When the U.S. Supreme Court aborted Florida's statewide recount on Dec. 9, Justice Antonin Scalia explained, "The counting of votes that are of questionable legality does in my view threaten irreparable harm to [George W. Bush], and to the country, by casting a cloud upon what he claims to be the legitimacy of his election. Count first, and rule upon legality afterwards, is not a recipe for producing election results that have the public acceptance democratic stability requires." By smothering the recount before hearing the case and then declining to resolve the validity of the disputed ballots, the court concocted an even less confidence-inspiring recipe: Don't rule upon legality, settle the election first, and count afterward.
Two weeks later, reporters are counting the ballots, and Republicans are trying to stop them. The GOP argues that the press recounts, like the partial official recounts that preceded them, are subjective and inconsistent, with different reporters applying different standards and reaching different results. To some extent, that's true. But accusations of subjectivity and inconsistency are inherently self-limiting. To know that reporters are judging ballots subjectively, you need objective evidence: the ballots. To know that the identity of the winner varies according to the standard by which ballots are counted, you need a uniform understanding of what each standard is and how many ballots for each candidate would meet it. In short, you can't know how subjective the news organizations' recounts are, or whether differences among their methods significantly affect the result, until they have examined, categorized, and added up the ballots. If the debate over the official recounts exposed the limits of the naive view that everything can be known, the debate over the unofficial recounts is exposing the limits of the cynical view that nothing can be known.
A month ago, "Frame Game" observed: "When the number of ballots requiring interpretation exceeds the number of votes separating the candidates, determining who won is no longer a matter of investigation. It's a matter of interpretation." Based on Al Gore's specification and characterization of 13,000 contested ballots in two counties, I concluded, "Now that [Gore] has told us how many votes he would gain against Bush, depending on the standard by which those ballots are counted, the choice of president comes down to the choice of standard." But two premises of that analysis have changed. The statewide media recounts are targeting at least 60,000 and possibly as many as 180,000 ballots that were rejected by the official vote-counting machines. This enlargement of the pool of disputed ballots raises the possibility (a small one, in my estimation) that the number of cleanly punched votes separating the candidates will turn in Gore's favor and will exceed the number of ballots requiring interpretation. Furthermore, the purpose of the media recounts is to understand what happened, not to choose the president. This project doesn't require us to resolve objectively which vote-counting standard is the fairest. Having dispensed with that threshold question, we're free to investigate how many votes each candidate would gain or lose, depending on the standard by which the ballots are counted.
Bush's aides and supporters have tried in various ways to bury the recounts. On Dec. 7, Bush campaign lawyer William Scherer won a court order barring public inspection of Broward County's ballots until each of the 88 candidates on the ballot was notified. On Dec. 9, Rep. Jennifer Dunn, R-Wash., one of Bush's convention co-chairs, falsely asserted: "The Freedom of Information Act will not allow people to subpoena [Florida] ballots. Those ballots are going to be sealed right after the election." On Dec. 14, a week before Bush named her to his Cabinet, Gov. Christine Whitman, R-N.J., said the ballots should be sealed for a decade, lest Bush's legitimacy be undermined by subjective "interpretation" of the ballots. On Dec. 18, Bush spokesman Tucker Eskew accused the Orlando Sentinel, which was about to report a net gain of 130 votes for Gore among 6,000 machine-rejected ballots, of "mischief making." On Dec. 19, House Deputy Majority Whip Mark Foley, R-Fla., flanked by two Republican officials, showed up at a media recount in Broward County to charge that the ballots were "not really a public record" and that "further reporting" might "undermine the legitimacy of the presidency" by raising doubts about Bush's victory. Gov. Jeb Bush said of the media's review: "I've got a problem with it. … I don't know how human beings can divine the intent of another voter … by looking at a ballot." Secretary of State Katherine Harris advised county officials that although they were permitted to facilitate the inspection process by separating machine-rejected ballots from the state's 6 million already-counted ballots, they weren't required to do so. And in an ABC interview aired on Christmas Eve, former President George Bush said of the ballot inspection, "I'd like to see it go away."
Let's set aside the embarrassing suggestion by Whitman, Foley, Eskew, and others that we shouldn't disturb our national "healing" or peace of mind by unearthing unpleasant facts. The GOP's more substantial argument is that the variety of possible standards for judging ballots makes the recount a confusing and meaningless attempt to "speculate" and "divine" the intent of each voter. That's exactly backward. The media, unlike county canvassing boards, aren't mysteriously declaring each ballot a vote or a non-vote. They're dividing that process into two intelligible parts. First, they're organizing the ballots into precise categories: dimple with hole, dimple without hole, three-corner chad, one-corner chad, marked with an "X," etc. Then they're calculating how many votes each candidate would gain depending on which of these standards is applied. The Palm Beach Post and Tampa Tribune have already published figures showing different results based on different possible standards in early batches of reviewed ballots.
A media recount that yields eight possible statewide results based on eight different standards doesn't resolve "the truth" about who won Florida. But it does resolve eight truths, which is eight more than we have today. It can distinguish ballots that are utterly inscrutable from those that clearly convey voters' intentions, while describing to readers—as the Sentinel and Miami Herald have begun to do—exactly how some ballots, through duplicate marks or handwritten comments, manifest those intentions.
Republicans protest that Democrats are misrepresenting the recounts as a resolution of "the truth" about who won. They also complain that different recounts by 20 different news organizations and interest groups will yield different results. Both charges are true, but the second problem solves the first. The best way to puncture the pretense that one publication's count is perfect is to check it against another's. Reading articles in 20 newspapers about a campaign event, like seeing a sports replay from five camera angles, gives you a sharper picture of what happened. The object becomes more complete and distinct as each subjective presentation challenges others, exposing distortions and uncertainties. Why shouldn't the same principle apply to an examination of ballots? Knowledge, like a jewel, can't be perfected in a single stroke. It has to be chiseled from many angles.
The most enlightening scene from the recounts so far was the Dec. 19 encounter between Rep. Foley and Tom Fitton, the president of the conservative legal watchdog group Judicial Watch, at a media inspection of Broward County ballots. While Foley denounced the inspection, Fitton wisely replied, "The more information is out there, the better off we'll be." Fitton explained to USA Today that by conducting its own recount, Judicial Watch would be able to "match what others say against our figures." That kind of mutual scrutiny has already begun in Lake County, where the Sentinel conducted its recount. Republican monitors say the Sentinel included as many as 29 "Gore-Cheney" or "Gore-Nader" write-in ballots in Gore's tally. The Sentinel says that's false. I don't know which account is correct. But because both sides were in the room, at least I know that there's a dispute—and that it involves less than a quarter of the votes Gore picked up.
Ballots, while susceptible to interpretation, have objective limits. You can exaggerate dimples, but you can't argue with a cleanly punched hole or a cleanly filled circle. So while the recounts may confirm some preconceptions, they're likely to shatter others. Arguing before the U.S. Supreme Court, for example, Bush's attorneys complained that the official statewide recount omitted "overvote" ballots, which machines had rejected because they appeared to register votes for two different presidential candidates. Gore's attorney dismissed the suggestion that such ballots might, on further review, show clear intent to vote for one candidate. On Dec. 15, Wall Street Journal commentator John Fund protested that media recounts would be "terribly misleading" if they excluded these ballots in counties that voted for Bush. Four days later, after examining such ballots in pro-Bush Lake County, the Sentinel reported that 12 percent of them actually showed clear votes for one candidate—and that by giving Gore a net gain of 130 votes, they would have shaved Bush's statewide lead to 24 votes.
Further investigation will also unearth additional votes for Bush and expose ballots improperly counted for Gore. The Herald has discovered that at least 445 felons—three-quarters of whom were registered Democrats—voted illegally in Florida, as did at least 144 unregistered voters in Dade County. Bush will gain votes from non-postmarked overseas military ballots that have been declared legal since the statewide returns were certified. And the 567 net votes Gore picked up in Broward County, where Republicans accused Democratic officials of padding Gore's total with ambiguous ballots, are finally facing independent scrutiny. The Herald found that "the Broward canvassing board's standard of what constituted a vote was inconsistent. The news organizations examined several absentee ballots … that were not counted as votes even though they featured the same type of dimpled and barely pressed chads that the canvassing board declared as official votes for Gore or Bush on other ballots." The Post reported that if the board had required chad to be punched through in at least two corners, Gore would have gained only four net votes among the first batch of ballots, instead of the 155 he was awarded.
It's true, as Republicans say, that the recounts will be somewhat subjective, inconsistent, and inaccurate. But they can't be more subjective, inconsistent, or inaccurate than the certified returns on the basis of which Bush was declared president-elect. The precise, uniform, automated count Republicans keep claiming to have won never existed. The acknowledged rate of error in vote-counting machines like those used in Florida exceeds the margin of difference between Bush and Gore. Last Friday, the Herald documented apparent errors by 13 specific machines in Dade County. The state's certified "automatic" tally includes ballots that were hand-counted by officials in several Republican counties because machines had failed to tabulate them. And contrary to the Bush campaign's assurances, at least 16 Florida counties never recounted all their ballots. Eskew, Bush's spokesman, claims the Sentinel recount is fraudulent because overvotes are illegal. "To publish illegal votes as legal votes would be to mislead the readers and the public," he warned the paper. The Sentinel then pointed out that "ballots exactly like those rejected in Lake [County]—and now called 'illegal' by Eskew—were counted by canvassing boards in places such as Orange and Seminole counties and are now part of the certified totals."
Nothing in this world is purely objective. But for that very reason, no judgment, standard, or process can be disqualified just because it is, like everything else, partly subjective. What matters is the degree to which objective factors limit the role of interpretation in that judgment or process. We know that machines missed some ballots in Florida. We don't know how many ballots or how clearly they were marked. We can sit here all day exchanging opinions about machine counts, hand counts, chad, and dimples, or we can look at the most objective available evidence: the ballots. They may be hard to read, but they're all we've got. Will they tell us who won Florida? I don't know. All I know is that if we bury them, nothing will.