Initial Phonological Analysis of Genç Zazaki Recordings

These notes describe recurring phonological and phonetic patterns in the Genç recordings from the May 2026 Zazaca Öğren export. They are not a full grammar of Genç Zazaki, and they should not be read as claims about every speaker from the area. The scope is narrower: document patterns that are well supported in this speaker’s recorded wordset data, compare them with relevant Zazaki dialectology and contact-language literature, and provide audio examples that can be checked directly.

The safest descriptive term for many of the cross-variety patterns below is phonological correspondence, not historical sound change. Directionality is clear in some loanword adaptations, but for these recordings versus other Zazaki varieties it is usually better to describe the observed correspondence unless independent historical evidence establishes a shift.

Corpus Basis

This analysis uses the live-site-scoped May 5, 2026 Genç dataset. The current transcription layer has practical orthography and IPA for all live Genç recording rows. The current IPA policy represents all literal t and d stop symbols as dental: t̪, d̪, t̪͡..., and d̪͡....

This dental convention is analytical and practical. Visual evidence strongly favors dental articulation for the recorded coronal stops, and the audio-only evidence is not reliable enough here to support two competing coronal-stop IPA layers.

Segmental Inventory Notes

Zazaki is commonly described with an eight-vowel system, and Todd’s Dimili grammar also gives an inventory of 30 consonants and 8 vowels with phonemic stress (Kırkan 2020, 5-6; Todd 2008 [1985], 20). Kırkan and Kızıl likewise describe an eight-vowel system while emphasizing regional differences in actual vowel realization and frequency (Kırkan and Kızıl 2020, 22).

For this dataset, the working IPA inventory is more phonetic than phonemic. The transcriptions preserve aspiration, secondary labialization/palatalization, affrication, uvulars, pharyngeal/laryngeal material, and rhotic allophony where they are useful for checking the audio. Those details should not all be promoted automatically into practical orthography.

Dental Coronal Stops

the IPA layer for this dataset treats ordinary t/d stops as dental by default:

IPA pattern	Audio examples
t̪ʰ	Play audiokûm qıc hê hêdi qalkên Play audiosegma rwece cûmê yo têm vinênsı ma vûn cûmê şıma xêrlib Play audioOw yo sıtıl naylûnin.
d̪	Play audioDarê guincu ha ça da? Play audioÊya yo mûryera. Play audioDarê palmya ha kûm ca da?
t̪͡ʃ	Play audioGuinc Play audioDarê guincu ha ça da? Play audioMûryer ha ça da?
d̪͡ʒ	Play audioÊyi guinc Play audioDarê guincu ha ça da? Play audioDarê palmya ha kûm ca da?

There is no separate IPA category in this dataset for a plain alveolar coronal stop. The current transcription policy uses t̪/d̪, including affricate onsets such as t̪͡... and d̪͡.... This is a policy for this current corpus, not a claim that every Zazaki variety lacks a coronal place contrast.

The tw/dw Labialization Series

the recordings have a recurring dental-plus-labial-release series that the current practical orthography represents with tw and dw. In IPA this series is usually written with reduced coarticulated stop symbols such as t̪͡p and d̪͡b. In clearer trill-like cases, it can be written t̪͡ʙ̥ or d̪͡ʙ.

Practical form	IPA pattern	Audio examples
twap / twapê	usually t̪͡p, sometimes more trill-like t̪͡ʙ̥	Play audioho twap mwecnen Play audioêya yo twapê ‘hewêna Play audiotwapê ‘hewên ha ça da
Twyer	reduced/coarticulated voiceless realization	Play audioTwyer Play audioTwyer ha ça da? Play audioÊya yo twyera.
dwılav / related dw- forms	voiced member of the same series	Play audiobwız dwılavi ha ça da Play audiodwir

For the current transcription policy, t̪͡p is used for the reduced or coarticulated voiceless realization. t̪͡ʙ̥ is reserved for cases where a sustained voiceless bilabial trill is clearly audible. The extra-short symbol t̪͡ʙ̥̆ and the aspirated spelling t̪͡pʰ are not ordinary inventory symbols in this dataset. Waveform/spectrogram checks of candidate labialization-series aspiration did not show a clean contrastive aspiration signal.

