Vowel Harmony and Cyclicity in Eastern Nilotic

Abstract

No abstract available.

ATR Harmony in African Languages

“…Dominant harmony languages are based on a phonological (in particular a featural) asymmetry: the principle that [−ATR] vowels assimilate to [+ATR] vowels, with the distinction between roots and affixes being essentially irrelevant (see, for example, Anderson 1980: 9, Vago 1980: XIII). This is sometimes described (see, for example, Aoki 1968: 143, Baković 2000: ii, 2001: 2; Krämer 2001) by saying, in effect, that in such a language a [+ATR] vowel anywhere in the word causes all other vowels to become [+ATR]. Root‐controlled languages are based on a different principle of morphological asymmetry, which dictates simply that affixes assimilate to roots.…”

“…More recently, constraints embodying these markedness relationships have figured prominently in synchronic theoretical treatments of [ATR] phenomena. Foundational works in this area include Archangeli and Pulleyblank (1994) and Calabrese (1988Calabrese ( , 1995, while subsequent applications or extensions of the fundamental ideas include, among others, Akinlabi (2006), Bakovic 2000, andLeitch (1996). Arguments against the view that the high [−ATR] vowels are especially marked are found in Casali (1995Casali ( , 2002.…”

“…In practice, the crucial diagnostic for categorizing a language as of the dominant or root-controlled type has almost always been the presence or absence of dominant [+ATR] affixes. Thus, Igbo, which lacks such affixes, is commonly referred to as a root-controlled language (Anderson 1980;Vago 1980: xiii), while Diola-Fogny, which has them, is referred to as a dominant harmony language (Anderson 1980;Bakovic 2000). It seems abundantly clear, however, from the way the distinction between dominant and root-controlled harmony is commonly presented that the distinction is intended to signify something more consequential than the mere presence or absence of a particular type of morpheme.…”

“…Consider first the common assumption that the root-affix distinction is relevant to root-controlled languages, but not dominant harmony languages. A fact that has sometimes been overlooked in introductory descriptions of dominant harmony (and may not be universally known) is that dominant affixes in ATR harmony languages are almost always suffixes; with only extremely rare and marginally productive exceptions, there are no dominant [+ATR] prefixes (Bakovic 2000;Clements 2000;Casali 2003). Instead, prefixes in both root-controlled and dominant harmony languages are consistently recessive, simply alternating in their [ATR] value in agreement with the root.…”

“…1986; Van der Hulst 1988; Clements 1989, 1991; Parkinson 1996; Salting 1998, 2005; Schane 2005), and, more recently, various developments within optimality theory (Smolensky 1993; Akinlabi 1994, 1997, 2006; Casali 1996; Leitch 1996; Pulleyblank 1996, 1997, 1998, 2002; Pulleyblank and Turkel 1996; Pulleyblank et al. 1997; Cahill 1999; Causley 1999; Baković 2000, 2001, 2002, 2003; Noske 2000; Krämer 2001; McCrary 2001; 2001a,b; O’Keefe 2003, 2007; Kaplan 2006, 2007; Lahne 2006; Ussishkin 2006; Finley 2007; Lesley‐Neuman 2007; Sasa 2007; Hansson forthcoming; among others).…”