Consumption, exchange and production at the Great Settlement Shang: bone-working at Tiesanlu, Anyang.

Campbell, Roderick B. ; Li, Zhipeng ; He, Yuling 等

[ILLUSTRATION OMITTED]

Introduction

King's City (Chang 1985), pivot of the four quarters, centripetalizing ceremonial center (Wheatley 1971) are all names and descriptors for the Bronze Age mega-site located at Anyang, China, long the source of social evolutionary counter-narratives (Chang 1980, 1983, 1985, 1990). Shang civilisation, according to the reigning paradigm, was founded on political and religious rather than economic innovation and its economy, unlike that of West Asian polities, was under-developed and essentially tributary (Chang 1990; Liu & Chen 2003; Underhill & Fang 2004). Yet if one looks more closely at the empirical foundations of this paradigm it is clearly an edifice built on sand. Social and economic research has not been on the agenda of Chinese archaeology until very recently and it might be safer to say that outside of certain aspects of elite practice we know almost nothing of Shang production, exchange or consumption.

Bronze casting has been the flagship of ancient Chinese production study (indeed is virtually the only ship in the fleet). Claimed to be indexical of early Chinese civilisation (Bagley 1999), it was central to the elite symbolic economy and the ancestral- sacrificial complex, involving multiple, large-scale resource flows and sitting at the pinnacle of technology. The nature of this economy, moreover, is widely held to be redistributive and tributary, defined by inter-elite exchanges and close control of production (Liu & Chen 2003; Li 2005; Underhill & Fang 2004). This view, stemming from both elite- biased, post-Shang textual sources and anthropological models of wealth economies, creates the paradox that Shang China seems to have had a chiefly gift-exchange economy yet, considering the kilo-tonnage of cast bronze, on a scale precocious even for an 'early state' (Chang 1980). Despite this tendency to interpret the Central Plains Bronze Age economy in terms of this one industry, there is reason to believe that this is not the whole picture. Chen (2005), for instance, despite being a prominent proponent of the bronze elite redistributive production model, notes, based on stone tool crafting at Huizui, that there may have been separate economies for non-elite goods. Campbell (2007) has argued, based on epigraphic evidence, for expansive, ad hoc, as well as routine and intensive Shang economic networks, not all of which were directly controlled by the court.

The large bone-working areas discovered at Anyang present an interesting and understudied window into Shang production. Previous research (ZSKY 1987, 1994; ZSKYAG 1992; Meng & Xie 2006) suggests that their main products were hairpins, an artefact not necessarily destined for elites. On the other hand, the most common raw material used in this production, the limb bones of cattle, derived from an animal of immense royal and ritual importance. Moreover, the location and scale of the bone-working areas mimics the patterns of bronze-working at the site: massive, multiple, redundant production sites located within apparent production zones (Figure 1). These apparent contradictions and broad linkages between different facets of Shang society suggest that Anyang's bone-working areas may offer novel perspectives on Shang social economies.

Up till now, archaeologists have reconstructed Shang civilisation largely from mortuary deposits, or the afterlife of things. Production studies, on the other hand, afford an opportunity to focus on the birth of things and the effects engendered by their entry into the social matrix. While much of the literature on production has been centred on neo-evolutionary attempts to make production serve as a proxy for social complexity, more recently interest has shifted to such topics as production traditions (Lemonnier 1986; Feinman & Nicholas 2004), philosophies of technology (Sillar 1996; Boivin 2008) and ritual economies (McAnany & Wells 2008), at the same time that artefacts are increasingly seen as things with agency (Gell 1998; Latour 2005).

[FIGURE 1 OMITTED]

Bone-working at Anyang

The excavations at Tiesanlu in Anyang provided an opportunity to study intensively a vast assemblage of production debris and its varied depositional contexts. In 2006 a road repair salvage excavation revealed a large bone-working area consisting of hundreds of middens, 151 tombs, 12 structures and over 34 000kg of animal bone (Figure 2), most of it worked and all of it dating between Anyang phase II and phase IV (c. 1200-1050 BC) corresponding with the zenith of the Shang mega-centre (ZSKY 2003). Although this excavation produced one of the world's largest archaeological collections of worked bone, it was essentially a 10m wide trench through a much larger site and, in turn, is only one of at least three major bone-working areas at Anyang (ZSKY 1987, 1994; ZSKYAG 1992; Meng & Xie 2006) and probably not the largest (Li et al. forthcoming).

