AUTHOR: Lebeaux, David TITLE: Where Does Binding Theory Apply? SERIES TITLE: Linguistic Inquiry Monographs PUBLISHER: MIT Press YEAR: 2009
Yosuke Sato, Department of English Language and Literature, National University of Singapore
David Lebeaux (henceforth, L) is widely known over the last 20 years for his extremely influential ideas on Binding Theory, phrase structure and language acquisition. The volume under review is a long-awaited monograph in the field that articulates and extends all these ideas with substantive conceptual and empirical arguments. L's central proposal here is that Principles B and C of Binding Theory (henceforth, BT (B) / BT (C)) are negative conditions that apply homogeneously throughout the syntactic derivation to filter out all illicit syntactic configurations whereas Principle A (henceforth, BT (A)) is a positive condition that applies solely at Logical Form (LF) to give a coherent semantic interpretation for the output syntactic representation. When combined with the unexpected lack of otherwise prevalent BT (C) effects down A-chains, this proposal leads to radical reconceptualization of the standard phrase structure theory, Case and Theta systems and the lexical vs. functional split. Specifically, L proposes that Theta and Case modules constitute two independent representations, Theta Subtree and Case Frame: the former consists solely of open-class items while the latter consists of closed-class Case-assigning and Case-receiving items with little pro's in all argument positions. The two representations are later fused into a single structure after A-movement has taken place, a position that L supports with ample evidence from child language acquisition, speech errors and passivizability of V+ DP idioms, among other arguments. These central proposals are all summarized in a succinct fashion in the Preface of the monograph.
The present volume is composed of nine relatively short chapters, followed by Notes, References and Index, with each chapter making specific arguments on which its successive chapter builds. Chapter 1 (''Introduction'') lays out the central proposal of the volume: BT (B) and BT (C) apply continuously throughout the derivation, unlike BT (A), which applies solely at LF. Chapter 2 (''Reconstruction Down A-Chains, and the Single Tree Condition) provides a variety of arguments from trapping effects, embedded anaphors, quantificational binding, the PRO-gate phenomenon and double binding in co-predicational structures in favor of A-reconstruction. Chapter 3 (''More on the Single Tree Condition'') argues on the basis of trapping effects for the Single Tree Condition: A) an element must be interpreted consistently in one position of the chain and B) all positive interpretive conditions, including idiom interpretation, BT(A), Quantifier interpretation, and pronominal binding must be met uniformly at LF. Having established the reality of A-reconstruction and the Single Tree Condition, Chapter 4 (''Condition C Applies Everywhere'') convincingly shows, based on binding effects in A-chains, as manifested in (1a, b), that both BT (B) and BT (C) apply uniformly throughout the derivation.
(1)a. * He seems to John's mother t to be expected t to win. (p. 23, L's (1a)) (1)b. * John believed him to be expected t to win. (p. 27, L's (9))
Chapters 5-8 discuss a problem for the homogenous theory of BT (C) posed by the lack of its otherwise prevalent effects under A-chains and presents a solution to it by radical reorganization of the phrase structure system into Theta Subtree and Case Frame. Chapter 5 (''The Structure of the Reconstruction Data (A Hole in Condition C)'') discusses a curious empirical hole in BT (C). Given the homogeneity version of BT (C), examples (2a, b) are wrongly predicted to be ungrammatical because A-reconstruction should cause a BT (C) violation, on a par with ungrammatical examples (3a, b). Note further that BT (C) effects obtain under A'-chains, as shown in (4). (I do not include indices in the examples in what follows but the binding pattern intended should be clear from the context of discussion.)
