Nostratic, Eurasiatic, Mitian, ...

Discussion of natural languages, or language in general.
Post Reply
User avatar
WeepingElf
Smeric
Smeric
Posts: 1630
Joined: Wed Mar 08, 2006 5:00 pm
Location: Braunschweig, Germany
Contact:

Nostratic, Eurasiatic, Mitian, ...

Post by WeepingElf »

These topics have arisen in several places on the ZBB, especially in the Great Proto-Indo-European Thread, so here I'm starting a thread for them (also meant as a stand-in for the late, lamented Nostratic-L mailing list).

So, what is this thread about? This thread is about the macro-family hypotheses mentioned in the title. Nostratic was first proposed by Holger Pedersen in 1903, who connected Indo-European, Uralic, Yukaghir, Altaic, Eskimo-Aleut and Afrasian (then Semitic and "Hamitic"), and revived in the 1960s by the Russian scholars Vladislav Illich-Svitych and Aaron Dolgopolsky, who counted Indo-European, Uralic, Altaic, Kartvelian, Dravidian and Afrasian as Nostratic languages. Eurasiatic is a similar proposal by Joseph Greenberg encompassing Indo-European, Tyrrhenian, Uralic, Yukahgir, Altaic, Korean, Japanese, Ainu, Nivkh, Chukotko-Kamchatkan and Eskimo-Aleut. Some Nostraticists such as Allan Bomhard combine both proposals by taking Eurasiatic as a subgroup of Nostratic. Mitian is a proposal by me, using a name coined by John Bengtson (if I am not mistaken), which encompasses Indo-European, Uralic, Turkic, Mongolic, Tungusic, Yukaghir, Chuktoko-Kamchatkan and Eskimo-Aleut.

The authors use different methods. The Nostraticists use (or claim to use) the traditional comparative method applied to lexicon; they are, however, often criticized for allowing much semantic latitude, imprecise matches (especially Dolgopolsky's work is riddled with cover symbols for only partly known phonemes), reaching down into daughter languages and other problems. Greenberg used mass lexical comparison, which is to the comparative method what asking the question is to working out the answer - the point where actual comparative work begins. Myself, I am trying to use the traditional comparative method, looking first at the morphologies, but my work is still in an inceptive stage. So far, I am focusing on the westernmost two families of Mitian: Indo-European and Uralic, under the working hypothesis that these form a node within Mitian - which may be wrong.

So the discussion is now open. Have fun!
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A

User avatar
Salmoneus
Sanno
Sanno
Posts: 3197
Joined: Thu Jan 15, 2004 5:00 pm
Location: One of the dark places of the world

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Salmoneus »

Well, not to be frightfully old-fashioned and boring about things, but perhaps a starting place might be: is there a shred of evidence for the existence of "Mitian", let alone "Eurasiatic" or "Nostratic"?
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]

But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by mèþru »

If there aren't enough identifiable cognates (not borrowings!) for Altaic to remain a widely accepted hypothesis, then I don't see how Mitian can be proved. And I think that Indo-Uralic has the same problem. I'm not saying that there is no relationship, I'm just questioning whether it passes the test of falsifiability.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

Šọ̈́gala
Lebom
Lebom
Posts: 172
Joined: Sun Oct 27, 2013 10:58 pm

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Šọ̈́gala »

How does Aquan fit in?

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: Nostratic, Eurasiatic, Mitian, ...

Post by KathTheDragon »

Salmoneus wrote:Well, not to be frightfully old-fashioned and boring about things, but perhaps a starting place might be: is there a shred of evidence for the existence of "Mitian", let alone "Eurasiatic" or "Nostratic"?
Afaict "Mitian" is based on the MT isogloss... which is probably about as skimpy as it gets.

User avatar
WeepingElf
Smeric
Smeric
Posts: 1630
Joined: Wed Mar 08, 2006 5:00 pm
Location: Braunschweig, Germany
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by WeepingElf »

KathTheDragon wrote:
Salmoneus wrote:Well, not to be frightfully old-fashioned and boring about things, but perhaps a starting place might be: is there a shred of evidence for the existence of "Mitian", let alone "Eurasiatic" or "Nostratic"?
Afaict "Mitian" is based on the MT isogloss... which is probably about as skimpy as it gets.
Yes, it is that isogloss (the pronoun roots *mi 'I' and *ti 'thou'), which is what it is named for. That's of course very skimpy evidence, but if eight language families occupying a (large but) contiguous area of the world, which also has all the characteristics of a spread zone as Johanna Nichols defined it, share such a pair of pronouns, descent from a common ancestor seems like the most plausible explanation. Of course, we cannot rule out borrowing, but 1st and 2nd person pronouns are not what you expect to behave as Wanderwörter at all.

There are of course other bits of evidence. Some of the correspondence sets in Dolgopolsky's and Bomhard's Nostratic etymological dictionaries may actually be valid (at least some must be spurious, though, because they posit different sound correspondences and are thus incompatible with each other!); the phonologies and morphologies of Uralic and Eskimo-Aleut look more similar than one would expect from languages spoken so far away from each other, and Indo-European seems to share in at least some parts of that; and some other things of that kind. Maybe this just shows how common chance resemblances are, but I still consider it likely that at least some of these seeming resemblances mean something.

Then, languages such as Proto-Indo-European and Proto-Uralic did not arise ex nihilo (though I once found an article by a creationist who claimed that these languages arose from the Confusion of Tongues at Babel, and therefore any deeper comparative work was meaningless), but evolved from earlier languages, and there are almost certainly relationships which are simply too deep to have been uncovered yet.

Yet, this is certainly a tough nut to crack, even a probably comparably close relationship as Indo-Uralic. Juho Pystynen (Tropylium) has written a roadmap article discussing the issues at hand. The bottom line: This is not something an enthusiastic amateur with a few handbooks and etymological dictionaries could fix up in a few months! This is a research project which could support several academic linguists' entire careers.

Accordingly, I don't seriously expect to be able to crack this riddle! All I can do is to think about it and contribute some ideas to discussions like this one. (Cases like Michael Ventris show that amateurs can make meaningful contributions to such endeavours if they are knowledgeable, well-networked, respectful of - and building on - the work of others, and willing to abandon hypotheses that fail to yield useful results, though.)

---

As for Aquan, it is unclassified, as there is too little known about it! All we have are a few hundred river names which look as if they were formed from the same set of roots and suffixes, but we don't know their original meanings, such that we cannot meaningfully compare them to known languages; and a few hundred words in Western and Central European IE languages that lack convincing etymologies, which may be from the same language family - or may even be from many different families. We just don't know. It seems more parsimonious to conjecture that all of this is from one family, but we simply can't tell.

