A Very Short History of Distributive Justice

In A Short History of Distributive Justice Samuel Fleischacker (1994) discusses the intellectual history of “distributive justice” from Plato through Rawls (and a bit beyond). It is an interesting idea and it provides a useful perspective, but it is not altogether successful. It is not an easy task to condense the thoughts on (distributive) justice of Plato, Aristotle, Cicero, Augustine, Aquinas, Grotius, Pufendorf, Locke, Hume, Adam Smith, Rousseau, Kant, Babeuf, the utilitarians, Marx, the positivists, Rawls, and various post-Rawls critics into 133 pages.

His task is made easier by his central thesis: the current notion of distributive justice, which many take to have an ancient pedigree is, in fact, mainly a product of the last two centuries. According to Fleischacker, it did not exist in the ancient world. Moreover, he views the modern concept to require five premises (enumerated below). That thesis allows him to focus on a few aspects of the vast body of arguments he surveys. But the constraint he has placed on himself has led to a high degree of selectivity and simplification. He narrowly focuses on the intellectual history of distributive justice and seems to look at only those passages which mention “justice”, not at those which are concerned about “distribution”. As a result, his presentation is somewhat polemical, and a bit narrow. But he does make a number of interesting points.

The five premises required to arrive at the modern concept of distributive justice are, according to him:

• 1. Each individual, ..., has a good that deserves respect, and individuals are due certain rights and protections in their pursuit of that good;

• 2. Some share of material goods is part of every individual’s due...;

• 3. The fact that every individual deserves this can be justified rationally, in purely secular terms;

• 4. The distribution of this share of goods is practicable ...;

• 5. The state, ..., ought to be guaranteeing the distribution. (p. 7)

I will proceed by offering a very abbreviated and selective alternative interpretation of the writings Fleishacker has highlighted and develop a short outline of a different historical sketch. I emphasize four things Fleishacker ignores: (1) other aspects of Aristotle’s writings; (2) a different conception of Rawls’ contribution; (3) works in both the philosophical literature and in the social sciences that add just deserts and merit to a conception of distributive justice and tie these to the (re)development of democracy; and (4) how distributive justice may be related to social welfare in modern democracies.

Aristotle

Fleischacker starts by trying to demonstrate that Aristotle’s thought never incorporated premises of the sort he deems necessary for a modern conception of distributive justice:

Aristotle never put the problem of how to “allocate scarce resources” under the heading of distributive justice, nor did he regard need as the basis of any claim to property (p. 1)

In its original Aristotelian sense, “distributive justice” referred to the principles ensuring that deserving people are rewarded in accordance with their merits... (p. 2)

While Fleischacker is, strictly speaking, correct in the last assertion, he is correct only under the most generous interpretation of the former. A broader look at Aristotle’s thought shows that Fleischacker has missed the forest for the trees. He has looked only at Aristotle’s writings under the rubric “ethics”, and examined passages only explicitly using the term “justice”. An examination of Aristotle’s notion of the ideal state, which surely has a normative component and implicitly must represent a “just” arrangement of the state, presents an entirely different picture of Aristotle’s notion of how to distribute (at least some resources) fairly, and whether only “merit” implies entitlement.

Some quotes from Aristotle’s Politics (Great Books of the Western World, Volume 9, Aristotle II, 1952) may illustrate how his thought contained the seeds of most of the premises that Fleischacker deems necessary for the modern conception:

(In a democracy one aspect of liberty is that) ... every citizen ... must have equality. ... Another is that a man should live as he likes (Book VI Chapter 1)

(A) state is not a community of living beings only, but a community of equals, aiming at the best possible life. (Book VII Chapter 8)

These are not terribly far from Fleischacker’s first premise.

