South Asia Syntax-Semantics Newsletter I am happy to announce that we have a new member on the editorial staff of SASSN. Rajesh Bhatt has joined me as an associate editor. Rajesh is a second year student in the Linguistics department at University of Pennsylvania. Before coming to the United States he got his B.Tech in Computer Science and Engg. from IIT, Kanpur. I am naturally very happy to have another person to share the responsibilities of this venture and I expect the readers will soon feel the impact of this change too. I think his energy and enthusiasm will manifest itself in improved coverage of news pertinent to South Asianists working in syntax and semantics. While we will both work on SASS, I will request contributions to be sent to me in order to ensure organizational order. Also, we have decided to keep Rutgers as the home base in the interests of maintaining continuity. In the last issue, I mentioned that SASS Newsletter has been supported for the last three years by the Department of Linguistics, Rutgers University and that there is a need now for making it financially self-supporting. One way of doing this is by cutting down on the costs. With this issue we are exploring the possibility of email transmission of do not have email addresses for all our subscribers so do let us know if you are on email. Please send email to Rajesh (email: bhatt@babel.ling.upenn.edu). As a test case, we shall send copies by email to those subscribers whose email addresses we do have. We will also continue to send out hard copies for now unless you let us know that you prefer to receive your SASS issues by email. However, we might have to adopt more stringent measures in the future. Another option is to create a SASS fund to help with the costs of sending out hard copies. I had suggested in the last issue that subscribers could send checks in the name of "SASSN, Rutgers University" to the address below. I would suggest checks in the amount of $5 since anything lower than that is probably not worth the trouble of processing. Let me reiterate that this donation is completely voluntary and no one will be taken off the mailing list. However, if you are in a position to make a donation, I would urge you to do so in order to ensure long term viability of this project. Thanks to all of you who have contributed, particularly readers from India whose work would otherwise remain unknown to us here in the US. I now have a copy of Lalitha Murthy's dissertation (reported in SASS 3.1) available for distribution. For wider availability of this newsletter we are planning to make it available also via the South Asia Gopher(SAG), located at Columbia University. We thank David Magier for extending this offer to us. Veneeta Dayal Editor, SASS Department of Linguistics Rutgers University 18 Seminary Place New Brunswick, NJ 08903 Phone: (908) 932-6903 email:dayal@zodiac.rutgers.edu Rajesh Bhatt Associate Editor, SASSN Department of Linguistics University of Pennsylvania Room # 619, Williams Hall Philadelphia, PA 19104 email:bhatt@babel.ling.upenn.edu DISSERTATION ABSTRACTS Rakesh Mohan Bhatt Word Order and Case in Kashmiri Doctoral Dissertation, University of Illinois at Urbana-Champaign, 1994. Thesis Supervisor: Prof. H. James Yoon This thesis explores the verb second phenomenon in Kashmiri and positions in the Kashmiri phrase structure where grammatical functions are licensed. V2 phenomenon is found elsewhere only in the languages belonging to the Germanic family. The standard account of V2 has the verb moving to Comp in matrix clauses -- in subordinate clauses there is no V2 because of the presence of an overt complementizer. In Kashmiri, Icelandic and Yiddish, however, V2 is found even in subordinate clauses with overt complementizers. The thesis investigates in detail the "Kashmiri Vorfeld" and concludes that there is a focus position immediately preceding the finite verb which in turn is (optionally) preceded by a (base-generated) topic position. Working within the Principles and Parameters framework of (Chomsky 1981-92) where UG is assumed to offer a range of principles and options which can be parameterized in different langauges, it develops a model of V2 which restricts the availability of variation in V2 languages to the possibilites offered by it. A related issue, discussed in generative accounts of V2 phenomenon, is the correlation between verb movement and nominative Case assignment to subject. This motivation for verb movement is untenable for Kashmiri V2 because not all subjects in Kashmiri are nominative -- experiencers and subjects of perfect predicates pose a serious challenge to an account of V2 motivated by Case-theoretic requirements. The assignment of nominative Case in Kashmiri cannot be tied to either Comp or Infl. In dative and ergative subject constructions the subject NP is non-nominative whereas the object is nominative. Such data challenge some current assumptions of the PP framework, particularly the uniqueness of the subject position in a clause, and the claim that subjects stand in a