JONATHAN WOLFF

MARX AND EXPLOITATION1 (Received 11 January 1999; accepted 30 April 1999)

ABSTRACT. The discussion of the adequacy of Karl Marx’s definition of exploitation has paid insufficient attention to a prior question: what is a definition? Once we understand Marx as offering a “reference-fixing definition in a model” we will realise that it is resistant to certain objections. A more general analysis of exploitation is offered here and it is suggested that Marx’s own definition is a particular instance of the general analysis which makes a number of controversial moral assumptions. KEY WORDS: coercion, definition, exploitation, exploitation of circumstances, Marx, vulnerability

Many people have commented on Karl Marx’s supposed definition of exploitation. It has been questioned whether he intended to make out any type of moral use of the concept of exploitation; it has been argued that he did but his definition fails; and, more rarely, some have argued that he does – broadly speaking – give a correct moral definition of exploitation.2 To my knowledge there has been very little attention to what may be a prior question: what is a definition? My case here is that when we pay more attention to different possible types of definition, and when we come to a fully general understanding of exploitation, we will come to understand that Marx’s definition of exploitation is resistant to many standard objections, given its limitations. It is limited to relations of economic exchange under conditions where there is no justified right to capital; that is, to earn money purely in virtue of one’s property holdings.

1 Much of this paper derives from my 1985 University of London MPhil thesis:

Exploitation. I should repeat here the thanks given there to G.A. Cohen, W.D. Hart and Chris Hull. An earlier version of this paper was delivered at the University of Virginia and I would like to thank the participants on that occasion for valuable discussion, and especially Talbot Brewer for his detailed and extremely helpful written and oral commentary. 2 Some of the best recent work has been collected in Kai Nielsen and Robert Ware (ed.), Exploitation (New Jersey: Humanities Press, 1997) and in Andrew Reeve (ed.), Modern Theories of Exploitation (London: Sage, 1987). In addition to the papers collected in these volumes, Richard Arneson, “What’s Wrong with Exploitation?” Ethics 91 (1981), pp. 207– 227 and Allen Buchanan, “Exploitation, Alienation and Injustice,” Canadian Journal of Philosophy 9 (1979), pp. 121–139 have helped shape my thought on this topic. The Journal of Ethics 3: 105–120, 1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands.

106

JONATHAN WOLFF

Section 1 looks at attempts to give a specification of Marx’s definition of exploitation, while in Section 2 issues of definition in general are addressed. Section 3 proposes a general analysis of exploitation and Section 4 draws the earlier discussions together in explaining the conditions under which Marx’s definition should operate.

1. M ARX ’ S D EFINITION

OF

E XPLOITATION

Traditionally Marx’s definition of exploitation is given in terms of the theory of surplus value, which in turn is taken to depend on the labour theory of value: the theory that the value of any commodity is proportional to the amount of “socially necessary” labour embodied in it. Yet the labour theory of value is generally now thought to be false. From the point of view of creation of value, there is nothing special about labour.3 Some, notably Robert Nozick, argue that with the collapse of the labour theory of value, the Marxist theory of exploitation collapses too.4 Others have felt that more argument than this is necessary, and that it is possible to restore the intuitive core of the theory independent of the labour theory of value.5 John Roemer, to take one leading case, states: “Marxian exploitation is defined as the unequal exchange of labour for goods: the exchange is unequal when the amount of labour embodied in the goods which the worker can purchase with his income . . . is less than the amount of labour he expended to earn that income.”6 Here the intuitive thought is this. Suppose I work eight hours to earn my wages. With this perhaps the best thing I can buy is a coat. But imagine that the coat took only a total of four hours to make. Therefore I have exchanged my eight hours work for only four hours of other people’s work. Therefore, on this view, I am exploited. The definition, no doubt, requires some refinement. One thing about it is that it is not relational: whether one is exploited depends on facts about the economy as a whole, rather than one’s particular contract of employment. Roemer seems to present this as an advantage – or at least as a positive feature – of this account of exploitation, although intuitively it may seem 3 John Roemer, “Should Marxists Be Interested in Exploitation?” in Nielsen and Ware

(ed.), Exploitation, p. 127. 4 Robert Nozick, Anarchy, State, and Utopia (Oxford: Basil Blackwell, 1974) pp. 253– 262. 5 G.A. Cohen argues that the relation between the labour theory of value and the concept of exploitation is one of “mutual irrelevance,” G.A. Cohen, “The Labour Theory of Value and the Concept of Exploitation,” in Nielsen and Ware (ed.) Exploitation, p. 94. 6 Roemer, “Should Marxists Be Interested in Exploitation?,” p. 122.

