http://worldlibrary.net/eBooks/Wordtheque/la/BembEtna.txt
All right, I suppose I can accept this one as evidence, though I wish there were something official on this matter. This so often seems to be the problem with mediaeval Latin.
Well, tho' I share your sentiments in part, I'll recall how very few, in fact none of the ancient "Roman" authors whose writings survive to us today, were in fact native Romans except Julius Caesar. And some of the greatest Latin authors did not speak Latin as a first language, including Ovid and Vergil. And they sure coined a hell of a lot!
Perhaps instead of "first language" I should have said "language learned from childhood". I believe most of the authors we know who wrote well in Latin were able to acquire the language at that critical age when language comes so easily. Anyways, the point I meant to make was that in a setting like we have today, where Latin is a communal language but not really native to any (besides a few exceptional cases), any attempt to change its grammar would have to be conscious (and therefore artificial) or based on error (and therefore not really a change to the language itself). It would require a large community of fluent speakers to effect any significant change in the language, and while we do have fluent speakers today, they are spread so thin that we can hardly call them members of a single Latin community. This is all besides the point since Bembo was writing at a time when such a Latin community did exist, which is why I find it easier to accept his usage.
flebile nescio quid queritur lyra, flebile lingua murmurat exanimis, respondent flebile ripae