Banaji’s Theory as History provides an incisive analysis of pre-capitalist modes of production, demonstrating that no aspect of human history is irrelevant to the present.

jarius banaji

Jairus Banaji, Theory as History: Essays on Modes of Production and Exploitation (Brill 2010), Historical Materialism Book Series, volume 25, xix, 406pp.

Did the first proletarian communists appear in the era of the Roman Empire? In fourth-century Numidia, what is now Algeria, bands of seasonally-hired labourers began to form roving groups which attacked the rich and their property, while preaching their brand of radical Christianity. A signature action would be to force the well-to-do from their carriages to run behind them in place of their slaves. These people called themselves Agonistici, ‘those who engage in struggle’, ‘fighters’, or as Jairus Banaji nicely translates it ‘militants’. Their enemies called them Circumcellions (and so they are known today) and accused them of all manner of inexplicable barbarities (uncritical, not to say gullible, modern commentators buy the notion that they were a suicide cult).

Apart from being wage labourers attacking the propertied ruling class, the Agonistici were the radical wing of the Christian Donatist heresy and rejected the existence of private property. They appear to have taken seriously all the implications of the Gospels that the leaders of the Christian Church, then a key institution of the late Roman Empire, sought to marginalise.

A simple answer to the opening question is, yes; the Circumcellions were a kind of early communist, and they deserve remembrance in the history of resistance to class rule. However to say that they were proletarian communists should strike the reader as obviously untenable: this was the Roman Empire not a modern capitalist economy, and so a proletariat could not exist. Yet it is not an exaggeration to note that the logic of some current approaches to pre-capitalist modes of production would force the conflation of formally free wage-labourers with modern proletarians, in effect ignoring the differences between the eras at the level of the totality.

The logic begins with the argument that ultimately there are only three ways of exploiting labour, through slavery, the seizure of rent or taxes from peasants, or the extraction of surplus value from wage labour. The consequence is that, with the exception of brief moments when slavery may have been the dominant form of exploitation, only one class-based mode of production could have existed before capitalism (see for example Haldon, 1993, pp.76-7, and Harmon, 2006, pp. 188-9). The unfortunately reductive implications of this position should immediately be clear. Despite widespread and significant use of wage labour in Late Roman north Africa, capitalist relations of production were certainly not emerging there (unless you commit the other error of identifying capitalism with monetised trade in itself). Marx’s historical concept of the mode of production involves a totality of social relations; it is not a positivist identification of a part of a system as diagnostic of the whole.

Debates over pre-capitalist modes of production might strike some as being abstruse and of only academic interest, but our understanding of these questions actually does directly affect our present politics. In Jairus Banaji’s case the implication for Indian politics is clear. The reductive and positivist approach above, identifying the dominant category of exploited labour as definitional of the mode of production, is a damaging error. The consequence is that the existence of a mass of peasant labour in India defines its economy as feudal. If modern India remains feudal at least in part, that condition defines the key forms and directions of class struggle and socialist politics. Capitalism remains a ‘progressive’ force in this scenario, and Stalinist developmentalist politics follow.

Genuine revolutionary socialist politics needs to be able to recognise the totality of social relations in which small peasant labour is exploited. In modern India, that exploitation takes place within a capitalist mode of production. Banaji argues that here capitalist market relations ultimately reward small peasants with a hidden wage, and subject them to the formal domination of capital (Banaji, pp.326-30). The problem for the Indian peasant is not the lack of capitalist development. Rather, it needs to be recognised that the international development of capitalist relations of production can freeze in place ‘backward’ social forms in some areas, actually as a consequence of this overall development. There is no particular reason why the capitalist system should be unique in this respect. As a consequence, the precise form of the exploitation of labour is not equivalent to relations of production, and is not diagnostic of a mode of production.

Whether his whole analysis of capitalism in India is accepted or not, Banaji’s methodological argument here is crucial. He is fighting to prevent Marxism from being absorbed into the mainstream as a form of modernisation theory. If modes of production are defined solely in terms of modes of exploitation, as Banaji formulates the distinction (p.4), then capitalism does not arrive until modern factory-based industry. The problems of third world economies become the problem of an insufficient development of capitalism, and various supposed blockages to the modernisation of these economies. Yet it should be clear enough that the problem of poverty and exploitation in the developing world is a product of the worldwide dynamics of capitalist relations of production, beginning at least as far back as the seventeenth century. The reductionist approach to relations of production thus threatens the clear understanding of the role of imperialism in the historic forced underdevelopment of the Third World by western polities (to adapt Walter Rodney’s memorable title: How Europe Underdeveloped Africa).

