Bactrian Documents from Ancient Afghanistan

Professor Nicholas Sims-Williams of the School of Oriental and African Studies (SOAS), University of London, who is the leading expert of the Sogdian and Bactrian languages, gave a lecture on the discovery and decipherment of Bactrian documents, written in the little-known Iranian language of Ancient Afghanistan in modified Greek script, at the Ancient Orient Museum in Ikebukuro, Tokyo, on September 23.


Prior to the talk he was awarded, jointly with his colleague Dr Joe Cribb of the British Museum, this year's Hirayama prize from Professor HIRAYAMA Ikuo, Director of the Institute of Silk Road Studies in Kamakura for their work on the decipherment and interpretation of the newly discovered Rabatak inscription in Northern Afghanistan.

The text of the lecture follows:


New Findings in Ancient Afghanistan

--- the Bactrian documents discovered from the Northern Hindu-Kush

Prof. Nicholas Sims-Williams (University of London)

Bactrian, the ancient language of Bactria in northern Afghanistan, is unique among the Iranian languages in being written by means of the Greek alphabet --- a legacy of the conquest of Bactria by Alexander the Great in the 4th cent. B.C. From this period onwards the Greek language, written in the Greek script, was for a long time the exclusive language of culture and administration in Bactria. When Bactria was overrun by nomadic peoples from the north, its new rulers, the Kushans, at first continued the use of the Greek language for administrative purposes, but soon they came to use the Greek script to write the local language, Bactrian. A crucial moment in the history of this language was the decision of the Kushan ruler Kanishka to adopt Bactrian as the language of his coinage. After the first issues of Kanishka, Greek disappears from the coinage once and for all, to be replaced by Bactrian.

During the first centuries of the Christian era, Bactrian could legitimately have been ranked amongst the world's most important languages. As the language of the Kushan kings, Bactrian must have been widely known throughout a great empire, in Afghanistan, Northern India and part of Central Asia. Even after the collapse of the Kushan empire, Bactrian continued in use for at least six centuries, as is shown by the ninth-century inscriptions from the Tochi valley in Pakistan [Slide 1 9KB] and the remnants of Buddhist and Manichean manuscripts found as far away as the Turfan oasis in western China. (This slide, for instance [Slide 2 12KB], shows the unique fragment of a Bactrian text written in Manichean script, which forms part of the Turfan collection in Berlin.) The career of Bactrian as a language of culture thus lasted for close to a thousand years.

Until forty years ago virtually nothing was known of the Bactrian language except for the legends on the coins of the Kushans and their successors. The Kushan coins are inscribed in Greek letters of an angular type, apparently imitating a style of writing used for monumental inscriptions. In principle these legends are not particularly difficult to read, but their content is limited to the names and titles of kings and deities. The coins of the later rulers of Bactria --- Kushano-Sasanians, Kidarites, Hephthalites, Turks, and so on --- are written in a cursive script, imitating manuscript styles, which has proved much more difficult to decipher. Some tiny scraps of manuscripts in a similar cursive script were also known, but they were too few and too incomplete to offer any realistic prospect of interpretation.

These prospects were transformed in 1957 by the discovery at Surkh Kotal near Baghlan of the first substantial Bactrian inscription [Slide 3 23KB]. The text, written in the monumental script already known from the Kushan coins, could be read without much difficulty; its interpretation was much more problematic, since the names and titles known from the coins provided only a minimal vocabulary and hardly a hint of the grammatical structure of the language. Nevertheless, the essential points were immediately recognized by W. B. Henning: the text refers to the foundation of a sanctuary by the emperor Kanishka, its abandonment as a result of problems with the water-supply, and its re-establishment by a high official named Nukunzuk in the year 31 of the era of Kanishka, that is, early in the reign of his successor Huvishka.

