A simple phonetic respelling for the English language
Perl Haskell Makefile
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
data
samples
scripts
src
.gitignore
LICENSE.txt
Makefile
README.md
Setup.hs
phoneng.cabal

README.md

Lytspel, a Simple Phonetic Respelling for the English Language

The idea: Spelling should be fun, not a burden. The traditional English spelling system is afflicted with exceptions and conflicting rules, making writing and reading texts unnecessary hard.

Lytspel is a proposal for reforming the English spelling in order to make it strictly follow the alphabetic principle. The alphabetic principle means that there is a predictable relationship between written letters and spoken sounds. When you see a written word, you know how to pronounce it (even if you don't know the word itself), and vice versa.

Traditional English spelling (tradspell) deviates from the alphabetic principle in many ways. The mapping between written and spoken words is very complex and riddled with irregularities and exceptions. Lytspel (for «lytwait speling» or ‹lightweight spelling›) is a proposal to change this while stilling keeping as much visual similarity to tradspell as reasonably possible.

Lytspel not only creates an unambiguous mapping between the sounds used in English and their written representations, it also indicates which syllable in each word is the stressed one.

In addition to the reform proposal, there will be a full dictionary and a converter that translates traditional spelling into Lytspel. Both are currently work in progress.

Here are two example paragraphs written in Lytspel. The first is The North Wind and the Sun, one of Aesop's fables that's frequently used as an example in phonetics, while the second is made up of two nonsense sentences that contain most of the sounds of English.

Dhe North Wind and dhe Sun wur dis'piuting wich wos dhe strongger, wen a traveler caim a'long rapd in a worm cloak. Dhay a'greed dhat dhe won hu furst su'xeeded in maiking dhe traveler taik his cloak of shood bee con'siderd strongger dhan dhe udher. Dhen dhe North Wind blu as hard as hi cood, but dhe mor hi blu dhe mor cloassli did dhe traveler foald his cloak e'round him; and at last dhe North Wind gaiv up dhe a'tempt. Dhen dhe Sun shynd out wormli, and i'meediatli dhe traveler took of his cloak. And so dhe North Wind wos o'blyjd to con'fess dhat dhe Sun wos dhe strongger ov dhe tu.

Dhat quik baizh fox jumpd in dhe air oaver eech thin dog under a caam autum muun. Look out, y shout, for hi's foild iu yet a'gen, cri'aiting cayoss.

Here are the same paragraphs written in tradspell. Lytspel might look a bit unusual at first, but should be easy to get used too. Just try reading it out aloud and you'll soon get the hang of it.

The North Wind and the Sun were disputing which was the stronger, when a traveler came along wrapped in a warm cloak. They agreed that the one who first succeeded in making the traveler take his cloak off should be considered stronger than the other. Then the North Wind blew as hard as he could, but the more he blew the more closely did the traveler fold his cloak around him; and at last the North Wind gave up the attempt. Then the Sun shined out warmly, and immediately the traveler took off his cloak. And so the North Wind was obliged to confess that the Sun was the stronger of the two.

That quick beige fox jumped in the air over each thin dog under a calm autumn moon. Look out, I shout, for he's foiled you yet again, creating chaos.

Conventions and Abbreviations

The following formatting conventions are used in this document:

  • «double guillemets» for Lytspel (and spellings that haven't changed)
  • ‹single guillemets› for traditional spellings
  • /slashes/ for phonetics, written in the SAMPA alphabet (with some slight modifications to make it more suitable for English)

Abbreviations used:

  • RP: Received Pronunciation, the standard form of English spoken in the United Kingdom, especially in England and Wales
  • GA: General American, English as typically spoken in the United States

The Rules of Lytspel

The syllable that carries the primary stress in a word is preceded by an apostrophe, e.g. «dis'cuss, in'telijent, co'miti, cri'ait» ‹discuss, intelligent, committee, create›. No stress marker is used if the word only has a single syllable or if the first syllable carries the primary stress, e.g. «big, week, camel, garden».

The short vowels are written as follows:

  • «a» /a/ as in «cat»
  • «e» /E/ as in «pen»
  • «i» /I/ as in «big»
  • «o» /Q/ as in «dog»
  • «oo» /U/ as in «book»
  • «u» /V/ as in «club»

The long vowels and diphthongs are written as follows:

  • «aa» /A/ as in «paam, faadher» ‹palm, father›
  • «ai» /eI/ as in «aim, saint, faiss» ‹...face›
  • «au» /O:/ as in «auther, paus» ‹author, pause›
  • «ee» /i:/ as in «tree, teem» ‹...team›
  • «oa» /oU/ as in «boat, hoam» ‹...home›
  • «oi» /OI/ as in «oil, point, a'void» ‹...avoid›
  • «ou» /aU/ as in «mouth, doun» ‹...down›
  • «uu» /u:/ as in «muun, ex'cluud, gruup» ‹moon, exclude, group›
  • «iu» /ju:/ as in «valiu, kiut, com'piuter» ‹value, cute, computer›
  • The spelling «ue» is used in words that are typically spoken with /ju:/ in RP, with /u:/ in GA, e.g. «due, nuetrel, tuen» ‹...neutral, tune›
  • «y» /aI/ as in «pryss, styl, dry» ‹price, style...›

