The study of electoral realignments is one of the most influential and intellectually stimulating enterprises undertaken by American political scientists. Realignment theory has been seen as a science able to predict changes, and generations of students, journalists, pundits, and political scientists have been trained to be on the lookout for "signs" of new electoral realignments. Now a major political scientist argues that the essential claims of realignment theory are wrong - that American elections, parties, and policy making are not (and never were) reconfigured according to the realignment calendar. David R. Mayhew is Sterling Professor of Political Science at Yale University.
Mayhew summarizes all the realignment literature into 15 key claims and evaluates whether they are empirically valid, coming to the conclusion that they range from mediocre to poor. He posits that we should think of realignment as a gradual process rather than a punctuated equilibrium model. It is difficult to create overarching theories concerning realignment due to contingency and valence issues, as well as alternative stories of bellicosity, race, and economic growth. Overall, it was useful to have the literature succinctly condensed and critiqued. I agree with many of his concerns and his general view of realignment.
A persuasive case for dismissing the whole genre. That said, I came to the opposite conclusion. Or, at least, think that Mayhew's work calls for reinterpreting rather than abandoning the idea of realignment.