The reductionist argument has further unfortunate consequences for Marxist theory, since, as noted, it effectively means that it is not possible to distinguish between different modes of production before capitalism. Tribute, tax and rent are all interchangeable, while slavery, for many, only ever played a temporary and subordinate role. If all pre-capitalist formations are really contingent variants of a generalised ‘feudalism’, then the Marxist historical method loses considerable explanatory force. Certainly at least the category ‘mode of production’ becomes next to useless, and the form of state, law, culture or ideology is not meaningfully determined by forms of social production and reproduction, but by merely contingent events. Marxism cannot explain, or at least explore, the huge differences between ancient and medieval states and polities from China, India, the Middle East, western Europe and so forth. Banaji’s argument, in contrast, points to a way forward for re-establishing the intellectually creative potential in the Marxist historical method.

The issue of distinguishing modes of exploitation and modes of production, Banaji’s starting point, needs some elaboration here, because it does have an impact on a whole number of debates, from the nature of feudalism, to the character of new world slavery in early capitalism. Banaji points out that phenomena like ‘wage labour’ were defined by Marx as ‘simple categories’ which can be found across various historical systems of production. Proletarian wage labour under capitalism is therefore a more complex category and cannot be reduced to the empirical appearance of wages in the historical record. Simple categories like rent or wages cannot therefore in themselves define a wider mode of production. Indeed Lenin’s own analysis of the Moscow pottery industry is an example of where capitalist relations of production ‘far from transforming bondage and serf-forms of exploitation into specifically capitalist forms of exploitation . . . intensify the existing backward forms of exploitation’ (Banaji, p.56). At this point, although he does not raise the issue, it would be possible to connect Banaji’s methodology with the theories of combined and uneven development and permanent revolution.

The ‘simple category’ cannot be used as an easy empirical signal through which the mode of production can be ‘read off’. Nor should modes of production be expected to follow a canonical pattern of development from slavery through serfdom to capitalism. Banaji clearly lays out the formalistic errors of writers like Hindiss and Hirst, or Perry Anderson’s perplexing descent into ‘metaphysical reasoning’: the Byzantine Empire remained ‘transfixed between slave and feudal modes of production… in a social deadlock’… for nine centuries (pp. 8-9)! Banaji is not the first to find Anderson ludicrous on this point. The limitations of sheer empiricism in a scholar like Maurice Dobb are also carefully revealed. All of these Marxists stirred important debates, and Banaji’s discussions here are all worth comment, however the key point lies in the level of analysis or abstraction to which Banaji directs us. He shows with copious reference to Marx’s analyses in the Grundrisse and Capital, that a mode of production has to be grasped at the level of the totality of its social productive relations. Each has to be seen as a dynamic system rather than being a set of parts built up from ‘simple categories’ into a static structure.

Banaji’s careful reading of Marx’s procedure in Capital allows him to make some illuminating distinctions between the dynamics of a feudal compared to a capitalist economy. For example, where production in capitalism needs to realise the value of the goods it produces to remain profitable, and therefore even to reproduce production, the same calculation does not apply to a feudal economy. Crops were, Banaji argues: ‘more or less “profitable”, not according to their monetary rate of profit but according to the volume of cash which they brought in. Any increment in crop-production which increased the total volume of these receipts, however marginal its contribution, was therefore “profitable”. That is to say, for a given distribution of productive forces, even if it made sense to sell a given output at the highest price, sales would nonetheless continue over a wide range of prices below this bound’ (Banaji, p.78). Working from this basis, Banaji is able to suggest a complex ‘law of motion’ for feudalism, which deserves serious exploration and confirmation.

This analysis is not merely of academic interest for the dynamics of production of the medieval economy. It is the starting point of a demonstration that bourgeois laws of economics are not universal and natural, but only came into existence in a definite historical context. The consequence is not only that production once did not obey the nostrums of today’s capitalist economics, but that it would not have to do so in the future either. Closely linked with the differentiation of the dynamics of production between feudalism and capitalism, is the place of trade and commerce within feudalism. Here it is again an issue of historicising that which bourgeois ideology holds to be natural and immutable. In mainstream discussions, it is easy to get away with the implicit assumption that trade is in itself constitutive of capitalism. Where you have industry of any kind linked to trade of any level, this is at least capitalism in embryo; the difference is merely one of quantity.

