Dean v. District of Columbia

653 A.2d 307 (1995)

Craig Robert DEAN and Patrick Gerard Gill, Appellants,
v.
DISTRICT OF COLUMBIA, et al., Appellees.

No. 92-CV-737.

District of Columbia Court of Appeals.

Argued November 2, 1993. Decided January 19, 1995.

William N. Eskridge, Jr., Georgetown University Law Center, with whom Craig Robert Dean, Washington, DC, was on the brief, for appellants.

James C. McKay, Jr., Asst. Corp. Counsel, with whom John Payton, Corp. Counsel at the time the brief was filed, and Charles L. Reischel, Deputy Corp. Counsel, Washington, DC, were on the brief, for appellee.

Laura A. Foggan, with whom Richard A. Gross, Washington, DC, was on the brief, for amicae curiae Elizabeth A. Leader and Barbara R. Lewis.

Arthur B. Spitzer, W. Stephen Smith, and Micki M. Chen, Washington, DC, filed an amicus curiae brief for the American Civil Liberties Union of the Nat. Capital Area, on behalf of appellants.

Evan Wolfson and A. Christopher Wieber, New York City, filed an amicus curiae brief for Lambda Legal Defense and Educ. Fund, Inc., on behalf of appellants.

*308 Before FERREN, TERRY, and STEADMAN, Associate Judges.

PER CURIAM:

The judgment of the trial court is affirmed pursuant to Parts I., II., III., and V. of Judge FERREN's opinion and the concurring opinions of Judges TERRY and STEADMAN.

              TABLE OF CONTENTS FOR OPINION OF JUDGE FERREN
                                                                                 Page
INTRODUCTION                                                                      309
   I. PROCEEDINGS TO DATE                                                         309
  II. THE MARRIAGE STATUTE CLAIM                                                  310
      A. Legislative History of the Marriage Statute                              310
      B. Statutory Definition of "Marriage"                                       312
      C. The Marriage Statute as Part of a Larger Legislative Scheme, Including
          the Divorce Statute                                                     314
      D. The Traditional Understanding of "Marriage"                              315
      E. Case Law from Other Jurisdictions                                        315
      F. The Anti-Sex Discriminatory Language Act of 1976                         316
      G. The 1982 Gender Rule of Construction                                     317
  III. THE HUMAN RIGHTS ACT CLAIM                                                 318
  IV. THE CONSTITUTIONAL ISSUES: PROCEDURAL AND ANALYTICAL PREREQUISITES          320
     A. Whether Constitutional Issues Have Been Properly Raised                   320
     B. Standard of Review                                                        321
     C. Relevance of the Distinction Between "Adjudicative Facts" and
        "Legislative Facts"                                                       322
       1. In General                                                              322
       2. "Adjudicative" and "Legislative" Facts Distinguished 323
       3. Judicial Process of Legislative Fact-Finding 326
   V. CONSTITUTIONAL DUE PROCESS: IS SAME-SEX MARRIAGE A "FUNDAMENT
      RIGHT"?                                                                     331
    A. Definition of "Fundamental Right"                                          331
    B. Appellants' Due Process Claim                                              332
  VI. EQUAL PROTECTION: ARE HOMOSEXUALS A "SUSPECT" OR " QUASI-SUSPE(
      CLASS?                                                                      333
    A. The Trial Court's Ruling                                                   333
    B. Introduction: Discrimination and Equal Protection of the Laws              334
    C. The Attributes of Marriage Justifying an Equal Protection Inquiry          335
    D. Summary Judgment for Appellants Inappropriate Assuming, for the
        Sake of Argument, That the Rational Basis Test Applies                    336
    E. Summary Judgment for the District Inappropriate Assuming, for the
        Sake of Argument, that Strict Scrutiny Applies                            337
    F. Constitutionally Protected Classes: United States v. Carolene
       Products Co.                                                          337
    G. Equal Protection After Carolene Products Co.: "Suspect"
       and "Quasi-Suspect" Classes                                                338
    H. The Implications, If Any, of Bowers v. Hardwick for
       Equal Protection Analysis                                                  340
    I. The Factors Applicable to Determining "Suspect" and "Quasi-Suspect"
       Class Status                                                               344
       1. History of Purposeful Discrimination                                    344
       2. Deep-Seated Prejudice Causing Inaccurate Stereotypes That Do
            Not Reflect Class Members' Abilities                                  345
       3. Immutability                                                            346
       4. Political Powerlessness                                                 349
    J. Whether Homosexuals Comprise a "Suspect" or "Quasi-Suspect" Class          351
       1. Three Easily Applied Factors                                            351
       2. Immutability                                                            351
       3. The Prevention/Immutability Distinction                                 352
    K. Whether the District Has a "Substantial" or "Compelling" State Interest
       in Barring Same-Sex Marriage                                               355
    L. Proposed Disposition: Reversal and Remand for Trial                        356
VII. POSTSCRIPT: RESPONSE TO MAJORITY ON EQUAL PROTECTION                         359

*309 FERREN, Associate Judge, concurring in part and dissenting in part:

Plaintiff-appellants, two homosexual men, want to marry each other. They appeal from the trial court's order granting summary judgment for the District of Columbia, rejecting their complaint for an injunction to require the Clerk of the Superior Court to issue them a marriage license. Appellants contend the trial court erred in three respects: (1) ruling that the District of Columbia marriage statute, D.C.Code §§ 30-101 to -121 (1993), prohibits the Clerk from issuing marriage licenses to same-sex couples; (2) ruling that the Clerk did not unlawfully discriminate against appellants under the District of Columbia Human Rights Act, D.C.Code §§ 1-2501 to -2557 (1992), by refusing to issue them a marriage license; and (3) interpreting the marriage statute in a way that unconstitutionally deprives same-sex couples of the right to marry.

We find no statutory violation or denial of due process, but, unlike the majority, I believe there are genuine issues of material fact precluding summary judgment on appellants' constitutional claim that they have been denied equal protection of the laws. Specifically, I conclude that a trial is required to determine whether same-sex couples comprise a "suspect" or a "quasi-suspect" class entitled either to "strict" or to "intermediate" scrutiny of governmental discrimination against them—in this case denial of the right to marry. Furthermore, if, as a result of the trial, the court decides that same-sex couples do comprise such a protected class, the trial will also be needed to determine whether the District, despite such constitutionally protected status for same-sex couples, has a "compelling," or at least a "substantial," governmental interest in keeping the marriage limitation to heterosexual couples as is. If it does, the District would prevail, otherwise not.

Recognition that homosexuals—like racial minorities and women—are entitled to special constitutional protection, therefore, would not necessarily mean that homosexuals are constitutionally entitled to marry one another; not all governmental discrimination against constitutionally protected groups is forbidden. But I do not believe that this court can properly conclude at this point—as a matter of law without benefit of a trial— that appellants have failed to proffer an equal protection claim. Accordingly, unlike my colleagues, who dispose of the matter summarily, I would reverse and remand this case for trial on the equal protection issue.

I. PROCEEDINGS TO DATE

On November 13, 1990, appellants Craig Robert Dean and Robert Gerard Gill applied for a marriage license from the Clerk of the Superior Court, as required by D.C.Code § 30-110. The Clerk, acting as head of the Marriage License Bureau, denied their application, explaining his action in a memorandum to the Director of the court's Family Division:

Title 30-110 of the District of Columbia Code authorizes the Clerk of the Court to grant or deny applications for marriage licenses.
The sections of the District of Columbia Code governing marriages do not authorize marriage between persons of the same sex. Therefore the application for a marriage license in this case is respectfully denied.

Appellants filed an action in Superior Court seeking declaratory and injunctive relief requiring the Clerk to issue them a marriage license. In their amended complaint, appellants alleged that they qualified for the license because the marriage statute, D.C.Code §§ 30-101 to -121, is "gender-neutral" and, further, because the Clerk, by refusing to issue the license, had discriminated against them on grounds of sex or sexual orientation, in violation of the District of Columbia Human Rights Act, D.C.Code §§ 1-2501 to -2557.

The District moved to dismiss for failure to state a claim for relief. Judge Bowers granted summary judgment for the District. He explained that all definitional sources for "marriage"—the legislative history of the Marriage and Divorce Act, D.C.Law 1-107, 1977 D.C.Stat. 114; the various references to *310 gender in relevant provisions of the District of Columbia Code; the common law of the District of Columbia; decisions of appellate courts in other states; references to marriage in the Bible; and dictionary definitions of "marriage"—show that marriage inherently requires one male and one female participant. Furthermore, according to Judge Bowers, the "City Council consciously chose not to make the language of the Human Rights Act applicable to the regulation of the marital relationship." He drew that conclusion from the fact that, only a few months before the Council enacted the Human Rights Act, it had rejected a proposal expressly to permit same-sex marriages. Judge Bowers concluded:

[P]laintiffs were denied a marriage license because of the nature of marriage itself, requiring, as it does, that the parties thereto be a male and a female. What the plaintiffs herein sought a license to enter into, by definition, simply was not a "marriage." Any change in that definition must come from the legislature—not this Court.

Appellants moved for reconsideration. They asserted that, in ruling that same-sex couples did not have the right to marry, the trial court interpreted the marriage statute and the Human Rights Act in ways that violated appellants' constitutional rights to due process and equal protection of the laws under the Fifth Amendment, and violated the establishment of religion clause of the First Amendment as well. In an opinion issued on June 2, 1992, Judge Bowers granted the motion for reconsideration but rejected appellants constitutional claims. Appellants filed a timely appeal on both statutory and constitutional grounds.[1]

II. THE MARRIAGE STATUTE CLAIM

Appellants contend that the Clerk should be required to issue them a marriage license because the marriage statute is gender-neutral and does not expressly prohibit same-sex marriages. We cannot agree. The language and legislative history of the marriage statute demonstrate that neither Congress nor the Council of the District of Columbia has ever intended to define "marriage" to include same-sex unions.

A. Legislative History of the Marriage Statute

On March 3, 1901, Congress enacted the first District of Columbia Code. See An Act to Establish a Code of Law for the District of Columbia, 31 Stat. 1189, ch. 854 (1901). That Act read: "[A]ll acts of Congress by their terms applicable to the District of Columbia... in force at the date of the passage of this act shall remain in force except in so far as the same are inconsistent with, or are replaced by, some provision of this code." Id., ch. 1, § 1, at 1189. Chapter 43 of the Code addressed marriage;[2] Chapter 22 dealt with *311 divorce.[3] The current marriage provision, D.C.Code §§ 30-101 to -121 and, for the most part, the present divorce provisions, D.C.Code §§ 16-901 to -924 (1989 & Supp. 1993), are essentially the same as those enacted in 1901.[4]

The only significant changes in the marriage and divorce provisions since 1901 occurred in the Marriage and Divorce Act of 1977, D.C.Law 1-107, 1977 D.C.Stat. 114. See generally Samuel Green & John V. Long, The Real and Illusory Changes of the 1977 Marriage and Divorce Act, 27 CATH.U.L.REV. 469 (1978).[5] Before that Act was adopted, however, Councilmember Arrington Dixon had introduced substantially different legislation, Bill No. 1-89, the "District of Columbia Marriage and Divorce" bill, which would have completely repealed and redrafted D.C.Code §§ 30-101 to -121 (1973) (marriage) and §§ 16-901 to -924 (1973) (divorce). See The District of Columbia Marriage and Divorce Act, Bill 1-89 (May 6, 1975, with amendments proposed July 7, 1975) (hereafter Original Bill 1-89). In particular, Bill No. 1-89 would have changed § 30-101 to read: "A marriage between two persons which is licensed, solemnized and registered as provided in this Act is valid in the District of Columbia." (Emphasis added.) Id. During a public hearing on Bill 1-89, Councilmember Dixon explained that the bill would permit marriages between persons of the same sex. See Councilmember Arrington Dixon, Opening Statement at Public Hearings on Bill No. 1-89 2 (June 7-8, 1975). Although the language of the bill did not directly authorize same-sex marriages, proposed § 30-112 expressly referred to such unions and thus indirectly acknowledged their validity:

(a) The court shall enter its decree declaring the invalidity of a marriage entered into under the following circumstances:
* * * * * *
(2) a party lacks the physical capacity to consummate the marriage by sexual intercourse, and at the time the marriage was solemnized the other party did not know of the incapacity; provided that *312 this clause shall not apply to married persons of the same sex.

Original Bill 1-89, § 30-112 (Emphasis added).

If Bill 1-89 had become law, same-sex marriages clearly would have been authorized in the District. Because of the fervent debate generated by Bill 1-89,[6] however, Councilmember Dixon moved to substitute for original Bill 1-89 another bill proposed by the bar associations. This substitute bill was the one Council enacted as the Marriage and Divorce Act of 1977. See 1977 D.C.Stat. 119. Rather than entirely repealing and redrafting D.C.Code §§ 30-101 to -121 (1973) (marriage) and §§ 16-901 to -924 (1973) (divorce), the 1977 Act merely amended existing code provisions. See COMM. ON THE JUDICIARY AND CRIM.LAW, REPORT ON BILL No. 1-89, THE DISTRICT OF COLUMBIA MARRIAGE AND DIVORCE ACT, at 1-5 (June 24, 1976). The 1977 Act contained no reference to same-sex marriage. In fact, it made only the following two changes in the long-standing marriage chapter:

Sec. 113. (a) Section 1291 of the Act of March 3, 1901 (D.C.Code, sec. 30-110) is amended by striking "names, ages and color" and inserting in lieu thereof "names and ages."[[7]]
(b) Section 1296 of the Act of March 3, 1901 (D.C.Code, sec. 30-116) is repealed.[[8]]

[O. 7] 1977 D.C.Stat. 119. In sum, the 1977 Act made no change germane to the issue before us; we are left to interpret "marriage" as understood by the Congress that enacted and codified the marriage statute in 1901 and, later, as understood by the Council that implicitly reconfirmed existing provisions of that statute while amending others through the 1977 Act.

B. Statutory Definition of "Marriage"

Because the present marriage statute is essentially the same as the 1901 statute, which was derived from even earlier legislation, see supra note 2, the initial question is: what did Congress intend by the word "marriage" when it enacted the marriage statute in 1901?[9]

Citing our well-known interpretive criteria, appellants stress that we should focus, first, on the plain words of the statute. See Peoples Drug Stores v. District of Columbia, 470 A.2d 751, 753 (D.C.1983) (en banc). They contend that since its inception the marriage statute, standing alone, has been essentially gender-neutral and, for that reason, has always authorized same-sex marriages.

*313 One provision of the present marriage statute—and only one—has always used genderspecific language.[10] The so-called consanguinity provision, D.C.Code § 30-101, refers to marriages of a "man" with a "wife," and of a "woman" with a "husband."[11] Appellants maintain that the policy underlying this provision is the prevention of genetic birth defects and mental retardation, a policy that necessarily would not apply to same-sex couples because they cannot produce children together. From this policy premise, appellants argue that this one statutory exception to gender neutrality in the marriage statute actually substantiates their argument that the right to marry extends to same-sex couples. According to appellants, because the only gender references are contained in a public health provision, relevant only to opposite-sex couples, it follows, perforce, that the omission of gender references in all other provisions of the marriage statute necessarily implies that same-sex marriages are permitted. See McCray v. McGee, 504 A.2d 1128, 1130 (D.C.1986) (basic rule of statutory construction is "that when a legislature makes express mention of one thing, the exclusion of others is implied, because there is an inference that all omissions should be understood as exclusions").

Appellants' argument, by its own terms, only has force if the consanguinity provision is limited to prohibitions against biological inbreeding. It is not. While that concern is obvious in the prohibition of a man's marrying his sister or a woman's marrying her father, there is no genetic danger in other prohibited situations; for example, the prohibitions against a man's marrying his son's wife or a woman's marrying her stepfather. See D.C.Code § 30-101(1) and (2), supra note 11. The consanguinity provision, therefore, reflects taboos—indeed moral judgments about improper marriage relationships—that transcend genetic concerns.

The use of gender-based terminology in § 30-101 to prohibit certain marriages, therefore, reflects a legislative understanding that marriage, as understood by Congress at the time of original enactment and thereafter, is inherently a male-female relationship. If that were not so, some of the statutory prohibitions not based on genetic reproductive concerns either would not be there or, to be consistent, would have been extended, for example, to prohibit a man's marrying his stepfather (just as a man cannot lawfully marry his stepmother) or to prevent a woman's marrying her wife's father (just as a woman cannot lawfully marry her husband's father). See D.C.Code § 30-101, supra note 11.

If appellants were to prevail in their statutory interpretation, the law would permit same-sex couples to enter into some kinds of marriage relationships that the statute forbids for opposite-sex couples, even though *314 such relationships would not be genetically dangerous for any kind of marriage. Indeed, if men could marry men, § 30-101 would not preclude a bi-sexual man who may have had a biological son from marrying that son, or from marrying his own father or brother. We do not believe that Congress, almost a century ago, envisioned such possibilities, given the consanguinity prohibitions imposed on opposite sex couples. See supra note 11. Nor is there any indication that more recent Congresses, or the Council in amending the marriage statute, ever modified the fundamental legislative understanding that "marriage" is limited to opposite-sex couples. Appellants' argument that § 30-101 reflects merely public health limitations, leaving room for all genetically safe marriages regardless of gender, accordingly fails.

In sum, to conclude that Congress intended to permit same-sex marriages would mean that Congress in 1901 intended to permit various categories of genetically safe, same-sex marriages that were denied, though genetically safe, to opposite-sex couples. There is no evidence this was the case; the consanguinity provisions, far from supporting appellants' argument, actually reinforce the government's position that the legislature never had same-sex marriages in mind when adopting, codifying, or amending the marriage statute.

C. The Marriage Statute as Part of a Larger Legislative Scheme, Including the Divorce Statute

Our conclusion that the marriage statute does not authorize same-sex marriages is buttressed by looking at the larger statutory scheme of which it is a part. See Citizens Ass'n of Georgetown v. Zoning Comm'n, 392 A.2d 1027, 1033 (D.C.1978) (en banc) ("It is a canon of statutory interpretation that one looks at the particular statutory language within the context of the whole legislative scheme when legislative intent is to be determined."); see also 2A NORMAN J. SINGER, SUTHERLAND ON STATUTORY CONSTRUCTION § 46.05 (5th ed. 1992). As indicated earlier, see supra notes 2 and 3, the marriage and divorce statutes originally were enacted at different times, but the modern statutes were both enacted in 1901. See supra notes 2 and 9. Nonetheless, these statutes from the beginning have been placed in different chapters of the D.C.Code, perhaps suggesting that they are not part of the same legislative scheme. On the other hand, there is, necessarily, a logical relationship between the marriage and divorce statutes; if same-sex couples can marry then, presumably, they must be able to divorce. Thus, as elaborated below, we can look at the divorce statute for clues as to how Congress, as well as the Council, has understood "marriage."

The divorce statute, traceable to congressional legislation before codification in 1901, see supra note 3, is replete with gender-distinctive references. See D.C.Code §§ 16-901 to -924 (1989). More specifically, the following provisions contain the terms "husband" and "wife": D.C.Code's §§ 16-904(d)(1) (annulment granted if either party has a husband or wife living), -911 (husband or wife must pay alimony to other spouse pending divorce), -912 (husband or wife may retain right of dower in other's estate), -913 (husband or wife may be required to pay alimony when divorce is granted), -916 (court may decree permanent alimony if husband or wife fails to maintain needy spouse).[12] Accordingly, when the marriage statute is read in context with the broader legislative scheme that includes divorce, one cannot say that marriage is gender-neutral.

Although we attribute corroborative, not determinative, significance to the divorce statute's pervasive use of gender terminology, it is significant that Congress enacted and codified the divorce chapter at the same *315 time it enacted and codified the marriage chapter, in 1901, using gender-specific terminology in each. See District of Columbia v. Thompson, 593 A.2d 621, 630 (D.C.1991) (when legislature enacts two statutes at same time and the statutes have similar subject matter and purpose, principle of in pari materia dictates that the statutes should be read with reference to each other). That basic language has been carried forward ever since, both by Congress and by the Council of the District of Columbia. See supra note 3. This statutory evolution, therefore, strongly suggests a consistent legislative understanding and intent that "marriage" means—and thus is limited to—unions between persons of opposite sexes.

D. The Traditional Understanding of "Marriage"

Our statutory understanding is further confirmed by the ordinary sense and meaning traditionally attributed to the word "marriage" when used to indicate an intimate relationship. See Barbour v. District of Columbia Dep't of Employment Servs., 499 A.2d 122, 125 (D.C.1985) ("Words of a statute must be construed by their common meaning and their ordinary sense."); In re Estate of Shutack, 469 A.2d 427, 429 (D.C.1983) ("The words of the statute should be construed according to their ordinary sense and with the meaning commonly attributed to them."). Black's Law Dictionary defines marriage as the "[l]egal union of one man and one woman as husband and wife." BLACK'S LAW DICTIONARY 972 (6th ed. 1990) (emphasis added). The second edition of Black's Law Dictionary printed in 1910—presumably reflecting the common understanding at the time the marriage statute was enacted in 1901—defined marriage as "the civil status of one man and one woman united in law for life, for the discharge to each other and the community of the duties legally incumbent on those whose association is founded on the distinction of sex." BLACK'S LAW DICTIONARY 762 (2d ed. 1910) (emphasis added). Similarly, Webster's Dictionary from 1902 defined "marry" as follows: "[t]o unite in wedlock or matrimony; to join, as a man and woman, for life; to make man and wife." WEBSTER'S MODERN DICTIONARY 281 (1902) (emphasis added). The same dictionary today defines marriage as "the state of being united to a person of the opposite sex as husband or wife." WEBSTER'S THIRD NEW INTERNATIONAL DICTIONARY 1384 (1986) (emphasis added). Although this edition of Webster's, as a third definition, defines marriage as "an intimate or close union," see id., and commitment and union are undoubtedly fundamental elements of a marriage, we are satisfied that the ordinary understanding of the word "marriage"—both at the turn of the century when the marriage statute was enacted and in modern times when that statute was amended—means the union of two members of the opposite sex.

Of course, the meanings of words are continually evolving, and we do not overlook the fact that the terms "marriage" and "gay marriage" are used colloquially today to refer to long-term same-sex relationships between gays and between lesbians. See Cory & LeRoy, Homosexual Marriage, 29 SEXOLOGY 660 (1963). Our task, however, is to determine what the legislature intended "marriage" to mean when the marriage statute was enacted, codified, or amended. Given the statutory language used, buttressed by the usual definition of "marriage," we cannot conclude that any legislature for the District of Columbia that has addressed the marriage statute has ever intended to authorize same-sex unions.

E. Case Law from Other Jurisdictions

Although not at all dispositive here, we note that the cases from other jurisdictions with marriage statutes similar to the District's—neither expressly prohibiting nor expressly authorizing same-sex marriages— have uniformly interpreted "marriage," by definition, as requiring two members of opposite sexes. The Supreme Court of Minnesota, for example, explained:

Minn.St. c. 517, which governs "marriage," employs that term as one of common usage, meaning the state of union between persons of the opposite sex. It is unrealistic to think that the original draftsmen of our marriage statutes, which date from territorial days, would have used the term in any different sense. *316 Baker v. Nelson, 291 Minn. 310, 191 N.W.2d 185, 185-86 (1971), appeal dismissed, 409 U.S. 810, 93 S. Ct. 37, 34 L. Ed. 2d 65 (1972). See also Jones v. Hallahan, 501 S.W.2d 588, 589 (Ky.1973) (same-sex couple incapable of entering into marriage as the term is defined); M.T. v. J.T., 140 N.J.Super. 77, 355 A.2d 204, 208 (App.Div.1976) ("requirement that marriage must be between a man and a woman ... is so strongly and firmly implied from a full reading of the statutes that a different legislative intent, one which would sanction a marriage between persons of the same sex, cannot be fathomed"); Singer v. Hara, 11 Wash.App. 247, 522 P.2d 1187, 1191 (1974) (marriage statute "clearly founded upon the presumption that marriage, as a legal relationship, may exist only between one man and one woman"); Peter G. Guthrie, Annotation, Marriage Between Persons of the Same Sex, 63 A.L.R. 3d 1199, 1199 (1975) ("In all cases so far discovered which have considered the question whether persons of the same sex may marry each other, the view has been taken that since the marriage relationship has always been the union of a man and a woman as husband and wife, there may be no valid contract entered into between persons of the same sex").[13]

While these cases do not deal with our local statute, they at least reflect the interpretive approach we apply here and thus provide precedent in the sense of analytical support for the result we reach.

F. The Anti-Sex Discriminatory Language Act of 1976

Appellants do not rest on their own interpretation of the marriage statute. They argue that the Council itself, through two separate pieces of legislation, has indirectly confirmed (or reinterpreted) the statutory definition of marriage in ways that guarantee the right to same-sex marriages.

The first interpretive legislation was the Anti-Sex Discriminatory Language Act of 1976. See 1976 D.C.Stat. 194.[14] This Act, which among other things amended the 1901 marriage statute, sought "to achieve equality under the law for men and women by eliminating sex-based distinctions in the District of Columbia Code, so that the rights and responsibilities of persons under D.C. law will not be different solely on the basis of their sex." COMM. ON THE JUDICIARY AND CRIM.LAW, REPORT ON BILL No. 1-36, THE ANTI-SEX DISCRIMINATORY LANGUAGE ACT, at 2-3 (May 20, 1976) (hereafter COMM. REPORT ON BILL 1-36). The Anti-Sex Discriminatory Language Act made only one change in the marriage statute:

Sec. 32. Section 1292 of the Act of March 3, 1901 (D.C.Code, sec. 30-111), is amended by striking out "unless the father of such persons, or if there be no father, the mother," and inserting in lieu thereof "unless a parent."[[15]]

1976 D.C.Stat. 201.

Because Councilmember Dixon's Bill 1-89 recognizing same-sex marriages was pending at the time the Council was discussing Bill 1-36 to establish the Anti-Sex Discriminatory Language Act, the Committee on the Judiciary and Criminal Law clarified the relationship between the two bills:

It is true that Bill 1-36 makes substantive changes in the domestic relations law. However, every such change is designed to achieve only one result, i.e., to make the law equal in effect for males and females.... Comprehensive revision of the divorce and marriage laws is contemplated in another Council bill, Bill No. 1-89, the "District of Columbia Uniform Marriage and Divorce act." That bill would make major revisions to local domestic relations law and would do so in a non-sex-discriminatory *317 manner.... [T]he Council should promptly enact Bill 1-36 which would do only one thing—enact the principle of sex equality into the D.C.Code including the domestic relations law. Thereafter, the Council may proceed to consider a more comprehensive revision of the domestic relations law....

COMM. REPORT ON BILL 1-36, at 5-6. This comment clarifies that the Anti-Sex Discriminatory Language Act served a limited purpose: to make the law equal in effect for men and women vis-a-vis each other; for example, it gave mothers a right equal to that of fathers to consent to marriage by a child under 18. See supra note 15. There was not a hint that the legislation was intended to give one class of males, e.g., gay men, an equality with another class of males, e.g., heterosexual men. Thus, the 1976 Act did not revise the substance of the marriage statute to redefine the term "marriage."

G. The 1982 Gender Rule of Construction

Finally, in 1982, the Council adopted a new Gender Rule of Construction, D.C.Law 4-111, § 2(a), 29 D.C.Reg. 1684 (1982), now contained in D.C.Code § 49-203 (1990). Appellants say this Rule conclusively requires interpretation of the marriage statute to authorize same-sex marriages.

