LINGUIST List 5.782

Thu 07 Jul 1994

FYI: ASCII version of IPA, v.1.12

Editor for this issue: <>


  1. David Prager Branner, An ASCII version of IPA, v.1.12

Message 1: An ASCII version of IPA, v.1.12

Date: Wed, 6 Jul 1994 21:00:53 -An ASCII version of IPA, v.1.12
From: David Prager Branner <>
Subject: An ASCII version of IPA, v.1.12

 I joined the net about six months ago and almost immediately
became frustrated by the limitations of the ASCII character set. I
devised a simplified system for representing IPA on an ASCII keyboard, in
the hopes of using this over email and on my tiny palmtop computer in the

 In June I saw a copy of Evan Kirshenbaum's system, which was
developed in 1992-1993 and has been in use on the sci.lang newsgroup. My
system is quite different from his - different enough to make it worth
posting, I think, and invite people's comments. My goals are rather
different from Kirshenbaum's and perhaps from those of many linguists on
the net. As a devoted fieldworker, I am only interested in having a
practical system, rather than an elegant one.

 I am now posting my own system and inviting comments. I have
borrowed a few details of Kirshenbaum's system, such as the vowels E and
O, and have heard suggestions from a number of other people, including
Miguel Carrasquer, Jonathan Evans, Karen Chung, and Mark Rosenfelder, to
whom I am grateful.



 To render the whole of the International Phonetic Alphabet in
ASCII characters, so as to aid the practical work of phonetic description
in the field when only simple keyboards are available, and to allow quick
IPA citations over email.


 1) As much as possible, reproduce the actual forms of IPA
symbols. Arbitrary assignments should as much as possible be avoided.
 2) When this is impossible, describe new characters using the
same kind of analysis used in IPA; i.e., do not introduce a new way of
analyzing a sound.
 3) Use only the most current version of IPA. Here I have used
the 1993 revision of Kiel, as published in the Journal of the IPA (1993)

 These principles are intended to keep the result *visually* as
close to current IPA as possible, within the very great limitations of the
ASCII system. If new ways of analyzing sound are introduced, conversion
to IPA will be made more difficult. So, for instance, central vowels
cannot simply be made up of non-central vowels plus the centralizing
diacritic. Retroflex consonants cannot just be plain consonants plus the
symbol for rhotacization. This system is designed to *render* IPA in
ASCII, not to replace it.

 A purely ASCII alphabet is extremely limited. Even the
typewriter, which was the tool of many American linguists during the
heyday of their school, can create new symbols by overstriking, but ASCII
is limited to the following 93 symbols, which cannot be overstruck:

 a b c d e f g h i j k l m n o p q r s t u v w x y z
 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
 1 2 3 4 5 6 7 8 9 0 - = ~ !  # $ % ^ & * () _ +
 [ ] { } \ | / < > ; : ' " , . ?

In order to reproduce IPA's much larger set of symbols and diacritics, I
have introduced a number of "operators", which alter a preceding symbol.
Operators always follow their operands. The following are the main
operators; note that some of them have purely typographic function, while
others have more phonological properties:

 1) - "Barred" characters, "crossed" characters.
 2) & "Turned" or otherwise altered character. Note that the
ampersand sign "&" is not used for the "ash digraph", though this has
become common usage among some groups using the net. In some cases where
the ampersand might be used, it has not been; for instance, the
unrounded u vowel, called "turned m" in IPA, could have been written
"m&", but instead I have written it "W".
 3) ^ Superscript. Found mainly among diacritics.
 4) ) General symbol for "special" phonological functions, such as
"r)", which distinguishes the retroflex consonants from the alveolar
series. IPA characters that are written with this symbol take up fully
three ASCII characters, which is perhaps too many; to shorten the
transcription, the ")" symbol itself can be left out as long as no
ambiguity results.
 5) " Graphically modified in an unspecified way.
 6) ! Click.
 7) $ Implosive. The 1993 version of IPA treats implosives as
inherently voiced, but it seems likely that voiceless implosives will be
assigned their own symbols some day soon. Voiced h could perhaps be
written "h$", on graphic grounds, but I have written it "*" instead.
 8) % Ejective.

 I have written glottal stop as "Q", following a Shanghainese
tradition going back to the beginning of this century (actually *final*
glottal stop in Shanghai, not to be confused with the Pinyin q- initial).
But glottal stop is often written with a question mark in common practice,
and I leave this as an optional alternate. But the question mark is such
a powerful symbol that I am afraid it could be misread as meaning
"uncertain" or "unknown" if it appeared in the middle of transcription.

 I have limited numerals to the description of tone, to prevent

 The following tables follow the arrangement of the chart published
in the Journal of the IPA, except that non-pulmonic consonants have been
arranged with pulmonic ones. An index is in preparation.



