>And on 'productive labour' - here's a beaut paragraph I've kept since
>Rakesh sent it some years ago:
>
>"As Geoffrey Kay whom I basically quote here has argued, productive labor as
>only that labor which produces surplus value is thus not a theoretical
>definition made for the purpose of studying capitalism, but a practical
>definition made by capitalism in its modus operandi. Hence valid objections
>to this practical defintion of productive labor that it is unrealistic (it
>would be realistic to include all the labor which makes indispensible
>contributions to surplus value production, eg., domestic work), irrational
>(labor producing armaments or luxuries can produce surplus value), or
>inconsistent (a nurse in a private hospital produces surplus value while
>one in a state hospital does not)--all these valid objections--are properly
>directed, as G Kay puts it, against capitalism itself."
>(Geoffrey Kay, The Economic Theory of the Working Class, St Martin's, 1979
>p. 133)
Excuse me, but this distinction between the practical and theoretical is silly. Nor are Kay's objections real, Rob (and Rakesh). They can be handled by a reasonably functional definition of productive labor, and have been by many writers. But before discussing that, we need to get on the same page. There are two distinct ways to use productive labor as an analytical tool: (1) to explain the existence of surplus value, and (2) for a theory of growth (laws of motion). Different definitions, it has been argued, are needed for different uses. To sort these out you need to pay attention to Marx's three types of goods (called departments): means of production (I), consumption goods (the basket of goods and services that constitute labor's (social) subsistence (II), and "luxuries" (III).
Marx's definition of productive labor is straightforward. That labor is productive which produces surplus value for capital, and thus works for the self expansion of capital. In production, of course, but in the sphere of exchange (distribution) too. Labor that is necessary to package, transport, handle, store, and distribute products for capital adds to value and surplus value. And some supervisory labor is also productive: that necessary to coordinate and unify production and distribution processes.
Not only is this construct necessary for Marx, it is central to his whole analytical system: The productive labor definition "expresses precisely the specific form of the labour on which the whole capitalist mode of production and capital itself is based" (Theories of Surplus Value I, p.396). Marx spent about 300 pages in Capital and Theories talking about it. Eric Olin Wright (Class, Crisis, and the State) added an important qualification to reflect current circumstances: today productive labor is best thought of as two dimensions of labor, often within the same job, rather than two types of jobs or workers.
But some (e.g., Joseph Gillman, Jacob Morris, Justin Blake, John Harrison, Ian Gough, and James Becker, in books and mainly the pages of the New Left Review and Science and Society, mostly in the 70s) have argued that while the definition is necessary to explain surplus value, it is not sufficient to analyze how capital evolves (theory of growth). For that, they say, you must distinguish between kinds of goods and services. Only those in Depts. I and II are part of system reproduction (i.e., they are goods necessary for the reproduction of capital and labor). Luxury goods (Dept. III) are not. When produced by capital, luxury goods do produce surplus value, but it is not part of the *reproduction* process, and so should not be part of any analysis of growth. Each writer, of course, has further twists on this idea, but you get the point.
One further complication should be mentioned. There are actually three types of labor: productive, unproductive, and noncapitalist. As I argued a while back when folks were discussing the role of teachers, noncapitalist labor that enhances surplus value in the first two departments cna be considered productive labor(e.g., as teachers do by adding value to the future labor power of students, assuming what they add exceeds the surplus value they use up).
So you can see, can't you, the lack of substance to Kay's "objections". Depending on the analytical purpose for which you use productive labor, each is taken care of. To the extent it enhances workers' ability to perform on the job, domestic work can be considered productive under the growth theory version, but not Marx's original. Labor producing armaments for capital and yielding surplus value is productive for Marx, but not for the growth theorists. There is nothing inconsistent from separating capitalist and noncapitalist labor, as in Kay's nurse example.
This debate about the meaning of productive labor raged mainly in the 70s, right before Kay wrote his book. Judging only by the characterization (unfair, of course) he seems to have missed it. I haven't read his book, but is it possible that he is one of those people that Carrol is (rightly) worried about: someone who confuses growth analytics with ruminations about class consciousness?