Internal Relations and Historical Materialism

Marx insisted that his dialectical method, much of which derived from Hegel, was essential to his work on social reality and history. G.A. Cohen and others have attempted to reconstruct some of Marx's key theories, including his theory of history, in entirely non-dialectical terms. Cohen's historical materialism is, by his own account, a version of technological determinism. In response, some Marxists who reject technological determinism (both as an interpretation of Marx and as a substantive theory) have claimed that abandoning Marx's method necessarily results in distortions of the content of his views, and inevitably results in a picture of social reality and history that is excessively "mechanical," and obscures the "larger totalities" that constitute social reality by, as Sean Sayers puts it, "fragmenting the world into a disconnected series of atomic particulars." According to Sayers, in order to understand Marx, and in order to develop an adequate theory of history, we must employ a logic of internal relations, within which,

"concrete and particular things are always and essentially related, connected to and interacting with other things within a larger totality...this context of relations is internal and essential to the nature of things, not external and accidental."

Now Sayers may be right that analytic philosophers who employ only the logic of "external relations" in reconstructing Marxian theories miss certain important aspects of Marx's own thinking. But the language of dialectics that was used by Marx, and is used by many contemporary Marxists, is often obscure (as in the Sayers quote above), and so it is difficult to state precisely what the dialectical alternative to technological determinism is, and how the logic of internal relations is supposed to support that alternative. Furthermore, it is not entirely clear how to understand what the logic of internal relations is supposed to imply about the Marxian categories used in the construction of historical materialism, and the particulars that fall into those categories.

Sayers claims that while it's fine for physicists and engineers to treat ordinary physical objects as isolated atomic particulars, social scientists cannot usefully do so, since they seek to understand those objects as, for example, productive forces or commodities. And an object is only a productive force or a commodity when it exists in certain contexts, within certain relations of production. As Sayers puts it,

"A machine is a productive force only in the context of certain relations of production in which it is employed productively...these relations are essential - that is to say internal and not merely external - to its being a productive force."

The forces and relations of production, two of the fundamental categories that Marx employs in constructing his theory of history, then, are, according to Sayers, internally related, since objects such as machines are only productive forces in the context of certain relations of production. But there's an important ambiguity here: it's not clear whether machines are supposed to count as productive forces only when they are actually employed productively (as seems to be the suggestion of the first part of the above quote) within the prevailing relations of production, or whether they count as productive forces so long as they could be so employed within the prevailing relations. The former interpretation is problematic because if unused machines do not count as productive forces, then the Marxian claim that epochal transformations occur in part due to contradictions between the existing forces and relations, including the failure to optimally employ the existing productive forces, is undermined. The latter, on the other hand, is problematic because it is not clear on what basis we could ever conclude that an object that has the physical characteristics necessary to aid in productive activity couldn't possibly, given the existing relations of production, actually do so. So on this view it seems that knowing which relations prevail will be unnecessary in order to conclude that an object is a productive force, and the thesis that forces and relations are internally related turns out to be false.

The dialectical response to my critique of the latter option is to claim that I'm employing an analytic, rather than a dialectical sense of 'could'. Of course we can imagine industrial machinery used in productive activity within feudal relations of production, but this is not a real historical possibility, and dialectical logic, including the thesis of internal relations, is designed precisely to demarcate real historical, rather than abstract logical possibilities.

But it seems to me that all that this move accomplishes is to present what are actually empirical claims about conditional possibilities (i.e. what is nomologically possible given some set of facts) in quasi-logical terms. And it's not clear what this adds to arguments on behalf of historical materialism. If Marx's theory of history is true, it's contingently true. That theory claims that the prevailing production relations in any historical period correspond to the level of development of the productive forces. The thesis of internal relations includes the claim that particular objects are only productive forces within the context of certain relations. But if we accept this than much of the "correspondence" between forces and relations will be ensured by the mere stipulation of when something counts as a productive force.

This is unacceptable. Historical theories, and especially those such as Marx's that purport to be scientific, must, like all scientific theories, be subject to empirical testing. The thesis of internal relations between forces and relations of production makes too much of Marx's theory true by definition alone.