ABSTRACT LABOR AND

COMMODITY PRODUCTION
John Weeks

First published in
Research in Political Economy,
ed. Paul Zarembtka,
Vol. 12, pp. 3-19, 1990.
From Marx to Mao

M
L
Digital Reprints
2016
www . marx 2 mao . com
ABSTRACT LABOR AND
COMMODITY PRODUCTION

John Weeks
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

I INTRODUCTION

The analysis of commodity exchange has produced three explanations for the
relative prices that manifest themselves in exchange: Ricardo’s embodied labor
theory, the marginalist/neoclassical general equilibrium theory (first formulated
by Walras), and Marx’s analysis based on the distinction between the use value
and exchange value of commodities. Each of these attempts to explain the
determinants of exchange value that underlie the observed fluctuations in market
prices. For Ricardo, the underlying determinant was the quantity of labor (past
and present) expended to produce each commodity. In the case of the neo-
classicals, transitory market prices represent fluctuations around a “long run”
equilibrium price, which is logically constructed on very restrictive assumptions.
The exact nature of Marx’s underlying determinant of relative (and absolute)
prices has been a matter of considerable controversy. In the 1970s a consensus
developed among Marxists in Europe and North America that Ricardo’s embod-
ied labor view was inconsistent with the method and analysis of Capital. But

––––––––––––––
Research in Political Economy, Volume 12, pages 3-19.
Copyright © 1990 by JAI Press Inc.
All rights of reproduction in any form reserved.
ISBN: 0-89232-990-4

3
4 JOHN WEEKS

while it was not difficult to demonstrate that Marx’s methodological break with
Ricardo lay precisely in differences over value theory, it has proved quite difficult
to reach agreement over what it was Marx did to make this break. In this paper I deal
with two prevailing—and perhaps most common—interpretations of
Marx’s law of value. I argue that in both cases the rejection of Ricardo’s ap-
proach has led to a retreat into bourgeois general equilibrium theory, implicitly
in one case and quite explicitly in the other. The cure proves worse than the dis-
ease, however, robbing the law of value of its power as a critique of capitalist
society.
Thus, the theme of this paper is a simple one: Ricardo’s embodied labor theory
of value should be rejected as theoretically unsound; but it should be replaced
with an analysis based on the distinction between use value and exchange value,
not by general equilibrium theory in disguised form. To develop this theme, first,
I clarify the concept of abstract labor, and in doing so consider the precise sense
in which (1) abstract labor is “the essence of value,” and (2) value arises in the
“interaction of production and exchange (distribution).” This clarification is
made in the context of what I call the “validation” school of interpretation of the
law of value, in which a methodological break is made with Ricardian value
theory by abandoning the sphere of production altogether. This approach proves
unsatisfactory because the validation school can produce no underlying determi-
nant of market values at all. It might be said that the validation school looks
below the level of appearances (market price) and finds nothing there. However,
even in theory the analysis cannot be left as the validation school does. The
implicit solution of market prices in its system can be found in the analysis of the
Uno school interpretation of Marx, which is considered in the subsequent sec-
tion. Here, a close relative of bourgeois general equilibrium theory enters as the
basis of the law of value. Criticism of this approach necessarily leads to treating
the question of whether the law of value should be understood as an equilibrium
theory, and in particular whether its robust conclusions refer to a state of general
equilibrium The last section of the article addresses this question, and argues
that Marx’s method involved the rejection of the concept of equilibrium, and that
the law of value can be (and should be) formulated in other terms. It is further
argued that the alternative formulation produces a value analysis that is determi-
nant, and that allows one to base the analysis of relative prices in the sphere of
production.
As a prelude to the main argument of the paper, a brief summary of the Marx-
ian objections to the Ricardian approach to value theory is necessary. Until the
early 1970s, one could summarize the standard Marxian formulation of the law
of value by the phrase, “value is created in production.” But in light of the
debates between Marxian writers and neo-Ricardians in subsequent years, this
simple statement could not be sustained without some elaboration, for it is open
to an embodied-labor interpretation of value. That is, the statement is consistent
with a system in which values are derived directly from the technical coefficients
Abstract Labor and Commodity Production 5

of production, which among other things implies that value in terms of labor time
can be deduced directly from the observation of labor-in-production. This treat-
ment of value, in addition to making the arbitrary assumption of homogeneous
labor, falls victim to the Sraffian “redundancy” criticism: if exchange values can
be deduced directly from production coefficients, then deriving values in labor
units is an unnecessary diversion (see Bandyopadhyay 1985).
The “redundancy” argument as such is less serious than other problems im-
plied by the Ricardian approach. More theoretically worrisome, the embodied
labor interpretation fails to capture a central feature of capitalist society, namely,
that production techniques are not uniform for each commodity, thus are not in
any sense “given.” Treating techniques as given adopts the neoclassical method
of considering production as purely a technical question, determined outside of
the system. DeVroey has correctly described this approach as “naturalism”
(DeVroey 1982). Similarly, Levine argues that assuming a system of given
techniques is the epitome of Ricardianism. It is analytically parallel to explaining
production and distribution on the assumption of given fertility of the soil
(Levine 1981, Introduction).
We have two distinct problems here. One is the fact that different commodities
are produced with different types of labor. The second is the fact that the same
commodity may be produced with essentially similar types of labor, but under
differing efficiencies in the use of those types of labor The first issue—hetero-
geneous labor as such—has received more emphasis in the debate over the
validity of the law of value (though the two are frequently subsumed or confused
within the same discussion). The vehicle to deal with the two problems has been
Marx’s insight that commodities have a two-fold nature, and, therefore, that the
labor that produces commodities also has a dual character. The group of writers
that uses the two-fold nature of commodities to deal with the problems men-
tioned above I call the “abstract labor” school. There are important differences
of emphasis within this school of interpretation, including differences over the
sense in which labor has a two-fold nature.

