Pluralism and the Standard of Living
Pluralism and the Standard of Living
Abstract and Keywords
The proper province of a government's economic policy is the enhancement of the standard of living of its citizens. Such a remark, while true, is hardly informative. For in what does the standard of living consist? The intention of this paper is to sketch part of an answer to this question, involving some necessary (though not sufficient) conditions that characterize the concept. In particular, I want to examine how a notion of the standard of living, intended as a criterion of use to governments implementing policy, might be sensitive to a relatively pluralist theory of society. By a pluralist theory I understand an account of how society should be arranged that incorporates the possibility of multiple and non‐trivially divergent views of the good life for individual human beings. The rider about non‐trivial divergence is important: a theory of society is not pluralist merely because it assigns some citizens to be philosophers and others to be accountants, snake‐charmers, or car‐park attendants. A genuinely pluralist theory must count it a social good that there may exist multiple views of the individual good that are not subsumable under an encompassing theory of the individual good. This does not imply that a pluralist social theory must be neutral between all or even many theories of the individual good; for it to count as pluralist, it is enough for it to be compatible with the assertion of more than one theory of the individual good. I shall not argue for the correctness of such a point of view. In this paper I want to see instead whether it is possible for a government's economic policies to incorporate it if it is correct. Or does the very notion of an economic policy presuppose that a government has its own comprehensive theory of the individual good? Nevertheless, although I do not consider the correctness of pluralism as such, I shall argue that the possibility of a pluralist economic policy has important implications for the coherence of pluralism even in principle....
The proper province of a government's economic policy is the enhancement of the standard of living of its citizens. Such a remark, while true, is hardly informative. For in what does the standard of living consist? The intention of this paper is to sketch part of an answer to this question, involving some necessary (though not sufficient) conditions that characterize the concept. In particular, I want to examine how a notion of the standard of living, intended as a criterion of use to governments implementing policy, might be sensitive to a relatively pluralist theory of society. By a pluralist theory I understand an account of how society should be arranged that incorporates the possibility of multiple and non‐trivially divergent views of the good life for individual human beings. The rider about non‐trivial divergence is important: a theory of society is not pluralist merely because it assigns some citizens to be philosophers and others to be accountants, snake‐charmers, or car‐park attendants. A genuinely pluralist theory must count it a social good that there may exist multiple views of the individual good that are not subsumable under an encompassing theory of the individual good. This does not imply that a pluralist social theory must be neutral between all or even many theories of the individual good; for it to count as pluralist, it is enough for it to be compatible with the assertion of more than one theory of the individual good. I shall not argue for the correctness of such a point of view. In this paper I want to see instead whether it is possible for a government's economic policies to incorporate it if it is correct. Or does the very notion of an economic policy presuppose that a government has its own comprehensive theory of the individual good? Nevertheless, although I do not consider the correctness of pluralism as such, I shall argue that the possibility of a pluralist economic policy has important implications for the coherence of pluralism even in principle.
The argument proceeds in four stages. In the first I shall argue for the general propriety of allowing a moral or political theory to depend intrinsically on the practical limitations upon implementing its recommendations. By this I mean more than the fatuous point that a theory should not make impossible recommendations. I mean that our choice of a moral or political theory should not always be significantly influenced by our intuitive reactions to what it recommends in impossible or sufficiently unlikely circumstances. In the second stage I shall use this conclusion, along with a series of specific constraints on implementation, (p.394) to help characterize the notion of the standard of living, giving some necessary conditions that any adequate standard must meet. In the third stage I shall draw analogies with other arguments in philosophy and economics that illuminate the connection between this notion and a pluralist theory of society. In the fourth stage I shall use the notion thus characterized to answer some specific questions about living standards and pluralism. These include the question of how the standard of living can incorporate values that are specific to certain cultures without thereby becoming relativist, and the question of whether economic policies need be intrinsically biased towards what is sometimes called commoditization, and to the market economy.
1 The Limitations on Moral Theories
It is a common practice among philosophers and social theorists to test the adequacy of a general theory by searching for possible counter‐examples. Robustness against counter‐example is definitive of a successful general theory, and on some views definitive of a theory, tout court. In ethics and social philosophy, where the data of our theories consist of our intuitive reactions to different situations, this search frequently takes the form of developing elaborately artificial situations designed to abstract away from the messy and confusing detail of every day. Thus we may situate moral dilemmas on Mars or behind a veil of ignorance or on tramlines stacked with Nobel Prizewinners, rather than in the Cabinet or in EastEnders1 or in our own families. The process by which the intuitions in these abstract cases inform our more general judgements has been variously characterized, and John Rawls's notion of ‘reflective equilibrium’ represents one of the clearest and best known of such accounts.2 While I have no intention of disputing the frequent usefulness and clarificatory value of such exercises,3 I do want to argue that we can sometimes reasonably object to the result of such a thought experiment that the situation described is improbable or bizarre. The reason for this is not that I wish to reject the idea that a general theory must be general. ‘Not Fa’ is still and will always be conclusively incompatible with ‘For all x, Fx’. The objection instead is to the idea that the judgements delivered by intuition in improbable cases must always amount to counter‐examples.4
This line of argument makes best sense within a broadly naturalistic view of ethics. Our ethical intuitions are the product of a long process of both biological and cultural evolution. This evolution has been shaped in a major way by the (p.395) need to cope with certain recurrent problems and difficulties faced by human beings in society. It has been shaped much less by very rare problems and difficulties that have affected very few people. And, at the risk of labouring the point, I should add that it has not been shaped at all by the need to cope with situations that have never happened or could never happen. It should not be surprising, therefore, if the intuitions we have concerning improbable or impossible situations turn out to be quixotic or mercurial. It is unwise to trust them very far.
