The Great Proto-Indo-European Thread's Sequel

[WeepingElf]

This discussion of revisions of the standard model reminds me of the discussion of revisions of another standard model

[abahot]

To be fair, isn't this entire thread about revisions of the standard model?

[WeepingElf]

Sure, it is. However, as I have written before, I think that the standard model accounts for the non-Anatolian IE languages reasonably well, and so far, most attempts at revising it as the latest common ancestor of those languages have made more problems than sense. But what regards the prehistory of that common ancestor, there are open questions that suggest that it emerged, perhaps after the separation of Anatolian, from something else.

For instance, the *T and *Dh sets of stops do not form a natural class at the exclusion of *D in the standard model but behave like one, which calls for an explanation - probably, they did once form such a class before things changed. Back when I was a glottalist, I felt that it makes no sense to posit the glottalist system for Late PIE but only for some earlier stage that shifted to something like the standard model later. Now I am no longer a glottalist, and think that an aspirated *T set does the job quite nicely, though I think that that set lost its aspiration before the breakup of Late PIE (in most dialects at least; the dialects ancestral to Germanic and Armenian may perhaps have preserved it).

Likewise, it is pretty clear that Late PIE was a nominative-accusative language with the peculiarities that neuter nouns had the same form for the nominative and accusative cases, neuter transitive subjects were avoided, and two stative inflectional categories of the verbs, the perfect and the middle, employed personal endings unrelated to those of the other categories, as in the standard model, but the question is legitimate how these peculiarities came to be, and it IMHO makes sense to assume that some earlier stage was some kind of split-ergative language, maybe an active-stative one.

So modifications of the standard model IMHO have their place in Early PIE or Pre-PIE, but for Late PIE, the standard model fits the facts well.

[hwhatting]

Or (following Adams, Hamp, and Fellner) that Tocharian didn't split off early.

Interesting. What I've seen of the verbal system looks like it has an earlier break-off point than the other non-Anatolian languages, but as I said above, I need to read up much more on Tocharian in order to have a grounded opinion on that.

[hwhatting]

So modifications of the standard model IMHO have their place in Early PIE or Pre-PIE, but for Late PIE, the standard model fits the facts well.

As dicussed above, it fits for the phonology, as long as you're agnostic to actual realisations, but for the morphology, define what "Standard model" means -
if you include e.g. the tripartite system or three genders, it doesn't fit. I'd rather sys that currently, there is no consensus and therefore no "Standard model", at least if you think of something as detailed as the Traditional model.

[WeepingElf]

So modifications of the standard model IMHO have their place in Early PIE or Pre-PIE, but for Late PIE, the standard model fits the facts well.

As dicussed above, it fits for the phonology, as long as you're agnostic to actual realisations, but for the morphology, define what "Standard model" means -
if you include e.g. the tripartite system or three genders, it doesn't fit. I'd rather sys that currently, there is no consensus and therefore no "Standard model", at least if you think of something as detailed as the Traditional model.

By "standard model" I mean what is presented in most handbooks, such as the three-gender system and the tripartite verb aspect system. I know that there are minor differences between different handbooks. And I think it does fit the non-Anatolian languages, even if some of them appear to have lost some of these categories, and only Vedic and Avestan have preserved everything. There may be minor issues with it, but one should not throw out the baby with the bathwater. The Anatolian languages - I don't know how often I have stated that here - are a different matter.

[Ketsuban]

I feel like however well the standard model works for the classical Indo-European languages it should be treated as schematic rather than representative of any historical point in the family. "Classical" Brugmannian Indo-European has its origins in philologists well familiar with Latin and Greek encountering Sanskrit for the first time and falling in love for its rich morphology ("a wonderful structure; more perfect than the Greek, more copious than the Latin, and more exquisitely refined than either"), with Anatolian and Tocharian (beyond surface elements like confirming the broad strokes of the laryngeal hypothesis) assumed to have branched earlier seemingly more because they were discovered later and don't "fit" the standard model.

[hwhatting]

By "standard model" I mean what is presented in most handbooks, such as the three-gender system and the tripartite verb aspect system. I know that there are minor differences between different handbooks. And I think it does fit the non-Anatolian languages, even if some of them appear to have lost some of these categories, and only Vedic and Avestan have preserved everything. There may be minor issues with it, but one should not throw out the baby with the bathwater. The Anatolian languages - I don't know how often I have stated that here - are a different matter.

