1Mr Administrator,
Your Excellencies, Mr Ambassadors,
My dear colleagues,
Ladies and gentlemen,

2The creation of a Chair of History and Cultures of Pre-Islamic Central Asia at the Collège de France, first initiated by Gérard Fussman and Jean Kellens, to whom I would like to express my gratitude, was not entirely innovative. A Chair with a fairly similar title, though without the “Pre-Islamic” specification, and held by Louis Hambis, had already existed from 1965 to 1977.

3The scientific object then referred to as Central Asia was very different to what it has since become. It corresponded to Paul Pelliot’s studies of the area, as part of the teaching he offered as Chair of Languages, History and Archaeology of Central Asia – but a Central Asia seen almost exclusively from the angle of what it had given to and received from China. The archaeology that could be drawn on at the time was almost exclusively one of Buddhism, from Afghanistan where the Délégation Archéologique Française en Afghanistan (French Archaeological Delegation in Afghanistan, DAFA) was working, and from Chinese Turkestan, explored by the early twentieth-century great European expeditions, including a French mission led by Pelliot himself. The results of Soviet excavations remained virtually unknown to Pelliot, and while Hambis had recognized their importance, he had no contact with the field. As for the written sources that were used, these were almost exclusively Chinese: chronicles and the accounts of Buddhist pilgrims, the only sources then thought to provide reliable chronological markers. Particular attention was also paid to thirteenth-century European travellers’ accounts, in which the Mongolian Empire was perceived as the great opening up of Eurasia. The emblematic undertaking in this respect was the publication of Marco Polo’s writings, with commentaries, to which Pelliot and then Hambis devoted much energy.

4The Collège de France’s Chair of Sinology had in fact always paid particular – and we could even say exclusive – attention to these land neighbours, whom the Chinese had called the “Barbarians of the West” before suffering, from 1840, other, more devastating invasions that arrived by sea on gunboats. In 1820, Jean-Pierre Abel-Rémusat, the first Chair of Chinese and Tartar-Manchu Languages and Literature, published Histoire de la ville de Khotan (History of the Town of Khotan), based on Chinese documents. A little later, Stanislas Julien produced the first translation of the pilgrim Xuanzang’s travel stories, one of the main sources on Central Asia and Northern India in the seventh century. Meanwhile in Russia, Nikita Bitchurin, a former Orthodox missionary in Peking who went under the name of Brother Hyacinth, published the collection of notes on Central Asia contained in Chinese chronicles. Since then our Russian-speaking colleagues have adopted this collection as a standard reference.1 Stanislas Julien and Brother Hyacinth regularly corresponded with each other, thus foreshadowing the friendships between Russian and French scholars that would later contribute substantially to the progress of Central Asian studies. By chance, I once came across correspondence in which they both lamented the fact that Marco Polo was then accessible only through Pauthier’s edition, considered to be a poor source. Pauthier was a scholar about whom Pelliot – severe as he was capable of being – later wrote that he had spent his life trying to learn Chinese without ever quite succeeding. In 1903, in Saint-Petersburg, Édouard Chavannes also published his admirable Documents sur les Tou-kiue (Turcs) occidentaux2, a work tool that was never outdated. Then came Pelliot, his disciple, the first French specialist on Central Asia to have worked in the field during his mission in Chinese Turkestan. His major contribution was however the collection of texts he brought back from Dunhuang. Of all those who drew from the jumble of the monk Wang’s manuscripts, he was the only one who had some grasp of the content of the manuscripts written in languages other than Chinese and Sanskrit. Thanks to Pelliot’s wise choices, the young Émile Benveniste, who was to teach comparative grammar at the Collège de France, was able to use the wealth of Sogdian material brought back to Paris, to take his first steps as a philologist of Iranian languages, introducing into Central Asian studies a culture that became a focus of attention much later. This future was also announced by Pelliot himself, who published a precursory article in 1916, entitled “Le ‘Cha Tcheou Tou Fou T’ou King’ et la colonie sogdienne de la région du Lob Nor”.3 It was only during the post-war years that scholars such as Edwin Pulleyblank reverted to the theme of Sogdian colonies. It is now at the heart of research on trans-Asian routes.

Figure 1. Tartary on Mercator’s Atlas.

5This reminder of what Central Asian studies owe to the Collège de France’s Chair of Sinology was necessary, as we are about to celebrate its bicentenary. At the same time, along with the contributions of the Chair of Indology, which likewise warrants acknowledgement, it teaches us about the approach that was taken to this cultural area for a long time, from the outside looking in. This approach was also adopted in the earliest historical geography research of Tomaschek and von Schwartz, which were based primarily on Alexander’s Greco-Latin historians, and that of Markwardt and Barthold, based on Arab, Persian, and Armenian geographies. Expressions specific to the peoples of Central Asia, whether literary or artistic, took a long time to escape from the grip of influences.