Kırkan’s broader work is useful here because it warns that some regionally real allophones are not represented directly in standard alphabetic practice (Kırkan 2020, 6; Arslan 2018, 5; Kırkan and Kızıl 2020, 18, 21). The recorded tw/dw series should therefore be treated as real phonetic/phonological evidence from this speaker, while keeping the orthographic representation simple.

Aspiration

Aspiration is frequent in the IPA layer, especially on voiceless stops:

Segment type	Audio examples
pʰ	Play audioÊya yo darê palmyawa. Play audioDarê palmyya Play audioÊya yo ampûla.
t̪ʰ	Play audiokûm qıc hê hêdi qalkên Play audiosegma rwece cûmê yo têm vinênsı ma vûn cûmê şıma xêrlib Play audioOw yo sıtıl naylûnin.
kʰ	Play audioIna yo darê kabaxa. Play audioKabax Play audioDarê kabax ha ça da?
qʰ	Play audioOw yo mâqes darûno. Play audioMâqês darûn Play audioMâqês darûn ho ça da?

For this practical orthography, aspiration is usually phonetic detail rather than a separate spelling contrast. The open question is narrower: whether the labialization-series t̪͡p is normally aspirated phonetically even when we write it without ʰ.

Affricates and Palatalization

The IPA transcriptions frequently show affricates such as t̪͡ʃ, d̪͡ʒ, and c͡ç. These are important for checking practical ç, c, and sometimes k-like spellings. Arslan notes that dialectal differences in Zazaki are especially dense in phonology, including consonantal alternations and palatalization patterns (Arslan 2018, 5, 9); Karacan also reports regional correspondences such as Palu z/c beside Siverek j in some forms (Karacan 2020, 8, 11).

IPA segment	Audio examples
t̪͡ʃ	Play audioGuinc Play audioDarê guincu ha ça da? Play audioMûryer ha ça da?
d̪͡ʒ	Play audioÊyi guinc Play audioDarê guincu ha ça da? Play audioDarê palmya ha kûm ca da?
c͡ç	Play audioêya yo kirazyera Play audioKirazyer Play audiokirazyer ha ça da

Laryngeal, Pharyngeal, and Uvular Material

The recordings include audible laryngeal or pharyngeal material in some words, especially in loans, emphatic forms, and words where the practical orthography uses an apostrophe or h-like marker. The IPA layer includes ʔ, ʕ, ħ, and occasionally more stop-like pharyngeal/epiglottal realizations.

IPA segment	Audio examples
ʔ	Play audioÊya yo twyera. Play audioIna darê çûma. Play audioow yo ‘elerebo
ʕ	Play audioma êy ra vûn ne’ene Play audioNa’ani Play audioOw yo ‘âmzik.
ħ	Play audioêya yo twapê ‘hewêna Play audiotwapê ‘hewên ha ça da Play audioOw yo destê re’hûno.
ʡ	Play audioÊyi ‘ârd. Play audio‘Ârd Play audio‘Ârd hê ça da?

Uvular and postvelar material is also productive in the data:

IPA segment	Audio examples
qʰ	Play audioOw yo mâqes darûno. Play audioMâqês darûn Play audioMâqês darûn ho ça da?
qʷ	Play audioOwe qwâp şêytûno. Play audioQwâp şêytûn Play audioQwâp şêytûn ho ça da?
qʷʰ	Play audioOw yo qwılbo. Play audioQwılbe Play audioQwılbe
χ	Play audioIna yo darê kabaxa. Play audioKabax Play audioDarê kabax ha ça da?

These details are important when comparing these forms with Arabic, Turkish, Kurmanji, or dictionary forms. They should be evaluated lexeme by lexeme rather than discarded mechanically.

Rhotics

the IPA layer for this dataset shows several rhotic realizations, including r, ɾ, ɹ, and occasionally voiceless or labialized variants. This is useful phonetic detail, but it should not automatically trigger practical spelling changes.

