NOTE-262: A Paradigm Shift in Theorizing about Justice? A Critique of Sen

Contribution

Valentini gives a page-referenced reconstruction of Sen’s methodological critique of Rawls in The Idea of Justice and the 2006 Journal of Philosophy article. She then gives a point-by-point Rawlsian rebuttal.

• Her thesis. Sen’s call to move from “transcendental institutionalism” to “realization-focused comparison” is unwarranted. “Most of Sen’s criticisms are in fact either based on a misrepresentation of the Rawlsian approach, or correct but of little consequence” (p. 2).
• What is needed instead. “not a paradigm shift, but a more nuanced understanding of the paradigm Sen and others criticize” (p. 2).

Key insight

Sen’s target conflates two things: a theory’s logical form (categorical, “X is perfectly just iff…”) and a theory’s practical idleness.

Rawls’s theory has the first. That is the only fair sense of “transcendental”, and “categorical” would be a better label (p. 6). But it does not have the second.

• It yields comparisons. Its principles already support many comparisons, through lexical priority and the inherently comparative difference principle (pp. 9–10).
• It is not tied to one set of institutions. It prescribes no single ideal institutional set: institutions vary with the society, for example affirmative action where there is prejudice (p. 6).
• It is revisable. It sits inside reflective equilibrium, which is exactly the revisable back-and-forth Sen says transcendental theory lacks (pp. 13–14).

Assumptions

• The target is Rawls’s method, not his substantive principles. She disclaims endorsing the latter (p. 14).
• Rawlsian theory aims to be action-guiding. So the “practical irrelevance” charge cannot be waved away as it could by a G. A. Cohen-style view that justice concerns what to think, not do (p. 5, citing Cohen 2003 and Valentini 2009).
• The circumstances of justice. Theories of justice exist because of scarcity and disagreement (the Humean circumstances; Rawls, TJ pp. 126–30) (p. 8).
• Late Rawls. Political Liberalism and The Law of Peoples are treated as part of the paradigm. They are used to answer the inflexibility charge (pp. 12–13, fn 44).

Key results

What The Idea of Justice claims, as Valentini reconstructs it (page references as she gives them)

This is second-hand for the book. Attribute it “per Valentini”.

