Studies in the History and Language of the Sarmatians
J. Harmatta


1. From the Cimmerii up to the Sarmatians

The significance of the nomadic Iranian peoples, the Cimmerii, Scythians and Sarmatians, emerges with ever-increasing clarity from the point of view of the evolution of Slav culture and ethnic characteristics. Thus the more recent Soviet historical science, as compared with the older trend which began the history of Russia only with the Varangians, in dealing with the antecedents of the formation of the first Russian state, goes back, at least as far as the Scythian epoch. Accordingly B. D. Grekov emphasizes the importance of Scytho-Sarmatian culture from the aspect of the Eastern Slavs [1] and P. I. Lyashchenko too deals in detail with these two peoples in his economic history of the USSR. [2] An even more far-reaching significance is ascribed to the Sarmatians in connection with the Southern Slavs by G. Vernadsky, according to whose theory Slav and Sarmatian tribes had been living together as early as pre-Christian times in Southern Russia. This gave rise to the later Russians with regard to ethnic character and culture. In his opinion even the name "rus" derives from the name of a Sarmatian tribe. [3] To some extent also the conception of V. V. Mavrodin tallies with this view. [4] Despite the recognition of the historical significance of the Sarmatians, their history nevertheless is obscure on many a point, in fact no unified picture could be formed of it. In the following we wish to throw light on one period of Sarmatian history which has hitherto not been elucidated.

The appearance of the Huns has been generally held responsible to have set in motion the large-scale movement of peoples that has been known by history as the Migration of Peoples. The appearance of the Huns in Europe was without doubt of decisive importance in history, yet it would be a mistake to believe that their entry to Eastern and Central Europe had been an entirely new and isolated phenomenon in the history of those parts of Europe. Over a century ago A. Hansen already saw clearly that the Migration of Peoples had begun a thousand years earlier with the appearance of the Scythians, [5] and recent investigations have convincingly demonstrated that the migration of the Huns was only one episode in the long series of migrations in the course of which the equestrian nomads of the steppes moved from east to west, and that the movement spread for over more than two thousand years. The process

Additional Notes

To p. 7. During the last two decades an admirable research work was done by Soviet archaeologists who elucidated many problems concerning the history and material culture of the Iranian peoples of Eastern Europe. Much has been done as regards the Cimmerian and Scythian Epoch in Hungary too.

1. B. D. Grekov, The Culture of Kiev Rus. Moscow, 1947. pp. 18.

2. P. I. Lyashchenko, Istoriya narodnogo khozyaistva SSSR. Vol. I (1947), 38—40.

3. G. Vernadsky, Ancient Russia. New Haven, 1943. pp. 74, passim. See my remarks on the matter RHC. N. S. V (1947), pp. 230.

4. V. V. Mavrodin, Obrazovanie drevnerusskogo gosudarstva. Leningrad, 1945. 390.

5. A. Hansen, Beiträge zur Geschichte der Völkerwanderung. I. Ost-Europa nach Herodot. Dorpat, 1844.


set in with the appearance of the Cimmerii at a thousand years B. C. The earliest known seat of the Cimmerii was in the Caucasus and on the adjoining steppes lying north to it; subsequently they penetrated further west and entered South Hungary in the course of the 8th century B. C. The archaeological remains relating to these peoples, contain a great number of horse trappings; one find of a reflex-bow, identified recently, clearly points to a warrior people with equestrian bowmen. The Cimmerii were, therefore, the first people who introduced to Europe a nomad type of warfare that employed equestrian bowmen in large numbers. The migration of the Cimmerii swept along with them a number of peoples who belonged to other ethnic groups, but there can be no doubt about it that the ruling classes of the Cimmerii must have spoken an Iranian language judging from the names of their rulers. It is, therefore, highly probable that they had originally come from somewhere in the steppes of Kazakstan which was supposedly the cradle of the Iranian peoples. [6]

A new equestrian nomad people appeared soon in the footsteps of the Cimmerii: the Scythians, who in the course of their westward movement put an end to the the power of the Cimmerii. At the end of the 6th century B. C. the Scythians had already invaded and conquered the South Russian steppes and penetrated further into the western borderlands of the Eurasian steppe belt. Judging from the archaeological remains the remnants of the Cimmerii settled in two separate lots, namely in Transylvania and along the river Tisa.The number of Cimmerii settled in Hungary at that time must, however, have been so small that they soon became merged into the indigenous Daco-Mysian tribes and into the Celtic peoples who had come to Hungary from the west. Like the Cimmerii, the Scythians also spoke an Iranian language, and so their arrival, one incident in the migration of the equestrian nomad peoples, again increased the preponderance of Iranian elements on the East European steppes. [7]

In the course of the migrations taking place in the Eurasian steppe belt, a new Iranian people, the Sarmatians, followed the Scythians to South Russia in the last centuries B. C. According to the current view we can trace their origin and history as far back as the 5th century B. C. [8] It was at this time that the contemporary account of Herodotus reported (IV. 21) that eastwards to the Scythians and beyond the river Tanais (=Don), there settled a people called the Sauromatae. The Sauromatae of Herodotus have generally been thought the same peoples as the Sarmatae

Additional Notes

To p. 8. In spite of the old opinion, wide-spread among archaeologists even to-day (adopted also by me in the 1st edition) we cannot presume the immigration of the Scythians into Hungary. A summary of my recent views about the history and material culture of the Cimmerii and Scythians was published in Ant. Tan. 13 (1966) 107 foll.

6. For the latest stand of the investigations concerning the Cimmerii see J. Harmatta, AÉ 7/8 (1946-48), pp. 107 ff.

7. The outstanding publications on the Scythians are: E. H. Minns, Scythians and Greeks. Cambridge, 1913.; M. Ebert, Südrußland im Altertum. Bonn—Leipzig, 1921.; M. Ebert in RLV MIL, pp. 52 ff.; M. Rostovtzeff, Iranians and Creeks in South Russia. Oxford, 1922.; M. Rostowzew, Skythien und der Bosporus. I. Berlin, 1931. On the Scythian archaeological remains found in Hungary see N. Fettich, Bestand der skythischen Altertümer Ungarns, in Rostowzew's Skythien und der Bosporus, vol. f., pp. 494 ff. Since the latter publication many new Scythian remains from the Tisza region have become known. For these see M. Párducz, Dolgozatok (Studies) 16 (1940), pp 79 ff., AÉ 4 (1943), pp. 50 ff., AÉ 5/6 (1944-45)., pp 62 ff. For literature on the Scythians in Hungary see P. Reinecke in AÉ 17 (1897), pp. 9ff.; V. Pârvan, Getica. O Protoistorie a Daciei. Bucarest, 1926., pp. 6ff.; V. G. Childe, The Danube in Prehistory. Oxford, 1929., pp. 394 ff. Rostowzew, Skythien und der Bosporus. Vol. I., pp. 530 ff. J. Nestor, Bericht der römisch-germanischen Komm. 22 (1932), pp. 143 ff. N. Fettich, La trouvaille scythe de Zöldhalompuszta in AH III. Budapest, 1929, and the same author's Der skythische Fund von Gartschinowo. AH XV. Budapest, 1934. On the ancient tribes of South Russia see S. A. Žebelev in VDI., 1938 I. pp. 149 ff.

8. See among others Ebert, Südrußland im Altertum, pp. 339 ff. and his contribution in RLV XIII., p. 61. K. Kretschmer in RE II. R. I., pp. 2545 ff., M. Vasmer, Die Iranier in Südrußland. Leipzig, 1923., pp. 23 ff., and his contribution in RLV XII. p. 237.


of a later date. [9] But Rostovtzeff, one of the foremost authorities on Scythian and Sarmatian archaeology, went so far as to deny that the two peoples had anything in common apart from a superficial similarity in their names. The description of the Sauromatae by Herodotus (IV. 110—117) shows obvious traces of a matriarchy or gynaecocracy, and Rostovtzeff adduces this as an argument to prove that this Iranian tribe had absorbed a great many local ethnic elements of the land. On the other hand, as Rostovtzeff points out, not the slightest traces of a social organization can be recovered that would point to a matriarchy with the Sarmatians. [10] Rostovtzeff's arguments have been rejected by practically all the investigators, [11] and in his latest summary of the question he himself has undertaken a certain modification of his original attitude on the dissimilarity of the Sauromatae and the Sarmatae. [12]

It cannot be maintained that the position Rostovtzeff had originally taken, was the best way to get rid of the difficulties, though it must be also admitted that not much was gained either by identifying the two peoples or by looking upon the two tribal names as simply being doublets. It must on no account be forgotten that it has so far not been unequivocally established what the names Sauromatae and Sarmatae connote ethnically; such a delimitation has not even been attempted though, it stands to reason, that without attempting such a definition, the question can never be solved in a satisfactory way. As soon as we set about to remedy this deficiency, we shall find already at the outset that the problem is far more complicated than either Rostovtzeff or his antagonists have ever imagined. The name Sauromatae as employed by Herodotus (IV. 21, 110—17), seems to suggest that it was used as a designation of an Iranian tribe whose seats lay east of Scythia, and that an attempt was made by him to delimit their actual seats with some accuracy by means of cartographical terms. Hardly a century had passed after the time of Herodotus when Ephoros widened the term of Sauromatae, [13] while his successors employed the name to denote a number of actual and mythical peoples. [14]