MARX AND EXPLOITATION

107

unmotivated. We will come back to this. Further, if I am taxed for the benefit of those unable to work, I will be exploited by the above definition, but this doesn’t seem to be what the definition of exploitation was intended to capture. Worse still, if there is one person exploited much more gravely than anyone else in the economy, then it may turn out that no-one else is exploited. Suppose someone is paid very little to produce hideous tourist novelties which are sold cheaply. Imagine this person works eight hours to produce one such novelty, and I, the second worst paid person in the economy, can purchase it with my daily wage, and still have some money left over. In such a case, as I can command more labour than I expended it would seem that I am not exploited, even if, in fact, all the things I actually want to buy contain less labour than I expended. It is not difficult to see how the definition should be altered to take account of this factor. Perhaps we need to say that what matters is not what you can purchase, but what you do purchase. Or better, to talk about the average labour content of the things one can buy. Roemer prefers to talk of a “grey area” of people who are neither exploited or exploiters, which, assuming a global economy, would include almost all the working class in the developed world, who can command a great deal of labour of those in the third world. This has its theoretical advantages too. The case of taxes for welfare benefits needs to be handled in a different way: the simplest proposal would be to say that everyone should be taken as benefiting indirectly from the welfare state, and thus a certain amount of labour goes to pay one’s “insurance premium.” However we need to be careful to ensure that it is not true by definition that welfare redistribution is never exploitative, for perhaps it might be. But we will not let this detain us here. The general point to focus on is that Roemer’s Marxian definition sees exploitation as involving unequal exchange of labour and goods. The exploited person just never gets back what he or she puts into the labour process. Now there may be, in particular cases, a great deal to be said about why this is perfectly acceptable from a moral point of view. However it is interesting that in the absence of a further story there seems something prima facie wrong about such a situation. It requires justification, and without justification it may seem proper to talk of exploitation. But is it? One nagging doubt is that, somehow, there just seems to be more to exploitation than unequal exchange. Even if it is true that unequal exchange is often exploitative, it may be that there are cases of exploitation which do not involve unequal exchange, or, even, that this is not the worst thing about cases that do. We will return to these questions.

108

JONATHAN WOLFF

2. W HAT I S

A

D EFINITION ?

Now we have a sketch of a definition of Marxian exploitation, although we have accepted that it needs further refinement, and even then may seem in some ways unsatisfying. But we must attend to a prior question before returning to the main theme: what is a definition? A dictionary is full of them. In general it is an attempt to convey the sense of a word, using different words. The idea of a dictionary definition is to be able to explain to someone unacquainted with the word how it is used and how it is to be understood. Dictionary definitions standardly have two related limitations. First, they generally only give you enough information to “get started” with the word; one has to see it in use, in a variety of contexts, before one fully gets the hang of it. Second, a dictionary does not provide a philosophical analysis of a concept, in the sense of specifying a complete set of necessary and sufficient conditions for the application of the term. It could be, of course, that dictionaries are limited in this way because no more can be done. Why, after all, should a language contain a word if its entire sense can be conveyed by the use of other words? Of course we might introduce a term as a deliberate abbreviation for a longer statement – “mother-in-law” is a standard example – in which case the analysis is trivial, at least at first (but do unmarried long-standing partners, or gay couples, have mothers-in-law?). But whether a complete definition of any “natural” term can be given is an experimental question, and so far the experiment seems to have failed in almost every, if not every, instance. Note, though, that I said that a dictionary tells us how a word is to be used and how it is to be understood. We owe to Saul Kripke the clarification of the thought that these can be different enterprises.7 To give a complete explanation of how a term is to be used is to fix the reference: it is to give an infallible way of picking out that object or those objects to which the term refers. In itself doing this may convey little understanding, or only partial understanding. Another task is to give, or we might better say explain, the meaning of the term. To give an example, a mathematician might define an infinite set as a set which has at least one proper subset with as many members as the set itself. A proper subset of a set is a subset which leaves out at least one member of the initial set. So the set of even numbers is a proper subset of the set of natural numbers. Yet the set of even numbers and the set of natural numbers can be placed in one-one correspondence, and this is taken to show that they are the same size. This apparently paradoxical feature characterises all and only infinite sets. The set of even numbers is a proper subset of the set of natural numbers, yet 7 Saul Kripke, Naming and Necessity (Oxford: Basil Blackwell, 1980), pp. 53–58.

