Marxism and Theoretical Overkill

Courtesy: Weekly Workers, Communist Party of Great  Britain

Mike Macnair reviews Jairus Banaji's 'History as theory: essays on modes of production and exploitation' Historical Materialism books series, Vol 25, Leiden, 2010, pp406.

Part I
In March 2010 the Indian novelist, Arundhati Roy, published in the journal Outlook India a substantial and sympathetic report of the activities of the Naxalite (Indian Maoist) guerrillas in Chattisgarh state in eastern India.Roy’s report has been very widely circulated on the web. It has also been the subject of furious attacks from Indian establishment politicians and the threat of prosecution under ‘anti-terrorism’ laws (though more serious threats to prosecute Roy, this time for sedition under the Indian penal code, have been made in relation to another article which supported the secession of Kashmir).

Shortly after Roy’s article was published, leftist and academic Jairus Banaji posted a short sharp critique of it on the Indian political blog Kafila. If Roy’s original article was savaged by the Indian political establishment, comrade Banaji’s critique has given rise to almost equally sharp polemics on the Indian left.Banaji has elaborated his critique in a substantial article, ‘The ironies of Indian Maoism’ in the autumn 2010 issue of the Socialist Workers Party’s theoretical journal, International Socialism.

Why is this current political debate relevant to History as theory, Banaji’s collection of essays written between 1976 and 2009, mainly on the problems of Marxist interpretation of ancient and medieval history? The answer is that Banaji’s theoretical arguments are in the last analysis targeted on those used by Indian ‘official communists’ and Maoists in support of their respective political lines.

‘Official communists’ argue, the world over, for a strategic alliance between the working class movement and sections of the bourgeoisie. In the old central imperialist countries this is usually presented as an ‘anti-monopoly’ alliance. In the countries which were formerly colonised, in contrast, the argument is that capitalism is not fully developed, because of colonial or neo-colonial subordination: there are significant ‘survivals of pre-capitalist relations of production’. ‘Official communists’ claim that it is therefore necessary to ally with the ‘national’ bourgeoisie against imperialism and/or with the ‘democratic’ bourgeoisie against the old landlords and similar classes to ‘complete the bourgeois revolution’.

Maoists classically argued that the same ‘survivals of pre-capitalist relations of production’ mean that the prime revolutionary class is the peasantry. Just as - according to Maoists - the working class of the imperialist countries forms a labour aristocracy relative to that of the colonial countries, so the urban working class of the colonial countries forms a labour aristocracy relative to the rural exploited classes. The strategy for revolution is therefore to ‘surround the cities’. It is this strategy that the Naxalites have been attempting to apply, with very varying levels of success, in parts of India - mainly in eastern states - since the 1970s.

There are very substantial objections to the arguments both of the ‘official communists’ and of the Maoists which can fall within the same general framework of the development of capitalism out of pre-capitalist societies, and the idea that some pre-capitalist relations of production survive in ‘third world’ countries, including India.

For example, both the ‘old Bolsheviks’ of Lenin’s time and Trotsky alike argued: (a) that the capitalist class would not seek to overthrow the pre-capitalist state (because it was more afraid of the rising working class than of the declining pre-capitalist classes); and (b) that the peasantry could only play a revolutionary role if the urban proletariat took the lead. They differed as to whether contradictions between the urban proletariat and the peasantry would mean that the resulting regime would fail in the absence of immediate support from the western proletariat taking power (Trotsky) or whether a ‘democratic dictatorship of the proletariat and peasantry’ could be (relatively) stable (Lenin). But neither would have agreed with a strategic class alliance with the bourgeoisie (the line of the Russian Mensheviks) or with peasant leadership in the revolution (the line of the Russian Socialist Revolutionary Party or Narodniks).

Equally, but less immediately dependent on Russian debates, it could be argued: (c) that the global course of events since 1945 has shown that the ‘national’ or ‘democratic’ bourgeoisie is an utterly untrustworthy ally for the working class; or (d) that the Maoists’ narrative of the Chinese revolution as a guerrilla struggle based on the peasantry and ending with ‘surrounding the cities’ is false: the Chinese Red Army in the 1940s was a large regular field army controlling substantial territory and supplied with munitions by the USSR. And the Chinese Communist Party of this period, if in a sense it based itself on the peasantry, continued to recruit cadre from the urban classes.

Banaji’s objections are more fundamental than these. In ‘Ironies’ he argues that a substantial part of what the Naxalites identify as ‘peasants’ are in reality already rural proletarians. Hence the Naxalites succeed - in their base-building phases - when they build what are in substance local rural proletarian mass movements. And hence they fail - in their bids to hold onto and govern territory against the Indian state - when they try to follow the Chinese example.

In History as theory, Banaji goes further. He argues that the whole ‘traditional Marxist’ scheme of differences between modes of production which are defined by the mode of exploitation - slavery in classical antiquity, serfdom under feudalism, wage labour under capitalism - is to be rejected. This scheme is, he says at several points, “teleological” (without explaining what he means by that). The objections are backed by depth empirical research, which he claims has been lacking in many Marxist writers who support the mode of exploitation schema.

This running argument makes History as theory more than ‘selected essays’. It ties together into a single argument chapter 2, ‘Modes of production in a materialist conception of history’ (1977); chapter 3, ‘Historical arguments for a logic of deployment in pre-capitalist agriculture’ (1992); chapter 4, the previously unpublished ‘Workers before capitalism’; chapter 5, ‘The fictions of free labour’ (2003); chapter 6, ‘Agrarian history and the labour-organisation of Byzantine large estates’ (1999); chapters 7 and 8, two critiques of Chris Wickham’s Framing the early middle ages (Oxford 2005), one new and one from 2009; chapter 9, ‘Islam, the Mediterranean and the rise of capitalism’ (2007); chapter 10, ‘Capitalist domination and the small peasantry; the Deccan districts in the late 19th century’ (1977); and two new concluding chapters, 11 and 12, ‘Trajectories of accumulation or ‘transitions’ to capitalism’, and ‘Modes of production: a synthesis’.

