The study of feudal society, its birth and its decline is central to Marxist historiography. But what if a foundational theoretical mistake has governed these inquiries for decades, a mistake that has profound implications for Marxist history-writing more generally? This theoretical error is evident in the work of two of the foremost Marxist historians, Perry Anderson and Jairus Banaji. Both Anderson and Banaji can be treated as exemplars of this wider tradition of historians for whom it is axiomatic that something called ‘feudalism’ once operated as a real social totality. Following in the footsteps of Marx’s analysis of capitalism, both of these scholars aim to locate the single economic relationship from which the entirety of feudal society unfolded. As such, neither seriously considers the possibility raised by several twentieth century Marxist theorists of capitalism: that capitalism is in fact unique in possessing such a unifying relation – the commodity structure. This insight has been missing for too long when Marxist historians study non-capitalist social forms. Incorporating it might meaningfully improve our understanding of the specificity of capitalism, casting doubt on some of the more formulaic means by which Marxists study most of human history as well as clarifying the nature of Marx’s objection to capitalism.
One of the crucial claims made by Marx and Engels is that capitalist society forms a unitary totality. Its unity is derived from the fact that its economic relations determine every other aspect of social life. Engels, in an 1890 letter clarifying the Marxist analysis of capitalism, wrote that economic relations are determinant ‘in the last instance.’ For Marx, in his 1857 introduction to the Grundrisse, ‘the simplest economic category ... can never exist other than as an abstract, one-sided relation within an already given concrete, living whole.’ Capitalism exists not just as a collection of different analytic categories but as ‘concrete whole’ given unity by its determinant economic relation. In one traditional view, then, a key task for medieval historians in the Marxist tradition is to mirror Marx’s analysis of capitalist society for the feudal case more rigorously than Marx himself managed: that is, to identify the economic relations that governed this different complex whole.
Perry Anderson’s Passages from Antiquity to Feudalism represents one attempt to form a coherent analysis of the putative feudal totality, notable not least for its creative Althusserian theoretical underpinnings. In For Marx, Louis Althusser offers a distinction crucial to Anderson’s theorisation of feudalism. He differentiates between the determinant contradiction – the economic relations of a society – and the dominant contradiction. For Althusser, the dominant or principal contradiction in a society at any given moment need not always be the economic relations.  While economic relations ultimately determine the complex whole we call a ‘society’, they do not necessarily dominate the structure of that society. There is a second distinction which is crucial for Anderson, and that is Althusser’s distinction between ‘modes of production’ – that term ubiquitous in Marxist historiography to refer to the different ‘epochs’ of human history, but which Althusserians treat only as ‘abstract, pure ideal objects’ – and concrete ‘social formations.’ In a social formation, diverse relations of production co-exist. But as Marta Harnecker clarifies, a social formation ‘is not a combination of modes of production, of abstract or ideal social totalities; it is a concrete, historically determined reality, structured beginning with the form in which the different relations of production which co-exist at the level of the economic structure are combined.’ Read through these categories, the complex whole of the feudal social formation contains multiple relations of production but is structured in such a way that the feudal relations dominate. This provides one solution to the problem of defending the unity of feudal society, a solution put to work adeptly in Anderson’s writing.
