The Greeks in Bactria and India c10.s5: Chapter VII. The Nomad Conquest of Bactria
The Greeks in Bactria and India c10.s5
- Language
- English
Chapter VII. The Nomad Conquest of Bactria
c10.s5.2
AFTER the death of Eucratides his eldest son Heliocles ultimately acquired control of the Bactrian kingdom. It may be supposed that the country rallied to him as the one effective force against the Parthians; the fact that the title on his coins,¹ *δίκαιος*, The Just, was adopted from the Euthydemids may conceivably mean that he tried to reconcile the Euthydemid partisans. Doubtless the return of the rest of Eucratides’ army from India gave him an accession of strength; but whether Mithridates of Parthia was expelled or bought off, or whether the virtual defection of his Sacas (Chap. v) induced him to quit the hostile country which he could hardly have held, and make sure of Seistan and Arachosia which he *could* hold, does not appear. Certainly the outlying provinces of the Bactrian empire in Iran were lost; the Parthian frontier was again the Arius, and (though there is no evidence) Mithridates presumably retained Herat, or he would have had no through communication between Parthia proper and Seistan. Heliocles probably retained Merv; otherwise his kingdom was apparently reduced to Bactria and southern Sogdiana. What happened to northern Sogdiana, the plain of the Jaxartes, will be considered later.
c10.s5.3
¹ *BMC* pp. 21 *sqq.*, p. 166; Whitehead *NC* p. 321. One of his was found in the Mandali hoard, east of the Tigris, which was buried c. 90–85 B.C., a long way for a Bactrian coin to travel: E. T. Newell, *NC* 1924 pp. 151, 160.
c10.s5.4
Justin,² speaking of the result of the wars of Eucratides, says that Bactria bled to death; the statement may be true, though it is put too early, for Heliocles’ conquests in India show that the country must still have possessed a fair degree of strength; he would of course retain Eucratides’ mercenaries. Heliocles had probably kept what Eucratides had had south of the Hindu Kush, Alexandria-Kapisa and the Paropamisadae, for, as has been seen, there is no real reason to suppose that Menander ever ruled that country; if so, Heliocles held the gateway into India and could, as he subsequently did, invade it when he pleased. Looking at what is now known, it seems that he ought to have conserved his strength for the coming shock from the north; but this may hardly have come into view by the time that he invaded India, and he may have thought that the possession of north-west India would strengthen him; probably too he regarded Gandhāra, which his father had conquered and lost again, as his by right. The idea that he wanted India as a refuge against the day when he should be driven out of Bactria may be disregarded; Greek kings did not practise what in modern jargon is called defeatism. The beginning of his reign was probably occupied in getting rid of the Parthians, and in India he made no move while Menander lived; that monarch must have seemed, and perhaps really was, too strong to be attacked. But when he died, leaving his widow regent for his young son, Heliocles at once invaded Gandhāra; some of his coins are found overstruck upon those of Agathocleia’s first issue as regent, and others upon those of Strato after he took over the rule.¹ This war then lasted three or four years at least, and its ultimate consequences were in every way disastrous; Heliocles, as the sequel shows, weakened rather than strengthened himself, while Menander’s vast and loosely-hung realm began to break up; its maintenance had obviously depended upon his own personality. Heliocles, as his coin-types show, conquered and annexed Gandhāra and Taxila,² which drove Strato back upon Menander’s original frontier, the Jhelum; and in northern India Strato retained only the eastern Punjab and the country south-eastward to Mathurā. It became more difficult for his government at Sāgala to communicate with Sind and Surastrene; whether he governed them by sub-kings or satraps, the southern provinces may gradually have begun to shift for themselves till they fell a prey to the Sacas; that Barygaza continued indefinitely to use or copy the coins of Apollodotus and Menander and nothing later is eloquent, but really nothing is known of the southern provinces.
c10.s5.5
² XLI, 6, 3.
c10.s5.6
¹ Rapson, Corolla Numismatica 1906 pp. 246-7. The belief that Strato overstruck a coin of Heliocles is unfounded, *ib.*
c10.s5.7
² Square bronze bilingual coins with type of humped bull and elephant. The humped bull makes Gandhāra certain; on the elephant as representing Taxila see pp. 163 sq.
c10.s5.8
Heliocles, when he invaded India, left the younger Eucratides (Eucratides II) to govern Bactria, with the royal title; whether he was his son or younger brother is unknown, but on the ages the latter is much more probable. He struck the coins with a youthful head remarkably like that of Heliocles and the type of Apollo with his bow, holding an arrow in his right hand, and the simple legend ‘Of King Eucratides’. ³ These coins used to be called an early issue of Eucratides (I);¹ but it has been suggested that the face is too young and too like that of Heliocles to be meant for Eucratides himself,² and the matter seems settled by a unique coin which recently reached Germany from Russia.³ The description of it gives the same reverse type, Apollo with his bow and arrow, but on the obverse a ‘youngish portrait bust’ and the unique legend βασιλέως σωτῆρος Εὐκρατίδου, ‘Of King Eucratides, Saviour’. Had Eucratides really used the title σωτήρ at the start, it is inconceivable that he would have discarded it when he adopted the type of the Saviour gods, the Dioscuri; and this coin seems sufficient proof of the existence of a second Eucratides.
c10.s5.9
³ *CHI* p. 460; see Pl. IV, 8 and 9.
c10.s5.10
¹ *BMC* p. 13 nos. 1-5, Pl. V, 4 and 5.
c10.s5.11
² By Sir G. Macdonald in *CHI* p. 460.
c10.s5.12
³ Auctioned by Schlessinger in Berlin, 4 Feb. 1935. The particulars are from the sale catalogue, communicated to me by Mr E. S. G. Robinson. I do not know who bought it.
c10.s5.13
The question was long ago raised whether there was not also a second Heliocles, to whom the bronze bilingual coins struck in India would belong.⁴ Certainly the face on the Indian coins is not much like that of Heliocles;⁵ but little reliance can be placed on the portraiture upon coins struck in India, the types are those of Heliocles, and no one but himself can have been the conqueror of Gandhāra and Taxila. What would be conceivable, however, is that Heliocles was doing exactly what Demetrius had done; when he invaded India he might have left his eldest son Eucratides II to govern Bactria and then left his second son, another Heliocles, to govern the Paropamisadae, and the younger Heliocles might have coined for his father precisely as Demetrius II had coined for Demetrius, striking coins with his father’s types and legend but with his own head. The difficulties in the way of believing this are however considerable. Heliocles’ invasion of Gandhāra cannot be later than about 145, if as late, and it is unlikely (see p. 218) that he himself can have been old enough at that time to have had two sons of an age to govern; and even were he older than I have supposed, the head on the Indian coins is not that of a particularly youthful man.
