A Review Of Archaeological Reports: Southern Bactria And Northern India Before Islam

Despite a huge and ever-increasing bibliography, the history of northern India from the death of Asoka to the first inroads of the Moslem armies is still imperfectly known. About its social history we can only state that new peoples kept coming from Iran and Central Asia and were, in the course of time, integrated into an Indian social organization about which we have very little incontrovertible data.[1] Its economic history is summed up by lists of commodities, some indications about currencies and monetary policies, and imprecise records of its trade with China and the Roman Empire. Thanks to recently discovered inscriptions and sculptures, the complex relationship between Buddhism and early Hinduism now appears in a new light, but this new data comes from widely separated Places (mainly Gandhara and Mathura) and its interpretation may be disputed.[2] The political history of northern India still consists of bare lists of names, with an often unsure relative chronology and a still more unsure absolute chronology. These chronological uncertainties cannot but have a bearing on the history of early Indian Art which, despite some advances,[3] has not yet been established on a sure footing.

The extant Indian, Western, and Chinese literatures have been so carefully sifted that new important revelations are not to be expected. New inscriptions and coins are published almost every year, but they are more often than not stray finds whose whereabouts are imperfectly known.[4] Thus, the only hope for the historian of early India lies in regular excavations. Indeed, as early as 1903, (Sir) John Marshall planned to start excavations in Taxila, partly because its location and history reminded him of ancient Greece, but mostly because he wanted to recover all kinds of data "on the political and religious history of the northwest ... and its material culture during lengthy periods between 500 B.C. and A.D. 500."[5] The trend was followed by (Sir) Mortimer Wheeler, who came to India in February 1944 with a plan of systematic excavations for recovering India's past.[6] But he had to leave India (now Pakistan) without being able to implement his plan fully. The Archaeological Survey of India and the Department of Archaeology of Pakistan did not follow in his steps. But his ideas were taken up by three archaeological missions whose heads had impressive academic backgrounds: the Delegation Archeologique Francaise en Afghanistan (DAFA), whose postwar director, Daniel Schlumberger, had the same taste for history as its founder, Alfred Foucher;[7] the Istituto Italiano per il Medio ed Estremo Oriente (IsMEO), whose chairman, Giuseppe Tucci, in 1955, earmarked places in the Swat valley of Pakistan for excavations with the specific purpose of unraveling the history of Buddhism in northwestern India and discovering the sequence of the Hellenized Buddhist art of Gandhara;[8] and the Archaeological Mission of the Museum of Indian Art (Berlin) which began to dig at Sonkh, near Mathura, in 1966 under the leadership of Herbert Hartel, then director of the museum, a pupil and heir of Professors Luders and Waldschmidt, to "collect material information on the early history of the once-flourishing State of Mathura, one of the most important cultural centres of ancient India."[9]

This was how, during the sixties, three large excavations were conducted in Afghanistan, Pakistan, and the Republic of India which should have helped to solve, but did not, the vexed questions of the origins of early Buddhist art, the creation of the anthropomorphic representation of Buddha, and the subsequent developments of early Indian art: Ai Khanum in northern Afghanistan (1964-1978), started by D. Schlumberger but almost entirely led by P. Bernard, which was to have given important clues to the Hellenization of northwestern India; Butkara I in Swat, "chosen by Prof. Tucci after careful study of historical sources continually checked by inspection of the ground,"[10] entirely dug out by Domenico Faccenna (1956-1962), where it was hoped that careful stratigraphical excavations could help to establish on a sure footing the chronology of Gandharan art; and Sonkh, excavated by Prof. Hartel in eight seasons (1966-1974), which should have been the counterpart of Butkara I in the Gangetic valley. These were large-scale excavations, backed by local authorities, conducted with great care by first-class scholars with relatively important budgets and a good staff of technicians and young promising scholars. We now know that they did not solve the difficult problem of the sources and chronologies of the Gandharan and Mathuran Buddhist arts. Despite the wealth of well-recorded finds, the new advances came mainly from stray finds,[11] and the best overviews are still Schlumberger 1960 and Lohuizen de Leeuw 1949. There is no reason to blame this failure on the excavators. Luck simply was not with them. They cannot be held responsible for not having dug out inscriptions giving a genealogy of Bactrian kings or ascribing a fine sculpture to the reign of a Saka sovereign. Although the problems of early Buddhist art were not solved during these excavations, we know for sure that Ai Khanum will remain for many years to come the standard site for Greek Bactria, that Butkara is the best recorded Buddhist site in northwestern India and that Sonkh will in the coming years be the great reference site for Uttar Pradesh. These excavations are not failures. On the contrary, they are text-book instances of the importance of well-planned and well-conducted excavations for our knowledge of the past of Central Asia and India.

This is not the place to review the important data they brought to light. They have been known for years through many interim reports[12] and exhibitions; the academic world has no doubts whatever about their importance. Now, when the final reports have been published, or are being published, it would be interesting to compare the strategies of excavation and publication adopted by the leaders of these huge projects. Although I owe a great debt to Gardin 1979 and agree with most of his conclusions, this will not be a study in theoretical archaeology. I have no taste nor skill for theory. This will be an empirical study, made by a historian who often needs to look into these reports and who is also faced with the necessity of conducting such projects - even if they are on a much smaller scale - and publishing them in due time.

In its restricted sense, Ai Khanum is a Hellenistic walled city located in northeastern Afghanistan, close to the border of current Tajikistan, at the meeting point of the Oxus (Amudarya) and Kokca rivers. It consists of two parts, a natural acropolis, 60 m high, mainly used for defensive purposes, and a lower town which was the inhabited part of the city (1,800 x 1,500 m in all). Part of the acropolis and more than one-third of the lower town were excavated. The main buildings there are a huge palace or administrative quarters, a large temple, a mausoleum and a heroon, a gymnasium, a theater and an arsenal. Great houses, with built-in courtyards or gardens, were located south of the palace (in the southwestern part of the city) and outside the city, nearby its northern wall. There were few such houses: it seems that there were many empty spaces inside the walled town, which was founded and built on virgin soil either c. 329 or 305 B.C. and ceased to exist as a town c. 146 B.C. There is no doubt that Ai Khanum was one of the main urban centres of Greek Bactria.

North and east of the city lies a small plain, 27 km long, less than 9 km wide, 220 [km.sup.2] in all, which French scholars call "Ia plaine d'Ai Khanum."[13] A survey, begun on J.-Cl. Gardin's initiative in 1974, with P. Bernard's agreement, ten years after the first dig in Ai Khanum city, demonstrated that this plain had been irrigated since the bronze age and could even boast a mature Harappan settlement (Gardin 1978, 130; Francfort 1989, 57-58). There were important settlements until the time of the Mongol invasions. The existence of the Greek city of Ai Khanum is thus only one important and, in any case, the most spectacular episode in the long history of the plain. It should have been studied accordingly, but this did not happen, as can be clearly seen from the scheme of publications adopted, a scheme quite inconvenient both for readers and librarians. Apart from the many interim reports, the results of the digs in Ai Khartum city have been published in the lavish series Memoires de la Delegation Archeologique Francaise en Afghanistan, under the responsibility of P. Bernard; the results of the surveys and excavations in the Ai Khanum plain were first published, under the direction of J.-Cl. Gardin, in cheaply printed series of the Centre National de Ia Recherche Scientifique (CNRS) or Editions Recherche sur les Civilisations, now in a series called Memoires de la Mission Archeologique Francaise en Asie Centrale. There is almost no relationship between the two sets of publications. The puzzled reader may wonder why one and the same French project is published in two separate series. I often do.

It is easier though to discern the scientific differences between the two series. The studies published in Memoires de la Delegation Archeologique Francaise en Afghanistan, as well as the interim reports by P. Bernard, are almost entirely devoted to studies of the Hellenistic architecture and the decorative art of the Greek city. The keywords of P. Bernard's papers, as listed by himself (Guillaume 1983, xi), are: Hellenistic, Greek, city, urbanism. History in the modern sense of the term almost never appears.[14] The scope of the J.-Cl. Gardin's series is much wider - Central Asia. There are often no chronological limitations (Gentelle 1978); the emphasis is mainly on non-Greek periods and never on art.

