High Vowel Fricativization and Chain Shift

Abstract

UC Berkeley Phonology Lab Annual Report (2014) High vowel fricativization and chain shift Matthew Faytak University of California, Berkeley December 24, 2014 Introduction ∗ Chain shifts are a thoroughly studied phenomenon by which intra-categorical changes occur in several cate- gories in tandem, such that one adjustment appears to drive another. I argue for expanding the typology of chain shifts to include chain shifts that involve the raising of non-high vowels in tandem with the fricativiza- tion of high vowels. I refer to the latter sound change here as high vowel fricativization (HVF), an unusual but sporadically attested change in which a reconstructible rounded or unrounded front vowel becomes a voiced strident coronal fricative (e.g. [i y] → [z z w ]), or a back rounded vowel becomes a voiced labiodental fricative (e.g. [u] → [v ]). HVF is due, in some better-studied cases, to allophonic coarticulation of high vowels with immediately preceding consonantal fricatives (e.g. [si] → [sz ]). However, still other cases of HVF appear to involve the movement of an entire high vowel category to production as a voiced fricative, even in the absence of a potential conditioning environment (e.g. [Oi] → [Oz ]). It is the latter set of vowel fricativizations that interest us here. I posit that HVF is, in these cases, involved as one intra-categorical change in a chain shift of one or more vocalic phonemes: in other words, that it takes part in a raising chain shift (i.e. e → i → z ). To bolster this argument, I examine the attestation of the fricativized vowels in question and reconstruct their associated raising chains in a small but genetically diverse set of languages. I argue from the attested changes that an explanation for HVF lies in the aerodynamics of speech: high vowels, particularly peripheral high vowels, tend to be realized with some turbulent airflow given their narrow aperture, which provides the “seed” for the eventual phonologization of fricative noise as ∗ This paper would not be possible without the contributions of others: thanks go to Larry Hyman and Andrew Garrett for reading my early drafts, and Keith Johnson for his comments on earlier presentations that gradually became sections of this paper. Similar thanks must go to audiences at the 2014 LSA meeting in Minneapolis and Stanford’s P-Interest workshop. I single out Paul Kiparsky, Meagan Sumner, Greg Finley, Katie Franich, and Jack Merrill for their useful feedback in these and other settings. I am also grateful to Bruce Connell, Katya Chirkova, Jeff Good, David Mortensen, and Chris Donlay for references, discussion, or data pertaining to their respective linguistic areas, which have informed my perspective on HVF even if they are not directly referenced in this paper. Some credit also must be given to Sarah Bakst, who lends this paper its unofficial subtitle, “When Push Comes to [Sv ].”