The World Loanword Database project (WOLD), edited by Martin Haspelmath and Uri Tadmor, is now online and freely available to users. It’s a remarkable resource compiled with the purpose of analyzing language contact at the lexical level. Over fifty linguists (including my colleague Martha Ratliff here at Wayne State) have provided mini-vocabularies of languages (41 in total) including information for thousands of words on borrowing, attested age, and analyzability, creating cross-linguistic indices that measure the degree to which particular words (and types of words) tend to be borrowed, and from which languages they are borrowed. So, for instance, you can find that more languages have a borrowed word for father than for mother, or that Hawai’ian borrowed several terms for continental North American animals (prairie dog, skunk, kingfisher) from Ute (an indigenous language of Colorado). It’s a rich and highly functional database, and my only ‘complaint’ is that I’d like to see hundreds more languages covered! I need to stop playing with it right now or my day is going to be shot.
Of course, being who I am and doing what I do, the first place I turned was to the numerals, and I immediately noticed two significant things:
– Ordinal numerals seem to be less frequently borrowed than cardinals; first is borrowed less often than one; second less often than two; and third less often than three.
– Fifteen is borrowed less often than five or ten. Fifteen is far more analyzable than either ten or five (most often as ’10+5′) – so how does this make sense?) I’ll have to look at the data more closely to figure this one out … but not today, work calls!