This 1982 legislation amended the language of former D.C.Code § 49-203, traceable to the 1901 Code, 31 Stat. 1189, ch. 854, Preamble, § 2 (Second), which had provided: "Words importing the masculine gender shall include all genders, except where such construction would be absurd or unreasonable." The new Rule also amended the language of D.C.Code § 1-230, adopted in 1975, which had read: "For the purposes of any act or resolution of the Council of the District of Columbia, unless specifically provided otherwise—... (3) words importing one gender include and apply to the other gender as well." 22 D.C.Reg. 1990 (1975). In adopting the new Gender Rule of Construction in 1982, therefore, the Council amended two provisions of the D.C.Code (among others not relevant here). It amended § 49-203 to say: "Unless the Council of the District of Columbia specifically provides that this section shall be inapplicable to a particular act or section, all the words thereof importing 1 gender include and apply to the other gender as well." D.C.Code § 49-203 (1990). And it amended § 1-230(3) to read: "With regard to resolutions, words importing 1 gender include and apply to the other gender as well." D.C.Code § 1-230(3) (1992).

Appellants argue that, when the Council removed the words "except where such construction would be absurd or unreasonable" from the 1901 rule of construction in former § 49-203, without adding a disclaimer with respect to the marriage statute, the Council implicitly authorized same-sex marriages. We cannot agree.

The legislative history of the Gender Rule of Construction makes clear that the Council enacted the Rule in 1982 for one purpose: "to create a consistent gender rule of legislative construction throughout the D.C.Code." COMM. ON PUBLIC SERVICES & CONSUMER AFFAIRS, COMMENTS ON BILL 4-374, THE ANTI-SEX DISCRIMINATORY LANGUAGE ACT AMENDMENT ACT OF 1981, at 1 (Feb. 16, 1982). Wilhelmina J. Rolark, then Chair of the Committee on Public Services & Consumer Affairs, elaborated:

Presently, there are two separate, somewhat inconsistent rules for interpreting statutory words having a gender meaning. Section 49-203 of the Code (31 Stat. 1189, enacted March 3, 1901) provides that "words importing the masculine gender shall be held to include all genders, except where such construction would be absurd or unreasonable." By contrast, Code section 1-230 (D.C.Law 1-17, 22 DCR 1990, effective September 23, 1975) provides that for the purposes of any act or resolution of the Council of the District of Columbia, unless specifically provided otherwise "words importing one gender include and apply to the other gender as well."

Id.

Not only was the original (1901) § 49-203 deficient because it did not apply to feminine gender words—e.g., the word "steward" in a statute would be construed to include a "stewardess," whereas the word "stewardess" in a statute would not be construed to apply to a "steward"—but also that former *318 § 49-203 had "different criteria" from those in former § 1-230 "for determining when the gender rule of construction [would] not apply." COMM. ON THE JUDICIARY, REPORT ON BILL No. 4-374, THE ANTI-SEX DISCRIMINATORY LANGUAGE ACT OF 1981, at 3 (Feb. 10, 1982). The Report to the Members of the Committee on the Judiciary explained:

The "absurd or unreasonable" criteria is found in sec. 49-203. In sec. 1-230, the gender rule of that section applies "unless specifically provided otherwise" by the legislature. Section 2(a) of Bill 4-374 [to adopt the gender Rule of Construction] would make the latter criteria the rule for all statutes in the Code.

Id. Thus, the Council merely intended to make D.C.Code § 49-203 (traceable to 1901) consistent with § 1-230 (adopted in 1975) when, in 1982, it adopted the Gender Rule of Construction to remove the words "except where such construction would be absurd or unreasonable" from former § 49-203.

Had the Council intended to enlarge the statutory definition of "marriage" to include same-sex unions, it surely would have mentioned such a significant intention in the legislative history of the statute implementing the new Gender Rule of Construction. See National Org. for Women v. Mutual of Omaha Ins. Co., 531 A.2d 274, 276 (D.C.1987) (if Council intended such a dramatic change in law, "it is reasonable to assume that there would have been at least some specific reference to it in the language of the Act or, at least, within its legislative history") (hereafter NOW). The Council did not do so. The Council merely intended to resolve a conflict between two inconsistent rules of construction on the books. We conclude, accordingly, that the Gender Rule of Construction, D.C.Code § 49-203, § 1-230, does not require recognition of same-sex marriages in the District. See NOW, 531 A.2d at 276.

* * *

The Marriage and Divorce Act of 1977 and the Anti-Sex Discriminatory Language Act of 1976, as well as the Gender Rule of Construction enacted in 1982, did not substantively change the central provisions of the marriage statute enacted and codified in 1901. The commonly understood meaning of "marriage" in those years was limited to a union between a man and a woman. Congress, in enacting and then codifying the marriage statute, used gender-specific language which no later Congress or the Council of the District of Columbia has ever changed. The logically related divorce statute, loaded with even more gender-specific language and codified at the same time as the marriage statute—without subsequent material change affecting the definition of "marriage"—reflects the contemporary, common legislative understanding that marriage requires a man and woman. Indeed, the common societal understanding of marriage, reflected in legal and ordinary dictionary definitions from the last century until today, presupposes a heterosexual union. The courts of other jurisdictions take this view as well.

All this convinces us that no legislature for the District of Columbia—Congress or Council—has ever intended to sanction same-sex marriages. The trial court, therefore, did not err in granting summary judgment for the District on appellants' claim under the marriage statute.

III. THE HUMAN RIGHTS ACT CLAIM

Appellants next contend that, by refusing to issue them a marriage license, the Clerk discriminated against them because of their sex or sexual orientation, in violation of the Human Rights Act, D.C.Code §§ 1-2501 to -2557 (1992). Specifically, appellants argue that when the Marriage License Bureau, a place of public accommodation under the Clerk of the Superior Court, refuses to issue marriage licenses to same-sex couples, gays and lesbians are unlawfully denied an "equal opportunity" to participate in marriage, an important "aspect of life." See D.C.Code §§ 1-2511, -2519, -2532.

Human Rights Act § 1-2519(a)(1) makes it "an unlawful discriminatory practice" for one to "deny, directly or indirectly, any person the full and equal enjoyment of the goods, services, facilities, privileges, advantages, and accommodations of any place of public accommodations," if the denial is "wholly or partially for a discriminatory reason based *319 on the ... sex ... [or] sexual orientation ... of any individual." A "place of public accommodation" is defined in D.C.Code § 1-2502(24) to include "wholesale and retail stores, and establishments dealing with goods or services of any kind," as well as "public halls and public elevators of buildings and structures." Appellants contend that the Marriage License Bureau is a "place of public accommodation" because it is an "establishment dealing with goods or services of any kind" and is located in the Superior Court building—a "public hall."

On appeal, the government assumes for the sake of argument that the Marriage License Bureau is a place of public accommodation. Furthermore, Elizabeth A. Leader and Barbara R. Lewis, District Human Rights Commissioners who, in their individual capacities, filed an amicae curiae brief on behalf of appellants, argue that all District of Columbia agencies are places of public accommodation, within the meaning of the Human Rights Act, because they provide goods and services to District residents. See also In the Matter of Kevin S. Dickerson v. District of Columbia Department of Human Services, District of Columbia Commission on Human Rights, No. 89-465-PA(N), Final Decision and Order (May 23, 1991) (District of Columbia Department of Human Services is place of public accommodation under Human Rights Act because it provides services to District residents). We, too, assume, without formally deciding, that the Marriage License Bureau is a place of public accommodation for purposes of our analysis.

The Council of the District of Columbia enacted the Human Rights Act of 1977 to "underscore the Council's intent that the elimination of discrimination within the District of Columbia should have the highest priority and that the Human Rights Act should therefore be read in harmony with and as supplementing other laws of the District." COMM. ON PUBLIC SERVICES AND CONSUMER AFFAIRS, REPORT ON BILL No. 2-179, THE HUMAN RIGHTS ACT OF 1977, at 3 (July 5, 1977) (citations and internal quotation marks omitted) (hereafter COMM. REPORT ON BILL 2-179). The Council undoubtedly intended the Human Rights Act to be a powerful, flexible, and far-reaching prohibition against discrimination of many kinds, including sex and sexual orientation.

The Council, however, did not intend the Act to prohibit every discriminatory practice. For example, in NOW, 531 A.2d at 277-78, we concluded that the Council did not intend the Human Rights Act to apply to the gender-discriminatory actuarial pricing practices of insurance companies. We explained:

It is true that it can be argued with some persuasion that the "plain language" of the Act prohibits discrimination based on gender in the services offered by insurance companies. See D.C.Code §§ 1-2502(24), -2519 (1987). Significantly, however, the statute contains no language purporting explicitly to regulate insurance premium practices. If the Council had intended to effect such a dramatic change in insurance rate-setting practices, it is reasonable to assume that there would have been at least some specific reference to it in the language of the Act or, at least, within its legislative history.

Id. at 276.

Our analysis in NOW, determining whether the Council intended the Human Rights Act to preclude gender-based rate differentials in the insurance statute, dealt with a much clearer case than the one currently before us. In NOW, although the Committee Report and other legislative history never specifically referred to actuarial rating practices, the Council clearly had heard testimony regarding legislation in other states prohibiting discrimination by insurance companies, when the Council considered the 1973 regulation Governing Human Rights, see 34 DCRR (1973) ("Title 34. Human Rights Law"), which is virtually identical to the present Human Rights Act. See NOW, 531 A.2d at 277. Furthermore, we noted that, throughout the effective period of the Regulation and at the time the Human Rights Act was adopted, the insurance statute expressly allowed a three-year set-back[16] for calculating *320 life insurance premiums for women, but not for men. See id. Finally, we considered the fact that, soon after adopting the Human Rights Act, the Council requested an opinion from the Corporation Counsel as to whether the life insurance set-backs violated the Act. See id. at 278. The Corporation Counsel issued an opinion saying such action would be lawful, and the Council accordingly increased the permissible set-back for women to six years. See id. Under these circumstances, we concluded that the Act did not preclude such differentials. We emphasized: "If the Council had intended to effect such a dramatic change ..., it is reasonable to assume that there would have been at least some specific reference to it in the language of the Act or, at least, within its legislative history." Id. at 276.

Although the Council undoubtedly intended the Human Rights Act to be read broadly to eliminate the many proscribed forms of discrimination in the District, we cannot conclude that the Council ever intended to change the ordinary meaning of the word "marriage" simply by enacting the Human Rights Act. Had the Council intended to effect such a major definitional change, counter to common understanding, we would expect some mention of it in the Human Rights Act or at least in its legislative history. See NOW, 531 A.2d at 276. There is none. See COMM. REPORT ON BILL 2-179. This is not surprising, however, for by legislative definition—as we have seen—"marriage" requires persons of opposite sexes; there cannot be discrimination against a same-sex marriage if, by independent statutory definition extended to the Human Rights Act, there can be no such thing. See Singer, 522 P.2d at 1190-95 (Washington's Equal Rights Amendment does not require the state to authorize same-sex marriage because such relationships are outside definition of marriage).

Furthermore, in 1977, the same Council was considering both the Human Rights Act and the Marriage and Divorce Act legislation. Council members were keenly aware of the gay marriage debate and presumably would have stated their intentions expressly if they had wanted the Human Rights Act, instead of the Marriage and Divorce Act, to expand the marriage statute to authorize same-sex unions. See NOW, 531 A.2d at 277 ("References to other statutes is particularly appropriate when the statutes were enacted by the same legislative body, at the same session."). We therefore cannot conclude that the Council intended the Human Rights Act to change the fundamental definition of marriage. The trial judge properly granted summary judgment for the District on appellants' Human Rights Act claim.

IV. THE CONSTITUTIONAL ISSUES: PROCEDURAL AND ANALYTICAL PREREQUISITES

A. Whether Constitutional Issues Have Been Properly Raised

Because of the way appellants presented their constitutional claims, there is a threshold question whether these claims are properly before this court. The amended complaint refers exclusively to statutory claims under the marriage law and under the Human Rights Act. In appellants' memorandum in support of their motion for summary judgment, however, they argued that the marriage statute "should be read, if it can be, so as to avoid difficult and sensitive constitutional questions." Gay Rights Coalition of Georgetown University Law Center v. Georgetown University, 536 A.2d 1, 16 (D.C. 1987) (en banc) (lead opinion). In other words, appellants argued as a fallback that, even if the trial court rejected their plain language and legislative history arguments, the statute could not survive a limitation to opposite-sex marriages unless that interpretation satisfied the Constitution. Specifically, appellants urged the court to use the approach we employed in Gay Rights Coalition: construe the statute in a way that "saves" its constitutionality—an analysis that inherently requires deciding whether a statute that bars same-sex marriages could withstand due process and equal protection challenges. See id., 536 A.2d at 49 (Ferren, J., concurring in the result in part and dissenting in part) (desire to save statute from constitutional infirmity means "constitutional *321 analysis determines statutory analysis," in contrast with "quite different ... doctrine favoring statutory over constitutional ground for decision when both are independently available").

In its first opinion granting summary judgment for the District, the trial court ignored the constitutional issues plaintiff-appellants had presented. Appellants moved for reconsideration on constitutional grounds, presenting comprehensive arguments, supported by case law and other authorities, explaining why the marriage statute, if upheld, would violate appellants' right to due process and equal protection of the laws. In response, the District contended that, "until the filing of their motion for reconsideration the plaintiffs consistently maintained that their case did not focus upon constitutional issues." The District accused appellants of an "eleventh hour attempt to recast themselves as victims of unconstitutional action." The trial court, however, granted appellants' motion for reconsideration and expressly ruled upon—and rejected—the constitutional claims. The District has not questioned on appeal the propriety of our reaching these issues; and, in any event, given appellants' and the trial court's presentation and resolution, respectively, of the constitutional claims, we conclude they are appropriately before us.

At this point in the constitutional analysis, I write only for myself until Part V., which Judge STEADMAN expressly joins. Neither Judge TERRY nor Judge STEADMAN, however, joins in Part VI.

B. Standard of Review

I turn to our standard of review. This case concerns cross-motions: the District moved to dismiss for failure to state a claim; plaintiff-appellants, presenting affidavits of compliance with all marriage statute application requirements and of rejection by the Clerk of the Superior Court, filed a motion for summary judgment. The trial court granted summary judgment for the District. See Bell v. Jones, 566 A.2d 1059, 1060 (D.C. 1989) (motion for judgment on pleadings requires treatment as motion for summary judgment "when it relies on matters outside the pleadings"); Clay v. Hanson, 536 A.2d 1097, 1100 n. 3 (D.C.1988) (when judge considers materials beyond pleadings, motion to dismiss for failure to state claim is converted into motion for summary judgment).

"A motion for summary judgment shall be granted `if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits ... show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law.'" Kurth v. Dobricky, 487 A.2d 220, 224 (D.C. 1985) (quoting Super.Ct.Civ.R. 56(c)). Courts frequently have emphasized the importance of an adequate record in cases presenting complex constitutional issues, and accordingly they have stressed that, in such cases, summary judgment should be granted sparingly. See, e.g., Felix v. Young, 536 F.2d 1126, 1135 (6th Cir.1976) ("The adequacy of the record is particularly important where the court is called on to decide questions of constitutional law without benefit of a trial."); Waldie v. Schlesinger, 166 U.S.App.D.C. 175, 177, 509 F.2d 508, 510 (1974) ("a full development of the facts of these [equal protection] cases is essential to any meaningful assessment of appellant's claim"); see generally 6 Part II JAMES WILLIAM MOORE, MOORE'S FEDERAL PRACTICE ¶ 56.1710 (1988); 10A CHARLES ALLEN WRIGHT, ARTHUR R. MILLER & MARY KAY KANE, FEDERAL PRACTICE AND PROCEDURE § 2732.2 (1983).

The parties here do not dispute the basic, historical facts giving rise to the claim: appellants applied for a marriage license; if they had been an opposite-sex couple they would have qualified under the marriage statute; the clerk denied their application because they both are men.[17] No party, *322 moreover, has asked for a trial. Thus, the only question the parties present is whether, on the undisputed facts of record, the District or the appellants should be "entitled to a judgment as a matter of law," Super.Ct.Civ.R. 56(c) (1993)—i.e., of constitutional law.

If, as it turns out, neither side is entitled to summary judgment, we must reverse the judgment for the District and remand the case for trial, even though no party has asked for one; each side's request for a summary remedy in its favor does not nullify the right to a trial if summary judgment is denied. See Super.Ct.Civ.R. 56(d) (elaborating procedure when summary judgment is "not rendered upon the whole case or for all the relief asked and a trial is necessary"); Estate of Wells v. Estate of Smith, 576 A.2d 707, 709 & 709 n. 1 (D.C.1990) (ordinarily upon reversal of summary judgment, appellate court remands for trial unless "remand would be futile," e.g., if the "only persons with knowledge of the [relevant] historical facts ... are deceased") (citations omitted).

C. Relevance of the Distinction Between "Adjudicative Facts" and "Legislative Facts"

The fact that this case is presented on appeal purely as a question of law, based on undisputed facts, is troublesome, if not deceptive, because there are two kinds of fact at issue: "adjudicative facts" and "legislative facts."

1. In General

Courts use the term "adjudicative fact" to describe the events which have happened between the parties. See State v. Erickson, 574 P.2d 1, 4 (Alaska 1978) ("Adjudicative facts ... are those facts which explain who did what, when, where, how, and with what motive and intent.") (citation omitted); FED. R.EVID. 201 advisory committee's note (a) (1994) ("Adjudicative facts are simply the facts of the particular case"); see also Lewis v. United States, 408 A.2d 303, 311 n. 11 (D.C.1979) (calling adjudicative facts "legal" facts); DONALD L. HOROWITZ, THE COURTS AND SOCIAL POLICY 45 & n. 58, 275 (1977) (calling adjudicative facts "historical facts").

"Legislative facts," in contrast, are patterns of social, economic, political, or scientific behavior or other data that a court inevitably uses to inform and shape the policy judgments it often has to make in deciding newly-presented questions of law. See Lewis, 408 A.2d at 311 n. 11; Erickson, 574 P.2d at 5; HOROWITZ, supra, at 45, 275; see generally JOHN MONAHAN & LAURENS WALKER. Social Authority: Obtaining, Evaluating, and Establishing Social Science in Law, 134 U.PA.L.REV. 477, 482-84 (1986). Courts, for example, may have to evaluate the impact of a proposed cross-racial adoption on a child's sense of identity, in order to decide the child's best interest, see In re R.M.G., 454 A.2d 776 (D.C.1982), or the court may have to ascertain the accuracy of a particular test for drugs, as it bears on guilt or innocence of a charged drug offense, see Jones v. United States, 548 A.2d 35 (D.C.1988). Such instances require the court to find social or scientific facts that transcend the individual case but determine, sometimes conclusively, how the case shall be decided. While deciding cases day-to-day, therefore, courts cannot help finding, and relying on, legislative facts; that process is inherent in the courts' answering questions of law and thus is not left exclusively to legislatures. See Erickson, 574 P.2d at 5-6; FED.R.EVID. 201 advisory committee's note (a).[18]

*323 As explained later, the question whether the state invidiously discriminates against homosexuals by withholding from same-sex couples the right to marry inevitably presents sub-questions about the nature and causes of homosexuality and, as a result, confronts this court with issues of legislative fact-finding. It is therefore necessary to explore in greater detail the judicial process of finding legislative facts. As discussed more fully below, courts traditionally answer questions of legislative fact, and thus questions of law, not only by referring to evidence of record but also by considering non-record sources such as scientific and social science studies found in law reviews and other journals. See generally Lewis, 408 A.2d at 311 n. 11; Erickson, 574 P.2d at 5; 2 MCCORMICK ON EVIDENCE § 331 (4th ed. 1992), at 398-402; HOROWITZ, supra, at 45, 275. The extent to which courts can properly use non-record sources in this way is a difficult and sometimes controversial subject, as we shall see.

In any event, as a prelude to discussion of the constitutional issues in this case, I believe I not only must elaborate the distinctions between adjudicative and legislative facts, but also must evaluate and apply the respective processes required for each kind of fact-finding—all in aid of deciding whether this case, factually, is ripe for summary judgment or requires a remand for trial.

2. "Adjudicative" and "Legislative" Facts Distinguished

In this case, all the undisputed facts plaintiff-appellants have supported by affidavits, see supra note 17, and presented for summary judgment as a matter of law are "adjudicative" facts: appellants are both men; they are residents of the District; they are not disqualified by any of the enumerated prohibitions under the marriage statute; they applied for a marriage license from the District's Marriage License Bureau, presenting valid blood tests and the name of an authorized person willing to perform the marriage ceremony; the Clerk of the Superior Court denied them a marriage license solely on the ground that the District of Columbia Code does not authorize marriage between persons of the same sex; they would have been issued a marriage license if they were a heterosexual couple; and the denial of a marriage license potentially denies them an extraordinary number of tangible benefits, based upon marital status, enumerated in the District of Columbia Code.[19]

*324 The first question, then, is whether these adjudicative facts are enough to resolve, through summary judgment, a same-sex couple's constitutional challenge to the marriage statute. The most useful way to answer this question is to focus, for purposes of illustration, on equal protection of the laws, for which the ultimate legal question in this case is whether the marriage statute discriminates, without sufficient justification, against members of a constitutionally protected class (allegedly, unrelated adult homosexual couples). Compare Plyler v. Doe, 457 U.S. 202, 223-24, 102 S. Ct. 2382, 2397-98, 72 L. Ed. 2d 786 (1982) (children of illegal aliens comprise protected class) with Massachusetts Bd. of Retirement v. Murgia, 427 U.S. 307, 312, 96 S. Ct. 2562, 2566, 49 L. Ed. 2d 520 (1976) (police officers over age 50 do not comprise protected class).

This statement of the ultimate issue requires the court to frame questions of adjudicative fact in a particular way: Are the plaintiffs actually members of the allegedly protected class? (Yes; they are both homosexual men.) Have they been treated in a discriminatory manner? (Yes; they have been denied a marriage license that would have been issued to an adult heterosexual couple.)[20]

The questions continue: Are members of the allegedly protected class (homosexual couples) entitled to greater constitutional protection against discriminatory treatment than members of other groups? Depending on the level of protection required, has the law unfairly discriminated against members of the allegedly protected class (homosexual couples)? These formulations are, fundamentally, questions of law, not questions of fact, since they focus on legal determinations: "entitlement" to "greater protection," and on "fairness" or "unfairness" of discriminatory treatment. As I have indicated, however, the answers to these questions are determined, ultimately, not only by reference to the relevant adjudicative facts outlined above, but also by consideration of so-called "legislative facts." More specifically, the legal question whether the state, in withholding the marriage statute from same-sex couples, violates their constitutional right to equal protection of the laws may turn to an appreciable extent on whether homosexuality is, to take the extremes, a genetically determined *325 or a learned orientation. It is therefore critical to understand how such questions are, if at all, to be answered. This requires in-depth understanding of the court's legislative fact-finding process.

Legislative facts—for example, the sub-category "social facts"—often are the critical facts used for answering major questions of constitutional law. See, e.g., Brown v. Board of Education, 347 U.S. 483, 494, 74 S. Ct. 686, 691-92, 98 L. Ed. 873 (1954) (demonstrable social fact that racial segregation of public schools "generates a feeling of inferiority as to [Negro children's] status in the community that may affect their hearts and minds in a way unlikely ever to be undone"); In re R.M.G., 454 A.2d at 787-88 (in "significant number of instances" persons responsible for adoption decisions "will not be able to focus adequately on an adoptive child's sense of identity, and thus on the child's best interest, without considering race") (opinion of Ferren, J.); id., 454 A.2d at 802 (Newman, J., dissenting) ("A degree of race-consciousness is permissible to the achievement of [child's best] interest, because certain potential future hardships to the child arise when the parents are of a different race.").

Similarly, legislative facts can be outcome-determinative "political" or "economic" facts. See, e.g., Baker v. Carr, 369 U.S. 186, 82 S. Ct. 691, 7 L. Ed. 2d 663 (1962) (recognition, based on political facts, of political disincentives to legislative cure for malapportionment of legislative representation); SEC v. Capital Gains Research Bureau, Inc., 300 F.2d 745, 750-51 (2d Cir.1961) (judicial notice of economic facts tending to prove advice tendered by small advisory service could not influence stock market), rev'd, 375 U.S. 180, 84 S. Ct. 275, 11 L. Ed. 2d 237 (1963) (judicial notice of economic facts tending to prove the contrary).

Legislative facts also include "scientific facts," providing answers, for example, to the question whether "`DNA' profiling evidence is admissible to corroborate the identification of a defendant in a criminal case," United States v. Porter, 618 A.2d 629, 630 (D.C. 1992), or to the question whether drug-testing with the so-called EMIT system has general acceptance in the scientific community, see Jones, 548 A.2d at 39-47, or to the question whether cocaine could properly be characterized as a "narcotic" under a criminal statute, requiring inquiry into both pharma-cological properties and psychological impact, see Erickson, 574 P.2d at 4-10.

The Supreme Court of Alaska has nicely summarized the uses courts make of legislative facts:

Legislative facts come into play when the court is faced with the task of deciding the constitutionality of a statute, statutory interpretation or the extension or restriction of a common law rule upon grounds of policy. These policy decisions, as in the case at hand, often hinge on social, political, economic, or scientific facts, most of which no longer fall within the classification of irrefutable. Cases involving such decisions cannot be decided adequately without some view by the court of the policy considerations and background upon which the validity of a particular statute or rule is grounded.
MCCORMICK ON EVIDENCE, § 334 (2d ed. 1972), alludes to the proposition that the topic of judicial notice, particularly as to legislative facts, does not conveniently fit within the structured confines of the law of evidence, but rather is more appropriately categorized in the more general area of judicial reasoning. A distinction must be made between evidence of the particular facts of a case, which can be accepted only through prescribed methods calculated to assure credibility, and those [legislative] facts which are of greater policy significance in that they describe aspects of our larger environment and form the basis upon which adjudicative facts are evaluated. It is a slippery distinction at best; but it is one that has been drawn by [Professor Kenneth Culp] Davis, incorporated into the federal rules, and at least implicitly recognized in nearly every situation where a court has been called upon to address a question of policy in evaluating the rationality or reason behind a statute or rule.

Erickson, 574 P.2d at 5-6 (footnotes omitted) (emphasis added).

The Federal Rules of Evidence also stress the distinction between adjudicative and legislative *326 facts. The Advisory Committee Note points out that FED.R.EVID. 201 is limited to "judicial notice of adjudicative facts," and that no federal rule of evidence "deals with judicial notice of legislative facts." FED. R.EVID. 201 advisory committee's note (a). Far from disparaging court involvement in legislative fact-finding, however, the Advisory Committee simply notes that only adjudicative facts can meet the "high degree of indisputability [that] is the essential prerequisite" to formally taking "judicial notice." The Committee then quotes from Professor Davis, see supra note 18, and others, who encourage judges to make their own findings of legislative fact, not only from the trial record but also from sources outside the record as needed, to inform their judicial reasoning—even though admittedly such facts, when found, will not be "indisputable." Under this view, legislative fact-finding is akin to legal research, except that the judge is free to consult non-legal sources as well (sometimes found in law reviews featuring interdisciplinary articles). Seen in this way, a judge who studies journals of social science, for example, in aid of constitutional rulings is, perhaps, better characterized as engaging in the process of answering questions of law than of taking judicial notice of legislative facts. See 2 MCCORMICK, supra, § 331.

