Talk:Tone letter

It makes the comparison to a musical staff, does that mean that the difference between the top tone and the bottom tone in a tonal language is a major 5th?


 * No, it depends on the speaker's voice, intonation, emotion, etc. There are five slots because that's just about the max in human languages. kwami (talk) 02:10, 29 August 2008 (UTC)

list of all unicode tone letters
Is there a list of all the tone letters that Unicode is able to display on most people's computers?--Sonjaaa (talk) 02:53, 3 January 2009 (UTC)


 * Depends on what you mean by "most people's". If you have installed a font with full tone-letter support, like SIL Charis, then all of them. kwami (talk) 03:05, 3 January 2009 (UTC)


 * It also depends on what you mean by "letter" I recently learned that tone letters which show contours are actually ligatures which are created by the font, not unique unicode characters. So I think all tone characters do show on people's computers, just not all ligatured combinations show unless the computer user is viewing at the tone character combination via a font which creates the ligatured combinations. On OS X the default font for the OS does not support tone letter ligatures. Therefore sequential tone letters appear as sequences rather than as angles. Hugh Paterson III (talk) 16:35, 9 January 2015 (UTC)

Chao tone numbers
I noticed that there seem to be no pages that clearly explain the correspondence between tone letters and Chao tone numbers (which are very commonly used in Chinese linguistics, including many pages on Wikipedia), so I've added a table. I've included a line for the Africanist usage of tone numbers (i.e., opposite of Chinese usage), but I know virtually nothing about African linguistics (e.g., whether Africanists use double numbers to describe tones), so please feel free to correct me if I'm wrong. If necessary, we could split the table so that it shows only Chinese usage. Talu42 (talk) 05:12, 25 March 2013 (UTC)


 * No, Africanists do not use double numbers, which Sinologists do because of the need to differentiate arbitrary tone numbers (e.g. Wade-Giles pin4 etc.).
 * Not everyone uses a 1–5 scale, however. 1–3 is common in Mesoamerica, for example – same polarity as in Asia, but with 3 = high. I've also seen 1–4 and in the case of Omotic 1–6, and there is no good way to convert either to Chao. Sometimes 0 is not a neutral tone but the level below 1: for example, the end-point of falling tones that drop below the level of low tone. — kwami (talk) 22:24, 25 March 2013 (UTC)


 * Do you have any citations for numeric notation for 0-5 or 1-6? I am simply looking for examples. --Hugh Paterson III (talk) 23:37, 7 January 2015 (UTC)Hugh Paterson III

Mac rendering of contour tone ligatures


More details at Talk:Tone_(linguistics). —Pengo 11:39, 8 May 2013 (UTC)

Citations Needed
Does anyone mind if I add citations? I see that one is called for on the Omotic languages of Ethiopia. I could add it, but am not a regular wikipedia contributor and am unsure what this page's community feels about edits without prior discussion.

Languages claimed to have six level tones: ISO 639-3 code: [cry] ISO 639-3 code: [bcq] Hugh Paterson III (talk) 17:14, 9 January 2015 (UTC)
 * Dihoff, Ivan. 1976. Aspects of the tonal structure of Chori. Doctoral dissertation, University of Wisconsin.
 * Wedekind, Klaus. 1983. A Six-Tone Language in Ethiopia: Tonal analysis of Benč⁴ non⁴ (Gimira). Journal of Ethiopian Studies vol. 16: 129-56.


 * What are we supposed to do if there is a citation with the wrong details? Chao's original article appears in vol.30 not vol.45. 71.193.151.118 (talk) 03:37, 15 December 2018 (UTC)

Chao tone letters and the Unicode Standard
We occasionally use sequences of two identical Chao tone letters to denote level tones, here and on Tone (linguistics), cf. (and a few others) by user:Kwamikagami. I believe this is a mistake. The current Unicode Standard stipulates on page 323: "To represent contour tones, the tone letters are used in combinations. The rendering of contour tones follows a regular set of ligation rules that results in a graphic image of the contour (see Figure 7-8)." There is no mention of combining tone letters to represent level tones, so fonts and other software are not expected to render sequences of identical tone letters as ligatures. We should simplify those sequences to single letters. Love —LiliCharlie (talk) 05:16, 8 June 2020 (UTC)


 * I used to think it was a mistake too. But Unicode is not a RS, and level tones are contour tones. With a register system, you don't characterize tones as level, you only say that when contrasting a level contour with a moving contour.


