About FEWL and FLEWSY

FEWL (the FLEWSY Encoded Word List) is a collection of fairly common English words, encoded in a notation called FLEWSY (Fonemic Latin-1 English Writing SYstem).  FLEWSY has been designed as an encoding of phonemic and morphological information about English words which facilitates transliteration into many reformed spelling systems.  The vocabulary of FEWL is derived from the 12dicts project (see this page).  The FLEWSY notation is ultimately derived from my spelling system MCM.

This page describes the fourth version of FEWL, dating from February, 2005.

Here is a tiny fragment of the FEWL list:

   ajè'nda  - agenda
   E'j(ent  - agent
   E'j-O'ld  - age-old
   E'jë$  - ages
   aglò·merE'Xon  - agglomeration

Each line is composed of a FLEWSY encoding followed by the separator "  - " and the traditional spelling of the word.  For some simple reform systems, it may be possible to transform the FEWL list into a list of respelled words using nothing more than a number of "Replace" commands in a good text editor.  For more complex systems, some programming will be necessary to perform the necessary transformations.  The level of programming expertise required for this purpose is not great.  Anyone familiar with the use of "regular expressions" will almost certainly be able to construct a suitable transformation program.  The level of effort required will of course depend on the complexity of the target system.  See this page for a discussion for programmers of how to implement a FEWL converter, including a link to a .zip file containing sample programs in Python.

FEWL is available for download here.

The FLEWSY Notation

FLEWSY is an encoding system for English words, using the Latin-1 character set.  It encodes four kinds of information: 

• Phonemic information, based primarily on American pronunciation
  
• Traditional spelling information
  
• Stress information
• Morphological information (inflections, prefixes, suffixes and compounding)

FLEWSY is too complex to be of much use itself as a spelling system.  But it contains much information that is relevant to common reformed spelling schemes.  When FLEWSY is transformed into another encoding, the information which is not relevant to the target scheme can be simply ignored in the transformation process.

The heart of FLEWSY is its encoding of English phonemes.  It is easiest to present FLEWSY as a phonemic encoding, and then to present the remaining elements as embellishments of that encoding.  Most examples below are encoded only in the phonemic part of FLEWSY for simplicity - their actual encodings in the FEWL list are necessarily more complicated.

Though FLEWSY's complexity makes it unsuitable for use as a spelling system, I have devised a practical (if odd-looking) system FLOSS (Fonemic Latin-One Spelling System) based rather closely on the phonemic parts of FLEWSY.  Despite its high degree of unfamiliarity and use of an extended mixed-case character set, it is surprisingly readable, though not as phonemic as MCM, from which it was ultimately derived.

Consonants

Vowels

FLEWSY uses the following symbols for English vowels:

Notes:

1. All five unaccented vowel letters represent the schwa sound.  The letter used in a particular word is taken from its traditional spelling.  Additionally, the symbol µ represents the schwa sound where the sound results from destressing the U or V sound, and the symbol 3 represents the schwa when the English spelling shows no vowel (as with bubble or chasm).

2. The symbols â and ê are used for the unstressed short i sound when it is traditionally spelled with a or e.  These symbols are used only when the sound is pronounced distinctly.

3. The symbolls ä, ë, ï, ö and ü are used for an unstressed, indistinct sound which varies from speaker to speaker between the short i sound and the schwa.  This sound frequently occurs in suffixes and inflectional endings such as -ity, -ness, -age and -ed.  As with the schwa representations, the letter used is derived from the corresponding traditional spelling.

4. The symbol @ is used for the short a (/{/) sound in American English when the corresponding British English pronunciation is /A:/.  In all other cases, the symbol à is used for this sound.
   

5. The symbols A and ò represent the same sound in American English.  The ò indicates that the British pronunciation of the word uses the British short o sound (Sampa /Q/), and A that the British pronunciation uses the American sound.  If the British pronunciation is neither /Q/ nor /A:/, the symbol corresponding to the standard spelling is used.  A British version of FLEWSY should use ò exclusively for the British short o.

