Looking at Indo-European studies has for a while now been giving me an impression that the usual vowel system reconstruction has unnoticed flaws in it.
They are different issues from the long-running debate on the reconstruction of the stop system, though. The traditional *i *e *a *o *u, easily attestable around the world, surely has nothing wrong in it in terms of synchronic phonology. Adding in the syllabic resonants *m̥ *n̥ *r̥ *l̥ won’t be a major typological problem, either. Rather… weird things start to pile up once we instead survey the development of this vowel system in the IE languages.
For a starting point, let’s consider Anatolian. I claim no particular expertise in the area though, so instead of getting my hands dirty with data, my commentary here follows fairly closely some short overviews by H. Craig Melchert. [1] He ends up positing (in an update to earlier views about a simpler 4+4 system) a vowel system almost identical to PIE for Proto-Anatolian: five short vowels *i *e *a *o *u, their long counterparts including *ā < *eh₂, as well as an unpaired long vowel *ǣ < PIE *eh₁ (early on also a later redacted “*ẹ̄” < PIE *ey). All of these yield their own distinct correspondence sets, and I would not try to claim that we need to merge or split some of these phonemes. But there are some imbalances in how some different contrasts develop. Melchert does not go into featural phonology, but if we are to trust his transcription, both *e and *o would be mid vowels. Their development tendencies however diverge. There is one general similarity: most Anatolian languages seem to show a trend of qualitatively simplifying the vowel system, towards plain *i *a *u. This is completed only in Luwian, but elsewhere, too, the mid vowels have a tendency to merge with other stuff. Short *e often yields *i in some kind of raising contexts: e.g. following *j, or when pretonic (kind of resembling Germanic). In a few other positions, there are conditional developments to *a, such as before *n in Hittite and Lydian. However, by contrast, there seems to be no evidence for a raising development *o > **u. Most Anatolian languages have generally merged *o and *ō into *a and *ā. Melchert only reports three features that allow distinguishing *o and *a:
- In Hittite, stressed *o in closed syllables yields long /ā/, while *a remains short /a/.
- In Lydian, stressed /o/ is found next to a labiovelar (either a stop *Kʷ or the glide *w).
- In Lycian, the general treatment is *o > /æ/ (transcribed e; no comment on what happens to *ō).
If, in a language family elsewhere, we were faced with two correspondence sets — one of them *a ~ *a ~ *a, the other *a/ā ~ *æ ~ *a/o — I would definitely not conclude that we are to reconstruct *a and *o respectively. And I would assume that Melchert, too, only ends up reconstructing a mid vowel *o, because this is what the second Anatolian vowel corresponds to in traditional PIE, not because the reflexes so demand. Even /o/ in Lydian looks like it might represent some kind of an assimilation from the adjacent labiovelars, rather than the preservation of original rounding.
The long vowel situation seems even more worrying. We would definitely expect to see a raising *ō > **ū at least somewhere, at minimum in languages like Lycian or Luwian where *ē > *ī, if these two had made up a similar class of long mid vowels. But apparently we only get /ā/ everywhere. Melchert reports for this contrast but a single distinguishing feature: apparently *dwō- yields /dā-/ in Hittite, versus no such loss of the glide for *dwā-. This seems to me much too iffy grounds for setting up a separate *ō.
Ignoring traditional PIE for the moment and instead reconstructing *a₁ (in place of *a) versus *a₂ (in place of “*o”), there would seem to be more promising options for phonological interpretation available. In terms of height, I’d assume that it was actually the latter that was the more open vowel *[a]. This is fairly directly suggested by the different treatment in Hittite: all other things being equal, more open vowels tend to be realized as longer. No clear evidence seems to exist for a difference in backness; *a₁ remains stable-ish (though I presume there would have been some variation on if a stands for central [a] or back [ɑ]), while *a₂ has both clearly fronted (Lycian) and backed (Lydian) reflexes. This, however, provides another reason to suspect a lower value for *a₂, given that backness contrasts tend to be more labile among lower vowels.
What this seems to leave available for *a₁, then, is some kind of a weaker vowel value still prone to lowering, like [ɐ], [ɜ] or [ə], probably both [-front] and [-round]. It seems a bit curious how this has not been retained as such anywhere, [2] but hardly any more so than the failure of *o to surface consistently anywhere (or any other family-wide “sweep” development, such as *s > *h in early Iranian or *a *ā *u *ū > *o *a *ъ *ɨ etc. in early Slavic).
Compare to this e.g. the develoment of English short a (Early Modern /a/) and laxed u (Early Modern /ə/): the former has split over the last few centuries into a variety of lengthened (BATH lexical set, father, “tense æ”), fronted (TRAP set) and/or backed (PALM and WATER sets) reflexes, while a new neutral short /a/ is in numerous varieties filled in from earlier †/ə/ > †/ʌ/. Similar vowel histories can be found moreover e.g. in Samic varieties (old *ā > á being more heavily split in allophones etc. versus lowered *ë > â/a remaining more neutral) or in Samoyedic (old *å *a yielding a large variety of reflexes versus *ə, often lowered, remaining more neutral).
Melchert’s most recent work also mentions the recent discovery of a “new” /o/, /ō/ for several Anatolian languages, in earlier work conflated with u, ū. The short version mainly evolves from labiovelar + syllabic resonant, the long version from *aw, *ow, both also from *u next to laryngeals (thus this /ō/ corresponds to late PIE *ū, from *uH; remaining cases of Hittite /ū/ are instead from *ew, or from stressed open-syllable lengthening of *u). These are therefore clearly distinct from traditional PIE *o, *ō. If this new *o could have been in place already in Proto-Anatolian (apparently plausible at least in non-final syllables), it’s all the more reason to not suppose also the simultaneous retention of old *o.
Given that Anatolian retains numerous archaisms, and the possibility of it being the earliest split-off of Indo-European entirely, we can also proceed to ask an important question: would Proto-Anatolian *ɜ *a or traditional PIE *a *o be the more archaic state of affairs? I would end up preferring the former: a chain shift a > ɑ > o, ə > ɜ > a is more typical than the opposite.
As soon as we’ve formed this hypothesis for a “skew triangular” (or perhaps even “square”? [3]) vowel system *i *e *ɜ *a *u for not only Proto-Anatolian, but also Early PIE altogether, there will be numerous immediate implications. I will not go into listing all of these just yet… But to mention one, this will nicely amount to addressing the now and then raised typological objections about the rarity (and possible absense entirely before laryngeal coloring) of traditional PIE *a. In the new system, this turns out to translate into the rarity of the more marked vowel *ɜ, while the proper cardinal open vowel *a is quite frequent indeed.
[1] 1992, “Relative Chronology and Anatolian: The Vowel System”, in Rekonstruktion und Relative Chronologie. Akten der VIII. Fachtagung der indogermanischen Gesellschaft, ed. Robert Beekes;
1993, “Historical Phonology of Anatolian” 1993, Journal of Indo-European Studies 21/3-4;
and 2015, “Hittite Historical Phonology after 100 Years (and after 20 years)“, in . I have not yet seen his 1994 monograph Anatolian Historical Phonology, but the 2015 paper seems to summarize the main points.
[2] In writing at least. It should be kept in mind that epigraphic evidence does not actually constitute phonetic evidence.
[3] Since *e may have well been half-open [ɛ] rather than half-close [e].
Freelance reconstruction 
Fascinating! I’ll wait for the other parts before I comment on the proposal itself, but so far I find it very promising.
:-) Unlike Navajo, German does not have sibilant harmony: Gesellschaft.
That should be the default assumption. As far as I’m concerned, it’s one of the IPA’s greatest flaws that the symbols e o were chosen for [e o] instead of for the much more common [ɛ ɔ].
Wouldn’t be the first time I stumble on this, won’t be the last. (“Hey, shall we have a brainstorming sesson on nice shushi places to go grab some dzuicy fish shnacks at?”)
(Perhaps you’ve seen this, but) Kümmel has tried to reinterpret the vowel system as well, coming to the conclusion that the *o:*e contrast represents an earlier **ā:**a contrast, predating the emergence of classical *a. He mentions some of this Anatolian evidence, Brugmann’s law, etc., in favour of the **longer value for *o, but confines most of his discussion to internal reconstruction of ablaut and gives a Late PIE system where *o [ɒ] had already been shortened, leaving unaddressed whether Brugmann’s law is supposed to still somehow be a retention. But as his *o is still lower than his *e [ɛ], your mechanism would obviate the comparative need for his earlier stage, it seems.
https://www.academia.edu/1538887/Typology_and_reconstruction_The_consonants_and_vowels_of_Proto-Indo-European
Looking forward to your future installments.
Here, especially on slide 15, is another; and Miguel Carrasquer-Vidal has a similar one on his Academia page that I don’t have time to look for right now.
I have seen both, yes. I will be commenting at least on Kümmel’s proposal in the next post in this series. Currently I have no plans on going into the details of different ablaut patterns with this idea though… if I find time to dwell more on IE reconstruction, first I’d like to dig into just how much ablaut exactly is strictly reconstructible in the first place. My main hunch from looking at typical descriptions of PIE is that “ablaut” is typically presented in an overly phonological sense and probably lumps together various morphologically unrelated vowel alternations.
There are actually better grounds to set up separate PAnat. *o *ō, namely the contrast between PIE *ó causing lenition, while PIE *h₃é does not. Since long vowels cause lenition and short vowels do not, this is good evidence for a development *ó > *ṓ, *h₃é > *ḫo.
Thank you! A good addition.
This appears to be quite compatible with what I have outlined here — nothing in this argues for reconstructing *o as specifically /o/, only for reconstructing *h₃-coloring (or, perhaps rather, the merger of colored *e₃) as a later change than the lengthening of the former vowel. It may have interesting further implications for relative chronology, though. Melchert dates medial lenition as Proto-Anatolian, which would then suggest that there have been two layers of stressed vowel lengthening in Hittite. I would wonder if we instead should consider medial lenition an areal Anatolian innovation? It is after all often widespread once it takes root, as is the case e.g. in most of Iranian and most of Uralic.
Or, alternatively, for reconstructing *o as inherently long and *e, even when colored, as inherently short, so that they stayed different phonemes in PIE and just happened to have the same quality. Partial loss of the inherent length would have happened only later in the non-Anatolian branch of IE (and of course full loss still hadn’t happened when Brugmann’s law operated).
…eh, sorry, that’s what you mean by “the merger of colored *e₃”. I’ve had my morning tea, but I seem to be wholly insensitive to caffeine somehow.
Very interesting observation. I am readily convinced that there is an issue, but I am not yet 100% happy with your alternative suggestion.
I stumbled over
“In the new system, this turns out to translate into the rarity of the more marked vowel *ɐ”
My (perhaps too uninformed) impression is that such “semi-reduced” vowels, while occurring only in a restricted set of languages each, are rather frequent once they are featured in a specific language. (Example: ă in Rumanian. If I am not mistaken, Rumanian ă is identical with ɐ and thus contained in the range you give for *a₁.)
This objection is, of course, not a severe issue, since it does not cause more incompatibilities, it just fails to resolve another one; so overall your suggestion may still look better.
However, I see another problem in the context with laryngeals (if you bear with me for a moment, you might find this on topic in the end). Unless we want to keep laryngeals as abstract symbols for “unknown sound, different from all others and one another”, we need to assign concrete phonological values, and while I do not know of a specific assignment that were generally accepted, the idea to assign both a consonant and a vowel value to each laryngeal type.
Now, I personally think that vowel values are indispensable (if you disagree, say *pXtḗr with a non-sibilant non-sonorant consonant X of your choice). But is it realistic to assume that the speaker of a language automatically switch between a certain consonant and a certain vowel if the phonological neighbourhood requires so, keeping this correspondence intact all the time? Yes, if we talk about y/i, w/u, or (in a wider sense) r/r̥ and similarly for the other resonants. But otherwise, I clearly plead “no”.
If it were not for Anatolian, there were no reason to assume anything but (reduced) vowels at all for the so-called laryngeals. Now it is no need to ignore the Anatolian evidence. Perhaps it is at least worth asking whether ḫ may as well be explained as a “glide” consonant inserted in certain neighbourhoods (next to the alleged laryngeals, that is), though that does not seem likely. A more promising approach starts with the observation that there is, to my knowledge, no evidence of different _consonant_ reflexes of h₁, h₂, h₃ in any Indoeuropean language at all. It therefore suffices to assume three different vowel qualities ə₁, ə₂, ə₃ (to be specified) and read h₁ = Hə₁, h₂ = Hə₂, h₃ = Hə₃ in every position. Here, H is the same consonant in each case; most simply a glottal stop, perhaps already the Hittite ḫ (and thus an actual laryngeal).