Several further Bactrian inscriptions have been discovered since that of Surkh Kotal, but most of them are too poorly preserved to add significantly to our knowledge of the language. However, in 1993 a new inscription of fundamental importance was discovered by chance at a site named Rabatak, not far from Surkh Kotal [Slide 4 18KB]. The inscription of Rabatak describes events of the first year of Kanishka in words strikingly reminiscent of those of Darius the Great in the inscription of Bisitun. Since Joe Cribb and I have already published a detailed study of this inscription in the most recent issue of Silk Road Art and Archaeology, I will limit my remarks to a few of the most important points.

The opening lines refer to Kanishka as "the great salvation, the righteous, just autocrat, worthy of divine worship, who has obtained the kingship from Nana and from all the gods, who has inaugurated the year one as the gods pleased". Then comes the significant statement: "He issued(?) an edict(?) in Greek and then he put it into the Aryan language". In principle, any of the Indo-European languages of Iran or India could be called "Aryan"; but when Kanishka refers to "the Aryan language" he surely means Bactrian, the language of this inscription, just as Darius meant Old Persian, the language of his inscription, when he wrote: "By the grace of Auramazda, I made another text in Aryan, which previously did not exist". It is difficult not to associate Kanishka's emphasis here on the use of the "Aryan language" with the replacement of Greek by Bactrian on his coinage. The numismatic evidence shows that this must have taken place very early in Kanishka's reign, quite possibly in his very first year.

Lines 4-7 of the Rabatak inscription give a list of the chief cities of north India which were controlled by Kanishka. Four of the five names can be identified: Saketa, Kausambi, Pataliputra, and Champa. [Slide 5 12KB] The wording of the inscription does not make it clear whether Champa is mentioned as belonging to the area ruled by Kanishka or as the first city beyond his eastern border. Even in the latter case, the statement that he ruled northern India as far as Pataliputra is sufficiently striking.

The major part of the inscription concerns the foundation of a temple, perhaps at Rabatak itself, which seems to have been an extensive site. Lines 9 and 10 name the divinities who are to be worshipped in the temple. This list is very intriguing. On the one hand it includes two Zoroastrian deities who are never portrayed on the Kushan coinage. On the other hand, it omits many names which are well attested on these same coins, such as Ma, the moon, and Ardukhsh, the goddess of plenty. Above the list of Iranian divinities some words have been added in smaller letters [Slide 6 22KB], which seem to identify some or all of them with Indian equivalents such as Mahasena and Visakha.

Apparently the temple was intended to contain statues of kings as well as gods. Kanishka lists four kings: Kujula Kadphises his great- grandfather, Vima Taktu his grandfather, Vima Kadphises his father, and himself, Kanishka. This list is extremely informative. In the first place, it bears witness to the existence of two kings named Vima, rather than one. Several inscriptions previously attributed to Vima Kadphises, notably the Bactrian inscription of Dasht-e Nawur [Slide 7 9KB], can now be ascribed to his father Vima Taktu. In all probability the coins of the anonymous king Soter Megas "the great saviour", which come between Kujula and Vima Kadphises in the numismatic sequence, should also be attributed to this newly-discovered Vima the First. Moreover, the indication that Kujula Kadphises was the great-grandfather of Kanishka evidently has a bearing on the oft-debated issue of the date of Kanishka. The fact that Kanishka belongs to the third generation after Kujula clearly imposes certain limits on the manner in which the early chronology of the Kushans may be reconstructed. Although I would not go so far as to say that the new facts are only compatible with a single chronological system, it is clear that the solutions previously proposed will now have to be reconsidered.

Lines 14-17 mention the officials charged with executing the orders of Kanishka. Amongst them is a certain Nukunzuk who is probably the same person who was later responsible for the works described in the Surkh Kotal inscription. At Rabatak, 30 years earlier, he does not yet bear the title karalrang "margrave" and seems to occupy a subordinate position. The last part of the inscription in which a continuous text can be read contains a wish for Kanishka's health and happiness and perhaps also --- if my reading is correct --- expresses the hope that his reign may last for a thousand years.

