Why doesn't anyone care about the tones in PIE?
Why doesn't anyone care about the tones in PIE?
People have written up millions of pages worth of studies dedicated to the consonants and vowels in proto-Indo-European, but the purported pitch accent usually just gets one line or at most a paragraph essentially stating that it existed. I would think that since at least two branches of IE (Greek and Indo-Iranian) had directly observable reflexes of the tones, that people would be trying to reconstruct them for PIE too. (Plus there's also Balto-Slavic, but I've heard it said that the tones of Serbian, Sorbian, Lithuanian, etc are not reflexes of the PIE tones, but just happened to appear due to the loss of long vowels.)
Reconstructing the pitch accent might help piece out the problems of the inconsistent reflexes of the laryngeals ... say, maybe H̩1 died when it was after a syllable with a high tone, but survived if the tone was low. Or even if there were no reflexes of any kind in any daughter languages, it'd still be nice to have tonal reconstructions for PIE for its own sake. How come people hardly ever mention them? Are there people who think that maybe they didnt even exist?
Reconstructing the pitch accent might help piece out the problems of the inconsistent reflexes of the laryngeals ... say, maybe H̩1 died when it was after a syllable with a high tone, but survived if the tone was low. Or even if there were no reflexes of any kind in any daughter languages, it'd still be nice to have tonal reconstructions for PIE for its own sake. How come people hardly ever mention them? Are there people who think that maybe they didnt even exist?
Sunàqʷa the Sea Lamprey says:
-
- Avisaru
- Posts: 275
- Joined: Wed Dec 15, 2004 9:05 am
- Location: Nottingham, England
- Contact:
Probably because it's a lot harder to reconstruct them?
Try the online version of the HaSC sound change applier: http://chrisdb.dyndns-at-home.com/HaSC
It looks questionable as to whether PIE had tone at all. If Vladislav Illich-Svitych is correct then the Balto-Slavic accent doesn't go back to PIE. Unless I'm mistaken that leaves only Indo-Iranian and Greek with evidence for tone. If the Graeco-Aryan Hypothesis is correct then Greek and Indo-Iranian are especialy close languages and their similar tone systems could be an innovation specific to that group.
On the flip side Dybo, Starostin, and Nikolayev propose that PIE had two tones. Unfortunately that paper doesn't really go into details, but is more of a plausibility argument based on the study of accent in some Caucasian languages.
On the flip side Dybo, Starostin, and Nikolayev propose that PIE had two tones. Unfortunately that paper doesn't really go into details, but is more of a plausibility argument based on the study of accent in some Caucasian languages.
Sounds too nonsensical to be commented on. To begin with, Lithuanian retains PIE vowel quantity (more-less)...Soap wrote:(Plus there's also Balto-Slavic, but I've heard it said that the tones of Serbian, Sorbian, Lithuanian, etc are not reflexes of the PIE tones, but just happened to appear due to the loss of long vowels.)
Sure, they've posted in this thread :)Soap wrote:Are there people who think that maybe they didnt even exist?
Because they know they have little to say for certain?Soap wrote:How come people hardly ever mention them?
There is a decent accentological reconstruction for Balto-Slavic, and it points more-less unambiguously to contrastive tones attached to syllables (or perhaps morphemes) - rather than word-level pitch accent or anything of that sort. (Note that Balto-Slavic is in all probability not a valid genetic grouping, or at least there's been a strong tendency lately to claim, not without grounds IMO, that most similarities between Baltic and Slavic are secondary.) Reconstructed B.-Sl. tones sometimes depend on the presence of a laryngeal, but not always (not on short vowels, at any rate). Besides, there was a tonal alternation (on nominal roots at least) which seems to have been part of (residually) productive morphology at some point.
Otherwise not much is known for certain. There is Vedic accent (not contrastive pitch or tone, that's secondary) and reflexes of a similar system in extant Indo-Iranian languages; there is moraic accent in Greek, which partly corresponds to Vedic; and there are numerous split reflexes in segments (at least in Germanic, Latin and Celtic) that somehow correspond to either Balto-Slavic or Vedic suprasegmentals. Nobody has ever succeeded in putting all pieces of this puzzle together.
