November 16, 2006

Ear accidents

I'm in Manchester for a couple of days, serving as one of the external moderators for an "externally moderated reflective self-examination" at the National Center for Text Mining (NaCTeM). Although the formal process doesn't start until tomorrow morning, I spent this afternoon learning about some of the technical work in and around the center. One stop on the tour was Bill Black's office, where he and Jock McNaught told me about the CAFETIERE information-extraction system. Jock was describing his experience in supervising undergraduate students to adapt this system to new topic areas, and one of the examples involved creating a database of information drawn from news stories about ear accidents.

As I tried to assimilate his description into my gradually-developing image of how the system works and what it would like to adapt it to a new domain, it came to me that I had stumbled on an interesting cultural difference between the U.S. and the U.K. I don't believe that I've ever seen a news story in a U.S. publication about an ear accident. No doubt such accidents do occur -- piercings gone wrong, and the like -- but in my experience, they don't make the news. Could this be an aspect of pub culture that has so far escaped my notice; or an unanticipated side effect of playing cricket? I looked puzzled, I guess, because Jock repeated the phrase, "news reports about ear accidents". "Ear accidents", I echoed meditatively, tugging on my earlobe.

Well, of course, Jock was talking about "air accidents".

[Update -- Simon Cauchi writes:

I wonder if Jock McNaught is by any chance a New Zealander.

See Margaret Betterham's article, "The apparent merger of the front centring diphthongs — EAR and AIR — in New Zealand English", in New Zealand English, edited by Allan Bell & Konraad Kuiper (Victoria University Press, 2000), pp. 111-145.

No, Jock is a Scot.

And I don't think he merges ear and air (though I didn't check). Rather, I believe that he pronounces air as [er], with a vowel perhaps on the high side of [e] and a trilled [r], whereas I'm used to hearing something closer to [ɛɹ] or perhap [ɛeɹ] , with a much lower nucleus.


Posted by Mark Liberman at November 16, 2006 12:37 PM