The Great Proto-Indo-European Thread
- Hallow XIII
- Avisaru
- Posts: 846
- Joined: Sun Nov 04, 2012 3:40 pm
- Location: Under Heaven
Re: The Great Proto-Indo-European Thread
As it happens, this is precisely the reason why I was interested in Colarusso's original paper. If any sort of useful arguments in this direction are to be found, it would probably be there.
陳第 wrote:蓋時有古今,地有南北;字有更革,音有轉移,亦勢所必至。
Read all about my excellent conlangsR.Rusanov wrote:seks istiyorum
sex want-PRS-1sg
Basic Conlanging Advice
Re: The Great Proto-Indo-European Thread
Are those features of Caucasic languages? I'd have picked out pharyngeal consonants and pharyngealization.suoenatroN wrote:How can it be Caucasoid? Is there any evidence of rounding and palatalization on non-velar consonants?
But that's kind of irrelevant - we really have no evidence for what NWC or NEC languages looked like 6000 years ago, or how similar or different they are to their modern descendants. Even the pharyngealization could have been due to contact with Arabic following the spread of Islam into the region.
The most significant features shared among both eastern and western languages (at least ancestrally) is the 4-way noun-class system. It's lost in some modern languages which have had extensive contact with Turkic languages, and simply due to sound changes affecting the segments which mark class --- and of course, other languages (Chechen, most notably) has expanded to 6 classes. But my not-terribly thorough analysis leads me to think that there was originally a 4-class system in place.
The question is though, does the presence of corresponding class markers in NEC and NWC indicate common ancestry, or some kind of sprachbund or other areal feature. Without more correspondences in morphology, grammar, or lexemes, I'd have to accept the latter.
Also, I'm considering getting as many papers together as possible and starting a North Caucasian thread.
- WeepingElf
- Smeric
- Posts: 1630
- Joined: Wed Mar 08, 2006 5:00 pm
- Location: Braunschweig, Germany
- Contact:
Re: The Great Proto-Indo-European Thread
Right. AFAIK, labialization and palatalization are a NWC feature which is not typical of NEC. Though NWC would of course be the family most relevant here, as it is closest (geographically) to the presumed PIE homeland north of the Black Sea. But indeed we don't know what those languages were like 6,000 years ago.Morrígan wrote:Are those features of Caucasic languages? I'd have picked out pharyngeal consonants and pharyngealization.suoenatroN wrote:How can it be Caucasoid? Is there any evidence of rounding and palatalization on non-velar consonants?
But that's kind of irrelevant - we really have no evidence for what NWC or NEC languages looked like 6000 years ago, or how similar or different they are to their modern descendants. Even the pharyngealization could have been due to contact with Arabic following the spread of Islam into the region.
What is typical of all Caucasian languages are, phonologically, uvular stops, ejectives and at least two series of sibilant affricates and fricatives. And that's not what PIE looks like. Uvular stops perhaps existed if the "plain velars" actually were uvulars; ejectives are of course the central tenet of the glottalic theory, which is controversial; but the only sibilant of PIE was */s/.
But again, we don't know whether the Caucasian languages were already like what they are now 6,000 years ago. (We don't even know whether the ancestors of the modern Caucasian languages were spoken in the Caucasus at all at that time! Some scholars assume them to have been pushed into the Caucasus by other languages, such as the IE languages on the Pontic-Caspian steppe.)
I think it is fairly well established that Proto-NEC had four noun classes.Morrígan wrote:The most significant features shared among both eastern and western languages (at least ancestrally) is the 4-way noun-class system. It's lost in some modern languages which have had extensive contact with Turkic languages, and simply due to sound changes affecting the segments which mark class --- and of course, other languages (Chechen, most notably) has expanded to 6 classes. But my not-terribly thorough analysis leads me to think that there was originally a 4-class system in place.
I don't know how much evidence there is for "North Caucasian", but NWC and NEC resemble each other less than IE and Uralic, if you ask me. Which may not mean much, though.Morrígan wrote:The question is though, does the presence of corresponding class markers in NEC and NWC indicate common ancestry, or some kind of sprachbund or other areal feature. Without more correspondences in morphology, grammar, or lexemes, I'd have to accept the latter.
That would be interesting, even though I know only little about those languages and won't have much to contribute.Morrígan wrote:Also, I'm considering getting as many papers together as possible and starting a North Caucasian thread.
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A
- Drydic
- Smeric
- Posts: 1652
- Joined: Tue Oct 08, 2002 12:23 pm
- Location: I am a prisoner in my own mind.
- Contact:
Re: The Great Proto-Indo-European Thread
Please, DO! That would be most awesome.Morrígan wrote:Also, I'm considering getting as many papers together as possible and starting a North Caucasian thread.