It may appear that, through legislative fact-finding, the trial or appellate judge is on a personal frolic, divorced too much from the parties' presentations, but that is an incorrect perception, and it would be wrong to say the traditional approach to legislative fact-finding is inappropriate or altogether unsatisfactory. The real concern is not whether judges should engage in legislative fact-finding; they inevitably do. No one can persuasively argue that the courts have no business deciding, for example, whether a particular drug fits the statutory definition of a "narcotic," see Erickson, 574 P.2d at 18-23, or whether race affects a child's sense of identity, for purposes of determining whether a cross-racial adoption will be in the child's best interest, see In re R.M.G., 454 A.2d at 791-94. Courts rarely can avoid such issues. The concern, rather, is that certain kinds of legislative facts may be too difficult to ascertain, or simply may be too controversial, for a court—rather than the legislature itself—to decide. I will return to this "difficulty" or "controversy" theme later; the point here is to make clear that there is quite a difference between saying a court should not engage in any legislative fact-finding and saying a court should stay away from particular legislative fact-finding. The former would be ill-advised and often impossible; the latter, on occasion, may be prudent.

3. Judicial Process of Legislative Fact-Finding

Overt legislative fact-finding is traceable to Muller v. Oregon, 208 U.S. 412, 419-20, 28 S. Ct. 324, 325-26, 52 L. Ed. 551 (1908), where the Supreme Court upheld the constitutionality of an Oregon law limiting women's factory work to ten hours a day after relying on amicus attorney Louis D. Brandeis's brief, which had presented 90 reports of committees, bureaus of statistics, commissioners of hygiene, and inspectors of factories detailing how long working hours, under the circumstances, were dangerous to women. Justice Holmes later summarized the principle: "A judge sitting with a jury is not competent to decide issues of fact; but matters of fact that are merely premises to a rule of law he may decide." Prentis v. Atlantic Coast Line Co., 211 U.S. 210, 227, 29 S. Ct. 67, 70, 53 L. Ed. 150 (1908).[21]

*327 Accordingly, whereas adjudicative facts are found exclusively on the basis of trial testimony by so-called fact witnesses and of properly verified documentary evidence, legislative facts are often found differently. The "process set up to establish the one is not necessarily adequate to ascertain the other." HOROWITZ, supra, at 45. Legislative facts commonly are found after hearing testimony, by qualified experts, but that testimony is usually buttressed by judicial attention not only to other court decisions but also to scientific or social science literature, including books, treatises, law reviews, and other journals containing useful information not of record.[22] Legislative facts also sometimes are found exclusively in non-record sources,[23] such as a party's "Brandeis brief" and the judge's own research, without help from expert testimony. See e.g., Brown, 347 U.S. at 489-94, 74 S. Ct. at 688-91; McLean v. Arkansas, 211 U.S. 539, 549-50, 29 S. Ct. 206, 208-09, 53 L. Ed. 315 (1909); Muller, 208 U.S. at 419-20, 28 S. Ct. at 325-26; Ribnik v. McBride, 277 U.S. 350, 363-72, 48 S. Ct. 545, 548-52, 72 L. Ed. 913 (1928) (Stone, J., dissenting).

The idea of gathering non-record information appears antithetical to the traditional understanding of sound judicial fact-finding; but, as we have seen with respect to legislative (though not adjudicative) facts, it is done every day as judges—especially appellate judges—search for authority to resolve difficult legal issues requiring policy resolution. Cross-examination of such non-record material, of course, is missing, except insofar as author contests author and the judge applies his or her own critical faculties to the debate. All the data is then tested against legal norms.

In short, the appellate judge tests the parties' proffers of legislative fact against reasoned, published authorities who have attempted to seek the truth through documented research, often critical of earlier efforts. This is the process that Justice Holmes called merely identifying premises for adopting a rule of law, see Prentis, 211 U.S. at 227, 29 S. Ct. at 69-70, and that Professors Morgan, Davis, and others have characterized as part of judicial reasoning and law-making rather than of traditional fact-finding. See, e.g., E.M. Morgan, Judicial Notice, 57 HARV. L.REV. 269, 270-71 (1944); K.C. Davis, An Approach to Problems of Evidence in the Administrative Process, 55 HARV.L.REV. 364, 403 (1942); J.B. THAYER, PRELIMINARY TREATISE ON EVIDENCE 279-80 (1898); Erickson, 574 P.2d at 5-6.

How can legislative fact-finding best be conducted? Ideally, several qualified experts on all sides of a social or scientific issue would appear in the trial court, subject to cross-examination not only about their own views but also about the views of other testifying experts, as well as the views of non-testifying authors of seminal articles on the subject. Indeed, ideally the very authors of the most probative studies would be among the experts who testify. In this way, highly qualified expert testimony would serve at least three important functions: focus the court's attention on the most relevant concerns, present the range of informed opinion on the subject, and both identify and critique the most probative literature. The court, *328 therefore, would achieve a sharpened, presumably reliable insight into complicated matters that, without such help, would be much more difficult for the judge to understand.

The dollar costs and time required for the ideal approach may be prohibitive, however, with the result that a typical hearing will involve only a few experts, often extreme proponents for the parties' respective views, who are likely not to be particularly well qualified and may not provide sufficient, let alone trustworthy, data permitting the court to rule with confidence. This court accordingly has stressed—for example, in connection with expert testimony on the question whether there is a consensus in the scientific community on the accuracy of EMIT drug testing—that non-record sources of information may be crucial to sound legislative fact-finding; "reference to such outside sources may be useful in exposing a proffered expert's bias or incompetence." Jones, 548 A.2d at 42.

Professor Donald L. Horowitz, who has substantial reservations about the capacity of courts to make social policy, is skeptical in part because he worries that the experts selected to testify as to legislative facts too often will lack sufficient qualifications, and that evidentiary rules—in particular FED. R.EVID. 803—will keep out of the trial record seminal studies the courts should know about. He therefore advocates rules changes

to admit books and articles on matters of social fact directly into evidence as exhibits, not require as a precondition that an expert refer to them in his [or her] testimony,[[24]] abolish the favored position of government reports,[[25]] and permit counsel to attack the reliability of the studies directly. This is no panacea, but it would bring the judge one step closer to the original materials, permit him [or her] more easily to check the statements of advocates and interpreters, and—since the studies will be more readily accessible to the judge—perhaps encourage expert witnesses to gear their presentations more closely to what the studies do and do not in fact show. So far, expert witnesses have had too much latitude to parade their own preferences as science.

HOROWITZ, supra, at 281. In short, Professor Horowitz has greater confidence in legislative fact-finding directly by judges than he does in building a trial record with expert testimony—unless, presumably, the judge can supplement the record by any means necessary to reach critical sources that will help the judge test the experts when counsel have not effectively completed the job. If, as others have said, trial and appellate judges, as fact-finders, are presently allowed unlimited access to non-record sources of legislative fact, then of course Professor Horowitz's concerns about the rules of evidence have become substantially moot.

Based on the foregoing considerations, it is questionable whether a hearing with expert testimony about issues of legislative fact "would reveal more reliable or higher quality information than is available by referring to authorities submitted in briefs by both sides, and, in appropriate cases, by additional research at the appellate level." Erickson, 574 P.2d at 6. The advantage of such a costly exercise is likely, on too many occasions, to be marginal at best, and the further away the hearing is from the ideal model I have posited (using the most highly qualified experts), the more the judge will confront distorted, perhaps even biased, presentations and thus have to rely primarily on many non-record sources for critical scientific or social science information. See generally Erickson, 574 P.2d at 4-7.

It is clear, of course, that a trial judge, or an appellate court reviewing a dismissal or summary judgment order, has discretion to order a hearing to help establish legislative facts. For example, in Chastelton Corp. v. *329 Sinclair, 264 U.S. 543, 44 S. Ct. 405, 68 L. Ed. 841 (1924), when a landlord claimed that an emergency justifying rent control was at an end and that continuing such regulation was unconstitutionally confiscatory, Justice Holmes opined:

It is conceivable that, as is shown in an affidavit attached to the bill, extensive activity in building has added to the ease of finding an abode. If about all that remains of war conditions is the increased cost of living, that is not in itself a justification of the act. Without going beyond the limits of judicial knowledge, we can say at least that the plaintiffs' allegations cannot be declared offhand to be unmaintainable, and that it is not impossible that a full development of the facts will show them to be true. In that case the operation of the statute would be at an end.
We need not enquire how far this Court might go in deciding the question for itself, on the principles explained in Prentis v. Atlantic Coast Line Co., 211 U.S. 210, 227 [29 S. Ct. 67, 69-70, 53 L. Ed. 150]. See Gardner v. Collector, 6 Wall. 499 [18 L. Ed. 890]. South Ottawa v. Perkins, 94 U.S. 260 [24 L. Ed. 154]. Jones v. United States, 137 U.S. 202 [11 S. Ct. 80, 34 L. Ed. 691]. Travis v. Yale & Towne Manufacturing Co., 252 U.S. 60, 80 [40 S. Ct. 228, 232, 64 L. Ed. 460]. These cases show that the Court may ascertain as it sees fit any fact that is merely a ground for laying down a rule of law, and if the question were only whether the statute is in force today, upon the facts that we judicially know we should be compelled to say that the law has ceased to operate. Here however it is material to know the condition of Washington at different dates in the past. Obviously the facts should be accurately ascertained and carefully weighed, and this can be done more conveniently in the Supreme Court of the District [a trial court] than here. The evidence should be preserved so that if necessary it can be considered by this Court.

Id. at 548-49, 44 S. Ct. at 406 (emphasis added). Justice Holmes, though mindful of the appellate court's authority to find its own legislative facts, concluded that, for the particular kind of facts at issue—conditions of the District of Columbia housing market over time—it would be more "convenient" for the trial court to do the work.

In every case, therefore, an appellate court has discretion, based on the nature of the inquiry and the trial court record, to determine how to decide questions of law when the record itself does not supply testimony or other documentation adequate for finding essential legislative facts. Certainly, the appellate court can remand for further proceedings as needed, see Chastelton Corp., 264 U.S. at 548-49, 44 S. Ct. at 406-07, but the court also can say, without fear of jurisprudential heresy, that the data proffered by the parties, as well as by the trial court, from non-record, uncross-examined sources, when supplemented by the appellate court's own like research, will suffice in a particular case for constitutional decision-making.

This prescribed free rein for judges to consult non-record sources for legislative facts to inform their policy judgments is not without limit, however.

While not necessarily undisputably true, it would appear that these legislative facts must at least appear to be more likely than not true if the opinion is going to have the requisite intellectual legitimacy upon which the authority of judge-made rules is ultimately founded.

2 MCCORMICK, supra, § 331, at 400 (emphasis added). Accordingly, despite the appropriateness, indeed the desirability if not necessity, of judges' conducting their own research in non-record, even non-proffered, sources, that exercise must reflect intellectual integrity such that the result, if not irrefutable, at least is unquestionably principled.

The point not to be forgotten, however, as Professor Horowitz has emphasized, is that the likelihood of expert testimony of record being superior to the court's own resourcefulness in finding legislative facts is often highly questionable. See HOROWITZ, supra, at 281. The quality of the paid experts too often will be limited to persons who have not done the most careful studies themselves and, in any event, may have an ax to grind that, absent very skillful cross-examination, can create a record that hides the real truth. See id. It *330 may be more "convenient" for an appellate court to remand for relatively easy legislative fact-finding, as in Chastelton Corp., where collection and presentation of historical economic data (housing market conditions) is likely to be straightforward, akin to adjudicative fact-finding. Arguably, the more complex the issue—for example, the nature and causes of homosexuality—the greater the risk that a hearing will yield no better, and perhaps less satisfactory, results than nonrecord sources, including a judge's own research into primary data, helped along, of course, by counsel's advocacy. See HOROWITZ, supra, at 281.

Whether the process I have been discussing is called legislative fact-finding or answering questions of law, the exercise is all the more difficult because legislative facts not only are rarely "indisputable" or "irrefutable" but also commonly change from time to time, whether they are "social," "political," "economic," or "scientific" facts. In contrast with adjudicative facts, which typically are static, see Lewis, 408 A.2d at 311 n. 11, legislative facts are "not necessarily immutable" because they typically involve patterns of behavior—or understandings of patterns of behavior—that can change over time. See HOROWITZ, supra, at 275.

Patterns can and sometimes do change, especially at the lower levels of analysis at which verifiable propositions are likely to be found. If law is to follow behavior, it must constantly monitor such changes—a most difficult undertaking.

Id. (quoted in Lewis, 408 A.2d at 311 n. 11).

The difficulties of legislative fact-finding— especially its differences from traditional, record-based adjudicative fact-finding—cause some lawyers and judges to say that, at least in difficult, controversial cases, legislative fact-finding should be left to the legislature. Perhaps they say this out of a belief that legislatures are better equipped to find such facts, since everyone even marginally interested in a matter has standing to be heard and, presumably, helps round out the record. But saying that proves too much, for the legislative process, virtually by definition, usually plays to majorities, and legislators can use any variety of protest or inconclusiveness of fact to ignore or defer attention to serious, often worthy claims of constitutional rights advanced by victimized minorities.

The truth is, a legislative hearing is not necessarily better suited than a court hearing is to making supportable findings, for example, about the nature and causes of homosexuality, including its degree of immutability. In fact, there is no guarantee that a legislature would ever find reason—or political courage—to schedule such a hearing. In contrast with legislatures, however, the courts are commissioned, among other things, to assure constitutional due process and equal protection of the laws for minorities, without fear of electoral consequences— especially in jurisdictions, such as ours, where judicial selection is immune from popular vote. Difficult as it may be to determine legislative facts for making social and legal judgments about the constitutional rights of homosexuals, the courts have been asked to do so, they are obligated to do so, and they are as equipped as any institution to do so.

* * * * * *

In sum, the question whether appellants' constitutional rights to due process and equal protection of the laws, while not presenting any genuine issue of material "adjudicative" fact in this case, requires "a most difficult undertaking," id.: ascertainment and application of "legislative" facts that include, as we shall see, some findings about the origins of homosexuality and the extent to which sexual orientation is immutable. For that fact-finding (or call it answering questions of law), both parties have relied exclusively on case law and on non-record sources: law reviews, scientific articles, social science compilations. No expert testimony has been proffered. Whether this particular invitation for judicial findings of legislative fact provides enough material, along with our own research, for this court to rule definitively here is a critical, very difficult question. I defer the answer until I explore the parties' presentations. In doing so, I will rely not only on case law but also on scientific and social science sources proffered by the parties—and found on my own—of the sort *331 judges traditionally rely on to decide constitutional questions of the magnitude presented.

V. CONSTITUTIONAL DUE PROCESS: IS SAME-SEX MARRIAGE A "FUNDAMENTAL RIGHT"?

Appellants contend that interpreting the marriage statute in a manner that denies same-sex couples the opportunity to marry violates their constitutional rights. Specifically, they argue, first, that even if marriage is traditionally understood and statutorily defined to include only opposite-sex couples, this limitation of the right to marry unconstitutionally burdens gays' and lesbians' "fundamental right" to marry as they choose—a right protected by the due process clause of the Fifth Amendment. I speak now for the division majority; we conclude that same-sex marriage is not a "fundamental right" protected by the due process clause, because that kind of relationship is not "deeply rooted in this Nation's history and tradition." Moore v. City of East Cleveland, 431 U.S. 494, 503, 97 S. Ct. 1932, 1938, 52 L. Ed. 2d 531 (1977).

A. Definition of "Fundamental Right"

The Supreme Court initially characterized as fundamental rights, entitled to heightened judicial protection under the due process clause, those privileges and immunities that belong to someone as a citizen of the United States—and thus cannot be denied by the states—because they are "implicit in the concept of ordered liberty." Palco v. Connecticut, 302 U.S. 319, 325, 58 S. Ct. 149, 152, 82 L. Ed. 288 (1937) (Fifth Amendment double jeopardy clause did not bar state's appealing second degree murder conviction for alleged evidentiary and instructional errors and, upon obtaining reversal, retrying defendant for first degree murder) (emphasis added). In rejecting the argument that the Fourteenth Amendment incorporates and forbids whatever the Fifth Amendment forbids, Palko announced the so-called selective incorporation approach. The Court characterized the fundamental rights that are so incorporated, and thus binding on the states, as those having their "source in the belief that neither liberty nor justice would exist if they were sacrificed." Id. at 326, 58 S. Ct. at 152.[26]

Over thirty years later, the Court held that "the Fourteenth Amendment guarantees a right of jury trial in all criminal cases which—were they to be tried in a federal court—would come within the Sixth Amendment's guarantee." Duncan v. Louisiana, 391 U.S. 145, 149, 88 S. Ct. 1444, 1447, 20 L. Ed. 2d 491 (1968). The Court premised its ruling on a belief "that trial by jury in criminal cases is fundamental to the American scheme of justice," id. (emphasis added)—a fundamental-right formulation the Court expressly recognized as somewhat different from Palko's. See Duncan, 391 U.S. at 149 n. 14, 88 S. Ct. at 1447-48 n. 14.[27]

Finally, in Moore v. City of East Cleveland (plurality opinion), the Court overturned a conviction under a housing ordinance that denied certain family members the right to live together. The Court premised its decision on the due process clause, which "protects the sanctity of the family precisely because the institution of the family is deeply rooted in this Nation's history and tradition." 431 U.S. at 503, 97 S. Ct. at 1938 (footnote omitted) (emphasis added).[28]

*332 This brings us to Bowers v. Hardwick, 478 U.S. 186, 106 S. Ct. 2841, 92 L. Ed. 2d 140 (1986), which presented the question "whether the Federal Constitution confers a fundamental right upon homosexuals to engage in [consensual] 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." Id. at 190, 106 S. Ct. at 2843. The Court answered "No," expressly reserving the question whether the Georgia statute criminalizing sodomy is constitutional as applied to consensual heterosexual sodomy. See id. at 188 n. 2, 106 S. Ct. at 2842 n. 2 ("We express no opinion on the constitutionality of the Georgia statute as applied to other acts of sodomy."). Writing for the majority, Justice White quoted both Palko and Moore (reflecting the range of "fundamental right" formulations under the due process clause) and concluded that "neither of these formulations would extend a fundamental right to homosexuals to engage in acts of consensual sodomy." Id. at 192, 106 S. Ct. at 2845.

As in Hardwick, we need not resolve where, along the continuum between Palko and Moore, the correct formulation falls. As elaborated below, we conclude that even under Moore's most inclusive definition of "fundamental right," appellants' due process claim fails.

B. Appellants' Due Process Claim

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 [and women]." Loving v. Virginia, 388 U.S. 1, 12, 87 S. Ct. 1817, 1824, 18 L. Ed. 2d 1010 (1967). More specifically, the Supreme Court has emphasized that "the right to marry is part of the fundamental `right of privacy' implicit in the Fourteenth Amendment's Due Process Clause." Zablocki v. Redhail, 434 U.S. 374, 384, 98 S. Ct. 673, 680, 54 L. Ed. 2d 618 (1978). In addressing a Wisconsin law that prohibited a couple from marrying if one of them had an outstanding child support obligation, the Court in Zablocki explained: "When a statutory classification significantly interferes with the exercise of a fundamental right, it cannot be upheld unless it is supported by sufficiently important state interests and is closely tailored to effectuate only those interests." Id. at 388, 98 S. Ct. at 682. The Court then concluded that an unpaid child support obligation did not give the state a sufficiently compelling reason to deny the couple the fundamental right to marry, and thus the Court declared the law invalid as an unconstitutional denial of equal protection of the laws. See id. at 388-91, 98 S. Ct. at 682-84.

An historical survey of Supreme Court cases concerning the fundamental right to marry, however, demonstrates that the Court has called this right "fundamental" because of its link to procreation. See Baehr v. Lewin, 74 Haw. 530, 852 P.2d 44, 55 (1993). The Court first discussed marriage as a fundamental right in Skinner v. Oklahoma, 316 U.S. 535, 62 S. Ct. 1110, 86 L. Ed. 1655 (1942), a case striking an Oklahoma statute that allowed the state to sterilize habitual criminals without their consent. In explicating *333 the rationale for decision, the Court stressed that "[m]arriage and procreation are fundamental to the very existence and survival of the race." Skinner, 316 U.S. at 541, 62 S. Ct. at 1113. In Zablocki, moreover, the Court explained:

It is not surprising that the decision to marry has been placed on the same level of importance as decisions relating to procreation, childbirth, child rearing, and family relationships. As the facts of this case illustrate, it would make little sense to recognize a right of privacy with respect to other matters of family life and not with respect to the decision to enter the relationship that is the foundation of the family in our society.... [I]f appellee's right to procreate means anything at all, it must imply some right to enter the only relationship in which the State of Wisconsin allows sexual relations legally to take place.

434 U.S. at 386, 98 S. Ct. at 681 (footnote omitted). See also Loving, 388 U.S. at 12, 87 S. Ct. at 1824 ("Marriage is one of the `basic civil rights of man,' fundamental to our very existence and survival.") (quoting Skinner, 316 U.S. at 541, 62 S. Ct. at 1113).

Although we recognize that gay and lesbian couples can and do have children through adoption, surrogacy, and artificial insemination, see Developments in the Law—Sexual Orientation and the Law, 102 HARV.L.REV. 1508, 1642-60 (1989) (hereafter Sexual Orientation and the Law), and that not all heterosexual married couples are able, or choose, to procreate, we cannot overlook the fact that the Supreme Court has deemed marriage a fundamental right substantially because of its relationship to procreation. Thus, in recognizing a fundamental right to marry, the Court has only contemplated marriages between persons of opposite sexes— persons who had the possibility of having children with each other. See Baehr, 852 P.2d at 56; see generally Note, Homosexuals' Right to Marry: A Constitutional Test and a Legislative Solution, 128 U.PA.L.REV. 193, 200-02 (1979) (hereafter Homosexuals' Right to Marry).

The question, then, is whether there is a constitutional basis under the due process clause for saying that this recognized, fundamental right of heterosexual couples to marry also extends to gay and lesbian couples. The answer, very simply, is "No." Even without reference to Hardwick's constitutional approval of statutes criminalizing consensual sodomy, we cannot say that same-sex marriage "is deeply rooted in this Nation's history and tradition." Moore, 431 U.S. at 503, 97 S. Ct. at 1938. Indeed, the District of Columbia marriage statute reflects an altogether different tradition. Accordingly, same-sex marriage cannot be called a fundamental right protected by the due process clause.

VI. EQUAL PROTECTION: ARE HOMOSEXUALS A "SUSPECT" OR "QUASI-SUSPECT" CLASS?

A. The Trial Court's Ruling

I write, once again, only for myself, and thus, in the balance of this opinion, respectfully dissent from the judgment to affirm.

Appellants also maintain that the statute limiting marriage to opposite-sex couples unconstitutionally discriminates against them as a gay couple, in violation of their Fifth Amendment right to equal protection of the laws.[29] The trial court rejected this contention in a Supplemental Memorandum Opinion and Order of June 2, 1992. The court concluded that homosexuals comprise neither a "suspect" class mandating "strict scrutiny" of the statutory bar against same-sex marriage, nor a "quasi-suspect" class requiring "intermediate scrutiny" of the marriage barrier. See Plyler v. Doe, 457 U.S. at 216 & n. 14, 217-18, 102 S. Ct. at 2394-95 & n. 14, 2395. The court accordingly held that the "rational basis" test applied and that the statute limiting marriage to heterosexuals is "rationally related" to three "legitimate state interest[s]":

*334 [1] fostering at a socially-approved point in time (i.e., during marriage), that which is essential to the very survival of the human race, namely, procreation.... [2] [prohibiting] ... the sexual conduct, to wit, sodomy, commonly associated with homosexual status—conduct deemed by society to be so morally reprehensible as to be a criminal offense in the District of Columbia and many other jurisdictions.[[30]] [3] Finally, the legislature could also rationally conclude that such authorization would constitute unprecedented and unwarranted "social tinkering" with one of the most sacred institutions known to mankind, namely, marriage.... [Footnotes omitted.]

The court then asserted in a footnote:

Indeed, the result would not be different even if plaintiffs were members of a "suspect" or "quasi-suspect" class. As previously noted, the state's interests and concerns are not only legitimate but compelling. And realistically there is no less restrictive means of adequately addressing them than by simply prohibiting the proposed union.

B. Introduction: Discrimination and Equal Protection of the Laws

The fact that same-sex marriage is not a fundamental right, entitled to due process protection, see supra Part V., does not end the constitutional inquiry; equal protection analysis is available to determine whether the classification at issue, unrelated adult homosexual couples,[31] can serve as a legitimate basis for excluding persons from a state benefit—in this case the right to marry— which is available virtually without limitation to unrelated adult heterosexual couples. The Supreme Court has explained that certain state-imposed limitations or exclusions can violate the equal protection guarantee either (1) by restricting the exercise of a fundamental right, see Skinner, 316 U.S. at 541, 62 S. Ct. at 1113—not the case here[32]—or (2) by discriminating, without sufficient justification, against members of a constitutionally protected class. See Plyler, 457 U.S. at 216-18, 102 S. Ct. at 2394-95; Massachusetts Bd. of Retirement v. Murgia, 427 U.S. at 312, 96 S. Ct. at 2566. This case concerns the latter, "discrimination" category.

It is important to say at the outset of this discussion what will become clearer as I proceed: the question I am addressing—whether homosexuals, see supra note 31, comprise a constitutionally protected class—must be answered for equal protection purposes generally, not just for a marriage case. That is to say, the answer to this "classification" question will be the same whether the issue in a particular case is alleged discrimination against homosexuals in employment, or in housing, or under the marriage statute; homosexuals either will, or will not, comprise a class entitled to special constitutional scrutiny of alleged discrimination against them.

What will not be the same from case to case, however, is the ultimate outcome resulting *335 from that classification. Suppose, for example, this court were to conclude here, or in another case, that homosexuals comprise a constitutionally protected class. This would do no more than shift to the government in a particular case the burden of proving that the alleged discrimination against homosexuals served a compelling, or at least an important, governmental interest. See Plyler, 457 U.S. at 216-18, 102 S. Ct. at 2394-95. Thus, classification of homosexuals as a constitutionally protected class would not grant them the right to marry one another. It is theoretically possible that the government would fail to carry its burden in a housing or employment, discrimination case but might succeed, nonetheless, in demonstrating why, in the public interest, only heterosexuals should be allowed to marry.

The point is: in the ensuing discussion of whether homosexuals comprise a constitutionally protected class—just as racial minorities and women are specially protected classes—I shall be dealing with a threshold inquiry that does not necessarily dictate the result of the case. Basically, I shall be dealing with the question (1) whether the plaintiffs-appellants should have the burden of showing why there is no rational basis for disallowing homosexuals to marry one another, or (2) whether instead the District should have the burden of showing a compelling, or at least substantial, governmental interest in preventing homosexuals from marrying one another.

It also is important to be clear from the beginning that, even though the state does not withhold a right deemed "fundamental" for constitutional purposes, a legislative classification that withholds other significant rights and benefits from a protected class of persons, while making those benefits available to others, can just as surely violate the equal protection clause. See, e.g., Plyler, 457 U.S. at 216-18, 102 S. Ct. at 2394-95 (Texas statute withholding from local school districts state funds for educating children who were not "legally admitted" violates equal protection clause). Thus, even though only heterosexual couples have a fundamental right to marry, the aspect of marriage that elevates it to a "fundamental" right under the due process clause—the capacity to have children together—does not gainsay the fact that marriage has other important attributes which, as the Supreme Court itself has recognized, can be significant enough on occasion to outweigh various interests the state may have in withholding the right to marry from one group or another.

