with whom Justice Ginsburg and Justice Breyer join, dissenting.
When the Court devises a new cause of action to enforce a constitutional provision, it ought to identify an injury distinguishable from the consequences of concededly constitutional conduct, and it should describe the elements necessary and sufficient to make out such a claim. Nothing less can give notice to those whose conduct may give rise to liability or provide standards for courts charged with enforcing the Constitution. Those principles of justification, fair notice, and guidance have never been satisfied in the instance of the action announced three Terms ago in Shaw v. Reno, 509 U. S. 630 (1993) (Shaw I), when a majority of this Court decided that a State violates the Fourteenth Amendment’s Equal Protection Clause by excessive consideration of race in drawing the boundaries of voting districts, even when the resulting plan does not dilute the voting strength of any voters and so would not otherwise give rise to liability under the Fourteenth or Fifteenth Amendments, or under the Voting Rights Act.
Far from addressing any injury to members of a class subjected to differential treatment, the standard presupposition of an equal protection violation, Shaw I addressed a putative harm subject to complaint by any voter objecting to an untoward consideration of race in the political process. Although the Court has repeatedly disclaimed any intent to go as far as to outlaw all conscious consideration of race in districting, after three rounds of appellate litigation seeking to describe the elements and define the contours of the Shaw cause of action, a helpful statement of a Shaw claim still eludes this Court. This is so for reasons that go to the conceptual bone.
The result of this failure to provide a practical standard for distinguishing between the lawful and unlawful use of race has not only been inevitable confusion in statehouses and courthouses, but a consequent shift in responsibility for *1046setting district boundaries from the state legislatures, which are invested with front-line authority by Article I of the Constitution, to the courts, and truly to this Court, which is left to superintend the drawing of every legislative district in the land.
Today’s opinions do little to solve Shaw’s puzzles or return districting responsibility to the States. To say this is not to denigrate the importance of Justice O’Connor’s position in her separate opinion, ante, at 990-992, that compliance with § 2 of the Voting Rights Act is a compelling state interest; her statement takes a very significant step toward alleviating apprehension that Shaw is at odds with the Voting Rights Act. It is still true, however, that the combined plurality, minority, and Court opinions do not ultimately leave the law dealing with a Shaw claim appreciably clearer or more manageable than Shaw I itself did. And to the extent that some clarity follows from the knowledge that race may be considered when reasonably necessary to conform to the Voting Rights Act, today’s opinions raise the specter that this ostensible progress may come with a heavy constitutional price. The price of Shaw I, indeed, may turn out to be the practical elimination of a State’s discretion to apply traditional dis-tricting principles, widely accepted in States without racial districting issues as well as in States confronting them.
As the flaws of Shaw I persist, and as the burdens placed on the States and the courts by Shaw litigation loom larger with the approach of a new census and a new round of redistricting, the Court has to recognize that Shaw’s problems result from a basic misconception about the relation between race and districting principles, a mistake that no amount of case-by-case tinkering can eliminate. There is, therefore, no reason for confidence that the Court will eventually bring much order out of the confusion created by Shaw /, and because it has not, in any case, done so yet, I respectfully dissent.
*1047I
As its text indicates and our cases have necessarily and repeatedly recognized,1 Article I of the Constitution places responsibility for drawing voting districts on the States in the first instance. See Art. I, § 2, cl. 1 (“The House of Representatives shall be composed of Members chosen every second Year by the People of the several States, and the Electors in each State shall have the Qualifications requisite for Electors of the most numerous Branch of the State Legislature”); Art. I, § 4, cl. 1 (“The Times, Places and Manner of holding Elections for Senators and Representatives, shall be prescribed in each State by the Legislature thereof; but the Congress may at any time by Law make or alter such Regulations”). The Court has nonetheless recognized limits on state districting autonomy when it could discern a strong constitutional justification and a reasonably definite standard for doing so, as, for example, in announcing the numerical requirement of one person, one vote, see Reynolds v. Sims, 377 U. S. 533 (1964).2 But the Court has never ignored the *1048Constitution’s commitment of districting responsibility to the political branches of the States and has accordingly assumed over the years that traditional districting principles widely accepted among States represented an informal baseline of acceptable districting practices. We have thus accorded substantial respect to such traditional principles (as those, for example, meant to preserve the integrity of neighborhood communities, to protect incumbents, to follow existing political boundaries, to recognize communities of interest, and to achieve compactness and contiguity); we have seen these objectives as entirely consistent with the Fourteenth and Fifteenth Amendments’ demands. See, e. g., id., at 578 (“A State may legitimately desire to maintain the integrity of various political subdivisions, insofar as possible, and provide for compact districts of contiguous territory in designing a legislative apportionment scheme”); White v. Weiser, 412 U. S. 783, 797 (1973) (“[T]he District Court did not suggest or hold that the legislative policy of districting so as to preserve the constituencies of congressional incumbents was unconstitutional or even undesirable”); Voinovich v. Quilter, 507 U. S. 146, 156 (1993) (“Because the States ... derive their reapportionment authority . . . from independent provisions of state and federal law, the federal courts are bound to respect the States’ apportionment choices unless those choices contravene federal requirements”) (internal quotation marks omitted; citation omitted).
The fundamental tenet underlying most of these constitutionally unobjectionable principles (respect for communities of interest or neighborhoods, say) is that voting is more than *1049an atomistic exercise.3 Although it is the law of the Constitution that representatives represent people, not places or things or particular interests, Reynolds, supra, at 562, the notion of representative democracy within the federalist framework presumes that States may group individual voters together in a way that will let them choose a representative not only acceptable to individuals but ready to represent widely shared interests within a district. Aleinikoff & Issacharoff, Race and Redistricting: Drawing Constitutional Lines After Shaw v. Reno, 92 Mich. L. Rev. 588, 601 (1993) (“It is only as collective partisans of the same political preference — whether that preference is defined by party or race or any other measure — that voters can assert their right to meaningful participation in the political process”). Hence, in respecting the States’ implementation of their own, traditional districting criteria, the Court has recognized the basically associational character of voting rights in a representative democracy.
