LINGUIST List 13.3172

Tue Dec 3 2002

Review: General Ling: Megerdoomian & Bar-el (2002)

Editor for this issue: Terence Langendoen <terrylinguistlist.org>

What follows is a review or discussion note contributed to our Book Discussion Forum. We expect discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in.

If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for review." Then contact Simin Karimi at siminlinguistlist.org.

Directory

  • Sharbani Banerji, WCCFL 20 (2002)

    Message 1: WCCFL 20 (2002)

     Date: Wed, 27 Nov 2002 15:16:15 +0530
    From: Sharbani Banerji < net. in">kajoli123now-india. net. in>
    Subject: WCCFL 20 (2002)


    Megerdoomian, Karine, and Leora Anne Bar-el, ed. (2002) WCCFL 20: Proceedings of the 20th West Coast Conference on Formal Linguistics. Cascadilla Press, 669pp, paperback ISBN 1-57473-043-6, $40.00.

    Sharbani Banerji, Ghaziabad, India

    INTRODUCTION This volume is an edited collection of 48 papers by different authors, that were part of the 20th West Coast Conference on Formal Linguistics, held at the University of Southern California on February 23-25, 2001. What follows is a description of the conclusions reached in each of the papers, and a very short evaluation of the book as a whole.

    SUMMARY 1. Phases and Interpretability, by David Adger and Gillian Ramchand The authors adopt a classical Chomsky (1977) approach to the construction of the semantics of relatives on the basis of a syntax using two features (Lambda and identifiability-features). This, according to them, captures a rich and complex system across Celtic Languages like Scottish Gaelic, Irish, Welsh, etc. They make a distinction between syntactic dependencies which are triggered by the requirements of uninterpretable features, and which therefore obey locality (here phasal) conditions, and a semantic binding dependency which is sensitive merely to c-command. The operation mediating syntactic dependency is Agree and not Move. The interpretability of a feature is dependent not only on the particular head that it occurs on, but also potentially on other formal features of that head. The two features relevant to Relativization are Lambda and Var: Lambda is interpreted at LF as something which creates a predicate from a proposition, so that a CP containing a Lambda feature will be interpreted as a predicate which abstracts over some variable. This is the syntactic correlate of 'lambda abstraction'. Var is one of a set of identifiability features, which appear on syntactic objects so that they can be semantically identified. The other prime case of identifiability features are phi- features. Var and phi are two complementary ways of identifying pronouns as variables at LF. Standardly (Chomsky 1977) relative clauses are assumed to involve movements of a null operator to the specifier of CP. The interpretation of this operator is essentially that of a 'lambda abstractor', which binds the position of a trace as in (1). (1) Op that the giant saw t The authors propose that the operator is a pro bearing a Var feature which is in its base position when the relative dependency is established. The relative complementizer bears a Lambda feature and agrees with the pro. They assume that 'pro' bears an uninterpretable Lambda feature [uLambda], and that C bears an uninterpretable Var feature [uVar]. This gives: (2) that[Lambda, uVar] the giant saw pro [uLambda, Var] In their approach, movement is driven only by the EPP feature. Since Celtic lacks such a feature, only Agree takes place, but not Move.

    2. Diathesis Alternations and Rule Interaction in the Lexicon, by Ra�l Aranovich and Jeffrey T. Runner Wasow (1977) argues that there are two different types of rules: lexical rules and transformational rules. Whereas adjectival passives are derived lexically, verbal passives are derived transformationally. The authors try to capture Wasow's insight in lexicalist framework like HPSG. They build on the analysis of Sag & Wasow (1999) who argue that there are two types of lexical entries: lexemes and words. They claim that there is a contrast between lexical rules that relate lexemes to lexemes (L-to L rules) and lexical rules that relate words to words (W- to-W rules). L-to-L rules may affect the lexical semantics of lexical items (viz. , changing the number and nature of semantic roles, or the semantic structure of a predicate), while W-to-W rules only affect its argument structure (i. e. , changing the grammatical functions assigned to the semantic roles). The authors use this distinction to explain certain differences between dative shift(DS) and Spray/Load(S/L) alternation. They claim that DS is a W-to-W rule and that S/L alternation is an L-to-L rule. Baker (1997) argues that DS is a transformational rule, and that S/L alternation a lexical rule. They however reject Baker's claim that it reflects a syntax vs lexicon dichotomy. According to them, they are two different lexical rule types.

    3. Conjunction Weakening and Morphological Plurality, by Ron Artstein This paper argues that multiple plurality outside the nominal domain should be taken as evidence that plural morphology on an expression excludes singularities from its denotation, and that cumulative conjunction is an operation that is generally available. The limited occurrence of cumulative conjunction is the result of the strongest meaning hypothesis and syntactic number of conjoined adjective phrases. How come coordinated singular APs form a singular AP, whereas coordinated singular NPs form a plural NP? The difference does not lie in the semantics of coordination, because cumulative coordination is available for plural adjectives. The author examines a characteristic called 'multiple plurality', where the conjunction of two morphologically plural predicates requires a subject whose denotation consists of at least four individuals. This should be taken as evidence that plural expressions only include plural entities in their extension, and as evidence for the existence of cumulative ('non- Boolean') conjunction on predicates. He assumes a structured domain of individuals, where plural objects are of the same type as singular individuals, namely type 'e'. His claim is that expressions that bear plural morphology only include plural objects in their extensions, and thus contrast with expressions that lack number marking.

    4. A [+interpretable] Number Feature on Verbs: Evidence from Squamish Salish, by Leora Bar-el, Peter Jacobs, and Martina Wiltschko The authors propose that in Squamish, the number feature on verbs is [+(int)erpretable]. In Chomsky's (1995) system, [+int] features do not have to be deleted. Since in Squamish, number on verbs is [+int], it does not have to be deleted. This means, number on verbs should receive an interpretation in Squamish. There are two types of number markers on verbs. (i)The verbal plural marker 'wa' pluralizes events and not individuals. (ii)plural on verbs can be marked by the same morpheme as plural on nouns (CVC reduplication). This proposal predicts that subject verb agreement in Squamish should be optional. It is optional in two ways:(I) plural agreement marking is optional (II) a plural marker can indicate the plurality of the subject or object or both. When reduplication attaches to verbs, it yields plural events. This plurality results in readings of habitual, iterative and intensive events.

