously or separately), the addition of a chart for theta roles and finally, the inclusion of a few more recent references by Chomsky for the interested reader.
In sum, this is an excellent book and I recommend it very highly for introductory courses to syntactic theory.
Abney, Steven. 1987. The English Noun Phrase in its Sentential Aspect. Ph.D. dissertation, MIT.
Chomsky, Noam. 1993. A Minimalist program for linguistic theory. In Kenneth L. Hale and Samuel J. Kayser (eds.), The View from Building 20: Essays in Honor of Sylvain Bromberger. Cambridge: MIT Press. Pp. 1-52.
Cowper, Elizabeth. 1992. A Concise Introduction to Syntactic Theory: The Government and Binding Approach. Chicago: Chicago University Press.
Haegeman, Liliane. 1994. Introduction and Binding Theory (2nd edition). Oxford: Blackwell.
Radford, Andrew. 1988. Transformational Grammar: A First Course. Cambridge: Cambridge University Press.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Lara Reglero is Assistant Professor of Spanish and Linguistics in the Department of Modern Languages and Linguistics at Florida State University. She has done research on wh-movement, information-theoretic notions such as topic and focus, ellipsis and adjacency effects.