The consonants are written as follows:

  • «b» /b/ as in «bed»
  • «ch» /tS/ as in «much»
  • «d» /d/ as in «desk»
  • «dh» /D/ as in «dhem» ‹them›
  • «f» /f/ as in «fat»
  • «g» /g/ as in «big, garden, gess» ‹...guess›
  • «h» /h/ as in «hot»
  • «j» /dZ/ as in «joy, dijit» ‹...digit›
  • «l» /l/ as in «leg»
  • «m» /m/ as in «mad»
  • «n» /n/ as in «now»
  • «ng» /N/ as in «long»
  • «p» /p/ as in «pop»
  • «r» /r/ as in «run»
  • «sh» /S/ as in «ship»
  • «t» /t/ as in «test»
  • «th» /T/ as in «thin»
  • «v» /v/ as in «ever»
  • «w» /w/ as in «west, wen» ‹...when›. Note: In some English accepts, the «w» in «wen» ‹when› is spoken differently than the «w» in «west», but since most accents (including RP and GA) no longer make this difference, it is not preserved in Lytspel (cf. wine-whine merger).
  • «y» /j/ as in «yet»
  • «zh» /Z/ as in «mi'raazh» ‹mirage›
  • The sound /k/ is written as «c» or «k», depending on context:
    • «k» is used before «e, i, y», e.g. «keep, king, kynd, calkyu'laition» ‹...kind, calculation›
    • «k» is also used before «h» to prevent confusion with the digraph «ch» /tS/, e.g. «lunkhed» ‹lunkhead›
    • «c» is used before any other letter, e.g. «cat, cost, crisp»
    • «c» is also used at the end of words ending in unstressed /ak/ or /Ik/, e.g. «mainiac, majic, public» ‹maniac, magic...›
    • «k» is used at the end of other words ending in /k/, e.g. «book, drink, quik, speek, naimsaik» ‹...quick, speak, namesake›
  • The sound /s/ is written as «c», «s», or «ss», depending on context:
    • «c» is used between any vowel and a vowel starting with «e, i, y», e.g. «deecent, ex'plicit, nececeri, de'cyd» ‹decent, explicit, necessary, decide›
    • «ss» is used between any vowel and a vowel starting with «a, o, u», e.g. «as'sault, epissoad, as'sumption» ‹assault, episode, assumption›. Note: If a stressed syllable starts with «ss», the stress marker is inserted before the second «s».
    • «ss» is also used at the end of words except after any of the consonants /f, k, p, t, T/, e.g. «less, miss, evidenss» ‹...evidence›
    • «s» is used anywhere else, e.g. «sun, desk, chips, up'sets» ‹...upsets›
    • «s» is also used in the prefixes «dis, mis» regardless of which letter follows, e.g. «disa'gree, disen'chantment, misin'turpret, misunder'stand» ‹disagree, disenchantment, misinterpret, misunderstand›
  • The sound /z/ is written as «s» or «z», depending on context:
    • «s» is used between two vowels, e.g. «visit, eesi, dusen, di'syr, bi'sar, ri'sult» ‹...easy, dozen, desire, bizarre, result›
    • «s» is also used at the end of words except after any of the consonants /f, k, p, t, T/, e.g. «his, wishes, meens, paus, quis» ‹...means, pause, quiz›
    • «z» is used anywhere else, e.g. «zeero, ig'zact, ob'zurv» ‹zero, exact, observe›

The schwa /@/:

  • The schwa is an unstressed and neutral vowel that occurs frequently in English words. Lytspel generally writes it as «e» or «u».
  • «u» is used whenever the traditional spelling contains this vowel, e.g. «album, nurvuss, sus'pend» ‹...nervous, suspend›.
  • «e» is used otherwise, e.g. «camel, problem, hiden, sentrel» ‹...hidden, central›.
  • This means that Lytspel doesn't distinguish between unstressed /V/ as in «funda'mentl» ‹fundamental› or /E/ as in «index» and the schwa /@/. However, since the unstressed form of these vowels sounds quite similar to the schwa and there is often no agreement between speakers whether to use one or the other, this shouldn't hurt.
  • A schwa immediately preceding the primarily stressed syllable can also be represented by one of the other three vowel letters («a, i, o», but not «y»), e.g. «a» in «a'bout» ‹about›, «i» in «imaji'naition» ‹imagination›, «o» in «com'piut» ‹compute›. The reason for this is that the vowel preceding the stressed syllable is often spoken so quickly that it's hard to decide whether a schwa or one of the short vowels is pronounced and not all speakers will agree on which one to use. For clarity, it's therefore better to stick with the original representation of the vowel in such cases, even if most speakers might reduce it to a schwa. Additionally, the reduced vowel often becomes clearly audible in related words, e.g. while the second «i» in «imaji'naition» ‹imagination› is often reduced to a schwa, in «i'majin» ‹imagine› it's clearly audible as /I/.
  • Between some consonant pairs at the end of words, the schwa is omitted in writing; since these pairs never occur without an intervening vowel at the end of words, there is no risk of confusion:
    • Final /s@l/ is written «ssl», e.g. «counssl, mussl» ‹council, muscle›
    • Final /z@l/ is written «sl», e.g. «pro'poasl, pusl» ‹proposal, puzzle›
    • The schwa is also omitted between any of /b, p, d, t, f, g, k/ and final /l/, e.g. «flexibl, simpl, midl, hospitl, litl, bafl, singgl, leegl, sta'tisticl, articl» ‹flexible, simple, middle, hospital, little, baffle, single, legal, statistical, article›. An exception is the suffix «ful», e.g. «helpful, meeningful» ‹...meaningful› (abbreviating this suffix to "fl" would be needlessly confusing and could lead to misreadings).
    • Final /z@m/ is written «sm», e.g. «criticism, sarcasm»
    • The schwa is also omitted between /D/ and final /m/, e.g. «algeridhm, ridhm» ‹algorithm, rhythm›

