| title: | Optimality Theory and Natural Language Interpretation |
| author(s): | Reinhard Blutner |
| affiliation: | Institut für deutsche Sprache und Linguistik -- Humboldt University Berlin |
| time and place: | Wednesday, 9.00, room C |
| abstract: | Optimality Theory as developed by Prince, Smolensky and others has been
fruitfully applied to phonology, morphology, and syntax. The attempt to
apply the ideas developed in these fields to the semantics/pragmatics
interface necessitates the conception of a bidirectional OT. The main
reason for assuming the idea of bidirectionality (combining expressive and
interpretive optimization) derives from a number of phenomena which, on the
one hand, demand the treatment of preferred interpretations, and, on the
other hand, suggest to take into account the existence of blocking effects.
In this talk, some general motivation for the idea of bidirectionality is
presented, the importance of this idea for the field of pragmatics
(conversational implicature) is investigated, and some applications of
bidirectional OT are outlined. A central discussion point concerns the distinction between two different conceptions of bidirection. Strong bidirection is the simultaneous realization of expressive and interpretive optimization. Weak bidirection, on the other hand, describes a more sensible interaction of the two modes of optimization – one that realizes Horn’s division of pragmatic labor: (un)marked expressions typically get an (un)marked interpretation). It is argued that strong bidirection describes the state of language after the OT learning algorithm has fully realized the equilibrium between expressive and interpretive optimization. The notion accounts for both unacceptability and ineffability, it is computationally tractable font=-2 (to the same extent as the unidirectional optimizations are), and it is able to deal with aspects of online processing. Weak bidirection, on the other hand, should best be considered as a principle describing the direction of language change (super-optimal pairs are tentatively realized in language change). In order to install this idea, recent work by Robert van Rooy may be helpful adopting an evolutionary setting. Finally, I consider it an exciting challenge for bidirectional OT to make the term “grammaticalization” more precise (discussing Hyman’s proposal to apply the term for "the harnessing of pragmatics by a grammar"). |
| title: | A unified theory of referential NPs |
| author(s): | Gennaro Chierchia |
| affiliation: | University of Milan |
| time and place: | Monday, 10.00, room C |
| abstract: | Noun phrases are generally classified as quantificational (e.g. every man) and referential (e.g., the boy). It is not straightforward to determine where indefinites fall with respect to this classification. They have, in general, a quantificational force; but they are also felt to have “referential” uses. I would like to address this question building on the recent debate on long distance indefinites. Consider, to start, referential NPs and, more specifically, definite descriptions. Their classical analysis is in terms of (something like) the i–operator. For example:
This looks like natural enough explanation for why the has a uniqueness presupposition. Let us turn now to indefinites. It seems plausible to maintain that their meaning differs minimally from that of definites. In particular, we may assume that the indefinite article too denotes (or can denote) a variable over choice functions. However such variable (unlike that associated with definites) must be existentially closed (a requirement triggered by indefinite morphology). What comes out of this is, of course, the Winter/Reinhart approach to indefinites. Except that we now can see that proposal (which arguably explains the peculiar scopal properties of indefinites) as a part of a unified theory of referential NPs (whether definite or indefinite). Summing up:
|
| title: | Discourse Syntax: the Growth of Logical Form |
| author(s): | Ruth Kempson |
| affiliation: | King's College, London |
| time and place: | Monday, 17.00, room C |
| abstract: | In this paper I argue that if we adopt a representationalist view of interpretation, modelling the process of how interpretation is built up in context as a left to right process of structural growth of logical form, the result is a mono-level theory of syntax - syntax as the growth of logical form along a left-right dimension. The case study is the characterisation of relative clause construal and the interaction between long-distance dependency and processes of anaphora resolution. I show that, within the proposed framework, we can model resumptive pronouns, correlative structures, and the distinction between restrictive/nonrestrictive relative clauses within an integrated typology of relative clause types, while retaining a unified and pragmatic account of anaphora. Essential to the characterisation is the analysis of NP construal in terms of variable-binding term-operators of type e (eg epsilon terms), and the building up of partial epsilon terms within the context of progressive logical form growth. In conclusion, I propose that we replace the previous question `Do natural language grammar formalisms need a level of semantic representation in addition to syntactic levels of representation?' with the new research question `Do natural language grammar formalisms need any level of syntactic representation in addition to the semantic representation of logical form?' |
