Phonotactic learning without a-priori constraints: Arabic root restrictions revisited
John Alderete, Paul Tupper, Stefan A. Frisch
In this article, we develop a connectionist model of learning phonotactics and apply it to the problem of learning root cooccurrence restrictions in Arabic. In particular, a multilayer network with a hidden layer is trained on a representative sample of actual Arabic roots using error-corrective back-propagation. The trained network is shown to classify actual and novel Arabic roots in ways that are qualitatively parallel to psycholinguistic studies of Arabic. Statistical analysis of network behavior also shows that activations of nodes in the hidden layer correspond well with violations of symbolic well-formedness constraints familiar from generative phonology. The larger finding is therefore that a sub-symbolic phonotactic system trained on realistic data can mirror the behavior of a symbolic-computational system typical of contemporary phonological analysis.
Keywords: sub-symbolic learning, connectionism, Optimality Theory, levels of explanation, Arabic, cooccurrence restrictions, dissimilation
Downloads: link to paper, math supplement
Full citation: Alderete, John, Paul Tupper, Stefan A. Frisch. 2014. Phonotactic learning without a-priori constraints: Arabic root cooccurrence restrictions revisited. In Andrea Beltrama, Tasos Chatzikonstantinou, Jackson L. Lee, Mike Pham, Diane Rak (eds.), 48th annual meeting of the Chicago Linguistics Society, pp. 1-16.