SPEAKING WITH SKILL, 2^ND EDITION, 1^ST DRAFT

Now that Speaking with Skill is readily available worldwide just in time for holiday gift-giving, I could allow a sigh of relief to flow egressively from my alveolar sacs; were it not for the fact that I managed to distribute liberally—throughout its pages—words,  sentences and sometimes entire paragraphs that need to be changed. Others might call these “errors” or “omissions” or “stupidities,” and some reviewer undoubtedly will, but I consider the opportunity to fix them an ongoing retirement “hobby.”

I offer therefore the following four corrections to the text. This is just for starters, mind you. I know that there are some typos, poorly shaped phrases, fuzzily expressed ideas and instances of questionable phonetic transcription in the book. I will let these pass until I actually do the second edition. (Though do But these are four that I am moved to correct right away; and that certainly need to be borne in mind by anyone teaching out of this book. Here they are, in descending order of egregiousness.

1. Why I got a C in First-Year Greek in college.  This one actually qualifies as stupidity. I have a long, boring and self-serving explanation as to why I came to write the two sentences that I will quote, and I will provide it upon written request, but I will not burden you with it here. I present them are as they appear on page 66, second paragraph, of Speaking with Skill: ‘It comes from the Greek word fdoggoV, meaning “sound”. Oh, by the way, voice the “th”.’ So wrong. Oh, so wrong. What that spelling would come up with, if spoken, would be pronounced approximately as “fdongos,” which is at least only mildly prurient. The correct Greek spelling of the word in question is fqoggoV. It is pronounced with the characteristic “phth” combination that one often finds in Greek. Oh, by the way, you don’t voice the “th”, so the sentence that breezily says oh by the way you do should be eliminated.

2. The missing approximants. This is a sin of omission. There is one line omitted in the Symbol Practice Exercise on page 143: the Lateral Approximants. Here it is. Place it right between Approximants and Other Symbols:

 l             ɭ             ʎ             ʟ      

3. Three lost paragraphs found.  On page 208, while discussing the disadvantages of the “vowel substitution” method for transcriptions of accent differences, I say grandly “But first we must examine. . .” allowing the ellipsis to build suspense all the way to the very next line that starts with the heading The Territory. The territory in the book, though, is a bit cramped, because I left out three paragraphs that actually explain what is going on. Here is the entire text that would continue onto page 209:

“The Territory. When we look at a vowel quadrilateral, divided into sections by lines and populated by symbols in suspiciously orderly array, we may get a hint that we are looking at a theoretical construct. And so we are. When famed British phonetician Daniel Jones established this version of a vowel chart (modifying earlier variants) in 1917, he did so by placing most of the front and back vowels in equidistant intervals from close to open. The few remaining vowel phonemes are also placed with some symmetry in between these so-called cardinal vowels. Jones constructed this formal graph to define the physical territory of the vowel chart, limited to its quadrilateral shape by the limits of vowel articulation within the oral cavity. By so doing, he was able to define physically the territory within the chart by providing specific points of physical reference, all related to degree of tongue arching/cupping or lip rounding/unrounding, the two chief ways of shaping acoustically the unobstructed flow to form vowels. He did not place the cardinal vowels in these positions because he was in any way suggesting that words using these phonemes must be pronounced at the cardinal positions.

“Physically you could enact the voiced, unobstructed flow anywhere within the limits of the vowel quadrilateral, including—but not limited to—the places where the cardinal vowels sit in isolated splendor.  What’s more, if we lump together all the world’s individual speakers’ patterns in all accents in all languages, every possible physical and acoustic exploration of the quadrilateral will happen. There are innumerable gradations of action and sound.

“Each vowel phoneme carries with it a unit of meaning that differentiates it from the other vowel phonemes and as long as we keep their realizations as sounds within that stern “cardinal” separation that Jones created, the “territory” is an easy one to navigate and there is no question about what phoneme you are saying. But in practice—and Jones knew this—each vowel phoneme exists within what I like to call its phoneme cloud that contains all the discrete possibilities for pronunciation variation that still convey the vowel phoneme’s unique meaning. Within the territory of the vowel quadrilateral there are (depending on what language you are speaking) a lot of phoneme clouds, each with its own wispy territory and its own shifting boundaries. On a clear day, the phoneme clouds remain separate or butt gently against one another. Sometimes, in cases of inclement linguistic weather, these clouds intersect and even flow into the territory of one another; these blended areas are where transcriptions that depend on vowel substitutions may be convenient; but they can also get you into real trouble, because if you don’t have a larger context within a sentence to define meaning your listener may not know what you’re saying. You may pronounce “sheep” like “ship”, or “ship” like “sheep”, for example; so if you write a purchase order that says “I want to buy a sheep” and phonetically transcribe (What purchase order is not written phonetically these days?) the last word as [ʃɪp] just because in your own pronunciation the phoneme cloud for [ʃip] has strayed southward, you may be surprised at what is delivered to your farm.

