Bowers, Attorney General of Georgia v. Hardwick et al.

Argued March 31, 1986

Decided June 30, 1986

CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE ELEVENTH CIRCUIT

*187 Michael E. Hobbs, Senior Assistant Attorney General of Georgia, argued the cause for petitioner. With him on the briefs were Michael J. Bowers, Attorney General, pro se, Marion O. Gordon, First Assistant Attorney General, and Daryl A. Robinson, Senior Assistant Attorney General.

Laurence H. Tribe argued the cause for respondent Hardwick. With him on the brief were Kathleen M. Sullivan and Kathleen L. Wilde.*

Briefs of amici curiae urging affirmance were filed for the State of New York et al. by Robert Abrams, Attorney General of New York, Robert Hermann, Solicitor General, Lawrence S. Kahn, Howard L. Zwickel, Charles R. Fraser, and Sanford M. Cohen, Assistant Attorneys General, and John Van de Kamp, Attorney General of California; for the American Jewish Congress by Daniel D. Levenson, David Cohen, and Frederick Mandel; for the American Psychological Association et al. by Margaret Farrell Ewing, Donald N. Bersoff, Anne Simon, Nadine Taub, and Herbert Semmel; for the Association of the Bar of the City of New York by Steven A. Rosen; for the National Organization for Women by John S. L. Katz; and for the Presbyterian Church (U. S. A.) et al. by Jeffrey O. Bramlett.

Briefs of amici curiae were filed for the Lesbian Rights Project et al. by Mary C. Dunlap; and for the National Gay Rights Advocates et al. by Edward P Errante, Leonard Graff, and Jay Kohorn.

*

Briefs of amici curiae urging reversal were filed for the Catholic League for Religious and Civil Rights by Steven Frederick McDowell; for the Rutherford Institute et al. by W. Charles Bundren, Guy O. Farley, Jr., George M. Weaver, William B. Hollberg, Wendell R. Bird, John W Whitehead, Thomas O. Kotouc, and Alfred Lindh; and for David Robinson, Jr., pro se.

JUSTICE WHITE delivered the opinion of the Court.

In August 1982, respondent Hardwick (hereafter respondent) was charged with violating the Georgia statute criminalizing *188 sodomy1 by committing that act with another adult male in the bedroom of respondent’s home. After a preliminary hearing, the District Attorney decided not to present the matter to the grand jury unless further evidence developed.

Respondent then brought suit in the Federal District Court, challenging the constitutionality of the statute insofar as it criminalized consensual sodomy.2 He asserted that he was a practicing homosexual, that the Georgia sodomy statute, as administered by the defendants, placed him in imminent danger of arrest, and that the statute for several reasons violates the Federal Constitution. The District Court granted the defendants’ motion to dismiss for failure to state a claim, relying on Doe v. Commonwealth’s Attorney for the City of Richmond, 403 F. Supp. 1199 (ED Va. 1975), which this Court summarily affirmed, 425 U. S. 901 (1976).

*189 A divided panel of the Court of Appeals for the Eleventh Circuit reversed. 760 F. 2d 1202 (1985). The court first held that, because Doe was distinguishable and in any event had been undermined by later decisions, our summary affirmance in that case did not require affirmance of the District Court. Relying on our decisions in Griswold v. Connecticut, 381 U. S. 479 (1965); Eisenstadt v. Baird, 405 U. S. 438 (1972); Stanley v. Georgia, 394 U. S. 557 (1969); and Roe v. Wade, 410 U. S. 113 (1973), the court went on to hold that the Georgia statute violated respondent’s fundamental rights because his homosexual activity is a private and intimate association that is beyond the reach of state regulation by reason of the Ninth Amendment and the Due Process Clause of the Fourteenth Amendment. The case was remanded for trial, at which, to prevail, the State would have to prove that the statute is supported by a compelling interest and is the most narrowly drawn means of achieving that end.

Because other Courts of Appeals have arrived at judgments contrary to that of the Eleventh Circuit in this case,3 we granted the Attorney General’s petition for certiorari questioning the holding that the sodomy statute violates the fundamental rights of homosexuals. We agree with petitioner that the Court of Appeals erred, and hence reverse its judgment.4

*190 This case does not require a judgment on whether laws against sodomy between consenting adults in general, or between homosexuals in particular, are wise or desirable. It raises no question about the right or propriety of state legislative decisions to repeal their laws that criminalize homosexual sodomy, or of state-court decisions invalidating those laws on state constitutional grounds. The issue presented is whether the Federal Constitution confers a fundamental right upon homosexuals to engage in sodomy and hence invalidates the laws of the many States that still make such conduct illegal and have done so for a very long time. The case also calls for some judgment about the limits of the Court’s role in carrying out its constitutional mandate.

We first register our disagreement with the Court of Appeals and with respondent that the Court’s prior cases have construed the Constitution to confer a right of privacy that extends to homosexual sodomy and for all intents and purposes have decided this case. The reach of this line of cases was sketched in Carey v. Population Services International, 431 U. S. 678, 685 (1977). Pierce v. Society of Sisters, 268 U. S. 510 (1925), and Meyer v. Nebraska, 262 U. S. 390 (1923), were described as dealing with child rearing and education; Prince v. Massachusetts, 321 U. S. 158 (1944), with family relationships; Skinner v. Oklahoma ex rel. Williamson, 316 U. S. 535 (1942), with procreation; Loving v. Virginia, 388 U. S. 1 (1967), with marriage; Griswold v. Connecticut, supra, and Eisenstadt v. Baird, supra, with contraception; and Roe v. Wade, 410 U. S. 113 (1973), with abortion. The latter three cases were interpreted as construing the Due Process Clause of the Fourteenth Amendment to confer a fundamental individual right to decide whether or not to beget or bear a child. Carey v. Population Services International, supra, at 688-689.

Accepting the decisions in these cases and the above description of them, we think it evident that none of the rights announced in those cases bears any resemblance to the *191 claimed constitutional right of homosexuals to engage in acts of sodomy that is asserted in this case. No connection between family, marriage, or procreation on the one hand and homosexual activity on the other has been demonstrated, either by the Court of Appeals or by respondent. Moreover, any claim that these cases nevertheless stand for the proposition that any kind of private sexual conduct between consenting adults is constitutionally insulated from state proscription is unsupportable. Indeed, the Court’s opinion in Carey twice asserted that the privacy right, which the Griswold line of cases found to be one of the protections provided by the Due Process Clause, did not reach so far. 431 U. S., at 688, n. 5, 694, n. 17.

Precedent aside, however, respondent would have us announce, as the Court of Appeals did, a fundamental right to engage in homosexual sodomy. This we are quite unwilling to do. It is true that despite the language of the Due Process Clauses of the Fifth and Fourteenth Amendments, which appears to focus only on the processes by which life, liberty, or property is taken, the cases are legion in which those Clauses have been interpreted to have substantive content, subsuming rights that to a great extent are immune from federal or state regulation or proscription. Among such cases are those recognizing rights that have little or no textual support in the constitutional language. Meyer, Prince, and Pierce fall in this category, as do the privacy cases from Griswold to Carey.

Striving to assure itself and the public that announcing rights not readily identifiable in the Constitution’s text involves much more than the imposition of the Justices’ own choice of values on the States and the Federal Government, the Court has sought to identify the nature of the rights qualifying for heightened judicial protection. In Palko v. Connecticut, 302 U. S. 319, 325, 326 (1937), it was said that this category includes those fundamental liberties that are “implicit in the concept of ordered liberty,” such that “neither *192 liberty nor justice would exist if [they] were sacrificed.” A different description of fundamental liberties appeared in Moore v. East Cleveland, 431 U. S. 494, 503 (1977) (opinion of POWELL, J.), where they are characterized as those liberties that are “deeply rooted in this Nation’s history and tradition.” Id., at 503 (POWELL, J.). See also Griswold v. Connecticut, 381 U. S., at 506.

It is obvious to us that neither of these formulations would extend a fundamental right to homosexuals to engage in acts of consensual sodomy. Proscriptions against that conduct have ancient roots. See generally Survey on the Constitutional Right to Privacy in the Context of Homosexual Activity, 40 U. Miami L. Rev. 521, 525 (1986). Sodomy was a criminal offense at common law and was forbidden by the laws of the original 13 States when they ratified the Bill of Rights.5 In 1868, when the Fourteenth Amendment was *193 ratified, all but 5 of the 37 States in the Union had criminal sodomy laws.6 In fact, until 1961,7 all 50 States outlawed sodomy, and today, 24 States and the District of Columbia *194 continue to provide criminal penalties for sodomy performed in private and between consenting adults. See Survey, U. Miami L. Rev., supra, at 524, n. 9. Against this background, to claim that a right to engage in such conduct is “deeply rooted in this Nation’s history and tradition” or “implicit in the concept of ordered liberty” is, at best, facetious.

Nor are we inclined to take a more expansive view of our authority to discover new fundamental rights imbedded in the Due Process Clause. The Court is most vulnerable and comes nearest to illegitimacy when it deals with judge-made constitutional law having little or no cognizable roots in the language or design of the Constitution. That this is so was painfully demonstrated by the face-off between the Executive and the Court in the 1930’s, which resulted in the repudiation *195 of much of the substantive gloss that the Court had placed on the Due Process Clauses of the Fifth and Fourteenth Amendments. There should be, therefore, great resistance to expand the substantive reach of those Clauses, particularly if it requires redefining the category of rights deemed to be fundamental. Otherwise, the Judiciary necessarily takes to itself further authority to govern the country without express constitutional authority. The claimed right pressed on us today falls for short of overcoming this resistance.

Respondent, however, asserts that the result should be different where the homosexual conduct occurs in the privacy of the home. He relies on Stanley v. Georgia, 394 U. S. 557 (1969), where the Court held that the First Amendment prevents conviction for possessing and reading obscene material in the privacy of one’s home: “If the First Amendment means anything, it means that a State has no business telling a man, sitting alone in his house, what books he may read or what films he may watch.” Id., at 565.

Stanley did protect conduct that would not have been protected outside the home, and it partially prevented the enforcement of state obscenity laws; but the decision was firmly grounded in the First Amendment. The right pressed upon us here has no similar support in the text of the Constitution, and it does not qualify for recognition under the prevailing principles for construing the Fourteenth Amendment. Its limits are also difficult to discern. Plainly enough, otherwise illegal conduct is not always immunized whenever it occurs in the home. Victimless crimes, such as the possession and use of illegal drugs, do not escape the law where they are committed at home. Stanley itself recognized that its holding offered no protection for the possession in the home of drugs, firearms, or stolen goods. Id., at 568, n. 11. And if respondent’s submission is limited to the voluntary sexual conduct between consenting adults, it would be difficult, except by fiat, to limit the claimed right to homosexual conduct *196 while leaving exposed to prosecution adultery, incest, and other sexual crimes even though they are committed in the home. We are unwilling to start down that road.

Even if the conduct at issue here is not a fundamental right, respondent asserts that there must be a rational basis for the law and that there is none in this case other than the presumed belief of a majority of the electorate in Georgia that homosexual sodomy is immoral and unacceptable. This is said to be an inadequate rationale to support the law. The law, however, is constantly based on notions of morality, and if all laws representing essentially moral choices are to be invalidated under the Due Process Clause, the courts will be very busy indeed. Even respondent makes no such claim, but insists that majority sentiments about the morality of homosexuality should be declared inadequate. We do not agree, and are unpersuaded that the sodomy laws of some 25 States should be invalidated on this basis.8

Accordingly, the judgment of the Court of Appeals is

Reversed.

1

Georgia Code Ann. § 16-6-2 (1984) provides, in pertinent part, as follows:

”(a) A person commits the offense of sodomy when he performs or submits to any sexual act involving the sex organs of one person and the mouth or anus of another. . . .

”(b) A person convicted of the offense of sodomy shall be punished by imprisonment for not less than one nor more than 20 years. . . .”

2

John and Mary Doe were also plaintiffs in the action. They alleged that they wished to engage in sexual activity proscribed by § 16-6-2 in the privacy of their home, App. 3, and that they had been “chilled and deterred” from engaging in such activity by both the existence of the statute and Hardwick’s arrest. Id., at 5. The District Court held, however, that because they had neither sustained, nor were in immediate danger of sustaining, any direct injury from the enforcement of the statute, they did not have proper standing to maintain the action. Id., at 18. The Court of Appeals affirmed the District Court’s judgment dismissing the Does’ claim for lack of standing, 760 F. 2d 1202, 1206-1207 (CA11 1985), and the Does do not challenge that holding in this Court.

The only claim properly before the Court, therefore, is Hardwick’s challenge to the Georgia statute as applied to consensual homosexual sodomy. We express no opinion on the constitutionality of the Georgia statute as applied to other acts of sodomy.

3

See Baker v. Wade, 769 F. 2d 289, rehearing denied, 774 F. 2d 1285 (CA5 1985) (en banc); Dronenburg v. Zech, 239 U. S. App. D. C. 229, 741 F. 2d 1388, rehearing denied, 241 U. S. App. D. C. 262, 746 F. 2d 1579 (1984).

4

Petitioner also submits that the Court of Appeals erred in holding that the District Court was not obligated to follow our summary affirmance in Doe. We need not resolve this dispute, for we prefer to give plenary consideration to the merits of this case rather than rely on our earlier action in Doe. See Usery v. Turner Elkhorn Mining Co., 428 U. S. 1, 14 (1976); Massachusetts Board of Retirement v. Murgia, 427 U. S. 307, 309, n. 1 (1976); Edelman v. Jordan, 415 U. S. 651, 671 (1974). Cf. Hicks v. Miranda, 422 U. S. 332, 344 (1975).

5

Criminal sodomy laws in effect in 1791:

Connecticut: 1 Public Statute Laws of the State of Connecticut, 1808, Title LXVI, ch. 1, § 2 (rev. 1672).

Delaware: 1 Laws of the State of Delaware, 1797, ch. 22, § 5 (passed 1719).

Georgia had no criminal sodomy statute until 1816, but sodomy was a crime at common law, and the General Assembly adopted the common law of England as the law of Georgia in 1784. The First Laws of the State of Georgia, pt. 1, p. 290 (1981).

Maryland had no criminal sodomy statute in 1791. Maryland’s Declaration of Rights, passed in 1776, however, stated that “the inhabitants of Maryland are entitled to the common law of England,” and sodomy was a crime at common law. 4 W. Swindler, Sources and Documents of United States Constitutions 372 (1975).

Massachusetts: Acts and Laws passed by the General Court of Massachusetts, ch. 14, Act of Mar. 3, 1785.

New Hampshire passed its first sodomy statute in 1718. Acts and Laws of New Hampshire 1680-1726, p. 141 (1978).

Sodomy was a crime at common law in New Jersey at the time of the ratification of the Bill of Rights. The State enacted its first criminal sodomy law five years later. Acts of the Twentieth General Assembly, Mar. 18, 1796, ch. DC, § 7.

New York: Laws of New York, ch. 21 (passed 1787).

At the time of ratification of the Bill of Rights, North Carolina had adopted the English statute of Henry VIII outlawing sodomy. See Collection of the Statutes of the Parliament of England in Force in the State of North-Carolina, ch. 17, p. 314 (Martin ed. 1792).

Pennsylvania: Laws of the Fourteenth General Assembly of the Commonwealth of Pennsylvania, ch. CLIV, § 2 (passed 1790).

Rhode Island passed its first sodomy law in 1662. The Earliest Acts and Laws of the Colony of Rhode Island and Providence Plantations 1647-1719, p. 142 (1977).

South Carolina: Public Laws of the State of South Carolina, p. 49 (1790).

At the time of the ratification of the Bill of Rights, Virginia had no specific statute outlawing sodomy, but had adopted the English common law. 9 Hening’s Laws of Virginia, ch. 5, § 6, p. 127 (1821) (passed 1776).

6

Criminal sodomy statutes in effect in 1868:

Alabama: Ala. Rev. Code § 3604 (1867).

Arizona (Terr.): Howell Code, ch. 10, § 48 (1865).

Arkansas: Ark. Stat., ch. 51, Art. IV, § 5 (1858).

California: 1 Cal. Gen. Laws, ¶ 1450, § 48 (1865).

Colorado (Terr.): Colo. Rev. Stat., ch. 22, §§ 45, 46 (1868).

Connecticut: Conn. Gen. Stat., Tit. 122, ch. 7, § 124 (1866).

Delaware: Del. Rev. Stat., ch. 131, § 7 (1893).

Florida: Fla. Rev. Stat., div. 5, § 2614 (passed 1868) (1892).

Georgia: Ga. Code §§ 4286, 4287, 4290 (1867).

Kingdom of Hawaii: Haw. Penal Code, ch. 13, § 11 (1869).

Illinois: Ill. Rev. Stat., div. 5, §§ 49, 50 (1845).

Kansas (Terr.): Kan. Stat., ch. 53, § 7 (1855).

Kentucky: 1 Ky. Rev. Stat., ch. 28, Art. IV, § 11 (1860).

Louisiana: La. Rev. Stat., Crimes and Offences, § 5 (1856).

Maine: Me. Rev. Stat., Tit. XII, ch. 160, § 4 (1840).

Maryland: 1 Md. Code, Art. 30, § 201 (1860).

Massachusetts: Mass. Gen. Stat., ch. 165, § 18 (1860).

Michigan: Mich. Rev. Stat., Tit. 30, ch. 158, § 16 (1846).

Minnesota: Minn. Stat., ch. 96, § 13 (1859).

Mississippi: Miss. Rev. Code, ch. 64, § LII, Art. 238 (1857).

Missouri: 1 Mo. Rev. Stat., ch. 50, Art. VIII, § 7 (1856).

Montana (Terr.): Mont. Acts, Resolutions, Memorials, Criminal Practice Acts, ch. IV, § 44 (1866).

Nebraska (Terr.): Neb. Rev. Stat., Crim. Code, ch. 4, § 47 (1866).

Nevada (Terr.): Nev. Comp. Laws, 1861-1900, Crimes and Punishments, § 45.

New Hampshire: N. H. Laws, Act. of June 19, 1812, § 5 (1815).

New Jersey: N. J. Rev. Stat., Tit. 8, ch. 1, § 9 (1847).

New York: 3 N. Y. Rev. Stat., pt. 4, ch. 1, Tit. 5, § 20 (5th ed. 1859).

North Carolina: N. C. Rev. Code, ch. 34, § 6 (1855).

Oregon: Laws of Ore., Crimes — Against Morality, etc., ch. 7, § 655 (1874).

Pennsylvania: Act of Mar. 31, 1860, § 32, Pub. L. 392, in 1 Digest of Statute Law of Pa. 1700-1903, p. 1011 (Purdon 1905).

Rhode Island: R. I. Gen. Stat., ch. 232, § 12 (1872).

South Carolina: Act of 1712, in 2 Stat. at Large of S. C. 1682-1716, p. 493 (1837).

Tennessee: Tenn. Code, ch. 8, Art. 1, § 4843 (1858).

Texas: Tex. Rev. Stat., Tit. 10, ch. 5, Art. 342 (1887) (passed 1860).

Vermont: Acts and Laws of the State of Vt. (1779).

Virginia: Va. Code, ch. 149, § 12 (1868).

West Virginia: W. Va. Code, ch. 149, § 12 (1868).

Wisconsin (Terr.): Wis. Stat. § 14, p. 367 (1839).

7

In 1961, Illinois adopted the American Law Institute’s Model Penal Code, which decriminalized adult, consensual, private, sexual conduct. Criminal Code of 1961, §§ 11-2, 11-3, 1961 Ill. Laws, pp. 1985, 2006 (codified as amended at Ill. Rev. Stat., ch. 38, ¶¶ 11-2, 11-3 (1983) (repealed 1984)). See American Law Institute, Model Penal Code § 213.2 (Proposed Official Draft 1962).

8

Respondent does not defend the judgment below based on the Ninth Amendment, the Equal Protection Clause, or the Eighth Amendment.

CHIEF JUSTICE BURGER, concurring.

I join the Court’s opinion, but I write separately to underscore my view that in constitutional terms there is no such thing as a fundamental right to commit homosexual sodomy.

As the Court notes, ante, at 192, the proscriptions against sodomy have very “ancient roots.” Decisions of individuals relating to homosexual conduct have been subject to state intervention throughout the history of Western civilization. Condemnation of those practices is firmly rooted in Judeo-Christian moral and ethical standards. Homosexual sodomy was a capital crime under Roman law. See Code Theod. 9.7.6; Code Just. 9.9.31. See also D. Bailey, Homosexuality *197 and the Western Christian Tradition 70-81 (1975). During the English Reformation when powers of the ecclesiastical courts were transferred to the King’s Courts, the first English statute criminalizing sodomy was passed. 25 Hen. VIII, ch. 6. Blackstone described “the infamous crime against nature” as an offense of “deeper malignity” than rape, a heinous act “the very mention of which is a disgrace to human nature,” and “a crime not fit to be named.” 4 W. Blackstone, Commentaries *215. The common law of England, including its prohibition of sodomy, became the received law of Georgia and the other Colonies. In 1816 the Georgia Legislature passed the statute at issue here, and that statute has been continuously in force in one form or another since that time. To hold that the act of homosexual sodomy is somehow protected as a fundamental right would be to cast aside millennia of moral teaching.

This is essentially not a question of personal “preferences” but rather of the legislative authority of the State. I find nothing in the Constitution depriving a State of the power to enact the statute challenged here.

JUSTICE POWELL, concurring.

I join the opinion of the Court. I agree with the Court that there is no fundamental right — i. e., no substantive right under the Due Process Clause — such as that claimed by respondent Hardwick, and found to exist by the Court of Appeals. This is not to suggest, however, that respondent may not be protected by the Eighth Amendment of the Constitution. The Georgia statute at issue in this case, Ga. Code Ann. § 16-6-2 (1984), authorizes a court to imprison a person for up to 20 years for a single private, consensual act of sodomy. In my view, a prison sentence for such conduct — certainly a sentence of long duration — would create a serious Eighth Amendment issue. Under the Georgia statute a single act of sodomy, even in the private setting of a home, is a *198 felony comparable in terms of the possible sentence imposed to serious felonies such as aggravated battery, § 16-5-24, first-degree arson, § 16-7-60, and robbery, § 16-8-40.1

In this case, however, respondent has not been tried, much less convicted and sentenced.2 Moreover, respondent has not raised the Eighth Amendment issue below. For these reasons this constitutional argument is not before us.

1

Among those States that continue to make sodomy a crime, Georgia authorizes one of the longest possible sentences. See Ala. Code § 13A-6-65(a)(3) (1982) (1-year maximum); Ariz. Rev. Stat. Ann. §§ 13-1411, 13-1412 (West Supp. 1985) (30 days); Ark. Stat. Ann. § 41-1813 (1977) (1-year maximum); D. C. Code § 22-3502 (1981) (10-year maximum); Fla. Stat. § 800.02 (1985) (60-day maximum); Ga. Code Ann. § 16-6-2 (1984) (1 to 20 years); Idaho Code § 18-6605 (1979) (5-year minimum); Kan. Stat. Ann. § 21-3505 (Supp. 1985) (6-month maximum); Ky. Rev. Stat. § 510.100 (1985) (90 days to 12 months); La. Rev. Stat. Ann. § 14:89 (West 1986) (5-year maximum); Md. Ann. Code, Art. 27, §§ 553-554 (1982) (10-year maximum); Mich. Comp. Laws § 750.158 (1968) (15-year maximum); Minn. Stat. § 609.293 (1984) (1-year maximum); Miss. Code Ann. § 97-29-59 (1973) (10-year maximum); Mo. Rev. Stat. § 566.090 (Supp. 1984) (1-year maximum); Mont. Code Ann. § 45-5-505 (1985) (10-year maximum); Nev. Rev. Stat. § 201.190 (1985) (6-year maximum); N. C. Gen. Stat. § 14-177 (1981) (10-year maximum); Okla. Stat., Tit. 21, § 886 (1981) (10-year maximum); R. I. Gen. Laws § 11-10-1 (1981) (7 to 20 years); S. C. Code § 16-15-120 (1985) (5-year maximum); Tenn. Code Ann. § 39-2-612 (1982) (5 to 15 years); Tex. Penal Code Ann. § 21.06 (1974) ($200 maximum fine); Utah Code Ann. § 76-5-403 (1978) (6-month maximum); Va. Code § 18.2-361 (1982) (5-year maximum).

2

It was conceded at oral argument that, prior to the complaint against respondent Hardwick, there had been no reported decision involving prosecution for private homosexual sodomy under this statute for several decades. See Thompson v. Aldredge, 187 Ga. 467, 200 S. E. 799 (1939). Moreover, the State has declined to present the criminal charge against Hardwick to a grand jury, and this is a suit for declaratory judgment brought by respondents challenging the validity of the statute. The history of nonenforcement suggests the moribund character today of laws criminalizing this type of private, consensual conduct. Some 26 States have repealed similar statutes. But the constitutional validity of the Georgia statute was put in issue by respondents, and for the reasons stated by the Court, I cannot say that conduct condemned for hundreds of years has now become a fundamental right.

*199 JUSTICE BLACKMUN, with whom JUSTICE BRENNAN, JUSTICE MARSHALL, and JUSTICE STEVENS join, dissenting.

This case is no more about “a fundamental right to engage in homosexual sodomy,” as the Court purports to declare, ante, at 191, than Stanley v. Georgia, 394 U. S. 557 (1969), was about a fundamental right to watch obscene movies, or Katz v. United States, 389 U. S. 347 (1967), was about a fundamental right to place interstate bets from a telephone booth. Rather, this case is about “the most comprehensive of rights and the right most valued by civilized men,” namely, “the right to be let alone.” Olmstead v. United States, 277 U. S. 438, 478 (1928) (Brandeis, J., dissenting).

The statute at issue, Ga. Code Ann. § 16-6-2 (1984), denies individuals the right to decide for themselves whether to engage in particular forms of private, consensual sexual activity. The Court concludes that § 16-6-2 is valid essentially because “the laws of . . . many States . . . still make such conduct illegal and have done so for a very long time.” Ante, at 190. But the fact that the moral judgments expressed by statutes like § 16-6-2 may be ” `natural and familiar . . . ought not to conclude our judgment upon the question whether statutes embodying them conflict with the Constitution of the United States.’ ” Roe v. Wade, 410 U. S. 113, 117 (1973), quoting Lochner v. New York, 198 U. S. 45, 76 (1905) (Holmes, J., dissenting). Like Justice Holmes, I believe that ”[i]t is revolting to have no better reason for a rule of law than that so it was laid down in the time of Henry IV. It is still more revolting if the grounds upon which it was laid down have vanished long since, and the rule simply persists from blind imitation of the past.” Holmes, The Path of the Law, 10 Harv. L. Rev. 457, 469 (1897). I believe we must analyze respondent Hardwick’s claim in the light of the values that underlie the constitutional right to privacy. If that right means anything, it means that, before Georgia can prosecute its citizens for making choices about the most intimate *200 aspects of their lives, it must do more than assert that the choice they have made is an ” `abominable crime not fit to be named among Christians.’ ” Herring v. State, 119 Ga. 709, 721, 46 S. E. 876, 882 (1904).

I

In its haste to reverse the Court of Appeals and hold that the Constitution does not “confe® a fundamental right upon homosexuals to engage in sodomy,” ante, at 190, the Court relegates the actual statute being challenged to a footnote and ignores the procedural posture of the case before it. A fair reading of the statute and of the complaint clearly reveals that the majority has distorted the question this case presents.

First, the Court’s almost obsessive focus on homosexual activity is particularly hard to justify in light of the broad language Georgia has used. Unlike the Court, the Georgia Legislature has not proceeded on the assumption that homosexuals are so different from other citizens that their lives may be controlled in a way that would not be tolerated if it limited the choices of those other citizens. Cf. ante, at 188, n. 2. Rather, Georgia has provided that ”[a] person commits the offense of sodomy when he performs or submits to any sexual act involving the sex organs of one person and the mouth or anus of another.” Ga. Code Ann. § 16-6-2(a) (1984). The sex or status of the persons who engage in the act is irrelevant as a matter of state law. In fact, to the extent I can discern a legislative purpose for Georgia’s 1968 enactment of § 16-6-2, that purpose seems to have been to broaden the coverage of the law to reach heterosexual as well as homosexual activity.1 I therefore see no basis for the *201 Court’s decision to treat this case as an “as applied” challenge to § 16-6-2, see ante, at 188, n. 2, or for Georgia’s attempt, both in its brief and at oral argument, to defend § 16-6-2 solely on the grounds that it prohibits homosexual activity. Michael Hardwick’s standing may rest in significant part on Georgia’s apparent willingness to enforce against homosexuals a law it seems not to have any desire to enforce against heterosexuals. See Tr. of Oral Arg. 4-5; cf. 760 F. 2d 1202, 1205-1206 (CA11 1985). But his claim that § 16-6-2 involves an unconstitutional intrusion into his privacy and his right of intimate association does not depend in any way on his sexual orientation.

Second, I disagree with the Court’s refusal to consider whether § 16-6-2 runs afoul of the Eighth or Ninth Amendments or the Equal Protection Clause of the Fourteenth Amendment. Ante, at 196, n. 8. Respondent’s complaint expressly invoked the Ninth Amendment, see App. 6, and he relied heavily before this Court on Griswold v. Connecticut, 381 U. S. 479, 484 (1965), which identifies that Amendment as one of the specific constitutional provisions giving “life and substance” to our understanding of privacy. See Brief for Respondent Hardwick 10-12; Tr. of Oral Arg. 33. More importantly, the procedural posture of the case requires that we affirm the Court of Appeals’ judgment if there is any ground on which respondent may be entitled to relief. This case is before us on petitioner’s motion to dismiss for failure to state a claim, Fed. Rule Civ. Proc. 12(b)(6). See App. 17. It is a well-settled principle of law that “a complaint should not be dismissed merely because a plaintiff’s allegations do not support the particular legal theory he advances, for the court is under a duty to examine the complaint to determine if the allegations provide for relief on any possible theory.” *202 Bramlet v. Wilson, 495 F. 2d 714, 716 (CA8 1974); see Parr v. Great Lakes Express Co., 484 F. 2d 767, 773 (CA7 1973); Due v. Tallahassee Theaters, Inc., 333 F. 2d 630, 631 (CA5 1964); United States v. Howell, 318 F. 2d 162, 166 (CA9 1963); 5 C. Wright & A. Miller, Federal Practice and Procedure § 1357, pp. 601-602 (1969); see also Conley v. Gibson, 355 U. S. 41, 45-46 (1957). Thus, even if respondent did not advance claims based on the Eighth or Ninth Amendments, or on the Equal Protection Clause, his complaint should not be dismissed if any of those provisions could entitle him to relief. I need not reach either the Eighth Amendment or the Equal Protection Clause issues because I believe that Hardwick has stated a cognizable claim that § 16-6-2 interferes with constitutionally protected interests in privacy and freedom of intimate association. But neither the Eighth Amendment nor the Equal Protection Clause is so clearly irrelevant that a claim resting on either provision should be peremptorily dismissed.2 The Court’s cramped reading of the *203 issue before it makes for a short opinion, but it does little to make for a persuasive one.

II

“Our cases long have recognized that the Constitution embodies a promise that a certain private sphere of individual liberty will be kept largely beyond the reach of government.” Thornburgh v. American College of Obstetricians & Gynecologists, 476 U. S. 747, 772 (1986). In construing the right to privacy, the Court has proceeded along two somewhat distinct, *204 albeit complementary, lines. First, it has recognized a privacy interest with reference to certain decisions that are properly for the individual to make. E. g., Roe v. Wade, 410 U. S. 113 (1973); Pierce v. Society of Sisters, 268 U. S. 510 (1925). Second, it has recognized a privacy interest with reference to certain places without regard for the particular activities in which the individuals who occupy them are engaged. E. g., United States v. Karo, 468 U. S. 705 (1984); Payton v. New York, 445 U. S. 573 (1980); Rios v. United States, 364 U. S. 253 (1960). The case before us implicates both the decisional and the spatial aspects of the right to privacy.

A

The Court concludes today that none of our prior cases dealing with various decisions that individuals are entitled to make free of governmental interference “bears any resemblance to the claimed constitutional right of homosexuals to engage in acts of sodomy that is asserted in this case.” Ante, at 190-191. While it is true that these cases may be characterized by their connection to protection of the family, see Roberts v. United States Jaycees, 468 U. S. 609, 619 (1984), the Court’s conclusion that they extend no further than this boundary ignores the warning in Moore v. East Cleveland, 431 U. S. 494, 501 (1977) (plurality opinion), against “clos[ing] our eyes to the basic reasons why certain rights associated with the family have been accorded shelter under the Fourteenth Amendment’s Due Process Clause.” We protect those rights not because they contribute, in some direct and material way, to the general public welfare, but because they form so central a part of an individual’s life. ”[T]he concept of privacy embodies the `moral fact that a person belongs to himself and not others nor to society as a whole.’ ” Thornburgh v. American College of Obstetricians & Gynecologists, 476 U. S., at 777, n. 5 (STEVENS, J., concurring), quoting Fried, Correspondence, 6 Phil. & Pub. Affairs 288-289 (1977). And so we protect the decision whether to *205 marry precisely because marriage “is an association that promotes a way of life, not causes; a harmony in living, not political faiths; a bilateral loyalty, not commercial or social projects.” Griswold v. Connecticut, 381 U. S., at 486. We protect the decision whether to have a child because parenthood alters so dramatically an individual’s self-definition, not because of demographic considerations or the Bible’s command to be fruitful and multiply. Cf. Thornburgh v. American College of Obstetricians & Gynecologists, supra, at 777, n. 6 (STEVENS, J., concurring). And we protect the family because it contributes so powerfully to the happiness of individuals, not because of a preference for stereotypical households. Cf. Moore v. East Cleveland, 431 U. S., at 500-506 (plurality opinion). The Court recognized in Roberts, 468 U. S., at 619, that the “ability independently to define one’s identity that is central to any concept of liberty” cannot truly be exercised in a vacuum; we all depend on the “emotional enrichment from close ties with others.” Ibid.

Only the most willful blindness could obscure the fact that sexual intimacy is “a sensitive, key relationship of human existence, central to family life, community welfare, and the development of human personality,” Paris Adult Theatre I v. Slaton, 413 U. S. 49, 63 (1973); see also Carey v. Population Services International, 431 U. S. 678, 685 (1977). The fact that individuals define themselves in a significant way through their intimate sexual relationships with others suggests, in a Nation as diverse as ours, that there may be many “right” ways of conducting those relationships, and that much of the richness of a relationship will come from the freedom an individual has to choose the form and nature of these intensely personal bonds. See Karst, The Freedom of Intimate Association, 89 Yale L. J. 624, 637 (1980); cf. Eisenstadt v. Baird, 405 U. S. 438, 453 (1972); Roe v. Wade, 410 U. S., at 153.

In a variety of circumstances we have recognized that a necessary corollary of giving individuals freedom to choose *206 how to conduct their lives is acceptance of the fact that different individuals will make different choices. For example, in holding that the clearly important state interest in public education should give way to a competing claim by the Amish to the effect that extended formal schooling threatened their way of life, the Court declared: “There can be no assumption that today’s majority is `right’ and the Amish and others like them are `wrong.’ A way of life that is odd or even erratic but interferes with no rights or interests of others is not to be condemned because it is different.” Wisconsin v. Yoder, 406 U. S. 205, 223-224 (1972). The Court claims that its decision today merely refuses to recognize a fundamental right to engage in homosexual sodomy; what the Court really has refused to recognize is the fundamental interest all individuals have in controlling the nature of their intimate associations with others.

B

The behavior for which Hardwick faces prosecution occurred in his own home, a place to which the Fourth Amendment attaches special significance. The Court’s treatment of this aspect of the case is symptomatic of its overall refusal to consider the broad principles that have informed our treatment of privacy in specific cases. Just as the right to privacy is more than the mere aggregation of a number of entitlements to engage in specific behavior, so too, protecting the physical integrity of the home is more than merely a means of protecting specific activities that often take place there. Even when our understanding of the contours of the right to privacy depends on “reference to a `place,’ ” Katz v. United States, 389 U. S., at 361 (Harlan, J., concurring), “the essence of a Fourth Amendment violation is `not the breaking of [a person’s] doors, and the rummaging of his drawers,’ but rather is `the invasion of his indefensible right of personal security, personal liberty and private property.’ ” California v. Ciraolo, 476 U. S. 207, 226 (1986) (POWELL, J., dissenting), *207 quoting Boyd v. United States, 116 U. S. 616, 630 (1886).

The Court’s interpretation of the pivotal case of Stanley v. Georgia, 394 U. S. 557 (1969), is entirely unconvincing. Stanley held that Georgia’s undoubted power to punish the public distribution of constitutionally unprotected, obscene material did not permit the State to punish the private possession of such material. According to the majority here, Stanley relied entirely on the First Amendment, and thus, it is claimed, sheds no light on cases not involving printed materials. Ante, at 195. But that is not what Stanley said. Rather, the Stanley Court anchored its holding in the Fourth Amendment’s special protection for the individual in his home:

” `The makers of our Constitution undertook to secure conditions favorable to the pursuit of happiness. They recognized the significance of man’s spiritual nature, of his feelings and of his intellect. They knew that only a part of the pain, pleasure and satisfactions of life are to be found in material things. They sought to protect Americans in their beliefs, their thoughts, their emotions and their sensations.’ ..... “These are the rights that appellant is asserting in the case before us. He is asserting the right to read or observe what he pleases — the right to satisfy his intellectual and emotional needs in the privacy of his own home.” 394 U. S., at 564-565, quoting Olmstead v. United States, 277 U. S., at 478 (Brandeis, J., dissenting).

The central place that Stanley gives Justice Brandeis’ dissent in Olmstead, a case raising no First Amendment claim, shows that Stanley rested as much on the Court’s understanding of the Fourth Amendment as it did on the First. Indeed, in Paris Adult Theatre I v. Slaton, 413 U. S. 49 (1973), the Court suggested that reliance on the Fourth *208 Amendment not only supported the Court’s outcome in Stanley but actually was necessary to it: “If obscene material unprotected by the First Amendment in itself carried with it a `penumbra’ of constitutionally protected privacy, this Court would not have found it necessary to decide Stanley on the narrow basis of the `privacy of the home,’ which was hardly more than a reaffirmation that `a man’s home is his castle.’ ” 413 U. S., at 66. “The right of the people to be secure in their . . . houses,” expressly guaranteed by the Fourth Amendment, is perhaps the most “textual” of the various constitutional provisions that inform our understanding of the right to privacy, and thus I cannot agree with the Court’s statement that ”[t]he right pressed upon us here has no . . . support in the text of the Constitution,” ante, at 195. Indeed, the right of an individual to conduct intimate relationships in the intimacy of his or her own home seems to me to be the heart of the Constitution’s protection of privacy.

III

The Court’s failure to comprehend the magnitude of the liberty interests at stake in this case leads it to slight the question whether petitioner, on behalf of the State, has justified Georgia’s infringement on these interests. I believe that neither of the two general justifications for § 16-6-2 that petitioner has advanced warrants dismissing respondent’s challenge for failure to state a claim.

First, petitioner asserts that the acts made criminal by the statute may have serious adverse consequences for “the general public health and welfare,” such as spreading communicable diseases or fostering other criminal activity. Brief for Petitioner 37. Inasmuch as this case was dismissed by the District Court on the pleading, it is not surprising that the record before us is barren of any evidence to support petitioner’s claim.3 In light of the state of the record, I see *209 no justification for the Court’s attempt to equate the private, consensual sexual activity at issue here with the “possession in the home of drugs, firearms, or stolen goods,” ante, at 195, to which Stanley refused to extend its protection. 394 U. S., at 568, n. 11. None of the behavior so mentioned in Stanley can properly be viewed as ”[v]ictimless,” ante, at 195: drugs and weapons are inherently dangerous, see, e. g., McLaughlin v. United States, 476 U. S. 16 (1986), and for property to be “stolen,” someone must have been wrongfully deprived of it. Nothing in the record before the Court provides any justification for finding the activity forbidden by § 16-6-2 to be physically dangerous, either to the persons engaged in it or to others.4

*210 The core of petitioner’s defense of § 16-6-2, however, is that respondent and others who engage in the conduct prohibited by § 16-6-2 interfere with Georgia’s exercise of the ” `right of the Nation and of the States to maintain a decent society,’ ” Paris Adult Theater I v. Slaton, 413 U. S., at 59-60, quoting Jacobellis v. Ohio, 378 U. S. 184, 199 (1964) (Warren, C. J., dissenting). Essentially, petitioner argues, and the Court agrees, that the fact that the acts described in § 16-6-2 “for hundreds of years, if not thousands, have been uniformly condemned as immoral” is a sufficient reason to permit a State to ban them today. Brief for Petitioner 19; see ante, at 190, 192-194, 196.

I cannot agree that either the length of time a majority has held its convictions or the passions with which it defends them can withdraw legislation from this Court’s security. See, e. g., Roe v. Wade, 410 U. S. 113 (1973); Loving v. Virginia, 388 U. S. 1 (1967); Brown v. Board of Education, 347 U. S. 483 (1954).5 As Justice Jackson wrote so eloquently *211 for the Court in West Virginia Board of Education v. Barnette, 319 U. S. 624, 641-642 (1943), “we apply the limitations of the Constitution with no fear that freedom to be intellectually and spiritually diverse or even contrary will disintegrate the social organization. . . . [F]reedom to differ is not limited to things that do not matter much. That would be a mere shadow of freedom. The test of its substance is the right to differ as to things that touch the heart of the existing order.” See also Karst, 89 Yale L. J., at 627. It is precisely because the issue raised by this case touches the heart of what makes individuals what they are that we should be especially sensitive to the rights of those whose choices upset the majority.

The assertion that “traditional Judeo-Christian values proscribe” the conduct involved, Brief for Petitioner 20, cannot provide an adequate justification for § 16-6-2. That certain, but by no means all, religious groups condemn the behavior at issue gives the State no license to impose their judgments on the entire citizenry. The legitimacy of secular legislation depends instead on whether the State can advance some justification for its law beyond its conformity to religious doctrine. See, e. g., McGowan v. Maryland, 366 U. S. 420, 429-453 (1961); Stone v. Graham, 449 U. S. 39 (1980). Thus, far from buttressing his case, petitioner’s invocation of Leviticus, Romans, St. Thomas Aquinas, and sodomy’s heretical status during the Middle Ages undermines his suggestion that § 16-6-2 represents a legitimate use of secular coercive power.6 A State can no more punish private behavior because *212 of religious intolerance than it can punish such behavior because of racial animus. “The Constitution cannot control such prejudices, but neither can it tolerate them. Private biases may be outside the reach of the law, but the law cannot, directly or indirectly, give them effect.” Palmore v. Sidoti, 466 U. S. 429, 433 (1984). No matter how uncomfortable a certain group may make the majority of this Court, we have held that ”[m]ere public intolerance or animosity cannot constitutionally justify the deprivation of a person’s physical liberty.” O’Connor v. Donaldson, 422 U. S. 563, 575 (1975). See also Cleburne v. Cleburne Living Center, Inc., 473 U. S. 432 (1985); United States Dept. of Agriculture v. Moreno, 413 U. S. 528, 534 (1973).

Nor can § 16-6-2 be justified as a “morally neutral” exercise of Georgia’s power to “protect the public environment,” Paris Adult Theatre I, 413 U. S., at 68-69. Certainly, some private behavior can affect the fabric of society as a whole. Reasonable people may differ about whether particular sexual acts are moral or immoral, but “we have ample evidence for believing that people will not abandon morality, will not think any better of murder, cruelty and dishonesty, merely because some private sexual practice which they abominate is not punished by the law.” H. L. A. Hart, Immorality and Treason, reprinted in The Law as Literature 220, 225 (L. Blom-Cooper ed. 1961). Petitioner and the Court fail to see the difference between laws that protect public sensibilities and those that enforce private morality. Statutes banning *213 public sexual activity are entirely consistent with protecting the individual’s liberty interest in decisions concerning sexual relations: the same recognition that those decisions are intensely private which justifies protecting them from governmental interference can justify protecting individuals from unwilling exposure to the sexual activities of others. But the mere fact that intimate behavior may be punished when it takes place in public cannot dictate how States can regulate intimate behavior that occurs in intimate places. See Paris Adult Theatre I, 413 U. S., at 66, n. 13 (“marital intercourse on a street corner or a theater stage” can be forbidden despite the constitutional protection identified in Griswold v. Connecticut, 381 U. S. 479 (1965)).7

This case involves no real interference with the rights of others, for the mere knowledge that other individuals do not adhere to one’s value system cannot be a legally cognizable interest, cf. Diamond v. Charles, 476 U. S. 54, 65-66 (1986), let alone an interest that can justify invading the houses, hearts, and minds of citizens who choose to live their lives differently.

IV

It took but three years for the Court to see the error in its analysis in Minersville School District v. Gobitis, 310 U. S. *214 586 (1940), and to recognize that the threat to national cohesion posed by a refusal to salute the flag was vastly outweighed by the threat to those same values posed by compelling such a salute. See West Virginia Board of Education v. Barnette, 319 U. S. 624 (1943). I can only hope that here, too, the Court soon will reconsider its analysis and conclude that depriving individuals of the right to choose for themselves how to conduct their intimate relationships poses a far greater threat to the values most deeply rooted in our Nation’s history than tolerance of nonconformity could ever do. Because I think the Court today betrays those values, I dissent.

1

Until 1968, Georgia defined sodomy as “the carnal knowledge and connection against the order of nature, by man with man, or in the same unnatural manner with woman.” Ga. Crim. Code § 26-5901 (1933). In Thompson v. Aldredge, 187 Ga. 467, 200 S. E. 799 (1939), the Georgia Supreme Court held that § 26-5901 did not prohibit lesbian activity. And in Riley v. Garrett, 219 Ga. 345, 133 S. E. 2d 367 (1963), the Georgia Supreme Court held that § 26-5901 did not prohibit heterosexual cunnilingus. Georgia passed the act-specific statute currently in force “perhaps in response to the restrictive court decisions such as Riley,” Note, The Crimes Against Nature, 16 J. Pub. L. 159, 167, n. 47 (1967).

2

In Robinson v. California, 370 U. S. 660 (1962), the Court held that the Eighth Amendment barred convicting a defendant due to his “status” as a narcotics addict, since that condition was “apparently an illness which may be contracted innocently or involuntarily.” Id., at 667. In Powell v. Texas, 392 U. S. 514 (1968), where the Court refused to extend Robinson to punishment of public drunkenness by a chronic alcoholic, one of the factors relied on by JUSTICE MARSHALL, in writing the plurality opinion, was that Texas had not “attempted to regulate appellant’s behavior in the privacy of his own home.” Id., at 532. JUSTICE WHITE wrote separately:

“Analysis of this difficult case is not advanced by preoccupation with the label `condition.’ In Robinson the Court dealt with `a statute which makes the “status” of narcotic addiction a criminal offense . . . .’ 370 U. S., at 666. By precluding criminal conviction for such a `status’ the Court was dealing with a condition brought about by acts remote in time from the application of the criminal sanctions contemplated, a condition which was relatively permanent in duration, and a condition of great magnitude and significance in terms of human behavior and values. . . . If it were necessary to distinguish between `acts’ and `conditions’ for purposes of the Eighth Amendment, I would adhere to the concept of `condition’ implicit in the opinion in Robinson . . . . The proper subject of inquiry is whether volitional acts brought about the `condition’ and whether those acts are sufficiently proximate to the `condition’ for it to be permissible to impose penal sanctions on the `condition.’ ” Id., at 550-551, n. 2.

