Is the Notion of Mode of Production Universally Relevant?

Is the Notion of Mode of Production Universally Relevant? Is the Notion of Mode of Production Universally Relevant?
Author: Rousseau, Jérôme
Journal: Social Evolution & History. Volume 7, Number 2 / September 2008


The Marxist notion of mode of production has served as a building block for the comparison of social systems. Each mode of production is defined by a specific form of control over the means of production. This paper argues that control over the means of production is an emerging feature of social systems. Without control over the means of production there cannot be a mode of production. Therefore, many societies cannot be said to have a mode of production, and the notions of primitive, lineage, domestic, Asiatic, and ancient modes of production are unsatisfactory and unnecessary.

In my work on the Kayan of central Borneo, I found that I could not meaningfully delineate a mode of production among them, although the notions of surplus appropriation and class were immediately relevant (Rousseau 1979a, 1979b, 1990). The same problem arose while writing a book about social evolution in middle-range societies (Rousseau 2006). I have come to the conclusion that some societies do not have a mode of production; rather, modes of production arise within the process of social complexification.

Marx's main concern was to understand capitalist societies in order to transform them, but the evolutionary nature of his project required that he placed the capitalist mode of production within a broader framework (Marx 1964). Marx's evolutionary schema was sufficient for his purposes, as he was only trying to show what might have been the antecedents of the capitalist mode of production, rather than attempt a detailed study of social evolution. His conceptualization of the feudal mode of production has stood the test of time, because it was an essential step in understanding the rise of capitalism. By contrast, his other modes of production were sketchier and of more limited heuristic value, in particular for stateless societies.

Surplus appropriation

In a first step, we need to clarify our understanding of appropriation. I define surplus appropriation as follows. A productive unit controls, and uses for its own needs, part of what it produces; the rest, which is appropriated by another entity, is surplus1. Surplus is not defined according to the producer's needs: it corresponds to the product of labour not directly controlled by the producer. Part of the surplus may be used to satisfy the producer's needs (we may benefit from the taxes we pay, but taxes are not under our control). Savings are not surplus, because they remain under the producer's control.

Surplus appropriation can take several forms. Woodburn (1982) has shown that there are two basic economic frameworks: immediate- and delayed-return economies. In immediate-return systems, found only among simple hunter-gatherers, the basic principle is ‘From everyone according to their ability to everyone according to their needs’. As a consequence, their social groupings are flexible; individuals can choose with whom they associate; people are not dependent on specific others for their basic requirements; relationships between people stress sharing and mutuality, but there are no long-term commitments; access to territory is open to all. As a consequence, surplus appropriation takes the form of demand sharing, a form of foraging in which the desired object is obtained through a person rather than from nature. Demanding gifts is an expression of egocentrism: ‘I want, so give me’. The demand is not motivated by a desire to fit in a network of reciprocity. In some cases, it verges on tolerated theft (Blurton Jones 1987). In other cases, demand sharing occurs successfully because the giver and the recipient do not value equally what is shared. For instance, a hunter who has killed a large animal might be keen to secure a portion to feed his family, but the rest of the catch has a higher marginal value for others who want to eat now; the hunter cannot reasonably refuse them. To put it another way, in immediate-return societies, the fact of having worked to obtain a resource does not establish a right over it. As a consequence, there is no accumulation of personal possessions, even if these are small and portable. In immediate-return systems, demand-sharing evolves into generalized reciprocity, which is demand sharing to which has been added the ethical injunction to share with others.

Delayed-return systems arise in order to limit and ideally prevent demand sharing. All human societies except simple hunter-gatherers are regulated by delayed-return systems. In delayed-return systems, people hold rights over assets: technical facilities used in production (e.g., boats, nets, stockades); processed and stored food or materials; wild products improved by human labour (e.g., tended wild food-producing plants); and rights held by men over female kin who are bestowed in marriage on other men (Woodburn 1982: 432–433). Delayed-return systems ‘imply binding commitments and dependencies between people’ (Ibid.: 433), because they require long-term organization. Producers require the support of others in order to secure the product of their labour. Co-operation is no longer a series of discrete events, but a chain of exchanges. A consequence of co-operation is the development of established social groupings (permanent villages, kinship groups, clans, and established marital exchanges) (Ibid.: 433). Co-operation increases group stability, because it is easier to co-operate with the same people over a long time. Stability sharpens group identity; distinct groups can develop divergent interests (e.g., mutually exclusive village territories).

