I am building a conlang, which is very likely going to be an isolating language. As such, I decided to make it a tonal language.

But there is a problem. This language is to be sung very often, so using pitch as a suprasegmental would compromise understanding lyrics.

So instead of tones, I decided to pick another kind of suprasegmentals. Candidates in my mind were:

  • Phonation, e.g. Voiceless [ḁ] / Breathy [a̤] / Modal [a] / Creaky [a̰].

  • Nasalization, e.g. Oral [a] / Nasal [ã].

  • Tongue root position, e.g. Advanced [a̘] / Relaxed [a] / Retracted [a̙].

I took the third option. Yet I doubt any natural language does this. Is there any?

  • 3
    Could you explain what it means for you for tongue root position to be "suprasegmental"? Usually, this refers to a feature that applies on a larger scope than one phonemic segment. Tongue root position is often part of a vowel harmony system, which I would say fits the definition of a suprasegmental property. But I'm not sure what your viewpoint is on what makes something suprasegmental vs. non-suprasegmental Oct 15, 2021 at 1:56
  • @brasstacks By "suprasegmental", I consider only those making minimal pairs. Oct 15, 2021 at 1:59
  • I'm still confused. Say a language has a tongue root position contrast between the following two sets of harmonizing vowels: +ATR [i u e o ə] and -ATR [ɪ ʊ ɛ ɔ a]. For simplicity, let's say that vowel harmony applies without exception on the word level (in real languages, there are usually exceptions, making it a bit more complicated than that). Then there will be minimal pairs between +ATR words and -ATR words (i.e. between the TR values of different harmonic domains), but not between +ATR and -ATR vowels within the same harmonic domain (within a word). Oct 15, 2021 at 2:15
  • For example, in the model language described above, we could have a minimal pair like vi vs. vɪ, but not a minimal pair like ivi vs. ivɪ. However, ivi vs. ɪvɪ would be a possible contrast, and assuming you treat +ATR/-ATR as a single word-level feature, could be considered a minimal pair. (How) does that differ from the situation that you are thinking of? Oct 15, 2021 at 2:17
  • @brasstacks One would distinguish [i] vs. [ɪ] by dorsal position, not laryngeal position. My language would make +ATR/0ATR/-ATR distinction by actively using laryngeal muscles to extreme. My language has 6 plain vowels, a/e/i/o/u/y. By adding the suprasegmental, my language has 6×3=18 vowels in total. Oct 15, 2021 at 2:24

1 Answer 1


Insofar as you've put creaky and breathy voice in one bin, and a three-way distinction in "ATR" in a second, you have described a situation that doesn't exist in any known language. There are languages with a two-way contrast which might be described as normal vs. Retracted Tongue Root (certain Tungusic languages), or normal vs. Advanced Tongue Root (various Nilotic languages), but no language has been found with ATR, RTR and Ø. In general, languages with ATR do not also exploit phonatory contrasts, instead phonatory differences are often a predictable concomitant property of one or both of the extant ATR contrasts. It would be difficult to argue that a putative example of a three-way contrast was indeed ATR/Ø/RTR and not ATR/Ø/creaky, for example.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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