privileged agreement relation with the verb. It presents an account of Case in Kashmiri which is independent of the V2 phenomenon; an account that explains how lexically assigned (dative/ergative) Case on the subject NP and nominative on the object NP is accomplished. It also provides a straightforward account of verb agreemnt. Finally, it motivates movement of lexically Case-assigned NPs to the "canonical" subject position. Veena Dhar Dwivedi Syntactic Dependencies and Relative Phrases in Hindi Doctoral Dissertation, University of Massachusetts, Amherst, 1994. Thesis Supervisor: Prof. Peggy Speas This thesis explores various syntactic dependencies in relative phrases in Hindi and shows that besides Scrambling, Topicalization and Left Dislocation, a fourth A' dependency must be recognized: Topic Dislocation. This relation is limited to referential Noun Phrases which are related to null pronominals. Since it is a non-movement relation, it is not subject to Subjacency. As a result, it may occur in construction with complex Noun Phrases. Further, it explores extraction phenomena from the correlative construction and claims that the relative phrase is not adjoined to the main clause (cf. Hale 1976, Srivastav 1991); it is asymmetrically co-ordinate to it. Following Williams 1994, it claims that this is a 'double headed' construction. Using this characterisation plus the fact that left conjuncts in Hindi are transparent to extraction, it is able to account for the transparency of both clauses of the correlative construction. Further, it claims that the 'that' clause in Hindi is also asymmetrically co-ordinate to the main clause. It is not subordinate to it at all. It shows that typical c-command effects do not obtain between the matrix and embedded clause. Finally, it posits the existence of 'afterthought restrictors'. These are relative phases to the right which have been assumed to be right adjoined. I assume a bar on all right adjunction, and instead claim that these phrases are linked at the Discourse level to the main clause. This proposal accounts for the fact that such phrases are completely opaque to any external syntactic dependency, e.g. Topicalization, Topic Dislocation and Left Dislocation. Since these phrases are not syntactically connected to the main clause, elements in them may not be c-commanded by the constituents in the main clause. Without c- command, none of the above mentioned relations is licensed. The assumption that right adjunction is barred in Hindi is supported by and supports Kayne 1993, where it is proposed that right adjunction is universally disallowed. Mona Singh Perfectivity, Definiteness, and Specificity: A Classification of Verbal Predicates in Hindi Doctoral Dissertation. University of Texas at Austin, 1994 Thesis Supervisor: Prof. Manfred Krifka The study of events in natural language is of prime importance in linguistics. Though there has been recent progress on formal theories of events, these theories do not address certain syntactic and semantic properties peculiar to languages such as Hindi. In this thesis many useful intuitions about events and the objects that participate in them are captured in terms of a small number of semantic features such as whether an object exists independently of an event, whether it is totally afected by the event, and so on. The intuitions about events concern two varieties of perfectivity; those about objects concern definiteness and specificity. I motivate some of these primitives, which have not been studied in the formal literature. I show how these primitives can be formalized in an algebraic framework and used in a categorial grammar to derive the properties of verbal and nominal predicates. The result is an integration of descriptive semantics with algebraic theories of objects and events. ABSTRACTS OF ARTICLES Josef Bayer (University of Stuttgart): "On the Origin of Sentential Arguments in German and Bengali" Papers from the 7th Comparative Germanic Syntax Workshop, Stuttgart" edited by H.Haider, S.Olsen and S.Vikner. This paper addresses the challenge of sentential complementation facts in German and Bangla to Baker's Uniformity of Theta Assignment Hypothesis. It argues against an extraposition analysis and in favor of positing two alternative A-positions for the licensing of clausal complements, one to the left and one to the right of V. This allows the validity of UTAH to be maintained. The existence of these positions is the result of the relation that the verb has with the CP's inner architecture. The impact of this analysis on Wh movement at SS and LF is also explored. Veneeta Dayal (Rutgers University): "Scope Marking as Indirect WH Dependency" Natural Language Semantics, 2.2, 1994, pp. 137-70. In languages such as Hindi, Bangla, German, Romani and Iraqi Arabic, scope marking structures are used to express long distance wh dependencies along with or instead of the more familiar extraction structure. The existence of these two strategies raises an interesting question for the mapping between syntactic structure