>However, Glen, geographical proximity also increases the possibility >of
>borrowing. Also, the match is rather less than perfect and holds
>between Mongolic and Finno-Ugric rather than Altaic and Uralic.
True, but I think I have to touch a bit more on our differences of
methodology and create a really long response again. :)
First, I try to keep very distinct the methodologies of physical science
versus theoretical science. Sometimes I get the impression that you blend
the two together into a strange hodgepodge system where questions enter a
revolving door and mysteriously vanish.
It's easy to "prove" things in a physical science like Physics, Mathematics,
Chemistry, etc. because they all deal with tangible things and the answers
to a given question are always absolute responses: "Yes", "No", "10%",
In theoretical sciences like Psychology, Sociology or Comparative
Linguistics (and all the humanities), there is nothing tangible that one can
prove. Like I tell John repeatedly, you won't find a mummy with a piece of
prehistoric language clutched in his hands.
So continuing to use the word "proof" as you do seems nonsensical and
arbitrary to me. "Proof" in such sciences is a matter of opinion only, since
it depends on the individual's tolerance to probability. All one can do as a
theoretical scientist is to travel down the most probable paths to arrive at
a conclusion, which by nature will forever be probabilistic and uncertain to
a degree... that includes our beloved IE hypothesis, of course. The point is
to make one's "hypothesis", which is a theoretical science's version of
"proof", the _most_ probable given the current information.
To give an example, if we were to ask ourselves "What is the closest
language related to Gilyak?", how might we go about this answer? If we went
about it your way, Piotr, by "proving" somehow that Gilyak is related to
another language, we'd certainly never get anywhere because by insisting on
"proof", we are insisting on an absolute answer that can't exist because of
the very nature of the question. Thus, as you do, we say "Don't know" or "We
can't say" which isn't an answer at all.
We need to find the _most probable_ answer of those available in order to
properly answer it (which is the whole point of asking a question!). From my
knowledge, the answer here appears to be "any one of the Turkic, Mongolic,
Tungusic, Japanese and Korean languages" according to common consensus,
which itself is based on similarities of pronouns, grammar and vocabulary
between those languages and Gilyak. Likewise, despite the unknowns
concerning Etruscan, the most probable answer for its origins remains linked
with the languages called Lemnian, Rhaetic and ultimately, IndoEuropean.
I cannot agree that a blank answer is a scientific answer at all because the
whole point of science is to ask questions as well as _answer_ them. Since
theoretical sciences can only have probabilistic answers to its questions,
an answer with 5% probability is still 5% better than no answer at all,
which is the most improbable solution one can put forth. With any question,
there is ALWAYS a most probable answer, even if it is a list of equal
possibilities. The key is to understand and acknowledge the relative
probability of a given answer.
Now, I know what you're thinking. You're wondering if Gilyak's similarities
to other Altaic languages is a matter of genetic relationship or simply
borrowing and how would I address that. As far as I'm concerned, two
languages which share a common set of pronouns and other grammatical
paradigms (declension, conjugation, etc), ignoring common vocabulary, is on
average _more_ probably a product of inheiritance rather than areal
influence. If we can't judge a relationship securely on common every-day
vocabulary with its high possibility of loanwords, grammar seems like the
securest connection we could ever make. In this case, the most probable
solution is that Gilyak is most closely related to the Altaic languages and
Korean until further evidence dismounts this most probable view.
In this case, it would be more complex to say that Gilyak isn't related to
anything, since it most certainly HAS to be, unless we have sufficient
reason to believe that Gilyak is a language distinct from all other human
languages! We also can't say that its relationship is "unknown" since the
possibility of Altaic relationship is already laid out on the table. Since
it is the only relationship I've heard of connected to Gilyak and since
these connections are readily visible, it remains the sole and most probable
idea. The null hypothesis doesn't serve anyone.
>I'd apreciate a more detailed account of the sound changes in question
> >plus some good comparative material to substantiate them.
Alright, for instance with AltaicGilyak languages, there are very distinct
tendencies that set it apart from other Steppe language groups:
- merger of *kW, *k:W and *gW to *q (Altaic *k)
Steppe *kWi "who?"
