By Anthony de Jasay
Though this book leans on political philosophy, economics, and history, it leans on each lightly enough to remain accessible to the educated general reader, for whom it is mainly intended. Its central theme—how state and society interact to disappoint and render each other miserable—may concern a rather wide public among both governors and governed. Most of the arguments are straightforward enough not to require for their exposition the rigour and the technical apparatus that only academic audiences can be expected to endure, let alone to enjoy…. [From the Preface]
First Pub. Date
Indianapolis, IN: Liberty Fund, Inc.
The text of this edition is under copyright. Picture courtesy of the author.
- Authors Note
- 1.1 Violence, Obedience, Preference
- 1.2 Title and Contract
- 1.3 The Contours of the Minimal State
- 1.4 If States Did Not Exist, Should They Be Invented
- 1.5 Inventing the State: The Social Contract
- 1.6 Inventing the State: The Instrument of Class Rule
- 1.7 Closing the Loop by False Consciousness
- 2.1 Repression, Legitimacy and Consent
- 2.2 Taking Sides
- 2.3 Tinker's Licence
- 2.4 The Revealed Preference of Governments
- 2.5 Interpersonal Justice
- 2.6 Unintended Effects of Producing Interpersonal Utility and Justice
- 3.1 Liberalism and Democracy
- 3.2 Through Equality to Utility
- 3.3 How Justice Overrides Contracts
- 3.4 Egalitarianism as Prudence
- 3.5 Love of Symmetry
- 3.6 Envy
- 4.1 Fixed Constitutions
- 4.2 Buying Consent
- 4.3 Addictive Redistribution
- 4.4 Rising Prices
- 4.5 Churning
- 4.5 Towards a Theory of the State
- 5.1 What Is to Be Done
- 5.2 The State as Class
- 5.3 On the Plantation
The Revealed Preference of Governments
2. The Adversary State
Nothing distinguishes interpersonal comparisons of utility to determine the best public action from the government “revealing its preference” for certain of its subjects.
When the state cannot please everybody, it will choose whom it had better please.
While deriving the goodness of an action from that of its consequences is the feature that most visibly sets utilitarianism apart from explicitly intuitionist moral philosophies, I would argue that even this apartness is only virtual and that at the end of the day utilitarianism is swallowed up by intuitionism. The steps in this argument lead once more to the realm of unintended effects. The nominal priority accorded to individual values leads, through the subordination of the lesser utility of some persons to the greater utility of others, to the
exercise of state “intuition” to compare utilities, and to the enhancement of state power.
Defining good actions as those which have good consequences defers the question and asks at one remove, Which consequence is a good one? The received answer is partly dross: the word useful (utile) has pedestrian, mundane and narrowly hedonistic connotations which indicate a value system lacking nobility, beauty, altruism and transcendence. Some utilitarians, not least Bentham himself, bear the guilt for letting this false understanding get into the textbooks. Strictly, however, it ought to be discarded. In a suitably general form, utilitarianism tells us to regard a consequence as good if it is liked, no matter whether it is “pushpin or poetry” and no matter why; certainly not exclusively, and perhaps not at all, because it is useful. The liked consequence is synonymous with the satisfaction of a desire as well as with the fulfilment of an end, and it is “the measure of right or wrong.” The subject whose liking, desire or end qualifies a consequence, is always the individual. Arguments aimed at the good of the family, the group, the class or the whole society must first somehow satisfy individual criteria—they have to be derived from the several goods of the persons composing these entities. The individual person is sovereign in his likes and dislikes. No one chooses his ends for him and no one has a brief to dispute his tastes (although many utilitarians choose to restrict the domain of utility, in effect postulating that ends must be worthy of rational and moral man). Moreover, as it is clearly possible for individuals to like liberty, justice or, for that matter, divine grace, their attainment is productive of utility in the same way as, say, food and shelter. It is, therefore, possible to treat utility as a homogeneous resultant, a general index of end-attainments in which their plurality is in some unspecified manner synthesized in the individual mind. Such a view presupposes that there are no absolute priorities, that for each person every one of his ends is continuous, and suitably small bits of it can be traded off at some rate against bits of any other end. Though convenient, this treatment is somewhat arbitrary and possibly wrong. Besides, merging such ends as liberty or justice into an index of universal utility would conjure away some of the important questions political theory wants to ask.
