Diafon
One alphabet for the sounds of many languages.
One sound, /t͡ʃ/, spelled five ways. Diafon writes it c̦ in each of them.
One alphabet for many languages
A phonemic Latin alphabet. Each letter stands for one sound and is read much the same way across languages, so a new word can be read close to how it is said.
The 26 Latin letters as easy input
Everything is typed with the plain Latin letters, without dead keys. There are a few letter combinations to learn.
Your language's own spelling too
Text can be converted to and from the usual spelling of the supported languages, Pinyin included.
One keyboard layout for fast typing
The supported languages can all be typed on one layout, without switching between keyboards.
One sound, one letter¶
On its own, c is /t͡s/, as in Czech, Polish and Hungarian. The comma below it, called the anchor, makes it /t͡ʃ/: c̦. The anchor changes the other letters of the same family in the same way:
| Letter | s /s/ | z /z/ | c /t͡s/ | x /d͡z/ |
|---|---|---|---|---|
| With the anchor | ș /ʃ/ | z̦ /ʒ/ | c̦ /t͡ʃ/ | x̦ /d͡ʒ/ |
The letters¶
Twenty letters each stand for a sound: 5 vowels and 15 consonants. These are the bases. The other six, c j q w x y, are operators. You type an operator after a base to change its sound. It never stands alone.
Some operators also appear as letters in the script: c is /t͡s/, x is /d͡z/,
q is /q/, j is /j/ and w is /w/. You type them with a base: tc gives
c, dc x, kq
q, ix j and ux
w.
The bases are the letters with a sound no other letter covers. The operators are the letters Latin spelling wastes: c and q repeat k or s, x is only k plus s, and j, w and y are i and u in another role.
Typing a sound¶
Type a sound as a base, then up to two operators, then w if it is long.
Tone languages add a digit after the syllable. An operator never starts a
sound, so everything you type turns into script one way only.
The six operators are always typed in this order. Each has one job after a consonant and one after a vowel, and shows as a small mark. The examples give the letter, what you type and its usual sound.
sj/ʃ/z̦zj/ʐ/țtj/tʲ/n̦nj/ɲ/ij/ɪ/ẹej/ɛ/ọoj/ɔ/ạaj/ə/tx/θ/ḍdx/ð/ḳkx/x/ṇnx/ŋ/ḥhx/ʔ/ix/j/wux/w/tc/ts/xdc/dz/ic/y/öec/ø/ïuc/ɨ/äac/ɒ/tq/tʃ/x̦dq/dʒ/eq/œ/ụ̈iq/ʏ/r̈aq/ɚ/ty/ʈ/n̖ny/ɳ/s̖sjy/sˤ/ay/ɑ̃/èey/ẽ/òoy/õ/sjw/ɕː/aw/ɑː/íiw/iː/éew/eː/- A second operator works on the first, as the first works on the base:
sjis ș,sjyis s̖. - A long vowel takes an acute,
awá. A long consonant is written double,sjwșș. - Aspiration and breathy voice are not operators. Type them with h, as they
sound:
th/tʰ/,bh/bʱ/.
The marks¶
Each mark means one thing on every letter, and none is taken from a national alphabet. A consonant takes one mark. A vowel takes at most one above and one below.
On consonants¶
| Mark | Shape | Operator | What it does |
|---|---|---|---|
| anchor | a long comma joined to the letter | j | hushed or soft: ț n̦ l̦ r̦ ș c̦ |
| dot | a single dot | x | a fixed nearby sound: ṭ ḍ ṇ ḳ ġ ḥ /θ ð ŋ x ɣ ʔ/ |
| reverse anchor | the anchor turned the other way | y | the language's own variant: retroflex t̖ d̖ n̖ (Hindi, Bengali, Tamil), the back hush s̖ z̖ c̖ x̖ /ʂ ʐ tʂ dʐ/ (Polish, Russian, Ukrainian), emphatic t̖ d̖ s̖ z̖ k̖ g̔ /tˤ dˤ sˤ ðˤ ħ ʕ/ (Arabic) |
The anchor is Diafon's own mark: a comma-like stroke joined to its letter, written in one movement by hand. It sits on the thin letters too: ț ș n̦ l̦ r̦ f̦.
On g and p the mark goes above, because the tail below is already taken: p̓ ġ. Capitals have no tail below, so they keep the mark below: P̦ G̣. This is the float rule.
On vowels¶
| Mark | Shape | Operator | What it does |
|---|---|---|---|
| acute | ´ | w | long |
| double dot | ¨ | c | lip rounding swapped: ü ö /y ø/; on a, the back ä /ɑ/ |
| double acute | ˝ | c + w | rounding swapped and long: ű /yː/ |
| grave | ` | y | nasal: à /ã/ |
| underdot | ◌̣ (below) | j | more open: ị ụ ẹ ọ /ɪ ʊ ɛ ɔ/, only in languages that have both; on a, the schwa ạ /ə/ |
The double acute is rounding swap and length in one mark. The dot below combines with one mark above: ẹ́ /ɛː/, ẹ̀ /ɛ̃/, ọ̈ /œ/.
Each letter covers an area¶
Diafon aims to be phonemic, not phonetic. A phonetic spelling records exactly how a sound is said. A phonemic spelling records only which sound it is, among the sounds a language tells apart. The p in English pin has a puff of air and the p in spin does not, but no English word depends on it, so both are p.
So each letter covers a small area of sounds, and reads a little differently from one language to the next, always close. Each language's table shows how its letters are read.
