The study presents the preliminary computational simulations, which are carried out within a broader research (Rosi in preparation) about the acquisition of Tense-Aspect system in Second Language Acquisition. The research aims to contribute to the debate about the acquisition of Tense-Aspect system in Second Language Acquisition by comparing the development of Aspect morphology and verb semantics in Italian L2 between human learners and computational simulations as neural networks, namely Self-Organizing Maps (Kohonen 2001). The comparison intends to shed light on the cognitive principles and mechanisms that guide learners to acquire linguistic structures in a second language. The literature on Tense-Aspect acquisition in First and Second Language, henceforth L1 and L2, (Andersen and Shirai 1994; Bardovi Harlig 2000) has documented the interaction between the morphological category of Aspect and the semantic category of Actionality in the development of the interlanguage of learners, who initially tend to restrict Aspect forms to specific Actionality classes and, then, gradually spread the Aspect markers following implicational developmental stages from most cognitively salient to least cognitively salient associations between Aspect and Actionality (fig.1.1). Recent surveys (Li and Shirai 2000; Li 2005) interpret the initial interaction between Aspect and Actionality in L1 Acquisition as the result of the children’s analyses of the co-occurrence probabilities between morphological forms and semantic value of predicates in the linguistic input and verify this hypothesis in a connectionist model. This study intends to test empirically if a neural network can display the acquisitional pattern of Aspect morphology and the interaction in this development between Aspect and Actionality also in L2 Acquisition, in order to help understanding, on the one hand, the correlations between frequency effects and cognitive mechanisms, on the other, the role of the prior linguistic knowledge of L1 in L2 Acquisition.
The acquisition of verb semantics and morphology in Italian as L2: toward computational simulations / Rosi, Fabiana. - (2006), pp. 226-238. (Intervento presentato al convegno European Summer School of Language, Logic, and Information tenutosi a Università di Malaga, Spagna nel 31 luglio - 11 agosto 2006).
The acquisition of verb semantics and morphology in Italian as L2: toward computational simulations
ROSI, Fabiana
2006-01-01
Abstract
The study presents the preliminary computational simulations, which are carried out within a broader research (Rosi in preparation) about the acquisition of Tense-Aspect system in Second Language Acquisition. The research aims to contribute to the debate about the acquisition of Tense-Aspect system in Second Language Acquisition by comparing the development of Aspect morphology and verb semantics in Italian L2 between human learners and computational simulations as neural networks, namely Self-Organizing Maps (Kohonen 2001). The comparison intends to shed light on the cognitive principles and mechanisms that guide learners to acquire linguistic structures in a second language. The literature on Tense-Aspect acquisition in First and Second Language, henceforth L1 and L2, (Andersen and Shirai 1994; Bardovi Harlig 2000) has documented the interaction between the morphological category of Aspect and the semantic category of Actionality in the development of the interlanguage of learners, who initially tend to restrict Aspect forms to specific Actionality classes and, then, gradually spread the Aspect markers following implicational developmental stages from most cognitively salient to least cognitively salient associations between Aspect and Actionality (fig.1.1). Recent surveys (Li and Shirai 2000; Li 2005) interpret the initial interaction between Aspect and Actionality in L1 Acquisition as the result of the children’s analyses of the co-occurrence probabilities between morphological forms and semantic value of predicates in the linguistic input and verify this hypothesis in a connectionist model. This study intends to test empirically if a neural network can display the acquisitional pattern of Aspect morphology and the interaction in this development between Aspect and Actionality also in L2 Acquisition, in order to help understanding, on the one hand, the correlations between frequency effects and cognitive mechanisms, on the other, the role of the prior linguistic knowledge of L1 in L2 Acquisition.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione