Ruhlen's way with words
Larry Trask
dishonesty. Now, as time permits, I'd like to post a few examples of
his so-called "linguistic work".
I'll start with his methodology. Ruhlen maintains that chance
resemblances among languages are vanishingly rare -- of "truly
minuscule" probability, in his words. He attempts to demonstrate this
by performing a calculation.
He attempts to calculate the probability that fourteen unrelated
languages will all exhibit the form PUT for the meaning 'vulva'
entirely by chance. He calculates that this probability is about one
in ten octillion.
Well, his calculation is completely wrong, since he has no idea how to
do probability calculations. Among other failures, he fails to
realize that it makes a difference whether we are calculating for 14
languages out of 14, or for 14 languages out of 14,000. He also
assumes that each language has only one word for 'vulva'.
But these are minor quibbles. The real idiocy lies elsewhere. In
doing his calculation, Ruhlen counts *nothing* as a resemblance except
perfect identity. So, only PUT counts as a resemblance, while even
POT is rejected as not constituting a resemblance. And, of course, he
counts only words meaning 'vulva', and nothing else. With these
maximally rigorous constraints on what counts as a resemblance, he has
no difficulty in concluding that chance resemblances are very rare.
But then he presents his evidence that all the world's languages are
related and descended from a common ancestor, "Proto-World". One of
his PW "etyma" is PUTI 'vulva'. And what does he count as a match for
this PUTI?
Anything and everything. Among many other forms, he counts the
following as "resemblances": bid, -ngoboti, bitt, pwt, futo, fido,
fugi, pundu, pos, piru, pidy, hütügün, potorro, sebud, besh, -psh,
lapus, dibis, vith, hapichatt, petaistapcca, poru, silfhuta,
hibitikope, shapsi, pirri... You get the idea.
And, of course, he doesn't insist on the sense 'vulva'. Among the
senses he accepts as matches are these: womb, anus, vagina, genitals,
penis, hole (in general), buttocks, girl, whore, rectum, colon, tube,
pubis, and clitoris.
Naturally, since he is working with no constraints at all, he has no
difficulty in collecting a long list of words which he counts as
"resemblances" to PUTI 'vulva', and therefore as obvious "cognates"
demonstrating common ancestry.
So, he works with maximally rigorous constraints in calculating the
probability of chance resemblances, but he works with no constraints
at all in collecting "cognates". Then he tells us that his collection
of resemblances must be true cognates because he has proved that
chance resemblances are virtually non-existent.
You can decide for yourself whether we're looking at monumental
stupidity or at flagrant dishonesty. It must be one or the other --
or maybe both.
Back another day with some examples of the master's technique.
Larry Trask
lar...@sussex.ac.uk
I
lar...@sussex.ac.uk (Larry Trask) wrote:
>Well, his calculation is completely wrong, since he has no idea how to
>do probability calculations. Among other failures, he fails to
>realize that it makes a difference whether we are calculating for 14
>languages out of 14, or for 14 languages out of 14,000.
More accurately, 14 languages randomly picked (out of 6000 or whatever)
versus 14 languages "cherry-picked" _after_ they are found to fit.
Another very important issue is that Ruhlen (following Greenberg) uses a
calculation which applies to _independent_ samples. This not true for a
lot of Greenberg's, and Ruhlen's, examples. To explain what I mean:
Suppose we want to demonstrate that Nahuatl <teotl> (meaning roughly
"god", as in Teotihuacan; actually deriving from something like
"awe-inspiring") is related to Greek <theos>. Well, there're only two
languages here, and this could be a chance similarity. So instead, we
compare Nahuatl to IE, and we have Latin <deus>, Lithuanian <dievas>,
Sanskrit <devah>, etc. etc. -- six languages, and counting. With so many
languages, surely this couldn't be a coincidence...
Likewise, I could demonstrate the Greek-Nahuatl connection alone, using,
say, 7 Nahuatl dialects and 7 Greek dialects.
(The deus/teo-tl similarity didn't escape the missionaries' attention,
either, and they found it quite useful.)
I think the above two issues can be a more fundamental problem for
Greenberg-style mass comparison than semantic/formal laxity. Suppose we
pick a meaning in 10 random, independent languages. Suppose the form and
meaning match, fairly laxly, across all these languages. Suppose that
the probability of a random match between any two languages is a giant
75% (due to said laxity). Then the probability of all these words
matching by chance would be 0.75 ^ 10, which comes to only 5% or so.
That said, Ruhlen is being so lax, that the probability of chance
similarity in his samples might be even more than 75%.
Jarel
>More accurately, 14 languages randomly picked (out of 6000 or whatever)
>versus 14 languages "cherry-picked" _after_ they are found to fit.
Here are my random 14 languages:
Margany
Kitanemuk
Arem
Babuyan
Bashkarik
Lenyima
Baropasi
Abasakur
Dohoi
Gwandara
Molo
Mbesa
Arawete
Khoirao
>Well, there're only two languages here, and this could be a chance similarity.
So instead, we compare Nahuatl to IE,
Instead of comparing Nahuatl and Proto-IE, wouldn't Proto-Uto-Aztecan or just
Proto-Aztecan be a better comparison with Proto-IE?
Jarel
Sous ses lois l'Amour veut qu'on jouisse
D'un bonheur qui jamais ne finisse;
Tendres coeurs venez tous
En jouir avec nous.
-Jean Galbert de Campistron (1686)
from the pastorale Acis et Galatee
by Jean-Baptiste Lully
I
nah...@aol.comnotojunk (Jarel) wrote:
> Nowhere said:
>
>>More accurately, 14 languages randomly picked (out of 6000 or whatever)
>>versus 14 languages "cherry-picked" _after_ they are found to fit.
>
> Here are my random 14 languages:
>
>Margany, Kitanemuk, Arem [...]
Right. So even if you had full dictionaries for these 14 languages
(which I assume belong to different families), and picked some meaning
at random, it'd be very unlikely to share form across all 14. If you
used reasonable standards for similarity in meaning and form, you
wouldn't find any words at all shared across the 14.
>>Well, there're only two languages here, and this could be a chance similarity.
>>So instead, we compare Nahuatl to IE,
>Instead of comparing Nahuatl and Proto-IE, wouldn't Proto-Uto-Aztecan or just
>Proto-Aztecan be a better comparison with Proto-IE?
I think you missed my point. If we compare PUO and PIE (which I am sure
have some chance similarities between them) that would be a binary
comparison, not the 'multilateral comparison' which Ruhlen claims to be
using.
Jarel
>I think you missed my point. If we compare PUO and PIE (which I am sure have
some chance similarities between them) that would be a binary comparison, not
the 'multilateral comparison' which Ruhlen claims to be using.
OK, I do miss the point. Who would use such a thing as this "multilateral
comparison"? Is this a joke? From their lists of comparisons, most contain
many language universals such as nursery words and onomatopoeia. Using
Ruhlen's or Greenberg's methods, I could connect any two languages in the
world.
I
nah...@aol.comnotojunk (Jarel) wrote:
> OK, I do miss the point. Who would use such a thing as this "multilateral
>comparison"? Is this a joke?
Greenberg (who I believe coined the term), and Ruhlen after him, among
others. And no.
As I understand Ruhlen (I don't have Greenberg's book), Greenberg's idea
was to look at many languages together, as oppposed to more modest
attempts to connect a few groups of American languages at a time. They
claimed that the more languages you look at, the lesser the effect of
chance similarities. I am saying that the effect is exactly the opposite.
> From their lists of comparisons, most contain
>many language universals such as nursery words and onomatopoeia.
A minority of their forms, not that they should have been included
anyway.
>Using Ruhlen's or Greenberg's methods, I could connect any two languages in the
>world.
Or more easily, any 6000 languages in the world.
David Thomas
nah...@aol.comnotojunk (Jarel) writes:
> Nowhere said:
>
>>I think you missed my point. If we compare PUO and PIE (which I am sure have
>some chance similarities between them) that would be a binary comparison, not
>the 'multilateral comparison' which Ruhlen claims to be using.
>
>
> OK, I do miss the point. Who would use such a thing as this "multilateral
>comparison"? Is this a joke? From their lists of comparisons, most contain
>many language universals such as nursery words and onomatopoeia. Using
>Ruhlen's or Greenberg's methods, I could connect any two languages in the
>world.
From what I understand, their methodology is a joke... at least, to serious
linguists.
- Vae
Jarel
>A minority of their forms, not that they should have been included anyway.
Words like:
milk, breast, nipples, etc
blow, spit, suck, nurse, split, drink, lick, smell, etc
mother, father, etc
pronouns
rooster, dog, etc
All of these should be excluded. If I remember, much of their wordlists
consisted of words like these above.
For good etymologies, give me some lower numerals! There is a good
correlation between numeral cognates and languages being related. One finds
numeral cognates in Iroquoian, but not in a larger unit like Keresiouan (which
is not valid).
>Or more easily, any 6000 languages in the world.
I'd say there are at least 9,000 languages. 6,000 is too low.
Douglas G. Kilday
"Jarel" <nah...@aol.comnotojunk> wrote in message ...
>
> [...]
>
> For good etymologies, give me some lower numerals! There is a good
> correlation between numeral cognates and languages being related. One
finds
> numeral cognates in Iroquoian, but not in a larger unit like Keresiouan
(which
> is not valid).
Whoa! Be careful with lower numerals, which can be borrowed! We all know
that Hungarian has Iranian numerals. If memory serves, Nahali has native
'1', but '2-5' are borrowed from Dravidian, and '6-10' from Indic. The forms
for '6' and '7' are strikingly similar in IE, Hebrew, Coptic, and Etruscan,
but that is not enough to construct a non-kooky Indo-Afro-Asio-Etruscan
theory. It is likely that trading in countable units (such as cattle)
between alloglottic societies encourages replacement of syllabically lengthy
numerals by shorter loan-numerals.
DGK
John Atkinson
"Jacques Guy" <jg...@alphalink.com.au> wrote...
> Larry Trask wrote:
>
> > You can decide for yourself whether we're looking at monumental
> > stupidity or at flagrant dishonesty. It must be one or the other --
> > or maybe both.
>
> Both. There is no reason to believe that those two cardinal
> virtues are mutually exclusive.
>
> But I am surprised: you sound as if had discovered
> Ruhlen only this morning!
>
> However, I would tend to ask the same question about
> his publishers: monumental stupidity, flagrant dishonesty,
> or both?
Neither. They are in the business of selling books. They decided these
books would sell. They were right. They would have been stupid (or
dishonest to their shareholders)
if they *hadn't* published them.
So now we come down to the people who buy them. Monumental stupidity, ...
or what?
John.
Des Small
> "Jacques Guy" <jg...@alphalink.com.au> wrote...
>
> > Larry Trask wrote:
> >
> > > You can decide for yourself whether we're looking at monumental
> > > stupidity or at flagrant dishonesty. It must be one or the other --
> > > or maybe both.
> >
> > Both. There is no reason to believe that those two cardinal
> > virtues are mutually exclusive.
> >
> > But I am surprised: you sound as if had discovered
> > Ruhlen only this morning!
> >
> > However, I would tend to ask the same question about
> > his publishers: monumental stupidity, flagrant dishonesty,
> > or both?
>
> Neither. They are in the business of selling books. They decided
> these books would sell. They were right. They would have been
> stupid (or dishonest to their shareholders) if they *hadn't*
> published them.
Not necessarily. Publishers and imprints are brands, and the damage
to the brand's prestige resulting from publishing rubbish may outweigh
the immediate gain.
I was curious enough to check Amazon to see who publishes Ruhlen's
stuff, and John Wiley & Sons won't be selling me any linguistics books
in the future that haven't been vouched for by a reliable source.
> So now we come down to the people who buy them. Monumental stupidity, ...
> or what?
Orbiting mind-control lasers.
Des
has his aluminium foil lined hat on, of course.
--
des....@bristol.ac.uk
Larry Trask
[on Ruhlen and Greenberg]
> From what I understand, their methodology is a joke... at least, to serious
> linguists.
Absolutely right. But the broad nature of "multilateral comparison"
is hardly the worst of it. In assembling his comparanda, Ruhlen is
grossly incompetent and grossly dishonest. He chooses his conclusions
in advance, and then he goes looking for confirming data. He
suppresses all unwelcome data, and picks out only the things he likes.
Even the data that make it onto the page are not presented honestly:
they are twisted, mangled and mutilated in order to force them into
the required fit.
I propose to work through Bengtson and Ruhlen's co-authored paper
'Global etymologies', presented in Ruhlen's 1994 Stanford book. This
is perhaps the central publication in defense of "Proto-World". B&R
present 27 putative "words" from "Proto-World", and then present
supposed descendants of those items in the world's languages. I'll
look at the eight items for which they present Basque comparanda.
Background. An ancestral form of Basque is sparsely recorded in the
Roman period. Otherwise, Basque is recorded in fragments from the 9th
century on, and the fragments become substantial in the late medieval
period. The first long connected text is a personal letter of 1537.
The first book was published in 1545, since when publication has been
continuous. The phonological prehistory of the language has been
reconstructed in great detail back to the Roman period.
4. CHU(N)GA 'nose; to smell'
Basque <su-dur> 'nose', <sun-da> 'smell'
Do you know of a language in which 'nose' and 'smell' are derived from
the same stem? B&R are just piling up vaguely related meanings in
order to make chance resemblances easier to find.
Basque <sudur> is the universal word for 'nose', but the segmentation
offered is arbitrary and indefensible. We might as well segment
English 'nose' as 'no-se'. And <sudur> is supposed to continue
CHU(N)GA, is it? How convincing.
But <sunda> is another matter. First, the gloss is wrong. The word
doesn't mean 'smell' in general: it means 'stench, stink'. Second,
the word is not general: it is confined to a small area in the west.
Third, it is not recorded before about 1800 -- rather late by Basque
standards. Fourth, the segmentation offered is preposterous, and
there is no more reason to link the word to <sudur> than to <suge>
'snake'. English 'ear' and 'hear' are a much more striking pair, but
are unrelated.
And this is by far the *best* of the Basque comparanda.
7. KATI 'bone'
Basque <gar-khotx(e)> 'nape' (<gara> = 'skull')
The Basque word for 'bone' is <hezur>, from earlier *<enazur>. This
is not helpful, so B&R go trawling.
The item selected is preposterous.
First, Basque has a number of regional words for 'nape', all of them
compounds and none of them showing signs of any antiquity. The one
presented here is confined to the Zuberoan dialect at the far eastern
margin of the country, on the French side, and is not recorded before
the late 19th century.
Second, no Basque word of any antiquity begins with /k/. In the Roman
period,
*/b g/ were the only plosives that could occur word-initially.
Nothing with initial /k/ can possibly be of any antiquity.
Third, the nape is not a bone, and it's not even very bony. The
proposal is semantically ludicrous.
Fourth, the element <gar-> is familiar; it means roughly 'high place,
high part', and by extension (in compounds) 'head', but not 'skull'.
Fifth, the first thing any halfway competent linguist would do in
contemplating an etymology for <garkhotxe> is to ask whether Zuberoan
has a word <khotxe>. It does. And what does this word mean? It
means 'basin, bowl'. And what shape is the nape? It's concave.
The standard (but unfinished) etymological dictionary of Basque
derives <khotxe> from Occitan, and it analyzes <garkhotxe> as
containing this <khotxe>. All this information is suppressed by B&R
-- not that they looked for it anyway. They found <garkhotxe> in a
dictionary, and their investigations stopped at that point. They made
no effort to find out anything about the word. They never do.
That's enough for now. Back later with some more gems. I can assure
you the story gets more and more deranged as it goes along.
Larry Trask
lar...@sussex.ac.uk
Larry Trask
> "Jacques Guy" <jg...@alphalink.com.au> wrote...
>
> > Larry Trask wrote:
> >
> > > You can decide for yourself whether we're looking at monumental
> > > stupidity or at flagrant dishonesty. It must be one or the other --
> > > or maybe both.
[JG]
> > Both. There is no reason to believe that those two cardinal
> > virtues are mutually exclusive.
For some reason, Jacques Guy's posting has failed to show up on my
reader, and I've seen only this quotation.
> > But I am surprised: you sound as if had discovered
> > Ruhlen only this morning!
Hardly. I've been having the Ruhlen and Bengtson experience for many
years. But I have the clear impression that the extent of Ruhlen's
dishonesty is not well known, even among historical linguists.
Incompetence, certainly, but not dishonesty. I've been scratching my
head for years for a way of publicizing his fraudulent ways, but
without success. I tried to squeeze an account into the popular book
on language change I'm doing for Cambridge, but the editor wouldn't
let me do it -- "too technical", even though the anonymous readers
loved it.
So I'm just letting off a little steam here on sci.lang.
Larry Trask
lar...@sussex.ac.uk
Larry Trask
9. KUAN 'dog'
Basque <haz-koin> 'badger' (lit. 'bear-dog')
B&R apparently believe our ancestors domesticated dogs before they
learned to speak. How else could the first human language have a word
for 'dog'?
The Basque for 'dog' is <(t)xakur>, in origin a diminutive of earlier
<zakur>, now specialized to 'big dog'. This has apparently displaced
another word, <or>, recorded in the 16th century at both ends of the
country.
Not useful. So B&R seize on the word for 'badger', which they
declare, without support, to be a compound of Basque <hartz> 'bear'
and a fanciful word for 'dog'.
Now, once again, no item beginning with /k/ can possibly be ancient in
Basque -- not even a real one, and this *<koin> is strictly the
product of B&R's overheated imagination.
As it happens, the Basque word for 'badger' occurs in 25 regional
variant forms -- but *<hazkoin> is not one of them. B&R have invented
this form. Are they so stupid that they can't copy a word off the
page correctly? Or have they simply decided that the real forms are
not good enough for their purposes, and replaced them with a phony
form they like better?
I guess the second. In another article, Ruhlen cites the word as
<harzkoin> -- which also doesn't exist, but which makes his etymologhy
look more plausible.
Anyway, it has been established that all the variants go back to an
earlier
*<azkone>, which is therefore the only form available for comparison.
But this form looks suspiciously similar to late Latin *<taxonem>
'badger', the source of Spanish <tejón> and of other Romance names for
the animal. This etymology is not ironclad, since loss of the initial
plosive would be irregular, but it has been taken seriously since the
days of Schuchardt. And etymologies involving real words are a hell
of a lot better than etymologies based on impossible phantasms.
14 MANA 'to stay (in a place)'
Basque <min> 'to place, set up, settle'
No such Basque verb exists, in any sense at all. B&R have invented it
out of thin air.
No such verb could possibly exist. All ancient Basque verbs are
obliged to conform to certain structural patterns, and <min> does not
conform. There can no more be a Basque verb of the form <min> than
there can be a Spanish or a French verb of this form. But B&R are too
dumb to know this, or else they hope their readers are too ignorant to
know it.
There is no limit on Ruhlen's dishonesty. When the real words don't
oblige, he is happy to invent phony ones and to present them with a
straight face as "evidence".
Larry Trask
lar...@sussex.ac.uk
Jarel
>If memory serves, Nahali has native
>'1', but '2-5' are borrowed from Dravidian, and '6-10' from Indic.
Actually, 2-4 are Dravidian and the rest are Indic. Also, those are obvious
loans though and do no cause any problem. What causes more headaches are those
which are more difficult to tell. For example, in Romani, are the numerals for
6 and 10 from Dardic? Japanese has many numeral cognates with Tungustic and
Mongolian. I think that is a strong case for relationship than anything else.
Peter T. Daniels
>
> > > But I am surprised: you sound as if had discovered
> > > Ruhlen only this morning!
>
> Hardly. I've been having the Ruhlen and Bengtson experience for many
> years. But I have the clear impression that the extent of Ruhlen's
> dishonesty is not well known, even among historical linguists.
> Incompetence, certainly, but not dishonesty. I've been scratching my
> head for years for a way of publicizing his fraudulent ways, but
> without success. I tried to squeeze an account into the popular book
> on language change I'm doing for Cambridge, but the editor wouldn't
> let me do it -- "too technical", even though the anonymous readers
> loved it.
>
> So I'm just letting off a little steam here on sci.lang.
Konrad Koerner isn't exactly kind to fools -- so maybe *Historiographia
Linguistica* or *Diachronica* would be a venue for an article on R's
place in linguistics.
--
Peter T. Daniels gram...@att.net
I
which has to be proven before it is claimed in print. I suspect that R&B
did not straight out invent the data you mention; with so many chance
resemblances available, why would they have to do that? At least in the
one case you mention, "hazkoin"/"harzkoin" looks fairly similar to the
protoform *azkone which you mention. Couldn't "hazkoin" be their own
attempt at reconstruction, however bad and undocumented?
Nevertheless, if you really think they made up some data, it would be
perfectly reasonable for you to contact the publishers (Stanford?), and
get them to ask the authors for their sources. From a publisher's point
of view it's one thing to publish bad science (which can always be
defended as "controversial" or as "sparking debate"); it's quite another
thing to publish fraudulent science.
In article <48c7f19.03090...@posting.google.com>,
lar...@sussex.ac.uk (Larry Trask) wrote:
[...]
>As it happens, the Basque word for 'badger' occurs in 25 regional
>variant forms -- but *<hazkoin> is not one of them. B&R have invented
>this form. Are they so stupid that they can't copy a word off the
>page correctly? Or have they simply decided that the real forms are
>not good enough for their purposes, and replaced them with a phony
>form they like better?
>
>I guess the second. In another article, Ruhlen cites the word as
><harzkoin> -- which also doesn't exist, but which makes his etymologhy
>look more plausible.
>
>Anyway, it has been established that all the variants go back to an
>earlier
>*<azkone>, which is therefore the only form available for comparison.
>14 MANA 'to stay (in a place)'
>
>Basque <min> 'to place, set up, settle'
>
>No such Basque verb exists, in any sense at all. B&R have invented it
>out of thin air.
[...]
David Thomas
lar...@sussex.ac.uk (Larry Trask) writes:
{I've snipped Monsieur Trask's criticism for brevity only.}
>vael...@aol.comUspamo (David Thomas) wrote in message
>news:<20030908204345...@mb-m20.aol.com>...
>
>[on Ruhlen and Greenberg]
>
>> From what I understand, their methodology is a joke... at least, to serious
>> linguists.
>
>Absolutely right. But the broad nature of "multilateral comparison"
>is hardly the worst of it. In assembling his comparanda, Ruhlen is
>grossly incompetent and grossly dishonest. He chooses his conclusions
>in advance, and then he goes looking for confirming data. He
>suppresses all unwelcome data, and picks out only the things he likes.
> Even the data that make it onto the page are not presented honestly:
>they are twisted, mangled and mutilated in order to force them into
>the required fit.
This sounds alot like what has been done in order to 'prove' such
pseudosciences as homeopathy, magnetic therapy, and Uri Geller!
Scientific dishonesty apparently takes the exact same shape--experimental
fudging of data--whenever it appears.
- Vae
David Thomas
{Again, I've snipped for brevity...}
>A little more Ruhlen, from the same article.
>
>9. KUAN 'dog'
I wonder if he's a Tolkien fan out to prove something?
See: 'khugan, khug' in Tolkien's etymologies...
[^h indicates superscript h]
(*k^hugan >) *chu3an > *chuan "hound" (Exilic Quenya huan)
Perhaps I should start checking his glosses for 'Proto-World' to make sure
they're not all derived from Elven...
Or maybe he *would* seriously suggest that the languages are related, if he
didn't know Quenya's fictional... maybe someone should ask him to find out?
- Vae
David Thomas
Perhaps you should publish a whole book on it?
Even a short one?
I'm certain you could find enough material for it--maybe one chapter for each
major assertion made in these phony ways?
- Vae
mb
...
> > However, I would tend to ask the same question about
> > his publishers: monumental stupidity, flagrant dishonesty,
> > or both?
>
> Neither. They are in the business of selling books. They decided these
> books would sell. They were right. They would have been stupid (or
> dishonest to their shareholders)
> if they *hadn't* published them.
>
> So now we come down to the people who buy them. Monumental stupidity, ...
> or what?
Perhaps: Some people can't stand the aggravation to be told (even by
kooks) that they should shut up if they didn't read it firsthand, so
if the stuff isn't in a nearby library, one buys. Some of my money
even went to Faucounau, of all people, despite all the early signs.
Monumental stupidity, but on the other hand how can one skip checking?
Arindam Banerjee
> <zakur>, now specialized to 'big dog'. This has apparently displaced
> another word, <or>, recorded in the 16th century at both ends of the
> country.
The Bengali word for "dog" is "kukur".
John Atkinson
"Des Small" <des....@bristol.ac.uk> wrote ...
> "John Atkinson" <john...@bigpond.com> writes:
>
> > "Jacques Guy" <jg...@alphalink.com.au> wrote...
> > >
> > > However, I would tend to ask the same question about
> > > his publishers: monumental stupidity, flagrant dishonesty,
> > > or both?
> >
> > Neither. They are in the business of selling books. They decided
> > these books would sell. They were right. They would have been
> > stupid (or dishonest to their shareholders) if they *hadn't*
> > published them.
>
> Not necessarily. Publishers and imprints are brands, and the damage
> to the brand's prestige resulting from publishing rubbish may outweigh
> the immediate gain.
Well, it "may". But this consideration didn't stop Elsevier Science from
publishing "Flying Saucers Have Landed" back in 1973. I guess they somehow
managed to live with the loss in prestige.
John.
Larry Trask
> Dishonesty - to the extent of fabrication - is a very serious charge,
> which has to be proven before it is claimed in print.
I have already documented Ruhlen's dishonesty, to the extent of
fabricating non-existent words and forms, and I will be documenting
this behavior further in my next several postings. The charge is
already proven.
> I suspect that R&B
> did not straight out invent the data you mention; with so many chance
> resemblances available, why would they have to do that?
Don't ask me to explain B&R's motives. But I have good reason to
surmise that
B&R find the genuine chance resemblances to be disappointingly sparse,
and so they have decided to manufacture some additional ones.
> At least in the
> one case you mention, "hazkoin"/"harzkoin" looks fairly similar to the
> protoform *azkone which you mention. Couldn't "hazkoin" be their own
> attempt at reconstruction, however bad and undocumented?
No. They do not describe the form as a reconstruction, and they do
not attach an asterisk to it. They present it as a genuine Basque
form. In fact, they present it as *the* Basque form, since they fail
to mention the existence of any variant forms.
You are being far too kind to these crooks.
> Nevertheless, if you really think they made up some data, it would be
> perfectly reasonable for you to contact the publishers (Stanford?), and
> get them to ask the authors for their sources. From a publisher's point
> of view it's one thing to publish bad science (which can always be
> defended as "controversial" or as "sparking debate"); it's quite another
> thing to publish fraudulent science.
Sadly, I can see little point in doing this. The book is already out,
and has been since 1994. And this book is hardly the only one of
Ruhlen's publications containing fraudulent data. I'll talk about
another book later.
Larry Trask
lar...@sussex.ac.uk
David Thomas
adda...@bigpond.com (Arindam Banerjee) writes:
Yet another wonderous relation!
Heh...
I wonder what other sorts of languages we can string together?
Perhaps we should get someone's conlang on here and start 'proving' that all
modern language actually descends from a remote dialect of it!
- Vae
Athel Cornish-Bowden
>
> > Larry Trask wrote:
> >
> > > You can decide for yourself whether we're looking at monumental
> > > stupidity or at flagrant dishonesty. It must be one or the other --
> > > or maybe both.
> >
> > But I am surprised: you sound as if had discovered
> > Ruhlen only this morning!
> >
> So now we come down to the people who buy them. Monumental stupidity, ...
> or what?
>
history) to know? I bought two of Ruhlen's books after reading a very
favourable
account of his ideas in a book by Luigi Luca Cavalli Sforza, and,
having read
them, I didn't find any _obvious_ evidence of dishonesty. Cavalli
Sforza is
certainly not monumentally stupid, and is a leader in his field,
widely
respected if not universally agreed with. It may be that he is
inclined to look
favourably on work in other fields that agrees with his own ideas, but
then,
who is not?
At the risk of exposing my monumental stupidity, I should say that I
found
Ruhlen's account of words for vulva convincing, but I was even more
struck
by his account of words similar to "akwa" for water, but then, I have
no expert
knowledge that would enable me to recognize if his examples were
fraudulent, and
if those who do have such expert knowledge refrain from publicizing
it, how is
the ordinary reader to know? The best I could do was to check some of
his
examples relating to South American languagesin a grammar of Mapudungu
that I have: the result didn't really support his thesis, but then,
Mapudungu
is merely one of many South American languages, and I don't know how
typical it
is.
The point of all this is that experts (in any field) have an
_obligation_ to draw
attention to errors (and still more to fraudulent data) in books in
their subjects
that achieve popular success, how ever distasteful this may be, and
how ever
much easier it may be just to ignore the books they think are packed
with
nonsense.
Incidentally, I noticed the absurd probability calculation noted
earlier in
this thread when I read Ruhlen's account of it, but I didn't give it
special
importance, because in my experience very few people have much idea
how
to do a probability calculation, and when I see an incompetent one in
a paper
that I have to assess the first reaction is not "this guy is an idiot
(or dishonest),"
but "this guy is as incompetent as most scientists at estimating
probabilities,
and is only doing it because the journal editor expects it." It is
hard to escape
the feeling that even if one did the calculation correctly the results
for
"akwa" = water would be significant (if the data are accurately
reported). Has
anyone done this sort of calculation in a statistically valid way?
athel
--
Athel Cornish-Bowden
at...@ibsm.cnrs-mrs.fr
http://bip.cnrs-mrs.fr/bip10/homepage.htm
Larry Trask
16. MENA 'to think (about)'
Basque <mun> 'medulla', <munak> (pl.) 'brains'
First the linguistics; then the reality check.
The word being cited has the variant forms <muin> (by far the most
widespread form), <fuin>, <mun>, <hun>, <gun>, <un>. All these are
descended, by familiar developments, from an earlier *<bune>, which is
already phonologically useless for B&R. We can't even reconstruct an
*/m/ for the Pre-Basque of some 2000 years ago.
The word means 'pith, marrow' everywhere.