Despite historical views of homosexuality, it is no longer viewed by mental health professionals as a “disease” or disorder. See Brief for American Psychological Association and American Public Health Association as Amici Curiae 8-11. But, obviously, neither is it simply a matter of deliberate personal election. Homosexual orientation may well form part of the very fiber of an individual’s personality. Consequently, under JUSTICE WHITE’s analysis in Powell, the Eighth Amendment may pose a constitutional barrier to sending an individual to prison for acting on that attraction regardless of the circumstances. An individual’s ability to make constitutionally protected “decisions concerning sexual relations,” Carey v. Population Services International, 431 U. S. 678, 711 (1977) (POWELL, J., concurring in part and concurring in judgment), is rendered empty indeed if he or she is given no real choice but a life without any physical intimacy.

With respect to the Equal Protection Clause’s applicability to § 16-6-2, I note that Georgia’s exclusive stress before this Court on its interest in prosecuting homosexual activity despite the gender-neutral terms of the statute may raise serious questions of discriminatory enforcement, questions that cannot be disposed of before this Court on a motion to dismiss. See Yick Wo v. Hopkins, 118 U. S. 356, 373-374 (1886). The legislature having decided that the sex of the participants is irrelevant to the legality of the acts, I do not see why the State can defend § 16-6-2 on the ground that individuals singled out for prosecution are of the same sex as their partners. Thus, under the circumstances of this case, a claim under the Equal Protection Clause may well be available without having to reach the more controversial question whether homosexuals are a suspect class. See, e. g., Rowland v. Mad River Local School District, 470 U. S. 1009 (1985) (BRENNAN, J., dissenting from denial of certiorari); Note, The Constitutional Status of Sexual Orientation: Homosexuality as a Suspect Classification, 98 Harv. L. Rev. 1285 (1985).

3

Even if a court faced with a challenge to § 16-6-2 were to apply simple rational-basis scrutiny to the statute, Georgia would be required to show an actual connection between the forbidden acts and the ill effects it seeks to prevent. The connection between the acts prohibited by § 16-6-2 and the harms identified by petitioner in his brief before this Court is a subject of hot dispute, hardly amenable to dismissal under Federal Rule of Civil Procedure 12(b)(6). Compare, e. g., Brief for Petitioner 36-37 and Brief for David Robinson, Jr., as Amicus Curiae 23-28, on the one hand, with People v. Onofre, 51 N. Y. 2d 476, 489, 415 N. E. 2d 936, 941 (1980); Brief for the Attorney General of the State of New York, joined by the Attorney General of the State of California, as Amici Curiae 11-14; and Brief for the American Psychological Association and American Public Health Association as Amici Curiae 19-27, on the other.

4

Although I do not think it necessary to decide today issues that are not even remotely before us, it does seem to me that a court could find simple, analytically sound distinctions between certain private, consensual sexual conduct, on the one hand, and adultery and incest (the only two vaguely specific “sexual crimes” to which the majority points, ante, at 196), on the other. For example, marriage, in addition to its spiritual aspects, is a civil contract that entitles the contracting parties to a variety of governmentally provided benefits. A State might define the contractual commitment necessary to become eligible for these benefits to include a commitment of fidelity and then punish individuals for breaching that contract. Moreover, a State might conclude that adultery is likely to injure third persons, in particular, spouses and children of persons who engage in extramarital affairs. With respect to incest, a court might well agree with respondent that the nature of familial relationships renders true consent to incestuous activity sufficiently problematical that a blanket prohibition of such activity is warranted. See Tr. of Oral Arg. 21-22. Notably, the Court makes no effort to explain why it has chosen to group private, consensual homosexual activity with adultery and incest rather than with private, consensual heterosexual activity by unmarried persons or, indeed, with oral or anal sex within marriage.

5

The parallel between Loving and this case is almost uncanny. There, too, the State relied on a religious justification for its law. Compare 388 U. S., at 3 (quoting trial court’s statement that “Almighty God created the races white, black, yellow, malay and red, and he placed them on separate continents. . . . The fact that he separated the races shows that he did not intend for the races to mix”), with Brief for Petitioner 20-21 (relying on the Old and New Testaments and the writings of St. Thomas Aquinas to show that “traditional Judeo-Christian values proscribe such conduct”). There, too, defenders of the challenged statute relied heavily on the fact that when the Fourteenth Amendment was ratified, most of the States had similar prohibitions. Compare Brief for Appellee in Loving v. Virginia, O. T. 1966, No. 395, pp. 28-29, with ante, at 192-194, and n. 6. There, too, at the time the case came before the Court, many of the States still had criminal statutes concerning the conduct at issue. Compare 388 U. S., at 6, n. 5 (noting that 16 States still outlawed interracial marriage), with ante, at 193-194 (noting that 24 States and the District of Columbia have sodomy statutes). Yet the Court held, not only that the invidious racism of Virginia’s law violated the Equal Protection Clause, see 388 U. S., at 7-12, but also that the law deprived the Lovings of due process by denying them the “freedom of choice to marry” that had “long been recognized as one of the vital personal rights essential to the orderly pursuit of happiness by free men.” Id., at 12.

6

The theological nature of the origin of Anglo-American antisodomy statutes is patent. It was not until 1533 that sodomy was made a secular offense in England. 25 Hen. VIII, ch. 6. Until that time, the offense was, in Sir James Stephen’s words, “merely ecclesiastical.” 2J. Stephen, A History of the Criminal Law of England 429-430 (1883). Pollock and Maitland similarly observed that ”[t]he crime against nature . . . was so closely connected with heresy that the vulgar had but one name for both.” 2 F. Pollock & F. Maitland, The History of English Law 554 (1895). The transfer of jurisdiction over prosecutions for sodomy to the secular courts seems primarily due to the alteration of ecclesiastical jurisdiction attendant on England’s break with the Roman Catholic Church, rather than to any new understanding of the sovereign’s interest in preventing or punishing the behavior involved. Cf. 6 E. Coke, Institutes, ch. 10 (4th ed. 1797).

7

At oral argument a suggestion appeared that, while the Fourth Amendment’s special protection of the home might prevent the State from enforcing § 16-6-2 against individuals who engage in consensual sexual activity there, that protection would not make the statute invalid. See Tr. of Oral Arg. 10-11. The suggestion misses the point entirely. If the law is not invalid, then the police can invade the home to enforce it, provided, of course, that they obtain a determination of probable cause from a neutral magistrate. One of the reasons for the Court’s holding in Griswold v. Connecticut, 381 U. S. 479 (1965), was precisely the possibility, and repugnance, of permitting searches to obtain evidence regarding the use of contraceptives. Id., at 485-486. Permitting the kinds of searches that might be necessary to obtain evidence of the sexual activity banned by § 16-6-2 seems no less intrusive, or repugnant. Cf. Winston v. Lee, 470 U. S. 753 (1985); Mary Beth G. v. City of Chicago, 723 F. 2d 1263, 1274 (CA7 1983).

JUSTICE STEVENS, with whom JUSTICE BRENNAN and JUSTICE MARSHALL join, dissenting.

Like the statute that is challenged in this case,1 the rationale of the Court’s opinion applies equally to the prohibited conduct regardless of whether the parties who engage in it are married or unmarried, or are of the same or different sexes.2 Sodomy was condemned as an odious and sinful type of behavior during the formative period of the common law.3 *215 That condemnation was equally damning for heterosexual and homosexual sodomy.4 Moreover, it provided no special exemption for married couples.5 The license to cohabit and to produce legitimate offspring simply did not include any permission to engage in sexual conduct that was considered a “crime against nature.”

The history of the Georgia statute before us clearly reveals this traditional prohibition of heterosexual, as well as homosexual, sodomy.6 Indeed, at one point in the 20th century, Georgia’s law was construed to permit certain sexual conduct between homosexual women even though such conduct was prohibited between heterosexuals.7 The history of the statutes cited by the majority as proof for the proposition that sodomy is not constitutionally protected, ante, at 192-194, *216 and nn. 5 and 6, similarly reveals a prohibition on heterosexual, as well as homosexual, sodomy.8

Because the Georgia statute expresses the traditional view that sodomy is an immoral kind of conduct regardless of the identity of the persons who engage in it, I believe that a proper analysis of its constitutionality requires consideration of two questions: First, may a State totally prohibit the described conduct by means of a neutral law applying without exception to all persons subject to its jurisdiction? If not, may the State save the statute by announcing that it will only enforce the law against homosexuals? The two questions merit separate discussion.

I

Our prior cases make two propositions abundantly clear. First, the fact that the governing majority in a State has traditionally viewed a particular practice as immoral is not a sufficient reason for upholding a law prohibiting the practice; neither history nor tradition could save a law prohibiting miscegenation from constitutional attack.9 Second, individual decisions by married persons, concerning the intimacies of their physical relationship, even when not intended to produce offspring, are a form of “liberty” protected by the Due Process Clause of the Fourteenth Amendment. Griswold v. Connecticut, 381 U. S. 479 (1965). Moreover, this protection extends to intimate choices by unmarried as well as married persons. Carey v. Population Services International, 431 U. S. 678 (1977); Eisenstadt v. Baird, 405 U. S. 438 (1972).

*217 In consideration of claims of this kind, the Court has emphasized the individual interest in privacy, but its decisions have actually been animated by an even more fundamental concern. As I wrote some years ago:

“These cases do not deal with the individual’s interest in protection from unwarranted public attention, comment, or exploitation. They deal, rather, with the individual’s right to make certain unusually important decisions that will affect his own, or his family’s destiny. The Court has referred to such decisions as implicating `basic values,’ as being `fundamental,’ and as being dignified by history and tradition. The character of the Court’s language in these cases brings to mind the origins of the American heritage of freedom — the abiding interest in individual liberty that makes certain state intrusions on the citizen’s right to decide how he will live his own life intolerable. Guided by history, our tradition of respect for the dignity of individual choice in matters of conscience and the restraints implicit in the federal system, federal judges have accepted the responsibility for recognition and protection of these rights in appropriate cases.” Fitzgerald v. Porter Memorial Hospital, 523 F. 2d 716, 719-720 (CA7 1975) (footnotes omitted), cert. denied, 425 U. S. 916 (1976).

Society has every right to encourage its individual members to follow particular traditions in expressing affection for one another and in gratifying their personal desires. It, of course, may prohibit an individual from imposing his will on another to satisfy his own selfish interests. It also may prevent an individual from interfering with, or violating, a legally sanctioned and protected relationship, such as marriage. And it may explain the relative advantages and disadvantages of different forms of intimate expression. But when individual married couples are isolated from observation by others, the way in which they voluntarily choose to conduct their intimate relations is a matter for them — not the *218 State — to decide.10 The essential “liberty” that animated the development of the law in cases like Griswold, Eisenstadt, and Carey surely embraces the right to engage in nonreproductive, sexual conduct that others may consider offensive or immoral.

Paradoxical as it may seem, our prior cases thus establish that a State may not prohibit sodomy within “the sacred precincts of marital bedrooms,” Griswold, 381 U. S., at 485, or, indeed, between unmarried heterosexual adults. Eisenstadt, 405 U. S., at 453. In all events, it is perfectly clear that the State of Georgia may not totally prohibit the conduct proscribed by § 16-6-2 of the Georgia Criminal Code.

II

If the Georgia statute cannot be enforced as it is written — if the conduct it seeks to prohibit is a protected form of liberty for the vast majority of Georgia’s citizens — the State must assume the burden of justifying a selective application of its law. Either the persons to whom Georgia seeks to apply its statute do not have the same interest in “liberty” that others have, or there must be a reason why the State may be permitted to apply a generally applicable law to certain persons that it does not apply to others.

The first possibility is plainly unacceptable. Although the meaning of the principle that “all men are created equal” is not always clear, it surely must mean that every free citizen has the same interest in “liberty” that the members of the majority share. From the standpoint of the individual, the homosexual and the heterosexual have the same interest in deciding how he will live his own life, and, more narrowly, how he will conduct himself in his personal and voluntary *219 associations with his companions. State intrusion into the private conduct of either is equally burdensome.

The second possibility is similarly unacceptable. A policy of selective application must be supported by a neutral and legitimate interest — something more substantial than a habitual dislike for, or ignorance about, the disfavored group. Neither the State nor the Court has identified any such interest in this case. The Court has posited as a justification for the Georgia statute “the presumed belief of a majority of the electorate in Georgia that homosexual sodomy is immoral and unacceptable.” Ante, at 196. But the Georgia electorate has expressed no such belief — instead, its representatives enacted a law that presumably reflects the belief that all sodomy is immoral and unacceptable. Unless the Court is prepared to conclude that such a law is constitutional, it may not rely on the work product of the Georgia Legislature to support its holding. For the Georgia statute does not single out homosexuals as a separate class meriting special disfavored treatment.

Nor, indeed, does not Georgia prosecutor even believe that all homosexuals who violate this statute should be punished. This conclusion is evident from the fact that the respondent in this very case has formally acknowledged in his complaint and in court that he has engaged, and intends to continue to engage, in the prohibited conduct, yet the State has elected not to process criminal charges against him. As JUSTICE POWELL points out, moreover, Georgia’s prohibition on private, consensual sodomy has not been enforced for decades.11 The record of nonenforcement, in this case and in the last several decades, belies the Attorney General’s representations *220 about the importance of the State’s selective application of its generally applicable law.12

Both the Georgia statute and the Georgia prosecutor thus completely fail to provide the Court with any support for the conclusion that homosexual sodomy, simpliciter, is considered unacceptable conduct in that State, and that the burden of justifying a selective application of the generally applicable law has been met.

III

The Court orders the dismissal of respondent’s complaint even though the State’s statute prohibits all sodomy; even though that prohibition is concededly unconstitutional with respect to heterosexuals; and even though the State’s post hoc explanations for selective application are belied by the State’s own actions. At the very least, I think it clear at this early stage of the litigation that respondent has alleged a constitutional claim sufficient to withstand a motion to dismiss.13

I respectfully dissent.

1

See Ga. Code Ann. § 16-6-2(a) (1984) (“A person commits the offense of sodomy when he performs or submits to any sexual act involving the sex organs of one person and the mouth or anus of another”).

2

The Court states that the “issue presented is whether the Federal Constitution confers a fundamental right upon homosexuals to engage in sodomy and hence invalidates the laws of the many States that still make such conduct illegal and have done so for a very long time.” Ante, at 190. In reality, however, it is the indiscriminate prohibition of sodomy, heterosexual as well as homosexual, that has been present “for a very long time.” See nn. 3, 4, and 5, infra. Moreover, the reasoning the Court employs would provide the same support for the statute as it is written as it does for the statute as it is narrowly construed by the Court.

3

See, e. g., 1 W. Hawkins, Pleas of the Crown 9 (6th ed. 1787) (“All unnatural carnal copulations, whether with man or beast, seem to come under the notion of sodomy, which was felony by the antient common law, and punished, according to some authors, with burning; according to others, with burying alive”); 4 W. Blackstone, Commentaries *215 (discussing “the infamous crime against nature, committed either with man or beast; a crime which ought to be strictly and impartially proved, and then as strictly and impartially punished”).

4

See 1 E. East, Pleas of the Crown 480 (1803) (“This offence, concerning which the least notice is the best, consists in a carnal knowledge committed against the order of nature by man with man, or in the same unnatural manner with woman, or by man or woman in any manner with beast”); J. Hawley & M. McGregor, The Criminal Law 287 (3d ed. 1899) (“Sodomy is the carnal knowledge against the order of nature by two persons with each other, or of a human being with a beast. . . . The offense may be committed between a man and a woman, or between two male persons, or between a man or a woman and a beast”).

5

See J. May, The Law of Crimes § 203 (2d ed. 1893) (“Sodomy, otherwise called buggery, bestiality, and the crime against nature, is the unnatural copulation of two persons with each other, or of a human being with a beast. . . . It may be committed by a man with a man, by a man with a beast, or by a woman with a beast, or by a man with a woman — his wife, in which case, if she consent, she is an accomplice”).

6

The predecessor of the current Georgia statute provided: “Sodomy is the carnal knowledge and connection against the order of nature, by man with man, or in the same unnatural manner with woman.” Ga. Code, Tit. 1, Pt. 4, § 4251 (1861). This prohibition of heterosexual sodomy was not purely hortatory. See, e. g., Comer v. State, 21 Ga. App. 306, 94 S. E. 314 (1917) (affirming prosecution for consensual heterosexual sodomy).

7

See Thompson v. Aldredge, 187 Ga. 467, 200 S. E. 799 (1939).

8

A review of the statutes cited by the majority discloses that, in 1791, in 1868, and today, the vast majority of sodomy statutes do not differentiate between homosexual and heterosexual sodomy.

9

See Loving v. Virginia, 388 U. S. 1 (1967). Interestingly, miscegenation was once treated as a crime similar to sodomy. See Hawley & McGregor, The Criminal Law, at 287 (discussing crime of sodomy); id., at 288 (discussing crime of miscegenation).

10

Indeed, the Georgia Attorney General concedes that Georgia’s statute would be unconstitutional if applied to a married couple. See Tr. of Oral Arg. 8 (stating that application of the statute to a married couple “would be unconstitutional” because of the “right of marital privacy as identified by the Court in Griswold”). Significantly, Georgia passed the current statute three years after the Court’s decision in Griswold.

11

Ante, at 198, n. 2 (POWELL, J., concurring). See also Tr. of Oral Arg. 4-5 (argument of Georgia Attorney General) (noting, in response to question about prosecution “where the activity took place in a private residence,” the “last case I can recall was back in the 1930’s or 40’s”).

12

It is, of course, possible to argue that a statute has a purely symbolic role. Cf. Carey v. Population Services International, 431 U. S. 678, 715, n. 3 (1977) (STEVENS, J., concurring in part and concurring in judgment) (“The fact that the State admittedly has never brought a prosecution under the statute . . . is consistent with appellants’ position that the purpose of the statute is merely symbolic”). Since the Georgia Attorney General does not even defend the statute as written, however, see n. 10, supra, the State cannot possibly rest on the notion that the statute may be defended for its symbolic message.

13

Indeed, at this stage, it appears that the statute indiscriminately authorizes a policy of selective prosecution that is neither limited to the class of homosexual persons nor embraces all persons in that class, but rather applies to those who may be arbitrarily selected by the prosecutor for reasons that are not revealed either in the record of this case or in the text of the statute. If that is true, although the text of the statute is clear enough, its true meaning may be “so intolerably vague that evenhanded enforcement of the law is a virtual impossibility.” Marks v. United States, 430 U. S. 188, 198 (1977) (STEVENS, J., concurring in part and dissenting in part).

Roy Romer, Governor of Colorado, et al., Petitioners, v. Richard G. Evans et al.

Argued October 10, 1995.

Decided May 20, 1996.

CERTIORARI TO THE SUPREME COURT OF COLORADO

*621 Kennedy, J., delivered the opinion of the Court, in which Stevens, O'Connor, Souter, Ginsburg, and Breyer, JJ., joined. Scalia, J., filed a dissenting opinion, in which Rehnquist, C. J., and Thomas, J., joined, post, p. 636.

Timothy M. Tymkovich, Solicitor General of Colorado, argued the cause for petitioners. With him on the briefs were Gale A. Norton, Attorney General, Stephen K. ErkenBrack, Chief Deputy Attorney General, John Daniel Dailey and Paul Farley, Deputy Attorneys General, and Rex E. Lee and Carter G. Phillips, Special Assistant Attorneys General.

Jean E. Dubofsky argued the cause for respondents. With her on the brief for respondents Evans et al. were Rod- *622 erick M. Hills, Jr., Matthew Coles, Steven R. Shapiro, Clyde J. Wadsworth, Suzanne B. Goldberg, Jeanne Winer, Gregory A. Eurich, David H. Miller, Darlene M. Ebert, Joseph N. de Raismes III, and Walter A. Smith, Jr. John P. Worcester and Edward M. Caswall filed a brief for respondents City of Aspen et al.*

*

 Briefs of amici curiae urging reversal were filed for the State of Alabama et al. by Charles J. Cooper,and by the Attorneys General for their respective States as follows: Jeff Sessions of Alabama, Daniel E. Lungren of California, Alan G. Lance of Idaho, Don Stenberg of Nebraska, Charles Molony Condon of South Carolina, Mark Barnett of South Dakota, and James S. Gilmore III of Virginia; for the American Center for Law and Justice Family Life Project by Jay Alan Sekulow and Keith A. Fournier; for the Christian Legal Society et al. by Steven T. McFarland, Samuel B. Casey, Gregory S. Baylor, and John K. Hulston Hall; for Colorado for Family Values by Robert K. Skolrood; for Concerned Women for America, Inc., by David J. Myers and Wendell R. Bird; for Equal Rights, Not Special Rights, Inc., by Michael A. Carvin, William L. McGrath, and Robert H. Bork; for the Family Research Council by Melissa Wells-Petry; for the Pacific Legal Foundation by Anthony T. Caso and Deborah J. La Fetra; and for the Oregon Citizens Alliance et al. by Lawrence J. Hall.

Briefs of amici curiae urging affirmance were filed for the State of Oregon et al. by Theodore R. Kulongoski, Attorney General of Oregon, Thomas A. Balmer, Deputy Attorney General, Virginia L. Linder, Solicitor General, Michael D. Reynolds, Assistant Solicitor General, and Rives Kistler, Assistant Attorney General, Thomas J. Miller, Attorney General of Iowa, J. Joseph Curran, Jr., Attorney General of Maryland, Scott Harshbarger, Attorney General of Massachusetts, Hubert H. Humphrey III, Attorney General of Minnesota, Frankie Sue Del Papa, Attorney General of Nevada, Christine O. Gregoire,Attorney General of Washington, and Garland Pinkston, Jr., Acting Corporation Counsel of the District of Columbia; for the City of Atlanta et al. by Louise H. Renne, Dennis Aftergut, Burk E. Delventhal, Julia M. C. Friedlander, Mary Carole Cooney, Robin Joy Shahar, Neal M. Janey, Stephen H. Clark, James K. Hahn, David I. Schulman, Eunice Gibson, Paul A. Crotty, Leonard A. Kerner, Jeffrey L. Rogers, Linda Meng, Janet E. Halley, Mark H. Sidran, Henry W. Underhill, Jr., and Susan S. Sher; for Affirmation: United Methodists for Gay, Lesbian and Bisexual Concerns et al. by Celeste McCollough; for the American Bar Association by George E. Bushnell, Jr.; for the American Association on Mental Retardation et al. by James W. Ellis and Maureen A. Sanders; for The American Federation of State, County and Municipal Employees, AFL—CIO, by John C. Dempsey and Larry P. Weinberg; for the American Friends Service Committee et al. by Stark Ritchie; for the American Psychological Association et al. by Paul M. Smith, James L. McHugh, Jr., and Richard G. Taranto; for the Asian American Legal Defense and Education Fund et al. by Eben Moglen and Pamela S. Karlan; for the Colorado Bar Association et al. by Stephen V. Bomse, Martha Minow, and Frances A. Koncilja; for the Gay and Lesbian Lawyers of Philadelphia by Cletus P. Lyman; for the NAACP Legal Defense and Educational Fund, Inc., et al. by Eric Schnapper, Elaine R. Jones, Theodore M. Shaw, Antonia Hernandez, Judith L. Lichtman, and Donna R. Lenhoff; for the National Bar Association by J. Clay Smith, Jr.; for the National Education Association et al. by Robert H. Chanin and John M. West; for James E. Andrews by Eric J. Graninger; and for Laurence H. Tribe et al. by Mr. Tribe, pro se, John Hart Ely, pro se, Philip B. Kurland, pro se, and Kathleen M. Sullivan, pro se.

Chai R. Feldblum filed a brief for the Human Rights Campaign Fund et al. as amici curiae.

*623 Justice Kennedy, delivered the opinion of the Court.

One century ago, the first Justice Harlan admonished this Court that the Constitution "neither knows nor tolerates classes among citizens." Plessy v. Ferguson, 163 U. S. 537, 559 (1896) (dissenting opinion). Unheeded then, those words now are understood to state a commitment to the law's neutrality where the rights of persons are at stake. The Equal Protection Clause enforces this principle and today requires us to hold invalid a provision of Colorado's Constitution.

I

The enactment challenged in this case is an amendment to the Constitution of the State of Colorado, adopted in a 1992 statewide referendum. The parties and the state courts refer to it as "Amendment 2," its designation when submitted to the voters. The impetus for the amendment and the contentious campaign that preceded its adoption came in large part from ordinances that had been passed in various Colorado municipalities. For example, the cities of Aspen and Boulder and the city and County of Denver each had *624 enacted ordinances which banned discrimination in many transactions and activities, including housing, employment, education, public accommodations, and health and welfare services. Denver Rev. Municipal Code, Art. IV, §§ 28-91 to 28-116 (1991); Aspen Municipal Code § 13-98 (1977); Boulder Rev. Code §§ 12-1—1 to 12-1—11 (1987). What gave rise to the statewide controversy was the protection the ordinances afforded to persons discriminated against by reason of their sexual orientation. See Boulder Rev. Code § 12-1—1 (defining "sexual orientation" as "the choice of sexual partners, i. e., bisexual, homosexual or heterosexual"); Denver Rev. Municipal Code, Art. IV, § 28-92 (defining "sexual orientation" as "[t]he status of an individual as to his or her heterosexuality, homosexuality or bisexuality"). Amendment 2 repeals these ordinances to the extent they prohibit discrimination on the basis of "homosexual, lesbian or bisexual orientation, conduct, practices or relationships." Colo. Const., Art. II, § 30b.

Yet Amendment 2, in explicit terms, does more than repeal or rescind these provisions. It prohibits all legislative, executive or judicial action at any level of state or local government designed to protect the named class, a class we shall refer to as homosexual persons or gays and lesbians. The amendment reads:

"No Protected Status Based on Homosexual, Lesbian or Bisexual Orientation. Neither the State of Colorado, through any of its branches or departments, nor any of its agencies, political subdivisions, municipalities or school districts, shall enact, adopt or enforce any statute, regulation, ordinance or policy whereby homosexual, lesbian or bisexual orientation, conduct, practices or relationships shall constitute or otherwise be the basis of or entitle any person or class of persons to have or claim any minority status, quota preferences, protected status or claim of discrimination. This Section of the Constitution shall be in all respects self-executing." Ibid.

*625 Soon after Amendment 2 was adopted, this litigation to declare its invalidity and enjoin its enforcement was commenced in the District Court for the City and County of Denver. Among the plaintiffs (respondents here) were homosexual persons, some of them government employees. They alleged that enforcement of Amendment 2 would subject them to immediate and substantial risk of discrimination on the basis of their sexual orientation. Other plaintiffs (also respondents here) included the three municipalities whose ordinances we have cited and certain other governmental entities which had acted earlier to protect homosexuals from discrimination but would be prevented by Amendment 2 from continuing to do so. Although Governor Romer had been on record opposing the adoption of Amendment 2, he was named in his official capacity as a defendant, together with the Colorado Attorney General and the State of Colorado.

The trial court granted a preliminary injunction to stay enforcement of Amendment 2, and an appeal was taken to the Supreme Court of Colorado. Sustaining the interim injunction and remanding the case for further proceedings, the State Supreme Court held that Amendment 2 was subject to strict scrutiny under the Fourteenth Amendment because it infringed the fundamental right of gays and lesbians to participate in the political process. Evans v. Romer, 854 P. 2d 1270 (Colo. 1993) (Evans I). To reach this conclusion, the state court relied on our voting rights cases, e. g., Reynolds v. Sims, 377 U. S. 533 (1964); Carrington v. Rash, 380 U. S. 89 (1965); Harper v. Virginia Bd. of Elections, 383 U. S. 663 (1966); Williams v. Rhodes, 393 U. S. 23 (1968), and on our precedents involving discriminatory restructuring of governmental decisionmaking, see, e. g., Hunter v. Erickson, 393 U. S. 385 (1969); Reitman v. Mulkey, 387 U. S. 369 (1967); Washington v. Seattle School Dist. No. 1, 458 U. S. 457 (1982); Gordon v. Lance,403 U. S. 1 (1971). On remand, the State advanced various arguments in an effort to show that *626 Amendment 2 was narrowly tailored to serve compelling interests, but the trial court found none sufficient. It enjoined enforcement of Amendment 2, and the Supreme Court of Colorado, in a second opinion, affirmed the ruling. 882 P. 2d 1335 (1994) (Evans II). We granted certiorari, 513 U. S. 1146 (1995), and now affirm the judgment, but on a rationale different from that adopted by the State Supreme Court.

II

The State's principal argument in defense of Amendment 2 is that it puts gays and lesbians in the same position as all other persons. So, the State says, the measure does no more than deny homosexuals special rights. This reading of the amendment's language is implausible. We rely not upon our own interpretation of the amendment but upon the authoritative construction of Colorado's Supreme Court. The state court, deeming it unnecessary to determine the full extent of the amendment's reach, found it invalid even on a modest reading of its implications. The critical discussion of the amendment, set out in Evans I, is as follows:

"The immediate objective of Amendment 2 is, at a minimum, to repeal existing statutes, regulations, ordinances, and policies of state and local entities that barred discrimination based on sexual orientation. See Aspen, Colo., Mun. Code § 13-98 (1977) (prohibiting discrimination in employment, housing and public accommodations on the basis of sexual orientation); Boulder, Colo., Rev. Code §§ 12-1—2 to —4 (1987) (same); Denver, Colo., Rev. Mun. Code art. IV, §§ 28-91 to —116 (1991) (same); Executive Order No. D0035 (December 10, 1990) (prohibiting employment discrimination for `all state employees, classified and exempt' on the basis of sexual orientation); Colorado Insurance Code, § 10-3—1104, 4A C. R. S. (1992 Supp.) (forbidding health insurance providers from determining insurability and premiums based on an applicant's, a beneficiary's, or an insured's *627 sexual orientation); and various provisions prohibiting discrimination based on sexual orientation at state colleges.26

Sweeping and comprehensive is the change in legal status effected by this law. So much is evident from the ordinances the Colorado Supreme Court declared would be void by operation of Amendment 2. Homosexuals, by state decree, are put in a solitary class with respect to transactions and relations in both the private and governmental spheres. The amendment withdraws from homosexuals, but no others, specific legal protection from the injuries caused by discrimination, and it forbids reinstatement of these laws and policies.

The change Amendment 2 works in the legal status of gays and lesbians in the private sphere is far reaching, both on its own terms and when considered in light of the structure and operation of modern antidiscrimination laws. That structure is well illustrated by contemporary statutes and ordinances prohibiting discrimination by providers of public accommodations. "At common law, innkeepers, smiths, and others who `made profession of a public employment,' were prohibited from refusing, without good reason, to serve a customer." Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U. S. 557, 571 (1995). The duty was a general one and did not specify protection for particular groups. The common-law rules, however, proved *628 insufficient in many instances, and it was settled early that the Fourteenth Amendment did not give Congress a general power to prohibit discrimination in public accommodations, Civil Rights Cases, 109 U. S. 3, 25 (1883). In consequence, most States have chosen to counter discrimination by enacting detailed statutory schemes. See, e. g., S. D. Codified Laws §§ 20-13-10, 20-13-22, 20-13-23 (1995); Iowa Code §§ 216.6-216.8 (1994); Okla. Stat., Tit. 25, §§ 1302, 1402 (1987); 43 Pa. Cons. Stat. §§ 953, 955 (Supp. 1995); N. J. Stat. Ann. §§ 10:5-3, 10:5-4 (West Supp. 1995); N. H. Rev. Stat. Ann. §§ 354—A:7, 354—A:10, 354—A:17 (1995); Minn. Stat. § 363.03 (1991 and Supp. 1995).

Colorado's state and municipal laws typify this emerging tradition of statutory protection and follow a consistent pattern. The laws first enumerate the persons or entities subject to a duty not to discriminate. The list goes well beyond the entities covered by the common law. The Boulder ordinance, for example, has a comprehensive definition of entities deemed places of "public accommodation." They include "any place of business engaged in any sales to the general public and any place that offers services, facilities, privileges, or advantages to the general public or that receives financial support through solicitation of the general public or through governmental subsidy of any kind." Boulder Rev. Code § 12-1—1(j) (1987). The Denver ordinance is of similar breadth, applying, for example, to hotels, restaurants, hospitals, dental clinics, theaters, banks, common carriers, travel and insurance agencies, and "shops and stores dealing with goods or services of any kind," Denver Rev. Municipal Code, Art. IV, § 28-92 (1991).

These statutes and ordinances also depart from the common law by enumerating the groups or persons within their ambit of protection. Enumeration is the essential device used to make the duty not to discriminate concrete and to provide guidance for those who must comply. In following this approach, Colorado's state and local governments have *629 not limited antidiscrimination laws to groups that have so far been given the protection of heightened equal protection scrutiny under our cases. See, e. g., J. E. B. v. Alabama ex rel. T. B., 511 U. S. 127, 135 (1994) (sex); Lalli v. Lalli, 439 U. S. 259, 265 (1978) (illegitimacy); McLaughlin v. Florida, 379 U. S. 184, 191-192 (1964) (race); Oyama v. California, 332 U. S. 633 (1948)(ancestry). Rather, they set forth an extensive catalog of traits which cannot be the basis for discrimination, including age, military status, marital status, pregnancy, parenthood, custody of a minor child, political affiliation, physical or mental disability of an individual or of his or her associates—-and, in recent times, sexual orientation. Aspen Municipal Code § 13-98(a)(1) (1977); Boulder Rev. Code §§ 12-1—1 to 12-1—4 (1987); Denver Rev. Municipal Code, Art. IV, §§ 28-92 to 28-119 (1991); Colo. Rev. Stat. §§ 24-34-401 to 24-34-707 (1988 and Supp. 1995).

Amendment 2 bars homosexuals from securing protection against the injuries that these public-accommodations laws address. That in itself is a severe consequence, but there is more. Amendment 2, in addition, nullifies specific legal protections for this targeted class in all transactions in housing, sale of real estate, insurance, health and welfare services, private education, and employment. See, e. g., Aspen Municipal Code §§ 13-98(b), (c) (1977); Boulder Rev. Code §§ 12-1—2, 12-1—3 (1987); Denver Rev. Municipal Code, Art. IV, §§ 28-93 to 28-95, 28-97 (1991).

Not confined to the private sphere, Amendment 2 also operates to repeal and forbid all laws or policies providing specific protection for gays or lesbians from discrimination by every level of Colorado government. The State Supreme Court cited two examples of protections in the governmental sphere that are now rescinded and may not be reintroduced. The first is Colorado Executive Order D0035 (1990), which forbids employment discrimination against "`all state employees, classified and exempt' on the basis of sexual orientation." 854 P. 2d, at 1284. Also repealed, and now forbidden, *630 are "various provisions prohibiting discrimination based on sexual orientation at state colleges." Id., at 1284, 1285. The repeal of these measures and the prohibition against their future reenactment demonstrate that Amendment 2 has the same force and effect in Colorado's governmental sector as it does elsewhere and that it applies to policies as well as ordinary legislation.

Amendment 2's reach may not be limited to specific laws passed for the benefit of gays and lesbians. It is a fair, if not necessary, inference from the broad language of the amendment that it deprives gays and lesbians even of the protection of general laws and policies that prohibit arbitrary discrimination in governmental and private settings. See, e. g., Colo. Rev. Stat. § 24-4—106(7) (1988) (agency action subject to judicial review under arbitrary and capricious standard); § 18-8—405 (making it a criminal offense for a public servant knowingly, arbitrarily, or capriciously to refrain from performing a duty imposed on him by law); § 10-3—1104(1)(f) (prohibiting "unfair discrimination" in insurance); 4 Colo. Code of Regulations 801-1, Policy 11-1 (1983) (prohibiting discrimination in state employment on grounds of specified traits or "other non-merit factor"). At some point in the systematic administration of these laws, an official must determine whether homosexuality is an arbitrary and, thus, forbidden basis for decision. Yet a decision to that effect would itself amount to a policy prohibiting discrimination on the basis of homosexuality, and so would appear to be no more valid under Amendment 2 than the specific prohibitions against discrimination the state court held invalid.

If this consequence follows from Amendment 2, as its broad language suggests, it would compound the constitutional difficulties the law creates. The state court did not decide whether the amendment has this effect, however, and neither need we. In the course of rejecting the argument that Amendment 2 is intended to conserve resources to fight discrimination against suspect classes, the Colorado Supreme *631 Court made the limited observation that the amendment is not intended to affect many antidiscrimination laws protecting nonsuspect classes, Romer II, 882 P. 2d, at 1346, n. 9. In our view that does not resolve the issue. In any event, even if, as we doubt, homosexuals could find some safe harbor in laws of general application, we cannot accept the view that Amendment 2's prohibition on specific legal protections does no more than deprive homosexuals of special rights. To the contrary, the amendment imposes a special disability upon those persons alone. Homosexuals are forbidden the safeguards that others enjoy or may seek without constraint. They can obtain specific protection against discrimination only by enlisting the citizenry of Colorado to amend the State Constitution or perhaps, on the State's view, by trying to pass helpful laws of general applicability. This is so no matter how local or discrete the harm, no matter how public and widespread the injury. We find nothing special in the protections Amendment 2 withholds. These are protections taken for granted by most people either because they already have them or do not need them; these are protections against exclusion from an almost limitless number of transactions and endeavors that constitute ordinary civic life in a free society.

III

The Fourteenth Amendment's promise that no person shall be denied the equal protection of the laws must coexist with the practical necessity that most legislation classifies for one purpose or another, with resulting disadvantage to various groups or persons. Personnel Administrator of Mass. v. Feeney, 442 U. S. 256, 271-272 (1979); F. S. Royster Guano Co. v. Virginia, 253 U. S. 412, 415 (1920). We have attempted to reconcile the principle with the reality by stating that, if a law neither burdens a fundamental right nor targets a suspect class, we will uphold the legislative classification so long as it bears a rational relation to some legitimate end. See, e. g., Heller v. Doe, 509 U. S. 312, 319-320 (1993).

*632 Amendment 2 fails, indeed defies, even this conventional inquiry. First, the amendment has the peculiar property of imposing a broad and undifferentiated disability on a single named group, an exceptional and, as we shall explain, invalid form of legislation. Second, its sheer breadth is so discontinuous with the reasons offered for it that the amendment seems inexplicable by anything but animus toward the class it affects; it lacks a rational relationship to legitimate state interests.

Taking the first point, even in the ordinary equal protection case calling for the most deferential of standards, we insist on knowing the relation between the classification adopted and the object to be attained. The search for the link between classification and objective gives substance to the Equal Protection Clause; it provides guidance and discipline for the legislature, which is entitled to know what sorts of laws it can pass; and it marks the limits of our own authority. In the ordinary case, a law will be sustained if it can be said to advance a legitimate government interest, even if the law seems unwise or works to the disadvantage of a particular group, or if the rationale for it seems tenuous. See New Orleans v. Dukes, 427 U. S. 297 (1976) (tourism benefits justified classification favoring pushcart vendors of certain longevity); Williamson v. Lee Optical of Okla., Inc., 348 U. S. 483 (1955) (assumed health concerns justified law favoring optometrists over opticians); Railway Express Agency, Inc. v. New York, 336 U. S. 106 (1949)(potential traffic hazards justified exemption of vehicles advertising the owner's products from general advertising ban); Kotch v. Board of River Port Pilot Comm'rs for Port of New Orleans, 330 U. S. 552 (1947) (licensing scheme that disfavored persons unrelated to current river boat pilots justified by possible efficiency and safety benefits of a closely knit pilotage system). The laws challenged in the cases just cited were narrow enough in scope and grounded in a sufficient factual context for us to *633 ascertain some relation between the classification and the purpose it served. By requiring that the classification bear a rational relationship to an independent and legitimate legislative end, we ensure that classifications are not drawn for the purpose of disadvantaging the group burdened by the law. See Railroad Retirement Bd. v. Fritz, 449 U. S. 166, 181 (1980) (Stevens, J., concurring) ("If the adverse impact on the disfavored class is an apparent aim of the legislature, its impartiality would be suspect").

Amendment 2 confounds this normal process of judicial review. It is at once too narrow and too broad. It identifies persons by a single trait and then denies them protection across the board. The resulting disqualification of a class of persons from the right to seek specific protection from the law is unprecedented in our jurisprudence. The absence of precedent for Amendment 2 is itself instructive; "[d]iscriminations of an unusual character especially suggest careful consideration to determine whether they are obnoxious to the constitutional provision." Louisville Gas & Elec. Co. v. Coleman, 277 U. S. 32, 37-38 (1928).

It is not within our constitutional tradition to enact laws of this sort. Central both to the idea of the rule of law and to our own Constitution's guarantee of equal protection is the principle that government and each of its parts remain open on impartial terms to all who seek its assistance. "`Equal protection of the laws is not achieved through indiscriminate imposition of inequalities.' " Sweatt v. Painter, 339 U. S. 629, 635 (1950) (quoting Shelley v. Kraemer, 334 U. S. 1, 22 (1948)). Respect for this principle explains why laws singling out a certain class of citizens for disfavored legal status or general hardships are rare. A law declaring that in general it shall be more difficult for one group of citizens than for all others to seek aid from the government is itself a denial of equal protection of the laws in the most literal sense. "The guaranty of `equal protection of the laws *634 is a pledge of the protection of equal laws.' " Skinner v. Oklahoma ex rel. Williamson, 316 U. S. 535, 541 (1942) (quoting Yick Wo v. Hopkins, 118 U. S. 356, 369 (1886)).

Davis v. Beason, 133 U. S. 333 (1890), not cited by the parties but relied upon by the dissent, is not evidence that Amendment 2 is within our constitutional tradition, and any reliance upon it as authority for sustaining the amendment is misplaced. In Davis, the Court approved an Idaho territorial statute denying Mormons, polygamists, and advocates of polygamy the right to vote and to hold office because, as the Court construed the statute, it "simply excludes from the privilege of voting, or of holding any office of honor, trust or profit, those who have been convicted of certain offences, and those who advocate a practical resistance to the laws of the Territory and justify and approve the commission of crimes forbidden by it." Id., at 347. To the extent Davis held that persons advocating a certain practice may be denied the right to vote, it is no longer good law. Brandenburg v. Ohio, 395 U. S. 444 (1969) (per curiam). To the extent it held that the groups designated in the statute may be deprived of the right to vote because of their status, its ruling could not stand without surviving strict scrutiny, a most doubtful outcome. Dunn v. Blumstein, 405 U. S. 330, 337 (1972); cf. United States v. Brown, 381 U. S. 437 (1965); United States v. Robel, 389 U. S. 258 (1967). To the extent Davis held that a convicted felon may be denied the right to vote, its holding is not implicated by our decision and is unexceptionable. See Richardson v. Ramirez, 418 U. S. 24 (1974).

A second and related point is that laws of the kind now before us raise the inevitable inference that the disadvantage imposed is born of animosity toward the class of persons affected. "[I]f the constitutional conception of `equal protection of the laws' means anything, it must at the very least mean that a bare . . . desire to harm a politically unpopular group cannot constitute a legitimate governmental interest." Department of Agriculture v. Moreno, 413 U. S. 528, 534 *635 (1973). Even laws enacted for broad and ambitious purposes often can be explained by reference to legitimate public policies which justify the incidental disadvantages they impose on certain persons. Amendment 2, however, in making a general announcement that gays and lesbians shall not have any particular protections from the law, inflicts on them immediate, continuing, and real injuries that outrun and belie any legitimate justifications that may be claimed for it. We conclude that, in addition to the far-reaching deficiencies of Amendment 2 that we have noted, the principles it offends, in another sense, are conventional and venerable; a law must bear a rational relationship to a legitimate governmental purpose, Kadrmas v. Dickinson Public Schools, 487 U. S. 450, 462 (1988), and Amendment 2 does not.

The primary rationale the State offers for Amendment 2 is respect for other citizens' freedom of association, and in particular the liberties of landlords or employers who have personal or religious objections to homosexuality. Colorado also cites its interest in conserving resources to fight discrimination against other groups. The breadth of the amendment is so far removed from these particular justifications that we find it impossible to credit them. We cannot say that Amendment 2 is directed to any identifiable legitimate purpose or discrete objective. It is a status-based enactment divorced from any factual context from which we could discern a relationship to legitimate state interests; it is a classification of persons undertaken for its own sake, something the Equal Protection Clause does not permit. "[C]lass legislation . . . [is] obnoxious to the prohibitions of the Fourteenth Amendment . . . ." Civil Rights Cases, 109 U. S., at 24.

We must conclude that Amendment 2 classifies homosexuals not to further a proper legislative end but to make them unequal to everyone else. This Colorado cannot do. A State cannot so deem a class of persons a stranger to its laws. Amendment 2 violates the Equal Protection Clause, *636 and the judgment of the Supreme Court of Colorado is affirmed.

It is so ordered.

26

" Metropolitan State College of Denver prohibits college sponsored social clubs from discriminating in membership on the basis of sexual orientation and Colorado State University has an antidiscrimination policy which encompasses sexual orientation.

"The `ultimate effect' of Amendment 2 is to prohibit any governmental entity from adopting similar, or more protective statutes, regulations, ordinances, or policies in the future unless the state constitution is first amended to permit such measures." 854 P. 2d, at 1284— 1285, and n. 26.

Justice Scalia, with whom The Chief Justice and Justice Thomas join, dissenting.

The Court has mistaken a Kulturkampf for a fit of spite. The constitutional amendment before us here is not the manifestation of a "`bare . . . desire to harm' " homosexuals, ante, at 634, but is rather a modest attempt by seemingly tolerant Coloradans to preserve traditional sexual mores against the efforts of a politically powerful minority to revise those mores through use of the laws. That objective, and the means chosen to achieve it, are not only unimpeachable under any constitutional doctrine hitherto pronounced (hence the opinion's heavy reliance upon principles of righteousness rather than judicial holdings); they have been specifically approved by the Congress of the United States and by this Court.

In holding that homosexuality cannot be singled out for disfavorable treatment, the Court contradicts a decision, unchallenged here, pronounced only 10 years ago, see Bowers v. Hardwick, 478 U. S. 186 (1986), and places the prestige of this institution behind the proposition that opposition to homosexuality is as reprehensible as racial or religious bias. Whether it is or not is precisely the cultural debate that gave rise to the Colorado constitutional amendment (and to the preferential laws against which the amendment was directed). Since the Constitution of the United States says nothing about this subject, it is left to be resolved by normal democratic means, including the democratic adoption of provisions in state constitutions. This Court has no business imposing upon all Americans the resolution favored by the elite class from which the Members of this institution are selected, pronouncing that "animosity" toward homosexuality, ante,at 634, is evil. I vigorously dissent.

*637 I

Let me first discuss Part II of the Court's opinion, its longest section, which is devoted to rejecting the State's arguments that Amendment 2 "puts gays and lesbians in the same position as all other persons," and "does no more than deny homosexuals special rights," ante, at 626. The Court concludes that this reading of Amendment 2's language is "implausible" under the "authoritative construction" given Amendment 2 by the Supreme Court of Colorado. Ibid.

