1. Section 188.029 of the Missouri statute—which specifies, in its first sentence, that a physician, before performing an abortion on a woman he has reason to believe is carrying an unborn child of 20 or more weeks gestational age, shall first determine if the unborn child is viable by using that degree of care, skill, and proficiency that is commonly exercised by practitioners in the field; but which then provides, in its second sentence, that, in making the viability determination, the physician shall perform such medical examinations and tests as are necessary to make a finding of the unborn child's gestational age, weight, and lung maturity—is constitutional, since it permissibly furthers the State's interest in protecting potential human life. Pp. 513-521.

(a) The Court of Appeals committed plain error in reading § 188.029 as requiring that after 20 weeks the specified tests must be performed. That section makes sense only if its second sentence is read to require only those tests that are useful in making subsidiary viability findings. Reading the sentence to require the tests in all circumstances, including when the physician's reasonable professional judgment indicates that they would be irrelevant to determining viability or even dangerous to the mother and the fetus, would conflict with the first sentence's requirement that the physician apply his reasonable professional skill and judgment. It would also be incongruous to read the provision, especially the word "necessary," to require tests irrelevant to the expressed statutory purpose of determining viability. Pp. 514-515.

(b) Section 188.029 is reasonably designed to ensure that abortions are not performed where the fetus is viable. The section's tests are intended to determine viability, the State having chosen viability as the point at which its interest in potential human life must be safeguarded. The section creates what is essentially a presumption of viability at 20 weeks, which the physician, prior to performing an abortion, must rebut with tests—including, if feasible, those for gestational age, fetal weight, and lung capacity—indicating that the fetus is not viable. While the District Court found that uncontradicted medical evidence established that a 20-week fetus is not viable, and that 231/2 to 24 weeks' gestation is the earliest point at which a reasonable possibility of viability exists, it also found that there may be a 4-week error in estimating gestational age, which supports testing at 20 weeks. Pp. 515-516.

(c) Section 188.029 conflicts with Roe v. Wade and cases following it. Since the section's tests will undoubtedly show in many cases that the fetus is not viable, the tests will have been performed for what were in fact second-trimester abortions. While Roe, 410 U.S., at 162, 93 S.Ct., at 731, recognized the State's interest in protecting potential human life as "important and legitimate," it also limited state involvement in second-trimester abortions to protecting maternal health, id., at 164, 93 S.Ct., at 732, and allowed States to regulate or proscribe abortions to protect the unborn child only after viability, id., at 165, 93 S.Ct., at 732-33. Since the tests in question regulate the physician's discretion in determining the viability of the fetus, § 188.029 conflicts with langu ge in Colautti v. Franklin, 439 U.S. 379, 388-389, 99 S.Ct. 675, 681-683, 58 L.Ed.2d 596, stating that the viability determination is, and must be, a matter for the responsible attending physician's judgment. And, in light of District Court findings that the tests increase the expenses of abortion, their validity may also be questioned under Akron, supra, 462 U.S., at 434-435, 103 S.Ct., at 2494-2495, which held that a requirement that second-trimester abortions be performed in hospitals was invalid because it substantially increased the expenses of those procedures. Pp. 516-517.

(d) The doubt cast on the Missouri statute by these cases is not so much a flaw in the statute as it is a reflection of the fact that Roe's rigid trimester analysis has proved to be unsound in principle and unworkable in practice. In such circumstances, this Court does not refrain from reconsidering prior constitutional rulings, notwithstanding stare decisis. E.g., Garcia v. San Antonio Metropolitan Transit Authority, 469 U.S. 528, 105 S.Ct. 1005, 83 L.Ed.2d 1016. The Roe framework is hardly consistent with the notion of a Constitution like ours that is cast in general terms and usually speaks in general principles. The framework's key elements—trimesters and viability—are not found in the Constitution's text, and, since the bounds of the inquiry are essentially indeterminate, the result has been a web of legal rules that have become increasingly intricate, resembling a code of regulations rather than a body of constitutional doctrine. There is also no reason why the State's compelling interest in protecting potential human life should not extend throughout pregnancy rather than coming into existence only at the point of viability. Thus, the Roe trimester framework should be abandoned. Pp. 517-520.

(e) There is no merit to Justice BLACKMUN's contention that the Court should join in a "great issues" debate as to whether the Constitution includes an "unenumerated" general right to privacy as recognized in cases such as Griswold v. Connecticut, 381 U.S. 479, 85 S.Ct. 1678, 14 L.Ed.2d 510. Unlike Roe, Griswold did not purport to adopt a whole framework, complete with detailed rules and distinctions, to govern the cases in which the asserted liberty interest would apply. The Roe framework sought to deal with areas of medical practice traditionally left to the States, and to balance once and for all, by reference only to the calendar, the State's interest in protecting potential human life against the claims of a pregnant woman to decide whether or not to abort. The Court's experience in applying Roe in later cases suggests that there is wisdom in not necessarily attempting to elaborate the differences between a "fundamental right" to an abortion, Akron, supra, 462 U.S., at 420, n. 1, 103 S.Ct., at 2487, n. 1, a "limited fundamental constitutional right," post, at 555, or a liberty interest protected by the Due Process Clause. Moreover, although this decision will undoubtedly allow more governmental regulation of abortion than was permissible before, the goal of constitutional adjudication is not to remove inexorably "politically devisive" issues from the ambit of the legislative process, but is, rather, to hold true the balance between that which the Constitution puts beyond the reach of the democratic process and that which it does not. Furthermore, the suggestion that legislative bodies, in a Nation where more than half the population is female, will treat this decision as an invitation to enact abortion laws reminiscent of the dark ages misreads the decision and does scant justice to those who serve in such bodies and the people who elect them. Pp. 520-521.

2. This case affords no occasion to disturb Roe § holding that a Texas statute which criminalized all nontherapeutic abortions unconstitutionally infringed the right to an abortion derived from the Due Process Clause. Roe is distinguishable on its facts, since Missouri has determined that viability is the point at which its interest in potential human life must be safeguarded. P. 521.

Justice O'CONNOR, agreeing that it was plain error for the Court of Appeals to interpret the second sentence of § 188.029 as meaning that doctors must perform tests to find gestational age, fetal weight, and lung maturity, concluded that the section was constitutional as properly interpreted by the plurality, and that the plurality should therefore not have proceeded to reconsider Roe v. Wade. This Court refrains from deciding constitutional questions where there is no need to do so, and generally does not formulate a constitutional rule broader than the precise facts to which it is to be applied. Ashwander v. TVA, 297 U.S. 288, 346, 347, 56 S.Ct. 466, 482-483, 483, 80 L.Ed. 688. Since appellees did not appeal the District Court's ruling that the first sentence of § 188.029 is constitutional, there is no dispute between the parties over the presumption of viability at 20 weeks created by that first sentence. Moreover, as properly interpreted by the plurality, the section's second sentence does nothing more than delineate means by which the unchallenged 20-week presumption may be overcome if those means are useful in determining viability and can be prudently employed. As so interpreted, the viability testing requirements do not conflict with any of the Court's abortion decisions. As the plurality recognizes, under its interpretation of § 188.029's second sentence, the viability testing requirements promote the State's interest in potential life. This Court has recognized that a State may promote that interest when viability is possible. Thornburgh v. American College of Obstetricians and Gynecologists, 476 U.S. 747, 770-771, 106 S.Ct. 2169, 2183-2184, 90 L.Ed.2d 779. Similarly, the basis for reliance by the lower courts on Colautti v. Franklin, 439 U.S. 379, 388-389, 99 S.Ct. 675, 681-683, 58 L.Ed.2d 596, disappears when § 188.029 is properly interpreted to require only subsidiary viability findings, since the State has not attempted to substitute its judgment for the physician's ascertainment of viability, which therefore remains "the critical point." Nor does the marginal increase in the cost of an abortion created by § 188.029's viability testing provision, as interpreted, conflict with Akron v. Akron Center for Reproductive Health, 462 U.S. 416, 434-439, 103 S.Ct. 2481, 2494-2497, 76 L.Ed.2d 687, since, here, such costs do not place a "heavy, and unnecessary burden" on a woman's abortion decision, whereas the statutory requirement in Akron, which related to previability abortions, more than doubled a woman's costs. Moreover, the statutory requirement in Akron involved second-trimester abortions generally; § 188.029 concerns only tests and examinations to determine viability when viability is possible. The State's compelling interest in potential life postviability renders its interest in determining the critical point of viability equally compelling. Thornburgh, supra, 476 U.S., at 770-771, 106 S.Ct., at 2183-2184. When the constitutional invalidity of a State's abortion statute actually turns upon the constitutional validity of Roe, there will be time enough to reexamine Roe, and to do so carefully. Pp. 525-531.

Justice SCALIA would reconsider and explicitly overrule Roe v. Wade. Avoiding the Roe question by deciding this case in as narrow a manner as possible is not required by precedent and not justified by policy. To do so is needlessly to prolong this Court's involvement in a field where the answers to the central questions are political rather than juridical, and thus to make the Court the object of the sort of organized pressure that political institutions in a democracy ought to receive. It is particularly perverse to decide this case as narrowly as possible in order to avoid reading the inexpressibly "broader-than-was-required-by-the-precise-f cts" structure established by Roe v. Wade. The question of Roe's validity is presented here, inasmuch as § 188.029 constitutes a legislative imposition on the judgment of the physician concerning the point of viability and increases the cost of an abortion. It does palpable harm, if the States can and would eliminate largely unrestricted abortion, skillfully to refrain from telling them so. Pp. 532-537.

REHNQUIST, C.J., announced the judgment of the Court and delivered the opinion for a unanimous Court with respect to Part II-C, the opinion of the Court with respect to Parts I, II-A, and II-B, in which WHITE, O'CONNOR, SCALIA, and KENNEDY, JJ., joined, and an opinion with respect to Parts II-D and III, in which WHITE and KENNEDY, JJ., joined. O'CONNOR, J., post, p. 522, and SCALIA, J., post, p. 532, filed opinions concurring in part and concurring in the judgment. BLACKMUN, J., filed an opinion concurring in part and dissenting in part, in which BRENNAN and MARSHALL, JJ., joined, post, p. 537. STEVENS, J., filed an opinion concurring in part and dissenting in part, post, p. 532-537.

William L. Webster, Jefferson City, Mo., for appellants.

Charles Fried, Washington, D.C., for U.S., as amicus curiae, supporting appellants, by special leave of Court.

Frank Susman, St. Louis, Mo., for appellees.

[Amicus Curiae Information from pages 497-499 intentionally omitted]

Chief Justice REHNQUIST announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I, II-A, II-B, and II-C, and an opinion with respect to Parts II-D and III, in which Justice WHITE and Justice KENNEDY join.

This appeal concerns the constitutionality of a Missouri statute regulating the performance of abortions. The United States Court of Appeals for the Eighth Circuit struck down several provisions of the statute on the ground that they violated this Court's decision in Roe v. Wade, 410 U.S. 113, 93 S.Ct. 705, 35 L.Ed.2d 147 (1973), and cases following it. We noted probable jurisdiction, 488 U.S. 1003, 109 S.Ct. 780, 102 L.Ed.2d 772 (1989), and now reverse.

* In June 1986, the Governor of Missouri signed into law Missouri Senate Committee Substitute for House Bill No. 1596 (hereinafter Act or statute), which amended existing state law concerning unborn children and abortions.1 The Act consisted of 20 provisions, 5 of which are now before the Court. The first provision, or preamble, contains "findings" by the state legislature that "[t]he life of each human being begins at conception," and that "unborn children have protectable interests in life, health, and well-being." Mo.Rev.Stat. §§ 1.205.1(1), (2) (1986). The Act further requires that all Missouri laws be interpreted to provide unborn children with the same rights enjoyed by other persons, subject to the Federal Constitution and this Court's precedents. § 1.205.2. Among its other provisions, the Act requires that, prior to performing an abortion on any woman whom a physician has reason to believe is 20 or more weeks pregnant, the physician ascertain whether the fetus is viable by performing "such medical examinations and tests as are necessary to make a finding of the gestational age, weight, and lung maturity of the unborn child." § 188.029. The Act also prohibits the use of public employees and facilities to perform or assist abortions not necessary to save the mother's life, and it prohibits the use of public funds, employees, or facilities for the purpose of "encouraging or counseling" a woman to have an abortion not necessary to save her life. §§ 188.205, 188.210, 188.215.

In July 1986, five health professionals employed by the State and two nonprofit corporations brought this class action in the United States District Court for the Western District of Missouri to challenge the constitutionality of the Missouri statute. Plaintiffs, appellees in this Court, sought declaratory and injunctive relief on the ground that certain statutory provisions violated the First, Fourth, Ninth, and Fourteenth Amendments to the Federal Constitution. App. A9. They asserted violations of various rights, including the "privacy rights of pregnant women seeking abortions"; the "woman's right to an abortion"; the "righ[t] to privacy in the physician-patient relationship"; the physician's "righ[t] to practice medicine"; the pregnant woman's "right to life due to inherent risks involved in childbirth"; and the woman's right to "receive . . . adequate medical advice and treatment" concerning abortions. Id., at A17-A19.

Plaintiffs filed this suit "on their own behalf and on behalf of the entire class consisting of facilities and Missouri licensed physicians or other health care professionals offering abortion services or pregnancy counseling and on behalf of the entir class of pregnant females seeking abortion services or pregnancy counseling within the State of Missouri." Id., at A13. The two nonprofit corporations are Reproductive Health Services, which offers family planning and gynecological services to the public, including abortion services up to 22 weeks "gestational age,"2 and Planned Parenthood of Kansas City, which provides abortion services up to 14 weeks gestational age. Id., at A9-A10. The individual plaintiffs are three physicians, one nurse, and a social worker. All are "public employees" at "public facilities" in Missouri, and they are paid for their services with "public funds," as those terms are defined by § 188.200. The individual plaintiffs, within the scope of their public employment, encourage and counsel pregnant women to have nontherapeutic abortions. Two of the physicians perform abortions. App. A54-A55.

Several weeks after the complaint was filed, the District Court temporarily restrained enforcement of several provisions of the Act. Following a 3-day trial in December 1986, the District Court declared seven provisions of the Act unconstitutional and enjoined their enforcement. 662 F.Supp. 407 (WD Mo.1987). These provisions included the preamble, § 1.205; the "informed consent" provision, which required physicians to inform the pregnant woman of certain facts before performing an abortion, § 188.039; the requirement that post-16-week abortions be performed only in hospitals, § 188.025; the mandated tests to determine viability, § 188.029; and the prohibition on the use of public funds, employees, and facilities to perform or assist nontherapeutic abortions, and the restrictions on the use of public funds, employees, and facilities to encourage or counsel women to have such abortions, §§ 188.205, 188.210, 188.215. Id., at 430.

