opacity and prominence in crimean...
TRANSCRIPT
Opacity and prominence in Crimean Tatar
Darya Kavitskaya Yale University
CUNY Conference on the Phonology of Endangered languages
January 14, 2011
The language • Crimean Tatar (CT) is an understudied and endangered language of the West
Kipchak branch of the Northwestern subgroup of the Turkic family (Johanson 1998).
• CT is spoken in Ukraine’s Crimean peninsula and also in Uzbekistan, Russia, Bulgaria, Romania, and Turkey (Samoilovich 1916, Bogoroditskii 1933, Sevortian 1966, Memetov 1993, Izidinova 1997, Useinov, Mireev & Sahadzhiev 2005, Kavitskaya 2010).
• The data come from the author’s fieldwork in 2002, 2003, 2009 in Crimea, Ukraine.
2
3
4
5
Processes
• Harmony – Backness – Rounding
• Syncope of high vowels – Initial – Medial
• Stress
6
CT vowels
-back +back -round +round -round +round +high i y ɯ u -high e ø a o • i and ɯ have undergone an almost complete
phonetic merger, but remain phonologically distinct.
7
Backness harmony bil-mek ‘know’ juv-maq ‘wash’ ket-mek ‘go’ qorq-maq ‘be afraid’ tyʃyn-mek ‘think’ qɯr-maq ‘rub’ tøk-mek ‘pour’ ajlan-maq ‘turn’
8
Rounding harmony • Triggered by any round vowel, targets high vowels only. dost-u ‘friend-3SG.POSS’ tʃift-i ‘pair-3SG.POSS’ • Rounding harmony is active only in the first two
syllables of a word. a. dost-um ‘friend-1SG.POSS’ kyz-lyk ‘autumn-ADJ.SUF’ bul-un-maq ‘find-PASS-INF’ b. tuzluɣ-ɯm ‘salt shaker-1SG.POSS’ syrgyn-lik ‘deportation-ADJ.SUF’ tykyr-in-mek ‘spit-PASS-INF’
9
Dialectal variation and harmony
• In the Southern dialect of CT, rounding harmony affects all high vowels in a prosodic word (low vowels are blockers), and in the Northern dialect of CT rounding harmony is lost; the feature [round] is licensed only in the initial syllable of the word (like in some Altaic languages, such as Vogul, Bashkir, Ostyak (Steriade 1995: 161-162)).
10
Syncope of high vowels • Syncope targets high vowels, in word-initial (a) and word-medial syllables
(b). • Syncope of a high vowel in an initial syllable can create word-initial onsets
that do not obey the CT phonotactics (a). • Word-medially, syncope is blocked if it results in structures not acceptable
by the phonotactics of the language. – In the native vocabulary, complex onsets are not allowed. – Complex codas are maximally CC and obey the SSP.
a. kitap [ktap] ‘book’ tɯʃlemek [tʃlemek] ‘to bite’ bilem [blem] ‘I know’ sɯkmaq [skmaq] ‘to push, press’ qɯsqa [qsqa] ‘short’ b. aldɯlar [aldlar] ‘they took’ otura [ot.ra] ‘s/he sits’ ketirip [ket.rip] ‘having brought’ øldyrmek [øldyrmek] *[øldrmek] ‘to kill’
11
Syncope of high vowels
• The leftmost vowel in a word deletes(c). • The vowel may delete even when it is the
absolute initial in a word (d). • Final (stressed) high vowels never delete (e). c. tyʃyrdik [tʃyrdik] ‘they dropped’ tykyrem [tkyrem] *[tykrem] ‘I spit’ piʃirem [pʃirem] *[piʃrem] ‘I cook’ d. iʃlemek [ʃlemek] ‘to work’ e. berdi [berdi] *[berd] ‘she gave’
12
A spectrogram of /tykyrmek/ ‘to spit’
13
Stress • Each word in Crimean Tatar has exactly one main stress. • The default stress position is word-final.
– It has been argued for Turkish (Levi 2005) that its default final stress is postlexical that seems to be the case for the related CT as well.
a. araˈba ‘cart’ araba-ˈlar ‘carts’ cart-PL araba-lar-ˈdan ‘from carts’ cart-PL-ABL b. baʃla-ˈdɯ-m ‘I began’ begin-PAST-1SG baʃ-lar-ɯmɯz-ˈnɯ ‘our heads’ head-PL-1PL.POSS-ACC
14
Stress
• Final stress is overriden by lexical stress in both roots and pre-stressing suffixes.
a. ˈnasɯl ‘which, how’ ˈmitlaqa ‘definitely’ ˈtezden ‘quickly’ b. aˈʃar-ɯm ‘I eat’ iˈtʃer-im ‘I drink’ c. geˈdʒe-lejin ‘at nights’ aʃɯq-tʃanˈlɯq-nen ‘in a hurry’ aˈna-dʒasɯna ‘in a motherly manner’ d. bar-ˈdɯ ‘he went’ ˈbar-ma-dɯ ‘he didn’t go’ bil-ˈmek ‘to know’ ˈbil-me-mek ‘to not know’ 15
An opaque interaction between harmony and syncope
• Harmony and syncope in rule terms: tyʃ-Ir-Em ‘fall-CAUS-1SG.PRES’
a. UR tyʃ-Ir-Em 1. Harmony tyʃyrem 2. Syncope tʃyrem Surface tʃyrem b. UR tyʃ-Ir-Em 1. Syncope tʃIrEm 2. Harmony tʃirem Surface *tʃirem
16
A classic OT account • LICENCERD(σσ) (after Walker 2005)
– Feature [round] must be associated to positions in two syllables. • DEP(round)
– Assign a violation mark for every instance of the feature [round] in the output that has no correspondent in the input (=don’t insert the feature [round]).
• *NUC/i,u,y,ɯ >> *Nuc/e,o,a,ø (informally, *Nuc/high >> *Nuc/low)
• (Gouskova 2003 on differential syncope, see also Prince and Smolensky 1993, de Lacy 2004, 2006).
• MAXV – Assign a violation mark for every input vowel that has no output
correspondent (=don’t delete a vowel).
17
CT opacity in classic OT
18
Harmonic serialism and vowel harmony
• Serial Harmony avoids some undesirable typological predictions with respect to feature spreading, present in classic OT (McCarthy 2009, to appear; Kimper 2008; Pruitt 2008; Wilson 2003, 2004, 2006; Wolf 2008, Zentz 2011).
• See, in particular, Padgett 1995, McCarthy 2003 on the sour-grapes property of local agreement constraints.
19
Assumptions of Serial Harmony (McCarthy 2009: 1-2)
• Distinctive features are privative (present/absent), and not equipollent (positive/negative).
• Harmony is motivated by a constraint on autosegmental representations, Share(F), that is violated by any pair of adjacent segments that are not linked to the same [F] autosegment.
• The input for the [tʃyrem] ‘I drop’ is: [round] | t y ʃ i r e m
20
Constraints • SHARE(F) (McCarthy 2009: 8)
– Assign one violation mark for every pair of adjacent segments that are not linked to the same token of [F].
• SHARE(back) • SHARE(round) • INITIAL(F) penalizes leftward spreading of a feature F (20),
and FINAL(F) penalizes rightward spreading of the feature F (McCarthy 2009).
• The harmony in CT is progressive, thus INITIAL(F) >> SHARE(F) >> FINAL(F) (where F is back and round).
• Harmonic Serialism has the same problem as classic OT with the analysis of counterbleeding opacity (see McCarthy 2007: 37).
21
OT with candidate chains (OT-CC, McCarthy 2007)
• The output is reached from the input via a series of steps (a candidate chain)
• Gradualness: one violation of one basic faithfulness constraint per step (a localized unfaithful mapping, LUM)
• The first step is the most harmonic faithful parse of the input
• Harmonic improvement: each step must improve harmony
• Each chain has a correspondent set of LUMs (the L-set) and an ordering of the elements in the set (rLUMSeq).
22
Valid chains for the input /tyʃ-ir-em/ ‘I drop’
a. <tyʃirem> Ø, Ø (faithful) b. <tyʃirem, tyʃyrem> {DEP(rd)@4}, Ø c. <tyʃirem, tʃirem> {MAXV@2}, Ø d. <tyʃirem, tyʃrem> {MAXV@4}, Ø e. <tyʃirem, tyʃyrem, tʃyrem> {DEP(rd)@4,
MAXV@2}, {<DEP(rd)@4, MAXV@2>} f. <tyʃirem, tyʃyrem, tyʃrem> {DEP(rd)@4,
MAXV@4}, {<DEP(rd)@4, MAXV@4>} 23
Opacity in OT-CC
• Within OT-CC, we account for opacity with a precedence constraint PREC(A, B), which requires that all violations of B are preceded by and not followed by violations of A.
• PREC(DEP(round), MAXV) requires violations of DEP(round) (harmony) to precede and not follow violations of MAXV (syncope).
24
An OT-CC tableau for the input /tyʃ-ir-em/ ‘I drop’
25
Prominence and the interaction of harmony and syncope
• In CT the prominence status of the initial syllable is different for different processes.
• The initial syllable is a common privileged position associated in the literature with phonological strength effects (see Barnes 2006; Beckman 1997; Kaun 1995, 2004). • Northern CT: roundness is limited to the initial syllable.
• The same position is also weak, and is thus the best syncope site, as it is the furthest away from the final stress. – CT does not show any evidence for secondary stress or further
footing. • The conflicting requirements on prominence are the source of
opacity in the system. • Support: a word nasɯl ‘which, how’ is stressed on the first syllable.
The second (high) vowel is reduced and often deleted.
26
An analysis
• To formalize the proposal, we modify OT-CC to include a family of constraints on the preference of the direction of iteration, PREFER(Fx, Fx+1), where F is a faithfulness constraint.
• PREFER(MAXx, MAXx+1) – Assign one violation mark for a candidate chain
that has a violation of MAX and a competitor chain in which this violation occurs earlier in the form.
27
A modified tableau for the input /tyʃ-ir-em/ ‘I drop’
28
Conclusions • Conflicting prominence in CT is the source of opacity. • Vowel harmony is driven by spreading of a feature from the
initial (most prominent) syllable. • Syncope of high vowels prefers the initial syllable since it is
the least prominent, being the furthest away from stress. • The decision between the initial and medial syncope
cannot be made by metrical constraints since there is no evidence for further footing in CT, beyond the final stressed syllable.
• In order to account for these data, we proposed a constraint on the preference of the direction of iteration.
29
Acknowledgements
• I thank Eric Ciaramella and Matt Wolf for their insightful comments on this paper. I am indebted to Remzije Berberova and to my other Crimean Tatar consultants for sharing their language with me.
30
Selected references • Beckman, J. 1997. Positional faithfulness, positional neutralisation and Shona
vowel harmony. Phonology 14: 1-46. • Berta, Árpád. 1998. West Kipchak languages. In L. Johanson, & E. Csato, eds., The
Turkic languages. New York: Routledge. 301–317. • Bogoroditskii, V.A. 1933. Dialektologicheskie zametki. V. O krymsko-tatarskom
narechii. Kazan. • de Lacy, Paul. 2004. Markedness conflation in Optimality Theory. Phonology 21:
145-199. • de Lacy, Paul. 2006. Markedness: reduction and preservation in phonology.
Cambridge, CUP. • Gouskova, Maria. 2003. Deriving economy: syncope in Optimality Theory. Ph.D.
dissertation, University of Massachusetts, Amherst. • Izidinova, S.R. 1997. Krymskotatarskii iazyk. In E.R. Tenishev (ed.), Iazyki mira.
Tiurkskie iazyki. Moscow: Indrik. • Johanson, Lars. 1998. The history of Turkic. In Johanson, L. & E. Csato (eds.) The
Turkic languages. New York: Routledge. 81–125. • Kaun, Abigail. 1995. The typology of rounding harmony: an Optimality Theoretic
approach. PhD dissertation, UCLA. • Kaun, Abigail. 2004. In Bruce Hayes, Robert Kirchner, and Donca Steriade, eds.
Phonetically Based Phonology. Cambridge University Press. • Kavitskaya, Darya 2010. Crimean Tatar. LINCOM Europa. • Kimper, Wendell. 2008. Local optionality and harmonic serialism. Unpublished
manuscript, University of Massachusetts, Amherst. ROA-988. • Memetov, A. 1993. Krymskie tatary: istoriko-lingvisticheskii ocherk. Simferopol:
Anaiurt. • Pruitt, Kathryn. 2008. Iterative foot optimization and locality in stress systems.
Unpublished manuscript, University of Massachusetts, Amherst. ROA-999.
31
• Levi, Susannah V. 2005. Acoustic correlates of lexical accent in Turkish. JIPA 35: 73-97.
• McCarthy, John J. 2003. OT constraints are categorical. Phonology 20: 75-138. • McCarthy, John J. 2007. Hidden Generalizations. Equinox, London. • McCarthy John J. 2008. The serial interaction of stress and syncope. NLLT 26: 499-
546. • McCarthy, John J. 2009. Harmony in harmonic serialism. Unpublished manuscript,
University of Massachusetts, Amherst. ROA-1009. • Samoilovich, A. N. 1916. Opyt kratkoi krymsko-tatarskoi grammatiki. Petrograd. • Sevortian, E. 1966. Krymskotatarskii iazyk. In N. Baskakov et al, eds., Iazyki narodov
SSSR 2, Nauka, 234-259. • Steriade, Donca. 1995. Underspecification and markedness. In John Goldsmith,
ed., The Handbook of Phonological Theory. Cambridge, MA: Blackwell Publishers. • Useinov, S., V. Mireev, & V. Sahadzhiev. 2005. Qırımtatar tilini ögreniñiz.
Simferopol: Ocaq. • Walker, Rachel. 2005. Weak triggers in vowel harmony. NLLT 23: 917-989. • Wilson, Colin. 2006. Unbouded spreading is myopic. In Workshop on Current
Perspectives on Phonology, vol. 23. • Wolf, Matthew A. 2008. Optimal interleaving: serial phonology-morphology
interaction in a constraint-based model. Ph.D. dissertation, University of Massachusetts, Amherst, MA. ROA-996.
• Zentz, Jason. 2011. Progressive front vowel harmony in Warlpiri: a Serial Harmony approach. Paper to be presented at the CUNY Conference on the Phonology of Endangered Languages, January 14, 2011. 32