At this point it is important to distinguish two conclusions that do not follow from this naturalistic line of reasoning, in order to protect from guilt by association the conclusion that does. First, from the fact that certain moral intuitions have arisen in us as a result of the need to cope with evolutionary contingencies it does not follow that these intuitions are morally defensible ones for us to have: aggression, vindictiveness, and jealousy all have comprehensible evolutionary roots. Instead, what the naturalistic argument illuminates is a question at the heart of the method of reflective equilibrium: do we have any reason to think that an equilibrium exists? Our intuitions can be considered as inputs to this reflective process. Certain requirements of consistency may have been imposed on those of our intuitions that have felt the evolutionary squeeze; others which have never done so may well be wildly inconsistent with what we believe in more familiar contexts. If we require a moral theory both to respect the requirements of reflective equilibrium and to be applicable to all logically possible situations, we may require the impossible.
A second unwarranted conclusion would be that this evolutionary reasoning implies an anti‐objectivist or even just a non‐cognitivist account of ethics. Many aspects of reality are too complex for many of reality's inhabitants to understand. The findings of modern physics are certainly beyond the comprehension of a dog, and we have no guarantee that the physical nature of reality will not turn out to be beyond our own comprehension too. To believe that we alone in the animal kingdom had reached the evolutionary plateau that would enable us to comprehend our own environment in its entirety would be a return to the image of ourselves as a chosen species, with no longer a God to be chosen by. So if likewise it turned out that we could not develop ethical theories to cover all possible contingencies without veering into contradictions, that would be a comment upon our moral capacities, not upon the existence or otherwise of objective moral truths.
The conclusion that does follow from this naturalistic argument is that, if we cannot be sure that there exists a moral theory that consistently satisfies our intuitions in all logically possible cases, we had better not throw out all candidates for a theory that happen to fail the test of our intuitions in certain improbable circumstances. We must find ways to choose among those theories that satisfy our intuitions in a sufficiently large number of cases. And here it seems to me inescapable to judge between cases not merely on grounds of number but also on grounds of importance. This is not the place to propose a detailed (p.396) criterion, but such a criterion will certainly tend, other things being equal, to ascribe greater weight to our intuitions concerning cases that are more central to our everyday practices of judging, evaluating, and acting, than to those that are not.5
This issue bears a close relation to an issue in the literature on social choice theory. That literature has been much concerned with the condition of unrestricted domain, which is a condition requiring any reasonable social welfare function or social decision function to apply to any logically possible ordering of individual preferences. To some it seems obvious that the question of what a defensible social decision function should be is independent of the particular decision problems it is applied to in any individual case: if a particular outcome is impossible, it is ruled out of the choice set as infeasible, and if a particular combination of preferences is impossible, it is ruled out as an input to the function. But at all events this infeasibility should not affect the choice of the original function. In other words, on this view you should first decide the moral theory and then apply it to some particular choices; there is no sense in allowing your choice of a moral theory to depend on what choices you will happen to face.
The problem with such a line of argument is that by imposing the condition of unrestricted domain along with other conditions (such as that the function respect transitivity), one effectively requires social preferences over feasible outcomes to be sensitive to what social preferences would be over outcomes that are improbable or physically impossible. This might not matter if we knew that social welfare functions compatible with all our intuitions existed. But the social choice literature has documented with depressing thoroughness the circumstances in which no social welfare function may exist and satisfy the restrictions that we might intuitively wish to place on it. It is clear that some of our intuitions somewhere may have to give. To insist upon retaining the condition of unrestricted domain is then to admit only those social welfare functions that fail to satisfy some of our other intuitions. The result is paradoxical: the price of requiring a theory to be compatible with all possible preferences over states consists in no longer requiring the theory to be compatible with even all actual preferences over theories.