*1050A
Accordingly, before Shaw I, the Court required evidence of substantial harm to an identifiable group of voters to justify any judicial displacement of these traditional district-ing principles. Such evidence existed in Reynolds v. Sims, supra, when the disparate weighting of votes was held unconstitutional, and it was present again when the Court recognized the unconstitutional consequences of vote dilution, see Whitcomb v. Chavis, 403 U. S. 124 (1971); White v. Regester, 412 U. S. 755 (1973). In the one case, the harm was mathematically identifiable; in the other, the arithmetic provided powerful circumstantial evidence of the impossibility of political success for the chosen candidate of a racial and numerical minority in an area with pervasive racial-bloc voting. In both cases, the complainants were from an easily identified group of voters; and even in cases of racial vote-dilution claims, which were conceptually more difficult to state than the principle of one person, one vote, there were readily recognized examples of the harm in question. Indeed, even when one acknowledged that voters would be served by a representative not of their own race and that the Constitution guaranteed no right to pick a winner, see Whitcomb, supra, at 153-155, it was impossible to see mere happenstance in the facts that the American voting-age population was 10.5% black, but the Congress that assembled in 1981 had only 17 black representatives out of 435 and no black senator. Statistical Abstract of the United States, 1982-83, p. 490 (103d ed. 1982) (Table 802); Black Americans: A Statistical Sourcebook 142 (L. Hornor ed. 1995) (Table 4.02); see also Parker, The Damaging Consequences of the Rehnquist Court’s Commitment to Color-Blindness Versus Racial Justice, 45 Am. U. L. Rev. 763, 770-771 (1996) (observing that “[pjrior to the latest round of redistricting after the 1990 Census,... [bjlacks, who constitute 11.1% of the nation’s voting age population, made up only 4.9% of the members of Congress”). The conclusion was inescapable that what we *1051know of as intentional vote dilution accounted for this astonishing fact,4 just as it is equally inescapable that remedies for vote dilution (and hedges against its reappearance) in the form of majority-minority districts account for the fact that the 104th Congress showed an increase of 39 black Members over the 1981 total. Minorities in Congress, 52 Cong. Q., Supplement to No. 44, p. 10 (Nov. 12,1994); see also Parker, supra, at 771 (noting “a fifty percent increase in the number of black members of Congress”).5
*1052Before Shaw I, we not only thus limited judicial interference with state districting efforts to cases of readily demonstrable harm to an identifiable class of voters, but we also confined our concern with districting to cases in which we were capable of providing a manageable standard for courts to apply and for legislators to follow. Within two years of holding in Baker v. Carr, 369 U. S. 186 (1962), that malappor-tionment was a justiciable issue, “the Court recognized that its general equal protection jurisprudence was insufficient for the task and announced an increasingly rigid, simple to apply, voting-specific mandate of equipopulousity.” Karlan, Still Hazy After All These Years: Voting Rights in the Post-Shaw Era, 26 Cumberland L. Rev. 287, 299 (1996) (hereinafter Karlan, Post-S/iaw Era). Likewise, although it is quite true that the common definition of a racial vote-dilution injury (“less opportunity ... to participate in the political process and to elect representatives . . . ,” 42 U. S. C. § 1973(b)) is no model of concrete description, the Court has identified categories of readily comprehensible evidence bearing on the likelihood of such an injury, including facts about size of minority population, quantifiable indications of political cohesiveness and bloc voting, historical patterns of *1053success or failure of favored candidates, and so on. See, e. g., Thornburg v. Gingles, 478 U. S. 30 (1986); White v. Regester, 412 U. S., at 766-770. The particularity of this evidence goes far to separate victims of political “inequality” from those who just happened to support losing candidates.
B
Shaw I, however, broke abruptly with these standards, including the very understanding of equal protection as a practical guarantee against harm to some class singled out for disparate treatment. Whereas malapportionment measurably reduces the influence of voters in more populous districts, and vote dilution predestines members of a racial minority to perpetual frustration as political losers, what Shaw I spoke of as harm is not confined to any identifiable class singled out for disadvantage. See Shaw v. Hunt, ante, at 923-925, 928 (Shaw II) (Stevens, J., dissenting) (noting the absence of a customary disadvantaged class and describing the Shaw I cause of action as a substantive due process, rather than an equal protection, claim). If, indeed, what Shaw I calls harm is identifiable at all in a practical sense, it would seem to play no favorites, but to fall on every citizen and every representative alike. The Court in Shaw I explained this conception of injury by saying that the forbidden use of race “reinforces the perception that members of the same racial group . . . think alike, share the same political interests, and will prefer the same candidates at the polls,” and that it leads elected officials “to believe that their primary obligation is to represent only the members of that group, rather than their constituency as a whole.” Shaw I, 509 U. S., at 647-648. This injury is probably best understood as an “expressive harm,” that is, one that “results from the idea or attitudes expressed through a governmental action, rather than from the more tangible or material consequences the action brings about.” Pildes & Niemi, Expressive Harms, “Bizarre Districts,” and Voting Rights: Evaluat*1054ing Election-District Appearances after Shaw v. Reno, 92 Mich. L. Rev. 483, 506-507 (1993); see also id., at 493 (“The theory of voting rights [that Shaw i] endorses centers on the perceived legitimacy of structures of political representation, rather than on the distribution of actual political power between racial or political groups”). To the extent that racial considerations do express such notions, their shadows fall on majorities as well as minorities, whites as well as blacks, the politically dominant as well as the politically impotent. Thus, as an injury supposed to be barred by the Equal Protection Clause, this subject of the “analytically distinct” cause of action created by Shaw I, supra, at 652, bears virtually no resemblance to the only types of claims for gerrymandering we had deemed actionable following Davis v. Bandemer, 478 U. S. 109 (1986), those involving districting decisions that removed an identifiable class of disfavored voters from effective political participation. See, e. g., Gomillion v. Lightfoot, 364 U. S. 339 (1960); Thornburg v. Gingles, supra.6
Just as the logic of traditional equal protection analysis is at odds with Shaw’s concept of injury, so the Court’s rhetoric of racially motivated injury is inapposite to describe the consideration of race that it thinks unreasonable. Although the Court used the metaphor of “political apartheid” as if to refer to the segregation of a minority group to eliminate its association with a majority that opposed integration, Shaw I, supra, at 647, talk of this sort of racial separation is not on point here. The de jure segregation that the term “political *1055apartheid” brings to mind is unconstitutional because it emphatically implies the inferiority of one race. See Brown v. Board of Education, 347 U. S. 483, 494 (1954) (“To separate [minority children] from others of similar age and qualifications solely because of their race generates a feeling of inferiority as to their status in the community”). Shaw I, in contrast, vindicated the complaint of a white voter who objected not to segregation but to the particular racial proportions of the district. See Karlan, Our Separatism? Voting Rights as an American Nationalities Policy, 1995 U. Chi. Legal Forum 83, 94 (hereinafter Karlan, Our Separatism) (noting the irony of using the term “apartheid” to describe what are “among the most integrated districts in the country”). Whatever this district may have symbolized, it was not “apartheid.” Nor did the proportion of its racial mixture reflect any purpose of racial subjugation, the district in question having been created in an effort to give a racial minority the same opportunity to achieve a measure of political power that voters in general, and white voters and members of ethnic minorities in particular, have enjoyed as a matter of course. In light of a majority-minority district’s purpose to allow previously submerged members of racial minorities into the active political process, this use of race cannot plausibly be said to affect any individual or group in any sense comparable to the injury inflicted by de jure segregation. It obviously conveys no message about the inferiority or outsider status of members of the white majority excluded from a district. And because the condition addressed by creating such a district is a function of numbers, the plan implies nothing about the capacity or value of the minority to which it gives the chance of electoral success.