| title: | Indeterminate Pronouns |
| author(s): | Angelika Kratzer |
| affiliation: | University of Amherst, Department of Linguistics |
| time and place: | Tuesday, 9.00, room C |
| abstract: | The term "indeterminate pronoun" is used by scholars of Japanese to refer to a class of pronouns that can take on existential, universal, interrogative, or negative interpretations, depending on the operator they come with. Building on recent work by Junko Shimoyama and Martin Haspelmath, I will explore the possibility that a Hamblin style alternative semantics might lead to a unified analysis of indeterminate pronouns, including the ones in Indo-European languages. |
| title: | Coevolution of Languages and the Language Faculty |
| author(s): | Ted Briscoe |
| affiliation: | Computer Laboratory -- University of Cambridge |
| time and place: | Monday, 14.00 - 16.45, room A In the thematic session on Semantics and Cognition |
| abstract: | Human language acquisition, and in particular the acquisition of grammar, is a partially-canalized, strongly-biased but robust and efficient procedure. A variety of explanations have been offered for the emergence of a partially-innate language acquisition device (LAD), such as exaption of a spandrel (Gould, 1987), biological saltation (Bickerton, 1990) or genetic assimilation (Pinker and Bloom, 1990). But none provide a coherent account of both the emergence and maintenance of a LAD in an evolving population of language users exposed to a variety of linguistic systems during the period of adaptation for the LAD. I will argue that a coevolutionary approach is the only coherent account, and that computational simulation suggests that genetic assimilation of grammatical information will occur despite epistasis and pleitropy and even in circumstances of rapid concurrent linguistic evolution (pace Deacon, 1997). |
| title: | In, On, Over and Between: Toward a functional geometry of spatial prepositions. |
| author(s): | Simon Garrod |
| affiliation: | |
| time and place: | Monday, 14.00 - 16.45, room A In the thematic session on Semantics and Cognition |
| abstract: | Locative expressions are few in number but allow for a wide range of uses
(Landau & Jackendoff, 1993). This discrepancy between the small number of
apparently simple spatial distinctions being made in language and the
frequency and variety of uses to which locative expressions are put presents
a challenge for semantic analysis. In particular, it makes it difficult to
give a straightforward geometric definition for any of these expressions
(Garrod & Sanford,’89; Herskovits, ‘86; Vandeloise, ‘91). The talk describes a series of experiments to show how locatives such as IN, ON, and BETWEEN denote location control relations defined in terms of a functional geometry (Garrod & Sanford, ’89; Garrod, Ferrier and Campbell,’ 99). Such relations capture the way in which objects are seen to control each other’s location by virtue of their spatial arrangement. For example, there is a control relation, which we refer to as fcontainment, by which the ground (i.e., a container) is seen to control the location of the figure (i.e., its content) by virtue of some degree of geometric enclosure. The first series of experiments I will discuss relate to the prepositions IN and ON. They show two things: (1) Confidence in descriptions containing IN and ON relate directly to judgements of the degree to which the ground controls the location of the figure, and, (2) Introducing dynamic information into scenes affects use of descriptions containing IN: Dynamic information consistent with location control of the figure by the ground enhances confidence in IN descriptions for scenes portraying equivalent geometric configurations; By contrast, inconsistent dynamic information reduces confidence in IN descriptions. The other experiment relates to the preposition BETWEEN. Here the control relation is more complicated. If X is BETWEEN Y and Z, then X is seen either to keep Y and Z separate or conversely to hold Y and Z together. This experiment tests the fconnection control relation. Scenes portraying the same geometric configuration of three objects are manipulated to include an alternative connector. Using the rationale behind Garrod et al.s’ (’99) Expt. 2 the prediction is that the presence of an alternative separator should reduce confidence in BETWEEN descriptions. The results supported the prediction for two levels of alternative control analogous to Garrod et al.’ (’99) findings with the preposition ON and alternative sources of support. I will discuss the results of these studies in relation to a functional geometric account for the semantic representation of these locatives. The account has the virtue of offering a simple quasi-geometric definition of the prepositions that can accommodate the wide range of spatial situations to which they can apply. It is also consistent with recent accounts of the perception of complex spatial relations in terms of the so-called “what?” and “where?” systems (Landau & Jackendoff, 1993) and with some accounts of physical imagery (Schwartz, 1999). References: Garrod, S.C. & Sanford, A.J. (1989). Discourse models as interfaces between language and the spatial world. Journal of Semantics, 6, 147-160. Garrod, S., Ferrier, G. & Campbell, S. (1999) In and On: Investigating the functional geometry of spatial prepositions. Cognition, 72, 167-189. Herskovits, A. (1986). Language and Spatial Cognition. Cambridge: Cambridge University Press. Landau, B. & Jackendoff, R. (1993). "What" and "where" in spatial language and cognition. Behavioural and Brain Sciences, 16(2), 217-265. Schwartz, D.L. (1999) Physical imagery: Kinematic versus dynamic models. Cognitive Psychology,38, 433-464. Vandeloise, C. (1991). Spatial Prepositions: A Case Study from French. Translated by Bosch, A.R.K. Chicago/London: The University of Chicago Press. |
