The Domestic Mode of Production: The Structure of Underproduction
This chapter is constructed on an observation in apparent contradiction to the pristine "affluence" I have just taken so much trouble to defend: the primitive economies are underproductive. The main run of them, agricultural as well as preagricultural, seem not to realize their own economic capacities. Labor power is underused, technological means are not fully engaged, natural resources are left untapped.
This is not the simple point that the output of primitive societies is low: it is the complex problem that production is low relative to existing possibilities. So understood, "underproduction" is not necessarily inconsistent with a pristine "affluence." All the people's material wants might still be easily satisfied even though the economy is running below capacity. Indeed, the former is rather a condition of the latter: given the modest ideas of "satisfaction" locally prevailing, labor and resources need not be exploited to the full.
In any event, there are indications of underproduction from many parts of the primitive world, and the first task of the essay is to give some sense of the evidence. Beyond any initial attempt at explanation, the discovery of this tendency—more precisely of several related tendencies of the primitive economic performance—seems of greater importance. I raise the possibility that underproduction is in the nature of the economies at issue; that is, economies organized by domestic groups and kinship relations.
Dimensions of Underproduction
UNDER USE OF RESOURCES
The major evidence for underexploitation of productive resources comes from agricultural societies, especially those practicing slash-and-burn cultivation. Probably this is a function of research procedures rather than a dubious special privilege of the subsistence type. Similar observations have been made of hunting and of herding economies, but anecdotally for the most part, and without benefit of a practicable measure. Slash-and-burn agriculture, on the other hand, uniquely lends itself to quantified assessments of economic capacity. And in almost all the cases so far investigated, still not numerous but from many different parts of the globe, especially where the people have not been confined to "native reserves," the actual production is substantially less than the possible.
Slash-and-burn, an agriculture of neolithic origin, is widely practiced today in tropical forests. It is a technique for opening up and bringing under cultivation a patch of forest land. The standing growth is first cleared by axe or machete and, after a period of drying out, the accumulated debris is burned off—thus the inelegant name, slash-and-burn. A cleared plot is cultivated for one or two seasons, rarely more, then abandoned for years, usually with a view toward restoration of fertility through reversion to forest. The area may then be opened again for another cycle of cultivation and fallow. Typically the period of fallow is several times the period of use; hence, the community of cultivators, if it is to remain stable, must always hold in reserve several times the area it has under production at any given moment. Measures of productive capacity must take this requirement into consideration; also the period of garden use, the period of fallow, the amount of land required per capita for subsistence, the amount of arable land within range of the community, and the like. So long as these measures are careful to respect the normal and customary practices of the people concerned, the final estimate of capacity will not be Utopian—that is, what might be done with a free choice of techniques—but only what could be done by the agricultural regime as it stands.
Nevertheless, there are inescapable uncertainties. Any "productive capacity" so estimated is partial and derivative: partial, because the investigation is restricted in advance to the cultivation of food, other dimensions of production left aside; derivative, because "capacity" takes the form of a population maximum. What research yields is the optimum number of people that can be supported by the existing means of production. "Capacity" appears as a determinate population size or density, a critical mass that cannot be surpassed without some change in agricultural practice or conception of livelihood. Beyond that point is a dangerous ground of speculation which daring ecolo-gists, identifying the optimum population as the "critical carrying capacity" or "critical population density," all the same do not hesitate to enter. "Critical carrying capacity" is the theoretical limit to which the population could be taken without degrading the land and compromising the agricultural future. But it is characteristically difficult to project from the existing "optimum" to the persisting "critical"; such questions of long-term adaptation are not decided by the short-term data. We have to be content with a more limited, if possibly defective, understanding: what the agricultural system as constituted can do.
W. Allan (1949, 1965) was the first to devise and apply a general index of population capacity for slash-and-burn agriculture. Several versions and variants of Allan's formula1 have since appeared, notably those of Conklin (1959), Carneiro (1960), and a complicated refinement fashioned by Brown and Brookfield for the New Guinea Highlands (1963). These formulas have been applied to specific ethnographic sites and, with less precision, to broad cultural provinces dominated by slash-and-burn production. Outside of reservations, in traditional agricultural systems, the results, although highly variable, are highly consistent in one respect: the existing population is generally inferior to the calculable maximum, often remarkably so.2
1. Following the slight rephrasing by Brown and Brookfield (1963), Allan's formula is: "carrying capacity" = 100 CL/P where Pis the percentage of arable available to the community, L is the mean acreage per capita under cultivation and Ca factor of the number of garden units needed for a full cycle, calculated as fallow period + cultivation period/fallow period. The result of 100 CL/P is the amount of land required to support one person in perpetuity. This is then converted into a density per square mile or square kilometer.
