See 3581262
I had reported that "The analytic element may only occur only within a biblStruct" included a duplication of "only". I now suspect that it was intended to say "may only occur once" and moreover that the RNG for biblStruct is wrong. It allows multiple analytic elements. I suspect it should allow only one.
The reference is to a bug, not feature request:
http://purl.org/tei/bug/3581262
In P4 the model for <biblStruct> was essentially the same, and this phrase was "The <analytic> element may occur only within bibliographic citation or reference elements". So I think your previous reading was correct, and the specification for biblStruct (from which the RNG is derived) should remain unchanged.
Is there anything that led you to this suspicion besides the fact that the Guidelines contain no examples with more than one <analytic>?
That must have been all it was. That and the "only occur only". I just read through the Guidelines and don't see any other language suggesting there can be only one <analytic>.
Okay, I'll close the ticket but leave it open for comments in case you come up with something.