European Commission logo
English English
CORDIS - EU research results

Differences in language structures as a guide to studying differences in cognitive representations: articles and gender

Final Activity Report Summary - GEND_ART_REPR (Differences in language structures as a guide to studying differences in cognitive representations: articles and gender)

The project concerned the relation between the dynamic processes of language evolution, acquisition, and on-line communication and the symbolic structures that constitute language in traditional linguistic descriptions. This topic was tackled both at the theoretical and empirical level.

On the theoretical level, building on the approach developed by H. Pattee, a particular relation between symbols and dynamics is proposed: that of mutual sustaining and control. On this view language's symbols have a value of physical and historical entities which harness the communicative, developmental and social dynamics. Dynamics becomes an inextricable element of the explanations of linguistic phenomena. Such understanding of the relation between symbols and dynamics in natural language is incongruent with the theories that presuppose the possibility of an independent description of language as a formal system of symbols, as well as with the view that an individual's cognitive competence will provide all (or most) elements for the explanation of linguistic structures. On the other hand, this approach recalls the functional theories of language of the 70s (Hymes, 1972; Halliday, 1975; Vygotsky; 1934) and resonates with the approaches to language that continue the functional trends (e.g. Schegloff, et al., 1996; Thibault, 2006, Cowley, 2006; Ross, 2004) and treat language as a social coordination tool.

On the empirical level, we demonstrated how the processes of conceptualisation and linguistic structures co-shape each other on three different time scales. We carried out five experiments and created a cross-linguistic grammatical gender database (1500 nouns, 7 languages). The first experiment involved 280 participants and investigated how the grammatical gender of an object's name in a given language (Polish and Italian) influences the descriptions that people used for this object. We created a Gender Load measure and collected its value for the adjectives used for descriptions. The results showed that Italians and Poles describe differently objects that have different grammatical genders in their native languages. More surprisingly however, the grammatical gender effects were much stronger, in both populations, for objects that had names congruent in grammatical gender in both languages. This suggests the existence of semantic biases in the distribution of grammatical gender within a language. The results of this study point also to the fact that in any experimental task that concerns processes of language-cognition interaction, effects are bidirectional, and due to multiple mechanisms that operate on different time scales: from on-line processes such as activation of frequent linguistic elements congruent in grammatical gender, to ontogenetic processes that shape the underlying concepts in accordance with linguistic categories and structures, to the cognitive biases of linguistic feature distribution. The on-line grammatical-gender effects were further investigated in a priming experiment, confirming the importance of semantics also on this time scale. In another experiment grammatical gender was shown to influence non-figurative object perception as well as behavior towards them, confirming the constraining role of language in interpersonal dynamics. Finally we used a different feature, articles, to show how the differences between Polish and Italian with respect to this feature modify speaker's strategies in dealing with articles in L2 (English).

Specifying the relation between symbols and dynamics along the lines of Pattee's theory can be of a great value for the development of functional approaches to language, not only because it provides a convenient framework but also because it allows for combining traditional psycholinguistic methods with a whole array of new tools that originate from the studies of self-organising dynamical systems. This view allows also seeing language as a natural, biological phenomenon, specific in its cultural dimension but nevertheless building on the shared properties of information transmission in nature.