Well, you talked about Late PIE, which would still be before the break-up, that's why I asked.

[WeepingElf]

By "standard model" I mean what is presented in most handbooks, such as the three-gender system and the tripartite verb aspect system. I know that there are minor differences between different handbooks. And I think it does fit the non-Anatolian languages, even if some of them appear to have lost some of these categories, and only Vedic and Avestan have preserved everything. There may be minor issues with it, but one should not throw out the baby with the bathwater. The Anatolian languages - I don't know how often I have stated that here - are a different matter.

Well, you talked about Late PIE, which would still be before the break-up, that's why I asked.

I apologize for the misunderstanding. I have the habit of using Late PIE for the stage of PIE just before the divergence of the non-Anatolian IE languages, and Early PIE for the stage of PIE just before the Anatolian/non-Anatolian split. But these terms are misleading; I have realized that many scholars use Late PIE for what I call Early PIE. So it seems we were just talking past each other, and I really should abandon my idiosyncratic usage.

[hwhatting]

All good

[Talskubilos]

The academic Indo-Europeanists have very good reasons to reconstruct Proto-Indo-European the way they do. Of course, the current model is not the final word, but it is the result of more than 200 years of work by hundreds of scholars, so they can't all be barking up the wrong tree all the time.

IMHO, the problem is the relative inadequacy of the classical tree model.

[WeepingElf]

I concur with you that the family tree model has its problems. The Indo-European family emerged from an expanding dialect continuum in which innovations spread and formed a complex pattern of intersecting isoglosses, so a kind of wave model describes things better.

[abahot]

On a related note, on the abundance of *m, I wonder if Pre-PIE had the sound change *-n# > *-m#. It would explain why *-m is so abundant in endings and why none end in *-n.

Putting this here for the future: in the course of my readings, I have found that the pre-PIE sequence **-Vn became PIE *-V̄, explaining the lack of word-final *n in Proto-Indo-European. If this is the case, this seems like a very strange phonological gap, that word-final *-m is present but word-final *-n is not. Maybe this explains the fact that word-final *-m is altered so frequently -- to nasalization in Latin (and Sanskrit?) and sometimes Germanic, to *-n in Greek and the rest of Germanic, and so forth.

[Zju]

I thought word final *n became *r in PIE, thus giving rise to heteroclitics? There's some oblique evidence in various 3PL verbal endings, too.

It would be very strange indeed only some nasal consonant (and not all) in coda position to give rise to nasal vowels.

[abahot]

I apologize -- I misread it, and it's only that **-ōn > *-ō, not all vowels.

As far as I know, the fate of word-final **-n is still not settled, though becoming *-r would make sense. What do you mean by the 3PL verbal endings? Like perhaps that stative 3PL *-ēr < **-ēn (-t)?

[WeepingElf]

I thought word final *n became *r in PIE, thus giving rise to heteroclitics? There's some oblique evidence in various 3PL verbal endings, too.

It would be very strange indeed only some nasal consonant (and not all) in coda position to give rise to nasal vowels.

The idea of *n > *r /_# explaining heteroclitics is something I entertain myself; this denasalization was blocked by another nasal before the *n, hence the *-men stems.

I apologize -- I misread it, and it's only that **-ōn > *-ō, not all vowels.

As far as I know, the fate of word-final **-n is still not settled, though becoming *-r would make sense. What do you mean by the 3PL verbal endings? Like perhaps that stative 3PL *-ēr < **-ēn (-t)?

An interesting idea which I haven't heard of yet, but it seems to make sense to me.

[WeepingElf]

I am trying to draw up an Indo-European family tree based on linguistic isoglosses and what geneticists and archaeologists have found out about the migrations of the people who probably spoke those languages.

The Yamnaya culture of the Pontic Steppe has been quite securely (not *perfectly* securely, but sufficiently so) as the speakers of Proto-Indo-European. After 3000 BC, these people began to expand in all directions, resulting in a handful of daughter cultures:

1. The Catacomb culture in the Pontic Steppe itself.
2. The Corded Ware culture in Central and Eastern Europe north of the Carpathians, from the Rhine to the Volga (west of the Vistula, in a mixed pattern with the Bell Beaker culture).
3. The Afanasievo culture far to the east near the Altai mountains.
4. The Bell Beaker culture in most of Western Europe.
5. The Ezero culture in the eastern part of the Balkan Peninsula.

