From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Linguistics / Phonetics  (Rated Start-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Linguistics, a collaborative effort to improve the coverage of Linguistics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by the Phonetics Task Force.


This is what you use to order natural healing ointment for home delivery (compare francophone. -- N. Utt

Does the deliverer travel the allopath? -phma

And do you sign the receipt with an allograph...? thefamouseccles


What is the difference between allophony and sandhi rules? (talk) 21:17, 19 November 2007 (UTC)

It's a good question, but since most phonology really can't account for most phonological variability--it certainly can't with just the concept of allophony--I doubt anyone will address this at the likes of a wiki article. —Preceding unsigned comment added by (talk) 07:15, 2 August 2008 (UTC)

Voiceless b[edit]

It is not quite true that Pinyin b, d, g are pronounced like unaspirated p, t, k. In the southern dialects of German (spoken in Austria, Switzerland, Bavaria and the like), b/d/g have become voiceless -- but, except in eastern Austria (like Vienna), unaspirated p/t/k (called "hard consonants") and voiceless b/d/g (called "soft consonants") stay two different sets of phonemes! You can trust me on this. I'm a native speaker. :-) The difference is that b, d, g are shorter and less intense than p, t, k. Pinyin b, d, g sound exactly like I pronounce German b, d, g (I learn Chinese).

This kind of sound also seems to occur at the beginning of Korean words, and apparently in Mongolian and Navajo.

David Marjanović 2005/8/6

My understanding of this is that actually the southern German dialects probably have /b/ as [p] and /p/ as [pp] i.e. they use a simple consonant for the soft voiceless unaspirated stops, and a geminate consonant for the hard voiclesess unaspirated stops. — Felix the Cassowary 05:03, 27 August 2005 (UTC)
That is not true. The most important difference is air pressure, intensity if you will. The easiest way to increase the air pressure of a stop is to prolong the hold, so the fortes are indeed normally longer than the lenes, but this also depends on other things like the surrounding sounds and the speed of speech. As an example, I hear the first word in this list of Lakhota audio samples as containing a long (IPA half-long?) voiceless [b]. It's not a [p].
Do you have Skype? :-)
David Marjanović | | 23:30 CEST | 2006/4/8
Since [b] is the "Voiced bilabial plosive", I don't understand what "voiceless b" could possibly mean.—Nat Krause(Talk!) 04:07, 16 October 2006 (UTC)
[b] is the voiced bilabial plosive, but /b/ has a range of phonetic difference as well as allophony. In French, it is an implosive [mb] and in English it is somewhat devoiced (although not completely).

Much of this discussion is doomed to the same old problem--you are categorically confusing articulation/production with perception. A lot of the differences in initial consonants are really about voice onset timing. It isn't that an initial [p] of English 'pill' does not voice, it voices well later than the initial [b] of 'bill'. Some languages are characterized by voicing of so-called 'voiced' initial consonants before the consonant is articulated--this, for example, marks a Spanish or Italian accent in English. —Preceding unsigned comment added by (talk) 07:10, 2 August 2008 (UTC)

'Night Rate' vs. Nitrate[edit]

I removed the comment about English speakers being able to hear the difference between t and its allophone ght as it is patently absurd. An English speaker can hear the difference between 'night rate' and 'ni-trate' because they are pronounced differently.

Isn't that precisely what it said before you removed it?—Nat Krause(Talk!) 04:04, 16 October 2006 (UTC)
Yeah, that removal didn't make sense. But to avoid confusion we might put some IPA. Nitrate is [ˈnaɪˌtʃɹeɪtˀ] and night rate is [naɪtˀ.ɹeɪtˀ]. I think... Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 04:26, 16 October 2006 (UTC)
Pharyngealisation? -Iopq 09:51, 16 October 2006 (UTC)
Glottalization. <ˤ> denotes pharyngealization. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 17:10, 16 October 2006 (UTC)
It might be clearer for the example if we transcribe nitrate as [ˈnaɪˌtʰɹeɪtˀ], which is also a valid pronunciation, and perhaps more common in well-enunciated speech.—Nat Krause(Talk!) 19:02, 20 October 2006 (UTC)
Well, I would remove the aspiration marker and make the r voiceless like this: [ˈnaɪˌtɹ̥eɪtˀ]. It's just simple spreading of voicelessness. Maybe that would be confusing. Is there another good example set? Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 19:20, 20 October 2006 (UTC)
I think "nitrate" sounds totally wrong without either aspiration or a ʃ. Possibly both.—Nat Krause(Talk!) 19:55, 20 October 2006 (UTC)
The voiceless r is due to the aspiration. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 22:18, 20 October 2006 (UTC)