 * I find the double tone letters for level tones to be annoying, and don't use that convention myself. I actually went through WP with AWB a couple years ago and replaced all the instances I could find with single tone letters. But, to my surprise, I recently came across a description of Chao's original system, at least as adopted by the IPA, and there there is a distinction between single and double tone letters. Single letters are apparently intended to be used for tones in a register system and for checked tones in a contour system, and double letters for unchecked level tones in a contour system, parallel to the conventional difference between /3/ and /33/ when using digits. So, while I don't like it, it would be POV to purge it from WP. — kwami (talk) 07:59, 8 June 2020 (UTC)


 * So what are ligatures of identical Chao letters supposed to look like? The same as single tone letters, to insure that (=double ) and  are the same width? For in the IPA Handbook's illustration of Chao letters for contour tone languages (p. 24) the letters for level, falling and rising tones are all the same width no matter if the syllable (in Cantonese) is checked or not. — My browser's IPA font (a derivative of Charis SIL) renders double  as a ligature twice the width of the  ligature, while my default browser fonts (Noto Sans; Noto Serif; Noto Sans Mono) display two separate letters. All these fonts show the expected ligature with a rising mid-to-top stroke for . To be honest: I think Google's Noto font series conforms to the Unicode Standard as well as to the Handbook of the IPA, and we don't. Unless you can point me to an IPA document that is as official as the Handbook and says otherwise.
 * P.S.: Most authors prefer diacritic accents instead of Chao letters for register tone languages, but this is of course not a requirement. Love —LiliCharlie (talk) 10:18, 8 June 2020 (UTC)
 * P.P.S.: If you need to show something that is not unambiguously defined in the Unicode Standard you cannot use Wikimedia's UTF-8 encoding of Unicode characters, but will have to use graphics. Love —LiliCharlie (talk) 11:27, 8 June 2020 (UTC)
 * Of course I can. There's almost nothing in Unicode that's unambiguous. They don't even understand basic typographic concepts such as what a digraph is, and they're full of duplicate, misidentified and badly defined characters. If fonts are badly designed, then readers need better fonts, but that's not something we can control. The new Liberation series is full of errors, but that only means they need to fix it, not that we have to avoid those characters. — kwami (talk) 11:42, 8 June 2020 (UTC)

Yes, they were double-width in the IPA adoption as well. I thought I saw that in a JIPA article, but I haven't tracked it down yet. But I did just come across this in the Report on the 1989 Kiel Convention, JIPA 19.2, p. 76:

"Tone letters" (following Chao 1933): these marks are to be placed before or after the segmental material. The report of the suprasegmentals groups also suggested that the Chao tone marks may be optionally attached to a vertical reference line, in which case narrow phonetic marks should precede the line and broad/phonological marks follow the reference line.

By the 'vertical reference line' they mean the the stem/staff that's now universal in fonts. I didn't realize that was optional! So it wasn't the Chao tone letters that were introduced in 1989, but only the option of adding a staff to them. And I'd never heard of the convention of left-facing for narrow and right-facing for broad/phonemic. I've only seen that for sandhi.

BTW, they also show a sequence of four grave/acute diacritics for a peaking-dipping tone, followed by "etc.", so evidently we're no more restricted to the combinations of tone diacritics on the chart than we are by the combinations of tone letters. — kwami (talk) 11:33, 8 June 2020 (UTC)


 * If you don't know if double   is expected to look the same as single, or should display as double width , or as two unligated letters you will have to resort to graphics, and perhaps propose a change to the Unicode Standard so that the graphics become superfluous one day. And if you do know what double   is expected to look like show me your source so I can request a change to the Noto fonts. Love —LiliCharlie (talk) 12:02, 8 June 2020 (UTC)
 * P.S.: I'm in favour of a Unicode proposal that uses combinations of the existing modifier letters rather than newly encoded characters. But such a decision requires some discussion among the Unicode community, doesn't it? Love —LiliCharlie (talk) 12:36, 8 June 2020 (UTC)
 * I see three options for the proposal: 1. To stipulate that combinations of two identical tone letters are as wide as combinations of two different tone letters, in which case a wider version requires a combination of three identical tone letters. 2. To stipulate that combinations of two identical tone letters are wider than combinations of two different tone letters. 3. To encode new characters. Love —LiliCharlie (talk) 13:18, 8 June 2020 (UTC)


 * Noto support of IPA is rather limited. I'm not familiar with the history of the convention, but presumably if someone takes the trouble to double a tone letter, they expect it to look different than a single letter does. — kwami (talk) 01:16, 10 April 2021 (UTC)

Numerical tones instead of tone sticks should be used for East and SE Asian languages
insists on using tone sticks instead of numerical tones to transcribe SE Asian languages, but this is currently not recommended by linguists working in the area. It is nearly equivalent to turning all IAST transcriptions into IPA in articles about Indo-Aryan languages. Please see the discussion at Wikipedia talk:WikiProject Languages. Lingnanhua (talk) 23:48, 9 April 2021 (UTC)


 * Local conventions are acceptable as an alt, but they should never be the default. Wikipedia is a global resource, and should use global conventions. I also go around converting imperial units to metric and AH dates to CE. Part of the problem is that your tone digits are backwards for many conventions, which can be really confusing for people outside of your walled garden. — kwami (talk)