The Court of Appeals for the Eighth Circuit affirmed, with one exception not relevant to this appeal. 851 F.2d 1071 (1988). The Court of Appeals determined that Missouri's declaration that life begins at conception was "simply an impermissible state adoption of a theory of when life begins to justify its abortion regulations." Id., at 1076. Relying on Colautti v. Franklin, 439 U.S. 379, 388-389, 99 S.Ct. 675, 681-683, 58 L.Ed.2d 596 (1979), it further held that the requirement that physicians perform viability tests was an unconstitutional legislative intrusion on a matter of medical skill and judgment. 851 F.2d, at 1074-1075. The Court of Appeals invalidated Missouri's prohibition on the use of public facilities and employees to perform or assist abortions not necessary to save the mother's life. Id., at 1081-1083. It distinguished our decisions in Harris v. McRae, 448 U.S. 297, 100 S.Ct. 2671, 65 L.Ed.2d 784 (1980), and Maher v. Roe, 432 U.S. 464, 97 S.Ct. 2376, 53 L.Ed.2d 484 (1977), on the ground that " '[t]here is a fundamental difference between providing direct funding to effect the abortion decision and allowing staff physicians to perform abortions at an existing publicly owned hospital.' " 851 F.2d, at 1081, quoting Nyberg v. City of Virginia, 667 F.2d 754, 758 (CA8 1982), appeal dism'd, 462 U.S. 1125, 103 S.Ct. 3102, 77 L.Ed.2d 1358 (1983). The Court of Appeals struck down the provision prohibiting the use of public funds for "encouraging or counseling" women to have nontherapeutic abortions, for the reason that this provision was both overly vague and inconsistent with the right to an abortion enunciated in Roe v. Wade. 851 F.2d, at 1077-1080. The court also invalidated the hospitalization requirement for 16-week abortions, id., at 1073-1074, and the prohibition on the use of public employees and facilities for abortion counseling, id., at 1077-1080, but the State has not appealed those parts of the udgment below. See Juris. Statement I-II.3

Decision of this case requires us to address four sections of the Missouri Act: (a) the preamble; (b) the prohibition on the use of public facilities or employees to perform abortions; (c) the prohibition on public funding of abortion counseling; and (d) the requirement that physicians conduct viability tests prior to performing abortions. We address these seriatim.

The Act's preamble, as noted, sets forth "findings" by the Missouri Legislature that "[t]he life of each human being begins at conception," and that "[u]nborn children have protectable interests in life, health, and well-being." Mo.Rev.Stat. §§ 1.205.1(1), (2) (1986). The Act then mandates that state laws be interpreted to provide unborn children with "all the rights, privileges, and immunities available to other persons, citizens, and residents of this state," subject to the Constitution and this Court's precedents. § 1.205.2.4 In invalidating the preamble, the Court of Appeals relied on this Court's dictum that " 'a State may not adopt one theory of when life begins to justify its regulation of abortions.' " 851 F.2d, at 1075-1076, quoting Akron v. Akron Center for Reproductive Health, Inc., 462 U.S. 416, 444, 103 S.Ct. 2481, 2500, 76 L.Ed.2d 687 (1983), in turn citing Roe v. Wade, 410 U.S., at 159-162, 93 S.Ct., at 729-731. It rejected Missouri's claim that the preamble was "abortion-neutral," and "merely determine[d] when life begins in a nonabortion context, a traditional state prerogative." 851 F.2d, at 1076. The court thought that "[t]he only plausible inference" from the fact that "every remaining section of the bill save one regulates the performance of abortions" was that "the state intended its abortion regulations to be understood against the backdrop of its theory of life." Ibid.5

The State contends that the preamble itself is precatory and imposes no substantive restrictions on abortions, and that appellees therefore do n t have standing to challenge it. Brief for Appellants 21-24. Appellees, on the other hand, insist that the preamble is an operative part of the Act intended to guide the interpretation of other provisions of the Act. Brief for Appellees 19-23. They maintain, for example, that the preamble's definition of life may prevent physicians in public hospitals from dispensing certain forms of contraceptives, such as the intrauterine device. Id., at 22.

In our view, the Court of Appeals misconceived the meaning of the Akron dictum, which was only that a State could not "justify" an abortion regulation otherwise invalid under Roe v. Wade on the ground that it embodied the State's view about when life begins. Certainly the preamble does not by its terms regulate abortion or any other aspect of appellees' medical practice. The Court has emphasized that Roe v. Wade "implies no limitation on the authority of a State to make a value judgment favoring childbirth over abortion." Maher v. Roe, 432 U.S., at 474, 97 S.Ct., at 2382-83. The preamble can be read simply to express that sort of value judgment.

We think the extent to which the preamble's language might be used to interpret other state statutes or regulations is something that only the courts of Missouri can definitively decide. State law has offered protections to unborn children in tort and probate law, see Roe v. Wade, supra, 410 U.S., at 161-162, 93 S.Ct., at 730-731, and § 1.205.2 can be interpreted to do no more than that. What we have, then, is much the same situation that the Court confronted in Alabama State Federation of Labor v. McAdory, 325 U.S. 450, 65 S.Ct. 1384, 89 L.Ed. 1725 (1945). As in that case:

"We are thus invited to pass upon the constitutional validity of a state statute which has not yet been applied or threatened to be applied by the state courts to petitioners or others in the manner anticipated. Lacking any authoritative construction of the statute by the state courts, without which no constitutional question arises, and lacking the authority to give such a controlling construction ourselves, and with a record which presents no concrete set of facts to which the statute is to be applied, the case is plainly not one to be disposed of by the declaratory judgment procedure." Id., at 460, 65 S.Ct., at 1389.

It will be time enough for federal courts to address the meaning of the preamble should it be applied to restrict the activities of appellees in some concrete way. Until then, this Court "is not empowered to decide . . . abstract propositions, or to declare, for the government of future cases, principles or rules of law which cannot affect the result as to the thing in issue in the case before it." Tyler v. Judges of Court of Registration, 179 U.S. 405, 409, 21 S.Ct. 206, 208, 45 L.Ed. 252 (1900). See also Valley Forge Christian College v. Americans United for Separation of Church & State, Inc., 454 U.S. 464, 473, 102 S.Ct. 752, 759, 70 L.Ed.2d 700 (1982).6 We therefore need not pass on the constitutionality of the Act's preamble.

Section 188.210 provides that "[i]t shall be unlawful for any public employee within the scope of his employment to perform or assist an abortion, not necessary to save the life of the mother," while § 188.215 makes it "unlawful for any public facility to be used for the purpose of performing or assisting an abortion not necessary to save the life of the mother."7 The Court of Appeals held that these provisions contravened this Court's abortion decisions. 851 F.2d, at 1082-1083. We take the contrary view.

As we said earlier this Term in DeShaney v. Winnebago County Dept. of Social Services, 489 U.S. 189, 196, 109 S.Ct. 998, 1003, 103 L.Ed.2d 249 (1989); "[O]ur cases have recognized that the Due Process Clauses generally confer no affirmative right to governmental aid, even where such aid may be necessary to secure life, liberty, or property interests of which the government itself may not deprive the individual." In Maher v. Roe, supra, the Court upheld a Connecticut welfare regulation under which Medicaid recipients received payments for medical services related to childbirth, but not for nontherapeutic abortions. The Court rejected the claim that this unequal subsidization of childbirth and abortion was impermissible under Roe v. Wade. As the Court put it:

"The Connecticut regulation before us is different in kind from the laws invalidated in our previous abortion decisions. The Connecticut regulation places no obstacles absolute or otherwise—in the pregnant woman's path to an abortion. An indigent woman who desires an abortion suffers no disadvantage as a consequence of Connecticut's decision to fund childbirth; she continues as before to be dependent on private sources for the service she desires. The State may have made childbirth a more attractive alternative, thereby influencing the woman's decision, but it has imposed no restriction on access to abortions that was not already there. The indigency that may make it difficult—and in some cases, perhaps, impossible—for some women to have abortions is neither created nor in any way affected by the Connecticut regulation." 432 U.S., at 474, 97 S.Ct., at 2382-83.

Relying on Maher, the Court in Poelker v. Doe, 432 U.S. 519, 521, 97 S.Ct. 2391, 2392 (1977), held that the city of St. Louis committed "no constitutional violation . . . in electing, as a policy choice, to provide publicly financed hospital services for childbirth without providing corresponding services for nontherapeutic abortions."

More recently, in Harris v. McRae, 448 U.S. 297, 100 S.Ct. 2671, 65 L.Ed.2d 784 (1980), the Court upheld "the most restrictive version of the Hyde Amendment," id., at 325, n. 27, 100 S.Ct., at 2692, n. 27, which withheld from States federal funds under the Medicaid program to reimburse the costs of abortions, " 'except where the life of the mother would be endangered if the fetus were carried to term.' " Ibid. (quoting Pub.L. 94-439, § 209, 90 Stat. 1434). As in Maher and Poelker, the Court required only a showing that Congress' authorization of "reimbursement for medically necessary services generally, but not for certain medically necessary abortions" was rationally related to the legitimate governmental goal of encouraging childbirth. 448 U.S., at 325, 100 S.Ct., at 2692-93.

The Court of Appeals distinguished these cases on the ground that "[t]o prevent access to a public facility does more than demonstrate a political choice in favor of childbirth; it clearly narrows and in some cases forecloses the availability of abortion to women." 851 F.2d, at 1081. The court reasoned that the ban on the use of public facilities "could prevent a woman's chosen doctor from performing an abortion because of his unprivileged status at other hospitals or because a private hospital adopted a similar anti-abortion stance." Ibid. It also thought that "[s]uch a rule could increase the cost of obtaining an abortion and delay the timing of it as well." Ibid.

We think that this analysis is much like that which we rejected in Maher, Poelker, and McRae. As in those cases, the State's decision here to use public facili ies and staff to encourage childbirth over abortion "places no governmental obstacle in the path of a woman who chooses to terminate her pregnancy." McRae, 448 U.S., at 315, 100 S.Ct., at 2687. Just as Congress' refusal to fund abortions in McRae left "an indigent woman with at least the same range of choice in deciding whether to obtain a medically necessary abortion as she would have had if Congress had chosen to subsidize no health care costs at all," id., at 317, 100 S.Ct., at 2688, Missouri's refusal to allow public employees to perform abortions in public hospitals leaves a pregnant woman with the same choices as if the State had chosen not to operate any public hospitals at all. The challenged provisions only restrict a woman's ability to obtain an abortion to the extent that she chooses to use a physician affiliated with a public hospital. This circumstance is more easily remedied, and thus considerably less burdensome, than indigency, which "may make it difficult—and in some cases, perhaps, impossible—for some women to have abortions" without public funding. Maher, 432 U.S., at 474, 97 S.Ct., at 2383. Having held that the State's refusal to fund abortions does not violate Roe v. Wade, it strains logic to reach a contrary result for the use of public facilities and employees. If the State may "make a value judgment favoring childbirth over abortion and . . . implement that judgment by the allocation of public funds," Maher, supra, at 474, 97 S.Ct., at 2382, surely it may do so through the allocation of other public resources, such as hospitals and medical staff.

The Court of Appeals sought to distinguish our cases on the additional ground that "[t]he evidence here showed that all of the public facility's costs in providing abortion services are recouped when the patient pays." 851 F.2d, at 1083. Absent any expenditure of public funds, the court thought that Missouri was "expressing" more than "its preference for childbirth over abortions," but rather was creating an "obstacle to exercise of the right to choose an abortion [that could not] stand absent a compelling state interest." Ibid. We disagree.

"Constitutional concerns are greatest," we said in Maher, supra, 432 U.S., at 476, 97 S.Ct., at 2383, "when the State attempts to impose its will by the force of law; the State's power to encourage actions deemed to be in the public interest is necessarily far broader." Nothing in the Constitution requires States to enter or remain in the business of performing abortions. Nor, as appellees suggest, do private physicians and their patients have some kind of constitutional right of access to public facilities for the performance of abortions. Brief for Appellees 46-47. Indeed, if the State does recoup all of its costs in performing abortions, and no state subsidy, direct or indirect, is available, it is difficult to see how any procreational choice is burdened by the State's ban on the use of its facilities or employees for performing abortions.8

Maher, Poelker, and McRae all support the view that the State need not commit any resources to facilitating abortions, even if it can turn a profit by doing so. In Poelker, the suit was filed by an indigent who could not afford to pay for an abortion, but the ban on the performance of nontherapeutic abortions in city-owned hospitals applied whether or not the pregnant woman could pay. 432 U.S., at 520, 97 S.Ct., at 2392; id., at 524, 97 S.Ct., at 2393-94 (BRENNAN, J., dissenting).9 The Court emphasized that the mayor's decision to prohibit abortions in city hospitals was "subject to public debate and approval or disapproval at the polls," and that "the Constitution does not forbid a State or city, pursuant to democratic processes, from expressing a preference for normal childbirth as St. Louis has done." Id., at 521, 97 S.Ct., at 2392. Thus we uphold the Act's restrictions on the use of public employees and facilities for the performance or assistance of nontherapeutic abortions.

The Missouri Act contains three provisions relating to "encouraging or counseling a woman to have an abortion not necessary to save her life." Section 188.205 states that no public funds can be used for this purpose; § 188.210 states that public employees cannot, within the scope of their employment, engage in such speech; and § 188.215 forbids such speech in public facilities. The Court of Appeals did not consider § 188.205 separately from §§ 188.210 and 188.215. It held that all three of these provisions were unconstitutionally vague, and that "the ban on using public funds, employees, and facilities to encourage or counsel a woman to have an abortion is an unacceptable infringement of the woman's fourteenth amendment right to choose an abortion after receiving the medical information necessary to exercise the right knowingly and intelligently." 851 F.2d, at 1079.10

Missouri has chosen only to appeal the Court of Appeals' invalidation of the public funding provision, § 188.205. See Juris. Statement I-II. A threshold question is whether this provision reaches primary conduct, or whether it is simply an instruction to the State's fiscal officers not to allocate funds for abortion counseling. We accept, for purposes of decision, the State's claim that § 188.205 "is not directed at the conduct of any physician or health care provider, private or public," but "is directed solely at those persons responsible for expending public funds." Brief for Appellants 43.11

Appellees contend that they are not "adversely" affected under the State's interpretation of § 188.205, and therefore that there is no longer a case or controversy before us on this question. Brief for Appellees 31-32. Plaintiffs are masters of their complaints and remain so at the appellate stage of a litigation. See Caterpillar Inc. v. Williams, 482 U.S. 386, 398-399, 107 S.Ct. 2425, 2432-2433, 96 L.Ed.2d 318 (1987). A majority of the Court agrees with appellees that the controversy over § 188.205 is now moot, because appellees' argument amounts to a decision to no longer seek a declaratory judgment that § 188.205 is unconstitutional and accompanying declarative relief. See Deakins v. Monaghan, 484 U.S. 193, 199-201, 108 S.Ct. 523, 528-529, 98 L.Ed.2d 529 (1988); United States v. Munsingwear, Inc., 340 U.S. 36, 39-40, 71 S.Ct. 104, 106-107, 95 L.Ed. 36 (1950). We accordingly direct the Court of Appeals to vacate the judgment of the District Court with instructions to dismiss the relevant part of the complaint. Deakins, 484 U.S., at 200, 108 S.Ct., at 528. "Because this [dispute] was rendered moot in part by [appellees'] willingness permanently to withdraw their equitable claims from their federal action, a dismissal with prejudice is indicated." Ibid.