Who of these spoke what? It seems quite clear that the Corded Ware culture spoke a dialect of PIE ancestral to Germanic, Balto-Slavic and it seems also to Indo-Iranian. Balto-Slavic appears to be closer to Indo-Iranian than to Germanic, but it shares some isoglosses (such as the development *bhy > *m observed in the dative plural) with Germanic, too. Germanic seems to be an outlier of sorts.

The Afanasievo culture probably spoke the ancestor of Tocharian.

Where to put Italo-Celtic and Greco-Armenian is a difficult matter. Some scholars identify Italo-Celtic with the Bell Beaker culture, but there seems to have been an older IE stratum in the area occupied by Bell Beaker which is not the ancestor of Italic and Celtic, as can be seen quite clearly from the fact that this substratum, which the Old European Hydronymy appears to be a trace of, apparently merged PIE *o with *a, which happened neither in Italic nor in Celtic. Also, the time depths of Italic and Celtic appear to be in the 1500-1000 BC range, which suggests that these languages were spread by the expansion of the Urnfield culture out of the Pannonian Basin about 1300 BC. So Italic and Celtic may descend from a language of descendants of the Catacomb culture who moved west, perhaps when Indo-Iranian pushed into the Pontic steppe.

Greco-Armenian seem to descend from the Catacomb culture as well, and thus it looks a bit as if Greco-Armenian forms a branch with Italo-Celtic. To me, at least, Greek looks as if it was closer to Latin than to Sanskrit, though it shares some traits with the latter, most prominently the verb system and the accent system, but both are archaisms and thus not diagnostic.

Anatolian probably came to Anatolia by way of the Ezero culture.

So what language did the Bell Beaker culture speak? Here we meet the most speculative part of the whole thing. IMHO they may have spoken a sister branch of Anatolian which I call "Southwest IE", the language of which traces are retained in the Old European Hydronymy (and forms the basis of my Hesperic conlang project).

So we would have the following scenario:

There was a dialectal division within PIE into a Northern and a Southern dialect prior to the Yamnaya expansions already. The common ancestor of these two dialects may have been spoken as early as 4000 BC. This division seems to correspond to a difference in genetics: Northern IE speakers would have R1a as the most common Y-DNA hapologroup, and Southern IE speakers R1b, though this was a rather gradual difference. (The two haplogroups are, as the nomenclature suggests, the closest relatives of each other, but the time depth is much greater than any reasonable guess of the time depth of the IE family.) The Northern group spread northward and broke up into three major branches: Germanic-Balto-Slavic-Indo-Iranian (Corded Ware), Tocharian (Afanasievo) and Italo-Celtic-Greco-Armenian (Catacomb) while the Southern group broke up into Anatolian (Ezero) and Southwest IE (Bell Beaker).

Of course, this is speculation, and has to be treated with caution as genes, material culture and languages do not always travel together.

[hwhatting]

Who of these spoke what? It seems quite clear that the Corded Ware culture spoke a dialect of PIE ancestral to Germanic, Balto-Slavic and it seems also to Indo-Iranian. Balto-Slavic appears to be closer to Indo-Iranian than to Germanic, but it shares some isoglosses (such as the development *bhy > *m observed in the dative plural) with Germanic, too. Germanic seems to be an outlier of sorts.

I already said that eleswhere, but I think you have that backward. Balto-Slavic is the outlier if you look at the verbal endings system; Slavic then became more similar to Indo-Iranian under Iranian influence. Germanic looks like common Western European and also has a significant number of lexical isoglosses with Italic; it then drifted farther away probably after its speakers moved from Central Europe to Scandinavia. The Balto-Slavic - Germanic isoglosses are minor and normally seen as areal phenomenon.

The Afanasievo culture probably spoke the ancestor of Tocharian.

That seems to be the common opinion, yes.

