The simple premise of an election, or so we have always thought, is that the outcome should be determined objectively, by the numbers. Both presidential candidates have publicly endorsed this ideal, demanding from their different perspectives that the count be "fair and accurate." There is a serious sense in which this has now become impossible.
The current election has created uncertainties of a sort that we as Americans have never before experienced, and that we are unlikely in our lifetimes to experience again. When, as in this case, we run up against the limits of electoral precision, the ambiguities and errors that plague every election can become crucial. One of these men, Bush or Gore, will become president. But will he take office with the legitimacy that a democratic election normally confers?
The disturbing tensions of the 2000 election derive from a remarkable statistical fluke. The difference in votes separating George W. Bush from Al Gore in the decisive state of Florida has been hovering around one one-hundredth of one percent, effectively a draw, and with no agreed-upon method for breaking the tie. Suddenly, the minor irregularities that plague almost every election have become epochal. The "chad" now looms very large.
The closeness of this election has furnished us with a lesson about the complications of counting. Until this month, most of us supposed that the process of tallying votes was automatic. (We might have learned something from the controversies about sampling in the 2000 census. But the census faced its own special problem: the difficulty of tracking down people who lack permanent residence or who may be nervous about contact with representatives of the government.) Election tabulation seems so straightforward: Individuals to be counted present themselves at a polling place or submit absentee ballots, which have only to be tallied. Or so you might think.
Any enumeration of people or objects in the millions is going to result in inaccuracies. But elections involve some special difficulties. The voting at the polls takes place on a single day, and the results must be processed quickly. Since we adhere to the principle of the secret ballot, it is tricky to check whether ballots have been properly filled out, and impossible to go back a week later and ask individual voters about their intentions. In addition, the United States has a tradition of radically decentralized decisions about polling practices, with many different systems and ballots, and very few checks to ensure that voters will find them comprehensible.
The uncertainties in our recent election were by no means confined to Florida. But the voting there involved peculiar problems from the beginning, some of them avoidable. Because the state recently relaxed its qualifying standards for presidential candidates, the ballot was packed with names. Officials in Palm Beach County responded by creating their now notorious butterfly ballot, which is reported to have misled thousands of Democratic voters into punching a tab for Pat Buchanan. In Duval County, which includes Jacksonville, the list of presidential candidates spilled onto a second page. Some 22,000 Duval ballots included two votes for president, one on each page. This "overvote" was largely the result, it appears, of a Democratic get-out-the-vote drive in which voters were instructed to vote (and vote Democratic) on every page.
The punching of ballots in Palm Beach and Duval counties each involved enough mistakes to have tipped the election to Gore if voters had done as they apparently intended. There is, on the other side, an appreciable number of uncounted ballots among the heavily pro-Bush overseas absentees, though no one has identified a reservoir of unsuccessful Bush voters to match those on the Gore side. Only a fraction of the uncounted ballots, namely those involving attached chads of various descriptions, can be salvaged. These tend to favor Gore, which of course is why the Bush campaign has worked so hard to keep them from being counted.
A certain rate of "overvote" (voting for two candidates) and "undervote" (voting for none) is typical in any election, particularly in punch-card voting. That rate was exceeded in Florida. In total, there were some 180,000 Floridians who turned out at the polls but recorded no vote for president, or had their votes annulled. This number, 180,000, is more than a hundred times greater than the largest margin separating the presidential candidates at any time during the recounts.
The electoral tally in Florida now inspires rage and despair. When there is rage in the United States, lawsuits surely follow, and thus we have witnessed the seemingly unceasing round of decisions from local and state courts in Florida. Their rulings have alternately stopped and started recounts, with the effect that it has become much harder to secure an accurate count than it would have been if a hand recount had proceeded from the beginning without interference. Instead, there have been obstacles at every point, including election observers who sense that delay may be the key to victory and partisans crowding the election halls.
In short, if counting millions of ballots involves some unavoidable inaccuracies, bitter wrangling has aggravated the problems immensely. Whatever the candidates and their spokesmen might say, some influential parties in this election do not want an accurate tally--they want the tally that gives them the victory.
I hasten to add that any surmise about who "should" have won Florida is speculative, and indeed will always remain so, even if a sincere effort can be made to count as accurately as possible. What we now know for sure is that we cannot know the true "victor" in this election. Against the standard of an ideal electoral process, in which every registered voter who turns out has the opportunity to record his or her choice, and to have it counted accurately, the Florida vote cannot possibly measure up. Should the final tally there happen to agree with what would have happened in an idealized election, it will be purely by accident. The number of votes lost due to errors in Florida is much greater than the spread between Bush and Gore.
After several years of buildup and many months of campaigning, this is the outcome that nobody expected--and nobody knows how to handle. The coin has landed neither with heads nor tails facing upwards, but on its rim. Such a result is perhaps especially problematic in America, for Americans are a peculiarly quantitative people. We invoke the accountants' "bottom line," sometimes literally and sometimes metaphorically, to justify the choices we make. Ours was among the very first countries to establish a regular census, which is required by the Constitution. We use it not only to set up congressional districts (the original purpose) but also to allocate all manner of public funds. We assent to this method of slicing the pie as fairer and more objective than the alternatives. Cost-benefit analysis, risk analysis and decision analysis, all of them embodying the view that proper decisions should be made according to the numbers, reflect a distinctively American outlook.
The unchallenged standing of elections as the only proper way to choose our leaders reflects, even more clearly, the same sensibility. An election should express the people's will, we believe, and should express it objectively. Vote-tallying machines embody, in a way, this idealization of an automatic, objective result. Witness the horror expressed over and over in the last two weeks by Bush campaign spokesmen about the possibility that this election might be decided not by the machines, but by a hand count. They kept saying it would introduce subjectivity into the process. And yet it is clear that the vote-tallying machines produced inaccuracies. Certainly machines cannot determine how to count the votes and which votes to count. Not a machine but a human face, perhaps the now-famous countenance of Florida Secretary of State Katherine Harris, will epitomize the electoral process of 2000.
A common vision of the future, containing elements of utopia and of anti-utopia, anticipates that humans will be replaced someday by machines we have made. The election in Florida shows why this is unlikely and why those machines often fail as an emblem of fairness and objectivity. The strict reliance on automation, often more appearance than reality, has been discredited, and machines cannot remove the danger that many Americans will regard the result of this election as illegitimate. The breakdown leaves no better alternative than to muddle through, using our best judgment and trying to count the ballots as fairly, as accurately, and as quickly as possible, in accordance with the law. But in a situation so politicized, "best judgment" and "fair counts" seem not to have much standing.
Theodore Porter, professor of history at UCLA, is the author of "Trust in Numbers: The Pursuit of Objectivity in Science and Public Life" (Princeton University Press), a historical study of the authority of numbers and calculation in the modern world.
