>>>There is a trace of this in chinese :
>>>"two" nyijH < b/nits
>>>"seven" tshit < b/s-hnit
>>
>>[...] but notice the similarities between *sNit/*bryat (or
>>s-hnit/brat, if you will). Now compare Basque zortzi/bederatzi.
>
>Well, there is some similarity, but can you explain me neatly the
>correspondance between chinese and basque for these numbers ?
Oh God, man, you don't know what you've asked! I apologize for the lengthy
letter but Guillaume is being a trouble maker :)
Alright first I'll show the numeral system in both Basque, ProtoVasconic and
the ST that I'm aware of. You can nibble at it from there. ProtoVasconic is
what I call the precursor to both Basque and the "Aquitanian" found scarcely
in some Roman texts via foreign, Basquish-looking names. Vasconic is the
stage of the language existant in pre-history. Forms that are not inherited
aren't included. I'm very unsure about the exact form of "nine" since for
one, it's unattested in Basque and has other phonetic problems so far.
Finally, please be aware that Basque "z" is /s/, "s" is /s^/ and "tz" is a
single phoneme /ts/ spelled *c in Vasconic:
Basque P-Vasconic ST DeneC.
three hiru *hirwu *gsum *(gu-)sul-mu,
*(gu-)hul-mu
four lau, laur *lawu(-l) *b-li *li-mu
five bost *bazi-l *bnga *pingu
six [sei] --- *drug *rutL
seven zortzi *surac "8" *snit *sul-rit,
*hul-rit
eight bederatzi *midrac "9" *bryat *mn-rit
nine -- -- *dgu (??)*ilkul
ten hamar *hamza-l *dci *hamsi
>(how come the -de- disappeared ?)
Let me consult my notes... I believe I said *kum for "ten". Mea culpa, I've
apparently without my knowledge come to a newer conclusion: *hamsi.
Sometimes the two hemispheres of my brain don't work in harmony. I'm weird
that way.
First thing first, Vasconic had a major vowel shift such that for the most
part we see *u > *i > *a in initial and medial positions. So we find DC *u
as Vasconic *i (Basque e or i) and DC *i/*a as Vasconic *a (Basque a or o).
Sometimes however loss of consonant in *-uC- would help to preserve the
quality of the vowel, resulting in Basque u or o as in DeneCaucasian *muhini
"brain" > Vasconic *muini > Basque buru-muin, buru "head"). Words with final
consonant were given *-i after the Vasconic stage that I reconstruct. Thus
*midrac "eight" > *midraci > bederatzi. Sometimes this added vowel shows up
and sometimes not which in itself would devote a page of explanation.
On the SinoTibetan front, we should expect another vowel shift here with
differences between open and closed/enclitic syllables:
*i > ST *i/*a and *u > ST *ei/*u
On to the show...
*(gu-)sulmu, (gu-)hulmu "three": Vasconic *hirwu // ST *g-sum
Based on the reconstruction I'm aware of, ST *g-sum might reflect an earlier
numerical prefix *gu-, based on Nostratic *gul and *gulmu which seems to
have infused it to the word. At any rate, there are also two forms of the
word, one with *s- and another with *h-. Verifying with my notes, *s- is
supposed to yield *s- in Vasconic as in *sun "who?" becoming Vasconic *sil
(Basque zer). Since Basque seems to have "h" in /hiru/, the form must have
*h- or *ng- in Vasconic. The loss of *-m- can be explained easily since it
becomes
*-w- in *lawu(l) "four" (DC *limu).
Thus we must reconstruct **hir(w)u or **ngir(w)u. However, in the term
*sulrit "seven" we see a variation of *(h)ulrit in NEC. Since Nostratic
loses *s- we can't expect it here either. Although not the most secure
cognate, *hirwu would seem best fitted to DC *sulmu if we accept a variant
*hulmu for "three".
*mnrit "eight": Vasconic *midrac // ST *br(y)at
The -de- in /bederatzi/ needs little explanation. I view the original form
to be *mnrit, a compound word probably meaning something like "two from
(ten)". We find *-rit also in *sulrit/*hulrit "seven". The *-nr- in *mnrit
understandably became *-dr- in Vasconic. Later, an -e- was inserted between
"d" and "r" to satisfy syllabic constraints that frowned upon medial -CC- in
polysyllabic words like this. The ubiquitous *-i later hitched a ride.
Vasconic *m- regularly becomes Basque b and is not surprising since Basque
words that now have /m-/ are all foreign words or abbhorations. ST has
*bryat or *brat with regular contraction of initial syllables and *mr > *br
for the same reasons as *-nr- in Vasconic.
*pingu "five": Vasconic *bazil // ST *bnga
*hamsi "ten": Vasconic *hamzal // ST *dci
Note that /bost/ "five" ends in a final stop within an intriguing final
consonant cluster. A small number of words in Basque suspiciously have
similar rare final consonant clusters like beltz "dark". It is also often
suggested that they derive from earlier two-syllable forms (hence *beletz or
*belec).
