(dissenting). In recognition that education is the only meaningful weapon society currently has to combat the spread of AIDS, the Commissioner of the State Education Department has promulgated regulations requiring all elementary and secondary schools within the State to develop courses of "appropriate instruction” on that subject as part of their health education programs (8 NYCRR 135.3 [b] [2]; [c] [2]). We have no quarrel with the Commissioner’s goal of instituting a program of mandatory AIDS education, nor do we question the Commissioner’s decision to implement that goal by requiring the establishment of community-based advisory councils to make "recommendations concerning the content [and] implementation” of the AIDS curriculum (id.). We do find it objectionable, however, that the Commissioner has seen fit to require the presence of "representatives from religious organizations” on these advisory committees (see, id.). In our view, this requirement has the inevitable and primary effect of both favoring and endorsing religion and, as such, violates a "core purpose of the Establishment Clause” (Grand Rapids School Dist. v Ball, 473 US 373, 389). Accordingly, we dissent.
*344Initially, we do not dispute that the tripartite Lemon test (Lemon v Kurtzman, 403 US 602) is the controlling standard for measuring the constitutional validity of the regulations challenged here. Nor do we find fault with the majority’s general discussion of the essential precepts of the Establishment Clause, including its acknowledgment that the Clause proscribes not only discrimination against some denominations but also discrimination in favor of religion in general (majority opn, at 338, citing Grand Rapids School Dist. v Ball, supra). Our quarrel with the majority lies, instead, with its application of these broad principles to the particular problem raised by the challenged regulations.
Since plaintiff does not argue that the regulations have a secular purpose, we will not here pursue that aspect of the tripartite Lemon analysis. Indeed, we find little reason to explore the degree to which the religious-representative requirements of 8 NYCRR 135.3 (b) (2) and (c) (2) have a goal that is sufficiently secular to pass muster under Lemon, since, contrary to the majority’s holding, we conclude that the regulations’ effect is one that is unacceptable under the second prong of the Lemon inquiry.
It is by now axiomatic that to avoid objection under the Establishment Clause a statute must have a " 'principal or primary effect * * * that neither advances nor inhibits religion’ ” (Mueller v Allen, 463 US 388, 394). It is also axiomatic that government may not adopt measures which have the effect of endorsing religion, even symbolically (Grand Rapids School Dist. v Ball, supra, at 389-390). The government must "maintain a course of neutrality among religions and between religion and nonreligion” (id., at 382). Finally, the area of government educational programs for children is one where special caution is warranted because children are highly impressionable and the symbolism of a union between church and State is likely to have a magnified impact (id., at 383, 390). When viewed in light of these principles, the challenged regulations cannot be regarded as anything other than a patent violation of the Establishment Clause.
Initially, the majority’s conclusion that the regulations do not have the "primary effect” of advancing religion appears to be based in large measure on the indisputable premise that the regulations serve a useful secular purpose (see, majority opn, at 340). In this respect, its analysis is simply incomplete, since Lemon v Kurtzman (supra) and its progeny require a *345showing of both a legitimate secular purpose and a "primary effect” that neither advances nor inhibits religion.
Furthermore, the majority’s analysis is constricted by an overly narrow view of the problem the challenged regulations present. The central focus of the majority’s opinion seems to be the degree to which the religious views of the advisory committees’ religious representatives are likely to have a concrete impact on the AIDS curriculum. Indeed, the entire tenor of the majority’s opinion suggests that it views the problem presented here solely as one involving a potential religious interference with school curriculum decisions (see, e.g,, Epperson v Arkansas, 393 US 97; Engel v Vitale, 370 US 421). Of course, if that were the only problem, the majority’s conclusion would be unobjectionable, since, as the majority states, it cannot be assumed that the religious views of the mandatorily designated committee members will necessarily find their way into the AIDS curriculum. However, as will be demonstrated below, that is only one minor aspect of the "effect” that these disputed regulations will have.
