MILLER v. ALBRIGHT (96-1060)
96 F.3d 1467, affirmed.
[ Stevens ]
[ O’Connor ]
[ Scalia ]
[ Ginsburg ]
[ Breyer ]
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version



[April 22, 1998]

Justice Breyer, with whom Justice Souter and Justice Ginsburg join, dissenting.

Since the founding of our Nation, American statutory law, reflecting a long-established legal tradition, has provided for the transmission of American citizenship from parent to child–even when the child is born abroad. Today’s case focuses upon statutes that make those children, when born out of wedlock, “citizens of the United States at birth.” 8 U.S.C. § 1401 and 1409. The statutes, as applied where only one parent is American, require the American parent–whether father or mother–to prove the child is his or hers and to meet a residency requirement. The statutes go on to require (1) that the American parent promise to provide financial support for the child until the child is 18, and (2) that the American parent (or a court) legitimate or formally acknowledge the child before the child turns 18–if and only if the American parent is the father, but not if the parent is the mother.

What sense does it make to apply these latter two conditions only to fathers and not to mothers in today’s world–where paternity can readily be proved and where women and men both are likely to earn a living in the workplace? As Justice O’Connor has observed, and as a majority of the Court agrees, “[i]t is unlikely … that any gender classifications based on stereotypes can survive heightened scrutiny.” Ante, at 7. These two gender-based distinctions lack the “exceedingly persuasive” support that the Constitution requires. United States v. Virginia, 518 U.S. 515, 530 (1996). Consequently, the statute that imposes them violates the Fifth Amendment’s “equal protection” guarantee. See Bolling v. Sharpe, 347 U.S. 497, 500 (1954).


The family whose rights are at issue here consists of Charlie Miller, an American citizen, Luz Peñero, a citizen of the Philippines, and their daughter, Lorelyn. Lorelyn was born out of wedlock in 1970 in the Philippines. The relevant citizenship statutes state that a child born out of wedlock shall be a “citize[n] of the United States at birth,” §1401, if the child is born to a father who “had the nationality of the United States at the time of the person’s birth,” if the “blood relationship between the person and the father is established by clear and convincing evidence,” if the father had been physically present in the United States for five years, and:

“(3) the father (unless deceased) has agreed in writing to provide financial support for the person until the person reaches the age of 18 years, and

‘‘(4) while the person is under the age of 18 years–

‘‘(a) the person is legitimated under the law of the person’s residence or domicile,

‘‘(b) the father acknowledges paternity of the person in writing under oath, or

‘‘(c) the paternity of the person is established by adjudication of a competent court.” 8 U.S.C. § 1409(a) and 1401(g).

Charlie Miller did not meet the requirements set forth in (3) and (4) above on time. And the question before us is whether the Constitution forbids the application of those requirements for the reason that the statute imposed them only where the child’s American parent is the child’s father, not the mother. In my view the Constitution does forbid their application.


I agree with Justice Stevens’ resolution of the Government’s three threshold claims. First, the Government takes issue with Lorelyn’s argument that provisions (3) and (4) unconstitutionally infringe the rights of her father, Charlie, an American citizen. Brief for Respondent 11. It adds that Charlie, not Lorelyn, should assert those rights himself and that Lorelyn lacks legal “standing” to do so. Id., at 11, and n. 2. This Court has made clear, however, that a party can “assert” the constitutional rights of another person, where (1) that party has “suffered an ‘injury in fact’ ”; (2) the party and the other person have a “close relationship”; and (3) “there was some hindrance” to the other person’s “asserting” his “own rights.” Campbell v. Louisiana, __ U.S. __ (1998) (slip op., at 4); see also Powers v. Ohio, 499 U.S. 400, 411 (1991). And these three requirements are met here.

Lorelyn has suffered an “injury in fact.” She has a “close” and relevant relationship with the other person, namely, her father. And there was “some hindrance” to her father’s asserting his own rights. Charlie began this lawsuit (originally filed in Texas) as a party, raising his own Equal Protection claim. The Government originally moved to dismiss the complaint, contending that Charlie “should be dismissed from this suit because he lacked standing.” Motion to Dismiss Plaintiff ’s First Amended Complaint, or, in the Alternative, Transfer Venue 6. The District Court agreed with the Government that Charlie lacked “standing,” and he was dismissed from the suit. App. 11a. Lorelyn remained as the sole plaintiff, and for reasons of venue, see 28 U.S.C. § 1391(e)(1), the court then transferred the case to the District of Columbia pursuant to §1406(a). App. 11a. The conclusion that the Government “hindered” Charlie’s assertion of his own rights in this case is irresistible.