In delayed-return systems, producers have a right over the product of their labour because it is the product of their labour. While the emergence of notions of property helps to limit scrounging, it does not prevent surplus appropriation. In fact, it can have the opposite effect. This is evident in the internal dynamics of domestic units.

Domestic units in delayed-return systems

Domestic units exist because spouses establish long-term commitments to each other and to their children. In all societies, a) a gender-based division of labour exists; b) children are not able to look after themselves; c) elderly parents may become less productive. In immediate-return systems, this internal diversity is easi-ly accommodated within the framework of generalized reciprocity: if it is permissible to scrounge from everyone, this is even more true within the domestic unit.

By contrast, delayed-return systems replace generalized reciprocity with accountable reciprocity: an account is kept of what has been received and given so that they are ideally equal. However, this is not feasible within domestic units because of the varying productivity of their members on the basis of gender and especially age. Therefore, generalized reciprocity persists within the domestic unit. In order to insulate the two forms of reciprocity, domestic units must become economic corporations in which all members share what has been produced by its active members.

Since the domestic unit is a corporation, there is an incentive to intensify production, because the unit retains the product of its labour. This can be achieved in a variety of ways: people can work harder; they can also seek to increase the number of economically-active members by retaining their adult children and the latter's spouses. Members of a domestic unit may interact in a spirit of mutuality and respect, but the fact that it is an economic corporation may become an opportunity for surplus appropriation within the domestic unit. For instance, in some societies of highland New Guinea, polygynous men benefited from their wives' labour in order to engage in activities from which they derived prestige and material advantage (Rousseau 2006: 117–135).

Surplus appropriation can also occur beyond the domestic unit on the basis of political differentials. Among the Kayan of central Borneo, commoners perform corvées for their aristocratic chiefs; this is justified by the chief's role in managing community affairs. The justification for surplus appropriation is entirely political: chiefs do not control the means of production. Indeed, when I did fieldwork there, the population density was very low (less than 0.5 per km2 in 1970); more agricultural land was available than people could use (Rousseau 2006: 171–178).

Surplus appropriation can exist without control over the means of production; the latter is an emerging feature that appears as a consequence of pre-existing surplus appropriation. Resource scarcity may be a trigger for developing control over the means of production, but in some cases, scarcity is more a pretext than the reason for it: people who have found ways to benefit from the labour of others realize the potential of controlling the means of production.

If control over the means of production is an emergent feature of social evolution, it follows that the notion of mode of production is not applicable to all societies, because control over the means of production is part of its definition.

Control over the means of production

The notion of mode of production hinges on control over the means of production. Hindess and Hirst's formulation can serve as a starting point.

A mode of production is an articulated combination of relations and forces of production structured by the dominance of the relations of production. The relations of production define a specific mode of appropriation of surplus-labour and the specific form of social distribution of the means of production corresponding to that mode of appropriation of surplus-labour (Hindess and Hirst 1975: 9–10).

On the basis of this general notion, specific modes of production have been identified: primitive, ancient, Asiatic, feudal, capitalist, and socialist. The feudal and capitalist modes of production are clearly conceived and specify precisely the relationship between surplus appropriation and control over the means of production. To varying degrees, the other notions are deficient in this respect.

The concept of primitive mode of production (also called foraging or communal mode of production) is characterized by the collective appropriation of surplus on the basis of collective control over the means of production. This conceptualization is both unconvincing and teleological. It is unconvincing because there rarely is a concept of ownership of the means of production in immediate-return societies; in any case, surplus extraction (in the form of demand sharing) is never justified by collective ownership. Demand sharing occurs at the intersection of individual wants and the need to maintain good relations with one's neighbours. The notion of primitive mode of production is teleological: because our society is framed by the capitalist mode of production, which replaced feudalism, Marxism has made the inference that one could extend the conceptual framework backward ad infinitum. The notion of primitive mode of production does not add anything to our understanding of such societies; indeed, it is less effective than ‘immediate return economy’ and ‘demand sharing’2.

