Religion by Any Other Name - Prohibitions on Same-Sex Marriage and the Limits of the Establishment Clause 132 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 RELIGION BY ANY OTHER NAME? PROHIBITIONS ON SAME-SEX MARRIAGE AND THE LIMITS OF THE ESTABLISHMENT CLAUSE* GARY J. SIMSON** Even the most casual observer of the debate in the United States whether to legalize same- sex marriage would be hard-pressed to ignore the significance of religion to that debate. Most obviously, a number of religious groups officially oppose or support same-sex marriage.' The Catholic Church, American Baptist Churches, Southern Baptist Convention, Islam, Orthodox Judaism, Mormon Church, and National Association of Evangelicals are among the groups expressly in opposition, while the United Church of Christ, Reform Judaism, and Unitarian Universalist Association of Congregations are among those expressly in support.2 Moreover, the various groups on record in opposition or support generally have not been bashful about * As is probably apparent, the question posed in the title borrows from Shakespeare. See WILLIAM SHAKESPEARE, ROMEO AND JuLIET, act 2, se. 2, lines 43-44 ("What's in a name? That which we call a rose / By any other name would smell as sweet."). ** Dean and Macon Chair in Law, Mercer University School of Law. For helpful suggestions, I am grateful to: the various participants in the March 2010 symposium at California Western School of Law, where I presented an early version of this piece; my Mercer colleague Jessica Feinbcrg; Jessie Hill, my former colleague at Case Western Reserve; Jenny Ding, article editor for this journal; and Rosalind Simson, philosophy professor at Mercer and, as always, my most rigorous and supportive critic. I See Religious Groups'Official Positions on Same-Sex Marriage, PEW FORUM ON RELIGION AND PUBLIC LIFE, July 27, 2010, http:/pewforum.org/Gay-Marriage-and-Homosexuality/Religious-Groups-Official-Positions-on-Sam-Sx-Marriage. aspx. 2 See id. The Pew Forum lists the Episcopal Church as officially uncommitted but suggests that it is leaning strongly toward support. The Pew Forum also characterizes both Hinduism and Buddhism as having no official position. According to the Pew Forum, some Hindus and Buddhists interpret their religion as being opposed to same-sex marriage, while others interpret it as being in support. The Pew Forum explains that Conservative Judaism does not explicitly endorse or disapprove of same-sex marriage, but instead leaves to individual rabbis the decision whether to perform such marriages in states that permit them. As far as the Presbyterian Church, the Pew Forum notes that (I) the church's General Assembly in 2004 encouraged states to permit same-sex couples to enter into civil unions and (2) although the General Assembly in 2010 reaffirmed the church's position defining marriage as between a man and a woman, it gave no indication that it was any less committed than in 2004 to same-sex couples' having a civil-union option. For insight into the firmness of one religion's official opposition to same-sex marriage, see Erik Eekholm, Southern Baptists Approve Steps Aimed at Diversity, N.Y. TIMES, June 16,2011, atA22 (explaining that, on the one hand, the Southern Baptist Convention, which grew out of"a break with Northern churches over their opposition to slavery" and remained "a defiantly white institution, often promoting segregation" until late in the last century, has begun taking "concrete steps" to reach out to African Americans; on the other hand, the Convention feels obliged not to modify its staunch opposition to homosexuality because, as explained by one of its leaders, "'We cannot compromise without disobeying the Scriptures."').. For an interesting illustration of the ambiguity of the Episcopal view, see Shaila Dewan, True to Episcopal Church s Past, Bishops Split on Gay Weddings, N.Y. TiMEs, July 19, 2011, at A I ("[The Episcopal Church] has not taken a firm position nationally on same-sex marriage, leaving local bishops with wide latitude to decide what priests may do when the law [authorizing same-sex marriage] takes effect in New York State. In the state, with six Episcopal dioceses, the bishops are split..."). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW 133 injecting themselves into the political fray and lobbying forcefully for the outcome that they see as right.3 According to a May 2011 Gallup Poll, a recent surge in public support for same-sex marriage has lifted the proportion of adults in support to slightly more than halfI However, if the adult population is divided into subgroups according to the importance of religion in one's life, some of those subgroups are far from evenly split between supporters and opponents of same-sex marriage. Of those surveyed in 2010 who said that religion is very important to them, only 27% favored legalizing same-sex marriage, and 70% opposed legalization. Conversely, of those surveyed who said that religion is not important to them, 71% favored legalization, and merely 27% were opposed.' Many opponents of legalizing same-sex marriage have made no secret of the fact that their opposition is rooted in religion. In some instances, their emphasis is the sanctity of the favored male-female bond. For example, according to one source, the Bible "maintained that in order to become fully human, male and female must join .... The union of male and female is not merely some lovely ideal; it is the essence of the Jewish and Christian outlooks on the 3 The events leading up to New York's enactment of a statute authorizing same-sex marriage provide a vivid illustration: Early in the week that ended with New York enacting same-sex marriage, the Rev. Anna Taylor Sweringen stood in a hallway just outside the State Senate chambers. She wore her clerical collar and held a sign saying, "Equality now." Around her gathered ministers and rabbis of similar sentiment, all in Albany to lobby and pray for the right of gay couples to wed. As Ms. Taylor Sweringen looked down the corridor, she saw the mirror image of mobilized clergy members, all irreconcilably opposed. One held a placard declaring, "God says no." Then the assemblage broke into a gospel song. "I told Satan to get thee behind," it went. "Victory today is mine." Among her allies, Ms. Taylor Sweringcn responded with a spiritual from the civil rights movement, "I'm Gonna Sit at the Welcome Table." Soon the dueling choirs were lining up along facing walls, barely inches apart, and the state police had to clear a path between them like a boxing referee. Samuel G. Freedman, How Clergy Helped a Same-Sex Marriage Law Pass, N.Y. TIMES, July 16, 2011, at A14. See also Abby Goodnough, Rhode Island Senate Approves Civil Unions After Marriage Measure Falters, N.Y. TIMES, June 30, 2011, at A 16 ("In an interview, [Governor] Chafec said Rhode Island's large elderly and Catholic populations helped explain why same-sex marriage has not gained traction. 'The church has been very active in calling the Legislature,' he said."). 4 See Frank Newport, For First ime, Majority of Americans Favor Legal Gay Marriage, GALLUP, May 20, 2011, http://www.gallup.com/po 147662/first-tim-majority-americans-favor-legal-gay-marriage.aspx (reporting a nine percent one- year increase in public support-from forty-four percent of a random sampling of the U.S. population aged eighteen and older to fifty-three percent of another such sampling). 5 See Jeffrey M. Jones, Americans'Opposition to Gay Marriage Eases Slightly, GALLUP, May 24, 2010, http://www. gallup.com/poll/128291/americans-opposition-gay-marriage-eases-slightly.aspx. 134 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 human experience." 6 In other instances, the argument focuses on the disfavored same-sex relationship. As one staunch critic of same-sex marriage explained, "I believe Christians must submit to the Bible's teachings, and I believe the Bible is unequivocal in its teaching that homosexual behavior is sinful. That being the case, it is impossible for me to accept same-sex marriage, which legitimizes a sinful behavior. ' 7 Often, though by no means always, legislators have been more reticent about citing 6 Defense of Marriage Act: Hearing on H.R. 3396 Before the Subcomm. on the Constitution, 104th Cong. 131 (1996) (prepared statement of Dennis Prager, author and talk show host). To similar effect, see Thomas J. Paprocki, Marriage, Same-Sex Relationships, and the Catholic Church, 38 Loy. U. CHI. L.J. 247, 250-51 (2007): The human person is a combination of body and soul, so men and women are focused on something higher-more than just the advancement of the species. Marriage cannot be just about propagation of the species-it is also about the soul and bringing each partner into a closer union with the other and closer to God .... 7 The Good Book and Gay Marriage, NEWSWEEK, Dec. 15, 2008, http://www.thedailybeast.com / newsweek/2008/12/15/the-good-book-and-gay-marriage.htm (remarks of Dr. Barrett Duke in electronically published e-mail debate). Along similar lines, see Thomas Kaplan, Settled in Albany, Gay Marriage Is Still Drawing Opposition, N.Y TIMEs, July 13, 2011, at A20 (recounting that, three days before New York's law authorizing same- sex marriage would go into effect, the town clerk in a small New York town submitted her letter of resignation, stating that she could not issue marriage licenses to same-sex couples because "there is a higher law than the law of the land.... It is the law of God in the Bible."); Religious Beliefs Underpin Opposition to Homosexuality: Republicans Unified, Democrats Split on Gay Marriage, Part 2: Gay Marriage, PEW FORUM ON RELIGION AND PUBLIc LIFE, Nov. 18, 2003, http://www.people-press.org/2003/l 1/18/part-2-gay-marriage/ ("Asked in an open- ended format their main reason for opposing gay marriage, more than a quarter of opponents (28%) explicitly cite the view that homosexuality is immoral, a sin, or inconsistent with biblical teaching, and another 17% say the idea simply is in conflict with their religious beliefs."). In addition, see Hon. Roderick L. Ireland, In Goodridge Wake: Reflections on the Political, Public, and Personal Repercussions of the Massachusetts Same-Sex Marriage Cases, 85 N.Y.U. L. REV. 1417 (2010), where one of the Massachusetts Supreme Judicial Court justices who, in Goodridge v. Department of Public Health, 798 N.E.2d 941 (Mass. 2003), recognized a state constitutional right to same-sex marriage, commented on his experience after that decision: We (and our decision) were called every name imaginable [including] "atheist liberals" who had "lost our minds" and who were bringing ruin on the United States, especially because "God destroyed Sodom and Gomorrah." ... Of course, some individuals were not satisfied by simply writing letters or engaging in public demonstrations. One group of people came to my church on back-to-back Sundays and disrupted the service.... The group that came sought to have the church rescind my membership and to obtain the congregation's assistance in petitioning to remove me from the bench.... Id. at 1421, 1431. COLUMBIA JOURNAL OF GENDER AND LAW religion to justify their opposition to same-sex marriage.' Some legislators whose opposition to same-sex marriage is rooted strongly in religion may be unwilling to admit it publicly out of a political calculation that citing religion would be apt to be unpopular and cost them votes. Particularly, however, in legislative districts heavily populated by adherents to one or more of the religions on record as opposed to same-sex marriage, political calculation seems an unlikely explanation for legislators' reluctance to cite religion as central to their opposition. A more plausible explanation appears to be constitutional concerns-specifically, concerns about running afoul of the First Amendment's ban on laws "respecting an establishment of religion."9 In this Article, I maintain that, regardless of what lawmakers opposed to same-sex marriage may be willing to state publicly as their reasons for voting against same-sex marriage, courts should find that laws prohibiting same-sex marriage violate the Establishment Clause 8 Interestingly, President Obama has squarely rested his longtime personal opposition to same-sex marriage on religious grounds. See Sheryl Gay Stolberg, Obaina ' Position on Gay Marriage Faces New Test, N.Y. TIMES, June 29, 2011 at Al 5 ("[A]s a Christian, he has said, he views marriage as the union of a man and a woman."). It appears, however, that he has been careful not to suggest that his personal, religiously based opposition would affect any decision that he might be required to make as chief executive. Indeed, he seems intent on keeping separate his personal and official views. Particularly noteworthy in this regard are (i) his decision that the Justice Department should stop defending in court the constitutionality of the portion of the Defense of Marriage Act (DOMA) that defines "marriage" for federal purposes as excluding same-sex marriages and (2) his justification of that decision in terms of(a) the logic under the Equal Protection Clause of subjecting any classification based on sexual orientation to an exacting standard of review and (b) the unconstitutionality of DOMA's definition of "marriage" under such a standard. See Charlie Savage & Sheryl Gay Stolberg, In Turnabout, U.S. Says Marriage Act Blocks Gay Rights, N.Y. TIMES, Feb. 24, 2011, at AI. to similar effect, note the Obama administration's decision a year later not to defend the constitutionality of federal laws denying marriage benefits to same-sex military spouses. See Charlie Savage, Justice Dept. Backs Equal Benefits for Gay Couples in Military, N.Y. TMES, Feb. 18, 2012, at A 16. 9 U.S. CONST. amend. I ("Congress shall make no law respecting an establishment of religion ... . Although the Establishment Clause, by its terms, applies only to federal lawmaking, the Supreme Court held in 1947 that the Due Process Clause of the Fourteenth Amendment made the Establishment Clause applicable to lawmaking by state and local government. Everson v. Bd. of Educ., 330 U.S. 1, 8 (1947). State Constitutions commonly contain one or more provisions similar in basic thrust to the Establishment Clause. See Linda S. Wendtland, Note, Beyond the Establishment Clause: Enforcing Separation of Church and State Through State Constitutional Provisions, 71 VA. L. REV. 625 (1985). Detailed analysis of the impact of those state constitutional provisions on laws prohibiting same-sex marriage is beyond the scope of the present Article. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW and uniformly should be struck down.'0 By way of general background, I begin in Part I by explaining the nonendorsement principle that the Supreme Court has recognized as central to the clause. I turn in Part II to the case law and commentary-or, more precisely, the general scarcity thereof-on the implications of the Establishment Clause for same-sex marriage prohibitions. After explaining in Part III the basic framework for the Establishment Clause analysis that follows, I examine in Part IV the various reasons that opponents of same-sex marriage have articulated with any frequency in support of a ban. I maintain that those reasons provide strikingly little justification for laws banning same-sex marriage. After discussing in Part V three substantial reasons for allowing same-sex couples to wed, I consider in Part VI the difficulty of making sense of a legislative decision to prohibit same-sex marriage-a decision that, in effect, gives priority to the flimsy reasons for prohibition over the weighty reasons for legalization. A covert purpose to endorse religion is not the only possible explanation for the legislature's decision. I suggest, however, that it is the most plausible one largely because, perhaps paradoxically, it is the one most in keeping with the deference that lawmakers are generally owed. In Part VII, I examine three possible objections to a conclusion that prohibitions on same- sex marriage violate the Establishment Clause. The first two objections that I address relate to ambiguities in the endorsement test. The third objection arises out of the possibility that, in the near future, the Supreme Court may replace the endorsement test with a less demanding test predicated on coercion. I consider both the likelihood of such a change in the applicable test and whether prohibitions on same-sex marriage are more apt to survive a coercion test than an endorsement test. Lastly, in Part VIII, I briefly examine the federal government's contribution 10 As of March 1, 2012, when Maryland Governor Martin O'Malley signed a bill authorizing same-sex marriage, eight states (Connecticut, Iowa, Maryland, Massachusetts, New Hampshire, New York, Vermont, and Washington) and the District of Columbia had legalized same-sex marriage by statute or judicial decision. Maryland Governor Signs Same-Sex Marriage Bill, CNN, Mar. 1, 2012, http://articles.cnn.com/2012-03- 01 /us/usmaryland-same-sex-marriage I hawaii-supreme-court-marriage-law-marriage-bill? s=PM:US. In addition, in a decision that may be reviewed en banc or by the U.S. Supreme Court, a federal appellate court struck down on federal equal protection grounds California's voter-adopted initiative amending the state constitution to ban same-sex marriage. Perry v. Brown, Nos. 10-16696 & 11-16577, 2012 WL 372713 (9th Cir. Feb. 7, 2012). Nonetheless, the great majority of states continue to prohibit same-sex marriage. For a useful compilation several years ago of the texts of relevant state provisions, see Andrew Koppelman, Interstate Recognition of Same-Sex Marriages and Civil Unions: A Handbook for Judges, 153 U. PA. L. REV. 2143, 2165-94 (2005). (Note, however, that the compilation precedes in time, and therefore does not reflect, the legalization of same-sex marriage in seven of the eight states-all but Massachusetts-that have legalized it). 136 23.1 COLUMBIA JOURNAL OF GENDER AND LAW to the same-sex marriage debate: the Defense of Marriage Act of 1996 (DOMA)." As I explain, despite some obvious differences, DOMA is very much akin to state laws prohibiting same- sex marriage. For much the same reasons that those laws are vulnerable to Establishment Clause challenge, DOMA is as well. I. The Nonendorsement Principle For lawmakers to invoke religion as the basis for adopting or retaining prohibitions on same-sex marriage treads on constitutionally perilous ground-at least as that terrain has been described by the Supreme Court over the past several decades. If, as Supreme Court precedent holds, the Establishment Clause prohibits government endorsement of one or another religion or religious belief, 2 lawmakers risk judicial invalidation by articulating their opposition to same-sex marriage in religious terms. Under the Supreme Court's longstanding "endorsement test," a law may be struck down on the basis of either a purpose or effect of endorsing religion. 3 The likelihood of judicial invalidation based on a purpose of endorsing religion is low because the Supreme Court has limited invalidation on that basis to instances in which the challenger can prove that a purpose of endorsement was the lawmakers' exclusive, or nearly exclusive, purpose in enacting the II As codified, DOMA appears at I U.S.C. § 7 (2006) and 28 U.S.C. § 1738C (2006), which are quoted infra notes 234 & 235. 12 See, e.g., Allegheny Cnty. v. ACLU, 492 U.S. 573, 592-94, 599-600 (1989); Wallace v. Jaffree, 472 U.S. 38, 56 & n.42 (1985). 13 Although the Court in the cases from the 1980s cited supra note 12 described the contours of the endorsement test more clearly than the Court had done before, the Court had already been treating a purpose or effect of endorsing religion as constitutionally problematic for more than two decades. See, e.g., Abington Sch. Dist. v. Schempp, 374 U.S. 203 (1963) (applying a two-part purpose-and-effect test, explicitly disclaiming that proof of coercion is needed to show an Establishment Clause violation, and invalidating the public school prayer exercises under review); see also Allegheny Cnty., 492 U.S. at 592 ("whether the challenged governmental practice either has the purpose or effect of 'endorsing' religion [is] a concern that has long had a place in our Establishment Clause jurisprudence"). 23.1 138 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 law.14 Given the difficulties inherent in trying to prove that lawmakers acted for purposes other than those that they are willing to admit, the task of proving that they were motivated entirely or almost entirely by an impermissible purpose-in this instance, one of endorsing religion-is almost always extremely difficult, even when there is very good reason to suspect that impermissible purpose played a crucial part.'5 The likelihood of judicial invalidation based on an effect of endorsing religion is considerably less remote. Under this aspect of the endorsement test, the challenger must persuade the court that a reasonable observer is likely to perceive the law in question as sending a message of government endorsement of religion. 6 The test does not require for invalidation that the impermissible effect of endorsement be exclusive or even nearly so. Instead, it is met by a showing that the impermissible effect is substantial. 7 Though not entirely clear, the requirement of a "substantial" effect of endorsing religion is probably best understood as focusing attention on whether a reasonable observer is apt to perceive a law as communicating 14 My characterization of the Court's test in terms of an "exclusive or nearly exclusive" purpose of endorsement reflects an ambiguity in the case law. See, e.g., McCreary Cnty. v. ACLU, 545 U.S. 844, 862 (2005) (requiring proof of a "predominantly religious purpose" to prevail on impermissible purpose grounds); Bowen v. Kendrick, 487 U.S. 589, 602 (1988) (requiring proof that the law is "motivated wholly by an impermissible purpose"). It is arguable-and, indeed, I have argued it, see Gary J. Simson, The Establishment Clause in the Supreme Court: Rethinking the Courts Approach, 72 CORNELL L. REv. 905, 908-11 (1987)--that the Court is wrong to require a challenger to make such a strong showing of a purpose of endorsing religion in order to secure invalidation of a law. If, as I believe, the Court's approach to impermissible purpose in the equal protection area, see, e.g., Arlington Heights v. Metro. Housing Dev. Corp., 429 U.S. 252 (1977), is sound, a law should be struck down on the basis of a purpose of endorsing religion if the law would not have been adopted but for that purpose. Whether such purpose was exclusive, nearly exclusive, or even primary should be seen as beside the point. For now, however, I need not press such objections, and for the remainder of this Article, I accept as is the Court's approach to invalidation on the basis of a purpose of endorsement. 15 See Paul Brest, Palmer v. Thompson: An Approach to the Problem of Unconstitutional Legislative Motive, 1971 SuP. CT. REv. 95, 119-24; Simson, supra note 14, at 922-23. 16 See, e.g., Lamb's Chapel v. Center Moriches Sch. Dist., 508 U.S. 384, 395 (1993); Allegheny Cnty., 492 U.S. at 592-94, 599-600; Gary J. Simson, Endangering Religious Liberty, 84 CALIF. L. REv. 441,469-70 (1996). 17 The three-prong (purpose-effect-entanglement) Establishment Clause test of Lemon v. Kurtzman, 403 U.S. 602, 612-13 (1971), which essentially includes the endorsement test as a subset of its first two prongs, see Simson, supra note 16, at 462 & n.65, would appear to condition invalidation for impermissible effect on a showing of "primary" effect. However, soon after Lemon, the Court expressly denied that its use of "primary" in the Lemon test should be taken literally. See Comm. for Pub. Educ. v. Nyquist, 413 U.S. 756, 783-84 n.39 (1973). Rather than require, as the Lemon test's use of "primary" would seem to demand, a determination whether an impermissible effect is a law's principal or predominant effect, the Court indicated that the presence ofa "substantial" impermissible effect is the relevant concern. See id. COLUMBIA JOURNAL OF GENDER AND LAW a substantial preference on the part of the state that people adhere to some religion (rather than none at all), a particular religion, or a particular religious belief.8 A reasonable observer's perception of the message communicated by government action only sensibly would turn largely on his or her perception of the purposes underlying the action. 9 If a reasonable observer were persuaded that the state acted without any purpose of endorsing religion, he or shewould not be apt to perceive the action as communicating a substantial pro-religion preference on the part of the state. Conversely, if a reasonable observer were persuaded that the state's exclusive purpose in acting was to endorse religion, he or she would readily perceive the action as communicating a substantial preference of this sort. A court's task in ascertaining a reasonable observer's likely perception when a law appears to be based on a variety of purposes, some religious and some not, is obviously more complex and requires a close assessment of the apparent purposes and the part that each appeared to play in the state's decision to act as it did. Because proof of a purpose of endorsing religion may be instrumental in proving an impermissible effect even if that purpose is not exclusive or nearly exclusive, lawmakers need to be considerably more secretive about their religious purposes than they would need to be if the only relevant inquiry under the endorsement test were the one into purpose. As I seek to demonstrate in this Article, however, secretiveness is no guarantee of success. II. The (Very Limited) Judicial and Scholarly Backdrop In light of the prominent role that religion has played in the public debate on same-sex marriage, one might expect that there would be a fair amount of case law and scholarly commentary addressing the implications of the Establishment Clause for same-sex marriage prohibitions. In fact, however, the case law is almost silent on those implications, and the little commentary in point does not explore them in depth. An unreported decision by a District of Columbia trial court, Dean v. District of Columbia,2° 18 See Simson, supra note 14, at 917-23. 19 For an illustration of this point in the context of abstinence-only sex education programs, see Gary J. Simson & Erika A. Sussman, Keeping the Sex in Sex Education: The First Amendment s Religion Clauses and the Sex Education Debate, 9 S. CAL. REv. L. & WOMEN'S STUD. 265, 294-95 (2000). 20 Civ. A. No. 90-13892, 1992 WL 685364 (D.C. Super. Ct. June 2, 1992), aff'd, 653 A.2d 307 (D.C. App. 1995). 23.1 140 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 appears to offer the only even moderately extensive judicial discussion of the issue. 