Marxist writing on development and underdevelopment, which barely a decade ago was largely confined to the shrill critiques of a few voices crying in the wilderness, seems well and truly now to have ‘taken off’.footnote^＊ Indeed, the growth of this new (or rediscovered) paradigm has been such that there seems to be almost as much variety of opinion and analysis within it as could be found among the bourgeois development theories that Marxists so trenchantly criticized.footnote¹ A few years ago it seemed appropriate to sketch out the distinctive features of a Marxist perspective as such, in comparison with other approaches.footnote² Today the observer is more likely to be struck by the controversies and debates going on between participants who would probably all claim to be in some sense Marxists, but who appear deeply and perhaps increasingly divided over fundamental issues. We can illustrate this briefly by looking at a range of recent Marxist attempts to grapple with what must be the most basic question for all such writers: namely, the proper characterization of ‘underdevelopment’ itself. A few (following Marx himself) continue radically to reject this problematic as such.footnote³ Warren, for instance, argues that the ‘Third World’ today is at an early stage (or various stages) of industrialization and the development of capitalism, precisely as we know these processes from the experience of ‘developed’ countries.footnote⁴ They are en route, even if the journey will be long and painful.

Against, this, of course, the great bulk of recent Marxist writing takes as its very point of departure the Third World’s not being en route, or at least not on that route. But even here there is considerable variety. Still with affinities (especially methodological) to the ‘classical’ view, there is the conception of underdevelopment as a transition blocked, a (‘normal’) process incomplete. This I take to be the sense of Kay’s dictum that ‘capital created underdevelopment not because it exploited the underdeveloped world, but because it did not exploit it enough’—a fact which he attributes to the unduly prolonged dominance in the Third World of merchant capital, unable as it is to revolutionize the mode of production.footnote⁵ Others go further: the ‘blocked transition’ becomes ‘peripheral capitalism’—a reality sui generis. For Amin, the ‘normal’ development of capitalism (as studied, and formally stated, by Marx in Capital) is but one variant—even if in some sense the ‘true’ one.footnote⁶ This is ‘autocentric’, based on a dynamic relationship between producer goods and consumer goods sectors (Marx’s ‘Departments i & ii’), and fuelled by home market demand. The Third World, by contrast, has from the beginning been extraverted, externally oriented: here the key sectors are export production and import consumption, again dynamically related but perversely so, and with no prospect of debouching into the ‘autocentric’ type.

For Frank, things are simpler: it is just capitalism.footnote⁷ (In fact Frank’s argument is often subtler in detail than his more programmatic statements suggest: even so, the stark simplicity of the latter played a crucial role in the development of a paradigm which enabled them to be transcended.) Capitalism is constituted by a uniform hierarchy of metropolis and satellite, expropriating and appropriating surplus upwards and outwards, nationally and internationally. In particular, as everyone knows, Frank will have none of any suggestion that the penetration of capitalism is in some sense partial or incomplete, so that underdevelopment should be understood as a form of combination of capitalism with something else. Hence his rejection equally of, on the one hand, conventional theories of ‘dualism’ and, on the other, the characteristic idea of Latin American Communist Parties (and others) that there are significant remnants of feudalism in their societies. It is important to understand Frank’s implacability on this point. Aside from the deleterious political consequences of what he regards as false conceptions, his stress on the ‘development of underdevelopment’ as an active process of appendagization and distortion entails absolute hostility to the idea of any currently existing social forms being seen as ‘traditional’ and hence extrinsic to this process. For Frank, as more recently for Wallerstein, there is but a single ‘world-system’;footnote⁸ and it is capitalist through and through.