Re: The Great Proto-Indo-European Thread
Ok will do, I'll get on that when I get home and can scrape together all my files. Also, I'll see if I can find the paper I wrote on patters of direct/oblique stem formation in NEC.Nessari wrote:Please, DO! That would be most awesome.Morrígan wrote:Also, I'm considering getting as many papers together as possible and starting a North Caucasian thread.
- Drydic
- Smeric
- Posts: 1652
- Joined: Tue Oct 08, 2002 12:23 pm
- Location: I am a prisoner in my own mind.
- Contact:
Re: The Great Proto-Indo-European Thread
Yaaaaaay!
*dances*
*dances*
Re: The Great Proto-Indo-European Thread
I would say PIE voiced aspirated were indeed voiced aspirates and stayed similar to that in eastern branches but changed into voiced fricatives in (the common ancestor of?) Albanian, Proto-Italo-Celtic and Proto-Germano-Balto-Slavic.
As for a connection with Proto-Uralic, I find it more reasonable for PU *tt to correspond to PIE *t and PU *t to correspond to PIE *d, since tt is more fortified than t, which itself is more fortified than d. In general, though, while the genetical relationship of PIE resembles PU more closely, it's the Caucasian languages it resembles more in terms of phonology. So it may be possible that some or even all of the ejectives in Pre-PIE were due to Caucasian contact, probably loanwords.
The laryngeals in the family words could theoretically be replaced by *@ an I have seen it as such in pre-laryngeal theory PIE reconstructions. What convinces me most of laryngeal theory is the presence in Hittite.
One other thing I'm curious about is the possibility that Gaulish -t- preterites are linked in any way to the Germanic preterite. There's obviously high likelihood that there's no connection at all but it would still be interesting to see if there is anything to it.
As for a connection with Proto-Uralic, I find it more reasonable for PU *tt to correspond to PIE *t and PU *t to correspond to PIE *d, since tt is more fortified than t, which itself is more fortified than d. In general, though, while the genetical relationship of PIE resembles PU more closely, it's the Caucasian languages it resembles more in terms of phonology. So it may be possible that some or even all of the ejectives in Pre-PIE were due to Caucasian contact, probably loanwords.
The laryngeals in the family words could theoretically be replaced by *@ an I have seen it as such in pre-laryngeal theory PIE reconstructions. What convinces me most of laryngeal theory is the presence in Hittite.
One other thing I'm curious about is the possibility that Gaulish -t- preterites are linked in any way to the Germanic preterite. There's obviously high likelihood that there's no connection at all but it would still be interesting to see if there is anything to it.
- WeepingElf
- Smeric
- Posts: 1630
- Joined: Wed Mar 08, 2006 5:00 pm
- Location: Braunschweig, Germany
- Contact:
Re: The Great Proto-Indo-European Thread
This may have been the case.jmcd wrote:I would say PIE voiced aspirated were indeed voiced aspirates and stayed similar to that in eastern branches but changed into voiced fricatives in (the common ancestor of?) Albanian, Proto-Italo-Celtic and Proto-Germano-Balto-Slavic.
What speaks against your proposed correspondence are questions of frequency and markedness. The PIE voiceless stops are the most frequent and least marked stops in the language. Uralic geminates, in contrast, are highly marked and much less common. Hence, it is very unlikely that these two classes of stops correspond to each other directly.jmcd wrote:As for a connection with Proto-Uralic, I find it more reasonable for PU *tt to correspond to PIE *t and PU *t to correspond to PIE *d, since tt is more fortified than t, which itself is more fortified than d. In general, though, while the genetical relationship of PIE resembles PU more closely, it's the Caucasian languages it resembles more in terms of phonology. So it may be possible that some or even all of the ejectives in Pre-PIE were due to Caucasian contact, probably loanwords.
But it may indeed be the case that Pre-PIE once was more similar to Uralic and was transformed by contact with Caucasian languages.
No, this does not work well. Sure, Szemerényi thought that way until his death, but he had to reintroduce voiceless aspirates and faced other problems, e. g. with the triple reflex in Greek, as well. The Greek triple reflex shows that there still were three different laryngeals at the time Late PIE broke up.jmcd wrote:The laryngeals in the family words could theoretically be replaced by *@ an I have seen it as such in pre-laryngeal theory PIE reconstructions. What convinces me most of laryngeal theory is the presence in Hittite.
I don't know where the Celtic t-preterite came from; it may indeed be a parallel development to the Germanic weak preterite, which is AFAIK not well-understood either.jmcd wrote:One other thing I'm curious about is the possibility that Gaulish -t- preterites are linked in any way to the Germanic preterite. There's obviously high likelihood that there's no connection at all but it would still be interesting to see if there is anything to it.
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A
- Salmoneus
- Sanno
- Posts: 3197
- Joined: Thu Jan 15, 2004 5:00 pm
- Location: One of the dark places of the world
Re: The Great Proto-Indo-European Thread
Another reason to think that laryngeals weren't just schwa: in some daughter, voiced+laryngeal>voiced aspirated. That's quite weird if laryngeals were vocalic. In fact, since it shows that eH and He had different reflexes, it's also pretty good evidence that there were indeed two elements not one (i.e. *eH wasn't actually just a single breathy-voiced vowel).
