Paul C. writes:
> I agree with Gil that one can not logically deduce from the
> exchange of equivalents that there must be a conserved substance
> in the exchange relation.
>
> One can show that the only metric on the space of commodity bundles
> that is consistent with
>
> 1. the physical and legal seperability of commodities
>
> 2. ensuring that all paths effected by exchange, which leave a point
> on one of the axes of commodity space, and return to that axis are
> closed
>
> is one that can be modeled by a mapping function from commodity bundles
> onto a one dimensional sub-space. It is this property of being topologically
> equivalent to 1 space, that accounts for the apparent conserved substance.
>
> However, given this property of commodity space as a system of representation,
> the question arises what real-empirical variable will capture the
> attractor of the space, not in all, but in the statistically dominant
> majority of cases.
>
> Given that mappings onto 1 space are characteristic of systems governed
> by conservation laws, what does this formal property of exchange select
> out in practice to be the thing that is conserved?
>
> The classical political ecnomists asserted that it was labour, Ricardo
> went so far as to quantify this as being 93onserved. The current
> literature seems to indicate that he was more or less spot on, with that
> estimate.
>
> Gil is justified in pointing out the inconsistencies in Marx's attempt
> to logically deduce the value substance from the form of value.
> But one can arrive at a correct conclusion using falty logic.
> If the conclusion is so evidently correct, it is easy to overlook
> the faults in the logic by which one arrived at it.
I agree that consistent and close correlation of values and
empirically based production prices suggests a _prima facie_
case for the explanatory significance of labor values.
But I also think something significant is lost when the empirical
correlation is severed from any (necessary) theoretical underpinning.
Because then the explanatory status of labor values can be challenged
by any other "value" measure which can be shown to be as strongly
correlated with such prices.
Example 1: Take any good which is either part of the wage bundle or
used in the production of goods which are part of the wage bundle.
One can then calculate an alternative value measure based on this
good--e.g., "corn values" or "steel values". If labor values are
strongly correlated with prices, then so must these alternative
values, since they reflect the same underlying data.
Example 2: I've seen presentations by Anwar's students in which they
demonstrate a strong correlation between labor values and prices
calculated in terms of the Sraffian "standard commodity", which
suggests that the latter would be strongly correlated with production
prices as well. Does this suggest that Sraffian standard commodities
have explanatory significance?
Alternatively, "correct" theory may, at least initially, offer poorer
empirical predictions than an incorrect theory. Paul Feyerabend
offers several examples of this in his AGAINST METHOD.
Nothing urgent or underhanded here--I'm just curious.
Gil