Date: Sun, 21 Aug 2005 16:18:18 +0400 From: Anna Grashchenkova Subject: A Minimalist Approach to Scrambling
AUTHOR: Karimi, Simin TITLE: A Minimalist Approach to Scrambling SUBTITLE: Evidence from Persian SERIES: Studies in Generative Grammar 76 PUBLISHER: Mouton de Gruyter YEAR: 2005
Anna Grashchenkova, Department of Theoretical and Applied Linguistics, Moscow State University, Moscow, Russia.
The book under review (as stated in the introductory chapter) follows three interrelated goals. The first one is to provide in-depth analysis of syntactic structure of Persian, and primarily to account for freedom of constituent order. The second is to relate inferences from Persian data to that provided by other languages in which scrambling occurs. The third one, as stated by the author is "to situate the results ... within the framework of the Minimalist Program, specifically phase theory"
The book consists of seven chapters.
Chapter 1: Introduction. The first chapter serves an introduction to the monograph. It provides a brief overview of Persian syntax and summarizes the basic theoretical assumptions, underlying the proposed investigation.
Providing empirical reasons Simin Karimi rejects derivational approach to word order, advanced by Kayne (1994), and follows traditional parametric approach, considering Persian a SOV language underlyingly. She argues nevertheless that Tense Phrase (TP) is head-initial in this language. This assumption is based primarily on the placement of sentential arguments of the verb, which follow the verb obligatory. Moreover, she states, that all functional heads appear in an initial position in this language.
Analysis advanced in the volume is based on phase theory. Following Chomsky (2001), Karimi distinguishes two strong phases: lexical phase vP and operator/discourse phase CP. The relevance of such a configuration is discussed in the following chapters. Along with basic theoretical assumptions of Minimalist Program (henceforth MP) in section 4 she introduces the main ideas underlying the theory of Distributed Morphology (henceforth DM), to which help she resorts in chapter 2, when analyzing (non)specific object marking.
Some empirical data are also introduced in this chapter. It includes the description of Persian basic word order, rather detailed analysis of complex predicates. In section 3 a descriptive discussion of elements that undergo scrambling is offered. At the end of the chapter Karimi introduces the notion of Specificity that is extremely relevant to the analysis advanced in the volume. She argues that specificity is responsible for case marking of direct objects, as well as for other different properties of specific and nonspecific noun phrases in object and certain subject positions. The detailed discussion of the phenomenon is offered in chapter 3.
In section 5 the outline of the content of the monograph is provided.
Chapter 2: Literature on Scrambling. As stated in the introductory lines, this chapter is devoted to a review of literature on scrambling in the last two decades. The two major approaches to scrambling are discussed, namely base-generation approach (henceforth BGA) and movement approach (henceforth MA).
BGA within the minimalist framework is associated primarily with Bošcović and Takahashi (1998) paper on Japanese. These authors suggest that "... scrambled elements are directly base-generated in their surface positions and undergo LF movement (lowering in most cases) to the positions where they receive theta roles". Saito and Fukui (1998) argue within the same lines, suggesting that scrambling is an optional operation, and thus should be treated as a special case of Merge. Both papers maintain thesis that scrambling is semantically vacuous.
In section 2.3. Karimi analyses Persian data and provides counter evidence with respect to the outlined theories. First of all she argues that scrambling is not semantically vacuous: a scrambled element in Persian may be interpreted as topic or focus based on its stress, moreover scrambling provides scope ambiguity even in case of long- distance scrambling (henceforth LDS), and thus the copy of the scrambled element plays a role in the interpretation of a sentence. The most vulnerable point in Bošcović and Takahashi (1998) paper is that only arguments are subject to scrambling. Karimi shows that adjuncts in Persian may also undergo LSD and create ambiguity as well (for similar evidence from Russian see for example Bailyn (2001)). The third argument against BGA to scrambling has to do with extraction out of islands. That is, if one assumes that scrambling is the result of Merge, than the impossibility of occurrence of scrambled elements within syntactic islands (attested in Persian as well as in other scrambling languages) cannot be accounted for.
Karimi further reviews the literature on scrambling from a movement point of view. She presents a brief survey of clause-bound scrambling into Case position, revealing properties of A(rgument)-movement (such as raising and passive constructions), such as locality, anti- Weak Crossover (WCO) effects, FQ (Floating Quantifiers), binding relations and lack of reconstruction.
The discussion in this chapter also suggests that scrambling exhibits some characteristics of A'- movement (such as wh-movement), such as licensing a parasitic gap and reconstruction. Then attested instances of atypical A and A' movement are discussed. Karimi argues for example, following May (1977, 1985), that reconstruction sometimes is possible from an A-position. She also provides evidence of Persian LDS that allow FQ, and other instances of movement that exhibit properties of A as well as A'-movement
She also sketches Webelhuth's (1992) proposal (which suggests that the landing site of scrambled elements exhibits mixed properties) and its criticism. She concludes discussion of MA pointing out that such an approach in terms of A-A' distinction as well as BGA also faces some problems.