Going beyond the enumeration of different Marxist characterizations of underdevelopment, it is possible to sketch out a more systematic account of changing emphases and foci within this paradigm. The resurgence of interest in Marxist approaches in the late 1960s was principally mediated via the Latin American ‘dependency’ school, and Frank in particular.footnote⁹ Subsequent evolution and criticism has largely proceeded in terms of reappraising the limitations of the ‘dependency’ perspective. Three lines of criticism may be distinguished. First, there was a problem of scope or scale. ‘Dependency’ might well suggest a macro-framework, but it did not easily manage the shift from general statements to micro-fieldwork.footnote¹⁰ Concepts like Frank’s ‘metropolis’ and ‘satellite’ were in their own way scarcely less elusive or easier to pin down than Rostow’s ‘stages of growth’. An operationalizing problem, then, was early detected; and social scientists attracted by the ‘dependency’ perspective often found that in practice they could use it as little more than a charter. Almost at once other approaches began to be sought, both for the detailed study of the local level and for understanding its linkages with the wider society. Two such candidates have been theories of ‘brokerage’ on the one hand, and ‘modes of production’ on the other.footnote¹¹

In the second place, the operational difficulty was interpreted in some quarters as implying that the dependency approach was not merely too broad in scope, but downright confused and contradictory: it lacked conceptual rigour.footnote¹² This objection, unlike the former, was principally a Marxist response, and it took various forms. At worst, it could be a palaeo-Marxist reaction to unfamiliar and hence threatening ideas, especially those of dubious intellectual parentage (and it may well be true that ‘dependency’ theory is the unhappy progeny of vaguely Marxist ideas coupled with Latin American bourgeois nationalism). Thus there were and are those whose response is to reiterate mechanically as Marxist dogma such themes as ‘stages of development’ (even though, as Amin has observed, this utterly Rostovian idea has nothing to do with the dialectics of Marxfootnote¹³). More originally, if not always much more usefully, others under the banner of Althusser have attempted rigorously to theorize in what they regard as properly Marxist terms the phenomenon loosely described as ‘underdevelopment’. As with the first group of critics of ‘dependency’, in practice this has led to a focusing of attention on modes of production.

It is this concern with modes of production, and what has recently been often called their ‘articulation’, which constitutes the third aspect of the attempt to advance beyond ‘dependency’—and indeed forms the main subject of this article. The essential conceptual differences between the ‘dependency’ and the recently fashionable ‘mode of production’ approaches can be seen in the well-known critique of Frank by Laclau.footnote¹⁴ Although the terminology may not be the same (in particular, Laclau does not use the word ‘articulation’ in the sense that is discussed here), the key distinctions and debates are all there. Against Frank’s ubiquitous and homogeneous ‘capitalism’, Laclau posits not a dualistic model (he too speaks of an ‘indissoluble unity’), but a structured and differentiated whole, the ‘economic system’—others will call it ‘social formation’—which is indeed capitalistic. However, this level of operation is constituted by market relations: for Laclau, what is more important are relations of production, and on this basis he maintains that there were and are substantial elements of feudalism in Latin America. Yet—and here is the twist—these exist not exogenous to capitalism, nor as pockets of decline, but as an intrinsic and structured part of a wider system. In Latin America, as in the ‘second serfdom’ of Eastern Europe, it was precisely the impact of an external market which—so far from dissolving—intensified or even invented feudal and other precapitalist modes of production.

We thus have the paradox of capitalism’s relation to other modes of production being conceived not (or not simply) as succession or evolution (as in the ‘stages’ model: primitive communal, ancient, slave, feudal, capitalist modes of production, with the ‘Asiatic’ awkwardly at a tangent). Nor yet as some kind of dialectical transcendence and dissolution (one could debate whether internal or external, interstitial or marginal, as in the classic Dobb-Sweezy polemic on feudalism and capitalismfootnote¹⁵). Nor even as a transition (unless prolonged to the point of analytical vacuity). On the contrary, this capitalism neither evolves mechanically from what precedes it, nor does it necessarily dissolve it. Indeed, so far from banishing pre-capitalist forms, it not only coexists with them but buttresses them, and even on occasions devilishly conjures them up exnihilo.