For Anderson, while the typical feudal relation is ‘seigneurial jurisdiction over an enserfed rural mass’, the existence of alternative relations of exploitation such as slavery do not challenge feudalism’s unity. They simply demonstrate that feudalism never existed in ‘a pure state’ but rather as the kind of concrete, complex social formation. Anderson then builds on the distinction between domination and determination in his analysis of what constitutes feudal society. While, in the last instance, the dynamic determining the development of feudal society is an economic relation – the conflict between the lord and the peasant over surplus distribution – feudal society is also constituted by a whole set of political relations which take on a dominant character. Anderson writes that ‘the parcellisation of sovereignty was constitutive of the feudal mode of production.’ He is here referring to the manner in which feudal states had no central government with a monopoly on power (especially judicial power), since individual lords’ and church courts also held a great deal of sway. This political relation then structured feudal economic relations; because there was no central power sufficiently strong to comprehensively reorder patterns of land ownership, communal village lands and peasant-controlled allods endured, enabled by the feudal division of sovereignty. Feudalism requires, for Anderson, a specific mode of economic exploitation married with fragmentation of political power. On this definition, feudalism did not come into being until the late ninth century with the decomposition of Carolingian state power and the increased subjection of the peasantry. Anderson builds a model of feudal totality in which unity is maintained by the feudal economic relation dominating other extant relations. This picture has deep empirical problems and only applies fully to France – the highly centralised Angevin monarchy of twelfth century England presents problems for Anderson’s model. But I raise him here as a theoretically distinctive exponent of what we might term trans-historical Marxism, the view that Marx’s economically determined model of (capitalist) social totality has trans-historical relevance as a way of understanding all social forms.
Jairus Banaji provides an important counterpoint as another historian armed with a relatively developed theory of the feudal totality. Banaji sidesteps the issue of slavery in quite a different way from Anderson: in his case, by moving away from a focus on conditions of exploitation – how exactly the surplus is extracted. He is committed to uncovering ‘the laws of motion’ which he sees as defining each epoch. Just as with Anderson, pre-capitalist societies exist as totalities in his understanding. He writes that Marx analysed laws of motion at two levels: ‘at the level of each enterprise and at the level of social totality of enterprises.’ It is worth noting that his claim that capitalist labour could exist under feudalism, much criticised by defenders of Marxist orthodoxy, does not distract from an essentially traditional, if slightly altered, theorisation of unified modes of production. Banaji posits that the law of motion for feudal society was that ‘the lord’s consumption constituted the only “motor-force” of expansion in the feudal economy.’ Notably and deliberately this makes the exact mode of exploitation unimportant in defining the mode of production. Banaji has no qualms in talking about slavery as existing under feudal production. As long as slaves (or any other agricultural labour force) are producing just to meet the needs of their owner rather than in pursuance of profit for its own sake – as in capitalist societies – then they are taking part in feudal production. This definition leads Banaji to identify a far earlier start to feudalism than Anderson. He argues that both the later organisation of bipartite demesnes, which Anderson sees as the start of feudalism, and the earlier post-Roman division into serf plots of land were feudal.
The problem then becomes explaining how the specific relation isolated by either writer can be said to determine the logic of a whole society. That is to say, on what basis can we assert that a freeholding peasant, or any one of those peasants readily identified by Anderson as operating outside his feudal ideal-type, had their mode of living determined by an economic relation in which they were not directly involved? The same question also applies to those living and working in urban areas. It is notable that nowhere in Anderson’s work is there a clear exposition of the mechanism by which one relation dominates the others to form a complex whole. In Anderson’s treatment of the dynamics at work in feudal society, forms of agricultural production that do not meet the feudal definition cited earlier are given a passing reference to show that feudalism was never ‘pure’ in practice, but there is no notion of how the relationship between a master and a slave was ‘dominated’ by the feudal relationship. The ‘unity of the complexity itself’ is simply assumed. Anderson’s model fails to provide a mechanism by which one relation can unify society by dominating all others. Likewise, Banaji does not devote much time to defending his proposition that a feudal mode of production existed because the organisation of production on estates determined the social whole.