6Central Asia as a cultural area was recognized at a late stage, for its emergence as a geographical reality was very slow. Today still, when talking about Samarkand, even to an informed interlocutor, the inevitable comment “Samarkand, what a dreamy name” immediately raises questions about its belonging to recognisable geographic, political, and linguistic frameworks. The term Central Asia was coined in around 1825, simultaneously in Russia, by the political agent Georges de Meyendorff, and in France, by Julius Klaproth, a somewhat nefarious Sinologist. Klaproth, who was propelled by Jan Potocki, author of the Manuscrit trouvé à Saragosse, readily fabricated false travel stories though he lacked his mentor’s literary talent. This term soon prevailed, alongside its variant Middle Asia that is sometimes preferred in Russia. Reading this nineteenth-century geographic literature, it soon becomes apparent that behind rationalizations about climates and so-called natural borders lay a deep misunderstanding between the European vision and the Russian one. From the European point of view, Central Asia was defined as that which was neither Russia, nor China, nor Persia, nor what was becoming British India: it was a sort of in-between destined to receive buffer States, the only one ultimately being Afghanistan. The Russian perspective, on the contrary, had since Peter the Great seen it as a contiguous space, the natural continuation of the Russian-Siberian steppe, and which the empire of the tsars had set out to control, then annex, and finally colonize. In its various manifestations, this project was fuelled by strategic necessities, a sense of cultural superiority, or a great ideological design. In Pan-Slavic circles, it was informed by the dream of reconquering the cradle of the Aryan people, a theme currently resurging among some scholars of independent Tajikistan. The recognized fact that this space did not have limits which could be scientifically proven conveniently supported several Russian and then Soviet attempts to violate it, be it in Chinese Turkestan on several occasions, in Iranian Khorasan during the last war, or in Afghanistan more recently. Among scholars as well, the notion long remained geometrically variable. Louis Hambis, who worked extensively on Siberia and Mongolia, gave it a very broad meaning. In my teaching, I personally tend to refocus Central Asia within the narrower limits on which archaeologists have agreed for several decades. These encompass the five ex-soviet republics of Turkmenistan, Uzbekistan, Tajikistan, Kazakhstan, and Kirghizstan, along with Afghanistan and, additionally, an interest in the autonomous Chinese region of Xinjiang, especially during the periods where it shared strong cultural affinities with the countries to the west of the Pamir. However, I will remain open to studying Iran, particularly Sasanian Iran, a powerful neighbour, at times a partial conqueror, and a reference in certain domains, a point to which I will revert shortly. I will also not systematically observe the distinction between Pre-Islamic and classical Islam.

7It is worth considering first how this space was defined in the West before the invention, somewhat by default, of the term Central Asia. What it came to replace on our maps is actually the notion of Tartary, or Great Tartary (Fig. 1), in conformity with the historically accurate understanding of the whole area’s pre-colonial political structuring as dating back to the sharing of the Mongol Empire between the sons and grandsons of Genghis Khan. This terminological past has had its role to play in the way these people have been perceived in the long term. Tartary was the land of the Tartars, whom Saint Louis said had left Tartar and should be sent back. In view of many other historical events, the most recent one being 11 September 2001, one might be tempted to think that the people of Central Asia are somewhat like Léo Ferré’s anarchists: they are only noticed when they are feared. As early as 842 in Bagdad, Caliph al-Wathiq was concerned when he learnt that nomads were about to pass the Gog and Magog Wall mentioned in the Quran, and sent a reconnaissance mission there. On its return, the mission described one of the long walls that had indeed been erected by the sedentary peoples at the edge of the steppe. It is not actually known which wall this was, the Great Wall of China being only one of a number of possibilities. Earlier still, in the fifth century, the empire of the Hephthalite Huns whose aristocracy may have been related to Attila’s clan (Fig. 2) was formed in current Afghanistan and to its north. In 484, the Sasanian sovereign Pērōz went to fight them and died in battle with the best of his cavalry. Reading what a contemporary of these wars, the Armenian chronicler Lazar of Pharp, wrote about them, one would think he was describing the Scourge of God or the Mogul terror:

Even in times of peace, the mere sight or the mere mention of a Hephthalite terrified everyone. [When Pērōz entered their land], his troops advanced more like men sentenced to death than warriors marching to war.

Figure 2. The Hephthalite sovereign Javukha (c. 490).

