1 G.A. Cohen, Karl Marx's Theory of History: A Defence (Princeton, NJ: Princeton University Press 1978); hereafter cited as KMTH. Earlier versions of this paper were read at the Interamerican Conference on Social Philosophy and the Pacific Division Meeting of the American Philosophical Association. We appreciate especially the critical contributions of Anatole Anton and Zachary Schiffman. We also wish to acknowledge the support of the National Endowment for the Humanities and Lafayette College, whose funds made this article possible.

2 Cohen, KMTH, x. Beyond the initial spate of reviews sparked by Cohen's book, a recent and valuable survey of the debate is provided by Allen E. Buchanan, Marx, Morality, and History: An Assessment of Recent Analytical Work on Marx, Ethics 98 (October, 1987), 104-19. Rapidly becoming a classic of exposition, Allen Wood's Karl Marx (London: Routledge & Kegan Paul1981) also offers a very readable introduction; see in particular 63-81 and 101-10. David Conway unsympathetically reviews the debate in A Farewell to Marx: An Outline and Appraisal of Hit Theories (London: Penguin Books 1987), 52-81, rejecting historical materialism, along with most of Marx's theories, as notwithstanding critical scrutiny, while in Part II of Making Sense of Marx (Cambridge: Cambridge University Press 1985), Jon Elster presents the most extensive philosophical discussion of both the sub stantive and analytical claims of historical materialism since KMTH. Unlike Con way, Elster sympathizes with the work begun by Cohen even while criticizing many of Cohen's specific positions and arguments.

3 Thus, for instance, the consonant remark of Allen Wood: Marx's thesis that the production relations of a society are determined by its productive powers admit: of two related applications: First, it can explain the economic structure of a given society at a given time. Second, it can explain the changes which economic structures undergo in the course of history (Karl Marx, 75).

4 Karl Marx and Frederick Engels, The German Ideology, in Karl Marx and Frederick Engels, Collected Works, Vol. 5 (New York: International Publishers 1976), 32-81, and Karl Marx, Grundrisse: Foundations of the Critique of Political Economy, trans. with forward by Martin Nicolaus (New York: Vintage Books 1973), 471-514

5 Note Elster's comment that for Marx events must necessarily evolve towards the rule first of the bourgeoisie and then of the proletariat. Most recently Elster has characterized historical materialism as a secular theodicy, claiming it very difficult to escape the conclusion that Marx was under the sway of a teleological conception of history. See Further Thoughts on Marxism, Functionalism and Game Theory, in Marx en Perspective, Actes du colloque organis par l'cole des Hautes tudes en Sciences Sociales, Paris, dcembre, 1983, textes reunis par Bernard Chavance (Paris: cole des Hautes tudes en Sciences Sociales 1985), 640, and Marxism and Individualism, in M. Dascal and 0. Gruengard, eds., Knowledge and Politics: Case Studies on the Relationship Between Epistemology and Political Philosophy (forthcoming from Westview Press). See also his discussion of Marx's teleological view of history in Making Sense of Marx, 29-31, 109-18, and 241-317 passim.

6 Marx, Grundrisse, 540-1

7 To be sure, as commentators sometimes note, in Marx's discussions of epochs there exists a tension between these conceptions. However, two points need mentioning here. First, Marx himself drew the distinction between describing chronological processes and explaining them. See, for example, his claim in the Grundrisse (100-1) that the process by which concrete reality comes into existence is not the process by which concrete reality is explained, and his letter to the editorial board of the Otechestvenniye Zapiski in which he states that in Part Eight, Book One of Capital he is describing, not explaining, the historical transition from feudalism to capitalism. Selected Correspondence, I. Lasker, trans.; S.W. Ryazanskaya, ed. (Moscow: Progress Publishers 1969), 293-4. Second, in reconstructing an account of the epochs we are keeping within the two constraints suggested by Cohen at the outset of the debate: what Marx wrote; the standards of clarity and rigor represented by contemporary analytic philosophy. We believe our reading of the epochs, based on an ownership criterion (developed below) accomplishes this task. An alternative, less viable approach is provided by Anthony Giddens, A Contemporary Critique of Historical Materialism: Volume I, Power, Property and the State (Berkeley: University of California Press 1981), 69-89.

8 Eric J. Hobsbawm, Introduction to Karl Marx, Pre-Capitalist Economic Formations, Jack Cohen, trans.; Eric J. Hobsbawm, ed. (New York: International Publishers 1965), 36-7

9 Marx's comments on Russia notwithstanding, the major qualification to this pattern is, arguably, the revolutionary transition from the epoch of capitalism to socialism. On some readings this transition requires that the analytical and chronological characterizations coincide, for such a change is purportedly global and dependent upon the revolutionizing and universalizing features of capitalism itself. See John E. Elliott, Marx's Grundrisse: Vision of Capitalism's Creative Destruction, Journal of Post Keynesian Economics 1 (Winter, 1978-79), 148-69.

10 This is our construction, not Marx's explicit definition. The expression ownership criterion is not to be reduced to legal relationships alone. Rather, it is to be seen as a convenient way of referring to various forms of effective control over the use and situation of some thing. The language of control can be effectively mapped onto legal language, provided we remember, as Marx insists, that legal use of the language of property functions so as to refer to existing relations of production. See the Preface to A Contribution to the Critique of Political Economy, Maurice Dobb, ed. (New York: International Publishers 1970), 19-23, and Cohen, KMTH, 34-5, 63. See also the German Ideology, where Marx identifies the division of labor with property, and its different forms with the stages of development. These stipulative, theory-laden identifications are possible only because of a common referent that joins them: an institutionally established and politically enforced system of ownership, specifying controls over both what producers do and what they produce. Thus Marx approvingly cites the modem economists use of property as the power of disposing of the labour-power of others and later in the Grundrisse refers to property as the appropriation of an alien will, that is of the instruments of production (Marx and Engels, German Ideology, 38, 52; Marx, Grundrisse, 500-1).

11 Again, the epochs are not empirical approximations that merely describe different stages of history; rather they constitute the theoretical frameworks that enable one to carry out empirical investigations. Important here is Maurice Mandelbaum's distinction between sequential and explanatory structures in historical accounts. See The Anatomy of Historical Knowledge (Baltimore: Johns Hopkins University Press 1977), chap. 2.

12 For an attempt to conceptualize ownership for such purposes and the construction of two general models of ownership relevant to this discussion, see George E. Panichas, Prolegomenon to a Political Theory of Ownership, ARSP (Archiv fr Rechts - und Sozialphilosophie) 44 (1978), 333-56.

13 Marx, Grundrisse, 474

14 Ibid., 474-5

15 This extension of exclusionary powers becomes codified in Roman law where tools are referred to as silent instruments, oxen as semi-vocal instruments, slaves as vocal instruments. For Marx, the concomitant advance of exclusionary control and productivity is, accordingly, the historical significance of Greco-Roman slavery (ibid.).

16 Cohen, KMTH, chap. 6

17 Marx, Grundrisse, 483

18 The money form acquired by this exchange became one of the galvanizing agencies in the march from feudalism to capitalism (Marx, Grundrisse, 483-5).

19 Karl Marx, Capital, v. I, Samuel Moore and Edward Aveling, trans.; Frederich Engels, ed. (New York: International Publishers 1967), 217

20 See, for example, Perry Anderson, Lineages of the Absolutist State (London: New Left Books 1974), 403-4.

21 Karl Marx, Capital, v. II, Frederich Engels, ed. (New York: International Publishers 1967), 34

22 Cohen, KMTH, 198. Elster constructs an alternative typology tied to production relations and consisting of three main stages: production for use, production for exchange, and production for surplus-value. To this tripartite schema he adds two intermediary stages, external trade and internal trade, the former mediating the passage from production for use to production for exchange, the latter from exchange to surplus-value. He then claims that Marx sees this five-stage sequence occurring twice in history, once culminating with commercial slavery of the ancient epoch, once culminating in capitalism. Clever though Elster's arrangement might be it overlooks the critical point that for Marx there is in every epoch a surplus product, which is a consequence of both the exercise of ownership controls and the epochal level of productivity permitted by those controls. Thus an epoch identified by production for use, with no surplus product, would be a contradiction in terms. See Making Sense of Marx, 302-17.

