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
Love of Symmetry
3. Democratic Values
Wanting equality for its own sake is no reason for wanting one equality rather than another.
One-man-one-pay and one-man-one-vote are not rules providing their own justification.
Everybody is bound to like ultimate goods like liberty, utility or justice. Not everybody is bound to like equality. If the democratic state needs consent and obtains some by producing some equality (a rather summary description of one type of political process, but it will have to do for my present purpose), it is the function of liberal ideology to inculcate the belief that this is a good thing. The high road leading to harmony between state interest and ideological prescription is to establish a deductive link, a causal relation or a reciprocal implication between ends which nobody disputes, such as liberty, utility and justice on the one hand, and equality on the other. If the latter produces the former, or if the latter is indispensable for producing the former, it becomes a simple matter of consistency, of plain common sense, not to dispute equality any more than one would dispute, say, justice or well-being.
Hearsay has it that there
are such deductive links: that freedom presupposes an equal sufficiency of material means; that social welfare is maximized by redistributing income from rich to poor; or that rational self-interest induces people unanimously to mandate the state to look after the least privileged. On examination, however, the detailed arguments from which the hearsay is distilled, prove unsuccessful. Like most hearsay, they have influence without quite silencing controversy and doubt. Far from establishing its universal validity to which men of good will cannot help but agree, it leaves the ideology vulnerable just as a religion which has the misplaced ambition of claiming the validity of logical deduction or scientific truth for its beliefs, is vulnerable. A less ambitious way, invulnerable to refutation, is to postulate that people do like equality for its own sake (so that its desirability need not be
deduced from the desiredness of anything else), or at least they would if they recognized its essential character.
People love symmetry, their senses expect it, they identify it with order and reason. Equality is to a system of rules as symmetry is to a design. The essence of equality
is symmetry. It is the basic presumption, it is what people visually or conceptually expect to find. For asymmetry as for inequality, they naturally look for a sufficient reason and are disturbed if there is none.
This line of reasoning tells people that it is inherent in their nature to approve of such rules as one-man-one-vote, to each according to his needs and the soil to him who tills it. In each of these rules, there is a clear symmetry which would be spoilt if some men had two votes and others one or none, if some (but only some) were given more than their needs and if some land belonged to the tiller and other land to the idle landlord.
However, if the choice is not between symmetry and asymmetry but between one symmetry and another, which is it inherent in human nature to prefer? Take the design of the human form, which must accommodate two arms and two legs. The arms can be placed symmetrically on either side of the spine, or symmetrically above and below the waist, and so can the legs. Between vertical and horizontal symmetry, which is right? A human figure with two arms on the right shoulder and hip and two legs on the left shoulder and hip would strike us as rather off-putting, not because it was asymmetrical (it would not be), but because its symmetry violated another to which our eye has become accustomed. Similarly, the preference for one order over another, one rule over another, one equality over another does not in any obvious manner spring from the depths of human nature, even if the preference for order over disorder may be plausibly held to do so.
The choice of a particular order, symmetry, rule or equality over its alternatives needs either habit, custom, or the force of substantive argument to explain it; if it is the former, political theory gets swallowed up in history (which might be a well-deserved fate) and if it is the latter, we will be back to square one, making
derivative cases for a liberty-securing, a utility-maximizing or a justice-dispensing equality rather than proving the claim that equality
for its own sake is intrinsically desirable.
It is worth spelling out that one equality crowds out another and that, as a corollary, the resulting inequality can always be said to have some equality as its reason and indeed its justification. (The adequacy of such a justification may have to be established, but this is very different from establishing the superiority of equality over inequality.) Take, for example, one of the central preoccupations of egalitarianism, the relations of symmetry or otherwise that prevail between workers, work, pay and need. One possible relation is equal pay for equal work, an equality which can be extended into the proportionality that more or better work should earn more pay.
*50 If this rule is good, it is a sufficient reason for inequality of remunerations. Another rule which suggests itself is to keep symmetry, not between work and pay, but between work and the satisfaction of the worker’s needs; the more children a worker has or the further away he lives from his place of work, the more he should be paid for equal work. This rule would yield unequal pay for equal work. Further “dimensions” can always be invented so that symmetry in one implies asymmetry in some or all the others, e.g. the importance or responsibility of the work done. Equal pay for equal responsibility will then (except for cases of purely accidental overlap) generally displace the equality between any two of the remaining characteristic dimensions of the relationship between worker, work, pay and need.
