[PS-3.39] How distributed semantic knowledge gives rise to a network of object nouns?

Kaoutar, S.

LIST Laboratory, FST, Abdelmalek Essaadi?s University Tangier, Morocco

We propose a model of lexical semantic memory based on three main concepts: emergence, distribution and integration. The model assumes that: 1)- lexical (i.e. word forms) and semantic aspects (i.e. the mental image of objects/actions the words refer to) of words emerge from distinct brain networks (Sporns, 2010), 2)- Semantic aspects are represented by distributed semantic knowledge organized in different brain areas (Martin, 2007) and 3)-Lexical aspects of words are linked to their semantic aspects through convergence zones (Damasio et al. 1989; Patterson et al. 2007) or simulation without the need of convergence zones (Barsalou 2008; Hauk et al. 2004). The presentation of an object noun (eg. cat) as stimulus evokes the activation of its meaning represented as a collection of semantic features (eg has_fur, has_tail, has_legs, Meows). As many object nouns share similar semantic features (eg. dog, horse, zebra, donkey), each feature can links an arbitrary number of object nouns giving rise to a network of object nouns (eg. Has_leg links all four-legged animals).The analysis of the network of object nouns through graph theory (eg. clustering coefficient, path length, and centrality) provides a powerful tool to understand how the brain computes lexical semantic cognitive operations.