PPIE

A forum for discussing linguistics or just languages in general.
Post Reply
DV82LECM
cuneiform
cuneiform
Posts: 182
Joined: 16 Dec 2016 03:31

PPIE

Post by DV82LECM »

Got a question, has anyone ever figured to derive a reasonably viable protoform for PIE?

Speaking bluntly, PIE, for what we think it is, is typologically odd (of course, there are stranger): breathy voiced consonants without the accompanying voiceless set, a horizontal vowel system with ablaut, and syllabic consonants. The first is something I believe that I have seen attested in one language in Africa. The second could be as reasonably assumed as other arrangements (I can't think of any attested language with a horizontal system). The final one is certainly the most common feature across many differing families (usually manifesting as things like prenasalization and consonants with co-articulation).

Now, sure, we don't possess every value of every sound, but IS it possible to reason out what its protoform might have been to account for its strange characteristics? Could it have been something other than fusional? perhaps a stage where morphological elements were more isolating? (That is my thinking.)
Nortaneous
greek
greek
Posts: 633
Joined: 14 Aug 2010 13:28

Re: PPIE

Post by Nortaneous »

yes, Roland Pooth and various other people at Leiden
DV82LECM wrote: 07 May 2021 02:11 breathy voiced consonants without the accompanying voiceless set
also attested in Cao Bang (hence the "Cao Bang theory" for pre-PIE)
a horizontal vowel system with ablaut
probably actually vertical (*a *e = a ə) with a later chain shift, which IIRC would make it similar to NWC. also in NWC schwa is acoustically similar to [ɛ] it seems possible that in some Anatolian languages the vowel transcribed <e> was still schwa, but I don't know if anyone has looked into this particular question
User avatar
Creyeditor
MVP
MVP
Posts: 4335
Joined: 14 Aug 2012 19:32

Re: PPIE

Post by Creyeditor »

I think there has been internal reconstruction for Pre-Indo European/Early PIE, but it's probably hidden on the second page of the google search results for it.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
Salmoneus
MVP
MVP
Posts: 2350
Joined: 19 Sep 2011 19:37

Re: PPIE

Post by Salmoneus »

To be frank, there's a lot of blurring between "PPIE" and "PIE".

Most of the "PIE" you'll see is actually just LPIE - that is, Late Proto-Indo-European, rather than Proto-Indo-Hittite. I don't think we can really securely reconstruct PIH, so a lot of the PIE you'll see is basically LPIE, with some footnotes speculating about PIH - for instance, it seems likely that the PIH verbal system was completely different from the LPIE one, and we can guess at what it must have been like, but those guesses are a lot less secure than the conventional reconstructions of LPIE.

And because of that, there's a blurring between PIH and PPIE - because while we can see the patterns that presumably were once there before LPIE, we often can't tell whether they were still fully there in PIH, or whether they're purely PPIE.

Similarly with something like Glottalic Theory - many people who support GT still accept something like the conventional reconstruction for LPIE, but posit GT as a previous stage. But is that previous stage still there in PIH (i.e. in PIE proper), or is it only in PPIE?

[as you may be able to tell, I think we should be talking about Proto-Indo-Hittite as the family, and use 'PIE' for what is now called 'LPIE'... but annoying that ship appears to have sailed]

Anyway, in answer to your question: no. We cannot reason out what the protoforms must have been, because we don't have enough evidence. We can spot weird things, as you say, and speculate on ways they could have arisen from less-weird things. And we can also spot patterns, as in the verbal system, that suggest change over time, but these are only isolated hints. And if we assume genetic or areal similarities with other languages - perhaps Uralic, perhaps Northwest Caucasian - we can try to think of ways those languages could have been more similar in the past. So yes, we can "internally reconstruct" a possible PPIE - or a host of them - but this is basically informed speculation at best, with no way to 'test' the hypotheses as we can for genuine reconstruction.
DV82LECM
cuneiform
cuneiform
Posts: 182
Joined: 16 Dec 2016 03:31

Re: PPIE

Post by DV82LECM »

Salmoneus wrote: 07 May 2021 12:44 To be frank, there's a lot of blurring between "PPIE" and "PIE".

