Logic, Language, and Computation

Frontmatter

A Computational Grammar for Georgian

Abstract

In this paper, I give an overview of an ongoing project which aims at building a full-scale computational grammar for Georgian in the Lexical Functional Grammar framework and try to illustrate both practical and theoretical aspects of grammar development. The rich and complex morphology of the language is a major challenge when building a computational grammar for Georgian that is meant to be more than a toy system. I discuss my treatment of the morphology and show how morphology interfaces with syntax. I then illustrate how some of the main syntactic constructions of the language are implemented in the grammar. Finally, I present the indispensable tools that are used in developing the grammar system: fst; the xle parsing platform, the LFG Parsebanker, and a large searchable corpus of non-fiction and fiction texts.

Paul Meurer

The Structure of Rigid Frames of Depth 3 Only

Abstract

In this paper we classify all rigid rooted IPC-frames of depth 3. Among other things we show that these have at most 3 maximal elements. The interest in rigid frames arose from the paper [5]. In this paper quasi-characterizing inference rules were discussed. These rules are built on the pattern of Jankov-formulas of finite rooted frames but a Jankov-formula of the form ϕ→p is transformed into a quasi-characterizing rule ϕ/p. Such a rule is called self-admissible if it is admissible in the logic generated by the frame corresponding to the rule itself. The important results of [5] are that self-admissible rules are admissible in IPC itself, and that such a quasi-characterizing inference rule is self-admissible iff the frame it derives from is not rigid. The classification of rigid frames thus becomes of interest.

Tahsin Oner, Dick de Jongh

Three Kinds of Event Nominal Negation in Russian

Abstract

This paper is about Russian event nominals, their negation and their meaning under negation. The paper shows that there are three ways to (semantically) combine negative marker with an event nominal, depending on the meaning of the nominal itself and requirements of the context. The main factors driving these differences can be shown to be of an aspectual nature; both aspectual characteristics of the initial nominal and the (contextually-driven) properties of the resulting nominal are important. As for the range of possibilities, one can find negated stative nominals derived from positive stative nominals, negated stative nominals derived from non-stative telic nominals, and negated non-stative telic nominals derived from non-stative telic nominals. These three types of negated nominals differ not only aspectually, but also with respect to where the negation is attached and how the denotation of the whole nominal is evaluated.

Anna Pazelskaya

A Dynamic Conceptual Model for the Linguistic Structuring of Space: Georgian Preverbs

Abstract

For structuring of space relations in the Georgian language three dimensions are valuable: 1. Point of View (speaker’s or teller’s position); 2. Geographic Space (various directions and distance dichotomy); 3. Communicational Space (Ego and Alter Spaces). ‘Point of View’, ‘Ego space’ and ‘Distance dichotomy’ are flexible: They can be changed according to the speaker’s (or teller’s) attitude, while abstract relations of ‘Geographic Space’ are stable. Various combinations of the dimensions are represented in Georgian by the preverbs: There are 9 simple and 7 complex preverbs. The paper proposes a dynamic conceptual model of space structuring for Modern Standard Georgian and examines the possibilities of its linguistic representation.

Rusudan Asatiani

even in Horn Space

Abstract

In the first part of the paper, I argue that current pragmatic theories of NPI licensing fail to capture the distinction between strong and weak NPIs. Specifically, I will show that an analysis in terms of covert even alone can not account for the limited distribution of strong NPIs. In the second part, I investigate the implicatures of even sentences in weak licensing contexts. I show that they give rise to a minimal-achievement implicature which can be used to derive the markedness of strong NPIs in weak licensing contexts.

Regine Eckardt

On -c and ḳi Particles in Georgian

Abstract

This paper considers the Georgian particles ḳi and -c. These particles are frequently used, separately as well as together. They have different meanings, but both of them can have a focusing function: emphasizing a word or a phrase they are attached to. In spite of having similar or even the same semantic features, the particles ḳi and -c cannot substitute for each other. One reason for this is that -c is a bound form and ḳi is not. They never substitute for each other but they very often occur together and they are much more emphatic when they are used together. The dominating element in building up the meaning of -c ḳi is -c, which is stronger in emphasis. -c ḳi is used to emphasize something unexpected or surprising. These particles are rendered in English by ‘even’.

Anna Chutkerashvili

Dealing with Polysemy in Russian National Corpus: The Case of Adjectives

Abstract

The paper describes a research carried out in the Russian National Corpus project (www.ruscorpora.ru). We discuss a method of word sense disambiguation, which is now being applied to polysemous adjectives in the RNC. The approach implies formulating rules to select the appropriate sense of the adjective by using co-occurrence restrictions observed in the corpus. The disambiguating filters operate with various kinds of grammatical and semantic information on the adjectives and the nouns modified. Our results demonstrate that the semantic filters are effective for WSD.