Section 188.029 of the Missouri Act provides:

"Before a physician performs an abortion on a woman he has reason to believe is carrying an unborn child of twenty or more weeks gestational age, the physician shall first determine if the unborn child is viable by using and exercising that degree of care, skill, and proficiency commonly exercised by the ordinarily skillful, careful, and prudent physician engaged in similar practice under the same or similar conditions. In making this determination of viability, the physician shall perform or cause to be performed such medical examinations and tests as are necessary to make a finding of the gestational age, weight, and lung maturity of the unborn child and shall enter such findings and determination of viability in the medical record of the mother."12

As with the preamble, the parties disagree over the meaning of this statutory provision. The State emphasizes the language of the first sentence, which speaks in terms of the physician's determination of viability being made by the standards of ordinary skill in the medical profession. Brief for Appellants 32-35. Appellees stress the language of the second sentence, which prescribes such "tests as are necessary" to make a finding of gestational age, fetal weight, and lung maturity. Brief for Appellees 26-30.

The Court of Appeals read § 188.029 as requiring that after 20 weeks "doctors must perform tests to find gestational age, fetal weight and lung maturity." 851 F.2d, at 1075, n. 5. The court indicated that the tests needed to determine fetal weight at 20 weeks are "unreliable and inaccurate" and would add $125 to $250 to the cost of an abortion. Ibid. It also stated that "amniocentesis, the only method available to determine lung maturity, is contrary to accepted medical practice until 28-30 weeks of gestation, expensive, and imposes significant health risks for both the pregnant woman and the fetus." Ibid.

We must first determine the meaning of § 188.029 under Missouri law. Our usual practice is to defer to the lower court's construction of a state statute, but we believe the Court of Appeals has "fallen into plain error" in this case. Frisby v. Schultz, 487 U.S. 474, 483, 108 S.Ct. 2495, 2501, 101 L.Ed.2d 420 (1988); see Brockett v. Spokane Arcades, Inc., 472 U.S. 491, 500, n. 9, 105 S.Ct. 2794, 2800, n. 9, 86 L.Ed.2d 394 (1985). " 'In expounding a statute, we must not be guided by a single sentence or member of a sentence, but look to the provisions of the whole law, and to its object and policy.' " Philbrook v. Glodgett, 421 U.S. 707, 713, 95 S.Ct. 1893, 1898, 44 L.Ed.2d 525 (1975), quoting United States v. Heirs of Boisdore, 8 How. 113, 122, 12 L.Ed. 1009 (1849). See Chemehuevi Tribe of Indians v. FPC, 420 U.S. 395, 402-403, 95 S.Ct. 1066, 1071-1072, 43 L.Ed.2d 279 (1975); Kokoszka v. Belford, 417 U.S. 642, 650, 94 S.Ct. 2431, 2436, 41 L.Ed.2d 374 (1974). The Court of Appeals' interpretation also runs "afoul of the well-established principle that statutes will be interpreted to avoid constitutional difficulties." Frisby, supra, 487 U.S., at 483, 108 S.Ct., at 2501.

We think the viability-testing provision makes sense only if the second sentence is read to require only those test that are useful to making subsidiary findings as to viability. If we construe this provision to require a physician to perform those tests needed to make the three specified findings in all circumstances, including when the physician's reasonable professional judgment indicates that the tests would be irrelevant to determining viability or even dangerous to the mother and the fetus, the second sentence of § 188.029 would conflict with the first sentence's requirement that a physician apply his reasonable professional skill and judgment. It would also be incongruous to read this provision, especially the word "necessary,"13 to require the performance of tests irrelevant to the expressed statutory purpose of determining viability. It thus seems clear to us that the Court of Appeals' construction of § 188.029 violates well-accepted canons of statutory interpretation used in the Missouri courts, see State ex rel. Stern Brothers & Co. v. Stilley, 337 S.W.2d 934, 939 (Mo.1960) ("The basic rule of statutory construction is to first seek the legislative intention, and to effectuate it if possible, and the law favors constructions which harmonize with reason, and which tend to avoid unjust, absurd, unreasonable or confiscatory results, or oppression"); Bell v. Mid-Century Ins. Co., 750 S.W.2d 708, 710 (Mo.App.1988) ("Interpreting the phrase literally would produce an absurd result, which the Legislature is strongly presumed not to have intended"), which Justice BLACKMUN ignores. Post, at 545-546.

The viability-testing provision of the Missouri Act is concerned with promoting the State's interest in potential human life rather than in maternal health. Section 188.029 creates what is essentially a presumption of viability at 20 weeks, which the physician must rebut with tests indicating that the fetus is not viable prior to performing an abortion. It also directs the physician's determination as to viability by specifying consideration, if feasible, of gestational age, fetal weight, and lung capacity. The District Court found that "the medical evidence is uncontradicted that a 20-week fetus is not viable," and that "231/2 to 24 weeks gestation is the earliest point in pregnancy where a reasonable possibility of viability exists." 662 F.Supp., at 420. But it also found that there may be a 4-week error in estimating gestational age, id., at 421, which supports testing at 20 weeks.

In Roe v. Wade, the Court recognized that the State has "important and legitimate" interests in protecting maternal health and in the potentiality of human life. 410 U.S., at 162, 93 S.Ct., at 731. During the second trimester, the State "may, if it chooses, regulate the abortion procedure in ways that are reasonably related to maternal health." Id., at 164, 93 S.Ct., at 732. After viability, when the State's interest in potential human life was held to become compelling, the State "may, if it chooses, regulate, and even proscribe, abortion except where it is necessary, in appropriate medical judgment, for the preservation of the life or health of the mother." Id., at 165, 93 S.Ct., at 732.14

In Colautti v. Franklin, 439 U.S. 379, 99 S.Ct. 675, 58 L.Ed.2d 596 (1979), upon which appellees rely, the Court held that a Pennsylvania statute regulating the standard of care to be used by a physician performing an abortion of a possibly viable fetus was void for vagueness. Id., at 390-401, 99 S.Ct., at 683-689. But in the course of reaching that conclusion, the Court reaffirmed its earlier statement in Planned Parenthood of Central Mo. v. Danforth, 428 U.S. 52, 64, 96 S.Ct. 2831, 2838-39, 49 L.Ed.2d 788 (1976), that " 'the determination of whether a particular fetus is viable is, and must be, a matter for the judgment of the responsible attending physician.' " 439 U.S., at 396, 99 S.Ct., at 686. Justice BLACKMUN, post, at 545, n. 6, ignores the statement in Colautti that "neither the legislature nor the courts may proclaim one of the elements entering into the ascertainment of viability—be it weeks of gestation or fetal weight or any other single factor—as the determinant of when the State has a compelling interest in the life or health of the fetus." 439 U.S., at 388-389, 99 S.Ct., at 682. To the extent that § 188.029 regulates the method for determining viability, it undoubtedly does superimpose state regulation on the medical determination whether a particular fetus is viable. The Court of Appeals and the District Court thought it unconstitutional for this reason. 851 F.2d, at 1074-1075; 662 F.Supp., at 423. To the extent that the viability tests increase the cost of what are in fact second-trimester abortions, their validity may also be questioned under Akron, 462 U.S., at 434-435, 103 S.Ct., at 2494-2495, where the Court held that a requirement that second-trimester abortions must be performed in hospitals was invalid because it substantially increased the expense of those procedures.

We think that the doubt cast upon the Missouri statute by these cases is not so much a flaw in the statute as it is a reflection of the fact that the rigid trimester analysis of the course of a pregnancy enunciated in Roe has resulted in subsequent cases like Colautti and Akron making constitutional law in this area a virtual Procrustean bed. Statutes specifying elements of informed consent to be provided abortion patients, for example, were invalidated if they were thought to "structur[e] . . . the dialogue between the woman and her physician." Thornburgh v. American College of Obstetricians and Gynecologists, 476 U.S. 747, 763, 106 S.Ct. 2169, 2180, 90 L.Ed.2d 779 (1986). As the dissenters in Thornburgh pointed out, such a statute would have been sustained under any traditional standard of judicial review, id., at 802, 106 S.Ct., at 2200 (WHITE, J., dissenting), or for any other surgical procedure except abortion. Id., at 783, 106 S.Ct., at 2190-91 (Burger, C.J., dissenting).

Stare decisis is a cornerstone of our legal system, but it has less power in constitutional cases, where, save for constitutional amendments, this Court is the only body able to make needed changes. See United States v. Scott, 437 U.S. 82, 101, 98 S.Ct. 2187, 2199, 57 L.Ed.2d 65 (1978). We have not refrained from reconsideration of a prior construction of the Constitution that has proved "unsound in principle and unworkable in practice." Garcia v. San Antonio Metropolitan Transit Authority, 469 U.S. 528, 546, 105 S.Ct. 1005, 1015, 83 L.Ed.2d 1016 (1985); see Solorio v. United States, 483 U.S. 435, 448-450, 107 S.Ct. 2924, 2931-2933, 97 L.Ed.2d 364 (1987); Erie R. Co. v. Tompkins, 304 U.S. 64, 74-78, 58 S.Ct. 817, 820-822, 82 L.Ed. 1188 (1938). We think the Roe trimester framework falls into that category.

In the first place, the rigid Roe framework is hardly consistent with the notion of a Constitution cast in general terms, as ours is, and usually speaking in general principles, as ours does. The key elements of the Roe framework—trimesters and viability—are not found in the text of the Constitution or in any place else one would expect to find a constitutional principle. Since the bounds of the inquiry are essentially indeterminate, the result has been a web of legal rules that have become increasingly intricate, resembling a code of regulations rather than a body of constitutional doctrine.15 As Justice WHITE has put it, the trimester framework has left this Court to serve as the country's "ex officio medical board with powers to approve or disapprove medical and operative practices and standards throughout the United States." Planned Parenthood of Central Mo. v. Danforth, 428 U.S., at 99, 96 S.Ct., at 2854 (opinion concurring in part and dissenting in part). Cf. Garcia, supra, 469 U.S., at 547, 105 S.Ct., at 1015-16.

In the second place, we do not see why the State's interest in protecting potential human life should come into existence only at the point of viability, and that there should therefore be a rigid line allowing state regulation after viability but prohibiting it before viability. The dissenters in Thornburgh, writing in the context of the Roe trimester analysis, would have recognized this fact by positing against the "fundamental right" recognized in Roe the State's "compelling interest" in protecting potential human life throughout pregnancy. "[T]he State's interest, if compelling after viability, is equally compelling before viability." Thornburgh, 476 U.S., at 795, 106 S.Ct., at 2197 (WHITE, J., dissenting); see id., at 828, 106 S.Ct., at 2214 (O'CONNOR, J., dissenting) ("State has compelling interests in ensuring maternal health and in protecting potential human life, and these interests exist 'throughout pregnancy' ") (citation omitted).

The tests that § 188.029 requires the physician to perform are designed to determine viability. The Stat here has chosen viability as the point at which its interest in potential human life must be safeguarded. See Mo.Rev.Stat. § 188.030 (1986) ("No abortion of a viable unborn child shall be performed unless necessary to preserve the life or health of the woman"). It is true that the tests in question increase the expense of abortion, and regulate the discretion of the physician in determining the viability of the fetus. Since the tests will undoubtedly show in many cases that the fetus is not viable, the tests will have been performed for what were in fact second-trimester abortions. But we are satisfied that the requirement of these tests permissibly furthers the State's interest in protecting potential human life, and we therefore believe § 188.029 to be constitutional.

Justice BLACKMUN takes us to task for our failure to join in a "great issues" debate as to whether the Constitution includes an "unenumerated" general right to privacy as recognized in cases such as Griswold v. Connecticut, 381 U.S. 479, 85 S.Ct. 1678, 14 L.Ed.2d 510 (1965), and Roe. But Griswold v. Connecticut, unlike Roe, did not purport to adopt a whole framework, complete with detailed rules and distinctions, to govern the cases in which the asserted liberty interest would apply. As such, it was far different from the opinion, if not the holding, of Roe v. Wade, which sought to establish a constitutional framework for judging state regulation of abortion during the entire term of pregnancy. That framework sought to deal with areas of medical practice traditionally subject to state regulation, and it sought to balance once and for all by reference only to the calendar the claims of the State to protect the fetus as a form of human life against the claims of a woman to decide for herself whether or not to abort a fetus she was carrying. The experience of the Court in applying Roe v. Wade in later cases, see supra, at 518, n. 15, suggests to us that there is wisdom in not unnecessarily attempting to elaborate the abstract differences between a "fundamental right" to abortion, as the Court described it in Akron, 462 U.S., at 420, n. 1, 103 S.Ct., at 2487, n. 1, a "limited fundamental constitutional right," which Justice BLACKMUN today treats Roe as having established, post, at 555, or a liberty interest protected by the Due Process Clause, which we believe it to be. The Missouri testing requirement here is reasonably designed to ensure that abortions are not performed where the fetus is viable—an end which all concede is legitimate—and that is sufficient to sustain its constitutionality.

Justice BLACKMUN also accuses us, inter alia, of cowardice and illegitimacy in dealing with "the most politically divisive domestic legal issue of our time." Post, at 559. There is no doubt that our holding today will allow some governmental regulation of abortion that would have been prohibited under the language of cases such as Colautti v. Franklin, 439 U.S. 379, 99 S.Ct. 675, 58 L.Ed.2d 596 (1979), and Akron v. Akron Center for Reproductive Health, Inc., supra. But the goal of constitutional adjudication is surely not to remove inexorably "politically divisive" issues from the ambit of the legislative process, whereby the people through their elected representatives deal with matters of concern to them. The goal of constitutional adjudication is to hold true the balance between that which the Constitution puts beyond the reach of the democratic process and that which it does not. We think we have done that today. Justice BLACKMUN's suggestion, post, at 538, 557-558, that legislative bodies, in a Nation where more than half of our population is women, will treat our decision today as an invitation to enact abortion regulation reminiscent of the Dark Ages not only misreads our views but does scant justice to those who serve in such bodies and the people who elect them.