In a few localities, the word enters into a compound with a word
meaning 'head' to express 'brain(s)': <garun>, <burumuin>. Now, just
as English 'eyelids' and 'eyelashes' are sometimes clipped to 'lids'
and 'lashes', these compounds are occasionally clipped to their second
element in the sense of 'brain(s)'. But Azkue's 1905 dictionary, which
is apparently the unidentified source being used here, expressly
states that such clipping is unusual, because of the obvious
ambiguity, and that 'brain(s)' is commonly expressed by the full
compounds.
All this information is silently suppressed by B&R. The two forms
cited by B&R are severely localized, and they have been carefully
picked out because they suit B&R's purposes. This is not just
dishonest: it's criminal.
Now the reality check. B&R apparently believe that the speakers of
the first human language had a good understanding of the connection
between brains and thinking. Nonsense. This connection is a recent
discovery. Even in classical Europe, people had no idea about this.
They typically assumed that the seat of the intellect was the heart,
which is why we learn things "by heart". The brain was assumed to do
something else, such as cooling the blood.
The first person in Europe -- and probably in the world -- to propose
that the brain was the seat of the intellect was the Greek physician
Galen, in the 4th century AD. But Galen was laughed at, and centuries
passed before this understanding became widespread.
So, the linguistic crimes aside, B&R's proposal is imbecilic.
19. PAR 'to fly'
Basque <pimpirina> 'butterfly' (< *<pir-pir->)
The correct spelling is <pinpirina>.
Now, the first result ever obtained in the study of the prehistory of
Basque was this: nothing beginning with /p/ is ancient in Basque.
This result was obtained by Hugo Schuchardt, the father of Basque
historical linguistics, in his very first publication on the language,
in 1887. And it has stood up. The consonant */p/ was at best very
rare in Pre-Basque, and it may not have existed at all. But it
certainly never occurred word-initially.
Moreovere, nothing containing the cluster /np/ can possibly be ancient
either, since plosives were uniformly voiced after /n/ in the Middle
Ages. For example, Latin <tempora> 'times', borrowed as *<denpora>,
appears today as <denbora> 'time'. B&R invoke an *ad hoc*
dissimilation, but there is absolutely no parallel in Basque for such
a process. They are merely twisting the data to suit themselves.
Now, words for 'butterfly' are remarkably unstable in Basque. Many
words are recorded, and the words recorded in our earliest texts are
not the words in use today. The most widespread word today is
<tximeleta>, which is not recorded before 1912. Even Azkue's 1905
dictionary doesn't list this, in spite of the fact that it's the usual
word today in Azkue's home town.
The word <pinpirina> is entirely confined to the small Lapurdian
dialect on the French side of the Pyrenees -- a fact concealed by B&R.
Besides 'butterfly', the words also means 'bud (of a flower)',
'garfish' and 'little girl dressed up in her finery'. Well, the last
sense might be a metaphor, but the others are not, and the signal
failure of buds and garfish to take to the air is rather awkward for
B&R. So, of course, they simply suppress this information, even
though they can hardly have failed to see it in the dictionary entry.
The word has variants with /l/ in place of /r/, such as
<pinpilinpauxa>, but these do not suit B&R's purposes, so they too are
suppressed.
Now, the crunch. The Lapurdian dialect, uniquely among the dialects
of Basque, is fond of expressive formations in <panp->, <pinp-> and
<punp-> (but never
<penp-> or <ponp->). Any dictionary which includes <pinpirina> will
list dozens of others, such as <pinpa> 'bounce of a ball', <pinperrez>
'inside out',
<pinpili-panpala> 'favorite', <pinpillun-punpullun> 'head over heels',
<pinpin> 'wagtail', <pinpin> 'a certain children's finger game',
<pinpina> 'elegant',
<pinpingaratxa> 'bluetit', <pinpinka> 'hopping', <pinpiro> 'garfish',
<pinpi-panpa> 'sound of a blow', <pinpirinakeri> 'putting on airs',
<pinpoil> 'somersault'...well, you get the idea, and this is only a
modest sample.
B&R can hardly have failed to notice that their chosen word was
sitting in the dictionary in the middle of a long list of these
transparent expressive formations, but of course they have suppressed
this information too.
Even more than their occasional fabrication of non-existent items, it
is this constant and systematic suppression of inconvenient data which
demonstrates B&R's fundamental dishonesty.
Larry Trask
lar...@sussex.ac.uk
Torsten Poulin
> The Bengali word for "dog" is "kukur".
And <kukur> is also Danish for 'cuckoo clock'.
--
Torsten
Jarel
> I bought two of Ruhlen's books after reading a very favourable account of his
ideas in a book by Luigi Luca Cavalli Sforza
Cavalli Sforza is not a linguist, so he is not a qualified authority to judge
Ruhlen's work.
>, but I was even more struck by his account of words similar to "akwa" for
water,
The various words for "water" to not make any convincing argument for all
languages being related regardless if the data is faulty or not. Certain kinds
of words are very similar in many languages across the world. These including
but not limited to: onomatopoeia, nursery words, etc.
Merlijn De Smit
Oh, but Tolkien was a linguist. A lot of the Quenya and Sindarin word
material is inspired by Nordic languages, Finnish, but many other
languages as well.
PIE *k´uon --> Greek kúon, German hund, etc. So not only did our
distant forefathers in the Rift Valley domesticate dogs some 150,000
years ago at the latest, the word stayed virtually unchanged for more
than 144,000 years right up until PIE and later, if you count ancient
Greek. Now Ruhlen has often cited the famous Romanian "nepot"
'nephew', I think it was, but this is just a bit too far-fetched.
One of the many weak points of multilateral comparison, but a very
important one, is that of course there are no objective criteria for
superficial similarity - and the method does not go much further than
lumping words together on the aubjective basis of a perceived
similarity. That´s why some people came up with the Comparative Method
in the 19th century. Greenberg/Ruhlen-type multilateral comparison
pushes the clock back to those times in which research showed that all
languages descended from Dutch.
Merlijn de Smit
Sebastian Hew
> In article <890e65ea.03090...@posting.google.com>,
> adda...@bigpond.com (Arindam Banerjee) writes:
>
> > > The Basque for 'dog' is <(t)xakur>, in origin a diminutive of
> > > earlier <zakur>, now specialized to 'big dog'. This has
> > > apparently displaced another word, <or>, recorded in the 16th
> > > century at both ends of the country.
> >
> > The Bengali word for "dog" is "kukur".
>
> Yet another wonderous relation!
>
> Heh...
>
> I wonder what other sorts of languages we can string together?
Perhaps we can add Sinitic languages to the list, with 'quan' and 'huen'
for 'dog' from Mandarin and Cantonese respectively. ;-)
Sebastian.
mb
...
> One of the many weak points of multilateral comparison, but a very
> important one, is that of course there are no objective criteria for
> superficial similarity - and the method does not go much further than
> lumping words together on the aubjective basis of a perceived
> similarity. That´s why some people came up with the Comparative Method
> in the 19th century. Greenberg/Ruhlen-type multilateral comparison
> pushes the clock back to those times in which research showed that all
> languages descended from Dutch.
Good point. However, there were a number of honest investigators even
then. When the Abbott Hervás y Panduro did his multilateral comparison
in 1787 with the more than 110 languages he could document for a very
long list of words, he used all of them, did not try tricks with forms
or meanings, and even though he was obviously relying on
impressionistically perceived similarities he grouped these according
to the number of similarities. This allowed him to correctly indicate
the main groupings (without reconstitution, as required by Ruhlen!)
and what, statistically speaking, were likely to be chance
similarities. He was testing the Babel tower hypothesis, and came up
with an honest evaluation: If mankind spoke a single language before
Babel, he says, nothing is left to recognize it, and even the most
familiar words like 'mother' are not reliably there; so the Lord must
have given the original nations completely different tongues from the
get-go.
His method was way more scientific and honest than that of G&R. The
G&R method and way of argumentation compare, sometimes unfavorably,
with the 16th century treatises hell bent on deriving everything from
Hebrew (or, as you say, Dutch).
Arindam Banerjee
> Arindam Banerjee wrote:
>
> > The Bengali word for "dog" is "kukur".
>
> And <kukur> is also Danish for 'cuckoo clock'.
Maybe, but a "cuckoo clock" is not a dog, nor were cuckoo clocks
around 10,000 years ago.
Torsten Poulin
> Torsten Poulin wrote:
>> Arindam Banerjee wrote:
>>
>> > The Bengali word for "dog" is "kukur".
>>
>> And <kukur> is also Danish for 'cuckoo clock'.
>
> Maybe, but a "cuckoo clock" is not a dog, nor were cuckoo clocks
> around 10,000 years ago.
Of course not, but the comparison was about as relevant as your
implied comparison of Bengali <kukur> and Basque <(t)xakur> or
<zakur>.
--
Torsten
David Thomas
adda...@bigpond.com (Arindam Banerjee) writes:
That was likely his point...
- Vae
David Thomas
isol...@hotmail.com (Merlijn De Smit) writes:
::snip me::
>Oh, but Tolkien was a linguist. A lot of the Quenya and Sindarin word
>material is inspired by Nordic languages, Finnish, but many other
>languages as well.
Most certainly agreed, and competent, knowledgeable people would know; that is
why I would expect someone in the same vein as Ruhlen to be able to make such a
transparently misinformed connection.
>PIE *k´uon --> Greek kúon, German hund, etc. So not only did our
I had noted this interesting similarity when I began learning Greek... as well
I've noticed many similarities between elven words in Tolkien's corpus and
others in my linguistic studies... though there's not a good way to know if any
of the similarities are due to a true relation--that is, due to Tolkien
hand-picking a few roots or somesuch here or there...
::snip::
>One of the many weak points of multilateral comparison, but a very
>important one, is that of course there are no objective criteria for
>superficial similarity - and the method does not go much further than
>lumping words together on the aubjective basis of a perceived
>similarity. That´s why some people came up with the Comparative Method
>in the 19th century. Greenberg/Ruhlen-type multilateral comparison
>pushes the clock back to those times in which research showed that all
>languages descended from Dutch.
Has Ruhlen ever used 'multilateral comparison'--hardly 'multilateral' in any
sense, methinks!--on any paradigms? I'm betting not.
- Vae
Brian M. Scott
Cornish-Bowden) wrote:
[...]
>The point of all this is that experts (in any field) have an
>_obligation_ to draw
>attention to errors (and still more to fraudulent data) in books in
>their subjects
>that achieve popular success, how ever distasteful this may be, and
>how ever
>much easier it may be just to ignore the books they think are packed
>with nonsense.
It's nice when they do, but it's not something that can
realistically be expected: debunking for laymen what any expert
can see is patent nonsense takes time away from research. This
is an especially grave concern for a young academic trying to get
tenure.
>Incidentally, I noticed the absurd probability calculation noted
>earlier in
>this thread when I read Ruhlen's account of it, but I didn't give it
>special
>importance, because in my experience very few people have much idea
>how
>to do a probability calculation, and when I see an incompetent one in
>a paper
>that I have to assess the first reaction is not "this guy is an idiot
>(or dishonest),"
>but "this guy is as incompetent as most scientists at estimating
>probabilities,
>and is only doing it because the journal editor expects it."
But that certainly isn't the case here, since it's one of
Ruhlen's stock responses.
>It is hard to escape
>the feeling that even if one did the calculation correctly the results
>for
>"akwa" = water would be significant (if the data are accurately
>reported). Has
>anyone done this sort of calculation in a statistically valid way?
Some people have tried. Don Ringe, for instance, has 'On
calculating the factor of chance in language comparison',
Transactions of the American Philosophical Society 82:1-110,
1992, which I haven't seen, and 'How hard is it to match
CVC-roots?', Trans. Phil. Soc. 97:213-244, 1997, which I have
seen (and which, unfortunately, is *not* free of error, though
the conclusions are qualitatively correct). Ann Kumar & Phil
Rose, 'Lexical Evidence for Early Contact between Indonesian
Languages and Japanese', Oceanic Linguistics v. 39, nr. 2, Dec.
2000, uses similar techniques to make a statistical case for
borrowing. And of course there's our own Mark Rosenfelder's 'How
likely are chance resemblances between languages?' at
<http://www.zompist.com/chance.htm>, which ought to be required
reading for anyone interested in the subject.
Brian
Phil Healey
>
> Obviously related to "cur" and "corgi". Come to think
> of it, French "coeur" too. Yes, I know it means "heart"
> but the semantic shift heart<->dog is well attested:
> see Arabic kalb "dog", qalb "heart".
>
> Will Stanford University Press find its way to my log
> cabin soon?
I'm afraid you've been beaten to publication by Mikhail Bulgakov in his
treatise "Sobach'e Serdtse" - "Heart of a Dog."
--
Phil Healey
'It's happened' I panted 'There's a Lion in my room'
'I'm afraid any discussion would have no value' he hung up.
Athel Cornish-Bowden
> Cornish-Bowden) wrote:
>
> [...]
>
> >The point of all this is that experts (in any field) have an
> >_obligation_ to draw
> >attention to errors (and still more to fraudulent data) in books in
> >their subjects
> >that achieve popular success, how ever distasteful this may be, and
> >how ever
> >much easier it may be just to ignore the books they think are packed
> >with nonsense.
>
> It's nice when they do, but it's not something that can
> realistically be expected: debunking for laymen what any expert
> can see is patent nonsense takes time away from research. This
> is an especially grave concern for a young academic trying to get
> tenure.
in other fields. As Miguel de Unamuno said many years ago (and
at the risk of his life) there are times when to remain silent is to lie.
It doesn't of course have to be done by young academics trying to
get tenure -- why not by their well established seniors?
>
[ ... ]
> >It is hard to escape
> >the feeling that even if one did the calculation correctly the results
> >for
> >"akwa" = water would be significant (if the data are accurately
> >reported). Has
> >anyone done this sort of calculation in a statistically valid way?
>
> Some people have tried. Don Ringe, for instance, has 'On
> calculating the factor of chance in language comparison',
> Transactions of the American Philosophical Society 82:1-110,
> 1992, which I haven't seen, and 'How hard is it to match
> CVC-roots?', Trans. Phil. Soc. 97:213-244, 1997, which I have
> seen (and which, unfortunately, is *not* free of error, though
> the conclusions are qualitatively correct). Ann Kumar & Phil
> Rose, 'Lexical Evidence for Early Contact between Indonesian
> Languages and Japanese', Oceanic Linguistics v. 39, nr. 2, Dec.
> 2000, uses similar techniques to make a statistical case for
> borrowing. And of course there's our own Mark Rosenfelder's 'How
> likely are chance resemblances between languages?' at
> <http://www.zompist.com/chance.htm>, which ought to be required
> reading for anyone interested in the subject.
Athel Cornish-Bowden
>
> > I bought two of Ruhlen's books after reading a very favourable account of his
> ideas in a book by Luigi Luca Cavalli Sforza
>
> Cavalli Sforza is not a linguist,
but if it was not I'm happy to make it explicit: Cavalli Sforza
is not a linguist, but he is the world's leading expert on genetic
diversity, and as such ...
>so he is not a qualified authority to judge
> Ruhlen's work.
on genetic diversity.
Athel Cornish-Bowden
>Athel Cornish-Bowden wrote:
>>
>> [ ... ]
>> >
>> > > But I am surprised: you sound as if had discovered
>> > > Ruhlen only this morning!
>> > >
>> [ ... ]
>> >
>> > So now we come down to the people who buy them. Monumental stupidity, ...
