The dynamics of wealth and poverty in the Trans-egalitarian societies of Southeast Asia

Brian Hayden*

Antiquity; Sep 2001; 75, 289

* Archaeology Department, Simon Fraser University, Burnaby BC, Canada V5A 1S6. Received 5 October 2000, accepted 21 February 2001, revised 31 May 2001 ANTIQUITY 75 (2001): 571-81

Understanding how differential wealth develops between households in villages is one key to understanding how socioeconomic inequalities develop; a key theoretical issue for archaeologists. An ethnoarchaeological perspective in Southeast Asia gains major insights into how socioeconomic inequalities emerged and were maintained, with numerous implications for other cereal and stock based prehistoric cultures, such asthose of Neolithic Europe.

Key-words: vveallh, feasting, transegalitarian, Southeast Asia, domestic: animals

Introduction

The goal of this article is to describe a major sector of the wealth-producing systems of tribal Southeast Asia and to understand the major constraints in wealth production and accumu¬lation. Four domains exist where wealth is gen¬erated in traditional, subsistence and agriculture-based economies: agriculture, do¬mestic animals, regional exchange and feast¬ing. Feasting and agriculture are analysed elsewhere (see papers in Dietler & Hayden 2001), while the mechanics and dynamics of regional exchange have been studied since Malinowski (1961). In contrast, there is much less informa¬tion available on aspects of animal domestica¬tion that generate inequalities. Therefore, in this article, I will focus on domestic animals as a source of wealth.

Domestic animals provide an instructive discussion of wealth for several reasons. First, from a Southeast Asian emic viewpoint, do¬mestic animals are strongly identified with wealth. Second, in Southeast Asia they play critical roles in the entire feasting complex upon which considerable socioeconomic and political inequality is based (see Clarke 1998; 2001). Third, the specific and detailed constraints on household animal production in Southeast Asia are very poorly understood by archaeologists. And fourth, traditional animal-raising strate¬gies seem uniform throughout most of tribal Southeast Asia.

Raising and using animals

I have argued previously (Hayden 1990; 1992; 1995) that animals were domesticated prima¬rily to increase fat content and desirability (and availability) for feasting. Perhaps nowhere else in the world is the importance of domesticated animals for feasting more evident than in South¬east Asia, although I suspect transegalitarian uses of animals follow a similar pattern almost everywhere (e.g. Keswani 1994). Indeed, in tribal Southeast Asia, often the only traditional use of domestic animals (cattle, buffalo, pigs, ducks, and chickens) is for feasting and sacrifices (water buffalo were also used for traction where paddy fields existed). Otherwise, meat from domes¬tic animals simply does not seem to have been consumed. Keeping pigs in Oceanic societies has even been called a 'luxury occupation' or 'ostentatious waste' (see Strathern 1971: 131); while in Thailand, all hill-tribe domestic animals were traditionally under-utilized for work func¬tions or secondary products, and they were viewed as a 'supplement to subsistence' (Kunstadter 1978: 105). However, acquiring economic, social and political benefits from feasting, and being suc¬cessful at feasting, was highly dependent on success in raising animals (Strathern 1971:134-5; Clarke 1998; 2001).

In addition, there appears to be a general consensus that medium and large domestic animals were used as a means of providing insurance against bad crop years or other mis- fortunes. Surpluses were used to raise pigs or care for cattle which could be exchanged or eaten in times of need (Strathern 1971: 131). Raising live stock was viewed like having a bank account (Falvey 1977: 22-3, 38, 40, 86; Shubert 1986: 81). However, unlike modern banking, there were major risks and possible liabilities in the form of damage compensations entailed in the investments in animals. In this respect and others, a better analogy for raising domes- tic animals would be like investing in the 'stock' market rather than a bank.

During fieldwork in northern Thailand and Vietnam, I was confronted by a number of critical problems in trying to understand why poor families did not seem able to increase their production of domestic animals. The main dif¬ficulty in modelling wealth production arose from observations that cattle, water buffalo and pigs (the mainstays of the feasting complex) largely foraged for themselves in forests. They apparently did not require much, if any, fod¬der or care from their human owners. If this was true, how did they really differ from wild animals? What limited the animal populations within village territories? How was ownership exerted over animals? How were they retrieved from forests when needed? And why did more poor households not raise these animals?