[FIGURE 2 OMITTED]

Due to the size of the assemblage, a dual scalar approach was adopted to obtain an overall understanding of the distribution of bone debris and finished artefacts over the entire site and to analyse intensively a single, deeply stratified excavation unit. The goals of the tine scale of analysis were to work out reduction paths, taxa, element, part and product ratios and their potential changes over time. After sorting bone from ceramic and lithic assemblages and weighing the bone according to context, the distribution of bone debris could be plotted across the site (Figure 3). Observation during this phase of work suggested that the assemblage was dominated by worked cattle limb bone fragments. The bone deposits occurred both in middens and unassociated with features and showed three major (units 1-4, 6, 8-11) and one minor concentration (units 15-19) extending approximately 200m north-south. An analysis of all finished and semi-finished artefacts collected during excavation (n = 571) showed that they also occurred in concentrations but not isomorphic with each other or the distribution of general bone debris.

Unit 5-especially stood out as an area with relatively little in the way of dump remains and a disproportionately high degree of finished and semi-finished products. It also contained portions of two building foundations and it is possible that unit 5 was an early phase (II-III) workshop area. Unit 5 included portions of 21 large and small middens, sections of two rammed earth structures, 18 tombs and two layers of deposit dating from Yinxu phase II-IV.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

The majority of the bone deposits were located in the early contexts (Layer 5 and midden H21) with phase IV middens (H1 & H7) next in terms of quantity (Figure 4). Of the 8153 fragments analysed, a high percentage showed evidence of sawing and/or grinding (42% by fragment; 51% by weight). Large bovines, deer and pigs comprise the majority of the mammalian taxa (Table 1) as well as the total assemblage, and were virtually the only taxa that were worked.

Medium mammals include fragments identified only to size and class: dogs, sheep and goats. Other includes tiger, buffalo, horse and human. The small mammal fragments include those identified only to size and class and small cervids. Deer and pig were almost entirely represented by head elements, over 99% of which was antler in the case of deer, and mandibles (remains of tusk extraction) in the case of pigs (Figure 5).

Cattle element distribution (Figures 5 & 6) shows an overall predominance of metapodials, especially metacarpals, reflecting a probable opportunity cost trade-off between food (grease and marrow) and bone-working utility. Thus, distal elements were favoured over proximal and forelimbs over hindlimbs. The cattle head elements were nearly all mandibles, apparently for spade production (Figure 7; Table 2).

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

Provision for sacrifice

Cattle, pigs and deer reached their points of consumption by different mechanisms, but especially via sacrifice. Pigs, prominent in domestic economies since the Neolithic in China, were likely raised locally, in close proximity to humans, consuming their refuse and in turn being consumed by them (Yuan & Flad 2002). The ritual use of pigs had gradually declined in the centuries before Anyang's construction (Yuan & Flad 2005) and their mention in oracle-bone divinations concerning royal sacrifice declined markedly after phase II (Campbell forthcoming). The Tiesanlu pigs then, need not have been involved in elite ritual and were possibly from a variety of local sources.

The deer, on the other hand, were hunted, some possibly at distant locales (Keightley 2000), either on a small scale or in the frequent large-scale hunts enjoyed by the Shang kings such as this oracle-bone records:

 Wu Wu day cracked, Ke Divined: We (should) hunt at Gui, (for if we
 do, we will capture (game). On this day (the King) hunted, and
 indeed captured (game). (We) [caught] tiger, one; deer, forty;
 "foxes", [two] hundred and sixty four; "antler-less deer" one
 hundred and fifty nine. (heji 10198)

While the Shang kings may not have held a hunting monopoly as some have claimed (Fiskesj6 2001), it is the case that wild animals make up a tiny proportion of residential faunal assemblages at Anyang but appear in great quantity and variety in the palace-temple area (Li 2009), suggesting that deer and other wild animals were largely consumed by elites.

