dissenting:
In VMI I, 976 F.2d 890, we held unanimously that the Commonwealth of Virginia’s official policy of allowing only men to be educated at state-supported Virginia Military Institute violated the Equal Protection Clause. Specifically, we held, applying intermediate level scrutiny under developed Equal Protection jurisprudence, that if, as the Commonwealth then asserted, the “important governmental objective” its policy served was the provision for its citizens of a diverse array of educational opportunities, ineluding single-gender education, then providing one single-gender institution for men but none for women could not be deemed “substantially related to achievement” of that objective. Id. at 892, 899. Furthermore, we expressed doubt that the asserted diversity-of-educational-opportunities objective could stand scrutiny as the actual reason for maintaining VMI’s male-only policy. Id. at 899 (pointing to lack of any state-announced policy of providing single-gender education as part of overall “diversity” goal; to the failure of the Commonwealth to defend the policy in this litigation; and to the fact that the actual policy being overwhelmingly followed by the Commonwealth’s colleges and universities was coeducation rather than single-gender education).
In any event, whether because the asserted governmental objective of “diversity” was not a credible reason for the policy, or because, if it were, maintaining one male-only *1243institution in the overall system could not be deemed substantially related to such an objective, we held the policy violative of equal protection guarantees. And, in keeping with established judicial policy where comparable forms of systemic state action have been found violative of equal protection, see, e.g., White v. Weiser, 412 U.S. 783, 794-95 (1973) (electoral redistricting), we remanded with directions to allow the Commonwealth to make the first attempt at remedy. VMI I, 976 F.2d at 900. In doing so, we noted the two obvious remedies: admitting women to VMI (going co-ed) or foregoing further state support (going private). Id. Additionally— and I believe prudently, if with risk — we noted the possibility, without pre-judging the validity of any effort to realize it, of establishing “parallel institutions or programs.” Id.
The Commonwealth opted for the “parallel program” possibility as an attempted remedy and submitted to the district court the proposed plan summarized in the majority opinion. Finding it adequate if properly implemented over time to satisfy equal protection guarantees, the district court adopted it in the form of an injunctive decree that directed compliance “with all deliberate speed.” 852 F.Supp. 471, 485 (W.D.Va.1994). The panel majority has now affirmed the district court’s decree and the critical findings and conclusions on which it is based.
With all respect, I would not do so. I do not believe the proposed remedial plan, whose judicial adoption in unrealized form obviously does not bring Virginia into present compliance with equal protection guarantees, has any real and effectively measurable capacity to do so over foreseeable time.
I therefore dissent. I would hold that the proposed remedial plan fails, as did the policy rejected in VMI I, to pass equal protection muster under the appropriate intermediate level of scrutiny. Accordingly, I would reject the plan, declare the VMI men-only policy still in violation of the Equal Protection Clause, and order that the violation be ended either by abandoning the policy or by foregoing further state support for the institution.
I
Though the legal framework is well known and the general historical background of this litigation is not in dispute, a brief summary is needed to aid in identifying the exact constitutional issue that is now before us.
When Virginia Military Institute was founded in 1839 as a state-supported military school for men only, it is inconceivable that any thought was given by the founders to the possibility that women should not be denied its intended benefits. No conscious governmental choice between alternatives therefore dictated the original men-only policy; it simply reflected the unquestioned general understanding of the time about the distinctively different roles in society of men and women. See Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 725 n. 10 (1982) (noting numerous examples from that era of “legislative attempts to exclude women from particular areas simply because legislators believed women were less able than men to perform a particular function”). Since that time and until this litigation (so far as anything before us reveals) no conscious governmental choice had ever been made by the Commonwealth of Virginia to reexamine that original policy. So far as can be told, the gender-role premises of its origins were those that continued over time to sustain it as official state policy.
