ex·act·ly iɡˈzak(t)lē/

If the t is in brackets, does it mean that I don't have to pronounce it? I'm asking because I'm a native English speaker and I never pronounce the "t" in "exactly".

  • 6
    It most likely means that it is optional. It's just a shorter way of writing two variant pronunciations.
    – curiousdannii
    Nov 4, 2016 at 13:53
  • Are you asking about the notation? Or are you asking why t appears to delete? You should clarify what your actual question is.
    – user6726
    Nov 4, 2016 at 16:17
  • I mean. If the t is in brackets, does it mean that I don't have to pronounce it? I'm asking because I'm a native English speaker and I never pronounce the "t" in "exactly".
    – James
    Nov 4, 2016 at 22:57
  • 1
    They failed to put brackets around (ig), but you also don't have to pronounce the first syllable. You pretty much don;t have to do anything.
    – user6726
    Nov 4, 2016 at 23:19

2 Answers 2


I managed to find the example you gave, /iɡˈzak(t)lē/, on both google dictionary (googling "define exactly") and the Merriam-Webster online dictionary. Google offers no explanation at all for its convention, and unforuntately Merriam-Webster's pronounciation key only explains the sounds used, not all aspects of the notation.

So it's hard to say exactly what was intended by the parentheses since the source is unwilling to comment. However, all is not lost! A lot tends to be borrowed from the conventions of IPA (just more "English-friendly" symbols are used when the subject matter is all English), and we can look at the IPA for "exactly" to know precisely what we're dealing with:


Here you can see the familiar parentheses around 't' appear again. In fact, all the punctuation appears to be the same (stress has been marked with an apostrophe before the stressed syllable and the word is encased in slashes) so that's a promising sign.

With this transcription we can say exactly what each part means, since IPA is very standardized. Parentheses in IPA do not indicate that a sound is optional per se, but rather that it's barely voiced. In practice the result is similar, as with anything there is some variance between each speaker and for some "barely" may become "not at all" without the word sounding noticeably different. In fact, if you lost the sound entirely and merely moved your mouth as if you were going to pronounce the t the IPA wouldn't change -- they use parentheses for silent mouthing when the transcription has been derived from lip-reading. The "ghost" of the sound, however, remains.

You yourself are a perfect example, as you say you never pronounce the 't' in exactly. I do, but subtly, and I certainly wouldn't notice if you said the word to me without it.

Given the prevalence of IPA, and the fact that it mirrors all the source's punctuation in its transcription (including the parentheses), it's most likely your original source meant the same thing by them.

You can find information on this and other "punctuation marks" in IPA under the transcription section on their Wikipedia page. It's an interesting read and very helpful for understanding levels of precision in various IPA transcriptions.

We needn't only compare to IPA, you can easily check the various other pronounciations available from free dictionaries online. It is, after all, just one word, and explanations which have the different sources agree with each other are most likely to be correct. None of them offer a pronunciation, optional or otherwise, in which the 't' is completely omitted.

  • Your answer might be relevant if there was any reason to think the transcription is in IPA, which it obviously isn't.
    – user6726
    Nov 5, 2016 at 4:57
  • 2
    Can you provide any evidence it means that the sound is barely pronounced, rather than being for two alternate pronunciations? There are other ways of writing barely pronounced sounds.
    – curiousdannii
    Nov 5, 2016 at 9:58
  • 3
    That was the purpose of my link to IPA conventions, and of checking how it's written in IPA. We can't magically know what some symbol is meant in any instance, but if we know how it's written in a very standardized format, and know what the parentheses mean in that very standardized format, then we know how the word is pronounced and what is done with the sound in the parentheses. There are other ways of writing two alternate pronunciations as well (AS two alternate pronunciations).
    – Emmabee
    Nov 5, 2016 at 14:14
  • Well, I think my answer may be ludicrously long now, but I've included as much evidence as I can on the sound not being alternately pronounced or unpronounced in two variant pronunciations. Hopefully this is satisfactory now.
    – Emmabee
    Nov 5, 2016 at 15:43
  • 1
    @Emmabee Long answers are more than welcome! :D Also, welcome to the site.
    – Alenanno
    Nov 7, 2016 at 16:15

The loss of t in "exactly" is a special case of a general rule that deletes an alveolar stop (t, d, n) in English between consonants. At least, this general formulation works for my own speech, except perhaps when the preceding consonant is a liquid (r, l), nasal (m, n), or glide (w, y), so probably the preceding consonant must be an obstruent for the deletion to happen.

It happens across word boundaries, by which I mean that the consonant following the t that is lost can be the first sound in the following word. For instance, try saying phrases like "exact same", "act stupid", "fact check", "last call" and see whether the t can be lost.

It is interesting that you notice this loss of t, because it bears on whether the rule that deletes it is an allophonic rule or a morphophonemic rule. Allophonic rules can produce non-phonemic results, but morphophonemic rules act on and produce phonemic forms. We expect morphophonemic rules to produce results that are readily perceptible to native speakers (because phonemes are easy to hear), but we do not expect such rules to apply across word boundaries.

Evidently this rule of English does work on phonemic forms and produces phonemic results, and you have noticed it. So it seems to be morphophonemic, yet it applies across word boundaries.

  • I would hazard a guess that you cannot freely delete inter-consonantal /t/ when it is preceded by a voiced phoneme, or when the /t/ is not in the same syllable as the preceding consonant. This is because, of course, you can freely delete inter-consonantal /d/ whenever the preceding consonant is voiced - regardless of whether it is an obstruent or not so long as they are both in the same syllable. Nov 7, 2016 at 14:30
  • Well, syllable onset t does not delete, at any rate. For instance, in string. The conditions you give for deletion of d would allow deletion in hardly.
    – Greg Lee
    Nov 7, 2016 at 17:20
  • Sorry, was giving conditions loosely in SSBE as opposed to Gen Am. More pecisely the rule for SSBE for dealveloar assimilation would be that the alveolar C would precede a morpheme boundary, the preceding C would match in terms of voicing. John Wells gives it as ALVEOLAR PLOSIVE ELISION / s+c / may be dropped when syllable-final if preceded by a consonant agreeing in voicing AND followed by a consonant So, yes, according to Wells, syllable onset clusters do not count. :) Nov 7, 2016 at 20:51

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.