Most of the "PIE" you'll see is actually just LPIE - that is, Late Proto-Indo-European, rather than Proto-Indo-Hittite. I don't think we can really securely reconstruct PIH, so a lot of the PIE you'll see is basically LPIE, with some footnotes speculating about PIH - for instance, it seems likely that the PIH verbal system was completely different from the LPIE one, and we can guess at what it must have been like, but those guesses are a lot less secure than the conventional reconstructions of LPIE.

And because of that, there's a blurring between PIH and PPIE - because while we can see the patterns that presumably were once there before LPIE, we often can't tell whether they were still fully there in PIH, or whether they're purely PPIE.

Similarly with something like Glottalic Theory - many people who support GT still accept something like the conventional reconstruction for LPIE, but posit GT as a previous stage. But is that previous stage still there in PIH (i.e. in PIE proper), or is it only in PPIE?

[as you may be able to tell, I think we should be talking about Proto-Indo-Hittite as the family, and use 'PIE' for what is now called 'LPIE'... but annoying that ship appears to have sailed]

Anyway, in answer to your question: no. We cannot reason out what the protoforms must have been, because we don't have enough evidence. We can spot weird things, as you say, and speculate on ways they could have arisen from less-weird things. And we can also spot patterns, as in the verbal system, that suggest change over time, but these are only isolated hints. And if we assume genetic or areal similarities with other languages - perhaps Uralic, perhaps Northwest Caucasian - we can try to think of ways those languages could have been more similar in the past. So yes, we can "internally reconstruct" a possible PPIE - or a host of them - but this is basically informed speculation at best, with no way to 'test' the hypotheses as we can for genuine reconstruction.
Very interesting. As for your assertion, in regards to their geographic proximity, I do believe there might be some genetic relationship between PIE and PNWC. However, the only reconstruction I've seen for the latter insists on upwards of 200 individual phonemes; PIE has less than 30. To me, a relationship insists characteristic similarity, and the two, for the discrepancy in number of phonemes, seem silly to compare, at all. As much, and this might be a bygone misconception, but I remember hearing that "caveman" (going back beyond the last Ice Age) speech was supposedly not that advanced (perhaps such is comparable to clicks being considered a primitive sound). I mention this for the fact that these ancestral languages were so much more complex -- almost to the point of MORE advanced -- compared to ours, in both phoneme count and morphological complexity, i.e. Archi. Other than explained away by extreme phoneme loss, what process might explain this? I am not going Sapir-Whorf here, but something for this presupposed complexity in relation to the assumed primitive nature of the people has to be addressed.
User avatar
Pabappa
greek
greek
Posts: 490
Joined: 18 Nov 2017 02:41
Contact:

Re: PPIE

Post by Pabappa »

I think that's just a stereotype .... one that I fell for myself, at one point, and in fact underlies my main conlang, Poswa. "Sounds so simple, but we can't make heads or tails of it .... how can they even learn this language?!"

But in the real world, there's no meaningful pattern to be found .... i think to some extent the myth that languages get simpler over time is just coming from looking through the tunnel of Indo-European ... a language that really did get simpler over time, at least on average (Slavic & Baltic arguably did not).

Likewise, I've seen the argument that phoneme count must have been very high in the world's first languages because humans at that time would have had no concept of phonemes and every possible sound would have been taken as part of the language, even if it was, for example, "a rising-then-falling pharyngealized egressive half-voiced half-high half-front half-long half-rounded breathy-voiced glottalized /o/", where one person created the word and everyone else imitated it exactly. But human language is very old, so even if that were true, we're talking about an era so far back that our proto-languages like Proto-Afro-Asiatic, Proto-Indo-European, and so on are indistinguishable from the modern languages we speak today. There's no reason to expect traces of the primeval human language in languages spoken 20,000 years ago, let alone 2,000 years ago.