    5. Integrity: A Syntactic Constraint ON Quantificational Scoping, by Chris Barker The paper proposes that quantifier scope alternations respect syntactic constituency according to the syntactic constituent integrity scoping constraint (Integrity) defined as: (1) If there is a syntactic constituent that contains B and C but not A, then A must take scope over both B and C or neither. The thrust of Integrity is that quantifier scoping sticks much closer to surface constituency than suggested by traditional theories of quantifier scope, thus indicating that we are one step closer to eliminating Logical Form as an indispensable level of grammatical representation. The need to represent the full range of quantifier scopings is one of the leading motivations for theories that posit a level of Logical Form(LF) distinct from surface structure. Allowing LF to differ significantly from surface syntax severely weakens the empirical force of the Principle of Compositionality. So, a theory that does without LF is strongly preferred. If Integrity is correct, and if there are compositional theories that respect integrity without resorting to LF, then recognizing integrity calls into question the need for LF as a distinct level of grammatical representation. If Integrity is an empirically valid theory, it argues in favor of a in- situ theory.

    6. Partial Copying and Emergent Unmarkedness in Igbo Reduplication, by Jill Beckman The Correspondence based Optimality Theory (OT) approach to reduplication developed in McCarthy & Prince (1995) provides an elegant mechanism for capturing a variety of effects in the phonology of reduplication. In this paper, the author examines the reduplication patterns of �n�c� Igbo. In this dialect, high-ranking markedness constraints require the reduplicant vowel to be [+high]. When the base vowel itself is [+high], full copying and featural identity result. However, when the base vowel is non-high, copying fails, resulting in a reduplicant vowel which is determined entirely by markedness constraints. That is, the Igbo reduplication system is captured and predicted by a model such as that of McCarthy and Prince (1995), in which segment-level copying and feature level identity are governed by separate and separately rankable constraints.

    7. The Implications of Rich Agreement: Why Morphology Doesn't Drive Syntax, by Jonathan David Bobaljik This paper offers a critical evaluation of, and an alternative to, the Rich Agreement Hypothesis: (1) 'Rich' agreement is the cause of (overt) verb movement to Infl. Its corollary would be: (2) Loss of 'rich' inflection causes the loss of verb movement. The leading idea is that morphological variation correlates with syntactic variation because the syntax is projected from morphology. In this paper, the author challenges the validity of (1). He gives counter examples from Germanic languages to show that verb movement to Infl can take place in the absence of rich morphology. However, the one-way implication holds. viz. (3) rich inflection -> verb movement to Infl. The correctness of (3) follows from an approach that takes the syntax to feed the morphology. He argues for the 'late insertion' model of the framework of Distributed Morphology. On this model, overt morphological variation can be seen not as a cause but the consequence of syntactic variation. That is, languages vary in the inventory of functional projections that make up the IP domain.

    8. Resumptives as Derivational Residues, by Cedric Boeckx The paper deals with the relation between a resumptive pronoun(RP) and it's antecedent. The author claims that resumption is the result of stranding under A-bar movement. Thus, the derivation would be like: (1) [CP Whi[. . [VP. . [DP ti'[D ti]]]]] Expectedly then, the RP can be found in intermediate landing sites, as witnessed in Hebrew. The consequences of this proposal are the following: (a) The binder and the bindee start off as one constituent and are separated in the course of the derivation by movement (Kayne 2000). (b) Once the role of D projection is clearly defined, it would account for the general absence of adverbial projections. (c) The structure (1) is similar to the structure of D-linked WhPs, and also to clitic doubling structures. (d) The stranding approach converges on Kayne's structure of 'that- relatives'. (e) Wh-elements that can appear in a resumptive configuration turn out to be rich enough to enter into partitive structure. That is, resumption seems to be restricted to 'richer/D-linked WhPs'. Though a structure like (1) violates the general ban on extraction out of left branches, he explains the phenomenon in terms of agreement vs non-agreement effects on extraction, by citing examples from a variety of languages, and proves his thesis convincingly.

    9. Maintaining Contrast in Nxa'amxc�n Reduplication, by Ewa Czaykowska- Higgins and Suzanne Urbanczyk Languages tend to encourage phonological contrasts where semantic contrasts also exist. This paper works on the data from Nxa'amxc�n (Interior Salish)in an Optimality Theoretic analysis of reduplicative patterns, and argues for a new constraint DISTINCT FORM which evaluates the relationship between form and meaning that underlies patterns of contrast. They posit that the segmental difference between two CVC shaped reduplicative morphemes are best explained by a requirement that different functions have different forms. That is, DISTINCT FORM demands that distinct meaning should have distinct forms. Comparison with other such Anti-Identity models are also made.

    10. Wh-Movement as Noun Incorporation in Nuu-Chah-Nulth, by Henry Davis and Naomi Sawai This paper provides a syntactic account of Wh-questions in Nuu-chah- nulth, which they claim is a Wh-movement language, such that overt wh- movement is motivated by a morphosyntactic requirement in which the matrix C has a strong Wh-feature, which must be checked by the Wh- feature on the interrogative element. Though traditionally this movement is analyzed as phrasal movement of the Wh-XP to [Spec CP], they claim that in Nuu-chah-multh, Wh-questions are derived solely by head movement, the result of which is feature checking in Head-Head configuration. They argue that Wh-questions in this language are derived via three separate processes (i)incorporation of the Wh-word into an incorporating predicate (ii)successive-cyclic movement of the Wh-(pred)icate complex to Mood (iii) further raising of the Pred-Mood complex to Comp.