Answers to these questions proved difficult to obtain given the abstract nature of the con¬cepts involved, problems in communication and substantial changes in traditional economies and land tenure everywhere in Southeast Asia in the last three decades. Initially, there seemed to be no obvious constraints on animal food resources (or at least no way of controlling who had access to grazing lands), no labour con¬straints and no major capital requirements ex¬cept modest ones related to the largest domestic animals. With almost no apparent constraints, it was a mystery why more people did not raise animals. Several regional ethnographers sug¬gested that the poor were poor simply because they were not interested in working to acquire wealth; the poor were not interested in feast¬ing and the extra work that it entailed; or that the poor were not even interested in raising animals using minimal extra efforts. Other fac¬tors that I thought might limit the production of animals included high risks of animal losses, costs associated with damages animals might cause, care costs such as tending animals or building shelters, start-up costs, hidden labour or feeding costs, and drug or alcohol addictions of household members.

The magnitude of inequality and ownership

According to expressed emic views, there are no rich or poor households in most tribal com¬munities. Everyone is supposed to be the same. Yet there are very apparent differences between households that even casual observers can note: differences in house size, house quality (e.g. dirt floors vs raised pilings), house contents and quality of clothes and jewellery. In addition, Clarke (1998) has documented major differences between households in the; magnitude and fre¬quency of hosted feasts. Numerous other au¬thors have noted fairly substantial differences between households in wealth levels (Alting von Gesau 1983; Shubert 1986: 55ff). The Shubert study reports that 44% of all house¬holds are in debt, i.e. cannot sustain themselves by their own productive efforts. Chamberlain (1996: 31—2) also reports rice deficits as com¬mon among the hill tribes of central Laos, where rice is the preferred staple. Moreover, while most households had eaten meat (most probably in a feast) in the previous 10-20 days, 10% of the households had not had any meat for over 30 days, indicating some substantial nutritional differences between poor and well-off house¬holds. As well, pig fat (a key component for traditional cooking) was absent in 28% of the households in his survey.

Because medium and large domestic animals probably constitute the major form of wealth in tribal Southeast Asia, it is especially instruc¬tive to examine data concerning animal own¬ership. Falvey (1977: 27-36, 39) provides the most detailed data from a very large sample of ethnically varied communities. His data are particularly interesting because they were col¬lected before major economic change. In gen¬eral, only 30-50% of households in most communities owned stock (FIGURES 1 & 2). Stock owners were more wealthy than non-stock owners. While median household herd size was less than 5, 30% of the village herds were gen¬erally owned by a single household, usually that of the village headman or other influen¬tial people (FIGURE 3). Village herd sizes usu¬ally varied between 40-100 head. In fact, Falvey (1977: 23, 53) noted that the village headman's house could often be identified from the high density of manure surrounding it, a potentially important detail for archaeological investiga¬tions. He also notes a much smaller propor¬tion of poor households to rich or moderately well-off households than is the case for more stratified societies (FIGURE 4), as well as a much higher proportion of 'rich' (i.e. stock owning) households (Falvey 1977: 41), although there is still a large gap between the average 'rich' households and the richest of the rich. This is a pattern that seems to be prevalent through¬out tribal Southeast Asia (e.g. Leach 1954; Condominas 1977; Clarke 1998). I suggest that Falvey's portrayal of the inverted pyramid dis¬tribution of wealth resources is probably typi¬cal of transegalitarian wealth distributions in general (Hayden 1997).

Given a general distribution of animals such as Falvey describes, why are there animal-poor and animal-rich households? One of the main issues that must be addressed is whether some individuals have decided voluntarily to opt out of participating in a surplus-demanding, time-consuming, traditional ritual and feasting com¬plex, or whether they have been manipulated and manoeuvered into a marginalized socio-economic position by more powerful members of their community. The suggestion must also be considered that dependence on alcohol or opium may account for the lack of motivation in poor households.

While alcohol and opium dependency cer¬tainly do account for many cases of poverty, Cooper (1984: xix-xx) has noted that many wealthy households also use opium regularly and that opium addiction is usually not a prob¬lem unless the household is poor. Aside from addictions, it is difficult to imagine many peo¬ple willing to renounce their economic and political stakes in a community by opting out of the traditional feasting complex. The resulting marginalization would simply be too risky and costly, potentially leading to high fines, indebt¬edness, forced use of poor quality land, inabil¬ity to marry and even enslavement or death. For instance, Condominas (1977: 94, 338-9) describes how one woman's father had bought a large quantity of fish from neighbouring peo¬ple. Her father had paid for the fish with a bill¬hook machete. But when her father died, the neighbouring vendors claimed that they had never received the machete. The administra¬tor and judges of the woman's village agreed to pay their neighbour a sabre, vest, and tur¬ban to settle the dispute, writh most of the costs being born by the head of the lineage. The lin¬eage head then demanded reimbursement from the widow who had no wealth. The widow, as well as her three young children and her young sister, were unable to pay these inflated sums and were therefore sold into slavery for a total of 10 gongs and 24 water buffaloes — recom¬pense out of all proportion to the original debt that was claimed. Thus, opting out of the feasting and support network may leave one open to socioeconomic predation.