A very brief overview: reconstructed B.-Sl. tones are known to correspond to vowel shortenings that happened in Celtic and Latin, as well as (in a somewhat different fashion) in Germanic. Besides, the intervocalic glide gemination (or lack thereof) in Germanic seems to correspond to B.-Sl. tones as well. On the other side, Vedic accent tends to correlate with Greek accent and B.-Sl. tones, but "exceptions" are too many to be discarded, and obviously the reconstructed B.-Sl. tones create a more powerful feature combinatorics than the accent of Vedic taken as is; in other words, we simply aren't anywhere close to having a working reconstruction that would explain both Vedic and B.-Sl. facts at least. Besides that, there are two somewhat different accent systems (Western+Scandinavian and Gothic) known for Germanic (as reflected by the outcomes of Verner's law in the respective languages); this one, again, correlates with Vedic accent, but (oddly) the most striking cases (e. g. consonant alternations in verbs due to Verner's law) correspond to those Vedic accent shifts which are probably secondary (since accent placement in Vedic verbs, unlike nouns, is mostly deducible from their segmental composition in PIE).
I think you misread what the article says....Etherman wrote:It looks questionable as to whether PIE had tone at all. If Vladislav Illich-Svitych is correct then the Balto-Slavic accent doesn't go back to PIE.
(Emphasis added by me)Wikipedia article wrote:It was shown, however, by Vladislav Illich-Svitych in 1963 that the Balto-Slavic accent does not match with that presupposed PIE accent reconstructed on the basis of Vedic and Ancient Greek—the Greek-Vedic barytones correspond to Balto-Slavic fixed paradigms, and Greek-Vedic oxytones correspond to Balto-Slavic mobile paradigms.
The article says that what Vedic and B.-Sl. have in common is not the immediately observable position of stress (which was the wrong presumption of most researchers before Illich-Svitych), but the distribution of lexemes among accentuation classes. As it was demonstrated later on (mostly by Vladimir A. Dybo), the factor that determines the choice of "accent paradigm" in B.-Sl. looks like a syllable-level tonal contrast; since the distribution of Vedic nouns between the barytonal and oxytonal types correlates with accent paradigms in B.-Sl., it is natural to suppose that it reflects the same PIE feature, in all probability the tone of the root morpheme.
Etherman wrote:Unless I'm mistaken that leaves only Indo-Iranian and Greek with evidence for tone.
Not syllable tone, at any rate. Neither Indo-Iranian nor Greek have "tone" in this sense.
Not until it is demonstrated how the contrastive accent positions developed e. g. from the pre-existing features of segmentals.Etherman wrote:If the Graeco-Aryan Hypothesis is correct then Greek and Indo-Iranian are especialy close languages and their similar tone systems could be an innovation specific to that group.
Basilius
@ Basilius: What's your opinion on Balto-Slavic, in more detail? IMO, there must have been a period of B-Sl unity, there are too many shared innovations:
- Common development in phonology and accent paradigms
- Loads of shared lexical innovations
- Shared (present) stem formations in verbs
- Shared innovations in pronominal and nominal morphology
The main difficulty for me consists in the significant discrepancies concerning the system of verbal endings and the tense-mood systems. Either we assume that 1) Balto-Slavic was a genetic unit with a convoluted verbal system still containing all the potential for developing both the Slavic and the Baltic verbal sytems or 2) two separate IE language groups converged in all the points I listed above, but kept their verbal systems different. I prefer 1) and tend to assume that the bigger closeness of reconstructible Proto-Slavic to the "classical" IE tense-mood system is due to Iranian influence, but I'd like to see your opinion.
- Common development in phonology and accent paradigms
- Loads of shared lexical innovations
- Shared (present) stem formations in verbs
- Shared innovations in pronominal and nominal morphology
The main difficulty for me consists in the significant discrepancies concerning the system of verbal endings and the tense-mood systems. Either we assume that 1) Balto-Slavic was a genetic unit with a convoluted verbal system still containing all the potential for developing both the Slavic and the Baltic verbal sytems or 2) two separate IE language groups converged in all the points I listed above, but kept their verbal systems different. I prefer 1) and tend to assume that the bigger closeness of reconstructible Proto-Slavic to the "classical" IE tense-mood system is due to Iranian influence, but I'd like to see your opinion.
According to computational and statistical phylogenic studies like this one (PDF), Balto-Slavic is a valid grouping. The next higher level includes Greek and Armenian; further up is Germanic and Albanian.
So the languages with pitch accent do form a subgrouping.
So the languages with pitch accent do form a subgrouping.
Last edited by Morrígan on Wed Apr 28, 2010 9:18 am, edited 1 time in total.
