Extraction of linguistically relevant pragmatic contrast through language learning

This study explores whether pragmatic contrast that is not expressed as relevant in the subjects' native language is extracted through language learning. Miniature artificial languages were used. In Experiment 1, a contrast was created such that four persons appeared together in each situation, with three persons performing the same activities and the fourth one performing a different target activity. Under this presentation condition pragmatic contrast was shown to be rather easily extracted. In Experiments 2 and 3, a single person appeared in each situation and contrast was created serially by changing the position in which the person performing the target activity was placed. Under these presentation conditions sufficient evidence for the extraction of pragmatic contrast was not found. Experiment 2 also examined the effect of different experience of instances on the extraction of pragmatic contrast. Results showed that although different experience did not produce a differential effect as expected, a definite conclusion was not drawn for a low level of performance. It was formulated that the two presentation conditions differed in the degree of transparency for a mapping relation between the linguistic structure and the pragmatic contrast. Implications of the findings were suggested in relation to a natural learning situation.Deep gratitude is due to Professor D. D. Steinberg for his reading of an earlier version of this paper.