As we shall see shortly, beside the risks of not being able to repay debts, there are several other more practical constraints on animal rais¬ing that clearly do exist and that have the ef¬fect of increasing the cost of animal breeding well above the level most poor households could cope with. These constraints may have been intentionally manipulated (and undoubtedly were so manipulated in some cases) by the more powerful members of the community for their own advantage (see Condominas 1977). Thus, while there are undoubtedly a few individuals who genuinely lack motivation or aspiration to defend their own interests in the commu¬nity or to participate in the feasting-social com¬plex, and while addictions certainly affect some households' ability to produce wealth, it is also apparent that there are very real practical con¬straints that prevent many interested and as¬piring households from becoming fully enfranchised in their communities. Let us dis¬cuss these constraints.

Forage and feed

Impressionistically, it always appears that there are far fewer cattle and buffalo than might be supported in village forested areas and regen¬erating forest of old swidden plots. This im¬pression is reinforced by the observation that forage rights outside actively farmed swiddens never seem to be restricted or controlled in any way within a village's territory. Quantitative data also seem to confirm this assessment. Falvey (1977: 55) uses an estimate of 15 ha (a range of 7-25 ha) of forest forage required for every free-range head of cattle or buffalo. Theera Visitpanich (pers. comm.) uses a similar esti¬mate of about 20 ha per head (about 50 head per square kilometre). These estimates compare favourably with the 12 ha per head of decidu¬ous forest that Gregg (1988:106) uses for Neo¬lithic Europe. In Southeast Asia, the more grass [Imperata) graze that is available, either natu¬rally or from cultural effects, the higher the density and herd size a village will be able to maintain, reaching densities of up to one head per 1-5 ha. Thus, Falvey (1977:107) reports that (presumably rich) villagers often oppose refor- estation programmes in order to retain as much grazing land as possible for their stock. If we assume a mix of forest and grass (Gregg assumes a 3:1 mix) and a village spacing of 5 km or so (creating village areas of about 7 sq. km], it is apparent that most villages should be able to sustain herds many times larger than they, in fact, do. Therefore, what is the limiting factor in herd size?

Cattle and water buffalo are not fattened or fed, although they may be given minimal amounts of feed. The general attitude is that cattle should take care of themselves (Falvey 1977: 67). Cattle are periodically given salt. While this is done more to control animals than for nutritional purposes, adding salt to rumi¬nant diets can increase production by as much as 30% (Falvey 1977: 52; pers. comm. 1998).

Treatment of pigs is very similar to stock, with a few important differences. Pigs are gen¬erally raised on a free-range basis and they obtain much of their food by themselves in the sur¬rounding forest. However, unlike stock, there are three important consequences if pigs are allowed to forage entirely for themselves. First, pigs are much more difficult to keep out of gar¬dens. Since pigs generally only forage in a daily radius of 2—3 km, if they can be 'tethered' in one way or another to the village, fields can be placed farther than 3 km from the village, ef¬fectively beyond the foraging range of pigs. Second, if pigs are not given any food supple¬ment to their forest forage, they only grow at about half the rate of pigs that are regularly fed (it takes 2-5 years for free-range pigs to reach the same weight as regularly fed pigs reach in 9-12 months — Visitpanich & Falvey 1980: 264). In turn, the longer the growth period, the higher the risk of losing the pig to disease (a major recurring source of loss) as well as having to use an undersized pig for a ritual feast prema¬turely. A third consequence of letting pigs for¬age entirely by themselves is that pig nutrition is so poor that mortality rates increase substan¬tially and fertility may be adversely affected as well (Visitpanich pers. comm.).

