Should vanden be a particle like vande and vander? (Pieter Vanden Broeck gives middle 'Vanden', family 'Broeck') #1239

parse("Pieter Vanden Broeck") gives given 'Pieter', middle 'Vanden', family 'Broeck'. Vanden is the Flemish contraction of van den, written as its own word (Vanden Berghe, Vanden Eede, Vanden Bossche). nl.toml lists the other two contractions, vande and vander, but not vanden. tools/vocab/query.py --word vanden finds nothing, not even an exclusion.

Measured (2026-11-08):

• Belgian parliament lists (OpenSanctions be_flemish_parliament, be_senate): 2 of 249 names, both read with middle 'Vanden'.
• ORCID: 65 Belgian family boxes and 1 Dutch one open with Vanden.
• config_diff.py score default 'expr:Parser(lexicon=Lexicon.default().add("vanden", kind="preposition"))' --group BE --group NL --group Belgian --group Dutch:
  • ORCID BE: 95.10 → 95.28 (36 fixed, 1 broken)
  • PatentsView BE: 97.28 → 97.34 (19 fixed, 8 broken)
  • PSC Belgian: 2 fixed, 0 broken
  • NL groups: unchanged
• On PatentsView BE and NL, 24 of 26 names holding Vanden move, all vanden middle → family.
• Bearers: US SSA 154 male births, US Census under 100, Wikidata none found.

Questions:

• Never-given, or ambiguous like vande (which carries a given reading on the conservative default) and vander (US SSA 1,828)?
  • Never-given folds a leading Vanden Broeck Pieter into one family (P1, the never-given particle fold).
  • Ambiguous keeps Vanden given there and reports the fork.
  • Both read Pieter Vanden Broeck right.