Darwin-L Message Log 8:8 (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.

---

<8:8>From DARWIN@iris.uncg.edu  Tue Apr  5 13:53:26 1994

Date: Tue, 05 Apr 1994 15:50:18 -0500 (EST)
From: DARWIN@iris.uncg.edu
Subject: "Cladistics" and "typology"
To: darwin-l@ukanaix.cc.ukans.edu
Organization: University of NC at Greensboro

This comes from Sally Thomason, who was having temporary mail problems.

Bob O'Hara (darwin@iris.uncg.edu)

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

   Bob O'Hara suspects that the eyes of us non-systematists will be
glazing over at the discussion of what is & isn't cladistics, and
why: not so, at least for this non-systematist -- I'm finding this
discussion really educational, and if I keep reading long enough I
might actually understand what's going on in that area.  Understanding
the biological systematists is crucial for figuring out the
relationships between the tree-building enterprises in biology and
linguistics (and elsewhere), after all.  Bob's revised definition
of cladistic analysis -- the method based on the idea that only
innovations count in identifying clades -- is THE obvious link between
systematics and historical linguistics: shared innovations has always
("always" = since the late decades of the 19th century, when the
comparative method got into high gear) been the primary criterion for
subgrouping languages in a family tree.  (And since the comparative
method, which most of us believe is the only tried & true method for
establishing language families, hasn't been shown to be valid for
time depths anywhere near the presumed origin of human language, our
numerous language families are each, for all practical purposes, the
equivalent of the systematists' single tree.)

   A brief comment on Bob's mention of the different connotations of
typology in linguistics and systematics (and apologies if this repeats
comments made in earlier discussions of typology): typology is absolutely
respectable in linguistics, but NOT as a historical methodology.  It is
in a synchronic field of study that seeks to identify common, not so
common, and nonexistent structural patterns in languages of the world,
with a view toward understanding the functionally and/or genetically
determined (or at least influenced) properties of human language.  It's
true that typological arguments have been brought to bear, by a number
of scholars, on the question of what to reconstruct, for a parent language,
from the patterns found in a bunch of daughter languages.  One line of
argumentation (maybe the main line) goes like this: (almost) no language in
the world has pattern X; therefore, we cannot reconstruct pattern X
for the ancestor of these languages.  Arguments of this strong type
are highly controversial and have not been generally accepted.  The best-
known instance, probably, is the Glottalic Theory, a typological proposal
for the reconstruction of Proto-Indo-European stops.  But in a weaker
form, typology IS used by all of us in reconstructing bits of proto-
languages: when, for instance, you have two sounds that correspond
regularly in the only two daughter languages of a family -- say, h and
p -- and you have to make a guess about which (if either) of these two
sounds represents the parent language's sound...or, more precisely,
phoneme...you guess the p, essentially on typological grounds: we all
know of quite a few sound changes of the type *p > h, but we know of
few or no changes of the type *h > p; since, therefore, *p > h seems
a more likely change, we reconstruct *p for the proto-language's
phoneme.  [Some historical linguist who is more knowledgeable than I
am will promptly post fifteen examples of sound changes from *h to p!
So I'll respond to that posting now: the methodological point is valid,
even if the particular example is no good.  Pretty weak answer, eh?
But I think most historical linguists would agree with it.]

  Notice, though, that using typology in this way in reconstructing
proto-languages has nothing to do with establishing language families.
Probably no current historical linguist would argue that shared
structural features in themselves provide evidence for language-family
grouping.

   Sally Thomason
   sally@pogo.isp.pitt.edu

---

Your Amazon purchases help support this website. Thank you!