    11. On Inclusive Reference Anaphora: New Perspectives from Hungarian, by Marcel Den Dikken, Anik� Lipt�k, and Zs�fia Zvolenszky Sentences in which an object overlaps in reference with a c-commanding noun phrase in its local domain never involve a reflexive in English, viz. , (1)a-c are all ungrammatical. (1) a. *We saw/voted for/elected myself. b. *I saw/voted for/elected ourselves c. * he saw/voted for/elected ourselves. When the reflexive in (1) is replaced by a pronoun, as in (2), constructions are acceptable with additional semantic factors. (2) a. We saw/voted for/elected me. b. I saw/voted for/elected us c. He saw/voted for/elected us. Both Principle A of Chomsky's (1981) Binding Theory and Condition A of Reinhart & Reuland's (1993) 'reflexivity' approach to binding guarantee the ill formedness of (1). But to explain the data in (2), the paper presents binding facts from Hungarian to show that: (a) languages do not systematically rule out inclusive reference with reflexives: Hungarian allows the counterpart of (1b). (b) On the other hand, Hungarian accepts the counterpart of (2b) with an out-of-the-ordinary type of object agreement only. The paper exploits Hungarian data to make a case against the semantic predicate based approach to Condition B of Reinhart & Reuland (1993), and to argue for a syntactically complex representation of first person plural pronouns, in terms of a comitative like structure. That is, the standard Binding Theory of Chomsky (1981) can accommodate the inclusive reference anaphora facts of Hungarian, and Principle B can make the right predictions for the Hungarian pronoun cases given an independently motivated analysis of first (and second) person plural pronouns in terms of a comitative structure.

    12. On Certain Head-Final Effects in Vietnamese, by Nigel Duffield The Vietnamese modal element 'can' receives three distinct interpretations depending on its clausal position. (a) as a deontic modal in preverbal position (b) as an aspectual element indicating accomplishment or achievement, when it occurs immediately postverbally (c) as an alethic (sometimes epistemic) modal, indicating ability (possibility) in post VP position It is the distribution of the latter that poses a challenge to the proposed universals of phrase structure (Kayne 1995) and modal adverb placement (Cinque 1999). Both previous treatments, one based on Vietnamese data, the other on 'Thai', reconcile Vietnamese data with Universal hierarchies through derivational analyses that have distinct empirical consequences. This paper compares the formal contexts in which the facts of Vietnamese and Thai crucially differ. It shows that to account for the data, both sentential topic analysis and Control analysis are required. The paper also highlights the need for a cautious examination of areally and typologically related languages, like Thai, which show subtle instances of microparametric variations.

    13. On the Semantics of Pronouns and Definite Articles, by Paul Elbourne Though contemporary semantic theory recognizes three distinct uses of third person pronouns, viz. bound-variable, referential and E-type, a unified account of all the three types does not exist. An unified account of the semantics of bound-variable and referential pronouns however exists, in which they are all construed as variables interpreted as variable assignments. E-type pronouns have been treated as something altogether different. But, no language, according to the author, makes any kind of lexical or morphological distinction between referential and bound variable pronouns on one hand, and E-type pronouns on the other. The author shows that it is possible to give an unified treatment of pronouns. He bases his theory on the hypothesis that pronouns have the semantics of definite articles. He shows that pronouns are Fregean definite articles whose NP complement must be phonologically null. If the complement is an NP-deleted common or garden NP, we get E-type anaphora. If it is an index, we get referential or bound variable anaphora. The pronouns and definite articles are similar, but not the same lexical items. Definite articles in English take two arguments, the inner one, an index, and the outer one a common NP.

    14. Paradigmatic Restrictions on Anaphors, by Martin Evaraert The paper discusses how to account for the absence of nominative reflexives in certain languages (Germanic, Romance) as in: (1) a. *Mary(i) thinks that [herself(i) loves Bill] b. *Gianni(i)i vuole che sei/se stesso(i) scriva un libro. Gianni wants that himself writes a book. To exclude nominative/subject anaphors, the author stipulates the incompatibility of anaphors and agreement, the Anaphor Agreement (AA) effect, first proposed in Rizzi (1990), and further developed in Woolford (1999). The alternative approach makes crucial use of the typology of anaphoric elements in the Reflexivity framework Reinhart & Reuland (1993) combined with some basic assumptions of Chomsky (1995). It is claimed that there is no AA effect other than the failure of some anaphors, lacking full specification of phi-features, to enter into a proper agreement relation with the verbal complex. It depends on the verbal inflectional paradigm whether the underspecified or unspecified anaphors can be licensed.

    15. Getting Results: Motion Constructions in Italian and Scottish Gaelic, by Raffaella Folli and Gillian Ramchand The authors compare path of motion accomplishment structures in English, Scottish Gaelic and Italian to argue in favour of (a) the linguistic reality of the logical decomposition of accomplishments and (b) the existence of a cluster of parameters which mediate the ways in which accomplishments can be composed in a particular language. They adopt the neo-Davidsonian semantic representation and also the decompositional view of Hale & Keyser (1993) and others. They are specifically concerned with the way 'telicity' is constructed in natural language. They argue that result interpretations in natural language are always represented semantically by the subevent structure <process, telos>, called the telic pair.

    16. Syncope Induced Metrical Opacity as a Weight Effect, by Mathew Gordon This paper shows that Optimality Theory can handle cases of stress opacity triggered by syncope, given a sufficiently rich set of weight constraints combined, in the case of Central Alaskan Yupik, with hierarchically ordered faithfulness constraints sensitive to morphology and position within a morpheme. Syncope has been shown to involve deletion of the lightest syllables: open syllables containing short vowels in Bedouin Hijazi Arabic and open syllables containing schwa in Central Alskan Yupik. Many weight distinctions which are more clearly evident in languages with relatively rich syllable structures reveal themselves in other languages only when syncope creates the relevant syllable types. These can be captured through a set of weight sensitive stress constraints referring directly to stress rather than metrical feet.

    17. Split Scrambling: Barriers as Inviolable Constraints, by Maria Gouskova In contrast to English, Russian allows movement out of a DP. This kind of movement is called 'split scrambling'. Both Topic and Foci can split scramble. Topics moving to sentence-initial position, and Foci to sentence-final position. Author argues that whereas DP is a barrier in English, it is not so in Russian. The central question of the paper is what determines the status of phrases as barriers cross-linguistically. She claims that barriers are violable universal constraints in an Optimality Theoretic grammar. Barrier constraints from a fixed hierarchy, such that movement out of a DP or PP is cross-linguistically more marked than movement out of a VP. The analysis predicts that no language allows movement out of a DP or PP but bans the same type of movement out of VP. Split scrambling arises from the conflict of two opposing forces: constraints that demand movement on the one hand, and constraints that oppose it on the other. In English, DP BARRIER dominates the Foc/Topic alignment constraints, whereas in Russian, the Alignment constraints force its violations. The barrier constraints are a necessary addition to the set of universal constraints in OT syntax. Adding them makes the right predictions if they are in a partially fixed hierarchy. The barrier typology supports this hierarchy.