Now there is no epistemological conflict between classical archaeology, with its emphasis on architecture, art, and political history, and the so-called new archaeology which aims at solving clearly defined historical problems through specific and fast procedures. We need both, and each complements the other. One can be interested in history of art and at the same time one may want to know about the everyday life of the artists and their servants. Although the DAFA did not lack money and although many scholars, with a wide range of interests, were working under its umbrella, obviously classical archaeology could not be reconciled with the new archaeology. The two main series of Memoires, the one under the editorial responsibility of P. Bernard for the Ai Khanum volumes, and the other under the editorship of J.-Cl. Gardin, mirror these scientific differences.

As far as I know, there is no preconceived publication plan for the Ai Khanum subseries of final reports in Memoires de la Delegation Archeologique Francaise en Afghanistan. The volumes follow each other treating building after building in a haphazard way: propylons of the main street, main temple, coins, fortifications, gymnasium, treasury.[15] Indeed the excavations are published as they were made, mound after mound, without any systematic plan of studying the grid of streets, the function of the empty spaces, the zoning of the city, the reasons for the location of buildings like the gymnasium or the arsenal. It does not seem that a systematic study of town-planning will ever be published.[16] The only overview of the city - very good if somewhat brief - is a comment on the often reproduced map of the Ai Khanum excavations (Bernard 1981, 110-14). There is no detailed study of the aerial views, which now could be supplemented by satellite imagery, and there was no surface prospection nor soundings (or very few) made for the specific purpose of ascertaining where the artisans or the servants dwelt.[17] We do not even know whether there was a market, or a bazaar, even less where it was located.

We are given no demonstrative reasons about the purpose of building such a city in such a location.[18] Nobody will blame the excavator for not being able to give the Greek name of the city, for not having discovered any sure evidence about the exact date of its foundation (already in the time of Alexander or under Seleukos the First), nor for being unable to state whether the palace was a royal residence or only administrative quarters inhabited by a governor and his retinue. The excavations were conducted with the utmost care, but no decisive inscription was discovered. The walls had been razed to the ground, and the buildings had been emptied of their contents. But we could have expected some kind of geopolitical study. In its place we are only given a very short set of unsupported statements. The first interim report (Bernard 1973) does not even address the question. One year later, an explanation is given: Ai Khanum was located in a strategic position and was used as a military outpost to keep in check the inroads of nomad enemies. It also derived some wealth "from direct farming colonization by (Greek) gentlemen-farmers" (Bernard 1974, 102).

When the first surveys of the plain of Ai Khanum by Gardin's team demonstrated that the land was irrigated long before the advent of the Greeks, another set of explanations was given, the final one as far as I know: the city was established because it could use the produce of a rich territory, well irrigated, well farmed, and inhabited by large numbers of skilled peasants. It could also derive some profit from the mineral wealth of the high valley of the Kokca river (or Badaxsan). Lastly, located in a strategic position against both hill tribes and steppe nomads, it was an indispensable fortified support area against them.[19]

Ai Khanum was indeed a fortified city, but all Greek cities were walled. I am not a general or a strategist, but I find it hard to believe that Ai Khanum was mainly a military outpost or fortress against the nomads. A partly empty city with a wall too long to be effectively manned is not a fortress. And if it had been a fortress, why, as it seems, was it vacated even before the advent of the nomads it was supposed to hold in check? The mineral wealth of the hilly country could be tapped anywhere in the plain, even north of the Hindukush (as in modern times). It probably added to the resources of the city, but it does not explain its location. The exploitation of farming possibilities on a well-irrigated countryside, associated with the possibility of building a large fortified town on a convenient site (two rivers, an acropolis), is thus the best explanation. Trade could also have played a role.[20] But these explanations will remain undemonstrated as long as the relationship between the countryside and the city is not explored. Who were the farmers? Were they affluent? Did they use the same kind of pottery as the citizens? Were there artisans and traders in the city?

These questions cannot be answered because they were never addressed. No use was made of Gardin's surveys for initiating a study of this kind. In confirmation I can adduce two facts. The maps published in the Memoires de la Delegation Archeologique Francaise en Afghanistan, even in their latest issue (Rapin 1992), never show Ai Khanum within its territory (either the plain of Ai Khanum or a larger part of northeastern Bactria). They are either large maps of Bactria, or a plan of the city which stops at the northernmost excavated buildings, i.e., 300 m north of the northern wall. Moreover, very few people know that two km north of the northern wall of the city lies a walled circular establishment, large enough to be called "circular town," and which, to my knowledge, is almost never mentioned by P. Bernard,[21] although sherds ranging from Achaemenid to Islamic times were collected from its surface.[22] If this town was a provincial headquarters before the advent of the Greeks, the historical and geopolitical problem of the location and function of Ai Khanum city should be addressed in a very different way.

This long discussion is not intended to deny the value of the Ai Khanum excavations. They were well done and they produced outstanding results. Without them, what would we know of Greco-Bactrian planning, architecture, sculpture, pottery, etc.? Without them, Gardin's surveys would not have been as successful as they were. These excavations brought to light a new world which the French archaeologists had relentlessly sought after for forty years. One cannot complain that the emphasis was on art history and traditional historiography.[23] This is what I do myself most of the time and it is necessary. It is indeed often a prerequisite. But it is to be regretted that, although there was no lack of funding or manpower, the excavations in Ai Khanum city were not planned in a way which perhaps (i.e., with some amount of luck) would have also allowed us to solve questions which are at the heart of contemporary historiography, such as demographic history, economic history, social history, agrarian history, etc. The Ai Khanum excavations demonstrate the truth of the old adage: you only find what you search for.

The final reports published in the Ai Khanum subseries, Memoires de la Delegation Archeologique Francaise en Afghanistan, are of a very high scholarly standard, with a wealth of maps, plans, drawings and photographs. But there is a bewildering lack of consistency, even in the format.[24] I have already noted the absence of an overall plan. There is no introductory volume, no historical account of the excavations, no volume on town-planning proper and, worst of all, no overall chronology. So most authors feel compelled to give their own chronology, either in a few lines (Francfort 1984, 2-3; Veuve 1987, 100-101) or in a long chapter (Rapin 1992, 281-94). This would be of no consequence if there were no discrepancies. But this is not so.

It is generally agreed that the town was vacated by the Greeks c. 145 B.C. under pressure from the nomads, although no incontrovertible evidence that it was attacked, burnt, or sacked by these nomads was ever found.[25] P. Bernard only made the noncommittal statement: "the Greeks of Ai Khanum were driven from their city by nomad invaders" (Bernard 1982, 148).[26] After a while the deserted city was pillaged by later inhabitants who settled in some of the ruined buildings (Bernard 1982, 110; Francfort 1984, 2), but Ai Khanum no longer existed as a town. So one is puzzled to learn that the town was probably vacated after an attack (Leriche 1986, 57); that it was destroyed by Yuezhi nomads (Rapin 1992, 291) and that there was a Kushan occupation (Leriche 1986, 99-101). What is P. Bernard's own opinion? Is there any evidence of these facts in other excavated buildings? And why are we not given the possibility of judging whether the arrowheads discovered against the northern wall belonged to Greek troops (who could have enlisted Central Asiatic archers) or to nomadic tribes?[27] Can we not know from the ceramic finds whether the city was inhabited in Kushan times or not?[28]

This is only one instance of the inconsistencies between the series. Apparently no guidelines were given to the authors, so that some reports are quite matter-of-fact (Guillaume 1983, Guillaume 1987, Veuve 1987), others look like a demonstration of classical erudition (Francfort 1984). The classification of finds in Guillaume 1987 is not the same as in Francfort 1984. Some volumes are Ph.D. theses (Bernard 1985, Rapin 1992) - which is a nice way of publishing archaeological reports, if the Ph.D. dissertation is shortened before publication. But this was not the case. In a series of final reports, it would have been more useful to reprint the coin hoards rather than to annex long disquisitions which are understandable in a Ph.D. thesis but have no connection whatever with the Ai Khanum excavations.[29]

Similarly, Rapin should have avoided publishing in this series, well after D. Schlumberger and P. Bernard, his own history of the Hellenization of Bactria, propounding little that is new[30] and encumbering it with a profusion of useless over-erudite footnotes.[31] He should have shortened - among others - his Sakuntala story (pp. 192-97), found in all histories of Indian literature, and I would have expected him to have been much more cautious in discovering a depiction of this same Sakuntala story in the so-called "plaque indienne."[32] Greater concision would have made this excellent volume much better and less expensive.

