Amartya Sen's Justice vs the Court's Justice

When Amartya Sen published The Idea of Justice in 2009, the book was read in India mostly as a long-running disagreement with John Rawls. It is that, but it is also something larger. Sen was arguing, across four hundred pages, against a way of doing justice that an entire profession in this country had taken as a moral premise. The profession is the Indian constitutional bar. The premise is that justice begins with the identification of an ideal arrangement of institutions and is realised through the policing of that arrangement by a Court. That is the project the Indian Supreme Court has run, with increasing self-confidence, since 1973. It is exactly the project Sen spent his career criticising. We owe both Sen and the Court the honesty of saying so.

What Sen actually rejects

Sen's opening move in The Idea of Justice is to split modern political philosophy into two traditions.¹ The first he calls transcendental institutionalism. It runs from Hobbes through Locke and Rousseau to Kant and finds its modern apex in Rawls. Its method is to ask what perfectly just institutions look like, build them on the foundation of a hypothetical original agreement, and then read off the standards by which actual societies are to be judged. The second tradition is comparative and realisation-focused. It runs from Smith and Bentham through Wollstonecraft and Mill to Marx and Sen himself. Its method is to compare the actual lives that actual institutions produce and to find ways of reducing manifest injustice without first agreeing on a perfect arrangement.

The distinction is not casual. Sen presses it because he thinks transcendental institutionalism has two structural problems that no amount of refinement can fix. The first he calls the feasibility problem: even under conditions of perfect impartiality, more than one set of principles can survive reasoned scrutiny. The flute parable is his compact illustration.² Anne can play the instrument. Bob has no toys. Carla made it. There is a defensible, impartial reason behind each child's claim. The transcendental method must declare one of them correct and the others mistaken. Sen says this is the wrong shape for the question. The right question is not whose abstract claim wins under ideal conditions, but how, in this society, with these children, we can move from a state of injustice to one less unjust.

The second is what Sen calls the redundancy problem. A complete theory of perfectly just institutions is not needed for most of what justice has actually had to do in human history. We did not need to agree on the contours of the just society to abolish slavery.⁷ We needed to agree that slavery was unjust. The decisive moves in the moral history of the modern world — the abolition of slavery, the extension of the franchise, the criminalisation of marital rape, the recognition that famine is preventable — were made through comparative judgment, not through the deduction of consequences from a first-position contract. Transcendentalism, on this view, is a long detour to conclusions reachable more directly.

Sen distils the distinction through an older pair of terms drawn from Sanskrit: niti and nyaya.³ Niti is the correctness of rules and institutions — the architecture of statutes and procedures and constitutional structures. Nyaya is the realised justice of human lives, the actual condition that the rules are meant to produce. Both matter, but the priority is reversed. Niti is the means; nyaya is the end. The transcendental method, in Sen's reading, forgets this. It treats niti as the whole subject.

The Court's project, named correctly

Now hold Sen's diagnostic next to what the Indian Supreme Court has been doing for the last fifty years.

The basic-structure doctrine, in its 1973 formulation in Kesavananda Bharati, is an act of transcendental institutional reasoning of the purest kind.⁴ It does not ask: in this case, before this Court, has injustice been done that can be remedied? It asks the prior question: are there ideal features of the constitutional arrangement that no Parliament, no matter how clear its mandate, can disturb? The seven-judge majority answered yes. Khanna J's reasoning, which tipped the balance, turned on a conceptual distinction between amendment and abrogation, between change that preserves the identity of the document and change that destroys it. As Madhav Khosla puts it in his Oxford Handbook chapter, this was a turn away from textual interpretation and towards a small number of fixed values around which the Constitution was to be anchored.⁶ Federalism. Secularism. Separation of powers. Judicial review. Rights. Democracy. A list, illustrative rather than exhaustive, of the conditions that any constitutional order worth the name must satisfy.

This is the Court's niti. Its job, on this view, is not to settle particular disputes through the application of general rules. It is to guard the architecture of perfect institutional design against the encroachments of imperfect politics. By Minerva Mills in 1980 this had become explicit.⁵ Judicial review of constitutional amendments was itself basic structure; the Court that polices the architecture cannot be excluded from policing itself. By IR Coelho in 2007, the doctrine had migrated from amendments into the Ninth Schedule's treatment of ordinary laws.⁸ By the time the National Judicial Appointments Commission was struck down in 2015, the doctrine could be deployed against any institutional design that diluted judicial primacy.

At each step, the Court reasoned the way Rawls reasons. It asked what features a competent constitutional order must have. It declared those features beyond the reach of ordinary politics. It assigned itself the role of guardian. The substantive claims were almost always defensible in isolation: of course free elections matter, of course judicial review matters, of course the rule of law matters. The form of the reasoning, however, was unmistakably transcendental. The Court was identifying perfect-justice features and then policing the constitutional order against deviations from them.

I am not the first person to notice this. Khosla observes that the basic-structure doctrine "asks what values and principles must exist for constitutionalism itself to exist" — that is, it asks a transcendental question, not a comparative one. What I want to add is the next sentence, the one we have been reluctant to say. The question Sen spent a career arguing was the wrong question to ask is the question the Indian Court has built its self-image on answering.

Where the Court's niti stops looking at nyaya

To see why the disagreement matters, consider what the Indian Court has actually delivered on the comparative dimensions of justice — the realisation side of the Sen/nyaya ledger — during the years it has been refining its transcendental machinery.

Take the criminal process. The basic-structure doctrine reaches the question of whether Parliament can dilute judicial review of detention. It does not reach the question of whether undertrials have spent more time in prison than the sentence they would have received on conviction — a condition that, by the National Crime Records Bureau's own numbers, describes a very large fraction of the Indian prison population. The first is a niti question and the Court has answered it with extraordinary doctrinal sophistication. The second is a nyaya question and the Court's answer, to the extent it has one, is a Public Interest Litigation here and a continuing mandamus there, with very limited follow-through on the actual life conditions.

Take fundamental rights jurisprudence. Maneka Gandhi and the post-1978 line of cases produced an Article 21 of impressive doctrinal architecture — the right to dignity, to shelter, to a clean environment, to privacy, eventually to be forgotten. The architecture is a niti achievement. The realisation has been uneven. The right to shelter coexists with the largest urban homeless population in the democratic world. The right to a clean environment coexists with Delhi's air. I am not blaming the Court for these failures — they are largely failures of executive capacity and legislative will — but I am pointing out that the Court's instinct, when faced with the gap between the architecture and the lives, is to refine the architecture further. To declare another right, locate another doctrine, identify another basic feature. The instinct is Rawlsian. It treats the gap as a defect of niti rather than as the substance of nyaya.

Or consider the doctrine's distributional silence. The basic-structure doctrine has, in fifty years, struck down precisely zero amendments because they failed to deliver on the Directive Principles. The Constitution's commitment to a fair distribution of the material resources of the community, in Article 39(b), is not basic structure. The independence of the judiciary, by contrast, is. We can have an argument about whether this hierarchy is correct. What we should not pretend is that the hierarchy was forced on the Court by the text. The text is silent. The hierarchy is the Court's own ranking of which features of the constitutional order matter most. It is a transcendental ranking, and it has come out where Rawls's would come out: with the procedural-institutional virtues at the top and the distributive-comparative ones below.

The honest counter-argument

The strongest defence of the Court's transcendentalism is not the one Court enthusiasts usually make. It is not that the doctrine has, in fact, prevented this or that bad amendment. It is institutional. India, the argument runs, does not have the conditions Sen's comparative method assumes. Sen's method works when there is, in a society, a thick and demanding public reasoning, a Smith-style impartial spectator who can be approximated by open democratic argument, a free press doing its job, a Parliament that takes its constitutional duties seriously, an executive that responds to deliberative pressure. Where these conditions exist, the comparative reduction of injustice can proceed without a Court playing guardian. Where they do not, the comparative method has nothing to grip. In their absence, transcendental institutionalism — a Court declaring certain features of the order beyond political reach — is not a luxury or a philosophical preference. It is the structure that holds the rest in place.

This is a serious argument and Sen himself would feel its force. He spends a substantial part of The Idea of Justice on the conditions for public reason, drawing especially on Smith's impartial spectator and the moral demands of reasoning across one's own boundary.⁹ He does not deny that institutional design matters. What he denies is that institutional design is where justice begins and ends. And the institutional defence of the Indian Court runs into two problems even on its own terms.

The first is that it proves too much. If the basic-structure doctrine is justified because no other check works, then we have not designed a constitutional order — we have appointed a guardian. The Constitution as drafted contemplated mutual checks: Parliament restraining the executive through confidence and budgets, the executive restraining Parliament through assent and reserve powers, the judiciary checking both through review of action and statute. The basic-structure doctrine, in its expanded form, has gradually substituted itself for the first two. We have a system in which Parliament is checked by the Court, the executive is checked by the Court, and the Court is checked by — the Court. This is not what the institutional argument promises. This is what happens when the institutional argument runs unchecked.

The second is that the institutional defence cannot, by its own logic, answer the question of when to stop. Every expansion of the doctrine — into ordinary statutes, into executive action, into the appointment of judges — has been justified by the institutional weakness of the alternative. There is no point in this argument at which the Court reaches an institutional check on itself that it must respect. The argument provides reasons to expand; it provides none to contract. This is what Sen meant when he warned that transcendental institutionalism, having identified itself as the guardian of perfect arrangement, can no longer take the comparative measure of its own performance. The guardian, on this view, is not subject to nyaya. It is its source.

