When a Passamaquoddy unstressable schwa, that's a mora
Paul Hagstrom, MIT
July, 1995

This paper discusses patterns of stress and syncope in the Algonquian language
Passamaquoddy, proposing an analysis in terms of Optimality Theory (Prince &
Smolensky (1993)) and the Prosodic Hierarchy (Selkirk (1984) and later work of
several authors).  The particular proposal involves a slight modification to the
Prosodic Hierarchy and focusing on the constraint *Peak(schwa), which avoids
schwa as the nucleus/peak of a syllable. The constraint itself was introduced
implicitly in Prince & Smolensky (1993) but can be ranked high enough to have
interesting effects, as the Passamaquoddy data show.  By allowing a structure in
which a mora (e.g., dominating a schwa in phonological representation) can be
directly dominated by a foot, Passamaquoddy has two relevant means by which it
can satisfy the *Peak(schwa) constraint: either syncope of the schwa segment, or
a mora-foot structure of the sort just described.

Another set of related phenomena involve a constraint which appears to favor
geminates, Min(mora).  In particular, if a schwa appears between identical
consonants in underlying form, the surface form will be a geminate due to syncope
of the schwa segment.  The Min(mora) constraint and its relatives also have
fairly diverse effects in the constraint interactions.

With these tools, this paper attempts to provide an analysis of a broad set of
empirical facts which had previously been analyzed extensively (though under
differing assumptions) in LeSourd (1993).  Other possible crosslinguistic
evidence for Min(mora) and *Peak(schwa) is also briefly suggested, although the
main focus of the present work is evidence internal to Passamaquoddy.

[Cited references can be found in the paper]