In reaching this conclusion, the Court considers it unnecessary to decide the validity of the State's argument that Amendment 2 does not deprive homosexuals of the "protection [afforded by] general laws and policies that prohibit arbitrary discrimination in governmental and private settings." Ante, at 630. I agree that we need not resolve that dispute, because the Supreme Court of Colorado has resolved it for us. In the case below, 882 P. 2d 1335 (1994), the Colorado court stated:

"[I]t is significant to note that Colorado law currently proscribes discrimination against persons who are not suspect classes, including discrimination based on age, § 24-34-402(1)(a), 10A C. R. S. (1994 Supp.); marital or family status, § 24-34-502(1)(a), 10A C. R. S. (1994 Supp.); veterans' status, § 28-3—506, 11B C. R. S. (1989); and for any legal, off-duty conduct such as smoking tobacco, § 24-34-402.5, 10A C. R. S. (1994 Supp.). Of course Amendment 2 is not intended to have any effect on this legislation, but seeks only to prevent the adoption of anti-discrimination laws intended to protect gays, lesbians, and bisexuals. Id., at 1346, n. 9 (emphasis added).

The Court utterly fails to distinguish this portion of the Colorado court's opinion. Colorado Rev. Stat. § 24-34-402.5 (Supp. 1995), which this passage authoritatively declares not to be affected by Amendment 2, was respondents' primary *638 example of a generally applicable law whose protections would be unavailable to homosexuals under Amendment 2. See Brief for Respondents Evans et al. 11-12. The clear import of the Colorado court's conclusion that it is not affected is that "general laws and policies that prohibit arbitrary discrimination" would continue to prohibit discrimination on the basis of homosexual conduct as well. This analysis, which is fully in accord with (indeed, follows inescapably from) the text of the constitutional provision, lays to rest such horribles, raised in the course of oral argument, as the prospect that assaults upon homosexuals could not be prosecuted. The amendment prohibits special treatment of homosexuals, and nothing more. It would not affect, for example, a requirement of state law that pensions be paid to all retiring state employees with a certain length of service; homosexual employees, as well as others, would be entitled to that benefit. But it would prevent the State or any municipality from making death-benefit payments to the "life partner" of a homosexual when it does not make such payments to the long-time roommate of a nonhomosexual employee. Or again, it does not affect the requirement of the State's general insurance laws that customers be afforded coverage without discrimination unrelated to anticipated risk. Thus, homosexuals could not be denied coverage, or charged a greater premium, with respect to auto collision insurance; but neither the State nor any municipality could require that distinctive health insurance risks associated with homosexuality (if there are any) be ignored.

Despite all of its hand wringing about the potential effect of Amendment 2 on general antidiscrimination laws, the Court's opinion ultimately does not dispute all this, but assumes it to be true. See ante, at 630. The only denial of equal treatment it contends homosexuals have suffered is this: They may not obtain preferential treatment without amending the State Constitution. That is to say, the principle underlying the Court's opinion is that one who is accorded *639 equal treatment under the laws, but cannot as readily as others obtain preferential treatment under the laws, has been denied equal protection of the laws. If merely stating this alleged "equal protection" violation does not suffice to refute it, our constitutional jurisprudence has achieved terminal silliness.

The central thesis of the Court's reasoning is that any group is denied equal protection when, to obtain advantage (or, presumably, to avoid disadvantage), it must have recourse to a more general and hence more difficult level of political decisionmaking than others. The world has never heard of such a principle, which is why the Court's opinion is so long on emotive utterance and so short on relevant legal citation. And it seems to me most unlikely that any multilevel democracy can function under such a principle. For whenever a disadvantage is imposed, or conferral of a benefit is prohibited, at one of the higher levels of democratic decisionmaking (i. e., by the state legislature rather than local government, or by the people at large in the state constitution rather than the legislature), the affected group has (under this theory) been denied equal protection. To take the simplest of examples, consider a state law prohibiting the award of municipal contracts to relatives of mayors or city councilmen. Once such a law is passed, the group composed of such relatives must, in order to get the benefit of city contracts, persuade the state legislature—unlike all other citizens, who need only persuade the municipality. It is ridiculous to consider this a denial of equal protection, which is why the Court's theory is unheard of.

The Court might reply that the example I have given is not a denial of equal protection only because the same "rational basis" (avoidance of corruption) which renders constitutional the substantive discrimination against relatives (i. e., the fact that they alone cannot obtain city contracts) also automatically suffices to sustain what might be called the electoral-procedural discrimination against them (i. e., *640 the fact that they must go to the state level to get this changed). This is of course a perfectly reasonable response, and would explain why "electoral-procedural discrimination" has not hitherto been heard of: A law that is valid in its substance is automatically valid in its level of enactment. But the Court cannot afford to make this argument, for as I shall discuss next, there is no doubt of a rational basis for the substance of the prohibition at issue here. The Court's entire novel theory rests upon the proposition that there is something special —something that cannot be justified by normal "rational basis" analysis—in making a disadvantaged group (or a nonpreferred group) resort to a higher decisionmaking level. That proposition finds no support in law or logic.

II

I turn next to whether there was a legitimate rational basis for the substance of the constitutional amendment—for the prohibition of special protection for homosexuals.1 It is unsurprising that the Court avoids discussion of this question, since the answer is so obviously yes. The case most relevant to the issue before us today is not even mentioned in the Court's opinion: In Bowers v. Hardwick, 478 U. S. 186 (1986), we held that the Constitution does not prohibit what virtually all States had done from the founding of the Republic until very recent years—making homosexual conduct a crime. That holding is unassailable, except by those who *641 think that the Constitution changes to suit current fashions. But in any event it is a given in the present case: Respondents' briefs did not urge overruling Bowers,and at oral argument respondents' counsel expressly disavowed any intent to seek such overruling, Tr. of Oral Arg. 53. If it is constitutionally permissible for a State to make homosexual conduct criminal, surely it is constitutionally permissible for a State to enact other laws merely disfavoring homosexual conduct. (As the Court of Appeals for the District of Columbia Circuit has aptly put it: "If the Court [in Bowers] was unwilling to object to state laws that criminalize the behavior that defines the class, it is hardly open . . . to conclude that state sponsored discrimination against the class is invidious. After all, there can hardly be more palpable discrimination against a class than making the conduct that defines the class criminal." Padula v. Webster, 822 F. 2d 97, 103 (1987).) And a fortiori it is constitutionally permissible for a State to adopt a provision not even disfavoring homosexual conduct, but merely prohibiting all levels of state government from bestowing special protections upon homosexual conduct. Respondents (who, unlike the Court, cannot afford the luxury of ignoring inconvenient precedent) counter Bowers with the argument that a greater-includes-the-lesser rationale cannot justify Amendment 2's application to individuals who do not engage in homosexual acts, but are merely of homosexual "orientation." Some Courts of Appeals have concluded that, with respect to laws of this sort at least, that is a distinction without a difference. See Equality Foundation of Greater Cincinnati, Inc. v. Cincinnati, 54 F. 3d 261, 267 (CA6 1995) ("[F]or purposes of these proceedings, it is virtually impossible to distinguish or separate individuals of a particular orientation which predisposes them toward a particular sexual conduct from those who actually engage in that particular type of sexual conduct"); Steffan v. Perry, 41 F. 3d 677, 689-690 (CADC 1994). The Supreme Court of Colorado itself appears to be of this view. See 882 P. 2d, at *642 1349-1350 ("Amendment 2 targets this class of persons based on four characteristics: sexual orientation; conduct; practices, and relationships. Each characteristic provides a potentially different way of identifying that class of persons who are gay, lesbian, or bisexual. These four characteristics are not truly severable from one another because each provides nothing more than a different way of identifying the same class of persons ") (emphasis added).

But assuming that, in Amendment 2, a person of homosexual "orientation" is someone who does not engage in homosexual conduct but merely has a tendency or desire to do so, Bowers still suffices to establish a rational basis for the provision. If it is rational to criminalize the conduct, surely it is rational to deny special favor and protection to those with a self-avowed tendency or desire to engage in the conduct. Indeed, where criminal sanctions are not involved, homosexual "orientation" is an acceptable stand-in for homosexual conduct. A State "does not violate the Equal Protection Clause merely because the classifications made by its laws are imperfect," Dandridge v. Williams, 397 U. S. 471, 485 (1970). Just as a policy barring the hiring of methadone users as transit employees does not violate equal protection simply because some methadone users pose no threat to passenger safety, see New York City Transit Authority v. Beazer, 440 U. S. 568 (1979), and just as a mandatory retirement age of 50 for police officers does not violate equal protection even though it prematurely ends the careers of many policemen over 50 who still have the capacity to do the job, see Massachusetts Bd. of Retirement v. Murgia, 427 U. S. 307 (1976) (per curiam), Amendment 2 is not constitutionally invalid simply because it could have been drawn more precisely so as to withdraw special antidiscrimination protections only from those of homosexual "orientation" who actually engage in homosexual conduct. As Justice Kennedy wrote, when he was on the Court of Appeals, in a case involving discharge of homosexuals from the Navy: "Nearly any *643 statute which classifies people may be irrational as applied in particular cases. Discharge of the particular plaintiffs before us would be rational, under minimal scrutiny, not because their particular cases present the dangers which justify Navy policy, but instead because the general policy of discharging all homosexuals is rational." Beller v. Middendorf, 632 F. 2d 788, 808-809, n. 20 (CA9 1980) (citation omitted). See also Ben-Shalom v. Marsh, 881 F. 2d 454, 464 (CA7 1989), cert. denied, 494 U. S. 1004 (1990).

Moreover, even if the provision regarding homosexual "orientation" were invalid, respondents' challenge to Amendment 2—which is a facial challenge—must fail. "A facial challenge to a legislative Act is, of course, the most difficult challenge to mount successfully, since the challenger must establish that no set of circumstances exists under which the Act would be valid." United States v. Salerno, 481 U. S. 739, 745 (1987). It would not be enough for respondents to establish (if they could) that Amendment 2 is unconstitutional as applied to those of homosexual "orientation"; since, under Bowers, Amendment 2 is unquestionably constitutional as applied to those who engage in homosexual conduct, the facial challenge cannot succeed. Some individuals of homosexual "orientation" who do not engage in homosexual acts might successfully bring an as-applied challenge to Amendment 2, but so far as the record indicates, none of the respondents is such a person. See App. 4-5 (complaint describing each of the individual respondents as either "a gay man" or "a lesbian").2

*644 III

The foregoing suffices to establish what the Court's failure to cite any case remotely in point would lead one to suspect: No principle set forth in the Constitution, nor even any imagined by this Court in the past 200 years, prohibits what Colorado has done here. But the case for Colorado is much stronger than that. What it has done is not only unprohibited, but eminently reasonable, with close, congressionally approved precedent in earlier constitutional practice.

First, as to its eminent reasonableness. The Court's opinion contains grim, disapproving hints that Coloradans have been guilty of "animus" or "animosity" toward homosexuality, as though that has been established as un-American. 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—murder, for example, or polygamy, or cruelty to animals—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. The Colorado amendment does not, to speak entirely precisely, prohibit giving favored status to people who are homosexuals; they can be favored for many reasons—for example, because they are senior citizens or members of racial minorities. But it prohibits giving them favored status because of their homosexual conduct —that is, it prohibits favored status for homosexuality.

But though Coloradans are, as I say, entitled to be hostile toward homosexual conduct, the fact is that the degree of hostility reflected by Amendment 2 is the smallest conceivable. *645 The Court's portrayal of Coloradans as a society fallen victim to pointless, hate-filled "gay-bashing" is so false as to be comical. Colorado not only is one of the 25 States that have repealed their antisodomy laws, but was among the first to do so. See 1971 Colo. Sess. Laws, ch. 121, § 1. But the society that eliminates criminal punishment for homosexual acts does not necessarily abandon the view that homosexuality is morally wrong and socially harmful; often, abolition simply reflects the view that enforcement of such criminal laws involves unseemly intrusion into the intimate lives of citizens. Cf. Brief for Lambda Legal Defense and Education Fund, Inc., et al. as Amici Curiae in Bowers v. Hardwick,O. T. 1985, No. 85-140, p. 25, n. 21 (antisodomy statutes are "unenforceable by any but the most offensive snooping and wasteful allocation of law enforcement resources"); Kadish, The Crisis of Overcriminalization, 374 The Annals of the American Academy of Political and Social Science 157, 161 (1967) ("To obtain evidence [in sodomy cases], police are obliged to resort to behavior which tends to degrade and demean both themselves personally and law enforcement as an institution").

There is a problem, however, which arises when criminal sanction of homosexuality is eliminated but moral and social disapprobation of homosexuality is meant to be retained. The Court cannot be unaware of that problem; it is evident in many cities of the country, and occasionally bubbles to the surface of the news, in heated political disputes over such matters as the introduction into local schools of books teaching that homosexuality is an optional and fully acceptable "alternative life style." The problem (a problem, that is, for those who wish to retain social disapprobation of homosexuality) is that, because those who engage in homosexual conduct tend to reside in disproportionate numbers in certain communities, see Record, Exh. MMM, have high disposable income, see ibid.;App. 254 (affidavit of Prof. James Hunter), and, of course, care about homosexual-rights issues much *646 more ardently than the public at large, they possess political power much greater than their numbers, both locally and statewide. Quite understandably, they devote this political power to achieving not merely a grudging social toleration, but full social acceptance, of homosexuality. See, e. g., Jacobs, The Rhetorical Construction of Rights: The Case of the Gay Rights Movement, 1969-1991, 72 Neb. L. Rev. 723, 724 (1993) ("[T]he task of gay rights proponents is to move the center of public discourse along a continuum from the rhetoric of disapprobation, to rhetoric of tolerance, and finally to affirmation").

By the time Coloradans were asked to vote on Amendment 2, their exposure to homosexuals' quest for social endorsement was not limited to newspaper accounts of happenings in places such as New York, Los Angeles, San Francisco, and Key West. Three Colorado cities—Aspen, Boulder, and Denver—had enacted ordinances that listed "sexual orientation" as an impermissible ground for discrimination, equating the moral disapproval of homosexual conduct with racial and religious bigotry. See Aspen Municipal Code § 13-98 (1977); Boulder Rev. Municipal Code §§ 12-1—1 to 12-1—11 (1987); Denver Rev. Municipal Code, Art. IV, §§ 28-91 to 28-116 (1991). The phenomenon had even appeared statewide: The Governor of Colorado had signed an executive order pronouncing that "in the State of Colorado we recognize the diversity in our pluralistic society and strive to bring an end to discrimination in any form," and directing state agency-heads to "ensure non-discrimination" in hiring and promotion based on, among other things, "sexual orientation." Executive Order No. D0035 (Dec. 10, 1990). I do not mean to be critical of these legislative successes; homosexuals are as entitled to use the legal system for reinforcement of their moral sentiments as is the rest of society. But they are subject to being countered by lawful, democratic countermeasures as well.

*647 That is where Amendment 2 came in. It sought to counter both the geographic concentration and the disproportionate political power of homosexuals by (1) resolving the controversy at the statewide level, and (2) making the election a single-issue contest for both sides. It put directly, to all the citizens of the State, the question: Should homosexuality be given special protection? They answered no. The Court today asserts that this most democratic of procedures is unconstitutional. Lacking any cases to establish that facially absurd proposition, it simply asserts that it must be unconstitutional, because it has never happened before.

"[Amendment 2] identifies persons by a single trait and then denies them protection across the board. The resulting disqualification of a class of persons from the right to seek specific protection from the law is unprecedented in our jurisprudence. The absence of precedent for Amendment 2 is itself instructive . . . .

"It is not within our constitutional tradition to enact laws of this sort. Central both to the idea of the rule of law and to our own Constitution's guarantee of equal protection is the principle that government and each of its parts remain open on impartial terms to all who seek its assistance." Ante, at 633.

As I have noted above, this is proved false every time a state law prohibiting or disfavoring certain conduct is passed, because such a law prevents the adversely affected group— whether drug addicts, or smokers, or gun owners, or motorcyclists—from changing the policy thus established in "each of [the] parts" of the State. What the Court says is even demonstrably false at the constitutional level. The Eighteenth Amendment to the Federal Constitution, for example, deprived those who drank alcohol not only of the power to alter the policy of prohibition locally or through state legislation, but even of the power to alter it through state constitutional amendment or federal legislation. The *648Establishment Clause of the First Amendment prevents theocrats from having their way by converting their fellow citizens at the local, state, or federal statutory level; as does the Republican Form of Government Clause prevent monarchists.

But there is a much closer analogy, one that involves precisely the effort by the majority of citizens to preserve its view of sexual morality statewide, against the efforts of a geographically concentrated and politically powerful minority to undermine it. The Constitutions of the States of Arizona, Idaho, New Mexico, Oklahoma, and Utah to this day contain provisions stating that polygamy is "forever prohibited." See Ariz. Const., Art. XX, par. 2; Idaho Const., Art. I, § 4; N. M. Const., Art. XXI, § 1; Okla. Const., Art. I, § 2; Utah Const., Art. III, § 1. Polygamists, and those who have a polygamous "orientation," have been "singled out" by these provisions for much more severe treatment than merely denial of favored status; and that treatment can only be changed by achieving amendment of the state constitutions. The Court's disposition today suggests that these provisions are unconstitutional, and that polygamy must be permitted in these States on a state-legislated, or perhaps even local-option, basis—unless, of course, polygamists for some reason have fewer constitutional rights than homosexuals.

The United States Congress, by the way, required the inclusion of these antipolygamy provisions in the Constitutions of Arizona, New Mexico, Oklahoma, and Utah, as a condition of their admission to statehood. See Arizona Enabling Act, 36 Stat. 569; New Mexico Enabling Act, 36 Stat. 558; Oklahoma Enabling Act, 34 Stat. 269; Utah Enabling Act, 28 Stat. 108. (For Arizona, New Mexico, and Utah, moreover, the Enabling Acts required that the antipolygamy provisions be "irrevocable without the consent of the United States and the people of said State"—so that not only were "each of [the] parts" of these States not "open on impartial terms" to polygamists, but even the States as a whole were not; *649polygamists would have to persuade the whole country to their way of thinking.) Idaho adopted the constitutional provision on its own, but the 51st Congress, which admitted Idaho into the Union, found its Constitution to be "republican in form and . . . in conformity with the Constitution of the United States. " Act of Admission of Idaho, 26 Stat. 215 (emphasis added). Thus, this "singling out" of the sexual practices of a single group for statewide, democratic vote—so utterly alien to our constitutional system, the Court would have us believe—has not only happened, but has received the explicit approval of the United States Congress.

I cannot say that this Court has explicitly approved any of these state constitutional provisions; but it has approved a territorial statutory provision that went even further, depriving polygamists of the ability even to achieve a constitutional amendment, by depriving them of the power to vote. In Davis v. Beason, 133 U. S. 333 (1890), Justice Field wrote for a unanimous Court:

"In our judgment, § 501 of the Revised Statutes of Idaho Territory, which provides that `no person . . . who is a bigamist or polygamist or who teaches, advises, counsels, or encourages any person or persons to become bigamists or polygamists, or to commit any other crime defined by law, or to enter into what is known as plural or celestial marriage, or who is a member of any order, organization or association which teaches, advises, counsels, or encourages its members or devotees or any other persons to commit the crime of bigamy or polygamy, or any other crime defined by law . . . is permitted to vote at any election, or to hold any position or office of honor, trust, or profit within this Territory,' is not open to any constitutional or legal objection." Id., at 346-347 (emphasis added).

To the extent, if any, that this opinion permits the imposition of adverse consequences upon mere abstract advocacy of polygamy, *650 it has, of course, been overruled by later cases. See Brandenburg v. Ohio, 395 U. S. 444 (1969) (per curiam). But the proposition that polygamy can be criminalized, and those engaging in that crime deprived of the vote, remains good law. See Richardson v. Ramirez, 418 U. S. 24, 53 (1974). Beason rejected the argument that "such discrimination is a denial of the equal protection of the laws." Brief for Appellant in Davis v. Beason, O. T. 1889, No. 1261, p. 41. Among the Justices joining in that rejection were the two whose views in other cases the Court today treats as equal protection lodestars—Justice Harlan, who was to proclaim in Plessy v. Ferguson,163 U. S. 537, 559 (1896) (dissenting opinion), that the Constitution "neither knows nor tolerates classes among citizens," quoted ante, at 623, and Justice Bradley, who had earlier declared that "class legislation . . . [is] obnoxious to the prohibitions of the Fourteenth Amendment," Civil Rights Cases, 109 U. S. 3, 24 (1883), quoted ante, at 635.3

*651 This Court cited Beason with approval as recently as 1993, in an opinion authored by the same Justice who writes for the Court today. That opinion said: "[A]dverse impact will not always lead to a finding of impermissible targeting. For example, a social harm may have been a legitimate concern of government for reasons quite apart from discrimination. . . . See, e. g., . . . Davis v. Beason, 133 U. S. 333 (1890)." Church of Lukumi Babalu Aye, Inc. v.Hialeah, 508 U. S. 520, 535 (1993). It remains to be explained how § 501 of the Idaho Revised Statutes was not an "impermissible targeting" of polygamists, but (the much more mild) Amendment 2 is an "impermissible targeting" of homosexuals. Has the Court concluded that the perceived social harm of polygamy is a "legitimate concern of government," and the perceived social harm of homosexuality is not?

IV

I strongly suspect that the answer to the last question is yes, which leads me to the last point I wish to make: The Court today, announcing that Amendment 2 "defies . . . conventional [constitutional] inquiry," ante, at 632, and "confounds [the] normal process of judicial review," ante, at 633, employs a constitutional theory heretofore unknown to frustrate Colorado's reasonable effort to preserve traditional American moral values. The Court's stern disapproval of "animosity" towards homosexuality might be compared with what an earlier Court (including the revered Justices Harlan and Bradley) said in Murphy v. Ramsey, 114 U. S. 15 (1885), rejecting a constitutional challenge to a United States statute that denied the franchise in federal territories to those who engaged in polygamous cohabitation:

"[C]ertainly no legislation can be supposed more wholesome and necessary in the founding of a free, self-governing *652 commonwealth, fit to take rank as one of the co-ordinate States of the Union, than that which seeks to establish it on the basis of the idea of the family, as consisting in and springing from the union for life of one man and one woman in the holy estate of matrimony; the sure foundation of all that is stable and noble in our civilization; the best guaranty of that reverent morality which is the source of all beneficent progress in social and political improvement." Id., at 45.

I would not myself indulge in such official praise for heterosexual monogamy, because I think it no business of the courts (as opposed to the political branches) to take sides in this culture war.

But the Court today has done so, not only by inventing a novel and extravagant constitutional doctrine to take the victory away from traditional forces, but even by verbally disparaging as bigotry adherence to traditional attitudes. To suggest, for example, that this constitutional amendment springs from nothing more than "`a bare . . . desire to harm a politically unpopular group,' " ante, at 634, quoting Department of Agriculture v. Moreno, 413 U. S. 528, 534 (1973), is nothing short of insulting. (It is also nothing short of preposterous to call "politically unpopular" a group which enjoys enormous influence in American media and politics, and which, as the trial court here noted, though composing no more than 4% of the population had the support of 46% of the voters on Amendment 2, see App. to Pet. for Cert. C-18.)

When the Court takes sides in the culture wars, it tends to be with the knights rather than the villeins—and more specifically with the Templars, reflecting the views and values of the lawyer class from which the Court's Members are drawn. How that class feels about homosexuality will be evident to anyone who wishes to interview job applicants at virtually any of the Nation's law schools. The interviewer may refuse to offer a job because the applicant is a Republican; because he is an adulterer; because he went to the wrong *653 prep school or belongs to the wrong country club; because he eats snails; because he is a womanizer; because she wears real-animal fur; or even because he hates the Chicago Cubs. But if the interviewer should wish not to be an associate or partner of an applicant because he disapproves of the applicant's homosexuality, then he will have violated the pledge which the Association of American Law Schools requires all its member schools to exact from job interviewers: "assurance of the employer's willingness" to hire homosexuals. Bylaws of the Association of American Law Schools, Inc. § 6-4(b); Executive Committee Regulations of the Association of American Law Schools § 6.19, in 1995 Handbook, Association of American Law Schools. This law-school view of what "prejudices" must be stamped out may be contrasted with the more plebeian attitudes that apparently still prevail in the United States Congress, which has been unresponsive to repeated attempts to extend to homosexuals the protections of federal civil rights laws, see, e. g., Employment NonDiscrimination Act of 1994, S. 2238, 103d Cong., 2d Sess. (1994); Civil Rights Amendments of 1975, H. R. 5452, 94th Cong., 1st Sess. (1975), and which took the pains to exclude them specifically from the Americans with Disabilities Act of 1990, see 42 U. S. C. § 12211(a) (1988 ed., Supp. V).

* * *

Today's opinion has no foundation in American constitutional law, and barely pretends to. The people of Colorado have adopted an entirely reasonable provision which does not even disfavor homosexuals in any substantive sense, but merely denies them preferential treatment. Amendment 2 is designed to prevent piecemeal deterioration of the sexual morality favored by a majority of Coloradans, and is not only an appropriate means to that legitimate end, but a means that Americans have employed before. Striking it down is an act, not of judicial judgment, but of political will. I dissent.

1

 The Court evidently agrees that "rational basis"—the normal test for compliance with the Equal Protection Clause—is the governing standard. The trial court rejected respondents' argument that homosexuals constitute a "suspect" or "quasi-suspect" class, and respondents elected not to appeal that ruling to the Supreme Court of Colorado. See 882 P. 2d 1335, 1341, n. 3 (1994). And the Court implicitly rejects the Supreme Court of Colorado's holding, Evans v. Romer, 854 P. 2d 1270, 1282 (1993), that Amendment 2 infringes upon a "fundamental right" of "independently identifiable class[es]" to "participate equally in the political process." See ante, at 625.

2

 The Supreme Court of Colorado stated: "We hold that the portions of Amendment 2 that would remain if only the provision concerning sexual orientation were stricken are not autonomous and thus, not severable," 882 P. 2d, at 1349. That statement was premised, however, on the proposition that "[the] four characteristics [described in the Amendment—sexual orientation, conduct, practices, and relationships] are not truly severable from one another because each provides nothing more than a different way of identifying the same class of persons. " Id., at 1349-1350 (emphasis added). As I have discussed above, if that premise is true—if the entire class affected by the Amendment takes part in homosexual conduct, practices, and relationships—Bowers alone suffices to answer all constitutional objections. Separate consideration of persons of homosexual "orientation" is necessary only if one believes (as the Supreme Court of Colorado did not) that that is a distinct class.

3

 The Court labors mightily to get around Beason, see ante, at 634, but cannot escape the central fact that this Court found the statute at issue— which went much further than Amendment 2, denying polygamists not merely special treatment but the right to vote —"not open to any constitutional or legal objection," rejecting the appellant's argument (much like the argument of respondents today) that the statute impermissibly "single[d] him out," Brief for Appellant in Davis v. Beason, O. T. 1889, No. 1261, p. 41. The Court adopts my conclusions that (a) insofar as Beason permits the imposition of adverse consequences based upon mere advocacy, it has been overruled by subsequent cases, and (b) insofar as Beason holds that convicted felons may be denied the right to vote, it remains good law. To these conclusions, it adds something new: the claim that "[t]o the extent [Beason] held that the groups designated in the statute may be deprived of the right to vote because of their status, its ruling could not stand without surviving strict scrutiny, a most doubtful outcome." Ante, at 634. But if that is so, it is only because we have declared the right to vote to be a "fundamental political right," see, e. g., Dunn v.Blumstein, 405 U. S. 330, 336 (1972), deprivation of which triggers strict scrutiny. Amendment 2, of course, does not deny the fundamental right to vote, and the Court rejects the Colorado court's view that there exists a fundamental right to participate in the political process. Strict scrutiny is thus not in play here. See ante, at 631. Finally, the Court's suggestion that § 501 of the Revised Statutes of Idaho, and Amendment 2, deny rights on account of "status" (rather than conduct) opens up a broader debate involving the significance of Bowers to this case, a debate which the Court is otherwise unwilling to join, see supra, at 640-643.

John Geddes Lawrence and Tyron Garner, Petitioners, v. Texas.

Argued March 26, 2003.

Decided June 26, 2003.

CERTIORARI TO THE COURT OF APPEALS OF TEXAS, FOURTEENTH DISTRICT

*559 *560 *561 KENNEDY, J., delivered the opinion of the Court, in which STEVENS, SOUTER, GINSBURG, and BREYER, JJ., joined. O'CONNOR, J., filed an opinion concurring in the judgment, post, p. 579. SCALIA, J., filed a dissenting opinion, in which REHNQUIST, C. J., and THOMAS, J., joined, post, p. 586. THOMAS, J., filed a dissenting opinion, post, p. 605.

Paul M. Smith argued the cause for petitioners. With him on the briefs were William M. Hohengarten, Daniel Mach, Mitchell Katine, Ruth E. Harlow, Patricia M. Logue, and Susan L. Sommer.

Charles A. Rosenthal, Jr., argued the cause for respondent. With him on the brief were William J. Delmore III and Scott A. Durfee.*

*

Briefs of amici curiae urging reversal were filed for the Alliance of Baptists et al. by Robert A. Long, Jr., and Thomas L. Cubbage III; for the American Psychological Association et al. by David W. Ogden, Paul R. Q. Wolfson, Richard G. Taranto, Nathalie F. P. Gilfoyle, and Carolyn I. Polowy; for the American Public Health Association et al. by Jeffrey S. Trachtman and Norman C. Simon; for the Cato Institute by Robert A. Levy; for Constitutional Law Professors by Pamela S. Karlan and William B. Rubenstein; for the Human Rights Campaign et al. by Walter Dellinger, Pamela Harris, and Jonathan D. Hacker; for the Log Cabin Republicans et al. by C. Martin Meekins; for the NOW Legal Defense and Education Fund by David C. Codell, Laura W. Brill, and Wendy R. Weiser; for Professors of History by Roy T. Englert, Jr., Alan Untereiner, and Sherri Lynn Wolson; for the Republican Unity Coalition et al. by Erik S. Jaffe; and for Mary Robinson et al. by Harold Hongju Koh and Joseph F. Tringali.

Briefs of amici curiae urging affirmance were filed for the State of Alabama et al. by William H. Pryor, Jr., Attorney General of Alabama, Nathan A. Forrester, Solicitor General, and George M. Weaver, and by the Attorneys General for their respective States as follows: Henry D. McMaster of South Carolina and Mark L. Shurtleff of Utah; for Agudath Israel of America by David Zwiebel; for the American Center for Law and Justice by Jay Alan Sekulow, Stuart J. Roth, Colby M. May, James M. Henderson, Sr., Joel H. Thornton, and Walter M. Weber; for the American Family Association, Inc., et al. by Stephen M. Crampton, Brian Fahling, and Michael J. DePrimo; for the Center for Arizona Policy et al. by Len L. Munsil; for the Center for Law and Justice International by Thomas Patrick Monaghan and John P. Tuskey; for the Center for Marriage Law by Vincent P. McCarthy and Lynn D. Wardle; for the Center for the Original Intent of the Constitution by Michael P. Farris and Jordan W. Lorence; for Concerned Women for America by Janet M. LaRue; for the Family Research Council, Inc., by Robert P. George; for First Principles, Inc., by Ronald D. Ray; for Liberty Counsel by Mathew D. Staver and Rena M. Lindevaldsen; for the Pro Family Law Center et al. by Richard D. Ackerman and Gary G. Kreep; for Public Advocate of the United States et al. by Herbert W. Titus and William J. Olson; for the Texas Eagle Forum et al. by Teresa Stanton Collett; for Texas Legislator Warren Chisum et al. by Kelly Shackelford and Scott Roberts; for the Texas Physicians Resource Council et al. by Glen Lavy; and for United Families International by Paul Benjamin Linton.

Briefs of amici curiae were filed for the American Bar Association by Alfred P. Carlton, Jr., Ruth N. Borenstein, and Beth S. Brinkmann; for the American Civil Liberties Union et al. by Laurence H. Tribe, James D. Esseks, Steven R. Shapiro, and Matthew A. Coles; for the Institute for Justice by William H. Mellor, Clint Bolick, Dana Berliner, and Randy E. Barnett; and for the National Lesbian and Gay Law Association et al. by Chai R. Feldblum, J. Paul Oetken, and Scott Ruskay-Kidd.

*562 JUSTICE KENNEDY delivered the opinion of the Court.

Liberty protects the person from unwarranted government intrusions into a dwelling or other private places. In our tradition the State is not omnipresent in the home. And there are other spheres of our lives and existence, outside the home, where the State should not be a dominant presence. Freedom extends beyond spatial bounds. Liberty presumes an autonomy of self that includes freedom of thought, belief, expression, and certain intimate conduct. The instant case involves liberty of the person both in its spatial and in its more transcendent dimensions.

I

The question before the Court is the validity of a Texas statute making it a crime for two persons of the same sex to engage in certain intimate sexual conduct.

In Houston, Texas, officers of the Harris County Police Department were dispatched to a private residence in response to a reported weapons disturbance. They entered an apartment where one of the petitioners, John Geddes Lawrence, *563 resided. The right of the police to enter does not seem to have been questioned. The officers observed Lawrence and another man, Tyron Garner, engaging in a sexual act. The two petitioners were arrested, held in custody overnight, and charged and convicted before a Justice of the Peace.

The complaints described their crime as "deviate sexual intercourse, namely anal sex, with a member of the same sex (man)." App. to Pet. for Cert. 127a, 139a. The applicable state law is Tex. Penal Code Ann. § 21.06(a) (2003). It provides: "A person commits an offense if he engages in deviate sexual intercourse with another individual of the same sex." The statute defines "[d]eviate sexual intercourse" as follows:

"(A) any contact between any part of the genitals of one person and the mouth or anus of another person; or   "(B) the penetration of the genitals or the anus of another person with an object." § 21.01(1).

The petitioners exercised their right to a trial de novo in Harris County Criminal Court. They challenged the statute as a violation of the Equal Protection Clause of the Fourteenth Amendment and of a like provision of the Texas Constitution. Tex. Const., Art. 1, § 3a. Those contentions were rejected. The petitioners, having entered a plea of nolo contendere, were each fined $200 and assessed court costs of $141.25. App. to Pet. for Cert. 107a-110a.

The Court of Appeals for the Texas Fourteenth District considered the petitioners' federal constitutional arguments under both the Equal Protection and Due Process Clauses of the Fourteenth Amendment. After hearing the case en banc the court, in a divided opinion, rejected the constitutional arguments and affirmed the convictions. 41 S. W. 3d 349 (2001). The majority opinion indicates that the Court of Appeals considered our decision in Bowers v. Hardwick, 478 U. S. 186 (1986), to be controlling on the federal due process aspect of the case. Bowers then being authoritative, this was proper.

*564 We granted certiorari, 537 U. S. 1044 (2002), to consider three questions:

1. Whether petitioners' criminal convictions under the Texas "Homosexual Conduct" law—which criminalizes sexual intimacy by same-sex couples, but not identical behavior by different-sex couples—violate the Fourteenth Amendment guarantee of equal protection of the laws.   2. Whether petitioners' criminal convictions for adult consensual sexual intimacy in the home violate their vital interests in liberty and privacy protected by the Due Process Clause of the Fourteenth Amendment.   3. Whether Bowers v. Hardwick, supra, should be overruled? See Pet. for Cert. i.

The petitioners were adults at the time of the alleged offense. Their conduct was in private and consensual.

II

We conclude the case should be resolved by determining whether the petitioners were free as adults to engage in the private conduct in the exercise of their liberty under the Due Process Clause of the Fourteenth Amendment to the Constitution. For this inquiry we deem it necessary to reconsider the Court's holding in Bowers.

There are broad statements of the substantive reach of liberty under the Due Process Clause in earlier cases, including Pierce v. Society of Sisters, 268 U. S. 510 (1925), and Meyer v. Nebraska, 262 U. S. 390 (1923); but the most pertinent beginning point is our decision in Griswold v. Connecticut, 381 U. S. 479 (1965).

In Griswold the Court invalidated a state law prohibiting the use of drugs or devices of contraception and counseling or aiding and abetting the use of contraceptives. The Court described the protected interest as a right to privacy and *565 placed emphasis on the marriage relation and the protected space of the marital bedroom. Id., at 485.

After Griswold it was established that the right to make certain decisions regarding sexual conduct extends beyond the marital relationship. In Eisenstadt v. Baird, 405 U. S. 438 (1972), the Court invalidated a law prohibiting the distribution of contraceptives to unmarried persons. The case was decided under the Equal Protection Clause, id., at 454; but with respect to unmarried persons, the Court went on to state the fundamental proposition that the law impaired the exercise of their personal rights, ibid. It quoted from the statement of the Court of Appeals finding the law to be in conflict with fundamental human rights, and it followed with this statement of its own:

"It is true that in Griswold the right of privacy in question inhered in the marital relationship. . . . If the right of privacy means anything, it is the right of the individual, married or single, to be free from unwarranted governmental intrusion into matters so fundamentally affecting a person as the decision whether to bear or beget a child." Id., at 453.

The opinions in Griswold and Eisenstadt were part of the background for the decision in Roe v. Wade, 410 U. S. 113 (1973). As is well known, the case involved a challenge to the Texas law prohibiting abortions, but the laws of other States were affected as well. Although the Court held the woman's rights were not absolute, her right to elect an abortion did have real and substantial protection as an exercise of her liberty under the Due Process Clause. The Court cited cases that protect spatial freedom and cases that go well beyond it. Roe recognized the right of a woman to make certain fundamental decisions affecting her destiny and confirmed once more that the protection of liberty under the Due Process Clause has a substantive dimension of fundamental significance in defining the rights of the person.

*566 In Carey v. Population Services Int'l, 431 U. S. 678 (1977), the Court confronted a New York law forbidding sale or distribution of contraceptive devices to persons under 16 years of age. Although there was no single opinion for the Court, the law was invalidated. Both Eisenstadt and Carey, as well as the holding and rationale in Roe, confirmed that the reasoning of Griswold could not be confined to the protection of rights of married adults. This was the state of the law with respect to some of the most relevant cases when the Court considered Bowers v. Hardwick.

The facts in Bowers had some similarities to the instant case. A police officer, whose right to enter seems not to have been in question, observed Hardwick, in his own bedroom, engaging in intimate sexual conduct with another adult male. The conduct was in violation of a Georgia statute making it a criminal offense to engage in sodomy. One difference between the two cases is that the Georgia statute prohibited the conduct whether or not the participants were of the same sex, while the Texas statute, as we have seen, applies only to participants of the same sex. Hardwick was not prosecuted, but he brought an action in federal court to declare the state statute invalid. He alleged he was a practicing homosexual and that the criminal prohibition violated rights guaranteed to him by the Constitution. The Court, in an opinion by Justice White, sustained the Georgia law. Chief Justice Burger and Justice Powell joined the opinion of the Court and filed separate, concurring opinions. Four Justices dissented. 478 U. S., at 199 (opinion of Blackmun, J., joined by Brennan, Marshall, and Stevens, JJ.); id., at 214 (opinion of Stevens, J.,joined by Brennan and Marshall, JJ.).

The Court began its substantive discussion in Bowers as follows: "The issue presented is whether the Federal Constitution confers a fundamental right upon homosexuals to engage in sodomy and hence invalidates the laws of the many States that still make such conduct illegal and have done so *567 for a very long time." Id., at 190. That statement, we now conclude, discloses the Court's own failure to appreciate the extent of the liberty at stake. To say that the issue in Bowers was simply the right to engage in certain sexual conduct demeans the claim the individual put forward, just as it would demean a married couple were it to be said marriage is simply about the right to have sexual intercourse. The laws involved in Bowers and here are, to be sure, statutes that purport to do no more than prohibit a particular sexual act. Their penalties and purposes, though, have more far-reaching consequences, touching upon the most private human conduct, sexual behavior, and in the most private of places, the home. The statutes do seek to control a personal relationship that, whether or not entitled to formal recognition in the law, is within the liberty of persons to choose without being punished as criminals.

This, as a general rule, should counsel against attempts by the State, or a court, to define the meaning of the relationship or to set its boundaries absent injury to a person or abuse of an institution the law protects. It suffices for us to acknowledge that adults may choose to enter upon this relationship in the confines of their homes and their own private lives and still retain their dignity as free persons. When sexuality finds overt expression in intimate conduct with another person, the conduct can be but one element in a personal bond that is more enduring. The liberty protected by the Constitution allows homosexual persons the right to make this choice.

Having misapprehended the claim of liberty there presented to it, and thus stating the claim to be whether there is a fundamental right to engage in consensual sodomy, the Bowers Court said: "Proscriptions against that conduct have ancient roots." Id., at 192. In academic writings, and in many of the scholarly amicus briefs filed to assist the Court in this case, there are fundamental criticisms of the historical premises relied upon by the majority and concurring opinions *568 in Bowers. Brief for Cato Institute as Amicus Curiae 16-17; Brief for American Civil Liberties Union et al. as Amici Curiae 15-21; Brief for Professors of History et al. as Amici Curiae 3-10. We need not enter this debate in the attempt to reach a definitive historical judgment, but the following considerations counsel against adopting the definitive conclusions upon which Bowers placed such reliance.

At the outset it should be noted that there is no longstanding history in this country of laws directed at homosexual conduct as a distinct matter. Beginning in colonial times there were prohibitions of sodomy derived from the English criminal laws passed in the first instance by the Reformation Parliament of 1533. The English prohibition was understood to include relations between men and women as well as relations between men and men. See, e. g., King v. Wiseman, 92 Eng. Rep. 774, 775 (K. B. 1718) (interpreting "mankind" in Act of 1533 as including women and girls). Nineteenth-century commentators similarly read American sodomy, buggery, and crime-against-nature statutes as criminalizing certain relations between men and women and between men and men. See, e. g., 2 J. Bishop, Criminal Law § 1028 (1858); 2 J. Chitty, Criminal Law 47-50 (5th Am. ed. 1847); R. Desty, A Compendium of American Criminal Law 143 (1882); J. May, The Law of Crimes § 203 (2d ed. 1893). The absence of legal prohibitions focusing on homosexual conduct may be explained in part by noting that according to some scholars the concept of the homosexual as a distinct category of person did not emerge until the late 19th century. See, e. g., J. Katz, The Invention of Heterosexuality 10 (1995); J. D'Emilio & E. Freedman, Intimate Matters: A History of Sexuality in America 121 (2d ed. 1997) ("The modern terms homosexuality and heterosexuality do not apply to an era that had not yet articulated these distinctions"). Thus early American sodomy laws were not directed at homosexuals as such but instead sought to prohibit nonprocreative sexual activity more generally. This does not suggest approval of *569 homosexual conduct. It does tend to show that this particular form of conduct was not thought of as a separate category from like conduct between heterosexual persons.

Laws prohibiting sodomy do not seem to have been enforced against consenting adults acting in private. A substantial number of sodomy prosecutions and convictions for which there are surviving records were for predatory acts against those who could not or did not consent, as in the case of a minor or the victim of an assault. As to these, one purpose for the prohibitions was to ensure there would be no lack of coverage if a predator committed a sexual assault that did not constitute rape as defined by the criminal law. Thus the model sodomy indictments presented in a 19th-century treatise, see 2 Chitty, supra, at 49, addressed the predatory acts of an adult man against a minor girl or minor boy. Instead of targeting relations between consenting adults in private, 19th-century sodomy prosecutions typically involved relations between men and minor girls or minor boys, relations between adults involving force, relations between adults implicating disparity in status, or relations between men and animals.

To the extent that there were any prosecutions for the acts in question, 19th-century evidence rules imposed a burden that would make a conviction more difficult to obtain even taking into account the problems always inherent in prosecuting consensual acts committed in private. Under then-prevailing standards, a man could not be convicted of sodomy based upon testimony of a consenting partner, because the partner was considered an accomplice. A partner's testimony, however, was admissible if he or she had not consented to the act or was a minor, and therefore incapable of consent. See, e. g., F. Wharton, Criminal Law 443 (2d ed. 1852); 1 F. Wharton, Criminal Law 512 (8th ed. 1880). The rule may explain in part the infrequency of these prosecutions. In all events that infrequency makes it difficult to say that society approved of a rigorous and systematic *570 punishment of the consensual acts committed in private and by adults. The longstanding criminal prohibition of homosexual sodomy upon which the Bowers decision placed such reliance is as consistent with a general condemnation of nonprocreative sex as it is with an established tradition of prosecuting acts because of their homosexual character.

The policy of punishing consenting adults for private acts was not much discussed in the early legal literature. We can infer that one reason for this was the very private nature of the conduct. Despite the absence of prosecutions, there may have been periods in which there was public criticism of homosexuals as such and an insistence that the criminal laws be enforced to discourage their practices. But far from possessing "ancient roots," Bowers, 478 U. S., at 192, American laws targeting same-sex couples did not develop until the last third of the 20th century. The reported decisions concerning the prosecution of consensual, homosexual sodomy between adults for the years 1880-1995 are not always clear in the details, but a significant number involved conduct in a public place. See Brief for American Civil Liberties Union et al. as Amici Curiae 14-15, and n. 18.

It was not until the 1970's that any State singled out same-sex relations for criminal prosecution, and only nine States have done so. See 1977 Ark. Gen. Acts no. 828; 1983 Kan. Sess. Laws p. 652; 1974 Ky. Acts p. 847; 1977 Mo. Laws p. 687; 1973 Mont. Laws p. 1339; 1977 Nev. Stats. p. 1632; 1989 Tenn. Pub. Acts ch. 591; 1973 Tex. Gen. Laws ch. 399; see also Post v. State, 715 P. 2d 1105 (Okla. Crim. App. 1986) (sodomy law invalidated as applied to different-sex couples). Post-Bowers even some of these States did not adhere to the policy of suppressing homosexual conduct. Over the course of the last decades, States with same-sex prohibitions have moved toward abolishing them. See, e. g., Jegley v. Picado, 349 Ark. 600, 80 S. W. 3d 332 (2002); Gryczan v. State, 283 Mont. 433, 942 P. 2d 112 (1997); Campbell v. Sundquist, 926 S. W. 2d 250 (Tenn. App. 1996); Commonwealth v. Wasson, *571 842 S. W. 2d 487 (Ky. 1992); see also 1993 Nev. Stats. p. 518 (repealing Nev. Rev. Stat. § 201.193).

In summary, the historical grounds relied upon in Bowers are more complex than the majority opinion and the concurring opinion by Chief Justice Burger indicate. Their historical premises are not without doubt and, at the very least, are overstated.

It must be acknowledged, of course, that the Court in Bowers was making the broader point that for centuries there have been powerful voices to condemn homosexual conduct as immoral. The condemnation has been shaped by religious beliefs, conceptions of right and acceptable behavior, and respect for the traditional family. For many persons these are not trivial concerns but profound and deep convictions accepted as ethical and moral principles to which they aspire and which thus determine the course of their lives. These considerations do not answer the question before us, however. The issue is whether the majority may use the power of the State to enforce these views on the whole society through operation of the criminal law. "Our obligation is to define the liberty of all, not to mandate our own moral code." Planned Parenthood of Southeastern Pa. v. Casey, 505 U. S. 833, 850 (1992).