2. This conclusion is framed for the population, globally considered, practicing a determinate form of agriculture; it does not preclude that localized subgroups (families, lineages, villages), under the given rules of recruitment and land tenure, will not experience "population pressure." Such of course is a structural problem, not posed by technology or resources per se.
Table 2.1 summarizes a certain number of ethnographic studies of population capacity from several world areas of shifting agriculture. Two of these studies, those of the Chimbu and Kuikuru, merit special comment.
Table 2.1. Relation of actual to potential population, swidden cultivators
Location
Population (size or density)
Actual as Percentageof Potential
Group
Actual
Potential Maximum
Source
Naiegu Chimbu
New Guinea
288/m2
453/m2
Brown and Brookfield 1963
Tsembaga * (Maring)
New Guinea
204 (local pop)
313-373
55-65
Rappaport 1967
Yagaw Hanaoo
Philippines
30/km2 (arable)
48/km2 (arable)
Conklin 1957
Lamett
Laos
2.9/km2
11.7-14.4/km2
20-25
Izikowitz 1951
Iban
Borneo
23/m2
(Sut Valley)
14/m2 (Baleh)
35-46m2
50-66 (s) 30-40
Freeman 1955
Kuikum
Brazil
145 (village)
Cameiro 1960
Ndembu (Kanongesha Chiefdom)
N. Rhodesia
3.17/m2
17-38/m2
8-19
Turner 1957
W. Lala*
N. Rhodesia
<3/m2
4/m2
<75
Allan 1965: 114
Swaka*
N. Rhodesia
<4/m2
10+/m2
<40
Allan 1965: 122-123
Dogomba*
Ghana
25-50/m2
50-60/m2
42-100
Allan 1965: 240
*Mean population capacity, between maximum and minimum pig-herd, here tabulated.
** The Lamet figures are calculated from Izikowitz's rough estimates, with the further assumption that only five percent, of the countryside is arable. The results are probably far from accurate. However, we have the ethnographer's assurance that Lamet villages have considerably more land at their disposal than they need (use) (1951, p. 43).
***Allan presents data on several African populations, confined to reserves or otherwise subjected to disturbances of colonialism, that are over the capacity of the traditional system. These are excluded here. The Serenji Lala, however, may be an exception. (Most of Allan's estimates seem more approximate than the other studies tabulated above.)
The Chimbu example is indeed theoretically privileged, not only for the unusually sophisticated techniques developed by the investigators, but because these techniques were tested on a system of peak density in one of the most densely occupied areas of the primitive world. The Naregu section of Chimbu studied by Brown and Brookfield certainly upholds the New Guinea Highlands' reputation: a mean density of 288 people/square mile. Yet this density is only 64 percent of the prevailing agricultural capacity. (The result of 64 percent is an average for 12 clan and subclan territories of Naregu; the range was from 22 to 97 percent of capacity; Table 2.2 gives the breakdown by territory.) Brown and Brookfield also made wider but less precise estimates for the 26 tribal and subtribal sections of Chimbu as a whole, yielding conclusions of the same order: mean population at 60 percent of capacity.3
3. Four of the 26 groups were above capacity. All four, however, fall into the two lowest of four categories of data-reliability developed by Brown and Brookfield. Only Naregu received the highest classification of reliability. Groups in the second highest had the following indexes of actual to potential population: 0.8 (two cases), 0.6, 0.5, 0.4, and 0.3.
The Kuikuru, on the other hand, illustrate another kind of extreme: the scale of the disparity that may exist between potential and reality. The Kuikuru village of 145 persons is only seven percent of the calculable maximum population (Carneiro, 1960). Given the Kuiku-rus' agricultural practices, their present population of 145 is supported from the cultivation of 947.25 acres. In fact, the community has a base of 13,350 acres (arable), sufficient for 2,041 persons.
