Talk:English phonology

From Wikipedia, the free encyclopedia

You have new messages (last change).
This article is part of WikiProject Phonetics, an attempt to build a comprehensive and detailed guide to phonetics and descriptive phonology on Wikipedia. If you would like to participate, you can edit the article attached to this page, or visit the project page, where you can join the project and see a list of open tasks.


I have deleted the following passages, because they were confusing, irrelevant or inaccurate. If anyone would like to restore any of them, let's talk about it here first.

  • None of the twenty-six letters of the Roman alphabet used in the English-speaking world can be relied upon for consistency. The letters, as mentioned above, can have either hard or soft sounds. They can also be used in combinations that either make multiple pronunciations possible (like the English words hegemony and phthisis, with nine possible pronunciations each) or exceptional (like the bird known as a chough, which is pronounced chuff). The most common vowel sound in English, in fact, isn't even a standard vowel sound; the schwa, printed as an upside down and backwards e, is more of a murmur. But the schwa appears as at least one vowel sound in almost every multisyllabic word of English.
  • Most linguistic authorities agree that English possesses more sounds than any other language; [...] By contrast, Italian has just twenty-seven sounds and Hawaiian only thirteen.
  • Part of the complexities of speaking English "correctly" comes from the tendency of English-speakers to slur words and expressions together that they are familiar with. Often this manifests itself in the spellings (for example, glimpse was originally glimsen, messenger used to be messeger, and pageant was pagean) due to subconscious inclusion of a sound; over time, the listeners to these spoken words became aware of something missing -- a letter or letter combination to "complete" the word.
  • Perhaps the most pronounced change in English pronunciation came in an almost imperceptible way to those who lived through it.

--Angr 19:03, 21 Feb 2005 (UTC)

I came to this page through a redirect from "English Pronunciation". Although the article, as it stands now, does a good job (how good I don't know, I really don't know much about phonology) for its title, what I was looking for when I got here was the kind of thing that's touched upon on that first paragraph you deleted; I wanted to see some comentary on all the irregularities or discrepancies between writing and pronunciation. If this is not the correct article to discuss this, can someone please create the proper article? -- tmegapscm, 2005-08-25

Try English orthography.
Peter Isotalo 22:52, 31 August 2005 (UTC)

Contents

[edit] Frequency

Vowel Frequency

Vowel %
ə 10.74
ɪ 8.33
ɛ 2.97
aɪ 1.83
ʌ 1.75
eɪ 1.71
i 1.65
əʊ 1.51
æ 1.45
ɒ 1.37
ɔ 1.24
u 1.13
ʊ 0.86
ɑ 0.79
aʊ 0.61
ɜː 0.52
ɛə 0.34
ɪə 0.21
ɔɪ 0.21
ʊə 0.21

I found this table at page 239 - "The Cambridge Encyclopedia of the English Language" by David Crystal, published by Cambridge University Press 1995. Would it be a copyright infringement if it were used? There is a similar table for consonants on page 242. Izehar 15:11, 26 November 2005 (UTC)

According to Wikipedia:Verifiability, only what can be found in reliable sources can be used - does that mean that we are violating someone's copyright every time we cite a source? Izehar 15:13, 26 November 2005 (UTC)
I don't think there's a problem with using the table as long as it's properly cited. If it's included it should be made clear that it applies to Received Pronunciation (I assume)... Quite a large number of English speakers don't have a number of those vowels and diphthongs at all (I, for instance, have a 0.00% frequency for [əʊ], [ɒ], [ɜː], [ɛə], [ɪə], or [ʊə], but I do have a number of vowel phonemes not listed here at all (see here for my manifesto on that, if you're really interested))... Cheers, Tomertalk 02:32, 14 February 2006 (UTC)

[edit] Factual Accuracy

From the article: "The following short vowel sounds cannot occur without a coda in a single syllable word: /e/, /æ/, /ɒ/ and /ʌ/". What about the word "yeah". That ends in æ. ColinKennedy 17:35, 11 January 2006 (UTC)

It may be a little too strong, but it's not that far off. (Yeah is also an exception for me, but it has /ɛ/, so it's like yet without the /t/.) See checked and free vowels, which does give a caveat.--JHJ 18:09, 12 January 2006 (UTC)
"yeah" [jæ:] is definitely an exception in Wisconsin, and, in fact, the only exception (based on my personal original research :-p). I've never heard anyone pronounce it [jɛ], and I'm pretty sure that if I did hear it, that I'd assume "you" was meant, not "yeah". Tomertalk 22:05, 1 February 2006 (UTC)
Also what about "paw," "saw," etc? They end in /Q/.
The above words can be thouɡht of as ending in /ɔ/ as opposed to /ɒ/. Also, as I pronounce it, ˈyeahˈ sounds like [jæə].ChillinChaz 00:19, 18 October 2006 (UTC)

I've amended the wording to re-assert that it refers to RP. All of the vowels you're referring to are long in RP. Gailtb 08:36, 6 May 2006 (UTC)

[edit] non-onset /r/

Are the following sentences in the current version correct?