My personal hypothesis is that we are dealing with a language family related to Indo-European here bacause some Celtic and Germanic words look like IE words but with the phonology, semantics or both somewhat "off", which looks as if they were borrowed from a related language, and the phonology of the Aquan languages seems to resemble the one I reconstruct for PIE1 (the pre-ablaut stage of PIE). This would mean that the languages are Mitian or whatever IE is a branch of; but this hypothesis may be completely wrong! Other scholars have proposed completely different relationships; the most prominent of these is Theo Vennemann who considers this substratum to be "Vasconic", i.e. a family of which Basque is the last survivor.

What I no longer hold is that the language of the Linearbandkeramik (LBK) culture was related to Indo-European. LBK and Yamnaya (the best candidate for the PIE-speaking community) are not particularly close, neither archaeologically nor genetically. I now fancy that the Aquan languages descend from the first wave of "Kurgan expansions" around 4500 BC, while Anatolian came from the second wave around 3500 BC and the non-Anatolian IE languages from the third wave around 3000 BC.

Finally, a clear line must be drawn between Aquan, the hypothetical natlang family, and Hesperic, my conlang family based on the hypothesis laid out above; Hesperic is designed to be a cognate entity to Indo-European and a kind of "missing link" between IE and Uralic, but that's merely a flight of fancy, and may have nothing to do with what was actually spoken in pre-IE Central Europe!
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A

User avatar
Salmoneus
Sanno
Sanno
Posts: 3197
Joined: Thu Jan 15, 2004 5:00 pm
Location: One of the dark places of the world

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Salmoneus »

WeepingElf wrote:
KathTheDragon wrote:
Salmoneus wrote:Well, not to be frightfully old-fashioned and boring about things, but perhaps a starting place might be: is there a shred of evidence for the existence of "Mitian", let alone "Eurasiatic" or "Nostratic"?
Afaict "Mitian" is based on the MT isogloss... which is probably about as skimpy as it gets.
Yes, it is that isogloss (the pronoun roots *mi 'I' and *ti 'thou'), which is what it is named for. That's of course very skimpy evidence, but if eight language families occupying a (large but) contiguous area of the world, which also has all the characteristics of a spread zone as Johanna Nichols defined it, share such a pair of pronouns, descent from a common ancestor seems like the most plausible explanation. Of course, we cannot rule out borrowing, but 1st and 2nd person pronouns are not what you expect to behave as Wanderwörter at all.
Well, for one thing, the M/T mass comparison doesn't even seem that convincing on the face of it. Take Eskimo-Aleut, for instance - where are the m/t pronouns? WALS lists Yupik as not having them. In Inuit, apparently, the first person subject marker is -unga with an indefinite object, and -agit and -ara with 2nd and 3rd person definite objects respectively. Oh, I see, apparently EA counts because in Sirenik - and no other EA language - the 1s pronoun has /m/ in it (oh, and it's in SOME forms of suffixed agreement in Aleut). [And is Sirenik the EA language most likely to be subject to siberian sprachbund effects? yes!]. And the 2nd person forms often have /l/ in them which is KIND of like T....

Next up is CK. Where are the m/t pronouns? Well, there M and T in Chukchi, if you don't mind them being the final consonants in /G@m/ and /G@t/. Some dialects of Itelmen have plain T- without that "prefix", but those dialects don't have any M. Several languages have M and T in the plural. And sure, maybe the ancestral pronoun only shows up in the plural, or only shows up with a "prefix" (read: inconvenient phonemes) in front of it. But expanding the search parameters in this way also make it much more likely for chance resemblance to play a role. Particularly when you define 'M' and 'T' pretty broadly! And is, say, W. Itelmen /k@zza/ really more T-ish than, say, Basque "zu"?

In Altaic, btw, the "M" is often /b/, and it's apparently widely believed that there must have been borrowing of pronouns between the three altaic families. And even so, it's only Mongolic that shows evidence of T - in Turkic and Tungusic it's a fricative (which, OK, could be derived from the plosive, but... again, shifting the parameters!).

Regarding the non-borrowability of pronouns: leaving aside the pronouns that have been borrowed, we also see processes of imitation, in which a pronoun in one language develops from within the language in a way that mirrors the pronoun of a neighbouring language. An example of this is the Finnish third person pronoun "hän", suspiciously similar to neighbouring Swedish "han", and not present in other Uralic languages, but not believed to have been directly borrowed.
the phonologies and morphologies of Uralic and Eskimo-Aleut look more similar than one would expect from languages spoken so far away from each other
Except that both Uralic and Eskimo-Aleut can be assumed to have originated in eastern Siberia, which puts them pretty close together. PIE, on the other hand, was thousands of miles away at the other end of the continent at the time.
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]

But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by mèþru »

I thought Proto-Uralic is native to Europe? There's evidence for both really. And Pre-Proto-Uralic could have migrated west towards the homeland of Proto-Uralic.

Also, with such broad terms, Afro-asiatic seems much more likely as a member than Siberian languages.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Soap
Smeric
Smeric
Posts: 1228
Joined: Sun Feb 16, 2003 2:57 pm
Location: Scattered disc
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Soap »

Salmoneus wrote:Well, for one thing, the M/T mass comparison doesn't even seem that convincing on the face of it. Take Eskimo-Aleut, for instance - where are the m/t pronouns? WALS lists Yupik as not having them. In Inuit, apparently, the first person subject marker is -unga with an indefinite object, and -agit and -ara with 2nd and 3rd person definite objects respectively.
I think the assumption is that the m/t/Ø paradigm in Eskimo-Aleut is reflected in the object markers, not the subject markers, whether because of sound change or perhaps due to their having been of an entirely different origin. They seem to also appear in the intranstive conjugation however, which is, in the singular, -unga/-utit/-uq for 1ps, 2ps, 3ps with sometimes an additional consonant before the /u/. I agree it's less than convincing, but I could easily imagine an older setup where the 1ps was *-umqa, and an even older system, perhaps, where the endings were monosyllabic -um/-ut/-u, the later extended suffixes having come from some type of agreement morpheme.

I think the part of the 1ps pronoun that means "me" is the -nga suffix, not the (variable) stem.
Sunàqʷa the Sea Lamprey says:
Image

User avatar
Salmoneus
Sanno
Sanno
Posts: 3197
Joined: Thu Jan 15, 2004 5:00 pm
Location: One of the dark places of the world

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Salmoneus »

mèþru wrote:I thought Proto-Uralic is native to Europe? There's evidence for both really. And Pre-Proto-Uralic could have migrated west towards the homeland of Proto-Uralic.
Proto-Uralic per se is believed to have been located on one or other side of the Urals, yes.