8During the excavation of the site of Ancient Merv, now in Turkmenistan, which was the starting point of Pērōz’s campaign, Soviet archaeologists found the place where the contingents brought from the South Caspian mountains – which were to Iran’s armies what Switzerland was to those of the Renaissance – had spent their pay on the eve of the final battle. Copper coins issued by a workshop in their country had been left in a building with small cells, which our colleagues first thought was a Christian monastery before identifying it as a morally less commendable establishment.4

9This minor example at least has the merit of showing that archaeologists’ task today is as much to substantiate the accounts conveyed by ancient authors as it is to demystify them when needed. The archaeology of Central Asia is very young in comparison to others. Aside from the great explorations of Chinese Turkmenistan which, except for Mark Aurel Stein’s work, were mostly collections carried out without excavation methods worthy of the name, it mainly dates back to two schools that formed in the pre-war years: the French school of Afghan archaeology, and the Soviet school of the Central Asian republics. These two schools only converged during the 1980s, and have now merged to a large extent in the field.

10The Délégation Archéologique Française en Afghanistan (DAFA) was created in 1922 by the Indianist Alfred Foucher, and long retained the monopoly on excavations in that country. Its aim was both to study the great number of Buddhist monuments, and to rediscover archaeological traces left by the Greek kingdoms that succeeded Alexander, which were still known only by their currencies, including some of the most beautiful specimens of Hellenic coinage. While the first objective was largely met from the start and resulted in a series of remarkable publications, the second was punctuated with failures until this long quest was very richly rewarded by the discovery of the city of Ai-Khanoum in 1964. Though this city is still referred to with the site’s current name, there is now agreement that it is actually the ancient city of Eucratideia, the capital of Eucratides (171-144 bce), one of the last Greek kings of Bactria. In the meantime, the DAFA had seized the opportunity to shed new light on the Kushan Empire, which was contemporary to the Roman Empire. While at the time it was especially known to shelter Buddhist establishments, more diverse aspects of this empire surfaced on two occasions. The first was Joseph Hackin’s discovery, in 1937, of the Begram treasure, which was long taken for a merchant’s cache but is now recognized as a walled royal treasury. It was in any case a highly eclectic depot, a mix of glassware from Alexandria, sculpted ivory from India, and Chinese lacquers: categories of objects that are all represented here at the Musée Guimet. Then during the 1950s there was the excavation of Surkh Kotal’s temple by Daniel Schlumberger, enlightened by the first major known inscription in Bactrian – the Iranian language of Bactria written in Greek letters –, which revealed an official religious expression that was not Buddhist. Whereas initially it was thought to be a unique kind of dynastic fire temple, we now know that this building was the first to be known in a series of image temples also found in Sogdiana. They belonged to what can be called the Central Asian variant of Zoroastrianism, which differed from the West Iranian variant through a strong iconic component based on Greek and Indian models. The DAFA had actually already encountered this particular Zoroastrianism before Surkh Kotal: the careful examination of the painting that rose above one of Bamiyan’s great Buddhas, which was also subsequently destroyed by the Talibans’ idiotic fury, later allowed it to recognize Mithra, the Iranian sun god, on his chariot on the crest of the Hindu Kush at dawn.

11The DAFA archaeologists were a handful of enthusiasts working with very few resources and sometimes without much support from the French authorities. Soviet archaeology in Central Asia, more so than anywhere else in the USSR, was a State enterprise involving very large teams, albeit with heterogeneous methods and resources. Some of them were considerable. For instance, in Khorezm, the Chorasmia of ancient authors, the land south of the Aral Sea, excavator Sergej Tolstov had access to Red Army aeroplanes to take aerial pictures that allowed him to draw the first diachronic map of ancient irrigation systems ever produced on the scale of an entire country. This endeavour, unprecedented in the history of archaeology worldwide, was motivated by various factors. First, it was fuelled by the encyclopaedist spirit driving the young Soviet science, thanks to which almost all the archaeological expeditions were interdisciplinary, as shown for example by the Khorezm Archeological and Ethnographic Expedition’s very name. A sense of urgency was also at play, coupled with an awareness that the fossil landscapes, strewn with ancient castles that were still standing even though they were built of mud, were going to disappear under the effects of renewed irrigation (Fig. 3). Finally, very present though rarely openly expressed was a concern with restoring to these peoples a historical memory which was in no way indebted to Islam. All the regions of Central Asia, from the oases to the deserts and most remote valleys of the Pamir and the Altai, benefited from this relentless effort until the 1970s, when it began to run out of steam. Today still, no student can claim to specialize in the region’s archaeology without first immersing themselves in these already old publications from which crucial data are still drawn. Admittedly, along with its brilliant successes this scientific epic also had its grey areas: publications almost exclusively in Russian, a tendency to over-interpret religious remains or remains assumed to be so, limited access to foreign publications, as well as individual disasters, resulting in destroyed careers and the absence of publications on many excavations. Western archaeologists, myself included, long thought that the archaeologist profession was well protected in the USSR, providing an outlet for non-conformist or Jewish intellectuals, quite often both at the same time. Since the publication in Saint-Petersburg of Humans and Fates. Bibliographic Dictionary of the Orientalist Victims of Political Repression during the Soviet Era, this illusion no longer holds.5

Figure 3. The Pre-Islamic castle of Gyaur-Kala in Khorezm (2nd-1st century bce).