This logic is agreed by Marx to be valid up to and including the “first phase of communist society” (though, to cheer up last-ditch egalitarians it ceases to be valid in the second phase):
The right of the producers is proportional to the labour they supply…. This
equal right is an unequal right for unequal labour. It recognizes no class differences, because everyone is only a worker like everyone else; but it tacitly recognizes unequal individual endowment and thus productive capacity as natural privileges.
It is, therefore, a right of inequality, in its content, like every right. Right by its very nature can consist only in the application of an equal standard; but unequal individuals (and they would not be different individuals if they were not unequal) are measurable only by an equal standard in so far as they are brought under an equal point of view, are taken from one definite side only, for instance, in the present case, are regarded
only as workers and nothing more is seen in them, everything else being ignored. Further, one worker is married, another not; one has more children than another, and so on and so forth. Thus, with an equal performance of labour, and hence an equal share in the social consumption fund, one will in fact receive more than another, one will be richer than another, and so on. To avoid all these defects, right instead of being equal would have to be unequal.
But these defects are inevitable in the first phase of communist society…. I have dealt… with “equal right” and “fair distribution”… in order to show what a crime it is to attempt… to force on our Party again, as dogmas, ideas which in a certain period had some meaning but have now become obsolete verbal rubbish… ideological nonsense about right and other trash so common among the democrats and French Socialists.
Quite apart from the analysis so far given, it was in general a mistake to make a fuss about so-called
distribution and put the principal stress on it.
True to form, clearer and more to the point, Engels blurts out:
The idea of socialist society as the realm of
equality… should now be overcome, for it only produces confusion in people’s heads.
Take two “dimensions” of comparison, like pay on the one hand, and the return on investment in education on the other. If pay in every job is equal, the return on the cost of getting educated for a particular job must be unequal (if educational requirements for various jobs differ, which they often do), and
vice versa. These two equalities are mutually exclusive. Asked to choose the more egalitarian of the two alternative rules, many if not most people would name one-man-one-pay, rather than one-education-one-pay. There may be a multitude of good reasons for giving priority to the one or the other; but it seems impossible to claim that love of symmetry, order and reason can weigh in favour of either one. The symmetry between education and pay (the neuro-surgeon getting far more than the car-wash attendant) and the symmetry between the man and the pay (neuro-surgeon and car-wash attendant both getting a man’s pay), cannot be ordered in terms of their greater or lesser symmetry, order or reasonableness.
When one equality, symmetry, proportionality, can only prevail at the cost of upsetting another, equality itself is patently useless as a criterion for giving precedence to one or the other. Love of equality is no better as a guide for choosing between alternative equalities than love of children is for adopting a particular child. The appeal of rationality merely calls for some order and not for one particular order to the exclusion of another. This has been put with great clarity by Sir Isaiah Berlin in his 1956 essay, “Equality”: “unless there is some sufficient reason not to do so, it is…
rational to treat every member of a given class… as you treat every other member of it.” However, “since all entities are members of more than one class—indeed of a theoretically limitless number of classes—
any kind of behaviour can be safely subsumed under the general rule enjoining equal treatment—since unequal treatment of various members of class A can always be represented as equal treatment of them viewed as members of some other class.”
Symmetry requires that all workmen be paid the same living wage; among “workmen” there are “skilled men” and “unskilled men,” and among “skilled men” there are hard workers and loafers, long-service men and newcomers, and so forth. Enough heterogeneity can be found within the “workmen” category for reasonable men to hold that the initial rule of equality between workmen, or simply men, should be replaced by other rules of equality between skilled workmen with equal length of service, equal industry, etc. each rule establishing equality within the class to which it relates. While one can break up any class into any number of other classes, the substantive reason for breaking up the class “workmen” and replacing one equality with several, is that the class is arguably too heterogeneous and a more
nuancé classification corresponds better to merit and yields more rational equalities.
But this is just our say-so; another reasonable man might argue the opposite; we would both be displaying Berlin’s “love of order,” the sense of symmetry which is the basis of the presumption for equality. We say “black” and he says “red,” and no third person called in to adjudicate can refer us to some mutually agreed criterion which will help decide which of the equalities we champion is more rational, more symmetrical.
Berlin warns that since one can always find a reason for permitting an inequality, the rational argument for equality is reduced to a “trivial tautology” unless the argument comes complete with the reason to be admitted as sufficient.