MARX AND EXPLOITATION

109

there are as many even numbers as natural numbers. Therefore the set of natural numbers is infinite. By using the criterion I can now with confidence pick out all and only infinite sets. Yet it would be hard to make out the case that, in knowing this, I now know the meaning of the words “infinite set,” or understand the nature of the infinite. Rather I know a property which is had by all infinite sets and only by infinite sets. Another example: I can fix the reference of the number 13 by saying that it is the second smallest prime number larger than 10. In using this account I will never go wrong in attempting to identify the number 13. But it is hard to believe that, if this is all I know, I now know what “13” means, as distinct from what it refers to. So we have a general distinction between fixing the reference of a term and the rather more mysterious idea of giving the meaning of a term. Yet we need a further distinction. Sometimes a mathematician, an economist, or scientist will set up a simplified model of reality and, within that, model, provide a reference-fixing definition of a term. Trying to apply the definition outside the model may lead us into error, as the terms and conditions of the model implicitly contribute to the definition, and so transplanted into another context the implicit content of the definition is lost. Here is a very simple example. In explaining to children how to detect even numbers we tell them that, however large the number, if (and only if) it ends in an even numeral it is even. Thus we have given a way of fixing the reference of “even number.” However this reference-fixing makes one big assumption: that we are operating in base 10 (or at least, an even base). Were we to be operating in base 9, then this definition would not work. In base 9, the numerals 12 refer to our base 10 number eleven. So we can say that the definition offered fixes the reference of the even numbers in the model of base 10, but we have to be very careful in attempting to apply it elsewhere. Much methodology in economics consists of first setting out simplified models, which may or may not later be complicated by bringing in more “real-world” factors. This is often known as the method of abstraction, and was Marx’s method in Capital. The reader should now be able to see what is coming: I will claim that the best way of reading Roemer’s reformulation of the Marxian definition of exploitation is to view it not as a way of “giving the meaning” of exploitation, or of giving a fully general referencefixing definition, but as a way of fixing the reference within a model. This being so, some counter-examples to the definition of exploitation will be no more relevant to the case in hand than counter-examples provided by

110

JONATHAN WOLFF

the smart kid who points out that our standard way of determining whether a number is even doesn’t work in base 9.8

3. T HE A NALYSIS

OF

E XPLOITATION

What is it to exploit someone? Here I shall outline a fully general analysis which applies to both economic and non-economic contexts. There can be no doubt that the idea of exploitation is closely related to some notion of use, and so, as a start, exploiting someone is using someone. Generally when we say “x used y,” where y is a person, we intend this to be a form of criticism of x’s behaviour, although people are quick to point out that this is not invariably the case. We say that directors use actors, and we don’t (always) mean this as a criticism. The term “exploit” has a similar range: in general to say that x exploited y is a criticism of x, but we may be able to think of exceptional contexts where it isn’t. However “exploits” as a moral criticism seems very close to “improperly uses,” although there are some improper uses of people – for example as a punchbag – which would not normally be thought of as forms of exploitation. We also have to be careful to note that we use the term “exploits” in inter-personal contexts in more than one way. For example a teenage child may say that her parents exploit her dependence on them to enforce a strict code of conduct. But let us suppose that they do this out of genuine concern for her, and believe that in the long term she will come to appreciate their efforts on her behalf. Now we might still be unhappy about their treatment of her; it would be much better if they used reason rather than power. But exploiting their power over her alone does not seem sufficient to lead to a justified charge of exploitation of her. They exploit her circumstances, but not her. Suppose, however, they derive an economic benefit from this treatment: perhaps they are trying to curry favour with a religious, elderly, wealthy relative, and they enforce a strict code on the daughter purely as a means of impressing the relative, as a way of securing a good slice of the estate. Here, I think, we would have much less doubt that they are exploiting her, in addition to exploiting their power over her.9 8 Indeed I shall suggest later that Roemer himself goes wrong in failing to understand the nature of the definition he offers. 9 Note that if we accept this example then exploitation is not always a matter of goods flowing from one person to another. For in this case one person’s harm – following “rigid rules” – does not directly correspond to another’s benefit, for the exploiter’s economic gain comes from an outside source.

