Saliens i ordlæring i tvetydige situasjoner
Gabrielle Weidemann, førsteamanuensis i psykologi ved Western Sydney University, kommer til Forum for klinisk lingvistikk og språktilegnelse 5. september for å holde foredraget «The influence of word and object salience on word learning under ambiguous conditions». Arrangementet foregår på engelsk og er åpent for alle.
In learning new words an association must be formed between a word and its referent, however in many circumstances there are many words and many possible referents, and associations are formed across time and contexts. This process has been modelled experimentally using cross-situational word learning paradigms where participants are presented with multiple words and multiple referents across multiple trials. However there is some debate about the underlying mechanisms, whether it is a form of statistical associative learning or a memory-based inference mechanism or some hybrid of the two. In a series of experiments using a cross-situational word learning paradigm, we manipulated the salience of words, and object referents in isolation and in combination in order to examine the mechanisms of association formation. We found that within experiments manipulations of word salience increased learning for both salient and non-salient words, whereas with experiment manipulations of object salience in isolation and in combination with word salience manipulations learning was impaired. Modelling suggests that manipulations of word salience increased learning through increasing word discriminability and that object salience captured attention but did not facilitate association formation, indicative of the operation of an a more inferential process.
Gabrielle Weidemann er førsteamanuensis i psykologi ved Western Sydney University. Hun har en doktorgrad innen nevrovitenskap fra University of New South Wales, og forsker blant annet på assosiativ læring hos mennesker. Hun kombinerer metoder som måling av reaksjonstid med fysiologiske målinger som hudrespons, hjerterytme, EMG, blunking og MR.