In truth, there is a claim (though a relatively subterranean one) about the means by which feudal relations dominate that we can reconstruct in both Anderson and Banaji. Crucially, these two historians both see feudalism as defined by a certain structuring of the ruling class. For Banaji, feudalism exists when a ruling class gears production towards its own needs. For Anderson, with more of an emphasis on specific conditions of exploitation and political formations, feudalism exists when landowners hold certain forms of juridical authority over their enserfed labourers. In both cases, the ability of this relation to determine everything else relies upon the power of the aristocracy. Chris Wickham’s analysis of towns in early Medieval Europe elucidates this point. He writes that one of the crucial determinants in the growth of towns was the presence of aristocratic wealth. Put simply, the more aristocrats in a settlement, the more other people there, since the presence of aristocrats enabled settlements to ‘operate as a production and exchange centre with an immediate market right on the spot.’ It follows that the richer aristocrats were, and the higher their desire for certain tradable goods, the more towns would grow. Wickham thus gives us an account of how traders and artisans in cities could be influenced by the relations pertaining on a lord’s estate. Importantly though, the problem is that this form of determination is agential rather than uniform. Whether or not all spheres of society are determined by the feudal relation is dependent on the position and the contingent level of power exercised by one specific agent – the feudal aristocracy. It is certainly possible that there were moments when this agent held such a total monopoly on wealth that only they could really determine the growth or contraction of towns. But to claim that this was true for the entire ‘feudal epoch’ – however one wants to date that – is simply not historically tenable. The same is true for freeholding peasants, whom archaeological evidence demonstrates did exist in fairly high proportions during the Early Medieval period, most notably in northern Spain. Only at some specific moments were the aristocracy powerful enough to determine their ways of living with anything like the same hegemonic level of control that those aristocrats wielded on their own feudal estates. Reliant on the strength of agents alone, feudalism could not constitute a unified social totality over a long span of time and space.
That should serve as the point of departure for a quite different conclusion than the one ventured by these Marxists: an analysis of the radical difference between the Medieval world and the capitalist totality. Capitalism is uniquely constituted by an impersonal economic relation – powered by ‘abstractions’, in Marx’s rendering in the Grundrisse, and not the whims of the ruling class alone. The social theorist Moishe Postone unpacked this argument especially forcefully. Here is Postone, in a 2016 interview:
I think analysing Marx’s argument in Capital calls into question the notion that you have any unified modes of production before the historical emergence of capital, which is unified in the sense that you can begin with a singular principle, the commodity, and you can unfold that to encompass the whole. You cannot find something analogous in other forms of social life, in part because the possibility of unfolding the social whole from a singular point of departure is possible only because, in capitalism, the mode of mediation is uniform.
For Marx, the commodity is an object in which ‘the social character of men’s labour appears to them as an objective character stamped upon the product of that labour.’ The exchange-value of the commodity appears to us in the circulation process as one of its intrinsic attributes, rather than as the result of labour. Capitalism is a sea of such commodities, all labour produces them, so that life becomes dominated by what György Lukács termed the ‘commodity structure’: wherein ‘a relation between people takes on the character of a thing and thus acquires a “phantom objectivity”.’ All economic relationships between individuals – whether one is employing the other or one is selling the other something – are represented as and so appear to be a relation between objects. Employment appears as the act of the labourer selling the commodity labour-power to their employer. This commodity structure gives to capitalist society a unity absent from previous social formations. To return to Lukács: commodities are ‘constitutive’ of society, ‘they penetrate society in all its aspects and remould it in its own image.’ This is importantly a logic structured into participation in capitalist society, which in this sense is law-governed rather than relying entirely on the ability of any particular agent to discipline society. No matter the relative power of the bourgeoisie, every individual in capitalist society has to buy, sell and produce commodities. While lords with fiefs ruled in most conceptions of feudalism, capitalists do not rule in any simple sense under capitalism. Instead the commodity structure rules by constituting all forms of life for both the capitalist and the worker. As Marx put it, ‘capital itself rules the capitalist.’ Only when the personal domination constitutive of the Medieval world gave way to the abstract domination of the commodity structure could a social totality come into being. Only at this point did the degree of unity of a society cease to be contingent on the power of its ruling class. There is no uniform mode of mediation like the commodity constituting society in the Medieval period. Therefore, the search for any other all-determining economic relation is bound to be elusive.