c10.s5.14
⁴ By P. Gardner in *BMC* p. 23.
c10.s5.15
As Menander’s death occurred between 150 and 145, Heliocles’ war with Strato probably takes us a good deal of the way to 141, when for the last time the Bactrian kingdom is mentioned in literature. While Heliocles was waging civil war, Mithridates I of Parthia had acquired and was consolidating a very considerable empire, embracing most of Iran (except Bactria) from the Caspian Sea to the Persian Gulf. Late in 142 he made himself master of Babylonia,¹ and in 141 the Seleucid Demetrius II attempted to form a coalition against Parthia to recover that country, his principal ally being ‘the Bactrians’, which presumably means Heliocles.² It is possible that Demetrius II of Syria still claimed to be Heliocles’ overlord (p. 206); but if he did it was a claim which had long ceased to have any meaning, like *Fid. Def.* on the British coinage. What he and Heliocles were really attempting was to carry out the plan of Antiochus IV and Eucratides and crush Parthia by a combined attack on two fronts; but what might once have been a simple matter had become impossible, for Parthia was now far too strong and vital to be overthrown by kingdoms past their prime. Mithridates allowed Demetrius II to defeat his generals and recover Babylonia while he himself met Heliocles’ invasion of Parthia proper; he was in ‘Hyrcania’ in December 141 and mastered the situation there quickly enough to allow him to return to Babylonia and defeat and capture Demetrius late in 140 or early in 139. How serious Heliocles’ defeat may have been cannot be said, but it may be assumed that it did not do Bactria any good.
c10.s5.16
¹ See Tarn, *CAH* IX pp. 579-80, where the chronology used in this section is discussed.
c10.s5.17
² Justin XXXVI, I, 4.
c10.s5.18
Heliocles is usually supposed to have been the last Greek king of Bactria because it was his coins which the nomad conquerors continued to copy, and no doubt he was the last king who effectively ruled the country; but it is possible—the suggestion has been made—that Antialcidas, who succeeded him in the rule of his Indian provinces between the Hindu Kush and the Jhelum (p. 313), may for a time have had some connection with Bactria. Most of Antialcidas’ coinage, like that of every Greek king who reigned eastwards of the Hindu Kush, is bilingual, and the silver was struck on the reduced standard usual among the Greek rulers in India; but he struck one silver issue of good coins on the Attic weight-standard which are not bilingual but bear his Greek legend only, ‘Of King Antialcidas the Victorious’.³ It was not a Bactrian coinage; it was issued in India. But the Attic standard and the absence of a Kharoshthi legend may point to this coinage being intended for use in Bactria, and it might mean that, though Bactria and the mint there had been lost, he for a time retained some territory to the north of the Hindu Kush; the title νικηφόρος, ‘The Victorious’, might refer to some success against the nomad conquerors.¹ This silver issue is the last appearance of the Attic standard in the Farther East.
c10.s5.19
³ *BMC* p. 25 no. 1, βασιλέως νικηφόρου ᾿Αντιαλκίδου.
c10.s5.20
¹ It is no objection to this view that Apollodotus, who never ruled any part of Bactria, struck some silver on the Attic standard (BMC p. 34) which is shown by the humped bull on it to have been issued at Pushkālavatī; for this issue was bilingual while that of Antialcidas was not.
c10.s5.21
In 141 the curtain falls on Greek Bactria, to rise again in 128 upon new peoples and new names; somewhere between these two dates lies the end of the Greek kingdom. The latter date² is the year in which the Chinese Chang-k’ien, general and diplomat, was in Bactria, and it is his Report to his emperor which supplies such knowledge as we have of the country immediately after the conquest. It is our misfortune that that acute observer could not have visited Bactria fifty years earlier; in a small way he resembles Strabo, who also deals with a world—the world of the Hellenistic kingdoms—which had just passed away, but he had not Strabo’s resource of drawing copiously upon his predecessors, for he had none; to China he was the discoverer of a new world. Though this book is meant to be Greek history, it is pertinent to give a few pages to the conquest and the conquerors of Greek Bactria;³ the Greek evidence should help to straighten out the story.
c10.s5.22
² I follow the Chavannes-Hirth dating for Chang-k’ien; but the year he spent with the Yueh-chi has also been given as 127–6 and 126. Nothing turns on this for my purpose.
c10.s5.23
³ See Appendix 20 on the Chinese sources and the translations used.
c10.s5.24
Before coming to the details of that conquest one must just glance at the cause of the nomad invasion which swept away the Greek kingdom of Bactria and came near to overrunning Parthia also. It has been powerfully urged⁴ that most of the historical invasions of the settled lands of Asia by nomads from the Eurasian steppes, including the one to be described in this chapter, followed fixed laws and were due, in whole or in part, to a regular climatic period,⁵ an alternation of humidity and aridity which worked in 600-year cycles; the periodic increase in the fertility of the soil led to an increase in the nomadic population,¹ and the subsequent decrease of fertility and consequently of the food supply forced the hungry nomads, whose grazing grounds no longer sufficed them, to invade the settled lands. But in the case we have to consider, the movements of the various Saca peoples we shall meet (except the Massagetae) are fully accounted for by the disturbances due to the advent of the Yueh-chi. No doubt the sequel in India shows that the Saca world *was* over-populated, but had it been underpopulated the movements would have been the same; for the ultimate reason why the Sai-wang and the Sacaraucae—one may add the Yueh-chi themselves, for that matter—came south was that they were driven by superior force. The Massagetae alone appear to have been free agents; Bactria and Khorasan had attracted them from the time of Euthydemus (p. 117) and doubtless earlier, and what brought them south was not compulsion but opportunity, an opportunity which they did not mean to leave entirely to others.
c10.s5.25
⁴ A. J. Toynbee, A Study of History III (2nd ed. 1935) pp. 395–452. He frankly gives the objections to his view.
