(16) Exemplar effects in Dutch plural and past tense inflection

Poster session 1
Monday, September 5, 17:30
Emmanuel Keuleers & Dominiek Sandra
University of Antwerp
emmanueldotkeuleersatuadotacdotbe
Two experiments involving plural and past tense inflection in Dutch offer support for exemplar models of inflectional morphology and question the need for a default process in these inflection systems. In a first experiment we asked participants in two age-groups (35- or- younger and 55-or-older) to make forced choices between plural forms of stems that can take the two productive and regular Dutch plural suffixes (-en and -s). Results showed that the relative frequencies of these two plural forms in a corpus were highly predictive for the plural choice. We also found a significant effect of age, the younger group having a consistently higher preference for the -s plural. We conclude that -s plurals are becoming more frequent in Dutch, and that, as a result, both age-groups have different relative frequencies for double plural forms. We suggest that single- mechanism exemplar models can naturally explain the age effect and speakers' sensitivity to the relative frequencies of two acceptable regular plurals , while dual-mechanism models could only explain it by assuming full storage for both forms, which would be incompatible with the proposal that both -en and -s plural inflection are default processes (Pinker, 1999). In a second experiment, a replication of Prasada & Pinker (1993, experiment 2), we collected acceptability ratings for regular and irregular past tense forms of pseudowords. We obtained results similar to Prasada & Pinker: irregular past tense forms were rated less acceptable as their phonological similarity to existing irregulars decreased, ratings for regulars were less affected by phonological similarity. However, contrary to the previous authors' conclusion that a default process is required to explain these results, we offer an explanation in which ratings are not only based on phonological similarity but also on lexical characteristics such as neighbourhood density, which differentiate regular and irregular Dutch verbs (e.g. Tabak, Schreuder, & Baayen, 2005). We demonstrate how an exemplar-based model can make use of these lexical characteristics to explain the data of the rating experiment.

References

Pinker, S. (1999). Words and Rules. London: Phoenix.

Prasada, S., & Pinker, S. (1993). Generalizations of Regular and Irregular Morphological Patterns. Language and Cognitive Processes, 8, 1-56.

Tabak, W., Schreuder, R., & Baayen, R. H. (2005). Lexical statistics and lexical processing: semantic density, information complexity, sex and irregularity in Dutch. In S. Kepser & M. Reis (Eds.), Linguistic Evidence Computational Perspectives. Berlin: Mouton de Gruyter.