The RSS West Midlands local group held a meeting at the University of Warwick on 9 October 2014 with invited speaker Professor Andrew Bevan of the Institute of Archaeology at UCL, who gave a talk titled: 'Imperial logistics by numbers: the case of the Chinese terracotta warriors'.
Andrew began by discussing the importance of statistical methods in archaeology and some of the key statistical problems the discipline faces at the moment. Like in many other areas, technological advances in gathering, storing and sharing data have resulted in the creation of many new datasets which are rich in information but difficult to analyse. Andrew spoke about the increasing use of remote sensing, via satellite or plane, to analyse large geographic areas, by way of example.
The main topic of the talk was Andrew's work on the famous terracotta warriors, in collaboration with a team at the Emperor Qin Shihuang's Mausoleum Site Museum and others at UCL. The warriors are part of a large tomb complex for the emperor completed around the time of his death in 210 BC.
There are a number of ways that the warriors can be studied. One important area of enquiry concerns the degree of standardisation in the methods of production employed to create the warriors. Given its immense scale, learning about how the tomb complex was constructed is likely to be informative about large-scale production in general in Qin dynasty China. In particular, the weapons the warriors are equipped with are real and battle-ready, hence they were probably produced in the same way as weaponry for the real army.
Looking at the composition of the metal in crossbow bolts is one way evidence on this issue has been gathered. Andrew reported about how Marcos Martinón-Torres and the UCL/Xian team found that there was less variation between bolts found in the same bundle (corresponding to the quiver of one warrior), than between bolts found in different bundles. This suggests the bolts were produced in small batches by production cells working in parallel, rather than a single large production line.
Another approach developed by a UCL colleague, Janice Li, and the team was to conduct a shape analysis of the metal crossbow triggers and combine this with the spatial data recording where each trigger was found. By taking a variety of measurements of each trigger, they were able to identify clusters of similar triggers that were probably produced at the same place and time. These clusters appear to correspond to certain areas in the pit, suggesting that warriors in different places were equipped at different times, and possibly by different teams of workers, again supporting the idea of a cell-like organisation of labour.
More recent work has made use of the structure-from-motion and multiview-stereo technique whereby an accurate 3D model of an object can be constructed from a set of digital photographs taken of the object from many different angles. This technique has been applied to some of the warriors, allowing them to be analysed at a deeper level of detail than was previously possible. One intriguing new direction taken by Andrew and his collaborators has been to study the facial features of the warriors using these 3D models, and in particular their ears, to see what can be learned about the process by which sculptors added the facial details unique to each warrior.