Reconstruction talk:Proto-Germanic/kumþiz

mft
I don't see a particular need to reconstruct -mft- here. It's a German-specific change. —CodeCat 20:56, 23 July 2015 (UTC)
 * That's what I thought. But Gothic shows two reflexes of *-mþiz, one as -mþs, the other as -mts. Kroonen lists the reconstructions as -mþiz but references Rasmussen who favors the reconstruction as -mftiz. According to Rasmussen the cluster -mþ- saw epenthesis of -ɸ-, turning it to -mft-, but this obviously postdated the Verner's alternation of -md-/-nd-. He says therefore that Gothic -mts must be simplified -m(f)ts, while the -mþs would be easier to explain as analogical. would also have seen this type of epenthesis.
 * I don't really mind undoing this if you prefer the reconstruction as -mþiz, though. Anglom (talk) 21:17, 23 July 2015 (UTC)
 * These explanations are nice, but I don't see much evidence. At least for there is clear and overwhelming evidence across Germanic for a cluster -ft-. But for these remaining ones, I don't see nearly the same kind of clear evidence. For, German indeed has -mft-, but Dutch has -mst-, and Gothic has -mþ-. If there were an attestation of long vowel + -ft- in an Ingvaeonic language, that might be more convincing. It has also been suggested (by Ringe, at least) that the regular outcome of -þVs and -dVs in Gothic was actually -ts in first instance, but this later got restored back to a fricative -þs by analogy. One place where analogy could not operate is in the second-person dual ending, where Ringe's proposed ending -diz does indeed become -ts in Gothic. So the -ts found in some nouns may well just be the original outcome. I certainly don't see it as evidence for -mft-, where did the -f- go otherwise? —CodeCat 21:33, 23 July 2015 (UTC)
 * is only attested in West Germanic. Unfortunately, none of the rest of the pertinent words are attested in any other West Germanic language. I don't really mind the reconstructions being one form or the other, we don't really have enough evidence to say which one is the clear Proto-Germanic reflex, but I feel we should be consistent in the reconstruction we choose. *kumþiz and *samþijaz or *kumftiz and *samftijaz. It feels noncommittal to reconstruct two different outcomes for the same environment. Anglom (talk) 22:01, 23 July 2015 (UTC)
 * But can we even be sure these are the same environment at all? At the very least, if Gothic has -mþ- preserved, that speaks for the change to -mft- being post-Proto-Germanic. Also, compare Middle Dutch . —CodeCat 22:14, 23 July 2015 (UTC)
 * It should be, unless we think *samftijaz received epenthesis earlier. Its supposed cognate Sanskrit shows the preform should be, I think.
 * The Dutch forms are interesting, and should point to -mþiz, but how can we be sure that they aren't secondary formations in (productive after stems ending in nasals) and . The latter was still very productive in Dutch, yes? Anglom (talk) 22:42, 23 July 2015 (UTC)
 * Yes, and is not the natural descendant of . The true descendant would not have a final -e. But I'm not sure where the -st- ending comes from. It's also found in  and German . It's entirely possible that languages felt a need for a separating fricative to break up the uncomfortable sequence of labial nasal + interdental fricative. Different languages may then have chosen different ways: some inserted a labial fricative as a "bridge", others used a sibilant. This can be compared to Latin  from . —CodeCat 23:45, 23 July 2015 (UTC)

This gliding f after original m occurs more often in High German: Brunft, Zunft (both ti-abstracta as well). The dissimilation mf > nf is secondary (and not always a reality in pronunciation). After n the gliding sound is s: Brunst, Kunst. 77.13.49.118 16:24, 31 March 2021 (UTC)