> As I read Marx, he defines the value of a commodity as the
> labor time socially necessary for its production. But
> exchange is not production. Therefore the aggregate of
> commodity value is "conserved" by exchange, simply because
> exchange by definition does not affect that aggregate one
> way or the other.
I agree with Gil's reading, and also agree that the conservation
of value in exchange, in this sense, is not an empirical
proposition.
> At the disaggregative level, if "equality" means that goods
> of equal labor value are exchanged in any case, the claim
> isn't true in general (I know we all know this, I'm just
> covering the bases).
It's certainly not true *exactly and in every case*, and so any
formulation of the labour theory of value that implies that
proposition must be false. But it is *close to true in the
great majority of cases*. I think we'll find that most
conservation laws are of the latter character (e.g. the famous
conservation of energy is not exactly satisfied in the case of
quantum phenomena).
Paul C. said earlier that the (close approximation to)
conservation of labour-time in exchange is "a brute fact", and
not deducible a priori. I have a slight difference with this; I
agree that it is verified empirically (as a good statistical
approximation), but I think it can be supported by a priori
argument.
Adam Smith had the right general idea when he said (I'm
paraphrasing from memory here) that labour is the "original
purchase price" of useful goods from nature. Prior to any
detailed statistical work, and on the basis of very broad
observation of the economy (not *purely* a priori, admittedly)
one can arrive at the idea that the amount of human labour
required to produce something represents its "true cost" from a
social point of view (for all reproducible goods). Why?
Because all other inputs - besides inherently non-reproducible
natural resources, which give rise to the well-understood
"deviation" from the labour theory of value called "rent" - can
be produced with a sufficient application of labour-time. And
because we humans (also) like to use our time for things beside
labour, so that labour-time is a cost from our point of view.
Note that the second-last sentence above is not true if you
substitute *anything else at all* for "labour-time". Labour-time
can produce steel, but steel cannot produce labour-time, even if
steel enters the workers' "consumption bundle". Labour is not a
commodity produced under capitalist conditions of production.
Labour is "out of the loop" of inter-industry input-output
relations (and abstract production theories that purport to
treat labour as "just another commodity" are grossly
misleading).
Allin Cottrell.