>Like all other universal preferences, the tendential
>association of sound with function can easily be overriden.
>It does have statistical validity, though counterexamples
>like those in your posting are easy to find. The phonetic
>encoding of meaning is in principle arbitrary.
Exactly, the phonetic encoding of meaning is very much arbitrary from one
language to the next. Language is a highly abstract form of communication.
We never question that there is inheirant meaning in computer bits, and yet
different machines use different combinations of these bits for the same
instructions. We can't possibly expect that phonemes should have any sort of
inheirant meaning whatsoever or that our DNA has somehow gone to all the
trouble of encoding this phonemic meaning in our genome.
These supposed universal phonemic resemblances are assumed to be
"tendencies" and are not proven because in order to prove them as
evolutionary tendencies conclusively, we must know how all languages relate
to each other in the first place. This knowledge is of course unknown and
remains to this day the holy grail of long-range comparison.
On the other hand, by taking examples within a particular language family
where linguistic relationship is uncontraversial between its members, we
find that no such tendency exists at all.
For example, Hindi. What's up with the pronoun /aap/? Doesn't look like
Hindi is following our imaginary universal trend now, does it? If it were,
we'd expect a statistically relevant number of pronominal replacements where
another word with the same phoneme took hold. Instead, pronouns even when
replaced with new forms appear to tolerate just about any phoneme or
Counterexamples are so easy to find simply because of the fact that these
universal googoo tendencies are most likely non-existant. They are readily
assumed without proof by those so bent against long-range comparison. There
is simply no logical way to rule out the possibly that Uralic and
IndoEuropean share 1ps *m- due to relationship rather than universal
tendency. Neither can we sift out whether all the world languages who have
nasal phonemes in their 1ps pronoun are a result of tendency or
Outright uncertainty - this is the reality of this hypothesis. So we can't
possibly pretend that this tendency is written in stone and exists with 100%
or even 50% certainty. The likelihood of the googoo theory being true is too
overwhelmingly small to contemplate... which is not to say that other
universal tendencies don't exist, just that this particular "tendency" is
worthy of the waste disposal.
>Even highly marked segments may occasionally occur in grammatical
>morphemes, but that won't happen very often. One could say
>that unmarked segments have a selective edge on other
>consonants in this environment.
I agree with backgrounding.
>Big and small things are often contrasted using back versus
>front vowels (and/or coronal versus labial or dorsal
>consonants), the volume of the oral cavity symbolising size,
>as in itsy-bitsy or teeny-weeny versus boom or googol. This
>correlation is, I suppose, intuitively obvious, though any
>number of counterexamples resulting from historical accident
>could be quoted.
This doesn't appear to be intuitively obvious to me. Another assumption that
can hardly be provable in a rational way.
>To give a more iconic example, the word for a cuckoo is very
>often something like /kuku/,
Many world languages don't even have a word for "cuckoo". This phenomenon
has nothing to do with the googoo hypothesis and is called onomatopoeia. Of
course, in long-range comparison one must be extra careful in recognizing
potentially onomatopoeic terms like these.
>The same holds for *mama- and other nursery noises.
Not exactly. Etruscan and Karvelian, for example, do not conform to the
"mama" pattern. The "mama" phenomenon can't be used as proof of some
inheirant phonemic symbolism either. However, I do agree that there is
always the chance that these much-used terms can be borrowed or at the very
least "backgrounded" over time. So, by themselves, they can't be used to
prove genetic relationship between languages. In concert with other proof it
can be an extra boost to genetic relationship though. Afterall, IE *mate:r
is praised as an uncontraversial IE term, no? :) I don't see why a Nostratic
*ama can't eventually be lauded in the same way over time and through
>Nobody says that any 1sg pronoun _must_ contain a nasal, but
>they _do_ contain nasals often enough for linguists to
>wonder why that should be the case. I gave the example of
>proto-Bantu *mi-, and there are plenty of New World examples
>(discussed at length in Campbell 1997 [American Indian
>Languages: the Historical Linguistics of Native America.
>Oxford: OUP]). There's nothing "bow-wow" or "goo-goo" about
Of course there are many examples (ProtoAlgonquian 1ps in *n-: Objibway nin,
Cree ni-, etc). But I'm shocked at you, Piotr. This couldn't be any more
googoo an idea! Linguists are not yet knowledgeable enough to be an adequate
judge of what is due to genetic relationship and what is due to universal
phonemic tendency on a global scale. This fact alone automatically provides
reasonable doubt to the theory. The rest of the doubt lies in the
unintuitiveness of the meanings often attributed to these phonemes like /m/
for "motherhood" or "safety" or "selfishness" or perhaps it's "madness"??,
/d/ for "demonstratives"??, etc. It's "anything goes" with unsubstantiated
pet theories like that.
Of course I definitely am aware that nasal phonemes are exceedingly common
for the 1ps singular but is it truely due to phonemic symbolism or is it due
to the strong tendency of these pronouns to resist replacement in general
over exceedingly long periods of time? I side with the latter.
Yes, I'm aware of the cornucopia of 1ps forms in Japanese and Vietnamese for
not only person and number, but also gender, social rank, age, etc, which
surely are relatively recent replacements that drown out any trace of any
originally "nasal" pronoun but overall it would appear that replacement is
not a common, everyday occurence of world languages, happening every few
millenia or so in any particular linguistic line.
>This is not to say that pronouns are useless when language
>relationships are discussed. But they should be used with
>more caution than shown by Bomhard or Dolgopolsky (they both
>explicitly insist that pronoun resemblances of the "mi/ti"
>kind are particularly strong indicators of genetic
You're twisting things a little here. Pronouns can be a bit of a useful
indicator, albeit not a terribly strong indicator, true. However, when a
three-person pronominal set is comparable between "unrelated" languages, the
indication of relationship is far stronger. The languages I call ProtoSteppe
(Bomhard's Eurasiatic) aren't just "mi/ti" resemblances. Ignoring the added
declensional relationships, they not only have an entire set intact
consisting of three persons in the singular [*mu/*tu/*a,*i] with plural
forms derived directly from the singular forms in Uralic, IndoEuropean and
Altaic (and probably Etruscan too) but they also show a very characteristic
conjugational system through suffixation of these same pronouns together
with another very phonemically different set that are not derivable from the
Bomhard does more than offer empty "mi/ti resemblances" by mentioning this
split conjugational system and endings of these languages. Unfortunately, he
doesn't explicitly outline how this came about (so I had to figure it out by
myself, boohoo, sob).
Do any African or New World languages conform to the pattern I've just
mentioned above? How likely would all these characteristics combine together
by chance in unrelated languages?
I guess what we both agree on then is that making quick and easy connections
by eye without a full analysis of the implications is the wrong way to go.
>Whatever one thinks of the validity of Nostratic, I'm sure you agree >that
>Ruhlen's or Ryan's Proto-World pronominal systems hardly deserve >to be
>discussed at all.
Agreed. In fact, Ryan inevitably runs the average thinking individual into a
logical brick wall where one is forced to contemplate how such a monstrosity
of a proto-language could possibly have developed in itself. Rather than
solving the language origin problem, he presents us with an even greater
problem. How do we get from grunting Australopitheci to this?? The ol'
chicken and the egg problem.
What Ryan and others of his ilk can't possibly fathom is a world where
language is inheirant to all living animals, evolving billions of years
before homo sapiens ever walked on the earth. The forms and complexity of
language have changed from the single-celled amoeba to humankind, but
everything remains the same in the end. To trace the origins of language is
to trace the very origin of life itself, quite literally. Eventually, we'll
all realize that Neanderthals were just as fully capable of communication
within their society (whether it be spoken or gestural) as Homo Sapiens and
other primates are, but until then, there will be many Patrick Ryan's out
there, doing their schtick, grappling in the dark for simplistic answers.
Nighty night yo'll.
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com