Elsewhere in Politics we find:

Let us then enumerate the functions of the state: ... First there must be food... (loc. cit. Chapter 8) (Later, in discussing distribution he says:)

Let me discuss the distribution of the land... for I do not think that property should be held in common, ..., but only that there be friendly consent that there should be a common use of it; and that no citizen be in want of subsistence. As to common meals, there is general agreement that a well-ordered city should have them ... They ought, however, be open to all the citizens. And yet, it is not easy for the poor to contribute the requisite sum out of their private means, and to provide also for their household. ... The land must therefore be divided into two parts, one public and the other private, ... part of the public land being ... used to defray the cost of the common meals... (Loc. Cit. Chapter 9)

This last excerpt seems to imply entitlement (of all citizens) to material support without regard to merit or via the invocation of religion, and it also demonstrates not only its practicability, but also the precise means of insuring adequate means for the redistribution.

So all of Fleischacker’s premises for the modern conception of distributive justice appear to have been met in Aristotle’s description of the ideal state. He has missed these aspects of Aristotle’s thought by his exclusion of the political. This seems to undermine the main line of his argument, but it doesn’t quite. After his slip regarding Aristotle, he gets on slightly firmer ground in discussing subsequent arguments which do not rely on Aristotle’s thought and proceeds along different lines. Nevertheless, his continued narrow reading of discussions of “distributive justice” make this latter part less useful than it otherwise would be.

Rawls: Fleischacker’s view and an alternative

I will not attempt to provide a full précis of Fleischacker’s argument, but rather will try to tie a number of his observations into a parallel discussion of the modern conception of distributive justice and provide an alternative analytic and historical sketch of its development.

For these purposes it is perhaps better to start at the end, rather than at the beginning of the story. For Fleischacker, John Rawls’ account of distributive justice is the first coherent and complete modern characterization of the concept. Indeed, he compares Rawls’ accomplishment in this sphere to those of Peano, Dedekind, and Cantor: three mathematicians who provided axiomatic bases for three number systems: the natural numbers, the real numbers, and transfinite numbers. He is lavish in his praise of Rawls for defining his principles and “...delineating their domain of application, with a mathematician’s precision.” (p. 113) He is equally impressed with Rawls’ use of an extant definition of societal well-being: Pareto optimality, which he characterizes as being a mathematical definition. He is further awed by Rawls’ use of other “mathematical” terms such as “classes”, “orderings”, “chain-connectedness”, “pair-wise comparisons”, and “close-knitedness” in producing “... a remarkably precise definition of “distributive justice” in its modern sense.” (p. 114). Although it is easily granted that Rawls’ work is the most influential contribution to the discussion of distributive justice in the last half of the 20th Century, Rawls’ argument is not an axiomatic mathematical derivation. His principles of justice cannot be deduced rigorously from the premises he offers, nor can they be derived from his justly (pun-intended) famous thought experiment of choice from behind the veil of ignorance.

Back in the mid 1980s, Joe Oppenheimer and I, (in part aided by Cheryl Eavey) devised and conducted several series of experiments designed to approximate the condition of impartial reasoning, the operative principle behind the veil of ignorance (Frohlich and Oppenheimer, 1992 summarizes a series of studies). Our objective was to see whether real individuals, operating under conditions approximating the veil of ignorance in this respect, would choose the difference principle over other competing principles when considering alternative distributions of income. Our results were not supportive of Rawls’ conjecture. The difference principle was the least favored of the competitors, and one “mixed” principle, maximizing average income with a floor constraint, fared qualitatively better than any other principle. In a substantial majority of the discussions in the experiments, “naïve” subjects were able to come to consensus on this principle. Listening to their discussions and reading the transcripts, it became clear why that principle was the most favored. It gave weight not only to the need of the worst off, but also to the entitlements of those who worked to produce income, and also to a concern that incentives be preserved in society in order to promote efficiency in production. The height of the floor (the income support level) moderated the tradeoff among the three competing ethical concerns. The higher the floor, the more weight was given to need as opposed to just deserts and incentives for efficiency, etc. References by different authors to all three of those concerns appear throughout the writings Fleischacker surveys.