[FIGURE 7 OMITTED]

The cattle, though domestic, were probably not all of local origin as the following oracle-bone inscription indicates:

 Yi Hai (day) cracked, (Diviner) Zheng divined: "(We should perform
 the) Yousacrifice (with) the cattle brought by the Wei Fang from
 (Ancestor) Shang Jia (on down). First Month. (heji 10084)

This inscription not only shows the king divining about a gift of cattle, but those sending them in are long-time enemies of the Shang court. What is more, aside from human captives, cattle were the kings' most frequently divined form of tribute (Campbell 2007, 2009) while the oracle-bones refer to lao (penned cattle), which may have been fattened on millet and became the most common sacrificial victim in the phase IV royal ritual. At the same time, a study of the inventory notations on the cattle scapulas used in royal divination suggest that they--sometimes in the form of live cattle--came to the Great Settlement as requisitions from nearby places securely under royal control (Campbell 2007, 2009).

 The potential scale of sacrifice is suggested by the following
 inscription:

 Divined: (In performing) the exorcism ritual, it should be cattle,
 three hundred (that we offer). (heji 300)

The symbolic significance of cattle is further indicated by its use as one of two media of royal divination with the royal consorts themselves sometimes preparing the scapulas, as in the following oracle-bone inventory notation:

Consort Jing prepared thirty (scapula). (Signed, Diviner) Zheng.

Nevertheless, recent work (Li 2009) has shown that beef consumption was widespread at Anyang suggesting massive redistributive mechanisms for royal sacrificial flesh--the above sacrifice would have produced around 75 000kg of meat--more localised consumption of lineage-owned cattle--perhaps also in sacrificial feast--or, most likely, both.

The bone industry: acquisition

The subsequent paths of pigs, deer and cattle partially converged at the TSL bone-working area. They only partially converge, because the element ratios above show that the animals were butchered elsewhere and only selected elements where taken to TSL. The relative quantity of unworked riders such as phalanges, carpals and tarsals suggests that, in the case of cattle, at least 5% of the time, articulated limbs were brought in. Since 15.5% of joint elements and phalanges show rodent gnawing, as opposed to 1.7% of the long bones, the riders were either exposed in a way that the worked bone was not prior to deposition, or they were deposited uncleaned. In either case, it appears that the cleaning and degreasing processes took place at the work site.

Looking at the issue of the kind of exchange that brought the bone crafters their raw materials, there are several lines of evidence to consider. First, the widespread use of cattle scapula in elite divination across Anyang (Flad 2008) and the paucity of non- oracle-bone cattle scapula at TSL, or other loci at Anyang, suggest that there were well- developed mechanisms for post-butchery bone distribution apart from the bone artefact industry. Second, although pigs were widespread, the large-scale use of antler and cattle bone may suggest a connection to royal practices of hunting and sacrifice. Nevertheless, the multiple, redundam distribution of large industrial areas around Anyang argues against direct, centralised management (Li 2005; Campbell 2007). Indeed, the current understanding of the Great Settlement as being composed of an amalgamation of lineage settlements (Tang 2004) and burial grounds would suggest that industrial areas or individual industries were likely in the hands of particular lineages, as later texts suggest, perhaps managed on behalf of the king. Moreover, there is evidence for domestic bone artefact production suggesting that access to bone was widespread and not monopolised. Given these things, it seems likely that there was a large centralised distribution mechanism for royal activity-related animal by-products, smaller ones for lesser lineages and ad hoc bone collection as well.

Production

The most remarkable aspect of the bone-working at TSL is its scale. The chief excavator of TSL, He Yuling, conservatively estimates the main bone-working area was approximately 17 600m2 (Figure 1) based on the 2002 and 2006 excavations. Estimating the number of cattle used at the site, unit 5 had a Minimum Number of Individuals (MNI) of 155 and contained 1% of the total 2006 assemblage by weight, yielding a total MNI of 15 500 for the 2006 excavation. Extrapolating over the estimated bone-working area yields (15 500/2400[m.sup.2]) 17 600[m.sup.2] = 113 667 head of cattle, or roughly two per day for the 150 years of bone-working activity. Using the element distributions worked out for unit 5 this would mean a minimum of 14 cattle bones to be cleaned, degreased, sawed into blanks, ground into shape then polished, perforated or carved with bronze and stone tools, not including antler or tusk working.

