Skip to content
Thursday, 19 November 2026 Today’s Dispatch
Ask
Thu 19 Nov 2026
Early colour photograph of a tiled minaret and ribbed dome of the Shir-Dor madrasa in Samarkand, covered in geometric patterns and a band of calligraphy in Arabic script.
Sergei Prokudin-Gorskii, Shir-Dor madrasa, Samarkand, between 1905 and 1915. Library of Congress, Prokudin-Gorskii Collection (LC-DIG-ppmsc-04440), no known restrictions on publication.
Dispatch · III.BA.CONT.01

Uzbekistan Is Rewriting Its Alphabet for a Different Machine

On 21 October Uzbekistan's president signed a law that replaces Oʻ, Gʻ, Sh and Ch with Ö, Ğ, Ş and Ç from April 2027. Officials defended it above all with what software does to an apostrophe-like mark that Uzbeks type in more than ten ways. The argument from machines is older than the law: three decades ago, the keyboard was part of the case for putting the apostrophe in.

9 November 2026 · 17 min read

On 21 October Uzbekistan’s president signed a law that replaces Oʻ, Gʻ, Sh and Ch with Ö, Ğ, Ş and Ç from April 2027. Officials defended it above all with what software does to an apostrophe-like mark that Uzbeks type in more than ten ways. The argument from machines is older than the law: three decades ago, the keyboard was part of the case for putting the apostrophe in.

On 21 October 2026, Uzbek Language Day, President Shavkat Mirziyoyev signed Law OʻRQ-1184, which gives Uzbek a Latin alphabet of 28 letters and one apostrophe-like sign in place of 26 letters and three letter combinations. Oʻ becomes Ö, Gʻ becomes Ğ, Sh becomes Ş and Ch becomes Ç. It enters into force on 22 April 2027.12

Uzbekistan’s 2026 reform is an explicit case of a state redesigning its script around what software can process: officials cited search, indexing, keyboards and speech technology more than any other reason. The argument is not new. In 1995 the keyboard helped bring the apostrophe into Uzbek; in 2026 search and language software take it out. Machines have become a reader whose needs governments cite when they set a spelling, and the machine that counts changes with each generation.

Three tests can falsify it:

  • The reform. The official text must give the date, the bodies that passed it, the letters changed and a transition period.
  • The justification. Digital reasons must be a main rationale in the law and in official statements. If identity or Turkic alignment dominates, the thesis narrows to rhetoric.
  • The pattern. If no earlier Uzbek script change was argued from machines, 2026 is the first machine-driven reform; if one was, the novelty lies only in which machine.

The first test passes. The second passes with a qualification. The third fails in its strong form, and the thesis above is the narrower version the evidence supports.

What Law OʻRQ-1184 changes

The law amends Law No. 931-XII of 2 September 1993, “On introducing the Uzbek alphabet based on the Latin script”. The Legislative Chamber adopted the amendment on 7 July 2026 and the Senate approved it on 10 September; the president signed it in Tashkent on 21 October, and it was entered in the National Legislation Database the same day.1 Kun.uz reported that the bill was not put to public discussion before the lower house voted, and that the session was streamed only on the chamber’s own website.3

Article 1 sets out the alphabet. Its letters are A, B, D, E, F, G, Ğ, H, I, J, K, L, M, N, O, Ö, P, Q, R, S, Ş, Ç, T, U, V, X, Y and Z, followed by the tutuq belgisi, the apostrophe that marks a glottal stop or a lengthened vowel, shown in the official table as U+02BC.1 The combination Ng loses its place as a unit of the alphabet. The two letters remain in writing, and their use is left to the spelling rules.4 The other letters do not change.

Article 2 governs the transition. Documents, banknotes and securities issued with the old letters stay valid and in circulation. State symbols, the relief of the state emblem on buildings and signs bearing place names may keep the old forms until they wear out, or until a deadline set in an implementation programme.1 Article 3 gives the Cabinet of Ministers two months to adopt that programme, and Article 4 sets entry into force at six months after official publication, which the database records as 22 April 2027.15

The schools move on their own timetable. Azizbek Turdiyev, deputy minister of preschool and school education, told the Senate that first-grade textbooks would be printed in the new alphabet once the law took effect in 2027, and that all textbooks would be in it by 2031. The change follows the existing replacement cycle: first-grade books are renewed every year, books for grades two to four every three years, and books for grades five to eleven every five years.6 Saida Mirziyoyeva, head of the Presidential Administration, described the approach at the signing as the path of “minimising costs”.7

III · THE KNOWLEDGE — BABEL

Four sounds written with two characters each will be written with one
Letters changed by Law OʻRQ-1184 (signed 21 October 2026, in force 22 April 2027), with their 1993 and 1995 forms and Unicode code points
1993 law1995–2027 formCode points (1995 form)2027 letterCode points (upper / lower)
ÕOʻU+004F + U+02BBÖU+00D6 / U+00F6
ĞGʻU+0047 + U+02BBĞU+011E / U+011F
ŞShU+0053 + U+0068ŞU+015E / U+015F
ÇChU+0043 + U+0068ÇU+00C7 / U+00E7
ŃNg (unit of the alphabet)U+004E + U+0067n + g, no longer a unitn/a
Letter count: 1993, 31 letters and an apostrophe; 1995, 26 letters and 3 combinations; 2027, 28 letters and one apostrophe-like sign
Source: Law OʻRQ-1184 and Law 931-XII in its 1993 and 1995 versions, lex.uz; code points computed by Titanica with the Unicode Character Database 15.1 (Python 3.13). The 1993 forms are as printed in the 1993 text on lex.uz. · Exhibit 01 · III.BA.CONT.01

The 1993 column matters for what follows. The first Latin alphabet of independent Uzbekistan, in the text adopted on 2 September 1993, had 31 letters and an apostrophe and already wrote three of these sounds with single marked letters, Ş, Ç and Ğ, along with Õ for the rounded vowel and Ń for the nasal.8 The version of 1995 replaced them with digraphs and apostrophe letters.89 In three of the four cases the 2026 law returns to a letter of the early 1990s. In the fourth it chooses Ö, the form used in Turkish and Azerbaijani, over the Õ of 1993.

The reasons on the record

The rapporteur in the Senate, Senator Odiljon Mamatkarimov, made three arguments. The first was phonetic: the reform brings Uzbek closer to “one sound, one letter”.6 The second concerned input: typing Oʻ and Gʻ on a computer can take up to six actions, according to a summary of the session by the Tashkent law firm Legal Alliance.4 The third concerned what happens after the text is typed: “modern information technologies do not perceive words with these letters as single words”. The mark is written in more than ten ways, and search engines and programs treat the same word as different words. The new system, he said, would allow national keyboards and artificial-intelligence models to be developed three times faster, and the alphabet would conform to ISO standards.10 He also said the reform “is not just a technical change”.11

The officials who answered senators’ questions spoke in the same register. Nodirjon Kholbutaev, head of the Department for the Development of the State Language under the Cabinet of Ministers, called the current Oʻ and Gʻ “a splinter in the body of our alphabet” and a barrier to digitising the language, and said a single key would replace the “five or six actions” now needed to type them.12 He said the draft was the fifth version prepared in ten years.13 Nurulla Abdullaev, adviser to the Minister of Digital Technologies on the state language, said that after adoption a single technical standard would be drawn up. State databases would be audited and Oʻ, Gʻ, Sh and Ch replaced automatically, data-exchange systems and electronic document workflows adapted, citizens’ data migrated, and search and identification mechanisms improved.13

The preamble of the law states its purpose in two parts: to refine the alphabet and the spelling rules on the basis of research by specialists, and to widen “the opportunities for using and digitalising the state language”.1 At the signing, Saida Mirziyoyeva said the law corrected “a long-standing historical mistake”, that there had been no other way to remove the shortcomings of the written language, and that it would “strengthen the position of the native language in the digital space”.7 A government report of a press conference at the Academy of Sciences on 29 October described the law’s aim as removing technical difficulties and making the language easier to adapt to digital technologies and computer programs.14

Turkic compatibility is also on the record, with less weight and fewer names. The Times of Central Asia reported that officials said the changes would improve compatibility with other Turkic languages that use the Latin script, and noted that Turkish and Azerbaijani already use several of the same letters.6 Krym.Realii, a service of Radio Free Europe/Radio Liberty, reported that alignment with other Turkic Latin-script states was expected to ease the exchange of books, translation, and cultural, scientific and economic links.15 In the reports consulted for this Dispatch, no named official made Turkic alignment the principal reason.

III · THE KNOWLEDGE — BABEL

Six of seven named official statements cite a digital reason; Turkic compatibility is attributed to unnamed officials
Reasons for the alphabet change in the law and in official statements, December 2023 to October 2026, as reported
StatementDateDigital or technicalTurkic compatibilityHistory, identity or phonetics
Saida Mirziyoyeva, then presidential assistant, conference on Jadidism11 Dec 2023No (the country’s name spelled in different ways)NoYes (the nation’s name and pride)
Sen. Odiljon Mamatkarimov, rapporteur10 Sep 2026Yes (word recognition, 10+ variants, keyboards, AI, ISO)Not in reports consultedYes (one sound, one letter)
Nodirjon Kholbutaev, State Language Department10 Sep 2026Yes (digitisation; one key for five or six actions)NoNo
Nurulla Abdullaev, Ministry of Digital Technologies10 Sep 2026Yes (database audit, search and identification)NoNo
Law OʻRQ-1184, preamble21 Oct 2026Yes (digitalising the state language)NoYes (specialists’ research on spelling)
Saida Mirziyoyeva, head of the Presidential Administration21 Oct 2026Yes (position in the digital space)NoYes (a historical mistake)
Government report, Academy of Sciences press conference29 Oct 2026Yes (technical difficulties, software)NoYes (ambiguity in spelling rules)
Unnamed officials, as reported by The Times of Central Asia and Krym.Realii10–11 Sep 2026YesYesn/a
Source: lex.uz; Gazeta.uz; UzDaily; Kun.uz; gov.uz; The Times of Central Asia; Krym.Realii (RFE/RL). Classification by Titanica from the reported wording; “No” means the reason does not appear in the reports consulted. · Exhibit 02 · III.BA.CONT.01

On this record the second test passes. Every named statement made while the bill was before parliament and at its signing cites a digital or technical reason, and the law writes digitalisation into its preamble. The qualification is that the technical reason never stands alone. It is paired with phonetics, with spelling, and at the signing with history. The earliest statement in the table, Saida Mirziyoyeva’s speech of December 2023, framed the problem through the country’s own name, which, she said, was being written in different ways in different places, and called for the alphabet question to be closed.16