MARX AND EXPLOITATION

111

Accordingly I believe that we must make a distinction between exploiting someone’s circumstances – that is, exploiting the power one has over another – and exploiting that person. The notion of having exploitable circumstances seems equivalent to that of being vulnerable, but it is easier to give a list of causes of vulnerability than to explain its nature. Typically you are vulnerable if (other things being equal) you are poorer, more ignorant, less intelligent, less cunning, or less ruthless than another, or have some other bargaining weakness with respect to them. One’s vulnerability is exploited if the other person uses this weakness to obtain agreement to, or at least acquiescence in, a course of action that one would not have accepted had there not been this asymmetry in power. Exploitation of a person’s circumstances is not sufficient for exploitation of that person, as we have already seen, for there can be paternalistic exploitation of circumstances, and paternalism is not exploitation. Indeed there can be purely malicious exploitation of circumstances: bullying for no reason other the enjoyment of the exercise of one’s own power over another, for example. But is exploitation of circumstances necessary for exploitation proper? Perhaps, although I am unsure. We can say, though, that typically exploitation of the person only happens because that person has exploitable circumstances. But what is it to exploit a person? Sometimes it is said that exploitation is akin – both structurally and morally – to coercion. If this is right it would give insight both into the nature of exploitation and provide one account of what is wrong with it. Or if not an account, a proof that it is wrong, for who would deny that coercion is an evil? There may be, of course, those who deny exactly this, but there is a greater weakness in the argument: there are some important differences between coercion and some cases of exploitation. As we have seen exploitation is typically a matter of using another person’s vulnerability to your own advantage. Coercion, on the other hand, typically proceeds by first creating another’s vulnerability and then exploiting it. It involves the difference, then, between happening upon exploitable circumstances and generating them. Coercion and exploitation turn out to be the same only if there is no moral difference between harming someone (creating a vulnerability) and not helping them out of their vulnerability even at some cost to yourself. This seems a highly implausible view. Nevertheless, although there is an analytic distinction to be drawn here, it may well be that the most interesting economic cases of exploitation turn out to be cases of coercion on this account; or at least they would do if we take a “class” interpretation of exploitation. That is, if workers as a group are in a vulnerable position as a result of the previous and continuing

112

JONATHAN WOLFF

actions of capitalists, then it is arguable that we would have crossed the divide from pure exploitation to what we could call coercive exploitation.10 But we should also keep in mind that not all exploitation is coercion. So what is the nature of exploitation, and where lies the wrong? I suggest that in this case, as in so many, we turn to Kantian moral philosophy, or at least that part of Kantian moral philosophy which has so often been appropriated by political philosophers: Kant’s formulation of the Categorical Imperative in terms of the “End in Itself.” To exploit someone is to treat that person purely as a means to your own ends, and not as an “end in themself.”11 This idea is suggestive, but requires further elaboration. What is it to treat someone as a means only, and not as an end? One may begin with the idea that a tool or an instrument is a type of means, and so one way of treating someone as a means is to treat them as a tool. But what more can we say? Sometimes it is said that to treat something as a means is to treat it without regard to its individuality, and the test would be that one can replace that object with another of the same kind, which performs the same function, without loss. However it seems to be the case that by this test we treat some tools with regard to their individuality. A craftsman might have a favourite chisel, or even a favourite screwdriver, and would be upset to lose it – perhaps they have been through so much together, or it was a present from a now deceased parent – even if it was instantly replaced by another of the same type. In such a case we might be tempted to say that the craftsman is not treating that tool as mere means. The tool is treated as valuable in itself. Yet this does not seem sufficient for it to be considered as being treated as an end in itself. What can we say to go further? The three great traditions of moral philosophy divide at this point. The Kantian tradition will emphasise the idea of will or choice: an end is capable of agreeing or disagreeing with proposed courses of action. The Aristotelian tradition will emphasis the idea that ends have their own good: things can go well or badly for that entity. It can flourish or wither. Finally the utilitarian tradition lights on the idea of pleasure and pain. So the Kantian form of the idea of exploitation can be given Kantian, Aristotelian or Utilitarian content.12 10 I thank Talbot Brewer for insisting on the importance of this. 11 Immanuel Kant, Groundwork of the Metaphysic of Morals, in The Moral Law, ed.

