Introduction to Uyghur — slides from Polyglot Conference Fukuoka 2019, with commentary
Last month, I attended the 2019 edition of the Polyglot Conference™ in Fukuoka, Japan, which was reportedly attended by as many as 400 people this time around. I also gave a presentation at the conference myself — an introduction to the Uyghur language — and the slides for that can be found here.
Since the slides themselves are a bit short on explanations — I threw most of them together in one evening a week before the event, and filled in the gaps during my live talk — I decided I should add in some of that missing content in a blog post (i.e. this post) for easier reading. A video of my talk should also be uploaded to the Conference’s Youtube channel some time in the coming year.
Scroll to the bottom for some additional thoughts I had on this year’s conference — and next year’s too!
Part 1: Uyghur Scripts
Let’s start with one of the more remarkable aspects of the Uyghur language, historically speaking: the variety of different scripts the language has been written in over the years. This will also serve as a quick overview of the history of the language.
Here we have an example of the oldest script to ever be referred to as “Uyghur” — the Old Uyghur alphabet, a vertical script used in what is now Chinese Central Asia between the 8th and 17th centuries. This script was apparently adapted from Aramaic and came to this part of the world from the Near East with the transmission of religions like Manichaeism and Nestorianism, and later became the basis for Mongolian and Manchu scripts as well.
The specific example in this slide was collected by Finnish (and imperial Russian, and ethnically Swedish) Baron Carl Gustaf Emil Mannerheim during his expedition through Xinjiang after the Russo-Japanese War, and is basically a financial document. I don’t read this script yet (I might eventually, if I ever get serious about learning Mongolian), but in the Romanization on the right you can see the interesting word “Kitai”, which is the modern Russian word for China and has a Turkic origin.
Now, the mainstream (i.e. Wikipedia) academic view seems to be that Modern Uyghur, the subject of this talk, is not a direct descendant of Old Uyghur. Old Uyghur also happens to be one of the oldest attested Turkic languages and a close relative of the“Old Turkic” that was once used in Mongolia and elsewhere. According to the mainstream view, Old Uyghur evolved into what is now the Western Yugur language of Gansu, China instead.
On the other hand, one linguist I consulted with had a different view on this issue, so I think it’s worth including here:
“Uyghur linguists regard the Old Uyghur as the direct ancestor of Modern Uyghur,” Mustafa Aksu of the University of Indiana told me. “There are many words that we are still using these days that you may not be able to find in other Turkic languages. For example, the word omaq which means cute. In Altun Yaruq(金光明经), and Irq Bitig [Old Uyghur texts], there are also many lexical items which currently available in modern Uyghur. Phonetically, all the phonemes in Old Uyghur are still available in Modern Uyghur.”
Jumping forward a few centuries, and adding a significant amount of Islamic (and therefore Arabic and Persian) influence, we come to a language that is generally considered to be the common ancestor of both the Uyghur and Uzbek languages — Chagatai. This was also the language of the early Mughals who eventually conquered most of India, and the sample above shows the opening of the Baburnama — the autobiography of the the first Mughal emperor Babur, which was written in Chagatai, also known simply as Turki at the time.
I have tried studying a bit of Chagatai in the past, but so far only in Romanized form. Original Chagatai texts are often written in a dense calligraphic style of Arabic script which I still have a hard time deciphering. The font used in the slide above — from a version printed in Russia in 1857 — makes it easier for me to identify the individual letters, but the elaborate grammatical style still makes it a bit hard to understand.
Skipping forward a bit further, here we have some examples of Turki as documented by British linguists in the 19th century. This is a lot closer to modern Uyghur, but still has a few features that are less common nowadays like the -ghay future suffix (in bolmighaylar, “they will not obtain”, fourth word from the end).
People who know Arabic script will also notice that Arabic/Persian loans in this sample are still written as they are in the source language, whereas in modern Uyghur words are written phonetically. For example عرض (ʿarḍ, “representation”) in line 3 of this text would be written as ئەرز (erz) today.
Next up, this is where things get weird — Uyghur in Cyrillic script! This was used in China briefly in the early 1950s (before the Sino-Soviet split, basically) and remains in use somewhat among Uyghurs who live in Kazakhstan and Kyrgyzstan.
The sample in the slide above is from a book on the modern Uyghur language published in Alma-Ata in the Kazakh SSR (now Almaty, Kazakhstan). The specific paragraph is about Chinese loanwords in Uyghur (“Uyghur tilidiki xensu tilidin özleshken sözler”), using the politically-correct term for ethnic Han Chinese, xensu (from 汉族 Hànzú) rather than the traditional term Xitay.
(On a related note, the official name for the People’s Republic of China in Uyghur is Jungxua Xelq Jumhuriyiti or Junggo for short, where Jungxua and Junggo come from Mandarin 中华 and 中国 respectively. Those words are officially used in Kazakh and Kyrgyz in China, as well.)