Seven marks that look like one

The claim about variants can be tested. The correct character for the mark in Oʻ and Gʻ is U+02BB, MODIFIER LETTER TURNED COMMA; the tutuq belgisi is U+02BC, MODIFIER LETTER APOSTROPHE.17 Neither is easy to type on common keyboard layouts, and Uzbek websites, including government sites, often use the left single quotation mark U+2018 or the typewriter apostrophe U+0027 instead.1718 Kursiv Uzbekistan noted in August that the substitutes look almost identical and are stored as different strings, which complicates exact-match search, sorting and data cleaning unless software normalises the text.19

Titanica wrote the word Oʻzbekiston with seven of these marks and ran two tests. The first used Python’s standard Unicode database and its default word pattern, the kind of rule a basic tokenizer applies. The second used SQLite’s FTS5 full-text index with its default tokenizer, a widely deployed search engine that applies the same kind of rule to every document it stores.20

III · THE KNOWLEDGE — BABEL

The same word written with seven look-alike marks is stored as seven different strings
The word Oʻzbekiston written with each mark: Unicode properties, a standard word pattern and a full-text search index, tested 9 November 2026
MarkCode point and nameUnicode categoryEqual to the U+02BB form, before or after NFKCPython word pattern \w+FTS5: documents found by a search for this form
ʻU+02BB modifier letter turned comma (official)Lm (letter)ItselfOne wordOnly itself
ʼU+02BC modifier letter apostropheLm (letter)NoOne wordOnly itself
‘U+2018 left single quotation markPi (punctuation)No“O” + “zbekiston”The five punctuation and symbol forms
’U+2019 right single quotation markPf (punctuation)No“O” + “zbekiston”The five punctuation and symbol forms
‘U+0027 apostrophePo (punctuation)No“O” + “zbekiston”The five punctuation and symbol forms
`U+0060 grave accentSk (symbol)No“O” + “zbekiston”The five punctuation and symbol forms
´U+00B4 acute accentSk (symbol)No“O” + “zbekiston”The five punctuation and symbol forms
ÖU+00D6, the 2027 letterLu (letter)n/aOne wordÖzbekiston and Ozbekiston (default setting folds diacritics)
Source: computed by Titanica with Python 3.13.18 (Unicode Character Database 15.1.0) and SQLite 3.53.1 FTS5, default unicode61 tokenizer, on a test index of nine one-line documents. · Exhibit 03 · III.BA.CONT.01

The results bear out the rapporteur’s description and add a detail. None of the seven forms equals another, and the compatibility normalisation NFKC, which software uses to merge characters that differ only in presentation, leaves all seven distinct. The marks also fall into two families. The two modifier letters count as parts of a word, so the word survives as one token, but each matches only itself. The five punctuation and symbol marks split the word in two. In the search index those five forms find one another, because each is reduced to the fragments “o” and “zbekiston”, and none of them finds the word written with the official character. In this index a user who types any of the five substitutes misses every document written with U+02BB, and the reverse.

The variants are visible in official writing. The title of Law OʻRQ-1184 on lex.uz, as indexed on 9 November 2026, writes the law’s own number with U+2018 and the five other instances of the letter, including the word for “Uzbek”, with U+02BB.20 A government headline about the reform on gov.uz of 29 October uses U+2018 twice.1420 A headline of the national news agency UzA on the Senate vote uses U+2018 for the letter and U+2019 for the tutuq belgisi, and the web address of an UzDaily report on the lower-house vote uses U+02BB and U+02BC.20 The consolidated 1995 text of the 1993 law, as displayed on lex.uz, renders the mark as a grave accent.8 These are five examples, not a sample of Uzbek text, and they show only that the problem described in the Senate appears in the state’s own publications.

The test also shows the limits of the fix. Each new letter has a single precomposed code point, which removes the ambiguity of the comma-shaped mark.19 An Ö can still be typed as O followed by a combining diaeresis, but the standard normalisation NFC turns that sequence into the precomposed letter, which no standard normalisation does for the seven marks above.20 In its default setting, though, the same search index folds diacritics, so a search for Özbekiston also finds Ozbekiston, and Şahar finds Sahar; the setting can be switched off.20 The reform turns a problem without a standard remedy into one with a standard remedy, provided software is configured for Uzbek. Kursiv’s caveat applies to both alphabets: keyboards, fonts, databases, search indexes and older content still have to be updated.19

Five scripts in a century

Uzbek has been written in a reformed Arabic script, a Latin alphabet of the Soviet 1920s, Cyrillic, the Latin alphabets of 1993 and 1995, and now a revised Latin alphabet, and each change was argued at the time. In the early 1920s a reformed Arabic orthography added vowels that the classical script left unwritten.1721 After the First Turkological Congress in Baku approved a Latin alphabet in 1926, Latin letters entered Uzbek newspapers, journals and secondary schools from 1928, and a 34-character alphabet was approved in Samarkand in May 1929.1722 The 1993 law itself dates the first Latin period to “1929–1940”.8