A century and a half will have to elapse after Ephoros before the name Sarmatae crops up for the first time in its historically accepted form, but even then the evidence contained in this first mention is so scanty that it is hardly sufficient to define what ethnic features went with the name Sarmatae. [15] It was only considerably later that a picture was given of the ethnic background of the name Sarmatae by Strabo (VII, 5, 18) in an information that can be traced back to Artemidoros. Strabo called Sarmatae a number of tribes or peoples who in his time made their first entry into classical literature. He used the term in a rather general sense and employed it to call by that name a number of tribes that were newcomers on the stage of history. The name received an even wider range of application later in the first centuries A. D. when it came to

Additional Notes

To p. 9. The problem of the identity of the Sauromatae and Sarmatians was much discussed by Soviet scholars during the last decades. Regarding my argumentation expounded in the 1st edition as essentially correct even to-day, I only refer to the view-point of K. F. Smirnov which seems to be only right. He writes:

Cf my exposition on p. 9—10.

9. In addition to the literature quoted in the previous footnote see on this question the works of J. Marquart, Ērānšahr. Berlin, 1901. p. 155. E. Herzfeld in AMI I (1929—30), p. 102, footnote I. H. H. Schaeder, Iranica. Berlin, 1934., p. 50.

10. See Rostovtzeff, Iranians and Greeks in South Russia, pp. 32 ff., his Skythien und der Bosporus, vol. I., p. 101, and his book on The Animal Style in South Russia and China. Princeton (1929), pp. 44 f.

11. On the position taken by other scholars see among others Altheim—Szabó in WaG 2(1936), p. 318, footnote 24, and J. Junge, Saka-Studien. Leipzig, 1939. 9, footnote 2, pp. 73 f.

12. Cf. CAH XI., pp. 91 f.

13. See J. Harmatta, Quellenstudien zu den Skythika des Herodot. Budapest, 1941, pp. 18 f.

14. An attempt of this nature can be seen among others with Mela (I 116) who included among the Sauromatae peoples like the Budini, Thyssagetae and Iyrcae. See J. Harmatta, Quellenstudien zu den Skythika des Herodot, pp. 8 f., pp. 11 f. and p. 19.

15. The Sarmatae were first mentioned without any doubt by Polybios who included Gatalos, king of the Sarmatae, as one of the parties to a treaty concluded in 179 B. C. The passage can be found in Polybios XXV 2.


be applied to peoples who formerly used to be well-known in geographical literature but who had since then been entirely lost sight of. [16]

This brief survey in itself will suffice to convince that the ethnic entities associated with the names Sauromatae and Sarmatae, may not be identified without reservations, not even if proofs were forthcoming that both names happened to be identical. [17] Such an erroneous identification would lead to a number of difficulties. How are we going to account for it in a satisfactory way why the name Sauromatae, that had already acquired a rather general application in the 4th century B. C., should come to be used in a narrower sense by Strabo in the form Sarmatae to denote a number of Iranian tribes that had but shortly been brought to the notice of the contemporaries? This latter fact undoubtedly suggest that a new wave of migration had by then broken over the steppes of South Russia. Such a belief receives confirmation from archaeological evidence, too. It was no other than Rostovtzeff himself who examined a portion of South Russian archaeological material from the last two centuries B. C., and in reference to the gilded silver phalerae, that characterised one group of finds, he came to the conclusion that the style of these phalerae stood in a rather close relation to Graeco-Indian art. [18] In view of the great number of relevant finds, this relationship, according to Rostovtzeff, can only be accounted for by assuming that the phalerae must have been used by tribes that had formerly been settled in the east in a close vicinity to Indo-Scythian tribes from whom the style of workmanship had been adopted and brought to South Russia.

Premissa like the foregoing make it rather likely that in the last centuries B. C., there had appeared a number of new Iranian tribes coming from the east. This again involves that, speaking ethnically, the names Sauromatae and Sarmatae must on no condition be identified, not even if it is assumed that the two names happened to be identical; on the other hand we may surmise that the ethnic entities of the Sauromatae—Sarmatae had undergone a change in the intervening period.