Those two additional elements of distributive justice, just deserts and efficiency, are not featured centrally in Rawls’ conception. As a result, when they appear in the historical literature, they are noted in Fleischacker’s analysis as examples not conforming to the “modern” conception. For example, his reference to Aristotle, cited above, identifies just deserts as the prime component of the term “distributive justice” and so he interprets this usage as outside the modern conception of distributive justice. Whenever notions of desert and efficiency appear, they are treated as examples of why the historical notions are so different from the “modern” conception.

A moment’s reflection would surely lead one to the conclusion that any reasonable conception of distributive justice should take some form of just deserts into account and should also be concerned with efficiency. Indeed, the relative low priority accorded these considerations in Rawls’ arguments may be what provoked the arguments of some of his major critics (viz. Nozick’s Anarchy, State and Utopia, 1974).

Joe Oppenheimer and I were rather surprised at the direction and consistency of the results of our early experiments. We had occasion to chat with Rawls at a workshop at which they were presented. He said that the conditions of the experiments resembled what he had in mind when he developed the notion of the “veil of ignorance”. As to the poor performance of the difference principle in the experiments he said: “If the results hold up, it may be that the difference principle cuts across the grain of human nature.”¹ Those experiments have since been replicated in a wide variety of cultures (Cruz-Doña and Martina, 2000 – Philippines; Frohlich and Oppenheimer, 1992 – Canada, USA, and Poland; Jackson and Hill, 1995 – Australia; Oleson, 1997 – USA; Saijo et al., 1996 – Japan but see a recent dissenting view by Michelbach et al., 2003), indicating that the modern conception of distributive justice may be more complex than anticipated by Rawls and Fleischacker. Moreover, a growing body of evidence from experimental economics using other protocols and addressing related questions have reinforced the notion that in matters of distributive justice, need is not the whole story: merit and efficiency also must be taken into account (Kahneman et al., 1986; Konow, 1996, 2000, 2001).

A wider reading of Fleischacker’s own analysis shows how these other values have constituted essential elements of the historical debate, and how the debate can be interpreted more naturally in light of these additional concepts. Fleischacker, himself, elaborates the concept of merit in his presentation of the historical debate. Locke and Adam Smith are cited as the first authors to associate labor with merit. “Locke ... (provided a) powerful formulation of the intuition that labor constitutes the primary source of “merit” by which anyone can rightly claim to deserve material goods.”(p. 25). A long quotation from Smith implicitly notes the injustice of the material conditions of the common laborer in then contemporary Britain. These arguments are, on the one hand, used as the justification for re-allocating adequate resources to the poor laborer, and on the other as an argument for private property. Hume and Smith are characterized as presenting “... the system of private property as standing under a presumption of unfairness because of the way the poor suffer to provide luxury for the rich.” (p. 39). Fleischacker concludes that the laws, or writings on justice, did not recognize any rights of the poor to property, and that the efficiencies engendered by private property may have been necessary to provide even the meager livings which they enjoyed. So efficiency is noted, implicitly, as a value in the arguments of the time.

Kant is credited with introducing the modern notion that all human beings are equally worthy in the sphere of morality. Kant claims that we have a duty to aid other rational beings and even introduces obligations that presage Sen’s capabilities theory (1983): “(Kant) speaks of the obligations we all have to develop our ‘talents’ or ‘gifts’ ... This provides a moral foundation for ... the bringing of all human capacities, by way of political, economic, and educational progress, to their fullest form.” (p. 74). Here again, there is an implicit argument that all individuals need “primary goods” mixed with an implicit concern about the utilization of human capabilities to produce various products efficiently.

