Also from the previous post:
the Dahalo word for ‘mother’, which is /jáájo/ and famous for being the only word in the language that has /j/
Looking over Ehret's work on Southern Cushitic also shows that this factoid is, alas, false news; he lists plenty of Dahalo words with /j/, also e.g. jààgo 'cow', jaar- 'to grow old', jab- 'to save', jálala 'mane', jaʔaṯu 'gonorrhea', jéjoʔo 'wild dog', jem- 'to stay in a place'…
I have indeed noted an issue in phonological descriptions of Dahalo where they go "we did not document word X so we don't have grounds to think a phoneme that supposedly occurs there exists", as if phonological description should be only ever determined on the basis of one research session with a few informants (ignoring previous research). And the sessions by the Maddieson et al. team, from whom the "/j/ as ultramarginal" claim comes from, must have been superficial indeed if they didn't even capture a fairly basic term like 'cow'!