LexIt: A Computational Resource on Italian Argument Structure (original) (raw)
Related papers
The goal of this paper is to introduce T-PAS, a resource of typed predicate argument structures for Italian, acquired from corpora by manual clustering of distributional information about Italian verbs, to be used for linguistic analysis and semantic processing tasks. T-PAS is the first resource for Italian in which semantic selection properties and sense-in-context distinctions of verbs are characterized fully on empirical ground. In the paper, we first describe the process of pattern acquisition and corpus annotation (section 2) and its ongoing evaluation (section 3). We then demonstrate the benefits of pattern tagging for NLP purposes (section 4), and discuss current effort to improve the annotation of the corpus (section 5). We conclude by reporting on ongoing experiments using semiautomatic techniques for extending coverage (section 6).
Argument structures, verb patterns and dictionaries
Atti Del Xii Congresso Internazionale Di Lessicografia Torino 6 9 Settembre 2006 Vol 2 2006 Isbn 88 7694 918 6 Pags 1169 1180, 2006
This paper addresses the following two questions: 1) how much can research on the interactions between the syntactic behaviour of verbs and their lexical semantic properties be relevant from a lexicographic point of view?; 2) how far can the integration of lexicological research and lexicographic practise go in this respect? After pointing out some of the main difficulties that theoretical studies still confront, I discuss concrete problems that arise when valency-based models are adopted in the presentation of specific verb classes in Italian monolingual dictionaries. With the help of the analysis of these specific cases, I intend to draw conclusions that are valid from a general perspective.
Toward Disambiguating Typed Predicate-Argument Structures for Italian
English. We report a word sense disambiguation experiment on Italian verbs where both the sense inventory and the training data are derived from T-PAS, a lexical resource of typed predicate-argument structures grounded on corpora. We present a probabilistic model for sense disambiguation that exploits the semantic features associated to each argument position of a verb.
Lexical Approaches to Argument Structure
2014
"This paper compares various approaches to argument structure. We start out presenting the lexical proposal that we want to defend in this paper. We then introduce phrasal proposals that are common in Construction Grammar. A historical section describes the oscillation between early lexical proposals in Categorial Grammar, phrasal approaches in phrase structure grammar (early Transformational Grammar, GPSG) back to lexical approaches in HPSG and Minimalism. We argue that there were good reasons for returning to lexical models and that the respective issues are not addressed in phrasal approaches. We go on discussing approaches that assume that sematnically compatible verbs are inserted into phrasal constructions and point out that lexical specification of valence plays an important role in various levels of description and that phrasal models can not account for this. A similar criticism applies to so-called Neodavidsonian approaches, which are discussed in a separate section. We will show that certain relations between constructions cannot be captured with inheritance or unification but require transformations or lexical rules, and hence in non-transformational syntax the lexical approach is the only option. Three sections are devoted to arguments from language acquisition, psycholinguistics, and statistics. We show that contrary to frequent claims the respective experiments do not provide evidence for phrasal constructions. We conclude that argument structure properties should be represented together with lexical items. "
2019
The annual conference CLIC–it (''Italian Conference on Computational Linguistics'') is an initiative of the ''Italian Association of Computational Linguistics'' (AILC – www.ai-lc.it) which is intended to meet the need for a national and international forum for the promotion and dissemination of high-level original research in the field of Computational Linguistics (CL), with particular emphasis on Italian. The volume gathers the Proceedings of the ''Third Italian Conference on Computational Linguistics'' (CLiC–it 2016), held in Naples on 5-6 December 2016. The CLiC–it 2016 papers cover a wide range of topics in the area of computational linguistics and natural language (both written and spoken) processing, by targeting state–of–art theoretical results, experimental methodologies, technologies and application perspectives, and by addressing challenges, open issues and new perspectives related to current and novel trends of the discipline
Clips, a Multilevel Italian Computational Lexicon: a Glimpse to Data
CLIPS is a multi-layered Italian computational lexicon based on the PAROLE-SIMPLE model. In this paper we briefly recall the main characteristics of the model and devote our attention to issues emerging from the encoding of large quantities of data, especially in relation to those types of syntactic and semantic information specific to our lexicon and that reflect innovative features of the underlying model. At syntactic level, we show how alternating structures may be encoded in a linguistically more elegant way by using framesets. We illustrate the connection between syntactic and semantic information, and show how the SIMPLE Italian lexicon approach to predicate selection has been refined in CLIPS. At semantic level, we illustrate the richness of information types encoded in a word sense description and the way such a wealth of data can be exploited. We stress in particular the expressive power of the Extended Qualia Structure yet mentioning some of its problematic aspects. We show that queries on qualia relations allow to retrieve lexical collocates, to extract domain specific information, semantic networks, and help interpreting modifying PPs in complex nominals. Finally, we show that features, which cut across the type hierarchy, have a stronger expressive power with respect to semantic types in identifying selectional preferences.
Categorial Type Logic meets Dependency Grammar to annotate an Italian Corpus
In this paper we present work in progress on the annotation of an Italian Corpus (CORIS) developed at CILTA (University of Bologna). We induce categorial type assignments from a dependency treebank (Torino University treebank, TUT) and use the obtained categories with annotated dependency relations to study the distributional behavior of Italian words and reach an empirically founded part-of-speech classification.