| title: | The Emergence of Metacommunicative Meaning |
| author(s): | Jonathan Ginzburg |
| affiliation: | |
| time and place: | Monday, 14.00 - 16.45, room A In the thematic session on Semantics and Cognition |
| abstract: | The issue of how human language evolved is commonly taken to reduce to the issue of how *grammar* has evolved. *Grammar* is taken to be, essentially, a disembodied system of expressions with a compositional semantics. In this talk I will argue for the need to consider a somewhat different perspective on evolution and ontogeny, one where grammar is taken to be a system of types of spatio-temporally located speech events, embedded within a system of information states of interacting agents. I will adduce a number of motivations for this: first, certain semantic phenomena characteristic of mature adult linguistic competence can only be described within the latter type of approach. A striking example of this are the readings displayed by fragments used to acknowledge understanding or request clarification of an utterance made by a previous interlocuter. These constitute an instance of semantic complexity that arises without concomitant syntactic complexity. Second, by considering language as a property of a communicative interaction system one can distinguish between the contingently existing communicative system of a given species of agents and the potential competence of this species. This is clearly true for human neonates, but is also relevant with respect to evidence of (limited) language learning among primates. Third, such an approach is required in order to account for some fundamental characteristics of how certain metacommunicative meaning actually emerges, for instance the ubiquity of partial repetition by novices of competent speakers' utterances. |
| title: | Module Interaction, Language Origins, Language Diversity | ||||||||||||
| author(s): | Pieter Muysken | ||||||||||||
| affiliation: | |||||||||||||
| time and place: | Monday, 14.00 - 16.45, room A In the thematic session on Semantics and Cognition | ||||||||||||
| abstract: |
Linguists must deal with the study of language origin from the perspective of the nature of language itself. What is it about language that might inform us about its possible genesis? The most serious problem with most structural approaches to language, particularly those within the structuralist and generative traditions, is that they have tended to view the language capacity as a single monolithic whole. This monolithic view stands in the way of the gradualist perspective of language evolution that is required if we bring the origin of language in line with the genesis of other human capacities, both cognitive and more generally neurological. A more promising approach takes the modular organization of the language capacity as its starting point. In human language, (at least) four essentially different modules intersect: (a) the structure-building and -processing capacity (syntax), (b) the sign forming and using capacity (semiotics), (c) the capacity to engage in sustained exchange of information (interaction), and (d) the capacity to form complex representations of information (cognition). This claim of four different interacting modules remains empty unless we manage to isolate the formal properties of these modules. These include:
This modular view of the language system has another crucial property: it allows us to account for the fact of language diversity. Why is there diversity at all, and what are its limits? This question is less interesting perhaps in a purely culturalist or semiotic approach to language, in which there is no claim made for a biologically conditioned human language capacity. However, when we view language from a biological perspective, the diversity encountered is a bit of a mystery. Why don't human languages resemble each other much more than they appear to do? Diversity can emerge because there are different ways the different processing systems involved in language can interact. These differences result from their different formal properties. Differences between languages are due to differential access to features defined in other modules. I will focus here on one subtheory, that of grammaticalization: semiotic, cognitive or interactional properties and oppositions become `visible' to syntactic operations through feature sharing at the interface. The relevant metalanguage involves notions such as visibility of features, compatibility of representations, and optimalization of matching. I will illustrate this account by briefly discussing phenomena of nominal classification in some Amazonian languages. |
| title: | Some Uses of Games in Logic |
| author(s): | Rohit Parikh |
| affiliation: | Department of Computer and Information Science, CUNY |
| time and place: | Wednesday, 14.00 - 17.30, room A In the thematic session on Games in Logic and Language |
