Did Al Gore win Florida? We'll never know. The problem goes deeper than the difficulty, cited by several analysts, of finding every lost ballot or computer glitch. Even if all the ballots were assembled before a team of neutral inspectors, the winner couldn't be objectively resolved. 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.
This is the latest twist in the unraveling of political truth. The elements of an election—which candidate is winning, what the voters think, whom they ultimately decide to vote for—used to seem definite. But the more closely they're examined, the less solid these elements—like electrons, quarks, and photons—turn out to be. The first casualty of scientific scrutiny was elite conventional wisdom, debunked by polling. Then poll "results" were found to be loaded with speculative assumptions. Then pollsters' erroneous projections of how the election would turn out were one-upped by erroneous Election Night projections of how the election had turned out. Then the standard by which those projections were judged inaccurate—the official vote tally—was corrected by machine recounts that identified previously uncounted ballots. Now the decisive ballots themselves are proving inscrutable. The ultimate certainty of politics—a marked ballot staring you in the face—turns out to be neither ultimate nor certain.
Florida's certified returns show George W. Bush beating Gore by 537 votes. In his complaint contesting those returns, Gore cites two groups of ballots sufficient to overturn Bush's margin. First,
4,000 ballots in Palm Beach County that were marked by the voter with an indentation but which were not (in most cases at least) punctured that the Palm Beach County Canvassing Board reviewed but did not count as a vote for any presidential candidate and which have been contested. If discernible indentations on such ballots were counted as votes, Al Gore and Joe Lieberman would receive more than 800 net additional votes.
9,000 ballots in Miami-Dade County that have not been recorded as a vote for any presidential candidate and which were never counted manually because the Miami-Dade County Canvassing Board prematurely ceased its manual count with only approximately 20 percent of the precincts counted. If these approximately 9,000 uncounted ballots result in the same proportional increase in net votes as the ballots that were counted by the Board before it stopped counting, these ballots would result in approximately 600 net additional votes for Gore/Lieberman.
Gore describes Miami-Dade ballots on which voting machines discerned no vote for president—he says there were 10,750 such ballots, including the 9,000 that weren't re-examined by hand—as "uncounted." Later he asserts, "The Palm Beach Board's uncompleted manual count resulted in a total of 8,222 uncounted votes." Referring to these ballots in his speech Monday night, Gore declared, "Many thousands of votes that were cast on Election Day have not yet been counted at all, not once." But how does Gore know the exact number of such ballots? Because they have been counted. They were counted as registering no vote for president—in effect, as ballots for none of the above.
Gore says these ballots should be searched manually for dents that might be construed as votes for one of the presidential candidates. His surrogates describe this process as giving each voter "the benefit of the doubt." But the logic is circular. To know whether this way of resolving a dubious ballot truly benefits the voter, you first need to know whether that voter intended to support a candidate. If she didn't, and you interpret a dent in her ballot as a vote for Gore, you're not counting her vote. You're changing it. The "benefit of the doubt" obscures the doubt of the benefit.
How exactly can the voter's intent be determined? When you punch the hole designating the candidate of your choice, the ballot instructions say you must detach the punched-out flap of paper—the "chad"—from the ballot. Otherwise, your punch isn't a vote. Some jurisdictions will count it as a vote if the chad is still hanging by one or two of its four corners. Others will still call it a vote if the chad is only slightly perforated and remains attached at three corners. If the voter has failed to punch out the chad completely, there's no objective way to decide which of these standards is the best measure of a voter's intent. But each of them, by specifying some number of detached corners, is more objective than the "standard" advocated by Gore. He proposes to call it a vote even if the chad isn't punctured at all.
Gore faults the Palm Beach board for using a "rigid rule" to weed out inadvertent dents. What was that rule? The board "failed to count ballots with indentations or dimples for a presidential candidate unless the ballot also revealed similar indentations, falling short of complete perforations, in other races." On its face, this rule seems a rational way of measuring the voter's intent: A dent doesn't "mean" a vote unless that voter generally supplied dents where votes were requested. Gore never explains what's wrong with this standard. His sole objection is that a judge said the board "could not apply rigid rules that would result in the rejection of validly marked ballots." But how do you judge whether a ballot is "validly marked" without applying rules?