“The further a phoneme cloud gets into the territory of another phoneme cloud, the more the resilience of meaning relies on these larger linguistic structures – grammar, other words in the sentence, tone of voice, emphasis, voice quality, intonation, and so on. For all these reasons, vowel substitutions are a limited way to show the phonetic variation that we find in accents. They are convenient, though.”

4. Pronunciation Exercise Phonetics Revised.   Pages 314 – 315. I have revised the phonetic transcription for this passage, cleaning up some errors and inconsistencies. I hope you don't have to squint too much at the font size.

If you are using Speaking with Skill as a text in your classes, I hope you will take these corrections to heart.  Especially that "phthong" spelling in Greek. 

         

Oh, those wandering GOATs

I think it's high time we stopped representing the So-Called General American GOAT phoneme as /oʊ̯/.  I don't think it can be very controversial to suggest that [oʊ̯], phonetically-speaking, would be a very unusual vowel to hear from most American speakers in 2012.  The only Americans who come close to a fully-rounded [o] in realizing this phoneme are from what dialectologists call the North Central dialect region—the Dakotas, Minnesota, Michigan's UP, and parts of Wisconsin, Montana & Iowa.  (We might also add Alaska's Mat-Su Valley, where Sarah Palin hails from—much of which was settled in the 1930s by transplants from Minnesota).

I can think of two arguments on behalf of retaining /oʊ̯/.  The first of these is that it is the established convention.  /oʊ̯/ has been used to represent the vowel in American pronunciations of GOAT for a very long time—at least since Kenyon and Kurath.  While choosing to buck the convention might create some confusion, however, I think it's a small price to pay for greater phonetic accuracy.  (I feel similarly about using /ɜ˞/ to represent the American NURSE vowel, and have long taken to using and teaching it as /ɘ˞/.)  Perhaps, eventually, we might change the convention.  In the long run, I think such a change would lessen student confusion.

Spelling is the other argument I can imagine in favor of keeping /oʊ̯/.  It has an 'o' in it, after all.  The letter 'o' is pronounced with the GOAT vowel, and stands in the popular imagination as the phoneme itself.  I can see how we might lose something by re-phonemicizing GOAT without an 'o.'  However, as with the argument about convention, I think there is more to be gained than lost.

Let's suppose I've persuaded you that we should do away with /oʊ̯/.  But what do we replace it with?  After all, another argument in favor of keeping it might be that we can find no likely replacement candidate.  There are a large variety of American realizations of this phoneme, after all; things like [ɔ̜ʊ̯̽, ɘʊ̯, ʌɯ̯̽, e̽ʏ̯, ɤu̯] and a great many more besides.

I'd like to propose /ɤʊ̯/.  I am perhaps biased, as this is more or less the way I would describe my own realization*.  Nevertheless, amid a sea of variation, I think this makes a certain sense.  I'd argue, first of all, that it is important that the first element be unrounded.  This, in fact, is the main thing that makes /oʊ̯/ so unsatisfactory to me.  The majority of Americans simply do not round the first element of this diphthong.  (Or if they do, it isn't much.)  I would say that the phoneme is characterized, in fact, by a progressive rounding of the lips through the diphthong.  And though there are many Americans who don't round the lips much or at all anywhere in the realization of this vowel—not even in the coda—lip-rounding is still common enough that I think we can leave the ʊ unchanged.

A final note—this is far from an academic argument.  Though we are generally more interested here in phonetics than in phonemics, the phonemes make a difference to how we and our students think about and teach things.  If I am teaching a So-Called General American accent to a non-native speaker, it makes much more sense to teach the phoneme as /ɤʊ̯/ and then teach to that phonetic target.

So let's do it!  All in favor of /ɤʊ̯/ raise your hands!  (Counter-suggestions welcome, of course.  Let's have a debate about it, by all means.  But I do think it's high time for a change.)

*Doug Honorof has suggested to me that my GOAT vowel starts lower—more open-mid than close-mid.  After a lot of introspection and phthong-testing™, I still think my nucleus is close-mid in height.  He's invited me up to Haskins Lab at Yale, where he will norm my vowel space and definitively measure the height of my GOAT nucleus.  Only one of us can be right!  (Unless it turns out to be perfectly -mid, I suppose.  We could call that a draw.)  Stay tuned.  If the suspense is killing you, call me and I'll murmur some GOATs in your ear.  You can make up your own mind.