Re: The Great Proto-Indo-European Thread
Posted: Tue Nov 26, 2013 1:28 pm
Oh, shut up, I was being partially sardonic anyway.
WE ARE MOVING - see Ephemera
http://www.incatena.org/
It all depends on the source of the absence. If the absence of the above system is due to a principled grammatical reason (as opposed to merely low probability), then this sort of argument is valid. You're right though that a universal argument from negative evidence is rather weak: all it takes is one piece of counterevidence, and universality is broken. This sort of thing is pretty common in linguistics: take the widely assumed principle that onset consonants can't contribute to syllable weight, which is untrue in several languages.Chagen wrote:http://rationalwiki.org/wiki/Improbable_things_happenMoreover, the traditional three-way phoneme system - unvoiced unaspirate : voiced unaspirate voiced aspirate - is not merely unnatural, it seems to be unattested in any known language, as noted by Jakobson in 1958 (Miller 1977).
Seriously. Improbable things happen. This "PIE couldn't have X because X is rare" is a non-sequitur. We don't care about what happens in other languages, we care about what happened in PIE.
Yes. Many "universals" were posited for which counterexamples were soon found. I have read somewhere that there are known counterexamples for every single item on Greenberg's famous list of morphosyntactic universals. There are about 6,000 languages in this world, belonging to perhaps 300 families; the number of possible combinations of features is much larger. Does it surprise anyone that some perfectly valid and plausible combinations just accidentally fail to occur?kodé wrote:It all depends on the source of the absence. If the absence of the above system is due to a principled grammatical reason (as opposed to merely low probability), then this sort of argument is valid. You're right though that a universal argument from negative evidence is rather weak: all it takes is one piece of counterevidence, and universality is broken. This sort of thing is pretty common in linguistics: take the widely assumed principle that onset consonants can't contribute to syllable weight, which is untrue in several languages.Chagen wrote:http://rationalwiki.org/wiki/Improbable_things_happenMoreover, the traditional three-way phoneme system - unvoiced unaspirate : voiced unaspirate voiced aspirate - is not merely unnatural, it seems to be unattested in any known language, as noted by Jakobson in 1958 (Miller 1977).
Seriously. Improbable things happen. This "PIE couldn't have X because X is rare" is a non-sequitur. We don't care about what happens in other languages, we care about what happened in PIE.
Just that. Personally, I lean towards a two-stage model. Late PIE had the system as traditionally reconstructed; but an early stage may have had the system posited by the glottalists. This is sufficient to explain the root structure constraints and the rarity of */b/, yet keeps the well-established phonological trajectories of the IE branches intact. Early PIE (as it was spoken when Anatolian started going its own way) may have been some sort of intermediate. The pathway may have been either Pre-PIE */t t' d/ > Early PIE */th t d/ > Late PIE */t d dh/ or Pre-PIE */t t' d/ > Early PIE */t d' d/ > Late PIE */t d dh/ (*/d'/ = implosive stop). I lean towards the former, though the latter has the advantage of not requiring a back-and-forth movement of the voiceless pulmonic stops (but aspiration is a common feature of such stops in languages that also have ejectives, e.g. in Georgian; so it would just be a phonemicization and later dephonemicization of a feature that may have been there all the time, and the stops in question did not change at all phonetically).kodé wrote:Of course, the whole edifice gets shakier when it comes to reconstructed languages, like PIE. We really can't have any certain idea of the laryngeal features of the stops; we can merely hazard an educated guess. I personally think the glottalic theory makes more sense typologically, but, there are various problems with it, the biggest one being the lack of explanation for the appearance of breathy voiced stops. In my mind, that's not enough to invalidate the theory, but certainly enough to fuel decades of bitter debate.
1) Stop making things up.What I was going to say was that there's something to be said for amateurs who are curious having more incentive (a hunger for knowledge) than professional linguists (money, probably) to do any sort of work. I certainly would work on PIE reconstruction myself if I had the basic knowledge to start with.
Even so, it proved rather unstable; The system collapsed in every daughter language, after all.Seriously. Improbable things happen. This "PIE couldn't have X because X is rare" is a non-sequitur. We don't care about what happens in other languages, we care about what happened in PIE.
So, there you go, my device went epileptic to stop me from saying something stupid (and then failed.)Terra wrote:1) Stop making things up.What I was going to say was that there's something to be said for amateurs who are curious having more incentive (a hunger for knowledge) than professional linguists (money, probably) to do any sort of work. I certainly would work on PIE reconstruction myself if I had the basic knowledge to start with.
2) Linguistics is more than just reconstructing PIE.
3) Reconstructing is hard.
Fair. If you are like me and do both historical linguistics and conlanging, you must always be aware which hat you are wearing in the moment. This is especially important if your linguistic research serves as input for your conlanging. You must restrict the information flow to one direction: from your linguistic studies to your conlang. If you let slip, you end up like Octaviano - a crackpot.Terra wrote:1) Stop making things up.What I was going to say was that there's something to be said for amateurs who are curious having more incentive (a hunger for knowledge) than professional linguists (money, probably) to do any sort of work. I certainly would work on PIE reconstruction myself if I had the basic knowledge to start with.
2) Linguistics is more than just reconstructing PIE.
3) Reconstructing is hard.
Sure. Indo-Aryan is the only branch which kept the breathy-voiced stops intact - apparently it stabilized them by innovating a set of voiceless aspirated stops, leading to a balanced four-way system in which voicing and aspiration are orthogonal. Everywhere else, it collapsed. The most common kind of collapse was to merge the *Dh and *D grades. The exceptions can be easily numbered:Terra wrote:Even so, it proved rather unstable; The system collapsed in every daughter language, after all.Seriously. Improbable things happen. This "PIE couldn't have X because X is rare" is a non-sequitur. We don't care about what happens in other languages, we care about what happened in PIE.
The problem with indo european comparative linguistics is that it's an 'academical island', so to speak. It seems to largely ignore developments in other fields within linguistics.vec wrote:What do you mean by this statement? Are you being facetious or do you have some specific opinions on the matter?sirdanilot wrote:PIE was not a language.
The problem with indo european comparative linguistics is that it's an 'academical island', so to speak. It seems to largely ignore developments in other fields within linguistics.
With stuff like generative linguistics, I think that's a good thing (because that's another field entirely), but not with things such as contact languages. IE assumes that languages descend from each other only in a genetic way; any borrowing is simply discarded as 'not fitting the system'. The reality is, as we are increasingly understanding, that contact phenomena are at least as important as the genetic relationships between langauges, and in many cases even more important.
What is also very annoying, is that they use their own terminology, rather than terminology that is common in other fields within linguistics. The IE terminology is often decades old, and any developments within linguistic typology go completely unnoticed.
Focus is also mostly on bare lexical items, rather than sentences. Historical syntax is a topic that is hardly written about.
So why was PIE not a language? Because reconstructions are not languages. Reconstructions are useful to make a 'common denominator' between cognate lexical items within a language group. But even a reconstruction of, say, modern-day West-Germanic languages was never at one time actually spoken.
Any articles on how the 'phoneme inventory of PIE was unstable' make little sense; of course it was unstable because PIE never actually existed as a spoken language.
How do you know this?Focus is also mostly on bare lexical items, rather than sentences. Historical syntax is a topic that is hardly written about.
I was going to guess this had happened in some dialect of Armenian, but it's apparently the only possibility that didn't, other than merging all three: Standard Eastern Armenian keeps all three distinct with *T *D *Dʰ > Tʰ T D, Standard Western Armenian has *T *D *Dʰ > Tʰ D Tʰ, and the Istanbul dialect has *T *D *Dʰ > Tʰ D D.WeepingElf wrote:(How about an IE conlang where *D devoiced but *T did not go out of the way so that both grades merged?)
From sequences of a voiceless stop plus a laryngeal, IIRC. For example, the second person dual ending in Sanskrit is -thaḥ, while the third person dual ending is -taḥ. These are reconstructed as *-th₁es and *-tes.Nortaneous wrote:Is there any reason not to posit a tʰ-t/d-dʱ distinction of the sort that currently exists in some dialects of Armenian for PIE, other than that Armenian is the absolute worst language in all of IE to base anything on? [even though said dialects perfectly reflect the phonation distinction reconstructed for PIE, except with aspiration on the t-series]
Where did Indo-Aryan get its unvoiced aspirates?
I don't know when I'll have time to read the paper, but does he say anything on why the outcomes of Grassmann are different in Greek and Sanscrit? In Sanscrit, Grassmann works on the (in the traditional model) unchanged voiced aspirates and produces unaspirated voiced stops, while in Greek it works on the devoiced aspirates and produces unvoiced stops (e.g. Scr. dadha:mi vs. Gr. tithe:mi < PIE *dhVdhe:mi "I put"). Also, in Greek Grassmann also works on /h/ from PIE */s/ (which cannot happen in Sanscrit because it generally keeps /s/ in positions where it could be affected by Grassmann). These facts are normally taken as evidence that Grassmann was a parallel development, not inherited, so he ought to to address that issue.Terra wrote: - He argues that the occurrence of Grassman's Law in both Greek and Sanskrit shows that it was originally a PIE thing, not a separate and identical innovation.
Indeed, PIE and PU were geographically probably quite close to each other. As you say, PU is somewhat younger than PIE. The similarities that speak most clearly in favour of a common ancestor are the resemblances in morphology, especially pronouns and verbal suffixes, but also the accusative singular in *-m. While it is not out of the question that elements of morphology get borrowed, when entire paradigms match, that is a less likely explanation than the assumption that the two languages share a common ancestor.Herr Dunkel wrote:Proto-Uralic is, at least for me, the most plausible next step.
Both groups have their Urheimaten in close vicinity of one another, both share some morphological and syntactical features and tendencies and possibly also some vocabulary. Their age matches up more or less (that is, I think it was Proto-Indo-European that had already split up somewhat when Proto-Uralic began splitting), they have sound systems that generally show similarities etc.
Yes. Gamkrelidze and Ivanov also try to project Grassmann's Law back to PIE, but that too doesn't make sense for the same reason you state above. There are some other bits in G&I's book Indo-European and the Indo-Europeans which IMHO do not make sense, such as the Armenian homeland and the extra phonemes (uvular and labialized dental stops, and a labialized and a palatalized sibilant). What IMHO does make sense, but only for a Pre-PIE and not for PIE proper, are the glottalic theory and the idea that the language was active-stative (though G&I's argumentation for the latter is flawed).hwhatting wrote:I don't know when I'll have time to read the paper, but does he say anything on why the outcomes of Grassmann are different in Greek and Sanscrit? In Sanscrit, Grassmann works on the (in the traditional model) unchanged voiced aspirates and produces unaspirated voiced stops, while in Greek it works on the devoiced aspirates and produces unvoiced stops (e.g. Scr. dadha:mi vs. Gr. tithe:mi < PIE *dhVdhe:mi "I put"). Also, in Greek Grassmann also works on /h/ from PIE */s/ (which cannot happen in Sanscrit because it generally keeps /s/ in positions where it could be affected by Grassmann). These facts are normally taken as evidence that Grassmann was a parallel development, not inherited, so he ought to to address that issue.Terra wrote: - He argues that the occurrence of Grassman's Law in both Greek and Sanskrit shows that it was originally a PIE thing, not a separate and identical innovation.
Valid.hwhatting wrote:I don't know when I'll have time to read the paper, but does he say anything on why the outcomes of Grassmann are different in Greek and Sanscrit? In Sanscrit, Grassmann works on the (in the traditional model) unchanged voiced aspirates and produces unaspirated voiced stops, while in Greek it works on the devoiced aspirates and produces unvoiced stops (e.g. Scr. dadha:mi vs. Gr. tithe:mi < PIE *dhVdhe:mi "I put"). Also, in Greek Grassmann also works on /h/ from PIE */s/ (which cannot happen in Sanscrit because it generally keeps /s/ in positions where it could be affected by Grassmann). These facts are normally taken as evidence that Grassmann was a parallel development, not inherited, so he ought to to address that issue.Terra wrote: - He argues that the occurrence of Grassman's Law in both Greek and Sanskrit shows that it was originally a PIE thing, not a separate and identical innovation.
Ossetian is Iranian, but it's Eastern Iranian - compared to Persian, the batshit crazy branch. And it's the descendant of Scythian/Saramatian (probably moreso the latter), the Steppe lords of old.Herr Dunkel wrote:Since Satemization is an areal thing, it depends on the location. Going east, you can expect /v/ as a reflex of /*w/, a merger of the velar and labiovelar rows, RUKI backing of /s/ and the transformation of palatovelars into sibilants. Since Centum languages do not share any characteristic sound changes, you can only expect the merger of palatovelars and velars since everything else was pretty much a separate development. Anatolian languages preserved the laryngeals and the three rows of dorsals, Tocharian merged the rows into a simple velar system.
Edit: you could take a look at Albanian, Armenian and Ossetian. They all are oddball languages in the IE phylum, although Ossetian is an Iranian IIRC language.
Duh, it means in a subfamily of its own.Salmoneus wrote:Anyway, I don't understand the premise of "IE but not related to any IE group".