Turning away, for the moment, from Wikipedia's material on vowels (I'm still struggling with WP's ways of uploading and downloading sound and picture files), I've been having another look at the article on Prosody and trying to pin down why I find it so disappointing.

The lead paragraph is ok, I think. Note that is preceded by a warning from WP that the article needs more references to back up some of the statements. That should be easy enough to fix. However, instead of surveying the components of prosody one by one, we are led off into some rather marginal areas. The first section is headed "Classification" and takes us off into the use of different rhythmical types for language typology. This subject area, a bit of a minefield, is covered more fully  in the article on Isochrony, which is gradually getting better. The last sentence in the Classification section reads  "The classification of languages is done under the assumption that a language has "isochronous rhythm", meaning that there is an equal amount of time between stressed syllables, symbols (sic), or moras, depending on the category of language". Yes, well I think that needs some rewriting.

The next section is called "Acoustic attributes". What I want to say here may strike some readers as being a bit picky: I was brought up to distinguish between acoustic and auditory phonetics, where the former is concerned with physical, measurable properties of speech that can be studied in the laboratory, and the latter deals with subjective impressions and is concerned with the listener's experience in listening to speech (more about this in a future post). So when I find this section talking about "syllable weight, loudness and pitch", I feel that we are in the area of auditory, not acoustic attributes. Some of this material is not very scientific. We get some rather poetic stuff about tone and intonation in Mandarin, then the text moves on to a mention of stress: we are told about the difference between 'dessert' and 'desert' (arid place). The writer could, I feel, have used stress marks to make the point clearer. Then we get the example of the child screaming "give me dessert!" ("when the entire word is stressed "), an example I remember seeing in another WP article. I simply can't understand what point is being made here. At the end of the section the writer starts talking about vowel formants (rather than vowel quality, which is more or less the auditory equivalent and would have been a more appropriate term).


The next section concerns itself with the Domain of prosody, whatever that means. Early in this section we are told "these prosodic units are the actual phonetic "spurts", or chunks of speech. They need not correspond to grammatical units such as phrases and clauses, though they may; and these facts suggest insights into how the brain processes speech." I can see why WP feels that some references are needed to back up vague generalizations such as these. Later, we are told "In English, falling intonation indicates a declarative statement while rising intonation indicates an interrogative statement". As we know, this claim does not stand up when we study natural speech, but still the myth lingers on. Then there's an odd little paragraph about Hebrew morphophonology that seems to have wandered in from somewhere else. At the end of the section we get a para on the prosody of sarcasm, which doesn't seem to me to have much to do with the subject of the section.

And so to the next section, which is Emotion. We are told that the main WP article on this topic is Emotional Prosody, but that article and the section we are looking at here, though they seem to duplicate each other, seem to have been written from completely different standpoints. Here, in contrast with the sometimes woolly generalizations that we have had in earlier sections, we get a lot of indigestible facts and figures from some psychological studies of listeners' ability to recognize emotions in speech. My aversion to this is, I suppose, personal - I once got involved in research in this area, and hope never to go near it again! For the rest of this article the topic switches to neurophysiological aspects of prosody.

We should remember that each aspect of speech that is traditionally treated as part of prosody gets an article to itself in Wikipedia, so it's instructive to look at articles on Intonation, Stress, Rhythm, Speech Tempo and so on. But I feel that if a student just read the Prosody article in preparation for writing a paper on that subject they would end up with a pretty confused picture.

Carrying on from my previous posting about Wikipedia's treatment of Cardinal  Vowels (I'm never sure whether it's appropriate to use capital letters for these), I have submitted a short article about phonetic training. This hasn't been reviewed by WP editors yet, but the draft can (I think) be seen here. I'd be glad of suggestions for improvement - there is still time to make changes.