Now it remains to find suitable vowel qualities ə₁, ə₂, ə₃. Since they tend to be rather elusive, especially next to a ordinary vowel, it makes sense to consider reduced (“ultra-short”) vowels. The vowel reflexes, in the case where they differ, already suggest a certain indication what the vowel should look like: e-colouring –> standard shwa, a-colouring –> ɐ as above, o-colouring –> something similar to French unstressed -e- (‘celui’).
Now why did I bring this up? We might need *ɐ for the “laryngeal” set! To reconcile these two thoughts, we would need to assign a value sufficiently different from that to your *a₁. Moreover, I think there needs to be a significant difference in “reducedness” between regular IE short vowels and the “shwa series” above (where “reducedness” may refer to quantity, tenseness, something else, or a combination of these.)
I suppose that there is still a way for such a reconciliation, but I do not have a concrete idea how.- So these are my 2 cents to this, and I hope that this is acceptable to your discussion in spite of its obvious amateur origin.
Alas, this not quite the case: Anatolian ‹ḫ› only continues *h₂ and in some positions possibly *h₃, while *h₁ does not leave any reflex. There’s also Cowgill’s Law in Germanic, which suggests something similar: *h₃w and possibly *h₂w > *gw > *kʷ, after a sonorant.
It is regardless a good observation that what we could call the “syllabic laryngeal” *ə₂ and “a-colored” *e₂ need to be kept apart. This is actually the main reason why I use here symbols like /ɜ/ or /ɐ/ for the latter, rather than outright /ə/. For “syllabic” *ə₁ though, if I was going to explain these as reduced vowel phonemes (allophones?), I would probably posit instead something fronted, perhaps a vowel similar to English /ɪ/.
At some later point I am going to have to collate laryngeal-theory-related complications like these in general. We will see if I can put together something that is at least as coherent than standard PIE.
Thanks for your reply!
As to different reflexes of the laryngeals, I was aware of the different behaviour in Anatolian (though not Germanic). My point was not questioning the differentiation between laryngeals 1,2,3, but that there are never different types of consonants occurring in the place of an old laryngeal (in Anatolian ḫ(ḫ) or nothing, in in Germanic k or nothing; rather than, say, *h₂ > **k, *h₃ > **h in Germanic). Of course ‘nothing’ is a very different reflex than something. Sorry for putting this completely wrong.
I meant to say reflex structure seems to allow a reconstruction with a common consonant (h₁ = Hə₁, h₂ = Hə₂, h₃ = Hə₃ or rather h₁ = ə₁H, h₂ = ə₂H, h₃ = ə₃H).
Not necessarily. [ɐ], ignoring the diphthongs that end in it, was clearly very rare in Bavarian-Austrian dialects such as mine before non-rhoticity swept through and made [ɐ] common.
On “syllabic laryngeals”, this work on PIE syllabification should be required reading. tl;dr: on the morphophonemic level, PIE had lots of consonant clusters that were phonotactically illegal; some of these (the rules are in the paper) were repaired by deletion of a segment, others by insertion of a non-phonemic epenthetic vowel that developed differently in different branches, often merging with a vowel phoneme.
“Cowgill’s law”, BTW, only has one good example, *unk- “us (dative)”. All the others, and more, are better explained by Seebold’s law (*/jw/ > */jgw/ before Grimm).
One consonant with three different vowels doesn’t explain the voicing effect of *h₃ (not just in the “drink” word, but also in words with the the “Hoffmann suffix”), or the fact that *h₁ never yields Anatolian ḫ.
I’m not familiar with a Seebold’s Law. How does this account for seeming counterexamples like *blīwą ‘lead’, *hlaiwą ‘tomb’, *Tīwaz ‘Tyr’ (if it, as you imply, still covers *kʷikʷaz ‘quick’)?
I’ll try to look it up soon.
No, that one seems to have a quite different origin.
Haven’t found it yet, but here’s a whole thesis on PIE syllabification, by the same author as the link above.
The solution is simple: I had gravely misremembered Seebold’s law, which is much more specific – between a non-syllabic “sonorant” and (pre-)Germanic */u/, */w/ becomes pre-Grimm */g/ by dissimilation. Sorry for the confusion.