II. PRIVATE AND SOCIAL LABOR: A CRITIQUE

Although many of the problems with which I subsequently deal are quite subtle,
the central problem of value analysis is straightforward. One can state the prob-
lem clearly: (1) the labor carried out in production is heterogeneous (both in the
sense of different types of skills of labor and different efficiencies in the use of
labor); (2) commodities exchange via money, which expresses an equivalence
among those commodities; therefore, (3) if we maintain that in reality the ex-
change of commodities is governed by their content of labor, then we must
explain how heterogeneous labor-in-production (concrete labor) is expressed as
homogeneous labor-in-exchange (abstract labor). To take an example, because
6 JOHN WEEKS

the work of carpentry is in fact different from the work of spinning cloth, how
can we say that the rate at which chairs exchange for a length of cotton is
determined by the labor time it takes to make each? Ricardo and Ricardians avoid
this problem by presuming all labor to be the same (or reducible to such in some
manner). It has become an identifying characteristic of Marxian value analysis to
reject this approach. Yet the problem remains for the abstract labor school—how
is heterogeneous production labor rendered into homogeneous labor in
exchange?
Two more-or-less polar solutions have been offered within the abstract labor
school of analysis. The first treats the metamorphosis from concrete to abstract
labor as occurring in exchange. By this interpretation, the twofold nature of
commodities involves the distinction between private and social labor. This
distinction leads to a stress upon the necessity for private labor to be “validated”
in exchange. This interpretation I call the “validation” argument, and is present-
ed quite convincingly by Mohun (1984–1985). If one adopts this view, it be-
comes problematical indeed to treat value as arising or created in production,
except in a tautological sense. Commodities must be produced; they have value
(because they exchange), but they cannot be said, in Marx’s words, “to enter the
market with their values stamped upon them” Rather, it is the market that
stamps them as values. If we follow this interpretation, exchange asserts itself as
dominant over production in the formation of values, and the relation between
value and production is said to be qualitative, but not quantitative. That is,
because value is price, it cannot be said that value determines price. The essence
of this interpretation is its stress on the necessity for labor-in-production to be
transubstantiated into labor-in-exchange.
The second approach places the process of abstraction (particular to general
labor) in production. It is argued that exchange renders concrete (private) labor
abstract only in the formal and definitional sense of establishing equivalence.
The real or essential process of abstraction occurs in production itself. In this, the
“production” argument, exchange only reflects what has occurred prior to ex-
change, in the labor process. The essence of abstractness is the real homogeneity
of labor created by the development of capitalism (Gleicher 1985). The argument
is that as capitalism develops, workers are deskilled, rendered general labor in
practice, and the equivalence of labor in exchange reflects its actual sub-
stitutability in the production of different commodities. By this analysis labor is
rendered homogeneous in production, and this homogeneity is expressed (one is
tempted to say, “merely” expressed) in exchange. Certainly the deskilling point
is a valid one. Capitalism tends to eliminate artisanal labor, and a convincing
argument can be made that this might be generalized as a process that renders the
working class into a largely undifferentiated mass of laboring potential. It is a
fact, for example, that a worker in his or her life may assume many jobs involv-
ing the production of many different commodities.
Much empirical evidence indicates that this second argument is correct. How-
Abstract Labor and Commodity Production 7

ever, it does not refer to abstract labor, but to concrete labor. In effect, it accepts
the neo-Ricardian interpretation that the major difficulty in relating labor-in-
production to labor-in-exchange is the reduction of complex labor to simple
labor. The abstract labor theorists are concerned with a different issue: how is
social reproduction achieved in a commodity-producing society? In other words,
one must explain how the labor performed in particular workplaces translates
into the general, value-creating labor for capitalist society. This refers to the
differing efficiencies in the use of labor by different producers. Whether workers
are different or identical in fact or fantasy is not directly relevant to this question.
The deskilling argument implies that concrete labor is rendered homogeneous
over time, and in and of itself has nothing to say about abstract labor.
Recall that our purpose is to investigate the basis of relative exchange values
within the Marxian methodology. Let us now turn to the validation theorists, and
the reason for the identifying modifier, “validation,” will become clear. I begin
by referring to two quotations, the first from Himmelweit and Mohun, the second
from Mohun. My critical comments are not intended to be a general rejection of
the overall positions of these two authors. Both articles represent important
contributions to our understanding of the law value. However, the interest here is
in what is missing in their approach. First, consider the treatment of social labor,

It is the process of exchange on the market that manifests the social character of individual
labors, establishes the social connections between independent commodity producer . . .
(Himmelweit and Mohun 1981, p. 233).1

Writing on his own, Mohun carries this argument further ,

The answer to the question of how private labor becomes social labor does not lie in an
analysis of the production process itself . . . Only through exchange are the products of
different labors commensurated, and hence only through exchange can it be determined a
posteriori whether the concrete labors of a particular production process are to count as a
portion of society’s total labor and to achieve social validity thereby . . . The moment of
exchange is thus the critical one (Mohun 1984–1985, p. 401).