The main reason, then, for not necessarily requiring a moral theory to satisfy our intuitions in improbable or impossible circumstances is that a theory compatible with all our intuitions may not exist. A subsidiary reason, familiar from considerations in the philosophy of language, is that it may be dubiously intelligible (p.397) to ascribe to creatures capable of intentional action preferences and beliefs bizarrely different from our own. It is not always evident what may count as a genuine possibility, and certain situations may appear intelligible until we inquire closely just what circumstances would make us judge that such a possibility had indeed occurred.6 This is a particularly powerful argument against the condition of unrestricted domain of preferences in the social choice literature. Here it underlines that in addition to the uncertain reliability of our intuitions in improbable or impossible circumstances, we must consider the uncertain intelligibility of those hypothetical circumstances themselves.
In the context of the present argument the upshot is this. A political theory, which is one kind of moral theory, deploys a variety of moral concepts in accordance with our intuitions about their application. One with which this paper will be much concerned is the concept of a contract, but there are of course many others, concepts like utility, the good life, self‐respect, dignity, obligation, right, the citizen, and so on. I am arguing that it will often be legitimate to characterize a political theory in terms of its adequacy to these concepts as we actually have them, not as we might have had them. For example, it is not in itself an objection to retributive theories of punishment that under conceivable circumstances (those described by Nagel (1976) and Parfit (1976) in their discussions of brain bisection, for instance) the boundaries between persons might be uncertain and the question ‘Who committed this crime?’ might have no answer. Our theories of punishment have developed to deal with persons as they are. And as they are, persons are distinct entities characterized by physical continuity and a high degree of psychological continuity. If grotesque possibilities become actual we may, of course, need to develop our theories to cope with them. But unless there is some real chance of their doing so we would be unwise to allow our choice of a theory of punishment to be determined by their mere possibility.
There is nothing unusual or original in this form of argument. The emphasis in Rawls's recent work on the political, not metaphysical nature of justice as fairness7 can best be understood as calling attention to the need to fashion justice to the needs of human communities that already have enough cohesion and common purpose to make justice possible at all. Using a theory of justice to speak to societies so fragmented that they cannot begin to co‐operate is a futile exercise, one which therefore teaches us nothing about the kind of justice that more unified communities require. We can say more: if certain rules could never be implemented in any community that was recognizably like our own, then a theory of justice had better respect that constraint, and not seek to judge communities by their degree of resemblance to communities in which such rules are possible.
It was stated earlier that a government's economic policy should concern itself with the standard of living of its citizens. I want now to turn this around and claim that the way to understand the notion of the standard of living is as that aspect of individuals' well‐being that falls within the proper sphere of society's concern. My perspective in making this claim will be broadly contractarian.8 Society is the institution of co‐operation between individuals to their mutual benefit (it may not be only that but it is at least that). In particular, society makes possible the implementation of a framework—involving, for example, property rights and the organization of both production and consumption—without which individuals could not enjoy many of the activities that are central to their well‐being. I am not here concerned with the difficult question as to whether an adequate theory of justice must ensure that all individuals are at least as well off in society as they would be on their own (the greatest problem for such a view being posed not by some of the very badly off, whose position is certainly unjust, but by able and law‐abiding members of society who in the state of nature would be successful cocaine barons).9 It is enough for my purposes to observe that in taking part in the institutions of production, consumption, and exchange as they in fact exist, all individuals gain benefits from the co‐operation that underlies these institutions. We can therefore reasonably ask what features of these institutions would have emerged from a process of agreement, or more formally of contract, between the individuals concerned, if society had indeed emerged from agreement instead of higgledy‐piggledy from history.
Many such features will depend upon how precisely such a contract is specified (for instance, whether individuals are behind a veil of ignorance or aware of some features of their situation in life). For present purposes all that matters is that it is a contract. But—and here lies the connection with the first stage of argument of this paper—it is a contract that is recognizably like contracts made by real people in the real world. In the real world, for example, contracts are frequently made contingent on certain specified circumstances (‘if the insured suffers a financial loss the insurer undertakes to indemnify him against such loss’). But the circumstances must always be, at least in principle, publicly verifiable. An insurance company may indemnify an individual against a loss of money, which is verifiable by recourse to the individual's bank, or against a loss of health, which is verifiable by recourse to his doctor. But it will not—it could not—indemnify him against a loss of utility, except in so far as a loss of utility might happen to be dependent on something that is publicly verifiable. So if we are to make sense of a contract at all, we must understand it as something (p.399) public, and the contingencies on which it might depend as being public too. Members of society could and perhaps would contract to insure themselves against the hazards of birth and circumstance, but only against the publicly verifiable hazards. They would not contract to insure themselves against all inequalities in utility. A contract that did that would be no contract in any sense that we understand the term; it would be a kind of divine intervention.
Similarly, all contracts specify some kind of exchange, and no party can contract to offer something which it is not in his power to supply. This has two aspects. First, I cannot contract to give you something which I cannot ever deliver to anyone. So I cannot offer you my utility or my state of health or my relationships with my friends, though I can offer you my meditation techniques or my patent medicine or my well‐thumbed copy of How to Win Friends and Influence People. Second, I cannot contract to give you something that must be given freely and without requiring a return if it is given at all. So I cannot make a contract promising you my respect or my love or my agreement with the conclusions of your latest book (I can promise to treat you with respect, but that is something different). And members of society contracting with each other to share the benefits of co‐operation cannot make such contracts either. Perhaps this is unfortunate, for the kinds of good we cannot trade by contract include many of the most important elements in the flourishing of human beings. But the fact remains that these goods are outside the sphere of contracts as we know them or even conceivably could know them.