In the final section of this chapter the author examines properties of scrambling languages and their differences from non-scrambling ones. She reviews several approaches (Fukui (1993), Müller and Sternefeld (1993), Bošcović and Takahashi (1998)), designed to account for this problem. She suggests that previously proposed factors (such as adjunction sites) cannot be parameterized to account for these differences. Following the most recent trends in scrambling theory, Karimi advances a hypothesis that scrambling is a feature-driven movement, and thus is not an optional syntactic operation. Under this assumption all differences between scrambling and non-scrambling languages are reduced "to the choice of selecting a certain type of feature from the lexicon".
Chapter 3: Local Scrambling and A-movement. This chapter is devoted to local scrambling in Persian and its clause structure in general.
Section 2 provides an in-depth analysis of Persian subjects. Discussing different types of construction (including unaccusatives, so- called passives as well as raising, subjectless, ECM (exceptional case marking) and 'tough' constructions) Karimi comes to the conclusion that subject in Persian is base-generated within the vP phase. She argues that the theme of unaccusatives and so-called passives are merged inside the PredP (complement of v). Based on position of prepositional phrases and vP adverbials with respect to subjects in these constructions, Karimi suggests an asymmetry between specific and non-specific subjects (similar to that of objects). She further argues that only specific subjects move out of PredP to Spec of vP to receive specific interpretation. Case and Agreement are also checked in this position. Non specific subjects are argued to stay in-situ, in PredP domain that is neutral with respect to Case and Agreement (this assumption is supported by necessary evidence). Subjects of transitive verbs are claimed to be merged directly in the Spec of vP (this is supported by the fact that agents are obligatorily [+ Specific]).
Additional evidence for the vP-internal hypothesis comes from absence of expletives (as argued by Karimi overt as well as covert). Karimi claims that "no element needs to appear in the vP external position in a situation when the entire propositional phrase is focused". Based on this assumption she further argues that EPP in Persian, similar to Nom Case is satisfied by the rich morphological inflection on Persian verb. This assumption in its turn calls in question the existence of pro and PRO. Karimi argues that their existence is justified by two independent factors. pro is suggested to be required by Nom Case feature of the verb that must be checked. The presence of PRO is argued to be unrelated to Nom Case and determined by control.
In section 3 the analysis of direct objects is proposed. Similar to subjects, direct objects are also claimed to be generated within the PredP. As well as subjects only specific objects escape the domain of existential closure and move into the lower Spec of vP in order to receive interpretation and to check case features. Their nonspecific counterparts remain in-situ and are not marked by Case, as long as Acc Case feature is locally checked by Agree between the specific object in the lower Spec of vP and v.
In section 4 Karimi advances the hypothesis that Persian is a topic- prominent language. This assumption is based primarily on the fact that Persian does not use grammatical constructions (such as passive) to extract a non-subject topic as subject-prominent languages do. She argues (and discusses it in detail in chapter 4) that elements extracted out of vP are discourse marked. She further argues that T in Persian optionally selects the feature [+Topic], and thus Spec of TP is a topic position in this language.
In conclusion Karimi briefly surveys the nature of the landing site of the vP internal scrambling in Persian (Spec of vP) and comes to the conclusion that it cannot be considered a typical A-position. (It is discussed in more detail in chapters 5, 7). And thus local scrambling is not a typical A-movement as has been suggested throughout the literature.
Chapter 4. Operator/Discourse Domain and A'-Scrambling. This chapter addresses syntactic properties of the operator/discourse phase. Karimi proposes the following phrase structure for this domain:
[CP [TopP [FP [TP [T' [...]]]]]]
Discussion in this and previous chapters is based on distinction of two different types of EPP: the first one is based on Chomsky's (1982) original idea regarding the requirement that every sentence have a subject. This type is labeled *grammatical* EPP (EPPg). The other EPP, based on Chomsky (2000), is suggested to be a strong feature responsible for movement of phrasal categories. This type is called *syntactic* EPP (EPPs). In chapter 3 Karimi advanced the idea that EPPg in Persian is satisfied by rich verbal inflection. In this chapter she argues that movement out of vP (lexical phase) is trigged by EPPs, which places the XP in the Spec of FocP or one of the two topic positions. In the case of the former, the XP receives a contrastive focus (as apposed to informational focus that does not involve movement) interpretation. The higher topic position (Spec of TopP) is reserved for the switched topic, and the lower one (Spec of TP) represents the background topic. The claim that all these movements are feature-driven is supported by the fact that the movement of two (for example, scope-bearing) elements belonging to the same category is subject to MLC (Minimal Link Condition (Chomsky 1995)).