In both these quotations the reference is to private or individual labor, which

implies that the product of that labor belongs to the producers. The emphasis that
labor carried out in production is private I take to mean privately owned, and
executed individually, not in direct concert with others.
Let us be clear about the argument being presented by the validation theorists.
Value only expresses itself in exchange (price), and is only determinant after (a
posteriori) exchange has occurred.2 If this is the case, can value be said to exist
separate from price as the determinant of price? Obviously not. This conclusion,
that value has no determinate existence prior to exchange, rejects Marx’s argu-
ment that price is merely the phenomenal form of value. No longer is value the
hidden, underlying determinant of price. We have looked beneath the price form
8 JOHN WEEKS

and nothing was there to be found. Should we be repelled by this conclusion,

Mohun seeks to reassure us, saying,

This is not to obliterate the distinction between value and exchange value [price]; nor is it to
assert the domination of circulation over production (Mohun, 1984-1985. p. 401).

However, if words have definitive meanings, and Mohun is an author who

writes with great clarity and careful choice of words, the distinction between
value and price has been obliterated and the domination of circulation clearly
asserted (“the moment of exchange is the critical one”; see above).
Valuable as Mohun’s insights are, I argue that his specification of the value
relation is flawed. Although value must by necessity take the form of price, it has
an ex ante existence with regard to exchange. It is, I shall argue, the underlying
determinant of price, unseen, but nonetheless real. Price itself is the superficial
expression of production relations, where the term “production relations” refers
to both the technological and social aspects of the labor process. The theoretical
difference between this position and the validation argument has a clear origin.
Those who equate price with value and, therefore, reduce value determination to
exchange are in effect considering value in the context of simple commodity
production, a situation in which value has no determining role.3 The difficulty of
separating value and price disappears once one considers value in the context of
capitalist production; that is, within the circuit of capital.4
Capitalist society is based on the separation of labor from the means of produc-
tion, which implies the progressive destruction of petty-commodity or individual
production by artisans and peasants. Capitalism throws workers together into
workplaces in increasing numbers, where their labor is collective labor. The
work of a laborer employed in a plant with a thousand other workers can in no
sense be described as private, not as individual, the terms used in the validation
argument to characterize commodity-producing labor. Capitalist production in-
volves collective, cooperative labor, directly social, consciously directed and
controlled—the collective and cooperative power of the working class domi-
nated by and subsumed under the authority of capital. That which is private and
individual is not labor-in-production, but the commodity that is the result of the
production process. Thus, labor-in-production itself is social labor; the collec-
tive, organic functioning of a class. It is the product of that labor, because it is a
commodity, that is private; labor appropriated by a dominant class that must be
exchanged as a commodity. At the highest level of abstraction, we can summa-
rize the mistake of the validation argument by saying that it treats capitalism as a
commodity circulating society (its appearance), rather than a commodity produc-
ing society (which is what it is).
Labor-in-production is collective, social labor, while labor in commodities is
appropriated labor, private in form. This is the basic contradiction in capitalist
society. Labor is carried out collectively, but is appropriated in the form of its
Abstract Labor and Commodity Production 9

antithesis, private labor, the private property of the capitalist. It is for this reason
that Marx commented that the worker’s product presents itself to the worker as
something alien to him or her, as the property of another. This, of course, is the
fetishism of commodities. Social relations among people (the class relations of
production) present themselves as exchange relations among inanimate objects
(commodities). The collective labor of the working class appears as the private
labor of individual capitalists. The value that workers create in production ap-
pears to gain validation from exchange. The all-pervasive illusions of commodity
fetishism generate the impression that it is exchange that is the “critical mo-
ment” in the determination of value.
Let us now shift to a less abstract criticism of the validation argument. When
one considers labor performed in different workplaces, it is not correct to see
these different workplaces as being integrated through exchange. Rather, ex-
change is the form in which the underlying social integration of different work-
places is expressed.
That exchange is not the real integrating factor or moment in capitalist com-
modity production can be shown by first considering an absurd hypothesis: that
the basis for the integration of different production processes is geographical
proximity, how close they are together. Consider the production of an auto-
mobile. The workers (and the capitalists) involved in automobile production are
directly linked to the production of steel, rubber, plastics, and so on. In some
cases, all of these products are made under the same roof (the famous River
Rogue Plant in Detroit). But the fact that the products are made in different
locations, be it one hundred yards down the assembly line, across the street or far
away, is largely irrelevant for their organizational unity from the point of view of
a capitalist. Under capitalism the separation of the steps in the production process
Is a question of property and ownership. Two plants may be on different conti-
nents, but if they are part of the same ownership unit, then they are directly
integrated. They may lie side-by-side and be “separate,” “independent,” if
owned by different capitals.
The hypothesis that production units are integrated through exchange is equal-
ly absurd. Consider, on the one hand, a corporation that produces automobiles
and also owns a steel plant; on the other, an automobile corporation that must
purchase all of its steel from another, “independent” corporation. In both cases,
the two products, steel and the automobile, are directly linked in a system of
special reproduction. In the second case, separate ownership of the two produc-
tion facilities (property relations of capital) creates the impression that these
facilities are “independent.” Exchange is merely the form that the distribution of
these products must take in a commodity-producing (capitalist) society. In cap-
italist society exchange is the general form by which production units are socially
integrated, but not the only form, since no commodity circulation as such occurs
within the same ownership unit (though monetary bookkeeping entries are
involved).
10 JOHN WEEKS

