AUTHORS: Epstein, Samuel David; Seely, T. Daniel TITLE: Derivations in Minimalism SERIES: Cambridge Studies in Linguistics, 111 PUBLISHER: Cambridge University Press YEAR: 2006
Shishir Bhattacharja, Institute of Modern Languages, University of Dhaka.
In this book, intended for an advanced readership in Generative syntax, the authors point to the shortcomings of some of the previous approaches that deal with derivation in syntax, and then go on to show how the facts of syntactic relations can be accounted for in their Derivational Model (hereinafter DM), a new theory consonant with the Minimalist program.
The book is divided into five chapters. Chapter 1, ''Orientation and Goals'' contains in fact, a brief description of two different topics discussed in the book: i) Whether the Argument chains (A-chains) constituted of a moved argument and its trace (see (1)) and Extended Projection Principle (EPP) (a configurational requirement, that a clause must have a subject) (see (2)) can be abandoned, and ii) How DM accounts for all the relationships between syntactic objects.
1. [Mary i [was arrested t i]] 2. *[was arrested]
In Chapter 2: ''On the Elimination of A-Chains,'' the authors argue that A-chains are not needed because i) they are not syntactic objects, and ii) are therefore inaccessible to syntax and invisible to Logical Form (LF); iii) the chains are redundant because the syntactic operation constituted of merge and move contains the information encoded in chains. Following Chomsky (2000) the authors claim that a syntactic object is either a lexical item, or a projection of lexical item, or features of syntactic objects already formed, and that nothing but the syntactic objects are visible to and manipulable by the computational system of human language. Thus for example, the chain constituted of the lexical item 'Mary' and its trace in (1) is not a syntactic object because it is neither a lexical item, nor does it represent features of any syntactic object. One of the components of this chain is an intermediate projection I' (e.g. [was arrested t i] in (2)) which is not a projection of the required sort, and therefore not accessible to syntactic operations.
The authors further show that the chains cannot be positionally defined in terms of sisters or mothers as Chomsky (2001a) & Lasnik (2002) presume, or even creatable given the X' invisibility hypothesis developed in Chomsky (1995:242): ''A category that does not project any further is a maximal projection XP, and one that is not a projection at all is a minimal projection Xmin; any other is an X', invisible at the interface and for computation.'' They convincingly show that chain formation (e.g. Mary + its trace) is redundant because the information represented by the trace tail of the chain is already part of the derivation resulting from operations like MERGE and MOVE (e.g. the relation between ARRESTED and the trace of Mary in (1) is expressed by 'Concatenate/Merge theta marker V and DP').
In Chapter-3, ON THE ELIMINATION OF THE EPP, the authors argue that EPP must be abandoned because it is ill-defined and highly redundant with other principles of Universal Grammar (UG) such as Case/Agreement, Theta theory, Locality conditions of UG, etc. They note that although originally EPP was the configurational requirement that clauses must have a subject, later it became a morpho-syntactic requirement of feature checking. In (3) for example, Mary is raised to Specifier (Spec) of Tense Phrase (TP) to satisfy EPP but also to check Case and Agreement features. The authors claim that in all finite clauses, the EPP overlaps with Case/Agreement checking in a specifier-head (Spec-head) configuration, a fact that points to the redundancy of EPP. However, there are few non-redundant cases of EPP in English (e.g. (4)) in which the Determiner Phrase (DP) moves to satisfy TO and THERE is inserted as a pure EPP checker), but the authors adopt the empirical hypothesis that that all such cases make wrong predictions.
3. _ has [Mary left] 4. *there seems a man to be outside.
In Chapter 4: ''More Challenges to the Elimination of the EPP: Some Movement Cases,'' the authors discuss approaches like, i) Lasnik (1999, 2003) who argues in favour of EPP and ii) Boskovic (2002) who argues against EPP but wants nevertheless to retain A-movement because it is a property of the movement itself. In (5) for example, DP moves to Spec, To, not for the purpose of checking EPP but to satisfy the principle of the shortest link, a requirement for movement. However, in the Generative approach in general, no movement can take place unless it is motivated by the need to check some features. Therefore, the question that remains to be answered is the following: what motivates the movement of DP to Spec, Inflexion Phrase (IP)?
For Lasnik, the DP movement to Spec IP in (5) & (6) is motivated by EPP since Spec IP in these examples is not a Case position. The authors recognize that Lasnik's argument for the EPP does present a challenge for DM. However, they also remind the reader that although their analysis is no better without EPP, it is perhaps no worse because, I suppose, EPP has its own problems.
5. A man seems IP[a man to a man be outside] 6. She proved him IP[him to be him guilty]
In Chapter 5 ''Exploring Architecture'', the authors explore certain aspects of DM, particularly their principle hypothesis, that linguistic entities are necessarily evaluated by LF and Phonetic Form (PF) at every point of derivation. They argue, consonant with the minimalist program, that the operations merge and move each take two objects, join them together (as a set) and then project one or the other, hereby creating a label for the resulting object. Hence, if X and Y are merged creating C, then C is necessarily the input to both LF and PF, which interpret as much of C as possible, while C may serve as input to subsequent derivational operations. They describe DM as a phasal process of syntactic derivation, and as is the case with phasal processes in general, a violation at one point does not necessarily crash the whole process. In (7) for example, LIKE and SEED are merged at the first stage creating representation LIKE SEED which is evaluated immediately by both PF and LF. This representation clearly violates the theta criterion/Full interpretation but (7) is not ungrammatical because DM allows violations at the intermediate levels of derivation. Each generated representation (i.e. syntactic object) has its own PF and LF properties and they may or may not be convergent. If a is non-convergent it does not follow that b containing a will also crash. Therefore, single-rule-application derivations (e.g. VP[like seed]) which the authors call unit derivation can be continued, creating bigger derivations.
7. [Birds VP[like] [seed]].
Now, there are two questions to be answered: i) How are Case and Agreement features checked and ii) What is the function of THERE in examples like (8)? In Chomsky (1995), the DP ('a man') covertly checks the Case of T via LF movement to Spec, TP whereas the Probe-goal analysis in Chomsky (2001a & 2001b) postulates that the Case of DP is checked in situ. Under both analyses, THERE is described as a pure EPP checker. But as the authors argue, if one treats THERE as the EPP checker, it will be difficult to explain why (9) in which THERE occupies Spec To position is unacceptable.
8. There will be a man outside (declarative) 9. *It would be likely (for) [there to seem that Fred left] 10. * T will have [a man slept] 11. A man will have a man slept]
The authors reject Chomsky's hypothesis that Case can be checked in situ and/or through asymmetric C-command because with this hypothesis it is difficult to explain why (9) in which the DP does not move, is unacceptable. In order to force the raising of DP in (10) the authors argue that symmetric C-command (sisterhood) is required for Case-checking. If the DP moves to Spec, T, then DP and T C-command each other and the sentence becomes acceptable (11).
12. There seems to be a man outside. 13. It is likely that there is a man outside.
DM also rejects the assumption that THERE is the EPP checker. For DM THERE locally checks only Case. Both (5) and (12) are acceptable because in (5) the DP moves to Spec, T in order to check the Case feature, and in (12), the DP remains in situ and the case is checked by THERE.
Now the question is how does the expletive THERE check the Case feature of T? If THERE first merges into Spec, T, it symmetrically C-commands T but T fails to C-command THERE. As it has been mentioned above, T must C-command THERE in order to check Case under symmetry. The authors argue that this problem can be solved either by assuming that i) the expletive merges at a lower position and then raises to Spec, T or that ii) the expletive merges in Spec T and then moves to the position of Complementizer (C). However, in certain constructions like (13), C is occupied (by THAT) and THERE cannot necessarily move to C. The authors recognize that such examples are problematic for DM.
14. *Will be a man outside (declarative) 15. A man will be outside. 16. *There seems a man to be outside.
Consonant with the minimalist approach, DM allows no non-purposeful movement. But what is new with DM is that it allows no covert movement either. (14) is unacceptable because, the DP ('a man') remains in situ and therefore, Case cannot be checked, although the agreement feature can be checked through probe-goal. (15) is grammatical because the DP moves overtly to check its Case. In (16) the DP has performed an illegal movement to Spec, To because no feature is checked in that position. Both (12) and (13) are acceptable because i) the DP remains in situ, ii) the Agreement feature of the DP is checked through Probe-goal and iii) THERE appears as Case checker.
In this book, the authors have clearly shown why EPP as well as A-chain formation should be eliminated from the Minimalist analysis, and have convincingly argued for adopting DM. The only thing that remains unclear (to me at least) is whether the postulation of an adverb like THERE as a nominative case checker of DP (and in other approaches, an EPP checker) can be justified on historical grounds although I am aware of the fact that such an assumption apparently helps explain certain facts of syntax in particular theoretical models. But overall, Derivations in Minimalism is a well-argumented and groundbreaking book on syntactic derivation within the minimalist framework. Unfortunately, the book does not contain a glossary of technical terms, probably because it is intended for an advanced readership. However, my opinion is, the inclusion of such a glossary would have been helpful for readers.
Boskovic, Zeljko (2002) A-movement and the EPP. Syntax 5,167-218.
Chomsky, Noam (1995) The minimalist program. Cambridge, MA: MIT Press.
_____ (2000) Minimalist inquiries: the framework. In Roger Martin, David Michaels and Juan Uriagereka, eds. Step by Step: Essays in honor of Howard Lasnik, 89-155. Cambridge MA: MIT Press, pp.
_____ (2001a) Derivation by phase. In Michael Kentowicz, ed. Ken Hale: A life in language, 1-55. Cambridge, MA: MIT Press.
_____ (2001b) Beyond explanatory adequacy, ms. MIT. A revised version to appear in Adriana Belletti ed. Structures and beyond: current issues in the theory of language, Oxford: Oxford University Press. Lasnik Howard (1999) Chains of arguments. In Samuel D. Epstein and Norbert Hornstein, eds. Working minimalism, 189-216. Cambridge, MA: MIT press.
_____ (2002) On the Extended Projection Principle, ms. University of Maryland, College Park.
_____ (2003) Minimalist investigation in linguistic theory. London: Routledge.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Shishir Bhattacharja was trained in Linguistics and Indology at the Sorbonne, Paris. Although his main area of research is syntax, he has recently finished a morphological description of Bengali and submitted it as his Ph.D. thesis in the University de Montreal. His most recent book is _Sanjanani Byakoron_ (1998), a collection of articles, written in Bengali, on generative syntax. He has been working at the University of Dhaka as an Assistant Professor of French since 1995.