To understand how many artefacts could have been created from those bones, however, we must determine what was being created from them. Beginning with the 571 finished and semi-finished artefacts collected during excavation, perforators (including decorative pins and awls) dominate the assemblage (90%), as was expected based on previous studies of Anyang bone-working sites (ZSKY 1987, 1994; Figure 7; Table 2).

Nevertheless, factors such as ease of breakage during manufacture might create a bias in favour of more fragile or difficult pieces. Indeed, looking at the manufacturing debris for unit 5, (debitage 87%; blanks 6%; finishing 7%; finished <1%), and despite the comparatively large number of semi-finished artefacts, debitage still comprises the vast majority of the worked bone. Systematically comparing the debitage to the blanks, the most common reduction sequences for each element can be reconstructed (Figure 8).

Comparing blanks to semi-finished artefacts it was also possible to deduce which artefacts were produced from which raw materials and estimate how many could be produced per element, or in the case of antler, per unit equivalent (Figure 8; Table 2).

[FIGURE 8 OMITTED]

Combining this with the Minimum Number of Elements (MNE) for each element, a Maximum Production Estimate from Debitage (MPED) was calculated. To see how well these figures agreed with the deposition of artefacts at the site, bone artefacts recovered from 11 034[m.sup.2] of excavations at a variety of residential loci (ZSKY 1987, 2004) as well as 1424 lineage cemetery tombs (ZSKY 1979, 1987, 1998) at Anyang were tabulated and plotted against the MPED for unit 5 and the total finished and semi-finished artefacts recovered from TSL (Figure 9).

Looking at Figure 9, the MPED for perforators (the combined score for decorative pins/indefinite perforators/awls) is 65%, and though likely to be biased toward production sequences that produce large quantities of debris, it agrees well with the figure for perforators recovered at Anyang (decorative pins [45%] + awls [21%] = 66%), but is lower than the perforator score for TSL recovered artefacts (decorative pins [48%] + indeterminate perforators [35%] + awls [7%] = 90%). The high TSL figure for recovered perforators may be due to the relative fragility of decorative pins and thus the higher likelihood of breakage and discard during production. The high MPED measure for awls stems from the fact that the different forms of perforator are all produced from the same kind of blank and thus are not distinguished in the MPED calculation.

The difference between the TSL and Anyang awl scores is likely due to the fact that awls are also domestically produced and are less likely to break during manufacture than ornamental pins. The MPED and TSL under-representation of spatulas is likely related to the fact that when produced from ribs they leave little debitage and are unlikely to break during production. Tusk ornaments present the opposite scenario in being over-represented by the MPED. Since the MPED calculation was based on counts of nearly complete pig mandibles, however, it is likely to be accurate and means that either unit 5 is not representative of the TSL production as a whole or that tusks were used in contexts outside the scope of our Anyang residential and mortuary sample. The 210 pieces used as inlay on a lacquer vessel in royal tomb 1001 (Kao 1976), or the hundreds of tusk plaques used to decorate horse fittings in some chariot burials (Yue Hongbin, pers.comm.) gives some credence to the latter possibility. The discrepancy between the recovery of artefacts in the category of other at TSL and other sites at Anyang is possibly due to domestic production of some tool types, that small artefacts (such as beads) are unlikely to be recovered in non-mortuary contexts without screening and that many of the artefact types would not break as easily during production as ornamental pins, biasing their recovery at production sites.

