Reconstruction talk:Proto-Germanic/-umnijō

Etymology
The preform would probably have been, as the cluster -mny- would presumably have given **-munjō, similar to how it developed in. Unless there were two forms *-umnij- and **-munj- depending on whether the preceding syllable was heavy or light, and then the latter was leveled out.

Gothic seems to show different genders, representing *-umnijō and *-umniją. I wasn't really sure whether to reconstruct both, or if Gothic shows independent development of one earlier form. Old Saxon reflects the former form, while Old English  reflects the latter. Maybe the former derives action nouns while the latter derives abstracts? Anglom (talk) 23:52, 3 November 2015 (UTC)
 * I found a source discussing this which I have listed on the page, but I feel a bit out of my depth when dealing with PIE. So here are some relevant comments:

The source of this suffix is Gmc. *-ubn(i)ja/ō / *-munja/ō from *-mn-yó- / *-mn-yéh2. Dialectal IE *wid-mࣹnyó-, a derivative of *wid-mén- (cf. Ved. vid-mán- ‘wisdom’) to *weid- ‘know’, gave Goth. witubni* (n) ‘knowledge’. The derivation is as follows: *weyd-men-yó- > *wy/id-mn-yó-. The maximize onset principle predicts *wi.dmࣹ.nyó- (with syllabic [mࣹ]). The alternative *wid-mnyó- (with only i and o syllabic) obeys maximize onset for mnyo but leaves a coda in wid, so the only way maximize onset can consistently apply to the entire word is *wi.dmࣹ.nyó-, the source of pre-Gmc. *witumn(i)jan whence, by nasal dissimilation, Gmc. *witubn(i)ja(n). In Germanic, Prokosch’s Law (the preference for heavy initial syllables) came into competition with maximize onset (Riad 1992: 45–62). For Gothic syllabications like [niþ.jis] ‘male cousin’, [nas.jis] ‘you save’, and line-end word divisions like rod-ja ‘I speak’, swis-tar ‘sister’, neþ-los ‘of a needle’, wiþ-rus ‘wether, lamb’, see §§2.11f. (cf. GGS 48; Bennett 1960: 28; Frey 1989; Suzuki 1995). The alternative syllabication is attested in Goth. lauhmuni* (lauhmoni Lk 17:24) ‘lightning; blaze’ (acc lauhmunja Lk 10:18, Bl 2r.12, dat lauhmon[j]ai 2Thess 1:8A) < pre-Goth. *l(a)uh.mun-jō < dialectal IE *l(o)uk-mnࣹ-yéh2 [*leuk- ‘light’] (NWG 277, 279) beside a syllabication *lou.kmࣹ.nyéh2 > Gmc. *lau.hum.nyō > *lau.hub.nyō, which possibly underlies ME levene, levin ‘lightning’ levin (cf. KM 130,AHDR 49).
 * Hope this is of use. — Mnemosientje (t · c) 09:47, 25 May 2020 (UTC)