Tuesday February 24, 2004
Over the last few days, I've been poking around in the Ethnologue database for a typology project, in particular their classification of the world's languages. It's got all the obvious categories. Indo-European? Check. Sino-Tibetan? Check. Austronesian? Check. But there's all sorts of interesting oddities in there, too.
I went to the Language Isolates phylum looking for Ling 101 favorites Basque and Japanese, only to find that both of those are top-level phyla. In the Ethnologue classification, Basque has 3 sub-languages and Japanese has 12. It'd be interesting to know how consistent the division of languages and dialects is—are the Japanese varieties really more diverse than, say, the Korean dialects, which are listed as a single isolate language?
The top-level Unclassified phylum—which isn't a phylum, exactly, but you know what I mean—contains a lot of languages I've never heard of, mostly in places I'd consider off the beaten path. (Eurocentric? Guilty.) But: Polari in the UK, "[a]n in-group language among theatrical and circus people. Second language only." Traveller Scottish in the UK, "[a] blend language of High Romani and Elizabethan Cant." Quinqui in Spain, "[a] blend language of certain urban ex-nomadic groups. It contains elements of Calo and Germania argot." (Is that pronounced [kiŋki]?) I was surprised that any European languages were still poorly studied enough that they remain unclassified at this late date. I suppose it's because all three are related to Romani, and one (Polari) served as British gay slang—and the European margins are just as marginal as anywhere else. Finally, the most interesting-sounding unclassified language: Haitian Vodoun Culture Language.
There's two groups of sign languages in Ethnologue. Deaf sign languages includes 114 sign languages from all over the world. (When I first saw this list, I was momentarily annoyed because all the entries have names of the form "Foo sign language"—why wasn't Ethnologue using the appropriate native names for the sign languages? Then my brain clicked on.) However, they've also got two languages in the separate Sign languages phylum: Monastic Sign Language, used in the Vatican ("[a] second language means of communicating while maintaining vows of silence") and Plains Indian Sign Language.
They've got an odd little phylum called Cant that has exactly one language in it: Pitcairn-Norfolk, a mixture of English and Tahitian, "[d]eveloped from mutineers settling on Pitcairn in 1790." That's the mutineers from HMS Bounty. Cool! (The entry seems a little confused, though. It's not clear whether they think Pitcairn-Norfolk is just a dialect of Pitcairn or not—if so, there ought to be a separate entry for Pitcairn, right?)
Phyla like Cant, Pidgin, and Creole raise an interesting question for typologies like Ethnologue's: is a tree really the right structure to use when describing the relationships between natural languages? The analogy with biological species is obvious, but it doesn't hold up. If two populations of living organisms are isolated long enough that they are no longer interfertile, they've speciated and that's the end of it. With languages, though, there's nothing that prevents two unrelated languages from being thrown together by the flow of history and, people being people, the almost inevitable result is a new blended language. I suspect you need at least a directed graph rather than a tree, and it's probably even more complicated than that.
(Isn't language interesting?)
TrackBack URL for this entry:
Listed below are links to weblogs that reference Interesting Oddities:
Of course, quinqui is pronounced [kiŋki].
Posted by: at Dec 10, 2005 3:36:04 PM