12I have just mentioned the isolation that long weighed on this archaeology. It was only partially interrupted by a few parthfinders such as Roman Ghirshman, a French archaeologist of Russian origin, who headed the DAFA for a while. Ghirshman published substantive abstracts of Tolstov’s books on Khorezm and to a large extent integrated the contributions of the Soviet urban excavations into his great book Iran. Parthes et Sassanides6, which was intended for both specialists and the general public. In 1965 the ice began to melt: for the first time, French and Soviet archaeologists worked side by side on the Ai-Khanoum excavation to comply with a political requirement by the Afghan authorities, who were rather well inspired on this occasion. Paul Bernard, who led the excavation, then asked all his young colleagues who wished to continue in this field to learn Russian, and encouraged them to undertake PhDs drawing mostly on the Soviet missions’ material. I was fortunate to be one of them. Having soon thereafter begun my trips to excavation sites in the USSR, which have since never ceased, I can attest to the interest that the discoveries of Ai-Khanoum aroused everywhere. No equivalent of this city has been preserved anywhere else in Central Asia. While its high cultural expressions are very Greek – the theatre, the gymnasium (Fig. 4), the stone carving – it is also specifically Bactrian in its modes of production and daily life. Perhaps also in its place in the religious geography, since it was established on the Ochus, a tributary of the Oxus, which Zoroastrian tradition identified with the “Good River” on the banks of which Zoroaster had received his revelation. Once the initial enthusiasm had subsided, Ai-Khanoum was of little interest to Hellenists, who were somewhat disappointed to find a cheap version of Greece and virtually no epigraphic material. These regrettable years of scant curiosity came to an end when economic inscriptions and a papyrus with fragments of a philosophical dialogue, perhaps one of Aristotle’s lost treatises, were discovered during the excavation of the palace. Pierre Hadot devoted a whole year of his seminar at the Collège de France to deciphering the philosophical papyrus. But these discoveries came too late, on the eve of the events in Afghanistan that forced the site to be abandoned before it was ultimately wiped out by the pillages.7

Figure 4. The Ai-Khanoum gymnasium (final state, mid-2nd century bce).

13The lost ground was soon recovered, though elsewhere. Owing to the perestroika, Soviet archaeologists could finally offer foreign colleagues what had always seemed unthinkable, even during the Tsarist era: opening up collaborative excavation sites on their home ground. In 1989, with the support of the CNRS and of the French Ministry of Foreign Affairs, I set up the Mission Archéologique Franco-soviétique de Sogdiane (Franco-Soviet Archaeological Mission of Sogdia) in Samarkand. Two years later, newly-independent Uzbekistan was intent on maintaining this openness, and the mission was renamed “Franco-Uzbek”. After the Afghan people’s fraternal qualities, we learnt to love those of the Uzbek people. Colleagues from Ai-Khanoum, former excavators from the Soviet era, young Uzbek archaeologists, and, increasingly, young students from various countries who come to train, continue to work together in this country, sharing methods and publishing jointly. Another mission initiated by the DAFA had preceded us in Tajikistan, where it is still excavating the key site of Sarazm dating back to the Chalcolithic era. French, European, Australian, Japanese, and, once again, Russian missions are working on other sites and following the same principles of transnational cooperation.

14The choice of Samarkand as the field of our second chance was informed by the renewal of already ancient knowledge, which fully bore its fruit only at that time. This renewal came with the discovery of the Sogdian civilization of Antiquity and the High Middle Ages, which was mainly linked to the excavation of the town of Panjikent started by the Ermitage museum just after the war, and which is still ongoing today. Since Antiquity, Sogdiana had been the name of the land of Samarkand and Bukhara, extending south until the Amu-darya, the Oxus of the Ancients. Panjikent, founded in the fifth century ce, 60 kilometres east of Samarkand and now situated in Tajik territory, has sometimes been compared to Pompeii. While this is a commonplace comparison in archaeology, it is in this case based on a few relevant analogies. This was also a medium-sized town, which may have accommodated 5,000 to 6,000 people at most, including its suburbs. An exceptionally high proportion of its inhabitants were aristocrats or even simply wealthy, partly because Panjikent escaped the 712 Arab conquest for 10 years longer than Samarkand, and because during those years it hosted a large number of aristocratic families from the capital, driven by a strong spirit of social competition and all intent on building themselves small palaces decorated with paintings. Like Pompeii, Panjikent was in a way sealed, not under volcano ashes but under the collapse of its mud walls after being rapidly abandoned by the first Islamized generation. Finally, like Pompeii, half of its surface is now being excavated, leading to a threshold effect in the possibilities of interpretation of the remains. I would add that this excavation has been successively led by two of the greatest minds of modern Russian Orientalism. The first was Aleksandr Belenitskij, initially an eminent specialist on Arabic and Persian, who urgently had had to change over to archaeology to escape the purge threatening the St Petersburg Orientalist School during the dying tremors of Stalinism. His initial training left him with an unmatched capacity to articulate iconographic documents to literary work. My friend Boris Marshak, who died at the excavation in 2006, had succeeded him. He had the ability to maintain this intellectual legacy by backing it with the most stringent methods of field archaeology and of formal analysis. All archaeologists of Central Asia knew that Panjikent was the reference. However it should be pointed out that Sogdiana’s major contribution to the history of the fifth to eighth centuries, its organizing role at all stages of trans-Asian trade, from China to Byzantium, from the Mongol steppe to the Himalayan passes, is hardly to be found in Panjikent, where the displayed values relate to aristocratic sociability and war prowess. Étienne de La Vaissière’s book, published at the Collège de France, recently shed decisive light on this history of the great trade, in large part thanks to written documents: merchants’ letters, Chinese customs documents, graffiti by caravaneers on their way to India discovered in the passes of the Upper Indus.