On another note, I think that a Mandarin speaker would not be able to discern the two too well either, because Mandarin does not have entering tones (where words end in unaspirated consonants p, t, k). The only possible ending consonants in Mandarin are n, ng. Compare this with Cantonese, which has the full compliment of Middle Chinese finals (ending in consonants m, n, ng and the corresponding p, t, k). (talk) 15:41, 31 May 2014 (UTC)

A good example for allophones is comparing English to French![edit]

This article is still a bit of a stub ... Well, let's compare the words coffee (en) and café (fr). The French pronunciation sounds fairly like "guff-fay" because the c is NEVER aspirated in French. That's a 100% rule, and will distinguish a good foreign French speaker from a bad one. One less common language where you find this is Finnish: kaappi means cupboard, and the k is always very smooth in sound, because it's never aspirated either. Also listen how F1 pilots' names like Räikkönen are pronounced in Kimi's home country. The k never sounds hard. -andy 16:30, 2 December 2006 (UTC)

That's not allophony, that's phonetic realization. So English and French both have /k/ but in English it is [kʰ] (in that position) while in French it is [k].Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 21:17, 2 December 2006 (UTC)

The point about the free variation form of allophony is not just that sounds show variations, but that these variations bear a relationship to one another within the overriding homogeny of a language community. It is not helpful to use the same terminology across language boundaries. Of course you could say that that is arbitrary, since we do do it across dialect boundaries, and linguistics has never been able to define the difference between separate dialects and separate languages. But nevertheless, we don't do it. --Doric Loon 10:32, 3 December 2006 (UTC)

Phonetic variation and allophony are clearly related. All those people trying to describe a phonology in terms of allophones simply give up due to the inadequacy of the tool and ignore far too much variation that might be explained better by a more comprehensive model of articulation and perception of language. Voice onset timing analysis might prove more illuminating in the case of initial [k] in French, English, etc. Why, for example, do English speakers hear a [g] when the foreign speaker of English might think they are making a [k]? —Preceding unsigned comment added by (talk) 07:34, 2 August 2008 (UTC)

I'm a native speaker of Midwestern American English who has been observing native French for 40 years, and I have never perceived the phonetic realization of French /k/ as [g], nor have I ever witnessed a situation in which another anglophone has. —Preceding unsigned comment added by (talk) 14:24, 28 January 2009 (UTC)


Is that the correct pronunciation of nitrate? Is that ʃ supposed to be there? Jon Harald Søby 17:47, 11 December 2006 (UTC)

Yeah. Many speakers affricate and retract /t/ before /r/. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 18:36, 11 December 2006 (UTC)
As in "train", "tree", and "try"? Any references on where these "many" speakers live? Or could we change the IPA to reflect some other dialect than a non-standard one (since /t/ -> /tʃ/ before /r/ is certainly not standard). Tesseran 22:59, 2 March 2007 (UTC)
I don't have any references but my understanding is that it's largely idiolectal. That is, there isn't a region per se where people speak it. I'm not sure how un-standard it is since speakers generally can't even tell the difference even if you point it out to them (at least in my experience).
We could always say "for some speakers" to clarify the matter for people. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 01:53, 3 March 2007 (UTC)
I'm more incredulous than anything else ... am I understanding the sounds involved correctly? Stepping away from IPA because maybe that's the problem, you're saying that for some people, the word "tree" begins with the first sound of the word "chair"? And that they're not aware of this? That just really surprises me. Can you pin down e.g. AmE vs. BrE vs. AusE?
In regard to the article, the problem (as I see it) is that the example of night rate vs. nitrate works better for speakers who don't have this /t/ -> /tʃ/ shift. For that reason, I think a transcription without would be better. (Also, any phenomenon that is idiolectal [idiolectical?] seems like a bad candidate for mention, no matter how widespread.) However, if I'm missing something, and this shift is necessary for nitrate to be an example, then I withdraw the suggestion. Tesseran 09:11, 3 March 2007 (UTC)
(By the way, what I find surprising is not that some speakers might affricate /t/ in this one context, which seems quite possible, but that it would be part of a larger shift. I'm partially surprised just that I've never heard of such a phenomenon.) Tesseran 09:21, 3 March 2007 (UTC)
Yeah, you're understanding it correctly. I think I see your point and I can agree that it might be confusing to some people. I simply chose that pronunciation because it shows a third allophone of /t/ (at least for some speakers). But the point is sufficient without it.
You're saying that you're surprised that the affrication of /t/ before /r/ is part of a larger shift? A shift of what? Is it also surprising that it may be affricated in words like twin and dwarf? Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 17:46, 3 March 2007 (UTC)

