Tuesday, November 21, 2006

I break holiday radio silence briefly ...

... to note that Mickey Kaus has attempted to explain something that has long puzzled me (indeed, I've posted about it here, but am too lazy to look up the post), which is: how can an exception prove a rule? (Scroll down to his November 20 discussion of the "incumbent rule" and Menendez's victory in New Jersey, following this link.)

According to Kaus, an exception can "prove" a rule if the "exception" can be shown to be in some basic way distinguishable from other possible exceptions. Specifically, Kaus posits the "incumbent rule" in predicting election results, which holds that if the incumbent has less than 50% in the final pre-election poll, he or she will lose, because undecideds always break against the incumbent. He then says Menendez is "the exception that proves the rule .. because ... he had only been in office a few months, having been appointed in January, 2006 to the seat vacated by now-Gov. Corzine."

Well, as the nerds in my law school classes always said when the professor called on them, it seems to me that Menendez doesn't prove the "incumbent rule", he simply provides a basis for narrowing it to exclude those whose incumbency has been brief, or possibly all those whose incumbecy is a result of appointment not election. More importantly, it seems that the incumbent rule isn't a "rule" that can be proven (in the sense that a Euclidean theorem can be proven), but rather a theory that can be supported, though never "proved", by empirical means, and can be falsified by contrary evidence. What Kaus is really saying is that Menendez doesn't falsify the incumbency rule if "incumbency" is defined in a strict way. Of course, the "rule" would be falsified if, in a future election, someone who has less than 50% in the last pre-election poll and has held office for a full term or more, still manages to win.

So, I'm still looking for an exception that proves a rule. Anyone who can give me a satisfactory one may have my mid 1960s vintage edition of Copi & Gould's Introduction to Logic (provided I can find it).

Update: Keifus, helpful as always, suggests that an exception "proves" a rule in the sense that it tests it, rather as a new kind of car is put on a test track to prove its driving characteristics. So, what Kaus is saying is that Menendez's re-election "proves" the incumbency rule in this sense, which differs from my more formal definition of proof.

I think this is kind of like saying, "What doesn't kill you makes you stronger", no?

1 comment:

  1. I always thought the expression invoked a definition #2 of "prove," meaning "to test." (Often enough, this was done out on the proving grounds.)

    So the old expression means that exceptions put the rule through its paces (which can bound validity, if not grant it).

    This is good, because now I don't feel a need to read the linked Mickey Kaus article. Close one.