Abstract
For coins, the chronology is currently being improved, following the liquidation of certain impossible theories on Kushan and post-Kushan coinage (Göbl, Zejmal' ) and the (very relative) refinement of knowledge on the small silver coinages of the Sogdian principalities up to the 6thcentury , but finds associated in stratigraphy with ramparts are rare for these periods. For ceramics, precision generally doesn't go below a century or a half-century, and even less so for the highly uninnovative period from the 2nd to the5th century , which imposes a looser grid than the political chronology. This unfortunate situation is particularly noticeable in Punjikent, where the partial destruction and rebuilding of ramparts, sometimes with temporary urban abandonment, followed a jerky rhythm throughout the 6th and 7thcenturies , with no precise knowledge of the city's political history at the time. In view of the overview of events presented in the preceding lectures, we would expect to find in Sogdiana the violent traces of three main invasions: the Chionites around 360, the Hephtalites around 480 (or perhaps only in 509, according to another interpretation of the Chinese data), the Turks in 556; moreover, we know that the population of Samarkand was rounded up by a Turkish qaghan in 640. It is possible that certain traces of destruction found at Punjikent and other sites correspond to one or other of these episodes, but we have no certainty.
A first stopping point: the Chionite episode of the 4thcentury . The archaeological impression around 400 (also the time of the sack of Rome by Alaric) is that the great cities inherited from Antiquity (Bactria, Samarkand, Merv) have regressed or are regressing in surface area, without however disappearing. The best-studied medium-sized cities (see the 2013-2014 lecture, https://journals.openedition.org/annuaire-cdf/11929) appear to have been partially abandoned (Dal'verzintepe, Dil'berdzhin, Termez, Toprak-Kala). In some cases, the ramparts are no longer maintained; in others, on the contrary, the urban function disappears and the site becomes a fort (Toprak-Kala). In some cases, all that remains inside are Buddhist monasteries (Dal'verzintepe, Termez) or Hinduized "local Zoroastrian" temples (Dil'berdzhin, where the temple has its own rampart). Abandoned areas were reoccupied by necropolises (Samarkand, Dal'verzintepe, Toprak-kala, Termez - in the Buddhist caves of Kara-Tepe, there were mass burials that were neither Zoroastrian nor Buddhist: an epidemic episode?). In eastern Tokharestān, at Chaqalak-Tepe, the only village excavated, the previously unfortified site received a double rampart and three fires marked the5th-6thcentury period.