| abstract: | We will survey some uses of games to understand propositional and
first order logics. The topics will include Ehrenfeucht-Fraisse
games, D-structures, and Ehrenfeucht's *-semantics as well as
IF-logic and the more recent FI-logics (where FI stands for
`finite information'.) D-structures generalize the games used in conventional first order logic and can be used to explain classical as well as *-semantics. Most of this work is old, done jointly with Mayberry, de Jongh and Goodman. But it suddenly has contemporary relevance. FI-logics similarly are a variant of the IF-logics studied by Hintikka and Sandu as well as Hodges, Janssen and Vaananen, but seem to correspond to procedures occuring in ordinary life and possess both the finite model property and decidability. |
| title: | Games and Dynamics |
| author(s): | Gabriel Sandu |
| affiliation: | Department of Philosophy, Helsinki |
| time and place: | Wednesday, 14.00 - 17.30, room A In the thematic session on Games in Logic and Language |
| abstract: |
| title: | The role of salience in the emergence of signalling conventions |
| author(s): | Robert Sugden |
| affiliation: | School of Economic and Social Studies, East Anglia |
| time and place: | Wednesday, 14.00 - 17.30, room A In the thematic session on Games in Logic and Language |
| abstract: | In Convention, David Lewis argues that languages can be understood as conventions. His argument works by showing that conventions by which particular signs indicate particular meanings can emerge as equilibrium solutions to signalling games (a sub-class of coordination games), and by arguing that such conventions are rudimentary languages. A possible criticism of this argument is that Lewis's theory assumes that the players of signalling games share common conceptions of salience; if the existence of such common conceptions depends on the prior existence of a language community, Lewis has not explained language 'from outside'. In Evolution of the Social Contract, Brian Skyrms claims to resolve this problem by reconstructing Lewis's model in a way that dispenses with salience; conventions evolve within a population of inductive learners by the amplication of initially random variations. I argue that Skyrms fails to take account of the problem that led Lewis to invoke salience: the unlimited number of potential signals. If there is an infinity of conceivable regularities in experience, inductive learning requires the prior privileging of a small number of these regularities as ones that, if observed, would count as patterns and not as random noise. Thus, inductive learning depends on a form of salience. However, the common conceptions of salience necessary for the emergence of signalling conventions do not require the prior existence of a language community. |
| title: | Pragmatics of propositional attitudes |
| author(s): | Maria Aloni |
| affiliation: | Department of Philosophy, ILLC -- University of Amsterdam |
| time and place: | Tuesday, 14.00, room B |
| abstract: | The article discusses pragmatic aspects of our interpretation of propositional attitude reports. First it shows that our evaluation of these constructions may vary relative to the ways in which the relevant individuals are identified in the context of use. It then gives this insight a precise formalization from a model- and proof-theoretic perspective in the framework of modal predicate logic. Finally, an account of actual evaluations of these constructions in contexts is proposed in terms of optimization games. |
| title: | Monotonicity and relative scope entailments |
| author(s): | Alon Altman (a), Ed Keenan (b) and Yoad Winter (a) |
| affiliation: | (a) Technion Israel Institute of Technology (b) UCLA Linguistics |
| time and place: | Tuesday, 12.00, room B |
| abstract: | This paper explores the hypothesis that simple monotonicity properties of quantifiers in natural language determine to a large extent the entailment relations between their wide/narrow scope readings. We prove that the disjunctive normal form of upward monotone quantifiers using principal ultrafilters correlates with whether an object narrow scope reading entails an object wide scope reading. This result naturally extends the familiar entailment relations between "Exists, For all" and "For all, Exists" quantification in first order logic into arbitrary ``finitely based'' upward monotone determiners, which are precisely defined. |
| title: | Game theoretical foundations for Gricean constraints |
| author(s): | Nicholas Asher, Itai Sher and Madison Williams |
| affiliation: | Department of Philosophy, University of Texas at Austin |
| time and place: | Tuesday, 12.00, room A |
| abstract: | Gricean maxims of quality, quantity and relevance have an intuitive appeal and would appear to be a part of any adequate pragmatic theory. Previous work in formal pragmatics has gone some way in giving a formalization of some of these principles. Given these formalizations, we can ask and begin to respond to a question that goes to the heart of pragmatics as a theory of rational action in discourse: do pragmatic constraints of a Gricean flavor have a deep connection to established models of rational behavior? We give a partial affirmative answer for one formalization of a pair of Gricean defaults using both standard and evolutionary game theory. |
| title: | Types for Linguistic Typologies. A case study of polarity items |