(2)a. John seems to himself t to like cheese. (2)b. John's mother seems to him t to be wonderful. (p. 32, L's (10a, b))
(3)a. * It seems to him that John likes cheese. (3)b. * It seems to him that John's mother is wonderful. (p. 32, L's (11a, b))
(4) *Which pictures of John does he like t? (p. 32, L's (vi))
Here, L lays out the basic solution to this problem drawing on the notion of staggered lexical insertion/overlay. More concretely, every nominal structure contains pro's (not necessarily identical to pro's proposed in the generative literature for pro-drop languages like Italian and Spanish), which can be moved by A-movement (but not by A'-movement) and later overlaid by a contentful DP. Under this analysis, (2a), for example, can circumvent the BT (C) violation in the manner seen in (5a-d).
(5)a. e seems to himself pro to like cheese - A-movement → (5)b. pro seems to himself t to like cheese. - Lexical Insertion → (5)c. John seems to himself t to like cheese. - Bind → (5)d. John seems to himself t to like cheese. (p. 36, L's (19a-d))
Chapter 6 (''Two Interesting Constructions'') examines two constructions, illustrated in (6a, b) and (7a, b), that show a tight interaction between quantificational binding, BT (C), the Single Tree Condition, and the late insertion of adjuncts and full DPs.
(6)a. Which paper that he gave to Bresnan did every student think t that she would like t? (With 'he' bound to 'every student'; 'she' coreferent with 'Bresnan')
(6)b.*Which paper that he gave to Bresnan did she think t that every student would like t? (With 'he' bound to 'every student'; 'she' coreferent with 'Bresnan') (p. 44, L's (5a, b))
(7)a. (His mother's) bread seems to every man t to be known by her t to be the best there is. (7)b.?*(His mother's) bread seems to her t to be known by every man t to be the best there is. (p. 47, L's (13a, b))
In (6a), there is a single position (namely, the intermediate trace/copy site) in which the c-command requirement on pronominal binding and the obviation of the BT (C) effect can be met, in conformity with the Single Tree Condition. (6b) is bad since there is no such position. Note that (6a) supports the late insertion of the adjunct relative clause after A'-movement, for otherwise the derivation would immediately trigger the BT (C) violation in the base position. Exactly the same argument can be constructed on the basis of (7a, b) if we assume the late insertion of the full DP after A-movement. That is, a pro is generated in the most deeply embedded position and moves successively cyclically into the highest CP. The insertion of a full DP applies to the intermediate trace/copy position (after 'every man'). This insertion then correctly derives the result that the BT (C) effect is obviated in (7a). (7b) is bad because this late insertion option cannot evade this effect: the pronominal binding requirement forces the DP 'his mother' to be overlaid in the intermediate trace/copy position (after 'every man'), thereby triggering the BT (C) violation. Chapter 7 (''The Architecture of the Derivation'') proposes a principled explanation for the optional late insertion of full DPs in A-chains through separation of the phrase structure as commonly conceived of into two separate tiers: Theta Subtree and Case Frame. The former is a pure representation of thematic relations consisting solely of open-class elements whereas the latter is a pure representation of Case, including both Case assigners and Case receivers such as (the accusative Case-assigning property of) V, P, and D. The two representations are later fused into a single tree by an operation dubbed Project-α. The crucial idea here is that A-movement takes place on the Case frame. This proposal directly explains the lack of BT (C) effects in (2a, b) because this tier contains only pro's, not full DPs. The presence of such effects in (4) also falls out if lexical insertion/overlay occurs prior to A'-movement. In this chapter, L adduces a wide variety of arguments for the two separate tiers from telegraphic speech, syntactic speech errors, the Case-sensitive nature of passivization, and the correlation between the free determiner distribution of V + NP idioms and their passivizability. Chapter 8 (''Another Negative Condition'') suggests that the Stray Affix Filter of Lasnik (1981) is also a negative condition that applies throughout the syntactic derivation. This suggestion, L argues, can correctly derive the observation that stems move upward to affixes rather than affixes moving downward to stems or their affixal selectors. Chapter 9 (''Conclusion'') summarizes high points of the monograph.