In the Nucleus section:

The following can occur as the nucleus:
/r/ in rhotic varieties of English (eg General American) in certain situations

and in the Coda section:

The following can occur as the coda:
The single consonant phonemes except /h/, /w/, /j/ and, in non-rhotic varieties, /r/

If the nucleic /r/ means /ɝ/ or /ɚ/, then I think it is not a consonant but a vowel. Likewise, the coda /r/ is not a consonant but a part of a diphthong such as /ɪɚ/. - TAKASUGI Shinji 08:11, 6 April 2006 (UTC)

I think most people agree that the nucleus of rhotic bird is [ɝ] and the coda simply [d], but whether the rhotic portion of beard is in the nucleus or the coda is still unclear, with evidence from different phenomena pointing in different directions. I wrote a paper once arguing that after a high vowel, r is in the coda, but after a non-high vowel it's in the nucleus (evidence was from language games in American English), so that the r of beard is in the coda but the r of board and bard is in the nucleus. Similar results are found in Abigail C. Cohn's dissertation, where she measures how quickly nasalization stops after a nasal consonant: the [ɪ] portion of the diphthong in noise still has slight nasalization, while the [o] of disyllabic neo has none; similarly, the [r] portion of north still has slight nasalization, while the [r] portion of near has none. On the other hand, fluent backward talkers treat diphthongs and affricates as units (backward-talk of join is [nɔɪdʒ]) but treat r-colored vowels as two segments (backward-talk of turn is [nrʌt]. So it's still clear, and for this article, it's probably safest to say [r] appears in both the nucleus (e.g. bird) and the coda (e.g. near) in rhotic accents. Angr (talk • contribs) 08:34, 6 April 2006 (UTC)
Thank you very much for your explanation. Isn't it because of the spelling that turn becomes nrut in backward-talk? Bird becomes drib, perhaps? The reason why I asked the question is that I'd like to import a vowel chart in International Phonetic Alphabet for English to the Japanese version of Wikipedia. I'll add a note on the ambiguity of [ɹ]. - TAKASUGI Shinji 15:51, 6 April 2006 (UTC)
I don't know what backward-talk for bird is, and I'm willing to believe [nrʌt] is at least partially influenced by the spelling, but the speakers in the article I read (I don't have the ref to hand I'm afraid) were otherwise reversing phonemes without regard to spelling (mice -> [saɪm], not to anything that could be rendered ecim). Angr (talk • contribs) 16:02, 6 April 2006 (UTC)
It's interesting both Shavian and Deseret alphabets treat [ɝ] as a sequence of a vowel and /r/. This suggests some native English speakers think [ɝ] is not a single phoneme. - TAKASUGI Shinji 16:26, 6 May 2006 (UTC)

[edit] add audio files ?

It may be useful, especially for non-native english speakers, to add some ogg audio files. 86.204.29.12 11:27, 3 May 2006 (UTC)

[edit] What about 4 consonant codas?

For example "glimpsed".

What about them? Sixths is another one. Angr (talk • contribs) 07:47, 6 May 2006 (UTC)

4-consonant codas which are formed with the -s/-z or -t/-d morphemes as the fourth consonant are explained in the introduction to the section on codas. Gailtb 08:24, 6 May 2006 (UTC)

Oops, missed that... Linguofreak 15:49, 7 May 2006 (UTC)

[edit] Non-Assimilated Onsets?

Note: A few onsets occur infrequently making it uncertain whether they are native pronunciations or merely non-assimilated borrowings, e.g. /sv/ (svelt), /sr/ (Sri Lanka), /vr/ (oeuvre), /ʃw/ (schwa), /smj/ (smew), and /sfr/ (sphragistics).

