I have heard a few English speakers — native — say the word “three” with what sounds like a flapped r. This might include other words that begin with “thr”, but I can’t remember. It’s just been bothering me for years. Why is this? Is this a documented thing? I don’t think these people have been from the same area. All from the USA, though. Anyone else notice this? I have looked it up, but to no avail.

In case it is useful: I’m from the South, and perhaps it’s less common down here? Maybe that’s why it seems like such an anomaly to me.

  • 1
    US th is usually interdental, and the r is usually retroflex. That has the tongue-tip moving across pretty much it's full range of (comfortable) motion in a very brief period of time, so it's not surprising that both phonemes retaining their usual sounds would be unstable. In particular, the shortest path between both locations has the tip of the tongue intersect with the alveolar ridge, so something like [θ͡ɾ͡ɻ] would almost be expected for /θɻ/. That said, I don't hear an obvious tap in the examples here: youglish.com/pronounce/three/english/us?
    – Tristan
    Apr 20, 2023 at 15:17
  • @Tristan Is the /r/ really retroflex for most Americans? I'm from Chicago, and I'm pretty sure my rhotic is nearly always realized as the molar (bunched) velar approximant - based on other things I've read on the Internet, I assumed this was true for most Americans, with the retroflex and apical (post)alveolar variants more common in limited regions (the South for retroflex, and I'm not sure where for the postalveolar) . Since the molar doesn't require the tongue tip to be turned upwards at all, an epenthetic alveolar tap seems superfluous and unlikely. I certainly don't use one.
    – Graham H.
    Apr 20, 2023 at 17:28
  • 7
    When you say ‘flipped r’ (which is not an established term), do you mean a flapped r, that is, the [ɾ] sound heard for <r> in Spanish pero or for <tt> in colloquial American bottom? A ‘flipped r’ could also be the symbol ɹ (literally the letter r flipped 180°), which is used in IPA for a (post)alveolar approximant, one of the fairly common realisations of English /r/, especially in the UK. Apr 20, 2023 at 20:05
  • I'm surprised other AmEng speakers here are unfamiliar with the flap realization of the rhotic in words like three -- I think of it as being pretty common (and associate it with Black speakers, though I could be wrong about that).
    – Tom Recht
    Apr 20, 2023 at 20:22
  • 1
    If you call it "flipped" instead of "flapped", we don't know what you mean. Phonetics is not worth discussing in text form. Apr 21, 2023 at 18:27

1 Answer 1


There is a "flapped r" i.e. IPA [ɾ] in American English which is an intervocalic allophone of /t,d/ as in "writing; riding", but this does not happen in onset clusters. This is, however a way of pronouncing the /r/ phoneme in some UK dialects. It is attested in the vocal-tract outputs of some Americans – actors – from about 100 years ago, but it was an individual speech affectation that was not part of general US English. I can't say that I have heard a person talk that way in person.

  • 3
    This "answer" doesn't discuss the feature about which the OP was asking at all; instead, it mentions some very basic things of which I'm certain the OP is already aware. They're clearly not asking about the rotated glyph, and I'm sure the feature they notice is a real thing that deserves to be answered in detail.
    – Graham H.
    Apr 20, 2023 at 17:12
  • Since there is no such thing as a "flipped r", your presumption is completely unjustified. I do discuss the plausible linguistic feature that might be assumed to be in question.
    – user6726
    Apr 20, 2023 at 17:19
  • 3
    I'm sorry if my comment was rude. They're referring to an alveolar tap as an occasional allophone of /r/ in clusters beginning with dental fricatives, and you don't discuss that specifically. I've seen this allophony mentioned in other places on the Internet before, so it's not an unknown thing. I don't know enough about it to give a full answer, and I wonder if you might not either - but I'll try not to make any certain presumptions.
    – Graham H.
    Apr 20, 2023 at 17:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.