In effect, the argument that exchange is the mode of social integration is the
consequence of viewing exchange as an isolated act among producers (com-
petitors) of the same commodity, an analysis of what neoclassical economists call
the “product market.” Stressing the problem of “producers” selling their com-
modity in a market with other competing ‘producers,” to the exclusion of
considering the production links among “producers,” is to treat capitalist com-
modity production as the simple circulation of commodities. Indeed, the very use
of the term “producers” carries with it the obfuscations of the fetishism of
commodities. For we are not in the contemporary world dealing with “indepen-
dent commodity producers” (Himmelweit and Mohun 1981, p. 233), not even
with capitalist producers, but with capitalist nonproducers, who appropriate the
product of the working class, then must circulate it as a commodity. Further,
exchange is only one historically specific manner in which this social labor can
express itself. The same system of production could function in the absence of
exchange (e.g., in socialist society).
Let me be clear about what has been established. Production occurs collec-
tively in capitalist society; this is not an assumption or an interpretation, but a
statement of fact. Different workplaces are interrelated through inputs and out-
puts, which derive from the nature of the products, not from exchange. Thus,
exchange expresses the underlying social nature of labor and relates capitalists
one to the other as the personification of the capital relation.
With this in mind, we can consider the question of exchange itself and the
problem of the validation of labor-in-production. Implicit in the formulation of
Himmelweit and Mohun, and explicit in the work of DeVroey, is the argument
that labor-in-production must be validated in exchange. That is, when the cap-
italist organizes production and oversees the execution of the work process, the
value of a commodity is not determined; rather, one must await the moment of
exchange in order for the labor-in-production to be endorsed by the market as
social. Let us accept this formulation of the problem, though the validation
argument presents a stylized view of the commodity seller suspiciously akin to
the neoclassical concept of a “price-taker.” That is, it implicitly characterizes the
capitalist as facing a market situation over which he/she has no control, and
whose decision must be accepted passively.
It is certainly correct that value implies validation, but this does not in turn
imply that value determination is subsequent to exchange. For when a capitalist
comes to sell his or her commodity, the labor which went to produce it has
already been validated by exchange. This point, which I have made elsewhere
and has been explained concisely by Shaikh (Shaikh 1985; Weeks 1981, Chap.
2), can be demonstrated by first considering a hypothetical case of simple com-
modity production in which the independent producer purchases none of his or
her inputs. In this case, when the producer comes to sell the product, he or she has no
idea what its value will be (except on the basis of past knowledge of the
Abstract Labor and Commodity Production 11

market for similar products). This is because the elements of production have not
had to stand the test of exchange, but entered the production process as use
values, not as commodities.
The situation is entirely different for the capitalist. In this case, all of the
ingredients of production (except those provided directly by nature and the
working time of the capitalist) have entered the production process as com-
modities, as values. The ingredients of production represent validated labor,
abstract labor. For a capitalist producing automobiles, the steel in the auto-
mobile is validated as abstract, socially necessary labor when bought, as occurs
in the case of the automobile itself when it is sold. The sale of the product
represents the attempt by the capitalist to recapture value, not establish value.
Whether the capitalist will be successful in doing so takes us into the analysis of
aggregate demand failure, which lie beyond the scope of this note (see Weeks
1982, 1983).
I have criticized the formulation of abstract labor as presented in the validation
argument, but this criticism should not be interpreted as general disagreement.
Indeed, my overall interpretation of the law of value is quite close to that of
Himmelweit and Mohun, so the commentary is intended to be a constructive one.
I conclude this part by suggesting the origin of the difference between my
approach and the validation argument. The problem of the validation approach is
that it incorrectly distinguishes the twofold nature of commodities. Commodities
are, on the one hand, objects of use; on the other, objects of exchange. This
implies a twofold character for labor: that it is both use value producing (concrete
labor), and value producing (abstract labor). Although this distinction is certainly
noted, and even stressed in Himmelweit and Mohun, it quickly becomes sub-
sumed under a distinction between private and social labor, with the consequence
that the more fundamental distinction becomes lost.
The concrete versus abstract dichotomy is the basic one in capitalist produc-
tion, while private versus social is a distinction in the sphere of exchange and
distibution. Validation refers to the conversion of concrete labor into abstract
labor, which involves not just the sale of the commodity, but also the capital
advanced for the ingredients of production. The private versus social distinction
refers to the contradictory appearance of collective labor as labor-in-production,
on the one hand, and as objectified in commodities, on the other hand, with the
latter being the vehicle for private appropriation.
The consequence of the validation approach is to be left with no value theory
at all. Price is value, and therefore price is indeterminate, set by conditions of
demand. This is an unsatisfactory conclusion, for it abandons any serious attempt
to account for the quantitative aspect of exchange. Two routes are open to the
validation theorists: either they can discard the law of value as a dead end (the
choice of DeVroey), or they can turn to general equilibrium theory as the Uno
school as done.
12 JOHN WEEKS