Both appellants and the United States as amicus curiae have urged that we overrule our decision in Roe v. Wade. Brief for Appellants 12-18; Brief for United States as Amicus Curiae 8-24. The facts of the present case, however, differ from those at issue in Roe. Here, Missouri has determined that viability is the point at which its interest in potential human life must be safeguarded. In Roe, on the other hand, the Texas statute criminalized the performance of all abortions, except when the mother's life was at stake. 410 U.S., at 117-118, 93 S.Ct., at 709. This case therefore affords us no occasion to revisit the holding of Roe, which was that the Texas statute unconstitutionally infringed the right to an abortion derived from the Due Process Clause, id., at 164, and we leave it undisturbed. To the extent indicated in our opinion, we would modify and narrow Roe and succeeding cases.

Because none of the challenged provisions of the Missouri Act properly before us conflict with the Constitution, the judgment of the Court of Appeals is

Reversed.

Justice O'CONNOR, concurring in part and concurring in the judgment.

I concur in Parts I, II-A, II-B, and II-C of the Court's opinion.

* Nothing in the record before us or the opinions below indicates that subsections 1(1) and 1(2) of the preamble to Missouri's abortion regulation statute will affect a woman's decision to have an abortion. Justice STEVENS, following appellees, see Brief for Appellees 22, suggests that the preamble may also "interfer[e] with contraceptive choices," post, at 564, because certain contraceptive devices act on a female ovum after it has been fertilized by a male sperm. The Missouri Act defines "conception" as "the fertilization of the ovum of a female by a sperm of a male," Mo.Rev.Stat. § 188.015(3) (1986), and invests "unborn children" with "protectable interests in life, health, and well-being," § 1.205.1(2), from "the moment of conception. . . ." § 1.205.3. Justice STEVENS asserts that any possible interference with a woman's right to use such postfertilization contraceptive devices would be unconstitutional under Griswold v. Connecticut, 381 U.S. 479, 85 S.Ct. 1678, 14 L.Ed.2d 510 (1965), and our subsequent contraception cases. Post, at 564-566. Similarly, certain amici suggest that the Missouri Act's preamble may prohibit the developing technology of in vitro fertilization, a technique used to aid couples otherwise unable to bear children in which a number of ova are removed from the woman and fertilized by male sperm. This process often produces excess fertilized ova ("unborn children" under the Missouri Act's definition) that are discarded rather than reinserted into the woman's uterus. Brief for Association of Reproductive Health Professionals et al. as Amici Curiae 38. It may be correct that the use of postfertilization contraceptive devices is constitutionally protected by Griswold and its progeny, but, as with a woman's abortion decision, nothing in the record or the opinions below indicates that the preamble will affect a woman's decision to practice contraception. For that matter, nothing in appellees' original complaint, App. 8-21, or their motion in limine to limit testimony and evidence on their challenge to the preamble, id., at 57-59, indicates that appellees sought to enjoin potential violations of Griswold. Neither is there any indication of the possibility that the preamble might be applied to prohibit the performance of in vitro fertilization. I agree with the Court, therefore, that all of these intimations of unconstitutionality are simply too hypothetical to support the use of declaratory judgment procedures and injunctive remedies in this case.

Similarly, it seems to me to follow directly from our previous decisions concerning state or federal funding of abortions, Harris v. McRae, 448 U.S. 297, 100 S.Ct. 2671, 65 L.Ed.2d 784 (1980), Maher v. Roe, 432 U.S. 464, 97 S.Ct. 2 76, 53 L.Ed.2d 484 (1977), and Poelker v. Doe, 432 U.S. 519, 97 S.Ct. 2391, 53 L.Ed.2d 528 (1977), that appellees' facial challenge to the constitutionality of Missouri's ban on the utilization of public facilities and the participation of public employees in the performance of abortions not necessary to save the life of the mother, Mo.Rev.Stat. §§ 188.210, 188.215 (1986), cannot succeed. Given Missouri's definition of "public facility" as "any public institution, public facility, public equipment, or any physical asset owned, leased, or controlled by this state or any agency or political subdivisions thereof," § 188.200(2), there may be conceivable applications of the ban on the use of public facilities that would be unconstitutional. Appellees and amici suggest that the State could try to enforce the ban against private hospitals using public water and sewage lines, or against private hospitals leasing state-owned equipment or state land. See Brief for Appellees 49-50; Brief for National Association of Public Hospitals as Amicus Curiae 9-12. Whether some or all of these or other applications of § 188.215 would be constitutional need not be decided here. Maher, Poelker, and McRae stand for the proposition that some quite straightforward applications of the Missouri ban on the use of public facilities for performing abortions would be constitutional and that is enough to defeat appellees' assertion that the ban is facially unconstitutional. "A facial challenge to a legislative Act is, of course, the most difficult challenge to mount successfully, since the challenger must establish that no set of circumstances exists under which the Act would be valid. The fact that the [relevant statute] might operate unconstitutionally under some conceivable set of circumstances is insufficient to render it wholly invalid, since we have not recognized an 'overbreadth' doctrine outside the limited context of the First Amendment." United States v. Salerno, 481 U.S. 739, 745, 107 S.Ct. 2095, 2100, 95 L.Ed.2d 697 (1987).

I also agree with the Court that, under the interpretation of § 188.205 urged by the State and adopted by the Court, there is no longer a case or controversy before us over the constitutionality of that provision. I would note, however, that this interpretation of § 188.205 is not binding on the Supreme Court of Missouri which has the final word on the meaning of that State's statutes. Virginia v. American Booksellers Assn., Inc., 484 U.S. 383, 395, 108 S.Ct. 636, 644, 98 L.Ed.2d 782 (1988); O'Brien v. Skinner, 414 U.S. 524, 531, 94 S.Ct. 740, 743-44, 38 L.Ed.2d 702 (1974). Should it happen that § 188.205, as ultimately interpreted by the Missouri Supreme Court, does prohibit publicly employed health professionals from giving specific medical advice to pregnant women, "the vacation and dismissal of the complaint that has become moot 'clears the path for future relitigation of the issues between the parties,' should subsequent events rekindle their controversy." Deakins v. Monaghan, 484 U.S. 193, 201, n. 5, 108 S.Ct. 523, 529, n. 5, 98 L.Ed.2d 529 (1988), quoting United States v. Munsingwear, Inc., 340 U.S. 36, 40, 71 S.Ct. 104, 107, 95 L.Ed. 36 (1950). Unless such events make their appearance and give rise to relitigation, I agree that we and all federal courts are without jurisdiction to hear the merits of this moot dispute.

In its interpretation of Missouri's "determination of viability" provision, Mo.Rev.Stat. § 188.029 (1986), see ante, at 513-521, the plurality has proceeded in a manner unnecessary to deciding the question at hand. I agree with the plurality that it was plain error for the Court of Appeals to interpret the second sentence of § 188.029 as meaning that "doctors must perform tests to find gestational age, fetal weight and lung maturity." 851 F.2d 1071, 1075, n. 5 (CA8 1988) (emphasis in original). When read together with the first sentence of § 188.029—which requires a physician to "determine if the unborn child is viable by using and exercising that degree of care, skill, and proficiency commonly exercised by the ordinary skillful, careful, and prudent physician engaged in similar practice under the same or similar conditions"—it would be contradictory nonsense to read the second sentence as requiring a physician to perform viability examinations and tests in situations where it would be careless and imprudent to do so. The plurality is quite correct: "the viability-testing provision makes sense only if the second sentence is read to require only those tests that are useful to making subsidiary findings as to viability," ante, at 514, and, I would add, only those examinations and tests that it would not be imprudent or careless to perform in the particular medical situation before the physician.

Unlike the plurality, I do not understand these viability testing requirements to conflict with any of the Court's past decisions concerning state regulation of abortion. Therefore, there is no necessity to accept the State's invitation to reexamine the constitutional validity of Roe v. Wade, 410 U.S. 113, 93 S.Ct. 705, 35 L.Ed.2d 147 (1973). Where there is no need to decide a constitutional question, it is a venerable principle of this Court's adjudicatory processes not to do so, for "[t]he Court will not 'anticipate a question of constitutional law in advance of the necessity of deciding it.' " Ashwander v. TVA, 297 U.S. 288, 346, 56 S.Ct. 466, 483, 80 L.Ed. 688 (1936) (Brandeis, J., concurring), quoting Liverpool, New York & Philadelphia S.S. Co. v. Commissioners of Emigration, 113 U.S. 33, 39, 5 S.Ct. 352, 355, 28 L.Ed. 899 (1885). Neither will it generally "formulate a rule of constitutional law broader than is required by the precise facts to which it is to be applied." 297 U.S., at 347, 56 S.Ct., at 483. Quite simply, "[i]t is not the habit of the court to decide questions of a constitutional nature unless absolutely necessary to a decision of the case." Burton v. United States, 196 U.S. 283, 295, 25 S.Ct. 243, 245, 49 L.Ed. 482 (1905). The Court today has accepted the State's every interpretation of its abortion statute and has upheld, under our existing precedents, every provision of that statute which is properly before us. Precisely for this reason reconsideration of Roe falls not into any "good-cause exception" to this "fundamental rule of judicial restraint. . . ." Three Affiliated Tribes of Fort Berthold Reservation v. Wold Engineering, P.C., 467 U.S. 138, 157, 104 S.Ct. 2267, 2279, 81 L.Ed.2d 113 (1984). See post, at 532-533 (SCALIA, J., concurring in part and concurring in judgment). When the constitutional invalidity of a State's abortion statute actually turns on the constitutional validity of Roe v. Wade, there will be time enough to reexamine Roe. And to do so carefully.

In assessing § 188.029 it is especially important to recognize that appellees did not appeal the District Court's ruling that the first sentence of § 188.029 is constitutional. 662 F.Supp. 407, 420-422 (WD Mo.1987). There is, accordingly, no dispute between the parties before us over the constitutionality of the "presumption of viability at 20 weeks," ante, at 515, created by the first sentence of § 188.029. If anything might arguably conflict with the Court's previous decisions concerning the determination of viability, I would think it is the introduction of this presumption. The plurality, see ante, at 515, refers to a passage from Planned Parenthood of Central Mo. v. Danforth, 428 U.S. 52, 64, 96 S.Ct. 2831, 2839, 49 L.Ed.2d 788 (1976): "The time when viability is achieved may vary with each pregnancy, and the determination of whether a particular fetus is viable is, and must be, a matter for the judgment of the responsible attending physician." The 20-week presumption of viability in the first sentence of § 188.029, it could be argued (though, I w uld think, unsuccessfully), restricts "the judgment of the responsible attending physician," by imposing on that physician the burden of overcoming the presumption. This presumption may be a "superimpos[ition] [of] state regulation on the medical determination whether a particular fetus is viable," ante, at 517, but, if so, it is a restriction on the physician's judgment that is not before us. As the plurality properly interprets the second sentence of § 188.029, it does nothing more than delineate means by which the unchallenged 20-week presumption of viability may be overcome if those means are useful in doing so and can be prudently employed. Contrary to the plurality's suggestion, see ante, at 517, the District Court did not think the second sentence of § 188.029 unconstitutional for this reason. Rather, both the District Court and the Court of Appeals thought the second sentence to be unconstitutional precisely because they interpreted that sentence to impose state regulation on the determination of viability that it does not impose.

Appellees suggest that the interpretation of § 188.029 urged by the State may "virtually eliminat[e] the constitutional issue in this case." Brief for Appellees 30. Appellees therefore propose that we should abstain from deciding that provision's constitutionality "in order to allow the state courts to render the saving construction the State has proposed." Ibid. Where the lower court has so clearly fallen into error I do not think abstention is necessary or prudent. Accordingly, I consider the constitutionality of the second sentence of § 188.029, as interpreted by the State, to determine whether the constitutional issue is actually eliminated.

I do not think the second sentence of § 188.029, as interpreted by the Court, imposes a degree of state regulation on the medical determination of viability that in any way conflicts with prior decisions of this Court. As the plurality recognizes, the requirement that, where not imprudent, physicians perform examinations and tests useful to making subsidiary findings to determine viability "promot[es] the State's interest in potential human life rather than in maternal health." Ante, at 515. No decision of this Court has held that the State may not directly promote its interest in potential life when viability is possible. Quite the contrary. In Thornburgh v. American College of Obstetricians and Gynecologists, 476 U.S. 747, 106 S.Ct. 2169, 90 L.Ed.2d 779 (1986), the Court considered a constitutional challenge to a Pennsylvania statute requiring that a second physician be present during an abortion performed "when viability is possible." Id., at 769-770, 106 S.Ct., at 2183. For guidance, the Court looked to the earlier decision in Planned Parenthood Assn. of Kansas City, Mo., Inc. v. Ashcroft, 462 U.S. 476, 103 S.Ct. 2517, 76 L.Ed.2d 733 (1983), upholding a Missouri statute requiring the presence of a second physician during an abortion performed after viability. Id., at 482-86, 103 S.Ct., at 2520-2523 (opinion of Powell, J.); id., at 505, 103 S.Ct., at 2532 (O'CONNOR, J., concurring in judgment in part and dissenting in part). The Thornburgh majority struck down the Pennsylvania statute merely because the statute had no exception for emergency situations and not because it found a constitutional difference between the State's promotion of its interest in potential life when viability is possible and when viability is certain. 476 U.S., at 770-771, 106 S.Ct., at 2183-2184. Despite the clear recognition by the Thornburgh majority that the Pennsylvania and Missouri statutes differed in this respect, there is no hint in the opinion of the Thornburgh Court that the State's interest in potential life differs depending on whether it seeks to further that interest postviability or when viability is possible. Thus, all nine Members of the Thornburgh Court appear to have agreed that it is not constitutionally impermissibl for the State to enact regulations designed to protect the State's interest in potential life when viability is possible. See id., at 811, 106 S.Ct., at 2205 (WHITE, J., dissenting); id., at 832, 106 S.Ct., at 2216 (O'CONNOR, J., dissenting). That is exactly what Missouri has done in § 188.029.

Similarly, the basis for reliance by the District Court and the Court of Appeals below on Colautti v. Franklin, 439 U.S. 379, 99 S.Ct. 675, 58 L.Ed.2d 596 (1979), disappears when § 188.029 is properly interpreted. In Colautti, the Court observed:

"Because this point [of viability] may differ with each pregnancy, neither the legislature nor the courts may proclaim one of the elements entering into the ascertainment of viability—be it weeks of gestation or fetal weight or any other single factor—as the determinant of when the State has a compelling interest in the life or health of the fetus. Viability is the critical point." Id., at 388-389, 99 S.Ct., at 682.