It goes without saying that neither the reign of Kanishka nor that of the Kushan dynasty approached 1000 years. In about 2Z4 A.D. the Sasanians came to power in Iran [Slide 8 9KB]. Within a few years the Sasanians had also conquered Bactria, which they ruled during part of the following period through a viceroy known as the Kushan-shah 'king of the Kushans", who was often a prince of the Sasanian royal family. Subsequently Bactria was invaded several times by nomads from the north. At different times the invaders are referred to under various names --- Chionites, Kidarite Huns, Hephthalites --- though it is not entirely clear whether all these names refer to the same or to different peoples. The next arrivals were the Turks, who in the middle of the sixth century allied themselves with the Sasanians to defeat the Hephthalites; and then finally all the local dynasties were swept aside by the coming of Islam and the Arabs.

The history of the period from the Sasanian Kushan-shahs to the arrival of the Arabs is illustrated by a second discovery. In December 1991, I was shown photographs of a newly discovered Bactrian document on leather. The document was inscribed on both sides with 28 lines in cursive Bactrian script, making it by far the most substantial example of cursive Bactrian so far known [Slide 9 12KB]. The document was clearly a letter, beginning with conventional phrases of address and greeting almost identical to those used in Sogdian letters: "To your lordship 1,000 and 10,000 times greeting and homage from so-and-so your servant. Having heard that your lordship is healthy I am happy; but I should be still happier if I myself might see your lordship in good health and pay homage ...". To judge by personal names such as Ohrmuzd and Khwasraw, the document belonged to the Sasanian period. Another striking name was Purlang-zin, evidently meaning "the man with the panther's skin" --- a clear reference to the zin-e palang of Rustam, one of the heroes of the Persian epic.

One such document was a revelation in itself. But it was as nothing compared to what was to come. Within five years the corpus of Bactrian documents had grown to a hundred, most of which are now in London, in the collection of Dr David Khalili. These documents have passed through the hands of many different dealers and collectors. In most cases there is no record of their original provenance, though a couple of them are said, quite plausibly, to have been found in Samingan. From internal evidence, especially the recurrence of the same names in several documents, it seems clear that most if not all of them ultimately derive from a single source.

Many of the documents are letters, some of them still sealed and therefore perfectly preserved. This slide [Slide 10 4KB] shows a letter sealed with a clay bulla, with a few words of address written on the outside; and here [Slide 11 10KB] is the same letter after opening, from which you can see the standard layout of the text, with a wide left-hand margin, and the way the seal is attached to a strip of leather cut along the bottom edge of the letter without being completely detached from it.

One of the less well-preserved letters is particularly interesting because it mentions a Kushan-shah [Slide 12 12KB]. This letter can hardly be later than the latter half of the fourth century, when the rule of the Kushan-shahs came to an end. The Kushan-shah here seems to be named as Warahran, though the reading of the name is not quite clear [Slide 13 10KB]. Since Warahran (or Bahram) was the name of the last Kushan-shah --- or the last two Kushan-shahs --- known from the coins, the letter probably belongs to the very end of the Kushano-Sasanian period. The sender of the letter was the daughter of a princess named Dukht-anosh, a Middle Persian name which is attested on a seal in Paris. The contents are not yet clear, but concern a eunuch with the remarkable name Dathsh-mareg "servant of the Creator" --- a compound of the Bactrian mareg "servant" and the Avestan genitive dathusho "of the Creator", which was probably used in the Zoroastrian calendar of Bactria to refer to a day dedicated to the Creator.

The letter illustrated on the next slide [Slide 14 9KB] is important for two reasons. Firstly, it is written by a representative of the Shahan-shah or "king of kings", the ruler of Iran, and must therefore belong to one of the periods when Bactria was under Sasanian control. Secondly, it is dated, though the era is not specified.

