Page 4 of 4

Posted: Fri Jul 16, 2010 12:03 pm
by WeepingElf
Tropylium wrote:
WeepingElf wrote:
Tropylium wrote:BTW, for a minor thredjack, how much of the laryngeals' fate is pre-PIE and how much post-PIE? Similarly, how much of it would be common post-exo-Anatolian? I never see this explained clearly, with an implication that it's all branch-specific, but yet changes like h2e :> h2a, or eh1 :> e: appear to be shared by everything.
The three laryngeals certainly were still there after the break-off of Anatolian, as the details of their loss are different in the various branches; and they could not have merged into one either, as they yield different vowels in syllabic position in Greek.
No, that only proves that "syllabic laryngeals", whatever those were, couldn't have merged yet, and cross-linguistical evidence considered, them having been actual syllabic voiceless spirants seems almost impossible. I'm tempted to think of those three as phonetically something like [ɪ ə ʊ], and if so, there should've been no reason for them to merge simultaneously with their consonantal counterparts.

And what you alluded to is exactly what I'm asking here: to what extent are the details of their loss different, and what prevents the non-different parts then from being inherited common innovations? I think it's oversimplifying the situation to just state "*eh2 :> a:" as if that were a single step. Phonetically, this would appear to have gone thru at least three or so stages: *eh2 :> ah2 :> ah :> a:. This kind of a detailed view allows both for, say, to retain a separate laryngeal phoneme until PBS and PII (perhaps elsewhere too, but those are the two I've explictly seen to require laryngeal retention), while still avoiding having to postulate the previous steps as shared but nonherited.
Indeed, the non-syllabic laryngeals, after colouring adjacent vowels (which went the same way in all branches and thus probably predates break-up), behave all the same in non-Anatolian IE languages, so they may indeed have merged into */h/ before the break-up of non-Anatolian IE. Yet, phenomena such as the Vedic and Avestan "laryngeal hiatus" (where certain long vowels behave metrically as if they were two vowels with a consonant between), and some peculiarities of accentuation in Balto-Slavic hint at the existence of laryngeals (though not necessarily three distinct ones) at a stage when the branches had already begun to diverge. Only for the syllabic laryngeals we really have to posit distinctness after break-up, though the Greek "prothetic vowels" (vowels which reflect initial pre-consonantal laryngeals - *h1 as /e/, *h2 as /a/ and *h3 as /o/ - which are lost in most other languages) may pose a problem here.

Posted: Fri Jul 16, 2010 5:16 pm
by Soap
This is a dumb question but where did the plain /h/ in Sanskrit come from? No easily accessible resource seems to tell me. I assume it's not a direct reflection of a laryngeal because that would contradict the idea that they appear only in Anatolian.

Posted: Fri Jul 16, 2010 5:42 pm
by Morrígan
Soap wrote:This is a dumb question but where did the plain /h/ in Sanskrit come from? No easily accessible resource seems to tell me. I assume it's not a direct reflection of a laryngeal because that would contradict the idea that they appear only in Anatolian.
ǵʰ > h
gʰ,gʷʰ > h /_e,i

Posted: Fri Jul 16, 2010 6:04 pm
by Soap
d'oh. I think I knew that and just forgot. Thanks. There still is a letter for /j_h/ in 'skrit, though, so I assume it wasnt an unconditioinal shift.

Posted: Fri Jul 16, 2010 8:39 pm
by Etherman
Tropylium wrote: That only follos if PIE actually had aspirates.
True. I'm not fully convinced of it myself. In fact I lean towards the Breathy Voice Hypothesis.
It seems plausible it's a later development from original mediae, and in the branches that clearly have or had them, there either also develops a /h/ (Greek, IA), or the aspirates swiftly become something else (Italic, Iranian).

(As for what the traditional mediae were then, my money's on voiced preglottalized.)
I derive the 3 PIE series of stops from an earlier system of two series; plain voiceless and voiceless aspirate. The aspirate series probably had breathy voice as an allophonic feature.

BTW, for a minor thredjack, how much of the laryngeals' fate is pre-PIE and how much post-PIE? Similarly, how much of it would be common post-exo-Anatolian? I never see this explained clearly, with an implication that it's all branch-specific, but yet changes like h2e :> h2a, or eh1 :> e: appear to be shared by everything.
I think the vowel coloring effects are pre-PIE. The various reflexes of the laryngeals are all post-PIE. For example the laryngeals are vocalized differently across the various language (though mostly as /a/).
One reason I ask is that there are posited loanwords in Uralic that reflect initial laryngeals as *k, but these are, interestingly enuff, not supposedly from PIE (too limited in distribution for that; some exist only in Finnish), but some separate and otherwise unattested arcaic northwestern branch. Yet it seems to me initial laryngeals were lost alreddy at the exo-Anatolian stage, so that sounds quite fishy.
Seems likely that there were unattested branches of PIE, but how will we ever know for sure?

Posted: Fri Jul 16, 2010 10:57 pm
by Dewrad
Etherman wrote:
Tropylium wrote: That only follos if PIE actually had aspirates.
True. I'm not fully convinced of it myself. In fact I lean towards the Breathy Voice Hypothesis.
Surely that's not just a hypothesis, but what "voiced aspirates" actually are. A voiced consonant accompanied by a voiceless release is unattested in PIE's descendant languages, so why reconstruct it at all for the parent?

Posted: Sat Jul 17, 2010 2:26 am
by chris_notts
Dewrad wrote:
Etherman wrote:
Tropylium wrote: That only follos if PIE actually had aspirates.
True. I'm not fully convinced of it myself. In fact I lean towards the Breathy Voice Hypothesis.
Surely that's not just a hypothesis, but what "voiced aspirates" actually are. A voiced consonant accompanied by a voiceless release is unattested in PIE's descendant languages, so why reconstruct it at all for the parent?
Yes. The wikipedia article on Glottalic theory makes the same bizarre statement:

"Hopper (1973) also proposed that the aspiration that had been assumed for the voiced stops bh, dh, gh could be accounted for by a low-level phonetic feature known to phoneticians as "breathy voice." "

Surely that's a tautology. Breathy voiced means exactly the same thing as voiced aspirate.

Posted: Sat Jul 17, 2010 10:36 am
by Nortaneous
maybe they meant some sort of other nonmodal voicing that could've shifted to breathy voice in sanskrit or whatever after the formation of the voiceless aspirates?

Posted: Sat Jul 17, 2010 11:08 am
by Tropylium
Or whoever wrote that (in WP) just didn't kno what they were talking about. Frankly it sounds like a some kind of attempt to weasel oneself out from the issue of traditional PIE's typological unbalance.
Etherman wrote:

BTW, for a minor thredjack, how much of the laryngeals' fate is pre-PIE and how much post-PIE? Similarly, how much of it would be common post-exo-Anatolian? I never see this explained clearly, with an implication that it's all branch-specific, but yet changes like h2e :> h2a, or eh1 :> e: appear to be shared by everything.
I think the vowel coloring effects are pre-PIE. The various reflexes of the laryngeals are all post-PIE. For example the laryngeals are vocalized differently across the various language (though mostly as /a/).
What of common exo-Anatolian, tho? Seems that stage includes at least
-initial loss
-merger elsewhere.

(Does Tocharian do anything interesting laryngeal-wise?)
One reason I ask is that there are posited loanwords in Uralic that reflect initial laryngeals as *k, but these are, interestingly enuff, not supposedly from PIE (too limited in distribution for that; some exist only in Finnish), but some separate and otherwise unattested arcaic northwestern branch. Yet it seems to me initial laryngeals were lost alreddy at the exo-Anatolian stage, so that sounds quite fishy.
Seems likely that there were unattested branches of PIE, but how will we ever know for sure?
Well, these etymologies include stuff like tehdas "factory", supposedly from *dʰeH-tos. However, the original meaning is "site", "delimited space", "span", with the modern meaning only influenced by *teke- "to do" (which DOES seem related to *dʰeH- however, but as a loan or cognate?), and there's a competing Germanic loan origin.

Posted: Sat Jul 17, 2010 11:39 am
by Tropylium
Soap wrote:This is a dumb question but where did the plain /h/ in Sanskrit come from? No easily accessible resource seems to tell me.
For future reference:
http://en.wikipedia.org/wiki/Indo-European_sound_laws

Posted: Sat Jul 17, 2010 11:47 am
by Etherman
Dewrad wrote: Surely that's not just a hypothesis, but what "voiced aspirates" actually are. A voiced consonant accompanied by a voiceless release is unattested in PIE's descendant languages, so why reconstruct it at all for the parent?
Isn't there supposed to be some Armenian dialects with true voiced aspirates?