Where to put Italo-Celtic and Greco-Armenian is a difficult matter. Some scholars identify Italo-Celtic with the Bell Beaker culture, but there seems to have been an older IE stratum in the area occupied by Bell Beaker which is not the ancestor of Italic and Celtic, as can be seen quite clearly from the fact that this substratum, which the Old European Hydronymy appears to be a trace of, apparently merged PIE *o with *a, which happened neither in Italic nor in Celtic. Also, the time depths of Italic and Celtic appear to be in the 1500-1000 BC range, which suggests that these languages were spread by the expansion of the Urnfield culture out of the Pannonian Basin about 1300 BC. So Italic and Celtic may descend from a language of descendants of the Catacomb culture who moved west, perhaps when Indo-Iranian pushed into the Pontic steppe.

To me it looks rather like there was a Central European dialect continuum, of which Italic, Celtic, Germanic were a part, sharing different isoglosses. All shared with Graeco-Aryan the development of the tripartite tense-aspect system, which then was rebuilt / merged in different ways in the three branches.

Greco-Armenian seem to descend from the Catacomb culture as well, and thus it looks a bit as if Greco-Armenian forms a branch with Italo-Celtic. To me, at least, Greek looks as if it was closer to Latin than to Sanskrit, though it shares some traits with the latter, most prominently the verb system and the accent system, but both are archaisms and thus not diagnostic.

This only makes sense if you assume the Classical Verbal system for PIE, which IMHO would be totally misguided. It rather looks like Greek,Indo- Aryan, Armenian (and some languages like Phrygian, of which we have only traces) were the core of a group of languages which developed the "Classical" verbal system, with, among others, the -o /-oi ending set in the medium, the long thematic vowel subjunctive, and the augment. The Western IE languages (Celtic, Italic, Germanic) only partially participated in these developments, and the oldest layer of Balto-Slavic not at all (while, as said above, the Slavic verbal system later moved closer under Iranian influence).
And BTW, Graeco-Armenian is one of those old canards that don't seem to want to die, although Clackson showed convincingly 30 years ago that most mooted Greek-Armenian isoglosses are either not exclusive or shared retentions.

[WeepingElf]

I already said that eleswhere, but I think you have that backward. Balto-Slavic is the outlier if you look at the verbal endings system; Slavic then became more similar to Indo-Iranian under Iranian influence. Germanic looks like common Western European and also has a significant number of lexical isoglosses with Italic; it then drifted farther away probably after its speakers moved from Central Europe to Scandinavia. The Balto-Slavic - Germanic isoglosses are minor and normally seen as areal phenomenon.

You said that before. But what does that "outlier" do right in the centre of the Indo-European world? That makes no sense to me. Also, you are the only person who says that.

And BTW, Graeco-Armenian is one of those old canards that don't seem to want to die, although Clackson showed convincingly 30 years ago that most mooted Greek-Armenian isoglosses are either not exclusive or shared retentions.

Fair. Indeed, Greek and Armenian do not look much alike, but Armenian has undergone many pervasive changes which make it hard to tell where it belongs.

[hwhatting]

You said that before. But what does that "outlier" do right in the centre of the Indo-European world? That makes no sense to me.

Because where languages are now is not where they were necessarily at the beginning? Except if you believe the fringe "Celtic from the West" theory, Celtic was orginally spoken somewhere in Western Central Europe (Hallstatt culture), while it's now spoken on the Western fringe of Europe. Germanic must have been spoken in the vicinity of Italic, also somewhere in Central Europe, before the ancestors of the Germanic speakers migrated North into Scandinavia and the ancestors of the Italic speakers South into Italy, which is supported by a bunch of Germanic-Italic isoglosses. These things are not controversial. The result of that is the Balto-Slavic was originally at the Northern edge of IE, with IE neighbours only to the South (South-West, South-East). Plus some of their original neigbours to the South were probably displaced when the Iranian speakers entered the Pontic Steppe from the East.

Also, you are the only person who says that.

I'm building here on things I learnt from my IE studies teacher, the late Erich Neu. The problem is that most studies of the Balto-Slavic verb assume the traditional model of the verbal system, and try to shoe-horn the Balto-Slavic evidence into it. Partially, because they were done over 50 years ago when the traditional model still reigned supreme. Nowadays, the Baltic verbal system is sometimes used to support reconstructions that deviate from the traditional model, but it's still more usual that the weirdness especially of the ending system is ignored or tried to be explained away.