III. VALUE AND GENERAL EQUILIBRIUM

The Uno school of Marxism seeks to resolve the difficulty of the validation
theorists in the tradition of Walras. I take as representative of the school the
approach of Sekine, for he is quite clear in his use of equilibrium analysis. For
the basis of my critical remarks to be clear, I begin with what I hope is a faithful
rendering of his analysis, which he presented in a 1984–1985 article in the form
of a rhetorical “seminar.” After defining value (question 1, “what is value . . ?”
and introducing the “value form” (question 2, “what does the value-form theory
do?”), he turns to the issue of money and the measure of value (‘what is the
measure-of-value function of money . . ?”—question 3) (Sekine 1984–1985, pp.
419–422). In this discussion we are told the sense in which values determine
prices, and it is necessary to follow the presentation in some detail. He assumes
at the outset that there is something called “the market,” and producers arrive at
this market with an endowment of commodities.

All commodities appear in the market with supply prices which their sellers (owners) quote
tentatively in expressing their values. The owners of money, representing the force of demand
in the market, buy immediately if the price is low and temporize if the prier is high. This
compels the sellers to increasingly adjust [sic] their supply prices to demand prices . . . The
price at which the commodity is sold varies . . . but eventually settles to a normal level . . .
What I call a normal price is, in fact, the same thing as an equilibrium price (Sekine 1984–
1985, pp. 422–423).

We are, however, cautioned that the process described does not imply “that
commodity values are proportional to normal prices.” 5 Here in a formal state-
ment is the type of exchange behavior that the validation theorists have in mind,
individual commodity producers/sellers meeting in a market, seeking to discover
if what they have produced will be validated in social commerce. The quotation
from Sekine involves a model of economic behavior that has the following
characteristics:

1. Sellers come to the market with a given endowment of commodities. This

must be the case, or the subsequent trading prices are indeterminate,
because the “normal price” is derivative from particular quantities offered
by sellers. Because quantities supplied are set, we are dealing with a
“market period” in which no production occurs.
2. All potential buyers and sellers are in the market (no exchanges take place
outside of it). This also must be the case if the process is to produce a
determinate price.
3. Buyers and sellers respond to excess demand (price of a commodity “too
low”) or excess supply (commodity price “too high”) by adjusting price,
“groping” for a market-clearing set of relative money prices.
Abstract Labor and Commodity Production 13

4. The outcome of this “groping” is a set of equilibrium prices, for which all
buyers and all sellers are satisfied, in the sense that when the “normal
price” is reached no one wishes to recontract.

The model of economic behavior that Sekine has described is well-known in

neoclassical economic theory: Walrasian general equilibrium analysis. Also well-
known are the conditions under which this model will generate an equilibrium
price vector (“normal prices”). And these conditions are so restrictive that an
entire school of mainstream economics had abandoned the model in the 1960s
and early 1970s,6 though the right-ward shift of the profession brought it re-
newed political (but not intellectual) respectability.
The determining characteristic of this model is that it involves a multicom-
modity system and reaches an equilibrium state in which there are no excess
demands or excess supplies (commodities are not “overproduced” or “under-
produced”). It is beyond the scope of my discussion to do justice to the the-
oretical critique of the Walrasian system, but I shall indicate the most telling
points. Of the many critiques of the model, Clower’s, made almost 25 years ago,
is the seminal and devastating (Clower 1965). It is to be noted that equilibrium in
the model requires that there be no excess demands or excess supplies (“Walras’
Law”). However, as Clower shows, the demands and supplies that must be
equated are not actual (often called “effective”) ones, but notional demands and
supplies, that is, the ex ante plans to buy and sell by the participants in the
market. Only when notional demands and supplies are matched is the market in
equilibrium.
This has a profound and devastating implication for the model: equilibrium
cannot be achieved if any trades occur at nonequilibrium prices (see Lei-
jonhufvud 1968, chap. 2). Walras himself recognized this basic contradiction in
his model, and offered as a solution what must rank as one of the more charming
and ludicrous deux ex machina in the history or economic thought. To avoid any
disequilibrium trading, this hypothetical market for all commodities is provided
with an “auctioneer,” whose task it is to call out various vectors of relative
prices, then adjust them if supplies do not match demands. As a part of this
process, the auctioneer draconianly enforces the rule that no trades can be made
until all commodity prices are in equilibrium (Harris 1981, pp. 271–272). If
disequilibrium trading occurs (called “false trading”), the result is by definition
“quantity responses” (buyers and sellers adjusting their notional demands to
effective demands), rather than “pure” price responses (adjusting prices accord-
ing to notional supply and demand). The consequence of disequilibrium trading
is cumulative movement away from the general equilibrium price vector, render-
ing the clearing of markets impossible.7
Furthermore, “false trading” only begins a long list of contradictions within
the Walrasian model. To mention two others, it appears that only under the most
restrictive assumptions is Walras’ Law (all excess demands and supplies zero)
14 JOHN WEEKS