Before «r», the vowels are spelled as follows:

  • «air» /e@/ as in «pair, shair, vairi'aition» ‹...share, variation›
  • «ar» /Ar/ as in «dark, argyument» ‹...argument›
  • «arr» /ar/ as in «carri, em'barress» ‹...embarrass›
  • «eer» /i@/ as in «cheer, yeer» ‹...year›. Note: GA speakers often pronounce this in the same way as the «ir» /Ir/ in «mirer» ‹mirror› (mirror–mere merger).
  • «er» /@`/ as in «number, modern» (r-colored schwa). Note: In RP and other non-rhotic accents, this usually sounds like a normal schwa; likewise, «r» after other vowels is often inaudible or reduced to a schwa.
  • «err» /Er/ as in «merry, errer» ‹...error›
  • «ir» /Ir/ as in «mirer, iritait» ‹mirror, irritate›
  • «oar» /or/ as in «poart, ig'noar, keeboard» ‹port, ignore, keyboard›
  • «or» /O:r/ as in «north, order, iuniform» ‹...uniform›. Note: Most RP speakers pronounce this in the same way as the «oar» /or/ in «board», but the two sounds are distinguished in GA.
  • «orr» /Qr/ as in «sorri, to'morro» ‹sorry, tomorrow›. Note: Most GA speakers pronounce this in the same way as the «or» /O:r/ in «north», but the two sounds are distinguished in RP.
  • «oor» /Ur/ as in «poor, toor, in'shoorenss» ‹...tour, insurance›
  • «iur» /jUr/ as in «Iurep, piur, ob'skiur» ‹Europe, pure, obscure›
  • The spelling «uer» is used in words that are typically spoken with /ju:r/ or /jUr/ in RP, with /Ur/ in GA, e.g. «duering, nuerel» ‹during, neural›
  • «our» /aUr/ as in «flour, our» ‹flour, our or hour›
  • «ur» /3:/ as in «turn, purfect, furst, con'surn, obzur'vaition» ‹...perfect, first, concern, observation›. This sound only occurs before «r» and can be considered a stronger (and often stressed) variant of the schwa.
  • «urr» /Vr/ as in «hurri, current» ‹hurry...›
  • «yr» is spoken /aI@`/ at the end of words (e.g. «fyr, in'spyr» ‹fire, inspire›), /aIr/ elsewhere (e.g. «vyruss, yreni, spyrel» ‹virus, irony, spiral›)

Some vowels are spelled differently at the end of words or before other vowels:

  • /eI/ is spelled «ay» at the end of words and before other vowels (instead of «ai»), e.g. «play, holiday, layer, cayoss» ‹...chaos›.
  • /aI/ is spelled «ai» before other vowels (instead of «y»), e.g. «dai'ameter, quaiet, vaielenss, aiern, haierarki, Ha'waii» ‹diameter, quiet, violence, iron, hierarchy, Hawaii›. This prevents confusion with the consonant /j/ (as in «yet») which is also spelled «y» but only occurs before vowels. There is also no risk of confusion with /eI/ (as in «aim, layer») which is always spelled «ay» in this position.
  • The same spelling is used if /aI/ occurs after a vowel at the end of a word, e.g. «raidiai» ‹radii›.
  • /O:/ is spelled «aw» at the end of words and before other vowels (instead of «au»), e.g. «law, drawing».
  • Final /E/ is spelled «eh» (instead of «e»), e.g. «eh».
  • Unstressed /i:/ is spelled «i» at the end of words and before other vowels (instead of «ee»), e.g. «fansi, ne'ceciti, vidio, cri'aition» ‹fancy, necessity, video, creation›. This means that Lytspel cannot distinguish between unstressed /i:/ and unstressed /I/ (as in «big») in these positions, but that shouldn't hurt, as /I/ is rarely used in these positions and dictionaries often disagree on which of these sounds to use anyway.
  • This spelling is also used in the personal pronouns ending in /i:/ (as these are often less stressed than other words in a sentence): «hi, shi, wi, mi» ‹he, she, we, me›. And it's used at the end of prefixes such as «anti, semi» (spoken with /i:/ by many speakers, with /aI/ by others) as well as «di, pri, ri» ‹de, pre, re› (even if the prefix is stressed), e.g. «anti'aircraft, anti'soashel; semi'fynel, semiauto'matic; di'graid, di'bug; prima'choor, pri'requisit; ri'act, ripro'duess» ‹antiaircraft, antisocial; semifinal, semiautomatic; degrade, debug; premature, prerequisite; react, reproduce›.
  • /oU/ is spelled «o» at the end of words and before other vowels (instead of «oa»), e.g. «yelo, zeero, co'operait, po'etic, sho» ‹yellow, zero, cooperate, poetic, show›. This poses little risk of confusion, as /Q/ (as in «dog») occurs rarely or never before other vowels. (Final /Q/ could be written «oh», but in practice it never seems to occur.)
  • This spelling is also used at the end of prefixes such as «baio, imyuno, jio, macro, mycro, nio, nuero, pailio, spectro, suudo, thurmo» ‹bio, immuno, geo, macro, micro, neo, neuro, paleo, spectro, pseudo, thermo›, e.g. «baio'kemistri, imyunodi'fishenssi, jio'grafic, macroico'nomic, mycrowaiv, nio'classic, nuero'surjeri, pailioji'ogrefi, spectrofoa'tomiter, suudosaienss, thurmody'namics» ‹biochemistry, immunodeficiency, geographic, macroeconomic, microwave, neoclassic, neurosurgery, paleogeography, spectrophotometer, pseudoscience, thermodynamics›.
  • /OI/ is spelled «oy» at the end of words and before other vowels (instead of «oi»), e.g. «boy, en'joy, em'ployi» ‹...enjoy, employee›.
  • /aU/ is spelled «ow» at the end of words and before other vowels (instead of «ou»), e.g. «now, power, a'low» ‹...allow›.
  • /u:/ is spelled «u» at the end of words and before other vowels (instead of «uu»), e.g. «clu, ishu, sichu'aition» ‹clue, issue, situation›. This poses little risk of confusion, as /V/ (as in «bus») occurs rarely or never before other vowels.
  • Final /V/ is spelled «uh» (instead of «u»), e.g. «huh».
  • Final schwa /@/ is spelled «a» (instead of «e» or «u»), e.g. «extra, daita» ‹...data›. (Should unstressed /a/ (as in «cat») occur at the end of words, it could be written «ah», but in practice it never seems to occur.)
  • Schwa /@/ before other vowels is always spelled «e» (never «u») to prevent confusion with /u:/ (spelled «u» in this position). Schwa in this position is very rare.

Spellings involving several sounds:

  • The vowel combination /i:@/ or /I@/ is written «ia», e.g. «mis'teeriass, ma'teerial, meediam, obviass, vairiabl, airia, i'meediat» ‹mysterious, material, medium, obvious, variable, area, immediate›. In a few words, «ia» instead represents /i:a/, e.g. «en'thuesiasm, zoadiac» ‹enthusiasm, zodiac› (this distinction is not represented in writing).
  • The R-colored vowel combination /i:@`/ is written «ier», e.g. «thieri» ‹theory›.
  • The vowel combination /oU@/ is written «oe», e.g. «boe, coe'lition, ys'socroess» ‹boa, coalition, isochroous›. In a few words, «oe» instead represents /oUE/, e.g. «floem» ‹phloem› (this distinction is not represented in writing).
  • The vowel combination /u:@/ is written «ua», e.g. «acchual, indi'vijual, influanss, soo'purfluass, crual» ‹actual, individual, influence, superfluous, cruel›. In a few words, «ua» instead represents /u:a/, e.g. «bivuac» ‹bivouac› (this distinction is not represented in writing).
  • Likewise, /ju:@/ is written «iua», e.g. «am'bigiuass, maniual» ‹ambiguous, manual›.
  • The consonant combination /kw/ is written «qu», e.g. «quit, ri'quest» ‹...request›; the letter «q» only occurs in this combination.
  • The sound combination /ks/ is written «x», e.g. «mix, next, ex'plicit, axident» ‹...explicit, accident›.
  • Since «ng» represents the single sound /N/ (as in «long»), the sound combination /ng/ is written «n'g» if a stress marker is needed in this position (e.g. «en'gaij, in'grain» ‹engage, ingrain›), «nng» otherwise (e.g. «cairnngorm, martenngail» ‹cairngorm, martingale›).
  • Before /k/, «n» is pronounced /N/ rather then /n/ (e.g. «bank, tranquil, dis'tinct, bron'kytiss» ‹...distinct, bronchitis›), except if «n» and /k/ belong to different parts of a compound (e.g. «mankynd, reincoat» ‹mankind, raincoat›) or if a word starts with any of «con, en, in, non, un» followed by a /k/ sound (e.g. «con'cluusion, en'counter, in'compitent, increess, noncon'foarmist, un'cleer» ‹conclusion, encounter, incompetent, increase, nonconformist, unclear›).
  • If /Nk/ is spoken where the preceding rule suggests /nk/, the /N/ is written «ng», e.g. «congker, ingk, ingcling, ungcl, ungction» ‹conker or conquer, ink, inkling, uncle, unction›. Some words starting with «con» or «in» are pronounced with /n/ by some speakers, with /N/ by others; in such cases, Lytspel uses «n» for simplicity, e.g. «conquest, increment, concreet, inqui'sition» ‹...concrete, inquisition›.
  • /n/ is written «nn» when it occurs before /k/ in places where «n» would be spoken /N/, e.g. «canncan, melenncoli, noamennclaicher» ‹cancan, melancholy, nomenclature›.

