Talk:English phonology

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Aspirated p's and t's[edit]

Just curious as to why Icelandic phonology lists aspirated t's and p's as separate phonemes from t and p, but not this article? Doesn't English also make common distinctions between the two? I.e isn't the initial t in say, Thomas and total, two different sounds? Or the p in pot vs stop. What makes Icelandic so special in this regard? More to the point, why aren't they included here? Peter Greenwell (talk) 15:00, 25 July 2009 (UTC)

In English, aspirated and unaspirated stops are not separate phonemes, as they are in Icelandic. Thomas and total begin with the same (aspirated) allophone of the same phoneme, while pot and stop have two different allophones of the same phoneme. +Angr 18:06, 25 July 2009 (UTC)
At a rather broad level, there are three phones, [tʰ t d]. Historically, [tʰ t] are allophones. In child language acquisition, [t d] are allophones. That is, if we transcribe tie as /tʰai/, and die as /dai/, then for sty we would need to choose between ?/stʰai/ and ?/sdai/, not a happy choice for most people. Or we could transcribe die as /tai/ (as we would in Icelandic), also not a happy choice for most people. So we follow the historical derivation and avoid ?/tʰ/ altogether. kwami (talk) 20:35, 25 July 2009 (UTC)
Super super late response, is this allowed? But I wanted to add: basically, the difference between [pʰ] and [p] is automatic and meaningless. If you were to pronounce spin as aspirated [spʰɪn], people would barely notice the difference.
Granted, if you were to pronounce pin as unaspirated [pɪn], it could be potentially problematic because the phoneme /b/ may be partially or fully devoiced [b̥~p] so a very telling difference is in fact the aspiration, and it could be mistaken for the word bin. (For example if a native English speaker were to listen to someone speaking a language where aspiration is the actual distinguishing factor - and I want to say for example Thai because it's a funny sounding language - sometimes some of their [p] will sound like a /b/). However /b/ is usually devoiced around other voiceless consonants: complementary distribution is the relationship where one element is found in one set of environments and the other element is found in a non-intersecting (i.e. complementary) set of environments. It often indicates that two superficially different elements are the same linguistic unit at a deeper level.
Icelandic, on the other hand, has the phonemes /p/ (written 'b' in the orthography) and /pʰ/ (written 'p'), which COULD be argued for English (though as mentioned, makes less sense than a voiced-voiceless /b/-/p/ distinction). The 3 different phones [b p pʰ] are definitely not all separate phonemes.
So in summary: the phonemes of Icelandic are unaspirated-aspirated /p/-/pʰ/ because Icelandic is a different language that is pronounced differently. (Disclaimer: I don't know what I'm talking about.) (talk) 19:27, 31 August 2015 (UTC)


Calling sphragistics, sclerosis, sthenics, phthalic, and thlipsis "loanwords" seems pointless. All of these words have been neutralized and are no longer foreign words. There is an obvious distinction between these words and words like "bwana" and "schvartze". The American Heritage Dictionary defines any word borrowed from another language as a loanword, including words like "very".[1] This means that prize, language, shrimp, pure, beautiful, tube, during, cute, argue, music, view, suit, Zeus, huge, lurid, skill, sphere, scream, square, student, and skewer are all loanwords. The aforementioned Greek loanwords clearly belong in this class. Perhaps the best solution is to use a term other than "loanword" to refer to words like "bwana" and "schvartze". — The Man in Question (gesprec) · (forðung) 20:36, 8 September 2009 (UTC)

"Recent loanword" might suffice. Or loanwords from X language since the __ century/era. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 22:53, 8 September 2009 (UTC)
The distinction I understand to be made here is that between inherited clusters, and clusters which only exist in loanwords. The choice of examples is not relevant. The likes of "Zeus" would be included by virtue of the cluster having arisen during the development of English, even if originally a loanword.
As a non-nativ speaker my opinion may not count, but "sphragistics" thru "thlipsis" are definitely Greek to me! At best I would be willing to concede them to be part of very specialized registers of English, but by that criterion "bwana", "zloty" or "schmuck" would be in just as well (if some not more so).--Trɔpʏliʊmblah 18:50, 9 September 2009 (UTC)
I think most native speakers would agree that thlipsis is not "English". Certainly 99%++ of native speakers would see a /θl/ onset as foreign, and be utterly unable to come up with an example on their own. /sf/, on the other hand, is fairly common, even if mostly in literary or technical language. Any child who's had geometry in school, or watched documentaries on Egypt has /sf/. That can't be said for most of these obscure loans. kwami (talk) 00:58, 10 September 2009 (UTC)
Ah, I see that the OED agrees with me. It marks thlipsis with the symbol "||", which stands for "not naturalized, alien". I think we definitely need to distinguish normal English phonotactics, meaning what people actually use when they speak, from Scrabble hunts for odd sequences that most of us have never heard spoken in our lives. kwami (talk) 01:01, 10 September 2009 (UTC)
There is some degree of arbitrariness to this, tho. If /pr/, /kr/, /pl/, /kl/, /spr/, /skr/, /spl/ are valid, shouldn't that mean that (s)(p/k)(r/l) is valid, and that /skl/, while only occurring in the loanword scleroris, should still be consider'd a part of "normal English phonotactics"? FWIW, Urban Dictionary has a few dozen entries under scl- and skl-, but very few under eg. sth- and many of them acronyms. (Also, tons of onomatopoeia under bw-.)--Trɔpʏliʊmblah 11:27, 10 September 2009 (UTC)

(outdent) Now that the topic is up, let's also mention that I've seen many sources tangent on an analysis of /ju/ being a difthong rather than constituting onset clusters involving /j/. (Eg. [2] mentioning "Davis & Hammond '95" but without further details). --Trɔpʏliʊmblah 11:27, 10 September 2009 (UTC)

I don't think we should present a theoretical abstraction as reality. Otherwise what's to stop s.o. from giving (s)(p/t/k)(r/l), and claiming that *stleet as a parallel to street is just an accidental gap? (After all, in quick enunciation delete may become [dli:t].)
Onomatopoeia has its own phonotactics in many languages. If we're going there, we'd have to say English is a click language, and that we have syllable obstruents like /S/ and /z/.
You have a good point with /ju:/. Historically that's a diphthong, though it's arguable today. I suggest we separate out those onsets and note that they depend on one's interpretation of /ju:/. kwami (talk) 20:27, 10 September 2009 (UTC)

I find /skl/ (but not /stl/) persuasive, not just bcz of sclerosis but bcz a presumably linguistics-naive author (or perhaps an actual child whose usage suggested her fictional child's usage) created a little girl fond of the neologisms "to sklathe" and "sklathing". (Am i foolish in believing the final phoneme is clearly voiced, or is that implicit in spelling it so, rather than "sklaith"?)
--Jerzyt 18:38, 10 November 2015 (UTC)

Could somebody translate this article into English, please?[edit]

I regret to say that most of this article (as with many of the linguistic articles in Wikipedia) is virtually unitelligible to the vast majority English speakers. The problem is twofold. First, the article assumes knowledge of linguistics terms of art (rhotic? voiceless dental fricative?) that no one outside the field of linguistics understands. The second problem is the use of phonetic symbols that mean nothing to anyone besides professional linguists. I really don't think that I should be required to learn what amounts to a whole new alphabet to read an article that is ostensibly in English. —Preceding unsigned comment added by (talk) 21:06, 22 September 2010 (UTC)

I see the remedy to the first problem being wikilinks and to the latter being lexical sets. Of course, knowledge of the IPA and intrinsic understanding of terminology provides a potentially richer experience (or more sophisticated critique of low quality articles). We're always open to new ways of framing this kind of specialist knowledge, though I suspect that people unwilling to learn a phonetic alphabet also don't care enough about the topic to learn much about it, anyway. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 21:40, 22 September 2010 (UTC)

Your attitude shows that you clearly have no respect for linguistics as an academic discipline. Would you make the same complaints about physics or mathematics articles? They too are chock-full of terminology on the subject that would prove too dense for the average reader, yet I for one wouldn't complain: it's an article about the subject, you should have some prior knowledge. Anyway, if you don't understand a term, it's probably hyperlinked to the article explaining it. As for not using IPA to describe the sounds, how else would you do it? You can hardly use the English spelling system as a phonetic guide (if you HAD understood the article you'd realize how varied pronunciation is, so the 'phonetic' spelling of words would be pronounced differently by different people, thus rendering it useless). If you are not interested in learning more about phonology in general, then what kind of explanation of English phonology were you looking for? — Preceding unsigned comment added by (talkcontribs)

I like the comparisons to physics and mathematics. We might also compare linguistics articles to those on music; if you don't have the vocabulary, they aren't "in English" either. GeorgeTSLC (talk) 03:55, 28 August 2012 (UTC)
I really like your comparison; it may be even better. After all, music theory is every bit as arcane and complex (and confusing to the lay reader!) as linguistics, although everybody (except for the profoundly deaf or isolated) listens to music (voluntarily or not) and almost everybody likes (some) music, music being just as common (and commonly discussed) in everyday life as language, but to millions of music lovers and amateur musicians, who can barely read notes, music theory (especially advanced music theory) is utterly alien and comes across a lot like advanced mathematics.
Moreover, the analysis of music, especially familiar music, tends to baffle and even antagonise people – why can't those weirdo academics simply shut up and listen to the beautiful sounds? Why do they need to cut it up into little parts and process it with abandon? Can't they appreciate music on its own, as it is? Aren't they destroying the magic of music with all this prattle?
Music is all around us, as a cultural phenomenon, and we are well aware of its existence (unlike, for example, bacteria), though not its details (seeing as we usually perceive it in a holistic fashion), it has a complex structure, like language, and we tend to have even stronger feelings about it than about language. This makes the place of music in our culture and daily lives quite comparable to that of language, especially a common language like English, and studying either as a (broadly speaking) scientific or academic subject seems stranger and "nerdier" to people than studying phenomena of lesser everyday relevance, as language and music are pretty unique in that everybody deems themselves a sort of expert in at least some forms of language and music, and has strong opinions and judgments about them. So the way specialised academics study language and music appears curious, especially given that schools do not give much insight into either field of study. People learn about the orthography and grammar of the dominant written language and perhaps one or a few others in a traditional way, and they learn a bit about the history of music and some other basics, and may learn to play an instrument and read notes, but neither has a lot to do with the way language and music are studied in academia.
This parallel was even more obvious prior to the decline of traditional and amateur music, especially acoustic home music, due to the rise of popular music and the entertainment industry thanks to the availability of recordings: most people could carry a tune, and many even play an instrument, which caused not only consuming, but also making music to be almost as natural as talking, and everybody could and can speak at least one language; at the same time, many or even most people were illiterate in musical notation, and many were even illiterate in reading and writing, let alone phonetic notations of the day.
Vocal pedagogues and many professional singers even use IPA! Actually, to learn IPA (especially more advanced IPA, where advanced phonetic knowledge becomes necessary) and musical notation (especially more advanced notation, where advanced theoretical knowledge becomes necessary) is pretty similar. If you cannot read notes at all, you would not expect to understand a treatise, or Wikipedia article, about a music-theoretical subject. Similarly, without a knowledge of IPA, you're lost in any article about a phonetical subject – even the phonetics of your own native language. --Florian Blaschke (talk) 22:57, 1 February 2015 (UTC)

Why bother reading this article if you do not understand IPA? It makes no sense. — Preceding unsigned comment added by (talk) 13:02, 8 October 2012 (UTC)

voiceless palatal fricative [ç] in English words[edit]

Whenever people on the net or wikipedia talk about this sound [ç] appearing in English, they only talk about words like e.g human, hue, huge, humid, humanity, humour etc.

What about words like: heat, heap, heed, heel, helium, haem etc. ? — Preceding unsigned comment added by Farazcole (talkcontribs)

There's usually very little if any palatal friction in them; consider also that for many (most?) speakers /iː/ is actually a diphthong [ɪi], so there's little reason to expect heat to have any more friction than hit. A. di M. (talk) 18:11, 26 November 2010 (UTC)
What about those who pronounce [i:] as a monophthong? —Preceding unsigned comment added by (talk) 18:19, 26 November 2010 (UTC)
Technically speaking the h in human isn't necessarily a palatal fricative; there is little acoustic distinction between a voiceless approximant and a voiceless fricative. On top of this, /h/ is often articulatorally placeless so that it adopts characteristics of neighboring vowels. In this sense, it's more palatal next to [i], more pharyngeal next to [ɑ], etc. In that sense it's more [ç]-like in words like heat than in words like heart but less like it than in words like huge. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 20:17, 26 November 2010 (UTC)
Oh yes. I know why only hue words are transcribed with a 'ç'. [h] must sound like the voiceless version of the vowel after it. In the word 'hue', the 'h' doesn't sound like the vowel (oo-sound) after it (unlike 'who'), and the [j]-sound that comes after it isn't a vowel, but a semivowel (and some people say hue is [çu:] and not [çju:] ). But in 'heat', the [h] sounds like the vowel after it, so it can be transcribed as [h].
In anology, in dialects that don't have the wine-whine merger, the 'wh' sound as in 'whine' is transcribed as [ʍ] because it doesn't sound like the vowel after it, but the 'wh' sound as in 'who' is transcribed as [h] because it DOES sound like the vowel after it. The whine-wine merger and the hue-you merger are similar to each other. Farazcole (talk) 18:26, 25 December 2010 (UTC)
I think you've got the gist of it. [h] is also described as being a voiceless version of neighboring sonorants so that both hue and whine can be described as having a cluster of /h/ and a sonorant. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 18:48, 25 December 2010 (UTC)
But do 'hue' and 'whine' have [j] and [w] sounds, respectively? Whine is transcribed as [ʍ], not [ʍw], and some people say that hue is [çu:] and not [çju:]. And is there any other allophone of [h], apart from its different sounds due to the sonorants next to it? —Preceding unsigned comment added by (talk) 10:02, 26 December 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── It kind of depends on speaker. While some people pronounce hue like [çuː], others pronounce it more like [çjuː]. In our Wikipedia transcriptions of English, we do away with [ʍ] altogether, since it looks like m, and transcribe words like whine with /hw/.

A number of people pronounce words like ahead with [ɦ], but other than that I can't think of any other allophones of [h] (in English, anyway). — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 16:54, 26 December 2010 (UTC)
I think you only MUST transcribe hue with a "ç" if someone pronounces it without a [j], but if someone pronounces hue with a [j] then you can use [h] in IPA or if they say 'heat' then you can use [h] in IPA, because the h-sound sounds like the neighbouring sonorant in the last two cases, but not the first one.
So.... should hue be transcribed as [hju:], [çju:] or [çu:]? — Preceding unsigned comment added by Farazcole (talkcontribs) 19:49, 26 December 2010 (UTC)
Do you mean for English? There's too much variation to have one authoritative pronunciation. Otherwise it's kind of up to your personal style. Here are some other options:
Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 20:57, 26 December 2010 (UTC)
But WHY is it written on the page that [h] becomes [ç]-like before [j] if it's MEANT to do that??? —Preceding unsigned comment added by (talk) 15:03, 27 December 2010 (UTC)
I don't think we've established that [h] is universally, or even crosslinguistically, [ç]-like before [j]. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 15:47, 27 December 2010 (UTC)
By [ç]-like I meant sounding like the voiceless version of [j] or [i] [voiceless palatal approximant). —Preceding unsigned comment added by (talk) 23:21, 27 December 2010 (UTC)
I'm not sure why. That's how literature explains it. If you want more explanation than that, I can only speculate. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 03:22, 28 December 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── But on the page it says:

"Although regional variation is very great across English dialects, some generalizations can be made about pronunciation in all (or at least the vast majority) of English accents: The voiceless stops /p t k/ are aspirated at the beginnings of words (for example tomato) and at the beginnings of word-internal stressed syllables (for example potato). They are unaspirated after /s/ (stan, span, scan) and at the ends of syllables. For many people, /r/ is somewhat labialized in some environments, as in reed [ɹʷiːd] and tree [tɹʷiː]. In the latter case, the [t] may be slightly labialized as well.[1] /h/ becomes [ç] before [j], as in human [ˈçjuːmən] or [ˈçuːmən]."