C. The Attributes of Marriage Justifying an Equal Protection Inquiry

In Turner v. Safley, 482 U.S. 78, 107 S. Ct. 2254, 96 L. Ed. 2d 64 (1987), the Court held that a state law withholding from prisoners the right to marry violated due process because four "important attributes of marriage," id. at 95, 107 S. Ct. at 2265, outweighed any penological concern the state could articulate to support the marriage ban. In particular, the Court said:

The right to marry, like many other rights, is subject to substantial restrictions as a result of incarceration. Many important attributes of marriage remain, however, after taking into account the limitations imposed by prison life. [1] First, inmate marriages, like others, are expressions of emotional support and public commitment. These elements are an important and significant aspect of the marital relationship. [2] In addition, many religions recognize marriage as having spiritual significance; for some inmates and their spouses, therefore, the commitment of marriage may be an exercise of religious faith as well as an expression of personal dedication. [3] Third, most inmates eventually will be released by parole or commutation, and therefore most inmate marriages are formed in the expectation that they ultimately will be fully consummated. [4] Finally, marital status often is a precondition to the receipt of government benefits (e.g., Social Security benefits), property rights (e.g., tenancy by the entirety, inheritance rights), and other, less tangible benefits (e.g., legitimation of children born out of wedlock). These incidents of marriage, like the religious and personal aspects of the marriage commitment, are unaffected *336 by the fact of confinement or the pursuit of legitimate corrections goals.

Id. at 95-96, 107 S. Ct. at 2265.

If these attributes of marriage are relevant to the needs and aspirations of gays and lesbians—as public education (not a fundamental right) was relevant to alien children (a quasi-suspect class) in Plyler, 457 U.S. at 217-18 & n. 16, 223-24, 230, 102 S. Ct. at 2395 & n. 16, 2397-98, 2401-02—we have the basis for inquiring whether a marriage statute that excludes homosexuals from the right to marry one another meets equal protection requirements.

Appellants proffer that, given the nature of homosexuality, Turner's attributes of marriage—emotional support, religious or spiritual significance, physical consummation, and government and other benefits—are as relevant and important to same-sex couples as to heterosexual couples. I perceive no basis for doubting that appellants can make such a showing.[33] Moreover, appellants buttress their argument by noting that many heterosexual couples are not able to have children, or may choose not to do so, whereas homosexual couples, absent state law or policy impediments,[34] can and do elect parenthood through adoption, surrogacy, or artificial insemination—the result being that parenthood, and even the benefits of procreation, are not necessarily limited to formally united heterosexual couples.

After considering pertinent legislative facts and applying relevant case law, I must conclude—as elaborated later—that the trial court erred in deciding, as a matter of law, that homosexuals do not comprise a "suspect" or "quasi-suspect" class, and thus the court erred in concluding as a matter of law that the rational basis test applies to this case under the equal protection clause. On the other hand, I am unable to ascertain to the required degree of certainty from the record, supplemented by my own study, the legislative facts necessary for deciding as a matter of law whether homosexuals are entitled under the equal protection clause to special scrutiny of their claimed right to marry. Thus, as explained below, I believe a trial will be required to decide the classification issue: whether the rational basis test, or a higher form of scrutiny, applies.

D. Summary Judgment for Appellants Inappropriate Assuming, for the Sake of Argument, That the Rational Basis Test Applies

I do not believe this court can resolve the matter in appellants' favor by assuming, for the sake of argument, that the rational basis test applies. Although I do not subscribe to the trial court's reasoning, see supra Part VI.A., I also cannot say as a matter of law that the limitation of marriage to heterosexual couples would not survive the traditional "rational basis" test under the equal protection clause.

Under that test, the government's action— in this case, the marriage statute limitation to heterosexual couples—

must be upheld against equal protection challenge if there is any reasonably conceivable state of facts that could provide a rational basis for the classification. A State, moreover, has no obligation to produce evidence to sustain the rationality of a statutory classification. "[A] legislative choice is not subject to courtroom factfinding *337 and may be based on rational speculation unsupported by evidence or empirical data."

Haller v. Doe, ___ U.S. ___, ___ _ ___, 113 S. Ct. 2637, 2642-43, 125 L. Ed. 2d 257 (1993) (citations omitted); see Cleburne v. Cleburne Living Center, Inc., 473 U.S. 432, 442, 105 S. Ct. 3249, 3255-56, 87 L. Ed. 2d 313 (1985); Graham v. Richardson, 403 U.S. 365, 371, 91 S. Ct. 1848, 1851-52, 29 L. Ed. 2d 534 (1971); Backman v. United States, 516 A.2d 923, 927 (D.C.1986). Although this case raises serious questions about the legitimacy, under the equal protection clause, of limiting marriage to apposite-sex couples, it appears that the Supreme Court has seen marriage as having a traditional principal purpose: to regulate and legitimize the procreation of children. See Zoblocki, 434 U.S. at 385-86, 98 S. Ct. at 680-81; Skinner, 316 U.S. at 541, 62 S. Ct. at 1113. Although the Court has also made clear that marriage embraces much more, see Turner, 482 U.S. at 95-96, 107 S. Ct. at 2265, I believe that this central purpose of the marriage statute—this emphasis on childbearing —provides the kind of rational basis defined in Heller, ___ U.S. at ___ _ ___, 113 S. Ct. at 2642-43, permitting limitation of marriage to heterosexual couples. I therefore cannot conclude that appellants are entitled to prevail even if the rational basis standard applies.

E. Summary Judgment for the District Inappropriate Assuming, for the Sake of Argument, that Strict Scrutiny Applies

I also conclude that this court cannot resolve the matter in the District's favor by assuming, for the sake of argument, that discrimination against homosexuals is subject to strict scrutiny; for In cannot conclude, as a matter of law, that the District has a compelling or even a substantial interest in reserving marriage for heterosexual couples. The trial court did not explicate the reasons why it so concluded, other than referring to "previously noted' reasons. See supra Part VI.A. As I read the court's opinion, those reasons are the same as those cited under the court's "rational basis' analysis. They are simply too conclusory for equal protection analysis; summary judgment is inappropriate on this issues.

The question then remains; whether (1) the "rational basis" test under equal protection analysis applies, or whether instead (2) a more rigorous scrutiny of the District's policy excluding same-sex marriage is required because homosexuals comprise a scenically protected class. In either event, absent any basis for summary judgment, I conclude the judgment for the District should be reversed and the case remanded for trial.

F. Constitutionally Protected Classes: United States v. Carolene Products Co.

An historical approach will be useful to the classification analysis. The idea that the legislative treatment of particular classes of persons requires greater scrutiny under the equal protection clause than other persons receive is traceable to dictum in a footnote to the Supreme Court's opinion in United States v. Carolene Products Co., 304 U.S. 144, 152 n. 4, 58 S. Ct. 778, 783-84 n. 4, 82 L. Ed. 1234 (1938) (upholding federal statute prohibiting interstates shipment of filled milk under rational basis test). In that footnote, Justice Stone indicated that "prejudice against discrete and insular minorities may be a special condition ... curtail[ing] the operation of those political processes ordinarily to be relied upon to protect minorities, and [so] may call for a correspondingly more searching judicial inquiry." Id. (emphasis added).

Racial discrimination presents the paradigm case under Carolene Products. A particular racial minority, such as Africa-Americans, is "discrete" in the sense intended; they are visible in a way that makes them "relatively easy for other s to identify." Bruce A. Ackerman, Beyond Carolene Products, 98 HARV.L.REV. 713, 729 (1985). African-Americans also are "insular," meaning they tend to "interact" with each other "with great frequency in a variety of social contexts," such as neighborhoods, churches, clubs. See id. at 726. African-Americans, moreover, are nationally, if not always locally, in the "minority." Finally, the history of civil rights litigation and legislation reflects "prejudice" which has resulted in invidious *338 discrimination against African-Americans, requiring various court-ordered remedies.

The premise underlying Carolene Products' call for "more searching judicial inquiry" into allegedly prejudicial treatment of "discrete and insular minorities" is that such minorities lack sufficient political power to fend for themselves in a democratic process that should, but fails, to "generate[] outcomes systematically more favorable to minority interests." Ackerman, supra, 98 HARV.L.REV. at 716. There are problems with the Carolene Products formulation, however. It is underinclusive, as later Supreme Court decisions have made clear. Women, for example, are subject to prejudicial discrimination while comprising a diffuse, not insular group. And, of course, women are not a minority. Furthermore, discrete and insular minorities are not necessarily less able to effectuate their interests through the legislative process than other groups or even disorganized majorities. Racial minorities, in any event, appear to have greater political muscle in most instances than other disadvantaged groups, such as illegitimate children or homosexuals or the poor, all of which tend on the whole to be less identifiable and more diffuse than African-Americans, for example. See id. at 728-31.

Finally, the kind of prejudice reflected in the Carolene Products footnote is also under-inclusive. There are at least two kinds of prejudice the Supreme Court has recognized: (1) lack of effective participation in the political process, as emphasized in Carolene Products, and (2) stigma, i.e., a mark of shame that invites demeaning treatment regardless of one's strength at the polls. See, e.g., Plyler, 457 U.S. at 223, 102 S. Ct. at 2398 ("stigma of illiteracy" affecting undocumented school-age children); Frontiero v. Richardson, 411 U.S. 677, 685, 93 S. Ct. 1764, 1769, 36 L. Ed. 2d 583 (1973) (plurality opinion) ("gross, stereotypical distinctions between the sexes"). As explained below, equal protection analysis has evolved beyond Carolene Products to the point where courts intervene to remedy both kinds of prejudice.

This discussion of Carolene Products is intended as background for discussion of how the Supreme Court has developed and applied the idea of intensified scrutiny to a variety of groups not limited to "discrete and insular minorities," and how the Court has expanded protectable prejudice from lack of effective participation in the political process to prejudice from stigmatizing and stereotyping—from devaluing—particular groups of human beings.

G. Equal Protection After Carolene Products Co.: "Suspect" and "Quasi-Suspect" Classes

In the years since Carolene Products, the Supreme Court has identified two kinds of legislative classifications that require intensive equal protection analysis. These commonly have been called, respectively, "suspect" and "quasi-suspect" classes (meaning the legislative classifications, not the people in them, are "suspect"). The first, "suspect" classification—which the Supreme Court has used to resolve complaints alleging discrimination based on race,[35] alienage,[36] and national origin[37]—must receive "strict scrutiny" from the courts; i.e., for legislative use of the classification to survive, the state must "demonstrate that its classification has been precisely tailored to serve a compelling governmental interest." Plyler, 457 U.S. at 217 & n. 15, 102 S. Ct. at 2395 & n. 15; see Cleburne, 473 U.S. at 440, 105 S. Ct. at 3254 (classifications based on "race, alienage, or natural origin . . . will be sustained only if they are suitably tailored to serve a compelling state interest").

The second, "quasi-suspect" classification—which the Court has used to decide cases alleging discrimination based on gender[38]*339 and illegitimacy[39]—requires "intermediate" or "heightened" scrutiny; i.e., the state must show that legislative use of "the classification reflects a reasoned judgment consistent with the ideal of equal protection" that "further[s] a substantial interest of the State." Plyler, 457 U.S. at 217-18 & n. 16, 102 S. Ct. at 2395 & n. 16; see Cleburne, 473 U.S. at 441, 105 S. Ct. at 3255 ("gender classification fails unless it is substantially related to a sufficiently important governmental interest"; classification based on illegitimacy will survive if "substantially related to a legitimate state interest").

The Supreme Court has not yet addressed whether homosexuals, let alone homosexual couples, constitute a "suspect" or "quasi-suspect" class. See Rowland v. Mad River Local School Dist., 470 U.S. 1009, 1014, 105 S. Ct. 1373, 1376-77, 84 L. Ed. 2d 392 (1985) (Brennan, J. dissenting from denial of certiorari). Nor has the Court clearly defined the differences between these two classifications requiring, respectively, "strict" and "intermediate" (or "heightened") scrutiny.[40] The Court, however, has focused from case to case on several factors to guide the analysis applied to either classification, although the Court has not addressed every factor in every case. As we shall see, these factors have one thing in common: they reflect various ways of evaluating whether intensive court scrutiny and, perhaps, intervention will be necessary, under the equal protection clause, to help substantially powerless classes of people maintain their dignity and receive important rights in the face of harmful, indeed invidious, discrimination by the state (meaning the popular majority).[41]

Specifically, the Court will ask: (1) Has the group suffered a history of purposeful discrimination?[42] (2) Is the class the object of such deep-seated prejudice that it is often subjected to disabilities based on inaccurate stereotypes that do not truly reflect the *340 members' abilities?[43] (3) Is the class defined by the presence of an immutable trait that is beyond a class member's control and yet bears no relation to the individual's ability to contribute to society?[44] (4) Is the group a politically powerless minority?[45]

Interestingly, the first two of these factors reflect a concern for stigma—for unfair stereotyping. The latter two focus on the ability of the group to avoid the claimed disadvantage through self-help—the classic Carolene Products concern.

H. The Implications, If Any, of Bowers v. Hardwick for Equal Protection Analysis

Before considering whether any combination of factors, if satisfied, would require intensive judicial scrutiny of the prohibition against homosexual marriage, it is necessary to note that four federal courts of appeals have ruled—primarily by reference to the Supreme Court's due process decision in Bowers v. Hardwick, supra—that homosexuals do not comprise a suspect or quasi-suspect class. See High Tech Gays v. Defense Industrial Security Clearance Office, 895 F.2d 563, 570-73 (9th Cir.1990); Ben-Shalom v. Marsh, 881 F.2d 454, 564-66 (7th Cir. 1989); Woodward v. United States, 871 F.2d 1068, 1076 (Fed.Cir.1989); Padula v. Webster, 261 U.S.App.D.C. 365, 370-72, 822 F.2d 97, 102-04 (1987).

In Padula, the United States Court of Appeals for the District of Columbia Circuit affirmed summary judgment for the government in appellant's suit claiming the FBI had violated the equal protection clause by refusing to hire appellant because she was a homosexual. The court concluded that homosexuals did not comprise a class deserving intensive scrutiny because the Supreme Court, in Hardwick, had approved state laws criminalizing homosexual conduct, and thereby implicitly had precluded any special protection of homosexuals.

If the Court [in Hardwick] was unwilling to object to state laws that criminalize the behavior that defines the class, it is hardly open to a lower court to conclude that state sponsored discrimination against the class is invidious.

Id. at 371, 822 F.2d at 103 (emphasis added); see Dronenburg v. Zech, 239 U.S.App.D.C. 229, 238-39, 741 F.2d 1388, 1397-98 (1984) (private, consensual, homosexual conduct is not constitutionally protected).

Padula's premise—that homosexual "behavior... defines the class," id.—appears facially overboard; homosexuals as a class are defined by reference to sexual orientation, which does not necessarily imply particular *341 conduct, even in marriage. See supra note 33; see generally Cass R. Sunstein, Sexual Orientation and the Constitution: A Note on the Relationship Between Due Process and Equal Protection, 55 U.CHI.L.REV. 1161, 1178 n. 85 (1988). As the D.C. Circuit itself had noted earlier, "[Hardwick] did not reach the difficult issue of whether an agency of the federal government can discriminate against individuals merely because of sexual orientation." Doe v. Casey, 254 U.S.App. D.C. 282, 296, 796 F.2d 1508, 1522 (1986) (CIA employee alleged he was dismissed because of homosexual orientation, and court concluded that if, indeed, CIA action reflected policy of terminating employment of all homosexuals, CIA would have to justify why this policy was necessary in interests of United States).[46]

It is important to note, however, that in Padula itself, plaintiff-appellant herself rejected the "conduct"/"orientation" distinction, premising her constitutional argument on a definition of homosexuals "as persons who engage in homosexual conduct." 261 U.S.App.D.C. at 370, 822 F.2d at 102. The court accordingly balked at finding invidious discrimination against a class of persons who, by definition, engaged in conduct that, consistent with the Constitution, could be criminalized.

Three other federal circuits followed suit. In Woodward v. United States, the court sustained dismissal of a naval reserve officer from active duty on the ground that he was an admitted homosexual. In rejecting the claims that the Navy's action violated Woodward's constitutional right to privacy, as well as his right to equal protection of the laws, the court relied on Hardwick. In holding that homosexuals were not members of a suspect or quasi-suspect class, the court merely asserted the legislative fact—without citing any authority—that "[m]embers of recognized suspect or quasi-suspect classes, e.g., blacks or women, exhibit immutable characteristics whereas homosexuality is primarily behavioral in nature." Woodward, 871 F.2d at 1076.

Similarly, in Ben-Shalom v. Marsh, the court of appeals, reversing the district court, upheld the Army's refusal to reenlist an admitted lesbian. Relying on Hardwick, the court concluded that, "[if] homosexual conduct may constitutionally be criminalized, then homosexuals do not constitute a suspect or quasi-suspect class entitled to greater than rational basis scrutiny for equal protection purposes." Ben-Shalom, 881 F.2d at 464 (footnote omitted). The court also appeared to be influenced by the fact that it had been asked to overturn an Army regulation, which the court believed would be an unjustified intrusion into military affairs. See id. at 465. The court added that "homosexuals are proving that they are not without growing political power." Id. at 466.

Finally, in High Tech Gays v. Defense Industrial Security Clearance Office, a class action brought by homosexual applicants for employment by the Defense Department, the court of appeals, reversing the district court, rejected plaintiffs' arguments that the Department's refusal to grant security clearances to known or suspected gay or lesbian *342 applicants violated the equal protection clause. The court relied on Hardwick, Ben-Shalom, and Padula and, in particular, asserted as legislative fact—without citing any authority—that "[h]omosexuality is not an immutable characteristic; it is behavioral and hence is fundamentally different from traits such as race, gender, or alienage, which define already existing suspect and quasi-suspect classes." High Tech Gays, 895 F.2d at 573. The court also noted that "homosexuals are not without political power." Id. at 574.

It is critical to understanding the equal protection issue for us to recognize, very frankly, that even if homosexuals as a class could be defined, at least in part, by reference to homosexual conduct, the federal courts in Padula, Woodward, Ben-Shalom, and High Tech Gays—as elaborated below— all misapplied Hardwick, a due process case in which the Supreme Court expressly noted it was not addressing equal protection issues. See Hardwick, 478 U.S. at 196 n. 8, 106 S. Ct. at 2847 n. 8 ("Respondent does not defend the judgment below based on the Ninth Amendment, the Equal Protection Clauses or the Eighth Amendment."). Furthermore, two of these courts, in Woodward and High Tech Gays—apparently recognizing that Hardwick was not dispositive—applied equal protection analysis by merely asserting that homosexuality is not immutable, entirely ignoring a substantial body of scientific research to the contrary. See infra note 49. I have considerable respect for the federal courts that issued these decisions, but this court owes them no deference, and we would abandon our judicial review responsibility if we accepted what, in my view, is critically flawed reasoning.

The discussion above in Part V., concluding that same-sex marriage is not a fundamental right under the due process clause, made clear that, to the extent the due process clause protects substantive rights, it characteristically upholds tradition. As Professor Cass R. Sunstein has noted:

From its inception, the Due Process Clause has been interpreted largely (though not exclusively) to protect traditional practices against short-run departures. The clause has therefore been associated with a particular conception of judicial review, one that sees the courts as safeguards against novel developments brought about by temporary majorities who are insufficiently sensitive to the claims of history.

Sunstein, supra, 55 U.CHI.L.REV. at 1163. Professor Sunstein then stressed that, whereas the due process clause reinforces tradition, the equal protection clause is forward-looking; it is intended to invalidate traditions, however longstanding, that become invidiously discriminatory as times change and disadvantaged groups call attention to their treatment.

The Equal Protection Clause ... has been understood as an attempt to protect disadvantaged groups from discriminatory practices, however deeply engrained and longstanding. The Due Process Clause often looks backward; it is highly relevant to the Due Process issue whether an existing or time-honored convention, described at the appropriate level of generality, is violated by the practice under attack. By contrast, the Equal Protection Clause looks forward, serving to invalidate practices that were widespread at the time of its ratification and that were expected to endure. The two clauses therefore operate along different tracks.

Id.

Professor Sunstein accordingly concluded that, although Hardwick sustained, against due process clause attack, a statute criminalizing consensual sodomy as applied to homosexuals, that ruling does not necessarily foreclose a discrimination claim by homosexuals who contend that the equal protection clause entitles them to the same treatment that heterosexuals receive. See id. at 1163-64; Watkins v. United States Army, 875 F.2d 699, 711, 716-20 (9th Cir.1989) (en banc) (Norris, J., concurring). As applied in this case, Sunstein's analysis suggests that, even though Hardwick says the state can outlaw consensual sodomy between homosexuals without violating the due process clause, this does not necessarily mean the state can deny homosexual couples the right to marry, while *343 allowing heterosexual couples to do so, without violating the equal protection clause.[47]

In concluding that homosexuals did not have a privacy right to engage in sodomy protected by the due process clause, Hardwick did not decide whether the state constitutionally could deny consenting heterosexuals the same right, see id., 478 U.S. at 188 n. 2, 106 S. Ct. at 2842 n. 2, although it is hard to imagine that the Court would find heterosexual sodomy "deeply rooted in this Nation's history and tradition," Moore, 431 U.S. at 503, 97 S. Ct. at 1938, and thus constitutionally protected by the due process clause. See Watkins, 875 F.2d at 717-718 (Norris, J., concurring). If the state can deny that right to heterosexuals and homosexuals alike by criminalizing consensual sodomy by everyone, consistent with due process, based on a tradition of criminal sodomy statutes that have drawn no distinction between homosexual and heterosexual sodomy, see Watkins, 875 F.2d at 718 (Norris, J., concurring), then such conduct provides no basis by itself for denying marriage to homosexual couples without also denying, if not revoking, that right for all heterosexual couples who intend, or are found, to practice consensual sodomy. If, on the other hand, the state cannot ban sodomy for consenting heterosexual couples, consistent with their constitutional right to privacy, then it is not readily apparent why the state could lawfully discriminate against homosexual couples, presumed to be engaging in unlawful consensual sodomy, by denying them a formal status, marriage, that heterosexual couples use to legitimize their own consensual sodomy (among other needs). See id.; C.R. Sunstein, supra, 55 U.CHI. L.REV. at 1169-70.

Lurking in this analysis is a subissue: the assumption that the state, despite Hardwick, cannot constitutionally prohibit consensual heterosexual sodomy by a married couple, consistent with the constitutional right to privacy, does not necessarily mean the state cannot constitutionally prohibit sodomy by a consenting unmarried heterosexual couple, just as the state criminalizes fornication. See, e.g., D.C.Code § 22-1002 (1989 Repl.) (fornication). On that assumption, the question then becomes: whether same-sex couples (presumed for this purpose to be a constitutionally protected class) can use the equal protection clause to claim a constitutional right to marry—absent a compelling or substantially state interest to the contrary— when they admittedly engage in conduct, consensual sodomy, which the state can lawfully proscribe for a large measure of the heterosexual population, namely all unmarried opposite-sex couples.

The answer, I believe, is yes; equal protection is available. In the first place, Hardwick, in dealing only with "consensual homosexual sodomy," 478 U.S. at 188 n. 2, 106 S. Ct. at 2842 n. 2, arguably drew the line between homosexual and heterosexual sodomy, not between unmarried and married conduct. But, even if Hardwick left room for constitutionally protecting consensual sodomy in marriage while permitting criminal penalties for consensual sodomy outside *344 marriage, that would only reenforce appellants' equal protection argument here. Heterosexual couples could validate their conduct and escape criminal prosecution, while attending to their other physical, emotional, and spiritual needs, by entering a marriage; homosexual couples engaging in the same conduct, and having the same other needs, could not. This difference arguably amounts to invidious discrimination because the state would be withholding from homosexual couples a status that heterosexual couples could elect to legitimize for themselves the very same conduct—to the point of curing otherwise criminal conduct—that homosexual couples would be helpless to legitimize. If marriage can make behavior acceptable—and constitutionally protectable—that would otherwise be unacceptable and unprotectable, this means that marriage can legitimize behavior that may be contextually, but not inherently, unacceptable. If, therefore, marriage (under the assumptions considered here) is the only variable that distinguishes between acceptable and unacceptable consensual sodomy, then a law permitting marriage only between opposite-sex couples would appear to discriminate invidiously against members of the only class of unmarried, unrelated, adult couples—homosexual couples—who are disqualified from legitimizing behavior, and from attending to other important needs, that similarly situated heterosexual couples can lawfully arrange to undertake and satisfy.

* * *

In light of the foregoing analysis, Hardwick's due process-approved ban on consensual homosexual sodomy begs the question presented by the equal protection clause: can the state discriminate against homosexual couples by denying them a status, marriage, solely because of conduct, consensual sodomy, which heterosexual couples either (1) have the constitutional right to engage in, at least when formalized by marriage, or (2) do not have the constitutional right to engage in—and are even vulnerable to criminal prosecution for doing so—but yet, despite doing so, retain the right to marry? Unencumbered by Hardwick, therefore, I proceed with the equal protection inquiry by returning to the classification issue.

I. The Factors Applicable to Determining "Suspect" and "Quasi-Suspect" Class Status

Unlike Padula, the three other federal circuit court cases that withheld special recognition for homosexuals under the equal protection clause did not rely exclusively on Hardwick; they also applied the factors the Supreme Court has used to ascertain suspect and quasi-suspect classes. I address these factors separately.

1. History of Purposeful Discrimination

The first issue is whether homosexuals as a group have suffered a history of purposeful discrimination. See supra note 41. In dissenting from the denial of certiorari in Rowland, 470 U.S. at 1014, 105 S. Ct. at 1377, Justice Brennan remarked that "homosexuals have historically been the object of pernicious and sustained hostility, and it is fair to say that discrimination against homosexuals is `likely ... to reflect deep-seated prejudice rather than ... rationality.'" See High Tech Gays, 895 F.2d at 573 ("homosexuals have suffered a history of discrimination"); Ben-Shalom, 881 F.2d at 465 ("Homosexuals have suffered a history of discrimination and still do, though possibly now in less degree."); Equality Foundation of Greater Cincinnati v. Cincinnati, 860 F. Supp. 417, 435 (S.D. Ohio 1994) ("gays and lesbians have suffered a history of invidious discrimination based on their sexual orientation"). Indeed, "[b]eing identified with homosexuality has been the basis of refusals to hire, the ruin of careers, undesirable military discharges, denials of occupational licenses, denials of the right to adopt, to the custody of children and visitation rights, denials of national security clearances and denials of the right to enter the country." Elvia R. Arriola, Sexual Identity and the Constitution: Homosexual Persons as a Discrete and Insular Minority, 10 WOMEN'S RTS.L.REP. 143, 157 (1988) (footnotes omitted). Judge Norris has summarized:

Discrimination against homosexuals has been pervasive in both the public and the *345 private sectors. Legislative bodies have excluded homosexuals from certain jobs and schools, and have prevented homosexual[] marriage. In the private sphere, homosexuals continue to face discrimination in jobs, housing and churches. See generally Note, An Argument for the Application of Equal Protection Heightened Scrutiny to Classifications Based on Homosexuality, 57 S.Cal.L.Rev. 797, 824-25 (1984) (documenting the history of discrimination). Moreover, reports of violence against homosexuals have become commonplace in our society. In sum, the discrimination faced by homosexuals is plainly no less pernicious or intense than the discrimination faced by other groups already treated as suspect classes, such as aliens or people of a particular national origin. See, e.g. Cleburne, 473 U.S. at 440, 105 S. Ct. at 3254....

Watkins, 875 F.2d at 724 (en banc) (Norris, J. concurring). I am satisfied on the basis of these judicial observations that no judge could reasonably conclude as a matter of law, for purposes of summary judgment, that homosexuals have not been the objects of purposeful discrimination.