Also, anyone proposing a 200-phoneme inventory is wrong, with no further explanation needed. I would love to see this reconstruction.
I'll take the theses, and you can have the thoses.
Salmoneus
MVP
MVP
Posts: 2350
Joined: 19 Sep 2011 19:37

Re: PPIE

Post by Salmoneus »

DV82LECM wrote: 07 May 2021 15:14
Very interesting. As for your assertion, in regards to their geographic proximity, I do believe there might be some genetic relationship between PIE and PNWC.
Not my assertion! Just a common theory. NWC and PIE look quite similar, and would have been spoken in or around the same location. Either a genetic link or areal influence would be prima facie plausible.
However, the only reconstruction I've seen for the latter insists on upwards of 200 individual phonemes; PIE has less than 30.
The 'reconstruction' of NWC is bollocks. Is that Starostin or the like again? The problem is, whenever you decide that two languages are related, you can always just posit a proto-language that they can be derived from, by multiplying the number of phonemes (or morpheme length). If you see that /k/ in one word in Alpha seems to correspond with /k/ in one word in Beta, you say "oh, there was a phoneme /k/ in the ancestor"... and if you then see that /k/ in another word in Alpha instead seems to correspond with /t/ in the other word in Beta, then you can say, "hmm, it seems like Proto-Alpha-Beta actually had TWO velar series, one more palatal than the other (which merged with alveolars in Beta)".

Please bear in mind: no actual human language recorded has ever had more than 86 non-click phonemes, and even that (Ubykh) is a massive, massive outlier. It is vanishingly unlikely that NWC would actually have had over 200 phonemes, but that's exactly what it would look to have if you created a phoneme inventory through this sort of lazy multiplication approach to reconstruction.
To me, a relationship insists characteristic similarity, and the two, for the discrepancy in number of phonemes, seem silly to compare, at all.
But closely-related languages can appear superficially very phonologically different. Finnish apparently can be analysed with only 14 consonants, while Northern Sami - spoken nearby and closely related - may have nearly 40. The languages of New Caledonia, in the Southern Oceanic subfamily, often have between 30 and 40 consonants, while many languages of neighbouring Vanuatu, also in the Southern Oceanic subfamily, have fewer than 20.
As much, and this might be a bygone misconception, but I remember hearing that "caveman" (going back beyond the last Ice Age) speech was supposedly not that advanced (perhaps such is comparable to clicks being considered a primitive sound).
Yes, this is a misconception.

Well, at some point it's likely that some "cavemen" - pre-humans - hundreds of thousands of yeas ago may have had a stage of language between that of chimpanzees (minimal) and humans (extensive). Though whether that semi-language stage would have lasted millennia or decades is unknowable.

However, that stage would have been long, long, long, long, long before the last Ice Age.

There have been theories around the intellectual and/or laryngeal capabilities of Neanderthals (who were not 'humans' in the modern sense, and not the primary ancestor to humans, although many humans have some Neanderthal genes due to interbreeding between species). More research seems to suggest that actually they were pretty smart. Some have suggested that despite being smart, their throats weren't capable of human speech... but this is speculative, and doesn't necessarily mean they weren't capable of something recognisable as language (who knows, perhaps they even used sign-language!).

Anyway, the key point is: whatever we theorise about the initial evolution of human language, that evolution must have taken place so very, very long ago that the distance in time between us and PIE would be just a blip by comparison. Evolution does not appear relevant to the timescales of serious practical linguistics.

I mention this for the fact that these ancestral languages were so much more complex -- almost to the point of MORE advanced -- compared to ours, in both phoneme count and morphological complexity, i.e. Archi.
Archi is not an "ancestral language" spoken by "cavemen" - it's a modern language, spoken by Russians. Contrary to propaganda, Russians are not actually primitive proto-humans, despite many of their languages have complicated phonologies.

