[OPE-L:1110] Response on TSS, part II: Details

Allin Cottrell (cottrell@wfu.edu)
Sun, 18 Feb 1996 17:53:33 -0800

[ show plain text ]

This is the second part of my response to Andrew's postings of
960214-960217. It is in answer to Andrew's request that I "deal with the
substance" of his remarks. I pull no punches, but I will say that I have
made a concerted effort to understand what Andrew is saying and have tried to
avoid "easy" responses that score points (or not) by placing a vulnerable
interpretation on his remarks. The indented text below is Andrew; the
flush-left is me.

[H]as anyone refuted, or conversely conceded, that simultaneous valuation
is incompatible with the determination of value by labor-time? No.

Here is a refutation. Simultaneous valuation means v = Av + l. The value of
each and every commodity in the system is determined by the sum of the direct
and indirect labour-time required to produce it under current conditions of
production. Consider v, A and l all to be dated (as of the same "period").
A change in the A matrix or the l vector from one period to another will
result in a new set of values, corresponding to the new sum of direct plus
indirect labour-time required to produce the commodities under the new
conditions of production (which will differ from the historically embodied
labour-time). This is precisely "the determination of value by labour-time",
on a reasonable -- I do not say a uniquely "correct" -- reading of
the phrase.

Andrew's worry seems to be this: Production takes real time; there is a
definite positive interval between the purchase of inputs and the sale of
outputs. Yet this system gives just one value per period to each commodity
(as input at the beginning of the period and as output at the end). The
value is therefore "frozen" for one period. But so long as it is frozen it
can't be influenced by changes in labour-time requirements.

But the "frozen values" idea is the wrong way to look at it. It's not that
values are "not allowed to change" from the beginning of the period to the
end (despite changes in labour requirements that might be going on); it's
that, on this accounting, any changes in labour requirements that have
occurred in the current period relative to the last are applied
retrospectively to the valuation of the inputs. That is, what would appear
as a difference between this period's input values and this period's output
values, on historic labour-embodied accounting, is transposed into a
difference between this period's input values and _last_ period's output
values. In both cases the difference stems from a change in direct or
indirect labour-time requirements (and nothing else). Whether one ought to
account required labour-time on a current reproduction basis is debatable (I
agree with Paul's point that this depends on the context, on the sort of
production process one is looking at). But one can't say that it is
"incompatible with the determination of value by labour-time."

Somewhat related:

And why is this notion of Torrens having assumed unchanged production
conditions cropping up again? Do you have any shred of evidence for this

Sauce for the gander -- it's the assumption that makes sense of Torrens'
argument. He obviously knew that the price of corn at harvest time could
differ from that at seed time. He was no idiot.


Has anyone refuted the claim that the TSS interpretation vindicates the
internal coherence of Marx's theoretical conclusions far more than any
other interpretation, or, conversely, conceded the validity of the claim.

I have conceded that the TSS interpretation saves Marx's statements in vol.
III regarding aggregate price-value invariances. But "internal coherence" in
a broader sense? No, because I believe that the TSS view contradicts Marx's
often-repeated and clear statements of what the value of a commodity amounts
to. See my previous posting.


As to Allin's reinterpretation of pp. 264-65 of Vol. III, it doesn't work.
The reason is that Marx doesn't have two systems of values and prices the
way you do.

Begs the question. I offered a detailed reading of a specific passage that
supported the "two systems" reading.


[Allin] says [Andrew] should be bothered by his matrix example, because the
"definition" of the value vector v = L{(I-A)**-1} is the correct,
unambigous vector of embodied labor-times.

Come on now! This is precisely what I deny.

Andrew, how can you? In the simple, static example I gave (where the
distinction between valuation at historic and at reproduction cost
disappears), v = L{(I-A)**-1} just is the labour-content. It is simply
another way of writing the sum of the direct labour inputs required to
produce the gross outputs of the commodities. "Value" is a technical term,
and potentially subject to competing interpretations; but "the labour-time
required to produce X" is plain ordinary English and cannot be assigned a
different meaning at will, or all discussion is at an end.