Seebold’s law allows us to connect the “squirrel” words of Germanic and Balto-Slavic, for example: PGmc. */aikurna/- < */ajgurna/- < */ajwurna/- < */ajwr̩no/- (zero-grade plus suffix *-no), Proto-Slavic */væver/- < */waiwer/- </ */wajwer/- (e-grade without suffix). The */k/ in the PGmc. e-grade form */aikwerna/- would have to be analogical, and I have no clue whether the initial */w/ was dissimilated away in Gmc. or copied in in BSl..
“this work on PIE syllabification”
That was interesting reading. Thank you for the link!
The author gives plausible reasons why, among other things, laryngeals are so often deleted or vocalized.
And of course, as in any modern novel, the fault lies not in their own characters, but in the influence from their environment (i.e. admissible clusters/syllables).
He starts with the explicit but casual assumption that “h₂ was a consonant and not [+syllablic]”.
(For stylistic reasons obscure to me, he does not write either “… and not syllablic” or “and [-syllabic]”.)
Somewhere halfway into his parade of (insightful!) rules on IE syllable structure, he mentions “This analysis requires that the laryngeals were more sonorous than stops.”
So he seems quite serious about this. In particular, I understand that he implies that IE *h₂ster was pronounced with three leading consonants, with h₂ up in the air similarly to the “s” in English “thieves” or the “p” in Greek “pteron” (“extrasyllabic element”).
So far I can follow and may accept it (even though **pster, **kster, **fster and others seem a lot more pronouncable and recognizable than *ʔster, *hster or even *ḫster to me).
But this only works for some examples. The author discusses the genitive of “daughter” at length; but how would you pronounce the nominative?
You have the cluster in *dugh₂ter (or *-ters or *-terr or *tēr, perhaps all of these at different times). According to Byrd, there is no problem with this word, no need for schwa insertion. So how do you pronounce it?
Curiously, no IE daughter language seems to agree, all of them deleted the laryngeal (including H-Luvian) even there – if I understand him correctly, the laryngeal at least worked by implying a syllable boundary so that the *-g- could develop independently without being assimilated to the *-t-.
So he makes a point about when *h₂ morphes – even at IE times – into *h₂ə, and when into *∅. But he speaks little of the case when *h₂ can, according to his rules, stays in place unaltered.
So although this analysis may be very insightful in predicting when vocalisation of laryngeals takes place and when it doesn’t.
But I am still not convinced that the Proto-IE speakers had reached a mastery of consonant clusters that none of even their most distant successors are able to live up to.
Let’s turn back to *h₂ster. Again, Byrd rules out the need for vocalisation / epenthesis. As far as I can see, all reflexes except in Greek and Anatolian can be explained by a simpler root **ster-.
Hittite has ḫa-, Greek has a-; both of which look more like *h₂ester than *h₂ster, and this pair fits into the usual ablaut pattern.
But is it really adequate to assume *h₂ster, or should we rather assume a duality of stems **ster-/**h₂ester?
There may be other processes than ablaut at work (e.g. prefixing), and ablaut itself may have several variants and special cases that lack from the the extremely systematic reconstruction.
Reconsider the (random, isolated) examples thiev-es, p-teron for extrasyllabic elements above: Both have arisen through a deletion of a vowel; the former as deletion after continued reduction, the latter through an ablaut process.
Shouldn’t we assume that a language featuring such awful sound clusters as *dugh₂ter, *ph₂ter, *h₂ster has developped from another one where these clusters were adorned with a few extra vowels?
It is rather arbitrary where to draw the line between Proto-IE and Pre-Proto-IE, but I do think it is important to take into account that we may be able to have a glimpse on several different stages of Proto-IE.
So far I view the traditionally reconstructed laryngeals as well-justified and well-distinguished placeholders.
I have no problem with an analysis like Byrd’s article that (in consequence) finds criteria for different types of reflexes in all daughter languages (where he thinks of *h_i vs. *∅ vs. *h_iə).
But I haven’t read a convincing suggestion for a realisation of these symbols.
That’s because you speak languages with /f/ but not with /ʔ/, right?
…[dʱugˈχtɛːr]?
I assure you, I pronounce consonant clusters beyond the reach of PIE on a daily basis. “X-rayed” in German is geröntgt*, and because I grew up too far south to interpret g as [ç], I pronounce tgt as a word-final plosive chain without further ado.
* Etymology: the name of Conrad Wilhelm Röntgen, the discoverer of what he called X-rays, looks like a verb, so it became one and is now conjugated normally.
We should assume unexplained variation on no further evidence? :-/
Sure. But the burden of evidence is on you, because you propose a less parsimonious hypothesis.