It's not clear that "ancestral" languages really were "complex". Etruscan, for instance, probably had under 25 phonemes - four cases in nouns (five in pronouns), male and female pronouns but no nominal gender, two numbers, with case+number being agglutinative rather than fusional, no inflection of adverbs and adjectives, two tenses, two moods (indicative and imperative), two voices. Perfectly bland and uncomplicated.

But in any case, don't just assume that morphological complexity is the same as total complexity - it is generally at least partly compensated for by reduced word order complexity.

Nor can 'complex' be equated to 'advanced' - there's nothing better or worse about having more phonemes, or more cases.
Other than explained away by extreme phoneme loss, what process might explain this?
Random chance is the big one: some areas have lots of phonemes, others have few.

If there is any underlying factor aside from luck, it's probably insularity. Isolated languages with small numbers of speakers are able to evolve more pointless complexity - language is a game, in essence, and as the same small number of people keep playing, the rules can get more complicated. Phonemes and cases accumulate in an in-group in the same way that in-jokes do. [but they don't have to - some small languages have simple phonologies]. Conversely, languages spoken by large and transient populations, with high percentages of non-native learners, will probably tend toward phonological and morphological simplicity, because this is easier for new learners, and helps to preserve intelligibility over larger areas.
I am not going Sapir-Whorf here, but something for this presupposed complexity in relation to the assumed primitive nature of the people has to be addressed.
What needs to be addressed is the unpleasant notion that speakers of Sanskrit or Latin, or indeed Archi, must be assumed to have a "primitive nature" compared to speakers of English. There is no reason to think that this is true.
User avatar
Vlürch
sinic
sinic
Posts: 394
Joined: 09 Mar 2016 21:19
Location: Finland
Contact:

Re: PPIE

Post by Vlürch »

DV82LECM wrote: 07 May 2021 15:14(perhaps such is comparable to clicks being considered a primitive sound)
Well, clicks are articulatorily complex sounds and that's probably why they're so rare. I don't think it's impossible that clicks existed in some absolutely insanely ancient "caveman" languages (although "cavemen" in the stereotypical sense were never really a thing), but the theory I've heard is that clicks originate from consonant clusters, so that'd only mean that at an even earlier period in time they didn't have clicks... and it just makes sense that languages get more complex and then less complex and then more complex again with time over and over again in a neverending cycle.
Salmoneus wrote: 07 May 2021 16:24The problem is, whenever you decide that two languages are related, you can always just posit a proto-language that they can be derived from, by multiplying the number of phonemes (or morpheme length). If you see that /k/ in one word in Alpha seems to correspond with /k/ in one word in Beta, you say "oh, there was a phoneme /k/ in the ancestor"... and if you then see that /k/ in another word in Alpha instead seems to correspond with /t/ in the other word in Beta, then you can say, "hmm, it seems like Proto-Alpha-Beta actually had TWO velar series, one more palatal than the other (which merged with alveolars in Beta)".
Even as a lumper, that annoys me as well. The length of time that has passed from those absolutely ancient hypothetical proto-languages to their daughter proto-languages is so extreme that acting like irregular sound changes (and/or reborrowings) couldn't have happened to explain all those weird semi-correspondences just makes no sense. Acknowledging that irregular sound changes happened doesn't mean just throwing your hands in the air like "oh well, we shouldn't even bother trying to reconstruct anything past proven proto-languages", it just means you won't be able to figure out which of the options the original sounds were closer to (unless some extraordinary piece of evidence comes along from a previously unattested language that's like a "missing link" between the families, but...)
Salmoneus wrote: 07 May 2021 16:24Archi is not an "ancestral language" spoken by "cavemen" - it's a modern language, spoken by Russians. Contrary to propaganda, Russians are not actually primitive proto-humans, despite many of their languages have complicated phonologies.
Maybe it's not true (and the last thing I want is another argument yet again), but in some travel show on TV where some dude travelled in the Caucasus it was said that the people throughout the region (except Cossacks) don't like being associated with Russia and that in practice the Russian government doesn't even have much power in the rural parts.