It is clear then that it was this litigation that prompted the Commonwealth’s first official re-examination of the policy and its underlying premises in light of the Fourteenth Amendment’s requirement that the states provide the equal protection of their laws to all persons subject to them. That obligation, as imposed in 1868, has from earliest times been understood by the courts to expose gender-classifications to equal protection judicial scrutiny. Early on, that scrutiny was almost completely deferential to the legislative prerogative, asking only whether the classification served any reasonably conceivable, legitimate governmental purpose. See, e.g., Bradwell v. Illinois, 83 U.S. (16 Wall), 130, 141 (1872) (standard applied to uphold law prohibiting women from practicing law). Since 1976, however, the Supreme Court, confirming a trend toward some degree of *1244heightened scrutiny that started in the early 1970’s, see, e.g., Reed v. Reed, 404 U.S. 71 (1971) (invalidating state law that preferred men over women as administrators of decedents’ estates), has interpreted the Clause to require a significantly more stringent standard, “intermediate” between the “strict” scrutiny required for racial and other historically “suspect” classifications, and the most deferential “rational basis” scrutiny originally applied to gender-based classifications. As expressly adopted in Craig v. Boren, 429 U.S. 190 (1976), this intermediate level of scrutiny asks whether the state’s gender-classification “serves important governmental objectives” and is “substantially related to achievement of those objectives,” id. at 197. Under this standard, states seeking to uphold such classifications “carry the burden of showing an ‘exceedingly persuasive justification’ for [it]” by demonstrating both that the governmental objectives it asserts for the classification are “important” ones and that “the discriminatory means employed are substantially related to achievement of those objectives.” Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 724 (1982).
It was this intermediate level of scrutiny that we applied in VMI I in holding the original male-only policy violative of equal protection. That holding still stands. Unless and until it is overruled, the original policy — 'which still remains in effect — remains unconstitutional. The district court’s decision that we now review does not of course purport to hold otherwise. It assumes, as it must, the continuing unconstitutionality of that policy, but holds that the violation may be effectively remedied by the state’s compliance with the injunctive decree entered by the court in adopting the state’s proposed “parallel program” plan for women only at Mary Baldwin:
Several important things emerge from those developments. The first is that the remedial plan proposed by the Commonwealth and adopted as remedy by the district court simply involves a new gender-classification which now has become the proper subject of the heightened scrutiny mandated by Craig and its progeny. Cf. White v. Weiser, 412 U.S. 783, 795 (1973) (judicial review of remedial redistricting plan asks only whether plan meets constitutional requirements, not whether it provides best possible remedy for original violation).
The next point of importance is that this new gender-classification (in its projected form) is of a type that has not yet been definitively subjected to equal protection scrutiny: it involves a state’s provision of separate single-gender educational institutions for men and women which it is claimed will meet equal protection requirements by providing substantially equal, though separately administered, benefits. This could raise a threshold question whether separate state-supported educational facilities for men and women, like those for white and black students, are so “inherently unequal,” by reason of their stigmatic implications, see Brown v. Bd. of Educ., 347 U.S. 483, 495 (1954), that the new classification violates equal protection per se and warrants no further scrutiny.
If the answer to that threshold question is, however, “no”, so that intermediate scrutiny must proceed in detail, a final point of importance about the new classification must be faced. It is that one of the two critical elements in its separate-but-equal arrangement, the women-only program at Mary Baldwin, is only a plan and not a present reality. This creates a difficult problem for Equal Protection analysis. Must we assume, without question, that the stated goals of the women’s program are actually achievable and that the fact of their achievement is subject to judicial verification when it occurs, so that we should, on that assumption, (though conditionally) assess the plan in its proposed ultimate form?1 Or may we, in intermediate scrutiny, question either or both the achieva-bility of the program’s stated objectives and the ability of the courts effectively to assess their achievement? If we undertake condi*1245tional assessment of the plan on the stated assumption, what is the proper equal protection test for allowable separate-but-equal state-supported educational institutions? What is the proper measure of equality for that purpose?
Each of these inescapable problems raises for us issues of first impression in application of equal protection jurisprudence to the resolution of this case.
II
The logical first question is whether separate single-gender undergraduate educational facilities for men and women are “inherently unequal” so that the proposed plan, even if perfectly realized in time, would be per se violative of equal protection. Cf. id. The question has not been addressed by the Supreme Court, see Hogan, 458 U.S. at 720 n. 1, or by this court, hence is an open one. Under the disposition I believe proper, it could remain open, for I would decline to address it, and hold that even if some separate-but-equal arrangement might pass equal protection muster, the one here proposed would not.2
III
This leads to the next question: whether the particular separate-but-equal arrangement proposed by the Commonwealth and adopted by the district court can survive intermediate equal protection scrutiny.