• The central question and its two components. Contemporary theory asks “What is a just society?”. Those who centre that question are “transcendental institutionalists” (ch. 1).
  • “Transcendental” means aiming to identify a perfectly just society.
  • “Institutionalist” means establishing perfect institutional arrangements “without paying much attention to the conduct of individuals”. Real societies are unjust to the extent that they fail to match that institutional perfection (p. 3, citing IoJ ch. 1).
  • “Transcendental” is primarily “absolute”, that is, non-comparative (IoJ p. 6).
  • “transcendental institutionalism concentrates primarily on getting the institutions right, and it is not directly focused on the actual societies that would ultimately emerge” (IoJ p. 6, quoted p. 6).
• Lineage. The approach is traced to Hobbes, Locke, Rousseau and Kant. Rawls’s theory is “[t]he most powerful and momentous exposition” of it (IoJ pp. 7–8; see p. 8).
  • The marks Sen finds in Rawls: (i) a unique, definitive set of principles; (ii) principles that select institutions and do not apply to individual behaviour; (iii) a society satisfying them is perfectly just (p. 3).
  • Sen also “acknowledges that Rawls is not only interested in the right institutions, but also in the right behaviour” (IoJ pp. 7, 78). Yet he says that “[i]n the Rawlsian system of justice as fairness direct attention is bestowed almost exclusively on just institutions” (IoJ p. 67; also p. 46) (fn 26).
• Sen’s alternative. “Realization-focused comparison” (p. 2). It admires Rawls (IoJ acknowledgements and ch. 2, fn 3).
• Six complaints. The book lists six complaints against transcendental institutionalism (IoJ p. 90, fn 9). Valentini groups them into three.
• Complaint A: neither necessary nor sufficient for comparisons (IoJ pp. 98–106).
  • Not necessary. One can know that the “iniquities of hunger, illiteracy, torture, arbitrary incarceration, or medical exclusion” are injustices without an account of perfect justice (IoJ p. 96; JPhil 2006 p. 218).
  • The famines passage. “[T]he injustice of continuing famines in a world of prosperity, or of persistently grotesque subjugation of women” is detectable without a complete picture (IoJ p. 103).
  • Not sufficient. Knowing the ideal does not by itself rank non-ideal alternatives. A metric of distance is also needed (IoJ pp. 98–101; JPhil pp. 219–21).
  • Trade-offs. Rawls’s theory does not say how violations of different basic liberties trade off (IoJ p. 99; JPhil p. 220).
  • Sen concedes one point. The difference principle “gives us ground enough to rank other alternatives in terms of the respective advantages of the worst-off” (IoJ p. 97).
• Complaint B: parochial and status-quo-biased (IoJ pp. 24–27 and ch. 6; JPhil pp. 226–8).
  • Scope. Perfect justice is realisable only where state-like institutions exist. So Rawls (The Law of Peoples) and Nagel (2005) deny global distributive justice, and they confine public reasoning to domestic communities.
  • Public reasoning. The views of those outside the polity should enter it (IoJ pp. 138–45). “the demands of objectivity not only require avoiding a ‘personal slant’ (as Rawls noted), but also national parochialism” (JPhil p. 235; IoJ pp. 149–52).
• Complaint C: inflexible (IoJ pp. 106–7, 46).
  • Plurality. There is an “inescapable plurality of competing principles” (IoJ pp. 106–7). The original position may not select a unique set (IoJ p. 46).
  • Reduction. Theories perform “radical surgery that reduces them all into one tidy box of complete and well-fitted demands” (IoJ p. 46). Rawls-style theory insists on “exacting and highly demanding rules” (IoJ p. 107).
  • The dotted-line passage. Principles signed up to “on a dotted line” meet unforeseen cases, and “We then have to decide what has to give and why”. Social-choice theory may seem “too permissive and indecisive”, but “mainstream theories of justice, like Rawls’s or Nozick’s, does not give the idea of justice its due” (IoJ p. 107).
• Sen knows Rawls later allowed a plurality of reasonable liberal conceptions (IoJ pp. 11–12 and especially p. 58, fn 44).
• From the 2006 article. Abolishing slavery, hunger or illiteracy can be shown comparatively to advance justice while leaving a society “far away from the transcendental requirements of a fully just society” (JPhil p. 217, quoted p. 7).

Valentini’s rebuttal

• §II: the characterisation of Rawls (pp. 5–7).
  • “Transcendental”. Taken as “beyond human reach”, the label is false of Rawls, who builds in moderate scarcity, limited altruism and “realistic utopia”. Taken as “categorical/non-comparative”, it is true but a poor label.
  • “Institutionalist”. Taken as “one set of perfect institutions”, it is false. Rawls’s principles apply to the basic structure, but the institutions that realise them vary by society. Per Pogge (2000, p. 165), Rawls assesses a basic structure by “the distribution it would tend to produce in the actual social system”. So Rawls, like Sen, cares about “the actual society that would ultimately emerge”.
• §III: dispensability (pp. 7–10). Sen needs two things: (A) that comparisons are what we most want, and (B) that Rawlsian theory is neither necessary nor sufficient for them.
  • Against (A). Knowing perfect justice matters. Structurally, justice may have a cut-off point, unlike utilitarianism’s unbounded sum (Pogge 1989, p. 41). Practically, we need to know when no more improvement is required.
  • (B), necessity. True but inconsequential for obvious cases such as arbitrary arrest or torturing children. Those are the starting points of theory, not its conclusions. Theory is needed where intuitions diverge: arbitrary arrest against poverty, the death penalty against crime rates. “The work of abstraction … is a way of continuing public discussion when shared understandings of lesser generality have broken down” (Rawls, PL pp. 45–6).
  • (B), sufficiency. “Taken literally, this claim is false”. Lexical priority ranks liberty violations over fair-equality-of-opportunity violations over difference-principle violations, and the difference principle is comparative. The correct nearby claim is that Rawls’s theory is incomplete. “More work needs to be done … but, once again, no paradigm shift is called for” (p. 10).
• §IV: scope (pp. 10–12).
  • The global restriction. Rawls and Nagel restrict justice’s domain on grounds of moral appropriateness (justice applies only given certain relations, coercive or cooperative), not feasibility. So Sen must engage the relational premise and the empirical claim that such relations are absent globally. Transcendental institutionalism “has little to do with this domain restriction” (p. 11).
  • Open public reasoning faces a dilemma. Either it is “so inclusive as to become almost empty”, admitting views that deny moral equality, for example that women are inferior, or it presupposes liberal commitments such as the moral equality of persons, “in which case Sen’s position is not that far from Rawls’s after all”. “no coherent theory of justice can be anti-parochial all the way down” (p. 12).
• §V: inflexibility (pp. 12–14).
  • Plurality. Later Rawls himself allows “many liberalisms” (PL pp. 223, 227; Law of Peoples). So excess ambition in early Rawls is not a reason for a paradigm shift. Some substantive account (constitutional essentials) remains necessary. “taking this modesty to the extreme would make theories of justice entirely useless” (p. 13).
  • Unrevisability. Rawls’s architecture sits inside reflective equilibrium, where “fixed points” are only provisionally fixed (TJ pp. 17–8, 42–3, 506–9). The original position’s conditions are themselves revisable (TJ p. 18, quoted). The process Sen describes in the dotted-line passage “turns out to be nothing other than reflective equilibrium” (p. 13). The charge may fit Nozick better (fn 47). As applied to Rawls, it is “misguided if not inexplicable” (p. 14).
• Verdict (Conclusion, p. 14).
  • Against Sen. The Rawlsian paradigm “needs to be better understood, and further developed, but is far from being fundamentally flawed”. “Sen has simply chosen the wrong target … If there are transcendental institutionalists, Rawls is not one of them.”
  • For Sen. She grants that Sen’s “emphasis on the importance of advancing justice in the real world, and his focus on comparative judgements of justice are timely and laudable”. His critique is “somewhat misdirected”.

