Sen’s Idea of Justice: A Puzzle?

By Anjum Altaf

I started reading Amartya Sen’s latest book The Idea of Justice in which he suggests we reduce injustice in the world we live in rather than attempt to create an ideally just world – he characterizes the contrasting perspectives as ‘realization-focused’ versus ‘arrangement-focused’ approaches to justice. For South Asians, the parallels are two different concepts of justice from early Indian jurisprudence – niti and nyaya. The former relates to ‘organizational propriety as well as behavioral correctness’ whereas the latter is concerned with ‘what emerges and how, and in particular the lives that people are actually able to lead.’

The distinctions, and Professor Sen’s preference, are quite clear and one can agree or disagree with his choice. Here I am concerned with the example that Sen uses to motivate his argument and to explain why I find it puzzling. I would like readers to reflect on the example and to comment on its appropriateness.

Sen uses an illustration called ‘Three Children and a Flute’ to make the point that it is not possible to find an unambiguous principle of justice that everyone can agree upon. The illustration asks the reader to decide which of three children – Anne, Bob and Carla – should get a flute about which they are quarreling. Anne claims the flute on the ground that only she knows how to play it; Bob on the ground that he is the only one among the three who is so poor that he has no toys of his own; Carla on the ground that she has been working diligently for many months to make the flute with her own labor. None of the individual claims are contested.

The argument Sen makes is that these are competing claims for justice and that there are no obvious reasons for preferring any one over the others:

The general point here is that it is not easy to brush aside as foundationless any one of the claims based respectively on the pursuit of human fulfillment, or removal of poverty, or entitlement to enjoy the products of one’s own labor. The different resolutions all have serious arguments in support of them, and we may not be able to identify, without some arbitrariness, any of the alternative arguments as being the one that must invariably prevail.

What puzzles me is the following: Why does Sen pose this problem in terms of an issue of justice? To me it comes across much more as a problem of distribution that is made complex by the fact that we need to allocate one discrete commodity among three contenders. It is this discreteness of the commodity that turns the problem into one of choice, which then calls for a principle to govern that choice.

It is not clear to me why we should have our hands tied by the discreteness of the flute. If we relax this artificially imposed constraint we could consider a number of other solutions to the problem. For example, the flute could be sold and the proceeds distributed amongst the three claimants. This would not resolve the problem completely – Sen would surely ask for the principle that would govern the distribution of the proceeds – but it would certainly make the solution more tractable.

But even the limitation of discreteness need not preclude alternative solutions. The three contenders could agree that in the absence of any prior claims or rights, a just solution could be a fair lottery. Or they could agree on a cooperative solution in which each would get to keep the flute for a period of time with the order of the rotation determined by a fair lottery.

My argument is that this is not an issue of justice since no manifest injustice has been done to any of the three individuals. It is a simpler problem of distribution and it seems possible to find a cooperative solution if we do away with the stumbling block of the discreteness of the flute either by converting it into a divisible commodity (money) or by dividing its use over time. Once we do that we might even be able to find a single principle of fairness, e.g., egalitarianism, to govern the allocation of the divisible commodity. (This need not be the most efficient allocation. The more ambitious might try for allocations that make each claimant equally happy or maximize total happiness but these would again open up the debate over the merits of rival claims.)

I even do not see how the nature of prior claims, a dimension Professor Sen has ignored in his illustration, adds to the complexity of the problem. If Carla has made the flute and owns it, the other two have no claim to it. If Carla made the flute as a gift for the father who has left it as an inheritance, either the father’s will or the rules of inheritance would govern the allocation. Such rules vary across societies but in general have legitimacy amongst members of a society or there are accepted rules to resolve disagreements.

My question is as follows: Did Professor Sen choose a good illustration to motivate his argument? Should he have specified the source of the quarrel amongst the claimants? Is some prior information necessary to understand the context of the problem and its relevance to justice?

See Ten Unacceptable Things for our ongoing exercise based on Professor Sen’s suggestion.

Back to Main Page

Tags: , ,

26 Responses to “Sen’s Idea of Justice: A Puzzle?”

  1. ercelan Says:

    yet to read the book, described as sen’s attempt to warn that the best may be the enemy of the good.

    i am a rawlsian — hence also hostile to attempts, sen included, that dilute, deprecate, dismiss the basic postulate of rawls, in opposition to pareto, that justice advances only and only if the worst off progress towards the best off.

    for the islamists: eik roti to hamsaya ka adha haq.

