Since there are languages with consonant clusters and languages with aspirated consonants, in principle there could be a language that has a surface contrast between [Ch] and [Cʰ]. Word-internally it might be possible to treat any such surface distinction as distinctive syllabification, i.e. [VC.hV] versus [V.ChV], but a surface contrast between #thV and #tʰV, or Vth# and Vtʰ# would provide pretty reasonable prima facia evidence for a contrast between a sequence versus a unit phoneme. I don't know of any language that has such a contrast – is such a language attested?

  • Is there some special quality about aspiration that is important here? Clusters of various sorts fall together with complex phonemes all the time in all kinds of languages.
    – jlawler
    Commented Nov 28, 2020 at 22:58
  • 1
    Long story, it's analogous to kw vs kʷ. I'm fairly sure I encountered a contrast in Lushootseed, where the cue was when the labialization starts. The ultimate question is approximately, "can one ever justify such a difference within a language based solely on pronunciation?"
    – user6726
    Commented Nov 29, 2020 at 0:21
  • Distinctive aspiration in Lushootseed? On what phones?
    – jlawler
    Commented Nov 29, 2020 at 2:30
  • No, rounding... analogous. kw:kʷ::kh:kʰ. I have the kw/kʷ part filled in, Im'm looking for the aspiration analogs.
    – user6726
    Commented Nov 29, 2020 at 2:40
  • Lushootseed individual variation is pretty frequent between front and back velars; there are /ts/ and /c/ which are distinguishable morphological but probly not phonologically, now that there are no more native speakers.
    – jlawler
    Commented Nov 29, 2020 at 15:59

2 Answers 2


It has been proposed that this contrast (aspiration vs. cluster consonant + [h]) does not exist (Kehrein & Golston 2004). The formal explanation is that aspiration is a property of the onset, not of individual segments. So, in an onset aspiration can occur once and it may be realized on the stop or as an [h] (or variably as both), but a contrast is not possible.

  • codas can also be aspirated. I'm not sure how this analysis handles cases of a coda immediately preceding an onset (i.e. whether it's onsets & codas that have the property of aspiration, or clusters, which may consist of both coda and onset)
    – Tristan
    Commented Nov 30, 2020 at 14:09
  • Phonologically, it's prosodic categories (onset, nucleus, coda) that have aspiration, but phonetically assimilation between these units may take place, so that they agree for aspiration on the surface
    – JMW64
    Commented Dec 2, 2020 at 4:48

If you allow for clusters across syllable boundaries, English could be an example

what hand = /wɒt.hænd/ what and = /wɒtʰ.ænd/

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.