| author(s): | Raffaella Bernardi |
| affiliation: | Utrecht Institute of Linguistics OTS |
| time and place: | Tuesday, 15.30, room A |
| abstract: | In this paper we show how categorial type logic can contribute to the study of linguistic typologies. In particular, we employ the derivability relations among types holding in the base logic -- the logic of residuated and Galois connected operators -- to account for the classification of polarity items (PIs) based on the distinction between veridical and non-veridical contexts. Crosslinguistic differences between Greek and Italian PIs distribution are accounted for by means of lexical type assignments. |
| title: | Quantified hybrid logic and natural language |
| author(s): | Patrick Blackburn (a) and Maarten Marx (b) |
| affiliation: | (a) INRIA Lorraine (b) ILLC -- University of Amsterdam |
| time and place: | Monday, 12.15, room B |
| abstract: | Quantified Hybrid Logic (QHL) is an extension of first-order modal logic in which it possible to reason about individual states. Recent work has shown that QHL is better behaved logically than ordinary first-order modal logic. In this paper we argue that QHL is also well-motivated from the perspective of natural language semantics, by examining the interaction of tense, temporal reference, and quantification. |
| title: | Duality and anaphora |
| author(s): | António Branco |
| affiliation: | Departement de Infomática -- University of Lisbon |
| time and place: | Monday, 11.15, room A |
| abstract: | The four constraints on sentential anaphoric binding, known as binding principles, are observed to form a square of oppositions. With the formal tools of phase quantification, these constraints are analysed as the effect of phase quantifiers over reference markers in grammatical obliqueness hierarchies. The four quantifiers are shown to be organized in a square of duality. The impact of this result on the distinction quantificational vs. non quantificational NPs and on the semantics of nominals in general is discussed. |
| title: | How pronominal are expletive pronouns? |
| author(s): | Ronnie Cann |
| affiliation: | Department of Theoretical and Applied Linguistics -- University of Edinburgh |
| time and place: | Tuesday, 10.15, room B |
| abstract: | This paper takes a fresh look at expletive constructions in English and presents analyses of 'it' and 'there' constructions within Dynamic Syntax. Whether a pronoun can be used expletively is ascribed to whether nodes decorated by such pronouns may be further developed, but no difference is postulated between expletive pronouns and their anaphoric counterparts. The properties of the former are shown to follow from those of the latter. |
| title: | Counting Concepts |
| author(s): | Cleo Condoravi (a), Dick Crouch (a) and Martin van den Berg (b) |
| affiliation: | (a) XEROX PARC (b) FXPAL |
| time and place: | Tuesday, 14.30, room B |
| abstract: | Singular indefinite NPs in prevention statements give rise to an ambiguity between a general and a specific reading. In order to account for this ambiguity, we extend Zimmermann's proposal for referential concept NPs to also allow for quantificational concept NPs. We further motivate the need for quantificational concept NPs on the basis of the interpretation of indefinite plural NPs with numeral determiners in prevention contexts. We treat numerals as generalized determiners quantifying over concepts and propose a way of counting concepts by counting maximally specific instantiated concepts. |
| title: | Type raising, continuations, and classical logic |
| author(s): | Philippe de Groote |
| affiliation: | INRIA Lorraine |
| time and place: | Monday, 11.15, room B |
| abstract: | There is a striking analogy between Montagovian type raising and the notion of continuation that has been developped in programming language theory. On the other hand, the notion of continuation allows classical logic to be given a Curry-Howard interpretation. Putting the pieces of the picture together, it is possible to use ``classical extensions'' of the lambda-calculus in order to express the semantic components of the lexical entries of type logical grammars. This solution offers the advantage of not burdening the syntax by enforcing type raising to the worst case. |
| title: | Meanwhile within the Frege boundary |
| author(s): | Paul Dekker |
| affiliation: | Department of Philosophy, ILLC -- University of Amsterdam |
| time and place: | Tuesday, 11.00, room B |
| abstract: | With this {talk/paper/squib} I want to contribute to understanding and improving upon (Keenan 1992)'s intriguing equivalence result about reducible type <2> quantifiers. I give an alternative proof of his result which generalizes to type |
| title: | Learning categorial grammars from semantic types |
| author(s): | Daniela Dudau-Sofronie, Isabelle Tellier and Marc Tommasi |
| affiliation: | LIFL-Grappa Team and Université Charles de Gaulle - Lille 3 |
| time and place: | Wednesday, 11.30, room B |
| abstract: | This paper investigates the inference of Categorial Grammars from a new perspective. To learn such grammars, Kanazawa's approach consists in providing, as input, information about the structure of derivation trees in the target grammar. But this information is hardly arguable as relevant data from a psycholinguistic point of view. We propose instead to provide information about the semantic type associated with the words used. These types are considered as general semantic knowledge and their availability is argued. A new learning algorithm from types is given and discussed. |