consistent with money exchanges. Indeed, only with great agility can it be shown
that money plays any role in the model (Harris 1981, pp. 289ff). Finally, when
the system is not in equilibrium, buyers and sellers no longer take prices as
given, since they must adjust their offer prices to meet the excess demands, and
are no longer price-takers. But at the same time these prices must confront
traders as “parameters” in order that there be no quantity adjustments (which
have the consequences mentioned above).8 It is this apparent contradiction that is
resolved by the mythical auctioneer.
All of these criticisms are well-known by neoclassical economists—and well-
recognized by the practitioners of Walrasian models. The best of these econo-
mists, Hahn (1981), for example, concede that the model is only consistent when
in equilibrium, and provides no vehicle for the analysis of adjustment processes
when markets are out of equilibrium. Sekine (1984–1985) seems unconcerned
with this extremely restrictive nature of the model of value determination which
he employs. Further, his formulation of it is logically inconsistent. Because he
permits disequilibrium trading (“the price at which the commodity is sold varies
from time to time”), his own postulated process does not converge on equi-
librium (“settle to a normal level”).
Even should Sekine reformulate his model so it is consistent, the law of value
is indeed in great theoretical difficulty if it must be developed within a Walrasian
general equilibrium framework. This would represent a major step backward,
embracing a model that the best of the neoclassical economists consider applica-
ble to a very narrow set of problems. How far back we would be taken can be
shown by indicating the extremely restrictive nature of value (“socially neces-
sary labor”) in his model. In his framework. “socially necessary labor” does not
refer to the labor time necessary to produce a unit of a commodity under average
production conditions (Marx’s definition), but rather to the total labor time neces-
sary to produce the entire supply of a commodity when the markets for all commodi-
ties are in long run equilibrium.9 Such a definition of value is indis-
tinguishable from the neoclassical concept of long-run supply prices in perfectly
competitive equilibrium, a situation in which even neoclassicals would agree that
the cost of producing commodities can be reduced to their labor content. At
most, the labor theory of value becomes a subset of neoclassical value theory, the
special case of fixed coefficients of production.
We can now develop further the close link between the views of the validation
school and those of Sekine’s rendition of the Uno interpretation of the labor
theory of value. Assume, for example, that apples sell for one dollar in the
morning and ten dollars in the afternoon. Following the validation argument, we
would have to say that in each exchange the value of apples is expressed with
equal validity. To escape the tautology of appearances, one must introduce some
concept of “normal” price (or perhaps “value-price”), around which observed
and occasional fluctuations move. Because the validation argument has rejected
the sphere of production as the “critical moment,” it is left with Sekine’s Walra-
Abstract Labor and Commodity Production 15

sian general equilibrium as a haunting familiar. DeVroey recognizes this. Hold-

ing to the validation argument, but rejecting the world of general equilibrium as a
relevant framework in which to analyze capitalism, he abandons value all to-
gether.10 One might say that by their stress upon the theory of the form of value,
the validation writers have lost any analysis of the essence of value and thrown
themselves into nihilism or the arms of the general equilibrium theory.

IV. ALTERNATIVE TO EQUILIBRIUM ANALYSIS

The concept of equilibrium is an essential ingredient in neoclassical economic

theory. Presentations of Marx’s analysis as if he, too, were an equilibrium theo-
rist are quite common, not limited to Sekine’s rendering of the Uno School. I
conclude by treating this question, in particular whether capitalism is correctly
treated as a system tending toward equilibrium. Prior to doing so, I must clarify
several points.
First, capitalist society tends to reproduce itself; that is, it does not continu-
ously demonstrate a tendency toward systemic collapse. On the contrary, violent
fluctuations are relatively rare. For the most part, the production and circulation
of commodities result in the gainful employment of the vast majority of the
laboring population, the survival and even expansion of most capitalist enter-
prises, and most important, the reproduction of the class relations of capital. On
the other hand, capitalist society is subject from time to time to extreme in-
stability, which takes the form of high levels of unemployment, accumulations of
unsold commodities, and inflation in the exchange value of money. One’s analy-
sis must account for both the relative stability and the manifestations of
instability.
Explaining the first, stability, does not imply equilibrium analysis, for equi-
librium is a construction of the mind, while the relative stability of capitalist
reproduction is an aspect of reality that cannot be denied. To an extent, one can
account for the relative stability with neither a theory of the value of commodities
nor a theory of accumulation. First, there is a tendency for the production of
commodities to adjust to their sales: if inventories build up excessively, produc-
tion is reduced; if unfilled orders accumulate, production expands. Capitalists
who do not react in this manner will not continue to produce for long.
However, this tendency for production to adjust to sales is not the same as
market equilibrium, or even the same as market clearing. It is a generalization
based on what actually occurs—capitalists do not tend to expand production when
sales are falling. When we look at the system as a whole, we see that the
tendency to adjust does for the most part result in a rough stability (for example,
between production and inventories).
On closer inspection, this real tendency for production to adjust to sales proves
to be quite complicated. One complication is to account for how the actions of
16 JOHN WEEKS

individual capitalists relate to the reaction of the aggregate. It is not valid to

