ABSTRACT
This chapter examines the analysis of transparent segments in nasal harmony, that is, segments which are produced with a raised velum within a nasal span. This realization of a truly oral segment within a nasal spreading domain is problematic because it presents a case which appears to deny the claim that feature spreading is segmentally strictly local. Chapter 2 maintained that a spreading nasal feature propagates only between immediately adjacent segments; skipping a segment is not a possible outcome in spreading. This result follows from the wellmotivated assumption that the gapped configuration is universally illformed: a representational consequence of interpreting a multiply-linked feature as a continuous property or gesture. In the previous chapter, evidence was adduced from the typology of nasal harmony in support of the claim that descriptively transparent segments should be analytically grouped with undergoers of nasal spreading. Some antecedent derivational or sequential multi-level accounts of truly transparent segments have maintained the strict locality of spreading by positing a level of representation at which the transparent segment undergoes spreading (e.g. Clements 1976; Vago 1976; Walker 1996; Nt Chiosain and Padgett 1997; among others). A subsequent rule or constraint then applies to this representation to change the feature specification on the transparent segment to realize its surface transparency. More generally, this kind of approach analyzes true transparency as an instance of a 'derivational opacity effect' (Kiparsky 1971, 1973), in the sense of an outcome that is derived through an opaque interaction of rules or
constraints. For transparent segments in nasal harmony, I follow this core idea by analyzing transparency as the outcome of an opaque interaction of optimality-theoretic constraints.
