I don't know if I really understood where an anonymous commenter was coming from yesterday, but our exchange sparked some thoughts that were meaningful at least to me. I believe that it can be very difficult for what I might call "evidentiary" thinkers and "religious" thinkers both to communicate and understand each other.
Wittgenstein, a mid-twentieth century philosopher, suggested that religious language was a special kind of language game, one that had an internal logic and coherence but did not really refer to supernatural realities at all. That is to say, he didn't actually believe Christians were talking about God as a real being outside ourselves. He believed that this language worked for Christians whether there was an actual God or not. A key indication of this claim to him was his sense that no amount of evidence would be sufficient to convince a religious thinker that his or her core religious beliefs were false.
I don't fully agree with Wittgenstein, but he may have captured an element of truth. It may very well be the case that many Christians use religious language at least partially in this way. I have observed, for example, that some Christians--even pastors--pray in a way that sounds more like they are talking to the congregation or themselves rather than to some Being with an objective existence beyond the church walls. And some people talk of belief in God as if it were a matter of personal choice rather than something that is either true or false.
I'm calling this way of thinking "religious thinking." It is leans heavily toward a non-evidence or non-referential type of thinking that is largely non-falsifiable on the basis of external evidence or attack. A lot of political thinking is of this sort too, especially when it comes to a rabid Democrat or Republican. Their ideas are not about what can be demonstrated, but data is used to reinforce what is already believed.
Evidentiary thinking, in this context, is thinking that can be modified on the basis of engagement with the data of the external world. Ideas can be measured against the "real world," including whether a set of ideas seems to work in predicting outcomes or whether excuses and rationalizations regularly have to be made.
Religious thinking often involves a set of interpretations of scriptural texts. These interpretations may have little relationship to the actual meaning of those texts, but nevertheless function within the overall language system of the group in question. They have an internal logic rather than a referential one.
In the contemporary church, you may find evidentiary thinkers and religious thinkers sitting next to each other in the pew. The evidentiary thinkers may find it very difficult to figure out what is going on or what exactly the preacher is saying. The pastor may be using words in a way that is quite difficult for them to penetrate, like the sermon is speaking a foreign language (of which "Christianese" is a part). Meanwhile, the religious thinker may think the evidentiary thinker lacks spiritual depth, is like the seed that fell on the path and was snatched by the birds, or is even predestined to be damned.
An evidentiary thinker completely outside a religion may find its adherents unintelligible (e.g., Richard Dawkins). Such individuals can speak condescendingly to the religious, completely unaware of the profound way in which religious language functions within a religious community. Unaware of their own assumptions, such individuals ignorantly try to throw a mountain of data at the religious, like the person that stupidly tries to communicate to a foreign speaker by shouting louder in English.
The only way to communicate to a religious thinker is to enter into his or her own language game, just as the only way to communicate to an evidentiary thinker is to speak the language of evidence. The most effective way to change a faulty religious language is to enter it and set it upon itself by demonstrating incoherence. The most effective way to change an evidentiary thinker is either to provide counter-evidence or uncover unexamined assumptions (such as the untenable assumption that anyone can evaluate evidence objectively).
But as it is, these two types of thinkers more often than not just talk past each other. Of course, religious thinkers from differing "religions" also talk past each other. We see this dynamic in the lead up to every presidential election...