I’m reading (finally—I bought it 15 years ago!) Janet Abu-Lughod’s Before European Hegemony: The World System A.D. 1250-1350, and she mentions that in the 8th century the Chinese called the Arabs Ta-shih (or Dashi in pinyin). Anybody know the etymology?
While I’m at it, I’ll just complain about the absurd mix of toponyms in this passage:

This is understandable because, according to [Pegolotti’s] itinerary, it will take 25 days by ox-wagon to go from Tana to Astrakhan, another 20 days by camel to reach Organci, another 35-40 days by camel-wagon to reach Otrar, 45 days by pack-ass to Armalec, another 70 days with asses to reach Camexu on the Chinese frontier, 45 more days to the river that leads to Cassai (Kinsai or Hangchow), and then finally 30 days overland to Peking (Khanbalik).

To be consistent, the names should be Tana (Azov), Gintarchan (Astrakhan), Organci (Urgench), Oltrarre (Otrar), Armalec (?Kulja), Camexu (Ganchau), and Garnalec (Khanbalik, modern Peking). (Cassai is OK as is.) What a mess!


  1. DeFrancis’ ABC Chinese-English dictionary has this entry:
    大食 Dàshí n.〈hist.〉Arab empire
    Not going to try to guess at the etymology, but from the Modern Chinese literal meaning (“big food”) maybe the Chinese just caught on early that the Arabs had wicked tasty chow 🙂

  2. Doc Rock says

    大食 is the Chinese attempt to reproduce the name ‘Tazhik’ from Persian.

  3. Doc Rock says

    大食 is the Chinese attempt to reproduce the name ‘Tazhik’ from Persian.

  4. Further on Doc Rock’s explanation, Mandarin pronounciations are not very helpful in understanding etymologies from this period because Mandarin has lost most coda consonants and done various other weird things: its the French of Sinitic. When the term was created, the second character “to eat” was pronounced something like /sik/. The final /k/ is preserved in dialects like Cantonese, as well as in Sino-Japanese shoku and Sino-Korean /sik/. This is probably just a phonological spelling – I wouldn’t read anything into it about cuisine. In fact, except for some of the spices, I don’t think that the Chinese were very taken with Arab or Persian food.

  5. I have the answer somewhere in one of 60 boxes. It would only take me 2-3 hours to find it.

  6. caffeind says

    Googling 大食 gets the Chinese Wikipedia pages for Arab Empire and Abbasid Dynasty. Click autotranslate and you get the translations “Great food States” and “big black clothing was fresh”. You have to read the original to get these equivalences:
    Black Clothes Dàshí: Abbasid Caliphate
    White Clothes Dàshí: Umayyad Caliphate
    Green Clothes Dàshí: Fatimid Caliphate

  7. Central Asian history is, obviously, cluttered with multiple names (problems of transliteration aside). Transoxonia has Latin, Arabic, Persian, Chinese, Turkish, and Mongol names, and sometimes they’re related and sometimes not. The Mongols called N. China Khitad (like Cathay, and like the Russian and Muslim names) and called S. China Namgia or the like, which is the Mongol version of the Chinese word meaning “Southern house”. The word Tabgach (Taugast), from the tribal people the Chinese called Toba who ruled the Northern Wei until 535 or so, was still used almost a millenium later.
    The Chinese call themselves Zhungguo ren, Hanren, or Tangren, but foreign names for China have often been the names of conquest dynasties. “China” does come from the Qin dynasty — I think, if it wasn’t the Jin — but the Chinese do not honor the Qin or identify with them.
    Abu-Lugoud’s great book depended on a multitude of secondary sources, and she’s lucky if she didn’t end up taking two names for the same people as two different peoples somewhere.

  8. John Emerson says

    “Tazhik” used to mean “Arab”, IIRC, but now means “Persian of Central Asia”.
    In Ibn Battuta, the the translation I’ve seen, “Arab” means “Bedouin” and usually “bandit”. Battuta was a Berber but I don’t think that that’s the reason. At that time individuals identified with Islam, with their sect, with their clan, or with their city, but not with “the Arab people”.

  9. I am not sure about “Tajik” meaning “Arab”, although I do recall something of that sort in Mu’jam al-buldan. Now there’s a book I need to finish … :o)
    As for “Arab” = “bedouin”, I second that one. I remember my professor telling us a story of one evening he spent as a guest of the local Bedu somewhere in or around Sinai. At one point, one of his hosts proclaimed proudly “Ihna ‘l-arab, ihna mish fallahhiin!”.

  10. The origin of the term “Tajik” is somewhat unclear. Today, most historians believe that the word “Tajik” – first mentioned by the Turkish historian Mahmoud Al-Kāshgharī – is an old Turkish expression referring to all Persian-speaking peoples of the region who are of Iranian origin. Alternatively some believe that it is a term of Eastern Iranian origin, which may originally have been applied by the inhabitants of Central Asia to the Arab conquerors of the region, and that its etymology is linked to the tribe of Tayy, whilst from the 11th century it came to be applied principally to Iranians.
    In the Turco-Persian culture of the conquerors Timur and Babur, the word “Tajik” referred to the Persian-speaking clerks who were schooled in Arabic.

  11. “Today, the most numerous Iranian-speaking population of Central Asia is the Tajik, who live mainly in the Tajik SSR …. Tajik is a term which probably goes back to a rather old Persian designation of the Arabs, Tazi. Later, this term was used throughout Central Asia in a much broader sense, meaning all people of Muslim faith, regardless of nationality. Thus, it also came to be known as Da-shy (Ta-shih) in China, where it is used in the same vague sense, sometimes meaning Turkistan, at other times the realm of the caliphate.”
    Karl H. Menges, p. 65 of ‘People, Languages and Migrations’ in the book Central Asia, edited by Edward Allworth (Duke, 1995).

  12. “The Tajik, who live mainly in the Tajik SSR”
    The Soviets deliberately drew national boundaries to divide peoples, and the present boundaries are unchanged. There are Tajiks throughout the Turkish republics. Canfield calls the area “Turko-Persia”.

  13. Jimmy Ho says

    During the Tang dynasty, the meaning of Dashi 大食 started switching from what the Chinese perceived as a Persian ethnicity to Arabs. In its description of the “Western territories” 西域, the Xin Tang shu 新唐書 describes the Dashi (men with long nose, dark skin and beard; women with white skin, always covering their face when they go out) as being from “Bosi” 波斯 (Persia).

  14. Jimmy Ho says

    And for what it’s worth, the official name for the Tajik minority (mostly concentrated in Xinjiang) is Tajike 塔吉克 (Wade “T’a-chi-k’e”).

  15. Jimmy Ho says

    As for the “etymology” of the term, most foreign names were transcribed phonetically, without much regard to the characters used. That’s why it is so frequent to see multiple renderings for the same loan word, most notoriously for Buddhist terms borrowed from Sanskrit and Pali (for an other Tang example, see the variations around the word “Uyghur”: Huihe 回紇, Huigu 回鶻, later evolving to Weiwu’er 畏吾兒 and finally the current Weiwu’er 維吾爾; under the Yuan, a person as eminent as Grand Master Phags-pa had his name transcribed in five or six different ways; modern scholars settled on Basiba 八思巴, but the Yuanshi 元史 isn’t so consistent).
    The idea that there should be a unified system that considers both the semantic and phonetic value of the characters is actually quite modern (and illustrated in the most obvious fashion nowadays by the “translations” of foreign brand names). Somehow, the fact that Chinese characters may have been used in a “purely” phonetical way is disturbing to many people, Chinese or not.

  16. Wow—thanks, everyone! The LH readership comes through again. (And by “etymology” I simply meant what turns out to be “from Tajik“; I’m aware that the particular characters used are essentially arbitrary.)

  17. Doc Rock says

    Thanks to Bill Poser for clarifying my all too brief etymology. I’d just add that the final “k” in southern Chinese was a glottally stopped “k,” a glottally stopped “entering tone.” The glottally stopped finals were “k,” “p,” and “t.” In Middle Korean the “t” stops changed to “l” in Chinese loans into Korean (i.e., Sino-Korean words). So Chinese Mandarin 国 guo, becomes Korean 国 국 guk, and Japanese 国 こく koku;while Chinese 十 shi becomes Korean 十 십 sip; and Chinese Mandarin 蜜 mi became 蜜 믿 mit in Old Korean which became 밀 in Middle and Modern Korean and is 蜜 in Japanese.
    But, then again, mitsu is an essay in itself!
    Doc Rock

  18. caffeind says

    Yes, the p, t, k finals are in Sino-Japanese, Sino-Korean, Sino-Vietnamese, Yue, Min, and Hakka, while Wu, Gan (Jiangxi), and Jin (Shanxi) still have an undifferentiated final glottal stop in most dialects. Only Mandarin and Xiang (Hunan) have lost nonnasal final consonants completely.

  19. Doc Rock says

    By the way, the t–>l transition in Middle Korean of the Sino-Korean glottal stop explains the Korean name for Viet-nam, wOlnam 越南,which was wOtnam in Middle Korean. Doc Rock

Speak Your Mind