c10.s5.26
⁵ The theory worked out by Dr Elsworth Huntington in a number of studies, which has not escaped criticism; see the note by G. F. Hudson in Toynbee op. cit. pp. 453–4.
c10.s5.27
¹ On the ease with which a nomad population may increase, given favourable circumstances, see Prof. F. A. Lindemann in Toynbee *op. cit.* p. 436.
c10.s5.28
It would seem, then, that if the movement was really due to the operation of a climatic cycle, that cycle must have operated by starting the Yueh-chi on the long trek which was the cause of all the disturbance. But as the Yueh-chi and the Hiung-nu had long been neighbours, climatic changes which affected the one people should have affected the other; nevertheless the historical record (p. 276) is that the Yueh-chi had cared nothing for the Hiung-nu (*i.e.* were too strong for them) until in the final struggle they suffered a great defeat at their hands and were driven from their grazing grounds, which started the whole movement. The Hiung-nu followed this up by the creation of a regular empire, which might mean that *their* motive force was not so much hunger as imperialism; it looks, too, as if the cause of the final defeat of the Yueh-chi, previously the stronger horde, was not that the Hiung-nu were starving but that they had received a large accession of strength by incorporating some other people or peoples. I have neither the desire nor the knowledge to question either the fact or the potency of the climatic cycle, and the Saca peoples, who were not all nomads but included cultivators and oasis-dwellers, had obviously become too numerous for their lands and were ripe for trouble; I only wish to point out that the whole of the events to be described in this chapter could be accounted for quite simply by Hiung-nu imperialism, a foretaste of that of Chingis and his Mongols later, which would mean that the whole thing originated not in natural causes but in the mental processes of some individual man or group of men.
c10.s5.29
I must now turn to the particular consideration of the conquest of Bactria. To Chang-k’ien, that conquest was the work of a single great nomad horde, the Ta Yueh-chi. The Yueh-chi¹ (the name is still unexplained) first appear in history in Kan-su, in the north-west of China, where apparently they had been for some time; a struggle between them and another great horde, the Hiung-nu, usually supposed (though it has been doubted) to have been the people known later to the western world as Huns, culminated in 176 or 174 B.C. in their complete defeat, and they quitted Kan-su and set out westward. Part of the horde, called by Chinese writers the Little Yueh-chi (Siao Yueh-chi,² in contrast to the larger body, the Ta Yueh-chi or Great Yueh-chi), unable or unwilling to follow, turned southward into the Tarim valley and settled among the Ki’ang, apparently a general term for the border peoples of China in that region; it used to be thought that they formed two kingdoms there, Turfan and Kucha,³ but that may now be doubtful (see p. 289). The main horde, going westward, fell on the Wu-sun, killed their king, and must have attempted to occupy their grazing lands and been driven out again, presumably by the Hiung-nu. Still going westward, somewhere before 160 they attacked a people called Sai-wang⁴ about Lake Issyk Kul and the plain northward of the Alexandrovski range and attempted to occupy their lands; the Sai-wang, or some part of them, fled southward (pp. 277 sq.). But in or just before the year 160⁵ the Yueh-chi were again attacked by the son of the dead Wu-sun king with the help of the Hiung-nu and were driven out of the Sai-wang country, of which the Wu-sun took possession; the main body of the Yueh-chi again went westward, though some remained and were ultimately absorbed by the Wu-sun, as were also some Sai-wang elements who had not joined in the flight southward.¹ After 160 the Yueh-chi vanish from view for a generation—a fact which has not been noticed—till we meet them again shortly before 128; their steady progress westward was doubtless primarily to escape from the Hiung-nu, but it is noticeable that they were retracing the road which, on one theory, the Tochari section of the horde may have originally followed from Europe (p. 290).
c10.s5.30
¹ This story, primarily from Ssu-ma Ch’ien, is given, whether fully or otherwise, in many modern books. See especially E. Chavannes’ translation of Ssu-ma Ch’ien, *Les mémoires historiques de Se-ma Ts’ien* vol. 1, 1895, Introduction pp. lxx sqq.; Franke, p. 13 and *passim*; *CHI* pp. 565 sqq.; E. H. Minns, *Scythians and Greeks* pp. 92, 121 sq.; de Groot pp. 9, 16, 24; Herzfeld, *Sakastan* pp. 14 sqq.; Konow, *CII* pp. liii sqq. De Groot spells the name Goat-si.
c10.s5.31
² Hirth p. 97; Wylie p. 40. Notices collected by F. W. K. Müller, *Berlin SB* 1918 pp. 570-4.
c10.s5.32
³ Grousset pp. 210, 213.
c10.s5.33
⁴ I follow Konow in keeping the Chinese name Sai-wang (Franke’s view), literally ‘King of the Sai’, so as to beg no questions as to who they were. De Groot, p. 25, would prefer to render the name Sak-ke or Sik-ke (*i.e.* Sacas), but admits that Franke may be right; Franke in a note maintains his view. All I want myself is a word which will distinguish this Saca people from others.
c10.s5.34
⁵ This date, established by Franke p. 55, is now a fixed point.
c10.s5.35
¹ Wylie part 2 p. 84; de Groot p. 125.
c10.s5.36
What ultimately turned them southward is not known. The brief Chinese statement that after the defeat of 160 they went westward and occupied Ta-hia (Bactria)² is really two separate statements with a whole generation in time between them, the second statement merely giving the end of a lost chain of events. All we know is that these lost years were occupied in fighting and in attempts to settle somewhere; for in 128 they told Chang-k'ien that they were weary of trekking and fighting and only wanted to live in peace,³ and this does not refer to the story we know, for by 128 a new generation had grown up to whom the defeats of 174 and 160 were at most only memories of childhood. Sometime between 141 and 128—we shall see that it cannot have been much, if at all, before 130—they crossed the Jaxartes westward of Ferghana,⁴ went southward, and put an end to the Greek kingdom of Bactria; in 128 the horde was camped somewhere between Samarcand and the Oxus,⁵ not having yet moved across the Oxus into Bactria proper (Ta-hia), though it was theirs, *i.e.* tributary.
c10.s5.37
² Wylie p. 34; de Groot p. 87; cf. Hirth p. 97.
c10.s5.38
³ Hirth p. 94.