2. Deep-Seated Prejudice Causing Inaccurate Stereotypes That Do Not Reflect Class Members' Abilities

Second, we must consider whether gays and lesbians have been targets for so much prejudice that they are often presented as inaccurate stereotypes that do not truly reflect their abilities.[48] See supra note 42. The danger from such stereotyping is not only its unfair assault on feelings but, more significantly, its prejudicial impact on rights and opportunities: inaccurate stereotyping typically withholds recognition of one's capacity to be a productive member of society. See Frontiero, 411 U.S. at 686, 93 S. Ct. at 1770; Equality Foundation of Greater Cincinnati, 860 F.Supp. at 437; see also LAURENCE H. TRIBE, AMERICAN CONSTITUTIONAL LAW 1616 (2d ed. 1988) ("in contrast with a characteristic like mental retardation, homosexuality bears no relation at all to the individual's ability to contribute fully to society"); Harris M. Miller, II, Note, An Argument for the Application of Equal Protection Heightened Scrutiny to Classifications Based on Homosexuality, 57 S.CAL.L.REV. 797, 814 (1984).

A commonly advanced stereotype of gays and lesbians, for example, suggests that they are sexually promiscuous and do not want to "settle down" in long-term, committed relationships to raise families. See JUDITH A. BAER, EQUALITY UNDER THE CONSTITUTION: RECLAIMING THE FOURTEENTH AMENDMENT 226-28 (1983); ALLEN P. BELL & MARTIN S. WEINBERG, HOMOSEXUALITIES: A STUDY OF DIVERSITY AMONG MEN AND WOMAN 81 (1978); ALBERT D. KLASSEN, ET AL., SEX AND MORALITY IN THE U.S. 171-73, 179-83 (1989) (many believe that homosexuals are dangerous, wanting to seduce children and colleagues). And yet there is powerful evidence to the contrary: "Approximately three million gay men and lesbians in the United States are parents, and between eight and ten million children are raised in gay or lesbian households." Sexual Orientation and the Law, 102 HARV.L.REV., at 1629 (citing ABA Annual Meeting Provides Forum for Family Law Experts, 13 Fam.L.Rep. (BNA) 1512, 1513 (Aug. 25, 1987)).

This promiscuity stereotype obviously contributes to undermining serious consideration of legitimating same-sex marriages. Viewing the record in the light most favorable to appellants, as we must, I cannot say as a *346 matter of law that gays and lesbians are not the victims of inaccurate stereotyping.

3. Immutability

There is a third question: whether homosexuals can be defined, as a class, by an immutable trait that is beyond a class member's control. This inquiry is important because a characterization that is "not within [a person's] control," such as race, gender, and illegitimacy, and that "bears no relation to the individual's ability to participate in and contribute to society," Lucas, 427 U.S. at 505, 96 S. Ct. at 2762, should not readily serve to justify discrimination by the state. See Frontiero, 411 U.S. at 686-87, 93 S. Ct. at 1770-71.

The degree to which an individual controls, or cannot avoid, the acquisition of the defining trait, and the relative ease or difficulty with which a trait can be changed, are relevant to whether a classification is "suspect" or "quasi-suspect" because this inquiry is one way of asking whether someone, rather than being victimized, has voluntarily joined a persecuted group and thereby invited the discrimination. See Plyler, 457 U.S. at 216-17 n. 14, 102 S. Ct. 2394 n. 14 ("legislation imposing special disabilities upon groups disfavored by virtue of circumstances beyond their control suggests the kind of `class or caste' treatment that the Fourteenth Amendment was designed to abolish" (emphasis added)).

Although the Supreme Court has focused on immutability in a number of cases, see supra note 44, it has "never held that only classes with immutable traits can be deemed suspect." Watkins, 875 F.2d at 725 (Norris, J., concurring); see also Miller, supra, at 813. Indeed, on occasion the Court has referred more broadly, to whether members of the class "exhibit obvious, immutable, or distinguishing characteristics that define them as a discrete group," Lyng v. Castillo, 477 U.S. 635, 638, 106 S. Ct. 2727, 2729, 91 L. Ed. 2d 527 (1986).

On the other hand, I am aware of no Supreme Court decision in which a suspect class has not had an immutable characteristic: race, alienage, national origin. See supra notes 35, 36, and 37. Furthermore, the classes specifically deemed quasi-suspect have reflected immutability as well: gender, illegitimacy. See supra notes 38 and 39. In light of this case law and the long-standing public policy reflected in the marriage statute, I would be hard pressed to say that same-sex couples belonged to a suspect or quasi-suspect class for equal protection purposes if sexual orientation were not virtually immutable.

If homosexuality has a genetic origin, like race or gender, any court—aware of the history of purposeful discrimination against homosexuals—would have to be sympathetic to arguments that any statute forbidding same-sex marriage should be subject to "strict," or at least "intermediate," scrutiny, with the result that the District must show a "compelling," or at least a "substantial," governmental interest in forbidding homosexuals to marry one another. Presumably, the same would hold true if sexual orientation were substantially determined by parental hormonal influences. If sexual orientation, however, were entirely a learned, and thus psychological, phenomenon—and were subject to change through a program of predictably successful, and safe, therapy—then the statute limiting marriage to heterosexual couples, reflecting traditional values, arguably would be reviewable under the "rational basis" test, unless the other factors relevant to determining suspect or quasi-suspect classes, when taken together, would compel more rigorous scrutiny. Cf. Plyler, 457 U.S. at 219 n. 19 & 220, 102 S. Ct. at 2396 n. 19 & 2396 (adult illegal aliens not a suspect class; undocumented status is not "an absolutely immutable characteristic since it is the product of conscious, indeed unlawful, action").

There is no scientific consensus about the origin of sexual orientation, although much has been learned quite recently. There is substantial literature to the effect that sexual orientation is formed at an early age, has a genetic or hormonal basis, and is highly resistant to change once established.[49] One *347 study buttresses its finding that homosexuality is a deep-seated, probably biologically based, virtually unchangeable condition by comparing it with bisexuality, which was found to result, to some extent, from "learning and social experiences."[50] Still other, mostly older studies suggest that homosexual orientation is probably a result of hormonal predispositions interacting with social and environmental factors.[51] Finally, in contrast with most of the recent scientific opinion, some studies suggest that, whatever its source, homosexual orientation is probably culturally determined[52] and, perhaps, can be changed through religious conversion or, with varying degrees of success, through sustained personal commitment and intensive shock aversion therapy or other unusual *348 treatments.[53] Several courts, citing research before the most recent genetic studies, see supra note 49, have taken a middle position, concluding that "the source of sexual orientation is still inadequately understood and is thought to be a combination of genetic and environmental influences." Opinion of the Justices, 129 N.H. 290, 530 A.2d 21, 25 (1987); see Baker v. Wade, 553 F. Supp. 1121, 1129 (N.D.Tex.1982) (same).

Whatever the answers are to questions about the origins of sexual orientation and about the kinds of efforts required to prevent or change homosexual orientation (if possible at all), there is substantial authority to the effect that any effort to change homosexual orientation, once in place, requires traumatic, perhaps even emotionally self-destructive, work toward that end. A 1981 study published by the Alfred C. Kinsey Institute for Sex Research concluded by saying:

Homosexuals, in particular, cannot be dismissed as persons who simply refuse to conform. There is no reason to think it would be any easier for homosexual men or women to reverse their sexual orientation than it would be for heterosexual readers to become predominantly or exclusively homosexual.

ALAN P. BELL, ET AL., SEXUAL PREFERENCE: ITS DEVELOPMENT IN MEN AND WOMEN 222 (1981). Plainly, the very idea of equal protection of the laws stands squarely in the way of any argument that a gay or lesbian is obliged to make what could be an emotionally destructive effort to change sexual orientation rather than receiving constitutional protection of his or her sexual persona as is. See Watkins, 875 F.2d at 725-26 (Norris, J., concurring).[54]

It is interesting to note that a federal district court recently has had a full-blown evidentiary hearing on the nature and causes of homosexuality. As a result, the court declared homosexuals and bisexuals a quasi-suspect class and concluded, on the basis of substantial expert testimony, that "homo-, hetero-, and bisexual orientation is a characteristic beyond the control of the individual." Equality Foundation of Greater Cincinnati, 860 F.Supp. at 437. More specifically, said the court, "sexual orientation is set in at a *349 very early age-3 to 5 years—and is not only involuntary, but is unamenable to change." Id. at 426. At the very least, therefore, I cannot say as a matter of law that homosexuality is not immutable, as that concept is to be understood in equal protection analysis. See supra notes 49, 50, and 53.[55]

4. Political Powerlessness

Finally, there is the question drawn directly from Carolene Products: whether gays and lesbians are a politically powerless minority. See supra note 45. This issue, like immutability, focuses on the power of the putative "suspect" or "quasi-suspect" class to avoid discrimination without help of the court. See Rodriguez, 411 U.S. at 28, 93 S. Ct. at 1294 (whether class is "relegated to such a position of political powerlessness as to command extraordinary protection from the majoritarian political process"). Political power, however, is measured not only by the extent to which a minority group, for example, is represented in legislative bodies, see Frontiero, 411 U.S. at 686 n. 17, 93 S. Ct. at 1770 n. 17, but also, more subtly, by the extent to which "deep-seated prejudice" prevents the group's full participation in the political process, see Plyler, 457 U.S. at 216-17 n. 14, 102 S. Ct. at 2394 n. 14.

One writer indicated almost fifteen years ago that "the gay rights movement is gaining strength in certain parts of the country, and its increasing momentum will likely mobilize the homosexual vote." Homosexuals' Right to Marry, 128 U.PA.L.REV. at 204. This writer also noted that "in areas with large gay populations, avowed homosexuals are beginning to run for public office, and some have even been victorious in city elections." Id. The conclusion, nonetheless, is that "homosexuals are still relatively powerless as a political group." Id.; see generally Kevin A. Zambrowicz, Comment, "To Love And Honor All The Days Of Your Life": A Constitutional Right to Same-Sex Marriage, 43 CATH. U.L.REV. 907, 938-39 (1994).

Measuring the political power of gays and lesbians is further complicated because the stigma of homosexuality has caused many "to conceal their difference and thus to hide part of their identities from the rest of society." BAER, EQUALITY UNDER THE CONSTITUTION at 226. Prejudice has prevented some homosexuals from coming out of the "closet" and joining gay rights organizations that can increase their political power. One commentator summarized:

Prejudice effectively silences homosexuals, and renders them unable to counter and remedy invidious government discrimination caused by that prejudice. Public officials sympathetic to the plight of homosexuals, or themselves homosexual, are also silenced by fear of damage to their political futures.

John C. Hayes, The Tradition of Prejudice Versus the Principle of Equality: Homosexuals and Heightened Equal Protection Scrutiny after Bowers v. Hardwick, 31 B.C.L.REV. 375, 461 (1990) (footnote omitted).

Professor Bruce A. Ackerman has addressed the relative political powerlessness of homosexuals when compared with Carolene *350 Products' most protected "discrete and insular minority":

[C]ompare the problem faced by black political organizers with the one confronting organizers of the homosexual community. As a member of an anonymous group, each homosexual can seek to minimize the personal harm due to prejudice by keeping his or her sexual preference a tightly held secret. Although this is hardly a fully satisfactory response, secrecy does enable homosexuals to "exit" from prejudice in a way that blacks cannot. This means that a homosexual group must confront an organizational problem that does not arise for its black counterpart: somehow the group must induce each anonymous homosexual to reveal his or her sexual preference to the larger public and to bear the private costs this public declaration may involve.
Although some, perhaps many, homosexuals may be willing to pay this price, the fact that each must individually choose to pay it means that this anonymous group is less likely to be politically efficacious than is an otherwise comparable but discrete minority. For, by definition, discrete groups do not have to convince their constituents to "come out of the closet" before they can engage in effective political activity.

Ackerman, supra, 98 HARV.L.REV. at 730-31.

The political power of gays and lesbians is the one relevant factor the trial court did briefly examine:

Of perhaps equal significance to this Court in reaching a similar finding of no "suspect class" or quasi-suspect class" is the reality that homosexuals today are not so lacking in political power as to warrant enhanced constitutional protection. Witness, for instance, the recent passage by the City Council and signing by the Mayor of the Domestic Partnership Bill. Gays and lesbians are, in the 1990's, a political force that any elective officeholder may ignore only at his or her peril.

Since the trial court issued its opinion, however, the United States House of Representatives effectively vetoed the domestic partnership legislation, codified in D.C.Code §§ 36-1401 to -1408 (1993 Repl.), by preventing the District from spending any money to implement it. See D.C. Supplemental Appropriations and Rescissions Act, Pub.L. No. 102-382, 106 Stat. 1422, 1422 (1992); Kent Jenkins, Jr., House Votes Referendum on D.C. Death Penalty; Lawmakers Gut City's Domestic Partners Law, WASHINGTON POST, Sept. 25, 1992, at Al. District Delegate Eleanor Holmes Norton was quoted as saying in response:

The scare word here was "homosexual marriages." Several members [of Congress] told me they saw 30-second commercials coming saying they supported homosexual marriages. The closer we get to an election, the worse the fortunes of any controversial legislation.

Id.

There can be no question that the political power of gays and lesbians has grown over the years in this community and elsewhere. Locally, as noted earlier in this opinion, the gay community in 1975 had the support of Council member Dixon, who sponsored a bill for same-sex marriage legislation that, while eventually withdrawn, received a serious and fair hearing. And, as the trial court noted, the Council adopted and the Mayor signed domestic partnership legislation,[56] although Congress effectively killed it.[57] Two federal courts, moreover, in denying special scrutiny for homosexuals, noted their increasing political power. See Ben-Shalom, 881 F.2d at 466 ("In these times homosexuals are proving that they are not without growing political power."); High Tech Gays, 895 F.2d at 574 ("[L]egislatures have addressed and continue to address the discrimination suffered by homosexuals on account of their sexual orientation *351 through the passage of anti-discrimination legislation.").

For perspective, however, as Professor Ackerman has pointed out, homosexuals— because of their tendency, overall, toward anonymity and diffusion, rather than discreetness and insularity—tend to have considerably less political power than African-Americans, a protected racial minority. See Ackerman, supra, 98 HARV.L.REV. at 730-31. Furthermore, for purposes of evaluating constitutional norms, the focus on political power, or powerlessness, has to be national, not local, lest constitutional rights vary from city to city. Two recent decisions in other jurisdictions attest to the fact that a voter majority will enact state constitutional and city charter amendments in efforts to preclude adoption of legislation to protect gays and lesbians against discrimination on the basis of their sexual orientation. See Equality Foundation of Greater Cincinnati, supra (city charter amendment); Evans v. Romer, 882 P.2d 1335 (Colo.1994) (en banc) (Evans II) (state constitutional amendment); Evans v. Romer, 854 P.2d 1270 (Colo.1993) (en banc) (Evans I) (same).

Accordingly, in light of such developments, including congressional action negating domestic partnership legislation to benefit same-sex couples (and others) in this jurisdiction, see supra note 57, I cannot say the political power of gays and lesbians locally or nationally is strong enough for me to conclude that the trial court was correct as a matter of law in ruling that gays and lesbians, as a class, have the kind of political power that conclusively cuts against their characterization as a suspect or quasi-suspect class. See Equality Foundation of Greater Cincinnati, 860 F.Supp. at 437-39.

J. Whether Homosexuals Comprise a "Suspect" or "Quasi-Suspect" Class

1. Three Easily Applied Factors

Comes now, finally, the dispositive question: on the basis of the non-record sources supplied by the parties and augmented by this court's own research, are we in a position to decide—by applying constitutional norms to legislative facts—whether homosexuals are a suspect or quasi-suspect class entitled to strict or intermediate scrutiny of the marriage statute that allegedly discriminates against them, in violation of the equal protection clause?

As discussed above, with respect to the four relevant factors, there is really no dispute about the first two: homosexuals have suffered a history of purposeful discrimination, and they have been the object of such deep-seated prejudice that they are often subjected to disabilities based on inaccurate stereotypes that do not truly reflect their abilities. I am also satisfied that a hearing with live testimony would add little, if anything, to evaluation of the fourth factor. At least when compared with racial minorities (in particular, African-Americans) and women—two groups constitutionally entitled to intensified scrutiny of alleged discrimination against them—I can say to a virtual certainty that homosexuals evidence no greater political power and would appear, on balance, to evidence less. I doubt a court will learn more from a hearing on this factor than it can from noting the interplay of political forces, documented in the sources cited in this opinion, concerning political advocacy of so-called gay rights.

With this said, I note that a federal district court has found testimonial evidence on the political power issue useful, concluding, after the hearing, that "gays, lesbians and bi-sexuals do not enjoy that type of legislative success, political representation, or political alliances building capability necessary to be considered a politically powerful group." Equality Foundation of Greater Cincinnati, 860 F.Supp. at 439.

2. Immutability

This brings us back to the third factor: immutability. Clearly, this factor is especially critical here because, unlike other groups comprising suspect or quasi-suspect classes to date, there is a serious division of opinion as to whether homosexuality is an immutable trait. Thus, there is a serious question whether homosexuals can escape from that orientation as a matter of will, thereby avoiding the scorn and discrimination that serves as the basis for an equal protection claim. *352 Some homosexuals may contend that immutability simply should not be an issue, i.e., that even if sexual orientation is significantly a matter of choice, not genetic dictation, they are entitled to substantial constitutional protection against discrimination directed at their preference. Even if we assume, solely for the sake of argument, that they are correct about what would be fair under that free-choice premise, mere fairness does not determine equal protection analysis; immutability is a critical factor, for good reason. Were it not, all kinds of groups with all sorts of preferences would demand special protection for behaviors that run counter to legitimate mores of the public-at-large. The Constitution does not afford special treatment for whims.

So what can be said about immutability here? From the sources consulted I can say, with confidence, that homosexuality is not a matter of whim; it falls within a range from biological (genetic and/or hormonal) to psychological predisposition that is very difficult, if not impossible, to reverse. Indeed, the federal district court that has taken extensive expert testimony on the subject, as indicated earlier, has concluded that homosexuality "is not only involuntary, but is unamenable to change." Equality Foundation of Greater Cincinnati, 860 F.Supp at 426. See also Sexual Orientation and the Law, 102 HARV. L.REV. at 1567-68; supra notes 49, 50, and 53. Virtually all the materials I have reviewed tend to show that, at the very least, short of undergoing intensive, prolonged, and traumatic (usually shock) therapy (or religious conversion or commitment that cannot be legally compelled), homosexuals cannot rid themselves of their sexual orientation—and there is no guarantee, in any event, that such therapy will be successful. See supra note 53. In fact, there is plenty of evidence it will not be. See Richard Green, The Immutability of (Homo) Sexual Orientation: Behavioral Science Implications for a Constitutional (Legal) Analysis, 16 J. PSYCHIATRY & LAW 537, 568-69 (1989).

Under these circumstances, I cannot say as a matter of law that homosexuality is not as immutable as race or gender for purposes of equal protection analysis, for I am not willing to say that traumatic, possibly emotionally destructive self-help, rather than constitutional protection, is the price homosexuals must pay (assuming such self-help would be effective, which I strongly doubt) to avoid pernicious discrimination. Indeed, the increasing use of gene therapy and drugs to manipulate health and human behavior suggests the quite scary spectre of enforcing a public policy for "curing" homosexuals—an Orwellian road not to be traveled.[58]

3. The Prevention/Immutability Distinction

On the other hand, the source and nature of homosexuality, while much better understood than even a few years ago, see supra notes 49 through 53, are still largely unknown; much is yet to be learned. For example, although research now may show that homosexuality is virtually immutable in the sense of being very difficult, if not impossible, to reverse once in place, there is a different, indeed threshold question that may be less easy to answer on the basis of what is known today: can homosexuality be prevented from happening? At some point early in a child's development, is the child at a juncture where sexual orientation can be affected, if not determined, by various environmental influences, including lifestyle examples *353 (such as marriage) the child can observe and eventually choose to copy?

This possible distinction between preventing homosexual orientation, on the one hand, and recognizing it is immutable once in place, on the other—I'll call it a prevention/immutability distinction—may or may not reflect a meaningful scientific distinction; the fact is, some courts have found it so. For example, the Supreme Court of New Hampshire asserted, without citation or analysis, that "[f]or purposes of federal equal protection analysis, homosexuals do not constitute a suspect class, nor are they within the ambit of the so-called `middle tier' level of heightened scrutiny...." Opinion of the Justices, 530 A.2d at 24. The court then applied the rational basis test to conclude that, for constitutional purposes, legislation barring homosexuals from adoption or foster parenting had a "legitimate government purpose" of providing "appropriate role models [for children]." Id. at 24, 25. The court explained: "Given the reasonable possibility of environmental [as well as genetic] influences, we believe that the legislature can rationally act on the theory that a role model can influence the child's developing sexual identity." Id. at 25. (Interestingly, the court limited its analysis to the "parent-child or other familial context," id.; the court declined to find a rational basis for excluding homosexuals from operating day care centers, see id.) Furthermore, studies showing that bisexuality may substantially reflect a learned experience, whereas exclusive homosexuality tends to reflect a biological beginning, may be relevant here. See supra note 50. These studies at least suggest the possibility that public policies which can be seen as positively endorsing homosexuality, in contrast with policies more clearly limited simply to forbidding discrimination against homosexuals, may have some bearing on how free an impressionable youth may feel to engage in homosexual experiences, if not to assume that orientation.

It is important to stress, however, that there is powerful evidence that so-called role models do not influence sexual orientation. Research that is becoming significant in cases concerning efforts of same-sex couples to adopt children indicates there is "little ground" for concern that "children might become homosexual if raised in a lesbian or gay household." Joseph Harry, Gay, Male and Lesbian Relationships in CONTEMPORARY FAMILIES AND ALTERNATIVE LIFESTYLES 216, 229 (Eleanor D. Macklin & Roger H. Rubin, eds., 1983). According to one comprehensive report, "every study on the subject has revealed that the incidence of same-sex orientation among the children of gays and lesbians occurs as randomly and in the same proportion as it does among children in the general population; as they grow up, children adopt sexual orientations independently from their parents." Steve Susoeff, Comment, Assessing Children's Best Interests When a Parent Is Gay or Lesbian: Toward a Rational Custody Standard, 32 U.C.L.A.L.REV. 852, 882 (1985).[59]

*354 For example, in Adoption of Tammy, 416 Mass. 205, 619 N.E.2d 315 (1993), the Supreme Judicial Court of Massachusetts held that the Massachusetts adoption statute did not preclude same-sex cohabitants—one of whom was the natural mother of a child fathered through artificial insemination by the other cohabitant's biological cousin— from jointly adopting the child, and that adoption was in the child's best interest. The court noted that a Harvard Medical School clinical professor of psychiatry, who had conducted a clinical assessment, "reviewed and referenced literature on child psychiatry and child psychology which supports the conclusion that children raised by lesbian parents develop normally." Id. 149 N.E.2d at 317. See also Adoption of B.L.V.D. and E.L.V.B., 160 Vt. 368, 628 A.2d 1271 (1993) (Vermont law permits adoption of child by same-sex couple, one of whom was natural mother impregnated by anonymous sperm donor). Considerable research focused on the adoption context, therefore, tends to be useful for contradicting arguments that availability of same-sex marriages, any more than same-sex adoptions, would effectively invite election of a homosexual lifestyle.

Furthermore, there is revealing expert testimony in two lawsuits brought to enjoin voter-enacted constitutional or city charter amendments to bar legislation that would preclude discrimination on the basis of sexual orientation. In the Cincinnati case, the federal district judge found, among other things, on the basis of expert testimony, that "children raised by gay and lesbian parents are no more likely to be gay or lesbian than those children raised by heterosexuals," that "[t]here is no connection between homosexuality and pedophilia," and that "[h]omosexuality is not indicative of a tendency towards child molestation." Equality Foundation of Greater Cincinnati, 860 F.Supp. at 426.

In the Colorado case, the state Supreme Court affirmed the trial court's entry of a permanent injunction barring enforcement of a state constitutional amendment, concluding there was no compelling state interest that would justify such an interference with the fundamental right of gay men and lesbians to participate equally in the political process. Evans II, 882 P.2d at 1350. The state had asserted, among other things, a compelling interest in "allowing the people themselves to establish public social and moral norms"—in particular, norms "preserv[ing] heterosexual families and heterosexual marriage and . . . send[ing] the societal message condemning gay men, lesbians, and bisexuals as immoral". Id. at 1346. The court concluded that, "even recognizing the legitimacy of promoting public morals as a governmental interest," the voter-enacted amendment barring antidiscrimination laws protecting homosexuals was not necessary to achieve that purpose. Id. at 1347. The court referred to relevant expert testimony, concluding that the state's assertion that gay rights laws

will undermine marriages and heterosexual families because married heterosexuals will "choose" to "become homosexual" if discrimination against homosexuals is prohibited. . . files in the face of the empirical evidence presented at trial on marriage and divorce rates.

Id. Considerable litigation to date, therefore, lends substantial support to the proposition that same-sex marriages would not effectively promote a homosexual lifestyle in others.

Inherent in the prevention/immutability discussion are actually two related questions: whether environmental factors influence sexual orientation, and, even if so, whether a marriage statute permitting same-sex marriages would be such an influence. These are two major questions of legislative fact that may become very significant here. One cannot know for sure from all the studies I *355 have cited whether the prevention/immutability distinction has validity or any practical utility. The fact is, though, that fair-minded persons naturally wonder about it.

Even if the prevention/immutability distinction does have meaning, however, it would appear to be irrelevant to this appeal; that distinction does not necessarily cut against ruling that homosexuals constitute a specially protected class under the equal protection clause, because, despite possible preventability, there is considerable evidence that homosexuality, once in place, is virtually unchangeable and thus presumably in need of constitutional protection. Seen in this way, the prevention/immutability distinction reinforces, rather than detracts from a finding that homosexuality is an immutable trait for equal protection purposes. Accordingly, the possibilities (if any) for state action deterring or preventing homosexual orientation would come into play not in "suspect" or "quasi-suspect" class analysis but in "substantial" or "compelling" state interest analysis—to which I now turn.[60]

K. Whether the District Has a "Substantial" or "Compelling" State Interest in Barring Same-Sex Marriage

I assume, for the sake of argument, that homosexuals have an irreversible sexual orientation and thus (when all relevant factors are considered) comprise a quasi-suspect or suspect class. The state, nonetheless, may attempt to demonstrate a substantial, if not compelling, interest in withholding marriage from same-sex couples simply because of a concern that such marriages, if deemed legitimate, could influence the sexual orientation and behavior of children, to the extent choice plays a role. I repeat: the "deterrence" scenario may—or may not—be scientifically far-fetched; this issue of legislative fact is perhaps the most complex of any presented here. But, if there can be any truth to the deterrence rationale at all, the state's interest in preventing same-sex marriages arguably may be substantial enough to deprive same-sex couples of that right, even though not substantial enough to allow discrimination against homosexuals based, for example, on housing or employment.