The courts below, on the interpretation of § 188.029 rejected here, found the second sentence of that provision at odds with this passage from Colautti. See 851 F.2d, at 1074; 662 F.Supp., at 423. On this Court's interpretation of § 188.029 it is clear that Missouri has not substituted any of the "elements entering into the ascertainment of viability" as "the determinant of when the State has a compelling interest in the life or health of the fetus." All the second sentence of § 188.029 does is to require, when not imprudent, the performance of "those tests that are useful to making subsidiary findings as to viability." Ante, at 514 (emphasis added). Thus, consistent with Colautti, viability remains the "critical point" under § 188.029.

Finally, and rather halfheartedly, the plurality suggests that the marginal increase in the cost of an abortion created by Missouri's viability testing provision may make § 188.029, even as interpreted, suspect under this Court's decision in Akron, 462 U.S., at 434-439, 103 S.Ct., at 2494-2497, striking down a second-trimester hospitalization requirement. See ante, at 517. I dissented from the Court's opinion in Akron because it was my view that, even apart from Roe's trimester framework which I continue to consider problematic, see Thornburgh, supra, 476 U.S., at 828, 106 S.Ct., at 2213-14 (dissenting opinion), the Akron majority had distorted and misapplied its own standard for evaluating state regulation of abortion which the Court had applied with fair consistency in the past: that, previability, "a regulation imposed on a lawful abortion is not unconstitutional unless it unduly burdens the right to seek an abortion." Akron, supra, 462 U.S., at 453, 103 S.Ct., at 2504 (dissenting opinion) (internal quotations omitted).

It is clear to me that requiring the performance of examinations and tests useful to determining whether a fetus is viable, when viability is possible, and when it would not be medically imprudent to do so, does not impose an undue burden on a woman's abortion decision. On this ground alone I would reject the suggestion that § 188.029 as interpreted is unconstitutional. More to the point, however, just as I see no conflict between § 188.029 and Colautti or any decision of this Court concerning a State's ability to give effect to its interest in potential life, I see no conflict between § 188.029 and the Court's opinion in Akron. The second-trimester hospitalization requirement struck down in Akron imposed, in the majority's view, "a heavy, and unnecessary, burden," 462 U.S., at 438, 103 S.Ct., at 2497, more than doubling the cost of "women's access to a relatively inexpensive, otherwise accessible, and safe abortion procedure." Ibid.; see also id., at 434, 103 S.Ct., at 2494-95. By contrast, the cost of examinations and tests that could usefully and prudently be performed when a woman is 20-24 weeks pregnant to determine whether the fe us is viable would only marginally, if at all, increase the cost of an abortion. See Brief for American Association of Prolife Obstetricians and Gynecologists et al. as Amici Curiae 3 ("At twenty weeks gestation, an ultrasound examination to determine gestational age is standard medical practice. It is routinely provided by the plaintiff clinics. An ultrasound examination can effectively provide all three designated findings of sec. 188.029"); id., at 22 ("A finding of fetal weight can be obtained from the same ultrasound test used to determine gestational age"); id., at 25 ("There are a number of different methods in standard medical practice to determine fetal lung maturity at twenty or more weeks gestation. The most simple and most obvious is by inference. It is well known that fetal lungs do not mature until 33-34 weeks gestation. . . . If an assessment of the gestational age indicates that the child is less than thirty-three weeks, a general finding can be made that the fetal lungs are not mature. This finding can then be used by the physician in making his determination of viability under section 188.029"); cf. Brief for American Medical Association et al. as Amici Curiae 42 (no suggestion that fetal weight and gestational age cannot be determined from the same sonogram); id., at 43 (another clinical test for gestational age and, by inference, fetal weight and lung maturity, is an accurate report of the last menstrual period), citing Smith, Frey, & Johnson, Assessing Gestational Age, 33 Am.Fam.Physician 215, 219-220 (1986).

Moreover, the examinations and tests required by § 188.029 are to be performed when viability is possible. This feature of § 188.029 distinguishes it from the second-trimester hospitalization requirement struck down by the Akron majority. As the Court recognized in Thornburgh, the State's compelling interest in potential life postviability renders its interest in determining the critical point of viability equally compelling. See supra, at 527-528. Under the Court's precedents, the same cannot be said for the Akron second-trimester hospitalization requirement. As I understand the Court's opinion in Akron, therefore, the plurality's suggestion today that Akron casts doubt on the validity of § 188.029, even as the Court has interpreted it, is without foundation and cannot provide a basis for reevaluating Roe. Accordingly, because the Court of Appeals misinterpreted § 188.029, and because, properly interpreted, § 188.029 is not inconsistent with any of this Court's prior precedents, I would reverse the decision of the Court of Appeals.

In sum, I concur in Parts I, II-A, II-B, and II-C of the Court's opinion and concur in the judgment as to Part II-D.

Justice SCALIA, concurring in part and concurring in the judgment.

I join Parts I, II-A, II-B, and II-C of the opinion of the Court. As to Part II-D, I share Justice BLACKMUN's view, post, at 556, that it effectively would overrule Roe v. Wade, 410 U.S. 113, 93 S.Ct. 705, 35 L.Ed.2d 147 (1973). I think that should be done, but would do it more explicitly. Since today we contrive to avoid doing it, and indeed to avoid almost any decision of national import, I need not set forth my reasons, some of which have been well recited in dissents of my colleagues in other cases. See, e.g., Thornburgh v. American College of Obstetricians and Gynecologists, 476 U.S. 747, 786-797, 106 S.Ct. 2169, 2192-2198, 90 L.Ed.2d 779 (1986) (WHITE, J., dissenting); Akron v. Akron Center for Reproductive Health, Inc., 462 U.S. 416, 453-459, 103 S.Ct. 2481, 2504-2508, 76 L.Ed.2d 687 (1983) (O'CONNOR, J., dissenting); Roe v. Wade, supra, 410 U.S., at 172-178, 93 S.Ct., at 736-739 (REHNQUIST, J., dissenting); Doe v. Bolton, 410 U.S. 179, 221-223, 93 S.Ct. 739, 762-63, 35 L.Ed.2d 201 (1973) (WHITE, J., dissenting).

The outcome of today's case will doubtless be heralded as a triumph of judicial statesmanship. I is not that, unless it is statesmanlike needlessly to prolong this Court's self-awarded sovereignty over a field where it has little proper business since the answers to most of the cruel questions posed are political and not juridical—a sovereignty which therefore quite properly, but to the great damage of the Court, makes it the object of the sort of organized public pressure that political institutions in a democracy ought to receive.

Justice O'CONNOR's assertion, ante, at 526, that a " 'fundamental rule of judicial restraint' " requires us to avoid reconsidering Roe, cannot be taken seriously. By finessing Roe we do not, as she suggests, ibid., adhere to the strict and venerable rule that we should avoid " 'decid[ing] questions of a constitutional nature.' " We have not disposed of this case on some statutory or procedural ground, but have decided, and could not avoid deciding, whether the Missouri statute meets the requirements of the United States Constitution. The only choice available is whether, in deciding that constitutional question, we should use Roe v. Wade as the benchmark, or something else. What is involved, therefore, is not the rule of avoiding constitutional issues where possible, but the quite separate principle that we will not " 'formulate a rule of constitutional law broader than is required by the precise facts to which it is to be applied.' " Ante, at 526. The latter is a sound general principle, but one often departed from when good reason exists. Just this Term, for example, in an opinion authored by Justice O'CONNOR, despite the fact that we had already held a racially based set-aside unconstitutional because unsupported by evidence of identified discrimination, which was all that was needed to decide the case, we went on to outline the criteria for properly tailoring race-based remedies in cases where such evidence is present. Richmond v. J.A. Croson Co., 488 U.S. 469, 506-508, 109 S.Ct. 706, 727-729, 102 L.Ed.2d 854 (1989). Also this Term, in an opinion joined by Justice O'CONNOR, we announced the constitutional rule that deprivation of the right to confer with counsel during trial violates the Sixth Amendment even if no prejudice can be shown, despite our finding that there had been no such deprivation on the facts before us—which was all that was needed to decide that case. Perry v. Leeke, 488 U.S. 272, 278-280, 109 S.Ct. 594, 599-600, 102 L.Ed.2d 624 (1989); see id., at 285, 109 S.Ct., at 602 (KENNEDY, J., concurring in part). I have not identified with certainty the first instance of our deciding a case on broader constitutional grounds than absolutely necessary, but it is assuredly no later than Marbury v. Madison, 1 Cranch 137, 2 L.Ed. 60 (1803), where we held that mandamus could constitutionally issue against the Secretary of State, although that was unnecessary given our holding that the law authorizing issuance of the mandamus by this Court was unconstitutional.

The Court has often spoken more broadly than needed in precisely the fashion at issue here, announcing a new rule of constitutional law when it could have reached the identical result by applying the rule thereby displaced. To describe two recent opinions that Justice O'CONNOR joined: In Daniels v. Williams, 474 U.S. 327, 106 S.Ct. 662, 88 L.Ed.2d 662 (1986), we overruled our prior holding that a "deprivation" of liberty or property could occur through negligent governmental acts, ignoring the availability of the alternative constitutional ground that, even if a deprivation had occurred, the State's postdeprivation remedies satisfied due process, see id., at 340-343, 106 S.Ct., at 679-681 (STEVENS, J., concurring in judgment). In Illinois v. Gates, 462 U.S. 213, 103 S.Ct. 2317, 76 L.Ed.2d 527 (1983), we replaced the pre-existing "two-pronged" constitutional test for probable cause with a totality-of-the-circumstances approach, ignoring the concurrence's argument that the same outcome could ave been reached under the old test, see id., at 267-272, 103 S.Ct., at 2347-2350 (WHITE, J., concurring in judgment). It is rare, of course, that the Court goes out of its way to acknowledge that its judgment could have been reached under the old constitutional rule, making its adoption of the new one unnecessary to the decision, but even such explicit acknowledgment is not unheard of. See Commonwealth Edison Co. v. Montana, 453 U.S. 609, 101 S.Ct. 2946, 69 L.Ed.2d 884 (1981); Perez v. Campbell, 402 U.S. 637, 91 S.Ct. 1704, 29 L.Ed.2d 233 (1971). For a sampling of other cases where the availability of a narrower, well-established ground is simply ignored in the Court's opinion adopting a new constitutional rule, though pointed out in separate opinions of some Justices, see Michelin Tire Corp. v. Wages, 423 U.S. 276, 96 S.Ct. 535, 46 L.Ed.2d 495 (1976); Pointer v. Texas, 380 U.S. 400 (1965); and Mapp v. Ohio, 367 U.S. 643, 81 S.Ct. 1684, 6 L.Ed.2d 1081 (1961). It would be wrong, in any decision, to ignore the reality that our policy not to "formulate a rule of constitutional law broader than is required by the precise facts" has a frequently applied good-cause exception. But it seems particularly perverse to convert the policy into an absolute in the present case, in order to place beyond reach the inexpressibly "broader-than-was-required-by-the-precise-facts" structure established by Roe v. Wade.

The real question, then, is whether there are valid reasons to go beyond the most stingy possible holding today. It seems to me there are not only valid but compelling ones. Ordinarily, speaking no more broadly than is absolutely required avoids throwing settled law into confusion; doing so today preserves a chaos that is evident to anyone who can read and count. Alone sufficient to justify a broad holding is the fact that our retaining control, through Roe, of what I believe to be, and many of our citizens recognize to be, a political issue, continuously distorts the public perception of the role of this Court. We can now look forward to at least another Term with carts full of mail from the public, and streets full of demonstrators, urging us—their unelected and life-tenured judges who have been awarded those extraordinary, undemocratic characteristics precisely in order that we might follow the law despite the popular will—to follow the popular will. Indeed, I expect we can look forward to even more of that than before, given our indecisive decision today. And if these reasons for taking the unexceptional course of reaching a broader holding are not enough, then consider the nature of the constitutional question we avoid: In most cases, we do no harm by not speaking more broadly than the decision requires. Anyone affected by the conduct that the avoided holding would have prohibited will be able to challenge it himself and have his day in court to make the argument. Not so with respect to the harm that many States believed, pre-Roe, and many may continue to believe, is caused by largely unrestricted abortion. That will continue to occur if the States have the constitutional power to prohibit it, and would do so, but we skillfully avoid telling them so. Perhaps those abortions cannot constitutionally be proscribed. That is surely an arguable question, the question that reconsideration of Roe v. Wade entails. But what is not at all arguable, it seems to me, is that we should decide now and not insist that we be run into a corner before we grudgingly yield up our judgment. The only sound reason for the latter course is to prevent a change in the law—but to think that desirable begs the question to be decided.

It was an arguable question today whether § 188.029 of the Missouri law contravened this Court's understanding of Roe v. Wade,* and I would have examined Roe rather than examining the contravention. Given the Court's newly contracted abstemiousness, what will it take, one must wonder, to permit us to reach that fundamental question? The result of our vote today is that we will not reconsider that prior opinion, even if most of the Justices think it is wrong, unless we have before us a statute that in fact contradicts it—and even then (under our newly discovered "no-broader-then-necessary" requirement) only minor problematical aspects of Roe will be reconsidered, unless one expects state legislatures to adopt provisions whose compliance with Roe cannot even be argued with a straight face. It thus appears that the mansion of constitutionalized abortion law, constructed overnight in Roe v. Wade, must be disassembled doorjamb by doorjamb, and never entirely brought down, no matter how wrong it may be.

Of the four courses we might have chosen oday—to reaffirm Roe, to overrule it explicitly, to overrule it sub silentio, or to avoid the question—the last is the least responsible. On the question of the constitutionality of § 188.029, I concur in the judgment of the Court and strongly dissent from the manner in which it has been reached.

Justice BLACKMUN, with whom Justice BRENNAN and Justice MARSHALL join, concurring in part and dissenting in part.

Today, Roe v. Wade, 410 U.S. 113, 93 S.Ct. 705, 35 L.Ed.2d 147 (1973), and the fundamental constitutional right of women to decide whether to terminate a pregnancy, survive but are not secure. Although the Court extricates itself from this case without making a single, even incremental, change in the law of abortion, the plurality and Justice SCALIA would overrule Roe (the first silently, the other explicitly) and would return to the States virtually unfettered authority to control the quintessentially intimate, personal, and life-directing decision whether to carry a fetus to term. Although today, no less than yesterday, the Constitution and the decisions of this Court prohibit a State from enacting laws that inhibit women from the meaningful exercise of that right, a plurality of this Court implicitly invites every state legislature to enact more and more restrictive abortion regulations in order to provoke more and more test cases, in the hope that sometime down the line the Court will return the law of procreative freedom to the severe limitations that generally prevailed in this country before January 22, 1973. Never in my memory has a plurality announced a judgment of this Court that so foments disregard for the law and for our standing decisions.