The last bit isn't true because [h] becomes a voiceless palatal approximant, not fully a fricative. It should be :[ç˕].

It depends. And, like I said, there's really very little difference between a palatal fricative and a voiceless palatal approximant. You'll probably find variation within individual speakers on that count. — Ƶ§œš¹ [aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi] 16:24, 28 December 2010 (UTC)
Well someone deleted the phrase that [h] becomes [ç] before [j], because some dialects drop the [h]. Even those who don't, not everyone pronounces the h as in hue as a complete fricative. The distinction between voiceless approximants and voiceless fricatives is very difficult anyway. —Preceding unsigned comment added by (talk) 17:07, 28 December 2010 (UTC)
A recent edit has produced the following sentence in Consonants: "In some dialects, /ç/ is allophonic with /h/. Examples are words such as hue [ç(j)uː].". Given what is said in the note numbered 4 above this, this sentence is both redundant and meaningless. I propose it should be deleted. RoachPeter (talk) 20:19, 21 June 2013 (UTC)
I've taken it out. — Lfdder (talk) 20:38, 21 June 2013 (UTC)

The Rhotic consonant in the phonemes section[edit]

Why is the sound transcribed as an alveolar trill, when only the Scots use this phone. If I'm not mistaking, the sound is (at least for General American) a retroflex approximant. There's not even a section in the chart for retroflex sounds, I know it's just one, but the author included a palatal column. Finally, what about using this symbol instead ɻ?≈ — Preceding unsigned comment added by (talk) 00:28, 10 July 2014 (UTC)

The reason is that we use the symbol /r/ to represent the phoneme. The other symbols mentioned are phonetic symbols used to represent different allophones of the /r/ phoneme.
re deletion of comments on allophones of /r/, I'm afraid I don't think these changes will do.
    • alveolar approximant [ɹ], the most common. (no citation to back this up)
    • labiodental approximant [ʋ], some Southern English speakers (e.g. Johnathan Ross). (no citation for "some Southern English speakers" - what about other parts of the world?; wrong spelling for Jonathan)
    • alveolar flap [ɾ], parts of Ireland (no citation; why only mention parts of Ireland?)
    • alveolar trill [r], many ESL speakers. (this article is not about the pronunciation of learners of English) RoachPeter (talk) 19:48, 14 August 2014 (UTC)

Looking at the latest version, I feel it is awkward that some of the information about /r/ is presented as Note 5 under Consonants while some more comes up in three bullet points under Sonorants. Is there a reason for dividing up the information in this way, or would a reorganization be desirable? RoachPeter (talk) 08:31, 7 October 2015 (UTC)

I agree. The note about transcription is still something to keep as a note under the table, but everything else can be covered in the sonorants section (which we can link to in note 5). — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 17:01, 7 October 2015 (UTC)

"Clear" and "dark" /l/ revisions[edit]

The account of "clear l" isn't improved by changing "before vowels" to "before or between". Being before a vowel is a sufficient condition for /l/ to be realized as clear, so the "or between" is redundant. The only restrictions that might need adding are (1) that the /l/ is most likely to be clear if it precedes a vowel within the same syllable, and (2) that /l/ will be realized as a voiceless [l] if preceded in a syllable by /p/, /t/ or /k/. The "other positions" referred to in the previous version as environments where dark /l/ is found then refer just to cases where /l/ precedes a consonant or is in final position. Saying that dark /l/ occurs after vowels is probably true in many cases, but it is what follows that matters. the /l/ in 'feeling', for example, has /l/ following a vowel, but the /l/ is clear. RoachPeter (talk) 10:42, 4 January 2015 (UTC)


The recent change of the transcription of "synchronic" from [sɪŋˈkɹɒnɨk] to [sɪŋˈkʰɹɒnɨk] and from [sɪnˈkɹɒnɨk] to [sɪnˈkʰɹɒnɨk] is wrong. There would never be an aspirated release of /k/ in such a context: perceptually, it would probably result in an extra syllable. If someone wants to add allophonic detail where it isn't really relevant, the ɹ symbol should have a diacritic beneath it to indicate that that the /r/ is voiceless. RoachPeter (talk) 11:01, 4 January 2015 (UTC)

There's a brief period of undisturbed airflow while my tongue moves from [k] to [r]. I don't think it'd be exactly wrong to transcribe [k] as aspirated. [But not in this article, obviously.] Alakzi (talk) 01:36, 8 January 2015 (UTC)

Child acquisition[edit]

The more I look at the section on Child Acquisition, the more I feel it doesn't belong in this article. While the phonology of children's English speech is clearly a topic of interest, this section is concerned with things like physiological explanations for differences from adult speech, types of intervention and various purely phonetic matters such as formant frequencies. There is very little in it that I would consider to belong to the topic of phonology. I would like to suggest moving this section out. There is a very good article on Language Acquisition, but surprisingly there is very little in it about phonology, so maybe that could provide a new home. RoachPeter (talk) 18:02, 8 January 2015 (UTC)

Looks as if nobody disagrees with me about this section, so I propose to remove it. RoachPeter (talk) 09:35, 11 January 2015 (UTC)
That makes sense. Is there a better place for it? — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 18:54, 11 January 2015 (UTC)
I was thinking that it might go into Language Acquisition, but taking a big chunk out of one article and inserting it into a different one is rather above my pay-grade, I think! As I have noted at [[3]] there is a real need for a phonology section in Language Acquisition, but there is already a rough draft of such a section sitting doing nothing in Talk:Language Acquisition. I don't think the Acquisition material in English Phonology on its own gives an adequate coverage. It's all a bit of a mess, and I'm not sure where to go from here. I suppose one solution would be for me to take Child Acquisition out of English Phonology and attempt to write some sort of synthesis of that and the draft in Talk:Child Acquisition, adding in my own limited knowledge of the subject where necessary, and putting that into Language Acquisition - but that would entail quite a lot of work. RoachPeter (talk) 10:20, 12 January 2015 (UTC)
I have now removed the Child acquisition material, and added a link to Phonological development, which is where such matters are properly covered. RoachPeter (talk) 12:37, 16 January 2015 (UTC)


The analysis of [ʍ] as underlying /hw/ is also supported by the realisation of Spanish ju (before vowels), i. e., [xw], in Caribbean and some Mexican accents typically [hw], as in Tijuana and Ciudad Juárez, as [w] in American English pronunciation, which is otherwise rather odd; at least I found it puzzling until I realised the connection with [ʍ] (this pronunciation, after all, still being retained by some American speakers, especially in the South). --Florian Blaschke (talk) 20:24, 1 February 2015 (UTC)


Tmesis is listed under the "onset" heading as an example of an unusual consonant cluster onset where the first phoneme is dropped and /tm/ is not listed under possible onset clusters. However, all dictionaries I've checked prefer the t to be pronounced. Longman has ˈtmiːs əs ˈmiːs-, $-əs, tə ˈmiːs-  ; Oxford /t(ə)ˈmēsəs/ ; Collins (təˈmiːsɪs; ˈmiːsɪs) ; Merriam Webster \(tə-)ˈmē-səs\ ; our own wiktionary /t(ə)ˈmiːsɪs/, /ˈmiːsɪs/ Afasmit (talk) 04:00, 21 March 2015 (UTC)

Affricate aspiration[edit]

As an Armenian speaker (Vahagn Petrosyan) told me, the English affricate /tʃ/ gets aspirated in the same way that stops /p t k/ do: that is, it's aspirated at the beginning of a syllable, and sometimes aspirated at the end: child [tʃʰaɪɫd], match [mætʃ] or [mætʃʰ]. I haven't seen any English phonetician note this, though. I can understand why, because a fricative release makes hearing aspiration difficult.

However, it's a big oversight phonologically, because affricates are similar to stops phonologically and can be expected to undergo similar phonetic changes under allophony. It's also a cross-linguistic faux-pas (so to speak), since Mandarin has an aspiration contrast in affricates, between /ts tɕ tʂ/ and /tsʰ tɕʰ tʂʰ/, and English clearly has an aspirated affricate initially, more similar to Mandarin /tɕʰ tʂʰ/ than /tɕ tʂ/ (excluding the place-of-articulation difference). Noting the phonetic similarity of these sounds is important, since there are many Mandarin speakers in the world, some of whom will no doubt be reading this article. Same with Armenian speakers, and speakers of other languages with aspirated affricates.

