Re: [tied] Re: Earth and Thorn

From: Patrick Ryan
Message: 39176
Date: 2005-07-11

 
----- Original Message -----
Sent: Sunday, July 10, 2005 8:20 PM
Subject: [tied] Re: Earth and Thorn

--- In cybalist@yahoogroups.com, "Patrick Ryan" <proto-
language@......> wrote:

[Jens:]
>   You don't? Fair enough. The word for "earth" is Hitt. tekan,
gen.
>   tagnas, which must represent a more original form of the
paradigm
>   which was changed in the other branches by introduction of the
>   product of the cluster *d(h)g^h- as it had been in the locative,
**d
>   (h)g^h-ém(-i) > IE *g^h{th}ém(-i) (Ved. ks.ámi). This is one of
the
>   mainstays of the understanding of the "thorn" clusters in IE. It
is
>   of course also one of the basic arguments for an Indo-Hittite
model,
>   indicating as it does that Anatolian was the first branch to
split
>   away from the IE unity. This is all classical knowledge by now.

[Patrick:]
>   Some very competent linguists of the past (Benveniste, for one)
looked at the disconnect between Hittite and _ALL_ the other IE-
derived languages, and concluded that Hittite introduced the
metathesis; so the original form was *g^h-Dem- (I will use -D for
bar-d, thorn).

[Jens:]
Did he really? Then he was wrong, and everyone who has chosen to
follow him has been wrong too. 
 
***
Patrick:
 
And Benveniste was backed in his opinion by Pokorny, who, might be said, to have been somewhat knowledgeable.
 
As usual, whoever does not share your opinion is simply wrong. No real arguments necessary since you have looked at all the evidence and made up your mind for us all.
 
By the way, I went along with Miguel's calling ð thorn so as not to seem to be knitpicking but since you decided to name your thread with 'thorn', I guess I am going to have to point out that the word we have been discussing (*g^hðém-) has 'edh' not 'thorn', which is Þ.
 
Of course, I already anticipate your reponse: you think the word is better reconstructed with Þ than ð, am I right?
 
***
 


[Patrick:]
>   In order for _ALL_ the other IE-derived languages to have made
the metathesis instead of Hittite, we are forced to assume a
theoretical unity of _ALL_ non-Hittite languages which is barely
theoretically possible but highly unlikely.

[Jens:]
That amounts to a very strong load of praise to those who saw this,
the only viable solution, anyway. It goes to the credit of a series
of scholars, including Paul Kretschmer, Brandenstein, Burrow and
Schindler. The still living representatives Jón Gunnarsson and
V.V.Ivanov might take a bow on behalf of all.  I wouldn't say  with
reference to any of these or any other scholars, " And if you have
never heard his name, you are simply underread for this discussion
and discussion list: this is not Sprachenkindergarten." I for one am
not above giving a free class to someone who needs it; I would have
preferred a more pleasant climate of discussion, though.
***
Patrick:
 
"the only viable solution", why there we have it again! How could Pokorny and Benveniste have been so idiotically blind as not to see, as you can, there was _only one_ answer: YOURS.
 
I am all for a free class once in a while but I do not do brain implants.
 
***
 

[Jens:]
The fact is that the Anatolian paradigms are of a more archaic
makeup than the one unanimously pointed to by the other branches. In
fact, the Hittite paradigm has precisely the ablauting
(amphikinetic) structure one would assume as the starting point of
the levelled "zero-grade of root plus ablaut of suffix"-structure
seen in Greek, Indo-Iranian, Old Irish, Balto-Slavic and Albanian.
You may do wisely to consult the relevant paragraphs of the
introductions by Szemerényi, Meier-Brügger or Fortson. Our mutual
hero Sihler uses it correctly too.
 
***
Patrick:
 
There we go again. "The fact is ..." I guess no one on the other side of the question had any facts at all. Well, let us thank our lucky stars that at least ONE PERSON is smart enough to recognize a "fact" when he sees it.
 