As earlier noted, the fact that the women-only component of this arrangement exists now only in plan form presents a difficult analytical problem: whether its consummation in fact should be assumed, with scrutiny then confined to the consummated overall plan, or whether the possibility of effective, judicially verifiable consummation of the plan may itself be questioned. Because I believe that even were the VWIL proposal to be substantially consummated in foreseeable time the resulting two-component arrangement would not pass equal protection muster, I would proceed on that assumption, though with some reservations to be expressed about the practical enforceability of the injunctive decree that embodies the proposal.
A
In its fully consummated form, the Commonwealth’s proposed arrangement would consist of two separate single-gender undergraduate institutions, one for men only, the other for women only. The basic structure of each has been accurately summarized in the majority opinion and is not in dispute. A brief recapitulation of the core aspects suffices here.
The men-only component would be Virginia Military Institute, a justly famous and distinguished state-supported four-year liberal arts college organized and operated since 1839 in the classic “military school” model, featuring a student body now numbering around 1,300 men organized as a quasi-military “Corps of Cadets” and a distinctive “adversative” social and educational methodology designed to produce a distinctive type of “citizen-soldier” particularly suited for military and civic leadership.
The women-only component would be the Virginia Women’s Institute for Leadership *1246(VWIL) operated under contract with, and funded by, the Commonwealth of Virginia, as part of the undergraduate program at the otherwise privately-funded Mary Baldwin College. This “Institute,” whose essential structure and stated mission are accurately summarized in the majority opinion, ante at 1233-35, would have come into existence in the Fall of 1995 at the earliest, around a century and a half after VMI’s founding. While its future enrollment is necessarily uncertain, it would be expected to start up with about 25 to 30 students and expand as a concededly problematic demand for its highly specialized program allowed.
As indicated, for purposes of decision here, I would lay aside all concerns about whether the VWIL program would actually ever work out substantially as proposed, assume that it would be, and subject the resulting two-component arrangement to intermediate level equal protection scrutiny. That is, I would ask whether the resulting provision of separate state-supported men-only and women-only educational opportunities at VMI and VWIL respectively could meet that standard.
B
The first step in that process is to identify the precise governmental objective(s) the Commonwealth asserts to be the “important” one(s) justifying the proposed double gender-classification under which women will continue to be denied admission to VMI and men to VWIL. For in intermediate level scrutiny, unlike rational-basis scrutiny, we are limited to consideration of the objectives specifically advanced by the state, and may not look beyond those to any our imaginations might seize upon as justification. See L. Tribe, American Constitutional Law, § 16-32, pp. 1604-06 (2d ed.1988).
Though usually the governmental objectives relied upon to justify gender (and other) classifications are plainly enough articulated by their state defenders, that is not so true here. There is a real problem of identification in this case, for the Commonwealth seems uncertainly to advance a number as alternative or cumulative free-standing possibilities. Three might be identified: (1) providing separate single-gender educational facilities for both men and women because of the intrinsic value to some in both genders of such a social environment for education (“intrinsic value”);3 (2) producing both men and women particularly suited for leadership roles as “citizen-soldiers” by providing separate single-gender educational programs for each that are designed to accommodate their different psychological and emotional strengths and weaknesses in becoming effective leaders in either domain (“gender-adapted leadership training”);4 and (3) providing separate single-gender educational facilities for men and women as part of an overall objective of providing a diverse array of state-supported higher-education opportunities (“system-diversity”).5
If these be, alternatively or together, the “governmental objectives” now asserted by the Commonwealth, we are entitled at the outset to inquire as to whether they are the “actual purposes,” and to reject them if the record draws their reality as the true motivations for the policy sufficiently in doubt.6 This was exactly what the Supreme Court did in rejecting the State of Mississippi’s assertion in Hogan that its primary objective in maintaining its School of Nursing for women only was to compensate for past discrimination against them. Looking to the history of the School’s founding and subsequent operation, to statistics respecting the actual dominance of women in the nursing profession throughout that history, and to state legislative history, the Court concluded that “although the state recited a ‘benign, compensatory purpose,’ it faded to establish that the alleged objective is the actual purpose underlying the discriminatory classification.” Hogan, 458 U.S. at 730. The real *1247purpose behind the original policy and its continuation through history was implicitly recognized by the Hogan court as being simply the carrying through of a “stereotyped view of nursing as an exclusively women’s job.” Id. at 729. See also Califano v. Goldfarb, 430 U.S. 199, 212-17 (1977); Weinberger v. Wiesenfeld, 420 U.S. 636, 648 (1975) (noting that “the mere recitation of a benign’ compensatory purpose is not an automatic shield which protects against any inquiry into the actual purposes underlying a statutory scheme”).