Next up, here we have an example written in the modern Uyghur alphabet, which — unlike Arabic script as it is more commonly used — really is an alphabet and not an abjad, because every vowel has its own letter. That makes it a bit like the writing system for modern Sorani Kurdish.
The sample above comes from Uyghur Texts in Context: Life in Shinjang Documented from Public Spaces by Frederick de Jong, a really interesting book with samples of modern Uyghur texts with English translations. This specific sample is just instructions for how to install language software on your computer, but most texts in the book are a bit more political and reflective the current atmosphere in Xinjiang.
To make things easier for audience members who don’t read Arabic, all remaining examples of Uyghur in this presentation are provided in the Uyghur Latin alphabet, whose creators meant for it to be used “solely in computer-related fields as an ancillary writing system” and not as a replacement of the Arabic script. It’s generally pretty intuitive, and you can check the link above for more details on the IPA values of the letters and how they correspond with letters in other Uyghur scripts.
Part 2: Sounds
Next, let’s take a look at how the phonology of Uyghur compares to its relatives in the Turkic family. For the most part, cognates in different Turkic language tend to be quite close to each other and easy to recognize (more so than, say, cognates in French and Spanish), but sometimes you have words like this which have drifted apart a bit more:
Without knowing the common root of these two words, the connection might be a bit hard to make.
As I’ve touched on in my previous posts about Uyghur and Turkish, one of the big differences between Uyghur and Turkish (which I’m using as the benchmark because a lot more people in the audience know Turkish than any other Turkic language) is that “q” and “gh” sounds have fallen out in a lot of places in Turkish, such as in the “-qan/-ghan/-ken/-gen” past tense suffix or the “-qa/-gha/-ke/-ge” indirect object suffix. Other examples include Uyghur emgek vs. Turkish emek for “labor”, for example.
The above chart shows some of the other prominent differences between Turkish and Uyghur, with cognates from Tatar and Kyrgyz included for additional context. The first two rows show how Turkish — like its Oghuz relatives, Azeri and Turkmen, and unlike most other major Turkic languages — frequently voices the first consonant, turning “t” into “d” and “k” into “g”. The next two rows show how there are some cases of an “r” being dropped in various situations, in both Turkish and Uyghur.
Row 5 of the chart shows how general Turkic initial “b” has often shifted into “v” in Turkish, and Row 6 shows that general Turkic “t” will often become a “ch” sound before front vowels. Finally, the last row shows how Turkish “ev” often corresponds to a diphthong in other Turkic languages, including Uyghur.
The next slide demonstrates two sound changes that are very widespread in Uyghur but otherwise not commonly seen in Turkic languages — vowel reduction and vowel fronting.
With vowel reduction, “a” and “e” sounds become neutralized into a short “i” sound before many suffixes. (Unlike Turkish, which has both a front, dotted “i” and a back, undotted “ı” sound, Uyghur has just one neutral “i”.)
With vowel fronting, “a” and “e” sounds become a more fronted “é” before other endings, including the passive suffix “-il” and the polite imperative “-ing”.
Part 3: Words
Now let’s take a look at loan words in Uyghur, which come from a wide range of sources.
First of all, for historical and cultural reasons, Uyghur is unsurprisingly full of Arabic and Persian loans words. The Arabic words are mainly borrowed via Persian, and reflect the Persian way of pronouncing Arabic words — although Uyghur has added a few innovations of its own.
For example, in the word xewer above, you can see how an original Arabic “b” became “w” between two vowels. In the word jawab, both “a”s have become long even though in Arabic the first one is short. In paaliyet, you have an original Arabic “f” becoming a “p” in Uyghur.
More recently, Uyghur has of course come under lots of Chinese influence as well. But what comes as a bit more of a surprise to many people is that there has been a significant amount of Russian influence, thanks to Russian involvement in Central Asia over the past couple of centuries.
Also worth noting is that Chinese had already had some influence on Uyghur long before modern times. For example, the Uyghur word baxshi (“shaman, sorceror”) comes from Chinese 博士 (bóshì in Mandarin, hakushi or hakase in Japanese, baksa in Korean), which originally meant “court scholar” but now means “doctor, PhD” in Chinese. The meaning has shifted a bit in Uyghur, as you can see. (This word is also related to Mongolian bagsh, which means “teacher”.)
Finally, here’s a quick look at the rich system Uyghur has for deriving words from other words. As the “with loanwords” examples show, this system is just as productive for borrowed vocabulary as it is for native, Turkic words.
Part 4: Sentence construction
Uyghur is a Turkic language, so its sentence structure is broadly similar to other Altaic languages like Mongolian and Manchu, as well as Japanese and Korean (regardless of whether you think these languages are actually related or not).
One thing that does set Turkic languages apart from Mongolian, Manchu, Japanese and Korean is that Turkic verbs are conjugated for person. The chart above show just a small portion of all the finite (conjugated) forms of the verb “to do”, leaving out negation, question marking, and evidentiality, for example.
The first tense shown above covers both general present and future, a lot like the non-past tense in Japanese, for example. The past continuous tense ultimately comes from a construction that means something like “I do and lie down” (“qil-ip yat-i-men” became “qil-i-wat-i-men”). You’ll see a construction like this in Kyrgyz, Kazakh and Uzbek too, whereas in Turkish the present continuous comes from “I do and run” (“ed-ip yür-ür-üm” became “ed-iyor-um”).
“Past I” and “Past II” vaguely correspond with simple past (“I did”) and past perfect (“I have done”), but not really. Past I is also used for things that happened in the immediate past, while Past II gets used more for historical events, for example, but the nuances can get a bit complicated.
Next, something that Uyghur shares with all Altaic languages (and Japanese/Korean) is its use of case endings to show the role of nouns in a sentence. The slide above shows examples of: accusative (direct object) marker “-ni”, which is only used for definite objects; genitive (“of”) marker “-ning”, which is only used for actual possession, mostly; and ablative (“from”) and allative/dative (“to”) markers “-din” and “-gha/-qa/-ge/-ke”.
If you know Turkish, you’ll notice that the corresponding Turkish endings for “-ni”, “-ning” and “-gha” are all missing the initial consonant compared to Uyghur, which is closer to the Old Turkic forms. Uyghur’s “-din” ending is also closer to the Old Turkic form than the “-dan/-den” of Turkish.
(Another significant difference which I didn’t touch on in my live talk was that Uyghur doesn’t insert an “n” between a possessive suffix and case ending the way Turkish does, so while “in his book” would be kitab-ı-n-da in Turkish, it’s just kitab-i-da in Uyghur.)
Uyghur also has a highly developed system for deriving all sorts of parts of speech from verb roots. The slide above shows how you can derive past, present and future tense modifiers from verbs in Uyghur, for example, quite similar to what you can do inKorean (e.g. ha-n, ha-neun, ha-l).
You can also form a variety of noun-like forms from verbs in Uyghur, and a lot of these can serve similar functions to relative clauses in English.
And last of all, here are some examples of adverbial forms derived from verbs.
Part 5: Sample texts
I wrapped up my talk with a few examples from Omniglot.
Addendum: Uyghur Learning Materials
This wasn’t included in my original presentation, but since several people have asked me about this, here are some places to get started with learning Uyghur yourself.
First, there’s Greetings from the Teklimakan: a handbook of Modern Uyghur, an introductory textbook provided for free online (with audio!) by the University of Kansas. This was the first textbook I used myself.
Another textbook that was very useful for me was Spoken Uyghur from University of Washington Press.
Beyond that, translator and writer Gene Bunin has made a large amount of Uyghur language material available on his website, in English, Chinese, Russian, Turkish, German, Japanese, and Uyghur. I found the Uyghur Reader to be particularly valuable for taking my Uyghur to the next level, and it includes texts on a wide range of subjects. If you read Chinese, 维吾尔语语法 (Uyghur Grammar) from Minzu University of China Press is the most in-depth Uyghur grammar I’ve been able to find anywhere, and was the source of many of the examples in my presentation.
Finally, I’ve mentioned this book earlier, but I just wanted to reiterate that Frederick de Jong’s Uyghur Texts in Context: Life in Shinjang Documented from Public Spaces is a great source for more timely Uyghur texts (with translations), once you’re ready for a bit more advanced reading.
All in all, I had a great time at the first Polyglot Conference to be held in Asia. The location helped draw a lot more language enthusiasts from the region that we usually get to see at events like this in North America and Europe, and it was a nice change of focus.
Because presentations were held in three parallel tracks, I only got to see less than half of the talks on offer in person, and my own consumption was heavily biased towards more technical talks about specific languages: Júlio César Pereira da Silva’s talk on Japanese dialects, Jonty Yamisha’s two talks on his attempts to revive the Circassian language, Yulia Savushkina’s talk on Tuvan, Syauqi Ahmad Zulfauzi Stya Lacksana’s talk on Sundanese, and Brian Loo’s talk comparing the situations of Chamorro and Santali.
Beyond those, Marty McNutt’s talk on possible links between Turkish and Japanese languages introduced me to the super-intriguing Transeurasian theory, which I hadn’t heard of before, and Alex Rawlings’s talk on “Personalities and Polyglottery” added some more compelling personal observations to the probably never-ending debate over whether speaking a different language can change your personality.
I also heard great things about many of the talk I didn’t make it to, and look forward to catching up on them when the recordings hit Youtube at some point.
And next year, the Polyglot Conference is headed to Cholula, Mexico! Considering that I’ve never been to Mexico yet (and just one Latin American country overall) despite living in the States for over a decade, I’m absolutely planning on attending, and I’m hoping it’ll give me a bit more motivation to push my Spanish beyond the (upper?) intermediate limbo it’s been in for the past several years.
Since the next Polyglot Conference is likely to have a Western Hemisphere focus, I’m also thinking about breaking out my presentation on Sranan Tongo (which I did in Montreal for LangFest last year) for another run. 🤔
