Languages and writing systems of Uzbekistan

Uzbekistan today is home to at least fifteen daily-spoken languages, though four dominate the ear on the street. Uzbek is the state language and mother tongue of roughly 84% of the country's ~37 million residents^[1]. Russian, though it holds no legal status, remains the lingua franca of business and cross-ethnic communication — around 70% of urban adults can hold a conversation in it, and much of higher education, technical publishing, and Tashkent's white-collar workplace still runs in Russian^[9].

Karakalpak, a Kipchak Turkic language much closer to Kazakh than to Uzbek, is co-official inside the Republic of Karakalpakstan, the autonomous republic that covers a third of Uzbekistan's territory. Roughly 800,000 people speak it^[10]. Tajik — a Persian language, mutually intelligible with the Farsi of Iran and Dari of Afghanistan — is the everyday home language of a substantial fraction of Samarkand, Bukhara, and Surkhandarya families; official figures put Tajik speakers at 5% of the population, but community estimates run to 10–15% in the western oases.

Beyond the big four, Uzbekistan hosts several communities' languages. Around 175,000 Koryo-saram — Koreans deported from the Soviet Far East by Stalin in 1937 — live mostly in the Tashkent region and preserve a distinctive Koryo-mar dialect^[11]. Bukharan Jews still maintain synagogues in Bukhara and Samarkand, though the Bukhori (Judeo-Tajik) language they historically spoke is now endangered, with most fluent speakers having emigrated to Israel or Queens, New York^[12]. Kazakh, Kyrgyz, Turkmen, Tatar, and Uyghur are spoken in the border regions and specific mahallas. Uzbek Sign Language, distinct from ASL and partly influenced by Russian Sign Language, serves an estimated 40,000–50,000 people.

Then there are the living relict languages — small speech communities carrying older tongues. Yaghnobi, still spoken by roughly 12,000 people in the mountains of Tajikistan just across the border (and by a Yaghnobi diaspora in Uzbekistan), is the last living descendant of Sogdian — the Silk Road language of the caravan cities^[13]. Parya, an Indo-Aryan language brought by 18th- and 19th-century Indian migrants, is still spoken in a handful of villages of Kashkadarya and Surkhandarya by around 2,000 people — the only Indo-Aryan language of Central Asia^[14].

The Turkic language family has around 200 million speakers and spans a broad crescent from the Bosphorus to Yakutia^[1]. Linguists divide it into six branches; Uzbek belongs to the Karluk (or Southeastern) branch, which also includes modern Uyghur (spoken in China's Xinjiang region) and the extinct Chagatai literary language that unified Central Asian Turkic writing from the 15th to the 19th centuries^[6].

Kazakh, Kyrgyz, and Karakalpak belong to the Kipchak branch — mutually intelligible with each other but only partly with Uzbek. Turkish, Azerbaijani, and Turkmen belong to the Oghuz branch. An Uzbek speaker can typically read Turkish with effort but not carry a fluent conversation; the closest large language to modern Uzbek is Uyghur, which shares much of the Karluk lexicon and grammar.

One striking feature of standard modern Uzbek — the variety spoken in Tashkent, Samarkand, and Bukhara — is that it has lost the vowel harmony that most Turkic languages retain. This is a substrate effect: those cities are old Sogdian-Iranian oasis cities where Turkic-speaking newcomers settled on top of centuries of Persian and Sogdian speech, and the Iranian phonological system left permanent traces^[6]. Kipchak-influenced Uzbek dialects (spoken by more nomadic-descended communities in the northern steppes) still retain partial vowel harmony.

Uzbek's core is Turkic, but every wave of history that passed through Central Asia deposited a lexical layer on top. The result is one of the most contact-rich vocabularies in the Turkic family — an Uzbek speaker who says an ordinary sentence probably uses words from four different language families in the same breath.

The Persian layer is by far the largest — roughly a third of everyday vocabulary. Words for many concrete objects (non 'bread', oshxona 'kitchen', bog' 'garden'), abstract concepts (dil 'heart', jon 'soul', dost 'friend'), most professions ending in -kor and -chi, and the numerals for prices in bazaars are all Persian in origin, deposited during the long millennium when Persian was the region's high-culture language^[16]. The Arabic layer arrived on top with Islam in the 8th–10th centuries: religious vocabulary (masjid 'mosque', ilm 'knowledge', kitob 'book'), abstract intellectual terms, the days of the week, and much of the vocabulary of law and administration are Arabic.

The Russian layer is more recent and thinner. Roughly 3–5% of everyday Uzbek is Russian-derived, concentrated in Soviet-era technology and institutional vocabulary (avtobus, poyezd 'train', velosiped 'bicycle', teatr, univermag 'department store'). Modern Uzbek is actively — and often successfully — replacing some of these with older Turkic or Persian equivalents, though many stick. And the twenty-first century has added a small English layer through IT and business (kompyuter, internet, marketing).

Uzbek has been officially written in four alphabets in a hundred years, each transition driven by politics as much as linguistics.

For a millennium before the twentieth century, Uzbek (as Chagatai Turkic) was written in a variant of the Perso-Arabic script. This was the alphabet of Alisher Navoi, Babur, and the Bukharan and Khivan chanceries; most Uzbek pre-modern manuscripts — the al-Bukhari Institute of Oriental Studies alone holds around 25,000 volumes — are in this script.