| title: | VP ellipsis by tree surgery |
| author(s): | Katrin Erk (a) and Alexander Koller (b) |
| affiliation: | (a) Programming Systems Lab, (b) Department of Linguistics -- Saarland University |
| time and place: | Tuesday, 14.30, room A |
| abstract: | We present jigsaw parallelism constraints, a flexible formal tool for replacing parts of trees with other trees. Jigsaw constraints extend the Constraint Language for Lambda Structures, a language used in underspecified semantics to declaratively describe scope, ellipsis, and their interaction. They can be used, among other things, to model certain classes of ellipses where existing approaches run into overgeneration problems. |
| title: | Foundational belief revision in update semantics |
| author(s): | Anthony Gillies |
| affiliation: | Department of Philosophy -- University of Texas at Austin |
| time and place: | Tuesday, 11.30, room A |
| abstract: | There are intuitive examples which seem to favor some sort of foundationalist theory of belief change. This is problematic for those of us who also find semantic models of belief change intuitively compelling since it is far from obvious how we can construct semantic models which are at the same time foundationalist. I offer one way of doing this from within the framework of Update Semantics. One moral is that our foundationalist intuitions are more general than many have thought. |
| title: | Indefinites and sluicing |
| author(s): | Gerhard Jäger |
| affiliation: | Department of Linguistics -- Humboldt University Berlin |
| time and place: | Tuesday, 14.00, room A |
| abstract: | We propose to extend the Lambek calculus with two additional implications, where the first one models anaphora and the second one indefiniteness. Both pronouns and indefinites are interpreted as identity functions, but they give rise to different types and are thus subject to different interpretation strategies. This leads to a straightforward surface compositional analysis of scopal behavior of indefinites and of Sluicing. |
| title: | Implicit slashing in IF-logic |
| author(s): | Theo Janssen |
| affiliation: | Department of Computer Science, ILLC -- University of Amsterdam |
| time and place: | Wednesday, 10.15, room B |
| abstract: | Independent choices arise in Hintikka's IF logic: the quantifier ?y/x used in IF logic says that y must be chosen independent of x. However, it has been argued that Hintikka's semantics for Independence Friendly logic does not formalize the intuitions about independent choices. One of the arguments is that 'signaling' is possible: to transfer the value of x by means of another variable. Hintikka adopted a convention (`implicit slashing') which prevents signalling. In this contribution it will be argued that this convention introduces several new conflicts with intuition and another solution will be proposed. |
| title: | Learning word-to-meaning mappings in logical semantics |
| author(s): | Makoto Kanazawa |
| affiliation: | Interfaculty Initiative in Information Studies -- University of Tokyo |
| time and place: | Wednesday, 11.00, room B |
| abstract: | I generalize Siskind's problem of learning word-to-meaning mappings to logical semantics, and formulate the heart of the problem as a problem in typed lambda calculus. I show that every instance of this `mapping problem' has infinitely many solutions, but many of them are equivalent in a certain sense. I show an algorithm for finding a reasonably small solution with respect to a certain `definability' relation between terms. |
| title: | Tense probabilism properly conceived |
| author(s): | Stefan Kaufmann |
| affiliation: | Graduate School of Informatics -- Kyoto University |
| time and place: | Tuesday, 10.15, room A |
| abstract: | While it is often observed that present counterfactuals are related to earlier predictive indicatives, they do not in general have the same probabilities at their respective reference times. I argue that they are equivalent without being equiprobable. The probability of an indicative predictive conditional is the expectation of the various values of the counterfactual it may ``turn into.'' |
| title: | Categorial-hybrid logical grammar |
| author(s): | Geert-Jan Kruijff |
| affiliation: | Department of Computational Linguistics -- Saarland University |
| time and place: | Monday, 11.45, room B |
| abstract: | A categorial grammar describes the relation between surface form and the (linguistic) meaning it realizes. Traditionally, categorial grammar has captured meaning using a typed lambda calculus, building a representation in parallel to the categorial inference. In this paper, a resource-sensitive categorial proof theory is proposed in which meaning is represented as a term in a hybrid (modal) logic. The system maintains compositionality, but differs from traditional systems linguistically in that meaning is conceived of as an ontologically rich, relational structure. |
| title: | Modality in comparatives |
| author(s): | Cecile Meier |
| affiliation: | Institut für Deutsche Sprache und Literatur II -- Frankfurt University |
| time and place: | Wednesday, 11.30, room A |
