If a text is published in all italics, that’s style/font choice. If a text is published using italics and roman contrastively and consistently, and everybody else is doing it pretty much the same way, that’s a convention.
But not all conventions are deemed worth of plaintext encoding.
English use of italics on isolated words to disambiguate the reading of some sentences is a convention. Everybody who does it, does it the same way. Not supported in plain text.
German books from the Fraktur age used Antiqua for Latin and other foreign terms. Definitely a convention that was rather universally applied (in books at least). Not supported in plain text.
In the first example, the mere need for disambiguation tells you that contrastive use should be possible: while some cases might not be truly ambiguous, but just misleading the reader, the ambiguity implies that more than one alternate reading may be possible and thus the use of italics would be contrastive.
In the second example, some foreign words use the same spelling as German words; the convention makes clear which is intended, and dropping it where the author relies on it, might well introduce ambiguity. Most of that convention wouldn't be contrastive, but in some cases it easily would be.
In either case, you lose information that's related to content and not merely to dressing up the text in a "pretty" way.
Still, not supported in plain text (unless you abuse the math alphabets for things they were not intended for).
Like so many general statements relating to Unicode, even this carries its exceptions.
A./
This archive was generated by hypermail 2.2.0 : Wed Jan 09 2019 - 01:30:35 CST