If this is clearly just overreacting and going to the opposite end to compensate for my past insensitivity, sorry.
Salmoneus wrote: 07 May 2021 16:24What needs to be addressed is the unpleasant notion that speakers of Sanskrit or Latin, or indeed Archi, must be assumed to have a "primitive nature" compared to speakers of English. There is no reason to think that this is true.
Yeah, I don't get why people tend to assume people in the past were that different from people today (or even people today who live outside of the western(ised) world). I mean, of course technology was less advanced and practical day-to-day things were "simpler" for the common peasant or whatever, but at their core they were just people and it's not like we're some kind of futuristic post-animals completely detached from our literal biological and neurological humanity either.
User avatar
WeepingElf
sinic
sinic
Posts: 237
Joined: 23 Feb 2016 18:42
Location: Braunschweig, Germany
Contact:

Re: PPIE

Post by WeepingElf »

Been there, done that, didn't work. I tried to build a homebrew internal reconstruction of Pre-PIE for the purpose of deriving a conlang - Old Albic - from it, but I ran into many difficulties, and had to fill in a lot of things creatively. Of course I did - if it was so easy that a bleeding amateur like me could achieve it, riddles like the origin of PIE ablaut would have been cracked long ago! Even reconstructing Early PIE (a.k.a. "Proto-Indo-Hittite") is difficult since with just two entities compared (Late PIE and Proto-Anatolian), and no reliable outgroup, is difficult - it is hard to decide which of the two has innovated where they differ.

These problems were one of the reasons why I recently decided to rebuild Old Albic as an IE language related to Hittite about as closely as, say, Latin is to Sanskrit. Other reasons were, of course, that Old Albic looked too artificial, and that my ideas about the origin of the Elves had changed in the light of new findings especially by geneticists studying ancient DNA. Old Albic as it has been harks back to the days when I placed the common ancestor of PIE and Old Albic on the northern shore of the Black Sea prior to the Black Sea Flood proposed by Ryan and Pitman. In my current model, IE consists of two primary branches, Northern and Southern. Northern IE consists of all the non-Anatolian IE languages known to us, while Southern IE consists of Anatolian and some unknown languages including that of the Bell Beaker people from whom the Elves descended, so Old Albic is such a Southern IE language. So I no longer need an internal reconstruction of Pre-PIE anymore, not even a reconstruction of Early PIE, as I can use Hittite as my starting point.

That said, a few glimpses of earlier stages of PIE can be obtained, but there are many open questions. Nobody knows for sure which phonetic values the PIE phonemes had, especially in case of the laryngeals. And while the Early PIE phonology seems to have been structurally much the same as the Late PIE one (since the latter can account for the phonological developments in Anatolian quite well), the phonetic values of the phonemes may have changed. I used to be a follower of the glottalic theory, but only in some early stage and not in Late PIE; but since then, I have found that what the glottalic theory seeks to explain can be accounted for by a less radical reinterpretation of the PIE stops: the voiceless stops were aspirated. That's all.

And what regards relatives of PIE, I still think that Uralic is the best candidate, as both families appear to share quite a chunk of their inflectional morphology, even if the 100-odd lexical resemblances are better explained as loanwords - the IE consonants are matched to their closest equivalents in Uralic, and the Uralic vowels faithfully reflect such things as ablaut grades and even vowel-colouring effects of laryngeals. And some of the morphological resemblances encompass further language families, namely Turkic, Mongolic, Tungusic, Yukaghir, Chukotko-Kamchatkan and Eskimo-Aleut. These languages, which I call "Mitian" after the distinctive 1st and 2nd person pronoun roots (the term was AFAIK coined by John Bengtson, though), may form a very ancient family, perhaps descending from a group that fanned out across northern Eurasia from somewhere near Lake Baykal at the end of the last ice age. There is some evidence from genetics which seems to confirm this, if we can tell anything about languages from that. However, PIE seems to have been a very atypical Mitian language, so a substratum perhaps related or areally connected to NWC may have been involved here.
... brought to you by the Weeping Elf
User avatar
qwed117
mongolian
mongolian
Posts: 3921
Joined: 20 Nov 2014 02:27