This claim has deep implications for Marxist historiography. In a 1968 lecture, Theodor Adorno argued that the exchange of commodities makes capitalist society a totality rather than just ‘an ordered agglomeration of facts.’ The latter, then, is what the Medieval world becomes once we acknowledge its lack of a unifying relation. This means that not only that the quest to find such a relation is redundant, but also that the emphasis on that quest hinders an accurate reconstruction of past society. To return to Anderson as an example, once he has isolated ‘the feudal dynamic’ he then sees its hand behind all developments in the Medieval social formation. This method needs to be adjusted significantly. So too does the very understanding of historical time in his work. The transition to feudalism is understood as simply a precursor to the High Medieval period; in the transition tendencies are set in motion that will come to fruition later. This is by no means an error limited to Marxist work; across Anglophone historiography the period c. 400-900 is known as either Late Antiquity or the Early Medieval period. Both titles presuppose that the era existed as either the death throes of Rome or the birth of Medieval Europe. However, a conception of history which sees the social organisation of these centuries as a transition or, in Banaji’s eyes, a less developed version of what was to come, is most susceptible to writing off whole swathes of history as mere introductory moments. Thus, the specificity of the ‘Early Medieval’ world is lost, its phenomena are not analysed as categories in their own right but rather awkwardly bundled into a presupposed schema that runs from slave mode of production to feudalism to capitalism. If we no longer see history as a series of unified epochs, with moments in between functioning solely as transitions from one totality to another, then how we periodise history and indeed our very understanding of what it means to speak of an ‘historical period’ changes a great deal. Finally, historical investigation proceeds not just by the collection of facts but also by choices of emphasis. Historians choose which phenomena in the past they deem most worthy of investigation. For Marxists, the question of the organisation of dominant forms of surplus extraction – agricultural estates for feudal society – has frequently been taken for granted as the central topic of study. Herein, supposedly, lay the key to wider society. If all other ways of living and producing – from commercial trade to independent peasant production – are no longer seen as secondary, then they might become far more important topics for Marxist historians to study.
The conception of capitalist totality advanced here has followed the Hegelian model embraced by the Frankfurt School, but they are not the only Marxists who have stressed the historical narrowness of Marx’s analytical object. In 1977, Althusser responded to the unfolding ‘Crisis of Marxism’, with these words:
Marxist theory (if we leave aside the temptation of the philosophy of history to which Marx himself sometimes succumbed, and which dominated in a crushing fashion the Second International and the Stalinist period) is inscribed within and limited to the current existing phase: that of capitalist exploitation.
The image of Marx ‘sometimes succumbing’ to the philosophy of history is a useful one. There is no clear definition of the mode of production in his work. It is far and away most commonly used to describe capitalism but generations of historians, some taking Althusser’s earlier writing as their inspiration, have attempted to formulate theories of pre-capitalist modes of production, taking their existence as a given. In doing so, they not only impoverish their discussions of non-capitalist worlds. They also trans-historicise a unique feature of capitalism, its unity around one determinant relation. They take a theory designed to explain the workings of capitalist society and attempt to apply to it altogether different phenomena (borrowing from a tendency Marx and Engels sometimes displayed, it is true). Marxist historians might object that this line of criticism simply rules out the usefulness of studying non-capitalist social forms, but in truth it gives that study a novel political importance for anyone interested in the supersession of capitalist society. If a study of feudalism highlights rather than obscures the exceptional nature of capitalism, then it is a valuable exercise even for those who have no interest in kings and castles.
Hugo Raine studies Medieval history at Wadham College Oxford
 F. Engels, trans. S. Hook, ‘Engels on Historical Materialism’ accessed at: https://www.marxists.org/history/etol/newspape/ni/vol01/no03/engels.htm
 K. Marx, trans. M. Nicolaus, Grundrisse, Foundations of the Critique of Political Economy (Penguin Books: London, 1993), p. 101.
 For Althusser, society forms a ‘complex whole’ which has ‘the unity of a structure articulated in dominance’, but ‘it is economism that identifies eternally in advance the determinant contradiction-in-the-Iast-instance with the role of the dominant contradiction’. See L. Althusser, trans. B. Brewster, For Marx, (London: Verso Books, 2005), pp. 202, 213.