The reasons offered then mixed ease of learning with politics. William Fierman’s study of Uzbek language planning, as cited by Ryan Schweitzer, quotes a leader of the Communist Party of Uzbekistan in 1932: “We know that the Arabic alphabet and Arabic writing were difficult – and so we discarded them.” The same speaker went on to present difficulty as only one of the reasons.2223 Eurasianet has summarised the background as Moscow’s association of the Arabic script with Islam.24 A decree of 8 May 1940 moved Uzbek to Cyrillic. The Diplomat describes the aims as distancing the Turkic republics from a Westernising Turkey and from pan-Turkism, and making Russian easier to teach.21

III · THE KNOWLEDGE — BABEL

Each change of script had a stated reason; arguments from communication enter the record in 1993
Scripts and alphabet versions for Uzbek, 1920s–2031, with the reason recorded at the time or by later accounts
YearChangeReason recordedDeadline and outcome
Early 1920sReformed Arabic orthographyVowels left unwritten in the classical scriptReplaced by 1929
1926–1929Latin: Baku congress 1926; schools and press from 1928; alphabet approved May 1929Arabic writing “difficult”, among other reasons (party leader, 1932); Moscow’s view of the Arabic script and IslamEnded 1940
1940Cyrillic, decree of 8 MayDistance from Turkey and pan-Turkism; easier teaching of RussianOfficial script until 1993; still widely used
1993Latin with diacritics, Law 931-XII of 2 SeptemberProgress of the republic and “entry into the system of world communication”Transition initially due by 2000
1995Latin with digraphs and apostrophe letters, amendment of MayFit the alphabet to the Latin keyboard (think-tank account)Deadline 1 Sep 2005, later 2010
2019–2021Drafts closer to 1993; road map to 2030Not consistently statedOfficial documents by 1 Jan 2023, not met
2026Law OʻRQ-1184: Ö, Ğ, Ş, ÇDigitalisation, word recognition, search, keyboards, AI; one sound, one letter; a historical mistakeIn force 22 Apr 2027; textbooks 2027–2031
Source: Law 931-XII (1993 and 1995 versions) and Law OʻRQ-1184, lex.uz; Schweitzer, CAP Paper 231, citing Fierman (1991); The Diplomat (2023); Eurasianet (2017); Eurasian Research Institute (2020); The Times of Central Asia (2026); Wikipedia, “Uzbek alphabet”. · Exhibit 04 · III.BA.CONT.01

Independence brought the argument from communication. The preamble of the 1993 law presents the Latin alphabet as creating conditions for the progress of the republic and for “its entry into the system of world communication”.8 Eurasianet, reviewing the reform in 2017, listed among the supporters’ reasons that familiarity with Latin letters would ease the adoption of modern technology such as computers.24 The transition was initially due to be completed by 2000.25

The revision of May 1995 replaced the marked letters with digraphs and apostrophe letters. The texts consulted for this Dispatch give no official reason for it. The Eurasian Research Institute, a think tank in Almaty, has written that the amendment aimed at the Latin keyboard, with double letters and apostrophes standing in for the special signs of 1993.9 Schweitzer’s paper for George Washington University’s Central Asia Program adds that Uzbekistan declined to follow the Turkish alphabet closely after relations with Turkey deteriorated in 1993.22 On the first account, the alphabet that the 2026 law corrects was shaped in part by a machine, the keyboard of the 1990s.

The deadlines then moved. The 1995 version set 1 September 2005 for the use of the Latin alphabet throughout the country; the date was later postponed to 2010.925 A presidential decree of October 2020 set a road map for language policy to 2030, and in February 2021 the government set 1 January 2023 as the date for official documents to switch, a date that also passed.91721 Kholbutaev’s remark that the 2026 bill was the fifth version in ten years describes this period.13

The third test therefore fails as originally framed. The 2026 law is not the first Uzbek script reform argued from machines. The 1993 change invoked world communication and, in its supporters’ case, computers, and the 1995 change, by the best available account, followed the keyboard. What changed in 2026 is the machine in question. The 1995 problem was entering characters on a keyboard built for 26 Latin letters. The 2026 problem is what software does with characters once they have been entered, at a time when Unicode gives Ş, Ç, Ğ and Ö a single code point each. The apostrophe eased the first problem and created the second.

Kazakhstan ran the argument in both directions

Kazakhstan offers the closest parallel and the clearest evidence that the technical argument cuts both ways. In April 2017 President Nursultan Nazarbayev called for a switch to the Latin alphabet, which the Astana Times reported as a move to better accommodate “the technological environment, communications, scientific and educational processes”.26 Decree No. 569 of 26 October 2017 approved an alphabet that avoided diacritics and digraphs and marked nine letters with apostrophes.27 The official line, as described by the anthropologists Alex Warburton and Meghanne Barker, was that the letters could be typed on a standard keyboard and entered into a computer at once.28

The apostrophe alphabet lasted four months. Technology workers pointed out that web addresses, hashtags, logins, passwords and e-mail addresses do not accept apostrophes.28 Linguists complained that apostrophes divided some words into four or five parts. A revised version, published on 20 February 2018 under Decree No. 637, replaced the apostrophes with diacritical marks, and the minister said the change followed recommendations from IT specialists and other experts.2927 The encoding question had already surfaced. On the Unicode Consortium’s public mailing list in January 2018, the participant Christoph Päper noted that official material circulated in Microsoft Word format showed the curly U+2019 instead of the keyboard apostrophe U+0027.30