It seems very likely that the era used here is the same as that of the Bactrian inscriptions of the Tochi valley in Pakistan [Slide 15 11KB]. These inscriptions, written in Arabic, Sanskrit, and Bactrian, contain dates expressed in three different eras. The vital evidence for determining the starting-point of the Bactrian era is provided by two bilingual inscriptions, the first being in Arabic and Sanskrit [Slide 16 13KB]. The Arabic text is naturally dated by means of the Hijri calendar, which is blessedly unambiguous and enables one to complete the date of the Sanskrit version, which is expressed in an abbreviated form omitting the thousands and hundreds. The second bilingual is in Sanskrit and Bactrian. Again the Sanskrit date is abbreviated; but, if one assumes that this inscription belongs to the same century as the Arabic-Sanskrit bilingual, the missing figures can be reconstructed, giving a date which can be computed as 863 A.D. The Bactrian version is dated with Greek numerals. Helmut Humbach, the first editor of the Bactrian texts, read these numerals as 632. On the basis of newly-discovered materials I would interpret the last digit as 1 rather than 2, but a difference of one year is of minor significance. In either case, the Bactrian date indicates the existence of an era beginning early in the Sasanian period, in 232 (according to Humbach) or 233 A.D. (accord to me). I am inclined to follow Humbach in regarding this as a "Kushano-Sasanian" era, whose starting-point was probably the Sasanian conquest of the Kushan empire.

Let us return to the Bactrian letter, whose date can be read as 239. If this era began in 233 A.D., the year 239 will correspond to 471 A.D., during the rule of Peroz, who spent much of his reign fighting the Hephthalites and who eventually perished in battle against them. Is it a coincidence that the word preceding the title Shahan-shah in the letter is in fact Piroz? Unfortunately the Middle Persian word peroz is not only a personal name but also a common adjective meaning "victorious", so that it is difficult to decide whether one should translate "Peroz, the king of kings" or merely "the victorious king of kings".

Only two of the newly-discovered Bactrian documents seem to be religious in inspiration. Both of them are Buddhist texts containing invocations to Buddhas, bodhisattvas, and so on. The one illustrated on this slide [Slide 17 16KB] is particularly interesting. Unusually, it is written on cloth rather than leather. As you see, it is illustrated with two drawings, perhaps representing a Buddha and an ascetic. The text begins with homage to "all the Buddhas" and to five or six named Buddhas ending with the historical Buddha Shakyamuni. Then homage is paid to a series of six Bodhisattvas, followed by the kings of the yakshas, rakshasas, kinnaras, nagas, pishacas, etc. and finally Shakra, the king of the gods, and the "great Brahma". I have not succeeded in deciphering all the names, but the outline of the text is quite clear up to this point. The last three lines are more obscure, but contain a reference to a Buddhist monastery (vihara} and a temple.

Apart from a few tears and small holes, some of which appear to have been made deliberately, the document seems to be complete. I suppose it to be a kind of votive offering, which would have been tied to a pole like a banner and fixed in a holy place, or perhaps an amulet.

After the letters, the largest group of texts consists of legal contracts and similar documents: deeds of sale, leases, guarantees, receipts, and deeds of gift. One particularly interesting document --- in fact the earliest of the dated documents --- is a marriage contract, in which one woman engages herself to marry two brothers at the same time [Slide 18 26KB]. The practice of polyandry, discussed by Prof. Enoki Kazuo in his well-known article "On the nationality of the Hephthalites" and here confirmed by first-hand evidence, was apparently typical of the region. Another unusual document [Slide 19 20KB], which now belongs to the Institute of Silk-Road Studies here in Kamakura, is a deed of manumission, recording the freeing of a slave in return for the purchase of a substitute.

Unlike the letters and Buddhist texts, legal documents such as these are always dated. So far I know of more than twenty documents with dates ranging from 110 to 549, that is (if we assume that the Bactrian era began in 233 A.D.), from 342 to 781 A.D. This span of more than four centuries covers the Chionite, Kidarite, Hephthalite, and Turkish periods, and extends well into Islamic times. As we shall shortly see, the contents of the documents provide many details against which one can test this chronological framework.