6. The symbols Ö and Ø represent the same sound in American English.  The Ö indicates that the British pronunciation of the word uses the British short o  sound (Sampa /Q/), and Ø that the British pronunciation uses some other sound.
   

7. The symbol & represents a stressed er sound (Sampa /3`/), as in herd or eternal.  An unstressed er sound (Sampa /@r/) is represented as a schwa followed by r, as in drYver [driver] or arìTmetìk [arithmetic].

8. The symbol ¥ is used to indicate the shorter less distinct form of long e which frequently occurs at the end of words and before other vowels, and which is commonly spelled y.  This vowel is often represented in Sampa as /i/, as opposed to the full long e sound represented as /i:/.

Vowels for RP (British) Compatibility

FLEWSY uses the following symbols for vowel sounds used before the letter r.  In American English, these sounds are generally merged with one of the vowel sounds above.  By permitting these sounds to be spelled differently depending on how they are spoken in RP, FEWL can better support systems which attempt to handle British as well as American English.

Notes:

1. Some American dialects have a similar sound.  The word "Mary", if pronounced differently from "marry" (màr¥) and "merry" (mèr¥) is the standard example.

2. In many words ending with -ary, such as "secretary", the American pronunciation is a stressed /Eri/, while the British pronunciation is an unstressed /@ri/.  Because this pronunciation is compatible with the -ary spelling, such words are rendered with Ër¥ rather than èr¥.
   

3. Ordinarily, the American Vr spelling is pronounced /U@(r)/ in RP.  The Ür spelling is used for the unusual cases where the /Ur/ pronunciation is also used in American English.
   

Other pronunciation symbols

A few FLEWSY characters indicate optional pronunciation elements, or sounds which depend on context.  These are as follows:

• The ? symbol is used to indicate an optional sound, that is, one pronounced by some speakers and not by others.   This symbol is used in such words as: dy?Ut¥ [duty], krìmï?nal [criminal], hìsto?r¥ [history], klOD?z [clothes], gìvand?tEk [give-and-take].
• The * symbol is used to indicate that the following letter should be interpreted in one of two specific ways.  The sequence *l indicates either l or al, as in the word màjìk*l¥ [magically].  The sequence *y indicates either y or ¥, as in the word petrOl*yum [petroleum].  The sequence *w indicates either w or hw, as in *wìs3l [whistle].  The sequence *X indicates either X or C, as in the word atèn*Xon [attention].
  
• The $ symbol is used to indicate a plural, which may be pronounced either as s or z depending on the preceding symbols.  The symbol is also used in some compound words formed from plurals.  Examples: àntìk$ [antics],  gl@së$ [glasses], h&d$man [herdsman].
• Similarly, the þ symbol is used to indicate a past tense, which may be pronounced as t or d depending on the preceding symbols.  It is also used in compound words formed from past tenses, and in words using the -ed ending with a noun.  Examples: wØrpþ [warped], rêlEtëþ [related], bÏrdëþ [bearded], brOkenhArtëþ [brokenhearted].
• It should also be noted that the sequence àr (as in pàrot and pòpyµlàrït¥) is frequently pronounced as èr.
  

Stress indicators

FLEWSY uses two symbols to indicate stress: ' and · .  The apostrophe indicates primary stress, and the raised dot represents secondary stress.  More than one syllable may be shown with primary stress, especially in compound words.  The stress marker always follows a vowel.  Examples:  kà't [cat], Ø·rganïzE'Xon [organization], sè'kondhà'nd [secondhand].  A small number of words, which are normally pronounced unstressed, are shown with no stress: ov [of], De [the], tµ [to].

