Why QazaqKey?

The stated intent

The Cyrillic Kazakh alphabet in use today was not a neutral piece of linguistic evolution. Archival Soviet-era footage of the decision states its purpose outright:

«Новый алфавит призван сблизить культуру русского и казахского народов»
"The new alphabet is meant to bring the cultures of the Russian and Kazakh peoples closer together" — in practice, Kazakh moving toward Russian, not the reverse.

Source: Azattyq (RFE/RL Kazakh service), "Как казахский язык переходил на кириллицу", whose own description reads: "Архивное видео о том, как это было" ("Archival video of how it was"). Azattyq is the Kazakh-language service of Radio Free Europe/Radio Liberty, a US state-funded international broadcaster operating since 1949. This is not inference or reconstruction — it is the policy's own stated purpose, preserved on film, at the time.

Before Cyrillic: Kazakh's own choice

Kazakh had already been written in reformed Arabic script (Akhmet Baitursynov's töte jazu, from 1912) and then in a Latin script (Jaŋalif, from 1929), and in both traditions Kazakh's own linguists chose to fully absorb loanwords into native phonology rather than cite them in their foreign spelling. The 1924 First Congress of Kazakh Scholars put it as one rule: foreign words should be spelled according to the rules of the Kazakh language.

Telzhan Shonanuly formalized this at the 1929 Qyzylorda orthography conference with 31 principles for handling foreign words — all assimilationist. A few of his examples: ф became p; и́nitial г became ç; ц became s (apeser for офицер, stansa for станция); the Russian -ция ending became -sa/-se (milijse for милиция); -ия became -a/-ä (burçuvaza for буржуазия). This was a deliberate, considered linguistic choice, made by Kazakh scholars, in the Latin-script era, before Cyrillic ever entered the picture.

Overridden by decree, not persuasion

In 1938, loan letters "for use in Russian words" (Ф, Х) were added and the nativized spellings were replaced with Russian citation forms — sovet, kolxoz-type spellings — by administrative decree, the same year Russian became a compulsory subject in Kazakh schools. The switch to Cyrillic in 1940 continued that decree; it was not a separate, later decision.

The same pattern, in the same years, played out in Kazakh's sibling languages. Kyrgyz linguist Konstantin Tynystanov's 1934 draft orthography nativized Soviet vocabulary into Kyrgyz phonology (qalqoz, doqtur), and in 1938 those spellings were purged as "distortions" and replaced with Russian forms (kolxoz, doktor) — the same year, the same justification, a different language. Nogai had moved even earlier: in 1931–33 several letters were added to its Latin script specifically to spell loan sounds, only for some to be stripped again by a 1935 workers' meeting under the same administrative pressure.

A live example of how the harm gets measured

This project's own process ran into the same bias directly, and it's worth describing honestly rather than editing out. Early analysis of one Kazakh spelling rule — how a Russian "soft sign" (ь) should be rendered — used frequency data from mainstream Kazakhstani text and concluded that a certain vowel-fronting pattern was rare, so the "correct" spelling should ignore it (медаль as medal, not medäl). That conclusion was wrong, and wrong in a specifically diagnostic way: the corpus was measuring Russified pronunciation and mistaking it for the baseline. The fronting is real and common — in Kazakh communities that didn't go through decades of Russian-medium schooling and media, in Mongolia, China, Iran, and Türkiye.

This is the general problem with using "what's common in today's Kazakhstani corpus" as ground truth: any method built that way will systematically reproduce the Russification it should be correcting for. QazaqKey instead treats diaspora and pre-reform practice as the reference point, and treats mainstream-corpus frequency as evidence of the damage, not a target to match.

