Thread by @azforeman

Neither Dante nor anybody else in the Late Middle Ages seems to have had any idea that the Romance languages were the descendants of Latin.

One of those things that just seems so obvious in retrospect, but they really had no idea.

During the 9th-11th centuries, people gradually came to think of Romance and Latin as separate languages, rather than simply different registers of the same language (as they had been thought of before).

At that point, the idea that they were (or had been) in some sense the same thing vanished. By the late 12th and early 13th centuries, when the sense emerged that Romance varieties were themselves distinct languages from each other, their similarity to was regularly remarked upon

But the idea of these as being descended from Latin is never once mentioned, and does not seem to have occurred to anyone.
For speakers in the Late Middle Ages, then, Romance had no necessary connection to Latin apart from sharing a lot of stuff with it.

This was so self-evident to people that there was rarely reason to state it overtly. Many, if not most, seem to have simply taken it for granted that the Ancient Romans wrote Latin but spoke Vernacular, just like they themselves did.

For Romans to have left no trace of this mysterious "Vernacular" wouldn't be surprising to thinkers who lived in a world in which most vernaculars they knew of were rarely if ever written down. Such things — they might assume — could easily have not gotten copied.

They understood, better than we usually give them credit for, how manuscript culture worked. But I doubt they even had to think through it that far most of the time.
Dante for his part seems to have thought of Latin grammar as an invention in an almost literal sense.

Latin to his mind was basically a Conlang that the Romans put together because they needed a stable, changeless medium to write in, whose structure — unlike the Vernacular — could be precisely formulated in rules.

This is not that strange. Belief that the higher-prestige liturgical and literary language was formulated by conscious effort is not that unusual cross-culturally. The first works even vaguely approaching grammatical description of Romance were produced after Dante's death.

He didn't seem to think Romance was governed by rules of any kind. At least not in the sense we usually mean when we say "grammatical rules."
This too is not at all a strange attitude.

Lots of Americans don't think AAVE has grammar either, and probably the majority of Arabic-speakers even today will find the notion that spoken Arabic has its own grammar to be silly (even though suspicions that it does have grammar are documented as early as the 14th century)

It is good to bear this in mind when one considers the language of the Commedia, where Dante feels free to alter a noun's gender or concoct new participles to eke out his rhymes, while throwing in words of his own manufacture according to stylistic necessity or personal fancy.

Dante probably didn't think he was breaking the rules of Tuscan vernacular, or using it incorrectly. I doubt it ever occurred to him that it was even possible to do so. For him, there were no correct or incorrect ways to write Tuscan Romance.

Perhaps, instead, there was only what worked (what suited his communicative and creative needs) and what didn't work (whether because it might be misunderstood, because it would not sit well with his audience's attitudes, or because it would sound incoherent.)

It is only in the early 15th century that some men like Poggio Bracciolini and Leonardo Bruni came to understand that Italian and other Romance varieties were what Latin had now become as a result of linguistic change.

16th century writers on French pronunciation, remarking on the relationship between French words' pronunciation and that of Latin cognates, develop various ideas as to the possible relationship to Latin. The idea that Romance is Latin's descendant was contested for centuries yet.

Not till the 1700s was there was anything approaching a "scholarly consensus" that Spoken Latin had turned into Romance.

Even now many scholars still have a hard time accepting that the Roman Republic was not diglossic.

A millennium and a half of cultural baggage presses down quite heavily on the brain.