UYGHUR BACKNESS HARMONY There are generally two classes of vowels - - PowerPoint PPT Presentation

UYGHUR LOCALITY, TRANSPARENCY, BACKNESS HARMONY

• There are generally two classes of vowels for harmony: regular

and neutral vowels (blockers and transparent vowels).

• Regular vowels undergo and spread harmony

Turkish ɑltɯ-lɑr-dɑ six-PL-LOC ɑltɯ-gen-ler-de six-PLYGN-PL-LOC

[+bk] [+bk] [-bk]

• There are generally two classes of vowels for harmony: regular

and neutral vowels (blockers and transparent vowels).

• Regular vowels undergo and spread harmony
• Blockers do not undergo harmony, but spread their own value of [F].

Turkish ɑltɯ-lɑr-dɑ six-PL-LOC ɑltɯ-gen-ler-de six-PLYGN-PL-LOC

[+bk] [+bk] [-bk]

• There are generally two classes of vowels for harmony: regular

and neutral vowels (blockers and transparent vowels).

• Regular vowels undergo and spread harmony
• Blockers do not undergo harmony, but spread their own value of [F].

Turkish ɑltɯ-lɑr-dɑ six-PL-LOC ɑltɯ-gen-ler-de six-PLYGN-PL-LOC

[+bk] [+bk] [-bk]

• Transparent vowels are skipped by harmony.
• They do not undergo, but do not impede feature spreading.

Hungarian pɔpiːr-nɔk paper-DAT hiːd-nɔk bridge-DAT

[+bk] [+bk]

• Transparent vowels are skipped by harmony.
• They do not undergo, but do not impede feature spreading.

Hungarian pɔpiːr-nɔk paper-DAT hiːd-nɔk bridge-DAT

[+bk] [+bk]

• There is descriptive and theoretical work suggesting that all

harmony is local.

• Harmony affects all vowels within its domain.
• “Transparent” vowels only seem transparent. They actually covertly alternate for

harmony.

• There is phonetic and phonological evidence that /iː/ does exhibit low-level

alternations in Hungarian.

Benus & Gafos 2007; Rebrus et al. 2012; Rebrus & Törkenczy 2016; Szeredi 2016

• Furthermore, some contend that harmony is strictly local.
• Strict locality demands that harmony affects everything within its

domain, including consonants.

• Turkic has figured largely in this body of work.
• The question is thus: do reported cases of transparency actually

involve transparency?

• I will argue, contrary to previous descriptions, that there is no

transparency, but rather strict locality in Uyghur backness harmony.

Dzhunisbekov 1980; Boyce 1990; Kaun 1995; Gafos 1999; Ní Chiosáin & Padgett 2001; McCollum 2015

Analysis Characteristics Attested Language (with articulatory and/or acoustic evidence for claim) Transparency neutral vowels do not exhibit any differences due to harmony Wolof, Uyghur Covert locality neutral vowels exhibit slight differences due to harmony Hungarian Locality all vowels exhibit salient differences due to harmony Kinande Strict locality all segments exhibit differences due to harmony Kazakh

Benus & Gafos 2007; Dye 2015; Gick et al. 2006; Dzhunisbekov 1980; McCollum 2015

Analysis Characteristics Attested Language (with articulatory and/or acoustic evidence for claim) Transparency neutral vowels do not exhibit any differences due to harmony Wolof Covert locality neutral vowels exhibit slight differences due to harmony Hungarian Locality all vowels exhibit salient differences due to harmony Kinande Strict locality all segments exhibit differences due to harmony Kazakh, Uyghur

Benus & Gafos 2007; Dye 2015; Gick et al. 2006; Dzhunisbekov 1980; McCollum 2015

• Uyghur is the only Turkic language reported to exhibit

transparency.

• Moreover, it is reported to exhibit transparency across a number
• f intervening neutral vowels.

Lindblad 1990; Hahn 1991, 1998; Vaux 2001; Yakup 2005

pɑlti-lir-i-dɑ axe-PL-POSS.3-LOC

• Uyghur is the only Turkic language reported to exhibit

transparency.

• Moreover, it is reported to exhibit transparency across a number
• f intervening neutral vowels.

Lindblad 1990; Hahn 1991, 1998; Vaux 2001; Yakup 2005

• The most common

harmonic alternation is [æ]~[ɑ]

• initial [i] triggers [-bk]

in some lexemes, but [+bk] in others

Suffix Backness Word Gloss PL [-back] bæl-lær waist-PL køl-lær lake-PL si-lær you-PL [+back] bɑl-lɑr face-PL jol-lɑr road-PL ʒil-lɑr year-PL LOC [-back] bæl-dæ waist-LOC køl-dæ lake-LOC [+back] bɑl-dɑ honey-LOC jol-dɑ road-LOC

• The most common

harmonic alternation is [æ]~[ɑ]

• initial [i] triggers [-bk]

suffixes in some lexemes, but [+bk] suffixes in others

Suffix Backness Word Gloss PL [-back] bæl-lær waist-PL køl-lær lake-PL si-lær you-PL [+back] bɑl-lɑr face-PL jol-lɑr road-PL ʒil-lɑr year-PL LOC [-back] bæl-dæ waist-LOC køl-dæ lake-LOC [+back] bɑl-dɑ honey-LOC jol-dɑ road-LOC

• The most common

harmonic alternation is [æ]~[ɑ]

• initial [i] triggers [-bk]

suffixes in some lexemes, but [+bk] suffixes in others

Suffix Backness Word Gloss PL [-back] bæl-lær waist-PL køl-lær lake-PL si-lær you-PL [+back] bɑl-lɑr face-PL jol-lɑr road-PL ʒil-lɑr year-PL LOC [-back] bæl-dæ waist-LOC køl-dæ lake-LOC [+back] bɑl-dɑ honey-LOC jol-dɑ road-LOC

• Non-initial /i/ vowels come in 3 types: underlying, epenthetic,

and derived

Suffix type Backness Word Gloss Underlying [-back] xæt-imiz-dæ letter-POSS.1P-LOC [+back] ɑt-imiz-dɑ horse-POSS.1P-LOC Epenthetic [-back] sællæ-m-dæ turban-POSS.1S-LOC xæt-im-dæ letter-POSS.1S-LOC [+back] ɑt-im-dɑ horse-POSS.1S-LOC

• Non-initial /i/ vowels come in 3 types: underlying, epenthetic,

and derived

Suffix type Backness Word Gloss Underlying [-back] xæt-imiz-dæ letter-POSS.1P-LOC [+back] ɑt-imiz-dɑ horse-POSS.1P-LOC Epenthetic [-back] sællæ-m-dæ turban-POSS.1S-LOC xæt-im-dæ letter-POSS.1S-LOC [+back] ɑt-im-dɑ horse-POSS.1S-LOC

• Non-initial /i/ vowels come in 3 types: underlying, epenthetic,

and derived

Suffix type Backness Word Gloss Underlying [-back] xæt-imiz-dæ letter-POSS.1P-LOC [+back] ɑt-imiz-dɑ horse-POSS.1P-LOC Epenthetic [-back] sællæ-m-dæ turban-POSS.1S-LOC xæt-im-dæ letter-POSS.1S-LOC [+back] ɑt-im-dɑ horse-POSS.1S-LOC

• Transparent vowels may also derive from low vowel raising
• In medial open syllables, low vowels raise to [+hi]

Backness Word Gloss [-back] sællæ turban sælli-lær turban-PL sælli-lir-imiz-dæ turban-PL-POSS.1P-LOC [+back] pɑltɑ axe pɑlti-lɑr axe-PL pɑlti-lir-imiz-dɑ axe-PL-POSS.1P-LOC

Hahn 1991; Vaux 2001; Yakup 2005

• Transparent vowels may also derive from low vowel raising
• In medial open syllables, low vowels raise to [+hi]

Hahn 1991; Vaux 2001; Yakup 2005

Backness Word Gloss [-back] sællæ turban sælli-lær turban-PL sælli-lir-imiz-dæ turban-PL-POSS.1P-LOC [+back] pɑltɑ axe pɑlti-lɑr axe-PL pɑlti-lir-imiz-dɑ axe-PL-POSS.1P-LOC

• Transparent vowels may also derive from low vowel raising
• In medial open syllables, low vowels raise to [+hi]

Hahn 1991; Vaux 2001; Yakup 2005

Backness Word Gloss [-back] sællæ turban sælli-lær turban-PL sælli-lir-imiz-dæ turban-PL-POSS.1P-LOC [+back] pɑltɑ axe pɑlti-lɑr axe-PL pɑlti-lir-imiz-dɑ axe-PL-POSS.1P-LOC

• Transparent vowels may also derive from low vowel raising
• In medial open syllables, low vowels raise to [+hi]

Hahn 1991; Vaux 2001; Yakup 2005

Backness Word Gloss [-back] sællæ turban sælli-lær turban-PL sælli-lir-imiz-dæ turban-PL-POSS.1P-LOC [+back] pɑltɑ axe pɑlti-lɑr axe-PL pɑlti-lir-imiz-dɑ axe-PL-POSS.1P-LOC

• Is /ɨ/ in the Uyghur inventory?
• A number of analyses argue for the inventory below:
• But they contend that /ɨ/ never surfaces due to an absolute neutralization

rule, |ɨ| → [i].

• back

+back

• round

+round

• round

+round +high i y ɨ u

• high, -low

(e) ø

• +low

æ ɑ

Linblad 1990; Hahn 1991; Yakup 2005; see also Vago 1973, 1976

• Is /ɨ/ in the Uyghur inventory?
• Vaux argues that /ɨ/ is not part of the inventory.
• Since /ɨ/ is not in the inventory, /i/ is unspecified for [back].
• Uyghur spreads only [-back]; [+back] is assigned late in the derivation
• back

+back

• round

+round

• round

+round +high i y u

• high, -low

(e) ø

• +low

æ ɑ

Linblad 1990; Hahn 1991; Yakup 2005; see also Vago 1973, 1976

Analysis Abstract /ɨ/ Non-contrastive [i]~[ɨ] Contrastive /i/-/ɨ/ Size of vowel inventory 9 at UR, 8 at SR 8 at UR and SR 9 at UR and SR Harmony characteristics Binary spreading of [±back] from all vowels Unary spreading of [-back] from contrastive vowels Spreading from all vowels Transparency Yes Yes No Lindblad 1990 Vaux 2001

Analysis Abstract /ɨ/ Non-contrastive [i]~[ɨ] Contrastive /i/-/ɨ/ Size of vowel inventory 9 at UR, 8 at SR 8 at UR and SR 9 at UR and SR Harmony characteristics Binary spreading of [±back] from all vowels Unary spreading of [-back] from contrastive vowels Spreading from all vowels Transparency Yes Yes No Lindblad 1990 Vaux 2001

Analysis Abstract /ɨ/ Non-contrastive [i]~[ɨ] Contrastive /i/-/ɨ/ Size of vowel inventory 9 at UR, 8 at SR 8 at UR and SR 9 at UR and SR Harmony characteristics Binary spreading of [±back] from all vowels Unary spreading of [-back] from contrastive vowels Spreading from all vowels Transparency Yes Yes No Lindblad 1990 Vaux 2001 This presentation

• Data was collected from 9

speakers (6 females, 3 males; age range 19-63, mean 44.4) from Shonzhy, Kazakhstan

• There are 3 main dialects

(Central, Southern, Eastern)

• Most research focuses on the

Central dialect

• Consultants speak the Central

dialect

Hahn 1998; Yakup 2005

• Stimuli were shown as randomly ordered pictorial prompts
• Stimulus items were mono- and disyllabic roots
• Speakers were then taught to associate certain visual cues with

grammatical categories to produce paradigms

• F2 was measured at vowel midpoint
• F2 was z-score normalized
• The data were analyzed using a linear mixed effects model with

random effects for speaker.

Lobanov 1971

• Previous analyses argue the following:
• 1. Surface [i] and [ɨ] are not contrastive.
• 2. (Triggers) There is no correlation between the

frontness/backness of initial [i]~[ɨ] and backness of following suffixes.

• 3. (Targets) Non-initial /i/ does not alternate for root backness.

Hahn 1986, 1991; Lindblad 1990; Vaux 2001; cf. Yakup 2005; Abdurehim 2013

• In initial syllables, there is a lot
• f variation.
• There is no obvious way to

define an allophonic relationship here.

• In particular, note the

difference between [ʃɨlɨm] and [ilim].

• In initial syllables, there is a lot
• f variation.
• There is no obvious way to

define an allophonic relationship here.

• In particular, note the

difference between [ʃɨlɨm] and [ilim].

Baker 1994 i ɪ ɨ ɯ

• In initial syllables, there is a lot
• f variation.
• The difference between [ʃɨlɨm]

and [ilim] is the best evidence against allophony.

i ɪ ɨ ɯ Baker 1994

• In initial syllables, there is a lot
• f variation.
• The difference between [ʃɨlɨm]

and [ilim] is the best evidence against allophony.

i ɪ ɨ ɯ Baker 1994

• In initial syllables, there is a lot
• f variation.
• The difference between [ʃɨlɨm]

and [ilim] is the best evidence against allophony.