In 1929 the early Soviet leadership pushed a modernising Latin alphabet across the Turkic USSR — a plan called Latinisation, part of a wider secularising and modernising programme that (not by coincidence) drove a wedge between Turkic peoples and the wider Arabic-Islamic world^[8]. Uzbek was written in Yañalif — 'new alphabet' — from 1929 to 1940. Then, in a sudden reversal, Stalin abruptly imposed Cyrillic across the Turkic republics in 1940 — one of many measures that would deepen Soviet integration.

Cyrillic was Uzbek's official alphabet from 1940 until 2023. A new Latin alphabet was declared official on 2 September 1993, revised in 1995 (introducing the distinctive apostrophe-letter system: o', g', sh, ch, ng), then transitioned over three decades before becoming mandatory for all state functions on 1 January 2023^[8][17].

Below the modern Latin surface, Uzbekistan is one of the world's densest palimpsests of dead writing. The Achaemenid Persian empire (c. 550 – 330 BCE) administered its Bactrian and Sogdian satrapies in Aramaic — a Semitic script that would go on to seed almost every writing system used in Central Asia. The table below traces the full inventory.

Of all the scripts on that list, one has an afterlife out of all proportion to its origins. The Sogdian script — used to write the Iranian language of the Sogdian merchants who ran the eastern Silk Road out of Samarkand and Panjikent between roughly 100 and 1000 CE — is the direct ancestor of the Old Uyghur, Mongolian, and Manchu scripts, which is to say that a script devised in Sogdiana is the same script that Genghis Khan's chancery used to write Mongol in 1206, and the same script the Manchu Qing dynasty used to administer China until 1912^[5][15].

The chain is startlingly precise. Sogdian merchants adopted the Aramaic script through the Persian imperial administration around the beginning of the Common Era, adapting it to write Iranian Sogdian. In the 8th century the Uyghur Kaganate (in what is now Mongolia) adopted the Sogdian script from Manichaean and Buddhist Sogdian missionaries; over the following centuries, Uyghur scribes rotated the writing direction 90 degrees so that lines ran top-to-bottom, a choice apparently made under Chinese influence. In 1204, Genghis Khan captured a Uyghur scribe named Tata-tonga and had him teach the Uyghur script to Mongol scribes. That script — the same one Sogdian merchants had used a millennium earlier — became the writing system of the Mongol empire^[15]. In 1599, the founder of the Manchu dynasty commissioned a further modification of the Mongol script for Manchu. The result is that when you look at a Manchu seal on a Qing-dynasty imperial edict, you are looking at a distant descendant of the alphabet a Sogdian trader used in Samarkand two thousand years earlier.

The Sogdian language itself has one small living descendant: Yaghnobi, spoken by around 12,000 people in the Yaghnob Valley of Tajikistan^[13]. Yaghnobi is the last direct linguistic heir of the Sogdian world.

Before Turkic settlement in the medieval period, Central Asia spoke Iranian languages — an entire family of related tongues that has since been almost entirely displaced. And into that old Iranian world, Buddhism, Islam, Manichaeism, Nestorian Christianity, and much later Judaism each added distinct linguistic layers. Almost all of these languages are now either extinct or seriously endangered.

The two closest large languages to Uzbek are Uyghur (its Karluk-branch sister, spoken in China) and Turkish (its Oghuz-branch cousin, spoken across the Bosphorus). Mutual intelligibility is variable but real. An educated Uzbek reading Turkish news headlines gets roughly 40–50% of the meaning; understanding spoken Istanbul Turkish is much harder because Turkish has undergone heavy vocabulary purification since 1928 and has different vowel-harmony rules. With Uyghur the story flips — an Uzbek can read modern Uyghur (still written in Perso-Arabic script) with training, and spoken understanding is much easier because both are Karluk.

Independence in 1991 forced Uzbekistan's language question open. Cyrillic had been Uzbek's alphabet for 50 years by then, but the young state wanted a symbolic reset. The reset took three decades.

On the street: primary Latin script on new government signs, road signs, national newspapers, currency, museum labels, and airports. Older commercial signage, many older restaurants, many older newspapers, and most Russian-language contexts still use Cyrillic. Bookshops carry both. Younger people (school-age since roughly 2010) read Latin fluently; older adults often read Cyrillic more comfortably and switch to Cyrillic when writing by hand.

The Uzbek Latin alphabet has 29 letters plus the apostrophe. The five letters most likely to confuse a visitor are: O'/o' (like 'aw' in 'saw', a rounded back vowel), G'/g' (a soft French uvular r), CH/ch (as in English 'church'), SH/sh (as in English 'shine'), and NG/ng (as in English 'sing', usually only at the end of words). The apostrophe (') is a full letter, not a punctuation mark — it changes the meaning: o (as in 'hot') and o' (a longer, rounder 'aw') are distinct sounds.

Uzbek uses both Q and K, and both X and H — the pairs are not interchangeable. Q is a deep uvular sound (as in Arabic qalb); K is a normal K. X is a Scottish-loch-style back fricative (as in Arabic khaf); H is a light English H. Words like Qishloq ('village'), Xalq ('people'), and Xush kelibsiz ('welcome') all use these distinctive letters — so once you learn them, you can pronounce almost any Uzbek word in Latin script correctly on sight.

A little Uzbek is disproportionately warmly received. The below use the current 2023 Latin standard.