Indeed, from another point of view it is not unfortunate at all, for there would be something distinctly unattractive about even the attempt to think of these goods as within the contractual sphere. One can think of many kinds of motivation that are not only inappropriate for their intended effect but become in the process of conception ridiculous. Human beings need self‐respect, and in order to attain self‐respect they must have the respect of their fellow human beings. But it would be absurd if they sought to do so by contracting to have respect for each other. In like manner, there can be something by turns comic and sinister about politicians' undertaking to create greater altruism or to improve the quality of family life, as distinct from ensuring that individuals have the means at their disposal to create these conditions themselves. Not only is this something they cannot deliver, but they ought not even to try.
In fact, the sphere of contracts consists more or less of rights over scarce10 physical commodities (including rights of access to such things as library books and land), and of rights to the performance of services, these being publicly verifiable, intentional activities performed by or caused by other human beings. Social contracts must be contracts for the exchange of these commodities or services, contracts that are contingent on publicly verifiable circumstances. And (p.400) the standard of living of citizens is represented by the command that they have over these commodities or services, given these circumstances. More precisely, one may say that the standard of living of individuals consists of those components of their well‐being the enhancement of which would be the appropriate subject of a social contract between individuals wishing to share the benefits of social co‐operation.
It will be evident by now how this account makes room for pluralism. There are a great many aspects of the good life for individuals the enhancement and distribution of which are simply not society's business. This is because, on the contractarian account I am outlining, nothing is society's business unless it could be the subject of an appropriate hypothetical social contract. Thus it is not the business of society at large whether people have happy marriages or believe in God, because these are not the kinds of thing people could contract to do. It may of course be the business of their friends or of their priests, individuals with whom their relationship is not contractual. And it will certainly be society's business whether people have the resources to make happy marriages possible (such as the ability for a couple to work and live in reasonable proximity to each other), and the ability to call on the state's resources to defend them from attacks on their worship of their own God in their own way. But their standard of living represents their command over these resources, and not the outcome that results from it. A pluralist social theory, which allows for the possibility of fundamentally different conceptions of the ends to which these resources may be put, will be superior to a theory which requires society to evaluate social outcomes in every detail, when many aspects of these outcomes are not ones society should be in the business of evaluating at all. The reason they are not is that the kind of world in which these aspects could be the subject of contract would be a world so unlike our own (peopled by individuals completely transparent to each other, and able to trade mental states with each other at will) as to be no fit point of comparison with the world we know.
In describing the standard of living as involving command over resources rather than the outcomes that result, I should make one point clear. The value of command over given resources may well be dependent on circumstances, provided these circumstances are publicly verifiable. Thus it may readily be granted that a disabled person has a lower living standard than a fully healthy person commanding the same resources, because the disability is a publicly verifiable circumstance. Amartya Sen has argued that this shows that the concept of the standard of living must be concerned with evaluating outcomes rather than control over resources,11 but I hope it is by now clear that, on the present account, it need show nothing of the kind. If standards of living were concerned with evaluating outcomes in their entirety, we should be obliged to state that whenever two individuals derive different levels of happiness from a given (p.401) endowment of commodities, their standards of living are different. The present theory differs from Sen's in allowing only some reasons for the divergence of utility levels to count in standard of living comparisons—namely, those that are sufficiently publicly observable to be the basis of a contract. Other claims, such as that one individual may be of a cheerful disposition while another is somewhat morose, may be true without being relevant to the standard of living.
So far a fairly stark contrast has been drawn between those components of the good life for individuals that are, and those that are not, society's business. But it is evident on reflection that there exists a spectrum of possibilities rather than a sharp division. First of all, there is no sharp line between what is and what is not a publicly verifiable event, certain kinds of mild mental disability being a good example of a borderline case. There is nothing wrong with the idea that our conception of the standard of living might be sensitive to developments in diagnostic medicine, for example. But the emphasis on verifiability will tend to impose a burden of proof on borderline cases proposed for inclusion. Two individuals commanding the same resources will be held to have the same standard of living unless a convincing case can be made for their differing in some verifiable respect that might itself be the subject of a contingency clause in a social contract. So two things will need to be established: first, that the difference is verifiable and, second, that it would itself be relevant to the contract. Consider the case of differences in sex: these are (normally) readily verifiable. But are they such as to justify the claim that men and women with the same command over resources differ in standard of living, in a way that society might perhaps wish to correct? This will depend on the degree to which inequalities between the sexes consist of resource inequalities as opposed to inequalities in other respects, and the answer to this is likely to differ between societies and over time. If existing inequalities consisted primarily of unequal access to resources, there would be a less compelling case for adjusting living standard measures to reflect gender differences for given levels of command over resources.