Re: PPIE

Post by qwed117 »

WeepingElf wrote: 09 May 2021 13:26 That said, a few glimpses of earlier stages of PIE can be obtained, but there are many open questions. Nobody knows for sure which phonetic values the PIE phonemes had, especially in case of the laryngeals. And while the Early PIE phonology seems to have been structurally much the same as the Late PIE one (since the latter can account for the phonological developments in Anatolian quite well), the phonetic values of the phonemes may have changed. I used to be a follower of the glottalic theory, but only in some early stage and not in Late PIE; but since then, I have found that what the glottalic theory seeks to explain can be accounted for by a less radical reinterpretation of the PIE stops: the voiceless stops were aspirated. That's all.

And what regards relatives of PIE, I still think that Uralic is the best candidate, as both families appear to share quite a chunk of their inflectional morphology, even if the 100-odd lexical resemblances are better explained as loanwords - the IE consonants are matched to their closest equivalents in Uralic, and the Uralic vowels faithfully reflect such things as ablaut grades and even vowel-colouring effects of laryngeals. And some of the morphological resemblances encompass further language families, namely Turkic, Mongolic, Tungusic, Yukaghir, Chukotko-Kamchatkan and Eskimo-Aleut. These languages, which I call "Mitian" after the distinctive 1st and 2nd person pronoun roots (the term was AFAIK coined by John Bengtson, though), may form a very ancient family, perhaps descending from a group that fanned out across northern Eurasia from somewhere near Lake Baykal at the end of the last ice age. There is some evidence from genetics which seems to confirm this, if we can tell anything about languages from that. However, PIE seems to have been a very atypical Mitian language, so a substratum perhaps related or areally connected to NWC may have been involved here.
Honestly, I find the connections with Northwest Caucasian to be likely as compared to Uralic, but I honestly think that there's a good reason that most macrolanguage families are seen as very spurious, it's just hard to get that level of accuracy. Anyways, the ablaut system of PIE is very interesting (and iirc, can be used to suggest that PPIE had "one" vowel, which is only comparable to Ubykh's two vowels to my knowledge). I think it'd be difficult to retain that quality in a PPIE linked to another language (unless you argue that the syllabic-consonant nature of /i u/ descends from an earlier vowel being eroded).

Anyways, I guess semi-comparable to that is that I had to derive a form of Arcado-Cypriotic Greek from PIE and Proto-Greek for Ekuprios.
Spoiler:
My minicity is Zyphrazia and Novland
What is made of man will crumble away.

The SqwedgePad
User avatar
WeepingElf
sinic
sinic
Posts: 237
Joined: 23 Feb 2016 18:42
Location: Braunschweig, Germany
Contact:

Re: PPIE

Post by WeepingElf »

The similarities between IE and NWC are IMHO not as great as some people claim, and of the kind that may emerge by contact or substratum influence (if the latter is a thing at all; in Romance linguistics, at least, the idea turned out to be not very useful and is now out of favour). The Yamnaya culture (who probably spoke PIE) and the Maykop culture (who probably spoke PNWC) were close neighbours and influenced each other to a substantial degree, so one would expect some linguistic interaction, too. What also speaks for a contact situation rather than a descent from a common ancestor is that ancient DNA studies have revealed that the two cultures were borne by genetically quite distinct populations. Of course, genes don't really say much about languages, so this may be deceptive.
... brought to you by the Weeping Elf
Post Reply