Weston Ruter

The symbols defined in the International Phonetic Alphabet are used extensively in Phonetics and Phonology, and it is therefore essential that the symbols be easily accessible and displayed. Before Unicode, this was difficult: linguists were forced to create custom ASCII fonts (such as SIL IPA93), use ASCII glyphs themselves (eg. SAMPA), or drop using text and use images of IPA symbols instead (see Graphical IPA Keypad). For a 1999 snapshot of these problems, see Matthew Brooks. With Unicode, which is now widely used and supported by browsers such as Firefox, Internet Explorer, and Opera, IPA symbols can be displayed efficiently and semantically. To showcase the capabilities of Unicode, I replicated the chart created by the International Phonetic Association, which appears in most of the Phonetics and Phonology texts. The chart is written in fully compliant XHTML 1.1 and CSS2, and is released under the GPL.

View the chart →

Additionally, I’ve also built an IPA “keyboard” off of the chart in order to facilitate copying the Unicode symbols and making transcriptions. It keeps track of recently used symbols and allows you to insert the symbols encoded in HTML entities. It works best in Firefox and Opera 9+, but also works in Internet Explorer 6 (but not in version 5).

View the “keyboard” →

Changelog

2008-06-12: Peter

Peter writes: “For the labiodental flap, your chart is using U+F25F, which is a private-use character.  According to Wikipedia, Unicode 5.1 defines U+2C71 for this.” Updated!

2007-10-22: Paul

Paul wrote in noticing that “[t]here are just a couple of clitches in the ‘Tone and Word Accents’ area. For example, the mid bar is linked to a lower bar in the pop-up. I think that the alt=”" values are off set by one value.” I have resolved this issue.

2006-03-13: Isaac C.

Per the suggestions of Kirk and another very insightful email from Isaac C. (see below), I have made the following changes:

• For the Pulmonic Consonant table, utilized the semantic caption element instead of a generic paragraph. Reverted for now.
• Added the new Charis SIL font as the preferred font; if it is not installed, the existing Doulos SIL and new Gentium are the next preferred fonts; I added a recommendation for these two new SIL fonts in the footer.
• Added the new IPA symbol, the labiodental flap: U+F25F “LATIN SMALL LETTER V WITH RIGHT HOOK” [1]
• Replaced the letter schwa with the rhotic hook diacritic with U+025A “LATIN SMALL LETTER SCHWA WITH HOOK”.
• Added contour tone letters.
• Added remaining contour tone diacritics, but they will remain commented out until Unicode 5.0.
• In keyboard mode, substituted level diacritics base character from U+2002 “EN SPACE” to U+25CC “DOTTED CIRCLE”

Additional changes I have made:

• Changed the keyboard frameset from scrolling=yes to auto; this removes the disabled horizontal scrollbar in Firefox and Opera.
• Changed the symbol for the Velarized or pharyngealized diacritic from the simple tilde to the combining.
• For the over- and under-tiebars, I added generic examples of both with U+25CC DOTTED CIRCLE; clicking this in keyboard mode will insert the tie bar by itself.
• Added PayPal donate link to IPA Chart footer.
• The keyboard now works in Opera, version 9 (and perhaps lower).

Issues remaining to be resolved:

• Tones and Word Accents: add more tone contour letters?
• There is too much whitespace.

This round of changes was prompted by a very insightful email I received from Isaac C. on March 12th, 2006:

Thanks for the IPA keyboard. I must first say that you’ve done a great job getting this application up. I especially like the arrangement of the content–it looks just like the original IPA chart itself. However, I found that the part about bottom ties not supported is untrue. If you look at the Unicode Standard 4.1, there is, at U+035C, a “COMBINING DOUBLE BREVE BELOW”, which essentially is the bottom tie bar you’re looking for. This is supported by Doulos SIL too. You might want to consider adding it in. (Technical note: If you use Character Map, Doulos SIL will show the bottom tie bar under the Private Use Area (PUA) at U+F176 and not U+035C. This is not a problem because within the font file both codes are mapped to the same glyph (dual-mapping).) I read your changelog and apparently Kirk already told you that “Your comment that the contour tone diacritics and tone letters are not supported by Unicode is misleading.” Doulos SIL indeed does support it, and I’ve included the following Unicode values you can use to display the characters:

• Rising: U+02E9 U+02E5 (˩˥)
• Falling: U+02E5 U+02E9 (˥˩)
• High Rising: U+02E7 U+02E5 (˦˥)
• Low Rising: U+02E9 U+02E7 (˩˨)
• High Falling: U+02E5 U+02E7 (˥˦)
• Low Falling: U+02E7 U+02E9 (˨˩)
• Peaking: U+02E5 U+02E4 U+02E5 (˧˦˧)
• Dipping: U+02E5 U+02E6 U+02E5 (˧˨˧)

There must not be any space between these tone characters or they will not display. Though Doulos SIL supports the diacritics for contour-tones, it is Unicode 4.1 unofficial (the codes are in the PUA). It is going to be in Unicode 5.0:

• U+1DC4 COMBINING MACRON-ACUTE
• U+1DC5 COMBINING GRAVE-MACRON
• U+1DC6 COMBINING MACRON-GRAVE
• U+1DC7 COMBINING ACUTE-MACRON
• U+1DC8 COMBINING GRAVE-ACUTE-GRAVE
• U+1DC9 COMBINING ACUTE-GRAVE-ACUTE

You might want to include these as a precursor to the upcoming new standard. Speaking of diacritics, you might want to use Unicode standard U+25CC “DOTTED CIRCLE” as the placeholder for the them instead of U+2002 “EN SPACE”. This I feel will bring more brevity to where the diacritics be applied unto another character. The rhotic schwa could be more accurately represented as U+025A “LATIN SMALL LETTER SCHWA WITH HOOK” rather than the letter schwa with the rhotic hook diacritic. I have some reservations on whether this is necessary because in the future such compound atomic characters might be obselete, as glyph mutation through different code sequences is already supported (quintessentially shown by the contour-tone case), just not so widely used. Perhaps fonts next time will automatically display this glyph when the schwa and rhotic hook are put together respectively? You might want to consider placing it on your keyboard anyway, since it is not done in Doulos SIL or Charis SIL. Have you considered adding the ExtIPA characters into your keyboard? I thought this might be a useful extension of what you’ve done now. I believe most of the characters required for ExtIPA exist in the Unicode specification now.

Thank you Isaac!

2005-12-11: Kirk M.

Kirk sent me a few comments for the chart. The corrections he suggested and what I did with them appear below:

1. Could you replace the <g> with its IPA equivalent? There’s no need of your chart just for traditional <g>.

[In an email conversation with Kirk] You are right, the IPA Handbook provides both on page 179. The ASCII <g> is referred to as “#210: Looptail G,” which Unicode classifies as “U+0067: LATIN SMALL LETTER G.” The one shown on the IPA chart <ɡ> is referred to as“#110: Opentail G,” which Unicode classifies as “U+0261: LATIN SMALL LETTER SCRIPT G.” Additionally, Unicode comments that this character U+0261 is a “voiced velar stop,” a comment which is not provided for U+0067. As you have mentioned, both the IPA Handbook and the IPA Unicode Test Page state that these two characters are equivalent:

IPA #110 (above) and IPA #210 (left) are equivalent “Voiced velar plosive” symbols. IPA #110′s symbol name is “Opentail G” while #210′s is “Looptail G”.

Being that these two character are equivalent, which one is preferable? Opentail G would seem to be preferred as it is the one which appears on the IPA chart. However, the traditional Looptail G is obviously supported by more fonts being that it has the benefit of being within the ASCII range. The IPA Palette decided it was necessary to replace “regular ‘latin small letter g’ (U+0067) with ‘latin small letter script g’ (U+0261).” You also give a good reason for this replacement in that “it would be nice to be able to get the other one using” the chart. I will go ahead and replace U+0067 with U+0261.

2. Your comment that the contour tone diacritics and tone letters are not supported by Unicode is misleading. They need to be rendered in the IPA font, just as most other sequences of diacritics. Doulos SIL has now made a start on this.

See above.

3. (Minor point.) Upstep and downstep are not supported by Unicode. What you have are ingressive and egressive airflow, used by the ExtIPA. Hopefully superscript arrows will be supported soon.

[action pending...]

Thank you Kirk!

2005-07-28: Sylvester

Added diacritics for rising (U+030C: Combining Caron) and falling (U+0302: Combining Circumflex Accent) tones; thanks to Sylvester.

2005-01-24: Mark D.

Corrected spelling of “plumonic” to “pulmonic”. Many thanks to Mark D. for bringing this to my attention.

External Resources

• Doulos SIL Unicode Font, recommended for my chart
• IPA in Unicode, John Wells’ technical reference
• IPA Charts, provided by the University of Victoria; includes excellent PDF charts and a Unicode chart, which LinguistList’s uses
• Unicode and Multilingual Support, Alan Wood’s resources