Fleischacker provides a brief overview of four streams of thought (19th century reactionaries, such as Malthus and Spencer, positivists, Marx, and Utilitarians) and concludes that Rawls undermined utilitarianism, accepted Marxian and positivistic critiques, and adapted Kant’s moral egalitarianism in developing his conception of distributive justice. But Fleischacker concludes that Rawls’ conception is the definitive modern statement. Rawls’ most interesting innovation is claimed to be “... that justice ought to be concerned only with the distribution of ‘primary goods’—goods that are necessary for the pursuit of practically any human end—and should set aside the question of what constitutes the ultimate human good.” (p. 111). And, as noted above, Rawls’ difference principle places inordinate moral weight on the plight of the worst off to the virtual exclusion of other values.

Democracy and the socio-economic context

There are, of course, alternative readings of the facts underlying the story presented by Fleischacker and alternative conclusions which can be drawn. One is that Rawls’ is not the definitive statement of the modern conception of distributive justice, but rather a major step towards the development of a deeper understanding of the concept. Fleischacker is a philosopher, and as such, he has naturally presented an intellectual history of the concept of distributive justice. A social and behavioral scientist would be more inclined to take into account the underlying social and economic conditions at play during the periods he covers, and to utilize contemporary empirical findings relevant to distributive justice. That different perspective and additional data could produce both different explanations of the concept’s trajectory and its likely future content.

For example, it could be argued that Aristotle’s notions of equality of individuals, their entitlement to basic support, and their rights to live as they like, flow from the underlining premises of the democratic polity in which he lived and of which he approved. At the simplest level, the notion of one citizen one vote accords each citizen equal status in the political arena.

But there are deeper epistemological assumptions that make their weight felt in the modern conception of democracy which may be thought of as clarifying and codifying what was implicitly hinted at by Aristotle. Recognizing and legitimating the individual as the appropriate judge of his own welfare implicitly involves some ontological and epistemological assumptions regarding the moral universe. Democracy ensures that political decisions are made by the people via some voting procedures. This implies that individual citizens have not only the right to but also the power to choose. The choice of the community’s citizens is deemed to represent what is good for the community. That communal good is decided by (or in a republican, or indirect democracy, it is seriously informed by) the decentralized choices of the individuals within the community. For that to make sense, it must be the case that the decisions are based on matters that can be potentially knowable by the voting citizen. This bespeaks an implicit epistemology regarding the Good: no-one, in general, is in a better position than the individual voter, to gain direct knowledge of what is good (at least for herself) based on observation, discussion, consultation, and inward reflection. This is not to say that the individual necessarily has any absolute knowledge of the Good. Rather, it is presumed to be better for mature individuals to make judgments on what is best for themselves because they can have (better) knowledge about their own welfare than anyone else.

In legitimating those judgments, democracy reinforces the normative assumption that individual welfare constitutes a major component of the Good. If the Good is knowable at all, it is the individual’s right to seek it for herself or to delegate the authority to recognize it to someone whom she reasonably believes has better tools to determine it (a doctor, a politician, etc.). This amounts to a literal empowerment of the citizen and recognition of the citizen as a moral agent.

The recognition that the individual’s judgment of her own Good is of value not only to herself but to society as a whole, which has an obligation to protect that conception, has implications. When coupled with the assumption that all individuals are, at some level, equal, these two premises provide a possible basis for arguments that the state should make some provision to ensure that each individual is in a position to pursue that Good, at least to a minimal extent.²

Perhaps that is why Aristotle’s thought (as reflected in the excerpts I have provided) has such a relatively modern cast to it. It may also explain why these notions appear to be absent in the subsequent writings that span the non-democratic periods into the late 18th century. A hierarchical society, in which social mobility (choice) is limited and everyone has a place on the static order, is not conducive to thoughts of the equality of all humans, nor of their rights to, or capabilities of, making choices regarding how best to lead their lives. Indeed, in economic conditions in which labor and capital are not mobile and a broad range of investment opportunities is not present, one might also expect less sensitivity to questions of productivity and merit: there is much less scope for individuals to demonstrate wide differentials in their economic performance. So discussions of distribution on the basis of any of the three principles might be less likely. The rise of democratic politics, along with the mobility of labor and capital in the 19th and 20th centuries may be an explanation for why these principles came to have more currency.