One can also wonder what logic explains that the excavations of some buildings are published with their associated finds (Rapin 1992), some without them (Veuve 1987, Leriche 1986), some with only a few associated finds (Guillaume 1983, 41-42). One also wonders why some finds were published long before the publication of a detailed description of the dig (Francfort 1984), a situation that makes it difficult to check the important - and indeed very useful - references to the stratigraphy, and why other, quite similar artifacts were published with minimal indications of their provenance (Guillaume 1987, 84). Will these last finds be published anew when the building they came from is published? And will these buildings ever be published? Seventeen years after the end of the excavations, we are still awaiting a full publication of the palace, the two temples, the houses, the arsenal, the theater, the cemetery....

The surveys and excavations made by J.-Cl. Gardin and his team in eastern Bactria during the very short period (1974-78) when it was still possible to travel in Afghanistan are published in a new series called Memoires de la Mission Archeologique Francaise en Asie Centrale. As far as I know, the reasons for creating a new series are never given. Apart from the obvious fact that the field work was financed by specific grants, not by the main budget of the DAFA, the substitution of the words Delegation and Afghanistan by Mission and Asie Centrale has some significance. The disappearance of the word Delegation obviously implies a break with the French tradition of the DAFA, i.e., a permanent establishment with a permanent staff (it was founded in 1922, when the journey from France to Kabul took about three months), having a special relationship with the Afghan government, and smacking of imperial times.

Asie Centrale implies a shift in scholarly interests. It denotes a temporary break with the Indian tradition of the DAFA, and an emphasis on a geographic, cultural, and geopolitical entity (Central Asia) whose existence is longer and historically more significant than that of recent states like Uzbekistan or Afghanistan. It implies also that the same problematics could be used in other parts of Central Asia. Indeed, the same French research team was and is still conducting explorations and digs in Tajikistan, Kazaxstan and Chinese Turkestan (or Xinjiang). For its Afghan irrigation surveys, it follows a trend first set by Soviet archaeologists.[33]

There are other obvious differences with former DAFA practices. The surveys were meant to be relatively cheap. They were made by a restricted team (four French scholars,[34] three to four Afghan assistants), acting on a written agreement with the Afghan government. This agreement stipulated the scientific objectives, the methodology, the relatively short duration (four years) of the project and insisted on early publication of the results.

The scientific objectives, the methodology, and the epistemological hypotheses are clearly stated in the first pages of the interim and final reports.[35] The purpose of the surveys, and of the consecutive dig of Shortughai, is clearly historical. The concern is not with art history, nor political history; it is not a search for the footsteps of Cyrus or Alexander. The historical questions J.-Cl. Gardin and his team address are the same as those which the Annales school tries to answer, along with more contemporary prehistorians. When did people settle in eastern Bactria? Why, i.e., for which ecological, geographical, economic reasons, did they settle there? How and to what extent did they modify the landscape? How did the population increase or decrease, and why? No time limit is set; that is to say, each so-called historical period is as valuable as any other for the historian, whether times of "barbarity" or times belonging to the "Greek miracle." To solve such problems, no artifact is more "noble" than any other: a sherd may tell us more than a long inscription.

The means used for solving these problems owe much to contemporary prehistory: surveys backed by a good knowledge of local geography, paleogeography, and even geology; collection of surface artifacts; limited soundings; the use of modern technology (satellite imagery,[36] physical and chemical analyses, paleontology or paleozoology, etc.). The working hypotheses were the following. There had been no substantial change in the climate since neolithic times. Accordingly, in ancient times (as in modern ones) agriculture was only possible with the help of channel irrigation. Tracing and dating old channels is therefore equivalent to tracing and dating old agricultural settlements.

Tracing old and dry channels depends on the ability and the eye of the surveyor, who should be also skilled in reading maps and aerial photographs. The method, experimented with in 1974 and 1975 during the surveys of the plain of Ai Khanum (Gardin 1976; Gentelle 1978), was developed to enable a very fast survey of the major part of eastern Bactria (Gardin 1978; Lyonnet 1995). It is specific to northern Afghanistan, which was still thinly populated and where wide tracts of arable land were no longer cultivated or only sporadically sown when the rains were good (lalmi, dry cultivation). As for the dating of the channels, it should be consistent with the date of the sherds collected on the ground of settlements whose economic activity was dependent on the functioning of these same channels. The validity of this postulate was demonstrated by a limited number of small digs and by the Shortughai excavations.

The results of these well-planned surveys are impressive. They demonstrated that, as postulated, the climate had not changed much during the past 5,000 years; that agriculturists had been settled in Bactria at least since the early third millennium B.C.; that channel irrigation was practiced on a large scale, and sometimes under very difficult technical conditions, long before the Greek conquest; that Ai Khanum was built on an already densely populated territory; that some changes in population could be inferred from the distribution of channels and sherds, and possibly correlated with previously known historical data.[37] They were also a very successful exercise in new archaeology, showing how a well-devised plan of surveys and excavations could solve historical problems at a cheaper cost, if these problems were precisely worked out beforehand.[38]

Early publication was an important part of Gardin's "new archaeology" challenge. Two main interim reports were published quickly, before the forced cessation of the surveys (Gardin 1976 and 1978). Four final reports have now been printed (Gentelle 1978, Francfort 1989, Gentelle 1989, Lyonnet 1996). A fifth and last is promised for the not too distant future.[39] These publications are also a challenge for J.-Cl. Gardin and his team, insofar as they have been advocating for years a more formal (logicist) argumentation in drawing historical conclusions from archaeological data and changes in publication practices.[40] As there are some acknowledged discrepancies in the historical conclusions of the interim and final reports,[41] I shall only review these.

The survey of the plain of Ai Khanum (1974-76) was published quite fast (Gentelle 1978).[42] It is a classical and matter-of-fact study. Starting from the assumption that there had been no major changes in climate, P. Gentelle first made a study of the geography and modern agriculture of the plain. He thus learned what were the preconditions for cultivation of the soil and for irrigation, whether in antiquity or today. He was then able to trace the former canals and even discover the remains of former fields and farms. As a result, we are given a full study of the hydro-history of the Ai Khanum plains. from ancient times to 1974, illustrated by many maps and photographs, including a first dating of the irrigation channels and a first estimate of the extent of the cultivated areas. The dating of the sherds found in the old channels or in associated archaeological sites made it possible to date these channels and to demonstrate that the plain was well cultivated before the coming of the Greeks, even in Harappan times. The irrigated areas were largest in Greek times, but as some channels were probably in use in pre-Greek, Greek, and post-Greek times, it was not possible to say whether there was an increase in cultivation in Greek times. A preliminary study of the former settlements suggested that in many parts of the surveyed area the population inhabited scattered large farmhouses.[43]

There is nothing specific in the outward appearance of the publication. It is cheaply produced (offset printing), with many maps and drawings (fifty-six in all) and good photographs. Compared with other archaeological publications, e.g., the Memoires de la Delegation Archeologique Francaise en Afghanistan, the most perceptible improvements are cautious wording, a minimal use of conclusive hypotheses, and the absence of useless pseudo-erudite footnotes.

The experience gained during this first survey was used for a much larger survey of all the arable lands east of the Kunduz river.[44] New heuristic hypotheses were formulated and tested in the field to take into account the much greater expanse of the surveyed area and the differences in population and cultivation.[45] The survey was planned for a four-year period (1977-80). It had to stop after two years. The traces of ancient channels are mapped in Gentelle 1989, which is very similar to Gentelle 1978, only better, with good studies of the geology and geography of the surveyed area. It is the indispensable geographical preliminary to the historical studies which are to follow. The differences with the former volume are improved printing (computer-aided), much better (three color) maps and photographs, and a series of studies by specialists on ancient vegetation and soils. These specialized studies establish, with great profit, a link with prehistoric archaeology, where such studies are systematically used.