A further version in January 2021 introduced ä, ö, ü, ğ, ū, ş and a nasal letter that became ñ in April, bringing Kazakh closer to the common Turkic alphabet, and the deadline moved from 2025 to 2031.2731 In April 2026 the Astana Times described the transition as continuing in phases toward completion in 2031.32

III · THE KNOWLEDGE — BABEL

The keyboard argument added apostrophes; later revisions removed them
Latin-script alphabets for Uzbek and Kazakh, 1995–2027, and the common Turkic alphabet of 2024: form chosen and reason given
AlphabetForm for special soundsReason givenOutcome
Uzbekistan, May 1995Apostrophe letters (Oʻ, Gʻ) and digraphs (Sh, Ch, Ng)Fit the Latin keyboard (think-tank account)In use 1995–2027
Kazakhstan, Decree 569, Oct 2017Nine letters with apostrophesTyped on a standard keyboard, “no hooks”Replaced after four months
Kazakhstan, Decree 637, Feb 2018Acute accents and digraphsApostrophes split words; advice of IT specialistsRevised 2019–2021
Kazakhstan, 202131 letters, including ä, ö, ü, ğ, ū, ş, ñCloser to the common Turkic alphabetCompletion targeted for 2031
Common Turkic alphabet, Sep 202434 letters, including Ç, Ğ, Ö, Ş, ÑUnderstanding among Turkic peoplesA commission proposal
Uzbekistan, Law OʻRQ-1184, 2026Ö, Ğ, Ş, Ç; Ng dropped as a unit; no ÑDigitalisation, search, keyboards, AIIn force 22 Apr 2027
Source: Eurasian Research Institute (2020); Wikipedia, “Kazakh alphabets”; Warburton and Barker, Society for Linguistic Anthropology (2018); RFE/RL (20 Feb 2018); Anadolu Agency (22 Apr 2021; 11 Sep 2024); Wikipedia, “Common Turkic alphabet”; Law OʻRQ-1184, lex.uz. · Exhibit 05 · III.BA.CONT.01

The sequence in Kazakhstan reproduces in four years the Uzbek sequence of three decades. A government chose apostrophes because they could be typed, and then found that a letter easy to type was hard to use in addresses, logins and long words. Both countries moved to letters with diacritics, and in both the endpoint resembles the common Turkic alphabet agreed in Baku in September 2024 by a commission working with the Turkic Academy of the Organization of Turkic States.3334

Where the technical case ends

The counter-argument

The strongest objection runs as follows. Alphabet reforms in this region have always been political, from the Latin of the 1920s to the Cyrillic of 1940, and the technical argument of 2026 is the acceptable language for a choice of alignment with Türkiye and the Turkic states and of distance from the Soviet script. The software problem had a software remedy, since a search engine or a government database can map every variant of the mark to U+02BB, so the decision to change the letters themselves is a decision about identity. The outcome also points one way: all four new letters appear in the common Turkic alphabet, Turkish and Azerbaijani already use several of them, and Ö was chosen over the Õ of 1993.

Much of this is supported by the record. Turkic compatibility is part of the official case, as reported.615 Saida Mirziyoyeva’s language at the signing, “a historical mistake”, is the language of correction, and her 2023 speech tied the issue to the nation’s name and pride.716 Kursiv’s report confirms that the variant problem can be handled by normalisation, and the test in Exhibit 03 shows that the new letters also need configuration.19 Uzbek commentary has placed the law in the regional story: Umida Yakupova, writing in Anhor.uz on 25 October, traced the reform to the Jadid reformers and set it beside the 2024 Baku agreement and Kazakhstan’s decree, while noting that a shared alphabet remains an aspiration.35

Other parts of the record cut the other way. The 2026 law does not adopt the common Turkic alphabet. That alphabet has Ñ for the nasal sound. Uzbekistan dropped Ng as a unit and kept the two letters, and Krym.Realii reported the reason given as that the combination does not begin Uzbek words.1534 The digital reasons are the most specific in the record: they name the number of keystrokes, the variants of the mark, word recognition, database audits and speech technology, while the Turkic argument is reported without a named speaker. The 1995 alphabet, which according to Schweitzer avoided following Turkish letters at a time of strained relations with Ankara, was by the Eurasian Research Institute’s account a keyboard alphabet.922

The balance supports a narrower conclusion than the working thesis. The reform has more than one motive and its letters serve more than one purpose. What distinguishes the 2026 record is the order of the argument. In the Senate and in the law’s preamble the problem was presented first as one of machine-readability, and the remedy was argued in terms a software engineer would recognise. Whatever weight the Turkic and historical motives carry, the government chose to make its case in the language of search, indexing and speech technology.

What the transition leaves open

The law does not end the use of two scripts, or now of three alphabets. Documents, money and signs in the old forms remain valid.1 Cyrillic has remained common in government paperwork, newspapers and publishing since 1993, and Krym.Realii reported in September that the Latin script is used mainly by young people while a significant part of the adult population still uses Cyrillic.2215 The Times of Central Asia noted that during the transition the same Uzbek words could appear in several written forms.6 The variant problem that the law addresses at the level of one mark will exist for some years at the level of whole spellings: Özbekiston, Oʻzbekiston with its look-alike marks, and Ўзбекистон.

