The Great Proto-Indo-European Thread

Discussion of natural languages, or language in general.
User avatar
Tropylium
Avisaru
Avisaru
Posts: 512
Joined: Sun Aug 07, 2005 1:13 pm
Location: Halfway to Hyperborea

Re: The Great Proto-Indo-European Thread

Post by Tropylium »

We've been throwing around the idea that PIE *h₂w was actually a single phoneme *h₂ʷ (maybe a more consistent notation would be to call it either *ḫʷ or *h₄). Turns out that this has indeed already been proposed! I've just run into Edgar Suter (2014), Das hethitische Phonem /xʷ/. He also reconstructs 'to wash' as √lexʷ- (LIV: *lewh₃-).
[ˌʔaɪsəˈpʰɻ̊ʷoʊpɪɫ ˈʔæɫkəɦɔɫ]

User avatar
WeepingElf
Smeric
Smeric
Posts: 1630
Joined: Wed Mar 08, 2006 5:00 pm
Location: Braunschweig, Germany
Contact:

Re: The Great Proto-Indo-European Thread

Post by WeepingElf »

Tropylium wrote:We've been throwing around the idea that PIE *h₂w was actually a single phoneme *h₂ʷ (maybe a more consistent notation would be to call it either *ḫʷ or *h₄). Turns out that this has indeed already been proposed! I've just run into Edgar Suter (2014), Das hethitische Phonem /xʷ/. He also reconstructs 'to wash' as √lexʷ- (LIV: *lewh₃-).
Interesting. So the labialized counterpart of *h2 would not have been *h3, but a phoneme *xw (or "*h2w") that merged with *w in Nuclear IE. This may be behind the *wl- and *wr- clusters of Nuclear IE, where we have been suspecting that something was going on all the time.
...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

Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Re: The Great Proto-Indo-European Thread

Post by Travis B. »

That still leaves the question of what happened to all those instances of *b, that has been previously hypothesized as having turned into *w. If many or most of the extraneous *w came from *xw and not *b, where did all the *b go to, or why did they never exist in the first place?
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: The Great Proto-Indo-European Thread

Post by KathTheDragon »

I've long supposed that *pibh₃eti can be alternately explained by initial devoicing of *b, rather than voicing of *p, so that could be employed if it stands up to scrutiny.

User avatar
Soap
Smeric
Smeric
Posts: 1228
Joined: Sun Feb 16, 2003 2:57 pm
Location: Scattered disc
Contact:

Re: The Great Proto-Indo-European Thread

Post by Soap »

KathTheDragon wrote:I've long supposed that *pibh₃eti can be alternately explained by initial devoicing of *b, rather than voicing of *p, so that could be employed if it stands up to scrutiny.
*looks at signature*
Well, that would explain, at least a little bit, why the word for "drink" shows up with diuble /p/ in some Italic languages but double /b/ in others ... including Latin. But if you're saying that all initial /b/ shifted to /p/, how would explain the non-initial ones? Or are there none? You have an interesting idea, but I'm not sure I believe it, especially given that all of the other voiced stops were apparently immune to this shift.
Sunàqʷa the Sea Lamprey says:
Image

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: The Great Proto-Indo-European Thread

Post by KathTheDragon »

I've never sat down to flesh out the idea because it's also based on precious little comparative evidence - just that one word. The fact of the matter is that we really don't know what happened to *b, and every theory I've ever seen is based on only one or two words.

User avatar
2+3 clusivity
Avisaru
Avisaru
Posts: 454
Joined: Fri Mar 16, 2012 5:34 pm

Re: The Great Proto-Indo-European Thread

Post by 2+3 clusivity »

I see wiki notes that Cowgill's Law in Germanic applies to *H3 and possibly to *H2.

I usually interpret the dual pronouns as having *H1, so I have wondered if a broader reading of Cowgill's Law in Germanic is possible.

Perhaps, PIE *RHw > P.GMC *Rgw (prior to grimm's law).

Are there any counter examples with H1 or H2?
linguoboy wrote:So that's what it looks like when the master satirist is moistened by his own moutarde.