However, genetically we know that the overwhelming majority of Proto-Uralic speakers came from east asia, and did so at a time that would have been putting them in the proto-uralic homeland around the time of proto-uralic. This gives us three possible hypotheses:

1) The Pre-Uralic People migrated from the east, speaking Pre-Uralic Language, which developed into Proto-Uralic. They came to the Urheimat and divided into the Uralic groups we know today. In the Urheimat (or before), they encountered IE speakers and borrowed some words from them, particularly relating to technology and environmental terms that were new to them.

2) The Pre-Uralic People migrated from the east, speaking An Unknown Language. They arrived in the Urheimat, and immediately adopted a Macro-European language en masse and without exception, despite the fact that they did not have any noticeable influx of Macro-European speakers into their gene pool, and despite the fact that they, so far as we know, would have been living in a different biome with a different culture and were not culturally dominated by their neighbours. The Macro-European language they adopted has, meanwhile, left no other trace of its existence.

3) The Pre-Uralic People migrated from the east, speaking a Macro-European language all along, because PIE had a close relative in Siberia, and the PUPs had adopted that en masse back in Siberia, and then just coincidentally migrated back to the PIE homeland. Additional difficulties in this scenario come from the fact that the most likely genetic migration route is northerly, while the known Macro-European people in that area would have been very southerly, and from timing problems: genetically the population would have to have been already in place in the west only shortly after the MEs had migrated east. It seems difficult to get these two migrating groups to meet up, let alone for one of them to completely adopt the language (but no other cultural similarities) of the other.

Now, fairness demands that we admit that all three scenarios are theoretically possible. But to me, 1 seems substantially more plausible and parsimonious than either 2 or 3. And so long as the evidence against it is that two neighbouring languages seemingly shared two single-phoneme morphemes, I'm going to find it hard to be convinced. Particularly when the "Mitian" hypothesis actively undermines Indo-Uralic - because if we do accept that the same morphemes have been hanging around across the whole of the step for tens of thousands of years in family after family, then the similarity between IE and Uralic ceases to be indicative of a direct connexion - if the same thing is in common with Chukotkan, then Uralic need be no more 'western' in origin than Chukotkan. [indeed, apparently eurasiatic enthusiasts rather like the fact that Chukotkan directly "retains" the IE "egom" pronoun too].
Also, with such broad terms, Afro-asiatic seems much more likely as a member than Siberian languages.
Afroasiatic doesn't have M/T, does it? And comes from Africa, not central asia. At least there are actual links between Ukraine and Siberia - the links between Ukraine and Sudan are rather more slender! [if we don't mention the obvious genetically oddity that's always best not mentioned lest our heads explode]
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]

But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!

User avatar
Tropylium
Avisaru
Avisaru
Posts: 512
Joined: Sun Aug 07, 2005 1:13 pm
Location: Halfway to Hyperborea

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Tropylium »

Salmoneus wrote:In Altaic, btw, the "M" is often /b/, and it's apparently widely believed that there must have been borrowing of pronouns between the three altaic families.
Which even means they might plausibly be a part of an original cluster of "W-pronouns" instead, as in English we. In Indo-European, of course, this is 1PP and not 1PS; some Nostraticists connect this to the 1PS pronouns in Chadic and Egyptian. This sounds hella skimpier than the evidence for M-pronouns, though.
Salmoneus wrote:Regarding the non-borrowability of pronouns: leaving aside the pronouns that have been borrowed, we also see processes of imitation, in which a pronoun in one language develops from within the language in a way that mirrors the pronoun of a neighbouring language. An example of this is the Finnish third person pronoun "hän", suspiciously similar to neighbouring Swedish "han", and not present in other Uralic languages, but not believed to have been directly borrowed.
Wrong. Finnish hän has cognates in all Finnic languages (from Veps hän to Livonian eņtš 'own'), and is normally also connected with 3rd person pronouns all over the rest of Uralic, such as Northern Sami son or Hungarian őn. Did you happen to have other examples of "imitation of pronouns" in mind — or better yet, some references about this? I don't think I've heard of the concept before.
Salmoneus wrote:However, genetically we know that the overwhelming majority of Proto-Uralic speakers came from east asia,
Uh, no we don't. We do not know the precise genetics of whoever spoke PU at all (there is more than one archeological culture as a candidate, and I don't think we have corresponding paleo-DNA from all of them). What are you going on about?

The "pre-Uralic must have come from east" meme — a speculative hypothesis which all too often gets treated as fact — is actually based on typology. Uralic has typological ties with the Altaic families, all of which have come from somewhere around the east (though I am not very sure about allegedly locating even Proto-Turkic in Manchuria). So if this typology goes back to the Proto-Altaic period or therearound, then also pre-Uralic would have to have been somewhere in the neighborhood.

This hangs on the assumption, though, that the modern "Ural-Altaic typology" is old enough in the first place. I am skeptical. For an example: it's been recently shown that e.g. Proto-Mongolic, Proto-Tungusic and Proto-Korean should be reconstructed with pharyngeal harmony (*u̘ ~ *ʊ̙ etc.), as still in Khalkha, instead of palatal harmony as in Uralic (*y ~ *u etc.). Turkic has also been proposed to have shifted from pharyngeal to palatal harmony recently; there are lines of evidence such as "front vowels" triggering initial voicing in Oghuz (= Turkic-Azeri-Turkmen), which makes zero phonetic sense, but which starts working quite well with +ATR vowels triggering voicing instead. The same also happens to be attested from Armenian right nearby.
— So, if at least one of the alleged "Ural-Altaic" typological parallels is demonstrably secondary, how many others might be as well? We know that "Altaic" has at any rate behaved as a language area for a couple thousand years. Once we reach deeper to something like 5000 years, the typological profile could have been relatively different really.

Also much of the original typological evidence is of very little value in the first place, since they're common-as-dirt features such as SOV word order, initial stress, lack of initial consonant clusters, or lack of ablaut. Many of these were originally defined in contrast to typologically weirder families like Semitic and Indo-European, but by now we know that it's these that are the weird ones, not "Ural-Altaic". Maybe you can put together an argument that there are enough of these that we can still statistically speak of a typological cluster, but any such group would probably also end up including at least Yukaghir and Eskimo-Aleut.
[ˌʔaɪsəˈpʰɻ̊ʷoʊpɪɫ ˈʔæɫkəɦɔɫ]

User avatar
Tropylium
Avisaru
Avisaru
Posts: 512
Joined: Sun Aug 07, 2005 1:13 pm
Location: Halfway to Hyperborea

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Tropylium »

KathTheDragon wrote:Afaict "Mitian" is based on the MT isogloss... which is probably about as skimpy as it gets.
The definition of a "Mitian" subset just on the basis of the 1st and 2nd person pronouns is of course far from airtight. E.g. modern English is not strictly "Mitian" since the 2P general pronoun is you; Hungarian is not strictly "Mitian" since the 1PS pronoun is én.