  2. Vikram Says:

    Do you think maybe Sen was talking more about civil rather than criminal law ? The notion of injustice is much stronger in criminal cases.

    • SouthAsian Says:

      Vikram: Yes, Sen was clearly referring to an issue of distribution seeking a principle of ‘just’ distribution that would be acceptable to all and arguing that it was not possible to do so. My point was that the distribution problem in the illustration was made more complex than necessary by assuming the flute to be indivisible. If sharing had been allowed a solution would have become easier. The self-imposed restriction made it an artificial problem.

  3. Bobby Says:

    Well are you not changing the problem and thereby ruining the moral point he is trying to make with it?

    • SouthAsian Says:

      Bobby: I don’t think so. The problem remains exactly the same – one of distribution. I am asking if this is an appropriate example to illustrate the dilemmas of distributive justice. I am not disagreeing with the proposition that there are such dilemmas.

      • Bobby Says:

        Well; the way i understand the problem, its not one of distribution. Its one of defining what would constitute a just action. Or rather “the most just action” in a given circumstance. So the circumstances cannot be changed, which is what you are trying to do.

        The question as far as I understand is to define the most just action.

        Does Justice demand that (a) the toy be given to the “needy”

        Or (b) that the guy who made the toy be given the due of his labour, whether or not there is some one who “needs it more”

        Or (c) Does the guy who can make use of it and do something constructive out of it ,have a greater claim on it.

        I think the analogy is perfect, given that Sen wants to make a case that one cannot actually “do justice”, because there is no uniformly held definition of what constitutes justice in the first place. Rather what one should do is to try and remove manifest injustice.

        • SouthAsian Says:

          Bobby: This is what Sen has to say after he has described the illustration: “I also want to draw attention to the fairly obvious fact that the differences between the three children’s justificatory arguments… [are] about the principles that should govern the allocation of resources in general.” And what he is trying to argue is that each of the arguments makes sense and “we may not be able to identify, without some arbitrariness, any of the alternative arguments as being one that must invariably prevail.”

          I do not see the objective of the exercise as finding out what justice demands in this case or to remove a manifest injustice. Rather, it is to show that we cannot find a unique principle that can resolve the competing claims each of which has merit. I agree with this position. What I am saying is that looking for a unique principle might not be the right way to address this problem. Or, putting it differently, this might not be the best example to illustrate the point that Sen is trying to make.

          • Bobby Says:

            I think I am saying exactly what you quoted Sen as saying, if you replace:

            Given Circumstance —–> resources.

            “Most just action”——-> unique principle

            I still feel that you are changing the nature of the problem, by adding other alternatives and possibilities.

          • SouthAsian Says:

            Bobby: I agree – we are focused at different levels of abstraction. The most just action in a given circumstance relates to a very concrete situation while a principle governing the just allocation of resources is a more meta consideration. The action has to be subservient to some principle. If you have not already seen it, you would find Michael Sandel’s video (Justice: What is the Right Thing to Do?) of interest.

  4. Vinod Says:


    Every analogy has its limitations or assumptions at the boundary. Analogies are given to make a point. Sen’s point is that there are multiple principles of justice that collide with each other in many cases. Yes, it is hard to give an analogy that cannot be picked apart (like you have) to illustrate the point. But the question back to you is, does Sen have a point despite the flawed analogy?

    • SouthAsian Says:

      Vinod: I am sorry I had missed your question and came upon it only after reading the discussion again on receiving AaNoBhadrah’s comment. Yes, of course, Sen is making a very important point that it is not possible to find a unique principle that can resolve competing claims each of which has merit. It follows that the search for unique principles or truths can lead to further problems or to an infinite regress of the type described by Galen Strawson in his essay on Free Will.

      As for AaNoBhadrah’s comment, I have a lot of respect for Arun Pillai’s understanding of the topic and hope that he will provide some input for further discussion.

  5. AaNoBhadrah Says:

    Wonderful blog! Coming to your question, I don’t think your alternatives resolve anything. It just widens the question – is it just to sell the flute because it is discrete? Is it justice if Carla who made the flute has no say as to whether or how much money you can trade it for? Who are you to decide to sell it anyway? What if Carla reacts to this suggestion like the proverbial mother reacted to Solomon’s verdict? Would you still sell it letting Carla suffer through her tears? Is that justice?