For /shri lanka/, there is also shrink, shroud, etc. For (sphragistics), what happens to sphinx and sphere? They don't seem to be in the table, but they are definitely assimilated. user:Phillipmackin 30/05/06

I think Sri Lanka is being assumed to have /sr/ rather than /ʃr/ like shrink, shroud, etc. I've heard both, and /ʃr/ is actually closer to the Sanskrit pronunciation. As for sphragistics, the point is that it has /sfr/; sphere and sphinx have only /sf/. Angr (talk) 19:55, 30 May 2006 (UTC)

Couldn't /vl/ as in Vladimir also go into this list? It is, as far as I know, pronounced /vl/ by most English speakers, but does not otherwise occur in English. Xyzzyva 15:32, 6 August 2006 (UTC)

[edit] æ-Tensing in RP

I have removed the following addition pending verification:

Old recording of British RP speakers (for instance, mid-century recordings of Queen Elizabeth) display evidence that a similar system existed in earlier versions of RP. However, such pronunciations died out after World War II.

In the absence of the citation of reliable sources, this sounds to me like original research. User:Angr 19:47, 2 August 2006 (UTC)

[edit] Minimal sets for onset contrasts?

I was wondering whether it might be helpful when giving example words for the various onsets to keep the codas and nuclei consistent (whenever possible). For instance, for "Voiceless fricative plus approximant other than /j/", the article could give "flee, sleep, free, three, shriek, sweep, thwack (for lack of a better word) and wheat" instead of the current examples. Then, for "/s/ plus voiceless plosive", it could give "speak, steed, ski", and so on. We could also do something similar with the coda lists. I think this would better highlight the contrasting sounds, but my concern is that it might make the section as a whole look too singsongy or hard to read. Does anyone else have an opinion? -- Calcwatch 04:21, 22 December 2006 (UTC)

[edit] Length and strengths

Is it correct to say that these words can have /ŋkθ/ or /ŋθ/ depending on the dialect? My own reasoning would say it is /ŋθ/ (ie the adjective with ablaut plus the nominal suffix /θ/) with an epenthetic [k]. Cf /wɔːmθ/ [wɔːmpθ]. What does the literature say? And if this is so, is there a principled reason for saying that words like prompt and glimpse have 3 consonants rather than 2 in their phonemic representation or do we just work on the basis of spelling? Gailtb 06:42, 27 January 2007 (UTC)

And if the [k] is epenthetic? It's still pronounced. OED has both pronunciations. Possibly it's justified by the 'g'; I pronounce the [k] in 'length' but not the [p] in 'warmth'. As for prompt, glimpse, everything I've looked at transcribes them with 3 consonants. –EdC 12:07, 27 January 2007 (UTC)
Possible justifications for the current /ŋθ/ vs. /ŋks/ (etc.) transcription:
  1. coda /md mz ŋd ŋz/ do not (AFAIK) become [mbd mbz ŋgd ŋgz]
  2. the rare cluster [mt] in dreamt (I don't think that's included in the list yet, BTW, but it's arguably just an allomorph of -ed and doesn't thus need an explicit mention)
OTOH some dialects have (again, AFAIK) [ŋks] vel. sim. even for intermorphemic instances such as gangster or even going somewhere, which illustrates that the epenthesis in question can be synchronical and not just a fossilized sound change result.
I think I'm bordering on OR here however. I agree that a literature quote would be good.
--Tropylium 23:09, 27 January 2007 (UTC)
I don't think we need to worry too much; it's a lot simpler to just report what (reputable, mainstream) dictionaries write as their transcriptions. (Though that does mean that [ŋks] is in: [1].) That sort of criterion is a lot easier to apply than trying to decide which transcriptions are epentheses; OR, as you say. –EdC 01:56, 28 January 2007 (UTC)

I was focusing on the difference between a phonemic transcription (which is what we're using) and a phonetic transcription. [ŋkθ] and [ŋθ] are definitely both used but that's not the same as saying that /ŋkθ/ and /ŋθ/ both exist. In the kind of Southern British dialects that I'm familiar with, when we move from a nasal (voiced) to a voiceless stop or fricative at a different place of articulation, we tend to insert a voiceless stop which is homorganic with the nasal. That's a phonological rule. (The /md mz ŋd ŋz/ examples don't fit this rule.) So I think the correct analysis for my speech is that /ŋθ/ is usually realised as [ŋkθ], though sometimes as [ŋθ]. But I don't know whether the literature agrees with my thinking, and I don't know if other dialects need a different analysis. Definitely not out for a major debate and happy to leave it as it is for the time being - I mostly wanted to satisfy my own intellectual curiosity. Gailtb 23:08, 28 January 2007 (UTC)

