The Greeks in Bactria and India c21: Appendix 9. Ki-pin (Kophen) and ‘Arachosia’

The Greeks in Bactria and India c21

Language: English

Appendix 9. Ki-pin (Kophen) and ‘Arachosia’

c21.2

The controversy whether the Chinese Ki-pin in the Kushan period meant Kapisa or Kashmir is now, or ought to be, ancient history; there is no reasonable doubt that the word meant the Kushan empire, which included Kapisa and Kashmir and much else; if you went to either place you went to Ki-pin.¹ But the Chinese had the word long before the Kushan period; and its earlier meaning has its bearing on the story of the Greeks in India.

c21.3

¹ Lévi concluded long ago (*JA* 1915 p. 102) that the Chinese mixed up Kapisa and Kashmir in their Ki-pin.

c21.4

It was seen long ago that the word Ki-pin imports a Greek place-name Kophen² or Kophene, obviously connected with the Kophen (Kabul) river. Cunningham, in the belief that this place-name does not actually occur in Greek (Kophene does not), gave the name Kophene, a properly formed eparchy name, to the district south of the lower Kabul river of which Purushapura (Peshawur) was the capital,³ thus making it one of the satrapies of Gandhāra; and in fact, though the names of several of the Gandhāra satrapies are known (p. 237), that of the Peshawur satrapy is missing (unless it were part of Peucelaitis and not a separate satrapy). But this cannot be right, for, though the province-name Kophene does not occur in Greek, a town-name Kophen does, and it cannot be Peshawur. Stephanus gives the name and identifies it with a town Arachosia ‘not far from the Massagetae’;⁴ his source for the identification is not given. By ‘Massagetae’ here he can only mean the people called Parsii (Pasiani of Apollodorus)⁵ who after the overthrow of the Bactrian kingdom had joined in the great Saca invasion of Parthia; his statement means that at the time of the invasion the Parsii were members of the Massagetae confederacy,⁶ for no other Massagetae had anything to do with Kophen. After the defeat of that invasion the Parsii had separated from the main body and had worked northward through Arachosia by the regular route which Alexander had once followed, by Ghazni to Kabul; this is clear from Ptolemy, who has preserved so much information from the Greek period in India (Chap. VI), and who, perhaps ultimately from ‘Trogus’ source’, records the Parsii by name as being in the Paropamisadae (which they could not have reached in any other way) and also gives two villages of theirs in that country, Parsia and Parsiana.⁸

c21.5

² Lassen, I² p. 29 n. 1; Gutschmid p. 109.

c21.6

³ *Geog.* pp. 38, 43, and see his Map 3.

c21.7

⁴ s.v. ’Αραχωσία: ’Αραχωσία, πόλις οὐκ ἄποθεν Μασσαγετῶν, ἥτις καὶ Κωφὴν ἐκαλεῖτο.

c21.8

⁵ For the Pasiani-Parsii see pp. 292 sqq.

c21.9

⁶ For another case of a particular member of the Massagetae confederacy being called ‘Massagetae’ compare Arr. IV, 16, 4 with Strabo XI, 513 (given p. 479 n. 1).

c21.10

⁷ Tarn, *SP Stud.* pp. 14 sqq.; *CAH* IX pp. 583 sqq.

c21.11

⁸ Ptol. VI, 18; Parsiana is the adjectival form, like Pasiani, p. 292.

