Re: hunt

From: dgkilday57
Message: 65408
Date: 2009-11-12

--- In cybalist@yahoogroups.com, "Richard Wordingham" <richard.wordingham@...> wrote:
>
> --- In cybalist@yahoogroups.com, "Torsten" <tgpedersen@> wrote:
>
> > > In order to explain these alleged doublets as AA loanwords, you
> > > need either two distinct branches of AA with divergent treatment of
> > > Old PAA stops,
>
> > Carleton Hodge has a list of examples of that, but I forgot in what book.
>
> The article is available at <http://www.nostratic.ru/books/(202)The%20implications%20of%20lislakh%20for%20nostratic.pdf> .

Hodge's methodology leaves a great deal to be desired. What he recognizes as "an unsatisfactory state of affairs" involving "problems in the establishment of regular sound correspondences" in Afro-Asiatic, between AA and Indo-European, and beyond (p. 237) is not remedied by his approach. Indeed, such problems as there are become even worse.

It is astounding that the following (p. 250) even got published, unless the reviewers were on Valium:

"Note. IE /w/ corresponding to other /b/ is normal ... There are sporadic survivals of /b/ (Skt. <balam> 'strength'), and there are conditioned survivals. An example of the latter is the /b/ of Lat. <bis> 'twice' from **dbis, where cluster simplification results in Gk. <dis> and Lat. <bis>. The alternate Latin form <dwis> shows that ordinary Indo-European /b/ alternates with /w/. LL [Livlakh, i.e. Indo-Afro-Asiatic] **b is the source of both AAs /b/ and IE /b/w/, and the scarcity of /b/'s as reconstructed for Proto-Indo-European is due to the general shift to /w/. ..."

If this sort of poppycock were posted to Cybalist, its author would quickly be laughed out of the forum. We all know that Lat. <bis> and Grk. <dís> are perfectly regular reflexes of IE *dwís. The Dvenos vase and archaisms like <Dvelonai> (S.C. de Bacch.) show that the Latin shift /dw/ > /b/ occurred in historical times. Hodge has misrepresented a secure Latin protoform as a synchronic "alternate form" in order to justify his ludicrous **dbis by circular argument. Moreover, the claim of a "general shift to /w/" to account for the scarcity of IE /b/ merely doubles the trouble, since the "sporadic" survivals of original /b/ remain unexplained. Anyone attempting long-range comparisons which include Indo-European should first acquire enough IE background to avoid such farcical fumbling.

The list of 32 sets of comparanda, which Hodge calls "core vocabulary related etyma" (pp. 240-5), consists of a mishmash of good AA comparanda (pronouns, body parts, basic verbs and nouns), bad AA comparanda (arbitrarily justified conflation of vaguely similar forms), and with few exceptions ugly IE pseudo-comparanda. The latter were supposedly collected from Szemerényi, an astonishingly bizarre choice for AA-IE comparative work, since Sz. was one of the least laryngeal-friendly IEists of the second half of the last century. Here I will briefly comment on the sets which contain good comparanda for more than two AA branches.

1. Hodge notes (p. 246) that the Egyptian consonant transcribed /3/ actually had the phonetic value [l] in the Old and Middle Kingdoms. We thus have Egy. <p-l> 'fly up', <n-p-l-p-l> 'flutter' which can reasonably be compared with Hausa <filfilwàà> 'fluttering', Ometo <pal-> 'fly', and Cushitic *pal- 'flutter'. (Semantically close, though not mentioned by Hodge, is Semitic *p-l-t, Arabic <falata> 'flee, escape', and perhaps Sem. *p-l-s, Ge`ez <falasa> 'emigrate'.) Also, Egy. <p-r-t> 'fruit' can reasonably be compared with Sem. *pary- 'id.' (Hebrew <pri:>). But there is no basis for relating this 'fruit' root to the 'fly' root, simply because fruit flies like bananas. Hodge attempts to bridge these senses with Egy. <p-r-?> 'go out', High East Cushitic *ful- 'id.', and Sem. *-prur- 'flee'. But the presumed relation between <p-r-?> and *ful- contradicts that already assumed between Egy. <p-l> and Cush. *pal-, and throwing in *-prur- helps nothing. Likewise, connecting Chadic *p-r 'fly, jump' with Cush. *par- 'id.' and Berber *f-r-f-r 'fly' (Touareg <fereret> 'take flight') makes good sense, but including these with Egy. <p-l> and the rest assumes an arbitrary r/l-alternation. That seems to be the heart of the problem with this sort of research. To me it appears that Hodge has conflated three distinct AA roots:

1a. *p-l 'fly', frequentative *p-l-p-l 'flutter', in Egy., Chad., Omot., and Cush., possibly in Semitic 'move swiftly' with root-extensions.

1b. *p-r 'jump, take flight', freq. *p-r-p-r 'fly', in Ber., Chad., and Cush.

1c. *p-r 'fruit', with nominal suffixation (not root-extensions) in Egy. and Sem.

The other words listed here by Hodge have only gratuitous similarity. His inclusion of IE *per- 'fly' (actually 'pass over'), *per- 'forward', *per- 'bear offspring', and *pel- 'thrust' is too silly for comment.

3. Second singular pronouns (absent from Omotic). Not disputed as inherited from AA. Other than the easily explained assibilation of the Egyptian feminine, all have */k/, which should warn researchers against arbitrarily adding glottalization or other features in selected AA groups in order to manufacture pseudo-cognates.

5. Egy. <b-n-w-t> 'sandstone', Sem. *?abn- 'stone' (Heb. <?even>, cf. <ba:na:h> 'build'), Tam. <bnu> 'build', etc. Probably an AA verbal root *b-n 'build' with old preformative (implemental?) */?/ in Sem. and Cush. Citing IE *wer- 'heavy' (actually *(s)wer-) makes no sense whatsoever, not even under the false assumption of earlier **ber-.

10. Egy. <n-g-b> 'turn aside', Ge`ez <gabbaba> 'be bent', Afar. <gu:b> 'id.', Omot. *gub? 'knee'. Not bad as an AA set, but Hodge's citation of IE *ge:u- 'bend' is purely gratuitous. It requires a laryngeal leaving no corresponding trace in AA at all.

11. Egy. <n-f-t> 'breath' (Coptic <nif(e)>), Tou. <unfas> 'id.', Cush. *nAp- 'id.', Sem. *?anp- 'nose' (Heb. <?af>), etc. Again, not bad as an AA set, but including IE *pneu- 'breathe' requires arbitrary metathesis, which fails the laugh test.

14. Sem. *baraq- 'lightning' (Heb. <ba:ra:q>), Omot. *b-r-k? 'id.', Cush. *bark? 'flash'. Not bad, but Egy. <b-l-q> 'bright' and Chad. *b[H]-l 'burn' do not belong here, nor does IE *bHereg- 'shine'. Since he has no actual soundlaws, Hodge could equally well have cited IE *bHelg- 'flash', which has in fact produced a 'lightning' word, Lat. <fulgur>.

16. Egy. <s-n-w-y> 'two' (Copt. <snau>, fem. <sn.te>), Sem. *T-n-y 'id.' (Heb. <s^nayim>, fem. <s^tayim>), Ber. *si:n 'id.', Chad. *s-r 'id.' Hodge suggests comparing with IE *sen(i)- 'apart' (Lat. <sine> 'without', OE <sundor> 'apart', etc.). Finally a plausible AA-IE comparandum!

17. Egy. <s-n> 'smell', Ber. *s-n 'know', Chad. *sun@ 'nose', Omot. *sint?- 'id.', etc. An interesting group. Hodge wants to connect with IE *sent- 'head for, go, feel', requiring an IE root-extension (or retention of an old dental only in IE and Omotic). Another AA-IE proposal worth considering, and a possible connection with Pre-Germanic substrate as well, given all the Germanic 'nose' words in *sn-.

21. Third plural pronouns (again absent from Omotic). Again undisputed AA, and again the stable consonantism should warn against arbitrary substitutions with other words.

22. Egy. <s-n> 'whatever is signified by the ARROWHEAD hieroglyph', Sem. *s^inn- 'tooth' (Heb. <s^e:n>, <s^inn->), Chad. *san- 'id.', etc. A good AA set, but Hodge's inclusion of IE *ser- 'sickle' (actually *serp-) is useless.

24. Egy. <?-n-b>, <?-l-b> 'heart', Sem. *libb- 'id.' (Heb. <le:v>, <libb->), Chad. *l-b 'belly', Galila <liBa> 'id.', and Cush. *l-b- 'chest' make a reasonable set. Ber. *wilih 'heart' and IE *reuto- 'intestines' have no business here. The Egy. form with /n/ is probably older, and this nasal has been lateralized in the other AA branches.

25. Egy. <n-s> 'tongue' (Copt. <las>), Sem. *lis^a:n 'id.' (Heb. <los^o:n>, Aramaic <lis^s^a:n>), etc. Here again Egy. has the old nasal, elsewhere lateralized as later in Coptic.

27. Egy. <m-w> 'water' (Copt. <moou>, Suidas <môu>), Sem. *ma:y- 'id.' (Heb. <mayim>, <me:->, <mo:->), Cush. *ma?- 'be wet', IE *ma:- 'damp'. The cited Ber. and Chad. forms do not belong here. Since the IE form is actually *meh2- (as Hodge should have known), this is a very important long-range gloss.

28. Interrogative pronouns. The Omotic form cited obviously does not belong here.

29. Egy. <m-w-t> 'die' (Copt. <mou>), Chad. *m-w-t 'id.', Sem. *mawt- 'death' (Heb. <mowet>), etc. A good AA set.

Now, the obvious remedy for problems with AA sound-correspondences is to discard the problematic forms from the cognate-sets, as I have done above. Unfortunately Hodge retains a hodgepodge of vaguely similar forms along with the good comparanda. This insistence on inclusion, this philological greed, this obsessive-compulsive hoarding of lexemes, is what motivates Hodge's "consonant ablaut hypothesis", a hopelessly jumbled mishmash of miscellaneous forms. Instead of carefully selecting plausible AA-IE comparanda (above only 16, 17, and 27), he lumps together everything bearing the slightest conceivable phonetic and semantic resemblance, and the plague spreads from his AA to his IE. Thus under his proto-root *p-l (p. 253) he includes IE *(s)pel- 'split', *per- 'forward', *bHel- 'bloom', *bHre(u)- 'sprout', *bHen- 'break', *mel- 'limb', *men- 'project', and Hittite <panniya-> 'drive away', as well as a similar assortment of AA forms. "There is no attempt at a complete inventory of forms known to be from this root." Indeed! It is not hard to see that any language could be reduced to one or two dozen proto-roots this way. This is no longer historical linguistics at all. It is the mindless sorting of linguistic roots into buckets on the basis of gross phonology, with no attention paid to historical plausibility, and having no scientific value whatsoever.

> Of course, sporadic examples are common - consider the English doublets _cage_ and _gaol_ (the second word does have a suffix), _Tyson_ and _Dyson_, and, allegedly, _plonk_ and _blank_.

Yes, the whole fox/vixen thing, and my pet peeve is that too many scholars project recently formed doublets back to a distant proto-stage, or use bogus doublets (and multiplets) to infer phantom oscillation.

DGK