In similar fashion, I propose that the /bost/ comes from a Vasconic form
*bazil with a suffix *-l (originally the DC pronominal *-n ?) that appears
in many of these numerals. The Vasconic *z, possibly a "zh"-sound, derived
from DeneCaucasian *-h-, *-ng- or *-s-. Well, *bazil became *bazili > *bazli
> *bazdi > Basque bost (with loss of *-i). Similarly DC *hamsi "ten" to
Vasconic *hamzal became *hamzali > *hamali > hamar but without syllabic
contraction perhaps because of intervening *-z- which would have caused
**-CCC-... YUCKEROO! Again, loss of final *-i. In ST we find *bnga with the
regular contraction of initial syllables but wrong vowel and *dci for "ten"
with the final tail of the original word preserved.
Last note, the Vasconic has *b from earlier DC *p because words in Basque
p-, like the words in m-, are not original to earlier forms of the language.
Again, p-words are foreign or recent innovations.
*sulrit "seven": Vasconic *surac // ST *snit
The *-t in both DC *sulrit and DC *mnrit became *-c. The *-t was always
preserved in ST as in *snit but was confused with the word for "two" g-hnis,
out of misanalysis. We should expect ST **sryat. In Vasconic the word is
*surac with the labial *-u- quality preserved as we should expect after the
loss of *-l-. So, *surac > *suraci > *surci > Basque zortzi.
*limu "four": Vasconic *lawu, *lawul // ST *bli
SinoTibetan *bli is an irregular formation by analogy with *bnga "five" but
is otherwise regular, stemming from a DC form *li without
the additional *-mu which is also found in DC *sul-mu "three". The DC form
*limu is to be found in Nostratic both as *lilmu AND *lil.
Basque sometimes has -r sometimes not. Until Basque makes up its damn mind,
I'm reconstructing both Vasconic *lawu and *lawul with numerical suffix.
>I think the final -zi is a suffix for numeral, isn't ?
Not proven in the least bit. It is only found in bederatzi and zortzi. If
you are going to develop this idea, you may as well understand that "tz" is
a single phoneme and thus you should be saying **-tzi. You'll have to agree
that the idea I present is much better or at the very least, more
comprehensive.
>Anyway, I am still convinced that ancient people coun't count.
"How ancient?", is the real question. I think it's unwise to avoid
linguistic evidence in favor of assumption of a negative that can't be
proven, don't you? I understand your arguement about numbers and the modern
cases you supply, of which I'm aware, but it would seem that our ancestors
weren't as mathematically simple as you think and it doesn't necessarily
have to be so just because of a few cultures here and there who don't feel
the need to invent any higher numbers.
>Besides, having a common word for a numeral throughout a language >family
>does not prove it was inherited.
True, it doesn't. In combination with grammatical comparison however, it
does, as the case of IndoEuropean shows us. I supply all of this for
perusal. What else do you need to convince you? When numbers undergo regular
sound changes with the rest of the words, they're probably inherited. If we
hesitate on this methodological principle, we'll be in the dark for
centuries to come.
>ST r ? who said that ? Never heard of it.
Well apparently someone I read thinks that ST had a uvular /R/, I'll find
out who soon enough. I distinctly remember this and I wouldn't bother making
it up. So I know you can't be too caught up on theory of SinoTibetan.
Perhaps this is not the case in Chinese which I agree would have to have a
coronal "r". What about TibetoBurman? Uvular or nyet?
>Anyway, we can continue discuss DC / AC and IE, but I suggest we
>discuss lower levels than DC. Which language family do you think is
>most closely related to chinese (apart from TB) ?
Lower than DC? Ugh. Well, TibetoBurman would be the closest but even I am
not too read on that family all too well. Taking TB away, there's nothing
but remote language families like Na-Dene and Burushaski-Yeneseian. They are
both vastly different from ST but my position is that ST simplified alot of
DC grammar. I believe Burushaski does have a causitive s-. Na-Dene languages
do use word class prefixes as well. ProtoAthabascan has *se- for first
person (DC *ti, not related to ST) and *ng@- for second person (DC *ngu).
The latter correlates with ST *nei (or *njyyeeaeX, if you will ;)
I don't believe that one can relate the pronouns found in ST with those of
Austronesian, of all things. Honestly, you have a better chance with
Australian languages :P DeneCaucasian remains the best bet and you haven't
been able to argue otherwise yet.
It would be easier if ST had many grammatical "sets" to go by other than
pronouns, like declension or conjugation, but even DC isn't like that.
Regardless, the numerals, even if they can't be used as a solid proof of
relationship with DC, show at the very least that these words are present.
This still shows a DC flavour in ST. It's about time you and your peers
acknowledge this.