Some sound combinations are written in a special way if they occur at the end of words:

  • Final /O:l/ is written «all», e.g. «ball, in'stall» ‹...install›. This spelling is preserved in derived words, e.g. «walls, calling».
  • Final /S@n/ is written «tion», e.g. «recog'nition, obzur'vaition, section, per'mition, moation, oation, miu'sition» ‹recognition, observation, section, permission, motion, ocean, musician›.
  • Final /Z@n/ is written «sion», e.g. «vision, con'fiusion, i'quaision» ‹...confusion, equation›.
  • The «tion» and «sion» spellings aren't used in words that are derived by appending «n» or «en» to another word, e.g. «ashen, freshen, Rushan, Purzhan» ‹ashen, freshen, Russian, Persian› from «ash, fresh, Rusha, Purzha» ‹ash, fresh, Russia, Persia›.
  • The «tion» and «sion» spellings are preserved in derived words (e.g. «ri'laitionship, tra'ditionel, o'caisionel» ‹relationship, traditional, occasional›), even if the derived form is irregular (e.g. «nationel» ‹national› from «naition» ‹nation›).

Disambiguating multigraphs (sequences of multiple characters representing a single sound or an r-colored vowel):

  • Spellings involving digraphs (such as «ch, sh, ss, ai, oa, oy, ur») and trigraphs (such as «arr, eer») are generally read from left to right: the first letter sequence that can be read as a multigraph should be read as such; e.g. «dueel» ‹dual or duel› is read as «d-ue-e-l» (/dj'u:@l/ or /d'u:@l/), «layer» is read as «l-ay-er» (/l'eI@`/).
  • But the trigraphs «iai, uai» are read as «i» /i:/ or «u» /u:/ followed by «ai» (spoken /aI/ at the end of words, /eI/ otherwise), e.g. «a'preeshiait, flucchuait, raidiai» ‹appreciate, fluctuate, radii›.
  • «oic» at the end of words is always spoken /oUIk/ (two separate vowels), e.g. «hi'roic, stoic» ‹heroic...›.
  • Since final /OI/ is spelled «oy», final «oi» unambiguously represents the two vowels /oUi:/. This combination occasionally occurs in derived words, e.g. «shoi» ‹showy› from «sho» ‹show›.
  • Double «rr» always forms a trigraph with the vowel to its left, e.g. «Februerri» ‹February› is read as «F-e-b-r-u-err-i».
  • In cases where the previous rules would lead to a misreading, Lytspel recommends inserting a hyphen to indicate the correct reading, e.g. «Cro-at, po-it, co-in'syd, co-ope'raition, su-er, Lu-eesi'ana, Aalts-hymer, sheeps-hed» ‹Croat, poet, coincide, cooperation, sewer, Louisiana, Alzheimer, sheepshead›.
  • Those who prefer a more classical form may omit the hyphen between two vowels and place a diaeresis over the second vowel, e.g. «Croät, poït, coïn'syd, coöpe'raition, acchuëri, Luëesi'ana».
  • Multigraphs ending in «r» may be interrupted by a stress marker before the final «r» and are still read as multigraphs, e.g. «co-ope'raition, hor'rific» ‹cooperation, horrific›. The multigraph «ss» is broken in the middle, e.g. «as'sumption» ‹assumption›. Other multigraphs are never interrupted by a stress marker, hence in such cases a marker has the same effect as a hyphen: it indicates that the letters to the left and to the right of it should be read in isolation, e.g. «co'operait» ‹cooperate›.
  • Hyphens are generally not inserted between a base word and a suffix or prefix, e.g. «truer» from «tru» ‹true›, «duer» ‹doer› from «du» ‹do›, «baioastre'nautics» ‹bioastronautics› from «astre'nautics» ‹astronautics›, «mycroa'naleciss» ‹microanalysis› from «a'naleciss» ‹analysis›. Neither are they inserted between the parts of a compound, e.g. «shorthand». If writers feel that confusion might result, they may, of course, insert a hyphen (or a diaeresis).

Rules for derived and related words:

  • In front of suffixes starting with «e, i, y», final «c» is changed to «k», e.g. «trafiking» ‹trafficking› from «trafic» ‹traffic›.
  • Final «c» remains unchanged in front of such suffixes if its pronunciation changes from /k/ to /s/, e.g. «publicist, publicys» ‹...publicize› from «public».
  • In front of suffixes consisting in a single consonant («d, n, s»), final «i» (unstressed /i:/) becomes «ie» to mark it as long, e.g. «studied» from «studi» ‹study›, «vairies» ‹varies› from «vairi» ‹vary›. Otherwise forms such as «*studid» would look like «splendid», but be spoken quite differently.
  • To prevent misreadings, final «o» /oU/ is changed back to «oa» and final «u» /u:/ to «uu» in front of the same suffixes, e.g. «foload» ‹followed› from «folo» ‹follow›, «chuud» ‹chewed› from «chu» ‹chew›, «throan» ‹thrown› from «thro» ‹throw›, «struun» ‹strewn› from «stru» ‹strew›, «po'taitoas» ‹potatoes› from «po'taito» ‹potato›, «shuus» ‹shoes› from «shu» ‹shoe›.
  • A schwa /@/ at the start of suffixes is written as «a» if the base word ends in /i:/ or /u:/, e.g. «eesiast» ‹easiest› from «eesi» ‹easy›, «vairiabl» ‹variable› from «vairi» ‹vary›, «duabl» ‹doable› from «du» ‹do›, «a'greeabl» ‹agreeable› from «a'gree» ‹agree›. This ensures consistency with the combined spellings «ia» and «ua» used elsewhere (see above) and prevents confusion with «ie» in words such as «studied». But note that R-colored schwa remains «er», e.g. «eesier» ‹easier›.
  • As noted above, the schwa is omitted between certain consonants and final «l» or «m». To prevent misreadings, the omitted schwa is restored as «e» in front of the suffixes «ait/it, erri, ism, ist, oid, uss, ys» ‹ate, ary, ..., ous, ise/ize› as well as before stressed «ee», e.g. «en'capselait» ‹encapsulate› from «capssl» ‹capsule›, «piupelerri» ‹pupillary› from «piupl» ‹pupil›, «vandelism» ‹vandalism› from «vandl» ‹vandal›, «iuni'vurselism» ‹universalism› from «iuni'vurssl» ‹universal›, «voakelist» ‹vocalist› from «voacl» ‹vocal›, «cristelys, cristeloid» ‹crystallize, crystalloid› from «cristl» ‹crystal›, «lybeluss» ‹libelous› and «lybe'lee» ‹libelee› from «lybl» ‹libel›. The obvious exception are derived words where no schwa is spoken, e.g. «syclys» ‹cyclize› from «sycl» ‹cycle›, «simplism» from «simpl» ‹simple›.
  • Otherwise, end-of-word spellings are preserved in all derived forms as long as of the base word doesn't change (except for possible shifts of stress), e.g.
    • «c» representing /k/: «publicli, publi'caition» ‹publicly, publication› from «public»; «me'canics, me'canicl» ‹mechanics, mechanical› from «me'canic» ‹mechanic›
    • «k» representing /k/: «weeks, weekli» ‹...weekly› from «week»
    • «ss» representing /s/: «classic» from «class»; «con'vinssd, con'vinssing» ‹convinced, convincing› from «con'vinss» ‹convince»; «a'nounssment» ‹announcement› from «a'nounss» ‹announce›; «prinssess» ‹princess› from «prinss» ‹prince›
    • «a» representing /@/: «extras» from «extra», «A'merrican» ‹American› from «A'merrica» ‹America›
    • «aw» representing /O:/: «draws, drawn, drawing» from «draw»
    • «i» representing unstressed /i:/: «sitis» ‹cities› from «siti» ‹city›; «fansiful, fansiing» ‹fanciful, fancying› from «fansi» ‹fancy›; «hapiness» ‹happiness› from «hapi» ‹happy›
    • «o» representing /oU/: «yeloer, yeloest» ‹yellower, yellowest› from «yelo» ‹yellow›; «vurchus» ‹virtues› from «vurchu» ‹virtue›
    • «oy» representing /OI/: «de'stroys, de'stroyd, de'stroyer» ‹destroys, destroyed, destroyer› from «de'stroy» ‹destroy›
    • «y» representing /aI/ (unchanged even before vowels): «hyer» ‹higher› from «hy» ‹high›; «su'plyer» ‹supplier› from «su'ply» ‹supply›
    • «yr» representing /aI@`/: «in'spyrd» ‹inspired› from «in'spyr» ‹inspire›
  • They are also preserved in compounds, e.g. «bakground, extrate'restriel, fyrwurks›» ‹background, extraterrestrial, fireworks›.
  • If the pronunciation of the base word itself changes (beyond mere shifts of stress), the spelling reflects this, e.g. «cri'aition, creecher» ‹creation, creature› from «cri'ait» ‹create›; «dis'cution» ‹discussion› from «dis'cuss» ‹discuss›; «con'cluusion» ‹conclusion› from «con'cluud» ‹conclude›; «acchual» ‹actual› from «act»; «as'sumption» ‹assumption› from «as'suum» ‹assume›; «men'taliti» ‹mentality› from «mentl» ‹mental›.
  • The past tense of regular verbs is formed by appending «ed» if /@d/ is spoken (e.g. «se'lected, waisted» ‹selected, wasted›), «d» if /d/ or /t/ is spoken (e.g. «ri'veeld, con'siderd, payd, fixd, con'vinssd» ‹revealed, considered, paid, fixed, convinced›). The past tense of irregular verbs is spelled as it's spoken, e.g. «kept» from «keep».
  • While in tradspell a noun and a related verb are often written the same even if spoken differently, in Lytspel the spelling reflects the actual pronunciation, e.g. «a suspect, to sus'pect; a rekerd, to ri'coard; a houss, to hous» ‹a suspect, to suspect; a record, to record; a house, to house›.
  • Normally Lytspel doesn't use double consonants, except in those cases where a double consonant is spoken differently than a single one (e.g. «ss» spoken /s/ in certain contexts where «s» is spoken /z/). However, double consonant can also occur if they belong to different parts of a compound or if one belongs to a prefix or suffix and the other to the base word, e.g. «toatelli, suuper'reejennel, un'nececeri» ‹totally, superregional, unnecessary›.

