[OPE-L:1057] Re: Definitions and Tautologies

Gilbert Skillman (gskillman@mail.wesleyan.edu)
Wed, 14 Feb 1996 11:44:09 -0800

[ show plain text ]

I have argued that the expression for a commodity's value in the TSS
system is inconsistent with Marx's stipulations in Vol. I, Ch. 1;
specifically: "Commodities which contain equal quantities of labour,
or which can be produced in the same time, have therefore the same
value" and "The value of a commodity would therefore remain constant
if the labour-time required for its production also remained constant." [p.

The reason I gave is that variations across commodities and over time in the
prices of input commodities can cause variations in the value of a commodity
[as defined in the TSS approach] even though the time required to produce
the commodity does not change.

In the following I argue that Andrew's response to this critique
[OPE-L 1044] assumes what must be proven. That is, if we don't
*begin* by accepting that prices enter directly into the
determination of a commodity's value, then Andrew's argument falls to
the ground. The problem begins in the following passage: [...]

> Gil says that the value of a commodity, according to the TSS interpretation,
> won't be constant if the labor-time required for its production is constant.
> Thus, the TSS interpretation contradicts Marx. I agree that if the value
> would not remain constant, then this would indeed contradict Marx. But
> I don't agree that the value can change independently of the labor-time
> required for production--when the latter is understood in its full
> complexity.

The last phrase sets off immediate warning bells. Putting aside for
a moment questions concerning the definition of value, the labor-time
required for production of a commodity is in principle quite
straightforward and unambiguous, subject to the caveat that the choice
of production technique is in general influenced by input commodity prices.
I'll abstract from this problem in what follows by assuming fixed coefficients
production technology.

For a given production technique, then, the labor required for
production of a given commodity can be measured either as the
sum of direct and indirect labor ("living" and "dead" labour)
required to produce the commodity, or as the total living labor
required to produce the inputs expended in producing the commodity.
These two measures are equal if the technology is unchanged over
time. This last qualifier can also be dropped if we restrict
ourselves to the *socially necessary* labor time required for
production; in that case the two measures are always equal under
general conditions, as Morishima shows.

Thus, the notion of the labor time necessary to produce a commodity
is not "complex" at all in principle, and in particular does not
depend in any way on commodity prices for its expression.

Andrew continues:

> Abstracting from fixed capital, the vector of constant capitals in the
> TSS interpretation is c(t) = p(t)*A(t). Gil says that p(t) can vary
> arbitrarily with changes in demand conditions, even if the labor time
> required for the commodity(ies) production changes. I agree that p(t)
> can vary with changes in demand conditions (e.g., some producers can get
> more surplus-value when demand for their studd [their what?--GS]
> is strong, thus raising their output prices). But notice that changes in
> p(t) lead c(t) to change. And c(t) is part of the value of the commodity:

> v(t+1) = c(t) + L(t)

This last statement manifestly begs the question at issue. If one
doesn't accept that prices enter into the definition of the value of
a commodity, then one doesn't accept that c(t) = p(t)*A(t) is part of
the expression of a commodity's value Thus Andrew is here quite
clearly assuming what must be proved.

Andrew continues:

> In other words, the value of a capitalistically produced commodity has
> TWO parts. Not only the new value added, but the old value preserved and
> transferred. Both parts are required for the production of the commodity,
> and both parts are *labor-time*. p(t) is a vector of labor-times per unit
> of the commodities, just as L(t) is.

One can accept the first statement without accepting that prices
enter directly into the determination of a commodity's value. The
last sentence is problematic at best. Prices p(t), as they are actually
found in the world, are measured in money units, not labor time units.
One can *transform* the money-unit measures into labor-time measures,
but in order to avoid fundamental confusions this transformation must be
explicit, e.g. via multiplying money prices p(t) by the "value of money"
(which must itself be independently defined). See my long post in
response to Alan on this issue.

Andrew continues:

> Hence, if p(t) changes then the labor-time required for the commodity's
> production changes (or course, p(t) must itself reflect only socially
> necessary labor, and that is assumed in my formula).

No it doesn't. We saw above that the labor-time required for a
commodity's production can be--must be--defined quite independently
of prices. All Andrew has shown above is that if p(t) changes, then
the labor-time measure corresponding to p(t)*A(t) (which must be
derived by multiplying p(t)*A(t) by the value of money) changes. But
that is a tautology, true virtually by definition. Just because this
measure and the labor-time required for a commodity's production are
both measured in labor hours does not mean they are the same thing.
In asserting that they are, Andrew again quite clearly is assuming
what must be proved.

> Hence, as I and others understand labor-time required for production, then
> Gil's conclusion doesn't follow. Similarly, when he notes that two
> commodities with different vectors of input commodities will have different
> values, according to the TSS interpretation, even if they can be produced
> with the same amount of labor-time, I will say that their values differ
> because they're produced with different amounts of labor-time.

The reasoning that produces this conclusion manifestly begs the
question at issue. See above.

> The key issue, of course, is whether c(t) is an amount of labor-time
> required for production even if c(t) not = L{(I-A)**-1}A. I say that
> Marx says it is. Several references could be brought, and I cited Vol.
> III (Vintage), p. 265.

First: even if Marx said this, it doesn't make it correct. As Andrew
defines it, c(t) is manifestly *not* "an amount of labor-time
required for production"; it's a particular dollar figure which can
be transformed into a particular figure measured in labor time. But
the connection between this figure and *labor time required for
production* must be established, not assumed, no matter *who* asserts
this connection. Second: I don't see any passage on p. 265 that
supports the interpretation Andrew offers here. Which passage are
you referring to, Andrew?

