WHO WERE THE ARIMASPEANS?

MICHAEL HEANEY

Bodleian Library, Oxford, England OX1 3BG, U.K. [1993]

Reproduced from Folklore, volume 104 (1993), pp. 53-66 © 1993

Web version with minor additions: Michael Heaney 2004.

[This article originally appeared in Folklore, volume 104 (1993), pp. 53-66, as part of a joint article with Adrienne Mayor, 'Griffins and Arismaspeans', pp. 40-66, of which the first part, 'What were the Griffins?' (pp. 40-53) was by Adrienne Mayor. In the first part of the article Adrienne Mayor demonstrated that the legend of the gold-guarding griffins is probably an aetiological myth intended to explain finds of fossil Protoceratops bones among gold-bearing deposits in the Altai Mountains in Central Asia. The legend that these griffins fought off marauding Arismaspeans comes down to us from Aristeas who, according to the reports of his work which have survived, claimed to have visited the Issedonians in Central Asia around 675 BC and learned the tale from them.]

Looking at the petrified battlefield, the Issedonians naturally cast around for the antagonists of the monstrous griffins. According to the tale which has come down to us, the Issedonians averred that their neighbours the Arimaspeans fought the griffins. If we consider the gold-guarding griffins to be flights of Issedonian fancy, it is natural to see the Arimaspeans as a real tribe about whom the Issedonians told tall tales. Once we realise that it is the griffins which are real (or at least have a solid base in reality), the explanation which the Issedonians gave about the demise of the griffins raises a new question. If the Issedonians were seeking a rational explanation for the presence of the bones of ceratopsian dinosaurs, how and why could they attribute them to battles with their 'neighbours' the Arimaspeans?

If the Issedonians had any social intercourse with the Arimaspeans, they could not be under the delusion that the Arimaspeans were the cause of the griffins' deaths. The Arimaspeans may have had their own explanation for the presence of dinosaurs' bones, but they would not have told the Issedonians, 'we fight with them'. It is plain that the chain of thought was not 'there are Arimaspeans, who fight griffins'. but 'there are dead griffins; someone must have fought with them; the deaths must be the result of battles with the Arimaspeans'. The Arimaspeans come into the picture as part of the attempt to rationalise the presence of dinosaur bones.

The Arimaspeans lived beyond the Issedonians, and beyond the Arimaspeans lived the griffins.(41) They are, therefore, a people inhabiting the approaches to the Altai Mountains -that is, northern Sinkiang and western Mongolia - who have no contact with the Issedonians although geographically close to them. Issedonians and others involved with the gold deposits probably lived near the oases along the trade route, and travelled through the lands where Arimaspeans dwelt, without ever meeting them, in order to reach the gold.

Classical sources are more reticent about the Arimaspeans than they are about the griffins. All the sources are presumed to derive ultimately from Aristeas. Aeschylus (Prometheus Bound, 803 ff.) characterises them as one-eyed horsemen; Herodotus agrees, and adds that they steal the gold and that they drove the Scythians west. The only other writer to add more information is John of Tzetzes, a late Byzantine writer who is generally acknowledged to have had access to earlier Greek sources now lost. He characterises the Arimaspeans as strong warriors, good horsemen rich in flocks of cattle and sheep and goats; they are one-eyed, 'shaggy with hairs, the toughest of men'. The hairiness of the Arimaspeans is confirmed by the unique Kelermes mirror, which appears to show two Arimaspeans in the grypomachy. In the mirror, although the Arimaspeans are hirsute they show no sign of monocularity. They are shown in profile, so only one eye is visible on each; but they face each other, and the eyes are placed correctly for binocularity. (42)

Arimaspeans on the Kelermes mirror [Added 2004]

The Arimaspeans are wild, hairy, and one-eyed; they are horsemen, and possess rich herds of cattle, sheep, and goats. What is remarkable is that these traits are variously attributed to a 'people' said to inhabit the same area today, with a very similar name: almas, almasty, or albasty. The name in its various forms is widespread in the Mongolian, Turkic, and Iranian languages of Central Asia and the Caucasus.(43) The tales told about the people vary from extremely naturalistic to wildly improbable. Some scholars give credence to the idea that these are real creatures, being representatives of a primitive hominid species. The most prominent researcher of the topic, the late B.F. Porshnev, assembled a wealth of material and rediscovered the forgotten works of a number of previous researchers, most notably of the Mongolian scholar Tsyben Zhamtsarano. Porshnev's work is the best source for much of our knowledge of tales and legends about these creatures, but he is undeniably controversial, partisan, and thereby given to overstating his case.(44) We shall therefore cite a more balanced summary exposition of Mongolian beliefs by the Mongolian scholar Damdinzharyn Maidar: (45)

Almases, according to the stories of witnesses, appear half animal, half human, with reddishblack hair. The face is hairless, the stomach covered with sparse growth. The head seems pointed at the occiput, the forehead flattened back with projecting brow ridges, and prominent cheekbones. They are the height of an average person. The almas walks with half-bent knees, is round-shouldered and pigeon-toed. It has broad shoulders and long arms. The women have long breasts. Almases are timid, suspicious, but not aggressive, and lead a nocturnal way of life. No-one has heard their speech.

They have been encountered in the majority of cases at dusk and dawn. They feed on roots and plants. Almases seek out uninhabited places, so are found in the mountains far from populated spots and nomads' field camps, but stay close to wild sheep [argali] and wild goats [iangir iama]. When herders, for example, in Tangal somon in Baian Ulegei aimak [in western Mongolia] move their herds from their mountain winter quarters to the distant summer pastures, their places at the warm time of the year are occupied by wild sheep and goats, and following them almases usually appear. Herders say that they meet almases in pairs, or alone, and one once saw a small almas child.

Tsyben Zhamtsarano, according to Porshnev, researched the distribution of reports of almases and detected a shrinking area of distribution: at the beginning of the nineteenth century reports came from all over southern Mongolia and the Gobi desert regions of Inner Mongolia; by the time of his researches, in the 1920s, reports came only from the Gobi Desert and the Khovd region of western Mongolia, which includes the Altai Mountains.(46) Zhamtsarano's pupil Rinčen adduces the many place names in southwestern Mongolia involving almases, e.g. Almasyn Ulan Khada, 'Red Cliff of the Almases'; Almasyn Dobo, 'Almas Hills'; Almasyn Ulan Oula, 'Red Mountains of the Almases'.(47)

A. M. Markova, a researcher at a southern Kazakh agricultural institute, who investigated the folklore attached to horses by the local people, reported that Kazakh herders believed that almases sucked the milk from their mares at night, and plaited the horses' manes. (48)

If we move further from the area of the almases' supposed habitation, more fantastic elements may be found. Rinčen notes this for northern Mongolia; (49) but the creature so named has been especially well discussed in relation to the beliefs of the peoples of the Pamirs region, some nine hundred miles southwest of the region we are proposing for the Issedonians, the Arimaspeans, and the Mongolian almases. B. A. Litvinskii describes the albasty as follows: (50)

In general ... albastys appear to people in the form of a woman (rarely, a man). It is usually a very fat, ungainly woman, covered in hair, or a gaunt swarthy woman with dishevelled yellow or black hair and drooping breasts, or a dirty old woman in torn clothing with matted hair. The albasty's breasts are so long that she sometimes throws them over her shoulder. Her body is black. The albasty's zoomorphic manifestations are as a lion, a tiger, bear, cat and dog. The dog-albasty is big and ungainly, with matted hair (often yellow) with two spots above the eyes. The Mountain and Pamiri Tajiks suggest at times that the albasty has the form of a wild, foul-smelling man with the muzzle of a wild boar, crooked fangs and a hairy black body. A characteristic feature of the albasty are numerous breasts, with an infant at eaoh of them. The breasts secrete venom, which she tries to give to the infants, but they also contain milk as sweet as honey, which she often feeds to them. The albasty harms not only women in labour and infants, but smothers people, horses and sheep in their sleep.

N.A. Kisliakov writes that the Mountain Tajiks of Vakhan and Ishkashim believed 'that almastys are encountered with two eyes or with one eye on the forehead'.(50a). Another writer on the mythology of the Pamirs, A. K. Rozenfel'd, characterises the 'almasty or albasty' as 'a werewolf harming women in labour and young children'. She also discusses another anthropomorphic figure of Pamir mythology, the gul'biiaban, but goes on to say: (51)

There is a widespread independent cycle of tales about the wild man (adam dzhapaisy), only sometimes intermingling with the tales about the gul'biiaban. The basis of this cycle is the concept of a naked man, without clothing, herbivorous, avoiding people and never coming into contact with them. While the gul'biiaban is said to live in the mountains of the Pamirs, the 'wild people' are firmly said not to live in the Pamirs, but 'they have been seen in western China'. Conceptions about the 'wild people' are very hazy; usually Kirghizes reply to questions about them by saying, 'Just as there are wild horses, camels and other animals, so there are wild people too'.