The state's interest in deterring homosexual lifestyles, of course, would be premised on the general public's adherence to traditional values favoring heterosexual orientation—majoritarian values which homosexuals question. If, however, at trial no scientifically credible evidence were presented tending to prove that a public policy, such as the present marriage statute, can help deter the advent of homosexual orientation and behavior—and I have seen no such evidence to date—then I do not believe the government would have demonstrated a substantial, let alone compelling, state interest in enforcing traditional heterosexual values, to the substantial prejudice of those who cannot share those values. That is to say, if the government cannot cite actual prejudice to the public majority from a change in the law to allow same-sex marriages, such as a predictable increase in antisocial homosexual behavior, then the public majority will not have a sound basis for claiming a compelling, or even a substantial, state interest in withholding the marriage statute from same-sex couples; a mere feeling of distaste or even revulsion at what someone else is or does, simply because it offends majority values without causing concrete harm, cannot justify inherently discriminatory legislation against members of a constitutionally protected class—as the history of constitutional rulings against racially discriminatory legislation makes clear.[61]

*356 Suppose, on the other hand, that scientifically credible "deterrence" evidence were forthcoming at trial, so that either the heterosexual majority or the homosexual minority would be prejudiced in some concrete way, depending on whether the marriage statute was, or was not, available to homosexual couples. See supra note 61. In that case, the ultimate question of whose values should be enforced, framed in terms of what a substantial or compelling state interest really is, would pose the hardest possible question for the court as majority and minority interests resoundingly clash.[62]

L. Proposed Disposition: Reversal and Remand for Trial

Despite familiarity with a substantial body of scientific literature, from which comes the range of possibilities I have identified, I am not comfortable opining about a subject so elusive, and so controversial, as the nature, causes, preventability, and immutability of homosexuality without benefit of a trial record with the right kind of expert testimony, subject to cross-examination. Such expert testimony would have to include—and this is important—examination and cross-examination about the most probative, up-to-date literature. This court was leery about relying at all, let alone exclusively, on scientific literature to determine whether the EMIT drug testing system met with general acceptance in the scientific community. See Jones, 548 A.2d at 39-47; supra note 22. I am even more leery about relying exclusively on scientific and other literature in this case of constitutional magnitude without benefit of questions germane to the issue of immutability, including the causes of homosexuality, asked of leading experts subject to cross-examination.

I have explained earlier, see supra Part IV.C., my doubts about the efficacy of trial court proceedings to ascertain legislative facts, given the likelihood that the parties will use too few truly qualified experts who are familiar with all relevant sources of information and who themselves do not have discernible biases. Thus, I cannot be sanguine about the merits of a remand for legislative fact-finding, as part of the trial, unless the parties arrange for the "ideal" kind of hearing I have called for. See supra Part IV.C.

On the assumption that there would be close to the ideal hearing necessary to address the complex issues presented, I see at least five potential benefits from reversing summary judgment and remanding the case for trial, rather than disposing of the case ourselves at this time on the basis of this court's own legislative fact-finding.

First, the court can gain whatever benefit there can be from testimony by truly knowledgeable experts, subject to examination and *357 cross-examination on the most revealing, up-to-date sources of information available about the causes of homosexuality and their effect on the issue of immutability. See Equality Foundation of Greater Cincinnati, supra; Evans II, supra. The non-record sources I already have reviewed, of course, will be available not only for examination at trial but also for this court's scrutiny once again (assuming appeal) in the event the trial record itself proves unsatisfactory.

Second, there is always the possibility that the fact-finding process at a trial will reveal reasons why the "rational basis" test, not scrutiny of a higher order, should apply. Based on the sources I have reviewed, however, most of which are cited in this opinion, I cannot say (as the trial court has) that the rational basis test applies as a matter of law, and I am skeptical, to say the least, about the likelihood of that test's surviving further inquiry. See Gay Rights Coalition, 536 A.2d at 36 (lead opinion) ("sexual orientation appears to possess most or all of the characteristics that have persuaded the Supreme Court to apply strict or heightened constitutional scrutiny to legislative classifications under the Equal Protection Clause"). At the same time, I am not yet comfortable saying—without more detailed, hopefully helpful examination corroborating or challenging my understanding of relevant social and scientific facts—that the traditional test assuredly does not apply. At least a trial will give the government its best opportunity to advocate the rational basis test—an opportunity the government deserves in this complex constitutional case, just as the plaintiffs-appellants deserve the opportunity to demonstrate the contrary.

Third, on the assumption that my present, tentative rejection of the rational basis test holds after trial, the anticipated cross-examined expert testimony, when added to existing, non-record sources, may be important to deciding whether homosexuals are a "suspect," or instead a "quasi-suspect," class entitled to "strict" or only "intermediate" (or "heightened") scrutiny. The Supreme Court to date has recognized only three suspect classes: race, alienage, and national origin; see supra notes 35, 36, 37, and a plurality of the Court once placed gender in that category, see Frontiero, 411 U.S. at 682, 93 S. Ct. at 1768 (plurality opinion), although gender today, like illegitimacy, receives intermediate scrutiny as a quasi-suspect classification. See supra note 38, 39. Women, of course, do not comprise a minority, and illegitimacy does not reflect the kind of discriminatory stereotyping historically experienced by racial minorities—and by homosexuals. see Mathews, 427 U.S. at 506, 96 S. Ct. at 2762-63. If a trial were to confirm that homosexuality, once in place, is no less immutable than race, or is virtually no less so, then there may be reason, when considering all factors, to consider homosexuals a suspect class. See Watkins, 875 F.2d at 724-28 (Norris, J., concurring) (concluding homosexuals comprise a suspect class); TRIBE, supra, at 1616 ("homosexuals in particular seem to satisfy all of the Court's implicit criteria of suspectness"); ELY, supra, at 162-64 ("Homosexuals for years have been the victims of both `first-degree prejudice' and subtler forms of exaggerated we-they stereotyping"). At this point, however, I am not prepared to make that judgment.

Fourth, given the predictable negative public response to a decision that ultimately could require the District to recognize same-sex marriages—absent a compelling, or at least substantial, governmental interest in continuing to ban them—it is important to be sure that the contending parties have every opportunity to make the strongest possible cases, pro and con, observed by all interested persons, so that procedurally there appears to be the maximum legitimacy to court decisions ultimately made. There is the danger, of course, as I have discussed, that the parties will not call truly knowledgeable experts, such as persons who have conducted critical studies, see supra notes 49 through 53, or who are truly informed about all probative studies and are unquestionably qualified to testify about them. But I believe it is important to give the parties the opportunity to do so, to everyone's benefit. Ultimately, of course, the reviewing court—beginning with this court (assuming eventual appeal)—would have responsibility for determining whether the trial record added anything useful to what we already know.

*358 Finally, even if this court were to conclude at this time that appellants comprise a suspect or quasi-suspect class, we would still have to remand for trial on whether the District can show a "compelling," or at least a "substantial" (or "important"), governmental interest in denying same-sex marriages. Cf. Evans I, supra (sustaining grant of preliminary injunction and remanding for determination whether voter-enacted constitutional amendment that precluded anti-discrimination legislation protecting "homosexual[s], lesbian[s], or bisexual[s]"—held to infringe on fundamental right under Equal Protection Clause to participate equally in political process—was supported by compelling state interest and narrowly drawn to achieve that interest in least restrictive manner possible). Interestingly, as we have seen, much of the "immutability" evidence germane to determining whether homosexuals comprise a suspect or quasi-suspect class is likely to be relevant to resolving whether the government has a compelling or at least a substantial interest in preventing same-sex marriages. See Evans II, 882 P.2d at 1347 (in holding that constitutional amendment precluding antidiscrimination legislation to protect homosexuals did not serve compelling state interest, court concluded that state's assertion gay rights laws would undermine marriages by enforcing heterosexuals into gay lifestyles "flies in the face of the empirical evidence presented at trial"). Consequently, as long as there has to be a trial that deals with evidence about the nature and causes of homosexuality, there is every reason to benefit from the trial court's initial application of that evidence (in addition to applying the legislative fact data I have compiled) to the threshold question whether homosexuals are entitled to "suspect" or "quasi-suspect" class status.

As far as I am aware, in virtually all other cases concerning suspect or quasi-suspect classifications, the immutability factor has not been contested—indeed, it has been self-evident. Accordingly, appellate courts, including the Supreme Court, typically have resolved the classification issue with judicial analysis of the relevant factors based exclusively on their own legislative fact-finding. See supra notes 43, 54. I have found one exception in this jurisdiction, however. Twenty years ago in Waldie v. Schlesinger, 166 U.S.App.D.C. 175, 509 F.2d 508, the United States Court of Appeals for the District of Columbia Circuit, confronted by a constitutional challenge to the "men only" admissions policies of the service academies, reversed the District Court's grant of summary judgment for the government and remanded the cases for a full trial on the merits, including the question whether the rational basis test applied to discrimination against women.

[W]e are not nearly as certain as the District Court that the Supreme Court has settled on the "rationality" standard for testing sex-based equal protection claims. Rather, we think this area of constitutional law is still evolving and is often highly dependent on the facts of each case. Accordingly, a full development of the facts of these cases is essential to any meaningful assessment of appellants' claim against the rapidly changing, and variously interpreted, case law.

Id. at 177, 509 F.2d at 510.

As in Waldie, in Equality Foundation of Greater Cincinnati, and in the Evans litigation in Colorado, a trial court record developing the legislative facts relevant to this case would be helpful here.[63] Furthermore, although the immutability issue would be the reason for the remand, I would not limit the parties or the trial court to that issue. All legislative facts applicable to all factors relevant to equal protection analysis would be fit subjects for trial.

Accordingly, I would reverse summary judgment for the District and remand the case for trial, at which the court would decide (1) the level of scrutiny constitutionally required and, if strict or intermediate scrutiny were called for, (2) whether the District has demonstrated a compelling or substantial enough governmental interest to justify refusing appellants a marriage license.

*359 VII. POSTSCRIPT: RESPONSE TO MAJORITY ON EQUAL PROTECTION

My colleagues in the majority very simply reject appellants' equal protection claim either (1) because marriage, as traditionally and statutorily defined, does not discriminate against homosexuals in fact since "marriage," conceptually understood, cannot include same-sex couples (Judge TERRY); or (2) because the marriage statute, even if it discriminates against homosexuals in fact, does not discriminate against them as a matter of constitutional law, since the statute does not reflect a discriminatory purpose (Judge STEADMAN).

Judge TERRY's response begs the question. Given appellants' undisputed proffer that many if not most homosexuals have the same emotional, spiritual, physical, and public-benefit needs that lead heterosexuals to marry, see supra note 33, there is no convincing basis for saying either that these needs do not exist or that, if they do, marriage could not satisfy them—the only reasons I can think of for saying that marriage and homosexuality, by definition, cannot fit together. This analysis is akin to the premise of the trial court opinion the Supreme Court rejected in Loving: that a divine natural order forbids racial intermarriage to the point of making it conceptually unthinkable.[64]

If homosexuals comprise a suspect or quasi-suspect class, which Judge TERRY does not question, then they cannot lawfully be denied the right to marry on the constitutionally unprecedented ground that this claimed right, by definition, is impossible to confer. See supra note 20.

Judge STEADMAN's principal argument is also unsound. He essentially rests his position on the incontestable fact that the legislature, when enacting the marriage statute for heterosexual couples, would not have dreamed that anyone might consider it discriminatory legislation against homosexuals. From that premise Judge STEADMAN concludes that the marriage statute cannot be held to reflect the discriminatory purpose essential to denial of equal protection of the laws. Compare Washington v. Davis, 426 U.S. 229, 96 S. Ct. 2040, 48 L. Ed. 2d 597 (1976) (upholding District of Columbia qualifying test for metropolitan police officers, neutral on its face but with discriminatory impact on black applicants, since test, having valid objective, did not reflect discriminatory purpose) and Personnel Administrator v. Feeny, 442 U.S. 256, 99 S. Ct. 2282, 60 L. Ed. 2d 870 (1979) (sustaining Massachusetts veterans preference for state civil service employment against equal protection challenge based on discriminatory impact on female civil service applicants, because there was no purposeful discrimination against women, some of whom were eligible veterans) with Yick Wo v. Hopkins, 118 U.S. 356, 6 S. Ct. 1064, 30 L. Ed. 220 (1886) (striking down San Francisco ordinance, neutral on its face, limiting commercial laundries to brick or stone buildings, absent Board of Supervisors' permission to use other materials, since ordinance was intended to discriminate against Chinese laundries typically housed in wooden buildings).

There are fundamental defects in this analysis. The line of cases that begins with Davis, sustaining statutes and other governmental actions, neutral on their face, for lack of a perceived discriminatory purpose, is inapplicable here for two reasons. First, the governmental actions at issue in those cases did not purport to exclude the affected classes altogether from the benefits sought. In Davis, racial minorities who passed the *360 qualifying test were eligible for appointment as police officers; in Feeny, female war veterans were entitled to the civil service employment preference. In short, the impacts were disproportionate but not exclusionary. In contrast, in the case of homosexuals, marriage is altogether forbidden to same-sex couples.

Second, the fact that the legislature, in adopting the marriage statute, did not have homosexuals in mind does not mean the statute lacks a discriminatory purpose; an absolute prohibition, whether explicit or implied, resulting in discrimination in fact, has an inherent discriminatory purpose, even if the legislature did not recognize it. As already indicated, when the Virginia legislature enacted an antimiscegenation statute, it was premised on the belief that couples of different races were inherently incapable of marriage, because the very idea of interracial marriage was an oxymoron—a perceived abomination that violated divine natural law. See Loving, 388 U.S. at 3-7, 87 S. Ct. at 1819-21. Similarly, when railway passenger cars were racially segregated by law, there was no perceived discriminatory purpose; though "separate," they were "equal." Plessy v. Ferguson, 163 U.S. 537, 16 S. Ct. 1138, 41 L. Ed. 256 (1896).[65] The Supreme Court came to recognize, however, that as the idea of equality changed, the purposes inherent in once-benign, supposedly nondiscriminatory statutes were seen to change. See Brown, 347 U.S. at 489-95, 74 S. Ct. at 688-92 (concluding that "[s]eparate educational facilities are inherently unequal"); Loving, 388 U.S. at 12, 87 S. Ct. at 1823 ("restricting the freedom to marry solely because of racial classifications violates the central meaning of the Equal Protection Clause").

Accordingly, if, as I believe is ineluctably true, the marriage statute discriminates against homosexuals in fact because they claim the same emotional, spiritual, and physical needs for marriage that heterosexuals have, see supra notes 20 and 33, then a court, in my opinion, cannot legitimately say that the statute—which leaves no room for qualifying homosexual couples—does not reflect a discriminatory purpose. Judge STEADMAN's "statute of inclusion of opposite-sex couples," post at 122, has become a statute of exclusion of same-sex couples.

Judge STEADMAN advances another argument against appellants' equal protection claim: because heterosexual couples, but not homosexual couples, have a fundamental right to marry, "the state may give separate recognition solely to that institution through a marriage act as here." Post at 125 (footnote omitted). I have pointed out at length, see supra Part VI. H., that "fundamental right" analysis under the due process clause does not dictate analysis applicable to a "suspect" or "quasi-suspect" class under the equal protection clause. Furthermore, I cannot accept my colleague's proposition (necessarily premised on at least a "quasi-suspect" class assumption) that the relationship between marriage and procreation, childbirth, and child rearing creates an "important" (if not "compelling") governmental interest that *361 justifies—as a matter of law—the statutory limitation of marriage to heterosexual couples, even if the statute were held to be discriminatory. As I have stressed earlier in Part VI. K. of this opinion, any governmental interest that might trump appellants' discrimination claim must be a matter for trial, not a matter for summary judgment. That claimed governmental interest has never been elaborated in the trial court or on appeal.

Taken together, my colleagues in the majority seem to be saying that, although the equal protection clause may not permit the state to discriminate against homosexuals in some areas, such as employment, any constitutional concern evaporates when marriage becomes the issue simply because marriage is different: it is conceptually limited by its traditional definition to opposite-sex couples—a limitation that inherently, therefore, cannot reflect discrimination against homosexual couples in fact or purpose. This definitional defense against an equal protection claim has failed before when substantial human rights, asserted in connection with a claimed right to marriage, have been at issue. See Loving. It should fail here. See generally James Trosino, American Wedding: Same-Sex Marriage and the Miscegenation Analogy, 73 B.U.L.REV. 93 (1993).

In concluding its opinion in Griswold v. Connecticut, 381 U.S. 479, 486, 85 S. Ct. 1678, 1682, 14 L. Ed. 2d 510 (1965), in another context, the Supreme Court stressed:

We deal with a right of privacy older than the Bill of Rights—older than our political parties, older than our school system. Marriage 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.

I see no basis on this record, or in law, for concluding summarily that plaintiffs-appellants, a homosexual couple, cannot lay equal claim to such a "noble" relationship, sustained by law.

TERRY, Associate Judge:

I join in parts I-III of Judge Ferren's opinion. I also join in Judge Steadman's opinion. Accordingly, I vote to affirm the judgment of the trial court in all respects, although not precisely for the reasons stated by the trial court.

The outcome of this case, in my view, turns on the definition of "marriage." Shakespeare in his 116th Sonnet wrote of "the marriage of true minds." In the game of pinochle, the king and queen of the same suit are referred to as a "marriage" when those cards are held by the same player; if that suit is trump, the combination of king and queen is a "royal marriage." But these and similar expressions are only metaphors, figures of speech derived from the literal meaning of the word that serves as the fulcrum of this case. Judge Ferren, in parts II and III of his opinion, cogently demonstrates that the word "marriage," when used to denote a legal status, refers only to the mutual relationship between a man and a woman as husband and wife, and therefore that same-sex "marriages" are legally and factually— i.e., definitionally—impossible.

This conclusion necessarily disposes of the equal protection issue that Judge Ferren goes on to discuss in part VI of his opinion. That is, if it is impossible for two persons of the same sex to "marry," then surely no court can say that a refusal to allow a same-sex couple to "marry" could ever be a denial of equal protection. I am willing to assume, for the purposes of this discussion, that homosexuality is an immutable trait; indeed, recent scientific literature strongly suggests that this is so, as Judge Ferren tells us, ante at 346-347 & nn. 49-52. But if two people are incapable of being married because they are members of the same sex and marriage requires two persons of opposite sexes, as Judge Ferren has shown, then I do not see how it makes any difference that the District of Columbia, or any agency of its government, discriminates against these two appellants by refusing to allow them to enter into a legal status which the sameness of their gender prevents them from entering in the first place. Thus Judge Ferren's discussion of "adjudicative" versus "legislative" facts in *362 part IV, while fascinating, is ultimately irrelevant to the outcome of this case,[1] and the equal protection issue is moot.

It seems obvious that the remedy for the dilemma facing these appellants lies exclusively with the legislature. The Council of the District of Columbia can enact some sort of domestic partners law, bestowing on same-sex couples the same rights already enjoyed by married couples, whenever it wants to. But no court can order a legislature to enact a particular statute so as to achieve a result that the court might consider desirable, or to appropriate money for a purpose that the court might deem worthy of being funded. See Zahn v. Board of Public Works, 274 U.S. 325, 328, 47 S. Ct. 594, 594-95, 71 L. Ed. 1074 (1927); Hart v. United States, 118 U.S. 62, 67, 6 S. Ct. 961, 963, 30 L. Ed. 96 (1886); cf. Reeside v. Walker, 52 U.S. (11 How.) 272, 289-290, 13 L. Ed. 693 (1851) (mandamus will not lie against the Secretary of the Treasury to pay a claim when Congress has not appropriated money to pay it). The separation of powers doctrine prohibits such action by a court. Nor can a court alter or expand the definition of marriage, as that term has been understood and accepted for hundreds of years. Thus the Council, and only the Council, can provide Messrs. Dean and Gill with the relief they seek.

Having concluded unanimously that it is impossible for two persons of the same sex to marry, this court cannot also conclude that it is—or even may be—a denial of equal protection to refuse to allow such persons to marry. The two conclusions are inherently inconsistent.[2] If these appellants cannot enter into a marriage because the very nature of marriage makes it impossible for them to do so, then their quest for a marriage license is a futile act, and the District's refusal to issue a license to them is legally and constitutionally meaningless. They are, of course, free to refer to their relationship by whatever name they wish. But it is not a marriage, and calling it a marriage will not make it one.

STEADMAN, Associate Judge, concurring:

I join Judge FERREN'S compelling analysis of appellants' several arguments in Parts I., II., III., and V. of his comprehensive opinion. However, in my judgment, the marriage statute must be sustained as well against the challenge under constitutional equal protection, applicable within the District of Columbia through the due process cause of the Fifth Amendment. See Bolling v. Sharpe, 347 U.S. 497, 499-500, 74 S. Ct. 693, 694-95, 98 L. Ed. 884 (1954).

My initial difficulty with a postulate of appellants' analysis, reflected in Judge FERREN'S discussion of equal protection, is its treatment of the marriage statute as the equivalent of a statute expressly addressed to an assertedly suspect class. The marriage statute is simply not the same as, say, a statute prohibiting the employment of homosexuals. Cf. Evans v. Romer, 882 P.2d 1335 (Colo.1994) (holding that Amendment 2, which prohibited the state and municipalities from passing legislation to protect homosexuals, infringed on plaintiffs' right to vote in violation of the Equal Protection Clause). Rather, it is a statute of inclusion of opposite-sex couples who may wish to enter a particular legal status recognized by the state. To the extent it is exclusive, it is exclusive evenly of all same-sex couples, who may, for whatever reason, wish to enter that legal *363 status.[1] I think it would take a considerable stretch to find, in such circumstances, the requisite "purposeful" or "invidious" legislative discrimination addressed to homosexuals. See Personnel Administrator v. Feeney, 442 U.S. 256, 274, 99 S. Ct. 2282, 2293-94, 60 L. Ed. 2d 870 (1979) (when a neutral law is challenged as having a disparate impact on women, plaintiffs must show purposeful discrimination); Washington v. Davis, 426 U.S. 229, 239-40, 96 S. Ct. 2040, 2047-48, 48 L. Ed. 2d 597 (1976) (plaintiffs who alleged a racially disproportionate impact from governmental action must show invidious discrimination on the part of governmental actors).[2]

But even assuming that the marriage statute should be analyzed as one of unequal application to homosexuals and assuming further that homosexuals are a quasi-suspect class, as Judge FERREN suggests may be the case,[3] I fail to see an unconstitutional transgression of equal protection. As Judge FERREN demonstrates, the right to marry is a fundamental right only in application to opposite-sex couples. While plainly the marriage state involves far more,[4] the Supreme Court teaches that at bottom the institution reflects considerations "fundamental to the very existence and survival of the [human] race," Skinner, supra, 316 U.S. at 541, 62 S. Ct. at 1113, and bound up with sexual relations, procreation, childbirth and child rearing. Zablocki, supra, 434 U.S. at 386, 98 S. Ct. at 681.[5] It seems to me apparent that *364 much the same considerations that elevate opposite-sex marriage to the status of a fundamental right constitute the requisite substantial relationship to an important governmental interest[6] of a statute designed to recognize and promote that fundamental right. Surely, if only opposite-sex marriage is a fundamental right, the state may give separate recognition solely to that institution through a marriage act as here.[7]

These and like considerations have led, so far as I am aware, every appellate court in the land presented with the issue to reject federal constitutional challenges to opposite-sex marriage statutes.[8] I am led to the same conclusion.

NOTES

[1] Appellants have not renewed on appeal their argument that the trial court's (and possibly the Council's) reliance on the Bible to understand the meaning of "marriage" violated appellants' First Amendment rights.

[2] The marriage statute enacted and codified in 1901 was a new statute derived from a 1777 Maryland statute entitled "An Act Concerning Marriages," 1777 Md.Laws, ch. 12, in 1 CLEMENT DORSEY, THE GENERAL PUBLIC STATUTORY LAW AND PUBLIC LOCAL LAW OF THE STATE OF MARYLAND, at 130 (1840) (hereafter DORSEY). See D.C.Code Encyclopedia 30-101 (1968). The original, 1977 Maryland statute had been amended by "An Act to Repeal Part of the Act Concerning Marriage," 1785 Md.Laws, ch. 35, in 1 DORSEY, at 196, and by "An Act to Repeal So Much of the Act Concerning Marriages as is Therein Mentioned," 1790 Md.Laws, ch. 20, in 1 DORSEY, at 256.

After the District of Columbia was created in 1801, Maryland law continued to apply to the part of the District ceded by Maryland. See D.C.Code vol. 1 at 46 (1991). In 1866, Congress enacted "An Act Legalizing Marriages and for Other Purposes in the District of Columbia," 14 Stat. 236, ch. 240 (1866), which permitted "[c]ertain colored persons in the District of Columbia, living together as man and wife, ... to be deemed husband and wife." In 1870, "An Act to Further Amend the Law of the District of Columbia in Relation to Judicial Proceedings, and Preserve Records of Marriages Therein," 16 Stat. 146, ch. 115 (1870), was enacted primarily to set forth the form and record of marriage licenses. Provisions requiring the clerk of the court to "examine and ascertain the full names, ages, color," prior marital status, and relationship of the parties desiring to marry prior to issuing a marriage license were added in 1896 in "An Act To Regulate Marriages in the District of Columbia," 29 Stat. 118, ch. 177 (1896-97).

[3] Congress enacted the District's first divorce statute in 1860 entitled "An Act to Authorize Divorces in the District of Columbia, and for Other Purposes," 12 Stat. 59, ch. 158 (1860). This act was the precursor of the divorce statute enacted in 1901. See D.C.Code Encyclopedia 16-901 (1966). "An Act Concerning Divorces in the District of Columbia," 16 Stat. 147, ch. 116 (1870), added habitual drunkenness, cruelty, and willful desertion to the grounds for divorce.

[4] From time to time between 1901 and 1977, the marriage statute was amended in minor respects; however, none of the changes has significance for the present case. See, e.g., 32 Stat. 543, ch. 1329 (1902) (added age of consent requirement for males and females to § 30-103, "Marriages Void from Date of Decree"; corrected misspelling in § 30-104); 50 Stat. 626, ch. 596 (1937) (increased age of consent for males from sixteen to eighteen and for females from fourteen to sixteen); 80 Stat. 264, Pub.L. 89-493, § 13(b) (1966) (added sentence to § 30-106 to permit deputy court clerks to celebrate marriages); 84 Stat. 570, Pub.L. 91-358, Title I, § 155(a) (changed "District of Columbia Court of General Sessions" to "Superior Court of the District of Columbia").

Although more significant amendments were made from time to time to the divorce provisions, those changes also are irrelevant to the present case. See e.g., 41 Stat. 567, ch. 153, § 983(a) (1920) (added section setting the effective date of an annulment or divorce decree); 75 Stat. 515, Pub.L. 87-246, § 3 (1963) (restored wife's right of dower and established a statutory right of dower in the husband); 77 Stat. 560, Pub.L. 88-241, § 1 (1963) (defined the term "court" as used in the chapter; substantively changed residency requirements; changed phraseology and arrangement of several sections).

[5] The minor changes made to the marriage and divorce statutes before 1977 focused primarily on arrangement and phraseology. See supra note 4. The Marriage and Divorce Act of 1977, however, made substantive changes in the law. As the Chairperson of the Committee on the Judiciary and Criminal Law reported to members of the Council of the District of Columbia:

In particular, the subject bill would reduce the residency requirement for eligibility to file a divorce or annulment action from the current standard of one year to six months. The bill would add to the D.C.Code a new "no fault" ground for obtaining a divorce or legal separation, namely, "if both parties to the marriage have lived separate and apart without cohabitation for a period of one year next proceeding the commencement of the action." ... [T]he bill [also] modifies the current fault grounds for divorce by removing desertion for one year and imprisonment for a felony conviction as grounds."

COMM ON THE JUDICIARY AND CRIM. LAW, REPORT ON BILL No. 1-89, THE DISTRICT OF COLUMBIA MARRIAGE AND DIVORCE ACT 5-6 (June 24, 1976).

