Repository navigation
Should Mahdi Karimi keep Mahdi as its given name, and Bard Larsen keep Bard? #867
Description
mahdi and bard are title words, so a name that starts with either loses its given name:
>>> parse("Mahdi Karimi") # title Mahdi, given '', family Karimi
>>> parse("Karimi, Mahdi") # title Mahdi, given '', family Karimi
>>> parse("Bard Larsen") # title Bard, given '', family Larsen
>>> parse("Ola Bard") # given Ola, family Bard (a trailing title word stays a name)
parse() records no ambiguity for any of them. This is not new: 1.4.0 reads Mahdi Karimi and Bard Larsen the same way, and Karimi, Mahdi as first Karimi with no last name. Both words came in with the 2014 title list (1a51555) with no recorded reason, and #737/#773 carried them into ar.toml (religious) and en.toml (occupation).
They surfaced in the groups view of the jrc-names local corpus (#865). For 17 persons there, some spellings read mahdi as a title and others don't; for 33, bard. Most title words in that list are real titles that some spellings leave out (sheikh, sultan, sir), so only these two are listed here.
Who bears them
tools/name_usage.py Mahdi Bard Bård, 2026-10-24:
| given | family | |
|---|---|---|
Mahdi |
SSA 2,398+ (US boys); Wikidata 141+ (Iran 43, Iraq 9); nam_dict band 5, Arabia/Persia | Census 1,241; Wikidata 13+ |
Bard |
SSA 218+; Wikidata 10+ | Census 5,953; Wikidata 100+ |
Bård |
nam_dict band 5, Norway | — |
-
Which box filers use: in
psc-individuals(Companies House; 1,034,888 distinct names in the 1-in-8 sample, each with the boxes the filer typed), no filer putMahdiorBardin the title box.Mahdiwent in a name box 195 times (forename 85, middle 66, surname 44) andBard16 (surname 13). -
Leading
Bardis mostlyBård: injrc-names, 31 of the 36 persons with a spelling that starts withBardalso have one that starts withBård. It is the Norwegian given name with its accent dropped. -
Distinct names reading the word as a title today:
Local corpus mahdibardjrc-names92 41 psc-individuals84 1 fec-individuals11 9 ctgov-officials6 0 No committed corpus holds either word, so the committed snapshot will not show a change.
As titles
- Mahdi is a messianic title, as in "the Mahdi" or
al-Mahdi. It is written as an epithet with the article or after a name, not as an honorific before a person's own name, so the leading-title reading serves no name we know of. - Bard: as an honorific before a name, I know of none.
Neither fits the exception in decisions/salutation-titles, which ships a title borne as a name only rarely in the leading position. Mahdi commonly leads a real name, and so does Bård written without its accent. AGENTS.md ("Titles permanently shadow first names") says to exclude such a word.
Proposal
- Remove both title readings, and record each word's counts as an
excluded = { title = "…" }entry, so a later contributor meets the reason as a build error. - Update the
mahdiline indocs/design/decisions/vocabulary-model.md(the list of Islamic titles moved toar.toml). - Add a release-log bullet and a
migrate.rst"Changed in" entry. 1.4.0 through 2.5 read both as titles.
Open: whether Mahdi should be a given-name title (addresses = "given") instead, so Mahdi Karimi keeps the given name.
Activity