21 The plaintiffs, Craig Dean and Patrick Gill, challenged on due process, equal protection, and Establishment Clause grounds a D.C. statute that, as interpreted by the trial court in an earlier order,22 restricted marriage to only opposite-sex couples. The court rejected the challenge on all three grounds. The plaintiffs' Establishment Clause argument centered on (1) the trial court's statement in issuing the prior order that its interpretation of "marriage" as used in the D.C. statute would turn on a determination of which unions "historically" were "encompassed within the common understanding and usage of the term" 23 and (2) the court's quotations from, and reliance upon, the Bible in making its historical determination. 24 The Establishment Clause challenge actually alleged two independent constitutional violations. The plaintiffs claimed that the trial court's reliance on the Bible in rendering its interpretation of the statute violated the clause. In addition, they claimed that if, as the court's opinion could be read to suggest, the D.C. Council in enacting the statute was proceeding on a Biblical understanding of the term "marriage," then the Council violated the clause as well. The court began its Establishment Clause analysis by reciting the three-part test that the Supreme Court had announced as broadly applicable to Establishment Clause problems in 1971 in Lemon v. Kurtzman.25 That test, which from the outset has included as a subset and central component the endorsement test discussed in Part I and which in the past two 21 The trial court's opinion in Dean is highly distinctive in even mentioning the Establishment Clause in its discussion of the constitutionality of the D.C. same-sex marriage prohibition. For leading decisions on the constitutionality of prohibiting same-sex marriage that, presumably reflecting the litigants' failure to raise an Establishment Clause challenge, make no mention of possible Establishment Clause concerns, see, for example, Perry v. Brown, Nos. 10-16696 & 11-16577, 2012 WL 372713, at *38 (9th Cir. Feb. 7, 2012) (invalidating California's Proposition 8, a voter-enacted state constitutional amendment to prohibit same-sex marriage, on federal constitutional equal protection grounds); Vamrnum v. O'Brien, 763 N.W.2d 862,904-06 (Iowa 2009) (invalidating on state constitutional equal protection grounds state's ban on same-sex marriage, and discussing religious opposition to same-sex marriage but solely for purpose of rejecting possible argument "left unspoken" by state that such opposition gives rise to state interest sufficient to justify unequal treatment); Lewis v. Harris, 908 A.2d 196, 220-21 (N.J. 2006) (holding that state constitution's equal protection guarantee requires state to dispense benefits and privileges equally to committed same-sex and married opposite-sex couples); Goodridge v. Dep't of Pub. Health, 798 N.E.2d 941, 968 (Mass. 2003) (invalidating state's prohibition on same-sex marriage under state constitution's individual liberty and equality guarantees). 22 See Dean at * (describing order of Dec. 30, 1991). 23 See id. at *4 n. 19 (quoting the court's opinion in issuing the prior order). 24 See id. (describing, and quoting from, the court's earlier opinion). 25 403 U.S. 602, 612-13 (1971). COLUMBIA JOURNAL OF GENDER AND LAW decades has become far less important to the Court's Establishment Clause approach than the endorsement test,26 lays down three requirements for a law to survive Establishment Clause review: the law must have a secular purpose; its primary effect must neither advance nor inhibit religion; and it must not give rise to excessive entanglement between government and religion." Surprisingly, rather than apply the Lemon test, the court turned its attention to the metaphor of a "wall of separation between church and state" that the Supreme Court has invoked from time to time in describing the purpose of the Establishment Clause.28 Moreover, the court found in that metaphor dispositive proof that the Establishment Clause had not been breached. "Whatever may be the exact contours of the separation," the court explained, "it is inconceivable that the 'wall' is so impregnable as to preclude judicial or legislative resort to the Bible merely as a historical reference. 29 After underlining that it had already said all that needed to be said,30 the court added for 26 See supra note 17. Although the Supreme Court has never said that it no longer continues to abide by the Lemon test, it has made clear on various instances that it does not feel strongly constrained by the test. See, e.g., Zelman v. Simmons-Harris, 536 U.S. 639 (2002) (upholding Ohio's school voucher statute and discussed in detail in Gary J. Simson, School Vouchers and the Constitution-Permissible, Impermissible, or Required?, I CORNELL J. L. & PUB. POL'Y 553, 566-75 (2002)). In addition, several of the Justices-most notably Justice Scalia-have been vocal in calling for its official demise. See, e.g., Lamb's Chapel v. Ctr. Moriches Union Free Sch. Dist., 508 U.S. 384, 398 (1993) (Scalia, J., concurring in judgment) (likening the Lemon test to "some ghoul in a late-night horror movie that repeatedly sits up in its grave and shuffles abroad, after being repeatedly killed and buried"). In contrast, as the Lemon test waned in force, the endorsement test retained its vitality, continuing to command five firm votes on the Court. See, e.g., McCreary Cnty. v. ACLU, 545 U.S. 844 (2005); Capitol Square Review & Advisory Bd. v. Pinette, 515 U.S. 753 (1995); Allegheny Cnty. v. ACLU, 492 U.S. 573 (1989). As discussed infra text accompanying notes 199-206, it is unclear whether a majority of the Court still subscribes to the endorsement test. 27 As a practical matter, the Court in Agostini v. Felton, 521 U.S. 203 (1997), eliminated the entanglement prong of the test as having any independent importance. After discussing the similarity in the factors considered under the effect and entanglement prongs, the Court essentially redefined the entanglement prong out of existence. "[l]t is simplest," the Court explained, "to recognize why entanglement is significant and treat it... as an aspect of the inquiry into a statute's effect." Id. at 233. 28 Dean at *5 (quoting passages using that metaphor in Everson v. Bd. of Educ., 330 U.S. 1, 16 (1947); McCollum v. Bd. of Educ., 333 U.S. 203, 213 (1948) (Frankfurter, J., concurring); and Lynch v. Donnelly, 465 U.S. 668, 673 (1984)). 29 Dean at *5. 30 See id. 23.1 142 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 good measure--"if further illumination be deemed warranted"31'-a "further explicat[ion]"32 of its views on the Establishment Clause. The court returned at this point to the test it had recited at the start of its Establishment Clause analysis. it launched into an extended discussion of what constitutes a purpose of advancing or endorsing religion and found that a prohibition on same-sex marriage could not be struck down for failure to meet Lemon's secular purpose requirement. Without attempting to catalog the various doctrinal and analytical flaws in the Dean court's "further illumination" of its views,33 suffice it to say that the court became so preoccupied with the purpose prong that it never even got to the effect one. As I discuss below, prohibitions on same-sex marriage are far more seriously threatened by the Supreme Court's approach under the Establishment Clause to effect than its approach to purpose. In essence, the Dean court rejected the claimants' Establishment Clause challenge without reaching the issue that gave that challenge genuine force.34 Only three articles appear to offer more than incidental analysis of a possible Establishment Clause challenge to laws prohibiting same-sex marriage. The earliest of the three, DOMA: An Unconstitutional Establishment of Fundamentalist Christianity by James M. Donovan,35 focused on the federal Defense of Marriage Act (DOMA). DOMA defines "marriage" for 31 Id. 32 Id. 33 Among the more glaring of those flaws is the court's drawing its understanding of permissible and impermissible purposes under the Lemon and endorsement tests from a dissenting opinion by the member of the Court, Justice Scalia, probably most hostile to both of those tests. See id. at *5. Another serious shortcoming is the court's treating as a core principle of Establishment Clause doctrine a sentence in one of the Supreme Court's earliest Establishment Clause opinions--"We are a religious people whose institutions presuppose a Supreme Being" (Zorach v. Clauson, 343 U.S. 306, 313 (1952))-that sits uneasily, to say the least, with the thinking behind the school prayer cases of the early 1960s and various other cases that came to comprise the core of Establishment Clause precedent. See Dean at *7. Though decided in 1952, Zorach came early in the development of Establishment Clause doctrine because the Supreme Court and lower courts had little occasion to interpret the clause prior to the Court's holding in Everson v. Board of Education, 330 U.S. 1, 8 (1947), that the Fourteenth Amendment's Due Process Clause made the Establishment Clause applicable to state and local government. For critical analysis of Zorach, see Simson, supra note 14, at 927-28. 34 Although the unsuccessful plaintiffs filed an appeal of the trial court's decision, they did not raise on appeal their Establishment Clause challenge below. See Dean v. District of Columbia, 653 A.2d 307, 310 n. I (D.C. App. 1995) (Feren, J., concurring in part and dissenting in part). As a result, although the appellate court's opinion, unlike the trial court's, was reported, there is no reported opinion in the case addressing the Establishment Clause challenge. 35 4 MICH. J. GENDER & L. 335 (1997). COLUMBIA JOURNAL OF GENDER AND LAW federal purposes as exclusively heterosexual.36 It also purports to negate any state full-faith- and-credit obligation to recognize a same-sex marriage lawfully performed out of state.37 DOMA is not a prohibition, as such, on same-sex marriage. However, as an expression of governmental antipathy to same-sex marriage, DOMA has much in common with same-sex marriage prohibitions and invites similar Establishment Clause objections. According to Mr. Donovan, DOMA violates the Establishment Clause for lack of any genuine secular purpose. As I demonstrate in this Article, the Establishment Clause question posed by prohibitions on same-sex marriage (or by a measure like DOMA that strongly disfavors same-sex marriage) is substantially more complex than Mr. Donovan assumes. In addition, a determination of unconstitutionality cannot fairly be premised on an analysis of purpose alone. The most recent of the three articles is a student note by Ben Schuman, Gods & Gays: Analyzing the Same-Sex Marriage Debate from a Religious Perspective.3" Mr. Schuman reaches a very different conclusion than Mr. Donovan. Interestingly, analyzing what he regards as the principal secular justifications that opponents of same-sex marriage have offered for banning same-sex marriage, Mr. Schuman takes a view of those justifications similar to Mr. Donovan's. According to Mr. Schuman, they are devoid of content, no more than "pretexts"3 9 for the religious reasons that actually motivated the lawmakers. Even more interestingly, in a brief section devoted to applying his analysis of religious and secular arguments to the Supreme Court's Establishment Clause tests, 40 Mr. Schuman concludes that laws prohibiting same-sex marriage should survive not only an inquiry into impermissible purpose but also one into impermissible effect. In part, Mr. Schuman errs by relying heavily on a lower court precedent-the Dean case discussed above41-that is not deserving of nearly as much weight as he assigns it. In addition, the secular purposes invoked on behalf of same-sex marriage cannot be discounted as readily as Mr. Schuman assumes, and the perceptions of government endorsement of religion that they invite are far more constitutionally problematic than he recognizes. The final article in this group of three is one of my own: Beyond Interstate Recognition in 36 1 U.S.C. § 7 (2006). 37 28 U.S.C. § 1738C (2006). 38 96 GEO. L.J. 2103 (2008). 39 Id. at 2113. 40 Id. at 2129-34. 41 See supra text accompanying notes 20-34. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW the Same-Sex Marriage Debate, published in 2006.42 The bulk of that article is addressed to an interstate recognition question that courts and scholars have treated as a question of choice of law: if a same-sex couple from a state that prohibits same-sex marriage (State A) gets married in a state that allows same-sex marriage (State B) and then returns to live in State A, should a State A court decide the validity of the couple's marriage by applying the law of State A or State B? After arguing in that article that, under any creditable choice-of-law approach, the State A court should apply the law of State A, I maintained that the recognition question under discussion could not properly be decided simply as a matter of choice of law. As I explained, that question implicates more than choice of law because laws prohibiting same-sex marriage may reasonably be found to overstep federal constitutional bounds-specifically, the bounds set by the Due Process, Equal Protection, and Establishment Clauses. Out ofregard for the article's overall scope and balance, I kept my analysis of Establishment Clause problems relatively brief, and limited myself to outlining its principal components.43 I recognized that to be fully persuasive, the analysis would need to be amplified substantially both by fleshing out a number of points already made and by addressing various possible objections that fairly could be raised. I decided, however, that a more comprehensive analysis was best left to a later time. Several years later, prompted by an invitation to contribute in person and in print to a symposium on same-sex marriage, I came to see that my Establishment Clause analysis in the 2006 article was even less complete than I recognized at the time.44 Indeed, although I continue to believe that same-sex marriage prohibitions should .be struck down under the Establishment Clause, I also believe that my earlier analysis is sufficiently vulnerable to attack because of incompleteness as to invite skepticism of the conclusion that it urged. For that reason, along with the inherent importance and richness of the issue, I seek to provide now the 42 40 U.C. DAVIS L. REv. 313 (2006). 43 Id. at 375-82. 44 See Gary J. Simson, Religion, Same-Sex Marriage, and the Defense of Marriage Act, 41 CAL. W. INT'L L.J. 35 (2011). .23.1 COLUMBIA JOURNAL OF GENDER AND LAW comprehensive and nuanced analysis that the issue deserves.45 To the extent that litigants and scholars have raised federal constitutional objections to prohibitions on same-sex marriage, they have done so almost exclusively in terms of equal protection and due process.46 In part, the inattention to Establishment Clause constraints may simply reflect the legal community's much greater conversance in general with equal protection and due process doctrine than with Establishment Clause doctrine. The proportion of litigators reasonably well-versed in basic principles of equal protection and due process law may not be enormous, but it surely dwarfs the proportion of those who can fairly claim a solid understanding of Establishment Clause law. Among constitutional law scholars, virtually all have a firm grasp of key equal protection and due process concepts-concepts that form a major part of almost any basic Constitutional Law course. A substantially smaller group, however, is conversant with Establishment Clause principles, which rarely find their way into the basic course and instead typically are the focus of an upperclass seminar or a minor portion of an upperclass First Amendment elective. 45 Beyond the three law review pieces discussed in the above text, legal scholarship largely ignores the implications of the Establishment Clause for prohibitions on same-sex marriage. To the extent that those implications are addressed at all, it is only in passing. See, e.g., Larry Cata Backer, Religion as the Language of Discourse of Same Sex Marriage, 30 CAP. U. L. REV. 221, 245-48 (2002) (acknowledging that the "dangers of the identity of secular and religious systems of organization are great" in today's society and stating with apparent concern that "law is being bent to the will of religious communities," but exploring possible Establishment Clause problems no further and instead urging proponents of same-sex marriage to learn from their opponents' tactics and invoke for their cause "the language of religion within the institutional frameworks of religion"); David B. Cruz, Disestablishing Sex and Gender, 90 CALIF. L. REv. 997, 1078 (2002) (maintaining that "[c]ivil marriage is one of the last great bastions of resistance to the disestablishment of religion in the United States," and suggesting that negativity toward same-sex marriage is largely "grounded in people's religious beliefs that marriage means, simply must be, and was instituted by God as, a union of one man with one woman," but pursuing that train of thought no further than as an explanation for same-sex marriage opponents' tendency to try to justify their opposition in secular terms); Josh Friedes, Can Same-Sex Marriages Coexist with Religion?, 38 NEW ENG. L. REv. 533, 534 (2004) (characterizing as "the paradox of the establishment clause" how that clause applies to prohibitions on same-sex marriage in light of the fact that "the universal norm from religion [of limiting marriage to the union of one man and one woman], which essentially was adapted into the civil law of states, is no longer in fact a universal norm, but continues to be applied by the state without question," but not attempting to resolve the paradox). 46 See, e.g., Perry v. Schwarzenegger, 704 F. Supp. 2d 921,929-30 (N.D. Cal. 2010), aff'd sub norm. Perry v. Brown, Nos. 10-16696 & 11 -16577, 2012 WL 372713 (9th Cir. Feb. 7, 2012) (summarizing plaintiffs' federal constitutional arguments-ones predicated on the Due Process and Equal Protection Clauses-for challenging California's Proposition 8); ANDREW KOPPELMAN, THE GAY RIGHTS QUESTION IN CONTEMPORARY AMERICAN LAW (2002); MARK STRASSER, LEGALLY WED: SAME-SEX MARRIAGE AND THE CONSTITUTION (1997); Nelson Tebbe & Deborah A. Widiss, Equal Access and the Right to Many, 158 U. PA. L. REv. 1375 (2010); Charlie Savage, Suits on Same Sex Marriage May Force Administration to Take a Stand, N.Y. TIMES, Jan. 29, 2011, at A 14. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW The scarcity of case law and commentary on the implications of the Establishment Clause for laws banning same-sex marriage also may reflect a sense that an Establishment Clause challenge to same-sex marriage bans faces a more complex set of hurdles than an equal protection or due process challenge. In and of itself, the fact that an Establishment Clause challenge, unlike an equal protection or due process one, calls for inquiry into lawmakers' actual purposes adds a layer of complexity that those other challenges lack. Ultimately, an Establishment Clause challenge may be no less potent than any that can be made under the Equal Protection or Due Process Clause. Indeed, it may be even more potent. That does not deny the possibility, however, that the greater complexity of an Establishment Clause challenge may have deterred more than a few from attempting to make it. Lastly, perhaps some litigators and scholars have avoided attacking same-sex marriage bans on Establishment Clause grounds because they have taken too much to heart the chastening experience of those who invoked the Establishment Clause as a basis for invalidating laws prohibiting abortion. After all, no less a constitutional law luminary than Harvard Professor Laurence Tribe felt obliged to recant in print his argument in 1973 in a Harvard Law Review "Foreword" that the Establishment Clause is fundamentally at odds with laws prohibiting abortion prior to fetal viability.47 Perhaps some who have come later have overcompensated by refraining, even when unwarranted, from sounding an Establishment Clause alarm with regard to other types of laws that plainly owe their existence at least in part to religious purposes but that do not explicitly take religion into account. In urging attention to the Establishment Clause, I emphasize that I in no way mean to suggest that the arguments for invalidating prohibitions on same-sex marriage under the Equal Protection and Due Process Clauses are anything less than sound. Indeed, in an earlier article, I spelled out in some detail the logic under each of those clauses of striking down such prohibitions. 48 However, unless and until the Supreme Court holds, or the lower courts are unanimously agreed, that laws banning same-sex marriage are constitutionally deficient, it is important to make clear the multiple ways in which such laws overstep constitutional bounds. Broadly speaking, the articulation of multiple substantial grounds for invalidation enhances the likelihood that the courts will strike down same-sex marriage prohibitions and that they 47 Compare Laurence H. Tribe, The Supreme Court, 1972 Term - Foreword: Toward a Model of Roles in the Due Process of Life and Law, 87 HARV. L. REV. 1, 18-32 (1973) (relying on the Establishment Clause to articulate an alternative defense of the Court's due process decision in Roe v. Wade, 410 U.S. 133 (1973)), with LAURENCE H. TRIBE, AMERICAN CONSTITUTIONAL LAW § 15-10, at 1350 (2d ed. 1988) (repudiating "on reflection" his earlier view for three reasons). As Professor Tribe notes, see id. at 1350 n.88, the recantation in the 1988 edition of his treatise reiterates one that he made in 1978 in the first edition of the book. 48 Simson, supra note 42, at 365-68 (due process), 368-74 (equal protection). 23.1 COLUMBIA JOURNAL OF GENDER AND LAw will do so sooner rather than later. It allows, for example, for the possibility that one judge may be persuaded by one argument for unconstitutionality, while others may be convinced by another. Less obviously, it creates momentum for, and expedites, judicial invalidation under one or another ground by highlighting the incompatibility of laws prohibiting same-sex marriage with the basic fabric of federal constitutional law. 49 Il. The Analytical Framework My ultimate conclusion that laws prohibiting same-sex marriage violate the Establishment Clause rests on a detailed analysis of the purposes and effects of such laws. My basic thesis, however, is quite simple. Laws prohibiting same-sex marriage are extremely difficult to understand in secular terms and extremely easy to understand in religious terms. As a result, a reasonable observer is likely to perceive-indeed, highly likely to perceive-that the state in banning same-sex marriage is sending a message endorsing a particular religious view. Before embarking on an analysis of purposes and effects, I note a few assumptions that I make throughout in order to enhance the significance of my ultimate conclusion that prohibitions on same-sex marriage violate the Establishment Clause. Each assumption makes proof of a constitutional violation more difficult than it would be if the contrary assumption were made. First, I assume that, in enacting the same-sex marriage prohibition, the lawmakers avoided making statements that rather clearly concede that their purpose was to endorse a religious belief that same-sex marriage should be banned. Concededly, lawmakers have made statements of that sort, and when they do so, they ease a challenger's burden of proof in seeking invalidation of the same-sex marriage ban. However, by assuming that lawmakers have not made such statements, I frame my analysis in a way that promises to broaden the applicability of my ultimate conclusion of unconstitutionality. Second, although some prohibitions on same-sex marriage have been adopted by popular referendum, I assume that the prohibition under discussion is the product of the legislative process. A reasonable observer is at least arguably more apt to perceive that a lawmaker acted for religious purposes if the relevant lawmaker is the voting public rather than the people's 49 See CHARLES L. BLACK, JR., STRUCTURE AND RELATIONSHIP IN CONSTITUTIONAL LAW 7 (1969) (urging the importance in constitutional interpretation not only of the commonly invoked "method of purported explication or exegesis of the particular textual passage considered as a directive of action" but also of the commonly neglected "method of inference from the structures and relationships created by the Constitution in all its parts or in some principal part"). For an interesting perspective on the same-sex marriage debate as a vehicle for understanding the viability of a leading theoretical approach to constitutional law, see Jane S. Schacter, Ely at the Altar: Political Process Theory Through the Lens of the Marriage Debate, 109 MICH. L. REv. 1363 (2011). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW elected representatives. Because the Oath-of-Office Clause of the Constitution" requires all legislators, state and federal, to take an oath to obey the Constitution, and because legislators are generally more conversant than the average person with federal constitutional bounds, legislators may reasonably be perceived as less apt in lawmaking than the general electorate to overstep Establishment Clause bounds. By using legislation rather than popular referendum as the prototype of a same-sex marriage prohibition, I therefore again frame my analysis in a way that helps maximize the significance of my ultimate conclusion of unconstitutionality. Third and lastly, I assume that the purposes on which a reasonable observer would focus in thinking about a same-sex marriage prohibition are the purposes that explain its continued existence. Under this view, purposes that played a part in the adoption of a prohibition are beside the point if they fail to explain why the prohibition remains on the books."' Because some prohibitions derive from states' early marriage laws, and because those early laws commonly were strongly rooted in religion, a focus on purposes at the time of adoption, rather than the purposes behind current retention, would militate in favor of a finding of unconstitutionality. A focus on purposes behind current retention over purposes behind adoption therefore is more in keeping with my desire in the analysis at hand to make an ultimate conclusion of unconstitutionality more broadly applicable. IV. An Assessment of Reasons Offered for Prohibiting Same-Sex Marriage In seeking to identify the components of a legislative decision to prohibit same-sex marriage, it is helpful to begin with two possible reasons for prohibition that the Supreme Court has placed off limits. After discussing those two impermissible purposes in section A, I consider in sections B through G the various permissible reasons that opponents of same-sex marriage have offered with some frequency to justify prohibitions. I conclude that, though perhaps impressive in number, those reasons do not provide, even in the aggregate, significant justification for prohibiting same-sex marriage. 50 U.S. CONST. art. VI, cl. 3 ("The Senators and Representatives before mentioned, and the Members of the several State Legislatures, and all executive and judicial Officers, both of the United States and of the several States, shall be bound by Oath or Affirmation, to support this Constitution."). 51 As discussed infra note 193 and accompanying text, I believe that any constitutional analysis in which purpose is a central concern most logically gives priority to purposes that explain a law's continued existence over purposes that prompted a law's adoption. 23.1 23.1 COLUMBIA JOURNAL OF GENDER AND LAW 149 A. The Simple Lessons of Romer and Lawrence Romer v. Evans12 in 1996 and Lawrence v. Texas53 in 2003 easily qualify as two of the Supreme Court's most controversial decisions of the past two decades.54 In both instances, a strongly worded majority opinion by Justice Kennedy provided the occasion for a vehement dissent by Justice Scalia on behalf of himself, Justice Thomas, and Chief Justice Rehnquist. In addition, both decisions fueled substantial debate outside the Court.5" The Court in Romer struck down under the Equal Protection Clause a Colorado constitutional amendment that, in the Court's words, had the practical effect of "prohibit[ing] all legislative, executive or judicial action at any level of state or local government designed to protect.., gays and lesbians."56 According to the Court, the Colorado measure placed "a broad and undifferentiated disability"57 on gays and lesbians that was "inexplicable by anything but animus toward the class that it affects."58 As the Court further explained, a law rooted in animus singles out a group of persons for disadvantage "not to further a proper legislative end 52 517 U.S. 620 (1996). 53 539 U.S. 558 (2003). 54 Professor Klarman has suggested that, as a result of the growing popular support for gay and lesbian rights, Lawrence, like Brown v. Board of Education, 347 U.S. 483 (1954), before it, is apt to evolve from a lightning rod for disagreement to a bedrock decision of the sort that no Supreme Court nominee serious about getting confirmed would dare to question. See Michael J. Klarman, Brown and Lawrence (and Goodridge), 104 MICH. L. REV. 431 (2005). 55 For a variety of scholarly views on Romer, see Akhil Reed Amar, Attainder andnAmendment 2: Romer Rightness, 95 MICH. L. REV. 203 (1996); Lynn A. Baker, The Missing Pages of the Majority Opinion in Romer v. Evans, 68 U. CoLo. L. REV. 387 (1997); Richard F. Duncan, Wigstock and the Kulturkampf" Supreme Court Storytelling, the Culture War and Romer v. Evans, 72 NOTRE DAME L. REv. 345 (1997); Daniel Farber & Suzanna Sherry, The Pariah Principle, 13 CONST. COMMENT. 257 (1996); Roderick M. Hills, Jr., Is Amendment 2 Really a Bill of Attainder? Some Questions About Professor Amar s Analysis of Romer, 95 MICH. L. REv. 236 (1996). For a range of perspectives on Lawrence, see Randy E. Barnett, Justice Kennedy ' Libertarian Revolution: Lawrence v. Texas, 2002-2003 CATO SuP. CT. REv. 21 (2003); Lino A. Graglia, Lawrence v. Texas: Our Philosopher-Kings Adopt Libertarianism as Our Official National Philosophy and Reject Traditional Morality as a Basis for Law, 65 Ono ST. L.J. 1139 (2004); Pamela S. Karlan, Foreword: Loving Lawrence, 102 MICH. L. REv. 1447 (2004); Cass R. Sunstein, What Did Lawrence Hold? OfAutonomy, Desuetude, Sexuality, and Marriage, 2003 Sup. CT. REv. 27; Laurence H. Tribe, Lawrence v. Texas: The "Fundamental Right" That Dare Not Speak Its Name, 117 HARV. L. REv. 1893 (2004). 56 517 U.S. at 624. 57 Id. at 632. 58 Id. COLUMBIA JOURNAL OF GENDER AND LAW but to make them unequal to everyone else,"59 and for the state to regard unequal treatment as a good thing in and of itself violates the core meaning of the Fourteenth Amendment's guarantee of equal protection of the laws.6° In Lawrence, the Court invalidated under the Fourteenth Amendment's Due Process' Clause a Texas statute that criminalized homosexual sex under the rubric of "deviate sexual intercourse" between persons of the same sex.61 Seventeen years earlier, the Court in Bowers v. Hardwick62 had dismissed as "facetious" 63 the argument that the constitutional right of privacy recognized in cases such as Griswold v. Connecticut 4 and Roe v. Wade65 encompassed sexual conduct between same-sex partners. Revisiting and overruling Bowers, the Court in Lawrence both renounced and denounced the Bowers majority's reasoning. Bowers's "continuance as precedent," the Court maintained, "demeans the lives of homosexual persons." According to the new majority, gays and lesbians have a "protected right" to "engage in intimate, consensual conduct. 67 However controversial Romer and Lawrence may be on and off the Court, no majority opinion of the Court has suggested that either decision is no longer good law. Moreover, despite changes in the Court's composition since Lawrence, Romer and Lawrence appear to continue to command a majority of votes on the Court.6 As authoritative statements of 59 Id. at 635. 60 See id. 61 539 U.S. at 563 (quoting TEX. PENAL CODE ANN. § 21.06(a) (2003)). 62 478 U.S. 186 (1986). 63 Id. at 194. 64 381 U.S. 479 (1965). 65 410 U.S. 113 (1973). 66 Lawrence, 539 U.S. at 575. 67 Id. at 576. 68 The Court's composition did not change between Romer and Lawrence. Since 2005, however, four members of the Court that decided Romer and Lawrence have left and been replaced. Three of the six Justices who voted in those cases to invalidate the state laws under review-Justices O'Connor, Souter, and Stevens-are gone in favor of, respectively, Justices Alito, Sotomayor, and Kagan. One of the three Justices who constituted the minority in those cases--Chief Justice Rehnquist-is gone in favor of Chief Justice Roberts. If, as expected, Justices Sotomayor and Kagan and Chief Justice Roberts take a view on these matters similar to the Justice whom he or she replaced and Justice Alito takes a contrary view, there continue to be at least five votes on the Court for adhering to Romer and Lawrence. 23.1150 COLUMBIA JOURNAL OF GENDER AND LAW constitutional law by the highest court in the land, the two cases establish the invalidity of two purposes that at least arguably underlie some lawmakers' opposition to same-sex marriage. First, Romeroutlaws a purpose ofdisadvantaging gays and lesbians.Alaw that disadvantages gays and lesbians as a means to a lawful end may or may not survive judicial review. The answer depends on (1) the standard of review applicable to laws treating people differently on the basis of sexual orientation-a matter that the Supreme Court has yet to resolve6 9- and (2) whether or not the state's justification for that difference in treatment is sufficiently weighty to satisfy the applicable standard of review. However, a law that disadvantages gays and lesbians as an end in and of itself-a law that, in other words, has the disadvantaging of gays and lesbians as a purpose-is an entirely different matter. It wrongly counts as a plus-as a benefit in the cost-benefit balance that underlies any legislative decision-that a particular group of people has been made to suffer disadvantage.70 Romer leaves no question that, as a purpose, singling out gays and lesbians for disadvantage is constitutionally out of bounds. For lawmakers to oppose same-sex marriage for the purpose of disadvantaging a group, gays and lesbians, whom they dislike is therefore constitutionally deficient as well. Second, Lawrence makes illegitimate a purpose of deterring homosexual sex. If, as Lawrence holds, gays and lesbians have a constitutional right under the Due Process Clause to engage in consenting sexual relations with a partner of the same sex, then lawmakers cannot legitimately defend their opposition to same-sex marriage in terms of a purpose of minimizing the frequency of homosexual sex. The Constitution forbids their treating as a plus in their decision-making process the deterrence of gays' and lesbians' exercise of their due process rights. B. Preventing an Immoral Union Laws against same-sex marriage have been defended as a means of preventing an immoral 69 For the view that laws treating people differently on the basis of sexual orientation should be treated as suspect and subjected to the most rigorous level of review, see WILLIAM N. ESKRIDGE, JR., THE CASE FOR SAME-SEx MARRIAGE: FROM SEXUAL LIBERTY TO CIVILIZED COMMITMENT 172-82 (1996); Simson, supra note 42, at 368-74. For the view that such laws trigger no more than rational basis review, see Lynn D. Wardle, A CriticalAnalysis of Constitutional Claims for Same-Sex Marriage, 1996 BYU L. REv. 1, 62-75, 88-95. 70 See Brest, supra note 15, at 116-17. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW union.7' One difficulty with such a defense stems directly from Lawrence. More specifically, after Lawrence, how much, if any, room is left for a moral objection of any force to same-sex marriage? As indicated above, Lawrence essentially tells lawmakers that regardless of how much they may dislike homosexual sex and regardless of why, they have no business trying to stop it. Gays and lesbians have a constitutional right to engage in homosexual sex. To the extent that lawmakers' moral objection to same-sex marriage is based on moral objection to homosexual sex, Lawrence makes such an objection constitutionally impermissible. In the abstract, it is perhaps conceivable that lawmakers may have a moral objection to same-sex marriage for reasons having nothing to do with a moral objection to homosexual sex. As a practical matter, however, it is difficult to pinpoint what the basis for such a moral objection would be. Surely lawmakers cannot constitutionally rest the objection on a belief that gays and lesbians are simply morally inferior persons and, as such, undeserving of marriage and other societal benefits that more morally sound individuals should be allowed to enjoy. Whatever the precise scope of Romer's holding, it plainly places such a belief outside of constitutional bounds. If lawmakers instead were to try to justify a moral objection to same- sex marriage on the ground that they draw their morality from their religion and their religion disallows same-sex marriage,72 they would avoid the pitfalls of Romer However, they would rush headlong into a virtual concession of unconstitutionality under the Establishment Clause. 71 For example, in reporting out the bill that became the Defense of Marriage Act, the House Judiciary Committee maintained that "[c]ivil laws that permit only heterosexual marriage reflect and honor a collective moral judgment about human sexuality." H.R. REP. No. 104-664, at 15 (1996), reprinted in 1996 U.S.C.C.A.N. 2905, 2919. The committee went on to quote with approval the committee chair's statement in markup that same-sex marriage "legitimates a public union, a legal status that most people.., feel ought to be illegitimate .... And in so doing it trivializes the legitimate status of marriage and demeans it by putting a stamp of approval ... on a union that many people... think is immoral." Id. at 16, 1996 U.S.C.C.A.N. at 2920 (ellipses in original). 72 In defending the proposed Defense of Marriage Act's definition of marriage for federal purposes as limited to heterosexual unions, the House committee that reported out the proposed Act argued along such lines: It is true, of course, that the civil act of marriage is separate from the recognition and blessing of that act by a religious institution. But the fact that there are distinct religious and civil components of marriage does not mean that the two do not intersect. Civil laws that permit only heterosexual marriage reflect and honor a collective moral judgment about human sexuality. This judgment entails both moral disapproval of homosexuality, and a moral conviction that heterosexuality better comports with traditional (especially Judeo-Christian) morality.... Id. at 15-16, 1996 U.S.C.C.A.N. at 2919-20. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW C. Encouraging Procreation No one doubts that a state legitimately may act for the purpose of promoting childbirth. Far more problematic, however, is whether that purpose offers substantial justification for laws prohibiting same-sex marriage. In reporting out the bill that became the Defense of Marriage Act, the House Judiciary Committee cited four governmental interests that in its view were furthered by the bill's two- pronged counterattack to, in the committee's words, "the orchestrated legal campaign by homosexual groups to redefine the institution of marriage. '73 First among the interests that it listed was "defending and nurturing the institution of traditional, heterosexual marriage,"74 and it explained the significance of this interest in terms of encouraging procreation: [From the] nexus between marriage and children spring the true source of society's interest in safeguarding the institution of marriage: "Simply defined, marriage is a relationship within which the community socially approves and encourages sexual intercourse and the birth of children. It is society's way of signaling to would- be parents that their long-term relationship is socially important-a public concern not simply a private affair." That, then, is why we have marriage laws. Were it not for the possibility of begetting children inherent in heterosexual unions, society would have no particular interest in encouraging citizens to come together in a committed relationship .... 7 To similar effect, in holding that "spouse," as used in a provision of federal immigration law, should be interpreted to exclude a person of the same sex, a California federal court relied heavily on the centrality, in its view, of procreation to marriage: The legal protection and special status afforded to marriage (being defined as a union of persons of different sex) has historically . . . been rationalized as being for the purpose of encouraging the propagation of the 73 Id. at 12, 1996 U.S.C.C.A.N. at 2916. 74 Id. 75 Id. at 14, 1996 U.S.C.C.A.N. at 2918 (quoting Council on Families in America, Marriage in America: A Report to the Nation, in PROMISES TO KEEP: DECLINE AND RENEWAL OF MARRIAGE IN AMERICA 303 (David Popenoe et al. eds., 1996)). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW race .... [I]f propagation of the race is basic to the concept of marriage and its legal attributes, "marriage" is again impossible and unthinkable between persons of the same sex.76 The nexus, however, between encouraging procreation and banning same-sex marriage is far more attenuated than the above committee report and judicial opinion assume. Obviously, a same-sex marriage does not hold as much promise of producing offspring as an opposite-sex marriage because same-sex couples, unlike opposite-sex couples, will not be having children by sexual intercourse with one another. For present purposes, however, the relevant question is not whether a same-sex marriage is more apt to produce offspring than an opposite-sex marriage. Rather, it is whether a society that allows same-sex marriage is less apt to procreate than a society that prohibits it. Almost certainly, the answer is "no." In fact, the society that allows same-sex marriage is apt to procreate more. First of all, if not allowed to marry, same-sex couples are highly unlikely to seek fulfillment of their individual desires to wed by breaking up, pairing off with and marrying opposite-sex partners, and producing offspring with them. To believe otherwise is possible only if one refuses to take seriously the growing consensus that sexual orientation is not readily subject to change. Studies over the past two decades strongly suggest that sexual orientation has a substantial genetic component.77 Moreover, the available evidence indicates even more powerfully that sexual orientation is sufficiently set by early adulthood, if not sooner, that for all practical purposes someone old enough to be seriously contemplating marriage is already past the point when his or her sexual orientation may be in his or her capacity to change or 76 Adams v. Howerton, 486 F. Supp. 1119, 1123 (C.D. Cal. 1980), aff'd, 673 F. 2d 1036 (9th Cir. 1982). For other judicial affirmations of procreation as the core of marriage and of marriage as therefore limited to opposite-sex partners, see Baker v. Nelson, 191 N.W.2d 185, 186 (Minn. 1971), appeal dismissed, 409 U.S. 810 (1972); Singer v. Hara, 522 P.2d 1187, 1195 (Wash. App. 1974). 77 See, e.g., Dean'H. Hamer et al., A Linkage Between DNA Markers on the X Chromosome and Male Sexual Orientation, 261 SCIENCE 321 (1993); Simon LeVay, A Difference in Hypothalamic Structure Between Heterosexual and Homosexual Men, 253 SCIENCE 1034 (1991); Nicholas Wade, For Gay Men, Different Scent of Attraction, N.Y. TIMES, May 10, 2005, at Al (discussing Swedish study); Science and Technology: Dyed in the Womb, Homosexuality, ECONOMIST, Oct. !1, 2003, at 99 (discussing British study). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW control.78 Based on the above discussion, one most reasonably would conclude that instituting same-sex marriage would have minimal, if any, effect on the total number of births through sexual intercourse by individuals who would take advantage of same-sex marriage if it were made available. With same-sex marriage, such individuals would be highly unlikely to procreate through sexual intercourse. (If the focus were exclusively the same-sex couple, "highly unlikely" could be changed to "entirely unlikely." Because extramarital opposite- sex sexual intercourse is not wholly outside the realm of possibility, I say "highly unlikely" instead.) Without same-sex marriage, such individuals are at most marginally more likely to have children by sexual intercourse. The above discussion, however, gives only a partial picture of the likely difference in procreation between a society with same-sex marriage and one without. It fails to take into account the probable increase in procreation by same-sex couples through modem reproductive technologies if those couples are allowed to marry. Of course, even if not permitted to marry, same-sex couples, no less than a single individual on his or her own, can take advantage of those technologies and have children. The likelihood, though, that a same-sex couple would decide to avail themselves of reproductive technologies and seek to have children seems appreciably higher if they are given the opportunity to formalize their relationship with state- sanctioned marital vows. Though less the norm than years ago, it is still most common for a man and a woman in a serious relationship to delay starting a family until they are married. This customary decision to delay childrearing until marriage stems partly from the fact that delaying until marriage does not mean delaying indefinitely. Unlike same-sex couples in all but the few states that authorize same-sex marriage, they have marriage as an option whenever they decide that they 78 See, e.g., Marshall Forstein, Overview of Ethical and Research Issues in Sexual Orientation Therapy, in SEXUAL CONVERSION THEPAPY: ETHICAL, CLINICAL AND RESEARCH PERSPECTIVES 167, 167 (Ariel Shidlo et al. eds., 2001); Douglas C. Haldeman, Gay Rights, Patient Rights: The Implications of Sexual Orientation Therapy, 33 PROF. PSYCHOL.: RES. & PRACTICE 260 (2002). Concededly, those who argue the benefits of conversion therapy for gays and lesbians would disagree. See Michael Schroeder & Ariel Shidlo, Ethical Issues in Sexual Orientation Conversion Therapies: An Empirical Study of Consumers, in SEXUAL CONVERSION THERAPY, supra, at 131 (discussing the claims made by conversion therapists). However, in light of the controversiality of conversion therapy, the fact that its proponents take issue with the claim that sexual orientation is very difficult, and at times virtually impossible, to change is not a serious strike against that claim. The American Psychiatric Association, the American Psychological Association, and many individual experts in the field are on record as highly skeptical of the purported benefits of conversion therapy. See id. at 132 (summarizing various individual and organizational views). In addition, some critics of conversion therapy have expressed deep concern that conversion therapy can inflict serious harm. See id. at 160-62 (views of Drs. Michael Schroeder and Ariel Shidlo); Douglas C. Haldeman, Therapeutic Antidotes: Helping Gay and Bisexual Men Recoverfrom Conversion Therapies, in SEXUAL CONVERSION THERAPY, supra, at 117. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW wish to go that route. To a significant extent, however, the common opposite-sex decision to wait until marriage to have children is based on considerations pertinent for same-sex couples even if those couples lack the option of marrying. Many opposite-sex couples, for example, will delay having children until marriage because they do not want to have children until they feel quite certain that their relationship will endure. Although marriage is obviously no guarantee of an enduring relationship, it does call for a solemn public promise of commitment that offers some assurance to each member of the couple that the other intends to be there for the long haul, even if not necessarily 'til death do they part. If marriage is not a possibility for same-sex couples, some couples may decide to go forward with the decision to have children even though they will never have the assurance of continuity provided by an exchange of marital vows. However, for others, the inability to secure that assurance through a state-sanctioned marriage may well prove to be critical. Denied the opportunity to marry, they will be unwilling to take the leap of faith in an enduring relationship that is for them a necessary condition to starting a family. The preceding analysis of a procreation rationale for prohibiting same-sex marriage suggests that such a prohibition is at best of minimal utility in terms of maximizing procreation by the two persons in a same-sex couple and that it may well be counterproductive in terms of that goal. The cogency of a procreation defense of prohibiting same-sex marriage is also undercut by the inordinate underinclusiveness of such a prohibition as a means of encouraging procreation. If lawmakers are genuinely worried that allowing certain couples to marry may inhibit procreation, it hardly makes sense for them to worry only about marriage between same-sex couples. Instead, they should try to prevent opposite-sex marriage that is unlikely to produce children. For example, they might require, as a condition of heterosexual marriage, that the man and women provide medical records attesting to their procreative abilities. Similarly, they might deny opposite-sex couples the right to marry unless the man and woman both affirm in a notarized writing that they wish to have offspring.7 9 D. Promoting Children's Welfare Childrearing by same-sex couples in the United States was a reality well before 79 The report of the House Judiciary Committee in support of the proposed Defense of Marriage Act noted both of these possibilities, but swept them aside as somehow unimaginable: "Surely no one would propose requiring couples intending to marry to submit to a medical examination to determine whether they can reproduce, or to sign a pledge indicating that they intend to do so. Such steps would be both offensive and unworkable." H. REP. No. 104-664, at 14 (1996), reprinted in 1996 U.S.C.C.A.N. 2905, 2918. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW Massachusetts became the first state to legalize same-sex marriage in 2003. By 2000, one- third of female same-sex couple households and almost one-quarter of male same-sex couple households included at least one child under 18.10 Nonetheless, there is little question that legalizing same-sex marriage would make parenting by same-sex couples substantially more prevalent." If parenting by same-sex couples were harmful to children, lawmakers therefore would have good reason to hesitate before encouraging it by authorizing same-sex marriage. Of all the purposes that lawmakers can cite as justification for one or another of its various legislative activities, the promotion of children's well-being is one that virtually no one would characterize as less than very important. The claim that parenting by same-sex couples is disadvantageous to children has been made with some frequency. A 2004 study, citing sources from a decade earlier, observed that "[m] ajor theories of human development have often been interpreted as predicting that youngsters living with same-sex parents would encounter important difficulties in their personal, social, and sexual adjustment, especially during adolescence."82 Similarly, summarizing the relevant research as a prelude to adopting a series of resolutions pertaining to same-sex parenting, the American Psychological Association in 2004 specified three types of child-related "concerns" that have "often been voiced" regarding same-sex parenting: (1) concerns that the children of same-sex parents will "experience more difficulties in the area of sexual identity" than other children; (2) concerns involving elements of those children's "personal development other than sexual identity," such as "more adjustment difficulties and behavior problems" and greater susceptibility to "mental breakdown"; and (3) concerns that such children will have "difficulty in social relationships," including being ridiculed by peers and being abused sexually by a parent or a parent's friends.83 80 See American Psychological Association, Sexual Orientation, Parents, & Children (2004), http://www. apa.org/about/govemance/council/parenting.aspx. 81 Among other things, legalizing same-sex marriage would facilitate adoption by same-sex couples. See Sabrina Tavemise, Adoptions by Gay Couples Rise, Despite Barriers, N.Y TIMES, June 14, 2011, at A I1 (reporting that, according to an advocacy group for families with same-sex parents, "[miost of the legal obstacles facing gay couples intending to adopt stem from prohibitions on marriage .... In most states, gay singles are permitted to adopt."). 82 Jennifer L. Wainright, Stephen T. Russell & Charlotte J. Patterson, Psychosocial Adjustment, School Outcomes, and Romantic Relationships ofAdolescents With Same-Sex Parents, 75 CHILD DEV. 1886, 1897 (2004). 