Linguistic complexity

From: Piotr Gasiorowski
Message: 4584
Date: 2000-11-05

With all due respect to Stephen Wurm, there is only one good reason why a grouping of languages is called a "phylum": the evidence of common descent is not sufficient to reconstruct the protosystem, the intermediate stages corresponding to the nodes of the putative "family tree", and the regular changes needed to derive the attested modern systems. Nobody calls Austronesian or Semitic a phylum; the term is reserved for loose and controversial groupings like Penutian, Khoisan or Niger-Congo (purists might want to add Altaic and Afroasiatic).
 
NG holds more than 60 relatively well-established "Papuan" (= non-Austronesian) families, though even this division is tentative: much more work is required on documenting local languages (some are still known from word-lists alone) and disentangling genetic and areal similarities. Wurm and his ANU team did a vast amount of really careful fieldwork and came up with a classificatory scheme that distinguishes as many as six major phyla and leaves a residue of thirty-odd isolates and languages grouped into itsy-bitsy families (2-8 members each). A Herculean labour indeed, but scarcely one that solves all problems.
 
Trans-NG (more than 500 members), Sepik-Ramu (ca. 100 members), Torricelli (ca. 50 members), etc., are typical portmanteau groupings, based on a protocol of typological and lexical agreements, not on a rigorous reconstruction. Archaeological evidence doesn't mend matters, as an early expansion of agriculture need not have involved speakers of just one language ("Proto-Trans-NG"). The phyla in question group languages that are no doubt "related" in some way, but the nature of the relationship is far from clear, as in the case of IE and Uralic.
 
Piotr
 
 
----- Original Message -----
From: John Croft
To: cybalist@egroups.com
Sent: Sunday, November 05, 2000 7:32 AM
Subject: [tied] Re: IE & linguistic complexity

John wrote
[on the impossibility of reconstructing large families in NG]:

This is not really true.  What is required however, is a very
intimate detailled knowledge of a large number of local languages.  I
was able to follow "trees" linking Enga, Huli, Angal Heneng, Kewa,
Samberigi and Wiru languages in the Southern and Western Highlands
(Central West Highlands Family of the Trans Papuan Phylum of Stephen
Wurm), and Bosavi, Biami, Foi'i and Fasu languages of the Western
Family of the same Phylum, in 1980-83.

Your explanation of "micro-explosions" is very real, however, and
seems to have been linked with the introduction of waves of new
cultivars and the effective exploitation of these.  The first one
mentioned above seems to have been due to the origins of new cold
resistent crops in the Enga Region of the Papua New Guinean highlands.

The extent of the Trans Papua Phylum (and to related interior
languages in New Britain, New Ireland, and elsewhere) seems due to a
very early expansion of agriculture, extending outwards from New
Guinea into the Solomon Islands to the East and into Timor and
Halmahera to the west.  This seems parallel (and even earlier) to the
Middle Eastern Explosion associated with post glacial grain
production.