Why not! But that is, at best, the next step of reconstruction. First we need to get PIE in the strictest sense right, the language at the moment of breakup into Anatolian and Non-Anatolian.
h₂: [χ] or possibly [ħ]
h₃: [ʁ] or possibly [ʕ]
h₁: either [h] or [ʔ]
h₁ is hard. It isn’t preserved as such anywhere, except maybe in Hieroglyphic Luwian and even less certainly as an inconsistent spelling convention in Hittite in initial position. The proponents of this argue that the sound in question was [ʔ]. In Greek, however, *h₁j- shows up as h, possibly [hj] > [ç] > [h]; and in Indo-Iranian it apparently aspirates preceding plosives just like h₂ does, hinting at [h] again. The [ʔ] of Classical Nāhuatl has become [h] in some of its descendants, so it’s entirely possible that h₁ was [ʔ] in PIE but [h] in the last common ancestor of the IE languages that survive today, for example. Another factor to consider is that some people think all IE words – or roots – had to begin with a consonant, prompting them to reconstruct *h₁ in front of all supposedly vowel-initial roots; if that is correct, it strongly hints at [ʔ], but whether it’s correct is kinda hard to test. Its behavior as a fricative or a plosive (on the sonority scale and suchlike) is ambiguous/unclear, judging from Byrd’s thesis including footnotes.
h₂ is easy: it colors *e to *a, it’s a fricative, and there’s no evidence of voice. That leaves [χ] and [ħ] as the top two candidates, with [h] a distant third. The Lycian spellings alternate between kappa, qoppa and chi, making [χ] a more probable candidate than [ħ]; on top of that, [χ] is the literal reading of (cuneiform) ḫ. And, BTW, there are hints that at least one of the sources of *h₂ was a Proto-Nostratic *q.
h₃ likewise behaves as a fricative, voices preceding plosives and shows interesting behavior in Anatolian, where all obstruents were voiceless: in some contexts it disappears, between vowels it becomes the short version of the long reflex of *h₂, in places where consonants aren’t allowed to contrast for length (…or, perhaps, where consonant length can’t be written in cuneiform…) it merges with *h₂ altogether. This looks like h₃ differed from h₂ only by being voiced. The coloring is different (*e to *o), but even without our esteemed host’s new proposal this doesn’t mean it had to be labialized as many have believed – the consonants that are known to be labialized (*w, *gʱʷ, *gʷ, *kʷ) had no coloring effect, which is why the queen’s wedding isn’t the quoon’s wadding, as someone put it. Instead, the voice could have had a tonogenesis-like effect like those that caused the vowel-system splits of Khmer or (Old > Middle) Chinese.
Moar crazy IE consonant clusters: Warsaw East Station is Warszawa Wschodnia, with word-initial [fsx] – three fricatives, more than PIE could apparently deal with.
Perhaps, David, you will win me over with you comprehensive arguments. So far, I’m afraid, that’s not yet done.
“But is it really adequate to assume *h₂ster, or should we rather assume a duality of stems **ster-/**h₂ester?
We should assume unexplained variation on no further evidence?”
Actually, it is exactly the converse:
The evidence only points to *ster and *h₂ester; more precisely, even to *ster and *h₂aster, and without Anatolian, *ster and *aster.
Deriving *h₂ster is a theoretical extra step beyond the evidence.
While variations between full and zero grades are well attested, nobody guarantees that applying this here makes any sense: The crucial difference to, say, *sed- is that there is no evidence that there was a variant like *h₂ster with all of these consonants in word onset.
And when I talked about unrealistic clusters, I didn’t mean to emphasize the sheer number of consonants, but rather the type of the clusters.
Perhaps I’m simply missing something, but is there any (single !) word in any language in the world starting (!) with χst-, or pχt-?
And while we’re at it, there is also no evidence for **ph₂tēr, only for *pə₂tēr.
Now you have three options:
i) Accept the dogma that *ə₂ always equals *h₂ə (or *əh₂), and that these combinations have directly yielded the vowel reflexes in the individual daughter languages.- This implies assuming a consonant here that is nowhere attested as a consonant, but vanished everywhere independently. Moreover, it involves the assumption of an unplausible consonant cluster.
ii) a) Assume some kind of “reduced” vowel *ə₂ still on a Proto-IE stage that is always derived from *h₂ə (or *əh₂). This is a better explanation for the absense of the consonant *h₂ in reflexes of this word everywhere, and you still get the completely systematic phonetic system on an earlier stage of Proto-IE.
ii) b) Assume some kind of “reduced” vowel *ə₂ that may have derived from earlier *h₂ə (or *əh₂), but not necessarily in this case. This makes the phonetic system of Proto-IE somewhat less systematic, but it is closer to the actual evidence.