User avatar
Znex
Lebom
Lebom
Posts: 226
Joined: Sun Aug 25, 2013 2:45 am
Location: Australia

Re: The Great Proto-Indo-European Thread

Post by Znex »

Reposting this from the CBB:
Ashtâr Balînestyâr wrote:Are there any credible reconstructions for the dual endings of the perfect? I’ve been using Donald Ringe’s -wé for the first person, and my admittedly makeshift -éth₁r̥ and -étr̥ for the second and third persons, backformed from the Sanskrit endings -áthuḥ and -átuḥ. For the 2.du. and 3.du., Greek has -aton, which is just the -a- from the 1.sg. with the primary ending -ton. The Gothic 2.du. ends in -ts, again just a primary ending, and the Avestan endings are unattested.
Native: English || Pretty decent: Ancient Greek || Alright: Ancient Hebrew || Eh: Welsh || Basic: Mandarin Chinese || Very basic: French, Latin, Nisuese, Apsish
Conlangs: Nisuese, Apsish, Kaptaran, Pseudo-Ligurian

Sumelic
Avisaru
Avisaru
Posts: 385
Joined: Sat Mar 28, 2015 7:05 pm

Re: The Great Proto-Indo-European Thread

Post by Sumelic »

I know I've seen people reconstructing *e and *o as [ə] and [a] (or [ä, ɑ]; e.g., it's argued for in The Indo-European Vowel System and the Qualitative Ablaut by E. G. Pulleyblank (1965). The one thing that seems to kind of be explained better by *e [ə] would be the origin of ablaut-alternations with *o [a]--is that the main argument for it?

All of its reflexes in the major branches I can think of point to frontness.

Typologically, I guess /i u ə a/ is a common 4-vowel system but /i u e a/ isn't that unusual either, from what I understand. (I'm ignoring *a since it's so commonly explained as a conditioned allophone of *e.)

Edit: guess I should have searched this thread before asking, sorry:
Tropylium wrote:*e as [+front] finds confirmation in every PIE branch AFAIK.

User avatar
Soap
Smeric
Smeric
Posts: 1228
Joined: Sun Feb 16, 2003 2:57 pm
Location: Scattered disc
Contact:

Re: The Great Proto-Indo-European Thread

Post by Soap »

I think it's usually positioned the other way around. /e/ comes from earlier /a/, and /o/ (the less common vowel) from earlier /ə/. This makes more sense since in languages with a phonemic schwa patterning as a full vowel, it tends to be a rare vowel. It could be that this /ə/ itself was once just an unstressed allophone of /a/, and therefore that there was a three-vowel system. See https://en.wikipedia.org/wiki/K%CA%B7etw%C3%B3res_rule for one example of this theory in applied use.

However, Ive always been sympathetic to the idea that most proto-language reconstructions based entirely on internal evidence tend to produce skimpy phonologies, and that there may have been independent /e/ and /o/ vowels all along that simply merged with the /a/ and /ə/ at some point in the late history of PIE. This would explain those few /o/'s that dont seem to be due to either ablaut or laryngeal influence. The phonology would be a lot messier with this setup, but with a messy proto-language it's a lot easier to explain the messes in the early daughter languages.
Sunàqʷa the Sea Lamprey says:
Image

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: The Great Proto-Indo-European Thread

Post by KathTheDragon »

The idea of *o as [ə] is hard to defend in light of it showing up as a long vowel in both Hittite and Indo-Iranian, under certain conditions. Probably the only good reconstruction is *e [a] *o [aː] (and subsequently [æ ɑː], as argued for by Kümmel [for which I will grab a link later]. However, he does not discuss the further possibility of contrasting [i iː u uː] as well; I propose that [iː] = *i (e.g. in the old ins.pl. *-is), [uː] = *u (e.g. in *bʰuH-). while and correspond to "non-apophonic" *e and *o, as in the nom.pl. *-es, and *poti-.