If these arguments of Banaji’s are right, those of the Indian ‘official communists’ and Naxalites are not merely falsified in the way that Trotskyist or old Bolshevik objections would falsify them. They fall to the ground as irrelevant to reality, because based on a false a prioriconstruction about historical development.

Overkill

At the same time, however, if the full effect is given to Banaji’s negative critique of the ‘traditional Marxist’ scheme, but no positive alternative scheme of general historical development is put in its place, Marx’s and Engels’ core arguments for the leading role of the proletariat in the struggle against capitalism also fall to the ground and for the same reason. What is left is merely an ethical or utopian socialism. This ethical or utopian socialism may prioritise the working class, as Banaji’s actual politics does. But it lacks serious and solid grounds for supposing that working class self-activity under capitalism points towards a future without capitalism. The result, in other words, is theoretical overkill.

There is a sense in which theoretical overkill is predictable from Banaji’s history. Banaji studied classics at Oxford University in the 1960s and went on to masters-level postgraduate work there before returning to India in 1972. As a student he became a member of International Socialism, the precursor of the SWP. In India, campus activism at Jahawarlal Nehru University Delhi in 1972-74 was followed by labour research and organising in Bombay in the late 1970s-80s.

In the late 1980s Banaji returned to Oxford and to classics to write a doctoral thesis on the late antique agrarian economy in Egypt, presented in 1992, which was published in a revised form in 2001 asAgrarian change in late antiquity: gold, labour and aristocratic dominance (Oxford). Agrarian change, though its origins as a doctoral thesis make it tightly argued and densely documented, is plainly part of the same general project on agrarian relations and modes of production as the essays in Theory as history. Since the 90s Banaji has held a range of senior research posts in various universities.

His joining IS when he was a student in the late 60s/early 70s cannot have been an ‘only show in town’ decision. Oxford University at the time had a significant Communist Party with a real intellectual life. The Healyite ‘orthodox Trotskyist’ Socialist Labour League was mainly based at the car factory, but also had a significant presence on campus. Later a mad sect, the SLL at this period intervened seriously in the academic left as well as the trade unions. The Mandelite International Marxist Group started its presence in the town at Ruskin, the trade unionists’ college, but by 1969-70 was significantly present on the university campus. The Maoist CPB (Marxist-Leninist) had a small but active branch in the city. And, of course, there were the usual recurring, ephemeral, semi-organised groups of anarchists, libertarian socialists, situationists, etc, who were and are found in every university town. The Oxford far left had various common projects in which they worked together, polemicised with each other and so on. Banaji’s choice to go with IS must have reflected not merely general radicalisation and activist commitment, but an active preference for IS over the alternatives.

The SWP is today a fairly standard ex-Trotskyist group evolving towards a sectarian left version of ‘official communism’. But in the late 1960s to very early 1970s, before the ‘Bolshevisation’ of the mid-1970s and the ‘party turn’ of 1977, the IS was something quite different. Its international politics were closer to those of today’s Alliance for Workers’ Liberty. Its ‘class struggle strategy’ for Britain and Europe was closer to that of today’s Commune group.

The IS’s origins were in a line of argument most clearly summed up in Tony Cliff’s State capitalism in Russia (1955): the argument that the USSR was a ‘state capitalist’ regime. Some other authors have suggested that ‘state capitalism’ in Russia was an extreme form of the protectionism and high-level state control common to transitions from feudalism to capitalism. For Cliff, in contrast, it was an expression of capitalist decline, a stage of monopolised capitalism beyond imperialism.

This argument was - as I suggested above that Banaji’s argument inHistory as theory may be - theoretical overkill. The post-war ‘official’ Trotskyists, following Trotsky’s 1939-40 arguments on the partition of Poland, said that the Sovietisation of eastern Europe and the Chinese revolution was somehow ‘progressive’ (whatever the explanation). Cliff’s theory rejected this in the most categorical way possible: the regime was part of the obvious enemy, capitalism.

But to achieve this result involved identifying as ‘capitalism’ a regime without unemployment (instead there was massive make-work and low labour productivity), which suffered from endemic and episodic sectoral underproduction, not from cyclical crises of overproduction.

It also involved casually conflating capitalism with pre-capitalist modes of production and these with each other. Thus Cliff at one point inState capitalism in Russia made an analogy between the Soviet economy and those of ancient China, Egypt and Babylonia (traditionally described by Marxists as examples of the ‘Asiatic mode of production’); at another he cites state ownership of the land under the Mamluk regime as ‘Arab feudalism’, showing that class society is compatible with the absence of private property in land.

New left and ‘teleology’

As well as this theoretical overkill, the IS in the late 1950s to early 1970s was deeply influenced by the ‘new left’, which emerged after the crisis in the western communist parties caused by ‘de-Stalinisation’ and the 1956 Hungarian revolution. In the 70s some ISers - notably Banaji’s slightly younger contemporary at Oxford, Alex Callinicos - were also influenced by French left-‘official communist’ theorist, Louis Althusser.

The question of ‘modes of production’ was problematic for the ‘new left’ in four ways, two coming from ‘official communism’ and two from the western academy. The first was that ‘official communist’ doctrine justified the tyrannical character of the Soviet and similar regimes as a regrettable necessary stage in the transition from capitalism to communism proper. The second was that this doctrine was also used to justify further ‘necessary stages’: the ‘advanced democracy’ which was supposed to be the outcome of the ‘anti-monopoly alliance’ in the imperialist countries, and the necessary capitalist stage in the colonised/neo-colonised countries.

