Reconstruction talk:Proto-Germanic/ahaz

Ax
Does Old Norse ax also come from ahaz? Bgagaga 02:37, 22 February 2012 (UTC)
 * No, normally h disappears in Old Norse except at the beginning of a word, so ahaz would have become ár. But in the combination -hs- it becomes -ks- (x) so 'ax' would probably come from Germanic ahs- or similar. —CodeCat 17:50, 22 February 2012 (UTC)
 * OK, so do you think ahaz and ahs (or similar) were synonyms in ancient Germanic? Bgagaga 19:05, 23 February 2012 (UTC)
 * I don't know. They do seem related but how the relationship is I don't know exactly. —CodeCat 19:20, 23 February 2012 (UTC)

I would posit that for Old Norse *ahaz > (**ahz >) ax (i.e. /ɑxs/). We know that the later vowel would be elided in ON., and I find it very dubious that -hz- could be sustained even for a year, so it would have been -hs- throughout almost immediately. We also know that -hs- (/-xs/) is generally kept intact, in words like ( > lax, not *láss). Therefore I do not see the need for a separate reconstruction for North Germanic at the proto-stage. Otherwise, we could put ON. ax under a separate *ahsan, but I advise against it. – Krun (talk) 17:39, 26 July 2012 (UTC)
 * That is a good point. But it depends on whether the loss of -h- preceded or followed the change -az > -z. What does the runic evidence show about this? Even so, though, it's hard to tell one way or another without other cases that are similar. Are there any other Germanic words that show -haz > -x in Old Norse? 18:06, 26 July 2012 (UTC)
 * I have not found any direct analogues yet, but (or *akwizī) is somewhat similar (at least in oblique cases and plural), with *-kwiz- / *-kuz- > *-ks- (*-hs-). /x/ and /k/ are similar sounds and might have worked the same. Other sounds that might have caused the devoicing of the z are /p/, /f~ɸ/, /t/ and (obviously) /s/. – Krun (talk) 18:38, 26 July 2012 (UTC)
 * I don't think you can really compare them. They are both results of a loss of vowels, yes, but that's where the similarity ends. The word *akwisī, for starters, already contains two voiceless consonants, so there is nowhere for devoicing to occur. It may be argued that the devoicing did occur to the voiced -z- in the other case forms, but the fact that several West Germanic languages have -s- despite not having lost the intermediate vowel (a requirement for voicing assimilation to occur) suggests that those forms generalised the -s- in favour of the -z- in all the forms. Gothic is the only descendant to have any trace of the -z-, and it too eliminated the consonant alternation but in the opposite direction.
 * Regarding *ahaz, there is another important point to make and that is the fact that in many Category:Proto-Germanic consonant stem nouns, voicing assimilation probably did occur in Germanic, but was actually undone in Old Norse. See for example, which had -s in Germanic but -r in Old Norse. That is of course an analogical change, but if it happened in *mēnōþs, why not also in *ahs < *ahaz? And also one further note... Germanic didn't have or allow the combination -ks- (because of the Germanic spirant law), so if it arose in Old Norse, it must have been formed after the spirant law had ceased to be effective.  20:51, 26 July 2012 (UTC)
 * Regarding *akwisī I concede that based on West Germanic there must have been an existing s; however, I don't believe regularization based on nominative alone is very common in Old Norse, and there are also no alternative inflected forms with r (e.g. *ǫkri, *ǫkrar). This seems to me to point to a regular change of the z in the inflected forms (although the indication is perhaps not very strong – this is all assuming the reconstruction is correct). I also never said or implied that -ks- occurred in Proto-Germanic; I assume the vowel loss occurred long after Proto-Norse began to diverge from Germanic. In any case, -ks-/-x(-) can occur in ON., likely pronounced /xs/ (original gs, hs, and ks probably all merged into /xs/). As for consonant stems such as *mēnōþs, these are very different in that the -s only appears in the nominative, and as this is a rare class, it is very susceptible to the analogical change you mention. In *ahaz and other z-stems, it is an essential quality of the z that it appears in all inflected forms and doesn't change and they regularly merged with -an neuters to form one endingless strong neuter class. Therefore, *ahs cannot undergo the same change as *mēnōþs, because: a) the -s in *ahs is not analogous to the -az (now -z, i.e. -ʀ) nominative ending; it is simply the end of the stem in a neuter class with no nominative ending (just like barn or any other original -an or -az neuters) and is present throughout the paradigm; and b) -hs is now a cluster, one which is regularly kept intact (*ahs should at this time function exactly the same as *fahs < ). – Krun (talk) 00:41, 27 July 2012 (UTC)
 * Update: I've since discovered that aha > ā must have happened before z > r, based on *fahaz > *fāʀ > . This one, however, also has a related form *fahsą > . – Krun (talk) 00:01, 29 August 2012 (UTC)