| abstract: | In this paper, I investigate more closely the contribution of modal operators (interpreted in a Kratzer-style semantics) to the semantics of comparatives and I show that there is no need for a maximality or minimality operator. I demonstrate that the ordering source reduces a set of possible degrees to a single degree that is most (or least) wanted or expected, i.e., maximality and minimality readings of comparative constructions are an effect of the pragmatic meaning of the modal. |
| title: | Lambda grammars and the syntax semantics interface |
| author(s): | Reinhard Muskens |
| affiliation: | Department of Linguistics -- University of Tilburg |
| time and place: | Tuesday, 16.30, room A |
| abstract: | In this paper we discuss a new perspective on the syntax-semantics interface. Current approaches to semantics are all based on the idea that syntactic objects are somehow prior and that semantics must be parasitic on those syntactic objects. We challenge this idea and develop a grammar in which syntax and semantics are treated in a strictly parallel fashion. Using grammatical signs that are tuples of lambda-terms, it turns out possible and even practicable to define grammars in which all dimensions of the grammar are combined simultaneously in the way terms are usually combined in semantics. The grammar we consider is related to LFG and to CG. |
| title: | Adverb order in type logical grammar |
| author(s): | Øystein Nilsen |
| affiliation: | Utrecht Institute of Linguistics OTS |
| time and place: | Tuesday, 16.00, room A |
| abstract: | Adverb order in type-logical grammar Empirical evidence concerning adverb ordering is presented which 1) cannot be dealt with syntactically in terms of a sequence of functional projecitons (Cinque 1999, Alexiadou 1997) and 2) suggests that (some) adverb ordering phenomena should be dealt with on a par with polarity sensitive expressions and scope phenomena, rather than analyzed with the aid of enrichment of the semantic ontology (Ernst 2001, Bartsch 1976). The account is cast in terms of Type-Logical Grammar. |
| title: | What psycholinguistics tells us about the semantics/pragmatics interface: the case of pronouns |
| author(s): | Massimo Poesio |
| affiliation: | Department of Computing Science |
| time and place: | Monday, 12.15, room A |
| abstract: | Whereas most theories of the meaning of pronouns view them as having a single, underspecified lexical meaning, psychological evidence about pronoun resolution and lexical access suggests that the lexicon does not assign pronouns an initial interpretation; rather, their interpretation is specified by competing defaults operating incrementally. We propose an incremental theory of the semantics and pragmatics of pronoun interpretation based on this evidence, that assumes that interpretation involves competing defaults activated in parallel. |
| title: | Epistemic implicatures in yes/no questions |
| author(s): | Maria-Isabel Romero and Chung-Hye Han |
| affiliation: | Department of Linguistics -- University of Pennsylvania |
| time and place: | Tuesday, 16.30, room B |
| abstract: | A yes/no-question with preposed negation -e.g., "Doesn't John drink?"-- gives rise to the implicature that the speaker expects the answer to be in the affirmative, contrary to its non-preposed negation version -e.g., "Does John not drink?". We propose that preposed negation involves focus on the negative polarity, and that its interaction with the semantics of yes/no-questions and with the discourse hierarchy of questions derives the desired epistemic implicature. |
| title: | Partial adjectives vs. total adjectives |
| author(s): | Carmen Rotstein and Yoad Winter |
| affiliation: | |
| time and place: | Wednesday, 12.00, room A |
| abstract: | This paper argues that the semantic behavior of modifiers such as almost, slightly and nearly is sensitive to the typology of total and partial adjectives. These are pairs of adjectives such as safe-dangerous or clean-dirty, where the first ("total") adjective in each pair describes lack of danger, dirt etc., while the second ("partial") adjective describes the existence of such properties. We analyze the truth and acceptability conditions of constructions such as almost safe/?dangerous and show some systematic contrasts between total and partial adjectives. These contrasts are formally accounted for by giving total adjectives and partial adjectives a different scale structure. |
| title: | A referential analysis of conditionals |
| author(s): | Philippe Schlenker |
| affiliation: | Department of Linguistics -- University of Southern California |
| time and place: | Wednesday, 11.00, room A |
| abstract: | We analyze if-clauses as definite descriptions of worlds. Stalnaker¹s Selection Function analysis of conditionals and von Heusinger¹s Choice Function analysis of definite descriptions are seen to be two sides of the same coin: "if p" selects the most highly ranked p-world(s) under a measure of similarity; "the P" selects the most highly ranked P-individual(s) under a measure of salience. Both expressions are referential terms, which (i) may yield referential failures, (ii) are subject to syntactic constraints on coreference, (iii) may be topicalized, and (iv) may be doubled by a pronoun. The indicative vs. subjunctive distinction is reanalyzed in terms of a general system of referential classification. |
| title: | Two kinds of long-distance indefinites |
| author(s): | Bernhard Schwartz |
| affiliation: | Department of Linguistics -- University of Texas at Austin |