I could be wrong about English phoneticians not noting it, though. Peter238, do you know? — Eru·tuon 20:16, 1 April 2015 (UTC)

No source I'm aware of discusses this, including the latest Gimson's Pronunciation of English and Collins & Mees (2003), which both describe RP/GA consonants in great depth. The only thing they say is that /tʃ, dʒ/ have a mandatory fricative release, nothing about the aspiration.
Maybe one or more of these contains the information you're looking for:
- Roca & Johnson - A Course in Phonology
- Peter Roach - English Phonetics and Phonology
- John C. Wells - Accents of English volume 1
You'll probably have to go to the library though. If you want to save your time, check Roach's book first. I have limited access to the rest, and on pages I could access, there was nothing about the aspiration of /tʃ/. Peter238 (talk) 21:35, 1 April 2015 (UTC)
I looked briefly at Roach. (I know he's here on Wikipedia, and will probably read this.) Just like Collins and Mees, he doesn't mention aspiration of /tʃ/ in his section on fortis consonants, only of /p t k/. Both he and C&M mention glottalization of the affricate, but not aspiration. I'll see if I can get ahold of those other sources. I'd just write affricate aspiration into the article, but it would be OR, unless I see a source that mentions it. — Eru·tuon 01:14, 2 April 2015 (UTC)
I'm away from my books at the moment, but as far as I remember it, the picture is that for syllable-initial /p t k/ the glottis is opened extremely wide (as wide as for a sharp intake of breath) during the closed phase of the stop. The voice onset time is then the time taken for the vocal folds to come together in the following vowel. In a voiceless affricate we do not get such a wide opening (I can't remember a ref. on this just now, but I have observed this in my own larynx seen through a laryngoscope), and consequently the glottis is not so wide open at the onset of the vowel. I would certainly not say that there could never be any aspiration in English /tʃ/, but I don't think it's as noticeable as that for /p t k/. In clusters like /sp st sk/ the maximum glottal opening occurs nearer to the beginning (i.e. during [s]), consequently shorter (or zero) VOT. In clusters like /tm tn/ it's unlikely there would be regular devoicing of the nasal. You might get a pronunciation of 'at night' with the /t/ apparently moved to the beginning of the second syllable, but the great majority of the speakers of the accents we are looking at would glottalize the /t/, thus reducing the possibility of a delay in voice onset. I think I could add a bit to Aspirated consonant when I get enough refs. RoachPeter (talk) 14:58, 2 April 2015 (UTC)
Hmm, interesting. I have never observed aspiration from the articulatory perspective, only the acoustic one. So aspiration is articulatorily produced by glottal wideness at the end of the stop or affricate release, and English affricates have less wideness than stops, and therefore are less aspirated? That sounds reasonable, but still it's not the best comparison to make (apples to oranges). Better to compare initial affricates in English with those between vowels after a stressed syllable, or affricates in English with those in languages with an aspiration contrast, to see what the differences in glottal wideness are. Perhaps aspirated affricates have less glottal wideness after their release in Mandarin as well as English, but that there's still more glottal wideness than in tenuis affricates, so a clear distinction is made. Has this type of comparison been made? — Eru·tuon 20:57, 2 April 2015 (UTC)


I think we need a table of phones. I suggested this to Maunus on my talk page for the English language article, and he seems to be unconvinced, but it may be more appropriate here. The value of a phones table is to show all the possible sounds that actually occur in English. At the moment, the table only includes phonemes, and does not reflect the fact that most dialects of English have aspirated and preglottalized stops, a glottal stop, voiceless sonorants, that some dialects have an alveolar flap or trill, voiceless w, and so on. This is somewhat unhelpful to speakers of languages that have these sounds as separate phonemes, and recognize their occurrence in English.

Here's a phones table with notes. I'm not sure about the arrangement of the aspirated stops, and I didn't include preglottalized stops. Arrangement may need to be tweaked. Also, I'm not sure whether this table counts as OR, or is an appropriate synthesis of information provided by sources. — Eru·tuon 20:42, 1 April 2015 (UTC)


This table shows the consonant sounds in English. Phonemes are unmarked, but allophones are enclosed in parentheses, and dialectal phones are marked with asterisks.

Major consonant phones of English
Bilabial Labio-
Dental Alveolar Post-
Retroflex Palatal Velar Glottal
Nasal              m              n                ŋ
Stop     (p)   b t̪ʰ* ()*  ()*     (t)   d ʈ*   ɖ*     (k)   ɡ     ʔ
Affricate tʃʷʰ    (tʃʷ)   dʒʷ
Fricative f     v       θ    ð       s    z          ʃʷ     ʒʷ     x*     h
Flap              ɾ*
Trill              r*
Approximant               ɹ̠ʷ         ɻʷ (ç)    j        ʍ*   w
Lateral (ɫ)       l
  • Fortis stops and affricates /p t tʃ k/ are always voiceless. In most dialects, they are aspirated [pʰ tʰ tʃʰ kʰ] when they occur alone at the beginning of stressed syllables, as in pin [pʰɪn], but are unaspirated in other cases, as in spin [spɪn]. At the end of words, they are frequently preglottalized, as in nip [nɪˀp].
  • Lenis stops and affricates /b d dʒ ɡ/ are always unaspirated. In most dialects, they are partially voiced at the beginning and end of words, as in bin [b̥ɪn] and nib [nɪb̥], and fully voiced between vowels, as in about [əˈbaʊt].
  • The dental fricatives /θ, ð/ are lost in some dialects, and instead pronounced as labiodental fricatives [f, v], dental stops [t̪ d̪], or alveolar stops [t, d] (th-fronting or th-stopping). These pronunciations occur in Southern England, Ireland, and in African American Vernacular English.
  • The alveolar stops /t, d/ are often pronounced as retroflex [ʈ, ɖ] in Indian English. They are both pronounced as an alveolar flap [ɾ] between vowels in the United States, Canada, Australian, and New Zealand (intervocalic alveolar-flapping).
  • The alveolar stop /t/ is pronounced as a glottal stop [ʔ] before nasals in most dialects, as in button [ˈbʌʔ.n̩], and after vowels in Southern England, as in butter [ˈbʌʔə] and what [ˈwɒʔ]. The glottal stop also occurs in the interjections uh-oh and uh-uh.
  • The rhotic consonant /r/ is the approximant [ɹ] in most dialects, but sometimes a trill or flap [r ɾ] in Scottish, Irish, and Indian English.
  • Postalveolar and retroflex consonants /tʃ dʒ r/ are somewhat labialized (pronounced with rounded lips) in most dialects: [tʃʷʰ tʃʷ dʒʷ ʃʷ ʒʷ ɹ̠ʷ ɻʷ].
  • In RP, the lateral approximant /l/ is pronounced as clear or plain [l] before vowels, but dark or velarized [ɫ] after vowels at the end of syllables. In some dialects, dark l is pronounced as a labiovelar approximant [w], and in American and Scottish English most cases of /l/ are pronounced as dark [ɫ].
  • Conservative dialects like Scottish English contrast a voiceless [ʍ] in whine, typically analyzed as the sequence /hw/, with the voiced [w] in wine. The voiceless sound has merged with voiced [w] in most dialects.
  • [ç] is the pronunciation of the sequence /hj/ as in huge.
  • Sonorants /j, l, r, w/ become voiceless [ç, l̥, m̥, n̥, ɹ̥, ʍ] after voiceless obstruents (stops, affricates, fricatives): please [ˈpl̥iːz], Cockney [ˈkɒkn̥i].
  • The glottal fricative /h/ is lost in Cockney and Yorkshire English.