This ‘common sense’ perspective once again naturalises capitalism into a universal constant of human society. Yet it is not so; trade and industry do not follow the same economic laws or dynamics in all societies, and commerce in feudal Europe, for example, should be seen in terms of the particular economic logic of that society. The reappearance of a monetary-based economy in twelfth-century Europe does not mean that capitalist relations of production were in any sense emergent from that time, despite unexamined assumptions that such was the case, for example in a classic work on the period like Lester K. Little, Religious Poverty and the Profit Economy.

Banaji warns of a series of mistakes to avoid in coming to terms with the rise of capitalism as a mode of production; Maurice Dobb, for example, mistook modern industry for the category ‘mode of production’, and searched for capitalism empirically in a strictly national context. From the first Banaji argues, capitalism has to be understood as an international system (pp.255-8). Nonetheless, none of this is easy to follow through systematically, and it seems as if Banaji’s position has changed somewhat over time; some chapters are quite new, but others reprint articles from 1977 onwards. Thus in some discussions Banaji appears to be inconsistent with his argument in other parts of the book, or less than clear in terminology. Banaji suggests at one point ‘that we see the twelfth to fifteenth centuries as the period of growth of capitalism in Europe (‘Mediterranean capitalism’) and the sixteenth to eighteenth centuries as the period of Company-capitalism marked by more brutal methods of accumulation and competition’ (pp.255-6). Yet at other points he appears to make it clear that capitalist relations of production do not exist, neither are they emergent unless several elements are articulated together (pp.29-30).

The ambiguity can be resolved. The first passage referred to above is a comment ending a very useful discussion on Marx’s concept of commercial capital, and the need to understand that concept differently in epochs before mature capitalism. Therefore, read carefully in context, Banaji may mean simply that from the twelfth century there is a growth of trade and industry, and therefore of certain kinds of ‘capital’ in western Europe, not ‘capitalism’ in the sense of a nascent mode of production (compare also with the comments and quotation from Marx on p.252). The trade networks and industries of the west up to about 1400 seem to me better illuminated by their analysis as an integral part of a maturing feudal economy (arguably perhaps with the momentary exception of fourteenth-century Florence). If such distinctions are not made carefully, the analysis would seem to slip back into a view that the emergence of trade and industry, necessarily therefore of ‘capital’ in a simple sense, is synonymous with ‘capitalism’ as a social system. Similarly Banaji’s comments on medieval Arab economic writing and activity certainly demonstrate the sophistication of the social formation, but would benefit from a clearer distinction between a monetised economy and incipient capitalist relations of production (pp.262-8).

In contrast, Banaji makes a persuasive demonstration of the particularity of the dynamics of trade and industry within each particular historical system in his brief discussion of the Mughal Empire. Here he sketches a ‘law of motion’ of a kind for the economy of the Empire which shows its growth to be self-limiting by the tendency of regional prosperity to dissolve the larger polity (pp.34-8). Banaji implies here a very different underlying set of social relations which govern commercial and industrial dynamics than could be found in contemporary western Europe. There is again some inconsistency in comments here, as he characterises Mughal India as a representative of a coherent ‘tributary mode of production’, (p.40), usefully distinguished from feudalism, while elsewhere he does appears to accept it as ‘feudal’ (p.91). This is in the end a relatively trivial problem. Nonetheless, a longer analysis here, engaging with Irfan Habib’s re-working of the ‘Asiatic mode’ for the development of Indian societies (Habib 2002), surely would have huge potential to re-invigorate Marxist historical theory. The discussion has tended to be bogged down in category definitions, rarely showing signs of Marx’s own dialectical approach to the workings of capitalism.

Banaji manages to cover nearly every major area of controversy among Marxists over pre-capitalist modes of production, but a very large proportion of the book is taken up with a scholarly polemic against the Marxist historian Chris Wickham and his Framing the Early Middle Ages: Europe and the Mediterranean 400-800 (Oxford 2007). The overall argument of Wickham’s book is a key application to historical practice of the approach to modes of production being critiqued by Banaji. It is a major work of scholarship, but remains disappointing from a theoretical perspective. It effectively sidelines rather than solves the problem of the transition from antiquity to feudalism. Still less does it explain why the eastern half of the Roman empire survived for around thousand years (depending on your criteria) after the fall of the western half. Part of the reason for this is Wickham’s acceptance of the position defined by John Haldon in The State and the Tributary Mode of Production (Verso 1993). Here it is argued that it is not possible to hold that rent, tax and tribute belong to different modes of production, as the shifts between them appear too frequently and easily. Certainly it is true that in early medieval England, for example, it is difficult to pin down various exploitative relations as they appear to any one of those three terms in a definitive sense before, perhaps, the ninth century. Another important example for this general argument is medieval Anatolia, where if tax and rent are held to define modes of production, then the region would have switched back and forth between ‘modes’ over the course of a bare century or so.

