Reconstruction talk:Proto-Germanic/dajjaną

Inflected forms

The combination -ji- would not have been allowed in Proto-Germanic, so a form like *dajisi would have to be *daisi. On the other hand, is it possible that the verb followed the strong conjugation instead? Then the inflection would be *dajjizi, which might have been possible. Or maybe *daīzi is more likely. Gothic is no help in this case, because it attestes the verb only as a present participle, which doesn't tell us anything about its inflectional class. So the two conjugational patterns would be:

• Strong class 7: present dajjizi? daīzi?, past dedajj? dedaī? dedai?
• Weak class 1: present daisi?, past daidē?

The combination -jj- is so exceedingly rare in Germanic that it's hard to say much about it at all. In fact I know of no other verbs, strong or weak, that contain it. Do you? —CodeCat 21:56, 20 August 2012 (UTC)

I have been able to find only two others: *hnajjōnan (to neigh) and *plajjanan (to move, play). I am leaning toward the weak 1 conj, with the alteration/elision. Anomlalies are a natural and normal part of every language, so I have no issue with it. Leasnam (talk) 22:15, 20 August 2012 (UTC)

Anomalies do occur in every language, but they should always be treated with suspicion when they lead to grammatical anomalies. Anomalous forms in words that are not commonly used tend to be regularised much quicker, and something like a dajj-/dai- alternation is obviously anomalous to the point of likely being unique. Those two verbs are interesting though, especially *hnajjōną. What languages are they found in? And is the past tense attested in those languages at all? —CodeCat 22:27, 20 August 2012 (UTC)

Well, I would love to report that Gothic has a descendant, but we only have one verb, OE hnǣġan, and a Norse deverbal in gnegg. I can't find any descendants (or derivative for that matter) for *plajjanan. Leasnam (talk) 22:33, 20 August 2012 (UTC)

I spoke too soon. Old Norse does have gneggja (< ga-hneggja) "to neigh (like a horse)" ( I hope this is not a denominal), which interestingly enough has returned to hneggja in the Modern language. I have not found a past in OE. I will check ON. There is one additional attest, which is a verbal noun in Middle Low German neiinge. Leasnam (talk) 22:40, 20 August 2012 (UTC)

I just can't figure out where OE deon goes: it is a weak 1, like *dajjanan is conjectured to be, but the form is hard to explain (unless dēon = *dēan < *dǣan/*dājan. The attested form lines up more with the weak 2 *dījōnan, and OE has a presend participle in dīende, suggesting an unattested alternative form *dīon, but the class is not right. If it were not for the fact that this root is a very important PIE root, I wouldn't worry too much over it; but it is useful for pointing to a lot of other English words like fetus, feminine, filial, etc. Leasnam (talk) 22:28, 20 August 2012 (UTC)

One way to check is to look at rhymes. Words that rhymed in Germanic would presumably continue to rhyme in Old English, so if they don't, then that is strong evidence that they are not cognate. There is one verb that rhymes with *dijōną, which is *frijōną. That word has been modified in Old English in the typical way for class 2 weak verbs, by replacing -on with -ojan > -ian, which in the case of this verb results in *frijōn > *frijōjan > frēoġan. That word doesn't rhyme with dēon, so it is likely not a cognate. The second alternative *dēaną (which I moved btw) rhymes with *blēaną and *sēaną. Verbs of this type (which are all class 7 strong for some reason) normally have different glide consonants inserted in them in the later languages, but differently in each. Old English normally inserts -w-, giving blāwan and sāwan. So, a hypothetical *dēaną ought to have ended up as *dāwan in Old English, which also doesn't match dēon.

So that leaves *dajjaną. If hnǣġan is a descendant of *hnajjōną, then we know at least that the combination -ajj- developed as though it was -aij-, with ai > ā in OE and then umlauted to ǣ. It probably developed like this in OE: *hnajjōną > *hnaijōn > (ai > ā, -ōn > -ōjan) *hnājōjan > (umlaut ā > ǣ, -ōjan > -ian) *hnǣjian > (simplify -ji- > -j-) hnǣġan. A similar development probably should have happened in *dajjaną then, so that it would have resulted in OE *dǣġan. The difference between that and dēon could be a result of the different conjugation class of the two verbs, with -ajjō- > -ǣġ- but -ajja- > ēo. We have no comparative evidence for or against it though, so it is still a blind guess. —CodeCat 22:52, 20 August 2012 (UTC)