The first set of historical studies relative to this survey is now in print.[46] It is a chronological study of the ceramics collected along the ancient irrigation channels and in associated archaeological areas (25 tons). The study of these surface artifacts is made according to Gardin's morpho-technological criteria.[47] The time continuum is divided into a number of historical periods for which a specific sample (assemblage) of ceramics is selected.[48] The ceramics are correlated with other ceramics found in other parts of Eurasia (mainly Soviet Central Asia), thus enabling the author to draw some conclusions about their date and provenance. Dated ceramics date the channels and the associated areas where they were found. It is thus possible to draw a map of irrigated cultivation for each selected period. Some changes in shape may indicate changes in population.

There are four main prerequisites for such a study. The first one is that the survey have value. J.-Cl. Gardin has often explained very convincingly how his procedure produced representative samples[49] and we may trust him. The second is that the technical study of ceramics have value. I am no specialist but have no doubt whatsoever that it is one of the best ever produced in that part of the world. Thanks to Lyonnet 1996, we will have the first systematic conspectus of northern Afghanistan ceramics. Lyonnet 1996 will become the reference work for any new Afghan or Central Asiatic (in the former Soviet sense) survey, or for any comparative study of the ceramics of that area.[50] The third prerequisite is a perfect knowledge of the scientific literature relative to ceramics. The bibliography of Lyonnet 1996, and especially the way she uses it, does not leave any doubt as to the excellence of her knowledge. Indeed, now that the eastern Bactria ceramics are well dated by the comparative studies she made, her book will be used as a criterion for dating ceramics everywhere in Central Asia, even the ones she used for comparison.

The last prerequisite is sound historical reasoning and cautious wording. Indeed every caution is taken when discussing remote periods about which we have no written evidence. It is shown that in the third millennium B.C., eastern Bactria belonged to a Baluc cultural sphere, without implying that it was a political entity or that there was any ethnic unity. The same is said about the Harappan period (2500-1500), at a time when close contacts with the Indus civilization are attested by surveys and the consecutive Shortughai dig (Francfort 1989, below): B. Lyonnet expresses some reservations about any explanation founded upon economic or ethnic colonization. She clearly says that the eastern Bactria surveys do not help to solve the vexing problem of the relationship between Turkmenia, Bactria, and Balucistan, and that there is no field evidence for the historically sure arrival of Indo-Iranian or Indo-Aryan tribes. In the same way there is no ceramic evidence for the inclusion of Bactria in the Persian Achaemenid empire, although the existence of a Bactrian Achaemenid satrapy is a well-known fact.

For later periods, B. Lyonnet did not always resist the temptation to correlate political and cultural events or areas. The chronological periodization she adopted suggests this at first glance. I am quite confident that she has every reason to see changes or intrusions in the ceramics tradition c. 330 B.C., c. 145 B.C. and c. 270 A.D. But she gives political names to these divisions: Hellenistic period (330-145); age of nomad invasions (145-40); Kushan period (Yueh-chi-Kushan, mature Kushan, Kushano-Sasanian and later), although the correlation may be less than evident.

The notion of nomadic invasions is a complex one. According to Chinese sources, Bactria at that time was a kind of patch-work of autonomous towns paying tribute to the invaders,[51] so that it cannot be believed that Hellenistic pottery came to a sudden end in 145 B.C. This is rightly said by B. Lyonnet herself. But it follows that some sites where only Hellenistic sherds have been found might date from Yuezhi times, so that the political denomination of the ceramics series is somewhat deceptive. In the same way, sherds from Kushan times may be found in places which never belonged to Kushan dominions. This is true of many Indian provinces (which are outside the scope of B. Lyonnet's book), but could also be true of some places in western Bactria. Moreover, economic trends during the Kushan period do not exactly follow political trends. There was a high phase of the monetary cycle apparently beginning with the last period of Wima's reign and ending sometime during the reign of Huviska, and economic trends for ceramics may be more important than political ones. "Kushano-Sasanian and later" mixes together very different times, among them many invasions of eastern tribes (Huns, etc.).

A better use of logic(ism) would have prevented B. Lyonnet from correlating the phases in Hellenistic pottery with political events, Alexander's conquest (white ceramics) and Seleukos I's reconquest (grey-black ceramics). One may believe that the difference is rather between unsettled and settled conditions, as unsettled conditions are not very conducive to big changes in fashion or culture. I do not believe either that a study of ceramics would help to solve the difficult problem of the borders of the Bactrian kingdom or of its administrative units (Bactria and Sogdia). The extent of these political entities changed from time to time, and without further evidence it is always risky to correlate longstanding cultural areas and fast-changing political units. This was demonstrated long ago by J.-Cl. Gardin (and now by B. Lyonnet) for Achaemenid times (Gardin 1977). In the same way, B. Lyonnet is probably right in ascribing some shapes (gobelets a piedouche and bottled-shaped pots) to nomad invaders, less so in attributing them respectively to the Saka (Sai) and Yuezhi, because Sai and Yuezhi are Chinese names for shifting confederations of tribes without any linguistic, ethnic (i.e., racial), and probably cultural, unity. These confederations no longer existed as political entities when they reached Bactria. This fact explains why Greek authors, who name four groups of tribes, do not know them. Perhaps there were more, among them some who never belonged to the former Sai or Yuezhi confederations.[52] So that the addition of new and specific shapes to the former characteristic Hellenistic ones, although denoting in this specific case (where we can use some written evidence) the arrival of new populations, cannot be correlated with historical names, be it only because the written evidence is patchy.[53]

These are the only flaws in an otherwise remarkable book, whose importance for the historian lies in the huge amount of economic and demographic data it supplies. We have for the first time a history of Bactria spanning a long stretch of time. The main economic impulse dates back to the second millennium B.C.: neither the Achemenids nor the Greeks were responsible for the irrigation of the larger part of Bactria, and the probable increase in population which ensued. There was no decisive increase in cultivation with the advent of the Greeks. Hence Greek colonization was not responsible for the proverbial agricultural wealth of Bactria. The second-century B.C. conquest of Bactria by eastern nomads apparently meant a decrease in cultivation,[54] which was never entirely remedied, even in later (Kushan) times. Further research will help to explain these bare facts or their consequences. Given the fact that there was no major increase in Greek times either in the cultivated areas or in agricultural techniques, and consequently no major increase in agricultural revenue, one might assume that the construction of huge monuments in Ai Khanum before any Indian conquest meant that the Greek officials or landlords took a much bigger share of the local agricultural produce than their predecessors. Other statements raise questions to which we cannot give any answers at this time. As an instance, it is now quite certain that even in remote villages locally made Hellenistic ceramics soon replaced earlier ceramics. This major and long-lasting cultural phenomenon can be explained in very different and even contradictory ways. One may advocate technological reasons (sudden improvements in quality and productivity which made it advantageous for local producers to stop their traditional production and to adopt new and foreign shapes); economic reasons (a new organization of the markets which enabled Greek potters to outsell local producers); political reasons (the political will of the new masters to impose the Greek way of life on the country; or the enslavement of the former population who could no longer choose their everyday pottery; or even the replacement of the earlier population either by western free colonists used to Hellenistic ceramics, or by slaves who were provided with pots by their Greek masters), etc. The importance of Lyonnet's book lies not only in the novelty and accuracy of the many statements it contains; it also lies in the novelty of the many questions it raises.

The final report of the Shortughai excavations, which is a revised Ph.D. dissertation (Francfort 1989), would deserve a longer review, both for the importance of the finds and the way they are published. The excavations, made after the discovery of Harappan sherds during the survey of the plain of Ai Khartum, brought to light the remains of a Harappan settlement, either a trading or colonial outpost (Francfort), or evidence for the existence of a huge cultural zone including both Central Asia and the Indus valley (Lyonnet), and proved the use of channel irrigation in Central Asia as early as the third millennium B.C. The excavations were apparently conceived as a challenge. They were supposed to validate the methods of the survey and the deductions of the surveyors, which they did. They were supposed to demonstrate the usefulness of the recently created Centre de Recherches Archeologiques (CRA), which put at the excavators' disposal a pool of full-time specialists in paleo-sciences who could help in maximizing the results of diggings. Shortughai is thus the only French excavation in Afghanistan where some use was made of methods now commonplace in prehistorical and even historical diggings: a geophysical survey was first made (that only indicated that there was no cemetery); plant, animal and human remains were collected and studied with care by specialists; minerals were analyzed, etc.

