Some of my students have a disagreement about transcribing the pronunciation of "trade" in American English. Some say it's (a) [t͡ʃeɪd] while others (and they point to dictionaries that support them) say it's (b) [treɪd]. I thought it was the former; the latter sounds a little like a fast "tirade". But now as I try to articulate it more, it seems like (c) [t͡ʃreɪd]. Any insights on this?

  • 6
    There is considerable variation in how much palatalization is introduced in a transition between an apical stop and a following /r/. For one thing, the actual pronunciation of /r/ varies widely, even in the USA. For another, there is no /tʃr/ cluster for /tr/ to constrast with, so it has spread into the palatal area (much the same way English /ə/ spreads out to encompass all central vowels, or English /h/ to encompass all voiceless vowels). It is certainly common, and counts as a sociolinguistic variation; you can make a teaching point out of it. Jan 18, 2014 at 19:03
  • 4
    @FumbleFingers Actually, it’s perfectly common in North America, too. A tree out of the mouths of babes is often enough a chree.
    – tchrist
    Jan 18, 2014 at 19:04
  • 2
    (a) is definitely not correct. That would be ‘chade’. I would say that the ‘rub-off’ of the r on the t is more retroflex than palatal, so /tʂɹʷ/ is probably more accurate for American English. Jan 18, 2014 at 19:17
  • 3
    I take it the implication of @John's there is no /tʃr/ cluster for /tr/ to constrast with means native speakers wouldn't particularly notice (slight) variations in pronunciation, since it never makes any difference to meaning. Jan 18, 2014 at 19:36
  • 5
    There's always a lot of dialects to account for; as far as I know, there is no list of lects that palatalize /tr/, nor correlation with any other phenomenon. It's, as I said, quite variable. Explanations await properly instrumented phoneticians and sociolinguists; in the meantime, it's a good classroom example of what we do say and we don't hear. Jan 18, 2014 at 21:24

1 Answer 1


The answer to your question depends to a large extent on whether you are using IPA to represent a phonemic transcription of the word, or a phonetic transcription.

In either case, your option (a) is absolutely not correct. IPA /t͡ʃeɪd/ (or /t͡ʃɛɪd/, depending on how you choose to transcribe diphthongs) would as regularly as possibly be represented orthographically as ‹chade›, even though this is not a word, as far as I know. A minimal pair can easily be found, though: ‹chuck› /t͡ʃʌk/ vs. ‹truck› /trʌk/ (phonemically).

Phonemic transcription

If you wish to transcribe the word phonemically (between /slashes/), according to the underlying phonemes it contains, John Lawler’s comment is immensely pertinent: there is no phonemic opposition in English between initial /tr/ and /t͡ʃr/. Whether the sound cluster found initially in ‹trade› is underlyingly made up of /t/ + /r/ or of /t͡ʃ/ + /r/ cannot really be determined.

As such, it doesn’t really matter which one you choose to use, as long as you do so consistently. You can write /treid/ or /tʃreid/ as you see fit. The former is probably more likely (adhering to Occam’s razor) and is what is normally used in dictionaries and other works that include phonemic transcription; but the latter is at least a theoretical possibility.

Phonetic transcription

If you are rather intending to use IPA to transcribe the word phonetically (in [square brackets]), describing in detail the actual sounds that your speech organs produce when saying the word, it is irrelevant what underlying phoneme the word is made up of—but it becomes then necessary to define which pronunciation of the word you are talking about. Some dialects will have differing pronunciations of more or less any word in any language. In this particular word, though, the main differences are twofold:

  1. The quality of the phoneme /r/. Some dialects (most forms of BrE, SAE, NZE, and variably AusE) have a simple, alveolar or retroflex approximant [ɹ] or [ɻ]; others (especially most forms of AmE and IrE) consistently labialise this and have [ɹʷ] or [ɻʷ]. The exact place of articulation varies, but it is usually retracted enough to count as at least semi-retroflex. (Some, like Scottish, tend to have an actual [ɾ], but let’s leave those out of this altogether.)

  2. The quality of the first part of the diphthong /ei/. Some dialects have [eɪ], others have [ɛɪ], some even have [æɪ] or [aɪ] (most notably AusE), and some (like Scottish and Irish, and some BrE dialects) monophthongise the sequence to [ɛː] or [eː].

If we focus on AmE, the first is almost always [ɻʷ]; and the second varies, even from speaker to speaker, between [eɪ] and [ɛɪ], but this difference does not make a difference for the initial consonant cluster, so I’ll just be lazy and write /eɪ/.

The transition from the plosive /t/ to the approximant /r/ causes the cluster to become an affricate, which is an almost universal case of assimilation. Since in AmE the /r/ is nearly always retroflex or near-retroflex, the assimilation causes the /t/ to move towards a retroflex place of articulation, i.e., the tip of the tongue is held against the hind part of the alveolar ridge and the tongue moves backwards upon the release of the plosive: [t͡ʂ].

In contrast, the phoneme /tʃ/ is postalveolar or palato-alveolar, meaning that the blade of the tongue (the part just behind the tip) is against the front part of the alveolar ridge, while the dorsum (the back of the tongue) is domed upwards towards the hard palate. The tongue moves forward upon release.

In other words, an actual phonetic sequence [t͡ʃɻʷ] is quite difficult to produce: it involves having the tongue start at the front of the alveolar ridge, with the dorsum domed upwards, and then upon release move backwards and curl upwards, i.e., both moving the front of the tongue and switching the curvature from downward to upward.

It is much easier to pronounce [t͡ʂɻʷ], which is also what initial /tr/ in AmE quite distinctly sounds like to me.

So in purely phonetic transcription, I would write ‹trade› (in AmE) as [t͡ʂɻʷeɪd].

  • 3
    I keep wanting to upvote your individual points. :)
    – tchrist
    Jan 18, 2014 at 20:52
  • Beautiful answer. I don’t have rounding in this context (nor such a high front vowel at the start of the diphthong). Interesting to know that (some) Americans do. (I'm antipodean.) Jan 18, 2014 at 22:02
  • Just a comment ... while I use [ɹʷ] for initial /r/, as in raid, I don't believe I have any labialization in trade or grade. It seems quite difficult for me to do. Apr 1, 2015 at 18:57
  • @PeterShor Interesting. It's the opposite for me, almost: I find it easy not to labialise the /r/ in raid or grade (as in Irish English, for example), but I find it very difficult not to labialise the one in /trade/. Apr 1, 2015 at 19:03
  • Huh... I recently realized that I consistently have labiovelar coarticulation for onset or intervocalic /r/, so I guess for me the phonetic realization would be [t͡ʂʰɻ^ʋɛɪ̯ˑd̥] or something of the sort.
    – herisson
    Aug 19, 2015 at 4:57

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.