[edit] Syllable structure organization

What would you think of more schematic onset/coda tables? Here's a quick example for the former:

m n 0 r l w
sm sn
b br bl d dr dw g gr gl gw dZ
p pr pl t tr tw k kr kl kw tS
sp spr spl st str sk skr skl skw s Sr sl sw
f fr fl T Tr Tw h (hr) (hl) hw S
v (vr) (vl) D Z z

The empty columns represent missing lab+w and alv+l series. I could have done three such ones for the palatals, too, but as they don't cluster to start with, it doesn't seem as informativ. Also, I'm not sure where to stick /z/, /j/, /sf/, and clusters with /j/ in there, but that's just a 5-minute sketch obviously. --Tropylium 23:55, 27 January 2007 (UTC)

Yes, that could definitely work better than the current mess. We should still include examples, though – perhaps using Template:Explain? –EdC 02:08, 28 January 2007 (UTC)
The general idea of a table which will show the parallels and with examples sounds good. Gailtb 23:17, 28 January 2007 (UTC)
I thought up a solution for onset /z/, which required a slight rearrangement. Also: a sketch for codas. This time there're more hard-to-place ones: /rl/, /ft sp st sk/, /dz/, /rst lst/, and /kst/. (Might be the beginnings of an /st/ series in there...) Missing /T/-final clusters would probably be better listed separately, since they're more of extrasyllabic than phonotactic nature. --Tropylium 22:29, 22 February 2007 (UTC)
m lm rm n ln rn N 0 l r
b lb rb d nd ld rd dZ ndZ ldZ rdZ g rg
p mp lp rp t nt lt rt tS ntS ltS rtS k Nk lk rk
pt mpt lpt rpt kt Nkt lkt rkt
ps mps rps ts lts rts ks Nks lks s ns ls rs
f mf lf rf T nT lT rT S lS rS
v lv rv D z nz

[edit] Caught in the cot-caught merger

I was born in Hollywood, I grew up in the San Fernando Valley, and I have spent over 80% of my life in California. In other words, I speak Californian like a native. While I agree that "father" and "bother" have the same vowel, and "cot" has the same vowel as those, "caught" is different, a lower sound than the other three. The sounds are close, perhaps so close that the distinction can not be reliably heard, but I wouldn't call them the same. — Randall Bart 01:50, 13 February 2007 (UTC)

While it's not impossible that you are in fact pronouncing them in two distinct ways (you would presumably be part of a very small minority in your area), it's impossible to determine in what way they are different with a description like "lower sound", which implies a difference in stress or tone. Perhaps you could learn how to describe your accent in IPA. 61.25.248.86 06:07, 14 March 2007 (UTC)
I interpreted "a lower sound" in the above comment to mean "a lower vowel" i.e. a more open vowel. While phoneticians speak of vowels as "close" or "open", phonologists (at least in the U.S.) usually say "high" or "low". —Angr 06:16, 14 March 2007 (UTC)

[edit] /tθ/

Is /tθ/, as in eighth, an affricative? Thanks. Kylmcd 14:54, 14 February 2007 (UTC)

You mean, is it an affricative or a stop-fricative sequence? Phonetically, it's an affricative because the stop [t] doesn't have a release before the [θ] starts, but phonemically it could be regarded as a stop-fricative because it's constructed from the morphemes /eɪt/ + /θ/ (ordinal marker). –EdC 15:14, 14 February 2007 (UTC)
Not to be pedantic, but what it is phonetically is an affricate, not an "affricative". —Angr 18:18, 14 February 2007 (UTC)

[edit] ambiguous examples

"For example, in his play The Taming of the Shrew, shrew rhymed with row."

The preceding is an example of several ambiguities to be found in the examples given throughout the article. This one is particularly vague because "row" could refer to the English noun which means "an arrangement of objects or people side by side in a line" which I pronounce [ɹəʊ] or the noun meaning "quarrel, angry dispute" which I pronounce [ɹaʊ]. Which one rhymed with "shrew" for Shakespeare, I still don't know.

If possible, IPA transcriptions should accompany all examples, or at least a citation of the referent dialect. Joshua Crowgey 06:46, 23 February 2007 (UTC)

I just looked through the play to see the offending rhyme, but the word "row" doesn't appear even once, though this line looked suspicious:
Your husband, being troubled with a shrew,
Measures my husband's sorrow by his woe:
...[2]
Perhaps it was actually "woe" that rhymed? Though though that doesn't give any hints as to the pronunciation of "-row". Either way, this has to be backed up with a reliable source. 61.25.248.86 06:25, 14 March 2007 (UTC)
Retrieved from "http://en.wikipedia.org../../../e/n/g/Talk%7EEnglish_phonology_60b8.html"