---

3. THE DECISIONS

---

4. A BADLY FLAWED ELECTION

Ronald Dworkin

The 2000 election has finally ended, but in the worst possible way—not with a national affirmation of democratic principle but by the fiat of the five conservative Supreme Court justices, Chief Justice Rehnquist and Justices Kennedy, O'Connor, Scalia, and Thomas, over the fierce objection of the four more liberal justices, Justices Breyer, Ginsburg, Souter, and Stevens. The conservatives stopped the democratic process in its tracks, with thousands of votes yet uncounted, first by ordering an unjustified stay of the statewide recount of the Florida vote that was already in progress and then by declaring, in one of the least persuasive Supreme Court opinions that I have ever read, that there was no time left for the recount to continue. It is far from certain that Gore would have been elected if the recounts had been completed; some statisticians believe that Bush would have picked up more additional votes than Gore. But the Court did not allow that process to continue, and its decision ensured both a Bush victory and a continuing cloud of suspicion over that victory.

Though it took six opinions for all the justices to state their views, the argument of the five conservatives who voted to end the election was quite simple. The Florida Supreme Court had ordered a recount of undervotes across the state, but instead of adopting detailed rules for determining whether a ballot that the counting machine had declared to have no vote for president was actually a vote for one candidate—rules that might have specified, for example, that if not a single corner of the chad of a punchcard ballot had been detached, the ballot could not count as a vote—it had directed only that counters count a vote if they found, considering the ballot as a whole, a "clear intention" of the voter. The five conservatives noted that this more abstract standard had been applied differently by counters in different counties and might be applied differently by different

---

counters within a single county, and they therefore held that the use of the standard denied voters the equal protection of the law required under the U.S. Constitution's Fourteenth Amendment.

The natural remedy, following such a ruling, would be to remand the case to the Florida court to permit it to substitute a more concrete uniform counting standard. Breyer, in his dissenting opinion, suggested that course. "[The] case should be sent back for recounting all undercounted votes," he said, "in accordance with a single uniform standard." But the conservatives declared that since the Florida legislature intended to take advantage of the "safe harbor" provision of federal law, which provides that election results certified by states to Congress by December 12 are immune from congressional reexamination, any further recount the Florida court ordered would have to be completed by that date—which ended two hours after the Supreme Court handed down its judgment. The conservatives had remanded the case to the Florida court for "proceedings consistent with" their opinion and then told them that no proceedings could possibly be consistent with their opinion. The election was over, and the conservative candidate had won.

The 5–4 decision would hardly have been surprising, or even disturbing, if the constitutional issues were ones about which conservatives and liberals disagree as a matter of constitutional principle—about the proper balance of authority between the federal and state governments, for example, or the criminal process, or race, or the character and extent of individual rights, such as abortion rights or rights of homosexuals, against state and national authorities. But there were no such constitutional issues in this case: the conservatives' decision to reverse a state supreme court's rulings on matters of state law did not reflect any established conservative position on any general constitutional question. On the contrary, conservatives have been at least as zealous as liberals in protecting the right of state courts to interpret state legislation without second-guessing by federal courts, and on the whole they have been less ready than liberals to appeal to the Fourteenth Amendment to reverse state decisions.

It is therefore difficult to find a respectable explanation of why all and only the conservatives voted to end the election in this way, and the troubling question is being asked among scholars and commentators whether the Court's decision would have been different if it was Bush, not Gore, who needed the recount to win—whether, that is, the decision reflected not ideological division, which is inevitable, but professional self-interest. The five conservatives have made this Supreme Court the most activist Court in history. They aim to transform constitutional law not, as the Warren Court did, to strengthen civil liberties and individual rights but rather to expand the power of states against Congress, shrink the rights of accused criminals, and enlarge their own powers of judicial intervention.^[1]

---

For three of them—Rehnquist, Scalia, and Thomas—the agenda presumably includes finally abolishing the abortion rights that were first established in Roe v. Wade over a quarter of a century ago, a decision they have never ceased insisting should be overruled. The prospects of future success for the conservatives' radical program crucially depend on the Court appointments that the new president will almost certainly make. Those appointments will determine whether the conservatives' activism will flourish (even adding, perhaps, the two new votes that would be needed to overrule abortion rights so long as O'Connor and Kennedy refuse to take that step) or whether it will be checked or reversed. Bush long ago signaled, in naming Scalia his favorite justice, his intention that it flourish.

We should try to resist this unattractive explanation of why the five conservative justices stopped the recount process and declared Bush the winner. It is, after all, inherently implausible that any—let alone all—of them would stain the Court's reputation for such a sordid reason, and respect for the Court requires that we search for a different and more creditable explanation of their action. Unfortunately, however, the legal case they offered for crucial aspects of their decisions was exceptionally weak. Their first major ruling, on Saturday, December 9 (soon after the recounting began), was to halt the recount even before they heard argument in Bush's appeal of the Florida Supreme Court decision ordering that recount. The ruling was in itself lethal for Gore. Even if the Court had ultimately rejected Bush's appeal and allowed the recount to resume, it could not possibly have been completed by December 12, the date that the conservatives later declared the final deadline.

Scalia argued that this serious injury to Gore was necessary to prevent irreparable harm to Bush: he said that Bush would be harmed if the recounts continued because if the Court later decided that the recount was illegal, the public's knowledge of the results would cast a "cloud" over "the legitimacy of his election." That bizarre claim assumes not only that Bush would have lost in the recount but also that the public is not to be trusted. Public knowledge that Gore would have won, if the recounts had continued and been accepted, would produce doubt about a Bush election only if the public disagreed with the Court's judgment that the recount was illegal; and it is constitutionally improper for the Court to keep truthful information from the public just because the information might lead it to conclude that the election was a mistake or that the Court was wrong.^[2]

The conservatives' second major decision was that the Florida court's "clear intention of the voter" standard for manual recounts violated the Equal Protection Clause because different counties and counters would interpret that standard differently. Two of the more liberal justices—Breyer and Souter—agreed,^[3] but the other liberal justices, Ginsburg and Stevens, rejected the argument, and they had the better case. The Equal Protection

---

Clause forbids voting procedures or arrangements that put particular people or groups at an electoral disadvantage. The Court has struck down poll taxes that discriminated against the poor, for example, and, citing a "oneperson, one-vote" electoral standard, has prohibited electoral districts of very different size because these give each voter in larger districts less impact on the overall election result than voters in smaller districts have. But a general standard for counting undervotes that may be applied differently in different districts puts no class of voters, in advance, at either an advantage or a disadvantage. If a voter's county uses a more permissive test to determine "clear intent," then he risks having his ballot counted when he did not intend to vote; if it uses a strict standard, he risks having his ballot ignored when he did intend to vote. One cannot say, in advance, that either a permissive or a strict test is more accurate and therefore cannot say that a system that combines both within a single state puts any identifiable group at an automatic disadvantage.^[4]

As Gore's counsel, David Boies, pointed out in oral argument, Florida's use of different voting machinery in different counties is much more arguably a violation of equal protection, because some types of machine are well known to be much less accurate than others. Punch-card ballot readers, which are used in counties with a high minority population such as Miami-Dade, ignore more than three times as many ballots as optical-scan ballot readers do and therefore give voters in those counties systematically less chance of having their votes counted.

The Court's equal protection decision is surprising in another way. The one-person, one-vote principle applies not just to presidential elections but also to elections for every federal and state office, major or minor, across the country. I do not know how many states use nothing more concrete than a "clear intent of the voter" standard for manual recounts, but several do, and the Supreme Court has now declared that they have all been acting, no doubt for many decades, unconstitutionally. This ruling alone may require substantial changes in the nation's electoral laws, and the Supreme Court may well regret having made it.

The conservatives' equal protection claim is defensible, however, and, as I said, two of the more liberal justices also accepted it. But the conservatives' third and by far most important major decision is not defensible. The most natural remedy for the supposed equal protection violation, as all the dissenters insisted, would be to remand the case to the Florida court so that it could establish uniform recount standards and attempt to complete a recount by December 18, when the Electoral College votes. But the conservatives held that since the Constitution gives the Florida state legislature authority over its own election law and since that legislature would wish to take advantage of the federal safe harbor law that guarantees a state certification of presidential electors immunity from congressional challenge

---

if the certification is made by December 12, any recounts beyond that date, even those necessary to ensure that all valid votes were counted, would automatically be unconstitutional.

But the safe harbor provision is not mandatory; it does not provide that a state loses its electoral votes if these are not submitted by December 12 but only that its votes, if submitted after that date, might conceivably be challenged in Congress, if reason can be found to challenge them. Certainly the Florida legislature would wish to meet the December 12 deadline if it fairly could, and its legislation should be interpreted, as the Florida Supreme Court said that it did interpret it, with that aim in mind.

But it goes far beyond that safe assumption to declare, as the five U.S. Supreme Court conservatives did, that the Florida legislature meant to insist that the optional deadline be met at all costs, even if it was necessary to ignore the principles of accuracy and fair treatment that underlie the rest of the election code. That would be a bizarre interpretation of any state's election law—what legislature would wish to be understood as purchasing an immunity it would almost certainly never need at the cost of sacrificing its basic commitments of justice?—and there is no evidence that the Florida legislature has ever made that choice.^[5] Even if the conservative justices thought this bizarre interpretation plausible, moreover, it would still be wrong for them to impose that interpretation on the Florida Supreme Court, which, according to the most basic principles of constitutional law, has final authority in interpreting its own state's law so long as its interpretation is not absurd. At most, the conservatives should have asked the Florida court to decide for itself whether Florida law, properly understood, declares that the safe harbor must be gained no matter what unfairness to Florida voters is necessary to gain it.

Even the two dissenters who had agreed with the majority that the Florida court's recount scheme violated the Equal Protection Clause thought it absurd to insist on the December 12 deadline, and all of the dissenters feared the impact on the Court itself of so weak an argument for so politically divisive a decision. Justice Stevens said the decision "can only lend confidence to the most cynical appraisal of the work of judges throughout the land." "We do risk a self-inflicted wound," Breyer added, "a wound that may harm not just the court, but the nation," and he also noted, pointedly, that the time pressure the conservatives cited was "in significant part, a problem of the [Supreme Court's] own making."^[6] We must try, as I said, not to compound the injury to the Court with reckless accusations against any of its members. But those of us who have been arguing for many years that the Supreme Court makes America a nation of principle have a special reason for sorrow.

The deeply troubling Supreme Court decision in Bush v. Gore makes even plainer the urgency of radically changing how we elect our presidents.

---

Our present system is an eighteenth-century antique: it presupposes a starkly elitist conception of government that was popular then but which no politician would dare endorse today. The Constitution's authors did not trust the people to elect the president directly; they expected the members of the Electoral College to be distinguished and independent citizens who would make up their own minds, after collective deliberation, about who the president and vice president should be. It was not as important as it later became how those electors were chosen, because, in principle, the selecting body would have no control over or even confidence in the opinions of the independent electors they chose. There was nevertheless some initial disagreement at the Constitutional Convention about the selection of electors: some delegates wanted them elected by popular vote, some by Congress, and some by the state legislatures.

The Convention finally decided, by way of compromise, not to establish any electoral method itself but to delegate the choice of methods to the state legislatures. Each state was assigned a number of electors equal to the total number of that state's representatives and senators in Congress (that formula was a concession to smaller states that had fewer representatives but the same number of senators as much larger states), and the state legislatures were directed to decide how their state's electors would be chosen. A majority in the legislature might select the electors themselves or provide for a popular vote within the state to select them (which might be by a statewide vote for all the electors, or by districts, or by proportional representation). Or, presumably, it might direct that electors be selected by lot. Once all the electors were selected in whatever ways the various state legislatures chose, they would meet in their states, deliberate, and vote for a president; if no presidential candidate received a majority of their votes, then the House of Representatives would choose a president, with each state delegation having one vote, so that the smallest state had as much influence as the largest.

America has long since rejected the intellectual premises of this baroque system. Now we embrace the very different principle that the point of elections—particularly the election of a national president, the one office we all elect together—is to determine and reflect the people's will. Electors are no longer expected to exercise their own judgment: it is candidates, not electors, whose names are on the ballot and it would be a scandal if the electors chose someone other than the candidate to whom most of them were pledged. A partisan majority in a state legislature still has the constitutional power, under Article II, to cancel presidential elections in its state and choose the state's electors by themselves. But if any legislature tried to exercise that power, its action would undoubtedly provoke a constitutional amendment ending that power.

We have been lucky not to have been seriously damaged by the Electoral

---

College system long before this election made its anachronism intolerable. It is dangerous to retain a constitutional structure when its principled base has been so thoroughly repudiated, because the structure then becomes a legal loose cannon. It generates pointless complexities and obstacles, and it is vulnerable to partisan manipulation and bizarre interpretation that cannot be checked by appealing to the structure's purpose, since it now has none. The legal battles in Florida and in the Supreme Court were dominated by a series of deadlines—the Court elected Bush by insisting on the importance of December 12—that are significant only because the eighteenth-century arrangement decreed a stately series of certifications, meetings, and pronouncements that are now only charades. The Republican strategy in Florida of delaying recounts through any means possible, including not only legal challenges but also noisy demonstrations outside counting rooms, was made possible only by those pointless deadlines. It makes no sense to demand that a breathtakingly close election be decided by any magic date in December in order that a new president be chosen by January 20.

The original decision to leave the manner of presidential elections to state legislatures corrupts elections in a different way. The one-person, onevote principle would suggest, as I said, that we elect presidents through uniform voting methods, with at least roughly equal accuracy, supervised by a national election commission under principles established by Congress. The eighteenth-century compromise guarantees, to the contrary, that different methods of recording votes, which vary dramatically in their accuracy, will be used not only in different states but in different counties within states as well. It also guarantees that inevitable uncertainties and ambiguities in election law will have to be faced anew in each close election, because even if Florida's law is clarified now, the next set of contests will arise in an entirely different state with an entirely different structure of law and ambiguity.

The present system means, moreover, that politics will play an inevitably ugly role in close elections. It is surely unacceptable that the Florida state legislature, dominated by Republicans, should have the power themselves to elect a set of electors pledged to the Republican candidate whenever they deem this to be necessary because the result of the election is uncertain. Many of the most consequential decisions in Florida were made by political officials whose future might depend either on who won the presidential election or on whether powerful Florida politicians, including Bush's brother, who is Florida's governor, would approve what they had done. Katherine Harris, the Florida secretary of state whose several erroneous rulings contributed enormously to the delays that prevented a fair recount, had been co-chair of Bush's campaign in Florida, and the New York Times reported that the Democratic mayor of Miami had been subject to a

---

great deal of local pressure just before the Miami-Dade County Canvassing Board reversed itself and decided to halt manual recounts.^[7] It would be a mistake to assume wrongdoing or improper motives in any such case, but it would certainly be better to vest critical decisions in nonpartisan federal election officials who would be much less likely to attract suspicion.

We now have the best chance ever to junk the anachronistic and dangerous eighteenth-century system. The public should demand that Congress begin a process of constitutional amendment that would eliminate that system, root and branch, and substitute for it the direct election of the president and vice president by a plurality of the national popular vote. The amendment should direct Congress to establish uniform election procedures and machinery across the nation, and that body might then design and finance computers with screens that clearly display a voter's tentative choices and ask the voter to confirm his votes before they are recorded.^[8] (It might be possible to allow people with computers, including absentees, to vote through them at home, although special digital identification and security precautions would have to be developed and care taken to avoid unfairness to voters with no access to computers.)

Congress should further require that voting booths be open for the same twenty-four-hour period across the different time zones of the country, so that voting stops simultaneously everywhere, and that television networks do not report results in one time zone while voting continues in another; and it should establish a national election commission with general supervisory power over national elections. Challenges and disputes would no doubt still arise, but these could be adjudicated by officials of such an agency, who would be appointed to provide nonpartisanship, subject to review by federal judges with life tenure rather than by state political officials and elected state judges.

The nation would benefit in other ways from the change. It does not matter, under the Electoral College system, who won the national popular vote, but that fact is nevertheless widely reported and widely thought significant; a president who has won in the Electoral College but lost the popular vote, even by a relatively slim margin, is thought by many people less legitimate for that reason. Making the popular vote decisive would end the possibility of such a situation. Would we lose anything by the shift? It is said to be a benefit of the Electoral College system that it forces candidates to campaign across the nation rather than only in a few highly populated regions with huge media markets and the largest number of potential votes. But in fact the system does not produce genuinely national campaigns. Candidates wholly ignore states that they are very likely either to win or to lose—few presidential campaign ads appeared in the New York media market in this election, for example—and devote most of their time and money to those relatively few states in which the race appears to be

---

close. If the national popular vote were decisive, they would not campaign just in the major population centers—there are too many votes elsewhere— but wherever they thought they could persuade a substantial number of as yet undecided voters.

It is also said that the Electoral College is necessary to protect regional interest groups that are powerful within certain states, and thus important to those states' electoral votes, but not nationally. But interest groups are now much more dispersed across the nation than they once were: many states that were formerly dominated by agricultural interests, for example, now have a more mixed economy, and farmers might be better protected by voting rules that made their absolute number important even if they were geographically dispersed.

It might appear that the Electoral College system reduces the number of postelection challenges and contests in close presidential races because it gives candidates no incentive to seek to correct mistakes in a state that the other candidate won so heavily that he would take its electoral votes anyway. Under a popular vote system, however, a candidate who lost the popular vote by a very small margin might canvas the entire country looking for a series of challenges that could yield only a few votes in each case but might change the overall result collectively.

But there is no reason to think, in advance, that a change from the Electoral College to a popular vote standard would produce more postelection challenges or contests. If a national election is close, then the election in states whose electoral votes are crucial is also likely to be close, and many fewer vote changes are needed to make a difference in the state than in the nation. Gore needed only to add a few hundred votes in Florida through challenges, but, even in this exceptionally close popular vote contest, Bush would have had to add more than 300,000 votes to his total to win, and there is no indication of irregularities elsewhere in the nation that affected, even cumulatively, that many votes.

The moment seems propitious, as I have said, for pressing for a constitutional amendment: politicians in either party would have great difficulty claiming that the system we have has worked well or supplying any principled rationale for it. But we must recognize that it is extremely difficult and normally takes many years to amend the Constitution. Short of a new constitutional convention, an amendment requires a two-thirds vote of both houses of Congress and then approval by the legislatures of three quarters of the states, and the pressure for an amendment may weaken before that long process has been completed. In any case, no amendment can succeed without the consent of many of the smaller states whose citizens benefit unfairly, in the ways I have described, from the Electoral College system that the amendment would end.

It is therefore important to consider how much of the gain that an

---

amendment would bring could be achieved at once without one or while one is pending. One gain I described—a twenty-four-hour election day ending simultaneously across the country—could be adopted by Congress now, because the Constitution assigns it authority to fix the time of presidential elections. More could be achieved through a model uniform election code that Congress might endorse and propose to the states, agreeing to finance elections for national office, including providing accurate electronic voting machinery, for those states that adopted that code. The model code would no doubt be adopted in somewhat different form in different states, but Congress could identify core provisions that guaranteed uniform voting machinery and mechanisms of challenge and review, for instance, that could not be changed without forfeiting the benefits Congress offered. There could be no objection under Article II to a state legislature adopting the model code; a legislature would of course be free to repeal the code later, but it would presumably face great political pressure not to do so.

These are extraordinary measures, and many people will be understandably timid about altering a constitutional structure that has been, as a whole, dramatically successful. But the Constitution's original design for elections, rooted in an elitism that is no longer tolerable, has proved its most unsuccessful feature. We have had to amend it before—in 1913, when the power to choose senators was taken away from the state legislatures and given to the people—in order to keep faith with our most basic constitutional conviction, which is that the Constitution creates and protects genuine democracy. We have now witnessed new and frightening challenges to that assumption, culminating in a deeply regrettable Supreme Court decision, and we must again change the Constitution in order to sustain our deep respect for it and for the institutions that guard it.

NOTES

---

5. EXCHANGE BETWEEN RONALD DWORKIN AND CHARLES FRIED

RESPONSE TO RONALD DWORKIN, A BADLY FLAWED ELECTION

CHARLES FRIED

I cannot claim to be a disinterested observer of the events Ronald Dworkin comments on in his essay "A Badly Flawed Election." I was counsel of record to the Florida legislature in the two Supreme Court cases spawned by the tabulation of the vote in Florida. In our second brief my Harvard colleague, Einer Elhauge, and I presented arguments that closely paralleled the Court's opinion as well as the concurring opinion of the Chief Justice. In spite of that involvement—maybe because of it—I readily concede that this was a difficult case with two sides. Quite unjustified, however, is Dworkin's high dudgeon and barely concealed innuendo that the Court had acted injudiciously out of a partisan zeal to protect its own agenda against future unsympathetic appointments. On the contrary, I see the Court as having reluctantly done the job its commission required of it.

In its first opinion of December 5 the Court reminded the Florida Supreme Court that its work in this matter was not solely a matter of state law (as Dworkin repeatedly suggests) but that it was the Constitution (in Art. II, sec. 1, cl. 2, dealing with the choice of the president) that committed the matter to the state legislature and a federal statute (3 U.S.C. § 5) that assumed that disputes regarding presidential electors were to be resolved by rules established prior to the election in question. So it was a premise of that first opinion that the faithfulness of the Florida Supreme Court to the directions of the state legislature and to preexisting rules was a question of federal law and thus ultimately a proper subject for review by the United States Supreme Court. That opinion was unanimous. I say the Court showed a proper reluctance about becoming involved because it was at

---

pains to achieve unanimity and because having issued its reminder, it remanded the case to the Florida Supreme Court.

The second opinion, which Dworkin writes about, came a week later, when the Florida court, to which the Supreme Court had shown traditional deference, had—so it seemed to many, including three of the seven justices of the Florida court—refused to take the hint and come down with a decision that merits at least as much criticism as Dworkin directs at the United States Supreme Court. It is true that the second time around the Court showed a good deal less deference to the Florida court, but that is often the case when a lower court appears to the justices to be taking its direction in less than a wholehearted spirit.^[1] Thus Dworkin's repeated characterization of this as an unprecedented and unwarranted interference in a matter of state law is misleading and incorrect. He writes that it is "difficult to find a respectable explanation of why all and only the conservatives voted to end the election in this way" and, becomingly, that "we should try to resist this unattractive explanation" that the Court majority was acting only to protect its own radical "conservative" agenda. Dworkin does not try hard enough. Instead he relentlessly casts the disagreement as one between the five "conservative" and the four "liberal" justices, with only the former moved by partisan motives. Even if the divide were as neat as he says, one wonders why exactly the same charge of partisanship could not be leveled against the four dissenters. But in fact the divide is not at all neat. For instance, on the most bitterly contested issue dividing the Court for several decades now, the right to choose established in Roe v. Wade, two members of the majority are committed to a version of the same position Dworkin espouses and the dissenters favor. Indeed, there are ideas and whole phrases in the O'Connor, Kennedy, Souter joint opinion in the Casey case that might have come straight out of Dworkin's writings. The same might be said about Justice O'Connor's opinion in the "right-to-die" cases. Surely neither she nor Justice Kennedy can fairly be readily relegated to some caricaturial conservative pigeonhole. And for that matter, Justices Scalia and Thomas are a good bit more "liberal" (if one must use these degraded and inaccurate labels) than Justices O'Connor and Breyer on a number of issues, such as free speech. Many commentators who share Dworkin's outrage cite, as if it proved something, Justice Stevens's dissent saying that the Court's decision will endanger public respect for the judiciary. But this is just the kind of thing he and Justice Scalia are sometimes inclined to say when they lose (e.g., Scalia in Evans v. Romer, the Colorado anti—gay rights initiative). Vehemence in dissent is traditional, but fouling your own nest always seems desperate.

Dworkin disagrees with the Court's judgment that the kind of recount ordered by the Florida Supreme Court was a denial of equal protection because it "puts no class of voters, in advance, at either an advantage or a

---

disadvantage." But the Supreme Court has made clear—as recently as last year in a unanimous opinion in a jejune case involving one family's sewer connection^[2]—that disparate treatment may violate the Constitution's guarantee of equal protection even if no identifiable class of persons is the target of the intentional disparity. The Florida court had explicitly ordered a procedure to take place that treated persons' votes in a senselessly variable manner.^[3] Dworkin argues that in fact there was one, general uniform standard: whether each ballot, taken as a whole, showed a clear intent to vote for one or another of the candidates for president. Discerning intent from a will or contract, a statute, or even the Constitution, taken as a whole, is a familiar and appropriate task for legal interpretation. Like Dworkin, I too am a fan of hermeneutics applied to such texts. Applied by many scores of variously trained, instructed, and supervised ballot counters to punched pieces of cardboard, such a concept is manifestly out of place, to say the least.^[4] In such stylized settings only a stylized system will do, and that system can and therefore should be uniform. But this is, for some of the reasons Dworkin gives, a question with two sides. In the end only two justices agreed with Dworkin, so this is hardly a cause for fulmination, dire warnings of the sort issued by Justice Stevens, or imputations of dishonestly partisan motives.

Although Dworkin finds the equal protection argument "defensible," agreeing with Justices Souter and Breyer, he argues that the Court's decree shutting down the election was not. The Court based its conclusion on what it discerned as the intent of the Florida law on no account to miss the December 12 safe harbor deadline, a deadline that evidently would not allow a second try by the Florida Supreme Court on that very day. This was the least convincing portion of the Court's opinion, but it too does not justify the depth of Dworkin's or the dissenters' scorn. They would have had the Court remand the case to the Florida Supreme Court to fashion a remedy that met the equal protection objection. But Dworkin and the two justices who dissented from the Court's remedy take it as a given that a recount on those terms would in any event have to have been completed by December 18, the day on which by federal law the electoral votes must be reported. But such a recount could not be completed in six days any more than in twenty-four hours. That is because that recount would go forward under the contest provisions of Florida law, and those envisage not a simple tally but a full-blown legal process, complete with briefing, oral argument, and a full recourse to appellate process. Such contests in Florida have been known to require sixteen months. So imagine what would have happened if Dworkin's and Breyer's solution had been adopted. There would have been further arguments in the Florida Supreme Court on remand, followed by an opinion from that court—which may have occasioned further review in the U.S. Supreme Court. Then the recount would

---

have taken place and there would have to have been still more process about that. If miraculously all this had been compressed into six days that fact itself would have occasioned a complaint to the Supreme Court that the Florida court had once again failed to comply with the preexisting standards of Florida law.^[5] Would such a continuation of the legal proceedings, inevitably leading to an indeterminate outcome, really have been a more satisfactory course? Surely if that was the alternative, the Court did well to shut the thing down then and there.

Finally, I think that the three concurring justices, whose views Dworkin does not discuss, were on sounder ground than the seven who found an equal protection violation. They argued that the Florida Supreme Court had not just interpreted some ambiguous language in the Florida statutes in a questionable way—a disagreement that perhaps the Supreme Court would have been well advised to let lie—but had turned that scheme completely on its head. Since fidelity to preexisting Florida law is a requirement of federal law—both statutory and constitutional—such a radical departure called for correction. The argument about Florida law is intricate, but its crux—as Elhauge shows in the portion of our brief for which he was principally responsible—is the Florida Supreme Court's premise that an interpretive manual recount is always preferable to a mechanical one and that in a close election an interpretive recount must always be had even if there was no evidence of fraud or mechanical breakdown. To argue that Florida law requires such a recount whenever the outcome might be affected (i.e., every close election) is to beg the question. Florida law insists that all legally cast ballots be counted, and it was the contention of the secretary of state that all such votes had been counted, while the famous undercounts with imperfectly perforated chads were by hypothesis not legal votes.

But in the end all this high dudgeon is unjustified for a deeper reason. This election, as any statistician will tell you, was in effect a tie. A difference of 0.5 percent in an election in which a hundred million votes were cast—at various times, under diverse circumstances, by a wide variety of means— exceeds our present capacity for accurate tabulation. The mantra of the Gore people, that we should keep counting until we can be sure that every vote has been registered, would have brought us more and more laborious recounts, with different results from each but no greater accuracy.^[6] So I agree with Dworkin's proposals that for the next presidential election we standardize and modernize the machinery, schedule, and procedures. (Such improved machinery might have seen Richard Nixon and not John F. Kennedy president in 1960, but we rarely hear that the latter's presidency was illegitimate.) As for this election, what we saw was a range of institutions—from local canvassing boards to the U.S. Supreme Court—struggling with a freakish situation beyond the capacity of any to resolve to everybody's satisfaction.

---

REPLY TO CHARLES FRIED

RONALD DWORKIN

In the article to which Charles Fried responds, I recommended a constitutional amendment abolishing the Electoral College and providing for the direct election of the president. Many readers warned, as I had suggested, that the smaller states, which have more electoral votes per citizen than the larger ones, would block that amendment because it would eliminate that advantage. But small states would not have that reason for rejecting the Model Uniform Election Code I also recommended, which would be offered for adoption to states one by one in return for federal financing of advanced voting machinery.

The model code would provide that each state's present number of electoral votes be split among presidential candidates as near as possible in proportion to the popular vote for each candidate in that state. If most states adopted that provision, the national electoral vote would match the national popular vote much more closely than it does now, and the risk that the winner in the popular vote would lose in the Electoral College would be much reduced. If Florida had enacted a code with that provision, then even without a recount Gore would have won twelve of the state's twenty-five electoral votes and would therefore have won the presidency.

Other readers expressed disappointment that I was not more critical of the five Supreme Court justices who made Bush president; it is disingenuous, they said, to try to find creditable explanations for what was so obviously a crude political decision. But the sense of legitimacy that the Supreme Court enjoys—demonstrated, once again in this case, by the fact that no one hesitated to accept its verdict even though a great many thought that verdict plainly wrong—is very important to the rule of law and principle in America, and we should not jeopardize that legitimacy out of anger at one apparently indefensible decision. It may well be, as one reader suggested, that whatever the five justices do in the rest of their careers they will be remembered, by the public and in the academic literature, mainly for their part in this decision. But that makes it all the more important to find what I said I could not yet find: at least an ideological rationale rather than one of mere self-interest for what they did.

I therefore looked forward to the comments of Fried, who was Reagan's solicitor general and later a judge on the eminent Supreme Judicial Court of Massachusetts before he returned to Harvard Law School. Fried was a coauthor of one of the many "friend of the court" briefs that urged the Supreme Court to overturn the Florida Supreme Court's recount order. Unfortunately, his essay does not provide the defense I had hoped for, and his failure will only deepen suspicion that no decent defense can be found.

---

The conservative majority made two claims: first, that the Florida Supreme Court's recount order, which specified only that ballots were to be inspected to determine the "clear intention" of the voter, violated the Fourteenth Amendment's requirement of equal protection of the laws for everyone, because different vote counters would interpret that abstract standard differently; and second, that any new recount, conducted in accordance with more concrete standards, would have to be completed by December 12, which was the very day on which, late in the evening, the Supreme Court issued its opinion.^[7] Fried attempts to defend both of these holdings, at least in substance, and then argues that because the election was so close anyway, it does not really matter whether the Court was right in either of them.

The Equal Protection Clause, as I pointed out, was designed to protect people against discrimination: it condemns, not any difference in the way a state's law treats different citizens, but only certain distinctions that put some citizens, in advance, at a disadvantage against others.^[8] The Florida court's "clear intention" standard (taken from Florida statutory law) puts no one at a disadvantage even if it is interpreted differently in different counties. Voters who indent a chad without punching it clean through run a risk that a vote they did not mean to make will be counted if they live in a county that uses a generous interpretation of the "clear intention" statute; or they run a risk that a vote they meant to make will be discarded if they live in a county that uses a less generous interpretation. But since neither of these risks is worse than the other—both threaten a citizen's power to make his or her vote count—the abstract standard discriminates against no one, and no question of equal protection is raised.^[9]

It was a striking embarrassment for the conservative majority's equal protection ruling that it cited not a single case in which a nondiscriminatory law had been held to violate the Equal Protection Clause.^[10] In his response, Fried cites Willowbrook v. Olech, in which the Court held that a town had denied a home owner equal protection when it demanded a thirty-threefoot easement over her property, as a condition for connecting a water supply to the house, although it demanded only a fifteen-foot easement in the case of other houses.^[11] But this citation shows that he has misunderstood the problem. The town had certainly discriminated against the home owner in that case—Justice Breyer, in his concurring opinion, noted, as important, the fact that the home owner had alleged that the town wanted deliberately to injure her because she had sued it in another case.

Many people have said that even if it was not unconstitutional for the Florida court to have used the abstract "clear intention" standard, rather than a more precise set of instructions about chads and indentations, using such an abstract standard was nevertheless unnecessary and unwise.^[12] They believe that even though the law often instructs judges and juries to determine

---

a person's intention with no more precise directions—to determine the intention of a deceased testator in writing the will he did, for example— such an abstract standard is out of place, as Fried puts it, "applied by many scores of variously trained, instructed, and supervised ballot counters to punched pieces of cardboard."

But judges and juries are also variously trained and supervised, and the difference between a ballot and a will is at best a matter of degree. Indeed, it might well be harder to set out in advance sensible criteria for interpreting the visual clues on a "punched piece of paper" than for interpreting the words of a will: it might be reasonably obvious from comparing a slight depression next to the name of a presidential candidate with the much more forceful indentations elsewhere on the ballot that a voter did not mean to vote for president, for example, though very hard to formulate, in a mechanical rule, a test for the comparative force of indentations. Legislatures must often choose between the imprecision of an abstract standard, like "clear voter intent," which risks differing interpretations, and the opposite dangers of a mechanical test that is almost certain to produce mistakes in particular cases. The claim that it is irrational to make either of these choices whenever the evidence in question is visual rather than verbal is surely wrong. Many states besides Florida, we must remember, have chosen the same abstract standard for manual recounts.

In any case, the question for the Supreme Court was not whether the Florida court's choice of standard was wise but whether it violated the Equal Protection Clause, and Fried has done nothing to support the majority's holding that it did or to cast doubt on his own earlier decision, while a Massachusetts judge, upholding and applying the very standard that he now believes to violate the Constitution.^[13] His failure, indeed, persuades me that the majority's decision was not even defensible, as I said it was. He next discusses the Court's even more plainly indefensible decision that any new recount under a single standard would have had to be completed by the end of December 12 so that any new recount was therefore made impossible. Fried concedes that this critical holding "was the least convincing portion of the Court's opinion" and does not try to defend it. But he insists that the Court's error made no difference, because, he says, a new recount could not even have been concluded by December 18, six days later, when the Electoral College was scheduled to vote.

His judgment about the feasibility of a recount in six days is speculative: the Florida Supreme Court might well have heard arguments and declared new standards within a day or two, and it could then have ordered that new recounts proceed as expeditiously as possible and that these recounts continue while any party who objected to the new standards appealed to the Supreme Court yet again. In any case, however, it would have been wrong for the Court to prejudge the question of whether a December 18 deadline

---

could have been met. Florida already had a slate of electors pledged to vote for Bush, and they would have voted for him on that date if the recount had not been completed. If it were completed after that date and showed that Gore had actually won, then a rival set of electors might have sent their votes to Congress, and that body would then have decided whose electoral votes to accept. That is the procedure the Constitution specifies, and the Court had no right to subvert it.

Fried next suggests that the strongest argument for what the Court did was one that in fact appealed only to the three most conservative justices, Rehnquist, Scalia, and Thomas. They said, and he agrees, that the Florida Supreme Court did not simply attempt to interpret Florida law, as it had a responsibility to do, but instead radically revised it. The Florida statute provides, as one ground for contesting an election, that the certified count has rejected a "number of legal votes sufficient to … place in doubt the result of the election." The key question is whether that provision permits a court to order a manual recount to determine how many of the ballots the vote-counting machines had rejected because of imperfect perforation actually demonstrated a clear intention to vote for one candidate or the other. Fried's brief argued (relying mainly on a debatable interpretation of provisions governing not the challenge stage of a postelection proceeding but the earlier protest stage) that the Florida statute assumed "by hypothesis" that imperfectly perforated ballots were not legal votes, so that the Florida court "begged the question" by assuming that a manual recount of such ballots was permissible even in a challenge proceeding.

But the Florida court's contrary interpretation is surely, at a minimum, defensible. That court attributed a perfectly sensible purpose to the challenge provisions of the statute: those provisions aimed, it supposed, to ensure that machine limitations should not change the result of a very close election by rejecting ballots that were legally cast and showed a clear intention to vote. That interpretation is attractive in itself and fits comfortably into the contest part of the Florida scheme, where success requires showing that the alleged error would have changed the overall result. Bush's own witness in the contest trial, John Ahmann, admitted that the machines used to tabulate punch-card ballots, which he had helped to design, often make mistakes, so that manual recounts are desirable in close elections.^[14] If we accept that understanding of the statute's purpose, such ballots are indeed "legal votes," and it is Fried, not that court, who begs the question. He and his colleague, Elhauge, may think their own interpretation better. But they have no basis for the extraordinary conclusion that the Florida court's interpretation was so demonstrably wrong as not to count as an interpretation at all.^[15]

Fried's closing observations are unfortunate. He says that it does not really matter whether the Supreme Court's decision was defensible because

---

the election was so close anyway that it might as well have been Bush. He treats the fact that five of the nine Supreme Court justices are political conservatives as, in effect, a tie-breaker like a coin flip. But it will make a very great difference to people everywhere that it is Bush rather than Gore who is America's president. Bush's cabinet appointments (which include, as attorney general, John Ashcroft, the candidate of the religious right) have already refuted the optimistic assumption that he will try to govern from the middle, and there is ample reason to worry that his future Supreme Court appointments will be equally aimed at pleasing the extreme right of his party.

When a presidential election is close, particularly when so much turns on the outcome, it is more and not less important that the rules in place be followed punctiliously in deciding who actually won. If Gore has won not only the national popular vote but also the popular vote in Florida, and hence the Electoral College as well, then it is not only unfortunate that we will be governed by Bush's policies and constituencies but unfair. Florida's vote-counting machines, many of which are conceded to be inaccurate, particularly in counties with a high proportion of minority and poor voters, declared 3 percent of the state's ballots nonvotes. (The average in the rest of the country was 2 percent.) Fried insists that it is beyond "our present capacity" to achieve a more accurate tabulation. But the careful unofficial recounts now being conducted separately by the Miami Herald and by a group of other prestigious newspapers, which Fried declares a foolish exercise, will presumably show that his surprising pessimism is unfounded.

NOTES

Fried's Response to Dworkin

---

Dworkin's Reply to Fried

---

6. BUSH V. GORE:
THREE STRIKES FOR THE CONSTITUTION,
THE COURT, AND DEMOCRACY,
BUT THERE IS ALWAYS NEXT SEASON

Michel Rosenfeld

I

In the often quoted words of Justice Stevens's strong dissent in Bush v. Gore,^[1] "Although we may never know with complete certainty the identity of the winner of this year's Presidential Election, the identity of the loser is perfectly clear. It is the Nation's confidence in the judge as an impartial guardian of the rule of law."^[2]

Actually things were much worse on that fateful day, December 12, 2000, when, in a 5–4 decision, a deeply divided Supreme Court put an end to the election and anointed George W. Bush president. Indeed, not only was the Supreme Court's majority decision patently unprincipled, but it also revealed a troubling flaw in the Constitution and potentially severe cracks in American democracy. As the Court noted, the Constitution does not give American citizens any right to vote for president.^[3] Currently, all fifty states provide their citizens with the right to vote for the state electors who become part of the Electoral College charged with the election of the president and vice president. But nothing in the American Constitution prevents any or all of them from taking that right away from their citizens and choosing electors as their respective legislatures may in their sole discretion direct.^[4] And, in fact, the Republican majority in the Florida legislature was well on its way to naming its own slate of electors regardless of the final tally of votes and could have thus single-handedly sealed the fate of the 2000 election, had the U.S. Supreme Court not done so beforehand.^[5]

America's original 1787 Constitution did not provide a right to vote either for senators or for the president. The Constitution was amended in 1913, however, to afford citizens a right to vote directly for their senators.^[6] Nothing of the sort has occurred regarding presidential elections, thus

---

leaving the citizenry without a constitutional right to a voice in the election of the country's only truly national representatives. Whatever virtues this constitutional scheme may have had in 1787, it is certainly hard to justify in 2000 that citizens of the most powerful democracy have no secure right to participate in the selection of the person who is to occupy the most important political office in the world.

One may debate whether reliance on the Electoral College—which gives disproportionate power to smaller states—should be currently considered constitutionally undesirable. It is clear, however, as vividly illustrated by the 2000 election, that the prevailing constitutional scheme does not necessarily go hand in hand with the promotion of democracy. Although periodically discussed as an odd but improbable possibility, for the first time in more than one hundred years, the winner of the Electoral College majority (assuming for the sake of argument that Bush was the clear winner of Florida's twenty-five electoral votes) was the loser of the popular vote nationwide. Overall, Gore obtained in excess of half a million votes more than Bush.^[7] Ironically, had it not been for the Florida debacle, this undoubtedly would have been the principal focus of public discussion in the aftermath of the 2000 election. America has changed dramatically since the late nineteenth century, the last time that the winner of the popular vote lost the presidential election. Today, the president's political legitimacy for wielding the vast powers that come with the office is primarily derived from having won the support of a majority (or at least a plurality) of the citizens who cast ballots on election day. Thus, even with an incontestable Bush win in Florida, the 2000 election still would have resulted in a significant setback for American democracy.

At a minimum, democracy requires that elections be won by the candidate who garners the greatest number of cast votes and that each vote that is cast be given equal weight. Recourse to the Electoral College may on occasion run counter to the first of these requirements, but it always runs afoul of the second one. The vote of a citizen of Montana or Wyoming weighs more than that of a citizen of California or New York.^[8] Particularly when the winner of the popular vote also wins a majority of the Electoral College, the constitutionally sanctioned departures from the grant of equal weight to all voters may result in a tolerable constraint on democracy for the sake of federalism—although this argument sounds much less persuasive if one remembers that one of the principal reasons for settling on the Electoral College in the first place was to boost the relative electoral power of the slave-owning states via-à-vis the rest of the Union.^[9] Accepting deviations from the grant of an equal weight to each vote within the same state, however, lacks any plausible justification and merely weakens democracy. Yet that is exactly what happened in Florida (and many other states that

---

were not subjected to the same intense scrutiny) as a consequence of reliance on countywide rather than statewide control of presidential elections.

Florida's sixty-seven counties were divided, for purposes of the 2000 election, among those using optical scanning voting systems and those using punch-card ballots. Of the approximately 6 million votes cast in Florida, 3.7 million were cast under the punch-card system while 2.3 million were cast under the optical scanning system.^[10] The rate of "undercounts" under the former system was three times as high as the rate under the latter, with the consequence that for every 10,000 votes cast under the punch-card systems there were 250 more nonvotes than under the optical scanning system.^[11] Accordingly, because of this discrepancy the relative weight of a vote cast in a punch-card county was measurably lower than that cast in an optical-scan county. Furthermore, to the extent that punch-card counties tended to be those with greater proportions of African-American voters and that allegations of intimidation and exclusion of African Americans from the voting process were accurate, the discrepancies in the weight of votes cast in Florida were not merely random. Instead—and this is even more troubling—the weight of the vote in Florida varied to some significant degree depending on the race of the voter.^[12]

II

Democracy, equality, and federalism all play important roles in American presidential elections but do not always pull in the same direction. Furthermore, the Constitution provides a framework for conducting and settling elections, but that framework is far from perfect or complete. Because of these factors, the U.S. Supreme Court was faced with a limited set of options when it was confronted with the conflicts that arose from the Florida election. A fair evaluation of the Court's handling of Bush v. Gore should properly account for the limitations in question. In the end, however, in spite of these limitations, the Court still deserves full blame for abruptly putting an end to the election. Indeed, under existing constraints, the Court had but one viable option that would have fully vindicated its institutional integrity: declining to intervene. Had that occurred and had the counting of contested Florida ballots been completed, then any remaining issues could have been settled by the U.S. Congress as specifically prescribed by the Constitution.^[13]

To better appreciate the full magnitude of Court's ill-advised intervention, let us briefly imagine what reasonable advice Americans might have had for another country having experienced something akin to what happened in Florida. Suppose that in a foreign country with a fragile democracy and with a history of discrimination in voting against certain ethnic

---

minorities, there were a close election under the same circumstances as those that existed in Florida during the 2000 presidential election. Suppose specifically that a large number of members of an ethnic minority having suffered widespread voting discrimination in the past were turned away from the polls, intimidated, falsely accused of having a criminal record that would disqualify them as voters, or given ballots that were so confusing that many cast them incorrectly, leading to an inordinately high number of invalid votes. Imagine further that on top of all that our fragile democracy, located somewhere in the Balkans, had an experience similar to that of Palm Beach County with the butterfly ballots. Because of that ballot's confusing design, more than three thousand Gore supporters who went to the polls to vote for him unwittingly registered a vote for Buchanan, a fact corroborated in many postelection interviews and acknowledged by Buchanan himself.^[14] To compound matters, a very large percentage of those 3,000 votes were cast by Jewish voters for whom Buchanan was a most improbable choice given widely held perceptions of him as having anti-Semitic leanings. The 3,000 votes at stake would have been sufficient to hand the presidential election to Gore as the official Bush lead in Florida had been set at 537 votes. Given that and the particular circumstances surrounding the butterfly ballot fiasco, its practical and symbolic place in the Florida election loom large. Moreover, it is easy to imagine that Americans committed to freedom and democracy would be quite upset if a similar set of circumstances had proved decisive in our frail and potentially explosive putative Balkan democracy.

The incidents described above and their likely impact on the 2000 election should certainly be considered an embarrassment for a country that regards itself as an example to be emulated everywhere. Bad as they are, however, these incidents were by no means the only ones that marred the presidential election in Florida. Indeed, partisan politics further damaged the integrity of the Florida vote, in relation to both the conduct of the election and the determination of its outcome.

In view of the closeness of the votes and of the problems already mentioned, it was far from ideal that the person in charge of supervising and certifying the vote count in Florida was that state's secretary of state, Katherine Harris, a partisan Republican who served as co-chair of the Bush Florida campaign committee. Even worse, her boss, the state's chief executive, Governor Jeb Bush, was not only the presidential candidate's younger brother, but he had also publicly given assurances before the election that the state would go to his brother, George, notwithstanding that polls taken as election day neared indicated that Gore might well win Florida. Although no impropriety has been associated with Jeb Bush's preelection statements and although he officially recused himself from the postelection vote counting and certification, behind the scenes he played an active and intense

---

role on behalf of his brother after November 7. Among other things, Jeb Bush mobilized his political associates in Florida to work assiduously and relentlessly against Gore and mustered his vast resources and contacts within the state systematically to assist his brother's efforts to secure Florida's twenty-five electoral votes.^[15]

For her part, Harris, who had promised to carry out her duties in a neutral fashion, made a series of crucial decisions, all of which helped Bush and hurt Gore. Moreover, in her zeal to help her political cause, she did not shy away from acting inconsistently. Thus, for example, Harris contradicted her own strict vote counting standards, which had thoroughly frustrated Gore, and shifted to a much more liberal standard when it came to counting absentee ballots, which predictably allowed Bush to pick up a significant number of crucial votes.^[16]

Another partisan action that cast a shadow on 25,000 absentee ballots occurred before the election when Republican Party workers improperly took it upon themselves to fix several thousand absentee ballot requests filed by Republican voters in Seminole and Martin Counties. Had these ballot requests not been illegally fixed, the corresponding ballots would not have been issued, costing Bush potentially 4,000 votes. After the election, Democrats brought suit to invalidate all ballots that might have been tainted by the "fixed" applications. Although these lawsuits failed because the Florida courts reasoned that the improprieties involved related to the ballot application rather than to the ballots themselves, the Florida Supreme Court chastised the Republican officials involved for having clearly violated Florida's Election Code.^[17] Had these officials not acted illegally Gore might have won the election. In any event, this incident further undermined vindication of the principle that all voters are entitled to equal treatment and to an equal right to vote.

After the election, both political parties mobilized on a national as well as statewide basis to influence the postelection proceedings in Florida. The Republicans were far more effective than the Democrats, which would not have mattered in and of itself. The Republicans, however, controlled both the executive and the legislative branches in Florida, and some of their most reprehensible actions may well have had a decisive impact on the ultimate outcome of the Florida election. The most glaring example of Republican misconduct concerns the suspension of the manual recount of 650,000 ballots in heavily Democratic Miami-Dade County in the aftermath of an intense and briefly violent demonstration by Republican partisans on November 22, four days before the recount deadline set by the Florida Supreme Court. The Republican demonstrators had set to pressure the canvassers proceeding with the recounts, and at one point, they physically assaulted the chairman of the Miami-Dade Democratic Party who had to be escorted to safety by police. Shortly thereafter, the recount was halted

---

to the delight of the protesters. A nonpartisan supervisor of elections stated in an interview that the demonstrators had indeed influenced his decision to suspend the recount. Democrats accused Republicans of deliberately using intimidation and violence to procure a cessation of the recount. Republicans countered that the protest had been nothing but the spontaneous expression by Republican sympathizers of frustration at the seemingly endless recounts.^[18] Contrary to this claim of spontaneity, however, the protesters had been led by staff workers sent from the offices of U.S. Senate Majority Leader Trent Lott and U.S. House Majority Whip Tom DeLay. Again, if we transport this sequence of events to our putative Balkan democracy, most Americans would find the organized partisan intimidation reprehensible if not downright unacceptable.

Taking into account the cumulative effect of what has been discussed thus far, only one solution to the Florida election mess would have been truly fair: to conduct a new election under uniform standards throughout the state and to entrust the subsequent vote count and certification to nonpartisan institutions and officials. Unfortunately, that solution was not available for various reasons, chief among them that federal law prescribes that presidential elections must be held simultaneously throughout the United States.

In the face of limited options and a need to minimize the evils of excessive localism and excessive partisanship, the ultimate resolution of the 2000 election was left to the courts. Although courts are supposed to remain above politics, the two courts that played the most important role in the controversy, the Florida Supreme Court and the U.S. Supreme Court, were both branded as partisan. The Florida Supreme Court judges were accused of acting as partisan Democrats and the U.S. Supreme Court majority that in effect put an end to the election as partisan Republicans. As we shall see below, there are ample grounds to conclude that the U.S. Supreme Court majority was excessively partisan. The same, however, cannot be fairly said of the judges on the Florida Supreme Court.

III

To appreciate why the five U.S. Supreme Court justices who handed the election to Bush were excessively partisan and why the Florida Supreme Court judges were not, it is necessary to consider briefly certain basic issues concerning judges and justice. Ideally, justice should be perfect and complete, and judges should be impartial and above all politics. In reality, however, this is rarely, if ever, possible. In a complex modern legal system like that of the United States, dispensing justice depends on interpreting and applying laws, and laws are by no means always good, clear, or complete.

---

The Florida electoral law that figured so prominently in Bush v. Gore is a good case in point. As we will see more fully below, this law was fragmentary and contradictory and fell far short of providing the necessary means to secure justice—justice in this case being fair and equal voting for all eligible Florida voters and accurate, uniform, and complete counting of the votes after the election. If we add to this the imperfections and lacunae of the relevant constitutional provisions and applicable federal laws, full justice in Bush v. Gore would have been impossible even if all judges involved had been superhuman.

Human judges, moreover, cannot achieve complete impartiality or remain above all politics. After all, judges are human beings with particular histories, beliefs, identities, prejudices, and political leanings—at least in the sense of leanings concerning what is best for society as opposed to preferences in partisan politics. For example, a judge who had been a civil rights lawyer may have a different outlook on race relations or gender issues and be more sensitive to certain discrimination claims than a judge with no comparable experience.

Judges may not be able to achieve absolute impartiality or to ascend above all politics, but that does not mean that all kinds or degrees of partiality or of politicization of the judiciary should be regarded as being equivalent. It is obvious that the kind of partiality at stake when a judge has a personal financial interest in a case or a close family relationship to one of the litigants is altogether different from the kind of partiality attributable to being generally more or less sympathetic to feminist claims. Accordingly, whereas it is impermissible for a judge to decide a case in which he or she has a personal financial stake or in which a relative is a litigant, no one would seriously suggest that a judge should be forbidden to decide a women's rights case because that judge has views about feminism. Similarly, it is not the same for a judge to decide cases based on partisan politics as it would be to decide based on broader or more general political commitments, such as the pursuit of greater freedom, equality, or solidarity in society.

Judges should be impartial as regards the litigants before them even if they cannot be completely neutral in their approach to the issues raised by the litigants' claims. In a case such as Bush v. Gore, however, it is particularly difficult to achieve or maintain the requisite impartiality inasmuch as judges like their fellow citizens have a stake and presumably some level of preference in presidential elections—and even if they were indifferent, it would be difficult to avoid the appearance of having preferences. Moreover, in such a case resolution of the legal claims seems inextricably linked to partisan politics, thus exacerbating the possibility of a decision based on such politics. Accordingly, it is hardly surprising that Bush and his supporters

---

accused the Florida Supreme Court of having caved to partisan politics or that Gore supporters criticized the U.S. Supreme Court for having sacrificed principle to partisan politics.

Because of all this, it would be best if courts could refrain from intervening in presidential elections. That is not always possible, however, and in the absence of a more suitable forum for resolving legal issues bearing on the integrity of an election, not always desirable. Thus the Florida courts properly intervened to address the legal claims arising out of the state's 2000 election as no other suitable state forum was available for that purpose. In contrast, the U.S. Supreme Court's intervention was unnecessary and ill advised and even arguably constitutionally unwarranted as the U.S. Congress is given the ultimate authority to settle disputed presidential elections.^[19] Moreover, if courts intervene in elections, because of the danger of becoming engaged in partisan politics, they would do well to avoid settling the election outright. Instead, if possible, courts ought to limit their intervention to minimizing injustices and establishing procedures, leaving it to others to determine the winner. That is what the Florida Supreme Court did, for notwithstanding all the accusations of partisanship, no one was in a position to predict the ultimate outcome of the vote counts it ordered and it was entirely possible that at the end of the process Bush would emerge as the incontestable winner.

In sharp contrast, the U.S. Supreme Court squarely put an end to all vote counting and in effect handed the election to Bush. Even worse, before its final decision on December 12, in granting the stay that ordered halting all vote counting in Florida on December 9, the Court for all practical purposes ensured the election of Bush. The Court did so by running out the clock on the vote recounts that had been proceeding throughout Florida before hearing oral argument on, or giving full consideration to, the merits of Bush's claims. This reeks of political partisanship as made plain by Justice Scalia's own words in support of the stay for which he and four of his fellow justices voted: "The counting of votes that are of questionable legality does in my view threaten irreparable harm to [Bush], and to the country, by casting a cloud upon what he claims to be the legitimacy of his election."^[20] As Justice Stevens observed in his dissent from the grant of the stay, "Counting every legally cast vote cannot constitute irreparable harm," but the stay itself indeed can as it may be "tantamount to a decision on the merits in favor of [Bush]."^[21]

Courts are generally reluctant to grant legal relief prior to considering a case on the merits and only do so on a showing of "great likelihood of irreparable injury" absent such interventions. For instance, if I have a legal dispute with my neighbor over ownership of the house I live in, and he shows up one morning with bulldozers, announcing that he is ready to demolish the house to build a bigger and better one, then I should be able

---

to persuade a court that I would suffer irreparable harm if my neighbor were allowed to proceed prior to resolution of our legal dispute. Indeed, if the house is destroyed and I eventually win the lawsuit on the merits, I will have suffered irreparable harm as money compensation that would allow me to purchase a comparable house could never make up for the emotional loss of no longer living in the house I grew up in, and so on. It seems, however, entirely ludicrous to claim that Bush would have suffered a comparable loss had the vote counting been allowed to proceed until the U.S. Supreme Court decided the case on the merits. What Bush wanted was to be declared the winner of the election and to have his election considered legitimate. Not only would a win on the merits before the U.S. Supreme Court have granted him that, but the very contests and controversies in Florida, of which the ongoing vote recount was but a part, had already cast strong doubts on the legitimacy of the slim official lead he held on December 9, the day the stay was granted. Accordingly, Justice Scalia's reason for granting the stay not only seems purely partisan, it is virtually inexplicable given what the controversy before the Court was all about. On the other hand, if one allows for the supposition that the Court's majority had not already decided the case on the merits prior to receiving written briefs and to hearing oral argument and consider it possible for one of the five justices leaning toward Bush to have changed his or her mind on full consideration of the case on the merits, then the five precious days lost between the grant of the stay and the decision on the merits may have made it impossible to complete the vote counting process to which Gore would have been entitled, thus in fact clearly causing him and the American people irreparable harm.

While there may be a difference between partisan politics and less pointedly focused politics, if judges cannot rise above politics, does it really matter all that much that five justices came down on the side of Bush? If judicial outcomes relating to issues of great public interest cannot avoid being either more liberal or more conservative, for example, then the loser will always feel that justice was not well served while the winner will be satisfied that it has been rightly vindicated. If this is true, then the controversy over Bush v. Gore ultimately should not be considered different from those that tend to arise when the Court decides cases on such divisive issues as abortion, affirmative action, or flag burning. Whatever the decision, a significant portion of citizenry is bound to feel that the case was wrongly decided because of what they perceive as the Court's liberal or conservative bias, as the case may be.

Some people, chief among them Justice Scalia, on the other hand, altogether deny that judgments have to be in any way political, provided that judges stick to the "plain meaning" of the laws they are charged with interpreting.^[22] Ironically—in view of how political the majority's decision in

---

Bush v. Gore would turn out to be—candidate Bush announced during the presidential campaign that if elected he would name judges in the mold of Justice Scalia and Justice Thomas to the Supreme Court because he was committed to having judges that "interpret the law rather than making the law." Although, as we shall see, this is but a vacuous slogan, it sounds attractive because it suggests that no matter how political a law may be, a judge interpreting it consistent with its "plain meaning" would not resort to any politics, as he or she would be exclusively guided by what the law in question clearly commands. Were this true, moreover, while the relevant constitutional provisions and legal rules in Bush v. Gore may have been somewhat defective, the Court's majority decision should be viewed as a faithful application of the pertinent legal norms to the controversy at hand, thus lifting the majority justices above the political fray. And if any lingering doubts remain, it must be because of an erroneous projection of the laws' inherent bias onto the judges.

On careful consideration, this last argument is but a smoke screen. A judge confronting the controverted constitutional issues raised in Bush v. Gore could not genuinely expect that the case could be resolved through any kind of reference to the "plain meaning" of the legal texts involved. It does not follow from that, however, that because any plausible resolution of the case would in some sense be political, all political outcomes would be equivalent, or that the use of certain traditional constraints on the politics of judging, most notably those imposed by the requirements of consistency and integrity, could or should not be used.

The constitutional provision that played the pivotal role in Bush v. Gore is the Equal Protection Clause of the Fourteenth Amendment, which provides that no state shall "deny to any person within its jurisdiction the equal protection of the laws." What that clause means has been highly contested since its adoption in 1868 after the conclusion of the Civil War. In the sole context of racial equality, the U.S. Supreme Court's interpretation of equal protection has fluctuated from sanctioning officially mandated racial segregation as acceptable to finding it unconstitutional and even in some cases, but not others, to accepting affirmative action programs involving preferential treatment based on race as constitutionally permissible.^[23]

It does not take a legal norm whose meaning is as contested as that of the Equal Protection Clause to realize that reliance on any "plain meaning" is bound to be illusory in most if not all cases. Take, for instance, the seemingly simple legal command "No vehicles shall be allowed in this park" posted at the entry of an urban park used for recreation. Suppose further that you are dispatched to the entry of this park with the duty of securing compliance with the prohibition in question. Should you deny entry to the park to a person in a wheelchair? Or in a motor-powered wheelchair? Should you prohibit a three-year-old from riding her tricycle? What answer

---

should you give to someone who asks you whether he can use his roller skates in the park? Suppose further that the park has a pond with rowboats, which had traditionally been rented out by the hour for recreation. In view of the new law, should you order that the use of these boats be terminated? Or imagine that a park visitor has just suffered a heart attack two hundred yards from the park entrance. Should an ambulance be allowed to enter the park to assist the victim? What about a taxicab hailed by the victim's friend to speed up transportation to the nearest hospital?

Even these few questions in this relatively simple situation indicate that it is illusory to expect most laws to have a "plain meaning." Moreover, even assuming that such "plain meaning" were possible—for example, all would automatically agree without the slightest hesitation on the precise meaning of "vehicle" it would still be undesirable to apply the law "literally." If "no vehicles" means "no vehicles ever regardless of the circumstances," then the ambulance should not be allowed into the park even if that would result in the death of the heart attack victim. But is that what the legislator would have wanted had she considered this possibility? Or, if we do not know or consider relevant the intent of the legislator, is that the best or most desirable interpretation of the law? Suppose the most logical assumption about the purpose of the law is that it was meant to promote enjoyment of the park by those who use it while at the same time ensuring their safety. Would it then not squarely be against the purpose of the law to refuse entry into the park to the ambulance?

Although there may be no "plain meaning" of "vehicle," it does not follow that any meaning will do, or that some meanings are not much more plausible than others. Meanings depend on context in terms of both the situation in question—for example, enjoyment of a park—and the relevant sociocultural setting. Indeed, in any given community at any given time certain meanings are fairly settled and others contested. For example, today it is well settled in the United States that equal protection is incompatible with state-mandated racial segregation, but this was not the case at the turn of the twentieth century. Consistent with this, moreover, whereas there may not be fixed or plain meanings, judges face constraints inasmuch as certain meanings are virtually unanimously shared throughout the polity while others are contested, allowing for a range of plausible interpretations. But even in the latter case, a judge would not be justified in straying far beyond the existing range.

Even interpretation of the law prohibiting vehicles in the park cannot avoid politics—albeit such politics may be minor and uncontroversial. For example, in deciding a case on the legality of an ambulance entering the park in a medical emergency, the judge must take a position concerning the best or most rational policy considering the issues of enjoyment, safety, emergency, and recreation raised by the law and the facts before her. In

---

addition, before reaching her decision, the judge must weigh whether a "literal" application of the law or one better suited to further its most probable purpose over which there is broad-based consensus ought to be adopted. This last issue, moreover, is one of judicial politics. Should judges have greater or lesser latitude when confronting issues not directly foreseen by the legislator?