15Paradoxically, the discoveries on Sogdian civilization afforded Central Asian studies new opportunities to return to India and to China, which had been their starting point at the beginning of the last century. With regard to Northern India, the archaeology of Sogdiana and Bactria led to a reappraisal of the “Huns” and Hephthalites period, described in the same gloomy terms by contemporary Buddhists as by the Armenian chronicler I mentioned earlier. In the long term, this was actually a period of particularly fruitful interaction in the new common political space that had formed, overlapping India and Central Asia.8 Central Asia experienced a new wave of urbanization in medium-sized fortified towns, and the mural painting schools born in the Kushan Empire migrated to Sogdiana, where they produced an art that could illustrate the Zoroastrian pantheon as well as the Iranian epic, or tales drawn from multiple traditions. In Northern China, from the 1990s, the repeated discoveries of late sixth century tombs sculpted for established Sogdian merchants did not only revolutionize knowledge about the status of Central Asian elites in the Northern Zhou and Sui empires, they also provided the most detailed known images of Zoroastrian myths on the beyond.9 After long being reluctant to recognize these foreign components, our Chinese colleagues now fully accept them and are resorting to the expertise of specialists on Sogdiana.

16Much is now known about the Sogdians and about their neighbours from Bactria and Khorezm. Yet much remains unknown, and some of the new information we have is in itself contradictory, as well as being at odds with certain notions that were thought to be established facts. I will limit myself to reviewing a few of these paradoxes, all challenges which need to be addressed.

17A first paradox relates to power structures and the role of royalty. The Central Asia contemporary to Sasanian Iran was long considered as a sort of cultural extension thereof. We now know that while there is an element of truth to this, it relates almost only to the symbolics of power. The sovereigns of the various principalities of Central Asia presented themselves and most likely perceived themselves as equivalents of the Sasanian kings, particularly from the time that the Iranian royalty ceased to exist there. In 660, nine years after the assassination of last King of Kings who had escaped in Merv, the King of Samarkand, allied with the Tangs’ China, commissioned or inspired a cycle of paintings in a wealthy home of his capital. The paintings depicted him as a sovereign of universal peace presiding over the New Year festivities and receiving the embassies, in other words in the political public function which was quintessentially that of Sasanian kings.10 On the south wall of the room, precisely in the direction attributed to the King of Kings in the cosmic symbolics, he leads the procession to his ancestors’ mausoleum (Fig. 5). We know however that his dynasty lasted for the time that Sogdian dynasties generally lasted: no more than two or three generations. The faces on the painting were struck out during a violent episode linked to political change. At the time of the Umayyad conquest the whole area was split between rival principalities where the real power lay in the hands of Turkish tribal groups and land or trade oligarchies that were themselves divided between what Marshak compared to the Guelphs and Ghibellines, as a case in point the Arab side and the Chinese side.

Figure 5. Samarkand, Ambassadors’ Painting (c. 660).