>> > or what?
>> >
>> Well maybe, but how is the ordinary reader (not an expert in language
>> history) to know? I bought two of Ruhlen's books after reading a very
>> favourable
>> account of his ideas in a book by Luigi Luca Cavalli Sforza, and,
>> having read
>> them, I didn't find any _obvious_ evidence of dishonesty. Cavalli
>> Sforza is
>> certainly not monumentally stupid, and is a leader in his field,
>> widely
>> respected if not universally agreed with. It may be that he is
>> inclined to look
>> favourably on work in other fields that agrees with his own ideas, but
>> then,
>> who is not?
>
>such a monumental fraud as the double-winged chart of the genetic tree
>"aligned" with Ruhlen's linguistic "tree"? Quite aside from the fact
>that Ruhlen's tree is absurd in the first place (which you wouldn't be
>expected to know),
>
I think you'll find that great many people give respect to Cavalli-Sforza,
and not just biologists: linguists also. For example, at
http://www.cogs.susx.ac.uk/users/larryt/basque.faqs.html you can
read the following words:
"Recently, however, the geneticist Luiga Luca Cavalli-Sforza has completed
a gene map of the peoples of Europe, and he finds the Basques to be strikingly
different from their neighbors. The genetic boundary between Basques and
non-Basques is very sharp on the Spanish side. On the French side, the
boundary is more diffuse: it shades off gradually toward the Garonne in
the north. These findings are entirely in agreement with what we know of
the history of the language."
These words don't read to me as if written by someone who gives no
respect at all to Cavalli-Sforza.
Torsten Poulin
> Jarel wrote:
>> Cavalli Sforza is not a linguist,
> [...]
>> so he is not a qualified authority to judge Ruhlen's work.
> ... is well qualified to refer to work in other fields that
> impinges on genetic diversity.
What makes you think typology etc. impinges on genetic diversity?
--
Torsten
Torsten Poulin
>> Maybe, but a "cuckoo clock" is not a dog, nor were cuckoo
>> clocks around 10,000 years ago.
> That was likely his point...
Indeed.
By the way, the Danish word is not even monomorphematic. It is
a compound of the onomatopoeia <kuk> '(the call of the) cuckoo'
and <ur> 'clock', 'watch' (through Low German and Old French from
Latin <hora>)
--
Torsten
Peter T. Daniels
>
> b.s...@csuohio.edu (Brian M. Scott) wrote in message news:<3f5fb96f...@enews.newsguy.com>...
> > On 10 Sep 2003 05:40:54 -0700, at...@ibsm.cnrs-mrs.fr (Athel
> > Cornish-Bowden) wrote:
> >
> > [...]
> >
> > >The point of all this is that experts (in any field) have an
> > >_obligation_ to draw
> > >attention to errors (and still more to fraudulent data) in books in
> > >their subjects
> > >that achieve popular success, how ever distasteful this may be, and
> > >how ever
> > >much easier it may be just to ignore the books they think are packed
> > >with nonsense.
> >
> > It's nice when they do, but it's not something that can
> > realistically be expected: debunking for laymen what any expert
> > can see is patent nonsense takes time away from research. This
> > is an especially grave concern for a young academic trying to get
> > tenure.
>
> It's not unrealistic, and it's exactly what researchers do do
Name some. S. J. Gould said he didn't have the time to devote to
debating creationists.
> in other fields. As Miguel de Unamuno said many years ago (and
> at the risk of his life) there are times when to remain silent is to lie.
> It doesn't of course have to be done by young academics trying to
> get tenure -- why not by their well established seniors?
> >
> [ ... ]
>
> > >It is hard to escape
> > >the feeling that even if one did the calculation correctly the results
> > >for
> > >"akwa" = water would be significant (if the data are accurately
> > >reported). Has
> > >anyone done this sort of calculation in a statistically valid way?
> >
> > Some people have tried. Don Ringe, for instance, has 'On
> > calculating the factor of chance in language comparison',
> > Transactions of the American Philosophical Society 82:1-110,
> > 1992, which I haven't seen,
Absolutely fundamental, and thoroughly refuted by Manaster Ramer in an
article entitled "Glass Houses: Greenberg, Ringe, and Innumeracy in
Comparative Linguistics," an article that I have only on disk but that I
have seen references to, so it's definitely been published.
> > and 'How hard is it to match
> > CVC-roots?', Trans. Phil. Soc. 97:213-244, 1997, which I have
> > seen (and which, unfortunately, is *not* free of error, though
> > the conclusions are qualitatively correct). Ann Kumar & Phil
> > Rose, 'Lexical Evidence for Early Contact between Indonesian
> > Languages and Japanese', Oceanic Linguistics v. 39, nr. 2, Dec.
> > 2000, uses similar techniques to make a statistical case for
> > borrowing. And of course there's our own Mark Rosenfelder's 'How
> > likely are chance resemblances between languages?' at
> > <http://www.zompist.com/chance.htm>, which ought to be required
> > reading for anyone interested in the subject.
>
> Thanks for these references.
--
Peter T. Daniels gram...@att.net
Peter T. Daniels
Larry will certainly speak for himself, but what the bloody hell does
the fact that Basque people are genetically distinct from the people
around them have to do with the fact that he was taken in by and
accepted the total fraud foisted on him by Ruhlen? At the very least, it
calls into question his critical abilities, and his publication of the
fraudulent chart suggests he has no capacity whatsoever for interpreting
two-dimensional diagrams.
[I don't know why my posting does not appear when I click the
"Reference" to it in the posting to which I reply, but I append the end
of my posting, which you failed to reproduce.]
>
> The two wings of the tree simply are not anywhere close to congruent.
>
> All they did was rotate all the twigs around the nodes of the branches
> so that the list down the middle came out in the same order (mostly) for
> the two sides.
>
> This was first published in 1988, in the Proc. Nat. Acad. Sci. (whose
> integrity was recently questioned anyway in this newsgroup), and has
> been reprinted endlessly ever since.
>
> Jacques Guy, incidentally, went to the trouble of posting details of the
> fraud on LINGUIST List once upon a time, but it's easier to check it in
> any version simply by imagining a mirror down the middle -- and noting
> that it doesn't produce the "reflection" shown.
>
> (My note on this is in my review of *Atlas of Languages* in *Language in
> Society* 27 (1998).)
Larry Trask
[on Ruhlen's failings]
> Well maybe, but how is the ordinary reader (not an expert in language
> history) to know?
The ordinary reader has no way of knowing, unless informed
commentators make the truth public. But most of the blame for
Ruhlen's prominence must be laid at the door of the journalists who
promote his ideas merely because those ideas are spectacular. If
those journalists would only do their job properly, and talk to the
professionals before they splashed Ruhlen's idiocies across their
pages, Ruhlen would now be little more than one more anonymous
linguistic crank among hundreds.
> I bought two of Ruhlen's books after reading a very
> favourable
> account of his ideas in a book by Luigi Luca Cavalli Sforza, and,
> having read
> them, I didn't find any _obvious_ evidence of dishonesty. Cavalli
> Sforza is
> certainly not monumentally stupid, and is a leader in his field,
> widely
> respected if not universally agreed with. It may be that he is
> inclined to look
> favourably on work in other fields that agrees with his own ideas, but
> then,
> who is not?
Is C-S really a "leader in his field"? He is certainly prominent.
I'm not competent to judge genetic work, but I have heard persistent
rumors that other geneticists do not think much of C-S's work.
And he's "not monumentally stupid"? Wanna bet?
C-S has read a bit of Greenberg and Ruhlen, or at least (more likely)
he's read about them in the newspapers. As a result, he fancies
himself an authority on language, and he pontificates on the subject
at length. But, in truth, he doesn't know his linguistic ass from his
elbow, and his linguistic writings consist of an unrelieved sequence
of hilariously imbecilic blunders. Look at ch. 7 of his book The
Great Human Diasporas.
Here's a sample of his insights, p. 193.
Before the Great Vowel Shift, English had only about seven vowels, "in
line with other Latin-derived languages". Today it has about 20
vowels, plus a number of diphthongs.
Boy, that was some shift.
All his writings on language are at this level of imbecility. But he
warmly endorses Greenberg and Ruhlen, and he has the gall to chide
professional linguists for our stick-in-the-mud attitude and our
failure to embrace G and R.
> At the risk of exposing my monumental stupidity, I should say that I
> found
> Ruhlen's account of words for vulva convincing,
Ah, you shouldn't. In my next posting, I'll present B&R's treatment
of the Basque word for 'vulva'. That should open your eyes.
The point is this: YOU CANNOT TRUST THE DATA PRESENTED BY RUHLEN.
He's a shyster, and his data are cooked. They are not honest data.
> but I was even more
> struck
> by his account of words similar to "akwa" for water, but then, I have
> no expert
> knowledge that would enable me to recognize if his examples were
> fraudulent, and
> if those who do have such expert knowledge refrain from publicizing
> it, how is
> the ordinary reader to know?
We don't refrain from publishing. I would dearly love to publish an
exposé of Ruhlen's crimes, but where? The New York Times is simply
not going to be interested. Journalists do not run stories along the
lines of "Nothing very interesting going on after all".
> The point of all this is that experts (in any field) have an
> _obligation_ to draw
> attention to errors (and still more to fraudulent data) in books in
> their subjects
> that achieve popular success, how ever distasteful this may be, and
> how ever
> much easier it may be just to ignore the books they think are packed
> with
> nonsense.
And that's exactly what I'm doing here.
> Incidentally, I noticed the absurd probability calculation noted
> earlier in
> this thread when I read Ruhlen's account of it, but I didn't give it
> special
> importance, because in my experience very few people have much idea
> how
> to do a probability calculation, and when I see an incompetent one in
> a paper
> that I have to assess the first reaction is not "this guy is an idiot
> (or dishonest),"
> but "this guy is as incompetent as most scientists at estimating
> probabilities,
> and is only doing it because the journal editor expects it." It is
> hard to escape
> the feeling that even if one did the calculation correctly the results
> for
> "akwa" = water would be significant (if the data are accurately
> reported).
But Ruhlen's incompetent calculation is not the point. The point, as
I explained, is that his calculation is of no relevance to his
practice. But then he lies and assures us that it is.
Larry Trask
lar...@sussex.ac.uk
Toni Keskitalo
> Arindam Banerjee wrote:
>> The Bengali word for "dog" is "kukur".
>
>
> Obviously related to "cur" and "corgi". Come to think
> of it, French "coeur" too. Yes, I know it means "heart"
> but the semantic shift heart<->dog is well attested:
> see Arabic kalb "dog", qalb "heart".
>
Newsflash: The Finnish word for 'dog' is "koira"! It does have an
uncanny resemblance, doesn't it?
Toni
--
# Replace .invalid with .fi for personal mail only #
"Jos aivastan, koko seinä kaikuu." (Tommi Liimatta)
Larry Trask
21. PUTI 'vulva'
Basque <poto-rro> 'pubis, vulva'
The universal Basque word for 'vulva' is <alu>, recorded from 1562.
No use to
B&R. So they've dragged up <potorro>, which means 'vulva' but not
'pubis'. The segmentation, as always, is arbitrary and indefensible.
This word exists nowhere but in a small area in the west of the
province of Gipuzkoa. According to the Basque Academy's huge new
dictionary, this is "a rather indecent word, used by men in
companionable contexts" (my translation).
In other words, this is an obscene slang term, used by men engaging in
guy talk over a few drinks. And you'll recall that nothing beginning
with /p/ is ever ancient in Basque. In fact, <potorro> is nowhere
recorded before 1977 -- a curious characteristic for a word descended
directly from Proto-World.
And this late attestation doesn't result from squeamishness. Azkue's
1905 dictionary lists a number of slang words for 'vulva', including
at least one he labels <malsonante> 'nasty, indecent'.
So, B&R have singled out a severely localized obscene slang term
coined only a few years ago, and they want us to believe that it is a
direct survival from Proto-World. This makes as much sense as
concluding that a Proto-World GAZ 'breast' survives in English
'gazongas'.
By the way, <pot->, with its palatalized variants <pott-> and <potx->,
is a common expressive stem in Basque. It's used to construct names
of sexual organs (male and female), names of small creatures, and
other senses typical of expressive forms in Basque. And this stem is
not even native: it's borrowed from Spanish. A big Spanish dictionary
will list a number of expressive formations in <pot-> and <poch->.
26. TSUMA 'hair'
Basque <zam-ar(r)> 'lock of wool, shock of hair'
Basque <zamar>, to give the word its correct spelling (B&R can't even
get that right), is recorded in a number of senses. But 'lock of
wool' and 'shock of hair' are not among them. B&R have manufactured
these glosses out of whole cloth. And, as usual, they've inserted an
arbitrary segmentation.
Anyway, most of the senses recorded for this word are not attested
before the late 19th century or the 20th century. The only sense
recorded in the 16th, 17th and 18th centuries, and in the first half
of the 19th century, is 'warm lined jacket', usually 'sheepskin
jacket'. This is therefore the only sense available for comparison.
The various other modern senses show developments in two directions:
'covering' and 'fleece'. The closest recorded sense to B&R's
invention is 'unruly mop of hair'. But this sense is not recorded
before 1961; it occurs only in the French Basque Country; and it is
undoubtedly calqued on French <toison>, which has this sense and which
shares other meanings with <zamar>.
The earliest records of the Basque word show a form <zamarra>. The
modern form therefore results from loss of the final <a>, a very
familiar and well-understood process in Basque.
Consequently, we may safely identify this word with Spanish <zamarra>
'sheepskin jacket, sheepskin'. Now, Basque loans into Spnaish are
vanishingly rare, while Spanish loans into Basque run into the
thousands.
Further, what appears to be the same word is recorded also in
Portuguese, Catalan, Occitan, Old French and even Italian. There is
no parallel for the diffusion of a Basque word into such a huge area
of Romance. On top of this, the earliest attestation of the word
occurs in the first half of the 13th century, in the north of
Languedoc -- hardly consistent with a Basque origin.
The eminent Romanist Corominas, in contemplating the word, considers a
Basque origin but rejects this as indefensible. He finally declares
the word to be "pre-Roman" -- Romanist-speak for "we haven't got a
clue".
So, B&R have metamorphosed a Romance word for 'sheepskin jacket' into
a fantastically ancient Basque word for 'hair'.
That concludes the Basque items presented in B&R's Proto-World
article. Back another time with some further gems from the world's
greatest linguistic clowns.
Larry Trask
lar...@sussex.ac.uk
Merlijn De Smit
>
> > I bought two of Ruhlen's books after reading a very favourable account of his
> ideas in a book by Luigi Luca Cavalli Sforza
>
> Cavalli Sforza is not a linguist, so he is not a qualified authority to judge
> Ruhlen's work.
>
> >, but I was even more struck by his account of words similar to "akwa" for
> water,
>
> The various words for "water" to not make any convincing argument for all
> languages being related regardless if the data is faulty or not. Certain kinds
> of words are very similar in many languages across the world. These including
> but not limited to: onomatopoeia, nursery words, etc.
But would words for "water" also fall in one of these categories? From
the top of my head, I cannot think of any word meaning water (or
river, lake, rain, etc.) in any language which might be iconic.
As a first-year student I once argued that Finnish 'vesi' might be
onomatopoetic, since if you leave the vowels out and try to pronounce
it it sounds like water running by, kind of. I had forgotten at the
time that the sibilant is secundary, though, which destroyed my nice
idea.
Merlijn de Smit
Brian M. Scott
Cornish-Bowden) wrote:
>b.s...@csuohio.edu (Brian M. Scott) wrote in message news:<3f5fb96f...@enews.newsguy.com>...
>> On 10 Sep 2003 05:40:54 -0700, at...@ibsm.cnrs-mrs.fr (Athel
>> Cornish-Bowden) wrote:
>> [...]
>> >The point of all this is that experts (in any field) have an
>> >_obligation_ to draw
>> >attention to errors (and still more to fraudulent data) in books in
>> >their subjects
>> >that achieve popular success, how ever distasteful this may be, and
>> >how ever
>> >much easier it may be just to ignore the books they think are packed
>> >with nonsense.
>> It's nice when they do, but it's not something that can
>> realistically be expected: debunking for laymen what any expert
>> can see is patent nonsense takes time away from research. This
>> is an especially grave concern for a young academic trying to get
>> tenure.
>It's not unrealistic, and it's exactly what researchers do do
>in other fields.
Oh? How many research biologists do you see spending their time
debunking creationism? How many researchers bother to debunk the
likes of Erich von Däniken, Graham Hancock, and Robert Bauval?
How many serious historians and archaeologists bother to write up
detailed refutations of the crap that Adrian Gilbert puts out?
In each case you can probably find a tiny handful, but they are
very much the exception.
[...]
Brian
Des Small
> On 11 Sep 2003 01:08:35 -0700, at...@ibsm.cnrs-mrs.fr (Athel
> Cornish-Bowden) wrote:
>
> >b.s...@csuohio.edu (Brian M. Scott) wrote in message news:<3f5fb96f...@enews.newsguy.com>...
>
> >> On 10 Sep 2003 05:40:54 -0700, at...@ibsm.cnrs-mrs.fr (Athel
> >> Cornish-Bowden) wrote:
>
> >> [...]
>
> >> >The point of all this is that experts (in any field) have an
> >> >_obligation_ to draw attention to errors (and still more to
> >> >fraudulent data) in books in their subjects that achieve popular
> >> >success, how ever distasteful this may be, and how ever much
> >> >easier it may be just to ignore the books they think are packed
> >> >with nonsense.
>
> >> It's nice when they do, but it's not something that can
> >> realistically be expected: debunking for laymen what any expert
> >> can see is patent nonsense takes time away from research. This
> >> is an especially grave concern for a young academic trying to get
> >> tenure.
>
> >It's not unrealistic, and it's exactly what researchers do do in
> >other fields.
>
> Oh? How many research biologists do you see spending their time
> debunking creationism?
Richard Dawkins's "The Blind Watchmaker" doesn't pull many punches,
and it is surely impossible to read much Gould and believe it's
reconcilible with creationism.
> How many researchers bother to debunk the
> likes of Erich von Däniken, Graham Hancock, and Robert Bauval?
> How many serious historians and archaeologists bother to write up
> detailed refutations of the crap that Adrian Gilbert puts out?
> In each case you can probably find a tiny handful, but they are
> very much the exception.
I'm not a particularly keen kook-watcher, but isn't von Däniken famous
for claiming that aliens helped build the pyramids and so on?
Ruhlen's proposals not only violate no laws of physics, they sound far
from superficially absurd. How many of your favourite archeological
kooks have free reign in the NYT, and have fooled Murray Gell-Mann and
allegedly prominent geneticists? If a creationist scam reached these
levels of apparent respectability I'd bet there'd be a lot more
biologists than Dawkins going ballistic.
I would say that the "cold fusion" affair is a closer match in terms
of perception by raw lay-persons, and those guys certainly got stomped
on pretty good. If Ruhlen is as bad as people here are saying (and
Larry Trask is putting together a convincing case) then it is surely
at least a shame that this is not better known. Even having the
details documented here is a start, of course.
Des
has correctly reconstructed proto-World but the margins are too small, hélas.
> [...]
>
> Brian
Jarel
>But would words for "water" also fall in one of these categories? From
the top of my head, I cannot think of any word meaning water (or river, lake,
rain, etc.) in any language which might be iconic.
I think the words for water are rather iconic. Most begin with vowels or
have labial consonants, semivowels and nasal. perhaps the sound of a stream
gently bubbling by or water being poured. Here are some words for water in
other languages:
Indonesian: air
Ainu: wakka
Jaqaru: uma
Quechua: yaku
Proto-IE: akwa/widur
Cherokee: ama
Dakota: mni
Japanese: mizu/oyo(cold water)
Klammam: klú?
Alabama: oki
Inuit: imiq
Tamil: am
See, there is one of those language universals going on, not that all human
languages have a cognate for water!
Brian M. Scott
wrote:
[...]
>Here are some words for water in
>other languages:
[...]
>Proto-IE: akwa/widur
That should be *wodr (oblique *wedn-, root *wed-), *wedo:r;
*akWah2- is, I believe, securely attested only in Germanic and
Italic and so quite possibly not PIE at all.
[...]
Brian
Brian M. Scott
>> >> [...]
First, one would expect a few more debunking efforts in this
particular field than in most, given the high profile of
creationism and crypto-creationist notions like intelligent
design (ID) these days. Secondly, 'not reconcilable with
creationism' is very different from 'directed debunking of
creationist nonsense'.
>> How many researchers bother to debunk the
>> likes of Erich von Däniken, Graham Hancock, and Robert Bauval?
>> How many serious historians and archaeologists bother to write up
>> detailed refutations of the crap that Adrian Gilbert puts out?
>> In each case you can probably find a tiny handful, but they are
>> very much the exception.
>I'm not a particularly keen kook-watcher, but isn't von Däniken famous
>for claiming that aliens helped build the pyramids and so on?
Von Däniken claimed an amazing variety of nonsense; some of it is
superficially credible. Some of Gilbert's stuff is in my opinion
entirely comparable to Ruhlen's. And the fact remains that this
sort of crap has a considerable following, however absurd you or
I may find it.
>Ruhlen's proposals not only violate no laws of physics, they sound far
>from superficially absurd. How many of your favourite archeological
>kooks have free reign in the NYT, and have fooled Murray Gell-Mann and
>allegedly prominent geneticists? If a creationist scam reached these
>levels of apparent respectability I'd bet there'd be a lot more
>biologists than Dawkins going ballistic.
I'm not sure that ID hasn't reached a comparable level of
apparent respectability in the lay world, at least in the U.S.,
but I still don't see many research biologists spending a lot of
time on public debunking.
>I would say that the "cold fusion" affair is a closer match in terms
>of perception by raw lay-persons, and those guys certainly got stomped
>on pretty good.
But I don't think that cold fusion is a close match in terms of
professional perception. As I understand it, the stomping was so
enthusiastic at least partly because it and its 'discoverers'
couldn't -- quite -- be dismissed out of hand.
[...]
Brian
Peter T. Daniels
> Anyway, most of the senses recorded for this word are not attested
> before the late 19th century or the 20th century. The only sense
> recorded in the 16th, 17th and 18th centuries, and in the first half
> of the 19th century, is 'warm lined jacket', usually 'sheepskin
> jacket'. This is therefore the only sense available for comparison.
> The earliest records of the Basque word show a form <zamarra>. The
> modern form therefore results from loss of the final <a>, a very
> familiar and well-understood process in Basque.
> Further, what appears to be the same word is recorded also in
> Portuguese, Catalan, Occitan, Old French and even Italian. There is
> no parallel for the diffusion of a Basque word into such a huge area
> of Romance. On top of this, the earliest attestation of the word
> occurs in the first half of the 13th century, in the north of
> Languedoc -- hardly consistent with a Basque origin.
Itinerant Basque haberdashers? After all, you used to could see those
wonderful Hudson's Bay Company coats (and blankets) far beyond Canada.
> That concludes the Basque items presented in B&R's Proto-World
> article. Back another time with some further gems from the world's
> greatest linguistic clowns.
HOpefully it won't take much to gather these postings and submit them as
an article to one of the journals I mentioned, or, heck, *Language* as a
"discussion note."
David Thomas
>
>> Maybe, but a "cuckoo clock" is not a dog, nor were cuckoo clocks
>> around 10,000 years ago.
>
>have argued that Bengali kukur and Danish kukur were
>proof positive that cuckoo clocks existed 10,000 years ago.
Yes, ancient man domesticated what is now the modern cuckoo clock some many
thousands of years ago...
...their relatives, wild cuckoo clocks, can still be found roaming the
Himalayas, chiming the hours of the night...
- Vae
David Thomas
>
>> By the way, the Danish word is not even monomorphematic. It is
>> a compound of the onomatopoeia <kuk> '(the call of the) cuckoo'
>> and <ur> 'clock', 'watch' (through Low German and Old French from
>> Latin <hora>)
>
>
>a link to... what was it? kukur in Bengali? Yes, "dog" in
>Bengali. Cuckoos have feather, dogs fur, the word for "feathers"
>in most Austronesian languages is also good for "hair" and
>"fur". Cuckoo the feathered one, kukur the furred one,
>natural semantic shift... Danish cuckoo clock = Bengali dog!
And they're clearly the same thing!
Heh...
- Vae
David Thomas
Scott) writes:
>>It's not unrealistic, and it's exactly what researchers do do
>>in other fields.
>
>Oh? How many research biologists do you see spending their time
>debunking creationism? How many researchers bother to debunk the
>likes of Erich von Däniken, Graham Hancock, and Robert Bauval?
>How many serious historians and archaeologists bother to write up
>detailed refutations of the crap that Adrian Gilbert puts out?
>In each case you can probably find a tiny handful, but they are
>very much the exception.
This is why we need people like James Randi...
I don't know that he's ever done anything on linguistic work, though. He seems
more concentrated on the physics/medicine aspect of pseudoscience, since these
are the ones that sell more to the, well, dumb masses.
- Vae
Hirofumi Nagamura
> Larry Trask wrote:
>
>>21. PUTI 'vulva'
>
>>Basque <poto-rro> 'pubis, vulva'
>
>
> Thirty years ago and more I shocked a Japanese
> visitor to ANU telling him that "Hotoke" (Buddha)
> meant "pubic hair" (hoto = penis/vulva, ke = hair).
Are there any attested cases of "hoto" being used for "penis"? I've
never seen it used for anything except "vulva".
> Gifted with a sense of humour (like most Japanese)
> he soon recovered, adding "you are polluting
> our language!"
I wonder just what our ancestors were thinking of when they imported
Buddhism. Perhaps they were looking to buy the latest pornography
from the continent, and got a religion instead.
"Hoto" is a dead word these days, but back in the days when it was
still alive, words like "hotoke" and "hotohoto"(!) must have been a
rich source of awful puns.
:-)
Cheers,
--
Hirofumi Nagamura
Kobe, Japan
Hirofumi Nagamura
>
> Japanese: mizu/oyo(cold water)
"Oyo"? Really?
Are you sure this isn't "o-yu" (honorific o- + yu "hot water")?
Brian M. Scott
Thomas) wrote:
They're also more dangerous, and they're where his expertise is
most useful.
Brian
Jarel
>Are you sure this isn't "o-yu" (honorific o- + yu "hot water")?
>
Yeah, you are right I looked at the wrong line, mizo is cold water and oyu is
hot water. Hmmm, if the o- is honorific, what is yu? Mizo is not a Chinese
loan is it?
David Thomas
Scott) writes:
Very true... don't know many people who've died from bad linguistics, though I
know it certainly puts a headache in my brain.
- Vae
Jarel
>You must be half asleep for confusing water with soup! ("miso")
My Japanese is terrible despite being half Japanese!
>In Norman French "water" is "yow". That's where your "yu" comes from,
obviously.
That has to be it! When the Normans invaded England is 1066, a small group
sailed to the Japanese. These Normans shortly took over and established a new
dynasty. Doesn't everyone know that? Look at these cognates:
Norman: Japanese: English:
yow yu water
dis to: ten
houit hachi eight
sčt shichi seven
yeun yon (4) one
Obviously Japanese "yon" was originally 'one' until the semantic shift to
'four'.
;) LOL
Athel Cornish-Bowden
> >
> > b.s...@csuohio.edu (Brian M. Scott) wrote in message news:<3f5fb96f...@enews.newsguy.com>...
> > > On 10 Sep 2003 05:40:54 -0700, at...@ibsm.cnrs-mrs.fr (Athel
> > > Cornish-Bowden) wrote:
> > >
> > > [...]
> > >
> > > >The point of all this is that experts (in any field) have an
> > > >_obligation_ to draw
> > > It's nice when they do, but it's not something that can
> > It's not unrealistic, and it's exactly what researchers do do
>
> Name some. S. J. Gould said he didn't have the time to devote to
> debating creationists.
>
As has already been remarked in this thread, you can't go away from his
books with the feeling that he has much time for creationism. Actually he
and other leading theoreticians in evolutionary biology have devoted
a great deal of time and energy to combatting creationism. To ask, as
Brian M. Scott does, "How many research biologists do you see spending
their time debunking creationism?", with the implied answer "none" , is
either profoundly ignorant or a slander.
As for Gould, he devoted a lot of energy to discussing the views of people
he disapproved of, sometimes usefully, as in "The Mismeasure of Man",
and sometimes scurrilously, as when he associated himself, his fame and
his prestige with the hounding of E. O. Wilson. In any case, I think you
have misunderstood the remark of Gould that you quote: I don't think it
meant that he didn't have the time to take apart the arguments of
creationists in his written work (an interpretation that is clearly belied
by the facts), but that he wasn't prepared to go to public meetings
packed with religious nutters.
But in any case I wasn't thinking of Gould, whose standing in his field is
a lot lower than readers of his popular books imagine. I was thinking
more of people like Dan Dennett, who devoted quite a few pages in one
of his books to a discussion of some of the weak logic in Gould's work,
and of John Maynard Smith, quoted at the beginning of the article by
John Tooby and Leda Cosmides at http://cogweb.ucla.edu/Debate/CEP_Gould.html,
and indeed Tooby and Cosmides themselves.
One could also mention Noam Chomsky, who at the time of his attack on
B. F. Skinner was by no means the famous person he is now, being far
less well known than his target.
In medicine, there are certainly some serious researchers who devote
some of their energies to the losing battle against the deluge of popular
books purveying quackery of one kind or another.
As for myself, I don't have the standing in evolutionary biology that would
justify my attacking purveyors of misconceptions about it (unless they
are very biochemical in nature, which most of them are not). However,
when the occasion arises and it is appropriate I certain spend time
trying to popular correct errors and misconceptions in fields that I
know about. On my web site there are at least four pages
(http://bip.cnrs-mrs.fr/bip10/hoyle.htm,
http://bip.cnrs-mrs.fr/bip10/enthalpy.htm,
http://bip.cnrs-mrs.fr/bip10/rose.htm,
http://bip.cnrs-mrs.fr/bip10/valevfaq.htm) that are primarily
concerned with this, three of them based on publications; the fourth is
not because it concerns nonsense that has itself only been presented in
news groups, not in print.
Des Small
Does Gould really never go into polemical debunking mode? In any
case, a better example would have been talk.origins, where there
really were proper scientists bebunking creationist nonsense
(including this "intelligent design" malarkey, which hasn't caught on
over here, hoorah) on a regular basis when I used to read it a couple
of years ago.
> >> How many researchers bother to debunk the
> >> likes of Erich von Däniken, Graham Hancock, and Robert Bauval?
> >> How many serious historians and archaeologists bother to write up
> >> detailed refutations of the crap that Adrian Gilbert puts out?
> >> In each case you can probably find a tiny handful, but they are
> >> very much the exception.
>
> >I'm not a particularly keen kook-watcher, but isn't von Däniken famous
> >for claiming that aliens helped build the pyramids and so on?
>
> Von Däniken claimed an amazing variety of nonsense; some of it is
> superficially credible. Some of Gilbert's stuff is in my opinion
> entirely comparable to Ruhlen's. And the fact remains that this
> sort of crap has a considerable following, however absurd you or
> I may find it.
Sure. But in the UK, at least, that kind of stuff is shelved (in
bookshops, libraries, etc) under "esoteric" or "unsolved mysteries" or
something. There's a kind of unspoken agreement on all sides that
this stuff and proper archeology/science have, let's say,
complementary niches in the market.
> >Ruhlen's proposals not only violate no laws of physics, they sound far
> >from superficially absurd. How many of your favourite archeological
> >kooks have free reign in the NYT, and have fooled Murray Gell-Mann and
> >allegedly prominent geneticists? If a creationist scam reached these
> >levels of apparent respectability I'd bet there'd be a lot more
> >biologists than Dawkins going ballistic.
>
> I'm not sure that ID hasn't reached a comparable level of
> apparent respectability in the lay world, at least in the U.S.,
> but I still don't see many research biologists spending a lot of
> time on public debunking.
The main battleground in the US, as I understand it, is what's taught
in high-schools, and the activism on both sides is mostly conducted at
ground roots level - election of school governors, and so on. The
talk.origins FAQ, a little Dawkins and a reasonable level of science
understanding should be enough for a DIY job.
(Intelligent design is addressed at
<http://www.talkorigins.org/faqs/cosmo.html> and there's a whole new
sister site dedicated to it at <http://www.talkdesign.org>)
> >I would say that the "cold fusion" affair is a closer match in terms
> >of perception by raw lay-persons, and those guys certainly got stomped
> >on pretty good.
>
> But I don't think that cold fusion is a close match in terms of
> professional perception. As I understand it, the stomping was so
> enthusiastic at least partly because it and its 'discoverers'
> couldn't -- quite -- be dismissed out of hand.
You are surely right about professional reception, but what about
_perceived_ professional perception? One of the more interesting
aspects of the cold fusion was precisely that it was announced at a
Big Event press conference before peer review got a shot at it, and so
a misclaim escaped into the wild and had to be hunted down and shot
before it wreaked too much havoc. Persons here have in the past
remarked that Ruhlen eschews publication in peer-reviewed journals
(who would presumably thrust eschewment on him if he didn't) in favour
of interdisciplinary conference proceedings, where he's the closest
thing to a linguist in the first place.
The modern semi-plausible pseudoscientist is learning to take his
"case" directly to the public, and not without success. And while
there are limits to how much pro bono policing of the public mind
scientists can be expected to engage in, I will certainly feel better
when Larry Trask succeeds to get his debunking in print, and better
still if it's in a book suitable for the laity.
[...]
Des
wants to hear less nonsense about language from the NYT and BBC.
--
des....@bristol.ac.uk
Athel Cornish-Bowden
>
> [on Ruhlen's failings]
>
> > Well maybe, but how is the ordinary reader (not an expert in language
> > history) to know?
>
> The ordinary reader has no way of knowing, unless informed
> commentators make the truth public.
greeted with such hostility.
[ ... ]
> Is C-S really a "leader in his field"? He is certainly prominent.
> I'm not competent to judge genetic work, but I have heard persistent
> rumors that other geneticists do not think much of C-S's work.
not think in the Manichaean terms that pervade much of this
thread: they agree with some of the things people say, disagree about
others, sometimes profoundly, sometimes only about certain
details. So, I don't think you're going to find any population geneticists
who will say that his writings are the voice of God, or any who will
characterize him in the sort of words you use about Ruhlen.
>
[ ... ]
> > fraudulent, and
> > if those who do have such expert knowledge refrain from publicizing
> > it, how is
> > the ordinary reader to know?
>
> We don't refrain from publishing. I would dearly love to publish an
> exposé of Ruhlen's crimes, but where? The New York Times is simply
> not going to be interested. Journalists do not run stories along the
> lines of "Nothing very interesting going on after all".
immediately above. You don't need the approval of the New York Times
to set up a web page, and the development of the World Wide Web
probably represents the biggest extension of free speech the world
has yet seen.
> > The point of all this is that experts (in any field) have an
> > _obligation_ to draw
> > attention to errors (and still more to fraudulent data) in books in
> > their subjects
> > that achieve popular success, how ever distasteful this may be, and
> > how ever
> > much easier it may be just to ignore the books they think are packed
> > with
> > nonsense.
>
> And that's exactly what I'm doing here.
illuminating, educational and stimulating (as I have found the two of
your books that I have read). And I intend to keep reading what you may
write in the future in this group (even if I think twice in future about
lifting my head above the parapet). But you still need to write it up
somewhere where ordinary readers are likely to read it. I suspect
that not many non-linguists follow this group, and those that do are
not made to feel very welcome if they ask any questions.
Des Small
> Des Small wrote:
>
> > Ruhlen's proposals not only violate no laws of physics
>
> They violate laws of elementary mathematics.
The public will have a secure grip on the differential
geometry of infinite dimensional spaces long before it twigs to
probability. (Probably.)
> They violate "laws" of elementary comparative linguistics,
> so elementary that any layman can rediscover them himself
> without outside help, without prior knowledge of linguistics,
> comparative or other.
Well, apparently not, isn't it? It's not as if Ruhlen has Eternal Life
or Alien Intelligence or a Priviledged Place In The Cosmos for Your
Particular Tribe, or any of the usual reality distortion fields that
so often make kookery transparently self-serving. Even as Profound
Esoteric Wisdom it isn't all that.
Des
keeps his word lists under a pyramid facing north, for added rigour.
--
des....@bristol.ac.uk
David Thomas
>
>> This is why we need people like James Randi...
>
>> I don't know that he's ever done anything on linguistic work, though.
>
>but not legerdemot.
Heh heh.
He's not much of a word magician, that's true.
- Vae
Peter T. Daniels
James Randi was used by Thomas A. Sebeok to derail his own conference on
the Clever Hans phenomenon at the New York Academy of Sciences ca. 1980.
When the primate communication researchers arrived in New York and
discovered that Randi was the featured speaker, some refused to
participate at all.
He presented a very entertaining magic show, but the relevance of
debunking Uri Geller to people who thought Washoe signed ASL was never
explained.
Peter T. Daniels
That's _exactly_ what I said. He did not debate creationists.
> But in any case I wasn't thinking of Gould, whose standing in his field is
> a lot lower than readers of his popular books imagine. I was thinking
> more of people like Dan Dennett, who devoted quite a few pages in one
> of his books to a discussion of some of the weak logic in Gould's work,
> and of John Maynard Smith, quoted at the beginning of the article by
> John Tooby and Leda Cosmides at http://cogweb.ucla.edu/Debate/CEP_Gould.html,
> and indeed Tooby and Cosmides themselves.
>
> One could also mention Noam Chomsky, who at the time of his attack on
> B. F. Skinner was by no means the famous person he is now, being far
> less well known than his target.
Chomskysreviewofskinner'sverbalbehavior _made_ him a celebrity in
linguistics. *Syntactic Structures* certainly didn't; it took Robert
Lees's review of it to make it comprehensible to linguists.
> In medicine, there are certainly some serious researchers who devote
> some of their energies to the losing battle against the deluge of popular
> books purveying quackery of one kind or another.
>
> As for myself, I don't have the standing in evolutionary biology that would
> justify my attacking purveyors of misconceptions about it (unless they
> are very biochemical in nature, which most of them are not). However,
> when the occasion arises and it is appropriate I certain spend time
> trying to popular correct errors and misconceptions in fields that I
> know about. On my web site there are at least four pages
> (http://bip.cnrs-mrs.fr/bip10/hoyle.htm,
> http://bip.cnrs-mrs.fr/bip10/enthalpy.htm,
> http://bip.cnrs-mrs.fr/bip10/rose.htm,
> http://bip.cnrs-mrs.fr/bip10/valevfaq.htm) that are primarily
> concerned with this, three of them based on publications; the fourth is
> not because it concerns nonsense that has itself only been presented in
> news groups, not in print.
Relevance?
Peter T. Daniels
> > We don't refrain from publishing. I would dearly love to publish an
> > exposé of Ruhlen's crimes, but where? The New York Times is simply
> > not going to be interested. Journalists do not run stories along the
> > lines of "Nothing very interesting going on after all".
>
> I didn't say "publishing", I said "publicizing", as quoted by you
> immediately above. You don't need the approval of the New York Times
> to set up a web page, and the development of the World Wide Web
> probably represents the biggest extension of free speech the world
> has yet seen.
Good grief. Just look at what the crackpots who turn up here (I think
that's what Larry calls "nutters") say, citing websites. How do you
propose to distinguish the legitimate sites from the kooks?
Larry Trask
> > Cavalli Sforza is not a linguist,
>
> but if it was not I'm happy to make it explicit: Cavalli Sforza
> is not a linguist, but he is the world's leading expert on genetic
> diversity, and as such ...
> >so he is not a qualified authority to judge
> > Ruhlen's work.
>
> on genetic diversity.
C-S is in no way qualified to comment on linguistic work, and he makes
an utter fool of himself every time he tries.
I've already posted one example of his imbecilic remarks on language,
on the Great Vowel Shift. Here are a few more, from the same chapter.
Icelanders can understand Swedish and Danish.
(Hell; the *Swedes* can hardly cope with Danish.)
The reconstructed PIE word for 'mother' is <ma> (no asterisk).
(What is he reading?)
The common origin of languages is more safely judged by syntactic
features, such as word order, than by "phonetics and semantics".
(A novel idea.)
Phonetics (by which he plainly means 'phonology': he doesn't
understand the difference) is "quite impervious to the influence of
other languages."
(So *that* is why English failed to acquire a phoneme /v/ after the
Norman
Conquest.)
The large number of Latin words in English results in the first
instance from the Roman occupation of Britain.
(No joke: he really asserts this. Anybody still willing to defend
C-S's
right to pontificate on language matters?)
Basque is related to the South Caucasian languages.
(This is not even the Ruhlen-Bengtson line: C-S can't even
understand
the simple falsehoods peddled by the Gruesome Twosome.)
There is "only one system of personal pronouns" in the 600-odd
languages assigned by Greenberg to his "Amerind".
(Even Greenberg never went this far. C-S can't read.)
Study of the documents in the extinct "Elamitic" [sic] language, in
their "cuneiform alphabet" [sic], have proved that the language was
Dravidian.
(Even McAlpin never went that far.)
Nearly all the Dravidian languages have disappeared. Only a few
tribes continue to use them.
(Dravidian has over 160 million speakers and nearly 30 languages.)
The Romans never conquered Wales, which is why Welsh has survived.
(This guy really has a problem with the Romans.)
The French and the Catalans have dropped the last syllables of their
words because these people have a strong preference for short words.
(Ah, I suppose the Catalans must get the heebie-jeebies every time
they
have to pronounce the name of their capital city. All those
syllables...)
Phonetic laws have been proved and applied on many occasions,
particularly in the 19th century.
(When's the last time *you* applied a phonetic law? Were you
satisfied
with the outcome?)
There are many more of these. Moreover, at various places in this
chapter, C-S attempts to characterize the ordinary working practices
of historical linguists, of which he disapproves, since these differ
from what Greenberg and Ruhlen do. But his "account" is a monument to
ignorance and stupidity. It is an imbecilic and incoherent travesty
bearing no relation to linguistic work.
I see no way round it. Cavalli-Sforza is an idiot.)
Larry Trask
lar...@sussex.ac.uk
Larry Trask
sidekick?
On p. 65 of his 1994 Wiley book, Ruhlen gives a table of
"Dene-Caucasian" "cognates". About half of the 13 Basque items are
faulty, but the worst is the word for 'bark'. The real word is
<azal>, but Ruhlen doesn't give this. He gives <laka>, a word which
does not exist and which, once again, he has manufactured out of thin
air.
On pp. 26-27 of his 1994 Stanford book, and elsewhere, Ruhlen gives a
much larger table of "Dene-Caucasian" items. Let's look at a couple
of the Basque entries.
TONGUE
The Basque word doesn't fit, so he cites this: <mili-> 'lick'.
First, the verb he is trying to cite is <milikatu>, whose stem is
<milika->, not *<mili->. Ruhlen has unscrupulously chopped off the
last syllable, because it doesn't match anything in the other
languages. This is typical of Ruhlen: he mangles and mutilates his
data, in order to force "matches" that don't exist.
The form of this item shows that it cannot possibly be ancient, and in
fact the word is not recorded before 1800. Anyway, it is
transparently an expressive formation.
The verb <milikatu> is not isolated. It belongs to a little group of
words all built on the same stem. Investigation shows that the
central sense of this stem is 'be picky, be choosy, be fussy'. The
stem has given rise to a group of words centering around the notion of
eating sparingly or reluctantly. Among the senses are 'nibble at
one's food', 'eat sparingly', 'taste, sample', 'sniff' and 'lick'.
Since all these items occur together in the dictionary, Ruhlen can
hardly have failed to notice them. But only the 'lick' sense fits his
comparative purposes, so he singles that one out and silently
suppresses all the rest.
You *cannot* trust anything you see on Ruhlen's pages. It may look
impressive to untutored eyes, but it's fraudulent.
PAW
Ruhlen cites Basque <hatz> 'paw'.
Any dictionary will list half-a-dozen senses for <(h)atz>, but Ruhlen
has suppressed all the ones he finds inconvenient.
The word <(h)atz> has been thoroughly investigated. Its original
sense was 'track, trail, trace'. This sense survives today in only a
few places, but it was clearly once universal in the language, because
all varieties retain this sense in old compounds. For example, the
Bizkaian dialect, in which the free word lacks this sense, has
<oinatz> 'footprint' (<oin> 'foot') and <lorratz> 'track left by a log
which is dragged' (<lor> 'log which is dragged').
South of the Pyrenees, the word was used to construct postpositions
and adverbs like <atzean> 'behind' (<-an> locative) and <atzera> 'to
the rear, backward'
(<-ra> allative). This is a frequent grammaticalization process in
the world's languages. From these items, a new noun <atze> 'space
behind' has been extracted.
Otherwise, the free word has undergone a number of semantic
developments in various parts of the country. Among the regional
senses it has acquired are 'footprint', 'finger', 'itch', 'mange,
scabies', 'bottom edge of a garment', and finally 'paw'.
But 'paw', which is confined to the northern dialects, is the most
recent development. This sense is not recorded before 1852. Today
<hatz> is the ordinary northern word for 'paw'. But, in the three
centuries before 1852, northern literature -- and we have a great deal
of this -- does not know the word, and 'paw' is constantly expressed
by the two words <oin> 'foot' and <zango> 'leg'.
None of this is a secret. But Ruhlen doesn't do linguistics. He
decides in advance what he wants to find. He leafs through bilingual
dictionaries of languages he knows nothing about, searching for what
he requires. He picks out anything that suits him, and he suppresses
everything that doesn't fit. And he makes no attempt to find out
anything about his chosen words. In fact, I know from experience that
he angrily rejects any such information when it is offered. He
doesn't want to know the truth, because the truth is not convenient.
Larry Trask
lar...@sussex.ac.uk
Brian M. Scott
news:yyrjisny...@pc156.maths.bris.ac.uk in sci.lang:
>> >> >> [...]
Not having read anything like his entire output, I really
don't know. It doesn't matter, as I've not claimed that *no*
active scientists engage in popular debunking.
> In any case, a better example would have been talk.origins,
> where there really were proper scientists bebunking
> creationist nonsense (including this "intelligent design"
> malarkey, which hasn't caught on over here, hoorah) on a
> regular basis when I used to read it a couple of years ago.
I've read it for several years. Yes, there certainly are some
real scientists there. But that's a spare time activity, very
different from actually writing a book or getting up on stage.
>> >> How many researchers bother to debunk the
>> >> likes of Erich von Däniken, Graham Hancock, and Robert
>> >> Bauval? How many serious historians and archaeologists
>> >> bother to write up detailed refutations of the crap
>> >> that Adrian Gilbert puts out? In each case you can
>> >> probably find a tiny handful, but they are very much
>> >> the exception.
>> >I'm not a particularly keen kook-watcher, but isn't von
>> >Däniken famous for claiming that aliens helped build the
>> >pyramids and so on?
>> Von Däniken claimed an amazing variety of nonsense; some
>> of it is superficially credible. Some of Gilbert's stuff
>> is in my opinion entirely comparable to Ruhlen's. And the
>> fact remains that this sort of crap has a considerable
>> following, however absurd you or I may find it.
> Sure. But in the UK, at least, that kind of stuff is
> shelved (in bookshops, libraries, etc) under "esoteric" or
> "unsolved mysteries" or something. There's a kind of
> unspoken agreement on all sides that this stuff and proper
> archeology/science have, let's say, complementary niches in
> the market.
Not true here, except for the very worst of it.
>> >Ruhlen's proposals not only violate no laws of physics,
>> >they sound far from superficially absurd. How many of
>> >your favourite archeological kooks have free reign in the
>> >NYT, and have fooled Murray Gell-Mann and allegedly
>> >prominent geneticists? If a creationist scam reached
>> >these levels of apparent respectability I'd bet there'd
>> >be a lot more biologists than Dawkins going ballistic.
>> I'm not sure that ID hasn't reached a comparable level of
>> apparent respectability in the lay world, at least in the
>> U.S., but I still don't see many research biologists
>> spending a lot of time on public debunking.
> The main battleground in the US, as I understand it, is
> what's taught in high-schools, and the activism on both
> sides is mostly conducted at ground roots level - election
> of school governors, and so on. The talk.origins FAQ, a
> little Dawkins and a reasonable level of science
> understanding should be enough for a DIY job.
It *ought* to be ...
[...]
> The modern semi-plausible pseudoscientist is learning to
> take his "case" directly to the public, and not without
> success. And while there are limits to how much pro bono
> policing of the public mind scientists can be expected to
> engage in, I will certainly feel better when Larry Trask
> succeeds to get his debunking in print, and better still if
> it's in a book suitable for the laity.
Oh, so will I, especially since he writes well. But if he
were just starting out, I don't think that he could afford
even to think about writing such an account.
Brian
Jarel
> Will you send me a copy, signed by your good self then?
You know the information is so good I want to share it. Why don't you
publish it. LOL
Stephen C. Carlson
>> Peter T. Daniels wrote:
>>>How can you give any respect at all to Cavalli-Sforza, when he
publishes
>>>such a monumental fraud as the double-winged chart of the genetic
tree
>>>"aligned" with Ruhlen's linguistic "tree"? Quite aside from the
fact
>>>that Ruhlen's tree is absurd in the first place (which you wouldn't
be
>>>expected to know),
Quite apart from the absurdity of Ruhlen's tree, upon which I do not
wish to comment, the congruence of Cavalli-Sforza's published trees
has been investigated by others with favorable results.
Here's one article with abstract:
Penny,D., Watson, E.E., Steel, M.A., 1993
Trees from Languages and Genes are very similar
Sys. Biol. 42(3):321-328
Tree diagrams have been used to represent relationships between both
human populations and languages but only recently have the trees been
compared (Figure 1) and the claim made "of considerable parallelism
between genetic and linguistic evolution" (Cavalli-Sforza et al.
1988:6002). However in rejecting this claim it was pointed out
(O'Grady et al. 1989) that the original study did not have an
objective measure to support their conclusion of high similarity. But
the lack of an objective measure invalidates the denial (O'Grady et
al. 1989) just as much as it invalidates the claim that the trees are
similar (Cavalli-Sforza et al. 1988; 1989). The original work has
recently been supported by computer simulation to estimate the
probability of finding such agreement between the gene and language
trees. Tree comparison metrics are perhaps a more natural way of
comparing trees quantitatively and we have recently extended knowledge
of such metrics (Steel and Penny, 1993). We report here that an
objective tree comparison measure can resolve the issue of the
similarity of the gene and language trees from Cavalli-Sforza et al.
(1988) showing that they are indeed far more similar than expected by
chance.
Stephen Carlson
Peter T. Daniels
>
> Stephen C. Carlson wrote:
>
> > Quite apart from the absurdity of Ruhlen's tree, upon which I do not
> > wish to comment, the congruence of Cavalli-Sforza's published trees
> > has been investigated by others with favorable results.
I have no opinion whatsoever as to the "congruence" of C-S's trees
(whatever you mean by that). Even if C-S's trees are a perfect
representation of the entire lineage of humanity, and even if R's tree
were a perfect representation of the entire lineage of human languages,
the simple fact is that THEY DO NOT AGREE EVEN A TINY BIT.
> > Here's one article with abstract:
> >
> > Penny,D., Watson, E.E., Steel, M.A., 1993
> > Trees from Languages and Genes are very similar
> >
> > Sys. Biol. 42(3):321-328
> >
> > We report here that an
> > objective tree comparison measure can resolve the issue of the
> > similarity of the gene and language trees from Cavalli-Sforza et al.
> > (1988) showing that they are indeed far more similar than expected by
> > chance.
Of course. They were manipulated SO THAT THE TWIGS APPEAR IN THE SAME
ORDER. There is no agreement IN THE BRANCHING OF THE TWO TREES.
> I don't buy that. I did a careful study of the two trees
> in the Scientific American article and found at best no
> correlation, at worst a mildly negative correlation. Rummage
> through the LINGUIST and you'll find it. That Penny
> at al. got published in Sys. Biol. whatever does not
> impress me. Scientific mags do publish preposterous
> rubbish too.
You yourself published the LINGUIST List link here not so long ago,
which is how I found out you'd done it -- you did it some time before I
knew about LINGUIST List or sci.lang.
Brian M. Scott
<jg...@alphalink.com.au> wrote:
>Still can't find my article on LINGUIST, [...]
<http://linguistlist.org/issues/3/3-81.html#1>
Brian
Douglas G. Kilday
"Brian M. Scott" <b.s...@csuohio.edu> wrote in message ...
> On 11 Sep 2003 17:22:53 GMT, nah...@aol.comnotojunk (Jarel)
> wrote:
>
> [...]
>
> >Here are some words for water in
> >other languages:
>
> [...]
>
> >Proto-IE: akwa/widur
>
> That should be *wodr (oblique *wedn-, root *wed-), *wedo:r;
> *akWah2- is, I believe, securely attested only in Germanic and
> Italic and so quite possibly not PIE at all.
>
The Russian river-name <Oká> (one in Irkutsk, another flowing from Kursk
into the Volga) is a plausible Slavic reflex. I agree with your point,
however: this second water-root cannot be securely assigned to PIE.
DGK
Stephen C. Carlson
> Jacques Guy wrote:
> > Stephen C. Carlson wrote:
> > > Quite apart from the absurdity of Ruhlen's tree, upon which I do not
> > > wish to comment, the congruence of Cavalli-Sforza's published trees
> > > has been investigated by others with favorable results.
>
> I have no opinion whatsoever as to the "congruence" of C-S's trees
> (whatever you mean by that). Even if C-S's trees are a perfect
> representation of the entire lineage of humanity, and even if R's tree
> were a perfect representation of the entire lineage of human languages,
> the simple fact is that THEY DO NOT AGREE EVEN A TINY BIT.
I wanted you to have a citation of an opposing study. What
you do with it is your business.
> > > Here's one article with abstract:
> > >
> > > Penny,D., Watson, E.E., Steel, M.A., 1993
> > > Trees from Languages and Genes are very similar
> > >
> > > Sys. Biol. 42(3):321-328
> > >
> > > We report here that an
> > > objective tree comparison measure can resolve the issue of the
> > > similarity of the gene and language trees from Cavalli-Sforza et al.
> > > (1988) showing that they are indeed far more similar than expected by
> > > chance.
>
> Of course. They were manipulated SO THAT THE TWIGS APPEAR IN THE SAME
> ORDER. There is no agreement IN THE BRANCHING OF THE TWO TREES.
You've said something like this before, but it is hard to understand
due to the use of non-standard tree terminology (e.g. "manipulate[]"
and "twigs"). By "manipulat[ing] so that the twigs appear in the
same order", do you mean a transformation of the tree on the left
from Example 1 to Example 2?
Example 1:
|-- A B --|
|---| |---|
| |-- B A --| |
| |
|------ C C-------|
Example 2:
|-- B B --|
|---| |---|
| |-- A A --| |
| |
|------ C C ------|
> Jacques Guy wrote:
> > I don't buy that. I did a careful study of the two trees
> > in the Scientific American article and found at best no
> > correlation, at worst a mildly negative correlation. Rummage
> > through the LINGUIST and you'll find it. That Penny
> > at al. got published in Sys. Biol. whatever does not
> > impress me. Scientific mags do publish preposterous
> > rubbish too.
(For some reason, Jacques Guy's posts do not show up my newsreader.)
Needless to say, a more persuasive dismissal of the Penny et al.
article may be one that engages with the article rather than brings
up the fact that journals occasionally publish rubbish. I suppose
the LINGUIST list publishes rubbish too.
As to your LINGUIST study, although it used the term "correlation",
it provided no formula for calculating the "correlation" nor any
citation to literature pertaining to calculating the "correlation"
of different trees. Penny et al., on the other hand, did provide
a formula for their measure plus cites to the relevant literature
justifying their formula.
Stephen Carlson
Douglas G. Kilday
"Jarel" <nah...@aol.comnotojunk> wrote in message ...
> Douglas said:
>
> >If memory serves, Nahali has native
> >'1', but '2-5' are borrowed from Dravidian, and '6-10' from Indic.
>
> Actually, 2-4 are Dravidian and the rest are Indic. Also, those are
obvious
> loans though and do no cause any problem. What causes more headaches are
those
> which are more difficult to tell. For example, in Romani, are the
numerals for
> 6 and 10 from Dardic? Japanese has many numeral cognates with Tungustic
and
> Mongolian. I think that is a strong case for relationship than anything
else.
I just looked at the numerals on Zompist, and the only obvious resemblance I
could find was between Old Japanese *<nana> '7' and Proto-Tungus *<nadan>
'7' (with attested reflexes); nothing between OJ and Mongolian. One rather
lax "match" out of 20 possible isn't very impressive.
Now, it may be that I am simply not seeing more subtle phonetic patterns
shared by these sets of numerals, or that the "many numeral cognates"
involve forms other than the most common citations for '1-10' tabulated by
Mr. Rosenfelder. But even if Old Japanese _numerals_ can be convincingly
related to those in Tungusic and Mongolian, a strong case for common descent
of the _languages_ cannot so easily be made. If the numerals are indeed
related, one can envision continental Proto-Japanese tribes associating with
Proto-Tungusic and/or Proto-Mongolian tribes and acquiring their numerals,
just as Proto-Hungarians presumably associated with Iranians (Scythians?
Sarmatians?) and acquired theirs.
DGK
Peter T. Daniels
>
> Stephen C. Carlson wrote:
>
> > (For some reason, Jacques Guy's posts do not show up my newsreader.)
>
> show in dejanews (now taken over by google). And they
> don't. I have no idea why. I use Netscape 3.0 for
> safety reason, and there is nothing in Netscape 3.0
> that lets you fiddle so that your posts are not
> archived, or whatever. So, once again, search me.
I have Netscape 3.04 and have no problem seeing your posts. It must be
an ISP thing. Or, for David's benefit, thang.
Peter T. Daniels
Yes.
> Example 1:
>
> |-- A B --|
> |---| |---|
> | |-- B A --| |
> | |
> |------ C C-------|
>
> Example 2:
>
> |-- B B --|
> |---| |---|
> | |-- A A --| |
> | |
> |------ C C ------|
But what the C-S/R chart does is more like
>
> |-- B B --|
> |---| |---|
> | |-- A A --| |-------------------
> | |
> |------ C C -----=|=-----|
> | |
> |-- D D ------| |
> |---| |
> | |-- E E -------------|------------
> | |
> |------ F F --| |
> |----------|
> |-- G G --|
> |---|
> | |-- H H ---------------------------
> |
> |------ I I --|
> |----------|
> |-- J J --| |-------------
> |---| |
> | |-- K K -------------|
> |
> |------ L L ---------------------------
The branchings simply do not agree.
> > Jacques Guy wrote:
> > > I don't buy that. I did a careful study of the two trees
> > > in the Scientific American article and found at best no
> > > correlation, at worst a mildly negative correlation. Rummage
> > > through the LINGUIST and you'll find it. That Penny
> > > at al. got published in Sys. Biol. whatever does not
> > > impress me. Scientific mags do publish preposterous
> > > rubbish too.
>
> (For some reason, Jacques Guy's posts do not show up my newsreader.)
>
> Needless to say, a more persuasive dismissal of the Penny et al.
> article may be one that engages with the article rather than brings
> up the fact that journals occasionally publish rubbish. I suppose
> the LINGUIST list publishes rubbish too.
What possible defense could Penny et al. for claiming that two trees
whose branchings don't agree are "very similar"?
> As to your LINGUIST study, although it used the term "correlation",
Are you saying you tracked down Jacques's list of disagreements? A link
would be polite.
> it provided no formula for calculating the "correlation" nor any
> citation to literature pertaining to calculating the "correlation"
> of different trees. Penny et al., on the other hand, did provide
> a formula for their measure plus cites to the relevant literature
> justifying their formula.
Peter T. Daniels
Thanks ... this time I've bookmarked it. If Mark were still among us, we
could ask him to put it in the FAQ.
Peter T. Daniels
>
> Peter T. Daniels wrote:
>
> > You yourself published the LINGUIST List link here not so long ago,
> > which is how I found out you'd done it -- you did it some time before I
> > knew about LINGUIST List or sci.lang.