> AG *qi
> Altaic *ki (Turkish /ki/, CJap /ka/)
Steppe *gWir "to heat up" (cf. IE *ghWer-)
> AG *q�r-�
> Altaic *k�r� (Turkish /k�z-/)
- softening of *-t to *-c and then *-r
(IndoTyrrhenian-AltaicGilyak isogloss perhaps?)
- vowel harmony within a root caused first by
progressive (de)labialisation and then
regressive (de)palatalisation, thus arriving
at a six-vowel system with four high vowels
and two low ones
- softening of all remaining plain voiceless stops
(at least within roots dominated with labial vowels)
*p > (*h)
*k > (*h)
*t > *c
Steppe *tun "you" > AG *cun [*t > *c]
Altaic *cun (Mong. c^i, Turk. sen)
Gilyak *ci-/ce- "your"
Steppe *kit:u "five" > AG *(h)�t� [*k > (*h)]
Altaic *�t� (Mong. tav, CJap itu-)
- a shift of tense stops to plain (*t: > *t)
and preservation of voiced stops
- change of *r- to *N- (Altaic *n-)
Then, within Altaic itself, there is a tendency to change *m- and *n- to *b-
and *d- in "labial" roots:
AG *mun "I" > *bun (Turkish /ben/, Mong. /bi/)
AG *qur(mun) "3" > *kur(bun) (CMong. qurban)
AG *n�r "4" > *d�r (Turkish /d�rt/, Mong. /d�r�v/)
but within "non-labial" roots... nada, zilch, zippo...
AG *mir "we" > *mir (Mong. manai "ours")
AG *nana "seven" > *nana (CJap. nana-, Tung. *nadan)
Note also that Turkish does kooky things like changing *-r to *z (*mir "we"
> biz) or making *t- into /NULL/ or /y-/ as in Turkish /on/ "ten" (*tuwon)
or yar� "half"/y�rm� "twenty" (*t�r� "two"). No wonder it doesn't look very
>If you select the one daughter that seems to have a potential cognate >of a
>word attested elsewhere (and ignore all the sister branches), you >ARE
>reaching down. [...]
>Regarding the Dravidian, Egyptian and Semitic reflexes of *munri, I'd >like
>to see more concrete information, something I can >evaluate. "Might be" is
>not enough AFAI'm concerned.
As I say, in re of Steppe *munra, there may be a correlation with Dravidian,
assuming Eurasiatic *munra > *minTa > *(w)iNT-u > Dravidian *eNTu. Also
perhaps *mi- "what?" > *(w)I- > Dravidian *ya:-, thereby merging in form and
function with *ya- "what?". Dunno. Just a thought.
So... If I understand you right... In the case of Nostratic, it isn't
"reaching down" since there is more than one sister language. In the case of
reconstructing Steppe, it is?? At any rate, note that other reconstructed
Dravidian numerals also conform nicely to its sister's numerical set:
1 *t:u *oru
2 *t:ui *iru
3 *kul *mu-
4 *nil *na:lu <--- !!!
5 *kit:u *cay-
6 *ru *a:rru <---
7 *rara *eRu <---
8 *munra *eNTu <--- if *m > NULL
9 *nukura *toL
10 *t:ukam *paTTu
(capital letters in Dravidian indicate "retroflex")
As for Semitic/Egyptian attestation, I can only surmise that there could be
a link if the phonetics of "eight" were affected by "six" and "seven". The
latter two are numbers that start with a sibilant. If we were to whip up an
ad hoc reconstruction of AA numerals for a moment to illustrate:
(Stage I) (Stage II)
six *sWats- sWats-
seven *sab- sab-
eight *mun- samun-
Just a thought... a very, very tentative thought. And I'm sure that, in no
way, does Nostratic *r- correlate with AfroAsiatic initial sibilants :)
>A reliable etymology should link more than just a couple of nodes in a
> >large family tree. Non-robust agreement may well be accidental.
This is true, I know. However, sometimes it's not possible to provide robust
evidence. According to the probabilistic manner in which I think, what I
give is still the likeliest answer currently from what I know, even though
the answer is of perhaps lesser probability because of your valid arguement.
Concerning IE's attestation of *kit:u...
>You don't quite address my objection. I'm not saying that irregular
> >development is unthinkable, but that its assumption destroys a case
> >based on weak evidence anyway.
Tough crowd. I can't think of how I might remedy this.
More on *kit:u "five"...
