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