It is a great privilege for me to be here today. I am thankful to the organizers for inviting me to participate in this panel and also to be here for my twenty-fifth law school reunion in such august company. I feel honored to be speaking here today with Justice Ginsburg. In March 1993, five months before her elevation to the Supreme Court, Ruth Bader Ginsburg delivered the Madison Lecture at New York University Law School, entitled Speaking in a Judicial Voice.1 It was a tremendously important lecture for me to read at the time. In the space of twenty-four pages, not only did Justice Ginsburg articulate her own theory of how the decision on the right to abortion should have been made in Roe v. Wade,2 but she also enunciated her distinctive account of how judges should go about adjudication and what the Court’s role is in America’s constitutional scheme.3 I found her argument compelling on all counts and was humbled by her uncanny ability to set it out so lucidly in such a small space; it had a substantial impact on my subsequent work.4

Justice Ginsburg argued in that Madison Lecture that the abortion decision should have been thought about by reference to equal protection—not the privacy doctrine taken over from Griswold v. Connecticut5 and Eisenstadt v. Baird,6and embraced by the majority in Roe.7It should have been embedded in a larger understanding of women’s equality that would take account not only of abortion but also of pregnancy, out-of-wedlock birth, gender discrimination, and related issues. Had an equal protection approach been taken, not only might the decision itself have been more secure, but some other issues, particularly relating to Medicare reimbursements and other aspects of abortion funding, might well have been better handled thereafter.8

In that same Madison Lecture, as I indicated, Justice Ginsburg developed an account of the role of the Court in relation to the other branches. In particular, she argued that when the Court plays a checking function, it should be primarily reactive, inviting a dialogue with the other branches as the Court exercises its independent role in saying what the law requires.9 And so she contrasted Justice Blackmun’s sweeping decision in Roe with the more judicious, one might say, decision in Brown v. Board of Education,10 which had restricted its focus to education, had not ruled in such a proactive way, and had left many other issues open to future litigation.11 Whereas Justice Blackmun in effect wrote his own statute on abortion regulation for different trimesters of pregnancy in what turned out to be his vain hope that this would settle the question once and for all,12 the Brown Court struck down “separate but equal” as unconstitutional without saying what should replace it.13 It was up to the Southern state legislatures to respond, if they chose, with something else—keeping the conversation going.

Ironically, by the time Justice Ginsburg joined the Court, a majority of the Justices had begun moving in the direction of her approach to the abortion question. In 1992, in Planned Parenthood of Southeastern Pennsylvania v. Casey, the majority drew a line under Roe with its strong stance in favor of stare decisis doctrine, shrouding Roe’s theoretical underpinnings in a fog of irrelevancy.14 At the same time, the Court affirmed a standard for evaluating abortion regulations that Justice O’Connor spelled out by reference to the idea of an “undue burden.”15 A legislature could regulate abortion for various purposes, including the health of the mother and vindication of its own interest in protecting potential life, but it could not do so in ways that placed undue burdens on women.16 This led to considerable debate in subsequent cases and the legal literature on what constitutes an undue burden, inevitably pushing abortion jurisprudence away from talk about privacy and toward talk about equality. This should not be surprising, because it is not possible to think about what burdens might be due or undue without appealing to the idea of equality.17

There are several dimensions to the undue-burden test, but one of particular interest here concerns risks to maternal health. It seems inevitable that the undue-burden test must require at a minimum that in circumstances where an abortion is legitimate, if a safer procedure is available, a woman cannot be required to undergo a less safe procedure. To deny this would be to impose an unnecessary burden on women that men do not have to bear. There are other dimensions of the idea of undue burden, but this one turns out to be consequential for the only two subsequent cases where the Court wrestled with what constitutes an undue burden in practice: Stenberg v. Carhart18 and Gonzales v. Carhart.19 Both dealt with the issue that is notoriously known as partial-birth abortion or late-term, partial-birth abortion.20

In Stenberg, the Court upheld the decision of a trial court that had struck down a Nebraska statute, partly because it was overbroad—it prohibited a variety of procedures other than the controversial dilation-and-extraction (D&X) procedure, and it would also punish individuals who performed inadvertent abortions.21 I will set those issues to one side. More important for us here, the Court also struck down the Nebraska statute on the ground that it contained no exception for the health of the mother.22 This bears on the undue-burden consideration that, if a safer procedure is available, no woman should be required to undergo a less safe one. Jeopardizing women’s health unnecessarily would amount to the gratuitous imposition of a burden that men do not have to bear.