Chief Justice Burger joined the opinion for the Court in Bowers and further explained his views as follows: "Decisions of individuals relating to homosexual conduct have been subject to state intervention throughout the history of Western civilization. Condemnation of those practices is firmly rooted in Judeo-Christian moral and ethical standards." 478 U. S., at 196. As with Justice White's assumptions about history, scholarship casts some doubt on the sweeping nature of the statement by Chief Justice Burger as it pertains to private homosexual conduct between consenting adults. See, e. g., Eskridge, Hardwick and Historiography, 1999 U. Ill. L. Rev. 631, 656. In all events we think that our laws and traditions in the past half century are of *572 most relevance here. These references show an emerging awareness that liberty gives substantial protection to adult persons in deciding how to conduct their private lives in matters pertaining to sex. "[H]istory and tradition are the starting point but not in all cases the ending point of the substantive due process inquiry." County of Sacramento v. Lewis, 523 U. S. 833, 857 (1998) (Kennedy, J., concurring).

This emerging recognition should have been apparent when Bowers was decided. In 1955 the American Law Institute promulgated the Model Penal Code and made clear that it did not recommend or provide for "criminal penalties for consensual sexual relations conducted in private." ALI, Model Penal Code § 213.2, Comment 2, p. 372 (1980). It justified its decision on three grounds: (1) The prohibitions undermined respect for the law by penalizing conduct many people engaged in; (2) the statutes regulated private conduct not harmful to others; and (3) the laws were arbitrarily enforced and thus invited the danger of blackmail. ALI, Model Penal Code, Commentary 277-280 (Tent. Draft No. 4, 1955). In 1961 Illinois changed its laws to conform to the Model Penal Code. Other States soon followed. Brief for Cato Institute as Amicus Curiae 15-16.

In Bowers the Court referred to the fact that before 1961 all 50 States had outlawed sodomy, and that at the time of the Court's decision 24 States and the District of Columbia had sodomy laws. 478 U. S., at 192-193. Justice Powell pointed out that these prohibitions often were being ignored, however. Georgia, for instance, had not sought to enforce its law for decades. Id., at 197-198, n. 2 ("The history of nonenforcement suggests the moribund character today of laws criminalizing this type of private, consensual conduct").

The sweeping references by Chief Justice Burger to the history of Western civilization and to Judeo-Christian moral and ethical standards did not take account of other authorities pointing in an opposite direction. A committee advising the British Parliament recommended in 1957 repeal of laws *573 punishing homosexual conduct. The Wolfenden Report: Report of the Committee on Homosexual Offenses and Prostitution (1963). Parliament enacted the substance of those recommendations 10 years later. Sexual Offences Act 1967, § 1.

Of even more importance, almost five years before Bowers was decided the European Court of Human Rights considered a case with parallels to Bowers and to today's case. An adult male resident in Northern Ireland alleged he was a practicing homosexual who desired to engage in consensual homosexual conduct. The laws of Northern Ireland forbade him that right. He alleged that he had been questioned, his home had been searched, and he feared criminal prosecution. The court held that the laws proscribing the conduct were invalid under the European Convention on Human Rights. Dudgeon v. United Kingdom, 45 Eur. Ct. H. R. (1981) ¶ 52. Authoritative in all countries that are members of the Council of Europe (21 nations then, 45 nations now), the decision is at odds with the premise in Bowers that the claim put forward was insubstantial in our Western civilization.

In our own constitutional system the deficiencies in Bowers became even more apparent in the years following its announcement. The 25 States with laws prohibiting the relevant conduct referenced in the Bowers decision are reduced now to 13, of which 4 enforce their laws only against homosexual conduct. In those States where sodomy is still proscribed, whether for same-sex or heterosexual conduct, there is a pattern of nonenforcement with respect to consenting adults acting in private. The State of Texas admitted in 1994 that as of that date it had not prosecuted anyone under those circumstances. State v. Morales, 869 S. W. 2d 941, 943.

Two principal cases decided after Bowers cast its holding into even more doubt. In Planned Parenthood of Southeastern Pa. v. Casey, 505 U. S. 833 (1992), the Court reaffirmed the substantive force of the liberty protected by the Due Process Clause. The Casey decision again confirmed *574 that our laws and tradition afford constitutional protection to personal decisions relating to marriage, procreation, contraception, family relationships, child rearing, and education. Id., at 851. In explaining the respect the Constitution demands for the autonomy of the person in making these choices, we stated as follows:

"These matters, involving the most intimate and personal choices a person may make in a lifetime, choices central to personal dignity and autonomy, are central to the liberty protected by the Fourteenth Amendment. At the heart of liberty is the right to define one's own concept of existence, of meaning, of the universe, and of the mystery of human life. Beliefs about these matters could not define the attributes of personhood were they formed under compulsion of the State." Ibid.

Persons in a homosexual relationship may seek autonomy for these purposes, just as heterosexual persons do. The decision in Bowers would deny them this right.

The second post-Bowers case of principal relevance is Romer v. Evans, 517 U. S. 620 (1996). There the Court struck down class-based legislation directed at homosexuals as a violation of the Equal Protection Clause. Romer invalidated an amendment to Colorado's Constitution which named as a solitary class persons who were homosexuals, lesbians, or bisexual either by "orientation, conduct, practices or relationships," id., at 624 (internal quotation marks omitted), and deprived them of protection under state antidiscrimination laws. We concluded that the provision was "born of animosity toward the class of persons affected" and further that it had no rational relation to a legitimate governmental purpose. Id., at 634.

As an alternative argument in this case, counsel for the petitioners and some amici contend that Romer provides the basis for declaring the Texas statute invalid under the Equal Protection Clause. That is a tenable argument, but we conclude *575 the instant case requires us to address whether Bowers itself has continuing validity. Were we to hold the statute invalid under the Equal Protection Clause some might question whether a prohibition would be valid if drawn differently, say, to prohibit the conduct both between same-sex and different-sex participants.

Equality of treatment and the due process right to demand respect for conduct protected by the substantive guarantee of liberty are linked in important respects, and a decision on the latter point advances both interests. If protected conduct is made criminal and the law which does so remains unexamined for its substantive validity, its stigma might remain even if it were not enforceable as drawn for equal protection reasons. 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. The central holding of Bowers has been brought in question by this case, and it should be addressed. Its continuance as precedent demeans the lives of homosexual persons.

The stigma this criminal statute imposes, moreover, is not trivial. The offense, to be sure, is but a class C misdemeanor, a minor offense in the Texas legal system. Still, it remains a criminal offense with all that imports for the dignity of the persons charged. The petitioners will bear on their record the history of their criminal convictions. Just this Term we rejected various challenges to state laws requiring the registration of sex offenders. Smith v. Doe, 538 U. S. 84 (2003); Connecticut Dept. of Public Safety v. Doe, 538 U. S. 1 (2003). We are advised that if Texas convicted an adult for private, consensual homosexual conduct under the statute here in question the convicted person would come within the registration laws of at least four States were he or she to be subject to their jurisdiction. Pet. for Cert. 13, and n. 12 (citing Idaho Code §§ 18-8301 to 18-8326 (Cum. Supp. 2002); La. Code Crim. Proc. Ann. §§ 15:540-15:549 *576 (West 2003); Miss. Code Ann. §§ 45-33-21 to 45-33-57 (Lexis 2003); S. C. Code Ann. §§ 23-3-400 to 23-3-490 (West 2002)). This underscores the consequential nature of the punishment and the state-sponsored condemnation attendant to the criminal prohibition. Furthermore, the Texas criminal conviction carries with it the other collateral consequences always following a conviction, such as notations on job application forms, to mention but one example.

The foundations of Bowers have sustained serious erosion from our recent decisions in Casey and Romer. When our precedent has been thus weakened, criticism from other sources is of greater significance. In the United States criticism of Bowers has been substantial and continuing, disapproving of its reasoning in all respects, not just as to its historical assumptions. See, e. g., C. Fried, Order and Law: Arguing the Reagan Revolution—A Firsthand Account 81-84 (1991); R. Posner, Sex and Reason 341-350 (1992). The courts of five different States have declined to follow it in interpreting provisions in their own state constitutions parallel to the Due Process Clause of the Fourteenth Amendment, see Jegley v. Picado, 349 Ark. 600, 80 S. W. 3d 332 (2002); Powell v. State, 270 Ga. 327, 510 S. E. 2d 18, 24 (1998); Gryczan v. State, 283 Mont. 433, 942 P. 2d 112 (1997); Campbell v. Sundquist, 926 S. W. 2d 250 (Tenn. App. 1996); Commonwealth v. Wasson, 842 S. W. 2d 487 (Ky. 1992).

To the extent Bowers relied on values we share with a wider civilization, it should be noted that the reasoning and holding in Bowers have been rejected elsewhere. The European Court of Human Rights has followed not Bowers but its own decision in Dudgeon v. United Kingdom. See P. G. & J. H. v. United Kingdom, App. No. 00044787/98, ¶ 56 (Eur. Ct. H. R., Sept. 25, 2001); Modinos v. Cyprus, 259 Eur. Ct. H. R. (1993); Norris v. Ireland, 142 Eur. Ct. H. R. (1988). Other nations, too, have taken action consistent with an affirmation of the protected right of homosexual adults to engage in intimate, consensual conduct. See Brief for Mary *577 Robinson et al. as Amici Curiae 11-12. The right the petitioners seek in this case has been accepted as an integral part of human freedom in many other countries. There has been no showing that in this country the governmental interest in circumscribing personal choice is somehow more legitimate or urgent.

The doctrine of stare decisis is essential to the respect accorded to the judgments of the Court and to the stability of the law. It is not, however, an inexorable command. Payne v. Tennessee, 501 U. S. 808, 828 (1991) ("Stare decisis is not an inexorable command; rather, it `is a principle of policy and not a mechanical formula of adherence to the latest decision' " (quoting Helvering v. Hallock, 309 U. S. 106, 119 (1940))). In Casey we noted that when a court is asked to overrule a precedent recognizing a constitutional liberty interest, individual or societal reliance on the existence of that liberty cautions with particular strength against reversing course. 505 U. S., at 855-856; see also id., at 844 ("Liberty finds no refuge in a jurisprudence of doubt"). The holding in Bowers, however, has not induced detrimental reliance comparable to some instances where recognized individual rights are involved. Indeed, there has been no individual or societal reliance on Bowers of the sort that could counsel against overturning its holding once there are compelling reasons to do so. Bowers itself causes uncertainty, for the precedents before and after its issuance contradict its central holding.

The rationale of Bowers does not withstand careful analysis. In his dissenting opinion in Bowers Justice Stevens came to these conclusions:

"Our prior cases make two propositions abundantly clear. First, the fact that the governing majority in a State has traditionally viewed a particular practice as immoral is not a sufficient reason for upholding a law prohibiting the practice; neither history nor tradition could save a law prohibiting miscegenation from constitutional *578 attack. Second, individual decisions by married persons, concerning the intimacies of their physical relationship, even when not intended to produce offspring, are a form of `liberty' protected by the Due Process Clause of the Fourteenth Amendment. Moreover, this protection extends to intimate choices by unmarried as well as married persons." 478 U. S., at 216 (footnotes and citations omitted).

JUSTICE STEVENS' analysis, in our view, should have been controlling in Bowers and should control here.

Bowers was not correct when it was decided, and it is not correct today. It ought not to remain binding precedent. Bowers v. Hardwick should be and now is overruled.

The present case does not involve minors. It does not involve persons who might be injured or coerced or who are situated in relationships where consent might not easily be refused. It does not involve public conduct or prostitution. It does not involve whether the government must give formal recognition to any relationship that homosexual persons seek to enter. The case does involve two adults who, with full and mutual consent from each other, engaged in sexual practices common to a homosexual lifestyle. The petitioners are entitled to respect for their private lives. The State cannot demean their existence or control their destiny by making their private sexual conduct a crime. Their right to liberty under the Due Process Clause gives them the full right to engage in their conduct without intervention of the government. "It is a promise of the Constitution that there is a realm of personal liberty which the government may not enter." Casey, supra, at 847. The Texas statute furthers no legitimate state interest which can justify its intrusion into the personal and private life of the individual.

Had those who drew and ratified the Due Process Clauses of the Fifth Amendment or the Fourteenth Amendment known the components of liberty in its manifold possibilities, they might have been more specific. They did not presume *579 to have this insight. They knew times can blind us to certain truths and later generations can see that laws once thought necessary and proper in fact serve only to oppress. As the Constitution endures, persons in every generation can invoke its principles in their own search for greater freedom.

The judgment of the Court of Appeals for the Texas Fourteenth District is reversed, and the case is remanded for further proceedings not inconsistent with this opinion.

It is so ordered.

JUSTICE O'CONNOR, concurring in the judgment.

The Court today overrules Bowers v. Hardwick, 478 U. S. 186 (1986). I joined Bowers, and do not join the Court in overruling it. Nevertheless, I agree with the Court that Texas' statute banning same-sex sodomy is unconstitutional. See Tex. Penal Code Ann. § 21.06 (2003). Rather than relying on the substantive component of the Fourteenth Amendment's Due Process Clause, as the Court does, I base my conclusion on the Fourteenth Amendment's Equal Protection Clause.

The Equal Protection Clause of the Fourteenth Amendment "is essentially a direction that all persons similarly situated should be treated alike." Cleburne v. Cleburne Living Center, Inc., 473 U. S. 432, 439 (1985); see also Plyler v. Doe, 457 U. S. 202, 216 (1982). Under our rational basis standard of review, "legislation is presumed to be valid and will be sustained if the classification drawn by the statute is rationally related to a legitimate state interest." Cleburne v. Cleburne Living Center, supra, at 440; see also Department of Agriculture v. Moreno, 413 U. S. 528, 534 (1973); Romer v. Evans, 517 U. S. 620, 632-633 (1996); Nordlinger v. Hahn, 505 U. S. 1, 11-12 (1992).

Laws such as economic or tax legislation that are scrutinized under rational basis review normally pass constitutional muster, since "the Constitution presumes that even improvident decisions will eventually be rectified by the *580 democratic processes." Cleburne v. Cleburne Living Center, supra, at 440; see also Fitzgerald v. Racing Assn. of Central Iowa, ante, p. 103; Williamson v. Lee Optical of Okla., Inc., 348 U. S. 483 (1955). We have consistently held, however, that some objectives, such as "a bare . . . desire to harm a politically unpopular group," are not legitimate state interests. Department of Agriculture v. Moreno, supra, at 534. See also Cleburne v. Cleburne Living Center, supra, at 446-447; Romer v. Evans, supra, at 632. When a law exhibits such a desire to harm a politically unpopular group, we have applied a more searching form of rational basis review to strike down such laws under the Equal Protection Clause.

We have been most likely to apply rational basis review to hold a law unconstitutional under the Equal Protection Clause where, as here, the challenged legislation inhibits personal relationships. In Department of Agriculture v. Moreno, for example, we held that a law preventing those households containing an individual unrelated to any other member of the household from receiving food stamps violated equal protection because the purpose of the law was to "`discriminate against hippies.'" 413 U. S., at 534. The asserted governmental interest in preventing food stamp fraud was not deemed sufficient to satisfy rational basis review. Id., at 535-538. In Eisenstadt v. Baird, 405 U. S. 438, 447-455 (1972), we refused to sanction a law that discriminated between married and unmarried persons by prohibiting the distribution of contraceptives to single persons. Likewise, in Cleburne v. Cleburne Living Center, supra, we held that it was irrational for a State to require a home for the mentally disabled to obtain a special use permit when other residences—like fraternity houses and apartment buildings—did not have to obtain such a permit. And in Romer v. Evans, we disallowed a state statute that "impos[ed] a broad and undifferentiated disability on a single named group"—specifically, homosexuals. 517 U. S., at 632.

*581 The statute at issue here makes sodomy a crime only if a person "engages in deviate sexual intercourse with another individual of the same sex." Tex. Penal Code Ann. § 21.06(a) (2003). Sodomy between opposite-sex partners, however, is not a crime in Texas. That is, Texas treats the same conduct differently based solely on the participants. Those harmed by this law are people who have a same-sex sexual orientation and thus are more likely to engage in behavior prohibited by § 21.06.

The Texas statute makes homosexuals unequal in the eyes of the law by making particular conduct—and only that conduct—subject to criminal sanction. It appears that prosecutions under Texas' sodomy law are rare. See State v. Morales, 869 S. W. 2d 941, 943 (Tex. 1994) (noting in 1994 that § 21.06 "has not been, and in all probability will not be, enforced against private consensual conduct between adults"). This case shows, however, that prosecutions under § 21.06 do occur. And while the penalty imposed on petitioners in this case was relatively minor, the consequences of conviction are not. It appears that petitioners' convictions, if upheld, would disqualify them from or restrict their ability to engage in a variety of professions, including medicine, athletic training, and interior design. See, e. g., Tex. Occ. Code Ann. § 164.051(a)(2)(B) (2003 Pamphlet) (physician); § 451.251(a)(1) (athletic trainer); § 1053.252(2) (interior designer). Indeed, were petitioners to move to one of four States, their convictions would require them to register as sex offenders to local law enforcement. See, e. g., Idaho Code § 18-8304 (Cum. Supp. 2002); La. Stat. Ann. § 15:542 (West Cum. Supp. 2003); Miss. Code Ann. § 45-33-25 (West 2003); S. C. Code Ann. § 23-3-430 (West Cum. Supp. 2002); cf. ante, at 575-576.

And the effect of Texas' sodomy law is not just limited to the threat of prosecution or consequence of conviction. Texas' sodomy law brands all homosexuals as criminals, thereby making it more difficult for homosexuals to be treated in the same manner as everyone else. Indeed, Texas *582 itself has previously acknowledged the collateral effects of the law, stipulating in a prior challenge to this action that the law "legally sanctions discrimination against [homosexuals] in a variety of ways unrelated to the criminal law," including in the areas of "employment, family issues, and housing." State v. Morales, 826 S. W. 2d 201, 203 (Tex. App. 1992).

Texas attempts to justify its law, and the effects of the law, by arguing that the statute satisfies rational basis review because it furthers the legitimate governmental interest of the promotion of morality. In Bowers, we held that a state law criminalizing sodomy as applied to homosexual couples did not violate substantive due process. We rejected the argument that no rational basis existed to justify the law, pointing to the government's interest in promoting morality. 478 U. S., at 196. The only question in front of the Court in Bowers was whether the substantive component of the Due Process Clause protected a right to engage in homosexual sodomy. Id., at 188, n. 2. Bowers did not hold that moral disapproval of a group is a rational basis under the Equal Protection Clause to criminalize homosexual sodomy when heterosexual sodomy is not punished.

This case raises a different issue than Bowers: whether, under the Equal Protection Clause, moral disapproval is a legitimate state interest to justify by itself a statute that bans homosexual sodomy, but not heterosexual sodomy. It is not. Moral disapproval of this group, like a bare desire to harm the group, is an interest that is insufficient to satisfy rational basis review under the Equal Protection Clause. See, e. g., Department of Agriculture v. Moreno, 413 U. S., at 534; Romer v. Evans, 517 U. S., at 634-635. Indeed, we have never held that moral disapproval, without any other asserted state interest, is a sufficient rationale under the Equal Protection Clause to justify a law that discriminates among groups of persons.

*583 Moral disapproval of a group cannot be a legitimate governmental interest under the Equal Protection Clause because legal classifications must not be "drawn for the purpose of disadvantaging the group burdened by the law." Id., at 633. Texas' invocation of moral disapproval as a legitimate state interest proves nothing more than Texas' desire to criminalize homosexual sodomy. But the Equal Protection Clause prevents a State from creating "a classification of persons undertaken for its own sake." Id., at 635. And because Texas so rarely enforces its sodomy law as applied to private, consensual acts, the law serves more as a statement of dislike and disapproval against homosexuals than as a tool to stop criminal behavior. The Texas sodomy law "raise[s] the inevitable inference that the disadvantage imposed is born of animosity toward the class of persons affected." Id., at 634.

Texas argues, however, that the sodomy law does not discriminate against homosexual persons. Instead, the State maintains that the law discriminates only against homosexual conduct. While it is true that the law applies only to conduct, the conduct targeted by this law is conduct that is closely correlated with being homosexual. Under such circumstances, Texas' sodomy law is targeted at more than conduct. It is instead directed toward gay persons as a class. "After all, there can hardly be more palpable discrimination against a class than making the conduct that defines the class criminal." Id., at 641 (SCALIA, J., dissenting) (internal quotation marks omitted). When a State makes homosexual conduct criminal, and not "deviate sexual intercourse" committed by persons of different sexes, "that declaration in and of itself is an invitation to subject homosexual persons to discrimination both in the public and in the private spheres." Ante, at 575.

Indeed, Texas law confirms that the sodomy statute is directed toward homosexuals as a class. In Texas, calling a person a homosexual is slander per se because the word "homosexual" *584 "impute[s] the commission of a crime." Plumley v. Landmark Chevrolet, Inc., 122 F. 3d 308, 310 (CA5 1997) (applying Texas law); see also Head v. Newton, 596 S. W. 2d 209, 210 (Tex. App. 1980). The State has admitted that because of the sodomy law, being homosexual carries the presumption of being a criminal. See State v. Morales, 826 S. W. 2d, at 202-203 ("[T]he statute brands lesbians and gay men as criminals and thereby legally sanctions discrimination against them in a variety of ways unrelated to the criminal law"). Texas' sodomy law therefore results in discrimination against homosexuals as a class in an array of areas outside the criminal law. See ibid. In Romer v. Evans, we refused to sanction a law that singled out homosexuals "for disfavored legal status." 517 U. S., at 633. The same is true here. The Equal Protection Clause "`neither knows nor tolerates classes among citizens.'" Id., at 623 (quoting Plessy v. Ferguson, 163 U. S. 537, 559 (1896) (Harlan, J., dissenting)).

A State can of course assign certain consequences to a violation of its criminal law. But the State cannot single out one identifiable class of citizens for punishment that does not apply to everyone else, with moral disapproval as the only asserted state interest for the law. The Texas sodomy statute subjects homosexuals to "a lifelong penalty and stigma. A legislative classification that threatens the creation of an underclass . . . cannot be reconciled with" the Equal Protection Clause. Plyler v. Doe, 457 U. S., at 239 (Powell, J., concurring).

Whether a sodomy law that is neutral both in effect and application, see Yick Wo v. Hopkins, 118 U. S. 356 (1886), would violate the substantive component of the Due Process Clause is an issue that need not be decided today. I am confident, however, that so long as the Equal Protection Clause requires a sodomy law to apply equally to the private consensual conduct of homosexuals and heterosexuals alike, such a *585 law would not long stand in our democratic society. In the words of Justice Jackson:

"The framers of the Constitution knew, and we should not forget today, that there is no more effective practical guaranty against arbitrary and unreasonable government than to require that the principles of law which officials would impose upon a minority be imposed generally. Conversely, nothing opens the door to arbitrary action so effectively as to allow those officials to pick and choose only a few to whom they will apply legislation and thus to escape the political retribution that might be visited upon them if larger numbers were affected." Railway Express Agency, Inc. v. New York, 336 U. S. 106, 112-113 (1949) (concurring opinion).

That this law as applied to private, consensual conduct is unconstitutional under the Equal Protection Clause does not mean that other laws distinguishing between heterosexuals and homosexuals would similarly fail under rational basis review. Texas cannot assert any legitimate state interest here, such as national security or preserving the traditional institution of marriage. Unlike the moral disapproval of same-sex relations—the asserted state interest in this case— other reasons exist to promote the institution of marriage beyond mere moral disapproval of an excluded group.

A law branding one class of persons as criminal based solely on the State's moral disapproval of that class and the conduct associated with that class runs contrary to the values of the Constitution and the Equal Protection Clause, under any standard of review. I therefore concur in the Court's judgment that Texas' sodomy law banning "deviate sexual intercourse" between consenting adults of the same sex, but not between consenting adults of different sexes, is unconstitutional.

*586 JUSTICE SCALIA, with whom THE CHIEF JUSTICE and JUSTICE THOMAS join, dissenting.

"Liberty finds no refuge in a jurisprudence of doubt." Planned Parenthood of Southeastern Pa. v. Casey, 505 U. S. 833, 844 (1992). That was the Court's sententious response, barely more than a decade ago, to those seeking to overrule Roe v. Wade, 410 U. S. 113 (1973). The Court's response today, to those who have engaged in a 17-year crusade to overrule Bowers v. Hardwick, 478 U. S. 186 (1986), is very different. The need for stability and certainty presents no barrier.

Most of the rest of today's opinion has no relevance to its actual holding—that the Texas statute "furthers no legitimate state interest which can justify" its application to petitioners under rational-basis review. Ante, at 578 (overruling Bowers to the extent it sustained Georgia's antisodomy statute under the rational-basis test). Though there is discussion of "fundamental proposition[s]," ante, at 565, and "fundamental decisions," ibid., nowhere does the Court's opinion declare that homosexual sodomy is a "fundamental right" under the Due Process Clause; nor does it subject the Texas law to the standard of review that would be appropriate (strict scrutiny) if homosexual sodomy were a "fundamental right." Thus, while overruling the outcome of Bowers, the Court leaves strangely untouched its central legal conclusion: "[R]espondent would have us announce . . . a fundamental right to engage in homosexual sodomy. This we are quite unwilling to do." 478 U. S., at 191. Instead the Court simply describes petitioners' conduct as "an exercise of their liberty"—which it undoubtedly is—and proceeds to apply an unheard-of form of rational-basis review that will have far-reaching implications beyond this case. Ante, at 564.

I

I begin with the Court's surprising readiness to reconsider a decision rendered a mere 17 years ago in Bowers v. Hardwick. *587 I do not myself believe in rigid adherence to stare decisis in constitutional cases; but I do believe that we should be consistent rather than manipulative in invoking the doctrine. Today's opinions in support of reversal do not bother to distinguish—or indeed, even bother to mention— the paean to stare decisis coauthored by three Members of today's majority in Planned Parenthood v. Casey. There, when stare decisis meant preservation of judicially invented abortion rights, the widespread criticism of Roe was strong reason to reaffirm it:

"Where, in the performance of its judicial duties, the Court decides a case in such a way as to resolve the sort of intensely divisive controversy reflected in Roe[,] . . . its decision has a dimension that the resolution of the normal case does not carry. . . . [T]o overrule under fire in the absence of the most compelling reason . . . would subvert the Court's legitimacy beyond any serious question." 505 U. S., at 866-867.

Today, however, the widespread opposition to Bowers, a decision resolving an issue as "intensely divisive" as the issue in Roe, is offered as a reason in favor of overruling it. See ante, at 576-577. Gone, too, is any "enquiry" (of the sort conducted in Casey) into whether the decision sought to be overruled has "proven `unworkable,'" Casey, supra, at 855.

Today's approach to stare decisis invites us to overrule an erroneously decided precedent (including an "intensely divisive" decision) if: (1) its foundations have been "ero[ded]" by subsequent decisions, ante, at 576; (2) it has been subject to "substantial and continuing" criticism, ibid.; and (3) it has not induced "individual or societal reliance" that counsels against overturning, ante, at 577. The problem is that Roe itself—which today's majority surely has no disposition to overrule—satisfies these conditions to at least the same degree as Bowers.

*588 (1) A preliminary digressive observation with regard to the first factor: The Court's claim that Planned Parenthood v. Casey, supra, "casts some doubt" upon the holding in Bowers (or any other case, for that matter) does not withstand analysis. Ante, at 571. As far as its holding is concerned, Casey provided a less expansive right to abortion than did Roe, which was already on the books when Bowers was decided. And if the Court is referring not to the holding of Casey, but to the dictum of its famed sweet-mystery-of-life passage, ante, at 574 ("`At the heart of liberty is the right to define one's own concept of existence, of meaning, of the universe, and of the mystery of human life'"): That "casts some doubt" upon either the totality of our jurisprudence or else (presumably the right answer) nothing at all. I have never heard of a law that attempted to restrict one's "right to define" certain concepts; and if the passage calls into question the government's power to regulate actions based on one's self-defined "concept of existence, etc.," it is the passage that ate the rule of law.

I do not quarrel with the Court's claim that Romer v. Evans, 517 U. S. 620 (1996), "eroded" the "foundations" of Bowers' rational-basis holding. See Romer, supra, at 640-643 (Scalia, J., dissenting). But Roe and Casey have been equally "eroded" by Washington v. Glucksberg, 521 U. S. 702, 721 (1997), which held that only fundamental rights which are "`deeply rooted in this Nation's history and tradition' " qualify for anything other than rational-basis scrutiny under the doctrine of "substantive due process." Roe and Casey, of course, subjected the restriction of abortion to heightened scrutiny without even attempting to establish that the freedom to abort was rooted in this Nation's tradition.

(2) Bowers, the Court says, has been subject to "substantial and continuing [criticism], disapproving of its reasoning in all respects, not just as to its historical assumptions." Ante, at 576. Exactly what those nonhistorical criticisms are, and whether the Court even agrees with them, are left *589 unsaid, although the Court does cite two books. See ibid. (citing C. Fried, Order and Law: Arguing the Reagan Revolution —A Firsthand Account 81-84 (1991); R. Posner, Sex and Reason 341-350 (1992)).1 Of course, Roe too (and by extension Casey) had been (and still is) subject to unrelenting criticism, including criticism from the two commentators cited by the Court today. See Fried, supra, at 75 ("Roe was a prime example of twisted judging"); Posner, supra, at 337 ("[The Court's] opinion in Roe... fails to measure up to professional expectations regarding judicial opinions"); Posner, Judicial Opinion Writing, 62 U. Chi. L. Rev. 1421, 1434 (1995) (describing the opinion in Roe as an "embarrassing performanc[e]").

(3) That leaves, to distinguish the rock-solid, unamendable disposition of Roe from the readily overrulable Bowers, only the third factor. "[T]here has been," the Court says, "no individual or societal reliance on Bowers of the sort that could counsel against overturning its holding...." Ante, at 577. It seems to me that the "societal reliance" on the principles confirmed in Bowers and discarded today has been overwhelming. Countless judicial decisions and legislative enactments have relied on the ancient proposition that a governing majority's belief that certain sexual behavior is "immoral and unacceptable" constitutes a rational basis for regulation. See, e. g., Williams v. Pryor, 240 F. 3d 944, 949 (CA11 2001) (citing Bowers in upholding Alabama's prohibition on the sale of sex toys on the ground that "[t]he crafting and safeguarding of public morality . . . indisputably is a legitimate government interest under rational basis scrutiny"); Milner v. Apfel, 148 F. 3d 812, 814 (CA7 1998) (citing Bowers for the proposition that "[l]egislatures are permitted to legislate with regard to morality . . . rather than confined *590 to preventing demonstrable harms"); Holmes v. California Army National Guard, 124 F. 3d 1126, 1136 (CA9 1997) (relying on Bowers in upholding the federal statute and regulations banning from military service those who engage in homosexual conduct); Owens v. State, 352 Md. 663, 683, 724 A. 2d 43, 53 (1999) (relying on Bowers in holding that "a person has no constitutional right to engage in sexual intercourse, at least outside of marriage"); Sherman v. Henry, 928 S. W. 2d 464, 469-473 (Tex. 1996) (relying on Bowers in rejecting a claimed constitutional right to commit adultery). We ourselves relied extensively on Bowers when we concluded, in Barnes v. Glen Theatre, Inc., 501 U. S. 560, 569 (1991), that Indiana's public indecency statute furthered "a substantial government interest in protecting order and morality," ibid. (plurality opinion); see also id., at 575 (Scalia, J., concurring in judgment). State laws against bigamy, same-sex marriage, adult incest, prostitution, masturbation, adultery, fornication, bestiality, and obscenity are likewise sustainable only in light of Bowers' validation of laws based on moral choices. Every single one of these laws is called into question by today's decision; the Court makes no effort to cabin the scope of its decision to exclude them from its holding. See ante, at 572 (noting "an emerging awareness that liberty gives substantial protection to adult persons in deciding how to conduct their private lives in matters pertaining to sex" (emphasis added)). The impossibility of distinguishing homosexuality from other traditional "morals" offenses is precisely why Bowers rejected the rational-basis challenge. "The law," it said, "is constantly based on notions of morality, and if all laws representing essentially moral choices are to be invalidated under the Due Process Clause, the courts will be very busy indeed." 478 U. S., at 196.2

*591 What a massive disruption of the current social order, therefore, the overruling of Bowers entails. Not so the overruling of Roe, which would simply have restored the regime that existed for centuries before 1973, in which the permissibility of, and restrictions upon, abortion were determined legislatively State by State. Casey, however, chose to base its stare decisis determination on a different "sort" of reliance. "[P]eople," it said, "have organized intimate relationships and made choices that define their views of themselves and their places in society, in reliance on the availability of abortion in the event that contraception should fail." 505 U. S., at 856. This falsely assumes that the consequence of overruling Roe would have been to make abortion unlawful. It would not; it would merely have permitted *592 the States to do so. Many States would unquestionably have declined to prohibit abortion, and others would not have prohibited it within six months (after which the most significant reliance interests would have expired). Even for persons in States other than these, the choice would not have been between abortion and childbirth, but between abortion nearby and abortion in a neighboring State.

To tell the truth, it does not surprise me, and should surprise no one, that the Court has chosen today to revise the standards of stare decisis set forth in Casey. It has thereby exposed Casey's extraordinary deference to precedent for the result-oriented expedient that it is.

II

Having decided that it need not adhere to stare decisis, the Court still must establish that Bowers was wrongly decided and that the Texas statute, as applied to petitioners, is unconstitutional.

Texas Penal Code Ann. § 21.06(a) (2003) undoubtedly imposes constraints on liberty. So do laws prohibiting prostitution, recreational use of heroin, and, for that matter, working more than 60 hours per week in a bakery. But there is no right to "liberty" under the Due Process Clause, though today's opinion repeatedly makes that claim. Ante, at 567 ("The liberty protected by the Constitution allows homosexual persons the right to make this choice"); ante, at 574 ("`These matters . . . are central to the liberty protected by the Fourteenth Amendment' "); ante, at 578 ("Their right to liberty under the Due Process Clause gives them the full right to engage in their conduct without intervention of the government"). The Fourteenth Amendment expressly allows States to deprive their citizens of "liberty," so long as "due process of law" is provided:

"No state shall . . . deprive any person of life, liberty, or property, without due process of law." Amdt. 14 (emphasis added).

*593 Our opinions applying the doctrine known as "substantive due process" hold that the Due Process Clause prohibits States from infringing fundamental liberty interests, unless the infringement is narrowly tailored to serve a compelling state interest. Washington v. Glucksberg, 521 U. S., at 721. We have held repeatedly, in cases the Court today does not overrule, that only fundamental rights qualify for this so-called "heightened scrutiny" protection—that is, rights which are "`deeply rooted in this Nation's history and tradition,'" ibid. See Reno v. Flores, 507 U. S. 292, 303 (1993) (fundamental liberty interests must be "so rooted in the traditions and conscience of our people as to be ranked as fundamental" (internal quotation marks and citations omitted)); United States v. Salerno, 481 U. S. 739, 751 (1987) (same). See also Michael H. v. Gerald D., 491 U. S. 110, 122 (1989) ("[W]e have insisted not merely that the interest denominated as a `liberty' be `fundamental' . . . but also that it be an interest traditionally protected by our society"); Moore v. East Cleveland, 431 U. S. 494, 503 (1977) (plurality opinion); Meyer v. Nebraska, 262 U. S. 390, 399 (1923) (Fourteenth Amendment protects "those privileges long recognized at common law as essential to the orderly pursuit of happiness by free men" (emphasis added)).3 All other liberty interests may be abridged or abrogated pursuant to a validly enacted state law if that law is rationally related to a legitimate state interest.

*594 Bowers held, first, that criminal prohibitions of homosexual sodomy are not subject to heightened scrutiny because they do not implicate a "fundamental right" under the Due Process Clause, 478 U. S., at 191-194. Noting that "[p]roscriptions against that conduct have ancient roots," id., at 192, that "[s]odomy was a criminal offense at common law and was forbidden by the laws of the original 13 States when they ratified the Bill of Rights," ibid., and that many States had retained their bans on sodomy, id., at 193, Bowers concluded that a right to engage in homosexual sodomy was not "`deeply rooted in this Nation's history and tradition,'" id., at 192.

The Court today does not overrule this holding. Not once does it describe homosexual sodomy as a "fundamental right" or a "fundamental liberty interest," nor does it subject the Texas statute to strict scrutiny. Instead, having failed to establish that the right to homosexual sodomy is "`deeply rooted in this Nation's history and tradition,'" the Court concludes that the application of Texas's statute to petitioners' conduct fails the rational-basis test, and overrules Bowers' holding to the contrary, see id., at 196. "The Texas statute furthers no legitimate state interest which can justify its intrusion into the personal and private life of the individual." Ante, at 578.

I shall address that rational-basis holding presently. First, however, I address some aspersions that the Court casts upon Bowers' conclusion that homosexual sodomy is not a "fundamental right"—even though, as I have said, the Court does not have the boldness to reverse that conclusion.

III

The Court's description of "the state of the law" at the time of Bowers only confirms that Bowers was right. Ante, at 566. The Court points to Griswold v. Connecticut, 381 U. S. 479, 481-482 (1965). But that case expressly disclaimed any reliance on the doctrine of "substantive due *595 process," and grounded the so-called "right to privacy" in penumbras of constitutional provisions other than the Due Process Clause. Eisenstadt v. Baird, 405 U. S. 438 (1972), likewise had nothing to do with "substantive due process"; it invalidated a Massachusetts law prohibiting the distribution of contraceptives to unmarried persons solely on the basis of the Equal Protection Clause. Of course Eisenstadt contains well-known dictum relating to the "right to privacy," but this referred to the right recognized in Griswold —a right penumbral to the specific guarantees in the Bill of Rights, and not a "substantive due process" right.

Roe v. Wade recognized that the right to abort an unborn child was a "fundamental right" protected by the Due Process Clause. 410 U. S., at 155. The Roe Court, however, made no attempt to establish that this right was "`deeply rooted in this Nation's history and tradition' "; instead, it based its conclusion that "the Fourteenth Amendment's concept of personal liberty . . . is broad enough to encompass a woman's decision whether or not to terminate her pregnancy" on its own normative judgment that antiabortion laws were undesirable. See id., at 153. We have since rejected Roe's holding that regulations of abortion must be narrowly tailored to serve a compelling state interest, see Planned Parenthood v. Casey, 505 U. S., at 876 ( joint opinion of O'CONNOR, KENNEDY, and SOUTER, JJ.); id., at 951-953 (Rehnquist, C. J., concurring in judgment in part and dissenting in part)—and thus, by logical implication, Roe's holding that the right to abort an unborn child is a "fundamental right." See 505 U. S., at 843-912 ( joint opinion of O'CONNOR, KENNEDY, and SOUTER, JJ.) (not once describing abortion as a "fundamental right" or a "fundamental liberty interest").

After discussing the history of antisodomy laws, ante, at 568-571, the Court proclaims that, "it should be noted that there is no longstanding history in this country of laws directed at homosexual conduct as a distinct matter," ante, *596 at 568. This observation in no way casts into doubt the "definitive [historical] conclusio[n]," ibid., on which Bowers relied: that our Nation has a longstanding history of laws prohibiting sodomy in general—regardless of whether it was performed by same-sex or opposite-sex couples:

"It is obvious to us that neither of these formulations would extend a fundamental right to homosexuals to engage in acts of consensual sodomy. Proscriptions against that conduct have ancient roots. Sodomy was a criminal offense at common law and was forbidden by the laws of the original 13 States when they ratified the Bill of Rights. In 1868, when the Fourteenth Amendment was ratified, all but 5 of the 37 States in the Union had criminal sodomy laws. In fact, until 1961, all 50 States outlawed sodomy, and today, 24 States and the District of Columbia continue to provide criminal penalties for sodomy performed in private and between consenting adults. Against this background, to claim that a right to engage in such conduct is `deeply rooted in this Nation's history and tradition' or `implicit in the concept of ordered liberty' is, at best, facetious." 478 U. S., at 192-194 (citations and footnotes omitted; emphasis added).

It is (as Bowers recognized) entirely irrelevant whether the laws in our long national tradition criminalizing homosexual sodomy were "directed at homosexual conduct as a distinct matter." Ante, at 568. Whether homosexual sodomy was prohibited by a law targeted at same-sex sexual relations or by a more general law prohibiting both homosexual and heterosexual sodomy, the only relevant point is that it was criminalized—which suffices to establish that homosexual sodomy is not a right "deeply rooted in our Nation's history and tradition." The Court today agrees that homosexual sodomy was criminalized and thus does not dispute the facts on which Bowers actually relied.

*597 Next the Court makes the claim, again unsupported by any citations, that "[l]aws prohibiting sodomy do not seem to have been enforced against consenting adults acting in private." Ante, at 569. The key qualifier here is "acting in private"—since the Court admits that sodomy laws were enforced against consenting adults (although the Court contends that prosecutions were "infrequen[t]," ibid.). I do not know what "acting in private" means; surely consensual sodomy, like heterosexual intercourse, is rarely performed on stage. If all the Court means by "acting in private" is "on private premises, with the doors closed and windows covered," it is entirely unsurprising that evidence of enforcement would be hard to come by. (Imagine the circumstances that would enable a search warrant to be obtained for a residence on the ground that there was probable cause to believe that consensual sodomy was then and there occurring.) Surely that lack of evidence would not sustain the proposition that consensual sodomy on private premises with the doors closed and windows covered was regarded as a "fundamental right," even though all other consensual sodomy was criminalized. There are 203 prosecutions for consensual, adult homosexual sodomy reported in the West Reporting system and official state reporters from the years 1880-1995. See W. Eskridge, Gaylaw: Challenging the Apartheid of the Closet 375 (1999) (hereinafter Gaylaw). There are also records of 20 sodomy prosecutions and 4 executions during the colonial period. J. Katz, Gay/Lesbian Almanac 29, 58, 663 (1983). Bowers' conclusion that homosexual sodomy is not a fundamental right "deeply rooted in this Nation's history and tradition" is utterly unassailable.

Realizing that fact, the Court instead says: "[W]e think that our laws and traditions in the past half century are of most relevance here. These references show an emerging awareness that liberty gives substantial protection to adult persons in deciding how to conduct their private lives in matters pertaining to sex." Ante, at 571-572 (emphasis *598 added). Apart from the fact that such an "emerging awareness" does not establish a "fundamental right," the statement is factually false. States continue to prosecute all sorts of crimes by adults "in matters pertaining to sex": prostitution, adult incest, adultery, obscenity, and child pornography. Sodomy laws, too, have been enforced "in the past half century," in which there have been 134 reported cases involving prosecutions for consensual, adult, homosexual sodomy. Gaylaw 375. In relying, for evidence of an "emerging recognition," upon the American Law Institute's 1955 recommendation not to criminalize "`consensual sexual relations conducted in private,'" ante, at 572, the Court ignores the fact that this recommendation was "a point of resistance in most of the states that considered adopting the Model Penal Code." Gaylaw 159.

In any event, an "emerging awareness" is by definition not "deeply rooted in this Nation's history and tradition[s]," as we have said "fundamental right" status requires. Constitutional entitlements do not spring into existence because some States choose to lessen or eliminate criminal sanctions on certain behavior. Much less do they spring into existence, as the Court seems to believe, because foreign nations decriminalize conduct. The Bowers majority opinion never relied on "values we share with a wider civilization," ante, at 576, but rather rejected the claimed right to sodomy on the ground that such a right was not "`deeply rooted in this Nation's history and tradition,'" 478 U. S., at 193-194 (emphasis added). Bowers' rational-basis holding is likewise devoid of any reliance on the views of a "wider civilization," see id., at 196. The Court's discussion of these foreign views (ignoring, of course, the many countries that have retained criminal prohibitions on sodomy) is therefore meaningless dicta. Dangerous dicta, however, since "this Court... should not impose foreign moods, fads, or fashions on Americans." Foster v. Florida, 537 U. S. 990, n. (2002) (Thomas, J., concurring in denial of certiorari).

*599

IV

I turn now to the ground on which the Court squarely rests its holding: the contention that there is no rational basis for the law here under attack. This proposition is so out of accord with our jurisprudence—indeed, with the jurisprudence of any society we know—that it requires little discussion.

The Texas statute undeniably seeks to further the belief of its citizens that certain forms of sexual behavior are "immoral and unacceptable," Bowers, supra, at 196—the same interest furthered by criminal laws against fornication, bigamy, adultery, adult incest, bestiality, and obscenity. Bowers held that this was a legitimate state interest. The Court today reaches the opposite conclusion. The Texas statute, it says, "furthers no legitimate state interest which can justify its intrusion into the personal and private life of the individual," ante, at 578 (emphasis added). The Court embraces instead Justice Stevens' declaration in his Bowers dissent, that "`the fact that the governing majority in a State has traditionally viewed a particular practice as immoral is not a sufficient reason for upholding a law prohibiting the practice,'" ante, at 577. This effectively decrees the end of all morals legislation. If, as the Court asserts, the promotion of majoritarian sexual morality is not even a legitimate state interest, none of the above-mentioned laws can survive rational-basis review.

V

Finally, I turn to petitioners' equal-protection challenge, which no Member of the Court save Justice O'Connor, ante, at 579 (opinion concurring in judgment), embraces: On its face § 21.06(a) applies equally to all persons. Men and women, heterosexuals and homosexuals, are all subject to its prohibition of deviate sexual intercourse with someone of the same sex. To be sure, § 21.06 does distinguish between the sexes insofar as concerns the partner with whom the sexual *600 acts are performed: men can violate the law only with other men, and women only with other women. But this cannot itself be a denial of equal protection, since it is precisely the same distinction regarding partner that is drawn in state laws prohibiting marriage with someone of the same sex while permitting marriage with someone of the opposite sex.

The objection is made, however, that the antimiscegenation laws invalidated in Loving v. Virginia, 388 U. S. 1, 8 (1967), similarly were applicable to whites and blacks alike, and only distinguished between the races insofar as the partner was concerned. In Loving, however, we correctly applied heightened scrutiny, rather than the usual rational-basis review, because the Virginia statute was "designed to maintain White Supremacy." Id., at 6, 11. A racially discriminatory purpose is always sufficient to subject a law to strict scrutiny, even a facially neutral law that makes no mention of race. See Washington v. Davis, 426 U. S. 229, 241-242 (1976). No purpose to discriminate against men or women as a class can be gleaned from the Texas law, so rational-basis review applies. That review is readily satisfied here by the same rational basis that satisfied it in Bowers —society's belief that certain forms of sexual behavior are "immoral and unacceptable," 478 U. S., at 196. This is the same justification that supports many other laws regulating sexual behavior that make a distinction based upon the identity of the partner—for example, laws against adultery, fornication, and adult incest, and laws refusing to recognize homosexual marriage.

JUSTICE O'CONNOR argues that the discrimination in this law which must be justified is not its discrimination with regard to the sex of the partner but its discrimination with regard to the sexual proclivity of the principal actor.

"While it is true that the law applies only to conduct, the conduct targeted by this law is conduct that is closely correlated with being homosexual. Under such circumstances, Texas' sodomy law is targeted at more than conduct. *601 It is instead directed toward gay persons as a class." Ante, at 583.

Of course the same could be said of any law. A law against public nudity targets "the conduct that is closely correlated with being a nudist," and hence "is targeted at more than conduct"; it is "directed toward nudists as a class." But be that as it may. Even if the Texas law does deny equal protection to "homosexuals as a class," that denial still does not need to be justified by anything more than a rational basis, which our cases show is satisfied by the enforcement of traditional notions of sexual morality.

