The Great Proto-Indo-European Thread's Sequel

[WeepingElf]

Hmm, I'm reading stuff like this that implies that most of the similarities between PIE and Proto-Uralic are likely either borrowing or chance, and that the Urheimat of PU is likely as far east as Yakutia, contrary to past views that set it in the Urals. As for Mitian, I'm seeing it suggested that possibly Mitian is a real thing but does not include IE at all.

That guy has his own axe to grind, namely that IE was a branch of Afroasiatic - which is hardly compatible with the Indo-Uralic hypothesis which he thus had to find an argument against. To me, PIE looks more like a sister of Proto-Uralic than like an Afroasiatic language, but things may be deceptive. My guess at the Mitian Urheimat is in southern Siberia, probably somewhere near Lake Baykal, at the end of the last ice age.

As for Indo-Uralic, the findings of this paper are usually interpreted as supporting a Caucasian origin of "Indo-Anatolian", but the genes of course say nothing about the languages the people spoke, even if they can be used to trace prehistoric migrations. My interpretation is that the Caucasian people ancestral to the Yamnaya spoke Pre-Proto-NWC, those that moved into Anatolia spoke a language that evolved into Hattic (a language about which not much is known, but there apparently is evidence of a relationship to NWC), and PIE is a language of the East European ancestors of the Yamnaya related to Uralic but heavily altered by a NWC-related substratum. (And Anatolian evolved from a conservative peripheral dialect of PIE on the western end of the Yamnaya horizon, and moved into Anatolia from the northwest shortly after 3000 BC.) This model (which C. C. Uhlenbeck already proposed in 1935, on purely linguistic grounds of course) IMHO explains things best.

Not an expert by any means but I‘ve taken an interest in Uralic linguistics recently and it seems like people are increasingly skeptical of even PIE > PU loans, at least on the Uralicist side. I think the consensus is that PU is a considerably younger language than PIE and therefore the timelines don‘t match up, leading to the reinterpretation of most early loans as being from (Proto-)Indo-Iranian.

Yes, PU seems to be younger than PIE, perhaps about 2500 BC or so. The (seeming) IE loanwords in it indeed show some traits that point at a PIE dialect ancestral to Indo-Iranian.

As Travis mentioned, if an eastern origin of Uralic is accepted, this makes the possibility of PIE > PU loans even less likely, although I’m not sure what the state of the research on that question is.

So far, it is little but speculation, I think. Genetics allow to trace ancient migrations, but the genes, as I said above, say nothing about the languages those people spoke. Which is shown by how the Lazaridis et al. paper I linked to above can be interpreted in (at least) two utterly different ways with regards to the origin of the PIE language. And even if Uralic came from the east, why can't IE come from the east too, as part of the very same migration? IE and Uralic would then have separated around the time Indo-Uralic entered Europe. (In fact, this is precisely what I think on this matter.)

[Lērisama]

Treating [i,u] and [y,w] as positional variants is usual; AFAIK, the somewhat more frequent position is to treat the glide variant as the archiphoneme and only have the ablauting vowels /e/, /o/ (and perhaps /a/) as "real" vowels.

That is what I intended to imply, did I miss something?

BTW, gsandi is He/Him.

Thanks¹

¹ Although the obligatory note that I would naturally use them in real life for people I knew the gender of – unless someone makes a thing about it it's just not something I'd consider could be a problem, except from people who still insist on parroting the prescriptions of Victorian grammarians, who thankfully don't seem to notice as much in speech and are getting rarer anyway, although details of the assumptions implied in pronoun use probably belong somewhere else.

[WeepingElf]

Treating [i,u] and [y,w] as positional variants is usual; AFAIK, the somewhat more frequent position is to treat the glide variant as the archiphoneme and only have the ablauting vowels /e/, /o/ (and perhaps /a/) as "real" vowels.

That is what I intended to imply, did I miss something?

I don't think you missed anything. It is crystal clear that /i e a o u/, short and long, plus syllabic resonants and laryngeals, was the (surface) vowel inventory of Late PIE. These are the input to the sound change rules of the individual IE languages. All discussions about different vowel inventories refer to earlier stages (or "deep structures", which are essentially the same) of PIE. As I said before, there are a few instances of /a/ in (Late) PIE that cannot be conveniently ascribed to laryngeals (such as the 'goose' word), which are likely loanwords from other languages, but they nonetheless make /a/ a real phoneme in Late PIE, just as /f/ is a real phoneme in Common Slavic even though it occurs only in loanwords. That with /i~j/ and /u~w/, the consonantal allophone is often considered the basic one is for consistency with the resonants and laryngeals, where the consonantal allophone is considered the basic one, too.

[Travis B.]

Hmm, I'm reading stuff like this that implies that most of the similarities between PIE and Proto-Uralic are likely either borrowing or chance, and that the Urheimat of PU is likely as far east as Yakutia, contrary to past views that set it in the Urals. As for Mitian, I'm seeing it suggested that possibly Mitian is a real thing but does not include IE at all.

That guy has his own axe to grind, namely that IE was a branch of Afroasiatic - which is hardly compatible with the Indo-Uralic hypothesis which he thus had to find an argument against. To me, PIE looks more like a sister of Proto-Uralic than like an Afroasiatic language, but things may be deceptive. My guess at the Mitian Urheimat is in southern Siberia, probably somewhere near Lake Baykal, at the end of the last ice age.

Okay, suggesting a link between IE and Afroasiatic is definitely crackpottery. But I have seen elsewhere that a link between PIE and PU is doubted by many, including many Uralicists, today. As fringe macrofamilies go, I've seen much more convincing links suggested between Uralic and Eskaleut of all things.

As for Indo-Uralic, the findings of this paper are usually interpreted as supporting a Caucasian origin of "Indo-Anatolian", but the genes of course say nothing about the languages the people spoke, even if they can be used to trace prehistoric migrations. My interpretation is that the Caucasian people ancestral to the Yamnaya spoke Pre-Proto-NWC, those that moved into Anatolia spoke a language that evolved into Hattic (a language about which not much is known, but there apparently is evidence of a relationship to NWC), and PIE is a language of the East European ancestors of the Yamnaya related to Uralic but heavily altered by a NWC-related substratum. (And Anatolian evolved from a conservative peripheral dialect of PIE on the western end of the Yamnaya horizon, and moved into Anatolia from the northwest shortly after 3000 BC.) This model (which C. C. Uhlenbeck already proposed in 1935, on purely linguistic grounds of course) IMHO explains things best.

From the reading I've done the PIE Urheimat is likely placed in present-day Ukraine, which would place it close to the Caucasus.

Not an expert by any means but I‘ve taken an interest in Uralic linguistics recently and it seems like people are increasingly skeptical of even PIE > PU loans, at least on the Uralicist side. I think the consensus is that PU is a considerably younger language than PIE and therefore the timelines don‘t match up, leading to the reinterpretation of most early loans as being from (Proto-)Indo-Iranian.

Yes, PU seems to be younger than PIE, perhaps about 2500 BC or so. The (seeming) IE loanwords in it indeed show some traits that point at a PIE dialect ancestral to Indo-Iranian.

I've seen similar suggestions of "PIE > PU" loans likely more Proto-Indo-Iranian or Pre-Proto-Indo-Iranian loans into PU.

As Travis mentioned, if an eastern origin of Uralic is accepted, this makes the possibility of PIE > PU loans even less likely, although I’m not sure what the state of the research on that question is.

So far, it is little but speculation, I think. Genetics allow to trace ancient migrations, but the genes, as I said above, say nothing about the languages those people spoke. Which is shown by how the Lazaridis et al. paper I linked to above can be interpreted in (at least) two utterly different ways with regards to the origin of the PIE language. And even if Uralic came from the east, why can't IE come from the east too, as part of the very same migration? IE and Uralic would then have separated around the time Indo-Uralic entered Europe. (In fact, this is precisely what I think on this matter.)

