Comment overflow from Crooked Timber, on this article by Paul Krugman in the New York Times (reproduced at the end):
Krugman quotes from Wikipedia:
A conspiracy theory, says Wikipedia, "attempts to explain the cause of an event as a secret, and often deceptive, plot by a covert alliance."
I think there are problems here - one is the assumption of pragmatic explanatory status (previously discussed), in which the word 'theory' plays a role, since it is capable of suggesting a parallel with paradigmatically explanatory endeavours like scientific theories. (The word often also, or instead, imports a presumption of mere speculation, or of a lack of adequate evidence. I note with approval that Wiki doesn't include any such evaluative element in its definition.)
Relatedly, the assumption that some reasonably compact and dicrete event must be involved is inaccurate, though as a defender of 'good' conspiracy theories, and strategically an aspiring rehabilitator of the term 'conspiracy theory', I don't mind this too much since those 'conspiracy theories' which are short on determinate and significant events obviously tend to be vague about all sorts of things, as well as defective in other ways - and if they are booted out of the nominal category 'conspiracy theory' so much the better.
Another more substantive problem is the 'covert alliance' bit - secrecy and/or deceptiveness of the plan is one thing (though it can be challenged as inessential to conspiracy both as a matter of definition and as a matter of salience in this context; the same goes, less plausibly and importantly, for the requirement that an 'alliance' be involved). But specifying that the 'alliance' itself must be covert is wrong - especially on a natural reading which places 'covert' at the extreme end of the secretiveness spectrum (the Bilderberg Group provides a good illustration of this spectrum with its progressive relaxation of secrecy, thanks largely to the efforts of 'conspiracy theorists').