Talk:фонарь

@Benwing2 I'm curious how do headword and declension templates know about the stress position here? --Anatoli T. ^{(обсудить}/^вклад) 06:37, 23 January 2016 (UTC)

Well, one thing is that stress types b and d default to last-syllable stress on the stem, and stress type f defaults to first-syllable stress on the stem (hence nom pl го́ловы [type f] not *голо́вы, but nom pl сапо́жки [type d] not *са́пожки). фона́рь is type b so the last-syllable stress might be coming from here. But I think it's rather coming from the fact that, as type b, the stress ought to be on the ending in the nom sg, but the ending is non-syllabic, so the stress migrates to the closest syllable, which is the last one. This would explain why the gen pl of голова́ is голо́в not *го́лов even though there's stress on the stem here. It seems to go like this: Pattern f calls for stem stress in the nominative plural (and inanimate accusative plural), and ending stress elsewhere in the plural. The stem stress in the nom pl reveals the underlying stem stress, which is first-syllable for pattern f, hence го́ловы. In the genitive plural, however, pattern f calls for ending stress but the ending is non-syllabic, so it's forced onto the last syllable of the stem, and the underlying stem stress doesn't apply, hence голо́в. Does that make sense?

BTW if you omit the stress in the lemma with types a, c and e, you get an error, but for types b, d and f, which are ending-stressed in the nominative, this is allowed, and defaults as above. Benwing2 (talk) 06:55, 23 January 2016 (UTC)