H.J. Paton (London: Hutchinson, 1948), pp. 90–91. 12 As an account of the moral community all of these suggestions have well-known disadvantages. Not one of them can account for the common assumptions that all human beings have equal moral standing; the higher animals rather less; but mammals more than

MARX AND EXPLOITATION

113

For the moment, in any case, we have the rough idea that to exploit someone is to make use of their circumstances in a way which fails properly to acknowledge their standing as an end in themself. Is such treatment a harm to the exploited individual? Here one might naturally think so, but we now confront what surprisingly has not been called the paradox of exploitation. The point is that exploitation is not always coercion; people appear to agree to accept exploitative arrangements. If they didn’t economic exploitation would be unlikely in the extreme, given freedom of contract. To see this the other way round, suppose you had the power to take a single exploitative contract of employment and make it illegal or otherwise prevent it. Who would thank you? Neither party would be better off without the trade. So unless some other trade, more advantageous to the exploited party (and there is no assurance that this will happen), replaces the now illegal transaction, both parties are made worse off by the prohibition. The argument, then, is that if people agree to a course of action they surely cannot consider it a harm. And if it is not a harm then it is not exploitation. Therefore there can be no such thing as (non-coercive) economic exploitation. To answer this we have to be clear that while exploitation involves making someone worse off, it cannot be defined in terms of making someone worse off than they would have been without the exploitative arrangement. But worse off than what? The best alternative is the view that the person is made worse off than they ought to be. But what determines how well-off someone should be? This is too general a question; we should restrict our attention to how well-off someone should be with respect to the particular transaction under consideration. David Miller, for example, suggests that the market in equilibrium, under suitable assumptions, sets the right baseline for comparison in economic cases.13 I have argued against that view elsewhere,14 but it illustrates the idea of setting a norm or baseline against which we can determine whether or not an arrangement is exploitative. What we need is a more normative approach. Here each of the Kantian, Aristotelian and Utilitarian perspectives can be of help. The Kantian approach would say that the baseline for comparison is some form of fish, and fish more than plants, which may have none. However this difficulty is not specific to the analysis of exploitation and so a failure to solve it here is not especially embarrassing, and indeed the flaw may be in common sense morality. 13 David Miller, “Exploitation in the Market,” in Reeve (ed.), Modern Theories of Exploitation, pp. 149–165. 14 Jonathan Wolff, “The Ethics of Competition” in A. Qureshi, G. Parry, and H. Steiner (eds.), The Legal and Moral Aspects of International Trade, Freedom and Trade, Volume III (London: Routledge, 1998), pp. 82–96.

114

JONATHAN WOLFF

agreement. Now if this agreement was simply what was actually agreed to, then we would be back with the “paradox,” so we need some notion of hypothetical agreement: what would be agreed to under some conditions. Under what conditions? Perhaps from a position of equal bargaining strength. This, though, is a notoriously difficult notion to define in a non-circular way. Better would be from ignorance of one’s position in the transaction; in other words from some version of a Rawlsian veil of ignorance. Finally, then, we end up with the main thought: a Kantian interpretation of the baseline for exploitation – motivated by taking a Kantian interpretation of what it is to be an end in itself – leads us to a “fairness” norm. Kantian exploitation, then, to a first approximation, is to use another person’s vulnerable circumstances to obtain their actual compliance with a situation that violates norms of fairness. It may be fairly simple to see how “Aristotelian” and “Utilitarian” accounts of exploitation may now be derived. They are, respectively, and again to a first approximation, to use another’s vulnerable circumstances to obtain their actual compliance with a situation which interferes with their flourishing in some way; or with a situation which involves them in some sort of avoidable suffering. There is no need to choose between these accounts: they are all forms of exploitation. Often they will coincide, in which case exploitation may seem particularly deep and unpleasant, for here exploitation may involve a form of degradation and suffering as well as unfairness. Indeed, putting this line of thought together with the earlier discussion allows us to discern three different dimensions of strength, or moral seriousness, of exploitation. First, which particular norms are violated? Second, how seriously are they violated? Third, is the situation generated or sustained by coercion? Arguably a paradigm case of the deepest type of exploitation would be the employment of young children at very low wages in extremely hazardous and life-shortening jobs, by those who have dispossessed the parents of land on which the family previously subsisted. Marx’s examples of the employment of young children in life-threatening occupations are well-known:15 it is equally well-known that similar practices still widely exist around the world. Such examples violate the first two norms to a grave degree, and, in many cases, are created or sustained by coercion. An example at the other end of the spectrum – what we could call “shallow exploitation” – might be the trader who, purely as a matter of brute good fortune, has large stocks during a temporary shortage, and who hikes the price simply because local consumers now have no alternative but to pay up. This case involves no coercion and violates just one norm, 15 Karl Marx, Capital, Volume 1 (London: Penguin, 1976), pp. 353–358 and elsewhere.

