This week's New Republic editorial gloats over the corpse of the alleged Bush/Rove electoral realignment that was supposed to make the Republicans into the dominant political party of our era. Quite so. But to push things more broadly, I would hope one result of the past several years would be to help political journalists to break their addiction to the whole concept of electoral realignments. As Yale political scientist David Mayhew points out in his book Electoral Realignments: A Critique of an American Genre there's very little reason to believe that this is a useful concept denoting a real phenomenon.
If you look backwards through American history with a strong ex ante commitment to believe in the idea that electoral realignments occur every 36 years, you can appear to find them. In particular, you'll come to the conclusion that 1896 was a realigning election that occurred precisely 36 years between the realigning elections of 1860 and 1932. You'll then find yourself puzzling over why the 1968-2004 period doesn't quite seem to fit the model, and possibly expecting that the country is due for another realignment sometime around now.
As Mayhew points out, though, if you don't just start looking backwards with the assumption of a 36 year realignment cycle, then it's not clear that 1896 had any especially fundamental significance as opposed to 1912 or several other election years between 1860 and 1932. And then you'll see that the "problem" of identifying post-1932 realignments isn't a real problem at all. And, indeed, you'll start to realize that there's no compelling theoretical reason to believe that electoral politics in the United States necessarily consists of alternating cycles of partisan dominance based on 36-year cycles. Instead, it just so happens to be the case that 1860 and 1932 were dramatic years in American political history and that those elections occurred 72 years apart from each other.
At any rate, if you don't want to read Mayhew's whole book, this article of his in the Annual Review of Political Science offers a shorter version of the argument.
Recent American political history tends, I think, to strongly support Mayhew's view of things in which electoral outcomes are highly contingent and largely dependent on somewhat transient events and weird stuff. It's just, in other words, one thing after another. 1996 gives the country divided government. Then, in 1998 voters react negatively to the Clinton impeachment and Democrats gain ground. In 2000, generally positive feelings about the direction of the country under Democratic rule result in further Democratic congressional gains and a popular majority for Al Gore; but Ralph Nader's strong third party bid, the electoral college, and some weird shit in Florida puts George W. Bush in the White House. Then, in 2002, 9/11 and the response to it pave the way for GOP gains. In 2004, the Democrats wind up flubbing the Iraq issue so Bush gets reelected; but the underlying congressional trends favor the Democrats; but canny Republican redistricting leads the GOP to actually pick up seats.
Then, in 2006 disgruntlement over the Iraq War and with Republican corruption swings the Democrats into power. This then leaves us with 2008 where the outcome will in all likelihood be highly influenced by events that haven't yet occurred and about which there's actually very little to say.