Darwin-L Message Log 4:7 (December 1993)

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.

---

<4:7>From sally@pogo.isp.pitt.edu  Fri Dec  3 17:23:13 1993

To: darwin-l@ukanaix.cc.ukans.edu
Subject: Re: linguistic drifts or "imbalances"
Date: Fri, 03 Dec 93 18:26:41 -0500
From: Sally Thomason <sally@pogo.isp.pitt.edu>

 The term "drift" is so vague that it isn't surprising if
nonlinguists find it unclear; it isn't what Gerard Donnelly
Smith has in mind, though.  I haven't been keeping up with
postings lately, so maybe other linguists have discussed
drift already.  I'm sorry if I'm repeating things.

 Imbalances (a loose term that is mainly an abbreviation for
a lot of different structural conditions) in a linguistic system
are, ultimately, things that are hard to learn because they're
irregular, or hard to hear, or whatever. A hole in a pattern
can also be an imbalance: an example that has , I think, been
mentioned here before is the new second person plural pronoun
that appears in various dialects -- not always the same form,
but always the same function, to disambiguate 2nd person
reference: yunz (here in Pittsburgh), y'all (for some Southern
speakers), etc.

 But many structural imbalances are not gaps in patterns.
Irregular forms, especially those that aren't very commonly used,
are often vulnerable to regularization.  So, for instance, most
young Americans have only _dreamed_ as the past tense of _dream_;
the new past is still competing with the old irregular past tense
_dreamt_ for many English speakers.  Another recent example of
regularization is the plural of _cow_: now it's _cows_, but it
used to be something like "ki" (I forget just what the vowel
would have been in Modern English) -- or, rather, it would have
been "ki" (or so) if _kine_ hadn't been formed instead, with the
same suffix as in _oxen_.

 Still another type of "imbalance" is often considered a
trigger for sound change -- things that are hard to hear, such
as final syllables in long words which are stressed on the initial
syllable.  (Well, maybe I'm using the term "imbalance" more
broadly than some other linguists might.  My point is that the
notion covers a heterogeneous bunch of phenomena. There are others
besides the ones I've mentioned.)

 The idea behind drift, in historical linguistics, is this:
if a language A is in the process of splitting, or has recently
split, into two daughter languages B and C, B and C have of course
inherited all the structures of A, including the hard-to-learn
things -- patterns with gaps, hard-to-hear sounds, irregular
forms, etc.  Because they have inherited the same structures,
and because changes that are motivated in part by difficulty
of learning are often quite similar, we are likely to find a
sizable number of identical, or very similar, changes in B and C.

 This fact has methodological implications.  Our Comparative
Method, by means of which we can reconstruct (parts of) an
unattested, i.e. literally prehistoric, parent language of a family
of related languages, assumes -- among other things -- that
features shared by all the daughter languages can safely be
reconstructed for the parent language.  Obviously, the results of
changes due to drift, that is, changes that occur AFTER B and C (the
only daughter languages of A) have diverged independently from A,
are not features that were present in A; so such features have
the potential of misleading us into reconstructing things for
A that A didn't have.  Sometimes we can find evidence that
will keep us from making this mistake: for instance, the so-called
"yers" of late Proto-Slavic were extra-short vowels that underwent
very similar changes in the various Slavic languages.  But the
results of those changes, while similar, differ in detail from
language to language; so we know that they post-dated the
splitting of Proto-Slavic into the various Slavic languages.
Sometimes, though, we don't have such evidence from differences
in the details of changes, and so we are bound to reconstruct
some things for a proto-langauge (parent language of a family)
that it didn't have.  No way to check.  And, of course, we
are also unable to reconstruct some things for a proto-language
that it DID have -- things that have been lost in all the
daughter languages, maybe due to drift.  Our methods have
limitations.  But isn't that true in all the historical sciences?

 Sally Thomason
 sally@pogo.isp.pitt.edu

---

Your Amazon purchases help support this website. Thank you!