RajeshA wrote:I'll try to explain what I find disconcerting in PIE.
These are very valid counterpoints, and can be called objections to not just PIE but the discipline of historical linguistics in general.
a) These sound change patterns are turned into sound change axioms giving them power. They don't remain simply observed data and data patterns, but become tools of theoretical projection. As such they change from simply analysis tools into synthesis tools.
What is axiomatic is : the human articulation won't change in large quanta, but builds up as small mispronunciations - based on how adjacent sounds affect articulation of a certain phone.
Other than that, I don't find any other axioms.
b) Also the sound change process, the evolutionary process, the circumstances, the reasons are completely ignored. Only the end result is deemed interesting.
Actually they aren't ignored. Just that they do not fall entirely under the study of phonetics but human behaviour. Eg. "Principles of Historical Linguistics", HH Hock, chap 20 is entirely devoted to studying these causes.
My personal impression is that humans imitate the accents of people with influence in their community. Actually not just accent, even other things.
- two or more sibling languages being direct children of a common hypothetical parent language. They could be cousins of various degrees, rather than siblings. This has an effect on what sound change "axioms" are used.
Actually, not "direct children", in many cases, eg. we know from those very sound changes that Indo-Iranian was a single language, not independently separated from the parent. There are other conclusions too - eg. Slavic was a single language.
the most notorious assumption is on the directionality of sound change between two languages, especially if one is hypothetical language. Just because a sound change pattern is found in one direction is some unrelated language pair and no sound change pattern has been found in the other direction between living/legacy languages, does not mean such a directionality is impossible. Secondly the other directionality could become probable if one uses other intermediate sound change possibilities.
I know; being from science/math background, we are used to see LHS = RHS and mathematical symbols are substitutable. But phonetics doesn't work like that.
I thought I explained with example of palatalization
, why it is considered irreversible. I remember you had one objection, which I did reply to
. If you have any other objections, do let me know.
We already know that the Indo-European language speaking people belong(ed) to different races. The Ancestral North Indian group is hardly ethnically synonymous with Germans. So somewhere along the way, one ethnic group family has transmitted its language to another ethnic group family. It is obvious that the other ethnic group family was earlier speaking a different tongue.
All evidence from Indian texts points that more than race (as in skin colour/features etc.), language was a greater marker of ethnicity in vedic period. So at least I won't go so far as to say, there was transmission between races (the term as used today).
So there is bilingualism here at work and there are substrata languages here at work.
Again, bilingualism is not ignored. It definitely is one of the strong contenders for causes of sound change. See chapter 16 of Hock's book.
Due to substratum language(s) it is actually impossible to discern the directionality of any sound change axiom at work.
I think even here, the basic phonetic rules do not change - if I learn a new language, eg German, I won't pronounce the 'r' sound uvular, like a native German, I will still pronounce it in retroflex or alveolar, the way I speak in my native Indian language.
1) There can be intermediate sound changes which allow an opposite process to palatalization of sounds with front vowels. It need not be a single stage process. It can be a multiple stage process.
Ok, start by proposing it then.
2) Just because the palatalization of sounds with front vowels has been observed in some languages does not mean the opposite cannot be there in some other languages.
The problem here is the corroborating evidence of front vowel. If indeed, as you say, it is a reversible process, let's call it "velarization", wherein an original palatal becomes a velar (c > k), we won't expect the 'k' standing next to a front vowel like e/i most of the time. But it does in Greek. If velarization was true, we would find the 'k' standing next to a back vowel like 'a'. But we don't.
In case you want to further dissect this directionality, I found a freely available book ( on books.google.com ) that explains it with some examples ...
pp 56 - 58 in "Historical Linguistics" By Theodora Bynon
3) Then there is the distant possibility that kʷ need not be the parent or intermediate sound of both c and p in this case, and there can be theoretically some other intermediate sounds moving from c to p than kʷ.
Ok, not that it hasn't been tried before. The late Satya Swarup Misra did propose such sound changes in ...
S.S. Misra, 1992, The Aryan Problem: A Linguistic Approach, New Delhi.
and a shorter paper in "The Indo-Aryan Migration Debate", E. Bryant.
But unfortunately, he does not address the problem posed by the front vowel in palatalization.
If PIE is to be believed, PIE proponents need to show the presence of substratum language(s) in Sanskrit and Greek which
- did not allow the pronunciation of kʷ
At least for Greek, that evidence for the transformation exists (Epic v's Mycenaean), also for English (Old English-> New English). However, in Sanskrit, there is no epigraphic evidence of labiovelars. There is only trace of the labio-velars which can be discerned by what is called u-colouring of liquids (r/l). Eg. sanskrit 'guru' appears as Latin 'gravis' and Greek 'barus'. The initial labiovelar gʷ causing the u-colouring in sanskrit.
- not whether kʷ can change into c or p, but rather why kʷ changed into c and p in Sanskrit and Greek respectively.
- what is the justification for migrating sound change patterns observed elsewhere to create hypothetical languages as if these were not just patterns but some laws.
Like I said, substrate can be just one of the influencing factors in sound change. I'm afraid linguists can never recover the 'why' here
Also Kazanas speak of heeding more the presence of roots than stems in finding out the "parent" language. The language, where one can show more "root" sounds, would be much closer to the PIE, or synonymous with it, rather than a language which can only show "stems". That way Sanskrit surely wins hands down.
He hasn't claimed more roots, only more regular forms. No linguist makes the correlation between regularity and antiquity. In fact a robust grammatical tradition as in India has resulted in many iterations of Sanskrit (Vedic v/s Pāṇinian forms) which have only kept improving regularity.
We aren't taught this in Sanskrit classes, but several sanskrit words like 'नीड' (nest) have also lost their roots. These can easily be recovered by theory of sound change.