c10.s5.39
⁴ Wylie p. 40; de Groot p. 16; they passed Ta-yuan and attacked Ta-hia on the west. So Hirth p. 97; they crossed (the Jaxartes) to the west of Ta-yuan.
c10.s5.40
⁵ Hirth p. 97; they were living north of the Oxus and subsequently had their capital there.
c10.s5.41
The defeated Sai-wang, who are supposed to have been Sacas of some sort, are said in the Ch'ien-han-shu to have moved southward to Ki-pin,⁶ or ‘over the Hanging Pass’ to Ki-pin.⁷ It has been sufficiently shown that no horde with its flocks and herds could ever have crossed the Hanging Pass;¹ this is a mistake of Pan-ku’s own, who knew that it became later the most direct route for Chinese envoys and traders. Also, in the middle of the second century B.C., there was no such place or thing as Ki-pin (see App. 9): it was Pan-ku’s usual term for the later Saca realm in India, which did not come into existence on the Indus till somewhere about 110 B.C. and did not extend to northern India till somewhere about 80 B.C.² The statement, then, that at some date prior to 160 the Sai-wang went southward to Ki-pin is exactly like the parallel statement that in 160 the Yueh-chi went westward to the Ta-hia: it is two separate statements, of which the second represents the culmination of a long and a lost chain of events: the Sai-wang are, to us, lost to view for at least two generations, but some of them ultimately ended up in Ki-pin, which means that among the Sacas who subsequently invaded India were Sai-wang elements.
c10.s5.42
⁶ Wylie p. 34, ‘moved south and ruled over Ki-pin’; de Groot p. 87, ‘moved south and made themselves masters of Ke-pin’.
c10.s5.43
⁷ Wylie part 2 p. 84 (who translated it as ‘Hindu Kush’); Franke p. 58; de Groot p. 125. For the Chinese accounts of the Hanging Pass see de Groot pp. 72, 76-7.
c10.s5.44
¹ Herzfeld, *Sakastan* p. 20.
c10.s5.45
² See Chap. VI p. 233, Chap. VIII pp. 320 *sqq.*, and App. 16.
c10.s5.46
The flight of the Sai-wang southward (which probably from the lie of the land means south-westward) would take them across the Jaxartes, to Ferghana or Chodjend. But there was no vacant country south of that river for them to occupy; and as I understand that they are never again mentioned by name in Chinese literature³ they must have ceased to exist as a separate horde. Some might have joined a horde called K’ang-kiu, whose grazing grounds were the Tashkent plain;⁴ but those who subsequently went to Ki-pin must have joined the Sacaraucae⁵ (p. 291), who occupied the country about Chodjend and the steppes to the west of it, for it does not appear how otherwise they could have reached India. A horde (or a fraction of a horde) which joined another may sometimes have preserved its separate identity for a considerable time; this supposition is necessitated by (for example) the two components of the Yueh-chi horde (p. 286) and by Ammianus sometimes calling the Alans Massagetae after they had, at two removes, absorbed part of the latter people (p. 307).
c10.s5.47
³ Franke p. 61.
c10.s5.48
⁴ Franke p. 67. Chang-k’ien found the K’ang-kiu north (say, rather, north-west, as Herzfeld, *Sakastan* p. 22) of Ta-yuan (Ferghana); Hirth p. 95 §22.
c10.s5.49
⁵ A. Herrmann, *Sacaraucae* in PW, identified the Sai-wang with the Sacaraucae; often followed, but it will not work in that absolute form.
c10.s5.50
The remainder of the Sai-wang horde apparently seized the Greek province of Ferghana (p. 83 and App. 10), their name to Chinese writers becoming merged in that of the province, Ta-yuan. There they set up the Saca, or rather nomad, government which Chang-k’ien found in 128 (see p. 308); they are represented as distinct from the K’ang-kiu but on good terms with them.¹ It was easy at this time to occupy Ferghana: Eucratides had just overthrown the Euthydemid dynasty, he himself with his army was in India, and in 159 he met his death; the government in Bactria, whatever it was—possibly for a time Mithridates—can in the years after 160 have had no troops to spare for the north; in or just after 159 the Sai-wang could have set up their rule in Ferghana with impunity, and Heliocles, preoccupied first with the recovery of Bactria and then with the invasion of India, must have let this outlying province go. It is of course possible that Saca rule in Ferghana was not set up till c. 130; but it is impossible to see what body of Sacas could then have occupied it, while c. 159 it is easy. There may have been Sacas in Ferghana since Darius’ day (p. 475); but if so they had seemingly long ceased to be nomads, and what Chang-k’ien describes is a ruling stratum of nomads imposed upon an agricultural population.²
c10.s5.51
¹ Ssu-ma Ch’ien says that a little later (106-101 B.C.) the K’ang-kiu were their allies in their war with China; Hirth p. 113.
c10.s5.52
² There is a difficulty (discussed p. 308) in making out from Ssu-ma Ch’ien exactly what the political position was in Ferghana when Chang-k’ien saw it in 128; but Pan-ku deliberately corrects the difficulty, and as both historians had the same evidence (Chang-k’ien’s Report) I follow him.
c10.s5.53
We must now turn to Chang-k’ien.³ He was sent in 138 by the Han emperor, Wu-ti, as his envoy to the Yueh-chi to solicit their alliance against the common enemy, the Hiung-nu. Where the Yueh-chi were at the time, and what route Chang-k’ien took, are not recorded, but it is generally supposed that he followed the northern route through Chinese Turkestan to Kashgar; thence he would have taken the route by Irkishtam and the Terek pass to Ferghana (p. 84), which probably shows that he expected to find the Yueh-chi still north of the Jaxartes. On his way through Chinese Turkestan he was captured by the Hiung-nu and kept in more or less honourable captivity for some ten years; finally he escaped with his attendants and proceeded on his mission as though nothing had happened, a fact which illustrates the man’s force of character. He reached Ta-yuan (Ferghana); the Saca government passed him through to the K’ang-kiu and they in turn to the Yueh-chi,⁴ then camped between Samarcand and the Oxus. This shows that part of the K’ang-kiu, who had perhaps received an accession of strength from the Sai-wang, had, like the Yueh-chi, crossed the Jaxartes and now lapped Ferghana round on the north-west and west; probably the Yueh-chi had driven part of the horde across before them, for Chang-k’ien says that the K’ang-kiu to the south of the river were ‘under the political influence of’ (i.e. subject to) the Yueh-chi as those north of the river were to the Hiung-nu.¹ They may possibly have extended to the Samarcand country,² though if they did it was probably later. But the reason that Ta-yuan entrusted Chang-k’ien to their safe-conduct, which would mean for him a considerable détour to the westward, more probably was, not merely that they were vassals of the Yueh-chi, but that Samarcand was still maintaining itself in some sort of quasi-independence and blocking the direct road; the nomads had no chance of taking a properly fortified town by assault, and could only bring it to terms by ravaging its lands year by year till it agreed to pay tribute for them, as some cities in Ionia had bought off the Galatae. Chang-k’ien failed to obtain the alliance of the Yueh-chi, who told him that they were tired of fighting and trekking and only wanted a peaceful life in the rich country which they had at last secured, and returned to China by the more difficult southern route from Badakshan over the Pamirs and so through Chinese Turkestan; he was again captured by the Hiung-nu, but after a year’s captivity he reached China in 126. In 115 he was sent on a mission to the Wu-sun, then apparently about Lake Issyk Kul (pp. 276 *sq.*) and from there sent out subordinate envoys to visit the Western Countries up to and including Parthia, a country he himself never saw. He died in 114, a year after his return to China.
