On the Fetishism of Primary Social Goods
All Animals Are Equal, But Some Animals Are More Equal Than Others.
-George Orwell, Animal Farm
In his Tanner Lecture “Equality of What?”, Amartya Sen considers several possible answers to the question of what exactly an egalitarian society should aim to distribute equally. One of these answers is Rawls’ conception of primary social goods (PSGs). An objection Sen levies against PSGs is their so-called fetishism. I will describe this charge in more detail below; in essence, it amounts to the objection that measuring distributive shares with regard to PSGs - or more generally, any resource-based metric - misses the forest for the trees: it is not the quantity of resources one possesses which determines how well off one is, but rather the basic capabilities enabled by those resources.
In this essay, I will explore the extent to which a theory of social justice such as Rawls’ from A Theory of Justice is answerable to Sen’s charge of fetishism. In Section 1 I discuss both primary social goods and Sen’s objection to them in more detail, and I describe Rawls’ justification for using PSGs to measure shares. In Section 2 I argue that principles of distributive justice, in the sense Rawls intends in A Theory of Justice, need not be concerned a priori with the distribution of "capabilities". Rather, a theory of distributive justice couched in terms of an independently defensible measure of shares such as PSGs is acceptable. The reason for this is that social justice is concerned with the distribution of the benefits of social cooperation. Insofar as capabilities fall into this category, the distinction between capabilities and PSGs is merely semantic; insofar as Sen’s conception departs from Rawls’, it falls beyond the purview of pure social justice, at least viewed naïvely.
I say “naïvely” because in Section 3, drawing on ideas of Scanlon, I describe another way in which egalitarianism enters the picture in discussions of social justice - namely, questions of individual responsibility and the value of choice; I argue that these considerations impose restrictions on the distribution of capabilities in order for the basic structures of a society to be just. As a result, principles of justice framed merely in terms of a resource-based measure of shares like PSGs are indeed liable to Sen’s objection of fetishism, although the manner in which the objection holds force is not quite that of Sen’s article.
1. Equality of What?
In this section we consider Sen’s question: Equality of what? We then turn to Rawls’ answer (primary social goods) and Sen’s objection to that answer (fetishism).
Sen considers four answers to his own question: marginal utility, total utility, PSGs, and capabilities. Equalizing marginal utility is mathematically equivalent to maximizing total utility, so this view leads to classical utilitarianism; this will not concern us here. The other three answers involving equality in shares of some quantity. It should be noted that when Sen discusses “equality” he means it in a broad sense. Given any particular metric for distributive shares, an egalitarian policy might call for complete equality between all persons or classes with regard to this quantity, or it might call for a lexicographic maximization of the share of the worst-off person or class (“leximin”) a la Rawls’ Difference Principle, or it might call for some other strategy entirely. Sen’s question amounts to asking which, if any, of these quantities makes the most sense to use in discussions of distributive justice, independently of the particular strategy used to distribute it “equally”. In the context of G.A. Cohen’s analysis of the same problem ("On the Currency of Egalitarian Justice", Ethics vol. 99, July 1989, pp. 906-44) Sen is making "weak equalisandum" claims. We will not consider Sen’s objections to a total utility metric; our focus is on PSGs.
Primary Social Goods. In A Theory of Justice, Rawls makes the simplifying assumption that there is a class of so-called primary goods that “every rational man is presumed to want.” These include natural primary goods - such as strength, intelligence, and health - as well as primary social goods - such as rights, liberties, opportunities, wealth, and self-respect. What distinguishes PSGs from other natural primary goods is the degree to which their distributions are directly under the control of the basic structure of society, rather than merely indirectly influenced by that structure.
In A Theory of Justice Rawls appeals to primary social goods, as opposed to a total utility metric for distributive shares, to make his system of “justice as fairness” neutral with regard to varying conceptions of the good, within very broad limits. (See pp. 92-4 of Theory; the limits being referred to are those of the "Aristotlean" conception of the good as the satisfaction of rational desire.) Rawls’ argument rests upon the foundational idea underlying his theory of justice, that principles of justice are acceptible if rational persons in the original position would agree to them. Persons in the original position are ignorant of their conception of the good, and so, Rawls suggests, they would desire more rather than less of any primary social good, since primary goods enable the satisfaction of any set of rational desires. (See loc. cit. In contrast, the utility associated with the satisfaction of their desires need not be something everyone in society agrees upon. For this would require some way to evaluate various conceptions of the good, which opens up a whole host of difficult questions....) Consequently, persons in the original position would agree to use PSGs as a metric for distributive shares, as “the most feasible way to establish a publicly recognized objective measure, that is, a common measure that reasonable persons can accept" (p. 95). In A Theory of Justice, Rawls does not take up the question of why natural primary goods would not be an appropriate component of the metric for distributive shares; however, on this matter, see Section 2 below.