Because of all these concerns, pigs are gen¬erally fed food supplements (mainly rice bran, maize and banana stalk) which require crush¬ing or boiling (Visitpanich & Falvey 1980: 263). By feeding each pig about 0-5 kg of maize each day for 3-6 months (i.e. as long as surpluses generally last), a pig will attain 60 kg in a year, vs only 30 kg without supplements. The amount given generally increases to 2-3 kg per day during the month before slaughtering. This fattening frequently takes place between the harvest and the tribal New Year celebrations, a time when crop surpluses are most abundant and pigs are in most demand for feasting. By feeding pigs daily, they return to their owner's houses every night. Without feeding, it would be impossible to get pigs to return to the vil¬lages and they would become feral, creating considerable crop damage (Visitpanich pers. comm.; Bounserm Cheva-Isarakut pers. comm.). Given this situation, the domestic pig popula¬tion is obviously heavily dependent on the ability of households to produce surplus food. Cattle density appears to be constrained by other factors.

Risks and costs

As noted earlier, raising animals is generally viewed as an investment of surpluses for fu¬ture contingencies. However, being able to gen¬erate surpluses may not be an option for some households due to lack of labour, incapacita-tion of household members or economic margin-alization. Even if some households can generate modest surpluses, there are important risks to consider if small surpluses are to be used for raising animals. There is also an economy of scale to consider. Bounserm Cheva-Isarakut (pers. comm.) argues that raising only one or two cows is simply not worth the trouble for most farmers unless perhaps they are very ambitious. The long period of waiting for in¬vestment returns to come in from raising one or two cows may be a major disadvantage, and the effort involved in raising one cow may be the same as the effort needed for raising 20 head, but the risks of loss and the possible risk of needing to make compensation payments for crop damage may also be too high to make raising one or two cows worth while.

Costs

The costs of starting to raise animals are quite high for newly established or poor families al¬though this can vary considerably depending on proximity of borders and lowland markets (Falvey pers. comm. 1998). In 1976, cattle cost from 2500-4500 Thai baht each (Falvey 1977: 89), with no returns for 2 years. He viewrs the high investment costs (often the equivalent of a household's annual income) and long period of no cash flow as one of the major deterrents to raising cattle. Similar factors probably af¬fect pig raising, although initial costs and re¬turn intervals would be quite a bit lower. I have no data for pigs on this matter, but Visitpanich (pers. comm.) argues that even the more mod¬est initial capital investment for pigs is too high for most poor households. He points out that poor families cannot grow enough food for their own annual consumption and almost certainly could not afford the half-kilogramme of grain per pig per day considered to be the minimum necessary for successfully raising pigs.

While there are no feeding costs associated with stock raising, there are salt costs, and ini¬tial capital costs would have been even more prohibitive for poor families than initial costs for pigs. Salt today is used primarily to calm cattle and attract them back to the owner's house¬hold at periodic intervals or to gather them to¬gether in their forage range for monitoring or moving them to other locations. Opinions dif¬fer as to what extent salt would have been used to control cattle before roads made industrial salts available, even though salt is important for proper cattle growth. Certainly, if salt was used in the past, its cost as a long-distance trade item would have been far greater than it is to¬day (possibly significantly increasing the value of cattle) and much smaller quantities may have been used.

Because of the substantial start-up costs, most cattle (and pig?) raising is begun by inheritance or agistment (the loaning of animals to others to take care of in exchange for a share of the resulting offspring). Terms of agistment con¬tracts vary, but typically, the borrower receives every second calf born. The borrower often takes on considerable responsibility for risks involving the animals. Contracts are witnessed by third parties. Although there are no definitive state¬ments on the matter, it seems highly likely that cattle-rich individuals would be most interested in agisting some cattle to junior lineage mem¬bers who were trustworthy and responsible. Cattle owners are always leary of agisting cat¬tle to unrelated or poor individuals who may not take proper care of the animals or who might allow high-risk situations to develop. At present, a common pattern is for lowland paddy farm¬ers, with cattle, to agist some of their herd with people in the highlands. Most lowland terrain is used for paddy, and there is little forage. Using valuable paddy land to grow grass for grazing is very costly in terms of lost rice production. In contrast, forest is abundant in the highlands and cattle can be raised at much lower cost. Therefore, the highlands are a significant res¬ervoir of cattle for the lowlands and there tends to be a net flow of cattle from the highlands to the lowlands with exchange of cattle for rice or silver or other prestige objects offsetting the trade imbalance. Such a pattern undoubtedly extends far back into prehistory and probably explains the presence of archaeological bronze prestige trade artefacts in the Vietnamese and other highland areas that are normally marginal areas for agriculture.