• There is a significant

difference in F2 for these two roots, β = -0.48z, p < .001

i ɪ ɨ ɯ Baker 1994

• Previous analyses argue the following:
• 1. Surface [i] and [ɨ] are not contrastive.
• 2. (Triggers) There is no correlation between the backness of

initial [i]~[ɨ] and the backness of following suffixes.

• 3. (Targets) Non-initial /i/ does not alternate for root backness.

• Most of these lexemes

variably trigger backness harmony on low suffixes.

Root # [+bk] suffixes /total % [+bk] suffixes qiʃ 41/41 100% ʤɨl 39/39 100% ʃɨlɨm 41/48 85% ʧiʃ 39/48 81% pɨl 28/35 80% ilim 23/37 62%

• There is a moderate

correlation between F2 of the root and F2 of the suffix, r=.45, p < .001.

• This correlation is the

strongest for the roots, /ʃɨlɨm/, /pɨl/, and /ilim/.

• There is a moderate

correlation between F2 of the root and F2 of the suffix, r=.45, p < .001.

• This correlation is the

strongest for the roots, /ʃɨlɨm/, /pɨl/, and /ilim/.

• In categorical terms, F2
• f the root is also

predictive of the backness of a following low suffix, z= 4.2, p < .001.

• This tendency would be

surprising if the difference between [i] and [ɨ] was purely allophonic.

• It is not clear that the relationship between [i] and [ɨ] is

allophonic.

• Importantly, the surface backness of these vowels does correlate

with suffix backness.

• This would be surprising if differences were only allophonic.

• Previous analyses argue the following:
• 1. Surface [i] and [ɨ] are not contrastive.
• 2. (Triggers) There is no correlation between the backness of

initial [i]~[ɨ] and the backness of following suffixes.

• 3. (Targets) Non-initial /i/ does not alternate for root backness.

Hahn 1986, 1991; Lindblad 1990; Vaux 2001; cf. Yakup 2005; Abdurehim 2013

• Recall from earlier that reported transparent vowels come in

three types: underlying, epenthetic, and derived.

Vowel type Suffix Form Position Examples with /bɑl/ ‘honey’ Underlying ABL

• din

__{C,V} bɑl-din POSS.3

• i

__{C,V} bɑl-i-dɑ __# bɑl-i ACC

• ni

__# bɑl-ni Epenthetic POSS.1S

• im

__{C,V} bɑl-im Derived

• i

XC__CV bɑlilɑr bɑliliri

• Recall from earlier that reported transparent vowels come in

three types: underlying, epenthetic, and derived.

Vowel type Suffix Form Position Examples with /bɑl/ ‘honey’ Underlying ABL

• din

__{C,V} bɑl-din POSS.3

• i

__{C,V} bɑl-i-dɑ __# bɑl-i ACC

• ni

__# bɑl-ni Epenthetic POSS.1S

• im

__{C,V} bɑl-im Derived

• i

XC__CV bɑlilɑr bɑliliri

• Recall from earlier that reported transparent vowels come in

three types: underlying, epenthetic, and derived.

Vowel type Suffix Form Position Examples with /bɑl/ ‘honey’ Underlying ABL

• din

__{C,V} bɑl-din POSS.3

• i

__{C,V} bɑl-i-dɑ __# bɑl-i ACC

• ni

__# bɑl-ni Epenthetic POSS.1S

• im

__{C,V} bɑl-im Derived

• i

XC__CV bɑlilɑr bɑliliri

***

*** ***

Buckley 1997 *** **

• Underlying high

vowels significantly differ in F2 based on root backness

• β = -0.70z, p < .001
• Backness is neutralized

in absolute word-final position

• Takeaway: underlying

high vowels alternate for harmony

Buckley 1997

• Recall from earlier that reported transparent vowels come in

three types: underlying, epenthetic, and derived.