The classical Marxist framework made a broad jump from primitive to ancient or Asiatic modes of production, which anthropologists found unsatisfactory, because it treated as an undifferentiated mass very diverse middle-range societies. In the 1960s, Claude Meillassoux made an attempt to fill this gap by developing the notion of lineage mode of production (1960; see also Rey 1975). According to its proponents, the lineage mode of production is organised around the lineage3, which is the unit of production, consumption, and reproduction. An individual exists as a social agent by virtue of membership in the lineage. Marriages are arrangements between lineages. The lineage is under the control of its elders who derive their power internally from their stewardship over the means of production, and externally from their control over marital arrangements. The lineage mode of production is characterised by a division of labour on the basis of age and sex, which implies the subordination of women and junior men. Production is primarily for use rather than exchange. A lineage mode of production exists when lineages exchange valuables and the subordinate members of the lineage give to the senior members in order for this exchange to take place. The exchange appears to be the raison d'être of the transfer of goods, when it is only the pretext (Testart 1985: 239–240). The notion of lineage mode of production seems to satisfy the general criteria for defining a mode of production, insofar as elders have control over the means of production. However, ‘control over the labor of individual human beings is more important than control over the means of production in defining the relations of production’ (Yanagisako and Collier 1987: 20).

In other words, the link between surplus extraction and control over the means of production is tenuous. While the notion of lineage mode of production implies that lineage elders control junior members because of their control over the means of production, it is more economical to turn the statement around: lineage elders have some control over the means of production because they control junior members. In other words, surplus extraction occurs because of a political relation, as it does among the Kayan, not because of control over the means of production. The notion of lineage mode of production is problematic because it is not linked to effective control over the means of production. Therefore, it does not describe a mode of production, but a form of domination: lineage elders control juniors because they control the political process (including marital arrangements). The notion of lineage mode of production exists only because of an a priori assumption that the mode of production framework is essential to understand all societies.

As an alternative to the primitive and lineage modes of production, Marshall Sahlins has proposed the domestic mode of production (1972), in which ‘production is entirely directed towards the household's internal requirements’ (Layton 1997: 134), hence there is no surplus extraction. The domestic mode of production has no heuristic value, because it argues ‘a priori that all kinds of domestic production are determined by kinds of values qualitatively different from, say, capitalist production’ (Donham 1981: 528). It is also radically flawed because surplus extraction already exists among the simplest hunter-gatherers in the form of demand sharing and generalised reciprocity. Furthermore, Sahlins assumes that the domestic mode of production is characterised by ‘economies organized by domestic groups and kinship relations’ (Sahlins 1972: 41). In fact, the economic importance of domestic groups and kinship varies significantly from society to society (Rousseau 2006: 190–193).

The domestic mode of production has been used by non-Marxist anthropologists to characterize a broad variety of stateless societies, but it is particularly deficient in explaining economic change. For instance, a proponent of the concept might say that, in 1970, the Kayan operated mostly on the basis of a domestic mode of production, insofar as the bulk of their production was for familial use and supra-familial exchange systems also satisfied domestic needs. In particular, when a hunter killed a large animal, it was shared between several domestic units on the basis of a formal distribution system in which each unit received a specific portion of the catch. In 1972, this system collapsed when traders arrived in boats equipped with refrigerators to buy meat and fish. Within a few months of the traders' arrival, meat distribution ceased to be an obligation and became merely an occasional virtuous distribution on the part of a few generous hunters. From the viewpoint of the domestic mode of production, this change must be seen as the consequence of new mercantile values. In fact, it can be explained much more efficiently on the basis of trading practices already present in the ‘traditional’ economy but not applicable to game until refrigerated boats came on the scene.