and semantic representation. Should apparent semantic equivalence be taken as a guide and syntactic parallelism posited at an abstract level of syntax? Or should the surface syntactic distinction between them be maintained and an alternative explanation sought for the similarity in meaning? This paper shows that theoretical as well as empirical considerations argue against the first approach. It presents a syntactic analysis of scope marking structures in which the dependency between wh expressions is indirect (in contrast to extraction structures which encode direct wh dependencies). It draws attention to certain differences between scope marking and extraction structures which show that they are not really equivalent. The interpretive procedure given for indirect wh dependencies derives the considerable similarity in meaning between the two structures while maintaining the necessary distinctions. Wayne E. Harbert (Cornell University) and Veneeta Dayal (Rutgers University). "Arbitrary Apnaa", South Asian Language Review, IV. 1, 1994, pp. 75-88. It is widely assumed that anaphors -- reflexives and reciprocals and the like -- are elements which have to be bound. It is shown that this assumption is theoretically unfounded and empirically invalid. Certain lexical anaphors, Hindi apnaa among them, need not be bound when they occur in syntactic contexts of a particular form. This is in keeping with the claim in Manzini (1983) that PRO, an anaphor, has arbitrary interpretation if the syntactic context in which it occurs exempts it from the binding requirement. The role of genericity in determining the interpretation of an unbound anaphor, the types of anaphors that can yield arbitrary readings and the relevance of the accessibility provision of the Binding Theory in the grammar of Hindi are some of the issues explored. Hans Hock. Wiener Zeitschrift fur die Kunde Sudasiens 35 Supplement 1993:9-29. Argues, against Hettrich (1988): Untersuchungen zur Hypotaxe im Vedischen, that appositive (or non-restrictive) relative clauses are not dying out in post-Rig-Vedic Sanskrit but continue into the classical language and into the modern Indo-Aryan languages. Also points out two pragmatic conditions which make appositive RCs felicitous and shows that these conditions are met more readily in certain types of texts than in others. Hans Hock. What's a nice word like you doing in a place like this? Syntax vs. Phonological Form. SLS 22:1.39-87 (1992) [Argues that P2 clitics and particles, including those of Vedic Sanskrit, are placed according to prosodic, rather than purely syntactic criteria. Hans Hock, Review of Tikkanen (1987): The Sanskrit gerund: A synchronic, diachronic, and typological analysis. Kratylos 37.62-68 (1992) Argues against Tikkanen's claim that early Vedic absolutives in -tva/tvi/ya are only, or predominantly, used in anterior/perfective function. Like their early Dravidian counterparts they seem to be temporally and aspectually unmarked. Hans Hock, Aspects of Sanskrit agreement. Proceedings of the Panel on Agreement in South Asian Languages, 1991 South Asia Meeting, Madison, Wisconsin. (In Press) Sanskrit agreement is best accounted for in a post-syntactic, morphological component. Syntactic frameworks subscribing to the strict autonomy-of-syntax hypothesis cannot satisfactorily account for Sanskrit agreement. Hans Hock,Discourse linkage in Sanskrit narratives with special emphasis on the story of Nala. Papers from the 1993 SALA Meeting, ed. by A. Davison et al. (In Press.) In addition to absolutives, noted by scholars like Bloch and attributed to Dravidian influence, Sanskrit uses a variety of other morpho-syntactic devices for discourse linkage, specifically in epic narrative. These include conjunctions and adverbial quasi-conjunctions, topicalized NPs, fronted demonstratives, fronted non-finite verbs other than absolutives, and finite-verb fronting. Hans Hock,Who's On First: Toward a prosodic account of P2 clitics. Papers from the July 1993 P2 Clitic Workshop, ed. by A. Halpern and A. Zwicky. (In press). Presented at the Workshop, resubmitted in substantially revised form. P2 clitics and particles, and in Vedic Sanskrit whole 'strings' of fronted elements cannot be satisfactorily accounted for syntactically. Mark Hale's attempt to do so fails on empirical grounds; fixing it up would require an immense, and quite ad hoc, proliferation of left-peripheral 'landing sites'. More significantly, in Vedic these strings occur not only clause-initially, but also line-initially, and after caesura, i.e., after prosodically defined breaks, for which there are no credible syntactic correlates. Further, the syntactic and prosodic breaks accommodating string placement also permit finite-verb and vocative accentuation in main clauses. Since clause breaks usually coincide with prosodic breaks, the best generalization for string placement and finite-verb and vocative accentuation is one in terms of prosody, not syntax. Anoop Mahajan (UCLA). "Against the Relevance of Subjacency at LF: The Case of Hindi