I believe that a comparable inquiry here could properly support a like rejection of the various governmental objectives suggested by the Commonwealth — on the basis that they demonstrably are rationalizations compelled by the exigencies of this litigation rather than the actual overriding purpose of the proposed separate-but-equal arrangement. Such an inquiry — looking realistically to the historical record, taking judicial notice of much of relevance that is known to the whole world and of which we are not compelled to feign ignorance, see Watts v. Indiana, 338 U.S. 49, 52 (1948), and holding the Commonwealth to its appropriate stringent burden of justification, see Hogan, 458 U.S. at 724 (must be “exceedingly persuasive”) would, I believe, reveal a quite different actual purpose. Specifically, I think it would support a confident and fair conclusion that the primary, overriding purpose is not to create a new type of educational opportunity for women, nor to broaden the Commonwealth’s educational base for producing a special kind of citizen-soldier leadership, nor to further diversify the Commonwealth’s higher education system — though all of these might result serendipitously from the arrangement — but is simply by this means to allow VMI to continue to exclude women in order to preserve its historic character and mission as that is perceived and has been primarily defined in this litigation by VMI and directly affiliated parties.7
To reach such a conclusion would no more question the good faith of the Commonwealth in advancing these claimed governmental objectives in this litigation than did the Supreme Court’s rejection of the objectives advanced by Mississippi in the Hogan litigation. It would simply involve the same realistic recognition that the objectives advanced represent after-the-fact rationalizations that, quite understandably, may be advanced by any state required in litigation to justify a gender-classification whose seeds were planted long before equal protection jurisprudence had come into being or had evolved to the point of drawing it in question. Cf. Cleveland Board of Education v. LaFleur, 414 U.S. 632, 653 (1974) (Powell, J., concurring in result) (urging analysis of mandatory pregnancy leave policy under equal protection doctrine, and rejecting under such an analysis “most of the after-the-fact rationalizations *1248proposed by the [state agency defendant]” as “unsupported by the record”).
A conclusion that the actual, overriding purpose of the proposed separate-but-equal arrangement remains the preservation by that means of the original 1839 policy of excluding women from VMI, a policy that unquestionably has been driven unchanged since its origins by a stereotyped view of the proper role and capabilities of women in society, would of course require declaring the proposed arrangement violative of equal protection without further inquiry into specifics. See Hogan, 458 U.S. at 729, 730.
Although, as indicated, I believe a decision on that ground would be proper, I would not decide the case on that basis alone, or even primarily. There are unique circumstances here that were not present in Hogan or in any other case of which I am aware in which a state’s asserted objectives have been rejected at the threshold as demonstrably not the “actual purpose” of a challenged gender-classification. Uniquely, the gender-classification under specific challenge here is one now defended by the Commonwealth as a proposed judicially required remedy for a so-' far unsuccessfully defended prior gender-classification. The real position of those who defended the original policy remains that the Commonwealth should not have been required to undertake any remedial action, hence that it need have no justifying objectives for the new remedial gender-classification it proposes. See supra, note 7. Though the Commonwealth does not press the point, I think it fair to recognize that in these circumstances, unlike those where the gender-classification being defended is one prompted entirely by voluntary state action, courts should be especially cautious about rejecting as not “actual” the objectives advanced for involuntarily undertaken remedial action. Accordingly, though I believe the Commonwealth must defend its remedial plan under the usual intermediate scrutiny standard, I think it is entitled to have its proposed separate-but-equal gender-classification assessed for the substantiality of its “fit” to the remedial objectives it now asserts.