*54 This is his typically courteous way of saying that the rabbit has to be put in the hat first. What reasons anyone finds sufficient for overruling one equality in favour of another depends obviously on his value judgments, of which his conception of justice will form a part; for it is now surely clear that the application of preference-less, value-free principles of rationality, order, symmetry, etc. can always be made to yield more than one, mutually conflicting rule of equality.
There are rules, such as a person’s right to his property, which are plainly anti-egalitarian in one variable (property) while egalitarian in another (the law). Most egalitarians would then hold that equality before the law must be upheld, but the law must be changed as regards property rights. This means that there must be no discrimination between rich and poor in the application of the law, and in order for this rule not to clash with the rule that all men should have the same property, the rich must be eliminated (without discriminating against them). While this promises a field day for pirouettes of sophistry either way, it is clear that for some unstated reason, priority is being given to one equality over another.
Another aspect of symmetry, that having to do with the relation between an activity and its inherent purpose or “internal goal,” has also been proposed as an argument leading to egalitarian results.
*55 If the rich buy medical care and the poor would but cannot, the purpose of medicine, which is to heal (rather than heal the rich) is deformed. It is irrational for medicine to heal rich people who are ill and not poor ones. Their needs with respect to medicine are the same and symmetry demands that they should receive the same treatment. To repair the irrationality, arrangements need be made to equalize rich and poor with regard to their access to the best medical care. If
only access to medical treatment is equalized,
the remaining riches of the rich may continue to deform the purpose of some other essential activity, which will create a need for equalizing with respect to
that activity, and so on,
until no rich and no poor are left.
But the rich’s being rich, and the poor’s being poor, may itself be found to correspond to the “internal goal” of some other essential activity, such as lively competition in the economy for material riches. Equalizing the prizes between winners and losers would defeat
its purpose and be irrational, etc. We now have one rationality entailing at least one irrationality, and while most egalitarians would have no trouble sorting this one out, their choice could not be based on the criterion of symmetry or reason. The “love of symmetry” argument and its developments, which show that equality is preferred for its own sake, depend on the alternative to equality being inequality. This is, however, a special case obtaining in artificially simplified situations only.
*56 If the alternative is generally
another equality, the argument is interesting but unimportant.
*57 Order in place of chaos may provide its own justification, but order as conformity to one rule in place of conformity to another does not entail the superiority of either rule; unless one rule can be proven to be “better,” more conducive to an agreed value than the other, the choice between them is best regarded as a matter of taste.
A population whose members are unequal to each other in an indefinitely great number of respects can be ordered in conformity to indefinitely many alternative rules, ordering them by the colour of their hair generally excluding, except by coincidence, a ranking by any other characteristic; symmetry between treatment and colour of
hair will imply asymmetry between treatment and
age or treatment and
education. However, there is usually quite wide agreement that for any given “treatment,” say the allocation of housing, only a handful of the indefinitely many dimensions in which applicants for housing may differ ought to be considered at all, e.g. rank on the waiting list, present accommodation, number of children and income. A rule of equality (proportionality, symmetry) can arbitrarily be laid down with respect to one of the four (generally entailing unequal treatment with respect to each of the remaining three), or a composite of all four may be formed with the aid of arbitrary weights, entailing unequal treatment with respect to any one but some rough-and-ready correspondence to the rational “sum” of all.
The agreement on what dimensions of a population ought to be considered at all for choosing a rule of equality, is a matter of the political culture. Thus, in a certain culture there may be wide consensus that steelworkers’ pay should not depend on how well they sing, yet students’ stipends should depend on how well they play football.
When a certain equality becomes an uncontroversial, generally agreed rule, the surrounding political culture can be taken to have become, in a sense, monolithic, for it has obliterated as irrelevant all the other dimensions, with respect to which alternative rules might have been formulated. One-man-one-vote in the democratic culture is the perfect example. It may be argued that each voter is a single individual, the rule of proportionality requiring that each should have a single vote. It may, on the contrary, be held that political decisions concern different individuals to different degrees (the paterfamilias vs the bachelor being a possible example), so that the proper rule should be: equal-concern-equal-vote, implying greater-concern-multiple-vote.
*58 On the other hand, one could maintain with the
Representative Government of John Stuart Mill that some people are more competent to make political judgements, including judging candidates for office, than others, which calls for the rule: equal-competence-equal-vote, greater-competence-more-votes. Such arguments used to find some practical expression in most nineteenth-century electoral laws with provisions for property and educational qualifications (contested as they were most of the time, not least by the “false consciousness” of the propertied and the educated). Obviously, the more the belief is eroded that some people legitimately have a greater stake in political decisions than others, or that everybody is not as good as everybody else at judging political issues and candidates, the less these inequalities can serve as relevant dimensions for ordering people’s voting rights. In the limiting case only one-man-one-vote is left, beginning to look very much like the self-evident, the only
conceivable symmetry of man and his vote.
By contrast, there is no consensus about the analogous role of one-man-one-pay, a rule calling for everybody getting the same pay either because they are all equal, one man being as good as another, or because their inequalities are not relevant to questions of pay. A great many rival rules continue to compete, suggesting variously that pay ought to be proportional to “work” or to “merit” (however defined), or to responsibility, seniority, need, educational accomplishment and so on, or possibly to hybrid composites of some of these or other variables.
It is anybody’s guess whether some or most of these rival rules will be obliterated from the political culture with the passage of time, possibly leaving a single surviving one which will then look as self-evident as one-man-one-vote does today. Liberal ideology, at all events, does not yet seem to have made its choice. Unlike socialism, which would give to each according to his
effort, pending the fullness of time when it can give to each according to his
needs (but which, in actual fact, simply gives to each according to his
rank), liberal thought is perfectly pluralistic in what sort of symmetries should prevail between people and their remunerations, finding much to be said for merit, responsibility, unpleasantness of the work and any number of other rules of proportionality, as long as it is
principles which prevail rather than the blatant ”
caprice of market contingencies.“
Where does this leave equality? The answer, I think, is a fascinating lesson in how a dominant ideology, totally unconsciously and without anybody’s directing design, adapts to the interests of the state. Liberalism only accords its respect to truly free contracts among equals, undistorted by “concealed duress” and “disguised oppression” (cf. pp. 120-1). Hence it would certainly not accept that people’s pay should simply be what it
is; it is deeply concerned by what it
ought to be, and its concern revolves around notions of justice and equity. However, as it tolerates a large number of mutually contradictory rules of equality, condemning few as unjust and inequitable, it will also tolerate a structure of remunerations where not only is everybody’s pay not equal to everybody else’s, but where it is not proportional either to any single most-logical, most-just (or perhaps most-useful, most-moral or most-anything) dimension of people’s inequalities. Whatever it will be, it will not be a “patterned” distribution.
This is just as well, for
if it were, what would be left for the state to
correct? Its redistributive function, which it must keep exercising to earn consent, would be
violating order and symmetry, upsetting the approved pattern in the act of levying taxes, giving subsidies and providing welfare in kind. On the other hand, if the pre-tax distribution is simply what it is without conforming to any one dominant norm of equality, the state has a great role to fulfil in imposing symmetry and order. This is why the pluralistic tolerance of a more or less patternless pre-tax distribution is such a precious feature of the liberal ideology. (By the same token, it is clear that the socialist ideology must not be pluralistic in this respect but must know right from wrong; for it is not serving a redistributive state which finds a pre-tax distribution determined by private contracts and
improves upon it, but rather a state which directly decides factor incomes in the first place and can hardly propose to correct its own handiwork by redistribution.
*60 “To each according to his efforts on behalf of society,” is the rule which must be claimed to characterize the whole distribution as decided by the socialist state, whatever other rules may shape it in reality. It is impolitic to invoke “to each according to his needs.”)
At the same time, liberal ideology fosters the claim that certain rules of equality are still better (more just, or more conducive to other undisputed values) than others, its preference being for distributions which favour the many over the few. If this claim sticks (though as I have tried to show on pp. 150-85, there is no good reason why it should), it is the warrant for redistributive moves which meet the democratic criterion of attracting more self-interested votes than they repel. It bears repeating that redistribution meeting the Janus-faced purpose of favouring the many and getting its instigator elected, is not necessarily “egalitarian” in the everyday sense of the word. Starting off with an initial distribution far removed from the equality of the one-man-one-pay kind, it will be a move
towards it; starting off with a distribution where such a rule is already being obeyed, it would be a move
away from it and towards some other kind of equality.