18Many paradoxes are also found in the religious domain. It has traditionally been thought that Buddhism was mainly supported by the urban merchant class. This can perhaps be said of Bactria where it held an important though not hegemonic position. But in Sogdiana, the most mercantile of these societies, the exact opposite has been found: archaeological traces of Buddhism are very rare. The Chinese pilgrim Xuanzang, who passed through Samarkand in 630, noted that there were almost no more monks and that the last ones were being hunted down in their monasteries by Zoroastrian Zealots brandishing fires of purification. While Sogdian Buddhists certainly left abundant writings, these Buddhists existed virtually only in China, where they had converted. Though Zoroastrianism was in principle an element of continuity with Sasanian Iran where it was the religion of the king and the State, it took on a very different form in Central Asia, as I have already mentioned. There its most spectacular originality was the important role it gave to cult images. The religious art created in Iran was desperately poor, with all the inventiveness going to monarchical art. Central Asian Zoroastrianism was no less “authentic”, if such a term can have any meaning in the history of religion; it in any case originated from the region, informed by the same texts as in Iran and sharing some of the same rituals. Yet, in contrast, this Zoroastrianism never shied away from endorsing the iconographic models provided by Greek colonizers, and then by contact with the Roman Empire, and later Hinduism.11 Far from having been subserviently transposed, these models were reshaped to suit the figures of the Iranian pantheon to which they were ascribed, with a clear concern for completeness. Perhaps it will one day be confirmed that each of the 27 gods of the calendar had their own image. There were also original creations closely inspired by the texts. Finally, there were elements foreign to the Iranian religion, nevertheless integrated into the festive calendar: the Mesopotamian cult of Ishtar, called Nana in Central Asia, and a surprising discovery made in recent years, a cult of Demeter associated with the cult of Ishtar in seasonal celebrations. This transplantation of the mysteries of Eleusis can only be identified as a legacy of the Greek era, even if no Greek coin from Bactria includes the image of the associated divinities.

19Another important difference – a rather pleasant one in fact – from Sassanid Zoroastrianism, which often involved sinister demonstrations of intolerance, was that the Sogdian religion did not persecute, except Buddhism, probably because it had never received the full support of the State apparatus. It always had to accommodate the presence of other faiths. In Bactria some of these, like Buddhism and Hinduism, had once enjoyed the favour of the Kushan rulers, which nevertheless displayed virtually only Zoroastrian divinities on its coinage. Unlike in Iran, no names of Jewish, Christian or Manichean martyrs for Pre-Islamic Central Asia have been handed down to us. While this obviously does not prove that there never were any, it is clear that such persecution did not play a massive role that might have provided material for Martyrologues.

20Another paradox worth considering relates to the use of writing. Thanks to successive discoveries, the ancient peoples of Central Asia have started speaking to us in their own words again, not only in those of others. The Sogdian language, virtually extinct from the eleventh century, is now quite well known. I even have three colleagues who speak it, albeit with rather different pronunciations, the one English, the other Japanese, and the third Russian. As for Bactrian – the existence of which had previously been attested to by only a few inscriptions –, it was resuscitated from 1991 when a mass of parchment archives issued by a small mountain kingdom of Afghanistan between the fourth and eighth centuries emerged from a cave that had been reopened to be turned into an arms cache.12 Admittedly, these written documents did not provide us with everything we could have hoped for. While the Sogdian texts certainly contain a large number of literary texts, these are overwhelmingly translations of writings from the major missionary religions at work at the time on China’s roads: Buddhism, Nestorian Christianity, and Manichaeism. Yet the Chinese recorded that the Sogdians sang on the roads, accompanying themselves on musical instruments. If they sang, it means they had poems. We are still looking for these. The walls of Panjikent’s wealthy houses are covered with painted strip cartoons featuring epic scenes, as well as small pictures with condensed depictions of all sorts of apologues and fabliaux. Only one of the many identifiable subjects, an episode of the feats of the hero Rostam (Fig. 6), was found in a preserved piece of Sogdian literature. This literature also includes various tales transcribed by Manicheans, and inspired in turn by Aesop, the Bible, and the Indian Panchatantra on which part of La Fontaine’s Fables drew. While all these components are recognizable or can be guessed in the range of tales illustrated by the paintings, no direct correspondence has yet been identified.

Figure 6. Panjikent, detail of Rostam’s feats (c. 740): fight against a dragoness.

21When it comes to Bactrian writings, the situation is even more frustrating. A medievalist colleague to whom I was showing photographs of the parchments recently found exclaimed: “These look like medieval deeds!” to which I replied, “they are medieval deeds” (Fig. 7). Our medievalists are happy to have deeds, but they would be less so if the literature of the Middle Ages was limited to them, as is virtually the case in Bactria, with the exception of a few texts of personal correspondence. People’s names offer a glimpse, in flashes, of an underlying literary culture. A fifth-century seal conveyed the features of a young princess and her name, Scheherazade (Fig. 8): this is the only known Scheherazade before the Arabian Nights. Did she have something interesting to tell? We would like to imagine so. More seriously, another protagonist of the Bactrian documents was called Purlangzin, in other words “The Man in the Panther’s Skin”: this was the nickname of the hero Rostam, whom we just encountered in Panjikent and whom we find again in The Knight in the Tiger Skin of the Georgian epic.13

Figure 7. Bactrian sales deed (Document J) dated 518 ce.

Figure 8. Tsiurazad, the Bactrian Scheherazade (c. 400-430).

22We should trust future discoveries. Even if these literatures were for the most part oral only, there is a real probability that pieces of it will reappear, perhaps in the form of historiated rolls that supposedly served as models for Sogdian painters. There seems to be little hope that the same will apply to music, even if ancient instruments are increasingly better known.