Criticism has concentrated on method and cost. Komil Jalilov, an education expert quoted by the Times of Central Asia, asked whether the effects had been studied enough, called for comparative studies of reading and comprehension across age groups, and raised the costs for publishers and businesses and the burden on people already literate in the current alphabet. The newspaper observed that the Senate’s public explanation focused on technical problems and implementation and did not detail such studies.6 Nigora Umarova, a staff writer at Kursiv Uzbekistan, expected adaptation to take at least a generation. The translator Nazarbek Nazarov said he had first opposed the reform on cost and then changed his view, while warning that implementation on the old pattern would leave people unsure which alphabet to use and when.19 The literary scholar Dilbar Khaydarova told the Times of Central Asia in August that the reform would produce results only if it began at the school level.25

The implementation programme required by Article 3 is the next document to watch. The Ministry of Digital Technologies has described a single technical standard and automatic replacement in state databases as the first steps.113 Maqsud Asadov, director of the Institute of the Uzbek Language, Literature and Folklore, said at the Academy of Sciences that no document would have to be changed before its normal replacement date.14 Whether the change takes hold will be measured in the textbook cycle that ends in 2031, in the databases the ministry has promised to convert, and in the share of everyday writing that moves from Cyrillic.

The apostrophe entered the Uzbek alphabet in 1995 because, among other reasons, it could be typed. It leaves in 2027 because, among other reasons, software could not be relied on to read it as a letter. In both decisions the government cited the needs of a machine. The machine of 2026 does more than receive keystrokes. It indexes, searches and listens, and the four letters signed into law on Uzbek Language Day were chosen, on the government’s own account, partly so that it would recognise a word as one word.