IPA segment	Audio examples
r	Play audioIna yo darê incirra. Play audioDarê incirr Play audio‘ela rebe ho ça da
ɾ	Play audioDarê guincu ha ça da? Play audioÊya yo mûryera. Play audioMûryer
ɹ	Play audioÊya yo nêrdıgûna. Play audioNêrdıgûn Play audioNêrdıgûn ha ça da?
ɾ̥	Play audioMûryer Play audioMûryer ha ça da? Play audioTwyer

Rounded-Vowel and Glide Correspondences

The strongest cross-variety pattern to foreground is not a single sound shift but a network of rounded-vowel/glide correspondences. The recorded data supports several useful working correspondences, especially rounded-vowel forms with we, o/we alternations, û corresponding to other o/a forms, and w/v or rounded-glide behavior.

Turkish or Other Rounded-Vowel Forms with we

the recordings have forms like çwel for Turkish çöl, and werdek for ördek:

Comparison	Recorded form	Audio examples
Turkish çöl	çwel / Çwêl	Play audioÇwel Play audioÇwel Play audioÇwêl
Turkish ördek	werdek	Play audioWerdek Play audioWerdek ha ça da? Play audiolirê werdek

This fits a broader Çewlîg/Bingöl pattern where rounded vowels often correspond to diphthongal or glide-bearing forms. Beltekin gives Çewlîg examples such as gual, gueş, and kue beside non-diphthongal standard spellings (Beltekin 2020, 8). Kırkan and Kızıl specifically connect the lower frequency of /o/ in Çewlîg to diphthongization and describe diphthongization as a distinctive feature of that dialect area (Kırkan and Kızıl 2020, 18, 21). Kırkan’s syllable work also notes that some heavier syllable patterns arise after /û/ changes to /w/ and that Bingöl material shows diphthongal syllables (Kırkan 2021, 8, 14).

In these recordings, çwel should therefore be treated as a regular-looking adaptation/correspondence rather than as an isolated transcription surprise.

The same loanword adaptation pattern is visible in the front rounded vowels. In the reviewed inventory subset, [ø] is rare and appears in Turkish-loan contexts, while Turkish ö is usually adapted with we, as in werdek, çwel, and gwel. The high front rounded vowel [y] also appears only in Turkish-loan contexts in the reviewed subset, and even there the dataset often adapts Turkish ü away from [y], either toward wı or toward final i; compare the project form wıti for Turkish ütü.

o / we and xw / w Patterns

The current Genç dataset includes dwe, kwel, and many xwe forms:

Recorded form	Audio examples
dwe	Play audioDwe Play audioDwe ho ça da?
kwel	Play audioKwel Play audioKwel bêr ho ça da?
xwe	Play audioXwe nimnen Play audioOw qıc ho bırê xwe ra xwe nimnen. Play audioHo twap hârzen bırê xwe.

Karacan reports regional w/v correspondences within Zazaki and also notes Kurmanji xw- forms corresponding to Zazaki forms where x drops before w (Karacan 2020, 10-11). That is not the same claim as the dataset’s xwe forms, but it is relevant background for why w, v, and rounded-glide material must be handled carefully in this dataset.

û Corresponding to Other o/a Forms

the recorded forms vûn, nûn, and sayû correspond to forms reported or expected elsewhere as von/non/sayon and van/nan/sayan. The current Genç dataset has many vûn and nûn examples, and sayû occurs in the “Adam elma satıyor” set:

Recorded form	Comparison set	Audio examples
vûn	compare von / van-type forms	Play audioma yemeg da vûn afyetıb Play audioSeg yo ten dest berz masê ma tı ra vûn ‘hûma razıb. Play audioÊya kerra ra ma vûn safir.
nûn	compare non / nan	Play audioNûn Play audioNûn ha ça da? Play audioHo nûn vırazen.
sayû	compare sayon / sayan	Play audiomwerik ho sayû rweşen Play audiosayû rweşen Play audiokûm ho sayû rweşen