Wh". Linguistic Inquiry, 25.1, 1994, pp. 171-179. This squib argues against the view advanced in Srivastav (1991) that Subjacency effects can be clearly observed at LF in Hindi wh questions. It notes four problems with Srivastav's account of the phenomenon of wh movement and proposes an alternative in which wh expressions adjoin to IP as an instance of Quantifier Raising. The conclusion it reaches is that the issue of whether Subjacency applies at LF or not remains an open question. Anoop Mahajan (UCLA). "On Gamma Marking Adjunct Traces in Hindi". UCLA Occassion Papers in Linguistics, 11, 1993, pp. 171-179. Lasnik and Saito (1984) distinguish between arguments and adjuncts along the following lines: argument traces created at SS must be gamma-marked at SS while gamma-marking for adjunct traces must be delayed till LF. This paper examines some aspects of adjunct extraction in Hindi that seem to indicate that such a distinction would yield undesirable results. The evidence suggests that adjunct traces should be gamma-marked in the same fashion as argument traces. This implies that Lasnik and Saito's analysis must be modified. The paper suggests a minor revision of Lasnik and Saito's theory that has the desired result. This revision does not appeal to SS/LF asymmetry in gamma-marking of arguments and adjuncts and is therefore compatible with theories such as Chomsky (1992) which dispense with the level of SS. Kashi Wali (Syracuse and Cornell Universities) and Ashok Kumar Koul (Brown University). "Kashmiri Clitics: The Role of Case and CASE". Linguistics, 32.5. 1994. Kashmiri pronominal clitics show a dual relationship -- of form and order -- with their coreferent. Forms of clitics are linked with lexical cases in a biunique fashion. However, the E-forms are linked with ergative subjects as well as null cased objects forming an exception. It is argued that lexical cases are a complex of positive and negative features and under the feature system the ergative and null cased objects share identical feature complexes. The clitics and coreferent NPs are now linked under the uniform condition that clitics match the lexical case features of their coreferent. BOOKS Edited by Hans Hock. Studies in Sanskrit syntax. Motilal Banarsidass, Delhi, 1991. Contents: Ashok Aklujkar: Syntactic gleanings from Bhartrhari's Trikandi R. N. Aralikatti: A note on word order in modern spoken Sanskrit and some positive constraints Vit Bubenik: Nominal and pronominal objects in Sanskrit and Prakrit Madhav Deshpande: Paninian syntax and the changing notion of sentence AND On the meaning of tumartha Hans Henrich Hock: Possessive agents in Sanskrit? AND Causees, passive agents, or instruments? Instrumental NPs with causatives in early and later Vedic Prose Stephanie Jamison: The syntax of direct speech in Vedic Brian D. Joseph: A diachronic phonological solution to the syntax of Vedic negative particles Jared S. Klein: Syntactic and discourse correlates of verb-initial sentences in the Rigveda K. Meenakshi: The genitive in Panini and in Epic Sanskrit Steven Schaufele: Single-word topicalization in Vedic Prose: A challenge to Government & Binding? AND Verb-medial clauses in Vedic: Some theoretical implications Bertil Tikkanen: On the syntax of Sanskrit gerund constructions: a functional approach A bibliography of writings on Sanskrit syntax by M.M. Deshpande and H. H. Hock. Some peculiarities of Vedic-Prose relative clauses. PAPERS PRESENTED AT RECENT CONFERENCES Linguistic Society of America, Annual Winter Meeting January 1994, Boston: Nalini Rau (University of Illinois-Urbana): "Reflexives in Kannada and the HPSG binding theory" contrasts evidence for a subject-antecedent rule for the binding properties of reflexive anaphors in Kannada with counter evidence for the rule. She offers an alternative analysis where both kinds of data are constrained by a single rule: the reflexive's antecedent is the least oblique argument of the verb. Anjum Pervez Saleemi(National University of Singapore) "Derivation constraints in early Urdu syntax" explains the systematic incompleteness of children's early language in terms of the following three factors, using the minimalist framework: (a) the grammar allows derivations that may converge prematurely, (b) the syntax is generally LF oriented (c) the overt syntax is degenerate in respect of language-particular elements and processes. George Giannakis (University of California-Los Angeles) "The system of the Karakas in Panini and its position in the history of linguistics" discusses the notion of "karaka", isolating the similarity and differences with modern theories of case grammar. In modern terminology, a karaka stands for the deep or underlying relation of an NP to the VP, whereas vibhakti (case form) is the representation of the karakas in the surface structure. Editors' note: These conference reports are prepared on the basis of secondary information. Any errors or omissions are regretted. If you wish to to ensure accuracy, please send me a three to four line summary of papers presented by you.