Irregular words:

  • The vowels in the articles «dhe, a, an» ‹the, a, an› retain their traditional spelling.
  • «ey» ‹eye› is written somewhat irregularly to distinguish it from «y» ‹I› and for easier recognition in compounds such as «fishey» ‹fisheye›.
  • «noa» ‹know› is written somewhat irregularly to distinguish it from «no».
  • «oa» ‹owe› is written somewhat irregularly since a verb with just one letter could be confusing (and also to distinguish it from the interjection «o» ‹oh›).
  • «soa» ‹sew or sow› is written somewhat irregularly to distinguish it from «so».
  • «to» remains irregular to distinguish it from «tu» ‹too or two›.

Uses of the apostrophe:

  • Lytspel uses an apostrophe to mark the syllable that carries the main stress (provided it's not the first one). Additionally, as in tradspell, apostrophes are used to mark genitives and contractions.
  • The genitive is marked by appending «'s», e.g. «mudher's car, Linda's sister, children's toys» ‹mother's car...›. The genitive form of a regular plural is formed by appending «'» only, e.g. «taxpayers' iunien, dogs' tails» ‹taxpayers' union...›.
  • «'s» is also used for contractions with «is, has», e.g. «dhair's» ‹there's›.
  • «'d» is used for contractions with «had, wood, shood» ‹...would, should›, e.g. «it'd».
  • «'l» ‹'ll› is used for contractions with «wil, shal» ‹will, shall›, e.g. «shi'l» ‹she'll›.
  • The contraction «y'm» ‹I'm› is short for «y am» ‹I am›.
  • «'r» ‹'re› is used for contractions with «ar» ‹are›, e.g. «dhay'r» ‹they're›.
  • «'v» ‹'ve› is used for contractions with «hav» ‹have›, e.g. «iu'v» ‹you've›.
  • «n't» is used for contractions with «not» where the /Q/ is not spoken, e.g. «shoodn't» ‹shouldn't›. If the first word ends in «n», the second «n» is omitted: «can't» from «can not». In a few cases, the first word may be further contracted: «ain't» for one of «am/is/ar/has/hav not» ‹am/is/are/has/have not›, «han't» for «has/hav not» ‹has/have not›, «shan't» for «shal not» ‹shall not›, «woan't» ‹won't» for «wil not». The pronunciations of «doan't» ‹don't› and «musn't» ‹mustn't› differ from the uncontracted forms, and the spelling reflects this.
  • In all these cases, the genitive or contraction marker is inserted after the last vowel and hence cannot be confused with a stress marker.
  • In other cases, a traditional contraction marker doubles as a stress marker in Lytspel, e.g. «o'clok» ‹o'clock› for «ov dhe clok» ‹of the clock›.
  • This also applies to Irish names starting with «O'», e.g. «O'Coner, O'Neel» ‹O'Connor, O'Neil›. The capitalized «O'» is spoken /oU/ (as in «boat»). Of course, proper names can be spelled as they please, so many names won't be regularized according to the Lytspel rules.
  • In the nonstandard expression ‹y'all›, short for «iu all» ‹you all›, the contraction marker is placed in front of the stressed vowel, marking the original word boundary.
  • In a few cases, words are contracted by eliminating a syllable or a sound for brevity or for literary effect. In these cases, Lytspel recommends placing a contraction marker before the final consonant in monosyllabic words («ai'r, nai'r, oa'r, ee'n, ma'm, bru'r» ‹e'er, ne'er, o'er, e'en, ma'am, br'er› from «ever, never, oaver, eeven, madem, brudher» ‹..., over, even, madam, brother›) and just using a regular stress marker otherwise («wair'air, wairso'air, nor'eester» ‹where'er, wheresoe'er, nor'easter› from «wair'ever, wair'soever, north'eester» ‹wherever, wheresoever, northeaster›).

Inner capitalization instead of stress marker:

  • Various Scottish and Irish names start with ‹Mc› or ‹Mac› /m@k/ followed by a capitalized and stressed second syllable. Lytspel recommends writing these names with «Mc» and omitting the stress marker, e.g. «McAdem, McDonel» ‹McAdam, McDonnell›. If the stressed syllable starts with a /k/ sound, that letter is capitalized (and repeated after the initial «Mc»), e.g. «McCarti, McKee, McQueen» ‹McCarty...›. If any other syllable is stressed, Lytspel recommend writing the word regularly without inner capitalization, e.g. «Makintosh» ‹Macintosh or McIntosh›.

English as an International Language

As an international language, English is currently written and spoken quite differently in various parts of the world.

On the one hand, by using a phonetic spelling, Lytspel eliminates most differences that traditionally exist between written British (BE) and American English (AE), e.g. «culer, senter, orgenys, anelys, lycenss, catalog, aijing, program, chek, gray» ‹colour/color, centre/center, organise/organize, analyse/analyze, licence/license, catalog/catalogue, ageing/aging, program/programme, cheque/check, grey/gray›.

On the other hand, tailoring Lytspel to faithfully represent just one regional variety would cause many deviations from the alphabetic principle for everyone else. And creating different variants of Lytspel each fitting just one regional variety would introduce even larger differences between different regional varieties of written English than traditionally exist between BE and AE.

Instead of going down either of theses routes, Lytspel had been designed as some kind of "global compromise." It aims to keep the phonetic principle reasonably intact for all varieties of spoken English, and particularly for RP (since the English language originates in England) and GA (since the United States are the country with the highest number of native English speakers).

Various details of the Lytspel spellings have already been motivated above as representing sound distinctions that exist either in RP or in GA, and the «ue» and «uer» spellings are a compromise used in words spoken differently in RP and GA. But such compromises are not almost possible without making the written language extremely complex and unwieldy. In other cases, Lytspel therefore resolves differences between RP and GA by preferring the spelling that keeps the written language more similar to tradspell:

  • «a» is used if one variety speaks /a/ and the other /A/, e.g. «last»
  • «o» is used if one variety speaks /Q/ and the other /O:/, e.g. «cloth»
  • «o» is also used if one variety speaks /Q/ and the other /V/, e.g. «from»

XXX Describe additional compromise decisions, e.g.

  • The one that's more similar to the traditionel spelling is preferred.
  • A short vowel such as /@, E, I/ is preferred over a long one such as /aI, i:/.
  • If one variety speaks /Vr/ and the other /3:/, «urr» is used if the traditional spelling includes ‹rr› (e.g. «current»), otherwise «ur» is used (e.g. «curij» ‹courage›)?
  • «er» is used if RP has /@/ and GA has /y@`/, e.g. «figer» ‹figure›?
  • Document that stress marker--free variant is preferred in doubt, e.g. «adult» rather than «*a'dult», «to dictait» ‹to dictate› rather than «to *dic'tate», «sabetaazh» ‹sabotage› rather than «*sabo'taazh».

Cf. American and British English pronunciation differences and Different spellings for different pronunciations.

Overview

TODO This section is older and partially outdated.

This phoneng program suite (short for: Phonetic English) shows how English is pronounced and offers an alternative spelling system that is more consistent and easier to learn and use than the traditional spelling.

TODO These programs aren't implemented yet. The pronounce command shows how English texts are pronounced. The lytspel command converts them into a simplified spelling.

The provided tools can also be used to implement your own spelling reform proposals or to adapt the chosen respellings as needed.

Installation

The phoneng program suite is written in Haskell. To build it from source, you need the Cabal build system. If you use a Debian-based system, install the cabal-install package to get it.

Afterwards clone this repository from GitHub and run the following commands in the main directory:

cabal configure && cabal build && cabal install

The compiled programs should now be in your path and ready to run.

Usage

TODO document

Files and File Formats

All files are in UTF-8 format (some of them may use just the ASCII subset).

Line files (extension: .txt) have one entry per line; line breaks in entries are therefore not allowed.

Key-value files (extension: .txt) are line files where each line represents a key/value pair. Keys and values are separated by ':'; trailing comments introduced by '#' are stripped. No escape syntax is supported, hence keys cannot contain ':', values cannot contain '#', and neither can contain line breaks.

Files in data Directory

TODO Update this section.

  • cmudict-phonemes.txt: key-value file containing a mapping from the phonemes used in cmudict to the corresponding Phonetic English phonemes. Used by the dictbuilder program.

  • custom.csv: CSV file listing those words for which a specific pronunciation should be used. The case of the words listed in the first field is ignored. The second field may contain the following values:

    • B: use British (RP) pronunciation
    • A: use American (GA) pronunciation
    • P: use the PhonEng pronunciation
    • D: don't add the word to the output dictionary (used for rare foreign words or names)
    • O: leave the spelling of the word unchanged
    • Alternatively, a custom pronunciation may be given which is then used to generate the final spelling

    Manually created file; used by lytspelify.

  • moby-phonemes.txt: key-value file containing a mapping from the phonemes used in Moby to the corresponding Phonetic English phonemes. Used by the dictbuilder program.

  • words-not-in-scowl.txt: Line file containing words that aren't listed in SCOWL but should become part of the pronunciation dictionary. Used by the dictbuilder program.

  • phonetic-dict.txt: Line file containing a mapping from words to their pronunciations. If there is just a single pronunciations, the entry is written as word: pron. If the pronunciation of a word depends on which POS (part-of-speed) it is, it is written as word/n: pron1; v: pron2 (where "n", "v" etc. are POS tags). Redirects are written as word:> target, e.g. colour:> color. Generated by the dictbuilder program.

History: Steps used to Generate the Phonetic Dictionary

Some of the following steps require manual intervention. They are described here to document the history of phoneng.

Downloaded and installed knowledge sources:

  • Downloaded SCOWL and VarCon from SCOWL And Friends -- version 2014.08.11 was used to create the distributed dictionary. Unzipped both of them within the data directory and renamed the resulting subdirectories to scowl and varcon.
  • Downloaded the CMU Pronouncing Dictionary -- version 0.7a was used to create the distributed dictionary. It's enough to download the file cmudict.0.7a and store it in a new data subdirectory named cmudict.
  • Downloaded the Moby Pronunciation List by Grady Ward. Created a data subdirectory named moby and unzipped it there.

Then run make from within the data directory, that should handle the rest. TODO Document PATH requirements.

TODO Or execute the following commands manually (the rest of this section is outdated and should be deleted or possibly cleaned up):

Invoked the dictbuilder program within the data directory. This writes a file called phonetic-dict.txt.

Invoked the csvdict script. This writes a file called phoneng-espeak-dict.csv.

Invoked the mergeprons script. This writes a file called phonetic-dict.csv.

Invoked the lytspelify script. This writes a file called lytspel-dict.csv.