JUSTICE O'CONNOR simply decrees application of "a more searching form of rational basis review" to the Texas statute. Ante, at 580. The cases she cites do not recognize such a standard, and reach their conclusions only after finding, as required by conventional rational-basis analysis, that no conceivable legitimate state interest supports the classification at issue. See Romer v. Evans, 517 U. S., at 635; Cleburne v. Cleburne Living Center, Inc., 473 U. S. 432, 448-450 (1985); Department of Agriculture v. Moreno, 413 U. S. 528, 534-538 (1973). Nor does JUSTICE O'CONNOR explain precisely what her "more searching form" of rational-basis review consists of. It must at least mean, however, that laws exhibiting "a desire to harm a politically unpopular group," ante, at 580, are invalid even though there may be a conceivable rational basis to support them.

This reasoning leaves on pretty shaky grounds state laws limiting marriage to opposite-sex couples. JUSTICE O'CONNOR seeks to preserve them by the conclusory statement that "preserving the traditional institution of marriage" is a legitimate state interest. Ante, at 585. But "preserving the traditional institution of marriage" is just a kinder way of describing the State's moral disapproval of same-sex couples. Texas's interest in § 21.06 could be recast in similarly euphemistic terms: "preserving the traditional sexual mores of our society." In the jurisprudence JUSTICE O'CONNOR *602 has seemingly created, judges can validate laws by characterizing them as "preserving the traditions of society" (good); or invalidate them by characterizing them as "expressing moral disapproval" (bad).

* * *

Today's opinion is the product of a Court, which is the product of a law-profession culture, that has largely signed on to the so-called homosexual agenda, by which I mean the agenda promoted by some homosexual activists directed at eliminating the moral opprobrium that has traditionally attached to homosexual conduct. I noted in an earlier opinion the fact that the American Association of Law Schools (to which any reputable law school must seek to belong) excludes from membership any school that refuses to ban from its job-interview facilities a law firm (no matter how small) that does not wish to hire as a prospective partner a person who openly engages in homosexual conduct. See Romer, supra, at 653.

One of the most revealing statements in today's opinion is the Court's grim warning that the criminalization of homosexual conduct is "an invitation to subject homosexual persons to discrimination both in the public and in the private spheres." Ante, at 575. It is clear from this that the Court has taken sides in the culture war, departing from its role of assuring, as neutral observer, that the democratic rules of engagement are observed. Many Americans do not want persons who openly engage in homosexual conduct as partners in their business, as scoutmasters for their children, as teachers in their children's schools, or as boarders in their home. They view this as protecting themselves and their families from a lifestyle that they believe to be immoral and destructive. The Court views it as "discrimination" which it is the function of our judgments to deter. So imbued is the Court with the law profession's anti-anti-homosexual culture, that it is seemingly unaware that the attitudes of that *603 culture are not obviously "mainstream"; that in most States what the Court calls "discrimination" against those who engage in homosexual acts is perfectly legal; that proposals to ban such "discrimination" under Title VII have repeatedly been rejected by Congress, see Employment Non-Discrimination Act of 1994, S. 2238, 103d Cong., 2d Sess. (1994); Civil Rights Amendments, H. R. 5452, 94th Cong., 1st Sess. (1975); that in some cases such "discrimination" is mandated by federal statute, see 10 U. S. C. § 654(b)(1) (mandating discharge from the Armed Forces of any service member who engages in or intends to engage in homosexual acts); and that in some cases such "discrimination" is a constitutional right, see Boy Scouts of America v. Dale, 530 U. S. 640 (2000).

Let me be clear that I have nothing against homosexuals, or any other group, promoting their agenda through normal democratic means. Social perceptions of sexual and other morality change over time, and every group has the right to persuade its fellow citizens that its view of such matters is the best. That homosexuals have achieved some success in that enterprise is attested to by the fact that Texas is one of the few remaining States that criminalize private, consensual homosexual acts. But persuading one's fellow citizens is one thing, and imposing one's views in absence of democratic majority will is something else. I would no more require a State to criminalize homosexual acts—or, for that matter, display any moral disapprobation of them—than I would forbid it to do so. What Texas has chosen to do is well within the range of traditional democratic action, and its hand should not be stayed through the invention of a brand-new "constitutional right" by a Court that is impatient of democratic change. It is indeed true that "later generations can see that laws once thought necessary and proper in fact serve only to oppress," ante, at 579; and when that happens, later generations can repeal those laws. But it is the premise of our system that those judgments are to be made *604 by the people, and not imposed by a governing caste that knows best.

One of the benefits of leaving regulation of this matter to the people rather than to the courts is that the people, unlike judges, need not carry things to their logical conclusion. The people may feel that their disapprobation of homosexual conduct is strong enough to disallow homosexual marriage, but not strong enough to criminalize private homosexual acts—and may legislate accordingly. The Court today pretends that it possesses a similar freedom of action, so that we need not fear judicial imposition of homosexual marriage, as has recently occurred in Canada (in a decision that the Canadian Government has chosen not to appeal). See Halpern v. Toronto, 2003 WL 34950 (Ontario Ct. App.); Cohen, Dozens in Canada Follow Gay Couple's Lead, Washington Post, June 12, 2003, p. A25. At the end of its opinion—after having laid waste the foundations of our rational-basis jurisprudence —the Court says that the present case "does not involve whether the government must give formal recognition to any relationship that homosexual persons seek to enter." Ante, at 578. Do not believe it. More illuminating than this bald, unreasoned disclaimer is the progression of thought displayed by an earlier passage in the Court's opinion, which notes the constitutional protections afforded to "personal decisions relating to marriage, procreation, contraception, family relationships, child rearing, and education," and then declares that "[p]ersons in a homosexual relationship may seek autonomy for these purposes, just as heterosexual persons do." Ante, at 574 (emphasis added). Today's opinion dismantles the structure of constitutional law that has permitted a distinction to be made between heterosexual and homosexual unions, insofar as formal recognition in marriage is concerned. If moral disapprobation of homosexual conduct is "no legitimate state interest" for purposes of proscribing that conduct, ante, at 578; and if, as the Court coos (casting aside all pretense of neutrality), "[w]hen *605 sexuality finds overt expression in intimate conduct with another person, the conduct can be but one element in a personal bond that is more enduring," ante, at 567; what justification could there possibly be for denying the benefits of marriage to homosexual couples exercising "[t]he liberty protected by the Constitution," ibid.? Surely not the encouragement of procreation, since the sterile and the elderly are allowed to marry. This case "does not involve" the issue of homosexual marriage only if one entertains the belief that principle and logic have nothing to do with the decisions of this Court. Many will hope that, as the Court comfortingly assures us, this is so.

The matters appropriate for this Court's resolution are only three: Texas's prohibition of sodomy neither infringes a "fundamental right" (which the Court does not dispute), nor is unsupported by a rational relation to what the Constitution considers a legitimate state interest, nor denies the equal protection of the laws. I dissent.

1

This last-cited critic of Bowers actually writes: "[Bowers] is correct nevertheless that the right to engage in homosexual acts is not deeply rooted in America's history and tradition." Posner, Sex and Reason, at 343.

2

While the Court does not overrule Bowers' holding that homosexual sodomy is not a "fundamental right," it is worth noting that the "societal reliance" upon that aspect of the decision has been substantial as well. See 10 U. S. C. § 654(b)(1) ("A member of the armed forces shall be separated from the armed forces . . . if . . . the member has engaged in . . . a homosexual act or acts"); Marcum v. McWhorter, 308 F. 3d 635, 640-642 (CA6 2002) (relying on Bowers in rejecting a claimed fundamental right to commit adultery); Mullins v. Oregon, 57 F. 3d 789, 793-794 (CA9 1995) (relying on Bowers in rejecting a grandparent's claimed "fundamental liberty interes[t]" in the adoption of her grandchildren); Doe v. Wigginton, 21 F. 3d 733, 739-740 (CA6 1994) (relying on Bowers in rejecting a prisoner's claimed "fundamental right" to on-demand HIV testing); Schowengerdt v. United States, 944 F. 2d 483, 490 (CA9 1991) (relying on Bowers in upholding a bisexual's discharge from the armed services); Charles v. Baesler, 910 F. 2d 1349, 1353 (CA6 1990) (relying on Bowers in rejecting fire department captain's claimed "fundamental" interest in a promotion); Henne v. Wright, 904 F. 2d 1208, 1214-1215 (CA8 1990) (relying on Bowers in rejecting a claim that state law restricting surnames that could be given to children at birth implicates a "fundamental right"); Walls v. Petersburg, 895 F. 2d 188, 193 (CA4 1990) (relying on Bowers in rejecting substantive-due-process challenge to a police department questionnaire that asked prospective employees about homosexual activity); High Tech Gays v. Defense Industrial Security Clearance Office, 895 F. 2d 563, 570-571 (CA9 1988) (relying on Bowers' holding that homosexual activity is not a fundamental right in rejecting—on the basis of the rational-basis standard—an equal protection challenge to the Defense Department's policy of conducting expanded investigations into backgrounds of gay and lesbian applicants for secret and top-secret security clearances).

3

The Court is quite right that "`[h]istory and tradition are the starting point but not in all cases the ending point of the substantive due process inquiry,'" ante, at 572. An asserted "fundamental liberty interest" must not only be "`deeply rooted in this Nation's history and tradition,'" Washington v. Glucksberg, 521 U. S. 702, 721 (1997), but it must also be "`implicit in the concept of ordered liberty,'" so that "`neither liberty nor justice would exist if [it] were sacrificed,'" ibid.Moreover, liberty interests unsupported by history and tradition, though not deserving of "heightened scrutiny," are still protected from state laws that are not rationally related to any legitimate state interest. Id., at 722. As I proceed to discuss, it is this latter principle that the Court applies in the present case.

JUSTICE THOMAS, dissenting.

I join JUSTICE SCALIA'S dissenting opinion. I write separately to note that the law before the Court today "is . . . uncommonly silly." Griswold v. Connecticut, 381 U. S. 479, 527 (1965) (Stewart, J., dissenting). If I were a member of the Texas Legislature, I would vote to repeal it. Punishing someone for expressing his sexual preference through noncommercial consensual conduct with another adult does not appear to be a worthy way to expend valuable law enforcement resources.

Notwithstanding this, I recognize that as a Member of this Court I am not empowered to help petitioners and others similarly situated. My duty, rather, is to "decide cases `agreeably to the Constitution and laws of the United States.'" Id., at 530. And, just like Justice Stewart, I "can find [neither in the Bill of Rights nor any other part of the *606 Constitution a] general right of privacy," ibid., or as the Court terms it today, the "liberty of the person both in its spatial and more transcendent dimensions," ante, at 562.

Hillary Goodridge & others1 vs. Department of Public Health & another.2

Suffolk.

March 4, 2003.

November 18, 2003.

Present: Marshall, C.J., Greaney, Ireland, Spina, Cowin, Sosman, & Cordy, JJ.

*310Mary Lisa Bonauto (Gary D. Buseck with her) for Hillary Goodridge.

Judith S. Yogman, Assistant Attorney General, for Department of Public Health.

The following submitted briefs for amici curiae:

Joseph P.J. Vrabel, Mark D. Mason, & Martin W. Healy for Massachusetts Bar Association.

Leslie Cooper & James D. Esseks, of New York, Jon W. Davidson & Shannon Minter, of California, Elliot M. Mincberg & Judith E. Schaeffer, of the District of Columbia, & John Reinstein, Sarah R. Wunsch, Paul Holtzman, & Hugh Dun Rappaport for Urban League of Eastern Massachusetts & others.

Paul Benjamin Linton, of Illinois, & Thomas M. Harvey for Robert J. Araujo & others.

Dwight G. Duncan for Massachusetts Family Institute, Inc., & others.

Glen Lavy, of Arizona, Stephen W. Reed, of California, & Bertin C. Emmons for National Association for Research and Therapy of Homosexuality, Inc., & others.

Robert W. Ash & Vincent P. McCarthy, of Connecticut, & Philip E. Cleary for The Common Good Foundation & others.

*311Don Stenberg, Attorney General of Nebraska, Mark L. Shurtleff, Attorney General of Utah, Brent A. Burnett, Assistant Attorney General of Utah, & Mark Barnett, Attorney General of South Dakota, for the State of Utah & others.

Chester Darling & Michael Williams for Massachusetts Citizens Alliance & another.

Daniel Avila for The Catholic Action League of Massachusetts.

Joshua K. Baker, of California, & Robert G. Caprera for José Martín de Agar & others.

Wendy J. Herdlein, of California, & James R. Knudsen for the Honorable Philip Travis & others.

Steven W. Fitschen, of Virginia, for The National Legal Foundation.

Jeffrey A. Shafer & David R. Langdon, of Ohio, William C. Duncan, of Utah, & Wendy J. Herdlein, of California, for Marriage Law Project.

Lisa Rae, Kenneth Elmore, Arthur Bemey, & Josephine Ross for The Religious Coalition for the Freedom to Marry & others.

Ann DiMaria for The Ethics & Religious Liberty Commission & others.

Anthony Mirenda, Vickie L. Henry, Lucy Fowler, John M. Granberry, Rachel N. Lessem, & Gabriel M. Helmer for Robert F. Williams & others.

Kenneth J. Parsigian for Peter W. Bardaglio & others.

David Cruz, of New York, John Taylor Williams, Carol V. Rose, Debra Squires-tee, Christopher Morrison, & Mami Gold-stein Caputo for William E. Adams & others.

Martin J. Newhouse & Katharine Bolland for Coalition gaie et lesbienne du Québec & others.

Joseph Ureneck, pro se.

Teresa S. Collett, of Texas, & Luke Stanton for Free Market Foundation.

Peter F. Zupcofska, L. Tracee Whitley, Heidi A. Nadel, & Corin R. Swift for Boston Bar Association & another.

Mary Jo Johnson, Jonathan A. Shapiro, & Amy L. Nash for The Massachusetts Psychiatric Society & others.

*312Tony R. Maida, Nina Joan Kimball, & Justine H. Brousseau for Libby Adler & others.

Daryl J. Lapp, Kevin D. Batt, & Katharine Silbaugh for Monroe Inker & another.

David Zwiebel, Mordechai Biser, & Nathan J. Diament, of New York, & Abba Cohen, of the District of Columbia, for Agudath Israel of America & others.

1

Julie Goodridge, David Wilson, Robert Compton, Michael Horgan, Edward Balmelli, Maureen Brodoff, Ellen Wade, Gary Chalmers, Richard Linnell, Heidi Norton, Gina Smith, Gloria Bailey, and Linda Davies.

2

Commissioner of Public Health.

Marshall, C.J.

Marriage is a vital social institution. The exclusive commitment of two individuals to each other nurtures love and mutual support; it brings stability to our society. For those who choose to marry, and for their children, marriage provides an abundance of legal, financial, and social benefits. In return it imposes weighty legal, financial, and social obligations. The question before us is whether, consistent with the Massachusetts Constitution, the Commonwealth may deny the protections, benefits, and obligations conferred by civil marriage to two individuals of the same sex who wish to marry. We conclude that it may not. The Massachusetts Constitution affirms the dignity and equality of all individuals. It forbids the creation of second-class citizens. In reaching our conclusion we have given full deference to the arguments made by the Commonwealth. But it has failed to identify any constitutionally adequate reason for denying civil marriage to same-sex couples.

We are mindful that our decision marks a change in the history of our marriage law. Many people hold deep-seated religious, moral, and ethical convictions that marriage should be limited to the union of one man and one woman, and that homosexual conduct is immoral. Many hold equally strong religious, moral, and ethical convictions that same-sex couples are entitled to be married, and that homosexual persons should be treated no differently than their heterosexual neighbors. Neither view answers the question before us. Our concern is with the Massachusetts Constitution as a charter of governance for every person properly within its reach. “Our obligation is to define the liberty of all, not to mandate our own moral code.” Lawrence v. Texas, 123 S. Ct. 2472, 2480 (2003) (Lawrence), quoting Planned Parenthood of Southeastern Pa. v. Casey, 505 U.S. 833, 850 (1992).

Whether the Commonwealth may use its formidable regula*313tory authority to bar same-sex couples from civil marriage is a question not previously addressed by a Massachusetts appellate court.3 It is a question the United States Supreme Court left open as a matter of Federal law in Lawrence, supra at 2484, where it was not an issue. There, the Court affirmed that the core concept of common human dignity protected by the Fourteenth Amendment to the United States Constitution precludes government intrusion into the deeply personal realms of consensual adult expressions of intimacy and one’s choice of an intimate partner. The Court also reaffirmed the central role that decisions whether to marry or have children bear in shaping one’s identity. Id. at 2481. The Massachusetts Constitution is, if anything, more protective of individual liberty and equality than the Federal Constitution; it may demand broader protection for fundamental rights; and it is less tolerant of government intrusion into the protected spheres of private Ufe.

Barred access to the protections, benefits, and obfigations of civil marriage, a person who enters into an intimate, exclusive union with another of the same sex is arbitrarily deprived of membership in one of our community’s most rewarding and cherished institutions. That exclusion is incompatible with the constitutional principles of respect for individual autonomy and equafity under law.

I

The plaintiffs are fourteen individuals from five Massachusetts counties. As of April 11, 2001, the date they filed their complaint, the plaintiffs Gloria Bailey, sixty years old, and Linda Davies, fifty-five years old, had been in a committed relationship for thirty years; the plaintiffs Maureen Brodoff, forty-nine years old, and Ellen Wade, fifty-two years old, had been in a committed *314relationship for twenty years and lived with their twelve year old daughter; the plaintiffs Hillary Goodridge, forty-four years old, and Julie Goodridge, forty-three years old, had been in a committed relationship for thirteen years and lived with their five year old daughter; the plaintiffs Gary Chalmers, thirty-five years old, and Richard Linnell, thirty-seven years old, had been in a committed relationship for thirteen years and lived with their eight year old daughter and Richard’s mother; the plaintiffs Heidi Norton, thirty-six years old, and Gina Smith, thirty-six years old, had been in a committed relationship for eleven years and lived with their two sons, ages five years and one year; the plaintiffs Michael Horgan, forty-one years old, and Edward Balmelli, forty-one years old, had been in a committed relationship for seven years; and the plaintiffs David Wilson, fifty-seven years old, and Robert Compton, fifty-one years old, had been in a committed relationship for four years and had cared for David’s mother in their home after a serious illness until she died.

The plaintiffs include business executives, lawyers, an investment banker, educators, therapists, and a computer engineer. Many are active in church, community, and school groups. They have employed such legal means as are available to them — for example, joint adoption, powers of attorney, and joint ownership of real property — to secure aspects of their relationships. Each plaintiff attests a desire to marry his or her partner in order to affirm publicly their commitment to each other and to secure the legal protections and benefits afforded to married couples and their children.

The Department of Public Health (department) is charged by statute with safeguarding public health. See G. L. c. 17. Among its responsibilities, the department oversees the registry of vital records and statistics (registry), which “enforce[s] all laws” relative to the issuance of marriage licenses and the keeping of marriage records, see G. L. c. 17, § 4, and which promulgates policies and procedures for the issuance of marriage licenses by city and town clerks and registers. See, e.g., G. L. c. 207, §§ 20, 28A, and 37. The registry is headed by a registrar of vital records and statistics (registrar), appointed by the Commissioner of Public Health (commissioner) with the approval of the public health council and supervised by the commissioner. See G. L. c. 17, § 4.

*315In March and April, 2001, each of the plaintiff couples attempted to obtain a marriage license from a city or town clerk’s office. As required under G. L. c. 207, they completed notices of intention to marry on forms provided by the registry, see G. L. c. 207, § 20, and presented these forms to a Massachusetts town or city clerk, together with the required health forms and marriage license fees. See G. L. c. 207, § 19. In each case, the clerk either refused to accept the notice of intention to marry or denied a marriage license to the couple on the ground that Massachusetts does not recognize same-sex marriage.4,5 Because obtaining a marriage license is a necessary prerequisite to civil marriage in Massachusetts, denying marriage licenses to the plaintiffs was tantamount to denying them access to civil marriage itself, with its appurtenant social and legal protections, benefits, and obligations.6

On April 11, 2001, the plaintiffs filed suit in the Superior Court against the department and the commissioner seeking a judgment that “the exclusion of the [pjlaintiff couples and other *316qualified same-sex couples from access to marriage licenses, and the legal and social status of civil marriage, as well as the protections, benefits and obligations of marriage, violates Massachusetts law.” See G. L. c. 231 A. The plaintiffs alleged violation of the laws of the Commonwealth, including but not limited to their rights under arts. 1, 6, 7, 10, 12, and 16, and Part II, c. 1, § 1, art. 4, of the Massachusetts Constitution.7,8 The department, represented by the Attorney General, admitted to a policy and practice of denying marriage licenses to same-sex *317couples. It denied that its actions violated any law or that the plaintiffs were entitled to relief. The parties filed cross motions for summary judgment.

A Superior Court judge ruled for the department. In a memorandum of decision and order dated May 7, 2002, he dismissed the plaintiffs’ claim that the marriage statutes should be construed to permit marriage between persons of the same sex, holding that the plain wording of G. L. c. 207, as well as the wording of other marriage statutes, precluded that interpretation. Turning to the constitutional claims, he held that the marriage exclusion does not offend the liberty, freedom, equality, or due process provisions of the Massachusetts Constitution, and that the Massachusetts Declaration of Rights does not guarantee “the fundamental right to marry a person of the same sex.” He concluded that prohibiting same-sex marriage rationally furthers the Legislature’s legitimate interest in safeguarding the “primary purpose” of marriage, “procreation.” The Legislature may rationally limit marriage to opposite-sex couples, he concluded, because those couples are “theoretically . . . capable of procreation,” they do not rely on “inherently more cumbersome” noncoital means of reproduction, and they are more likely than same-sex couples to have children, or more children.

After the complaint was dismissed and summary judgment entered for the defendants, the plaintiffs appealed. Both parties requested direct appellate review, which we granted.

n

Although the plaintiffs refer in passing to “the marriage statutes,” they focus, quite properly, on G. L. c. 207, the marriage licensing statute, which controls entry into civil marriage. As a preliminary matter, we summarize the provisions of that law.

General Laws c. 207 is both a gatekeeping and a public records statute. It sets minimum qualifications for obtaining a marriage license and directs city and town clerks, the registrar, and the department to keep and maintain certain “vital records” of civil marriages. The gatekeeping provisions of G. L. c. 207 are minimal. They forbid marriage of individuals within certain *318degrees of consanguinity, §§ 1 and 2, and polygamous marriages. See G. L. c. 207, § 4. See also G. L. c. 207, § 8 (marriages solemnized in violation of §§ 1, 2, and 4, are void ab initia). They prohibit marriage if one of the parties has communicable syphilis, see G. L. c. 207, § 28A, and restrict the circumstances in which a person under eighteen years of age may marry. See G. L. c. 207, §§ 7, 25, and 27. The statute requires that civil marriage be solemnized only by those so authorized. See G. L. c. 207, §§ 38-40.

The record-keeping provisions of G. L. c. 207 are more extensive. Marriage applicants file standard information forms and a medical certificate in any Massachusetts city or town clerk’s office and tender a filing fee. G. L. c. 207, §§ 19-20, 28A. The clerk issues the marriage license, and when the marriage is solemnized, the individual authorized to solemnize the marriage adds additional information to the form and returns it (or a copy) to the clerk’s office. G. L. c. 207, §§ 28, 30, 38-40 (this completed form is commonly known as the “marriage certificate”). The clerk sends a copy of the information to the registrar, and that information becomes a public record. See G. L. c. 17, § 4; G. L. c. 66, § 10.9’10

In short, for all the joy and solemnity that normally attend a marriage, G. L. c. 207, governing entrance to marriage, is a licensing law. The plaintiffs argue that because nothing in that licensing law specifically prohibits marriages between persons of the same sex, we may interpret the statute to permit “qualified same sex couples” to obtain marriage licenses, thereby avoiding the question whether the law is constitutional. See *319School Comm. of Greenfield v. Greenfield Educ. Ass’n, 385 Mass. 70, 79 (1982), and cases cited. This claim lacks merit.

We interpret statutes to carry out the Legislature’s intent, determined by the words of a statute interpreted according to “the ordinary and approved usage of the language.” Hanlon v. Rollins, 286 Mass. 444, 447 (1934). The everyday meaning of “marriage” is “[t]he legal union of a man and woman as husband and wife,” Black’s Law Dictionary 986 (7th ed. 1999), and the plaintiffs do not argue that the term “marriage” has ever had a different meaning under Massachusetts law. See, e.g., Milford v. Worcester, 7 Mass. 48, 52 (1810) (marriage “is an engagement, by which a single man and a single woman, of sufficient discretion, take each other for husband and wife”). This definition of marriage, as both the department and the Superior Court judge point out, derives from the common law. See Commonwealth v. Knowlton, 2 Mass. 530, 535 (1807) (Massachusetts common law derives from English common law except as otherwise altered by Massachusetts statutes and Constitution). See also Commonwealth v. Lane, 113 Mass. 458, 462-463 (1873) (“when the statutes are silent, questions of the validity of marriages are to be determined by the jus gentium, the common law of nations”); C.P. Kindregan, Jr., & M.L. Inker, Family Law and Practice § 1.2 (3d ed. 2002). Far from being ambiguous, the undefined word “marriage,” as used in G. L. c. 207, confirms the General Court’s intent to hew to the term’s common-law and quotidian meaning concerning the genders of the marriage partners.

The intended scope of G. L. c. 207 is also evident in its consanguinity provisions. See Chandler v. County Comm’rs of Nantucket County, 437 Mass. 430, 435 (2002) (statute’s various provisions may offer insight into legislative intent). Sections 1 and 2 of G. L. c. 207 prohibit marriages between a man and certain female relatives and a woman and certain male relatives, but are silent as to the consanguinity of male-male or female-female marriage applicants. See G. L. c. 207, §§ 1-2. The only reasonable explanation is that the Legislature did not intend that same-sex couples be licensed to marry. We conclude, as did the *320judge, that G. L. c. 207 may not be construed to permit same-sex couples to marry.11

in

A

The larger question is whether, as the department claims, government action that bars same-sex couples from civil marriage constitutes a legitimate exercise of the State’s authority to regulate conduct, or whether, as the plaintiffs claim, this categorical marriage exclusion violates the Massachusetts Constitution. We have recognized the long-standing statutory understanding, derived from the common law, that “marriage” means the lawful union of a woman and a man. But that history cannot and does not foreclose the constitutional question.

The plaintiffs’ claim that the marriage restriction violates the Massachusetts Constitution can be analyzed in two ways. Does it offend the Constitution’s guarantees of equality before the law? Or do the liberty and due process provisions of the Massachusetts Constitution secure the plaintiffs’ right to marry their chosen partner? In matters implicating marriage, family life, and the upbringing of children, the two constitutional concepts frequently overlap, as they do here. See, e.g., M.L.B. v. S.L.J., 519 U.S. 102, 120 (1996) (noting convergence of due process and equal protection principles in cases concerning parent-child relationships); Perez v. Sharp, 32 Cal. 2d 711, 728 (1948) (analyzing statutory ban on interracial marriage as equal protection violation concerning regulation of fundamental right). See also Lawrence, supra at 2482 (“Equality of treatment and the due process right to demand respect for conduct protected by the substantive guarantee of liberty are linked in important respects, and a decision on the latter point advances both interests”); Bolling v. Sharpe, 347 U.S. 497 (1954) (racial *321segregation in District of Columbia public schools violates the due process clause of Fifth Amendment to United States Constitution), decided the same day as Brown v. Board of Educ. of Topeka, 347 U.S. 483 (1954) (holding that segregation of public schools in States violates equal protection clause of Fourteenth Amendment). Much of what we say concerning one standard applies to the other.

We begin by considering the nature of civil marriage itself. Simply put, the government creates civil marriage. In Massachusetts, civil marriage is, and since pre-Colonial days has been, precisely what its name implies: a wholly secular institution. See Commonwealth v. Munson, 127 Mass. 459, 460-466 (1879) (noting that “[i]n Massachusetts, from very early times, the requisites of a valid marriage have been regulated by statutes of the Colony, Province, and Commonwealth,” and surveying marriage statutes from 1639 through 1834). No religious ceremony has ever been required to validate a Massachusetts marriage. Id.

In a real sense, there are three partners to every civil marriage: two willing spouses and an approving State. See DeMatteo v. DeMatteo, 436 Mass. 18, 31 (2002) (“Marriage is not a mere contract between two parties but a legal status from which certain rights and obligations arise”); Smith v. Smith, 171 Mass. 404, 409 (1898) (on marriage, the parties “assume[] new relations to each other and to the State”). See also French v. McAnarney, 290 Mass. 544, 546 (1935). While only the parties can mutually assent to marriage, the terms of the marriage — who may marry and what obligations, benefits, and liabilities attach to civil marriage — are set by the Commonwealth. Conversely, while only the parties can agree to end the marriage (absent the death of one of them or a marriage void ab initia), the Commonwealth defines the exit terms. See G. L. c. 208.

Civil marriage is created and regulated through exercise of the police power. See Commonwealth v. Stowell, 389 Mass. 171, 175 (1983) (regulation of marriage is properly within the scope of the police power). “Police power” (now more commonly termed the State’s regulatory authority) is an old-fashioned term for the Commonwealth’s lawmaking authority, as bounded by the liberty and equality guarantees of the Mas*322sachusetts Constitution and its express delegation of power from the people to their government. In broad terms, it is the Legislature’s power to enact rules to regulate conduct, to the extent that such laws are “necessary to secure the health, safety, good order, comfort, or general welfare of the community” (citations omitted). Opinion of the Justices, 341 Mass. 760, 785 (1960).12 See Commonwealth v. Alger, 7 Cush. 53, 85 (1851).

Without question, civil marriage enhances the “welfare of the community.” It is a “social institution of the highest importance.” French v. McAnarney, supra. Civil marriage anchors an ordered society by encouraging stable relationships over transient ones. It is central to the way the Commonwealth identifies individuals, provides for the orderly distribution of property, ensures that children and adults are cared for and supported whenever possible from private rather than public funds, and tracks important epidemiological and demographic data.

Marriage also bestows enormous private and social advantages on those who choose to marry. Civil marriage is at once a deeply personal commitment to another human being and a highly public celebration of the ideals of mutuality, companionship, intimacy, fidelity, and family. “It is an association that promotes a way of life, not causes; a harmony in living, not political faiths; a bilateral loyalty, not commercial or social projects.” Griswold v. Connecticut, 381 U.S. 479, 486 (1965). Because it fulfils yearnings for security, safe haven, and connection that express our common humanity, civil marriage is an esteemed institution, and the decision whether and whom to marry is among life’s momentous acts of self-definition.

Tangible as well as intangible benefits flow from marriage. The marriage license grants valuable property rights to those who meet the entry requirements, and who agree to what might otherwise be a burdensome degree of government regulation of their activities.13 See Leduc v. Commonwealth, 421 Mass. 433, *323435 (1995), cert, denied, 519 U.S. 827 (1996) (“The historical aim of licensure generally is preservation of public health, safety, and welfare by extending the public trust only to those with proven qualifications”). The Legislature has conferred on “each party [in a civil marriage] substantial rights concerning the assets of the other which unmarried cohabitants do not have.” Wilcox v. Trautz, 427 Mass. 326, 334 (1998). See Collins v. Guggenheim, 417 Mass. 615, 618 (1994) (rejecting claim for equitable distribution of property where plaintiff cohabited with but did not marry defendant); Feliciano v. Rosemar Silver Co., 401 Mass. 141, 142 (1987) (government interest in promoting marriage would be “subverted” by recognition of “a right to recover for loss of consortium by a person who has not accepted the correlative responsibilities of marriage”); Davis v. Misiono, 373 Mass. 261, 263 (1977) (unmarried partners not entitled to rights of separate support or alimony). See generally Attorney Gen. v. Desilets, 418 Mass. 316, 327-328 & nn.10, 11 (1994).

The benefits accessible only by way of a marriage license are enormous, touching nearly every aspect of life and death. The department states that “hundreds of statutes” are related to marriage and to marital benefits. With no attempt to be comprehensive, we note that some of the statutory benefits conferred by the Legislature on those who enter into civil marriage include, as to property: joint Massachusetts income tax filing (G. L. c. 62C, § 6); tenancy by the entirety (a form of ownership that provides certain protections against creditors and allows for the automatic descent of property to the surviving spouse without probate) (G. L. c. 184, § 7); extension of the benefit of the homestead protection (securing up to $300,000 in equity from creditors) to one’s spouse and children (G. L. c. 188, § 1); automatic rights to inherit the property of a deceased spouse who does not leave a will (G. L. c. 190, § 1); the rights of elective share and of dower (which allow surviving spouses certain property rights where the decedent spouse has not made adequate provision for the survivor in a will) (G. L. c. 191, *324§ 15, and G. L. c. 189); entitlement to wages owed to a deceased employee (G. L. c. 149, § 178A [general] and G. L. c. 149, § 178C [public employees]); eligibility to continue certain businesses of a deceased spouse (e.g., G. L. c. 112, § 53 [dentist]); the right to share the medical policy of one’s spouse (e.g., G. L. c. 175, § 108, Second [a] [3] [defining insured’s “dependent” to include one’s spouse), see Connors v. Boston, 430 Mass. 31, 43 (1999) [domestic partners of city employees not included within term “dependent” as used in G. L. c. 32B, § 2]); thirty-nine week continuation of health coverage for the spouse of a person who is laid off or dies (e.g., G. L. c. 175, § 110G); preferential options under the Commonwealth’s pension system (see G. L. c. 32, § 12 [2] [“Joint and Last Survivor Allowance”]); preferential benefits in the Commonwealth’s medical program, MassHealth (e.g., 130 Code Mass. Regs. § 515.012 [A], prohibiting placing lien on long-term care patient’s former home if spouse still lives there); access to veterans’ spousal benefits and preferences (e.g., G. L. c. 115, § 1 [defining “dependents”] and G. L. c. 31, § 26 [State employment] and § 28 [municipal employees]); financial protections for spouses of certain Commonwealth employees (fire fighters, police officers, and prosecutors, among others) killed in the performance of duty (e.g., G. L. c. 32, §§ 100-103); the equitable division of marital property on divorce (G. L. c. 208, § 34); temporary and permanent alimony rights (G. L. c. 208, §§ 17 and 34); the right to separate support on separation of the parties that does not result in divorce (G. L. c. 209, § 32); and the right to bring claims for wrongful death and loss of consortium, and for funeral and burial expenses and punitive damages resulting from tort actions (G. L. c. 229, §§ 1 and 2; G. L. c. 228, § 1. See Feliciano v. Rosemar Silver Co., supra).

Exclusive marital benefits that are not directly tied to property rights include the presumptions of legitimacy and parentage of children bom to a married couple (G. L. c. 209C, § 6, and G. L. c. 46, § 4B); and evidentiary rights, such as the prohibition against spouses testifying against one another about their private conversations, applicable in both civil and criminal cases (G. L. c. 233, § 20). Other statutory benefits of a personal nature available only to married individuals include qualification for *325bereavement or medical leave to care for individuals related by blood or marriage (G. L. c. 149, § 52D); an automatic “family member” preference to make medical decisions for an incompetent or disabled spouse who does not have a contrary health care proxy, see Shine v. Vega, 429 Mass. 456, 466 (1999); the application of predictable rules of child custody, visitation, support, and removal out-of-State when married parents divorce (e.g., G. L. c. 208, § 19 [temporary custody], § 20 [temporary support], § 28 [custody and support on judgment of divorce], § 30 [removal from Commonwealth], and § 31 [shared custody plan]; priority rights to administer the estate of a deceased spouse who dies without a will, and the requirement that a surviving spouse must consent to the appointment of any other person as administrator (G. L. c. 38, § 13 [disposition of body], and G. L. c. 113, § 8 [anatomical gifts]); and the right to interment in the lot or tomb owned by one’s deceased spouse (G. L. c. 114, §§ 29-33).

Where a married couple has children, their children are also directly or indirectly, but no less auspiciously, the recipients of the special legal and economic protections obtained by civil marriage. Notwithstanding the Commonwealth’s strong public policy to abolish legal distinctions between marital and non-marital children in providing for the support and care of minors, see Department of Revenue v. Mason M., 439 Mass. 665 (2003); Woodward v. Commissioner of Social Sec., 435 Mass. 536, 546 (2002), the fact remains that marital children reap a measure of family stability and economic security based on their parents’ legally privileged status that is largely inaccessible, or not as readily accessible, to nonmarital children. Some of these benefits are social, such as the enhanced approval that still attends the status of being a marital child. Others are material, such as the greater ease of access to family-based State and Federal benefits that attend the presumptions of one’s parentage.

It is undoubtedly for these concrete reasons, as well as for its intimately personal significance, that civil marriage has long been termed a “civil right.” See, e.g., Loving v. Virginia, 388 U.S. 1, 12 (1967) (“Marriage is one of the ‘basic civil rights of man,’ fundamental to our very existence and survival”), quoting Skinner v. Oklahoma, 316 U.S. 535, 541 (1942); Milford v. *326Worcester, 7 Mass. 48, 56 (1810) (referring to “civil rights incident to marriages”). See also Baehr v. Lewin, 74 Haw. 530, 561 (1993) (identifying marriage as “civil right[ ]”); Baker v. State, 170 Vt. 194, 242 (1999) (Johnson, J., concurring in part and dissenting in part) (same). The United States Supreme Court has described the right to marry as “of fundamental importance for all individuals” and as “part of the fundamental ‘right of privacy’ implicit in the Fourteenth Amendment’s Due Process Clause.” Zablocki v. Redhail, 434 U.S. 374, 384 (1978). See Loving v. Virginia, supra (“The freedom to marry has long been recognized as one of the vital personal rights essential to the orderly pursuit of happiness by free men”).14

Without the right to marry — or more properly, the right to choose to marry — one is excluded from the full range of human experience and denied full protection of the laws for one’s “avowed commitment to an intimate and lasting human relationship.” Baker v. State, supra at 229. Because civil marriage is central to the lives of individuals and the welfare of the community, our laws assiduously protect the individual’s right to marry against undue government incursion. Laws may not “interfere directly and substantially with the right to marry.” Zablocki v. Redhail, supra at 387. See Perez v. Sharp, 32 Cal. 2d 711, 714 (1948) (“There can be no prohibition of marriage except for an important social objective and reasonable means”).15

Unquestionably, the regulatory power of the Commonwealth *327over civil marriage is broad, as is the Commonwealth’s discretion to award public benefits. See Commonwealth v. Stowell, 389 Mass. 171, 175 (1983) (marriage); Moe v. Secretary of Admin. & Fin., 382 Mass. 629, 652 (1981) (Medicaid benefits). Individuals who have the choice to marry each other and nevertheless choose not to may properly be denied the legal benefits of marriage. See Wilcox v. Trautz, 427 Mass. 326, 334 (1998); Collins v. Guggenheim, 417 Mass. 615, 618 (1994); Feliciano v. Rosemar Silver Co., 401 Mass. 141, 142 (1987). But that same logic cannot hold for a qualified individual who would marry if she or he only could.

B

For decades, indeed centuries, in much of this country (including Massachusetts) no lawful marriage was possible between white and black Americans. That long history availed not when the Supreme Court of California held in 1948 that a legislative prohibition against interracial marriage violated the due process and equality guarantees of the Fourteenth Amendment, Perez v. Sharp, 32 Cal. 2d 711, 728 (1948), or when, nineteen years later, the United States Supreme Court also held that a statutory bar to interracial marriage violated the Fourteenth Amendment, Loving v. Virginia, 388 U.S. 1 (1967).16 As both Perez and Loving make clear, the right to marry means *328little if it does not include the right to marry the person of one’s choice, subject to appropriate government restrictions in the interests of public health, safety, and welfare. See Perez v. Sharp, supra at 717 (“the essence of the right to marry is freedom to join in marriage with the person of one’s choice”). See also Loving v. Virginia, supra at 12. In this case, as in Perez and Loving, a statute deprives individuals of access to an institution of fundamental legal, personal, and social significance — the institution of marriage — because of a single trait: skin color in Perez and Loving, sexual orientation here. As it did in Perez and Loving, history must yield to a more fully developed understanding of the invidious quality of the discrimination.17

The Massachusetts Constitution protects matters of personal liberty against government incursion as zealously, and often more so, than does the Federal Constitution, even where both Constitutions employ essentially the same language. See Planned Parenthood League of Mass., Inc. v. Attorney Gen., 424 Mass. 586, 590 (1997); Corning Glass Works v. Ann & Hope, Inc. of Danvers, 363 Mass. 409, 416 (1973). That the Massachusetts Constitution is in some instances more protective of individual liberty interests than is the Federal Constitution is not surprising. Fundamental to the vigor of our Federal system of government is that “state courts are absolutely free to interpret state constitutional provisions to accord greater protection to individual rights than do similar provisions of the United States Constitution.” Arizona v. Evans, 514 U.S. 1, 8 (1995).18

The individual liberty and equality safeguards of the Mas*329sachusetts Constitution protect both “freedom from” unwarranted government intrusion into protected spheres of life and “freedom to” partake in benefits created by the State for the common good. See Bachrach v. Secretary of the Commonwealth, 382 Mass. 268, 273 (1981); Dalli v. Board of Educ., 358 Mass. 753, 759 (1971). Both freedoms are involved here. Whether and whom to marry, how to express sexual intimacy, and whether and how to establish a family — these are among the most basic of every individual’s liberty and due process rights. See, e.g., Lawrence, supra at 2481; Planned Parenthood of Southeastern Pa. v. Casey, 505 U.S. 833, 851 (1992); Zablocki v. Redhail, 434 U.S. 374, 384 (1978); Roe v. Wade, 410 U.S. 113, 152-153 (1973); Eisenstadt v. Baird, 405 U.S. 438, 453 (1972); Loving v. Virginia, supra. And central to personal freedom and security is the assurance that the laws will apply equally to persons in similar situations. “Absolute equality before the law is a fundamental principle of our own Constitution.” Opinion of the Justices, 211 Mass. 618, 619 (1912). The liberty interest in choosing whether and whom to marry would be hollow if the Commonwealth could, without sufficient justification, foreclose an individual from freely choosing the person with whom to share an exclusive commitment in the unique institution of civil marriage.

The Massachusetts Constitution requires, at a minimum, that the exercise of the State’s regulatory authority not be “arbitrary or capricious.” Commonwealth v. Henry’s Drywall Co., 366 Mass. 539, 542 (1974).19 Under both the equality and liberty guarantees, regulatory authority must, at very least, serve “a *330legitimate purpose in a rational way”; a statute must “bear a reasonable relation to a permissible legislative objective.” Rushworth v. Registrar of Motor Vehicles, 413 Mass. 265, 270 (1992). See, e.g., Massachusetts Fed’n of Teachers v. Board of Educ., 436 Mass. 763, 778 (2002) (equal protection); Coffee-Rich, Inc. v. Commissioner of Pub. Health, 348 Mass. 414, 422 (1965) (due process). Any law failing to satisfy the basic standards of rationality is void.

The plaintiffs challenge the marriage statute on both equal protection and due process grounds. With respect to each such claim, we must first determine the appropriate standard of review. Where a statute implicates a fundamental right or uses a suspect classification, we employ “strict judicial scrutiny.” Lowell v. Kowalski, 380 Mass. 663, 666 (1980). For all other statutes, we employ the “ ‘rational basis’ test.” English v. New England Med. Ctr., 405 Mass. 423, 428 (1989). For due process claims, rational basis analysis requires that statutes “bear[] a real and substantial relation to the public health, safety, morals, or some other phase of the general welfare.” Coffee-Rich, Inc. v. Commissioner of Pub. Health, supra, quoting Sperry & Hutchinson Co. v. Director of the Div. on the Necessaries of Life, 307 Mass. 408, 418 (1940). For equal protection challenges, the rational basis test requires that “an impartial lawmaker could logically believe that the classification would serve a legitimate public purpose that transcends the harm to the members of the disadvantaged class.” English v. New England Med. Ctr., supra at 429, quoting Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 452 (1985) (Stevens, J., concurring).20

The department argues that no fundamental right or “suspect” *331class is at issue here,21 and rational basis is the appropriate standard of review. For the reasons we explain below, we conclude that the marriage ban does not meet the rational basis test for either due process or equal protection. Because the statute does not survive rational basis review, we do not consider the plaintiffs’ arguments that this case merits strict judicial scrutiny.

The department posits three legislative rationales for prohibiting same-sex couples from marrying: (1) providing a “favorable setting for procreation”; (2) ensuring the optimal setting for child rearing, which the department defines as “a two-parent family with one parent of each sex”; and (3) preserving scarce State and private financial resources. We consider each in turn.

The judge in the Superior Court endorsed the first rationale, holding that “the state’s interest in regulating marriage is based on the traditional concept that marriage’s primary purpose is procreation.” This is incorrect. Our laws of civil marriage do not privilege procreative heterosexual intercourse between married people above every other form of adult intimacy and every other means of creating a family. General Laws c. 207 contains no requirement that the applicants for a marriage license attest to their ability or intention to conceive children by coitus. Fertility is not a condition of marriage, nor is it grounds for divorce. People who have never consummated their marriage, and never plan to, may be and stay married. See Franklin v. Franklin, 154 Mass. 515, 516 (1891) (“The consummation of a marriage by *332coition is not necessary to its validity”).22 People who cannot stir from their deathbed may marry. See G. L. c. 207, § 28A. While it is certainly true that many, perhaps most, married couples have children together (assisted or unassisted), it is the exclusive and permanent commitment of the marriage partners to one another, not the begetting of children, that is the sine qua non of civil marriage.23

Moreover, the Commonwealth affirmatively facilitates bringing children into a family regardless of whether the intended parent is married or unmarried, whether the child is adopted or bom into a family, whether assistive technology was used to conceive the child, and whether the parent or her partner is *333heterosexual, homosexual, or bisexual.24 If procreation were a necessary component of civil marriage, our statutes would draw a tighter circle around the permissible bounds of nonmarital child bearing and the creation of families by noncoital means. The attempt to isolate procreation as “the source of a fundamental right to marry,” post at 370 (Cordy, J., dissenting), overlooks the integrated way in which courts have examined the complex and overlapping realms of personal autonomy, marriage, family life, and child rearing. Our jurisprudence recognizes that, in these nuanced and fundamentally private areas of life, such a narrow focus is inappropriate.