(With the pretentiousness which makes the language of the social sciences sometimes so tiresome, “liking” is invariably transformed into its derivative “preference.” Texts on “social choice” usually talk of preferring, even when they do
not mean liking
better. This usage is now a
fait accompli and I will conform to it as long as I do not also have to say “betters” when I mean “goods.” It would be a relief, though, if accepted practice did not oblige us to employ the comparative where the simple affirmative would suffice.)
Private actions, often, and public ones nearly always, have consequences for several persons, typically for entire societies. Since the unit of reference is the individual, the measure of their goodness is the
algebraic sum of the utilities which they cause to accrue to each individual they affect. (Vaguer rankings of goodness can serve for very limited purposes only.) We are, in other words, dealing with the sum of the utilities gained by the gainers
less that lost by the losers. If the public good is to be maximized, the choice between mutually exclusive public policies must favour the one which causes the greater net positive utility. How do we tell?
The two easy cases, where we can simply ask all concerned and take their answer (or watch what they do in order to read off the preferences they reveal), are unanimity and so-called Pareto-superior choices, the latter being cases where at least one of the people concerned prefers (the consequences of) policy
A and none prefers policy
B. In all other cases the choice, whatever it is, could be disputed
either because some of the people concerned would opt for
A and others for
B or—doubly open to dispute and more realistic as a description of political life—because there is no
practicable way of reliably consulting everybody even on the most important choices that would affect them, nor of causing each person to reveal his preference in other convincing ways. Let me stress again in passing that the unit of reference is still the individual; he alone has desires to satisfy and hence preferences to reveal.
To consign utilitarianism as a political doctrine to the
oubliette, we can take the position that disputes arising out of conflicting views about the net balance of utility, are matters for knocking heads together, for there are no intellectually more acceptable means for resolving them. Consequently, unless some other doctrine is agreed for justifying its taking sides, the state ought to
lean over backwards to avoid putting itself in a position where it
must make choices pleasing some of its subjects and displeasing others. This leaning over backwards is, of course, the stance of the capitalist state which we have derived, from quite different premises, in chapter 1 (pp. 31-5).
The adversary state, on the contrary, positively needs occasions for taking sides, for reducing the satisfactions of some people, as this is the available coin with which to buy the support of others. To the extent that state policy and dominant ideology must advance more or less in step, dropping utilitarianism in the
oubliette could have left the democratic state temporarily out on a limb, to be rescued eventually by the rise of substitute doctrines. It is not altogether clear whether this has, in fact, happened. Many strains of political thought, while protesting to have broken with utilitarianism, reason by what amounts to all practical purposes to the utilitarian calculus. Perhaps only properly trained socialists (who do not deal in satisfactions), are not unconscious “closet utilitarians.” Many, if not most, liberals abjure interpersonal comparisons, yet advocate actions by the state on quintessentially interpersonal utility-maximizing grounds.
The uncompromising view of interpersonal comparisons that would deny the least place to political utilitarianism, is that to add one man’s quiet contentment to the exuberant joy of another, to deduct a woman’s tears from another woman’s smile, is a
conceptual absurdity which will not bear examination but, once stated, collapses of its own. When children learn that they must not try to add apples to pears, how can grown-ups believe that, if only they are performed carefully enough and buttressed by modern social research, such operations could serve as the guide to the desirable conduct of the state, to what is still fondly called ”
A revealing private confession by Bentham himself about the honesty of the procedure has been discovered in his private papers by Elie Halévy. Ruefully, Bentham declares, “‘Tis in vain to talk of adding quantities which after the addition will continue as before, one man’s happiness will never be another man’s happiness…
you might as well pretend to add twenty apples to twenty pears… This addibility of the happiness of different subjects… is a postulatum
without the allowance of which all practical reasoning is at a stand.”
*66 Amusingly, he was prepared to admit
both that the “postulatum of addibility” is logical wickedness
and that he could not do without it. This might have caused him to pause and reflect on the honesty or otherwise of the “practical reasoning” he wished to promote. However, there could be no question of letting “practical reasoning come to a stand.” He accepted pretence, intellectual opportunism ”
pour les besoins de la cause,” rather in the manner of the atheist priest or the progressive historian.
Conceding that the utilities of different persons are incommensurate, so that utility, happiness, well-being cannot be interpersonally integrated is, at the same time, an admission that social science operating with utilitarian premises cannot be invoked to validate claims about one policy being “objectively” superior to another (except in the rare and politically almost insignificant case of “Pareto-superiority”). Utilitarianism then becomes ideologically useless. If policies still need rigorous intellectual advocacy, they have to be argued for in some other, less convenient and less seductive doctrinal framework.
