Should "Silva, Junior Carlos" read given 'Junior', rather than suffix 'Junior Carlos' and no given name? #1039

What happens. junior has one reading, the generation suffix (mul.toml since #1030). Leading the name it is read as a given name, but after a family comma it is read as the suffix even where nothing else is left for the given part:

parse("Junior Mensah")         →  given 'Junior', family 'Mensah'
parse("Kofi Junior")           →  given 'Kofi', suffix 'Junior'      (given-or-family reported)
parse("SILVA, JUNIOR")         →  family 'SILVA', suffix 'JUNIOR'    (no given name)
parse("Silva, Junior Carlos")  →  family 'Silva', suffix 'Junior Carlos'   (suffix-or-name reported)

The last row loses a given name the writer stated.

Bearers. Junior is a common given name: 43,182+ US SSA births. Wikidata gives it as a given name for 435+ people (US 60, Brazil 53, Peru 34) and as a family name for 68+ (Brazil 37). The FEC corpus writes JUNIOR 66 times, 52 of them after a comma. Brazil's TSE donors write it 15,370 times, mostly as the trailing generational word.

The question. Under decisions/vocabulary-collisions, does junior take a given reading, which would make it a member of the ambiguous suffix class the words-to-spare count decides (rules/S11)? What would that move? The trailing Brazilian Junior is the population to measure it against: snapshot.py --local tse-donors, plus the comma shapes in fec-individuals.

Part of #753.