83 American Psychological Association, supra note 80. For examples of commentary urging children's welfare as a prime reason for banning same-sex marriage, see George W. Dent, Jr., "How Does Same-Sex Marriage Threaten You?," 59 RUTGERS L. REV. 233, 240-45 (2007); Trayce Hansen, Same-Sex Marriage: Not in the Best Interest of Children, THE THERAPIST (May/June 2009), http://www.drtraycehansen.com/Pages/ writings notinthebest.html; Amy L. Wax, Traditionalism, Pluralism, and Same-Sex Marriage, 59 RUTrERS L. REv. 377, 402-09 (2007). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW The various harms to children identified by critics of same-sex parenting are disturbing to say the least. Even more disturbing, however, is the disparity between the critics' rather harrowing characterizations of same-sex parents' influence on children and the actual reality. Empirical support for the critics' claims is sorely lacking. Instead, study after study negates the validity of those claims. For example, the 2004 study noted above-a study that examined a sizable national sample of adolescents having same-sex parents and ones having opposite-sex parents-found no support for the negative outcomes that some had predicted based on interpretations of, and extrapolations from, certain theories of human behavior.14 Rather, the study's findings strongly suggested that the key to adolescents' well-being lies in whether they have "close and satisfying relationships" with theirparents, whatever their parents' sexual orientation may be. 85 Similarly, after articulating the concerns that others had raised, the American Psychological Association flatly stated that the "[r]esults of social science research have failed to confirm any of these concems. '8 6 Instead, according to the Association, "research has shown... that the children of lesbian and gay parents are as likely as those of heterosexual parents to flourish. 8' The Association has adopted a resolution that "it is unfair and discriminatory to deny same-sex couples legal access to civil marriage and to all its attendant benefits, rights, and privileges. 88 Other authoritative sources are very much in accord. For example, in endorsing same-sex marriage, the American Psychiatric Association in 2005 was unequivocal about the relative merits of same-sex versus opposite-sex parenting: "[N]o research has shown that the children raised by lesbians and gay men are less well adjusted than those reared within heterosexual relations. '89 In addition, in an interesting and important variation on the child welfare theme, the group emphasized the benefits to children of legalizing same-sex marriage. In their view; the many children being raised by same-sex couples are adversely affected when those couples 84 Wainwright, Russell & Patterson, supra note 82, at 1897. 85 Id. 86 American Psychological Association, supra note 80. 87 Id. 88 Id. 89 American Psychiatric Association, Support ofLegalRecognition ofSame-Sex Civil Marriage: Position Statement (2005), http://www.psych.org/Departments/EDU/Library/APAOfficialDocumentsandRelated/ PositionStatements/200502.aspx. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW are denied the opportunity to marry. According to the group's 2005 position statement: a "stable, adult partnership" has an "invariably positive influence" on the "health of all family members"; heterosexual couples have "a legal framework for their existence through civil marriage, which provides a stabilizing force"; and the children of same-sex parents living in a state that neither authorizes same-sex marriage nor recognizes same-sex marriages validly formed out of state "do not have the same protection that civil marriage affords the children of heterosexual couples." 90 In 2008, three Michigan State researchers published the results of their "comprehensive meta-analysis"'" of the studies to date on the children raised by same-sex and opposite-sex couples. According to the authors, their study, which was based on a search for "relevant studies" that "was as exhaustive as possible and included both published and unpublished literature," '92 "confirmed previous findings ... that parent sexual orientation was not a salient predictor for children's development." 93 More specifically, "children raised by same-sex and heterosexual parents were found to not differ significantly in terms of their cognitive development, gender role behavior, gender identity, psychological adjustment, or sexual preferences."'94 The following year, in a speech accepting the American Psychological Association's annual Award for Distinguished Contributions to Research in Public Policy, University of Virginia Professor Charlotte Patterson offered a very similar overview: More than 25 years of research on the offspring of nonheterosexual parents has yielded results of remarkable clarity. Regardless of whether researchers have studied the offspring of divorced lesbian and gay parents or those born to lesbian or gay parents, their findings have been similar. Regardless of whether researchers have studied children or adolescents, they have reported similar results. Regardless of whether investigators have examined sexual identity, self-esteem, adjustment, or qualities of social relationships, the results have been remarkably consistent. In study after study, the offspring of lesbian and gay parents have been found to be at least 90 Id. 91 Alicia Crowl, Soyeon Ahn & Jean Baker, A Meta-Analysis of Developmental Outcomes for Children of Same-Sex and Heterosexual Parents, 4 J. GLBT FAM. STUD. 385, 389 (2008). 92 Id. 93 Id. at 398. 94 Id. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW as well adjusted overall as those of other parents.95 Without canvassing the various additional sources to similar effect, suffice it to say that the available research on the relative well-being of children of same-sex and opposite-sex couples is dramatically at odds with a child-welfare defense of prohibitions on same-sex marriage. E. Protecting the Institution of Traditional, Heterosexual Marriage In reporting out the bill that became the Defense of Marriage Act, a majority of the House Judiciary Committee cited as a prime reason for rejecting same-sex marriage the need to safeguard "the institution of traditional, heterosexual marriage. '96 The committee majority has not been alone in invoking that objective as a basis for opposing same-sex marriage.97 The cogency of opposing same-sex marriage on that basis, however, is tenuous at best. First of all, it is unclear that the objective of protecting the institution of traditional, heterosexual marriage is entitled to significant weight. Surely it is not entitled to such weight on the theory that the defense of any tradition is a weighty goal. Although Justice Scalia, for one, would probably disagree,98 the fact that something is a tradition does not settle that protecting it has special value. Concededly, some traditions are worth preserving, but some clearly are not. The tradition, for example, in much of the United States until Brown v. Board ofEducation99 of racially segregated schools is one that few people today would want to honor. In addition, well before Brown, the Thirteenth Amendment'0 brought to an end a tradition in many states that people today see as even less deserving of preservation-the white majority's 95 Charlotte J. Patterson, Children of Lesbian and Gay Parents: Psychology, Law, and Policy, 64 AM. PSYCHOLOGIST 727, 732 (2009). 96 H.R. REP. No. 104-664, at 12 (1996), reprinted in 1996 U.S.C.C.A.N. 2905, 2916. 97 To similar effect, see, for example, Dent, supra note 83, at 254-56; John M. Finnis, Law, Morality and "Sexual Orientation," 69 NOTRE DAME L. REv. 1049, 1070 (1994).' 98 Justice Scalia holds tradition in such high esteem that he has urged, often quite emphatically, that constitutional importance be given to it in a variety of contexts. See, e.g., Lee v. Weisman, 505 U.S. 577, 631-36 (1992) (Scalia, J., dissenting) (arguing that tradition calls for upholding the constitutionality of graduation prayer under the Establishment Clause); Burnham v. Superior Court, 495 U.S. 604, 608-16 (1990) (plurality opinion of Scalia, J.) (arguing that tradition calls for upholding the constitutionality of transient presence as a basis for personal jurisdiction under the Due Process Clause). 99 347 U.S. 483 (1954). 100 Adopted in 1865, the Thirteenth Amendment provides: "Neither slavery nor involuntary servitude, except as a punishment for crime whereof the party shall have been duly convicted, shall exist within the United States, or any place subject to their jurisdiction." U.S. CoNsT. amend. XIII, § I. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW enslavement of African Americans. It is also difficult to argue that protecting the institution of traditional, heterosexual marriage is important because it is a subset of an objective that is itself important - specifically, the objective of protecting the institution of traditional marriage. In the abstract, protecting the institution of traditional marriage may sound like an important objective. In reality, however, its importance is belied by states' general lack of regard for it. Although the couple's heterosexuality is undoubtedly an aspect of traditional marriage in the United States, so is permanence, procreation, and sexual exclusivity. 0' If states have honored the latter three aspects of traditional marriage, the case for honoring the fourth aspect, permanence, would be quite strong. Particularly in recent years, however, nothing of the sort has occurred. First, as evidenced by the general trend in state laws since the late 1960s away from relatively narrow, fault-based grounds for divorce and toward substantially broader, no-fault grounds, 102 states in the past several decades have attached increasingly less importance to the permanence aspect of traditional marriage. Although states continue to regard the impermanence of a marriage as unfortunate, they generally have come to see permanence, in and of itself, as an inadequate guidepost for state action. In essence, in addressing the unhappy situation of a married couple who are no longer compatible and a marriage that has broken down, state divorce laws today worry far more about extricating the couple from the marriage with as little discord and long-term damage as possible than about perpetuating the marriage out of service to some ideal of permanence. Second, states have long since detached the institution of marriage from any requirement or even expectation that the couple will procreate. Most obviously, if the procreation aspect of traditional marriage were truly intact, states would commonly bar or at least openly discourage, marriage by women past childbearing age. Plainly, however, states do neither." 3 Third and lastly, although laws against adultery typically remain on the books, states' almost uniform nonenforcement of them today leaves little doubt that states' commitment to the sexual exclusivity aspect of traditional marriage is far more superficial than real. The House Judiciary Committee's report offers yet another basis for claiming that the protection of the institution of traditional, heterosexual marriage is a weighty goal - its 101 For detailed discussion of the components of traditional marriage in the. United States and states' increasing disregard for them, see Jessica R. Feinberg, Exposing the Traditional MarriageAgenda, Nw. J.L. & Soc. POL'Y (forthcoming 2012). 102 See JOHN DE WITT GREGORY, PETER N. SWISHER & SHERYL L. WOLF, UNDERSTANDING FAMILY LAW 236-38 (3d ed. 2005); ROBERT E. OLIPHANT & NANCY VER STEEGH, FAMILY LAW 94-96 (3d ed. 2010). 103 More generally, see the discussion of procreation supra Part IVC. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW connection to the states' "deep and abiding interest in encouraging responsible procreation and child-rearing." The connection to the latter interest, however, fails to establish that the cited goal is a weighty one, because the connection is flimsy at best. As noted immediately above, potential to procreate is plainly not a precondition for marriage under the laws of any state. Nor do any states make an effort to bar or even discourage couples from getting married who have no desire whatsoever to have children. The connection is no more impressive if the link to "responsible" child-rearing is the focus. As already discussed, study after study belies the notion that same-sex couples are any less likely than opposite-sex couples to be good parents.104 Even assuming for purposes of argument that the objective of protecting the institution of traditional, heterosexual marriage is a weighty one, it does not provide a substantial basis for opposing same-sex marriage, because a prohibition on same-sex marriage is a poor means of serving it. Prohibiting same-sex marriage may be thought to serve as a means of protecting heterosexual marriage under two theories, neither of which deserves much credence. Under one theory, prohibiting same-sex marriage protects heterosexual marriage by removing a possible disincentive to individuals to enter into, and remain in, a heterosexual marriage. The thinking would be that if same-sex marriage is an option, some individuals who would otherwise marry, and stay married to, someone of the opposite sex would instead enter into a same-sex marriage. Such thinking is minimally rational at best. Among other things, it proceeds on an assumption of the mutability and impermanence of sexual orientation that is at odds with the great weight of recent genetic and psychological studies.°5 In addition, it depends for its logic on the dubious assumption that individuals who would otherwise marry, and stay married to, someone of the opposite sex would be significantly more likely to opt out of entering into, or remaining in, a heterosexual marriage if the opt-out held the possibility of a same-sex marriage rather than only the possibility of living together in a same-sex relationship. Under an alternative theory, prohibiting same-sex marriage protects heterosexual marriage by eliminating a possible form of marriage that some opposite-sex couples would find so repugnant that they would avoid marrying, or dissolve their marriage, rather than exist in a relationship that bears the same name. I have no doubt that some individuals are so repulsed by the thought of sexual relations between same-sex couples that they could not help but think of the state's tacitly approving such relations by authorizing same-sex marriage as demeaning the entire institution of marriage. I am far less persuaded that the state's authorizing same-sex marriage would cause those individuals to adopt such an unfavorable 104 See supra Part IV.D. 105 See sources cited supra notes 77 & 78 and accompanying text. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW view of marriage that they would forgo the benefits to themselves of heterosexual marriage. Even assuming, however, that some such individuals would, it is questionable whether their behavior is deserving of substantial weight. In particular, to give it substantial weight reaffirms a way of thinking about sexual relations between same-sex couples that, under the Supreme Court's reasoning in Lawrence, is constitutionally anathema. 0 6 Just as the Supreme Court in Palmore v. Sidoti'07 rejected racial prejudice as a basis for preferring adoption by uniracial, over interracial, couples, so should courts, in keeping with Lawrence, refuse to give weight to arguments against same-sex marriage rooted in prejudice against lesbians and gays. F. Guarding Against the Erosion of Other Societal Prohibitions Some critics of same-sex marriage have warned that the state's legitimating same-sex marriage undermines the existing laws and social norms against certain types of marriage and against sexual relationships widely viewed as antithetical to the general welfare. Very simply, if same-sex marriage is fine, then why aren't polygamous and incestuous marriages and even marriage with animals-along with the sexual relationships contemplated by such marriages-fine as well?' 08 From these critics' perspective, there is no logical stopping point between approving same-sex marriage and approving those other forms of marriage and the sexual relationships entailed. In their view, if same-sex marriage wins official approval, the legal barriers to those other forms of marriage and sexual relationships inevitably will atrophy; even if some such barriers remain on the books, society's moral opposition to them will have 106 In explaining its reasons for overruling the holding of Bowers v. Hardwick that a state may criminally punish a same-sex couple for engaging in private in sexual relations, the Court in Lawrence maintained: When homosexual conduct is made criminal by the law of the State, that declaration in and of itself is an invitation to subject homosexual persons to discrimination both in the public and in the private spheres .... [Bowers's] continuance as precedent demeans the lives of homosexual persons. Lawrence v. Texas, 539 U.S. 558, 575 (2003). 107 466 U.S. 429 (1984). 108 See, e.g., George W. Dent, Jr., Traditional Marriage: Still Worth Defending, 18 BYU J. PUB. L. 419, 440-43 (2004) (maintaining that authorizing same-sex marriage would "logically necessitate" authorizing "[p] olygamy, endogamy, and marriage with minors and animals"); The Good Book and Gay Marriage, supra note 7 ("If we redefine marriage contrary to how the Bible defines it and to how society has developed it, why should we stop with homosexual couples? What prevents it from eventually including polygamy? Nothing.") (excerpt from comments of Dr. Barrett Duke, vice president for public policy and research at Southern Baptist Convention's Ethics and Religious Liberty Commission); see also Jay Michaelson, Chaos, Law, and God: The Religious Meanings of Homosexuality, 15 MICH. J. GENDER & L. 41, 53 (2008) (discussing the "slippery slope arguments that acceptance of same-sex relationships will lead to the erosion of taboos against polygamy, bestiality, and incest"). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW eroded so substantially that they will go unenforced. One possible response to the above argument is that at least some of the predicted repercussions of approving same-sex marriage are not all that bad. For example, in the eyes of some groups, polygamy serves a number of useful functions, and laws in the United States fail to give polygamy its due.'09 By the same token, although some of the marital and sexual relationships prohibited by incest laws (for example, parent-child) are ones that virtually no one would champion, there are others (for example, cousin-cousin) that historically have gotten a more mixed reception in the United States."' At least to some people, a narrowing of incest laws to make room for some marital and sexual relationships now typically prohibited might seem a good thing. However, even someone fully persuaded of the value of maintaining and enforcing current legal prohibitions on polygamy, incest, and bestiality should be unimpressed by an argument to prohibit same-sex marriage as a means of avoiding the erosion of those other prohibitions. The logical connection between authorizing same-sex marriage and undermining those prohibitions is attenuated at best. It appears otherwise only if one ignores the differences between the individual and governmental interests at 'stake. On the one hand, the individual interest in entering into a same-sex marital relationship is very different in nature from the individual interests in entering into a polygamous, incestuous or bestial marital relationship. Like traditional heterosexual marriage, same-sex marriage revolves around what the Lawrence Court called an "intimate, consensual""' relationship between two individuals. The same cannot be said for any of the other three types of marriage. Unlike same-sex and traditional heterosexual marriage, polygamous marriage entails multiple partners. Also, largely because of that multiplicity in partners, it promises materially less intimacy. Incestuous marriage entails closeness in consanguinity and, in many instances, disparities in power that make it demonstrably less consensual than the typical same-sex or traditional heterosexual marriage. Marriage with an animal obviously is one person short of the two persons central to a same-sex or traditional heterosexual marriage. Indeed, the claim made, explicitly or implicitly, by some critics of same-sex marriage that it and bestial marriage are functionally indistinguishable is so outlandish and wildly insulting that I hesitate 109 See Adrienne D. Davis, Regulating Polygamy: Intimacy, Default Rules, and Bargaining for Equality, 110 COLUM. L. REv. 1955, 1969-75 (2010) (discussing religious defenses of polygamy by non-LDS Mormons, Muslims, and others, as well as secular defenses by radical feminists, black nationalists, and others). 110 See Frederic P. Storke, The Incestuous Marriage-Relic of the Past, 36 U. COLO. L. REv. 473 (1964). Ill Lawrence, 539 U.S. at 576. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW to address it at all." 2 On the other hand, the governmental interests in prohibiting polygamous, incestuous or bestial marriages are very unlike those militating against recognition of same-sex marriage. Prohibitions on polygamous marriage have been defended largely in terms of concerns about children's well-being, possible exploitation of women, and interference with the desired relation in a democracy of the people to their government. "3 Only the first of these concerns has any relevance in the same-sex marriage context, and it is fairly seen as much less problematic in that realm." 4 Prohibitions on incestuous marriage principally rest on concerns about the likelihood of genetic abnormalities in children and of serious disparities in power between the partners to the marriage."' Neither concern provides a plausible basis for opposition to same-sex marriage. Lastly, laws against bestial marriage are based on concerns, ranging from animal cruelty to the psychological well-being of the human would-be partner, so different in kind and degree from those implicated by same-sex marriage as to make irrational any attempt to equate the two sets of concerns. G. Preserving Institutional and Individual Religious Liberty Another argument offered at times for prohibiting same-sex marriage is that legalization seriously threatens the ability of religious institutions to carry out their religious missions and the ability of religiously observant individuals to live in accordance with their religious beliefs. Focusing on the arguments made by two prominent commentators-Roger Severino, legal counsel for the Becket Fund for Religious Liberty, and law professor George Dent, I6 -- I 112 The Supreme Court long ago made clear its view that the individual interest in polygamy is devoid of constitutional force. See Reynolds v. U.S., 98 U.S. 145, 165-66 (1878). Moreover, although the Court has yet to rule on the constitutional stature of the individual interests in incest and bestiality, its implicit understanding of those interests as constitutionally inconsequential is surely beyond any doubt. 113 See Davis, supra note 109, at 1975-78. With regard to the concern about the possible exploitation of women, I note that "polygamy" does not necessarily mean one husband and multiple wives. It is group marriage in whatever form. However, with rare exception, the configuration of one husband with multiple wives-"polygyny"-- is overwhelmingly the form of polygamy present in cultures that permit it and the form practiced by the 30,000 to 100,000 families that illegally practice polygamy in the United States. See id. at 1966-68. 114 As discussed supra Part IV.D, there is no good reason to believe that children's well-being is any more at risk with same-sex parents than with parents of the opposite sex. In contrast, there is evidence suggesting that children of polygamous marriages may be especially at risk. See Davis, supra note 109, at 1976-77 & nn.59-62. 115 On the concerns underlying prohibitions on incestuous marriage, see Storke, supra note 110, at 476-78. 116 See Dent, supra note 108, at 443-45; Roger Severino, Orfor Poorer? How Same-Sex Marriage Threatens Religious Liberty, 30 HARV. J.L. & PUB. POL'¥ 939(2007). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW maintain below that the threat posed to institutional and individual religious liberty provides no more than meager support for a legislative decision to oppose same-sex marriage. 1. Institutional Religious Liberty Several years ago, Mr. Severino published an article cataloguing the "multiplicity of serious risks for religious institutions" posed by legal recognition of same-sex marriage." 7 According to Severino, "The specific consequences that will likely flow from legalizing same-sex marriage include both government compulsion of religious institutions to provide financial or other support for same-sex married couples and government withdrawal of public benefits from those institutions that oppose same-sex marriage."' 8 He warns, for example, that legalizing same-sex marriage is apt to lead to suits by same-sex couples (a) charging religious institutions with violation of state or federal employment discrimination laws and fair housing laws" 9 or (b) seeking to bar such institutions from equal access to such governmental benefits as tax-exempt status and government-funded social service contracts. 20 Severino makes clear that he believes that the aggregate risks to institutional religious liberty constitute a strong counterweight to any arguments for legislative adoption of same-sex marriage. In his view, "when weighing the benefits and costs of adopting as fundamental social change as same-sex marriage, particularly close consideration must be given to its impact on religious freedom,"' 2 and in this instance, "the likely cost to religious liberty is a high one indeed."'' 22 As discussed below, Severino's argument fails for two reasons. He does not adequately take into account the extent to which federal and state constitutional protections reduce the possible adverse effect of legalizing same-sex marriage on religious institutions' religious liberty. Even more importantly, he fails to factor into his analysis the latitude that legislatures enjoy to use statutory exemptions to protect against any such adverse effect. a. Federal and State Constitutional Protections In the introduction to his article, Severino acknowledges that "[r]eligious institutions 117 Severino, supra note 116, at 957. 118 Id. at 943. 119 Id. at 958-63. 120 Id. at 972-76. 121 Id. at 979-80. 122 Severino, supra note 116, at 980. 