Reconstruction talk:Proto-Germanic/sa

Forms with þ- and forms with h-
This article should distinguish better between Gmc forms stemming from þ- and those stemming from s-. The current version does not distinguish clearly. It gives the impression that for instance Dutch de, English the etc is cognate to Proto Gmc *sa, which is not correct.

A similar article (about *hiz) says correctly that only the h- forms of Dutch zij stem from Proto Gmc *hiz. This article should list these forms (haar etc) and omit zij because zij is etymologically unrelated.

--MaEr 10:57, 21 December 2010 (UTC)


 * In the case of the forms with h- or without, there are really two completely separate paradigms. However in this case, there is a single suppletive paradigm whose lemma form happens to be *sa. German shares all forms with *sa, only the forms  and  are the result of regularising the paradigm in favour of the þ- forms. But they are still cognates, since they descend from the same paradigm.  —CodeCat 15:50, 21 December 2010 (UTC)


 * This is not my understanding of descendant. Of course I see that one paradigm contains elements with completely different etymologies. But nevertheless, in my understanding of the term descendant there must be an etymological relationship, not just a "paradigmatical" one. Maybe we need a clear definition of the term descendant (maybe from a native speaker of English, which I am not).
 * By the way, we have a similar case in . This verb contains elements of two or three different sources, if I remember well. In the article, the descendants are correct, as far as I can see: Latin fu-, Slavic by- and Germanic wes-/was- are not listed, even if they belong the paradigm.
 * --MaEr 19:26, 21 December 2010 (UTC)


 * In a highly inflected language like Proto-Germanic, each form of a word is still perceived to be that word. There is no doubt that Proto-Germanic speakers perceived *sa and *þat to be the same underlying word, just different variations. And so there is only one lemma form, which for pronouns and adjectives is taken to be the masculine nominative singular *sa. It is pointless to have two separate lists of descendants, one for some forms and one for the rest. We don't say 'sing descends from OE singan, sang from OE sang' and so on. They are part of the same paradigm so listing the lemmas is enough.
 * Furthermore, the difference with *sa is that it already had the same suppletive paradigm in PIE itself. PG simply carried over that same paradigm without change. It didn't somehow inherit *sa separately from the þ-forms. They were inherited as a single common system.
 * Another issue is that there is no place anywhere else to list descendants of the þ- forms on. Which means we'd end up deleting a lot of useful information simply because there is no other place to put it. —CodeCat 19:31, 21 December 2010 (UTC)
 * I agree that some of the suppletives lack a clear "headword-worthy" form, so having a place to list them is needful. The only place these can be displayed is in the paradigm entry. But I do see the logic in the other side of the argument as well...167.1.146.100 20:02, 21 December 2010 (UTC)
 * Sorry about that. That last edit was mine. Somehow I got logged out. Leasnam 20:04, 21 December 2010 (UTC)

þē > thee
It says here that the descendant of Old English "þē" is English "thee". When I tried to edit the page to fix this, it seems that this gets generated by the code which automatically lists the descendants of Old English "þē". The problem is, there are two different Old English words "þē", one which became "thee" and one which became "the". They should have two different "desctree" templates. Can someone more knowledgeable about this template stuff fix this? AndreRD (talk) 11:11, 22 November 2019 (UTC)

*sai or *siz
This page claims that OE sē and OS sē both derive from *sai, but in OE it says that it comes from *siz. The derivation from *sai only works for OS, since PG*/ai/ > /ɛː/, but in Old English the expected form would be *sā. The OE form could derive from either *siz or *sa(through brightening maybe). Can anyone shed some clarity on this topic? 71.80.97.83 12:37, 29 April 2023 (UTC)