Nor in my memory has a plurality gone about its business in such a deceptive fashion. At every level of its review, from its effort to read the real meaning out of the Missouri statute, to its intended evisceration of precedents and its deafening silence about the constitutional protections that it would jettison, the plurality obscures the portent of its analysis. With feigned restraint, the plurality announces that its analysis leaves Roe "undisturbed," albeit "modif[ied] and narrow[ed]." Ante, at 521. But this disclaimer is totally meaningless. The plurality opinion is filled with winks, and nods, and knowing glances to those who would do away with Roe explicitly, but turns a stone face to anyone in search of what the plurality conceives as the scope of a woman's right under the Due Process Clause to terminate a pregnancy free from the coercive and brooding influence of the State. The simple truth is that Roe would not survive the plurality's analysis, and that the plurality provides no substitute for Roe's protective umbrella.

I fear for the future. I fear for the liberty and equality of the millions of women who have lived and come of age in the 16 years since Roe was decided. I fear for the integrity of, and public esteem for, this Court.

I dissent.

* THE CHIEF JUSTICE parades through the four challenged sections of the Missouri statute seriatim. I shall not do this, but shall relegate most of my comments as to those sections to the margin.1 Although I disagree with the Court's consideration of §§ 1.205, 188.210, and 188.215, and am especially disturbed by its misapplication of our past decisions in upholding Missouri's ban on the performance of abortions at "public facilities," its discussion of these provisions is merely prologue to the plurality's consideration of the statute's viability-testing requirement, § 188.029—the only section of the Missouri statute that the plurality construes as implicating Roe itself. There, tucked away at the end of its opinion, the plurality suggests a radical reversal of the law of abortion; and there, primarily, I direct my attention.

In the plurality's view, the viability-testing provision imposes a burden on second-trimester abortions as a way of furthering the State's interest in protecting the potential life of the fetus. Since under the Roe framework, the State may not fully regulate abortion in the interest of potential life (as opposed to maternal health) until the third trimester, the plurality finds it necessary, in order to save the Missouri testing provision, to throw out Roe's trimester framework. Ante, at 518-520. In flat contradiction to Roe, 410 U.S., at 163, 93 S.Ct., at 731, the plurality concludes that the State's interest in potential life is compelling before viability, and upholds the testing provision because it "permissibly furthers" that state interest. Ante, at 519.

At the outset, I note that in its haste to limit abortion rights, the plurality compounds the errors of its analysis by needlessly reaching out to address constitutional questions that are not actually presented. The conflict between § 188.029 and Roe's trimester framework, which purportedly drives the plurality to reconsider our past decisions, is a contrived conflic : the product of an aggressive misreading of the viability-testing requirement and a needlessly wooden application of the Roe framework.

The plurality's reading of § 188.029 is irreconcilable with the plain language of the statute and is in derogation of this Court's settled view that " 'district courts and courts of appeals are better schooled in and more able to interpret the laws of their respective States.' " Frisby v. Schultz, 487 U.S. 474, 482, 108 S.Ct. 2495, 2500, 101 L.Ed.2d 420 (1988), quoting Brockett v. Spokane Arcades, Inc., 472 U.S. 491, 499-500, 105 S.Ct. 2794, 2799, 86 L.Ed.2d 394 (1985). Abruptly setting aside the construction of § 188.029 adopted by both the District Court and Court of Appeals as "plain error," the plurality reads the viability-testing provision as requiring only that before a physician may perform an abortion on a woman whom he believes to be carrying a fetus of 20 or more weeks gestational age, the doctor must determine whether the fetus is viable and, as part of that exercise, must, to the extent feasible and consistent with sound medical practice, conduct tests necessary to make findings of gestational age, weight, and lung maturity. Ante, at 514-517. But the plurality's reading of the provision, according to which the statute requires the physician to perform tests only in order to determine viability, ignores the statutory language explicitly directing that "the physician shall perform or cause to be performed such medical examinations and tests as are necessary to make a finding of the gestational age, weight, and lung maturity of the unborn child and shall enter such findings" in the mother's medical record. § 188.029 (emphasis added). The statute's plain language requires the physician to undertake whatever tests are necessary to determine gestational age, weight, and lung maturity, regardless of whether these tests are necessary to a finding of viability, and regardless of whether the tests subject the pregnant woman or the fetus to additional health risks or add substantially to the cost of an abortion.2

Had the plurality read the statute as written, it would have had no cause to reconsider the Roe framework. As properly construed, the viability-testing provision does not pass constitutional muster under even a rational-basis standard, the least restrictive level of review applied by this Court. See Williamson v. Lee Optical Co., 348 U.S. 483, 75 S.Ct. 461, 99 L.Ed. 563 (1955). By mandating tests to determine fetal weight and lung maturity for every fetus thought to be more than 20 weeks gestational age, the statute requires physicians to undertake procedures, such as amniocentesis, that, in the situation presented, have no medical justification, impose significant additional health risks on both the pregnant woman and the fetus, and bear no rational relation to the State's interest in protecting fetal life.3 As written, § 188.029 is an arbitrary imposition of discomfort, risk, and expense, furthering no discernible interest except to make the procurement of an abortion as arduous and difficult as possible. Thus, were it not for the plurality's tortured effort to avoid the plain import of § 188.029, it could have struck down the testing provision as patently irrational irrespective of the Roe framework.4

The plurality eschews this straightforward resolution, in the hope of precipitating a constitutional crisis. Far from avoiding constitutional difficulty, the plurality attempts to engineer a dramatic retrenchment in our jurisprudence by exaggerating the conflict between its untenable construction of § 188.029 and the Roe trimester framework.

No one contests that under the Roe framework the State, in order to promote its interest in potential human life, may regulate and even proscribe nontherapeutic abortions once the fetus becomes viable. Roe, 410 U.S., at 164-165, 93 S.Ct., at 732. If, as the plurality appears to hold, the testing provision simply requires a physician to use appropriate and medically sound tests to determine whether the fetus is actually viable when the estimated gestational age is greater than 20 weeks (and therefore within what the District Court found to be the margin of error for viability, ante, at 515-516), then I see little or no conflict with Roe.5 Nothing in Roe, or any of its progeny, holds that a State may not effectuate its compelling interest in the potential life of a viable fetus by seeking to ensure that no viable fetus is mistakenly aborted because of the inherent lack of precision in estimates of gestational age. A requirement that a physician make a finding of viability, one way or the other, for every fetus that falls within the range of possible viability does no more than preserve the State's recognized authority. Although, as the plurality correctly points out, such a testing requirement would have the effect of imposing additional costs on second-trimester abortions where the tests indicated that the fetus was not viable, these costs would be merely incidental to, and a necessary accommodation of, the State's unquestioned right to prohibit non-therapeutic abortions after the point of viability. In short, the testing provision, as construed by the plurality, is consistent with the Roe framework and could be upheld effortlessly under current doctrine.6

How ironic it is, then, and disingenuous, that the plurality scolds the Court of Appeals for adopting a construction of the statute that fails to avoid constitutional difficulties. Ante, at 514, 515. By distorting the statute, the plurality manages to avoid invalidating the testing provision on what should have been noncontroversial constitutional grounds; having done so, however, the plurality rushes headlong into a much deeper constitutional thicket, brushing past an obvious basis for upholding § 188.029 in search of a pretext for scuttling the trimester framework. Evidently, from the plurality's perspective, the real problem with the Court of Appeals' construction of § 188.029 is not that it raised a constitutional difficulty, but that it raised the wrong constitutional difficulty—one not implicating Roe. The plurality has remedied that, traditional canons of construction and judicial forbearance notwithstanding.

Having set up the conflict between § 188.029 and the Roe trimester framework, the plurality summarily discards Roe's analytic core as " 'unsound in principle and unworkable in practice.' " Ante, at 518, quoting Garcia v. San Antonio Metropolitan Transit Authority, 469 U.S. 528, 546, 105 S.Ct. 1005, 1015, 83 L.Ed.2d 1016 (1985). This is so, the plurality claims, because the key elements of the framework do not appear in the text of the Constitution, because the framework more closely resembles a regulatory code than a body of constitutional doctrine, and because under the framework the State's interest in potential human life is considered compelling only after viability, when, in fact, that interest is equally compelling throughout pregnancy. Ante, at 519-520. The plurality does not bother to explain these alleged flaws in Roe. Bald assertion masquerades as reasoning. The object, quite clearly, is not to persuade, but to prevail.

* The plurality opinion is far more remarkable for the arguments that it does not advance than for those that it does. The plurality does not even mention, much less join, the true jurisprudential debate underlying this case: whether the Constitution includes an "unenumerated" general right to privacy as recognized in many of our decisions, most notably Griswold v. Connecticut, 381 U.S. 479, 85 S.Ct. 1678, 14 L.Ed.2d 510 (1965), and Roe, and, more specifically, whether, and to what extent, such a right to privacy extends to matters of childbearing and family life, including abortion. See, e.g., Eisenstadt v. Baird, 405 U.S. 438, 92 S.Ct. 1029, 31 L.Ed.2d 349 (1972) (contraception); Loving v. Virginia, 388 U.S. 1, 87 S.Ct. 1817, 18 L.Ed.2d 1010 (1967) (marriage); Skinner v. Oklahoma ex rel. Williamson, 316 U.S. 535, 62 S.Ct. 1110, 86 L.Ed. 1655 (1942) (pro reation); Pierce v. Society of Sisters, 268 U.S. 510, 45 S.Ct. 571, 69 L.Ed. 1070 (1925) (childrearing).7 These are questions of unsurpassed significance in this Court's interpretation of the Constitution, and mark the battleground upon which this case was fought, by the parties, by the United States as amicus on behalf of petitioners, and by an unprecedented number of amici. On these grounds, abandoned by the plurality, the Court should decide this case.

But rather than arguing that the text of the Constitution makes no mention of the right to privacy, the plurality complains that the critical elements of the Roe framework—trimesters and viability—do not appear in the Constitution and are, therefore, somehow inconsistent with a Constitution cast in general terms. Ante, at 518-519. Were this a true concern, we would have to abandon most of our constitutional jurisprudence. As the plurality well knows, or should know, the "critical elements" of countless constitutional doctrines nowhere appear in the Constitution's text. The Constitution makes no mention, for example, of the First Amendment's "actual malice" standard for proving certain libels, see New York Times Co. v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964), or of the standard for determining when speech is obscene. See Miller v. California, 413 U.S. 15, 93 S.Ct. 2607, 37 L.Ed.2d 419 (1973). Similarly, the Constitution makes no mention of the rational-basis test, or the specific verbal formulations of intermediate and strict scrutiny by which this Court evaluates claims under the Equal Protection Clause. The reason is simple. Like the Roe framework, these tests or standards are not, and do not purport to be, rights protected by the Constitution. Rather, they are judge-made methods for evaluating and measuring the strength and scope of constitutional rights or for balancing the constitutional rights of individuals against the competing interests of government.

With respect to the Roe framework, the general constitutional principle, indeed the fundamental constitutional right, for which it was developed is the right to privacy, see, e.g., Griswold v. Connecticut, 381 U.S. 479, 85 S.Ct. 1678, 14 L.Ed.2d 510 (1965), a species of "liberty" protected by the Due Process Clause, which under our past decisions safeguards the right of women to exercise some control over their own role in procreation. As we recently reaffirmed in Tho nburgh v. American College of Obstetricians and Gynecologists, 476 U.S. 747, 106 S.Ct. 2169, 90 L.Ed.2d 779 (1986), few decisions are "more basic to individual dignity and autonomy" or more appropriate to that "certain private sphere of individual liberty" that the Constitution reserves from the intrusive reach of government than the right to make the uniquely personal, intimate, and self-defining decision whether to end a pregnancy. Id., at 772, 106 S.Ct., at 2184. It is this general principle, the " 'moral fact that a person belongs to himself and not others nor to society as a whole,' " id., at 777, n. 5, 106 S.Ct., at 2187, n. 5 (STEVENS, J., concurring), quoting Fried, Correspondence, 6 Phil. & Pub. Aff. 288-289 (1977), that is found in the Constitution. See Roe, 410 U.S., at 152-153, 93 S.Ct., at 726. The trimester framework simply defines and limits that right to privacy in the abortion context to accommodate, not destroy, a State's legitimate interest in protecting the health of pregnant women and in preserving potential human life. Id., at 154-162, 93 S.Ct., at 727-31. Fashioning such accommodations between individual rights and the legitimate interests of government, establishing benchmarks and standards with which to evaluate the competing claims of individuals and government, lies at the very heart of constitutional adjudication. To the extent that the trimester framework is useful in this enterprise, it is not only consistent with constitutional interpretation, but necessary to the wise and just exercise of this Court's paramount authority to define the scope of constitutional rights.

The plurality next alleges that the result of the trimester framework has "been a web of legal rules that have become increasingly intricate, resembling a code of regulations rather than a body of constitutional doctrine." Ante, at 518. Again, if this were a true and genuine concern, we would have to abandon vast areas of our constitutional jurisprudence. The plurality complains that under the trimester framework the Court has distinguished between a city ordinance requiring that second-trimester abortions be performed in clinics and a state law requiring that these abortions be performed in hospitals, or between laws requiring that certain information be furnished to a woman by a physician or his assistant and those requiring that such information be furnished by the physician exclusively. Ante, at 518, n. 15, citing Simopoulos v. Virginia, 462 U.S. 506, 103 S.Ct. 2532, 76 L.Ed.2d 755 (1983), and Akron v. Akron Center for Reproductive Health, Inc., 462 U.S. 416, 103 S.Ct. 2481, 76 L.Ed.2d 687 (1983). Are these distinctions any finer, or more "regulatory," than the distinctions we have often drawn in our First Amendment jurisprudence, where, for example, we have held that a "release time" program permitting public-school students to leave school grounds during school hours to receive religious instruction does not violate the Establishment Clause, even though a release-time program permitting religious instruction on school grounds does violate the Clause? Compare Zorach v. Clauson, 343 U.S. 306, 72 S.Ct. 679, 96 L.Ed. 954 (1952), with Illinois ex rel. McCollum v. Board of Education of School Dist. No. 71, Champaign County, 333 U.S. 203, 68 S.Ct. 461, 92 L.Ed. 649 (1948). Our Fourth Amendment jurisprudence recognizes factual distinctions no less intricate. Just this Term, for example, we held that while an aerial observation from a helicopter hovering at 400 feet does not violate any reasonable expectation of privacy, such an expectation of privacy would be violated by a helicopter observation from an unusually low altitude. Florida v. Riley, 488 U.S. 445, 451, 109 S.Ct. 693, 697, 102 L.Ed.2d 835 (1989) (O'CONNOR, J., concurring in judgment). Similarly, in a Sixth Amendment case, the Court held that although an overnight ban on attorney-client communication violated the constitutionally guarant ed right to counsel, Geders v. United States, 425 U.S. 80, 96 S.Ct. 1330, 47 L.Ed.2d 592 (1976), that right was not violated when a trial judge separated a defendant from his lawyer during a 15-minute recess after the defendant's direct testimony. Perry v. Leeke, 488 U.S. 272, 109 S.Ct. 594, 102 L.Ed.2d 624 (1989).