Affixes and compound words

FLEWSY indicates, in a somewhat imperfect way, when words are combined or extended through the use of affixes or compounding.  The following symbols and conventions are used for this purpose:

• The symbols - and = are used to separate parts of compound words.  The hyphen is used when it seems reasonably clear that the word is indeed a compound, while the equal sign is used for more figurative or speculative compounds, or even for nonsense words with the form of compounds.  The choice of which symbol to use, if any, was in many cases rather arbitrary, and possibly obviously wrong to those more etymologically knowledgeable than myself.  Examples:  rU'm-mE·t [roommate], mI'n=*wY·l [meanwhile], lì·kët¥=splì't [lickety-split].  Derivatives of compound words will still show the compounding.  For instance "broad-mindedness" is written brØ·d-mY'ndëþnës, despite the non-existence of the word "mindedness".
• The symbols / and \ are used when a word is contained within another, but the word as a whole is not a true compound.  The symbol / is followed by the contained word, while \ is preceded by it.  \ is often used following standard English prefixes.  The situation with suffixes is rather more complicated, and is discussed below in a separate topic.  As with - and =, derivatives of words transcribed with a / or \ will retain the character.  Examples: à'tmo/sfÏ·r [atmosphere], mù'l/bè·r¥ [mulberry], CE'nj\lìG [changeling], mì·s/bêhE'v [misbehave], spØ'rt$\k@·ster [sportscaster].
• The handling of suffixes is complicated in FLEWSY, because the effects of suffixes are much more varied than prefixes.  There are two different reasons for marking suffixes.  One is to allow the spelling of a root word to be preserved when a suffix is attached, especially if its pronunciation is unchanged.  The other is to allow standardization of the spelling of the suffix.  The characters \ and } are used in FLEWSY for the first purpose, and the characters ( and { for the second.  \ and } look backwards, to the preceding root, while ( and { look forward to the suffix itself.  A marking from each class may appear at the same point in a word, as in vì'zit\(or [visitor].  The backward notation always precedes the forward notation in such cases.
  
• \ and } are used to mark root words preceded by a suffix.  \ is used if the pronunciation of the root is unchanged, except possibly for changes between primary and secondary stress.  } is used when the pronunciation of the root is changed.  Such changes are limited to vowel changes, and/or insertion or removal of a schwa before a liquid.  Inflections not marked with a single $ or þ character are marked with \ or } where appropriate.  Examples: hà'p¥\nës [happiness], hà'pï}l¥ [happily], bù'ter\¥ [buttery], hù'Ggr}¥ [hungry], dêvO't\ëþ [devoted], ny?Utrà'l}(ït¥ [neutrality].  When a word contains several suffixes suitable for marking with \, they are usually treated as a single suffix, especially for very common suffixes, as in frè'nd\l¥nës [friendliness] and alè'j\ëþl¥ [allegedly].
  
• The symbols ( and { are used to mark the use of certain standard suffixes.  Most suffixes are not marked in this way.  A suffix is marked when it is reasonably common and when it has more than one FLEWSY representation, either due to pronunciation differences or due to FLEWSY's use of the traditional spelling in its representation.  For instance, the suffix -man is marked because it may be represented as either mà·n or man depending on stress, while the suffix -able/ible is marked because it may be represented as either ab3l or ib3l.  In most cases, the ( character is used to mark the suffix.  The { marker is used instead for borderline cases, which could reasonably be construed as an instance of the suffix, but which perhaps do not meet the strict definition.  For instance, the word "terrible" is encoded as tè'r{ib3l because, unlike "visible", it does not indicate a capability, but yet, unlike complete deceptions like "mandible", it is an adjective which is very similar to more literal -ible adjectives.  Examples of suffix markings:  kò'Ggrës\(man [congressman], mà'd\(mà·n [madman], hù'nt\(er [hunter], sE'l\(or [sailor], mE'{or [mayor], pE'X(ens [patience], dêfY'\(ans [defiance].  Note that suffixes may be indicated in components of a compound word, as in *wI·l\{er-dI'l\(er [wheeler-dealer].  The specific suffixes marked in this way in FLEWSY are: -ability/ibility/ubility, -able/ible/uble, -ably/ibly/ubly, -ally, -ance/ence, -ancy/ency, -ant/ent, -ar/er/or, -ee, -efy/ify, -ety/ity, -ication, -ification, -iful, -ist and -man.  The adjectival suffix -ous is also marked with a { whenever a backwards marker does not also appear.  This is to allow uses of this suffix to be distinguished from the many other words in which a schwa-s ending does not represent this suffix, such as "chorus", "furnace", "princess" and the like.  Similarly, the suffix -ics associated with sciences or occupations are marked with {ìks, if there is no corresponding -ic word (such as mà·themà't{ìks), or with ik\s, if there is (è·konò'mik\s).
  