Fetishism and Narrowness. In developing his critique of Rawls’ PSGs, Sen makes use of an example which will also be helpful to us in the sequel. Consider a disabled person X who is unable to move independently - perhaps he is paralyzed below the neck - without an expensive wheelchair. Suppose X has all the rights and liberties afforded to other members of his society, and that the government even makes arrangments for transportation to the polling booth so as not to effectively disenfranchise X. Suppose further that X has found satisfying employment which he is able to pursue despite his disability; perhaps he is a moral philosopher, and can write articles from home using a computer controlled via blinking. He does not lack steady income, self-esteem, or any other of Rawls’ primary social goods. Nonetheless, through no fault of his own, X is at a severe disadvantage in terms of utility. For, unable to afford the wheelchair, he cannot pursue those of his rational desires which require independent mobility. Despite this state of affairs, his government, modelled after Rawls’ "well-ordered society", does not see fit to purchase him the wheelchair. Indeed, so long as he has the requisite allotment of PSGs, the government is entirely unconcerned with whether or not X gets the wheelchair. Following Sen, we note the strong intuition that justice is not being served with regard to X. That is, in this particular case, using PSGs as a metric of distributive shares seems to yield an undesirable result.
Extrapolating from the example of X, Sen poses two related objections to the use of PSGs. One of these is their narrowness: in cases like X’s, PSGs seem to be too narrow a framework in which to measure a person’s well-being. Due to other factors, people’s lives can vary quite significantly - in ways which seem relevant to the extent to which a social arrrangement is egalitarian - despite their possessing equivalent, or even identical, bundles of primary social goods. The same problem can be recast more abstractly thus: It seems fetishistic to couch principles of distributive justice in terms of a PSG metric, because as mere goods PSGs themselves are just a means to the end of fulfilling rational desires -- they are not the fulfillment of those desires themselves. As Sen puts it, this results in the utility associated to such fulfillment being relegated to complete irrelevance, as regards the sense of urgency which gives questions of distributive justice their resonance. This seems undesirable, insofar as there is an underlying pre-theoretical principle which holds that the measurement of a person’s distributive share must be correlated in some way with how happy they are. Not only PSGs themselves, but also what they do to people, are important.
Capabilities. Sen’s positive theory, developed as an alternative both to fetishist PSGs and to similarly undesirable "welfarist" (utility-based) metrics for distributve shares, is founded upon capability sets. These are defined as the sets of certain basic things people are able to do. The relevant example in the case of the disabled man X is the capability to move about independently, which X lacks but might obtain if someone bought him his wheelchair.
Cohen, in the article mentioned parenthetically above, suggests that capability is an “infelicitous” term for what Sen really means, which is actually more along the lines of something “goods do to or for people”. Indeed, Sen’s defense of his view, as “a natural extension of Rawls’ concern with primary goods, shifting attention from goods to what goods do to human beings” bears out this reading. The essential point is that capabilities offer a way to preserve the desirable neutrality of PSGs as an egalitarian metric, while eliminating their fetishism. For principles of justice couched in terms of capabilities would cast no judgment about which of their capabilities a person can or should exercise; all such principles would require is that each person have a legitimate expection of some set of capabilities, to exercise as she sees fit.
2. Fetishism and the Scope of Distributive Justice
I now turn to the question of the extent to which Rawls is answerable to Sen’s objection. In this section, I shall argue that in fact the charge of fetishism holds no intrinsic force as an objection to the use of PSGs as a metric for distributive shares, within the context of a theory of social justice like Rawls’.