166 23.1 COLUMBIA JOURNAL OF GENDER AND LAW will be able to assert a wide range of substantial First Amendment defenses"'123 to litigation that, in one way or another, attempts to require them to act inconsistently with their faith- based opposition to same-sex marriage. He then proceeds, however, to offer a one-paragraph summary of the potentially relevant First Amendment guarantees with no attempt to describe the extent to which each guarantee promises to relieve religious institutions of any burdens that the legalization of same-sex marriage might appear to place on them.'24 In addition, immediately after reciting the guarantees, he essentially sweeps them aside as only marginally relevant. "It is difficult," he maintains, "to predict the ultimate effectiveness of these constitutional defenses after several years of precedents eroding religious liberty."'25 Rather than wrestle with this difficulty, he is content to intimate broadly that, in light of general trends in Supreme Court case law on religious liberty and gay rights, the constitutional defenses are apt to be ineffective and therefore need not be explored in any detail for purposes of the present discussion. 26 As he explains in a footnote, an "in-depth analysis" of the defenses "is beyond the scope of this Article."'27 Unfortunately, it is not nearly as easy as Severino seems to believe for him to justify giving such limited attention to First Amendment and other constitutional defenses. In an article that claims, as Severino's does, to make the case that potential interference with religious institutions' religious liberty should count in the legislative cost-benefit balance as a cost that is a "high one indeed,"'28 it is essential to examine carefully any such defenses that may materially reduce that cost. Most notably, as a student author, Fredric Bold, Jr., more recently has demonstrated in some detail, 29 the First Amendment's guarantee of freedom of expressive association is of 123 Id. at 944. 124 Id. at 944-45. 125 Id. at 945. 126 See id. 127 Id. at 944 n. 13. 128 Id. at 980. 129 Fredric J. Bold, Jr., Comment, Vows to Collide: The Btgeoning Conflict Between Religious Institutions and Same-Sex Marriage Antidiscrimination Laws, 158 U. PA. L. REV. 179, 215-29 (2009). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 central importance in this regard. 30 Although the First Amendment does not explicitly mention this freedom, the Supreme Court has long held that the First Amendment implicitly elevates it to constitutional stature. 3' Under the Court's decision in 2000.in Boy Scouts ofAmerica v. Dale,3 2 religious institutions enjoy a freedom of expressive association that shields them from any laws that, if enforced against them, would have the effect of changing substantially the institution's self-definition. 33 A law that enables same-sex couples to marry and that redefines "marriage" for purposes of any state law or regulation that bears on a religious organization's activities and structure surely may have such an effect. The Free Exercise Clause of the First Amendment may not afford religious institutions as much protection from the repercussions of legalizing same-sex marriage as the freedom of expressive association affords, but the protection that it does provide is significant nonetheless. The Supreme Court's decision in 1990 in Employment Division v. Smith 134 interpreted the Free Exercise Clause much more narrowly than the Court had been construing the clause in prior 130 Rather remarkably, although Severino devotes only a sentence to the Supreme Court's opinion in Dale as a basis for a religious institution's freedom of association defense to particular requirements that might arise from legalizing same-sex marriage, see Severino, supra note 116, at 967, he spends two pages discussing a statutory question that the New Jersey Supreme Court resolved in the course of deciding Dale on its way to the U.S. Supreme Court. See id. at 966-67. The New Jersey high court's resolution of that question was of no legal effect, because the U.S. Supreme Court reversed on other grounds. Moreover, the New Jersey high court resolved the question in a way that, by Severino's own representation, seems not to merit substantial exposition. He cites four cases'in other jurisdictions that disagree with the New Jersey Supreme Court's interpretation and only one case in agreement. Id. at 966 n. 122. 131 See NAACP v. Button, 371 U.S. 415,428-29 (1963); NAACP v. Alabama, 357 U.S. 449,460-61 (1958). 132 530 U.S. 640 (2000). 133 Dale held that the Boy Scouts of America's freedom of expressive association called for the rejection of New Jersey's attempt, pursuant to its public accommodations law, to require the Boy Scouts to accept an openly gay individual as an assistant scoutmaster. The majority in Dale deferred to the Boy Scouts' representations that (a) homosexual conduct is contrary to the organization's values and (b) the presence of an openly gay assistant scoutmaster would substantially impair the organization's desired message that homosexual conduct is bad. -The Supreme Court's decision in 2010 in Christian Legal Society v. Martinez, 130 S. Ct. 2971 (2010), does not detract from the protection afforded by Dale. In Christian Legal Society the Court held that a state law school was free to withdraw official recognition from the Christian Legal Society (CLS) because CLS had failed to abide by the law school's policy that student organizations must accept all comers. CLS did not allow any students to join whose religious beliefs did not conform to the group's "Statement of Faith" or who engaged in homosexual conduct. The Court distinguished Dale on the ground that CLS, unlike the Boy Scouts, was not being forced to include an unwanted member. According to the Court, CLS could exclude whomever it wished as long as it was willing to operate without the benefits of official law school recognition. 134 494 U.S. 872 (1990). COLUMBIA JOURNAL OF GENDER AND LAW years. In Smith, a sharply divided Court held that the Free Exercise Clause does not protect religious claimants from any burden that a generally applicable law may happen to place on them as a result of their religious beliefs.'35 Prior to Smith, the Court had long interpreted the clause to authorize courts to carve out exemptions from generally applicable laws. 136 If a law could not be enforced without imposing a substantial burden on a claimant's religious liberty, a court would prohibit enforcement against the religious claimant unless the state could show that such enforcement was necessary to serve a compelling state interest. Although the Court's opinion in Smith adopted an approach to free exercise less protective of free exercise claimants than the preexisting approach, it did not render free exercise protection inconsequential. In particular, seemingly out of a desire to avoid overruling certain cases that one or more of the five Justices subscribing to the majority opinion wished to preserve, the Court in Smith created several exceptions to the basic approach that it lay down. 37 The Court left room, for example, for courts to carve out religious exemptions where a generally applicable law substantially burdens both free exercise and another constitutional right. 1 38 As one might expect of exceptions designed primarily to preserve particular cases rather than promote analytical coherence, the exceptions are quite indefinite in scope. Moreover, as Professor Nelson Tebbe recently has observed, lower courts have been willing to take 135 The opinion of the Court was written by Justice Scalia and joined by four others. Justice O'Connor wrote an opinion concurring in the judgment and explicitly repudiating an approach to the Free Exercise Clause that she characterized as "dramatically depart[ing] from well-settled First Amendment jurisprudence" and "incompatible with our Nation's 'fundamental commitment to individual religious liberty." Smith, 494 U.S. at 891 (O'Connor, J., concurring). Justice Blackmun, joined by Justices Brennan and Marshall, wrote a dissenting opinion in which he subscribed to Justice O'Connor's criticisms of the Court's approach but disagreed with the result that she reached in applying the approach that she and the dissenters agreed should apply. See id. at 909 (Blackmun, J., joined by Brennan and Marshall, JJ., dissenting). For a sampling of the copious scholarly commentary critical of Smith, see JESSE H. CHOPER, SECURING RELIGIOUS LIBERTY 54-61 (1995); DANIEL 0. CONKLE, CONSTITUTIONAL LAW: THE RELIGION CLAUSES 91-95 (2d ed. 2009); Douglas Laycock, The Remnants of Free Exercise, 1990 SuP. CT. REV. 1; Michael W. McConnell, Free Exercise Revisionism and the Smith Decision, 57 U. CHI. L. REv. 1109 (1990). For a recent defense of Smith, including an attempt to demonstrate that the widely shared view of Smith as a major departure from prior precedent is wrong, see Marci A. Hamilton, Employment Division v. Smith at the Supreme Court: The Justices, the Litigants, and the Doctrinal Discourse, 32 CARDOZO L. REv. 1671 (2011). 136 See, e.g., Hobbie v. Unemployment Appeals Comm'n, 480 U.S. 136, 141-42 (1987); Wisconsin v. Yoder, 406 U.S. 205, 212 (1972); Sherbert v. Verner, 374 U.S. 398, 406 (1963). 137 See Nelson Tebbe, Smith in Theory and Practice, 32 CARDOZO L. REV. 2055, 2057-58 (2011) (describing four apparent exceptions and discussing their apparent roots in a desire on the part of some Justices to preserve certain precedent). 138 See Smith, 494 U.S. at 881-82. For discussion of other exceptions, see Laycock, supra note 135. 23.1 170 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 advantage of the Smith exceptions' fuzzy-at-best contours to grant relief to religious claimants whose circumstances strongly favor it.'39 Accordingly, if legalizing same-sex marriage would have the effect of making a generally applicable law apply to a religious institution in a manner that substantially and unnecessarily burdens the institution's religious liberty, a court could be expected to seize upon one of Smith's ill-defined exceptions to exempt the institution from application of the law. Severino also fails entirely to take into account the importance of state constitutional law to the validity of his basic thesis that legalizing same-sex marriage entails a "cost to religious liberty" that is a "high one indeed."' 4 Every state constitution contains language that, in essence, guarantees the free exercise of religion.' 4' In some instances, the language closely tracks the language of the federal constitutional clause; in other instances, there are more significant differences, including differences that at times facilitate an interpretation of free exercise under the state constitution more protective than Smith's interpretation of free exercise under the federal Constitution. 42 Even in states where the state constitutional language guaranteeing free exercise does not vary materially from the federal clause, Smith is no obstacle to the state courts' rejecting as inadequate under the state constitution an interpretation of free exercise that, like the Court's in Smith, makes no room for courts to 139 See Tebbe, supra note 137, at 2059-60. 140 Severino, supra note 116, at 980. 141 See Nicholas P. Miller & Nathan Sheers, Religious Free Exercise under State Constitutions, 34.J. CHURCH & STATE 303,322 (1992) (table listing citations for the free exercise provisions in all fifty states' constitutions). 142 According to a study of state constitutional provisions done not long after Smith: Four basic clauses providing free exercise protection appear widely in state constitutions. These are: (I) a provision for "no interference" with worship and conscience; (2) a "peace and safety" limitation on religious activity; (3) aprohibition against persecution "on account of' religion; and (4) a "free exercise" clause identical to that found in the federal constitution .... [Some] states have constitutions that contain combinations of these clauses. Miller & Sheers, supra note 14 1, at 312. Citing as authority the Minnesota Supreme Court's decision shortly after Smith in State v. Hershberger, 462 N.W2d 393 (Minn. 1990)--a decision that interpreted the state constitution's free exercise provision as much more protective of individual rights than the federal provision in the wake of Sniththe authors of the study maintained that provisions of either of the first two varieties described above facilitated a substantially more claimant-friendly interpretation of state, than federal, constitutional free exercise rights. See Miller & Sheers, supra, at 312-15. The authors further argued that clauses of the third variety were also fairly read as inviting an interpretation of state free exercise rights more generous to claimants than a Smith-like interpretation, see id. at 315-18, and they concluded by emphasizing that even states with clauses identical to the federal one were by no means bound in interpreting the state constitutional clause to follow the lead of Smith, see id. at 318. COLUMBIA JOURNAL OF GENDER AND LAW carve out exemptions to generally applicable laws. As the ultimate authority on the meaning of their respective state constitutions, state courts are free to interpret their state constitutions' provisions protecting free exercise in a manner considerably more protective of free exercise than Smith.'43 Smith was so strongly criticized and so widely unpopular that it led within three years to Congress's virtually unanimous enactment of legislation to undo it.'" Although that federal legislation, the Religious Freedom Restoration Act, failed to survive Supreme Court review except insofar as it authorizes free exercise exemptions from federal laws,"a5 courts in various states have taken advantage of their interpretive freedom in construing their state constitutions to distance themselves from Smith.' In those states, religious institutions have broader opportunity than under federal free exercise alone to secure judicial protection from inroads on their religious liberty resulting from the enforcement of laws allowing same-sex marriage. b. Statutory Exemptions In considering the impact of legalizing same-sex marriage on religious institutions' religious liberty, Severino addresses the possible significance of statutory exemptions in much the same way as he addresses the possible significance of constitutional guarantees. As with constitutional guarantees, he acknowledges that statutory exemptions exist that mitigate the impact on institutional religious liberty, 47 but he disclaims any obligation on his part to analyze carefully the magnitude of the mitigating effect. According to Severino, an "in-depth analysis" of this sort is "beyond the scope" of his article.' 48 Furthermore, as with 143 See Gary J. Simson, Reflections on Free Exercise: Revisiting Rourke v. Department of Correctional Services, 70ALB. L. REV. 1425, 1427 (2007). 144 For discussion of the background of the Religious Freedom Restoration Act of 1993, see Douglas Laycock, Free Exercise and the Religious Freedom Restoration Act, 62 FORDHAM L. REv. 883 (1994); Simson, supra note 16, at 442-43. 145 See City of Boeme v. Flores, 521'U.S. 507 (1997) (holding that the Act was unconstitutional as applied to state laws for lack of congressional authority to enact such legislation under Section 5 of the Fourteenth Amendment). 146 See, e.g., State v. Hershberger, 464 N.W.2d 393 (Minn. 1990); Catholic Charities v. Serio, 859 N.E.2d 459 (N.Y. 2006); Rourke v. Dep't of Corr. Servs., 603 N.Y.S.2d 647 (Sup. Ct. 1993), aff'd, 615 N.Y.S.2d 470 (App. Div. 1994); First Covenant Church v. Seattle, 840 P.2d 174 (Wash. 1992). For an account of the Rourke case and its importance by one of the claimant's attorneys, see Simson, supra note 143. 147 See, e.g., Severino, supra note 116, at 959 ("[B]oth federal and state law specifically exempt religious institutions from prohibitions on religious discrimination."). 148 Id. at 944 n.13. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 constitutional guarantees, he does not let his disclaimer of careful analysis stand in the way of his offering a sweeping generalization that strongly suggests that the mitigating effect of statutory exemptions is apt to be of little importance. 49 I have argued that Severino could not reasonably, on the one hand, deny the need to analyze closely the mitigating effect of constitutional guarantees and, on the other hand, claim that he has shown that the legalization of same-sex marriage threatens to exact a cost on institutional religious liberty that is "a high one indeed."' ° His denial of the need to analyze closely the mitigating effect of statutory exemptions is, if anything, even more misguided. Unlike constitutional guarantees, which are beyond the legislature's power to alter, statutory exemptions are well within that legislative power. As a result, if he wishes to try to claim a high cost to religious liberty, he can be faulted for failing to analyze closely not only the mitigating effect of existing statutory exemptions but also that of statutory exemptions likely to be adopted. Severino's failure to consider statutory exemptions that are likely to be adopted is so problematic because of the high probability that a legislature that decides to legalize same-sex marriage will resort to exemptions to minimize or eliminate entirely any significant adverse effects of legalization on religious institutions' religious liberty.'5' I suggest that legislatures are highly likely to resort to exemptions for two reasons. 149 See id. at 960 ("As state legislatures increasingly grant protection for sexual orientation through anti- discrimination laws, these traditional religious exemptions may be modified or omitted by legislatures or narrowed by courts to the point of vanishing."). 150 Id. at 980. 151 Recent legislative activity in Rhode Island illustrates in the context of a civil union law how very effective lobbyists for religious exemptions can be. In the face of opposition led by the State Senate president, state lawmakers supportive of same-sex marriage decided not to press for enactment of a same-sex marriage bill and threw their support behind a civil union bill. However, when the civil union bill came to the House floor, an amendment was introduced and adopted that broadly exempted religious and religiously affiliated organizations from any obligation to provide services, facilities or accommodations to a couple joined in a civil union. Advocates for gay and lesbian rights protested that, for example, "a Catholic hospital could choose not to allow a lesbian to make medical decisions on behalf of her partner, and a Catholic university could deny family medical leave to gay employees." Ray Sullivan, the campaign director for Marriage Equality Rhode Island, acknowledged that "[m]ost civil union and gay marriage bills offer some religious protections-allowing a minister not to perform a gay marriage ceremony if he so chooses, for example." He also stated that he and his organization .'support common-sense exemptions."' He maintained, however, that the exemptions in the Rhode Island law went much too far. See Goodnough, supra note 3. Governor Chafee, who had made clear at the outset his support for a same-sex marriage bill, signed the amended civil union bill into law. In doing so, however, he expressed serious reservations about the religious exemption, characterizing it as one of "unparalleled and alarming scope." See Dana Rudolph, Governor Signs Rhode Island Civil Union Law, But Pleases No One, KEEN NEWS SERVICE, July 3, 2011, http://www. keennewsservice.com/2011/07/03/govemor-signs-rhode-island-civil-union-law-but-pleases-no-one/. COLUMBIA JOURNAL OF GENDER AND LAW First, by creating exemptions that ensure that legalizing same-sex marriage does not materially interfere with institutional religious liberty, lawmakers who vote for same-sex marriage help defuse religious opposition to their votes. They do so both by softening the blow felt by religious institutions opposed to same-sex marriage and by demonstrating that they are respectful of religion. Second, the religions whose institutional religious liberty is apt to be threatened by the legalization of same-sex marriage are ones that, in the aggregate and, in some instances, individually, are well represented among the electorate and capable of wielding significant political power. It is one thing for lawmakers to vote to legalize same-sex marriage and run the risk of incurring political retribution at the hands of religions opposed to legalization. It is quite another for them to refuse to make reasonable accommodation by statutory exemptions for the special burdens that legalization may place on those religions' institutional religious liberty. Depending on the particular state, legislative action of the former variety is more or less politically bold. Legislative action of the latter variety, however, is almost invariably politically reckless. Concededly, the Establishment Clause places some limits on what a legislature can do in the way of statutory exemptions in order to protect religious institutions from possible adverse effects that legalizing same-sex marriage may have on their religious liberty. However, under the relevant Supreme Court precedent, those limits are marginal in the extreme. As Justice O'Connor, almost certainly the Justice most influential in shaping the endorsement test,'52 explained in a concurring opinion, the government is neither fairly understood nor reasonably perceived as endorsing religion when "it lifts a government-imposed burden on the free exercise of religion."' 53 In essence, when the government acts in this manner, its purpose should be understood as one of affirming the values underlying the Free Exercise Clause, not as one of endorsing religion.'54 In addition, the effect of the government's action should be perceived as one of affirming free exercise values, not endorsing religion.'55 The Court has made clear that it regards burden-lifting of the sort described by Justice O'Connor as a nonendorsement of religion even when the burden that the government is lifting is not one that a court appropriately could have ordered it to lift as a matter of compliance with the 152 See CONKLE supra note 135, at 125-28. 153 Wallace v. Jaffree, 472 U.S. 38, 83 (1985) (O'Connor, J., concurring in judgment). 154 See id. 155 See id. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW demands of the Free Exercise Clause. 156 Under the preceding approach, legislatures have broad latitude to protect religious institutions by statutory exemptions from any burdens that the legalization of same-sex marriage may place on their religious liberty. As long as the statutory exemption is reasonably well-tailored in scope to the government-created burden on free exercise that it is designed to lift, the exemption would appear to be immune from Establishment Clause attack. 2. Individual Religious Liberty Professor George Dent has urged attention to the adverse consequences that the legalization of same-sex marriage may have for individuals who devoutly adhere to religions that oppose same-sex marriage.'57 According to Dent, legalization threatens to undermine whatever constitutional protection such adherents may enjoy to express religiously based opposition to homosexual sex and same-sex marriage. 58 Also according to Dent, legalization may erode whatever constitutional protection such adherents currently enjoy to refuse to participate in diversity training and other activities that, contrary to their religion, teach that treating gays and lesbians less favorably than others is discriminatory and wrong. 159 Professor Dent's arguments are conjectural at best. In essence, his arguments consist of (a) reciting the details of two cases in which the religious adherents' claims of constitutional protection were vindicated and then (b) ominously speculating that the legalization of same- sex marriage might somehow cause the constitutional protection to "vanish."' 6 Neither case had anything to do with marriage of any kind, same-sex or opposite-sex. In one case, public employees sued successfully on First Amendment grounds when they received written reprimands for reading their Bibles to themselves during a mandatory diversity training session 156 See, e.g., Corp. of Presiding Bishop v. Amos, 483 U.S. 327, 334-37 (1987) (assuming "for the sake of argument" that the exemption under review-an exemption of secular nonprofit activities of religious organizations from the Title VII prohibition against employment discrimination based on religion-went further in accommodating religion than the Free Exercise Clause required, but upholding the exemption nonetheless against a charge that it constituted government endorsement of religion); Walz v. Tax Comm'n, 397 U.S. 664, 672-73 (1970) (upholding property tax exemptions for places of worship as a "reasonable and balanced attempt to guard against" the "latent dangers inherent in the imposition of property taxes," and underlining that the "limits of permissible state accommodation are by no means co-extensive with the noninterference mandated by the Free Exercise Clause"). 157 Dent, supra note 108, at 443-45. 158 Id. at 444. 159 Id. at 443-44. 160. ld. at 444. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW as a silent protest against the session's message regarding gay and lesbian coworkers. 16' In the other case, a Christian minister prevailed on First Amendment grounds when a billboard display that he had put up was removed in response to pressure from a local official who objected to the display's apparent message of condemnation of same-sex sexual relations. 62 Whatever connection Professor Dent may see between legalizing same-sex marriage and diminishing the First Amendment freedoms vindicated in the two cases under discussions, he says nothing to assist the reader in ascertaining what that connection might be. The unpersuasiveness of Professor Dent's examples does not establish that legalizing same-sex marriage would never pose a potential danger to the religious liberty of individuals who adhere to religions opposed to same-sex marriage. To the extent, however, that legalization may pose such dangers, there is good reason in the individual religious liberty context, as in the institutional religious context discussed above, to believe that such dangers are greatly diminished by federal and state free exercise protections and by existing and probable statutory exemptions.'63 Professor Dent says nothing to the contrary that might lend support to his belief that the legalization of same-sex marriage sounds the death knell for religious liberty. V. Three Good Reasons for Legalizing Same-Sex Marriage To determine the role that religion has played in legislators' decisions to prohibit same- sex marriage, it is necessary to examine, as I did in Part IV, the weightiness of the nonreligious reasons that lawmakers have offered or would be apt to offer in support of prohibition. It is also necessary to examine, as I do below, the weightiness of the reasons for legalizing same- sex marriage that lawmakers presumably considered and found less weighty than the reasons for prohibition. Even if I did no more than reflect upon my own experience, I could easily name numerous benefits of marriage and discuss them at length. For present purposes, however, it seems more than adequate simply to call attention to a few commonly enjoyed benefits of obvious substantiality. As I explain below, taken together, those benefits are sufficiently weighty to make a decision to prohibit same-sex marriage virtually incomprehensible in terms of the reasons discussed in Part IV 161 Altman v. Minn. Dep't of Corrs., 251 F.3d 1199 (8th Cir. 2001). If not withdrawn, the written reprimands would have rendered the plaintiffs ineligible for promotion for two years. Id. at 1202. 