[6] Original Bill No. 1-89 generated substantial debate. Gay rights groups agreed that the proposed gender-neutral language of the statute authorized same-sex marriages; however, they asked the Council to amend the bill to provide expressly for same-sex marriages. In his testimony before the Committee on the Judiciary on July 7, 1975, Cade Ware, President of the Gay Activists Alliance of Washington, D.C. proposed that § 30-101 be amended to read:

A marriage between two persons, regardless of race, creed, sex or religious affiliation, licensed, solemnized, and registered as provided in this act is valid in the District of Columbia.

The District of Columbia Uniform Marriage and Dissolution Act, 1975: Hearing on Bill 1-89 Before the District of Columbia Council Comm. on Judiciary and Criminal Law, July 7, 1975, at 8 (testimony of Cade Ware, President of the Gay Activists Alliance of Washington, D.C.).

Groups on the other side of the debate were concerned with the bill's failure to prohibit same-sex unions expressly. The Catholic Archdiocese of Washington submitted a letter to the Judiciary Committee stating that such a sweeping change in the law—the implicit authorization of same-sex marriages in proposed § 30-101—could not be inferred from "the law's failure to prohibit expressly something which has otherwise never been permitted under a marriage statute." The letter nonetheless recommended amending § 30-101 to read:

A marriage between a man and a woman which is licensed, solemnized and registered as provided in this act is valid in the District of Columbia.

Letter from William W. Baum, Archbishop of Washington, to David Clarke, Chairman of the Committee on the Judiciary and Criminal Law, at 8 (Aug. 27, 1975).

[7] D.C.Code § 30-110 describes the duties of the Clerk to obtain certain information from marriage license applicants.

[8] Former D.C.Code § 30-116 dealt with slave marriages.

[9] We reiterate: the District's 1901 marriage statute, incorporated into the first D.C.Code, was a new statute that replaced previous legislation, see 31 Stat. at 1189, ch. 1, § 1. It was not merely a law carried over from before codification. See supra note 2.

[10] Former D.C.Code § 30-116 discussing slave marriages traceable to the 1901 Code, used the terms "husband and wife"; however, the Marriage and Divorce Act of 1977, 1977 D.C.Stat. 119, repealed § 30-116. See supra note 8.

[11] D.C.Code § 30-101, titled "Marriages void ab initio—In general," provides:

The following marriages are prohibited in the District of Columbia and shall be absolutely void ab initio, without being so decreed, and their nullity may be shown in any collateral proceedings, namely:

(1) The marriage of a man with his grandmother, grandfather's wife, wife's grandmother, father's sister, mother's sister, mother, stepmother, wife's mother, daughter, wife's daughter, son's wife, sister, son's daughter, daughter's daughter, son's son's wife, daughter's son's wife, wife's son's daughter, wife's daughter's daughter, brother's daughter, sister's daughter;

(2) The marriage of a woman with her grandfather, grandmother's husband, husband's grandfather, father's brother, mother's brother, father, stepfather, husband's father, son, husband's son, daughter's husband, brother, son's son, daughter's son, son's daughter's husband, daughter's daughter's husband, husband's son's son, husband's daughter's son, brother's son, sister's son;

(3) The marriage of any persons either of whom has been previously married and whose previous marriage has not been terminated by death or a decree of divorce.

This provision was adopted in 1901 and was derived from the 1777 Maryland statute, 1 DORSEY at 130-31 § 1. See supra note 2. Although the Maryland provision was worded differently and provided that one who violated the statute must pay 500 pounds or be banished from the state forever, the relationships forbidden in the 1777 statute were exactly the same—and even were listed in the same order—as in the current D.C. statute.

[12] Two provisions in other chapters of the Domestic Relations title—Property Rights and Uniform Support—also use the terms "husband" and "wife." See D.C.Code §§ 30-201 (section does not affect ownership of property held by husband and wife as tenants by the entireties), -318 (husband and wife shall be competent witnesses). Section 30-201 was codified in 1901, "An Act to Establish a Code of Law for the District of Columbia," 31 Stat. 1189, ch. 854 (1901). Section 30-318 was not enacted until 1957, "An Act To improve and extend, through reciprocal legislation, the enforcement of duties of support in the District of Columbia," 71 Stat. 288, Pub.L. 85-94 § 18 (July 10, 1957).

[13] Although the Supreme Court of Hawaii, in Baehr v. Lewin, 74 Haw. 530, 852 P.2d 44 (1993), recently reversed a trial court decision barring same-sex marriages and remanded for further proceedings, the court's opinion was premised on state constitutional grounds, not on statutory interpretation applying an evolving definition of marriage.

[14] The Anti-Sex Discriminatory Language Act, as such, does not appear in the D.C.Code. Rather, the Act amended dozens of existing provisions of the D.C.Code "to provide legal equality for men and women." COMM. ON THE JUDICIARY AND CRIM. LAW. REPORT ON BILL No. 1-36, THE ANTI-SEX DISCRIMINATORY LANGUAGE ACT, at 3 (May 20, 1976).

[15] D.C.Code § 30-111 explains who can give consent to marry for persons under age 18.

[16] A three-year set-back permits life insurance companies to "calculate premiums for women according to an age not more than three years younger than their actual age, due to the fact that women's life expectanices are longer than men's." NOW, 531 A.2d at 277.

[17] In support of their motion for summary judgment, appellants submitted Plaintiffs' Statement of Material Facts not in Dispute (Dean and Gill are a gay male couple who did not suffer from any of the prohibitions of marriage enumerated in the Marriage Act); an Affidavit of Craig Robert Dean; an Affidavit of Patrick Gerard Gill; an Affidavit of Councilmember David Clark regarding his sponsorship of the Human Rights Act; Testimony of Dr. Franklin E. Kameny, President of the Mattachine Society, at a Council Hearing discussing the Marriage and Divorce Act of 1977; Councilmember Arrington Dixon's Memorandum to the Judiciary Committee regarding the Marriage and Divorce Act of 1977; a Memorandum on the History of Same-sex Marriage; a Post-Hearing Submission on the Passage of the Amendments to the District of Columbia Marriage and Divorce Act; and a Memorandum on the District of Columbia Domestic Partnership Act.

[18] The "legislative facts" terminology was introduced by Professor Kenneth Culp Davis in An Approach to Problems of Evidence in the Administrative Process, 55 HARV.L.REV. 364, 404-407 (1942). Professor Donald L. Horowitz suggests using the terms "historical facts" and "social facts" in lieu of "adjudicative facts" and "legislative facts," respectively, because the latter terms inaccurately imply a strict division of labor between courts and legislatures. See HOROWITZ, supra, at 45.

[19] See, e.g., D.C.Code §§ 1-622.7 (1992) (public employee health insurance), -623.7 (public employee life insurance), -624.10 (public employee disability insurance), -627.4(6)(E) (public employee retirement benefits); 2-1502 (1994) (power to make anatomical gifts), -1507(a) (power to dispose of body after making anatomical gifts), -2813 (power to claim human remains); 5-825(b) (1994) (priority in leasing real property for business purposes after displacement of business by District of Columbia); 6-1967 (1989) (ability to give informed consent in medical emergency), -2023(a) (ability to request information regarding a mental health client); 10-225(f) (1989) (succession to deceased's retail distributor marketing agreement); 14-309 (1989) (right to claim clergy privilege); 15-503(d) (bankruptcy proceeding rights); 16-308 (1989) (waiver of investigation in adoption of spouse's child), -577 (wage garnishment limits not to apply to spousal support obligations), -910 (power of court to dispose of property on divorce), -911 (alimony pendente lite), -913 (alimony), -916 (maintenance), -1005(b) (right to claim marital privilege to prevent spouse's testimony in domestic relations case from being used in criminal proceedings), -2306 (right to be served with neglect petition for spouse under 18 years of age), -2312(b) (right to notice of minor spouse's hearing for detention or shelter care), -2701 (right to maintain wrongful death action and recover damages), -2921 to -2925 (dower rights); 19-101 to -115 (1989) (rights of survivorship), -301 to -304 (rights of intestacy); 20-303(a) (1989) (priority in appointment as personal representative in probate case); 21-107 (1989) (preference in appointment as guardian of spouse under 14 years of age), -301(11) (spouse included in definition of "member of minor's family" in section dealing with transfer of mentally ill minor), -511 (power to hospitalize mentally ill spouse who is a minor), -522 (right to notice of minor spouse's admission to hospital), -541 (right to petition for hospitalization), -546 (right to request mental health examination), -564 (right to notice of incompetence), -565 (right to statement regarding procedures of release and adjudication for incapacitated spouse); 21-902(c) (1989) (right to notice of apprehension of mentally ill spouse), -2042(a) (right to notice of guardianship proceeding), -2043(c)(1) (priority in serving as guardian for incapacitated spouse), -2055 (court powers to be exercised for benefit of mentally ill person's immediate family), -2057 (preference in serving as conservator), -2210 (priority in giving substituted consent to health care in absence of durable power of attorney); 30-507(c)(7) (1993) (enforceability of spousal support order when part of child support obligation); 35-507(b)(3) (1993) (life insurance benefits), -514(a)(7)(E) (group insurance policy benefits); 36-308(4) (1993) (worker's compensation disability payments); 40-102(e)(4) (1990) (joint ownership of motor vehicles), -703(j)(1) (tax exemption on gift to one spouse of motor vehicle owned by the other spouse); 46-101(19) (1990) (rights of dependents to unemployment compensation); 47-845 (1990) (tax deferral to continue on conveyance of real property to spouse), -1801.4(26) (difference in standard tax deduction for single and married persons), -1805.1(e) (right to file joint tax returns to qualify for tax benefits), -1805.2(2) (status as fiduciary for filing joint tax returns), -1806.2(b) (spousal tax exemption for non-earning spouse), -1806.2(e) (additional spousal exemption for non-earning spouse over 65 years of age).

[20] There are those, including my colleague Judge TERRY, who argue that denial of a marriage license to a same-sex couple is not "discriminatory" because, as demonstrated in Parts II. and III, above, marriage, by definition, has been an institution traditionally limited to heterosexual couples—especially so because child-bearing is at the heart of the institution. That argument begs the question. Appellants, a same-sex couple, proffer the same psychological and physical needs to marry that heterosexual couples have, see Turner v. Safley, 482 U.S. 78, 95-96, 107 S. Ct. 2254, 2265, 96 L. Ed. 2d 64 (1987), and they accordingly proffer that sexual differences should neither affect their right to consummate a marriage in their own way nor automatically withhold the statutory benefits of marriage, including the possibility of having children with the help of surrogates or of adopting abandoned children. This proffer of prejudice from denial of the right to marry makes clear that the marriage statute discriminates in fact against same-sex couples—a claimed impact that is distinguishable from the question whether such discrimination is legal. Put another way, denial of a marriage license to a same-sex couple is "discriminatory" in fact, whether or not it is "invidiously" (unconstitutionally) discriminatory at law. I therefore believe this court cannot properly avoid coming to grips with the constitutional issue appellants have posed simply by defining that issue away as though the prejudicial fact of the alleged discrimination does not exist. Recall (more on this later) that in Loving v. Virginia, 388 U.S. 1, 87 S. Ct. 1817, 18 L. Ed. 2d 1010 (1967), the Supreme Court struck down Virginia's antimiscegenation law against a defense that marriage was intrinsically a same-race institution and thus, as such, was not susceptible to an equal protection challenge. See id. at 3, 7, 87 S. Ct. at 1819, 1821.

[21] Professor Sanford H. Kadish, in Methodology and Criteria in Due Process Adjudication—A Survey and Criticism, 66 YALE L.J. 319, 360 (1957), has summarized the approach:

To some degree data of this kind [social, political, economic and scientific] has reached the Court through lower court direct testimony; but the principal access to it has been through the device of judicial notice, combined with the Court's independent researches or the presentation of factual data by counsel in the form of the Brandeis brief. (footnotes omitted).

Similarly, Professor (later Judge) Jack B. Weinstein, in Judicial Notice and the Duty to Disclose Adverse Information, 51 IOWA L.REV. 807, 822-23 (1951), has addressed consideration of "legislative facts" in the following way:

Information useful to the court in deciding what the rule of law ought to be is supplied in a variety of ways, including the "Brandeis brief" giving extensive factual background materials. The court acts creatively to make new law either in its interpretation of statutes or in extending prior case doctrine. Its problems in determining what is a sound rule are much like those faced by the legislature.

[22] See, e.g., Porter, 618 A.2d at 630-31 (remanding case, on basis of post-trial National Research Council Report, on question whether DNA testing has "requisite consensus" in scientific community); Jones, 548 A.2d at 39-47 (sustaining validity of EMIT drug testing system on basis of expert testimony in trial record of another D.C. Superior Court case, coupled with judicial decisions from other jurisdictions, leaving open question whether appellate court may ascertain social/scientific facts "exclusively by reference to scientific and legal literature, absent relevant expert testimony of record"); Erickson, 574 P.2d at 4-7 (in determining pharmacological properties and psychological impact of cocaine, in connection with deciding whether cocaine was "narcotic" within meaning of criminal statute, appellate court properly could refer to scientific articles not considered by trial court in addition to expert testimony of record).

[23] Whether the judge reviews books or articles proffered by the parties or finds relevant sources in a library, these are not part of the trial record because their authors have not appeared and been subjected to cross-examination. Such materials are therefore appropriately called "extra record" or "non-record" sources of information. See K.C. Davis, supra note 18, at 404-06; 2 McCORMICK, supra, § 331, at 399 n. 16; Erickson, 574 P.2d at 6.

[24] FED.R.EVID. 803(18) exempts from the hearsay rule statements in "treatises, periodicals, or pamphlets on a subject of history, medicine, or other science or art," to the extent they are called to the attention of an expert witness, provided the statements are "read into evidence" and "not ... received as exhibits."

[25] FED.R.EVID. 803(8) exempts from the hearsay rule most "Records, reports, statements, or data compilations" of "public offices or agencies," "unless the sources of information or other circumstances indicate lack of trustworthiness."

[26] In articulating its definition, the Court drew on earlier formulations: "principle of justice so rooted in the traditions and conscience of our people as to be ranked as fundamental," Palko, 302 U.S. at 325, 58 S. Ct. at 152 (quoting Snyder v. Massachusetts, 291 U.S. 97, 105, 54 S. Ct. 330, 332, 78 L. Ed. 674 (1934)); "fundamental principles of liberty and justice which lie at the base of all our civil and political institutions," Palko, 302 U.S. at 328, 58 S. Ct. at 153 (quoting Herbert v. Louisiana, 272 U.S. 312, 316, 47 S. Ct. 103, 104, 71 L. Ed. 270 (1926)).

[27] One treatise has seen the Duncan test as being somewhat broader than Palko's concept of ordered liberty:

This new test meant that the Court would be willing to enforce values which the justices saw as having a special importance in the development of individual liberty in American society, whether or not the value was one that was theoretically necessary in any system of democratic government.

JOHN E. NOWAK, ET AL., CONSTITUTIONAL LAW 455 (1983).

[28] In a footnote, Justice Powell responded to Justice White who, in dissent—although agreeing that Palko's formulation had by then been considered too restrictive—rejected as too broad Moore's reformulation of the kind of liberty interest protected as a fundamental right under the due process clause. See Moore, 431 U.S. at 541, 546, 549-50, 97 S. Ct. at 1957, 1959-60, 1961-62 (White, J., dissenting). According to Justice Powell:

Although he agrees that the Due Process Clause has substantive content, MR. JUSTICE. WHITE in dissent expresses the fear that our recourse to history and tradition will "broaden enormously the horizons of the Clause." Post, at 549-550. To the contrary, an approach grounded in history imposes limits on the judiciary that are more meaningful than any based on the abstract formula taken from Palko Connecticut, 302 U.S. 319 [58 S. Ct. 149, 82 L. Ed. 288] (1937), and apparently suggested as an alternative. Cf. Duncan v. Louisiana, supra, [391 U.S.] at 149-150, n. 14 [88 S.Ct. at 1447-48, n. 14] (rejecting the Palko formula as the basis for deciding what procedural protections are required of a State, in favor of a historical approach based on the Anglo-American legal tradition). Indeed, the passage cited in MR. JUSTICE WHITE's dissent as "most accurately reflect[ing] the thrust of prior decisions" on substantive due process, post, at 545, expressly points to history and tradition as the source for "supplying ... content to this Constitutional concept." Poe v. Ullman, supra, [367 U.S. 497] at 542 [81 S. Ct. 1752, 1776, 6 L. Ed. 2d 989 (1961)] (Harlan, J., dissenting).

Id. at 504 n. 12, 97 S. Ct. at 1938 n. 12.

[29] The constitutional equal protection guarantee, applicable to the states through the Fourteenth Amendment, is extended to the District of Columbia through the due process clause of the Fifth Amendment. See Bolling v. Sharpe, 347 U.S. 497, 74 S. Ct. 693, 98 L. Ed. 884 (1954).

[30] After the trial court ruled, the District's statute criminalizing even consensual sodomy by adults was amended "to eliminate consensual sexual acts between persons who are above the existing age of consent." See Right to Privacy Amendment Act of 1993, D.C.Law, 10-14, 40 D.C.Reg. 3007 (1993); D.C.Code § 22-3502 (1994 Supp.).

[31] Although the class at issue is limited to adult homosexual couples who (in order to avoid the unchallenged consanguinity bar) are not related, this case, fundamentally, deals with the question whether homosexuals are entitled to "suspect" or "quasi-suspect" class status under equal protection analysis. For convenience, therefore, I shall refer simply to the rights of homosexuals or of gays and lesbians, not more narrowly to unrelated adult homosexual couples. No party has suggested that, because this is a marriage case, the constitutional analysis of the protected class at issue is somehow narrower than in other cases where sexual orientation receives constitutional consideration. If there were any significance to the fact that only a subclass of homosexuals is involved here, that presumably would have received attention by the government in defending the marriage statute.

[32] Appellants have not argued that same-sex marriage, though not a "fundamental right" under the due process clause, is a "fundamental right" for equal protection purposes. I therefore do not address that possibility. Professor Sunstein has adverted to this question, pointing out possible distinctions between "fundamental rights" in due process and equal protection analyses, in Cass R. Sunstein, Sexual Orientation and the Constitution: A Note on the Relationship Between Due Process and Equal Protection, 55 U.CHI.L.REV. 1161, 1168-69 (1988).

[33] First, appellants stress that the need for emotional support for an intimate relationship, as well as the need for an official blessing having religious or spiritual significance for the relationship, is self-evidently not limited to heterosexuals.

Next, appellants contended in the trial court that physical consummation of a same-sex marriage can occur without violating the sodomy law, through mutual masturbation. On appeal, appellants note that the law criminalizing adult, consensual sodomy has been repealed. See supra note 30. Accordingly, we judicially know that, in the District, couples of marriageable age may now lawfully consummate their unions, if they choose to do so, through sodomy.

Finally, appellants have catalogued what we can judicially know: the D.C.Code confers on married couples a substantial number of benefits not available to unmarried couples. See supra note 19.

[34] In this opinion I do not consider whether same-sex couples, if not foreclosed from marriage, are entitled to constitutional protection for the exercise of various derivative actions that would affect third parties, such as adoption of children.

[35] See Loving v. Virginia, 388 U.S. at 11, 87 S. Ct. at 1823.

[36] See Graham v. Richardson, 403 U.S. at 372, 91 S. Ct. at 1852.

[37] See Oyama v. California, 332 U.S. 633, 644-46, 68 S. Ct. 269, 274-75, 92 L. Ed. 249 (1948); Korematsu v. United States, 323 U.S. 214, 216, 65 S. Ct. 193, 194, 89 L. Ed. 194 (1944); Hirabayashi v. United States, 320 U.S. 81, 100, 63 S. Ct. 1375, 1385, 87 L. Ed. 1774 (1943).

[38] See Mississippi University for Women v. Hogan, 458 U.S. 718, 724, 102 S. Ct. 3331, 3336, 73 L. Ed. 2d 1090 (1982); Craig v. Boren, 429 U.S. 190, 197, 97 S. Ct. 451, 456-57, 50 L. Ed. 2d 397 (1976); but see Frontiero v. Richardson, 411 U.S. at 682, 93 S. Ct. at 1768 (plurality opinion) (classifications based on sex are "inherently suspect").

[39] See Mills v. Habluetzel, 456 U.S. 91, 99, 102 S. Ct. 1549, 1554-55, 71 L. Ed. 2d 770 (1982); Lalli v. Lalli, 439 U.S. 259, 265, 99 S. Ct. 518, 523, 58 L. Ed. 2d 503 (1978); Trimble v. Gordon, 430 U.S. 762, 767, 97 S. Ct. 1459, 1463-64, 52 L. Ed. 2d 31 (1977); Mathews v. Lucas, 427 U.S. 495, 510, 96 S. Ct. 2755, 2764-65, 49 L. Ed. 2d 651 (1976).

[40] "The Court has never provided a coherent explanation of the characteristics which, either overtly or covertly, trigger intermediate review." LAWRENCE H. TRIBE, AMERICAN CONSTITUTIONAL LAW 1614 (2d ed. 1988).

[41] One federal district court has held that homosexuals and bisexuals comprise a quasi-suspect class. See Equality Foundation of Greater Cincinnati v. Cincinnati, 860 F. Supp. 417 (S.D.Ohio 1994) (holding, among other things, that gays, lesbians, and bisexuals comprise quasi-suspect class, with result that any law discriminating against class would have to be substantially tailored to sufficiently important governmental interest; and, applying that test, permanently enjoining voter-enacted city charter amendment that would prohibit adoption or enforcement of any protection based on sexual orientation, status, conduct, or relationship, since amendment would violate equal protection clause, having no legitimate governmental purpose). See also Watkins v. United States Army, 875 F.2d 699, 724-28 (9th Cir.1989) (en banc) (Norris, J. concurring) (opining that homosexuals comprise a suspect class). See generally Homosexuals' Right to Marry, Note, 128 U.PA.L.REV. at 202-06; TRIBE, supra, at 1614-18; Note, The Constitutional Status of Sexual Orientation: Homosexuality as a Suspect Classification, 98 HARV.L.REV. 1285, 1297-1305 (1985); Harris M. Miller II, Note, An Argument for the Application of Equal Protection Heightened Scrutiny to Classifications Based on Homosexuality, 57 S.CAL.L.REV. 797, 812-16 (1984); John C. Hayes, The Tradition of Prejudice Versus the Principle of Equality: Homosexuals and Heightened Equal Protection Scrutiny After Bowers v. Hardwick, 31 B.C.L.REV. 375, 405-24 (1990); Kevin A. Zambrowicz, Comment, "To Love and Honor All The Days of Your Life": A Constitutional Right To Same-Sex Marriage, 43 Cath.U.L.Rev. 907, 935-40 (1994).

[42] See Lyng v. Castillo, 477 U.S. 635, 638, 106 S. Ct. 2727, 2729, 91 L. Ed. 2d 527 (1986) (primary household's close relatives, excluded from federal food stamp program, "have not been subjected to discrimination"); Cleburne, 473 U.S. at 443, 105 S. Ct. at 3256 (mentally retarded not victims of "continuing antipathy or prejudice"); Murgia, 427 U.S. at 313, 96 S. Ct. at 2567 (police officers over age 50, and the aged generally, have not experienced "`history of purposeful unequal treatment'"); San Antonio School Independent Dist. v. Rodriguez, 411 U.S. 1, 28, 93 S. Ct. 1278, 1294, 36 L. Ed. 2d 16 (1973) (minority and poor schoolchildren not subjected to "history of purposeful unequal treatment").

[43] Cleburne, 473 U.S. at 440, 105 S. Ct. at 3254 (statutory classifications by race alienage, and national origin are "deemed to reflect prejudice and antipathy"); Mississippi University for Women, 458 U.S. at 725, 102 S. Ct. at 3336 ("Care must be taken in ascertaining whether the statutory objective itself reflects archaic and stereotypic notions."); Murgia, 427 U.S. at 313, 96 S. Ct. at 2566 (class of police officers over age 50, "have not experienced a "history of purposeful unequal treatment, or been subjected to unique disabilities on the basis of stereotyped characteristics not truly indicative of their abilities").

[44] Lyng, 477 U.S. at 638, 106 S. Ct. at 2729 (close relatives "do not exhibit obvious, immutable, or distinguishing characteristics that define them as a discrete group"); Cleburne, 473 U.S. at 442, 105 S. Ct. at 3255-56 (mentally retarded are different from other people, "immutably so, in relevant respects"); Plyler, 457 U.S. at 220, 102 S. Ct. at 2396 (children of illegal aliens, unlike their parents, have "legal characteristic[s] over which children can have little control"); Lucas, 427 U.S. at 505, 96 S. Ct. at 2762 (1976) (legal status of illegitimate children "is, like race or natural origin, a characteristic determined by causes not within the control of the illegitimate individual"); Frontiero, 411 U.S. at 686, 93 S. Ct. at 1770 ("sex, like race and natural origin, is an immutable characteristic determined solely by the accident of birth").

[45] Lyng, 477 U.S. at 638, 106 S. Ct. at 2729 (close relatives of primary household are "not a minority or politically powerless"); Cleburne, 473 U.S. at 445, 105 S. Ct. at 3257 (lawmakers have been adopting remedial legislation in manner that "negates any claim that the mentally retarded are politically powerless in the sense that they have no ability to attract the attention of the lawmakers"); Rodriguez, 411 U.S. at 28, 93 S. Ct. at 1294 (class of minority and poor schoolchildren has not been "relegated to such a position of political powerlessness as to command extraordinary protection from the majoritarian political process"); see also Plyler, 457 U.S. at 216 n. 14, 102 S. Ct. at 2394 n. 14 (quoting Rodriguez); Murgia, 427 U.S. at 313, 96 S. Ct. at 2566-67 (quoting Rodriguez).

[46] Indeed, "any attempt to criminalize the status of an individual's sexual orientation would present grave constitutional problems." Watkins, 875 F.2d at 725 (Norris, J., concurring) (citing Robinson v. California, 370 U.S. 660, 82 S. Ct. 1417, 8 L. Ed. 2d 758 (1962)). See also Rowland, 470 U.S. at 1016, 105 S. Ct. at 1378 (Brennan, J., dissenting from denial of certiorari) (even if adverse state action based on homosexual conduct were valid under equal protection principles, "such approval would not answer the question, posed here, whether the mere nondisruptive expression of homosexual preference can pass [constitutional] muster").

Recently, in Steffan v. Perry, 41 F.3d 677 (D.C.Cir.1994) (en banc), the United States Court of Appeals for the District of Columbia Circuit sustained the constitutionality of regulations pursuant to which a former Navy midshipman was discharged from the Naval Academy after admitting he was a homosexual. A majority of the court concluded that the government had not punished Steffan for homosexual status but, rather, for inferable conduct—homosexual acts—attributable to his announced sexual orientation. The dissenters, rejecting the inference of conduct from sexual orientation alone, found unconstitutional punishment based solely on sexual orientation. For purposes of the case, Steffan had conceded that the government had a legitimate interest in keeping individuals who engage, or intend to engage, in homosexual conduct out of the military; accordingly, he did not contest the court's equal protection review under the rational basis test.

[47] Judge Norris has offered a useful hypothetical, based on the facts of Moore v. City of East Cleveland, supra, that makes clear why due process analysis does not dictate an equal protection result.