Whatever the answers to the above questions, the judge will nonetheless have to adhere to certain constraints in order to fulfill her duties legitimately. For example, the political decision as to whether opening the park to motorcycle races would ultimately constitute a better use for it would clearly be an inappropriate one for the judge to consider under the existing law. In short, just as certain meanings are excluded because of the prevailing sociocultural consensus, certain political choices would be inappropriate to the extent that they are implicitly or explicitly incompatible with the purpose of the applicable law or its implementation.

To the extent that the law leaves open certain political choices, however, judges are not precluded from following their convictions in deciding cases, and hence under those conditions both liberal and conservative solutions would be equally legitimate. But even then, judges ought to be further constrained by requirements of consistency and integrity. Even where liberal or conservative leanings may be entirely proper, a judge should not jump from one to the other solely to achieve particular results in individual cases. Finally, consistency is not an ultimate value inasmuch as there ought to be room for a judge to recognize past mistakes or changes due to evolving circumstances or a bona fide shift in convictions. Breaks in consistency, however, should be assessed in terms of a standard of integrity. While frequent lack of consistency to further a partisan agenda would lack integrity, rare breaches of consistency because of principled reconsiderations would clearly comport with preservation of integrity.

IV

Consistent with the above standards and constraints, the Florida Supreme Court decisions meet with judicial consistency and integrity in crucial ways that the U.S. Supreme Court's decision in Bush v. Gore does not. The Florida Supreme Court had to deal with a state law that was unsatisfactory, incomplete, and contradictory as well as inconsistent with federal law. Nevertheless, the Florida court managed to draw consistent and defensible conclusions, even if they were not the only plausible ones. In other words, the applicable legal norms were sufficiently open-ended to allow judges with different judicial philosophies to reach different results while remaining equally faithful to the requirement of judicial integrity and to widely accepted canons of judicial interpretation.

---

In the broadest terms, the controversy before the Florida courts revolved at first around two issues concerning the "protest" phase (as opposed to the "contest" phase). The protest phase is that which precedes the official certification of election results by the secretary of state. The contest phase begins after certification. The two issues in question were: (1) the extent to which extensive manual vote recounts were proper; and (2) the requisite deadline for certification in the face of such recounts and the extent of the secretary of state's discretion in fixing that deadline before completion of ongoing vote recounts.

The Florida Election Code, which applies to all state elections including the presidential one, provides both that election results shall be certified seven days after election day and that full manual recounts are appropriate if errors in vote tabulation could affect the outcome of the election. The code, however, does not set any deadline for filing additional or corrected vote totals resulting from such a recount.^[24] Moreover, two provisions, one dealing with the state officials in charge of counting the votes and the other dealing with deadlines for submitting vote counts to the secretary of state, were inconsistent. The first mandated ignoring counts submitted after the deadline; the second, in contrast, permitted but did not require ignoring such counts. Furthermore, the first provision had been enacted in 1951 and the second in 1989.

The code also provided for submission of counts of absentee ballots just like all other votes by the end of the seventh day after the election. Federal law, however, requires that in elections for federal office absentee ballots from overseas must be counted so long as they were signed no later than election day and were received in the state within ten days after the election. This federal requirement, moreover, allowed for an anomalous situation in an election as close as that for president in 2000. If the results were certified within seven days and the overseas absentee ballots received three days later tipped the election to the certified loser, then the latter was entitled to be declared the winner. But if properly ordered manual recounts completed between the seventh and the tenth day after the election would when added together with absentee overseas ballots tip the election in favor of the candidate who was trailing the seventh day after the election, then the secretary of state could in effect choose the election's winner by using her discretion to accept or reject the state votes tallied after the seventh day following the election.

The Florida Supreme Court dealt with the above-mentioned inconsistency in the Florida Election Code through reliance on traditional and wellsettled rules of legal interpretation. When a state law is inconsistent, a more specific provision has priority over a more general one, and a more recently adopted provision supersedes a conflicting older provision. In this case, the more permissive provision, which allowed inclusion of vote recounts concluded

---

after the seventh day following the election, was both the most recent and the most specific of the two.

In concluding that the secretary of state had exceeded her discretion in rejecting amended returns that would result from manual recounts completed more than seven days afer the date of the election, the Florida Supreme Court also remained within the bounds of legitimate judicial interpretation. The secretary had rejected the recounts on the grounds that there had been no election fraud and that the mere possibility that the recount could affect the outcome of the election was insufficient to warrant extending the deadline. Stressing that the Florida Constitution reserves a prominent place to the right of suffrage, the Florida Supreme Court construed the secretary's realm of discretion more narrowly than she had. Given that neither the Florida voters' nor the state's interest in participating to the fullest possible extent in the election of the president would be harmed by extending the certification beyond seven days, the added gain of allowing for a more accurate count in a close election outweighed whatever loss would result from somewhat narrowing the scope of the secretary's discretion. Or at least that was as plausible an interpretation of Florida's electoral and constitutional scheme as one that put greater stress on executive discretion and lesser weight on the democratic principle of making every vote count.

Subsequent to the Florida Supreme Court's decision regarding the protest phase, Bush was certified the winner by 537 votes on November 26. On November 27, Gore set in motion the contest process, which culminated with the December 8 Florida Supreme Court 4–3 decision in Gore v. Harris.^[25] Florida law allows a candidate to contest an election by bringing a lawsuit in circuit court. Gore did so and lost, but the Florida Supreme Court reversed the trial judge's decision as "clearly erroneous." Relying on a 1982 precedent and refusing to investigate the factual evidence introduced by Gore—which included 9,000 so-called undervotes (i.e., punched ballots that registered no vote for president while registering votes for other offices during the machine count) never subjected to a manual count from heavily Democratic Miami-Dade County—the trial judge, N. Sanders Sauls, flatly rejected Gore's challenge.

What made Judge Sauls's decisions "clearly erroneous" according to the majority on the Florida Supreme Court was that in relying on the 1982 precedent, he had failed properly to account for crucial changes in Florida law that had been adopted by the state legislature in 1999. Chief among these was the addition of the following ground as a basis for contesting an election: "Receipt of a number of illegal votes or rejection of a number of legal votes sufficient to change or place in doubt the result of the election."^[26] Clearly, even standing alone the 9,000 Miami-Dade undervotes

---

"placed in doubt" the result of the election as Bush's lead had shrunk to less than 200 votes by that time.

Since machines typically undercount a certain percentage of valid votes and since the Florida legislature had clearly indicated, in laws adopted before the 2000 election, that no vote should be ignored so long as "the clear intent of the voter" could be ascertained, the court ordered a statewide manual count of the undervotes. Gore had requested manual counts only in certain traditionally Democratic counties, and Bush did not ask for any such counts as his strategy was to obtain a judicial order ending all counts while he was still ahead. The court felt, however, that since the overriding legal concern was that all legal votes should be counted, the fairest and likeliest way to accomplish this would be through a statewide count of all the undervotes that bore "a clear indication of the intent of the voter."

This standard comported with the law of Florida as well as that of many other states. Moreover, inclusion of vote counts that had been rejected by the secretary of state for not having been submitted in a timely manner was consistent with recent Florida precedent. In a contest relating to an election to the U.S. House of Representatives, a unanimous Florida Supreme Court rejected the contesting candidate's request to exclude 11,000 ballots that had not been submitted on time.^[27] And in so doing, the court decided that failure to abide by time limits similar to those present in the presidential election did not constitute sufficient grounds to disenfranchise thousands of voters. In both cases, the Florida Supreme Court determined that the need for accuracy and the attempt to make every vote count outweighed the need for prompt finality. Whether or not one agrees with this, the Florida Supreme Court certainly acted with consistency and integrity and within accepted bounds of legal interpretation. In short, the Florida law may not have provided a clear and unequivocal road map to the resolution of Gore's contest, but any fair-minded person would have to conclude that the court's decision was comfortably within the bounds of legitimacy.

V

The same cannot be said, however, of the U.S. Supreme Court's majority decision. As already mentioned, the grant of the stay that stopped all vote counts on December 9 is difficult to explain except in purely partisan terms (whether deliberate or not). And so is the same majority's decision on December 12 that no time remained to remedy the equal protection violation found by the Court. Finally, some aspects of the majority opinion, and the main thrust of the concurring opinion by Chief Justice Rehnquist

---

and Justices Scalia and Thomas, contravene fundamental canons of judicial integrity and consistency.

Significantly, the five justices in the majority have been responsible for a veritable judicial revolution reversing a trend going back to the New Deal in order greatly to expand states' rights and powers at the expense of the national government. Indeed, making proof of an activism that easily matches if it does not exceed that embraced by the Warren Court in its protection of fundamental individual rights, the Rehnquist Court's majority has systematically invalidated as impermissibly intrusive on state powers widely supported laws enacted by Congress in numerous different areas, including gun possession in schools,^[28] overtime pay for state probation officers,^[29] age discrimination,^[30] and violence against women.^[31] Yet in Bush v. Gore this same majority turned against one of the most fundamental tenets of American federalism, namely, that state courts are the authoritative interpreters of state law and that federal courts—including the U.S. Supreme Court—must defer to state court interpretations of state law.^[32] Thus not only was the majority on the U.S. Supreme Court acting in a way that was glaringly inconsistent with their active promotion of states' rights, but they did so with respect to a well-settled and deeply entrenched principle of federalism supported alike by both partisans and opponents of expanded states' rights. Under the circumstances, the Rehnquist majority acted not simply in an inconsistent manner but in a way so at odds with the position they had steadfastly embraced before this case, and to which they would return immediately thereafter,^[33] as clearly to breach the most elementary requirements of the principle of integrity.

Specifically, the U.S. Supreme Court was confronted with two issues: (1) whether the Florida Supreme Court's interpretation of that state's electoral code amounted to the creation of new legal standards to resolve presidential election contests in violation of Article II, section 1, clause 2, of the federal Constitution; and (2) whether the "clear intent of the voter" standard used to carry out the manual count ordered by the Florida Supreme Court violated the Equal Protection Clause. The first of these issues was addressed only in the concurring opinion, and it tracked the Bush camp's attack against the Florida Supreme Court on the grounds that it had in essence written an entirely new law, thus usurping the role of the legislature instead of faithfully interpreting the law that the latter had previously enacted. On the other hand, the second issue, which was to prove decisive for the five justices who put an end to the election, had been virtually ignored in all the previous judicial proceedings leading to this decisive round. Actually, that was hardly surprising given the potential contradictory and explosive consequences that might follow from a serious and principled application of the Equal Protection Clause to the Florida presidential election.

---

In a nutshell, the U.S. Supreme Court's decision turned on two legal conclusions: (1) application of the "clear intent of the voter" standard in the manual count of undervotes amounted to an equal protection violation because it failed to give equal weight to every vote to the extent that ballots with equivalent physical features were treated differently by various persons entrusted with the count; and (2) the combination of federal and state legal norms applicable to presidential elections mandated that the Florida electors be finally ascertained by December 12, six days before these electors were required to cast their votes for president. As we shall see, the first of these conclusions is at best barely tenable and clearly unwise. The second, however, is downright indefensible, and as such it seems a mere pretext for purposes of securing a purely partisan result.

In the context of voting, equal protection requires an equal opportunity to vote for all eligible voters and granting an equal weight to all votes cast. As already indicated in the discussion of the problems that plagued the Florida presidential election and its aftermath, neither of these requirements was satisfied. Indeed, the differences between the optical-machine system and the punch-card system when added to the butterfly ballots, to the obstacles encountered by certain African-American would-be voters, and to the inconsistent standards for accepting votes as valid, sanctioned by Katherine Harris, clearly combine into significant departures from equal opportunity to vote and to a vote-counting process bound to result in an unequal weighing of all legally cast votes. Moreover, even if these problems had not occurred, the mere fact that responsibility for organizing, monitoring, and tallying the Florida vote was left to each of the sixty-seven Florida countries made it impossible for the state (as the Equal Protection Clause clearly requires) uniformly to guarantee or achieve an equal opportunity to vote or an equal weighing of all cast votes.

In comparison to these major problems, the equal protection concerns actually addressed by the Court seem minor if not downright trivial. Indeed, given the choice of leaving a large number of votes entirely uncounted (Bush partisans insisted that these votes had been counted because they were subjected to machine counts, but inasmuch as these machines did not register some of the votes that had been properly completed, the latter votes can be fairly said to have remained uncounted) or counting all such votes subject to certain inconsistencies, it plainly seems that the second option was by far preferable among the two imperfect alternatives actually available.

It is also important to note that contrary to suggestions in the Court's per curiam opinion to the effect that the Florida manual count was being conducted in the absence of any standard, the Florida Supreme Court had in fact imposed a uniform statewide standard. That standard, which had been adopted by the Florida legislature and which was the standard in a

---

majority of the states—including Texas where it had been signed into law by none other than Bush himself—required counting all ballots that evinced "the clear intent of the voter."^[34] The problem, therefore, was not the lack of a standard but rather that the standard was an imperfect means to the end of ascribing an equal weight to all cast votes.

The law often provides imperfect means to clearly defined ends, and in many cases it must do so because there is no plausible alternative. Thus, for example, in the area of criminal trials where the uncontested purpose is to acquit the innocent and convict the guilty, the means employed, namely, "proof beyond a reasonable doubt," are not only imperfect—both because they are skewed in favor of acquitting some of the guilty in order to minimize the chances of convicting anyone who is innocent and because in spite of that they cannot prevent the conviction of a certain number of innocent persons—they are also bound to vary in their implementation from one jury to the next. Because there is no way to quantify what should amount to proof beyond a reasonable doubt, the actual standard for convictions is bound to fluctuate from one jury to the next. Such fluctuations, however, are not deemed to amount to violations of equal protection. So long as the standard is applied honestly, inequities resulting from a different understanding of what that standard requires do not alone give rise to constitutional violations. Even when what is at stake is life imprisonment or the death penalty.

It seems that use of the intent-of-the-voter standard for purposes of granting equal weight to each legally cast vote is analogous from an equal protection standpoint to reliance on the beyond-a-reasonable-doubt standard in the context of criminal trials. Notwithstanding this analogy, one can raise two objections as the two situations under comparison are arguably different in crucial respects. The first objection is that whereas proof beyond a reasonable doubt may be the best we can do in the criminal trial context, the same is not true about the intent-of-the-voter standard in the context of elections. Thus, for example, instead of using the intent-of-thevoter standard, which is bound to lead to certain inconsistencies, one could adopt a physical standard—such as a clear punch through the ballot leaving no hanging chads—which could be applied consistently to all votes subjected to a manual counting. The second objection, which was relied on by the Court's majority to buttress its conclusion that the count ordered by the Florida Supreme Court violated equal protection, is that the analogy between the above-mentioned standards is inappropriate. This is because the fact finder in a criminal trial must deal with testimony by persons and must assess their credibility, whereas the vote counter confronts a "thing" and can therefore ascertain intent through "specific rules designed to ensure uniform treatment."^[35]

Neither of these objections is particularly persuasive, however, as the

---

first depends on a policy choice over which reasonable minds may disagree and the second relies on an overly simplistic and hence erroneous conception of interpretation. In dealing with how to handle vote counts, a legislature could choose between a criterion relying exclusively on physical marks on ballots or one focused on intent of the voter. Whereas the first alternative would promote greater certainty, the second would give greater weight to the will of the voter. Accordingly, any legislature that believed that greater fidelity to a voter's intent is a more important objective would be justified in settling for an intent-of-the-voter standard.

As to the second objection, the premise adopted by the Court's majority that specific rules provide better means for ascertaining intent when dealing with an inanimate object, such as a ballot, is plainly wrong. An undervote that is a punch-card ballot bearing certain full perforations, certain partial ones, and perhaps other signs marked by a voter must be viewed as a text conveying a message. Like every other text, the message conveyed by a perforated punch-card ballot may be immediately obvious or it may depend on placing the various signs involved in proper context. For example, should an incomplete perforation relating to the vote for president be treated differently according to whether it is completely unlike any other perforation on the same ballot or whether it bears close resemblance to some of the other perforations? Arguably, the answer is yes, for it is entirely possible that the same physical mark may amount to a clear intent to vote in one context but not in another. And if this is true, then there is no difference in kind between interpreting the testimony of a trial witness and interpreting the intent of a voter from a "reading" of his or her ballot. The main difference between the two is just one of degree.^[36]

Although the equal protection ruling handed down by the Court's majority may have been unsupported by precedent and unwise, it does not of itself exceed the bounds of legitimate judicial interpretation. When put in proper context, however, this ruling cannot meet the minimum standards of judicial consistency or integrity. One of the cornerstones of judicial integrity consists in accepting that the rule that determines the results in the case being adjudicated will have precedential value in subsequent cases raising the same or similar issues. Accordingly, in deciding a case, a judge must be mindful that the rule he or she devises for the occasion will be applicable to subsequent cases and determine in whole or in part the winners and losers in those cases in ways that the judge who is now imposing the rule may not be able to anticipate. This requirement encourages judges to think in terms of more general rules suitable for certain classes of cases rather than rules that are so narrow as to appear fit only to lend support to a particular result in the case at hand. For example, a determination that equal protection requires rejection of all requests for manual recounts in presidential elections would certainly have benefited Bush in 2000 but

---

could well work to the advantage of a Democratic candidate in some subsequent election. In Bush v. Gore, in contrast, the Court's majority carved out a very broad rule linking precise uniform statewide standards to satisfaction of the constitutionally mandated requirement of granting equal weight to all cast votes but expressly limited its application to the unique circumstances of the 2000 Florida election.^[37] In thus attempting to avoid the broad precedential implications of their novel decision, the justices in the majority adverted to the "many complexities" relating to application of equal protection to elections. As further refinement of rules to accommodate unanticipated circumstances is part and parcel of applying precedents, however, the Court's majority efforts to undercut the precedential value of their ruling strongly suggest a breach of integrity for purposes of achieving an expedient result without having to live with its clearly foreseeable consequences.

This lack of integrity is exacerbated in the case of the three justices who joined the concurring opinion. Indeed, these justices had a long history of interpreting the Equal Protection Clause in the narrowest possible terms, only invalidating laws that discriminated intentionally on the basis of race, national origin, or religion or that involved affirmative action benefiting racial minorities. Accordingly, the three justices' embrace of a completely inconsistent broad view of equal protection combined with the disavowal of its precedential value only exacerbate the impression that their only concern was to secure Bush's victory.

Of the seven justices who found an equal protection violation, only two, Breyer and Souter, acted with integrity. Indeed, both of them had embraced a broader conception of equal protection in the past, and both wanted to give the Florida courts an opportunity to tailor an adequate statewide standard that would allow the manual count to proceed. Justices Breyer and Souter were thus prepared to live with the consequences of their decision, and had their position prevailed either Bush or Gore might have won depending on subsequent actions by the Florida courts and on the outcome of the vote count conducted under new standards.

In contrast, the five justices who in effect put an end to the election did not leave themselves open to the slightest possibility that Gore might eventually prevail. And since there was no basis in law for locking in Bush's victory—as even the expansive view of equal protection that they had embraced for the occasion did not foreclose concluding the manual recount process—these five justices resorted to a veritable deus ex machina. They invoked a phantom December 12 vote counting deadline required neither by state nor by federal law. Ironically, had they not arbitrarily ordered the vote count halted on December 9, it might well have been completed by December 12 (making the complaints regarding a lack of uniform standards either moot, if Bush prevailed, or amenable to resolution in the

---

Florida courts, if Gore prevailed, while fully preserving the possibility of an appeal to the U.S. Supreme Court). And had these five justices not imposed the fabricated December 12 deadline making it explicit that any remedy the Florida court might grant to get over the equal protection hurdle would be unconstitutional, the vote count, even under a new standard, might well have been completed by December 18, the first plausible legal deadline.

The arguments offered by the five justices to support their December 12 final deadline can only be characterized as highly disingenuous. These arguments revolve around two relevant sources of law: 3 U.S.C. § 5, the federal law designed to provide a safe harbor shielding states against possible rejection of their selected presidential electors by the U.S. Congress; and the Florida Supreme Court's interpretation of that state's election code. Although the federal law in question is far from clear in all respects, it is rather plain in one respect that is crucial regarding the Florida election dispute. What is clear under 3 U.S.C.§5is that if a state acting under a law adopted before the day of the election has settled on a slate of presidential electors six days before the date fixed for the meeting of such electors throughout the nation (in 2000 that latter date was December 18), then that slate must be accepted as legitimate by the U.S. Congress when it meets on January 6 to certify the votes of the Electoral College. Thus, by settling on its electors by December 12, Florida could avail itself of a "safe harbor" and ensure against challenges in the U.S. Congress. What is also clear from consideration of all applicable constitutional and legal provisions is that a state is perfectly entitled to select its electors up to the date set for the convening of electors (December 18 in 2000) and that the vote of such electors would still count in full unless Congress decided to uphold a challenge against them. In short, nothing in federal law prohibited Florida from settling on its slate of electors at least up to December 18. The only thing that would be lost by extending the process beyond December 12 was immunity from possible challenges before the U.S. Congress.

What is not clear about 3 U.S.C. § 5, however, is the scope of the immunity it affords to electors coming within its safe harbor. It is obvious that such a scope cannot be unlimited as the Constitution itself forbids certain persons—for example, a senator or the secretary of defense—from being electors, and if a state had such a person in its slate selected within the safe harbor period, the Congress would not only be entitled to reject that person's vote, it would be compelled to do so.^[38] Therefore, the most plausible interpretation of 3 U.S.C.§5 would consist in construing the safe harbor provision as immunizing a state's slate of electors against challenges under state law but not against those grounded on the Constitution or federal law.^[39]

Federal law thus did not require Florida to wrap up its selection of electors by December 12. The five-justice majority in the U.S. Supreme Court

---

decreed, however, that time had run out under Florida state law as interpreted by the Florida Supreme Court. As the Florida Election Code does not explicitly address this issue, the sole authority on which the five justices could rely were the pronouncements made by the Florida Supreme Court in its December 8 decision ordering the recounts. It is true that the Florida court indicated that consistent with its interpretation of state law, the ordered recount should be concluded by December 12 so as not to lose the benefits of the safe harbor provided by federal law. But that was at a time when the Florida court expected that the manual count could be completed by December 12, before the December 9 stay put a stop to the count and before the December 12 U.S. Supreme Court decision decreed that the count ordered by the Florida court was in violation of the Equal Protection Clause.

Nowhere in the Florida Supreme Court's December 8 decision was there a conclusion that the December 12 deadline imposed an absolute bar to any further remedy under the Florida Election Code. Nor was there any discussion of whether falling within the safe harbor was more important under Florida law than making every vote count, or whether a proper vote count extending to December 18 would not be more consonant with Florida's law than holding fast to an illusory safe harbor as Florida's selection of electors had become vulnerable to equal protection challenges. Indeed, had the U.S. Supreme Court sent back the case to the Florida courts without having determined that the clock had run out on December 12, it is entirely possible that applying the broad equal protection standards articulated for the first time by the U.S. Supreme Court, the Florida courts would have found a whole new host of constitutional violations in the handling of the 2000 election by Florida officials. Consistent with this hypothesis, moreover, the Florida courts could have reasoned that the December 12 deadline was of little use because Congress would be free to overturn Florida's choice on the basis of federal constitutional violations. Or the Florida courts could have simply concluded that in their understanding of Florida law remedying constitutional flaws was more important than seeking safe harbor protection.

The December 12 deadline thus by no means required putting an end to the Florida presidential election aftermath under either federal or state law. Significantly, as one of the judges on the Florida Supreme Court stated in his concurring opinion in the final dismissal of Gore's case in conformity with the U.S. Supreme Court December 12 decision, "December 12 was not a ‘drop dead’ date under the Florida election scheme. … [It] was simply a permissive ‘safe harbor’ date to which the states could aspire. It certainly was not a mandatory contest deadline under the plain language of the Florida Election Code (i.e., it is not mentioned there) or this Court's prior rulings."^[40] As this judge further noted, the U.S. Supreme Court simply

---

misinterpreted Florida law.^[41] But even if that had not been the case, the U.S. Supreme Court should still have left it to the state's court to interpret state law and had no basis to usurp that power. In view of all this, there is but one inescapable conclusion: the five justices in the majority threw away all consistency and integrity to ensure that the state courts would not be able to legitimately exercise their exclusive powers in any way that might threaten the certainty or finality of Bush's victory.

Although the concurring opinion of Chief Justice Rehnquist joined by Justices Scalia and Thomas did not directly figure in the disposition of the case, it deserves a brief comment inasmuch as it lacks even the most tenuous links to consistency or integrity and as it clearly betrays these justices' pretense to textualism and "strict constructionism." One can only surmise that one of the principal reasons for writing the concurring opinion was that the three justices felt uncomfortable joining an equal protection opinion standing against what they had consistently and vigorously fought against throughout their tenure on the High Court. It seems that these justices felt compelled to join Justices O'Connor and Kennedy to form the per curiam majority for purely political reasons. Bush still would have won if the three justices refused to join the per curiam opinion, but there would have been no majority support for any legal ground that would justify the outcome of the case—a situation potentially much more politically explosive than that triggered by the actual Court decision. Under this scenario, four justices (O'Connor, Kennedy, Souter, and Breyer) would have agreed that there was an equal protection violation, but five justices (Rehnquist, Stevens, Scalia, Thomas, and Ginsburg) would have disagreed; two justices (O'Connor and Kennedy) would have agreed that the equal protection violation could not be remedied, but seven justices would not have agreed that that was dispositive (the five who would not have agreed there was an equal protection violation plus Souter and Breyer, who did agree but thought it could be remedied); and finally, three justices (Rehnquist, Scalia, and Thomas) would have ordered the end of the Florida election on the extraordinary grounds discussed below, but six justices (Stevens, O'Connor, Kennedy, Souter, Ginsburg, and Breyer) would have disagreed with disposition of the case on those grounds. In sum, Bush still would have won because three justices would have favored him on some grounds and two on other grounds. While this happens in a number of ordinary cases before the Court without much fanfare, imagine the impact of the headline "U.S. Supreme Court Orders End of Florida Recount Making Bush Winner: No Majority on the Court Agrees on the Reasons Why."

At bottom, the concurring opinion echoed Bush's often repeated charge that the Florida Supreme Court had in effect enacted an entirely new election law, thus illegitimately usurping the role of the legislature rather than acting as a court. As we have seen above, this charge is simply false as the

---

Florida Supreme Court decisions fall well within the range of entirely acceptable judicial interpretations. Moreover, being the staunchest and most steadfast defenders of states' rights on the Court, the justices who joined the concurring opinion could not simply ignore the undisputed constitutional principle that federal courts must defer to their state counterparts when it comes to interpreting state law. Instead, the three justices had to appeal to federal law but had to do so in a bizarre and largely incoherent manner because otherwise they simply could not have come up with any cogent reason for stopping the Florida vote count.

Ordinary recourse to federal law (including constitutional law) in order to invalidate a state law or a state judicial decision involves establishing an irreconcilable conflict between federal law and state law as enacted by the legislature or as interpreted by courts. A clear example of this is the per curiam opinion's invalidation of the court-ordered Florida manual count of undervotes as contrary to what is prescribed by the Equal Protection Clause. The concurring opinion, however, does not follow a similar path but instead invokes federal constitutional law as a means to prevent the Florida court from frustrating the Florida legislature. On the surface, the concurring opinion's rather unusual undertaking may nonetheless seem to lead to a plausible result. Concretely, since Article II of the U.S. Constitution explicitly empowers state legislatures in presidential elections, the legislatures arguably become federal agents for purposes of the presidential election. This would be analogous to—if for the sake of argument we do not assume it to be unconstitutional—a state governor being temporarily appointed by the president as a special envoy to a foreign nation. Clearly, under those circumstances, it would be logical for the governor to be bound by the president's instructions rather than by any laws to the contrary promulgated by his or her state's legislature.

On closer analysis, however, the Constitution does not merely make the state legislatures federal agents, and accordingly the above analogy becomes highly misleading. The relevant constitutional language is: "Each state shall appoint, in such manner as the legislature thereof may direct, a number of electors."^[42] If the legislature directed that it would itself choose the electors, then the above analogy may or may not be appropriate, depending on whether "legislature" is merely understood as the sum of all legislators or as an institution within a political framework carved out by a state constitution (in which case, for example, a valid legislative act would require the acquiescence of the governor, or if he or she were to veto the slate of electors approved by a majority of state legislators, a subsequent two-thirds majority vote by such legislators). In the case of Florida, as the legislature enacted an election code and entrusted the selection of presidential electors to the state's eligible voters, questions concerning interpretation

---

and application of the code were bound to arise. Were these questions to be answered by the courts? By the legislature?

The concurring opinion acknowledges that under the Florida election scheme as fashioned by the state legislation there is room for court interpretation of the relevant law. But it also goes on to say that in view of the federal constitutional design, "the text of the election law itself, and not just its interpretation by the courts of the States, takes on independent significance."^[43] This last statement, however, amounts to nonsense as no meaning can be derived from a text without interpretation. Indeed, no text speaks for itself as made plain by the fact that a person who does not understand English cannot extract any meaning from the Florida Election Code or any other law written in English.

Once one recognizes that all legal texts require some interpretation before they can become intelligible, then the concurring opinion's position collapses into incoherence. Indeed, the three concurring justices do not recognize an exclusive right of interpretation for the state legislatures or the federal courts, nor do they point to any independent federal norms— such as the Equal Protection Clause—that would rule out certain interpretations of state election law as being in conflict with existing federal law. Accordingly, shorn of all its deceptive accoutrements, the concurring justices' position boils down to the creation out of the blue of a new power for federal judges. As a practical matter, this new power available in the context of presidential elections allows federal judges to set aside state court decisions on state election law whenever federal judges feel like it, as all that is needed is an assertion that the state court interpretation departs from the text of the state law. While this may not, strictly speaking, give federal judges absolute discretion, it nearly does so. That is because only those laws that would be given automatically the exact same interpretation by all those who understood the language in which they were written could remain immune from any deviation between the text and its interpretation.

What is perhaps most remarkable about the concurring opinion is that while it posits texts as having meaning independent of any interpretation, it manages to undermine this very hypothesis in its attempt to demonstrate how its premises necessarily lead to its conclusions. Specifically, the concurring justices provide some aberrant readings of the relevant constitutional text and radically depart from their well-established prior (and based on decisions handed down shortly after Bush v. Gore soon to become subsequent)^[44] commitment to affording broad protection to state prerogatives threatened by congressional legislation—even if such legislation is designed, like the one at stake in the present case, to further prescriptions contained in the Constitution itself.

As I have already emphasized, the Florida Election Code is so full of

---

complexities, inconsistencies, and lacunae that the idea that its text may be given any cogent meaning without elaborate interpretation is preposterous. Moreover, as also already mentioned, the relevant constitutional text mandating selection of presidential electors as the state legislature "shall direct" is far from transparent. Most important, this text does not in and of itself specify whether the term "legislature" should be interpreted as meaning a detached self-contained body acting by and for itself or as part and parcel of an organically linked set of interrelated institutions designed to act in a coordinated fashion within a commonly shared normative framework carved out by the state's constitution. Yet the difference between these two interpretations is crucial for the principal argument on which the concurring opinion relies. Indeed, if the second of the above interpretations is the correct one, then the entire thrust of the concurring opinion is completely absurd. The Florida Constitution and election law make it clear that the Florida courts (including the Florida Supreme Court) are competent to interpret the state's election law and to adjudicate disputes under it. Consequently, since the Florida legislature is a creature of the state's constitution and since that legislature did not exclude Supreme Court review of contests in presidential elections in the election code it adopted, then the only logical conclusion is that the Florida legislature "directed" before the 2000 election that the selection of presidential electors should proceed according to the prescriptions in the election code as interpreted and applied by the Florida courts. And consistent with this, the concurring opinion's stated grounds for protecting the state legislature from the state's judiciary amount to no more than a completely arbitrary unjustifiable departure from the principle of respect for state court interpretations of state law. Furthermore, this departure is utterly inconsistent with the aggressive and highly activist record of the three justices involved regarding zealous protection of state prerogatives in general and state judicial prerogatives in particular in the context of congressional legislation designed to further federal constitutional objectives.^[45]

Whereas the preceding remarks should satisfy any fair-minded person that neither Article II of the Constitution nor the Florida Election Code "speak for themselves," it may still be arguably possible to justify the concurring opinion's conclusions based on adoption of the first of the two above-mentioned interpretations of the relevant constitutional language. In other words, the concurring justices may have been utterly inconsistent and may have paid lip service to a theory of interpretation that they themselves were squarely contradicting, but they may have nonetheless reached a conclusion that may have satisfied the requirements of consistency and integrity if it had been articulated by liberal judges with much more restrictive views of the proper scope of states' rights.

On closer examination, however, not even a liberal justice could put

---

forward a credible justification for interpreting Article II as suggesting, much less requiring, that state legislatures ought to be understood as operating completely independently from all other institutions to which they are otherwise intrinsically linked pursuant to the dictates of their state's constitution. Moreover, that no judge could plausibly or consistently justify such a conclusion underscores the following crucial point. It does not follow from the recognition that legal texts do not speak for themselves, that they may be construed as meaning virtually anything, or that all plausible meanings are likely to be equivalent.

There are several reasons why interpretation along the lines suggested immediately above ought to be rejected. First, nothing in Article II suggests that "legislature" should be understood as an entity extracted from its customary context. While this suggests that there is no positive reason for leaning in favor of the interpretation under consideration, it does not provide a negative reason for rejecting such an interpretation outright. Such a negative reason is readily available, however, and it can be established through application of a well-established rule of constitutional interpretation.

The rule in question holds that whenever constitutional text may be plausibly interpreted in more than one way, preference should be given to those interpretations that are more likely to uphold the consistency of the Constitution (or of one of its articles or clauses as the case may be) taken as a whole. In other words, the Constitution should be construed as an internally coherent document. In particular, seemingly small differences in formulation should not be considered as superfluous, as the Constitution's drafters are not to be assumed to have used language gratuitously. For example, in Article I the Constitution authorizes exercise of powers that are characterized as "necessary" in one place and as "absolutely necessary" in another—a difference noted by the U.S. Supreme Court in its interpretation of the scope of the powers that qualify as "necessary."^[46] Along similar lines, in contrast to Article II of the Constitution, which provides for appointment of presidential electors as the state legislature "may direct," Article I, section 3, provides that the two senators form each state shall "be chosen by the legislature thereof." Although this latter language was superseded in 1913 by the Seventeenth Amendment, which provides for direct elections of senators, while it was in force it endowed state legislatures with a federal power that was completely independent from their powers under the constitution of their own state. Indeed, even if the state constitution were silent on, or worse, against, the power of the legislature to choose senators, such a legislature would still have it as a matter of federal constitutional law. The power to choose senators implied, therefore, that if necessary state legislatures could operate independently from, or in contradiction to, the state constitutional order in which they were otherwise

---

embedded. In contrast, appointment of presidential electors as the state legislature "may direct" under Article II does not require powers outside, or in contradiction to, the state constitution, or at the very least does not require such powers if the legislature, such as Florida's did before the 2000 election, already "directed" a mode of appointment that is consistent with state constitutional constraints. Moreover, this mode of appointment embodied in the Florida Election Code was meant to be carried out like all other legislation within a state constitutional framework that contemplated recourse to the state's judiciary as a matter of course.

In the end, therefore, the concurring justices not only sharply departed from what was—and would shortly become again—their radical devotion to the vindication of seemingly ever-expanding conceptions of state rights, but they also embraced for the occasion a path of constitutional interpretation that even their most liberal colleagues would have had a very difficult time trying to explain.

VI

The 2000 Florida presidential election and the U.S. Supreme Court decision in Bush v. Gore reveal serious flaws in the American Constitution, in American democracy, and in the Court's handling of a politically charged case in which the justices did not (and perhaps could not) dispel the perception that they were ultimately acting out of partisan self-interest. Indeed, according to a report in Newsweek, Justice O'Connor and her husband expressed annoyance when on the day of the election the television networks initially projected Gore the winner in Florida and thus had him well on his way to the presidency. O'Connor told friends that she was disappointed because she wanted to resign from the Court and go back to her native Arizona but that the expected Gore victory would force her to postpone her plans as she did not want a Democrat to nominate her successor.^[47] There had also been much talk about the chief justice wanting to resign but preferring that his successor be nominated by a Republican. It also seems reasonable to surmise that Justices Scalia and Thomas would be interested in consolidating their powers on the Court, a prospect that would be much brighter with Bush as president, particularly in view of his announcement during the campaign that he would nominate to the court persons in these two justices' very image. Finally, it also seemed in the interest of Justice Kennedy, the fifth justice in the series of 5–4 decisions expanding states' rights, to have a Republican president who would be more likely to perpetuate or enlarge the Court's existing majority. In short, on various levels ranging from the personal to the institutional, the five justices who put an end to the election had much to gain from a Bush presidency.

---

In spite of the problems with the Constitution, democracy, and the Court, it would be possible to institute many worthy improvements by the next presidential election. Unfortunately, six months after the 2000 election, very little has been done (except in Florida and Georgia where significant reforms have been enacted), and many proposals generated shortly after the election debacle are at this writing languishing unattended.^[48] Significantly, in his first presidential budget sent to Congress, Bush has not asked for any funds for electoral reforms. Although, as former President Carter has observed, voter registration is far superior in Peru than in the United States, not much is likely to be done about this as all reform proposals tend to become hostage to partisan bickering among Democrats and Republicans.^[49]

Little is likely to be done, though much could be accomplished through fairly simple and straightforward reform. From the standpoint of the Constitution, the two flaws that ought to be corrected are the lack of a constitutional right to vote for president and the existence and role of the Electoral College. The first of these could be overcome through a constitutional amendment, which, like the Seventeenth Amendment did for elections to the Senate, would grant the right to vote for president and vice president directly to the citizens rather than "in such manner as the [state] Legislature … may direct." Since all states already grant their citizens the right to vote for president, this proposed amendment should not be controversial, and it seems fair to assume that if proposed it would ultimately succeed. Although citizens already vote, this amendment would not be superfluous. For one thing, it would enshrine the right to vote for president in the Constitution; for another, it would make it impossible for state legislatures to take matters into their own hands after elections as the Florida legislature began doing before the December 12 U.S. Supreme Court decision.

The elimination of the Electoral College, however, is bound to be much more controversial. Also, because small states would lose power if the Electoral College were abolished and since constitutional amendments must be ratified by the legislatures of (or conventions in) three quarters of the states, there is virtually no chance of abolishing the Electoral College anytime soon. Nevertheless, the arguments for abolishing it are quite compelling. Besides fostering the possibility of electing the loser of the popular vote president—which tends to undermine democratic legitimacy—the electoral vote system produces other pathologies. Thus, for example, the three largest states, California, Texas, and New York, were virtually left out of the 2000 presidential campaign as Bush enjoyed an insurmountable lead in Texas, Gore in the other two states. In contrast, toss-up states got an inordinate amount of attention, prompting an extraordinary focus on local issues affecting but a tiny percentage of the electorate and leading candidates to take inconsistent positions throughout the country as they adjusted

---

their promises to fit the needs of different localities. For example, in Michigan, where the election was considered a toss-up, Bush and Gore concentrated much effort on wooing certain blue-collar workers interested in strong unions and free guns. Gore was vulnerable to the extent that he was perceived as in favor of gun control, Bush to the extent that he was viewed as antiunion. Both, therefore, had an incentive to distort or downplay some of their positions to increase their appeal to this seemingly crucial constituency. But by carrying on a series of local campaigns rather than one national one, the candidates are subjected to great temptations to change their positions as they travel through the country, thus increasing public cynicism about their veracity and integrity.

Some argue that elimination of the Electoral College would level an unacceptable blow against federalism. So long as one agrees that the president should represent all Americans rather than the states that have combined to give him or her a majority of presidential electors, however, this argument is completely unpersuasive. Although Bush lost the popular vote, he has already acted in ways that appear designed to reward the states that voted for him and punish those that did not. For example, he has failed to take steps to alleviate California's energy crisis or to adjust his tax cut proposals to address a problem they would cause primarily in the northeastern states, which almost unanimously voted for Gore. To the extent that states can be punished for having voted for the loser in a presidential election, reliance on the Electoral College can clearly have an adverse effect on the operation of democracy on a national scale.

A truly national election for president would not compromise federalism. The interests of the states in the federal government are most directly represented in the Senate but also to a large extent in the House of Representatives. Indeed, although members of the House are supposed to represent the people rather than the states, every House district is encompassed within a single state, thus allowing voters to concentrate on state issues and concerns. Furthermore, issues of federalism, particularly those dealing with increasing or decreasing state powers vis-à-vis national ones can properly be made part of presidential campaigns. Accordingly, candidates for president could compete on how much power ought to be left to the states without having to pit one state against another.

The two constitutional changes suggested above would improve democracy without threatening federalism. Moreover, if they were adopted, the Twelfth Amendment, which provides for Congress to resolve disputed elections, would have to be replaced with a provision that would be consistent with the suggested changes and that would reinforce Congress's role in settling election disputes to the exclusion of the Court. There are two principal reasons for favoring Congress over the Court: (1) Congress is democratically

---

accountable, so that the electorate can refuse to reelect those representatives whom it feels did not deal with the election dispute fairly; and (2) the Court depends on the president for nominations of new justices, which makes decisions like Bush v. Gore inherently undesirable.

Nationalizing presidential elections by developing uniform voting conditions and uniform vote counting procedures throughout the land would also improve democracy, by replacing the current array of local countylevel standards by a single uniform standard applicable to all. Because of the improbability of eliminating the Electoral College, however, it may be necessary to set our sights lower and to seek improvements to democracy at the statewide level rather than at the national one. Ironically, in this respect the Court's otherwise ill-advised equal protection holding in Bush v. Gore may yet prove quite useful. Indeed, if the majority's efforts at suppressing its precedential value fail, then the equal protection rights first announced in Bush v. Gore could usefully be involved to give substance to equal opportunity to vote and to ensure granting an equal weight to all votes on a statewide basis. At least then equality of voting could be guaranteed within states even if not among them.

Finally, because of its blatant inconsistencies, pervasive lack of integrity, and inescapable airs of unmitigated partisanship, the Court's decision in Bush v. Gore forces us to rethink whether judicial decisions can ever be held to fair, acceptable, and feasible standards. Paradoxically, because its many vices are so obvious, the Court's decision in Bush v. Gore serves to underscore that in spite of inevitable imperfections and of the impossibility to rise above all politics, there are significant differences between principled and unprincipled decisions. Indeed, the above analysis demonstrates that it does not have to follow from the fact that legal texts do not have "plain meanings" or that human beings cannot ascend above politics that all judicial opinions are ultimately arbitrary and subjective. Judicial interpretation may be imperfect, but so long as it can be achieved with consistency and integrity, it may be political or even liberal or conservative without becoming illegitimate. What Bush v. Gore demonstrates is that the important line to draw is that between the purely political, the partisan, and the willingness to sacrifice consistency and integrity to secure a particular result, on the one hand, and the politically situated, morally committed, and historically confined judge who labors as a fallible human being with a steadfast resolve to upholding consistency and integrity, on the other. And if this line is clearly kept in mind, then Bush v. Gore may eventually be credited for having provided a salutary lesson. Perhaps the best we can expect from our judges is to render decisions that slow down political strife and that blunt its rougher edges. The majority opinion in Bush v. Gore did the exact opposite.

---

NOTES

I wish to thank Robert Bennett, Larry Kramer, and Rick Pildes for helpful comments.

---

7. THE UNBEARABLE RIGHTNESS OF
BUSH V. GORE

Nelson Lund

Bush v. Gore was a straightforward and legally correct decision. If one were familiar only with the commentary that ensued in the decision's wake, this claim might sound almost lunatic. This essay explains why the Supreme Court acted properly, indeed admirably, and why the ubiquitous criticisms that have been leveled at the justices from both the Left and the Right are at best misguided.

For almost forty years, the Supreme Court has treated the stuffing of ballot boxes as a paradigmatic violation of the Equal Protection Clause. Much subtler and more indirect forms of vote dilution have also been outlawed. Like some of those practices, the selective and partial recount ordered by the Florida Supreme Court may have been an inadvertent form of vote dilution. But that recount had effects that were virtually indistinguishable from those in the paradigmatic case. There is no meaningful difference between adding illegal votes to the count and selectively adding legal votes, which is what the Florida court was doing. The Supreme Court rightly concluded that the vote dilution in this case violated well-established equal protection principles.

Nor did the Supreme Court err in its response to this constitutional violation. Although the Court acted with unprecedented dispatch after the Florida court's December 8, 2000, decision, it was highly improbable that a legally proper recount could be conducted by the December 18 deadline set by federal law. And it was quite impossible for such a recount to meet the December 12 deadline that the Florida court itself had found in Florida law. Contrary to a widespread misconception, the U.S. Supreme Court properly accepted the Florida court's interpretation of state law and provided that court with an opportunity to reconsider its own interpretation

---

of state law. When the clock ran out, it was entirely due to mistakes and delays attributable to the Florida court.

The least known passages in Bush v. Gore are those in which the dissenters explain why the majority's legal analysis was erroneous. These passages are not well known because they do not exist. The best known passage, which comes from Justice Stevens's dissent, consists of a rhetorical flourish rather than an analysis:

What must underlie [George W. Bush's and the other defendants'] entire federal assault on the Florida election procedures is an unstated lack of confidence in the impartiality and capacity of the state judges who would make the critical decisions if the vote count were to proceed. Otherwise, their position is wholly without merit. The endorsement of that position by the majority of this Court can only lend credence to the most cynical appraisal of the work of judges throughout the land. It is confidence in the men and women who administer the judicial system that is the true backbone of the rule of law. Time will one day heal the wound to that confidence that will be inflicted by today's decision. One thing, however, is certain. Although we may never know with complete certainty the identity of the winner of this year's Presidential election, the identity of the loser is perfectly clear. It is the Nation's confidence in the judge as an impartial guardian of the rule of law.^[1]

This passage became famous because it has been read to mean that Stevens was impugning the integrity of the five justices who joined the majority opinion.^[2] But if Stevens was slyly encouraging this interpretation, he was careful not to say or imply any such thing. Indeed, if we take his statement at face value, Stevens's point is almost the opposite: cynical appraisals of the work of judges—any judges—are a threat to the rule of law.

Justice Stevens may well have been correct that the real loser in the 2000 election was "the Nation's confidence in the judge as an impartial guardian of the rule of law." Partisans on both sides accused judges of manipulating the law in order to assist the candidate they favored, and aspersions were cast on the integrity of some judges even before they ruled. For the vast majority of observers who lacked the time or expertise to form an independent judgment, it must have seemed unlikely that all the judges involved had behaved impartially. And many Americans may well have quietly concluded that they are all just a bunch of political hacks in robes.

That conclusion would be a mistake, if for no other reason than the impossibility of proving or disproving such charges. Justice Stevens, however, offered a very different reason for worrying about the reputations of the Florida judges: "It is confidence in the men and women who administer the judicial system that is the true backbone of the rule of law." A thoughtful citizen unschooled in legal folkways might regard this as a very odd notion. This citizen might suppose that the true backbone of the rule of law is

---

actual adherence to the law by the men and women who administer the judicial system. And the citizen might then suppose that refusals by judges to adhere to the law should be exposed and corrected. In fact, that is all Bush's lawyers asked for, and that is all the Supreme Court gave them.

Odd as Stevens's statement might seem to an ordinary citizen, it is quite consistent with a theory deeply embedded among sophisticated legal elites but seldom advocated in popular discourse. That theory essentially holds that the law is what the judges say it is, so that an aura of impartiality around judges would serve mainly to help them impose better laws on the nation than the people are willing to enact through their legislatures. This theory was rejected by the Bush v. Gore majority, which took an approach much more closely aligned with the ordinary citizen's view. That rejection provides the most important reason for defending the Court's decision, and I take up that defense here.

A thoughtful citizen's attitude is a useful tool when thinking about this case, but it will not be enough. Judges are expected to apply the law set out in the Constitution and statutes. For two main reasons, however, that law often cannot be applied with the same certainty offered, say, by the laws of algebra. First, many provisions of the law are ambiguous, which means that judgment has to be used in choosing among a range of possible interpretations. Second, our legal system has adopted a practice in which courts adhere to their previous decisions even when they have reason to believe that those decisions were wrong. This rule of stare decisis, however, is not completely inflexible, and courts must therefore also exercise judgment in applying it.

Although this means that colorable arguments can often be made on both sides of the legal questions with which courts are confronted, it does not mean that the distinction between legally right and legally wrong answers is a chimera. Nor does it mean, as the fashionable academic theories would have it, that judicial decisions should be judged on the basis of their political effects rather than their fidelity to the law. If there are some close legal questions, as there are, there are also such things as stronger and weaker legal arguments. And if there are some legal questions with no indubitably clear answers, as there are, there are also some questions that have right and wrong legal answers.

Simply stated, my claim is that Bush v. Gore should be evaluated as a legal decision and that it stands up very well when judged by appropriate legal standards. Conversely, whatever motivated the Florida judges whose decision was reversed, their ruling was indefensible as a legal decision. The criticisms that can most plausibly be leveled against the Supreme Court majority are essentially political criticisms of a kind that might more fittingly be directed against a Senate majority leader or an ambassador to China. Justice Stevens's rhetorical flight, in which the rule of law becomes conflated

---

with confidence in its supposed guardians, is one example. I say that the Florida Supreme Court grossly violated the law and that the U.S. Supreme Court properly acted to stop the travesty. That decision was the legally correct decision, and political criticisms of the Court are based on the corrosively sophisticated assumption that the justices cannot be anything except politicians in robes.

OVERVIEW OF THE FACTUAL SETTING

An extraordinary confluence of events presented the American judicial system with a genuinely difficult challenge in the aftermath of the voting that took place on November 7, 2000. The decisive election in Florida was so excruciatingly close that certainty about the outcome could not have been achieved under the best of conditions. In the final official counts of Florida's ballots, the difference between Bush and Gore was only 537 votes out of some 6 million, which is less than one one-hundredth of 1 percent. Even if there were some unerring and unambiguously correct way to tabulate the ballots, which there probably is not, certainty would still have eluded us because we would not know how many ballots were cast by ineligible or nonexistent voters. In an election this close, with such a large number of ballots cast, only God can know who "really" won.

As a practical and legal matter, this would not have mattered very much if there had been some clear and agreed-upon rules for determining which ballots to count for which candidates and how to tabulate the results. But there were not. The Florida statutes governing election disputes had apparently been drafted with local rather than statewide elections primarily in mind, and without considering the unique time constraints that federal law imposes on the resolution of disputes about the electoral college. This should not be too surprising. Statewide elections in a jurisdiction as populous as Florida have rarely if ever been close enough to have their outcomes turn on an interpretation of the rules for counting ballots. And who would have thought that this unlikely contingency would ever be compounded into freakishness by coming to pass in a state whose electoral votes were going to make the difference in a presidential election? When the freak event did occur, it turned out that the statutes drafted with local elections in mind did not fit a statewide election dispute very easily.

On top of everything else, federal law had its own share of problems and uncertainties. In the wake of the notorious Hayes-Tilden contest in 1876, Congress had enacted a number of provisions aimed at avoiding another such disorderly mess.^[3] But the meaning of these provisions, and their relationship to even older statutes, was not entirely clear. On the books for more than a century, they had never been tested in practice or in the courts. Meanwhile, the twentieth century had witnessed the independent development

---

of a complex and evolving body of constitutional election law in the federal courts.

Thus when George Bush and Al Gore entered what looked at times like mortal legal combat, there were lots of weapons scattered around the arena. Because Bush had more votes tallied in the initial count and in the automatic recount that Florida law provided for close elections, Gore could win only by attacking the official vote counts. Correlatively, Bush's self-interest dictated that he defend those same counts. Whether from pure self-interest or not, both candidates skillfully and relentlessly deployed all their legal weapons in a fight for victory.

Without blaming either candidate for his litigation strategy, one can note that this created an unusual problem for the courts. A great many novel legal issues were raised in a large number of lawsuits filed by the candidates and their supporters. Furthermore, unlike most other election disputes, those involving the Electoral College must be resolved very quickly. Bush v. Gore itself was probably decided faster than any comparably important decision in history,^[4] and it came at the end of a series of judgments that had themselves been made in unusual haste. What is perhaps most remarkable about the Supreme Court's opinion is how easily defensible it is.

OVERVIEW OF THE LEGAL SETTING

A full analysis, or even an adequate summary, of the legal disputes that set the stage for the decision in Bush v. Gore is beyond the scope of this essay. I therefore offer only the barest essentials here.

The Constitution requires each state to appoint, "in such Manner as the Legislature thereof may direct," a number of presidential electors equal to the size of the state's congressional delegation.^[5] The Constitution also requires these electors to meet in their states to cast their ballots and then to send a certified list of the votes to the president of the Senate (who in this case was Al Gore).^[6] He is required to open them in the presence of the Senate and House of Representatives, where they are then counted.^[7] An absolute majority of the "Electors appointed" is needed to win the election.^[8] Absent such a majority, the House of Representatives chooses the president under a rule that gives each state's delegation one vote.^[9] If no president is chosen by January 20, an acting president takes office "until a President shall have qualified."^[10]

Congress has attempted to fill in some of the details that are left unspecified by the Constitution. Two of those statutes are especially relevant. First, federal law required that presidential electors meet and give their votes on December 18, 2000.^[11] Second, federal law provided that if a state had enacted laws for resolving disputes before November 7, 2000, and used

---

those laws to resolve election disputes by December 12, 2000, such resolution would be treated as conclusive when the votes were counted in Congress.^[12]

Florida election law is much more complex and detailed. The procedure for dealing with disputes has four main elements. First, an automatic statewide recount is conducted in close elections.^[13] Second, a "protest" period occurs, during which certain kinds of challenges can be brought before county canvassing boards (which comprise two local elected officials and one local judge).^[14] Third, state officials accept election results from these county officials and "certify" a winner.^[15] Fourth, that certification can be challenged in court during the "contest" period.^[16]

THE LITIGATION BEGINS

Bush won the initial count by 1,784 votes, and he was still ahead by 327 votes after the automatic statewide machine recount.^[17] Gore then filed protests,^[18] demanding a hand recount of the ballots in four heavily Democratic counties, only three of which are relevant to the following discussion: Broward, Palm Beach, and Miami-Dade.^[19] Gore apparently chose these counties for one or both of two reasons.^[20] First, to the extent that errors by the counting machines were randomly distributed, Gore could expect to be a net gainer in these most heavily Democratic jurisdictions.^[21] Second, the hand recounts would be supervised by local elected officials, and the chances that such officials would be biased in Gore's favor (or at least not biased in Bush's favor) would be highest in the most heavily Democratic counties.^[22]

Gore's strategy was consistent with the letter of Florida law, at least in the sense that it permitted Gore to request recounts in selected counties, but it raised serious constitutional questions that had lurked unnoticed so long as the law had been applied only to local elections. If the law actually allowed one candidate to obtain a geographically biased recount in a statewide election, the Florida statute may have unconstitutionally (albeit inadvertently) run afoul of established principles requiring the fair and equal treatment of similarly situated voters. Accordingly, Bush promptly filed a lawsuit in federal court, in which he sought to stop the recounts that Gore had demanded.

The courts never addressed the merits of Bush's arguments, concluding instead that the relief he sought was premature.^[23] What proved to be the decisive litigation resulted instead from lawsuits brought by Gore in an attempt to overcome a series of obstacles in state law that threatened to frustrate his chosen strategy.

The first obstacle was a statutory provision requiring that the local officials provide a final tally to the secretary of state within seven days after the

---

election, whereupon the statewide result would be "certified." None of the three counties had finished their hand recounts by that deadline, and Secretary of State Katherine Harris concluded that they had not offered legally sufficient reasons for any further delay. In response to a lawsuit filed by Gore and others, a Florida court rejected Gore's claim that Harris had behaved illegally by refusing to accept tardy recounts. Gore then appealed to the Florida Supreme Court, which took the first of many highly questionable actions. Without being asked to do so by any litigant, the court issued an unexplained order forbidding state officials from "certifying" the results of the election.

This order, to which Gore raised no objection, had the practical effect of artificially extending the protest phase (in which preliminary decisions are made by local elected officials) and therefore necessarily shortening the contest phase of the legal process (in which final decisions are made by courts). As we shall see, the shortening of the contest period had fateful consequences.

BUSH V. PALM BEACH COUNTY CANVASSING BOARD ("BUSH I")

The unprompted decision by the Florida Supreme Court was the first in a series that culminated in the U.S. Supreme Court's decision that the Florida recount was being conducted in an unconstitutional manner. The next step was the Florida Supreme Court's decision, four days later, to reverse the trial court, thereby overturning two decisions made by the secretary of state, who had concluded (1) that manual recounts were legally available only to correct errors made by the voting or counting machines, not errors by voters; and (2) that conducting recounts based on voters' errors did not justify relaxing the statutory deadline for the counties to report their election returns.

The Florida court sought to justify its decision by resolving what it identified as three troublesome or ambiguous features of the state election statute. First, the statute allowed full manual recounts only to correct an "error in the vote tabulation," without specifying whether this would include a failure by the voter to mark or punch a ballot in the manner required to render the ballot machine-readable. Second, one statutory provision says that the secretary of state "shall" ignore late returns from the counties, while another provision says that she "may" ignore late returns. Third, one statute allows a candidate to request a recount at any time before the county returns are certified, while another requires the county officials to certify the returns within seven days; thus, cases might arise in which a candidate requested a recount just before the seventh day, leaving no time to conduct the recount.

The Florida court believed that it should resolve these issues so as to

---

facilitate the right of the voters to express their will, without abiding by "[t]echnical statutory requirements."^[24] The court rejected Secretary of State Harris's argument that an "error in vote tabulation" could only refer to malfunctioning machines, concluding instead that the statutory language also referred to cases in which a ballot was punched or marked in such a way that a human, but not a machine, could detect an intent to vote for a particular candidate.

For two reasons, this was the court's most important decision. First, the disputes about the deadline only became relevant on the assumption that there was a legal basis for the recounts in the first place. Second, and most important, it was this interpretation of the statutes that made Gore's cherrypicking strategy feasible, thus raising serious constitutional questions about those statutes.

The decision was also more far-fetched than it may at first appear. According to the court's interpretation, machine tabulations will always be erroneous if any voter failed to follow the instructions for marking the ballot, which always happens. Why then would the statutes provide for an automatic machine recount in close elections? Such a procedure would almost always be pointless because a hand recount to correct these inherently erroneous machine recounts would always be justified. It should therefore come as no surprise that recounts had never before been conducted to correct voters' errors.^[25]

The court's conclusions with respect to the other two issues were similarly implausible. The Florida court resolved the apparent conflict between the "shall ignore" and "may ignore" provisions by inventing a new meaning inconsistent with them both, namely, that the secretary of state may not ignore late returns. The court then went on to give the counties an entirely new deadline of nineteen days after the election. This deadline had no basis anywhere in the statutes, and it was adopted without any explanation except a vague allusion to "the equitable powers of this Court."^[26] The justification given for these conclusions was that "the will of the electors supersedes any technical statutory requirements."^[27]

Bush sought review in the U.S. Supreme Court, arguing that the Florida court had simply disregarded the statutes, thus violating the Constitution's command in Article II that electors be chosen "in such Manner as the Legislature [of the State] may direct." What the Florida court had done, Bush argued, was not to interpret the statutes but to rewrite them, in contravention of the U.S. Constitution.^[28]

Bush's argument presented the U.S. Supreme Court with a genuinely difficult question. The Constitution plainly says that the directions of the state legislature must be followed, and the Florida court was pretty plainly not following the legislature's directions. However, the decisions of state supreme courts are almost always treated as authoritative interpretations

---

of state law, no matter how implausible they may seem. It thus would not have been altogether unthinkable to assume that the Florida legislature had implicitly "directed" that electors be chosen in accordance with Florida law as interpreted by the Florida courts.

A unanimous Supreme Court avoided this difficult Article II question, and rightly so. In resolving what it saw as the troublesome features of the statutory scheme, the Florida court had appeared to rely in part on the notion that statutes should be interpreted so as to render them consistent with its own prior interpretation of the Florida Constitution, according to which "[u]nreasonable or unnecessary restraints on the elective process are prohibited."^[29] This would have seemed a normal approach to resolving a state law question because a state constitution has greater authority than state statutes.^[30] In this case, however, that seemingly normal approach may have been misplaced. McPherson v. Blacker, an 1892 Supreme Court case apparently overlooked by the Florida Supreme Court, had suggested (without deciding) that state constitutions are not authorized to constrain state legislatures in the special context of choosing presidential electors.^[31]

Thus the Supreme Court was confronted with a double uncertainty. First, it had previously said, but had not actually decided, that a very unusual relationship exists between state constitutions and state statutes in the context of selecting presidential electors. Second, the Florida Supreme Court had not made it clear that its construction of the state statutes was crucially dependent on the Florida constitution. If the Florida court were given a chance to construe the state statutes without reference to the state constitution, then the U.S. Supreme Court might not have to decide whether to adopt the suggestion made in McPherson. Accordingly, the Supreme Court vacated the decision and remanded the case so that the Florida court could clarify or reconsider its ruling.^[32]

In the course of its opinion, the Supreme Court also cautioned the Florida court to give attention to a federal statute that it had previously ignored. That statute, 3 U.S.C. § 5, provided that if a state resolved any election disputes by December 12, 2000, using laws in place before November 7, 2000, such resolution would be treated as conclusive when the votes were counted in Congress. The Supreme Court noted that "a legislative wish to take advantage of the ‘safe harbor’ [offered by this federal statute] would counsel against any construction of the [Florida] Election Code that Congress might deem to be a change in the law."^[33]

It certainly seems reasonable to suppose that Florida's legislature would want to take advantage of this safe harbor. And it might make sense to resolve statutory ambiguities so as to bring the state within the safe harbor, though there is no evidence that anyone in the Florida legislature had ever heard of 3 U.S.C.§5 before the 2000 election. But what did any of this have to do with the case before the Court? The fact is that it had no relevance

---

at all unless the U.S. Constitution required the Florida court to give effect to such a "legislative wish." As we have seen, however, the Supreme Court had never decided that the Constitution does impose this requirement, and the Court had taken pains to avoid deciding the question in this case.

Thus the message that the unanimous Court was sending to the Florida judges should have been quite clear: We are not anxious to decide difficult questions of federal constitutional law without giving you an opportunity to address those questions first. But you had better take federal law much more seriously than you did in your first opinion.

THE FLORIDA COURT CAREENS OUT OF CONTROL

Clear as it was, either the Supreme Court's message did not reach a majority of the Florida judges or they decided they could safely ignore it. Whatever the cause, those judges soon embarked on an extraordinary journey outside the bounds of federal law. To appreciate the necessity and the restraint of the Supreme Court's controversial decision in Bush v. Gore, one must first understand the sheer outlandishness of the Florida decision that provoked it.

Before coming to that, however, I need to summarize a few more facts. Florida, like many other states, has a decentralized system for conducting elections. Each of Florida's sixty-seven counties conducts elections under the supervision of local officials. These officials are bound by a number of rules established by state law, but many details are left to their discretion. Different counties, for example, have used different kinds of voting machines, and the counties have not been told by state law exactly what rules to use when conducting hand recounts.

As the whole nation learned in 2000, there is room for considerable debate about the proper way to classify punch-card ballots during a manual review. Without reviewing the intricacies of the controversies over matters such as hanging and dimpled chads, it should be enough to note that the three counties chosen by Gore for his protests used different standards of review and that one county actually changed its standard repeatedly during the recount. Although Democratic officials controlled all of the recounts, and despite the extra twelve days that the Florida Supreme Court had created for the recount process, Bush remained ahead when the new deadline arrived.

Accordingly, state officials "certified" Bush as the winner of the election, by a margin of 537 votes. Gore then invoked the contest provisions of state law, filing a lawsuit challenging this certification. To prevail in this suit, the statute required Gore to begin by proving the "receipt of a number of illegal votes or rejection of a number of legal votes sufficient to change or place

---

in doubt the result of the election."^[34] Gore claimed that he could do this, primarily on the basis of the following claims: (1) even though Palm Beach County had not completed its recount by the extended deadline and even though Miami-Dade had not completed its recount at all, Gore should be credited with the net gains he had thus far made in those counties (215 and 168 votes, respectively); and (2) more important, some 9,000 Miami-Dade "undervote" ballots—those on which the machine did not detect a choice for any candidate for the office of president—would change the result of the election if reexamined by hand.^[35]

After a trial at which Gore had the opportunity to establish his claims, Judge N. Sanders Sauls found that he had failed to do so. The most important element in Sauls's reasoning was that Gore had offered "no credible statistical evidence and no other competent substantial evidence" to establish that the certified result of the statewide election would be changed if further scrutiny of the Miami-Dade ballots were undertaken.^[36]

The fundamental difficulty confronting Gore was this: He had demanded a manual recount only in selected counties, and that demand was manifestly calculated to produce a shift in the statewide totals on the basis of chance alone. But even after the Florida Supreme Court had created an extra twelve days for Gore to pursue this constitutionally dubious strategy, the result of the election had remained the same. Thus Gore appeared to be locked into a losing game, even assuming that his strategy was permissible under federal law.

On December 8, 2000, however, the Florida Supreme Court created a whole new theory under which Gore might be able to get the outcome of the election changed. Reversing the trial court's decision by a vote of 4–3, the majority ordered the trial court to take the following actions:

• Add a net of 215 votes (or perhaps 176, depending on a factual issue
• that the judges did not resolve) to Gore's total, based on the Palm
• Beach recount, whose results were not reported to state officials within the judicially extended protest period.
• Add a net of 168 votes for Gore to the officially certified vote totals, based on the incomplete recount conducted by local election officials in Miami-Dade County.
• Conduct a manual recount of the 9,000 Miami-Dade ballots that Gore claimed might shift the statewide totals in his favor.
• Conduct a statewide recount of some kind, which the Florida Supreme Court strongly suggested should be limited to a recount of the undervote ballots in each county.^[37]

This was a truly bizarre ruling. First, it ignored the legal effect of the U.S. Supreme Court's decision in Bush I. On December 4 the Supreme Court had vacated the Florida court's November 21 decision in that case, thus

---

rendering it a legal nullity. The Florida court's December 8 decision, however, appeared to assume the validity of the nullified decision because it ordered additions to Gore's vote total that had been made possible only by the November 21 decision's extension of the statutory protest period.^[38]

Second, the statutory interpretation underlying the December 8 decision was even more questionable than that on which the November 21 decision rested. Florida's contest statute required Gore to prove the existence of errors sufficient to change or place in doubt the outcome of the election.^[39] The only evidence he had was the existence of some 9,000 undervote ballots that the Miami-Dade officials had found it impracticable to examine during the protest period.^[40] The court held that the mere existence of these ballots was sufficient to place the outcome of the statewide election in doubt, even though Gore had not proved that a recount of these ballots would even favor him.^[41] The assumption here seemed to be that in a very close election, almost anything could put the outcome in doubt.^[42] That has a certain plausibility, but the court also held, in a bizarre reversal of logic, that the statute did not require a recount of all Miami-Dade ballots (let alone all ballots statewide) because Gore had only put these 9,000 at issue.^[43] The absurdity of putting these two conclusions together was apparently obvious to the court itself, for it then spun off in a different direction, concluding without explanation that a recount could not be confined to Miami-Dade, though it could be confined to undervote ballots.^[44] How the court got this conglomeration of conclusions out of the statute is anyone's guess.

Thus the effect of this ruling by the Florida court was to raise exactly the same difficult constitutional question that the U.S. Supreme Court had carefully avoided in the Bush I case, namely, whether a state court's interpretation of state statutes can be so clearly untenable that it constitutes an impermissible departure from the legislative directions referenced in Article II of the U.S. Constitution.

Third, the court ordered the addition of 168 votes to Gore's certified totals, based on the partial recount in Miami-Dade. This order is worth pausing over because it is truly shocking. Whatever rationale one might use to justify conducting recounts in some jurisdictions but not others, stopping in the middle of a recount and definitively awarding one candidate the number of new votes he had picked up by that point simply defies explanation in terms of an effort to produce a more accurate count of the votes. What is worse, there was unrebutted evidence at trial that Miami-Dade had begun its recount with the most heavily Democratic precincts, which means that the partial recount was obviously biased in Gore's favor.^[45]

Fourth, the statewide remedy of reexamining undervote ballots had not been requested by any of the parties, it had no source in the Florida statutes, and the court provided no meaningful instructions for conducting

---

it.^[46] The court's decision, moreover, came only four days before the federal safe harbor deadline that was pointedly discussed in the U.S. Supreme Court's Bush I opinion. It was perfectly obvious, as the three Florida Supreme Court dissenters insisted, that the majority was "departing from the essential requirements of the law by providing a remedy which is impossible to achieve and which will ultimately lead to chaos."^[47] And, as Chief Justice Wells pointed out in his dissent, the lawlessness was so obvious that it seemed likely to "eventually cause the election results in Florida to be stricken by the federal courts or Congress."^[48]

What could have caused the majority to take this reckless action? Leaving cynical hypotheses aside and looking only at the justification offered by the Florida judges themselves, it turns out that their reasoning actually contradicted their actions. The majority purported to adopt what it called a "common sense" approach to the statute, summed up in the notion that the outcome of elections should be determined by "the will of the voters" rather than by "strategies extraneous to the voting process."^[49] The actual ruling, however, was based on a very different theory, which was never stated in the opinion and which was essentially the opposite of the stated theory.

The real theory went something like this. Once the ballots have been counted by machine, we will allow the loser to choose which ballots to reexamine by hand. Any changes in the vote totals resulting from this selective and partial recount, such as the 168 votes in Miami-Dade, will be adopted. But because this would so manifestly allow the outcome to turn on "strategies extraneous to the voting process," we will try to create what we regard as a tolerable approximation of evenhandedness by directing the trial court to make an effort to perform a somewhat less selective and somewhat less partial recount than the loser had at first demanded.

If the Florida Supreme Court had actually been seeking to ascertain the "will of the voters" of Florida, it would have designed a statewide recount that could believably be called more accurate or more reliable than the initial machine counts. At an absolute minimum, that would have required reexamining all the overvotes (where the machines detected a vote for more than one candidate and therefore recorded no vote) as well as the undervotes (where the machines detected no vote for any candidate). Once one assumes that the "intent of the voter" should be honored even when the voter failed to comply with the instructions on how to vote, these two categories of ballots become logically indistinguishable.^[50]

Furthermore, the need to treat undervotes and overvotes the same way is only the most obvious requirement of a recount aimed at determining the will of the voters. If one were actually serious about designing a recount that was more accurate than the machine counts, one would also have to recount all 6 million "legal votes." Whatever criterion is adopted for changing undervotes to legal votes (the presence of a hanging chad or the presence

---

of a dimpled chad, for example), that same criterion should be applied to ballots containing both a machine-readable hole and a hanging or dimpled chad. That means that some "legal votes" would have to be changed to "overvotes" and thus deducted from the vote totals. This could be quite significant, for ballots containing both a clean hole for one candidate and a dimpled or indented chad for another candidate were quite common.^[51] Alternatively, the court might have been justified in restricting a recount to undervote ballots if it had employed a standard designed to count the ballots of those voters whose efforts were frustrated by faulty machines, without counting the ballots of voters who failed to follow the instructions. But the Florida court insisted on the proposition that "a legal vote is one in which there is a ‘clear indication of the intent of the voter,’ " with or without evidence of a faulty machine.^[52]

Thus the Florida Supreme Court could not have been seeking to ascertain the will of the voters of Florida. Instead, it was seeking to ascertain the will of a peculiar subset of Florida voters, namely, those who had cast undervote ballots and those other voters who happened to reside in the counties Gore had selected for full recounts. The court gave no explanation for this illogical choice.^[53]

Not only did the Florida Supreme Court focus the statewide recount on amanifestly inappropriate subset of the ballots, it did not even indicate that the statewide recount of undervotes would actually have to be completed in order for Gore to prevail in his challenge. What the majority apparently contemplated was that it would stop the recount at some point (December 12? December 18? January 6? January 20?) and declare a winner on the basis of whatever new vote totals existed at that time. Although the court did not announce this, it is the logical inference from the majority's decision definitively to award Gore the 168 votes he had already picked up in the uncompleted recount in Miami-Dade. If that uncompleted recount was enough to justify changing the official vote count, why couldn't a similarly uncompleted statewide recount be used to justify changing the outcome of the election? And why wasn't the world told in advance when the recounting would stop?

THE SUPREME COURT'S DECISION IN BUSH V. GORE

The day after this amazing decision by the Florida court, the U.S. Supreme Court voted to halt the statewide partial recount that the Florida judges had initiated and to schedule a full hearing two days later.^[54] On December 12, only four days after the Florida court's decision, the Supreme Court held that the recount violated the Equal Protection Clause of the Fourteenth Amendment to the U.S. Constitution, with only two justices dissenting from this conclusion.

---

The majority's equal protection analysis was quite straightforward and firmly grounded in precedent. After a brief summary of the Court's votedilution jurisprudence, the majority described several ways in which the Florida recount entailed the uneven treatment of different voters: (1) varying standards for determining a voter's intent had been employed; (2) the statewide recount had been limited to undervotes, while the recounts in the Gore-selected counties had included all ballots; (3) the partial recount in Miami-Dade had been used for certification, and the Florida court evidently contemplated the future use of partial recounts; and (4) the statewide recount was being conducted by untrained personnel, without an opportunity for observers to make contemporaneous objections. Without saying that any one of these features of the recount process would by itself have been legally fatal, the majority concluded that the process as a whole failed to satisfy "the minimum requirement for non-arbitrary treatment of voters necessary to secure the fundamental right" to vote.^[55]

Not a single one of the Court's dissenters made any effort to show that the Florida recount satisfied the minimum requirements of equal protection. This should be no surprise, for reasons that will become clear when we take a closer look at the Court's precedents.

Under well-known and long-established case law, the right to vote has been treated as a fundamental right that must be extended equally to all citizens.^[56] This means that state governments cannot deny the vote to any citizen without an extremely powerful justification. The Court has also held that "the right of suffrage can be denied by a debasement or dilution of the weight of a citizen's vote just as effectively as by wholly prohibiting the free exercise of the franchise."^[57]

The Court has held, for example, that the seats in state legislatures must be equally apportioned on a population basis;^[58] that statewide elections may not be conducted under a "county unit" system resembling the federal Electoral College;^[59] and that a state may not require that a nominating petition for presidential elector include the signatures of at least two hundred qualified voters from each of at least fifty counties.^[60] Faced with such rules, which effectively gave more "weight" to the votes of those living in rural or sparsely populated areas of a state than to those living in more densely populated areas, the Court declared:

[T]he weight of a citizen's vote cannot be made to depend on where he lives. Population is, of necessity, the starting point for consideration and the controlling criterion for judgment in legislative apportionment controversies. A citizen, a qualified voter, is no more nor no less so because he lives in the city or on the farm. This is the clear and strong command of our Constitution's Equal Protection Clause. This is an essential part of the concept of a government of laws and not men. This is at the heart of Lincoln's vision of "government of the people, by the people, (and) for the people."^[61]

---

The application of the vote-dilution principle is not confined to any particular class of voting rules: "Weighting the votes of citizens differently, by any method or means, merely because of where they happen to reside, hardly seems justifiable. One must be ever aware that the Constitution forbids sophisticated as well as simple-minded modes of discrimination."^[62] And "sophisticated" modes of discrimination include those that are unintentionally discriminatory.^[63]

In this case, the Florida court devised an extremely complex system of weighting, in which certain kinds of ballots were more likely to be counted as legal votes in some places than in others, thus discriminating for and against different groups of voters based on where they happened to reside. Most obviously, voters who cast overvote ballots in Broward, Palm Beach, and Miami-Dade Counties were treated more favorably than those who cast similar ballots elsewhere. Similarly, voters who cast dimpled chad ballots in Broward were treated more favorably than those who cast similar ballots in Palm Beach. Voters living in the unrecounted (and more Republican) precincts of Miami-Dade were disadvantaged in comparison with those living in the recounted (and more Democratic) precincts. The complexity of the vote dilution involved did not convert it into something other than vote dilution.

Before Bush v. Gore, geographic vote denial and dilution controversies had arisen primarily in two kinds of cases: (1) where imbalances arose because legislatures had failed to reapportion in response to population shifts;^[64] and (2) where discriminatory arrangements had been adopted deliberately in order to serve what legislatures thought were overriding purposes, such as to protect the influence of certain constituencies or to create districts whose boundaries would coincide with preexisting political or geographic borders.^[65] In these cases, the Court applied what is often called "strict scrutiny," which requires that any inequality or discrimination be justified by legitimate and compelling government purposes and that the inequality not extend farther than those purposes require.^[66]

Bush v. Gore did not involve the application of a preexisting rule that systematically discriminated against an identifiable class of voters, such as those residing in more sparsely populated jurisdictions. But nothing in the rationale underlying the vote-dilution cases limits it to such cases. Unconstitutional vote dilution has been found, for example, where there is no systematic discrimination against a class of voters with shared political interests.^[67] Furthermore, the rationale of the decisions implies, if anything, that the application of new and discriminatory rules after an election has been held should receive an especially skeptical review by the courts because after-the-fact manipulation of voting rules is especially prone to abuse.^[68]

Indeed, the Court has frequently used the stuffing of ballot boxes as a

---

paradigmatic example of an obvious constitutional violation.^[69] But any distinction between adding illegal ballots to the count and selectively adding legal ballots in a way that favors one candidate over another would be entirely sophistic. The Court long ago ruled out such sophistry when it declared that the Fourteenth Amendment "nullifies sophisticated as well as simple-minded modes of discrimination."^[70]

Vote dilution obviously occurs when illegal ballots are counted along with legal ballots. It also occurs when legal ballots are counted for one candidate but not the other. It occurs when ballots are counted only from precincts with a history of favoring one party over the other. And it occurs when a special effort is made to find previously overlooked legal ballots in arbitrarily chosen subcategories. Nor does it make any difference whether such vote dilution proceeds from partisan motives. Thus, for example, if a vote count were inadvertently inflated with illegal ballots and a court arbitrarily refused to correct the count, it would not matter whether the judge was dishonest or just mistaken about his or her obligations. Similarly, it makes no difference whether the Florida judges were trying to help Gore or were simply the victims of confused thinking.

The discrimination in the Florida recount was novel, complex, and subtle, which helps to explain why it was unprecedented. No legislature would ever adopt a recount process like the one adopted by the Florida court, and no court had ever done so either. Whether this uniquely bizarre procedure resulted from bad faith (which I do not assert) or from a misunderstanding of the law, it should not even survive rational-basis scrutiny, let alone the strict scrutiny that the Supreme Court has previously employed in vote-dilution cases.

To see why this simply was not a close or debatable case, it is important to remember that the recount process designed by the Florida court was a substitute for the standardized, machine counts on which the secretary of state had sought to rely. Although the machine counts were undoubtedly imperfect, there could be no legitimate, let alone compelling, interest in substituting hand recounts unless those recounts could reasonably have been expected to be more accurate as a whole than the machine recounts. The Florida Supreme Court never made any attempt to show that its recount procedure would likely be more accurate, and any such effort would have been laughable.

The only justification the Florida Supreme Court ever offered for its orders was that some new "legal votes" (i.e., ballots containing evidence of an "intent to vote" undetected by the counting machines) would turn up in the various partial manual recounts. The underlying theory was apparently that any "legal votes" that happened to turn up in any of these selective recounts should be added to the totals generated by the machine counts.

But this completely misses the point of the equal protection cases—that

---

substantially the same rules, whatever those rules are, must be applied to all voters and all ballots. Suppose, for example, that undervote ballots containing evidence of an intent to vote for Gore were changed to legal votes, but similar ballots showing an intent to vote for Bush were not changed to legal votes. Such a recount would be "better" than the machine counts under the criterion employed by the Florida court because it would result in more legal votes being tabulated. But it would not be better in any constitutionally relevant sense,^[71] or indeed under any sane criterion. The difference between this hypothetical and the actual order of the Florida court is only one of degree, and a very slight degree at that.

It is true, as the U.S. Supreme Court has always recognized, that the law cannot and does not require perfect equality in the treatment of all voters. Every law has different effects on different people, but that does not mean that all laws are unconstitutional. Similarly, voting procedures do not have identical effects on everyone, but that does not make them unconstitutional.

Rural voters, for example, must on average travel farther to their polling places than urban voters, but the Court has not required that election officials somehow correct this inequality. Nor would the Court permit a "correction" that entailed a more pronouncedly unequal effect, such as the creation of malapportioned districts that gave greater weight to the ballots of rural voters. Similarly, there may be latent forms of inequality associated with particular kinds of voting machines, or in the use of different kinds of machines in different counties. But it does not follow that such relatively minor and speculative inequality can permissibly be "corrected" with the kind of gross and palpable inequality that pervaded the Florida court's recount process.

It should therefore come as no surprise that not a single member of the U.S. Supreme Court actually defended the Florida court's recount process against the equal protection critique articulated by the majority. Two of the dissenters (Souter and Breyer) acknowledged that the recount process could not be defended against equal protection objections.^[72] Justice Stevens, who refused to find the Florida recount unconstitutional, offered nothing more than an utterly anodyne allusion to the need for "a little play in the joints" of the machinery of government.^[73] That maxim could be used to defend any vote-dilution scheme, including all of those that have been invalidated by the Supreme Court in the past. Justice Ginsburg, for her part, merely offered an unsupported and unreasoned refusal to recognize the constitutional violation.^[74]

There is a good reason for the failure of the Bush v. Gore dissenters to offer any legal defense of what the Florida court did. It was simply indefensible under the principles established in the Supreme Court's equal protection jurisprudence.

---

WERE THERE LEGALLY PREFERABLE ALTERNATIVES
TO THE COURT'S APPLICATION
OF EQUAL PROTECTION ANALYSIS?

Although the dissenters did not provide any legal criticism of the majority's equal protection analysis, they did dissent. It is therefore worth considering whether their dissenting opinions contained or suggested any legally appropriate objections to the majority's disposition of the case.

Refusing to Review the Case

The most plausible legal objection offered by any of the Bush v. Gore dissenters was Justice Breyer's suggestion that the Twelfth Amendment assigns to Congress, not to the federal courts, the responsibility for correcting constitutional violations like those the Florida Supreme Court committed. The Twelfth Amendment does assign to Congress the authority and responsibility for counting electoral votes. And it seems undeniable that Congress must also have the authority to make decisions about the legal validity of votes that are submitted to Congress, most obviously in cases where more than one slate of votes is received from the same state.^[75] And it may well be that Congress is authorized to ignore judicial decisions that conflict with its own judgments about the legality of the electoral votes it receives.

For Justice Breyer's suggestion to have any merit in the context of this case, however, one would have to go even further and argue that the Constitution gives Congress the exclusive authority to rule on the legality of electoral votes, thereby depriving federal courts of the jurisdiction they would otherwise have to adjudicate claims arising under federal law. The constitutional text, however, does not by its terms provide such exclusive jurisdiction to Congress. An argument supporting such exclusivity would therefore have to rely on inferences from the structure and history of the Constitution and/or on the judicially developed "nonjusticiability" doctrine.^[76]

Breyer made no attempt to develop an argument along these lines in his Bush v. Gore dissent, probably because of one simple and powerful legal fact: the Supreme Court had previously held, in McPherson v. Blacker, that Congress does not have such exclusive authority.^[77] Breyer was obviously aware of this holding, since the Court had unanimously relied on dicta in the same case just a few days earlier in Bush I. It would have been quite a challenge to explain why the Court should overrule the holding in a case on whose dicta the justices had so recently and unanimously relied.

Accordingly, Breyer never quite asserted that the Court was legally forbidden to review the Florida court's judgment. Instead, he merely contended that the Twelfth Amendment, as well as various federal statutes that

---

had been enacted to guide the counting of electoral votes, somehow conveyed a counsel of "judicial restraint."^[78] In the end, Breyer did not and could not contend that the majority committed a legal error in agreeing to review the Florida court's decision. Instead, he offered a nakedly political critique of the majority: "[A]bove all, in this highly politicized matter, the appearance of a split decision runs the risk of undermining the public's confidence in the Court itself."^[79]

This political approach to the exercise of jurisdiction deserves some attention, for it goes to the core of the most commonly articulated criticism of the Bush v. Gore majority. The justices, we are often told, have a duty to preserve the institutional capital of the Court by avoiding entanglements in the "political thicket," where their reputation for impartiality might be sullied, fairly or not. As Breyer so eloquently put it:

[T]he public's confidence in the Court itself … is a public treasure. It has been built slowly over many years, some of which were marked by a Civil War and the tragedy of segregation. It is a vitally necessary ingredient of any successful effort to protect basic liberty and, indeed, the rule of law itself. We run no risk of returning to the days when a President (responding to this Court's efforts to protect the Cherokee Indians) might have said, "John Marshall has made his decision; now let him enforce it!" Loth, Chief Justice John Marshall and The Growth of the American Republic 365 (1948). But we do risk a self-inflicted wound—a wound that may harm not just the Court, but the Nation.^[80]

Breyer forgot to mention that this argument about avoiding the "political thicket" was one that the Court had rejected in the vote-dilution cases on which the majority relied.^[81] Moreover, the notion of a general duty to avoid decisions that might undermine the public's confidence in the Court is not one that anybody actually believes. In fact, many of the Court's most intensely admired decisions are exactly those that were most controversial when decided: Brown v. Board of Education, which forbade racially segregated schools; Engel v. Vitale, which forbade prayer in the schools; Miranda

v. Arizona, which forbade the use of voluntary confessions at trial unless preceded by a series of judicially created warnings; Reynolds v. Sims, which required equality of population in state legislative districts; Roe v. Wade, which established a right to abortion; Texas v. Johnson, which protected a right to desecrate the American flag.

Notwithstanding the sound of Breyer's rhetoric, the theory underlying his call for judicial restraint is actually not one that would preclude any of the decisions in the above list. On the contrary, it is a theory meant to foster just such controversial decisions, along with their frequently profound political effects, even or perhaps especially when those effects are so profound as to shake the public's confidence in the Court. The real theory, well

---

known to sophisticated students of law and political science, is that the Supreme Court should refuse to decide certain politically sensitive cases, especially those involving the constitutional allocation of power between the federal and state governments, in order to conserve the Court's political resources for more important tasks, especially those involving the protection of "individual liberties."^[82] In practice, what this means is that the Court should sometimes allow the Constitution to be violated when Congress infringes on the rights of the states while protecting judicially selected "individual liberties" that often have no basis in the Constitution.

This calculated, asymmetrical, and ultimately lawless concern with the maintenance and deployment of judicial political capital has been a hallmark of modern liberal jurisprudence. It is, in fact, a corollary of the political theory reflected in Justice Stevens's dissent, where the rule of law and the rule of judges become conflated. And it is very plainly the basis for Breyer's dissent. Even though he acknowledged that the Florida court's recount process was inconsistent with constitutional standards, Breyer contended that "the Court is not acting to vindicate a fundamental constitutional principle, such as the need to protect a basic human liberty."^[83]

Breyer attempted to justify his position by arguing that the Court could have waited to see whether the unconstitutional recount process would actually alter the election's outcome, thus giving the Florida courts an opportunity to address the constitutional issue "if and when it was discovered to have mattered."^[84] For reasons explored below, Breyer was wrong to assume that there was time left for the state court to correct the problem it had created. Even apart from that mistake, however, it is simply not the case that Justice Breyer believes that the Supreme Court should generally stay its hand until the very last moment before a constitutional violation becomes unquestionably irremediable.

Just a few months before Bush v. Gore, for example, Breyer himself had written the majority opinion in a case that had the following interesting features: (1) the Court was reviewing a state statute that had been deliberately drafted to be consistent with the Supreme Court's case law; (2) the Court rejected an interpretation of the statute that would have made it consistent with that case law, instead adopting a far-fetched interpretation that allowed the Court to invalidate the statute; (3) the state itself had argued before the Supreme Court in favor of the interpretation that the Court rejected; (4) the state courts had never been allowed to review the statute at all because the Supreme Court struck it down before it was ever applied to anyone; and (5) the Court's 5–4 decision exposed divisions within the Court whose bitterness easily exceeded what was expressed in Bush v. Gore.^[85]

And what was the "fundamental constitutional principle" at stake in this case, the likes of which were supposedly absent in Bush v. Gore? The right

---

to so-called partial birth abortion, a procedure that is deeply repugnant to many millions of American citizens, that had been outlawed by at least thirty states, and that Congress had twice voted by wide margins to forbid. Compared with preserving this truly important right, what is a little matter like conducting a presidential election in a constitutional manner and protecting the constitutional rights of those who voted in it? According to Justice Breyer, not much.

Perhaps it should come as no surprise that the left wing of the current Court would object to deciding a controversial, high-profile case involving partisan politics that would contribute nothing to the protection of the Left's favored individual liberties. Somewhat more surprisingly, however, the theory of judicial politics underlying Breyer's dissent in Bush v. Gore is one to which more conservative members of the Court have sometimes been attracted.

In 1992, for example, the Court reaffirmed the judicially created right to abortion, even while strongly hinting that some justices who voted to do so had serious misgivings about the decision's consistency with the Constitution.^[86] And just last year, the Court reaffirmed a constitutional right to so-called Miranda warnings,^[87] notwithstanding the fact that some members of the majority had previously said that such warnings are not required by the Constitution.^[88] In both cases, stare decisis was offered as the principal rationale for the decision, but neither decision can be explained on that ground, for they both reaffirmed some precedents while overruling other and more recent precedents.^[89]

Far more important in both cases than any supposed respect for precedent was an easily discernible concern with the Court's own public image and a fear of diminishing its own political capital. In the abortion case, for example, a majority of the justices issued one of the most grandiose expressions of the judicial self-importance on record:

Like the character of an individual, the legitimacy of the Court must be earned over time. So, indeed, must be the character of a Nation of people who aspire to live according to the rule of law. Their belief in themselves as such a people is not readily separable from their understanding of the Court invested with the authority to decide their constitutional cases and speak before all others for their constitutional ideals. If the Court's legitimacy should be undermined, then, so would the country be in its very ability to see itself through its constitutional ideals.^[90]

The Miranda-warning opinion, which was mercifully free of such rhetoric, confined itself to observing that Miranda "has become embedded in routine police practice to the point where the warnings have become part of our national culture."^[91] But the implications of this bland statement are no less troubling. Indeed, I would translate the Court's remark to mean something

---

like this: We would look pretty silly if the one rule of constitutional law that every American is familiar with, from watching untold numbers of copsand-robbers shows on television, were suddenly declared by the Supreme Court to be a figment of the Court's own imagination. The masses might even start to wonder where all the other rules of constitutional law are coming from. And who knows what would happen to our national culture then?

Three members of the Bush v. Gore majority had joined one or both of these opinions.^[92] It is therefore striking that all three rejected the temptation to conserve the Court's political capital by avoiding any involvement in Bush v. Gore. They could easily have avoided such involvement, simply by voting not to review any of the Florida election cases. Such refusals require no explanation and are without precedential effect. Indeed, it had been widely anticipated that this is exactly what would happen before the Court surprised the world by granting review in Bush I.

The Bush v. Gore majority had to know that a decision in Bush's favor would trigger an avalanche of scurrilous accusations and politically motivated attacks and endless insinuations about their personal integrity. They were thus faced with a very unpleasant choice: if they enforced the law, they ran the risk of acquiring a reputation for having done the opposite, but if they refused to enforce the law, they would preserve their reputation for judiciousness. In deciding to hear the case and then resolving it in accordance with the law, the majority demonstrated genuine integrity and impartiality in exactly those circumstances in which it is most difficult to practice.

In contrast to Justice Stevens's remarkable assault on George W. Bush for having had the temerity to defend himself against Vice President Gore's lawsuit, and in contrast to Stevens's emotional attack on his colleagues for agreeing to hear Bush's appeal, the majority treated this as a legal case that deserved to be treated as such by judges, even if others chose to use it as a political football:

None are more conscious of the vital limits on judicial authority than are the members of this Court, and none stand more in admiration of the Constitution's design to leave the selection of the President to the people, through their legislatures, and to the political sphere. When contending parties invoke the process of the courts, however, it becomes our unsought responsibility to resolve the federal and constitutional issues the judicial system has been forced to confront.^[93]

One might think that fulfilling their unsought responsibilities is just about the minimum that we ought to expect from Supreme Court justices, who are given life tenure for just this purpose. But when one reflects on the concept of judicial integrity that infuses the dissenting opinions in Bush v.

---

Gore, a simple willingness to enforce the law begins to look like a kind of heroism.

The Article II Argument

The majority's decision in Bush v. Gore relied entirely on an equal protection analysis. One line of criticism, particularly appealing to the conservative legal mind, is that the decision should have rested instead on the analysis set forth in Chief Justice Rehnquist's concurring opinion. That opinion resolved the question that the Court had avoided in Bush I.

Article II of the Constitution provides: "Each State shall appoint, in such manner as the Legislature thereof may direct, a Number of [presidential] Electors."^[94] Along with Justices Scalia and Thomas, Rehnquist argued that the Florida Supreme Court had violated the Constitution by discarding the election statutes written by Florida's legislature and writing a new election code that was inconsistent with the legislature's directions.

One can make a powerful argument for this conclusion. Even apart from its many strange "interpretations" of statutory language, the Florida court's crucial decisions—especially the order for a partial and selective statewide recount—were simply disconnected from anything in the statutes. Taken as a whole, moreover, the court's exposition of Florida law had results that were so absurd and inequitable that they could not possibly have been intended by the legislature. Whatever authority there might be for a state court to ignore the legislature's directions in other contexts, Article II of the Constitution appears on its face to forbid such judicial reshaping of the law in connection with the appointment of presidential electors. This straightforward textual argument has a kind of intellectual power that no equal protection analysis can match. A more detailed comparison of the majority's equal protection approach with the Article II approach, however, will show that the one is not so clearly preferable to the other as may first appear.

Consider first the weaknesses of the majority's analysis. The Supreme Court's entire equal protection jurisprudence is notoriously ill rooted in either the text of the Fourteenth Amendment or in the expectations of those who enacted it. The voting rights branch of equal protection law, moreover, is particularly vulnerable to criticism based on the text and history of the Constitution, as Justice Harlan demonstrated in his devastating and unanswered dissent in Reynolds v. Sims.^[95] The evolution of the law of equal protection, moreover, has been something less than a model of logical consistency. New doctrinal pathways have sometimes been opened up with scarcely a legal reason offered,^[96] while equally plausible lines of development have been foreclosed or suddenly stopped in their tracks without much more than a wave of the hand.^[97]

---

Over the decades, the Court has developed a complex scheme under which it requires varying degrees of justification for the inequalities associated with different kinds of laws. Both critics and proponents of the aggressive use of equal protection analysis have contended that this scheme does not constitute a set of preexisting rules that are applied to new factual situations as they arise but rather reflects a series of judgments made independently of the theoretical apparatus that is used to explain the results.^[98] Notwithstanding the very real difficulty of identifying a coherent set of principles that are applied in a principled manner throughout the Court's equal protection cases, however, it does not follow that every equal protection decision is an unprincipled exercise of political judgment.

For example, given that the Court has required states to apportion their legislatures on the basis of equal population in order to avoid diluting the votes of some citizens, consistency requires that this rule be applied to all states. Creating an arbitrary exception—such as one for states with two Republican senators, or one for states through which the Mississippi River passes—would clearly be unacceptable. Conversely, refusing to create such exceptions is appropriately principled. At the other extreme, the creation of some exceptions is clearly proper and therefore does not manifest an arbitrary or unprincipled approach to the law. The Court, for example, has recognized that perfect equality of population in state legislative districts would create enormous and possibly insurmountable practical difficulties and has therefore never demanded it.

Between these two extremes is a middle range, where more or less reasonable differences of opinion might arise. As the failure of the Bush v. Gore dissenters to mount any meaningful criticism of the majority's equal protection analysis suggests, this case is much closer to one extreme than the other. The Florida court's recount procedure was rife with differences in the treatment of various categories of ballots that were at best arbitrary, and there was no compelling or even legitimate reason to create an equal protection exception that would permit such capricious forms of inequality. Indeed, the truly unprincipled course of action would have been to create an exception on the basis of the legally flimsy or irrelevant grounds advanced by the Bush v. Gore dissenters.

Thus the application of equal protection analysis by the Bush v. Gore majority did not exhibit the sort of unprincipled, essentially political judgments that have rightly made legal conservatives uncomfortable with some of the Court's equal protection decisions.

A related but slightly different objection to the majority's analysis is that it will lead to a flood of socially undesirable litigation challenging a vast number of traditional election practices.^[99] Must every voter now use exactly the same kind of ballot, which will be counted by exactly the same kind of machine? If some ballots are counted or recounted by hand, must the same

---

treatment be given to all ballots in that election? Must voters who ask questions of officials at their local polling place receive exactly the same answer in precisely the same words? Must those who count the ballots and tabulate the results receive exactly the same training and conduct themselves according to exactly the same procedures? Is it even permissible for election officials to be chosen in partisan elections, as Florida's secretary of state and many of its local canvassing officials were?

These and a host of similar questions are now thought to be the inevitable subject of litigation in the wake of future elections, where they will inevitably produce new suspicions about judicial bias. That danger certainly does exist, even though the Court expressly limited its holding to cases involving a court-ordered statewide recount lacking even the rudimentary requirements of equal treatment and fundamental fairness.^[100] Defendants in future cases will cite the narrow statement of the Court's holding in order to show that Bush v. Gore does not compel radical and unwarranted changes in traditional election practices. But plaintiffs will nonetheless be able to argue that a great many of those practices have in one way or another crossed an ill-defined boundary between what is fundamentally unfair and what is permissibly or tolerably unfair.

It is too soon to know how many benefits will come at what cost as future courts wrestle with the questions that are certainly going to arise in an area that the majority freely acknowledged is fraught with "many complexities."^[101] But it is not too soon to recognize that this is nothing new in the jurisprudence of equal protection. Indeed, virtually every major equal protection decision has created the potential for similar consequences. When the Court held that segregated public schools are unconstitutional, it inevitably opened up a host of questions about the permissibility of other forms of official segregation and discrimination. Some of those questions continue to be litigated almost a half century later.^[102] Similarly, when the Court ruled that legislative districts must be apportioned equally on the basis of population, it opened the way for a great deal of ensuing litigation about the exact degree of equality that is required, about the possibility of special circumstances in which there might be good reason to relax the general rule, and about the application of the underlying principle of equal weighting for all votes to arguably analogous situations like that presented by gerrymandered districts.

If Bush v. Gore does lead to litigation that results in significant alterations of American election practices, that might merely indicate there is a real problem that needs to be addressed, as was clearly the case with Brown v. Board of Education. But that hardly seems the most likely outcome. For all their impressive industry and creativity, Vice President Gore's lawyers utterly failed to produce examples of the existence of election practices that even approached the level of arbitrary and unnecessary unfairness that

---

pervaded the court-ordered recount in Florida. There may be some existing practices that will come into serious question as a result of Bush v. Gore, such as the availability of recounts in selectively chosen jurisdictions under Florida's protest mechanism.^[103] But there is little reason to believe and nothing in the Court's opinion to suggest that any massive or inappropriate reconsideration of America's traditional, decentralized electoral system is about to be undertaken by the federal courts.

Thus, the majority's equal protection analysis is not quite so problematic as legal conservatives may be inclined to suppose, either with respect to its roots in prior case law or with respect to its implications for future case law. On the other side of the scale, Chief Justice Rehnquist's Article II analysis is not without significant difficulties of its own.

First, although Rehnquist's analysis is anchored in the text of the Constitution, that textual anchor is an ambiguous one. It is certainly quite plausible to read Article II to outlaw election procedures that are devised by courts in contravention of a state legislature's directions. But it is not inconceivable that the Constitution's reference to the legislature's directions could refer to state statutes as interpreted by state courts. The framers were well aware that statutes often do require judicial interpretation in order to be applied, and federal courts ordinarily assume that state statutes mean what state courts say they mean. It would not be outlandish to interpret Article II as incorporating the same background assumption.

Second, using Rehnquist's Article II theory would have been unprecedented. In one sense, that is unproblematic. An Article II objection to a state election had apparently never come before the Court, and every issue has to be a new issue once. In another sense, however, using the Article II argument would have created some tension with existing precedent. Let us assume that the Florida Supreme Court's application of the Florida election statutes was so far-fetched and untenable that it constituted an act of legislating rather than an interpretation of existing law. The same can be said of a significant number of decisions by the U.S. Supreme Court itself.^[104] Indeed, the dissenters in some of these cases have plausibly suggested that the Court was violating the Constitution by legislating from the bench.^[105] Although some of us would have been well pleased if Bush v. Gore had signaled the beginning of a new era of judicial respect for the text of the statutes that the Supreme Court is charged with interpreting, a decision relying on Article II would have exposed the Court to a colorable objection that it was holding the Florida court to standards of fidelity in statutory interpretation that it has not imposed on itself.

Third, an opinion based solely on Article II grounds might have suggested that the Florida statutes themselves were constitutionally unproblematic. In fact, however, substantial equal protection objections can be

---

raised against a statutory scheme under which the losing candidate can demand recounts in counties selectively chosen so as to tilt the incidence of random errors in his own favor. Although the majority did not need to reach this issue, its application of equal protection analysis suggests, more strongly than an Article II analysis would have, that such objections should be taken quite seriously.

Thus, the equal protection and Article II rationales for reversing the Florida Supreme Court have somewhat different legal strengths and corresponding weaknesses. In important ways, the equal protection rationale is less bold—and in that respect perhaps more judicious. Even if one is inclined to prefer an argument based on Article II, as I am, it is not preferable across all the relevant dimensions.^[106] And it is not on the whole so clearly preferable as to provide a good reason for criticizing the majority's use of equal protection. It should therefore come as no surprise that those members of the Court who joined Chief Justice Rehnquist's opinion also joined the majority opinion.

The Suspiciously Narrow Holding

Courts, especially appellate courts, are supposed to apply general rules and standards to particular cases. When an appellate court cannot or will not articulate its reasons at an appropriate level of generality, one is entitled to wonder whether it is being driven by something other than principle. For that reason, there is something immediately troubling about the Bush v. Gore majority's narrow statement of its holding:

The recount process, in its features here described, is inconsistent with the minimum procedures necessary to protect the fundamental right of each voter in the special instance of a statewide recount under the authority of a single state judicial officer. Our consideration is limited to the present circumstances, for the problem of equal protection in election processes generally presents many complexities.

The question before the Court is not whether local entities, in the exercise of their expertise, may develop different systems for implementing elections. Instead, we are presented with a situation where a state court with the power to assure uniformity has ordered a statewide recount with minimal procedural safeguards. When a court orders a statewide remedy, there must be at least some assurance that the rudimentary requirements of equal treatment and fundamental fairness are satisfied.^[107]

At least at first, this could sound rather like the statement of people who know what result they want but cannot quite say what their reasons are. Before jumping to that conclusion, it is important to remember that overly

---

broad holdings can be worse than those that are too narrow. Broad holdings may effectively decide future cases that are factually dissimilar in ways that should be legally distinguished. That danger is particularly acute in an area of the law, like equal protection, in which the Court is necessarily drawing lines between "too much" and "not enough" without the benefit of guidance from the Constitution itself.

Closer consideration of the Bush v. Gore holding reveals good reasons for a narrow holding. First, it was appropriate to limit the decision to recounts. Procedures employed after the decision makers know they are dealing with a close election, in which one candidate is provisionally the loser, present opportunities for abuse that are at the very least much less pronounced in other circumstances. This case provides an example, for nobody would have dreamed of proposing that an initial count of the ballots be conducted in the way that the court-ordered recount was proceeding in Florida. Without accusing or exonerating anyone in Florida of misconduct, it is obvious that the incentives to adopt inappropriately discriminatory procedures increases dramatically once the responsible officials know which candidate is more likely to be adversely affected by them.

Second, it was also appropriate to limit the decision to cases involving judicial recounts because these enjoy a finality that is not present when executive officials make decisions that are subject to judicial review. Similarly, it was probably appropriate to limit the decision to cases involving a recount by a single judge because (a) those are the cases in which there is least likely to be any good practical reason for tolerating significant differences in the way similar ballots are treated; and (b) those are the cases in which there is the least chance that arbitrary differences in the treatment of ballots will cancel each other out and thus leave the result of the election unaffected.

That leaves one limitation on the holding for which I cannot see a plausible justification: the restriction to statewide recounts. Although the inclusion of this limitation seems mistaken to me, it is a relatively small point that cannot justify the conclusion that the majority was result oriented or unprincipled. If this is the only error in the majority opinion, it is almost miraculous that an opinion written under such enormous time pressures would be so slightly blemished.

Third, it is worth emphasizing that the Court did not preclude the application of vote-dilution principles to other election procedures, such as statewide recounts or counts conducted by executive officials. The majority decided only that they lacked sufficient time and information to evaluate such procedures responsibly in the context of this case. If there is one thing about this extraordinary case that should be undeniable, this would seem to be it.

---

Fidelity to Federalism

Another common criticism of the Bush v. Gore majority is that they behaved hypocritically by interfering in Florida's resolution of its own state election. These are the same five justices who have been moving in what many consider an aggressive fashion to protect the states from federal interference in a variety of other contexts. What happened to their solicitude for states' rights in this case?^[108]

One might simply turn the question around and ask: What happened to the dissenters' solicitude for federal authority and the fundamental equal protection rights of voters? This kind of "so's your Mother" response, however, is both inadequate and inappropriate. It is inadequate because it is perfectly possible for everyone on the Court to be guilty of hypocrisy. It is not much of a defense to a charge of hypocrisy to show that someone else is hypocritical as well. And it is inappropriate because it distracts attention from the real question, which is whether the case was correctly decided.

The hypocrisy objection is never framed in precise legal terms, nor could it be, for there is no legal tension between the holding in Bush v. Gore and the holdings in any of the Court's other recent decisions. In legal terms, moreover, a "federalism objection" to Bush v. Gore would be ludicrous.^[109] States are required to conform their conduct with the U.S. Constitution, and the Supreme Court has jurisdiction to correct state court judgments that are inconsistent with the Constitution. No member of the Bush v. Gore majority has ever questioned these propositions, and none has ever suggested that the Supreme Court should stop enforcing either the Fourteenth Amendment or the Court's vote-dilution precedents.

The federalism objection thus turns out to be yet another criticism based on the premise that the Bush v. Gore majority were politically rather than legally misguided. But unless one can show that the majority were politically rather than legally motivated, which I think has not and cannot be done, this objection is simply another regrettable manifestation of the fashionably decadent view that judges cannot and should not be anything except robed politicians.

The Controversial Remedy

Among the more interesting criticisms of the Bush v. Gore majority is that they erred by declining to remand the case to the state court with instructions to conduct a recount under constitutionally permissible procedures. Justices Souter and Breyer, who thought the Court should have refused to hear the case at all, advocated this approach as the best way of dealing with the constitutional violation whose existence they could not deny.

---

To understand this criticism, it is important to recall that Bush v. Gore was decided on December 12, the deadline for Florida to take advantage of the safe harbor offered by federal law.^[110] Whether this nonbinding deadline was met or not, however, federal law required that all presidential electors meet and cast their votes on December 18.^[111] As a practical matter, it is almost inconceivable that the Florida courts could have established constitutionally adequate procedures and then used them to conduct a statewide, hand recount during this six-day period. And even if one supposes that this could somehow have been done, how could the loser have been given any meaningful right of appellate review?

These difficulties are exactly what made the Souter-Breyer approach look so politically attractive. Writing immediately after the Court's decision, Michael W. McConnell put it this way:

Such a disposition would have maintained the 7–2 majority for the entire holding, which the American public would find vastly more reassuring. To be sure, it is probably impossible to conduct a proper recount by [the statutory deadline of December 18], but by cutting off the possibility, the court encouraged critics to blame the court majority—rather than the passage of time—for the outcome.^[112]

I agree with McConnell that this disposition would have spared the Supreme Court some of the criticism it has received and that the institutional political interests of the Court might have been served by a nearly unanimous ruling that encouraged the Florida court to make yet another effort to find a "better" way to count the ballots than the initial machine counts had provided. It would then have become more clear to more people that the Florida court's project had been frustrated by simple reality, as well as by its own mistake in extending the protest period beyond the statutory deadline.

Although it is easy to see how this approach might have prevented some of the political criticism that the Court has received, it is also easy to see how such a stratagem could have blown up in the Court's face. First, the Florida court had already proved highly aggressive and irresponsible in dealing with federal law and with the U.S. Supreme Court. It is therefore quite possible that the next stab at a statewide recount would have been infected with new constitutional problems, which the U.S. Supreme Court would then have had to deal with under time pressures even greater than those it faced in Bush v. Gore.

Second, the passing of the December 12 safe harbor deadline would virtually have assured intervention by the Florida legislature. With the election results themselves still tied up in litigation and the absolute legal deadline of December 18 fast approaching, Florida would have been in real

---

danger of having made no clear choice of electors in time for the Electoral College to meet. Accordingly, the legislature was already gearing up to appoint a slate of electors directly. Given the makeup of the Florida legislature and the fact that Bush was the certified winner of the election, it is safe to predict that a slate of electors pledged to Bush would have been selected. That would have created a whole new swarm of legal and political controversies.

Good arguments can be made, on the basis of both the Constitution and a specific federal statute, that the legislature had a right, or even a constitutional duty, to step in and appoint electors.^[113] But this had never happened before, and the legal basis for it was anything but crystal clear. More litigation, probably beginning in the Florida courts, would therefore have ensued, and it is entirely possible that the U.S. Supreme Court would have been faced with a new set of difficult legal questions, which would have been posed in an atmosphere even more politically charged than before. And if the ongoing recount of Florida ballots had at some point tipped just once in Gore's favor, the political histrionics on both sides would probably have reached levels well beyond the very impressive exchanges of venom that we had already observed. Intervention by the Florida legislature also would have heightened the chances that Congress would have received votes from multiple slates of putative electors, as had happened in 1876. This would have generated yet more litigation, with all the added potential for the U.S. Supreme Court to be accused of politically motivated decisions, no matter how it ruled.

We have no way of knowing whether the Bush v. Gore majority had considerations like these in mind when they decided the case, let alone whether it would have been shrewder to accept the Souter-Breyer invitation.^[114] And there is no reason to criticize the majority for rejecting the Souter-Breyer approach. Apart from the questionable propriety of employing such calculations, they were entirely unnecessary.

My reason for offering this conclusion is quite simple: the Souter-Breyer approach was legally untenable, for exactly the reasons given by the majority. On December 11, just one day before the decision in Bush v. Gore, the Florida Supreme Court had finally issued its decision in response to the remand in Bush I. In that opinion, the Florida court had interpreted state law to allow the late filing of amended election returns by county officials in only two circumstances: when a late filing would preclude someone from exercising his rights under the statutory "contest" provisions and when the late filing would "result in Florida voters not participating fully in the federal electoral process, as provided in 3 U.S.C.§5 [the safe harbor provision of federal law]."^[115] Perhaps even more emphatically, the Florida Court said in the same opinion: "Although the [Florida Election] Code sets no specific

---

deadline by which a manual recount must be completed, the time required to complete a manual recount must be reasonable,"^[116] to which it added the following footnote:

What is a reasonable time required for completion will, in part, depend on whether the election is for a statewide office, for a federal office, or for presidential electors. In the case of the presidential election, the determination of reasonableness must be circumscribed by the provisions of 3 U.S.C. § 5, which sets December 12, 2000, as the date for final determination of any state's dispute concerning its electors in order for that determination to be given conclusive effect in Congress.^[117]

And again, in the same opinion:

As always, it is necessary to read all provisions of the elections code in pari materia. In this case, that comprehensive reading required that there be time for an elections contest pursuant to section 102. 168, which all parties had agreed was a necessary component of the statutory scheme and to accommodate the outside deadline set forth in 3 U.S.C. § 5 of December 12, 2000.^[118]

Thus, the Florida Supreme Court had already concluded, as a matter of state law, that recounts had to be concluded by December 12.^[119] If the U.S. Supreme Court had remanded the case on December 12 with instructions or encouragement to conduct a recount under constitutionally adequate procedures, it would have been ordering or inviting the Florida court to violate Florida law as construed by the Florida Supreme Court. The U.S. Supreme Court simply had no grounds for doing that because the ensuing violation of state law would not have been dictated by any requirement of federal law.

One might argue that the Florida court's discussion of the binding nature of the December 12 deadline came in the context of a discussion of the protest provisions of the Florida Election Code, whereas the issues in Bush v. Gore arose under the contest provisions. Nothing in the Florida court's December 11 opinion, however, suggested that this should make any difference at all. The Florida court's decision in the contest case, moreover, referenced the federal safe harbor statute, without mentioning any alternative possible deadlines.^[120] The U.S. Supreme Court simply had no basis at all for imagining that some deadline other than December 12 would be applicable under state law to the "contest" at issue in this case.

Still, one might say, the Supreme Court should at least have remanded the case to the Florida court so that it could reexamine the state law question itself. Perhaps that court would have concluded that state law ultimately subordinated the December 12 deadline to the goal of obtaining a constitutionally acceptable hand recount.

Fair enough. But that is exactly what the Supreme Court did. Contrary to a

---

widespread misperception, the Supreme Court did not forbid the Florida court from attempting to conduct a statewide recount under constitutionally permissible standards.^[121] That would have been the effect of a judgment that reversed the Florida court and remanded with instructions to dismiss the case.^[122] But the Court did not order the case dismissed. Instead, it reversed and remanded with instructions "for further proceedings not inconsistent with this opinion." The only statement in the Supreme Court's opinion that could conceivably be considered "inconsistent" with a new recount is the following:

Because the Florida Supreme Court has said that the Florida Legislature intended to obtain the safe-harbor benefits of 3 U.S.C. § 5, Justice breyer's proposed remedy—remanding to the Florida Supreme Court for its ordering of a constitutionally proper contest until December 18—contemplates action in violation of the Florida election code, and hence could not be part of an "appropriate" order authorized by Fla. Stat. § 102.168(8) (2000).^[123]

It is true that this statement assumes that Florida law had not changed between December 11 and December 12, and it assumes that the December 11 opinion meant what it appeared to say. But this statement does not purport to forbid the Florida court from concluding on remand that the U.S. Supreme Court had misinterpreted the statements it made on December 11. The Supreme Court's statement, for that matter, does not purport to forbid the Florida court from overruling its own December 11 interpretation of Florida law. Thus, as a legal matter, the Florida court was indeed left free to order the sort of recount that Justices Souter and Breyer suggested.

Gore's lawyers reportedly recognized that the Florida Supreme Court had been left free to order a new recount but decided on political grounds not to request one.^[124] There is not much to be gained by speculating about what the Florida court might have done in response to such a request,^[125] but it is important to recognize that the U.S. Supreme Court did not prevent Gore from continuing to litigate his case and that the U.S. Supreme Court did not dictate the interpretation of Florida law to the Florida courts.

Here again, as in every other aspect of this case, the majority simply applied the law. If that turned out to be bad politics, which is a pretty dubious proposition anyway, it at least had the merit of being the right thing for judges to do.

CONCLUSION

For several decades, constitutional law has held that states may not weight the votes of people according to where they reside without a legitimate and compelling public purpose. Such vote dilution permeated the recount process

---

designed by the Florida Supreme Court, and that court offered no coherent, let alone compelling, justification for the discrimination it was imposing on the election process. Nor was any justification offered by either of the two dissenting U.S. Supreme Court justices who claimed that they could not perceive the completely obvious constitutional problem identified by the other seven.

The Bush v. Gore majority opinion has been harshly criticized—by the dissenters and by a wide range of commentators—for a variety of supposed sins. The Court should have refused to hear the case for fear of creating an "appearance" of political partiality. The Court should have refused to apply its Fourteenth Amendment precedents for fear of having them taken seriously in future cases. The Court should have ignored the Florida court's one-day-old decision about the meaning of Florida law, thereby inviting that court to commit further violations of federal law. The Court should have refused to apply well-established federal law in this case because of a supposed commitment by the Court's conservatives to some notion of federalism imputed to them by people who have apparently never read their opinions.

None of these criticisms has the slightest legal merit. All of them are political criticisms, offered by people who have forgotten the distinction between law and politics, or who do not want the distinction to exist, or who do not want to be snickered at for defending the distinction. Once one surrenders that distinction, however, all of law becomes at best a decadent exercise in sophistry.

Faced with a gross violation of law by a subordinate court, the Bush v. Gore majority did exactly what an appellate court is supposed to do. It reversed the erroneous decision and upheld the law. That this action has provoked so much outrage and so little reasoned approval suggests that the history of our contemporary legal culture may have to be written by a Tacitus, or perhaps a Juvenal.

NOTES

Thanks to Peter Berkowitz, Douglas R. Cox, C. Boyden Gray, Mara S. Lund, John O. McGinnis, and Richard A. Posner for helpful comments, and to the Law & Economics Center at George Mason Law School for generous financial support. I am especially grateful to Stephen G. Gilles for his relentlessly skeptical and constructive criticisms of several preliminary drafts. A more detailed and fully annotated version of this chapter will appear in the Cardozo Law Review as part of its "Votes and Voices" symposium.

---

---

---

---

8. THE GHOSTWRITERS

Arthur J. Jacobson

One feature of the majority's opinion in Bush v. Gore^[1] struck a sophisticated observer of the Court immediately. "[T]he opinion," wrote Linda Greenhouse in the New York Times, "was labeled ‘per curiam,’ meaning ‘by the court,’ a label used by courts almost exclusively for unanimous opinions so uncontroversial as to not be worth the trouble of a formal opinion-writing process."^[2] The per curiam in Bush v. Gore was far from unanimous, hardly uncontroversial, and certainly the subject of a "formal opinion-writing process." If Greenhouse is right and the majority in Bush v. Gore not wrong, theirs must fall within the category of opinions suggested by "almost." Greenhouse does not speak directly about the Court, only courts, and does not go into any detail about the occasions on which the Court, or, more precisely, a majority on the Court, labels an opinion "per curiam." Without knowing this, a reader may be forgiven the impression that recourse to the per curiam device in Bush v. Gore was unusual if not aberrant and that only sinister motives explain it. So I decided to investigate the Court's per curiam practice in order to right any rhetorical injustice the press inadvertently may have inflicted on the majority by testing their use of the per curiam against such norms as may be implicit in the practice.

My review of the Court's use of per curiams does suggest norms justifying the majority in Bush v. Gore. But more than that, the norms put the shoe on the other foot. For it is the motives of the minority—specifically, Justices Breyer and Souter—not the majority, that the norms put in question. Had Justices Breyer and Souter complied with them, then seven justices would have joined the per curiam, not just five, and the result in Bush v. Gore would have been 7–2, not 5–4. The rhetorical, hence political, effect of the opinion would have been very different, depriving critics of the decision of

---

a host of denunciations tied to a one-vote victory by the conservative wing of the Court.

FROM ORIGINS …

The history of the Court's use of per curiams effectively begins in 1883. Before then the Court had designated opinions "per curiam" only twice, in 1839 and 1862. West v. Brashear^[3] and Mesa v. U.S.^[4] dealt with minor (albeit interesting) matters of Supreme Court procedure on which the Court's decisions were indeed unanimous. In the five years from January 1883 to November 1888, the Court labeled two more opinions "per curiam," both unanimous.^[5] Both dealt with the actual disputes between the parties and not with Supreme Court procedure. Over the next twelve years, until the end of the century, the practice accelerated, with the Court labeling thirty-four opinions "per curiam," all unanimous. It has continued to accelerate since. Why the sudden increase after 1883?

Two forces converged in the 1880s and 1890s to account for it. One is obvious, the other more subtle. The obvious force is simply the dramatic increase in the Supreme Court's business during the course of the nineteenth century but especially in the 1880s, after Congress significantly enlarged the Court's jurisdiction in 1875. The pressure to dispose of cases quickly without dissents requiring elaborate reasoned responses by the majority was enormous. Unanimity invites brevity. Yet brevity and unanimity do not by themselves commend attribution of authorship to the Court. They make it feasible. Other mechanisms were at work that explain the Court's particular response to the press of business.

First was expansion of the category of procedure. The per curiams in 1839 and 1862 dealt with minor matters concerning the Court's own rules. To be sure, ten of the thirty-six per curiams between 1883 and 1900 did involve just such matters.^[6] But seven were dismissals for want of jurisdiction, where what was at stake was the power of the Court to hear the case— neither a minor matter nor about the Court's own rules.^[7] Nevertheless, lawyers routinely classify jurisdiction under the rubric of procedure and would have considered labeling opinions that dispose of jurisdictional defects "per curiam" a natural extension of the Court's pre-1883 practice.

The second extension was more synthetic. Fourteen per curiams in the period 1883–1900, including the two mentioned above, addressed questions that were not obviously procedural. The cases were a version of what lawyers today would call "companion cases," raising questions similar to or identical with ones the Court had just answered in a previous case, often that very day.^[8] Yet these are procedural in a sense and thus natural extensions of the per curiam practice as well. After all, it was only an accident of procedure that led cases involving the same set of facts or questions of law

---

to the Court separately. Today, separate cases presenting common questions of law and fact can be pulled together for trial and come up together as one case on appeal. The second extension prefigures this idea. Also, the label fit. When the Court had several cases before it, all presenting the same question, it would have been silly for it to repeat the majority and dissenting opinions in each case, inserting different names. Certainly the dissenting justices were not conceding agreement with the majority by not dissenting in the companion cases. Nevertheless, they could (and did) agree with the majority that if the case chosen as showcase for the opinions came out one way, the others should as well. The per curiams were not really decisions on substance but concessions to procedural justice.

Obviously, the second extension of the idea of procedure is extremely elastic. What counts as a "similar" question or a sufficiently recent answer can be stretched (and was stretched starting in the 1930s).^[9] So the second extension ultimately led to unanimous summary dispositions of clearly nonmeritorious appeals—where the loser below had the right, not the wisdom, to demand review—in excruciatingly short per curiam opinions. "Clearly nonmeritorious" here means just that the Court had answered a question sufficiently similar to the question raised on appeal and done so sufficiently recently that the Court could adequately and efficiently explain its decision by a citation, and nothing more.

The third extension plunged even deeper into the territory of substance. In four cases between 1883 and 1900 the Court labeled its opinion "per curiam" when, for one reason or another, the case came back to it after it had already made a decision.^[10] For example, in the first, Hunt v. Blackburn, the appellee died after the Court had entered judgment against him two years before. So the appellant returned to the Court asking that its decree "be made absolute against the heirs and representatives of Sallie S. Blackburn, deceased," which it did, per curiam. Once again, the only reason the case was presented separately to the Court was procedural accident; the Court had to fix a decree entered previously. Once again, it made perfect sense for the justices to be unanimous, even if some of them had dissented in the first round, and it was not unreasonable for the Court to speak per curiam.

Thus the category of procedure together with these three extensions of it account for thirty-five of the thirty-six per curiams between 1883 and 1900. The odd case out presented what was clearly not a question of Supreme Court procedure, however capacious the category, and clearly did not fit the practice the Court had been following consistently since 1839. In Smith v. United States^[11] the Court was asked to review a conviction in the U.S. circuit court in Maryland for a murder committed on Navassa, a guano island in the Carribean (more about circuit courts soon). (Guano islands were immensely valuable in those days for the digging and exportation of

---

guano, "a natural manure found in great abundance on some sea-coasts … consisting of the excrement of sea-fowl.")^[12] In the Guano Islands Act of 1856 Congress had extended the criminal jurisdiction of the circuit courts to any guano island whose discoverer proved to the satisfaction of the State Department and ultimately the president that he was indeed the discoverer and had peaceably occupied the island. The act provided that an indictment may be tried in the district into which the offender was first brought. Smith's indictment was tried in the district of Maryland. The problem was that Haiti was contesting American control over Navassa and in 1858 dispatched two warships to disturb the discoverer's peaceable possession. President Buchanan stated that he considered Haiti's claim unfounded and ordered a force to the island to protect citizens of the United States from Haitian interference. These events led Smith to question whether the circuit court had jurisdiction over his crime. The circuit court decided that it did, and in an almost wordless per curiam "opinion" the Supreme Court affirmed.^[13]

Like the seven jurisdictional per curiams before it, Smith concerns judicial power, but it would be a mistake to put it in the same category. For the power at issue in Smith is not the power of the Supreme Court within a system of courts already exercising jurisdiction over a dispute but rather the power of the system of courts altogether. It is also not strictly or solely about the power of courts but about the power of the United States to extend the criminal jurisdiction of courts to territories discovered and occupied by its citizens. In that sense Smith is the first undeniably nonprocedural per curiam.^[14] While a natural extension of the jurisdictional per curiams, Smith steps over the line into substance.

Smith was decided on November 24, 1890. The timing of the decision suggests a subtler but nonetheless powerful force leading to the flowering of per curiams in this era and the sudden appearance of a nonprocedural per curiam sharply departing from the Court's previous practice. November 24 was a scant three months before the enactment (and smack in the middle of the crafting) of legislation that would profoundly alter the Supreme Court's relationship to the other courts in the federal system. It was this alteration, I submit, that transformed the Court's per curiam practice from unanimous opinions about minor questions of procedure to include major and ultimately controversial questions of substance, like the per curiam in Bush v. Gore.

On March 3, 1891, the Evarts Act^[15] became law, a key step in dismantling a structure that had persisted since the founding of the Republic. Throughout the nineteenth century Supreme Court justices "rode circuit," just as high court judges in England had done in the Middle Ages. In England circuit riding was meant to bring central control over judicial matters to the provinces. In the United States its purpose was to put the justices in

---

touch with the people, to lessen the distance between Washington and the states. Here is how the system worked. Instead of having an intermediate appellate court between the Supreme Court in Washington and the local district (trial) courts, Congress established regional circuit courts. Rather than have judges of its own, each circuit court was composed, at first, of two Supreme Court justices together with the local district court judge. These three would decide cases "appealed" (actually brought by "writs of error") from the district court and would also conduct trials in certain cases, as in Smith. Cases could then be brought from the circuit court to the Supreme Court. The circuit court held two sessions a year in each district. Though the system effectively brought them to the people, the justices hated it, and as judicial business increased over the course of the nineteenth century the system threatened to collapse.^[16] Congress tinkered with changes, such as reducing the number of justices riding circuit from two to one and increasing the number of circuits (hence the number of justices). But in 1869 it bit the bullet. The Act of 1869^[17] created special judgeships for the circuit courts, authorizing the appointment of one circuit judge for each circuit, and reduced the frequency of the circuits from two every year to one every two years.

The Act of 1869 had two significant effects, which the Evarts Act and the campaign for its passage would later enhance: it estranged the justices from other federal judges, and it sharpened the institutional distinction and hierarchical supremacy of the Supreme Court. Withdrawing the justices from riding circuit withdrew them from mingling with the people but also with the judges on the lower courts. Riding circuit muted the distinctions between the district courts and the Supreme Court. Curtailing it heightened those distinctions. On the old circuit court both district judges and Supreme Court justices often heard cases together—some on appeal and some in trial—and cases reaching the Supreme Court usually bore a justice's personal imprimatur. After 1869, with the appointment of specialized circuit court judges and the reduction of circuit riding, Supreme Court justices became less coworkers and more supervisors of judicial business on the appellate level. More decisions reached the Court that lacked a justice's imprimatur. The Supreme Court thus developed opinions, but also interests, unattached to collegial relations with lower court judges. The justices now faced not one but two cadres of lower court judges whose professional and personal affiliations were far more intense than their ties to the justices. In such an atmosphere it became ever more plausible for the justices to divorce their opinions from the relationship of signature, to speak impersonally, per curiam, for the Court.

That, coupled with the enlargement of the Court's jurisdiction in 1875, accounts for the rise in per curiams and the broadening of their scope from 1883 through 1889. It does not, though, by itself account for the

---

appearance of a nonprocedural per curiam in 1890, in Smith v. United States. The key to Smith lies in its appearance on the eve of passage of the Evarts Act, which further heightened the estrangement and distinction of the Court. The Evarts Act created genuine regional intermediate appellate courts alongside the circuit courts. These were the circuit courts of appeal, each staffed with three judges of its own. The act also introduced the key concept of discretionary review by "writ of certiorari" in a large number of cases, in contrast to the old "writ of error," which provided for review as a matter of right. It did not abolish the circuit courts, since Congress was still wedded to the vision of justices mingling with the people, but took away their appellate jurisdiction over the district courts instead. Congress was not to eliminate circuit riding and the circuit courts completely until the Judicial Code of 1911. But in 1891 the modern system of local district courts, regional intermediate appellate courts, and a Supreme Court with activities confined to Washington was effectively in place. The evolution of the Court as a distinct and supreme institution in American life had largely been accomplished.

Smith reflects this evolution. Why did the Court speak per curiam while saying virtually nothing except "affirmed"? Because the controversy in Smith engaged a profound interest of the United States as a government, its power to extend criminal jurisdiction (if not sovereignty) over contested and economically valuable territory. In particular, it engaged an interest of the judiciary as an institution and the Court as the pinnacle of the judiciary. A judgment asserting that interest ought not to be signed by a single justice, even unanimously joined by the others. It ought to be a judgment of the Court as a distinct and supreme institution of the government. And the Court was ready to act as a distinct and supreme institution because of the articulation, isolation, and elevation of its position in the judicial hierarchy in the Acts of 1869 and 1891.

The voice of the per curiam in Smith was, in a way, the same as the voice of the procedural per curiams preceding it, the voice of a decision implicating the interests of the Court as an institution distinct from the judicial and political visions and interests of any one justice or aggregation of justices. In the case of the procedural per curiams, the decisions were minor and peculiar to the Court by nature of their subject matter. In Smith the decision was limited but in no sense minor, and it was not at all peculiar to the Court but a general interest of the government as a whole. Thus the voice of the per curiam in Smith was also new. It was not a per curiam by default, not a concession to the requirement of authorship in a case otherwise unworthy of its claim. It was a per curiam asserting the distinctive contribution of the Court as an institution to the interests of the Republic in a political arena where no assertion succeeds by default and every contribution

---

must be earned. Such a per curiam is not the voice of an understudy but of a star. It does not speak for or instead of the justices. Rather, the justices speak for the per curiam, as ghostwriters. It is the voice of this per curiam, by lineal descent and accomplishment, for which the majority ghostwrote the per curiam in Bush v. Gore.

The distinctiveness of the Court's voice in Smith is underscored by a second innovation in the same era—the Court speaking without attribution of authorship, either to the Court or to an individual justice. In the wake of the Evarts Act's creation of discretionary jurisdiction the Court began to craft methods of handling petitions for writs of certiorari. (The leader of these innovations was Chief Justice Melville W. Fuller, whose term, significantly, began two years before Smith.)^[18] One was to record denials of these petitions in the Court's official reporter in one brief sentence—Petition for writ of certiorari denied" without explanation. (The first of these was in 1896, in American Sugar Refining Co. v. Steamship G. R. Booth.)^[19] Often the explanation was that the issues raised were without merit; often they had merit, but a sufficient number of justices felt that addressing them was inopportune. The important point is that the majority did not, in general, attach the label "per curiam" to summary denials of petitions for writs of certiorari. One reason was that some denials were not unanimous, not truly "by the court" as a whole. But many were unanimous, and still the Court did not label them "per curiam."

The absence of signature in certiorari denials contrasts with the practice, amplified in the years following Smith, of attaching the "per curiam" label to unanimous summary dispositions of clearly nonmeritorious appeals. Why the difference? Summary dispositions of appeals are decisions on the merits constrained by the Court's understanding of the law. Summary dispositions of petitions for writs of certiorari are not decisions on the merits but rather decisions whether to hear a case at all. They are discretionary decisions, unconstrained by law. By granting a petition and deciding to hear a case, the Court is not saying that the position of the petitioner necessarily has merit; exactly the opposite might well be true. Nor by denying a petition and deciding not to hear a case is the Court saying that the petition necessarily lacks merit, only that the Court is declining to make a decision for reasons that may have little to do with the merits of the petition. Decisions whether to grant or deny petitions for writs of certiorari express an institutional interest in making a decision rather than judgments of law. Of course, the Court's interest is often informed by legal judgment but only as refracted through broader considerations of interest—whether the Court wishes to develop legal doctrine in a certain area, whether it is politically expedient for the Court to intervene, whether the Court can tolerate the addition to its workload, and a host of other factors. From the

---

perspective of strict legality, therefore, decisions about petitions for writs of certiorari are grounded in what can only be characterized as whim. And one does not sign whim.

Both innovations of the 1890s—giving voice to the Court as an institution in substantive per curiams while taking it away in decisions about petitions for writs of certiorari—express the same thought: the Court has a distinct interest as an institution of government apart from its obvious role as a supreme arbiter of legal doctrine. They tell us that the Court's interest extends from minor issues peculiar to the Court as an institution—Court procedure, Court jurisdiction, companion cases, returning cases—to great issues of state implicating the Court's role in articulating and defending an entire system of governance and in the case of per curiams, to nothing in between.^[20]

… TO NORMS …

All was well until 1909. Then a curious and fateful event, launched by none other than Justice Oliver Wendell Holmes Jr., began to bend the per curiam opinion from a device necessarily expressing unanimity about uncontroversial cases to a very different aim: Holmes, joined by Justices William H. Moody and Edward D. White, dissented from a per curiam. The case came to the Court by a procedure that while different from appeal, had the same demanding character. In Chicago, Burlington, & Quincy Railway Company v. Williams,^[21] the Eighth Circuit Court of Appeals had "certified a question" to the Supreme Court. The certification procedure allows a circuit court of appeals to ask the Supreme Court to answer a "pure question of law," a question that does not require the Court to consider any factual disputes. The purpose of the procedure is to allow the lower court to dispose of a case on the basis of one issue and thus avoid deciding others it wishes to avoid. Over a dissent without opinion by Justice David J. Brewer, the Supreme Court, in an opinion written by the elder Justice John Marshall Harlan, had decided that the Eighth Circuit's question was not a pure question of law, which was grounds for dismissing the certificate. The Eighth Circuit tried again, this time refining the question into three. The Supreme Court, per curiam, dismissed all three, saying that "this certificate is essentially the same as that disposed of" previously.^[22] Holmes dissented:

When this case was here before I felt doubts, but deferred to the judgment of the majority, as I think one should when it does not seem that an important principle is involved or that there is some public advantage to be gained from a statement of the other side. But it seems to me that the present order is a mistake upon an important matter, and I am unwilling that it should seem to be made by unanimous consent.^[23]

---

Note that the per curiam itself was appropriate: the Supreme Court had decided the question previously and six of the justices undoubtedly felt that the Eighth Circuit was fiddling with them. The dismissal of the questions was not unlike the per curiams for returning cases or for summary dismissal of clearly nonmeritorious appeals. Holmes could not bring himself to support unanimity. The ranks of per curiam had been broken.

Over the next thirty-five years, through 1944, one or more justices dissented from eight per curiams. That is still not a deluge, but Williams had established that per curiam and dissent—the rhetorical assertion that the Court's opinion is unanimous accompanied by a clear statement that it is not—could coexist in the same decision. The deluge came after. From 1945 through 1959, dissents accompanied per curiams in 156 cases, or an average of 10.4 cases per year (compared to an average of 0.26 per year in the period 1909–44). From 1960 to the present, the reports show approximately 1,000 per curiam dissents, two and one-half times the rate in the previous period. What Holmes began in 1909 has obviously become accepted practice.

A practice laced with irony. When a majority labels an opinion "per curiam," it is saying, "We, though a mere majority and not unanimous, speak for the entire court," and when a minority dissents in the face of that, it answers, "No you don't!" The majority is avowing an opinion unanimous that manifestly is not. By persisting in labeling the opinion "per curiam," it is, in effect, taking the position that the decision ought to be unanimous. By persisting in dissent, the minority takes the position that it need not. Dissent here plays an unusual, dual role: it registers the minority's disagreement with the underlying decision but also its objection to the majority's implicit invocation of a norm of unanimity. This rhetorical struggle between the minority and the majority—a struggle without legal consequence^[24]— shifts our focus from the ordinary dispute over the rightness of the decision to the propriety of the per curiam and the legitimacy of dissent. Accompanied by dissents, per curiams suddenly must be justified.

The divergent behavior of the justices in the next two per curiams after Williams burdened by a dissent—Murphy v. Sardell^[25] and Donham v. West-Nelson Manufacturing Co.^[26]—inchoately expresses the normative struggle. Both were cases decided within four years of Adkins v. Children's Hospital,^[27] in which the Court had held that a District of Columbia law prescribing minimum wages for women violated due process. Eight of the nine justices thought them indistinguishable from Adkins and joined a brief per curiam, in each affirming the judgment of the court below. Once again, the per curiams fell comfortably within the Court's practice, specifically, the use of per curiams to dispose of cases sufficiently similar to a case previously decided. But Holmes, along with Chief Justice William Howard Taft and Justice

---

Edward T. Sanford, had dissented in Adkins. What were they to do? If they could not get a majority of their colleagues to vote to reverse Adkins— and the norm of stare decisis suggests that they should not even try—Adkins was the law, and it would be useless protest to dissent in Murphy and Donham. Indeed they did not. Taft and Sanford simply joined the per curiams. But Holmes did something interesting. In Murphy, decided two years after Adkins, he "request[ed] that it be stated that his concurrence is solely upon the ground that he regards himself bound by the decision in Adkins v. Children's Hospital."^[28] He did not even file a formal concurrence. In the second, two years later, he silently joined the majority.

Justice Louis D. Brandeis behaved quite differently. He dissented in both Murphy and Donham. Brandeis was one of the justices to whom one can trace what Robert Post has described as a decline in "institutional norms of unanimous decision-making" on the Court during the course of the twentieth century.^[29] Though he was on the Court, Brandeis had not participated in the decision in Adkins, and it would have been reasonable for him to express his disagreement in the two subsequent cases in order to underscore how he would have voted. Brandeis's dissents need not have been implicit criticisms of the majority's claim to unanimity. Moreover, the norm of unanimity still had considerable heft in those days. The acquiescence of Holmes, Taft, and Sanford was undoubtedly motivated by that norm, which would have governed whether or not the majority used the per curiam designation. It was not necessarily a response to the call for unanimity implicit in the designation.

All but one of the six remaining per curiam dissents in the series are similarly inconclusive.^[30] In United States v. Swift & Co.,^[31] however, Justice Robert H. Jackson dissented from a jurisdictional per curiam even more articulately than Holmes did in Williams. The case came to the Court as a direct appeal under the Criminal Appeals Act,^[32] which provided for direct appeal to the Court of a decision or judgment based on the invalidity or construction of the statute on which the indictment was founded. The reason for direct appeals to the Court—skipping intermediate review by a circuit court of appeals—is to expedite resolution of a dispute challenging a component of the frame of governance, exactly the spur to signing per curiam.^[33] The trial court had dismissed an indictment under the Sherman Antitrust Act, and the United States appealed to the Supreme Court, saying that the dismissal was based on a construction of the Sherman Act. That was not necessarily clear from the trial judge's rather sibylline account of his decision. Nevertheless, a bare majority of the Court decided, per curiam, to remand the case to the circuit court of appeals for regular treatment, on the ground that the trial court's dismissal of the indictment did not involve construction of the Sherman Act. Without even labeling their disagreement with the Court's disposition a "dissent," Justices Hugo L.

---

Black, William O. Douglas, and Frank Murphy asked the reporter to record that they "think that the ruling of the district court was based on a ‘construction’ of the Sherman Act and that this Court therefore has jurisdiction to review the judgment."^[34] Jackson agreed with them but nonetheless filed a concurring opinion. "I agree," he said,

with the dissenting Justices that the decision of the District Court is "based" upon the construction of the Sherman Act. …

If the Court is to dispatch its business as an institution, some accommodation of views is necessary and, where no principle of importance is at stake, there are times when an insistence upon a division is not in the interests of the best administration of justice.

Such a case I consider this to be. To persist in my dissent would result either in affirmance of the judgment by an equally divided Court [Justice Wiley B. Rutledge did not participate in the decision, so a 4–4 vote would have resulted in affirmance] or in a reargument. …

However the case may be disposed of, reargument seems to be in order, and I believe that the practical advantages favor rearguing it before the Circuit Court of Appeals, where there is no doubt that all of the questions can be decided.

Under these circumstances, to persist in my dissent would seem a captious insistence upon my reading of a District Court's informal opinion as to which there is reasonable ground for difference.

I should not desire to appear committed to this case as a precedent. I concur in the result only because it seems the most sensible way out of our impasse in the immediate case.^[35]

Jackson's statement in Swift helps to illuminate Holmes's explanation of his dissent in Williams. Recall that Holmes began by avowing his allegiance to the general norm of unanimity: "When this case was here before I felt doubts, but deferred to the judgment of the majority, as I think one should when it does not seem that an important principle is involved or that there is some public advantage to be gained from a statement of the other side." He then justified departure from the norm: "But it seems to me that the present order is a mistake upon an important matter, and I am unwilling that it should seem to be made by unanimous consent." "Consent" is curious. One does not ordinarily think of Supreme Court justices voting in favor of a position as "consenting" to it. They join it, they support it, they believe in it—anything but consent. Yet that effectively is also Jackson's characterization of his position in Swift: "I should not desire to appear committed to this case as a precedent. I concur in the result only because it seems the most sensible way out of our impasse in the immediate case." By voting for the majority position, he is not "supporting" it (if he were, the case could be used as precedent), just "consenting." Why did he, as Holmes would not, "consent" to the per curiam? "To persist in my dissent

---

would result either in affirmance of the judgment by an equally divided Court or in a reargument." He certainly did not wish to affirm the trial judge's dismissal of the indictment. To that extent he agreed with the per curiam. He disagreed only about the appropriate venue for initial reargument—Supreme Court or circuit court of appeals—not a momentous matter. So he relented and concurred. He voted for the per curiam position, not because he agreed with it, but "because it seems the most sensible way out of our impasse."

Jackson's account of his behavior in Swift puts some meat on the bones of Holmes's "consent." A justice consents to a position with which he does not agree when some interest of the Court as an institution demands it. The interest in Swift was to grant reargument to the government when a trial judge, possibly erroneously, dismissed an indictment under the Sherman Act without clearly saying why. Had Jackson not consented to the per curiam the government's interest in enforcing the act and the judiciary's interest in supervising enforcement would have been sacrificed. The reason for consenting to the per curiam was the same reason the majority had used the designation in the first place.

The forces at work at the end of the nineteenth century only became more intense in the twentieth. In the 1950s the Court began to use per curiams to announce major decisions on matters of substance, far exceeding in notoriety and importance the earliest substantive per curiam or any of those that followed in its wake. At the same time, the practice of dissent made it plausible for the Court to use per curiams even in contentious cases. The concurrence of these two lines of force produced a remarkable handful of cases in the second half of the century that weave some further normative threads into the per curiam fabric.

The modern history of the major substantive per curiam begins, I believe, with Cooper v. Aaron^[36] in 1958. The story is well known. In Brown v. Board of Education^[37] four years earlier the Supreme Court had rejected the "separate but equal" doctrine of Plessy v. Ferguson^[38] as it applied to public schools, beginning the long and as yet unfinished task of ending school segregation. A few days after Brown the school board of Little Rock, Arkansas, announced that it was studying the establishment of an integrated school system and eventually came up with a plan to desegregate gradually, starting with the high schools in the 1957–58 school year. The plan was approved by a federal district court in 1956^[39] but met massive popular resistance and obstruction at all levels of state government. Governor Faubus famously called out the National Guard in September 1957 to stop nine African-American students from entering Central High School in Little Rock. The students were eventually escorted into school under the protection of combat troops of the regular U.S. Army, but public unrest and incidents in the school required the continuing presence of federal troops

---

and eventually the federalized National Guard. In February 1958 the school board asked the federal district court for permission to suspend implementation of the plan. The trial court granted the request on June 20,^[40] but the Court of Appeals for the Eighth Circuit reversed on August 18.^[41] In a special August term, the Supreme Court granted the board members' petition for certiorari, heard argument on August 28 and September 11, and affirmed the Eighth Circuit's refusal to stay implementation of the plan on September 12 in a unanimous half-page per curiam opinion. The Court explained the rush and lack of a full opinion:

In view of the imminent commencement of the new school year at the Central High School of Little Rock, Arkansas, we deem it important to make prompt announcement of our judgment affirming the Court of Appeals. The expression of the views supporting our judgment will be prepared and announced in due course.^[42]

The full opinion followed on September 29.^[43] It was not per curiam. Rather, it was written by Chief Justice Earl Warren and signed by every member of the Court (including Justice Felix Frankfurter, who filed a concurring opinion on October 6).

Not only did the justices drop the voice of the per curiam, but the report of the case in the official United States Reports virtually suppresses its existence. Instead of printing the per curiam as a freestanding opinion, the reporter prints it as a footnote to the full opinion (five pages into the report). (West Publishing Company handled the two opinions differently in its unofficial Supreme Court Reporter, printing the per curiam starting on page 1399 of volume 78 and the full opinion starting on page 1401; thus when Chief Justice Warren cited the per curiam in his footnote, only the West citation was available, not the official one!) By this time the Supreme Court per curiam had a long and honorable history. It had been used in numerous decisions on matters of substance, and major ones at that. So sweeping it under the rug this way seems odd. Yet it is one thing to make a deliberate decision, as I did, to look back at the history of Supreme Court per curiams and quite another to experience what was effectively only an oral tradition about their appropriate use handed down over the generations of justices since the 1890s. Tradition undoubtedly took the procedural fictions of the early per curiams at face value, though procedure had long strayed into substance, and no justice could possibly have found it useful or convenient to track the fiction into the obscure cases that expose its underlying reality and that computerized legal research has recently made accessible. It is also likely that the justices in Cooper felt that signature by each and every one of them was a more emphatic way of meeting Arkansas's challenge to the authority of the federal government. The members of the Court undoubtedly justified use of the per curiam in their initial

---

affirmance as a procedural necessity, well within the tradition of the fiction. Yet listen to Warren's account of what was at stake in the first sentences of his full opinion:

As this case reaches us it raises questions of the highest importance to the maintenance of our federal system of government. It necessarily involves a claim by the Governor and Legislature of a State that there is no duty on state officials to obey federal court orders resting on this Court's considered interpretation of the United States Constitution.^[44]

It is only in retrospect, on scrutiny of the actual history of per curiams, that it becomes possible to assure the justices that a per curiam would have been appropriate in the full opinion as well and that the per curiam they did use need not have been swept under the rug.

Once Cooper was decided, however, per curiam was lifted from obscurity, and oral tradition was supplemented by the prominence of a decision on amajor matter engaging the interests of the Court as an institution, a decision rendered per curiam and subject to efficient recollection. That was the state of affairs when the Court confronted the effort of the United States to enjoin the New York Times and the Washington Post from publishing the Pentagon Papers in 1971.

New York Times Co. v. United States (the Pentagon Papers case)^[45] shared three characteristics with Cooper v. Aaron: it was decided in a rush, the Court extended its term to accommodate the decision, and the majorities labeled their opinions, which were exceedingly short, "per curiam." But it was very different in one significant respect: where both the per curiam and the full opinion were unanimous in Cooper, the per curiam in the Pentagon Papers case was burdened with no less than six concurring opinions and one dissent, written by the younger Justice John Marshall Harlan and joined by Chief Justice Warren E. Burger and Justice Harry A. Blackmun. David Rudenstine has speculated that "the reason for a per curiam opinion seems to have been that the justices lacked the time before the summer recess to hammer out a majority opinion they could endorse."^[46] But in Cooper the Court delayed going to press with the full opinion, first publishing the per curiam disposition of the judgment below. It could have done likewise in the Pentagon Papers case. That it did not had a supremely important consequence for the modern history of Supreme Court per curiams: in Cooper the justices could justify the per curiam as a procedural necessity, while the Pentagon Papers case was clearly a decision on substance without any procedural justification. All six justices who joined the per curiam could have signed it as nine justices did in Cooper. The per curiam in the Pentagon Papers case thus presents an unabashedly substantive per curiam without a shred of procedural justification and without the consolation of unanimity. None

---

of this was now concealed by an oral tradition; all of it had entered selfconscious history.

The stage was set for Buckley v. Valeo,^[47] a scant four years after the Pentagon Papers case, in which the Court held unconstitutional several provisions of the Federal Election Campaign Act of 1971,^[48] notably those limiting soft money and self-financing by candidates. Like Cooper and the Pentagon Papers case, Buckley was decided under tremendous (and, the dissenters insisted, gratuitous) time pressure.^[49] Unlike those cases, however, the per curiam in Buckley was not short: it occupies 136 pages in the official reports. It was also not without dissent: five justices—Chief Justice Burger and Justices Byron R. White, Thurgood Marshall, Blackmun, and William H. Rehnquist—separately concurred in parts of the per curiam opinion and dissented from other parts.

In contrast to the Pentagon Papers case, no justice dissented outright. Whatever disagreements each of the dissenters had with the per curiam— and there were many—all but Chief Justice Burger agreed with the Court's disposition of the lower court's judgment—affirmed in part and reversed in part." Under the shred that is left of the norm of unanimity, each of the justices who agreed with the disposition was compelled to concur in part with the per curiam, no matter how thoroughly they may have disagreed with its reasoning. Interestingly, Chief Justice Burger also concurred in part, even though he disagreed with the partial affirmance (he thought the act had created a comprehensive and interlinked scheme of regulation, the invalidation of parts of which should render the entire scheme void). Nevertheless, he overcame his objections rather than dissent, one supposes, from the Court's invalidation of parts of the act. In the Pentagon Papers case the three dissenting justices did not agree with the Court's disposition, however much they ultimately might have come to agree with its reasoning had the case been tried below in the ordinary manner and a record created, as they urged.

The trio of major modern per curiams—Cooper v. Aaron, the Pentagon Papers case, and Buckley v. Valeo— provide a model of judicial behavior in response to a majority's decision to designate an opinion per curiam. A justice who agrees with the per curiam's disposition of the lower court's judgment ought to join or concur with the per curiam to the extent that he agrees with it and dissent to whatever extent he does not. That is true as well for ordinary signed opinions, but even more true for a per curiam. Despite the disintegration in our age of the norm of unanimity, a per curiam designation, in one of its modes, is at least a rhetorical call for unanimity where a decision implicates the interests of the Court as an institution in a major issue affecting the frame of governance. In Cooper the issue was the contumacy of a state toward the authority of the federal government;

---

in the Pentagon Papers case it was the overreaching of the executive; and in Buckley it was the constitutionality of an act of Congress threatening the integrity of the process by which the people chooses its government. That each of the trio was decided in an atmosphere of urgency or even crisis only underscores the need for as much unanimity as possible. The reason the Court ought to strive for unanimity is exactly the reason for a per curiam: events demand that the Court speak with one voice and that the voice it speaks with be its own.

… TO BUSH V. GORE …

The Supreme Court issued not one but two per curiams in the course of the battle over the Florida recounts. Only the second was in Bush v. Gore, on December 12. The first was in Bush v. Palm Beach County Canvassing Board,^[50] a related case that the Supreme Court decided eight days before, on December 4.

As other essays in this volume have described in detail, the Palm Beach County case arose out of the protest phase of the Florida struggle, before certification of the election results by the Florida secretary of state, Katherine Harris. Gore and the Florida Democratic Party had challenged Harris's refusal to include late amended returns from several counties in her statewide certification. Wrestling with a rather jumbled Florida Election Code, the Supreme Court of Florida had ordered Harris to accept late returns through Sunday, November 26. Bush filed a petition for certiorari to review the Florida decision. The U.S. Supreme Court granted certiorari to decide whether the Florida Supreme Court had violated the U.S. Constitution or a federal statute providing that a state's resolution of any dispute over the choice of electors by December 12, using laws in place before November 7, would be treated as conclusive when the votes for the Electoral College were counted in Congress. After reviewing the opinion of the Florida Supreme Court, the U.S. Supreme Court found that there was considerable uncertainty as to the precise grounds for the Florida court's decision, vacated its judgment, and remanded the case to the Florida Supreme Court to clear up the uncertainty. The per curiam opinion was unanimous.

The Court's use of the per curiam in Palm Beach County was completely appropriate according to the norms implicit in the line of cases beginning with Cooper v. Aaron. Not only was the decision unanimous, but it was also a response to a great occasion of state engaging the interests of the Court as an institution. The Court's decision would affect the choice of a president; it would chasten and if necessary correct the role that the highest court of a state had chosen to play in the process. If Buckley v. Valeo was per curiam, then certainly Palm Beach County could be as well.

---

Four days later the Florida Supreme Court handed down its decision in Gore v. Harris.^[51] The case arose from the contest phase of the struggle, after Secretary Harris's certification of the statewide results. Gore and Lieberman had sued Harris in Leon County Circuit Court, asserting that Harris and the State Elections Canvassing Commission had erroneously included illegal votes and failed to include legal votes in their November 26 certification. Circuit Judge N. Sanders Sauls rejected the claim. Gore and Lieberman appealed. In a 4–3 decision, per curiam(!), the Florida Supreme Court agreed with Gore and Lieberman that they were entitled to completion of a manual count of the undervotes in one county, Miami-Dade, but agreed with Harris that ultimate relief would require counting undervotes in all the counties where the undervote had not been subject to manual tabulation.

This was the state of the play when Bush and Cheney asked the U.S. Supreme Court to stay the mandate of the Florida Supreme Court in Gore v. Harris. On December 9 the Court stayed the mandate but also granted certiorari, and on December 12 it rendered its decision. With five justices joining, the Court held, per curiam, that the Florida Supreme Court's remedy violated the Equal Protection Clause (and possibly the Due Process Clause) of the U.S. Constitution two ways—by ordering the inclusion of votes counted according to variant standards in different counties and by ordering a statewide recount with minimal procedural safeguards. That conclusion certainly warranted reversal of the Florida court's judgment, which is precisely what the per curiam did. But it would also warrant instructions to the Florida court to conduct a statewide recount according to uniform standards with adequate procedural safeguards. That the per curiam did not do, because the five justices joining it gleaned from the Florida Supreme Court's opinion in Gore v. Harris that under Florida law any count or contest designed to lead to a conclusive selection of electors must be completed by December 12—the very day of the U.S. Supreme Court decision. Hence, the U.S. Supreme Court reversed the judgment of the Florida Supreme Court and remanded the case, in the time-honored formula of respect for the courts of another sovereign, "for further proceedings not inconsistent with this opinion."

The per curiam in Bush v. Gore was well within the tradition established by the trio of per curiams beginning with Cooper v. Aaron, for the same reason as the per curiam in Palm Beach County. But it was appropriate for an additional reason: Bush v. Gore was a case returning to the Court for further action, not technically, to be sure, but in reality. If Palm Beach County was per curiam, it made sense for Bush v. Gore to be per curiam as well. Furthermore, the presence of dissents in Bush v. Gore does not at all detract from the propriety. Nor does the majority's use of the per curiam detract from the propriety of the dissents. Minorities on the Court have long seen

---

fit to disagree with the voice of the Court, and the voice of the Court to disagree with minorities, ever since Holmes in 1909.

What they have not done, however, is dissent when they agree,^[52] exactly what Souter and Breyer did in Bush v. Gore. Souter, whom Breyer joined, agreed with the per curiam that the Florida Supreme Court's remedy violated the Equal Protection Clause. He disagreed only with the majority's conclusion that Florida law required closure by December 12. He believed that the real deadline was December 18, the date set for the meeting of electors. Thus neither Souter nor Breyer disagreed with the Court's disposition of the Florida court's judgment, which was "reversed and remanded," only with the palaver accompanying it.

And palaver it was! The sole impact the majority's understanding or misunderstanding of Florida law had on their opinion was that it dissuaded them from instructing the Florida court to conduct a recount consistent with equal protection. As Nelson Lund points out in this volume (chap. 7), the Florida Supreme Court was free to chide the U.S. Supreme Court for getting wrong what they said in Gore v. Harris, or, if they felt that the U.S. Supreme Court got it right, to change their minds about Florida law. The U.S. Supreme Court is completely powerless to dictate Florida law to the Florida Supreme Court, and they did not do so in Bush v. Gore. They did not order the Florida court not to conduct a recount; they simply declined to instruct it to conduct one, as a consequence of their understanding, drawn from the Florida Supreme Court, of Florida law.

There was simply nothing in the per curiam's disposition of the Florida judgment inconsistent with Souter's and Breyer's wish for a recount consistent with equal protection. There was nothing in the disposition to prevent them from concurring in part with it and dissenting in part, exactly as a majority of the Court did in Buckley v. Valeo. It is what they should have done.

… AND FINALLY PRAISE

I want to say one final word—a discordant note of praise for all the justices. They performed under astonishing pressure as a result of the gravity of the events in which they were participating and the stark calendar dictated by the United States Constitution. All their opinions, agree with them or not, are cogently argued and tightly written. If I criticize Justices Souter and Breyer for not concurring in part with the per curiam and dissenting in part, it is on the basis of much reflection and intense discussion with extraordinarily able colleagues, reflection and discussion that circumstances did not permit the justices. I do believe that Souter and Breyer were the only justices on the Court who plainly stayed true to their jurisprudential colors. I do not, however, fault any of the others, especially those who spoke

---

through a voice not their own but the voice of the Court. Had Justices Souter and Breyer had the model of Buckley v. Valeo in mind, they might well have decided to join that voice, concurring in part and dissenting in part. No one can say. But on the basis of the history of the Court's use of per curiams, one can say that the majority's use of a per curiam was abundantly appropriate, indeed the right way.

NOTES

I would like to thank Rich Friedman, Nelson Lund, John McGinnis, Robert Post, and Michel Rosenfeld for their invaluable help and encouragement.

---

---

---

9. NOTES FOR THE UNPUBLISHED
SUPPLEMENTAL SEPARATE OPINIONS
IN BUSH V. GORE

Burt Neuborne

These opinions were prepared for discussion in a constitutional law class. They are, of course, wholly fictitious; I have attempted to capture the voice of an opinion, without the use of the usual legal citations and technical arguments. They are designed to explore the legal issues raised in the actual opinion and to provoke discussion on judicial role. They are also designed to rebut the notion that the decisions of the various judges can only be explained by efforts to advance a favored presidential candidate.

O'Connor and Kennedy, J.J.:

We write separately to explain why we believe that the Florida legislature adopted the federal December 12 "safe harbor" provision for certifying presidential electors as a firm, nonextendable deadline for completing the count of ballots for presidential electors and why it was permissible, indeed necessary, for us to have decided that very important question of Florida law. We take the extraordinary step of issuing this supplemental opinion in response to suggestions that our decision to join the Court's stay of the Florida hand recount and our decision to join the Court's per curiam opinion were unprincipled exercises of raw power designed to assure the election of a favored presidential candidate. Ordinarily, we would not respond to such criticism. In the unique circumstance of this case, however, we believe that respect for the integrity of the judiciary warrants a more complete explanation of our reasoning than was possible in the difficult circumstances surrounding the release of the Court's per curiam opinion.

We understand the intense disappointment felt by persons who argue that Vice President Gore not only won the national popular vote by almost 400,000 votes, but was denied the opportunity to demonstrate that a hand recount of the so-called undercounted Florida ballots would have shown him to be the winner of Florida's crucial 25 electoral votes. We also understand that reasonable persons may harbor different opinions concerning

---

the legal issues raised by Bush v. Gore, including the crucial questions of (1) whether the December 12 safe harbor date had been adopted by the Florida legislature as a firm deadline; (2) whether we were authorized to decide that question of Florida law; and (3) whether it was permissible for us to have stayed the manual recount ordered by the Supreme Court of Florida. We write to make clear that we engaged in the same process of statutory interpretation as did the Florida Supreme Court in its effort to resolve ambiguities in the Florida statutory scheme and that we—and every other judge who sought to perform his or her judicial duty in these cases— acted in accordance with our views of the law and not as politicians wearing black robes.

We begin with the relatively uncontroversial but crucial proposition that any hand recount of the so-called undercounted ballots in Florida that tallied "dimpled" ballots in one county but rejected them in the neighboring county or that tallied ballots with partially punched holes in one county but required a complete punch in another county would apply such divergent ground rules in measuring a valid vote as to deny Florida voters the equal protection of the laws within the meaning of the Fourteenth Amendment. Seven members of this Court appear to have endorsed that proposition. Requiring minimal standards of clarity and uniformity in connection with counting the ballots in a single election not only assures that all votes weigh equally but also provides an important prophylaxis against efforts by local officials to manipulate the application of an unduly discretionary standard for or against a particular candidate. In analogous areas of the law, we have often invalidated government programs that vest an impermissible level of discretion in local officials to make politically charged judgments in the areas of the First and Fourteenth Amendments.

If, as we believed, the standards governing the statewide court-ordered hand recount mandated by the Florida Supreme Court were fatally vague and nonuniform, the difficult decision to stay the recount was certainly defensible. A statewide recount carried out pursuant to unconstitutionally vague and nonuniform tabulation standards simply could not have resulted in a legally valid outcome. Unless stayed, the vote count as ordered by the Florida Supreme Court could only have culminated in a legally meaningless result fraught with confusion and misunderstanding.

Once we determined that it was impossible to reach a legally acceptable result using the vague and nonuniform vote-counting standards announced by the Florida courts, we were faced with the choice of whether to remand to the Florida Supreme Court, once again, with instructions to attempt to fashion an appropriate uniform standard under Florida law, or to enforce the December 12 safe harbor deadline on the choice of presidential electors that we understood to have been imposed by the Florida legislature. Ordinarily, we would have followed the course urged by Justices Souter and

---

Breyer, as well as by Justices Stevens and Ginsburg, and remanded to the Florida Supreme Court in order to permit the highest court of that state to make the decision as a matter of Florida law. Given the unique time constraints of this case, however, it would have been impossible for the Florida Supreme Court to have acted in time to place Florida's electors within the federal safe harbor. Thus, if we believed that the Florida legislature wished to impose a December 12 deadline as a matter of Florida law in order to ensure that Florida's electors would not be subject to congressional challenge, we had no choice but to act immediately.

In fact, the issue before us was quite similar to the issue that confronted the Florida Supreme Court in its initial decision on whether to extend the deadline for certification in order to permit hand recounts in four Florida counties at the request of Vice President Gore. The Florida court weighed the text and policies underlying the statutory seven-day deadline against the text and policies underlying the statute authorizing hand recounts and found that the Florida legislature probably intended the deadline to give way to the recount, as long as the ultimate safe harbor certification of electors was not imperiled. Our refusal to join in the separate opinion of Justices Scalia and Thomas and Chief Justice Rehnquist is premised on our belief that in making that judgment the Florida Supreme Court was acting within the parameters of legitimate judicial behavior in interpreting ambiguous or conflicting statutory provisions. We need not agree with the interpretation of Florida law advanced by the Florida Supreme Court to recognize that, under our federal system, the choice of interpretive philosophies adopted by a state's judicial organs is none of our business. We therefore reject the notion implicit in Justice Scalia's opinion that state courts in interpreting state statutes governing the selection of presidential electors are obliged to adopt a strict constructionist philosophy under Article II, section 1, of the U.S. Constitution. While Article II, section 1, grants power to state "legislatures" to control the selection of presidential electors, we read the clause as meaning the state "legislature" as it operates in the sovereign political and legal matrix chosen by each state. Any other reading would federalize both the state legislature and the state courts in cases involving the selection of presidential electors in ways that are deeply inconsistent with the Constitution's fundamental decision to treat the states as sovereign political entities.

Our decision required us to weigh the text and policies underlying a firm adherence to the December 12 safe harbor deadline against the text and the undeniably important policies underlying a continuation of the effort to count every vote. Thus we too were forced to make an educated guess about whether the Florida legislature intended the December 12 deadline to give way to additional efforts to complete the recount. This time, however, the significant concern with imperiling the selection of Florida

---

electors identified by the Florida Supreme Court as the outside limit for judicially extending the certification deadline was in play. If the certification of Florida electors had not been definitively settled by December 12 and if a hand recount had resulted in a judicial determination that Vice President Gore was the winner of the Florida election, it appears that two rival slates of Florida electors—one certified by the state legislature and governor and one certified by the Florida courts—would have been presented to the Electoral College. The first tie-breaking mechanism designed to choose between the rival slates, selection by both houses of Congress, probably would have failed, as the newly elected Senate likely would have voted 51–50 for the slate pledged to Vice President Gore, and the newly elected House of Representatives likely would have voted narrowly for the slate pledged to Governor Bush. The next tie-breaking mechanism, gubernatorial certification, probably also would have failed, since the imprimatur of the governor of Florida would have been of questionable validity because it would have been affixed in violation of the order of the Florida Supreme Court. If both tie-breaking mechanisms failed, the Electoral College would have met without electors from Florida, raising the question of whether an absolute majority of 270 electors is required to elect the president or whether a majority of the electors actually appointed is sufficient. If 270 is required, the presidential election would be thrown into the House of Representatives, where the delegations would ballot by state, with the probable winner being Governor Bush, since the Republican Party is in the majority in twenty-eight states. If a majority of the appointed electors is required, Vice President Gore would be the winner with 267 electors. Suffice it to say that there is no clear answer to the issue, since the text of the Twelfth Amendment may be read either way. Thus, if the December 12 safe harbor deadline had been breached, the nation would have been swept into the wholly uncharted waters of a genuine constitutional crisis. After considering the ambiguous Florida text and the observations of the Florida Supreme Court and the participants in the litigation, we believed that the Florida legislature wished to make such a divisive potential constitutional crisis impossible by imposing a firm December 12 deadline on the certification of Florida electors that would make it impossible to challenge the certified electors in Congress. After reviewing the same material, Justices Souter and Breyer appear to believe that Florida would have risked a congressional challenge in order to assure that every vote was counted. That is a plausible reading of Florida law. We believe, however, that it is even more plausible to believe that Florida's legislature wished to establish a firm deadline of more than five weeks after the election in order to cut off the possibility of the nightmare constitutional scenario described above. Ideally, of course, we would have left the question to the courts of Florida. December 12 was upon us, however, making it necessary for us to decide

---

the issue immediately in order to prevent the de facto breaching of the safe harbor deadline. Finally, although we acknowledge that contrary practical pressures were hydraulic, our remand to the Florida Supreme Court gave that court an opportunity to challenge our reading of Florida law in the event we misunderstood the importance of the December 12 deadline. Neither Vice President Gore nor the Florida Supreme Court acting sua sponte questioned the accuracy of our effort to apply Florida law.

In short, we believe that reasonable persons can disagree with the initial judgment of the Florida Supreme Court that Florida law authorized an extension of the initial certification beyond the statutory deadline. We also believe that reasonable persons can differ with our judgment that Florida law forbids a further extension beyond the December 12 safe harbor deadline. Indeed, the essence of judging in both cases involved good faith efforts to resolve questions concerning the meaning of statutory and constitutional text about which highly competent and well-meaning people can reasonably differ. In reaching decisions in both cases, the participating judges in both Florida and Washington were required to balance the mandate of adherence to textual command against the difficulty of knowing precisely what the command is, using appropriate methods to break ties when good faith doubt about the content of a textual command exists. We are confident that arguments of principle dictated the outcomes of both courts.

Souter and Breyer, J.J.:

We join in recognizing that principle, not political expediency, was the motivating force behind the decisions of the Supreme Court of Florida and the decisions of the justices of this Court. Irresponsible criticism of judges by persons who assert that outcomes are dictated by partisan loyalty misunderstands the complexity of the judicial process and badly disserves the rule of law. But acknowledging the good faith of the majority's per curiam opinion should not immunize the opinion from the scathing criticism it deserves.

To begin with, the per curiam badly mishandled the crucial equal protection issue. As the separate opinion of Justices O'Connor and Kennedy makes clear, a finding that the hand recount was proceeding under conditions that violated the Fourteenth Amendment was critical both to the grant of the stay and to the ultimate judgment to vacate the opinion of the Florida Supreme Court mandating a statewide hand recount of potentially uncounted presidential ballots. It was, however, both inaccurate and misleading to have viewed the court-supervised hand recount as though it were the genesis of the equal protection problem. In fact, Florida had violated the Equal Protection Clause by deploying voting machines of widely disparate accuracy throughout the state. In some counties, optical scanning machines were used with an undercount error rate of one in three hundred.

---

In many other counties, voting machines based on older punch-card technology were used with an undercount error rate of 1.5 percent and higher, to say nothing of an exponentially higher potential for confusion and double voting. Thus, voters in counties with outdated technology were five times more likely to have had their votes undercounted and far more likely to have their votes inaccurately recorded than voters in counties with newer machines. Moreover, if, as has been alleged, the outdated voting machines were concentrated in counties with low per capita income and high rates of racial minority voters, the disparity takes on an even greater significance under the Fourteenth and Fifteenth Amendments and section 2of the Voting Rights Act.

Whether or not use of the older machines, standing alone, would have raised questions of electoral fairness, surely such mechanically induced electoral inequality cannot be squared with the per curiam's intense concern with Fourteenth Amendment voting equality. Thus, when Florida courts intervened in an effort to cure the unconstitutional outcome of the unequal vote counting machinery, it was clearly improper for this Court to have blocked the judicial remedy in the name of equality while allowing the underlying Fourteenth Amendment violation to have gone unremarked. The Court was obliged to either mandate a federal remedy for the ongoing equal protection violation that would have superseded any state imposed deadline on counting votes or leave the less-than-perfect state remedy in place as the lesser of two constitutional evils. The one indefensible result was to have blocked the Florida courts from seeking to remedy the blatant federal violation of electoral equality caused by the unequal propensity of differing voting machines to undercount presidential ballots. The result is particularly indefensible because the failure of the Florida courts to have imposed a uniform standard for evaluating undercounted ballots was, in large part, caused by this Court's expression of concern in our initial unanimous per curiam opinion that Florida courts were not exhibiting sufficient deference to the will of the Florida legislature. If, in the teeth of the first per curiam, the Florida courts had adopted a uniform standard, they would have been accused of judicial legislation. Having refrained from judicial activism by adhering to the Florida legislature's textual standard in deference to the apparent wishes of this Court, it is an extraordinary whipsaw to be told that greater judicial activism was necessary to avoid violating the Fourteenth Amendment. Imposing statewide uniform standards risked invalidation under Article II, section 1, and Title 3 U.S.C. §5. Not imposing uniform standards risked invalidation under the Fourteenth Amendment. An observer might conclude that the Florida Supreme Court's effort to secure a hand recount was doomed no matter what the court did. Finally, the suggestion that the per curiam decision to terminate a judicially supervised statewide hand recount of undercounted ballots was

---

consistent with decisions invalidating the grant of standardless discretion to local police officials in the First Amendment area ignores the fact that the recount was being conducted, not by local canvassing boards subject to political influence, but by the Florida judiciary operating in two-judge teams, reporting to a single judge for the resolution of contested ballots. While the void-for-vagueness doctrine places some limits on a delegation of standardless discretion to judges, the Florida statutory standard falls well within the bounds of our traditional view of judicial power.

Nor does the separate opinion justify the per curiam's insistence on deciding that the Florida legislature had imposed a firm deadline of December 12 for the counting of ballots. All agree that issues of state law must, ordinarily, be finally determined by state courts. At most, this Court would have been authorized to remand the matter to the Florida Supreme Court to resolve the state law issue. The suggestion that by remanding the matter to the Florida Supreme Court at 11:00 P.M. on December 12, this Court provided the Florida court with a meaningful opportunity to disagree with this Court's reading of Florida law is wholly unpersuasive. If such an option was intended, the language of the per curiam rendered it terminally opaque.

Moreover, the effort to posit a necessity-based exception to our inability to resolve state law issues fails to persuade for two reasons. First, it was this Court's improper issuance of a stay that brought the case to the brink of December 12 without the completion of the statewide hand recount; and, second, the existence of a federal Fourteenth Amendment issue justified, if it did not mandate, the lifting of the December 12 deadline in order to complete the ballot count in accordance with Fourteenth Amendment standards.

Thus, while the separate opinion reveals a principled explanation for the action of the swing justices, it fails to provide a persuasive refutation of the charge that the opinion is extremely bad law, however well intentioned it may have been. This Court may have elected a president by one vote, but it has not written a defensible opinion.

Stevens and Ginsburg, J.J.:

The failure of the separate opinion of Justices O'Connor and Kennedy to provide a wholly persuasive legal justification for the action of the per curiam in terminating the hand recount of undercounted presidential ballots in Florida poses a profound question about the role of this Court. As the separate opinion of Justices Souter and Breyer demonstrates, the separate opinion of Justices O'Connor and Kennedy, while making it clear that arguments of principle can be deployed in support of the majority's position, fails to persuade on the issues of voting equality, standardless discretion, and respect for the power of Florida courts to interpret Florida

---

law. In fact, the real basis for the decision of the swing justices to join the per curiam opinion does not appear to have been the persuasive power of its legal reasoning but consequentialist concerns over the lack of an orderly process to determine which, if either, of two rival slates of electors would speak for Florida in the Electoral College if the statewide hand recount, using uniform standards, had been permitted to proceed to a conclusion favoring Vice President Gore. Faced with the prospect of dueling slates of electors and lacking a clear mechanism for choosing between the rival slates, the swing justices appear to have subscribed to a weakly reasoned opinion in order to hasten the emergence of a clear, formal winner of the 2000 presidential election. In effect, the swing justices appear to have projected their concerns over the risks inherent in a potential constitutional crisis onto the Florida legislature and found that the Florida legislature had mandated a firm deadline of December 12, the federal safe harbor date, in order to prevent the emergence of rival slates. To make the December 12 deadline stick, the swing justices were obliged to subscribe to a view of the Equal Protection Clause that invalidated an imperfect judicial cure for Florida's unequal mechanical ballot counting, leaving the underlying disease untreated, and to a view of federalism that preempted Florida's power to decide for itself whether to continue the counting after December 12.

The legitimacy of such judicial behavior turns on whether the legal arguments for or against the recount were in rough equipoise. If a judge finds herself confronted with powerful legal positions pulling in different directions, there is nothing wrong with using consequentialist reasoning to help break the legal tie. If, however, legal reasoning is used merely to provide a veneer of legitimacy for a judicial decision that is wholly based on consequentialist reasoning, the judicial process impermissibly merges into the political process. We believe that, in this case, the traditional legal arguments tilt so strongly against the per curiam's equality position, and so strongly against its federalism position, that it cannot be defended as an effort to break legal ties in favor of avoiding potentially disastrous consequences. Rather, it appears to be an example of well-meaning justices so concerned with potentially disastrous consequences that they are prepared to deploy weakly reasoned law to avoid those consequences.

Although we believe that it is inappropriate to bend the law in such a way even to avoid serious potential harms, it must be noted that such an effort to shield the nation from the risks inherent in dueling slates of electors and an indefinite prolongation of uncertainty over the winner of the 2000 presidential election is a far cry from the charges of narrow partisan decision making that have been leveled by many observers. At worst, the decision of the swing justices to join the Court's per curiam opinion is a misguided attempt to prevent the nation from sailing onto uncharted constitutional seas, not an exercise in personal self-indulgence.

---

In the particular circumstances of this case, we believe that the effort by the swing justices to turn the Court into a deus ex machina to spare the nation the agony of a protracted fight over the Florida electors was particularly unjustified. First, the per curiam sacrificed the value of democratic legitimacy to the value of stability and predictability. Preventing the completion of a count of the ballots that might well have determined the presidency is an enormous intrusion by judges into the democratic process. The swing justices apparently believed that the risk to democracy caused by instability and unpredictability outweighed the cost to democracy of failing to count all the votes in Florida. While such a belief was doubtless honestly held, it is uncomfortably close to arguing that it is acceptable to burn down a village in order to save it.

Second, the per curiam appears to have badly overestimated the likelihood of a constitutional crisis. It is true that the Florida legislature was poised to select a post—December 12 slate of electors pledged to Governor Bush that would have been presented to the Electoral College as an alternative to any slate pledged to Vice President Gore that might have emerged from the statewide judicial recount. But democratic mechanisms designed to choose between the rival slates were in place. Both slates would have been presented to Congress. If the Gore slate had received more votes in Florida pursuant to a fair recount procedure using uniform standards pursuant to a process that enjoyed the imprimatur of this Court, it is quite unlikely that either house of Congress would have insisted on recognizing the nondemocratic slate. Thus, by short-circuiting the operation of the democratic process, the swing justices prevented the mechanisms of democracy from defending themselves. With friends who exalt stability and predictability over democratic legitimacy even in settings where democracy can defend itself, democracy does not need enemies.

Third, in purely consequentialist terms, the per curiam actually weakened crucial American institutions. The presidency of George W. Bush was denied the democratic legitimacy that would have flowed from a victory in the statewide recount. He takes office by the grace of a single vote in the Supreme Court, having lost the nationwide popular vote by almost 450,000 votes and having possibly lost the Florida popular vote as well. In addition, the Supreme Court weakened its capacity to act as a forum of principle in unpopular cases. At best, the per curiam will be perceived as bad law; at worst, it will be perceived as bad faith. Even the rule of law was weakened. While the willingness of the nation to abide by the razor-thin vote of the Supreme Court is testimony to the enduring power of our respect for the rule of law, when, as here, formal law departs from the underlying principles that give it moral strength, respect for law is inevitably weakened. In this case, by using law to prevent the operation of the democratic process in a highly public display of power, the per curiam has forced the nation

---

to accept the formal rule of law but only at the expense of divorcing the law in this case from its moral underpinnings as the defender of democracy.

Scalia and Thomas, J.J., and Rehnquist, C.J.:

The struggle by our colleagues to decide whether particular exercises of judicial power are, or are not, legitimate highlights the fundamental weakness in the analytic approach of all three separate opinions. Each separate opinion explores the limits of judicial power in going beyond or ignoring the plain meaning of textual commands. The separate opinion of Justices O'Connor and Kennedy argues that the decisional processes of both the Florida Supreme Court and this Court were essentially identical, with each court involved in a free-floating decision whether to abide by a statutory deadline or to continue the vote count. According to the O'Connor-Kennedy opinion, the Supreme Court of Florida elected not to be bound by the statutory deadline, finding the judicial power to continue the recounts beyond the seven-day certification deadline in the values of the Florida constitution and in the alleged collision with other statutes authorizing a hand recount. Conversely, argue Justices O'Connor and Kennedy, this Court in its per curiam opinion elected to be bound by the December 12 statutory deadline at the expense of continuing the hand recount of undercounted ballots. According to Justices O'Connor and Kennedy, both exercises in reading text are within the range of legitimate judging, even though they come out 180 degrees apart. But, as described by Justices O'Connor and Kennedy, neither exercise of judicial power is legitimate. It is currently fashionable to describe exercises in reading text judicially as exercises in discretion, with most cases providing a court with a choice between cleaving closely to text or departing from text in search of a more perfect justice. In settings where text is so ambiguous as to provide no real guide to a court, such an exercise may well be necessary. But where, as here, the text is clear, courts simply have no power to depart from it, even in defense of important values such as democracy and fairness. The widely disparate views of the six justices who have declined to join this opinion demonstrate that once judges are embarked on the kind of open-ended search for justice that licenses them to disregard controlling text, the judicial process becomes hopelessly subjective. It is an irony that Justices O'Connor and Kennedy have chosen to defend an opinion designed to curb overly broad discretion in the counting of ballots by ascribing overly broad discretion to judges in the reading of text.

We concede that such a narrow conception of the role of a court confronted with a textual command is not widely held. Several generations of vigorous exercise of judicial power to operate as a full partner in the shaping of statutes have made it commonplace for courts to assume a duty to "perfect" statutory schemes. Ordinarily, it would be up to the Florida courts

---

to choose an interpretive philosophy. Were this a garden variety case, we might believe that the Florida courts, in ignoring statutory deadlines and inventing judge-made remedies, were acting beyond the scope of legitimate judicial activity, but that question would be for the people of Florida to determine. Cases involving the selection of presidential electors are not, however, garden variety cases. They are governed by a specific provision in the federal constitution, Article II, section 1, that grants power to the state "legislature" to control the selection of presidential electors. If we are to be true to that textual command, we must insist that state courts respect the primacy of state legislatures in selecting presidential electors. In one sense, the separate opinion of Justices O'Connor and Kennedy does just that by enforcing the Florida state legislature's wish that ballot counting stop on December 12. But the road to that result took them through a series of discretionary choices that are beyond the limited scope of judicial power.

The separate opinion of Justices Souter and Breyer takes the per curiam to task for ignoring the alleged equal protection violation caused by the use of voting machines with widely disparate error rates in tabulating ballots. Whatever the validity of such a Fourteenth Amendment claim, it is foreclosed in this case because no one raised it at any level. No claim of denial of federal equal protection was raised by Vice President Gore, either in the Florida courts, in the lower federal courts, or in this Court. If the parties choose to forgo raising federal claims as a matter of strategy in an effort to deprive this Court of jurisdiction, this Court can hardly be taxed with a failure to decide them.

Finally, the separate opinion of Justices Stevens and Ginsburg demonstrates the risks of going beyond the text in the search for justice. If, as they allege, consequentialist reasoning dictated the actions of the swing justices, it is difficult to separate the Court's thought processes from that of a legislature. If judging is merely a judicial policy preference dressed up as law, it does not deserve to be called adjudication. Thus, of the varying approaches presented by the separate opinions, only an approach that reads Article II, section 1, as commanding respect for state legislative policy preferences and denying state or federal judges an active role in the selection of presidential electors can be applied in a principled manner. The moment one goes beyond that strict legislative tether, judges assume a subjective and unjustified role in the selection of the president that cannot be squared with constitutional text.

Stevens, O'Connor, Kennedy, Souter, Ginsburg, and Breyer, J.J.:

We cannot allow these separate opinions to end without responding to Justice Scalia's insistence that Article II, section 1, imposes a single philosophy of judging on state courts confronted with a case involving the selection

---

of presidential electors. Justice Scalia argues that words in a statute often have a single meaning and that while judges ought not to depart from that meaning in any case involving text, they may not depart from textual plain meaning in any statutory case involving the selection of presidential electors. His separate opinion breaks down, however, at precisely the two points that his theory of judging often breaks down. First, words used in legal texts are notoriously ambiguous, making it almost impossible to insist that they carry an objectively knowable single meaning. Second, efforts to construe complex statutory schemes often reveal lacunae and disharmony that cannot be resolved by resort to literal readings. This case illustrates both problems.

Justice Scalia argues that the word "legislature" in Article II, section 1, plainly delegates plenary power to state legislatures to control the selection of presidential electors free from state judicial interference based on activist theories of statutory construction, or interpretations of the state constitution. While such a reading of "legislature" is plausible, it is far from the only possible reading. It is equally plausible to read "legislature" as that term is used in Article II, section 1, as a description of a political institution operating within a matrix of statutory and constitutional provisions that cause it to interact with other organs of state government. Why would the Founders, having just invented the idea of a legislature acting subject to judicially enforceable constitutional constraints, mandate that state legislatures be freed from state constitutional constraints when selecting electors? Moreover, how could such a deracinated reading of the term be squared with conceptions of state sovereignty that shield state institutions from being "commandeered" by the federal government? Forcing state courts to adopt a controversial judicial philosophy in cases involving presidential electors and untethering legislatures from their state constitutional moorings cannot be harmonized with a vision of the states as true sovereigns. Not surprisingly, despite Justice Scalia's confident insistence that the word "legislature" as used in Article II, section 1, must mean only what he says it means, six members of the Court declined to accept such a reading.

Similarly, Justice Scalia argues that the effort of the Florida Supreme Court to interpret Florida's election laws to determine whether the original certification deadline could be extended so as to permit precertification hand counts of the ballots, or whether power existed to order a postcertification statewide hand count of all undercounted ballots, violated Article II, section 1, and 5 U.S.C.§3because it constituted postelection judicial lawmaking in derogation of the power of the legislature to set preelection rules governing the selection of presidential electors. Under Justice Scalia's personal judicial philosophy, we have no doubt that the Florida court's aggressive efforts to construe the Florida Election Code against the background of an intense commitment to having each vote counted might appear

---

unduly activist. But nothing in Article II or the federal code entitles the members of the Supreme Court to impose a personal view about how statutes should be read by a state supreme court merely because presidential electors are involved. In fact, the Florida court acted well within the parameters of generally accepted norms of statutory construction in seeking to discern the meaning of the Florida election laws. In deciding the precertification issue, the Florida court was confronted with two statutory provisions that appeared to point in different directions—a seven-day deadline and an authorization of hand counts that, often, could not be completed in seven days. Justice Scalia insists that only one right answer existed—respecting the deadline. But, whatever one's views about what the right answer was, it is clear that the Florida Supreme Court was following a well-trodden path when it sought to harmonize two conflicting provisions of the Florida code.

Similarly, in deciding the postcertification issue, the Florida court was confronted with a general grant of remedial power with little guidance as to how the power should be used in specific situations. Once again, while Justice Scalia's narrow view of the judicial role would undoubtedly cause him to pause before imposing an innovative statewide remedy, it is clear that many courts would have done exactly what the Florida court did in seeking to forge a specific remedy armed only with general statutory authorization.

Thus, to make Justice Scalia's model work, we must be prepared to accept an arbitrary, highly undesirable reading of the term "legislature" in Article II, section 1, and to impose a single, narrow philosophy of statutory construction on all state courts in cases involving presidential electors. That we decline to do. Judging in hard cases is not reducible to a single objective formula. The best that we can do is to provide honest explanations for why we believe that text and precedent should be read in a particular way, provide a principled defense of that reading, and rely on the give-and-take of the adversary process, judicial debate, and public criticism to help us reach the best set of answers. We have widely different views concerning the best answers to the difficult legal issues raised in these cases. Those disagreements flow, however, not from efforts to advance a favored presidential candidate but from honest differences over judicial role, the relative weight of conflicting values, and divergent philosophies of reading text. There simply is no other way.