162 See Dent, supra note 108, at 444 (recounting case from news article). 163 In the individual liberty context, I make no mention of freedom of expressive association, which is an organizational, not individual, right. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW A. Love and Marriage When Frank Sinatra recorded the Sammy Cahn/Jimmy Van Heusen song "Love and Marriage" in 1955, " it sped to the top of the music charts and became a major hit.'65 "Love and marriage,' according to the song, "go together like a horse and carriage." Although the popularity of this first recording of the song no doubt stemmed in part from Sinatra's inimitable style and Van Heusen's catchy tune, it also was a tribute to how well Cahn's lyrics managed to capture a widely shared sentiment. Though not the most romantic or cerebral analogy that might be drawn, the analogy of love and marriage to a horse and carriage did succeed in communicating simply and succinctly the natural connection between love and marriage that people have long perceived. To be sure, love can and does exist without marriage. Nevertheless, for many couples today and in years past, marriage is the natural culmination and celebration of their love. By the same token, marriage can and does exist without love, but for many, marriage serves, and has served, both as a constant reminder of the love that led them to wed and as a nurturing force that helps sustain love and facilitate its growth. Although some opponents of same-sex marriage may believe otherwise,"6 there is no good reason to think that love between a same-sex couple is inherently any less strong or meaningful than love between an opposite-sex couple. Nor is there any good reason to think that the relationship between love and marriage for same-sex couples is any less natural than for opposite-sex couples.167 In thinking about the potential benefits of authorizing same-sex marriage, a reasonable legislator therefore would recognize that same-sex couples, like opposite-sex couples through the ages, would reap important emotional and other personal rewards from being allowed to link their love to marriage. He or she would count as a substantial benefit of authorizing same- sex marriage that it enabled same-sex couples to enjoy these rewards. 164 FRANK SINATRA, Love and Marriage, on THIS IS SINATRA! (Capitol Records 1956). (THIS IS SINATRA! is a compilation of singles that Sinatra had previously released, including "Love and Marriage" in 1955.) 165 See Love and Marriage, http://en.wikipedia.org/wiki/Love-andMarriage. 166 See, e.g., Dent, supra note 83, at 251-52, 258-59 (maintaining that same-sex marriages "tend to be short-lived," "few gays will marry" if given the opportunity, and "gay couples that do marry will tend to have a lower level of commitment than traditional couples"). 167 A wealth of anecdotal support is supplied by the news coverage of the first day on which same-sex couples could legally marry in New York. See, e.g., Michael Barbaro, With Wait Over Gay Couples Wed Across New York, N.Y. TIMES, July 25,2011, at Al; Elissa Gootman, A Busy Day for Nuptials Across New York, Beginning at the Stroke of Midnight, N.Y TIMES, July 25, 2011, at Al 9. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW B. Marriage as a Stabilizing Force One need only consult divorce statistics in the United States, particularly since 1970, to recognize that marriage is hardly a guarantee that a relationship will last. 68 By the same token, examples abound of couples who, as a matter of personal choice (for opposite-sex couples) or legal constraints (for same-sex couples), never married but spent their lives together. Nonetheless, there is much to support the view, expressed by the American Psychiatric Association in 2005 in a position statement in favor of legalizing same-sex marriage, that marriage "provides a stabilizing force."' 69 At the start of its position statement, the Association underlined that greater stability in a couple's relationship entails substantial rewards for the couple and for any children that they may have: As physicians who frequently evaluate the impact of social and family relations on child development, and the ability of adults and children to cope with stress and mental illness, psychiatrists note the invariably positive influence of a stable, adult partnership on the health of all family members. Sustained and committed marital and family relationships are cornerstones of our social support network as we face life's challenges, including illness and loss. There is ample evidence that long-term spousal and family support enhances physical and mental health at all stages of development. 70 168 As summarized by Professors Gregory, Swisher, and Wolf several years ago: Although divorce was once relatively rare in the United States, it is now commonplace. During the decade between 1970 and 1980, the divorce rate more than doubled, and more than a million divorces currently are granted each year, a figure that has remained fairly constant throughout the 1980s and I 990s. An estimated one-half or more of American marriages will end in divorce. The average duration of marriage was seven years in 1987, and remarriage of divorced persons accounts for nearly one-half of the marriages in this country. The continuing high divorce rate is one of the most dramatic alterations in contemporary American family life. GREGORY, SWISHER & WOLF, supra note 102, at 238-39. 169 American Psychiatric Association, supra note 89. 170 Id. Along similar lines, see Sabrina Tavemise, More Unwed Parents Live Together Report Finds, N.Y TIMES, Aug. 17, 2011, at A 13 (describing recent report indicating that parents who live together without marrying are "more than twice as likely to break up" as married parents and that children with cohabiting parents tend to have "less stable" lives, "perform worse in school," and "be less psychologically healthy" than children with married parents). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW In considering the benefits of legalizing same-sex marriage, lawmakers only sensibly would count as a substantial plus that legalization would bring greater stability to same-sex couples' relationships, which in turn would have positive consequences for the couples and their children. Though a relatively small proportion of society, same-sex couples and their children are numerous enough that their physical and mental health are matters of significant concern not only for them but also for society as a whole. Bolstering their physical and mental health is advantageous to society in a variety of ways. Perhaps most notably, it helps ensure that a significant segment of the citizenry is more productive and less likely to constitute a drain on state services and financial resources. Lawmakers pondering the benefits of legalizing same- sex marriage would recognize the importance of such societal rewards. C. Equal Respect, and Avoidance of Dignitary Harm We live in a society that treats, and has long treated, marriage as a preferred status. Illustrations abound in books, films, newspapers, and other elements of the popular culture. However, it is perhaps nowhere more apparent than in the plethora of federal and state laws that advantage married over unmarried couples on matters ranging from eligibility for death and disability benefits to inheritance rights to tax exemptions and more.' When government controls access to a preferred status like marriage and denies access to a particular group, the message communicated, intentionally or not, to those in the excluded group is almost inevitably one of disrespect and unequal worth. Laws excluding same-sex couples from the opportunity to marry enjoyed by opposite-sex couples strike directly at gays' and lesbians' sense of dignity and self-worth. The exclusionary laws tell them that they are second-class citizens in the eyes of the law. 171 In a letter early in 2004 to Bill Frist, the Senate Majority Leader at the time, the U.S. General Accounting Office reported thai, as of the end of 2003, it had been able to identify "a total of 1,138 federal statutory provisions classified to the United States Code in which marital status is a factor in determining or receiving benefits, rights, and privileges." Letter from Dayna K. Shah, Associate General Counsel, U.S. General Accounting Office, to Bill Frist, U.S. Senate Majority Leader, at 1 (Jan. 23, 2004), http://www.gao.gov/new.items/d04353r.pdf. Several appendices were attached to the letter, including one that summarized the nature of the laws falling within each of the thirteen .categories into which the Office had divided the laws. See id. at 16-18. For.a sampling of the array of state laws making marital status material, see Tara Siegel Bernard, How Gay Marriage Will Change Couples 'Financial Lives, June 27, 2011, http://bucks.blogs.nytimes.com/2011/06/24/how-gay-marriage-will-change-couples-financial- lives/. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW To borrow language from the Court's discussion in Brown v. Board of Education'7 2 of the impact on black schoolchildren of another type of exclusionary, unequal treatment: the denial of equal access to so central a societal institution as marriage "generates" in gays and lesbians "a feeling of inferiority as to their status in the community."'73 In addition, it tends to alienate them in a way that exacts intangible costs on society as a whole. 74 In assessing the benefits of giving same-sex couples the same access to marriage as opposite-sex couples have long enjoyed, a reasonable legislator would give substantial weight to the value to gays and lesbians and to society as a whole of avoiding the dignitary harm and alienation that unequal access inflicts. VI. Prohibitions on Same-Sex Marriage as Government Endorsements of Religion If, as I have argued in Parts IV and V, the reasons for legalizing same-sex marriage greatly outweigh the reasons for not doing so, how is one to make sense of the numerous state legislative prohibitions on same-sex marriage? The answer seems to lie in three unconstitutional reasons for prohibiting same-sex marriage that lawmakers almost never explicitly acknowledge as influencing their decision to vote for prohibition. In some instances, lawmakers may avoid articulating those reasons because they recognize that doing so would invite judicial invalidation of the prohibition. In other instances, they may avoid doing so because they have little, if any, conscious awareness of the influence that the reasons are having on their decisionmaking. Either way, whatever the source of lawmakers' silence about those reasons, their decisions to prohibit same-sex marriage become much more understandable if those 172 347 U.S. 483 (1954). 173 Id. at 494. A recent interview with Heather Purser, a twenty-eight-year-old woman who is a lesbian and a member of the Suquamish Tribe, offers interesting anecdotal support for the proposition stated in the accompanying text. At Ms. Purser's urging at the tribe's annual meeting, the tribal members in attendance voted to approve same- sex marriage. Subsequently, the tribal council did likewise, and the policy became tribal law. According to a news account of an interview with Ms. Purser: "Throughout my life I've not always felt a part of the tribe, even though I'm so involved in it," said Ms. Purser, who is in a relationship but does not plan to marry soon. "This acceptance of gay marriage, when the tribe said, 'O.K. we're behind you,' it took away so much of that old pain, and I really feel like I'm so much a part of the community again ...." William Yardley, A Washington State Indian Tribe Approves Same-Sex Marriage, N.Y. TIMES, Aug. 12, 2011, at A12. 174 Cf Simson, supra note 16, at 467 (discussing the alienating effect that government endorsement of a particular religion (or of religion over nonreligion) is apt to have on persons not adhering to the favored belief system, and observing that because "government relies for its effectiveness on broad-based participation and support," such an alienating effect "can be a burden for both nonadherents and their government"). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW reasons are taken into account. 175 First, lawmakers may be prohibiting same-sex marriage in order to try to deter, or minimize the frequency of, a type of conduct - sexual relations between individuals of the same sex - that the lawmakers regard as unhealthy, immoral, or inimical to the general welfare in some other way. Second, lawmakers may be prohibiting same-sex marriage in order to disadvantage two groups, gays and lesbians, whom they deeply dislike. If lawmakers indeed have a deep-seated prejudice or animus toward gays and lesbians, they would count as a reason in favor of prohibiting same-sex marriage that doing so disadvantages and inflicts harm on gays and lesbians. Third, lawmakers may be prohibiting same-sex marriage in order to endorse religion. From this perspective, the lawmakers are denying gays and lesbians the opportunity to marry enjoyed by others in order to codify into law certain religious beliefs that the lawmakers or their constituents, or both, hold true. As discussed in Part IV.A, the first of the above reasons violates the Due Process Clause as interpreted in Lawrence v. Texas,'76 and the second violates the Equal Protection Clause as interpreted in Romer v. Evans. 77 As discussed in Part I, the third reason violates the Establishment Clause under the long line of precedent going back at least as far as the landmark school prayer decision of 1963.178 Recognizing that one or more of these three impermissible reasons must have played an important role in a lawmaker's decision to prohibit same-sex marriage is crucial to making sense of the legislative decision to opt for prohibition. Otherwise, the only plausible explanation for why a legislature would opt for prohibition when the reasons against prohibition so clearly outweigh the constitutionally permissible reasons for prohibition, would be that the legislature is incapable of, or indifferent to, reasoned analysis. Even assuming, however, that a "reasonable observer"--the relevant frame of reference in 175 Commentators have urged in other constitutional contexts that the courts need to take into account the role that unconscious biases may play in legislative decisionmaking. See, e.g., John Hart Ely, The Wages of Crying Wolf A Comment on Roe v. Wade, 82 YALE. L.J. 920, 933 & n.85 (1973) (discussing the importance of unconscious bias to the equal protection doctrine of suspect classifications); Charles R. Lawrence III, The Id, the Ego, and Equal Protection: Reckoning with Unconscious Racism, 39 STAN. L. REV. 317 (1987) (discussing the importance of unconscious racial bias to the equal protection doctrine of discriminatory purpose). 176 See supra text following note 70. 177 See supra text accompanying notes 69 & 70. 178 See supra notes 13-15 and accompanying text. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW the Court's endorsement test'--would recognize that one or more of the three impermissible reasons must have played an important role in the legislature's decision to prohibit same- sex marriage, it remains unclear whether the reasonable observer would conclude that a prohibition of same-sex marriage has an unconstitutional effect of endorsing religion. The reasonable observer would only reach that conclusion if, as I now argue, he or she would also recognize that, of the three impermissible reasons that may have figured in the lawmaker's thinking in banning same-sex marriage, the impermissible reason of endorsing religion played the dominant role. A reasonable observer would make this further recognition because he or she would regard the endorsement-of-religion reason as substantially more consistent than either of the two other impermissible reasons with the general mindset of the people's elected representatives. I suggest that a reasonable observer most likely and reasonably proceeds on the assumption that lawmakers, though plainly not without flaws, are basically decent, well-meaning individuals mindful of their constitutional responsibilities. If so, then the less a particular impermissible reason is consistent with that perception of lawmakers, the less readily a reasonable observer will conclude that the lawmakers are motivated by that reason. Each of the three impermissible reasons that I have identified essentially charges lawmakers with disrespecting constitutional bounds. Nonetheless, the reasons are not equally problematic in terms of degree of inconsistency with a reasonable observer's perception of the people's elected representatives. Most obviously, to assume that legislators oppose same- sex marriage for the reason that they regard disadvantaging gays and lesbians as a good thing in and of itself is deeply at odds with the perception of lawmakers that I believe a reasonable observer would hold. To assume that lawmakers are acting out of such animus is tantamount to assuming that they are fundamentally bad people. Arguably, the reason of deterring, or minimizing the frequency of, sexual relations between persons of the same sex is somewhat more compatible than the preceding reason with the posited perception of lawmakers. After all, even Justice Scalia in his vehement, doomsday dissent in Romer was emphatic that he did not regard a desire to disadvantage gays 179 See supra text accompanying notes 16-19. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW and lesbians as a legitimate reason for government action.'80 In contrast, a majority of the Supreme Court in Bowers v. Hardwick in 1986 and the three dissenting Justices in Lawrence in 2003 were willing to treat as legitimate the use of criminal sanctions to deter or minimize same-sex sexual relations.'' Ultimately, however, I am unpersuaded that the deterrence-based reason is materially more consistent than the animus-based reason with the posited perception of lawmakers. Both reasons are deeply troubling. If a lawmaker acts for either reason even in only one instance, there is ample justification to question whether he or she is fit to be occupying a position of public importance and trust. The deterrence-based reason treats as irrelevant the Court's holding in Lawrence that the Due Process Clause requires the government to refrain from interference in gays' and lesbians' "private sexual conduct."'' 82 This disregard for constitutional constraints is problematic not only as a statement of disrespect for the authority of the highest court in the land. Though explicitly framed as an objection to particular conduct, a reason of seeking to minimize sexual relations between persons of the same sex is rooted in a substantial lack of respect for gays and lesbians as persons. Though hardly to be applauded, the reason of endorsing religion is substantially more compatible than either of the other two reasons with a reasonable observer's likely perception of lawmakers. First of all, although government endorsement of religion is a bad thing, religion itself surely is not. The Establishment Clause, like its companion clause, the Free Exercise Clause, is rooted in respect for individual religious liberty and for the value that religion has in 180 See Romer v. Evans, 517 U.S. 620, 644 (1996) (Scalia, J., dissenting): Of course it is our moral heritage that one should not hate any human being or class of human beings. But I had thought that one could consider certain conduct reprehensible ... and could exhibit even "animus" toward such conduct. Surely that is the only sort of "animus" at issue here: moral disapproval of homosexual conduct, the same sort of moral disapproval that produced the centuries-old criminal laws that we held constitutional in Bowers.... 181 See Lawrence v. Texas, 539 U.S. 558, 586-605 (2003) (Scalia, J., joined by Rehnquist, C.J. and Thomas, J., dissenting) (maintaining that a criminal prohibition on same-sex sodomy should be upheld); Bowers v. Hardwick, 478 U.S. 186 (1986) (opinion of the Court by White, J., joined by Burger, C.J., and Powell, Rehnquist, and O'Connor, JJ.) (same). 182 Lawrence, 539 U.S. at 578. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW many people's lives.'83 To charge lawmakers with acting in a particular instance out of fidelity to their religious beliefs is to charge them with failing on that occasion to act in accordance with their constitutional responsibilities in a system that includes the Establishment Clause as part of its governing charter. 84 Unlike, however, a charge that they acted once out of animus or basic disrespect toward gays and lesbians, this charge falls far short of one that they are bad people. Second, although government endorsement of religion is a bad thing, government promotion of religion is not, as long as it stops short of endorsement. Moreover, it is not always easy for lawmakers to recognize when promotion crosses over into endorsement and when it does not. On occasion, the Free Exercise Clause demands government accommodation of religion. 85 In addition, the Supreme Court has consistently held that the government's latitude to act on behalf of religious liberty is not limited to satisfying the requirements of free exercise. 86 The line between permissible, though not required, accommodation of religion and impermissible endorsement of religion can be a difficult one to draw. Reasonable Justices and scholars may often disagree as to where it should be drawn.'87 Many lawmakers lack legal 183 As Justice Goldberg once explained, the "two proscriptions are to be read together, and in light of the single end which they are designed to serve." Their "basic purpose" is "to promote and assure the fullest possible scope of religious liberty and tolerance for all and to nurture the conditions which secure the best hope of attainment of that end." Abington Sch. Dist. v. Schempp, 374 U.S. 203, 305 (1963) (Goldberg, J., concurring). To similar effect, see id. at 227 (Douglas, J., concurring) ("While the Free Exercise Clause of the First Amendment is written in terms of what the State may not require of the individual, the Establishment Clause, serving the same goal of individual religious freedom, is written in different terms."); Gail Merel, The Protection of Individual Choice: A Consistent Understanding of Religion Under the First Ainendinent, 45 U. CHI. L. REV. 805, 810-11 (1978): If there is any single unifying principle underlying the two religion clauses, therefore, it is that individual choice in matters of religion should remain free .... The establishment clause serves to protect individual choices from the pressures of an official viewpoint; the free exercise clause proscribes the more direct constraints of interference with particular beliefs.. 184 See Gary J. Simson, Mired in the Confirination Mess, 143 U. PA. L. REV. 1035, 1060-61 (1995). 185 See, e.g., Wisconsin v. Yoder, 406 U.S. 205 (1972); Sherbert v. Verner, 374 U.S. 398 (1963). 186 See, e.g., Emp't Div. v. Smith, 494 U.S. 872, 890 (1990); Corp. of Presiding Bishop v. Amos, 483 U.S. 327, 334-35 (1987). 187 For one Justice's very thoughtful discussion of the problem, see Wallace v. Jaffiee, 472 U.S. 38, 81-84 (1985) (O'Connor, J., concurring in judgment). For cases generating substantial debate among the Justices on the topic, see, for example, Bd. of Educ. of Kiryas Joel Vill. Sch. Dist. v. Grumet, 512 U.S. 687 (1994); Bd. of Educ. of Westside Cmty. Sch. v. Mergens, 496 U.S. 226 (1990); Texas Monthly, Inc. v. Bullock, 489 U.S. 1 (1989). For a range of scholarly views, see CHOPER, supra note 135, at 97-140; CONKLE, supra note 135, at 139-54; TRIBE, supra note 47, at 1166-79. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW training, and even those who have such training rarely have a sophisticated understanding of the Establishment and Free Exercise Clauses. All of this does not excuse legislators' failure to act in confonnity with the Establishment Clause and its prohibition on religious endorsements. However well-intentioned their efforts, their endorsement of religious beliefs cannot stand, even when those religious beliefs have as broad support across religions as does the belief that marriage between same-sex partners is wrong. Endorsements of religion commonly cause, directly or indirectly, a variety of constitutionally significant harms, ranging from alienation of nonadherents to discrimination against nonadherents to degradation of the ostensibly favored religion.88 Nonetheless, the fact remains that legislators' resistance to same-sex marriage for religious reasons-whether such reasons reflect their personal religious beliefs or instead religious beliefs widely shared among their constituents-lacks the virulence of legislative resistance based on animus toward gays and lesbians. It also does not entail the level of disrespect toward gays and lesbians as persons that is implicit in resistance based on a desire to help minimize same-sex sexual relations. Accordingly, if, as I have posited, a reasonable observer is apt to regard the people's elected representatives as essentially honorable, public-spirited persons, he or she would find it much easier to understand legislative opposition to same-sex marriage in terms of the impermissible reason of endorsing religion rather than either or both of the other two impermissible reasons discussed. So understood, legislative opposition to same-sex marriage communicates a substantial preference on the part of the state that people abide by the religious belief that same-sex marriage is wrong-a communication that epitomizes an unconstitutional effect of endorsing religion. VII. Possible Objections to a Holding of Unconstitutionality At least three objections may fairly be raised to my conclusion that laws prohibiting same- sex marriage have an unconstitutional effect of endorsing religion. Two of the objections arise from ambiguities in applying the endorsement test, and the third derives from uncertainty as to whether that test will persist much longer. Ultimately, as discussed below, I believe that, though not trivial, the three objections do not insulate prohibitions on same-sex marriage from successful constitutional attack. My conclusion of unconstitutionality under the Establishment Clause stands. 