    Further, I think you should concede that even if you think this is a problem of distribution (which I suspect it is not), you should concede the very real possibility that alternative solutions might actually be precluded. It is possible that the alternatives you suggest so considerably reduce the value of the flute to render it useless.

  6. Arun Pillai Says:


    In my view, this is a question of distributive justice. It is not justice as opposed to injustice, but the question of how to distribute something justly. I agree that South Asian’s response does not resolve the fundamental question Sen raises about the three incompatible principles because even if the good is made continuous rather than discrete, the same principles would apply and still be incompatible. Now there would be the additional possibility of some sharing but that does not change the fundamental question.

  7. SouthAsian Says:

    AaNoBhadrah: I have thought about your comment and Arun’s perspective on the question you have posed. I feel I would get a better sense if you described what you think is the nature of the problem posed by Sen if it is not that of distribution.

    I continue to feel the problem is one of distribution simply because Sen framed it in the context of an inheritance to be divided among children. Assuming that to be the case for the moment, I have the following observations:

    1. I agree fully with Sen that no over-riding principle can be found that can guarantee a fair distribution. My suggestions are not trying to refute Sen’s point and are not in search of an over-riding principle.

    2. I am following Sen’s distinction between niti and nyaya and wondering what can be done in this situation that would be feasible in the context of “lives that people are actually able to lead.” After all, lives go on and people are often able to find acceptable solutions to seemingly intractable problems.

    3. My suggestion is that a transformation that would render the asset continuous instead of discrete would widen the set of available solutions although they would not guarantee a solution.

    4. One such solution on which the three claimants might compromise is an equal sharing over time through rotation. This would offer a solution without requiring any one principle of distribution to be subordinated to another.

    5. This solution is better than one in which either no one gets anything or one person gets everything. Therefore, it is a rationally superior solution. However, we know that in issues of inheritance, rationality is often in short supply. That is also a part of the “lives that people are actually able to lead.”

  8. Arun Pillai Says:

    South Asian,

    I agree with the first four points you have made. To consider the fifth point, assume that any distribution is written (a,b,c) where a goes to Ann, b to Bob, and c to Carla. Then the distribution where no one gets anything is (0,0,0) and the distribution where one person gets everything would be either (1,0,0) or (0,1,0) or (0,0,1) where the “1” stands for the whole flute. The distribution where the flute is shared equally would be (1/3,1/3,1/3) which is (Pareto) superior to (0,0,0) but not to (1,0,0) or (0,1,0) or (0,0,1) because 0 < 1/3 < 1.

    Your principle of equal sharing is just one of many meta-principles one could use. There is no a priori reason each of the principles Sen has described should be given *equal* weight. Some substantive reason would have to be given for equality. For example, one could say: each of Sen's principles has the same validity and so needs to be recognized equally. Or one could use an analogue of the Principle of Insufficient Reason from probability theory.

    In my view, the field of ethics just has a lot of unsolvable problems. One reason for this is that the problems are typically stated in an abstract way. If one re-introduces the context in which the problem might actually reside in a real life situation, then sometimes (but not always) the context provides additional information that makes it possible to resolve the deadlock. This kind of ethics – which I call "situated ethics" or "situated justice" – is closer to Sen's realization-focused approach though it also has some features of the arrangement-focused approach.

    • SouthAsian Says:

      Arun: Doesn’t this point to the grievous hole in Paretian logic? If one person has all the wealth in the world (normalize it to 1) and everyone else has zero, wouldn’t any redistribution be Pareto inferior to the existing one? But does this abstraction have any relevance in a “situated” context? Is it a surprise that Paretian logic has found no resonance with ordinary people?

      There are other things that go into a utility function besides ownership of an asset that are associated with a sense of loss or regret. Even at an abstract level Kahneman-Tversky have tried to capture that in Prospect Theory – from any given reference point an equal loss yields greater disutility than the utility from an gain of equal magnitude.

      I am suggesting adding a variable for time. Then what seem two equivalent solutions become distinct in a “situated” context. The solution (1/3, 1/3, 1/3) is distinct from the solution [(1,0,0) for 4 months; (0,1,0) for 4 months; (0,0,1) for 4 months].