23It is important to bear in mind that these discoveries, particularly those of texts and paintings on perishable media, will not be made through regular excavations so much as through the antiques market, as in the case of Bactrian archives. In Afghanistan, even though archaeological activity has resumed and the DAFA was reopened in 2002, the archaeologist profession is less profitable than that of clandestine excavation entrepreneurs. Though this profession has not appeared in the former Soviet republics, at least not to my knowledge, that of the archaeologist has become financially unappealing and there can be legitimate concern as to who will take over from the current generation, which was very well trained in the previous period. The tightening of borders between the new republics is hindering personal contact between researchers. The loss of expertise is already felt alarmingly in the restoration of historical buildings. But are we ourselves always best placed to be preaching? Amongst many other things, the Cabinet des Médailles of our Bibliothèque Nationale holds a great number of numismatic and artistic treasures from Iran and Central Asia, and as I have just seen for myself, significant discoveries can still be made there. Yet this Cabinet has now been inaccessible for four years, for lack of resources and sufficiently well informed interlocutors in the State administration.

24Let us rather turn to more positive signals that amply make up for the others. As I speak, excavations are once again highly active in all the countries concerned. In Xinjiang, where the authorities were long reluctant to allow in foreign archaeologists, the latter are now in the field, particularly a French team that is making remarkable findings in the ancient oases covered in sand South of the Taklamakan desert. Also encouraging is the fact that the previously highly uncertain chronology of the first centuries of the Common Era is now well established, virtually putting an end to debates that took up too much energy in the last century. The international enterprise Sylloge nummorum, launched by a joint initiative in Vienna and Paris, is endowing Iran and Central Asia’s numismatics with the best possible tools. The ceramics of all the eras has also received or is currently receiving its own tools, in the form of reference textbooks and pottery collections, particularly owing to the successive efforts of Boris Marshak, Jean-Claude Gardin, and Bertille Lyonnet. Finally, Central Asian archaeology is starting to catch up in the use of the physical sciences. A convincing climatological model was recently proposed, linking the collapse of the first Turkish Empire of Mongolia to a volcanic event that occurred in 626.14 Another will soon be proposed, likely to shed new light on the tremendous expansion of Genghis Khan’s Empire and its subsequent fragmentation. DNA research is also making its contribution, perhaps a little noisily with regard to Xinjiang, where some Western researchers are forcefully seeking to de-Sinicize the region’s past. They are exhibiting red-haired mummies, which are actually from widely diverse periods, just as ancient Kushan poets are portrayed virtually as Celtic bards with respect to superficial linguistic parallels. The Collège de France will welcome these innovations with discernment and openness, in the study days that I hope to organize with scientific colleagues.

25It is probably no coincidence that several cumulative book projects seeking to strengthen the notion of Central Asia are being launched around the same time. I am part of this, and I foresee that my first few years of teaching at the Collège de France will contribute to this synthetic work. In particular, I will suggest drawing up an inventory of urban archaeological excavations between the Iron Age, the era when the major oasis centres that have survived until now emerged, and the beginning of the Islamic era. For the Bronze Age, which experienced initial urbanization followed by a more or less general collapse, I will invite colleagues who are experts in that particular field – and there are certainly many in our country – to my seminar. Several major excavations are now published, others partially so and will not realistically be published any further, while yet others are in the early stages of the process. Each site has its scientific personality linked to its own history, to the unequal state of the documentation, and to the methods and assumptions of the various teams of archaeologists. It seems that the time has come to line up these distinct urban histories and to ask common questions: when and why were these cities born, and when and why did they die or move? Were there exchanges between cities? Did the building programmes discovered by archaeologists always serve major functions of urban life? If we take the case of Samarkand, the answer is by no means obvious. Like the Samarkand in Hugo Pratt’s La Maison dorée de Samarkand, mentioned but never drawn, the pre-Mongolian urban centre whose remains we are currently bringing to light gives something of the sense of a virtual city where constant undertakings have rarely been completed, aside from the city walls, because each successive power has wanted to make its mark and has neglected that of the predecessor.

