ABSTRACT
This chapter surveys the literature of connectionist models which have been developed to address problems central to generative phonology. It provides the formal background necessary for understanding how connectionist phonology works. Phonological structure may be organized into syllables and other higher level units like prosodic feet that enrich the symbolic phonology with constituency. Syllables are canonically built up in three concrete steps. First, a syllable nucleus is built over high-sonority segments, typically vowels or other sonorants. Syllable onsets are then formed by grabbing a string of rising-sonority consonants and placing them in syllable-initial position. Finally, the residue is dumped into the syllable coda, a process that is subject to certain constraints. Goldsmith and Larson's analysis of Berber syllabification gives a very direct analysis of the claim that phonology is local. One of the clear advantages of connectionist approaches to phonology is that they are naturally gradients, so they give direct analyses of graded phonology and scales.