MARX AND EXPLOITATION

115

that of fairness. Thus although this case and the case of child labour are both examples of exploitation, they are poles apart. Deeper cases seem to involve much more than a monetary payment; they have a substantial effect on one’s behaviour or life. A further wrinkle. Could it be that one can be exploited yet, by accident as it were, come to no actual harm even according to any of the norms? Suppose an employer imposes extremely taxing conditions of work on three workers. For two of them the work is punishing and destroys their enjoyment of life, but for the third – and no one could have predicted or expected this – it is experienced as a source of self-affirmation. Does it seem odd to say that two are exploited while the third is not? I have to confess that I have been thinking about such issues for too long to have untutored intuitions here, but at least at one time I thought that if the two were exploited then so was the third. If this seems accurate then it seems also we are forced to consider exploitation as analogous to negligence: no harm need to be done, but insufficient precautions are taken to avoid harm. In other words, to be an exploiter is to use another’s circumstances to obtain their actual compliance with a situation without having sufficient regard to whether that situation violates fairness, flourishing, or suffering norms. To be exploited is to be treated in this way, whether or not actual harm is suffered. This is the final step in our fully general analysis of exploitation. Interestingly in his early use of the term “exploitation” in the German Ideology Marx refers to capitalism as a regime of “mutual exploitation.”16 One way of understanding this would be that sometimes person A exploits B, B exploits C, and C exploits A and so on. This is relatively uninteresting aside from the question of how economic power must be distributed in the type of society in which this is possible. However, if by “mutual exploitation” Marx had in mind both parties exploiting each other in a single transaction, this may, at first, be more difficult to understand. At first sight, it might seem necessary to adopt the counterfactual analysis of exploitation just given, for how can a transaction be unfair to both parties? So we might have a picture of both sides trying to press home a perceived advantage without regard to whether or not this is fair to the other party. However we are forced to the counterfactual analysis only if we think that exploitation always violates fairness norms (and that a single transaction cannot be unfair on both sides). This is almost certainly not what the early Marx would have had in mind. Rather mutual exploitation is likely to violate flourishing norms (through the notion of a species being). Each 16 Karl Marx and Frederick Engels, The German Ideology, ed. C. J. Arthur (London:

Lawrence and Wishart, 1970), p. 111.

116

JONATHAN WOLFF

party to an economic transaction under capitalism, according to Marx, degrades both their own nature and that of their trading partner. Each person is treated simply as a means to the pursuit of self-interest, rather than a member of the same community with needs of his or her own, and the capacity to confirm their own standing in the world as a speciesbeing.17 To put this another way, to lead a flourishing life, on Marx’s view, has both a biological and a social aspect. Economic trade substitutes the wrong type of social relations. Each person is forced, by the circumstance that they live under capitalist economic structures which make room for only certain sorts of interaction, to make a trade which leads both parties away from a properly flourishing life. In other words, while it is difficult to see how a transaction can be unfair to both parties, there is no problem in seeing how it can be degrading to both. It is a further step, of course, to imagine that this is true of all economic transactions. 4. M ARX

AND

E XPLOITATION

The last discussion suggests that there is more than one account of exploitation to be found in Marx’s work. The “official” account, though, is the one derived from Capital, as outlined in Section 1, involving unequal exchange of labour. What are the challenges to this? For the purposes of this paper I will focus on just one. Aesop’s fable of the Grasshopper and the Ants tells a story of the industrious ants, who spend the summer building a shelter and acting in a responsible, grown-up, fashion, preparing and cultivating the land, while the grasshopper spends the summer making music, singing and generally enjoying herself. When the winter comes the shivering, starving, grasshopper asks for the ants’ help, who tell her “if you spent the summer singing you can’t do better than spend the winter dancing” and go on with their business, chuckling to themselves.18 There ends the story. But let us imagination a slight variation, with somewhat less sadistic ants. In version two the ants say “if you spent the summer singing you should work for us on a subsistence wage while we spend the winter dancing.” In other words, the ants could use their power over the grasshopper to get a very advantageous bargain. Considered just in itself this looks like a case of non-coercive exploitation both in the official Marxist sense and according to the analysis 17 Karl Marx, “Economic and Philosophical Manuscripts,” in L. Colletti (ed.), Early

Writings (London: Penguin, 1975), pp. 279–400. See also Jonathan Wolff, “Playthings of Alien Forces,” Cogito 6 (1992), pp. 35–41. 18 Aesop’s Fables (Ware, Hertfordshire: Wordsworth, 1994), p. 12.