presume that the sum of individual reactions will be stabilizing. A second com-
plication involves sales, for the actual sales realized are not necessarily the only
information that the capitalists have upon which to base their expectations of the
future. At this point one enters the realm of theory, in which terms such as
“sales” take on a meaning other than a strictly empirical one.
Neoclassical equilibrium analysis involves specifying behavioral relationships
such that the tendency for production to adjust to sales can be interpreted as a
mechanism which tends to bring the system to harmonious rest. That is, it
interprets relative stability (absence of imminent collapse) as implying general
and harmonious interaction, and that the system tends quickly and smoothly to
correct any imbalances. This interpretation is achieved by creating certain ideal
constructions, such as the “market” for a commodity. The so-called “theory of
rational expectations” is the most extreme and ideological form of this
metaphysics.
In reality, commodities do not exchange in “markets,” be they neoclassical or
Marxist ones, but in a multitude of transactions separated in time and space. The
concept of “market” for a commodity involves collapsing actual exchanges in
time and space, so one can proceed with the fiction that buyers and sellers
confront each other directly, all at the same moment, all receiving the same
“signals.” Such an arrangement is a purely ideal construction, having no relation
to the actual process of buying and selling.11
In the absence of this fanciful market, the concept of equilibrium is dubious,
since transactions are an on-going process that is never at rest. This is the way in
which transactions are created in Marx’s analysis. In the first chapter of the first
volume of Capital he stresses that sale and payment are separate acts when
exchanges involve credit, a real distinction, not a creation of the mind. This
approach involves the use of money as means of payment as well as means of-
circulation, and places exchanges in chronological time. In Marxian analysis
there are no demand or supply schedules, no short-run equilibrium prices, and no
market clearing, because there are no fictitious markets that are the basis of these
concepts, and notional demands and supplies play no role. Certainly, there is
analytical simplification (abstraction) in Marx’s treatment of exchange. How-
ever, the simplification is achieved through analyzing the exchange of com-
modities within the circuit of capital (as commodity-capital), not by presuming
them to exchange in a hypothetical, timeless setting (which they do not).12
Having considered capitalists’ adjustment to demand fluctuations, let us turn to
another immanent tendency, the movement of capital in response to difference in
rates of return. This tendency is not assumed, but exists, as a reflection of the
development of capitalist institutions. In contemporary capitalist society, large
corporations are basically financial entities, and the material nature of the prod-
uct in which they invest is a matter of indifference.
Again, this actual tendency takes on a fantasy life in neoclassical theory. The
Abstract Labor and Commodity Production 17

movement of capital is treated in the absence of technical change, so that invest-

ments involve entry into an industry with existing techniques in conformity with
the assumption of the “representative firm.” The logical consequence of this,
when combined with appropriate assumptions about “long-run cost curves,” is
the expansion of output in a context of harmonious interaction of capitals. Some-
thing of the sort is implicitly assumed by Sekine, or he could not reach a general
equilibrium solution in his model.
Marxian analysis could not be more different. Each industry is the result of
historical development, so that in each there are capitals utilizing fixed means of
production of different vintage, involving different unit costs for the commodity
they produce. This implies that the capitals in an industry do not all earn the same
rate of profit, which, of course, is the case. The movement of capital in response
to differential profit rates is typically executed by investing in the technique that
provides the invading capital with the strongest competitive position. In conse-
quence, the movement of capital in response to differential profit rates is disrup-
tive to the system. It is the carrier of technological change, which devalues the
existing means of production in an industry and intensifies the competitive strug-
gle. What serves as an equilibrium mechanism in neoclassical theory is a de-
stabilizing process in Marxian analysis. It is the introduction of technical change
into the analysis that eliminates any concept of equilibrium. At its basis, equi-
librium is a concept relevant only to exchange divorced from production. Once
one enters the sphere of production, the concept becomes counterproductive to
thought.
The above discussion brings us to the question of competition. Here the
contrast between ideal states and real processes is quite clear. With regard to any
one commodity, the process of competition involves the struggle over shares of
the market, and this struggle is implemented in a number of ways, one of which
is price competition. This is precisely what Marxian analysis is able to consider,
because its theoretical treatment of an industry incorporates the uneven develop-
ment of capital. In contrast, neoclassical analysis of competition defines com-
petition in order to eliminate the possibility of conflict. A “perfectly com-
petitive” industry is one in which all firms are price takers and none has
sufficient installed capacity to expand and affect price and, therefore, affect the
market share of any other firm. In other words, the competitive struggle has been
assumed away—by assumption there is no price competition and by assumption
there can be no struggle over market shares (for an excellent discussion, see
Shaikh 1985). The result is a harmonious equilibrium, not because the industry is
competitive, but because competition is excluded.
In summary, one can say that in Marx’s analysis the possibility of the system
being “at rest” is precluded, because the tendencies for production to adjust to
sales and capital to move in response to differential profit rates do not have the
result of eliminating the cause of those tendencies. They are not self-correcting
mechanisms. Particularly in the case of the movement of capital among indus-
18 JOHN WEEKS

tries, the result is to reproduce the cause of the tendency. Thus, the movement of
capital is an adjustment process that does not result in correction of the initial
impulse to adjustment; on the contrary, it reproduces the impulse.
We have seen that the validation argument leads one down a road away from
the sphere of production, to a point where value merges with price, and the
escape from this tautology of exchange seems to be equilibrium theory. It may or
may not be that the neo-Ricardians have some validity in their argument that
values are redundant. But their arguments will not be countered by a treatment of
value in which exchange is the “critical moment.” Above all, the law of value
must derive from the insight that capitalism is primarily a commodity producing
society, and only secondarily a commodity exchanging society. Fruitful advances
in value analysis will be made by formulating the issues in terms of the circuit of
capital, not simple commodity circulation.

NOTES

1. DeVroey also makes the validation argument:

The main feature of the commodity system is that here social labor is formed in an indirect
way. Labor is first performed as private labor . . . It is transformed into social labor . . . only
through the sale of its product (DeVroey 1981, p. 176).

