Back to list

Detail of contribution

Auteur: Vsevolod KAPATSINSKI

Morphological schema induction by means of conditional inference trees

Abstract/Résumé: First-order schemas or constructions have been proposed to be the basic unit of morphology in constructionist and cognitive approaches to language (Booij 2010, Bybee 1985, 2001, Nesset 2008) and have received much experimental support (e.g., Kapatsinski 2012, 2013). First-order schemas are abstractions over form-meaning pairings rather than changes in context. However, no explicit computational models of schema extraction have been proposed. The present paper contributed to filling this gap by showing that first-order schemas or constructions can be extracted via non-parametric conditional inference (Hothorn ef al. 2006). The basic idea is that learners gradually zero in on the features that characterize forms that occur in a certain cell of a morphological paradigm. The main theoretical predictions of the proposal are 1) that first-order schemas grow more specific over the course of language acquisition, and 2) cross-over interactions between units should be hard to learn (e.g., Dell 2000, Kapatsinski 2009) and rare in languages, whereas non-crossover interactions should be easy to learn and common in languages. References: Booij, G. 2010. Construction Morphology. Oxford: Oxford University Press. Bybee, J. L. 2001. Phonology and language use. Cambridge: Cambridge University Press. Bybee, J. L. 1985. Morphology: A study of the relation between meaning and form. Amsterdam: John Benjamins. Dell, G. S., K. D. Reed, D. R. Adams, & A. S. Meyer. 2000. Speech errors, phonotactic constraints, and implicit learning: a study of the role of experience in language production. Journal of Experimental Psychology: Learning, Memory & Cognition, 26, 1355-67. Hothorn, T., K. Hornik, & A. Zeileis. 2006. Unbiased recursive partitioning: A conditional inference framework. Journal of Computational and Graphical Statistics, 15, 651-74. Kapatsinski, V. 2013. Conspiring to mean: Experimental and computational evidence for a usage-based harmonic approach to morphophonology. Language, 89(1). Kapatsinski, V. 2012. Which statistics do learners track? Rules, constraints, or schemas in (artificial) language learning. In S. Th. Gries & D. Divjak, eds. Frequency effects in language, 53-82. Berlin: Mouton de Gruyter. Kapatsinski, V. 2009. Testing theories of linguistic constituency with configural learning: The case of the English syllable. Language, 85, 248-77. Nesset, T. 2008. Abstract phonology in a concrete model: Cognitive Linguistics and the morphology-phonology interface. Berlin: Mouton de Gruyter.