SCYTHIAN LANGUAGE. The language of the Scythians, which is not attested by authentic continuous texts, but only in a quite fragmentary manner indirectly and chiefly by a variety of names, is one of the idioms spoken by the nomadic tribes of the Eurasian steppelands along the northern edge of the home of the Iranian peoples in the Old Iranian period. According to the traditional division it is part of the group of the North-Eastern Iranian languages because of the obvious connection with Ossetic. Written evidence of the Scythians and their language is recorded only by some neighboring peoples, the Assyrians, Persians, and Greeks, who had come into contact with them. From the cuneiform sources (both Neo-Assyrian [NAssyr.] and Old Persian) we know only some isolated names of persons or tribes. More information is provided by Greek sources and in particular, even for the language, by the historiographer Herodotus. One could have assumed that the Scythian policeman whom Aristophanes (ca. 450 BCE-388 BCE) puts on stage in his Thesmophoriazousai, would speak Scythian, but in reality he talks only some kind of comedians’ hotchpotch.

Since in Herodotus several Scythian tribes are mentioned and likewise, taking into account that “the Persians call all the Scythians Sakas” (as already Herodotus [7.64.2] did know), in the Old Persian royal inscriptions various tribes of Sakas are distinguished, we have to reckon also with different Scythian (Saka) dialects, which we cannot make out, however, in detail. Among those Scythian or Saka tribes is one, whose name is attested in both traditions, that of the (OPers.) Sakā Haumavargā, who correspond to the Skýthai Amýrgioi (Herodotus 7.64.2). But we have to consider that when Herodotus or any other Greco-Roman author talks about “Scythians,” there are meant the tribes living to the north of the Black Sea (from the lower Danube up to the Sea of Azov and the lower Don), for the Greeks of pre-Hellenistic times had practically not the slightest information on the countries and peoples further to the east apart from what they heard in the North-Pontic Greek cities. Nevertheless Herodotus gives some hints about different Scythian dialects, even if rather vague hints, for example about linguistic peculiarities (solecisms) of the Scythian-speaking Sauromátai (4.117). And the Gelōnoí, who are neighbors of the Scythians, are said to speak a partly Scythian, partly Greek, language (4.108.2). This reminds one of the ancient Greeks’ and Romans’ favorite interpretations as a hybrid or mixed language. But it is left open whether it was Greek with a strong Scythian influence or inversely a form of Scythian with many Graecisms.

The significance of Herodotus is the result of the fact that he is the only author (whose work is preserved) who wrote at a time clearly contemporaneous with the linguistic period of Old Iranian, whereas for all possible later literary or epigraphic evidence of non-Greek words or names attested for those regions the assignment to Old or Middle Iranian (i.e., to Scythian or Sarmatian [q.v.]) is possible only on the basis of a linguistic analysis and of unambiguous criteria, since material of both kinds may occur side by side. That interconnection can be illustrated by the following example: In 220 CE the anthroponym Baióraspos (i.e., OIr. *Baivar-aspa-, MPers. Bēwarasp “possessing myriads of horses”; see AŽDAHĀ i.) is attested in Tanais at the Don estuary, and in a clearly Old Iranian form by its phonetic development; but only a few years later (228 CE) we find there the name Phoúrtas with (as the metathesis of *θr to rt shows) a clearly Middle Iranian form comparable to Alanic phourt, Oss. furt/fyrt “son” in contrast to OIr. *puθra- (Av. puθra-, OPers. puça-). Those later successors of the Scythian language are attested in the idioms of the numerous Sarmatian, Alanic, and related tribes to the north of the Black Sea, the Caucasus, and the Caspian Sea. They are all known primarily from personal names in Greek inscriptions of Roman times and from words and names found in literary sources. The only direct evidence is the Alanic inscription (in Greek characters) from the banks of the river Zelenchuk (ca. 10th–12th centuries), by which the linguistic development leads to the modern Ossetic language.

Even for chronological reasons all these sources should not be mingled with the Herodotian material, which alone dates from the Old Iranian period. Besides, Herodotus used older literary sources like Aristeas of Proconnesus or Hecataeus and the reports of several informants whom he had met on his travels to the North-Pontic Greek cities (probably mainly Olbia) and who told him myths and legends of the oral tradition which he then quotes repeatedly. Here the reader gains also an insight into the pictorial language of that oral poetry, e.g., with the description of a snow flurry (4.7.3), “when earth and air are full of feathers (and thus intercept visibility)”.