Providing evidence of cooccurrence of wh-phrase together with complementizer, Karimi argues that wh-phrases do not undergo movement to Spec of CP, but may move to Spec of FocP to receive contrastive interpretation. She then distinguishes wh-arguments and wh-adjuncts, assuming that the former have a D-head, while the latter are purely quantificational (and thus move to a different position). In order to account for an interrogative interpretation of a sentence in the absence of a wh-phrase in the Spec of CP, Karimi assumes (following (Aoun and Li 1993)) that there is a wh-operator in the Spec of CP and the wh-feature moves to C for a local Agree relation with the wh-operator. It is further argued that the focus feature of non-wh- phrases moves to be adjoined to Foc. These claims are based on the fact that the movement of a Foc feature and wh-feature is blocked when the XP carrying one of them is in the domain of another scope bearing element (such as Negative Polarity Item).
Chapter 5: Scrambling, Scope and Binding. This chapter concentrates on the semantic impact of scrambling with respect to scope and binding relations.
Section 2 is devoted to scope marking. Karimi provides evidence that movement of quantificational elements local as well as long-distant alters their scope. In a similar way scrambling affects scope between a quantified element and a wh-phrase. Scope of adjuncts and interaction of negation with existential and universal quantifiers also contradict the assumption that LDS is subject to radical reconstruction. Analyzing these data Karimi arrives at the conclusion that scope relations are determined in overt syntax and that there is no need for covert XP movement, as suggested by Chomsky (1995).
Section 3 concentrates on scrambling with respect to anaphoric relations. Analyzing Persian data Karimi suggests that scrambling does not affect Principle A, feeds but does not bleed Principles B and C of the Binding Theory. However some typological data, presented in this chapter, evidence that at least some languages allow scrambling to feed Principle A. Karimi suggests that this possibility is subject to parametric differences between languages.
Karimi's analysis of binding relations (as well as that of scope interaction) contradicts the traditional assumption that scrambled elements are radically reconstructed.
Finally Karimi argues that movement theory based on phase cannot account for some cases of binding relations (such as relations between pronominal and its antecedent, when the latter is in the higher clause, or Principle C violations). She concludes assuming that "binding interpretations are not established by derivation, but rather by representation". This issue is discussed in more detail in chapter 7.
Chapter 6: Long Distance Scrambling and Island Constraints. In this chapter Karimi examines the differences between LDS and typical instances of operator movement such as structural wh-movement and topicalization.
In section 2 Karimi shows that LDS is not subject to the type of constraint that blocks wh-movement. LDS in Persian can dislocate several elements out of CP, without rendering a sentence ungrammatical. This contrasts with structural wh-movement in English and German, where wh-phrases move cyclically through the same intermediate position, namely Spec of CP. Karimi argues that in case of LDS when more than one element represents contrastive focus, they occupy multiple specifiers of FocP.
In section 3 Karimi provides a set of novel LDS data. She shows that an element cannot scramble into a higher clause if another element with the same grammatical function already exists in that clause or in the intermediate clause. Karimi formulates this condition as follows:
(1) Condition on LDS LSD is blocked in the configuration: *[phase YP1α XPα ....... [t1]], where α represents a specific grammatical function (e.g. subject)
She shows that this condition does not hold in case of a typical operator movement such as wh-movement in English. Karimi further argues that this condition cannot be accounted for on the basis of processing theories. Two processing strategies are introduced and rejected in section 4.
In section 5 additional data are provided in defense of (1). Karimi argues that (1) is not a restriction on A-movement into the argument position, based on facts that pro in the target clause does not block movement of the embedded subject, and that LDS of adjuncts is also subject to (1). She further suggests that this condition (when restated in terms of features) accounts for operator movement as well. The only distinction is the domain of application of this constraint: while CP is the island domain for operator movement, vP serves as the island domain for LDS. Karimi further argues, that since this condition accounts for both A and A' movements, it represents a problem for the typology of movement. This issue is discussed in more detail in the final chapter.
In section 6 Karimi discusses the instances of left-dislocation and ATB (Across the Board) and argues as in the previous chapter that interpretation cannot be solely based on cyclic derivations within a phase, but need to be applied representationally in certain cases.
Chapter 7: Theoretical Consequences. This chapter concentrates on the theoretical outcome of the analysis advanced in this work.
Section 2 reviews the typology of movement. Karimi examines three types of syntactic properties that have been assumed throughout the literature to be diagnostic for A/A' distinction. These are reconstruction (suggested to be exclusively the property of A'-movement), Anti-WCO effects and FQ (both considered to be possible only when A- movement is involved). The data provided in this section show, however that there is no clear cut distinction between A and A' movement based on these factors.