It is hard to tell whether ii)a) or ii)b) is more “parsimonious”, as a) saves phonems (in the earliest stage), while b) saves development steps (for single words).
However, I’m positive that both variants ii) are simpler than i).
But then, perhaps we’re both just examples of the Adam Scott razor: The explanation you want to believe will always seem the simplest to you …
As far as your suggestion *[dʱugˈχtɛːr] is concerned:
I don’t see why the syllable boundary shouldn’t be after the *χ (Byrd’s article doesn’t seem conclusive here).
And anyway in this position, I would expect assimilation *gχ > **kχ, so that we would have only reflexes of **dukter (and not have: this particular discussion).
Finally, there are some general problems with the concrete phonological values of the laryngeals.
First, and I don’t get tired of mentioning it, they vanish almost everywhere as consonants (give or take a few cases of Cowgill’s law).
That implies that they must have been a series of particularly “weak” consonants.
ʔ would do that, perhaps also h and ʕ.
But I wonder if there was a χ or ɣ among them, why they would not merge or at least interfere with, say, gʰ, in any of daughter languages.
The only way to make this situation more plausible is to lock Anatolian in as first split-off (without Tocharian), and assume the last act, where laryngeals color their neighbourhoods while dissolving like fizzy powder, for this “Late Common Proto-IE” stage.
But there is yet another issue:
The e-coloring is (arguably) the most important part of the job description for would-be laryngeals.
While I don’t see a problem with assuming sound shifts like *eχ > *aχ > *ah > *ā, the shift in the a-direction is by no means compellent; there are sufficiently many examples of /eχ/ in attested languages without any plans to dye it /**aχ/ any time soon, like Russian “grekh” (sin) or Spanish “ejercer” (practice, perform) or Swiss German “Recht” (right, law).
Therefore I don’t believe that shifts like this have happened in several branches independently. Again, this should make us to locate the coloring effects at a (later, but still complete) Proto-IE stage.
However, in the converse situation, why should *χe > *χa take place?
Are both of these sound shifts together attested anywhere?
More plausible shifts would be χe > çe or χe > çi.
Will you accept Polish chrzcić [xʂt͡ɕit͡ɕ] ‘baptise’ as precedent for [χst-]? I’d also offer pchnąć [pxnɔ̃ɲʨ̑] ‘push (perf.)’ towards [pχt-] but I expect you’ll deem a nasal to be too unlike a stop.
Byrd states unambiguously that his PIE syllabification rules allow only SSP-compatible -CC codas but a broader class of CC- onsets (item 10, section 3.7 of Predicting IE Syllabification Through Phonotactic Analysis). Fricatives aren’t less sonorous than stops, which rules out [dʱugχ.tɛːr] for him.
One thing that is bothering you is a practice of Indo-Europeanists which is indeed unjustified by the comparative method, namely, writing vowels in the e-grade of ablaut as *e even next to laryngeals in their reconstructions. For whatever reason, the done thing is to cite such forms in an earlier internally-reconstructed stage. I don’t know any Indo-Europeanist who thinks that laryngeal colouring actually postdates even the separation of Anatolian.
I didn’t have good Internet connections recently, and I still won’t for a few more days. I’ll have more to say later, especially about how vowel coloring works in attested languages. For now, I’ll just recommend forgetting about considering any reconstructed or misreconstructed consonant clusters of PIE at all extreme: this is what extreme looks like.
Finally! Sorry for the delay.
Alex Fink has provided an example even more complex than [χst]-. Concerning [pχt]-, note that it has a sonority rise in the middle of the cluster (plosive, fricative, back down to plosive), which was not allowed in PIE*; consequently, the *ph₂t- of the “father” word never surfaced as such.