A careful scrutiny of the results obtained concerning the identification of the Sauromatae and Sarmatae peoples, suggest the conclusion that the original seats of the Sarmatae have to be put considerably further east but on no account with the Sauromatae of Herodotus. The name Sarmatae could not have been the name of one single tribe only, it must have been much more a collective name for a number of tribes scattered over a wide area. This again implies that the ethnic background of the name Sarmatae included features widely divergent in time as well as in geographical distribution.

Among the epigraphical sources of the ancient history of South Russia the so-called Protogenes inscription has been given an outstanding significance. [19] The inscription was found on a memorial tablet dating from the beginning of the 2nd century B. C., [20] and was dedicated in honour of Protogenes, her much esteemed citizen, by the Greek town Olbia in grateful acknowledgement for the help received

Additional Notes

To p. 10. Following V. Latyschew D. M. Pippidi has dated the Protogenes inscription to the first half of the 3rd century B. C. recently (Epigraphische Beiträge zur Geschichte Histrias in hellenistischer und römischer Zeit. 15). That this dating is impossible and that we can only think of the end of the 3rd or the beginnings of the 2nd century B. C., was convincingly demonstrated by the careful study of T. N. Knipovich (VDI 1966/2. 142—149). The Galatae of the Protogenes inscription may represent the Bastarnae who can be regarded as a Celtic tribe according to the recent investigation by Z. Mády ( Moscow 1967. 179 foll.).

16. Cf. Pliny the Elder's Natural History VI 19. See also J. Harmatta, Quellenstudien zu den Skythika des Herodot, p. 11.

17. From a linguistic point of view the names were identified by Marquart, Untersuchungen zur Geschichte von Eran. II. Leipzig, 1905, p. 78 and in his Ērānšahr. p. 155. See also Vasmer, Die Iranier in Südrußland, p. 51. E. Herzfeld in AMI I (1929—30), p. 102, footnote l. Schaeder, Iranica, p. 50. Contrary to them N. S. Nyberg, Die Religionen des alten Iran. Leipzig, 1938, p. 250 considered the two names to derive from different roots.

18. Rostovtsev, Sarmatskiya i indoskifskiya drevnosti. Recueil-Kondakov. Prague, 1926, pp. 239 ff. See also N. Fettich, Die Metallkunst der landnehmenden Ungarn. AH XXXI. Budapest, 1937. pp. 142 ff.

19. Cf. Dittenberger, Sylloge Inscriptionum Craecarum. No 495.

20. See a recent article by Altheim—Szabó in WaG 2 (1936), p. 319.


in many of the crises that had confronted the community. The inscription gives us a close-up of the hard times that had come upon the once prosperous town. The flourishing and peaceful life of the town to which Herodotus bore testimony, had been a thing of the past by then. Numbers and numbers of new peoples threatened to sack and destroy the town (cf. lines 102 ff. of the Protogenes inscription). The inscription gives a list of these new peoples by name such as the Saii, Galatae, Skiri, Thisamatae and Saudaratae. From among these only the Galatae and the Skiri are known to us from other sources. The Galatae were a Celtic tribe of South Russia whose presence can be proved by archaeological evidence. [21] The Skiri [22] were a Teutonic tribe who were to play some part in the age of the Huns. The other tribal names mentioned on the inscription such as the Saii, the Thisamatae, the Saudaratae, have never been mentioned in any other sources.

There are, however, clues that contribute to our knowledge of these otherwise unknown peoples. The inscription includes the name Saitapharnes, king of the Saii, and his name can be established without doubt to have been an Iranian proper name. [23] This is a useful hint to establish the origin of the tribal name Saii, [24] which can be sufficiently explained as an Iranian derivative and its meaning is multi-coloured. [25] The adjective was often used as a proper name with a number of nomad tribes, and especially with the horsebreeding nomads it used to refer to the colour of the tribe's horses. [26] Thus among others we know a number of Turk tribes with the tribal name Bulaq ( = multi-coloured). [27] This is significant in so far as it may serve in a way as a hint to trace the origins of the Saii. A multi-coloured type of horse was known in Chinese records, [28] and may, therefore, be taken as a typical Asiatic equine variety. [29] Undomesticated specimens of this breed were still seen by Przewalsky in Asia. [30] If,

Additional Notes

To p. 11. I now prefer the interpretation of the name Saii from Old Iranian *xšaya- and that of Saitapharnes from Old Iranian *xšaita-farnah-. Cf. also p. 94—95, 107.