c21.12

One must first try to be clear about this town Arachosia, for no serious examination of the difficult geography of the Arachosian satrapy has been made since Droysen. Stephanus’ town Arachosia is the town Arachosia of Pliny VI, 92 and the Arachotos of Ptolemy VI, 20, and is also ‘the Arachosians’ of Eratosthenes in Strabo XI, 514¹ and of Pliny VI, 62, two notices which both go back to Alexander’s bematists and give their distances; the name ‘the Arachosians’ indicates the Persian capital of the Arachosian satrapy. ‘The Arachosians’ has nothing to do with any Alexandria, or with Kandahar or Ghazni; we have the bematists’ figures, and Droysen² proved conclusively that ‘the Arachosians’ was somewhere in the neighbourhood of Kalat-i-Gilzai, a natural enough situation for the seat of the Persian satrap who had to govern both the plains of Seistan and the hill country to the north and east. But Alexander separated Seistan from the Arachosian satrapy and left that satrapy nothing but the hill country, identical with the later Parthian satrapy of Arachosia in Isidore 19; and ‘the Arachosians’ ceased to be a good centre for government. Alexander, therefore, who had already provided for Seistan by the foundation of Alexandria-Prophthasia (p. 14) on the Hamun Lake, proceeded to found, much farther to the north than Kalat-i-Gilzai, an Alexandria which was to be capital of the hill country; the foundation of this Alexandria is recorded in Arrian III, 28, 4 as made *before* Alexander crossed the Caucasus into the Kabul valley but when he was already in the hills and in a district where the people were ‘Indians’ (*ib.* III, 28, 1); without any doubt therefore it is now represented by Ghazni. This Alexandria is the μητρόπολις ’Αραχωσίας of Isidore 19, which Droysen rightly saw was Ghazni;³ Isidore, in calling the Parthian satrapy of Arachosia ‘White India’, recalls Arrian’s ‘Indians’. It is also Alexandria ‘among the Arachosians’, no. 12 in Stephanus’ list, and may also be no. 15, ‘beside the Arachosians (or “the Arachosians”) but bordering on India’.⁴

c21.13

¹ Stephanus *s.v.* ’Αραχωτοί, citing Strabo XI, *i.e.* 514, calls ‘the Arachosians’ a city of India, which can only mean Isidore’s ‘White India’ (p. 53). As Isidore’s phrase belongs to the time of the Azes dynasty, *c.* 30 B.C.–A.D. 19, Stephanus shows that the name Demetrias had been lost and the old name had come back by that date, which agrees roughly with the period when the name Demetrias in Sind went out of use, p. 257 n. 3. The rest of the notice in Stephanus, for which he also cites Strabo XI, εἰσι δὲ καὶ ἄλλοι (’Αραχωτοί) πλησίον Μασσαγετῶν, has (as ἄλλοι shows) nothing to do with ‘the Arachosians’ or the city Arachosia; it is merely the ’Αραχωτοί on the Oxus of Strabo XI, 513, an obvious error for Apasiacae, see p. 91 n. 3.

c21.14

² *Hellenismus* III, 2, pp. 217–20 (second German ed., 1877).

c21.15

³ *Ib.* III, 2, p. 720. Writers since Droysen have usually called it Kandahar, and I fear I have done this myself; but it will not work once it is looked into.

c21.16

⁴ Παρὰ τοῖς ’Αραχώτοις, ὁμοροῦσα τῇ ’Ινδικῇ. This may be a duplicate of Alexandria no. 12, ἐν ’Αραχώτοις, from another source, or perhaps a duplicate of no. 5, Alexandria (of the Caucasus) in Opiane.

c21.17

This shows that the corruption in Isidore 18 and 19 goes deeper than has been supposed;¹ the name of the metropolis of Arachosia in 19 must be ’Αλεξάνδρεια, not ’Αλεξανδρόπολις, and ’Αλεξάνδρεια should not come in 18 at all but only ’Αλεξανδρόπολις, near (πλησίον) Sacastene (not in Sacastene, as Droysen loc. cit. and Tscherikower, p. 103 give it). Alexandropolis near Sacastene can only be Kandahar, and this settles the question of its foundation, for a place called Alexandropolis cannot have been a city founded by Alexander (p. 7), and in fact there is no record that he founded Kandahar; Alexandropolis at best was a military colony which (possibly quite correctly) attributed its settlement to him. I need not consider the conflicting opinions about the name Kandahar, whether it be derived from Alexander (Iskandr), from Gandhāra (for which there seems no historical reason), or (most probably) from the Parthian Gondophares (Gundofarr). It must however be supposed, for reasons given elsewhere (p. 94), that the old Persian capital, ‘the Arachosians’, was the point at (or representing) which Demetrius founded his name-city Demetrias.

