Matt's statement that "American political journalists act as if the political science underlying realignment theory is strong and sound." Do they? Who are "they?" At least for me, journalists who even discuss realignment elections speak of one in the 20th Century - FDR and the Democrat's 1932 election sweep which overturned 36 years of virtually unbroken Republican control of the Presidency AND the Congress. In the most basic understanding, it seems to me that calling the 1932 election a realigning election seems sound. Why wouldn't be? Did some folks in ivory towers build a whole house of cards on this fairly obvious observation? I dunno. Yglesias cites to Wikipedia, which doesn't give me much confidence:

Realigning election or realignment are terms from political history and political science describing a dramatic change in politics. It may center on a "critical election" or be spread out over several elections. More specifically, they refer to any one of several United States presidential elections in which there are sharp changes in the rules of the game (such as campaign finance laws or voter eligibility), new issues, new leaders and new bases of power for each of the two political parties, resulting in a new political power structure and a new status quo that will last for decades. The usual focus is on the transition between party systems, as between the First Party System and the Second Party System, and then to the Third Party System and so on.

The central holding of realignment theory, first developed in the political scientist V.O. Key's 1955 article, "A Theory of Critical Elections", is that American elections, parties, and policymaking routinely shift in swift, dramatic sweeps.

V.O. Key Jr., E.E. Schattschneider, James L. Sundquist, Walter Dean Burnham and Paul Kleppner, some of the most distinguished election scholars of the past two generations, studied the election returns going back 150 years, and found patterns so similar and so peculiar that at first they seemed difficult to believe. Though they differed on some of the details, it was concluded that not only do realigning elections occur, but that they occur on a regular schedule: once every 36-years or so.

The alignment of 1860, with Republicans winning a series of close presidential elections, yielded abruptly in 1896 to an era of more decisive GOP control, in which most presidential elections were blowouts, and Democratic Congresses were infrequent and brief. Thirty-six years later, that system was displaced by a cycle of Democratic dominance, lasting throughout the Great Depression and beyond.

Ahhh. So the nub is the theory that they happen regularly every 36 years or so. so Reagan in 1980 is a realignment and then we aren;t due for another for 10 more years. Well, if that is the theory, then duh, that really is stupid.

Why in heaven's name would the passing of 36 years or so be the trigger? Finding of coincidences does not a serious theory make. Did political scientists really spend a lot of time on this? Gawd I hope not.

Moral of the story, Yglesias sometimes posts about stupid stuff. And so do I. Anyway, tell me why I am wrong if anyone cares to.