Date: Tue, 23 Jul 1996 09:38:58 -0700


Subject: Re: WAFT and the Principle of Linguistic Entropy

On Tue, 23 Jul 1996, Ron Butters wrote:

My AMERICAN HERITAGE DICTIONARY tells me that one can pronounce WAFT to rime

either with FATHER (first choice) or PAT (second choice). I appreciate all

the psycholinguistic speculation going on here, and I would not care to imply

that there may not be some geosocial distribution behind the variablity of

the A in WAFT. But I also would like to suggest that real people (as opposed

to linguists) do indeed look up words in dictionaries and are influenced by

what they read there--when they can remember what they read. (Also,

dictionary makers look up pronunciations in people and write down what they

find.) You say TOE-MAY-TOE and I say TOE-MAH-TOE, you say POE-TAY-TOE and I


In other words, when a lexical item is as rare as WAFT, most people haven't

made up their minds and really don't care very much one way of the

other--especially when their dictionary tells them that they can go either

way. ( I can't even decide whether I want to use SWUM as the past participle

of SWIM, and I swim about five miles a week.) This is what Hjelmslev called

the "Principle of Linguistic Entropy."

Yes, but doesn't this just beg the question (in the original meaning of

that phrase)? The lexicographers who put the two pronunciations in the

dictionary got them from SOMEWHERE. And where they got them, ultimately,

as you point out, was from informants. Even if some of these informants

were influenced by a dictionary entry, at some point we still get back to

such things as, yes, perhaps psycholinguistic and geosocial factors, but

certainly to plain old phonology and regional variation.

Peter McGraw

Linfield College

McMinnville, OR