Reconstruction talk:Proto-Indo-European/-ónts

Question

Hi, @Victar. Are you sure this is a noun forming suffix, it seems to me that this creates active participles out of athematic verbal stems. Do you know of any other nouns with this suffix aside the one for tooth? Because that one is often explained as the nominalization of the participle of a verb that meant to eat. Tom 144 (talk) 01:17, 10 December 2017 (UTC)

Hey @Tom 144, thanks for working on the entry. I added some sources to the pages. --Victar (talk) 03:09, 11 December 2017 (UTC)

@Victar, Any time. --Tom 144 (talk) 03:12, 11 December 2017 (UTC)

@Tom 144: Sorry, now that I had time to look into it again, I can see you're confusing this with the nt-stem *-onts. This is a specific early suffix for creating body part nouns. I'll add some more examples and sources. --Victar (talk) 03:29, 11 December 2017 (UTC)

@Victar. *-onts looks rather odd to me. The suffix was added to stems not to roots, therefore it cannot be classified as inherent to an ablauting paradigm, since paradigms are exclusive to nouns and adjectives with the R + S +E structure, while this suffix could form participles with a structure of R + S + S + E and even R + S + S + S + E (if you consider that the feminine bears an extra suffix). Plus, Sihler describes it as been accented when athematic, with a static zero grade on the root.

I'm dubious about this body part forming suffix, since Sihler himself says the word is the nominalization of a participle, and is the only body part derived from it. But latin sons is evidently derived from (Ø)-ónts and not from (é)-onts.

I personally think is kind of obvious that these were the same suffix but depending on the verbal stem, this could have been accented on the verbal stem or the suffix. As you can see on the sanscrit reflexes (भवन् , bhávan= accented on the verbal stem, अदन्, adán= accented on the suffix).--Tom 144 (talk) 04:06, 11 December 2017 (UTC)

@Tom 144: I think {{R:ine:Beekes 2011|pages=196–197}} explains the 3 types of nt-stems the best. Have a look and let me know your thoughts. {{R:gem:Kroonen 2013}} cites a "wider cluster of original nt-stems denoting body parts", where in Germanic, it appears to have been semi-productive. For what it's worth, Sanskrit is well know for shifting stresses, especially when nominalizing adjectives. --Victar (talk) 04:57, 11 December 2017 (UTC)

@Victar, I don't have access to Kroonen, but I do have Beeks. He doesn't mention the morphological entymon of h₃dónts. I'm also suspicious about *mh₂-n̥t-éh₂. De Vaan claims those descendants come rorm the root *mon-, linking them to latin manus hand. He says that mancus handles proves that the n belonged to the stem, therefore the derivation should be *mn̥-t-éh₂. Others have also linked it to greek μάρη reconstructing an r/n stem heteroclitic. But it has been rejected since that would require a root composed of a single *m.--Tom 144 (talk) 14:31, 11 December 2017 (UTC)

@Tom 144: *mh₂-n̥t-éh₂ is taken from {{R:gem:Kroonen 2013}}, where he also reconstructs *mh₂-n-ós > Latin manus and *mh₂-r-éh₂ > Ancient Greek μάρη (márē), the latter demonstrating that the original root lacks an n, for which he proposes Proto-Indo-European *meh₂- (“to beckon”). I recommend you grab a copy, if you can. --Victar (talk) 15:05, 11 December 2017 (UTC)

Either way though, my point was that we have several body-part nd-suffix words in Germanic, like *harundō (“flesh”), *waisundō (“wind-pipe”), and *mundō (“hand”), that appear to be continuations of PIE nt-stems. --Victar (talk) 15:15, 11 December 2017 (UTC)

@Victar I've got to admit now that you've pointed it out, that Kroonen's solution seems more reasonable.

Conserning the *-ónts suffix. For the explamples you gave, it seems that the germanic type was only productive when it was in combination with the -éh₂ suffix. Then shouldn't be reconstructed as -n̥teh₂? I cannot think of any reflexes for -onts meaning body parts aside for tooth. But spanish garganta, which is probably borrowed, showes the -n̥téh₂ suffix, that would serve as an argument to claim that -n̥téh₂ predates germanic. Latin also has planta which also meant the part of the foot that touches the ground. --Tom 144 (talk) 16:06, 11 December 2017 (UTC)

@Tom 144: Perhaps, but it could also be Germanic innovation, like how BSL turned *h₃dónts into and i-stem using the accusative. --Victar (talk) 16:23, 11 December 2017 (UTC)

venter

@Victar, what are the sources for this etymology of venter, because De Vaan does not list it, and it doesn't seem to work either morphologically or phonologically. Thanks! —*i̯óh₁nC^[5] 23:29, 11 December 2017 (UTC)

Hah, you're right. I wasn't thinking. I've deleted it. --Victar (talk) 01:48, 12 December 2017 (UTC)

@Victar: Thanks. You're right that the semantics are really appealing. —*i̯óh₁nC^[5] 02:03, 12 December 2017 (UTC)

Declension of nt-stem nouns

FYI, I'm reading {{R:hit:Kloekhorst:2008}}, and according to him, this is the paradigm of nt-stem nouns: -n̥t-s (nom.sg.) ~ (Ø)-ént-m̥ (acc.sg.) ~ (Ø)-n̥t-ós (gen.sg.).

Otherwise, my main source for the declension of nt-stems is Kümmel (2017): (Ø)-ónt-s (nom.sg.) ~ (Ø)-ónt-m̥ (acc.sg.) ~ (Ø)-n̥t-ós/(Ø)-ónt-os (gen.sg.).--Victar (talk) 05:38, 12 December 2017 (UTC)

@Victar: Yeah, Kloekhorst is espousing the Leiden model. Kümmel's Erlangen model is far more standard. I've been thinking recently that we should maybe show multiple declensions, since there is so much disagreement. —*i̯óh₁nC^[5] 05:47, 12 December 2017 (UTC)

@JohnC5: Kümmel is probably happy to use the Erlangen model because it seems to align best with PII. Both models though reconstruct the gen.sg. with -ós, but we're using -és. Any reason for that? I tried using |gen_sg=*(Ø)-n̥tós, but I guess we don't have that set. --Victar (talk) 15:15, 12 December 2017 (UTC)

What would the conditioning be for using -és or -ós? If we can't tell when to use which, we won't get far. Ringe reconstructs -s when the ending lacks stress, -és when the ending is stressed, and -os when the ending lacks stress and the preceding stem ends with -s-. —Rua (mew) 19:03, 12 December 2017 (UTC)

There's a belief that *e > *o when unstressed and in a closed syllable. The issue is there are a lot of counterexamples, and if was a functioning rule, it probably occurred in pre-PIE. On the other hand, the Latin has the *-es genitive in all athematic nouns and AG, *-os. I think some people just give both genitives saying the evidence is inconclusive. —*i̯óh₁nC^[5] 21:40, 12 December 2017 (UTC)

Germanic has -iz on consonant stems as well. —Rua (mew) 22:26, 12 December 2017 (UTC)

nt-stem agent nouns

I think there is also another type of agent noun nt-stem, ex. *h₁s-énts > Sanskrit सन् (sán, “being”), *kh₂-r-énts > Proto-Celtic *karants (“friend”), *ḱli-énts > Latin cliēns (“client”), *h₂uh₂-énts > Hittite [script needed] (ḫuḫḫant, “(great)grandfather”), many of which are being built off nouns and adjectives, not verbal roots. --Victar (talk) 17:52, 12 December 2017 (UTC)

I'm not sure about this one. These just seem like denominal present participles to me. —*i̯óh₁nC^[5] 21:45, 12 December 2017 (UTC)

@JohnC5: Denominal? Do you mean deverbal? --Victar (talk) 22:43, 12 December 2017 (UTC)

@Victar: No, I mean denominal verbs which make participles. —*i̯óh₁nC^[5] 23:22, 12 December 2017 (UTC)

@JohnC5: Gotcha. It's certainly possible. --Victar (talk) 23:41, 12 December 2017 (UTC)

āašant-

@Victar, Luwian āašant- is certainly the Anatolian universal packager -ant- which is affixed to neuter nouns, especially given that the word a-a-aš-ša is attested. Maybe these two things are related, but to say that this has an PIE original seems very suspicious to me. —*i̯óh₁nC^[5] 21:59, 12 December 2017 (UTC)

@JohnC5: Universal packager? Do you mean a gender modifying suffix? --Victar (talk) 22:46, 12 December 2017 (UTC)

@Victar: no, I mean a universal packager. Anatolian had a universal packager -ant- that began to be used as an ergativity marker in Anatolian. In this case, it would be the ergative usage. —*i̯óh₁nC^[5] 23:23, 12 December 2017 (UTC)