An initial question to ask is, "To what extent do Rawls’ PSGs differ from Sen’s capabilities?" Insofar as primary goods simpliciter enable the pursuit of rational desires (their intended purpose), they occupy an important part of the conceptual niche filled by capability sets. For by enabling the pursuit of such desires, it is clear that primary goods endow their possessors with capabilities. Conversely, most important capabilities - such as the ability to move about, the ability to live a healthy life, the ability to read, etc. - derive from one primary good or another. This is not to say that they derive from a primary social good, for indeed many important capabilities (such as our example of independent movement) manifestly do not. (Instead, primary natural goods, such as health, seem more relevant.)
Now to simply identify a person’s share of primary goods with her share of capabilities would be to sneakily sidestep Sen’s charge of fetishism. And yet one may legitimately argue somewhat similarly. Wherever Rawls’ theory makes use of the phrase “primary good”, simply replace it with “capability.” Persons in the original position are concerned with primary goods (or rather, the social subset of them, the PSGs, to which I will turn in a moment) merely instrumentally, as a metric of distributive shares. Since - I have suggested - capabilities (or the lack thereof) are precisely the effects of primary goods (or the lack thereof) on persons, primary goods measure precisely the same thing as capabilities. Hence, as a metric of distributive shares, primary goods are entirely equivalent to capabilities.
Be this as it may, this does not fully justify the use of primary social goods in A Theory of Justice. For this purpose, it is necessary to consider that Sen’s objection occurs within a context somewhat more general than the questions of social justice with which Rawls is concerned. Sen is investigating the formulation of equality most appropriate within the very broad scheme of what he calls “discussions in moral philosophy”. One way of looking at this problem is the following: If a society were to equalize some measure of distributive shares, would it accurately capture our intuitions regarding an egalitarian society? Writing in a similar vein, Cohen identifies the primary egalitarian impulse as “to extinguish the influence on distribution of both exploitation and brute luck.”
In contrast, Rawls uses PSGs as a metric for distributive shares, not as a foundation for an egalitarian society, but as the only just (in the sense of justice as fairness, that which is chosen in the original position) basis for comparing basic structures and thus arriving at principles of distributive justice. It so happens that Rawls believes the resulting principles would have an egalitarian character (i.e. the Difference Principle), but this purpose should still be distinguished from the role of the egalitarian metric in Sen’s lecture and Cohen’s article. So long as PSGs provide a justifiable metric for the purposes of arriving at principles of distributive justice, Rawls is safe from Sen’s objections, even if they hold more generally.
Therefore, it suffices to defend Rawls’ restriction of his attention to distributive shares of PSGs - or equivalently, to the corresponding capabilities - rather than to distributive shares of primary goods (capabilities) in general, within the limited context of a theory of social justice. And this limited context is the crucial point. For Rawls carves out an explicitly limited scope for his theory in the introduction of his book:
The scope of our inquiry is limited. . . I am concerned with a special case of the problem of justice. I shall not consider the justice of institutions and social practices generally. . . There is no reason to suppose ahead of time that the principles satisfactory for the basic structure hold for all cases.
And what exactly is the intended scope? Rawls specifies it earlier in the introduction:
There is an identity of interests in a society since social cooperation makes
possible a better life for all than any would have if each were to live soly
by his own efforts. There is a conflict of interests since persons are not in-
different as to how the greater benefits produced by their collaboration are
distributed, for in order to pursue their ends they each prefer a larger to a
lesser share. A set of principles is required for choosing among the various
social arrangements which determine this division of advantages and for un-
derwriting an agreement on the proper distributive shares. These principles
are the principles of social justice. . .
Thus, social justice is not concerned with the proper division of distributive shares, in the broadest moral sense of Sen and Cohen, but rather with distributive shares of the advantages stemming from social cooperation. Note that, along with Robert Nozick in Anarchy, State, and Utopia (Basic Books, 1974), pp. 185ff, we may be skeptical regarding why it is social cooperation itself which gives rise to the problems of distributive justice. However, since Rawls writes under this presupposition, in objecting to his choice of PSGs as the metric for distributive shares along Sen’s fetishist lines, one must be charitable to Rawls on this point. Furthermore, in justifying his use of PSGs in A Theory of Justice, Rawls makes a convincing argument to the effect that all capabilities deriving from primary goods simpliciter- comprising the capabilities which enable the pursuit of rational desires - are, prima facie, “fair game” in this matter. It would be an unjustifiable breach of neutrality to restrict the purview of distributive justice to shares of capabilities which adhere to any particular conception of the good. Rather, any capability within the scope of “advantages stemming from social cooperation” must be apportioned according principles of distributive justice.