By requiring that all AIDS education advisory committees include representatives of religious organizations, the State is declaring, in the most direct terms, that, above all other interested groups in society, religion has a unique and indispensable role to play on the councils of government (cf., Bowen v Kendrick, 487 US 589, 605 [challenged legislation does not suggest "that religious institutions * * * are uniquely well qualified to carry out (funded) services”]). The unmistakable message is that the views of religious leaders enjoy a favored status and that relevant secular viewpoints, such as those of psychologists, social workers and business leaders are of substantially less significance.1 The "endorse*346ment” of religion, as distinguished from "nonreligion,” could not be clearer (see, Grand Rapids School Dist. v Ball, supra, at 382).
Furthermore, the effect of the challenged regulation goes beyond a merely symbolic endorsement of religion. By guaranteeing their leaders a seat on the schools’ AIDS education advisory panels, the State has given religious adherents automatic access to the governmental decision-making — a benefit not conferred on any other group of citizens. As a consequence, religious adherents, through their representatives, are afforded greater participatory privileges, in direct violation of the fundamental neutrality principle. Indeed, the twin evils that Justice O’Connor identified in Lynch v Donnelly (465 US 668, 688 [O’Connor, J., concurring]) are directly implicated here. 8 NYCRR 135.3 (b) (2) and (c) (2) "give [religious] institutions access to government or governmental powers not fully shared by nonadherents” and, as a consequence, those regulations "foster the creation of political constituencies defined along religious lines”2 (id., at 688).
Nor can it be said that the threat to the values implicit in the Establishment Clause is any less substantial because the role that religious leaders are granted is merely advisory (see, majority opn, at 340-341). Plainly, if a regulation directed all public school principals to consult with the parish priest or local rabbi before developing a history or science curriculum, few would dispute that the Establishment Clause had been vio*347lated. Yet, that is precisely the import of the regulations at issue here.
While it is true, as the majority notes, that the Establishment Clause does not "preclude government from ** * * taking religion into account” (Wallace v Jaffree, 472 US 38, 70), it certainly does not follow that government may give religion preferential access to its decision-makers’ ears. Nor does it follow that the Establishment Clause permits government to confer such favored status on religion over other viewpoints whenever some secular goal, such as that identified by the majority (majority opn, at 340), is advanced.3 Indeed, in its effort to find a "primary” secular effect to satisfy the second prong of the Lemon test, the majority has, ironically, itself adopted a position that the Establishment Clause forbids.
The majority argues that mandatory involvement of representatives of religious views is desirable and has a primarily secular effect because it will ensure that the resulting product will reflect "a greater societal consensus” and will thereby "neutraliz[e]” potential community opposition (majority opn, at 340). However, that is simply another way of saying that the regulations are valuable because they enable government to co-opt the views of the religious community by incorporating, or at least appearing to incorporate, them in its educational programs. Once again, even this purported "secular effect” that the majority has identified has as its true "primary” effect the endorsement of religion, in direct violation of the *348Supreme Court’s prohibition in Lemon v Kurtzman (supra, at 612-613) and Grand Rapids School Dist. v Ball (supra).
Moreover, by relying on the benefits that an alliance with religious leaders might have for government, the majority has fallen into one of the traps that are associated with governmental efforts to use religion to accomplish its own secular goals. As one commentator has noted, when the government uses religion, or religious leaders, as a tool, it is actually exploiting a “borrowed aura of * * * legitimacy,” which, because of the "fundamental importance that religion holds for many people,” may lend its policies an undeserved level of credibility (Tribe, American Constitutional Law § 14-15, at 1287 [2d ed]).
Yet another problem with the majority’s proffered justification for upholding the regulations is its insensitivity to the feelings and opinions of those who are not religious adherents. The majority’s assumption that a curriculum adopted after consultation with religious leaders will have “greater secular effectiveness” and will “neutraliz[e]” possible community opposition overlooks the fact that, in addition to religious adherents, the community includes individuals who not only do not adhere to a particular religious view but are also opposed to educational programs that, directly or indirectly, reflect the viewpoints of religious leaders. Thus, the majority’s analysis itself accomplishes precisely what Justice O’Connor cautioned against in Lynch v Donnelly (supra, at 688) — "send[ing] a message to nonadherents that they are outsiders, not full members of the political community, and an accompanying message to adherents that they are insiders, favored members of the political community.”