From the use of OPers. Saka- in the Achaemenid royal inscriptions and the Herodotian passage quoted above it is clear that Gk. Skýthai is an alloethnonym and not the self-designation of this people. The same ethnic name is attested also in NAssyr. Aš-ku-za-/Iš-ku-za-, i.e., original /Škuza-/ without the prothetic vowel, as well as in Hebrew ʾšknz, which evidently has to be emended into ʾškwz (as one of Gōmer’s sons in Gen. 10.3 and a kingdom like Urartu and that of the Mannaeans in Jer. 51.27 are called). The original source of these forms is disputed, however (for neither *Skuča- nor *Skuθa- nor *Skuδa- “shooter, marksman” [thus Szemerényi, pp. 16–23] are fully convincing), and cannot be reconstructed with certainty, because Sem. z and Gk. th are incompatible (and most likely are based on ). But because the Greek form is attested already about 700 BCE in Hesiod, frag. 150.15 M.–W., one might assume also that the Greeks did learn this Near Eastern name not via Asia Minor, but via the Thracians or some other people living at the Black Sea as intermediaries.

Since the Scythian words and names differ in their structure and development clearly from the later, “Sarmatian” evidence (see above) and alone are on the Old Iranian stage of development, it is methodologically sound and suitable to regard as Scythian only the material attested in Herodotus or even before him. This restriction has been demanded rightly by M. Vasmer (1923, p. 8, etc.), and the most cautious and reliable studies on the Scythian language (e.g., by Mayrhofer or Pinault) follow this principle. Other scholars, however, understand the term “Scythian” in a wider sense, either diachronically including all the evidence up to the 5th century CE (thus Abaev, 1949, pp. 147–244, and, only with a changed title, Abaev, 1979) or diatopically (thus Zgusta, pp. 245–64 distinguishing two different dialects in the personal names of the Pontic-Greek colonies), and these two approaches create further confusion.

The names attested in cuneiform inscriptions are (1) NAssyr. Iš-pa-ka-a-a (the leader of the Scythian troops defeated about 675 BCE by king Esarhaddon) = OIr. (“Scyth.”) *Spak-aya-, a hypocoristic based on *spaka- “dog” (see Schmitt, 2009, pp. 93f.); (2) NAssyr. Bar-ta-tu-a (the king of the Scythians at roughly the same time, who asked Esarhaddon for the hand of one of his daughters) = Gk. Protothýēs (in Herodotus 1.103.3 the father of the Scythian king Madýēs, who fought with the Assyrians against the Medes), perhaps from OIr. *Pr̥θu-tavah- “with far-reaching powers” (see Schmitt, 2009, pp. 64–66); (3) OPers. Skunxa- (the leader of the Sakas, who rebelled against Darius I), perhaps related to Oss. skoyxyn/æsk’wænxun “to distinguish oneself”.

It is a matter of course that the indirect evidence (and onomastic as well) causes much uncertainty, and therefore it is not adequate to characterize the Scythian language very precisely. But several phonological and morphological features characteristic of Scythian can be made out: The compositional vowel has been preserved (cf. Aria-peíthēs, Sparga-peíthēs, etc.); the old diphthongs are not yet changed to long vowels; OIr. *sp (from Indo-Ir. *ćṷ) is preserved as in most of the other Iranian languages (but not Old Persian and Khotanese Saka).

We should not forget, however, that Herodotus was neither a philologist nor a linguist. This becomes quite clear, where he mentions etymological interpretations of names, as they occur sometimes in his work: The name of the one-eyed Arimaspoí, whom “we (the Greeks) call in Scythian (!) Arimaspoús” (as Herodotus 4.27 has it, perhaps based on Aristeas), is said to contain árima “one” and spoû “eye”; but this obviously is based on the Greek accusative form, and the two alleged Iranian words are pure invention, so that a connection with OIr. *aspa- “horse” may be preferred, perhaps a compound *Aram-aspa- “doté de chevaux adaptés”, as proposed by Pinault, pp. 125–36. And the Amazons are called Oiórpata “man-killing” by Herodotus (4.110.1) with oiór “man” and patá “to kill”; at best, however, only oior (if from *oiro) could be derived from Ir. *vīra- “man”, but even this would be in contrast to Herodotus’ normal rendering of Ir.*vi- by Gk. hy-, so that doubts on this interpretation, which only provides the aition for the following story, are allowed (cf. a more recent attempt, based though on a conjecture, in Hinge, pp. 95f.).

Quite problematic are also the Scythian theonyms listed by Herodotus 4.59.2 (see Humbach and Faiss, pp. 4-8). The counterparts of Zeus, Apollon, Aphrodite, and Poseidon are rather unclear, but easier to understand are Apí for the “Earth” () and Tabití for the hearth-goddess Hestía: Tabití can be derived from the root Ir. *tap “to burn, heat”, probably as a participle (fem.) *tapatī (originally “heating” as an epithet of the fire); and for Apí “Earth”, that reminds one of Ir. *āp- “water”, it has been suggested that some confusion happened, most likely in an enumeration of fire–heaven–water–earth or similar, as we find it sometimes in the Avesta.