A second kind of borderline problem concerns what kinds of service activity could be the subject of a contract. The problem is most perplexing within associations that are either not contractual or only partly contractual, like families. But here we may distinguish between problems of inclusion and problems of allocation. On the present account the time spent on housework, for example, even though it is frequently performed unpaid and without even a verbal contract, would count as a (negative) component of the standard of living, because it is the kind of activity that households need to perform and some reduction in which they would usually be willing to contract to achieve. Some problem of inclusion still arises because of the difficulty of distinguishing some such activities from those considered hobbies, where the activity is itself valuable. If, for example, I am a keen gardener (the example is hypothetical) I may not be willing to contract for anyone else to do my gardening for me, because it would then in some sense not be my garden any more. But in general there is no doubt that one significant component of poverty in many countries is the (p.402) time required for unpaid household tasks, a component that the familiar measures of the standard of living routinely ignore.
Separate from the problem of inclusion is a problem of allocation that may arise between members of the household. Suppose, for example, that household tasks are shared very unequally. We would be inclined to say that the standard of living of individuals within the household is to that extent unequal. But suppose it were objected that relations within the household are not an appropriate subject for a contract, that there could not be a social contract aiming to equalize time spent on household tasks? This is a difficult area, but if the premiss of the objection were accepted we should have to withdraw the judgement of inequality within the household. Of course, that premiss may be disputed: one of the features of modern feminism has been precisely to draw attention to the tacitly contractual way in which many aspects of relations between the sexes may be understood. Certainly, the theory being advanced here is in no way undermined by the reflection that the concepts of a contract and of what is fit material for contractual exchange are themselves variable and evolving over time.
As the example of household work suggests, the limitation of the notion of the standard of living to the sphere of commodities and services, broadly conceived, does not restrict it to marketed commodities and services. In fact, it includes many services of a very general kind. So, for example, the extent to which members of a society enjoy freedom of speech and of association is certainly a part of their standard of living. This is because these freedoms are safeguarded by the resources of society in that people are protected from assaults upon their exercise of free speech or free association. Such freedoms (and the resources to protect them) they could and would contract with each other to guarantee (though not, of course, to an unlimited extent). But the extent to which they use these freedoms to say and do worthwhile things is not part of their standard of living, though it may be an important part of what gives value of their lives. Two societies that devote equal effort and resources to television do not differ in their standard of living merely because one society is innovative and creative while the other broadcasts rubbish. One cannot contract to have good taste.
Nevertheless, non‐marketed resources do raise problems of a different kind, namely those of valuation. The valuation of marketed resources in measures of the standard of living is commonly made at their market price. How reasonable is this, and how should non‐marketed resources be treated? The argument so far has confined itself to considering what aspects of individuals' lives should be counted as part of their standard of living. For all practical purposes, though, what is required is not a complete description of the standard of living of each individual in society, but some aggregation or summary of the information this contains. Two kinds of aggregation are needed: first, aggregation over the many different components of an individual's standard of living, in order to compare individuals with each other. Second, aggregation over the many different individuals in society for the purpose of comparing one society with another or with (p.403) itself at different times or in different circumstances. The basic conceptual problems these forms of aggregation raise are very similar. All aggregation involves throwing away information. The best way to aggregate cannot be determined independently of the use to which the information is to be put. Different uses may require different methods of aggregation, and different degrees of aggregation: for example, for some purposes we may need to reduce the measure of the standard of living to a single scalar number; for others we may be happy to reduce it to a vector so that we can make comparisons in several dimensions. There is no reason whatever to believe that there is a single optimal aggregate measure of the standard of living either of an individual or of a society.
This point has been obscured by the implicit utilitarianism of much of the economic literature on the standard of living, and the related literature on such questions as the measurement of inequality. It follows from the standard axioms of individual rational choice that individuals choose their consumption so as to equate the ratio of the marginal utility of each commodity to the marginal utility of money with the price of that commodity in money terms. A measure of the standard of living can be constructed adding quantities of commodities traded and using market prices as weights. This measure represents the individual's consumption, and the sum of these measures across all individuals is the national consumption. If—and it is a big if—utility in this sense represents the appropriate maximand for society, and if all consumers have the same marginal utility of money, then a government that acts to maximize the national consumption will thereby maximize the utility of marginal changes in the availability of goods and services to the population.
There are numerous and well‐rehearsed problems with this line of reasoning, of which here are six:
1. One may deny that individual consumers maximise anything.
2. Even if they do, one may deny that what they maximize is utility in any interesting sense.
3. Even if it is, one may deny that society should be concerned only with utilities.
4. Even if it should, one may deny that society should be concerned only with the sum of utilities and not with their distribution.
