I'm getting started with creating a language and I run into trouble when I started using rules I devised for protolanguage (e.g. only open syllables).

However I run into trouble when I tried to apply it to string [epse] which I'd syllabilized as [ep.se] rather than [e.pse]. I'm assuming I did something wrong with sonority hierarchy but I failed to find resources about what exactly.

EDIT: I would prefer generic rules for syllabication but for the Y problem - the rules I set so far is (C)CV syllable where:

  • If there are two consonants exactly one of them is a fricative (s, x or ʂ)
  • Fricative must follow plosives, nasals and trills but must precede lateral approximants (l)
  • 1
    It would help if you could give us the details of your rules, and why you think they are giving you a bad result.
    – curiousdannii
    Commented Jul 8, 2019 at 4:04
  • @curiousdannii I edited the question. Commented Jul 8, 2019 at 7:13

3 Answers 3


I think as a general rule, a syllable comprises some combination of the following kinds of one or more of the sound types of your language: vowel, liquid, consonant.

You've decided on (C)CV. And if there are two C then one must be a sibilant and the sibilant must follow a stop. So you can have syllables like sa, ta, tsa but apparently not sta. You might then have a words like tasa & satsa & tsatsa.

In English, we'd probably break that into SAT-SA. Your language will likely differ, because everyone knows that a syllable can't end with a consonant! So, they'd much more likely break the word thus: SA-TSA!

I would argue that what little you've described of your language and its rules speaks in favour of E-PSE rather than EP-SE. The only thing you "did wrong" was to apply the English rule rather than the rules that apply to your invented language! The second thing you did wrong was to second guess yourself based on the English rule!


You are the conlang designer, and you are in power to decide that epse is syllabified as e.pse by design of the language. This is not as unnatural as it might look at first sight for a native speaker of English, in Classical Greek, the letter Ψ (psi) can occur word-initially and is always considered a unit for syllabification.

  • In Greek prosody, psi is always treated as two consonants and makes the preceding syllable heavy, even when it occurs word-initially
    – b a
    Commented Jul 8, 2019 at 13:48
  • 1
    So the rules of syllabification are 'arbitrary and my division into ep.se was influenced by my language/accent rather than something inherited in this division? Commented Jul 8, 2019 at 17:17
  • At least, the rules of syllabification allow for some leeway that a language designer can use, I won't go so far to say that they are completely arbitrary.
    – Sir Cornflakes
    Commented Jul 8, 2019 at 18:00

Syllables are inherent to the spoken language and a native of that language will take a certain grouping of letters and turn it into a set of syllables according to that language’s rules.

In German, hyphenation generally follows syllables boundaries. About a decade ago, an article in a newspaper talked about the Chinese city of Shenyang but hyphenated it Sheny-ang. From a German point of view that could have been an acceptable syllable grouping if y is considered a vowel as ya (or ia) is not an acceptable diphthong, but Shen-yang (y as a half-vowel) would also have worked. For Chinese, only Shen-yang is possible because of the very restricted set of syllables. In Japanese, Shen-yan(g) or She-nyan(g) (strictly speaking, Japanese doesn’t have an ng sound unless k or g follow) would be possible with a y in the spelling; She-ni-an(g) or Shen-i-an(g) would strictly require an i in there.

As you maybe see, what will be seen depends on the language, and only what is possible will be produced. Another example is Japanese again, which only has open syllables or those ending in n. So when a Japanese tries to pronounce startling, that tends to end up somewhere near sutartolingu, a five-syllable (seven-mora) word. Thus, a speaker of your conlang will not even consider ep-se a possibility but will default to e-pse, if that is all the language does.

For the record, some languages with loose syllable rules still tend to end up with ambiguity. For example, the Finnish place Yli-Ii will have to have a hyphen to signify the short-long sequence. The German word Bakterien has an unusual syllable boundary between i and e (Bak-te-ri-en) while ie usually just signifies a long /i/ sound. Those things will need to be learnt at school, there is no way to predict.

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.