>
>
> -------------------------------------------------------------------
> Message 2: Cavalli-Sforza et al. in "Le Monde"
>
> Date: Mon, 13 Apr 92 09:51:49 ESCavalli-Sforza et al. in "Le Monde"
> From: Jacques Guy <j....@trl.oz.au>
> Subject: Cavalli-Sforza et al. in "Le Monde"
>
> There was in "Le Monde" of 26 March an article by Yvonne Rebeyrol
> entitled "Cousins proches, mais uniques" in which the phylogenetic
> theories of the Sforzas and Allan Wilsons are discussed. It is
> extremely well written, clear, and informative. It is difficult, after
> having read it, to still believe that Wilson, Sforza et al. did not
> deliberately refuse to take into account known evidence, and known
> genetic mechanisms that did not fit their aims. In particular, you will
> discover in that article that mitochondrial DNA does carry male genetic
> information too, contrary to what is claimed in that famous Scientific
> American article, and that geneticists had known it since the 70s. The
> editors of "Le Monde" thought the topic important enough to include
> Yvonne Rebeyrol's article in their "Selection hebdomaire"
> (international edition). It's in there, p.10 (Le Monde, Selection
> hebdomadaire No.2265 du jeudi 26 mars au mercredi 1er avril).
> ------------------------------------------------------------
>
> I had completely forgotten about that article of mine.
> I supposed that there others floating about, equally
> forgotten.
I don't see a ref. to an article by JG, unless you mean this exact
paragraph from (presumably) LINGUIST.
Larry Trask
> >such a monumental fraud as the double-winged chart of the genetic tree
> >"aligned" with Ruhlen's linguistic "tree"? Quite aside from the fact
> >that Ruhlen's tree is absurd in the first place (which you wouldn't be
> >expected to know),
That chart is indeed a disgrace. Anyone who looks at it can see that
the correlations claimed by the authors do not exist. This is another
reason for entertaining grave doubts about C-S's trustworthiness.
> I think you'll find that great many people give respect to Cavalli-Sforza,
> and not just biologists: linguists also. For example, at
> http://www.cogs.susx.ac.uk/users/larryt/basque.faqs.html you can
> read the following words:
>
> "Recently, however, the geneticist Luiga Luca Cavalli-Sforza has completed
> a gene map of the peoples of Europe, and he finds the Basques to be strikingly
> different from their neighbors. The genetic boundary between Basques and
> non-Basques is very sharp on the Spanish side. On the French side, the
> boundary is more diffuse: it shades off gradually toward the Garonne in
> the north. These findings are entirely in agreement with what we know of
> the history of the language."
>
> These words don't read to me as if written by someone who gives no
> respect at all to Cavalli-Sforza.
I wrote those words many year ago, when C-S had not yet published most
of his linguistic idiocies, or at least when I hadn't yet seen them,
and when I was taking it for granted that C-S was a reliable
authority. I still have no particular reason to doubt C-S's findings
in this case, and in any event I am not qualified to judge work in
genetics -- though recall those rumors about C-S and his fellow
geneticists. As it happens, for reasons I won't go into here, I have
been unable to update my Website.
But I think you can safely assume that my recent words supersede my
earlier words, in any cases in which there is a conflict.
Larry Trask
lar...@sussex.ac.uk
Jarel
>I just looked at the numerals on Zompist, and the only obvious resemblance I
could find was between Old Japanese *<nana> '7' and Proto-Tungus *<nadan> '7'
(with attested reflexes); nothing between OJ and Mongolian.
I guess you aren't really looking. There are quite a few cognates:
Jap. Tun. Mon. Tur
1 pito bir
4 yo/do dü- dör- tört/dört
5 itu tu- ta-
6 mu ñö-
7 nana nada-
8 ya/da jap-
9 kokono xüyä-
10 towo juwan
To save room, I have only included the more relevant first part of the
numeral word. OK, 9 might be a stretch, but Japanese numerals share the most
cognates with Tungus. I once did a paper and I found regular sound
correspondences between Japanese and Tungus.
Different numeral roots in one language family is not uncommon.
Proto-Indo-European has two roots for "one", "four", and some Iranian languages
have a different word for "nine". Yes it is possible that Japanese borrowed
from Tungus or that Tungus borrowed some from Mongolian. I have looked at many
many numerals systems and for the lower numerals, borrowing is not uncommon,
but it is usually very obvious. Off the top of my head here are the only IE
languages that have borrowed lower numerals:
Pomak (From Turkish)
Romani (From Greek and Dardic)
Domari (From Persian)
A few Munda languages borrowed from Indic.
Look at this East African situation:
Hadza Sandawe Asa Nandi
3 samakapi somki-x samak somok
Asa is Rift Cushitic and the other languages in Rift C. have tam/tham/tami
for 3. Nandi is Southern Nilotic and all the languages have a similar root for
3.
Daic only retains native 1, 2, and I think 5.
I would love to know what is going on in Molale and Cayuse! The two are not in
the same family, so are the similarities due to borrowing?
Brian M. Scott
<fuf...@chorus.net> wrote:
>"Jarel" <nah...@aol.comnotojunk> wrote in message ...
[...]
>> Japanese has many numeral cognates with Tungustic and
>> Mongolian. I think that is a strong case for relationship than anything
>> else.
Roy Andrew Miller makes a point of the difficulty of the numeral
words, in contrast to the usefulness for reconstruction of the
pronouns and interrogatives.
>I just looked at the numerals on Zompist, and the only obvious resemblance I
>could find was between Old Japanese *<nana> '7' and Proto-Tungus *<nadan>
>'7' (with attested reflexes); nothing between OJ and Mongolian. One rather
>lax "match" out of 20 possible isn't very impressive.
In Miller's view '[t]he words for '4' offer the most complete
representation of a uniform, original numeral root throughout the
Altaic languages, including Japanese', though even here the
Turkish forms show modification from what would have been
expected. From his Table 7:
Old Turkish (Orkhon): tört
Chuvash: tãvatã
Written Mongol: dörben
Dagur: durube
Proto-Tungus: dügün
Manchu: duin
Old Japanese: yö-tu
Proto-Altaic: *dör-
Old Turkish: tört (loanword)
Proto-Mongol: *dör-
Proto-Tungus: *du-
Old Japanese: yö-
Here <ã> is a-breve. The first part of the table evidently
contains the data, the second part the reconstructions or
isolated morphemes. I don't know what the Proto-Tungus in the
first part of the table is (and haven't looked very hard for an
explanation). According to Miller, OJap /y-/ is the regular
reflex of */d-/ before original front vowels, so that all of this
except OTurk is regular.
>Now, it may be that I am simply not seeing more subtle phonetic patterns
>shared by these sets of numerals, or that the "many numeral cognates"
>involve forms other than the most common citations for '1-10' tabulated by
>Mr. Rosenfelder. But even if Old Japanese _numerals_ can be convincingly
>related to those in Tungusic and Mongolian, a strong case for common descent
>of the _languages_ cannot so easily be made. If the numerals are indeed
>related, one can envision continental Proto-Japanese tribes associating with
>Proto-Tungusic and/or Proto-Mongolian tribes and acquiring their numerals,
>just as Proto-Hungarians presumably associated with Iranians (Scythians?
>Sarmatians?) and acquired theirs.
Another quotation from Miller:
In this same connection, it is important to notice that
the Altaic numerals, including those of Japanese,
provide excellent evidence showing that numerals
in many, if not most, languages are in no way 'basic',
in the sense of being particularly resistant to or
inhibiting of borrowing. The concept that numerals
are somehow a part of a 'basic' stratum of
vocabulary, and that they are less susceptible to
borrowing than other segments of the lexicon, has
its origins in Indo-European studies; but it is
plainly mistaken as well as misleading when taken
over bodily and applied to the study of languages
in other parts of the world, particularly in the Far
East. The Chinese numerals have been borrowed
into many other languages -- Japanese and Thai
are both excellent cases in point; and there is
additional evidence from as far afield as Dravidian
(Emeneau 1957), all pointing in the same direction.
Numerals in this part of the world seem fully as
likely to undergo borrowing as are any other sets
of linguistic forms; and it is clear that they are in
no sense of the term 'basic vocabulary'.
The discussion that follows of how numeral terms have been
modified, especially by pairing, is quite interesting.
Brian
Brian M. Scott
<jg...@alphalink.com.au> wrote:
>Stephen C. Carlson wrote:
>> (For some reason, Jacques Guy's posts do not show up my newsreader.)
>Search me. Yusuf Gursey also complained that they do not
>show in dejanews (now taken over by google). And they
>don't. I have no idea why. I use Netscape 3.0 for
>safety reason, and there is nothing in Netscape 3.0
>that lets you fiddle so that your posts are not
>archived, or whatever. So, once again, search me.
Here are the headers from the post to which I'm replying, minus
the Path and References headers:
Message-ID: <3F63E5...@alphalink.com.au>
Date: Sat, 13 Sep 2003 20:50:33 -0700
From: Jacques Guy <jg...@alphalink.com.au>
Reply-To: jg...@alphalink.com.au
Organization: rather disorganized
X-Mailer: Mozilla 3.0 (Win95; I; 16bit)
MIME-Version: 1.0
Newsgroups: sci.lang
Subject: Re: Ruhlen's way with words
[References goes here]
Content-type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
NNTP-Posting-Host: d60-ds2-mel.alphalink.com.au
X-Original-NNTP-Posting-Host: d60-ds2-mel.alphalink.com.au
X-Trace: news.alphalink.com.au 1063450234 202.161.99.252 (13 Sep
2003 20:50:34 +1000)
Lines: 42
Admin: ad...@alphalink.com.au
Abuse: ab...@alphalink.com.au
X-No-Archive: Yes
[Path goes here]
Xref: spln sci.lang:214623
As you can see, somehow the X-No-Archive header has been set to
'Yes', and Google does honor it.
[...]
Brian
Brian M. Scott
wrote:
> Douglas said:
>>I just looked at the numerals on Zompist, and the only obvious resemblance I
>could find was between Old Japanese *<nana> '7' and Proto-Tungus *<nadan> '7'
>(with attested reflexes); nothing between OJ and Mongolian.
> I guess you aren't really looking. There are quite a few cognates:
> Jap. Tun. Mon. Tur
>1 pito bir
>4 yo/do dü- dör- tört/dört
There seems not to be a do- form in OJap.
>5 itu tu- ta-
>6 mu ñö-
Miller says that OJap mu- has no secure Altaic connections and
marks Proto-Tungus *ñöN- as a loanword; according to the text
it's a borrowing of a Proto-Mongol compound '2 times 3'.
>7 nana nada-
Miller considers nada- a borrowing into Proto-Tung. from
Proto-Mongol.
>8 ya/da jap-
Similar comment as for '4'. Miller gives *3^a- (where '3' is
yogh and '^' is hachek) for PTung.; this would be cognate with
OJap ya-
>9 kokono xüyä-
Miller takes OJap kökönö- and PTung *xüjägün to be parallel
compounds, both meaning essentially '3 times 3'.
>10 towo juwan
OJap töwo, PTung *3^uw-. Miller sees them as cognate but also
sees some problems: the OJap /t-/ goes with the back vocalism of
the PTung root, but the /ö/ is a front vowel. He wonders whether
the OJap was originally *towo, giving a possible reason for the
change.
[...]
Brian
Jarel
>There seems not to be a do- form in OJap.
I thought P-Japanese y- went back d-. The Koguryo data is interesting and
obviously Japanese. 10 is "tek". It is funny because Ruhlen had the Koguryo
data as Old Korean!!!!!
Jarel
>In this same connection, it is important to notice that the Altaic numerals,
including those of Japanese,
I don't think Altaic is a valid language family anyway. I have never seen
great evidence like regular sound correspondences. I just think: Tungus,
Mongolian, Turkic, Korean, and Japanese form(ed) a linguistic area and not a
language family.
>The discussion that follows of how numeral terms have been modified,
especially by pairing, is quite interesting.
Yes, the Japanese case was: mittsu/muttsu - 3/6 yottsu/yattsu - 4/8. I
think Miller is the one that mentions that Japanese two, putatsu may have been
*tutatu? I wonder if pitotu "1" some how contaminated "*tutatu" and also too
many t's like kokonotu had too many k's.
Jouni Filip Maho
>
> Merde alors. Look, when Yusuf Gursey said that I did go through
> all the options in the "Options" tab of Netscape 3.0 and I found
> nothing. Has anyone any idea what's going on there and how I can
> fix it? (short of using Internet Explorer of course)
See if you have a program called "Internet Config" on your computer. If
you do, open it, click on "News" and delete anything appearing under
"News headers".
If that doesn't work, post a query on some approriate newsgroup.
---
jouni maho
benlizross
>
> Stephen C. Carlson wrote:
>
> > (For some reason, Jacques Guy's posts do not show up my newsreader.)
>
> show in dejanews (now taken over by google). And they
> don't. I have no idea why. I use Netscape 3.0 for
> safety reason, and there is nothing in Netscape 3.0
> that lets you fiddle so that your posts are not
> archived, or whatever. So, once again, search me.
>
Hmm...while we're on this subject...I use Netscape 3.04, as it happens.
Jacques' posts come through fine, but the time displayed on the inbox
listing is out of whack with everyone else's -- apparently 19 hours
ahead of the time given in the date line in the message. Thus the
displayed times are often in the future relative to when I'm reading
them. This may just be a consequence of the extremely advanced thinking
embodied in Dr.Guy's contributions. Or is there some other explanation?
Ross Clark
Peter T. Daniels
That suggests he hasn't set the time zone on his computer -- the date &
time look right to him, but not to the rest of the world. (I found out
about that quirk when my messages were doing the same thing.) It might
have shipped set for San Francisco or something.
Kai Henningsen
> Scientific dishonesty apparently takes the exact same shape--experimental
> fudging of data--whenever it appears.
No, that's only one of the three usual forms, the second being fudging
statistics (and from what people say, this seems also active in this
case).
The third, of course, is plagiarism.
Kai
--
http://www.westfalen.de/private/khms/
"... by God I *KNOW* what this network is for, and you can't have it."
- Russ Allbery (r...@stanford.edu)
Yusuf B Gursey
there isn't anything irregular of OTurk. . oghuz (incl. turkish) has
do"rt
(*do":rt). oghuz voiced intials are also found in brahmi script old
uighur and elsewhere. it just that some turkic languages avoid voiced
initals or perhaps they were only weakly voiced. for example, 19th
cent. arabic script uzbek has them, earlier uzbek and cyrillic uzbek
doesn't.
Yusuf B Gursey
> Brian said:
>
>
> >In this same connection, it is important to notice that the Altaic numerals,
> including those of Japanese,
>
> I don't think Altaic is a valid language family anyway. I have never seen
> great evidence like regular sound correspondences. I just think: Tungus,
> Mongolian, Turkic, Korean, and Japanese form(ed) a linguistic area and not a
> language family.
an alternative theory is that the group just broke up earlier than IE,
and so some of the expected cognates have been lost.
I
nah...@aol.comnotojunk (Jarel) wrote:
> I don't think Altaic is a valid language family anyway. I have never seen
>great evidence like regular sound correspondences. I just think: Tungus,
>Mongolian, Turkic, Korean, and Japanese form(ed) a linguistic area and not a
>language family.
As I recall, Vovin claims some Altaic body part etymologies, and also
that the Korean-Japanese-Tungus connection is more secure than those
with Turkic and Mongolian. I don't know anything about the subject, but
Vovin seems careful and conscientious. Some of his comments on Altaic
are on old Linguist List messages.