• The symbol " indicates a position in which the traditional spelling uses an apostrophe.  This makes the programming easier for systems which want to use apostrophe like TS.  Examples: mà'"m [ma'am], hà'v\3n"t [haven't], and lI'vY·"$ [Levi's].
  
• A few words like U-turn and x-ray are normally spelled with a single initial letter rather than strictly phonetically.  The initial letter for such words is set off with colons in FLEWSY, as in :U:t&·n and :x:rE·.
  

Note that all of these symbols (other than the colon and the quote mark) are used in a quite imprecise way - while they may provide useful guidance to whether a word is related to other words, they must not be taken too seriously.  The main value of these notations is that they allow systems which vary the representation of sounds by position to take into account connections between words that are not obvious.  For instance, a system which represents vowel sounds differently at the end of a word might want to treat the a in sofabed as if it ended the word, but this is hard for a converter to handle automatically unless the compounding is noted in the FLEWSY representation.

How FEWL was compiled

The FEWL list was derived from the TARDRE dictionary for my spelling system DRE.  The pronunciations listed in FEWL reflect what I call CLAEP, Consensus Lexicographic American English Pronunciation.  For each word, the pronunciation shown was determined from the first listed pronunciation in each of three dictionaries, resolved by majority vote.  The three dictionaries are the CD-ROM version of The Merriam-Webster Collegiate Dictionary 11th edition, the fourth edition American Heritage Dictionary CD ROM, and the Longman Pronunciation Dictionary.  In the event these sources did not agree, the Random House Unabridged Dictionary CD ROM was used as a tiebreaker.  The listed pronunciations often do not reflect my own.  Some words have two or three different pronunciations, depending on part of speech or meaning, in which case one pronunciation for each distinct form is listed, and an indication of which form it applies to follows the traditional spelling in parentheses.

It is worth noting how a few specific notations were determined:

• The distinctions between ò and A, à and @, èr and Ër, etc. were based on the British pronunciations listed in Longman.
• The notations ï, ë, etc. were used when the pronunciations shown for a specific vowel were divided between the short i and the schwa, with no other options.
• Most uses of y?, showing an optional y before one of the vowels U, V or µ, were derived from the British pronunciations in Longman.  Most other uses of the ? notation were based on Merriam-Webster.  In any event, the adding of ? was less than systematic.
  
• The ¥ symbol generally corresponds to the use of /i/ rather than /i:/ in Longman.

For various reasons, a few words in FEWL are shown with a pronunciation which differs from the CLAEP pronunciation.  These "signature words" are listed and discussed below.

The FEWL word list was constructed in a number of steps, each step adding more detail to the previous step.  The results of the earlier steps must be considered to be more reliable than the later steps, since they have been more thoroughly checked.  The steps were as follows:

1. The process began with the DRE dictionary.
2. Four ambiguities in the DRE dictionary were resolved: distinguishing the two sounds of th (D and T), the two sounds of x (ks and gz), the three sounds of ng (G, Gg and ng), and the two sounds of nk (Gk and nk).
   
