It has been argued that Wittgenstein's later thought, though perhaps not overtly fideistic, nevertheless lends itself to fideistic interpretation. According to this interpretation, religion is a self-contained—and primarily expressive—enterprise, governed by its own internal logic or “grammar.” This view—commonly called Wittgensteinian Fideism—is variously characterized as entailing one or more of the following distinct (but arguably inter-related) theses: (1) that religion is logically cut off from other aspects of life; (2) that religious discourse is essentially self-referential and does not allow us to talk about reality; (3) that religious beliefs can be understood only by religious believers; and (4) that religion cannot be criticized.
and it sort of seems correct. (It seems plausible, at least.) However, I would have to disagree with all four points--and fairly strenuously at that.
1) Religious belief overlaps with many facets of life, not all of them clothed in the language of religious discourse: Charity, literature, family, legal affairs, architecture and so forth. Religion is not a mutually exclusive discourse to other discourses. (And from reading the fideism article, I would think trying to put within the rubric of logic one's religious attitude would already elude the spirit of fideism.)
2) Relgious discourse--like all discourse--is intimately tied to life, the Lebensform. Language is explained by examples, but its meaning lay in use.
3) This claim, that only religious believers understand religious beliefs, seems prima facie correct. But again, when I look at my life--or say, better yet, James Joyce's--when I look at Joyce... No, better to look at mine. I am not a religious believer but I think I understand religious beliefs. I can go to a Roman Catholic mass and do everything, I know why everything is done, I can read Aquinas and Augustine and follow.
4) Religion cannot be criticized by a religious believer: It would be like saying, 'This mountain wasn't here thirty seconds ago'--the entire system of one's beliefs wouldn't allow for such a proposition to have sense.
The encyclopedia entry goes on to further criticize fideism as a philosophic doctrine, comparing it (maybe in psychological terms) to relativism and anti-realism: Things one doesn't want to be labeled.
But there is a kind of anti-rational sense to Wittgenstein's work. In On Certainty, he notes something along the lines of the following. Moore says he knows he has two hands--OK. The book seems to be about debunking that sort of use of the phrase 'to know'. It would make no sense to say that because it couldn't be imagined what not knowing that would entail; i.e., what sort of system of knowledge would allow for one's not knowing he had or had not two hands. Using this systems of beliefs/knowledge epistemic outlook, Wittgenstein says it makes no sense to say to a Catholic, 'That is wine, not blood', 'That little cracker... why are you eating it?' and so forth. (I don't have the book in front of me, it's near approximately page 30.)
Wittgenstein says earlier if someone were to fail to believe in Napoleon he wouldn't know what to do; but if someone doubted the Earth's existence before his birth--he could deal with that. It's easier to understand disparate systems of knowledge than incongruous pieces of knowledge.
My point is that Wittgenstein, while not a fideist in my view, certainly (exuberantly, sometimes) allows for a religious epistemic system to co-exist along with--or even have primacy over--other epistemic systems. And I think philosophy of religion / ethics / moral epistemology is something that I tend to read into philosophers. (Husserl's later philosophy is lifted from the same basic template as the Christian Gospels.)