Since, in initial position, the English stop contrast is aspirate vs voiced, with tenuis being a conflation of the two, I would put [p t tS k] in parentheses and leave the aspirated stops unmarked.
There are many other phones in English. What are your criteria for what to include? — kwami (talk) 23:17, 1 April 2015 (UTC)
I guess I took [p t tʃ k] as the default forms, because they're used in phonemic representation, but I think your point regarding initial realization is a good reason to make the aspirated ones default.
No criteria; I included everything I can think of. It may as well be a complete list. I suppose some places of articulation for stops and nasals are missing. I'll modify the table accordingly. Let me know what's missing, and I'll add it. — Eru·tuon 01:09, 2 April 2015 (UTC)
Now that I'm adding all the random places of articulation, perhaps they would best not mentioned when they're just the result of assimilation. Perhaps the criterion could be whether the phones are noticeable, either in dialects or as allophones. Still kind of vague, but whatever. So, the table below can have all the random allophones, whatever their importance, but the one above will attempt to follow notability. — Eru·tuon 01:45, 2 April 2015 (UTC)

[ç] is the pronunciation of the sequence /hj/ as in huge.

This is not general in NAmEng, and the only sources I've seen that note it are about BrEng pronunciations. Curly Turkey ¡gobble! 05:12, 2 April 2015 (UTC)
Huh. I certainly say the word with [ç], or [çj], and Wiktionary agrees with me. What sound besides those would a NA speaker use? — Eru·tuon 06:28, 2 April 2015 (UTC)
I definitely don't pronounce it that way—and the difference when I hear a BrEng speaking say it is striking to me. The only sources for the Wiktionary entry are over a century old, so they obvious don't provide IPA. Curly Turkey ¡gobble! 06:39, 2 April 2015 (UTC)
How do you say it? Does the US pronunciation file (given on Wiktionary) sound odd to you as well? It seems to have [çj]: partially voiceless, partially voiced palatal continuant. — Eru·tuon 07:38, 2 April 2015 (UTC)
It doesn't sound strange, but (a) there's a distinct /j/ in there, and I think you'd be hard pressed to find a NAmEng speaker who didn't clearly pronounce it in all circumstances: /hj/ definitely doesn't collapse into [ç]; and (b) it could as easily be recording noise as palatalization, at least to my ears. I definitely don't pronounce the /h/ in "huge" the way I would the /h/ in Japanese hito [çi̥to] (which also sounds clearly more palatalized to me than what you may hear in the audio file above). Curly Turkey ¡gobble! 08:10, 2 April 2015 (UTC)
Okay. Well, phonetically speaking, there's no such thing as [h]: it's really just voicelessness of whatever sound precedes or follows. Hence, the /h/ in huge is realized as voicelessness of the sound that follows, /j/, which implies either [çj] or [ç]. If that's not the pronunciation, then there must be a pre-velar [x̟] or post-palatal fricative [ç₋] (a sound in-between velar and palatal) instead of [ç]: meaning, [x̟j] or [ç₋j]. There cannot actually be a glottal fricative, because that sound doesn't exist. Still, phoneticians might transcribe the pre-velar or post-palatal with the palatal symbol for simplicity, using the same symbol for a sound that's slightly different acoustically and articulatorily from the Japanese one. — Eru·tuon 19:31, 2 April 2015 (UTC)

Minor allophones[edit]

The table below is an attempt to show the "upper limit" of phones in English, the full number of phones that could be listed (maybe an impossible task). It's probably not helpful to readers, so the table above is a paring-back of this table to a list of phones that might actually be notable enough to mention to readers. This is a somewhat arbitrary criterion, but maybe we can refine it. Please comment only on the "major phones" table, not on this one. — Eru·tuon 20:24, 2 April 2015 (UTC)

Consonant phones of English
Bilabial Labio-
Dental Alveolar Post-
Retroflex Palatal Velar Glottal
Nasal              m              (ɱ)              ()                n                ŋ
Stop     (p)   b     () t̪ʰ*    ()*  ()*     (t)    d ʈ*   ɖ* ()  (c)  (ɟ)     (k)    ɡ     ʔ
Affricate (tsʰ)            tʃʷʰ    (tʃʷ)   dʒʷ
Fricative f     v θ    ð s    z          ʃʷ     ʒʷ     x*     h
Flap              ɾ*
Trill ()*    r*
Approximant        (ɹ̠̊)   ɹ̠ʷ        (ɻ̊)   ɻʷ      (ç)    j        ʍ*   w
Lateral (ɫ)  ()    l
Do you have a source for the voiceless nasals? AFAIK only /r, j, l, w/ can be devoiced and only in stressed syllables. Peter238 (talk) 02:03, 2 April 2015 (UTC)
Yes, I was going to say the same thing. Also, if you don't have assimilation in the top table, you wouldn't include [ç, l̥, ɹ̥, ʍ] after voiceless obstruents anyway. And what's going on here is assimilation to the aspiration: skew, splay, spray, squiggle do not have [ç, l̥, ɹ̥, ʍ]. Also, the dental d should not be in parentheses, should it?
The trill is also dialectical. Within the minor forms, you might want to distinguish a retroflex rhotic. Also, would we include Indian English? There are lots of Englishes out there. Or do we restrict ourselves to Englishes that descend as a mother tongue directly from England? — kwami (talk) 02:09, 2 April 2015 (UTC)
I'm not sure about the voiceless nasals; Maunus wrote them into the Phonology section of English language, with the example Cockney [kɒkn̥i]. I think you may be right, since sonorants are devoiced mainly by aspiration, and therefore voiceless sonorants occur only where aspiration does. Maunus's addition was probably mistaken, since we no longer have words like German Knie.
The confusing thing about dental [d̪] is that it's either from assimilation or dialectal (th-stopping). I suppose it would count as a dialectal phoneme or main allophone, but it's a minor allophone of alveolar d in most dialects in words like width, so that's why I used both brackets and an asterisk. Makes things confusing to include both minor allophones and dialectal phonemes.
Retroflex is technically included under postalveolar, but probably should add the retroflex approximant symbol and distinguish it from a retracted alveolar one. Yes, I think it would be best to include Indian and other Englishes significantly influenced by non-English phonologies. They are, after all, very important in number of speakers. I suppose that would require adding retroflex stops. Any other significant Englishes besides Indian? Nigerian, Singaporean? — Eru·tuon 04:50, 2 April 2015 (UTC)
  • Brinton and Brinton 2010 include voiceless nasals in their account of allophones, they state that all approximants and nasals are devoiced following voiceless consonants. Their examples are snore and smart. It is by the way the only source I have been able to find that gives an easily digestible overview of phonological processes in English.·maunus · snunɐɯ· 05:17, 2 April 2015 (UTC)
  • BTW I don't think a table of phones is a good idea. Especially not if it is not based on a clear source that gives such a table. A table of allophones is meaningful only if it is in the format phoneme/environment, so that it shows which phonemes have which allophones in which environments. A table of phones in a given language is pretty much meaningless unless it has some very specific criteria for which phones to include and why - there are probably very few phones that don't appear in some variety of English one way or the other.·maunus · snunɐɯ· 05:33, 2 April 2015 (UTC)
Not all existing phones occur in English, and not all are major allophones of phonemes, or occur in commonly spoken dialects, or are discussed by sources. As far as I know, some phones are rare or nonexistent: uvulars, pharyngeals, epiglottals, lateral fricatives, implosives. So, even when this table is complete, its inventory will be well short of the standard IPA consonant chart.
Certainly valid to question whether the table is WP:OR. I don't think it is, if all the phones shown are given by sources. Our lovely International Phonetic Alphabet chart for English dialects is cobbled together from many sources. I think sourcing can be handled by describing phones in the text below the table, and adding references there. — Eru·tuon 06:18, 2 April 2015 (UTC)
I don't actually think I have ever seen a list of phones given for any language. Allophones yes, but the problem with lists of allophones is that which ones seem "major" are just the ones that a given listener is likely to find particularly different from its prototypical pronunciation and that depends on one's own phonetical experience. Most English speakers would never hear the difference between a voiced and voiceless nasal, and the difference would seem minor, but for someone whose native language has that contrast it would seem like a major allophone. So my advice is, stick with conditioned allophones only, present them by conditioning environment, and stick to one accent and one source.·maunus · snunɐɯ· 15:11, 2 April 2015 (UTC)
@Maunus: Actually, voiceless sonorants are very rare cross-linguistically, and as far as I know, they don't occur in any of the top 10 languages by number of speakers. However, a tenuis–aspirated contrast in stops is pretty common, and occurs in four of the top ten languages, including Hindi and Mandarin, and a voiced-voiceless contrast occurs in at least three: Spanish, Hindi, and Portuguese. So only a handful of readers will have voiceless sonorants in their native phonemic inventories, while many will have a voiced–voiceless or tenuis–aspirated contrast. Hence, if we want to confuse the fewest readers, it's perfectly fine to omit voiceless sonorants from the allophone table, but very problematic to omit aspirated stops. — Eru·tuon 05:13, 3 April 2015 (UTC)
I’m afraid I have to agree with the reservations of Maunus above, in spite of all the work that has gone into this. It is likely to get very complicated, and the theoretical basis of what’s proposed seems rather unsure. What is proposed is not really a catalogue of phones (phones are independent of any phoneme), but a summary of the allophones of the phonemes of English. Such a summary is only useful if each allophone has an explanation of the context it occurs in (if it’s a context-dependent allophone) or the accent it occurs in (if it’s an allophone not found in the reference accent). I don’t think such explanation is possible in a compact table, and the result will be an ever-expanding list of not-very-readable footnotes. The best exposition of phonemes and allophones can be seen in the book that is the Bible for British phoneticians, Gimson’s Pronunciation of English, edited by Alan Cruttenden. The core of the book is a phoneme-by-phoneme description setting out all the allophonic variation found in British English, and this is a lot of material. It’s important to remember that the number of possible allophones (or phones) is infinite, so you need to have criteria for inclusion and exclusion. The heading “Minor Allophones” is confusing: what makes them minor? RoachPeter (talk) 10:05, 2 April 2015 (UTC)
I agree that the "minor phones" table is problematic. "Minor allophones" means those arising from simple assimilatory processes: labiodental stops and nasals near labiodental fricatives, dental stops and nasal near dental fricatives, voiceless sonorants after aspirated stops (and voiceless fricatives?). These are all cases where a sound changes a few phonological features to match a neighboring sound, and many speakers would not notice it.
Maunus, there is actually a major phones table in Icelandic phonology § Major allophones. That's where I got the idea for making such a table for English. To you and RoachPeter, I'm not proposing we add the "minor phones" table to the article, but rather the "major phones" table. The "minor phones" table is just an attempt to show the "upper limit" of phones we could list. I'm not sure which of your comments and criticisms are aimed at which table, so please comment only on the major phones table, in the section directly above this one. — Eru·tuon 20:16, 2 April 2015 (UTC)
The Icelandic situation is very different because there are different analyses of the inventory, and there is a broad literature about the potential analyses of preaspirated/voiceless sonorants. Note that it says "potentially contrastive phones (important phonetic distinctions which minimally contrast in some positions with known phonemes" - this means that there is doubt about what is the exact phoneme inventory of icelandic and this is what motivates the presentation of a table of phones that are potentially phonemic depending on analysis.·maunus · snunɐɯ· 02:31, 3 April 2015 (UTC)
That's rather like spider vs spied her in US English. [p t k] and the US tap are arguably major phones due to being conflations of phonemes. Someone might make a claim of phonemicity for [ç]. The others are harder to justify. And I've seen it argued that [ŋ] is not phonemic.
BTW, while snore might have a voiceless /n/ (I'd have run it through Praat to convince myself), I seriously doubt Cockney does, at least in US English. Perhaps they meant following a voiceless obstruent in syllable onset? — kwami (talk) 02:54, 3 April 2015 (UTC)
The question is not if the argument could be made, but if anyone has made it. As for Cockney I think that probably they didnt mean the rule to apply across syllable boundaries, I just didnt think of that when I chose the example. I've changed the example of the allophonically devoiced nasal to "snow". ·maunus · snunɐɯ· 03:02, 3 April 2015 (UTC)