The fact really is you have absolutely no way of really knowing if Hittite preserves a form that predates the split of Hittite and non-Hittite or not. The Hittite form, tikan, could be the Hittite solution to a consonant cluster; or, it could be only the way they wanted to spell /tkán/. Or it could be from a different root altogether, like *dhig^hóm.
 
I hope we all, at least, can agree that the root had the form *CVC before any suffixes were added.
 
If the root was *g^hedh-, adding -*óm would produce *g^hdhóm, the form we find reconstructed in Pokorny except that the pedants have insisted on ð for dh: *g^hðóm.
 
If the root was *dheg^h- as Jens and Miguel etal. say, adding -*óm would prodce *dhg^hóm.
 
Now *dhg^hóm could work for Tocharian tkaM but it cannot work for tikan.
 
Precisely, some will say. Hittite tikan shows an archaic form that pre-dates zero-grade Ablaut: *dheg^hóm.
 
But if it does, Tocharian tkaM cannot be directly related since it shows zero-grade in the root.
 
Also, as Peter kindly pointed out: "the Tocharian "earth" word appears as tkam (Toch A) but also as ksaise
(older Toch B) and this later form points to a pre-Tocharian kt- cluster, not tek-."
 
So, to believe that the Tocharian A and Hittite forms are related, we must assume that Tocharian A tkam derived from *dheg^hóm as did Hittite; and Tocharian (A and B) later independently innovated the zero-grade form for the root. Thereafter, Tocharian A kept the inherited coronal-dorsal order while the Tocharian B found them too hard to pronounce, and substituted by metathesis dorsal-coronal.
 
Or is it likelier that Tocharian A and B both inherited *g^hdhom-, which Tocharian B kept (ksaise) while Tocharian A metathesized kt into tk?
 
I am for the second choice because it involves a lot less typing.
 
We might also what you look at a real fact: several languages display -*n forms with *dheig^h-, and some derivatives mean 'earthen'.
 
If Hittite tikan represented PIE *dhig^hón-, the -*i- would be naturally explained as the zero-grade of -*ei-, and the expected Ablaut pattern could be seen. There would be no need to postulate the miraculous appearance of a pre-zero-grade Ablaut form; I wonder if Jens and Miguel know of any other such unicorns in Hittite?
 
[Patrick:]
>   As far as "classical knowledge" is concerned, "classical
knowledge" once held the world was flat although there were, at all
times, those who asserted it was round.

[Jens:]
And therefore you stick to good old preclassical lack of knowledge.
I refuse to follow.
***
Patrick:
 
At your peril. Many of the pre-classical natural philosophers held _correctly_ that the world was round.
 
The Academy thought that view "politically incorrect" so they were maligned and ignored; and science was held back for a thousand years.
 
***
 

[Patrick:]
>   Frankly, just the initial premise is preposterous.

[Jens:]
What do you mean by this? What premise?

[Patrick:]
But then to explain it, as you seem to be doing, by an _ALL_ non-
Hittite response to the phonological shape generated by the word in
the locative (really just an adjectival form) case, compounds
preposterousness with sheer incredibility.

[Jens:]
The locative of a word meaning `earth' is hardly something marginal.
In an amphikinetic paradigm the loc.sg. should have the structure *d
(h)g^h-ém(-i) with zero-grade of the root and accented e-grade of
the suffix. The form may be endingless or extended by the ending *-i
(Vedic has both ks.ám and ks.ámi).
 
***
Patrick:
 
Marginal? It is not even _on_ the margin! The nominative is _already_ *g^hdhém- or *g^hdhóm-. Forming a locative with or without -*i does not change anything.
 
