Another look at velar deletion in Turkish, with special ... - Linguistics
Another look at velar deletion in Turkish, with special ... - Linguistics
Another look at velar deletion in Turkish, with special ... - Linguistics
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
14<br />
(19)<br />
(nom<strong>in</strong><strong>at</strong>ive) Accus<strong>at</strong>ive gloss<br />
/kojn/ [ko.jun] [koj.nu] ‘bosom’<br />
koyun koynu<br />
/kojun/ [ko.jun] [ko.ju.nu] ‘sheep’<br />
koyun koyunu<br />
Second, suffix disharmony is a property of some roots even when epenthesis is not <strong>in</strong><br />
the picture. As noted by Lewis (1967:19-20) and Clements & Sezer (1982), <strong>in</strong>ter alia, a<br />
number of <strong>Turkish</strong> roots, ma<strong>in</strong>ly loans from Arabic and Persian, exceptionally take front<br />
vowel harmony on suffixes even though the root vowels are all back. This behavior traces<br />
back to a front-back dist<strong>in</strong>ction on root-f<strong>in</strong>al consonants, long s<strong>in</strong>ce lost, whose only<br />
synchronic reflex is <strong>in</strong> the vowel quality of suffixal (or epenthetic) vowels (20a). <strong>Turkish</strong><br />
does ma<strong>in</strong>ta<strong>in</strong> a synchronic pal<strong>at</strong>ality contrast <strong>in</strong> l<strong>at</strong>erals, and underly<strong>in</strong>gly pal<strong>at</strong>al /l/<br />
triggers front harmony on follow<strong>in</strong>g vowels <strong>in</strong> a transparent way (20b). A sampl<strong>in</strong>g of<br />
back-vowel roots tak<strong>in</strong>g front-harmonic suffixes is given below. Note the front harmony<br />
exhibited by underly<strong>in</strong>g suffix vowels and epenthetic vowels, alike: 8<br />
(20)<br />
nom<strong>in</strong><strong>at</strong>ive accus<strong>at</strong>ive (-/I/) 1sg possessive (-/m/)<br />
(underly<strong>in</strong>g vowel) (epenthetic vowel)<br />
a. /sa<strong>at</strong>/ [sa.<strong>at</strong>] [sa.a.ti] [sa.a.tim] ‘hour’<br />
sa<strong>at</strong> sa<strong>at</strong>i sa<strong>at</strong>im<br />
/hadd/ [had] [had.di] [had.dim] ‘limit’<br />
had haddi haddim<br />
/takʲaːt/ [ta.kʲ<strong>at</strong>] [ta.kʲaː.ti] [ta.kʲaː.tim] ‘strength’<br />
tak<strong>at</strong> tak<strong>at</strong>i tak<strong>at</strong>im<br />
b. /mentolʲ/ [men.tolʲ] [men.to.ly] [men.to.lym] ‘menthol’<br />
mentol mentolü mentolüm<br />
/sosjalʲ/ [sos.jalʲ] [sos.ja.li] [sos.ja.lim] ‘social’<br />
sosyal sosyali sosyalim<br />
Disharmony is clearly a property of <strong>in</strong>dividual roots. Clements & Sezer (1982) handled<br />
disharmony by associ<strong>at</strong><strong>in</strong>g a [-back] fe<strong>at</strong>ure to the f<strong>in</strong>al consonant of roots like sa<strong>at</strong>.<br />
Zimmer (1992) argued aga<strong>in</strong>st this analysis and <strong>in</strong> favor of a flo<strong>at</strong><strong>in</strong>g autosegmental [-back]<br />
fe<strong>at</strong>ure th<strong>at</strong> would l<strong>in</strong>k to any underspecified vowels, whether epenthetic or underly<strong>in</strong>g. We<br />
assume the correctness of the l<strong>at</strong>ter account, but regardless of how disharmony is analyzed,<br />
it clearly exists <strong>in</strong>dependently of epenthesis. In a form like /kabr/ (18), the flo<strong>at</strong><strong>in</strong>g [-back]<br />
8 L<strong>at</strong>erals, like <strong>velar</strong>s, are predictably pal<strong>at</strong>al <strong>in</strong> the environment of front vowels; here,<br />
pal<strong>at</strong>ality is transcribed only when not predictable, i.e. when <strong>in</strong> the environment of back<br />
vowels.