You can use this link if you want to have an additional citation for devoicing of nasals. It states that Australian English /m, n/ are partially devoiced by a preceding voiceless consonant. Peter238 (talk) 12:54, 3 April 2015 (UTC)

Instead of having a table of allophones presented within the IPA consonant chart, could we have something like the table at Fortis and lenis? It would give the context, which is sadly missing from the above proposals. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 03:42, 4 April 2015 (UTC)
Whether we have the IPA-style chart or not, I definitely think we need something like the fortis–lenis table, since a graphical representation is more helpful than simply discussing allophones in bullet points. — Eru·tuon 04:55, 4 April 2015 (UTC)

Ordinary English pronunciation[edit]

Including only the linguistic version of pronunciation is the single worst feature of English Wikipedia. I'll wager that no more than 30,000 of the 300,000,000 native speakers of English understand this specialized notation. Despite this, Wikipedia does not furnish the reader an intelligible alternative way to pronounce words. For example: consider Wikipedia's pronunciation of the word "English: "/ˈɪŋɡlɪʃ/". This tells me nothing. However, Wiktionary adds an easily-understood alternative: "enPR: ĭngʹglĭsh". How could it harm anything to include a comprehensible pronunciation? For experts, inclusion of the more technical and precise pronunciation would still be included, just as it is in Wiktionary.

The two following easily-pronounced words are merely two out of millions of possible examples, rendered in symbols comprehensible only to specialists: --football: /ˈfʊtbɔ(ː)l/ --Jersey /ˈdʒɜrzi/

Football: IPA(key): /ˈfʊtbɔ(ː)l/ ing' -glish or, often,-lish

Jersey /ˈdʒɜrzi/

Barrister noir (talk) 16:15, 21 August 2015 (UTC)

You fail to appreciate the unique logic behind the system. We use the specialized (IPA) symbols rather than comprehensible (respelling) transcriptions so that our transcriptions are accessible to everyone, not just to English native speakers. However, when it's pointed out that we use the IPA symbols in a bizarre self-invented way that will mislead anyone who actually knows the standard system, we're told that it's only in fact English native speakers who matter, and they won't be misled because they don't know the system anyway. Mad I know, but don't bother trying to argue about it - I know, I've tried - the ruling clique of Wikipedia linguists has made up its collective mind about this. (Actually it is possible to add comprehensible respelling transcriptions if you want, although there is also a slightly bizarre self-invented system for those, just to prevent anyone from making things too easy.) W. P. Uzer (talk) 17:28, 21 August 2015 (UTC)
Barrister, I think the place you're looking for is Help talk:IPA for English. While W.P. is right that doing away with IPA would be met with much resistance, I should note that there is a respelling system similar to what you've proposed at WP:RESPELL that is used alongside IPA transcriptions. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 19:04, 21 August 2015 (UTC)

Word-level rules?[edit]

I think the following point is slightly vague verging on wrong: Sequences of /s/ + C1 + V̆ + C1, where C1 is a consonant other than /t/ and V̆ is a short vowel, are virtually nonexistent. Spit? Stop? Spot? Slim? Skip? Scum? Sludge? (talk) 18:42, 31 August 2015 (UTC)

In the examples you cite, your second and third consonants are not the same. The "rule" described says that we don't have words where they are. So we don't have words like spip, scack, or slull. But I squint my eyes at this as a "rule." Seems more like a distributional overlap. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 18:57, 31 August 2015 (UTC)

What English nouns begin with /ð/?[edit]

What English nouns begin with /ð/? It seems to occur in many basic function words (the, that, those, they, with) but I can't seem to think of a single noun where it occurs initially. Equinox (talk) 15:10, 3 December 2015 (UTC)

How about thank?
You can check out this category list at wiktionary and see if any other th nouns listed there start with eth. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 15:40, 3 December 2015 (UTC)
All of them, but only if you speak with a thick West Country accent. I don't know any in standard English, and I'm sceptical about "thank" (unless it's a part of your regional accent). Peter238 (talk) 16:00, 3 December 2015 (UTC)
You might be right about that. I'm going to do a lot of nice things for people today and see how they say "thank you." — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 16:16, 3 December 2015 (UTC)
@Equinox: I'm pretty sure you're right, that only function words have initial /ð/. The contrast between [ð] and [θ] was allophonic in Old English, with the voiced sound occurring between voiced sounds, and the voiceless sound at the beginning of a word or when geminated (probably near voiceless sounds too?). Function words didn't count as their own phonological word, so they developed voiced initials. With the loss of final unstressed /ə/ in late (?) Middle English, a phonemic contrast between [ð] and [θ] developed at the end of the word (for instance, bath and bathe; the last one used to end in a vowel), and then the function words got their voiced [ð] allophone fossilized as a phoneme too. The same thing happened with the former allophones [s z] and [f v]. Relevant pages: Old English phonology § Intervocalic voicing, Phonological history of English consonants § Voiced/voiceless splits. It doesn't seem like this topic is described in much detail, unfortunately... — Eru·tuon 08:26, 13 September 2016 (UTC)

English phonology#Sonorants[edit]

As an Australian I disagree with the statement that in Australia /l/ is always dark. I say a word such as "lull" as [laɫ] where the final "l" sounds somewhat like a "w". In fact, deliberately mispronouncing it as [law] sounds almost exactly like "lull". While saying the first "l" the tip of the tongue touches the roof of the mouth but the final "l" has the tongue somewhat retracted. Can someone find some recent research to confirm this for Australian English? — Preceding unsigned comment added by Danielklein (talkcontribs) 13:22, 18 January 2016 (UTC)

I think I sometimes pronounce final l the same way, not touching my tongue to the roof of my mouth. (I speak American English.) That is an example of l-vocalization, a sound change that in English operates on dark l, not clear l. — Eru·tuon 00:55, 4 September 2016 (UTC)

Shouldn't we replace /r/ in the consonant table to [ɹ̠] or something else if it's almost never pronounced as [r]?[edit]

I know the reason is given below as "conventionally transcribed with the basic Latin letter ⟨r⟩", but since all the other symbols in that table are IPA's, using r there seems very confusing for readers.--fireattack (talk) 19:59, 2 October 2016 (UTC)

Are you sure it would be confusing? The only people I think who might be confused are new phonetics students who have just learned the IPA. And the confusion would be more of why were are using this convention, rather than whether English really has a trilled rhotic. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 20:16, 2 October 2016 (UTC)
>the confusion would be more of why were are using this convention
Yeah that's exactly what I meant, sorry for not being clear. --fireattack (talk) 06:17, 5 October 2016 (UTC)
I think it would be helpful to use the coronal approximant symbol, because it would show the fact that English usually has a different rhotic from other languages like Spanish and Russian, which actually have a trill. But this is a discussion that would have to take place at Help talk:IPA for English, since it concerns the Wikipedia diaphonemic transcription system. — Eru·tuon 20:22, 2 October 2016 (UTC)

In General American, the vowels [ə], [ʌ] and [ɜː] may be considered a single phoneme.[edit]

@Mr KEBAB: @Aeusoes1: This is information may be sourced. However, a user once said that:

  • If I've learned one thing over the years, it's never trust what British authors have to say about American English. (The exception is John C. Wells, but even with him you have to be a little careful.)

Therefore I am a little hesitant to make the claim that /ʌ/, /ə/, and /ɜː/ can be considered one phoneme in American English. For the same reason, I'd rather have an American source before we make that claim.LakeKayak (talk) 20:14, 22 January 2017 (UTC)

I think you need a more compelling reason to remove that information. I've contacted the editor who made that contribution (@Victor Yus:) because it sounds false, but close to a claim I encounter from time to time. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 20:23, 22 January 2017 (UTC)
Thank you.LakeKayak (talk) 22:25, 22 January 2017 (UTC)
Which part of it do you think false? You say yourself (and Wells and others seem to agree) that GAm lacks /ɜː/ as a phoneme (presumably that means that NURSE contains the /ʌ/ of STRUT). That /ʌ/ and /ə/ may be considered a single phoneme appears on Wells' p. 132. (I can't access online the page that is actually referenced in the text, or some of the other pages which might shed light on the matter.) W. P. Uzer (talk) 21:49, 22 January 2017 (UTC)
I like the way you've clarified it. I'm just hoping this is what Wells actually says. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 14:59, 24 January 2017 (UTC)
@W. P. Uzer: I can cite the relevant passage from p. 132, vol. 1, of Accents of English: In Wales, and in some (higher-prestige) midlands and north-of-England accents STRUT words have stressed [ə], in consequence of the STRUT-Schwa Merger (vol. 2, 5.1.3, 4.4.2). A central [ɜ] is found in some American southern accents. Even in GenAm it may well be considered that stressed [ʌ] and unstressed [ə] are co-allophones of one phoneme. The referred §5.1.3 (Wales) and §4.4.2 (northern England) are on pages 380-381 and 352-353 respectively. On p. 486, vol. 3, Wells gives a vowel diagram, where the places of the articulation of American /ʌ/ and /ɜ/ are shown clearly distinct.--Lüboslóv Yęzýkin (talk) 09:09, 25 January 2017 (UTC)
So the [ɜ] that Wells mentions is a diaphonemic variant of /ʌ/ and Wells doesn't even mention the NURSE vowel.
Which makes me wonder about the placement of this note, because it's designed to explain a feature of a table that is now uncited. Do we have sources that argue that the vowels of STRUT and NURSE are the same? — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 14:49, 25 January 2017 (UTC)
@Любослов Езыкин: I'm having trouble reconciling your above quotes with this edit. Where does Wells say that the vowel of NURSE is a pre-rhotic STRUT in American English? — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 01:48, 26 January 2017 (UTC)
Possibly on the page originally cited (p. 121), which Google Books seems not to provide online. I can try to look it up tomorrow if nobody has a copy of the book handy. OK, I looked it up in the library, what we have seems to be basically correct. I can provide more detail if anyone likes, but it would probably go beyond what's suitable for a note under the table (perhaps more suitable for the General American article). W. P. Uzer (talk) 19:53, 26 January 2017 (UTC)
"@Aeusoes1: See the citations below. The only contradiction I see here is the vowel chart on the following page 486 where /ʌ/ and /ɜ/ are located separately. But I believe more Wells's detailed explanations than charts.--Lüboslóv Yęzýkin (talk) 16:59, 27 January 2017 (UTC)
Ah, thank you. I am satisfied. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 01:35, 28 January 2017 (UTC)

@Aeusoes1: My real problem was that I think I hear a distinction between the three vowels. The page General American says that the diaphoneme /ɚ/ and /ɝ/ do merge. However, the page states a clear distinction between /ʌ/ and /ə/. So, it seems that this line alone may be inaccurate. To check its accuracy, I will look at the ANAE. Until then, I will not touch that information.LakeKayak (talk) 22:43, 22 January 2017 (UTC)

I have a suspicion that Wells is saying that the vowels we transcribe as /ə/ and /ʌ/ are the same (much like how [ɚ] and [ɝ] are the same phoneme) and that what we transcribe as /ɜːr/ would be a biphonemic cluster composed of this same vowel phoneme and /r/. I've seen people use ⟨ə⟩ for all three and it's a common enough claim, most notably highlighted in how the vowels are in complementary distribution from each other, that it's not all that controversial, as long as it's clear. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 19:03, 23 January 2017 (UTC)

In my speech, I think //ɪ// merges with //ə// (that is, stressed //ɪ// is //ə//). If so then //ʌ// cannot be merged with //ə//, or I would have no distinction between pin and pun. But this is just anecdotal unless a source says the same thing. — Eru·tuon 04:43, 23 January 2017 (UTC)

