A braille version of the IPA was first created by Merrick and Potthoff in 1934, and published in London. It was used in France, Germany, and anglophone countries. However, it was not updated as the IPA evolved, and by 1989 had become obsolete. In 1990 it was officially reissued by BAUK, but in a corrupted form that made it largely unworkable. In 1997 BANA created a completely new system for the United States and Canada. However, it was incompatible with braille IPA elsewhere in the world and in addition proved to be cumbersome and often inadequate. In 2008 Robert Englebretson revised the Merrick and Potthoff notation and by 2011 this had been accepted by BANA.[a] It is largely true to the original in consonants and vowels, though the diacritics were completely reworked, as necessitated by the major revisions in print IPA diacritics since 1934. The diacritics were also made more systematic, and follow rather than precede the base letters. However, it has no general procedure for marking tone, and not all diacritics can be written.
IPA Braille does not use the conventions of English Braille. It is set off by slash or square brackets, which indicate that the intervening material is IPA rather than national orthography. Thus brackets are required in braille even when not used in print.
Following a pattern found in many national alphabets, a few braille letters are reversed to represent a similar letter: thus reversed ⟨n, s, z⟩ produces ⟨ŋ, ʃ, ʒ⟩. The choice for ⟨ɹ⟩ may reflect the shape of that letter in print. Many of the vowels are used for modified vowels in national alphabets, such as French Braille.
A few other letters such as ⠹ occur, but only as parts of digraphs.
Other IPA letters are indicated with digraphs or even trigraphs usinɡ 5th-decade letters (letters from the punctuation row). The component letter ⠲ ".", for example, is equivalent to the tail of the retroflex consonants. This presumably derives from the old IPA practice of using a subscript dot for retroflex consonants. It also marks vowels which in print are formed by rotating the letter.[b] is treated as a rotated (script?) ⟨o⟩, and ⟨ɯ⟩ as a rotated ⟨u⟩ rather than ⟨m⟩, perhaps facilitated by braille ⟨u⟩ and ⟨m⟩ themselves being a rotated pair.
⠨ is used for fricatives written with Greek letters, using the English conventions for Greek letters in scientific notation. The basic braille letters ⠹ and ⠯, which do not occur on their own in IPA usage, appear here.
⠨ is also used with letters of the fifth decade for transcriber-defined symbols, which need to be specified for each text, as they have no set meaning. These are ⠨⠂, ⠨⠆, ⠨⠒, ⠨⠲, ⠨⠢, ⠨⠖, ⠨⠶, ⠨⠦, ⠨⠔, ⠨⠴.
Ligatures, regardless of whether these are written with a tie bar or as actual ligatures in print, are indicated by dot 5, so ⟨t͜ʃ⟩ and ⟨ʧ⟩ are both ⠞⠐⠱. This includes the historic ligatures ⟨ɮ⟩ ⠇⠐⠮ and ⟨ɚ⟩ ⠢⠐⠗. Ejectives are written as ligatures with an apostrophe, ⠄, so ⟨tʼ⟩ is ⠞⠐⠄.
IPA Braille diacritics are written in two cells. The first indicates the position: whether superscript, mid-line, or subscript. ⠻, for example, is the tilde. ⠁⠈⠻ is therefore nasal ⟨ã⟩, ⠁⠠⠻ is creaky-voiced ⟨a̰⟩, and ⠎⠐⠻ is pharyngealized ⟨s̴⟩. As noted above, dot 5 is used for the tie bar; the only diacritics which use it are the rhotic and ejective (apostrophe) diacritics seen above, and the velar/pharyngeal tilde just illustrated.
In other cases, such as the ring for voiceless ⟨l̥ ŋ̊⟩, a diacritic may superscript or subscript with no change in meaning. The distinction is therefore not necessary in braille, but can be maintained if the text is to be transliterated into print. The wedge, which indicates voicing when below a letter, can be placed above to represent Americanist notation such as ⟨š⟩ and ⟨č⟩. The placement dot distinguishes the tilde from the letter ⟨ð⟩, and the ring from the letter ⟨ŋ⟩.[e] Many diacritics are identical in form to the corresponding symbols in Unified English Braille, but always follow the letter in IPA Braille. When there is more than one diacritic, they are written in the order lowest to highest.
Superscript letters are simply the superscript placement dot plus the base letter. So: ⠀⠈⠓ ⟨◌ʰ⟩, ⠀⠈⠚ ⟨◌ʲ⟩, ⠀⠈⠇ ⟨◌ˡ⟩, ⠀⠈⠝ ⟨◌ⁿ⟩, ⠀⠈⠺ ⟨◌ʷ⟩, ⠀⠈⠨⠛ ⟨◌ˠ⟩, ⠀⠈⠖⠆ ⟨◌ˁ⟩, etc., though not all superscript letters are possible.[f] The other diacritics are as follows; any of these may be modified by a change in placement:
There is no general provision for IPA tone letters. Ligatures which did not happen to be chosen as illustrations in the 2005 edition of the IPA chart, such as the Mandarin dipping tone ˨˩˦ and various other tones used in the IPA Handbook, are not covered. The high and low falling tone letters above are extrapolated from the rising tone letters. ⠸⠲ transcribes peaking (rising–falling) tones in general, and so has no exact print equivalent; the parallel dipping tone may be expected from the patterns elsewhere. Since other tones are written as ligatures in print, the ligature mark ⠐ would presumably be appropriate, so perhaps Mandarin ˨˩˦ could be written ⟨⠸⠤⠐⠠⠤⠐⠉⟩.
⠰ opts out of IPA. A single instance indicates that the following cell is to be read in the national orthography. Doubled, ⠰⠰, it indicates that the following string is to be read in orthography, up to the opt-in indicator ⠰⠆, which marks the end of the non-IPA passage. These are only used for non-IPA (or nonstandard IPA) strings within IPA brackets. An example is the use of parentheses to mark an optional consonant or vowel, since parentheses are not defined for IPA Braille.
^For this purpose, ⟨ɟ⟩ is treated as a small capital J
^It is therefore not possible to transcribe superscript ⟨ᶞ⟩ or ⟨ᵑ⟩ without some additional convention, such as the voiceless diacritic always being subscript, and the superscript indicating an engma in for example ⟨ᵑǃ⟩. Or braille click notation could substitute the proper tilde and the voicing diacritics for the ad hoc superscript engma and gee generally used in print, as the diacritics do not have the legibility problems in braille that they have in print.
^The IPA letters which duplicate diacritics and which therefore cannot be written as superscripts are ⟨ɑ æ c d ð e ə i ɪ ŋ ʊ x ː⟩. Of these, superscript ⟨d ə i ŋ x⟩ are fairly common in print. ⠈⠆ is similarly potentially ambiguous between superscript ⟨ʔ⟩ and the extra-IPA diacritic ⟨◌̍⟩ that is sometimes seen in transcriptions of languages with four- or six-tone systems. See the previous note for two suggested solutions for rendering superscript ⟨ŋ⟩ in braille.