The Violence of Abstraction The Analytic Foundations of Historical Materialism

Marx's later writings extend the argument to the actual forms of law themselves, and anchor these more directly in the phenomenal forms of commodity circulation. Pashukanis (1978; cf. Corrigan and Sayer, 1981) has developed this theme in Marx with considerable brilliance (if also, at times, undue abstractness). ‘Capital-logic' theorists (who can sometimes be subjected to the same criticism) have argued similarly regarding the characteristic political forms of bourgeois democracy (Holloway and Picciotto, 1978). So I shall be brief. For the Grundrisse, the juridical subject - the free, autonomous, sovereign individual - is implicit in the relation of commodity exchangers: in the very act of exchange they are stipulated for each other as, and prove themselves to be, equals, [and] there enters, in addition to the quality of equality, that of freedom':

although individual A feels a need for the commodity of individual B, he does not appropriate it by force, nor vice versa, but they recognise one another reciprocally as proprietors, as persons whose will penetrates their commodities. Accordingly, the juridical moment of the Person enters here, as well as that of freedom, in so far as it is contained in the former. No one seized hold of another's property by force. Each divests himself of his property voluntarily. (1858a: 243; cf. 1867a: 84)

Marx goes on to argue that the exchange of exchange-values is the productive, real basis of all equality and freedom', and ‘as pure ideas they are merely the idealized expressions of this basis; as developed in juridical, political, social relations they are merely this basis to a higher power'(1858a: 245). Equality and freedom presuppose relations of production lacking in the medieval and ancient world, which rested on ‘direct forced labour' (ibid.; cf. 464-5). Capital argues similarly that the notion of human equality' can only ‘acquire the fixity of a popular prejudice' when ‘the dominant relation between man and man is that of owners of commodities' (1867a: 60). In this sense the market is ‘a very Eden of the innate rights of man', where ‘alone rule Freedom, Equality, Property and Bentham' (ibid.: 176).

The attributes of the juridical person', then, are precisely [those] of the individual engaged in exchange' (1858a: 246). However, Marx argues – and here we return to the ideality’ of superstructures – that the sphere of exchange is merely the surface process’ of bourgeois society, 'beneath which, however, in the depths, entirely different processes go on, in which this apparent individual equality and liberty disappear' (ibid.). He of course has in mind the eminently unequal ‘exchange' of labour-power for wages, out of which arises an unrequited surplus-value. Polemicizing against French socialists, who see this as a perversion’ of the original equality of the exchange relationship, Marx argues that capital is inherent in’ exchange-value, and wage-labour in commodity-producing labour. The possibility of the perversion' is in other words given in the original relationship itself. Failing to see this is a utopian inability to grasp the necessary difference between the real and the ideal form of bourgeois society', the latter being only the inverted projection [Lichtbild] of this reality” (ibid.: 249). Such are superstructures'.

VII

Let me now try to pull together the strands of this argument. Nothing I have said, or quoted from Marx, is in any way intended to deny the reality of organs, institutions or agencies of government or law as such. That would be absurd. It is the reality of their apparent separation from capitalism's economic structure' - its essential relations – which is at issue.

Marx's analysis of legal and political ‘superstructures', I believe, hinges upon two key points, both of which are very subversive of 'traditional historical materialism'. First, he seeks to establish that the regulative agencies of state and law are substantially internal to capitalism's 'base' itself. That is to say, they are but forms of the social relations within and between classes – division of labour and labour/capital relation - upon which the possibility of capitalist economy is predicated. Their appearances of independence notwithstanding, state and law are in reality 'the form in which the individuals of a ruling class assert their common interests, and in which the whole civil society of an epoch is epitomised' (1846a: 79–80). Marx insists, moreover, that state and law are necessary forms of bourgeois rule: the individuals who rule in these conditions - leaving aside the fact that their power must assume the form of the state – have to give their will - a universal expression as the will of the state, as law' (ibid.: 329, emphasis added). State and law are thus essential relations of capitalist production, in the sense elaborated in my last chapter: it could 'not proceed in their absence. We should pay close attention to Marx's language when he says that civil society must assert itself in its foreign relations as nationality, and inwardly must organise itself as State', and speaks of civil society in its action as State’ (ibid.: 50, 48). This does not sound like an external relation.