None of the names of the great North-Pontic rivers, that are first mentioned by Herodotus, can be analyzed precisely (e.g., Borysthénēs = Dnieper, “Scyth.” Pórata = “Gk.” Pyretós = Pruth, Tánaïs = Don, Týrēs = Dniester, Hýpanis = Bug), only Ir. *kapa- “fish” (Khot. kavā-, Oss. kæf, etc.) in Pantikápēs (a branch of the Dnieper, 4.18.2 etc.) is fairly clear. Likewise things are with the names of tribes and peoples, including the Sakā and Skýthai (see above) themselves. Only for the name of the Massagétai (q.v.) to the east of the Caspian Sea the analysis as the plural form (with suffix *-, Gk. -tai) of Ir. *Masi̯a-ka-, a derivation from *masi̯a- (YAv. masiia-) “fish”, seems to be certain, especially since Herodotus 1.216.3 or his source expressly states that they lived on their livestock and fishes. As to the so-called Argippaîoi at the edge of the earth, whom Herodotus (4.23.5) knew only from hearsay, it must be remarked, that the original form of the name was Orgimpaîoi and that Argippaîoi is nothing else than a Greek scholarly reinterpretation as “with swift (or: sparkling) horses” (see Schmitt, 2006). On the other hand the Orgim-paîoi may be related with the (OPers.) Sakā Hauma-vargā, as Pinault (pp. 121–23) plausibly has proposed.

The Scythian words and names of Herodotus and a selection of the younger material is collected (with chronological distinction, of course) and succinctly discussed by Mayrhofer as well as by Humbach and Faiss, pp. 8–18; for the personal names see also Schmitt, 2003 (and Schmitt, 2011a, s.vv.), for the common nouns also Brust (s.vv.), and in general the studies listed in the bibliography. As a result we may summarize the essential lexemes (see Mayrhofer, pp. 24f.; Pinault, pp. 108–16): *ari̯a- “Iranian” (in Aria-peíthēs); *aspa- “horse” (see above ad Arimaspoí; cf. Oss. jæfs, but Khot. aśśa-, Wakhi yaš); *gaiθā- “herd, possessions” and *sūra- “strong, mighty” (in the theonym Goitó-syros); *kapa- “fish” (see above); *madu- “intoxicating drink” (in Madýēs); *masi̯a- “fish” (see above); *paisah- “ornament, decoration” (in Aria-, Sparga-peíthēs, slightly Grecized); *para-dāta- “placed at the front” (in Paralátai, the name of a royal house; = YAv. paraδāta-); *pr̥θu- “far-reaching” and *tavah- “power, strength” (see above ad Bartatua/Protothýēs); *sparga- “scion, descendant” (in Sparga-peíthēs); *tap “to burn, heat” (see above ad Tabití); *vari- “cuirass” (in Órikos from *Vari-ka-); and *xšaya- “ruling over” (in Arpó-, Lipó-, and Kolá-xaïs). These last-mentioned names of three brothers, the sons of the first Scythian king Targítaos, listed in Herodotus 4.5.2 “Lipóxaïs, Arpóxaïs and Koláxaïs as the youngest,” still show an inherited stylistic device, the so-called “augmented triad,” i.e., an epithet only with the last element of an enumeration, which Herodotus ought to have owed to the Scythian oral tradition. This goes well with the fact that naming brothers with partly identical names (here *-xšaya-) reflects also an old custom inherited from Indo-European times. But we have to add the reservation that many of the personal names are not clear in every detail and that in several cases their form is Greek at all (e.g., Lýkos).

All in all, it can be summarized that the Herodotian material shows hardly any striking dialectal feature, but mostly is in accordance with the common Old Iranian stage of development (see Schmitt, 2003, p. 6; Mayrhofer, p. 26). Only by taking the post-Herodotian evidence into account, too, in an unacceptable manner (see above) one gets different and often rather speculative results (cf., e.g., Vitchak). Here should be added also, that the hypothesis of Lubotsky, according to which Scythian borrowings could be proved in other Old Iranian languages, too (e.g., in OPers. *farnah- “abundance”, duvarθi- “gate”, and *θigra- “garlic”), met with general disapproval.

An interesting special case of a different nature is the name of a typically Scythian food prepared of mare’s milk, i.e., some cheese named hippákē (first attested already in Aeschylus, frag. 198 R.). At first glance this word seems to be purely Greek, but it has been explained convincingly as the loan-translation of Scyth. *aspaka-, originally the epithet of some feminine noun (see Forssman).

Finally it should be stressed that the name of the Black Sea (for more details), Gk. Áxeinos Póntos, though often regarded as of Scythian origin, has nothing to do with that language. Gk. Áxeinos “inhospitable” (first in Pindar, with whom we find also the euphemistically remodeled Eúxeinos Póntos for the first time) reflects OIr.*axšaina- “of dark color, black”, which as part of a system, in which the cardinal points symbolically are described by color terms, also means “northern”. But if the Black Sea is the northern one, then the name must have been given by people living to the south of that sea. And since such names are not found with the peoples of the Ancient Near East (Babylonians, Assyrians, Elamites, etc.), their origin obviously must be in the Achaemenid Empire that extended in the south as far as the Persian Gulf (Gk. Erythrḕ Thálassa “Red Sea”) and in the north as far as the Black Sea (see explicitly Herodotus 4.37).



H. Humbach and K. Faiss, Herodotus’s Scythians and Ptolemy’s Central Asia. Semasiological and Onomasiological Studies, Wiesbaden, 2012.

M. Mayrhofer, Einiges zu den Skythen, ihrer Sprache, ihrem Nachleben, Vienna, 2006.

G.-J. Pinault, “La langue des Scythes et le nom des Arimaspes,” Académie des Inscriptions et Belles-Lettres. Comptes rendus 2008, pp. 105–36.

R. Schmitt, “Andere altiranische Dialekte,” in R. Schmitt, ed., Compendium Linguarum Iranicarum, Wiesbaden, 1989, pp. 92 f.

Idem, “Die skythischen Personennamen bei Herodot,” Annali dell’Università degli Studi di Napoli “L’Orientale” 63, 2003, pp. 1–31.

Idem, “Die Sprache(n) von Skythen, Sarmaten und Alanen. Bemerkungen zur nordiranischen Sprachgeschichte,” Philologia Fenno-Ugrica 16–17, 2010–11 [2012], pp. 1–27.

M. Vasmer, Untersuchungen über die ältesten Wohnsitze der Slaven. I: Die Iranier in Südrußland, Leipzig, 1923 (repr. in Vasmer, 1971, pp. 106–170).

Idem, “Skythen. B. Sprache,” in Reallexikon der Vorgeschichte XII, Berlin, 1928, pp. 236–51 (repr. in Vasmer, 1971, pp. 179–99).

Idem, Schriften zur slavischen Altertumskunde und Namenkunde I, Berlin, 1971.

Other works cited.

V. I. Abaev, Osetinskiĭ yazyk i fol´klor I (Ossetic language and folklore I), Moscow and Leningrad, 1949.

Idem, “Skifo-sarmatskie narechija,” (Scythian-Sarmatian dialects) in Osnovy iranskogo yazykoznaniya: Drevneiranskie yazyki (Fundamentals of Iranian linguistics: Old Iranian languages), Moscow, 1979, pp. 272–364.

M. Brust, Die indischen und iranischen Lehnwörter im Griechischen, 2nd ed., Innsbruck, 2008.

B. Forssman, “hippákē, ein verkapptes skythisches Lehnwort im Griechischen,” ZVS 79, 1965, pp. 285–90.

G. Hinge, “Herodot zur skythischen Sprache. Arimaspen, Amazonen und die Entdeckung des Schwarzen Meeres,” Glotta 81, 2005 [2006], pp. 86–115.

A. Lubotsky, “Scythian Elements in Old Iranian,” in N. Sims-Williams, ed., Indo-Iranian Languages and Peoples, Oxford, 2002, pp. 189–202.

R. Schmitt, “Griechische Umdeutung eines ‘skythischen’ Ethnonyms,” Historische Sprachforschung 119, 2006, pp. 186–89.

Idem, Iranisches Personennamenbuch VII/1A: Iranische Personennamen in der neuassyrischen Nebenüberlieferung, Vienna, 2009.

Idem, Iranisches Personennamenbuch V/5A: Iranische Personennamen in der griechischen Literatur vor Alexander d. Gr., Vienna, 2011a.

Idem, “Herodot und iranische Sprachen,” in R. Rollinger et al., eds., Herodot und das Persische Weltreich/Herodotus and the Persian Empire, Wiesbaden, 2011b, pp. 313–41.

O. Szemerényi, Four Old Iranian Ethnic Names: Scythian – Skudra – Sogdian – Saka, Vienna, 1980.

K. T. Vitchak, “Skifskiĭ yazyk: Opyt opisaniya,” (Scythian language: An attempt at a description), Voprosy Yazykoznaniya 1992/5, pp. 50–59.

L. Zgusta, Die Personennamen griechischer Städte der nördlichen Schwarzmeerküste: Die ethnischen Verhältnisse, namentlich das Verhältnis der Skythen und Sarmaten im Lichte der Namenforschung, Prague, 1955.

(Rüdiger Schmitt)

Originally Published: April 10, 2018

Last Updated: April 10, 2018

Cite this entry:

Rüdiger Schmitt, “SCYTHIAN LANGUAGE,” Encyclopædia Iranica, online edition, 2018, available at (accessed on 10 April 2018).