That numerous constitutional doctrines result in narrow differentiations between similar circumstances does not mean that this Court has abandoned adjudication in favor of regulation. Rather, these careful distinctions reflect the process of constitutional adjudication itself, which is often highly fact specific, requiring such determinations as whether state laws are "unduly burdensome" or "reasonable" or bear a "rational" or "necessary" relation to asserted state interests. In a recent due process case, THE CHIEF JUSTICE wrote for the Court: "[M]any branches of the law abound in nice distinctions that may be troublesome but have been thought nonetheless necessary: 'I do not think we need trouble ourselves with the thought that my view depends upon differences of degree. The whole law does so as soon as it is civilized.' " Daniels v. Williams, 474 U.S. 327, 334, 106 S.Ct. 662, 667, 88 L.Ed.2d 662 (1986), quoting Le Roy Fibre Co. v. Chicago, M. & St. P.R. Co., 232 U.S. 340, 354, 34 S.Ct. 415, 418, 58 L.Ed. 631 (1914) (Holmes, J., partially concurring).

These "differences of degree" fully account for our holdings in Simopoulos, supra, and Akron, supra. Those decisions rest on this Court's reasoned and accurate judgment that hospitalization and doctor-counseling requirements unduly burdened the right of women to terminate a pregnancy and were not rationally related to the State's asserted interest in the health of pregnant women, while Virginia's substantially less restrictive regulations were not unduly burdensome and did rationally serve the State's interest.8 That the Court exercised its best judgment in evaluating these markedly different statutory schemes no more established the Court as an " 'ex officio medical board,' " ante, at 519, quoting Planned Parenthood of Central Mo. v. Danforth, 428 U.S. 52, 99, 96 S.Ct. 2831, 2854, 49 L.Ed.2d 788 (1976) (opinion of WHITE, J., concurring in part and dissenting in part), than our decisions involving religion in the public schools establish the Court as a national school board, or our decisions concerning prison regulations establish the Court as a bureau of prisons. See Thornburgh v. Abbott, 490 U.S. 401, 109 S.Ct. 1874, 104 L.Ed.2d 459 (1989) (adopting different standard of First Amendment review for incoming as opposed to outgoing prison mail). If, in delicate and complicated areas of constitutional law, our legal judgments "have become increasingly intricate," ante, at 518, it is not, as the plurality contends, because we have overstepped our judicial role. Quite the opposite: the rules are intricate because we have remained conscientious in our duty to do justice carefully, especially when fundamental rights rise or fall with our decisions.

Finally, the plurality asserts that the trimester framework cannot stand because the State's interest in potential life is compelling throughout pregnancy, not merely after viability. Ante, at 519. The opinion contains not one word of rationale for its view of the State's interest. This "it-is-so-because-we-say-so" jurisprudence constitutes nothing other than an attempted exercise of brute force; reason, much less persuasion, has no place.

In answering the plurality's claim that the State's interest in the fetus is uniform and compelling throughout pregnancy, I cannot improve upon what Justice STEVENS has written:

"I should think it obvious that the State's interest in the protection of an embryo—even if that interest is defined as 'protecting those who will be citizens' . . .—increases progressively and dramatically as the organism's capacity to feel pain, to experience pleasure, to survive, and to react to its surroundings increases day by day. The development of a fetus—and pregnancy itself—are not static conditions, and the assertion that the government's interest is static simply ignores this reality. . . . [U]nless the religious view that a fetus is a 'person' is adopted . . . there is a fundamental and well-recognized difference between a fetus and a human being; indeed, if there is not such a difference, the permissibility of terminating the life of a fetus could scarcely be left to the will of the state legislatures. And if distinctions may be drawn between a fetus and a human being in terms of the state interest in their protection—even though the fetus represents one of 'those who will be citizens'—it seems to me quite odd to argue that distinctions may not also be drawn between the state interest in protecting the freshly fertilized egg and the state interest in protecting the 9-month-gestated, fully sentient fetus on the eve of birth. Recognition of this distinction is supported not only by logic, but also by history and by our shared experiences." Thornburgh, 476 U.S., at 778-779, 106 S.Ct., at 2188 (footnote omitted).

See also Roe, 410 U.S., at 129-147, 93 S.Ct., at 715-24.

For my own part, I remain convinced, as six other Members of this Court 16 years ago were convinced, that the Roe framework, and the viability standard in particular, fairly, sensibly, and effectively functions to safeguard the constitutional liberties of pregnant women while recognizing and accommodating the State's interest in potential human life. The viability line reflects the biological facts and truths of fetal development; it marks that threshold moment prior to which a fetus cannot survive separate from the woman and cannot reasonably and objectively be regarded as a subject of rights or interests distinct from, or paramount to, those of the pregnant woman. At the same time, the viability standard takes account of the undeniable fact that as the fetus evolves into its postnatal form, and as it loses its dependence on the uterine environment, the State's interest in the fetus' potential human life, and in fostering a regard for human life in general, becomes compelling. As a practical matter, because viability follows "quickening"—the point at which a woman feels movement in her womb—and because viability occurs no earlier than 23 weeks gestational age, it establishes an easily applicable standard for regulating abortion while providing a pregnant woman ample time to exercise her fundamental right with her responsible physician to terminate her pregnancy.9 Although I have stated previously for a majority of this Court that "[c]onstitutional rights do not always have easily ascertainable boundaries," to seek and establish those boundaries remains the special responsibility of this Court. Thornburgh, 476 U.S., at 771, 106 S.Ct., at 2184. In Roe, we discharged that responsibility as logic and science compelled. The plurality today advances not one reasonable argument as to why our judgment i that case was wrong and should be abandoned.

Having contrived an opportunity to reconsider the Roe framework, and then having discarded that framework, the plurality finds the testing provision unobjectionable because it "permissibly furthers the State's interest in protecting potential human life." Ante, at 519-520. This newly minted standard is circular and totally meaningless. Whether a challenged abortion regulation "permissibly furthers" a legitimate state interest is the question that courts must answer in abortion cases, not the standard for courts to apply. In keeping with the rest of its opinion, the plurality makes no attempt to explain or to justify its new standard, either in the abstract or as applied in this case. Nor could it. The "permissibly furthers" standard has no independent meaning, and consists of nothing other than what a majority of this Court may believe at any given moment in any given case. The plurality's novel test appears to be nothing more than a dressed-up version of rational-basis review, this Court's most lenient level of scrutiny. One thing is clear, however: were the plurality's "permissibly furthers" standard adopted by the Court, for all practical purposes, Roe would be overruled.10

The "permissibly furthers" standard completely disregards the irreducible minimum of Roe: the Court's recognition that a woman has a limited fundamental constitutional right to decide whether to terminate a pregnancy. That right receives no meaningful recognition in the plurality's written opinion. Since, in the plurality's view, the State's interest in potential life is compelling as of the moment of conception, and is therefore served only if abortion is abolished, every hindrance to a woman's ability t obtain an abortion must be "permissible." Indeed, the more severe the hindrance, the more effectively (and permissibly) the State's interest would be furthered. A tax on abortions or a criminal prohibition would both satisfy the plurality's standard. So, for that matter, would a requirement that a pregnant woman memorize and recite today's plurality opinion before seeking an abortion.

The plurality pretends that Roe survives, explaining that the facts of this case differ from those in Roe: here, Missouri has chosen to assert its interest in potential life only at the point of viability, whereas, in Roe, Texas had asserted that interest from the point of conception, criminalizing all abortions, except where the life of the mother was at stake. Ante, at 521. This, of course, is a distinction without a difference. The plurality repudiates every principle for which Roe stands; in good conscience, it cannot possibly believe that Roe lies "undisturbed" merely because this case does not call upon the Court to reconsider the Texas statute, or one like it. If the Constitution permits a State to enact any statute that reasonably furthers its interest in potential life, and if that interest arises as of conception, why would the Texas statute fail to pass muster? One suspects that the plurality agrees. It is impossible to read the plurality opinion and especially its final paragraph, without recognizing its implicit invitation to every State to enact more and more restrictive abortion laws, and to assert their interest in potential life as of the moment of conception. All these laws will satisfy the plurality's non-scrutiny, until sometime, a new regime of old dissenters and new appointees will declare what the plurality intends: that Roe is no longer good law.11

Thus, "not with a bang, but a whimper," the plurality discards a landmark case of the last generation, and casts into darkness the hopes and visions of every woman in this country who had come to believe that the Constitution guaranteed her the right to exercise some control over her unique ability to bear children. The plurality does so either ob ivious or insensitive to the fact that millions of women, and their families, have ordered their lives around the right to reproductive choice, and that this right has become vital to the full participation of women in the economic and political walks of American life. The plurality would clear the way once again for government to force upon women the physical labor and specific and direct medical and psychological harms that may accompany carrying a fetus to term. The plurality would clear the way again for the State to conscript a woman's body and to force upon her a "distressful life and future." Roe, 410 U.S., at 153, 93 S.Ct., at 727.

The result, as we know from experience, see Cates & Rochat, Illegal Abortions in the United States: 1972-1974, 8 Family Planning Perspectives 86, 92 (1976), would be that every year hundreds of thousands of women, in desperation, would defy the law, and place their health and safety in the unclean and unsympathetic hands of back-alley abortionists, or they would attempt to perform abortions upon themselves, with disastrous results. Every year, many women, especially poor and minority women, would die or suffer debilitating physical trauma, all in the name of enforced morality or religious dictates or lack of compassion, as it may be.

Of the aspirations and settled understandings of American women, of the inevitable and brutal consequences of what it is doing, the tough-approach plurality utters not a word. This silence is callous. It is also profoundly destructive of this Court as an institution. To overturn a constitutional decision is a rare and grave undertaking. To overturn a constitutional decision that secured a fundamental personal liberty to millions of persons would be unprecedented in our 200 years of constitutional history. Although the doctrine of stare decisis applies with somewhat diminished force in constitutional cases generally, ante, at 518, even in ordinary constitutional cases "any departure from . . . stare decisis demands special justification." Arizona v. Rumsey, 467 U.S. 203, 212, 104 S.Ct. 2305, 2310, 81 L.Ed.2d 164 (1984). See also Vasquez v. Hillery, 474 U.S. 254, 266, 106 S.Ct. 617, 624, 88 L.Ed.2d 598 (1986) ("[T]he careful observer will discern that any detours from the straight path of stare decisis in our past have occurred for articulable reasons, and only when the Court has felt obliged 'to bring its opinions into agreement with experience and with facts newly ascertained,' " quoting Burnet v. Coronado Oil & Gas Co., 285 U.S. 393, 412, 52 S.Ct. 443, 449, 76 L.Ed. 815 (1932) (Brandeis, J., dissenting)). This requirement of justification applies with unique force where, as here, the Court's abrogation of precedent would destroy people's firm belief, based on past decisions of this Court, that they possess an unabridgeable right to undertake certain conduct.12

As discussed at perhaps too great length above, the plurality makes no serious attempt to carry "the heavy burden of persuading . . . that changes in society or in the law dictate" the abandonment of Roe and its numerous progeny, Vasquez, 474 U.S., at 266, 106 S.Ct., at 625, much less the greater burden of explaining the abrogation of a fundamental personal freedom. Instead, the plurality pretends that it leaves Roe standing, and refuses even to discuss the real issue underlying this case: whether the Constitution includes an unenumerated right to privacy that encompasses a woman's right to decide whether to terminate a pregnancy. To the extent that the plurality does criticize the Roe framework, these criticisms are pure ipse dixit.

This comes at a cost. The doctrine of stare decisis "permits society to presume that bedrock principles are founded in the law rather than in the proclivities of individuals, and thereby contributes to the integrity of our constitutional system of government, both in appearance and in fact." 474 U.S., at 265-266, 106 S.Ct., at 624. Today's decision involves the most politically divisive domestic legal issue of our time. By refusing to explain or to justify its proposed revolutionary revision in the law of abortion, and by refusing to abide not only by our precedents, but also by our canons for reconsidering those precedents, the plurality invites charges of cowardice and illegitimacy to our door. I cannot say that these would be undeserved.

For today, at least, the law of abortion stands undisturbed. For today, the women of this Nation still retain the liberty to control their destinies. But the signs are evident and very ominous, and a chill wind blows.

Justice STEVENS, concurring in part and dissenting in part.

Having joined Part II-C of the Court's opinion, I shall not comment on § 188.205 of the Missouri statute. With respect to the challenged portions of §§ 188.210 and 188.215, I agree with Justice BLACKMUN, ante, at 539-541, n. 1 (concurring in part and dissenting in part), that the record identifies a sufficient number of unconstitutional applications to support the Court of Appeals' judgment invalidating those provisions. The reasons why I would also affirm that court's invalidation of § 188.029, the viability testing provision, and §§ 1.205.1(1), (2) of the preamble,1 require separate explanation.

* It seems to me that in Part II-D of its opinion, the plurality strains to place a construction on § 188.0292 that enables it to conclude: "[W]e would m dify and narrow Roe and succeeding cases," ante, at 521. That statement is ill advised because there is no need to modify even slightly the holdings of prior cases in order to uphold § 188.029. For the most plausible nonliteral construction, as both Justice BLACKMUN, ante, at 542-544 (concurring in part and dissenting in part), and Justice O'CONNOR, ante, at 525-531 (concurring in part and concurring in judgment), have demonstrated, is constitutional and entirely consistent with our precedents.

I am unable to accept Justice O'CONNOR's construction of the second sentence in § 188.029, however, because I believe it is foreclosed by two controlling principles of statutory interpretation. First, it is our settled practice to accept "the interpretation of state law in which the District Court and the Court of Appeals have concurred even if an examination of the state-law issue without such guidance might have justified a different conclusion." Bishop v. Wood, 426 U.S. 341, 346, 96 S.Ct. 2074, 48 L.Ed.2d 684 (1976).3 Second, "[t]he fact that a particular application of the clear terms of a statute might be unconstitutional does not provide us with a justification for ignoring the plain meaning of the statute." Public Citizen v. Department of Justice, 491 U.S. 440, 481, 109 S.Ct. 2558, 2580, 105 L.Ed.2d 377 (1989) (KENNEDY, J., concurring in judgment).4 In this case, I agree with the Court of Appeals, 851 F.2d 1071, 1074-1075 (CA8 1988), and the District Court, 662 F.Supp. 407, 423 (WD Mo.1987), that the meaning of the second sentence of § 188.029 is too plain to be ignored. The sentence twice uses the mandatory term "shall," and contains no qualifying language. If it is implicitly limited to tests that are useful in determining viability, it adds nothing to the requirement imposed by the preceding sentence.