I have a number of concerns about the WP article on Cardinal Vowels, some of which I have already mentioned. 
1. I think it's essential to make it clear that the system works through training in its use, something which is not mentioned in the article.
2. Although the CV diagram in the Cardinal Vowels article looks to me like the real thing, the vowel diagrams on all the dependent articles on specific vowels reproduce the Wikipedia version which has a lot of extra vowels added. In what I will call the WP-CV diagram you get this (note that WP labels this "IPA vowel chart"):
 

So in addition to the traditional primary and secondary vowels, we have a row of five Mid vowels, two open central vowels and two near-close central vowels ɪ̈ and ʊ̈. I can't see why these are thought to be useful.
3. The distinction between Primary and Secondary Cardinal Vowels is a very old one. I have never read a proper justification for the use of this distinction. Why are Primary vowels primary? Is it (as I suspect) because they are thought to be easier to learn and produce by speakers of European languages, while Secondary vowels are the hard ones for us? I think that in the context of the languages of the world it is wrong to imply that in some way Primary vowels are "unmarked" and Secondary "marked".


The next step in sorting out WP's treatment of CV's is to consider the recorded examples. Leaving aside the controversial Mid and Central vowels for now, it sounds to me as if some of these recordings need to be improved. [ɛ] has a lot of noise in the recording. [a] is not fully front and [ɑ ] is not fully back. [ɔ] is too close. [œ] is a poor-quality recording. The [Œ] vowel, admittedly a rare and difficult beast, sounds nowhere near what I would expect - it is not rounded enough, open enough or front enough. [ʌ] is not back enough. Finally, [ɨ] sounds rounded to me.

The Trained Phonetician

Listening to some of the vowels recorded for Wikipedia and published in some of the articles referred to in the previous posting brings up a problem. Since I got into the world of practical phonetics in the 1960’s I have usually believed that the trained phonetician has a privileged position in making claims about the quality of speech sounds and in comparing one sound with another. It is striking that the WP article on Cardinal vowels makes no mention of the process by which someone learns, and learns to make use of, these vowels, yet to phoneticians of my generation the intensive training in recognizing and in making these vowels is absolutely essential. Likewise, only intensive practical training in production and recognition has been thought to qualify one to make proper use of the IPA Chart. I have always believed that practical phonetic training is a useful thing to do (if horrendously expensive in terms of teaching hours – it has to be done in small groups or even one-to-one), but when I read material about vowels and consonants written by enthusiastic amateurs who may never have received any sort of practical training I am forced to examine what it is that we trained phoneticians can claim to have which others do not have. There is a mystique surrounding Cardinal Vowels, in particular, which is vaguely redolent of a sacred rite – the only true initiates are those who have been trained by Daniel Jones, or by one of his pupils or colleagues, or (these days) by one of his pupils’ pupils’ pupils. There has to be a laying-on of hands to pass the expertise down from teacher to pupil, preferably not too far from London WC1. I used to believe that the very best practical phoneticians could reliably distinguish and produce hundreds of different vowel qualities, but my faith was put to the test when I read Peter Ladefoged’s experimental work on vowels, published in the 1950’s and 1960’s – and one study in particular. For those who don’t know it, his study of how trained phoneticians coped with the vowels of a dialect of Scots Gaelic seems to me to be crucially important: he asked eighteen phoneticians to listen to a recording of ten words spoken by a native speaker of  Gaelic and to place the vowels on a cardinal vowel quadrilateral. He then studied the degree of agreement or disagreement among the phoneticians. Although Ladefoged himself draw attention to the fact that the phoneticians trained in the British tradition were closer to each other in their judgments than were those who had not had this training, what I take from the study is the great divergence of judgments among all the listeners in the case of vowels that were distant from Cardinal values.  (The best place to read about this is in PL’s Three Areas of Experimental Phonetics, OUP, 1967, Section 6, pp. 132-142 – see especially Figure 47 on p. 135). My scepticism was added to when I was teaching in the Linguistics and Phonetics Department at Leeds University. We held, for a while, regular meetings of the staff who taught practical phonetics, a good-sized gathering of very experienced phonetics teachers, to check that we were all making consistent sounds and judgments. The lack of agreement among the teachers was, to me at least, disturbing.