Publication was the main challenge. The final report, although quite elaborate, indeed much more elaborate than most final reports, was published very quickly. The site was discovered in 1975. A first trial trench was dug in 1976, the main excavations were made in three successive campaigns (1977-79). After two interim reports, a very detailed and richly illustrated final report was published in 1989 (Francfort 1989). Considering that H.-P. Francfort was busy with other publications and field work during these ten years, it is a wonderful achievement. The text is divided in two parts, a detailed report of the Shortughai excavations followed by a historical study of the economy and population of Central Asia before the Achaemenid conquest.

The report is written in a very unusual way. It is an illustration of the endeavors made by J.-Cl. Gardin's team to improve publication practices in archaeology? As such, it deserves careful reading even by archaeologists or historians who have no interest whatever in Central Asia. The main endeavor is the precise clarification of the objectives, the procedures, and the deductive reasoning used during excavations and publication, with a minimal use of conventional rhetoric. The book thus begins with a reminder of the limited aims of the excavations (a search for the entire Harappan historical sequence, i.e., a search for stratigraphy, evolution and ecology, not for buildings), the definitions of words whose meaning is usually taken for granted (level, layer, soil, wall, etc.) and statements concerning the deductions which can be made from the many possible combinations of the realities expressed by these words. This is presented in a few words and diagrams, similar to models found in mathematical or linguistic literature. Then comes a description of the dig, level by level, room by room, with almost no grammatical sentences. The data is always tabulated in the same order, followed by a study of the ceramics in Gardin's way,[55] [56] with all necessary comparisons with Central Asiatic and Indian potteries.

The evolution of the ceramics is studied with the help of a computerized data-base and statistics, according to a procedure which, due to my own limitations, I am at pains to follow and to explain. In this way "non-subjective" conclusions may be drawn about the relative chronology of the different levels and the evolution of the Shortughai ceramics during two millennia. Then follow detailed studies of other artifacts, plant and animal remains, and two human skeletons. The time has now come for syntheses (excavated area by excavated area, period by period, level by level) and limited ecological and sociological conclusions. Here ends the final report proper, a model of completeness, useful erudition, and clarity. When using it we know what was done, how it was done, what was found, which conclusions the author reached, and the exact procedure he followed to reach them. All the help we need (drawings, analyses', etc.) is given and the necessary reservations are made. Gardin's challenge is met and the demonstration is done.

The second part is a study of the economy, demography, and movements of population in Central Asia in the so-called Bronze and Iron Ages, backed in part by the results of the Shortughai excavations. H.-P. Francfort displays the same qualities of erudition, clearness, and caution. This is not the place to discuss, nor even sum up, the results of his study, which looks quite sound. Specialists will find many important hints for dating the Bronze (e.g., Namazga) and Iron (e.g., Jazd) Cultures more precisely and good discussions about the dilemma between the diffusion of cultures and population movements. H.-P. Francfort is most probably right in denying the existence of any archaeological evidence of the Indo-Aryan presence, although most of our (former) Soviet colleagues are ready to detect it. But as an historian, I do not feel much at ease in this second part. Although afraid of being thought too conservative and in some ways even backward, I feel I am an intruder in a club of colleagues familiar with the same problematics and language. Although these problematics and this common language are quite usual now in South Asian and Central Asiatic prehistorical studies, and even in some historical studies, they remain foreign to me. H.-P. Francfort feels it necessary to prove, in a very courteous, scholarly and computer-aided way, that most of the theories advocated by his colleagues do not hold water, which to my hypercritical mind seemed obvious at first glance. He devotes many pages to the often debated questions of the emergence of the state, the definition of a town and the processes of urbanization and deurbanization. But most of the problems arise from using technical terms that denote ever-changing realities and hence cannot have invariant meanings. Lacking absolute definitions, they cannot be grasped by one and the same set of criteria - pace the view of Gordon Childe, and many others since, that there is no such thing as the essence of a town. There is nothing in common between Hastinapura and New Delhi, even less between Hastinapura and Ibadan - except the fact that they are inhabited areas located on a continuous scale with megalopolises and conurbations at one end and isolated houses at the other, with huge time and place variables. Census-makers are familiar with this kind of problem and at each new census feel obliged to elaborate new criteria to express changing realities. Indeed many discussions would be cut short if fewer words with fewer connotations were used. It is easier to reach a consensus when talking about political organizations, or sizes and densities of inhabited areas, than when referring to the emergence of the state and growing urbanization.

The other reason why I do not favor such formalized and computer-aided discussions is because I do not believe that history is a science or should be a science. Archaeology, philology, linguistics are sciences and as such need to evolve formal procedures and methods of validation. History is not, as demonstrated by Veyne 1971, and there is no reason why it should be. It is a discourse whose main objective is the pleasure we take in writing or reading it. Lack of accuracy and faulty reasoning spoil this pleasure, but that does not make history a science. Most readers do not like faulty reasoning in detective novels either.

Afghan Turkestan was not an easy country; I suspect Swat was more difficult still. It is densely populated and cultivated. Everybody boasts at least one gun. Women are kept under strict seclusion. Nor do poppy cultivation, drug smuggling, weapon trafficking and lucrative illegal digging encourage archaeological surveys. But Pathans are a hospitable people and, if you do not break the rules, you may be welcome. Indeed G. Tucci succeeded in establishing an archaeological mission whose scope was very broad from its inception. Although, personally, mainly interested in the history of Buddhism, he launched a series of excavations with the avowed aim of unravelling the whole history of the Swat valley, from prehistoric times (G. Stacul) to nineteenth-century mosques (U. Scerrato), without neglecting inhabited areas like Udigram (G. Gullini) or Bir-kot-Ghwandai (P. F. Callieri).[57] Limited surveys were made and an archaeological map of the Jambil valley, supposed to be "section I of the Archaeological Map of Swat," was in preparation.[58]

Judging from Tucci's motivations[59] and the number of publications, the main object of research was the Buddhist "Sacred Precinct"[60] of Butkara I. The purpose of the excavation was to unravel the tangled history of Gandharan art and Buddhism. The site was carefully selected by G. Tucci: a huge area, between two streams, next to the main city of Buddhist Swat, probably the main monastery and the richest in that area. I also guess that the fact that the land lay unused and that the province headquarters were nearby played a role in the selection of this area. The site was excavated between 1956 and 1962, i.e., at a time when physical methods of analysis and paleo-sciences were not used widely in archaeology. Some luck befell the excavators. It was discovered that the main stupa was built in the third century B.C.; up to the tenth century A.D. it underwent five reconstructions, each new one encasing the previous. There was thus some chance of detecting changes in religious attitudes and of obtaining a relative chronology of Gandharan sculptures, building devices and decoration. But lying in low country, close to a growing modern town, the area was used as a source for building materials by the local contractors. While many Buddhist buildings still stand almost intact on the sides of the nearby hills, the Butkara I monuments were nearly razed to the ground. The loss in preserved elevations was more than compensated for by the gain in the building sequences.[61]

The excavations (c. 6,000 [m.sup.2]) were published in great detail: six volumes in-folio for the architecture alone (Faccenna 1980), with a great many drawings and photographs, and a wealth of tables and indexes. The plan of the publication is clear and very appropriate. We find a description of the main stupa, period by period, with an overall chronology; descriptions of the individual monuments, arranged in periods and typological classes, of floors, decoration and painting of the Sacred Precinct, and a few "urban" structures lying outside this Sacred Precinct. References are given whenever necessary, even when not so necessary. Reading and understanding are made easier by numerous drawings and by precise definitions and sketches of the technical vocabulary used by D. Faccenna, e.g., names of mouldings, etc. The chapter on the chronology of the monuments (vol. I, pp. 167-75) sums up the evidence in a very convincing way. The chronology depends on the stratigraphy and on the date of coins found in different layers and buildings. For absolute datings, D. Faccenna uses without a question mark the chronology of the Kushan kings elaborated by R. Gobl, who published the coins found during the excavations in an exemplary way (Gobl 1976). As everyone knows, this chronology is open to question, and careless readers who content themselves with looking at the table of contents (vol. I, pp. xix-xx) and the synoptic chart (vol. I, between pp. 174 and 175) must be warned about this peculiarity. Actually D. Faccenna's choice is no great problem, for in the text we are given all the necessary indications to fit the absolute Kushan chronology of our choice into the relative chronology of the Sacred Precinct.

I would go so far as to say that Butkara I is the best excavation ever made and published of a Gandharan Buddhist site. As a result, thanks to this very detailed report, we can use a wealth of reliable data for tracing the history of building techniques, decorative elements (mouldings, columns, etc.) and the painting of the Hellenized art of Gandhara. It is a wonder that nobody, up to now, has taken advantage of the evidence that is here provided in such a convenient way for further research. Faccenna 1980 is a welcome counterpart to Marshall's Taxila excavations, whose data, collected with less care,[62] are notoriously unreliable. For a student of Gandharan architecture and building techniques, Faccenna 1980 should definitely supplant Marshall 1951. I would like to add that, as a user, I am glad to be given an easy-to-use final report, with so many plans and large-scale drawings, color plates and long descriptions of each monument, however minute or damaged it might be, although more concise descriptions, and smaller-size drawings would have been sufficient in some instances.

Contrary to G. Tucci's expectations, apparently no evidence of religious changes was found. This does not mean that they did not take place, insofar as religious changes cannot be easily deduced from changes in building techniques.[63] Sculptures (infra) and inscriptions, which could have been used as evidence for such changes, do not exhibit anything decisive. In this instance luck was not with the excavator. But it is surprising that we are told nothing about the location of the monks' cells, and very little about the relationship between the Sacred Precinct and the adjacent town. From the panoramas published in Faccenna 1980, it seems that in 1956 the valley was not yet overbuilt. If it was really impossible at that time to discover any surface evidence of the location of the inhabited areas, it would have been better to say so in a few lines in the introduction.

At the end of the introduction, D. Faccenna indicates that "a study of the numerous finds the excavation has yielded up is not included in this volume. A remarkable collection of stuccoes and architectural and figured sculpture has been provided by the [Sacred Precinct] alone, while among the finds in the [Inhabited Area] pottery prevails.... Other categories of objects are made up of coins, figured terracottas, the contents of reliquary recesses (necklace beads, copper, silver, and gold items), and minor objects.... Obviously, it is only when all the material has been published - and here we wish to emphasize this fact - that the publication may be considered complete, and so at the present moment it is held to be still 'open'" (Faccenna 1980, I:3-4). In 1995, thirty-three years after the completion of the excavations, the publication is still open; these finds are only partly published in interim reports or specialized studies. The same can be said of the excavations of the small, but interesting, nearby hill monastery of Panr. The final report (Faccenna 1993) exhibits the same outstanding qualities as Faccenna 1980, but after thirty-one years, the publication is still "open." In this respect the final report of the third monastery excavated by our Italian colleagues in that area, the less important site of Saidu Sharif I (1977-82), represents a big improvement. Each find is recorded and published (Callieri 1989). But only one part of the excavations, the monks' living quarters, has been published. We still have to wait for the Sacred Area or stupa terrace, where most of the sculpture was found. Let us hope that the delay will be shorter than for Butkara I and Panr.

I especially regret the delay in publishing the sculptures. The bibliography devoted to the history of Gandharan art is tremendous and, alas, ever-growing. But for a few pieces (below), we have no sure evidence on which to establish a relative chronology of the sculptures, less so an absolute dating. Each date is dependent on the intuitive feeling of the researcher, and no two specialists feel things in the same way. The stratigraphy and relative locations of the Butkara I and Panr sculptures could be of great assistance in this respect. Starting from a relative chronology of the Swati sculptures, we could work out a relative chronology of the Gandharan and Taxilan finds. D. Faccenna is seemingly busy making a detailed study of the Butkara I finds, based on art-historical criteria and the location of the finds: "We have the possibility of determining several groups of sculptural productions marked by a close correspondence of (stylistic features) and of recognizing, within each group, several series of products of the same workshop.... On the basis of the discovery of some re-worked sculpture(s) these circumscribed groups can be arranged in their own internal and relative position(s). The slab of schist, worked on one site, is re-used later ... and worked on the back with another scene ..." (Peshawar Exh. 1982, 40-41). A catalogue and photographs of most of these sculptures were published long ago by M. Taddei (Faccenna [sic] 1961). Let us hope that D. Faccenna's study will now be published soon and will help to elaborate an overall chronology of Gandharan art free from subjectivity and prejudice. Scholars need no reminder of the importance of the Butkara I finds for this purpose. I recall that one of the major advances in our knowledge of early Gandharan art was the dating of an archaic series of sitting Buddhas by J. van Lohuizen de Leeuw from evidence gathered from the Butkara I excavations.[64]

One last observation should be made for those interested in the sociology of modern archaeology. On the cover of Faccenna 1980, despite the great number of drawings and their usefulness, only one name appears. On the cover of Faccenna 1993, seven names are printed.

The excavations at Sonkh, near Mathura, were conducted for eight years (1966-74) by a team of German archaeologists led by Prof. Hartel. With the objective of unravelling "structural remains to an extent sufficient for a reassessment of the antiquity of Mathura and the nature of early historical settlements in its environment" (Hartel 1993, 12), the relatively undisturbed mound of Sonkh (c. 20 m high) was selected. Two areas were fully excavated, down to the virgin soil, part of a village (4,800 [m.sup.2]) and a nearby low lying apsidal temple (c. 200 [m.sup.2]). It is obvious that the excavator was especially interested in the early periods (third century B.C.-fifth century A.D.) and that he would not have been averse to discovering cultic evidence and sculptures. But from the start the main dig, located on the mound, was intended to be a stratigraphic dig, unearthing each level on an area wide enough to get some idea of building techniques and shapes and village planning. Indeed, in each inhabited level excavated in the mound, four to ten living quarters ("blocks") and a number of roads or lanes could be traced. Forty levels, making eight periods, were unearthed, ranging from the eighth century B.C. to the eighteenth century A.D. The two excavated areas yielded a wealth of artifacts, among them some beautiful stone sculptures and two metal statuettes.

The publication is both detailed, precise and concise. It is divided in two parts: the citadel (mound) area and the low-lying apsidal temple.[65] After an introduction and a very useful and well-done summary of the results (table E, p. 17), the dig is described period by period, level by level, in concise style, well illustrated with good photographs and drawings in the text, so that there is no need to search for them at the end of the book. Then follows a chronology, with evidence adduced from radiocarbon data, ceramics and coins, four pages in all (pp. 85-88, 427), quite clear and giving all the necessary details without undue flourish. The finds are described in the same systematic way, category by category, with short descriptions following general introductions, and excellent photographs printed in the text, next to the description of the objects they refer to. An index closes the book. There is no study of plant and animal remains, hence no information on the crops and domestic herds nor on the villagers' diets.

What characterizes this report are its concision and its clarity, the one inseparable from the other, all made possible by a very clever use of lavish illustration. Much attention was given to the page layout, making this book very convenient to use. It is printed on a beautiful glazed paper, which makes the photographs still better and the text easy to read. The plans are small but drawn so that each detail and caption may be easily seen and it is convenient to have them in the text.[66] This is a perfectly produced book, up to usual Berlin standards.

People used to French reports will be surprised that Prof. Hartel did not avail himself of the results of the Mathura excavations to write a cultural history of northern India. Indeed, the catalogue of objects is conspicuous by its lack of references to finds made in other Gangetic sites or even outside India. References are only given to papers or books where the Sonkh object described was commented upon either by Prof. Hartel or by one of his (mainly Indian) colleagues. This final report looks like a series of "objective" descriptions without any other comments, save those needed for a better comprehension by the reader. In complete contrast with the French publications of the Ai Khanum or Shortughai digs and the eastern Bactria surveys, it is not meant to be a historical study made from new finds, but a tool for further historical studies.