2. DeVroey (1982, p. 184), makes the same point with clarity, “this [validation argument]
implies that abstract tabor is logically posterior to exchange.”
3. The issue of the theoretical and historical relevance of petty (or simple) commodity production
is a complex one which lies outside of the scope of this paper. I wish to thank David Laibman for
comments which led me to correct and temper my reference to petty commodity production.
4. Marx himself explicitly recognized the mistake involved in the private/social dichotomy. In
Chapter I of Volume I of Capital he abandoned his earlier usage of this dichotomy in favor of the
concrete/abstract dichotomy. His earlier approach is found in a manuscript entitled “The Com-
modity” (reprinted in Dragstedt 1976), where the editor does not comment on the contradiction of
this approach with that in Capital.
5. Because prices are measured in units of money (“the measurement of value thus occurs when
physical money acts as means of purchase.” [Sekine 1984–1985. pp. 422–423]), and values are
measured in some other, but as yet unspecified units (unless price is synonymous with value), it is
unclear what “proportional” might mean in this context. Some version of general (abstract) labor is
being implicitly assumed here in the third question.
6. This is the “reappraisal” school, whose members are sometimes called “post-Keynesians” or
the “new Keynesians.” For an excellent discussion of Walrasian general equilibrium theory and its
critics, see the survey of bourgeois monetary theory by the British Marxist Laurence Harris (1981,
chaps 4, 13).
7. These cumulative effects involve the famous “multiplier.” See Leijonhufvud (1968. Chap. 2)
and Harris (1981, Chap. 13).
8. The contradiction in the model between agents being active price-setters in disequilibrium and
passive price-takers in equilibrium is explained well in (Leijonhufvud 1981, chap. 4). The discussion
in this and previous paragraphs lends me to disagree with Mohun when he writes, “while neoclassical
theory . . . is certainly consistent, it is also the case that its consistency is achieved at the expense of
Abstract Labor and Commodity Production 19

a notorious air of unreality” (Mohun 1984–1985, p. 395). Neoclassical self-criticism has shown that
the theory, of both microeconomic and macroeconomic general equilibrium, is both unrealistic and
inconsistent.
9. Sekine states this quite clearly:

When commodities are produced in socially necessary (equilibrium) quantities, i.e., are
neither overproduced nor underproduced, we say that “socially necessary labor” has been
spent for their production. Hence, a precise statement of the labor theory of value is that
socially necessary labor forms the substance of commodity values (1984–1985, p. 425).

10. This is to be found in DeVroey’s recent writings in French, but his concern that Marx might
have been an equilibrium theorist appears in Capital and Class (1982). The validation argument has
been a strong current in French political economy in recent years, and led to a mass-migration away
from value theory, with a few notable exceptions, such as Suzanne DeBrunhoff.
11. Harris (1981, p. 271) has an excellent discussion of the absence of time in Walrasian
markets.
12. Marx used the term “market price,” of course, but this no more implies for him an idealistic
concept of markets than neoclassical theory’s use of the word “value” implies the labor theory of
value.

REFERENCES

Bandyopadhyay, P. 1984–1985. “Value and Post-Sraffa Marxian Analysis.” Science and Sociology
48 (4).
Clower, W. 1965. “The Keynesian Counterrevolution: A Theoretical Appraisal.” In The Theory of
Interest Rates, edited by F. H. Hahn and F. P. R. Breching. New York: St. Martin’s Press
DeVroey, M. 1981. “Value, Production, and Exchange.” In The Value Controversy, edited I.
Steedman and P. Sweezy. London: New Left Books.
_______. 1982. “On the Obsolescence of the Marxian Theory of Value.” Capital and Class
17 (July),
Dragstedt, A., ed. 1976. Value Studies by Marx. London: New Park.
Gleicher, D. 1985. “Note: A Rejoinder to Eldred.” Capital and Class.
Hahn, F. 1981. Equilibrium and Macroeconomics. Oxford: Basil Blackwell.
Harris, L. 1981. Monetary Theory. New York: McGraw-Hill.
Himmelweit, S. and S. Mohun. 1981. “Real Abstractions and Anomalous Assumptions.” In The
Value Controversy, edited by I. Steedman and P. Sweezy. London: New Left Books.
Leijonhufvud, A. 1968. On Keynesian Economics and the Economics of Keynes. New York Oxford
University Press.
_______. 1981. Information and Coordination: Essays in Macroeconomic Theory. New York: Ox-
ford University Press.
Levine, D. 1981. Economic Theory. Boston: Routledge & Kegan Paul.
Mohun, S. 1984–1985. “Abstract Labor and Its Value Form.” Science and Society 48 (4).
Sekine, T. T. 1984–1985. “An Uno School Seminar on the Theory of Value.” Science and Society
48 (4): 419-32.
Shaikh, A. 1982, “Neo-Ricardian Economics” Review of Radical Political Economics 14 (2).
Weeks, J. 1981. Capital and Exploitation. Princeton, NJ: Princeton University Press.
_______. 1981. “A Note on the Underconsumptionist Theory and the Labor Theory of Value.”
Science and Society 46 (1).
_______. 1983. “On the Issue of Capitalist Circulation and the Concepts Appropriate to Its Analy-
sis.” Science and Society 47 (2).