Tbilisi 2009: Eighth International Tbilisi Symposium on Language, Logic and Computation |
 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
Preliminary programProgram Grid
Workshop dedicated to the birthdays of Leo Esakia and Dick de JonghIn honour of Leo Esakia and Dick de Jongh, their importance for international logic and for the Georgian symposia, a workshop is held Tuesday September 22. Speakers may include:
(back to program) Detailed schedule for contributed talksAuthors, titles and abstractsTutorialsCarlos ArecesIn this tutorial we will start from the very beginning by introducing propositional logic, and using it as an example to discuss fundamental notions like model, proof, expressive power, satisfiability / unsatisfiability, contradiction / tautology / contingency. We will then introduce different inference tasks and define algorithms that solve them. We will discuss what these algorithms tell us about the logic, and which is their complexity. We will then show that propositional logic is, at the same time, very expressive and not expressive enough: even though we can represent complicated problems using propositional logic, we cannot properly represent `trivial' arguments like "If all my students are intelligent", and "I have at least one student", then "I have at least one intelligent student". We will then investigate ways in which we can extend the expressive power of our logic to overcome this limitation, and see that our fundamental notions (model, proof, satisfiability, etc.) still 'work,' and that we can even adapt our inference algorithms, in a simple way, to the new logics. By repeating the game of `what I can/cannot say in my logic' and extending its expressivity when needed, we will arrive to very complex languages (first and second order) in a uniform way (and on our way there, we will have learned one or two things about model theory, inference methods, expressive power and complexity theory).
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| sortal | et | indefinite, absolute |
| individual | e | definite, absolute |
| relational | eet | indefinite, possessive |
| functional | ee | definite, possessive |
The central proposal of the talk is this:
- The function of (non-quantificational) determiners is to indicate the logical type of the NP token; the token type is built upon the lexical type of the head N, but NP token type and lexical N type may differ.
- If the lexical N type coincides with the NP token type, it remains unchanged. This case constitutes the unmarked option.
- If the two types differ, the lexical type of N is shifted to the type indicated by the determiner. This is the marked option.
- Type shifts may involve regular lexical shifts, mechanisms of semantic composition, as well as the exploitation of given context information.
A logical analysis is offered for the system of possible combinations of noun types and types of determiners (all combinations are possible). Statistical and typological evidence will be presented which corroborates the proclaimed markedness phenomena.
Institut für Sprache und Information, Heinrich-Heine-Universität, Düsseldorf.(back to program)
Norbert Preining
Semantics for First Order Gödel Logics -- the 9-headed Hydra
Over the time span of about 50 years the same logic has been discussed
and reinvented again and again. Today generally called "Gödel Logic"
it is interesting out of many different reasons: On the syntax side
it is the only logic in the class of t-norm based logics allowing a
recursive axiomatization. And on the semantic side one can find a
big selection of different semantics, going from real valued truth-values,
via Kripke style semantics, to algebraic semantics. And each of these
semantics throws a different light on the underlying logic and its
strength and weakness.In this talk we present a general introduction to first order G¨o;del logics, present the different semantics and pick for each of the semantics one particularly interesting example where the (often present) gap between logic and (more general) mathematics is bridged.
Institute of Discrete Mathematics and Geometry, Vienna University of Technology.
(back to program)
Workshop
Lev Beklemishev
On topological interpretation of Japaridze's provability logic
The well-known polymodal provability logic GLP (introduced by
Giorgi Japaridze) is known to be incomplete w.r.t. Kripke frames. We study
natural topological models of GLP, in the spirit of Esakia, where
modalities correspond to derived set operators acting on a
poly-topological space. We show topological completeness for the fragment
of GLP with two modalities. On the other hand, for more than two
modalities, completeness w.r.t. natural ordinal spaces turns out to be
dependent on large cardinal axioms of ZFC and various facts on stationary
sets and reflection. In particular, it is consistent (relative to ZFC)
that GLP is incomplete. However, under the assumption of large cardinal
axioms one can establish at least some partial completeness results. Under
the assumption V=L we show that the bimodal fragment of GLP is complete
w.r.t. the cardinal $\aleph_\omega$ (taken with the interval topology and
the club filter topology). Partly joint work with Thomas Icard and Guram
Bezhanishvili.
Steklov Mathematical Institute, Russian Academy of Sciences, Moscow.