c21.18

¹ In 18 (Sacastene) he has καὶ πλησίον ’Αλεξάνδρεια πόλις καὶ πλησίον ’Αλεξαν-δρόπολις πόλις, and in 19 (Arachosia) he has ’Αλεξανδρόπολις, μητρόπολις ’Αραχωσίας. It has long been recognised that one of the two names in 18 must go out; and Arrian III, 28, 4 makes it certain that the name in 19 (Isidore is reproducing an official survey) should be ’Αλεξάνδρεια. ’Αλεξάνδρεια therefore has been transferred from 19 to 18, and must go out of 18 and back to its place in 19, which has been filled by ’Αλεξανδρόπολις, taken from 18. The real reading then of 18 is καὶ πλησίον ’Αλεξανδρόπολις πόλις and of 19 ’Αλεξάνδρεια, μητρόπολις ’Αραχωσίας.

c21.19

Stephanus’ notice of the town Arachosia and the Massagetae is therefore absolutely right. But his further identification of the town Arachosia with Kophen cannot be right, for Kophen cannot be separated from the Kophen (Kabul) river; what the identification does show is that there was some connection between the town Arachosia and Kophen, and this is shown again by Pliny VI, 94, who makes the Kophen a river of Arachosia, though he places on it a town with an Indian name.² Returning to the bematists’ route from Seistan to the Paropamisadae via Ghazni, there is no doubt that Ortospana, as universally believed, is Kabul,³ this being its Iranian name;⁴ and as the town Kophen has to be connected both with the town Arachosia and the Kophen river there is no place which it can be except Kabul, for there is no other place of importance on the route Seistan—Arachosia (Demetrias)—Ghazni (Alexandria)—Kabul (Ortospana)—τρίοδος (Alexandria) whose Greek name is missing.⁵ Kophen therefore is Ortospana.

c21.20

² Condigramma, which is Sanscrit -grāma; see other cases p. 244.

c21.21

³ Pliny VI, 61 (from the bematists), fifty Roman miles from Alexandria, i.e. from the τρίοδος, a fixed point (see App. 6); it is near enough. On the right reading of Strabo XI, 514, see p. 461: the distance between Ortospana and the τρίοδος has fallen out.

c21.22

⁴ Ptolemy has similar Iranian names in VI, 8, 13, Πορτόσπανα in Carmania, and VI, 4, 4, Πορύσπανα in Persis.

c21.23

⁵ I cannot use Hsüan Tsiang’s Hu-pi-na as evidence for the name Kophen, for opinion has always been divided as to whether Hu-pi-na was Kabul or whether

c21.24

Kabul, and Kophene, though the actual word does not occur, must have been the Greek name of the Kabul eparchy of the Paropamisadae satrapy under the Seleucids; several eparchy names in that satrapy are known (pp. 96 *sq.*), but the Kabul name is missing. Stephanus’ mistaken identification of the town Arachosia with Kophen (Kabul) arose therefore, not merely from the fact that the two towns were stations on the same great route from Seistan to Alexandria-Kapisa, but from the fact that in Demetrius’ reign both were for a time under the same sub-kings (Pantaleon, Agathocles, see pp. 157 *sq.*) and both subsequently formed part of the kingdom of the Parsii, a kingdom known to some source of Ptolemy (presumably ‘Trogus’ source’), if his Gazaka or Gauzaka¹ be the native name of Alexandria-Ghazni;² for he puts this place in the Paropamisadae, which (as Ghazni was certainly in Arachosia) here means the kingdom ruled from Kabul.

c21.25

¹ Ptol. VI, 18, 4, Γάζακα ἢ Γαύζακα.

c21.26

² E. Benveniste, *JA* 226, 1935, p. 143, reading Γάνζακα. This town has not the same co-ordinates as Alexandria (Ghazni), VI, 20, 4, but that means little; Ptolemy’s habit of making two places out of one is notorious (pp. 231 *sq.*).

c21.27

It can now be seen what happened. After the Parthians, somewhere between 124 and 115, broke the Saca invasion, the invaders, who comprised more than one people, entered India not by one route but by two. The movements of the main body, who ultimately came up the Indus to Taxila, are noticed elsewhere (Chap. VIII). But another body, the Parsii, went north from Sacastene through Arachosia by Alexander’s route—we have been looking at the record—and ultimately took Kophen-Kabul from the Greeks. From their first contact with China dates the origin of the name Ki-pin.