MARX AND EXPLOITATION

117

just given. If the grasshopper is performing all the winter labour, yet the ants are consuming some of her produce, then it seems clear that the grasshopper is expending more labour than the labour embodied in her consumption bundle. Further, the ants are using the grasshopper’s vulnerability – her lack of food and shelter in the coming winter – to propose a very one-sided and apparently unfair contract of employment. However the challenge is that if this turns out to be a case of exploitation, then so much the worse for the analysis. Surely the ants are quite entitled in the circumstances to propose the one-sided contract of employment. They have worked hard over the summer, and it seems only fair that they now get some return. Are they really expected to provide the grasshopper winter food and shelter for nothing, or even for a wage that does not allow them to profit in any way from their earlier labours? However such arguments also show why this is not properly a case of exploitation either in the Marxian sense, or in the sense proposed here. To take the latter first, provided that the ants do not require the grasshoppers to work in a degrading or overly exhausting fashion, then the transaction need not violate any norms. I agree that, considered without its history, the contract of employment may be considered unfair to the grasshopper, and would not be accepted behind the veil of ignorance. However it is obviously wrong to ignore historical factors in this case, and once these are taken into account both the ants and the grasshopper may well agree to the one-sided contract if they didn’t know which side they were on. In other words, once history is taken into account there is no unfairness. Obviously, though, there are limits. If the ants attempted to use their own one-season work to impose lifetime drudgery on the grasshopper while the ants sing in the summer and dance in the winter for the rest of their long lives (assuming an unnatural life-span of course), then the contract would have become unfair and contrary to the grasshopper’s flourishing and thus exploitative. So my analysis of exploitation is not challenged by this example once historical factors are taken into account. The case on Marx’s behalf is that (the amended Marxian) definition of exploitation is proposed as a reference fixing definition of exploitation in a model, and the model of capitalism presupposes that ownership of capital does not have the type of “clean history” had by the ants’ ownership of their shelter and cultivated fields. Thus we have to say that in Marx’s view the amended version of Aesop’s fable isn’t properly analogous to capitalism, and so the definition of exploitation is simply not intended to apply. We could speculate on how it would need to be adapted to apply to such cases but there is little or nothing in Marx to help here.

118

JONATHAN WOLFF

Indeed it is worth bringing out the two limitations with the Marxian definition (as reworked by Roemer) once it is interpreted as a “referencefixing in a model” definition. First a reference fixing definition of exploitation, if it is successful, provides a way of picking out all and only cases of where exploitation takes place, but it need not provide any particular insight into the nature of exploitation. I said earlier that there seems something wrong with the idea of defining exploitation as unequal exchange of labour for goods. I said that we have a “nagging feeling” that there is more to exploitation than this. But this is only to be expected with a referencefixing definition. To continue an example used earlier in this paper, there is, no doubt, more to the infinite than being the same size as a proper subset. A reference-fixing definition and an analysis are not the same thing. But however we might come to analyse the infinite, if the reference-fixing definition is a good one it will pick out the same set of objects as the analysis. Whenever a set has this feature it will have the other interesting features that infinite sets have. Why, then, bother with reference-fixing definitions? Only because they are precise and concise. Thus gold is defined as having an atomic number of 79. Although for most of us this is not exactly easy to apply, it does give perfectly precise conditions of application for those with the right knowledge and equipment. This, then, is the first limitation. A reference fixing definition need not pick out the most interesting, deepest, or most salient feature of the class of objects defined: it need only pick a property (or cluster of properties) held by all and only the objects under consideration. The second limitation we have already seen. Because, I claim, the amended Marxian definition should be taken as offering a reference fixing definition in a model, the assumptions of the model may add further, implicit, conditions. In a different model or context the definition may simply fail to apply.19 The best way of going further would be to provide a general analysis of exploitation and to show how the Marxian definition is an application of this general analysis for the particular circumstances of his account of capitalism, the most salient feature of which is that those who have accumulated large stocks of capital which give them market power have invariably done so through illegitimate means (in many cases illegitimate 19 In combination these limitations mean that one has to be very careful in proposing counter-examples to the definition. It may be possible to come up with examples where the “reference-fixing” feature (unequal exchange) becomes detached from the other features which normally accompany it (vulnerability, degradation, etc). If so, no particular conclusions can be drawn. Allen Wood makes related points in explaining why Roemer’s attempted counter-examples to the Marxian definition [in Allen Wood, “Should Marxists be Interested in Exploitation?” “Exploitation,” in Nielsen and Ware (eds.), Exploitation, fn. 15, p. 23] are “beside the point.”

MARX AND EXPLOITATION

119

even by the moral standards of capitalism). In short, in Marx’s model, there is no moral entitlement to private property. Compare the fable with another in which the ants have the shelter and fields only because they forced the aphids, who built and cultivated them, out of the region, and the grasshopper has no shelter for there are no more materials left with which to build. This is more like Marx’s model of capitalism. Under such circumstances it would be unfair for the ants to use their bargaining advantage over the grasshopper to extract surplus labour from her: from behind the veil of ignorance neither side would agree. This, then, is the basic way in which the Marxian account of exploitation is a special case of the general analysis.20 If Marx is right about the historical conditions of the development of capitalism, then fairness demands that those who find themselves holding private property have no right to benefit from it. Consequently contracts of employment in which surplus labour is required or at least expected violate norms of fairness, and this is enough to make the contracts exploitative. The fact that, on Marx’s view, the labour is alienated and involves great suffering compounds the case, to make exploitation under capitalism particularly deep. Finally, I mentioned that for Roemer, Marxian exploitation is not a relation between individuals, and we can see one reason why we should agree. Suppose Marx is right about capitalism and we consider one employer who employs a number of workers at the going market rate, which, let us suppose, involves unequal exchange of labour for money. We protest that the employer is exploiting these workers. The employer may well ask what she is expected to do: it would be impossible to remain in business yet pay a fair wage given Marx’s assumptions. And who would it help for her to go out of business? For Marx, this simply reveals a structural fact about capitalism: we are all locked into structures where all we can do to survive is to exploit others, 20 Talbot Brewer points out a methodological difficulty with the approach adopted here:

essentially an ordinary language analysis of the English term “exploitation” is used to attempt to illuminate Marx’s use of German term “ausbeutung” which does not translate perfectly into English. Some uses of the term “exploitation” – those concerning noncoercive advantage taking – would normally be translated by “ausnutzung.” Thus my attempt to provide a more general understanding of exploitation may, from the point of view of the German language, simply run together two different concepts. To complicate matters further, when Marx uses the notion of exploitation in the German Ideology he uses neither German term but the French expression “exploitation de l’homme par l’homme” (p. 110) which seems a fairly direct equivalent of the English. In the end, though, I make no claims about what was going through Marx’s head. Rather I want to provide a way of sympathetically elaborating his insight, while explaining the assumptions upon which it is best read as depending. I do not think that these (real) methodological worries about translation affect this main project.

120

JONATHAN WOLFF

and to degrade our own natures in the process. In the circumstances it is unhelpful to suppose that for each exploited person there is an exploiter. First of all, for technical reasons having to do with the transformation problem, this will simply not be the case; but second it suggests that a possible remedy lies in many individual capitalists taking a more humane attitude to their workers. But individual capitalists cannot do this while others are not, and therefore, can hardly be expected to sacrifice themselves in this way. Thus there are advantages in thinking of capitalist exploitation in class terms, even though in other cases it seems clearly a relation between individuals. Finally I should make it clear that I do not intend to endorse Marx’s model of capitalism here. This is not to say either that I want to defend capitalist ownership of private property in general either: my view is that cases differ. One task is simply to rescue Marx from some apparent criticism, by stating the conditions under which his definition is best read as operating. But placing Marx’s account in a more general context sets the scene for what we might think of a liberal democratic account of exploitation: there is still such a thing as economic exploitation – or rather many forms and depths of economic exploitation – even if we reject some of the moral assumptions driving Marx’s own account. Department of Philosophy University College London Gower Street London, WC1E 6BT UK