Three concrete tests where the methods diverge

The disagreement is not an argument about labels. The two methods produce different answers on cases the Indian Court has actually decided, and the differences are illuminating.

Consider I.R. Coelho, the 2007 Ninth Schedule case. The Court's transcendental reasoning was elegant. The Ninth Schedule, originally designed to immunise agrarian reform from rights challenges, had over the decades been used to immunise more than two hundred and eighty laws of every kind. The Court held that any law inserted into the Schedule after the date of Kesavananda would itself be subject to basic-structure review. The reasoning is institutional and architectural. The Court is policing the integrity of the constitutional design against a vehicle that had been emptied of its original meaning. A comparative-justice question would ask different things. Of the laws actually in the Schedule, which ones produced lives that were better or worse than the lives they were designed to alter? Which Schedule additions, if pulled out, would actually shift the conditions of identifiable people? Coelho does not ask. It is content to clear the doctrinal ground. The architecture is restored to integrity; the lives are someone else's problem.

Consider next the death-penalty jurisprudence. The Court's transcendental architecture, from Bachan Singh through Machhi Singh to the long line of mitigating-and-aggravating cases, identifies a rarest-of-rare standard that is impressively reasoned in the abstract. The realisation has been less reassuring. The same offence, on similar facts, has produced both confirmation and commutation depending on which bench heard the appeal. A Sen-style critic would say: stop refining the doctrine; start asking what the doctrine has produced. If the same architecture yields outcomes that look like the lottery, the architecture is not the answer to the injustice — it is a description of the injustice. The Court's instinct, on this question and others, is to write the next Constitution-Bench judgment that clarifies the standard. Sen would say the clarification of standards is doing different work from the reduction of injustice.

Consider, third, the right to food. Article 21, on the post-1978 reading, contains a right to food; the Court has said so. The realisation has come through PIL orders directing the executive to operationalise food-grain procurement and distribution. This is closer to a comparative-justice exercise than the previous two, and it is no accident that it is also the work of the Court that is admired least by transcendentalists and most by people who measure justice in calories rather than in doctrine. The right-to-food jurisprudence is not perfect. It has been criticised for putting the Court in an administrative role it cannot sustain. But it is the part of the Indian Court's work that most resembles what a Sen-shaped Court would routinely do: take a substantive deprivation, name it, and use such legal tools as are available to move identifiable lives in the direction of less of it. The fact that the doctrinal world has often regarded this work as embarrassing — too administrative, too non-justiciable, too undignified — tells you which method the doctrinal world prefers.

If Sen and the Court's method disagree this sharply, why is the disagreement so rarely named? Three answers, none flattering to the profession.

The first is disciplinary. Sen is read in economics departments and in the better political-theory courses. He is not on most Indian constitutional-law reading lists. When he is read by lawyers, he is read as a moral philosopher rather than as a critic of judicial method. The constitutional bar's exposure to the Court's transcendentalism is by contrast complete: every doctrine memorandum, every senior counsel's leading question, every law-school course on constitutional law restates the basic-structure architecture as the master frame. The Court's method has the field to itself in the only room where its critics would have to be heard.

The second is professional. The transcendental method is the method of constitutional counsel. It is what they were trained in, what they are paid to perform, what their reputations are built on. A method that demoted institutional architecture to a procedural backstop and elevated the comparative reduction of injustice to the central task would re-rank the profession. The senior counsel who has spent a career mastering the basic-structure cases would be less central. The clinical lawyer who actually moves the lives of undertrials would be more so. Re-ranking is not how guilds reform themselves; it is what guilds resist. The Indian constitutional bar is a guild.

The third is institutional self-image. The Court has, over fifty years, told itself a particular story about why it matters. The story is that it is the guardian of the basic features of the constitutional order against the depredations of imperfect politics. The story is true in important episodes and false in others, but the story is also the Court's emotional architecture. To set Sen against the Court is to ask the Court to consider whether its self-image describes the most important work it does. Institutions do not give up their self-images easily. They give them up only when the gap between the image and the lives the image is meant to dignify becomes intolerable.

What a Sen-shaped Court would look like

A lawyer reading this far will want to push back. Comparative justice, the objection runs, is the work of legislatures and the executive, not the work of courts. Courts apply general rules to particular facts; that is what makes them courts. Asking a court to do the comparative reduction of injustice in the lives of identifiable people is asking it to do administration in robes. The basic-structure doctrine, on this view, is not a substitute for nyaya. It is the niti that makes nyaya in a constitutional democracy possible. Without a stable architecture there is nothing for legislatures and executives to do their distributive work inside.

