Low-level situational cognitive models within the Lexical Constructional Model and their computational implementation in FunGramKB

  1. Ruiz de Mendoza Ibáñez, Francisco José
Libro:
Language processing and grammars: The role of functionally oriented computational models

ISBN: 9789027259158

Año de publicación: 2014

Páginas: 367-390

Tipo: Capítulo de Libro

DOI: 10.1075/SLCS.150.15IBA GOOGLE SCHOLAR

Resumen

This paper investigates the notion of low-level situational cognitive model, its role in linguistic description and its possible computational treatment in the knowledge base FunGramKB. Low-level situational models are exploited metonymically to produce situation-based implicatures. When such inferences become stably associated with a formal pattern, they give rise to implicational constructions. Other kinds of construction make use of different kinds of cognitive model. For example, argument-structure constructions are based on high-level non-situational cognitive models. The paper then provides a typology of low-level situational cognitive models, which can be roughly equated with Schank and Abelson’s now classical notion of script. Schank and Abelson’s classification into situational, personal and instrumental scripts is revised and refined to make it include a further division into simple, complex and composite scripts. Simple scripts capture sequences of actions, while complex scripts consist of chained sequences of subscripts and composite scripts are combinations of independent subscripts. Since scripts are cases of procedural knowledge, which is included in the so-called cognicon of FunGramKB, the paper explores the incorporation of this typology into the architecture of this part of the knowledge base. This incorporation is argued to endow the cognicon with greater descriptive parsimony, which results in a more efficient computational implementation. Finally, the paper uses the FunGramKB representation metalanguage COREL as an adequate way of supplying precise descriptions in terms of event structure variables and their sequencing, which turns out to be useful to enhance descriptive adequacy in the linguistic model itself.