It’s not exactly a question of logic, strictly speaking, but the dictum *entia non sunt multiplicanda praeter necessitatem* or *pluralitas non est ponenda sine necessitate* might aptly characterize what is wrong with George Will’s Sunday op-ed in the *Washington Post*. Even though such hefty Latinisms are no doubt familiar to the typical Will reader (cf his recent “Pennsylvania *est omnis divisa in partes tris*”), for the sake of those not used to Will’s Latin, it means, in paraphrase, “the simpler explanation is probably the better.” And that is just what we found in Sunday’s post; complicated explanations when simpler ones would probably do.

Part of the problem with the coming election, as Will sees it, consists in the “fraud-friendly” voter registration systems, such as the National Voter Registration Act, and their effect in a couple of key states–in particular, Ohio and Pennsylvania. In the wake of such measures, Will worries that

perhaps we should not be surprised that . . . since 1995, Philadelphia’s population has declined 13 percent but registered voters have increased 24 percent. Are we sure we should be pleased?

The unexamined belief that an ever-higher rate of voter registration is a Good Thing has met its limit in the center of the state that this year is the center of the political universe — Ohio. The Census Bureau’s 2003 estimate is that in Franklin County — Columbus — there are approximately 815,000 people 18 or older. But 845,720 are now registered.

A simple explanation quickly presents itself: voters whose status has changed–they moved, died, or went to prison–remain on the voter rolls. And, oddly, this is exactly what Will reports. However sufficient this explanation seems to be, Will immediately suggests something more sinister to be afoot:

One reason for such unacceptable numbers in various jurisdictions across the nation is that voter rolls are not frequently enough purged of voters whose status has changed. For example, in 2000, the Indianapolis Star’s Bill Theobald reported that “hundreds of thousands of names, as many as one in five statewide” were improperly on Indiana registration rolls “because the people behind those names have moved, died or gone to prison.” Unfortunately, there is reluctance, especially among Republicans, to support measures that might appear to have a “disparate impact” on minorities and therefore be denounced as racist.

This is a red herring. The question seems to be why–in one county in Ohio–so many people seem to be on the voter rolls. The simple explanation, people have changed their voting status, is converted into a concern–by someone in Indiana–that the Republicans are afraid of being called racists for purging the voter rolls. Before diverting out attention from the integrity of the voting process to the far more interesting (however groundless) reverse racism accusation, Will should follow through on the simpler explanation: what evidence is there for the claim that outdated voter rolls leads to electoral fraud? Will offers none. In fact, for someone to commit fraud, she would have to vote in more than one county on election day. And for this to amount to anything, thousands of people would have to do the same thing, fraudulently (voting in several counties or states by absentee voting or just by doing a lot of driving). It would have to be a plot of enormous complexity and secrecy for it to pretend to have any effect on the outcome of the election.

Against the existence of a plot of such devious and gratuitously baroque complexity, one might say that both political parties have perhaps come up with an equally sinister, but much simpler, scheme: register people to vote.