188 See Simson, supra note 16, at 464-68. 184 23.1 COLUMBIA JOURNAL OF GENDER AND LAW A. The Uncertain Perspective of a Reasonable Observer One possible objection relates to ambiguity in applying in this context the reasonable observer concept central to the endorsement test. In brief, according to my analysis, when a reasonable observer considers the reasons for and against same-sex marriage that legislators legitimately may have taken into account in enacting a prohibition, he or she would conclude that the reasons for prohibition are not weighty and pale in substantiality alongside the reasons against. Then, my analysis continues, seeking an explanation for the legislators' reaching a result seriously at odds with the weight of the legitimate reasons for and against, a reasonable observer would further conclude that the legislators must have given determinative weight to an unconstitutional reason of endorsing religion. It may be objected, however, that my analysis takes too idealistic a view of legislators' abilities and natural tendencies. Would a reasonable observer really assume that lawmakers thought about the reasons for and against same-sex marriage with a fair amount of care? Would a reasonable observer instead take a more cynical view of the legislative process? Even if a reason for prohibition, such as encouraging procreation, is revealed upon close examination to be entitled to little weight, is it unrealistic to assume that the lawmakers closely examined it? Would it be more appropriate to assume that the lawmakers thought about this and other reasons in a rather superficial and highly unsystematic way? An enactment or committee report often includes a statement of purposes, and a bill's sponsors frequently articulate one or more purposes in the course of legislative debate. Would a reasonable observer simply take at face value such identification of supporting reasons without regard to whether the reasons withstand careful scrutiny? Is a reasonable observer being unrealistic to assume that the legislative process operates with a significant degree of logic and thoughtful deliberation? I do not think questions of this sort are at all frivolous. In part, they reflect some ambiguity in the Supreme Court's concept of a reasonable observer. It is unclear how hard-headed and untinged by idealism a reasonable observer is expected to be. In addition, the questions reflect some ambiguity in our understanding of how legislatures actually work. Even assuming that a reasonable observer looks at the legislative process with a thoroughly detached, utterly realistic, eye, it is difficult to say with great specificity what the reasonable observer would see. This is so particularly because legislative processes vary substantially as a result of differences in structure, personnel, tradition, and more. Despite these uncertainties, I submit that the reasons against prohibiting same-sex marriage so clearly outweigh the legitimate reasons for prohibition that even a rather cynical observer of the legislative process would conclude that the impermissible purpose of endorsing religion must have played a crucial role in adopting any prohibition. To argue that a reasonable 23.1 185 COLUMBIA JOURNAL OF GENDER AND LAW observer could conclude otherwise, one would need to assume that the legislative process is operating so differently than one would hope in a representative democracy that rational discussion and debate have little, if any, meaningful role. B. A Problem of Timing A second objection to my conclusion that laws prohibiting same-sex marriage have an unconstitutional effect of endorsing religion distinguishes between prohibitions adopted before and after Lawrence. An important ingredient of my analysis is the assumption, rooted firmly in the language and holding of the majority opinion in Lawrence, that lawmakers cannot count as a legitimate reason for prohibiting same-sex marriage that such a prohibition would deter or help minimize sexual relations between persons of the same sex. 89 However, prior to deciding Lawrence in 2003, the Supreme Court had not told lawmakers that this reason was constitutionally off-limits. Indeed, quite the contrary. The Court in 1986 in Bowers v. Hardwick9 ' essentially gave lawmakers an open invitation to seek to deter or minimize same-sex sexual relations as much as they wished. The Bowers Court's holding that the Due Process Clause permits states to criminalize homosexual sodomy between consenting adults in the privacy of their home left little doubt that the Court took a very expansive view of the scope of state authority to suppress same-sex sexual relations. Moreover, to the extent that the holding alone left any doubt on this score, it was surely dispelled by the language of the opinion, particularly such stunningly insensitive remarks as "[T]o claim that a right to engage in such conduct is 'deeply rooted in the Nation's history and tradition' or 'implicit in the concept of ordered liberty' is, at best, facetious."'' 91 Under the circumstances, should prohibitions adopted prior to Lawrence and its explicit overruling of Bowers'92 be viewed more indulgently, in terms of a possible violation of the endorsement test, than prohibitions adopted after? I believe that the answer should be "no." As the Supreme Court has indicated on a number of occasions, any inquiry into the purposes underlying a law cannot always sensibly be limited to the purposes that prompted adoption of the law. In a system in which laws enacted by legislatures are not carved in stone, but rather are open to amendment and repeal, the relevant focus must be dynamic - in other words, not 189 See supra text following note 70. 190 478 U.S. 186 (1986). 191 Id. at 194 (quoting phrases the Court had used in earlier cases to describe interests that qualify as "fundamental" and that therefore trigger exacting judicial review). 192 Lawrence, 539 U.S. at 578. 23.1186 COLUMBIA JOURNAL OF GENDER AND LAW why a law was adopted, but rather why it remains on the books today. 93 To be sure, if a law's text or legislative history reveals that the legislature regarded certain purposes as important to its adoption, it is entirely appropriate to presume that those purposes remain important to the law's continued existence. However, particularly where circumstances have changed in a way that reasonably would be expected to make lawmakers think seriously about the wisdom of retaining the law, the inquiry into underlying purposes must take into account any differences in the purposes behind legislative adoption and retention of the law. For an illustration of changed circumstances of this sort, the same-sex marriage context is hard to beat. Legislators would have to be unreflective in the extreme not to be moved by Lawrence to think seriously about whether to retain a preexisting prohibition on same- sex marriage and, if so, why. To be sure, legislators may reasonably regard some Supreme Court decisions as rather forgettable-of interest, if at all, only to judges and scholars. As was immediately apparent, however, Lawrence is not arguably such a decision. 94 193 For example, in rejecting an Establishment Clause attack on Maryland's Sunday closing laws in McGowan v. Maryland, 366 U.S. 420 (1961), the Court explained: The present purpose and effect of most [states' Sunday closing laws] is to provide a uniform day of rest for all citizens; the fact that this day is Sunday, a day of particular significance for the dominant Christian sects, does not bar the State from achieving its secular goals. To say that the States cannot prescribe Sunday as a day of rest for these purposes solely because centuries ago such laws had their genesis in religion would give a constitutional interpretation of hostility to the public welfare rather than one of mere separation of church and State. Id. at 445. To similar effect, see Justice Brennan's concession in Lynch v. Donnelly, 465 U.S. 668 (1984)--itself a religious display case-that the Establishment Clause is compatible with the government's making Thanksgiving a national holiday: [O]ur cases recognize that while a particular governmental practice may have derived from religious motivations and retain certain religious connotations, it is nonetheless permissible for the government to pursue the practice when it is continued today solely for secular reasons.. . T]he mere fact that a governmental practice coincides to some extent with certain religious beliefs does not render it unconstitutional. Thanksgiving Day, in my view, fits easily within this principle, for despite its religious antecedents, the current practice of celebrating Thanksgiving is unquestionably secular and patriotic .... Id. at 715-16 (Brennan, J., dissenting). 194 See Linda Greenhouse, The Supreme Court: Homosexual Rights; Justices, 6-3, Legalize Gay Sexual Conduct in Sweeping Reversal of Court ' '86 Ruling, N.Y. TIMES, June 27, 2003, at AI (characterizing the Court's decision the prior day in Lawrence as "a sweeping declaration of constitutional liberty for gay men and lesbians," and observing that "few people on either side of the case expected a decision of such scope" and that "there was no doubt that the decision had profound legal and political implications"). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW Furthermore, it is fair to assume that by 2003, when Lawrence was decided, other developments already should have put lawmakers in the mindset of rethinking the justification for any existing prohibition on same-sex marriage. Two developments in particular stand out: the Hawaii Supreme Court's pathbreaking decision in 1993 in Baehr v. Lewin 95 - the first state high court decision to cast doubt on the validity of any state's prohibition; and Congress's enactment in 1996 of the Defense of Marriage Act.196 Also important, however, are various other judicial and legislative developments for and against same-sex marriage in the period between Baehr and Lawrence. 97 In addition, though not addressed to the same-sex marriage issue, the Court's opinion in Romer'98-the Court's first majority opinion evidencing significant sensitivity to discrimination against gays and lesbians-could not help but provoke some legislative reflection on the future and wisdom of same-sex marriage prohibitions. In short, in thinking about whether a reasonable observer would conclude that a prohibition on same-sex marriage is substantially based on religious purposes and, as a result, has an unconstitutional effect of endorsing religion, it is entirely realistic and appropriate to treat prohibitions adopted before and after Lawrence the same. Regardless of whether legislators adopting prohibitions before Lawrence felt constitutionally justified in doing so in order to deter or help minimize same-sex sexual relations, they or their successors in the legislature can fairly be held accountable after Lawrence for recognizing that purpose as impermissible and for deciding to retain the prohibition nonetheless. C. The Uncertain Future of the Endorsement Test When Justice O'Connor retired in 2006, the Supreme Court lost its leading proponent of 195 852 P.2d 44 (Haw. 1993). The Hawaii high court's decision, which remanded the case for a determination of whether the state's ban on same-sex marriage could survive a very demanding standard of review, see id. at 68, did not result in the ban's demise. Before the case could be decided on remand, Hawaii's voters approved a state constitutional amendment that banned same-sex marriage and rendered the case moot. 196 The Act, which is codified at I U.S.C. § 7 (2006) and 28 U.S.C. § 1738C (2006), is the subject of Part VIII infra. Another exceptionally high-visibility development came only several months after Lawrence in the fall of the same year: the Massachusetts Supreme Judicial Court's pathbrcaking decision in Goodridge v. Department of Public Health, 798 N.E.2d 941 (Mass. 2003). By striking down (under the Massachusetts Constitution) the state's statutory prohibition on same-sex marriage, the state high court made Massachusetts the first state in the United States to legalize same-sex marriage. 197 See Simson, supra note 42, at 315-17. 198 Romer v. Evans, 517 U.S. 620 (1996), discussed supra text accompanying notes 52-60,68-70. 23.1 23.1 COLUMBIA JOURNAL OF GENDER AND LAW the endorsement test. 199 In addition, with her retirement, the Court also may have lost the fifth vote needed for the test to retain majority support. As far back as the Court's 1963 decision invalidating Bible-reading and prayer in schools, 20° there has been disagreement among the Justices as to whether a showing of endorsement is sufficient to constitute an Establishment Clause violation or whether a more difficult showing-one of coercion-is required instead. Over Justice Stewart's solitary dissent,20 the Court in that 1963 decision explicitly disclaimed the need to show coercion,20 2 and the Court has yet to hold otherwise. However, it has been unclear since Justice O'Connor's retirement and replacement by Justice Alito whether, if presented with a case in which the outcome would differ depending on whether an endorsement test or coercion test were applied, a majority of the Court would abandon the endorsement test in favor of a coercion test.203 As with so many other issues, the balance of power on the Court on this issue appears to rest in Justice Kennedy's hands.2 0 4 His Establishment Clause opinions and alignments over 199 Two opinions by Justice O'Connor-Wallace v. Jafree, 472 U.S. 38, 67-84 (1985) (O'Connor, J., concurring), and Lynch v. Donnelly, 465 U.S. 668, 687-94 (1984) (O'Connor, J., concurring) -were especially influential in the evolution of the test. For discussion of Justice O'Connor and the endorsement test, see CONKLE, supra note 135, at 125-28, and Arnold H. Loewy, Rethinking Government Neutrality Towards Religion Under the Establishment Clause: The Untapped Potential ofJustice O'Connors Insight, 64 N.C. L. REv. 1049 (1986). 200 Abington Sch. Dist. v. Schempp, 374 U.S. 203 (1963). 201 See id. at 316 (Stewart, J., dissenting) ("In the absence of coercion upon those who do not wish to participate ... [the statutory] provisions [authorizing the religious exercises under review] cannot be held, in my view, to represent the type of support of religion barred by the Establishment Clause."). 202 See id. at 223 ("The distinction between the two clauses is apparent-a violation of the Free Exercise Clause is predicated on coercion while the Establishment Clause violation need not be so attended."). 203 See Simson, supra note 42, at 379-80 & n.215 (discussing the Justices' apparent alignment, after Justice Alito's appointment to the Court, with regard to retaining the endorsement test as opposed to replacing it with a coercion test). In the years since Justice Alito joined the Court, two Justices supportive of the endorsement test-Justices Souter and Stevens-have retired, and their seats have been filled by Justices Sotomayor and Kagan, respectively. If, as seems likely, Justices Sotomayor and Kagan would vote to retain the endorsement test, then their appointment has not altered the number of Justices for and against retaining the test. 204 In the years since Justice O'Connor's retirement, recognition of Justice Kennedy's role as the Court's "swing Justice" has become commonplace. See, e.g., Adam Liptak, A Significant Term, With Bigger Cases Ahead, N.Y. TIMES, June 29, 2011, at A12 ("[T]he number of 5-to-4 decisions in which the court's four liberals found themselves on one side and its four conservatives on the other was high: 12 of the 14 closely divided cases were configured that way, with Justice Anthony M. Kennedy casting the decisive vote ... Justice Kennedy, the court's swing justice, voted with its more conservative members in those dozen cases two-thirds of the time."); Mark Sherman, Justice Anthony Kennedys Influence on Supreme Court Likely to Grow, HUFFINGTON PosT, July II, 2010, http://www.huffingtonpost.com/2010/07/ I/anthony-kennedys-influenc_n_642149. html ("Justice Anthony Kennedy, who already decides whether liberals or conservatives win the Supreme Court's most closely contested cases, is about to take on an even more influential behind-the-scenes role with the retirement of Justice John Paul Stevens."). COLUMBIA JOURNAL OF GENDER AND LAW 23.1 the years suggest that he would favor replacing the endorsement test with a coercion test. However, those opinions and alignments are not unambiguous in this regard.21 5 Moreover, JusticeK nngdy's general inclination in constitutional interpretation to give substantial 205 For example, in his opinion in Allegheny County v. ACLU, 492 U.S. 573 (1989), Justice Kennedy maintained that the endorsement test "is flawed in its fundamentals and unworkable in practice," id. at 669 (Kennedy, J., concurring in judgment in part and dissenting in part), and went on to criticize it at length as inconsistent with precedent, tradition, and the need for standards inviting of reasonably objective application. See id. at 669-78. In light of the tenor of his opinion in Allegheny County, it was surprising, to say the least, to find Justice Kennedy not only voting to invalidate the graduation prayer in Lee v. Weisman, 505 U.S. 577 (1992), but also writing an opinion for the Court in Lee that avoided the question of whether the endorsement test should be abandoned and that characterized the latter question as a "difficult" one. Id. at 586. Of course, Justice Kennedy could not have hoped to get the four Justices who joined his opinion in Lee to join an opinion that jettisoned the endorsement test. All four of those Justices were sufficiently committed to the endorsement test that they wrote or joined separate opinions underlining their continued allegiance to the test. See id. at 606-09 & n. 9 (Blackmun, J., joined by Stevens and O'Connor, JJ., concurring); id. at 618-26 (Souter, J., concurring). Nonetheless, if Justicc Kennedy were as intent on abandoning the endorsement test as he appeared to be in Allegheny County, one would have expected him to take advantage in Lee of the recent change in the Court's membership. As a result of Justice Marshall's retirement and the appointment of Justice Thomas to succeed him, the Court at the time of Lee had four Justices, not including Justice Kennedy, who were plainly prepared to vote to abandon the endorsement test See id. at 641-42 (Scalia, J., joined by Rehnquist, C.J., and White and Thomas, JJ., dissenting) ("[Tlhere is simply no support for the proposition that the officially sponsored nondenominational invocation and benediction read by Rabbi Gutterman-with no one legally coerced to recite them-violated the Constitution."). Justice Kennedy could have written an opinion that announced the judgment of the Court, abandoned the endorsement test in a portion of the opinion joined by the four Justices who dissented from the judgment, and struck down the graduation prayer with the supporting votes of the four Justices who refused to abandon the endorsement test. (Compare Justice Powell's famous lead opinion announcing the judgment in the Court's first major affirmative action case, Regents of the University of California v. Bakke, 438 U.S. 265 (1978).) Subsequently, and further muddying the waters as far as Justice Kennedy's likely intentions toward the endorsement test, he joined without comment Justice Stevens's opinion for the Court in Santa Fe Independent School District v. Doe, 530 U.S. 290 (2000), striking down a Santa Fe high school's football game prayer practice. Part Ill of the Court's opinion (id. at 310-13) is devoted to a discussion of the test that Justice Kennedy had applied in Lee. However, the focus of Part 11 (id. at 301-10) is a question of endorsement, and it is difficult to understand Kennedy's willingness to join this part of Justice Stevens's opinion as anything other than a significant softening of the hard line that he took toward the endorsement test roughly a decade earlier in Allegheny County Justice Kennedy's later votes to uphold the Ten Commandments displays in McCreary County v. ACLU, 545 U.S. 844 (2005), and Van Orden v. Perry, 545 U.S. 677 (2005), make clear that if Justice Kennedy is indeed content to retain the endorsement test, he is disinclined to apply it as strongly as some members of the Court would apply it. At a minimum, he seems relatively untroubled by the threat of endorsement posed by a display that is "passive," as the plurality opinion joined by Justice Kennedy in Van Orden characterized the display there (see Van Orden, 545 U.S. at 686 (opinion of Rehnquist, C.J., joined by Scalia, Kennedy, and Thomas, JJ.)), or"unremarkable-if indeed.., noticed at all," as the dissenting opinion joined by Justice Kennedy in McCreary described that display (see McCreary, 545 U.S. at 903 (Scalia, J.,joined by Rehnquist, C.J., and Kennedy and Thomas, JJ., dissenting)). In that regard, even if, as I suggest, Justice Kennedy has gravitated away from some of his most strident criticisms of the endorsement test in his Allegheny County opinion, he continues to take a view of the type of religious displays under review in Allegheny County that is much like the view that he expressed in that case. As he put it in Allegheny County, when "the government's act of recognition or accommodation [of religion] is passive and symbolic ... any intangible benefit to religion is unlikely to present a realistic risk of establishment." 492 U.S. at 662 (Kennedy, J., concurring in judgment in part and dissenting in part). COLUMBIA JOURNAL OF GENDER AND LAW weight to the fact of precedent 2 ---an inclination probably more characteristic of him than of any of his current colleagues-makes hazardous any prediction that he would actually cast a deciding vote to unsettle settled law. Let us assume, for purposes of argument, that Justice Kennedy is prepared to replace the endorsement test with a coercion test and that his willingness to do so would provide a crucial fifth vote for that change. Even if both of these assumptions are true, there is good reason to believe that, in deciding whether a prohibition on same-sex marriage violates the Establishment Clause, Justice Kennedy, like the four Justices apparently favoring an endorsement test, would vote to strike down the prohibition. Of paramount importance in this regard is the nature of the coercion test that Justice Kennedy would be most likely to apply. In 1989, at the end of his first full Tern on the Court, Justice Kennedy in Allegheny County v. ACLU 07 wrote an opinion joined by Chief Justice Rehnquist and Justices Scalia and White that was sharply critical of the endorsement test. The opinion argued in favor of having as a "limiting principle" that "government may not coerce anyone to support or participate in any religion or its exercise." 08 Applying that principle to the two holiday displays at issue in the case-a creche inside a government building and a giant menorah on the grounds of another such building-Justice Kennedy had no hesitation finding both consistent with the Establishment Clause. In his view, the determinative factor was that "[n] o one was compelled to observe or participate in any religious ceremony or activity.""z In the course of his analysis, Justice Kennedy indicated that special deference is due "practices that are accepted in our national heritage."2t 0 Indeed, he went so far as to say that any interpretation of the Establishment Clause that "if applied with consistency, would invalidate longstanding traditions cannot be a proper reading of the Clause."2t' Three years later, Justice Kennedy in Lee v. Weisman2"' parted ways with the three 206 See, e.g., United States v. Lopez, 514 U.S. 549, 574 (1995) (Kennedy, J., concurring) (maintaining that stare decisis "operates with great force" to foreclose a very narrow reading of the Commerce Clause); Planned Parenthood v. Casey, 505 U.S. 833, 854-69 (1992) (opinion of Court by O'Connor, Kennedy, and Souter, JJ.) (maintaining that stare decisis militates strongly against overruling Roe v. Wade, 410 U.S. 113 (1973)). 207 492 U.S. 573 (1989). 208 M. at 659 (opinion of Kennedy, J., concurring in thejudgment in part and dissenting in part). 209 Id. at 664. 210 Id. at 663. 211 Id. at 670. 212 505 U.S. 577 (1992). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW Justices who had joined his opinion in Allegheny County. Lee involved an Establishment Clause challenge to a public secondary school's inclusion in its graduation ceremonies of invocation and benediction prayers by a member of the clergy. Justice Kennedy wrote an opinion of the Court for himself and for four Justices who also wrote separately to underline their continued allegiance to the endorsement test. He maintained that the constitutionality of the practice at issue in Lee could be decided without reaching the question of whether a coercion test should supplant the endorsement one. Instead, the Court would apply as a constitutional "minimum"213 the coercion test that he had articulated as a limiting principle in Allegheny County. As Justice Kennedy further explained, the need to strike down the practice at hand under that test was clear. In finding coercion to participate in a religious exercise, Justice Kennedy cited psychological research that "supports the common assumption that adolescents are often susceptible to pressure from their peers towards conformity, and that the influence is strongest in matters of social convention."'214 In light of the "subtle coercive pressure ' '2 5 at work, the state could not, according to Justice Kennedy, place students who might object to the invocation or benediction "in the dilemma of participating, with all that implies, or protesting. '21 6 In essence, the state through its public schools was "exact[ing] religious conformity from a student as the price of attending her own high school graduation," and this "[t]he Constitution forbids. '217 In a dissent joined by Chief Justice Rehnquist and Justices White and Thomas, Justice Scalia strongly criticized-or, perhaps more accurately, mocked-the coercion analysis in Justice Kennedy's opinion of the Court as "incoherent" and as an example of "psychology practiced by amateurs. 2 18 In distancing himself from the coercion test applied by the Court, Justice Scalia maintained that the Court had wrongly expanded the concept of coercion far beyond its historical origins. In his view, rather than concern itself with the "precious question" of possible psychological coercion, the Court should have considered only whether the graduation prayer practice under review implicated the type of coercion that was "a hallmark of historical establishments of religion. '2 9 According to Justice Scalia, such coercion took the 213 Id. at 587. 214 Id. at 593. 215 Id. at 592. 216 Id. at 593. 217 505 U.S. at 596. 218 Id. at 636 (dissenting opinion of Scalia, J.). 219 Id. at 640. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW form of "acts backed by threat of penalty. ' 220 Justice Scalia also censured the Court's opinion as "conspicuously bereft of any reference to history. ' 221 Highlighting with obvious ire what he regarded as an unconscionable change in thinking on Justice Kennedy's part, Justice Scalia began his dissenting opinion by quoting Justice Kennedy's affirmations in Allegheny County of the importance in Establishment Clause analysis of respect for "longstanding traditions. ' 222 "With nary a mention that it is doing so," Justice Scalia charged, the Court in Lee "lays waste a tradition that is as old as public school graduation ceremonies themselves, and that is a component of an even more longstanding American tradition of nonsectarian prayer to God at public celebrations generally. 223 I do not discuss Justice Scalia's criticisms of Justice Kennedy's coercion test because I find them particularly persuasive. In fact, I find them particularly unpersuasive. I discuss them, however, because they help bring out the differences between Justice Kennedy's coercion test and the one that Justice Scalia and probably three other Justices would apply today. If, as seems likely, Justice Kennedy currently, as at the time of Lee, holds the balance of power in Establishment Clause cases, and if Justice Kennedy were to apply a coercion test- whether as a constitutional "minimum" as in Lee or out of conviction that it should supplant the endorsement test-his more expansive understanding of that test would be determinative of the result reached by the Court. As discussed below, I believe that the constitutionality of prohibitions on same-sex marriage, like the constitutionality of the graduation prayer in Lee, would be decided very differently if Justice Kennedy's coercion test, rather than Justice Scalia's, were the one applied. Before turning to the question of the constitutionality of prohibitions on same-sex marriage under the Lee coercion test, I note that a decision eight years after Lee supports treating Justice Kennedy's conception of a coercion test in Lee as more accurately reflecting his current conception of the test than the less interventionist, more tradition-conscious conception that he offered in Allegheny County after little more than a year on the Court. In Santa Fe Independent School District v. Doe,2 4 Justice Kennedy joined a majority opinion by Justice Stevens that relied heavily upon the Court's opinion in Lee in invalidating a prayer practice that, as discussed below, would appear to be at least somewhat less coercive than the 220 Id. at 642. 221 Id. at 631. 222 505 U.S. at 631. 223 Id. at 631-32. 224 530 U.S. 271 (2000). 23.1 COLUMBIA JOURNAL OF GENDER AND LAW one in Lee.225 If at the time of Santa Fe Justice Kennedy were any less committed to the Lee conception of a coercion test than at the time of Lee, he almost certainly would have voted the other way. As required by the Santa Fe school board, each public high school in the district would conduct an election among its student body prior to the varsity football season as to (a) whether a student should be selected to give "a brief invocation and/or message" during pre- game ceremonies to "solemnize the event, to promote good sportsmanship and student safety, and to establish the appropriate environment for the competition," and if so, (b) which student off of a list of volunteers should give the invocation and/or message.226 Unlike the graduation prayer in Lee, which was initiated by the school principal and delivered by a member of the clergy, the football pre-game prayer in Santa Fe therefore would transpire only if the student body voted to have a student speaker and only if the student speaker opted to give an invocation, rather than a nonreligious "message." As such, the football pre-game prayer in Santa Fe would appear to entail less governmental coercion than the graduation prayer in Lee. Also suggestive of a lesser degree of coercion than in Lee is the difference in the forum where the invocation or benediction is received. It seems highly unlikely that students sitting in the bleachers while awaiting the start of a football game feel as much pressure to join in (or at least listen with respectful silence to) a prayer offered over the loud speaker as students sitting in an auditorium during their graduation with familial and other eyes upon them. Under the circumstances, Justice Kennedy's willingness to join an opinion striking down the football pre-game prayer strongly suggests that, at the time of Santa Fe, he conceptualized a coercion test at least as broadly as he had done in Lee. Prohibitions on same-sex marriage present a rather different kind of problem under Justice Kennedy's coercion test than the prayer practices in Lee and Santa Fe. On the one hand, while the coerciveness of the prayer practices in the two cases is open to debate, prohibitions on same-sex marriage are plainly coercive. By such prohibitions, the state is coercing same- sex couples to live their lives in accordance with a conception of marriage--one that treats marriage as properly limited to opposite-sex couples-that they may wholeheartedly reject. 225 Although the Court's opinion in Santa Fe contains discussions of application of the endorsement test without counterpart in the Court's opinion in Lee, see supra text accompanying notes 212 & 213, it also strongly affirms, both at the outset of its analysis and in Part III of the analysis, the importance of Lee. See Santa Fe, 530 U.S. at 301-02 ("Although this case involves student prayer at a different type of school function, our analysis is properly guided by the principles we endorsed in Lee."); id. at 310-13 (addressing and rejecting the school district's argument that the prayer practice under review is "distinguishable from the graduation prayer in Lee because it does not coerce students to participate in religious observance"). 226 Santa Fe, 530 U.S. at 298 n.6. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW On the other hand, while the prayer practices in Lee and Santa Fe are plainly religious, the religious nature of the conception of marriage codified by prohibitions on same-sex marriage is open to debate. Despite these differences between prohibitions on same-sex marriage and the prayer practices in Lee and Santa Fe, the Court's invalidation of the practices in the two cases strongly supports invalidation of same-sex marriage bans as well. First, in applying his coercion test in Lee, Justice Kennedy stated a principle of application with significant implications for applying the test to prohibitions on same-sex marriage: "Law reaches past formalism. '2 7 Justice Kennedy in Lee explicitly relied upon that principle in rejecting an argument that a student's option to forgo attending graduation negates any possible concern that he or she is being coerced to participate in the clergy-offered graduation prayer. As Justice Kennedy explained, "[T]o say that a teenage student has a real choice to attend her high school graduation is formalistic in the extreme." '228 Furthermore, although Justice Kennedy did not explicitly invoke that principle elsewhere in his opinion, it implicitly played a central role throughout. In discussing coercion in Lee with attention to relevant psychological research and the pressures to conform that students commonly feel, Justice Kennedy was similarly "reach[ing] past formalism." Tacitly, he was rejecting Justice Scalia's insistence on limiting coercion to instances of "acts backed by threat of penalty" 22 9 as "formalistic in the extreme." Formally, prohibitions on same-sex marriage are religion-neutral. They generally contain no explicit invocation of religion in limiting marriage to opposite-sex couples. The reality, however, is very different. As discussed in detail in Parts IV-VI above, prohibitions on same- sex marriage are most reasonably understood as endorsements of a religious tenet. To take them at face value and ignore their religious roots is to become mired in formalism, not to 227 Lee, 505 U.S. at 595. 228 Id. 229 Id. at 642 (Scalia, J., dissenting). 23.1 196 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 reach past it as Justice Kennedy's application of his coercion test strives to do.230 Justice Kennedy's application of his test in Lee militates for yet another reason in favor of finding that laws prohibiting same-sex marriage entail impermissible religious coercion. Although he does not directly respond to Justice Scalia's claim that longstanding traditions should be treated as essentially immune to Establishment Clause attack, he tacitly rejects it. If Justice Scalia's equation of longevity with constitutionality had carried the day in Lee, a finding that prohibitions on same-sex marriage entail impermissible religious coercion would be inconceivable regardless of how much logic might be mustered in support. After all, when it comes to longstanding traditions, the tradition of limiting marriage to opposite-sex couples makes the tradition in the United States of graduation prayer seem still in infancy. Interestingly, in parting ways implicitly with Justice Scalia (as well as with his own 230 An interesting question is whether, under Justice Kennedy's coercion test, the religious nature of governmental action should be judged from the perspective of the reasonable observer or instead from the perspective of those who are experiencing the coercion. On the one hand, it may be most in keeping with Justice Kennedy's focus in gauging coercion to gauge the religious nature of governmental action from the perspective of those experiencing the coercion. In deciding whether the students are coerced in Lee, Justice Kennedy appears to be trying to see things from the students' point of view as he delves into the psychological realities and takes cognizance of the "subtle coercive pressure" felt by students. Lee, 505 U.S. at 592. On the other hand, it would be most consistent with the Court's application of the endorsement test to gauge the religious nature of governmental action from the perspective of a reasonable observer. Obviously, the endorsement and coercion tests differ in fundamental ways. However, it is at least arguable that a frame of reference-the reasonable observer-that is appropriate for one test rooted in the Establishment Clause is also appropriate for another. For purposes of discussing the application of Justice Kennedy's coercion test to prohibitions on same- sex marriage, I need not resolve this ambiguity in the test. In my analysis in the text, I have assumed that the relevant perspective in gauging the religious nature of such prohibitions is that of a reasonable observer, and I have maintained that such prohibitions are religious from that perspective. If I had instead assumed that the relevant perspective is that of those experiencing the coercion-i.e., same-sex couples wishing to many but precluded from doing so--a conclusion that prohibitions on same-sex marriage entail religious coercion would follow even more easily. Presumably, same-sex couples experiencing this disadvantage would be especially sensitive to (a) the strength of the reasons for allowing same-sex marriage and the weakness of the permissible reasons for prohibiting it and (b) the likelihood that the impermissible reason of endorsing religion played a crucial role in the decision to adopt a prohibition. For criticism of the Court's focus on the reasonable observer in applying the endorsement test, see Simson, supra note 14, at 916 (maintaining that a focus on "the reasonable perception of persons who would feel pressured and alienated" by the state action would be logical in terms of helping to ensure that people are not made to endure the harms inflicted by government endorsement of religion). COLUMBIA JOURNAL OF GENDER AND LAW opinion three years earlier in Allegheny County231) on the capacity of tradition to insulate a practice from constitutional challenge, Justice Kennedy invoked "tradition" in a very different way: The lessons of the First Amendment are as urgent in the modem world as in the 18th century when it was written. One timeless lesson is that if citizens are subjected to state-sponsored religious exercises, the State disavows its own duty to guard and respect that sphere of inviolable conscience and belief which is the mark of a free people. To compromise that principle today would be to deny our own tradition and forfeit our standing to urge others to secure the protections of that tradition for themselves.232 If the relevant tradition in gauging the constitutionality of prohibitions on same-sex marriage is not the longstanding tradition of opposite-sex marriage but rather the constitutional tradition that Justice Kennedy discusses above, "tradition" is entirely consistent with a determination that those prohibitions entail impermissible religious coercion. VHI. A Brief Aside: The Indefensibility of the Defense of Marriage Act The Establishment Clause analysis in this Article has important implications for the federal Defense of Marriage Act even though that law is not a prohibition of same-sex marriage. DOMA contains two key provisions. One purports to relieve a state that does not authorize same-sex marriage from any obligation under the Constitution's Full Faith and 231 SeeAllegheny Cnty. v. ACLU, 492 U.S. 573,669-74 (1989) (Kennedy, J., concurring in judgment in part and dissenting in part) (maintaining that, to be viable, an Establishment Clause test generally must not threaten to invalidate traditional practices). 232 Lee, 505 U.S. at 592. 23.1 COLUMBIA JOURNAL OF GENDER AND LAW 23.1 Credit Clause233 to treat as valid a same-sex marriage performed in a state that permits such marriages.2 4 The other provision declares that any reference to "marriage" or "spouse" in any federal law should be understood as referring only to opposite-sex marriages and spouses. 235 Although the interstate recognition provision leaves each state free to decide whether or not to recognize a same-sex marriage lawfully performed in another state, that provision is not 233 Article IV, Section 1 of the Constitution provides: Full Faith and Credit shall be given in each State to the public Acts, Records, and judicial Proceedings of every other State. And the Congress may by general Laws prescribe the Manner in which such Acts, Records and Proceedings shall be proved, and the Effect thereof. U.S. CONsT. art. IV, § 1. In 1790, exercising its authority under the second sentence ofArticle IV, Section 1, Congress enacted a statute requiring a court in one state to give a judgment rendered in another state's court the "faith and credit" that the rendering court would give it. 28 U.S.C. § 1738 (2006). In contrast, Congress was silent until 1948 as to the "faith and credit" that a court in one state must accord to the laws of another state. In a number of cases prior to 1948, the Court made clear that it did not regard congressional action under the second sentence of Article IV, Section 1 as essential for the first sentence to act as a constraint on courts' decisions whether to apply the laws of another state rather than their own. See, e.g., Pac. Emp'rs Ins. Co. v. Indus. Accident Comm'n, 306 U.S. 493, 502 (1939) (rejecting a full faith and credit challenge to a California court's choice of California, rather than Massachusetts, law, and explicitly noting, but not treating as a basis for decision, the absence of congressional action implementing the first sentence of Article IV, Section 1 with regard to choice of law); Bradford Elec. Light Co. v. Clapper, 286 U.S. 145 (1932) (invalidating a New Hampshire court's choice of New Hampshire, rather than Vermont, law as a violation of the Full Faith and Credit Clause). Since Congress's amendment in 1948 of 28 U.S.C. § 1738 to encompass sister-state laws as well as judgments, the Court has given no indication that it regards the amendment as affecting its interpretation of the clause. Questions of the effect that the courts in one state should give a marriage performed in another have long been treated as questions of the respect due to the laws of another state. See, e.g., In re May's Estate, 114 N.E.2d 4 (N.Y. 1953); RESTATEMENT OF CONFLICT OF LAWS § 132 (1934). 234 See 28 U.S.C. § 1738C (2006): No State, territory, or possession of the United States, or Indian tribe, shall be required to give effect to any public act, record, or judicial proceeding of any other State, territory, possession, or tribe respecting a relationship between persons of the same sex that is treated as a marriage under the laws of such other State, territory, possession, or tribe, or a right or claim arising from such relationship. 235 See I U.S.C. § 7 (2006): In determining the meaning of any Act of Congress, or of any ruling, regulation, or interpretation of the various administrative bureaus and agencies of the United States, the word "marriage" means only a legal union between one man and one woman as husband and wife, and the word "spouse" refers only to a person of the opposite sex who is a husband or a wife. COLUMBIA JOURNAL OF GENDER AND LAW remotely neutral on the question of the wisdom of authorizing same-sex marriage. It comes down clearly on the side opposing authorization. It is hardly an accident that although state marriage laws differ in a variety of ways, Congress has only felt the need to speak up on behalf of one state's right to refuse recognition to a marriage performed in another state when the difference between the two states' laws pertains to the legality of same-sex marriage. As with state prohibitions on same-sex marriage, when one searches for an explanation of the thinking behind DOMA's nonrecognition provision, only one reason appears to have a great deal of explanatory value: religion. Congress's singling out same-sex marriages for disadvantage with that provision is difficult, if not impossible, to understand except as an effort to endorse the religious view that same-sex sexual relations are sinful and that government should not be encouraging or condoning such relations by allowing same-sex couples to marry. Congress's attentiveness to little, if anything, other than endorsing religion when it enacted the nonrecognition provision is evidenced by the fact that the provision accomplishes virtually nothing other than endorsing religion. Remarkably, the provision relieves states of a full faith and credit obligation that simply does not exist. Unless the Supreme Court decides to revise dramatically its longstanding interpretation of a state's obligation under the Full Faith and Credit Clause to decide choice-of-law questions by another state's law, a court has virtually unfettered freedom to apply forum law and refuse to treat as valid a same-sex marriage lawful under the law of the state where the marriage was formed. Full faith and credit limitations on state choice of law are not entirely nonexistent, but they are so close to being so that the right of interstate nonrecognition announced in DOMA is essentially beside the point.236 Unlike the nonrecognition provision, DOMA's definitional provision packs a powerful punch. 237 Regardless of a married same-sex couple's rights under the law of the state in which they were lawhfily wed and regardless of whether they may have lived in that state for their 236 See Gene R. Shreve, Choice of Law and the Fotgiving Constitution, 71 IND. L.J. 271, 274-76 (1996); Gary J. Simson, The Choice-of-Law Revolution in the United States: Notes on Rereading von Mehren, 36 CORNELL INT'L L.J. 125, 126-27 (2003). For a proposed interpretation of the Full Faith and Credit Clause that, unlike the Supreme Court's longstanding interpretation, treats the clause as imposing substantial constraints on choice of law, see GaryJ. Simson, State Autonomy in Choice of Law: A SuggestedApproach, 52 S. CAL. L. REv. 61, 6-80 (1978). 237 Early in 2011, U.S. Attorney General Eric Holder announced that the Obama administration had concluded that the Act's definition of marriage for federal purposes violates the Equal Protection Clause of the federal Constitution and that the administration therefore should no longer defend the constitutionality of that provision of the Act when it is challenged in court. Although other administrations at times had refused to defend the constitutionality of an act of Congress, the practice is plainly "rare." See Charlie Savage & Sheryl Gay Stolberg, In Turnabout, U.S. Says Marriage Act Blocks Gay Rights, N.Y. TMES, Feb. 24, 2011, at Al. More recently, the Obama administration announced its support for legislation sponsored by Senator Dianne Feinstein to repeal DOMA in entirety. See Helene Cooper, Obama Will Back Repeal of Law Restricting Marriage, N.Y TMES, July 20, 2011, at A 19. 23.1 199 COLUMBIA JOURNAL OF GENDER AND LAW entire lives, they are second-class citizens for purposes of federal law. Thanks to DOMA's definitional provision, the hundreds of federal laws giving one or another advantage to people in opposite-sex marriages give nothing to ones in same-sex marriages. 238 Though not as potent as a national prohibition on same-sex marriage-a measure that Congress had good reason to believe, and may well have believed, would meet with strong political opposition and judicial invalidation239 -the definitional provision is functionally closely akin to a prohibition for purposes of Establishment Clause review. If the apparent reasons for and against Congress's decision to throw this massive bucket of cold water on same-sex relations are analyzed, the conclusion to be drawn is much the same as with state prohibitions. The decision defies reasonable explanation unless one assumes that religion played a role that the Establishment Clause forbids. CONCLUSION Religion and prohibitions on same-sex marriage are obviously interrelated, but the question of the constitutionality of those prohibitions under the Establishment Clause requires careful and detailed analysis nonetheless. Anything less fails to give due regard to the complexities of both the lawmaking process and the applicable Establishment Clause doctrines. Ultimately, as I have argued, a thorough Establishment Clause analysis strongly points to the conclusion that prohibitions on same-sex marriage impermissibly endorse religion and coerce conformity to religious tenets. Accordingly, under either an endorsement test or a coercion test, such laws should be struck down. The Supreme Court should so hold, as should lower courts before the high court has spoken. Invalidation of prohibitions on same-sex 238 For more on the federal laws conferring benefits of various sorts on the basis of marriage, see supra note 171. Rather remarkably, in urging enactment of DOMA's definitional provision, the House Judiciary Committee report pointed to the "array of material and other benefits" that the federal government confers on married couples and underlined the value of discriminating against gay married couples in terms of "preserving scarce government resources." H.R. REP. No. 104-664, at 18 (1996), reprinted in 1996 U.S.C.C.A.N. 2905, 2922. The repeal of the military's "don't ask, don't tell" policy raises for the first time the question whether military spousal benefits, such as base housing and health insurance, extend to same-sex spouses. Although DOMA would appear to dictate a negative answer, Pentagon officials reportedly are "looking at 'gray areas' that might allow them to extend some benefits." Elisabeth Bumiller, A Final Phase for Ending 'Don 't Ask, Don Tell', N.Y. TIMES, July 23, 2011, at AI3. 239 1 have suggested elsewhere that Congress opted for the definitional provision rather than a prohibition for three reasons: Congress lacked an obvious source of power under the Constitution to enact a prohibition on same-sex marriage; a federal prohibition would substantially interfere with state prerogatives in a field-family law-traditionally regulated by state, rather than federal, law; and public support for a federal prohibition on same- sex marriage is significantly lower than public support for a state prohibition. See Simson, supra note 44, at 45-46. 200 23.1 COLUMBIA JOURNAL OF GENDER AND LAW marriage depends on conscientious application of existing Establishment Clause doctrines, not revision of those doctrines. Though framed in terms of federal constitutional law, the argument presented in this Article also has important implications for the validity of same-sex marriage bans under relevant state constitutional provisions analogous to the U.S. Constitution's Establishment Clause. For much the same reasons that state prohibitions on same-sex marriage should not survive review under the federal clause, they generally should be found wanting under comparable state constitutional constraints. Lastly, I emphasize that my analysis has important implications not only for courts, but for legislators as well. Legislators, whether state or federal, do not fulfill their duty under Article VI to "support this Constitution""24 simply by refraining from enacting legislation that the courts have made clear violates the Constitution. Rather, they themselves are also obliged to think seriously about whether contemplated legislative action satisfies constitutional norms."' In the instance of a prohibition on same-sex marriage, conscientious legislators should recognize that such a prohibition cannot reasonably be justified in terms of permissible policy considerations. They also should recognize that a same-sex marriage prohibition very neatly serves an impermissible objective of endorsing religious opposition to same-sex marriage. If predisposed to vote for such legislation, legislators mindful of their constitutional responsibilities should question whether their support for such legislation rests upon a reasoned assessment of legitimate policy considerations or, instead, in substantial part, on conscious or unconscious reliance on the impermissible objective of endorsing religion. Ultimately, legislators genuinely mindful of their duties under the Establishment Clause have only one real option: to vote against same-sex marriage bans. 240 U.S. CONST. art. VI, cl. 3 (providing that federal and state legislators, as well as members of the executive and judiciary, "shall be bound by Oath or affirmation, to support this Constitution"). 241 See Paul Brest, The Conscientious Legislator s Guide to Constitutional Interpretation, 27 STAN. L. REV. 585 (1975). 23.1