RetRIevINg RaWLs foR RacIaL justIce? ( 169 )Note what he does not say. He does not say that although race and gender
are not “explicitly included” among these “contingencies,” it would be easy
enough to add them, and then work out how the “two principles of justice”
(including FEO) would “apply” to citizens disadvantaged by race and/ or
gender. Instead he asserts explicitly that this exclusion is a principled (not
merely contingent) one, arising out of the fact that the two principles are
principles of ideal theory for a well- ordered society, while race and gender
problems fall under the different category of non- ideal theory. He does not
say it is just a matter of applying FEO, as Shelby thinks. Instead, he suggests
tentatively that what would be required is “a special form of the difference
principle.”^45 So this is in direct contradiction of Shelby’s claim.
In sum, insofar as Shelby is supposed to be giving us a sympathetic recon-
struction of how Rawls would extrapolate his principles to deal with racial
injustice, he would obviously have to explain (a) why his reconstruction
runs directly opposite to what Rawls himself says, to the limited extent that
he does make positive recommendations (that is, he endorses a modified
DP rather than FEO), and (b) why Rawls himself is so tentative and hesi-
tant in the more frequent textual locations where he raises the problem but
gives no positive recommendation, when according to Shelby it would just
be a simple and straightforward matter of extending FEO to include race.^46
The Importance of the Ideal Theory/ Non- Ideal Theory
DistinctionSo what explains Rawls’s tentativeness, and why does Shelby not appreci-
ate its significance? My suggestion is that though Shelby does mention the
ideal theory/ non- ideal theory distinction, he does not really attribute that
much weight to it. He believes either that non- ideal theory just involves
populating the terms of ideal theory with different variables, or that you can
pre- empt the need for non- ideal theory altogether by appropriately extrap-
olating ideal theory. In my opinion, both of these judgments are wrong and
Shelby— to use old- fashioned Rylean language— is guilty of a category
mistake.^47 As Thomas Nagel points out, affirmative action, that policy of
racial justice actually implemented in the United States, “is probably best
understood in Rawlsian terms as an attempt at corrective justice,” rectifying
violations of the basic liberties, and thus not part of ideal theory.^48 Similarly,
Samuel Freeman judges that “so- called ‘affirmative action,’ or giving prefer-
ential treatment for socially disadvantaged minorities, is not part of FEO
for Rawls, and is perhaps incompatible with it.”^49
The distinction between ideal and non- ideal theory is related to the dis-
tinction between distributive and rectificatory justice, a distinction which