There are other approaches too, though. For example, taking a few of the best-attested demonstrative and interrogative pronouns instead, we could set up "Takwean":
– Indo-European *to-, *kʷe-
– Semitic *ta-; *kam 'how much'
– Uralic *tä- (prox.) ~ *to- (dist.), *ke- ~ *ku-
– Yukaghir *ta-, *kin 'who' ~ *qa- 'which'
– Turkic *ti-, *kem 'who' ~ *kai 'which'
– Mongolic *te-, *ken
– Tungusic *ta-, *xai
The "core" bundle of IE-U-Y-Tk-Mg-Tg still tags along, but some other families fail to adhere as nicely (and in the absence of other good AA parallels, Semitic is perhaps best considered a chance resemblance).

If we switch to regular lexicon, though, the track record seems to be instantly worse. There are many good-looking pairwise matches, but not many that would cover three separate families. E.g. suppose we pick "Donamean", for 'give' and 'name', both well-established IE–Uralic parallels; then the only other family that fits in as well is Yukaghir (*tant-, ńuː ~ ńiw ~ nim).
Salmoneus wrote:Afroasiatic doesn't have M/T, does it?
It does if you datamine broadly enough: Bomhard alleges here to first person M-pronouns in Chadic, first person M-suffixes in Highland East Cushitic, and to widespread suffixal T-elements in second person pronouns.

But of course, we'd like to know how much of this system can be reconstructed for Proto-Afroasiatic itself, and if the functions of these elements line up correctly too. Bomhard claims also half a dozen other markers for 1st person (in Afrasian: Egyptian-Semitic-Berber-Cushitic suffixal #-ka, Chadic-Semitic-Ongota pronoun stem #na, Egyptian-Semitic-Berber-Cushitic #ʔija, and the aforementioned #wa) and 2nd person (Omotic #ni). It should be clear that not all of these can have coexisted with equal status in the proto-language.
[ˌʔaɪsəˈpʰɻ̊ʷoʊpɪɫ ˈʔæɫkəɦɔɫ]

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by mèþru »

But a few pronouns, whether personal or not (or even both), can not be the sole comparative evidence. You must have a massive corpus of linked cognates.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Salmoneus
Sanno
Sanno
Posts: 3197
Joined: Thu Jan 15, 2004 5:00 pm
Location: One of the dark places of the world

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Salmoneus »

Tropylium wrote:
Salmoneus wrote:Regarding the non-borrowability of pronouns: leaving aside the pronouns that have been borrowed, we also see processes of imitation, in which a pronoun in one language develops from within the language in a way that mirrors the pronoun of a neighbouring language. An example of this is the Finnish third person pronoun "hän", suspiciously similar to neighbouring Swedish "han", and not present in other Uralic languages, but not believed to have been directly borrowed.
Wrong. Finnish hän has cognates in all Finnic languages (from Veps hän to Livonian eņtš 'own'), and is normally also connected with 3rd person pronouns all over the rest of Uralic, such as Northern Sami son or Hungarian őn. Did you happen to have other examples of "imitation of pronouns" in mind — or better yet, some references about this? I don't think I've heard of the concept before.
Of pronouns specifically, no. But the general idea of areal affects, and of imitation, seems well-established. Off the top of my head, one common example given is the simultaneous development of "haben" perfects in Germanic and "habere" perfects in Romance, despite the fact that the two verbs have no genetic relationship. People like to imitate one another...

Regarding the specific pronoun: sorry, when I said "not present", I meant "not present in that role". I know that the morpheme is present in related languages in related roles - that's what I meant by "devolops from within the language" - a word shifts in its meaning in a way that brings the language closer to its neighbours. Iirc similar processes have been suggested for some anomolous developments in English that seem to have been influenced by Norse without necessarily direct borrowing.
Salmoneus wrote:However, genetically we know that the overwhelming majority of Proto-Uralic speakers came from east asia,
Uh, no we don't. We do not know the precise genetics of whoever spoke PU at all (there is more than one archeological culture as a candidate, and I don't think we have corresponding paleo-DNA from all of them). What are you going on about?
OK, 'know' is slightly too strong. However, we do know that most modern Uralic speakers, with the exception of Hungarians (who look like their neighbours) (and also the Estonians have had a fair bit of interbreeding with their neighbours too, though not to Hungarian levels), have very strong genetic similarities to one another, and that their genes are believed to have migrated from northeast asia. By looking at rates of mutation in different areas and influence from and into other groups, we can also get a firm theory that those genes arrived in the area of Proto-Uralic thousands of years before Proto-Uralic in the strict sense (the last common ancestor) is believed to have been spoken. It is true that, hypothetically, it is possible that Proto-Uralic speakers were not representative of this group that is known to have been in the region of Proto-Uralic at the time when Proto-Uralic was spoken, and that each Uralic group has independently suffered almost complete gene replacement by the same, otherwise unevidenced people occupying similar areas at similar times. That is possible. But it's really, really unlikely. So... we "can be very confident that" proto-uralic speakers came from the east (whether or not they were speaking the language at the time).

What's more, the pattern in the extent to which groups don't look genetically Uralic also militates against this more improbable solution. In essence, the further east the Uralic group, the more purely 'Uralic' they are genetically, while groups further west are more genetically mixed, and specifically they're mixed with typically European populations - the Nganasans work out to be pretty much proxies for uralic ancestry in all the other groups. This makes sense following the conventional explanation, in which Nganasans actually are Uralic and western Uralic groups interbred with locals; but if we assume that all this "Uralic" is actually a later group who replaced all the genes in modern groups independently without replacing the language, it would be a bizarre coincidence that we'd get exactly this pattern.

Specifically, Uralic speakers have very high levels of genetic markers otherwise not found in the west, but found voluminously in east asians, including Han Chinese. They DON'T appear to have high levels of association with the western or mid-siberian genetic traits found in the Ural region before the Uralics turned up (which are instead found in groups like the Yamnaya and the modern Ket) - or rather, they seem to have gained those traits only to the extent they interbred with Indo-Europeans (or Ket) later on.
The "pre-Uralic must have come from east" meme — a speculative hypothesis which all too often gets treated as fact — is actually based on typology.
But this runs around in circles. Because if you want to deny that the typological links between Uralic and Altaic can be old enough to reflect a genuine relationship, or even early areal effects, then you can't turn around and use the same typological links (but less extensive and persuasive!) to claim a particularly close relationship between Uralic and Indo-European, or indeed a looser but vastly older relationship across "Mitian" as a whole!
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]

But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!

