[Lingtyp] wordhood

[David Gil]

In response to my proposed definition of wordhood, Martin writes:
> If different criteria are used for different languages, how do we know 
> that we are measuring the same phenomenon across languages? 
But there is nothing at all incoherent or logically faulty about 
defining universally valid comparative concepts based on incommensurate 
language-specific categories.  Take the Dahl/Velupillai WALS "Past 
Tense" map. It distinguishes four feature values: past tense (i) absent; 
(ii) present, no remoteness distinctions; (iii) present, 2-3 remoteness 
distinctions; (iv) present, 4 or more remoteness distinctions. It is 
based on language-specific descriptions of 222 languages which may be 
and probably are at least partly incommensurate with respect to the 
nature of their past-tense forms.  Nevertheless, it makes perfect sense 
to compile these incommensurate descriptions and abstract away from 
them, in this case by simple counting, which yields an elementary formal 
property functioning as a comparative category: the number of remoteness 
distinctions in the language.  My proposed definition of word does the 
same.  And there's nothing very earthshakingly original about it; it's 
what we as typologists do all the time.  What I see my proposal as doing 
is merely taking an intuition about wordhood that I believe most of us 
share, and — motivated by Martin's original critique — trying to be a 
bit more careful than we usually are in making clear what aspects of it 
are language-specific and incommensurate, and what other aspects might 
indeed form the basis of a cross-linguistically valid comparative concept.

-- 
David Gil

Department of Linguistic and Cultural Evolution
Max Planck Institute for the Science of Human History
Kahlaische Strasse 10, 07745 Jena, Germany

Email: gil at shh.mpg.de
Office Phone (Germany): +49-3641686834
Mobile Phone (Indonesia): +62-81281162816

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20171112/0542d929/attachment.htm>