My interpretation of th plain language is supported by the structure of the statute as a whole, particularly the preamble, which "finds" that life "begins at conception" and further commands that state laws shall be construed to provide the maximum protection to "the unborn child at every stage of development." Mo.Rev.Stat. §§ 1.205.1(1), 1.205.2 (1986). I agree with the District Court that "[o]bviously, the purpose of this law is to protect the potential life of the fetus, rather than to safeguard maternal health." 662 F.Supp., at 420. A literal reading of the statute tends to accomplish that goal. Thus it is not "incongruous," ante, at 515, to assume that the Missouri Legislature was trying to protect the potential human life of nonviable fetuses by making the abortion decision more costly.5 On the contrary, I am satisfied that the Court of Appeals, as well as the District Court, correctly concluded that the Missouri Legislature meant exactly what it said in the second sentence of § 188.029. I am also satisfied, for the reasons stated by Justice BLACKMUN, that the testing provision is manifestly unconstitutional under Williamson v. Lee Optical Co., 348 U.S. 483, 75 S.Ct. 461, 99 L.Ed. 563 (1955), "irrespective of the Roe[v. Wade, 410 U.S. 113, 93 S.Ct. 705, 35 L.Ed.2d 147 (1973),] framework." Ante, at 544 (concurring in part and dissenting in part).

The Missouri statute defines "conception" as "the fertilization of the ovum of a female by a sperm of a male," Mo.Rev.Stat. § 188.015(3) (1986), even though standard medical texts equate "conception" with implantation in the uterus, occurring about six days after fertilization.6 Missouri's declaration therefore implies regulation not only of previability abortions, but also of common forms of contraception such as the IUD and the morning-after pill.7 Because the preamble, read in context, threatens serious encroachments upon the liberty of the pregnant woman and the health professional, I am persuaded that these plaintiffs, appellees before us, have standing to challenge its constitutionality. Accord, 851 F.2d, at 1075-1076.

To the extent that the Missouri statute interferes with contraceptive choices, I have no doubt that it is unconstitutional under the Court's holdings in Griswold v. Connecticut, 381 U.S. 479, 85 S.Ct. 1678, 14 L.Ed.2d 510 (1965); Eisenstadt v. Baird, 405 U.S. 438, 92 S.Ct. 1029, 31 L.Ed.2d 349 (1972); and Carey v. Population Services International, 431 U.S. 678, 97 S.Ct. 2010, 52 L.Ed.2d 675 (1977). The place of Griswold in the mosaic of decisions defining a woman's liberty interest was accurately stated by Justice Stewart in his concurring opinion in Roe v. Wade, 410 U.S. 113, 167-170, 93 S.Ct. 705, 733-35, 35 L.Ed.2d 147 (1973):

"[I]n Griswold v. Connecticut, 381 U.S. 479 [85 S.Ct. 1678, 14 L.Ed.2d 510], the Court held a Connecticut birth control law unconstitutional. In view of what had been so recently said in [Ferguson v.] Skrupa, [372 U.S. 726, 83 S.Ct. 1028, 10 L.Ed.2d 93 (1963),] the Court's opinion in Griswold understandably did its best to avoid reliance on the Due Process Clause of the Fourteenth Amendment as the ground for decision. Yet, the Connecticut law did not violate any provision of the Bill of Rights, nor any other specific provision of the Constitution. So it was clear to me then, and it is equally clear to me now, that the Griswold decision can be rationally understood only as a holding that the Connecticut statute substantively invaded the 'liberty' that is protected by the Due Process Clause of the Fourteenth Amendment. As so understood, Griswold stands as one in a long line of pre-Skrupa cases decided under the doctrine of substantive due process, and I now accept it as such.

* * * * *

"Several decisions of this Court make clear that freedom of personal choice in matters of marriage and family life is one of the liberties protected by the Due Process Clause of the Fourteenth Amendment. Loving v. Virginia, 388 U.S. 1, 12 [87 S.Ct. 1817, 1823, 18 L.Ed.2d 1010 (1967) ]; Griswold v. Connecticut, supra; Pierce v. Society of Sisters, [268 U.S. 510, 45 S.Ct. 571, 69 L.Ed. 1070 (1925) ]; Meyer v. Nebraska, [262 U.S. 390, 43 S.Ct. 625, 67 L.Ed. 1042 (1923) ]. See also Prince v. Massachusetts, 321 U.S. 158, 166 [64 S.Ct. 438, 442, 88 L.Ed. 645 (1944) ]; Skinner v. Oklahoma, 316 U.S. 535, 541 [62 S.Ct. 1110, 1113, 86 L.Ed. 1655 (1942) ]. As recently as last Term, in Eisenstadt v. Baird, 405 U.S. 438, 453 [92 S.Ct. 1029, 1038, 31 L.Ed.2d 349] [ (1972) ], we recognized 'the right of the individual, married or single, to be free from unwarranted governmental intrusion into matters so fundamentally affecting a person as the decision whether to bear or beget a child.' That right necessarily includes the right of a woman to decide whether or not to terminate her pregnancy. 'Certainly the interests of a woman in giving of her physical and emotional self during pregnancy and the interests that will be affected throughout her life by the birth and raising of a child are of a far greater degree of significance and personal intimacy than the right to send a child to private school protected in Pierce v. Society of Sisters, 268 U.S. 510 [45 S.Ct. 571, 69 L.Ed. 1070] (1925), or the right to teach a foreign language protected in Meyer v. Nebraska, 262 U.S. 390 [43 S.Ct. 625, 67 L.Ed. 1042] (1923).' Abele v. Markle, 351 F.Supp. 224, 227 (Conn.1972).

"Clearly, therefore, the Court today is correct in holding that the right asserted by Jane Roe is embraced within the personal liberty protected by the Due Process Clause of the Fourteenth Amendment." (Emphasis in original; footnotes omitted.)8

One might argue that the Griswold holding applies to devices "preventing conception," 381 U.S., at 480, 85 S.Ct., at 1679—that is, fertilization—but not to those preventing implantation, and therefore, that Griswold does not protect a woman's choice to use an IUD or take a morning-after pill. There is unquestionably a theological basis for such an argument,9 just as there was unquestionably a theological basis for the Connecticut statute that the Court invalidated in Griswold. Our jurisprudence, however, has consistently required a secular basis for valid legislation. See, e.g., Stone v. Graham, 449 U.S. 39, 40, 101 S.Ct. 192, 193, 66 L.Ed.2d 199 (1980) (per curiam ).10 Because I am not aware of any secular basis for differentiating between contraceptive procedures that are effective immediately before and those that are effective immediately after fertilization, I believe it inescapably follows that the preamble to the Missouri statute is invalid under Griswold and its progeny.

Indeed, I am persuaded that the absence of any secular purpose for the legislative declarations that life begins at conception and that conception occurs at fertilization makes the relevant portion of the preamble invalid under the Establishment Clause of the First Amendment to the Federal Constitution. This conclusion does not, and could not, rest on the fact that the statement happens to coincide with the tenets of certain religions, see McGowan v. Maryland, 366 U.S. 420, 442, 81 S.Ct. 1101, 1113, 6 L.Ed.2d 393 (1961); Harris v. McRae, 448 U.S. 297, 319-320, 100 S.Ct. 2671, 2689, 65 L.Ed.2d 784 (1980), or on the fact that the legislators who voted to enact it may have been motivated by religious considerations, see Washington v. Davis, 426 U.S. 229, 253, 96 S.Ct. 2040, 2054, 48 L.Ed.2d 597 (1976) (STEVENS, J., concurring). Rather, it rests on the fact that the preamble, an unequivocal endorsement of a religious tenet of some but by no means all Christian faiths,11 serves no identifiable secular purpose. That fact alone compels a conclusion that the statute violates the Establishment Clause.12 Wallace v. Jaffree, 472 U.S. 38, 56, 105 S.Ct. 2479, 2489, 86 L.Ed.2d 29 (1985).

My concern can best be explained by reference to the position on this issue that was accepted by the leaders of the Roman Catholic Church for many years. The position is summarized in a report, entitled "Catholic Teaching On Abortion," prepared by the Congressional Research Service of the Library of Congress. It states in part:

"The disagreement over the status of the unformed as against the formed fetus was crucial for Christian teaching on the soul. It was widely held that the soul was not present until the formation of the fetus 40 or 80 days after conception, for males and females respectively. Thus, abortion of the 'unformed' or 'inanimate' fetus (from anima, soul) was something less than true homicide, rather a form of anticipatory or quasi-homicide. This view received its definitive treatment in St. Thomas Aquinas and became for a time the dominant interpretation in the Latin Church.

* * * * *

"For St. Thomas, as for mediaeval Christendom generally, there is a lapse of time—approximately 40 to 80 days—after conception and before the soul's infusion. . . .

"For St. Thomas, 'seed and what is not seed is determined by sensation and movement.' What is destroyed in abortion of the unformed fetus is seed, not man. This distinction received its most careful analysis in St. Thomas. It was the general belief of Christendom, reflected, for example, in the Council of Trent (1545-1563), which restricted penalties for homicide to abortion of an animated fetus only." C. Whittier, Catholic Teaching on Abortion: Its Origin and Later Development (1981), reprinted in Brief for Americans United for Separation of Church and State as Amicus Curiae 13a, 17a (quoting In octo libros politicorum 7.12, attributed to St. Thomas Aquinas).

If the views of St. Thomas were held as widely today as they were in the Middle Ages, and if a state legislature were to enact a statute prefaced with a "finding" that female life begins 80 days after conception and male life begins 40 days after conception, I have no doubt that this Court would promptly conclude that such an endorsement of a particular religious tenet is violative of the Establishment Clause.

In my opinion the difference between that hypothetical statute and Missouri's preamble reflects nothing more than a difference in theological doctrine. The preamble to the Missouri statute endorses the theological position that there is the same secular interest in preserving the life of a fetus during the first 40 or 80 days of pregnancy as there is after viability indeed, after the time when the fetus has become a "person" with legal rights protected by the Constitution.13 To sustain that position as a matter of law, I believe Missouri has the burden of identifying the secular interests that differentiate the first 40 days of pregnancy from the period immediately before or after fertilization when, asGriswold and related cases establish, the Constitution allows the use of contraceptive procedures to prevent potential life from developing into full personhood. Focusing our attention on the first several weeks of pregnancy is especially appropriate because that is the period when the vast majority of abortions are actually performed.

As a secular matter, there is an obvious difference between the state interest in protecting the freshly fertilized egg and the state interest in protecting a 9-month-gestated, fully sentient fetus on the eve of birth. There can be no interest in protecting the newly fertilized egg from physical pain or mental anguish, because the capacity for such suffering does not yet exist; respecting a developed fetus, however, that interest is valid. In fact, if one prescinds the theological concept of ensoulment—or one acc pts St. Thomas Aquinas' view that ensoulment does not occur for at least 40 days—a State has no greater secular interest in protecting the potential life of an embryo that is still "seed" than in protecting the potential life of a sperm or an unfertilized ovum.

There have been times in history when military and economic interests would have been served by an increase in population. No one argues today, however, that Missouri can assert a societal interest in increasing its population as its secular reason for fostering potential life. Indeed, our national policy, as reflected in legislation the Court upheld last Term, is to prevent the potential life that is produced by "pregnancy and childbirth among unmarried adolescents." Bowen v. Kendrick, 487 U.S. 589, 593, 108 S.Ct. 2562, 2566, 101 L.Ed.2d 520 (1988); accord, id., at 602, 108 S.Ct., at 2570. If the secular analysis were based on a strict balancing of fiscal costs and benefits, the economic costs of unlimited childbearing would outweigh those of abortion. There is, of course, an important and unquestionably valid secular interest in "protecting a young pregnant woman from the consequences of an incorrect decision," Planned Parenthood of Central Mo. v. Danforth, 428 U.S. 52, 102, 96 S.Ct. 2831, 2856, 49 L.Ed.2d 788 (1976) (STEVENS, J., concurring in part and dissenting in part). Although that interest is served by a requirement that the woman receive medical and, in appropriate circumstances, parental, advice,14 it does not justify the state legislature's official endorsement of the theological tenet embodied in §§ 1.205.1(1), (2).

The State's suggestion that the "finding" in the preamble to its abortion statute is, in effect, an amendment to its tort, property, and criminal laws is not persuasive. The Court of Appeals concluded that the preamble "is simply an impermissible state adoption of a theory of when life begins to justify its abortion regulations." 851 F.2d, at 1076. Supporting that construction is the state constitutional prohibition against legislative enactments pertaining to more than one subject matter. Mo. Const., Art. 3, § 23. See In re Ray, 83 B.R. 670 (Bkrtcy.Ct., ED Mo.1988); Berry v. Majestic Milling Co., 284 Mo. 182, 223 S.W. 738 (1920). Moreover, none of the tort, property, or criminal law cases cited by the State was either based on or buttressed by a theological answer to the question of when life begins. Rather, the Missouri courts, as well as a number of other state courts, had already concluded that a "fetus is a 'person,' 'minor,' or 'minor child' within the meaning of their particular wrongful death statutes." O'Grady v. Brown, 654 S.W.2d 904, 910 (Mo.1983) (en banc).15

Bolstering my conclusion that the preamble violates the First Amendment is the fact that the intensely divisive character of much of the national debate over the abortion issue reflects the deeply held religious convictions of many participants in the debate.16 The Missouri Legislature may not inject its endorsement of a particular religious tradition into this debate, for "[t]he Establishment Clause does not allow public bodies to foment such disagreement." See County of Allegheny v. American Civil Liberties Union, Greater Pittsburgh ACLU, 492 U.S. 573, 651, 109 S.Ct. 3086, 3132, 106 L.Ed.2d 472 (STEVENS, J., concurring in part and dissenting in part).

In my opinion the preamble to the Missouri statute is unconstitutional for two reasons. To the extent that it has substantive impact on the freedom to use contraceptive procedures, it is inconsistent with the central holding in Griswold. To the extent that it merely makes "legislative findings without operative effect," as the State argues, Brief for Appellants 22, it violates the Establishment Clause of the First Amendment. Contrary to the theological "finding" of the Missouri Legislature, a woman's constitutionally protected liberty encompasses the right to act on her own belief that—to paraphrase St. Thomas Aquinas—until a seed has acquired the powers of sensation and movement, the life of a human being has not yet begun.17