Re: Genetics and language, ugh

From: jdcroft
Message: 13820
Date: 2002-06-11

Glen wrote language and Genes do not *ALWAYS* correlate. Glen, it is
good to see you are prepared to come some of the way. I have never
argued that they *always* did, only that they often do, and
sufficiently often to be statistically discernable within populations.

The question is "How frequently" do they vary in common. On the basis
of the work of the Human Genome Diversity Project, it is sufficiently
common to enable a comparison of genetic distance between differing
populations, and a linguistic phylogenic tree of languages to be
usefully compared.

You stated
> There is no "English" or "Chinese" gene so why do you insist on this
> crap?

Of course there is no English or Chinese "gene". I have never
remotely argued that there was. But Englishwomen and Chinese women
both have genes, and if they emigrate they tend to carry both their
language and there genes with them.

You wrote
> Your a) automatically demonstrates that genes and language
> DO NOT NECESSARILY correlate... This is my point!
>
> Whether we decide the scenario was a) or b) requires
> special attention of whatever individual language that one
> is studying. With IE, it appears to be a healthy mixture
> of both a) and b), and indeed, this is the most popular
> scenario for the spread of any language, whether it be
> English, Spanish, Hindi or Mandarin.

Agreed, but even in case a) there is likely to be "a statistically
significant spread" of genetic markers in a population too. English
speaking in India (at the level of the total Indian population) could
be seen to tend to correlate with Anglo-Indians, a sizable percentage
of whom would have Y chromosomal links with England.

To my point 4
> >4. People of higher social standing are generally able
> >to command access to more diverse and better resources
> >than those of low standing. They are therefore likely
> >to leave more surviving offspring (all else being equal).

Glen wrote
> Only if we assume that scenario b) is prevalent enough
> to make a) a logically insignificant scenario. You must
> first prove the above for every language on earth both
> alive and dead before you can assert that language and
> genes always correlate. Have fun with that proof and I'll
> see you in fifty years.

Not at all. I am not arguing that a) was logically *insignificant*.
Rather I am arguing that b) is sufficiently sufficient enough to be
(at the level of populations) to be measurable. Measurement is
currently occurring, in Oxford, Stanford and elsewhere.

> Again, genes and languages do not always correlate.

Agreed, Glen. I never said they always did. But they do often enough
to make meaningful conclusions (on a case by case basis) possible.

I wrote

> >Glen, whilst in 3(a) linguistics and genetics are totally
> >independent (as you keep asserting),
>
> Wrong, I'm asserting that linguistics and genetics are
> are not dependent enough to give us any meaningful
> information without examining specifics of the languages
> we are dealing with.

Already the "meaningfulness" of the information is proving its worth,
for instance in the Han Chinese populations. Genetically they seem to
form two genetically distinct groupings - a north Chinese grouping,
and a south Chinese grouping that have been separate from each other
for at least 60,000 years. The northern grouping has more in common
with Central Asian (and ultimately European) populations than it does
with the South Chinese who have more in common with Austro-Asiatic and
Austronesian (and ultimately Australoid and Indo-Pacific) populations.
It is not surprising that the genetic boundary and the linguistic
boundary between northern Mandarin and southern other languages
correlates closely to this genetic boundary.

Glen wrote
> I accept both your 3a and 3b but this
> is the very reason why your attempts to blindly link genes
> and languages are misguided.

Glen, I am not doing so *blindly*.

> Your assumptions of absolute correlation are similar to the
> technique of mass comparison. It may help us pinpoint
> general facts about a given language but it certainly tells
> us nothing meaningful about the specifics. Here, genes can
> only give us general ideas about the movement of languages.

Glen, please listen.... I am not arguing for any absolute correlation.
Even you seem to have come a long way in my direction by saying
"genes can only give us general ideas about the movement of
languages". Glen, drop the word *only* and this is what I am saying!
Genes *can* give us general ideas about the movement of languages -
not universally, not in every case, but generally.

> But there is an entire slew of other problems that need to
> be properly looked at such as social considerations, local
> economic issues, prehistorical disasters such as the Lake
> Euxine event, and most especially the actual study of
> linguistics which you continuously demonstrate to know
> little about. You're not concerned with addressing these
> problems and they are heavily important to understanding
> the spread of any language, whether historical or
> prehistorical.

I do consider these events. "Social considerations, local economic
issues, prehistoric disasters" have an effect in that they all impact
demographically - on birth rates, death rates and immigration. Their
cumulative impacts shows up demographically.

> You're losing out on a big aspect of comparative linguistics. It's a
> multidisciplinary field. Genetics alone won't help you.

Glen, read my fingers - I never argued for genetics alone. Just for
genetics as a suppliment and complement for the linguistic work. At
the same time, if the genetics shows one thing and the linguistics
shows something contrary it can mean -

a. The genetic anaylsis is wrong and the linguistic interpretation is
right.
b. The genetic analysis is right and the linguistic interpretation is
wrong.
c. They are both wrong.
d. They are both right, and some other factor left out explains why.

Based upon the "hardness" of genetic data, I would argue that b. and
d. tend to be more often correct than a. and c. Nevertheless,
individual cases would depend very much according to local contexts
and history.

> The spread of Latin was clearly the combination of
> demic, genetic and linguistic spread, for gods sakes.
> They do not correlate! Can you honestly tell me that
> there wasn't an advantage to speaking the lingua franca
> as a foreign speaking citizen of the empire??
>
> Allolinguistic populations, even with different genes
> other than "Italian genes" (to quote your racist,
> ultimately meaningless phrase), were most definitely
> adopting Latin. By the way, please identify what
> these Italian genes are that you've discovered, mister
> armchair-geneticist, or forfeit your statement. What
> constitutes an "Italian"? What period are you speaking of?
> Is it 2000 BC? 1000 AD? What? What borders from which time
> period are you speaking of that might genetically define
> "Italians"? You're way out there in left field at this
> point and sounding more neo-something by the minute.

Glen, Gallic and Britannic legions tended to be recruited from the
region of Tuscany. Genetic researches are finding that haplotype T
probably had its origins in the Tuscan region. Some of the spread of
this haplotype in areas of the Roman Limes (Frontiers) and Coloni is
explicable by this fact. These people were those who established
Vulgar Latin as a Lingua Franca that you speak of.

> All that I'm saying is common sense so I can only shake
> my head if you don't understand my points.
>
> To repeat, genetics and language DO NOT **ALWAYS** correlate.
> This is different from "Genetics and language never
> correlate" which is what your entire message was wastefully
> responding to.

My message was responding to your argument that genetics and
linguistics only produce "junk correlations" that have no basis in
fact. I was arguing that the correlations, when they do exist (and
they frequently can) *ARE* meaningful.

Regards

John