Classical expositions of historical materialism have presented the process of transition from one social form to another as endogenously generated.footnote1 But Marxists have not been unaware of the facts of the diffusion of both technical and scientific knowledge, on the one hand, and political institutions and social structures on the other. Indeed, much of the emphasis of twentieth-century Marxism has been on such diffusion, whether in the form of imperialism and colonization, or in that of the ‘structural assimilation’ of various countries to the Soviet model as a result of political-military domination by the ussr or emulation of the model by modernizing elites in the ‘Third World’. Exogenous factors were also in the minds of those Marxists who saw a struggle between rival social systems as assuming equal importance to (or even greater importance than) the struggle between bourgeoisie and proletariat. It is the contention of this article that by taking explicit account of the facts of intersocietal conflict within a revised historical materialism,footnote2 some of the unresolved difficulties of the classical model can be freshly and fruitfully addressed.

In an important article, Yuri I. Semenov has asked what it is that undergoes the sequence of progressive modes of production that Marx writes of in the Preface to A Contribution to the Critique of Political Economy.footnote3 He makes the point that no particular society—such as Roman or French society—has undergone the whole sequence. We can, of course, speak of French society undergoing the feudalism-to-capitalism transition. But there are some parts of the sequence that no society has gone through (the transition from ancient society to feudalism, for example). Rather than abandoning the unilinear story, as many have done, Semenov argues that the sequence applies to human history considered as a whole. He thereby adopts what Ernest Gellner has termed the ‘torch-relay’ theory of history, a view where particular cultures trailblaze for human development as a whole, only to pass on the task to others once the possibilities of a particular mode of development have been exhausted.footnote4 The Hegelian flavour of this is apparent: we have all the paraphernalia of world-historic cultures and non-historic nations.

It is not the purpose of this article to offer a detailed defence of Semenov, but rather to propose a similar ‘torch-relay’ theory of historical development. Historical materialism is an evolutionary theory. But although Marxism has often adopted modes of explanation that share features of Darwin’s theory of biological evolution, this has not been so for the historical process as such. Let us recall what the elements of the Darwinian theory are. First, there is an environment constituted both by the physical environment and by other biological organisms. Second, there are organisms which survive and reproduce in that environment. Many members of a species may be produced, but only those best adapted to the rigours and scarcities of the environment will survive. Genetic mutation gives rise to variation within a given population. An organism may become well-adapted to exploit a particular ‘niche’; it may develop sophisticated defences against predators; and so on. Of course the environment is not a static one, both because the physical environment is subject to change and because other species are changing.

Other evolutionary theories share this structure. For example, it is possible to construct a theory of the firm on quasi-Darwinian lines.footnote5 Firms face an environment constituted both by their competitors and by physical, geopolitical and other circumstances. They have no way of knowing in advance what the optimal structures and decision mechanisms would be, so they adopt rules of thumb (they guess, rely on habit, management folklore, and so forth). Firms that adopt suboptimal procedures fail, and the routines embraced by successful firms spread through the population. But of course the environment faced by firms is constantly changing, and emulation of yesterday’s optimal rules may spell disaster tomorrow. In this arena, combinations of chance and choice combine to impose both a pattern over time and rough adaptation of organisms to the environment they face.

This paper aims to extend this explanatory structure to the evolution of social systems.footnote6 If Marx is right, the most appropriate way to classify societies from a historical perspective is according to their social structure as given by the pattern of effective control over both human and non-human productive assets (labour-power, technical and scientific knowledge, physical plant and machinery, and so forth). At different levels of development of the material productive forces, different social structures are appropriate for their further development. Marxism asserts both an autonomous tendency for human productive power to increase, and the explanatory primacy of the productive forces over the social relations of production, so that when a social structure is no longer an appropriate developmental shell it will be replaced by another, better, one. Social structures are bound together by a political and legal superstructure. As human productive power increases, social forms succeed one another until the advent of a nonclass society in the shape of communism.footnote7