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