    18. Comparative Quantifiers and Plural Predication, by Martin Hackl The question asked is, to what extent the surface similarity of the highlighted expressions reflect a deep/structural similarity? (1) a. John read 'more than three books'. b. John read 'more than half of the books'. c. John read 'more books than papers'. d. John read 'more books than Bill (did/read)' e. John read 'more books than there are planets in the solar system'. The goal of the paper is to present an argument in favor of a uniform analysis of all expressions underlined in (1) as comparative constructions. The argument is based on the observation that comparative quantifiers and amount comparatives impose the same constraints on their environment. Both require that the NP and VP predicates they combine with range over pluralities. The analysis explains the observations as a consequence of the degree function MANY providing the semantic core of both comparative quantifiers and amount comparatives. Comparative determiners like 'more than three' receive a compositional treatment as comparative construction.

    19. Max-Position Drives Iterative Footing, by Nancy Hall The author argues that several prosodic phenomena are better captured if Max-Position constraints, rather than PARSE-sigma are the pressure behind iterative footing. A language with iterative footing builds as many feet per word as possible, while one with non-iterative footing builds only one foot per word, aligned to one of the word edges. The proposal is that PARSE-sigma should be abandoned. The pressure behind iterative footing and a number of foot-related phenomena is better captured by Max-position family of constraints. These constraints favor maximal packing of input material into prominent output positions such as onsets, root-initial syllables, and heads of feet. This approach ties together a number of phenomena involving the interaction of prosody and faithfulness, ascribing to them a single basic motivation. The intuition is that higher prosodic structure exists to give prominence to underlying material. Since underlying material is what distinguishes one lexical item from another, 'prominence' has the function of aiding in the recovery of the lexical entry.

    20. Negation, Focus and Alternative Questions, by Chung-hye Han and Maribel Romero It is observed that negative non-wh-questions with inverted negation do not have an alternative (alt-) question reading. In English, (1) has two possible readings: yes-no(yn) question reading (1a), and alt- question reading (1b). (1) Did John drink coffee or tea? a. 'Is it the case that John drank any of these two things, coffee or tea?' b. 'Which of these two things did John drink: coffee or tea?' In case of negative questions, with non-inverted negation, both readings are available as in (2). (2) Did John not drink coffee or tea? a. Yes, John did not drink coffee or tea (yn reading) b. John did not drink coffee (alt-reading) But in case of inverted negation, as in (3), though the yn-question reading is available, the corresponding alt-question reading is no longer available. (3) Didn't John drink coffee or tea? a. No, John did not drink coffee or tea (yn reading) b. *John did not drink coffee (alt-reading) The same interpretive asymmetry is found cross linguistically in a number of languages. The authors propose that inverted negation in yn- questions contributes focus marking on the polarity, and that the lack of alt-reading results as a by-product of the interaction of polarity focus and the LF syntax of alt-questions. According to the authors, (i) the alt-questions involve ellipsis, (ii) polarity Focus in C0 is always exhaustive, never contrastive, (iii) Focus-marked constituents cannot be deleted. In sum, inverted negation in a yn question contributes an extra focus that cannot be licensed under the alt-reading. The inverted negation contributes verum focus in yn-questions.

    21. Event-Related Adjuncts and the OV/VO Distinction, by Roland Hinterh�lzl The event related adverbs like Time, Place and Manner adverbs occur preverbally in the order T > P > M in OV languages, but in exact mirror image in VO languages as shown in (1): (1) a. C TPM-V OV lgs b. C V-MPT VO lgs Some of their properties are: (i) It is exactly these adjunct types, and not other adverbs, which are correlated with the positioning of arguments (ii) These adverb types are not rigidly ordered, given Cinque's (1999) seminal work on adverbs. (iii) These advPs are realized as PPs or bare NPs. The word order facts are explained as follows. The reason why event related adverbs appear postverbally in a VO-language like English should neither be explained by linking the placement of these adverbs to a directionality parameter (like the head-complement parameter) nor to their presumptive property of being event predicates. Instead, it is argued that this property should be related to the fact that event- related adverbs, in distinction to the other, especially higher adverbs, are realized in their majority as NPs and PPs. As PPs, they were prone to be placed in post verbal position by a peripheral rule in the OE period. Due to converging factors the situation got fixed when the post verbal order of these adverbs was reanalyzed as the result of successive cyclic VP-intraposition.

    22. EPP: Object Shift and Stylistic Fronting in Scandinavian, by Ken Hiraiwa One of the parametric differences between Icelandic and Mainland Scandinavian languages (MSc), Da(nish), Nor(wegian) and Swe(dish)) is the fact that Icelandic allows Object Shift(OS)of weak pronominals as well as full DPs, but MSc allows only the former. That is, the question is: (1) Why is full DP OS illicit in MSc, but licit in Icelandic? The author proposes an articulated theory of EPP, 'A Split EPP/Agree Theory', and its parameterization in Scandinavian, and shows that this provides a principled explanation for the apparent difference in OS between Icelandic and MSc. Another parametric difference in Stylistic Fronting (SF) in Scandinavian also reduces to the Split EPP/Agree parameter. Elaborating the notion of EPP and Agree, the author proposes the following: (2) The Split EPP/Agree Parameter: Satisfaction of EPP on T is (not) contingent on a syntactic operation Agree. Under this theory, languages vary in whether EPP satisfaction is contingent on Agree or not. Then the following parameterization follow: (3) Split EPP/Agree Parameterization on T in Scandinavian a. Icelandic: EPP is not contingent on the operation Agree b MSc: EPP is contingent on the operation Agree. (3) leads to the following feature constitution of T: (4) Probe features on T in Icelandic and MSc a. Icelandic: T[EPP, phiEPP] b. MSc : T[phiEPP]

    23. The Status of Voice in German, by Michael Jessen and Catherine Ringen The purpose of this paper is to present experimental results that bear on the issue of whether German has underlying voiced stops or [spread glottis]([sg]) stops. The voiced stops in German are different from the voiced stops in languages like Spanish, Russian and Hungarian. In spite of this, German is usually claimed to have stops that contrast in voicing just like Russian, Spanish and Hungarian. The experimental data that the authors present support an account of German stops that involve an underlying feature of [sg]. These facts are difficult to account for if the feature [voice] is assumed. To account for the data they have assumed two constraints which refer to [sg] segments: one which requires that obstruents in word final position be [sg] and second which requires that obstruents in codas be [sg].