c10.s5.54
³ See generally the Life, Wylie pp. 66 sqq.; and Hirth p. 93. There is a good account of Chinese expansion in Asia under the Han in Grousset pp. 208 sqq.
c10.s5.55
⁴ Hirth p. 94 §9; Wylie p. 66.
c10.s5.56
¹ Hirth p. 96 §27, where ‘east’ must really be north or north-east; ‘east’ would be Ta-yuan and the Yueh-chi.
c10.s5.57
² Herzfeld, *Sakastan* p. 23, and his map facing p. 24 (following de Groot pp. 106–8, see p. 307 n. 4). But Herzfeld’s whole scheme is coloured by his mistake about Ta-yuan (App. 10), and there is nothing in the sources to show that *at this time* the K’ang-kiu were southward of Ferghana at all: the accounts both of Ssu-ma Ch’ien (Hirth p. 95 §20) and Pan-ku (Wylie p. 44; de Groot p. 109) represent Chang-k’ien as putting them *northward* of it, and this is a passage in which Pan-ku is correcting a mistake of Chang-k’ien’s on another geographical point (p. 477).
c10.s5.58
His Report to the emperor on the western countries which he discovered, and the use made of it by Ssu-ma Ch’ien and Pan-ku, are noticed in Appendix 20. But, though Ssu-ma Ch’ien professedly gives his Report, the question arises how far the Report so given is really Chang-k’ien’s own, and I have not succeeded in finding any study of this subject.¹ Pan-ku says that the subordinates whom Chang-k’ien sent out from the Wu-sun country in 115 did not get back to China till after his death;² Chang-k’ien’s Report therefore should only contain such information as he acquired in 128. But it is certain that the Report *as we have it* in Ssu-ma Ch’ien includes information obtained in 115. In 128 the great Saca invasion of Parthia was in full swing.³ But in the Report An-si (Parthia) is bounded on the north by the An-tsai (Aorsi)⁴ who lay round the Aral northward and eastward; this means that Parthia was suzerain of the whole Massagetae country up to the Aral and the lower Oxus, which was only the case from the time when Mithridates II (who became king in 124) had overcome the invasion to his death in 87 B.C. Again, in the Report Merv is Parthian;⁵ but in 128 it was in the hands of the Saca invaders, and was never Parthian till, as part of the liquidation of the invasion, it was taken for Mithridates II by the Suren who struck the ‘campaign’ coins (p. 499), again later than 124. Finally, the Report says that the people of Parthia wrote on parchment in horizontal lines;⁶ this looks like the observation of an eye-witness, but Chang-k’ien himself never saw Parthia. If then Pan-ku be right about Chang-k’ien’s lieutenants, Ssu-ma Ch’ien has supplemented the Report himself and included in it information obtained in 115. But it will be noticed that all the above instances relate to Parthia alone, which Chang-k’ien never visited but which by the time Ssu-ma Ch’ien is supposed to have written had become very important to China on account of the through trade. It does not follow therefore that Ssu-ma Ch'ien did more than supplement the An-si section; and the things which matter to Greek history, the accounts of Ta-hia and Ta-yuan, I have taken—I have found nothing to the contrary—to represent Chang-k'ien's own observation.
c10.s5.59
¹ Some Japanese scholar may have done one.
c10.s5.60
² Wylie p. 70.
c10.s5.61
³ Dates: Tarn, *SP Stud.* pp. 14 *sqq.*; *CAH* IX pp. 583 *sqq.*
c10.s5.62
⁴ Hirth p. 97 §37; see Tarn, *CAH* IX p. 585.
c10.s5.63
⁵ Hirth p. 97 §34, An-si is ‘close to the Oxus’ (‘extends to the Oxus’, Wylie p. 39; ‘lies on the Oxus’, de Groot p. 93); this can only mean that Merv was already Parthian. Whether the name An-si itself be Antioch (Merv), or (spelling it An-sik) Arsak, has been much discussed. All the countries in the Report are names of peoples, and there is therefore no reason for supposing that in one case a king’s name was used, especially as the Report knows nothing of the Parthians proper, the aristocracy; but it has not been noticed that the *Ch’ien-han-shu* is conclusive in favour of Antioch. It says that to reach An-si from O-ik-san-li (Seistan, see p. 347) you go north and then somewhat east (de Groot p. 92), which exactly describes the great road from Seistan by Herat to Merv; to Pan-ku therefore An-si was originally Antioch-Merv, Parthia being a derivative meaning. This is confirmed by Merv (Mou-lou) being called Little An-si in the *Hou-han-shu* (Chavannes, *T’oung Pao* VIII p. 177).
c10.s5.64
⁶ Hirth p. 97 §37; Wylie p. 40.