5. Even if it should do that, one may deny that individuals have the same marginal utility of money.
6. Even if they do, one may deny that maximizing utility at the margin is equivalent to global maximization of utility.
The important point in this context is that only if it is true that society's goals can be appropriately represented by the maximization of some one quantity inherent in all commodities, will it follow that there is a single optimal way to aggregate the standards of living either of individuals or of societies.
This paper is concerned with the proper province of government policy, not with what the criteria of that policy should be. So the question of how to aggregate living standards, which requires answers to the latter problem, cannot be (p.404) addressed fully here. All the same, for many ordinary purposes market prices will provide a useful first approximation to a weighing system for those components of living standards for which markets exist. The reason for this is that if the standard of living comprises those commodities and services that can be exchanged by contract, one cultural indicator of their value will be the amounts of other commodities that citizens would require under the terms of some social contract in order to induce them to exchange. The justification for this need not be utilitarian (and typically will not be, since these amounts represent relative marginal, not average utilities). The justification might be directly contractarian: these amounts represent the relative weights that have been agreed as a result of the contractual process. And relative value by this criterion will be related to the amounts of other commodities that citizens in fact require in exchange under the terms of actual contracts. The two will not always or even usually be the same, partly for reasons of market imperfection that are familiar in the economics literature (such as the presence of externalities), and partly because the terms under which individuals make actual contracts are not the terms that would be specified by hypothetical social contracts. Actual contracts reflect substantial inequalities in power and wealth that social contracts might well seek to redress. But here as before the nature of real contracts between real individuals represents an important starting point in the analysis of what social contracts might be.
Non‐marketed resources naturally pose greater informational problems than do marketed resources, but the conceptual issues are very similar. Likewise, the valuation of rights such as the right to free speech depends upon what would be the terms of a social contract in which that right was guaranteed: how much of society's resources would be devoted to policing it in the optimal social contract? The more thoroughgoing the right the more expensive its defence, and the higher the valuation placed upon having that right. For those who blanch at the thought of ‘valuing’ rights in this fashion, I should repeat that the values thus derived represent purely a measure of the weight these rights have in that sphere of life that it is society's business to organize. Their overall importance in and contribution to the flourishing of individual citizens is something that society has no business to put a value upon.
3 Contractarianism and Pluralism
That there is a natural affinity between broadly contractarian views of the justification of social theories, and a pluralistic conception of the social good, has a well‐established pedigree in the history of social thought. The connection has not always been made explicitly, but one place where it is so made is John Locke's Letter Concerning Toleration. On the basis of the claim that ‘the commonwealth [is] a society of men constituted only for the procuring, preserving, and (p.405) advancing their own civil interests’, Locke argues that it can be no business of the state what religion its citizens believe in:
it appears not that God has ever given any such authority to one man over another, as to compel any one to his religion. Nor can any power be vested in the magistrate by the consent of the people, because no man can so far abandon the care of his own salvation as blindly to leave to the choice of any other, whether prince or subject, to prescribe to him what faith or worship he shall embrace. For no man could, if he would, conform his faith to the dictates of another. All the life and power of true religion consists in the inward and full persuasion of the mind; and faith is not faith without believing.
The irrelevance of contractual agreement to the procurement of spiritual well‐being contrasts with the more material sphere:
For those things that are necessary to the comfortable support of our lives are not the spontaneous products of nature, nor do offer themselves fit and prepared for our use. This part therefore draws on another care, and necessarily gives another employment. But the depravity of mankind being such that they had rather injuriously prey upon the fruits of other men's labours than take pains to provide for themselves, the necessity of preserving men in the possession of what honest industry has already acquired, and also of preserving their liberty and strength, whereby they may acquire what they further want, obliges men to enter into society with one another, that by mutual assistance and joint force they may secure unto each other their properties, in the things that contribute to the comfort and happiness of this life, leaving in the meanwhile to every man the care of his own eternal happiness, the attainment whereof can neither be facilitated by another man's industry, nor can the loss of it turn to another man's prejudice, nor the hope of it be forced from him by any external violence.12
Though Locke is not consistent in his demarcation between the contractual and the non‐contractual spheres (at some points treating it as coextensive with the distinction between this world and the next, and elsewhere exempting atheists from the right to religious toleration), the idea that there exists such a distinction is central to Locke's argument, as it is to the argument advanced here.
More recently than Locke, Ronald Coase's approach to the analysis of externalities in production and consumption13 illustrates another natural connection between contractualism and pluralism. Prior to Coase's work, it had been generally accepted that the appropriate response to an externality such as industrial pollution was to tax the polluting firm so as to bring private and social costs into equality at the margin. Coase pointed out that it was not evident that the polluting firm was the right agent to tax, because of the symmetrical character of externalities. For just as another firm downstream of the first might be said to suffer as a result of the actions of the polluting firm, so the polluting firm itself, if taxed, could be said to suffer as a result of the presence of another, a presence that made the pollution more costly in its effects (for the sake of the example, it is assumed that the damage due to the pollution is confined to (p.406) its effect on the downstream firm). The right solution, said Coase, was for the two agents to get together and bargain their way to an optimal outcome. It might be that this outcome would involve, not a reduction in the output of the polluting firm, but a (less costly) move by the other firm to another site where it would be unaffected by the first firm's activities.