Added to the anomalies of Shaw /’s idea of equal protection injury and the rhetoric of its descriptions, there is a further conceptual inadequacy in Shaw I. Whereas it defines injury as the reinforcement of the notion that members of a racial group will prefer the same candidates at the polls, the imme*1056diate object of the constitutional prohibition against the intentional dilution of minority voting strength is to protect the right of minority voters to make just such a preference effective. There would, for example, be no vote dilution by virtue of racial-bloc voting unless voters of a racial minority would themselves tend to stick together in voting for a given candidate (perhaps, though not necessarily, of their own race, as well). Indeed, if there were no correlation between race and candidate preference, it would make no sense to say that minority voters had less opportunity than others to elect whom they would; they would be part of the mainstream and the winners would be their own choices. When voting is thus racially polarized, it is just because of this polarization that majority-minority districts provide the only practical means of avoiding dilution or remedying the dilution injury that has occurred already. Shaw I has thus placed those who choose to avoid the long-recognized constitutional harm of vote dilution at risk by casting doubt on the legitimacy of its classic remedy; the creation of a majority-minority district “reinforces” the notion that there is a correlation between race and voting, for that correlation is the very condition on which the success of the court-ordered remedy depends. So it is that the Court’s definition of injury is so broad as to cover constitutionally necessary efforts to prevent or remedy a violation of the Fourteenth and Fifteenth Amendments and of § 2 of the Voting Rights Act.
One way to temper the overreach of the Court’s concept of injury (though it would not avoid the difficulty that there is no equal protection injury in the usual sense, discussed above, see supra, at 1050) would be simply to exclude by definition from Shaw I injury a use of race in districting that is reasonably necessary to remedy or avoid dilution; the Court’s move at least in this direction, see infra, at 1066-1069, is a sound one, as is its continuing recognition (despite its broad definition of harm) that not every intentional creation of a majority-minority district requires strict scrutiny. *1057See ante, at 958; ante, at 993 (O’Connor, J., concurring); cf. Miller, 515 U. S., at 928 (O’Connor, J., concurring). But the suggested qualification would fall short of eliminating the difficulty caused by the existing definition, for the uses of race to remedy past dilution or to hedge against future dilution are not the only legitimate uses of race that are covered, and threatened, by the overbreadth of the Shaw injury. This will become clear in examining the Court’s efforts to solve its definitional problems by relying upon the degree to which race is used in defining the injury it discerns.
C
The Court’s failure to devise a concept of Shaw harm that distinguishes those who are injured from those who are not, or to differentiate violation from remedy, is matched by its inability to provide any manageable standard to distinguish forbidden districting conduct from the application of traditional state districting principles and the plans that they produce. This failure, while regrettable, need not have occurred, for when the Court spoke in Shaw I of a district shape so “bizarre” as to be an unequivocal indication that race had influenced the districting decision to an unreasonable degree, Shaw I could have been pointing to some workable criterion of shape translatable into objective standards. Leaving Shaw’s theoretical inadequacies aside, it would have been possible to devise a cause of action that rested on the expressive character of a district’s shape, and created a safe harborin the notion of a compact district objectively quantified in terms of dispersion, perimeter, and population. See Pildes & Niemi, 92 Mich. L. Rev., at 553-575. Had the Court followed this course, the districts whose grotesque shapes provoke the sharpest reaction would have been eliminated in racially mixed States, which would have known how to avoid Shaw violations and, thus, federal judicial intrusion. Shaw would have been left a doctrinal incongruity, but not an unmanageable one.
*1058The Court, however, rejected this opportunity last Term in Miller v. Johnson, supra, when it declined to contain Shaw by any standard sufficiently quantifiable to guide the decisions of state legislators or to inform and limit review of districting decisions by the courts. The Court rejected shape as a sufficient condition for finding a Shaw violation, or even a necessary one. 515 U. S., at 915. See also Issacharoff, The Constitutional Contours of Race and Politics, 1995 S. Ct. Rev. 45, 56 (hereinafter Issacharoff, Constitutional Contours) (“Miller is rather categorical in its. refusal to limit the application of the equal protection clause to bizarre districts alone”)- Instead, it recharacterized the cause of action in terms devised in other cases addressing essentially different problems, by proscribing the consideration of race when it is the “predominant factor motivating the legislature],” 515 U. S., at 916, or when the use of race is “in substantial disregard of customary and traditional districting practices,” id., at 928 (O’Connor, J., concurring).
As a standard addressed to the untidy world of politics, neither “predominant factor” nor “substantial disregard” inspires much hope.7 It is true of course that the law rests certain other liability decisions on the feasibility of untangling mixed motives, and courts and juries manage to do the untangling. See, e.g., Mt. Healthy City Bd. of Ed. v. Doyle, 429 U. S. 274, 287 (1977) (employee’s burden to show that constitutionally protected conduct is a “substantial factor” in *1059decision not to rehire him; employer’s burden to show that it would have made same decision “even in the absence of the protected conduct”); Hunter v. Underwood, 471 U. S. 222, 228 (1985) (“Once racial discrimination is shown to have been a ‘substantial’ or ‘motivating’ factor behind the enactment of the law, the burden shifts to the law’s defenders to demonstrate that the law would have been enacted without this factor”); but see Arlington Heights v. Metropolitan Housing Development Corp., 429 U. S. 252, 265 (1977) (“Rarely can it be said that a legislature or administrative body operating under a broad mandate made a decision motivated solely by a single concern, or even that a particular purpose was the ‘dominant’ or ‘primary’ one”). At first glance, then, it may not seem entirely out of the question for courts to sort out the strands in Shaw cases. But even this cool comfort would be misplaced.
While a court may be entitled to some confidence that in most cases it will be able, for example, to distinguish the relative strength of an employer’s dissatisfaction with an employee’s job performance from his displeasure over a worker’s union membership, see NLRB v. Transportation Management Corp., 462 U. S. 393, 403-405 (1983), such confidence would be unwarranted in the districting context. It is not merely that the very nature of districting decisions makes it difficult to identify whether any particular consideration, racial or otherwise, was the “predominant motive,” though that is certainly true:
“Districting plans are integrated bundles of compromises, deals and principles. To ask about the reason behind the design of any one particular district is typically to implicate the entire pattern of purposes and trade-offs behind a districting plan as a whole. Searching for ‘the reason’ or ‘the dominant reason’ behind a particular district’s shape is often like asking why one year’s federal budget is at one level rather than another. Moreover, to require a coherent explanation for the spe-*1060ciñe shape of even one district is to impose a model of legalistic decisionmaking on the one political process that least resembles that model.” Pildes & Niemi, swpra, at 585-586 (footnote omitted).
