ellis@spar.UUCP (Michael Ellis) (04/01/85)
This seems to be a worthwhile topic for this newsgroup. Somebody should develop an ascii phonetic notation, and it might as well be us! =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= > >To start the ball rolling, I'll propose one that could be based >on ASCII characters. There is a problem, of course, in that we >have only 26 letters, and roughly three dozen phonemes, depending >on dialect. So obviously we use digraphs. It would be better if >we could use superposed diacriticals, but what can you do? > >First, the consonants. English consonants can be organized roughly >as follows: > > lab dnt alv ret vel phr >vd.stop b d g >vl.stop p t k >vd.affr c c~ >vl.affr j~ >vd.spir v d~ z z~ >vl.spir f t~ s s~ h >vd.nas m n n~ >vd.smiv w y l r > >Note some funny things, like 'c' meaning 'ts', with 'c~' for 'tsh'. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= English has no use for your symbol (c), which, incidentally, you've placed in the wrong row -- I believe you meant to switch c <=>j. And since digraphs in +h have been firmly established in many languages, why not use them? I'd modify your chart as below: lab l/d dnt alv ret vel phr vd.stop b d g vl.stop p t k vd.affr j vl.affr c vd.spir v dh z zh vl.spir f th s sh h vd.nas m n ng (-ng- in finger is (ngg)) vd.smiv l r Another possibility is to use (h) for (ng) since they have complementary distribution. Or maybe (q), which is otherwise unused. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= >Now for the vowels. They're a bit more confusing, because most >English-speaking people can't even define "vowel", and think that >many things are consonants when they're vowels. But here goes: > > front mid back >st.high i^ u^ ["Strong" high vowels] >wk.high i u ["Weak" high vowels] >st.mid e^ o^ >wk.mid e u~ o >low a^ a =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= My counterproposal again dispenses with the diacritics: front mid back semivowel y `yet' w `wet' st.high iy `beet' uw `boot' wk.high i `bit' u `put' st.mid ey `bait' ow `boat' wk.mid e `bet' o `bought' low.diph ay `bite' aw `bout' low ae `bat' ao `pot' aa `but' diph oy `boy' schwa a (first syllable in about) The above method using (y) and (w) to indicate English `long' vowels is already in common use in many linguistic texts. Ideally, (ae) and (ao) should be digraphs (written together as a single character). Since \a\ is never followed by \e\ or \o\ in any English dialect I know, no confusion should occur when they are written separately. One optional addition is to use (1) to represent unaccented (i), the final vowel of `happy'. Then accented syllables contain (i e ae aa ao o u) and unaccented syllables (a 1). =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= > >Note that with one exception, I've used '~ to mark consonants >and '^' to mark vowels. This pattern will continue with the >"special" vowels that double as consonant sounds: > > m^ n^ l^ r^ > =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= I'd write these sounds as (am) (an) (al) (ar) unaccented or (aar) accented. Note that vocalic (m) (n) (l) never occur in accented sylables. Incidentally, all diacritic marks have been dispensed with in my notation. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= >Dialect difference are to be handled in a straightforward manner, >by using the "superset" of phonemes that are distinguished in the >standard "generic" English, plus pronunciation rules that say how >each dialect maps the phonemic symbols into sounds. Thus, in my >native dialect (West Coast North American), the 'o' and 'a' sounds >are merged into one intermediate sound. But this doesn't change >the fact that the spelling system distinguishes them. I just have >to learn the spelling of the words that are homophones. [From this I assume you pronounce cot and caught identically] =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= This is actually an intriguing problem, but it may be harder to solve than you suggest. There are at least three kinds of problems: 1) In many cases, english dialects preserve phonemic distinctions even where there are differences in pronunciation. For example, British speakers usually drop (r) following vowels, but preserve the lost sound via compensatory vowel lengthening. British speakers would have the rule VOWEL+(r) => LONG VOWEL 2) Unfortunately, many dialects contain distinctions that are nonexistent in others. Words like `new', `dew', `tune', `sewer' (Br= nyuw dyuw tyuwn syuwar, Am= nuw duw tuwn suwar) would be viewed as containing `silent letters' by American speakers, following the rule (y) => 0 /(t,d,n,s) _ VOWEL 3) Some dialectical discrepencies require the addition of extra symbols to account words having different phonemes in different dialects. The most obvious phonemic shift I know of occurs in the vowels in (path) and (ox). I believe that at least two extra vowels will be required (here called (a') and (o')) to handle the words below: British American Common ------- -------- ------ cat kaet kaet kaet path paoth paeth pa'th ox oks aoks o'ks bought bot bot bot Another example is the (d)-like sound in American `pretty', `little'. Some British dialects make this a glottal stop, others pronounce it as an ordinary (t). Shall we call this (t') ? I wonder how many more such additions would be required to handle the majority of english dialects? =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Here's how I'd have spelt your examples: > >Some samples using this spelling system: >Su~m sa^mpl^z yu^zin~ d~is spelin~ sistm^: Saam saempalz yuwzing dhis speling sistam: [.. yuwz1ng .. spel1ng..] > > Old Nu^ Yet anaadhar sistam > --- --- --- ------- ------ > cat ka^t kaet > cot kat kaot > caught kot kot > coat ko^t kowt > curt kr^t kaart > coot ku^t kuwt > word wr^d waard [wurd] > weird wird wird > ward ward word > warred ward word [Du~z eni^wu~n pronawns d^i^z difrentli^?] [Daaz eniwan pranawns dhiyz difrantli ?] [..en1wan..difrantl1?] > they d~e^ [or d~ey] dhey > them d~em dhem > this d~is dhis > these d~i^z dhiyz > think t~in~k think [thiynk/thiyngk]* > button bu~tn^ baatan > bottom batm^ baotam > throttle t~ratl^ thraotal [thraot'al] > thirsty t~r^sti^ thaarsti [thaarst1, thurst1] > prince princ prin(t)s ** > prints prints prin(t)s ** > grep grep grep *(nk) == (ngk). Also, English only seems to allow 5 (6) vowels to occur before this sound (iy, ey, ae, aa, o), and (ao) in polysyllables: sing length long hang hung (congress) **Most dialects of English I've heard confuse all cases of: NASAL+FRICATIVE <=> NASAL+STOP+FRICATIVE as in the below examples: mf <=> mpf lymph (limf) or (limpf) ns <=> nts prince (prins) or (prints) nsh <=> nch mention (menshan) or (menchan) nth <=> ntth ninth (naynth) or (nayntth) ngsh <=> ngksh junction (jaangshan) or (jaangkshan) -michael