***
 
 
 Melchert has published an
interpretation of Luwian /inzgan/ as meaning `into the earth' in
which in- is the preposition *H1en- `in', and -z(a)gan is the old
accusative which he posits as Anatolian *dzg^ó:m, identical with
Ved. ks.á:m and Avest. zaNm. I find it more likely that -zgan is the
endingless locative, if need be in the syntactic use as a Wohin-
Kasus which is not uncommon in Hittite. In that case there would
have been some development in the direction of assibilation or
affrication in the locative, the only case where the oldest paradigm
had a real cluster, in all the IE we have. But even here the
transposition of the dental or sibilant part (the "thorn" part) to
the final position of the cluster was posterior to the separation of
Anatolian from the rest of IE.

***
Patrick:
 
Well, now Sturtevant is pre-classical: he does not know of a locative case for Hittite. But, of course, since it is endingless, who can say Jens is wrong? And the _fact_ that the locative case in PIE indicates _absence of motion_ is, I guess, also a pre-classical idea.
 
Since both PIE -*om (neuter) and -*m (accusative) show up as -an in Hittite, a form like z(a)gan, if it existed, would represent the nominative/accusative of most Hittite nouns. If Luwian in- means 'into', the accusative would be the likeliest case to expect.
 
Also, as an aside, those of you who read cuneiform know that it is impossible to write inzgan with it. Perhaps it is Hieroglyphic Luwian, which makes it even less possible. Annick Payne knows of no "inzgan" so she must also be pre-classical.
 
And by the way, it should be "edh" part rather than "thorn part".
 
 
***
 

[Patrick:] 
>   What on Earth leads you (or anyone else) to think that *g^h-Dém-
i- would, in any way, merit a different phonological deveopmental
response from *g^h-Dém-???  This, by itself, is highly suspect.

[Jens:]
Could you explain what you are referring to? It seems you should be
marched ack to your desk to read again and be ashamed. As a special
service, you need only look above. 
***
Patrick:
 
Ueni, uidi, uici
 
***
 

[]
[Jens:]
Out of kindness, I snip something here.
 
***
Patrick:
 
The mohell you say.
 
***


[Patrick:]
To my way of thinking, *-Dg^h- has no advantage whatsoever over *g^h-
D- as an initial cluster. Tell me what advantage you imagine here.
And forget about thorn. If our understanfing of thorn is based on
the premises mentioned above, the problem desperately needs a new
look.

[Jens:]
Your way of thinking would not account for the facts of Hittite:
nom.-acc. te:kan has a vowel between the dental and the velar, and
the two come in that order. The two consonants are not contiguous,
so there is no reason to assume they have switched places. That is
evidence for the order of the elements. Luwian has tiyammi-s from
*De:g-om- with the usual weakening of g-sounds. The order of the
elements also perspires, as Miguel points out, from the cases where
one of the stops is lost, because then it is always the velar that
remains which must then have been the second element. There is a
fine little list if examples of this; the ones I remember are: Ved.
turí:ya- 'fourth', IE *k^m.tóm '100', Gk. kteíno:/kaíno: 'kill', Gk.
khamaí (and cognates).
***
Patrick:
 
First off, you and I both know that tekan and tikan would be written the same in Hittite cuneiform.
 
As for Luwian tiyammi-s, it could certainly be cognate with Hittite tikan; but that does not prove PIE *dhg^hém- was the basis of metathesized *g^hdhém-; it could just as easily prove that a Hittite-Luwian source for both was PIE *dhig^hém/n-; in fact, that is far more likely.
 
I am well aware that it is usually the second element of an initial cluster that is retained although I am sure you also know that when an infant is learning to speak he says /piz/ not /*liz/ for 'please'.
 
But there are examples of words derived from *g^hdhém- where the -dh-, the second element, is retained: e.g. De:mé:te:r.
 
If this combination did not have special properties, there would have been no reason for people to play around with edh (ð) now, would there?
 

[Patrick:]
>   You asked in an earlier posting about the "proof" for <*dh> in
the word: Greek <th> is the regular response to PIE <*dh>; and the
Greek reflex is khthó:n.