In Stenberg, the lower courts had found that a substantial body of medical authority supported the attending physician’s judgment that the D&X procedure was the safest available in certain circumstances.23 That was contested by the dissenters: Justice Kennedy, joined by Chief Justice Rehnquist and Justices Scalia and Thomas.24 But reviewing courts do not listen to witnesses or hear expert testimony. As a result, they are generally expected not to second-guess the findings of fact below. Unless the record contains overwhelming evidence that those findings involved some kind of abuse of process or other egregious failure, the factual findings below are generally to be taken as given. Even when there is reason to doubt them, the appropriate remedy is to remand the case to the trial court for rehearing, not to make a different determination of the facts during the process of appellate review.

From this perspective, it is the dissenters in Stenberg, not the majority, who inappropriately relied on their own (lack of) expertise in the adjudication of contested factual questions. Rather than debate the conflicting views about the science with the dissenters, the majority would have done better to merely take the position that, as a disputed question of fact, the matter was best left as it had been settled by the trial court. If the Nebraska legislature had considered the evidence of the relative safety of different procedures, the time to establish this was at trial, and, if Dr. Carhart disputed the science on which its judgment was based, he too would have had to persuade the trial court. Perhaps the Nebraska legislature never considered the matter at all, in which case the notion that it was better situated than an appellate court to make this determination, as the dissenters contended, would be beside the point.

The Court was silent on the abortion question for seven years after Stenberg. In the interim, the Republicans took over Congress and began passing legislation to restrict abortion, including late-term, partial-birth abortion.25 President Clinton vetoed these bills.26 But the White House changed hands in 2001, and two years later Congress passed the Partial-Birth Abortion Act of 2003, which President George W. Bush signed into law.27 The new federal statute cured the issues concerning overbreadth and inadvertent abortions that had been partly responsible for the Nebraska statute’s demise in 2000,28 but it also contained a direct assault on the Court’s decision in Stenberg as it related to maternal health.29 As a result, it soon found its way into litigation in the federal courts.

Before passing the bill, Congress had held hearings on whether or not the D&X procedure is ever medically indicated as safer for the mother, and answered the question in the negative.30 Yet various lower courts struck down the 2003 Act on the basis that it violated the undue-burden standard that had been set out in Casey. How should we think about that in view of the fact that Congress had concluded after hearings that the procedure is never medically indicated for the health of the mother? More particularly, how should we think about that in light of Justice Ginsburg’s account of adjudication and the relations between the Court and legislatures, as laid out in her Madison Lecture? Part of her critique of Roe had been that the Court invited no dialogue with the Texas state legislature; instead Justice Blackmun had penned a sweeping decision that many believed had undermined the Court’s legitimacy while failing to settle the debate over the constitutionality of abortion. But here there had definitely been back-and-forth with both the Nebraska state legislature and Congress. Nebraska had passed its bill in response to the Court’s decision in Casey and Congress had passed its bill in response to Stenberg. Indeed, in Stenberg, Justice Scalia complained about “a 5-to-4 vote on a policy matter by unelected lawyers . . . [prevailing over] the judgment of 30 legislatures,” including the Nebraska state legislature.31

In reality there was no policy disagreement between legislatures and the Court in either Stenberg or Gonzales. Neither the Nebraska state legislature nor the United States Congress ever took issue with the undue-burden standard. Once the Court made it explicit in Stenberg that this included a health exception that encompassed the notion that women should not be required to undergo unnecessarily harmful procedures, Congress did not dispute that either. Indeed, the fact that Congress held hearings on the question of safety underscored its acceptance of the standard. Congress took a stand not on a question of law or policy but on a question of fact; to wit, it asserted that the procedure is never medically indicated for the health of the woman.32 The question is what the Court should have made of the stand that Congress took.