In section 3 the improper interaction of movement is considered. Karimi discusses impossibility of interaction of wh-movement with scrambling and topicalization in English and German (languages that exhibit structural movements for both wh-phrases and topic). She then presents two solutions previously offered to account for this problem: Epstein's (1992) approach, based on Economy of Derivation and Müller and Sternefeld's (1993, 1996) account based on Principle of Unambiguous Binding (PUB). She further proposes an alternative account, suggesting that the observed restrictions on interaction of different types of movement is determined by the position where XP receives interpretation, rather than by the typology of movement. That is if XP moves into a Spec of a functional head, where it receives its interpretation, it cannot move into a new position to receive a different interpretation. Based on this assumption the following constraint is suggested:
(2) Constraint on Interpretation (CI): Within the functional domain, If XP receives interpretation in α it cannot be interpreted in β.
Karimi's approach allows not to stipulate an LF movement to account for the problematic data. Section 4 concentrates on the interaction of phase theory with interpretation that was briefly discussed in chapter 5. She discusses several cases that pose problem to phase theory with respect to interpretation. Karimi argues that co-reference of an embedded pronoun with the left-dislocated DP cannot be accounted for within a purely derivational model based on phase. Additional problems are provided by ATB cases and resumptive pronouns in island constructions, where movement is not an option. Based on this data Karimi suggests "that UG (Universal Grammar) must allow interpretation based on representation, in addition to derivation, if phase theory is to be maintained".
Karimi concludes the final chapter with a brief discussion of unresolved issues, left for further research.
It is perfectly obvious that Karimi's book represents an important contribution to scrambling as well as to syntactic theory in general. Analysis advanced in this work addresses the main problems that scrambling poses to the syntactic theory. It is argued in this volume that scrambling is triggered by EPPs, has a semantic effect on the output of the derivation and is not syntactically optional. Moreover a set of novel interesting data, not previously attested in scrambling languages, is provided.
Certain imperfection of the work that pretends to postulate universal patterns seems to be the lack of typological data. Some generalizations (such as the [+Topic] feature of T), made by the author, call for a more representative sample of languages. It should be mentioned in conclusion that the book is highly reader-friendly. Each chapter opens with an introductory section and concludes with a brief summary of the most important issues. And the whole book is structured in the same way.
Aoun, Joseph and Yen-hui Audrey Li (1993) Wh-elements in-situ: Syntax of LF? Linguistic Inquiry 24 (2): 199-238.
Bailyn, John (2001) On scrambling: A reply to Bošcović and Takahashi. Linguistic Inquiry 32(4): 635-657.
Bošcović, Željko and Daiko Takahashi (1998) Scrambling and last resort. Linguistic Inquiry 29 (2): 347-366.
Chomsky, Noam (1982) Some Concepts and Consequences of the Theory of Government and Binding. Cambridge/London: The MIT Press.
Chomsky, Noam (1995) The Minimalist Program. Cambridge/London: The MIT Press.
Chomsky, Noam (2000) Minimalist inquiries: The framework. In Step by Step: Essays on minimalist syntax in honor of Howard Lasnik, R. Martin, D. Michaela, and J. Uriagereka (eds.), 89-155. Cambridge/London: The MIT Press.
Chomsky, Noam (2001) Derivation by phase. In Ken Hale: A life in language, M. Kenstowicz (ed.), 1-52. Cambridge/London: The MIT Press.
Epstein, David (1992) Derivational constraints on A'-chain formation. Linguistic Inquiry 23 (2): 235-259.
Fukui, Naoki (1993) Parameters and optionality. Linguistic Inquiry 24 (3): 399-420.
Kayne, Richard (1994) The Antisymmetry of Syntax. Cambridge/London: The MIT Press.
May, Robert (1977) The Grammar of Quantification, Ph. D. Diss., Department of Linguistics and Philosophy, Massachusetts Institute of Technology. MIT. [Published 1990, New York: Garland]
May, Robert (1985) Logical Form: Its Structure and Derivation. Cambridge/London: The MIT Press.
Müller, Gereon and Wolfgang Sternefeld (1993) Improper movement and unambiguous binding. Linguistic Inquiry 24 (3): 461-507.
Müller, Gereon and Wolfgang Sternefeld (1996) A'-chain formation and economy of derivation. Linguistic Inquiry 27 (3): 480-511.
Saito, Mamoru and Naoki Fukui (1998) Order in phrase structure and movement. Linguistic Inquiry 29 (3):439-474.
Webelhuth, Gert (1992) Principles and Parameters of Syntactic Saturation. New York/Oxford: Oxford University Press.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Anna Grashchenkova has defended her diploma work at the Moscow State University, Russia. Her graduate work deals with reflexive binding within the Adjective Phrase. Her research interests include syntax, syntactic theories and typology.