User avatar
WeepingElf
Smeric
Smeric
Posts: 1630
Joined: Wed Mar 08, 2006 5:00 pm
Location: Braunschweig, Germany
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by WeepingElf »

Tropylium wrote:
Salmoneus wrote:In Altaic, btw, the "M" is often /b/, and it's apparently widely believed that there must have been borrowing of pronouns between the three altaic families.
Which even means they might plausibly be a part of an original cluster of "W-pronouns" instead, as in English we. In Indo-European, of course, this is 1PP and not 1PS; some Nostraticists connect this to the 1PS pronouns in Chadic and Egyptian. This sounds hella skimpier than the evidence for M-pronouns, though.
AFAIK, the Altaic languages have a *bi ~ *min alternation in their 1st-person pronouns. The */m/ in the oblique form *min may be an assimilation of an earlier */b/ to the */n/, but it is also conceivable that the original phoneme was */m/ which regularly denasalized except when another nasal followed.

Afrasian almost certainly originated in Africa (somewhere around Sudan/South Sudan/Ethiopia, I think); sure, the Nostraticists place it in the Near East, but they of course have an axe to grind there, and actually Semitic is the only branch of Afrasian that has ever left Africa. Furthermore, the morphological structure of Afrasian is so remote from IE or the other Mitian languages that no resemblance can be observed. Together, these two points make any relationship hypothesis connecting the two seem utter random. While there are a few IE words that seem to resemble Semitic words, these are probably just Neolithic Wanderwörter that spread from the lost language of the first farmers both north into IE and south into Semitic together with the new economy.
Salmoneus wrote:Regarding the non-borrowability of pronouns: leaving aside the pronouns that have been borrowed, we also see processes of imitation, in which a pronoun in one language develops from within the language in a way that mirrors the pronoun of a neighbouring language. An example of this is the Finnish third person pronoun "hän", suspiciously similar to neighbouring Swedish "han", and not present in other Uralic languages, but not believed to have been directly borrowed.
Wrong. Finnish hän has cognates in all Finnic languages (from Veps hän to Livonian eņtš 'own'), and is normally also connected with 3rd person pronouns all over the rest of Uralic, such as Northern Sami son or Hungarian őn. Did you happen to have other examples of "imitation of pronouns" in mind — or better yet, some references about this? I don't think I've heard of the concept before.
Pronoun borrowings are not unknown, but AFAIK rare, especially between unrelated languages, and positing 1st and 2nd person pronouns as Wanderwörter in a continent-sized area is just ludicrous. And chance resemblance, well, for that we are dealing with too many families. Alas, nobody with a sane mind would build such a vast family proposal on just two pronouns, so research into other meaningful resemblances in morphology and lexicon is required here. The pronouns are just the most visible hint that there may be something going on.
Salmoneus wrote:However, genetically we know that the overwhelming majority of Proto-Uralic speakers came from east asia,
Uh, no we don't. We do not know the precise genetics of whoever spoke PU at all (there is more than one archeological culture as a candidate, and I don't think we have corresponding paleo-DNA from all of them). What are you going on about?

The "pre-Uralic must have come from east" meme — a speculative hypothesis which all too often gets treated as fact — is actually based on typology. Uralic has typological ties with the Altaic families, all of which have come from somewhere around the east (though I am not very sure about allegedly locating even Proto-Turkic in Manchuria). So if this typology goes back to the Proto-Altaic period or therearound, then also pre-Uralic would have to have been somewhere in the neighborhood.
Indeed, typological similarities between neighbouring language families mean nothing with regard to their relationships - especially in an area like northern Eurasia where people have been moving around at a large scale during all of the area's known history, and the linguistic map shows a complicated patchwork of several families.

In my opinion, if Proto-Mitian existed, the homeland would most likely have been somewhere in Central Asia at the end of the last ice age. The first split may have been into "Euro-Siberian" (IE, Uralic, Yukaghir, Chukotko-Kamchatkan and Eskimo-Aleut) and Altaic, but that's highly uncertain. I see no reason to assume that Proto-Uralic arose further east than that. And if both IE and Uralic have entered Europe from Central Asia, the question arises whether these were one wave of immigration (or diffusion) or two. Why can't there have been a Proto-Indo-Uralic, perhaps north of the Caspian Sea, about 8,000 to 10,000 years ago? I can think of at least one genetic marker in support of the Indo-Uralic hypothesis, and that is lactose tolerance (not universal in either speaker community, but quite high in both).

Also, research into a language relationship is not really invalidated by the initial omission of further related languages. At most the later recognition of such languages may ask for a renaming of the entity and a refinement of the reconstruction. If we find meaningful resemblances between IE and Uralic, there is always the possibility that further languages (such as Turkish, Yukaghir or Eskimo-Aleut) share in. Then our results would no longer be "Indo-Uralic", but still valid to whichever name we may assign to the macrofamily. Rasmus Rask, one of the founding fathers of Indo-European comparative linguistics, did not consider the Indo-Iranian languages, which later turned out so valuable to the discipline, in his Undersøgelse, yet he diagnosed the relationship between Germanic on one hand and Greek and Latin on the other hand in basically correct ways.

So even if the Indo-Uralic node may turn out to be invalid (as did Rask's "Thracian" node consisting of Greek and Latin), that doesn't mean that we can't start our endeavours into Northern Eurasian long-range comparative linguistics with this nice pair of families, which have the advantage of being well-studied, and just two families make for an overseeable endeavour where useful results are more achievable than when someone tries to compare half a dozen (or a full dozen) language families, including poorly known ones, in one course.
This hangs on the assumption, though, that the modern "Ural-Altaic typology" is old enough in the first place. I am skeptical. For an example: it's been recently shown that e.g. Proto-Mongolic, Proto-Tungusic and Proto-Korean should be reconstructed with pharyngeal harmony (*u̘ ~ *ʊ̙ etc.), as still in Khalkha, instead of palatal harmony as in Uralic (*y ~ *u etc.). Turkic has also been proposed to have shifted from pharyngeal to palatal harmony recently; there are lines of evidence such as "front vowels" triggering initial voicing in Oghuz (= Turkic-Azeri-Turkmen), which makes zero phonetic sense, but which starts working quite well with +ATR vowels triggering voicing instead. The same also happens to be attested from Armenian right nearby.
Things like vowel harmony can easily spread by diffusion between different families, and there is no need to reconstruct anything like that for Proto-Mitian. Also, the vowel harmony reconstructed for Proto-Uralic is very different from that found in Turkic, let alone Mongolic or Tungusic. Proto-Uralic had nothing like the famous Turkic vowel cube; it had a full vowel inventory in the first syllable, and (probably) a highly reduced vocalism consisting of just a higher and a lower vowel in non-first syllables, each with a front and a back allophone depending on the quality of the first syllable vowel.
— So, if at least one of the alleged "Ural-Altaic" typological parallels is demonstrably secondary, how many others might be as well? We know that "Altaic" has at any rate behaved as a language area for a couple thousand years. Once we reach deeper to something like 5000 years, the typological profile could have been relatively different really.
Indeed! Typological profiles of languages may change much over the course of their history. This can happen at a breathtaking pace - consider how the Insular Celtic languages changed from a conservative Indo-European profile to something the early Indo-Europeanists had trouble recognizing as IE at all within about (at most) 1,000 years!
Also much of the original typological evidence is of very little value in the first place, since they're common-as-dirt features such as SOV word order, initial stress, lack of initial consonant clusters, or lack of ablaut. Many of these were originally defined in contrast to typologically weirder families like Semitic and Indo-European, but by now we know that it's these that are the weird ones, not "Ural-Altaic". Maybe you can put together an argument that there are enough of these that we can still statistically speak of a typological cluster, but any such group would probably also end up including at least Yukaghir and Eskimo-Aleut.
Yes, these features are mostly common as dirt. SOV is the most common of the six possible basic word orders in the languages of the world. Initial stress is perhaps the most common stress pattern. Lack of initial consonant clusters is quite common. Many languages lack any kind of ablaut. Such features are found, for instance, in Dravidian, which seems to have nothing to do with Mitian (though the Nostraticists include it, but this inclusion opens up a keg of problems). No scholar worth his stripes believes in Ural-Altaic anymore.

