Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site ucla-cs.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!bellcore!decvax!ittatc!dcdwest!sdcsvax!sdcrdcf!ucla-cs!das
From: das@ucla-cs.UUCP
Newsgroups: net.nlang
Subject: Phoneme phrequencies
Message-ID: <8886@ucla-cs.ARPA>
Date: Tue, 11-Feb-86 14:18:06 EST
Article-I.D.: ucla-cs.8886
Posted: Tue Feb 11 14:18:06 1986
Date-Received: Fri, 14-Feb-86 02:22:34 EST
Organization: UCLA Computer Science Department
Lines: 17

[Background:  I was listening to the radio, and the DJ announced the next song,
 adding that the group that performed it was from Canada.  "Canajun, eh?",
 I thought, and listened carefully for phonetic evidence, focussing primarily
 on what I, a native Angeleno, pronounce as /au/.  There wasn't a single one
 in the entire song!  No "out", "about", "house", nothing!]

Anyone who has played around with cryptanalysis knows where to find single-
letter, digraph, trigraph, etc. frequencies for English and some other
Roman-alphabet European languages.

Is there a source listing phoneme frequencies?  From observation, it's clear
that unvoiced "th" is just about the least frequent consonant in English, and
that in my dialect, at least, the vowel in "caught" (which I distinguish from
"cot"), "saw", and "ought" is probably the least frequent vowel.

Is there any compilation of single-phoneme, diphthong, "diphone", "triphone",
etc. frequencies for spoken English?  Any dialect will do.