108

IDEAL SUPERSTRUCTURES

IDEAL SUPERSTRUCTURES

109

This argument implies, among other things, that the class relations of capitalism cannot adequately be delineated just at the economic' level, as Cohen's supposedly rechtsfrei ownership relations. State and law are intrinsic to the totality of social relations which actually make up the seemingly impersonal' power of bourgeois property, which maintain a social order in which domination can appear to take purely ‘economic' forms. They are constitutive of bourgeois class rule; it could not exist otherwise. The bourgeoisie, in Max Weber's useful phrase, is the national citizen class' (1966: 249); Marx himself observes of the notion of citizenship ‘it should read: domination of the bourgeoisie' (1846a: 215). Anderson, too, must be qualified here. I quoted him above on the novelty of ‘the economy' being a formally self-contained order' in capitalism, for which ‘superstructures' of state and law are ""external" preconditions' (p. 73). This is fine, and grasps an important contrast between capitalism and previous forms of society, providing we remember that although the novelty is real enough, the externality of these ‘spheres' is merely apparent. This brings me to Marx's second point.

For he is equally clear that in capitalism, these essential relations manifest themselves in Erscheinungsformen which conceal their true substance. The appearance of the economic sphere’ as an external, objective realm is one side of that world of illusion, whose fetishes we have sufficiently explored in previous chapters. As E.P. Thompson has observed, the very category of economics – the notion that it is possible to isolate economic from non-economic social relations, that all human obligations can be dissolved except the cash-nexus – was the product of a particular phase of capitalist evolution' (1965: 82–3). The appearance of the political state, the illusory community', peopled by abstract individuals, equal citizens of the polity and subjects of the law, is the other side of the same phenomenon. The link between the two is the real subject of the commodity exchange – the bourgeois. When Marx speaks of state and law as superstructures’, it is to this bewitched world of appearances that I believe he is referring. Far from inscribing in his own theory, via a metaphor of levels', the reifications of capitalist society, the base/superstructure metaphor is intended to show them to be just that. The real state is an essential dimension of capitalism's class relations, the combination of one class over against another' (1846a: 93). Its ideal appearance is of an illusory community [which] took on an independent existence' (ibid.). It is the latter Marx calls a superstructure', in order precisely to show up its illusoriness, its ideal' character, and deny its substantial independence.

Regrettably Marx never did write that 'Critique of Politics and Political Economy'. But his later writings are thoroughly consonant with the critique he developed in the 1840s, arguing both the internality of the real state to capitalism's class relations and the insubstantiality of its ideal'independence. Capital for instance asserts that the use of state power is an essential element of the so-called primitive accumulation'(1867a: 737), whose methods ‘all employ the power of the state, the concentrated and organised force of society' (ibid.: 751). For Grundrisse, likewise, speaking of England, governments, e.g. of Henry VII, VIII, etc. appear as conditions of the historic dissolution process fof feudal relations] and as makers for the conditions of existence of capital' (1858a: 507). Part 8 of Capital I details some of what Marx has in mind.13 Both The Civil War in France, and the Critique of the Gotha Programme, texts of the 1870s, continue to anchor the political state’ - the state in so far as it forms a special organism separated from society through the division of labour' (1875: 356), a State separate of and independent from society' (1871: 151) - in the specifically capitalist division of labour.

Marx also continues emphatically to deny the substantial reality of this separability. For The Civil War, the Second Empire whilst ‘at first view, apparently the usurpatory dictatorship of the governmental body over society itself, rising alike above and humbling all classes', is in fact ‘the only possible stateform in which the appropriating class can continue to sway it over the producing class' (ibid.: 196), ‘the last degraded and the only possible form of that class ruling' (ibid.: 198).14 The very separation of ‘civil society' and political state’ is an ideological superstructure on capitalism's component relations, an ideological form in which they are masked, which simultaneously idealizes individual and community. Two passages, one from 1843, the other from 1871, bring out well the underlying continuities – and the animating emancipatory vision – in Marx's thinking here:

110

IDEAL SUPERSTRUCTURES

IDEAL SUPERSTRUCTURES

111

Only when man has recognized and organized his forces propres’ as social forces, and consequently no longer separates social power from himself in the shape of political power, only then will human emancipation have been accomplished.(1843c:168)

[The Paris Commune] was .. . a revolution not against this or that, legitimate, constitutional, republican or Imperialist form of state power. It was a revolution against the State itself, of this supernaturalist abortion of society, a resumption by the people for the people of its own social life. (1871: 150)

Cohen at one point in his book remarks that ‘even the mature idea of the political superstructure has some analogy with the doctrine of commodity fetishism' (1978: 128). Too true. But why, then, perpetuate the fetish within the analytic categories of historical materialism itself?