Reports of wild people in Central Asia may be found throughout history. A thirteenth-century Armenian traveller reported that the desert was 'inhabited by naked wild men with horse-hair on their heads. The breasts of the females were extremely large and pendant'.(52) At the beginning of the fifteenth century Johann Schiltberger visited the Khanate of Siberia, and reported:(53)

There is a mountain in that country, which is thirty-two days' journey in extent [a reference either to the Altai or Tien Shan mountains - MH]. The people there themselves say that at the extremity of the mountain is a desert, and that the said desert is the end of the earth; and in this same desert nobody can have a habitation, because of snakes and wild beasts. On the same mountain there are savages, who are not like other people, and they live there. They are covered all over the body with hair, except the hands and face, and run about like other wild beasts in the mountain, and also eat leaves and grass, and any thing they can find. The lord of the country sent to Edigi [Edigei], a man and a woman from among these savages, that had been taken in the mountain.

So the idea of wild people in northern Sinkiang and western Mongolia is widespread, and in the immediate area of their supposed habitation they are often known by the name almas or a variant thereof. Further away from this area the same name is applied to a creature or creatures with various otherworldly attributes, though this creature is distiguished from the 'wild people'. The potential for conceptual confusion and intermixing is great, and it is not surprising to find that some of the tales of the almas in Mongolia exhibit some features reminiscent of the more fantastic elements of the tales current in the Pamirs. The Mongolian scholar Rinčen reported that in 1928 he spoke to an old woman of seventy who:(54)

as a baby left momentarily alone by her mother had been suckled by a naked woman covered with sparse reddish hair, whom her mother had surprised on her return to the yurt ... As the hairy woman ran off her mother noted that the woman's legs were bent, her arms unnaturally long, and that she was pigeon-toed.

Though outwardly naturalistic, this tale clearly contains elements of the belief that the almas is a threat to mothers and newborn children. The various beliefs and perceptions feed upon one another; and once the ancient Issedonian-Scythian tales reached Greek ears through Aristeas, Greek beliefs could also remodel the original elements of the story.(55)

In seeking to identify the Arimaspeans we need a hairy people in Central Asia, who were perhaps one-eyed, were rich in animal stock (notably sheep, cattle, goats and horses), but who had no contact with their neighbours. In the almases we have hairy people in Central Asia, avoiding contact with their neighbours, and whose name is also applied to a variety of mythological beings with a variety of traits, including monocularity. Almases are said to live in association with the wild sheep and goats, in the same region as the wild horse (Przewalski's horse), and are said to suck mares' milk. It is a short step from believing that almases live commensally with the indigenous wildlife to believing that they (being anthropomorphic) have dominion over them. This is suggested by a legend collected by Badzar Baradiin when travelling across the Gobi Desert: (56)

Our Alashanians told us that in the sandy regions of Alashan [an area of the Gobi Desert in Inner Mongolia - M.H.], especially its western half, there are sizable oases with brackish and sometimes even freshwater lakes with splendid pastures. Alashanians live with their herds very wealthily in these oases, which are lost in the most inaccessible depths of wide expanses of sand, and so successfully avoid the persecutions of the authorities, and heavy taxes and duties. The people of these places, are very wild and avoid the outside world, they say.

Although this report does not mention the almas, it seems to underlie later accounts which aver that Baradiin actually saw an almas at the time he wrote this journal entry.(57)

On their characteristics alone, almases make very good candidates for the Arimaspeans. Arimaspeans have been interpreted by classicists as shamanistic beings, an isolated real tribe, 'squinting archers' and miners with headlamps;(58) but there is no need to 'interpret' the sources in this way when their attributes can directly be matched with those of the almases. The remarkable coincidence of their respective names provides further compelling support for the argument. Their phonetic similarity has not been discussed previously, although the etymology of each word has been and continues to be the subject of much speculation.