The Government points out that Charlie might have appealed the adverse Texas District Court ruling. Brief for Respondent 11, n. 2. But appeals take time and money; the transfer of venue left the plaintiffs uncertain about where to appeal; the case was being heard with Lorelyn as plaintiff in any event; and the resulting comparison of costs and benefits (viewed prospectively) likely would have discouraged Charlie’s pursuit of the alternative appeal route. The Government’s successful dismissal motion thus had practical consequences that “hindered” Charlie at least as much as those we have elsewhere said create “hindrances” sufficient to satisfy this portion of the “third party standing” test. See, e.g., Campbell, supra, at __ (slip op., at 4) (criminal defendant can assert rights of racially-excluded petit jurors because of “arduous” process surrounding, and small benefits accruing to, juror effort to vindicate own rights); cf. Craig v. Boren, 429 U.S. 190, 193—194 (1976) (“decision . . . to forgo consideration of the constitutional merits . . . to await” another party’s identical claim would “foster repetitive and time-consuming litigation under the guise of caution and prudence”).

Second, the Government, citing United States v. Verdugo-Urquidez, 494 U.S. 259 (1990), and Johnson v. Eisentrager, 339 U.S. 763 (1950), argues that the Fifth Amendment does not protect an alien, such as Lorelyn, living outside the United States. Brief for Respondent 11—12. The rights to be vindicated here, however, are Charlie’s, not Lorelyn’s. And, in any event, those cases, as Justice Stevens points out, are irrelevant, for the matter at issue here is whether or not Lorelyn is a citizen. See Rogers v. Bellei, 401 U.S. 815 (1971) (considering on the merits a putative citizen’s claim that he was a citizen due to the operation of the Fifth Amendment, even though he apparently was living outside the United States at the time he filed suit).

Third, the Government argues that Lorelyn cannot succeed because a federal court lacks the power to grant her the relief she seeks, namely, a grant of citizenship. Brief for Respondent 43—50. As I shall later explain in more detail, however, this argument is beside the point, for, once the two unconstitutional clauses are excised from the statute, that statute operates automatically to confer citizenship upon Lorelyn “at birth.” 8 U.S.C. § 1401; see Part V, infra.

Justice O’Connor, joined by Justice Kennedy, says that Lorelyn cannot assert her father’s rights because “she has not demonstrated a substantial hindrance to her father’s ability to assert his own rights.” Ante, at 3. But the obstacles that the Government placed in her father’s path substantially hindered his efforts to do so in practice. See supra, at 3—4. Several of the cases mentioned in Justice O’Connor’s opinion involved the denial of standing, but none of those cases involved any “hindrance,” and Justice O’Connor does not claim that they do. See FW/PBS, Inc. v. Dallas, 493 U.S. 215, 234 (1990) (husband lacks standing to assert wife’s moot claim); Bender v. Williamsport Area School Dist., 475 U.S. 534, 544—545 (1986) (school board member lacks standing to defend on board’s behalf a claim that all other board members voted not to defend); Gladstone, Realtors v. Village of Bellwood, 441 U.S. 91, 112, n. 25 (1979) (nonresidents lack standing to challenge local real estate practices as discriminatory); Heald v. District of Columbia, 259 U.S. 114, 123 (1922) (District resident lacks standing to claim local tax unconstitutional as applied to bonds held by nonresidents outside District). I have previously pointed to cases in which the Court has found third-party standing where the “hindrance” was of the same kind and approximate degree as that present here. Supra, at 4. There are, of course, other cases finding standing that arguably involve even greater hindrance. See, e.g., Hodel v. Irving, 481 U.S. 704, 711—712 (1987); Carey v. Population Services Int’l, 431 U.S. 678, 684, n. 4 (1977); Singleton v. Wulff, 428 U.S. 106, 108 (1976); Craig, supra, at 192; Eisenstadt v. Baird, 405 U.S. 438, 446 (1972); NAACP v. Alabama ex rel. Patterson, 357 U.S. 449, 459 (1958); Barrows v. Jackson, 346 U.S. 249, 254 (1953). But they set no inner limit.

Nor do I agree with Justice O’Connor’s determination that “rational scrutiny” must apply to Lorelyn’s assertion of her own rights. Lorelyn belongs to a class made up of children of citizen-fathers, whom the law distinguishes from the class of children of citizen-mothers, solely on grounds of the parent’s gender. This Court, I assume, would use heightened scrutiny were it to review discriminatory laws based upon ancestry, say laws that denied voting rights or educational opportunity based upon the religion, or the racial make-up, of a parent or grandparent. And, if that is so, I am not certain that it makes a significant difference whether one calls the rights at issue those of Lorelyn or of her father. Allen v. Wright, 468 U.S. 737 (1984), does not hold to the contrary. Id., at 755 (black schoolchildren’s parents who claimed a “stigmatizing injury” due to Internal Revenue Service decision to grant tax exempt status to racially discriminatory private schools had not been “personally denied equal treatment,” and thus had not been injured).

Regardless, like Justice O’Connor, I “do not share,” and thus I believe a Court majority does not share, “Justice Stevens’ assessment that the provision withstands heightened scrutiny.” Ante, at 7. I also agree with Justice O’Connor that “[i]t is unlikely” that “gender classifications based on stereotypes can survive heightened scrutiny,” ibid., a view shared by at least five members of this Court. Indeed, for reasons to which I shall now turn, we must subject the provisions here at issue
to “heightened scrutiny.” And those provisions cannot survive.


This case is about American citizenship and its transmission from an American parent to his child. The right of citizenship, as this Court has said, is “a most precious right.” Kennedy v. Mendoza&nbhyph;Martinez, 372 U.S. 144, 159 (1963); see also Fedorenko v. United States, 449 U.S. 490, 507 (1981) (citizenship is a “priceless treasure”) (internal quotation marks omitted); Luria v. United States, 231 U.S. 9, 22 (1913) (“Citizenship is membership in a political society”); Afroyim v. Rusk, 387 U.S. 253, 268 (1967) (“[This Nation’s] citizenry is the country and the country is its

Further, the tie of parent to child is a special one, which in other circumstances by itself has warranted special constitutional protection. See, e.g., Wisconsin v. Yoder, 406 U.S. 205 (1972); Pierce v. Society of Sisters, 268 U.S. 510 (1925); Meyer v. Nebraska, 262 U.S. 390 (1923); see also Skinner v. Oklahoma ex rel. Williamson, 316 U.S. 535 (1942).

Moreover, American statutory law has consistently recognized the rights of American parents to transmit their citizenship to their children. See Act of Mar. 26, 1790, §1 1 Stat. 103; Act of Jan. 29, 1795, §3, 1 Stat. 415; Act of Apr. 14, 1802, §4, 2 Stat. 155; Act of Feb. 10, 1855, §1, 10 Stat. 604; Rev. Stat. §1993; Act of Mar. 2, 1907, §6, 34 Stat. 1229; Act of May 24, 1934, §1, 48 Stat. 797; Nationality Act of 1940, §201(g), 54 Stat. 1139; Immigration and Nationality Act of 1952, §301(a)(7), (b), 66 Stat. 235, 236, as amended, 8 U.S.C. § 1401; cf., e.g., 1 Oppenheim’s International Law §384 (R. Jennings & A. Watts 9th ed. 1992) (noting that in many States, children born abroad of nationals become nationals); 43 A. Berger, Encyclopedic Dictionary of Roman Law 389 (1953) (Roman citizenship was acquired principally by parentage); Sandifer, A Comparative Study of Laws Relating to Nationality at Birth and to Loss of Nationality, 29 Am. J. Int’l L. 248, 248—261, 278 (1935) (discussing citizenship laws throughout the world and noting the “widespread extent of the rule of jus sanguinis”); E. de Vattel, The Law of Nations 101—102 (J. Chitty trans., 1883) (1758).

Finally, the classification at issue is gender-based, and we have held that, under the equal protection principle, such classifications may not rest on generalizations about the different capacities of males and females when neutral categories would serve the legislature’s end. United States v. Virginia, 518 U.S., at 540—546.