Andrew talking about the "simultaneist" model:

[M]y theorem is true for values as well--the "value rate of profit" of the
simultaneists is uniquely determined by physical quantities and relative
values (this is likewise true of each sector's profit rate). Now imagine
the vector of unit living labor coefficients changes from L to kL, where k
is a scalar. Then the vector of unit values will change from

V = L{(I-A)**-1) to kL{(I-A)**-1) = kV

and all reltive values will remain the same. Hence, no matter how much L
changes, if the physical outlays and outputs remain the same, and each L
changes proportionately, each and every profit rate will remain the same!

On the standard view with which I am familiar, Andrew has not given us enough
data to determine what will happen to surplus value and profit. Suppose the
aggregate working day, and hence the total production of value, remains
constant. What has happened to the value of labour-power? If the
real-commodity wage remains unchanged, then the value of labour-power has
increased. Therefore the surplus value (and hence the profit) must have
fallen. This ought to be familiar. It is -- in essence -- Ricardo's theory
of the falling rate of profit. Suppose on the other hand that the workers
suffer a permanent cut in the real-commodity wage, sufficient to leave the
value of labour-power unchanged in the face of the increase in the
labour-time required to produce wage-goods. Then, and only then, will
surplus-value and profit remain unchanged. What is Andrew's alternative


In response to Paul's saying that simultaneity is a red herring:

[J]ust because [Paul] would be happy with temporally determined values,
that doesn't mean anyone else would be. Hence, my raising the issue of
simultaneity is certainly no red herring. ... [H]ow about my discussion
with Gil (and some others) on the meaning of Marx's statements that a
commodity's value is determined by the labor time needed for its production
or contained in it? Gil and others have claimed that the meaning of this
is univocal, that values are uniquely determined by current production
conditions. In other words, values are solely a function of the current A
matrix and L vector. [But] if there are *multiple possible
interpretations* of such statements, then one can no longer use their
alleged univocal meaning as the basis for saying that past distribution of
value has no effect on subsequent values.

While I can't speak for Gil, I strongly suspect that he, like Paul, would be
quite willing to accept v(t) = Av(t-1) + l(t-1), for the sake of argument at
any rate. I certainly would. As compared to the "simultaneist" version,
this is just a different way of adding up labour-content (in historic terms
rather than at current reproduction cost). It's compatible with the notion
of the labour "contained in" a commodity. It doesn't confuse labour-content
with labour-commanded, which, to my mind, is the main issue here. To accept
that there might be two alternative definitions of labour-content, which
diverge when the conditions of production are changing, is not to say that
anything goes.


On 960217 Andrew issued a

SECOND GENERAL REQUEST for those who are not yet convinced of the
superiority of the TSS interpretation to specify clearly the conditions
under which they would accept it. Again, if one cannot specify conditions
which I am in principle able to fulfill, i.e., not "if Marx rose from the
dead and said the TSS interpretation is right," then one is, like it or
not, a dogmatist.

"You are a dogmatist if you cannot specify conditions under which you
would concede that you are really Ivan the Terrible." Not obviously true,
is it? Some propositions are so much at odds with everything one knows
that one's inability to specify conditions under which one would count
them as correct is not a sign of dogmatism, but rather of sanity. Ok,
that's a general philosophical point. The case in hand is not so extreme.
But it's still pretty radical. Personally, about the only thing that
would convince me that Andrew's reading is right would be if a
long-forgotten notebook were found in which Marx plainly states something
like this: "When I say 'the labour contained in X' or 'the labour expended
in the production of X', or 'the labour required to reproduce X', I mean
'the sum of the direct labour required to produce X and the labour
commanded by the means of production used up in the production of X'." (I
might add that if this notebook were found, though I'd have to concede
Andrew's interpretive claim, Marx would come down substantially in my


Textual analysis -- Andrew on Capital:

[Was] Marx ... admitting "error" when he wrote that "if the cost price of
the commodity is identified with the value of the means of production
consumed in it, it is always possible to go wrong" (Vol. III, p. 265
Vintage)? To *me*, it is "natural" to read this passage as saying that the
value of the used-up constant and variable capital is not identical to the
value of the means of production and subsistence , precisely because the
deviations of prices from values--"this error in the past" (ibid.)-- then
affect the value of new capital outlays. And there are several other
passages in Vol. III that say the same thing. Plus, e.g., the passage on
p. 317 of Vol. I, where Marx explicitly says a change in the "price" of
cotton changes the amount of value transferred to the product (yarn). Plus
the greater part of Ch. 6 of Vol. III is just about this point--although
Marx is still assuming values = prices there, he says the analysis (effect
of price changes on the rate of profit) is valid no matter what the reason
for the price change.

