proto-proto-proto...

[Victor Steinbok]

The links have already been posted. I just wanted to point to two more
"stories". Eugene Volokh also posted recently in response to a comment
that suggested common confusion of "epitaph" and "epithet"--more
specifically, the alleged nearly complete replacement of "epithet" with
"epitaph". EV calls "bullshit" on this one, and rightfully so. As he
writes, most of the items that have incorrect replacement appear to be
comments complaining incorrect replacement. Now, had someone suggested a
common confusion between "epigram" and "epitaph", it might be worth an
investigation... or not.

The second "story" popped up last night and seems to be everywhere now.
For example, CSM posts, "Europeans all shared a common ancestor just
1,000 years ago, new genetic study reveals." Really? All Europeans
shared a common ancestor 1000 years ago? From which parody site are they
getting this nonsense? The Onion? The Daily Currant? It sounds like
someone screwed up the delivery on research data and all the MSM picked
up the same mistake without realizing it or giving it a plausibility
check. I suppose, one could blame the grammatically weak construction
for that sentence (quantifier error)--it's not so much that /all/
Europeans had a single common ancestor, but /any/ Europeans could be
shown to be related if one goes back far enough (and for this study,
"far enough" is 1000 years).

Sure enough, Nature phrases things a bit differently:

http://goo.gl/Ap4cZ
> Whether they are a Serb and a Swiss, or a Finn and a Frenchman, any
> two Europeans are likely to have many common ancestors who lived
> around 1,000 years ago. A genomic survey of 2,257 people from 40
> populations finds that people of European ancestry are more closely
> related to one another than previously thought, and could help to
> bring about new insights into European history.

Apparently, this is a new anthropological proof of the near-mathematical
proposition that 1=many.

More from Nature:

> Gene-sequencing companies such as 23andMe, based in Mountain View,
> California, use this property to connect distant cousins enrolled in
> their databases. Ralph and Coop looked for even more distant relatives
> by identifying stretches of the genome shared by people living
> throughout Europe. By looking at the length of these chunks, the
> researchers were able to determine approximately when distant cousins’
> common ancestor lived.
>
> They found common ancestors from as recently as 500 years ago mainly
> within populations. Older stretches of DNA, however, connected more
> geographically distant Europeans.

This is not even close to the claim that there was a single common
ancestor 1000 years ago! It's not even close to a claim that a group of
common ancestors from 1000 years ago is responsible for all current
Europeans (unless that "group" includes the /entire/ European population
of that period). All that this really says is that if you take two
seemingly unrelated people from a large but contiguous geographic area
(in this case, Europe) and go far enough back up the family trees,
you're bound to find a common ancestor. But that's pairwise common
ancestry not "a single ancestor for all Europeans".

Boggles the mind...

CSM does try to make sense of it all.

http://goo.gl/RASNG
> The researchers found that all Europeans shared a common ancestor just
> 1,000 years ago.
> There were also some regional surprises.
> For instance, Italians are slightly less related to one another than
> people from other European countries are to one another, perhaps
> because Italians have had a large, fairly stable population for a few
> thousand years.
> In addition, people from the United Kingdom are more related to people
> from Ireland than they are to other people from the United Kingdom.
> That's possibly because many people have migrated from the smaller
> country of Ireland to the bigger United Kingdom in the past several
> hundred years, Coop told LiveScience.
> The researchers also showed that people in Eastern Europe were
> slightly more related to each other than were those in Western Europe.

Some of this might make sense, although it's hard to say whether it's
coming from the paper or from the reporter. It gets worse.

> That may be the signature of Slavic expansion and migrations, such as
> those of the Huns and the Goths, about 1,000 years ago, said John
> Novembre, a population geneticist at the University of Chicago, who
> was not involved in the study.

I'd be willing to bet that this is /not/ what John Novembre said. In
particular, "the Huns and the Goths" are not "such as" the Slavs,
irrespectively of whether the Slavs were expanding and migrating or not.

What the study may actually be pointing to is the superregional
migrations, but not of entire populations. Think of this as a bunch of
overlapping "neighborhoods" (similar to the mathematical sense, I
suppose), where people tend to migrate and intermarry, so that within
each neighborhood most people are in some way related. Two overlapping
neighborhoods will eventually be completely interrelated as well,
because genetic material from the non-overlapping parts will eventually
migrate into the overlapping part and, from there, into the other
neighborhood. Now, with multiple chains of such neighborhoods, the
interrelation becomes more complex, but also takes much longer to
propagate. However, in the end, all the neighborhoods will have some
genetic relations to each other--some more, some less. It should not be
complicated to set up a purely theoretical dynamic stochastic model that
would replicate the genetic research. If this is what the research
actually shows, the paper is not particularly sensational. But, thanks
to the MSM science writers, we now know better.

To add insult to injury, CSM story is followed by a link to ther "Are
you scientifically literate?" quiz. Actually, there are three links to
that "quiz" on the page.

The original paper is here: Ralph, P. & Coop, G. PLoS Biol. 11, e1001555
(2013).
http://goo.gl/ZobL6

VS-)

PS: I initially became aware of the story through a link on
TalkingPointsMemo.com which has a headline substantially the same as CSM.

PPS: Please forgive the rant, as I'm not in a great mood after being
thoroughly soaked by the rain thanks to a very late-arriving
transportation. Posting this on a bus with a very intermittent internet
connection (no, not going tower to tower--they promote their wifi
service on each bus).

On 5/8/2013 9:12 AM, Amy West wrote:
> I was hoping to see some discussion of this here.
> Where can I find Eugene Volokh's comments? I'd like to read his response.
>
> ---Amy West

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org