The earliest etymology supplied for the root arimasp- is that given by Herodotus (4.27): he reports that arima is Scythian for 'one' and spu Scythian for 'eye'. This has no support in our present knowledge of Scythian; the Scythian for 'one' was aiva.(59) What is lacking in direct Scythian sources can be supplemented by our knowledge of other Iranian languages, notably Old Persian. Tomaschek points to an Iranian proper name Arimaçpo, supposedly meaning 'owner of wild/steppe horses', and adduces Avestan airima, 'desert', and the known Scythian aspu, 'horse' He suggests that the secondary association of 'desert' with 'solitude' is the source of the mistaken attribution by Herodotus of the meaning 'one' to the first element of the word. (60) Modern sources are more hesitant, accepting only the link with Iranian aspa, 'horse'.

'Desert horse' (there is no need to postulate the additional semantic element 'owner of . . .') may be acceptable as an etymology, given that almases are said to be found in the same region as the wild Przewalski's horse. Scythian tribes lived in a totemic society (Saka, the Ancient Persian appellation for the Scythians, is a totemic name meaning 'deer'), and their attribution of a similar animal name to the supposed wild people living in association with the wild horses of the desert is quite feasible. 'Desert horse' is certainly much more acceptable than Bolton's speculation of äräm däk, Mongolian for 'one-eyed'. (61) It also provides a basis for the belief that the Arimaspeans were 'horsemen', the phrase originally denoting totemistic affiliation later being interpreted figuratively. However, we must also consider the various etymologies suggested for almas and its variants, and the possible linguistic relation between arimasp- and almas.

Almas and its variants have, like arimasp-, been discussed extensively, with greatest emphasis on the forms found in the Pamirs. The most thorough treatment, which also reviews most of the earlier explanations, was published by G. A. Klimov and D. I. Edelman in 1979. (62) The most widely accepted etymologies up to the publication of Klimov's and Edelman's article were those of M. S. Andreev and E. Benveniste. Andreev related the word to the ancient divinity Al and Turkic basty, 'smother'; (63) Benveniste to the Indo-Iranian demon *Al ('*' denotes a hypothetical, unattested form) with the same Turkic root for the second half of the word.(64) Andreev dismisses the form in -m- as a 'variant in popular pronunciation'; Benveniste ignores it altogether. Both see the word as basically an Iranian importation from Turkic.

Klimov and Edelman draw attention to the inadequacies of these approaches, especially as regards the variants in -m-: almasty, etc. They point out that the majority of forms in -b- are in Turkic languages, where folk etymologies involving the root -basty identified by Andreev and Benveniste may be at work. Albasty may be explained as a derivative in folk etymology from almasty, but not vice versa. They infer that almasty is the primitive form, and derive that in turn from the name of the Iranian goddess Lamastu/Lamaštu, who, like the albasty, threatened breastfeeding children.

All of the above etymologies concentrate on the forms ending in -ty, presumably seeing the form almas as derived from them. However, Rinčen sees almas as primitive; he writes:

The origin of this old name is quite unknown to us: but we can recognise its variations in Turkish languages: Almasty, Albasty, etc. . . . Albast, Alboosty, also Almasty, are the Turkish forms of Mongol almas with an addition of suffix -ti/ty, which only [sc. alone] is explicable.(65)

Rinčen is not explicit, but he probably has in mind the Turkic denominative suffix -di/dy, giving the meaning 'the almas one'.Presumably, like Klimov and Edelman, he would derive albasty secondarily from almasty.

If arimasp- and almas are related, we must examine the mechanisms by which both may be derived from a common ancestor, or by which one may be derived from the other. Arimasp- is presented as a word used by the Issedonians, who may have been Scythians, or who may have spoken through Scythian intermediaries. The Iranian languages generally lost the phoneme /l/ at an early period in their development (second millennium BC), and it may be that arimasp- reflects a non-Iranian form *alimasp borrowed into an Iranian language. The history of the Scythian language, however, strongly suggests that arimasp- is a Scythian word.

