To add to that, the grammar (and thus the meaning) is sometimes driven by the meter, which is a concept pretty much entirely foreign to many familiar with English prose.

Word order in general being at the whim of the writer is the bane of the modern learner, although we can be sure that this is a stylistic device and that the spoken word order would have been much more standardised. There's a wonderful line in "The Loom of Language" (wonderful book, wish I could find my copy) berating "the crossword puzzles of Cicero" vs the language he'd use to bawl out his slave.

> the grammar (and thus the meaning) is sometimes driven by the meter

What do you mean by this? The closest I can come to making sense of it is that some Latin words obligatorily appear as the second word in the sentence, but it seems like you must mean something different?

I'm guessing that the issue is just that in poetry many authors are willing to use word orders and constructions that would otherwise be regarded as questionable or ungrammatical—"metri causa".

I'm pulling from my studies of ~10 years ago, but IIRC there are points (noun and adjective endings?) where diacritical marks can change the meaning of the sentence. Unlike, say, Spanish, these marks are not written, and frequently have to be interpreted from language cues, including the meter.

That's basically correct -- some case endings differ only by vowel length and are therefore not distinguished in standard Latin writing, which doesn't indicate vowel length in any way. (Old Latin writing sometimes did.) "Via", in nominative case, meaning "the road" as the subject of a sentence, is written identically to "via", in ablative case, meaning "by the road", but the 'a' of the ablative case ending is longer.

However, you can't describe this as "the meter driving the grammar". The grammar requires the same cases and vowel lengths regardless of the meter you're trying to achieve. Rather, we modern-day people reading Latin poetry can use the meter to help us determine what was originally written. All of the causal influence is from grammar to meter, not the other way around.

That's pretty much what I meant... A naturally fluent reader of Latin poetry would have had to follow the same process of inferences (drawing from a far more intuitive and internalized model of the meter, of course).

Quite. Though, if by any chance you have reached a sufficient mastery of English prosody to understand the iambic pentameter, the dactylic hexameter really isn't such a stretch, notwithstanding one being based on stress and the other weight.

I so enjoyed the pattern-matching puzzle that I found preparing for the AP Latin (Virgil) exam to be.

> you can't practise talking to real people

Not to native speakers, but to semifluent and (rarely) fluent 2L speakers, you can. While there's always an argument that the loss of native speakers radically impoverishes our sense of a language and its potential, and that what we can figure out from writing is just a portion of what the language used to consist of, you can definitely have fun practicing spoken Latin with experts and dramatically improve your understanding of the language.

There are several spoken Latin programs, some of which use immersion methods, mentioned in the article. I spent a week at one of Nancy Llewellyn's programs about a decade ago, and mehercule can she speak fluently and extemporaneously about anything she chooses in Latin. :-)

If you are in the Seattle area, you can practice Latin, among others, at the ancient language Meetup. Latin is by far our biggest meetup (average 6 people ha). All skill levels welcome!

> you can't practise talking to real people

I have not tried this myself, but there is a skype channel used to converse in Latin [1]. I'm sure the quality varies widely, but it is there.

[1] https://www.youtube.com/watch?v=1T9V04fDh6o