    24. A (covert)Long Distance Anaphor in English, by Christopher Kennedy and Jeffrey Lidz The focus of this paper is the distribution of strict and sloppy interpretations of reflexive pronouns in comparative stripping. The paper argues that the distribution of strict/sloppy readings in English comparative stripping constructions provides evidence for the existence of long distance anaphor(LDA) in English which occurs only under ellipsis. The long distance anaphor does not show up overtly, because English does not have a morphological realization of the feature combination corresponding to this object. The English LDA may appear in ellipsis however, because it is precisely in ellipsis that morphological instantiation is not required. This indicates that the set of feature structures generated in the syntax may be distinct from those that have corresponding vocabulary items: not every structure can be morphologically realized. That means, in order to identify a particular language's inventory of syntactic objects we must not only look at contexts where we expect them to be pronounced, but also look at contexts where they don't have to be.

    25. On Prerelatives and Appositives, by Cornelia Krause In this paper it is argued that there is a class of prerelatives -- prerelatives with genitive subjects -- that do not exhibit the syntactic properties of appositives nor appositive interpretations even when they modify proper names. Aside from a restrictive interpretation prerelatives on proper names can receive interpretations that are typical for Free Adjuncts. Prerelatives with nominative subjects however, can be appositives. This difference between prerelatives with genitive and nominative subjects can be derived from the fact that the former but not the latter are reduced relatives. According to the Main Clause Hypothesis, appositives must be full clauses. Prerelatives with nominative subjects can be appositives. This is because they, unlike prerelatives with genitive subjects are full clauses.

    26. Wh- and Focus are Not the Same Projection, by Felicia Lee Though Wh- and Focus movement have been argued to target the same projection across a range of languages, this paper argues that contrary to superficial appearances, wh- and Focus do not target the same projection. Instead, they target separate interacting projections: an interrogative projection IntP, which licenses interrogative elements, and FocP, which licenses focussed constituents. The co-occurrence restrictions involving focus and wh-movement are due to the semantics of wh-words: most wh-expressions encode focus, as well as interrogative features. Thus they pass through FocP before landing in IntP, preventing other constituents from appearing in FocP.

    27. A Way to Undo A-Movement, by Vivian Lin In this paper it is argued that A-movement is not subject to Coordinate Structure Constraint (CSC), since CSC does not exist as a constraint on movement operations. However, given the revised understanding of what CSC is -- A-movement is subject to CSC effects, which take the form of forced reconstruction. Specifically, evaluation of component structures requires reconstruction of independently A-moved subjects from surface subject position (Spec T) back to their VP-internal positions -- unless there is a resumptive pronoun in the second conjunct. The resumptive pronoun licenses the appearance of the moved subject high in the clause by allowing it to be incorporated into an operator-variable chain. The author adopts a version of CSC given in Fox (2000). According to him there is no independent Coordinate structure constraint. Instead CSC effects are derived from: (1) a. Extraction out of coordinate structure is possible only when the structure consists of two [or more] independent substructures, each composed of one of the coordinates together with material above it up to the landing site (component structures). b. Grammatical constraints are checked independently in each of the component structures. (Fox 2000: 50).

    28. Intonation, Scope, and Restrictions on Quantifiers, by Luisa Mart� This paper highlights the importance of intonation in semantic interpretation. It discusses the consequences of the interaction between intonation and scope for the domain restriction of quantifiers. The language of study is German. Consider (1) (English words for German) -- the subordinate 'because' clause is focused ('pressure' is the word that bears the actual stress). (1) the half the students is failed [because the dean to me under PRESSURE put has]F 'Half of the students failed because the dean put pressure on me'. Example (1) contains two scope bearing elements 'half of the students' and the 'because clause'. Thus, we can have two readings: (2) a. Half of the students have failed, and the reason for this is that the dean put pressure on me. b. For half of the students, they failed because the dean put pressure on me to fail those particular students. If (1) has a different intonation, scope possibilities change. (3) [THE HALF]T the students is failed, [because the dean to me under PRESSURE put has]F The stress on the quantifier 'half' marks a contrastive Topic. Now, the only available interpretation for (3) is where the subject takes scope over the subordinate clause. It is assumed that quantifiers are restricted by a contextually specified variable C. By default the variable takes the value 'being relevant'. The author argues that there are semantic and pragmatic reasons why the scope possibilities are more limited for (3) than they are for (1). The analysis is couched in a very modular view of grammar, where 'grammatical sentence' means that the sentence is phonologically, syntactically, semantically and pragmatically well formed.

    29. AspP-Shell Structure in VP-Ellipsis and ACD, by Ayumi Matsuo This paper argues that Antecedent Contained Deletion(ACD) is an operation targeting an (outer) AspP rather than a VP; hence, ACDs should not be treated as an instance of VP-deletion. ACDs have been traditionally treated on a par with VP-ellipsis and pseudo-gapping. However, the author provides evidence that ACDs are AspP-deletion (copying). It is also claimed that ACDs have a large number of properties in common with VP-fronting, which also target AspP. To support the claim, differences between ACDs and VP-ellipsis/pseudo- gapping w. r. t aspect parallelism are considered, and what kind of adverbs and modals each construction can occur with. Whereas ACDs and VP-fronting access part of an AspP, VP-ellipsis access something bigger or smaller than AspP; namely TP, AuxP and VP.