To conclude: analysis of the argument that love of symmetry, which is intrinsic in human nature, is tantamount to love of equality for its own sake, should have helped to focus attention on the multi-dimensional character of equality. Equality in one dimension typically entails inequalities in others. Love of symmetry leaves undetermined the preference for one sort of symmetry over another, one equality over another. Thus, one-man-one-vote is one equality, equal-competence-equal-vote is another. It is only in the limiting case, where all men are taken to have one (i.e. the same) competence, that they are not mutually exclusive.
Similarly, the rules “one-man-one-tax” or “from each, equally” (i.e. poll tax), “from each according to his income” (i.e. flat-rate tax) and “from each according to his capacity to pay” (i.e. progressive income tax with some putative proportionality between tax and the taxpayer’s residual means over and above his “needs”), are generally alternatives. Only in the limit where everybody’s incomes and needs are the same, are the three rules compatible.
There is no intelligible sense in which one of two alternative equalities is
more equal, or
bigger, than the other. As they are not commensurate (cannot be made to yield an algebraic sum), subtracting a lesser equality from a greater one so as to leave some residual equality is gobbledy-gook. Consequently, it cannot be affirmed that a policy change which enthrones one equality by violating another has, on balance, introduced more equality into the arrangements of society.
It makes perfect sense, however, to
prefer one equality to another and to defend this preference on the ground that
de gustibus non est disputandum (which is not the same as making an ethical judgement about their relative justice), as well as to allocate one’s own preference to that of the majority on the ground that respect for democracy demands it. As a practical matter, people do speak of social and political arrangements being (yes or no, more or less) egalitarian, and though it is not always very evident what they have in mind, we might as well suppose that most often it is this democratic criterion they are implicitly employing. None of this, however, makes the slightest contribution to establishing the claim (to which the “love of symmetry” argument is finally reduced) that what a majority will vote for also happens to be morally more valuable or corresponds more closely to the common good.
as well as for unequal work,” which seems contrary to the intention of the proposer. If he did
not want proportionality, he would have proposed “one man, one pay” regardless of the quantity or quality of the
Selected Works in One Volume, Moscow, 1968, pp. 320-1, italics in text.
Selected Works, p. 336, italics in text.
Concepts and Categories, 1978, pp. 82-3.
Philosophy, Politics and Society, 1962.
Equalities, 1981. Rae and his co-authors, very sensibly, want us to ask, not “whether equality” but “which equality?” (p. 19). They develop a “grammar” for defining and classifying equalities, and to provide some light relief, by permutation find no less than 720 sorts of equality (p. 189, note 3). However, they adopt the position that one situation can often, if not always, be diagnosed as more equal than another, i.e. that at least a partial ordering of social situations is possible, according to
how equal they are. My view is that ordering situations characterized by alternative equalities is inevitably done according to some other, often occult, criterion (e.g. justice or interest) and cannot be performed according to the criterion of equality itself.
“Concern” is an unsatisfactory explanation of why people bother to vote, but I am unaware of any more satisfactory rival ones; cf. the highly contrived “minimum regret” rule proposed by Ferejohn and Fiorina. For the basic statement that voting is irrational, see Anthony Downs,
An Economic Theory of Democracy, 1957, p. 274.
Abstention is, however, only a rough-and-ready approximation to the rule of greater-concern-more-vote. In this respect, Professor Lipset’s understandable mistrust of mass participation might find only very partial reassurance. For, although the extreme arbitrariness of one-man-one-vote is mitigated by the inclination to abstain of those who do not feel very concerned (and although their relative unconcern is a subjective feeling which need not coincide with the realities of their situation—perhaps they
should be concerned) the fact that the unconcerned
could vote if they felt like it, will still weigh in the political balance.
Suppose, for argument’s sake, that it is the
lumpenproletariat which habitually abstains. An electoral programme designed to attract the majority of the electorate
lumpenproletariat would always run the risk of being defeated by one designed to win over the majority of an electorate
lumpenproletariat, in case the latter were so roused that it did bother to go to the polls, after all. Hence, all competing programmes might take greater account of it than would be indicated by the paucity of the votes it habitually casts, and indeed by its apparent unconcern.
Anarchy, State and Utopia, p. 156. If all income from employment depended on the variable “work,” under the rule of proportional equality “equal pay for equal work, more pay for more work,” and all other income on one other variable, the distribution of total income would be “patterned.” If many contradictory rules are simultaneously at work and some incomes do not obey any obvious rule, the total distribution is “patternless”; at least this is my reading of Nozick’s use of this very suggestive and serviceable term.
Capitalism, Socialism and Democracy, 5th edn, 1976, pp. 198-9).