Nominative:
*|pχtɛrs| morphophonemic
*/ptɛːr/ phonemic
*[ptɛːr] phonetic
Genitive:
*|pχtrɛs| morphophonemic
*/pχtrɛs/ phonemic
*[pəχˈtrɛs] phonetic
Because their field is older than most of phonology, IEists have the quite counterproductive habit of randomly mixing morphophonemic, phonemic and phonetic transcriptions without saying so or being aware that they’re doing this – sometimes within the same reconstructed word.
* Except, marginally, when the fricative was *s. However, clusters of /s/ + plosive behave so strangely in IE generally that some have suggested, e.g. here, that they shouldn’t be considered clusters in the first place, but unitary phonemes.
Byrd puts a lot of emphasis on the Avestan nominatives ptā and ftā, of which at least the latter is the expected reflex of *[ptɛːr] after the laryngeal deletion that Byrd expects. Byrd goes on to hypothesize that the other branches mostly regularized the strange paradigm of *[ptɛːr], *[pəχˈtrɛs] by restoring the laryngeal with its epenthetic vowel to the nominative: *[pəχˈtɛːr], *[pəχˈtrɛs].
Alex Fink has answered this.
I’m not sure why this doesn’t happen. However, this is the easiest place to block it: at the beginning of a long consonant cluster, across a syllable boundary. It could be relevant that PIE had word-final voicing.
We do apparently find dissimilation of *gʷ…gʷ to *gʷ…h₃w in almost all branches. And Lycian liked to spell its laryngeal reflex with kappa and chi (as well as a specially devised Ж-shaped letter).
Absolutely. Whether the consonants influence the vowels or vice versa depends on the size of the vowel system.
Consider Greenlandic. It has three vowel phonemes, /a i u/, and two uvular consonants /q ʁ/. What happens to /qa qi qu ʁa ʁi ʁu/? The vowels are colored, [qɑ qe qo ʁɑ ʁe ʁo], and this is even mostly spelled out (qa qe qo ra re ro).
Now compare Lakhota. It has five vowel phonemes /a e i o u/ and a uvular consonant /ʁ/. /ʁi/ can’t be rendered as [ʁe], because /ʁe/ already exists; and /ʁe/ can’t be rendered as [ʁɑ] because /ʁɑ/ already exists. What happens? /ʁa ʁe ʁo ʁu/ are not colored, and /ʁi/ comes out as [ʀi]!
In PIE, only the most [ə]-like vowel, *e, underwent coloring. The others, of which only *o was common, were resistant. Within Anatolian, however, *u was in fact colored into a new [o] as mentioned in the OP.
(BTW, your Russian example is wrong: Russian has [x], not [χ]. And while Castilian Spanish has [χ] more or less across the board, Latin America uses a lot of [x], even [h] in places like Colombia.)
More fun with vowel coloring: what happens when a large consonant system (with several uvulars) meets a rather large vowel system (with /ə/)? Interesting things happen, that’s what!
Not as strongly as I thought. In Arapaho (incidentally a language without /a/…) the automatic prothetic consonant is /h/, not the also existing /ʔ/.
Oh – this draft-like thing explains the whole lengthened grade away as predictable from synchronic PIE phonology. The plot thickens, or something.
How does that proposal not destroy the nonsingular of the reduplicated perfect? (Are CVOO roots, O = obstruent, an insignificant class?)
(or RVC)
Similar-place avoidance is supposed to be a sporadic, irregular process, as the “paper” says… Is the root expected to be in zero-grade in the nonsingular of the reduplicated perfect?
Roots ending in two obstruents are uncommon enough that the second obstruent gets termed “root extension”. I don’t know if that’s rare enough to be insignificant for this purpose.
Yeah, there’s a classic alternation in the perfect indicative between o-grade singular and zero-grade nonsingular. For instance, the perfect ‘think, remember’ of *men- goes 3sg *memóne, 3pl *memnḗr. In this case, luckily, *mn- is a licit onset (labial resonant plus another resonant seems to have been OK) so the latter form can be syllabified *me.mnḗr and survive. But if the root had been *nem-, the 3pl would have had to syllabify as *nen.mḗr, which Sandell says should lose its second *n.
Roots with *-OO are essentially restricted to *-HO (78 in my working list of LIV’s roots) and *-Oh₂ (21, of which 14 *-th₂), with some spares of *-sD (8) or *-Ks (6), so that might not end up being much of a problem. *-HH is nonexistent (or just not yet reconstructed?), *-PP is essentially nonexistent (2 examples), as are *-Oh₃ and *-Oh₁ (1 example each).