Back when I was reporting this article for the January 2004 issue of The American Prospect, I became convinced that most "superdelegates" would actually be loathe to use their influence to reverse the outcome of the primaries. Thus, I've always thought that superdelegate support should probably be discounted at this point. There's nothing stopping the SDs pledged to Clinton from switching to Obama and vice versa. This Chris Bowers post did, however, suggest to me a way in which they really might come into play:
A campaign that is now on course to be down by more than 100 pledged delegates in two weeks didn't "tie." Just like Mitt Romney, any campaign that is talking about changing delegate allocation rules didn't "tie." A campaign that is plugging its website to try and raise money didn't "tie." A campaign that talks about stopping the momentum currently enjoyed by its opponent didn't "tie." That is a campaign back on its heels. As I wrote last night, this was not a tie, and Obama clearly has the edge.
Imagine a scenario in which Obama has the majority among pledged delegates, but Clinton has the lead among all delegate. Her superdelegates probably won't want to give the election to the candidate who lost. But they could use their majority at the convention to seat the Michigan and Florida delegations. And then Michigan and Florida could put Clinton over the top in a way that could be construed as more democratic than the alternative. Just idle speculation for now, but who doesn't like idle speculation?