Per the study that placed the original of the Uralic-speaking peoples in Yakutia, apparently the "Uralic" genes are found in Uralic-speaking peoples in general save the Hungarians (which are likely a case of Indo-European-speaking peoples adopting Uralic from a small foreign elite).If we reject the idea that these genes really are "Uralic" genes, then we have to explain how all the Uralic-speaking people except for the Hungarians happened to specifically get them.

[WeepingElf]

Don't throw out the baby with the bathwater, and don't try to prove or disprove language relationships with genetics. IE and Uralic resemble each other more than what could reasonably be ascribed to chance, no other language family looks more similar to IE than Uralic, though that is insufficient to prove a relationship - but there certainly was at least extended intense contact between PIE and PU, which could hardly have been if PU was in Yakutia. PU words that resemble PIE words are real, the question is only whether they are inherited from a common ancestor or borrowed. What concerns the similarities between Uralic and Eskimo-Aleut, they are undeniable, but my personal hypothesis is that these two families are just especially conservative, which may create the impression that they form a node within Mitian which they perhaps actually don't.

And genes and languages do not always travel together, so one cannot conclude that two genetically related populations spoke related languages, nor that two genetically distant populations spoke unrelated languages. Linguistic hypotheses need linguistic evidence to back them up; genetics can only give evidence of ancient migrations but doesn't tell us which languages those people spoke, and languages often enough travel without genes - language shifts are a thing, and not to be underrated.

[Travis B.]

As for PIE and PAA, I don't know why people even posit trying to unify the two. There simply are no real similarities between either, and to me it seems even less plausible than Indo-Uralic. The unifying factors in AA are its morphosyntax and pronouns, which PIE entirely lacks (hell Early PIE doesn't even have a feminine gender, much the less the classic AA /t/ morpheme). Yes, some may try to claim that IE ablaut is somehow similar to AA templatic morphology, but this seems entirely superficial and there seem to be no real connections between the two.

[WeepingElf]

As for PIE and PAA, I don't know why people even posit trying to unify the two. There simply are no real similarities between either, and to me it seems even less plausible than Indo-Uralic. The unifying factors in AA are its morphosyntax and pronouns, which PIE entirely lacks (hell Early PIE doesn't even have a feminine gender, much the less the classic AA /t/ morpheme). Yes, some may try to claim that IE ablaut is somehow similar to AA templatic morphology, but this seems entirely superficial and there seem to be no real connections between the two.

Concurred 100%. I think this is just the old notion that "both families are inflected", though they inflect their words utterly differently. Also, the tradition of attributing IE to Japheth, Semitic to Shem, and "Hamitic" to Ham - characters that almost certainly never lived, and thus cannot figure in a linguistic theory that isn't completely laughable. Furthermore, the two families have prestige as including languages of great ancient civilizations, but that too of course means nothing (what about Chinese or Mayan?), and can only be considered racist bias.

[Travis B.]

Don't throw out the baby with the bathwater, and don't try to prove or disprove language relationships with genetics. IE and Uralic resemble each other more than what could reasonably be ascribed to chance, no other language family looks more similar to IE than Uralic, though that is insufficient to prove a relationship - but there certainly was at least extended intense contact between PIE and PU, which could hardly have been if PU was in Yakutia. PU words that resemble PIE words are real, the question is only whether they are inherited from a common ancestor or borrowed. What concerns the similarities between Uralic and Eskimo-Aleut, they are undeniable, but my personal hypothesis is that these two families are just especially conservative, which may create the impression that they form a node within Mitian which they perhaps actually don't.

And genes and languages do not always travel together, so one cannot conclude that two genetically related populations spoke related languages, nor that two genetically distant populations spoke unrelated languages. Linguistic hypotheses need linguistic evidence to back them up; genetics can only give evidence of ancient migrations but doesn't tell us which languages those people spoke, and languages often enough travel without genes - language shifts are a thing, and not to be underrated.

What we do know is that Indo-Iranian spread east from the PIE Urheimat, and was spoken in a wide area which is now Turkic-speaking, and that the contact between IE and PU was likely really between PII or Pre-PII and PU rather than between PIE and PU, which is compatible with a view that the PU Urheimat was in the east and then PU rapidly spread west.