Claims

id	claim	strength	support
C1	Sen’s critique consists of three complaints (non-necessity/non-sufficiency for comparisons; parochialism; inflexibility), regrouped from six in IoJ p. 90	moderate (a reconstruction with page references; the grouping is hers)	§I, fn 9
C2	Rawls’s theory is “transcendental” only in the formal, categorical sense, not in the sense of beyond human reach	moderate (textual argument from TJ part III and Law of Peoples)	§II, p. 5–6
C3	Rawls’s principles do not select one ideal set of institutions; the institutions that realise them vary with the society	moderate (textual argument plus Pogge; affirmative-action example)	§II, p. 6–7
C4	An account of perfect justice has independent theoretical and practical value (justice may have a cut-off point)	weak–moderate (conceptual argument; depends on the substantive claim that justice is bounded, which she does not establish)	§III.A, pp. 7–8
C5	Sen’s non-necessity point is true but inconsequential: the obvious cases are theory’s inputs; theory is needed where intuitions conflict	moderate (informal argument, Rawls PL pp. 45–6)	§III.B, pp. 8–9
C6	Rawls’s principles support substantial comparative judgements (lexical priority; the difference principle is comparative, as Sen concedes, IoJ p. 97)	strong as a textual point	§III.B, pp. 9–10
C7	Rawls’s and Nagel’s global restriction rests on relational moral appropriateness, not on transcendental institutionalism	moderate (reading of Rawls and Nagel; overview via Abizadeh 2007)	§IV, p. 11
C8	Sen’s open public reasoning is either near-empty or presupposes liberal moral equality, making it close to Rawls’s	moderate (dilemma argument; Sen’s reply not considered)	§IV, pp. 11–12
C9	The inflexibility charge fails, because Rawls’s method is reflective equilibrium and later Rawls accepts plural liberal conceptions	strong as a textual point (TJ pp. 17–18, 42–3, 506–9; PL pp. 223, 227)	§V, pp. 12–14
C10	No paradigm shift is warranted; Sen’s critique misses Rawls	moderate (the conclusion of C2–C9; she allows it may hit other theorists, e.g. Nozick)	Conclusion, p. 14

Concepts

• Transcendental institutionalism — Sen’s term (book only), as reconstructed. A theory aims to identify perfectly just institutions, with little attention to individual conduct or actual outcomes (p. 3).
• Transcendental (two senses) — (i) beyond human reach; (ii) “absolute” or categorical: “Society X is perfectly just iff…”. Contrast comparative: “Society X is more just than Y if (or iff)…” (p. 6). Valentini would say “categorical”.
• Realization-focused comparison — Sen’s alternative. It ranks actual social realisations comparatively (p. 2).
• Absolute vs comparative justice — judgements of full justice versus judgements of relative justice (p. 4).
• Reflective equilibrium — Rawls’s holistic method, in which principles, judgements and the original position’s conditions are all revisable (p. 13).
• Scope vs site — whom justice applies to versus what it regulates (fn 39, Abizadeh 2007).