The third problem was the great emphasis placed on the supposedly teleological character of Marx’s account of history by Karl Popper and a broad range of sub-Popperian authors across several academic disciplines: authors who built on Max Weber’s ‘ideal types’, opponents of ‘historicism’ in anthropology, and so on. The fourth was the US state funding of social democratic politicians and authors in the cold war period. This meant - for example - widespread willingness to deploy Karl Kautsky’s theoretical objections to the ‘prematurity’ of the Russian Revolution, based on a theory of necessary stages, in favour of the idea of ‘Leninism opposed to Marxism’.

‘New left’ authors and the activists who used their ideas responded in two ways. One - as it were the right wing of the ‘new left’ - could perhaps be called ‘premature Eurocommunists’. ‘Marxist humanists’ like Roger Garaudy in France and EP Thompson in England called on the ethical and humanistic elements of the writings of the early Marx against the ‘scientism’ of the later Marx, Engels, Kautsky and Stalin. They accepted the general frame of capitalist constitutionalism as protecting important liberties that Stalinism destroyed; and they retained the general political framework of the people’s front policy, merely getting rid of the ‘inhuman’ role of the party.

The second line of approach was to resurrect the arguments of the revolutionary syndicalist, Georges Sorel, in The decomposition of Marxism (1908) - for the most part not directly. Rather the arguments used were those of authors within the socialist movement, but to some extent influenced by the revolutionary syndicalists, like Anton Pannekoek; and authors from the left wing of the early Comintern, like Karl Korsch and (in the early 1920s) Georg Lukács and Antonio Gramsci. Rosa Luxemburg, viewed pretty much exclusively through the prism of Reform or revolution (1900) and The mass strike, the political party and the trade unions (1906) became almost the totem of this sort of ‘new left’.

This trend’s political inheritance from Sorel was a syndicalist focus on the immediate class struggle at the point of production - strikes, and if possible unofficial ones. Its theoretical inheritance was the belief that the ‘historical materialist’ arguments shared by the young Marx and Engels (and, in reality, by the late Marx and Engels) could be discarded. The centre of Marxism was Capital and, in particular, the Hegelian dialectical exposition of the first part of volume 1 of Capital.These arguments provided the exclusive ground for the leading role of the working class and could be read to give a central role to the strike as the moment at which the working class, otherwise merely within capitalism, became an actor against it.

The ideas of this ‘left new left’ could be mixed up with elements taken from Maoism or from Che Guevara. By the late 1960s it had also had a profound influence not only on the IS, but also on the ‘official’ Trotskyist Unified Secretariat of the Fourth International. The ‘orthodox’ Trotskyist opponents of this influence, in so far as they did not collapse simply into ‘official communism’ (the US SWP, and so on) have since then largely collapsed into it themselves. It has thus shaped the ideas of the far left well beyond people who are conscious of its origins.

Louis Althusser was a left academic within the Parti Communiste Français somewhat sympathetic to Maoism, who deployed a highly selective version of the ‘left new left’ critique of historicism and ‘historical materialism’ to provide arguments against the PCF’s ‘Marxist humanist’ critics. Althusser in a certain sense preserved the idea of distinct modes of production, but he did so by effecting a complete severance between modes of production, which cease to form a general historical narrative. Rather Althusser adapted the ‘structural anthropology’ of Claude Lévi-Strauss, itself derived from the ‘structural linguistics’ of Ferdinand Saussure. The key move was that ‘synchrony’ (structural causation within a mode of production) overdetermines ‘diachrony’ (historical development). Like the ‘left new lefts’ Althusser insists that the only real Marxism is that of Capital, claiming that an ‘epistemological break’ lay between the early Marx and the Marx of Capital.

The intellectual context within which Banaji constructed his early arguments was thus one in which it was an orthodoxy not needing much explanation that the ‘Stalinist’, ‘Kautskyite’ or ‘Engelsian’ sequence of modes of production (primitive communism, Asiatic mode, slavery, feudalism, capitalism, communism in a higher form) was wrong, teleological, deterministic-automatistic, and so on, and that Marxism had to start from Capital Vol 1 and nowhere else.

Within this framework, references to ‘modes of production’ have to be read not as phases of a narrative, but as ‘ideal types’ in the style of Max Weber: or, to put it another way, as a historian’s equivalent of the ‘comparative statics’ which are foundational in the various forms of marginalist economics.

These assumptions remain present in Banaji’s work down to the present day as assumptions, not argued positions. What his work does is to offer empirical historical evidence, informed by economic analysis, against authors and tendencies who do use the sequence of modes of production in historical and political argument.

True or false?

What I have said so far perhaps helps to explain why Banaji’s arguments have taken the shape they have. But it certainly does not prove they are false. On the contrary, the essays are very high-quality historical work.

It seems to me that Banaji succeeds in demonstrating certain of his specific claims. In particular:

1. There was very substantial use of wage labour in agriculture (and elsewhere) in many pre-modern societies. (Chapters 3, 4 and 6).

2. There is a spectrum between the considerable degree of freedom (of movement, of choice of employer, etc) of many workers in the more developed capitalist countries and the total unfreedom of chattel slaves. Neither the chattel slavery of Africans in the early modern to 19th century plantation economies nor forms of indentured labour, debt-bondage, sharecropping and so on, then or more recently, can be said to show the existence of (in any strong sense) pre-capitalist social relations of production in a country (chapter 5).

3. Following the last two points, phenomena of labour relations at the point of production alone cannot be used to identify the mode of production in the larger sense or to describe the larger society as pre-capitalist (passim in the book).

4. Following on from all this, the ‘Brenner thesis’ that capitalism emerged in England as a result of a specific mutation in labour relations in agriculture is to be rejected. Rather capitalism, at least in its modern sense, emerged in the later middle ages in the Mediterranean interface of Catholic Christendom, Byzantium and the Dar al-Islam (chapter 9).

5. Indian agriculture in the 19th century was dominated by capitalistrelations, although these were mainly ones of (in Marx’s terminology) the formal subsumption of labour under capital (household commodity production dependent on and organised by merchants and moneylenders) rather than ones of the real subsumption of labour under capital (large-scale shipping, factory production and mechanised or semi-mechanised large-scale farming).

Any narrative of historical materialism will therefore have to take serious account of Banaji’s arguments and evidence on these issues.

On the other hand, a number of Banaji’s assumptions, and in some cases his formal claims, are more problematic. It will clarify what follows to state some points as briefly as possible. A second part of this review will provide more supporting argument for some of these points.

First and at the most superficial level, it seems to me that the line of argument connecting reformism and Stalinism to historical materialist ‘automatism’, ‘scientism’, etc is false as a characterisation of the history of the workers’ movement and leads to dead-end politics. I have made this argument elsewhere and will not elaborate further here, since, though I think it is part of the background to Banaji’s argument, it is not part of the argument itself.

Second. The argument that the idea of the sequence of modes of production is ‘teleological’ is unsound as a matter of epistemology and historical method, quite irrespective of whether the sequence of historical periods constructed is a Marxist, or any other, interpretation of history.

What we actually have from Marx and Engels on this topic are a few general sketches of the approach (in The German ideology, The Communist manifesto, The Contribution to the critique of political economy); some of Marx’s rough notes; a polemic by Engels against the rival ‘force theory’ (in the Anti-Dühring); Engels’ Origin of the family based on the contemporary anthropology and on aspects of classical antiquity; and some journalism and private correspondence. This material was all written before the publication of the vast bulk of the written sources for ancient and medieval society now available, let alone the information generated by archaeology. To treat Marx’s and Engels’ comments as holy writ for modern historical investigation is therefore obvious nonsense.

There are, however, solid non-historical grounds in human biological nature and our material needs for the core of historical materialism - that the ways in which historical societies produce their material subsistence constrain the sort of general social orders possible. Similar grounds support the rejection of methodological individualism (humans are a social species) and of marginalism (there are physical minimum subsistence levels, maximum working hours and maximum quantities of land). These grounds require the analysis of societies and their dynamics in terms of the social division of labour.

The productive character of this basic approach as a research paradigm in history is evident in historical work produced in the last century (Banaji’s work is a very distinguished example). This evidence also suggests that certain aspects of Marx’s and Engels’ specific arguments and comments on aspects of the past were insightful beyond the historical evidence available to them. But we have to be willing to reconstruct their specific theories and narratives very radically or replace them, so far as this is required by the evidence.The question is whether reconstruction on this sort of basis will produce similar arguments to those of Theory as history.

Origins of the present

Third. Because of his assumptions about teleology and so on, Banaji simply does not address in any systematic way the problems of grand-scale narrative of the origins of the present and of historical periodisation and transitions. Nor, apart from in the very concrete chapter 9 on the Mediterranean origins of capitalism, does he address the specific aspect of this problem which presents itself as the historical, economic and social-scientific debates about the origins of European priority in the development of capitalist modernity.

The result is that the most theoretical sections of the book - chapters 1 and 2, 10 and 11 - are relatively disappointing: their outcome seems to be merely negative critique, and fiddling round the edges of the classical Marxist scheme.

There is no engagement with the attempts at a general reconstruction by authors coming from the Marxist tradition, like - for example - Igor Diakonoff’s The paths of history (1999) or David Laibman’s Deep history (2006). Nor are fully non-Marxist attempts to address these issues tackled, like - for example - John A Hall’s Powers and liberties(1986), or Michael Mann’s Sources of social power (1986-93). There is also a large volume of neoclassical and ‘institutionalist’ economists’ writing on history from marginalist perspectives - much worthless, but some containing real insights - produced since the 1980s.

Fourth. This ‘missing link’ affects the plausibility of some of the interpretations in History as theory: in particular, the two chapters of critique of Chris Wickham on the early middle ages (chapters 7 and 8).

The starting point here is a persistent controversy between ‘modernist’ interpretations of the economy of the Roman empire (and, in particular, the late empire), which stress its similarities to European capitalism in the period before steam-driven industry, and ‘primitivist’ interpretations, which stress its differences.

Fashions among ancient historians have shifted on this topic. In the 19th century the ‘modernist’ view was largely dominant. In the 20th, the dominant interpretation shifted towards the ‘primitivist’ side. In the very recent past there has been a shift back towards ‘modernism’. Banaji places himself on the ‘modernist’ side of this dispute, while he argues that Wickham’s theoretical construction assumes ‘primitivism’.

The shifting fashions have an ideological aspect. But the absence of a clear, settled view is also partly due to the severe limitations of the historical sources for the shape of the Mediterranean and European economies before the central middle ages, when tax and judicial records, etc begin to survive in sufficient quantities to be plausibly representative.

In addition, the sources we do have may be analogous to the words linguists call ‘false friends’, where the same word is used in two languages with different meanings. The reason for this is that (as Marx observed in The 18th Brumaire of Louis Bonaparte) Europeans down to the 19th century were very prone to ‘copying’ classical antiquity. Words and forms may therefore be the same in appearance, but very different in content when read in context.

This is intensely true of legal sources. The same limited body of Roman legal texts was read from the 11th-12th century down to the 18th as the basis for medieval law; from the 16th-19th century as the basis for overthrowing medieval law in order to ‘return’ to an imagined ‘law of business Rome’ - ie, create law fit for capitalism; and from the 19th onward as evidence for scholarly interpretations of the Roman society and economy, whether ‘modernist’ or ‘primitivist’.

Hence, legal sources cannot be read as transparently expressing current economic practices. This is not only true of legislation and treatises, but also of written contracts and pleadings in disputes before judges: these documents are products intended to create legal results, and therefore using the language of the law, even if this language refers to long-obsolete practices or involves fictions. Using them as evidence for the social relations of production requires an interpretive context which includes the structure and evolution of the legal order as a whole - which, of course, poses that of the evolution of the social order as a whole - in order to disentangle norm and practice.

In Agrarian change in late antiquity Banaji’s use of legal sources is able to approach success in this task, because the main body of the work is narrowly focused on the specific evolution of the Egyptianagrarian economy in antiquity - which was always a special case and is much better evidenced than elsewhere. In the critiques of Wickham, which necessarily have a broader focus, some of the arguments are more problematic.

This is also, I think, because the underlying structure of Banaji’s negative critique of the relation between mode of production and mode of exploitation in History as theory, and his non-engagement with the larger ‘origins of the present’ issues, does not really allow for full integration of the questions of the evolution of legal norms.

Fifth. By the rejection of ‘teleology’ and so on, Banaji rules out in principle the interpretation of particular forms as transitional. He makes the superficially legitimate point that there is a risk that the ‘transitions’ will swallow up the ‘mode of production’. However, arguably we should read ‘modes of production’ as forms of social dynamics which rise and decline, and which only have strong direct descriptive value at the moment of apogee. (This moment should be placed for feudalism somewhere around the 11th-13th centuries; for classical antiquity somewhere before the fall of the Roman republic; for capitalism - probably - in the 19th century.)

If so, the transitional forms will include not only forms which are visible precursors of a new order, but also specific adaptations to decline which do not foreshadow the new, and blind-alley experiments which fail.

This point has more specific political implications. The formal subsumption of labour under capital in Marx’s writing is transitional between household production and capitalism proper. It is empirically observable that it does not in itself produce the same political dynamics as the real subsumption of labour under capital. This is because only the real subsumption of labour under capital forceslabour to recognise that it is engaged in a cooperative enterprise which is part of a general social division of labour. The implication of this observation is not that Banaji’s critique of Indian Maoism and ‘official communism’ is exactly false, but that, in positing simple labour self-organising as the alternative to these trends, it is not merely theoretical overkill, but also may fail to address issues that need to be addressed.

I will elaborate on some of these points in the second part of the review.

Part II

At the end of the first part of this review of Jairus Banaji's book I made a series of points critical of comrade Banaji's argument, and said I would elaborate on some of them further in the second part. In fact, there will be three parts in all. This part will discuss large-scale theoretical issues: 'teleology', and the grounds for historical materialism. A third part will apply the points made here to the problems of the 'direction' of history, historical periodisation, 'transitions' and 'transitional forms', and look at the concrete political implications of the differences between the 'formal subsumption of labour under capital' (merchant/ moneylender control of production on a household scale) and the 'real subsumption of labour under capital' (large-scale shipping, factory production and large-scale agriculture). Teleology is technically a branch of philosophy which deals with 'final ends' or, in more modern English, the ultimate purposes of life, of all sorts of entities, and so on. Greek philosophers Plato and Aristotle, and - following Aristotle - medieval theologian Thomas Aquinas, drew moral conclusions from reasoning as to the purposes of animals, humans, etc. In the 18th and 19th century teleology, in the form of the 'argument from design', was one of the major intellectual props supporting the idea of a creator-god. Marx and Engels employed 'teleology' in this sense - 'arguments from design' which justify the world as it is - among other uses as a criticism of Hegel on the state. They therefore welcomed Charles Darwin's Origin of species as a blow to teleology.

Karl Popper argued that the boot was on the other foot: Marx's and Engels' theory of history was a teleology whose 'final end' was general human emancipation. The idea was not original to Popper, since it was an element in Eduard Bernstein's attack on Marx's alleged residual Hegelianism in The preconditions of socialism(1898); but Popper became in the 1950s its main academic standard-bearer. Popper's arguments were hypothetically dependent on his general 'falsificationist' theory of knowledge. In practice they were made plausible by the apparent choice between Stalinism on the one hand, and on the other the post-war social democratic consensus, which appeared to show the 'piecemeal' social reforms Popper defended in The poverty of historicism at work.

But the idea that the traditional sequence of modes of production was a teleology was picked up much more widely than just among strict Popperians. For the 'new left', as we saw in the first part of this review, it was an additional stick to beat reformists and 'official' communists. For Althusser and his supporters it was a stick to beat the Marxist-humanists. In the university history departments, the idea of teleology as a vice of historical writing became merged with the cult of Herbert Butterfield's The Whig interpretation of history (1931) as an elementary methodology text for undergraduate history students, to inoculate them against Marxist and other ideas of 'progress' in history.

In this sub-Popperian sense 'teleology' shifted from its original meaning and became an objection (1) to the claim that it is possible to predict the human future on the basis of the human past and (2) to any theory of history which claims to explain long-range causes for the origins of the present (on the basis that such a theory potentially implies prediction of the human future on the basis of the past).

Popper

The purposive core of Popper's argument is the claim that it is impossible to predict the human future from the human past. Popper claimed that the ideas of Poverty of historicism were the first element of his work, originating at the time of his early 1920s break with his student communism. He also claimed that he had in the 1940s proved the claim by pure logic.

The problem posed is that the physical sciences also predict the future from the past. The logic of scientific discovery (1934) grows out of the project of proving the impossibility of prediction (or, as Popper called it, 'prophecy') by corralling off scientific predictions in two ways. In the first place, science proper makes claims not about the future but about the permanent: that is, matters not subject to time and change. Hence Popper never accepted that evolutionary biology was scientific. In his view study of objects which are themselves subject to change will involve change being indeterminate, and therefore a prohibition on unpredictability.

Second, for Popper science works by hypotheses which are not grounded on prior inductive inferences from regularities in the past. These emerge in other ways, but are tested for potential falsification by experiments. The arguments of this aspect of The logic of scientific discovery were reduced to absurdity by Feyerabend's Against method(1975) and have been criticised from various other directions; they are not now generally believed to offer a plausible account of scientific discovery.

If Popper's rejection of inductive inferences from regularities in the past in science is itself rejected, there ceases to be a serious ground for rejecting attempts to predict the future from the human past more generally. Of course, such predictions are subject to the same problems of complexity and sensitivity to initial conditions which affect attempts to predict climate and weather: short-term predictions can be a lot more detailed and categorical than long-term ones, and so on.

But in reality, all human perception and action involves inductive inferences from past to future; and this is just as much true of human actions as of the physical world. We infer that people living in England will speak English; we take it that there will be a general election by May 2015. Both predictions may turn out to be false, but they aregood enough to act on. The same may be true of much longer-term predictions which imply action in the present (like human-induced climate change, which involves not only predictions about natural processes, but also predictions about human behaviour). This last point - that predictions may be less certain than those of physics, but yet still be good enough to act on - is critical.

The poverty of historicism and the rest of Popper's theoretical construction was a body of highly elaborate abstract argument against revolution and in favour of political gradualism and - to some extent - in favour of political inaction. The short answer to this line of reasoning is Martin Niemöller's famous post-1945 judgment: "First they came for the communists, and I didn't speak out because I wasn't a communist. Then they came for the trade unionists, and I didn't speak out because I wasn't a trade unionist. Then they came for the Jews, and I didn't speak out because I wasn't a Jew. Then they came for me and there was no-one left to speak out for me." In other words, inaction as much as action has moral and practical consequences.

Popper had some difficulty getting published between leaving Austria in 1937 and the outbreak of the cold war, and had particular difficulty with The poverty of historicism. The most plausible explanation is that he was - rightly - seen from his submitted manuscripts as an adherent of the Austrian school of marginalist economics. And the Austrian school's 'do nothing' approach to economic crises was - equally rightly - given a significant moral responsibility for the mass unemployment of 1931-33 and, hence, the victory of the Nazis in Germany. Popper played down his Austrian-school connections in the published version of The poverty of historicism, and emphasised the consistency of his method with social democratic reformism. The real Austrian-school leaders - Mises and Hayek - remained intellectually marginal and disreputable until the generation who lived through the 1930s as adults began to die or retire.

While gradualism and partial reforms may have real advantages if they are actually available, and it is wrong to fetishise the moment of revolution, under some circumstances the overthrow and radical reconstruction of the state is plainly inevitable and necessary. Popper objects to Marx's talk of 'shortening the birth pangs' of the new society as involving 'prophecy'. But it is hard to avoid the conclusion that well-organised, forcible resistance to the Nazis in 1933, leading to a civil war in which the cities carried fire and the sword through the small towns and countryside, would at minimum have 'shortened the birthpangs' of what emerged after the Red Army and Allies carried fire and the sword through Germany in 1944-45 after a devastating global war. Under these circumstances promoting gradualism actually costsmany more lives than it saves.

Causation

If the Popperian objection to predicting the human future from the human past is rejected on epistemological and moral grounds, as it is here, it follows that Popper's and his followers' more general objections to the ideas of long-term causality and directionality in human history can also be rejected, since this objection is merely ancillary to the objections to predicting the future. This does not, of course, establish positively that there is long-term causality and directionality in human history: it merely means that 'teleological' is not a knock-down objection without a lot of further specification.

The question of directionality in history will have to wait until I have addressed that of the foundations of historical materialism. 'Teleology' in the minimal sense of long-term causality leading up to the present is actually indispensable to the historical enterprise. In the first place, all claims to have discovered transhistorical truths in reality depend on historical evidence. To say, for example, as Popper did, that the Austrian school of marginalist economics has the same sort of status as Newtonian physics would imply showing that marginalist laws didfunction in Roman antiquity, the European middle ages, pre-modern China and so on. Of course, when we make this sort of investigation, we discover that the role of subjective marginal utility in the theory actually has the effect of rendering marginalism ... unfalsifiable.

Secondly, it is simply and blindingly obvious that there is some long-term historical causation. To return to an earlier and simple example, most people in England today speak English, not French, a P-Celtic language (Welsh/ Cornish/ Breton), or a Romance language independently derived from late vulgar Latin. Any explanation of the fact is necessarily historical, and historical over the long term. To refuse long-term, causative explanations in history is therefore to refuse any explanation of - taking this single example - language diversity in the present.

Thirdly, indeterminacy objections to long-term causality in history are, in reality, also objections to short-term causality in history. The result, if they are taken seriously, is to reduce history increasingly to its medieval form, the chronicle - a narrative of effectively unexplained events. An egregious example is Anthony Fletcher's The outbreak of the English civil war (1981).

Alternatively, historians may - as in the case of the 'revisionist' school of early modern history of which Fletcher's book was part - argue that the actual outcome was highly unlikely. The deeper this sort of argument is carried, the closer it comes to the 'alternate history' science fiction of Harry Turtledove and similar writers. Or we may have a pinpoint description of a moment in the past which floats free of anything else, so that it might be travel writing for time-travellers - or hobbyists' antiquarianism.

Banaji's essays belong with the variety of historical work which tests and attempts to falsify by empirical evidence an interpretation of the origins of the present. He ends by making some quite limited criticisms of the traditional Marxist scheme. In this context his episodic use of 'teleological' as a 'boo word' is close to meaningless.

In a sense, like Popper, it belongs with the past of bourgeois ideology. Official discourse no longer counterposes reformist gradualism to revolution. 'Reform' in official discourse now means reaction - taking away the working class gains of the 20th century in the hope of returning to a (falsely) imagined 19th. And tipping points, phase transitions and so on are intellectually respectable, and even revolutions that overthrow states are entirely desirable as long as they are 'colour' revolutions to bring in neoliberal regimes. Fitting in with this shift, the study of early modern English history is crawling painfully out of the slough of 'revisionism', marginalists write (commonly wildly speculative) long-term histories of the origin of capitalism, and so on.

In spite of this I have discussed the point at length for two reasons. First, sub-Popperian hostility to 'teleology' is methodologically poisonous in 'left' as well as in conventional academic forms. Second, in Banaji's essays in History as theory the use of 'teleological' as a 'boo word' licenses his failure to construct an alternative general narrative of the origins of the present in which his specific studies could be integrated.

Materialist Foundation

I said in the first part that there are solid non-historical grounds in human biological nature and our material needs for the core of historical materialism - that the ways in which historical societies produce their material subsistence constrain the sort of general social orders possible. And similar grounds support the rejection of methodological individualism (humans are a social species) and of marginalism (there are physical minimum subsistence levels, maximum working hours and maximum quantities of land). These grounds require the analysis of societies and their dynamics in terms of the social division of labour. What follows will travel some distance from Banaji's arguments but will begin to return to them later.

To defend historical materialism it is necessary to get out of the way what it is not. I begin, therefore, by repeating something I have previously written. against the sort of materialism which bases itself on Lenin's Materialism and empirio-criticism (MEC) and therefore argues that ideas are merely 'reflections' of the material world and (in strong forms) for a fully determinist account of human history. This approach is opposed to the Marx who wrote in the first and 11th 'Theses on Feuerbach': "The main defect of all hitherto-existing materialism - that of Feuerbach included - is that the Object [Gegenstand], actuality, sensuousness, are conceived only in the form of the object [Objekt], or of contemplation [Anschauung], but not as human sensuous activity, practice [Praxis], not subjectively. Hence it happened that the active side, in opposition to materialism, was developed by idealism - but only abstractly, since, of course, idealism does not know real, sensuous activity as such." And: "Philosophers have hitherto only interpreted the world in various ways; the point is to change it." Or more snappily in the 18th Brumaire of Louis Bonapartethat "Men make their own history, but they do not make it as they please; they do not make it under self-selected circumstances, but under circumstances existing already, given and transmitted from the past."

'Materialism', in the sense of Marxist materialism, has more than one level. The most basic level is that it is unnecessary to suppose the existence of god or gods, a 'world-spirit', the Hegelian self-moving Idea, spirits, the existence of the soul, the élan vital, or an immaterial homunculus 'consciousness' which sits in the human body and drives it as a motorist drives a car. The phenomena can be adequately explained by the methods of the sciences without any such suppositions. The ideas in my head are electro-chemical phenomena in my brain which are part of an embodied consciousness, which has developed through the physical (Darwinian) and social evolution of the human species. The words I am writing are - as I write them - electrical patterns in the computer; when they are printed they will be patterns of ink on the printed page. They are just as material as trees, etc.

At a second level, within this framework, material forces in the real world vary in power. The power of the ideas in my head, or the words I write, is very limited. Using the methods of the sciences requires us to presuppose the real existence, or more exactly the recalcitrance, of the material world outside our heads. If I had the idea that I could walk on water, it would not prevent me getting wet. If I do not have the idea of a tree in front of me (because I am not looking where I am going) I will walk into the tree and injure myself. It is this fundamental point which Bogdanov and his co-thinkers denied, and which Lenin defended in a muddled way in MEC.

Hence, within the framework of praxis - of 'the active side' and "the point is to change it" - materialism implies that ideas are commonly more powerful to the extent that they are adapted to the external forces in the material world and applied to manipulate these forces. The idea of a stone hand-axe is a means to various human actions to change the world. From this small starting point begins what develops into the massive physical powers of modern technology (the forces of production ...). The idea of a hand-axe and of how to make one - together with the materials to do so - is more powerful than a dream of eating meat or spells cast by a shaman.

This leads in turn to the third level of materialism. This is that social orders and dynamics are in the last analysis governed by technology (the forces of production) and the material division of labour (the relations of production) as means to satisfy very basic human needs (food, shelter, etc). In the last analysis, because, for example, though 13th century England and Japan were both characterised by (in Marxist terms) feudal social orders, these were markedly different from one another, and even now, under globalisation, both Japanese capitalism and Japanese language and culture are profoundly different from their British equivalents.

In a passage which used to be commonly prescribed to students as an antidote to Marxism, John Plamenatz argued in 1963 that recognition of material basic necessities is a truism without practical consequences for social ordering. This argument is manifestly apologetic-ideological.Clear biological necessities for the continuation of the human species are food, protection from predators which may attack humans, and (in cold and temperate climates) clothing, shelter and heat; and the availability of these not only for the adult population, but also for the nurture of sufficient children for population replacement. Recognition of our biological character means that satisfaction of these basic needs have to be seen as preconditions for other activities. Further, both techniques and modes of organisation themselves create derivative necessities, which may become quite elaborate, but are necessary if there is not to be a regression to a less efficient technique and mode of organisation and a consequent reduction in the sustainable population: for a rather basic example, cities need systems for getting clean water and for disposing of human waste products. For the purposes of the present point it is not necessary to go too far beyond the basic biological necessities. In today's world large numbers of people lack access to some or all of the basic necessities, or feel their access to them to be insecure - including in such 'advanced' countries as the USA and Britain.

Division of Labour

Elementary and obvious features of human biology equally imply, as I have said above, outright rejection of methodological individualism, of marginal-utility economics, and of Say's Law as interpreted by the marginalists (that there is an equilibrium price at which every product or service would find a buyer).

In the first place, a species simply cannot exist without the existence of a breeding population: otherwise all that we have is a mutant or 'sport' who cannot pass on his or her peculiar characteristics. Secondly, if the human lifestyle under low-technology and low-population conditions found in the archaeological and anthropological evidence was analogous to that of the European wildcat (asocial except for mating and the early youth of offspring) methodological individualism might still be defensible. In fact it is not: humans characteristically live and breed in social groups (and this is predictable both from chimpanzee lifestyles and from the needs imposed by humans' rather limited teeth and claws).

If marginalism was right, the organisation of social groups could in spite of these observations fall to be explained largely by individuals' choices with a view to their own marginal utility. (Intra-family relations would still remain very problematic, because children are irretrievably 'downstream' of their parents: ie receive benefits for no obligatory return). Anything which didn't fit the paradigm could either be explained as really utility-maximising (as, for example, early 'Chicago school' economist Frank Knight argued that the armed robber's profit properly and legitimately reflected his investment in weapons and the risks he took on). Or it could be identified as a defect in the social order (insufficiently free-market).

The problem is quite simply that there are biological maximum working hours and minimum subsistence costs, given by biological human needs. In addition, the quantity of land is subject to absolute limits (we have only one planet). And the quantity of money is subject to necessary relative limits: if the quantity of money was not limited, it could not serve as a store of value and, given the time element in exchange, if it could not serve as a store of value it could not serve as a means of exchange (witness the Zimbabwe dollar). These limits mean that markets do not invariably or even generally clear. A common form is the so-called 'downward stickiness of wages'; but it is also the case that goods may remain unsold because the producer cannot afford to accept the price on offer, and eventually be dumped in landfill.

The result is that pure private-choice regimes simply do not work in the way the 'hidden hand' suggests. The larger the private-choice element of the economic order, the stronger the tendency towards polarisation (extremes of rich and poor) and in capitalism towards crises (periodic radical dislocations of production); in pre-modern society, towards famine vulnerability and hence periodic famines.

Now this may appear to be merely a point about capitalist markets, but it is not (which is why I expressed it as 'private-choice regimes') though it is most transparent in capitalism. It can apply equally if less intensely to the private choices of pastoralists, peasants or artisans, of slave-takers or feudal lords.

It is from these points that it follows that we have to analyse social orders in terms of the total material social division of labour in the society. Precisely because marginalism and pure private-choice regimes do not work, it is necessary to every society that this social division of labour contain both a familial/gender element (how the reproduction of the species takes place) and a public, state and/or religious element, involving two dimensions.

One of these dimensions is common and public infrastructure (public ways, etc), defence and emergency management, and measurements, which only have any productive utility as collective practices of the social group, including money issuance. Another dimension is that every society which is not merely 'tribal' or 'segmentary' (and some that are) necessarily contains redistributiveinstitutions (commonly religious) which mitigate the inherent tendency of private-choice regimes to produce violent social inequality and episodic general breakdown.

From this angle Banaji is perfectly right to insist that we cannot characterise social orders and their dynamics simply in terms of relations of exploitation at the point of production, as they appear directly as master-slave, landlord-tenant or employer-worker relations. He slightly suggests that we may be able to characterise them in terms of the incentives and aspirations of the elite or ruling classes. This is helpful - for example, in pointing to the land-hunger which landlord classes shared with peasants - but still deficient, because it does not analyse the material division of labour as a whole-society practice.

Base and Superstructure

What I have just said is as 'revisionist' of Marx and Engels as Banaji's argument is, so I should be explicit about it. It is conventional Marxism to divide the economic 'base' from the political, religious, cultural, etc, 'superstructure'. The point I am making is that the 'base' is the total material division of labour in the society, not those forms which are immediately analogous to the capitalist 'economy'. In this context, the family form and - more strikingly - the state and the institutions of redistribution (whether or not religious) are both elements of the material division of labour, hence part of the 'base', not part of the 'superstructure'.

Applying this approach involves real complexities. It is certainly the case that a large part of both law, and other forms of state self-image, and of religious doctrine is genuinely superstructural. To take an instance from law, in the medieval period both Christendom and Islam adopted fixed-share rules affecting inheritance. These rules tended to produce fragmentation of estates, and in both social orders landlord classes found ways to evade them. But the legal doctrines adopted for this purpose are wildly different: the Muslim waqf purports to be a perpetual charity for 'poor' relations; the continental fideicommissum imposes a personal obligation on one 'absolute' owner to hand over to the next in succession; the English entail/settlement divides ownership into time-slices. The problem posed is to disentangle the aspect of law which is expressive of underlying social relations of production, from the aspect of law which is expressive merely of the intellectual creativity of lawyers.

The same problem applies with equal or more force to religious doctrine and institutions (to the extent that legal and religious institutions can be themselves disentangled from one another, which is quite variable at different periods and in different countries).

It is, nonetheless, both possible and necessary that this work should be done. The reason is that neither states nor religious institutions are simple parasites on an imagined private-choice economy. They do real material jobs, and to understand the social division of labour, and hence the 'social relations of production', we need to distinguish the real jobs from the parasitic aspects.

One last point should be made in this context. Whatever their particular forms, private property, the family and class division set up competition between families and reasons to aspire to you or your family climbing rather than sliding down the greasy pole. This set-up is prima facie competitive and therefore a dynamic element in social order. The state, and religious redistribution, are in contrast prima facie conservative institutions. Only prima facie: states may exploit by conquest; temples or monasteries may compete among themselves. But this point provides a limited reason for supposing that we should characterise societies, or historical periods, primarily by their class forms - even if 'their class forms' does not mean counting the frequency of the relations of master and slave, lord and man, or employer and employee. We will return to this point in the third part of this review.