Allophones in Hawaiian[edit]

I heard that some phonemes in Hawaiian have highly unusual allophones. Maybe it would be worth mentioning (as well as examples from many other languages). —Preceding unsigned comment added by (talk) 19:11, 22 June 2008 (UTC)

Hawaiian phonology talks about the variations in the pronunciation of consonants, although most of the weird ones are sounds in free variation. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 19:26, 22 June 2008 (UTC)

Why not cut to a short dictionary-like article[edit]

Allophony would appear to be an operative concept in phonetics--the variation of realized sounds, across speakers, dialects, languages, even just across instantiations. It is also a key concept in largely outmoded phonemic accounts of phonology. Much of the rest of this article could just be covered in phonetics and phonology then. Since so much of Wiki is shite, really, less is much much more. —Preceding unsigned comment added by (talk) 07:13, 2 August 2008 (UTC)

deleted paragraph[edit]

I took out this paragraph as I feel it belongs in phoneme, and only confuses the issue here:

Speakers of a particular language perceive a phoneme as a distinctive sound in that language. This could support the theory that a language's phonology is comprised of an identifiable, stable set of phonemes. However, it might also indicate languages simply schematize and limit redundancy in how phonetics are used in a given spoken languages, thereby making spoken communication more informationally efficient. One theory is that phonemes are determinative of speech perception; however, that is not the only possible interpretation. Rather, phonemes may themselves be the result of successful speech perception (in other words, a phoneme can be identified because successful lexical recognition has already taken place). Usually, allophones falling under a specified phoneme are not considered distinctive, but rather a variant of that phoneme. In theory, changing the allophone won't change the meaning of a word, but the actual result may sound non-native, or be unintelligible to normal speech perception processes.

kwami (talk) 08:10, 2 August 2008 (UTC)

Examples of the different allophonic processes.[edit]

I think some word examples of the different allophonic processes would greatly help readers to better understand what is happening in their vocal tract by practicing with actual words.

Ex: Lack of plosion – In English a plosive (p, t, k, b, d, g) has no plosion when it is followed by another plosive or an affricate inside words or across word boundary. example: ditch /dɪtʃ/; judge /dʒʌdʒ/

(It might also be helpful to indicate the lack of plosion as a quasi glottal stop, but without creating a separate syllable.)

Nasal plosion – In English a plosive (p, t, k, b, d, g) has nasal plosion when it’s followed by nasal, inside a word or across word boundary. example: big noise /bɪgⁿ nɔɪz/

etc... Finitoultero (talk) 04:10, 12 August 2008 (UTC)

Can a better example be found?[edit]

"For example, [pʰ] (as in pin) and [p] (as in spin) are allophones for the phoneme /p/ in the English language."

I have said these words over and over, and I can detect no difference whatsoever in the "p" sounds. Can we have an example that is more obvious to the oridnary reader, please? (talk) 11:42, 8 September 2010 (UTC)

They sound distinct to native English speakers. If you pronounce /p/ as [p] in "pin", it will sound like "bin" to many English speakers unless they are exposed to non-native speakers. Could it be that your native language doesn't use aspiration? Dutch, French and Japanese are examples of this. If so, you will have trouble telling [p] and [pʰ] apart. We shouldn't change the example because aspiration is, in fact, one of the biggest sources of allophony in English. What we can do is add another example that uses something other than aspiration. How about /k/ in "skew" where it is pronounced [kj] ("skyew") vs in "scull" where it is pronounced as a simple [k]? From --Hunnjazal (talk) 15:30, 19 September 2010 (UTC)

I think very few native speakers will hear the difference between aspirated and unaspirated p. I am a native speaker, and have been teaching languages at university for nearly 20 years, but frankly I have great difficulty "hearing" the difference. I feel the difference if I put my hand in front of my mouth and concentrate on the puff of air. I feel it if I think about the tension in my lips. I understand theoretically how the s changes the airflow on the p. But I am really struggling to be aware of the audible difference in the p itself. There are far more obvious examples of allophony than this, ones that native speakers will immediately relate to without having other linguistic phenomena explained. --Doric Loon (talk) 17:28, 19 September 2010 (UTC)

Yes, that would be expected. As the article says - Speakers of a given language usually perceive one phoneme in their language as a single distinctive sound in that language and are "both unaware of and even shocked by" the allophone variations used to pronounce single phonemes. Similarly, Hindi-Urdu speakers will tell you that it is silly to talk about differences between [v] and [w]. "They sound the same." None of these sound the same. You yourself are pronouncing them differently ("I feel the difference if I put my hand in front of my mouth") but are unconscious of it. In your mind they are mapping to the same phoneme. That is *precisely* how allophony works. If English had explicit aspirated consonants, this would have been patent, similar to skew vs scull (I guess that's one of the more obvious examples). Feel free to add it, but do not remove the aspirated examples (pin-spin or nitrate-nightrate). Virtually every text I can find lists these as prime examples of English allophony. Please also be mindful of WP-NOR. --Hunnjazal (talk) 22:36, 19 September 2010 (UTC)
An addition, which may or may not be helpful to you (it might, because you mentioned that you teach languages, so might be comfortable with foreign notation), but does illustrate a related point. When Hindi speakers learn English, they infer phoneme-correspondence. There is an underlying assumption that a 1-to-1 or 1-to-none phoneme mapping exists. The problem is that it can be one-to-many. So, Hindi speakers learn that /t/ corresponds to /ट/. However, it doesn't. /t/ is allophonic and corresponds to either /ट/ or /ठ/. English-speakers use them allophonically and switch unconsciously. When a Hindi-speaker asks them "are you making a /ट/ sound or a /ठ/ sound," they look blankly back at them and say, "there's no difference, they sound the same to me." So, the Hindi-speaker assumes a default /t/=/ट/ equivalence. When the Hindi-speaker says "Tea" they say "टea" instead of the correct "ठea". Since, in this position, /t/ is always /ठ/, this sounds like "Dee." If the English-speaker has been exposed to Hindi-speakers, they are able to compensate and understand, thinking (incorrectly) "well, Indians can only speak flat English." If they are not, what they hear is "I want some Dee." This feels bizarre to Hindi-speakers because /d/ is /ड/ - "how could anyone possibly confuse the two?" But they can, if allophonic rules governing /t/ in English dictate that it is always aspirated in that position - the closest unaspirated sound is, in fact, /d/=/ड/. When an English-speaker mimics a Hindi-speaker he will say exactly stuff like this - "I binned a note to my dee-shirt." (I pinned a note to my tee-shirt). This sounds dishonest to a Hindi-speaker because he said [p] not [b] and [t] not [d]. The problem is that he should have said [pʰ] and [tʰ].
Hindi-speakers run into this for a bit and then think, aha!, it's always a /ठ/. Now they're saying "Nighठ Rate," instead of the correct "Nighट Rate," which makes it sound like "Nitrate." Although there are conflicts (such as the one just cited) with this, it is actually more intelligible to English-speakers. It sounds like you're putting on an unnatural accent though - to both other Hindi-speakers and English-speakers. The issue is a genuine disconnect. Hindi-speakers are pulling their hair out asking "what the heck do you want from me - /ट/ or /ठ/?" English-speakers are saying "there's no difference, you're just having some difficulty for some reason." The same is true in reverse for /v/ and /w/. English-speakers who learn Hindi want to know - "is /व/ equal to /v/ or /w/?" Hindi-speakers say, "they're the same, there is no difference." The English-speaker has either heard Hindi-speakers say something like "adwance" for "advance," so he assumes that /व/ = /w/. Now (in Hindi), the English-speaker reads व्रत as wrat when it should be vrat and says "Muslims keep a wrat in Ramzan (Ramadhan)." Instead of meaning, "Muslims keep a fast in Ramzan," this sounds like "Muslims keep a woman in Ramzan." Confusion ensues and the Hindi-speaker says "vrat, vrat - can't you just say vrat?" The English-speaker says "well is व va or wa?" The Hindi-speaker says "they are identical, what are you talking about?" When a Hindi-speaker mimics an English speaker, he will say "I am going to Oo-aa-ranasi" (I am going to Varanasi), which makes it sound like the speaker has some abdominal issues. This is offensive because the speaker said "I am going to Waranasi," and this feels like a deliberate attempt to exaggerate his accent.
But, you see, the best examples to demonstrate this unconscious switching of phones and adamant insistence that they are the same phoneme in Hindi will be things like [v]-[w], where Hindi-speakers will complain "they are identical - give me a different example." Ditto for English. Sorry, that was a long explanation, but I think it's all relevant. --Hunnjazal (talk) 06:13, 20 September 2010 (UTC)
Hunnjazal, thanks for your informative reply. I am actually a native English speaker, so your statement that "They sound distinct to native English speakers" doesn't quite apply to me. On the other hand, your comment that this is almost the whole point of allophony -- that speakers don't notice the difference, and find it hard to detect one -- seems very pertinent. In fact, though, after trying all sorts of antics that must have made me appear like lunatic, I think I may have found a way to make the difference detectable: by comparing "Look at its pin" with "Look at it spin". Does that work for anyone else? As far as your other example is concerned, the sounds in "skew" and "scull" seem obviously different to me -- too obvious for it to be allophony? "skew" and "skoo" could be different words (whereas "pʰin" and "pin" couldn't) so doesn't that disqualify "k" and "ky" from being allophones? (talk) 00:28, 29 September 2010 (UTC)
Glad you found it useful! "Look at its pin" versus "Look at it spin" works great. I guess the maddening thing about allophony is that native speakers (of every language I know of) are unconscious of the allophonic differences but actually quite Stalinist in preferring them (specifically talking about conditional allophones). When someone (non-native speaker) doesn't apply them as they ought to, native speakers have trouble understanding them (extreme case) or find the accent foreign (mild case). The insidiousness of it is that the native speaker cannot really pinpoint for the non-native what the problem is except to say unhelpful things like "try harder" or "listen carefully," followed by "no, that's not it. no, that's not it. yeah, that's it! no, that's not it - say it like you said it before! no, that's not it." Even after the non-native catches on to the whole allophonic thing, the work still lies ahead of him. Now he needs to figure out which allophonic variant is standard in which situation. Painful. On "skew" and "scull", I'd have thought the same thing, except that reference had it as an example. I agree it doesn't really feel like a good one. --Hunnjazal (talk) 03:28, 29 September 2010 (UTC)
As a native speaker of both English and Hindi, I feel that the "k" sound (Voiceless velar plosive) better shows the unnoticed difference in aspiration among English speakers. Eg. Can and scan. Interesting I think many native speakers can make the change to an unaspirated "can" when trying imitate an Indian or many European accents. GizzaDiscuss © 03:58, 10 November 2010 (UTC)

