Whilst worldbuilding, I decided to make a conlang. A lot of consonant assimilations happen in this language. This language is supposed to be spoken by angels.

My native language is Korean, btw. Korean also has many consonant assimilations.

Orthography and phonetics

My conlang uses Latin alphabets except W. No diacritics. All of them is pronounced as their IPA counterpart, except four:

  • C = [qʼ] (Voiceless uvular ejective)

  • N = [ɴ] (Voiced uvular nasal stop)

  • Q = [ʔ] (Glottal stop)

  • X = [ǃ] (Voiceless alveolar click)

Sounds exotic, huh? That makes 19 consonants and 6 vowels.

Plosives are tenuis by default.

Consonant assimilation

In my conlang, syllables have structure of (C)V(C). So two syllables can make two adjacent consonants. Every combination of two consonants can occur. That's where consonant assimilation happens, and also where this question arises. Let's call CC "former consonant" and "latter consonant", respectively:

  • Plosives as the former consonant get no release. Examples: "BK" to [b̚.k], "KB" to [k̚.b].

  • "H" as the former consonant usually assimilates to [x].

  • "V" as the former consonant usually assimilates to [ʋ].

  • "H" as the latter consonant...:

    • ...turns voiced plosives to implosives. Example: "BH" to [ɓ].

    • ...makes fricatives pharyngealized. Example: "FH" to [fˁ].

    • ...makes nasal stops voiceless. Example: "MH" to [m̥].

    • ...makes voiceless plosives aspirated. Example: "KH" to [kʰ].

  • "J" as the latter consonant usually makes the former consonant palatal or palatalized. Examples: "MJ" to [mʲ], "SJ" to [ɕ].

  • A plosive followed by "L" has lateral release.

  • "V" as the latter consonant usually makes the former consonant labialized. Example: "XV" to [ǃʷ].

  • "CH", "DJ", "DZ", "PF", "TJ", "TS" assimilate to affricates.

To show all combinations:

Consonant assimilations

I set these consonant assimilations to make pronunciation easier, but my concern is that these introduce too many new phones that lacks their own letters. Are they tolerable?

  • 1
    Why shouldn't they be? Consider that there are many natural languages whose orthography does not accurately reflect their ... phonography? phonology? (not sure what word is correct here - "It's not one sound per letter".) Commented Dec 5, 2020 at 14:54
  • This doesn't even rise to the level of "digraphs", this is straight up allophony! Unless your angels pay very close attention to the sounds their mouth parts produce, I doubt they'd even notice the difference.
    – No Name
    Commented Mar 27, 2023 at 8:41

1 Answer 1


I would argue that the orthography of your conlang is not just OK but really excellent given your description of the phonology.

Why is this the case? It keeps the number of letters needed to write the language in a very comfortable range. The additional sounds are given by digraphs, a method that is often preferable over adding more letters to the basic alphabet or adding diacritics to the base letters.

I also imagine that the orthography keeps morphemes recognisable assuming that at least some morphemes are just CVC syllables and combining them results in assimilation.

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.