When Samuel Alito replaced Sandra Day O’Connor on the Supreme Court in 2006, observers predicted that Anthony Kennedy would quickly become the key figure in the nation’s jurisprudence. And recent terms have confirmed those predictions: Across a wide range of controversial constitutional issues, Kennedy now holds the balance of power on an otherwise deadlocked court”particularly on abortion.

Writing for a 5-4 majority in last June’s Gonzales v. Carhart , Kennedy upheld the constitutionality of a federal statute banning the horrific practice known to abortion providers and the mainstream media as “intact dilation and extraction” and to everyone else, including Congress, as “partial-birth abortion.” The result was cause for at least modest celebration, particularly when contrasted with the court’s 2000 Stenberg decision, in which a different 5-4 majority struck down Nebraska’s ban on partial-birth abortions, over Kennedy’s dissent.

The decision in Carhart was achieved because Kennedy joined forces with Justices Scalia and Thomas (both of whom are open opponents of Roe v. Wade ) and Chief Justice Roberts and Justice Alito (who haven’t committed themselves on Roe but are presumably willing to limit abortion rights). For thirteen frustrating years after Ruth Bader Ginsburg replaced Byron White in 1993, it was O’Connor who played the role of swing justice on abortion, and Stenberg was the disastrous result. Now that Kennedy casts the deciding vote, we have Carhart , and partial-birth abortion can no longer be inflicted on living fetuses. Pro-life prospects are better than they were.

But how much better? Kennedy’s elevation makes it imperative to assess his views on abortion to determine which, if any, pro-life legislative initiatives have a realistic chance of winning his vote. On this score, the views of pro-life observers range from bitterly pessimistic to guardedly optimistic. But, if we take Kennedy at his word, Carhart could be the beginning of a string of incremental but significant pro-life victories in the Supreme Court. The opening for pro-life legislation may well be greater than even the optimists think.

The pessimists have one very big point in their favor: It was Kennedy, after all, who cast the deciding vote in Planned Parenthood v. Casey (1992) against overruling Roe . It would be miraculous if he were to reverse course at this late date.

According to Justice Blackmun’s posthumously released papers, Kennedy initially provided the crucial fifth vote in Casey to overrule Roe . Tragically, within weeks he changed his mind”and went on to coauthor the infamous “joint opinion” in Casey , which ­reaffirmed the “essential holdings” of Roe while allowing somewhat greater room for state regulation of ­abortion.

To make matters worse, Casey had the temerity to assert that the pro-life movement should stop asking the Supreme Court to change its mind and accept Roe and Casey as what Senator Arlen Specter has since dubbed “super-duper precedent.” Kennedy has never shown any willingness to reconsider this aspect of Casey or its central holdings. Indeed, he presented his opinion in Carhart as a faithful application of the principles of Casey . Living under Casey has not been easy for the pro-life movement to bear, and Kennedy’s defection in that case has understandably caused him to be viewed with suspicion and resentment in pro-life circles.

Nevertheless, there is also an often overlooked pro-life side to Kennedy’s message. Judging from his opinions, Kennedy strongly believes both that a pregnant woman should not be legally required to bring her child to birth and that society should be free to express its opposition to abortion through legislation designed to persuade her to choose life. Casey affirmed that the state has a “profound interest” in fetal life “throughout pregnancy,” and “measures designed to advance this interest will not be invalidated as long as their purpose is to persuade the woman to choose childbirth over abortion.” In Casey , this involved informed-consent measures requiring that women be provided with information about the nature and consequences of abortion. In Carhart , it involved banning an especially perverse method, to educate women about the grue­some consequences of second-trimester abortions”and hopefully avert some abortions as a result.

Kennedy also thinks the government interest in fetal life encompasses an interest in humane and respectful treatment of fetuses quite apart from whether their lives are saved. By affirming the “intrinsic value” of fetal life, including the lives of fetuses doomed to die via abortion, regulation of abortion methods teaches that abortion is morally wrong even if Casey requires that it remain legal.

Indeed, as Kennedy’s impassioned dissent in Stenberg attests, he thinks pro-life legislation is not just permissible but desirable: “The political processes of the State are not to be foreclosed from enacting laws to promote the life of the unborn and to ensure respect for all human life and its potential. The State’s constitutional authority is a vital means for citizens to address these grave and serious issues, as they must if we are to progress in knowledge and understanding and in the attainment of some degree of consensus.”

It would be an enormous mistake for the pro-life movement to ignore the opportunities that this statement presents. There are many initiatives worth considering. Legislation mandating that women considering abortion be offered a publicly funded ultrasound is likely to be constitutional and could save many lives. Legislation banning gender-selective abortion appears to be unconstitutional under current law, but Kennedy might be willing to entertain the idea that some limits can be placed on the reasons for an elective abortion. Legislation requiring that fetuses be given an anesthetic prior to second-trimester abortions is another intriguing possibility; while it is uncertain when fetuses become capable of experiencing pain, Carhart is emphatic that legislatures have “wide discretion to pass legislation in areas where there is medical and scientific uncertainty.”