C
Assuming then for purposes of this case that the governmental objectives earlier identified should be accepted as reflective of the “actual purposes” of the proposed plan despite my stated doubts about their reality as other than compelled remedial rationalizations, the next question is whether they have also been shown to be “important” and not merely “rational.” Here again, because of the conceptual difficulties presented by the remedial context of the case, I would assume arguendo8 the importance of the governmental objectives asserted by the Commonwealth and proceed to the second inquiry under the proper equal protection test: whether the Commonwealth has made an “exceedingly persuasive” showing that the gender-classification central to its proposed separate-but-equal arrangement .is “substantially and directly related to its proposed [remedial] objectives.” Hogan, 458 U.S. at 724, 730. I would hold not, as the primary ground for decision that the proposed plan does not pass constitutional muster.
What is the “substantial and direct relationship” — the “fit” — between means and asserted ends for which we search in intermediate scrutiny, and how do we look for it — in general, and particularly in this case? The general question whether a challenged classification “is substantially related to its asserted goals” has been characterized as “at best an opaque one.” Michael M. v. Superior Court of Sonoma County, 450 U.S. 464, 474, *1249n. 10 (1981) (plurality opinion) (emphasis in original). Though as this rightly observes, the substantive inquiry is likely to be difficult, two critical aspects of the inquiry process are plain enough.
1. The inquiry is one of law — of constitutional law — so that review of a lower court’s determination of the issue, (though not of any underlying factual predicates) is plenary. See, e.g., Wengler v. Druggists Mutual Ins. Co., 446 U.S. 142, 150-52 (1980) (plenary review of State Supreme Court holding).
2. The requirement that the relationship between discriminatory means and asserted goals be “substantial and direct,” rather than merely “rational for any conceivable purpose,” mandates an inquiry into available alternatives, including gender-neutral ones. This is not to determine whether the fit is the best one possible, but to ensure that the means chosen did not by-pass reasonably available alternatives less discriminatory or not at all discriminatory in their impact on the disfavored gender. See, e.g., Wengler, 446 U.S. 142, 151 (1980); Orr v. Orr, 440 U.S. 268, 283 (1979).
The specific issue thus becomes whether the Commonwealth has sufficiently shown within these principles, that its proposed separate single-gender school arrangement is directly and substantially related to the achievement of the three governmental objectives earlier identified in short-form as the “intrinsic value,” “gender-adapted leadership training,” and “system-diversity” objectives. Supra at 1246.
As earlier indicated, the question of the fitness of any such separate single-gender school arrangement to achieve any governmental objective apparently is one of first impression in contemporary equal protection jurisprudence. One aspect of the matter, however, seems clear at the outset to me: no such arrangement could be found substantially related to any conceivable governmental objective unless the benefits to be separately distributed by the arrangement were substantially equal across the board of the relevant criteria for evaluating educational institutions. The Supreme Court’s reference, in dicta, to the possibility of such an arrangement in Hogan seems to assume such an equality of benefits as a given, see Hogan, 458 U.S. at 720 n. 1, and I do not see how it could be otherwise under contemporary equal protection jurisprudence. Certainly, when separate-but-equal educational arrangements for the races were considered to be tolerable under the Equal Protection Clause, a basic prerequisite was that they be truly, substantially equal in all the relevant criteria, tangible and intangible, by which educational institutions are evaluated. See e.g., Sweatt v. Painter, 339 U.S. 629, 633-34 (1950) (requiring “substantial equality in educational opportunities” to justify separate state-supported law schools for white and black students, and not finding it upon considering both tangible resources such as “scope of library” and intangible resources such as “position and influence of the alumni,” “traditions and prestige”). Though race is a “suspect” classification and gender so far is not, I see no reason why the same requirement of substantial equality of benefits that was thought at one time to justify separate-but-equal schools for the different races should not apply to separate schools for men and women if that classification now. does, as race formerly but no longer does, permit separate-but-equal arrangements. If that be so, then no governmental objective whose achievement is specifically dependent upon the utilization of separate single-gender institutions to distribute educational benefits could possibly justify a significant discrimination between the two in terms of the basic content, or quality, or quantity of those benefits — tangible and intangible. Thus, I would think a state could not justify under intermediate scrutiny the provision of a men-only engineering school and a women-only nursing school as a means of achieving such asserted objectives as system-diversity, or the intrinsic value to some in each gender of single-gender educational environments, or the like. Neither could it justify the provision of separate graduate or undergraduate institutions having comparable educational programs and missions, but also having wide disparities favoring one gender over the other in matters of physical plant, annual funding, faculty or like commonly understood measures of value.
*1250Does this mean that there is no way that a state constitutionally could set about achieving such governmental objectives through the provision of separate single-gender institutions? I think it does not necessarily mean that, but the arguably acceptable means would seem to me to be very narrowly circumscribed. If we looked for the arrangement most likely to survive scrutiny, it presumably would involve simultaneously opened single-gender undergraduate institutions having substantially comparable curricular and extra-curricular programs, funding, physical plant, administration and support services, and faculty and library resources. Such an arrangement would involve no gender-line discrimination in terms of tangible benefits, nor of intangible benefits such as tradition, prestige and alumni influence — as to which each starts with none. Nor could there be any stigmatic implications arising from the substantially comparable content of its educational program. If any arrangement involving separate-but-equal single-gender institutions set in place to achieve governmental objectives of system-diversity, or of accommodating valid preferences in each gender for a single-gender educational environment, could survive equal protection scrutiny, it surely would be one such as that posited.
Taking it as the paradigm against which to measure the proposed arrangement reveals how far short the proposed plan falls from providing substantially equal tangible and intangible educational benefits to men and women. Without denigrating in any way the proposed VWIL program, nor certainly Mary Baldwin, the contrast between the two on all the relevant tangible and intangible criteria is so palpable as not to require detailed recitation. If every good thing projected for the VWIL program is realized in reasonably foreseeable time, it will necessarily be then but a pale shadow of VMI in terms of the great bulk, if not all of those criteria. Particularly is this obvious with respect to the intangibles such as prestige, tradition and alumni influence which the Supreme Court, looking for substantial equality of educational opportunities in Sweatt, thought “more important” even than tangible resources. 339 U.S. at 634. The student and eventual graduate of VWIL will not be able to call on the prestigious name of “VMI” in seeking employment or preference in her various endeavors; the powerful political and economic ties of the VMI alumni network cannot be expected to open for her; the prestige and tradition of her own fledgling institution cannot possibly ever achieve even rough parity with those of VMI. The catch-up game is an impossible one, as any honest reflection upon the matter must reveal.
The district court and the majority apparently seek to avoid the insurmountable problem of finding substantial equality of benefits by narrowly defining the relevant range of those to be considered. If inquiry is confined only to those benefits sought by those relatively few women who are expected to self-select VWIL primarily for its single-gender environment, it becomes very easy to ascribe not only substantial equality, but superiority, to the benefits available at VWIL. But that, of course, won’t do. The proper perspective from which to measure substantial equality of available benefits is that of the potential student who could be admitted to either school and has a choice. As was said in Sweatt, “[i]t is difficult to believe that one who had a free choice between [these] schools would consider the question close.” Id.
The implication of all this is, as I realize, a stark one. No separate single-gender arrangement that involved VMI as the all-mens’ school and any newly-founded separate institution (whether free-standing or an appendage) as the all womens’ component could pass equal protection muster. It could not provide substantially equal educational benefits or opportunities to both genders.
This may be most obvious when the proposed arrangement is tested for fit against the “system-diversity” and “intrinsic value” objectives. The “gender-adopted leadership training” objective poses a slightly different problem. The benefit upon which it concentrates is a projected outcome: that of being one especially suited for military and civilian leadership by virtue of training adapted to different gender-characteristics, as “citizen-soldier.”
*1251As to this particular objective, I will close by noting a process reservation beyond the substantive concerns for achieving substantial equality of the outcome goal. It seems to me too amorphous an objective to permit any principled judicial assessment as the VWIL program is expected to evolve. When can it first be assessed? Surely not earlier than the four years it will take to produce the first graduate presumably trained for that special leadership role. Must it not actually await an additional period for putting the training to test in the military and civilian domains? How will it be assessed even then: by comparing, on a proportional basis, the actual leadership positions achieved by graduates of the two schools? My pessimistic assessment is that one of two things will occur. One, this particular governmental objective — actually a critical one as advanced — will simply be allowed to fall out of sight in the judicial monitoring of results that is projected. Two, its attempted monitoring will generate an absolute quagmire of conflicting contentions about achievement of the objective.