Posted: Sat Jul 17, 2010 11:51 am
by Etherman
chris_notts wrote: Yes. The wikipedia article on Glottalic theory makes the same bizarre statement:

"Hopper (1973) also proposed that the aspiration that had been assumed for the voiced stops bh, dh, gh could be accounted for by a low-level phonetic feature known to phoneticians as "breathy voice." "

Surely that's a tautology. Breathy voiced means exactly the same thing as voiced aspirate.
Why would you think that? A voiced aspirate is a voiced stop with aspirate release. A breathy voiced stop is like a voice stop but the vocal cords are held further apart allowing greater airflow.

Posted: Sat Jul 17, 2010 12:31 pm
by Morrígan
Etherman wrote:
chris_notts wrote: Yes. The wikipedia article on Glottalic theory makes the same bizarre statement:

"Hopper (1973) also proposed that the aspiration that had been assumed for the voiced stops bh, dh, gh could be accounted for by a low-level phonetic feature known to phoneticians as "breathy voice." "

Surely that's a tautology. Breathy voiced means exactly the same thing as voiced aspirate.
Why would you think that? A voiced aspirate is a voiced stop with aspirate release. A breathy voiced stop is like a voice stop but the vocal cords are held further apart allowing greater airflow.
No, I believe chris is right. Both phenomena are caused by high subglottalic pressure (HSGP): on a voiceless stop, this is manifested as a puff of air because of the higher pressure, but if the stop is voiced, the voiced portion is affected by the HSGP and has a different sound than modal voice.

The only possibly difference is that the "classical aspirated stop" has a later voice-onset-time than the "classical breathy-voiced stop" which is pre-voiced. This part is pure speculation.

Posted: Sat Jul 17, 2010 12:39 pm
by Tropylium
Also: an example of an actual voiced aspirate would insted be those Khoi-San /dtʰ/ thingies.

Posted: Sat Jul 17, 2010 12:55 pm
by Etherman
Tropylium wrote: What of common exo-Anatolian, tho? Seems that stage includes at least
-initial loss
-merger elsewhere.
Albanian supposedly retains initial *h4. Armenian retains initial *h2, but not consistently. Kortland explains /h/ in Anatolian, Armenian, and Albanian the same way: *h1 is lost everywhere, *h2 and *h3 are preserved as /h/ before *e but are lost before *o. It's unlikely that these three form a genetic node so they were almost certainly independent developments. The developments of sonants in Greek near laryngeals depends on the kind of laryngeal.
(Does Tocharian do anything interesting laryngeal-wise?)
Not that I know of, but I don't know that much about Tocharian.
Well, these etymologies include stuff like tehdas "factory", supposedly from *dʰeH-tos. However, the original meaning is "site", "delimited space", "span", with the modern meaning only influenced by *teke- "to do" (which DOES seem related to *dʰeH- however, but as a loan or cognate?), and there's a competing Germanic loan origin.
Well this Finnish word looks like an IE borrowing but the semantics look iffy.

Posted: Sat Jul 17, 2010 1:08 pm
by tezcatlip0ca
Soap wrote:d'oh. I think I knew that and just forgot. Thanks. There still is a letter for /j_h/ in 'skrit, though, so I assume it wasnt an unconditioinal shift.
An aspirated palatal approximant? There is no letter for that.

Posted: Sat Jul 17, 2010 1:27 pm
by Morrígan
Aid'os wrote:
Soap wrote:d'oh. I think I knew that and just forgot. Thanks. There still is a letter for /j_h/ in 'skrit, though, so I assume it wasnt an unconditioinal shift.
An aspirated palatal approximant? There is no letter for that.
I imagine he meant jh झ [ɟ͡ʝʱ], though all evidence points to this being secondary, though I can't figure out how it was derived.

Posted: Sat Jul 17, 2010 1:45 pm
by Tropylium
Etherman wrote:
Tropylium wrote: What of common exo-Anatolian, tho? Seems that stage includes at least
-initial loss
-merger elsewhere.
Albanian supposedly retains initial *h4. Armenian retains initial *h2, but not consistently.
I'm reminded here how a number of Finnish words have added an unetymological /h/ before /a/ (eg. PU *ačka "eider" :> *aška :> haahka). This kind of a change doesn't necessarily have to reflect an original consonant, and indeed, if it doesn't match up with the evidence of laryngeals elsewhere, it might as well be a partially irregular innovation.
The developments of sonants in Greek near laryngeals depends on the kind of laryngeal.
Hmm. Going by the WP table, that seems like it's related to Greek's syllabic laryngeal distinction once again. Other languages also seem to have reflexes that point to something like common *R=H= (perhaps phonetically [əRəH]), not simply *R=H.