Constructing an alternate dictionary


Like some other spelling systems, my system DRE has a dictionary, TARDRE.  I am not using the classical meaning of "dictionary" here.  TARDRE is not a collection of words with definitions, etymologies, etc.  In the context of alternative spelling systems, I use the word "dictionary" to mean a list of English words, paired with their respellings.  TARDRE was a very time-consuming project, but I believe it was time well-spent.  Though it took a long time to produce TARDRE, DRE has a characteristic that made producing a dictionary much easier than it would have been for any of my other systems:  its similarity to TS.  About 1 in 3 English words is unchanged when spelled with DRE, and most words are changed only in a few letters.  This made generating the dictionary a much less laborious process than it would otherwise have been.

Having a dictionary for one's system provides the following benefits:

  1. It enables the writing of a conversion program, to automatically convert a TS text into the target system.

  2. The labor of composing the dictionary gives one a very strong hands-on experience with one's own system.  It is likely to expose flaws in the system which otherwise would have remained dormant for quite some time. 

  3. It gives the system designer experience with the patterns and subtleties of the English vocabulary that is hard to get any other way.


Much of the labor in constructing TARDRE had little to do with DRE per se - the hard part was establishing standard pronunciations, given that I could not trust my own pronunciation to be acceptable.  It occurred to me that it might be possible to modify the DRE dictionary so that it became a useful tool for other system designers, especially by allowing them to create their own dictionaries without having to repeat the pronunciation research.  I undertook such a project and eventually developed the FEWL dictionary, written in a specially designed notation called FLEWSY.  Using FEWL and FLEWSY, plus some programming, it is possible to produce a high-quality dictionary for a reformed spelling in weeks or even days rather than months.

FLEWSY, the Fonemic Latin-1 English Writing SYstem, may be thought of as a spelling system itself - though it is too complex to actually be useful in that capacity.  It encodes phonemic, morphological and traditional spelling information in a way that can be easily transformed by a program into a wide variety of alternate spelling systems.  FEWL, the FLEWSY Encoded Word List, is a list of relatively common English words, paired with their FLEWSY encodings.  To produce a dictionary for a new spelling system, all that is necessary is to write a program to convert the FLEWSY notation into the target system.  For those capable of the programming (or of finding someone else to do it for them), this can generate a dictionary in fairly short order.  Attributes of FEWL that make this process attractive include the following:


A final advantage of using FEWL/FLEWSY as the basis of a dictionary is that it offers the same opportunities as a more manual method for getting acquainted with the English lexicon, and for discovering weaknesses in one's system.  The difference is that, when a weakness is discovered and one's system is modified to avoid it, it is generally a trivial matter to update the programming and generate a new dictionary.  Often, one can try out several variants of a notation in a day, rather than having to laboriously correct the dictionary by hand each time.

As FEWL was being finalized, I was in the process of designing my spelling system WMM (a.k.a. GASS).  WMM made an excellent test case for the use of FEWL in building a dictionary.  The process exposed some serious design issues with WMM which otherwise might have been unnoticed for months, and the greatest part of the work was done in a day.  I would not have had the patience to develop a dictionary for WMM manually - because WMM is so unlike TS, the process would have taken much longer than for DRE.

WMM was finalized in a couple of weeks; without the use of the FEWL dictionary, it would surely have taken months to bring it to the same point, and there would never have been a WMM dictionary.

It must be noted that there are certain kinds of systems for which using FEWL to construct a dictionary may not be successful.  FEWL is not appropriate for systems which basically modify TS (like DRE), systems which depend on information not encoded by FLEWSY (such as syllable boundaries), and systems which insist on non-dictionary pronunciations.  In the first case, it may be possible to produce a dictionary by modifying the DRE dictionary.  In the other cases, one option is to do an approximate conversion using FEWL anyway, and then edit the results manually.  Whether this is practical or not will depend on how pervasive the trouble spots are.

Since I wrote this page, I've put out another list like FEWL, using a variant of Bob Boden's Bobdot notation (Bobdot-ph) in place of FLEWSY.  Bobdot-ph is much easier to describe and process than FLEWSY, but has some limitations.  If your system is based solely on American English, does not reflect secondary stress, and makes no special provisions for compound words, prefixes or suffixes, an approach based on Bobdot-ph is likely to be simpler and easier than using FEWL.  Also, I have another project similar to FEWL called CAAPR, oriented towards both RP and GA rather than solely towards American English.  Though CAAPR is, at this time, less developed than FEWL, it may be useful for systems not content to target a single variety of English.

For all the gory details about FEWL and FLEWSY, click here.  See here for infomation on Bobdot-ph, and here for information on CAAPR.




To comment on this page, e-mail Alan at wyrdplay.org

Go to wyrdplay.org home page
Go to wyrdplay.org spelling system roster