The Earliest Surviving Textiles in East Asia from Chertovy Vorota Cave, Primorye Province, Russian Far East

Article excerpt



The origin of textile production is among the most crucial issues in world archaeology (e.g. Good 2001; Gilligan 2010). It is clear that some types of textiles have been made and used by humans since the Upper Palaeolithic, about 26 500-32 000 BP (c. 30 000-37 000 cal BP) (e.g. Adovasio et al. 1996; Bar-Yosef et al. 2011). However, there are few actual finds of twisted fibres in pre-Neolithic contexts that may be considered direct evidence of textile production (Nadel et al. 1994; Bar-Yosef et al. 2011). Some of these fibres are strings, which may be among the earliest types of textile (Hardy 2008). In our definition of 'textile', we follow Adovasio (1996: 709): "... [it] includes not only flexible cloth with continuous-plane surfaces produced on frames or heddle looms (i.e. textiles proper), but also products as diverse as basketry, matting, bags, nettings, cordage, sandals, and related so-called perishable fiber artifacts."

The generally poor preservation of such perishable artefacts makes it difficult to study the earliest stages of their production and use. Most early textiles come from dry caves, cold regions and waterlogged sites (e.g. Good 2001). In Eurasia, the first evidence of well-developed textile production, represented by cordage, mats, baskets, knotted netting, fabrics and plain weaves, is known from Hahal Hemar Cave in the Levant and belongs to the Pre-Pottery Neolithic B (Bar-Yosef & Schick 1989), with several direct AMS [sup.14]C dates of c. 8500-9210 BP (c. 9010-11 190 cal BP) (Housley 1994). The Near East has the largest number of early textile finds from the Old World's Neolithic (e.g. Barber 1991).

In East Asia, direct traces of textile production in early periods are rarely encountered. It is suggested that some type of basketry was used in the Initial Neolithic of Japan and the Russian Far East (c. 9000-13 000 BP, or c. 10 500-15 000 cal BP; see Kuzmin 2003) based on impressions on potsherds (Zhushchikhovskaya 1997:160-62, 2005; see also Good 2001: 215), but actual remains of extant textiles are considerably later. In recent years direct AMS [sup.14]C dating of sub-milligram samples has made it possible to produce reliable results from even the smallest textiles (e.g. Jull & Burr 2006). In this paper we report calibrated AMS [sup.14]C dates of textiles from far eastern Russia in the tenth-ninth millennia cal BP and discuss their implications for the archaeology of East Asia and further afield.

Early textiles in East Asia: a brief overview

In Japan, almost all of the oldest textiles belong to the Early Jomon (e.g. Habu 2004). Although there may be still earlier textiles (Kobayashi 2005), there are no convincing data to confirm this. The oldest find of basketry is known from the Early Jomon site of Higashimyo (Kyushu Island; sec Figure 1) with a direct AMS [sup.14]C date of c. 6980 BP (c. 7710-7930 cal BP) (Nishida et al. 2006). Another example is from the Early Jomon component of the Torihama waterlogged site in western Honshu Island (e.g. Aikens & Higuchi 1982). Here plain-woven fabric (Ozeki 1996) is associated with [sup.14]C dates from the same cultural layer of c. 5500-6170 BP (c. 6210-7240 cal BP) (Keally & Muto 1982; Torihama Shellmound Research Group 1985). Several other Early Jomon sites on Honshu Island (such as Sannai Maryuama) contain evidence of textiles (e.g. Habu 2004:217-20).

In China, the earliest actual textile remains, tabby (plain) weave fabrics, are known from the Late Neolithic Liangzhu culture (e.g. Barnes 1999), with [sup.14]C dates of e. 4030-5870 BP (e. 4260-6930 cal BP) (Chang 1992a, 1992b). Non-perishable artefacts associated with weaving (thread and fabric beams, and shuttles) were recovered from the even earlier Middle Neolithic complex of Hemudu (Barnes 1999: 172), generally dated to c. 5740-6130 BP (c. 6320-7260 cal BP) (Chang 1992b). Information also exists concerning finds of textiles and weaving devices at the Caoxie Shan site closely related to the Hemudu complex (Wang 2001); it is not clear, however, whether actual pieces of loom were recovered or whether this conclusion is based solely on the presence of textiles (see Wang 2001:208-217). …