The reason that use of the predominant motive standard in reviewing a districting decision is bound to fail is more fundamental than that: in the political environment in which race can affect election results, many of these traditional districting principles cannot be applied without taking race into account and are thus, as a practical matter, inseparable from the supposedly illegitimate racial considerations. See Pildes & Niemi, supra, at 578 (“[R]ace frequently correlates with other socioeconomic factors. In evaluating oddly shaped districts, this correlation will require courts to attempt to untangle legitimate communities of interest from the now-illegitimate one of race. If blacks as blacks cannot be grouped into a ‘highly irregular’ district, but urban residents or the poor can, how will courts distinguish these contexts, and under what mixed-motive standard?”); Issacharoff, Constitutional Contours 58 (“Given the palpability of racial concerns in the political arena, [Miller’s causation standard could] either doom all attempts to distribute political power in multi-ethnic communities or . . . fail to provide a basis for distinguishing proper from improper considerations in redistricting”).
If, for example, a legislature may draw district lines to preserve the integrity of a given community, leaving it intact so that all of its members are served by one representative, this objective is inseparable from preserving the community’s racial identity when the community is characterized, or even self-defined, by the race of the majority of those who live there. This is an old truth, having been recognized every time the political process produced an Irish or Italian or Polish ward.
*1061“[Ejthnicity itself can tie people together, as volumes of social science literature have documented — even people with divergent economic interests. For this reason, ethnicity is a significant force in political life....
“. . . The creation of ethnic districts reflecting felt identity is not ordinarily viewed as offensive or demeaning to those included in the delineation.” Miller v. Johnson, 515 U. S., at 944-945 (Ginsburg, J., dissenting).
Or take the traditional principle of providing protection for incumbents. The plurality seems to assume that incumbents may always be protected by drawing lines on the basis of data about political parties. Cf. ante, at 967-968, 970-971. But what if the incumbent has drawn support largely for racial reasons? What, indeed, if the incumbent was elected in a majority-minority district created to remedy vote dilution that resulted from racial-bloc voting? It would be sheer fantasy to assume that consideration of race in these circumstances is somehow separable from application of the traditional principle of incumbency protection, and sheer incoherence to think that the consideration of race that is constitutionally required to remedy Fourteenth and Fifteenth Amendment vote dilution somehow becomes unconstitutional when aimed at protecting the incumbent the next time the census requires redistricting.
Thus, it is as impossible in theory as in practice to untangle racial consideration from the application of traditional districting principles in a society plagued by racial-bloc voting8 with a racial minority population of political significance, or at least the unrealized potential for achieving it. And it *1062is for just this fundamental reason that a test turning on predominant purpose is incapable of producing any answer when traditional districting principles are applied in the political environment in which Shaw I actions are brought.
II
Shaw Vs recognition of a misuse of race in districting even when no vote dilution results thus rests upon two basic deficiencies: first, the failure to provide a coherent concept of equal protection injury, there being no separably injured class and no concept of harm that would not condemn a constitutionally required remedy for past dilution as well as many of the districting practices that the Court is seeking to preserve; second, the failure to provide a coherent test for distinguishing a “predominant” racial consideration from the application of traditional districting principles in a society whose racial mixture is politically significant and where racial-bloc voting exists. The necessary consequence of these shortcomings is arbitrariness; it is impossible to distinguish what is valid from what is not, or to decide how far members of racial minorities may engage “in the same sort of pluralist electoral politics that every other bloc of voters enjoys.” Karlan, Our Separatism 103. Indeed, if one needed further proof of this arbitrariness, one need go no further than Justice Stevens’s dissent in this case. The plurality effectively concedes that Justice Stevens has not unfairly applied the principles governing the Shaw cause of action, cf. ante, at 971, n. (noting that “[i]n the application of our precedents to District 30, our disagreement with Justice Stevens’ dissent, [ante], at 1014-1031, is largely factual”); in my judgment he has faithfully applied those principles in the spirit intended by the plurality. And yet the conclusions that the two sides reach after applying precisely the same test could not be more different.
*1063Along with this endemic unpredictability has come the destruction of any clear incentive for the States with substantial minority populations to take action to avoid vote dilution. Before Shaw, state politicians who recognized that minority vote dilution had occurred, or was likely to occur without redistricting aimed at preventing it, could not only urge their colleagues to do the right thing under the Fourteenth Amendment, but counsel them in terrorem that losing a dilution case would bring liability for counsel fees under 42 U. S. C. § 1988(b) or 42 U. S. C. § 19732(e). See Issacharoff, Constitutional Contours 48 (“Minority political actors could leverage not only their political power but the enforcement provisions of Section 5 of the Voting Rights Act, and the threat of suit under Section 2 of the Act against adverse districting decisions”); cf. Hastert v. Illinois State Bd. of Election Commr’s, 28 F. 3d 1430, 1444 (CA7 1993) (awarding fees to the prevailing parties in a case in which the state legislature failed to draw congressional districts, over the Board of Elections’s objection that it had “no interest in the eventual outcome except that there be an outcome” for it to implement) (emphasis in original). But this argument is blunted now, perhaps eliminated in practice, by the risk of counsel fees in a Shaw I action. States seeking to comply in good faith with the requirements of federal civil rights laws “now find themselves walking a tightrope: if they draw majority-black districts they face lawsuits under the equal protection clause; if they do not, they face both objections under section 5 of the Voting Rights Act and lawsuits under section 2.” Karlan, Bost-Shaw Era 289. See ante, at 1037 (Stevens, J., dissenting) (“On one hand, States will risk violating the Voting Rights Act if they fail to create majority-minority districts. If they create those districts, however, they may open themselves to liability under Shaw and its progeny”). The States, in short, have been told to get things just right, no dilution and no predominant consider*1064ation of race short of dilution, without being told how to do it. The tendency of these conflicting incentives is toward a stalemate, and neither the moral force of the Constitution nor the mercenary threat of liability can operate effectively in this obscurity.
As a consequence, where once comprehensible districting obligations confronted the legislators and governors of the States, there is now a vacuum of responsibility in any State with the mixed population from which Shaw suits come. We can no longer say with the old assurance that such States have a duty to comply with federal requirements in district-ing, since a State, like an individual, can hardly be blamed for failing to fulfill an obligation that has never been explained. It is true, of course, that a State may suffer consequences if the ultimate arbiter decides on a result different from the one the State has put in place, but that bad luck does not change the fact that a State cannot be said to be obliged to apply a standard that has not been revealed. Because the responsibility for the result can only be said to rest with the final arbiter, the practical responsibility over districting has simply shifted from the political branches of the States with mixed populations to the courts, and to this Court in particular. “The Court has apparently set itself upon a course of . . . reviewing challenged districts one by one and issuing opinions that depend so idiosyncratically on the unique facts of each case that they provide no real guidance to either lower courts or legislatures.” Karlan, Post-Shaw Era 288. The tragedy in this shift of political responsibility lies not only in the fact of its occurrence in this instance, but in the absence of coherent or persuasive justifications for causing it to occur.
i — I H-1 H-i
Although today’s cases do not address the uncertainties that stem from Shaw’s underlying incoherence, they do aim to mitigate its inscrutability with some specific rules.