Vowel type Suffix Form Position Examples with /bɑl/ ‘honey’ Underlying ABL

• din

__{C,V} bɑl-din POSS.3

• i

__{C,V} bɑl-i-dɑ __# bɑl-i ACC

• ni

__# bɑl-ni Epenthetic POSS.1S

• im

__{C,V} bɑl-im Derived

• i

XC__CV bɑlilɑr bɑliliri

• Epenthetic vowels significantly

differ in F2 based on root backness

• β = -1.56z, p < .001
• Takeaway: Epenthetic vowels

alternate for harmony

***

• Recall from earlier that reported transparent vowels come in

three types: underlying, epenthetic, and derived.

Vowel type Suffix Form Position Examples with /bɑl/ ‘honey’ Underlying ABL

• din

__{C,V} bɑl-din POSS.3

• i

__{C,V} bɑl-i-dɑ __# bɑl-i ACC

• ni

__# bɑl-ni Epenthetic POSS.1S

• im

__{C,V} bɑl-im Derived

• i

XC__CV bɑlilɑr bɑliliri

• Raised vowels significantly

differ in F2 based on root backness

• β = -1.37z, p < .001
• Takeaway: Raised vowels

alternate for harmony

***

• All non-initial vowels examined, underlying, epenthetic, and

derived via raising exhibited alternations for backness.

• Some alternations were smaller, e.g. ABL
• Some were much more salient, e.g. raised vowels
• Thus, I

I conclu lude that har armony is is loc local, l, an and overt rtly ly so.

• Additional data supporting locality comes from consonant

alternations.

• It is well known that dorsal obstruents alternate for backness in

Uyghur.

• Velars occur with front vowels, and uvulars occur with back vowels.
• In neighboring Kazakh, all consonants are argued to alternate for

harmony.

• I measured F2 of laterals to further examine their status for

harmony.

bæl-gæ waist-DAT bal-ʁa honey-DAT

Hahn 1991; Dzhunisbekov 1980; McCollum 2015

• It is well known that dorsal obstruents alternate for backness in

Uyghur.

• Velars occur with front vowels, and uvulars occur with back vowels.
• In neighboring Kazakh, all consonants are argued to alternate for

harmony.

• I measured F2 of laterals to further examine their status for

harmony.

bæl-gæ waist-DAT bɑl-ʁɑ honey-DAT

Hahn 1991; Dzhunisbekov 1980; McCollum 2015

• It is well known that dorsal obstruents alternate for backness in

Uyghur.

• Velars occur with front vowels, and uvulars occur with back vowels.
• In neighboring Kazakh, all consonants alternate for harmony.
• I measured F2 of laterals to further examine locality in Uyghur.

bæl-gæ waist-DAT bɑl-ʁɑ honey-DAT

Hahn 1991; Dzhunisbekov 1980; McCollum 2015

• Laterals significantly differ in

F2 based on root backness

• β = -2.06z, p < .001
• Takeaway: Laterals alternate

for harmony, [l] ~ [ɫ].

***

• In roots, it is not clear that [i] and [ɨ] are allophones.
• Additionally, there is some evidence that these vowels behave

differently in the phonology.

• They tend to trigger suffixes in accordance with their backness.
• Non-initial high vowels (underlying, epenthetic, and derived) all

alternate for harmony.

• The lateral also alternates for [back], suggesting that harmony

may be strictly local in Uyghur.

• I propose that /i/ and /ɨ/ are contrastive, and as a result, the

Uyghur inventory looks like this:

• back

+back

• round

+round

• round

+round +high i y ɨ u

• high, -low

(e) ø

• +low

æ ɑ

• Why did my results differ so greatly from previous work?
• Dialectical differences?
• Perhaps contact from Kazakh is triggering phonological restructuring

among Uyghurs in Kazakhstan.

• Perhaps our results differ primarily to due to methodological differences.

Hahn 1998; Yakup 2005

• Why did my results differ so greatly from previous work?
• Dialectical differences?
• Contact from Kazakh may be triggering phonological restructuring.
• Perhaps our results differ primarily to due to methodological differences.

• Why did my results differ so greatly from previous work?
• Dialectical differences?
• Contact from Kazakh may be triggering phonological restructuring.
• Methodological differences?

Baker 1994

• More work on Uyghur is necessary to fully assess the nature of

the /i/-/ɨ/ contrast.

• Current results demonstrate very clearly that the propagation of

[back] is local, affecting all vowels and at least some types of consonants.

• Experimental work on other languages with reported

transparency should provide more insight into issues of locality more generally.

