Zompist Bboard Again

Salmoneus wrote: ↑Thu Sep 05, 2019 6:02 pm 166 is a lot, but 50 is hardly extreme. Many modern NWC languages have in the 45-70 consonant range.

But 50 is a good deal more than most "Nostratic" languages have. Most "Nostratic" proto-languages are in the range of 20 to 30 consonants (e.g., IE: 25; Uralic: 17-20), only Proto-Kartvelian and Proto-Afroasiatic lie above that (PAA: 31 according to Wikipedia, some other reconstructions have a few more; Proto-Kartvelian: 37).

But this is not the place to discuss the matter in depth. The whole point was that adding more phonemes to the proto-language is a cheap trick to gloss over unresolved questions.

WeepingElf wrote: ↑Fri Sep 06, 2019 9:24 am But this is not the place to discuss the matter in depth. The whole point was that adding more phonemes to the proto-language is a cheap trick to gloss over unresolved questions.

It can be dangerous, of course, just as reconstructing clusters can be.

But it can also be correct, just as reconstructing clusters can be.

Indeed, in the case of something like NWC, if the two branches weren't so close together geographically, I think you'd have to say it was probable that the parent had even more phonemes. Having 50 consonants is very unlikely to happen in multiple related languages due to independent inventory expansion, one would think, but very likely to happen independently in multiple daughters if the parent has even more distinctions!

[but in this case that principle gets messed up by the, in my view, more likely scenario that the massive inventories are due at least in part to areal influence, rather than independent development.]

About a dozen of Bomhard's proto-consonants are only based on the evidence of one branch; a few of these are nominally continued in a larger number, but in practice the overlap between their evidence is too small to make any difference.
– alveolar vs. postalveolar sibilants *cʰ, *cʼ, *dz, *s | *čʰ, *čʼ, *dž, š (only distinguished in Kartvelian)
– velar vs. uvular stops *kʰ, *kʼ, *g, *kʷʰ, *kʷʼ, *gʷ | *qʰ, *qʼ, *ɢ, *qʷʰ, *qʷʼ, *ɢʷ (distinguished in Kartvelian, Central Cushitic [a subfamily of a sub-family!] and Chukotko-Kamchatkan)
– velar vs. pharyngeal fricatives *x, *ɣ | *ħ, *ʕ (only distinguished in North Afrasian plus Southern Cushitic; the latter only in Semitic)
There's even a *ʔʷ posited in just one word, that seems to be based on no evidence at all and could be immediately merged with *w.

You could easily rack PIE up to 40-ish consonants by this type of maneuvering: e.g. reinstate the voiceless aspirates, reconstruct a series of "half-palatalized velars" whenever Baltic has plain velars corresponding to II first palatals, set up an *ɫ just for unclear reflexes in Albanian or Armenian…

Yes, that is precisely why I am skeptical of such bloated protolanguage phoneme inventories. As some of you know, I consider it likely that the "Mitian" langauges (IE, Uralic, Turkic, Mongolic, Tungusic, Yukaghir, Chukotko-Kamchatkan, Eskimo-Aleut) are related to each other, while I am doubtful about the relationship of Kartvelian, Dravidian or Afroasiatic to them. And of the protolanguages of the eight Mitian families, PIE has the largest consonant inventory. The other MItian inventories are quite similar to each other: a row of voiceless stops at four points of articulation (labial, dental, palatal, velar; to which ChK and EA add uvular) with a matching set of nasals and of voiced obstruents which may be either stops or fricatives; in addition to that some sibilants and liquids. One would guess that the Proto-Mitian inventory was similarly structured. It seems as if PIE was the "odd man out" here, and the Proto-Indo-Uralic inventory (if that's a valid node within Mitian) more like the Uralic one.

Proto-Indo-Uralic is a legitimate concept even if one can't work out what else may descend from it.

And it of course belongs to the Macrofamilies thread; I was going to propose continuing the discussion there, but Howl beat me to that.

On the CONLANG mailing list, we are currently discussing the PIE stop types. Benct Philip Jonsson (who knows his PIE quite well) suggested that the PIE *D set may have been stiff-voiced, while the *Dh set was slack-voiced. What do you think of this idea? I think it makes sense - more sense IMHO than the glottalic theory. At least, this would give the *D set a highly "marked" character, as we can observe from their low frequency and their distributional constraints. (Also keep in mind that in many languages with ejectives, such as most Caucasian languages, these constraints are not in force, with worlds like Georgian k'op'e 'ladle' - with two ejectives in the root, one of them labial - occurring there.)

It may be that PIE *b was so rare because it merged with *w in most contexts, which would explain the initial *wr- and *wl- clusters found in PIE. This may mean that the *D set had a rather weak articulation and speaks against them having been ejectives, as [p'] is a very salient "popping" sound that is unlikely to merge with [w]. I don't know, though, whether labial stiff-voiced stops have a propensity to weaken that way or not.

Now for an idea on nominal morphology. Is it a coincidence that the genitive singular ending *-s (probably *-sa before ablaut) which does not correspond to anything in Uralic (if IE and U are related at all, of course) resembles the 3rd person pronominal root? My idea, which certainly has occured to others before me already, is that the ending resulted from a kind of head-marking construction. In an early stage, the possessum would have been marked with a prefixed pronoun, and the possessor places before that without a case marker, i.e. 'man 3SG-house' for 'the man's house'. Later, the prefixed pronoun was reanalyzed as a suffix on the possessor noun, i.e. a genitive case ending.

Why on earth would you think that?

This would require:

- possessive prefixes, in a language without any general pattern of prefixing
- posessive prefixes that match the nominative pronouns
- the third person prefix - and only the third person prefix - become dissociated from the noun to become a free-standing possessive particle
- genitives not marked for case, in a generally case-marking language
- free-standing possessive particles then being bound to the preceding nouns as case affixes
- that particular case affix then losing its final vowel

None of these things are impossible per se. But none of them sound particularly parsimonious assumptions, and there's a whole string of them. I don't see any actual reason to think that any of these things happened, and I don't see what you gain by assuming that it did.

The only justifications you seem to raise are:
a) two semantically unrelated morphemes share a certain phoneme
and
b) two languages, very unlikely to be closely related, do not share a particular item of morphology

The obvious answer to a) is "it's a coincidence", and b) "they're not closely related".

But then, the thing is that even if you refuse to believe either of those, then you can deal with b) much more easily without this hypothesis, and it's not even essential for a) either!

-

WeepingElf wrote: ↑Sun Sep 29, 2019 2:57 pmwhich certainly has occured to others before me already

This kind of idea certainly isn't new. The usual comparison is actually with the nominative singular demonstrative *so, which has the benefits of actually being nominative singular. It would then be interpreted as a kind of focus particle on the topical subject: "this the X" -> "the X" -> "X". One slight issue that needs addressing by a theory that assumes a pre-ablaut form *-sV for the nominative singular ending is why there are vowel-final endings at all (non-neuter dual *-h₁e, *h₂e-conjugation (-> perfect) 1sg *-h₂e). Since there is actually the possibility of an apocope rule in forms *-VHe (compare consonant-stem *-h₁e, o-stem *-oh₁, neuter *-ih₁) you could assume that this apocope occurred in *-Vse as well, and the vowelless ending was then generalised (but why wouldn't the dual generalise it too?).

-

I've not seen the retention of laryngeals proposed as a strong characteristic of anatolian vs non-anatolian. At best, it's a weak support for the theory (i.e. we know they were retained in anatolian, and only suspect they may briefly have been retained elsewhere but aren't certain). And of course, their retention in Anatolian could just be, theoretically, areal rather than genetic.

But it's kind of a moot point, since nobody would ever deny that Anatolian must be the earliest branch. Everything, from phonology to morphology to syntax to lexicon demonstrates that it's not just highly divergent, but positively archaic, compared to all other branches.

I suppose it would be possible to have a theory like "there are TWO (or more) anatolian families, one of which split off before the other, so both are archaic, but are actually maximally remote from one another genetically and one is more closely related to NAIE than to the other, but they look very similar due to areal effects". But I don't think there'd be any reason to have tthat theory...

mae wrote: ↑Fri Oct 04, 2019 5:52 pm I assume the strongest phonological evidence for an Anatolian branch as separate from Everything-Else is the preservation of laryngeals.

That goes against the cladistic principal of using shared derived characteristics, rather than retentions, to define a group.

It's also claimed that Vedic Sanskrit preserved laryngeals (though as only a single phoneme). One basis is some of the sandhi rules. There are also claims that the thematic suffixes of verbs was He/Ho, and that that laryngeal makes (or can make) position in Vedic scansion. Bello Witzel, Vedic tape recorders are broken.

There are also claims for Greek showing distinct reflexes of different laryngeals.

In short, laryngeals generally lasted into the fragmentation of IE, so no deep grouping can be based on their continued presence.

I think mae is right, .... it's just that, while that may indeed be the strongest evidence, it's not the only evidence, so in a way you're right too. Just no need to rush to extremes here. Anatolian preserves its laryngeals as *consonants*, which no other branch does, so it would be striking if it were merely a subgroup of some other branch whose other attested descendants all had the typical IE vowel coloring reflex. But Anatolian also differs grammatically from the other branches in ways not seen in other branches, thus providing further evidence of its early branching date.

Pabappa wrote: ↑Sat Oct 05, 2019 2:59 am I think mae is right, .... it's just that, while that may indeed be the strongest evidence, it's not the only evidence, so in a way you're right too. Just no need to rush to extremes here. Anatolian preserves its laryngeals as *consonants*, which no other branch does, so it would be striking if it were merely a subgroup of some other branch whose other attested descendants all had the typical IE vowel coloring reflex. But Anatolian also differs grammatically from the other branches in ways not seen in other branches, thus providing further evidence of its early branching date.

There are a few questionable statements here, so let me clear things up.

Firstly, as Mae themself just said, there's strong evidence that consonantal laryngeals were still present in PIE all the way up to the complete separation of branches. So, "which no other branches does" is just plain wrong. Secondly, are you implying that Anatolian doesn't have vowel colouring? It does. Every IE language does. It could be taken as one of the defining characteristics of being specifically IE.

Laryngeals must have been preserved in Graeco-Armenian (the triple reflex) and may have been preserved in Germanic (Cowgill's law) and Indo-Iranian (Kümmel's proposal for Persian h- < *h2-); there's also the glottal-stop-or-whatever reflex of the laryngeals that's been proposed for Vedic Sanskrit, Avestan, and Proto-Balto-Slavic.

Is there a good source on proto or early Celtic pronouns?

There is (or was) a user here who knows a lot about Celtic: Dewrad. You should try PMing him.

This is a good place to start. Dewrad was in touch with these people as well at one time, I believe. This site is about Gaulish, but the pronoun forms are probably not much different from Proto-Celtic, and besides you can always compare them to Old Irish to fill in any gaps.