These circumstances mean that courts should not diminish the quality of review–that they should not apply specially lenient standards–when they review these statutes. The statutes focus upon two of the most serious of human relationships, that of parent to child and that of individual to the State. They tie each to the other, transforming both while strengthening the bonds of loyalty that connect family with Nation. Yet because they confer the status of citizenship “at birth,” they do not involve the transfer of loyalties that underlies the naturalization of aliens, where precedent sets a more lenient standard of review. See Fiallo v. Bell, 430 U.S. 787 (1977).

To the contrary, the same standard of review must apply when a married American couple travel abroad or temporarily work abroad and have a child as when a single American parent has a child born abroad out of wedlock. If the standard that the law applies is specially lenient, then statutes conferring citizenship upon these children could discriminate virtually free of independent judicial review. And as a result, many such children, lacking citizenship, would be placed outside the domain of basic constitutional protections. Nothing in the Constitution requires so anomalous a result.

I recognize that, ever since the Civil War, the transmission of American citizenship from parent to child, jus sanguinis, has played a role secondary to that of the transmission of citizenship by birthplace, jus soli. See Rogers v. Bellei, 401 U.S., at 828; see also Weedin v. Chin Bow, 274 U.S. 657, 669—671 (1927) (citing United States v. Wong Kim Ark, 169 U.S. 649, 674 (1898), and id., at 714 (Fuller, C. J., dissenting)). That lesser role reflects the fact that the Fourteenth Amendment’s Citizenship Clause does not mention statutes that might confer citizenship “at birth” to children of Americans born abroad. U.S. Const., Amdt. 14, §1 (stating that “[a]ll persons born or naturalized in the United States . . . are citizens”). But that omission, though it may give Congress the power to decide whether or not to extend citizenship to children born outside the United States, see Rogers v. Bellei, supra, at 835, does not justify more lenient “equal protection” review of statutes that embody a congressional decision to do so.

Nothing in the language of the Citizenship Clause argues for less close scrutiny of those laws conferring citizenship at birth that Congress decides to enact. Nor have I found any support for a lesser standard in either the history of the Clause or its purpose. To the contrary, those who wrote the Citizenship Clause hoped thereby to assure that courts would not exclude newly freed slaves–born within the United States–from the protections the Fourteenth Amendment provided, including “equal protection of the laws.” See, e.g., Afroyim v. Rusk, 387 U.S., at 262; id., at 283—284 (Harlan, J., dissenting); H. Flack, Adoption of the Fourteenth Amendment 83—97 (1908). They took special care, lest deprivation of citizenship undermine the Amendment’s guarantee of “equal protection of the laws.” Care is no less necessary when statutes, transferring citizenship between American parent and child, make the child a citizen “at birth.” How then could the Fourteenth Amendment itself provide support for a diminished standard of review?

Nor have I found any such support in the history of the jus sanguinis statutes. That history shows a virtually unbroken tradition of transmitting American citizenship from parent to child “at birth,” under statutes that imposed certain residence requirements. Supra, at 5—6; see also Bellei, supra, at 835. A single gap occurred when, for a brief period of time, the relevant statutes (perhaps inadvertently) failed to confer citizenship upon what must have been a small group of children born abroad between 1802 and 1855 whose citizen-fathers were also born between 1802 and 1855. See Montana v. Kennedy, 366 U.S. 308, 311—312 (1961); Weedin, supra, at 663—664; Wong Kim Ark, supra, at 673—674. But even then, some courts, recognizing the importance of the right, found common-law authority for the transmission to those children of their parent’s American citizenship. See Ludlam v. Ludlam, 26 N. Y. 356, 362—372 (1863); see also Lynch v. Clarke, 1 Sandf. Ch. 583, 659—663 (N. Y. 1844).

The history of these statutes does reveal considerable discrimination against women, particularly from 1855 to 1934. See ante, at 4—6 (Ginsburg, J., dissenting). But that discrimination then cannot justify this discrimination now, when much discrimination that the law once tolerated, including “de jure segregation and the total exclusion of women from juries,” is “now unconstitutional even though [it] once coexisted with the Equal Protection Clause.” J. E. B. v. Alabama ex rel. T. B., 511 U.S. 127, 143, n. 15 (1994).

Neither have I found case law that could justify use here of a more lenient standard of review. Justice Stevens points out that this Court has said it will apply a more lenient standard in matters of “ ‘immigration and naturalization.’ Ante, at 13, n. 11 (quoting Mathews v. Diaz, 426 U.S. 67, 82 (1976)). But that language arises in
a case involving aliens. The Court did not say it in-
tended that phrase to include statutes that confer citizenship “at birth.” And Congress does not believe that this kind of citizenship involves “naturalization.” 8 U.S.C. § 1101(a)(23) (“The term ‘naturalization’ means the conferring of nationality of a state upon a person after birth, by any means whatsoever”) (emphasis added). The Court to my knowledge has never said, or held, or reasoned that statutes automatically conferring citizenship “at birth” upon the American child of American parents receive a more lenient standard of review.

The Court has applied a deferential standard of review in cases involving aliens, not in cases in which only citizens’ rights were at issue. See Mathews, supra (rights of alien residents); Kleindienst v. Mandel, 408 U.S. 753 (1972) (citizens’ rights related to treatment of alien); Fiallo v. Bell, 430 U.S. 787 (1977) (citizens’ rights to obtain immigration preferences for relatives who are aliens). When the Court has considered the latter kind of case, it has not lowered the standard of review. See Bellei, supra, at 828—836 (evaluating due process challenge to citizenship statute under generally applicable standard).

In sum, the statutes that automatically transfer American citizenship from parent to child “at birth” differ significantly from those that confer citizenship on those who originally owed loyalty to a different nation. To fail to recognize this difference, and consequently to apply an unusually lenient constitutional standard of review here, could deprive the children of millions of Americans, married and unmarried, working abroad, traveling, say, even temporarily to Canada or Mexico, of the most basic kind of constitutional protection. See U.S. Dept. of Commerce, Bureau of Census, Statistical Abstract of the United States 53 (1997) (table 54) (reporting that, as of 1990, 1.86 million United States citizens were born abroad or at sea to American parents); see also Hearing before the Subcommittee on International Operations of the House Committee on Foreign Affairs, 102d Cong., 1st Sess., 114 (1991) (testimony of Andrew P. Sundberg) (“According to the most recent survey carried out by the State Department, 40,000 children are born abroad each year to a U.S. citizen parent”). Thus, generally prevailing, not specially lenient, standards of review must apply.


If we apply undiluted equal protection standards, we must hold the two statutory provisions at issue unconstitutional. The statutes discriminate on the basis of gender, making it significantly more difficult for American fathers than for American mothers to transmit American citizenship to their children born out of wedlock. If the citizen-parent is a man, the statute requires (1) a promise by the father to support the child until the child is 18, and (2) before the child turns 18, legitimation, written acknowledgment by the father under oath, or an adjudication of paternity. 8 U.S.C. § 1409(a). If the citizen-parent is a woman, she need not do either. §1409(c).

Distinctions of this kind–based upon gender–are subject to a “ ‘strong presumption’ ” of constitutional invalidity. Virginia, 518 U.S., at 532 (quoting J. E. B., supra, at 152 (Kennedy, J., concurring in judgment)). The Equal Protection Clause permits them only if the Government meets the “demanding” burden of showing an “ ‘exceedingly persuasive’ ” justification for the distinction. Virginia, supra, at 533; see also J. E. B., supra, at 136; Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 724 (1982); Personnel Administrator of Mass. v. Feeney, 442 U.S. 256, 273 (1979); Kirchberg v. Feenstra, 450 U.S. 455, 461 (1981). That distinction must further important governmental objectives, and the discriminatory means employed must be “substantially related” to the achievement of those objectives. Virginia, supra, at 533 (citing Mississippi Univ. for Women, supra, at 724). This justification “must be genuine, not hypothesized or invented post hoc in response to litigation.” Virginia, 518 U. S, at 533. Further, “it must not rely on overbroad generalizations about the different talents, capacities, or preferences of males and females.” Ibid.; see also J. E. B., 511 U.S., at 139—140, and n. 11; Craig, 429 U.S., at 201; Califano v. Goldfarb, 430 U.S. 199, 223—224 (1977) (Stevens, J., concurring in judgment); Weinberger v. Wiesenfeld, 420 U.S. 636, 643 (1975). The fact that the statutes “discriminat[e] against males rather than against females” is beside the point. Mississippi Univ. for Women, 458 U.S., at 723.

The statutory distinctions here violate these standards. They depend for their validity upon the generalization that mothers are significantly more likely than fathers to care for their children, or to develop caring relationships with their children. But consider how the statutes work once one abandons that generalization as the illegitimate basis for legislative line-drawing we have held it to be. Id., at 726, 730. First, assume that the American citizen is also the Caretaker Parent. The statute would then require a Male Caretaker Parent to acknowledge his child prior to the child’s 18th birthday (or for the parent or child to obtain a court equivalent) and to provide financial support. It would not require a Female Caretaker Parent to do either. The gender-based distinction that would impose added burdens only upon the Male Caretaker Parent would serve no purpose at all. Second, assume that the American citizen is the Non-Caretaker Parent. In that circumstance, the statute would forgive a Female Non-Caretaker Parent from complying with the requirements (for formal acknowledgment and written promises to provide financial support) that it would impose upon a Male Non-Caretaker Parent. Again, the gender based distinction that would impose lesser burdens only upon the Female Non-Caretaker Parent would serve no purpose.

To illustrate the point, compare the family before us–Charlie, Lorelyn, and Luz–with an imagined family–Carlos, a Philippine citizen, Lucy, his daughter, and Lenora, Lucy’s mother and an American citizen. Suppose that Lenora, Lucy’s unmarried mother, returned to the United States soon after Lucy’s birth, leaving Carlos to raise his daughter. Why, under those circumstances, should Lenora not be required to fulfill the same statutory requirements that here apply to Charlie? Alternatively, imagine that Charlie had taken his daughter Lorelyn back to the United States to raise. The statute would not make Lorelyn an American from birth unless Charlie satisfied its two conditions. But had our imaginary family mother, Lenora, taken her child Lucy back to the United States, the statute would have automatically made her an American from birth without anyone having satisfied the two conditions. The example suggests how arbitrary the statute’s gender-based distinction is once one abandons the generalization that mothers, not fathers, will act as caretaker parents.

Let me now deal more specifically with the justifications that Justice Stevens finds adequate. Justice Stevens asserts that subsection (a)(4) serves two interests: first, “ensuring reliable proof of a biological relationship between the potential citizen and its citizen parent,” ante, at 14, and second, “encouraging” certain relationships or ties, namely “the development of a healthy relationship between the citizen parent and the child while the child is a minor,” ante, at 16, as well as “the related interest in fostering ties between the foreign-born child and the United States.” Ibid. I have no doubt that these interests are important. But the relationship between the statutory requirements and those particular objectives is one of total misfit.

Subsection (a)(4) requires, for example, the American citizen father to “acknowledg[e]” paternity before the child reaches 18 years of age, or for the child or parent to obtain a court equivalent (legitimation or adjudication of paternity). Justice Stevens suggests that this require-
ment “produces the rough equivalent of the documentation,” such as a birth certificate memorialized in hospital records, “already available to evidence the blood relationship between the mother and the child.” Ante, at 14. But, even if I assume the “equivalency” (only for argument’s sake, since birth certificates do not invariably carry a mother’s true name or omit the father’s), I still do not
understand the need for the prior-to-18 legitimation-or-acknowledgment requirement. When the statute was written, one might have seen the requirement as offering some protection against false paternity claims. But that added protection is unnecessary in light of inexpensive DNA testing that will prove paternity with certainty. See Shapiro, Reifler, and Psome, The DNA Paternity Test: Legislating the Future Paternity Action, 7 J. Law & Health 1, 29 (1992—1993) (current testing methods can determine probability of paternity to 99.999999% accuracy); see also H. R. Rep. No. 98—527, p. 38 (1983).

Moreover, a different provision of the statute, subsection (a)(1), already requires proof of paternity by “clear and convincing evidence.” No one contests the validity of that provision, and I believe that biological differences between men and women would justify its imposition where paternity is at issue. In light of that provision, subsection (a)(4)’s protection against false claims is not needed. Indeed, the Government concedes that, in light of the “clear and convincing evidence” requirement, the “time limit for meeting the legitimation-or-acknowledgement requirement of Section 309(a)(4) must . . . reflect, at least in part, some other congressional concern.” Brief for Respondent 27 (emphasis added).