On the Celtic preterite: AIUI, the theory is that it's from disambiguation of the third person singular. A vowel was added to the end of it for some reason, and then the -*t- was reanalysed as a preterite marker and spread by analogy.
On the Celtic preterite: AIUI, the theory is that it's from disambiguation of the third person singular. A vowel was added to the end of it for some reason, and then the -*t- was reanalysed as a preterite marker and spread by analogy.
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]
But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!
But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!
Re: The Great Proto-Indo-European Thread
No, that doesn't follow. "A commonly goes to B" does not imply "B commonly comes from A".Salmoneus wrote:Surely the a/o mergers are good evidence for it being o? Since a/o mergers are immensely common.
Consider e.g. A B = /ʰt t/; A B = /ð d/; A B = /y i/. All of these would involve a more marked segment becoming a less marked one. In most such cases (proportional to just how rare A is in relation to B), a previous unmarked segment already existed, and was more common than the merging marked segment. This is the case for o → a as well; e.g. considering the AmEng. cot-caught merger, the LOT vowel was rather more common than the THOUGHT vowel to begin with.
(Also, how many o → a examples that AREN'T early IE langs could you come up with? Sure there are examples, but I don't think it's quite as common as you may think.)
Yet in most IE langs, the most common (or one of the most common, as is the case for Indo-Iranian) source of unmarked /a/ is precisely *o. *h₂e and vocalized laryngeals come in at #2 I think, followed by possible substratal cases of *a at #3, followed by possible original *a only at #4.
I believe that reinterpreting the unmarked vowel /a/ as a common segment would be sufficient typological payback to offset the assumptions that 1) this develops to /o/ in several branches (also common), 2) that *a was instead a "neutral but marked" vowel that everywhere eventually becomes *a (fairly typical behavior for central vowels of the *ə sort, and we're already assuming just about this for the "schwa secundum" of vocalized laryngeals), and 3) that the PIE vowel system (discounting syllabic resonants) was not the perfectly symmetric *i *e *a *o *u but the slightly less common *i *e *ə *a *u.
(The last assumption seems to find some areal support from the similar systems existing in e.g. Sumerian, Akkadian and Etruscan. Also /o/ is of course absent from NWC too, but so is /e/.)
If there was some prior stage where *h₂e + vocalized laryngeals + original *a had all been merged as, say, /ə/ before further developing to /a/ is not clear though. *h₂ seems to have been retained particularly long (long enough to postdate Proto-Indo-Iranian, since aspirated stops only turn up in Indo-Aryan) — so *h₂-coloring could well have also been a relatively late development that produced directly /a/.
It does seem hard to figure with any kind of a *e ~ *o = /ə/ ~ /a/ stage for ablaut. /e/ ~ /a/ for early PIE should be functional enough.Salmoneus wrote:The problem with that, to me, is that it seems counter-intuitive for all stressed vowels to become schwa, and unstressed vowels to shift from schwa to /a/...Hallow XIII wrote:That is another reason why I think the scenario of a Caucasoid (Pre-)PIE is rather likely. The reconstruction of two underlying vowels /e o/ that ablaut seems odd at first but if you consider that that may well have come from an earlier /ə a/ things suddenly make a lot more sense (and provide a good argument for [a] being underlying /eh2/ to boot; the original low vowel shifted to o and the vowel-laryngeal sequences shifted to create a functional five-vowel system).
[ˌʔaɪsəˈpʰɻ̊ʷoʊpɪɫ ˈʔæɫkəɦɔɫ]
- Salmoneus
- Sanno
- Posts: 3197
- Joined: Thu Jan 15, 2004 5:00 pm
- Location: One of the dark places of the world
Re: The Great Proto-Indo-European Thread
How about *e=/a/, *o=/@/?
*h2 would then simply back /a/ to /A/, and *h3 would back and round /a/ to /Q/. If /@/ then moves to be a back vowel, any backing and rounding effects on it would become invisible. You'd then have a very cluttered space, with /A/, /a/, /Q/ and /O/ (for the sake of argument). /Q/ and /O/ could merger, while /a/ moves forward to escape, ultimately to /E/, and /A/ remains as the only a-like vowel.
EDIT: the advantage of this is that it gives a pre-PIE with a neat i-a-u vowel system, with /a/ neutralising to schwa when unstressed.
Of course, this still leaves problems - if i and u are vowels, why the phonotactic constraints? What about a) roots believed to have *o in them, b) *o in the roots of certain derivatives, and c) root nouns with o/e alternation. Are these all just the result of analogies, or was there an actual additional vowel that just happened to merge with schwa later? And what about vowel lengthening in the root - is there any motivation for that?
