Date: Tue, 9 Sep 2003 11:22:37 -0400 From: Michael Barrie Subject: Minimalist Investigations in Linguistic Theory
Lasnik, Howard. (2003) Minimalist Investigations in Linguistic Theory. Routledge Leading Linguists. Routledge.
Reviewed by Michael Barrie, Department of Linguistics, University of Toronto
The silver embossed series title "Routledge Leading Linguistics" on the cover is merely a typo (one of the very few) for the correct series title "Routledge Leading Linguists."
This volume is a collection of nine articles published by Howard Lasnik between 1995 and 1999. Each article tackles a particular aspect of the Minimalist Program (MP), as it was developed from the early to late 1990's. Lasnik approaches a particular problem or set of problems with MP from both an empirical and theoretical perspective in each article. In his introduction, which is actually chapter one, Lasnik credits Chomsky for spurring his early interest in linguistics and for demonstrating to him how to formulate a linguistically viable argument. The entire introduction briefly traces the history of the development of Chomskian generative syntax, which mirrors Lasnik's personal growth as a linguist alongside the growth of the field. The discussion on Lasnik's calling to linguistics will definitely touch a chord with any generative linguist who reads this book. Below, I give a brief summary of each article, followed by some critical notes. The best overview of each chapter, though, is provided by Lasnik himself at the beginning of each chapter. Lasnik's summaries, composed presumably for this specific collection of his articles, provide a historical setting for each chapter. This is a much welcome feature of the volume, given the time frame during which the papers were written. The reader might otherwise encounter difficulty in flipping among the various theoretical instantiations of MP found in Chomsky (1993, 1994, 1995).
Chapter 2: Patterns of verb raising with auxiliary "be"
This paper is a write-up of a talk Lasnik presented at the 1995 University of Massachusetts, Amherst, Conference on African-American English. Lasnik addresses the problem of so-called affix-hopping and auxiliary raising in English. The central idea is that Infl can be either a bundle of features, which overtly attracts a verb, or an affix, which lowers to the main verb. The feature bundle in Infl attracts either "have" or "be", which then can raise to the head of CP in questions. If Infl is an affix, it lowers to the main verb, which cannot raise to the head of CP in questions, rather do-insertion must take place. Evidence from African-American English is presented, including the null copula and the habitual "be" form.
Chapter 3: Last Resort and Attract F
This paper discusses restrictions on movement within MP. Lasnik challenges Chomsky's economy constraint Greed and suggests that Enlightened Self-Interest is a more adequate formulation of this economy constraint. Lasnik considers feature movement in existentials, scope phenomena and antecedent-contained deletion (ACD) to support his arguments. In addition to dispensing with Greed in favour of Enlightened Self-Interest, Lasnik also concludes that the direct object in exceptional Case-marking (ECM) environments raises overtly to the specifier of AgroP, and consequently that the verb raises to a still higher position.
Chapter 4: Levels of Representation and the Elements of Anaphora
This chapter discusses what levels of representation impinge on the evaluation of binding theory. Traditional assumptions that binding relations are evaluated at S-structure must be reconsidered in MP, where the levels of D- structure and S-structure are dispensed with. Lasnik starts by reviewing early evidence that Condition C must hold at S-structure and contradictory evidence as to what level of representation Condition A holds at. He then questions the interpretation of this data in preparation for reformulating Binding Theory without reference to S-structure. The crucial evidence involves movement and binding in existentials versus ECM. It is assumed that the associate raises to the expletive at LF; however, the binding facts suggest this is not the case. The binding facts indicate that the object of ECM does, however, raise at LF. Lasnik discusses these facts and some prior solutions to this problem. Unfortunately, it is difficult to discuss the relevant data here without providing the theoretical context in which Lasnik's discussion is framed, which would require far more discussion than is appropriate for a review. As in the previous paper, Lasnik concludes, based on discussion by Koizumi (1993, 1995) that the subject raises overtly to AgroP in ECM constructions.
Chapter 5: Pseudogapping Puzzles
This chapter investigates several properties of pseudogapping in English. An example of pseudogapping is given in (1).
(1) John will select me, and Bill will __ you. (p. 56, ex 1)
Lasnik compares and contrasts two previous approaches to this phenomenon in the literature. The first approach, which he discounts, assumes that the overt post-verbal material undergoes heavy NP shift, then remnant VP deletion takes place. In example (1), "you" undergoes NP shift out of the VP, then the VP, which now contains only the verb "select" is deleted. The approach that Lasnik argues for assumes that the overt post-verbal material undergoes object-shift to the specifier of AgroP before remnant VP deletion. In example (1), "you" undergoes object-shift, then the remnant VP is deleted. The object-shift approach, Lasnik argues, accounts for a wider range of data than the heavy NP shift approach, but forces him to posit a split VP analysis, in which a VP appears on either side of the AgroP.