Connections

• Zwolinski’s review of The Idea of Justice. The two accounts of the book agree. Both report the transcendental/comparative contrast, the institutions-versus-realisations complaint and the pluralism theme.
  • Valentini’s page references for the necessity point. She cites IoJ pp. 96, 98–106, and for Rawls as institutionalist pp. 6–8, 67.
  • Zwolinski’s page references. He cites pp. 10–12 (difficulty of reasoned agreement on the ideal), pp. 95–102 (perfection not needed for comparative guidance) and pp. 20–22, 67–69 (the realisation focus).
  • Where they disagree. They disagree on the verdict, not on the reconstruction. Zwolinski says Sen’s suggestions “have the potential to change their field for the better”. Valentini says they miss Rawls.
• Nussbaum, “Capabilities as Fundamental Entitlements” (2003). It predates the book. Nussbaum complains that Sen’s capability approach is “typically comparative merely” and wants a threshold list: a partial categorical standard in Valentini’s sense. So Valentini’s defence of categorical principles, and Nussbaum’s demand for them within capability theory, point the same way. This is my connection; neither cites the other.
• Sen, “Equality of What?” (1979); SEP, IEP, Robeyns 2017. Not engaged here. Valentini’s paper is about method, not the metric of justice.
• Record. LIT-130 (Personal Identity and Ethics) mentions Rawls only on the separateness of persons. It has no bearing. Nothing else in the record discusses Rawls or ideal theory.

Bearing on the record

• Use as a second-hand source for The Idea of Justice. Its reconstruction section (§I, plus fns 4, 9, 26, 35, 44) is page-referenced and usable for the book’s methodological claims. Every claim must be marked “per Valentini”. She reads Sen as a critic, and she flags where Sen qualifies himself (fns 26, 44), which makes the reconstruction more trustworthy than a hostile paraphrase.
• Not a source for the capability approach.
• ML practice. It carries nothing for ML practice. The ideal-versus-comparative debate has a loose analogue in alignment discussions (target specification against incremental improvement), but the paper does not touch it.

Limitations

• Negative by design (p. 2). It does not assess whether a comparative theory could do better, or engage Sen’s social-choice apparatus: partial orderings and the “permissive and indecisive” point (IoJ p. 107) are mentioned but not analysed.
• §III.A rests on a premise it does not establish. Its main argument is that justice “may” have a cut-off point.
• The dilemma in §IV is not tested against Sen. It is not checked against Sen’s actual account of the impartial spectator or of positional objectivity, which Zwolinski reports as central to Part II of the book.
• Version. Read from the working paper. The published version (19 pp.) may differ.

Open questions

• Did the journal version change the argument? In particular: did it answer Sen’s impartial-spectator (“open impartiality”) account in §IV? Settled by reading Economics & Philosophy 27(3).
• Did Sen reply? Did Sen respond to Valentini specifically? Crossref shows a Sen reply in Social Philosophy Today 27 (2011), “The Idea of Justice: A Reply”. It is unread, and whether it addresses her is unverified.

Corrections to the seeded skim

• none (there was no seed or dossier)
• What the paper does not cover. The scoping report describes it as a “Critique of The Idea of Justice on transcendental institutionalism vs. comparative justice”. That is right. But the paper says nothing about capabilities, functionings or the capability approach. It explicitly sets aside “Sen’s own path-breaking contributions”, “whose value is not in dispute” (p. 2). It cannot be used as a secondary source on the capability approach, only on Sen’s justice methodology.
• Sen has six complaints, not three. Valentini notes that The Idea of Justice lists six complaints against transcendental institutionalism (p. 90). She regroups them into three (p. 3, fn 9). So “three complaints” is her reconstruction, not Sen’s count.
• The label differs between Sen’s two texts. “Transcendental institutionalism” appears only in the book. The 2006 article says just “transcendental” (fn 4).
• Volume and pages verified against Crossref: 27(3), 297–315, Nov 2011. The scoping report’s “(WP Nov 2010)” and its description of p. 14 as the conclusion are both accurate.