Nevertheless, H. Hartel took advantage of this opportunity to make a number of concise statements, indeed a series of short, well-informed and carefully worded studies. They are scattered through in the book, where they look like summaries: chronology of the Mitra and Datta kings of Mathura (pp. 85-86); stylistic studies of the terra-cotta figurines (pp. 88-89, 100, 112, etc.); categorization of votive tanks (p. 195); characteristics of minor religious sculptures (pp. 245, 248, 281); function of the apsidal temples and evidence of a naga cult. This is obviously a choice in the name of "objectivity" which we should, in principle, applaud. I nevertheless feel that Prof. Hartel is the most qualified person to use the results of the Sonkh excavations for larger syntheses. I hope he will do it. But I can understand his point and the preference he gave to first publishing, as soon as possible, in a very precise way, the remains he unearthed, for the finds are impressive. The Sonkh final report will remain for a long time the main, if not the only, source of comparison for Gangetic pottery, terracotta figurines, and stone and metal objects. Nowhere in the available scientific literature are these artifacts so well dated, described and illustrated. The Sonkh excavations will also remain the standard reference work for the culture of the Mathura country for years to come. They will be the basis for studies on building history, material culture, domestic cults. When compared with similar artifacts found or dug out elsewhere, the Sonkh finds will enable us to trace the connections between Mathura and other parts of India and to characterize more precisely the westernization or internationalization of the Gangetic culture in the centuries circa the Common Era. And it is not indifferent to this purpose that the final report was published relatively fast and printed in a very handy way.

An archaeologist is never entirely free to choose his excavation strategy. He has to make concessions to the demands of his sponsors and to negotiate the permission to dig or survey with local authorities. But he can choose his publication strategy - ultra-academic like Rapin 1992 or seemingly ultra-matter-of-fact like Hartel 1993; ultra-detailed and even over-sized like Faccenna 1980 or ultra-modernist like Francfort 1989. But his main task is to publish a thorough final report, more for the benefit of the reader than for his own glory. The sooner, the better.

Notes

[1] We may assume that, as in almost every society, Indian society at that time was made up of a number of groups. But we do not know whether these groups were knitted into a fixed and hierarchized social frame bearing some resemblance to the modern Hindu jati system or whether there was only one system of ranking, incorporating the many groups of followers of Hinduism, Buddhism, Jainism, and Iranian religions into one and the same scheme. If we can judge from some stray inscriptions, social groups of foreign origin, especially the ruling clans, kept to quite unshastric family systems and rights of inheritance (Fussman 1980, 23). Brahmins could be engaged in trade and their widows could act as family heads (Fussman 1993, 115). This scanty data is enough to show that around the Common Era, Indian society was still far from functioning like the ideal Hindu society depicted in the probably much later dharmasastras.

[2] For Buddhism, see now the summary in Fussman 1994. For Vaishnava bhakti cults in Mathura, the best overview is still Srinivasan 1988, although dating back to 1981.

[3] The chronology of early Mathuran art still rests on the excellent, though now half-a-century-old, Ph.D. thesis by Lohuizen de Leeuw (1949), conveniently summed up by Sharma 1984, 173-242. For early Gandharan art, see Fussman 1987.

[4] Stray finds bought on the antique market are very often given a provenance which may or may not be true. Local dealers, as a rule, never give the true provenance of artifacts they sell either because they are intermediaries who have every reason to keep the origin of the artifact secret, not to be bypassed by other dealers, or because they got them from middlemen who have the same reason for not divulging its source. Scholars ignorant of the rules governing the antique market are very often misled by such provenances.

[5] Marshall 1951, xv-xvi.

[6] Wheeler 1955, 188.

[7] See Olivier 1996.

[8] Tucci 1958, 284.

[9] Hartel 1993, 9.

[10] Faccenna 1980, I:10.

[11] See above, n. 3. On the import of the discoveries in Butkara I for the early Buddhist art of Gandhara, see below, pp. 254-55.

[12] Bibliography of the Ai Khartum interim reports by P. Bernard in Guillaume 1983, xi-xii; of Butkara interim reports in Faccenna 1980, I:1-4 (footnotes); of Sonkh interim reports in Hartel 1993, 471-72.

[13] More often, now, "plaine de Dasht-i Qala."

[14] See below, n. 23, about Bernard 1985.

[15] It may be convenient for the reader to reproduce here the list of the Ai Khanum volumes in this series:

XXI: P. Bernard, Fouilles d'Ai Khanoum, I (Campagnes 1965, 1966, 1967, 1968), 1973 (quarto)

XXVI: O. Guillaume, Fouilles d'Ai Khanoum, II: Les propylees de la rue principale, 1983 (folio).

XXVII: H.-P. Francfort, Fouilles d'Ai Khanoum, III: Le sanctuaire du temple a redans, 2: Les trouvailles, 1984 (folio).

XXVIII: P. Bernard, Fouilles d'Ai Khanoum, IV: Les monnaies hors tresors; Questions d'histoire greco-bactrienne, 1985 (quarto).

XXIX: P. Leriche, Fouilles d'Ai Khanoum, V: Les remparts et les monuments associes, 1986 (folio).

XXX: S. Veuve, Fouilles d'Ai Khanoum, VI: Le gymnase; Architecture, ceramique, sculpture, 1987 (folio).

XXXI: O. Guillaume et A. Rougeulle, Fouilles d'Ai Khanoum, VII: Les petits objets, 1987 (folio)

XXXIII: Cl. Rapin, Fouilles d'Ai Khanoum, VIII: La tresorerie du palais hellenistique d'Ai Khanoum, 1992 (big quarto).

(Vol. XXXII is the second volume of the Surkh Kotal final report [1990] which has been ready for publication for years.)

[16] A study of the way Ai Khanum was provided with water is nevertheless announced in a note by Gentelle 1978, 142.

[17] Nevertheless see Bernard 1978b, 42.

[18] See the motivations behind the survey of the plain of Ai Khanum as related in Gardin 1976, 60.

[19] " ... un point d'appui militaire essentiel pour la defense des marches orientales de la Bactriane" (Bernard 1981, 109-10); "a military outpost ... to control the eastern marches of Bactria and block a potential invasion route ..." (Bernard 1982, 148).

[20] The discovery of the Harappan outpost of Shortughai and of Sogdian inscriptions on the banks of the Indus river show that there existed a direct link between Central Asia and northern India through the Kokca valley and Chitral. But we have no evidence that it was used in Greek times.

[21] Referred to in Bernard 1978b, 15: "L'organisation administrative et la securite etaient assures par une autorite dont le siege etait une place forte importante.... Entre les deux remparts circulaires qui entourent la citadelle proprement dire a pu egalement s'abriter une petite agglomeration urbaine, lieu d'echange pour les produits de la terre et ceux d'un artisanat local."

[22] Gardin 1976, 78-79, pl. xix-xxiii, who, using the evidence from these surface collections of ceramics, surmises that "in Persian times, the circular town was the main settlement in the Ai Khanum plain." Also Gentelle 1978, 12-13, and 31(c) for a map; Gentelle 1989, 144 and front and back covers for a satellite image (commented on in Gentelle 1978, 28). The same satellite image in color in Francfort 1989, opposite the title-page.

[23] The historical questions addressed in Bernard 1985 (Questions d'histoire greco-bactrienne) are: Greek Arachosia and the Mauryas (a problem solved in the main since Foucher's times); the era of Eucratides; the Branchides in Central Asia; the causes of the revolt of the Greek mercenaries in 323; Bactria and Menander's Samian; Euthydemos I and Magnesia on the Meander.

[24] See n. 15.

[25] The only evidence of an attack would be five arrowheads and two lance-heads found against the northern wall (Leriche 1986, 56-57). See below n. 27.

[26] For more details, see Bernard 1973, 110-11 and Gardin 1976, 61.

[27] These arrowheads will be published later (when?) by F. Grenet (Leriche 1986, 117). According to Francfort 1984, 54, I, who quotes Grenet, the five arrowheads found in the temple could be Saka. but the same Francfort rightly adds that arrowheads of the same shape were found in the Ai Khanum arsenal and even in Greece proper.