Chapter 6: On Feature Strength
This chapter evaluates three approaches to feature checking based on the pseudogapping and sluicing data presented in previous chapters. Overt movement is argued to happen because of the presence of strong features, and failure of overt movement will result in the survival of these strong features. Lasnik then discusses the question of what level of representation it is that cannot tolerate the unchecked strong feature, thus causing the derivation to crash. The three options are called the PF crash theory, LF crash theory, and the Virus theory. The PF and LF crash theories assume that the presence of a strong feature at the PF or LF level, respectively causes the derivation to crash. The Virus theory assumes that a strong feature must be checked immediately when it enters the derivation. Lasnik shows that the pseudogapping and sluicing data are consistent with either the PF crash theory or the Virus theory, but that there are theory-internal reasons for preferring the PF crash theory. In doing so, Lasnik draws support for the notion that the strong feature is located on the higher element rather than the lower element. In other words, he argues for an "attract" versus "move" theory of overt movement. At the end of the chapter, Lasnik presents some Bulgarian data that is a challenge to the theory of overt movement presented, and leaves this problem for future research. He notes in the introduction, however, that subsequent work in Boskovic (1999) suggests a solution to this problem.
Chapter 7: A Gap in an Ellipsis Paradigm
In this chapter, Lasnik revisits his analysis of English verbal morphology presented in Chapter 2. It first appeared in the journal Linguistic Analysis in 1997 as a response to two replies to the analysis presented in chapter 2. Lasnik originally presented a hybrid analysis in which lexical verbs enter the derivation uninflected and Infl undergoes "affix-hopping" down to the verb, and auxiliary verbs are "lexicalist" in that they enter the derivation fully- inflected. Lasnik shows that the alternative analyses suffer from empirical and theoretical problems. The gap in the ellipsis paradigm is shown in (2), and is contrasted with a grammatical example of ellipsis in (3):
(2) *John was here, and Mary will, too. (p. 104, ex 5)
(3) John sleeps here and Mary should, too. (p. 104, ex 4b)
Lasnik employs the feature checking apparatus of Chomsky (1995) in an attempt to salvage the alternatives offered to his original hybrid analysis in chapter 2, but in the process destroys the account for the contrast in (2) and (3). Lasnik concludes that we must accept the hybrid analysis, until a better one presents itself.
Chapter 8: On a Scope Reconstruction Paradox
This chapter deals with a well-known paradox in the literature concerning scope reconstruction in subject raising contexts. The relevant examples are given below:
(5) everyone seems [t not to be there yet] (p. 118, ex. 2)
In (4), the quantifier can take scope under clausal negation, but this reading is absent in (5). The paradox, then, is why a lowered reading is available in (4) but not in (5). Lasnik discusses various attempts to explain the scope phenomena in (4) and (5). Quantifier Lowering (QL) and reconstruction are the two mechanisms employed in these discussions. One approach argues that reconstruction cannot take place in A-chains (as in 5), but that QL is available. Another approach argues that there is no difference between QL and reconstruction, and that the paradox is resolved in another fashion. The current account of this paradox starts out by critically analyzing the semantic characterizations of sentences with lowered readings. Lasnik points out that previous descriptions have relied on paraphrases that are not necessarily accurate. He tentatively concludes that the lowered readings may be due to some other semantic factor such as a theme-rheme distinction. In this light, the original assertion that A-chain reconstruction is unavailable may hold.
Chapter 9: Some Reconstruction Riddles
This chapter deals with various reconstruction riddles along the same lines as chapter 8, and supports the same conclusion that A-movement reconstruction does not exist. The first riddle concerns a complement/adjunct asymmetry that shows up as a Condition C violation. The relevant contrast is given below:
(6) Which report that John-i revised did he-i submit? (p. 126, ex. 1)
(7) *Which report that John-i was incompetent did he-i submit? (p. 126, ex. 4)
Under standard assumptions, the question phrases in (6) and (7) are merged below the co-indexed pronoun, resulting in a Condition C violation. Wh- movement salvages the violation in (6), but not in (7). This rather nice asymmetry is deconstructed by Lasnik's discussion on the grammaticality of sentences as in (7). His claim is that (7) is ungrammatical for other reasons (largely pragmatic) and offers several examples with noun complements as in (7) that are fully grammatical, once the pragmatic effects are filtered out. The second riddle revisits the lack of reconstruction with A-movement. Lasnik points out that the prohibition on reconstruction with A-movement seems stipulatory given Chomsky's (1995) formulation of chain-formation. He presents some further empirical evidence of Chomsky's for the lack of reconstruction in this environment, and proceeds to discuss the evidence. Specifically, Lasnik addresses the A-movement reconstruction paradox again, noting that the actual contrast concerns definite versus indefinite quantifiers. Examining the claim that traces of A-movement are deleted, hence eliminating the possibility for reconstruction, Lasnik uncovers another paradox, namely that traces in theta- positions must be present at LF for Full Interpretation. But these are the very traces that were just argued to be erased or absent at LF to prevent reconstruction. Lasnik does not resolve this final problem.
Chapter 10: Chains of Arguments
The final chapter of this volume deals solely with the absence of reconstruction effects in A-movement as dealt with in the previous two chapters. He begins by examining Chomsky's claim that reconstruction should be barred from A-chains on conceptual grounds, starting with some early arguments against the claim that intermediate traces can be deleted. He points out that some of Chomsky's (1995) arguments against A-chain reconstruction are not straightforward. Lasnik then discusses some empirical data concerning binding and scope. To understand the lack of reconstruction effects with A-movement, Lasnik takes a close look at ECM. He shows, using binding effects and pseudogapping that the embedded subject in ECM constructions raises to the matrix clause. These are the same arguments presented in earlier chapters. Lasnik then discusses scopal properties of ECM subjects and clausal negation and concludes that raising of the ECM subject is obligatory for pronouns and optional for full DPs. He suggests that the optionality of raising might be a reflex of the optionality of the presence of AgroP. Lasnik revisits Quantifier Lowering one last time, and concludes, tentatively, that it does not exist. Finally, discusses some arguments that suggest that A-movement does not leave a trace, hence the unavailability of Quantifier Lowering with A-movement.
The reader who chooses to sit and read this entire volume at once will be prone to bouts of deja vu, as many of the articles revisit the same topic, using the same data, several times. Although it may seem redundant to the reader to read through the volume in one fell swoop, it actually proves to be a rather insightful exercise, since the articles are cleverly ordered chronologically so as to highlight Lasnik's development of the ideas expressed in this volume. Indeed, after finishing the final chapter, the reader has an extremely good sense of the issues at stake, what Lasnik's views are on them, what competing views exist in the literature, and most importantly, how these views were arrived at over time. Most chapters have also been prefaced by recent advances in the topics in question since the original publication of the articles in question. This helps to make the task of reading this volume less of an exercise in examining the historical development of the issues, but educates the reader on the current state of affairs of the matters at hand. In general, I found very few problems with the text. Explanations were clear and well thought-out. As I mentioned above, Lasnik credits Chomsky for his learning how to lay out a linguistic argument. However, the point Lasnik is trying to demonstrate is sometimes obscured by the use of esoteric examples, which are telling of his background in mathematics:
(8) The mathematician made every even number out not to be the sum of two primes (p. 122, ex. 21)
Lasnik states that "the only reading available for  is the implausible one..." The implausibility of this reading may be obvious to mathematician, but not necessarily to a linguist. The problem wouldn't bear mentioning except that examples such as this one appear several times throughout the volume. Also, coreference between nominals is sometimes not made explicit. A particularly troublesome example is found on p. 115, ex. (59), where an index - i is found on a pronoun, but it is not clear what NP it is meant to refer to. The point Lasnik is trying to make crucially depends on the pattern of coreference, here. These problems, however, are minimal, and should not be a deterrent to the potential buyer.
Boskovic, Z. 1999. "On multiple feature checking: multiple Wh-fronting and multiple head movement," in S. D. Epstein and N. Hornstein (eds) Working minimalism, Cambridge: MIT Press.
Chomsky, N. 1993. "A minimalist program for linguistic theory," in K. Hale and S. J. Keyser (eds) The view from building 20: essays in linguistics in honor of Sylavain Bromberger, Cambridge: MIT Press.
Chomsky, N. 1994. "Bare phrase structure," MIT Occasional Papers in Linguistics 5, Cambridge, MIT.
Chomsky, N. 1995. The Minimalist Program. Cambridge: MIT Press.
Koizumi, M. 1993. "Object agreement phrases and the split VP hypothesis," in Papers on Case and Agreement I: MIT Working Papers in Linguistics 18, J. K. Bobaljik and C. Phillips (eds) 99-148. Cambridge: MITWPL, Department of Linguistics and Philosophy, MIT.
Koizumi, M. 1995. "Phrase structure in minimalist syntax," unpublished PhD dissertation, MIT.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER Michael Barrie is a PhD student of linguistics at the University of Toronto. His main research interests are noun-incorporation and clause structure of Iroquoian languages and Romance syntax, particularly the syntax of Portuguese clitics. His MA thesis was on the verb movement and clitic placement in European Portuguese. He is currently working on Control Theory in English and the properties of control in Oneida, an Iroquoian language.