[Jens:]
You don't say? Well, Greek has no opposition between the reflexes of
d + gh and dh + gh, so its output /khth/ cannot show whether, in
this case, it is from *dg^h or from *dhg^h. It is however possiblé
that Tocharian A tkam. (B kem.) excludes *d which is generally lost
in clusters; however, that language could still hold its surprises.
If reliable this would mean the denbtal involved in 'earth' is
aspirated *dh (*t is excluded because of the iE root structure
constraints).
 
***
Patrick:
 
Sometimes you miss the point altogether.
 
Greek<th> _is_ the regular result of PIE <*dh>. Tell me I am wrong.
 
Your example of what *d + *gh and *dh + *gh occasion might or might not be true (I would like to have an example) but it is irrelevant to the *g^h + dh which I am asserting.
 
***
 
 

[Patrick:]
>   It is an gross example of inbred pedantic scholastic thinking in
its very worst sense to reconstruct a PIE phone (thorn) to explain
the aberrant reponses to *dh in this cluster for the sake of three
or four words. I do not believe I have ever seen a reputable table
of PIE sounds that included it.
CORRECTION: not thorn but edh. PCR
 
 

[Jens:]
Out of sheer unkindness, I let those words stand as you wrote them.
***
Patrick:
 
Now I am _not_ disappointed.
 
***
 
[Jens:]
I have seen thorn being treated many times, practically always as a
problem of phonemicization. Brugmann posited [{th}], [{dh}], [{dh}h]
in complementary distribution; he didn't care much about phonemes,
and he assumed IE age of Bartholomae's law. The problem with "thorn"
is of course whether it should be ranked a phoneme, a question one
cannot really decide for an imperfectly known language. For the
earliest stage of the "Remainder-IE" left when Anatolian (and
Tocharian?) had left the unity, it is even more aggravating. Could
there be a case of *t + *k at that node that would not end up as
Greek /kt/, Sanskrit /ks./, Latin /s/? The Sanskrit sandhi product
of -t s'- is -c ch-, not anything like -k s.-, so the matter is not
just trivial. The core of it is perhaps just one of temperament: We
are still some who do not want to say more than we believe we know.
I have made it my habit to specify the thorn examples by using the
Icelandic "thorn" for all combinations, because there is no
opposition between "thorn" and "edh" or even "aspirated edh". I do
not want to write the presumed input to the clustering if I do not
know what it was.

[Patrick:] 
>   The proper method is to reconstruct it as you seem to be doing
above as *g^hdhém-, and then seek to explain the aberrant reflexes
of _this_ cluster. This in no way necessitates or even makes
desirable the postulation of a new phone which is simply diletante.

[Jens:]
You are really putting your foot in it, aren't you? A new "phone" it
certainly was, discussion may be had over its phonemic status. It
does not seem to be an independent *morphophoneme*.

[Patrick:] 
>   I have come to expect a much higher level of argumentation from
you, Jens, and this line of reasoning seems, itself, 'aberrant' to
me.

[Patrick:] 
>   I have my own ideas about this word and its original
significance; and I think it is mandatory to incorporate into the
semantic and phonological explanation the relationship of
*g^hdhyés, 'yesterday', and *g^hdhu:, 'fish' to *g^hdhem-.

[Jens:]
This sounds alarming; tell me nothing about it.
 
***
Patrick:
 
Again, no disappointment.
 
But I cannot believe that _all_ reads of this list will be able to brush aside even the question of a relationship among these three identically initialed words.
 
***
 


[Patrick:]
>   Sorry to have to be so negative but I am really disappointed.

[Jens:]
I wish I could say the same.
 
***
Patrick:
 
No disappointment there again. I guess you have just told me that you expect nothing valuable from anything I write. Well, I cannot say the same. So, even though your contempt is explicit now, I will continue to soldier on.
 
***
 

[Patrick:]
>   Now when you write back to tell me what an idiot I am, just
remember, you are tarring Benveniste with the same brush.

[Jens:]
*Nobody's* mistakes should be accepted when they have been made out
to be that. Benveniste earned his laurels elsewhere. 
 
***
Patrick:
 
Have your ever actually made one?
 
 
***