21. See Rostowzew, Skythien und der Bosporus, vol. I. p. 465 ff.

22. L. Schmidt deals with them in his Die Ostgermanen. München, 1941, pp. 47 f.

23. See Vasmer's Die Iranier in Südrußland. p. 50. It was justly pointed out by Tomaschek that the first element of the compound name may be compared with Avestan šaēta- (= Geld, Vermögen) and the second element to Avestan xvarənah-, Old Persian farnah- (= Ruhm, Ruhmesglanz, Herrlichkeit, Hoheit, Majestät). The name Saitapharnes may, therefore, be related to an Iranian *šaita-farn. This is an instance of the bahuvrīhi type of word-composition, and it may be rendered by "der durch Vermögen Herrlichkeit besitzt".

24. Tomaschek in his Die alten Thraker, I. p. 99., connected the word Saii with Avestan xšaya- which means "Herrscher, Fürst, König". Vasmer in Die Iranier in Südrußland, p. 50., doubts the possibility of such a comparison since the Greek transcription of the name points to an intitial s- or š-. Against this we have to point out that in some of the New Iranian languages a sound-change from - to š- is an established fact. Thus e. g. the outcome of Old Iranian xšaya- sounds in Wakhi and Šugni as follows: Wakhi šāi "fat, rich", Shughni šayēn "khans". Although the modern forms of Old Iranian *xšaya- entirely coincide, as far as phonetic development goes, with the tribal name Saii, yet this coincidence may be a fortuitous one since the Iranian dialects in South Russia have not so far yielded any evidence that would justify to assume a phonetic change from - to š-.

N. Jokl in RLV XIII, p. 281, pointed out the phonological identity between the word Saii and the Thracian tribal name Saii. Dittenberger, however, has proved that this contention is far from being likely. See the latter's Sylloge Inscriptionum Graecarum. I. p. 739, footnote 12.

25. Compare Avestan sāy- (= ungleichmäßig gefärbt, scheckig), sāyuždrī- "Eigenname eines Gläubigen". Specific meaning of latter: 'des weibliche Zugtiere scheckig sind' Bartholomae, Altiranisches Wörterbuch. 1569, 1572.

26. See J. Németh in KCsA 1. Ergänzungsband (1938) pp. 345 fT. J. Harmatta in MNy 42 (1938), pp. 27 ff.

27. See J. Harmatta in MNy 42 (1946), p. 31.

28. Cf. E. Chavannes, Documents sur les T'ou-Kiue (Turcs) occidentaux. St. Petersbourg. 1903, p. 29.

29. For further information on this point see J. Németh in KCsA l Erg. Bd. (1938). pp. 349 ff.

30. See Bretschneider, Mediaeval Researches. I. 168., p. 463 footnote.


therefore, the Saii had a peculiar breed of horses, and this may be assumed, then they themselves together with their horses, must have come into South Russia from the West-Asiatic steppes.

This evidence is of great importance since in the Saii we believe to have got hold of the first eastern tribe that had been pushed along by the new Iranian wave of migration. And this new Iranian wave seems to be significant. The name Saii covers, namely, not only one single tribe but rather a federation of tribes since the Protogenes inscription mentions their tribal chiefs in the plural number (cf. lines 43 f.). Further the name Saios [31] on one Panticapaeum inscription may be taken to witness to their subsequent spread eastwards and attest the fact of their survival.

The other tribal names on the Protogenes inscription: Thisamatae and Saudaratae, are not unlikely of Iranian origin, [32] that is to say, like the Saii tribe, we may also take these two peoples to have been Iranians. And since the inscription made separate mention of the Scythians, it is not unreasonable to assume that the two tribes did not belong to the Scythians; as we have done, with the Saii, we may take them also to have belonged to the new, eastern tribes of Iranian descent. It is of decisive importance, therefore, that the Protogenes inscription did not call any of these newcomers by the name of Sarmatae; nor can it be said that the name Sarmatae was not known in those days for Polybios mentioned it in the peace treaty of 179 as referred to above. There is one explanation open to account for this strange circumstance, and that is that the name Sarmatae was not a tribal proper name but only an appellation of a more general application, the use of which was spread by literary means. This, of course, makes it peremptory to search for the solution of the difficulties attached thereto, more on literary basis than by any other means.

31. See Vasmer, Die Iranier in Südrußland, p. 50.

32. See Vasmer, Die Iranier in Südrußland, p. 51.

[Previous] [Next]
[Back to Index]