Caledonian,

The part I have a problem with is where you go from the cardinality of the sets to a judgment of "equally probable".

Let me put it this way: you wrote,

In the English version, in any of the truncated (and sufficiently long) versions of the Library, the sequence "AB" is much more common than "CDEFG". It doesn't matter whether the texts are ten thousand letters long, or ten billion - the first is less complex and thus more probable than the second.

The "any" is the problem. I can construct a truncated version of the Library where your assertion doesn't hold, just like I can fiddle with the order of a conditionally convergent series to get any limiting value I want. So when you say, "In the FULL version...", you've left a key piece of information out, to wit, what is the limiting process which takes finite versions of the Library to the infinite version.