Nasalization, Neutral Segments, and Opacity Effects
 





Rachel Walker



UC Santa Cruz











This thesis explores cross-linguistic variation in nasal harmony.  The 
 

goal is to unify our understanding of nasal harmony so that patterns 
 

across languages conform to one basic character and to examine the wider 
 

implications of this account for phonological theory.
 





The analysis builds on generalizations from a comprehensive survey 
 

documenting variation in three descriptive sets of segments in nasal 
 

harmony: targets, which become nasalized, blockers, which remain oral 
 

and block spreading, and transparent segments, which remain oral but 
 

do not block.  The typological generalizations established by this study 
 

provide strong support for a unified view of nasal harmony in which 
 

variation is limited in a hierarchical fashion.







To capture cross-linguistic variation, this analysis draws on a 
 

phonetically-grounded constraint hierarchy ranking segments according 
 

to their incompatibility with nasalization (building on Schourup 1972; 
 

Pulleyblank 1989; Piggott 1992; Cohn 1993c; Padgett 1995c; Walker 1995).  
 

Constraint ranking and violability, fundamental concepts in Optimality 
 

Theory (Prince and Smolensky 1993), also play a crucial role.  Ranking 
 

a [nasal] spreading constraint at all points in relation to the 
 

hierarchy of violable nasalization constraints achieves precisely the 
 

attested set of patterns.







Another typological discovery is that transparent segments pattern with 
 

targets and should be regarded as belonging to this set of segments.  A 
 

theoretical consequence is that [nasal] spreading never skips a segment, 
 

finding new support for strict segmental locality (Ni Chiosain and Padgett 
 

1997; cf. Gafos 1996).  The resulting challenge is determining what 
 

produces surface-transparent outcomes.  Building on early derivational 
 

approaches (e.g. Clements 1976; Vago 1976), I propose to analyze segmental 
 

transparency as a derivational opacity effect.  Following McCarthy (1997) 
 

and extensions by Ito and Mester (1997a), I achieve derivational opacity 
 

effects in Optimality Theory through a correspondence relation between the 
 

actual output and a designated "sympathetic" (failed) member of the 
 

candidate output set.  Sympathetic correspondence realizes transparency 
 

by selecting the output most closely resembling the nasal character of the 
 

fully-spread sympathetic form, while respecting nasal incompatibility 
 

constraints for segments that behave transparent.  Importantly, by 
 

bringing segmental transparency under the wing of derivational opacity, 
 

transparency-specific representations can be eliminated from the theory.
 





Chapter 1 presents background.  In chapter 2, I develop a unified 
 

description and analysis of a cross-linguistic typology of nasal harmony.  
 

Chapter 3 turns to the analysis of transparent segments and a case study 
 

of nasal harmony in Tuyuca.  Chapter 4 presents an acoustic study of nasal 
 

harmony forms in Guarani which verifies that voiceless stops are truly 
 

surface-transparent.  In chapter 5 I consider other proposals for the 
 

analysis of transparent segments, and in chapter 6 I examine other 
 

phenomena that may be mistaken for [nasal] feature spreading.  Nasal 
 

agreement in Mbe forms a case study involving reduplication.