There was a claim that in English, /d/ becomes [dʒ] (actually /dʒ/) before "r". Could you put up a sound file of a dialect that actually does that? kwami 05:48, 2005 July 24 (UTC)
Pretty much all of American English does this, except in careful speech. Same with /t/->[tS] /_[r]. Any reasonably complete phonetic description of American English will note this. You can even hear some affrication in the extremely careful speech used for the pronunciation files at m-w.com (Merriam-Webster) [1]. Nohat07:16, 24 July 2005 (UTC)
[dʒ] is not standard American English. "Some affrication" (actually something between an affricate and an approximant) is not the same as [dʒ]. The fricative part (or semi-fricative part) is not a sibilant. And it's not /d/ before /r/, but the sequence /dr/ that does this. Although we don't have the sequence /dʒr/ in English, it's not hard to do, and the effect is distinct from /dr/. A transcription [dɹ̝] with a raised diacritic, to show that this isn't the sibilant [ʒ] would be acceptable, though it's really only fricated in the beginning of the /r/: [dɹ̝ɹ] (though that's probably too narrow a transcription for our purposes).
Whether the affrication is part of the /d/ or the /r/ is a moot point because the affrication is the realization of hiatus during the transition from [d] to [ɹ]. Second, the fricative part is most definitely [ʒ]. Try looking at an LPC spectrum of the [dʒ] in that link above—it distinctly has peaks well over 20dB below 2 kHz—quite similar, in fact, to the spectra for [ʃ] and [ɕ] on page 177 of Ladefoged & Maddieson. Furthermore, the place of articulation is decidedly post-alveolar, not alveolar, so [ɹ̝]—which would mean a lightly fricated [z]—is not a particularly accurate transcription. There is nothing about the IPA description of [ʒ] that mandates that it only be used to transcribe "sibilant" sounds; it only need be a voiced fricative articulated postalveolarly. I suppose you could use both the lowered and the retracted diacritic on [ɹ] to transcribe this sound: [ɹ̞̠], but two diacritics seems like overkill to transcribe what is actually [ʒ]. If you're really intent on describing the fact that this phone is fricated more lightly than normal [ʒ], you could use the lowered diacritic: [ʒ̞]. Finally, I think you misperceive the claim that /dr/ is realized as [dʒ]. In fact, that's not the case at all—it's just the /d/ that's realized as [dʒ]. The entire sequence is realized as [dʒɹ] (or [dʒ̞ɹ] if you think it is imporant to mark the fact that the [ʒ] is more lightly fricated).
Also, it's actually [dʒ] not /dʒ/ because we're describing an allophone of /d/, not the separate phoneme /dʒ/.
Yes, the original claim had it as /dʒ/.
You're also drawing a distinction between [t] and [d̥] for the last segment of washed. Is this different from [t] for some people? If not, we should just have [t], since [d̥] suggests partial voicing; there'd be no other reason that I can think of to have a second transcription for ordinary [t] since we're not getting into morphophonemics here. (If there is such a reason it should be explained.) kwami 08:20, 2005 July 24 (UTC)
No one is "drawing a distinction" between [t] and [d̥] as you describe. We are just being complete in indicating that both transcriptions are used to transcribe the sound at the end of washed. The two transcriptions are phonetically identical, and nothing in the prose indicates otherwise. Nohat18:54, 24 July 2005 (UTC)
Since they're not always used to mean the same thing, we should indicate that we're discussing transcription rather than sound here. I'll change it. kwami 19:03, 2005 July 24 (UTC)
Nohat, [ɹ̝] is standard notation for a voiced non-sibilant alveolar fricative. Check, say, Ladefoged. I'll leave in your can-IPA ref, but as an alternate claim, not scripture. kwami 22:33, 2005 July 25 (UTC)
Check the standard IPA chart: [ɹ̝] is so common that they use it as the example of the raising diacritic on a consonant. kwami 00:33, 2005 July 26 (UTC)