One way to characterize Coase's argument is this: using taxes and subsidies to equalize marginal private and social costs will bring about local efficiency, in the sense that no reallocation of resources at the margin will make at least one agent better off without making any other worse off. But locally efficient points may not be globally efficient,14 because there may exist large resource reallocations (like moving whole factories) that dominate them. In order to ensure that globally efficient points are attained it is necessary for agents to bargain. The analogy with the present argument is that to make non‐local comparisons of living standards (comparisons, that is, between individuals or societies that are very different in their values and ways of life), it may be necessary to consider explicitly what would be the terms of a bargain or contract between the different parties. For if we adopt the measures (such as measures of real income) that would be appropriate to local comparisons (within societies, perhaps), we may make incompatible judgements about their standards of living.15 Each society may have inferior living standards when viewed by the local criteria of the other. To compare them we must consider those terms of a bargain on which the diverse individuals or societies could agree.
4 Living Standards and Pluralism
Economic development has always provoked anxieties about conflicts of value, from fears about the provocation of decadence by riches to more subtle worries about the effect of development on the diversity of cultural traditions. In these matters, economics is often believed, with some justification, to represent philistinism incarnate. The account of living standards outlines here suggests that this need not always be so. First, consider the question whether measures of living standards can be sensitive to the different values of different cultures. In a simple sense they can certainly be so, because living standards have nothing to say about many important areas of human life, which individuals and associations are therefore free to order according to whatever values they hold dear. A more interesting question is whether those elements of individuals' lives that count as part of their standard of living, and the valuation of those elements, might themselves be sensitive to the different values of different cultures.
There is every reason why they should be, and the extent of this sensitivity will depend on the purpose for which the measures of living standards are (p.407) developed. A measure of living standards developed for comparing countries may differ from a measure used in determining the policy of one particular country. This will be so for two kinds of reason. First, different conceptions of the good life will place different emphasis on the commodities and services necessary for their pursuit.16 Second, different cultures vary to an extent in their views as to what is and what is not the proper subject of a contract. Anyone who has watched an American broadcast of televangelism will be aware that religion has a much more explicitly contractual character in some communities than in others. The point about a social contract is that it must rest on the agreement of all parties. Social contracts within reasonably homogeneous societies will therefore be able to take much more for granted than a social contract involving the citizens of the whole world.
There is no answer to the question ‘Which social contract is the right one to make?’ Considerations of justice affect relations within villages, within regions, within countries, within the world. To each of these questions a different social contract may be appropriate, and from each of these contractual exercises a different conception of the standard of living may emerge. The degree of divergence between these will depend in a fairly evident way on the degree to which parties to a social contract are required to abstract from the particularities of their own position. If they did not abstract at all the exercise would be futile, but there are grave doubts about the coherence of their abstracting from every value that gives individual character to their judgements—a point that has been familiar since Hegel's criticism of Kant. This is too large a question to explore here, but it may be observed that unless it is possible to delimit a set of values that must be held by all rational beings qua rational beings and that are sufficiently rich to characterize a social contract, then the outcome of a social contract will be sensitive to the specification of the parties involved. So, therefore, will a contractarian conception of the standard of living. This no more condemns such a conception to relativism than a road map is made relativist because the way to Tipperary depends on where you start from.
Finally, to what extent must governments' attempts to enhance the standard of living bias economic development towards the commoditization of society? Commoditization is a difficult concept to understand and is used in a variety of senses.17 Most straightforwardly, it refers to the tendency of trade (either trade in general or monetized trade in particular) to penetrate social institutions. Existing measures of national income undoubtedly incorporate a bias towards the measurement of those goods and services that are in fact traded. In the theory advanced here such a bias is quite unwarranted, and involves essentially a confusion between the terms of a social contract and the terms of actual contracts. Nevertheless, it may still be objected that the emphasis in this theory (p.408) upon those components of the good life that could be the subject of contracts incorporates the same bias in a more subtle disguise.
There are two things to say about this objection. The first is that it sometimes incorporates the suggestion that a bias of this kind is in some way responsible for the widely observed tendency of human societies to mediate more and more of their consumption through the market as they become more prosperous. I do not think this is a suggestion to be taken seriously, at least not in this form. For one thing, it is not even true at all levels of development: for example, one may cite the growth of large corporations and their tendency to determine many aspects of their employees' lives either through fringe benefits or as an intrinsic part of their working conditions. Corporations can be viewed as institutions that supersede the market mechanism, as Ronald Coase long ago pointed out.18 But even barring this exception, the spread of the market is due partly to certain general facts about human beings, about their tastes and about the technologies of production, that make the benefits from specialization in production tend dramatically to outweigh any benefits of specialization in consumption. This case must not be overstated, since it remains one of the most devastating elements of Marx's critique of capitalism that it has tended so to exaggerate specialization in production as to cramp and wither the human potential of all those who perform society's most routine and repetitive tasks. Serious as this criticism is, the causes of such a phenomenon are independent of whether the province of society's concern be restricted to commodities and services in the way I have argued here.
