One objection that you sometimes hear to more communicative methods is that those who are using historical languages productively will ‘get it wrong’ and thus be creating unrepresentative/incorrect language data for learners. This is another in my And especially that this is self-reinforcing – incorrect language features are used by incomplete learners, and create a hybrid monster that is “communicative historical language X” which actively harms people learning “historical language X” (feel free to substitute in your Latin, ancient Greek, biblical Hebrew, etc., as needed).
My response to this kind of objection is threefold: Yes, it is a danger; the benefit far outweighs the danger; teachers have a duty.
Firstly, I think there is a genuine danger here though I think it’s vastly over-represented by people who do make this objection. Speakers of Latin and ancient Greek are almost all incomplete learners. Some of us are more incomplete than others! And so mistakes are made. And sometimes mistakes are made and perpetuated, within the speaking community, in a way that makes me cringe.
However, no one I know is incognisant of this danger, no one cavalier, no one thinks it something to just shrug off. Perhaps there are people with that attitude, I don’t know. Our goal, almost always, is to pursue a representation of the language in our communication that reflects and approaches the patterns of the literature corpus we are interested in. That’s the norming norm for our situation.
The benefit of communicative approaches, and particularly in terms of actively speaking and conversing and writing in these languages far outweighs the danger of bad representation. If you think the danger is, “oh, a learner will think that XYZ is standard but it’s not”, what exactly damage is that going to cause? They’re unlikely to go and misread some text that does have the standard feature. They’re no more likely, when it comes to biblical studies for instance, to get things wrong than the woefully less complete and accurate representation of the language that a non-communicative student of equal attainment has. Essentially, someone who speaks Latin is still far and away going to end up a better reader of Latin, even if their speech perpetuates some non-standard features.
Thirdly, teachers have a duty. This goes back to the norming norm – insofar as we are aiming to reflect a particular corpus (e.g. classical Latin. historically broad Latin. Koine Greek. Classical Athenian Greek. Greek across 2000 years. that corpus can be narrower or broader), our language communicatively should aim to reflect that. Which means in terms of self-monitoring and editing and particular production of learning materials, I do think teachers have an obligation to be checking and correcting their own materials to those standards of the corpus. It’s why I’m always interested in the answer of, “who uses this structure?”, “is this attested?”, “attested where and by who?”.
In sum, we ought to strive to reflect an accurate representation of the language as we find it in our literatures, while not letting this paralyse us from actively using the language, which is a paramount way by which we improve our individual acquisition of that very same representation of the language.