Both these examples ought to be a clue that the wrong question is being asked of the evidence. Wickham, however, accepts the argument and proceeds on the basis that therefore all pre-capitalist modes of production, including that existing in the Roman empire, were ‘feudal’. The problem of transition is thus solved, as there was in fact no fundamental change, except where class society collapsed entirely in the fifth and sixth centuries, as in Britain. Yet, if there is no transition between modes of production, then Marxist theory cannot explain the survival of the eastern Empire beyond making some basic points, that it was richer for example, which any historical school might propose. In analysis that follows it sometimes appears that Wickham cannot find any particularly profound differences between the base economies of early medieval Egypt and Italy, while the overall contrasts between, say, Lombard Italy and early Islamic Egypt are quite striking.

The many different discussions across Theory as History show an alternative methodology that exposes the approach of Haldon and Wickham, amongst others, as being positivist and reductionist. Towards the close of the book, Banaji is able to explain straightforwardly: ‘Relations of production are not reducible to given forms of exploitation of labour’ (p.353). The point should really be obvious: wage labour, in production, has existed in most societies to one degree or another, but that does not mean that capitalism existed in embryo everywhere that wage-labour did. Nor is it sound to identify feudalism specifically with serfdom, as Maurice Dobb did. The use of tax or rent to exploit peasants indeed does not distinguish Song China from Byzantium or Capetian France, but that does not mean that these societies are identical at the level of their modes of production.

While the disagreement between Wickham and Banaji lies fundamentally on the methodological level, at points the controversy becomes one of empirical contrasts. Wickham is keen to maximise the extent to which tenant relations prevailed across the whole of the post-Roman world, east and west. Conversely Banaji is concerned to maximise the extent to which aristocratic domain-holding, based on the exploitation of landless labourers of one kind or another, prevailed across the whole region.

It seems likely that Banaji overestimates the importance of ‘landless’ labourers in late Roman Egypt. He fails to acknowledge the extent to which aristocratic grants of small plots of land (‘wages’ according to Banaji) are perhaps a sign that these people are better seen as a semi-bonded peasantry than a rural proletariat (see Banaji, 2001/7, pp.184-5). They are certainly not simply tenants, however, as Wickham’s analysis would seem to prefer. Wickham is more likely to be predominantly correct about the dominance of peasants and tenant relationships in the west of the post-Roman world, while Banaji is probably correct about the ability of the eastern aristocracy to virtually dispossess (and semi-enslave?) the peasantry in the east.

Where Banaji is ultimately more convincing, it is in his ability to interpret the sparse evidence in a way that shows the dynamic of social change, while Wickham’s picture is oddly static. The analysis of Egypt, which is given in much greater detail in Banaji’s earlier book Agrarian Change in Late Antiquity (Oxford 2001), shows an increasingly dominant aristocracy dispossessing lesser ‘middle class’ village landlords. This process helps explain the sudden growth of the movement of ascetic holy men, known subsequently as the Christian ‘desert fathers’. The best known of these figures, the hermit St Anthony of Egypt, was precisely from the category of lesser landholder. These ascetics were at least implicitly critical of the existing social order but never really escaped the position of quietistic religious withdrawal from a corrupt and sinful society. What movement dominated by former property owners was likely to do otherwise? Indeed, these figures are at the origin of the subsequent property-owning monastic orders.

Unlike Wickham’s presentation of apparently static tenant and landlord hierarchies, Banaji’s conception of a contradictory and rapidly developing totality of productive relations suggests connections to be made, and explanations for phenomena outside the strictly economic. If further distinctions can be made between Banaji’s east and Wickham’s west, using Banaji’s methodological approach, then the beginnings appear of a satisfying solution to the old question of the fate of the two halves of the Roman Empire.

One further important consequence of Banaji’s approach to this period, is that the class struggle in pre-capitalist societies becomes once more central to the workings of the system in questions of aristocratic control of property, and resistance to it. The now neglected arguments of E. A. Thompson (less famous than E. P. Thompson, he also left the CP in 1956) on the role of peasant revolts at the end of Empire therefore become relevant again (Wickham, pp.530-3 is oddly dismissive of these Bagaudae). Pre-capitalist class struggles are often, even by Marxists, sidelined as doomed or fatally limited. This is perhaps connected with tendencies towards static structuralist accounts of the past, which fail to see much meaningful change across different social systems.