The objection has force, and Sen would concede part of it. He is not arguing that courts should do everything. He is arguing that the question of what justice is should not be framed by what courts happen to be good at. The fact that judicial method privileges rule-and-architecture reasoning does not mean rule-and-architecture reasoning is what justice is. It means courts have an institutional bias towards a particular slice of the question. The trouble starts when that bias is mistaken for the substance of the discipline. Indian constitutional theory has, for fifty years, treated the slice as the cake.

There is a related worry — that comparative-justice reasoning, applied by judges, slides quickly into the kind of free-floating policy adjudication that the PIL critics rightly object to. Continuing-mandamus jurisprudence, sealed-cover procedure, the Court directing the executive on environmental policy or pandemic procurement — these are well-attested pathologies of judicial reach. If a Sen-shaped Court means more of this, the cure is worse than the disease. I take the worry seriously, but I think it gets the diagnosis backwards. The reason PIL has degenerated into administration-by-bench is not that the Court has thought comparatively; it is that the Court has tried to do comparative work without giving up its transcendental self-image. It still wants to be the guardian of the architecture. It still wants to write the next constitutional standard. But it now wants to do this while also moving lives. The combination is what produces the worst features of present-day PIL: doctrines that are too thin to constrain the bench yet too ambitious to admit failure. A Sen-shaped Court would not be more activist. It would be more honest about which of its work is doctrine-building and which is realisation, and clearer about the second's limits.

I am not arguing that the Indian Court should abolish the basic-structure doctrine. The Emergency, the 39th Amendment, the 42nd Amendment — these episodes show why a Court without the doctrinal tools to push back is a Court that has surrendered its function. The 1973 holding, narrowly read, has done its job. What I am arguing is that the Court should stop reading the doctrine as the principal substance of constitutional justice and start reading it as a procedural backstop.

A Sen-shaped Court would do four things differently. It would limit its transcendental reasoning to the cases where the alternative is constitutional collapse, and treat the rest of its docket as the comparative reduction of injustice through ordinary legal materials. It would treat the Directive Principles not as aspirations to be pushed onto the political branches but as evidence of what nyaya, in this constitutional order, was meant to look like. It would judge its own performance less by the doctrines it has elaborated and more by the lives those doctrines have actually altered. And it would be willing, when faced with the gap between architecture and realisation, to say that the right response is not a new doctrine but a sharper question.

None of this is a programme for tomorrow morning. The Court has fifty years of institutional investment in the transcendental method. The profession has its prestige tied up in mastery of that method. Sen is not, on most Indian law-school reading lists, a constitutional theorist; he is an economist that the constitutional lawyers have heard of. The result is that the most important challenge to the way the Court does justice has been kept at a polite distance from the way the Court is actually taught and discussed. I think we can do better than that.

What the disagreement is, finally, about

The Court treats justice the way Rawls did. Sen spent a career arguing this is the wrong way to think about justice at all. The disagreement is not between two doctrinal positions. It is between two pictures of what the activity called justice is for. The Indian Court has, in its strongest moments, asked: what arrangement of institutions would a just society have? Sen asks: which actual injustice can we, with the tools we possess, reduce in this lifetime?

These are not the same question and they do not yield the same answers. The first question makes guardians. The second makes citizens. The first question is about what a Court must do. The second is about what a society must do. The basic-structure doctrine answers the first. It does not, by its design, answer the second. We have spent fifty years treating its answer as the central achievement of Indian constitutional law. Sen would say that has been a mistake. Reading him alongside the Court's own cases, I find that I agree.

The honest claim is smaller than either side usually makes. The basic-structure doctrine is necessary as a procedural backstop. It is not a substitute for the comparative work of reducing actual injustice. The Court that delivered Kesavananda built a tool. The Court that holds it today has been mistaking the tool for the project. Sen, read carefully, lets us name that mistake without giving up what the tool was actually for. The basic-structure doctrine is one of the constitutional bar's proudest possessions. It deserves the conservation that comes with limits, not the protection that comes with reverence. Read alongside Sen, the doctrine becomes smaller in claim and steadier in use. That is the version of it worth defending.

If the Court took Sen seriously, the change would not be procedural. It would be a change in the questions the Court asks of itself at the end of a year's docket. The questions would no longer be about which doctrines have been elaborated, which architectures secured, which institutional encroachments resisted. They would be the harder, more uncomfortable questions about which specific injustices, in the lives of identifiable Indians, are less than they would have been without the Court's work. The doctrine would still be there, narrower than today and serving its original purpose. The self-image around the doctrine would have to go. That is a loss the Court can afford. It is also, I think, a loss it has put off for too long.

Get the week's biggest decisions, distilled.

A weekly digest for lawyers, students and the legally curious. Every Friday morning. Free.