Finally, the majority has not explained how school districts attempting to implement the challenged regulations can avoid the other vice that often accompanies rules that give religion a preferred position, i.e., favoring one denomination over another. Since the number of positions on the advisory boards mandated by 8 NYCRR 135.3 (b) (2) and (c) (2) is finite, school authorities will have to make choices among the leaders of the various sects in their communities. Some religions will be singled out for the favored status of committee membership, while others will necessarily be excluded. As a consequence, the principle of denominational neutrality will inevitably be violated (see, Tribe, op. cit., at 1286).
In sum, the religious-representative requirement of the *349Commissioner’s AIDS education rules runs directly counter to the primary objective of the Establishment Clause. Individual religious leaders, like the leaders of other segments of the community, undoubtedly have important secular contributions to make to the development of a program of AIDS education. However, when government mandates the participation of religious leaders as a class, the message is clear that what the government is seeking is a specifically religious perspective. That message represents a forbidden endorsement of religion and is inimical to the First Amendment’s goal of "prevent-ting], as far as possible, the intrusion of either [Church or State] into the precincts of the other” (Lemon v Kurtzman, supra, at 614). Accordingly, we dissent.
Chief Judge Wachtler and Judges Simons, Kaye and Hancock, Jr., concur with Judge Bellacosa; Judge Titone dissents and votes to reverse in a separate opinion in which Judge Alexander concurs.
Order affirmed, with costs.
. The majority takes us to task for referring occasionally to religious "leaders” rather than to religious "representatives” — the term used in the regulation (majority opn, at 341). The majority’s point in this regard is, at best, enigmatic. Although it is true that the regulations do not mandate that religious "leaders” be selected for service on the committees, it seems unreasonable to think that individuals who are little known or are not recognized as "leaders” within their religious groups would be chosen as "representatives.” Similarly, it would be naive to suppose that the selected "leaders” or "representatives” appointed to the committees will not at least be perceived as "representing,” and therefore expressing, a religious viewpoint. Indeed, if the appointed representatives were not "leaders” in some sense and were not perceived as representing a uniquely religious perspective on the question of AIDS education, there would be no purpose to mandating the presence of "representatives from religious organizations” as *346a class. Certainly, the purpose the majority identifies, i.e., "neutralizing or reducing possible objections from parents and subsequent opt-outs,” (majority opn, at 340) would not be served, since the achievement of that goal depends to a large extent on the public’s perception that the religious viewpoint has been taken into account.
. The majority’s assertion that religious representatives "are invested with [no] greater role than other included individuals” (majority opn, at 341) misses the point. Our concern here is not that the religious representatives have a greater voice on the committee vis-a-vis "other included individuals,” but rather that religious adherents, unlike all other community members, have guaranteed representation on the committees. While it is true that there is no fixed number of positions on the AIDS advisory committees, it is also true that the number of possible committee seats is necessarily finite and that, consequently, not all of the interest groups in the community will be able to be accommodated. Thus, legitimate interest groups such as planned parenthood organizations, local drug treatment programs and child welfare advocates will have to compete for representation, while, by virtue of the challenged regulations, religious adherents are assured of at least one seat. It is in this sense that the regulations confer a preference on religious adherents as a class and, for that reason alone, they are objectionable.
. Although it has quoted from Bowen v Kendrick (487 US 589), the majority has wisely avoided adopting respondent’s argument that that case is all but dispositive of the issues presented here. In fact, because the statute challenged in Bowen is so different from the statute challenged here, the Supreme Court’s decision in Bowen is of little analytical import in this appeal. The statute at issue in Bowen, which authorized Federal funding for private organizations to provide " 'services and research in the area of premarital adolescent sexual relations,’ ” required grant applicants to describe how they will " 'involve religious and charitable organizations,’ ” as well as families and other community groups (487 US, at 593, 596). Thus, unlike in this case, the statute upheld in Bowen did not mandate that government agencies consult directly with religious representatives. Furthermore, as the Supreme Court pointedly observed, there was "no requirement * * * that grantees be affiliated with any religious denomination” and there was no "suggestion that religious institutions * * * are uniquely well qualified to carry out [the funded] services” (id., at 604-605). It is precisely those characteristics that distinguish the regulations challenged here from the statute at issue in Bowen. And, it is precisely those characteristics that render these regulations constitutionally infirm.