> Here's the point. Say we have two sectors, one
> that produces commodities having a value of 100 and the other producing
> commodities having a value of 40. But they will sell maybe at prices of
> 90 and 50. Given only that total value = total price, then it is clear
> that the prices are just fractions of the total labor-time, albeit
> different fractions than the values.

The last sentence requires translation. If values are measured in
labor-time and the prices we actually find in the world are measured
in money units, then "total value=total price" is not only not
"given", it is nonsensical. Now, one can *define* the value of money
so as to *guarantee* that aggregregate prices = aggregate values,
measured in labor time, but then this isn't "given", it's dictated by
the definition of the value of money. And given that definition,
"prices [correspond to]fractions of the total labor time" as a
necessary consequence of the definition. But so what?

> Hence, the prices are themselves
> sums of labor-time, not giraffes, or Augustinian theology, or logarithms
> (of any color). This is not tautology--it depends on the conservation
> of value in exchange.

First, prices only correspond to "sums of labor-time" if one converts
them from money units into labor units, and the result, as we saw
above, is most definitely a tautology.

> Of course, this is not what most people have thought Marx meant, but that
> is precisely the problem. Their interpetation cannot make sense of the
> whole; it especially cannot obtain his major theoretical results. The
> above interpretation, I submit, can, and has done so on many fronts.
> Hence, it is preferable *as an interpretation of Marx's value theory*.

Given the foregoing, this remains to be seen, to say the least.

> One can of course keep trying to find missing pieces of the jigsaw
> puzzle that the TSS interpretation has lost under the rug, but I really
> think this is futile. The pieces are just fit together differently than
> in the standard interpretation, so that they *do* fit together.

By definition. And inconsistently with Marx's stipulations in
Volume I, Ch. 1, unless one assumes what must be proved.

> This is the 100th anniversary of Bohm-Bawerk's _Karl Marx and the Close of
> His System_. That whole century was marked by people, Marxists as well
> as non-Marxists, trying to find self-contradictions in Marx. I suggest
> a new beginning for the next century. Let's try to understand what he
> had to say. Instead of rushing to conclusions that A means X and B
> means Y, but X and Y are inconsistent, the way to *begin* to understand
> Marx--or anyone else--is to relect a moment and say, well, if X and Y
> are inconsistent, then maybe, just maybe, A does not mean X or B does
> not mean Y or both. Maybe, just maybe, there's another possible meaning
> that can make sense of things. Under what conditions would this person's
> conclusions follow from their argument? How can I change my inderstanding
> of the terms of their argument so that it makes sense?

Well, to the extent that the above refers to my own critique of
Marx's Vol. 1 Ch. 5 argument (which uses A's and B's in its condensed
form), even the TSS interpretation does not rescue Marx's argument
from its evident logical fallacies. But putting this aside, we've
seen that Andrew's story can be made consistent with Marx's
stipulation only if one either assumes what must be proved, or
assumes that logical inconsistencies can be removed by fiat.

> I do not think the proponents of the TSS interpetation are especially
> smart or clever or whatever.

Was this an issue?

> I think that we have been able to make sense
> out of Marx's work because we have *tried* to do so. It doesn't take all
> that much effort. Now, for a number of reasons, people don't want to
> look like religious fanatics, and trying to understand what Marx had to
> say by working backwards from conclusions to concepts (as well as forwards
> from concepts to arguments) may strike some people as apologetics. But
> I suggest that this is the way we all try to understand someone if we're
> really trying to understand them (e.g., somehow people manage to understand
> my posts despite all the horrible typos, because they infer the "real"
> words from the whole). And I suggest that we *must* do so if we are to
> understand Marx. Trying to find holes in his arguments and
> internal inconsistencies will not produce real understanding.

For what it's worth, I'm trying to understand what Marx said as well.
I believe that most if not all of us on this list are trying to do the
same thing. But I don't choose to ignore obvious logical fallacies
as a result.

> It has not. And the originators of this approach to the interpreation of
> Marx, quite frankly, did not want to.

Unlike present company.

> I do feel very strongly that people should be free to express their
> disagreements with Marx--that might lead to some real positive development.
> But his critics should express their disagreements as *disagreements* and
> not as *proofs* of "error" emitted inexorably by a disembodied, classless,
> logic machine.

What? Andrew, are you suggesting that the determination of a logical
fallacy (as opposed to an empirical or substantive error) depends on
what class one is in? May I go on record as rejecting this notion
unequivocally (with George Spencer-Brown's LAWS OF FORM as my
grounds), and suggesting that if this is true we should all give up
right now. We cannot possibly make progress.

> This is, at root, an ideological cover that serves the
> purpose of making it look like there's no one here except us disinterested
> scientists, rather than the clash of antagonistic interests.

Untrue. For example, in emphasizing the (unambiguous) logical
fallacy in Marx's Ch. 5 argument, I am unavoidably indicating that
certain issues addressed in CAPITAL are of paramount importance to
me. I've explained why I think they're important. I've also argued
why this fallacy must be confronted in order to understand Marx's
broader historical materialist account of capitalist exploitation.
But this interest is not primarily "antagonistic" to others on the
list since in principle we all share an interest in moving Marxian
political economy forward.

I can't see how ignoring manifest, and substantively pernicious,
logical fallacies in Marx's account serves this common interest.

I can't see how assuming what must be proved is a satisfactory method
of avoiding some of these logical inconsistencies.

In solidarity, Gil