(back to workshop)
Nick Bezhanishvili
Generalized Esakia duality and generalized Jankov-de Jongh formulas
The Esakia duality between the category of Heyting algebras and the
category of Esakia spaces---special order-topological spaces---provides
a good understanding of the phenomenon of the Jankov-de Jongh
formulas. In this talk, I will discuss a generalization of the Esakia duality.
This generalization will enable us to develop a similar approach
towards Zakharyaschev's canonical formulas and to view these formulas as
generalized Jankov-de Jongh formulas.Department of Computing, Imperial College, London.
(back to workshop)
David Gabelaia
Topology and Modal logic in Tbilisi
In this talk I will present a brief survey of research areas where
Modal Logic and General Topology meet. In particular, the talk will be
concerned with the two topological semantics for modal logic
(diamond-as-closure and diamond-as-derivative interpretations)
introduced by McKinsey and Tarski on the one hand and the
order-topological duality theory for S4-algebras on the other hand. I
will present a short overview of the results obtained and the methods
developed in these areas by Leo Esakia and his school in Tbilisi. I
will also present two new results - one about the C-logics generated
by subsets of the real line (joint work with Guram Bezhanishvili) and
another about the d-logics of the class of Stone spaces and the class
of compact Hausdorff spaces (joint work with Guram Bezhanishvili and
Leo Esakia).
Razmadze Mathematical Institute
(back to workshop)
Mai Gehrke
One-sided and two-sided completions: the relationship between presentations of dcpo algebras and canonical extensions
This talk reports on joint work with Jacob Vosmaer. The canonical extension of a lattice is in an essential way a two-sided completion. Domain theory, on the contrary, is primarily concerned with one-sided completeness. In this talk, after an introduction to one- and two-sided completions, we show two things. Firstly, that the canonical extension of a lattice can be given an asymmetric description in two stages: a free co-directed meet completion, followed by a completion by selected directed joins. Secondly, we show that the general techniques for dcpo presentations of dcpo algebras used in the second stage of the construction immediately give us the well-known canonicity result for bounded lattices with operators.
Institute for Mathematics, Astrophysics and Particle Physics, Radboud Universiteit, Nijmegen
(back to workshop)
Rosalie Iemhoff
Terms and quantifiers in nonclassical proofs
Skolemization, Herbrand's theorem, and the epsilon calculus are methods to compare quantifiers and terms in proofs. It has long been known that these methods, although applicable, do not behave well in a nonclassical setting. We present alternatives to the three mentioned methods that possess many of the useful properties that their classical counterparts have. We will apply the results to various constructive theories, such as the theory of equality, the theory of groups, and Heyting Arithmetic.
(Due to unfortunate circumstances this talked had to be cancelled.)
Department of Philosophy, Utrecht University.
(back to workshop)
D. Pataraia and Mamuka Jibladze
A new proof of total disconnectedness for compact Hausdorff boolean algebras
All known proofs of the fact stated in the title use the Peter-Weyl theorem which requires highly nontrivial use of measure theory, invariant integration, representation theory, etc.A proof avoiding all of this will be presented. Instead, the proof involves an interesting topological boolean algebra structure on the infinite-dimensional sphere previously considered by Gavin Wraith.
Mathematical Institute of the Georgian Acedemy of Sciences
(back to workshop)
Contributions
A.S. Andreyeva
Use of Non-Standard Lexical-Functional Correspondences in Machine Translation
Lexical functions (LFs) are used in systems of natural language processing, in particular in systems of machine translation, in several ways. The most important of them is providing proper translation equivalents. Information about LFs can be easily used for providing proper translation equivalents when corresponding word-combinations from two languages represent the same lexical function. But if an LF construction in one language does not correspond to the word-combination representing the same LF in the other language, information about lexical functions can still be used for providing proper translation equivalents. The paper describes cases when an LF construction corresponds to the LF construction representing the other LF or to the value of the other LF only. It gives the reasons for the existence of such correspondences between Russian and English and shows how these correspondences can be implemented in machine translation system.
Lomonosov Moscow State University.