Interlude: the giraffes among the acacias

What follows is a brief, but I hope a useful, digression from my main argument. It is entirely concerned with a series of issues arising out of G.A. Cohen's formulation of 'traditional historical materialism'.l This chapter, which is the most philosophically technical in this book, may be skipped by those persuaded of my arguments thus far and disinclined to pursue Cohen further, since the aspect of his work I want to take up here – his functionalism – rests on premisses I have already rejected. If my view of Marx's basic concepts is accepted, there is evidently no possibility of developing a general functional model (any more than there is of elaborating a general causal model) of the forces/relations or base/superstructure relationship, as Cohen attempts. The issue of functional explanation in Marxism is worth discussing independently none the less, for two main reasons.

First, Cohen is undoubtedly correct that many forms of Marxism are tacitly functionalist. These include certain internal relations perspectives, and I wish clearly to distinguish my own position from these.² Functionalist formulations are also endemic in some variants of Marxist-feminism. Second, and more importantly, Cohen maintains that Marx's claims in the 1859 Preface can only be rendered coherent if they are interpreted functionally. Only thus, he contends, can we reconcile Marx's recognition that specific superstructures are conditions of existence for economic structures, and economic structures for development of the productive forces, with his assertion of the respective explanatory primacy of the latter term in each pair. Marxism in Cohen's view entails an inescapable commitment to functionalism.

I believe Cohen to be right on this issue if - but only if - his definitions of the entities in question are accepted. Non-functionalist causal interpretations of the forces/relations and base/superstructure ‘correspondences' run into unavoidable problems of circularity, of the sort raised by Acton. To this extent Cohen renders Marxist debate a significant service in showing that an orthodox conception of Marx's basic concepts does entail a functional reading of his explanatory propositions - a conclusion many Marxists have been reluctant to accept, even whilst wishing to maintain their conceptual orthodoxy. Interestingly, it was a recognition that only a covert teleology could reconcile a conception of superstructures as conditions of existence of economic structures with the assertion of economic determination ‘in the last instance’ which led Hindess and Hirst³ eventually to argue the ‘necessary non-correspondence' of these supposed levels of the social formation.

T

112

IDEAL SUPERSTRUCTURES

5

For me, on the other hand, Cohen's demonstration of the unavoidability of functionalism within traditional historical materialism provides an additional, and compelling, reason for doing what I have argued for throughout this book, and rethinking these basic concepts themselves. For I shall argue that Cohen does not succeed in defending functional explanation. The alternatives are therefore to follow Hindess and Hirst out of the historical materialist framework entirely, in the direction of a systematic pluralism, or to seek to reconstitute that framework on foundations very different from those of 'tradition'. If we are to reject a functionalist version of Marx, in short, we must also abandon the familiar interpretation of his fundamental concepts.

1

Functional explanation, for Cohen, is ‘a distinctive explanatory procedure, in which reference to the effects of a phenomenon contributes to explaining it' (1978: 250). It is a causal explanation, but a ‘special type of causal explanation . . . deriving its peculiarity from generalizations of distinctive logical form' (ibid.). Cohen calls the logically distinctive generalizations on which functional explanations are based consequence laws'. He sets out the form of these laws as follows:

IF it is the case that if an event of type E were to occur at t1, then it would bring about an event of type F at t2, THEN an event of type E occurs at t3. (ibid.: 260)

In practice consequence-explanations are invariably also functional explanations - that is, the explanandum-event E is functional for something or other. Importantly, however, Cohen insists that this functionality is not a necessary feature of the structure of functional explanations, which is simply that of consequence-explanations in general. It is at least logically possible to imagine a consequenceexplanation in which the explanandum-event would be dysfunctional. We may therefore abstract from the functionality of the explanandum in characterizing functional explanations, and describe them sufficiently by reference to their logical form – that of the consequenceexplanation – alone. I assume Cohen insists on this in order to avoid standard objections to teleology.

Though consequence-explanations are causal, they are not, for Cohen, mirror-images of ordinary causal explanations' (ibid.: 261). The logical form of the consequence-laws on which consequenceexplanations rest differs from that of ordinary causal laws. He asserts, none the less, that ‘a consequence-statement explains when it relates to a consequence-law in whatever way an explanatory precedencestatement relates to a pertinent law' (ibid.: 259). This means that mutatis mutandis, a functional explanation can be cast in the familiar deductive-nomological (hereafter D-N) form, as set out classically by Hempel. In Cohen's words:

An elementary consequence-explanation, D-N construed, would contain as first premiss a consequence law and as second a statement asserting an instantiation of its major antecedent. The schema, where what is to be explained is the occurrence of an event of a certain type, would be as follows:

L IF it is the case that if an event of type E were to occur at tl then it would bring about an event of type F at t2, THEN an event of type E occurs at t3.