[FIGURE 9 OMITTED]

Returning to the issue of production scale, perforator and arrowheads combined account for roughly 90% of cattle bone based production at TSL whether derived from MPED, TSL or Anyang recovered artefacts (98%, 97% & 79% respectively). If the remaining 10% of cattle bone production was comprised of shovels, tubes and spatulas then total TSL production from the 14 elements per day calculated above would yield: 0.9([14.sub.elements] x [6.sub.artefacts/element]) 0.1([14.sub.elements] x [1.sub.artefacts/element]) = 77 artefacts produced per day, or around ([77.sub.artefacts] x 54 [750.sub.days]) = 4 215 750 artefacts over the lifetime of the production area from cattle bone alone. How many people would have been required to produce that many artefacts is difficult to determine without replication experiments, but surely there were dozens of craftspeople at work at TSL throughout its operation.

The polished, symmetrical form of even the most ordinary decorative pins and the technical difficulty manifest in the finest testifies to the presence of skilled craftspeople, while tare, intricately carved and inlaid pieces found in some high elite tombs speaks to the existence of master bone-carvers somewhere at Anyang: perhaps at TSL. Nevertheless, the cleaning, degreasing, blank preparation and even initial grinding could have been done by less skilled workers.

If there were many craftspeople and at least some of them were highly skilled, then a further question might be, 'how were they organised?' For even if they were members of a common lineage, as archaeological and textual evidence suggests (Chang 1980; Tang 2004; Zhu 2004), one could just as easily imagine a collection of independent family producers as a single integrated workshop. While it will take further work to determine this issue, even at this preliminary stage there is some tantalising evidence in favour of specialisation, division of labour and overall integration. As noted above, there are large, non-isomorphic concentrations of debitage and semi-finished artefacts possibly indicating the existence of areas associated with different production stages. Production techniques were also relatively standardised with common reduction sequences for each bone element. Perhaps even more suggestive is the large-scale production of decorative pins in batches as in the 49 broken phase II bird-form pins found in unit 5, midden H21, and the 107 phase IV composite pinheads from unit 9, layer 5 (Figure 10). Not only do the large numbers of these pin wasters suggest even larger numbers of unflawed pins and pinheads produced in stylistically uniform batches, they show standardised production steps.

Ledderose (2001) has argued that Chinese crafts have a long tradition of 'modular production' characterised by specialisation, standardisation, the production of things in sets and the use of modules that can be combined in different ways. These combine to enable the mass-production of high quality objects such as the First Emperor's terracotta army. If 'modularity' is a characteristically Chinese 'technological philosophy' (Lemonnier 1986; Sillar 1996), then we might expect to find modular production in Shang bone- working as indeed is suggested in the large batches of matching decorative pins produced in standardised steps and the componential design of the phase IV pins, with the heads and stems possibly produced in different areas.

In terms of wider social context, the presence of sections of large, superimposed, rammed-earth courtyard structures in units 12-14 and 13-15, sandwiched between two concentrations of debitage indicates the presence of elites or perhaps public architecture--such as a lineage temple--associated with the bone-working area. Sections of nine smaller, less well-preserved foundations and subterranean structures in units 1-9 may have served as workshops or dwellings for the non-elite craftsmen, while the intermingled tombs were likely their final resting places. The mixture of residences and burials of different size and wealth is also characteristic of other Anyang locales interpreted as lineage settlements (ZSKY 2003; Tang 2004).

The image tentatively sketched is one of a hierarchical bone-crafting lineage, operating a large-scale, integrated workshop, churning out batches of perforators among other things, and of those, primarily decorative pins. Nevertheless, the questions of who the products were for and what paths they took after production remain to be answered.

[FIGURE 10 OMITTED]

Exchange and consumption

The Institute of Archaeology's 1994 synthesis of Anyang archaeology (ZSKY 1994) states that over 24 000 bone artefacts had then been recovered, mostly from elite tombs. Of these, 85% were arrowheads and the majority of these artefacts came from royal or high elite tombs such as the 672 spatulas and 6583 arrowheads from HPKM1001 (Kao 1976) or the 499 hairpins from royal consort Fu Hao's tomb (ZSKY 1980). The impression that the bone workshops were primarily provisioning elite tombs is, however, misleading. Despite the prodigious consumption of bone artefacts on display in some elite burials, if one considers that the survey of residential areas and lineage tombs showed a bone artefact density of (1769/11 034 [m.sup.2]) = 0.16/[m.sup.2] and (210/[1424.sub.tombs]) = 0.15/tomb respectively, extrapolating over Anyang's 30k[m.sup.2], there should be over five million bone artefacts outside elite mortuary contexts.

