¹ John Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971); The Law of Peoples (Cambridge, MA: Harvard University Press, 2001)

² Rawls, Theory of Justice, 8.

³ David Hume, An Enquiry Concerning the Principles of Morals, [1751].

⁴ Ibid., section III. Emphasis added.

⁵ John Simmons, A., “Ideal and Nonideal Theory,” Philosophy and Public Affairs 38, no. 1 (2010): 5–36, at p. 8.CrossRefGoogle Scholar

⁶ Rawls, Theory of Justice, 241.

⁷ It is worth noting that the degree of complication that is introduced by what Simmons calls a “normal level of noncompliance” depends largely on the stringency of the strictures of justice that deem a level of noncompliance high, low, or “normal.” Noncompliance is not a dependent or endogenous variable within a theory of justice. I owe this observation to David Schmidtz.

⁸ Simmons, “Ideal and Nonideal Theory,” 22.

⁹ In light of approaches to justice as a moral ideal such as those of Estlund and Cohen, one might ask whether strictures on justice to which everyone always complied would still include strictures on just punishment of noncompliance. I leave this question to exponents of such an idealistic approach.

¹⁰ Ibid., 34.

¹¹ Amaryta Sen, The Idea of Justice (London: Penguin, 2009), 102.

¹² This claim is by itself debatable. If the Rawlsian “Everest” is unattainable, then neither of the “lesser peaks” of justice would be on a feasible path to it. Yet it would still be an important matter which is to be preferred in a theory of justice. I owe this point to David Schmidtz.

¹³ Simmons, “Ideal and Nonideal Theory,” 35.

¹⁴ Indeed, one may speculate that the evolution of speech in Hominins took advantage of this regrettable imperfection. I owe this observation to David Schmidtz.

¹⁵ David Schmidtz, “Ideal Theory: What It Is and What Ideally It Would Be,” Ethics 121 (2011): 775–76.

¹⁶ Sen, Amartya, Development as Freedom (New Haven, CT: Yale University Press, 1997), chap. 6.Google Scholar

¹⁷ Is it enough of a vindication of ideal theory that it offer a standard of justice that applies “temporarily” or to a time slice “snapshot” of social relations and institutions, enabling us to identify the highest peak of justice on the current landscape, the one closest to fulfilling the standards of ideal theory, and perhaps also enabling us to temporarily prioritize the gravest injustices to rectify? The trouble with this rather modest ambition for ideal theory is that movements in directions toward and away from the temporary, perhaps transitory “locations” identified using the standard, will shift the landscape itself. I give some examples below.

It is true that justice-enhancing amelioration requires some kind of measure on the dimensions of the space. This by itself does not make ideal theory necessary or even feasible. Following G. A. Cohen, Rescuing Justice and Equality (Cambridge, MA: Harvard University Press, 2008), one might set a higher ambition for ideal theory, one that identifies the highest point or altitude in the space of justice. But if such a point is not attainable without, so to speak, rending the elastic fabric of society altogether, the approach is of little more than academic interest.

¹⁸ At the risk of pushing the metaphor too far, we may initially think of the elastic surface proposed as having x,y coordinates reflecting social relations, and an orthogonal z-axis reflecting degrees of justice and injustice. I owe this observation to a referee. However, as the text notes, if justice is a multidimensional “quantity,” we need to expand the space well beyond three dimensions to a “hyperspace.”

¹⁹ Schmidtz, “Ideal Theory: What It Is and What Ideally It Would Be,” 775–76. Using a map for action-guiding purposes in the domain of justice complicates the elastic sheet metaphor further. The map cannot track fixed truths. The actions it guides, to move out of sinkholes or further up hillsides to locally more just outcomes, will also change the terrain on which the map is supposed to provide guidance. So, the normative map cannot just track a preexisting truth, as realism requires a theory/map to do in science.

²⁰ Rawls, Theory of Justice, 138. As Rawls reminds us in the first section of chapter IV, “Equal Liberty,” section 31, “The four stage sequence,” in the original position each agent is assumed to have enough knowledge to decide not only on the two principles of justice, but also on the characteristics of a just constitution, and similarly the features of just legislation. When it comes to constitutional and legislative strictures on commercial relations especially, the foresight demanded of agents in the original position will be weighty indeed. They will have to identify constitutional and legislative regimes that never provide perverse incentives to “game the system,” no matter how social and technological relationships change. For example, they will identify arrangements that guarantee Nash equilibrium strategies among parties that preserve the difference principle. Failing to do so will defeat Rawls’s objective of designing a sufficiently stable conception of justice, one that motivates the actions of agents. I owe this observation to Wayne Norman.

²¹ Ibid., 101, 107.

²² Cf. note 20.

²³ Ibid., 34ff. Cf. especially the long footnote, number 18 on that page.

²⁴ Samuel Freeman, The Cambridge Companion to Rawls (Cambridge: Cambridge University Press, 2003), 27. Brackets in the quoted material are Freeman’s.

²⁵ Rawls, Theory of Justice, 426–27.

²⁶ Rawls, John, “Justice as Fairness: Political Not Metaphysical,” Philosophy and Public Affairs 14, no. 3 (1985): 223–51.Google Scholar

²⁷ Rawls, John, Collected Papers, ed. Freeman, Samuel (Cambridge, MA: Harvard University Press, 1999), 329.Google Scholar

²⁸ As Rawls himself recognized. Cf. John Rawls, Theory of Justice, 493. I owe this point to a referee.