The study of the acquisition of pragmatic inferences as an element in the choice between theoretical models in dispute

Abstract

Phenomenon known for its ambivalent behavior, presupposition has been challenging some generations of linguists regarding two complementary problems: the presupposition projection problem and the problem about the source of presuppositions. Although the descriptive problem has received more attention, it is impossible to cope with it without assuming some position regarding the explanatory problem, associated to the source of presuppositional inferences. The noticeable divergence related to the presupposition projection descriptions is in contrast, however, to a certain consensus, historically built, about the nature of presuppositions. According to this consensus, presupposition is a conventional phenomenon, affected only in its projection by conversational linguistic aspects. Recently, that consensus has been shaken by a new wave of researches that, in different ways, has started to advocate the idea that presupposition is a phenomenon essentially conversational. Among those proposals, it deserves attention for its plausibility the one that associate presupposition to scalar quantity implicature. Both forms to take into account presuppositions, nevertheless, allow different predictions regarding their processing patterns. Focused on the confirmation of one of these hypothesis, some experimental studies had been carried out in order to assess how presuppositions and scalars implicatures are processed. The results obtained have been an important part in the discussion about the explanatory problem of presuppositions. This article revisits some experiments that explore the processing of presuppositions and scalar implicatures, and try to appreciate the importance, in the general context of the debate, of the works which investigate the way subjects in the acquisition stage process pragmatic inferences. What it shows is that the evidences obtained may be seen as supporting some models or, alternatively, as clues aimed to the improvement of the models which they, at a first glance, seems to falsify.