    30. Obligatory Scalarity (A Sliding Scale), by Ora Matushansky Scalar predicates can be described as having domains that are partially ordered according to some property that permits grading. (1) Millay is large/beautiful The predicate in (1) is vague in the sense that the degree to which the subject of the predication is large or beautiful is established by the context. This degree is an additional argument of such scalar predicates. The degree variable slot can be filled by an overt measure phrase, as in (2a), be restricted by a comparison class, as in (2b), or can also be quantified over by a comparative operator as in (2c). (2) a. Frank Sandow is 5 feet 10 inches tall. b. Frank Sandow is tall for a man. c. Frank Sandow is taller than Lady Karle. The claim is that semantic scalarity (presence of a degree variable slot) should be distinguished from its syntactic realization(projection of a DegP), and that the latter comes into play as a selectional restriction on the non-propositional complement of 'seem'. The perception verb 'seem', unlike the epistemic modal 'seem', selects a DegP complement. It is shown that semantic scalarity has to be extended to non-adjectival predicates, such as evaluative nouns and PPs denoting mental and emotional states, which also show similar syntactic and semantic behaviour. Additional data involving degree modification can be accounted for if the projection of DegPs is forced by degree movement, which is attested in exclamatives, interrogatives, topicalization, etc.

    31. Quantitative Processes in Trochaic Systems, by Evan Mellander The paper provides a unified account of a number of quantitative processes in trochaic systems, as well as a number of asymmetries between the manifestation of quantitative processes in trochaic and iambic systems. The principle of Head Government (HD-GOV) accounts for the basic foot inventory in (bounded) iambic and trochaic systems, the restriction of trochaic (LL)-creating processes to moraic trochee systems, and the non-attestation of parallel (LL) creating processes in iambic systems. Contrary to recent analysis, the uneven(HL)trochee is assumed to be a well formed phonological structure, and quantitative processes creating such feet are seen as phonological in nature. (1) Head Government (HD-GOV): Dependent elements within a foot must be governed by an adjacent head.

    32. A Union Function for Complex Coordinate Structures, by Michelle J. Moosally and L. Kirk Hagen Unification grammars lack the expressive power required to predict agreement in natural languages when targets are complex and non- isomorphic. Current theories of agreement are unable to provide a coherent theory that handles agreement in both coordinate and non- coordinate structures. This paper argues that theories that rely on principles of feature unification fail to predict the range of possible coordination patterns. The authors propose an alternative theory based on the union of sets of feature values and show how union succeeds where unification fails. This work, with a parser based framework, yields an approach that relies on union of feature matrices; this approach accurately predicts a wide array of complex agreement patterns without compromising intuitions about constituency.

    33. Interpreting Measure DP Adverbials, by Marcin Morzycki This paper deals with weak DP adverbials, or 'measure DP adverbials' as in (1): (1) a. It had been raining an hour. b. Clyde played the ukelele every day. c. Floyd slept the wrong way again. They are simultaneously DPs and adverbial modifiers; prototypical DPs are not modifiers, and prototypical adverbials are not DPs. These adverbials constitute a natural class distinguished not only by quantificational strength, but also by scope restrictions, distribution and an Aktionsart effect. It is shown that certain Aktionsart information is encoded in a feature in verbal functional structure responsible for licensing these adverbials. Independent evidence is adduced from true adverbs. It is suggested that part of the semantic contribution of a modifier is attributed instead to a fixed position in functional structure. Many modifiers for example, change interpretation with structural position in ways other than scope. Thus, the semantics of measure DP adverbials arise in part directly from their position, through the feature in functional structure that licenses them. This feature imposes a homogeneity requirement, occupies a fixed low position in the clause, and is implicated in interpretation of durative adverbs as well.

    34. Indefinites and Frozen Scope in Japanese: Restrictions on QR and Choice Functions, by Kimiko Nakanishi In English, the Direct Object(DO) cannot take scope over Indirect Object(IO) in IO-DO order, whereas in DO-IO order, either object can take scope over the other. Thus, in (1a), the scope is 'frozen', in that only surface scope reading is available. But in (1b), there is no scope-freezing effect. (1) a. The teacher assigned one student every problem. Scope freezing: IO > DO; *DO > IO b. The teacher assigned one problem to every student No scope freezing: IO > DO; DO > IO (Larson 1990) It is claimed that frozen scope is due to the specificity of the indefinite IO in IO-DO. This specificity is semantically encoded as a choice function interpretation in a similar way as 'a certain NP' in English. Why is the IO in IO-DO specific? The current proposal of the author points toward the same direction as the previous studies of pragmatics: the IO in IO-DO is 'non-dominant'(i.e. , generally definite), 'more topicworthy' or 'the possessor of the DO', whereas the IO in DO-IO does not necessarily satisfy these properties.

    35. A Distinctness Condition on Linearization, by Norvin Richards The author proposes a new well-formedness condition on the linearization statements used by Kayne's (1995) LCA. The new restriction on ordering statements will have the effect of making multiple syntactic nodes with the same label impossible to linearize if they are close together in the structure. This captures a number of recalcitrant syntactic phenomena that conform to a general pattern of avoidance of adjacent identical objects. This approach to LCA requires that ordering statements make reference to node types rather than particular tokens of syntactic objects, thereby limiting the class of ordering statements which can actually be used to impose an ordering on terminals.

    36. Licensing and Feature Interaction Processes in Child Language, by Yvan Rose This paper explores two feature interaction processes, consonant harmony and metathesis found in the outputs of Clara, a learner of Qu�bec French. It is argued that the patterns observed in Clara's data are best captured in an analysis based on (a) highly articulated prosodic representations and (b) licensing relationships taking place between the place features and heads of prosodic constituents. It is also shown that previous approaches to consonant harmony would have difficulty accounting for Clara's data, because none of these approaches make reference to prosodic domains. This approach draws a formal distinction concerning the prosodification of non-final and final consonants in French. Only reference to highly articulated prosodic representations enables one to establish such a distinction and predict its effects on licensing.

    37. A contrast to A Trace, by Uli Sauerland The following three structures have been proposed for a movement operation like Quantifier Raising, in the recent literature. (1) A girl danced with every boy a. [every boy]x a girl danced with x (copy + replace) b. [every boy]x a girl danced with [every boy](copy) c. [every boy]x a girl danced with [the(x) boy](copy+modify). Chomsky (1993) argues against the (copy + replace) theory in (1a) on the basis of Condition C data that show that moved material can behave as if it occupies the base position of movement. (1b) has its own problems of interpretation. This paper argues in favour of (1c), proposed by Chomsky (1993) and Fox (1999). That is, copying is followed by a trace modification operation that replaces the determiner of the moved DP with something else. The author assumes that this is an indexed definite determiner. The prediction of this analysis is that, under certain circumstances, a second occurrence of the moving quantifier could be contrasted with the trace of QR receiving narrow Focus on the determiner. e.g. , (2) [every boy]x . . the(x) boy . . [EVERY]F boy It is argued that the plain copy theory does not have a conceptual advantage when the interpretive system is considered. On the other hand, (copy + modify) theory involves minimal change required for the LF structures to be interpretable, and is better than copy+replace theory.