However, life is never secure and neither are investments. There are always risks of varying magnitude, especially for ambitious investors. The risks involved in raising cattle and pigs are significant, but there are few other options for using surpluses or for advancing in socio-economic standing. Pigs and cattle are the only really convertible form of surplus for access¬ing semi-precious metals or other items of wealth from distant regional centres. Raising animals is also the major way of converting surplus agricultural crops into the social and economic currencies created by feasting (conversion into alcohol and outright consumption being the main other alternatives). There are three main types of risks involved in raising animals: dis¬eases, theft and compensation payments for damages caused by animals to swidden crops, property, or other people.

Epidemic diseases (especially swine fever) frequently decimate pig herds, occurring with an average frequency of 1-6 years, and involv¬ing a mean mortality rate of 74% (Visitpanich & Falvey 1980: 264). Even chicken flocks un¬dergo decimation every two years (Shubert 1986: 80). On the other hand, cattle mortality from disease is relatively low, only about 10% per annum from epidemics (Falvey 1977: 64) al¬though this may be much more prevalent in other regions (Strathern 1971: 129, 130).

Theft of cattle amounts to about 4-5 head per year per village herd when herds are un-tended, which is the usual practice. However, when herds are closely monitored and shep¬herded, losses from theft are almost completely eliminated (Falvey 1977: 54). Thefts and sus¬picion of thefts create substantial conflict be-tween villages. Both theft and disease increase as herd size and animal densities increase (Falvey 1977: 66). This helps place a limit on the size of individual and village herds.

Finally, and most significantly, there are risks associated with damage caused by animals, especially to crops. Fields are generally fenced to prevent cattle from damaging crops, but as herd size and cattle density increase, so does the incidence of crop damage from cattle. Some people do have herds of over 100 head, but risks or management costs must reach maximum lim¬its close to this level. Keeping cattle at a dis¬tance from fields is one of the primary techniques used to keep cattle out of swidden plots, but it requires periodic monitoring and moving of herds (Falvey 1977: 53-5; pers. comm. 1998). Untended herds can graze up to 5-10 km from villages. Ditches, 1-2 m deep, were also used by the Lisu to keep cattle out of fields in some areas. Another technique for reducing crop damage involves the use of herders to monitor the animals. However, this method involves considerable labour costs. Monitored animals return to the villages nightly and generally graze within 2 km of the villages whereas most swiddens are farther than this range.

The seriousness of the risks and costs in¬volved in compensation for damages done by domestic animals is indicated by the fact that harvests could not be obtained from some dam¬aged fields. In some villages, systems of fines were established, and in other villages stock raising was completely abandoned due to fre¬quent crop damages (Falvey 1977: 43, 86). Crop damages must have been a recurring source of conflict and litigation within and between vil¬lages (Strathern 1971: 132; Visitpanich pers. comm.). Families with no domestic animals must have resented the extra labour needed to fence fields or locate them out of range of the domestic animals that other families turned loose in the village territory. For poor house¬holds (or even households with only one or two animals), risks of damage compensation claims against them could be devastating, po¬tentially leading to a life of indebtedness, clientship, enslavement, or violent retribution in some of the more traditional tribal areas of Southeast Asia (Condominas 1977: 123, 139, 151, 156, 338-9). Such risks must have inhib¬ited many poor or socially isolated households from embarking on any significant course of investment, leaving the field to the rich who could absorb the costs of misfortunes much more easily, and who had adequate political back¬ing to resist strong litigations and excessive claims so typical of transegalitarian societies (Douglas & Isherwood 1979: 96; Condominas 1977; Burch 1975: 209, 226; 1980: 267).

In a circular fashion, the use of animals in feasting was essential for establishing the kind of support networks (primarily with affines and agnates) that would agist animals, provide loans in the event of damage compensation claims, and provide political support to resist such compensation claims. Recently established households by newly married couples who were well connected to lineage social networks via kinship and feasting could expect such sup¬port. Poor households lacking feasting involve¬ments or support networks would run substantial risks in venturing into cattle or even pig raising. These and other aspects of tradi¬tional Hill Tribe feasting are discussed in de¬tail by Clarke (1998; 2001).

Labour

In considering labour constraints on cattle rais¬ing, most authors are struck by the very low labour inputs compared to the high value placed on animals (Falvey 1977: 52-4; Shubert 1986: 80-81; Kunstadter 1978: 102). Labour inputs do vary from almost nothing to constant moni¬toring or even penning in the case of pigs. By far, the minimal end of the labour spectrum is more typical of tribal groups (Visitpanich pers. comm.). Low monitoring may save on labour costs, but it increases costs from calf mortality, theft, and predation. Monitoring is generally con¬sidered low skill work and is often assigned to children except during planting and harvesting seasons when all available labor is required to maximize crop production (Falvey 1977: 55).

For cattle, minimal labour costs include the need to check on the location of the animals every week or two to ensure that they have not wandered into another village's territory and are not in danger of invading swidden fields. Since cattle can wander up to 10 km from vil¬lages, these trips can consume considerable time. Giving cattle salt on visits acts to attract them and make them more tractable if they need to be moved. Some households also build shel¬ters or corrals for cattle during inclement weather or for cattle that return nightly to the village — a more common occurrence during the late dry season when surface water and forage can be¬come scarce. Fencing or ditching fields should also probably be considered a labour cost of keeping cattle and pigs, but these costs are borne by each household for their own swidden plots. To an extent, the households that own cattle and pigs impose these costs on households without such animals.

Thus, labour costs related to cattle were tra¬ditionally limited, involving only occasional checks on their location and activities, providing them with salt on visits (or other attractive feed), the construction of field fences or ditches, and possibly the construction of shelters or corrals. Cattle were not generally fed supplements or fattened before slaughter. Movement of cattle to the lowlands for exchange might also be considered a labour cost, but all in all, labour costs are low.

Labour investment is considerably higher for pigs, but pork is preferred and pigs are prob¬ably the most frequently consumed domestic animal (Shubert 1986: 58; Visitpanich pers. comm.). Preference for pork may be due to its higher fat content. Some pigs are penned, but the more traditional villages generally lack pens. As noted earlier, feeding supplements to pigs is an essential part of keeping them, and this generally requires about two hours of feed prepa¬ration per day (Shubert 1986: 80). Some of the food used to feed pigs must also be collected or grown, especially maize, and this can be considered an additional labour input as well.

Summary of the animal raising system

Southeast Asia provides a particularly well documented and well understood system of domestic animal use, raising, distribution and constraints under traditional tribal conditions. Domestic animals serve as the primary vehi¬cles for transforming agricultural surpluses and labour via feasting into other useful currencies such as social, economic, and political mutual help or debt relationships, and wealth in the form of silver or other prestige goods acquired through exchanges.

There is considerable household variation in owning and raising animals with both poor and rich extremes. Poor households appear disadvantaged in most aspects of village life unless they are members in good standing of dominant lineages or power blocks. The poor must adapt to broadly supported aggrandizer strategies for raising domestic animals in vil¬lage communal lands. The poor must fence their swidden plots to protect them from depreda¬tions of free-ranging animals owned by richer households. It is also possible that the com¬petitive nature of some feasts (especially us¬ing as many animals as possible to sacrifice at funerals) may have been promoted by the rich as a ploy (justified on religious or ideological grounds, such as to empower the ancestors) to put pressure on other villagers to produce and exchange greater and greater quantities of sur¬pluses either by raising animals or buying them from the rich. Families who could not afford either to buy or to raise animals would be sub¬ject to public ridicule and censure. These are poor households in the fullest sense that Douglas & Isherwood (1979: 65, 80, 86, 95, 112, 132) use the term. That is, although they may re¬ceive some feasting food directly or indirectly on occasion, the poor are largely cut off from consumption rituals (feasting) and higher ex¬change spheres. They are thus excluded from building social networks, their dominant ac¬tivities are devoted to subsistence, they have little leeway in scheduling or for making mis¬takes, they do not participate much in activi¬ties outside the household, they have little scope or synthesis in their world views, they have little control in their information world and their consumption activities have a high pe¬riodicity but low value. The cost of improving their life constantly seems to be going up or to be a self-defeating undertaking.

In effect, the phenomenon of poverty begins with transegalitarian societies. Douglas & Isherwood may focus on the consumption of goods, but feasting operates in much the same manner, except that among societies with sub¬sistence-based economies, prestige goods can be rare, while having enough desirable food to eat for the entire year is often a major preoccu¬pation. Thus, in most traditional transegalitarian societies, the exchange and use of food surpluses — especially highly valued types of foods — plays the major role in social transactions rather than the exchange of prestige goods (as in Chi¬nese peasant communities — Yan 1996). This is particularly true of groups with low levels of surpluses such as despot types of trans¬egalitarian communities (Hayden 1995). From an archaeological viewpoint, it is interesting to note that feasting may be easier to identify archaeologically than consumption rituals in¬volving non-culinary types of goods due to the specialized preparation and serving vessels involved, as well as the sometimes abundant food remains.