 M. Harnecker, ‘Mode of Production, Social Formation and Political Conjuncture’ in Marxism Today, (February 1980), p. 26.
 P. Anderson, Passages From Antiquity to Feudalism, (London: Verso Books, 1978), p. 156.
 Ibid., p. 154.
 Ibid., p. 183.
 Ibid., p. 148.
 Ibid., p. 149.
 Ibid., p. 142.
 J. Banaji, Theory as History, (Chicago: Haymarket Books, 2011), p. 46.
 Ibid., p. 59.
 T. Brass, ‘Jairus Banaji's Mode of Production: Eviscerating Marxism, Essentialising Capitalism’, in Journal of Contemporary Asia, Vol. 42, No. 4, (November 2012), p. 709.
 Ibid., p. 75.
 Ibid., p. 74.
 P. Anderson, op. cit., pp. 182-184.
 L. Althusser, op. cit., p. 202.
 C. Wickham, Framing the Early Middle Ages, (Oxford: Oxford University Press, 2005), p. 595.
 A. Rio, Slavery After Rome, (Oxford: Oxford University Press, 2017), p. 181.
 M. Postone, ‘An Interview with Moishe Postone’ in Crisis and Critique, Vol. 3, No. 3, (June 2016), p. 509.
 K. Marx, trans. S. Moore & E. Aveling, Capital Volume I accessed at: https://www.marxists.org/archive/marx/works/1867-c1/ch01.htm#S4
 G. Lukacs, trans. R. Livingstone, History and Class Consciousness (Cambridge: The Mitt Press, 1922), p. 83.
 Ibid., p. 85.
 K. Marx, trans. T.B. Bottomore, Karl Marx: Early Political Writings, (London: Watts, 1963), p. 85.
 T. Adorno, Introduction to Sociology, (Stanford: Stanford University Press, 2000), p. 32.
 L. Althusser, trans. G. Lock, The Crisis of Marxism, accessed at: https://www.viewpointmag.com/2017/12/15/crisis-marxism-1977/[book-strip index="1" style="display"]
Adorno, Theodor, Introduction to Sociology, (Stanford: Stanford University Press, 2000).
Althusser, Louis, trans. Brewster, Ben, For Marx, (London: Verso Books, 2005).
Althusser, Louis trans. Lock, Graham, The Crisis of Marxism, accessed at: https://www.viewpointmag.com/2017/12/15/crisis-marxism-1977/.
Anderson, Perry, Passages From Antiquity to Feudalism, (London: Verso Books, 1978).
Banaji, Jairus, Theory as History, (Chicago: Haymarket Books, 2011).
Brass, Tom, ‘Jairus Banaji's Mode of Production: Eviscerating Marxism, Essentialising Capitalism’, in Journal of Contemporary Asia, Vol. 42, No. 4, (November 2012) pp. 707-716.
Engels, Friedrich, trans. Hook, Simon, ‘Engels on Historical Materialism’ accessed at: https://www.marxists.org/history/etol/newspape/ni/vol01/no03/engels.htm.
Harnecker, Marta ‘Mode of Production, Social Formation and Political Conjuncture’ in Marxism Today, (Feb. 1980), pp. 23-31.
Lukacs, Gyorgy, trans. Livingstone, Rodney, History and Class Consciousness (Cambridge: The Mitt Press, 1922).
Marx, Karl, trans. Nicolaus, Martin, Grundrisse, Foundations of the Critique of Political Economy (Penguin Books: London, 1993).
Marx, Karl, trans. Moore, Samuel & Aveling, Edmund, Capital Volume I accessed at: https://www.marxists.org/archive/marx/works/1867-c1/ch01.htm#S4.
Marx, Karl trans. Bottomore, Thomas, Karl Marx: Early Political Writings, (London: Watts, 1963).
Postone, Moishe, ‘An Interview with Moishe Postone’ in Crisis and Critique, Vol. 3, No. 3, (June 2016) pp. 501-517.
Rio, Alice, Slavery After Rome, (Oxford: Oxford University Press, 2017).
Wickham, Chris, Framing the Early Middle Ages, (Oxford: Oxford University Press, 2005).