For example, the next slide [Slide 20 12KB] shows a contract for the sale of land dated in the year 295, which I interpret as 527 A.D., during the period of Hephthalite domination. This agrees well with the statement of the text that the "Hephthalite tax" due on the property has been paid. The format of the document is typical, though this example is exceptionally well preserved. There are two complete copies of the text. One copy is left open to be read. The other is tightly rolled, tied with string, and sealed with five bullae. The first two bear impressions of the fingernails of the vendors; the others are impressed with the seals of three witnesses. Presumably it was intended that the sealed copy should be opened in the presence of a judge in case of a dispute. On the reverse of the document [Slide 21 5KB], the names of the vendors and witnesses are written beside the holes for the seal-strings.

In addition to dates, many of these legal documents contain place-names, including the names of the places where they were written [Slide 22 10KB]. Several documents state that they were written in Samingan, Rob (modern Ruy), Malr or Madr, or Kah (modern Kah-mard). All four places are apparently within the jurisdiction of a ruler who is referred to in many documents as "the khar of Rob". On the other hand, Tarmid (or Termez), to the north of the Oxus, and Bamiyan, which is separated from Kah and Madr by a considerable ridge of mountains, may well have been outside his kingdom. The khar of Rob is no doubt to be identified with the Ru'b-khan, the ruler of Ru'b and Siminjan, who helped Qutayba b. Muslim to defeat the Hephthalite rebel Nezak Tarkhan in the year 91 of the Hijra (710 A.D.), as mentioned by the historian Tabari.

I suspect that the title khar is an Iranian --- but not necessarily Bactrian --- dialect form derived from Old Iranian *xshathriya- "ruler" [Slide 23 11KB]. The true Bactrian form may be sher, which is mentioned by Muslim writers as the title of the rulers of Bamiyan, Gharchistan, and other places in the area around ancient Bactria. The ruler of Rob may at times have controlled a wider area than is indicated by the place-names mentioned so far. For instance, in a letter which probably dates from about 480, the khar of Rob is addressed, somewhat bombastically, as the "Hephthalite yabghu, ... scribe of the Hephthalite lords, and judge of Tukharistan and Gharchistan". Tukharistan is the land to the north of the Hindu-kush, including Rob and Samingan, but covering a considerably wider area; Gharchistan usually refers to the mountainous area to the west of Bamiyan, but since the Bactrian form of the name indicates that it means merely "the land of the mountaineers" it is possible that it refers more generally to the mountain areas to the south of Tukharistan.

Yet another document is written in a place named Warnu. This is surely to be identified with the Aornos; mentioned by Arrian as one of the two chief cities of Bactria. According to Paul Bernard and others, Aornos is to be located near Khulm or Tashkurgan, where the valley of the Khulm river, in which Rob and Samingan are situated, opens out into the plain. Since the whole collection of Bactrian documents almost certainly represents the royal archive of the kings of Rob, the mere presence of this document in the archive would suggest that Warnu too formed part of their dominions.

The next document which I would like to describe [Slide 24 11KB] is a contract for the sale of a slave, which begins: "It was the year 446, the month Ab, the day Wahman, when this sealed document, this purchase contract, was written here in the district of Samingan, at Marogan, the court of the khar of Rob". The year 446 should correspond to 678 A.D. By this time Turkish names and titles are common, as can be seen from the witness-list which follows: "Under the protection of the god Ram-set, the granter of favours, the granter of wishes, the wonderful, who is worshipped here at Marogan, at the court, under the protection of Zhun-lad son of Shabur, the successful, prosperous qaghan, the tapaghligh iltabir, the khar of Rob, and in the presence of Khusaru the tarkhan, and in the presence of Dev-raz, the brave chief justice of the khars of Rob, and also in the presence of the other noblemen who were present there amongst them and who bear witness concerning this matter". This document too was originally sealed with five seals, whose owners are named on the back: remarkably, the seals of the witnesses include that of the god Ram-set, who was presumably represented by his priest. In the same way, other documents include the god Wakhsh, the deified River Oxus, amongst the witnesses.