Ekaterina V. Rakhilina, Tatiana I. Reznikova, Olga Yu. Shemanaeva

Inquisitive Semantics: Two Possibilities for Disjunction

Abstract

We introduce an inquisitive semantics for a language of propositional logic, where the interpretation of disjunction is the source of inquisitiveness. Indicative conditionals and conditional questions are treated on a par both syntactically and semantically. The semantics comes with a new logical-pragmatical notion which judges and compares the compliance of responses to an initiative in inquisitive dialogue.

Jeroen Groenendijk

Implicatures of Irrelevant Answers and the Principle of Optimal Completion

Abstract

In this paper, we present a game–theoretic account of a subclass of ‘relevance’ implicatures arising from irrelevant answers. We show that these phenomena can be explained if we assume that interlocutors agree on production and interpretation strategies that are robust against small ‘trembles’ in the speaker’s production strategy. In this context, we argue for a new pragmatic principle which we call the principle of optimal completion. We also show that our model provides a parallel account of scalar implicatures which removes some limitations of previous accounts.

Anton Benz

Conceptualization of Pain: A Database for Lexical Typology

Abstract

The paper presents a study in lexical typology. We focus on the semantic domain of pain as one of the most universal and complex areas of human experience. The predicates of unpleasant bodily sensations are compared in a sample of 23 languages. The collected material demonstrates that the use of pain verbs is dependent on the range of factors of different nature. This data heterogeneity poses the problem of cross-linguistic comparability of pain predicates. As a way to overcome this problem, we propose the construction of a typological database. The multidimensional classifications implemented in the database allow for various cross-linguistic generalizations on pain and human body conceptualizations as well as on regularities of semantic shifts in different languages.

Anastassia Bonch-Osmolovskaya, Ekaterina Rakhilina, Tatiana Reznikova

Expressing Belief Flow in Assertion Networks

Abstract

In the line of some earlier work done on belief dynamics, we propose an abstract model of belief propagation on a graph based on the methodology of the revision theory of truth. A set of postulates is proposed, a dynamic language is developed for portraying the behavior of this model, and its expressiveness is discussed. We compare the proposal of this model with some of the existing frameworks for modelling communication situations.

Sujata Ghosh, Fernando R. Velázquez-Quesada

The Computational Complexity of Quantified Reciprocals

Abstract

We study the computational complexity of reciprocal sentences with quantified antecedents. We observe a computational dichotomy between different interpretations of reciprocity, and shed some light on the status of the so-called Strong Meaning Hypothesis.

Jakub Szymanik

Quantifiers in Japanese

Abstract

This paper argues that the generalized quantifier theory does not directly apply to Japanese quantifiers because of the following reasons: (i) the number of noun phrase (NP) arguments is underspecified and (ii) quantities are often expressed by predicative adjectives. It further claims that the word order changes the interpretation. For example, non-split quantifiers correspond to definite NPs that are unique in the domain of discourse, while split NPs are wide-scope indefinites. Adjectival quantifiers are polymorphic, and continuation-based combinatory categorial grammar [1] accounts for different meanings between (non)split quantifiers.

Sumiyo Nishiguchi

Exceptional Scope as Discourse Reference to Quantificational Dependencies

Abstract

The paper proposes a novel solution to the problem of exceptional scope (ES) of (in)definites, exemplified by the widest and intermediate scope readings of the sentence Every student of mine read every poem that a famous Romanian poet wrote . We propose that the ES readings have two sources: (i) discourse anaphora to particular sets of entities and quantificational dependencies between these entities that restrict the domain of quantification of the two universal determiners and the indefinite article; (ii) non-local accommodation of the discourse referent that restricts the quantificational domain of the indefinite article. Our account, formulated within a compositional dynamic system couched in classical type logic, relies on two independently motivated assumptions: (a) the discourse context stores not only (sets of) individuals, but also quantificational dependencies between them, and (b) quantifier domains are always contextually restricted. Under this analysis, (in)definites are unambiguous and there is no need for special choice-functional variables to derive exceptional scope readings.

Adrian Brasoveanu, Donka F. Farkas

Satisfaction and Friendliness Relations within Classical Logic: Proof-Theoretic Approach

Abstract

We present the logical friendliness relation in a proof-theoretic fashion as sequent system \(\boldsymbol{F}\). Then, the completeness theorem is proved. On the way to this theorem, we characterize the notion of satisfiability with respect to the classical two-valued semantics, in a proof-theoretic manner as system \(\boldsymbol{S}\), so that the latter becomes part of the definition of system \(\boldsymbol{F}\). Also, we obtain the strong compactness property for friendliness as a corollary of our completeness theorem.