 Bilabial Labio- Dental Alveolar Post- Retroflex
Articulation dental alveolar

Stop p b t d tr) dr)
Nasal m M n nr)
Trill B r
Flap/Tap d" r"
Fricative F V f v T D s z S Z sr) zr)
Lateral fric s" z"
Approximant v" r&
Lateral approx l lr)

Click p! t! l! s!
Voiced Impl. b$ d$
Ejective p% t% s%

 Palatal Velar Uvular Pharyngeal Glottal

Stop c j- k g q G Q (or ?)
Nasal nj) ng) N
Flap/Tap R
Fricative c" j" x g" X R& h- Q& h * (or h$)
Approximant j W"
Lateral appr y& L

Click c!
Voiced impl. j$ g$ G$
Ejective k%

Other symbols:

w& Voiceless labial-velar fricative
w Voiced labial-velar approximant
h& Voiced labial-palatal approximant
H Voiceless epiglottal fricative
Q&- Voiced epiglottal fricative
Q- Epiglottal plosive

c" z" Alveolo-palatal fricatives
l" Alveolar lateral flap
*" Simultaneous S and x
xy( Affricates and double articulations can be represented by two
symbols joined by a tie bar "(" if necessary.


 When vowels appear in pairs, the one to the right represents a
rounded vowel.

 Front Central Back
Close y i i- u- W u
 I Y U
Close-mid e o" e& o- e" o
Open-mid E oe) E& E" v& O
 ae) a&
Open a Oe) A A&

DIACRITICS (shown here after a dummy "x")

xV) Voiceless x* Breathy voiced x[ Dental
xv) Voiced x~ Creaky voiced x] Apical
xh^ Aspirated xp) Linguolabialized x[] Laminal
xu) More rounded xw^ Labialized x~^ Nasalized
xU) Less rounded xj^ Palatalized xn^ Nasal release
x+ Advanced xg^ Velarized xl^ Lateral release
x_ Retracted xQ&^ Pharyngealized x.) No audible release
x" Centralized x~) Velarized or pharyngealized
xx^ Mid-centralized x= Raised
x,) Syllabic x=" Lowered
x') Non-syllabic x< Advanced tongue-root
xr^ Rhoticity x> Retracted tongue-root


' Primary stress
, Secondary stress
: Long
; Half-long
;) Extra-short
. Syllable break
| Minor (foot) group
|| Major (intonation) group
=) Linking (absence of a break)

Tones and word accents:

\) Downstep / Global rise
/) Upstep \ Global fall

 Phonetic tone, of either level or contour type, should be
described using the practice common in Chinese, which is that numerals are
used to represent the Chao tone symbols. The range of the voice is
divided into five units, to which the numbers 1 through 5 are assigned,
with 1 low and 5 high. Tones are then described by the numbers
representing the beginning and end of the contour, along with any dips or
peaks in between. So the well known contours of standard Mandarin are
 55 (first tone: high level)
 35 (second tone: mid rising to high)
 213 (third tone: mid-low, dipping and rising to mid)
 52 (fourth tone: falling from high to mid-low).

Bidirectional tones are not at all rare in China, and even tridirectional
tones are occasionally encountered. The Foochow tone /6/, which rises
from mid to mid-high and then falls to mid-low, would be written [342].
The tone contours of very short syllables may be described with a single

 Phonemic tone category may be written with numerals, or with
letters or a combination of numerals and letters. It is of course
essential to distinguish phonemic from phonetic description in this
case. Alternately, tone categories could be marked in superscript.


 Phonemic forms can still be written between slashes in this
system, and phonetic forms between square brackets, but care must be
taken to distinguish them consistently from symbols ending in / or ]. In
Chinese this is not a problem, because every syllable ends in a tone
contour of some sort. But it may cause difficulty in non-tonal
languages. There may also be difficulty with the close-parenthesis sign

 The symbols #, {, and } have been omitted from this system, at
least for now. I believe it may be necessary to introduce them in order
to clear up the problems of /, ], and). The boundary marker # is not
widely used in field transcription now, and may be hard to introduce.

 There are a great many symbols that have been discarded from the
IPA, or that are common in particular linguistic traditions but are
unknown internationally. No effort has been made to include these symbols
above. For my own use in China, however, I have devised symbols
representing the four special vowels universally used in Chinese IPA:

 i" unrounded frictionless z
 y" rounded frictionless z
 I" unrounded frictionless zr)
 Y" rounded frictionless zr)

Chinese linguists distinguish a whole alveolo-palatal series of
consonants, at least in theory; the symbols for them are modelled after
the IPA alveolo-palatal fricatives. In practice, however, Chinese
linguists often use these symbols to write the palatal consonants. I have
not devised special symbols for them.


 I would very much appreciate hearing any reactions to this system,
especially from practicing fieldworkers.

David Prager Branner, Yuen Ren Society
Asian L&L, DO-21, University of Washington
Seattle, WA 98195 <>

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue