Recovering morphology from local phonotactic constraints
Basilio Calderone, Chiara Celata, Ivan Herreros · 2008
Implicit phonotactic knowledge emerges from learners’ exposure to language-specific distributional information (e.g. Albright & Hayes, 2003; Pierrehumbert & Coleman 1997). ‘Wordlikeness’ studies (Bailey & Hahn, 2001) distinguish between two different types of knowledge involved in production/judgement of novel words: (a) phonotactic (string level); (b) lexical (word level). Different statistical models are used to shape the two types of knowledge (bi/trigram transition probability for (a) and lexical neighborhood for (b)). In this work, we propose that the phonotactic level and the word level should not be conceived of as two independent, blind domains, but rather as interconnected systems interacting in a bottom-up fashion. The interplay of phonotactic local constraints and frequency-sensitive word patterns is powerful enough to enable overranked morphological patterns to be recovered from rough data.