Abdurehim, Esmael. 2014. The Lopnor dialect of Uyghur: A descriptive analysis. Publications of the Institute for Asian and African Studies 17. Archangeli, Diana B., and Douglas George Pulleyblank. 1994. Grounded phonology. Vol. 25. MIT Press. Baker, Collin F. 1994. Uighur Vowels: An Acoustic and Perceptual Study. In Annual Meeting of the Berkeley Linguistics Society, vol. 20: 68-81. Bakovic, Eric. 2000. Harmony, dominance and control. PhD dissertation, Rutgers University. Benus, Stefan, and Adamantios I. Gafos. 2007. Articulatory characteristics of Hungarian ‘transparent’vowels. Journal of Phonetics 35.3: 271-300. Boyce, Suzanne E. 1990. Coarticulatory organization for lip rounding in Turkish and English. The Journal of the Acoustical Society of America 88.6: 2584-2595. Buckley, Eugene. 1997. Against vowel length in Tigrinya. Studies in African Linguistics 26.1: 63-102. Dye, Amanda. 215. Vowel Harmony and Coarticulation in Wolof and Pulaar: An Ultrasound Study. PhD dissertation, New York University. Dzhunisbekov, Alimkhan. 1980. Singarmonizm v kazakhskogo jazyka. Nauka: Alma-Ata. Gafos, Adamantios I. 1999. The articulatory basis of locality in phonology. Garland. Gick, Bryan, Douglas Pulleyblank, Fiona Campbell, and Ngessimo Mutaka. 2006. Low vowels and transparency in Kinande vowel harmony. Phonology 23:1: 1- 20. Goldsmith, John A. 1976. Autosegmental phonology. PhD dissertation, MIT. Hahn, Reinhard F. 1986. Modern Uighur language research in China: four recent contributions examined. Central Asiatic Journal 30: 35-54. Hahn, Reinhard F. 1991. Spoken Uyghur. University of Washington Press. Hahn, Reinhard F. 1998. Uyghur. In The Turkic Languages, Johanson, Lars & Csató, Éva Ágnes eds., Routledge: 379-396. Hayes, Bruce, and Zsuzsa Cziráky Londe. 2006. Stochastic phonological knowledge: The case of Hungarian vowel harmony. Phonology 23.1: 59-104.

Kaun, Abigail. 1995. An Optimality-Theoretic typology of rounding harmony. PhD dissertation, UCLA. Lindblad, Vern M. 1990. Neutralization in Uyghur. MA thesis, University of Washington. Lobanov, Boris M. 1971. Classification of Russian vowels spoken by different speakers. The Journal of the Acoustical Society of America 49: 606-608. Mahanta, Shakuntala. "Directionality and locality in vowel harmony." PhD diss., Utrecht University, 2007. McCollum, Adam G. 2015. Labial Harmonic Shift in Kazakh: Mapping the Pathways and Motivations for Decay." In Proceedings of the Annual Meeting of the Berkeley Linguistics Society, vol. 41: 329-351. Nadzhip, Ėmir Nadzhipovich. 1971. Modern Uigur. Nauka: Moscow. Ní Chiosáin, Máire, and Jaye Padgett. 2001. Markedness, segment realization, and locality in spreading. Segmental phonology in Optimality Theory: Constraints and representations, 118-156. Odden, David. 1994. Adjacency parameters in phonology. Language 70.2: 289-330. Rebrus, Péter, Péter Szigetvári, and Miklós Törkenczy. 2012. Dark secrets of Hungarian vowel harmony. Sound, structure and sense: Studies in memory of Edmund Gussmann, 491-508. Rebrus, Peter and Törkenczy, Miklos. 2016. Types and degrees of vowel neutrality. Linguistica, 56.1:239-252. Ritchart, Amanda, and Sharon Rose. 2017. Moro vowel harmony: implications for transparency and representations. Phonology34.1: 163-200. Szeredi, Daniel. 2016. Exceptionality in vowel harmony. PhD dissertation, New York University. Vago, Robert M. 1973. Abstract vowel harmony systems in Uralic and Altaic languages. Language 49:3: 579-605. Vago, R.M., 1976. Theoretical implications of Hungarian vowel harmony. Linguistic Inquiry, 7.2: 243-263. Vaux, Bert. 2001. Disharmony and derived transparency in Uyghur vowel harmony. In Proceedings of the North East Linguistic Society, vol. 30: 672-698. Yakup, Abdurishid. 2005.The Turfan Dialect of Uyghur. Vol. 63. Otto Harrassowitz Verlag.