Similarly, I am not convinced that the notion of ancient mode of production is justified. For Hindess and Hirst (1975: 82), its surplus appropriation is established by right of citizenship: surplus is accumulated in a number of ways by the state and redistributed to citizens. This is certainly a form of domination, and appropriation undoubtedly takes place, but I don't see how this is a mode of production. In the same way, I do not think the notion of Asiatic mode of production adequately links surplus appropriation by the state (through the collection of taxes and tributes) with control over the means of production. It only specifies political control. Therefore, classes are probably present in all societies that have been seen as examples of the Asiatic mode of production, but the existence of a mode of production is doubtful.

In conclusion, the notion of mode of production is of limited heuristic value in understanding foraging and small-scale sedentary societies, because it arbitrarily links surplus extraction with the control over the means of production. Socially defined control over the means of production is not a given in all human societies, but an emerging property of social complexification. Control over the means of production by the dominant sector(s) of society is incomplete except for feudalism, capitalism, and socialism. Marx deve-loped the notion of mode of production in order to understand capitalism, where control over the means of production is crucial. While the concepts of primitive and lineage mode of production were developed to explain the antecedents of capitalism, they are inadequate conceptual tools to analyze these societies. The only reason they were developed was because of an a priori assumption that, if the notion of mode of production is so useful to understand feudal and capitalist societies, it must necessarily be useful for all societies. The evidence contradicts this.

Social class

If the notion of mode of production has a limited heuristic value for small-scale societies, this raises questions about the relevance of the concept of social class in understanding small-scale societies. Like ‘mode of production’, ‘class’ was first developed to understand industrial societies and was then extended to pre-industrial and stateless societies. There is a common core to Marxist definitions: classes are functional groupings that interact in an unequal manner on the basis of their positions in the relations of production. Poulantzas's (1973: 27–28) definition of class also accounts for the relationship between economy and ideology. For him, classes are groupings of socially-identified persons defined primarily but not exclusively by their place in the production process. Economy plays a central role in determining the nature of social classes, but they come into being through the combined effect of the economy, politics, and ideology. Classes are not abstract categories, but groupings of social agents who act collectively. Classes exist only if we can see their actions. A number of features – economic, political, and ideological – come together to form a specific class configuration. If a class analysis is not to be an artificial exercise, we must first identify social groupings with a collective action, and then we can find out how these groups came to be formed.

Common action is not sufficient to identify classes. For a grouping to be a class, its identity must be linked to the relations of production in that society4. Members of a class have common interests: some things are to their advantage because of their position in the relations of production and because of specific political and ideological factors. Members of a class are usually aware of these interests (‘class consciousness’). Within a given society, all classes, by definition, have distinct interests; some classes have contradictory interests, i.e., one exploits the other. I agree with all the above elements in the definition of ‘class’, except the need to link classes to control over the means of production.

As I argued in the previous section, modes of production are emergent structures that are rarely present in middle-range societies, although various forms of exploitation already exist among them. If surplus extraction can take place without control over the means of production, it is unwise to construct a concept of class strictly based on this link, as this makes it more difficult to understand historical transformations. On this basis, I would reformulate Poulantzas's definition as follows. Social classes are groupings of socially-identified agents defined primarily but not exclusively by their place in the process of exploitative surplus appropriation.

Similarly, Donald (1985: 241) suggests that

class is present in a society when significant segments of that society have relatively permanent differential access to resources and/or power. In addition, a class must, in principle, be capable of reproducing itself biologically. This precludes gender or age categories from being regarded as classes.

His formulation raises the issue of the relevance of gender and age to class differences. For instance, Rey and Testart hold opposing views about the presence of classes in lineage societies. For Rey (1979: 60), ‘The lineage mode of production … is characterised by class domination in which the chiefs who inherit power … exercise their power over local groups’. For Testart (1985: 246), there is no class opposition because the kinship relation between exploiter and exploited is a condition of exploitation in the lineage mode of production. The kinship relation is as important as the exploitative relationship, hence the latter does not constitute a fundamental rupture between exploiters and exploited. On this issue, I side with Testart. There is a significant threshold between societies in which exploiters and exploited are relatives and those where exploiters and exploited form separate sectors.