Although studies such as these remain few, the results they present do not appear to be exceptional nor limited to the instances in question. On the contrary, reputable and sober authorities have been tempted to generalize to the same effect about wide geographical areas with which they are familiar. Carneiro, for example (projecting from Kuikuru but in a way that presumes them unusually well off), considers that traditional agriculture in the South American Tropical Forest Zone was capable of sustaining village populations on the order of 450 people; whereas the modal community throughout this extensive area was only 51-150 (1960). The Congo forest of Africa, according to Allan, was likewise underpopulated over wide stretches—"well below the apparent carrying capacity of the land for the traditional systems of land use" (1965, p. 223). Again in West Africa, particularly Ghana before the cocoa boom, Allan reports that "population densities in the central forest zone were far below the critical levels" (p. 228; cf. pp. 229, 230, 240). J. E. Spencer frames a similar opinion of shifting cultivation in Southeast Asia. Impressed by the unusually high densities of upland New Guinea, Spencer is inclined to believe "most shifting-cultivator societies are operating at less than maximum potential so far as their agricultural system is concerned" (1966, p. 16). His interpretation is of interest:
Table 2.2. Actual and maximum population capacities ofNaregu Chimbu Groups*
(from Brown and Brookfield, 19631 pp. 117, 119)
Group
Total Population
Population Density per Square Mile
Proportion ofActual toMaximum Density
Actual
Maximum
Actual
Maximum
Kingun-Sumbai
0.49
Bindegu
0.91
Togl-Konda
0.82
Kamaniambugo
0.97
Mondu-Ninga
0.77
Sunggwakani
0.66
Domkani
0.58
Buruk-Maima,
0.80
Damagu
Komu-Konda
0.79
Bau-Aundugu
0.56
Yonggomakani
0.40
Wugukani
0.22
X = 288
X = 453
X = 0.64
*The capacities reported by Brown and Biookfield include a small allowance (.03 acres/capita) for a cash crop, coffee, as well as an allowance for a tree crop, pandanus (0.02 acres/capita). The food-crop requirement of 0.25 acres/capita also includes an amount for pig food and some food sold. The allowance for pigs, however, is not adjusted to maximum herd size.
Light areal density patterns of population are naturally associated with many groups following shifting cultivation because of their intrinsic social system. . . . This cultural tradition cannot be interpreted in terms of the carrying capacity of the land, so that the social phenomenon, rather than the literal carrying capacity of the land itself, has assumed the dynamic role of controlling population density (Spencer, 1966, pp. 15-16).
Let us underline the point, at the same time reserving it for fuller discussion later. Spencer says that the social-cultural organization is not designed after the technical limits of production, to maximize output, but rather impedes development of the productive means. If this position runs counter to a certain ecological thinking, it is nevertheless repeated by several ethnographers of underproduction. For the Ndembu, in Turner's view (1957), it is the contradictions of customary modes of residence and descent, coupled to an absence of political centralization, that set off village fission and population dispersal at a level inferior to the agricultural capacity. Izikowitz (1951), speaking of Lamet, and Carneiro of Amazonian Indians (1968) alike hold the weakness of the community polity responsible for an undue centrifugal segmentation. Quite generally among the tribal cultivators, the intensity of land use seems a specification of the social-political organization.
To return to the technical facts and their distribution: slash-and-burn agriculture is a major form of production among extant primitive societies, perhaps the dominant form.4
4. According to a recent FAO report, some 14 million square miles, occupied by 200 million people, are still exploited by slash-and-burn (cited in Conklin, 1961, p. 27). Of course, not all of this is primitive domain.
Inquiries in a number of communities, from several different world areas, confirm that (outside native reserves) the agricultural system is running below its technical capacity. More broadly, extensive areas of Africa, Southeast Asia, and South American occupied by swidden cultivators are authoritatively judged under-exploited. May we be permitted to conclude that the dominant form of primitive production is underproduction?5
5. The consistent discrepancy between population density and agricultural capacity, even where the former attains 200-plus people/square mile, raises more than one passionate theoretical question. What are we to make of the popular inclination to invoke demographic pressure on resources in explanation of diverse economic and political developments ranging from the intensification of production to the elaboration of patrilineal structure or the formation of the state? First of all, it is not evident that archaic economies know a tendency to reach, let alone exceed, the population capacity of their means of production. On the other hand, it is evident that current mechanistic explanations from demographic cause—or, conversely, the inference of "population pressure" from an observed economic or political "effect"—are often oversimplified. In any given cultural formation, "pressure on land" is not in the first instance a function of technology and resources, but rather of the producers access to sufficient means of livelihood. The latter clearly is a specification of the cultural system—relations of production and property, rules of land tenure, relations between local groups, and so forth. Except in the theoretically improbable case in which the customary rules of access and labor are consistent with optimum exploitation of land, a society may experience "population pressure" of various kinds and degrees at global densities below its technical capacity of production. Thus the threshold of demographic pressure is not an absolute determination of the means of production but is relative to the society at issue. Moreover, how this pressure is organizationally experienced, the level of the social order to which it is communicated, as well as the character of the response, also depend on the institutions in place. (This point is well made by Kelly's study of the problem in the New Guinea Highlands, 1968.) Hence both the definition of population pressure and its social effects pass by way of the existing structure. Consequently, any explanation of historical events or developments, such as warfare or the origin of the state, that ignores this structure is theoretically suspect.