As you mention Yukaghir and Eskimo-Aleut: I am still unsure about Uralic-Yukaghir; the lexical resemblances may be due to Samoyedic loanwords, and the morphologies look very different. Eskimo-Aleut shows a rather close structural resemblance to Uralic, but that may be merely due to both families being the most conservative members of the MItian bunch, and need not speak for a "Uralic-Eskimo" node which would be hard to justify giving the geographical separation.
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A

User avatar
WeepingElf
Smeric
Smeric
Posts: 1630
Joined: Wed Mar 08, 2006 5:00 pm
Location: Braunschweig, Germany
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by WeepingElf »

Tropylium wrote:
KathTheDragon wrote:Afaict "Mitian" is based on the MT isogloss... which is probably about as skimpy as it gets.
The definition of a "Mitian" subset just on the basis of the 1st and 2nd person pronouns is of course far from airtight. E.g. modern English is not strictly "Mitian" since the 2P general pronoun is you; Hungarian is not strictly "Mitian" since the 1PS pronoun is én.

There are other approaches too, though. For example, taking a few of the best-attested demonstrative and interrogative pronouns instead, we could set up "Takwean":
– Indo-European *to-, *kʷe-
– Semitic *ta-; *kam 'how much'
– Uralic *tä- (prox.) ~ *to- (dist.), *ke- ~ *ku-
– Yukaghir *ta-, *kin 'who' ~ *qa- 'which'
– Turkic *ti-, *kem 'who' ~ *kai 'which'
– Mongolic *te-, *ken
– Tungusic *ta-, *xai
The "core" bundle of IE-U-Y-Tk-Mg-Tg still tags along, but some other families fail to adhere as nicely (and in the absence of other good AA parallels, Semitic is perhaps best considered a chance resemblance).
These two pronouns also are of course also insufficient to build a family on, but the fact that "Takwean" seems to encompass most of the same languages as "Mitian" is another drop that contributes to filling the barrel. So we have four pronouns, which still isn't much - but better than two, of course!
If we switch to regular lexicon, though, the track record seems to be instantly worse. There are many good-looking pairwise matches, but not many that would cover three separate families. E.g. suppose we pick "Donamean", for 'give' and 'name', both well-established IE–Uralic parallels; then the only other family that fits in as well is Yukaghir (*tant-, ńuː ~ ńiw ~ nim).
The 'name' parallel between IE and Uralic is IMHO problematic. PIE *h1neh3mn 'name' clearly is not a single morpheme! We can extract the well-known suffix *-m(e)n 'way, means' from this, so we have the root *h1neh3-, which may mean 'to call' or something like that. We would have to assume that the Uralic word, if it is cognate, is composed of two cognate morphemes. If you ask me, it looks more like a chance resemblance.
Salmoneus wrote:Afroasiatic doesn't have M/T, does it?
It does if you datamine broadly enough: Bomhard alleges here to first person M-pronouns in Chadic, first person M-suffixes in Highland East Cushitic, and to widespread suffixal T-elements in second person pronouns.

But of course, we'd like to know how much of this system can be reconstructed for Proto-Afroasiatic itself, and if the functions of these elements line up correctly too. Bomhard claims also half a dozen other markers for 1st person (in Afrasian: Egyptian-Semitic-Berber-Cushitic suffixal #-ka, Chadic-Semitic-Ongota pronoun stem #na, Egyptian-Semitic-Berber-Cushitic #ʔija, and the aforementioned #wa) and 2nd person (Omotic #ni). It should be clear that not all of these can have coexisted with equal status in the proto-language.
Such a plethora of morphemes is not very convincing. No language has that many different 1st person markers! If these languages are indeed all related to each other, at least some of them must have innovated. The question is which ones.

There is no solid reconstruction of Proto-Afrasian yet. There are two major attempts on the market, one by Orel and Stolbova, and one by Ehret; like Dolgopolsky's and Bomhard's Nostratic reconstrutions, they use partly different sound correspondences and cannot both be right. I have once read a review of them which attests similar problems as usually found in macro-comparative work.

The reason why Afrasian is an accepted family is not the presence of a good reconstruction, but it seems to be lower standards in African language classification, where less evidence is apparently necessary to establish a classification than in Eurasia. The African language classification algorithm seems to run as this:

1. If it shows signs of being related to Semitic, it goes to Afrasian.
2. If it shows signs of being related to Bantu, it goes to Niger-Congo.
3. If it doesn't seem to be related to either, and has clicks, it goes to Khoisan.
4. If it doesn't seem to be related to either, and doesn't have clicks, it goes to Nilo-Saharan.

Yet, morphological resemblances between the six Afrasian families look quite nice, and it is at least very plausible that these are indeed related to each other, even if that relationship is almost as deep as Mitian.
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A

User avatar
Tropylium
Avisaru
Avisaru
Posts: 512
Joined: Sun Aug 07, 2005 1:13 pm
Location: Halfway to Hyperborea

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Tropylium »

Salmoneus wrote:However, we do know that most modern Uralic speakers, with the exception of Hungarians (who look like their neighbours) (and also the Estonians have had a fair bit of interbreeding with their neighbours too, though not to Hungarian levels), have very strong genetic similarities to one another, and that their genes are believed to have migrated from northeast asia.
Again, not really. Even if we discard the westernmost groups from consideration, the Volga area peoples have more in common with their Turkic neighbors than with Siberian peoples; the Siberian peoples in turn have more in common with their Siberian neighbors.

Maybe by "very strong" you just mean "discernible" (e.g. variants of Y-chromosomal haplogroup N1C).

