LINGUIST List 3.845

Thu 29 Oct 1992

Sum: ASCII IPA

Editor for this issue: <>

Directory

1. Richard Sproat, Summary Posting on ASCII IPA

A few weeks ago I posted a request for info on uniform two-byte ascii
encodings of the IPA. The advantages of such a representation are:

A) One byte per ipa symbol -- even assuming an extended (8 bit)
character set --- would not not be sufficient to represent all the
ipa symbols that one would want in order to handle a lot of languages.

B) Uniform n-byte codes (as opposed to codes which have differing
numbers of bytes to represent different symbols) simplify the problem
of parsing phonetic representations.

C) (7 bit) ascii makes display of the characters generally trivial.

It does not appear that exactly what I was after exists as such, but
there are at least a few apposite schemes around. This posting
summarizes the responses to the query:

1) There is the 16-bit ISO 10646 standard as used in Unicode. One
problem with this for the moment is the fact that these are 8-bit
bytes.

2) Also, there has been a discussion on the sci.lang newsgroup where
Evan Kirshenbaum <kirshenbaumhpl.hp.com> is working on a complete
representation of the IPA in ascii. This convention is basically
intended for network discussions, and it only maps a subset of the IPA
characters and diacritics to the ascii set, using a featural
representation for the rest. The representations are not necessarily
compact, therefore.

3) Finally, there is the following reference:

	J. C. Wells, _Computer Coded Phonetic Transcription_,
	Journal of the IPA, 17(2): 94-114, 1987.

For input, thanks to:

Chet A. Creider (creidertaptet.sscl.uwo.ca)
E. Dean Detrich (22743MGRmsu.edu)
Lee Hartman (GA5123SIUCVMB.SIU.EDU)
Lars Mathiesen (thorinndiku.dk)
Jon Whalen (jonhook.corp.mot.com)

(I hope I haven't forgotten anyone).


 Richard Sproat
 Linguistics Research Department
 AT&T Bell Laboratories	 | tel (908) 582-5296
 600 Mountain Avenue, Room 2d-451 | fax (908) 582-7308
 Murray Hill, NJ 07974, USA	 | rwsresearch.att.com