3. The result of step 2 was transcribed into the basic FLEWSY notation.  However, the symbol R was used for ar/er/ir/or/ur/3r.  This facilitated correcting R to & or vice versa.
4. Short vowels and schwas were distinguished.  The vowels à, è, ì, ò and ù were reduced to a, e, i, o, u and µ where appropriate.  Also, the vowels â and ê were marked where appropriate, and ¥ was marked at word end and before a few suffixes.
5. Occurrences of ä, ë, ï, ö and ü were identified and marked.
6. The ?, *, $ and þ notations were added.  ¥ was added where it occurred before another vowel.
7. Stress markings were added and compound words were identified.
   
8. Most of the variably spelled suffixes were noted and marked with ( or {.
   
9. The R symbol was replaced by ar/er/ir/or/ur/µr/3r, and the -ar/er/or suffix was noted and marked.  This point marked the release of version 1 of FLEWSY and FEWL.
   
10. Additional suffixes were identified and marked with \ and }, and apostrophes in the original spelling were noted using the " symbol.  The : notation for words like T-shirt was also added at this time.
11. The \ and } markings were extended to words embedded in related words (such as "simple" in "simpleton" or "merit" in "meritorious") which resist interpretation as simple suffixation.  This marked the release of version 2 of FLEWSY and FEWL.
    
12. The à symbol was replaced by @ when appropriate.  This marked the release of version 3 of FLEWSY and FEWL.
    
13. The symbols Ë, Ï, Ö, Ü and ú were introduced where appropriate.
    

Because it was compiled so late in the process, and because the criteria used were rather subjective, the information on compound words and suffixes must be considered considerably less reliable than the other parts of FEWL.

Signature words

The concept of signature words was originally introduced for commericial word lists (as for spelling checkers) to indicate unusual words added to the list, allowing unauthorized copies of the list to be recognized.  This concept has been broadened to indicate words on a list which technically do not belong.  For instance, the ENABLE word list (see here) is a large non-commercial list of words recognized as valid plays in Scrabble.  The ENABLE signature words are a small set of perfectly valid words which are not however permitted in Scrabble tournaments.  (Scrabble is a trademark of Hasbro, Inc.)

The FEWL signature words are words which are listed with a pronunciation other than their CLAEP pronunciation.  Most of these words were given a less common pronunciation to improve their representation in DRE, for instance, to allow homonyms to be distinguished, or to increase the regularity of the DRE spellings.

Very demanding users of FEWL might wish to reference the actual CLAEP-determined pronunciation - for this reason, the FEWL signature words and the way they "should" have been represented in FEWL are listed below:

   à'ber(ant  - aberrant  [should be àbè'r(ant]
   A'rkEì·z3m  - archaism  [should be A'rk¥i·z3m]
   ArC/dY'osIs  - archdiocese  [should be ArC/DY'osës]
   A'nt  - aunt  [should be @'nt]
   bolO'nya  - bologna  [should be bolO'n¥]
   bØ'rO  - borrow  [should be bò'rO]
   bØ'rO\(er  - borrower  [should be bò'rO\(er]
   kolè'sterØ·l  - cholesterol  [should be kolè'sterO·l]
   kò'mptrO·l{er  - comptroller  [should be kontrO'l{er]
   tsA'r  - czar  [should be zA'r]
   tsAr\I'na  - czarina  [should be zAr\I'na]
   dY'osIs  - diocese  [should be dY'osës]
   dØ'gØ'n  - doggone  [should be dò'gò'n]
   fè'brUË·r¥  - February  [should be fè'by?UË·r¥]
   hà·lOI'n  - Halloween  [should be hà·lowI'n]
   hù'H  - huh  [should be hu', or so they say]
   ì·/maCV'r  - immature  [should be ì·/maty?V'r]
   ì·/maCV'r\l¥  - immaturely  [should be ì·/maty?V'r\l¥]
   ì·/maCV'r\(ït¥  - immaturity  [should be ì·/maty?V'r\(ït¥]
   ì·n/dêpè'nd\(ens  - independence  [should be ì·n/dëpè'nd\(ens]
   ì·n/dêpè'nd\(ent  - independent  [should be ì·n/dëpè'nd\(ent]
   ì·n/dêpè'nd\(ent\l¥  - independently  [should be ì·n/dëpè'nd\(ent\l¥]
   jè'zy?U·ït  - Jesuit  [should be jè'JU·ït]
   jU'dEì·z3m  - Judaism  [should be jU'd¥ì·z3m]
   kimO'nO  - kimono  [should be kimO'no]
   lI'Ezò·n  - liaison  [should be lI'azò·n]
   lyA'ma  - llama  [should be lA'ma]
   maCV'r  - mature  [should be maty?V'r]
   maCV'r\l¥  - maturely  [should be maty?V'r\l¥]
   maCV'r\(ït¥  - maturity  [should be maty?V'r\(ït¥]
   mIlyU'  - milieu  [should be milyV']
   nAI'v  - naive  [should be nYI'v]
   nAI'v\l¥  - naively  [should be nYI'v\l¥]
   nAI·v\tE'  - naivete  [should be nYI·v\tE']
   nè'sësË·rï}l¥  - necessarily  [should be nè'sësè·rï}l¥]
   nè'sësË·r¥  - necessary  [should be nè'sësè·r¥]
   prI·maCV'r  - premature  [should be prI·maty?V'r]
   prI·maCV'r\l¥  - prematurely  [should be prI·maty?V'r\l¥]
   prYmË'rï}l¥  - primarily  [should be prYmè'rï}l¥]
   rE'X¥O·  - ratio  [should be rE'XO]
   rêkò'nasans  - reconnaissance  [should be rêkò'nazans]
   sà'karI·n  - saccharine  [should be sà'ka?rïn]
   sèrE't\ëþ  - serrated  [should be serE't\ëþ]
   sØ'rO  - sorrow  [should be sò'rO]
   sØ'rO\ful  - sorrowful  [should be sò'rO\ful]
   sØ'rO\fu?l¥  - sorrowfully  [should be sò'rO\fu?l¥]
   sØ'r¥  - sorry  [should be sò'r¥]
   stA'nC  - stanch  [should be stØ'nC]
   T&'z/dE  - Thursday  [should be T&'zd¥]
   tµmØ'rO  - tomorrow  [should be tµmò'rO]
   trØmà'\tìk  - traumatic  [should be tramà'}tìk]
   ty?U'z/dE  - Tuesday  [should be ty?U'zd¥]
   ù·n/àksè'pt\(ab3l  - unacceptable  [should be ù·n/aksè'pt\(ab3l]
   ù·n/àksè'pt\(abl¥  - unacceptably  [should be ù·n/aksè'pt\(abl¥]
   ùn/kù'mfort\{ab3l  - uncomfortable  [should be ùn/kù'mft{ab3l]
   ùn/kù'mfort\{abl¥  - uncomfortably  [should be ùn/kù'mft{abl¥]
   ù·n/wò'Xþ  - unwashed  [should be ù·n/wØ'Xþ]
   v&'tëbrI·  - vertebrae  [should be v&'tëbrE·]
   vò·luntË'rï}l¥  - voluntarily  [should be vò·luntè'rï}l¥]
   wÏ'r/wV·lf  - werewolf  [should be wË'r/wV·lf]
   wÏ'r/wV·lv$  - werewolves  [should be wË'r/wV·lv$]
   *wùt=è'ver  - whatever  [should be *wòt=è'ver]
   *wì't/sù'n/dE  - Whitsunday  [should be *wì't/sù'nd¥]
   yè'  - yeah  [should be yè'a]

For some practical recommendations on converting FLEWSY to your own notation, please proceed to this page.

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