Return
to AKSES links.
English
(Spelling-to-AKSES) Transliteration Tables
The example tables below are
presented as a conceptual database
for a program to convert traditionally-spelled text into AKSES phonemic
text. They do not include all words
required of a useful dictionary and, because they represent my personal
perception of a phonemic form for American English, they are not as
authoritative as if prepared by American lexical organizations that have
empirical word-element data bases collected from speakers and writers from all
parts of the nation.
Each table contains words of
the same initial letter in alphabetical order of the spelled words. The 2-column format is described below:
The first column contains
traditionally-spelled words; the Latinate letter is the first entry of each
table. Subsequent first-column entries
are alphabetically listed spelled words.
A starting list of about 10,000 uninflected words was augmented to over
46,000 by adding inflected forms and words found in a number of traditionally-spelled documents.
The second column contains AKSES written words corresponding
to the first column spelled words.
Spelled homographs are repeated on successive lines in order to display
all AKSES words associated with each traditional homograph. Examples are "read" (/rēd/
present tense, /red/ past tense) and "live" (/līv/ adjective,
/liv/ verb).
Click on these links to see example spelled-to-AKSES
writing tables. Use your browser’s
return botton:
Process of Transliteration - Converting
Spelled Text to AKSES Text
Most spelled words convert to 1 and
only 1 AKSES word. (Exceptions are
homographic words and proper names.)
Ignoring these exceptions for the moment, we see that the job of converting
a conventionally-spelled text document into an AKSES-written document is
exceedingly simple and involves kinds of tasks that computers do rapidly and
efficiently. Likewise, punctuation and
formatting are not a problem because, except for converting word numbers to Arabic
numbers, neither changes. The basic
process described has 3 parts: 1) Read
the spelled text as punctuation, numerical values, and text words; 2) Transfer
punctuation without change, convert numerical values from words to properly
punctuated Arabic numerical format where necessary, and substitute the AKSES
counterpart for each spelled word; and 3) Print the reassembled AKSES-form text
and print the final AKSES file using the original document’s formatting.
The conversion program must correctly
identify 2 exceptional kinds of words.
Firstly, the person doing the transliteration must select 1 of the AKSES
words provided that corresponds to the intended sense of each spelled
homograph. Secondly, he or she must mark
“words” that must not be changed, such
as foreign words, proper names which retain their spellings, and non-words like
letters used to organize lists and outlines or acronyms.
An authoritative AKSES dictionary must
be prepared and published by a recognized
lexical organization to be accepted by the public and academics alike. The dictionary publisher then combines the
AKSES wordlist with the traditionally-spelled wordlist from which the AKSES
words were derived to form the database for a transliteration program anyone
can use to convert correctly-spelled text files into correctly-written AKSES
text files.
Return
to AKSES links.
First
published 12/04/04 by J H Kanzelmeyer; last worked on 03/13/08 JHK