By what principles of phonology does the digraph < ng > in < sing > admit a monophonematical representation by the phoneme /ŋ/, while in the word < sink >, the final < k > is represented by its own phoneme /k/? Both /g/ and /k/ commute with 0, because /sɪn/ is a proper word, so I find more than dubious that these two words admit interpretations of 3 and 4 phonemes, respectively. Moreover, we have /sɪŋər/ for the person, but /sɪŋgl/ or /sɪŋgəl/ for < single > (song) which is even more mysterious...

  • Do you accept that /ŋ/ and /n/ are articulated at different places? How do you think "sink" and "sing" should be transcribed? You seem very adamant that the conventional wisdom is wrong but you offer no evidence, and it's unclear if your objection is phonetic or orthographic; you cannot infer pronunciation a priori (and simply saying you find something dubious is argument from incredulity, which is a logical fallacy).
    – Stuart F
    Commented Nov 13, 2023 at 23:48

1 Answer 1


The phonemic status of [ŋ] in English is debated. The simplest argument for it being a phoneme is the existence of the following pairs:

  • singer vs. sinner (establishes a contrast between [ŋ] and [n])

  • lo[ŋɡ]er "more long" vs. lo[ŋ]er "one who longs" (establishes a contrast between [ŋɡ] and [ŋ])

Those pairs involve [ŋ] in a different (intervocalic) context, but if we have established that si[ŋ]er and lo[ŋ]er contain a phoneme /ŋ/ in intervocalic position, it's pretty straightforward to then argue that the related monosyllables sing and long end with /ŋ/ in word-final position.

Arguments arise because of the generally low level of contrast between [ŋ] and [ŋɡ] and the fact that the difference is mostly predictable when word structure is taken into account. Some phonological traditions see it as convincing to analyze [ŋ] as still underlyingly a cluster /ŋɡ/ in modern English words like sing, long, singer: this reduces the number of phonemes and allows us to explain some cases where [ŋɡ] and [ŋ] alternate such as long, lo[ŋɡ]er (adj.), at the cost of requiring us to suppose that the English sound system includes rules that are sensitive to word structure, like "/ŋɡ/ is pronounced [ŋ] before a morpheme boundary, except for when followed by the comparative or superlative suffixes -er, -est". It is debated whether rules like that really exist, or whether English speakers just memorize the small number of words that show an alternation.

More discussion on Reddit: Is /ŋ/ a phoneme in English?

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.