I can see two sides to this.
There's a basic discrepancy between text in 3XX authority fields meant to categorize the entity in relation to a controlled vocabulary and text meant for use as a gramatically appropriate qualifier added to the authorized access point. PCC has favored the former use, which can entail the use of plural terms and limited term choice to match a controlled vocabulary. RDA favors the latter, preferring to enter 3XX terms as they would appear in a qualifier. The same text matching logic which drives the use of the controlled term also drives the need to match the uncategorized qualifier text in 100 $c with the text in a 374 (or some other sematically categorized field). Though I'm not sure the 368 "other designation" field adds much semantic specificity to a term, the other 3XX fields do; e.g., "374 $a Professor" indicates that the possible 100 $c term is for an occupation or profession. It might also follow personal usage more closely than the authorized category term, e.g., "Professor" rather than "374 $a College teachers $2 lcsh".
I agree that the notion that systems would be able to select and add qualifiers to a 1XX automatically is dubious; but being able to categorize terms in $c represents a different use case, and one not covered by the use of controlled vocabulary only. I also agree that controlled vocabulary terms have more general purpose uses than terms tailored to individual cases. If we have to choose between these two practices, then I'd side with using controlled vocabularies. But do we have to choose? Given that the data element is potentially supporting to two different purposes, maybe we should be more tolerant of parallel 3XX fields.