It has not stopped

Citation-mode Russification is not a single historical event — it is a pattern that keeps re-asserting itself, including after independence. Kyrgyz orthographic history traces this precisely: nativized pre-revolutionary loans were grandfathered into the 1940 Cyrillic rules, abolished outright in 1953, partially re-admitted in 2002 under a declared "phoneticism" principle, then walked back toward the 1953 stance again in 2008. Across the wider Kipchak language family, the pattern is the same: citation spelling is always the cheaper option for a writer under any kind of pressure, so left unchecked it expands over time rather than receding — regardless of who is nominally in charge of the orthography.

Discussion of a Kazakh Latin alphabet goes back to 1991, when newly-independent Turkic states met in Istanbul to consider a shared script; Kazakhstan held back for economic and political reasons, and a government commission's own 2006–2007 study — a phased 12–15 year transition plan — went nowhere. Kazinform, the state news agency, built its own Latin romanization system in anticipation of a switch that was never officially adopted. The formal decree process alone has run since 2017 through four public drafts — 2017 (apostrophes), 2018, 2019, and 2021 (diacritics) — each addressing complaints about the last without resolving the underlying tension: an alphabet designed by committee, in an environment where Russian remains the dominant administrative and media language, keeps re-arriving at compromises that favor legibility to Russian readers over fidelity to Kazakh's own phonology.

What QazaqKey does differently

Two decisions follow directly from this history. First, QazaqKey's loanword handling is deliberately narrow: recent, unassimilated loans and foreign proper nouns get a separate, tightly-triggered "citation mode" that never touches native vowel harmony, and the native pipeline stays the unmarked default — the opposite of how citation spelling has historically expanded at the native system's expense. Second, every rule that reconstructs pronunciation (vowel-glide restoration, soft-sign fronting, and others) is checked against authentic, un-Russified speech first, and only measured against the modern Kazakhstani corpus second, as a check on frequency and coverage — not as the arbiter of what's "correct."

References

1. Azattyq (RFE/RL Kazakh service), archival footage on the 1940 Cyrillic decision — azattyqasia.org/a/alfavit1940-video (also on YouTube)
2. Odagiri, K. (2015). "How to Deal with the Elements of Russian Origin: Developments of Orthographic Reforms for the Kyrgyz Language." Kansai University (PDF) — kansai-u.ac.jp/fl/publication/pdf_department/12/69odagiri.pdf
3. Tiltanym journal, №3(95) 2024, on Telzhan Shonanuly's orthographic principles (PDF) — tiltanym.kz/jour/article/download/1622/721
4. "Kazakh and Turkic Alphabet Reform 1900–1939: Change Without Change," University of Illinois pressbook — iopn.library.illinois.edu — chapter 10
5. Kazakh alphabets — Wikipedia — en.wikipedia.org/wiki/Kazakh_alphabets
6. Nogai alphabets — Wikipedia — en.wikipedia.org/wiki/Nogai_alphabets
7. Nogai language — Wikipedia — en.wikipedia.org/wiki/Nogai_language
8. Romanization of Kyrgyz — Wikipedia — en.wikipedia.org/wiki/Romanization_of_Kyrgyz
9. Karakalpak language — Wikipedia — en.wikipedia.org/wiki/Karakalpak_language
10. Imla (qaraqalpaq) — Qaraqalpaq Wikipedia — kaa.wikipedia.org/wiki/Imla_(qaraqalpaq)
11. Omniglot: Karakalpak — omniglot.com/writing/karakalpak.php
12. RFE/RL, "Common Turkic Alphabet" coverage (2024) — rferl.org — common Turkic alphabet
13. Eurasianet, "Kazakhstan: The ABCs of the Alphabet Debate" — eurasianet.org — the alphabet debate
14. Kazinform — Wikipedia (its independent pre-2017 Latin romanization system) — en.wikipedia.org/wiki/Kazinform
15. Radio Free Europe/Radio Liberty — Wikipedia (Azattyq's parent organization) — en.wikipedia.org/wiki/Radio_Free_Europe/Radio_Liberty

The full internal research notes behind this page, including the process record for how QazaqKey's own rules were checked against this history, are kept in the project repository's docs/ folder.