The Court’s basis in Gonzales for deferring to the congressional finding was not entirely clear. As Justice Ginsburg noted in dissent, the Congressional Record was replete with testimony contradicting the finding in the Act that the D&X procedure is never safer than the alternative—not to mention evidence to that effect from other professional sources such as the American College of Obstetricians and Gynecologists.33 The Act incorrectly asserts that no medical schools taught the D&X procedure when in fact many of the leading ones did.34 Moreover, as if to underscore the cursory and result-driven nature of the hearings, none of the physicians who testified before Congress had in fact performed the procedure.35

The federal courts have a long history of refusing to defer to manifestly implausible legislative findings, and the lower courts followed that precedent in this case. As Justice Thomas put it while serving on the D.C. Circuit before his elevation to the Supreme Court, “If a legislature could make a statute constitutional simply by ‘finding’ that black is white or [that] freedom [is] slavery, judicial review would be an elaborate farce.”36 If the Southern state legislatures had responded to Brown in the late 1950s by staging bogus hearings to conclude that their separate school districts were not in fact unequal, the federal courts would not have capitulated. Nor should they have done so. Yet, in Gonzales, the new majority on the Supreme Court deferred to Congress on the safety question, reversed the findings below, and held that in the face of professional disagreement over the medical merits of the procedure, the Court could not foreclose the legislative power of Congress.37

But no foreclosure of the legislative power was at stake. The point of sending the matter back to the legislative branches in Stenberg was to ask them to fashion a policy that comes within the constitutional constraint, not to subvert that constraint with legislation based on implausible factual assertions. The congressional findings at issue in Gonzales were readily impeachable at trial, as we have seen. That the trial court was persuaded by different testimony that contradicted the congressional findings is a signal to the legislature that such perfunctory factual hearings will likely be accorded the deference that they merit.

The trial court might, of course, have found differently had the government presented a more convincing case. In that event the Partial-Birth Abortion Ban Act of 2003 would have been upheld at trial. As things turned out, it was not, and the Supreme Court had no more business reversing that finding than it would have had reversing the trial court’s rejection of the Nebraska partial-birth abortion ban seven years earlier. The dissenters in Stenberg were part of the majority in Gonzales, but this was because the composition of the Court had changed, not the merits of the issue. It remained as true in 2007 as it had in 2000 that those who see and hear witnesses and testimony are best placed to judge their credibility, and that reviewing courts should therefore second-guess trial courts on questions of fact only when it is manifest that they have abused their factfinding authority. That proposition was never alleged by any litigant below or in any opinion in Gonzales.

What happened in Gonzales was that the trial court was persuaded by testimony that the D&X procedure is sometimes medically indicated for the woman. Consequently, it struck down the statute in line with Casey’s undue-burden test as further elaborated in Stenberg. All that was at issue was a question of fact. Justice Ginsburg’s dissent in Gonzales was consistent with her views on adjudication and on the Court’s ongoing dialogue with legislatures in America’s constitutional scheme. Dialogue must require more of Congress than inventing its own reality by legislative fiat. Reflecting on why this is so enables us to make explicit one aspect of the rules of grammar that should govern that dialogue going forward.

* * *

Justice Ginsburg offered the following response:

* * *

Ian Shapiro, who received a Ph.D. and J.D. from Yale, is the Sterling Professor of Political Science at Yale University.

Preferred citation: Ian Shapiro, Still Speaking in a Judicial Voice: Ruth Bader Ginsburg Two Decades Later, 122 Yale L.J. Online 257 (2013), http://yalelawjournal.org/forum/still-speaking-in-a-judicial-voice-ruth-bader-ginsburg-two-decades-later.html.