Linguasphere web site

Herb Stahlke HSTAHLKE at gw.bsu.edu
Tue Sep 21 21:05:40 UTC 1999


[ moderator re-formatted ]

I just ran across the web site www.linguasphere.org, prepared largely by David
Dalby at SOAS.  Dalby has been involved in a sort of areal-typological
classification scheme for over thirty years, including considerable work in
Africa.  His classificatory scheme claims not to be diachronic but synchronic,
looking at shared vocabulary and shared features of other sorts.  His lexical
similarity measure is a limited use of the Swadesh 100 and 200 word lists.
What he's hoping to come up with, as I read his material, is a theory-neutral
classification of the world's languages, which he numbers at around 12,000+,
into groups defined on the ground of geographical proximity and various sorts
of linguistic similarities.  He struggles a bit with terminology for different
degrees of relationship, from closely similar dialects to distinct languages.

His comparative theory looks rather like that of Malcolm Guthrie, in his
Comparative Bantu, who also claimed not to be doing genetic classification but
some sort of synchronic classification.  Guthrie's work was vey careful, and
his Bantu zone system is still found widely in comparative Bantu terminology.
However, his historical inferences have generally been rejected.  One of these
inferences, for example, was that the homeland of a language group is its
synchronic area of greatest similarity, which placed the Bantu homeland
somewhere south of the Great Lakes rather than in the Nigeria-Cameroun border
areas.

His overall classification is decimal, dividing the world into 5 geosectors and
5 phylosectors.  It looks as if each geosector is a collection of 10 sets, as
he calls them, which may be either geozones, unrelated groups inhabiting a
region, or phylozones, related groups inhabiting a region.  For example, his
geosector North America comprises Arctic (1), Athabaskic (2), Algonkic (3),
Pacific Coast geozone (24), Transamerica geozone (3), Gulf geozone (8),
Aztecotanic (2), Otomangic (1), Mayanic (1), and Meso-America geozone (9).  The
parenthesized numbers indicate the number of sets (next smaller unit of
classification representing Swadesh rates of 25-30+ shared vocabulary items.
He recognizes but does not address the problem of determining what items might
be shared vocabulary.  He does not use the comparative method at all, although
his groupings frequently represent the results of the CM.

Dalby lays out some high-minded humanitarian goals and motivations that he
hopes his global classification will further.  This is a curious scheme that a
lot of work and thought has gone into, but I'm not sure quite what comes out of
it, beyond some sort of theory-neutral global referential system for language
groups, languages, and dialects.

Herb Stahlke



More information about the Indo-european mailing list