One wonders why some languages easily drop nasals with nasalization only lasting a short time, while others keep it, why is it that French keeps nasalization that developed in the early middle ages, while for instance Germanic languages have more or less lost the Proto-Germanic nasalization completely, apart from maybe frisian? and dalecarlian.Chagen wrote:IIRC Sihler's New Comparative Grammar of Greek and Latin states that Latin final /m/ was pronounced as some kind of nasalized /w/ dropped in casual/informal speech, since it makes position at the end of a line in poetry but does not block elision. The Roman grammarians even had a term, "mytacism" whose true meaning is obscure but appears to be the mistake of pronouncing final /m/ as...an actual [m].
Does the amount of minimal pairs really matter? French is loaded with homophones.
It is interesting to see that PIE, a language making me think of Czech or Arabic, filled with "impossible" consonant clusters, developed into stuff like Danish and French, that just sound like garbled vowels and approximants.