User avatar
Soap
Smeric
Smeric
Posts: 1228
Joined: Sun Feb 16, 2003 2:57 pm
Location: Scattered disc
Contact:

Re: The Great Proto-Indo-European Thread

Post by Soap »

KathTheDragon wrote:The idea of *o as [ə] is hard to defend in light of it showing up as a long vowel in both Hittite and Indo-Iranian, under certain conditions. Probably the only good reconstruction is *e [a] *o [aː] (and subsequently [æ ɑː], as argued for by Kümmel [for which I will grab a link later]. However, he does not discuss the further possibility of contrasting [i iː u uː] as well; I propose that [iː] = *i (e.g. in the old ins.pl. *-is), [uː] = *u (e.g. in *bʰuH-). while and correspond to "non-apophonic" *e and *o, as in the nom.pl. *-es, and *poti-.

Im interested to see the paper, but I dont think it contradicts my theory. The people proposing /e/=[a] /o/=[ə] are talking about a period thousands of years before the attested Indo-Iranian and Hittite scripts. The /o/ could easily have swung around to be stronger than the original /e/ by that time, especially, as you say, under certain condiitons. Also I imagine that long vowels, like schwas, are "marked" and therefore rarer than the plain full-length vowels such as /a/, and therefore it would be likely for the rarer of two vowels to be the one that lengthens, even if it was originally a weaker vowel thousands of years in the past.
Sunàqʷa the Sea Lamprey says:
Image

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: The Great Proto-Indo-European Thread

Post by KathTheDragon »

Hm. The paper I have lying around by Kümmel is not the one where he talks about the PIE vowels. This is annoying because it means I have to go hunting for a paper I don't know the name of.
Found it

The conditions are not the same in Indo-Iranian and Anatolian - the former is Brugmann's law (open syllables), the latter is lengthening under the stress. Besides, *o is neither 'weaker' (cf. also Tocharian *e > *ä, *o > *e), an 'unstressed variant of *e' (cf. *wódr̩, *póds), or even 'rare', which leaves me justifiably skeptical about *o [ə].

There are other slightly more circumstantial arguments for a long *o. Namely, the o-stem nominal inflection is really easy to explain if we assume that many or all of the athematic endings have an inherent *e, which is in the zero-grade under the accent - we can then trivially have e.g. nom.sg. *-os < *-ās < *-a-as = thematic extension + athematic ending. Then there's a probable sound law *-eHe > *-oH (and *-ese > *-os is logical but there are no potential examples), for which cf. the dual endings *-h₁e (athematic) vs. *-oh₁ (thematic), and possibly also *-h₂e (1sg perfect) and *-oh₂ (1sg thematic) (though of course the theory where these are in fact cognate is non-standard). I fail to see how either of these are explicable under the theory of *o [ə].

User avatar
Pole, the
Smeric
Smeric
Posts: 1606
Joined: Sat Feb 11, 2012 9:50 am

Re: The Great Proto-Indo-European Thread

Post by Pole, the »

The conditions are not the same in Indo-Iranian and Anatolian - the former is Brugmann's law (open syllables), the latter is lengthening under the stress. Besides, *o is neither 'weaker' (cf. also Tocharian *e > *ä, *o > *e), an 'unstressed variant of *e' (cf. *wódr̩, *póds), or even 'rare', which leaves me justifiably skeptical about *o [ə].

[…] I fail to see how either of these are explicable under the theory of *o [ə].
Well, it could have been a non-weak /ə/.
The conlanger formerly known as “the conlanger formerly known as Pole, the”.

If we don't study the mistakes of the future we're doomed to repeat them for the first time.

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: The Great Proto-Indo-European Thread

Post by KathTheDragon »

Are there parallels? Essentially every stage in the derivation of *e *o from *a *ā is well-supported; can you say the same for *ə yielding long back round vowels?

User avatar
Pole, the
Smeric
Smeric
Posts: 1606
Joined: Sat Feb 11, 2012 9:50 am

Re: The Great Proto-Indo-European Thread

Post by Pole, the »

Well, it's not the exact same process, but for instance, Old Polish long /aː/ turned that way, yielding /ə~ɜ~ʌ/ in some dialects and /ɔ/ or /ɔu/ in some others, probably via something like /ɤː/ or /əː/. (It merged with short /a/ in Modern Polish.)
The conlanger formerly known as “the conlanger formerly known as Pole, the”.

If we don't study the mistakes of the future we're doomed to repeat them for the first time.

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: The Great Proto-Indo-European Thread

Post by KathTheDragon »

Yeah, that's not an examle of [ə] > some back long vowel.

User avatar
Pole, the
Smeric
Smeric
Posts: 1606
Joined: Sat Feb 11, 2012 9:50 am

Re: The Great Proto-Indo-European Thread

Post by Pole, the »

Yep, but PIE doesn't require it.
The conlanger formerly known as “the conlanger formerly known as Pole, the”.

If we don't study the mistakes of the future we're doomed to repeat them for the first time.

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: The Great Proto-Indo-European Thread

Post by KathTheDragon »

What?

User avatar
Pole, the
Smeric
Smeric
Posts: 1606
Joined: Sat Feb 11, 2012 9:50 am

Re: The Great Proto-Indo-European Thread

Post by Pole, the »

Well, you said *o couldn't have been /ə/, because /ə/ would have been “weaker” and inherently “unstressed”, which PIE *o isn't. I countered that by saying that /ə/ doesn't need to be inherently weaker, you can have a language with stressed schwas, or even long schwas.

Now about long back vowels: if we assume /ə/ behaved as a normal vowel and not as an unstressed/reduced allophone, it could have easily become a back vowel (by the virtue of a simple vowel shift), a long vowel (by the virtue of conditional lengthening) or both. There is nothing in PIE requiring this change to be atomic and unconditional.

Also:
However, he does not discuss the further possibility of contrasting [i iː u uː] as well; I propose that [iː] = *i (e.g. in the old ins.pl. *-is), [uː] = *u (e.g. in *bʰuH-). while [i] and correspond to "non-apophonic" *e and *o, as in the nom.pl. *-es, and *poti-.

There's a problem with that. As far as e : o (← [a aː] in your version) apophony in PIE is easily documented, analogical e : i (← [i iː]) and o : u (← [u uː]) is not shown to exist.
The conlanger formerly known as “the conlanger formerly known as Pole, the”.

If we don't study the mistakes of the future we're doomed to repeat them for the first time.

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: The Great Proto-Indo-European Thread

Post by KathTheDragon »

There's a good reason for that: *e ~ *o apophony is unrelated to the original length contrast - it's actually a later development related to stress. But since old *ī *ū are so rare, we wouldn't really expect to see any apophony.

Edit: I'll concede that *o [ə] is in principle possible, it's still hardly compelling and has essentially no explanatory power. *o [aː], however, does explain more than what it was posited for.

Another edit: Actually, another reason why we don't see any apophony from the old high vowels could be that by the time the shortening took place, *i and *ī (resp. *u and *ū) had a qualitative difference, not unlike in modern English. The result of shortened *ī would then be distinct from original *i, preventing a merger (cf. the not entirely parallel scenario where lengthened /ɪ/ in BrE yielded /ɪː/ a distinct phoneme from /iː/)

User avatar
kanejam
Avisaru
Avisaru
Posts: 257
Joined: Sat Mar 30, 2013 1:16 pm
Location: New Zealand

Re: The Great Proto-Indo-European Thread

Post by kanejam »

Just finished reading a rather long article on reduplication. The first half argues that the reduplicant starts off copying segments from the base towards using fixed segments in the reduplicant. The author then looks specifically at PIE and concludes that the present reduplication in Italo-Celtic, Greek and Sanskrit was a repair method when the loss of laryngeals ended up with stems that violated the biconsonantal stem pattern. The Hittite forms don't quite seem to correspond to the others, and the author argues it shouldn't be reconstructed for PIE proper.

Perfect reduplication, on the other hand, does go back to PIE was originally vowel copying, which is continued in forms such as spopondī and cucurrī (perfects of spondeō 'I promise' and _currō 'I run'_), rather than the usual analysis that these all started with /e/ and then assimilated, and was used where the usual o-grade would not have created a distinct perfect stem. This can be seen in most branches except Greek, which moved to fixed vowel reduplication using /i/: Greek τίθημι corresponds to Sanskrit dádhāti.
Last edited by kanejam on Sun Jun 18, 2017 8:48 pm, edited 1 time in total.
If you cannot change your mind, are you sure you have one?

Here's a thread on Oscan.

User avatar
Ser
Smeric
Smeric
Posts: 1542
Joined: Sat Jul 19, 2008 1:55 am
Location: Vancouver, British Columbia / Colombie Britannique, Canada

Re: The Great Proto-Indo-European Thread

Post by Ser »

Is the use of relative pronouns to connect sentences, as if they were practically demonstratives, as in Latin, attested anywhere else in (old) Indo-European? And if it is, have there been any thoughts that these kʷ- pronouns could've been demonstratives in the past? For example:
Cicero, [i]On the Nature of the Gods[/i] wrote:Sunt enim philosophi et fuerunt, qui omnino nullam habere censerent rerum humanarum procurationem deos. Quorum si vera sententia est, quae potest esse pietas, quae sanctitas, quae religio?
"There have been philosophers, after all, who have thought the gods don't take care of human affairs at all. If their thinking is true, what piety, what moral integrity, what religious zeal can there be?"

I think we can agree that in this case it'd be very weird to try to translate quorum as "whose". In fact, I personally can't even produce a grammatical translation of the sentences using "whose".
kanejam wrote:Perfect reduplication, on the other hand, does go back to PIE was originally vowel copying, which is continued in forms such as spopondī and cucurrī (perfects of spondeō 'I promise' and ), rather than the usual analysis that these all started with /e/ and then assimilated, and was used where the usual o-grade would not have created a distinct perfect stem. This can be seen in most branches except Greek, which moved to fixed vowel reduplication using /i/: Greek τίθημι corresponds to Sanskrit dádhāti.
"and currō 'I run'"

User avatar
kanejam
Avisaru
Avisaru
Posts: 257
Joined: Sat Mar 30, 2013 1:16 pm
Location: New Zealand

Re: The Great Proto-Indo-European Thread

Post by kanejam »

Serafín wrote:"and currō 'I run'"
Oops, thanks :oops:
If you cannot change your mind, are you sure you have one?

Here's a thread on Oscan.

User avatar
dhok
Avisaru
Avisaru
Posts: 859
Joined: Wed Oct 24, 2007 7:39 pm
Location: The Eastern Establishment

Re: The Great Proto-Indo-European Thread

Post by dhok »

kanejam wrote:Just finished reading a rather long article on reduplication. The first half argues that the reduplicant starts off copying segments from the base towards using fixed segments in the reduplicant. The author then looks specifically at PIE and concludes that the present reduplication in Italo-Celtic, Greek and Sanskrit was a repair method when the loss of laryngeals ended up with stems that violated the biconsonantal stem pattern. The Hittite forms don't quite seem to correspond to the others, and the author argues it shouldn't be reconstructed for PIE proper.

Perfect reduplication, on the other hand, does go back to PIE was originally vowel copying, which is continued in forms such as spopondī and cucurrī (perfects of spondeō 'I promise' and _currō 'I run'_), rather than the usual analysis that these all started with /e/ and then assimilated, and was used where the usual o-grade would not have created a distinct perfect stem. This can be seen in most branches except Greek, which moved to fixed vowel reduplication using /i/: Greek τίθημι corresponds to Sanskrit dádhāti.
Wait...τίθημι is a present, not a perfect; its perfect is τέθηκα.

Whitney notes (§643) that a root-vowel in the zero grade is always replaced by i in reduplicated presents, which is why the present of 'bear' is (at least in Vedic) bíbharmi, bíbharṣi, bíbharti. There are also (§661 onwards) a number of roots of the form √Cā that have i-reduplication: √śā, śiśāti 'sharpen', √mā, mimāti 'measure, bellow', √hā/jihīte 'remove', √sthā/tiṣṭhati 'stand'. Pretty clearly in these cases we have a root *CeH reduplicating to *CHCeHti, but a number of *CeH roots, particularly √dā (*deh₃-) and √dhā (*dʰeh₁-), reduplicate with -a- even from the earliest layer of the Vedas.

And none of this explains very well why Greek would have fixed ι-reduplication in presents, since it doesn't have /i/ as the reflex of interconsonantal laryngeals. I can't think of a single root *Cer(C)- that has a reduplicated present in Greek, but there probably is one...

Post Reply