*h2 would then simply back /a/ to /A/, and *h3 would back and round /a/ to /Q/. If /@/ then moves to be a back vowel, any backing and rounding effects on it would become invisible. You'd then have a very cluttered space, with /A/, /a/, /Q/ and /O/ (for the sake of argument). /Q/ and /O/ could merger, while /a/ moves forward to escape, ultimately to /E/, and /A/ remains as the only a-like vowel.
EDIT: the advantage of this is that it gives a pre-PIE with a neat i-a-u vowel system, with /a/ neutralising to schwa when unstressed.
Of course, this still leaves problems - if i and u are vowels, why the phonotactic constraints? What about a) roots believed to have *o in them, b) *o in the roots of certain derivatives, and c) root nouns with o/e alternation. Are these all just the result of analogies, or was there an actual additional vowel that just happened to merge with schwa later? And what about vowel lengthening in the root - is there any motivation for that?
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]
But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!
But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!
Re: The Great Proto-Indo-European Thread
Not implausible at some very distant point, but not that much for PIE proper. This doesn't seem to have any nice consequences for explaining the vowel developments in the descendants. *e as [+front] finds confirmation in every PIE branch AFAIK.Salmoneus wrote:How about *e=/a/, *o=/@/?
Most languages of the Lezgian branch at least have all sorts of labialized and sometimes palatalized consonants. Check out e.g. Archi, Lezgian and Tsakhur. Elsewhere in the family, e.g. Khwarshi has a lot of secondarily articulated stuff too.WeepingElf wrote:Right. AFAIK, labialization and palatalization are a NWC feature which is not typical of NEC.
To be contrary again, I'm starting to think that the best arguments for the glottalic theory actually come from the analysis of the voiced aspirates (if I were to release a full version of the theory, I'd rename it something like "secondary aspirate theory").WeepingElf wrote:This may have been the case.jmcd wrote:I would say PIE voiced aspirated were indeed voiced aspirates and stayed similar to that in eastern branches but changed into voiced fricatives in (the common ancestor of?) Albanian, Proto-Italo-Celtic and Proto-Germano-Balto-Slavic.
The typologically neutral direction of development appears to be B → Bʰ → Pʰ. Traditional PIE, however, required a widespread reversal of the 1st step. There's direct evidence for aspiration only in Greek and Indo-Aryan — and not even all of the latter; IIRC Dardic and some other varieties only show plain voiced stops. (Armenian has "direct evidence" for just about any hypothesis imaginable and doesn't count.) Indirectly perhaps also Italic / Venetic. but there could be other routes of explanation for those.
IA and Greek have incidentally also been in close contact with language families that already have aspirates: Greek with Tyrsenan, IA with Bodic. Coincidence?
It's also telling that there are essentially no reliable examples of *tt in Proto-Uralic. I think these should be considered in connection to the other stop+stop clusters, not as a "stop series".WeepingElf wrote:What speaks against your proposed correspondence are questions of frequency and markedness. The PIE voiceless stops are the most frequent and least marked stops in the language. Uralic geminates, in contrast, are highly marked and much less common. Hence, it is very unlikely that these two classes of stops correspond to each other directly.jmcd wrote:As for a connection with Proto-Uralic, I find it more reasonable for PU *tt to correspond to PIE *t and PU *t to correspond to PIE *d, since tt is more fortified than t, which itself is more fortified than d. In general, though, while the genetical relationship of PIE resembles PU more closely, it's the Caucasian languages it resembles more in terms of phonology. So it may be possible that some or even all of the ejectives in Pre-PIE were due to Caucasian contact, probably loanwords.
• *čk and *kt are quite common
• *tk, *čč, *pp and *pt are also fairly common
• *kk and *ćć are rarer but still well-supported
• There's spotty evidence for *tt, *ćk, *kč, *kć
A related reconstruction problem is that many of these widely collapse to single voiceless consonants. E.g. telling *tk and *čk apart from *tt and *čč is only possible if there's evidence from the languages west of Permic. The UEW for some reason defaults to *tt, but more probably these words had something else.
At any rate, root-medial two-obstruent clusters are such a major feature in Uralic that if Indo-Uralic holds, there would have to have been some similar collapse of this cluster type in pre-IE, or some innovation that created lots of them in pre-Uralic.
[ˌʔaɪsəˈpʰɻ̊ʷoʊpɪɫ ˈʔæɫkəɦɔɫ]
- WeepingElf
- Smeric
- Posts: 1630
- Joined: Wed Mar 08, 2006 5:00 pm
- Location: Braunschweig, Germany
- Contact:
Re: The Great Proto-Indo-European Thread
A number of interesting suggestions. I think the traditional reconstruction gives a very good picture of Late PIE; most attempts to reform it complicate the trajectories of the attested IE languages, usually requiring parallel innovations in a number of branches far away from each other. But Late PIE of course has a history itself, and we know so little about Anatolian phonology that Early PIE could have been almost everything, and even Early PIE did not spring fully girded from Jupiter's head but has a long prehistory.
