One of the unusual things about me is that I have learned both Latin and Russian through the American educational system, and have taught intermediate-level college courses in each language at the same university (the University of Southern Maine–a school which now, under the current “Metropolitan University” regime, offers no foreign languages. But I digress.) Now, after studying both Latin and Russian for more than 30 years, it is remarkable to me how differently I know them. And that, of course, is a consequence of how I was taught.
One of the reasons I persisted in studying Russian was that I wanted the experience of having to speak and understand a language with a case system and flexible word order like Latin. Russian, you see, has three genders for nominals and six cases. (the spoken language has a vocative in addition, I’d argue). And word order, particularly in the spoken language, is very free, although there are rules, of course. Linguists have  characterized the elements of a simple sentence in terms of a “theme/rheme” or “topic/comment” structure. Basically, this means that the more newsworthy the information, the later it comes in the sentence. I apologize to any readers for whom this is old hat.
I’d like to present three hypotheses about Latin that have occurred to me only because of my study of Russian. I can’t claim that these observations are original, because there is a lot of fundamental German scholarship in Classics that I don’t know at all (I read German badly–too busy learning Russian!). But I have found none of these hypotheses discussed in English language scholarship.

#1. The yes/no question marker -ne was optional in Classical Latin; Latin had a distinctive intonation pattern to mark a yes/no question.

We all teach -ne, nōnne, and num to our students, but I have never heard another Latinist remark that -ne was not required. The most famous example of this is probably In Cat. 1, 2: Senatus haec intellegit, consul videt; hic tamen vivit. Vivit? immo vero etiam in senatum venit … . There must have been a contrasting intonation for the second vivit that Cicero used to mark it as a question.
This second vivit is, of course, a “rhetorical question,” as well as an example of anadiplosis, and I have a sense that the -ne marker is very rare with rhetorical questions. Whether my sense is right or not falls outside the bounds of this CANE blog post.
The Russian connection concerns the yes/no question marker li. This enclitic is much more common to mark a question in the literary/written language than in the spoken language; the spoken language uses multi-syllable intonation patterns to mark a yes/no question as opposed to a statement.
It’s worth noting in passing that there are substantial differences between colloquial Russian and the formal, literary language which educated people often use in speech. These differences are largely borrowed foreign words, mostly from English, slang, and slurring or omission of syllables in longer words, not simplification or mistakes in the grammatical system. Case/verb conjugation mistakes are rare, and actually embarrass the speaker.
In any case, I’ve encountered many Russians on the street there who I struggle mightily to understand. Interestingly, virtually all of them are men. Didn’t Cicero remark somewhere that the Latinity of Roman aristocratic women was better than that of their husbands?

#2. The Latin case system is surprisingly ambiguous.

I have both heard and read classicists remark that Latin is a “precise” language, and I’ve always found this a peculiar thing to say. Words in a “precise language” might have fewer referents, for example. By that measure, Latin is very far from precise (in contrast to Ancient Greek or English, for example): the Romans had the consistently alarming habit of re-applying words to new and different (e.g. technical, military, marine) things. Testūdō, for example, means (1) a tortoise or turtle; (2) its shell; (3) a military maneuver with interlocked shields; (4) a wooden shed serving the same purpose as (3); (5) a roof having four sides converging to a point; and (6) some feature involved in the heating system of a bathhouse. Egad.
Does the case system make Latin more precise? Well, it might, if it didn’t have so many ambiguities itself. We all know the places in the nominal chart where a case form is identical to some other case form. Russian is informative in contrast: it does share with Latin the identity of all neuter nominatives and accusatives (this is an inherited feature); some nominative plurals and genitive singulars are the same for both masculine and feminine, but not all (they are distinguished by word stress, which is not fixed on the same syllable for the forms of some nouns (like Greek). Most importantly, dative and instrumental and dative and prepositional case forms are distinct for nominals (except some feminine singular nouns and adjectives). Here I should explain that there is no “ablative” in Russian; instrumental and prepositional (or locative) cases are distinct. Latin’s ablative case combines what were three distinct cases (ablative or ‘separational’, instrumental and locative) in the parent language.
In other words, there is no puzzling over famous phrases like Vergil’s “incute vim ventis” (Aeneid 1, 69) to decide if a word is dative or ablative (“duplex sensus est“, says Servius). Moreover, we need to take Servius’ words to heart. The way we learn Latin conditions us to think that every word fits one slot on a chart somewhere. I am quite confident that the vast majority of Romans could not name a noun’s case, even if their lives depended on it. And yet, they spoke “correct” Latin.
Talking to Russians about their language has convinced me of the rightness of this. There’s been universal education since the 20’s; they are all taught the names of the cases. But this information isn’t retained, except by those who pursue higher education/are part of the intelligentsia, because it has no purpose. They just know what sounds right.
Take this famous sentence from Caesar’s description of the invasion of Britain in book 3 (c. 24) of the DBG:
Erat ob hās causās summa difficultās, quod nāvēs propter māgnitūdinem nisi in altō cōnstituī nōn poterant, mīlitibus autem, īgnōtīs locīs, impedītīs manibus, māgnō et gravī onere armōrum oppressīs, simul et dē nāvibus dēsiliendum et in fluctibus cōnsistendum et cum hostibus erat pūgnandum, cum illī aut ex āridō aut paulum in aquam prōgressī, omnibus membrīs expedītīs, nōtissimīs locīs, audācter tēla cōicerent et equōs īnsuēfactōs incitārent.
Every time I read this sentence, I think militibus is ablative until I reach the gerundives. Then I rethink and conclude it must be dative of agent. But I highly doubt Caesar thought of it this way. Perhaps he meant it in Servius’ duplici sensu, meaning that the militibus functions both as an ablative and dative here.
Iphigenia is sacrificed by her father at Aulis. Or not. Both are real myths.

#3. Hyperbaton is a feature derived from colloquial speech in part. It is not entirely an artificial figure of rhetoric.

Take this sentence, translated word for word from Russian:
Avia! Tibi Tata donum pulchrum misit, librum!
Hardly eventful for a Latinist. But in Russian the neutral word order would be:
Avia! Tata tibi misit librum, pulchrum donum.
So, it’s clear in Russian that the first sentence is quite emotive–maybe a child is speaking.
So, my point: we should associate non-standard word order with emphatic or emotive utterances as much as we do with literary artifice. I’m not asserting that what’s true in Russian must be true in Latin, but rather that common tendencies do operate in languages with flexible word order. For words can only move earlier or later in a sentence. Moving them earlier, or “fronting” them, gives them emphasis; they pop out, almost as an interjection, before their normal place in the sentence. Moving them later suggests that they are a second thought.
Enjoy your warm and sustaining winter vegetables, everybody.