The Wikipedia vowel chart in IPA vowel chart with audio implies that we can recognize/produce at least thirty-five distinct vowels. As you can imagine, I feel pretty sceptical about this, and the recordings don’t reassure me. I would like to be able to claim that I can contribute some expertise to this work, but I find it very hard to find evidence that I could use to convince someone who has no previous contact with the world of the IPA. I can’t find anything in the Principles of the IPA that sets out the value of practical phonetic training, and in fact, in trying to write a short article on that subject for WP, I have found it very hard to turn up any published work that addresses the issue. There are practical guides for use in phonetic training, certainly – books such as Wells and Colson’s Practical Phonetics, Catford’s Introduction to Practical Phonetics or Pat Ashby’s Speech Sounds, as well as a lot of free material on the internet, but these do not give much in the way of explanation of the nature of phonetic training and its uses. Am I overlooking some work that does this? Did Jones, or Sweet. or some other influential figure, set out principles for the training of phoneticians?

It would be nice if the IPA would make freely available on the internet an authorized set of Cardinal Vowels and audio demonstrations of all the consonants and tones that appear on the IPA Chart. That would make a great impression on the many people who currently get their phonetic expertise from Wikipedia. Somehow, I don’t see it happening.

Audible vowels

On some Wikipedia pages you can find “click and listen” buttons associated with some vowels. If you look at Cardinal vowels you can see the familiar chart and underneath it a table listing all the vowels. Each entry in this table takes you to a separate WP article on that vowel (you may feel that one article per vowel is a bit excessive), and in the “info-box” at upper right on each page is a button that gives you the audio. I do not know who recorded these vowels, nor even if they were all recorded by the same person. At this stage I prefer not to comment on the quality. If you go to IPA vowel chart with audio you can hear the same recordings (I think) just by clicking the appropriate place on the chart. However, there are now additional vowel symbols on the chart: these are vowels which are not included in the IPA’s version of the Cardinal Vowel chart, but which WP seems keen to include. Some of these vowels are ones which I don’t think I have ever attempted in my life – not in public, anyway. These “off-piste” vowels do sound very strange to my ears. On the Talk page corresponding to this topic I have already complained about adding extra vowels to the IPA vowel chart, and got a reasonably sympathetic response. I notice on that page that “ … the article on the near-close central rounded vowel [ʊ̈] or [ʉ̞] doesn't have a recording yet”. Now there’s a challenge for someone to take up!

There’s a lot more to say on this subject, and I’ll return to it. After that, there’s a lot to say about the article IPA non-pulmonic consonant chart with audio

There is a Wikipedia article with this title here. It simply contains the names of eight institutions, and a few links to other places. Most of those listed don’t actually have a Department of Phonetics. The list raises an interesting question: in my young days there were many universities, particularly in Europe, which had departments with the title ‘Department of Phonetics’. I studied in one (UCL), but never worked in one – I taught phonetics in the Department of Linguistic Science at Reading University and in the Department of Linguistics and Phonetics at the University of Leeds. All the ones I used to know have now been absorbed into bigger units with longer titles, and I really wonder how many Departments of Phonetics actually exist anywhere in the world. Perhaps someone could put me right on this? The nearest I can see in this list is Maria Curie-Skłodowska University in Lublin (Poland), which I have had the pleasure of visiting, but that is a Department of Phonetics and Phonology. If you Google ‘Department of Phonetics’, the only names that come up are the University of Turku , Finland and the University of Zagreb, Croatia, neither of which appears in the WP article’s list. There is a Department of Phonetics and Spoken English at the English and Foreign Languages University, Hyderabad.

I am going to propose that the current WP article is of no use and gives a misleading impression of the place of phonetics in the academic world. If someone would produce a list of universities or institutions which teach phonetics it might be worth publishing that as a WP article, but it would be difficult to decide which institutions should be included. Would we list places which just teach one module on Speech Communication? A one-semester class on Pronunciation Teaching?