Ok, see this word entertainment has IPA of /en.təˈteɪn.mənt/. Ok, now in American English if /t/ is between 2 vowel sounds then it will become /d/ cos it is flap T.

But /t/ will become flap T only if the sound in the word is not stressed. Is that correct?

For example, for the above example, we can pronounce /en.dəˈteɪn.mənt/ but not /en.dəˈdeɪn.mənt/, right?

  • 1
    In AmE we're more likely to drop the consonant altogether. Entertainment becomes /en.əɻˈteɪn.mənt/ – Robusto Mar 31 '15 at 1:34
  • 1
    @Robusto: but we only drop the first /t/ because that /t/ is after an /n/, and not between two vowel sounds. – Peter Shor Mar 31 '15 at 1:35
  • The rule is intermedial unstressed. so 'letter' -> /'le der/, but 'deter' -> /dij 'ter/ – Mitch Mar 31 '15 at 2:05
  • 2
    Actually, a t will flap before a stressed vowel, provided that the t is at the end of a syllable. When t and the following stressed vowel are within the same word, the t is never at the end of a syllable, so it never flaps. However, when the following stressed vowel is in a different word, then the t remains in the same syllable as the preceding vowel, and it will flap, provided other conditions on flapping are met. – Greg Lee Mar 31 '15 at 3:33
  • 1
    You’re leaving out the R. You shouldn’t do that. We don’t talk that way. – tchrist Mar 31 '15 at 5:52

t/d/n become flaps when they are (1) after a vowel or glide (including r but not l), (2) before a vowel, and (3) at the end of a syllable. (Condition (3) is stated with the assumption that an intervocalic consonant before an unstressed vowel goes at the end of the preceding syllable.)

"Entertainment" has a rather involved derivation. After nasalizing the preceding vowel, the first n is lost by a rule that deletes nasal consonants before voiceless consonants at the same place of articulation. The first t is now between vowels and at the end of a syllable, so it flaps. The flap assimilates in voice to the preceding and following voiced vowels, and being a sonorant consonant, it also assimilates in nasality to the preceding nasal vowel, so we wind up with a nasal voiced flap between the first two vowels of "entertainment". Nothing remarkable happens in the "tainment" part of the word, except [n] optionally assimilates in position to following [m].

This follows the phonological treatment worked out a long time ago by my teacher David Stampe. Note that there is never a d at any stage of the derivation, and there is never an intervocalic n, either.


  • BRAVO! The punchline at the bottom is so well done you should put a # in front of it so it gets all big and bold. :) – tchrist Mar 31 '15 at 3:22
  • 1
    Should the last /n/ not also be deleted before the homorganic /t/? I definitely don't have an actual [n] at the end before the unreleased [t̚] in normal speech. I presume the nasal deletion rule (apart from being optional) is more conditioned than you explain here—it doesn't apply in words like hamper or until, for example, but can apply in bonkers. – Janus Bahs Jacquet Mar 31 '15 at 9:55
  • @JanusBahsJacquet, yes, it seems the last [n] should be lost. But in my own speech, it isn't. I don't know why. (I wrote schwa, but maybe it's syllabic [n].) In until, I think the [n] being in a different syllable than the [t] makes the deletion less applicable. I can delete [m] in hamper -- I don't know why it's not lost for you. – Greg Lee Mar 31 '15 at 14:34

As an American, I would say I pronounce the word quite differently than your rendition:


With both "t" sounds. However, please note, neither is actually between two vowel sounds. A better example would be letter.


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