C Were an event of type E to occur at t', it would bring about an event of type F at t"

E An event of type E occurs at t′′ (ibid.: 272)

In an article of 1983 Cohen offers a less technical account of functional explanation than in his book, which I have followed so far. If e is a cause and f its effect, the explanation is neither that e occurred because f occurred, nor that e occurred because it caused f. Either of these has the unacceptable consequence of requiring an effect temporally to precede its cause. The form of functional explanation

114

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

115

Cohen favours is: ‘e occurred because it would cause f, or, less tersely but more properly: e occurred because the situation was such that an event like e would cause an event like f (1983b: 115). In functional explanation, it is the fact that were an event of a certain type to occur, it would have a certain effect, which explains the occurrence of an event of the stated type' (1978: 261). Cohen calls this fact – that described by the words he italicizes - a ‘dispositional fact'. Thus in Cohen's Marxism the emergence of production relations of a given type is functionally explained by the dispositional fact that were such relations to emerge, they would promote development of the productive forces. Similarly with superstructures vis-à-vis bases.

Cohen emphasizes, additionally - and very importantly for my argument - that such explanations may be confirmed in the absence of a theory as to how the dispositional property figures in the explanation of what it explains ... we may have good reason for thinking a functional explanation is true even when we are at a loss to conjecture by what means or mechanism the functional fact achieves its explanatory role' (ibid.: 266). As his later article puts it: ‘to say that A explains B is not necessarily to indicate how A explains B . . . In a widely favoured idiom, [we] may not know the mechanism linking cause and effect, or, as I prefer to say, [we] may be unable to elaborate the explanation'(1983b: 118-19).

Thus specification of causal mechanisms is not for Cohen necessary to explanation. To use his own terms, explanations answer ‘whyquestions' and we may be able to say why an event occurs without knowing how it occurs. In a functional explanation, the answer to the question why an event occurs is necessarily and sufficiently furnished by citation of the relevant consequence-law and antecedent (dispositional) conditions, exactly as within the conventional D-N schema to say why an event occurs is simply to invoke a covering causal law whose antecedent conditions are satisfied. Identification of causal mechanisms - saying how the explanandum-event is actually brought about – pertains to what Cohen calls the elaboration' of explanations. Such elaboration – the provision of a theory of causal mechanisms – may be desirable. But it is not in any way mandatory. Functional explanations as such are sufficiently characterised by their formulability in terms of Cohen's modified D-N schema alone, and hence remain valid as explanations even in the complete absence of theoretical elaboration. This is, of course, the standard position adopted by positivists (and denied by realists) with regard to ordinary causal explanations.

II

Discussion may best be advanced from here by considering one of Cohen's examples of functional explanation in more detail. I shall treat it as a paradigm. It is this:

A population of giraffes with a mean neck length of six feet lives in an environment of acacia trees, on whose leaves they feed. The height of the trees makes it true that, if they now had longer necks, their survival prospects would be better. They subsequently come to have longer necks. So far all we have is evidence of a consequence generalization. But if Darwin's theory of evolution is true, then the fact that were they to have had longer necks, they would have fared better, contributes to explaining the elongation. The environment selects in favour of variants with longer necks precisely because it is an environment in which longer necks improve life chances. On no construal can that dispositional fact be reduced to an unexplanatory precursor of the acquisition of the feature. Its explanatory relevance to the elongation of the neck is entailed by the Darwinian theory. (ibid.: 269)

Cohen then confronts an imaginary sceptic, who argues that the dispositional property (that the environment is of a kind in which longer necks would improve life chances) merely correlates with what really explains the elongation. This is the presence of the trees, a quite undispositional circumstance' (ibid.: 270). Cohen replies as follows:

We can agree that it is the trees which, aided by the chance genetic variation, cause the neck of the species to grow. But what is it about the trees which makes them have that effect? Answer: where trees are of the given height giraffes with longer necks would prosper better. The dispositional fact is an essential element in the explanatory story. (ibid.)

Consistently with his disjunction between explanation and its elaboration', Cohen claims, additionally, that in this example the dispositional fact is essential to the explanation of the lengthening of giraffes' necks irrespective of whether we favour Darwinian or Lamarckian accounts of how the elongation occurs - the latter,

116

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

117

Cohen might have added, being a functionalist theory in the conventional teleological sense, and the former not. In Cohen's view, what we have in the Darwin/Lamarck dispute are rival theories of why a consequence-explanation holds, not rival alternatives to consequenceexplanation' (ibid.: 271). We see here the import of Cohen's denial that the functionality of explananda is necessary to structural characterization of functional explanation, and the way this connects with his explanation/elaboration distinction. In so far as the functionality of an explanandum enters into a putative causal account, as it does for Lamarck, it does so at the elaborative level. It is not necessary to the formulation of the consequence-laws upon which functional explanation as such rests.