      I like what Nassim Taleb has to say about “naive optimization“:

      The reason I tell people to avoid attending an (orthodox) economics class and argue that economics will fail us is the following: economics is largely based on notions of naive optimisation, mathematised (poorly) by Paul Samuelson – and these mathematics have contributed massively to the construction of an error-prone society. An economist would find it inefficient to carry two lungs and two kidneys – consider the costs involved in transporting these heavy items across the savannah. Such optimisation would, eventually, kill you, after the first accident, the first “outlier”. Also, consider that if we gave Mother Nature to economists, it would dispense with individual kidneys – since we do not need them all the time, it would be more “efficient” if we sold ours and used a central kidney on a time-share basis. You could also lend your eyes at night, since you do not need them to dream.

  9. Arun Pillai Says:

    South Asian,

    The Pareto principle is admittedly a very weak criterion for justice if it is used in isolation. The problem is that there is no compelling argument that can be made for stronger criteria: this is precisely the point of Sen’s example – should the principle for distributive justice be based on fulfillment or poverty or labor? The three are incompatible and no obvious meta-argument is available to choose among them. Equality – either of distribution itself or of principles of distribution – that is, either as a fourth principle or as a meta-principle – is not always compelling either.

    If you introduce some reference point as in Prospect Theory – then it becomes even harder to argue for equality because the reference points of each child may be quite different. The reference point would be a part of the context. Thus, different contexts, each with its own setup, could result in different arguments for one or other principle. For example, the child with talent may already have some access to other means of fulfillment etc. So her claim might be ignored and one could then divide the time equally between the other two resulting in a distribution of (0,1/2,1/2) as optimal. Or maybe the poverty of the second child might also be abject in which case one might find a distribution like (0,9/10,1/10) optimal, and so on.

    I used the Pareto principle only to make the point that one cannot claim superiority for equality without some substantive argument. So in my view the way to use Pareto optimality is only when it definitely eliminates some candidate distribution; otherwise one needs some further argument to justify superiority. It is incorrect to draw the conclusion that two Pareto equivalent candidates are equally acceptable. One needs further argument. I think the Pareto principle which is a very simple and obvious idea has been misused in economics to ideologically justify unequal distributions of wealth. This use of the principle has nothing to do with the principle itself which is rarely applicable and so simply points to the need for other arguments.

    I find Taleb always too carried away by a single idea and always overstating it. All I would say is that the behavioral revolution should have come much earlier because it was obvious to everyone that the assumption of rational economic man or homo economicus was absurd and yet people kept going with it.

    • SouthAsian Says:

      Arun: There is theory and there is real life. As you mention, the theory of rational economic man is absurd and yet it rules in theory. There are many other such examples which include this aspect of the theory of distributive justice we are discussing. This perspective implicitly posits a philosopher-king who is going to decide which meta principle or which combination of principles is going to be the most just for society – Sen rightly moves away from this. In real life, people with an inheritance to divide don’t go about thinking of the claims of fulfillment or poverty or labor. Most of the time the dominant criterion is that of fairness and in the situated context varying bases for a determination are found (or not found). The articulation of fairness as a criterion is so prominent in all walks of life and yet it hardly finds any mention in economic theory. Akerlof and Shiller talk about this in their new book, Animal Spirits: How Human Psychology Drives the Economy and Why it Matters for Global Capitalism.

      Akerlof and Shiller overstate this idea of animal spirits just as Taleb overstates his ideas but economic theory is in such a mess that overstatement may be needed to draw attention before the alternatives are articulated in more detail.

  10. Arun Pillai Says:

    South Asian,

    Not having read Sen’s book, I do not know how he tackles the problem. How does one decide what is fair? Fairness is like justice: everyone would agree that the distribution should be fair/just but each person in society may put forward a different principle of fairness/justice. Take a more realistic example. Suppose there are only two children, one of whom has a high paying job and another who is ill and cannot work. How should the inheritance be divided? What according to you would be fair?

    I respect both Akerlof and Shiller so I will take a look at the book. I agree that economic theory is in a complete mess but people like Taleb don’t seem to me to help much because they tend to speak in extreme rather than moderate ways. But I haven’t read the Black Swan either – I am going by hearsay.

    • SouthAsian Says:

      Arun: There is certainly an overlap between justness and fairness but there are significant nuances that I find intriguing. Justice is something that is dispensed by a neutral third-party according to some principles; fairness is something that the parties involved negotiate or work out amongst themselves based on all the peculiarities of their situation and context.

      Thus, in everyday life workers don’t typically ask employers for a just wage but they do expect a fair wage and any perception of unfairness has negative implications for effort. On the other hand, people do ask for justice but not fairness from authority perhaps because they intuit that a just solution need not necessarily be a fair one. So the issue is one of framing and it helps to use the frame that best fits a particular context. Theory that is too abstract frames issues in ways that render them artificial and thereby misleading.