*1065A
In each of today’s cases, the Court expressly assumes that avoiding a violation of the Voting Rights Act qualifies as a sufficiently compelling government interest to satisfy the requirements of strict scrutiny. See ante, at 977 (“As we have done in each of our three previous cases , we assume without deciding that compliance with the results test [of § 2 of the Voting Rights Act] . . . can be a compelling state interest”); Shaw II, ante, at 915 (“We assume, arguendo, for the purpose of resolving this case, that compliance with § 2 could be a compelling interest”). While the Court’s decision to assume this important point, arguendo, is no holding, see Seminole Tribe of Fla. v. Florida, ante, at 125 (Souter, J., dissenting), the assumption itself is encouraging because it confirms the view that the intentional creation of majority-minority districts is not necessarily a violation of Shaw I, ante, at 958 (strict scrutiny does not “apply to all cases of intentional creation of majority-minority districts”), and it indicates that the Court does not intend to bring the Shaw cause of action to what would be the cruelly ironic point of finding in the Voting Rights Act of 1965 (as amended) a violation of the Fourteenth Amendment’s equal protection guarantee. Cf. Pildes & Niemi, 92 Mich. L. Rev., at 498 (observing that “[i]f the Court believed there were serious constitutional questions with the fundamental structure of this scheme, the Court had numerous means to avoid permitting an unconstitutionally composed legislature to assume power,” and seeing the reservation of this question in Voinovich v. Quitter, 507 U. S., at 157, as “evidence that a majority of the Court is not prepared to find a general ban on race-conscious districting in the Constitution”). Justice O’Con-nor’s separate opinion, ante, at 990-992, bears on each of these points all the more emphatically, for her view that compliance with § 2 is (not just arguendo) a compelling state interest and her statement of that position virtually insulate the Voting Rights Act from jeopardy under Shaw as such.
*1066B
The second point of reference to come out of today’s cases is the rule that if a State begins its map-drawing efforts with a compact majority-minority district required by Gingles, the State may not rely too heavily on racial data in adjusting that district to serve traditional districting principles. While this rule may indeed provide useful guidance to state legislatures, its inherent weakness is clear from what was said above, supra, at 1060-1062: it is in theory and in fact impossible to apply “traditional districting principles” in areas with substantial minority populations without considering race. As to some of those principles, to be sure, the ban on the overuse of racial data may not have much significance; racially identified communities can be identified in other ways and will be, after today. But protecting a minority incumbent may be another matter, since we cannot assume, as the plurality does, that reliance on information about “party affiliation” will serve to protect a minority incumbent, and we cannot tell when use of racial data will go too far on the plurality’s view, ante, at 962-963. It therefore may well be that loss of the capacity to protect minority incumbency is the price of the rule limiting States’ use of racial data. If so, it will be an exceedingly odd result, when the whole point of creating yesterday’s majority-minority district was to remedy prior dilution, thus permitting the election of the minority incumbent who (the Court now seems to declare) cannot be protected as any other incumbent could be.
C
The third point of reference attributable to today’s cases is as yet only a possibility; a suggestion in the discussions of the narrow tailoring test that States seeking to avoid violating § 2 of the Voting Rights Act may draw the district that the Voting Rights Act compels, and this district alone. See Shaw II, ante, at 915-918 (rejecting North Carolina’s District 12 because it does not sufficiently coincide with the *1067assumed Gingles district); ante, at 1035 (Stevens, J., dissenting) (“[I]t now seems clear that the only way that a State can both create a majority-minority district and avoid a racial gerrymander is by drawing . . . within the ‘limited degree of leeway’ granted by the Court... the precise compact district that a court would impose in a successful §2 challenge”). If the Court were to say that a district drawn to avoid dilution must respond to the dilution threat in some geographically exact way, but see Shaw II, ante, at 916, n. 8 (suggesting that States may have flexibility in complying with § 2 of the Voting Rights Act); ante, at 1037 (Stevens, J., dissenting) (noting that States traditionally have enjoyed a broader discretion in drawing district lines), then presumably a district drawn in a race-conscious fashion could survive only if it was as compact as the Gingles district hypothesized for purposes of stating a vote-dilution claim, and positioned where the hypothetical district would be.
If the Court ultimately were to reach such a conclusion, it would in one respect be taking a step back toward Shaw I and its suggestion that a district’s shape might play an important, if not determinative role in establishing a cause of action. Such a step would, however, do much more than return to Shaw I, which suggested that a compact district would be a safe haven, but not that the district hypothesized under Gingles was the only haven. See, e. g., Shaw I, 509 U. S., at 646 (“The district lines may be drawn, for example, to provide for compact districts of contiguous territory, or to maintain the integrity of political subdivisions”).
I refer to this step as a “possibility” deliberately. The Court in Shaw II does not go beyond an intimation to this effect, and Bush raises doubt that the Court would go so far. See ante, at 977-978 (rejecting the argument made by Justice Stevens); see also ante, at 978 (“[T]he States retain a flexibility that federal courts enforcing § 2 lack .... And nothing that we say today should be read as limiting ‘a State’s discretion to apply traditional districting princi-*1068pies’ ”); but see ante, at 994 (O’Connor, J., concurring) (“[I]f a State pursues that compelling interest by creating a district that ‘substantially addresses’ the potential liability, and does not deviate substantially from a hypothetical court-drawn §2 district for predominantly racial reasons, its dis-tricting plan will be deemed narrowly tailored” (citations omitted)); but see also ante, at 977 (“We also reaffirm that the ‘narrow tailoring’ requirement of strict scrutiny allows the States a limited degree of leeway in furthering such interests. ... We thus reject, as impossibly stringent, the District Court’s view of the narrow tailoring requirement, that ‘a district must have the least possible amount of irregularity in shape, making allowances for traditional districting criteria’ ” (citation omitted)). Indeed, Bush leaves open the possibility that a State could create a majority-minority district that does not coincide with the Gingles shape so long as racial data are not overused, ante, at 962, 981, and it does not suggest that a Shaw claim could be premised solely on a deviation from a Gingles district.