The vowel system proposed by Tropylium may have been in place in Early PIE, but probably shifted to the more stable "classical" system on the way to Late PIE. The Uralic "geminates" may indeed have just been stop+stop clusters, quite common in Uralic, whose members just happened to have the same POA. If Indo-Uralic is real, it must indeed be found out what their IE cognates are - the only stop+stop clusters in PIE are AFAIK the "thorn" clusters of dental+velar; I prefer calling them "harmonic clusters" after a similar phenomenon in Kartvelian.
Kartvelian, BTW, has some traits that are quite reminiscent of Late PIE (such as a similar, but by far not identical, ablaut system; and a verb aspect system that makes similar distinctions as the Late PIE one); yet, there is little evidence in favour of an IE-Kartvelian relationship, so the similarities between Late PIE and Kartvelian are most likely due to contact, especially as Early PIE seems to have been less similar to Kartvelian (for instance, it did not yet have the Kartvelian-like verb aspect system).
The vowel system proposed by Tropylium may have been in place in Early PIE, but probably shifted to the more stable "classical" system on the way to Late PIE. The Uralic "geminates" may indeed have just been stop+stop clusters, quite common in Uralic, whose members just happened to have the same POA. If Indo-Uralic is real, it must indeed be found out what their IE cognates are - the only stop+stop clusters in PIE are AFAIK the "thorn" clusters of dental+velar; I prefer calling them "harmonic clusters" after a similar phenomenon in Kartvelian.
Kartvelian, BTW, has some traits that are quite reminiscent of Late PIE (such as a similar, but by far not identical, ablaut system; and a verb aspect system that makes similar distinctions as the Late PIE one); yet, there is little evidence in favour of an IE-Kartvelian relationship, so the similarities between Late PIE and Kartvelian are most likely due to contact, especially as Early PIE seems to have been less similar to Kartvelian (for instance, it did not yet have the Kartvelian-like verb aspect system).
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A
Re: The Great Proto-Indo-European Thread
Another specific indication that *o might not have been rounded is the de-labialization of the velars *kʷ *gʷʰ *gʷ only occurs with *u and not *o.
Despite this, I did have *o affect the labiovelars in Kuma-Koban, but it didn't sit completely right with me, and I may have to go back on this.
Despite this, I did have *o affect the labiovelars in Kuma-Koban, but it didn't sit completely right with me, and I may have to go back on this.
Re: The Great Proto-Indo-European Thread
If by "Late PIE" you mean "last common ancestor of anything non-Anatolian (and non-Tocharian?)", I don't think that would be a very good idea. It would just entail reconstructing a history *a → *o → *a for several separate branches.WeepingElf wrote:I think the traditional reconstruction gives a very good picture of Late PIE; most attempts to reform it complicate the trajectories of the attested IE languages, usually requiring parallel innovations in a number of branches far away from each other. But Late PIE of course has a history itself, and we know so little about Anatolian phonology that Early PIE could have been almost everything, and even Early PIE did not spring fully girded from Jupiter's head but has a long prehistory.
The vowel system proposed by Tropylium may have been in place in Early PIE, but probably shifted to the more stable "classical" system on the way to Late PIE.
If you were to admit an *e *ə *a system for early PIE, I think it follows that *ə *a → *a *o would be best considered a development affecting just the southwestern Celtic-Italian-Greek area, not everyplace. I'm not sure what "parallel innovations in branches far away" would be involved here?
*ā → *ō would also extend to at least Balto-Slavic and Armenian☆ (not unexpected; long vowels rise more commonly than short ones), which seems to imply in this model a late-ish date for this shift, later than the vocalization of laryngeals.
(☆ Possibly Albanian too: traditional *ō → Proto-Albanian ē while *ē → PAlb *ō, which ok could be routed by an early backing of *ē and a raising of *ā to *ē rather than *ō. But Albanian also has *ū → y, which I guess suggests that instead the route was *ō → *ø̄ → *ē?)
This could also be identified with the change *ā → *ō in Germanic, if *ə̄ → *ā had already happened by this point (i.e. the change would've spread areally to Germanic fairly late). It's worth noting too that this change happened in Finnic and Samic as well around the same timeframe, so that's confirmation that a change ā → *ō existed around the time and was spreading areally.
There are similar readjustments possible for most other PIE reinterpretations too. Consider e.g. the velar-labiovelar-uvular system: this assumption, even if it were initially based in pre-PIE, leads to an analysis where no Satemic fronting (or Satem neutralizations of the *ḱ → *k type) actually needs to have happened anywhere in the "Centum" area.