The “marriage is procreation” argument singles out the one unbridgeable difference between same-sex and opposite-sex couples, and transforms that difference into the essence of legal marriage. Like “Amendment 2” to the Constitution of Colorado, which effectively denied homosexual persons equality under the law and full access to the political process, the marriage restriction impermissibly “identifies persons by a single trait and then denies them protection across the board.” Romer v. Evans, 517 U.S. 620, 633 (1996). In so doing, the State’s action confers an official stamp of approval on the destructive stereotype that same-sex relationships are inherently unstable and inferior to opposite-sex relationships and are not worthy of respect.25

The department’s first stated rationale, equating marriage with unassisted heterosexual procreation, shades imperceptibly into its second: that confining marriage to opposite-sex couples ensures that children are raised in the “optimal” setting. Protect*334ing the welfare of children is a paramount State policy. Restricting marriage to opposite-sex couples, however, cannot plausibly further this policy. “The demographic changes of the past century make it difficult to speak of an average American family. The composition of families varies greatly from household to household.” Troxel v. Granville, 530 U.S. 57, 63 (2000). Massachusetts has responded supportively to “the changing realities of the American family,” id. at 64, and has moved vigorously to strengthen the modem family in its many variations. See, e.g., G. L. c. 209C (paternity statute); G. L. c. 119, § 39D (grandparent visitation statute); Blixt v. Blixt, 437 Mass. 649 (2002), cert, denied, 537 U.S. 1189 (2003) (same); E.N.O. v. L.M.M., 429 Mass. 824, cert, denied, 528 U.S. 1005 (1999) (de facto parent); Youmans v. Ramos, 429 Mass. 774, 782 (1999) (same); and Adoption of Tammy, 416 Mass. 205 (1993) (coparent adoption). Moreover, we have repudiated the common-law power of the State to provide varying levels of protection to children based on the circumstances of birth. See G. L. c. 209C (paternity statute); Powers v. Wilkinson, 399 Mass. 650, 661 (1987) (“Ours is an era in which logic and compassion have impelled the law toward unburdening children from the stigma and the disadvantages heretofore attendant upon the status of illegitimacy”). The “best interests of the child” standard does not turn on a parent’s sexual orientation or marital status. See e.g., Doe v. Doe, 16 Mass. App. Ct. 499, 503 (1983) (parent’s sexual orientation insufficient ground to deny custody of child in divorce action). See also E.N.O. v. L.M.M., supra at 829-830 (best interests of child determined by considering child’s relationship with biological and de facto same-sex parents); Silvia v. Silvia, 9 Mass. App. Ct. 339, 341 & n.3 (1980) (collecting support and custody statutes containing no gender distinction).

The department has offered no evidence that forbidding marriage to people of the same sex will increase the number of couples choosing to enter into opposite-sex marriages in order to have and raise children. There is thus no rational relationship between the marriage statute and the Commonwealth’s proffered goal of protecting the “optimal” child rearing unit. Moreover, the department readily concedes that people in same-sex couples may be “excellent” parents. These couples (includ*335ing four of the plaintiff couples) have children for the reasons others do — to love them, to care for them, to nurture them. But the task of child rearing for same-sex couples is made infinitely harder by their status as outliers to the marriage laws. While establishing the parentage of children as soon as possible is crucial to the safety and welfare of children, see Culliton v. Beth Israel Deaconness Med. Ctr., 435 Mass. 285, 292 (2001), same-sex couples must undergo the sometimes lengthy and intrusive process of second-parent adoption to establish their joint parentage. While the enhanced income provided by marital benefits is an important source of security and stability for married couples and their children, those benefits are denied to families headed by same-sex couples. See, e.g., note 6, supra. While the laws of divorce provide clear and reasonably predictable guidelines for child support, child custody, and property division on dissolution of a marriage, same-sex couples who dissolve their relationships find themselves and their children in the highly unpredictable terrain of equity jurisdiction. See E.N.O. v. L.M.M., supra. Given the wide range of public benefits reserved only for married couples, we do not credit the department’s contention that the absence of access to civil marriage amounts to little more than an inconvenience to same-sex couples and their children. Excluding same-sex couples from civil marriage will not make children of opposite-sex marriages more secure, but it does prevent children of same-sex couples from enjoying the immeasurable advantages that flow from the assurance of “a stable family structure in which children will be reared, educated, and socialized.” Post at 381 (Cordy, J., dissenting).26

No one disputes that the plaintiff couples are families, that many are parents, and that the children they are raising, like all children, need and should have the fullest opportunity to grow up in a secure, protected family unit. Similarly, no one disputes that, under the rubric of marriage, the State provides a *336cornucopia of substantial benefits to married parents and their children. The preferential treatment of civil marriage reflects the Legislature’s conclusion that marriage “is the foremost setting for the education and socialization of children” precisely because it “encourages parents to remain committed to each other and to their children as they grow.” Post at 383 (Cordy, J., dissenting).

In this case, we are confronted with an entire, sizeable class of parents raising children who have absolutely no access to civil marriage and its protections because they are forbidden from procuring a marriage license. It cannot be rational under our laws, and indeed it is not permitted, to penalize children by depriving them of State benefits because the State disapproves of their parents’ sexual orientation.

The third rationale advanced by the department is that limiting marriage to opposite-sex couples furthers the Legislature’s interest in conserving scarce State and private financial resources. The marriage restriction is rational, it argues, because the General Court logically could assume that same-sex couples are more financially independent than married couples and thus less needy of public marital benefits, such as tax advantages, or private marital benefits, such as employer-financed health plans that include spouses in their coverage.

An absolute statutory ban on same-sex marriage bears no rational relationship to the goal of economy. First, the department’s conclusory generalization — that same-sex couples are less financially dependent on each other than opposite-sex couples — ignores that many same-sex couples, such as many of the plaintiffs in this case, have children and other dependents (here, aged parents) in their care.27 The department does not contend, nor could it, that these dependents are less needy or deserving than the dependents of married couples. Second, Massachusetts marriage laws do not condition receipt of public and private financial benefits to married individuals on a demonstration of financial dependence on each other; the benefits are available to married couples regardless of whether *337they mingle their finances or actually depend on each other for support.

The department suggests additional rationales for prohibiting same-sex couples from marrying, which are developed by some amici. It argues that broadening civil marriage to include same-sex couples will trivialize or destroy the institution of marriage as it has historically been fashioned. Certainly our decision today marks a significant change in the definition of marriage as it has been inherited from the common law, and understood by many societies for centuries. But it does not disturb the fundamental value of marriage in our society.

Here, the plaintiffs seek only to be married, not to undermine the institution of civil marriage. They do not want marriage abolished. They do not attack the binary nature of marriage, the consanguinity provisions, or any of the other gate-keeping provisions of the marriage licensing law. Recognizing the right of an individual to marry a person of the same sex will not diminish the validity or dignity of opposite-sex marriage, any more than recognizing the right of an individual to marry a person of a different race devalues the marriage of a person who marries someone of her own race.28 If anything, extending civil marriage to same-sex couples reinforces the importance of marriage to individuals and communities. That same-sex couples are willing to embrace marriage’s solemn obligations of exclusivity, mutual support, and commitment to one another is a testament to the enduring place of marriage in our laws and in the human spirit.29

It has been argued that, due to the State’s strong interest in *338the institution of marriage as a stabilizing social structure, only the Legislature can control and define its boundaries. Accordingly, our elected representatives legitimately may choose to exclude same-sex couples from civil marriage in order to assure all citizens of the Commonwealth that (1) the benefits of our marriage laws are available explicitly to create and support a family setting that is, in the Legislature’s view, optimal for child rearing, and (2) the State does not endorse gay and lesbian parenthood as the equivalent of being raised by one’s married biological parents.30 These arguments miss the point. The Massachusetts Constitution requires that legislation meet certain criteria and not extend beyond certain limits. It is the function of courts to determine whether these criteria are met and whether these limits are exceeded. In most instances, these limits are defined by whether a rational basis exists to conclude that legislation will bring about a rational result. The Legislature in the first instance, and the courts in the last instance, must ascertain whether such a rational basis exists. To label the court’s role as usurping that of the Legislature, see, e.g., post at 394-395 (Cordy, J., dissenting), is to misunderstand the nature *339and purpose of judicial review. We owe great deference to the Legislature to decide social and policy issues, but it is the traditional and settled role of courts to decide constitutional issues.31

The history of constitutional law “is the story of the extension of constitutional rights and protections to people once ignored or excluded.” United States v. Virginia, 518 U.S. 515, 557 (1996) (construing equal protection clause of Fourteenth Amendment to prohibit categorical exclusion of women from public military institute). This statement is as true in the area of civil marriage as in any other area of civil rights. See, e.g., Turner v. Safley, 482 U.S. 78 (1987); Loving v. Virginia, 388 U.S. 1 (1967); Perez v. Sharp, 32 Cal. 2d 711 (1948). As a public institution and a right of fundamental importance, civil marriage is an evolving paradigm. The common law was exceptionally harsh toward women who became wives: a woman’s legal identity all but evaporated into that of her husband. See generally C.P. Kindregan, Jr., & M.L. Inker, Family Law and Practice §§ 1.9 and 1.10 (3d ed. 2002). Thus, one *340early Nineteenth Century jurist could observe matter of factly that, prior to the abolition of slavery in Massachusetts, “the condition of a slave resembled the connection of a wife with her husband, and of infant children with their father. He is obliged to maintain them, and they cannot be separated from him.” Winchendon v. Hatfield, 4 Mass. 123, 129 (1808). But since at least the middle of the Nineteenth Century, both the courts and the Legislature have acted to ameliorate the harshness of the common-law regime. In Bradford v. Worcester, 184 Mass. 557, 562 (1904), we refused to apply the common-law rule that the wife’s legal residence was that of her husband to defeat her claim to a municipal “settlement of paupers.” In Lewis v. Lewis, 370 Mass. 619, 629 (1976), we abrogated the common-law doctrine immunizing a husband against certain suits because the common-law rule was predicated on “antediluvian assumptions concerning the role and status of women in marriage and in society.” Id. at 621. Alarms about the imminent erosion of the “natural” order of marriage were sounded over the demise of antimiscegenation laws, the expansion of the rights of married women, and the introduction of “no-fault” divorce.32 Marriage has survived all of these transformations, and we have no doubt that marriage will continue to be a vibrant and revered institution.

We also reject the argument suggested by the department, and elaborated by some amici, that expanding the institution of civil marriage in Massachusetts to include same-sex couples will lead to interstate conflict. We would not presume to dictate how another State should respond to today’s decision. But neither should considerations of comity prevent us from according Massachusetts residents the full measure of protection available under the Massachusetts Constitution. The genius of our Federal system is that each State’s Constitution has vitality specific to *341its own traditions, and that, subject to the minimum requirements of the Fourteenth Amendment, each State is free to address difficult issues of individual liberty in the manner its own Constitution demands.

Several amici suggest that prohibiting marriage by same-sex couples reflects community consensus that homosexual conduct is immoral. Yet Massachusetts has a strong affirmative policy of preventing discrimination on the basis of sexual orientation. See G. L. c. 15IB (employment, housing, credit, services); G. L. c. 265, § 39 (hate crimes); G. L. c. 272, § 98 (public accommodation); G. L. c. 76, § 5 (public education). See also, e.g., Commonwealth v. Balthazar, 366 Mass. 298 (1974) (decriminalization of private consensual adult conduct); Doe v. Doe, 16 Mass. App. Ct. 499, 503 (1983) (custody to homosexual parent not per se prohibited).

The department has had more than ample opportunity to articulate a constitutionally adequate justification for limiting civil marriage to opposite-sex unions. It has failed to do so. The department has offered purported justifications for the civil marriage restriction that are starkly at odds with the comprehensive network of vigorous, gender-neutral laws promoting stable families and the best interests of children. It has failed to identify any relevant characteristic that would justify shutting the door to civil marriage to a person who wishes to marry someone of the same sex.

The marriage ban works a deep and scarring hardship on a very real segment of the community for no rational reason. The absence of any reasonable relationship between, on the one hand, an absolute disqualification of same-sex couples who wish to enter into civil marriage and, on the other, protection of public health, safety, or general welfare, suggests that the marriage restriction is rooted in persistent prejudices against persons who are (or who are believed to be) homosexual.33 “The Constitution cannot control such prejudices but neither can it *342tolerate them. Private biases may be outside the reach of the law, but the law cannot, directly or indirectly, give them effect.” Palmore v. Sidoti, 466 U.S. 429, 433 (1984) (construing Fourteenth Amendment). Limiting the protections, benefits, and obligations of civil marriage to opposite-sex couples violates the basic premises of individual liberty and equality under law protected by the Massachusetts Constitution.

IV

We consider next the plaintiffs’ request for relief. We preserve as much of the statute as may be preserved in the face of the successful constitutional challenge. See Mayor of Boston v. Treasurer & Receiver Gen., 384 Mass. 718, 725 (1981); Dalli v. Board of Educ., 358 Mass. 753, 759 (1971). See also G. L. c. 4, § 6, Eleventh.

Here, no one argues that striking down the marriage laws is an appropriate form of relief. Eliminating civil marriage would be wholly inconsistent with the Legislature’s deep commitment to fostering stable families and would dismantle a vital organiz*343ing principle of our society.34 We face a problem similar to one that recently confronted the Court of Appeal for Ontario, the highest court of that Canadian province, when it considered the constitutionality of the same-sex marriage ban under the Canadian Charter of Rights and Freedoms (Charter), part of Canada’s Federal Constitution. See Halpern v. Toronto (City), 172 O.A.C. 276 (2003). Canada, like the United States, adopted the common law of England that civil marriage is “the voluntary union for life of one man and one woman, to the exclusion of all others.” Id. at par. (36), quoting Hyde v. Hyde, [1861-1873] All E.R. 175 (1866). In holding that the limitation of civil marriage to opposite-sex couples violated the Charter, the Court of Appeal refined the common-law meaning of marriage. We concur with this remedy, which is entirely consonant with established principles of jurisprudence empowering a court to refine a common-law principle in light of evolving constitutional standards. See Powers v. Wilkinson, 399 Mass. 650, 661-662 (1987) (reforming common-law rule of construction of “issue”); Lewis v. Lewis, 370 Mass. 619, 629 (1976) (abolishing common-law rule of certain interspousal immunity).

We construe civil marriage to mean the voluntary union of two persons as spouses, to the exclusion of all others. This reformulation redresses the plaintiffs’ constitutional injury and furthers the aim of marriage to promote stable, exclusive relationships. It advances the two legitimate State interests the department has identified: providing a stable setting for child rearing and conserving State resources. It leaves intact the *344Legislature’s broad discretion to regulate marriage. See Commonwealth v. Stowell, 389 Mass. 171, 175 (1983).

In their complaint the plaintiffs request only a declaration that their exclusion and the exclusion of other qualified same-sex couples from access to civil marriage violates Massachusetts law. We declare that barring an individual from the protections, benefits, and obligations of civil marriage solely because that person would marry a person of the same sex violates the Massachusetts Constitution. We vacate the summary judgment for the department. We remand this case to the Superior Court for entry of judgment consistent with this opinion. Entry of judgment shall be stayed for 180 days to permit the Legislature to take such action as it may deem appropriate in light of this opinion. See, e.g., Michaud v. Sheriff of Essex County, 390 Mass. 523, 535-536 (1983).

So ordered.

3

For American appellate courts that have recently addressed this issue, see Standhardt v. Superior Court, 77 P.3d 451 (Ariz. Ct. App. 2003); Dean v. District of Columbia, 653 A.2d 307 (D.C. 1995); Baehr v. Lewin, 74 Haw. 530 (1993); Baker v. State, 170 Vt. 194, 242 (1999). Earlier cases include Adams v. Howerton, 486 F. Supp. 1119 (C.D. Cal. 1980), aff'd, 673 F.2d 1036 (9th Cir.), cert, denied, 458 U.S. 1111 (1982); Jones v. Hallahan, 501 S.W.2d 588 (Ky. Ct. App. 1973); Baker v. Nelson, 291 Minn. 310 (1971), appeal dismissed, 409 U.S. 810 (1972); Singer v. Hara, 11 Wash. App. 247 (1974). See also Halpern v. Toronto (City), 172 O.A.C. 276 (2003); Egale Canada, Inc. v. Canada (Attorney Gen.), 13 B.C.L.R. (4th) 1 (2003).

4

General Laws c. 207, § 37, provides: “The commissioner of public health shall furnish to the clerk or registrar of every town a printed list of all legal impediments to marriage, and the clerk or registrar shall forthwith post and thereafter maintain it in a conspicuous place in his office.” The record does not reveal whether any of the clerks’ offices that considered the plaintiffs’ applications for a marriage license had posted such a list of impediments, or whether such list included as an impediment that the applicants are of the same sex.

5

The plaintiffs alleged that they met all of the facial qualifications to obtain marriage licenses pursuant to G. L. c. 207, and the department does not contest this assertion.

6

The complaint alleged various circumstances in which the absence of the full legal protections of civil marriage has harmed them and their children. For example, Hillary and Julie Goodridge alleged that, when Julie gave birth to their daughter (whom Hillary subsequently coadopted) during a delivery that required the infant’s transfer to neonatal intensive care, Hillary “had difficulty gaining access to Julie and their newborn daughter at the hospital”; Gary Chalmers and Richard Linnell alleged that “Gary pays for a family health insurance policy at work which covers only him and their daughter because Massachusetts law does not consider Rich to be a ‘dependent.’ This means that their household must purchase a separate individual policy of health insurance for Rich at considerable expense. . . . Gary has a pension plan at work, but under state law, because he is a municipal employee, that plan does not allow him the same range of options in providing for his beneficiary that a married spouse has and thus he cannot provide the same security to his family that a married person could if he should predecease Rich.”

7

Article 1, as amended by art. 106 of the Amendments to the Massachusetts Constitution, provides: “All people are bom free and equal and have certain natural, essential and unalienable rights; among which may be reckoned the right of enjoying and defending their lives and liberties; that of acquiring, possessing and protecting property; in fine, that of seeking and obtaining their safety and happiness. Equality under the law shall not be denied or abridged because of sex, race, color, creed or national origin.”

Article 6 provides: “No man, nor corporation, or association of men, have any other title to obtain advantages, or particular and exclusive privileges, distinct from those of the community, than what arises from the consideration of services rendered to the public . . . .”

Article 7 provides: “Government is instituted for the common good; for the protection, safety, prosperity, and happiness of the people; and not for the profit, honor, or private interest of any one man, family or class of men: Therefore the people alone have an incontestable, unalienable, and indefeasible right to institute government; and to reform, alter, or totally change the same, when their protection, safety, prosperity and happiness require it.”

Article 10 provides, in relevant part: “Each individual of the society has a right to be protected by it in the enjoyment of his life, liberty and property, according to standing laws . . . .”

Article 12 provides, in relevant part: “[N]o subject shall be . . . deprived of his property, immunities, or privileges, put out of the protection of the law ... or deprived of his life, liberty, or estate, but by the judgment of his peers, or the law of the land.”

Article 16, as amended by art. 77 of the Amendments, provides, in relevant part: “The right of free speech shall not be abridged.” Part II, c. 1, § 1, art. 4, as amended by art. 112, provides, in pertinent part, that “full power and authority are hereby given and granted to the said general court, from time to time, to make, ordain, and establish all manner of wholesome and reasonable orders, laws, statutes, and ordinances, directions and instructions, either with penalties or without; so as the same be not repugnant or contrary to this constitution, as they shall judge to be for the good and welfare of this Commonwealth. ’ ’

8

The department claims that the plaintiffs have waived their art. 12 and art. 16 claims on appeal. Because our holding today does not turn on art. 12 or art. 16, we do not consider the department’s waiver argument.

9

The marital forms forwarded by the clerk or register must contain the “date of record, date and place of marriage, name, residence and official station of the person by whom solemnized; for each of the parties to be married the name, date and place of birth, residence, age, number of the marriage, as first or second, and if previously married, whether widowed or divorced, and the birth-given names of their parents.” G. L. c. 46, § 1.

10

“The record of a marriage made and kept as provided by law by the person by whom the marriage was solemnized, or by the clerk or registrar, or a copy thereof duly certified, shall be prima facie evidence of such marriage.” G. L. c. 207, § 45. A “certificate of the [c]ommissioner’s copy, signed by the [c]ommissioner or the [rjegistar, is admissible as evidence of the record.” Secretary of the Commonwealth v. City Clerk of Lowell, 373 Mass. 178, 181-182 (1977).

11

We use the terms “same sex” and “opposite sex” when characterizing the couples in question, because these terms are more accurate in this context than the terms “homosexual” or “heterosexual,” although at times we use those terms when we consider them appropriate. Nothing in our marriage law precludes people who identify themselves (or who are identified by others) as gay, lesbian, or bisexual from marrying persons of the opposite sex. See Baehr v. Lewin, 74 Haw. 530, 543 n.11, 547 n.14 (1993).

12

“The term public welfare has never been and cannot be precisely defined. Sometimes it has been said to include public convenience, comfort, peace and order, prosperity, and similar concepts, but not to include ‘mere expediency.’ ” Opinion of the Justices, 333 Mass. 773, 778 (1955).

13

For example, married persons face substantial restrictions, simply because they are married, on their ability freely to dispose of their assets. See, e.g., *323G. L. c. 208, § 34 (providing for payment of alimony and the equitable division of property on divorce); G. L. c. 191, § 15, and G. L. c. 189 (rights of elective share and dower).

14

Civil marriage enjoys a dual and in some sense paradoxical status as both a State-conferred benefit (with its attendant obligations) and a multi-faceted personal interest of “fundamental importance.” Zablocki v. Redhail, 434 U.S. 376, 383 (1978). As a practical matter, the State could not abolish civil marriage without chaotic consequences. The “right to marry,” id. at 387, is different from rights deemed “fundamental” for equal protection and due process purposes because the State could, in theory, abolish all civil marriage while it cannot, for example, abolish all private property rights.

15

The department argues that this case concerns the rights of couples (same-sex and opposite-sex), not the rights of individuals. This is incorrect. The rights implicated in this case are at the core of individual privacy and autonomy. See, e.g., Loving v. Virginia, 388 U.S. 1, 12 (1967) (“Under our Constitution, the freedom to marry or not marry, a person of another race resides with the individual and cannot be infringed by the State”); Perez v. Sharp, 32 Cal. 2d 711, 716 (1948) (“The right to marry is the right of individuals, not of racial groups”). See also A.Z. v. B.Z., 431 Mass. 150, 162 (2000), *327quoting Moore v. East Cleveland, 431 U.S. 494, 499 (1977) (noting “freedom of personal choice in matters of marriage and family life”). While two individuals who wish to marry may be equally aggrieved by State action denying them that opportunity, they do not “share” the liberty and equality interests at stake.

16

The department argues that the Loving decision did not profoundly alter the by-then common conception of marriage because it was decided at a time when antimiscegenation statutes were in “full-scale retreat.” But the relationship the department draws between popular consensus and the constitutionality of a statute oppressive to a minority group ignores the successful constitutional challenges to an antimiscegenation statute, initiated some twenty years earlier. When the Supreme Court of California decided Perez v. Sharp, 32 Cal. 2d 711, 728 (1948), a precursor to Loving, racial inequality was rampant and normative, segregation in public and private institutions was commonplace, the civil rights movement had not yet been launched, and the “separate but equal” doctrine of Plessy v. Ferguson, 163 U.S. 537 (1896), was still good law. The lack of popular consensus favoring integration (including interracial marriage) did not deter the Supreme Court of California from holding that that State’s antimiscegenation statute violated the plaintiffs’ *328constitutional rights. Neither the Perez court nor the Loving Court was content to permit an unconstitutional situation to fester because the remedy might not reflect a broad social consensus.

17

Recently, the United States Supreme Court has reaffirmed that the Constitution prohibits a State from wielding its formidable power to regulate conduct in a manner that demeans basic human dignity, even though that statutory discrimination may enjoy broad public support. The Court struck down a statute criminalizing sodomy. See Lawrence, supra at 2478 (“The liberty protected by the Constitution allows homosexual persons the right to make this choice”).

18

We have recognized that our Constitution may more extensively protect individual rights than the Federal Constitution in widely different contexts. See, e.g., Horsemen's Benevolent & Protective Ass’n v. State Racing Comm’n, 403 Mass. 692 (1989) (freedom from intrusive drug testing in highly regulated industry); Cepulonis v. Secretary of the Commonwealth, 389 Mass. 930 (1983) *329(inmates’ right to register to vote); Batchelder v. Allied Stores Int’l, Inc., 388 Mass. 83 (1983) (freedom to solicit signatures for ballot access in public election); Moe v. Secretary of Admin. & Fin., 382 Mass. 629 (1981) (right to State Medicaid payment for medically necessary abortions); Coffee-Rich, Inc. v. Commissioner of Pub. Health, 348 Mass. 414 (1965) (freedom to pursue one’s lawful business).

19

The Massachusetts Constitution empowers the General Court to enact only those orders, laws, statutes, and ordinances “wholesome and reasonable,” that are not “repugnant or contrary” to the Constitution, and that, in the Legislature’s judgment, advance the “good and welfare” of the Commonwealth, its government, and all of its subjects. Part n, c. 1, § 1, art. 4. See Opinion of the Justices, 360 Mass. 877, 883 (1971), quoting Jones v. Robbins, 8 Gray 329, 343 (1857) (powers vested in government are set down in Massachusetts Constitution “in a few plain, clear and intelligible *330propositions, for the better guidance and control, both of legislators and magistrates”).

20

Not every asserted rational relationship is a “conceivable” one, and rationality review is not “toothless.” Murphy v. Commissioner of the Dep’t of Indus. Accs., 415 Mass. 218, 233 (1993), citing Mathews v. Lucas, 427 U.S. 495, 510 (1976). Statutes have failed rational basis review even in circumstances where no fundamental right or “suspect” classification is implicated. See, e.g., Murphy v. Commissioner of the Dep’t of Indus. Accs., 415 Mass. 218, 226-227 (1993) (fee imposed on retention of counsel in administrative proceedings); Secretary of the Commonwealth v. City Clerk of Lowell, 373 Mass. 178, 186 (1977) (selection of surname for nonmarital *331child); Aetna Cas. & Sur. Co. v. Commissioner of Ins., 358 Mass. 272, 280-281 (1970) (automobile insurance ratesetting); Coffee-Rich, Inc. v. Commissioner of Pub. Health, 348 Mass. 414, 422 (1965) (sale of wholesome product); Mansfield Beauty Academy, Inc. v. Board of Registration of Hairdressers, 326 Mass. 624, 627 (1951) (right to charge for materials furnished to models by trade school); Opinion of the Justices, 322 Mass. 755, 760-761 (1948) (proposed statute concerning regulating cemeteries); Boston Elevated Ry. v. Commonwealth, 310 Mass. 528, 556-557 (1942) (legislation impairing contract right); Durgin v. Minot, 203 Mass. 26, 28 (1909) (statute authorizing certain board of health regulations).

21

Article 1 of the Massachusetts Constitution specifically prohibits sex-based discrimination. See post at 344-345 (Greaney, J,, concurring). We have not previously considered whether “sexual orientation” is a “suspect” classification. Our resolution of this case does not require that inquiry here.

22

Our marriage law does recognize that the inability to participate in intimate relations may have a bearing on one of the central expectations of marriage. Since the earliest days of the Commonwealth, the divorce statutes have permitted (but not required) a spouse to choose to divorce his or her impotent mate. See St. 1785, c. 69, § 3. While infertility is not a ground to void or terminate a marriage, impotency (the inability to engage in sexual intercourse) is, at the election of the disaffected spouse. See G. L. c. 207, § 14 (annulment); G. L. c. 208, § 1 (divorce). Cf. Martin v. Otis, 233 Mass. 491, 495 (1919) (“impotency does not render a marriage void, but only voidable at the suit of the party conceiving himself or herself to be wronged”); Smith v. Smith, 171 Mass. 404, 408 (1898) (marriage nullified because husband’s incurable syphilis “leaves him no foundation on which the marriage relation could properly rest”). See also G. L. c. 207, § 28A. However, in Hanson v. Hanson, 287 Mass. 154 (1934), a decree of annulment for nonconsummation was reversed where the wife knew before the marriage that her husband had syphilis and voluntarily chose to marry him. We held that, given the circumstances of the wife’s prior knowledge of the full extent of the disease and her consent to be married, the husband’s condition did not go “to the essence” of the marriage. Id. at 159.

23

It is hardly surprising that civil marriage developed historically as a means to regulate heterosexual conduct and to promote child rearing, because until very recently unassisted heterosexual relations were the only means short of adoption by which children could come into the world, and the absence of widely available and effective contraceptives made the link between heterosexual sex and procreation very strong indeed. Punitive notions of illegitimacy, see Powers v. Wilkinson, 399 Mass. 650, 661 (1987), and of homosexual identity, see Lawrence, supra at 2478-2479, further cemented the common and legal understanding of marriage as an unquestionably heterosexual institution. But it is circular reasoning, not analysis, to maintain that marriage must remain a heterosexual institution because that is what it historically has been. As one dissent acknowledges, in “the modem age,” “heterosexual intercourse, procreation, and child care are not necessarily conjoined.” Post at 382 (Cordy, J., dissenting).

24

Adoption and certain insurance coverage for assisted reproductive technology are available to married couples, same-sex couples, and single individuals alike. See G. L. c. 210, § 1; Adoption of Tammy, 416 Mass. 205 (1993) (adoption); G. L. c. 175, § 47H; G. L. c. 176A, § 8K; G. L. c. 176B, § 4J; and G. L. c. 176G, § 4 (insurance coverage). See also Woodward v. Commissioner of Social Sec., 435 Mass. 536, 546 (2002) (posthumous reproduction); Culliton v. Beth Israel Deaconness Med. Ctr., 435 Mass. 285, 293 (2001) (gestational surrogacy).

25

Because our laws expressly or implicitly sanction so many kinds of opposite-sex marriages that do not or will never result in unassisted reproduction, it is erroneous to claim, as the dissent does, that the “theoretical]” procreative capacity of opposite-sex couples, post at 391 (Cordy, J., dissenting), sufficiently justifies excluding from civil marriage same-sex couples who actually have children.

26

The claim that the constitutional rights to bear and raise a child are “not implicated or infringed” by the marriage ban, post at 371 (Cordy, J., dissenting), does not stand up to scrutiny. The absolute foreclosure of the marriage option for the class of parents and would-be parents at issue here imposes a heavy burden on their decision to have and raise children that is not suffered by any other class of parent.

27

It is also trac that civil marriage creates legal dependency between spouses, which is simply not available to unmarried couples. See Part HI A, supra.

28

Justice Cordy suggests that we have “transmuted the ‘right’ to marry into a right to change the institution of marriage itself,” post at 365 (Cordy, J., dissenting), because marriage is intimately tied to the reproductive systems of the marriage partners and to the “optimal” mother and father setting for child rearing. Id. That analysis hews perilously close to the argument, long repudiated by the Legislature and the courts, that men and women are so innately and fundamentally different that their respective “proper spheres” can be rigidly and universally delineated. An abundance of legislative enactments and decisions of this court negate any such stereotypical premises.

29

We are concerned only with the withholding of the benefits, protections, and obligations of civil marriage from a certain class of persons for invalid reasons. Our decision in no way limits the rights of individuals to refuse to *338marry persons of the same sex for religious or any other reasons. It in no way limits the personal freedom to disapprove of, or to encourage others to disapprove of, same-sex marriage. Our concern, rather, is whether historical, cultural, religious, or other reasons permit the State to impose limits on personal beliefs concerning whom a person should marry.

30

Justice Cordy’s dissenting opinion, post at 386-388 and nn.24-28 (Cordy, J., dissenting), makes much of the current “battle of the experts” concerning the possible long-term effects on children of being raised in households headed by same-sex parents. We presume that the Legislature is aware of these studies, see Mutual Loan Co. v. Martell, 200 Mass. 482, 487 (1909), aff’d, 222 U.S. 225 (1911), and has drawn the conclusion that a child’s best interest is not harmed by being raised and nurtured by same-sex parents. See G. L. c. 210, § 7. See also Adoption of Tammy, 416 Mass. 205 (1993); 110 Code Mass. Regs. § 1.09 (3) (2000) (“The Department [of Social Services] shall not deny to any person the opportunity to become an adoptive or foster parent, on the basis of the . . . sexual orientation ... of the person, or of the child, involved”). Either the Legislature’s openness to same-sex parenting is rational in light of its paramount interests in promoting children’s well-being, or irrational in light of its so-called conclusion that a household headed by opposite-sex married parents is the “optimal” setting for raising children. See post at 392 (Cordy, 1, dissenting). We give full credit to the Legislature for enacting a statutory scheme of child-related laws that is coherent, consistent, and harmonious. See New England Div. of the Am. Cancer Soc’y v. Commissioner of Admin., 437 Mass. 172, 180 (2002).

31

If total deference to the Legislature were the case, the judiciary would be stripped of its constitutional authority to decide challenges to statutes pertaining to marriage, child rearing, and family relationships, and, conceivably, unconstitutional laws that provided for the forced sterilization of habitual criminals; prohibited miscegenation; required court approval for the marriage of persons with child support obligations; compelled a pregnant unmarried minor to obtain the consent of both parents before undergoing an abortion; and made sodomy a criminal offense, to name just a few, would stand.

Indeed, every State court that has recently considered the issue we decide today has exercised its duty in the same way, by carefully scrutinizing the statutory ban on same-sex marriages in light of relevant State constitutional provisions. See Brause vs. Bureau of Vital Statistics, No. 3AN-95-6562CJ (Alaska Super. Ct., Feb. 27, 1998) (concluding marriage statute violated right to privacy provision in Alaska Constitution) (superseded by constitutional amendment, art. I, § 25 of Constitution of Alaska); Baehr v. Lewin, 74 Haw. 530, 571-580 (1993) (concluding marriage statute implicated Hawaii Constitution’s equal protection clause; remanding case to lower court for further proceedings); Baker v. State, 170 Vt. 194, 197-198 (1999) (concluding marriage statute violated Vermont Constitution’s common benefits clause). But see Standhardt v. Superior Court, 77 P.3d 451 (Ariz. Ct. App. 2003) (marriage statute does not violate liberty interests under either Federal or Arizona Constitution). See also Halpern v. Toronto (City), 172 O.A.C. 276 (2003) (concluding marriage statute violated equal protection provisions of Canada’s Charter of Rights and Freedoms); EGALE Canada, Inc. v. Canada (Attorney Gen.), 13 B.C.L.R. (4th) 1 (2003) (same).

32

One prominent historian of marriage notes, for example, that in the Nineteenth Century, the Reverend Theodore Woolsey led the charge against expanding the grounds for divorce, arguing that the “the only divinely approved (and therefore truly legitimate) reason for divorce was adultery” and that only the innocent party to a marriage terminated by reason of adultery be permitted to remarry. Cott, Public Vows: A History of Marriage and the Nation 106 (2000). See id. at 44-45, for a general discussion of resistance to the demise of antimiscegenation laws.

33

It is not dispositive, for purposes of our constitutional analysis, whether the Legislature, at the time it incorporated the common-law definition of marriage into the first marriage laws nearly three centuries ago, did so with the intent of discriminating against or harming persons who wish to marry another of the same sex. We are not required to impute an invidious intent to the Legislature in determining that a statute of long standing has no applicability *342to present circumstances or violates the rights of individuals under the Massachusetts Constitution. That the Legislature may have intended what at the time of enactment was a perfectly reasonable form of discrimination — or a result not recognized as a form of discrimination — was not enough to salvage from later constitutional challenge laws burdening nonmarital children or denying women’s equal partnership in marriage. See, e.g., Trimble v. Gordon, 430 U.S. 762 (1977) (nonmarital children); Angelini v. OMD Corp., 410 Mass. 653, 662, 663 (1987) (“The traditional common law rules which discriminated against children bom out of wedlock have been discarded” and “[w]e have recognized that placing additional burdens on [nonmarital] children is unfair because they are not responsible for their [status]”); Silvia v. Silvia, 9 Mass. App. Ct. 339, 340-341 (1980) (there now exists “a comprehensive statutory and common law pattern which places marital and parental obligations on both the husband and wife”). We are concerned with the operation of challenged laws on the parties before us, and we do not inhibit our inquiry on the ground that a statute’s original enactors had a benign or at the time constitutionally unassailable purpose. See Colo v. Treasurer & Receiver Gen., 378 Mass. 550, 557 (1979), quoting Walz v. Tax Comm’n of the City of N.Y., 397 U.S. 664, 678 (1970) (“the mere fact that a certain practice has gone unchallenged for a long period of time cannot alone immunize it from constitutional invalidity, ‘even when that span of time covers our entire national existence and indeed predates it’ ”); Merit Oil Co. v. Director of Div. on the Necessaries of Life, 319 Mass. 301, 305 (1946) (constitutional contours of State’s regulatory authority coextensive “with the changing needs of society”).

34

Similarly, no one argues that the restrictions on incestuous or polygamous marriages are so dependent on the marriage restriction that they too should fall if the marriage restriction falls. Nothing in our opinion today should be construed as relaxing or abrogating the consanguinity or polygamy prohibitions of our marriage laws. See G. L. c. 207, §§ 1, 2, and 4. Rather, the statutory provisions concerning consanguinity or polygamous marriages shall be construed in a gender neutral manner. See Califano v. Westcott, 443 U.S. 76, 92-93 (1979) (construing word “father” in unconstitutional, underinclusive provision to mean “parent”); Browne’s Case, 322 Mass. 429, 430 (1948) (construing masculine pronoun “his” to include feminine pronoun “her”). See also G. L. c. 4, § 6, Fourth (“words of one gender may be construed to include the other gender and the neuter” unless such construction would be “inconsistent with the manifest intent of the law-making body or repugnant to the context of the same statute”).

Greaney, J.

(concurring). I agree with the result reached by the court, the remedy ordered, and much of the reasoning in the court’s opinion. In my view, however, the case is more directly resolved using traditional equal protection analysis.

(a) Article 1 of the Declaration of Rights, as amended by art. 106 of the Amendments to the Massachusetts Constitution, provides:

“All people are bom free and equal and have certain natural, essential and unalienable rights; among which may be reckoned the right of enjoying and defending their lives and liberties; that of acquiring, possessing and protecting property; in fine, that of seeking and obtaining their safety and happiness. Equality under the law shall not be denied or abridged because of sex, race, color, creed or national origin.”

This provision, even prior to its amendment, guaranteed to all people in the Commonwealth — equally — the enjoyment of rights that are deemed important or fundamental. The withholding of relief from the plaintiffs, who wish to marry, and are *345otherwise eligible to marry, on the ground that the couples are of the same gender, constitutes a categorical restriction of a fundamental right. The restriction creates a straightforward case of discrimination that disqualifies an entire group of our citizens and their families from participation in an institution of paramount legal and social importance. This is impermissible under art. 1.

Analysis begins with the indisputable premise that the deprivation suffered by the plaintiffs is no mere legal inconvenience. The right to marry is not a privilege conferred by the State, but a fundamental right that is protected against unwarranted State interference. See Zablocki v. Redhail, 434 U.S. 374, 384 (1978) (“the right to marry is of fundamental importance for all individuals”); Loving v. Virginia, 388 U.S. 1, 12 (1967) (freedom to marry is “one of the vital personal rights essential to the orderly pursuit of happiness by free men” under due process clause of Fourteenth Amendment); Skinner v. Oklahoma, 316 U.S. 535, 541 (1942) (marriage is one of “basic civil rights of man”). See also Turner v. Safley, 482 U.S. 78, 95-96 (1987) (prisoners’ right to marry is constitutionally protected). This right is essentially vitiated if one is denied the right to marry a person of one’s choice. See Zablocki v. Redhail, supra at 384 (all recent decisions of United States Supreme Court place “the decision to marry as among the personal decisions protected by the right of privacy”).1

Because our marriage statutes intend, and state, the ordinary understanding that marriage under our law consists only of a union between a man and a woman, they create a statutory classification based on the sex of the two people who wish to marry. See Baehr v. Lewin, 74 Haw. 530, 564 (1993) (plurality opinion) (Hawaii marriage statutes created sex-based classification); Baker v. State, 170 Vt. 194, 253 (1999) (Johnson, J., concurring in part and dissenting in part) (same). That the classification is *346sex based is self-evident. The marriage statutes prohibit some applicants, such as the plaintiffs, from obtaining a marriage license, and that prohibition is based solely on the applicants’ gender. As a factual matter, an individual’s choice of marital partner is constrained because of his or her own sex. Stated in particular terms, Hillary Goodridge cannot marry Julie Goodridge because she (Hillary) is a woman. Likewise, Gary Chalmers cannot marry Richard Linnell because he (Gary) is a man. Only their gender prevents Hillary and Gary from marrying their chosen partners under the present law.2

A classification may be gender based whether or not the challenged government action apportions benefits or burdens uniformly along gender lines. This is so because constitutional protections extend to individuals and not to categories of people. Thus, when an individual desires to marry, but cannot marry his or her chosen partner because of the traditional opposite-sex restriction, a violation of art. 1 has occurred. See Commonwealth v. Chou, 433 Mass. 229, 237-238 (2001) (assuming statute enforceable only across gender lines may offend Massachusetts equal rights amendment). I find it disingenuous, at best, to suggest that such an individual’s right to marry has not been burdened at all, because he or she remains free to chose another partner, who is of the opposite sex.

The equal protection infirmity at work here is strikingly similar to (although, perhaps, more subtle than) the invidious discrimination perpetuated by Virginia’s antimiscegenation laws *347and unveiled in the decision of Loving v. Virginia, supra. In its landmark decision striking down Virginia’s ban on marriages between Caucasians and members of any other race on both equal protection and substantive due process grounds, the United States Supreme Court soundly rejected the proposition that the equal application of the ban (i.e., that it applied equally to whites and blacks) made unnecessary the strict scrutiny analysis traditionally required of statutes drawing classifications according to race, see id. at 8-9, and concluded that “restricting the freedom to marry solely because of racial classifications violates the central meaning of the Equal Protection Clause.” Id. at 12. That our marriage laws, unlike antimiscegenation laws, were not enacted purposely to discriminate in no way neutralizes their present discriminatory character.

With these two propositions established (the infringement on a fundamental right and a sex-based classification), the enforcement of the marriage statutes as they are currently understood is forbidden by our Constitution unless the State can present a compelling purpose furthered by the statutes that can be accomplished in no other reasonable manner.3 See Blixt v. Blixt, 437 Mass. 649, 655-656 (2002), cert, denied, 537 U.S. 1189 (2003); Lowell v. Kowalski, 380 Mass. 663, 667-669 (1980). This the State has not done. The justifications put forth by the State to sustain the statute’s exclusion of the plaintiffs are insufficient for the reasons explained by the court, to which I add the following observations.

The rights of couples to have children, to adopt, and to be foster parents, regardless of sexual orientation and marital status, are firmly established. See E.N.O. v. L.M.M., 429 Mass. 824, 829, cert, denied, 528 U.S. 1005 (1999); Adoption of Tammy, 416 Mass. 205, 210-211 (1993). As recognized in the court’s opinion, and demonstrated by the record in this case, however, *348the State’s refusal to accord legal recognition to unions of same-sex couples has had the effect of creating a system in which children of same-sex couples are unable to partake of legal protections and social benefits taken for granted by children in families whose parents are of the opposite sex. The continued maintenance of this caste-like system is irreconcilable with, indeed, totally repugnant to, the State’s strong interest in the welfare of all children and its primary focus, in the context of family law where children are concerned, on “the best interests of the child.” The issue at stake is not one, as might ordinarily be the case, that can be unilaterally and totally deferred to the wisdom of the Legislature. “While the State retains wide latitude to decide the manner in which it will allocate benefits, it may not use criteria which discriminatorily burden the exercise of a fundamental right.” Moe v. Secretary of Admin. & Fin., 382 Mass. 629, 652 (1981). Nor can the State’s wish to conserve resources be accomplished by invidious distinctions between classes of citizens. See Plyler v. Doe, 457 U.S. 202, 216-217, 227 (1982).4

A comment is in order with respect to the insistence of some that marriage is, as a matter of definition, the legal union of a man and a woman. To define the institution of marriage by the characteristics of those to whom it always has been accessible, in order to justify the exclusion of those to whom it never has been accessible, is conclusory and bypasses the core question we are asked to decide.5 This case calls for a higher level of legal analysis. Precisely, the case requires that we confront ingrained assumptions with respect to historically accepted roles of men and women within the institution of marriage and requires that we reexamine these assumptions in light of the *349unequivocal language of art. 1, in order to ensure that the governmental conduct challenged here conforms to the supreme charter of our Commonwealth. “A written constitution is the fundamental law for the government of a sovereign State. It is the final statement of the rights, privileges and obligations of the citizens and the ultimate grant of the powers and the conclusive definition of the limitations of the departments of State and of public officers . . . . To its provisions the conduct of all governmental affairs must conform. From its terms there is no appeal.” Loring v. Young, 239 Mass. 349, 376-377 (1921). I do not doubt the sincerity of deeply held moral or religious beliefs that make inconceivable to some the notion that any change in the common-law definition of what constitutes a legal civil marriage is now, or ever would be, warranted. But, as matter of constitutional law, neither the mantra of tradition, nor individual conviction, can justify the perpetuation of a hierarchy in which couples of the same sex and their families are deemed less worthy of social and legal recognition than couples of the opposite sex and their families. See Lawrence v. Texas, 123 S. Ct. 2472, 2486 (2003) (O’Connor, J., concurring) (moral disapproval, with no other valid State interest, cannot justify law that discriminates against groups of persons); Planned Parenthood of Southeastern Pa. v. Casey, 505 U.S. 833, 850 (1992) (“Our obligation is to define the liberty of all, not to mandate our own moral code”).

(b) I am hopeful that our decision will be accepted by those thoughtful citizens who believe that same-sex unions should not be approved by the State. I am not referring here to acceptance in the sense of grudging acknowledgment of the court’s authority to adjudicate the matter. My hope is more liberating. The plaintiffs are members of our community, our neighbors, our coworkers, our friends. As pointed out by the court, their professions include investment advisor, computer engineer, teacher, therapist, and lawyer. The plaintiffs volunteer in our schools, worship beside us in our religious houses, and have children who play with our children, to mention just a few ordinary daily contacts. We share a common humanity and participate together in the social contract that is the foundation of our Commonwealth. Simple principles of decency dictate that we *350extend to the plaintiffs, and to their new status, full acceptance, tolerance, and respect. We should do so because it is the right thing to do. The union of two people contemplated by G. L. c. 207 “is a coming together for better or for worse, hopefully enduring, and intimate to the degree of being sacred. It is an association that promotes a way of life, not causes; a harmony in living, not political faiths; a bilateral loyalty, not commercial or social projects. Yet it is an association for as noble a purpose as any involved in our prior decisions.” Griswold v. Connecticut, 381 U.S. 479, 486 (1965). Because of the terms of art. 1, the plaintiffs will no longer be excluded from that association.6

1

It makes no difference that the referenced decisions consider the right to marry in the context of the Fourteenth Amendment to the United States Constitution rather than in the context of our Constitution. As explained by the court, ante at 328 n.18, a fundamental right under the Federal Constitution enjoys at least a comparable measure of protection under our State Constitution. See Moe v. Secretary of Admin. & Fin., 382 Mass. 629, 651 (1981).

2

In her separate opinion in Baker v. State, 170 Vt. 194, 253 (1999) (Johnson, J., concurring in part and dissenting in part), Justice Johnson described the equal protection defect in Vermont’s marriage statutes in a slightly different, but no less persuasive, fashion:

“A woman is denied the right to marry another woman because her would-be partner is a woman, not because one or both are lesbians. Similarly, a man is denied the right to marry another man because his would-be partner is a man, not because one or both are gay. Thus, an individual’s right to marry a person of the same sex is prohibited solely on the basis of sex, not on the basis of sexual orientation. Indeed, sexual orientation does not appear as a qualification for marriage under the marriage statutes. The State makes no inquiry into the sexual practices or identities of a couple seeking a license.”

