We explore the implications of a particular approach to learning for an architectural question in phonology. The learning approach follows the principle of Minimum Description Length (MDL), which has recently been used for learning in both constraint-based and rule-based phonology. The architectural question on which we focus is whether the grammar allows language-specific statements to be made at the level of the lexicon, as was assumed in early generative phonology, or whether such statements are prohibited, as is commonly assumed within more recent work. We show that under MDL, the architectural question has real empirical implications: across a range of seemingly natural representational schemes, an ability to make language-specific statements about the lexicon is needed to ensure the learnability of an important aspect of phonological knowledge.

In Icelandic, and in most dialects of Faroese, the fundamental laryngeal stop contrast is one of aspiration. This contrast is neutralized in certain positions, following patterns which appear problematic for perceptually-based theories (licensing by cue). In postvocalic onset position, a cue-rich environment, the contrast is absent in Icelandic. Most dialects neutralize to unaspirated [p, t, k], whereas others have postaspirated [ph, th, kh] instead. In the latter case, the appearance of loanwords with intervocalic [p, t, k] is now creating an incipient aspiration contrast, undoing the previous neutralization pattern. The implications of these facts for Optimality Theory (Richness of the Base) are discussed. In certain dialects of Faroese, such postvocalic singleton onsets are realized as PREaspirated [hp, ht, hk]. Curiously, this preaspiration is suppressed in precisely the contexts where it ought to be most salient. An OT analysis is developed to account for this distribution.