[28] A first study of the ceramics was published by Gardin in Bernard 1973, 121-88. At that time, apparently, no Kushan pottery had been discovered. A refined chronology of ceramics was made by B. Lyonnet and J.-Cl. Gardin in 1978: it was never published, but was used as a working hypothesis by most of the scholars involved in the publication of the Ai Khanum volumes. According to that document, the bulk of Ai Khanum ceramics should be dated between 330 and 125 B.C. (Leriche 1986, 105-6). There were also some Bronze Age ceramics, discovered in early burials, but there is no evidence whatever of Kushan pottery as far as I know. Indeed, it is a pity that the Greek and post-Greek ceramics of Ai Khanum have not yet been published in full, no more than the Kushan ceramics from Surkh Kotal and the Kushano-Sasanian ceramics from Kohna-Masjid (of which there is a manuscript by S. Veuve). This pottery was on display in the Kabul Museum depot, which J.-Cl. Gardin and his colleagues organized from 1978 to 1982, and this depot was worth more than any publication: to study pottery, you need to handle it. But the Kabul Museum has burnt down, and as the Soviets are not responsible, nobody cares. See now Lyonnet 1996 (below, p. 250).

[29] Bernard 1985. See n. 23.

[30] The only thing new is the "discovery," quite unsubstantiated in my opinion, of evidence of two invasions by nomads (Rapin 1992, 287-94), which actually comes from a hint by B. Lyonnet (see now Lyonnet 1996). See above, p. 247, and below, p. 251.

[31] E.g., Rapin 1992, 186, n. 655.

[32] The identification rests on the fact that this much damaged object depicted a king, standing on a chariot with two standing attendants (probably the charioteer and a woman), entering a forest or a pleasure grove, and that a terra-cotta medallion from Bhita depicts perhaps the Sakuntala story. But what is to be seen in the poor remains of the Ai Khartum plaque (now well fitted and otherwise perfectly studied) is consistent with so many stories we read in the Jatakas, in the Epics, the narratives, etc.: Indian literature is replete with stories relating how a king went hunting or went into an asrama or a pleasure grove.

[33] E.g., Lisitsina 1969 and Zejmal' 1969.

[34] B. Lyonnet (ceramologist), J.-Cl. Gardin and H.-P. Francfort (archaeologists), P. Gentelle (geographer).

[35] Gardin 1976, 59-63; Gentelle 1978, 3-7; Gardin 1978, 99-108; Gardin in Gentelle 1989, 11-15; Francfort 1989, 13-14.

[36] SPOT satellite imagery, which did not exist in 1974, would have helped much, but at a much greater cost.

[37] See below, p. 251.

[38] The surveys were made under especially favorable conditions: J.-C. Gardin had traveled in Afghanistan for twenty-five years; he spoke fluent Persian; he had very good local connections; he was the best specialist in Afghan ceramics; he could make use of detailed (1:10.000 and 1:50.000) maps and he was working in a sparsely populated country where large tracts of arable land lay little cultivated. But the former DAFA almost never took advantage of these conditions for making surveys and previous surveys by other scholars were much less successful (Gardin, in Gentelle 1989, 12-13). Nor are luck or favorable conditions responsible for the choice and training of the scholars who participated in this research, the methodology elaborated for tracing and dating the channels, and the way the field data were worked out for publication.

[39] Now announced as Gardin (forthcoming) (with a slightly different title in Gentelle 1989, 12).

[40] Gardin 1992, 100. Full bibliography on p. 103.

[41] See, e.g., below, n. 54. Such discrepancies are a necessary consequence of scientific research: if we could publish fast results of surveys or digs without further research and detailed studies, there would be no need of elaborate final reports.

[42] P. Gentelle, who is a professional geographer, was responsible for the biggest part of the field-work. He adapted and improved the methodology for tracing former irrigation channels. The collected artifacts were sorted out and dated by J.-Cl. Gardin and B. Lyonnet, who were at the same time busy studying the pottery of the Ai Khanum excavations.

[43] Gentelle 1978, 117-29. Gardin 1978, 103 2.3.c, expresses some disappointments at these specific results, apparently because it was not possible at that time to date with precision the surface remains of fields. But the last word has not been said: a successful trench (when and if it is ever possible to go back to Ai Khanum) or new physical methods may in the future help to date these remains which are the only evidence we have for making an estimate of the types of property and the size of the economic units.

[44] The main surveyors in this second phase were J.-Cl. Gardin and B. Lyonnet. The survey also included a patch of land on the left (western) side of the Qunduz river: Gentelle 1989, 102-5.

[45] Gardin 1978, 104-8. Gardin, in Gentelle 1989, 13-15.

[46] Lyonnet 1995. This is a Ph.D. thesis which I was able to consult before publication, thanks to Mme. Lyonnet. I am not able to refer to the pages of the printed publication and there may be some differences in content between the manuscript I read and the book now being printed.

[47] E.g., in the category "open-mouthed pots," a first subdivision is made between plates, dishes, bowls, etc., called groups. Then further distinctions (series) are made within each group according to specific characteristics of shape, decoration, etc.

[48] These samples do not claim to show every shape used during that period: shapes which were in continual use for more than one historical period were deliberately excluded from these synthetic tabulations.

[49] See above, n. 45.

[50] With the reservation that it does not include all the ceramics in use at one time: see n. 48. See also n. 28.

[51] Fussman 1993, 123-24.

[52] Obviously, some Saka tribes which invaded Bactria and Iran had belonged to the Sai confederation but there is no reason to equate entirely Saka with Sai or with Scythian: these tribes were even able to adopt a new language quickly, like the former Yuezhi (or some of them) who took up Bactrian.

[53] See already Gardin 1990.

[54] This contradicts a previous assessment: "on n'apercoit aucune recession au moment de la conquete nomade et de l'eviction des Grecs" (Gardin 1978, 140-41). See above, n. 41. I am told that J.-Cl. Gardin still adheres to his previous (1978) statement. All the necessary evidence should be given in his next volume (Gardin, forthcoming). Until it is published, all that can be said is that a decrease in cultivation after a conquest by nomad tribes, who need pasture for their horses and livestock, would be no surprise. But some reservations should be expressed: the decrease may be less than stated, because Hellenestic pottery was still in use years after the conquest and there are very few ceramic markers of the presence of the nomads. It might have started before the conquest, during the civil wars which contributed to the end of Greek dominion.

[55] Gardin 1979 and above, n. 40.

[56] See n. 47.

[57] For an overview of the work of the Italian Archaeological Mission to Swat, see Peshawar Exh. 1982, with bibliography.

[58] Faccenna 1980, I:10, n. 1. I am afraid it is still not ready.

[59] See above, p. 244.

[60] "Sacred Precinct" (or "Sacred Area") is an expression coined by our Italian colleagues to denote the stupa area of a monastery as distinguished from the monks' quarters.

[61] It is a pity that D. Faccenna could not use the evidence from the almost adjacent site of Butkara III. This site, which could not be detected from surface evidence, was discovered after heavy rains and flooding. It was excavated by a team from Peshawar University in 1982 and 1985 (Rahman 1990), well after the completion of Faccenna 1980. It was buried under a thick deposit of mud, brought down by a swollen rivulet, and boasts a number of well-preserved stupas, complete up to the umbrellas, with sculptures still in place, some of them clearly archaic (e.g., Rahman 1990, 702, fig. 8). We can deeply regret that it will not be published in the same detailed manner as Butkara I.

[62] This is not a censure of Marshall's work, which is eminently laudable. But between 1913 and 1956, the excavation techniques had changed much, and for the better.

[63] Comparing a Gandharan stupa with a Tibetan mchod-rten does not teach us anything about changes in conceptions.

[64] Lohuizen de Leeuw 1981, using Faccenna 1974, 172-75 and personal information from D. Faccenna. Slightly revised chronology (based on a new dating of the coins) in Fussman 1987, 69 [section] 6.

[65] Or apsidal temple no 2: apsidal temple no 1 was discovered in the citadel area (period V, i.e., Kushan).

[66] Hartel, 1993, 10, complains that "caused by the growing expenses for the publication, only reduced versions of the (original) plans of the levels could be published in the report. They were redrawn from reproductions of the original drawings (1:50)." I would dare say that it is more convenient for the reader to be given these reduced versions of the plans in the text.

Copyright © 1998-2011 The Circle of Ancient Iranian Studies (CAIS)