[WeepingElf]

Don't throw out the baby with the bathwater, and don't try to prove or disprove language relationships with genetics. IE and Uralic resemble each other more than what could reasonably be ascribed to chance, no other language family looks more similar to IE than Uralic, though that is insufficient to prove a relationship - but there certainly was at least extended intense contact between PIE and PU, which could hardly have been if PU was in Yakutia. PU words that resemble PIE words are real, the question is only whether they are inherited from a common ancestor or borrowed. What concerns the similarities between Uralic and Eskimo-Aleut, they are undeniable, but my personal hypothesis is that these two families are just especially conservative, which may create the impression that they form a node within Mitian which they perhaps actually don't.

And genes and languages do not always travel together, so one cannot conclude that two genetically related populations spoke related languages, nor that two genetically distant populations spoke unrelated languages. Linguistic hypotheses need linguistic evidence to back them up; genetics can only give evidence of ancient migrations but doesn't tell us which languages those people spoke, and languages often enough travel without genes - language shifts are a thing, and not to be underrated.

What we do know is that Indo-Iranian spread east from the PIE Urheimat, and was spoken in a wide area which is now Turkic-speaking, and that the contact between IE and PU was likely really between PII or Pre-PII and PU rather than between PIE and PU, which is compatible with a view that the PU Urheimat was in the east and then PU rapidly spread west.

I understand. I have no problems with a Proto-Uralic homeland in SW Siberia, where it may have come into contact with Pre-Proto-Indo-Iranian; it would still be close enough to the Pontic-Caspian steppe for an Indo-Uralic relationship - and I have no problems with the Para-Uralic component of PIE originating there, unlike some racist nincompoops who insist on deep "racial" divisions between the peoples of Europe and Siberia. Yet, Yakutia is another matter. Indo-Iranian never was close to that.

[Travis B.]

Kortlandt considers the Vestjysk stød (preglottalization of voiceless stops in West Jutlandic), the aforementioned English (pre)glottalization, preaspiration in various North Germanic varieties, along with things such as Sindhi implosive voiced consonants, the dialectal Armenian glottalization we mentioned, and the Latvian glottalization to be relics of a glottalized past. He also considers aspiration in IE daughters to be separate innovations in different branches (whereas the original glottalic theory posits that both *T and *Dh were allophonically aspirated in PIE to explain doublets where *D and *Dh swap places in different IE daughters).

Still, why then is glottalization limited in IE, i.e. why does it not survive in onsets outside Sindhi, if the posited system is such a stable one? The lack of glottalized onsets is what makes me doubt this, and glottalized codas (e.g. English, West Jutlandic) could easily be explained as a separate development in IE daughters from unreleased fortis coda consonants, which are common crosslinguistically.

[rotting bones]

What's the difference between the Indo-Hittite hypothesis and the mainstream chronology besides naming a different point in the family tree "PIE"?

[WeepingElf]

What's the difference between the Indo-Hittite hypothesis and the mainstream chronology besides naming a different point in the family tree "PIE"?

It is mostly indeed a matter of terminology. But the term "Indo-Hittite" or "Indo-Anatolian" is usually associated with the notion that Anatolian broke off a few thousand years before the final breakup of of PIE, rather than just a few hundred years as many (most?) Indo-Europeanists assume, and "Proto-Indo-Anatolian" is often placed in the Caucasus (on the ground of the Lazaridis et al. paper I mentioned yesterday) while only "PIE proper" (i.e., the common ancestor of the non-Anatolian IE languages) is placed in the steppe about 3000 BC. Alas, I think that this interpretation of the genetic findings is misguided, and Anatolian just descends from an archaic peripheral dialect of PIE spoken somewhere near the Danube delta (while Tocharian descends from another peripheral, but less archaic, dialect on the opposite end of the Yamnaya horizon near the Ural mountains, and the "classical" IE languages originated closer to the Yamnaya heartland).