Suffice it to say for now that if the Court were to try to render Shaw more definite by imposing any such limitations on shape and placement, the added measure of clarity would either be elusive or it would come at an exorbitant price from States seeking to comply with the Voting Rights Act and the Fourteenth and Fifteenth Amendments. It would be elusive if the Court meant that race could be considered in alleviating racial dilution but not in applying any traditional districting principle: we have already seen that race is inextricably intertwined with some common districting principles when applied in a multiracial society. See supra, at 1060-1062. Or it would come at an exorbitant price, because no other districting principle would be allowed to affect the compactness or placement that would be required for purposes of Gingles. The Court would thus be cutting back on a State’s power to vary district shape through its application of the very districting principles that are supposed to pre*1069dominate in importance over racial consideration. That is, the Court would be reducing the discretion of a State seeking to avoid or correct dilution to the scope of a federal court’s discretion when devising a remedy for dilution. There could, of course, be no justification for taking any such step. While there is good reason to limit a federal court’s discretion to interfere, in a State’s political process when it employs its remedial power in dilution cases, cf. Voinovich v. Quilter, 507 U. S., at 156 (“Federal courts are barred from intervening in state apportionment in the absence of a violation of federal law precisely because it is the domain of the States ... to conduct apportionment”), there is no apparent reason to impose the same limitations upon the discretion accorded to a State subject to an independent constitutional duty to make apportionment decisions, see ibid. (“Because the States . . . derive their reapportionment authority . . . from independent provisions of state and federal law,... the federal courts are bound to respect the States’ apportionment choices unless those choices contravene federal requirements”) (internal quotation marks omitted). The principles of federalism that we have tried to follow strongly counsel against imposing any such limitations.
D
In sum, the three steps the Court takes today toward a more definite cause of action either fail to answer the objections to Shaw I or prompt objections of their own. Recognition of a State’s interest in complying with the Voting Rights Act does not address the practical impossibility courts will encounter in identifying a predominant use of race, as distinguished from some lesser, reasonable consideration of it, when a State applies its customary districting principles. The limitation on the use of racial data is unlikely to make much difference in practice except to jeopardize minority incumbency protection. And the possibility that the Court will require Gingles districts (or districts substantially close *1070to them) when compliance with § 2 of the Voting Rights Act is an object of districting would render a State’s districting obligation more definite only by eliminating its ability to apply the very districting principles traditionally considered to be important enough to furnish a theoretical baseline of reasonable districting practices.
> H-1
If today’s developments fall short of curing Shaw s unworkability, it must be said that options for addressing them are few. Assuming that Shaw is not to be overruled as a flawed experiment, the Court may select from two alternatives, depending on whether its weightier concern is to preserve traditional districting principles or to cure the anomalies created by Miller’s “predominant purpose” criterion.
If the Court’s first choice is to preserve Shaw in some guise with the least revolutionary effect on districting principles and practice, the Court could give primacy to the principle of compactness and define the limits of tolerance for unorthodox district shape by imposing a measurable limitation on the bizarre, presumably chosen by reference to historical practice (adjusted to eliminate the influence of any dilution that very practice may have caused in the past, cf. Pildes & Niemi, 92 Mich. L. Rev., at 573-574, n. 246 (discussing the egregious racial gerrymanders of the 19th century)) and calculated on the basis of a district’s dispersion, perimeter, and population. See id., at 553-575. This alternative would be true to Shaw I in maintaining that a point can be reached when the initially lawful consideration of race becomes unreasonable and in identifying appearance as the expression of undue consideration; and it would eliminate Miller’ s impossible obligation to untangle racial considerations from so-called “race-neutral” objectives (such as according respect to community integrity and protecting the seats of *1071incumbents) when the racial composition of a district and voter behavior bar any practical chance of separating them. The incongruities of Shaw’s concept of injury when considered in light of our customary equal protection analysis, our remedial practice, and traditional respect for state district-ing discretion would, of course, persist, but if Shaw were defined by measures that identified forbidden shape as the manifestation of unreasonable racial emphasis, we would at least provide the notice and guidance that are missing from the law today.
The other alternative for retaining a Shaw cause of action in some guise would be to accept the fact that, in the kind of polarized multiracial societies that will generate Shaw actions as presently understood, racial considerations are inseparable from many traditional districting objectives, making it impossible to speak of race as predominating. The consequence of facing this reality is that if some consideration of race is to be forbidden as supposedly unreasonable in degree, then the use of districting principles that implicate the use of race must be forbidden. That is, traditional dis-tricting practices must be eliminated. Such a result would, of course, be consistent with Shaw I’s concept of injury as affecting voters of whatever race. But cf. Valley Forge Christian College v. Americans United for Separation of Church and State, Inc., 454 U. S. 464, 489 (1982) (fact that some expressive harms are insufficient to satisfy Article III standing requirements does not allow for relaxation of those requirements). The result, in short, would be colorblindness in determining the manner of choosing representatives, either by eliminating the practice of districting entirely, or by replacing it with districting on some principle of randomness that would not account for race in any way.
While such is the direction in which Shaw and Miller together point, the objections to following any such course seem insurmountable. The first is the irony that the price *1072of imposing a principle of colorblindness in the name of the Fourteenth Amendment would be submerging the votes of those whom the Fourteenth and Fifteenth Amendments were adopted to protect, precisely the problem that necessitated our recognition , of vote dilution as a constitutional violation in the first place. Eliminating districting in the name of colorblindness would produce total submersion; random submersion (or packing) would result from districting by some computerized process of colorblind randomness. Thus, unless the attitudes that produce racial-bloc voting were eliminated along with traditional districting principles, dilution would once again become the norm. While dilution as an intentional constitutional violation would be eliminated by a randomly districted system, this theoretical nicety would be overshadowed by the concrete reality that the result of such a decision would almost inevitably be a so-called “representative” Congress with something like 17 black members. See supra, at 1050. In-any event,, the submergence would violate the prohibition of even non-intentional dilution found in §2 of the Voting Rights Act. The only way to avoid this conflict would be to declare the Voting Rights Act unconstitutional, a prospect hardly in harmony with the Court’s readiness to assume today that compliance with the Voting Rights Act qualifies as a compelling state interest for purposes of litigating a Shaw claim.
The second objection is equally clear. Whatever may be the implications of what I have called Shaw’s failings, the Court has repeatedly made it plain that Shaw was in no way intended to effect a revolution by eliminating traditional districting practice for the sake of colorblindness. Shaw I, 509 U. S., at 642 (“Despite their invocation of the ideal of a ‘color-blind’ Constitution, see Plessy v. Ferguson, 163 U. S. 537, 559 (1896) (Harlan, J., dissenting), appellants appear to concede that race-conscious districting is not always unconstitutional. . . . That concession is wise: This Court *1073never has held that race-conscious state decisionmaking is impermissible in all circumstances”); cf. Richmond v. J. A. Croson Co., 488 U. S. 469, 520-521 (1989) (Scalia, J., concurring in judgment) (criticizing the majority for rejecting a strict principle of colorblindness). Indeed, the very fear that led to the creation of the Shaw cause of action was that racial concerns were taking too heavy a toll on districting practices that had evolved over the years through the political process. Shaw I, 509 U. S., at 644-649. Justice O’Connor, moreover, has made it obvious that race has a legitimate place in districting, id., at 642 (“[R]ace-conscious redistricting is not always unconstitutional”); Miller, 515 U. S., at 928-929 (O’Connor, J., concurring); ante, at 993 (O’Connor, J., concurring), that the intentional creation of majority-minority districts is not forbidden by Shaw, Miller, supra, at 928 (O’Connor, J., concurring) (districts may be permissible “even though race may well have been considered in the redistricting process”); ante, at 990-992 (O’Connor, J., concurring), and that Shaw was aimed at only the exceptional district, 515 U. S., at 928-929 (“Application of the Court’s standard does not throw into doubt the vast majority of the Nation’s 435 congressional districts”). Of the present Court majority, only Justices Scalia and Thomas are on record as concluding that any intentional creation of a majority-minority district is a forbidden racial gerrymander. Ante, at 1001 (Thomas, J., concurring in judgment).