Now, I believe that in this example Cohen mis-states Darwin, and when Darwin's theory is properly expressed, dispositional properties are not essential to explanation. There is a construal within which such dispositional facts’ are non-explanatory precursors of the elongation at issue: and that construal is standard Darwinian theory, at least as I understand it. Let me elaborate. In confronting his imaginary opponent, Cohen attributes to him the view that it is the presence of the trees which explains the elongation of the giraffes' necks, and this is not a dispositional circumstance. Cohen easily rebuts this. But a Darwinian would not, in fact, be likely to advance this argument. He would be more likely to argue in two stages, which Cohen conflates.

First, he would say that the initial appearance of long-necked giraffes is not explained by anything in the environment, dispositional' or otherwise, but by genetic mutation. Such mutations are random, and (pace Lamarckian theories) altogether unconnected with environmental circumstance. It is this mutation which is the cause of longnecked variants initially emerging. And the presence or absence of acacias – importantly - makes it neither more nor less likely that this original adaptive mutation will occur, precisely because mutation is random. Indeed it is perfectly conceivable that the mutation of the neck might not have occurred. Some species, after all, do actually die out. Consider dinosaurs. We cannot then speak here of the operation of a consequence-law (if e would cause f, then e). The presence of acacias cannot therefore be said to dispose'the genes to mutate in an appropriately adaptive way. To suppose otherwise is to fall into the fallacy of reasoning post hoc ergo propter hoc.

Second, our Darwinian would say that the survival of the longnecked variant - a different issue - is undoubtedly then explained, ceteris paribus, by the presence of the trees whose leaves they are better equipped to eat than their short-necked competitors. But dispositional propositions are in no way ‘an essential part of the explanatory story’ at this level of the explanation either. It is rather the case that given the fact of the mutation, the presence of the trees then does sufficiently explain the enhanced survival chances of the long-necked giraffes. That the mutated giraffes can eat acacia leaves, and thus have a competitive advantage in the struggle for survival, in other words remains a totally contingent consequence of their having fortuitously acquired long necks for reasons that are entirely unconnected with the environment or its supposedly ‘dispositional' properties. This can be expressed in standard D-N form, without any recourse to consequence-laws':

LI Individuals well adapted to their environment will have competitive advantages over those less well adapted.

L2 Such advantages, over time, ensure the survival of the fittest’ and thus the generalization through a species of the relevant adaptive features.

CI Some giraffes have long necks as a result of fortuitous genetic mutation.

C2Giraffes live among tall acacia trees.

E Long-necked giraffes have a competitive advantage over short-necked in their natural habitat, and the long-necked variant will therefore prevail.

Certainly the character of the environment is an important part of the explanation of the generalization of long necks among the giraffe population, but not in the way Cohen supposes. Giraffes have long necks not because of the dispositional properties of the environment, but because of the complex interaction through time of genetic variation and environmental circumstance - these two factors, on Darwin's theory (though not Lamarck's) being contingently connected.

An adequate causal account of the acquisition of long necks by the giraffe species would thus have to combine a series of propositions concerning the population dynamics of giraffes, genetic variation in the giraffe population, and features of the environment, including the height of acacias. But none of these propositions would have to be dispositional' for the explanatory story' to make sense, as Cohen alleges. It is not the dispositional fact' that were giraffes to have long necks, they would fare better among the acacias, which in any sense explains why the feature was acquired by some giraffes in the first place, and the explanation of why this variant then prevails supposes this initial acquisition of the feature. Given the mutation, moreover, presence of trees then does sufficiently explain the survival of longnecked giraffes ceteris paribus.

118

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

119

On the contrary, what is being offered in Darwinian theory is an historical and relational – a material - explanation of what functionalism apparently explains, namely the evident adaptation of giraffes to life among the acacias. It was of course exactly for this reason that Marx hailed the publication of Darwin's Origin of Species for dealing 'the death-blow to teleology in the natural sciences' whilst setting forth its rational meaning in an empirical way' (1861). Darwin, Marx is saying, provided a material explanation of (and thereby makes rational) what superficially appears to be a teleological process. Cohen can construe the dispositional fact' that environments with tall trees favour long-necked mammals as in itself explanatory only by shortcircuiting this complex, material account of how long necks were both initially acquired and subsequently generalized through the species.

III

This line of criticism of functional explanation is hardly new, but Cohen does not confront it. What I am saying, in effect, is that explanations in terms of Cohen's ‘dispositional properties’ can be entirely rewritten in straightforward causal terms without any loss in content, but with a considerable gain in accuracy. Cohen's ‘dispositional properties’ are contingent factors (like presence of acacias) which mav feature as relevant antecedents in a perfectly conventional causal explanation (as of the long-necked giraffes’ survival) if certain other equally contingent events (like mutation) occur, but they do not thereby explain why those other events occur in the first place. They merely appear to do so, as long as we remain in ignorance of the real causal mechanisms at work.

Somewhat paradoxically, the possibility of thus rewriting functional explanations as orthodoxly causal explanations is a consequence of Cohen's own insistence on eliminating the functionality of explananda from his characterisation of ‘consequence-explanation'. I could not have argued this case for all varieties of functional explanation. For Cohen, dispositional facts explain in exactly the same way that causal antecedents do. But this is not the case, for instance, in avowedly teleological explanations in which ends are explanatory without being temporal antecedents. Such explanations self-evidently cannot be assimilated to orthodox (Humean) causal paradigms in the way Cohen attempts for his consequence-explanations'. Were the functionality of the explanandum-event a necessary feature of functional explanation, the latter indeed would be a distinctive form of explanation - if an obviously dubious one, except when applied to entities capable of acting to achieve goals, like people or gods – and the translation of functional into causal propositions would be impossible without loss of the specifically teleological element in their content. For Cohen, though, causes lie in antecedent dispositional facts, not the functionality of explananda as such, and it is these supposed dispositional facts I argue can be causally rewritten.

Now, it will be recalled that Cohen carefully distinguishes explanation and elaboration. The possibility of translating functional into causal propositions, however, depends precisely upon knowledge of what Cohen sees as extraneous to explanation strictu sensu, namely the causative mechanisms he relegates to the category of ‘elaboration’ of explanations. In his Darwinian example, it is only knowledge of random genetic variation - which actually causes long-necked giraffes to emerge - which allows us to rewrite what appears to be an explanation in terms of the dispositional properties of an environment, as an explanation in terms of material causes operating within that environment. In the absence of knowledge of the real causal mechanism at work, it is indeed only the ‘dispositional fact' that acacias favour species with long necks - precisely Darwin's explanandum – which appears capable of explaining the presence of giraffes.

It is worth spelling it out that, contrary to Cohen, this knowledge of causative mechanisms is not something through which we merely 'elaborate' explanation, supplying the how' to buttress the ‘why' we already know. It tells us rather than the functional answer to the question why - despite its initial plausibility - is wrong. Giraffes do not, except perhaps in a shorthand pardonable in biologists but inexcusable in philosophers, have long necks because their environment is such that if they did they would thrive in it. They have long necks by the luck of the genetic draw, operating within given environmental conditions, certainly, but not in any way which is functionally explained by those conditions. Indeed, knowledge of mechanisms shows that there are in fact no consequence-laws involved; as argued above, it in no way follows from the presence of acacias that the adaptive mutation actually will occur. If Darwinian theory is correct, we cannot say, for this area, if e would cause f, e will occur. We can reach this last conclusion however - in defiance of manifest correlations of species and environments - only because we know the adaptive mechanism in fact to be random. The causal ‘elaboration', in short, actually vitiates the supposed functional ‘explanation'; it shows the ‘dispositional' features of the environment not to be the explanatory factor, appearances notwithstanding.

120

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

121

By the same token, we should not suppose production relations and superstructures take the forms they do because they respectively serve’ productive forces and economic structures - even if they manifestly do - without a far more convincing defence of functional explanation than Cohen supplies. It may on the contrary be the case that, sav, a ‘superstructural' feature arising out of eminently noneconomic causes is quite contingently able to facilitate certain economic' developments, as the mutation does the survival of giraffes on the plains of Africa. In such cases the feature should properly belong, in the first instance, in the causal explanans of why a given economic structure is able to emerge, and not, by functional sleight of hand, be numbered among its explananda. Once the mode of production is on its feet, the environment it constitutes for the superstructural feature at issue might in turn help explain the latter's persistence, as the continued presence of acacias does the survival of long-necked giraffes. But if it did, it would do so materially, not functionally.