L's argumentation is admirably clear throughout the monograph, though it strikes me as repetitive at times. A-reconstruction exists. Hence, (1a, b) show that BT (B) and BT (C) must apply homogenously throughout the derivation, in contrast to the minimalist conjecture (Chomsky 1995) that all interpretive conditions apply solely at LF. Now, if this much is established, the hole in BT (C), illustrated in (2a, b), becomes a problem. This problem is solved by late insertion of lexical material in A-chains, not in A'-chains. This solution, in turn, necessitates the division of phrase structure into Case Frame and Theta Subtree, a division that receives substantial independent support. This division, therefore, calls for a major rethinking of the standard generative architecture of the grammar. L's argumentation is reminiscent of that made by Riemsdijk and Williams (1981) in favor of NP-Structure as an independent level of representation at which certain grammatical conditions and generalizations cluster together.
L's global theory of BT (B) and BT (C) raises a number of interesting issues for future research on Binding Theory and interpretive principles, more generally. First, where does the asymmetry between BT (A) and BT (B) / BT (C) come from? L suggests that positive conditions apply to yield a single coherent interpretation; negative conditions, on the other hand, are free from this constraint and hence can apply globally. Although this characterization of interpretive conditions makes a lot of intuitive sense and receives ample empirical support, it is still not quite clear why this dichotomy comes about in the grammar. Second, how can the global nature of BT (B) / BT (C) be stated in a principled fashion within the context of the Minimalist Program? Among major conceptually parsimonious guidelines of the Minimalist Program is the Inclusiveness Condition, namely, that the output of syntax cannot contain anything beyond rearrangement of the input lexical material. L's suggestion that any violation of the negative binding conditions is recorded through a star that cannot be removed clearly violates this condition. A major remaining issue, then, is how to reconcile the global nature of interpretive conditions within a contemporary approach to syntax as in recent Phase Theory (Chomsky 2004). This is an important point because resolution of this issue in one way or another should necessitate an even more dramatic rethinking of the whole grammatical architecture than L suggests in the present monograph.
There is every reason to believe that this monograph is going to be among the standard references for any future investigation of Binding Theory both in terms of empirical scope and theoretical significance. As an additional bonus, it also shines as an excellent example of modern linguistic argumentation: how detailed empirical paradigms affect the overall architecture of grammar and vice versa. I highly recommend this monograph to all syntacticians, graduate students and advanced undergraduate students that are interested in Binding Theory and the theory of grammar.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press. Chomsky, Noam. 2004. Beyond explanatory adequacy. In Structures and Beyond: The Cartography of Syntactic Structures, Volume 3, ed. by Adriana Belletti, 104-131. Oxford: Oxford University Press. Lasnik, Howard. 1981. Restricting the theory of transformations. In Explanation in Linguistics, ed. by Norbert Hornstein and David Lightfoot, 152-173. London: Longmans. [Reprinted in Lasnik (1990), Essays on Restrictiveness and Learnability. Dordrecht: Kluwer.] Riemsdijk, Henk van and Edwin Williams. 1981. NP-structure. The Linguistic Review 1: 171-217.
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Yosuke Sato received a PhD in Linguistics from the University of Arizona.
After serving as a Postdoctoral fellow at the University of British
Columbia, he has joined the Department of English Language and Literature
at the National University of Singapore/NUS. His specialties are in syntax
and its interface with morphology, semantics and phonology within the
framework of Generative Grammar. His work thus far has focused on the
grammar of Japanese, Indonesian, Javanese, and (Singapore) English, and his
thinking on linguistic interfaces is put together in his book 'Minimalist
Interfaces: Evidence from Indonesian and Javanese' (2010), John Benjamins.
His most recent interests revolve around issues within the Biolinguistic
Program (e.g. the design, development and evolution of the language
faculty, precursors to recursion, animal communication, game-theoretic
perspectives on cross-linguistic variation, evolutionary biology). He
currently teaches syntax, semantics, pragmatics and the lexicon of English