Date created: | Last Updated:

Category: Project

Description: Despite the rich theoretical and empirical literature on negative polarity items (NPIs) in general, understating NPIs like 'all that' or 'much' have received relatively little attention in psycholinguistics. In this paper, we investigate the comprehension, processing, and production of two such understating NPIs in German, namely 'sonderlich' ‘particularly’ and 'so recht' ‘really’. In a first experiment, using self-paced reading and naturalness ratings, we found that 'sonderlich', contrary to 'so recht', was rated as natural in affirmative contexts although this environment is incompatible with NPIs. The finding is subsequently extended to the domain of sentence production, demonstrating that 'so recht' was consistently used as NPI, but 'sonderlich' was not. The last two experiments investigate the factors underlying this finding, showing that the surprising patterns for 'sonderlich' may relate to its susceptibility for interference from form- and meaning-related lexical competitors during tasks that strain cognitive resources, and, to some extent, to individual differences in participants’ language aptitude measured through print exposure. Based on the novel empirical data, we discuss the theoretical status of 'sonderlich' and 'so recht' as understating NPIs, on the one hand, and the cognitive mechanisms affecting retrieval of their NPI-related lexical-semantic features, on the other.