(back to session)
Rusudan Asatiani
Functional Analysis of Conjunctions Showing Contrast in Georgian
Linguistic structuring of reality based on the notions `same-different' proceeds through `oppositions'. An opposition means that there are at least two items one of which is `marked' and another is `unmarked'. Structuring of information, its packaging, also proceeds through the foregrounding of such contrastive constituents: one part of information stands out against a background of the other part of information. As a result different formal models of various types of topics and focuses arise.The paper examines the role of conjunctions in the process of structuring of contrastive topics and/or focuses in Georgian. There are 8 conjunctions aramed?o?ond?tumca? magram?k'i//k'idev?xolo?tu?da in Georgian, which show different kinds of contrast. On the basis of functional-semantic analysis of the conjunctions seven binary differential semantic features are suggested: +/-[Contrast], +/-[Opposition], +/-[Correction], +/- [Expectation], +/-[Wish], +/-[At least one], +/-[Strong underlining]. Definite combinations of the features express main functions of the conjunctions, which are relevant for formal representations of contrastive topics and/or focuses during the unification of some constituents or sentences.
Institute of Oriental Studies.
(back to session)
Matthias Baaz and Agata Ciabattoni
On the limits of analiticity
In this lecture we present sufficient conditions for a first-order logic do not admit any analytic calculus of a reasonable kind.
TU Wien.
(back to session)
Anton Benz and Fabienne Salfner
Discourse Relations and Embedded Relevance Implicatures: Some Remarks and Observations about their Interdependencies
We look at the dependencies between rhetorical discourse structure and `relevance' implicatures. We follow Benz and v. Rooij (2007) and infer `relevance' implicatures from the assumption that an answer provides optimal information for solving an explicitly or implicitly given decision problem of the inquirer. Such a decision problem can be identified with a question raised in the conversation. Background questions not only depend on explicit questions under discussion but may also be raised by rhetorical relations such as CONCESSION or CONTRAST. In this talk, we are especially concerned with implicatures of embedded questions, and show by some examples that determining the rhetorical relation that connects two text spans and setting up a pragmatic model that explains the implicatures of embedded sentences interact with each other.
ZAS Berlin.
(back to session)
Marta Bilkova
On Normal Forms and Uniform Interpolants
The aim of this talk is to illustrate how uniform interpolants can be constructed naturally from suitable normal forms. We mainly consider modal logics, K being a basic example. In this case such normal forms lay in the hart of coalgebraic nature of modal logic. We will show how to use deep inference calculi to visualize construction of uniform interpolants via erasing atoms from a normal form, to do this we propose a new calculus for modal logic K. Next we revisit Pitts' proof of uniform interpolation for intuitionistic logic and discuss similar issues there.
Charles University in Prague, Department of Logic.
(back to session)
Elizaveta Bylinina (cancelled)
Wh-reduplication, possible worlds and distributivity
The focus of this work is the indefinite pronoun series formed by reduplication of wh-words. They are found in various languages and have very stable semantics typologically: they have narrow scope and tend to occur in non-specific contexts. To provide semantics for these items, we try to make use of the idea that reduplication and distributivity are connected. We use Landman's plurality theory of distributivity and show that treating predicates in intentional contexts as plural in their special way results in a good analogy with 'ordinary' distributivity. The sum/group shifting mechanism from Landman's theory gives us some of the particular language data for free, though some remains unaccounted for. Treating plurality of worlds as an instance of predicate plurality opens a promising view on non-specific indefinites and possible-world semantics in general: indefinites are plurals with distributive interpretation with respect to intentionally plural predicate.
Claudia Casadio and Mehrnoosh Sadrzadeh
Cyclic Pregroups and Clitic Movement in Persian and Italian
In this paper, we provide a general setting for clitic movement in Persian and Italian by adding a `clitic rule' to a pregroup, a calculus of (left and right) adjoints recently developed by J. Lambek as a non conservative extension of his Syntactic Calculus. This proposal is motivated by the non-commutative (additive, multiplicative) logic developed by Abrusci as a cyclic formulation of the Syntactic Calculus. We show that in this version of a pregroup, the types of the verbs with postverbal clitics are derivable from the types of the verbs with pre-verbal pronouns. Arguments are also presented for the fomulation of the `clitic rule' as a meta-rule (in the sense assumed in Lambek's pregroups) vrt. as a logical axiom of the calculus of a cyclic pregroup.
University `G. D'Annunzio', Chieti and Computing Laboratory, Oxford University.
(back to session)
Anna Chutkerashvili
The Functions of Discourse Clitics `mxolod' and `mart'o' in Georgian
The functions of the discourse clitics mxolod and mart'o are considered in the paper. These particles can both indicate to focus as well as the topic of the sentence. They often have the similar functions in discourse and can substitute each other, but they also have different meanings. In the paper these particles are analyzed in different types of sentences.