      The question that you pose about inheritance illustrates some of these considerations. An inheritance is not decided in this manner. Either, the person would have left a will incorporating his or her sense of a fair distribution based on knowledge of the context; or, failing that, the courts would decide based on precedent that reflects the consensus of the society concerned. It would not be dependent on my abstract notion of a principle of fairness or justice.

      The points that Taleb and Akerloff/Shiller make are fair ones although they do get carried away. In their defense, they are fighting a deeply entrenched and introverted priesthood that refuses to see or acknowledge any of the abuses of the Church.

  11. Arun Pillai Says:

    South Asian,

    I broadly agree with the distinctions you make between the uses of “just” and “fair.”

    Regarding the inheritance example, how would you suggest handling it if you were (1) the father making a will and (2) the brother with the job and (3) the brother who is ill? In another words, how should it be handled by the people concerned rather than an external theorist?

    It would be good to see the practical approach in action.

    • SouthAsian Says:

      Arun: This is a helpful question. Since the distribution of an inheritance is decided by the parent, I will deal only with the first scenario. If the parent dies without a will, the situations are much more messy and we can come to them later.

      Of course, every father is different and every father’s individual relationships with his children are different, so I can only speak for myself with the premise that in this hypothetical situation I am not biased towards either son for any extraneous reason. I will also assume that the amount of the available inheritance does not pose a binding constraint in the way of my desires, i.e., this is an unconstrained allocation.

      With the above caveats, I would divide the inheritance amount into two parts. The first (senior) part would constitute the amount needed to take care of the expenses of the child who is ill as long as he lives. This could be in the form of an endowment that would yield a monthly return sufficient to take care of the expenses of illness.

      The second (subordinated) part, the remainder, I would divide equally between the two sons provided that is the accepted norm in my community and if there is no reason to discriminate between the two due to considerations of age, future responsibility for the remaining spouse, etc.

      Three things need elaboration. First, the brother with the job might consider this allocation neither just nor fair but I think it would have the backing of the community. Second, more often than not, there is wisdom in community traditions that have evolved after repeated interactions with situations like this. Third, my thinking could be matched to a Rawlsian principle, i.e., that primacy must be given to the needs of the most disadvantaged before the needs of the others are considered. However, I own this principle and have internalized it; it is not being imposed on me by an external authority.

      Note that many refinements are needed in real life. For example, I could stipulate the distribution of the senior amount in case the ill child were to die or recover completely but these are matters of detail. Let us see what are the objections to this line of thinking and then elaborate the argument accordingly.

      Also note that we have had an extended discussion on the laws of inheritance in different societies (here and here) and it is clear that what is considered a fair distribution varies considerably across societies and that this perception changes over time.

  12. Arun Pillai Says:

    South Asian,

    I personally agree with the method you have used for the allocation. However, a theorist would say that all you have done is use Rawls’s principle. True, you have internalized it but your allocation is still describable theoretically.

    Also, the example was a relatively easy one. Sen’s example involves incompatible claims. How would you apply Rawls’s principle in his example? That is, how would you determine who is most disadvantaged – the child whose talent would be denied, or who is poor, or whose labor would go uncompensated?

    You also introduced a complicating factor – the approval of the community. What if the allocation does not meet with unanimous approval etc.?

    • SouthAsian Says:

      Arun: The fact that something sensible can be described theoretically should not be held as a strike against it. And the fact that the method adheres to a Rawlsian logic should not mean that I would mechanically apply the same logic to every situation. In the Sen example, I would use an entirely different method since none of the claimants is described as being disadvantaged relative to the others. I would propose sharing the flute over time by turns because that seems the most fair solution to me in the circumstances.

      I brought in the community only to stress the situated nature of these solutions. Communities face such situations repeatedly and the traditions that evolve embed, more often than not, what are considered the fairest solutions. However, if I felt there were circumstances that required my going against a tradition, I would feel free to do so. I would just keep in mind that the survivors would have to continue to live in the community.

  13. Arun Pillai Says:

    South Asian,

    Fair enough. I accept your solution. Let me just point out this is precisely what I take to be a situated approach to issues of fairness and justice i.e. that the circumstances or situation or context determines what principles are relevant. I don’t think this guarantees a fair solution for every ethical dilemma but it does seem to make more problems solvable.

  14. SouthAsian Says:

    Any comments

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: