Schrodinger's Election

by Jonathan Wallace

November 18, 2000-- In the quantum physics parable of Schrodinger's Cat, the cat is neither alive nor dead until you open the box, but in a probabilistic state embracing both. What we have before us today is Schrodinger's election, in which either Bush or Gore is president, and where we are seeking to open the box, and ascertain the truth, via hand recounts, rhetoric and litigation.

In college, I remember the sense of relief I felt when I realized that there must be one truth, always ascertainable if you have the equipment. I would debate people with a much fuzzier idea of truth, who said things like, "Well, one thing may be true for you, another for me." And I would respond with my hypothetical of Columbus aboard his ship: "If Columbus believed that the world was round, and a sailor named John believed equally fervently that it was flat and had an edge, would the ship simultaneously sail round the world and fall off the edge?"

Of course, it was not too long afterwards that I realized that while (as I continue to believe) truth itself must be unitary (and I do not pretend to understand quantum physics) we may not always have the equipment to perceive it. The dilemma that your mind becomes capable of wrestling with around age 12--the seeming impossibility of time or space being either finite or infinite-- is one that humans may simply not be smart enough to solve, any more than my pet cockatiel is capable of learning that transparent glass is a barrier he cannot pass through.

In another class of dilemmas, the data was of a type that can be easily understood by humans, but has passed beyond the event horizon and is no longer available to be analyzed. The state of mind of the Egyptair pilot who allegedly crashed the plane into the sea can be guessed at but can never now be known. Every day people die who are the last witnesses to--the last ones who know the truth about--events important and trivial. While many black boxes are opened (we now know that the Rosenbergs were in fact atomic spies) there are many we will never open (unless previously unknown documents still come to light, we will not know whether a secret deal was struck to resolve the Tilden-Hays electoral college dispute in 1876.)

The 2000 presidential election falls into the latter category of event: the data is of a kind to be easily understood, but it wasn't captured properly and is not really available for analysis. What happened is similar to a familiar problem from my business life: while profitability is a binary switch (a company must either be profitable or not, it can't be both), there are circumstances where it is impossible to know if an enterprise is profitable or not, because the full range of data necessary to make the determination hasn't been properly recorded.

Thus some millions of Florida voters either went to the polls or filled in absentee ballots, and if we had some way of reading their minds--if, in fact, we could take a snapshot across time of their state of mind as they recorded their vote--we would really know who won the election. Of course, the ballots themselves were supposed to be that snapshot. But they aren't even close to providing an objective record, in a close election. This is the lesson of the so-called "dimpled chads".

Unlike the antiquated but better designed voting machines I am used to in New York, where you pull small levers to display a set of x's next to the names of the candidates you intend to vote for, but the vote is not recorded until you pull a large lever, the punchcard process used in several Florida counties contains no built in error correction. The forty year old New York machines will not allow me to vote for more than one Presidential candidate, but the Florida approach produces ballots where two presidential candidates are punched, or a dimple next to a candidate's name must be interpreted to determine if it was a failed attempt to punch a hole there.

Looked at from that perspective, the Florida debacle is one of the most important design failures in American history, akin to the O-ring which snapped in the Challenger. The following elements of bad design have come to light in the casual reading I've done this week, and I'm sure this is only a partial list:

There is no reason why a ballot entered in a voting booth should be an analog form at this late date. Absentee ballots which are filled out and mailed in, by their nature, are analog forms which cannot contain error prevention protocols. However, there is really no reason why technology could not be designed to enable absentee voting by computer--if not from home computers today, with the corresponding encryption and security issues (which do not seem insurmountable) then from supervised "polling places" in U.S. consulates and military bases abroad.

Problems with absentee ballots included ones sent via military mail which were not postmarked and numerous ballots which failed to include the voter's ID number. In some cases, Republican party members who had filled out the ballots and sent them to voters for signature, were allowed to add the missing ID numbers after the ballots were received.

It is infuriating to blame voters for falling into the traps set for them by a bad design. There have been many jokes flying around this past week about the low intelligence of Florida Democrats. However, anyone, even an Einstein, can be misled by a poor user interface. For a fascinating book with numerous examples of bad interface design, see Donald Norman, The Psychology of Everyday Things, which contains numerous photos of confusing doors, switches and other technology. My favorite illustration shows an instrument panel in a nuclear plant, where two identical levers, placed next to each other, have radically different functions. The workers solved the problem by placing differently shaped beer kegs over the two levers so that they could easily distinguish between them in an emergency.

The result of the dismal design of voting systems in Florida is that it may ultimately not be possible to reconstruct who actually won the election. Systems with poor granularity are only useful, if at all, for capturing wide-margined data. A company with a terribly-designed accounting system may know it is losing money in an extreme situation because it cannot pay bills, but the system may not have the granularity to recognize that it is losing a few dollars every quarter. The Florida system can tell you who won by a landslide but became useless in an election where the margins are only a few hundred votes.

The technology has been available for decades to do a better job than this. Why didn't the people responsible in Florida take advantage of it? While many design failures, like the Challenger explosion, can be chalked up to human vanity and complacency, in this case there may be a subtext. Charles Perrow in Normal Accidents--a fascinating work, now unfortunately out of print, on catastrophic design failures--makes the observation that there is no technological reason why we cannot have marine traffic control systems similar to the ones we use for air traffic control. He surmises that the reason we do not is that wealthy, powerful and famous people tend to die in plane crashes, while unknown sailors, usually from foreign countries, die in marine disasters. (A typical nautical accident involves two ships at night, coming from opposite directions, on course to miss one another, which turn in towards one another and collide. The reason is that a ship approaching is misconstrued by the human navigator as one going away, who then attempts to steer inside the other, closer to shore.) My brother and sister-in-law, who live in Florida, inform me that the cost of installing better machines is considered prohibitive in Florida. But perhaps then the federal government should fund an upgrade. The feds pay for air traffic control systems, after all, rather than leaving these the individual responsibility of the counties through which the planes fly.

In the shipwreck that is the Florida election, there may be a natural but repellent desire at play: that for maximum human discretion and control, achievable only through vague analog systems. Florida election officials may be like the controller in a company with a completely manual accounting system, who sees his job, or at least his importance, threatened by automation. This theme was at play both in the wide discretion of the Florida secretary of state (a Republican) to decide whether to accept the result of manual recounts, and Bush's quite contradictory challenge to those same recounts because they involve too much human discretion.

Finally, it is worth noting the extremely strange and rather humorous spectacle of what is essentially a forensic attempt to reconstruct the intent of a living, accessible group of people who would be happy to clarify what they intended if given the opportunity. The process of trying to ascertain the intent of a voter who dimpled a ballot is similar to that of trying to determine the wishes of a deceased person from an incoherent or contradictory handwritten will--but in this case the voter can be asked to comment, as the deceased cannot. No doubt, the simplest solution would be to hold the Florida election again. It might not be less expensive but it would be a much more accurate proceeding than all this fighting and counting. There are only two arguments against it, closely related to one another, and neither of which I find persuasive: the first is that the Constitution forbids it, by requiring that the election take place everywhere on a single day; the second is that "you cannot step into the same election twice," and that the result on the second attempt may be something entirely new, rather than a clarification of what was intended ther first time.

On the first issue, as the Florida Supreme Court just pointed out, when the law contradicts itself, a court must decide which of two incompatible rules better serves justice, and remove the other. The Constitution certainly intends that a clear result be produced, indicating the intent of the states, and was not written to tolerate accident or fraud. If the goal of a clear result cannot, despite everyone's best efforts, be accomplished on a single day, I think it is more important to get the right result than a fast one. This is in fact exactly what the Florida court decided when it ruled that completing the recounts accurately was more important than meeting the law's deadline for certification of the results one week after the election.

The second argument, that another vote in Florida would be a new election rather than a clarification of the earlier one, is more persuasive. It is likely that many people would change their vote: certainly many Nader voters would vote for Gore this time, probably enough to put him over the top; some people, however, perceiving that their candidate has been behaving like an obsessive wiener since losing the initial count, might go the other way. However, to say "you can't step into the same election twice" begs the question that in this case, due to the badly flawed design, we couldn't step into it even once. If a hurricane had destroyed twenty or thirty percent of the Florida ballots before they could be counted, would we have proceeded to count the electoral vote without Florida (in which case Gore has a majority of the electors) or would we re-do the Florida election? Due to the very poor data capture, we are essentially in that situation.

November 25-- I put this essay aside for a week, and today, as I pick it up again two days after Thanksgiving, we still do not know who our president is. In the intervening week, there have been several court cases, including a decision of the Florida Supreme Court extending the tight deadline for completing manual recounts, and the Supreme Court has just agreed to take the Bush campaign appeal challenging this decision. In the meantime, Governor Bush is still ahead by several hundred votes after all the partial recounts, Gore has incurred some public enmity by targeting un-postmarked military ballots for elimination, and a Republican-instigated near-riot in Miami-Dade may have contributed to the Democratic cavassing board's decision (potentially crippling for Gore) to stop the recount.

Although (having voted for Nader) I would rather see Gore become President than Bush, I have extremely mixed feelings about all this recounting and litigation. Through the whole process, including every manual recount so far, Bush has maintained an informal lead that has never fallen below three hundred and twenty votes, and is presently up around six hundred. Gore, who has announced that he will continue to fight if Secretary of State Katherine Harris certifies the election for Bush tomorrow night, looks even to a lifelong Democrat (I voted the straight party line from McGovern in 1972 through Clinton in 1992) as if he is looking for votes under every rock instead of facing the truth, that Bush won Florida by a slim margin. If among the thousands of contested votes, he manages still to pick up a majority (something he has not yet done, even for a moment, in the two weeks since the election) it will look like a manipulation or pure blind luck; he does not seem to be a man looking to correct an error he knows has occurred.

Of course, to some committed Democrats the current process is nothing more than a search for the truth. I cannot accept it as such for two reasons. First, as I stated above, I believe the truth is not ascertainable, due to the extremely badly designed punch-card system for capturing the data. Secondly, it is Gore's responsibility to explain why he thinks he actually won, other than his own hope and desperation. A theory would be something single and consistent, like minority voters being disfranchised everywhere in Florida (an NAACP report says this was widespread, but for mysterious reasons has not been picked up and brandished by the Gore campaign.) Gore's arguments sound more like classic legalistic proliferating arguments in the alternative: the ballot was bad here, military votes shouldn't be counted there, we need to look at dimples over there.

The misstep on the military votes particularly illustrates that Gore's legal strategy has no ethical cornerstone: it is a battle to win at any cost and not a quest for truth. The military frequently does not postmark mail, and military voters cannot accordingly be held responsible for their mail not being postmarked, the way they could for failing to sign it. The Gore campaign accordingly has attempted to disenfranchise military voters for one reason alone: because they tend to vote predominantly Republican. The simple and compelling theme of "let every vote count," which would have been a good old fashioned battle cry for a Democrat's campaign, instead becomes, "Let's extract the maximum number of arguably Democratic votes while rejecting even unambigous Republican ones," which is not a cry that's going to make me want to run out of the trenches and face the bullets.

Gore's desire to count dimples is similarly a major embarrassment. (Let's put aside the fact that it is a worse national embarrassment that in 2000 A.D. we are still using voting equipment that can create dimples.) There turns out to be a significant body of precedent as to when dimples are to be counted as votes. This is typically left to discretion at the county level, and Florida counties are handling it differently. Reportedly, Gore is now looking primarily to the dimples for the hundreds of votes he needs for victory.

On the one hand, dimples are routinely counted under Texas law and in other places, and (a minor embarrassment for Bush) were just used in a Texas recount of a local race this week, peacefully and without inciting any dissension between the local Republicans and Democrats. On the other hand, dimples are inherently ambiguous, and the prevalent rule (echoed by the Florida court last week in its citation of an Illinois ruling on the subject) is simply to do the best you can in determining the intent of a voter. To see the inherent difficulty, think about the following possibilities: straight Democratic line, all dimples; straight democratic line, all punched through except the presidential choice, which is dimpled; straight Republican line for all other offices, but a dimple for the Democratic presidential choice; and so on. Sometimes a dimple may represent a false start, in a decision not to vote for a particular office at all; sometimes a dimple may occur during the handling of a ballot after the election.

If Gore gets in solely based on this highly ambiguous evidence, it is likely he will serve a single term as one of our weakest leaders, and be known forever as "President Dimple."

November 26--Litigation. In the last few days it seems to have become clear that Gore is determined to become president now, at any price, and without regard to the reputation he will have in 2004. I had assumed, on election day, that Gore, because he ran such a close race, would be his party's candidate again in the next election; Gore seems to assume no such thing, and to believe he has nothing to lose, by antagonizing his own voters via litigious behavior, so long as he becomes president now.

This is especially saddening because it would appear that whoever becomes president now will be very weak, and will be vulnerable in the next election. If Gore were a statesman, considering what is best for the country and for the party, he would concede now and lead an honorable opposition, aiming to recapture the Congress in 2002 and the presidency in 2004. Gore instead has announced he will not accept the results if Bush is certified today, even though all the recounting has yet to produce a Gore majority, no matter how ephemeral. This obsessive behavior boils down to plain bloody-mindedness, a refusal to accept reality coupled with a ferocious desire for victory at any price. From this perspective, Gore is enacting his own tragedy of the commons, a desire for maximum benefit today at the expense of future generations.

Is there permanent harm from this behavior? I think there is. Perhaps the most important lesson of this election has been the Heisenbergian uncertainty of the election process in determining a victor in a close race. This would be a healthy realization if the immediate reaction was to use modern technology to create a more reliable way of capturing the data. Instead, we may end up with a new tradition, that post-election day litigation is a normal part of the process.

Post-modern academic analysis specializes in the construction and decontruction of everything: for example, as per Erving Goffman, the social construction of the self, the persona we put on for other people. Democracy itself is quite a fragile construction: like William Gibson's cyberspace, a "consensual hallucination" of millions of people. This hallucination is based on quite a few "self-evident truths" that are not really so evident: that elections work as a way of determining the will of the majority; that the rules, such as the electoral college, make sense and serve the goals for which they were intended; that the whole system is more than a fragile human rule-book that can be changed at any time, but has some validity, some reality, exceeding the merely transient and convenient. The founders began this fiction with the Declaration's words that "we hold these truths to be self-evident", which I have described elsewhere as being merely a debater's trick of declaring victory.

The best democracy (a very loaded idea, which one day will warrant its own essay) would be an intellectual partnership of people conscientously committed to a particular set of rules, and to amending them when necessary. The idea of our democracy as something exceeding that, as engraved in some way in the fabric of the universe via some set of "inalienable rights", is reminiscent of the old testament parable of the golden idol being used as a trick to get people to raise their eyes to heaven.

Democracies cease to exist when not enough of their members believe in the rules and act in good faith to support and extend them. You might argue that the judiciary is simply part of the rulebook, and that involving them in every election is not as pernicious as bringing in troops. But the difference appears to me to be one of degree, not of kind: in both cases, routinely disregarding and attacking the legitimacy of the election process by reference to another branch of government. Again, the real task would be to improve the accuracy of election results, not ask the courts to be the arbiter every time. Hauling in the courts communicates, past a certain point, that the will of the electorate is impossible to determine (which is clearly not true) and asking the judges (who themselves all belong to a political party through whom they got their jobs, by election or appointment) to substitute their own judgement.

In case you think I only blame Gore: Bush is no better. Because he held the majority, and also perhaps because he is slower-witted, he has let Gore seize the litigation advantage; nonetheless, Bush got into court first, with his insidious law-suit to stop the manual recounts. Bush shows no sign of being on any different moral level than his adversary. There is no grandeur, no sense of honor, anywhere in this election.

This seems like the right place to mention the bullying of Democratic officials by an unruly crowd of Republican demonstrators in Miami-Dade, and the role of the party, including prominent elected officials, in inciting the violence. Though I do not share Phil Agre's apparent belief that this is the defining story of the election (nor is it clear that this intimidating behavior was the cause of the canvassing board's decision to stop counting), it is nonetheless a significant symbol of the breakdown of the process. U.S. democracy has been touted for years (I suppose mainly by Americans) as representing the most significant peaceful transition of power in the world. Like much of what is happening in Florida, assaults on members of another party simply for daring to recount the votes, sounds like something we would expect in a third world country, not here. If Democrats (as I have heard so many times this week) have a propensity for fraud, Republicans (particularly the "Second Amendment is the reset switch" set) have a clear predilection for violence. It is a fundamentalist tendency, an explosive mixture of absolutism and self-righteousness. It is there, between the lines, in a lot of the email flying around on public lists this week: if the recount elects Gore, "he won't be my president!"; Gore should concede because the Republican half of the electorate will be "far angrier" than the Democrats will be if they must endure Bush for four years; the constant references to Gore trying to "pull off a coup", even though he had a majority of the popular vote.

The Electoral College. I wrote in Democracy and Illusion two months back that we do not have as much democracy as we think we do. The electoral college is a good example of this. Although human failure to understand an institution is not in itself evidence of that institution's ineffectiveness, it can reveal a divergence between publicly held beliefs and a deliberately concealed reality.

In the United States most people have always believed that we elect the president by popular vote. This image, which is wildly inaccurate, is pervasive because consistent with other popularly held ideas about American "democracy". As I illustrated in Democracy and Illusion and Democracy and Stupidity, the founders didn't want a popular democracy; the Federalist Papers are full of somber, bloody references to the dangers of Athenian-style popular rule.

The electoral college is a prime illustration of the complex machinery the framers devised to avoid majority rule. A cautious, complicated compromise carefully negotiated among the representatives at the Philadelphia convention, the electoral college was adopted to ensure that the more populous cities of the North would not routinely capture the Presidency.

The electoral college consisted of one elector for each Senator and Congressman from each state. Since even states so tiny as to have only one Congressperson still have two Senators, the system ensured that the smaller states would have disprtoportionate representation. Further skewing things was the fact that the Constitution provided that slaves would be treated as three-fifths of a human being for the purposes of determining the number of Congressmen to which a state would be entitled--thereby also influencing the number of electors.

The states were left free to determine how electors are selected. Everywhere the competing parties appoint their own slates of electors, and the popular vote determines which slate is successful. The Constitution does not actually require that there be a popular vote, nor are the electors in many states required to vote for the candidate whose coattails they rode. There are a few examples, mentioned in the newspapers these last few weeks, of electors who broke from the pack and voted for someone different (including one Republican elector who voted for the Libertarian candidate for president.)

In forty-eight states, the winner of the popular vote gets all the state's electors, and in the other two, the electors are selected proportionally to reflect the popular vote.

The main consequence of this approach is that it makes it possible, as is about to happen in the 2000 election, that the winner will be the candidate defeated in the popular vote. Though rare, this has also happened before, most notably in the 1876 Hayes-Tilden race which bears so much resemblance to today's mess.

The idea of a candidate unwanted by a majority of Americans seems repugnant to everything we were ever taught about democracy. Nevertheless, as a complex compromise of the framers, it has always worked fairly smoothly (it has not ever led to the destruction of the system, and presidents succeed each other without our ever thinking too much about the electoral college.)

There is no way to do justice to the system of federalism in a few paragraphs. Under our system, certain things are left to the states, while others are the province exclusively of the federal government. To pick vivid examples, the state of Nevada can decide to legalize prostitution, but it cannot ban eighteen-wheelers from entering on interstate highways. The Supreme Court, more used over the years to overturning state laws for violating Constitutional principles, threw out a federal law barring guns from schoolyards a few years ago, on the grounds that this is a state matter. The Supreme Court's obscenity rulings provide for wide state variations in defining obscenity.

One common theme in all of these cases is the recognition that groups of people, in smaller units than the entire nation, should in many cases be able to decide how they want to live locally: they may want prostitution, or to ban obscenity. In other cases, the Constitution as interpreted bans local decisions: states can't bar eighteen wheelers, or (another case I remember vividly from law school) mandate that margarine be colored red. In each of these cases there is a principle that certain categories of decisions must only be handled at the federal level: decisions concerning interstate commerce and civil rights are two prime examples.

The electoral college system does not serve either goal, of local authority or federal pre-emption. It is the only Federalist relic which flouts national self-determination without protecting any local right. In every other example, we can point to some tangible group of people--the residents of Cincinnati, Ohio or Reno, Nevada--who are able to pursue their own local destiny, or not, depending on the category of decision. The electoral college, however, seems like another type of Federalist animal altogether. What the electoral college really means is that the people of the United States do not elect the President: the states do. But the logic here seems to break down irretrievably, because a state (despite all our legal fictions to the contrary) is not a person. It is a group of people with local rights. But the fiction promoted by the electoral college is that the set consisting of all states can have a different identity, and reach a different result, than the set consisting of all the people of the United States, and that just doesn't make any sense. Which is why I think we should abolish the electoral college.

November 27--Last night I watched Florida Secretary of State Harris certifying Governor Bush as the winner of Florida's 25 electoral votes. Gore vowed to fight on, advancing numerous challenges to the cessation of the Miami-Dade recount, Harris' failure to use the Broward recount which was not completed by the deadline, the failure in several counties to count dimpled ballots, and most significantly, the infamous butterfly ballot.

The attack on the butterfly ballot may mean that Gore is trying to knock out the Florida vote entirely, so that the election will be decided based on the electoral vote without Florida, of which Gore has a majority. If a court ruled that the butterfly ballots were so badly designed as to be completely unreliable in capturing the results (and I believe they were) it is hard to see the remedy. There is no way to determine that the anonymous voter who clearly, unambiguously voted for Buchanan (no chads) really meant to vote for Gore. So a recount of these ballots would be meaningless. Holding the election again would be a better solution, as I mentioned above, but no court is going to order this at this late date. So the only solution would be to throw out the results in the county or in all of Florida, disenfranchising those voters who voted for the candidate they intended.

I expect to see leading Democrats defect from Gore in the days to come, as his position is now untenable. All this litigating and recounting, and he has so far failed to reveal a hidden Florida majority. It has become impossible to reply to the Republican thesis that Gore just wants to keep recounting until some combination of dimples and technical challenges to Republican absentee ballots puts him over the top.

The time has come for Gore to concede and let everyone get on with their business. He is beginning to seem obsessive and unbalanced, the sore loser the Republicans have called him. I never felt good about him but his behavior of the last weeks especially makes me doubt his ability to govern.

Blaming Nader. There has been a lot of rage unleashed against Nader: that he stole the election from Gore, that his vanity kept him going when he should have withdrawn and asked his followers to vote Democratic, that no prominent Democrat will ever again return his phone calls, etc.

I find this rhetoric astonishingly immature. On one level, it involves a complete refusal of responsibility. The real question is not why Nader had the chutzpah to siphon off two million votes from the Democrats, but what the Democrats did to alienate two million mostly Democratic voters, and what they could do to get them back. And the answer is very simple: stop taking money from huge corporations which exact in return that you sell out all of your traditional constituencies, such as labor, minorities and environmentalists. In fact, the implication that we as voters can be stolen, rather than alienated, insults our intelligence. I failed to vote for Gore because he took illegal Asian money, because he failed to differentiate himself enough from Clinton both morally and politically, because he was quiet and complacent while bad decisions were made about the environment and world trade. I voted for Nader as a protest, with full foreknowledge of what the consequences might be, and I would do it again, without a second thought. Even if I lived in Florida.

The anger against Nader is hugely arrogant. Anyone building a third party has a right to do anything legal under the existing rules to build that party and assure its success. The idea that Nader should have conceded the race and thrown his votes to Gore presumes that only the two parties are legitimate, that anything else is mere play and wasted time, and that Nader should have gone back to his sandbox before he did any harm. This is as twisted a view of competition as you are ever likely to hear. Morally it is tantamount to saying that if Michigan needs to win one game to make the Rose Bowl, and Ohio State is not going to get there no matter what, that Ohio State should throw the game with Michigan rather than stand in the way. No, in that context we expect Ohio State to play just as fiercely as it can.

Accepting the view that Nader should have withdrawn would place us, the alienated Democrats, in a permanently abusive relationship with the Democratic party, where the party can completely disregard our views, take actions which harm us, and still expect our votes because it is confident that it is the "lesser of two evils." I won't play that game. By supporting the Greens I expect to contribute to the world being a lot better in thirty years than it would be otherwise. Even if it isn't, at least I will have tried. I would rather be an enthusiastic Green than a disappointed Democrat.