Examples of this might include, in the case of capitalism, Weber's Protestant ethic or the English common law. Few sociologists would deny the pertinence of either for capitalist development, but only a historical ignoramus would ascribe them capitalist origins. Many would accept that whatever their origins, the survival of both was considerably facilitated by the success of the capitalist economic structure they (arguably) helped give rise to. Christopher Hill argues this brilliantly for both Protestantism and the common law as summated in Coke's Institutes.4 The seductiveness - a fatal one - of Cohen's functionalist model is that (like much bad Marxism) it apparently allows us to have our cake and eat it. Technological and economic determinisms can be maintained in a dispositional’ form in which straightforward empirical challenges cannot be used to refute them. I would argue, however, that to hold, as Cohen must, that such superstructural’ factors as Protestantism and law are ‘selected for' by bases (or bases by productive forces) at best obscures explanation of their real material causes, while to excavate the latter – schisms in the Church, the strength of the Norman and Angevin monarchies, and suchlike eminently undispositional factors - is likely to reveal the alleged ‘selection' to be as contingent, and dispositional' explanation as spurious, as in Cohen's mistreatment of Darwin. To sever explanation' and ‘elaboration', in short, is to evade explanation, in any material sense, entirely. In the case of Marxism it is to substitute theory for history. Functional explanations can masquerade as explanatory only so long as material causes remain unexplicated.

This brings me to the major point I want to make here. What is most striking in Cohen's formulation of functional explanation, and has not generally been raised in debates around his book, is that he sees it as a variant of causal explanation – albeit of a distinctive kind – as the latter is understood within the mainstream positivist tradition. His use of Hempel's D-N schema is indicative. Within this tradition - which derives ultimately from Hume - a cause is understood in terms simply of constant conjunction (if p, then q) of terms between which no necessary connection obtains. To give a causal explanation of something is accordingly merely to show that it instantiates a covering law, the latter being a universal conditional relating empirical particulars. Cohen formulates the relation between his consequencelaws and consequence-statements in exactly the same way. His expunging of considerations of mechanism from explanation proper, and attempt to characterize the latter in terms of its logical form alone, is likewise wholly of a piece with the positivist project.

But Cohen ignores here the entire (and currently very influential) realist tradition in the philosophy of science, which has many adherents within or sympathetic to Marxism. Indeed there is now a large literature which seeks to annex Marx to the realist cause.5 Now for realism, expressibility in D-N form is certainly not sufficient, and for some realists like Bhaskar is not even always necessary, to qualify a series of propositions as an explanation. Another way of putting this is to say that what is entailed in giving an adequate explanation cannot be sufficiently described by giving an account of its logical form alone. The covering law model, accordingly, does not adequately characterize what scientists (or ordinary language) actually mean when they speak of a cause.

122

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

123

Keat and Urry (1975: 11-12) use a famous illustrative example. The onset of measles is invariably preceded by the presence of small white marks in the lining of the cheeks, known as koplik spots. The latter have both of the necessary characteristics of a Humean cause (temporal precedence, and invariant succession), and could therefore serve as a major premise to D-N explanation of the form ‘if koplik spots, then measles'. And if D-N formulability suffices to characterize what is required for an explanation, then we have no good reason not to explain measles as the outcome of koplik spots. Yet we would not ordinarily want to say that the spots cause the measles. The cause, rather, is invasion of bodily cells by a virus of the family paramixo viridae, of which koplik spots are an early manifestation. Certainly we may predict the imminent onset of measles from the presence of koplik spots, but the spots do not in any meaningful sense explain the disease.

Ascription of causes - explanation - for realist critics of positivism, necessarily entails giving an account of the actual mechanisms through which a given effect is seen as being really produced, in this case the activities of the virus. It is only in virtue of such an account that we can differentiate real causes from merely incidental correlations, or stop a potentially infinite causal regress of empirical antecedents. This in turn requires what Bhaskar calls 'ontological depth' - the invocation of real, and sometimes invisible entities distinct from the phenomena observed, whose characteristics, properties and powers are capable of explaining those phenomena in this materially causal sense. Such are viruses with measles, and genes with giraffes’ necks. This conception of causal explanation clearly deviates considerably from Hempel's authoritative characterization of the positivist model, which Cohen straightforwardly accepts, for it centrally involves going beyond mere empirical regularities. Indeed for realism such regularities – like the manifest adaptedness of species to their environments - are characteristically part of science's explananda. An instance in Marx would be the correlation between fertility of land and differential rent, which he does not accept as in itself explanatory (see Sayer 1983: 49–52).

This means that giving a causal explanation necessarily involves elaborating’ a theory of causal mechanisms. It is only such a theory which makes a series of propositions into an explanation. We cannot answer a why question except by saying how. We cannot give a properly causal explanation without elaboration, for it is precisely in the supposed ‘elaboration' - the development of a theory of mechanisms - that explanation, as distinct from unexplanatory (or insufficiently explanatory) correlations of the koplik spots/measles (or acacias/long necks) variety, finally consists. In my final chapter I shall argue that in Marx, the analogous mechanisms lie ultimately in the actions of real individuals, and that causal explanation of social phenomena must acordingly be historical. We do not explain such things as Protestantism or law by their functionality but their genealogy.

