I don't think that "warm" descended of "gʷʰer", because Gothic keeps the "g" as a "k" as in "𐌵𐌴𐌽𐍃" and also Proto-Germanic keeps the "g" as a "k" in “kwēniz".
- 𐌵𐌴𐌽𐍃 and *kwēniz both mean 'woman', not 'warm', they descend from PIE *gʷḗn. —CodeCat 15:56, 28 July 2012 (UTC)
I know, but I meant that:
Proto-Indo-European: "g" → Proto-Germanic: "k".
So I exclude "gʷʰer" from the etymology, as it'd be "kwenaz" in Proto-Germanic and as it'd be "quen" in English.
- It doesn't work that way. PIE *g did indeed develop into PG *k, but the sound here is not *g, it's *gʷʰ! The small extra symbols ʷ and ʰ don't just act as separate letters. They are there to indicate a modification of the base letter, and they form a separate phoneme altogether, which developed independently. So for that sound, there is a different rule. These rules together are known as w:Grimm's law and very shortly they can be explained as follows:
- p > f, t > þ, k > h, kʷ > hw
- b > p, d > t, g > k, gʷ > kw
- bʰ > b, dʰ > d, gʰ > g, gʷʰ > gw > w or g (except after n)
- There is another rule which applied after this, called w:Verner's law, which changed some of the sounds further. But I wonder one thing... Even if we assume (incorrectly) that the root was *gʷer- and not *gʷʰer-, why did you think that the r became n? —CodeCat 21:28, 28 July 2012 (UTC)
Thanks! I actually thought whether the "r" descended too and I attempted to write "kweraz" and "quer" in English, but that couldn't even have existed since it is too similar to "queer". The term "quen" might be too similar to "queen", too.
But I still propose "*wer", because it is a verb with one meaning, "to burn". So "warmaz" descended of the o-grade in the perfect or the past participle "wor-" meaning "burnt" and reduced the meaning a bit and put a "m" into it.
In "*gʷʰer" I don't know why it would have descended of the o-grade, since it's an adjective.
I also created a PDF-file for Grimm's law:  (download for higher resolution).