Learning non-concatenative morphology
Michelle Alison Fullwood, Tim O’Donnell · 2013
Recent work in computational psycholinguistics shows that morpheme lexica can be acquired in an unsupervised manner from a corpus of words by selecting the lexicon that best balances productivity and reuse (e.g. Goldwater et al. (2009) and others). In this paper, we extend such work to the problem of acquiring non-concatenative morphology, proposing a simple model of morphology that can handle both concatenative and non-concatenative morphology and applying Bayesian inference on two datasets of Arabic and English verbs to acquire lexica. We show that our approach successfully extracts the non-contiguous triliteral root from Arabic verb stems. 1