The text continues: "Now, I, Yaskul, and I, Yezdgird, the sons of Kaw, inhabitants of Khwastu, who are now present here in the district of Samingan, and our brothers and sons, have sold to you, Fanz, and to you, Winamarg, and to you, Pusk, the sons of Bag-mareg, you whose estate is called Gabaliyan, and to your brothers, sons and descendants, a certain boy belonging to us as brothers, this same boy who is called Khalas, for three Persian drachmas, since we are unable to keep him in plenty and famine. So now, may the boy mentioned above belong properly and well to you, Fanz, and to you, Winamarg, with your brothers, sons and descendants hereafter, from now to eternity".

I hope that these quotations may give you a flavour of the legal terminology of these documents --- for which, incidentally, one can find many parallels both in documents written in other Central Asian languages such as Sogdian and Turkish, and in the 5th cent. B.C. Aramaic documents from Elephantine in Egypt.

The next document [Slide 25 11KB] is dated in the Bactrian year 478, that is, 710 A.D., in the month "Second New Year". The document is a deed recording a gift of land and of a slave girl to the god Kamird and his ked or priest, apparently in return for the healing of a member of the donor's family [Slide 26 10KB]. Kamird means literally "head" or "chief"; it is perhaps the god's title rather than his name. The word ked is almost certainly the source of the Chinese term, ji-duo, older kej-ta, mentioned by Xuanzang as the name of the worshippers of Zhun, the god of Zabulistan, to the south of the Hindu-kush. This god was also known in the kingdom of Rob, as we can see from the name of its ruler in the previous document: Zhun-lad, literally "given by Zhun". In that case it is quite likely that Zhun is the god referred to by the title Kamird "the chief (god)". Perhaps he is the mysterious dedicatee of the cave-sanctuary of Nigar (Dukhtar-e Nushirvan) in the mountains south of Rob? [Slide 27 8KB]

The donor in the present document, the queen of a ruler with the Turkish name Qutlugh Tapaghligh Bilga Savug "fortunate, possessing service, wise, beloved", is also referred to as "the princess of the Khalas". Khalas, which was also the name of the slave-boy in the previous document, probably represents the Turkish tribal name Khalach, of which this would be one of the earliest occurrences.

One of the very latest Bactrian documents is a deed of sale dated in the year 525, that is 757 A.D. [Slide 28 13KB] I just want to draw attention to a couple of passages in this contract. The first is the series of clauses which describe the rights of the new owners of the property: "to have and to hold it, to sell it, to give it away, to pawn it, to offer it for rent, to exchange it for another piece of land, to give it for a son's wedding or a daughter's dowry, to make a monastery or temple, to make a *cemetery or *crematorium ...". Here the Indian term for a Buddhist monastery, vihara, is contrasted with the Bactrian word for a "temple", presumably referring to a non-Buddhist shrine. A similar contrast can be seen between the following pair of nouns, both of which may refer to places for disposal of the dead: laxmig would correspond to the Avestan daxma-, Middle Persian daxmag, terms which usually refer to a structure used for the Zoroastrian rite of exposure, but sometimes also to a grave, while laxshatanig, if it derives from the root daxsh- "to burn", would necessarily refer to a non-Zoroastrian, perhaps Indian, rite of cremation. These terms, together with the numerous theophoric personal names found in the documents, give us a glimpse of the variety of religious belief and practice in this area before the coming of Islam.

The independence of the kingdom of Rob was nearly at an end, however. While earlier contracts had expressed prices in gold diners or in Persian silver, the latest texts refer to "Arab silver dirhams", which seem to be specified as 'locally *current". In addition, the present text refers, for the first and last time, to the payment of taxes to the Arabs. Soon afterwards, Arabic must have replaced Bactrian as the written language of the area; and indeed, a small number of Arabic documents have come to light, which appear to form a continuation of the same archive [Slide 29 12KB].