Considering that the royal tombs have already been excavated and probably the majority of the high elite tombs accounted for, elite mortuary deposition of bone artefacts can account for only a small fraction of the total bone-working output. Moreover, despite the vast estimated size of bone artefact consumption at Anyang, if we consider that TSL is only one of three major bone-working areas (Li et al. forthcoming), and if our back-of-the-envelope calculation of over four million artefacts for TSL is accurate, then bone workshop production outstripped local consumption, perhaps by as much as 300%. While such a massive surplus could possibly have been dispersed to non-local subordinates and allies in elite gifting, and while different artefacts may well have followed different pathways of exchange, the widespread distribution of most of the TSL products makes the possibility of circulation through other mechanisms such as markets or informal exchange seem far more likely.

Social lives of things

The products of TSL, dispersed along many different social paths, would have participated in a variety of different practices. Spatulas, for instance, are generally associated with high-status burials and were probably used in elite sacrificial feasting. Sometimes elaborately carved, inlaid with turquoise or even inscribed, the social lives of spatulas intersected and helped perform foodways, ritual, status and commemoration. Awls, spades and likely some of the non-decorative pins, on the other hand, usually found in midden contexts, were unadorned production tools. As they were sometimes produced domestically, it is unclear if the TSL tools were meant for royal workshops or for wider distribution. In any case, they were facilitators of farming, corvee labour, drudgery and craft. Bone arrowheads, like their bronze counterparts, were widely distributed in tombs and middens, and participated in the key, twinned practices of hunting and war (Keightley 2000; Fiskesjo 2001; Campbell 2007, 2009). Decorative pins, on the other hand, while not part of standard Shang burial assemblages, occasionally occur in large numbers in the tombs of elite women where matching sets were used in elaborate hairstyles or headdresses (ZSKY 1980). Nevertheless, decorative pins are more commonly found in tombs singularly and with individuals of all ages and sexes, while the vast majority were deposited widely in residential areas, discarded or forgotten.

These lines of evidence suggest that while the finest decorative pins were likely made for elite Shang women, most of those produced at TSL would have found their way into the hands of ordinary people. One of the few surviving elements of Shang adornment, along with beads, jades and vanished fabrics, they were the mass-produced constituents of aestheticised bodies and graded social being.

Conclusion

Though superficially a collection of worked bone debris, the TSL assemblage bears mute testimony to a node of networks linking and transforming animals, people and things near and far, across a bewildering number of social fields in late second-millennium BC north China. Although still preliminary, the work done on TSL suggests a large-scale and integrated bone-working site, perhaps organised on a similar template to the bronze workshop on which it borders. It likely acquired its raw materials from a variety of sources, but was at least indirectly connected to royal ritual activities.

Contextual information suggests the workshop was under the control of a hierarchical lineage, while diverse lines of evidence suggest that centralised redistribution mechanisms would have played a role in provisioning the workshop and likely also in redistributing some of its products. More surprisingly, the scale of production and its widespread distribution suggests that much of it was meant for non-elite and perhaps even non-local consumption. If the admittedly preliminary production calculations presented here are correct, moreover, then we must seriously consider the possibility of Shang markets and non-elite economies that may have eclipsed the ritual economy of the high elites. In short, we may need to rethink the dominant model of Shang economy and polity.

This rethinking will have to take a closer look at the biases built into both the sources and practices of Shang archaeology and history. Non-elites, quotidian production and trade have been largely ignored in Shang archaeology as lack of evidence has been conflated with lack of importance. It now seems quite likely that Anyang was more precocious than hitherto imagined, its economy more developed and that enormous social and economic change and intensification had occurred in the Bronze Age centuries before its supposed Iron Age advent.

Acknowledgements

We would like to thank the Luce/ACLS East Asian Archaeology initiative for generous post-doctoral funding of the eight-month preliminary study at the Anyang station, the Institute of Archaeology, Chinese Academy of Social Sciences, for its ongoing support, and all the people at the Anyang station for their help and generosity. Campbell would especially like to thank Sue Alcock and the Joukowsky Insitute of Archaeology and the Ancient World for providing an institutional home during this research. The authors would like to thank Rowan Flad, Jessica Rawson and three anonymous reviewers for their helpful comments on previous versions of this paper.

References

BAGLEY, R. 1999. Shang archaeology, in M. Loewe & E.L. Shaughnessy (ed.) The Cambridge history of ancient China: from the origins of civilization to 221 BC: 124-231. Cambridge: Cambridge University Press.

BOIVIN, N. 2008. Material cultures, material minds: the impact of things on human thought, society and evolution. Cambridge: Cambridge University Press.

CAMPBELL, R. 2007. Blood, flesh and bones: kinship and violence in the social economy of the late Shang. Unpublished PhD dissertation, Harvard University.

--2009. Towards a networks and boundaries approach to early complex polities: the late Shang case. Current Anthropology 50(6): 821-48.

--Forthcoming. Guo zhi da shi: Shangdai wanqi zhong de lizhi gailiang, in J. Tang (ed.) Yinxu kexuefajue 80 nian xueshu jinianhui. Beijing: Science Press.

CHANG, K.C. 1980. Shang civilization. New Haven (CT): "(ale University Press.

--1983. Art, myth and ritual: the path to political authority in ancient China. Cambridge (MA): Harvard University Press.

--1985. Guanyu Zhongguo chuqi 'chengshi' zhege gainian. Wenwu 2: 61-7.

--1990. Ancient China and its anthropological significance, in C.C. Lamberg-Karlovsky (ed.) Archaeological thought in America: 155-66. Cambridge: Cambridge University Press.

CHEN, X. 2005. Lithic production of early states in China: an examination of the development of craft specialization. Paper presented at the Workshop on Early Chinese Civilization, University of British Columbia, 10-12 March 2005.

FEINMAN, G. & L. NICHOLAS. 2004. Unravelling the pre-Hispanic highland Mesoamerican economy: production exchange and consumption in the Classic period Valley of Oaxaca, in G. Feinman & L. Nicholas (ed.) Archaeological perspectives on political economies: 167-88. Salt Lake City (UT): University of Utah Press.

FISKESJO, M. 2001. Rising from blood-stained fields: royal hunting and state formation in Shang China. Bulletin of the Museum of Far Eastern Antiquities 73: 48-191.

FLAD, R. 2008. Divination and power: a multiregional view of the development of oracle bone divination in early China. Current Anthropology 49(3): 4O3438.

GELL, A. 1998. Art and agency: an anthropological theory Oxford: Clarendon Press.

KAO, C-H. (ed.) 1976. Hou Chia Chuang. Volume 2: HPKM 1001. Taipei: Institute of History and Philology, Academia Sinica.

KEIGHTLEY, D. 2000. The ancestral landscape: time, space and community in late Shang China (China Research Monographs 53). Berkeley (CA): Center for Chinese Studies.

LATOUR, B. 2005. Reassembling the social: an introduction to actor-network theory. Oxford: Oxford University Press,

LEDDEROSE, L. 2001. Ten thousand things: module and mass production in Chinese art. Princeton (NJ): Princeton University Press.

LEMONIERRE, P. 1986. The study of material culture today: towards an anthropology of technical systems. Journal of Anthropological Archaeology 5: 147-86.

LI, Y-T. 2005. The social-political implications of multiple craft-production areas at Yinxu. Paper presented at the Workshop on Early Chinese Civilization, University of British Columbia, 10-12 March 2005.

LI, Z. 2009. The study of faunal remains from Anyang, the capital site of late Shang. Unpublished PhD dissertation, Chinese Academy of Social Sciences (in Chinese).

LI, Z., Y. HE & Y. JIANG (with R. CAMPBELL). Forthcoming. Yinxu Wanshang zhigu zuofang yu zhigu shougongye de yanjiu huigu yu zai tantao. Sandai Kaogu.

LIU, L. & X. CHEN. 2003. State formation in early China. London: Duckworth.

MCANANY, P. & C. WELLS (ed.). 2008. Dimensions of ritual economy (Research in Economic Anthropology 27). Bingley: Emerald Group.

MENG, X. & S. XIE. 2006. Yin Shang zhigu. Yindu Xuekan 3: 8-16.

SILLAR, B. 1996. The dead and the drying: techniques for transforming people and things in the Andes. Journal of Material Culture 1: 259-89.

TANG, J. 2004. The social organization of late Shang China: a mortuary perspective. Unpublished PhD dissertation, University College London.

UNDERHILL, A. & H. FANG. 2004. Early state economic systems in China, in G. Feinman & L. Nicholas (ed.) Archaeological perspectives on political economies: 129-44. Salt Lake City (UT): University of Utah Press.

WHEATLEY, P. 1971. The pivot of the four quarters: a preliminary enquiry into the origins and character of the ancient Chinese city. Edinburgh: Edinburgh University Press.

YUAN, J. & R. FLAD. 2002. Pig domestication in ancient China. Antiquity 76: 724-32.

--2005. New zooarchaeological evidence for changes in Shang dynasty animal sacrifice. Journal of Anthropological Archaeology 24(3): 252-70.

ZSKY (Zhongguo Shehui Kexue Kaogu Yanjiusuo). 1980. Yinxu Fuhao Mu [The tomb of Fu Hao at Yinxu]. Beijing: Wenwu chubanshe.

--1987. Yinxu fajue baogao 1958-1961 [Excavation of Yinxu 1958-1961]. Beijing: Wenwu chubanshe.

--1994. Yinxu de faxian yu yanjiu [Archaeology, excavation and researches in the Yin ruins]. Beijing: Kexue chubanshe.

--2003. Zhongguo kaoguxue: Xia, Shangjuan [Chinese archaeology: Xia and Shang]. Beijing: Zhongguo shehui kexue chubanshe.

ZSKYAG (Zhongguo Shehuikexue Kaogu Yanjiusuo Anyang Gongzuodui). 1992. 1986-1987 nian Anyang Huayuanzhuang nandi fajue baogao. Kaoguxuebao 1.

(1) Merton College, Oxford University, Merton Street, Oxford OX1 4JD, UK (Email: roderickbcampbell@gmail.com)

(2) Institute of Archaeology, Chinese Academy of Social Sciences, 27 Wangfujing St., Beijing 100710, PR China (Email: lizhp@cass.org.cn; yuanjing@cass.org.cn)

(3) Anyang Workstation, Institute of Archaeology, Chinese Academy of Social Sciences, West Xiaotun Village, Anyang, Henan 455000, PR China (Email: stationay@gmail.com)

Roderick B. Campbell (1), Zhipeng Li (2), Yuling He (3) & Yuan Jing (2)

Received: 7 February 2011; Accepted: 5 Apri12011; Revised: 18 April 2011

Table 1. Relative frequency of mammalian taxa recovered from unit
5 based on specimens identified and bone weight.

 Number of identified
Taxa specimens (NISP) Weight (g)

Bos sp. 810 9.9% 45 394 21.9%
Indeterminate large bovine 3362 41.3% 88 464 42.6%
Indeterminate large mammal 1749 21.5% 29 886 14.4%
Pig 579 7.1% 27 583 13.3%
Cervid 276 3.4% 5907 2.8%
Other medium mammal 334 4.1% 3821 1.8%
Other 34 0.4% 1046 0.5%
Small mammal 11 0.1% 65 <0.1%
Indeterminate 994 12.2% 5253 2.5%
Total 8149 100% 207 419 100%

Table 2. Raw materials and production estimates based on analysis of
semi-finished artefacts and debitage.

 Mean estimated artefact
Products recovered NISP Related debitage per element/unit

Perforators 513 Limb bones (90%) 6
 Antler (10%)

Arrowheads 42 Limb bones (55%) 6
 Antler (45%)

Spatulas 8 Ribs (75%) 1
 Humerus (25%)

Shovels 4 Cattle mandible 1

Tusk ornaments 0 Pig mandible 1

Tubes 1 Tibia 1

Plaques 1 Ribs 10