    38. Subquestions and Quantificational Variability Effects, by Yael Sharvit and Sigrid Beck In the following, (1a-c), have a Quantificational Variability (QV) reading, where the adverb ('for the most part', 'partly', 'with few exceptions' etc. ,) quantifies over an answer to the interrogative complement of the main verb: (1) a. For the most part, John knows who cheated. 'For most x that cheated, John knows that x cheated'. b. John partly remembers who cheated. c. With few exceptions, John found out who cheated. The proposal is that in a QV structure, the adverb quantifies over semantic questions (Hamblin-intensions -- functions from possible worlds to sets of possible answers; Hamblin (1973). All question taking verbs may trigger QV effects. The questions quantified over need not correspond to members of the original Hamblin-extension.

    39. The End OF CED?, by Arthur Stepanov It is well known that in English and many other languages wh-extraction out of subjects and structural adjuncts is impossible, in contrast to extraction out of objects. This work puts the complement/non-complement distinction under closer scrutiny. It is shown that abandoning the distinction is empirically desirable. Abandoning the distinction leads to an 'eclectic' approach in which extractability out of subjects and adjuncts is regulated by different mechanisms of grammar in a non- overlapping manner. The 'eclectic' approach finds independent empirical support.

    40. What Can Child Japanese Tell Us About the Syntax of Scrambling?, by Koji Sugisaki and Miwa Isobe The study is an attempt to determine the validity of two competing syntactic analysis of the Direct Object(DO) -- Indirect Object(IO) order within the VP in Japanese, based on data from child language acquisition. One analysis suggests that this order is derived from the IO-DO order via application of movement operation called scrambling. The other analysis claims that both orders can be base generated. It is shown that under certain acquisitional assumptions, these two analysis make different predictions for child Japanese, and that experimental bear out the predictions of the movement analysis.

    41. Information Structure and Disambiguation in Japanese, by Satoshi Tomioka This paper argues that the interpretations of the Japanese existential construction are intimately influenced by information structure. The semantic impact of word order differences gives an impression that the disambiguation is a result of a syntactic constraint, but that impression is misleading. When Focus, topic marking and embedded sentences are taken into consideration, it becomes clear that a syntactic approach cannot provide a coherent explanation. The pragmatic approach advocated in this paper succeeds in deriving the facts a syntactic approach fails to account for.

    42. Between Mass and Count, by Lucia M. Tovena The present paper discusses a case where mass nouns appear to go together with singular count nouns rather than plural ones. A subset of mass nouns, identified as intensive quantifiers (IQs), can pair with singular count nouns rather than plural ones and the rest of mass nouns in certain quantificational contexts. It is assumed that the structure of the domain of countable nouns contains atoms, and mass nouns denote continuous entities. The author adds to this a is a third type of domain/level, made of weakly discrete units. It is argued that the domain of IQs contains weakly discrete units, in the sense that they can be quantified over in particular contexts. Weakly discrete units in IQs are understood as degrees. Thus finer grained classification of mass nouns allows us the possibility of predicting systematic different grammatical statuses for given det+N combinations.

    43. A Neo-Lexicalist Movement Analysis of Incorporation, by Takashi Toyoshima In this paper, the author scrutinizes Baker's (1996) refined analysis of the theory of incorporation, and proposes an alternative analysis in which incorporated words are formed by a pre-syntactic morpholexical process, and licensed in syntax by phrasal movement of 'pro'. The 'lexicalist' insight into 'classifier' incorporation, coupled with Sportiche's (1996) hybrid analysis of clitic placement can give a better account of Baker's (1996) new evidence. With inflection already being a morphological process in the minimalist framework, this seems to leave no empirical necessity for head-to-head adjunction. This neo- lexicalist movement analysis of incorporation contributes to the pursuit of the 'perfect' system of syntactic computation, by dispensing with unnecessary head-to-head adjunction.

    44. The Syntax of Transitivity and Its Effects: Evidence from Halkomelem Salish, by Martina Wiltschko This paper shows that in Halkomelem, unergative verbs are not concealed transitives, and that there is no syntactic distinction between unergative and unaccusative verbs: This is derived from the morphosyntax of transitivity. In Halkomelem, the projection of vP is dependent on the presence of the transitive marker and therefore unergative predicates are not to be analysed as concealed transitives. The cross-linguistic difference between English-type languages and Halkomelem provides evidence that unergatives are not universally concealed transitives (Chomsky 1995). Cross linguistic differences in morphosyntactic structure correlate with cross linguistic differences in the projection of arguments. This crucially supports Borer's (1994) syntactic predicate driven approach to argument projection.

    45. Measure Phrase Modification in Vector Space Semantics, by Yoad Winter Spatial and temporal measure phrases (MPs) such as ten meters, five years etc. , appear with various linguistic items, including locative prepositions, adjectives and comparatives. Three main facts about MP modifications are observed. (i) Some locative prepositions and degree adjectives allow MP modification while others do not. (ii) When an adjective allows MP modification, it loses the 'value judgement' part of it's meaning. (iii) In their comparative form, both positive and negative adjectives allow MP modifications. To account for such facts, this paper generalizes the modification condition of (Zwarts 1997) and (Zwarts & Winter 2000), which handles MP modification of locatives in Vector Space Semantics (VSS). The extended principle also governs the acceptability of Measure Phrase modification with absolute and comparative adjectives.

    46. An Argument for Category Neutrality?, by Rachel Wojdak The author shows that Nuu-chah-nulth is not a syntactically category neutral language. It has been demonstrated that there is a syntactic evidence for [+/-N] distinction which coincides with the morphological [+/-N] contrast previously known to exist in Nuu-chah-nulth. The author argues that the language allows nominals and non-nominals alike to serve as the base of an argument construction, which is derived syntactically via a determiner. Such an analysis challenges the notion that nominals differ universally from non-nominals in their ability to function as arguments. The underlying contrast between lexical categories in Nuu-chah-nulth is reflected in category dependent restrictions on modification.

    47. Agree: The Other VP-Internal Subject Hypothesis, by Susi Wurmbrand The paper provides empirical evidence for the notion of AGREE (Chomsky 1998, 2000) -- i.e. , an abstract feature matching relation between a functional head and a 'goal' in-situ. This is in contrast to the MOVE approach, whereby Case and Agreement features are checked in Specifier- head configuration. The argument is based on German topicalization constructions in which the subject (i.e. ,a nominative XP agreeing with the finite verb) is in a position lower than its case/agreement position (i.e. , Spec TP) at PF, and importantly, is trapped in this position at LF. Since in these contexts, movement to the specifier position of the licensing head cannot occur (neither overtly nor covertly),the grammaticality of these constructions suggests that Case and agreement licensing does not require a specifier-head configuration, which is compatible with the AGREE approach to feature licensing, but incompatible with the MOVE approach.

    48. On Distributional Differences Between Universal and Existential Quantifiers, by Kazuko Yatsushiro It is argued that the distribution of the universal particle 'mo' and the existential particle 'ka' in Japanese is explained by the semantics of these particles. Japanese universal and existential quantifiers consist of two parts: Indeterminate pronouns and quantificational particles and suffixes. The author adopts and further justifies the P- set analysis of 'mo-construction' proposed by Shimoyama (1999). The P- set analysis proposes that 'mo' universally quantifies over the P-set (Rooth 1985) of the sister constituent of 'mo'. It is also claimed that there is no covert movement or type shifting in Japanese.

    EVALUATION I can only give a general evaluation of the book, since, in a review of the sort presented here, it is impossible to evaluate each of the forty-eight papers separately. The papers cover a wide range of topics in syntax, semantics, morphology, and phonology. The data is drawn from many known, as well as lesser known languages, thus providing the reader an extensive scope for comparison. For example, personally, I found that many of the phenomena I thought occurred only in the languages I am working on, found support in the data presented in this book. The papers are well written, and just the sheer variety is a treat for any serious student of linguistics.

    REFERENCES Baker, Mark (1996) The Polysynthesis Parameter. Oxford University Press. Baker, Mark (1997) Thematic Roles and Syntactic Structure. In L. Haegeman, ed., Elements of Grammar, Kluwer, 73-137. Borer, Hagit (1994) On the Projection of Arguments. In: E. Bendicto and J. Runner, eds., 'Functional Projections'. University of Massachusetts Occasional Papers in Linguistics. 17. 19-48. Chomsky, Noam (1977) 'On Wh-movement'. In P. Culicover, T. Wasow and A. Akmaijan, eds, Formal Syntax, Academic Press, 157- 205. Chomsky, Noam (1981) Lectures on Government and Binding, Foris. Chomsky, Noam (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 Sylvain Bromberger, MIT Press. Chomsky, Noam (1995) The Minimalist Program, MIT Press. Chomsky, Noam (1998) Minimalist Inquiries: The Framework. MIT Occasional Papers in Linguistics 15. Chomsky, Noam (2000) 'Minimalist Inquiries: The Framework'. In R. Martin, D. Michaels, and J. Uriagereka, eds., Step by Step: Essays on Minimalist Syntax in Honor of Howard Lasnik, MIT Press, 89-155. Cinque, Guglielmo (1999) Adverbs and Functional Heads: A Cross- Linguistic Perspective, Oxford. Fox, Danny (1999) 'Reconstruction, Variable Binding and the Interpretation of Chains'. Linguistic Inquiry 30: 157-196. Fox, Danny (2000) Economy and Semantic Interpretation, MIT Press. Hale, Ken and Samuel J. Keyser (1993)'On Argument Structure and the Lexical Expression of Syntactic Relations'. In K. Hale and S. J. Keyser, eds. The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger, MIT Press. Hamblin, C. L. (1973) 'Questions and Montague English'. Foundations of Language 10:41-53. Kayne, Richard (1995) The Antisymmetry of Syntax, MIT Press. Kayne, Richard (2000) 'How Movement, Binding and Agreement are Related'. Talk given at WCCFL 19. Larson, Richard (1990) 'Double Objects Revised: Reply to Jackendoff'. Linguistic Inquiry 21: 589-632. McCarthy, John and Alan Prince (1995) 'Faithfulness and Reduplicative Identity'. In: University of Massachusetts Occasional Papers in Linguistics 18: Papers in Optimality Theory, J. Beckman, L. Walsh Dickey and S. Urbanczyk, eds., 249-384. Reinhart, Tanya and Eric Reuland (1993) 'Reflexivity'. Linguistic Inquiry 24. 657-720. Rizzi, Luigi (1990) 'On the Anaphor Agreement Effect'. Rivista di Linguistica 2: 27-42. Rooth, Mats (1985) Association with Focus. PhD Dissertation, University of Massachusetts, Amherst. Sag, Ivan and Thomas Wasow (1999) Syntactic Theory: A Formal Introduction, CSLI Publications. Shimoyama, Junko (1999) 'Complex NPs and Wh-Quantification in Japanese'. Proceedings of NELS 29, 355-365. Sportiche, Dominique (1996) 'Clitic Constructions'. 213-276. In J. Rooryck and L. Zaring, eds, Phrase Structure and the Lexicon, Kluwer. Wasow, Thomas (1977) 'Transformations and the Lexicon'. In: P. Culicover, T. Wasow, and A. Akmaijan (eds). Formal Syntax, Academic Press, 327-360. Woolford, Ellen (1999) 'More on the Anaphor Agreement Effect'. Linguistic Inquiry. 30: 257-287. Zwarts, Joost (1997) 'Vectors as Relative Positions: A Compositional Semantics of Modified PPs'. Journal of Semantics, 14: 57-86. Zwarts, Joost and Yoad Winter (2000)'Vector Space Semantics: A Model- Theoretic Analysis of Locative Prepositions'. Journal of Logic, Language and Information, 9: 169-211.

    ABOUT THE REVIEWER Sharbani's research interests include Morphology, Minimalist Syntax, Semantics, and their application in Computational Linguistics. net. in">