Although I have only been able to describe a small part of an immense new body of material, I hope that I have said enough to show that it will throw new light on many aspects of the history and culture of ancient Afghanistan. But as yet I have hardly mentioned its importance for Iranian historical linguistics, though for me personally this is its chief fascination .

This slide [Slide 30 17KB] shows a small selection of forms which illustrate the position of Bactrian amongst the Iranian languages. In particular I have chosen forms which show the connection between Bactrian and the languages of the surrounding area: medieval Sogdian and Choresmian; modern Pashto, Yidgha-Munji, and Ishkashmi. Such forms support the conclusion which Henning reached on first acquaintance with the new language that it is "in its natural and rightful place in Bactria" and justify his decision to name it Bactrian.

In many cases the new material confirms or contradicts views originally reached on the basis of limited evidence. For instance, Gershevitch's controversial interpretation of lruh-minan in the Surkh Kotal inscription as the plural of a putative *lruh-min "enemy" receives strong support from the contexts in which the later form druh-min occurs. It is particularly impressive that the new texts provide examples of many previously unattested Bactrian words whose existence had already been postulated by Martin Schwartz on the basis of their occurrence as loanwords in other languages of Central Asia. [Slide 31 17KB]

Of course the new texts also attest many forms for which there was previously no evidence at all. For instance, the only verbs in the Surkh Kotal inscription are a few forms of the simple past tense and the present optative. Now one can quote almost complete paradigms not only of these tenses, but also of the present indicative, subjunctive and imperative, and even a few forms of the perfect subjunctive and optative. Some features of Bactrian are quite unexpected, such as the existence of two types of infinitive as in Sogdian and Khotanese, or the tendency to fuse sequences of conjunctions, adverbs, prepositions, and pronouns into complex words such as o-ta-kald-men "and then when to us". The fact that many texts are dated makes it possible to trace historical developments in the language. For instance, in texts of the 7th century and later, where an "l" and an "r" come into direct contact, the "l" changes to "d", as in the example quoted before: lruh-min "enemy", later druh-min.

Up to now, Bactrian has been the poor relation amongst the Iranian languages --- the one with nothing to give and everything to receive. Now that Bactrian is no longer so obscure, it can start to provide solutions as well as problems [Slide 32 12KB]. For example, the traditional translation of Avestan axvareta- by Pahlavi agrift "not taken", which many scholars have regarded as a mistake, is justified by its Bactrian cognate. Similarly, the meaning of the much-debated Middle Persian term bun-xanag, literally "foundation-house", is clarified by the corresponding expression in Bactrian, which clearly implies "estate", that is, "house and lands". A place-name mentioned by the great Muslim scholar al-Beruni, which modern scholars have emended out of existence, is shown to be correct as it stands. In the infinitive migd "to exchange", Bactrian even attests an Indo-European root which is apparently not found in any language more closely related than Greek.

The complete elucidation of the Bactrian documents and inscriptions will require many kinds of expertise: in palaeography and epigraphy, in history, historical geography, history of religions, numismatics, sphragistics, Arabic, Turkish, Chinese ... Since no one individual could possibly be competent in so many fields, such a task demands collaboration between scholars in several disciplines. The starting-point for this collaboration must be the decipherment of the text and the drafting of a first, tentative translation. This is the business of the philologist, who employs his linguistic instinct and his knowledge of cognate languages to formulate hypotheses about the meanings of words and the grammatical structure of the language. Without his preliminary work there is in fact nothing for anyone else to study. Philology is a branch of scholarship which some regard as old-fashioned; but I am proud to call myself a philologist and glad to have had this opportunity to describe to you a body of material which so well demonstrates both the necessity and the rewards of philological research.

© N. Sims-Williams 1997


Larger photos of the documents shown in the text above.


To the Japanese Version
Back to Home Page