Transparency, Strict Locality, and Targeted Constraints

The claim that feature assimilation is strictly local, applying only between adjacent segments, appears to be contradicted by languages in which, descriptively speaking, vowel harmony passes through so-called 'transparent' vowels without affecting them. We adopt the particular approach to vowel harmony developed within OT by Bakovic (2000), according to which agreement constraints require only articulatorily adjacent (Gafos 1996, Ni Choisain & Padgett 1997) vowels to harmonize with one another. Adopting strict locality seems to force us to incorrectly predict that non-assimilating vowels block harmony -- are opaque to it -- in all languages. Our proposal is that transparency is optimal in some languages because an output candidate with a transparent vowel diverges minimally from an output candidate with full assimilation. These languages highly value full assimilation, but they also highly value avoidance of certain types of (marked) vowels. These two value systems interfere with one another, resulting in a pattern that is an optimal blend of assimilation and avoidance: namely, transparency. Formally, transparency arises from a targeted constraint (Wilson 2000; cited as Wilson, in preparation in the paper) that disprefers the full assimilation candidate relative to another candidate that is exactly the same except that one or more vowels (the transparent ones) have failed to assimilate. Because a potential candidate with opaque vowels is in turn disprefered relative to full assimilation by an agreement constraint, the end result is that transparency is optimal. The substantive basis for this targeted constraint is bipartite: first, transparent vowels and their assimilated counterparts are perceptually similar, and second, the assimilated counterparts of transparent vowels are articulatorily marked. The example discussed in the paper concerns the transparency of high vowels to [ATR] harmony: [-ATR] high vowels are articulatorily marked, and are perceptually similar to [+ATR] high vowels (Archangeli & Pulleyblank 1994).The definitive version of this paper was published in WCCFL 19 : proceedings of the 19th West Coast Conference on Formal Linguistics (2000) and is available at http://www.cascadilla.com/wccfl19.htmlBakovic, E. & Wilson, C. (2000). Transparency, Strict Locality, and Targeted Constraints. In R. Billerey & B.D. Lillehaugen (Eds.), WCCFL 19 : proceedings of the 19th West Coast Conference on Formal Linguistics (pp. 43-56.). Somerville, MA : Cascadilla Press