Perhaps the most important opportunity that Carhart offers, however, is legislation that bans another brutally violent second-trimester abortion method: the widely used standard dilation and extraction, or D&E. Carhart observes that this method of performing abortions “is a procedure itself laden with the power to devalue human life.” Judging from that and other hints in his opinion, Kennedy may be receptive to the constitutionality of a ban on both types of D&E abortions (standard as well as partial-birth). In addition to sending a strong signal about the evils of second-trimester abortion, this legislation would be an ideal vehicle for persuading the Court to adopt a suitably narrow interpretation of the maternal life-or-health exception that most restrictions on abortion are required by Casey to contain.

In other words, the federal ban on partial-birth D&E is good as far as it goes, but it doesn’t go far enough. The standard D&E abortion is, if anything, more barbaric. This procedure, which is performed roughly a hundred thousand times a year in America, should be called “dismemberment D&E,” because the abortionist literally dismembers the fetus using the woman’s cervix as a fulcrum and the forceps as a lever. Even if those authorities who maintain that second-trimester fetuses cannot experience pain are correct, dismemberment D&Es, like partial-birth D&Es, involve extreme violence that degrades and dehumanizes the fetus by mutilating its body beyond recognition.

Of course, as with the ban on partial-birth D&Es, a ban on all D&E abortions would not necessarily save fetal lives. To be constitutional under Carhart , it would be necessary for the legislation to permit second-trimester abortions by at least one “commonly used and generally accepted method.” The only practicable alternative is induced-labor abortions, which were the most popular method before the invention of D&E. If dismemberment D&E abortions were banned, many women would simply switch to induced labor instead (while others might substitute first-trimester vacuum abortions).

Still, by publicly expressing society’s judgment that fetuses are innocent human offspring who must not be killed in these brutal ways, a ban on all D&E abortions might lead some women to reject abortion. Moreover, induced-labor abortions, while gravely wrong, are less pernicious than D&E abortions. Unlike the 1970s version of induced labor, which chemically burned fetuses, contemporary induced-labor abortions use drugs that trigger labor without poisoning or defiling the fetus. The previable fetus dies of oxygen deprivation during or shortly after induced labor, but, rather than dying a violent, disfiguring death, it comes into the world “stillborn,” with its unique individual human form intact.

A ban on all D&E abortions would declare, then, that D&E abortions undermine respect for fetal life more than induced-labor abortions do and that aborted fetuses are at least entitled to a dignified and nonviolent death. Carhart suggests that such a legislative judgment would be legitimate and reasonable. What Kennedy said of partial-birth abortion in Stenberg applies with full force to all D&E abortions: States “have an interest in forbidding medical procedures which, in the State’s reasonable determination, might cause the medical profession or society as a whole to become insensitive, even disdainful, to life, including life in the human fetus.”

The abortion lobby will contend that a ban on D&E abortions violates Casey ‘s “undue burden” test for previability-abortion regulations. Induced labor requires hospitalization and is thus more expensive and available at fewer facilities. It is also reportedly emotionally more distressing for many women. Higher cost, greater suffering, and reduced availability, it will be said, add up to a substantial obstacle to obtaining a second-trimester abortion.

This argument unquestionably will carry the day for the four justices who support abortion rights. To keep Kennedy from joining them, it will be critical to convince him that it rests on a mistaken premise. Under Casey , there is a constitutional right to choose abortion before viability, but there is no parallel constitutional right to choose a particular abortion method .

For constitutional purposes, the correct comparison is not between forbidden and permitted abortion methods, but between the permitted abortion method and childbirth”at which point it becomes clear why the proposed legislation should be upheld: Premature induction of labor appears to be at least as safe as childbirth and is considerably less expensive. For women who want an abortion, the distress of induced-labor abortion is less than the distress of childbirth; and even if induced-labor abortions are not currently available in some locations, that would change as abortion doctors adapted to the new rules.

The cost differential between D&E and induced-labor abortion will presumably reduce the incidence of abortion somewhat, but that is permissible because the primary purpose or effect of the proposed ban is not to make abortion prohibitively expensive. Even if induced-labor abortions were cheaper than D&E abortions, the barbaric violence of D&E would still justify singling it out for prohibition. The undue-­burden challenge should therefore fail: As Carhart ­ reiterated, “the fact that a law which serves a valid ­purpose, one not designed to strike at the right itself, has the incidental effect of making it more difficult or more expensive to procure an abortion cannot be enough to invalidate it.”