[ˌʔaɪsəˈpʰɻ̊ʷoʊpɪɫ ˈʔæɫkəɦɔɫ]
Re: The Great Proto-Indo-European Thread
... u srs?jmcd wrote:I would say PIE voiced aspirated were indeed voiced aspirates and stayed similar to that in eastern branches but changed into voiced fricatives in .... Proto-Germano-Balto-Slavic.
Slava, čĭstŭ, hrabrostĭ!
Re: The Great Proto-Indo-European Thread
BTW, I should mention the vowel system reanalysis topic is not just something I came up out of the blue. Reconstructing *o as /a/ at an earlier date (perhaps phonetically [aˑ]) has been supported by Martin Kümmel recently, in a very interesting article Typology and Reconstruction of PIE.
(He has some interesting remarks in defense of a "compromise glottalic theory" as well.)
(He has some interesting remarks in defense of a "compromise glottalic theory" as well.)
[ˌʔaɪsəˈpʰɻ̊ʷoʊpɪɫ ˈʔæɫkəɦɔɫ]
- Drydic
- Smeric
- Posts: 1652
- Joined: Tue Oct 08, 2002 12:23 pm
- Location: I am a prisoner in my own mind.
- Contact:
Re: The Great Proto-Indo-European Thread
In the original contextR.Rusanov wrote:... u srs?jmcd wrote:I would say PIE voiced aspirated were indeed voiced aspirates and stayed similar to that in eastern branches but changed into voiced fricatives in .... Proto-Germano-Balto-Slavic.
it works. Now go scream at some turks or something.jmcd wrote:I would say PIE voiced aspirated were indeed voiced aspirates and stayed similar to that in eastern branches but changed into voiced fricatives in (the common ancestor of?) Albanian, Proto-Italo-Celtic and Proto-Germano-Balto-Slavic.
Re: The Great Proto-Indo-European Thread
Changed to voiced fricatives then magically reverted to stops again in all three of Germanic, Baltic, and Slavic.
Because sound changes are real fond of reversing themselves, especially in typologically unlikely ways
Because sound changes are real fond of reversing themselves, especially in typologically unlikely ways
Slava, čĭstŭ, hrabrostĭ!
- Drydic
- Smeric
- Posts: 1652
- Joined: Tue Oct 08, 2002 12:23 pm
- Location: I am a prisoner in my own mind.
- Contact:
Re: The Great Proto-Indo-European Thread
You apparently don't know that every Germanic language shows signs of the voiced stops being ambiguous between voiced stops and fricatives (à la modern Spanish) in their histories, and it is the state which is reconstructed for Proto-Germanic.R.Rusanov wrote:Changed to voiced fricatives then magically reverted to stops again in all three of Germanic, Baltic, and Slavic.
Because sound changes are real fond of reversing themselves, especially in typologically unlikely ways
Re: The Great Proto-Indo-European Thread
In intervocalic position.
b>v / V_V is a common change
v>b / #_, _#, _C, C_ isn't.
And Baltic? Slavic? Your reasoning is so west-centric it's not even funny.
b>v / V_V is a common change
v>b / #_, _#, _C, C_ isn't.
And Baltic? Slavic? Your reasoning is so west-centric it's not even funny.
Slava, čĭstŭ, hrabrostĭ!
- WeepingElf
- Smeric
- Posts: 1630
- Joined: Wed Mar 08, 2006 5:00 pm
- Location: Braunschweig, Germany
- Contact:
Re: The Great Proto-Indo-European Thread
By "Late PIE", I indeed mean the last common ancestor of anything non-Anatolian. Which may be a valid concept or not. The kind of dialectal division you mention makes sense. The traditional vowel system is preserved in Italo-Celtic (which may be a real grouping) and Helleno-Armenian (which may also be a real grouping); these may have formed a southwestern dialect area, so a common innovation in these branches is not out of the question. There indeed is no evidence of "classical" *o having been rounded in any other IE branch. It merged with *a in Germanic, Balto-Slavic and Albanian (and of course Indo-Iranian, and Anatolian, too), and Tocharian, as is well-known, did so much batfuckery with its vowels that we cannot say anything about their antecedents.Tropylium wrote:If by "Late PIE" you mean "last common ancestor of anything non-Anatolian (and non-Tocharian?)", I don't think that would be a very good idea. It would just entail reconstructing a history *a → *o → *a for several separate branches.WeepingElf wrote:I think the traditional reconstruction gives a very good picture of Late PIE; most attempts to reform it complicate the trajectories of the attested IE languages, usually requiring parallel innovations in a number of branches far away from each other. But Late PIE of course has a history itself, and we know so little about Anatolian phonology that Early PIE could have been almost everything, and even Early PIE did not spring fully girded from Jupiter's head but has a long prehistory.
The vowel system proposed by Tropylium may have been in place in Early PIE, but probably shifted to the more stable "classical" system on the way to Late PIE.
If you were to admit an *e *ə *a system for early PIE, I think it follows that *ə *a → *a *o would be best considered a development affecting just the southwestern Celtic-Italian-Greek area, not everyplace. I'm not sure what "parallel innovations in branches far away" would be involved here?
