As far as I know, most writing systems for tonal languages fall into one of four groups:

  • The writing system is not phonetic (e.g. Han logograms)
  • Tone is not generally indicated in writing (e.g. many African languages)
  • Tone is indicated with diacritics on the vowels (e.g. pinyin, Vietnamese)
  • Tone is indicated by a glyph after the syllable (e.g. Hmoob RPA, IPA tone letters)

Are there any other options in active use? For example, do any languages have distinct graphemes for /á/ versus /à/?

  • Depends how strictly you define tone, but many languages use o where others might use à. Alemannic, Yiddish, Tajik. Alemannic and to some degree Standard German contrasts ä and e. May 26, 2020 at 20:52
  • 1
    Would you consider (Lhasa) Tibetan and Panjabi (in either Gurmukhi or Shahmukhi), where the use of tonal "letters" for consonants is actually a reflex of a lost voicing contrast? Also Thai, which is mixed between this tonal consonant system and a vowel (syllable?) diacritic system?
    – Michaelyus
    May 26, 2020 at 21:01
  • @Michaelyus I'm very far from an expert on Tibetan, but I think that would count! If I understand right, /tá/ and /tà/ would be written with different syllabic glyphs?
    – Draconis
    May 26, 2020 at 21:20
  • 4
    @AdamBittlingmayer I know very little about most of those languages, but I didn't think Yiddish or Standard German had any tonal features?
    – Draconis
    May 26, 2020 at 21:21
  • @Draconis. I share your bemusement.
    – fdb
    May 26, 2020 at 21:22

2 Answers 2


Tone Marking using Distinct Consonants

Orthographies that employ this form of tone marking are generally used in languages that developed tones from (usually voiced) consonants during their documented written history. Hence these tone-marking consonants were usually markers of voicing, and so it is debatable whether this truly counts as dedicated tone marking.

One good example is Lhasa Tibetan, which is written in an abugida generally held to have been developed in the 7th century CE. Modern Lhasa Tibetan is tonal, although the number of tones is subject to varying analyses. It is certain that there are at least 2, high and low; most analyses also split these, forming 4: high level, high falling, low level/slow-rising, low rising/rise-fall. Some even argue for 6, based on the slightly different contours of checked syllables from non-checked short vowels.

The following example shows how high falling vs low rising/rise-fall contrast with each other:

  • ཕོད (Wylie: phod; IPA for Lhasa: /pʰøː˥˩/) = to dare
  • འཕོད (Wylie: 'phod; Lhasa IPA: /pʰøː˥˩/) = completion
  • བོད (Wylie: bod; IPA for Lhasa: /pʰøː˩˧˨/) = Tibet
  • པོད (Wylie: pod; IPA for Lhasa: /pøː˥˩/) = book, volume
  • འབོད (Wylie: 'bod; Lhasa IPA: /pøː˩˧˨/) = to call out, to shout

As can be seen from the Wylie transliteration, in Classical Tibetan the difference in the script encoded a three-way distinction between voiceless plain /p/ , voiceless aspirated /pʰ/, and voiced /b/.

With the diachronic devoicing that happened in most of medieval East and Southeast Asia, affecting specifically Central Tibetan including Lhasa, voiced consonants became voiceless aspirated + low tone, viz = /pʰ/ + a low tone, creating the minimal pair vs in the aspirated labial consonants. The same applies to the velar, palatal affricate, alveolar affricate, palatal/postalveolar fricative and alveolar fricative series.

Interestingly, this also extended to the unaspirated versions, because of the prefix consonant: the "a chung" , whose phonetic realisation in Classical Tibetan is still under debate (either a voiced /ɦ/ or a nasal). It appears in the script either as the root consonant (directly bearing the vowel diacritic) or as a prefix to voiced or voiceless aspirated consonants, not voiceless plain consonants. In prefixed clusters with orthographic voiced consonants, this creates the contrast vs འབ. Note that the a chung prefix does not form a contrast with orthographic aspirated consonants. Certain other consonants in prefix position like or also behave in this way.