c10.s5.65
For what he saw himself his evidence is excellent, but the things he only heard are not necessarily in the same category. No doubt he could speak the Yueh-chi language, for it was apparently customary in China for officials destined for any country to be taught its language before starting;¹ but outside of the Yueh-chi he may have depended on such interpreters as they could provide (as the Wu-sun in 115 provided interpreters for his subordinates),² and they themselves were new arrivals in a strange world, as he was. At the same time, if the leading Yueh-chi clans spoke Saca, as is now the dominant belief (p. 287), then Chang-k'ien could himself more or less understand the peoples he met with; this is reflected in his statement that from Ferghana to Parthia, though there were dialectical variations, the people could all understand one another.³ His one real mistake was that sometimes, over the relative positions of peoples, he got the points of the compass wrong,⁴ a mistake easy enough to make in a new world; Pan-ku, by whose time a mass of information on the subject was available, occasionally corrects him.⁵ One thing however the Yueh-chi could not provide, and that was an interpreter for Greek: from first to last he never notes the existence of Greeks⁶ though he records their works, just as neither he nor his subordinates ever discovered the Parthian aristocracy. He distinguishes most carefully, as any Chinaman of experience would, between nomads like the Yueh-chi and the K'ang-kiu, who followed their flocks, and settled countries like Ta-yuan, Ta-hia, and An-si (Parthia), where the people had ‘fixed abodes’ and practised agriculture. The strange thing about his Report as we have it is that it never mentions the Sacaraucae and Massagetae and their invasion of Parthia; this was going on while he was among the Yueh-chi and he must have heard of it. But by the time Ssu-ma Ch'ien wrote it was ancient history; direct intercourse by caravan between China and Parthia was well established, and if Ssu-ma Ch'ien edited the An-si section he may have omitted some reference to the invasion as being now meaningless; he does not give the Report as a separate formal document but incorporates its substance in his own work. But one cannot analyse from translations as though they were Greek texts, and the work must be left to Chinese scholars.
c10.s5.66
¹ Wylie part 2 p. 89; de Groot pp. 127–8. But what Wylie translates as ‘Academic Institute’ de Groot calls a park full of strange animals. Would officials learning a language really be sent to the Zoological Gardens? ² Wylie p. 69; de Groot p. 27. ³ Hirth p. 108 §101. Cf. Eratosthenes in Strabo xv, 724. ⁴ If all the points of the compass given in Chang-k'ien's Report in Ssu-ma Ch'ien be taken out, it will be found impossible to make them all fit each other. ⁵ An important instance in App. 10. ⁶ It is conceivable that such as remained were bilingual, but not very likely as early as 128.
c10.s5.67
I have said that Chang-k'ien is quite clear that the conquest of the Ta-hia (Bactria proper) was the work of the Yueh-chi.¹ But almost every modern writer known to me attributes that conquest to 'Sacas' driven southward by the Yueh-chi, who are supposed to have occupied the country until the Yueh-chi expelled or subdued them. Chang-k'ien, who was there, knows nothing about this, and no scrap of evidence for it exists; it arose originally from a misunderstanding of a simple passage in Strabo, and for many years one writer has just copied it from another, till it has become an obsession; every form of 'Saca'—Sacaraucae, Sai-wang, even Tochari—has been pressed into service, and the theories to which this belief has given rise have done more than anything else to obscure the history of the time. Certainly Strabo says that the Sacas occupied Bactria;² but the most cursory perusal of the context shows that throughout the whole section he is talking, not of the second century B.C., but of a time long before that—he calls it Achaemenid, but it was really the seventh century—the time of the great Saca invasion, well known from Assyrian sources, which had played its part in the fall of Nineveh and had penetrated as far as Armenia and Cappadocian Pontus.³ An attempt has indeed been made since the theory was started to found the supposed Saca conquest of Greek Bactria a little more plausibly by citing a passage in Trogus, but as a matter of Latin Trogus' text will not bear the interpretation put upon it.¹ There is in fact no reason of any kind for thinking that Chang-k’ien was mistaken;² and whatever happened to outlying parts of the Bactrian kingdom, the supposed Saca conquest of Greek Bactria proper is a myth.
c10.s5.68
¹ Hirth p. 96 §29, and the parallel passage Wylie p. 41; de Groot p. 95.
c10.s5.69
² Strabo XI, 511, τὴν Βακτριανὴν κάτεσχον. The latest quotation of this theory is by Konow, CII p. xxi, who has made it the basis of his scheme in JIH XII, 1933, p. 1. Herrmann, Sakai in PW col. 1788, has been exceptional in recognising that Strabo refers to the seventh century B.C.
c10.s5.70
³ A good commentary on the Strabo passage is E. H. Minns, CAH III pp. 188 sqq. and especially 190. See also J. Junge, Z. f. Rassenkunde III, 1936, p. 68.
c10.s5.71
¹ *Prol.* XLI, Saraucae et Asiani Bactra occupavere et Sogdianos, which has been taken by Marquart (*Ērānšahr*, Gött. Abh. III, 1901, p. 205), Konow (*CII* p. xxii, *JIH* 1933 p. 6), and Herzfeld (*Sakastan* p. 26), to mean that the Saraucae occupied Bactria and the Asiani Sogdiana. A Latin writer who meant this would have said so; Trogus’ sentence, from its form, can only be a perfectly general statement. Herrmann, *Sakai* in PW col. 1617, properly rejected Marquart’s view.
c10.s5.72
² The movements of the Sai-wang have been already dealt with.
c10.s5.73
It is time to consider the Greek writers.³ Apollodorus attributes the conquest of the Bactrian kingdom to four nomad peoples, Asii, Pasiani, Tochari, and Sacarauli;⁴ ‘Trogus’ source’ formally attributes it to two, Asiani and Saraucae,⁵ though subsequently he mentions the Tochari (p. 286). Taking ‘Trogus’ source’ first, one of his two names must represent Chang-k’ien’s Yueh-chi; and as the Saraucae (Sacaraucae), of whom something is known (p. 291), are out of the question, the Yueh-chi are the Asiani. The form Asiani is an (Iranian) adjectival form of Apollodorus’ Asii, which is the substantival form; the Asii therefore are the Yueh-chi, whether (as some have supposed) the two words be identical, or not.⁶ From 1918 to 1936 it was further believed that a Central Asian text gave the name of a people Arsi who spoke toxrï (Tocharian) and who were the Greek Asii;⁷ the Arsi were accordingly supposed to be the Yueh-chi (or the dominant stratum of the Yueh-chi, which I shall come to) and much has been written about them. It has now been argued, with an impressive wealth of evidence, that Central Asian texts know no such people as the Arsi.⁸ Naturally I do not know myself if this will win general acceptance, but to avoid questions I shall use the name Asii, though in quoting from others I must give the form they use, Arsi.
c10.s5.74
³ I considered all these names in 1930 in *SP Stud.* §1, in the light of what was then known, but there have been many important developments since.
c10.s5.75
⁴ Strabo XI, 511. There can be no possible doubt that Apollodorus is Strabo’s source, though he is not named; there is no one else.
c10.s5.76
⁵ Trogus, *Prol.* XLI. Saraucae for MSS Sarancae is certain.
c10.s5.77
⁶ On the various attempts to explain the name Yueh-chi see Konow in *CII* pp. lviii sq.; P. Pelliot, *JA* 224, 1934, p. 25 n. 2.
c10.s5.78
⁷ E. Sieg, *Berlin SB* 1918 p. 560; F. W. K. Müller, *ib.* p. 578. A clear account of the history of the question is given by P. Pelliot, *JA* 224, 1934, p. 23.
c10.s5.79
⁸ H. W. Bailey, *Ttaugara*, *BSOS* VIII, 1936, pp. 883, 905 sqq. He makes ārsi = Sanscrit ārya. Pelliot, *T’oung Pao* XXXII, 1936, p. 265, finds the rejection of the name Arsi convincing.
c10.s5.80
It has, however, not been noticed that the Hellenistic world knew a people called Arsi, even if Central Asia did not. The name occurs in a very curious list of peoples in Pliny. In chapter 16 (18) of book VI (46 sqq.) Pliny describes the lands from Hyrcania to Bactria; he gives first Hyrcania, then Margiane, then ‘Caucasus’, which he describes as extending from Margiane to Bactria, inhabited by the Mardi (of the Elburz) and some unknown peoples; beyond (ultra) comes the list in question (48); then a city in Aria; then the Derbices (a Massagetae clan) on the Oxus, and some unknown names; finally Bactria. His own ideas of this part of the world were therefore very hazy; but he thought that he was going eastward, and by ultra he meant east of Margiane and ‘Caucasus’, though we should really have to translate it ‘somewhere in the East’. For the list which is ultra is not one of his usual strings of unknown names, but is a refuge for a number of well-known peoples from all over Asia whose names were in his note-books but whom he did not know where to place—the Matiani from Armenia, the Cadusii from Gilan, the Chorasmii from the Aral,¹ the Gandari from India (Gandhāra), the Sarangae from Seistan,² the Paricani from Ferghana or Gedrosia,³ and the Arsi from—where? The name proves nothing; but it suggests that he had got the native name of the Asii, and, not knowing where to place them, put them in his ‘refuge’ list, just as Ptolemy got the name Tochari in connection with Bactria and, not knowing what to do with them, made them a Bactrian tribe.⁴
c10.s5.81
¹ Or from Parthia proper (App. 11), if he took this name and Gandari from Herod. VII, 93, where the two are brigaded together.
c10.s5.82
² Given again twice over, as Drangae and as Zarangae, in their right place, VI, 94; the spelling Sarangae is from Herod. III, 93. Pliny has reproduced names from three different sources without suspecting their identity.
c10.s5.83
³ Gedrosia has been the usual view. Ferghana: Herzfeld, *Memoirs of the Arch. Survey of India* XXXIV, 1928, p. 6.
c10.s5.84
⁴ Ptol. VI, 11.
c10.s5.85
The name Asii, then, represents the Yueh-chi. But very many scholars, from Richthofen to Herzfeld, have held that the Tochari were the Yueh-chi;⁵ and to-day that is certain, from the now known occurrences of the name Tochari. It has been identified with the Thagouroi of Ptolemy on the Silk Route who had a city Thogara;⁶ the Chinese knew of remnants of the Togara in Kan-su in the second century B.C.;¹ the name occurs in or about Kan-su in Thibetan texts;² and the same name, ttaugara, for a town in Kan-su, is found in a document in Khotan Saca as late as A.D. 800.³ It follows that the Tochari, who as we know from Apollodorus came to Bactria in the second century from somewhere else, came from Kan-su; but the Chinese historians show that what did come from Kan-su to Bactria at that time were the Yueh-chi. Again, we know from Apollodorus and Pliny (fully given p. 84) that, while there were no Tochari in the Tarim basin in the reign of Euthydemus (died about 190 B.C.), there were Tochari there later at the date of some source of Pliny's; and here again Chinese historians show that part of the Yueh-chi (the Little Yueh-chi) came and settled there at some time not long after 174 B.C. (p. 276); and Indian writers called both the Great and the Little Yueh-chi by the same name, Tukhāra (Tochari).⁴ Finally Ptolemy, all unknowingly, locates Tochari in several places where the Yueh-chi are known to have been on their journey (see App. 21, p. 517), which alone would be conclusive. In fact, wherever we meet Yueh-chi from the Chinese side, we also meet Tochari, whether the evidence comes from texts in Greek, Chinese, Indian, Thibetan, or Saca; that suffices.
c10.s5.86
⁵ Müller and Sieg, *Berlin SB* 1916 p. 395, said that no one had ever doubted this. If only that were true (see pp. 295 sq. for the other theory about the Tochari) how much unnecessary trouble and confusion would have been saved.
c10.s5.87
⁶ Ptol. VI, 16, 5 and 8, from the Maes Titianus itinerary, early second century A.D. See F. W. Thomas, *JRAS* 1931 pp. 834-5; A. Herrmann, Θαγουροι in PW,
c10.s5.88
¹ Bailey op. cit. p. 885.
c10.s5.89
² Thomas loc. cit., see Bailey ib.
c10.s5.90
³ Bailey op. cit. p. 884.
c10.s5.91
⁴ Lévi, JA 1897 i p. 10, 1933 p. 26; de la Vallée-Poussin pp. 336 sq. One modern theory brings Kanishka from the Little Yueh-chi.
c10.s5.92
The Yueh-chi horde, therefore, was composed of two different peoples, who appear in Greek as Asii and Tochari. ‘Trogus’ source’ gives the relationship: reges Thocarorum Asiani—the Asii are lords of the Tochari.⁵ That is, at the time of the conquest of Bactria the Asii were the dominant or ruling people,⁶ however the horde may have originally taken shape. The idea occasionally put forward that the Asii conquered the Tochari after the conquest of Bactria may be summarily dismissed. Had Trogus meant this, he (or any other Latin writer) would have said so, and would not have used a phrase of unexampled ambiguity and clumsiness for its (supposed) purpose; while the fact that the Chinese had the Tochari name in the second century B.C., but
c10.s5.93
⁵ Trogus, Prol. XLII.