Justice Stevens says that this “other concern” is a concern for the establishment of relationships and ties, to the father and to the United States, all before the child is 18. Ante, at 16—17. According to Justice Stevens, the way in which the requirement serves this purpose is by making certain the father knows of the child’s existence–in the same way, it says, that a mother, by giving birth, automatically knows that the child exists. Ibid.

The distance between this knowledge and the claimed objectives, however, is far too great to satisfy any legal requirement of tailoring or proportionality. And the assumption that this knowledge-of-birth could make a significant gender-related difference rests upon a host of unproved gender-related hypotheses. Simple knowledge of a child’s existence may, or may not, be followed by the kinds of relationships for which Justice Stevens hopes. A mother or a father, knowing of a child’s birth, may nonetheless fail to care for the child or even to acknowledge the child. A father with strong ties to the child may, simply by lack of knowledge, fail to comply with the statute’s formal requirements. A father with weak ties might readily comply. Moreover, the statute does little to assure any tie for, as Justice Stevens acknowledges, a child might obtain an adjudication of paternity “absent any affirmative act by the father, and perhaps even over his express objection.” Ante, at 12.

To make plausible the connection between the statute’s requirement and the asserted “relationship” goals, Justice Stevens must find a factual scenario where a father’s knowledge–equivalent to the mother’s knowledge that she has given birth–could lead to the establishment of a more meaningful parenting relationship or tie to America. It therefore points to what one might term the “war baby” problem–the problem created by American servicemen fathering children overseas and returning to America unaware of the related pregnancy or birth. The statutory remedy before us, however, is disproportionately broad even when considered in relation to that problem. Justice Stevens refers to 683,000 service personnel stationed in the Far East in 1970 when Lorelyn was born. Ante, at 17. The statute applies, however, to all Americans who live or travel abroad, including the 3.2 million private citizens, and the 925,000 Federal Government employees, who live, or who are stationed, abroad–of whom today only 240,000 are active duty military employees, many of whom are women. U.S. Dept. of State, Private American Citizens Residing Abroad (Nov. 21, 1997); U.S. Dept. of Commerce, Bureau of the Census, Americans Overseas in U.S. Censuses, Technical Paper 62, p. 62 (Nov. 1993) (1990 census figures); U.S. Dept. of Defense, Selected Manpower Statistics 23, 44 (DIOR/MO1—96 1996). Nor does the statute seem to have been aimed at the “war baby” problem, for the precursor to the provisions at issue was first proposed in a 1938 report and was first adopted in the Nationality Act of 1940, which was enacted before the United States entered World War II. Nationality Laws of the United States: Message from the President of the United States, 76th Cong., 1st Sess., pt. 1, pp. 17—18 (Comm. Print submitted to House Comm. on Immigration and Naturalization, 1939); Nationality Act of 1940, §205, 54 Stat. 1139.

Nor is there need for the gender-based discrimination at issue here, for, were Congress truly interested in achieving the goals Justice Stevens posits in the way Justice Stevens suggests, it could simply substitute a requirement of knowledge-of-birth for the present subsection (a)(4); or it could distinguish between caretaker and non-caretaker parents, rather than between men and women. A statute that does not do so, but instead relies upon gender-based distinctions, appears rational only, as I have said, supra, at 13—14, if one accepts the legitimacy of gender-based generalizations that, for example, would equate gender and caretaking–generalizations of a kind that this Court has previously found constitutionally impermissible. See, e.g., Virginia, 518 U.S., at 542, 546 (striking down men-only admissions policy at Virginia Military Institute even assuming that “most women would not choose VMI’s adversative method”); J. E. B., 511 U.S., at 139, n. 11 (invalidating gender-based peremptory challenges “[e]ven if a measure of truth can be found in some of the gender stereotypes used to justify” them); Craig, 429 U.S., at 201 (invalidating Oklahoma law that established different drinking ages for men and women, although the evidence supporting the age differential was “not trivial in a statistical sense”); Wiesenfeld, 420 U.S., at 645 (holding unconstitutional statutory classification giving to widowed mothers benefits not available to widowed fathers even though “the notion that men are more likely than women to be the primary supporters of their spouses and children is not entirely without empirical support”). Although Justice Stevens cites Lehr v. Robertson, 463 U.S. 248 (1983), for support, ante, at 19, that case was decided before the DNA advances described earlier.