There also is, of course, the Schwundtheorie ('loss theory'), according to which Anatolian is just an "ordinary" IE branch which just has lost the morphological categories it is missing at an unusually early date, though this is a minority opinion these days (which of course doesn't imply that it is wrong). There even have been scholars (such as Jaan Puhvel or Annelies Kammenhuber) who had tried to show that Anatolian was a western IE branch related to Italic, Celtic and Germanic, but those western languages acquired their allegedly Anatolian-like characteristics, mainly the simplification of the verb tense/aspect system to just a present vs. preterite distinction (on top of which Italic and Celtic later innovated new imperfects and futures), late; my conjecture is that this is due to a substratum related to Anatolian, spoken by the Bell Beaker people, but I admit that this is adventurous speculation.

[rotting bones]

It is mostly indeed a matter of terminology. But the term "Indo-Hittite" or "Indo-Anatolian" is usually associated with the notion that Anatolian broke off a few thousand years before the final breakup of of PIE, rather than just a few hundred years as many (most?) Indo-Europeanists assume, and "Proto-Indo-Anatolian" is often placed in the Caucasus (on the ground of the Lazaridis et al. paper I mentioned yesterday) while only "PIE proper" (i.e., the common ancestor of the non-Anatolian IE languages) is placed in the steppe about 3000 BC.

Thanks. Does this mean the Indo-Hittite hypothesis has no direct effect on the results of the comparative method?

Alas, I think that this interpretation of the genetic findings is misguided, and Anatolian just descends from an archaic peripheral dialect of PIE spoken somewhere near the Danube delta (while Tocharian descends from another peripheral, but less archaic, dialect on the opposite end of the Yamnaya horizon near the Ural mountains, and the "classical" IE languages originated closer to the Yamnaya heartland).

I can't comment on this until I have reviewed the archaeological evidence myself. Until then, I'm with the scholarly consensus by default.

[WeepingElf]

It is mostly indeed a matter of terminology. But the term "Indo-Hittite" or "Indo-Anatolian" is usually associated with the notion that Anatolian broke off a few thousand years before the final breakup of of PIE, rather than just a few hundred years as many (most?) Indo-Europeanists assume, and "Proto-Indo-Anatolian" is often placed in the Caucasus (on the ground of the Lazaridis et al. paper I mentioned yesterday) while only "PIE proper" (i.e., the common ancestor of the non-Anatolian IE languages) is placed in the steppe about 3000 BC.

Thanks. Does this mean the Indo-Hittite hypothesis has no direct effect on the results of the comparative method?

I don't know what you mean by that question. PIE as reconstructed in the handbooks is considered "PIE proper" by the adherents of the Indo-Hittite hypothesis, while "Indo-Hittite" is a different, older language. However, the two do not seem to differ much in terms of phonology; many adherents of the Indo-Hittite hypothesis, though, also adhere to the glottalic theory for "Indo-Hittite", so there would have been a major shift in the stop system from "Indo-Hittite" to "PIE proper"; also, it is often claimed that the laryngeals were lost on the way from "Indo-Hittite" to "PIE proper", but the loss of laryngeals went differently in different branches of IE (see such things as the "triple reflex" in Greek, which shows that the laryngeals had not even merged with each other in that branch, or metric irregularities in the oldest parts of both the Rigveda and the Avesta which can be fixed by restoring the laryngeals, which probably means that the laryngeals were lost only as late as about 1000 BC in these languages), so they must still have been there when "PIE proper" broke up.

Yet, Anatolian morphology is starkly different from that of the "classic" IE languages, and it is anyone's guess which of the two is more archaic. The followers of the Indo-Hittite hypothesis usually claim that Anatolian is more archaic, but that is a leap of logic: if "PIE proper" could innovate such things as the feminine gender or the three-way verb aspect system after the breakup of "Indo-Hittite", Anatolian could have lost them after the same event. However, Anatolian seems to represent an earlier stage, as the Anatolian cognates of the "Classic PIE" feminine and verb aspect markers have derivative functions, and derivations are more likely to evolve into inflections than vice versa. They look more like pre-stages than like residues of the "missing" categories.