Tbilisi State University.
(back to session)
Sabine Gründer
Aspectual Shifts With and Without Type Conflict
The paper offers a new type of approach to the semantic phenomenon of adverbial aspect shift. It accounts for standard data that resisted a full theoretical explanation so far and deals with some challenging new data, moreover. The theory combines concepts from the fields of theoretical computer science (regular languages), artificial intelligence (hierarchical abstraction), philosophical logic (supervaluations), and theoretical linguistics (syntactic phases; aspectual viewpoint). According to the proposal, temporal prepositions are dynamic presuppositions that can aspectually underspecify a situational concept. The simple algorithm underlying it derives the correct set of possible interpretations on the basis of lexical semantic input only, and, furthermore, may claim cognitive plausibility.
Department of Computer Science, University of Leipzig.
(back to session)
Marine Ivanishvili and Ether Soselia
Preverbs in Megrelian
- The preverb system of Megrelian is much more complicated and distinctive than the Georgian one.
- In contrast to Georgian, compound preverbs in Megrelian are derived from the simple ones by adding following elements: ko-, c'o-, to-, no-, la-, ?a-, da-.
- The semantics of a compound preverb is not a simple sum of the constituents.
- Some phonetic changes are caused by adding either positive or negative particle-prefixes (ko- and va- correspondingly) to the preverbial forms.
- Combinations of preverbs and verbal vowel prefixes are put to the phonetic changes according to the general phonetic rules in Megrelian: a + i > e; o + o > a; o + i > i.
- Preverbs: a-, ga-, da-, c'a-, ?a- are regarded as Georgian borrowings, and it seems quite natural due to the mutual Zan-Georgian isoglosses.
Acad.G.Tsereteli Institute of Oriental Studies, Tbilisi.
(back to session)
Maarten Janssen and Olga Borik
Past Perfective and Telicity
It has often been argued that some languages conflate the categories of tense and aspect in so-called "aspectual" tenses, such as the perfective and imperfective past in Romance languages (e.g. de Swart 1998). As shown by Cipria and Roberts (2000), these tenses are not telic, and the terminativity of these tenses has to be modelled in a different way. We take this claim further by showing that tense in fact never logically entail termination, but only strongly 'suggests' it. In other words, termination with the past tense is not an entailment, but a (strong) pragmatic inference.
IULA, Universitat Pompeu Fabra and CTL Universitat Autonoma de Barcelona.
(back to session)
Mircea Marin and Besik Dundua
System Description: PRhoLog
We describe a system for rule based programming in Prolog. The system is called PRhoLog, and is based on the RhoLog calculus that we proposed recently for rule based programing.The computational model has a logic programming semantics where program clauses define strategies that act on sequences of trees (also known as hedges). Besides, it has an expressive pattern matching mechanism with logical variables for terms, hedges, function symbols, and term contexts.
The outcome is a highly declarative extension of Prolog that is expressive enough to support concise implementations for specifying and prototyping deductive systems, solvers for various equational theories, tools for querying and translating XML, evaluation strategies, etc.
University of Tsukuba and Johannes Kepler University.
(back to session)
Paul Meurer
A finite state approach to Abkhaz morphology and stress
Although Abkhaz features a daunting number of verbal prefixes and suffixes that may combine to form an inflected verb form, the Abkhaz verb has rather regular agglutinative (slot) morphology. There is little suppletivism, and there are only a few phonological processes, the most prominent of them being assignment of word stress. The stress pattern of an Abkhaz word form determines the orthographic (and phonetic) realization of Schwa. Thus, one has to take stress into account even if only orthographic forms are to be analysed or generated. The derivation of the stress pattern of Abkhaz word forms is complicated an still not fully understood. Arie Spruit, building on work by V. Dybo, gives a detailed account of Abkhaz stress and its relationship to morphology. More recently, Loren Trigo tried to refine Spruit's analysis.
In my talk, I will show how Spruit's and Trigo's analysis can be exploited to build a morphological analyser for Abkhaz in the finite state framework.
Aksis/Unifob.