>What I can do is give you a limited list of those words that look >really
>cognate between TB and AC. I cite only tibetan when there is >a cognate.
>[...]
>This list is just a first draft, you get new comparisons in >subsequent
>letters, I can't do everything from memory.
Interesting. Starostin appears to acknowledge your "hmyj?" to some degree
under his ST *me:jH with *sm[e:]j?...
Chinese: *sm[e:]j? fire.
Tibetan: me fire.
Burmese: mih, LB *m[e]jh fire.
Kachin: mji2 fire (only in compounds).
Lushei: mei fire, KC *me�i.
Lepcha: mi� fire
Kiranti: *mi�
If "hmyj?" is correct, then ST *hmejx would be better in line.
Starostin also supplies a similar form for "to become warm, hot, catch fire"
in his North Caucasian *=He:wxV(n) (Chechen =owxa, Batsbi =apxe^, Ingush
=`ajxa). One could perhaps re-reconstruct it as *=hipx. Comparison with ST
would yield a very tentative DC *himuh for "fire" (Don't cringe, I saw you
cringe!). It would show the DC *m > NEC *b/p thing as we find in NEC Libu
(lHe) "three" < DC *limu "four" or *b-xwiL (?wil?i) from DC *m-hutL "eye".
So, *Libu, *=hipx, *b-xwiL. Hmmm.... Just thoughts.
>I think body parts, basic verbs of action, are the most conservative
>words, not numerals. As for pronoms and grammar words, the issue is >not
>settled.
Why do you ignore *m-hutL which is a wonderfully preserved body part term?
It shows regular change to Vasconic *mik: (> *mik:i > *bigi > Basque begi)
by the way just as *tLu becomes Vasconic *k:u (Basque gu). What do you need
in order for you to be satisfied with a) Dene-Caucasian and b) Sino-Tibetan?
Anyways, the *m-hutL saga goes on. I was in the library today and glimpsed
at a book "Atlas of Languages" (1996). I found quite by accident, a page
where it openly talked about "Austric" and coincidentally cited an example
of "eye":
Austronesian MonKhmer Kam-Thai
*mata *mat *taa
It said that the Kam-Thai example shows it's loss of the word class prefix
in this Austric hypothesis (hmm...) Now compare my Dene-Caucasian *m-hutL
for "eye". You start to wonder. That makes two tentative correspondances
that would show *tL = *t between DC and these SEast-Asia-based languages:
DeneCaucasian Austric
eye *m-hutL(a) *m-hutL *m-ata
we *tLu *tLu *ta
... BUT... just a light-headed impulsive thought for now. Fast food for
thought.
>About uralic-IE, do you know Jorma Koivulehto's book "uralische >Evidenz
>f�r die Laryngaltheorie" ? I haven't had time yet to p�ck it >up in the
>library. This book treats old IE loanwords in uralic.
No, haven't seen. Is it recent? If it's in German, I'll only be able to pick
it up at my University of Manitoba library and they are bad at getting new
books, especially the linguistic kind. They still have to fix the leaky
ceilings and exterminate the ants in the men's washroom of the Isbister
building. Canada doesn't fund education like it used to. Hell it doesn't
fund anything like it used to, not even our world-famous health care :(
>Well, as for uralic-IE, although I am neither competent to juge IE or
>uralic reconstruction, I must admit that I was surprised, while
>learning finnish, by the pronouns min�, sin�, that are reminiscent of
>IE, and words such as vesi that you cited earlier.
So what's the problem? Why do you fight the obvious with a catch-all excuse
of borrowing? How many is enough?? Argh, I'll ask on a similar note as
above: What will it take for you to accept Indo-Uralic (or rather
ProtoSteppe)? The pronouns, the subjective/objective in Uralic vs
imperfective/perfective in IE, many verb roots, etc. show relationship. The
fact is that IE preserves roots far better than complete nouns. There are
very few ancient nouns that one can confidentally reconstruct (like *hwegwts
"eye") but they often derive from verb roots in the end anyway. Uralic too
is innovative and so the only hope is connecting verb roots and not whole
nouns. Thus Uralic *wete is related to IE *wed- "to moisten".
...But then again, with all these grammatical and vocabulary connections,
they can all be reduced quite lazily as borrowings, so I guess I and all
other comparative linguists can never win. It's a wonder that the
IndoEuropean hypothesis was ever accepted in the end.
How DID it get accepted then?
I can never get out of people that refuse these relationships what it is
that will make them accept the hypothesis, aside from whipping and other
sources of torture. I get to reasoning that nothing will. They just like to
be stubborn because that's the human way. :) Damn these humans, why did my
friends leave me on this planet, .... ah, erh, I mean, Gee, I can be
stubborn at times, yeah, yeah, just like a human being is supposed to be
(Sshhh, ixnay on the aliensway)
- gLeN
______________________________________________________