Democracy, distributive justice, social welfare, and framing

The introduction of merit and efficiency into the definition of distributive justice turns the spotlight off the poorest and onto the whole society. Virtually everyone in the society has some claim to merit and shares in the spillover from growing efficiency and prosperity. So this broader conception of distributive justice begs the question: What is the relationship between distributive justice and social welfare, writ large? And once that question is posed, the complexity of a definitive answer becomes clear. There is an apparent tradeoff between efficiency and need. If the support level in a society for those who have difficulty providing for themselves is too high, it creates a moral hazard: there is an increased incentive for the worse off to free-ride on society’s handouts and not to seek to contribute to society’s productivity. By the same token, the taxes on those who do work must be raised, impinging on merit, to support those at the bottom, which reduces their incentives to work. Can one assume that the just income distribution in a society is the one that provides the exact tradeoff among the three principles enunciated (or any other principles that may constitute distributive justice) to produce the greatest social welfare?

The answer to that question may be more than difficult. Following Arrow’s General Possibility Theorem (1963), it may be impossible to achieve a stable optimal state of social welfare in a well-ordered democracy. Yet recent findings from experimental economics and social psychology may point the way out of this seeming impasse. For roughly the past 20 years, experimenters in both fields have been identifying new elements in individuals’ preference structures which can be construed as pertaining to distributive justice. People seem to care not only about what they get in any situation but how their payoffs relate to what others get, and to the relative merit and/or need of the others. This bespeaks some non-separable preferences: preferences that depend not only on the individual’s payoff, but also on the pattern of payoffs and the particular agency whereby that pattern is achieved. Both Arrow and Sen have anticipated that an enrichment of the neoclassical model of self-interested utility functions along with the possibility of consensus on the relative states of welfare of different individuals might ameliorate the problem of a well-defined social welfare function (Arrow, 1977; Sen, 1977, 2002).

The existence of non-separable aspects of preferences have been observed in a variety of contexts: Frohlich and Oppenheimer (1984), Kahneman et al. (1986), Fleischacker (1994), Konow (1996, 2000), Larrick and Blount (1997), Eckel and Grossman (1996), Fehr and Schmidt (1999), Frohlich et al. (2004), Rabin (1993, 1998), Ruffle (1998), Cox et al. (2001). On the face of it, one would think that the added complexity of non-separable preferences might make it more rather than less difficult to integrate a notion of distributive justice with the achievement of an optimal state of social welfare. However, if individuals do place a value on distributive justice, this amounts to a sort of intrasubjective comparability of welfare. This allows individuals to measure the relative weights of personal material advantage in relation to others’ welfare in terms of an underlying concept of distributive justice. If there is agreement on a conception of distributive justice, that opens the possibility that consensus might be achievable regarding a most preferred social outcome in a fashion parallel to that envisioned by Arrow (1977).

Recently Wittman (2002) conjectured that a consensual commitment to equality might be adequate to overcome cycling in a majority rule democracy where the issue was the equitable distribution of monetary payoffs. He identified the intensity of preferences for distributive justice (in his case equality) necessary to achieve equal distributions of income under majority rule. The requirements were relatively severe, but not outrageous, and demonstrated the theoretical possibility of consensus on distribution.

Wittman dealt with a very simple characterization of distributive justice. Only a concern for their own material conditions and equality entered into individuals’ evaluations of income distribution. But his approach is generalizable to the multi-attribute characterization of distributive justice discussed above. In Frohlich and Oppenheimer (2006), we build on the growing evidence for a multi-attribute characterization of distributive justice and define a justice vector, which has a single dimension for each attribute. We introduce individual utility functions which include dummy variable elements for each attribute of distributive justice. By positing consensus on a justice vector, we show the possibility of achieving stable majority rule support for a just income distribution if a majority of individuals has a sufficiently high concern for justice.