So the question then is, is pitch accent from PIE? Or if it isn't, did proto-Balto-Slavo-Indo-Iranio-Germano-Albanio-Greek have a reconstructable pitch accent?
And what about Hittite? I only have Beginning Hittite, from Slavica, which is kind of hard to follow, admittedly- the Hittite syllabary needed help, and the book isn't really laid out the way I wish it was. (My first language textbook was Hansen and Quinn's Ancient Greek book, and books in the style of that and Wheelock's Latin, or if we're talking reference grammars old-style stuff like Whitney's Sanskrit, is what I'm really used to.)
And what about Hittite? I only have Beginning Hittite, from Slavica, which is kind of hard to follow, admittedly- the Hittite syllabary needed help, and the book isn't really laid out the way I wish it was. (My first language textbook was Hansen and Quinn's Ancient Greek book, and books in the style of that and Wheelock's Latin, or if we're talking reference grammars old-style stuff like Whitney's Sanskrit, is what I'm really used to.)
Well, the ablaut patterns in PIE suggest a stress-accent at one point, but IIRC Balto-Slavic, Indo-Iranian, and Greek do seem to point to a consistent pattern.
Beekes reconstructs the accent in the nominal paradigm, and makes some comments concerning the effect of laryngeals in Baltic on the accent patterns.
Beekes reconstructs the accent in the nominal paradigm, and makes some comments concerning the effect of laryngeals in Baltic on the accent patterns.
- Salmoneus
- Sanno
- Posts: 3197
- Joined: Thu Jan 15, 2004 5:00 pm
- Location: One of the dark places of the world
hwh: I can't speak for this particular case, but in general 2 would seem far more likely. All those commonalities you list are things that are frequently areal features anyway.
TGM: you should just run those programs on a wider selection of languages. Give it a few minutes, and we'd have a conclusive and unarguable family tree of the whole of Nostratic. Simple!
TGM: you should just run those programs on a wider selection of languages. Give it a few minutes, and we'd have a conclusive and unarguable family tree of the whole of Nostratic. Simple!
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!
I would agree if it were only shared lexicon or shared phonological features. But for Balto-Slavic you can reconstruct matching accentuation patterns for many individual lexical items; on phonetic developments you have e.g. the fact that PIE syllabic liquids become vocalised either as iR or as uR according to rules that are not yet fully understood, but with matched outcomes between Baltic and Slavic, present-stem formations are often matched between both groups, despite big differences in the verbal system otherwise - this simply doesn't look like loaning or convergence to me, but more like common development.Salmoneus wrote:hwh: I can't speak for this particular case, but in general 2 would seem far more likely. All those commonalities you list are things that are frequently areal features anyway.
Last edited by hwhatting on Wed Apr 28, 2010 9:42 am, edited 1 time in total.
It is my understanding that these are exactly the kinds of features which are used to construct the statistical phylogenic model I posted earlier.hwhatting wrote:But for Balto-Slavic you can reconstruct matching accentuation patterns for many individual lexical items; on phonetic developments you have e.g. the fact that PIE syllabic liquids become vocalised either as iR or as uR according to rules that are not yet fully understod, but with matched outcomes between Baltic and Slavic, present-stem formations are often matched between both groups, despite big differences in the verbal system otherwise - this simply doesn't look like loaning or convergence to me, but more like common development.
They are; it just didn't seem relevant to the tone/pitch-accent discussion.jal wrote:Interesting. And where are the Celtic languages? Or weren't they used?
Italo-Celtic is a valid grouping under this model, and the third group to split off of "Common-PIE" after Anatolian and then Tocharian.
This may suggest that oblique cases in *-m might be original, and the ones with *bh-/p- are innovations. Or that both are innovations, and that PIE had fewer cases than the later languages. Which is less crazy than it sounds, given this model.
Thank you all for the replies. I admit I dont actually understand everything written here but it helps answer my questions even so. (I wasn't specifically asking "what was the tone system in PIE" so much as just asking why most linguists seem to hardly talk about it). I was originally expecting to use this information to help me with a conlang, but now that I look at it, the tone system of this conlang was so weak that it would be more difficult to explain how it could possibly keep such a system than to lose it. (It is a three-tone system, but contrast can only appear in accented syllables, and only if the syllable is open, and only if the vowel is /a/ /i/ or /u/. So, incredibly unstable really. I think I'll just create final consonants like /?/ and /h/ as a remnant of the tones since the syllables that have them are always open.)
Sunàqʷa the Sea Lamprey says: