Jennifer,
First, the word is spelled 'magisteria,' not 'majisteria.' My mistake.
Now, I think you missed the point by paying too much attention to my science example. If I say that B is a set of scientific statements, you can reliably infer that those statements are consistent and nontaulogical. If I say that A is a set of religious statements, then there does not seem to be anything which can be reliably inferred from that.
The semantic content of the word 'religion' is empty, it doesn't actually mean anything. Its only significance is that with which people treat it, but we treat it with significance only because everyone else treats it with significance, and so everyone must continue to treat it with significance, even though it doesn't actually mean anything.
1) One socalled religious belief is that God created the universe about 6000 years ago, and that all evidence to the contrary, from fossils to distant stars, has been placed there by God as a test of our faith. This belief is unscientific i.e. it is immune from empirical refutation, not only is every recorded fact consistent with the belief, but every possible recordable fact is consistent with it.
It has been noted by many that socalled religious beliefs are often unscientific, since the beliefs are structured in such a way as to immunise them from empirical refutation. It is not so much that they can't be "proven," but rather that they are always "proven," no matter what. The problem is that irreconcilable beliefs in God are similarly "proven" by exactly the same evidence, so no number of experiments could help us decide whether the Christian God or Allah is the one true God, nor that God exists at all.
So, we might try and say about set A, the set of religious statements, that every element of A is epirically untestable, in contrast to set B, where every element of B was epirically testable. However, this will not do. It may be a common characteristic of socalled religious beliefs that they are empirically untestable, but then that is a characteristic shared by the whole of mathematics, philosophy, metaphysics, and even metascience, yet few would consider those disciplines religious.
Furthermore, just because many socalled religious beliefs are empirically untestable, does not mean that all religious beliefs are empirically untestable, and indeed the belief that God created the universe 6000 years ago is extremely testable, it is only untestable when butressed with the belief that all evidence to the contrary was planted by God to test our faith. Though not everybody who believes the former necessarily believes the latter, and a great many people do actually revise their beliefs in the light of empirical evidence to the contrary.
2) So, we might say that set A is a set of revelatory beliefs, but then that would make nonrevelatory beliefs nonreligious, but the vast majority of beliefs seem to me to be traditions, passed on by the written or spoken word down generations. What about those socalled religions where leaders have a monopoly on revelation, and simply pass these revelations onto followers, are these followers nonreligious because they have no revelations?
In fact, what might we make of revelations people claim to have had concerning science, philosophy or math, was Descartes 'corgito ergo sum' a revelation, and therefore religious?
3) Another contender is faith though for this we have to change our set A, so it is no longer a set of religious statements, but a set of religious people. Here 'faith' is used to mean dogmatic or fideistic belief, and is irrational (i.e. uncritical). This idea is that religiosity is not a property of statements, but an attitude of people, so a belief is religious if it is held with an unflinching and irrational faith.
However, again we can only conclude that this is also inadquate, since it is quite possible to believe in almost all the teachings of the Bible and not be dogmatic about any of it, though under this definition, our hypothetical nondogmatic Christian, a church goer who sincerely believes in Christ is nonreligious. For an example of such an attitude, check out Mere Christianity by C. S. Lewis.
Conversely, there are many logicians and scientists who are dogmatically committed to the authority of senseobservation, or logical axioms. These people, who are often atheist, have faith no less powerful than a true believer of Mohammed or Christ.
There are more possible demarcations, but every one I entertain leads to a similar situation. It seems that from set A, we can conclude that they are either scientific or unscientific, dogmatic or undogmatic, rational or irrational, revelatory or nonrevelatory, concerning God or not conerning God, moral or immoral, etc. In other words, we can conclude absolutely bugger all.