All this doesn't tell us where PIE (or "Indo-Hittite") was spoken, but the homeland question has repercussions on the Indo-Uralic hypothesis, which works better if the earliest stage of PIE was already in the steppe than if it was in the Caucasus or Anatolia. And the Indo-Uralic hypothesis, though not proven, cannot be dismissed easily, as there are resemblances between IE and Uralic in morphology that can be better explained by a common ancestor than by chance or borrowing. I don't understand Travis's skepticism here, the arguments he invoked were strawmen, and I have the feeling that the goalposts are no longer where they used to be, though that feeling may be deceptive And as I said, the results of the geneticists which some (not all; Kloekhorst, for instance, doesn't!) proponents of the Indo-Hittite hypothesis interpret as evidence of "Indo-Hittite" in the Caucasus before 4500 BC can also be interpreted thus that PIE emerged when the immigrants from the Caucasus, speaking a language related to NW Caucasian, mixed with indigenous Eastern Europeans speaking a language related to Uralic, and the latter language prevailed but was altered by the former language acting as a substratum.

[rotting bones]

Thanks for the detailed discussion.

(Yes, I was asking whether the proto-language of the whole family, regardless of what you call it, would be the same.)

[Ketsuban]

As for PIE *a, I think it originally existed only as an allophone of *e next to *h₂ (I think the Early PIE values of *e, *a, *o and *h₂ were [æ], [ɑ], [ɒ] and [χ], respectively), but there are some items with *a that can't be explained by laryngeals, such as *gʰans- 'goose', which are probably loanwords from other languages, or onomatopoetic formations.

The thing that sticks in my craw about this is that it is simultaneously possible for the majority of instances of *a to be allophony and for there to be isolated instances of actual bona fide phonemic *a. (Marginal phonemes are a thing, and they're not always the result of loans and onomatopoeia.) This amount of ambiguity is fine if you just want Proto-Indo-European as a label for a smudge on a linguistic map of Neolithic Europe, but for Proto-Indo-European as a building block for higher-order relationships it's too flabby. I do not accept the logic (assumed in for example Wiktionary's reconstruction of *gʰh₂éns) that all *a must be the result of *h₂ because most apparent *a can be explained that way without more detail. For example, Wiktionary points to *gerh₂-ḗn "crane, heron" and notes that "there is some confusion" between *gerh₂- ~ *ǵerh₂- "cry hoarsely" and *ǵeh₂r- "cry, shout". The idea of *gʰéh₂n- as yet another variant underlying *gʰéh₂ns "goose" seems plausible; the multiplicity of forms is explained as onomatopoeia more comfortably than a lone inconvenient word.

So, what can we do about *bak-?

[hwhatting]

I'd say that Indo-Hittite nowadays is just a historical label.
The real question is whether the spread of IE was a kind of "starburst", where the branches split off more or less at the same time, or can we observe a series of split-offs, with the languages staying behind showing common developments? Originally, Indo-Hittite meant that one could basically ignore Anatolian and go on reconstructing core-PIE as in the classical Brugmannian model. The opponents of I-H came from two sides - those who maintained that Anatolian can be explained based on the classical model, just with having lost a lot of features, and those who maintained that Anatolian showed that the classical model was wrong. Both preferred the "starburst" model, because it took Anatolian out of its role of being somehow "special" and separate.
I'd say that common opinion nowadays is that Anatolian indeed split of first, and that the classical model needs to be modified based on the evidence of Anatolian; the question is to what degree. And most IEanists also assume that Tocharian split off second, with the remaining branches undergoing further common developments.
BTW, the idea that the female gender is an innovation in IE is an old one; it was postulated based on the fact that most declensions (the ones that are ot o/eH2) originally didn't seem to distinguish female and male, but only neuter and non-neuter, and that you can observe an increase of male-female differentiation in the declensions of many IE languages; on the existence of epicene o-stem adjectives (e.g., Greek o-stem adjectives that distinguish only neuter and non-neuter and have no female forms in -eH2) and of female o-stems, and on the similarity of the *(e)H2 female ending to the collective / neuter plural. Anatolian only made this idea more mainstream and provided additional evidence for it, and it changed the periodization of the development of the female gender from "some point in the prehistory of PIE" to "after the split-off of Anatolian".