23 Cohen, KMTH, 83, 185; Karl Marx, Theories of Surplus Value, Part III, Jack Cohen and S.W. Ryazanskaya, trans. (Moscow: Progress Publishers 1971), 431

24 Marx, Preface, 20-2

25 Cohen, KMTH, 134, and especially chaps. 6 and 9. More recently Cohen has narrowed the range of historical phenomena explained by historical materialism, as he analyses it, but he has continued to defend his general typology of historical epochs, the primacy thesis, and functional explanations (or, more precisely, consequence explanations) as providing an adequate account of both intraepochal and interepochal change. See Restricted and Inclusive Historical Materialism, in Chavance, ed., Marx en Perspective, 53-76, and for his refinements on functional explanations, Functional Explanation, Consequence Explanation, and Marxism, Inquiry 25 (March, 1982), 27-56.

26 For a succinct discussion of the definitions of forces and relations of production in Marx, see William Shaw, Marx's Theory of History (Stanford, CA: Stanford University Press 1978), chap. 1.

27 Mandelbaum, Anatomy, 156

28 Joshua Cohen argues that G.A. Cohen's defense of the primacy thesis fails, and that the model of explanation employed by G.A. Cohen is always either viciously circular or trivial (Joshua Cohen, Review of Karl Marx's Theory of History [The Journal of Philosophy 79 (May, 1982), 253-73]). But see G.A. Cohen and Will Kymlicka's response to this criticism, Human Nature and Social Change in the Marxist Conception of History, The Journal of Philosophy 85 (April, 1988), 171-91, as well as Cohen's Reply to Four Critics, Analyse & Kritik 5 (December, 1983), 195-222, and William Shaw, Historical Materialism and the Development Thesis, Philosophy of the Social Sciences 16 (1986), 197-210. We believe the defenses proffered by Cohen, Kymlicka and Shaw of KMTH's central theses succeed, but do so, for reasons advanced below, only within a specific historical epoch. Hence we agree that there is a successful theory of social or historical development in Marx (on G.A. Cohen's reading), but not, as will be argued in section IV below, a successful theory of interepochal, revolutionary change.

29 Cohen, KMTH, 233

30 Ibid., 278

31 Ibid., 260

32 In instantiating the general schema of a functional explanation, we have substituted the term form for type in reference to ownership relations. This terminological change does not alter the logic of the schema, but does enable us to maintain a consistent terminology, which uses form to refer to a dominant set of ownership relations and type to refer to sub-set of the dominant form. This latter distinction is substantively critical for understanding how functional explanations work to explain intraepochal development.

33 Charles Taylor, Review (of Cohen, KMTH), Canadian Journal of Philosophy 10 (1980), 329

34 A broad range of alternative forms of ownership relations occurring in, or dominating, respective epochs is logically possible. But any form of relation other than R (e.g. S, T, U, etc.) is a R form of relation. So no matter what alternative form of relation is chosen to instantiate the general schema of functional explanation, an epoch characterized by ownership relations other than R is dominated by a R set of ownership relations. If we understand him correctly, Cohen seems to admit this point in his reply to the criticisms of Milton Fisk, for he claims that during periods of transition from one social form to another more than one set of relations would, with cooperation, develop the forces, and it is logically impossible for more than one set to obtain. Further, he suggests, in such a period which relations prevail indeed becomes what Fisk calls a political question, and one whose answer is not settled by the state of the productive forces. Thus it appears in these remarks that Cohen holds (a) the logical incompatibility of different sets of production relations and (b) the inability of an appeal to productive forces to explain the supersession of one set by another. However, earlier in the same article he reiterates the claim here under challenge: the old relations go because they fail to develop the forces, from which we can infer that the new relations supervene, and persist, because and as long as they do develop the forces. And that is a functional explanation .. . (see Cohen, Reply, 204, 211).

35 In his highly influential work, The Nature of Explanation (Oxford: Oxford University Press 1983), Peter Achinstein defends the legitimacy of functional explanations as distinct from causal explanations in a way that corroborates our claim. For Achinstein, functional sentences of the form the function of x is to do y provide a basis for constructing an explanation of the form the reason that x exists is to do y; explaining why x exists. On this account it would be explanatory jibberish to assert the reason that x exists is to do y; explaining why x exists. It is equally nonsensical to assert the reason that dominant ownership relations of form R exist is to influence (advance or fetter) productive forces; explaining why dominant ownership relations of some other form (R) exist. Yet applying functional explanations to interepochal change entails just such an assertion (see 263-90).

36 Maurice Mandelbaum, History, Man and Reason, A Study in Nineteenth Century Thought (Baltimore: Johns Hopkins University Press 1971), 115-18

37 Ibid., 114