| time and place: | Monday, 11.45, room A |
| abstract: | This talk argues that indefinites taking long-distance scope come in two varieties. One kind, functional indefinites, should be analyzed in terms of free (skolemized) choice function variables (Kratzer). By contrast, long-distance scope of non-functional indefinites is better credited to long-distance scope shifts. Building on observations by Chierchia, it is shown that wide scope existential closure of (skolemized) choice function variables (Reinhart, Winter, Matthewson) overgenerates. |
| title: | Existential import |
| author(s): | Pieter Seuren |
| affiliation: | Max Planck Institute für Psycholinguistik, Nijmegen |
| time and place: | Tuesday, 10.15, room A |
| abstract: | This paper shows that Aristotelian Predicate Calculus (APC) is logically sound if the Conversions are replaced with one-way entailments; this revised APC corresponds better with natural semantic intuitions than either APC or the modern standard calculus. With the use of presupposition-preserving negation, the system is extendable to cover quantification over intensional objects with intensional predicates. |
| title: | A variable-free dynamic semantics |
| author(s): | Chung-chieh Shan |
| affiliation: | Harvard University |
| time and place: | Tuesday, 11.00, room A |
| abstract: | I propose a variable-free treatment of dynamic semantics. By
"dynamic semantics" I mean analyses of donkey sentences ("Every
farmer who owns a donkey beats it") and other binding and anaphora
phenomena in natural language where meanings of constituents are
updates to information states, for instance as proposed by
Groenendijk and Stokhof. By "variable-free" I mean denotational
semantics in which functional combinators replace variable indices
and assignment functions, for instance as advocated by Jacobson. The new theory presented here achieves a compositional treatment of dynamic anaphora that does not involve assignment functions, and separates the combinatorics of variable-free semantics from the particular linguistic phenomena it treats. Integrating variable-free semantics and dynamic semantics gives rise to interactions that make new empirical predictions, for example "donkey weak crossover" effects. |
| title: | Embedded quantifiers and quantificational variability |
| author(s): | Yael Sharvit |
| affiliation: | Department of Linguistics -- University of Conneticut |
| time and place: | Tuesday, 11.30, room B |
| abstract: | As observed in Szabolcsi (1997), many quantifiers in interrogatives embedded under verbs of the 'know'-class may be interpreted as having scope over the embedding verb. This paper argues that this wide scope is only apparent, and is a Quantificational Variability effect. The proposal predicts: (a) The unavailability of apparent wide scope for Quantificational Variability with such verbs; (b) Embedded 'no' cannot take apparent wide scope; (c) When the embedded quantifier takes apparent wide scope, the common noun part of the quantifier may be read 'de dicto'. |
| title: | Expressivity of extensions of DPL |
| author(s): | Balder ten Cate, Jan van Eijck and Juan Heguiabehere |
| affiliation: | ILLC -- University of Amsterdam |
| time and place: | Wednesday, 12.00, room B |
| abstract: | Dynamic predicate logic (DPL), presented in Groenendijk and Stokhof (1991) as a formalism for representing anaphoric linking in natural language, can be viewed as a fragment of a well known formalism for reasoning about imperative programming (Harel 1979). An interesting difference from other forms of dynamic logic is that the distinction between formulas and programs gets dropped: DPL formulas can be viewed as programs. In this paper we show that DPL is in fact the basis of a hierarchy of formulas-as-programs languages. |
| title: | Too |
| author(s): | Rob van der Sandt and Bart Geurts |
| affiliation: | Department of Philosophy -- University of Nijmegen |
| time and place: | Tuesday, 16.00, room B |
| abstract: | We propose an account of the presuppositional contribution of too which deviates from the orthodox view in that we divide the standard presupposition in two: a descriptive part encoding information which must be matched by parallel information in the discourse context, and a hidden pronominal element. The two components are resolved independently and may link up to different levels of discourse structure. This yields a straightforward account of Kripke's observation that too seems to give rise to a more specific presupposition than is predicted by the standard view. The analysis also enables us explain why the presupposition of too is reluctant to accommodate, may access positions that are otherwise inaccessible, and gives rise to fully transparent readings in attitude contexts. |
| title: | The Semantic Contribution of Particles |
| author(s): | Henk Zeevat |
| affiliation: | Department of Computational Linguistics -- University of Amsterdam |
| time and place: | Tuesday, 15.30, room B |
| abstract: | The paper proposes a dual approach to modal particles like doch and ja. They are taken to be primarily context markers (particles indicating that the sentence has one of a small number of relations towards the context in which it is uttered) but can develop into speech act markers, markers that indicate which speech act is expressed with the sentence. This allows a relatively easy treatment of these particles around a core meaning. |