[WeepingElf]

So, what can we do about *bak-?

This word contains both of the "problematic" phonemes, and for this reason is held to be a loanword (though perhaps already into PIE, rather than daughter languages) by many scholars. But as you say, marginal phonemes are a thing, and not always a sign of a non-inherited word, though they often are. One must not forget that PIE was just a language, with a long history of its own, with different dialects, probably with living relatives, and certainly in contact with other languages, even if we can only catch faint glimpses of these connections. All reconstructions of PIE are just models and not the real thing (hence the custom of marking reconstructed word forms with an asterisk)!

[WeepingElf]

I'd say that Indo-Hittite nowadays is just a historical label.
The real question is whether the spread of IE was a kind of "starburst", where the branches split off more or less at the same time, or can we observe a series of split-offs, with the languages staying behind showing common developments? Originally, Indo-Hittite meant that one could basically ignore Anatolian and go on reconstructing core-PIE as in the classical Brugmannian model. The opponents of I-H came from two sides - those who maintained that Anatolian can be explained based on the classical model, just with having lost a lot of features, and those who maintained that Anatolian showed that the classical model was wrong. Both preferred the "starburst" model, because it took Anatolian out of its role of being somehow "special" and separate.
I'd say that common opinion nowadays is that Anatolian indeed split of first, and that the classical model needs to be modified based on the evidence of Anatolian; the question is to what degree. And most IEanists also assume that Tocharian split off second, with the remaining branches undergoing further common developments.
BTW, the idea that the female gender is an innovation in IE is an old one; it was postulated based on the fact that most declensions (the ones that are ot o/eH2) originally didn't seem to distinguish female and male, but only neuter and non-neuter, and that you can observe an increase of male-female differentiation in the declensions of many IE languages; on the existence of epicene o-stem adjectives (e.g., Greek o-stem adjectives that distinguish only neuter and non-neuter and have no female forms in -eH2) and of female o-stems, and on the similarity of the *(e)H2 female ending to the collective / neuter plural. Anatolian only made this idea more mainstream and provided additional evidence for it.

Yep. As I have said before, the truth probably lies somewhere in the middle between the extremes of the Indo-Hittite theory and the Schwundtheorie, and the "starburst" model is merely a way of saying "We don't know how PIE broke up into the various branches". Some "internal nodes" are emerging, though. Indo-Iranian is rock solid; Balto-Slavic is also quite certain, even if some Lithuanians and Latvians doubt it for political reasons; Italo-Celtic is still controversial but seriously discussed, and looks good enough to me. There also seems to be a "Paleo-Balkan" group consisting of Greek, Albanian, Thracian (which may be the ancestor of Albanian), Phrygian and Armenian (of which we don't know yet how it got into its historical seats; I conjecture that it descends from Cimmerian, a language of which we know hardly anything, though). This brings the number of primary branches within "Post-Tocharian" IE down to just five. There furthermore seems to be a connection between Balto-Slavic and Indo-Iranian which have a few but significant shared innovations (such as the satem shift with RUKI rule), though Balto-Slavic also shares features (such as the dative plural in *-mos) with Germanic. It seems as if Germanic, Balto-Slavic and Indo-Iranian descend from the PIE dialect spoken by the Corded Ware people, with Germanic diverging earlier than the other two. We are approaching the limits of the family tree model here, and get into territory where the wave model does a better job - early IE was an expanding dialect continuum, with innovations springing up in various places and spreading across the dialects, resulting in intersecting isoglosses. You can see that in just about any dialect continuum, e.g. in German.

[hwhatting]

Hrrrm. Graeco-Phrygian keeps coming up, although it has been shown that all demonstrable isoglosses are either rententions (so proving nothing about a closer genetic relatedness) or shared with wider groups (e.g., the augment with what some people have called Graeco-Aryan - the core group that comes closest to the Brugmannian model.)
I agree that for anything after the split-off of Tocharian, we don't seem to have clean split-offs of single branches anymore, but a dialect continuum in which developments spread to a differing extent before it dissolved into indivudual branches.