Yes.Tropylium wrote:*ā → *ō would also extend to at least Balto-Slavic and Armenian☆ (not unexpected; long vowels rise more commonly than short ones), which seems to imply in this model a late-ish date for this shift, later than the vocalization of laryngeals.
(☆ Possibly Albanian too: traditional *ō → Proto-Albanian ē while *ē → PAlb *ō, which ok could be routed by an early backing of *ē and a raising of *ā to *ē rather than *ō. But Albanian also has *ū → y, which I guess suggests that instead the route was *ō → *ø̄ → *ē?)
This could also be identified with the change *ā → *ō in Germanic, if *ə̄ → *ā had already happened by this point (i.e. the change would've spread areally to Germanic fairly late). It's worth noting too that this change happened in Finnic and Samic as well around the same timeframe, so that's confirmation that a change ā → *ō existed around the time and was spreading areally.
Indeed, I consider it likely that the *ḱ-series were just plain velars, the *k-series were uvulars, and the *kʷ-series labialized uvulars. This doesn't really break anything in the framework of IE sound changes. Uvulars merging with velars are AFAIK commonplace enough to occur in a number of different IE branches.Tropylium wrote:There are similar readjustments possible for most other PIE reinterpretations too. Consider e.g. the velar-labiovelar-uvular system: this assumption, even if it were initially based in pre-PIE, leads to an analysis where no Satemic fronting (or Satem neutralizations of the *ḱ → *k type) actually needs to have happened anywhere in the "Centum" area.
...brought to you by the Weeping Elf
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A
Tha cvastam émi cvastam santham amal phelsa. -- Friedrich Schiller
ESTAR-3SG:P human-OBJ only human-OBJ true-OBJ REL-LOC play-3SG:A
- Drydic
- Smeric
- Posts: 1652
- Joined: Tue Oct 08, 2002 12:23 pm
- Location: I am a prisoner in my own mind.
- Contact:
Re: The Great Proto-Indo-European Thread
I never said v>b / #_, _#, _C, C_.R.Rusanov wrote:In intervocalic position.
b>v / V_V is a common change
v>b / #_, _#, _C, C_ isn't.
And Baltic? Slavic? Your reasoning is so west-centric it's not even funny.
- KathTheDragon
- Smeric
- Posts: 2139
- Joined: Thu Apr 25, 2013 4:48 am
- Location: Brittania
Re: The Great Proto-Indo-European Thread
Following what has been said about *o, is it then possible that it was something like [ʌ]?
Re: The Great Proto-Indo-European Thread
This kind of started out as a possible explanation for the origin of ablaut in PIE (no idea how plausible it might be IRL) but even if it's nowhere near close, I might end up using it in a possible future conlang:
Original Paradigm
Three phonemic vowels /i u a/ which appear as [e o ə] in unstressed syllables.
Diphthong Reduction
Schwa in a diphthong is dropped, leaving /ai au ia ua/ in alternation with unstressed . A similar change also occurs before resonants.
then lower before syllable-final resonants.
Laryngeal Colouring
Laryngeal colouring affecting allophones rather than phonemes, but only non-high allophones. /h/ caused no colouring, /x/ caused the vowels to back or lower (didn't affect back vowels) while /xʷ/ caused vowels to back and round (didn't affect back rounded vowels).
The same colouring affected the [a] and [ə] of diphthongs.
Laryngeals after vowels caused the same colouring, but caused the vowel to lengthen, thus, for example [exʷ] > [ɵ:xʷ]. If a vowel was preceded and followed by a laryngeal, the preceding one would colour the vowel while the following one would lengthen it, e.g. [xah] > [xɑh] > [xɑ:h]
Movement of Allophones
The resulting allophones at this stage are
[ʌ] merges into [o].
[ɒ] in a similar manner to [a], becomes [ɔ] before merging into [o].
Stressed [ə] merges into [o], except after /x/ when it instead becomes [a], merging into [ɑ] and stressec [ɵ] merges into [o] afterwards. Unstressed mid vowels drop before resonants.
The following alternations occur ([stressed]~[unstressed]
While before, [e], [o] and [ə] were limited to unstressed syllables and , and [a] to stressed syllables, the distribution has changed in such a way that and are now rare in stressed syllables, while [e] and [o] are common in all syllable types. Analogy could kick in, forcing stress off of and and onto a neighbouring vowel, causing and to be re-analysed as unstressed [Vi] and [Vu].
Original Paradigm
Three phonemic vowels /i u a/ which appear as [e o ə] in unstressed syllables.
Code: Select all
[i]~[e] [u~o]
[a]~[ə]
[ai]~[əi] [au]~[əu]
Diphthong Reduction
Schwa in a diphthong is dropped, leaving /ai au ia ua/ in alternation with unstressed . A similar change also occurs before resonants.
then lower before syllable-final resonants.
