Should el be listed as the article in Moroccan and Egyptian surnames? (Sara El Idrissi gives middle 'El', family 'Idrissi') #1240

parse("Sara El Idrissi") gives middle 'El', family 'Idrissi'. El is the Maghrebi and Egyptian spelling of the Arabic article that ar.toml lists as al, written as a separate word in surnames (El Idrissi, El Amrani, El Khattabi). el has no vocabulary entry. decisions/vocabulary-collisions has left it "needing judgement" since 2026-08-17 (the C-i correction's call to re-judge lo and el), and no issue tracks it.

Measured (2026-11-08), config_diff.py score with el as an article (never-given) and as an article plus a given reading (ambiguous, as al is):

• Never-given breaks about 300 Moroccan given names that open with El (El Mehdi, El Hachmi, El Mostapha), which P1's fold would turn into family. Ambiguous keeps them: El Mehdi Bennani gives given 'El', middle 'Mehdi', and reports particle-or-given.
• Algeria loses under both. Its 221 broken are compound given names with the article inside (Nour El Houda, Mohamed El Amine, Chems El Hoda). Ambiguous el joins those forward into the family: Nour El Houda Mansouri gives given 'Nour', family 'El Houda Mansouri'. Today the default reads middle 'El Houda', which the ORCID boxes score as right. This is P10's shape (a given name a particle completes), the Arabic counterpart of María del Carmen, and al presumably has the same cost today.

Questions:

• List el as an ambiguous article (with a given reading), as al is? Which file: ar.toml beside al, or a Maghrebi tag?
• Should the second words of Arabic compound given names (houda, hoda, amine, dine, khalil) get P10's after_particle reading, for el and al alike?
• Spanish el (Juan El ...) barely appears: ORCID ES gains 14 and breaks none.