In an earlier paper (Rousseau 1979a), I used the notion of class to understand Kayan stratification. I argued that, while the Kayan have four strata, one can identify three classes. This conclusion helped make sense of an apparent anomaly in Kayan stratification, whereby the hipuy, whom we might gloss as ‘lower aristocrats’, are ritually identified with the maren (chiefly stratum), but socially integrated with the panyin (commoners) (Rousseau 2006: 171–178). As the distinctions between chiefly stratum, commoners, and slaves form the basic structure of Kayan society, defined by surplus appropriation, the stratification system serves to legitimate the social structure. Strata are hereditary, with a preference for stratum endogamy. It becomes possible to believe that the different strata have different natures justifying their respective privileges and obligations. This corresponds to my definition of social classes as groupings of socially identified persons defined primarily by their place in the process of exploitative surplus appropriation, with reference to politics and ideology (including religious beliefs and practices). The hipuy are a consequence of this class situation. If there are too many maren, some of them are drones, and why should they deserve corvées? The hipuy stratum provides the solution. They are supernumerary maren who have ceased to play a chief role. By redefining them as hipuy, one protects the supernatural (and ‘natural’) definition of strata, because the hipuy use the same ritual devices as the maren. The hipuy stratum plays another role. The rigidity of the Kayan stratification systems seems to preclude upward mobility. In fact, if ambitious panyin can demonstrate the presence of aristocratic ancestors (maren or hipuy) in their pedigree, they can argue that they are really hipuy too. They do not make this claim verbally, but by performing aristocratic rituals. If nothing wrong happens to them in the following months, this is seen as a supernatural validation of their claim (Rousseau 1979a: 229–231).

In this analysis, social classes and strata are equally ‘real’. The strata are obvious social categories. The classes are also real groupings with specific action, not only constructs of the analyst. In my fieldwork, I routinely observed statements and actions demonstrating class consciousness. The maren knew they needed to maintain solidarity in order to continue to rule their followers (Ibid.: 232–233). Commoners (i.e., panyin and hipuy) also recognised that the chief's interests were often contradictory to theirs. This class-consciousness was not revolutionary, because they did not see the possibility of transforming the system. Since World War II, there have been attempts to do so politically by trying to form commoner villages (Ibid.: 232), and religiously through an indigenous religious reform that made away with ritual distinctions between commoners and aristocrats (Rousseau 1998: Ch. 2). Both failed. The breakaway villages were forced to return to their original communities; the maren and the priests of the old religion managed to take control of the religious reform; the ritual contrasts between aristocrats and commoners were reinstated.

Similarly, the notion of class has been useful in understanding Northwest Coast societies (Rousseau 1979b; Ruyle 1973). On the other hand, it is not applicable to such societies as the New Guinea Highlands, where exploitation takes place primarily within the domestic unit. Classes are not coterminous with technological or economic complexity. For instance, the trading networks of New Guinea Highland big men reach further than the local-level extraction of central Borneo chiefs. Many small-scale societies experience inequality without having classes. Social inequality is prior to, and broader than, class. The appearance of classes needs to be explained on the basis of pre-existing inequality. The Marxist argument about the class struggle being the motor for the evolution of societies is a coherent, if incomplete, explanation of state formation and the later vicissitudes of states, but it begs the question of what happened between the first delayed-return systems and the first class struggles. Surplus extraction precedes class struggles; it is itself predicated on the existence of social differentiation. The early forms of social differentiation are very different from those of class societies.


Surplus appropriation exists in all societies. Furthermore, exploitative surplus appropriation can be present without exploiters having control over the means of production. Exploitation precedes the emergence of classes. Classes are present when a segment of society can expect, because of its socio-political position, to appropriate the labour (or produce) of another sector, and where this cleavage identifies fundamental differences of interest recognized by the participants. In later stages, it may lead to control over the means of production.


* An early version of this paper is included in Rousseau (2006: 214–222). I thank McGill-Queen's University Press for permission to republish this argument in an expanded format.

1 The nature of the productive unit is identified in each social formation. It often is the domestic unit, but there are other possibilities.