>More detail please. How exactly do you derive these Altaic forms? Is >*t
>the only tangible trace of *kit:u/*kWut:u?
Unfortunately, in AltaicGilyak, yes. That is, if you mean that consonants
are the only tangible evidence for you. It's very easy for a VCV word to
become CV, but the initial vowel of Classical Japanese (itu-) shows that
there was an initial syllable present unless you can explain the vowel
differently? The Steppe *k-, being a voiceless stop, likes to aspirate
itself out of existence in AltaicGilyak so I can't help here.
>You bet I will. I also understand that your theory is a tentative >proposal
>(what theory isn't?), but I gave you my reasons for believing >that a
>tentative proposal should also meet certain reasonable
>standards. If it doesn't, people will reject it out of hand even if >you
>threaten to insult them.
I can only hope that this judgment isn't final - afterall, my theory isn't
either. I will, and always do, try to strengthen my ideas whenever I can.
Piotr, after submitting a long but interesting response, drudges up past
quotes like this long and ugly one that I didn't think anyone actually read:
> Glen: ... The evidence so far IS weak. This is because the >ones
>that are competent enough to provide solid evidence have thrown >the
>Nostratic Hypothesis out the window because it is too taboo >and "weak" for
>their liking, rather than actually honestly examining >the connections and
>putting aside irrelevant failures in the past by >lunatic amateurs. We all
>know how political linguistics can be. ><snip> ... </snip> Still, I
>maintain that the lack of success >regarding the Nostratic Hypothesis stems
>mainly from apathy and >politics more than anything. I hate to see this go
Piotr responds to the above quote of mine:
> Piotr: Attributing unspecified political motives to your >opponents is
>a very political thing to do -- a mere rhetorical trick, >and a very unfair
>one at that, just like insinuating that anyone who
>disagrees with you must be dumb, lazy or apathetic.
You're right but nonetheless, Nostratic or DeneCaucasian is still somewhat
taboo in comparison to IE et al, I think, because it's so easy to lump these
ideas with ProtoWorld, IE-Quechua, alien conspiracies and the like. Hell,
"Nostratic" has an uncanny resemblance to "Nostradamus" afterall.
> Piotr: Unfortunately, in Uralic the K-set refers to humans (Hungarian
>ki, Finnish kuka 'who?') and the M-set to "things" (Hungarian mi, Finnish
>mik�� 'what?'). I wonder how you explain this >mismatch between Uralic and
I don't. You're very correct - I've shamelessly mixed them up. <:S Steppe
*kWi- is "animate" and *mi- is "inanimate". Whoops! I will correct this
error immediately. Regardless for this ugly faux pas, the arguement for
animate/inanimate contrasts in Steppe still hold and nothing is lost. I
believe the source of my confusion was Semitic where the tricky *mi-
interrogative also exists for animate purposes.
>Glen:...but how do we falsify NON-relatedness, Piotr? That seems >frankly
>like a fruitless game in itself.
>Piotr: How? By showing beyond reasonable doubt that observable >systematic
>resemblances are genetic rather than areal or accidental. >It can be done.
>If the game were fruitless, there would be no IE
>reconstruction; and if you yourself believed it to be fruitless, why >the
>heck should you be wasting your time playing it?
I think IE as a theory exists the way it does because, in the beginning, the
methodology was different, based on probability (as it should be), rather
than "100% proof" which is impossible to come by in this field. The amazing
advancement of the IE hypothesis over the past 150 years shows the efficacy
of a probabilistic method. Nostratic is hardly as developed as IE, not
because the methodology used is any different, but because IEists now seem
to have turned the tables and are in expectation of hard "proof" based at
most on an arbitrary definition of it.
Was IE rung so deeply in the ringer as Nostratic is now? At what point was
IE "proven"? Year? Month? Day?
>Even you should be exercising scepticism with one half of your mind >while
>the other half is doing stratospheric flights of fancy. The >question "How
>can I be sure I'm not imagining things?" can't be asked >too often.
Believe me, I adopt and reject many ideas every day. I always ask if I'm
imagining things. Funny thing too - sometimes a voice answers back and I'm
the only one in the room! But don't worry, I'm close to the doctor's office
if anything goes wrong :)
>As for the ad hominem part, ego te absolvo if you do the same.
No harm done,
No tear shed.
Now if you'll excuse me,
I'm going to bed.
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com