Darwin-L Message Log 8:14 (April 1994)

Academic Discussion on the History and Theory of the Historical Sciences

This is one message from the Archives of Darwin-L (1993–1997), a professional discussion group on the history and theory of the historical sciences.

Note: Additional publications on evolution and the historical sciences by the Darwin-L list owner are available on SSRN.

---

<8:14>From sally@isp.pitt.edu  Wed Apr  6 18:40:40 1994

To: darwin-l@ukanaix.cc.ukans.edu
Subject: Re: "Cladistics" and "typology"
Date: Wed, 06 Apr 94 20:40:37 -0400
From: Sally Thomason <sally@isp.pitt.edu>

   Peter Cannell asks what a language family is -- good question!
I've been avoiding using our standard term, which is "genetic
relationship", for fear of confusing and/or annoying people who
deal with literal rather than metaphorical genes.  Anyway, the
concept of a language family is pretty clear, although -- not
surprisingly -- it's fuzzy around the edges: if one language
diverges over time into two or more daughter languages, that's
a language family.  The daughter languages are changed later
forms of their single parent language.  There are some methodological
assumptions which, though clearly invalid most of the time in
real life, are useful nevertheless; the major one is that, when
a lg. splits into two or more daughters, the split is CLEAN: that
is, no more contact between the daughters after the split.  You
don't get a split at all unless there's some breakdown/reduction
in communication between speakers of two or more dialects; but
cases like Romani, whose speakers had no more contact at all with
other Indic languages after they left India in ?1500 A.D.?, are
pretty rare, as far as we know.

   At shallow enough time depths, when there's still a lot of
evidence left (not too many changes in the daughter languages to
obscure the regularities of descendent lexical and grammatical
features), it's fairly easy to separate inherited material from
borrowed material (i.e. convergences), IFF the borrowing isn't
occurring between very closely related languages.

   But that's at shallow enough time depths.  Once you're several
thousand years away from the source, things get more difficult;
that's why historical linguists make an informal guess of ca. 10,000
years as the upper limit for establishing genetic relationships
among languages.  (Most historical linguists, anyway.)  After that
too many changes have left so few systematic traces that it can
be impossible to distinguish borrowed from inherited features, even
when you're pretty sure that there is SOME historical connection
between two or more languages.  So historical linguists don't expect
ever to construct a family tree for ALL human languages, assuming
that there was once one single original language; we have some
hundreds of language families currently, and though that number
will undoubtedly shrink as the laborious historical study of different
groups progresses, it's unlikely to shrink to just a few huge language
families -- not at the time depths that, to judge by the level of
diversity in the world's several thousand languages, must be
reckoned with.

   Language is significantly different from other cultural or
behavioral studies, though: language change is largely, though not
entirely, independent of speakers' or societies' desires and
intentions; language changes willy-nilly, which is why (for instance)
pundits who rail against the sloppiness of, say, modern English
speakers have been thriving for hundreds of years.  There are fads
in language, to be sure (like teen-age slang), but they are quite
superficial: the structure of the language tends not to be affected
by fads.  Other cultural and behavioral studies have nothing like the
regularity hypothesis of sound change, which is our main means of tracing
language history back through time.

     Sally Thomason
     sally@isp.pitt.edu

---

Your Amazon purchases help support this website. Thank you!