  • སྤོད (Wylie: spod, IPA: /pøː˥˩/) - spice/flavouring
  • རྦོད (Wylie: rbod, IPA: /pøː˩˧˨/) - incite

The "secondary level" contrast between contours is much rarer, but this is still an example where the orthographic information is coded by a consonant, this time a final consonant. The quintessential example is:

  • ཁམ (Wylie: kham, IPA for Lhasa: /kʰam˥/) [high level] - piece
  • ཁམས (Wylie: khams, IPA for Lhasa: /kʰam˥˩/) [high falling] - Kham region of Tibet

These are examples of the few minimal pairs, ending in /m/ or /ŋ/, where short syllables are signalled with a final orthographic consonant after the sonorant.

Both the overall pitch level and the contour are encoded in orthographic consonant contrasts, that were likely originally pronounced as such in Classical Tibetan. Hence, the tone is a reflex of both lost voicing distinctions and a lost final segment. In some other Tibetan varieties, e.g. Amdo Tibetan, phonemic tones have not emerged and voicing is preserved in initial position.

A major example of a mixed use of consonants and tone diacritics to show tone is Thai. This abugida is well known for its rules for working out which orthographic elements map to the five tones of Central Thai. The 13th century Thai inscriptions do have tone markers as diacritics, but that was Old Thai, and with the various changes through time, modern Central Thai tone marking relies on three elements: 1) initial consonant; 2) tone diacritic; 3) final consonant, plus the mapping procedure.

Tone Marking Using Distinct Syllable Glyphs

The one that comes to my mind is Northern Yi (Nuosu) in the Modern Yi script, a syllabary standardised in 1975 although it dates back to the Tang dynasty. The language can be written in Roman script, with the post-syllable Roman glyphs -t, -x, null, or -p (generally listed in that order), but the modernised version of the traditional script is quite fascinating.

The -x tone (apparently mid falling or high-mid level in Liangshan Yi) is derived from the null (which is mid level), which reflects its marginal status in the language. However -t (high level tone) and -p (low falling tone) orthographic series do not necessarily have obvious parallels with the null series; although occasional resemblances can be gleaned for some finals (e.g. o with op) the majority are unrelated.

Tone Marking Using Positional Variants

This is a feature of the Pollard script, an abugida originally written for the Hmongic language of Northeastern Yunnan A-Hmao. The vowel itself is determined by the shape of the vowel diacritic, whilst its tone is determined by the position of the diacritic with respect to its base consonant. I see four tones are encoded by the script, although I haven't done much research into how the tones actually correspond.

  • Don't really like to add to such an erudite answer but FWIW I had the following thoughts about Thai: (1) About a third of words have tone diacritics so the system is probably best regarded as being based on the consonant characters (which reflect lost voicing contrasts) (2) it is arguable whether the final consonant is part of the system for marking tones - it's not as though you can change the final consonant character and get the same syllable with a different tone. The differences that are attributable to the final consonant can be said to belong to the tonal system...
    – rchivers
    May 27, 2020 at 9:21
  • ... rather that the writing system (3) either way, you have to include glottal stops in the definition of "consonant" (4) what goes for Thai also goes for Lao.
    – rchivers
    May 27, 2020 at 9:25
  • Are the first two Tibetan examples (‘dare’ and ‘completion’) meant to be transcribed identically? Or should they have differing tones/consonants? May 27, 2020 at 13:21
  • @JanusBahsJacquet འཕོད 'completion' is an example of a chung not affecting anything with the orthographic aspirated series, so it does not differ in pronunciation at all from ཕོད 'dare'... in isolation at least.
    – Michaelyus
    May 27, 2020 at 17:00
  • Hmong alphabetic writing uses "silent" final consonants as tone markers.
    – jlawler
    Jun 15, 2023 at 19:07

the Yi syllabary is a good place to look: https://en.wikipedia.org/wiki/Yi_script#Syllabary

  • The Yi syllabary is a perfect example of what I'm looking for! Could you elaborate a bit on how it works (maybe with examples of different glyphs for different tones), and how that works?
    – Draconis
    May 27, 2020 at 17:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.