The Indo-European Hypothesis [was Re: The Neolithic Hypothesis]

[Ray Hendon]

RAY HENDON:
  The exposure side (rate of contact between infecteds and
  susceptibles) of the equation is easily identified in the linguistic
  community:  The relatively few Romans sent to govern England was not
  sufficient to generate a critical mass of exposure units for Latin
  to predominate.  And the Romans left local courts and laws stand,
  lessening the need for everyone to know Latin in order to get along.

Number might have something to do with it but what about languages or
dialects that become viewed as prestigious? I'm sure it's more
complicated than the above otherwise we would have had a secure model
100 years ago :)

Glen Gordon

--------------------------------------------

I think that prestige can be readily accommodated by the model.  A language
that is viewed as prestigious in a certain population would have a higher
average susceptibility to learning the language.  Thus the susceptibility
rate for a prestige-seeking population would be higher than for a population
of non-prestige-seekers.  Also, the exposure rate between the non-speakers
and speakers would be increased, as those wishing to learn the prestigious
language would actively seek to out those "infected" with the language so
they could learn it more quickly.

An example I think of first of this type of snob-appeal in a language was
exhibited by the elite Filipinos after Spain conquered them.  In the late
1980s I talked with a young woman from the Philippines who said that her
family spoke Spanish and was very proud of it.  It was definitely a prestige
factor in her family.  Certainly a population of prestige-seeking people
will exhibit a higher penetration rate of the prestigious language, and a
higher level of saturation than other, non-prestige-seeking populations.

Your second concern, that if mathematical modeling were applicable to the
field of linguistics it would already have been done, is one of which I am
quite skeptical.  Mathematical modeling was not developed until World War
II, and it was not being applied in civilian areas of research or
universities until after the war (with the exception of physics, where
modeling has been used for centuries--Newton, Einstein, e.g.).  The
development of multiple regression analysis, for instance, which was a great
boon to the ability to actually test the mathematical model developed, did
not occur until around 1955. The computers it took to make the incredibly
complex calculations for multiple regression were not generally available
until well after the 1950s.  The first test of the medical model of the
spread of infection was applied to heroin addiction in the late 1960's.
This is still a relatively new field.

Also, I do not know if the training for an academic linguist involves
advanced mathematics and statistics.  Are these subjects taken by Ph.D.
candidates?  If linguists are well trained in mathematics as another
language, then I would be surprised that no quantitative work would have
been done.  But if they are not as a group well versed in these disciplines,
I would be equally surprised if there had been much done.  I can see a
certain level of knowledge about statistics among some of the members of
this interest group.  All the scientific dating techniques of artifacts, for
example, must involve the use of statistical probability, so you must be
generally aware of the use of that tool.  But, so far, at least, I haven't
run into any work.  If I don't see some soon, I may take a look at the
journal literature in your field and see what I can find.

I think this approach,while perhaps limited in its applicability to
linguistics, may be of some help in forecasting the effects of invasions,
migrations and general population movements.  It may be an area that
deserves looking into.

Regards,
Ray Hendon
rayhendon at worldnet.att.net

[ Moderator's comment:
  I think that in the typical program in linguistics, whether undergraduate or
  graduate, the only real exposure to statistics is in experimental phonetics,
  where simple statistics (Student's T, chi-square, analysis of variance, etc.)
  are needed to assure the validity of results.  Few other subdisciplines use
  stats, since most of the work is done by individual introspection.
  --rma ]