(back to session)
Robert Samuel Ralph Myers
Applications of Tableaux for Multisorted Coalgebraic Modal Logic
We present generic tableau algorithms which build possibly multisorted coalgebras for endofunctors on Set^m and focus on some appli- cations. Two generic languages are considered: coalgebraic hybrid logic and a fragment of the coalgebraic mu-calculus. The hybrid language in- volves nominals, the global modality and the difference modality. It has applications in description logic and norm-governed multiagent systems. The fixpoint language may be thought of as a generalisation of Kleene's theorem, which associates deterministic automata with regular expres- sions. It has applications in automata synthesis, program analysis and reactive programming languages.
Imperial College London.
(back to session)
Sumiyo Nishiguchi
Neo-Davidsonian Extension of GL for Possessive Disambiguation
This paper proposes the elaboration of the qualia structure of the Generative Lexicon (GL) in Pustejovsky (1995) and the Extended Generative Lexicon theory (Lenci et al. 2000). I argue that qualia roles of nouns should be expanded to incorporate Neo-Davidsonian event semantics. My proposal is based on the Japanese genitive postposition no. The Japanese ``NP1-no NP2" construction expresses a wider range of relations between two entities than does the English possessive ``NP1's NP2," such that the Pustejovskian qualia roles encoded in NP2 do not supply the necessary relations between two entities, which Vikner and Jensen (2002) succeeded to certain degree. The genitive phrase modification detects event argument of nouns other than deverbal nouns discussed in Larson (1998).
Osaka University.
(back to session)
David Pearce, Pedro Cabalar and Sergei Odintsov
Axiomatic foundations for well-founded and partial stable semantics
Which logic yields an adequate foundation for the partial stable model semantics of Przymusinski (1991) and the well-founded semantics of Van Gelder, Ross and Schlipf (1991) in the sense that its minimal models (appropriately defined) coincide with the partial stable models of a logic program? Initial work on this problem was undertaken by Cabalar (2001) who proposed a frame-based semantics for a suitable logic which he called $HT^2$. Preliminary axioms for $HT^2$ were given in Cabalar et al (2006). In this paper we analyse $HT^2$ frames and identify them as structures of a logic $N^*$ having intuitionistic positive connectives and Routley negation and give a natural axiomatics for $HT^2$. We define a notion of minimal, total $HT^2$ model which we call partial equilibrium model and show that for logic programs these models coincide with partial stable models.
Universidad Politecnica de Madrid.
(back to session)
Michal Pelis and Ondrej Majer
Dynamic epistemic logic with questions
The aim of our article is to explore questions in the process of communication in a group of agents using the framework of dynamic epistemic logic. We start with representation of questions by means of the standard multiagent epistemic logic S5 and then we use the framework of public announcement logic to model information exchange in the process of asking questions in a group of agents. Our framework allows us to discuss complex processes like the flow of information in a group of agents seeking cooperatively an answer to a publicly announced question.
Institute Of Philosophy, Academy of Sciences of the Czech Republic.
(back to session)
Konstantine Pkhakadze, George Chichua, Lasha Abzianidze, Aleksandre Maskharashvili, Aleksandre Vashalomidze, Merab Chiqvinidze and Nika Pkhakadze
Toward Complete Mathematical and Mechanical Foundation of the Georgian Language and Thinking
We present the basic results of our researches for the complete mathematical and mechanical foundation of the Georgian Language and Thinking (GLT). Since 2003, the researches go under the theoretical and technological aims declared by the State Priority Program (SPP) `Free and Complete Inclusion of a Computer in the Georgian Natural Language System'. By now, we are working out the project `Foundations of Mathematical Theory of the Georgian Language and Thinking and Sub-systems of the 1-version of the Voice Managed Georgian Intellectual Computer System', which is funded by St. Andrew the First Called Georgian University of The Patriarchy of Georgia. This project is a current subproject of the SPP whose theoretical bases will be the main themes of our paper. More strictly, in the paper, we will concentrate on our lingual ideology, which we call as Georgian Lingual Ideology (GLI) and our new theoretical approaches, which are elaborated for the aim to create Georgian Reader-Listener System with User's Possibility to Build in an Own Synthetic Voice (GeoRLSwithUV).
I.Vekua Institute of Applied Mathematics of TSU.
(back to session)
Mikheil Rukhaia
Second Order Cut-Elinimation
This paper is about implementation of the second order cut elimination. We define LKII, second order calculus, extending LK with second order quantifier introduction rules. The goal was to extend CERES system for the LKII-proofs, but method CERES cannot be extended to the second order calculus in the straightforward way. We decided to extend Gentzen's method to LKII. We extended reduction rules for the second order quantifiers. We implemented reduction rules and specific algorithm to apply reduction rules to the LKII-proof and integrated it into the CERES system. So, system was extended to the second order calculus.