This pattern is strongly compatible with published Bingöl/Çewlîg descriptions. Kalçık reports frequent Bingöl û where other regions or the Vate standard may have o or a (Kalçık 2011, 3). Beltekin specifically describes final an varying across Çewlîg/Bingöl-area speech: central Çewlîg and Darahenî use ûn, Bongilan uses on, and some village varieties retain an; final n may also drop (Beltekin 2020, 5). This directly supports treating the dataset’s vûn, nûn, and sayû as part of a regional correspondence set rather than isolated irregularity.

wağırıb and oğırıb

The current dataset includes both Wağırıb and Oğır be:

Recorded / comparison form	Audio examples
Wağırıb	Play audioWağırıb
Oğır be	Play audioOğır be

The relationship to oğırıb is best described as an initial rounded-glide/vowel correspondence: the recordings have a wa- onset in one formula and an o- onset in another closely related formula. This should not yet be framed as a unidirectional historical shift. It belongs in the same analytic space as the o/we, û/o/a, and v/w correspondences discussed above.

Transcription Layer Reflected Here

The examples in this post reflect the current project transcription layer. In practical orthography, the labialization-series consonants are represented with tw/dw. In the IPA layer, the reduced voiceless member is represented as t̪͡p, while t̪͡ʙ̥ is reserved for cases where a sustained trill is represented in the current data.

All literal IPA t/d stop symbols in this dataset are represented as dental: t̪, d̪, t̪͡..., and d̪͡.... Aspiration, initial glottal stop, rhotic quality, and laryngeal or pharyngeal material are treated as phonetic details in the IPA layer unless a separate lexical or orthographic reason supports representing them in practical spelling.

The cross-variety comparisons in this post use “correspondence” rather than “sound change” unless independent historical evidence establishes directionality.

Open Questions

More analysis is needed before answering these questions confidently:

1. Is the labialization-series stop normally aspirated in the recorded speech, even when represented as t̪͡p?
2. What acoustic threshold should separate reduced/coarticulated t̪͡p from sustained-trill t̪͡ʙ̥?
3. Which vowel correspondences are stable lexical or dialectal patterns, and which are surface variation in the recorded speech?
4. Which laryngeal/pharyngeal details in loans should be represented in practical orthography rather than only in IPA?
5. How should wağırıb beside oğırıb be modeled after more same-family examples are collected?

References

Arslan, İlyas. 2018. “Zaza Dilinde Lehçe Farklılıkları.” Bingöl Üniversitesi Yaşayan Diller Enstitüsü Dergisi 4(7). 51-66. https://izlik.org/JA42NF25DG.

Beltekin, Nurettin. 2020. “Nuştişê Zazakî de Problemê Nuştişê Fekan -Nimûneyê Fekê Çewlîgî-.” The Journal of Mesopotamian Studies 5(Special issue). 49-63. https://doi.org/10.35859/jms.2020.794263.

Kalçık, Hasan. 2011. “Bingöl Ağzının Bazı Özellikleri.” Zazaki.net. Published May 11, 2011. https://www.zazaki.net/d/file/bingolagzi.pdf.

Karacan, Hasan. 2020. “Kurmanji and Zazaki Dialects: Comparative Study on Their Phonetics.” International Journal of Kurdish Studies 6(1). 35-51. https://doi.org/10.21600/ijoks.653812.

Kırkan, Ahmet. 2020. “Asta Segmentale û Suprasegmental de Yewine-Vengê Zazakî.” Nubihar Akademi 4(14). 29-46. https://doi.org/10.55253/nubihar.856537.

Kırkan, Ahmet. 2021. “Zazakî de Babetê Heceyanê.” Şarkiyat 13(1). 353-366. https://doi.org/10.26791/sarkiat.859667.

Kırkan, Ahmet, and Hayreddin Kızıl. 2020. “Goreyê Peymanê Fonolojî Analîzkerdişê Zazakî; Muqayesekerdişê Vakur, Başûr û Merkezî.” The Journal of Mesopotamian Studies 5(1). 48-70. https://doi.org/10.35859/jms.2020.682589.

Todd, Terry Lynn. 2008 [1985]. A Grammar of Dimili, Also Known as Zaza. 2nd ed., rev., electronic publication. Stockholm: Iremet Förlag. https://www.zazaki.de/files/Terry-Lynn-Todd-A-grammar-of-Dimili.pdf.