It will not work.
. The district court necessarily made that assumption. Implicit in its decision is the determination that (1) if the asserted objectives of the women-only program at Mary Baldwin are achieved, the result will be a separate-but-equal provision of benefits that passes equal protection muster and (2) achievement of the objectives is a realistic possibility capable of verification by the cotut when it occurs.
. The United States has expressly disclaimed any contention that any and all forms of state-supported single-gender education are per se viola-tive of equal protection. See VMI I, 976 F.2d at 898. This position would seem compelled by the Supreme Court’s recognition in Hogan that a single-gender educational institution might be justified on the basis of need to compensate the favored gender for past discrimination. 458 U.S. at 727, 728. But that is a different issue than the issue whether the provision of allegedly separate-but-equal single-gender facilities for purposes other than compensation for past discrimination against one of the genders would violate equal protection per se because — as in matters of race — such separateness is “inherently unequal.” The "question whether states can provide 'separate but equal' undergraduate institutions for males and females” was, in fact, expressly noted by the Supreme Court in Hogan as still an open one. Hogan, 458 U.S. at 720 n. 1. I do not therefore understand the United States' disclaimer to run as well to noncompensatory "separate but equal” arrangements, particularly in view of its suggestion of their necessary stigmatic implication. Appellant's Br. 20-22; Reply Br. 9, 10. For this reason, I do not think the issue whether non-compensatoiy "separate but equal” arrangements are per se violative can be avoided as waived, but, as indicated, would myself avoid it as unnecessary to decision in this case.
. Appellees' Br. at 4, 5.
. Id. at 2-4, 32-37.
. Id. at 14 & n. 5.
.Such an inquiry logically precedes inquiry into the "importance” of any objectives accepted as reflective of "actual purpose.” See Hogan, 458 U.S. at 730. As to how the "importance” inquiry would work out in this case, see infra, at 1248 & n. 8.
. As the record indicates, and as it is well to recall at this stage of the litigation, the Commonwealth of Virginia did not officially defend the original VMI men-only policy that we held to be unconstitutional in VMI I. The justification for that policy advanced in VMI I was exclusively shaped and actively conducted by VMI, its official governing board, that board's members and (as intervenors) VMI alumni organizations. See VMI I, 976 F.2d at 894 & n. 3. The justification then advanced by those parties is fairly and simply summarized:
VMI's distinctive educational program, featuring rigorous military discipline and an "adversative” methodology, is suitable only for men and not for women, to the point that the admission of any women into it would effectively destroy it; the demonstrated value to society of that program and those it has produced is too important to allow it to be destroyed in that way.
See VMI I, 976 F.2d at 896-97.
That justification and the perception underlying it has not been abandoned by those who advanced it. They continue to press it by "protective” cross-appeal on this appeal after having sought to challenge its rejection in the Supreme Court.
Taking judicial notice of matters surely of common knowledge in the Commonwealth, I would be prepared to conclude that (1) the perception underlying the policy justification advanced by VMI officials and alumni organizations remains alive and strongly held by those parties, and that (2) the prestige and influence of VMI and its justly loyal alumni and their organization in influencing any political decision affecting VMI's interests is sufficiently powerful to ensure that their overriding purpose in this matter effectively defines the actual governmental objective of the Commonwealth's proposed remedial plan. That overriding purpose remains the preservation of VMI as a state-supported educational institution for men only, with all other asserted purposes of the plan merely secondary means to that end.
. I make the assumption arguendo because resolution of the special conceptual problems respecting remedial objectives in intermediate-level scrutiny that would be required is not necessary to the decision I would reach. In making the assumption, I note, with all respect, that I do not agree with the majority’s assertion that when we do assess "importance” we owe great deference to legislative judgments on the matter. I believe instead that in intermediate scrutiny the state's burden of showing the "importance" of asserted objectives is, as with all aspects of its justification defense, a difficult one that provokes non-deferential judicial balancing of the objectives asserted against affected private interests. See, e.g., Craig, 429 U.S. 190, 196 (1976) (asserted governmental objective of administrative convenience not sufficiently "important” to justify gender-classification being challenged; no deference accorded legislative judgment).