We identify the conditions sufficient to guarantee a stable just outcome under majority rule, and show that these conditions vary as a function of the perspectives that individuals take on injustice. Under some perspectives, individuals need to have a stronger concern for justice in order for that aspect of their preferences to overwhelm the selfish component and turn their votes. The perspectives are variable across two dimensions: whether the individual evaluates injustice absolutely or relatively, and whether the individual takes a partial or impartial point of view in evaluating injustice. The former dimension distinguishes an individual’s evaluation of the justice of her own position in relation to some impersonal standard with that individual’s evaluation of her position in relation to the positions of others with whom she associates. Recent work has demonstrated the importance of this distinction and its relationship to injustice (Crosby, 1976; Runciman, 1966; Walker and Smith, 2002).

The second dimension of evaluating justice turns on whose interests count in the evaluation of injustice. Individuals can take an egocentric moral point of view and be concerned only with one’s own injustice (either relative or absolute). But it is also possible to take a broader moral point of view. This would involve taking others’ injustice into account directly. That would mean being concerned not only with the magnitude of one’s own injustice, but also with the injustice experienced by others. That would mean that the term to be entered into the individual’s utility function would be a function of the vector of injustices suffered by others.

The strength of individuals’ concern for justice needed to achieve a stable majority rule outcome varies as a function of those two dimensions. The most stringent requirement attaches to an absolutely partial point of view, and the least stringent to the relative, impartial point of view. The impartiality/partiality dimension has a stronger effect than the absolute/relative dimension, and from an impartial perspective the requirements for consensus are not particularly demanding. They are much less so than those required by Wittman (2002). So there is, perhaps, a possibility that a democratic society organized in terms of an underlying consensus on distributive justice could achieve a stable social welfare outcome that is not incompatible with optimal social welfare. But the argument needs to be developed further. And it must be noted that even if the underlying preferences and perspectives of the citizens of a democracy could in principle support a move toward more distributive justice, the structural aspects of the democracy, and the incentives of the political actors in that structure may thwart that movement.

The observation that the “perspective” taken on justice affects the intensity of preferences individuals must have to allow their sense of justice to operate effectively brings us to a final observation on the contribution of recent experimental findings to the distributive justice debate. Tversky and Kahneman (1979 and in many subsequent articles) have shown, famously, that individuals’ preferences vary as a function of the way in which decisions are framed.

The different perspectives on justice just sketched above are possibly subject to the same sort of framing effect, and the framing may turn on the structural aspects of the democracy. So, for example, the political debate in one democratic system may characterize the provision of health care as a problem of getting market efficiency. In another democracy the debate may be more explicitly in terms of distributive justice. This may well engender different perspectives on the part of voters. Put in terms of some of the discussion above, different perspectives on justice may be evoked in their utility functions, and the results of a vote in two societies having citizens with the same underlying preference structures could be quite different.

It may be important to evaluate democratic institutions on the basis of their propensities to frame issues in such a way that moral stances are evoked. As noted above, political actors can frame issues as having moral content or not, but it may be that certain institutional characteristics facilitate, or inhibit such behavior. The set of institutions (or discourses) that evokes only egoistic elements as opposed to those that also evoke normative elements and perspectives are likely to generate results less likely to reflect distributive justice. Clearly, the latter could be argued to be more desirable than the former. And this line of reasoning lays open the very difficult question: If different framings of issues can evoke different perspectives and preferences, then which precise perspectives and preferences are best? In other words, if mature individuals’ preferences are unstable in the sense that different cues can evoke different preferences, what appeals, or debates, or framing might serve to bring out the “best” in us to achieve the “best” outcome?

Taking note of the numerous paths left unexplored by Fleischacker, it would appear that a fuller history of the development of the concept of distributive justice is called for. Nevertheless, the historical perspective he presents offers a useful starting point. Looking ahead at the complexity of the issues emerging from our growing understanding of ourselves as cognitive and moral actors, the history of the debate on distributive justice is likely to be quite long.