For similar reasons, subsection (3) denies Charlie Miller “equal protection” of the laws. That subsection requires an American father to “agre[e] … to provide financial support” for the child until the child “reaches the age of 18,” but does not require the same of an American mother. I agree with the Government that this provision has as one objective helping to assure ties between father and child. Brief for Respondent 26. But I do not see why the same need does not exist with respect to a mother. And, where the American parent is the Non-Caretaker Parent, the need for such assurances would seem the same in respect to either sex. Where the American parent is the Caretaker Parent, there would seem no need for the assurance regardless of gender. Since either men or women may be caretakers, and since either men or women may be “breadwinners,” one could justify the gender distinction only on the ground that more women are caretakers than men, and more men are “breadwinners” than women. This, again, is the kind of generalization that we have rejected as justifying a gender-based distinction in other cases. Virginia, supra, at 540—546; J. E. B., supra, at 140, n. 11; Craig, supra, at 201; Wiesenfeld, supra, at 645.

For these reasons, I can find no “exceedingly persuasive” justification for the gender-based distinctions that the statute draws.


Justice Scalia argues that, if the provisions at issue violate the Constitution, we nonetheless are powerless to find a remedy. But that is not so. The remedy is simply that of striking from the statute the two subsections that offend the Constitution’s equal protection requirement, namely (a)(3) and (a)(4). With those subsections omitted, the statute says that the daughter, Lorelyn, of one who, like Charlie, has proved paternity by “clear and convincing evidence,” is an American citizen, and has lived in the United States for five years, is a “citize[n] of the United States at birth.” 8 U.S.C. § 1409(a) and 1401. Whatever limitations there may be upon a Court’s powers to grant citizenship, those limitations are not applicable here, for the Court need not grant citizenship. The statute itself grants citizenship automatically, and “at birth.” And this Court need only declare that that is so. INS v. Pangilinan, 486 U.S. 875 (1988), which Justice Scalia cites in support, is beside the point, for the plaintiffs in that case, conceding that the statute at issue did not make them citizens, asked the courts to confer citizenship as a remedy in equity. Cf. Bellei, 401 U.S., at 828—836 (assessing claim that statute conferred citizenship in the absence of a provision argued to be unconstitutional, without identifying any special remedial problems).

Of course, we can excise the two provisions only if Congress likely would prefer their excision, rather than imposing similar requirements upon mothers. Califano v. Westcott, 443 U.S. 76, 89—93 (1979); Welsh v. United States, 398 U.S. 333, 361 (1970) (Harlan, J., concurring in result). But, since the provisions at issue seem designed in significant part to address difficulties in proving paternity (along with providing encouragement for fathers to legitimate the child) and, since DNA advances have overcome the paternity-proof difficulties, I believe that Congress would have preferred severance.

Justice Scalia is also wrong, I believe, when he says that “the INA itself contains a clear statement of congressional intent” not to sever, ante, at 6, for the Act in fact contains the following explicit severability provision:

“If any particular provision of this Act, or the application thereof to any person or circumstance, is held invalid, the remainder of the Act and the application of such provision to other persons or circumstances shall not be affected thereby.” §406, 66 Stat. 281; see note following 8 U.S.C. § 1101 p. 38, “Separability.”

The provision cited by Justice Scalia says:

“A person may be naturalized as a citizen of the United States in the manner and under the conditions prescribed in this title and not otherwise.” §310(d), 66 Stat. 239, 8 U.S.C. § 1421(d).

As “naturalization” under this statute does not include the conferral of citizenship at birth, the provision does not apply here. See 8 U.S.C. § 1101(a)(23) (“The term ‘naturalization’ means the conferring of nationality of a state upon a person after birth”) (emphasis added).

Justice Scalia also says that the law, as excised, would favor fathers over mothers. Ante, at 7—8. The law, however, would require both fathers and mothers to prove their parentage; it would require that one or the other be an American, it would impose residency requirements that, if anything, would disfavor fathers. I cannot find the reverse favoritism that Justice Scalia fears.

For these reasons, I would reverse the judgment of the Court of Appeals.