To defend realism, whether as a philosophy of science or in relation to Marx's method, is beyond the remit of this book. I have made some gestures in this direction elsewhere (1983). But it is not irrelevant to my overall argument to suggest that what finally vitiates Cohen's account of functional explanation is exactly what realists argue undermines positivism, the lack of any material conception of cause. For Cohen's characterization of functional explanation is, as he goes to some lengths to show, a variant of Hempel's D-N model. Had Cohen operated with a material understanding of cause, he would never have made the mistake he did with the giraffes - excluding what actually makes their necks grow from the category of explanation proper, whilst ascribing a contingent empirical correlate (the dispositional properties’ of the environment) an illusory causal role. Nor would he have had the gall - or should I say the historical ignorance, comprehensible only in an analytic philosopher? - to suggest that, in effect, in the social world we have explained a phenomenon by demonstrating its functionality for another, whilst establishing how it was brought about (or is materially sustained) is a mere optional extra, irrelevant to the truth or falsity of our supposed explanations. To call such a procedure either historical or materialist is a travesty of either adjective, and about as far away from Marx as we could conceivably get.

124

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

INTERLUDE: THE GIRAFFES AMONG THE ACACIAS

125

The historicity of concepts

Let me now try to pull together some of the major arguments of this book. Since my main concern has been with the fundamental concepts of historical materialism - forces of production, economic structure/relations of production, and superstructure – it is worth reminding ourselves, to begin with, how Marx himself regarded theoretical categories. They were not free-floating analytic devices, innocent of historical content. Rather, for him ideas, categories' are but 'the abstract ideal expressions of . . . social relations. Indeed, the categories are no more eternal than the relations they express. They are historical and transitory products'(1846b: 189). Hence, 'in the succession of the economic categories, as in any other historical, social science, it must not be forgotten that their subject - here, modern bourgeois society - is always what is given, in the head as well as in reality, and that these categories therefore express the forms of being, the characteristics of existence'. Marx adds, with emphasis: 'this holds for science as well' (1857: 106). There is no theoretical Archimedean point, from which scientific analysis could commence, which lies outside the history and societies of which historical materialism tries to make sense. For Marx, our categories of analysis inescapably partake of the social reality they seek to depict; they bear the stamp of history' (1867a: 169).

This remains the case even for – and perhaps especially for - the most apparently pure and simple of abstractions, like the Hegelian concept of property, ostensibly the subject's simplest juridical relation', discussed in chapter 3. Such abstractions always presuppose some definite concrete substratum' (1857: 102). Similarly with the idea of the ‘abstract individual', discussed in chapter 4. Indeed Marx argues that ‘as a rule, the most general abstractions arise only in the midst of the richest possible concrete development, where one thing appears as common to many, to all' (ibid.: 104). ‘Simple' categories are latecomers on the historical stage. The historicity of such analytic categories is not, however, by any means always apparent. They appear as, precisely, abstractions - concepts whose content and validity are not circumscribed by time or place, which may therefore be applied quite legitimately to the analysis of any mode of production or social formation. Marx questions this appearance of pure' abstraction, and denies such universalistic pretentions.

In the 1857 General Introduction, he considers this question at length with regard to the conceptual fundamentals of political economy. He identifies a persistent ambiguity of reference in economists' categories. The abstraction of the category "labour", "labour as such", labour pure and simple', for instance, in one sense expresses an immeasurably ancient relation valid in all forms of society'. At the same time, Marx maintains that the concept ‘achieves practical truth as an abstraction only as a category of the most modern society'. It becomes thinkable’ in such abstraction only in ‘a form of society in which individuals can with ease transfer from one labour to another, and where the specific kind of labour is a matter of chance for them, hence of indifference' (1857: 104–5). In Capital Marx suggests that Aristotle was precluded from developing a concept of value based on labour because of the absence of this material presupposition (1867a: 59–60); Greek society was founded upon slavery, and had, therefore, for its natural basis, the inequality of men and their labour-powers'. Elsewhere he develops similar arguments to explain the physiocrats limitation of the category of value-producing labour to agricultural labour alone (1863a: ch. 2; 1857: 104), and Petty's restriction of the labour theory of value to labour which produces gold and silver (1859b: 54). The 1857 Introduction draws this general conclusion:

This example of labour shows strikingly how even the most abstract categories, despite their validity - precisely because of their abstractness - for all epochs, are nevertheless, in the specific character of this abstraction, themselves likewise a product of historic relations, and possess their full validity only for and within these relations. (1857: 105)

THE HISTORICITY OF CONCEPTS

127

6