Since a radical transformation of the political selection process in the name of colorblindness is out of the question, the Court’s options for dealing with Shaw’s unworkability are in truth only these: to confine the cause of action by adopting a quantifiable shape test or to eliminate the cause of action entirely. Because even a truncated Shaw would rest on the untenable foundation I have described, and the supposed, expressive harm Shaw seeks to remedy is unlikely *1074to justify the disruption that even a modified Shaw would invite, there is presently no good reason that the Court’s withdrawal from the presently untenable state of the law should not be complete. While I take the commands of stare decisis very seriously, the problems with Shaw and its progeny are themselves very serious. The Court has been unable to provide workable standards, the chronic uncertainty has begotten no discernible reliance, and the costs of persisting doubt about the limits of state discretion and state responsibility are high.
There is, indeed, an added reason to admit Shaw’s failure in providing a manageable constitutional standard and to allow for some faith in the political process. That process not only evolved the very traditional districting principles that the Court has pledged to preserve, but has applied them in the past to deal with ethnicity in a way that should influence our thinking about the prospects for race. It is difficult to see how the consideration of race that Shaw condemns (but cannot avoid) is essentially different from the consideration of ethnicity that entered American politics from the moment that immigration began to temper regional homogeneity. Recognition of the ethnic character of neighborhoods and incumbents, through the application of just those dis-tricting principles we now view as traditional, allowed ethnically identified voters and their preferred candidates to enter .the mainstream of American politics, see Miller, supra, at 944-945 (Ginsburg, J., dissenting); D. Judd, The Politics of American Cities: Private Power and Public Policy 70 (3d ed. 1988); see generally S. Erie, Rainbow’s End: Irish-Americans and the Dilemmas of Urban Machine Politics, 1840-1985 (1988), and to attain a level of political power in American democracy. The result has been not a state regime of ethnic apartheid, but ethnic participation and even a moderation of ethnicity’s divisive effect in political practice. For although consciousness of ethnicity has not disappeared from the *1075American electorate, its talismanic force does appear to have cooled over time.9 It took Boston Irish voters, for example, to elect Thomas Menino mayor in 1993.10
*1076There is, then, some reason to hope that if vote dilution is attacked at the same time that race is given the recognition that ethnicity has historically received in American politics, the force of race in politics will also moderate in time. There are even signs that such hope may be vindicated, even if the evidence is necessarily tentative as yet. See U. S. Comm’n on Civil Rights, The Voting Rights Act: Ten Years After, p. 155 (Jan. 1975) (“In many areas the great increase in minority registration and voting since the passage of the Voting Rights Act in 1965 has meant that politicians can no longer afford to ignore minority voters. This has brought about a significant decline in racial appeals by candidates and has made incumbents and candidates more responsive to minority needs”); Carsey, The Contextual Effects of Race on White Voter Behavior: The 1989 New York City Mayoral Election, 57 J. of Politics 221, 228 (1995) (reporting, in 1994, that “the contextual effects of race may not be so different from the contextual effects of factors like partisanship, ethnicity, or social class as we might have believed”); Sigelman, Sigelman, Walkosz, & Nitz, Black Candidates, White Voters: Understanding Racial Bias in Political Perceptions, 39 Am. J. of Political Science 243, 244 (1995) (“Over the years, white Americans have expressed increasing willingness to vote for black candidates”); Peirce, Fresh Air in City Hall, Baltimore Sun, Nov. 8, 1993, p. 7A (“In contest after contest, victory has gone to mayoral candidates who eschew talk of race”); see also Gingles, 478 U. S., at 56 (noting that crossover voting in favor of minority candidates is more common when minority incumbents stand for reelection); Collins v. Norfolk, 883 F. 2d 1232, 1243 (CA4 1989) (same). This possibility that racial politics, too, may grow wiser so long as minority votes are rescued from submergence should be considered in determining how far the Fourteenth and Fifteenth Amendments require us to devise constitutional common law to supplant *1077the democratic process with litigation in federal courts. It counsels against accepting the profession that Shaw has yet evolved into a manageable constitutional standard, and from that case’s invocation again today I respectfully dissent.
See, e. g., Growe v. Emison, 507 U. S. 25, 34 (1993) (“[T]he Constitution leaves with the States primary responsibility for apportionment of their federal congressional and state legislative districts”) (citing U. S. Const., Art. I, §2); Voinovich v. Quilter, 507 U. S. 146, 156 (1993); Reynolds v. Sims, 377 U. S. 533, 586 (1964).
Even in the no longer controversial instance of the one-person, one-vote rule, the adequacy of justification and standard was subject to sharp dispute, and some of the Court’s best minds expressed principled hesitation to go even this far into what has been called the political thicket, see id., at 615 (Harlan, J., dissenting) (“The Court’s elaboration of its new ‘constitutional’ doctrine indicates how far—and how unwisely—it has strayed from the appropriate bounds of its authority. The consequence of today’s decision is that in all but the handful of States which may already satisfy the new requirements the local District Court or, it may be, the state courts, are given blanket authority and the constitutional duty to supervise apportionment of the State Legislatures. It is difficult to imagine a more intolerable and inappropriate interference by the judiciary with the independent legislatures of the States”); Baker v. Carr, 369 *1048U. S. 186, 267 (1962) (Frankfurter, J., dissenting) (“The Court's authority— possessed of neither the purse nor the sword — ultimately rests on sustained public confidence in its moral sanction. Such feeling must be nourished by the Court’s complete detachment, in fact and in appearance, from political entanglements and by abstention from injecting itself into the clash of political forces in political settlements”).
As Professor Issacharoff notes, our vote-dilution eases acknowledged that “the right to cast an effective ballot implied more than simply the equal weighting of all votes .... To be effective, a voter’s ballot must stand a meaningful chance of effective aggregation with those of like-minded voters to claim a just share of electoral results. For this reason, any sophisticated right to genuinely meaningful electoral participation must be evaluated and measured as a group right . . . .” Issacharoff, Groups and the Right to Vote, 44 Emory L. J. 869, 883 (1995); see also Davidson, The Recent Revolution in Voting Rights Law Affecting Racial and Language Minorities, in Quiet Revolution in the South: The Impact of the Voting Rights Act, 1965-1990, p. 23 (C. Davidson & B. Grofman eds. 1994) (“Ethnic or racial vote dilution takes place when a majority of voters, by bloc voting for its candidates in a series of elections, systematically prevents an ethnic minority from electing most or all of its preferred candidates .... Vote dilution not only can deprive minority voters of the important symbolic achievement of being represented by preferred members of their own group, it can deprive them of a committed advocate in councils of government. . . [and] of the substantial benefits that government bestows.. .”).
