So, not to sound like a crank, but... I find a lot of details about reconstructed PIE a little hard to believe

[Travis B.]

The comparative method is the best we've got, and all this talk about substrata and adstrata and Wanderwörter is essentially pulling things out of thin air without even a pretense of attempting to be scientific about it. You have yet to give anyone any reason to believe that they have anywhere near as strong a basis as the reconstruction of PIE via the comparative method. Therefore, it is safe to say that they are best regarded as crankery at this point.

[WeepingElf]

The comparative method is the best we've got, and all this talk about substrata and adstrata and Wanderwörter is essentially pulling things out of thin air without even a pretense of attempting to be scientific about it. You have yet to give anyone any reason to believe that they have anywhere near as strong a basis as the reconstruction of PIE via the comparative method. Therefore, it is safe to say that they are best regarded as crankery at this point.

While things like Wanderwörter, substrata and adstrata exist, the burden of proof lies on who invokes them to explain something. It is quite instructive that after a boom in the early 20th century, substratum theory has pretty much fallen out of favour in Romance historical linguistics because the observed isoglosses in the Romance dialect continuum bear little resemblance to the pre-Roman linguistic landscape (as far as we know it) of Western Europe. The Romance languages owe only very little to languages such as Gaulish, Iberian or Etruscan. So why should substratum theory work any better on the Indo-European level, where we know virtually nothing about the pre-IE linguistic landscape?

[Travis B.]

I wonder if, two thousand years from now, future linguists studying the Anglic varieties of North America will come upon persons who put forward substrata and adstrata from Native American languages as significant contributors to said Anglic varieties, and who will point at placenames as evidence for said substrata and adstrata.

[Talskubilos]

The comparative method is the best we've got, and all this talk about substrata and adstrata and Wanderwörter is essentially pulling things out of thin air without even a pretense of attempting to be scientific about it. You have yet to give anyone any reason to believe that they have anywhere near as strong a basis as the reconstruction of PIE via the comparative method. Therefore, it is safe to say that they are best regarded as crankery at this point.

[bradrn]

[Talskubilos]

While things like Wanderwörter, substrata and adstrata exist, the burden of proof lies on who invokes them to explain something. It is quite instructive that after a boom in the early 20th century, substratum theory has pretty much fallen out of favour in Romance historical linguistics because the observed isoglosses in the Romance dialect continuum bear little resemblance to the pre-Roman linguistic landscape (as far as we know it) of Western Europe. The Romance languages owe only very little to languages such as Gaulish, Iberian or Etruscan. So why should substratum theory work any better on the Indo-European level, where we know virtually nothing about the pre-IE linguistic landscape?

Don't forget Latin itself has a significant number of loanwords from Etruscan and Gaulish, especially the former. But I'm afraid IE isn't anything like Romance languages, nor "PIE" is to Latin.

[keenir]

So why should substratum theory work any better on the Indo-European level, where we know virtually nothing about the pre-IE linguistic landscape?

Don't forget Latin itself has a significant number of loanwords from Etruscan and Gaulish, especially the former. But I'm afraid IE isn't anything like Romance languages, nor "PIE" is to Latin.

is that an argument for or against what you've been claiming?

[Travis B.]

The comparative method is the best we've got, and all this talk about substrata and adstrata and Wanderwörter is essentially pulling things out of thin air without even a pretense of attempting to be scientific about it. You have yet to give anyone any reason to believe that they have anywhere near as strong a basis as the reconstruction of PIE via the comparative method. Therefore, it is safe to say that they are best regarded as crankery at this point.

[snip]

What? Me saying that rejecting the comparative method for pulling things out of thin air for supposed substrata, adstrata, and Wanderwörter is crankery is somehow an example of tu quoque?

[Frislander]

PIE certainly had a phonetic contrast between *e *a *o, since laryngeal colouring preceded Anatolian's departure from the family, and hence was a feature of PIE. The convention of not writing it is both anachronistic and for morphonological reasons. Pre-PIE, on the other hand, if native PIE *a always comes from laryngeal colouring, almost certainly had *e and *o as low vowels, probably *æ and *ɑ (or *ɑ̄, which lets you further reconstruct *a *ā). This does support seeing *h₂ as just blocking the raising *æ > *e, incidentally.

Doesn't *e look like it was schwa-like? The kʷetwóres rule makes more sense if it was *kʷətwáres than *kʷætwɑ́ræs.

Or it could be *kʷatwā́ras.

This kinda gets into chicken-and-egg situations here, because sure *e *o could be a case of quality arising from quantity (which is the usual development at least in modern IE languages), but it could easily be the other way around, at least in the *ə *a reconstruction, with the Brugmannian long ā from *o in open syllables being the result of spontaneous lengthening of low vowels (which is a very common development in Uralic, and part of the reason the vowels move about so much in that family). Indeed, I would argue in favour of this development to be honest, as otherwise we don't appear to find any kind of evidence for *o being longer than *e in PIE, unless there's something in Anatolian that I'm ignorant of, so I'd be tempted instead to propose that this is an Indo-Iranian innovation, possibly even a result of the same early Uralic contact that saw a significant quantity of either II or Pre-II lexical items go in the other direction into Uralic.

[KathTheDragon]

unless there's something in Anatolian that I'm ignorant of

Kümmel provides all the evidence I'm aware of here.

[Vijay]

What? Me saying that rejecting the comparative method for pulling things out of thin air for supposed substrata, adstrata, and Wanderwörter is crankery is somehow an example of tu quoque?

(Okay, really late, but) I think Talskubilos is basically saying, "I think the comparative method is crankery, but you can't refute me! You're just saying my views are also crankery! Tu quoque! That's not a logical defense!" (In case it wasn't clear: Yes, it is).

[Moose-tache]

I’ve always been sceptical of the glottalic hypothesis, but somehow I managed to miss Clackson’s reworking of the theory using phonation (i.e. variable glottal restriction, but not full closure). While I still think the traditional system of stops is the best at describing the daughter languages, phonation is a really plausible explanation for what came immediately before and how PIE ended up (however briefly) with such an unusual system.

In the first stage, late pre-PIE, let’s imagine the primary distinction between stops is one of phonation. It could be a creaky-breathy, modal-breathy, or creaky-modal phonetic realization, but for now let’s just call it tense-lax. The lax stops were distinguished from one another by a secondary property that would later be realized as voicing, so let’s tentatively call it voice. This gives us the following arrangement:

Thorn: lax (modal/breathy?), secondarily voiceless
Tyr: tense (creaky/modal?), voice-neutral
Dag: lax (modal/breathy?), secondarily voiced

From here we only need one big change to get to the T-D-Dh system of the traditional reconstruction. The tense-lax distinction becomes inaudible in voiceless stops. This means that Thorn and the voiceless allophone of Tyr would be indistinguishable, and Tyr would only appear as phonetically voiced from then on. Since now only the voiced stops have a phonation distinction, we can say that voicing is the primary distinction among stops in the next stage.

Thorn: voiceless, phonation-neutral
Tyr: voiced, secondarily tense (creaky/modal?)
Dag: voiced, secondarily lax (modal/breathy?)

In functional terms, this is basically indistinguishable from the T-D-Dh setup. The phonation could be exaggerated until Dag becomes fully aspirated, as in Hellenic or Italic. Or it could be lost, causing Dag to merge with Tyr in Celtic or Slavic, unless a chain shift keeps Tyr out of reach as in Germanic or Armenian.

The benefits of this model are clear. It provides a plausible route to the T-D-Dh system, since phonation distinctions on voiced stops that are not mirrored by voiceless stops is not unusual in languages that have phonemic phonation contrasts. It explains why Tyr sometimes behaves like a glottalized consonant but ends up as an ordinary voiced stop in most daughter languages, without resorting to absurd sound changes like /t’/ > /d/. It also addresses most of the weaknesses of the glottalic hypothesis. For example, the Germanic and Armenian chain shifts are innovations, not ancestral features.

What do people think about the glottalic-pre-PIE but non-glattolic-PIE scenario?

EDIT The following two posts are the result of a terminological confusion rectified below, and may be ignored.

[bradrn]

Thorn: lax (modal/breathy?), secondarily voiceless
Tyr: tense (creaky/modal?), voice-neutral
Dag: lax (modal/breathy?), secondarily voiced

I’m pretty sure this is physically impossible. Voicing is phonation — you can arrange it on a scale roughly like so:

[glottis completely closed] glottal stop — creaky voice — stiff voice — modal voice (‘voiced’) — slack voice — breathy voice — voiceless [glottis completely open]

Due to this, it’s impossible to have something which is *‘breathy and secondarily voiceless’ or *‘breathy and secondarily voiced’, as these would need to be in two places on the glottal closure continuum at the same time.

I suppose you could maybe rescue this by giving pre-PIE a three-way contrast between glottalised/slack/stiff consonants (like Javanese with an extra series), where the slack consonants have secondary aspiration; then merge slack/stiff consonants into modal voicing and make the glottalised series voiceless. But then again, how likely is it for glottalised consonants (I’m thinking ejectives here) to become voiceless? They’re on opposite sides of the continuum, after all.

[Travis B.]

I’m pretty sure this is physically impossible. Voicing is phonation — you can arrange it on a scale roughly like so:

[glottis completely closed] glottal stop — creaky voice — stiff voice — modal voice (‘voiced’) — slack voice — breathy voice — voiceless [glottis completely open]

Due to this, it’s impossible to have something which is *‘breathy and secondarily voiceless’ or *‘breathy and secondarily voiced’, as these would need to be in two places on the glottal closure continuum at the same time.