3

Some might say that the use of the so-called strict scrutiny formula is too facile in the sense that, once a court focuses on the formula as a dispositional tool, the result is automatically preordained — the statute will fail because the State cannot possibly sustain its heavy burden to overcome the presumption of arbitrary and invidious discrimination. This is not so. See, e.g., Blixt v. Blixt, 437 Mass. 649, 656-657 (2002), cert, denied, 537 U.S. 1189 (2003) (concluding G. L. c. 119, § 39D, grandparent visitation statute, furthered compelling State interest in mitigating potential harm to children in nonintact families).

4

The argument, made by some in the case, that legalization of same-sex marriage in Massachusetts will be used by persons in other States as a tool to obtain recognition of a marriage in their State that is otherwise unlawful, is precluded by the provisions of G. L. c. 207, §§ 11, 12, and 13.

5

Because marriage is, by all accounts, the cornerstone of our social structure, as well as the defining relationship in our personal lives, confining eligibility in the institution, and all of its accompanying benefits and responsibilities, to opposite-sex couples is basely unfair. To justify the restriction in our marriage laws by accusing the plaintiffs of attempting to change the institution of marriage itself terminates the debate at the outset without any accompanying reasoned analysis.

6

Justice Cordy’s separate opinion points out, correctly, that, when art. 1 was revised by the people in 1976, it was not then intended to be relied on to approve same-sex marriage. Post at 377-379 (Cordy, J., dissenting). (Justice Spina adverts to the same proposition in his separate opinion, post at 355 [Spina, J., dissenting]). Decisions construing the provision cited in Justice Cordy’s opinion are interesting, but obviously inapposite because they have not dealt in any significant way with the issue before us. Nonetheless, the separate opinion concludes, from what was intended in 1976, and from various cases discussing art. 1, that the revised provision cannot be used to justify the result I reach.

In so reasoning, the separate opinion places itself squarely on the side of the original intent school of constitutional interpretation. As a general principle, I do not accept the philosophy of the school. The Massachusetts Constitution was never meant to create dogma that adopts inflexible views of one time to deny lawful rights to those who live in another. The provisions of our Constitution are, and must be, adaptable to changing circumstances and new societal phenomena, and, unless and until the people speak again on a specific subject, conformable in their concepts of liberty and equality to what is fair, right, and just. I am cognizant of the voters’ intent in passing the amendment to art. 1 in 1976. Were the revision alone the basis for change, I would be reluctant to construe it favorably to the plaintiffs, in view of the amendment’s recent passage and the voters’ intent. The court’s opinion, however, rests in part on well-established principles of equal protection that are independent of the amendment. It is on these principles that I base my opinion.

Spina, J.

(dissenting, with whom Sosman and Cordy, JJ., join). What is at stake in this case is not the unequal treatment of individuals or whether individual rights have been impermissibly burdened, but the power of the Legislature to effectuate social change without interference from the courts, pursuant to *351art. 30 of the Massachusetts Declaration of Rights.1 The power to regulate marriage lies with the Legislature, not with the judiciary. See Commonwealth v. Stowell, 389 Mass. 171, 175 (1983). Today, the court has transformed its role as protector of individual rights into the role of creator of rights, and I respectfully dissent.

1. Equal protection. Although the court did not address the plaintiffs’ gender discrimination claim, G. L. c. 207 does not unconstitutionally discriminate on the basis of gender.2 A claim of gender discrimination will lie where it is shown that differential treatment disadvantages one sex over the other. See Attorney Gen. v. Massachusetts Interscholastic Athletic Ass’n, 378 Mass. 342, 349-352 (1979). See also United States v. Virginia, 518 U.S. 515 (1996). General Laws c. 207 enumerates certain qualifications for obtaining a marriage license. It creates no distinction between the sexes, but applies to men and women in precisely the same way. It does not create any disadvantage identified with gender, as both men and women are similarly limited to marrying a person of the opposite sex. See Commonwealth v. King, 374 Mass. 5, 15-22 (1977) (law prohibiting prostitution not discriminatory based on gender because of equal application to men and women).

Similarly, the marriage statutes do not discriminate on the basis of sexual orientation. As the court correctly recognizes, constitutional protections are extended to individuals, not couples. Ante at 326 n.15. The marriage statutes do not disqualify individuals on the basis of sexual orientation from entering into marriage. All individuals, with certain exceptions not relevant here, are free to marry. Whether an individual chooses not to marry because of sexual orientation or any other reason should be of no concern to the court.

The court concludes, however, that G. L. c. 207 unconstitutionally discriminates against the individual plaintiffs because it *352denies them the “right to marry the person of one’s choice” where that person is of the same sex. Ante at 328. To reach this result the court relies on Loving v. Virginia, 388 U.S. 1, 12 (1967), and transforms “choice” into the essential element of the institution of marriage. The Loving case did not use the word “choice” in this manner, and it did not point to the result that the court reaches today. In Loving, the Supreme Court struck down as unconstitutional a statute that prohibited Caucasians from marrying non-Caucasians. It concluded that the statute was intended to preserve white supremacy and invidiously discriminated against non-Caucasians because of their race. See id. at 11-12. The “choice” to which the Supreme Court referred was the “choice to marry,” and it concluded that with respect to the institution of marriage, the State had no compelling interest in limiting the choice to marry along racial lines. Id. The Supreme Court did not imply the existence of a right to marry a person of the same sex. To the same effect is Perez v. Sharp, 32 Cal. 2d 711 (1948), on which the court also relies.

Unlike the Loving and Sharp cases, the Massachusetts Legislature has erected no barrier to marriage that intentionally discriminates against anyone. Within the institution of marriage,3 anyone is free to marry, with certain exceptions that are not challenged. In the absence of any discriminatory purpose, the State’s marriage statutes do not violate principles of equal protection. See Washington v. Davis, 426 U.S. 229, 240 (1976) (“invidious quality of a law claimed to be . . . discriminatory must ultimately be traced to a . . . discriminatory purpose”); Dickerson v. Attorney Gen., 396 Mass. 740, 743 (1986) (for purpose of equal protection analysis, standard of review under State and Federal Constitutions is identical). See also Attorney Gen. v. Massachusetts Interscholastic Athletic Ass’n, supra. This court should not have invoked even the most deferential standard of review within equal protection analysis because no individual was denied access to the institution of marriage.

2. Due process. The marriage statutes do not impermissibly burden a right protected by our constitutional guarantee of due *353process implicit in art. 10 of our Declaration of Rights. There is no restriction on the right of any plaintiff to enter into marriage. Each is free to marry a willing person of the opposite sex. Cf. Zablocki v. Redhail, 434 U.S. 374 (1978) (fundamental right to marry impermissibly burdened by statute requiring court approval when subject to child support order).

Substantive due process protects individual rights against unwarranted government intrusion. See Aime v. Commonwealth, 414 Mass. 667, 673 (1993). The court states, as we have said on many occasions, that the Massachusetts Declaration of Rights may protect a right in ways that exceed the protection afforded by the Federal Constitution. Ante at 328. See Arizona v. Evans, 514 U.S. 1, 8 (1995) (State courts afforded broader protection of rights than granted by United States Constitution). However, today the court does not fashion a remedy that affords greater protection of a right. Instead, using the rubric of due process, it has redefined marriage.

Although art. 10 may afford greater protection of rights than the due process clause of the Fourteenth Amendment, our treatment of due process challenges adheres to the same standards followed in Federal due process analysis. See Commonwealth v. Ellis, 429 Mass. 362, 371 (1999). When analyzing a claim that the State has impermissibly burdened an individual’s fundamental or other right or liberty interest, “[w]e begin by sketching the contours of the right asserted. We then inquire whether the challenged restriction burdens that right.” Moe v. Secretary of Admin. & Fin., 382 Mass. 629, 646 (1981). Where a right deemed “fundamental” is implicated, the challenged restriction will be upheld only if it is “narrowly tailored to further a legitimate and compelling governmental interest.” Aime v. Commonwealth, supra at 673. To qualify as “fundamental” the asserted right must be “objectively, ‘deeply rooted in this Nation’s history and tradition,’ [Moore v. East Cleveland, 431 U.S. 494, 503 (1977) (plurality opinion)] . . . and ‘implicit in the concept of ordered liberty,’ such that ‘neither liberty nor justice would exist if they were sacrificed.’ ” Washington v. Glucksberg, 521 U.S. 702, 720-721 (1997), quoting Palko v. Connecticut, 302 U.S. 319, 325, 326 (1937) (right to assisted suicide does not fall within fundamental right to refuse medical treatment because *354novel and unsupported by tradition) (citations omitted). See Three Juveniles v. Commonwealth, 390 Mass. 357, 367 (1983) (O’Connor, J., dissenting), cert, denied sub nom. Keefe v. Massachusetts, 465 U.S. 1068 (1984). Rights that are not considered fundamental merit due process protection if they have been irrationally burdened. See Massachusetts Fed’n of Teachers v. Board of Educ., 436 Mass. 763, 777-779 & n.14 (2002).

Although this court did not state that same-sex marriage is a fundamental right worthy of strict scrutiny protection, it nonetheless deemed it a constitutionally protected right by applying rational basis review. Before applying any level of constitutional analysis there must be a recognized right at stake. Same-sex marriage, or the “right to marry the person of one’s choice” as the court today defines that right, does not fall within the fundamental right to marry. Same-sex marriage is not “deeply rooted in this Nation’s history,” and the court does not suggest that it is. Except for the occasional isolated decision in recent years, see, e.g., Baker v. State, 170 Vt. 194 (1999), same-sex marriage is not a right, fundamental or otherwise, recognized in this country. Just one example of the Legislature’s refusal to recognize same-sex marriage can be found in a section of the legislation amending G. L. c. 15 IB to prohibit discrimination in the workplace on the basis of sexual orientation, which states: “Nothing in this act shall be construed so as to legitimize or validate a ‘homosexual marriage’. . . .” St. 1989, c. 516, § 19. In this Commonwealth and in this country, the roots of the institution of marriage are deeply set in history as a civil union between a single man and a single woman. There is no basis for the court to recognize same-sex marriage as a constitutionally protected right.

3. Remedy. The remedy that the court has fashioned both in the name of equal protection and due process exceeds the bounds of judicial restraint mandated by art. 30. The remedy that construes gender-specific language as gender-neutral amounts to a statutory revision that replaces the intent of the Legislature with that of the court. Article 30 permits the court to apply principles of equal protection and to modify statutory language only if legislative intent is preserved. See, e.g., Commonwealth v. Chou, 433 Mass. 229, 238-239 (2001) (judicial *355rewriting of gender language permissible only when Legislature intended to include both men and women). See also Lowell v. Kowalski, 380 Mass. 663, 670 (1980). Here, the alteration of the gender-specific language alters precisely what the Legislature unambiguously intended to preserve, the marital rights of single men and women. Such a dramatic change in social institutions must remain at the behest of the people through the democratic process.

Where the application of equal protection principles do not permit rewriting a statute in a manner that preserves the intent of the Legislature, we do not rewrite the statute. In Dalli v. Board of Educ., 358 Mass. 753 (1971), the court refused to rewrite a statute in a manner that would include unintended individuals. “To attempt to interpret this [statute] as including those in the category of the plaintiff would be to engage in a judicial enlargement of the clear statutory language beyond the limit of our judicial function. We have traditionally and consistently declined to trespass on legislative territory in deference to the time tested wisdom of the separation of powers as expressed in art. [30] of the Declaration of Rights of the Constitution of Massachusetts even when it appeared that a highly desirable and just result might thus be achieved.” Id. at 759. Recently, in Connors v. Boston, 430 Mass. 31 (1999), we refused to expand health insurance coverage to include domestic partners because such an expansion was within the province of the Legislature, where policy affecting family relationships is most appropriate and frequently considered. Id. at 42-43. Principles of equal protection do not permit the marriage statutes to be changed in the manner that we have seen today.

This court has previously exercised the judicial restraint mandated by art. 30 and declined to extend due process protection to rights not traditionally coveted, despite recognition of their social importance. See Tobin’s Case, 424 Mass. 250, 252-253 (1997) (receiving workers’ compensation benefits not fundamental right); Doe v. Superintendent of Schs. of Worcester, 421 Mass. 117, 129 (1995) (declaring education not fundamental right); Williams v. Secretary of the Executive Office of Human Servs., 414 Mass. 551, 565 (1993) (no fundamental right to receive mental health services); Matter of Tocci, 413 Mass. 542, *356548 n.4 (1992) (no fundamental right to practice law); Commonwealth v. Henry’s Drywall Co., 366 Mass. 539, 542 (1974) (no fundamental right to pursue one’s business). Courts have authority to recognize rights that are supported by the Constitution and history, but the power to create novel rights is reserved for the people through the democratic and legislative processes.

Likewise, the Supreme Court exercises restraint in the application of substantive due process “ ‘because guideposts for responsible decisionmaking in this unchartered area are scarce and open-ended.’ [Collins v. Harker Heights, 503 U.S. 115, 125 (1992).] By extending constitutional protection to an asserted right or liberty interest, we, to a great extent, place the matter outside the arena of public debate and legislative action. We must therefore ‘exercise the utmost care whenever we are asked to break new ground in this field,’ [id.], lest the liberty protected by the Due Process Clause be subtly transformed into the policy preferences of the Members of this Court, Moore [v. East Cleveland, 431 U.S. 494, 502 (1977)] (plurality opinion).” Washington v. Glucksberg, supra at 720.

The court has extruded a new right from principles of substantive due process, and in doing so it has distorted the meaning and purpose of due process. The purpose of substantive due process is to protect existing rights, not to create new rights. Its aim is to thwart government intrusion, not invite it. The court asserts that the Massachusetts Declaration of Rights serves to guard against government intrusion into each individual’s sphere of privacy. Ante at 329. Similarly, the Supreme Court has called for increased due process protection when individual privacy and intimacy are threatened by unnecessary government imposition. See, e.g., Lawrence v. Texas, 123 S. Ct. 2472 (2003) (private nature of sexual behavior implicates increased due process protection); Eisenstadt v. Baird, 405 U.S. 438 (1972) (privacy protection extended to procreation decisions within nonmarital context); Griswold v. Connecticut, 381 U.S. 479 (1965) (due process invoked because of intimate nature of procreation decisions). These cases, along with the Moe case, focus on the threat to privacy when government seeks to regulate the most intimate activity behind bedroom doors. The statute in question does not seek to regulate intimate activity *357within an intimate relationship, but merely gives formal recognition to a particular marriage. The State has respected the private lives of the plaintiffs, and has done nothing to intrude in the relationships that each of the plaintiff couples enjoy. Cf. Lawrence v. Texas, supra at 2484 (case “does not involve whether the government must give formal recognition to any relationship that homosexual persons seek to enter”). Ironically, by extending the marriage laws to same-sex couples the court has turned substantive due process on its head and used it to interject government into the plaintiffs’ lives.

1

Article 30 of the Massachusetts Declaration of Rights provides that “the judicial [department] shall never exercise the legislative and executive powers ... to the end it may be a government of laws and not of men.”

2

Article 1 of the Massachusetts Declaration of Rights, as amended by art. 106 of the Amendments, the Equal Rights Amendment, states: “Equality under the law shall not be denied or abridged because of sex, race, color, creed or national origin.”

3

Marriage is the civil union between a single man and a single woman. See Milford v. Worcester, 1 Mass. 48, 52 (1810).

Sosman, J.

(dissenting, with whom Spina and Cordy, JJ., join). In applying the rational basis test to any challenged statutory scheme, the issue is not whether the Legislature’s rationale behind that scheme is persuasive to us, but only whether it satisfies a minimal threshold of rationality. Today, rather than apply that test, the court announces that, because it is persuaded that there are no differences between same-sex and opposite-sex couples, the Legislature has no rational basis for treating them differently with respect to the granting of marriage licenses.1 Reduced to its essence, the court’s opinion concludes that, because same-sex couples are now raising children, and withholding the benefits of civil marriage from their union makes it harder for them to raise those children, the State must therefore provide the benefits of civil marriage to same-sex couples just as it does to opposite-sex couples. Of course, many people are raising children outside the confines of traditional marriage, and, by definition, those children are being deprived of the various benefits that would flow if they were being raised in a household with married parents. That does not mean that the *358Legislature must accord the full benefits of marital status on every household raising children. Rather, the Legislature need only have some rational basis for concluding that, at present, those alternate family structures have not yet been conclusively shown to be the equivalent of the marital family structure that has established itself as a successful one over a period of centuries. People are of course at liberty to raise their children in various family structures, so long as they are not literally harming their children by doing so. See Blixt v. Blixt, 437 Mass. 649, 668-670 (2002) (Sosman, J., dissenting), cert, denied, 537 U.S. 1189 (2003). That does not mean that the State is required to provide identical forms of encouragement, endorsement, and support to all of the infinite variety of household structures that a free society permits.

Based on our own philosophy of child rearing, and on our observations of the children being raised by same-sex couples to whom we are personally close, we may be of the view that what matters to children is not the gender, or sexual orientation, or even the number of the adults who raise them, but rather whether those adults provide the children with a nurturing, stable, safe, consistent, and supportive environment in which to mature. Same-sex couples can provide their children with the requisite nurturing, stable, safe, consistent, and supportive environment in which to mature, just as opposite-sex couples do. It is therefore understandable that the court might view the traditional definition of marriage as an unnecessary anachronism, rooted in historical prejudices that modem society has in large measure rejected and biological limitations that modem science has overcome.

It is not, however, our assessment that matters. Conspicuously absent from the court’s opinion today is any acknowledgment that the attempts at scientific study of the ramifications of raising children in same-sex couple households are themselves in their infancy and have so far produced inconclusive and conflicting results. Notwithstanding our belief that gender and sexual orientation of parents should not matter to the success of the child rearing venture, studies to date reveal that there are still some observable differences between children raised by opposite-sex couples and children raised by same-sex couples. *359See post at 386-387 (Cordy, J., dissenting). Interpretation of the data gathered by those studies then becomes clouded by the personal and political beliefs of the investigators, both as to whether the differences identified are positive or negative, and as to the untested explanations of what might account for those differences. (This is hardly the first time in history that the ostensible steel of the scientific method has melted and buckled under the intense heat of political and religious passions.) Even in the absence of bias or political agenda behind the various studies of children raised by same-sex couples, the most neutral and strict application of scientific principles to this field would be constrained by the limited period of observation that has been available. Gay and lesbian couples living together openly, and official recognition of them as their children’s sole parents, comprise a very recent phenomenon, and the recency of that phenomenon has not yet permitted any study of how those children fare as adults and at best minimal study of how they fare during their adolescent years. The Legislature can rationally view the state of the scientific evidence as unsettled on the critical question it now faces: Are families headed by same-sex parents equally successful in rearing children from infancy to adulthood as families headed by parents of opposite sexes? Our belief that children raised by same-sex couples should fare the same as children raised in traditional families is just that: a passionately held but utterly untested belief. The Legislature is not required to share that belief but may, as the creator of the institution of civil marriage, wish to see the proof before making a fundamental alteration to that institution.

Although ostensibly applying the rational basis test to the civil marriage statutes, it is abundantly apparent that the court is in fact applying some undefined stricter standard to assess the constitutionality of the marriage statutes’ exclusion of same-sex couples. While avoiding any express conclusion as to any of the proffered routes by which that exclusion would be subjected to a test of strict scrutiny — infringement of a fundamental right, discrimination based on gender, or discrimination against gays and lesbians as a suspect classification — the opinion repeatedly alludes to those concepts in a prolonged and eloquent prelude before articulating its view that the exclusion lacks *360even a rational basis. See, e.g., ante at 313 (noting that State Constitution is “more protective of individual liberty and equality,” demands “broader protection for fundamental rights,” and is “less tolerant of government intrusion into the protected spheres of private life” than Federal Constitution); ante at 322 (describing decision to marry and choice of marital partner as “among life’s momentous acts of self-definition”); ante at 326 (repeated references to “right to marry” as “fundamental”); ante at 327-328 (repeated comparisons to statutes prohibiting interracial marriage, which were predicated on suspect classification of race); ante at 328 (characterizing ban on same-sex marriage as “invidious” discrimination that “deprives individuals of access to an institution of fundamental legal, personal, and social significance” and again noting that Massachusetts Constitution “protects matters of personal liberty against government incursion” more zealously than Federal Constitution); ante at 329 (characterizing “whom to marry, how to express sexual intimacy, and whether and how to establish a family” as “among the most basic of every individual’s liberty and due process rights”); id. (“liberty interest in choosing whether and whom to marry would be hollow” if Commonwealth could “foreclose an individual from freely choosing the person” to marry); ante at 333 (opining that in “overlapping realms of personal autonomy, marriage, family life, and child rearing,” characterized as “fundamentally private areas of fife,” court uses “integrated” analysis instead of “narrow focus”). See also ante at 337 n.29 (suggesting that prohibition on same-sex marriage “impose[s] limits on personal beliefs”); ante at 339 n.31 (suggesting that “total deference” to Legislature in this case would be equivalent to “strip[ping]” judiciary “of its constitutional authority to decide challenges” in such areas as forced sterilization, antimiscegenation statutes, and abortion, even though all cited examples pertain to fundamental rights analyzed under strict scrutiny, not under rational basis test); ante at 339 (civil marriage as “a right of fundamental importance”); ante at 341 (noting State policy of “preventing discrimination on the basis of sexual orientation”); id. (prohibition against same-sex marriage inconsistent with “gender-neutral laws promoting stable families,” and “rooted in *361persistent prejudices against” homosexuals); ante at 342 (prohibition against same-sex marriage “violates the basic premises of individual liberty”). In short, while claiming to apply a mere rational basis test, the court’s opinion works up an enormous head of steam by repeated invocations of avenues by which to subject the statute to strict scrutiny, apparently hoping that that head of steam will generate momentum sufficient to propel the opinion across the yawning chasm of the very deferential rational basis test.

Shorn of these emotion-laden invocations, the opinion ultimately opines that the Legislature is acting irrationally when it grants benefits to a proven successful family structure while denying the same benefits to a recent, perhaps promising, but essentially untested alternate family structure. Placed in a more neutral context, the court would never find any irrationality in such an approach. For example, if the issue were government subsidies and tax benefits promoting use of an established technology for energy efficient heating, the court would find no equal protection or due process violation in the Legislature’s decision not to grant the same benefits to an inventor or manufacturer of some new, alternative technology who did not yet have sufficient data to prove that that new technology was just as good as the established technology. That the early results from preliminary testing of the new technology might look very promising, or that the theoretical underpinnings of the new technology might appear flawless, would not make it irrational for the Legislature to grant subsidies and tax breaks to the established technology and deny them to the still unproved newcomer in the field. While programs that affect families and children register higher on our emotional scale than programs affecting energy efficiency, our standards for what is or is not “rational” should not be bent by those emotional tugs. Where, as here, there is no ground for applying strict scrutiny, the emotionally compelling nature of the subject matter should not affect the manner in which we apply the rational basis test.

Or, to the extent that the court is going to invoke such emotion-laden and value-laden rhetoric as a means of heightening the degree of scrutiny to be applied, the same form of rhetoric can be employed to justify the Legislature’s proceeding with extreme caution in this area. In considering whether the *362Legislature has a rational reason for postponing a dramatic change to the definition of marriage, it is surely pertinent to the inquiry to recognize that this proffered change affects not just a load-bearing wall of our social structure but the very cornerstone of that structure. See post at 393 (Cordy, J., dissenting). Before making a fundamental alteration to that cornerstone, it is eminently rational for the Legislature to require a high degree of certainty as to the precise consequences of that alteration, to make sure that it can be done safely, without either temporary or lasting damage to the structural integrity of the entire edifice. The court today blithely assumes that there are no such dangers and that it is safe to proceed (see ante at 340), an assumption that is not supported by anything more than the court’s blind faith that it is so.

More importantly, it is not our confidence in the lack of adverse consequences that is at issue, or even whether that confidence is justifiable. The issue is whether it is rational to reserve judgment on whether this change can be made at this time without damaging the institution of marriage or adversely affecting the critical role it has played in our society. Absent consensus on the issue (which obviously does not exist), or unanimity amongst scientists studying the issue (which also does not exist), or a more prolonged period of observation of this new family structure (which has not yet been possible), it is rational for the Legislature to postpone any redefinition of marriage that would include same-sex couples until such time as it is certain that that redefinition will not have unintended and undesirable social consequences. Through the political process, the people may decide when the benefits of extending civil marriage to same-sex couples have been shown to outweigh whatever risks — be they palpable or ephemeral — are involved. However minimal the risks of that redefinition of marriage may seem to us from our vantage point, it is not up to us to decide what risks society must run, and it is inappropriate for us to arrogate that power to ourselves merely because we are confident that “it is the right thing to do.” Ante at 350 (Greaney, J., concurring).

As a matter of social history, today’s opinion may represent a great turning point that many will hail as a tremendous step *363toward a more just society. As a matter of constitutional jurisprudence, however, the case stands as an aberration. To reach the result it does, the court has tortured the rational basis test beyond recognition. I fully appreciate the strength of the temptation to find this particular law unconstitutional — there is much to be said for the argument that excluding gay and lesbian couples from the benefits of civil marriage is cruelly unfair and hopelessly outdated; the inability to marry has a profound impact on the personal lives of committed gay and lesbian couples (and their children) to whom we are personally close (our friends, neighbors, family members, classmates, and co-workers); and our resolution of this issue takes place under the intense glare of national and international publicity. Speaking metaphorically, these factors have combined to turn the case before us into a “perfect storm” of a constitutional question. In my view, however, such factors make it all the more imperative that we adhere precisely and scrupulously to the established guideposts of our constitutional jurisprudence, a jurisprudence that makes the rational basis test an extremely deferential one that focuses on the rationality, not the persuasiveness, of the potential justifications for the classifications in the legislative scheme. I trust that, once this particular “storm” clears, we will return to the rational basis test as it has always been understood and applied. Applying that deferential test in the manner it is customarily applied, the exclusion of gay and lesbian couples from the institution of civil marriage passes constitutional muster. I respectfully dissent.

1

The one difference that the court acknowledges — that sexual relations between persons of the same sex does not result in pregnancy and childbirth — it immediately brushes aside on the theory that civil marriage somehow has nothing to do with begetting children. Ante at 331-333. For the reasons explained in detail in Justice Cordy’s dissent, in which I join, the reasons justifying the civil marriage laws are inextricably linked to the fact that human sexual intercourse between a man and a woman frequently results in pregnancy and childbirth. Indeed, as Justice Cordy outlines, that fact lies at the core of why society fashioned the institution of marriage in the first place. Post at 381-382 (Cordy, J., dissenting).

Cordy, J.

(dissenting, with whom Spina and Sosman, JJ., join). The court’s opinion concludes that the Department of Public Health has failed to identify any “constitutionally adequate reason” for limiting civil marriage to opposite-sex unions, and that there is no “reasonable relationship” between a disqualification of same-sex couples who wish to enter into a civil marriage and the protection of public health, safety, or general welfare. Consequently, it holds that the marriage statute cannot withstand scrutiny under the Massachusetts Constitution. Because I find these conclusions to be unsupportable in light of the nature of the rights and regulations at issue, the presumption *364of constitutional validity and significant deference afforded to legislative enactments, and the “undesirability of the judiciary substituting its notions of correct policy for that of a popularly elected Legislature” responsible for making such policy, Zayre Corp. v. Attorney Gen., 372 Mass. 423, 433 (1977), I respectfully dissent. Although it may be desirable for many reasons to extend to same-sex couples the benefits and burdens of civil marriage (and the plaintiffs have made a powerfully reasoned case for that extension), that decision must be made by the Legislature, not the court.

If a statute either impairs the exercise of a fundamental right protected by the due process or liberty provisions of our State Constitution, or discriminates based on a constitutionally suspect classification such as sex, it will be subject to strict scrutiny when its validity is challenged. See Blixt v. Blixt, 437 Mass. 649, 655-656, 660-661 (2002), cert, denied, 537 U.S. 1189 (2003) (fundamental right); Lowell v. Kowalski, 380 Mass. 663, 666 (1980) (sex-based classification). If it does neither, a statute “will be upheld if it is ‘rationally related to a legitimate State purpose.’ ” Hallett v. Wrentham, 398 Mass. 550, 557 (1986), quoting Paro v. Longwood Hosp., 373 Mass. 645, 649 (1977). This test, referred to in State and Federal constitutional jurisprudence as the “rational basis test,”1 is virtually identical in substance and effect to the test applied to a law promulgated under the State’s broad police powers (pursuant to which the marriage statutes and most other licensing and regulatory laws are enacted): that is, the law is valid if it is reasonably related to the protection of public health, safety, or general welfare. See, e.g., Leigh v. Board of Registration in Nursing, 395 Mass. 670, 682-683 (1985) (applying rational basis review to question of State exercise of police power).

The Massachusetts marriage statute does not impair the exercise of a recognized fundamental right, or discriminate on the basis of sex in violation of the equal rights amendment to the Massachusetts Constitution. Consequently, it is subject to *365review only to determine whether it satisfies the rational basis test. Because a conceivable rational basis exists upon which the Legislature could conclude that the marriage statute furthers the legitimate State purpose of ensuring, promoting, and supporting an optimal social structure for the bearing and raising of children, it is a valid exercise of the State’s police power.

A. Limiting marriage to the union of one man and one woman does not impair the exercise of a fundamental right. Civil marriage is an institution created by the State. In Massachusetts, the marriage statutes are derived from English common law, see Commonwealth v. Knowlton, 2 Mass. 530, 534 (1807), and were first enacted in colonial times. Commonwealth v. Munson, 127 Mass. 459, 460 (1879). They were enacted to secure public interests and not for religious purposes or to promote personal interests or aspirations. (See discussion infra at 381-385 ). As the court notes in its opinion, the institution of marriage is “the legal union of a man and woman as husband and wife,” ante at 319, and it has always been so under Massachusetts law, colonial or otherwise.

The plaintiffs contend that because the right to choose to marry is a “fundamental” right, the right to marry the person of one’s choice, including a member of the same sex, must also be a “fundamental” right. While the court stops short of deciding that the right to marry someone of the same sex is “fundamental” such that strict scrutiny must be applied to any statute that impairs it, it nevertheless agrees with the plaintiffs that the right to choose to marry is of fundamental importance (“among the most basic” of every person’s “liberty and due process rights”) and would be “hollow” if an individual was foreclosed from “freely choosing the person with whom to share . . . the . . . institution of civil marriage.” Ante at 329. Hence, it concludes that a marriage license cannot be denied to an individual who wishes to marry someone of the same sex. In reaching this result the court has transmuted the “right” to marry into a right to change the institution of marriage itself. This feat of reasoning succeeds only if one accepts the proposition that the definition of the institution of marriage as a union between a man and a woman is merely “conclusory” (as suggested, ante at 348 [Greaney, J., concurring]), rather than the basis on which the *366“right” to partake in it has been deemed to be of fundamental importance. In other words, only by assuming that “marriage” includes the union of two persons of the same sex does the court conclude that restricting marriage to opposite-sex couples infringes on the “right” of same-sex couples to “marry.”2

The plaintiffs ground their contention that they have a fundamental right to marry a person of the same sex in a long line of Supreme Court decisions, e.g., Turner v. Safley, 482 U.S. 78 (1987); Zablocki v. Redhail, 434 U.S. 374 (1978); Loving v. Virginia, 388 U.S. 1 (1967); Griswold v. Connecticut, 381 U.S. 479 (1965); Skinner v. Oklahoma, 316 U.S. 535 (1942); that discuss the importance of marriage. In context, all of these decisions and their discussions are about the “fundamental” nature of the institution of marriage as it has existed and been understood in this country, not as the court has redefined it today. Even in that context, its “fundamental” nature is derivative of the nature of the interests that underlie or are associated with it.3 An examination of those interests reveals that they are either not shared by same-sex couples or not implicated by the marriage statutes.

Supreme Court cases that have described marriage or the right to marry as “fundamental” have focused primarily on the underlying interest of every individual in procreation, which, historically, could only legally occur within the construct of marriage because sexual intercourse outside of marriage was a *367criminal act.4 In Skinner v. Oklahoma, supra, the first case to characterize marriage as a “fundamental” right, the Supreme Court stated, as its rationale for striking down a sterilization statute, that “[m]arriage and procreation are fundamental to the very existence of the race.” Id. at 541. In concluding that a sterilized individual “is forever deprived of a basic liberty,” id., the Court was obviously referring to procreation rather than marriage, as this court recognized in Matter of Moe, 385 Mass. 555, 560 (1982). Similarly, in Loving v. Virginia, supra, in which the United States Supreme Court struck down Virginia’s antimiscegenation statute, the Court implicitly linked marriage with procreation in describing marriage as “fundamental to our very existence.” Id. at 12. In Zablocki v. Redhail, supra, the Court expressly linked the right to marry with the right to procreate, concluding that “if [the plaintiff’s] right to procreate means anything at all, it must imply some right to enter the only relationship in which the State . . . allows sexual relations legally to take place.” Id. at 386. Once again, in Turner v. Safley, supra, striking a State regulation that curtailed the right of an inmate to marry, the Court included among the important attributes of such marriages the “expectation that [the marriage] ultimately will be fully consummated.” Id. at 96. See Milford v. Worcester, 1 Mass. 48, 52 (1810) (purpose of marriage is “to regulate, chasten, and refine, the intercourse between the sexes; and to multiply [and] preserve . . . the species”). Because same-sex couples are unable to procreate on their own, any right to marriage they may possess cannot be based on their interest in procreation, which has been essential to the Supreme Court’s denomination of the right to marry as fundamental.

Supreme Court cases recognizing a right to privacy in intimate decision-making, e.g., Griswold v. Connecticut, supra (striking down statute prohibiting use of contraceptives); Roe v. Wade, 410 U.S. 113 (1973) (striking down statute criminalizing abortion), have also focused primarily on sexual relations and the decision whether or not to procreate, and have refused to recognize an “unlimited right” to privacy. Id. at 154. Massachusetts courts have been no more willing than the Federal *368courts to adopt a “universal!]” “privacy doctrine,” Marcoux v. Attorney Gen., 375 Mass. 63, 67 (1978), or to derive “controversial ‘new’ rights from the Constitution.” Aime v. Commonwealth, 414 Mass. 667, 674 n.10 (1993).

What the Griswold Court found “repulsive to the notions of privacy surrounding the marriage relationship” was the prospect of “allow[ing] the police to search the sacred precincts of marital bedrooms for telltale signs of the use of contraceptives.” Griswold v. Connecticut, supra at 485-486. See Moe v. Secretary of Admin. & Fin., 382 Mass. 629, 658 (1981), quoting L. Tribe, American Constitutional Law 924 (1978) (finding it “difficult to imagine a clearer case of bodily intrusion” than being forced to bear a child). When Justice Goldberg spoke of “marital relations” in the context of finding it “difficult to imagine what is more private or more intimate than a husband and wife’s marital relations [hip],” Griswold v. Connecticut, supra at 495 (Goldberg, J., concurring), he was obviously referring to sexual relations.5 Similarly, in Lawrence v. Texas, 123 S. Ct. 2472 (2003), it was the criminalization of private sexual behavior that the Court found violative of the petitioners’ liberty interest.

In Massachusetts jurisprudence, protected decisions generally have been limited to those concerning “whether or not to beget or bear a child,” Matter of Moe, 385 Mass. 555, 564 (1982) (see Opinion of the Justices, 423 Mass. 1201, 1234-1235 [1996] [“focus of (the Griswold and Roe cases) and the cases following them has been the intrusion . . . into the especially intimate aspects of a person’s life implicated in procreation and childbearing”]); how to raise a child, see Care & Protection of Robert, 408 Mass. 52, 58, 60 (1990); or whether or not to accept medical treatment, see Brophy v. New England Sinai Hosp., Inc., 398 Mass. 417, 430 (1986); Superintendent of Belchertown State Sch. v. Saikewicz, 373 Mass. 728, 742 (1977), none of which is at issue here. See also Commonwealth v. Balthazar, 366 Mass. 298, 301 (1974) (statute punishing unnatural and *369lascivious acts does not apply to sexual conduct engaged in by adults in private, in light of “articulation of the constitutional right of an individual to be free from government regulation of certain sex related activities”).

The marriage statute, which regulates only the act of obtaining a marriage license, does not implicate privacy in the sense that it has found constitutional protection under Massachusetts and Federal law. Cf. Commonwealth v. King, 374 Mass. 5, 14 (1977) (solicitation of prostitution “while in a place to which the public had access” implicated no “constitutionally protected rights of privacy”); Marcoux v. Attorney Gen., supra at 68 (right to privacy, at most, protects conduct “limited more or less to the hearth”). It does not intrude on any right that the plaintiffs have to privacy in their choices regarding procreation, an intimate partner or sexual relations.6 The plaintiffs’ right to privacy in such matters does not require that the State officially endorse their choices in order for the right to be constitutionally vindicated.

Although some of the privacy cases also speak in terms of personal autonomy, no court has ever recognized such an open-ended right. “That many of the rights and liberties protected by the Due Process Clause sound in personal autonomy does not warrant the sweeping conclusion that any and all important, intimate, and personal decisions are so protected . . . .” Washington v. Glucksberg, 521 U.S. 702, 727 (1997). Such decisions are protected not because they are important, intimate, and personal, but because the right or liberty at stake is “so deeply rooted in our history and traditions, or so fundamental to our concept of constitutionally ordered liberty” that it is protected by due process. Id. Accordingly, the Supreme Court has concluded that while the decision to refuse unwanted medical treatment is fundamental, Cruzan v. Director, Mo. Dep’t of Health, 497 U.S. 261, 278 (1990), because it is deeply rooted in our nation’s history and tradition, the equally personal and profound decision to commit suicide is not because of the absence of such roots. Washington v. Glucksberg, supra.

*370While the institution of marriage is deeply rooted in the history and traditions of our country and our State, the right to marry someone of the same sex is not. No matter how personal or intimate a decision to marry someone of the same sex might be, the right to make it is not guaranteed by the right of personal autonomy.

The protected right to freedom of association, in the sense of freedom of choice “to enter into and maintain certain intimate human relationships,” Roberts v. United States Jaycees, 468 U.S. 609, 617 (1984) (as an element of liberty or due process rather than free speech), is similarly limited and unimpaired by the marriage statute. As recognized by the Supreme Court, that right affords protection only to “certain kinds of highly personal relationships,” id. at 618, such as those between husband and wife, parent and child, and among close relatives, id. at 619, that “have played a critical role in the culture and traditions of the Nation,” id. at 618-619, and are “deeply rooted in this Nation’s history and tradition.” Moore v. East Cleveland, 431 U.S. 494, 498-499, 503 (1977) (distinguishing on this basis between family and nonfamily relationships). Unlike opposite-sex marriages, which have deep historic roots, or the parent-child relationship, which reflects a “strong tradition” founded on “the history and culture of Western civilization” and “is now established beyond debate as an enduring American tradition,” Wisconsin v. Yoder, 406 U.S. 205, 232 (1972); or extended family relationships, which have been “honored throughout our history,” Moore v. East Cleveland, supra at 505, same-sex relationships, although becoming more accepted, are certainly not so “deeply rooted in this Nation’s history and tradition” as to warrant such enhanced constitutional protection.

Although “expressions of emotional support and public commitment” have been recognized as among the attributes of marriage, which, “[tjaken together . . . form a constitutionally protected marital relationship” (emphasis added), Turner v. Safley, 482 U.S. 78, 95, 96 (1987), those interests, standing alone, are not the source of a fundamental right to marry. While damage to one’s “status in the community” may be sufficient harm to confer standing to sue, Lowell v. Kowalski, 380 Mass. 663, 667 (1980), such status has never been recognized as a *371fundamental right. See Paul v. Davis, 424 U.S. 693, 701 (1976) (mere damage to reputation does not constitute deprivation of “liberty”).

Finally, the constitutionally protected interest in child rearing, recognized in Meyer v. Nebraska, 262 U.S. 390, 399 (1923); Pierce v. Society of Sisters, 268 U.S. 510, 534-535 (1925); and Care & Protection of Robert, supra at 58, 60, is not implicated or infringed by the marriage statute here. The fact that the plaintiffs cannot marry has no bearing on their independently protected constitutional rights as parents which, as with opposite-sex parents, are limited only by their continued fitness and the best interests of their children. Bezio v. Patenaude, 381 Mass. 563, 579 (1980) (courts may not use parent’s sexual orientation as reason to deny child custody).

Because the rights and interests discussed above do not afford the plaintiffs any fundamental right that would be impaired by a statute Umiting marriage to members of the opposite sex, they have no fundamental right to be declared “married” by the State.

Insofar as the right to marry someone of the same sex is neither found in the unique historical context of our Constitution7 nor compelled by the meaning ascribed by this court to the liberty and due process protections contained within it, should the court nevertheless recognize it as a fundamental right? The consequences of deeming a right to be “fundamental” are profound, and this court, as well as the Supreme Court, has been very cautious in recognizing them.8 Such caution is required by separation of powers principles. If a right is found *372to be “fundamental,” it is, to a great extent, removed from “the arena of public debate and legislative action”; utmost care must be taken when breaking new ground in this field “lest the liberty protected by the Due Process Clause be subtly transformed into the policy preferences of [judges].” Washington v. Glucksberg, 521 U.S. 702, 720 (1997).

“[T]o rein in” the otherwise potentially unlimited scope of substantive due process rights, id. at 722, both Federal and Massachusetts courts have recognized as “fundamental” only those “rights and liberties which are, objectively, ‘deeply rooted in this Nation’s history and tradition,’ [Moore v. East Cleveland, supra at 503] . . . and ‘implicit in the concept of ordered liberty.’ ” Id. at 720-721, quoting Palko v. Connecticut, 302 U.S. 319, 325 (1937). See Dutil, petitioner, 437 Mass. 9, 13 (2002) (same). In the area of family-related rights in particular, the Supreme Court has emphasized that the “Constitution protects the sanctity of the family precisely because the institution of the family is deeply rooted.” Moore v. East Cleveland, supra.9

Applying this limiting principle, the Supreme Court, as noted above, declined to recognize a fundamental right to physician-assisted suicide, which would have required “reversing] centuries of legal doctrine and practice, and striking] down the considered policy choice of almost every State.” Washington v. *373Glucksberg, supra at 723. While recognizing that public attitudes toward assisted suicide are currently the subject of “earnest and profound debate,” the Court nevertheless left the continuation and resolution of that debate to the political arena, “as it should be in a democratic society.” Id. at 719, 735.

Similarly, Massachusetts courts have declined to recognize rights that are not so deeply rooted.10 As this court noted in considering whether to recognize a right of terminally ill patients to refuse life-prolonging treatment, “the law always lags behind the most advanced thinking in every area,” and must await “some common ground, some consensus.” Superintendent of Belchertown State Sch. v. Saikewicz, 373 Mass. 728, 737 (1977), quoting Burger, The Law and Medical Advances, 67 Annals Internal Med. Supp. 7, 15, 17 (1967). See Blixt v. Blixt, 437 Mass. 649, 662-663 n.22 (2002) (“social consensus about family relationships is relevant to the constitutional limits on State intervention”).

This is not to say that a statute that has no rational basis must nevertheless be upheld so long as it is of ancient origin. However, “[t]he long history of a certain practice . . . and its *374acceptance as an uncontroversial part of our national and State tradition do suggest that [the court] should reflect carefully before striking it down.” Colo v. Treasurer & Receiver Gen., 378 Mass. 550, 557 (1979). As this court has recognized, the “fact that a challenged practice ‘is followed by a large number of states ... is plainly worth considering in determining whether the practice “offends some principle of justice so rooted in the traditions and conscience of our people as to be ranked as fundamental.” ’ ” Commonwealth v. Kostka, 370 Mass. 516, 533 (1976), quoting Leland v. Oregon, 343 U.S. 790, 798 (1952).

Although public attitudes toward marriage in general and same-sex marriage in particular have changed and are still evolving, “the asserted contemporary concept of marriage and societal interests for which [plaintiffs] contend” are “manifestly [less] deeply founded” than the “historic institution” of marriage. Matter of the Estate of Cooper, 187 A.D.2d 128, 133-134 (N.Y. 1993). Indeed, it is not readily apparent to what extent contemporary values have embraced the concept of same-sex marriage. Perhaps the “clearest and most reliable objective evidence of contemporary values is the legislation enacted by the country’s legislatures,” Atkins v. Virginia, 536 U.S. 304, 312 (2002), quoting Penry v. Lynaugh, 492 U.S. 302, 331 (1989). No State Legislature has enacted laws permitting same-sex marriages; and a large majority of States, as well as the United States Congress, have affirmatively prohibited the recognition of such marriages for any purpose. See P. Green-berg, State Laws Affecting Lesbians and Gays, National Conference of State Legislatures Legisbriefs at 1 (April/May 2001) (reporting that, as of May, 2001, thirty-six States had enacted “defense of marriage” statutes); 1 U.S.C. § 7 (2000); 28 U.S.C. § 1738C (2000) (Federal Defense of Marriage Act).

Given this history and the current state of public opinion, as reflected in the actions of the people’s elected representatives, it cannot be said that “a right to same-sex marriage is so rooted in the traditions and collective conscience of our people that failure to recognize it would violate the fundamental principles of liberty and justice that lie at the base of all our civil and political institutions. Neither . . . [is] a right to same-sex marriage *375. . . implicit in the concept of ordered liberty, such that neither liberty nor justice would exist if it were sacrificed.” Baehr v. Lewin, 74 Haw. 530, 556-557 (1993). See Dean v. District of Columbia, 653 A.2d 307, 333 (D.C. 1995) (per curiam) (Ferren, J., concurring in part and dissenting in part); Baker v. Nelson, 291 Minn. 310, 312 (1971), appeal dismissed, 409 U.S. 810 (1972); Storrs v. Holcomb, 168 Misc. 2d 898, 899-900 (N.Y. Sup. Ct. 1996), dismissed, 245 A.D.2d 943 (N.Y. 1997).11 In such circumstances, the law with respect to same-sex marriages must be left to develop through legislative processes, subject to the constraints of rationality, lest the court be viewed as using the liberty and due process clauses as vehicles merely to enforce its own views regarding better social policies, a role that the strongly worded separation of powers principles in art. 30 of the Declaration of Rights of our Constitution forbids, and for which the court is particularly ill suited.

B. The marriage statute, in limiting marriage to heterosexual couples, does not constitute discrimination on the basis of sex in violation of the Equal Rights Amendment to the Massachusetts Constitution. In his concurrence, Justice Greaney contends that the marriage statute constitutes discrimination on the basis of sex in violation of art. 1 of the Declaration of Rights as amended by art. 106 of the Amendments to the Constitution of the Commonwealth, the Equal Rights Amendment (ERA).12 Such a conclusion is analytically unsound and inconsistent with the legislative history of the ERA.

The central purpose of the ERA was to eradicate discrimination against women and in favor of men or vice versa. See Attorney Gen. v. Massachusetts Interscholastic Athletic Ass’n, 378 *376Mass. 342, 357 (1979). Consistent with this purpose, we have construed the ERA to prohibit laws that advantage one sex at the expense of the other, but not laws that treat men and women equally, id. at 346-349 (assuming that “separate but equal” treatment of males and females would be constitutionally permissible). The Massachusetts marriage statute does not subject men to different treatment from women; each is equally prohibited from precisely the same conduct. See Baker v. State, 170 Vt. 194, 215 n.13 (1999) (“there is no discrete class subject to differential treatment solely on the basis of sex”). Compare Commonwealth v. King, 374 Mass. 5, 16 (1977) (law prohibiting prostitution applied to both male and female prostitutes and therefore did not discriminate), and Personnel Adm’r of Mass. v. Feeney, 442 U.S. 256, 274-275 (1979) (declining to characterize veterans’ preference as sex discrimination because it applied to both male and female veterans), with Attorney Gen. v. Massachusetts Interscholastic Athletic Ass’n, supra, and Lowell v. Kowalski, 380 Mass. 663 (1980) (where statutes and rules at issue advantaged one sex over another).

Of course, a statute that on its face treats protected groups equally may still harm, stigmatize, or advantage one over the other. Such was the circumstance in Loving v. Virginia, 388 U.S. 1 (1967), where the Supreme Court struck down a State statute that made interracial marriage a crime, as constituting invidious discrimination on the basis of race. While the statute purported to apply equally to whites and nonwhites, the Court found that it was intended and structured to favor one race (white) and disfavor all others (nonwhites). The statute’s legislative history demonstrated that its purpose was not merely to punish interracial marriage, but to do so for the sole benefit of the white race. As the Supreme Court readily concluded, the Virginia law was “designed to maintain White Supremacy.” Id. at 11. Consequently, there was a fit between the class that the law was intended to discriminate against (nonwhite races) and the classification enjoying heightened protection (race).

By contrast, here there is no evidence that limiting marriage to opposite-sex couples was motivated by sexism in general or a desire to disadvantage men or women in particular. Moreover, no one has identified any harm, burden, disadvantage, or *377advantage accruing to either gender as a consequence of the Massachusetts marriage statute. In the absence of such effect, the statute limiting marriage to couples of the opposite sex does not violate the ERA’S prohibition of sex discrimination.13

This conclusion is buttressed by the legislative history of the ERA, which was adopted by the voters on November 2, 1976, after being approved by constitutional conventions of the Legislature on August 15, 1973, (by a vote of 261-0) and May 14, 1975 (by a vote of 217-55).

In anticipation of its adoption, the Legislature enacted and, on June 21, 1975, the Governor approved a “Resolve providing for an investigation and study by a special commission relative to the effect of the ratification of the proposed amendments to the Constitution of the Commonwealth of Massachusetts and the Constitution of the United States prohibiting discrimination on account of sex upon the laws, business communities and public in the Commonwealth.” Res. 1975, c. 26. One of the principal tasks of the commission was to catalog the aspects of the General Laws that would have to be amended for the statutory code to comply with the mandate of the proposed amendment that equality not be abridged on the basis of sex.14

On October 19, 1976, just before the general election at which the amendment was to be considered, the commission filed its Interim Report, which focused on the effect of the Massachusetts ERA on the laws of the Commonwealth. 1976 Senate Doc. No. 1689. A section of the report, entitled “Areas Unaffected by the *378Equal Rights Amendment,” addressed some of the legal regimes that would not be affected by the adoption of the ERA. One such area was “Homosexual Marriage,” about which the commission stated:

“An equal rights amendment will have no effect upon the allowance or denial of homosexual marriages. The equal rights amendment is not concerned with the relationship of two persons of the same sex; it only addresses those laws or public-related actions which treat persons of opposite sexes differently. The Washington Court of Appeals has already stated that the equal rights amendment to its state constitution did not afford a basis for validating homosexual marriages. In Colorado, the attorney general has likewise issued an opinion that the state equal rights amendment did not validate homosexual marriage. There are no cases which have used a state equal rights amendment to either validate or require the allowance of homosexual marriages.” (Footnotes omitted.)

Id. at 21-22.15

The views of the commission were reflected in the public debate surrounding the passage of the ERA that focused on gender equality. See, e.g., Referenda reviewed, Boston Globe, Nov. 1, 1976, at 26; Voters’ guide on nine state referendum measures, Boston Herald American, Nov. 1, 1976, at 17. Claims that the ERA might be the basis for validating marriages between same-sex couples were labelled as “exaggerated” and “unfounded.” For example, before the vote, the Boston Globe published an editorial discussing and urging favorable action on the ERA. In making its case, it noted that “[t]hose urging a no vote . . . argue that the amendment would . . . legitimize marriage between people of the same sex [and other changes]. In reality, the proposed amendment would require none of these things. Mass, ballot issues ... 1 Equal Rights Amendment. Boston Globe, Nov. 1, 1976, at 29. And in the aftermath of the vote, the Boston Globe heralded the electorate’s acceptance of “the arguments of proponents that the proposal would not result *379in many far-reaching or threatening changes.” Referendums fared poorly, Boston Globe, Nov. 4, 1976, at 29.

While the court, in interpreting a constitutional amendment, is not bound to accept either the views of a legislative commission studying and reporting on the amendment’s likely effects, or of public commentary and debate contemporaneous with its passage, it ought to be wary of completely disregarding what appears to be the clear intent of the people recently recorded in our constitutional history. This is particularly so where the plain wording of the amendment does not require the result it would reach.

C. The marriage statute satisfies the rational basis standard. The burden of demonstrating that a statute does not satisfy the rational basis standard rests on the plaintiffs. It is a weighty one. “[A] reviewing court will presume a statute’s validity, and make all rational inferences in favor of it. . . . The Legislature is not required to justify its classifications, nor provide a record or finding in support of them.” (Citation omitted.) Paro v. Longwood Hosp., 373 Mass. 645, 650 (1977). The statute “only need[s to] be supported by a conceivable rational basis.” Fine v. Contributory Retirement Appeal Bd., 401 Mass. 639, 641 (1988). See Massachusetts Fed’n of Teachers v. Board of Educ., 436 Mass. 763, 771-772 (2002). As this court stated in Shell Oil Co. v. Revere, 383 Mass. 682, 687-688 (1981):

“[I]t is not the court’s function to launch an inquiry to resolve a debate which has already been settled in the legislative forum. ‘[I]t [is] the judge’s duty ... to give effect to the will of the people as expressed in the statute by their representative body. It is in this way . . . that the doctrine of separation of powers is given meaning.’ Commonwealth v. Leis, 355 Mass. 189, 202 (1969) (Kirk, L, concurring).

“This respect for the legislative process means that it is not the province of the court to sit and weigh conflicting evidence supporting or opposing a legislative enactment. . . .

“Although persons challenging the constitutionality of legislation may introduce evidence in support of their *380claim that the legislation is irrational . . . they will not prevail if ‘the question is at least debatable’ in view of the evidence which may have been available to the Legislature. United States v. Carolene Prods. Co., 304 U.S. 144, 154 (1938).”

The “time tested wisdom of the separation of powers” requires courts to avoid “judicial legislation in the guise of new constructions to meet real or supposed new popular viewpoints, preserving always to the Legislature alone its proper prerogative of adjusting the statutes to changed conditions.” Pielech v. Massasoit Greyhound, Inc., 423 Mass. 534, 539, 540 (1996), cert, denied, 520 U.S. 1131 (1997), quoting Commonwealth v. A Juvenile, 368 Mass. 580, 595 (1975).

In analyzing whether a statute satisfies the rational basis standard, we look to the nature of the classification embodied in the enactment, then to whether the statute serves a legitimate State purpose, and finally to whether the classification is reasonably related to the furtherance of that purpose. With this framework, we turn to the challenged statute, G. L. c. 207, which authorizes local town officials to issue licenses to couples of the opposite sex authorizing them to enter the institution of civil marriage.

1. Classification. The nature of the classification at issue is readily apparent. Opposite-sex couples can obtain a license and same-sex couples cannot. The granting of this license, and the completion of the required solemnization of the marriage, opens the door to many statutory benefits and imposes numerous responsibilities. The fact that the statute does not permit such licenses to be issued to couples of the same sex thus bars them from civil marriage. The classification is not drawn between men and women or between heterosexuals and homosexuals, any of whom can obtain a license to marry a member of the opposite sex; rather, it is drawn between same-sex couples and opposite-sex couples.

2. State purpose. The court’s opinion concedes that the civil marriage statute serves legitimate State purposes, but further investigation and elaboration of those purposes is both helpful and necessary.

*381Civil marriage is the institutional mechanism by which societies have sanctioned and recognized particular family structures, and the institution of marriage has existed as one of the fundamental organizing principles of human society. See C.N. Degler, The Emergence of the Modem American Family, in The American Family in Social-Historical Perspective 61 (3d ed. 1983); A.J. Hawkins, Introduction, in Revitalizing the Institution of Marriage for the Twenty-First Century: An Agenda for Strengthening Marriage xiv (2002); C. Lasch, Social Pathologists and the Socialization of Reproduction, in The American Family in Social-Historical Perspective, supra at 80; W.J. O’Donnell & D.A. Jones, Marriage and Marital Alternatives 1 (1982); L. Saxton, The Individual, Marriage, and the Family 229-230, 260 (1968); M.A. Schwartz & B.M. Scott, Marriages and Families: Diversity and Change 4 (1994); Wardle, “Multiply and Replenish”: Considering Same-Sex Marriage in Light of State Interests in Marital Procreation, 24 Harv. J.L. & Pub. Pol’y 771, 777-780 (2001); J.Q. Wilson, The Marriage Problem: How Our Culture Has Weakened Families 28, 40, 66-67 (2002). Marriage has not been merely a contractual arrangement for legally defining the private relationship between two individuals (although that is certainly part of any marriage). Rather, on an institutional level, marriage is the “very basis of the whole fabric of civilized society,” J.P. Bishop, Commentaries on the Law of Marriage and Divorce, and Evidence in Matrimonial Suits § 32 (1852), and it serves many important political, economic, social, educational, procreational, and personal functions.

Paramount among its many important functions, the institution of marriage has systematically provided for the regulation of heterosexual behavior, brought order to the resulting procreation, and ensured a stable family structure in which children will be reared, educated, and socialized. See Milford v. Worcester, 7 Mass. 48, 52 (1810) (civil marriage “intended to regulate, chasten, and refine, the intercourse between the sexes; and to multiply, preserve, and improve the species”). See also P. Blumstein & P. Schwartz, American Couples: Money, Work, Sex 29 (1983); C.N. Degler, supra at 61; G. Douglas, Marriage, Cohabitation, and Parenthood From Contract to Status?, in Cross Currents: Family Law and Policy in the United States and *382England 223 (2000); S.L. Nock, The Social Costs of Deinstitutionalizing Marriage, in Revitalizing the Institution of Marriage for the Twenty-First Century: An Agenda for Strengthening Marriage, supra at 7; L. Saxton, supra at 239-240, 242; M.A. Schwartz & B.M. Scott, supra at 4-6; Wardle, supra at 781-796; J.Q. Wilson, supra at 23-32. Admittedly, heterosexual intercourse, procreation, and child care are not necessarily conjoined (particularly in the modern age of widespread effective contraception and supportive social welfare programs), but an orderly society requires some mechanism for coping with the fact that sexual intercourse commonly results in pregnancy and childbirth. The institution of marriage is that mechanism.

The institution of marriage provides the important legal and normative link between heterosexual intercourse and procreation on the one hand and family responsibilities on the other. The partners in a marriage are expected to engage in exclusive sexual relations, with children the probable result and paternity presumed. See G. L. c. 209C, § 6 (“a man is presumed to be the father of a child . . . if he is or has been married to the mother and the child was bom during the marriage, or within three hundred days after the marriage was terminated by death, annulment or divorce”). Whereas the relationship between mother and child is demonstratively and predictably created and recognizable through the biological process of pregnancy and childbirth, there is no corresponding process for creating a relationship between father and child.16 Similarly, aside from an act of heterosexual intercourse nine months prior to childbirth, there is no process for creating a relationship between a man and a woman as the parents of a particular child. The institution of marriage fills this void by formally binding the husband-father to his wife and child, and imposing on him the responsibilities of fatherhood. See J.Q. Wilson, supra at 23-32. See also R Blumstein & P. Schwartz, supra at 29; C.N. Degler, supra at 61; G. Douglas, supra at 223; S.L. Nock, supra at 7; L. Saxton, supra at 239-240, 242; M.A. Schwartz & B.M. Scott, supra at 4-6; Wardle, supra at 781-796. The alternative, a *383society without the institution of marriage, in which heterosexual intercourse, procreation, and child care are largely disconnected processes, would be chaotic.

The marital family is also the foremost setting for the education and socialization of children. Children learn about the world and their place in it primarily from those who raise them, and those children eventually grow up to exert some influence, great or small, positive or negative, on society. The institution of marriage encourages parents to remain committed to each other and to their children as they grow, thereby encouraging a stable venue for the education and socialization of children. See P. Blumstein & P. Schwartz, supra at 26; C.N. Degler, supra at 61; S.L. Nock, supra at 2-3; C. Lasch, supra at 81; M.A. Schwartz & B.M. Scott, supra at 6-7. More macroscopically, construction of a family through marriage also formalizes the bonds between people in an ordered and institutional manner, thereby facilitating a foundation of interconnectedness and interdependency on which more intricate stabilizing social structures might be built. See M. Grossberg, Governing the Hearth: Law and Family in Nineteenth-Century America 10 (1985); C. Lasch, supra; L. Saxton, supra at 260; J.Q. Wilson, supra at 221.

This court, among others, has consistently acknowledged both the institutional importance of marriage as an organizing principle of society, and the State’s interest in regulating it. See French v. McAnarney, 290 Mass. 544, 546 (1935) (“Marriage is not merely a contract between the parties. It is the foundation of the family. It is a social institution of the highest importance. The Commonwealth has a deep interest that its integrity is not jeopardized”); Milford v. Worcester, 7 Mass. 48, 52 (1810) (“Marriage, being essential to the peace and harmony, and to the virtues and improvements of civil society, it has been, in all well-regulated governments, among the first attentions of the civil magistrate to regulate [it]”). See also Skinner v. Oklahoma, 316 U.S. 535, 541 (1942) (“Marriage and procreation are fundamental to the very existence and survival of the [human] race”); Maynard v. Hill, 125 U.S. 190, 211 (1888) (marriage “is an institution, in the maintenance of which in its purity the public is deeply interested, for it is the foundation of the family *384and of society, without which there would be neither civilization nor progress”); Murphy v. Ramsey, 114 U.S. 15, 45 (1885) (“no legislation can be supposed more wholesome and necessary in the founding of a free, self-governing commonwealth . . . than that which seeks to establish it on the basis of the idea of the family, as consisting in and springing from the union for life of one man and one woman ... the sure foundation of all that is stable and noble in our civilization; the best guaranty of that reverent morality which is the source of all beneficent progress in social and political improvement”); Reynolds v. United States, 98 U.S. 145, 165 (1878) (“Upon [marriage] society may be said to be built, and out of its fruits spring social relations and social obligations and duties, with which government is necessarily required to deal”).

It is undeniably true that dramatic historical shifts in our cultural, political, and economic landscape have altered some of our traditional notions about marriage, including the interpersonal dynamics within it,17 the range of responsibilities required of it as an institution,18 and the legal environment in which it exists.19 Nevertheless, the institution of marriage remains the principal weave of our social fabric. See C.N. Degler, supra at 61; A.J. Hawkins, Introduction, in Revitalizing the Institution of Marriage for the Twenty-First Century: An Agenda for Strengthening Marriage xiv (2002); C. Lasch, supra at 80; WJ. O’Donnell & D.A. Jones, Marriage and Marital Alternatives 1 (1982); L. Saxton, supra at 229-230, 260; M.A. Schwartz & B.M. Scott, supra at 4; Wardle, supra at 777-780; J.Q. Wilson, supra at 28, 40, 66-67. A family defined by heterosexual marriage continues to be the most prevalent social structure into which the vast majority of children are bom, nurtured, and prepared for productive participation in civil society, see Children’s Living Arrangements and Characteristics: March, *3852002, United States Census Bureau Current Population Reports at 3 (June, 2003) (in 2002, 69% of children lived with two married parents, 23% lived with their mother, 5% lived with their father, and 4% lived in households with neither parent present).

It is difficult to imagine a State purpose more important and legitimate than ensuring, promoting, and supporting an optimal social structure within which to bear and raise children. At the very least, the marriage statute continues to serve this important State purpose.20

3. Rational relationship. The question we must turn to next is whether the statute, construed as limiting marriage to couples of the opposite sex, remains a rational way to further that purpose. Stated differently, we ask whether a conceivable rational basis exists on which the Legislature could conclude that continuing to limit the institution of civil marriage to members of the opposite sex furthers the legitimate purpose of ensuring, promoting, and supporting an optimal social structure for the bearing and raising of children.21

In considering whether such a rational basis exists, we defer to the decision-making process of the Legislature, and must make deferential assumptions about the information that it might *386consider and on which it may rely. See Shell Oil Co. v. Revere, 383 Mass. 682, 688 (1981) (court considers “evidence which may have been available to the Legislature” [emphasis added]); Slome v. Chief of Police of Fitchburg, 304 Mass. 187, 189 (1939) (“any rational basis of fact that can be reasonably conceived” may support legislative finding); Mutual Loan Co. v. Martell, 200 Mass. 482, 487 (1909), aff’d, 222 U.S. 225 (1911) (“Legislature may be supposed to have known” relevant facts).

We must assume that the Legislature (1) might conclude that the institution of civil marriage has successfully and continually provided this structure over several centuries22; (2) might consider and credit studies that document negative consequences that too often follow children either bom outside of marriage or raised in households lacking either a father or a mother figure,23 and scholarly commentary contending that children and families *387develop best when mothers and fathers are partners in their parenting24; and (3) would be familiar with many recent studies that variously support the proposition that children raised in intact families headed by same-sex couples fare as well on many measures as children raised in similar families headed by opposite-sex couples25; support the proposition that children of same-sex couples fare worse on some measures26; or reveal notable differences between the two groups of children that warrant further study.27

We must also assume that the Legislature would be aware of the critiques of the methodologies used in virtually all of the comparative studies of children raised in these different environments, cautioning that the sampling populations are not representative, that the observation periods are too limited in time,28 that the empirical data are unreliable, and that the *388hypotheses are too infused with political or agenda driven bias. See, e.g., R. Lemer & A.K. Nagai, No Basis: What the Studies Don’t Tell Us About Same-Sex Parenting, Marriage Law Project (Jan. 2001) (criticizing forty-nine studies on same-sex parenting — at least twenty-six of which were cited by amici in this case — as suffering from flaws in formulation of hypotheses, use of experimental controls, use of measurements, sampling and statistical testing, and finding false negatives); Stacey, (How) Does the Sexual Orientation of Parents Matter, 66 Am. Soc. Rev. 159, 159-166 (2001) (highlighting problems with sampling pools, lack of longitudinal studies, and political hypotheses).

Taking all of this available information into account, the Legislature could rationally conclude that a family environment with married opposite-sex parents remains the optimal social structure in which to bear children, and that the raising of children by same-sex couples, who by definition cannot be the two sole biological parents of a child and cannot provide children with a parental authority figure of each gender,29 presents an alternative structure for child rearing that has not yet proved itself beyond reasonable scientific dispute to be as optimal as the biologically based marriage norm. See Baker v. State, 170 Vt. 194, 222 (1999) (“conceivable that the Legislature could conclude that opposite-sex partners offer advantages in th[e] area [of child rearing], although . . . experts disagree and the answer is decidedly uncertain”). Cf. Marcoux v. Attorney Gen., 375 Mass. 63, 65 (1978). Working from the assumption that a recognition of same-sex marriages will increase the number of children experiencing this alternative, the Legislature *389could conceivably conclude that declining to recognize same-sex marriages remains pmdent until empirical questions about its impact on the upbringing of children are resolved.30

The fact that the Commonwealth currently allows same-sex couples to adopt, see Adoption of Tammy, 416 Mass. 205 (1993), does not affect the rationality of this conclusion. The eligibility of a child for adoption presupposes that at least one of the child’s biological parents is unable or unwilling, for some reason, to participate in raising the child. In that sense, society has “lost” the optimal setting in which to raise that child — it is simply not available. In these circumstances, the principal and overriding consideration is the “best interests of the child,” considering his or her unique circumstances and the options that are available for that child. The objective is an individualized determination of the best environment for a particular child, where the normative social structure — a home with both the child’s biological father and mother — is not an option. That such a focused determination may lead to the approval of a same-sex couple’s adoption of a child does not mean that it would be irrational for a legislator, in fashioning statutory laws that cannot make such individualized determinations, to conclude generally that being raised by a same-sex couple has not yet been shown to be the absolute equivalent of being raised by one’s married biological parents.

That the State does not preclude different types of families from raising children does not mean that it must view them all as equally optimal and equally deserving of State endorsement and support.31 For example, single persons are allowed to adopt children, but the fact that the Legislature permits single-parent *390adoption does not mean that it has endorsed single parenthood as an optimal setting in which to raise children or views it as the equivalent of being raised by both of one’s biological parents.32 The same holds true with respect to same-sex couples — the fact that they may adopt children means only that the Legislature has concluded that they may provide an acceptable setting in which to raise children who cannot be raised by both of their biological parents. The Legislature may rationally permit adoption by same-sex couples yet harbor reservations as to whether parenthood by same-sex couples should be affirmatively encouraged to the same extent as parenthood by the heterosexual couple whose union produced the child.33

In addition, the Legislature could conclude that redefining the *391institution of marriage to permit same-sex couples to marry would impair the State’s interest in promoting and supporting heterosexual marriage as the social institution that it has determined best normalizes, stabilizes, and links the acts of procreation and child rearing. While the plaintiffs argue that they only want to take part in the same stabilizing institution, the Legislature conceivably could conclude that permitting their participation would have the unintended effect of undermining to some degree marriage’s ability to serve its social purpose. See Commonwealth v. Stowell, 389 Mass. 171, 175 (1983) (given State’s broad concern with institution of marriage, it has “legitimate interest in prohibiting conduct which may threaten that institution”).

So long as marriage is limited to opposite-sex couples who can at least theoretically procreate, society is able to communicate a consistent message to its citizens that marriage is a (normatively) necessary part of their procreative endeavor; that if they are to procreate, then society has endorsed the institution of marriage as the environment for it and for the subsequent rearing of their children; and that benefits are available explicitly to create a supportive and conducive atmosphere for those purposes. If society proceeds similarly to recognize marriages between same-sex couples who cannot procreate, it could be perceived as an abandonment of this claim, and might result in the mistaken view that civil marriage has little to do with procreation: just as the potential of procreation would not be necessary for a marriage to be valid, marriage would not be necessary for optimal procreation and child rearing to occur.34 In essence, the Legislature could conclude that the consequence *392of such a policy shift would be a diminution in society’s ability to steer the acts of procreation and child rearing into their most optimal setting.35 Hall-Omar Baking Co. v. Commissioner of Labor & Indus., 344 Mass. 695, 700 (1962) (“Legislative classification is valid if it is rational and bears some relationship to the object intended to be accomplished” [emphasis added]).

The court recognizes this concern, but brushes it aside with the assumption that permitting same-sex couples to marry “will not diminish the validity or dignity of opposite-sex marriage,” ante at 337, and that “we have no doubt that marriage will continue to be a vibrant and revered institution.” Ante at 340. Whether the court is correct in its assumption is irrelevant. What is relevant is that such predicting is not the business of the courts. A rational Legislature, given the evidence, could conceivably come to a different conclusion, or could at least *393harbor rational concerns about possible unintended consequences of a dramatic redefinition of marriage.36

There is no question that many same-sex couples are capable of being good parents, and should be (and are) permitted to be so. The policy question that a legislator must resolve is a different one, and turns on an assessment of whether the marriage structure proposed by the plaintiffs will, over time, if endorsed and supported by the State, prove to be as stable and successful a model as the one that has formed a cornerstone of our society since colonial times, or prove to be less than optimal, and result in consequences, perhaps now unforeseen, adverse to the State’s legitimate interest in promoting and supporting the best possible social structure in which children should be bom and raised. Given the critical importance of civil marriage as an organizing and stabilizing institution of society, it is eminently rational for the Legislature to postpone making fundamental changes to it until such time as there is unanimous scientific evidence, or popular consensus, or both, that such changes can safely be made.37

There is no reason to believe that legislative processes are *394inadequate to effectuate legal changes in response to evolving evidence, social values, and views of fairness on the subject of same-sex relationships.38 Deliberate consideration of, and incremental responses to rapidly evolving scientific and social understanding is the norm of the political process — that it may seem painfully slow to those who are already persuaded by the arguments in favor of change is not a sufficient basis to conclude that the processes are constitutionally infirm. See, e.g., Massachusetts Fed’n of Teachers v. Board of Educ., 436 Mass. 763, 778 (2002); Mobil Oil v. Attorney Gen., 361 Mass. 401, 417 (1972) (Legislature may proceed piecemeal in addressing perceived injustices or problems). The advancement of the rights, privileges, and protections afforded to homosexual members of our community in the last three decades has been significant, and there is no reason to believe that that evolution will not continue. Changes of attitude in the civic, social, and professional communities have been even more profound. Thirty years ago, The Diagnostic and Statistical Manual, the seminal handbook of the American Psychiatric Association, still listed homosexuality as a mental disorder. Today, the Massachusetts Psychiatric Society, the American Psychoanalytic Association, and many other psychiatric, psychological, and social science organizations have joined in an amicus brief on behalf of the plaintiffs’ cause. A body of experience and evidence has provided the basis for change, and that body continues to mount. The Legislature is the appropriate branch, both constitutionally and practically, to consider and respond to it. It is not enough that we as Justices might be personally of the view that we have learned enough to decide what is best. So long as the question is at all debatable, it must be the Legislature that decides. The marriage statute thus meets the requirements of the *395rational basis test. Accord Standhardt v. Superior Court, 77 P.3d 451 (Ariz. Ct. App. 2003) (marriage statutes rationally related to State’s legitimate interest in encouraging procreation and child rearing within marriage); Baker v. Nelson, 291 Minn. 310, 313 (1971) (“equal protection clause of the Fourteenth Amendment, like the due process clause, is not offended by the state’s classification of persons authorized to marry”); Singer v. Hara, 11 Wash. App. 247, 262-263 (1974) (“There can be no doubt that there exists a rational basis for the state to limit the definition of marriage to exclude same-sex relationships”).

D. Conclusion. While “[t]he Massachusetts Constitution protects matters of personal liberty against government incursion as zealously, and often more so, than does the Federal Constitution,” ante at 328, this case is not about government intrusions into matters of personal liberty. It is not about the rights of same-sex couples to choose to live together, or to be intimate with each other, or to adopt and raise children together. It is about whether the State must endorse and support their choices by changing the institution of civil marriage to make its benefits, obligations, and responsibilities applicable to them. While the courageous efforts of many have resulted in increased dignity, rights, and respect for gay and lesbian members of our community, the issue presented here is a profound one, deeply rooted in social policy, that must, for now, be the subject of legislative not judicial action.

1

The rational basis standard applied under the Massachusetts Constitution and the Fourteenth Amendment to the United States Constitution is the same. See Chebacco Liquor Mart, Inc. v. Alcoholic Beverages Control Comm’n, 429 Mass. 721, 722-723 (1999).

2

The same semantic sleight of hand could transform every other restriction on marriage into an infringement of a right of fundamental importance. For example, if one assumes that a group of mature, consenting, committed adults can form a “marriage,” the prohibition on polygamy (G. L. c. 207, § 4), infringes on their “right” to “marry.” In legal analysis as in mathematics, it is fundamentally erroneous to assume the truth of the very thing that is to be proved.

3

Casting the right to civil marriage as a “fundamental right” in the constitutional sense is somewhat peculiar. It is not referred to as such in either the State or Federal Constitution, and unlike other recognized fundamental rights (such as the right to procreate, the right to be free of government restraint, or the right to refuse medical treatment), civil marriage is wholly a creature of State statute. If by enacting a civil marriage statutory scheme Massachusetts has created a fundamental right, then it could never repeal its own statute without violating the fundamental rights of its inhabitants.

4

For example, see G. L. c. 272, §§ 14 and 18, the Massachusetts adultery and fornication statutes.

5

While the facts of Griswold v. Connecticut, 381 U.S. 479 (1965), involved a married couple, later decisions clarify that its holding was not premised on the marriage relationship. See Carey v. Populations Servs. Int’l, 431 U.S. 678, 687 (1977) (stating that Griswold rested on “right of the individual” to be free from governmental interference with child-bearing decisions [emphasis in original]); Eisenstadt v. Baird, 405 U.S. 438, 453-454 (1972) (same).

6

Contrast Lawrence v. Texas, 123 S. Ct. 2472 (2003), in which the United States Supreme Court struck down the Texas criminal sodomy statute because it constituted State intrusion on some of these very choices.

7

The statutes from which our current marriage laws derive were enacted prior to or shortly after the adoption of our Constitution in 1780, and “may well be considered ... as affording some light in regard to the views and intentions of [the Constitution’s] founders.” Merriam v. Secretary of the Commonwealth, 375 Mass. 246, 253 (1978).

8

Tobin’s Case, 424 Mass. 250, 252-253 (1997) (no fundamental right to receive workers’ compensation benefits); Doe v. Superintendent of Schs. of Worcester, 421 Mass. 117, 129 (1995) (no fundamental right to education); Williams v. Secretary of the Executive Office of Human Servs., 414 Mass. 551, 565 (1993) (no fundamental right to receive mental health services); Matter of Tocci, 413 Mass. 542, 548 n.4 (1992) (no fundamental right to practice law); Rushworth v. Registrar of Motor Vehicles, 413 Mass. 265, 269 n.5 (1992) (no fundamental right to operate motor vehicle); English v. New England Med. *372Ctr., Inc., 405 Mass. 423, 429 (1989), cert, denied, 493 U.S. 1056 (1990) (no fundamental right to recover tort damages); Commonwealth v. Henry’s Drywall Co., 366 Mass. 539, 542 (1974) (no fundamental right to pursue one’s business). Cf. Aime v. Commonwealth, 414 Mass. 667, 674 n.10 (1993) (recognizing right to be free from physical restraint “does not involve judicial derivation of controversial ‘new’ rights from the Constitution”). See generally Williams v. Secretary of the Executive Office of Human Servs., supra at 566 (recognizing fundamental right to receive mental health services “would represent an enormous and unwarranted extension of the judiciary into the [Department of Mental Health]’s authority”); Ford v. Grafton, 44 Mass. App. Ct. 715, 730-731, cert, denied, 525 U.S. 1040 (1998), quoting DeShaney v. Winnebago County Dep’t of Social Servs., 489 U.S. 189, 203 (1989) (“people of Massachusetts may choose by legislation to [provide remedies for “grievous harm”] . . . however, ‘they should not have [such remedies] thrust upon them by this Court’s expansion of the Due Process Clause . . .”).

9

See Michael H. v. Gerald D., 491 U.S. 110, 122-123 & n.3, 127 (1989) (plurality opinion) (limits on substantive due process rights center on “respect for the teachings of history”); Griswold v. Connecticut, 381 U.S. 479, 501 (1965) (Harlan, J., concurring) (same).

10

Compare Curtis v. School Comm. of Falmouth, 420 Mass. 749, 756 (1995), cert, denied, 516 U.S. 1067 (1996), quoting Wisconsin v. Yoder, 406 U.S. 205, 232 (1972) (“primary role of the parents in the upbringing of their children is now established beyond debate as an enduring American tradition”); Aime v. Commonwealth, supra at 676 (“right to be free from governmental detention and restraint is firmly embedded in the history of Anglo-American law”); Brophy v. New England Sinai Hosp., Inc., 398 Mass. 417, 430 (1986) (right to make decisions to accept or reject medical treatment “has its roots deep in our history” and “has come to be widely recognized and respected”); and Moe v. Secretary of Admin. & Fin., 382 Mass. 629, 649 (1981) (characterizing decision whether to bear child as “hold[ing] a particularly important place in the history of the right of privacy” and finding “something approaching consensus” on right to refuse unwanted infringement of bodily integrity), with Trigones v. Attorney Gen., 420 Mass. 859, 863 (1995), quoting Medina v. California, 505 U.S. 437, 445 (1992) (upholding statute that does not “offend some principle of justice so rooted in the tradition and conscience of our people as to be ranked fundamental”); Three Juveniles v. Commonwealth, 390 Mass. 357, 364 (1983), cert, denied sub nom. Keefe v. Massachusetts, 465 U.S. 1068 (1984) (declining to find fundamental right to child-parent privilege where “[n] either Congress nor the Legislature of any State has seen fit to adopt a rule granting [such] a privilege . . .”); Commonwealth v. Stowell, 389 Mass. 171, 174 (1983), quoting Roe v. Wade, 410 U.S. 113, 152 (1973) (declining to recognize right not “implicit in the concept of ordered liberty”).

11

Because of the absence of deep historical roots, every court but one that has considered recognizing a fundamental right to same-sex marriage has declined to do so. See, e.g., Standhardt v. Superior Court, 77 P.3d 451 (Ariz. Ct. App. 2003); Dean v. District of Columbia, 653 A.2d 307, 333 (D.C. 1995) (per curiam) (Ferren, J., concurring in part and dissenting in part); Baehr v. Lewin, 74 Haw. 530, 556-557 (1993); Baker v. Nelson, 291 Minn. 310, 312-314 (1971); Storrs v. Holcomb, 168 Misc. 2d 898, 899-900 (N.Y. Sup. Ct. 1996), dismissed, 245 A.D.2d 943 (N.Y. 1997). The one exception was the Alaska Superior Court, which relied on that State’s Constitution’s express and broadly construed right to privacy. Brause vs. Bureau of Vital Statistics, No. 3AN-95-6562CJ (Alaska Super. Ct. Feb. 27, 1998).

12

Article 106 is referred to as the Equal Rights Amendment.

13

Justice Greaney views Loving v. Virginia, 388 U.S. 1 (1967), as standing analogously for the proposition that just as a person cannot be barred from marrying another person because of his or her race, a person cannot be barred from marrying another person because of his or her sex. Ante at 346-347 (Greaney, J., concurring). While superficially attractive, this analogy does not withstand closer scrutiny. Unlike Virginia’s antimiscegenation statute, neither the purpose nor effect of the Massachusetts marriage statute is to advantage or disadvantage one gender over the other. This distinction is critical and was central to the Loving decision. More fundamentally, the statute at issue burdened marriage with a requirement that was both constitutionally suspect and unrelated to protecting either the underlying purposes or nature of the institution. In contrast, the limitation of marriage to one man and one woman preserves both its structure and its historic purposes.

14

The commission was composed of five State representatives, three State senators and three gubernatorial appointees. All of the gubernatorial appointees were attorneys.

15

The Washington case cited by the commission was Singer v. Hara, 11 Wash. App. 247 (1974).

16

Modem DNA testing may reveal actual paternity, but it establishes only a genetic relationship between father and child.

17

The normative relationship between husband and wife has changed markedly due to the overwhelming movement toward gender equality both at home and in the marketplace.

18

The availability of a variety of social welfare programs and public education has in many instances affected the status of the marital family as the only environment dedicated to the care, protection, and education of children.

19

No-fault divorce has made the dissolution of marriage much easier than ever before.

20

“It is important to distinguish the individual interests in domestic relations from the social interest in the family and marriage as social institutions.” Pound, Individual Interests in the Domestic Relations, 14 Mich. L. Rev. 177, 177 (1916). The court’s opinion blurs this important distinction and emphasizes the personal and emotional dimensions that often accompany marriage. It is, however, only society’s interest in the institution of marriage as a stabilizing social structure that justifies the statutory benefits and burdens that attend to the status provided by its laws. Personal fulfilment and public celebrations or announcements of commitment have little if anything to do with the purpose of the civil marriage laws, or with a legitimate public interest that would justify them.

21

In support of its conclusion that the marriage statute does not satisfy the rational basis test, the court emphasizes that “[t]he department has offered no evidence that forbidding marriage to people of the same sex will increase the number of couples choosing to enter into opposite-sex marriages in order to have and raise children.” Ante at 334. This surprising statement misallocates the burden of proof in a constitutional challenge to the rational basis of a statute (see supra at 379). It is the plaintiffs who must prove that supporting and promoting one form of relationship by providing (as is pointed out) literally hundreds of benefits, could not conceivably affect the decision-making of anyone considering whether to bear and raise a child. The department is not required to present “evidence” of anything.

22

See C.N. Degler, The Emergence of the Modem American Family, in The American Family in Social-Historical Perspective 61 (3d ed. 1983); AJ. Hawkins, Introduction, in Revitalizing the Institution of Marriage for the Twenty-First Century: An Agenda for Strengthening Marriage xiv (2002); C. Lasch, Social Pathologists and the Socialization of Reproduction, in The American Family in Social-Historical Perspective, 80 (3d ed. 1983); W.J. O’Donnell & D.A. Jones, The Law of Marriage and Marital Alternatives 1 (1982); L. Saxton, The Individual, Marriage and the Family 229-230, 260 (1968); M.A. Schwartz & B.M. Scott, Marriages and Families: Diversity and Change 4 (1994); Wardle, “Multiply and Replenish”: Considering Same-Sex Marriage in Light of State Interests in Marital Procreation, 24 Harv. J.L. & Pub. Pol’y 771, 777-780 (2001); J.Q. Wilson, The Marriage Problem: How Our Culture has Weakened Families 28, 40, 66-67 (2002).

23

See Rodney, Behavioral Differences between African American Male Adolescents with Biological Fathers and Those Without Biological Fathers in the Home, 30 J. Black Stud. 45, 53 (1999) (African-American juveniles who lived with their biological fathers displayed fewer behavioral problems than those whose biological fathers were absent from home); Chilton, Family Disruption, Delinquent Conduct and the Effect of Subclassification, 37 Am. Soc. Rev. 93, 95 (1972) (proportion of youth charged with juvenile offenses who were not living in husband-wife family was larger than comparable proportion of youth charged with juvenile offenses who were living in husband-wife family); Hoffmann, A National Portrait of Family Structure and Adolescent Drug Use, 60 J. Marriage & Fam. 633 (1998) (children from households with both mother and father reported relatively low use of drugs, whereas children from households without their natural mothers and from other family type households had highest prevalence of drug use). See also D. Blankenhorn, Fatherless America: Confronting Our Most Urgent Social Problem 25 (1995).

24

H.B. Biller & J.L. Kimpton, The Father and the School-Aged Child, in The Role of The Father in Child Development 143 (3d ed. 1997); H.B. Biller, Fathers and Families: Paternal Factors in Child Development 1-3 (1993); Lynne Marie Kohm, The Homosexual “Union”: Should Gay and Lesbian Partnerships be Granted the Same Status as Marriage? 22 J. Contemp. L. 51, 61 & nn.53, 54 (1996) (“[statistics continue to show that the most stable family for children to grow up in is that consisting of a father and a mother”).

25

See, e.g., Patterson, Family Relationships of Lesbians and Gay Men, 62 J. Marriage & Fam. 1052, 1060, 1064-1065 (2000) (concluding that there are no significant differences between children of same-sex parents and children of heterosexual parents in aspects of personal development).

26

See, e.g., Cameron, Homosexual Parents, 31 Adolescence 757, 770-774 (1996) (concluding results of limited study consonant with notion that children raised by homosexuals disproportionately experience emotional disturbance and sexual victimization).

27

See, e.g., Stacey, (How) Does the Sexual Orientation of Parents Matter?, 66 Amer. Soc. Rev. 159, 172, 176-179 (2001) (finding significant statistical differences in parenting practices, gender roles, sexual behavior but noting that “heterosexism” and political implications have constrained research). See also Coleman, Reinvestigating Remarriage: Another Decade of Progress, 62 J. Marriage & Fam. 1288 (2000) (concluding that future studies of impact of divorce and remarriage on children should focus on “nontraditional" stepfamilies, particularly same-sex couples with children, because impact of such arrangements have been overlooked in other studies).

28

In Massachusetts, for example, the State’s adoption laws were only recently interpreted to permit adoption by same-sex partners. Adoption of Tammy, 416 Mass. 205 (1993). It is fair to assume that most of the children affected by that ruling, who properly would be the subject of study in their teenage and adult years, are still only children today.

29

This family structure raises the prospect of children lacking any parent of their own gender. For example, a boy raised by two lesbians as his parents has no male parent. Contrary to the suggestion that concerns about such a family arrangement is based on “stereotypical” views about the differences between sexes, ante at 337 n.28, concern about such an arrangement remains rational. It is, for example, rational to posit that the child himself might invoke gender as a justification for the view that neither of his parents “understands” him, or that they “don’t know what he is going tMough,” particularly if Ms disagreement or dissatisfaction involves some issue pertaining to sex. Given that same-sex couples raising children are a very recent phenomenon, the ramifications of an adolescent cMld’s having two parents but not one of his or her own gender have yet to be fully realized and cannot yet even be tested in significant numbers. But see note 25, supra, regarding studies of children raised without parents of each gender.

30

The same could be true of any other potentially promising but recent innovation in the relationships of persons raising children.

31

The plaintiffs also argue that because the State requires insurance companies to provide coverage for diagnosing and treating infertility unrestricted to those who are married, G. L. c. 175, § 47H, limiting marriage to opposite-sex couples is contrary to its currently stated public policy and, therefore, no longer rational. This argument is not persuasive. The fact that the Legislature has seen fit to require that health insurers cover the medical condition of infertility, for all subscribers, is not inconsistent with the State’s policy of encouraging and endorsing heterosexual marriage as the optimum structure in which to bear and raise children. There is no rule that requires the State to limit every law bearing on birth and child rearing to the confines of *390heterosexual marriage in order to vindicate its policy of supporting that structure as optimal. Just as the insurance laws relating to infertility coverage cannot be said to be a State endorsement of childbirth out of wedlock, they cannot be said to represent an abandonment of the State’s policy regarding a preference that children be bom into and raised in the context of heterosexual marriage.

32

Indeed, just recently, this court reasoned that the Legislature could permissibly conclude that children being raised by single parents “may be at heightened risk for certain kinds of harm when compared with children of so-called intact families,” because such children “may not have or be able to draw on the resources of two parents” when having to cope with some form of loss. Blixt v. Blixt, 437 Mass. 649, 663, 664 (2002), cert, denied, 537 U.S. 1189 (2003). In that case, the differences between single parents and parents raising a child together sufficed to justify subjecting single parents to the grandparent visitation statute, G. L. c. 119, § 39D. Id. at 662-664. Because the statute implicated fundamental parental rights, its classifications had to survive strict scrutiny, id. at 660, not the mere rational basis test at issue in today’s opinion. The fact that single people can adopt children did not insulate them from differential treatment with respect to their parental rights.

33

Similarly, while the fact that our laws have evolved to include a strong affirmative policy against discrimination on the basis of sexual orientation, have decriminalized intimate adult conduct, and have abolished the legal distinctions between marital and nonmarital children, may well be a reason to celebrate a more open and humane society, they ought not be the basis on which to conclude that there is no longer a rational basis for the current marriage law. See ante at 332-333. To conclude the latter based on the former threatens the process of social reform in a democratic society. States must be free to experiment in the realm of social and civil relations, incrementally and without concern that a step or two in one direction will determine the outcome of the experiment as a matter of law. If they are not, those who argue “slippery slope” will have more ammunition than ever to resist any effort at progressive change or social experimentation, and will be able to put the lie to the arguments of the proponents of such efforts, that an incremental step *391forward does not preordain a result which neither the people nor their elected representatives may yet be prepared to accept.

34

The court contends that the exclusive and permanent commitment of the marriage partnership rather than the begetting of children is the sine qua non of civil marriage, ante at 332, and that “the ‘marriage is procreation’ argument singles out the one unbridgeable difference between same-sex and opposite-sex couples, and transforms that difference into the essence of legal marriage.” Ante at 333. The court has it backward. Civil marriage is the product of society’s critical need to manage procreation as the inevitable consequence of intercourse between members of the opposite sex. Procreation has always been at the root of marriage and the reasons for its existence as a social institution. Its structure, one man and one woman committed for life, *392reflects society’s judgment as how optimally to manage procreation and the resultant child rearing. The court, in attempting to divorce procreation from marriage, transforms the form of the structure into its purpose. In doing so, it turns history on its head.

The court compounds its error by likening the marriage statute to Colorado’s “Amendment 2,” which was struck by the United States Supreme Court in Romer v. Evans, 517 U.S. 620, 633 (1996). That amendment repealed all Colorado laws and ordinances that barred discrimination against homosexuals, and prohibited any governmental entity from adopting similar statutes. The amendment withdrew from homosexuals, but no others, legal protection from a broad range of injuries caused by private and governmental discrimination, “imposing a broad and undifferentiated disability on a single named group.” Id. at 632. As the Court noted, its sheer breadth seems “inexplicable by anything but animus toward the class it affects.” Id. The comparison to the Massachusetts marriage statute, which limits the institution of marriage (created to manage procreation) to opposite-sex couples who can theoretically procreate, is completely inapposite.

35

Although the marriage statute is overinclusive because it comprehends within its scope infertile or voluntarily nonreproductive opposite-sex couples, this overinclusiveness does not make the statute constitutionally infirm. See Massachusetts Fed’n of Teachers v. Board of Educ., 436 Mass. 763, 778 (2002) (“Some degree of overinclusiveness or underinclusiveness is constitutionally permissible . . .”). The overinclusiveness present here is constitutionally permissible because the Commonwealth has chosen, reasonably, not to test every prospective married couple for fertility and not to demand of fertile prospective married couples whether or not they will procreate. It is satisfied, rather, to allow every couple whose biological opposition makes procreation theoretically possible to join the institution.

36

Concems about such unintended consequences cannot be dismissed as fanciful or far-fetched. Legislative actions taken in the 1950’s and 1960’s in areas as widely arrayed as domestic relations law and welfare legislation have had significant unintended adverse consequences in subsequent decades including the dramatic increase in children bom out of wedlock, and the destabilization of the institution of marriage. See Nonmarital Childbearing in the United States 1940-99, National Center for Health Statistics, 48 Nat’l Vital Stat. Reps, at 2 (Oct. 2000) (nonmarital childbirths increased from 3.8% of annual births in 1940 to 33% in 1999); M.D. Bramlett, Cohabitation, Marriage, Divorce, and Remarriage in the United States, National Center for Health Statistics, Vital & Health Stat. at 4-5 (July 2002) (due to higher divorce rates and postponement of marriage, proportion of people’s lives spent in marriage declined significantly during later half of Twentieth Century).

37

“[T]he State retains wide latitude to decide the manner in which it will allocate benefits.” Moe v. Secretary of Admin. & Fin., 382 Mass. 629, 652 (1981). To the extent that the Legislature concludes that one form of social relationship is more optimal than another for the bearing and raising of children, it is free to promote and support the one and not the other, so long as its conclusion is rational, and does not discriminatorily burden the exercise of a fundamental right. Id. Cf. Rust v. Sullivan, 500 U.S. 173, 192-193 (1991) (“Government can, without violating the Constitution, selectively fund a program to encourage certain activities it believes to be in the public interest, without at the same time funding an alternative program which seeks to deal with the problems in another way”).

38

Legislatures in many parts of the country continue to consider various means of affording same-sex couples the types of benefits and legal structures that married couples enjoy. For example, in 1999 the California Legislature established the first Statewide domestic partner registry in the nation, and in each of the years 2001, 2002, and 2003 substantially expanded the rights and benefits accruing to registered partners. Cal. Fam. Code §§ 297 et seq. (West Supp. 2003). See also comments of Massachusetts Senate President Robert Travaglini to the effect that he intends to bring civil union legislation to the floor of the Senate for a vote. Mass. Senate Eyes Civil Unions: Move Comes as SJC Mulls Gay Marriages, Boston Globe, Sept. 7, 2003, at Al.