In miy experience, the contrast between aspirated and non-aspirated plosives is a peculiarity some varieties of English, expecially those spoken in England, whereas American or Canadian speakers don't make this distinction. Regarding Indian English, the idea that Indians should learn to aspirate the English /p/ /t/ /k/ is nonsense; Englis is a universally used language in India and with mixt marriage has also become a mother-tongue for some children. Indians (and Pakistainis0 pronounce English in the Indian way, such that linguistically, Indian English is a valid variety of English. Forcing Indians to aspirate plosives would alienate these speakers from the rest of the Indian population. Apart from this, I have heard some Pakistanis pronounce /θ/ in "this" as an aspirated plosive.  Andreas  (T) 12:48, 13 August 2012 (UTC)

Its wrong Andreas. Americans, Canadians and Australians also do. Indians perhaps don't, but it makes no sense writing "in non-indian English".·ʍaunus·snunɐw· 12:50, 13 August 2012 (UTC)

phones map to graphemes not only phonemes in example[edit]

Hi. In the section Allophone#Examples_in_English_vs._other_languages, concerning the the Hindi-Uru example part that says "write the two phones differently and treat them as completely distinct phonemes: [p] is written as 'प' (or 'پ'), while [pʰ] is written 'फ' (or 'پھ').[5]" , shouldn't it be added that these (phones) map to distinct graphemes too? - Scriber (talk) 14:51, 2 April 2011 (UTC)

Seems to me the sentence already says that twice. —Tamfang (talk) 02:56, 10 April 2011 (UTC)

the flowchart[edit]

Are the sounds phonetically similar? — Seems awfully subjective. —Tamfang (talk) 02:53, 10 April 2011 (UTC)

It's meant to cover cases such as English [h] and [ŋ] (as mentioned on the image description page). I didn't know of any way to state it more clearly that would have fit in the image at the chosen font size. AnonMoos (talk) 18:01, 10 April 2011 (UTC)

Nasal Vowels[edit]

Does English really have nasal vowel allophones, as stated in the last section? This is not mentioned in English phonology. I think Spanish might make a better example in that section. — Preceding unsigned comment added by Pokajanje (talkcontribs) 16:38, 26 September 2012 (UTC)

May be more obvious in closed syllables ("meant" instead of "many"), and somewhat dependent on dialect. AnonMoos (talk) 20:58, 26 September 2012 (UTC)