Did you mean to write "unstressed //ɪ// is //ə//"? This (the weak vowel merger) wouldn't necessarily invalidate the claim, though, since it could be interpreted as a neutralization applicable to weak syllables. W. P. Uzer (talk) 07:54, 23 January 2017 (UTC)
W. P. Uzer: No, I'm aware of the weak-vowel merger, and I am referring to a merger of both stressed and unstressed //ɪ// with schwa. — Eru·tuon 23:22, 23 January 2017 (UTC)
Are your parents from New Zealand by any chance? I've never heard of KIT-COMMA merger in North America. Mr KEBAB (talk) 00:35, 24 January 2017 (UTC)
Do you have the same first vowel in mirror and hurry? W. P. Uzer (talk) 06:58, 24 January 2017 (UTC)
@Mr KEBAB: No, my parents are both native Americans. My mom spent part of her childhood in Canada, so perhaps she was subtly influenced by the Canadian Vowel Shift (though her pronunciation doesn't have Canadian raising of ow).
@W. P. Uzer: No; I would transcribe my pronunciation of mirror as /ˈmiɹəɹ/ and hurry as /ˈhəɹi/. That is, the result of the mirrornearer merger should be identified phonologically as the FLEECE vowel, not the KIT vowel; the sound changes that happen to KIT do not happen to mirror. — Eru·tuon 22:23, 24 January 2017 (UTC)
@Erutuon: I see. Doesn't the Canadian VS shift [ɪ] to [ɛ] though (as the California VS does)? If what you're saying is true, I think that the reason for which you have the KIT-COMMA merger must be different. Mr KEBAB (talk) 00:44, 25 January 2017 (UTC)
@Mr KEBAB: Well, I really don't know much about the Canadian Vowel Shift, but the article says that KIT vowel is lowered or retracted according to different writers, perhaps depending on the region. My KIT vowel is both lowered and retracted, so maybe it's consistent with the shift and maybe not. — Eru·tuon 03:23, 25 January 2017 (UTC)
@Erutuon: Aren't you from the North by chance?--Lüboslóv Yęzýkin (talk) 12:08, 28 January 2017 (UTC)
@Любослов Езыкин: Yes, I'm from the North, but my accent is different, more North-Central if anything, though not very strongly so. I have cotcaught merger, which is not a feature of Northern Cities as I understand it. You can see a description of my vowels on one of my userpages. — Eru·tuon 22:22, 28 January 2017 (UTC)

Citations from Wells[edit]

OK, to finally clear the issue I decided to provide all the relevant passages from Accents of English by John C. Wells.

P. 121, vol. 1
If we treat [ɝ] as underlying /ɜr/, it is logical to treat [ɚ] as underlying [ər], thus further /ˈfɜrðər/ [ˈfɝðɚ].

The opposition between [ɜr] and [ər] is tenuous and may be absent (a possible minimal pair is foreword /ˈforwɜrd/ vs. forward /ˈforwərd/). This raises the further possibility of treating ‘/ɜ/’ and /ə/ as phonologically identical, so that [ɝ, ɚ] would be taken as underlyingly /ər/ (further /ˈfərðər/). And since there may be no real opposition between /ʌ/ (love /lʌv/) and /ə/, there is also the possibility of analysing [ɝ, ɚ] as /ʌr/.

P. 132, vol. 1
In Wales, and in some (higher-prestige) midlands and north-of-England accents STRUT words have stressed [ə], in consequence of the STRUT-Schwa Merger (vol. 2, 5.1.3, 4.4.2). A central [ɜ] is found in some American southern accents. Even in GenAm it may well be considered that stressed [ʌ] and unstressed [ə] are co-allophones of one phoneme.
P. 480–481, vol. 3
In words such as furrow, worry, hurry, courage, we would expect

to find /ʌ/, i.e. the vowel of STRUT, thus /ˈhʌri/ hurry, etc. It can be argued that phonemically speaking this is what we always do find in American accents. Phonetically, though, [ʌ] plus /r/ quite rare, being virtually restricted to New York City and scattered parts of the east and south. Elsewhere we get an r-coloured mid central vowel [ɝ], thus [ˈhɝi], or else a vowel of the [ɜ] type plus /r/, thus [ˈhɜɹi]; in each case this is phonetically the same as in words such as nurse [nɝs ~ nɜɹs], as can be seen from the fact that in such accents hurry rhymes with furry. Phonemically, one of two conclusions can be drawn: either hurry, etc., have the same /ɜ/ as NURSE, and we write /nɜrs/, /ˈhɜri/ (or, with Kenyon & Knott 1953, /nɝs/, /ˈhɝi/); or, alternatively, [ɜ] can be regarded as a positional allophone of /ʌ/ and [ɝ] as a realization of /ʌr/, from which it follows that there is no separate phoneme /ɜ/ (or [ɝ]) in this most usual type of GenAm. The latter seems the most economical analysis, and is logically unassailable; it means that nurse [nɝs] is to be phonemicized /nʌrs/. (Since GenAm also usually lacks a proper opposition between [ʌ] and [ə], it follows that the phoneme may equally well be written /ə/; which leads to the phonemicization of hurry [ˈhɝi] as /ˈhəri/. Following my usual practice of taking the pronunciation of the simple keyword as the guide to appropriate notation, though, I shall write the GenAm STRUT phoneme as /ʌ/, and hurry, therefore, as /ˈhʌri/. In view of the above argument, it is admittedly inconsistent to write NURSE words in GenAm with /ɜr/, as I have done everywhere else in these volumes; but it makes for easier comparison with other accents.)

I hope this helps.--Lüboslóv Yęzýkin (talk) 16:51, 27 January 2017 (UTC)

In certain American English dialects, the diphthongs /ɪə/ and /eə/ can be found in words such as ideas and rail, respectively.[edit]

I have a question about /ɪə/. In which dialects of American English is this vowel class not used for the "ea" in "idea"? Thank you.LakeKayak (talk) 22:49, 25 January 2017 (UTC)

Some dialects (maybe even most) definitely realize the "ea" in idea as [i:ə] (i.e. "eye-DEE-a"), with a longer middle vowel. Although maybe it gets closer to [ɪə] in rapid speech. Someone who is better with U.S. dialects will hopefully be able to provide more details. ¡Bozzio! 11:07, 27 January 2017 (UTC)
Note however that the article specifies the plural ("ideas"), and adding the /-s/ can definitely alter vowel length. ¡Bozzio! 11:09, 27 January 2017 (UTC)
@Bozzio: I did want to avoid examples of /ɪə/, being pronounced as [iːə]. This is a different production with the same realization. This is not a different vowel class being used altogether.LakeKayak (talk) 21:53, 27 January 2017 (UTC)

Page archiving[edit]

Is there a way to set up automatic archiving for this page? There are currently 62 sub-sections, which makes navigation difficult. ¡Bozzio! 11:02, 27 January 2017 (UTC)

Done with a bot. More than 3 years old discussions are supposed to be archived soon.--Lüboslóv Yęzýkin (talk) 17:29, 27 January 2017 (UTC)
@Любослов Езыкин: Hmm, it would have been better to use the existing archive box labeled by dates somehow... — Eru·tuon 20:36, 27 January 2017 (UTC)
I thought about making the archive into a chronological order but given the fact we have already got "Archive 1" and "Archive 2", I was not sure how the bot would react if I order it to make archives by year, especially when many discussions stretch for many years (see some examples above, where a topic started in 2009 got answers in 2015). So I ordered it to continue the established pattern. Anyway, feel free to correct.--Lüboslóv Yęzýkin (talk) 11:23, 28 January 2017 (UTC)

/ts/ or /t͡s/?[edit]

Is there a set rule where the sequence of /t/ and /s/ is pronounced as the affricate /t͡s/? I've observed that it only occurs in a syllable coda (e.g. cats, pits), and not in an onset (e.g. tsunami, as pronounced by most Americans) or between syllable boundaries (e.g. sight·seeing, but not Pitts·burgh). (talk) 03:41, 17 July 2017 (UTC)

It may be a phonetic affricate [t͡s], but it's not phonemic /t͡s/. Mr KEBAB (talk) 06:22, 17 July 2017 (UTC)
Still, there are rules governing whether it's pronounced with a [t] + [s] sequence. In the examples you gave, the syllable boundaries help govern the pronunciation. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 14:16, 17 July 2017 (UTC)

External links modified[edit]

Hello fellow Wikipedians,

I have just modified one external link on English phonology. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

You may set the |checked=, on this template, to true or failed to let other editors know you reviewed the change. If you find any errors, please use the tools below to fix them or call an editor by setting |needhelp= to your help request.

  • If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
  • If you found an error with any archives or the URLs themselves, you can fix them with this tool.

If you are unable to use these tools, you may set |needhelp=<your help request> on this template to request help from an experienced user. Please include details about your problem, to help other editors.

Cheers.—InternetArchiveBot (Report bug) 08:18, 21 September 2017 (UTC)