Yet, different trajectories of class struggle in east and west may have been crucial in the outcomes of the crisis of the Roman Empire. A more dialectical approach to the way particular institutions of exploitation, that is to say particular regimes of rent, taxation or tribute, are embedded in a social totality, would reveal much more about the differing patterns of history across the range of civilisations. A more nuanced approach to how the thorny categories ‘base’ and ‘superstructure’ are handled might also result. A case highly relevant to the transition from Antiquity to feudalism in the West and the Middle East is the role of the Christian Church, the only Empire-wide institution which survived the crisis of the Roman Empire in the west. Marxists have had a tendency to dismiss analysis in this area. This is once again an inheritance of the mechanical Marxism of the Second International and of Stalinism.

Religion was a central frame through which the class struggle was mediated at the end of the Roman Empire. The (Catholic) Church was a property-owning social institution which was to have a crucial role in the maintenance of private property. It preserved forms of the subjugation of labour in the post-Roman west, and it restored class society in the course, for example, of the conversion of the pagan Anglo-Saxon elite in the seventh century. The class society that emerges from the ruins of the Roman west was significantly dependent upon the Church not merely for ideological support, but in terms of the very organisation of private, non-communal landed property. In other words, those communally-living monks under vows of poverty were instrumental in re-imposing private property. The same dynamic was just not present either with Eastern branches of Christianity, or with the Islamic religion which followed. The consequences for the future history of both regions are not trivial, and there are still implications for present understanding of the very different roles of religion in each society.

An approach which would say that tribute to a seventh-century Northumbrian monastery, taxation due to the Caliphs of Baghdad, or money rent due to a fifth-century Egyptian aristocrat were all essentially the same thing, is a form of analysis that has little to offer. The distinctiveness of social institutions through which rent, tax and tribute were mediated, and the consequently various dynamics and possibilities of class struggle in different civilisations, are essential frameworks for working towards a full understanding of differences and similarities among pre-capitalist modes of production across the world. Given that ‘the tradition of all dead generations weighs like a nightmare on the brains of the living’, if we really are to make our own history in the future, we must understand the past in its totality.

Banaji’s Theory as History is a book written at the summit of a lifetime’s engagement with issues of Marxist theory and practice, by a political activist whose deep understanding of praxis led him to engage with what would appear to many to be quite abstruse issues. Yet Banaji’s work demonstrates that no aspect of human history is irrelevant to the present. His scholarship shows immense skill, depth and range. Yet it is his systematic use of Capital and the Grundrisse, to provide analytical guidance in the exploration of pre-capitalist modes of production, that really sets his analysis apart from many other contributions. Over and again, the analysis of pre-capitalist societies has been bedevilled by an acceptance of bourgeois historical assumptions, categories and methodologies. It is not the Marxist method that has been at fault, but the dominance of non-Marxist theory and method in the minds of Marxists. One final lesson therefore that Banaji may bring the reader is on the profundity of Marx’s observation that the thoughts of the ruling class are ruling thoughts. Banaji’s analysis is an immensely useful signpost on the road away from the dominance of those thoughts in our own minds.


Jairus Banaji, Agrarian Change in Late Antiquity: Gold, Labour and Aristocratic Dominance (Oxford, 2001 and revised 2007).
John Haldon, The State and the Tributary Mode of Production (Verso 1993).
Chris Harman, ‘Shedding new light on the Dark Ages’, International Socialist Journal 109 (2006), pp. 187-191. See also an earlier essay by Neil Davidson, ‘Comment’,, November 2004, which explicitly locates ‘modes of exploitation’ as key to defining ‘modes of production’. This passage is wholly non-dialectical; Davidson is searching for parts of a mode of production to define it rather than seeing how elements fit into a totality.
Lester K. Little, Religious Poverty and the Profit Economy (Paul Elek 1978).
E A Thompson, ‘Peasant Revolts in Late Roman Gaul and Spain’, Past and Present (1952), vol. 2, pp. 11-23.
Chris Wickham, Framing the Early Middle Ages: Europe and the Mediterranean, 400-800 (Oxford 2005).

Dominic Alexander

Dominic Alexander is a member of Counterfire, for which he is the book review editor. He is a longstanding activist in north London. He is a historian whose work includes the book Saints and Animals in the Middle Ages (2008), a social history of medieval wonder tales, and articles on London’s first revolutionary, William Longbeard, and the revolt of 1196, in Viator 48:3 (2017), and Science and Society 84:3 (July 2020). He is also the author of the Counterfire books, The Limits of Keynesianism (2018) and Trotsky in the Bronze Age (2020).