SOGDIANA i. The Name



i. The Name

The name of the ancient and medieval land around Samarqand, which is usually rendered in European languages as Sogdiana, Sogd, or Sogdia (sometimes written with gh, as commonly in the German adjective soghdisch), has been attested in a series of mutually related forms in many languages of Antiquity and the Middle Ages, including Sogdian itself.  These attestations have been collected recently by Xavier Tremblay (p. 133); for Chinese names, see details by Kurakichi Shiratori (esp., p. 143) and Kazuo Enoki.  A short outline may be presented as follows:

  1. The forms similar to  *suγd- / *suγδ­ are found in Avestan (Suγda), Old Persian (Sugda, Suguda), Akkadian (Su-ug-da), Elamite (Šu-ug-da), Greek (Sughdianē), Latin, Syriac, Parthian, Old Turkic (sγdʾq, soγdāq, see Kāšḡari, s.v. sγdʾq, soγdāq), Bactrian (family-name Sogdokano, see Sims-Williams, 2000, ag2-3, which can be either a transcription of a Sogdian name or a local Bactrian development; Sogdianagoon a Bactrian inscription from Afrāsiāb, the site of medieval Samarqand in the heart of Sogdiana; see Livshits, 2006, p. 71; see below, 4, for a different form), Pre-Tang Chinese *Su-yi (粟弋; Old Chinese pronunciation suk-dək), later Sute (粟特, Early Middle Chinese suawk-dək), Tibetan sog-dag (Kumamoto, p. 296), Tumshuquese Saka suḏana (cf. Henning, 1936,p. 13, with the remark of Konow in n. 3), Arabic (Ṣoḡd, Soḡd), Persian (Soḡd, never with the final “ḏ” in Persian or Arabic) and Sogdian itself.
  2. The form *s(u)γuδ- with an added vowel in the second syllable or*sγu­ with metathesis (cf. Gershevitch, sec. 421) is found in Old Persian, Sogdian, and possibly in Chinese Sage, Suoge, Xue-ge(薩索薛 葛; Tang-time pronunciation *sat-kat, sak-kat and similar; see Pulleyblank, pp. 343-44; Kiyohiro, pp. 59-60; these forms, however, do not agree with the vocalism of the original and, moreover, are applied not to Sogdians of Sogdiana, but to a Turkic tribe with supposed Central Asian origin, so their relation to the name Sogd is open to question).
  3. *subd- *suβd- *sovd- *sūd- was in use in Middle Persian, from which terms were also borrowed into medieval Greek, Syriac, Armenian (while **swt’yk in Sogdian is a ghost-word; see Yoshida, 2003, p. 37 n. 4).
  4. *Suγl- is found in the Chinese form Suli (速利, səwk-lih, Tang times), in the Manichean Middle Persian rendering swγlyy, and in the Bactrian place-name Bonosogoligo, if it means “land (bono)of a Sogdian.”
  5. *Sūl- is attested in Sanskit (Puranic śūlikā, see Gauthiot, p. 542, de la Vaissière, pp. 74-75, n. 15, 18), Niya Prakrit, Khotanese (sūl(i)ya-, probably corresponding to Tang Chinese Hu “western barbarian, [Sogdian] trader”; cf. Emmerick, p. 149, Kumamoto, pp. 191-19, 291-29) and Xuanzang’s rendering Suli (窣利, Early Middle Chinese Swət-lih).
  6. The Avestan form suxδəm (Yt 10. 14; AirWb., col. 1582) with the voiceless “x” standing apart.  Old Iranian forms witness that it was a thematic a-stem, while in Sogdian only suffixed forms are attested, which do not permit any decision of whether the stem was light or heavy, but from a formal point of view one would expect the former; moreover, the adjective sγwδyʾnʾk has the suffix –yʾnʾk, which is attested otherwise only attached to ʾyk- “?” and kš- “Kesh” (Sims-Williams, 1992, p. 44), the first stem is probably and the second one definitely light.

Several proposals have been made to etymologize the name Sogdiana, but none is accepted universally. As early as in 1877, Wilhelm Tomaschek (pp. 74 ff.) proposed deriving  it from Old Iranian *suxta- “burned, pure, holy” (as in Ossetic syǧdæg), and this etymology was repeated in several later major reference works. However, as soon as Sogdian was deciphered, it became clear that Old Iranian *xt normally developed into a cluster which was always spelt with the letters <γt> in the Sogdian national alphabet, and never <γδ­>, while we always observe δin the form swγδ-/ sγwδ- (see Szemerényi, pp. 34 ff.; Livshits, 2003, p. 80 n. 15; Tremblay, p. 134).  Oswald Szemerényi’s suggestion (pp. 39-40) that *sugda- is a syncopated form of the earlier *sukuda-, which is, in its turn, an anaptyctic variant of *skuda- “archer, Scyth,” meets a serious hindrance in the fact that the Old Iranian cluster *sk remains sk in Sogdian (Tremblay, p. 134).

O. I. Smirnova (1963, pp. 24-26, following a remark by Vjatkin; cf. also Khromov) proposed linking the land-name with the Persian (and Tajiki) soḡd meaning “low-lying ground where rain-water collects” and “low-lying swampy place; tilled locality rich in flowing water” (so in Tajiki, where it is pronounced suḡud).  This word occurs in Persian dictionaries, modern North-Tajiki dialects, and a few medieval texts from Transoxiana.  Although from the geographical viewpoint, the middle and lower part of the Zarafšān valley perfectly falls within this denomination, the absence of any comparanda, etymology and early attestations of Persian/Tajiki soḡd/soḡod renders Smirnova’s suggestion questionable (rather, the opposite, the land-name Soḡd with its abundant waters is responsible for the Persian common noun? cf. Jayḥun “the River Oxus,” used also in the sense of “large river” in several classical Persian texts; see Lurje, 2006, pp. 414-15).  Soḡoti / soḡdi, the name applied today to the agriculturalist population of the lower parts of the Hissar and Kaška Daryā valleys by their neighbors (see Oransky, p. 151; Karmysheva, pp. 56-59), probably originates in Tajiki soḡod.

Mention should also be made of Xavier Tremblay’s suggestion (esp. pp. 134-35) that *suγδa- represents Old Iranian *suxθa-, a parallel form for *suxta- “burnt.”  The development of *xθ into γδ is regular in Avestan, but since there is no other example known so far of how *xθ was realized in Sogdian, Tremblay’s etymology remains hypothetical as well.

Bibliography: see Sogdiana ii.

(Pavel Lurje)

Originally Published: August 10, 2017

Last Updated: August 10, 2017

Cite this entry:

Pavel Lurje, “SOGDIANA i. The Name SOGD,” Encyclopædia Iranica, online edition, 2017, available at (accessed on 10 August 2017).