Barnard (2004: 9) makes a related point when he suggests that we should talk of a ‘foraging mode of thought’ in order to ‘direct our attention away from production, in the narrow, Marxist sense, and towards an understanding of social relations’.

3 A lineage is constituted by consanguineal relatives who trace their descent to specific ancestors; they have a variety of rights and obligations because of their membership in their lineage.

4 For instance, Shivji's (1976) analysis of colonial and post-colonial Tanzania follows Poulantzas's conceptualisation. Tanzanian Asians form a class (‘commercial bourgeoisie’) not only because of their ethnic identity, but because this ethnic identity is articulated to the relations of production organising Tanzanian society. Asians came to Tanzania because of colonialism and were expected to occupy a specific position in the economy.


Barnard, A.

2004. Mutual Aid and the Foraging Mode of Thought: Re-reading Kropotkin on the Khoisan. Social Evolution & History 3(1): 3–21.

Blurton Jones, N. G.

1987. Tolerated Theft: Suggestions about the Ecology and Evolution of Sharing, Hoarding, and Scrounging. Social Science Information 26(1): 31–54.

Donald, L.

1985. On the Possibility of Social Class in Societies Based on Extractive Subsistence. In Thompson, M., Garcia, M. T., and Kense, F. J. (eds.),Status, Structure and Stratification: Current Archaeological Reconstructions. Proceedings of the Sixteenth Annual [Chacmool] Conference (pp. 237–244). Calgary: The Archaeological Association of the University of Calgary.

Donham, D. L.

1981. Beyond the Domestic Mode of Production. Man (N. S.) 16: 515–541.

Hindess, B., and Hirst, P. Q.

1975. Pre-capitalist Modes of Production. London: Routledge & Kegan Paul.

Layton, R.

1997. An Introduction to Theory in Anthropology. Cambridge: Cambridge University Press.

Marx, K.

1964. Pre-capitalist Economic Formations (with an Introduction by Eric J. Hobsbawm). New York: International.

Meillassoux, C.

1960. Essai d'interprétation du phénomène économique dans les sociétés traditionnelles d'autosubsistance. Cahier d'études africaines 1: 38–67.

Poulantzas, N.

1973. On Social Classes. New Left Review 78: 27–54.

Rey, P.-P.

1975. The Lineage Mode of Production. Critique of Anthropology 1(3): 27–79.

1979. Class Contradiction in Lineage Societies. Critique of Anthropology 4(13–14): 41–60.

Rousseau, J.

1979a. Kayan Stratification. Man (N. S.) 14: 215–236.

1979b. Stratification and Chiefship: A Comparison of Kwakiutl and Kayan. InTurner, D. H., and Smith, G. A. (eds.), Challenging Anthropo-logy (pp. 229–244). Toronto: McGraw-Hill Ryerson.

1990. Central Borneo: Ethnic Identity and Social Life in a Stratified Society. Oxford: Clarendon.

1998. Kayan Religion: Ritual Life and Religious Reform in Central Borneo. Leiden: KITLV Press.

2006. Rethinking Social Evolution: The Perspective from Middle-Range Societies. Montreal: McGill-Queens University Press.

Ruyle, E. E.

1973. Slavery, Surplus and Stratification on the Northwest Coast: The Ethnoenergetics of an Incipient Stratification System. Current Anthropology 14: 603–631.

Sahlins, M. D.

1972. Stone Age Economics. Chicago: Aldine.

Shivji, I.

1976. Class Struggles in Tanzania. New York: Monthly Review Press.

Testart, A.

1985. Le communisme primitif. Tome I. Economie et idéologie. Paris: Editions de la Maison des Sciences de l'Homme.

Woodburn, J.

1982. Egalitarian Societies. Man (N. S.) 17: 431–451.

Yanagisako, S. J., and Collier, J. F.

1987. Toward a Unified Analysis of Gender and Kinship. In Collier, J. F., and Yanagisako, S. J. (eds.), Gender and Kinship: Essays Towards a Unified Analysis (pp. 14–50). Stanford: Stanford University Press.