Vienna University of Technology.
(back to session)
Kerstin Schwabe and Robert Fittler
On the interpretations of embedded questions in German
The paper presents the semantic properties of German matrix predicates – the con¬sistency conditions – which explain the different interpretations of ob-forms [cf. F weiß/sagt, ob M kommt `F knows/tells whether ?' vs. F bedenkt ob ? `F considers whether ?'] as well as of wh-forms [cf. F weiß/sagt, wer kommt `F knows/tells who is coming' vs. F bedenkt (es), wer kommt `F considers (it) who is coming']. Verbs like wissen/sagen exhibit the external ob-form and the exhaustive wh-form since they are objective. Verbs like es bedenken/darüber nach-denken, on the other hand, only exhibit the internal whether-form and the non-exhaustive wh-form. The internal ob-form is only allowed if the consistency conditions do not exclude tautologies, the non-exhaustive wh-form is only possible if the consistency conditions license the internal ob-form or a factive interpretation together with an appropriate correlate.
ZAS Berlin and Mathematisches Institut, FU Berlin.
(back to session)
Luca Spada
Advances in the theory of fixed points in many-valued logics
This is a work in progress addressing the problem of nested fixed points in Lukaisewicz logic.
Department of Mathematics and Computer Science, University of Salerno.
(back to session)
Maria Spychalska
Experimental study on reasoning with quantifiers ``most'' and ``some''
We report on our experimental work concerning reasoning in natural language with quantifiers ``some'' and ``most'', with a special regard to scalar implicatures. We analyzed people's readiness to accept given direct inferences as correct. Our results included (1) high frequency of scalar implicatures of ``most'' and ``some'' in natural reasoning (2) higher frequency of accepting ``some'' as a conclusion from ``most'' than from ``all'' (3) higher frequency of evaluation ``Some A's are not B'' than ``Most A's are not B'' as a correct conclusion from ``Some A's are B''.Based on our results we propose how scalar implicatures of quantifiers ``some'' and ``most'' can be modeled with the use of fuzzy semantics.
University of Warsaw, Institute of Philosophy.
(back to session)
Tomoyuki Suzuki, Alexander Kurz and Emilio Tuosto
Nominal Kleene Algebras and Nominal Regular Languages
We study how to lift the classical theory of regular languages to the nominal case and define nominal regular languages. More precisely, we generalise Kleene algebras to nominal sets. This generalisation allows us to use the framework of History-Dependent automata HDA to characterise the regular expressions induced by the class of nominal Kleene algebras. In fact, HDA constitute a trait d'union between regular languages associated with nominal Kleene algebras and the resource aware computations as e.g. those expressed by the so-called nominal calculi.This research paves the way towards a new theory of formal languages with binders suitable for representing resource-aware computations.
The main contributions of this work are as follows. Firstly, we give the definition of the class of nominal Kleene algebras and of corresponding class of regular expressions and nominal regular languages. Secondly, we show the correspondence between regular expressions on nominal Kleene algebras and HDA. Finally, we characterise the class of nominal regular languages accepted by finite HDA.
University of Leicester.
(back to session)
Lucia M. Tovena
The imperfect measure of internally plural events
In this paper, we explore the hypothesis that internal pluractional verbs describe events as not being canonical, namely either as not full-blown or in excess of something, because the measure of one dimension of the event is explicitly made inappropriate for the purpose.
Université Paris VII.
(back to session)
Maria Tsyurupa (cancelled)
Modifier of Resultative State Duration
In this paper we discuss one of the verbal actionality parameters - modifier of resulting state duration (MRSD) in Russian. Modifier of duration (for an hour) and modifier of time of completion (in an hour) have been discussed in numerous research papers, but MRSD has recieved less attention. In part 2 of this paper, semantics of the modifier is discussed, as well as its aspectual impact: shifting the focus from the process to the resultative state, MRSD makes the whole VP momentative. In parts 3-4, we discuss some semantic restrictions the verb must meet to be compatible with MRSD. These restrictions are not purely thematic or aspectual.They are best accounted for if we assume that MRSD is in fact three different constructions.
Most examples the research is based on are from the Russian national corpus (ruscorpora.ru). Others are constructed for elicitation purposes and tested on native speakers.
VINITI RAS (All Russia Institute for Scientific and Technical Information).
(back to session)
Levan Uridia
Boolean Modal Logic wK4B - Doxastic Interpretation
We consider the boolean modal logic wK4$_B$ for which we prove topological and kripke completeness results. The main result is to show that wK4$_B$ is expressively equivalent to the extended language with common and distributed belief operators.
Universidad Rey Juan Carlos.
(back to session)
Nikolas Vaporis
The extension property up to $n$ and $T_n$-projectivity
Our contribution is a refinement of a well-known theorem of Ghilardi which establishes the equivalence between projectivity and the extension property. We first define the notion of $T_n$-projectivity by changing the underlying logic in the definition of projectivity from IPC to $T_n$, where $T_n$ are the Gabbay-de Jongh intermediate logics of finite $n$-ary Kripke tree frames. Then, we prove that a propositional formula is $T_n$-projective if and only if it has the extension property up to $n$, the last is a refinement of the extension property defined by Iemhoff.
University Utrecht.
(back to session)
Michal Walicki, Marc Bezem, and Clemens Grabmayer
Graph Kernels, Logic and Choice Axioms
Our contribution is threefold. First, we establish a link between the existence of kernels of digraphs and the satisfiability of corresponding theories of propositional logic. This correspondence relates infinitely branching digraphs to infinitary propositional logic, and finitely branching ones to the usual propositional logic. Second, for three set-theoretical axioms concerning different forms of `choice' we present reformulations in terms of digraph kernel existence: for the axiom of choice, the axiom of countable choice, and Weak Koenig's Lemma we give equivalent statements over the subsystem RCA of second-order arithmetic in terms of the existence of kernels for special classes of digraphs. Third, we show that the existence of kernels of recursive graphs is Sigma-1-1-complete. As a consequence the existence of kernels of recursive, finitely-branching trees is not provable in RCA.
Bergen University and Utrecht University.
(back to session)
Thomas Wier
Topic, Focus and the (Non)configurational Properties of Georgian Syntax
Georgian word order is traditionally described as 'free', whereas in fact there are two major constraints on it: clause boundaries and discourse structural effects of topic and focus. I will also examine how infinitival constructions based on nominalizations (masdars, -isa-s gerundives, and purposive infinitives based on the future participle) confound common nouns of word order in Georgian.
University of Chicago.
(back to session)
Yoad Winter and Eric Reuland
Binding without identity: Reference by proxy and the functional semantics of pronouns
Binding without Identity: reference by proxy and the functional semantics of pronouns One of the well-known properties of reflexive pronouns is their ability to have `proxy readings'. This is illustrated in (1) (Jackendoff 1992):
- (Upon a visit in a wax museum:) All of a sudden Ringo started undressing himself.
OTS, Utrecht University.
(back to session)
Grégoire Winterstein
Optional Additivity and Exhaustivity
This work deals with the optional nature of too in certain examples. We claim that although this appears as a counter-example to recent accounts of too in terms of obligatory presupposition, this can rather be explained by the interplay between these analyses and the optional exhaustive interpretation of an utterance. We further motivate our claim with empirical evidence. We then briefly look at the consequences our claim has on theories of exhaustification and on the relationship between additive and adversative markers.
LLF - Université Paris Diderot -- Paris 7.
(back to session)
Henk Zeevat
Bayesian Interpretation
The paper makes the case for applying bayes' theorem to the interrpetation of NL utterances. It discusses the repercussions for pragmatics, language production and language evolution.
ILLC/Universiteit van Amsterdam.
(back to session)
Dragisa Zunic, Silvia Ghilezan, Pierre Lescanne and Jelena Ivetic
Intuitionistic sequent-style calculus with explicit structural rules
In this paper we extend the Curry-Howard correspondence to the intuitionistic sequent calculus with explicit weakening and contraction. We study a system derived from lambda-Gentzen calculus of Espirito Santo by adding explicit operators for weakening and contraction, which we call linear-lambda-Gentzen and which contains only linear terms. For the proposed calculus we introduce the type assignment system with simple types. The presented system has a natural diagrammatic representation, which can be used for proving the subject reduction property. We prove the strong normalisation property by embedding our calculus into the simply typed lambda-lxr calculus of Kesner and Lengrand.
University Business Academy, Faculty of Engineering, University of Novi Sad, and Ecole Normale Superieure de Lyon.
(back to session)