The whole swathe of the Eurasian steppe was inhabited in the first millennium BC by Iranian tribes known under several names, but generally identified by Greeks as Scythians and by Persian sources as Sakas. They penetrated eastwards as far as the Tarim basin in southwestern Sinkiang. Despite their wide distribution it is probable that they all spoke closely related dialects of the same Scythian language; their nomadic way of life minimised the development of divergent speech patterns between communities.(66) This linguistic unity did not preclude the use of a variety of tribal names, such as Massagetes, Skolotes, etc. It is quite probable, given their apparent geographical location, that the Issedonians were one of these tribes; this is also suggested by the fact that Ptolemy refers to Issedonian 'trading stations', apparently near Turfan ('Issedon Scythica') and An-hsi (east of Lop Nor, 'Issedon Serica').(67) Some modern scholars locate the Issedonians west of the Dzungarian Gate, along the northern slopes of the Tien Shan, in the former Soviet Union; others follow Ptolemy's placement at the eastern foot of the Tien Shan, in oases along the old routes between the Gobi Desert, the Tarim basin, and the Dzungarian Gate.(68) Perhaps Ptolemy's two names do no more than reflect the presence of Scythians both north and east of the Tien Shan, which mark the border between the Eurasian steppe and Sinkiang.

The Tien Shan formed a barrier between the Scythians on either side of the range; the Scythians of Sinkiang would not have easily participated in the flow of movement which accounts for the linguistic unity of the Scythians of the Eurasian steppe. It is not surprising, therefore, that the language of these Scythians (who gave rise to the Khotan Saka state, with a considerable body of literature, in the second half of the first millennium AD), is thought to have diverged from 'standard' Scythian early on, from about the middle of the first millennium BC.

The evidence for the divergence of Khotan Saka is germane to this issue, because it rests on the fact that Khotan Saka did not participate in the sound change /r'/ > /l/, which took place in Scythian towards the end of the first half of the first millennium BC (e.g. aryana > alana) The evidence for the date of the change comes from Herodotus, who cites Scythian words not exhibiting the change (e.g. anaries < Sc. anarya) and words showing it (e.g. Kolaxais < Sc. hvaraxšaya). At the time of Aristeas, therefore, a Scythian word arimasp- may have existed, and could have been taken over into Greek. In that form it would have been frozen in the Greek language, not being subject to developments in Scythian phonology; but in the Scythian language it would have developed naturally into *alimasp-.(69)

If arimasp- developed into *alimasp-, we may match this against the known developments in Mongolian historical phonology. The difficulty is that history fails us. The known historical phonology of Mongolian takes us back no more than eight hundred years, and we must span two-and-a-half millennia. However, there is some evidence we can adduce. First, almas is not a form which necessarily arises as a result of known developments in Mongolian phonology: it may be ancient in exactly that form. Second, there is evidence that Mongolian has lost unstressed medial /ĭ/, e.g. literary Mongolian solbi < *solĭbu, cf. the root soli; tuγurbi < *tuγurĭbu.(70) Third, Mongolian now lacks word-final -sp; if it ever existed it has now gone. Ancient Mongolian (ca. AD 1000) lacked the phoneme /p/ entirely: in borrowings of that period, /p/ in isolation becomes /b/. (71) In other words, an original *alimasp- may also be expected to give almas in modern Mongolian.

The phonological path traced above from arimasp- to almas suggests that Tomaschek was right to seek an etymology in the Iranian languages. The most straightforward solution is that the name simply means 'desert horse'.The Scythian arimasp-, developing into *alimasp-, accounts for the Mongolian form almas. In the Turkic languages, the form almasti is probably the result of dissimilation from *almaspi, the dissimilation taking place either in Scythian, or after borrowing into Turkic. (72) The former is more likely, given that forms in -m- are commonest in the Iranian languages; but it should be pointed out that the present-day Iranian languages of the area are not the linear successors of Scythian, and cultural continuity cannot be presumed. In Turkic languages almasti would then have been subjected to the processes of folk etymology suggested by Klimov and Edelman, producing albasty.

The almases of Central Asia, be they mythical or real, match the Arimaspeans of Aristeas in name, appearance, habits, and geography. It is they whom the Issedonians chose as the slayers of the griffins; they who avoided humans, so that the Issedonians never saw the battles themselves, but found only the petrified remains of the vanquished griffins.

Bodleian Library
Oxford OX1 3BG

A NOTE ON TRANSLITERATION

Where there is no established English form for Russian, Mongolian and Chinese proper names, Library of Congress preferred forms are used. An exception is made in the case of Turkic-speaking Sinkiang, where the Library of Congress transliteration of the Turkic form of names is used in preference to the jurisdictionally correct but unfamiliar Chinese form. Where Roman-script translations or publications of works by Russian or Chinese authors are cited, the form given in the publication is used. In the linguistic discussion, a phonemic transcription is used.

NOTES