Morpheme replacement

[Marisa Lohr]

----------------------------Original message----------------------------
On Mon, 13 Oct 1997, Mikael Parkvall wrote:
 
> ----------------------------Original message----------------------------
> If I remember correctly, the items on the Swadesh list were said by Swadesh
> himself to be replaced at a rate of about 15 per milennium. Does anybody on
> the list have any idea regarding the differences between various types of
> items; in other words, would lexical morphemes be replaced at a higher rate
> than grammatical ones (not just those on the Swadesh list, but also bound
> morphemes), or vice versa? Or is there no difference at all between them?
>
> Mikael Parkvall
> parkvall at ling.su.se
 
Dyen (1964) demonstrated that different parts of list have different
replacement rates, and found statistically significant correlations
in the ordering of rates for particular parts of the list in a comparison
of Malayo-Polynesian, Indo-European, and, with Kruskal and Black (1973)
Philippine and Cushitic also.  Joos (1964) and Van der Merwe (1966)
suggested dividing the list into subsections with different retention
rates; while Dyen assumed a distinct rate for each word, sought to
calculate this, and put forward a method for combining this information to
give an overall rate (Dyen, James and Cole 1967; Kruskal, Dyen and Black
1973).
 
I have done some work on the retentiveness of basic vocabulary, examining
over 450 words, with a significant overlap with Swadesh's lists.  Although
I did not examine bound morphemes, I did not find a absolute distinction
between "more grammatical" and "more lexical" items in terms of
retentiveness.  For example, the most retentive set of words I found
included the 1sg. pers. pronoun, low numerals excluding "one", the 3sg.
pres. of the verb "to be", the verb "to lick" and the nouns "nail" (on
finger), "wolf" and "name".  However, for the list as a whole, the
proportion of "grammatical" items does grow smaller in comparison with
purely lexical items as the retentiveness of the words decreases.  I
should say that the work was based entirely on Indo-European data;
however, inasmuch as comparison with Kruskal, Dyen and Black's (1973)
figures for Philippine was possible, the ordering of replacement rates for
different sets of words showed a very strong correlation for the two
language families.
 
Refs.
 
Dyen, Isidore (1964) 'On the Validity of Comparative Lexicostatistics' in
"Proceedings of the Ninth International Congress of Linguists" ed. Horace
G Lunt (Mouton, The Hague), 238-52
 
Dyen, Isidore, Alan James, JWL Cole (1967) 'Language Divergence and
Estimated Word Retention Rate'
 
Kruskal, Joseph B, Isidore Dyen, Paul Black (1973) 'Some Results from the
Vocabulary Method of Reconstructing Language Trees' in "Lexicostatistics
in Genetic Linguistics.  Proceedings of the Yale Conference" ed. Dyen
(Mouton, The Hague), 30-55
 
Joos, Martin (1964) 'Glottochronology with Retention-Rate Inhomogeneity'
(abstract) in "Proceedings of the Ninth International Congress of
Linguists" ed. Lunt, 237
 
Van der Merwe (1966) 'New Mathematics for Glottochronology': Current
Anthropology 7, 485-500
 
 
Marisa Lohr