Thus, animals are the centre of the economic and social universe of Southeast Asian tribal villages. Since there are constant demands for animals for feasts, it may be difficult for poor families to accumulate enough surpluses to extricate themselves from poverty. The main constraints on the production of cattle appear to be costly initial investments (or lack of close connexions to owners willing to agist some of their stock), long periods with no return on investment or other benefits from owning stock, and high risks for significant economic losses and conflicts. For pigs, the major constraints are initial procurement costs (or ajistment connexions), production of surplus corn or other foods, and high risks. These constraints put poor and merely self-sufficient households at a dis¬tinct disadvantage in raising animals; and it places the rich in a clearly advantageous situ¬ation.

The above characteristics and generalizations may be common among most or all swidden cultures with stock and pig breeding. In par¬ticular, it might be interesting to conceptual¬ize the initial Neolithic colonization of central Europe in these terms (Gregg 1988). Among the many parallels between the Southeast Asian tribal societies described here and the early European Neolithic are:

1 the highly symbolic and economic value of cattle (Thomas 1991: 24);

2 the primary use of cattle and pigs in feasting contexts, especially at causewayed en¬closures and other Neolithic monuments (Thomas 1991: 22-4, 165);

3 the use of cattle as a major form of wealth (Whittle 1985: 61; Thomas 1991: 24);

4 the minimal role of cattle for secondary purposes such as dairy production (Sherratt 1981);

5 the use of other foods for normal meals including occasional hunted animals (Whit¬tle 1985: 109; Thomas 1991: 21,24).

Ultimately, in both Southeast Asia and Neolithic Europe, successful feasting and raising of do¬mestic animals must be predicated on agricultural surpluses (in the case of pigs), prior wealth accumulations through inheritance of agistments (also deriving from production of agricultural surpluses), and possibly abundant or good qual¬ity labour (for producing initial agricultural surpluses and for monitoring cattle). When cattle started to be used for traction and ploughing, surplus production undoubtedly increased (Bogucki 1993) and later Neolithic societies probably resembled the more stratified lowland Southeast Asian societies rather than highland tribal societies. Understanding how the living systems operate and are related will surely enhance our perceptions of prehistoric socie¬ties with similar socio-economic foundations.

Conclusions

It should be abundantly clear that there are a number of important constraints on the accu¬mulation of surpluses and wealth in the hill tribe communities of Southeast Asia. While in some cases poverty may be a choice of lifestyle by those who simply cannot be bothered with the extra efforts involved in producing surpluses for numerous feasts and attending or socializ¬ing at feasts, in other cases there are powerful constraints on the poor for improving their condition. In fact, the 'choice' of a poor life¬style by some individuals may be more the rec¬ognition of the futility of trying to effect changes in one's status, an aspect of poverty suggested by Douglas & Isherwood (1979: 62). However, their general conclusion (1979: 148) that the poor live the way they do simply because they have lower consumption standards cannot be applied to all households living in poverty. It may be true of some households, but we do not know what that proportion is. In many other cases, poor lifestyles probably result from de¬spair and depression resulting from feeling overpowered by the constraints established by the more powerful members of the community. While transegalitarian communities almost universally give lip-service to an ideology of equality, this rhetoric generally masks very powerful forces working to establish inequalities of wealth, resources, influence and power as documented in the preceding pages. This is why these societies are termed 'transegalitarian'. We can almost certainly find the sources of our own institutionalized inequalities in these societies where inequalities first emerge in furtive, tran¬sient fashions using a wide array of strategies.

Acknowledgements. Accumulating reliable data and obser¬vations on such a wide range of factors was a formidable task. Traditional ethnographies were universally silent con¬cerning most such issues. After experiencing considerable frustration in my attempts to grapple with these problems while interviewing household heads in the region, I sought out academic or other authorities that might be able to shed light on some of these questions. I was very fortunate in be¬ing able to discuss the most critical issues with a number of researchers at Chiang Mai University and the University of Melbourne who were involved in research conducted bv the Thai-Australian Highland Agronomy Project in the early 1970s. This project produced a number of useful publications. I am especially endebted to Dr Theera Visitpanich (Vice-Dean, Faculty of Agronomy, Chiang Mai University), and Dr Lindsay Falvey (Dean, Institute of Land and Food Resources, Univer¬sity of Melbourne) for sharing their insights. In addition, I would like to thank Ralana Maneeprasert of the Tribal Research Institute in Chiang Mai for helping in all aspects of ray work in Thailand. Funding for this research was provided by the Canadian Social Sciences and Humanities Research Council.