c21.28

Ki-pin was originally Kophen-Kabul.³ The name Ki-pin is unknown to Chang-k’ien and also to Ssu-ma Ch’ien, who is supposed to have finished the Shi-ki about 99 B.C. or somewhat later⁴ (Chap. CXXIII, which alone concerns us, contains information of the year 101); the name had therefore not reached China by *c.* 100 B.C. It first occurs in the Ch’ien-han-shu, written by Pan-ku late in the first century A.D. That work states that the first relations of China with Ki-pin (i.e. China’s first attempt to reach any land south of the Hindu Kush) began in the reign of the Emperor Wu-ti,⁵ who died in 87 B.C.; Kabul therefore was in Saca hands before that date, and the natural supposition is that the Sacas took it when the kingdom of Antialcidas, who must have been a king of some power, broke up on his death (p. 315); incidentally this furnishes another reason for supposing that he was not alive

c21.29

³ Lassen suggested Ki-pin might be Kabul, but it has never been worked out; Wylie pp. 33-4 assumed it.

c21.30

⁴ Hirth p. 91.

c21.31

⁵ Wylie p. 36; Franke p. 63; de Groot p. 88.

c21.32

it represented Opian, and there can be no certainty. See the discussion in T. Watters, *On Yuan Chwang’s travels in India* II, 1905, p. 266, with the itinerary p. 342 of V. A. Smith, who preferred Opian, and add to Watters’ references Wylie p. 34 n., who preferred Kabul.

c21.33

as late as 90 B.C. (p. 314). Wu-ti’s communication was certainly not with Antialcidas himself: the Chinese could never have been given, or obtained, the name Kophen as the name of Antialcidas’ extensive realm, for to the Greeks Kabul was of very secondary importance. At the same time, the great Saca state in northern India created by Maues had not formed by 87 (Chap. VIII); Wu-ti’s Ki-pin therefore was Kabul and whatever the Sacas (Parsii) of Kabul ruled at the time, which may have been a large part of the Paropamisadae (p. 332). This too is the meaning of Ki-pin in the story of Yin-mo-fu given in the Ch’ien-han-shu (set out pp. 339 sq.), which Pan-ku must have taken from Wen-chung’s Report, no other source being possible. But shortly after this episode communication with the Paropamisadae was abandoned by the Emperor Yuan-ti (48–33 B.C.), and a memorandum drawn up by Tu-k’in for the next Han Emperor Ching-ti (32–7 B.C.) said that it was impossible to reach that country effectively and that it was best to leave it alone (p. 350); Wen-chung was probably the last Chinese official to visit it.

c21.34

The name of this the first Saca state south of the Hindu Kush with which the Chinese had become acquainted was transferred, by a natural process, to the great Saca realm which soon afterwards absorbed it, and that is what it meant to Pan-ku himself, writing about a century later than that absorption; probably he did not know what Wen-Chung’s Ki-pin was, but he transcribed faithfully and gave us the chance of knowing. To Pan-ku personally Ki-pin meant the Saca realm in India;¹ later writers transferred the name to the realm of the Kushans who ultimately replaced the Sacas. It has been truly said that the Chinese never got clear ideas about this great kingdom of ‘Ki-pin’ in Saca times.² Pan-ku did not follow his usual rule of giving the numbers of families, of inhabitants, and of warriors, but merely said that the numbers were those of a great kingdom, his formula for states of which he knew little, and his own formal account of Ki-pin³ is a mixture of elements drawn from different parts of it and not necessarily all trustworthy; for its alleged currency⁴ numismatists have sought, and will seek, in vain.

c21.35

¹ So Herzfeld, *Sakastan*, p. 32 (cf. p. 35): Ki-pin must be treated as ‘die politische Einheit des Sakareichs’. Pan-ku, having transferred the Ki-pin name of Kabul to the Saca kingdom, introduced a new name for Kabul (Ko-fu, Kao-phou, Ko-hu) and wrongly made it the domain of one of the five *yabghus* of the Yueh-chi; the author of the *Hou-han-shu* pointed out the error (see de Groot pp. 101-2). The evidence given in this Appendix and in Chapter VIII shows that the *Hou-han-shu* is right.

Linked texts (3)

Naturalis historia — C. Plinius Secundus, Latin
Ἀλεξάνδρου ἀνάβασις — Λούκιος Φλάουιος Ἀρριανός, Ancient Greek
Corpus Inscriptionum Indicarum