To rebut the charge of fetishism, then, all Rawls requires is an argument to the effect that capabilities deriving from natural (non-social) primary goods lie beyond the scope of the benefits from social cooperation. And indeed, such an argument is readily forthcoming. Natural primary goods, such as health, good looks, or intelligence, are not, in general, products of social cooperation. One can come up with exceptions - such as a cheap vaccination against a grave illness, developed using government-funded research. In this case, the better health resulting from vaccination might be considered a product of social cooperation. But it also seems clear that this particular aspect of “health” cannot be considered a natural primary good in the intended sense; rather, it is manifestly a primary social good - precisely because it derives from social cooperation! - and would thus be distributed appropriately according to a Rawlsian theory of social justice. In general, the capabilities deriving from PSGs are precisely those things members of society have claim on as participants in the practices that produced them. Consequently, these are precisely the capabilities the distribution of which concerns Rawlsian distributive justice.
We may conclude that Rawls is quite safe from the charge of fetishism in his use of PSGs as a distributive metric. (However, we shall see below that there is a more sophisticated form of the objection which does hold force against Rawls.) One may observe that the precise principles of distributive justice involved in Rawls’ theory never come into this debate at all. Indeed, any theory of social justice - in the Rawlsian sense of the quoted passages - founded upon the metric of PSGs is permissable.
3. Responsibility, Choice, and Capabilities
But there is always a “but.” As remarked above, the charge of fetishism can be recast in a
subtler way as a serious objection to the exclusive consideration of PSGs in determining principles of distributive justice. As we shall see, this reformulation can account for the troubling intuitions Sen points out, regarding the case of the
disabled man X.
Responsiblity and the Value of Choice. A powerful objection to Rawls’ political philosophy (and liberal political philosophy in general) is that it relies upon a conception of individual responsibility which is thinner than the one which jives with many people’s considered judgments regarding morality - especially as regards retributive justice (crime and punishment). For an elaboration and analysis of this view, see Samuel Scheffler, "Responsibility, Reactive Attitudes, and Liberalism in Philosophy and Politics,” Philosophy and Public Affairs 21 (1992): 299-323; see also Scheffler’s “Justice and Desert in Liberal Theory,” California Law Review 88 (2000): 965-990. For, the argument runs, a consequence of Rawls’ rejection of societies which do not conform to the Difference Principle as unjust, is that in such societies worse-off individuals are not responsible for their position. Were the basic structures different, such individuals would have more; hence, by Rawls’ lights, such individuals deserve more - in an institutional sense of the word “deserve”, whereby people deserve what they can legitimately expect from just basic structures.This Rawlsian conception of desert is ennuciated in §48 of A Theory of Justice, pp. 310-315. If individuals lack responsbility for their economic situation in this manner, then they must also lack moral responsibility for their actions, such as crimes they commit. Hence, they do not deserve punishment for their crimes, and this seems to conflict with our considered judgments, and Rawls’ own remarks, regarding retributive justice. (Scheffler (2000) develops this argument in detail.)
To save liberal philosophy from this dilemma, Thomas Scanlon argues for a distinction between moral responsibility - which he identifies with the attributability of actions to moral agents - and substantive or social responsibility (this is the responsibility a person has for his fate when others are not obligated to do anything about it). In the same article and elsewhere, Scanlon suggests that this distinction rests upon different conditions for when people are responsible for their choices in each of the two senses. (See T.M. Scanlon, Jr., “Justice, Responsibility, and the Demands of Equality,” The Egalitarian Conscience (March 2006): 70-88; “The Significance of Choice,” The Tanner Lectures on Human Values, 1986. Online at http://www.tannerlectures.utah.edu/lectures/scanlon88.pdf.) People may be morally responsible for their actions even when they lack substantive responsibility for their position because their social institutions are open to legitimate criticism. By these lights, people lack substantive responsibility for their positions in unjust societies because their choices are not made insufficiently favorable conditions; their choices lack sufficient value. In a just society, people’s lack of economic benefits is only made legitimate by the fact that their choices do have sufficient value in such a society.