It is also not the Estonians who have "interbred with their neighbors". The current understanding is that there was a wholesale language shift from Baltic to Finnic, and instead the later expansion of Finnish, Karelian etc. involved interbreeding with eastern-type hunter-gatherer populations (these had already first shifted to speaking Sami at this point, but their ancestors had arrived in the Baltic region much earlier than Uralic did)
Salmoneus wrote:In essence, the further east the Uralic group, the more purely 'Uralic' they are genetically, while groups further west are more genetically mixed, and specifically they're mixed with typically European populations - the Nganasans work out to be pretty much proxies for uralic ancestry in all the other groups.
Sorry, but this is nonsense. The Nganasans have extensive Paleosiberian substrate ancestry, and this also shows up in the language, which has one of the lowest retention rates of Uralic or even Samoyedic native vocabulary.

There's a so-called "Nganasan component" in many autosomal genomic analyses, which turns up in many Uralic-speaking groups, and is at its highest among the Nganasans; but it does not mean that the Nganasans could be treated as unmixed Proto-Uralians. It does not even mean that the component is associated with the expansion of Uralic at all! The same component turns up all around Asia and even in America, and it's better interpreted as degree of interbreeding with Paleosiberian-type populations. Point is that the Nganasans are the westernmost population that has essentially zero western Eurasian ancestry, and therefore they're often used as a "calibration group" for genetic analyses of various other populations in Eurasia. See e.g. figures 4.2 and 4.4 here for a "Nganasan component" in various East Asian peoples, or figure 2 here for a "Nganasan component" in various Turkic-speaking peoples.
Salmoneus wrote:Because if you want to deny that the typological links between Uralic and Altaic can be old enough to reflect a genuine relationship, or even early areal effects, then you can't turn around and use the same typological links (but less extensive and persuasive!) to claim a particularly close relationship between Uralic and Indo-European, or indeed a looser but vastly older relationship across "Mitian" as a whole!
Yes, of course typology doesn't work for establishing relationships. (And even if it did, it does a very poor job for any attempts to link Indo-European to Uralic.)
[ˌʔaɪsəˈpʰɻ̊ʷoʊpɪɫ ˈʔæɫkəɦɔɫ]

User avatar
gach
Avisaru
Avisaru
Posts: 472
Joined: Mon Feb 17, 2003 11:03 am
Location: displaced from Helsinki

Re: Nostratic, Eurasiatic, Mitian, ...

Post by gach »

Salmoneus wrote:
Tropylium wrote:
Salmoneus wrote:Regarding the non-borrowability of pronouns: leaving aside the pronouns that have been borrowed, we also see processes of imitation, in which a pronoun in one language develops from within the language in a way that mirrors the pronoun of a neighbouring language. An example of this is the Finnish third person pronoun "hän", suspiciously similar to neighbouring Swedish "han", and not present in other Uralic languages, but not believed to have been directly borrowed.
Wrong. Finnish hän has cognates in all Finnic languages (from Veps hän to Livonian eņtš 'own'), and is normally also connected with 3rd person pronouns all over the rest of Uralic, such as Northern Sami son or Hungarian őn. Did you happen to have other examples of "imitation of pronouns" in mind — or better yet, some references about this? I don't think I've heard of the concept before.
Of pronouns specifically, no. But the general idea of areal affects, and of imitation, seems well-established. Off the top of my head, one common example given is the simultaneous development of "haben" perfects in Germanic and "habere" perfects in Romance, despite the fact that the two verbs have no genetic relationship. People like to imitate one another...

Regarding the specific pronoun: sorry, when I said "not present", I meant "not present in that role". I know that the morpheme is present in related languages in related roles - that's what I meant by "devolops from within the language" - a word shifts in its meaning in a way that brings the language closer to its neighbours. Iirc similar processes have been suggested for some anomolous developments in English that seem to have been influenced by Norse without necessarily direct borrowing.
Even the "not present in that role" isn't an accurate statement. Both the already mentioned North Saami son and Hungarian ő but also other cognates like Erzya son and Udmurt so all function in the SG3 pronoun role. The *s > h shift in Finnic is the same weak grade change that's responsible for turning the *s into either h or zero in most reflexes of the old illative suffix *-sen, such as in the Finnish allomorphs -seen in venee-seen (boat-ILL) vs. -hVn in maa-han (ground-ILL) and -Vn in vete-en (water-ILL).

I agree that the idea of non-cognate words converging across language boundaries by phonological imitation is an interesting one but it needs much more substance to back it. It's clearly different from borrowing grammatical constructions or taking loan words.
Tropylium wrote:This hangs on the assumption, though, that the modern "Ural-Altaic typology" is old enough in the first place. I am skeptical. For an example: it's been recently shown that e.g. Proto-Mongolic, Proto-Tungusic and Proto-Korean should be reconstructed with pharyngeal harmony (*u̘ ~ *ʊ̙ etc.), as still in Khalkha, instead of palatal harmony as in Uralic (*y ~ *u etc.). Turkic has also been proposed to have shifted from pharyngeal to palatal harmony recently; there are lines of evidence such as "front vowels" triggering initial voicing in Oghuz (= Turkic-Azeri-Turkmen), which makes zero phonetic sense, but which starts working quite well with +ATR vowels triggering voicing instead. The same also happens to be attested from Armenian right nearby.
Sounds interesting. Can you give the relevant references for weekend reading?

User avatar
WeepingElf
Smeric
Smeric
Posts: 1630
Joined: Wed Mar 08, 2006 5:00 pm
Location: Braunschweig, Germany
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by WeepingElf »

I wish to resume here the discussion of the origin of the PIE stop grades we had on the Great PIE Thread and which led me to start this thread. We don't know yet what the Proto-Indo-Uralic or Proto-Mitian consonant inventory was like, but I think that the one reconstructed by Fortescue for Uralo-Siberian may give an idea - a better one at least IMHO than the large inventories reconstructed by the Nostraticists which are bloated by the inclusion of Afrasian, Kartvelian and Dravidian and underaccounting of conditioned changes.

In such a system, it would seem plausible that the PIE voiceless stops descend from the PIU voiceless stops, and the PIE voiced aspirated stops (which may actually have been voiced fricatives IMHO) descend from voiced fricatives. But what about the PIE voiced unaspirated stops, which according to some scholars like Gamkrelidze, Ivanov and Bomhard may have actually been ejectives? I am not really sold by this "glottalic theory", but it seems as if we are dealing with a highly marked kind of phonemes here, if not glottalized then pharyngealized or affricates or whatever. One could speak of "emphatic" stops as I shall do here.

Before the split of the velar series into three (front, back, labialized) in the changes associated with "Great Vowel Collapse", there would have been just two emphatic stops, *t' and *k' (*p' seems to have been missing). Now if the PIU/Mitian inventory was like Fortescue's Uralo-Siberian one, *t' could have emerged from the affricate *c (this would have to have gone somewhere as PIE doesn't have any affricate). What could *k' come from? Perhaps PIU had a uvular stop *q (though Fortescue doesn't reconstruct one), which would have changed into *k' (in a reverse of a well-known Semitic sound change). Of course, a velar affricate is also an option; like *q, it could have gone to *k or *x in Uralic. Yet, velar affricates are rare in the world's languages.
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: Nostratic, Eurasiatic, Mitian, ...

Post by KathTheDragon »

As an alternative to ejectives, we might consider implosives for PIU, yielding the PIE plain voiced stops, and the PU nasals.

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by mèþru »

WeepingElf wrote:voiced aspirated stops (which may actually have been voiced fricatives
Which then turned into voiced stops, breathy voiced stops and aspirated stops in descendant languages? I think it is much more likely that they were some kind of stop, at least on the phonemic if not phonetic level.

I think, if you are doing a reconstruction of Uralic and Indo-European, then you should use an accepted version of the Proto-Uralic inventory and try doing a comparative analysis of the two proto-languages. Here's some stuff I found on Wikipedia that can serve as a starting point:
Among the sound correspondences which Čop did assert were (1972:162):
Uralic m n l r = Indo-European m n l r.
Uralic j w = Indo-European i̯ u̯.
Uralic sibilants (presumably s š ś) = Indo-European s.
Uralic word-initial voiceless stops (presumably p t č ć k) = Indo-European word-initial voiceless stops (presumably p t ḱ k kʷ), also Indo-European s followed by one of these stops.
Uralic word-initial voiceless stops (presumably p t č ć k) = Indo-European word-initial voiced aspirates (presumably bʰ dʰ ǵʰ gʰ gʷʰ).
Uralic ŋ = Indo-European g and ng.

I am of the opinion that the palatovelars were actually plain and that the plain velars were actually uvular (at least in early PIE). Maybe Uralic participated in the satem areal shift.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Soap
Smeric
Smeric
Posts: 1228
Joined: Sun Feb 16, 2003 2:57 pm
Location: Scattered disc
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by Soap »

WeepingElf wrote: Now if the PIU/Mitian inventory was like Fortescue's Uralo-Siberian one, *t' could have emerged from the affricate *c (this would have to have gone somewhere as PIE doesn't have any affricate). What could *k' come from? Perhaps PIU had a uvular stop *q (though Fortescue doesn't reconstruct one), which would have changed into *k' (in a reverse of a well-known Semitic sound change). Of course, a velar affricate is also an option; like *q, it could have gone to *k or *x in Uralic. Yet, velar affricates are rare in the world's languages.
I consider the ejectives original, and propose a shift of /t' k'/ > /c k/ in Uralic, /t q/ in proto-Eskimo-Aleut, and persistence of /t' k'/ in glottalic PIE.
Sunàqʷa the Sea Lamprey says:
Image

User avatar
gach
Avisaru
Avisaru
Posts: 472
Joined: Mon Feb 17, 2003 11:03 am
Location: displaced from Helsinki

Re: Nostratic, Eurasiatic, Mitian, ...

Post by gach »

If you take the plain voiced stops as the last stop series that PIE gained, why couldn't you explain their distribution simply by the distribution of their originating condition? You could for example easily imagine a Pre-PIE stop system with a plain two-way distinction of [-voiced] and [+voiced] stops where the ancestral [+voiced] series later gained secondary breathiness. If you then have some later split leading into the development of [+voiced, -breathy] stops that contrasted with the older [+voiced, +breathy] series, their distribution could simply reflect the distribution of their originating condition. I don't know how much evidence there is for such development history but the point is that a restricted distribution of a phoneme can reflect its history without the phoneme having to have any particularly special phonetic features.

I'll have to point out that hypothesising with more marked extra phonemes in Pre-Uralic to make unconditional ejectives possible in some branch of Para-Uralic (= PIE?) has some of the same feel than those bloated Nostratic inventories, though certainly in a much more tasteful scale. Just be careful that whatever you suggest remains justified.

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by mèþru »

@Soap
And how would you interpret it if you assume that neither Indo-European nor early Proto-Uralic had ejectives?
@WeepingElf
Are you trying to first prove a genetic relationship of Uralic and Indo-European or to first prove the entire Mitian language family?

Edit: I completely agree with what gach said.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
WeepingElf
Smeric
Smeric
Posts: 1630
Joined: Wed Mar 08, 2006 5:00 pm
Location: Braunschweig, Germany
Contact:

Re: Nostratic, Eurasiatic, Mitian, ...

Post by WeepingElf »

I think Indo-Uralic is a useful first step towards Mitian and should be established first, but I also think Indo-Uralic reconstruction is best done with having the state of affairs in likely further relatives in mind. Such considerations lead to the working hypothesis that the Proto-Uralic phonology is more conservative than the PIE one. Also, the PU agglutinating morphology looks more archaic than the PIE fusional one. See below for more on this.

I mean like, one would come up with an Indo-Uralic correspondence set, e.g. (I don't know yet whether this set is valid or not) PIE *dh : PU *ð. But what was the PIU sound like? Taking a look at languages such as Eskimo-Aleut, or Fortescue's Uralo-Siberian reconstruction attempt, one may guess that the Proto-Indo-Uralic sound may have been *ð. Or if we could show that PIE *d (< earlier *t'?) was corresponding to PU *c (*/ts/), the question is, was it PIU *t' > PU *c, or PIU *c > Pre-PIE *t' > PIE *d? Or yet something else? One could, by looking at other Mitian languages, guess that it was probably more like the second alternative.

After all, when the rest of Mitian was more like PIE, then PIU probably was more like PIE as well, and when it was more like PU, then PIU probably was more like PU, too, in order to keep the necessary changes to a minimum and avoid a "there and back again" change in one of the languages. Right now, we see a phonology somewhat resembling that of Proto-Uralic in Eskimo-Aleut, and - a bit less clearly - in Yukaghir and Chukotko-Kamchatkan (see the correspondence chart here); even the Altaic languages, in such things as having two rather than three grades of stops (though Starostin, Dybo and Mudrak reconstruct three for Proto-Altaic), are a tad closer to the Uralic than the IE type. No Mitian language family resembles IE as closely as EA resembles Uralic, so we can at least guess that Uralic is more archaic than IE.

In fact, classification comes before reconstruction. You can't reconstruct a protolanguage without having good evidence that the languages you are reconstructing the common ancestor of are related. Tropylium has made a blog post on this matter. There are several classifications on the market which are accepted by most relevant scholars yet lack a generally accepted reconstruction. Afrasian is an example; there are several reconstruction attempts already, but no consensus on which of these is correct.
Last edited by WeepingElf on Sat Aug 05, 2017 12:22 pm, edited 2 times in total.
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A

Post Reply