See Pildes, The Polities of Race, 108 Harv. L. Rev. 1359, 1369 (1995) (reviewing Quiet Revolution in the South: The Impact of the Voting Rights Act, 1965-1990 (C. Davidson & B. Grofman eds. 1994)) (noting that studies of Southern States demonstrate that, as a result of racial-bloc voting, “the probability of a district’s electing a Black representative was less than 1% regardless of a district’s median family income, its percentage of high school graduates; its proportion of residents who were elderly, urban, foreign-born, or who had been residents of the state for more than five years; or the region of the country in which the district was located”); id., at 1375 (finding similar results nationwide). There is, of course, reason to hope that conditions are improving. See infra, at 1076 (discussing elections in which crossover voting favors minority incumbents and in which racial issues have not played a significant role in the outcome). As I discuss in detail in Part IV, infra, I believe that these improvements may be attributed in large part to the effect of the Voting Rights Act, and thus to our willingness to allow race-conscious districting in certain situations.
I recognize, of course, that elsewhere we have imposed prohibitions on the consideration of race, but contexts are crucial in determining how we define “equal opportunity.” Consider our decisions on peremptory jury challenges. There, as in politics, one race may not have had a fair shake from the other. But the differences between jury decisionmaking and political decisionmaking are, I believe, important ones. Politics includes choices between different sets of social values, choices that may ultimately turn on the ability of a particular group to enforce its demands through the ballot box. Jury decisionmaking is defined as a neutral process, the impartial application of law to a set of objectively discovered facts. To require racial balance in jury selection would risk redefining the jury’s role. Without denying the possibility that race, especially as an imperfect proxy for experience, makes a difference in jury decisionmaking (and, in some cases, legitimately so), it seems to me that the better course is to *1052ensure a fair shake by denying each side the right to make race-based selections. The cost of the alternative is simply too great. It is an entirely different matter, however, to recognize that racial groups, like all other groups, play a real and legitimate role in political decisionmaking. It involves nothing more than an acknowledgment of the reality that our concepts of common interest, geography, and personal allegiances are in many places simply too bound up with race to deny some room for a theory of representative democracy allowing for the consideration of racially conceived interests. A majority of the Court has never disagreed in principle with this position. See, e. g., Shaw I, 509 U. S. 630, 642 (1993) (noting that race-conscious redistricting is not always unconstitutional); Miller v. Johnson, 515 U. S. 900, 928-929 (1995) (O’Connor, J., concurring) (consideration of race in the redistricting process does not always violate the Constitution); ante, at 958 (opinion of O’Connor, J.) (noting that strict scrutiny does “not apply merely because redistricting is performed with consciousness of race”).
Leaving aside the question whether such a catholic injury can be a violation of the Equal Protection Clause, there still might be a use of race that harms all district voters because it is used to an unreasonable degree. But see Valley Forge Christian College v. Americans United for Separation of Church and State, Inc., 454 U. S. 464, 489 (1982). But the Court has never succeeded in identifying how much is too much, having adopted a “predominant purpose” test that amounts to a practical repudiation of any hope of devising a workable standard. See Part I-C, infra.
See Cannon v. North Carolina State Bd. of Ed., 917 F. Supp. 387, 391 (EDNC 1996) (describing this “difficult area of the law” and predicting that it will “gain better definition by reason of an imminent decision by the Supreme Court of the United States [in Shaw II]”)', Briffault, Race and Representation After Miller v. Johnson, 1995 U. Chi. Legal Forum 23, 50 (1995) (“[I]t is unclear what work the adjectives ‘predominant’ and ‘overriding’ do in the Supreme Court’s test”); Karlan, Post-Shaw Era 287 (Miller “further unsettled the already unclear roadmap” of Shaw 7); Issacharoff, Constitutional Contours 60 (“[T]he Court’s facile reliance on standards of causation vaguely reminiscent of tort law does nothing to defer confronting the hard issue of acceptable standards of conduct”).
Even in areas where there is no racial-bloc voting, the application of certain traditional districting principles may involve a legitimate consideration of race.
See Karst, Paths to Belonging: The Constitution and Cultural Identity, 64 N. C. L. Rev. 303, 347, 350 (1986) (“[T]he surest path to assimilation is participation in the larger society’s activities and institutions. Voting is not just an expression of political preferences; it is an assertion of belonging to a political community....” “When legislative districts are defined in ways that exclude the possibility of significant minority representation, potential minority voters see that their votes are not worth casting. Yet electoral mobilization is vital... to the group members’ perceptions that they belong to the community”); Walzer, Pluralism in Political Perspective, in The Politics of Ethnicity 1, 18 (S. Thernstrom, A. Orlov, & O. Handlin eds. 1982) (“[Political life is in principle open, and this openness has served to diffuse the most radical forms of ethnic competition”); Kanto-wicz, Voting and Parties, in The Politics of Ethnicity, supra, at 29, 45 (noting that political successes and recognition made members of an ethnic group “feel that it belonged in the wider society . . . [and brought] them inside the political system”); Mintz, Ethnicity and Leadership: An Afterword, in Ethnic Leadership in America 198 (J. Higham ed. 1978) (concluding after reviewing several studies of ethnic politics that “we ignore at our peril the need to understand those processes by which being shortchanged ... politically can become any group’s motto or battle standard”); cf. Karlan, Our Separatism 102 (“two generations of communist suppression and ethnic and religious tension in Yugoslavia did little to ensure stability, tolerance, or integration”).
See, e. g., Nolan, Boston Mayoral Race Could Break Dominance of Ethnicity, Boston Globe, Apr. 9,1993, p. 40 (“When Boston finishes choosing a new mayor, the city may discover that after centuries of immigration, ethnicity is no longer the dominant factor in its politics”); Black, Once-Solid Voting Blocks are Splitting in Boston, Boston Globe, Nov. 1, 1993, p. 1 (commenting that voters consider Menino’s Italian descent “little more than a historical footnote” and observing that “ethnic voting has faded ... [a]s various groups enter the American economic and social mainstream ... [and] gain some semblance of [political] power”); DTnnocenzo, Gulotta Can’t Count on Ethnicity, Newsday, Oct. 19,1993, p. 97 (noting that “[t]he vowel at the end of Tom Gulotta’s name may not matter in this year’s county executive election as it once did” because “Italian-Americans in Nassau County are likely to go to the polls with more than ethnic favoritism in mind”; attributing the decline in ethnicity-based voting to the *1076fact that “Nassau Italian-Americans feel less marginali[zed] as an ethnic group”).