Capabilities, Revisited. What does all this have to do with fetishism and primary social goods? The answer to this lies in the factors which contribute to the value of a person’s choices under a given set of conditions. In some (but not all) cases,the collection of available alternatives contributes positively to the value of a particular choice. It seems plausible that, generally speaking, capability sets play an important role in guaranteeing the value of people’s choices.
Here, then, the quotation from Orwell at the beginning of this paper becomes relevant. Even within a society which is just in Rawls’ sense, in which the worst-off persons (with regard to PSGs) have more PSGs than they would under any other basic structure, the example of the disabled man X demonstrates that people’s capability sets may be quite different. Consequently, the value of their choices may differ. Insofar as Rawls’ principles of justice are to be viewed as egalitarian, this is a society in which all people are treated “equally”. And yet, some are more equal than others.
What insures that people’s choices are sufficiently valuable in this society that the decisions they make have legitimating force, and thus justify the economic outcome in this society itself? The principles of justice themselves must play this role, since they provide the only moral constraints on the basic structures. To the extent that capability sets influence the value of choice, there must therefore be some place for capabilities (in Sen’s sense) in questions of distributive justice, despite the discussion in Section 2 above. While distributive justice does not comprise the just distribution of capabilities themselves (they are not an appropriate metric for distributive shares in the benefits from social cooperation), it is indeed the case that principles of distributive justice must be somewhat constrained by the actual distribution of capabilities. For if under this distribution some people’s choices were sufficiently devalued, the guiding principles of justice would in fact lack legitimacy.
Fetishism, Revisited. In this context, Sen’s objection that PSGs are a narrow and fetishist metric for distributive shares becomes newly relevant. On the one hand, the considerations above provide a theoretical reason why a legitimate theory of social justice should be concerned with the distribution of capabilities. We should therefore be suspicious of fetishist principles of justice couched solely in terms of goods rather than the capabilities they endow.
On the other hand, the same considerations justify concern over the narrowness of the PSG metric. For there is every reason to believe that the value of people’s choices is greatly affected by all their capabilities, rather than merely those stemming from primary social goods. The example of X makes this clear: his lack of independent mobility devalues his choices to an alarming degree, but does not derive (essentially) from a lack of PSGs. (While it could be rectified with a greater share of PSGs, viz. more wealth, it is essentially a consequence of X’s deficiency in the natural primary good of a well-functioning body, and the corresponding capability of independent mobility.)
Now, it should be noted that this objection to Rawls’ principles of justice is rather weak. For example it seems quite possible that a just distribution of PSGs in Rawls’ sense automatically gives rise to a distribution of capabilities in which choices have sufficient value to legitimize the outcome. If this were in fact the case, then the problem of narrowness would be dissolved, and perhaps the fetishism concern could be similarly dispatched with. However, it is not at all clear a prioi that this “automatic legitimacy” actually obtains.
Moreover, I believe that the considered judgments or intuitions regarding the example of X, to which Sen originally appealed in formulating his objections to PSGs, should give the Rawlsian cause for concern. The objection posed in this section holds that the Rawlsian conception of justice, which results X’s remaining immobile. might be illegitimate on account of X’s devalued choices. This seems to coincide with the intuition that Justice, in a larger sense, has not been served with respect to X.
4. Conclusion
I have argued that an objection to a Rawlsian theory of social justice couched in terms of primary social goods, along the lines of Sen’s charges of narrowness and fetishism, fails on account of the limited scope of distributive justice. For this objection rests upon capabilities derived from natural primary goods, which capabilities are not part of the benefits from social cooperation, and are therefore not part of that bounty the just distribution of which concerns Rawlsian distributive justice.
However, the considerations Sen raises regarding narrowness and fetishism may nonetheless serve to destabilize Rawls’ theory, albeit by a different route. By calling into question the distribution of capabilities - and hence the value of choice - in a society which Rawls would consider just, these objections undermine the legitimacy (and hence the justice) of this outcome. Since issues regarding individual responsibility and desert force the Rawlsian to ensure his society’s legitimacy in this sense, the fetishism of primary social goods is therefore a potentially significant objection to Rawls’ theory of justice.