Naslov
Formal prototypicality assessment model and its application to morphological level of language processing

Autori
Galetić, Vedran ; Kelić, Maja ; Pavlinušić, Eva

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, ostalo, znanstveni

Skup
2nd Summer School of the International Max Planck Research School on Neuroscience of Communication: Function, Structure, and Plasticity

Mjesto i datum
London, Ujedinjeno Kraljevstvo, 06.07.2011. - 08.07.2011

Vrsta sudjelovanja
Poster

Vrsta recenzije
Nije recenziran

Ključne riječi
Cognitive Linguistics ; Prototype Theory ; acquisition of Croatian as L1 ; Croatian as L2 ; verbal morphology ; formal evaluation of linguistic competence

Sažetak
Cognitive Linguistics' theory of prototypicality stands for an important property of language, whether it be observed from the aspect of linguistic phenomena (such as polysemous lexical categories) or the organisation of space of linguistically expressive concepts as a cognitive model of reality. Prototypical organisations arise from language usage and experience through living in certain geographic surroundings. While designing various tests of infants' language development and educational aids for foreign language learners, it would be exceptionally beneficial if exhibited linguistic errors could be formally classified and evaluated. This work brings forth a formal methodology framework for designing a computationally amenable representation of prototypically organised conceptual space, thereby providing a means of enabling a formal analysis of various linguistic phenomena. A formal model is proposed that allows for assessment of category members' level of representativeness within a conceptual space. The model is primarily conceived for application on the semantic level covering various complexities, from lexemes to complex semantic structures. Concepts are represented as vectors in vector spaces spanned by basis vectors corresponding to characteristic semantic constituents. Categories are modelled as fuzzy sets, implying that the concepts' membership is not of binary character, but rather quantified with respect to contributions of their semantic specifications. The model can also be adjusted and applied to other levels of linguistic processing, such as syntax, morphology and phonology. Here we present the application of the model for formal classification and evaluation of specific errors within the process of acquisition of morphological properties of Croatian present tense.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo, Filologija

POVEZANOST RADA