c10.s5.94
⁶ So Müller, Berlin SB 1918 p. 579: the Arsi were the ruling caste of the Tochari.
c10.s5.95
1934; Bailey op. cit. p. 885. See further App. 21. Both Thomas and Herrmann have noticed that the identity must follow.
c10.s5.96
nevertheless called the horde which migrated from Kan-su not Tochari but Yueh-chi, shows that that horde, before it quitted Kan-su, already contained something more than Tochari, *i.e.* that the combination of Asii and Tochari into one horde had already taken place.
c10.s5.97
The race and language or languages of the composite Yueh-chi horde have been the subject of much discussion, chiefly philological.¹ A few years ago the dominant belief was that the Asii (Arsi) were a Turki people,² but to-day it is thought that they were Iranian Sacas;³ I may perhaps venture the remark that Asia is getting very full of Sacas, even if Pliny (VI, 50) did call them ‘innumerable’. One must take the Kushans as representing the Asii or say rather the dominant element among the Yueh-chi,⁴ whether the Kushans were a tribe or sept or whether Kushāna was not a tribal name but a family or dynastic title;⁵ and it is a strong argument that when the Kushan language was first reduced to writing on its coins, it was apparently the north-Iranian dialect now called Saca and not the Sogdian dialect of Bactria.¹ But language and race are not identical things. To Greeks, who were not scientific ethnologists, the matter was simple: if a horde spoke Saca it was a Saca horde. It is not so simple to-day; the one assurance we have is that every great horde must have been a racial mixture, whatever might be the preponderating element. No philological arguments can make some of the Kushan portraits look like Iranians or any other ‘Indo-European’ people;² and I have felt trouble myself—no one else seems to have done so—over the number five of the five Yueh-chi princes (*yabghu*), for five is not a typically Iranian number and one cannot well separate the five princes of the Yueh-chi from the five of the Turki Hiung-nu. But five was a dominant number in China,³ where it occurs with the same monotonous regularity as does seven in astrology; and it may be that both hordes had merely borrowed from China,⁴ with which they had long been in contact, and had passed the number on to others.⁵ But this is a guess, and the number where it occurs *might* indicate a Turki custom and a Turki element; the word *yabghu* is Turki. Be this as it may, the right view about the Asii in Bactria, I think, is to say that they probably spoke Saca, but to refrain from dogmatising about their racial quality.
c10.s5.98
¹ For the literature see the discussions in *CII* pp. xlix *sqq.*; Grousset p. 213; de la Vallée-Poussin pp. 303–8.
c10.s5.99
² The reasons are collected in *CII* p. 1. There seems to be some evidence for Turki septs on the Jaxartes much earlier; see Kiessling, *Hunni* in *PW* col. 2599, and a communication by Nöldeke to Gutschmid (*Gesch. Irans* p. 2 n. 1) pointing out that the name Carthasis in Curt. VII, 7, 1 is Turki.
c10.s5.100
³ *CII* pp. li *sqq.* and references; Grousset *op. cit.* p. 59 and references; O. G. von Wesendonk, *Klio* XXVI, 1933, p. 337 (who connects the name Arsi with the Iranian stem aršan, man or hero, seen also in aršaka). I gather from de la Vallée-Poussin (pp. 301, 303), who has seen the MS, that this will be the view of F. W. Thomas in vol. II of *CHI*.
c10.s5.101
⁴ I need not consider a new theory (Konow, *JIH* XII, 1933, pp. 8–14) which makes of the five Yueh-chi princes (the Kushan chief being one) five Saca princes of Bactria conquered by the Yueh-chi. It is claimed that Pan-ku is ambiguous; on this ground the plain account in the *Hou-han-shu* is thrown overboard, though Fan-ye must have known, and, as the Chinese always call the Kushans Yueh-chi, it is said that they were mistaken. The theory is one more unhappy offshoot of the elementary blunder over Strabo which started the belief in a Saca conquest of Greek Bactria (p. 283). Another version of the same theory, open to precisely the same objections, makes the five princes Ta-hia = Tochari: P. Pelliot, *JA* 224, 1934, pp. 38 *sqq.*; on the impossible equation of Ta-hia and Tochari see p. 296.
c10.s5.102
⁵ A tribe or sept: J. Kennedy, *JRAS* 1912 p. 670; Baron A. von Staël-Holstein, *Berlin SB* 1914 p. 645, *JRAS* 1914 p. 79; A. Herrmann, *Sacaraucae* in *PW* col. 1613; E. J. Rapson in *CHI* p. 583; de la Vallée-Poussin p. 306; Konow in *CII* pp. xlvi, xlix *sq.*, though later he has called it a family name, *Ep. Ind.* XXI, 1931, p. 59. A title: J. F. Fleet, *JRAS* 1914 pp. 369–81, 413; F. W. Thomas, *JRAS* 1915 p. 532; B. Laufer, *Language of the Yueh-chi* 1917 (not seen). A personal name: K. P. Jayaswal, *JBORS* XVI, 1930, p. 256.
c10.s5.103
¹ Eratosthenes called Bactria and Sogdiana ὁμόγλωττοι παρὰ μικρόν, Strabo xv, 724. H. W. Bailey, *op. cit.* p. 893, referred ‘Bactria’ here to the Kushans, on the ground that Strabo wrote a century after the conquest of Bactria. The passage was in fact written by Eratosthenes about a century *before* the conquest.
c10.s5.104
² Note the strong resemblance between the portrait of Wima Kadphises, *BMC* Pl. XXV, 6, and those of Ephthalite Huns in Whitehead *NNM* Pl. XII, 1 and *ASI* 1915-16 Pl. XXVI nos. 46 to 50.
First 103 of 252 passages. Read on
Linked texts (5)
- Ἀλεξάνδρου ἀνάβασις — Λούκιος Φλάουιος Ἀρριανός, Ancient Greek
- Naturalis historia — C. Plinius Secundus, Latin
- Milindapañha — Pāli
- Γεωγραφικά — Στράβων, Ancient Greek
- Historiae Philippicae — Pompeius Trogus, Latin