26One of the purposes of my seminar will be to contribute to what I will call “documentary watch”. In Central Asia, probably more so than in many other geographical areas, it is important to know how to welcome and to manage the unexpected. If I consider my scientific production of the last 15 years, at the very least a quarter of my articles were not planned at all. Who could have foreseen, even just a few months earlier, the reopening of Afghanistan to archaeological research in 2002, after it had been closed for twenty years15? Soon afterwards, I received photographs of the only Sasanian relief ever discovered in the east of Iran, sculpted on a cliff North of Kabul. Fearing that it would be destroyed, I put everything on hold to set up a mission. We were not expecting to find a Sasanian relief, and were even less prepared for what its study revealed: a relief showing Shapur I, the vanquisher of the Romans, hunting a rhinoceros under a mango tree in a region where there were neither rhinoceroses nor mango trees (Fig. 9). Through this choice of motives, Shapur was inviting us to look to India, which he had begun to conquer.16 A similar experience occurred in 2006, when I received photographs of inscriptions found in Kazakhstan, which had been circulating clandestinely for some time and were thought to be Hebraic. Suspecting that they were Sogdian, I sent them to Nicholas Sims-Williams who immediately deciphered them, confirming what I had thought: these were the oldest known documents in that language. They were proclamations of military and agricultural colonization on the edges of the steppe, which could be dated to the second and early third centuries, a time when we knew the Sogdians only as caravaneers on the route to China.17 This prompted me to get in touch with the excavator of the site from which these inscriptions came and to extend my mission’s area of interest to Kazakhstan. As a last example, in the same year of 2006, our Australian colleagues working in Khorezm obtained confirmation that the site of Kazakly-yatkan where they were excavating, and which had not been covered by the former Soviet mission, had been the kingdom’s first capital in the first two or three centuries of the Common Era. This knowledge was based primarily on their discovery of mural paintings, in the study of which they associated me.18 Probably set in the context of the New Year celebration, the paintings featured horse processions that are comparable to those of Persepolis, and a gallery of royal portraits that could be ancestors. For these we have only one point of comparison in the Iranian area, which is strictly contemporary but geographically strictly opposite: the double gallery of Iranian and Greek ancestors at the Nimrud Dagh royal tomb in Eastern Turkey.

Figure 9. The Sassanid relief of Rag-i Bibi in Afghanistan (c. 250-270): Shapur I hunting the rhinoceros.

27As some fields open, others are of course at risk of closing, without any further notice. But historical interactions between all the countries of Central Asia as well as a necessary methodological feedback loop between the field, texts, and images offer many fall-back and reorientation possibilities. All researchers working on the area have at some point in their career experienced these constraints, which can ultimately prove more stimulating that settling too comfortably somewhere.

28In this presentation I have sought to show that Central Asia could be discussed leaving out a major term. You were surely expecting to hear this name, but did not: the Silk Route. I am certainly not trying to suggest that silk played no historical role, nor that the notion of a Silk Route coined in 1877 by the geographer Ferdinand von Richthofen did not play a major heuristic role in the development of our studies. But the term now seems a little worn out from being used everywhere and anyhow. It conceals the now well-recognized fact that for its Chinese producers, silk was not an object of profit but simply a currency used to pay civil servants and gratify foreign sovereigns, particularly the menacing nomads. The Sogdian merchants were the ones who harnessed it on the road and turned it into an economic object. Even from their point of view, it does not seem that it was always perceived as the core of their activity. Reading letters they exchanged and the records of the customs posts they went through, they could just as well have called themselves the masters of the musk road, or of the sandalwood road. In any case, the quantities were absolutely tiny compared to the modern trade and the products were light, with a price that was measured almost exclusively in terms of symbolic gratification. Even if long-distance trade, through the contacts it afforded, hugely contributed to the religious, literary, and artistic culture of Central Asia, we should let go of the idea that it ever constituted its main economic base. This base, which in some places dates back to the Copper Age or the Bronze Age, was the never-ending work of the canal diggers – I am tempted to say hydraulic engineers, so advanced was their expertise compared to that of modern technology. It was also very special plant selection and acclimatization engineering, particularly with fruit. At the time of Sogdian trade, the “golden peaches” of Samarkand had acquired legendary renown in China (Fig. 10), and much later the Great Mughal Babur, conquering India, confided in his Memoirs that he cried the day he was once again able to savour a melon from Kabul that was brought to him in ice. Archeobotanics now confirms these data long provided by literature and images alone. The identity of Central Asia was also defined by the raw earth itself in all the forms in which it can be used in building; by the arts of ceramics, of precious metals, and of semi-precious stones – though hardly of building stones –; by the arts of wood, now so scarce, of carpets, of drape, and of urban comfort in general, with the notable exception of bathing facilities that disappeared between the Greeks and the arrival of Islam. Rooted in a character more specific to the area, it was furthermore shaped by relations of trade and complementarity between a sedentary world and a nomadic one, forged on the outskirts of each oasis and sometimes within them. In the long run, these relations mattered far more than hostile events, before the major rupture brought on by the Mongol invasion in 1220, causing terrible population loss detected everywhere in the field, with consequences comparable to the Great Plague of the West, which was most likely its distant consequence.

Figure 10. Panjikent, peaches and melons in a banquet scene (c. 740).

29Many young teams, which every year produce new material and experiment with new approaches, are now taking up this research. Far beyond the work I am carrying out and will continue to carry out with my colleagues in Samarkand and elsewhere, the Chair to which I have had the honour of being elected by the Collège de France will host this research, foster dialogue within it, and support it.