Laryngeal Colouring
Laryngeal colouring affecting allophones rather than phonemes, but only non-high allophones. /h/ caused no colouring, /x/ caused the vowels to back or lower (didn't affect back vowels) while /xʷ/ caused vowels to back and round (didn't affect back rounded vowels).
Code: Select all
[xe] > [xə]
[xo] > [xo]
[xa] > [xɑ]
[xə] > [xʌ]
[xʷe] > [xʷɵ]
[xʷo] > [xʷo]
[xʷa] > [xʷɒ]
[xʷə] > [xʷo]
Laryngeals after vowels caused the same colouring, but caused the vowel to lengthen, thus, for example [exʷ] > [ɵ:xʷ]. If a vowel was preceded and followed by a laryngeal, the preceding one would colour the vowel while the following one would lengthen it, e.g. [xah] > [xɑh] > [xɑ:h]
Movement of Allophones
The resulting allophones at this stage are
[a] initially moves forwards to [æ] before merging into [e].[(H)i (H)u]
[(x/h/0)e (x/h/0)ə (xʷ)ɵ (x/xʷ/h/0)o (x)ʌ]
[(h/0)a (x)ɑ (xʷ)ɒ]
[ʌ] merges into [o].
[ɒ] in a similar manner to [a], becomes [ɔ] before merging into [o].
Stressed [ə] merges into [o], except after /x/ when it instead becomes [a], merging into [ɑ] and stressec [ɵ] merges into [o] afterwards. Unstressed mid vowels drop before resonants.
The following alternations occur ([stressed]~[unstressed]
Code: Select all
[i] ~ [e] > [i] ~ [e] > [i] ~ [e]
[eR] ~ [eR] > [eR] ~ [eR] > [eR] ~ [eR]
[u] ~ [o] > [u] ~ [o] > [u] ~ [o]
[oR] ~ [oR] > [oR] ~ [oR] > [oR] ~ [oR]
[a] ~ [ə] > [a] ~ [ə] > [e] ~ [o]
[aR] ~ [əR] > [aR] ~ [R] > [e] ~ [R]
[ai] ~ [i] > [ai] ~ [i] > [ei] ~ [i]
[au] ~ [u] > [au] ~ [u] > [eu] ~ [u]
[hi] ~ [he] > [hi] ~ [he] > [hi] ~ [he]
[heR] ~ [heR] > [heR] ~ [heR] > [heR] ~ [heR]
[hu] ~ [ho] > [hu] ~ [ho] > [hu] ~ [ho]
[hoR] ~ [hoR] > [hoR] ~ [hoR] > [hoR] ~ [hoR]
[ha] ~ [hə] > [ha] ~ [hə] > [he] ~ [ho]
[haR] ~ [həR] > [haR] ~ [hR] > [he] ~ [hR]
[hai] ~ [hi] > [hai] ~ [hi] > [hei] ~ [hi]
[hau] ~ [hu] > [hau] ~ [hu] > [heu] ~ [hu]
[xi] ~ [xe] > [xi] ~ [xə] > [xi] ~ [xɑ]
[xeR] ~ [xeR] > [xəR] ~ [xəR] > [xɑR] ~ [xR]
[xu] ~ [xo] > [xu] ~ [xo] > [xu] ~ [xo]
[xoR] ~ [xoR] > [xoR] ~ [xoR] > [xoR] ~ [xoR]
[xa] ~ [xə] > [xɑ] ~ [xʌ] > [xɑ] ~ [xo]
[xaR] ~ [xəR] > [xɑR] ~ [xR] > [xɑ] ~ [xR]
[xai] ~ [xi] > [xɑi] ~ [xi] > [xɑi] ~ [xi]
[xau] ~ [xu] > [xɑu] ~ [u] > [xɑu] ~ [xu]
[xʷi] ~ [xʷe] > [xʷi] ~ [xʷɵ] > [xʷi] ~ [xʷo]
[xʷeR] ~ [xʷeR] > [xʷɵR] ~ [xʷɵR] > [xʷoR] ~ [xʷR]
[xʷu] ~ [xʷo] > [xʷu] ~ [xʷo] > [xʷu] ~ [xʷo]
[xʷoR] ~ [xʷoR] > [xʷoR] ~ [xʷoR] > [xʷoR] ~ [xʷoR]
[xʷa] ~ [xʷə] > [xʷɒ] ~ [xʷo] > [xʷo] ~ [xʷo]
[xʷaR] ~ [xʷəR] > [xʷɒR] ~ [xʷR] > [xʷo] ~ [xʷR]
[xʷai] ~ [xʷi] > [xʷɒi] ~ [xʷi] > [xʷoi] ~ [xʷi]
[xʷau] ~ [xʷu] > [xʷɒu] ~ [xʷu] > [xʷou] ~ [xʷu]
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.