Active Nader supporters are harder and harder to find in progressive circles these days. Rehabilitated Naderites, on the other hand, seem to be everywhere, offering up the same refrain: I've seen the light; Bush is a disaster; I'm voting for Kerry. So how to explain the fact that Nader's poll numbers remain effectively unchanged from 2000? In poll after poll, he consistently draws around 5 percent of the vote, which is almost exactly what he was polling four years ago--and significantly higher than the 2.7 percent of the vote he actually won in the last election.

There are essentially two ways to explain Nader's continued levels of strong support. The conventional theory is that Nader's backers are basically the same type of people (and probably the same people) who voted for him four years ago. That is, they identify as progressives and they genuinely think Nader best approximates their politics. But there's reason to believe that this theory doesn't explain everything about Nader's high level of support. In fact, the more credible possibility is one that few Democrats seem to have considered: that the folks supporting Nader aren't progressives--or, for that matter, potential Kerry voters--at all; rather, they're largely apolitical types who, when queried by pollsters, default to any available third-party candidate.

If you believe the first theory, then Nader's current support is coming mainly from those who backed Kucinich in the primaries and other left-wingers. These voters want a higher minimum wage, a U.S. withdrawal from Iraq, and universal health care. If Nader weren't in the race, the theory goes, they would have no choice but to vote for Kerry, whom they may not like, but who would obviously be the proverbial lesser of two evils. Some polls do support this premise. Last week's CBS/New York Times poll showed that when Nader was included as an option, he won 5 percent, while Kerry dropped from 46 percent to 41 percent. In a mid-April ABC News/Washington Post poll, Kerry pulled 43 percent, with Nader at 6 percent; but Kerry jumped to 48 percent without Nader in the picture. And an April 9 Newsweek poll had Kerry at 46 percent and Nader at 4 percent, with Kerry winning 50 percent in a Nader-free contest.

These polls suggest a pretty obvious conclusion. But there's reason to believe that conclusion is, at best, incomplete--and maybe flat-out wrong. If you believe the alternative theory about Nader's support, then 5 percent of voters aren't choosing him because they agree with his politics, or even because they know that much about him. Rather, they say they're voting for Nader because he is referred to by pollsters as an "independent," and they are independents. "I think you basically have people that are not paying that much attention," says liberal pollster Ruy Teixeira, "and they're asked a horse race question: Would you vote for Republican George Bush, Democrat John Kerry, or independent Ralph Nader? And people say, 'What the hell, independent Ralph Nader.'" This doesn't mean that they are going to actually vote for Ralph Nader when the time comes. Says Teixeira, "My feeling is that it is a very light preference." (For the record, the ABC and Newsweek polls both identify Nader as an "independent"; the CBS poll does not, but it does identify Kerry as "the Democrat" and Bush as "the Republican," so it seems reasonable to guess that most voters infer Nader--the only candidate of the three not identified with a party--to be an independent.)

Teixeira points out that there is precedent for a small number of independent voters behaving this way. In 2000, Pat Buchanan declared he was running for president on the Reform ticket, and before Nader entered the race, the right-winger was polling at 5 percent, too. (He eventually got less than 1 percent of the vote.) This suggests that there's a certain 5 percent of the electorate who will say they're voting for the non-major-party candidate six months before the election regardless of what that candidate's views actually are. In 2000, Nader ran on the Green Party ticket, but this year he is running as an independent--which makes it even more likely that apolitical voters are choosing him on the basis of a generic label alone.

Just as there are polls that show Nader voters moving exclusively into the Kerry column when forced to choose between the two front-runners, there are also polls that show these voters more evenly split between Bush and Kerry. A Los Angeles Times poll in late March showed Kerry and Bush each losing two points to Nader when he is added as an option, and the latest Pew Poll has Kerry gaining four points and Bush gaining two when Nader is removed from the picture.

But perhaps the most convincing piece of evidence for this theory is the poll where Nader does the worst: George Washington University's Battleground Poll. This is the one poll that asks likely voters the open-ended question of whom they would vote for, without prompting them with names. One would think that if Nader's poll numbers were the result of real commitment by ideologically driven voters, the open-ended format would have no effect. But in this poll, Nader draws 1 percent.

In the end, you're left with two competing theories, both of which have good polling data to back them up. But the second one seems far more convincing, if only because it squares firmly with the available anecdotal evidence about liberals being united behind Kerry. It jibes with the fact that Nader was unable to get 1,000 people to come to a recent rally in Oregon, which would have automatically put him on that state's ballot. And it explains the marked dissonance between what former Nader voters say they are planning to do and what the candidate's poll numbers continue to show.

Implicit in this theory are two important points--one with bad implications for the Democrats, the other with good implications. The bad news is that if Nader's supporters really were antiwar, anti-corporate leftists, they could, I think, be fairly easily brought around to voting for Kerry. Friends would flood their inboxes with emails about the latest Bush outrage, family members would stage interventions, and by Election Day a sizeable portion, the vast majority even, would hold their noses and vote for Kerry. If, however, these voters aren't left-wingers--but instead barely interested contrarians and independents--then there's really no telling whom they'll vote for. This means that Democrats can't mentally add 3 to 5 percent to Kerry's numbers every time they see a poll in which Nader does well.

But the good news for Kerry is that if these Nader voters are not really leftists--and therefore, not really votes that he ought to win--he can safely ignore them, and run a more or less centrist campaign. That should allow him to tailor his message to swing voters, who will likely decide the election, without paying a political price.

The irony, of course, is that Nader has been telling Democrats to chill out, and insisting with a straight face that he'll draw support evenly from Democrats, independents, and even Republicans. It turns out he may be right. But if that's true, indeed, precisely because that may be true, his supposed rationale for running--nudging the Democratic Party, and American politics, towards the left--isn't something he appears capable of doing. Which raises the question: If the people who share Ralph Nader's politics aren't supporting Ralph Nader, then what constituency does he represent--other than himself?