I cannot afford the time to contest Andrew's interpretation of every passage
in Marx. A sample will have to do. Let's take, for instance, ch. 6 of Vol.
III. (References are to the Progress edition.)

"If the price of raw material, for instance of cotton, rises, then the price
of cotton goods -- both semi-finished goods like yarn and finished goods like
cotton fabrics -- manufactured while cotton was cheaper, rises also. So does
the value of the unprocessed cotton held in stock, and of the cotton in the
process of manufacture. The latter because it comes to represent more
labour-time in retrospect and thus adds more than its original value to the
product which it enters, and more than the capitalist paid for it." (112)

Out of context, this does look as if it might be supportive of Andrew's
interpretation. But let's read carefully. That a change in the _price_ of
cotton should induce a change in the _price_ of cotton products (the first
sentence) is no problem for the standard view of value. The second sentence
looks like it might be a problem though: a price change seems to be inducing
a value change. But what does Marx means when he says that the higher-price
cotton "comes to represent more labour-time in retrospect"? Is this Andrew's
labour-commanded (i.e. it "represents more labour-time" in virtue of the fact
that its price has risen, period)? No, because on the next page we find
this: "We proceed in this entire analysis from the assumption that the rise
or fall in prices expresses actual fluctuations in value." (Andrew: I will
return to the sentence following this one in a moment.) The hypothesized
increase in the price of cotton was not just any old price increase; it was
an increase due to an increase in the labour-time required, under current
conditions of production, to produce the cotton. This increases the value of
existing stocks of cotton because Marx is operating with the dreaded
simultaneist conception of valuation:

"The value of every commodity -- thus also of the commodities making up the
capital -- is determined not by the necessary labour-time contained in it,
but by the social labour-time required for its reproduction. This
reproduction may take place under unfavourable or under propitious
circumstances, distinct from the conditions of original production. If,
under altered conditions, it takes double or, conversely, half the time, to
reproduce the material capital, and if the value of money remains unchanged,
a capital formerly worth L100 would be worth L200, or L50 respectively."
(p. 141)

Thus the original passage provides no support for Andrew's contention.

Addendum: When Marx says that he's assuming that price changes reflect value
changes in the "entire analysis" of chapter 6, he follows this with the
statement "But since we are here concerned with the effects such price
variations have on the rate of profit, it matters little what is at the
bottom of them." (113) Andrew jumps on this, but again let's read carefully:
it "matters little" what the price changes are caused by if one is purely
concerned with the question, What will happen to the rate of profit?
Clearly, a rise in the price of materials will reduce the capitalist's profit
regardless of its source. On the other hand, it matters a lot if one is
trying to use the passage to make inferences about Marx's conception of


Allin also says Marx didn't know linear algebra. Again, this begs the
question--according to WHOM does the transformation of values into prices
need to be "solved" by the methods of linear algebra? [T]o claim that Marx
knew his transformation account was erroneous is in effect to charge him
with blatant dishonesty, because the rest of Vol. III relies on the
conclusions he arrived at in Ch. 9.

My point about Marx and linear algebra was this: Since Marx was not familiar
with linear algebra, he was not in a position clearly to perceive the
problems associated with value-to-price-of-production transformation as seen
on the "standard" view. Therefore, the fact that he saw the transformation
as unproblematic is not, in itself, evidence that he _didn't_ hold the
standard view. I'm not saying (at all) that Marx knew his transformation to
be erroneous; I'm saying that if it is erroneous (as it is, to some degree,
on the standard view) it is quite understandable that Marx would not have
noticed this. This argument does not prove that Marx held the standard view
-- it just undercuts one reason for believing that he held a non-standard
view. Let there be two interpretations, I1 and I2, of some proposition P as
uttered by X. The Principle of Charity generally favours I1 over I2 if P is
true on I1 but false on I2. But this reason for favouring I1 is undercut if
it is shown that X was not in a position to see that P was false, as
interpreted by I2.

Allin Cottrell