Alexei Y. Muravitsky

Identification through Inductive Verification

Application to Monotone Quantifiers

Abstract

In this paper we are concerned with some general properties of scientific hypotheses. We investigate the relationship between the situation when the task is to verify a given hypothesis, and when a scientist has to pick a correct hypothesis from an arbitrary class of alternatives. Both these procedures are based on induction. We understand hypotheses as generalized quantifiers of types \(\left\langle 1\right\rangle\) or \(\left\langle 1,1\right\rangle\). Some of their formal features, like monotonicity, appear to be of great relevance. We first focus on monotonicity, extendability and persistence of quantifiers. They are investigated in context of epistemological verifiability of scientific hypotheses. In the second part we show that some of these properties imply learnability. As a result two strong paradigms are joined: the paradigm of computational epistemology (see e.g.[6,5] ), which goes back to the notion of identification in the limit as formulated in [4], and the paradigm of investigating natural language determiners in terms of generalized quantifiers in finite models (see e.g.[1]).

Nina Gierasimczuk

Enlarging the Diversity of Valency Instantiation Patterns and Its Implications

Abstract

In the prototypical case, arguments (=actants) are directly subordinated to their predicates and occupy positions of the subject and direct or indirect object. Valency slots filled in this way are called active. In non-prototypical cases, arguments can syntactically subordinate their predicate (passive valency slots) and even have no immediate syntactic link with it (discontinuous valency slots). These types of valency slots are mostly characteristic of adjectives, adverbs and nouns. A number of linguistic concepts are related, directly or indirectly, to the notion of actant. However, usually only prototypical – active – valency instantiation is taken into account. If one includes into consideration passive and discontinuous valency slot filling, the area of actant-related phenomena expands greatly. We discuss some of these phenomena and show that the notions of diathesis and conversion require broader generalization.

Igor Boguslavsky

The Modal Formula (†) $\square \diamondsuit p \supset \square \diamondsuit \square \diamondsuit p$ Is Not First-Order Definable

Abstract

The formula (†) above is shown not to be first-order definable. The result is obtained by complicating the construction introduced in [4]. Two motivations are given for why the question of the first-order definability of (†) matters, one from theoretical considerations relating to modal logic, the other from applications of modal logic to philosophy, namely logic of ability. Finally a comparison with a cognate notion in the literature is given.

Ali Karatay

Semantic Characterizations of German Question-Embedding Predicates

Abstract

The paper investigates the conditions under which German two-place verbs like wissen dass ‘know’ and bedauern dass ‘regret’ embed interrogatives. We present a necessary and sufficient condition for a dass-verb to have an ob-form. The corresponding verbs we call objective. An objective verb has a wh-form (F weiß, wer kommt ‘F knows who is coming’) if it satisfies a further condition stating that it has to be consistent with wissen dass. A non-objective dass-verb does not have an ob-form, but it can have a wh-form if it permits a da- or es-correlate and meets par ticular consistency conditions which render it factive or cognitive in the presence of the correlate (cf. bedauern ‘regret’ vs. annehmen ‘assume’) It turns out that the meaning of the wh-form of non-objective verbs deviates distinctly from the meaning of the wh-form of objective verbs. Unlike other approaches our rules are general and hold without exceptions.

Kerstin Schwabe, Robert Fittler

Processing Definite Determiners: Formal Semantics Meets Experimental Results

Abstract

Experiments on the online processing of linguistic utterances provide information about language processing in the first instance, and only indirectly about linguistic knowledge, while it has been linguistic knowl edge, and not linguistic processing, that has been the subject matter of theoretical linguistics. So how can such evidence be relevant to theoretical linguistics? Or how can linguistic theory inform a theory of language processing? – This issue is discussed here with respect to the processing and the formal semantics of the English definite determiner. I argue that the meaning of the definite determiner, as it shows up in experiments on online comprehension, can actually be accounted for in an incremental variant of current formal semantics.

Peter Bosch

Terminal Sequence Induction via Games

Abstract

In this paper we provide an alternative proof of a fundamental theorem by Worrell stating that the (possibly infinite) behaviour of an F-coalgebra state can be faithfully approximated by the collection of its finite, n-step behaviours, provided that F:Set→Set is a finitary set functor. The novelty of our work lies in our proof technique: our proof uses a certain graph game that generalizes Baltag’s F-bisimilarity game. Phrased in terms of games, our main technical result is that behavioural equivalence on F-coalgebras for a finitary set functor F can be captured by a two-player graph game in which at every position a player has only finitely many moves.

Clemens Kupke

Dutch Grammar and Processing: A Case Study in TLG

Abstract

The aim of this paper is to see type logical grammar (TLG) at work on an interesting linguistic case: the incremental processing of Dutch subordinate clause word order, namely the so-called cross-serial dependencies. With the help of proof net machinery adapted for the continuous and discontinuous Lambek calculus we are able to account for the increasing unacceptability of cross-serial dependencies with increasingly multiple embeddings.

Glyn Morrill, Oriol Valentín, Mario Fadda

Backmatter