The dangers of decompartmentalising toxic waste have been covered here before: Phil Goetz's classic Reason as memetic immune disorder. Vladimir Nesov hypothesises that this is why humans compartmentalise.
In the skepticsphere, decompartmentalising stupidity is considered the best hypothesis to explain the Salem hypothesis: that if a creationist touts scientific expertise in a supporter, said supporter is likely an engineer. But engineers in general are notorious for this sort of thing.
Related to: Reason as memetic immune disorder, Commentary on compartmentalization
On the old old gnxp site site Razib Khan wrote an interesting piece on a failure mode of nerds. This is I think something very important to keep in mind because for better or worse LessWrong is nerdspace. It deals with how the systematizing tendencies coupled with a lack of common sense can lead to troublesome failure modes and identifies some religious fundamentalism as symptomatic of such minds. At the end of both the original article as well as in the text I quote here is a quick list summary of the contents, if you aren't sure about the VOI consider reading that point by point summary first to help you judge it. The introduction provides interesting information very useful in context but isn't absolutely necessary.
Link to original article.
Introduction
Nerd Failure Mode
This section is the part most relevant to LessWrong:
In sum:
I bolded the note on mass literacy and participation because of the interesting historical conclusion that in the United Stated mass participation in democracy inevitably made the influence of religion on policy greater. It goes against a deep assumption shared by most educated people that "democratic elections" necessarily produce "liberal" or "secular" results. It was particularly evident among pundits and particularly easy to see as foolish with the recent upheavals in the Middle East.
This last rather minor seeming note is perhaps the most relevant part of the article for aspiring rationalist. Not only is it particularly salient for those us inclined to questioning the usefulness of the category "religion" in certain context, but because nearly all of us are not religious. Our bad axioms seem unlikely to originate directly from something like a religious texts, though obviously it is plausible many of our axioms ultimately originate from such sources.Not many of us are Communists either, but we are attracted to highly consistent ideologies. We seem likely to be particularly vulnerable to bad axioms in a way most minds aren't.
So if after some thought and examination you notice that a widely respected and universally endorsed axiom in your society has clear and hard to deny implications that are in practice ignored or even denounced by most people, you should be more willing to dump such axioms than is comfortable.