Consonants¶
Consonant areas are narrow, about one step of place or manner. For example, t is made at the teeth in Hungarian and a little further back in English. r covers the trilled /r/ and the English /ɹ/. ṛ covers both the tap and the flap, /ɾ ɽ/, as languages hardly ever tell them apart.
The reverse anchor marks a language's own variant of a consonant, such as the retroflex t̖ d̖ of Hindi or the emphatics of Arabic. The plain letter is a fair approximation of it.
Vowels¶
Vowel areas are wide. e is any front mid vowel, close or open. a is any open vowel, front or back. ü covers /y/ and /ʏ/. Most languages have one vowel in each area and write it bare.
A few languages have two vowels inside one area. Only they write the second one, with a mark that means the same in all of them: ẹ ọ /ɛ ɔ/ in French, Italian, Portuguese, Hindi, Thai and Vietnamese; ị /ɪ/ in Dutch and Ukrainian; ä /ɑ/ in Finnish, Estonian and Persian. A language that makes the split writes it everywhere, in long, nasal and diphthong vowels too: French /ɔ/ is ọ in porte and ọ̀ in bon.
When long and short vowels also differ in quality, length is enough. German Ofen is ó /oː/ and offen is o /ɔ/, so German needs no split. English is the same: bit i, beat í.
The schwa has its own area, ạ /ə/. It sits next to both a and e in too many languages to belong to either.
The limit¶
An area is kept small enough that its sounds are not easily told apart: no u for /y/, no voiced for voiceless. Dialects and fine detail are not written, as in any everyday spelling.
Capitals¶
Case is spelling, not sound. Type the base in uppercase: Sjip
Șip. Operators read the same in either case, so
SJIP is ȘIP. A long consonant capitalises only its
first letter.
Tones¶
A tone is typed as a digit after the syllable, and each digit is the same
tone in each language. The script shows it as a sign after the syllable,
using the Bopomofo tone marks: ma1 ma2 ma3 ma4
maˉ maˊ maˇ maˋ.
| Digit | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
|---|---|---|---|---|---|---|---|---|
| Tone | high | rising | dipping | falling | mid | low | glottal rising | glottal low |
| Sign | ˉ | ˊ | ˇ | ˋ | none | ˍ | ˜ | ˀ |
Mid tone has no sign, as in Thai and Vietnamese spelling. The sign never sits on the vowel, so the vowel keeps its own marks. A language's own spelling can still show tones its usual way, like the accents of Pinyin.
Text and fonts¶
Diafon text is ordinary Unicode, so any app, file or message can carry it. Each letter is a plain Latin letter plus standard combining marks:
| Mark | Character | On g and p |
|---|---|---|
| anchor | U+0326 comma below | U+0313 comma above |
| dot | U+0323 dot below | U+0307 dot above |
| reverse anchor | U+0316 grave accent below | U+0314 reversed comma above |
| acute, grave, double acute | U+0301, U+0300, U+030B | |
| double dot | U+0308 | |
| dot below a vowel | U+0323 | |
| tone signs | U+02C9 ˉ, U+02CA ˊ, U+02C7 ˇ, U+02CB ˋ, U+02CD ˍ, U+02DC ˜, U+02C0 ˀ |
Text is stored in the standard composed form (NFC). Where Unicode has a ready-made letter (ș ṭ ä á) that letter is used. Elsewhere the letter and its mark stay a pair (t̖ k̦ a̋). Both count as the same text.
The Diafon fonts (Serif, Sans and Mono) draw the marks as designed, with the anchor joined to its letter. Other fonts draw the same characters roughly in place, and the font picker on the Reference page shows how any font does.
All the sounds¶
Every sound of the system in one chart. Consonants run from the lips to the throat. Vowels sit as areas in the vowel triangle, with their splits. Each tile shows the letter, what you type and the sound, coloured by the operator key that makes it.
| labial | dental | alveolar | postalveolar | palatal | velar | uvular | glottal | |
|---|---|---|---|---|---|---|---|---|
| stop | pp/p/ bb/b/ | tt/t/ dd/d/ | țtj/c/ d̦dj/ɟ/ | kk/k/ gg/g/ | qkq/q/ | ḥhx/ʔ/ | ||
| affricate | ctc/t͡s/ xdc/d͡z/ | c̦tq/t͡ʃ/ x̦dq/d͡ʒ/ | ||||||
| fricative | ff/f/ vv/v/ | ṭtx/θ/ ḍdx/ð/ | ss/s/ zz/z/ | șsj/ʃ/ z̦zj/ʒ/ | ḳkx/x/ ġgx/ɣ/ | hh/h/ | ||
| nasal | mm/m/ | nn/n/ | n̦nj/ɲ/ | ṇnx/ŋ/ | ||||
| trill · tap | rr/r/ ṛrc/ɾ/ | |||||||
| lateral | ll/l/ |
input · /reading/colour: the operator that makes it, as on the keyboarda pair in one cell: voiceless | voiced, unrounded | roundedconsonants: lips to throat, the place marked in each column's profilevowels: front left, close at the top; a chip covers an area, /all it covers/; the small dashed chip is its split, written only by the languages named under itdashed: a composition shown for orientation (ț d̦ n̦)Glossary¶
| Term | Meaning |
|---|---|
| base | a letter that stands for a sound: 5 vowels and 15 consonants |
| operator | c j q w x y, typed after a base to change its sound |
| anchor | the comma-like mark joined below a consonant, typed j |
| reverse anchor | the anchor turned the other way, typed y: a language's own variant |
| float rule | on g and p the mark goes above |
| area | the range of sounds one letter covers |
| split | a second vowel inside an area, written only by languages that have both |
Diafon is named after the diaphoneme: one unit for matching sounds across languages, said a little differently in each.