Suppose a city passed a "single family occupancy" housing ordinance allowing only members of the immediate, nuclear family to live in the same house. Suppose further that a disproportionate number of black families in the community lived together in extended families that included, for example, cousins and grandparents. Finally, suppose the ordinance was motivated by a racially discriminatory purpose. A black family challenging the ordinance could raise a due process claim, arguing that the ordinance impermissibly intruded on "deeply rooted" family traditions. In real life, the Court found such a due process claim persuasive. But suppose the Court had rejected the due process claim. Suppose the Court had instead agreed with the city of East Cleveland that the privacy interests protected by the Constitution do not include extended family relationships—that the due process clause does not "give grandmothers any fundamental rights with respect to grandsons." In that event, the black family could still challenge the ordinance on equal protection grounds, arguing that the ordinance discriminated against blacks. Could anyone seriously maintain that the Court's hypothetical refusal to give due process protection to "extended family" living would have any bearing on the black family's equal protection claim? Of course not.

Watkins, 875 F.2d at 719 (Norris, J., concurring) (footnotes omitted).

[48] See Sexual Orientation and the Law, supra, at 1567 (the stereotypes "that gay men and lesbians are more likely to molest children, that they cannot be trusted to keep secrets, and that homosexuality is `contagious' ... have repeatedly been proven false") (footnotes omitted); see generally, Sharon B. Gurwitz & Melinda Marcus, Effects of Anticipated Interaction, Sex, and Homosexual Stereotypes on First Impressions, 8 J. APPLIED SOC. PSYCHOL. 47, 54 (1978) (In an experiment testing 96 college students' response to a questionnaire and videotape interview of a fellow student, "the Gay male was rated as having less positive personality traits than the Straight male as well as possessing stereotype-related traits to a greater extent."); Mary E. Kite & Kay Deaux, Attitudes Toward Homosexuality: Assessment and Behavioral Consequences, 7 BASIC & APPLIED SOC. PSYCHOL. 137, 156 (June 1986) ("Persons who are aware of a person's homosexuality are more negative toward homosexuals than are persons who are unaware, regardless of their attitude toward homosexuality.").

[49] See, e.g., Laura S. Allen & Roger A. Gorski, Sexual Orientation and the Size of the Anterior Commissure in the Human Brain, 89 PROC. NAT'L ACAD. SCI. U.S.A. 7199 (1992) (Neuroanatomical study reveals significant differences between homosexual and heterosexual men in the structure of the anterior commissure region of the brain); J. Michael Bailey & Richard C. Pillard, A Genetic Study of Male Sexual Orientation, 48 ARCH. GEN. PSYCHIATRY 1089, 1093 (Dec.1991) (Twin study suggests that "genetic factors are important in determining individual differences in sexual orientation"); Dean H. Hamer, et al., A. Linkage Between DNA Markers on the X Chromosome and Male Sexual Orientation, 261 SCIENCE 321, 321 (July 16, 1993) (Pedigree and linkage analysis indicates "a statistical confidence interval of more than 99 percent that at least one subtype of male sexual orientation is genetically influenced"); Simon LeVay, A Difference in Hypothalamic Structure Between Heterosexual and Homosexual Men, 253 SCIENCE 1034 (Aug.1991) (Neuroanatomical study reveals significant differences between homosexual and heterosexual men in the structure of the anterior hypothalamus region of the brain); D.F. Swaab & M.A. Hofman, An Enlarged Suprachiasmatic Nucleus in Homosexual Men, 537 BRAIN RESEARCH 141 (1990) (Neuroanatomical study reveals significant differences between homosexual and heterosexual men in the structure of the suprachiamatic nucleus region of the brain); see also ALAN P. BELL ET AL., SEXUAL PREFERENCE: ITS DEVELOPMENT IN MEN AND WOMEN 212-15, 216 (1981) (sexual preference formed at an early age; citing genetic and hormonal factors, "our findings are not inconsistent with what one would expect to find if, indeed, there were a biological basis for sexual preference"); Ray B. Evans, Physical and Biochemical Characteristics of Homosexual Men, 39 J. CONSULTING & CLIN. PSYCHOL. 140, 146 (1972) ("the differences in physical characteristics and in the biochemical values [of heterosexual and homosexual males] support the general thesis of a biological factor in the etiology of homosexuality in males"); Richard Green, The Immutability of (Homo)sexual Orientation: Behavioral Science Implications for a Constitutional (Legal) Analysis, 16 J. Psychiatry & Law 537 (1989) (comprehensively examines twin studies, family studies, hormonal influence studies, brain studies, and various treatments for homosexuality to conclude that homosexuality is immutable for equal protection purposes); Martin Hoffman, Homosexuality, in HUMAN SEXUALITY IN FOUR PERSPECTIVES 164-189 (Frank Beach, ed., 1976) (therapy should be used to help homosexuals accept themselves rather than to "cure" them because past efforts have shown that homosexuality is not curable in a large part of the homosexual population and therapists harm homosexual patients by making them believe that they are "sick"); James D. Weinrich, Is Homosexuality Biologically Natural?, in HOMOSEXUALITY SOCIAL, PSYCHOLOGICAL AND BIOLOGICAL ISSUES 197-208 (William Paul, et al., eds., 1982) ("human homosexuality is ... as biologically natural as is human heterosexuality").

[50] BELL ET AL., supra note 49, at 211. These authors from the Alfred C. Kinsey Institute for Sex Research elaborated:

Another notable difference we found has to do with whether the respondents were bisexual or exclusively homosexual. Exclusive homosexuality seemed to be something that was firmly established by the end of adolescence and relatively impervious to change or modification by outside influences. For the bisexuals, by contrast, a homosexual preference seemed to emerge later and to be more tied to learning and social experiences. These findings may well have implications for whatever help certain homosexuals seek to resolve their own ambiguities or feelings of guilt about their homosexuality. To therapists, we would suggest that exclusive homosexuality probably is so deeply ingrained that one should not attempt or expect to change it. Rather, it would probably make far more sense simply to recognize it as a basic component of a person's core identity and to help the client develop more-positive feelings about and respect for his or her sexual proclivities.

Id. (emphasis added).

[51] See WILLIAM MASTERS & VIRGINIA JOHNSON, HOMOSEXUALITY IN PERSPECTIVE 411 (1979) (critiqued in BELL., ET AL., supra note 49, at 213).

[52] Philip Blumstein & Pepper Schwartz, Intimate Relationships and the Creation of Sexuality, in HOMOSEXUALITY/HETEROSEXUALITY: CONCEPTS OF SEXUAL ORIENTATION 307-320, 309 (David P. McWhirter, et. al. eds. 1990) (asserting that "fundamental categorical desire" for sexual relations with persons "of only one specific gender" may not exist, and that "it is the culture that creates understandings about how people are sexual and determines whether people will be able to have only one sexual focus, to eroticize both sexes, or to experience categorical desire for one sex at one point in their lives and categorical desire for the other sex at another point in their lives").

[53] See WILLIAM AARON, STRAIGHT (1972) (autobiographical description of transition from homosexual to heterosexual lifestyle and marriage through commitment to Christianity); ALBERT ELLIS & ROBERT A. HARPER, A NEW GUIDE TO RATIONAL LIVING 52-59 (1975) (concluding that passive homosexuality is the result of an Oedipal complex with a corresponding fear of castration and that reindoctrination is possible with "better information and clearer thinking"); id. at 189-93 (describing homosexual's treatment through change in attitude); B.H. Fookes, Some Experiences in the Use of Aversion Therapy in Male Homosexuality, Exhibitionism, and Fetish-Transvestism, 115 BRIT, J. PSYCHIATRY 339 (1969) (six of nine males who had not experienced heterosexual intercourse prior to shock aversion therapy demonstrated decreased homosexual arousal and engaged in heterosexual coitus at three years' follow-up); William Freeman & Robert G. Meyer, A Behavioral Alteration of Sexual Preference in the Human Male, 6 BEHAV. THERAPY 206 (1975) (all nine subjects free of homosexual behavior one year after aversive shock treatment, and seven subjects free of homosexual behavior eighteen months after treatment); E. Mansell Pattison & Myrna Loy Pattison, "Ex-Gays": Religiously Mediated Change in Homosexuals, 137 AM.J.PSYCHIATRY 1553, 1562 (1980) (study of 11 of 30 homosexual men who changed their sexual orientation through Pentecostal church conversion concludes that "when homosexuality was defined as a changeable condition, it appears that change was possible"): KENT PHILPOTT, THE THIRD SEX? SIX HOMOSEXUALS TELL THEIR STORIES 172, 178-200 (1975) (concluding that homosexuality is "a choice to be and do what was not intended" and that people can be cured of homosexual temptations by turning to counselling and to Christianity); D.J. WEST, HOMOSEXUALITY RE-EXAMINED 251 (1977) (use of pornography and group sex may help troubled homosexuals get aroused in heterosexual situations); id. at 261 (aversion therapy which involved showing homosexuals erotic homosexual pictures and interrupting pleasure with unpleasant, punishing sensations to discourage them from homosexual arousal enjoyed considerable vogue in the thirties and forties).

[54] Two federal circuit courts of appeals have simply asserted, as though it were self-evident, that homosexuality is not immutable. See High Tech Gays, 895 F.2d at 573 ("Homosexuality is not an immutable characteristic; it is behavioral and hence is fundamentally different from traits such as race, gender, or alienage."); Woodward, 871 F.2d at 1076 ("Members of recognized suspect or quasi-suspect classes, e.g., blacks or women, exhibit immutable characteristics, whereas homosexuality is primarily behavioral in nature."). Not only do these cases fail to cite any authority for their conclusions, but also they were written before the most recent scientific studies concluding that sexual orientation appears to have a genetic basis. See supra note 49.

[55] In two cases where the classes at issue had immutable characteristics, the Supreme Court found reason not to apply strict or intermediate scrutiny to the legislation under attack. In Cleburne, which considered the constitutionality of a zoning ordinance that required permits for group homes for the mentally retarded, the Supreme Court struck down the ordinance under the rational basis test after rejecting the court of appeals' conclusion that mental retardation was a "quasi-suspect" classification. The Court did not dispute the immutability of mental retardation, see Cleburne, 473 U.S. at 442, 105 S. Ct. at 3255, but concluded that lawmakers took a benign view of the class that "negate[d] any claim that the mentally retarded are politically powerless" and "belie[d] a continuing antipathy or prejudice and a corresponding need for more intrusive oversight by the judiciary." Id. at 443, 445, 105 S. Ct. at 3256, 3257. In Murgia, in considering a state law mandating retirement of police officers at age 50, the Court sustained the legislation under the rational basis test, see Murgia, 427 U.S. at 314, 96 S. Ct. at 2567, concluding that "even old age does not define a `discrete and insular group' in need of `extraordinary protection from the majoritarian political process.'" Id. at 313, 96 S. Ct. at 2567 (citation omitted). Cleburne and Murgia, of course, are cases in which the Court found none of the relevant factors, except immutability, present for evaluating strict or intermediate scrutiny; in the other cases reflecting immutability, see supra notes 35-39, more stringent scrutiny was required because the factors, taken as a group, so indicated.

[56] Health Care Benefits Expansion Act of 1992, 39 D.C.Reg. 2861 (1992), codified in D.C.Code §§ 36-1401 to -1408 (1993 Repl.).

[57] See D.C. Supplemental Appropriations and Rescissions Act, Pub.L. No. 102-382, 106 Stat. 1422, 1422 (1992) (Congress prevented District from funding Health Care Benefits Expansion Act of 1992 in fiscal year 1993); D.C. Appropriations Act, Pub.L. No. 103-127, 107 Stat. 1336, 1349 (1994) (Congress prevented District from funding Health Care Benefits Expansion Act of 1992 in fiscal year 1994).

[58] There undoubtedly are those who will argue from principle, based on religious or other moral premises, that if there is any evidence therapy— even traumatic therapy—will succeed in removing one's tendency toward homosexuality, that is to be preferred—in the name of public morality—over legal legitimation of "unnatural" behavior proscribed, indeed criminally proscribed, throughout American history. Putting aside the observation that the need for traumatic therapy to reverse homosexuality would indicate homosexuality is not necessarily "unnatural," I will simply say that one person's principle is, from an alleged victim's viewpoint, unwarranted prejudice. I cannot imagine any constitutional basis for allowing the government to discriminate against homosexuals in any way premised on an allegation, coupled with a finding, that homosexuality is not immutable simply because traumatic therapy can change it. In context, that would be the equivalent of using "the rack and the screw"—torture society cannot use even to punish its worst criminals, consistent with the Eighth Amendment. Rochin v. California, 342 U.S. 165, 172, 72 S. Ct. 205, 209-10, 96 L. Ed. 183 (1952).

[59] The footnote supporting this conclusion is as follows:

192. See Green, Sexual Identity of 37 Children Raised by Homosexual or Transsexual Parents, 135 AM.J.PSYCHIATRY 692, 696 (1978) (reporting normal behavior and typical aspirations among the subject children); Green, The Best Interests of the Child with a Lesbian Mother, 10 BULL.AM.ACAD.PSYCHIATRY & L. 7, 14 (finding "no significant [gender identity] differences for the boys or the girls in either [the heterosexual or lesbian] set of families"); Cohen, Children of Homosexuals Seem Headed Straight, PSYCHOLOGY TODAY, NOV. 1978, AT 44-45; Hoeffer, Children's Acquisition of Sex-Role Behavior in Lesbian-Mother Families, 51 AM.J.ORTHOPSYCHIATRY 536, 542 (1981) (noting no significant difference in the acquisition of sex-role traits between the children of lesbian and heterosexual mothers and hypothesizing that children's peers have the greatest influence on their sex-role development); Weeks, Two Cases of Children of Homosexuals, 6 CHILD PSYCHIATRY & HUM. DEV. 26-32 (1975) (finding it impossible to distinguish specific aspects of the children's development that are directly related to their parents' sexuality); Note, The Avowed Lesbian Mother and Her Right to Child Custody: A Constitutional Challenge That Can No Longer Be Denied, 12 SAN DIEGO L.REV. 799, 861 (1975) (quoting psychiatrist George Weinberg):

Most homosexuals have had parents who are exclusively heterosexual or primarily so. As this fact suggests, homosexual men and women do not learn their sexual preferences by watching the sexual activities of their parents.... The occasional concern that a homosexual parent will rear homosexual children is unwarranted by the evidence. See also CALIFORNIA COMM'N ON PERSONAL PRIVACY, REPORT OF THE COMM'N ON PERSONAL PRIVACY 364 (it is as likely that the left-handed minority will "convert" members of the right-handed majority as it is that homosexuals can "convert" heterosexuals).

Other literature also supports this conclusion. See Susan Golombok et al., Children in Lesbian and Single-Parent Households: Psychosexual and Psychiatric Appraisal, 24 J. CHILD PSYCHOL. PSYCHIAT. 4:551 (1983); Richard Green et al., Lesbian Mothers and Their Children: A Comparison with Solo Parent Heterosexual Mothers and Their Children, 15 ARCHIVES OF SEXUAL BEHAV. 2:167 (1986); Mary B. Harris & Pauline H. Turner, Gay and Lesbian Parents, 12 J.HOMOSEXUALITY 2:101 (1986).

[60] The legislative history of the Marriage and Divorce Act of 1977, D.C.Stat. 119, shows that when Councilmember Dixon substituted the bill adopted as that Act, after withdrawing Bill 1-89, which would have authorized same-sex marriages, he did so in response to an intense debate between gay rights activists and opponents of Bill 1-89 led by the Catholic Archdiocese of Washington. See supra note 6. The Council of the District of Columbia itself, therefore, never rejected a same-sex marriage bill, and, in any event, there was no discussion or finding of a compelling or substantial state interest in barring same-sex marriages.

[61] Homosexual orientation, as such, is not antisocial; homosexual behavior, however—like heterosexual behavior—may, or may not, be antisocial. It is therefore not unlikely that the heterosexual majority would characterize a predictable increase of homosexual behavior of one kind or another as an increase in antisocial behavior, justifying legislation precisely tailored to deter, if not prevent, that behavior. The public majority, for example, might claim a substantial, or even compelling, state interest in withholding from same-sex couples the right to marry, in order to prevent creation of role models who, merely by virtue of a legitimate marriage, might encourage bisexual experimentation that could unsettle mind and body to the point, eventually, of making stable marriage relationships less likely. If such reasoning were proffered, the government would have the burden of demonstrating through credible evidence that the danger was real enough to justify the majority's claimed interest. Cf. Evans II, supra. On the other hand, if the public majority were to justify barring same-sex marriages simply because they "give the wrong message" or "offend public morality" or "debase our religious and cultural traditions," that is not the kind of concrete harm, reflecting an important or compelling state interest, that would justify government discrimination against homosexual marriage. Cf. id. The question whether an increase in homosexual behavior would amount to an increase of antisocial behavior, therefore, is a contextual one, not self-evidently answerable in the abstract.

[62] In concrete terms, the question theoretically might be: would the government have a substantial, or even compelling, state interest in barring same-sex marriages if the best evidence available tended to prove, for example, that there was a 25% to 50% chance that same-sex marriages, if amounting to 3% of all formalized marriages in a community, would influence .5% of all children (but probably no more) in the community before age 12 to experiment with homosexual behavior, if not settle on that sexual orientation? Evidence presented in such precise terms may not be forthcoming. I postulate it only to show as precisely as possible the kind of inquiry I believe is required in trying to formulate the state's interest. As indicated earlier, the state cannot characterize its interest as "substantial" or "compelling" merely by reference to what offends majority values, without concrete accompanying harm.

[63] Nothing, moreover, would preclude the parties or the trial court from referencing these other trial records for relevant expert testimony. See Jones, 548 A.2d at 45-46.

[64] In Loving, 388 U.S. at 3, 87 S. Ct. at 1819, the trial court had said:

Almighty God created the races white, black, yellow, malay and red, and he placed them on separate continents. And but for the interference with his arrangement there would be no cause for such marriages. The fact that he separated the races shows that he did not intend for the races to mix.

In sustaining this position, the Supreme Court of Appeals of Virginia had relied on its decision in Naim v. Naim, 197 Va. 80, 87 S.E.2d 749 (1955) (en banc), where the court had said, with reference to an Indiana decision upholding an antimiscegenation statute:

It was said in that case that the question was one of difference between the races, not of superiority or inferiority, and that the natural law which forbids their intermarriage and the social amalgamation which leads to a corruption of races is as clearly divine as that which imparted to them different natures.

Id. 87 S.E.2d at 752.

[65] In discussing the Fourteenth Amendment, the Court said:

The object of the amendment was undoubtedly to enforce the absolute equality of the two races before the law, but in the nature of things it could not have been intended to abolish distinctions based upon color, or to enforce social, as distinguished from political equality, or a commingling of the two races upon terms unsatisfactory to either. Laws permitting, and even requiring, their separation in places where they are liable to be brought into contact do not necessarily imply the inferiority of either race to the other, and have been generally, if not universally, recognized as within the competency of the state legislatures in the exercise of their police power. The most common instance of this is connected with the establishment of separate schools for white and colored children, which has been held to be a valid exercise of the legislative power even by courts of States where the political rights of the colored race have been longest and most earnestly enforced.

Plessy, 163 U.S. at 544, 16 S. Ct. at 1140 (emphasis added).

In response to the Thirteenth Amendment, the Court had a similar view:

A statute which implies merely a legal distinction between the white and colored races—a distinction which is founded in the color of the two races, and which must always exist so long as white men are distinguished from the other race by color—has no tendency to destroy the legal equality of the two races, or reestablish a state of involuntary servitude.

Id. at 543, 16 S. Ct. at 1140 (emphasis added).

[1] The same could be said about the due process section of Judge Ferren's opinion, part V, although I agree with him that the Supreme Court has made clear that "marriage" between two persons of the same sex is not a fundamental right, and hence that the denial of a right to enter into such a "marriage" is not a violation of due process.

[2] Judge Ferren suggests that my rationale is "akin to" the discredited notion that "a divine natural order forbids racial intermarriage," a notion which the Supreme Court quite properly laid to rest in Loving v. Virginia, 388 U.S. 1, 87 S. Ct. 1817, 18 L. Ed. 2d 1010 (1967). With all respect, Judge Ferren reads too much into this opinion. As a mere judge, I claim no ability to fathom the "natural order," and any discussion of a divinity that shapes its ends is far beyond the scope of this opinion or the competence of its author. All I am saying is that the very nature of the relationship that we call marriage, as it has been recognized and defined for centuries—indeed, millennia—necessarily excludes two persons of the same sex from entering into that relationship. I leave the theological issues to the theologians; the legal and constitutional issues that this case presents are difficult enough.

[1] As the Supreme Court of Hawaii has noted, just as not all opposite-sex marriages are between heterosexuals, not all same-sex marriages would necessarily be between homosexuals. Baehr v. Lewin, 852 P.2d 44, 51 n. 11 (Haw.), reconsideration granted in part, 74 Haw. 650, 875 P.2d 225 (1993). Cf., e.g., the statute struck down in Loving v. Virginia, 388 U.S. 1, 87 S. Ct. 1817, 18 L. Ed. 2d 1010 (1967), which specifically employed the suspect classification of race.

[2] I have a somewhat similar difficulty with the argument that discrimination on the basis of gender is involved here. The marriage statute applies equally to men and women. It seems to me to stretch the concept of gender discrimination to assert that it applies to treatment of same-sex couples differently from opposite-sex couples. This argument was rejected in Singer v. Hara, 11 Wash.App. 247, 522 P.2d 1187, 1196 review denied, 84 Wash.2d 1008 (1974), which discussed and distinguished Loving v. Virginia, supra, outlawing the ban on interracial marriages. (The Hawaii Supreme Court took a somewhat different view in ruling on the application of the gender discrimination clause of the Hawaii Constitution in Baehr, supra note 1.) In any event, appellants' broad argument is plainly focussed on the impact of the marriage statute upon homosexuals, not upon men and women in general, and, in this context, it would seem somewhat unrealistic to use gender discrimination as the basis for the analysis.

[3] Discrimination on the basis of gender calls for an analysis based on quasi-suspect status, see Craig v. Boren, 429 U.S. 190, 197, 97 S. Ct. 451, 456-57, 50 L. Ed. 2d 397 (1976), and I can perceive no warrant under current constitutional doctrine and present factual knowledge for placing homosexuality in a status calling for greater scrutiny. Indeed, as Judge FERREN points out, all federal appellate courts that have considered the subject have refused to recognize homosexuality even as much as a quasi-suspect class, even with respect to statutes or regulations addressed solely to that status.

[4] See Turner v. Safley, 482 U.S. 78, 95-96, 107 S. Ct. 2254, 2265, 96 L. Ed. 2d 64 (1987). I do not think that the existence of other aspects of marriage with which the state cannot impermissibly interfere negates the importance of the basic considerations expressed, e.g., in Zablocki v. Redhail, 434 U.S. 374, 386, 98 S. Ct. 673, 681, 54 L. Ed. 2d 618 (1978), and Skinner v. Oklahoma, 316 U.S. 535, 541, 62 S. Ct. 1110, 1113, 86 L. Ed. 1655 (1942).

[5] The sexual aspect of opposite-sex marriage would appear to relate in a similar way to same-sex marriage between homosexuals. If, as it may appear from Bowers v. Hardwick, 478 U.S. 186, 106 S. Ct. 2841, 92 L. Ed. 2d 140 (1986), the state could ban the commission of acts presumably to be expected in such a same-sex relationship, it is difficult to understand on what basis the state constitutionally could be forced to extend the recognition of marriage to that relationship, whatever view it might otherwise take of such acts. Cf. Steffan v. Aspin, 303 U.S.App.D.C. 404, 409-10, 8 F.3d 57, 62-63 (1993) ("we have noted the anomaly of according special protection to a class whose defining characteristic, homosexual conduct, can be made illegal"), judgment vacated, 41 F.3d 677 (D.C.Cir.1994) (en banc). In the marriage context, the much-mooted distinction between homosexual orientation and homosexual acts does not seem particularly relevant.

It is, of course, true that not all opposite-sex marriages involve procreation or even sexual relations and conversely that those activities occur without marriage. In the leading case of Baker v. Nelson, the Minnesota Supreme Court rejected this consideration thus:

Petitioners note that the state does not impose upon heterosexual married couples a condition that they have a proved capacity or declared willingness to procreate, posing a rhetorical demand that this court must read such condition into the statute if same-sex marriages are to be prohibited. Even assuming [I think the Minnesota court here makes a massive assumption] that such a condition would be neither unrealistic nor offensive under the Griswold rationale [Griswold v. Connecticut, 381 U.S. 479, 85 S. Ct. 1678, 14 L. Ed. 2d 510 (1965)], the classification is no more than theoretically imperfect. We are reminded, however, that `abstract symmetry' is not demanded by the Fourteenth Amendment.

291 Minn. 310, 191 N.W.2d 185, 187 (1971) (footnote omitted). The court then cites to several cases and quotes from Skinner v. Oklahoma, supra, 316 U.S. at 540, 62 S. Ct. at 1113: "The Constitution does not require things which are different in fact or opinion to be treated in law as though they were the same." 191 N.W.2d at 187 n. 4.

[6] See 3 RONALD D. ROTUNDA & JOHN E. NOWAK, TREATISE ON CONSTITUTIONAL LAW § 18.23 (2d ed. 1992) and cases cited.

[7] To the extent that appellants attack not the institution of marriage as recognized by statute but rather the special considerations given to that status by particularized statutes, as detailed by Judge FERREN in his opinion, it seems to me clear that appellants cannot claim the protection of the equal protection clause on the basis of homosexuals as a quasi-suspect class. Most if not all of such statutes "adversely" affect all unmarried couples of whatever status, and presumably would pass the rational relation test normally used in equal protection analysis. See, e.g., Pennell v. City of San Jose, 485 U.S. 1, 14, 108 S. Ct. 849, 858-59, 99 L. Ed. 2d 1 (1988). In any event, each such statute would have to be addressed on an individual basis.

[8] See Jones v. Hallahan, 501 S.W.2d 588, 590 (Ky.Ct.App.1973) (holding that two females are not capable of entering a marriage and therefore the denial of a marriage license to same-sex couples infringes upon no constitutional rights); Baker, supra, 191 N.W.2d at 186-87 (holding that the state's marriage statute did not authorize same-sex marriages and rejecting the plaintiffs' claims that the prohibition of same-sex marriages violated the Fourteenth Amendment's Equal Protection and Due Process Clauses); In re Cooper, 187 A.D.2d 128, 592 N.Y.S.2d 797, 799-800 (holding that a same-sex partner was not a "surviving spouse" and therefore had no right of election against a decedent's will and that exclusion of same-sex partners did not violate equal protection), appeal dismissed, 82 N.Y.2d 801, 604 N.Y.S.2d 558, 624 N.E.2d 696 (1993); Singer, supra, 522 P.2d at 1196-97 (holding that the state's refusal to issue a marriage license to a female-female couple did not violate equal protection under the state or federal constitution). A number of cases are extant rejecting other arguments relating to same-sex marriages. See, e.g., Jennings v. Jennings, 20 Md. App. 369, 315 A.2d 816, 820 n. 7 (1974) (noting that Maryland does not recognize same-sex marriages); Gajovski v. Gajovski, 81 Ohio App. 3d 11, 610 N.E.2d 431, 433 (1991) (holding that a female cannot live in concubinage with another woman because Ohio does not recognize homosexual marriages); DeSanto v. Barnsley, 328 Pa.Super. 181, 476 A.2d 952, 955-56 (1984) (holding that two persons of the same sex cannot enter a common-law marriage); Slayton v. State, 633 S.W.2d 934, 937 (Tex.Crim.App.1982) (holding that a marriage cannot exist between two persons of the same sex under Texas law). As already alluded to, see note 2 supra, the Hawaii Supreme Court has found that a state statute which denies same-sex couples the right to marry presumptively violated its state constitution unless the state showed it could survive strict scrutiny. Baehr, supra, 852 P.2d at 67. However, the constitutionality of the District's laws is judged solely under the federal constitution.