The second thing to say about this objection is that we should indeed be legitimately concerned about the development of our lives in all sorts of ways outside the broad sphere of commodities and services discussed here. If economic development were to lead to all social relations' becoming infected with the mentality of the market, that would be a matter for the gravest concern. We should indeed worry lest prosperity lead to what one might with some licence call the death of the soul (though by any standards the effect of poverty on the soul is of far more pressing concern in the world today). What is denied here is that governments may be legitimately concerned with such matters. It is hard to conceive that a theory which brought these concerns within the province of governments could retain any pretensions to pluralism at all.
Coase, Ronald (1937). ‘The Nature of the Firm’, Economica, 4, 386–405.
—— (1960). ‘On Problem of Social Cost’, Journal of Law and Economics, 3, 1–44.
Dasgupta, P., and Heal, G. (1979). Economic Theory and Exhaustible Resources, Cambridge: Cambridge University Press.
Farrell, Joseph (1987). ‘Information and the Coase Theorem’, Journal of Economic Perspectives, 1, 113–29.
Hart, Keith (1985). ‘Commoditisation and the Standard of Living’, in Sen, 1985.
Hart, Oliver, and Moore, John (1988). ‘Property Rights and the Nature of the Firm’, STICERD Discussion Paper, London: London School of Economics.
Locke, John (1689). A Letter Concerning Toleration. Oxford: Blackwell. 1956.
Nagel, Thomas (1976). ‘Brain Bisection and the Unity of Consciousness’, in Jonathan Glover (ed.), The Philosophy of Mind. Oxford: Oxford University Press.
Parfit, Derek (1976). ‘Personal Identify’, in Jonathan Glover (ed.), The Philosophy of Mind. Oxford: Oxford University Press.
Rawls, John (1971). A Theory of Justice. Oxford: Oxford University Press.
—— (1985). ‘Justice as Fairness: Political not Metaphysical’, Philosophy and Public Affairs, 14, 223–51.
Roemer, John (1985). ‘A General Theory of Exploitation and Class’, in Roemer (ed.), Analytical Marxism. Cambridge: Cambridge University Press.
Sen, A. K. (1985). The Standard of Living (The Tanner Lectures). Cambridge: Cambridge University Press.
Starrett, David (1972). ‘Fundamental Nonconvexities in the Theory of Externalities’, Journal of Economic Theory, 4, 180–99.
Many people have made helpful comments on an earlier version, both participants at the WIDER conference and people at meetings in Cambridge and at the University of East Anglia. The present version has benefited from a number of suggestions made in Derek Parfit's original discussion of the paper, as well as from comments by Tom Baldwin, Jane Heal, Christine Korsgaard, and Martha Nussbaum.
(1) A soap opera on British television.
(3) Indeed, the exercise of the imagination is an essential component in our moral thinking.
(4) It is of course true that our initial reactions to improbable situations may well change on further reflection. It is then trivially true that initial judgements do not always constitute counter‐examples. The claim here is the stronger one that settled judgements of improbable situations do not always constitute counter‐examples.
(5) This is not the same as a procedure that weights cases according to their probability of occurrence. Some cases might have a low probability of occurrence but still be sufficiently like actual cases to be a valuable test of our intuitions. Others might be equally improbable, but so different in nature from actual cases as to be useless to us. Examples of the former occur when we ask questions like: ‘If it were possible, by pressing a button, painlessly to kill an unknown and distant individual with no ties to family and friends, would it be wrong to do so?’ An example of the latter might be the question: ‘Would loyalty be so important a virtue if human beings reproduced asexually and did not age?’
(6) For instance, it may seem logically possible that there should exist a creature all of whose beliefs were false, until one begins to reflect what circumstances could lead us to attribute entirely false beliefs to a language user. We would cease to call such a creature a language user long before that point had been reached.
(8) This does not imply that I think the contractarian perspective to be without problems. But the present argument is of strictly limited scope: if contractarianism is correct, and if pluralism is correct, then economic policy‐making can be consistently pluralist.
(10) The rider about scarcity reflects the fact that for things that are in indefinite abundance there would be no gains from co‐operation and no need to make contracts. Such things (oxygen in the atmosphere, for example) will not be components of the standard of living, though they may become so if the possibility of their scarcity arises (through air pollution, for example).
(15) This point is considered explicitly in Christopher Bliss's paper in this volume.
(16) The commodities and services discussed here therefore represent a much broader spectrum than Rawls's primary goods, which are defined as ‘things that every rational man is presumed to want’ (Rawls, 1971: 62).