LINGUIST List 3.937

Sat 28 Nov 1992

Disc: Assessing Remote Linguistic Relationships

Editor for this issue: <>

Directory

1. "Bruce E. Nevin", assessing remote relationships
2. , 3.917 Greenberg, Classification and Historical

Regarding assessment of claims of remote genetic relationship of
languages with respect to chance resemblances, this weekend I
received some correspondence from Bob Oswalt (who is not on the
net). He sent me a photocopy of his 1991 paper on his computer
program for this purpose, which I believe is discussed in Don
Ringe's monograph (which I haven't seen yet). Bob's paper:

Oswalt, Robert L. 1991. A method for assessing distant
 linguistic relationships. In _Sprung from Some Common
 Source: investigations into the prehistory of languages_,
 ed. by Sydney M. Lamb and E. Douglas Mitchell, Stanford:
 Stanford University Press.

In his letter, Bob refers to "a considerable literature on the
estimation of chance resemblances -- and the nonestimation, just
assuming chance resemblances are small." There is reason to hope
that he may soon write a comprehensive survey article on the
subject, possibly in _Current Anthropology_ (where he has
published some comments on distant relationships proposed by
others in _CA_).

Among the references in the 1991 paper is:

Oswalt, R. L. 1970. The detection of remote linguistic
 relationships. _Computer studies in the humanities and
 verbal behavior_ 3: 117-129.

In his letter, bob says that this is a condensation of the one
submitted to _Language_ in 1965, which was an elaboration of a
talk given in 1962 to the Berkeley Linguistics Group. He
mentions two articles in the Proceedings of the Hokan-Penutian
Conference, but I don't have those citations at hand.

 Bruce Nevin
 bnbbn.com

 STEVEROYIDUI1.CSRV.UIDAHO.EDU
suggests that Greenberg and Ruhlen may be claiming that
"that "basic vocabulary" is somehow more
resistant to regular sound changes that over the very long time have
obscured the historical relationships than other words are."

However, that is not the case. Instead, they argue (or at least
Greenberg does) that instead of precise correspondences it is
enough if you establish broad resemblances (i.e., correspondences
between sets of sounds, is what it seems to amount to). In
principle, there is nothing wrong with THIS idea, although I
would argue that he has not carried it out correctly.