The Gettier argument (and its descendants) run thusly…
Someone thinks they know x.
However, due to factor y, they do not know x.
These sorts of thought experiments are used regularly to undermine different accounts of knowledge. Generally I think they are effective but there is one gray area that is under-appreciated.
When the thought experiment is introduced, it is generally assumed to be unproblematic: whoever is setting up the thought experiment is defining the situation and generally is allowed to do so as he or she pleases. However, in setting up a thought experiment that has to do with knowledge, we are inherently assuming an ability to create thought experiments. This means we are presupposing knowing how to do something order to analyze knowledge in general.
In this instance, due to the self reflexive nature of epistemological research, we are forced to accept a presupposition of knowledge of thought experiments when trying to explain knowledge. Therefore Gettier-style thought experiments beg the question by analyzing Knowledge while making you implicitly presuppose a form of knowledge.
Unless you are a skeptic, you are probably thinking that no one is denying that we have knowledge; we just can’t explain it yet. Gettier merely was pointing this out. Therefore it is fine that we have the knowledge of how to have thought experiments and the Gettier-style thought experiments stand as testament to the failure of the Justified-True-Belief account of knowledge.
As I said above, for the most part I agree. I say ‘for the most part’ because just about all the theories of knowledge I have seen don’t take presupposed knowledge as a problem that has to be dealt with, just explained. Hence the big upshot is: If an epistemology came along that started off by explaining thought experiments (and presupposed knowledge in general), then that theory would be a step ahead of Gettier. With a theory of presupposed knowledge you would have the opportunity to prevent Gettier-style thought experiments from becoming problematic. (Those theories would be able to have a retroactive thought-experiment abortion, a la The Terminator.)
Personally I default to my stated epistemological position. Still, for those who disagree with me (and as far as I can tell that is everyone but, like, 3, if I am lucky and it’s a good day), I offer this argument/ suggestion in hopes that it is useful.