How does this work with, say, English or languages with ejectives, where glottal closure is associated with voiceless consonants, not voiced consonants?

[Moose-tache]

I’m pretty sure this is physically impossible.

OK, this is my fault. You're confusing phonation on the consonant with phonation on the vowel, and I can see why. I am refering to phonation and voicing as phonemic characteristics of the consonant, not necessarily simultaneous phonetic realizations on the same segment. For example, in some Austroasiatic languages there are voiced consonants that are "creaky" and voiced consoants that are "breathy." How is this possible? The voicing refers to the consonant, and the phonation refers to the effect the consonant has on the vowel.

So when I say "creaky, secondarily voiced" what I mean is something like /d/, but the following vowel experiences a partial constriction of the glottis.

This type of terminology is not uncommon, but in retrospect I can see why someone would find it confusing. Sorry about that.

[bradrn]

I’m pretty sure this is physically impossible. Voicing is phonation — you can arrange it on a scale roughly like so:

[glottis completely closed] glottal stop — creaky voice — stiff voice — modal voice (‘voiced’) — slack voice — breathy voice — voiceless [glottis completely open]

Due to this, it’s impossible to have something which is *‘breathy and secondarily voiceless’ or *‘breathy and secondarily voiced’, as these would need to be in two places on the glottal closure continuum at the same time.

How does this work with, say, English or languages with ejectives, where glottal closure is associated with voiceless consonants, not voiced consonants?

This has always been a bit confusing to me — I said as much at the end of my post. But at least the fact that phonation can be arranged along a glottal closure scale is basic and uncontroversial articulatory phonetics.

I’m pretty sure this is physically impossible.

OK, this is my fault. You're confusing phonation on the consonant with phonation on the vowel, and I can see why. I am refering to phonation and voicing as phonemic characteristics of the consonant, not necessarily simultaneous phonetic realizations on the same segment. For example, in some Austroasiatic languages there are voiced consonants that are "creaky" and voiced consoants that are "breathy." How is this possible? The voicing refers to the consonant, and the phonation refers to the effect the consonant has on the vowel.

I thought that in SE Asian languages, descriptors like ‘creaky’ and ‘breathy’ referred to the consonant itself, with the phonation extending to the following vowel: /d̰a d̤a/ → [d̰a̰ d̤a̤]. But I do get what you’re saying now.

[Skookum]

But then again, how likely is it for glottalised consonants (I’m thinking ejectives here) to become voiceless? They’re on opposite sides of the continuum, after all.

Just speculation, but I don't think it's that unlikely from an acoustic point of view. In some languages ejectives are weak and can be really difficult to distinguish from plain voiceless stops. But I'm not sure how plausible that is in the context of your hypothetical of a glottalized/slack/stiff contrast in PIE. I feel like I would expect the glottalized and stiff series to merge as voiceless with slack going voiced, but I have no experience with languages with slack/stiff phonation contrasts.

[bradrn]

But then again, how likely is it for glottalised consonants (I’m thinking ejectives here) to become voiceless? They’re on opposite sides of the continuum, after all.

Just speculation, but I don't think it's that unlikely from an acoustic point of view. In some languages ejectives are weak and can be really difficult to distinguish from plain voiceless stops.

Well, yes, this is exactly what is confusing me so.

…but then again, on the other hand, maybe acoustics rather than articulation is the explanation here. Because even though the left and right sides of the hierarchy are very far apart articulatorily, they share the acoustic property of having no glottal vibration — of course, glottal vibration reaches a maximum at modal voicing (the ‘sweet spot’).

But I'm not sure how plausible that is in the context of your hypothetical of a glottalized/slack/stiff contrast in PIE. I feel like I would expect the glottalized and stiff series to merge as voiceless with slack going voiced, but I have no experience with languages with slack/stiff phonation contrasts.

I don’t have too much experience here either. But I don’t necessarily think my merger is any less probable than yours — it would all depend on whether the glottalised series becomes voiceless before the merger or after the merger.

[WeepingElf]

Thorn: lax (modal/breathy?), secondarily voiceless
Tyr: tense (creaky/modal?), voice-neutral
Dag: lax (modal/breathy?), secondarily voiced

This is similar to what I fancy:

"Thorn": [+breath][-voice]
"Tyr": [-breath][±voice]
"Dag": [+breath][+voice]

[Creyeditor]

I was thinking about a grand compromise, but it's a bit tongue-in-cheek:

Thorn: [-voiced,-glottalized]
Tyr: [-aspirated,-long]
Dag: [-glottalized,-long]