  1. Republic of Uzbekistan, Law No. OʻRQ-1184 of 21 October 2026, “On amending the Law of the Republic of Uzbekistan ‘On introducing the Uzbek alphabet based on the Latin script'”: adopted by the Legislative Chamber 7 July 2026, approved by the Senate 10 September 2026, entered in the National Legislation Database 21 October 2026; preamble and Articles 1–4; in force 22 April 2027 (official). https://lex.uz/docs/8548754 (accessed 9 Nov 2026) ↩︎
  2. Government portal of the Republic of Uzbekistan, “Президент подписал закон о совершенствовании узбекского алфавита на основе латинской графики” (21 Oct 2026) (official). https://gov.uz/ru/migration/news/view/236403 (accessed 9 Nov 2026) ↩︎
  3. Kun.uz, “Özbekistonda alifbo islohoti böyiça qonun qabul qilindi” (7 Jul 2026) (press report). https://kun.uz/news/2026/07/07/ozbekistonda-alifbo-islohoti-boyicha-qonun-qabul-qilindi-4cf07b (accessed 9 Nov 2026) ↩︎
  4. Legal Alliance, “Senate approves a new 28-letter Uzbek alphabet: Ö, Ğ, Ş and Ç to replace Oʻ, Gʻ, Sh and Ch” (10 Sep 2026) (commentary, law firm). https://legalalliance.uz/en/news/novyy-uzbekskiy-alfavit-28-bukv-senat (accessed 9 Nov 2026) ↩︎
  5. Norma.uz, “Утвержден обновленный узбекский алфавит” (22 Oct 2026), on Law ZRU-1184 (press report, legal information portal). https://www.norma.uz/novoe_v_zakonodatelstve/utverjden_obnovlennyy_uzbekskiy_alfavit (accessed 9 Nov 2026) ↩︎
  6. Sadokat Jalolova, “Uzbekistan’s Senate Approves Latin Alphabet Changes”, The Times of Central Asia (11 Sep 2026) (press report). https://timesca.com/uzbekistan-latin-alphabet-changes-senate/ (accessed 9 Nov 2026) ↩︎
  7. Gazeta.uz, “Президент подписал закон о новом варианте узбекского алфавита в День узбекского языка” (21 Oct 2026), with statements by Saida Mirziyoyeva (press report). https://www.gazeta.uz/ru/2026/10/21/uzbek-alphabet/ (accessed 9 Nov 2026) ↩︎
  8. Republic of Uzbekistan, Law No. 931-XII of 2 September 1993, “On introducing the Uzbek alphabet based on the Latin script”, comparison of the versions of 12 October 1993 and 17 June 1995, Uzbek text (https://lex.uz/en/docs/-112286?ONDATE=12.10.1993&ONDATE2=17.06.1995&action=compare) and Russian text (https://lex.uz/en/docs/112283?ONDATE2=17.06.1995&action=compare) (official). Accessed 9 Nov 2026. ↩︎
  9. Balzhan Eshmetova, “Uzbekistan’s Transition to Latin Alphabet and Lessons Learned”, Eurasian Research Institute (29 Nov 2020) (analyst estimate and commentary). https://eurasian-research.org/publications/uzbekistans-transition-to-latin-alphabet-and-lessons-learned (accessed 9 Nov 2026) ↩︎
  10. Gazeta.uz, “Сенат одобрил новый вариант узбекского алфавита из 28 букв” (10 Sep 2026), report of the Senate’s plenary session (press report). https://www.gazeta.uz/ru/2026/09/10/uzb-alphabet/ (accessed 9 Nov 2026) ↩︎
  11. Gazeta.uz, “Senate approves new version of Uzbek alphabet of 28 letters” (10 Sep 2026) (press report). https://www.gazeta.uz/en/2026/09/10/uzb-alphabet/ (accessed 9 Nov 2026) ↩︎
  12. Gazeta.uz, “Как предлагается разместить новые буквы узбекского алфавита на клавиатурах” (10 Sep 2026) (press report). https://www.gazeta.uz/ru/2026/09/10/keyboard-uzb/ (accessed 9 Nov 2026) ↩︎
  13. Anvar Umarov, “Госсистемы Узбекистана адаптируют под новый алфавит”, UzDaily.uz (10 Sep 2026) (press report). https://www.uzdaily.uz/ru/gossistemy-uzbekistana-adaptiruiut-pod-novyi-alfavit/ (accessed 9 Nov 2026) ↩︎
  14. Government portal of the Republic of Uzbekistan, “Yangi o‘zbek alifbosi: o‘zgartirilgan 4 ta harf” (29 Oct 2026), report of a press conference at the Academy of Sciences (official). https://gov.uz/oz/icc/news/view/239623 (accessed 9 Nov 2026) ↩︎
  15. Krym.Realii (Radio Free Europe/Radio Liberty), “Сенат Олий Мажлиса Узбекистана одобрил новый вариант узбекского алфавита из 28 букв на основе латиницы” (10 Sep 2026) (press report). https://ru.krymr.com/a/news-oliy-mazhlis-uzbekistana-uzbekskiy-alfavit-latinitsa-novyy-variant/33851716.html (accessed 9 Nov 2026) ↩︎
  16. Kun.uz, “Saida Mirziyoyeva: alifbo masalasiga qat’iy nuqta qöyişimiz zarur” (11 Dec 2023) (press report). https://kun.uz/news/2023/12/11/saida-mirziyoyeva-alifbo-masalasiga-qatiy-nuqta-qoyishimiz-zarur (accessed 9 Nov 2026) ↩︎
  17. Wikipedia, “Uzbek alphabet”: history section and section on the characters for oʻ, gʻ and the tutuq belgisi (commentary, encyclopedia). https://en.wikipedia.org/wiki/Uzbek_alphabet (accessed 9 Nov 2026) ↩︎
  18. Wikipedia, “Oʻ” (commentary, encyclopedia). https://en.wikipedia.org/wiki/O%CA%BB (accessed 9 Nov 2026) ↩︎
  19. Oleg Son, “Four-letter reform: Uzbekistan’s new Latin alphabet wins cautious support”, Kursiv Uzbekistan (25 Aug 2026) (press report). https://uz.kursiv.media/en/2026-08-25/four-letter-reform-uzbekistans-new-latin-alphabet-wins-cautious-support/ (accessed 9 Nov 2026) ↩︎
  20. Computed by Titanica, 9 Nov 2026. Python 3.13.18 with its unicodedata module (Unicode Character Database 15.1.0): code points, names, general categories, NFC and NFKC normalisation, and the default re pattern \w+. SQLite 3.53.1, FTS5 virtual table with the default unicode61 tokenizer (and with remove_diacritics 0 for comparison); nine test documents: Oʻzbekiston written with U+02BB, U+2018, U+0027, U+02BC, U+2019, U+0060 and U+00B4, plus Özbekiston and Ozbekiston, each followed by “respublikasi”; a phrase query for each form. Code-point counts from title strings returned by web search on 9 Nov 2026 for lex.uz (Law OʻRQ-1184), gov.uz (the headline cited in the text) and UzA, uza.uz (“Senatorlar lotin yozuviga asoslangan o‘zbek alifbosini takomillashtirishga qaratilgan qonunni ma’qulladi”), and from the percent-encoded address https://www.uzdaily.uz/uz/deputatlar-o%CA%BBzbek-lotin-alifbosidagi-islohotni-ma%CA%BCqulladilar/ after decoding. Five strings; not a sample of Uzbek text. ↩︎