Against this intransigent position, three stands rehabilitating interpersonal comparisons can be discerned. Each is associated with the names of several distinguished theorists, some of whom in fact straddle more than one position. It is as arbitrary to confine them to a single stand as it is sharply to demarcate one stand from another. Partly for this reason, and partly to avoid giving offence through what can be little better than vulgarized capsules unfit to contain the whole of a subtle and complex treatment, I will refrain from attributing specific positions to particular authors. The informed reader will judge whether or not the resulting
roman à clef represents fairly the thinly disguised real characters involved.
One stand restoring to utilitarianism its role of judging policy, is that interpersonal comparisons are obviously possible since we are making them all the time. Only if we denied “other minds” could we rule out comparisons between them. Everyday linguistic usage proves the logical legitimacy of such statements as ”
A is happier than
B” (level-comparison) and, at a pinch, presumably also ”
A is happier than
B but by less than
B is happier than
C” (difference-comparison). A degree of freedom is, however, left to interpretation, which vitiates this approach. For these everyday statements can, for all their form tells us, just as well be about facts (
A is taller than
B) as about opinions, tastes or both (
A is more handsome than
B). If the latter, it is no use linguistic usage telling us that interpersonal comparisons are “possible” (they do not grate on the ear), because they are not the comparisons utilitarianism needs to provide “scientific” support for policies. An equally crucial ambiguity surrounds the piece of linguistic testimony that tends to be invoked in direct support of redistributive policies: “a dollar makes more difference to
B than to
A.” If the statement means that the incremental utility of a dollar to
B is greater than it is to
A, well and good. We have successfully compared amounts of utilities of two persons. If it means that a dollar affects
B‘s utility more than
A‘s, we have merely compared the
relative change in
B‘s utility (“it has been vastly augmented”) and in
A‘s (“it has not changed all that much”), without having said anything about
B‘s utility-change being
absolutely greater or smaller than
A‘s (i.e. without demonstrating that the utilities of two persons
are commensurate, capable of being expressed in terms of some common homogeneous “social” utility).
Another integrationist stand confronts the issue of heterogeneity, as it were, head on, by proposing what I would call
conventions for getting rid of it, rather as if Bentham had announced that he was going to feel free to call both apples and pears “fruit” and perform additions and subtractions in terms of “fruit-units.” These conventions can be regarded as non-empirical, not-to-be verified postulates introduced to round out a non-empirical circle of argument. It is, for instance, said that the utilities of “isomorphic” persons, who are identical in all but one variable (e.g. income, or age) can be treated as homogeneous quantities, and it is further proposed to treat certain populations for certain purposes as isomorphic. A different convention would have us regard everybody’s utility as inextricably bound up with everybody else’s
via a relation of “extended sympathy.” Yet another approach transforms (roughly speaking) the utility functions of different people into linear transformations of one and the same function, by taking out of the parameters of preferences everything that makes them different, and putting back the differences “into the objects of the preferences.” There is also a proposal (which I personally find disarming), to put in place of people’s actual preferences the “moral” preferences they would have if they all identified themselves with society’s representative individual. A somewhat comparable convention is to regard different persons as “alternative selves” of the observer.
These and related conventions are
an sich harmless and acceptable alternative statements of what
would suffice to legitimize the integration of different persons’ utilities, happinesses or well-beings. They can be paraphrased to read: “The welfares of different individuals can be added into a social welfare function if they are agreed not to be different individuals.” Such conventions may well command agreement as to the sufficient condition which
would make the summation of personal utilities legitimate. They are, however, not to be mistaken for ways to legitimize the summation if it was not legitimate
to begin with.
A (to my mind) diametrically opposed stand is fully to accept that individuals are different, but to deny that this must render social welfare judgements arbitrary and intellectually unclean. This stand, like the “linguistic” one, seems to me to suffer from the ambiguity that the judgements it makes and the decisions it recommends (these being possibly two distinct functions) may be either questions of fact or matters of taste, without their form necessarily telling us which they are. If they are matters of taste—even if it is “taste” educated by practice and enlightened by information—there is little else to be said. We are manifestly in the hands of the sympathetic observer and all depends on who has the power to appoint him. Claims that one policy is better for society than another will rest upon authority.
On the other hand, if they are to be understood as verifiable, refutable matters of fact, interpersonal comparability must mean that any difficulties we may have with adding up are technical and not conceptual; they are due to the inaccessibility, paucity or vagueness of the required
information. The problem is how to get at and measure what goes on inside people’s heads and not that the heads belong to different persons. Minimal, widely accessible information about Nero, Rome and fiddling, for example, is sufficient for concluding that,
for a fact, there was no net gain of utility from the burning of Rome while Nero played the fiddle. Progressively richer, more precise information allows progressively more refined interpersonal findings. Thus we move forward from the non-addibility resulting from sheer lack of specific data to an at least quasi-cardinal utility and its at least partial interpersonal comparison.
*67 At least ostensibly, the contrast with proposals to ignore specificity and strip individuals of their differences, could not be more complete. The proposal here seems to be to start from admitted heterogeneity and approach homogeneity of individuals by capturing as many of their differences as possible in pairwise comparisons, as if we were comparing an apple and a pear first in terms of size, sugar content, acidity, colour, specific weight and so on through
n separate comparisons of homogeneous attributes, leaving uncompared only residual ones which defy all common measure. Once we have found the
n common attributes and performed the comparisons, we have
n separate results. These must then be consolidated into a single result, the Comparison, by deciding their relative weights.
Would, however, the admission that this procedure for adding up utilities is intellectually coherent, suffice to make it acceptable for choosing policies? If the procedure were to be operated, a host of debatable issues would first have to be somehow (unanimously?) agreed by everybody whose utility gain or loss was liable to be compared in the operation. What distinguishing traits of each individual (income, education, health, job satisfaction, character, spouse’s good or bad disposition, etc.) shall be pairwise compared to infer utility levels or utility differences? If some traits can only be subjectively assessed, rather than read off from Census Bureau statistics, who shall assess them? What weight shall be given to each characteristic in inferring utility, and will the same weight do for people of possibly quite different sensibilities? Whose values shall condition these judgements? If some “equitable” way were unanimously agreed for delegating powers for taking comparative readings and setting the weights, the delegate would either go insane, or would just produce
whatever result looked right to his intuition.*68
The long and short of it is that objective and procedurally defined interpersonal comparisons of utility, even if they are modestly partial, are merely a roundabout route all the way back to irreducible arbitrariness, to be exercised by authority. At the end of the day, it is the intuition of the person making the comparison which decides, or there is no comparison. If so, what is the use of making intuitive interpersonal comparisons of utility so as to determine the preference ranking of alternative state policies? Why not resort directly to intuition for telling that one policy is better than the other? Intuitively deciding what had best be done is the classic role assigned to the sympathetic observer who has listened to the arguments, looked at the facts and then for better or worse exercised his prerogative. Who else is he, albeit at one remove, if not the state?
Failing unanimity on how exactly to perform interpersonal comparisons, different descriptions of the choice of policy are simultaneously possible. It can be said that the state, marshalling its statistical resources, its knowledge, sympathy and intuition, constructed measures of its subjects’ utilities, enabling them to be added to and deducted from each other. On this basis, it calculated the effect of each feasible policy upon total utility, and chose the one with the best effect. Alternatively, the state can be said to have simply chosen the policy which it thought was the best. The two descriptions are mutually consistent and one cannot contradict or refute the other.
In an analogous manner, the two statements “the state found that increasing group
P‘s utility and decreasing that of group
R would result in a net increase in utility” and “the state chose to favour group
P over group
descriptions of the same reality. Nothing empirical distinguishes the two operations they refer to. Whichever description is employed, by its choice the state will have “revealed its preference.” This is not to try and argue that all inquiry must stop at this point, for it is not done to question the causes of a preference. It is, however, a plea not to explain the state’s partiality by some futile hypothesis which can, by virtue of the irreducible arbitrariness of interpersonal comparisons, never be falsified.
The Growth of Philosophical Radicalism, p. 495, quoted by Lord Robbins,
Politics and Economics, 1963, p. 15; my italics.
Revealed Preference of Governments, 1980, ch. 6.
I have borrowed the title of this perfectly dispassionate book to head the present section because its unintentional black humour conveys so well what I take to be the irreducible core of the utilitarian solution. The only preference which is ever “revealed” in the maximization of social welfare is
that of the maximizer, of the holder of sovereign power over society.