There remains, however, the question of the mother’s health. Casey reaffirmed that, even after viability, states must allow abortions that are “necessary, in appropriate medical judgment, for the preservation of the life or health of the mother.” As a corollary, Casey held that previability abortion ­regulations must exempt women who would be subjected to “significant health risks” if they complied with the law. For some women, induced labor would present a grave and imminent danger to life or health. Consequently, a ban on all D&E abortions would undeniably require a life-or-health exception.

Once the need for a statutory life-or-health exception is admitted, an important and long unresolved legal question comes to the fore: How broadly must that exception be defined to comply with the Court’s decisions? In Stenberg , the pro-abortion-rights majority shamelessly asserted that whatever abortion method is medically “appropriate” for the woman is also “necessary” to preserve her life or health. By the same token, Stenberg indicated that any maternal health risk that would lead an abortion provider to choose one method over another qualifies as a “significant health risk.” On this view, a ban on D&E abortions is unconstitutional, because the estimated health risks of induced-labor abortions, while quite low, appear measurably higher than the risks of D&Es.

Kennedy’s opinion in Carhart rejects these extreme views (as did his dissent in Stenberg ). Yet Carhart fails to spell out a definition of the life-or-health exception. (The Court was able to duck that question because it deferred to Congress’ finding that partial-birth D&E abortion is never the safest method.) What legislatures need is a carefully crafted statutory exception, narrow enough to preserve the ban’s regulatory bite but not so narrow that Kennedy balks. Here’s my proposed language: “An abortion (or an abortion method) is necessary to preserve the woman’s life or health if there is a substantial likelihood that continuing the pregnancy (or employing any alternative abortion method) would result in her death or in grave and irreparable damage to her physical or mental health.”

In challenging this definition, pro-abortion-rights groups will be acutely aware that there are now only four votes for Stenberg ‘s categorical rule that women’s health always trumps fetal life or dignity. They will therefore need to construct a less extreme argument calculated to pick up Kennedy’s vote. After viability, they might concede, an abortion should perhaps not be deemed necessary for health reasons unless childbirth would involve a substantial risk of death or serious health injury. But before viability, the balance should be struck differently, because the government’s interest in fetal life is less weighty.

The previability test, the argument goes, should be whether a restriction on abortion methods subjects women to “significant health risks” (in Casey ‘s phrase). That test, they will urge, should be interpreted to include all risks that would be a material factor in the decision of a reasonable woman and her abortion provider. Induced-labor abortions reportedly have significantly greater rates of mortality and morbidity than D&E abortions, and that is a material difference.

Given his commitment to Casey , there’s a good chance Kennedy will not want to jettison the formula of “significant health risks.” The key question is what meaning he will decide to assign to that vague and open-ended phrase. Kennedy’s Stenberg dissent gives us a window on his thinking. He argued there that women are constitutionally entitled to a safe abortion, not the safest possible abortion, and that “unsubstantiated and generalized health differences which are, at best, marginal do not amount to a substantial obstacle to the abortion right.” That language aptly describes the small statistical risks that supposedly make induced-labor abortions marginally less safe than D&E abortions. And unless one is prepared to say that childbirth is not safe, the conclusion is inescapable that induced-labor abortion (which has a slightly lower estimated mortality risk than childbirth) is safe as well.

There is good reason to hope, therefore, that Kennedy will reject the materiality approach and ultimately conclude that the small mortality and morbidity risks of induced-labor abortions do not subject women to “significant health risks.” To that end, it will be vital to develop a factual record attesting to the low rates of death and complications associated with induced-labor abortions. But it will also be vital to convince Kennedy that an abortion method should be deemed safe unless it entails a substantial chance of death or grave and irreparable damage to health, as opposed to a small (even if quantifiable) risk of such harm.

TThe key here is the holding in Casey that the government has important interests in fetal life and respect for human life throughout pregnancy. Even before viability, a “significant health risk” ought to mean a maternal health risk so dangerous that avoiding it properly overrides these important government interests. The pro-abortion “material risk” interpretation refuses to balance: The woman’s interest in her own health always trumps the government interests in fetal life and fetal dignity. Yet, in many other contexts, from mandatory vaccinations and the military draft to restrictions on the sale of human organs for transplants, individuals may be compelled by law to accept health risks they would strongly prefer to avoid. Abortion should be no different, and if Kennedy means what he said in ­ Stenberg, he should agree.

Legislation banning all D&E abortions would have several virtues. It would eliminate one of the most corruptingly violent abortion methods; it would powerfully affirm the human dignity of fetuses; and it would probably save some fetal lives by educating pregnant women and the general public about the nature and consequences of second-trimester abortions. If the Court upheld the legislation’s narrow life-and-health exception, that ruling would authorize the inclusion of similarly narrow exceptions in many other types of previability abortion regulations.

A favorable decision on both the D&E ban and the life-and-health exception would constitute a major victory by any standard”as well as an important milestone in the gradual process of undermining the already weakened doctrinal foundation on which Roe v. Wade rests.

Stephen G. Gilles is professor of law at Quinnipiac ­University School of Law.