Much less can be said about the performance of other common production types. There are suggestions that hunting-gathering may be no more intensive than slash-and-burn agriculture. But the interpretation of resource underuse among hunters presents special difficulties, even apart from the lack of a practicable measure. It is usually not possible to determine whether an apparent underproduction of the moment nonetheless represents a long-term adaptation to recurrent shortages, bad years when it would be possible to support only a fraction of the present population. All the more pertinent, then, the following remark of Richard Lee on /Kung Bushman subsistence, as the period of field observation included the third year of a prolonged drought such as rarely visits even the Kalahari Desert:
It is impossible to define "abundance" of resources absolutely. However, one index of relative abundance is whether or not a population exhausts all the food available from a given area. By this criterion, the habitat of the Dobe-area Bushmen is abundant in naturally occuring foods. By far the most important food is the Mongomongo (mangetti) nut… Although tens of thousands of pounds of these nuts are harvested and eaten each year, thousands more rot on the ground each year for want of picking (Lee, 1968, p. 33; see also pp. 33-35).
Woodburn's comments on Hadza hunting carry the same implication:
I have already mentioned the exceptional abundance of game animals in this area. Although Hadza, in common probably with all other human societies, do not eat all the types of animals available to them—they reject civet, monitor lizard, snake, terrapin among others—they do eat an unusually wide range of animals.. . . In spite of the large number of species which they are both able to hunt and regard as edible, the Hadza do not kill very many animals and it is probable that even in the radically reduced area they occupied in 1960 more animals could have been killed of every species without endangering the survival of any species in question (Woodburn, 1968, p. 52).
In a work primarily devoted to subsistence agriculture, Clark and Haswell (1964, p. 31) make a daring argument about preagricultural resource use that at least invites contemplation. Basing their calculations on certain data for East Africa summarized by Pirie (1962),6 and positing certain conservative assumptions about animal reproduction rates in the wild, Clark and Haswell estimate that the annual natural yield of meat is forty times greater than necessary to support a hunting population living at one person/20 square kilometers (1/7.7 square miles) and exclusively on animal foods—that is to say, the animal reproduction fully utilized would support five persons per square mile. This without diminishing the natural supply. Whether hunters need such a margin of safety is another, unanswered question, although Clark and Haswell rather think they do.
6. These Pirie had himself culled from the Arusha symposium on Conservation of Nature and Natural Resources in Modern African States (1961). This publication was not available to me on writing. Pirie's article, moreover, raises some question about the control of predators (p. 411), the significance of which is unclear but which may have bearing on the figures for wild animal yields.
A further implication of Pirie's East African figures is that the wild animal yield per area of natural grazing land is higher than the output of pastoral nomadism in adjacent regions (cf. Worthington, 1961). Again, Clark and Haswell generalize to an interesting judgment of pastoralist land use:
We should remind ourselves that the primitive pastoral communities, found where the land is not forested . . . live at a density of about 2 persons/sq. km. Though not so wasteful of the land and its resources as are the primitive hunting peoples, they nevertheless fall far short of fully exploiting the potential mean output of land, which Price estimates at 50 kg. liveweight gain/ha./year (5 tons liveweight gain/sq. km.). Even if we half this figure, as some would do, it seems clear that primitive pastoral peoples . . . are unable to exploit the full growth of grass in favourable seasons of the year (1964 ).
Without technical means of accumulating fodder, as the authors recognize, pastoralists are of course restricted to the livestock they can support in poorer rather than favorable seasons. Still, Clark and Haswell's conclusion finds some support from Allan. As a rough conjecture, Allan supposes that East African pastoralists know a "critical population density" on the order of seven persons per square mile. But from a series of actual cases, "It would seem that population densities of surviving pastoral peoples are usually well below this figure, even in the more favourable of the regions they still occupy" (Allan, 1965, p. 309).7
7. Allan, on the other hand, finds among pastoralists some urge to accumulate cattle that may outstrip pasturage capacities, and at least two peoples, Masai and Mukogodo, with an apparent "excess of livestock in relation to the economic requirements of simple pastoralism" (1965, p. 311).
We seem perilously close to that characteristic failing of interdisciplinary study—an enterprise which often seems to merit definition as the process by which the unknowns of one's own subject matter are multiplied by the uncertainties of some other science. But enough said at least to raise doubt about the efficiency of resource exploitation in the primitive economies.