The Lapse Constraint

First, it does the work of the traditional Lapse constraint (cf. Selkirk 1984) preventing a metrical foot from becoming too large and hence imposes an upper bound of binarity. Second, it prevents feet from drifting too far apart--at most one unparsed syllable can intervene (a constraint similar to this effect was independently discovered by Kager 1994); finally, another effect of Lapse unique to our formulation is that it prevents a foot from moving too far from the edge of the word: s('ss)s# satisfies the constraint but ('ss)ss# violates it with two successive unparsed syllables. Stated differently, the Lapse constraint ensures that stress is located in a three-syllable window at the edge of the word. In the rest of this paper we show how Lapse elucidates some of the classic data sets in the metrical stress literature including three-syllable window effects from Piraha and Turkish toponyms, minimal binarity in Maranungku, Banawa, and Passamaquody, and ternary stressing in Chugach Alutiiq, Estonian, and Cayuvava.