References

ALTING VON GEUSAU, L. 1983. Dialectics of Akhazan: the interiorizations of a perennial minority group, in ). McKinnon & W. Bhruksasri (ed.). Highlanders of Thai¬land: 243-77. Kuala Lumpur: Oxford University Press.

BOGGCKI. P. 1993. Animal traction and household economies in Neolithic Europe. Antiquity 67: 432-503.

BCRCH. E., JR. 1975. Eskimo kinsmen. SI Paul (MN): West Pub¬lishing. 1980. Traditional Eskimo societies in northwest Alaska. Senri Ethnological Studies 4: 253-304.

CHAMBERLAIN, J. 1996. The Ethno-economic context, in Envi¬ronmental and social management plan for Nakai-Mam Theun catchment and corridor areas. Vientiane: IUCN.

CLARKE, M. 1998. Akha feasting. Unpublished MA thesis, Ar¬chaeology Department. Simon Fraser University. Burnabv BC. 2001. Akha feasting: an elhnoarchaeological perspective, in Uietler & Hayden (ed.): 144-67.

CONDOMINAS, G. 1977. We have eaten the forest. New York (NY): Hill & Wang.

COOPER. R. 1984. Resource scarcity and the Hmong response. Singapore: Singapore University Press.

DlETLER, M. & B. HAYDEN. 2001. Feasts: archaeological and ethnographic perspectives on food, politics and power. Washington (DC): Smithsonian Institution Press.

DOLGLAS. M. & B. l.SHERWOOD. 1979. The world of goods. Lon¬don: Routledge.

FALVEY, L. 1977. Ruminants in the highlands of northern Thai¬land. Chiang Mai: Thai-Australian Highland Agronomy Project. Tribal Research Institute. Chiang Mai University.

GREGG. A. 1988. Foragers and farmers. Chicago (IL): Univer¬sity of Chicago Press.

HAYDEN. B. 1990. Nimrods, piscators, pluekers, and planters, Journal of Anthropological Archaeology 9: 31—69, 1992. Models of domestication, in A.B. Gebauer & T.D. Price (ed.). Transitions to agriculture in prehistorv: 11-19. Madison (WI): Prehistory Press

1995. A new overview of domestication, in T.D. Price & A. Gebauer. (ed.). Last hunters — first farmers: 273-99. Santa Fe (NM): School of American Research Press. 1997. The Pithouses of Keallev Creek. San Antonio (TX): Harcourt Brace.

HAYDEN, B. &TRAN Qror VIONG. N.d. Report on ethnographic work among the Ta Oi and Rhade.

KESVVANI. P. 1994. The social context of animal husbandry in early agricultural societies. Journal of Anthropological Archaeology 13: 255-77.

KUNSTADTER. P. 1978. Subsistence agricultural economies of Lua' and karen hill farmers, in P. Kunstadter. E. Chapman & S. Sabhasri (ed.). Farmers in the forest: 74-133. Hono¬lulu (HI): Universitv of Hawaii Press.

LEACH, E. 1956. The political systems of highland Burma. Bos¬ton (MA): Beacon Press.

MALINOWSKI, B.. 1961. Argonauts of the Western Pacific. New York (NY): E.P. Dutton.

SHERKATT, A.G. 1981. Plough and pastoralism: aspects of a sec¬ondary products revolution, in I. Hodder, G. Isaac & N. Hammond [ed.]. Pattern of the past: 261-305. Cambridge: Cambridge University Press.

SHLIBERT, B. 1986. Proposals for farming systems-oriented crop research of Waivi highland agricultural research station in northern Thailand. Berlin: Center for Advanced Training in Agricultural Development Technical University of Berlin.

STKATHEKN, A. 1971. Pig complex and cattle complex: some comparisons and counterpoints. Mankind 8: 129-36.

THOMAS. ). 1991. Bethinking the Neolithic. Cambridge: Cam¬bridge University Press.

VISITPANICH. T. & L. FALVEY. 1980. A survey of the highland pig industry, Thai Journal of Agricultural Science 13: 259—67.

WHITTLE. A. 1985. Neolithic Europe: a survey. Cambridge: Cam¬bridge Universitv Press.

YAN YUNXIANG. 1996. The flow of gifts. Stanford (CA): Stanford University Press.