  21. “Uzbekistan’s Drawn-out Journey From Cyrillic to Latin Script”, The Diplomat (January 2023), on the reformed Arabic script, the decree of 8 May 1940, the 1993 law and the 2023 target (commentary). https://thediplomat.com/2023/01/uzbekistans-drawn-out-journey-from-cyrillic-to-latin-script/ (accessed 9 Nov 2026) ↩︎
  22. Ryan Michael Schweitzer, “Alphabet Transition in Uzbekistan”, CAP Paper No. 231, Central Asia Program, George Washington University (undated; cites sources to 2020) (analyst paper). https://centralasiaprogram.org/wp-content/uploads/2020/05/CAP-Paper-231.pdf (accessed 9 Nov 2026) ↩︎
  23. William Fierman, Language Planning and National Development: The Uzbek Experience, Contributions to the Sociology of Language 60 (Berlin and New York: Mouton de Gruyter, 1991), publisher’s record; the 1932 quotation is cited here through Schweitzer, who gives Fierman as his source (peer-reviewed monograph). https://www.degruyterbrill.com/document/doi/10.1515/9783110853384/html (accessed 9 Nov 2026) ↩︎
  24. Eurasianet, “Latin Alphabet in Uzbekistan: To B or Not to Б” (29 Mar 2017) (press report). https://eurasianet.org/latin-alphabet-in-uzbekistan-to-b-or-not-to-b (accessed 9 Nov 2026) ↩︎
  25. Sadokat Jalolova, “Uzbekistan’s Alphabet Reform Revives a 33-Year-Old Debate”, The Times of Central Asia (20 Aug 2026) (press report). https://timesca.com/uzbekistan-alphabet-reform-latin-cyrillic/ (accessed 9 Nov 2026) ↩︎
  26. The Astana Times, “Kazakh President lays out principles for modernisation of nation’s identity” (18 Apr 2017) (press report). https://astanatimes.com/2017/04/kazakh-president-lays-out-principles-for-modernisation-of-nations-identity/ (accessed 9 Nov 2026) ↩︎
  27. Wikipedia, “Kazakh alphabets”, on Decrees No. 569 (26 Oct 2017) and No. 637 (19 Feb 2018) and the 2021 version (commentary, encyclopedia). https://en.wikipedia.org/wiki/Kazakh_alphabets (accessed 9 Nov 2026) ↩︎
  28. Alex Warburton and Meghanne Barker, “Putting an End to the Rain of Apostrophes”, Society for Linguistic Anthropology (23 Nov 2018) (commentary, scholars). https://linguisticanthropology.org/blog/2018/11/23/putting-end-rain-apostrophes/ (accessed 9 Nov 2026) ↩︎
  29. Radio Free Europe/Radio Liberty, “No Apostrophe: Nazarbaev Decree Spells Out Changes In New Latin Alphabet For Kazakhstan” (20 Feb 2018) (press report). https://www.rferl.org/a/kazakhstan-latin-alphabet-drop-apostrophes/29050511.html (accessed 9 Nov 2026) ↩︎
  30. Christoph Päper, “Re: 0027, 02BC, 2019, or a new character?”, Unicode mailing list (17 Jan 2018) (commentary). https://unicode.org/mail-arch/unicode-ml/y2018-m01/0039.html (accessed 9 Nov 2026) ↩︎
  31. Anadolu Agency, “Kazakhstan opens new Latin alphabet to public opinion” (22 Apr 2021) (press report). https://www.aa.com.tr/en/asia-pacific/kazakhstan-opens-new-latin-alphabet-to-public-opinion/2217193 (accessed 9 Nov 2026) ↩︎
  32. The Astana Times, “How Kazakhstan’s Writing System Evolved Through History” (19 Apr 2026) (press report). https://astanatimes.com/2026/04/how-kazakhstans-writing-system-evolved-through-history/ (accessed 9 Nov 2026) ↩︎
  33. Anadolu Agency, “Turkic states reach agreement on common 34-letter alphabet” (11 Sep 2024) (press report). https://www.aa.com.tr/en/world/turkic-states-reach-agreement-on-common-34-letter-alphabet/3327560 (accessed 9 Nov 2026) ↩︎
  34. Wikipedia, “Common Turkic alphabet” (commentary, encyclopedia). https://en.wikipedia.org/wiki/Common_Turkic_alphabet (accessed 9 Nov 2026) ↩︎
  35. Umida Yakupova, “Harflar ortidagi siyosat”, Anhor.uz (25 Oct 2026) (commentary). https://anhor.uz/uzl/ichkaridan-nazar/harflar-ortidagi-siyosat (accessed 9 Nov 2026) ↩︎
Read through the Method
Context
Law OʻRQ-1184 of 21 October 2026, signed on Uzbek Language Day, which replaces Oʻ, Gʻ, Sh and Ch with Ö, Ğ, Ş and Ç from 22 April 2027; the Senate debate of 10 September 2026; a Titanica test of seven look-alike marks in a standard tokenizer and a full-text search index.
Causes
Variant characters for the mark in Oʻ and Gʻ that software stores as different strings; the cost of typing the letters; the phonetic principle of one sound, one letter; the unfinished transition from Cyrillic since 1993; the 1995 adaptation of the alphabet to the keyboard; Turkic compatibility as a secondary stated aim.
Actors
President Shavkat Mirziyoyev; Saida Mirziyoyeva, head of the Presidential Administration; the Legislative Chamber and Senate of the Oliy Majlis; Senator Odiljon Mamatkarimov; the Department for the Development of the State Language and Nodirjon Kholbutaev; the Ministry of Digital Technologies and Nurulla Abdullaev; the Ministry of Preschool and School Education; the Institute of the Uzbek Language, Literature and Folklore; Kazakhstan's presidency and its alphabet decrees of 2017 and 2018; the Turkic World Common Alphabet Commission.
Perspectives
The official record read for the order and specificity of the reasons given, with Turkic alignment and historical correction weighed as rival explanations; the claim of novelty tested against the 1993 and 1995 reforms and against Kazakhstan's apostrophe alphabet; computed tests labelled as tests on a small index, and five observed strings kept apart from any claim about Uzbek text in general; critics of method and cost reported alongside officials.
Relationships
Opens the Babel Domain. Links to Inference (Nvidia's Board Now Investigates a National Security Question, https://titani.ca/dispatch/nvidias-board-now-investigates-a-national-security-question/, on technical systems becoming a question of state) and to Gnosis (Guy Fawkes Was Never Just One Symbol, https://titani.ca/dispatch/guy-fawkes-never-just-one-symbol/, on one sign carrying successive meanings).

One phenomenon each morning. Free.