Comments Welcome
George W. Grace
University of Hawaii

There is also a possibility that the phonetic detail rules can be eliminated or simplified if, instead of working out the phonetic details for each segment in each relevant context, a general account of the "basis of articulation" in the language can do that once and for all (Fillmore 1984: 130)

Some weeks ago I was reading Margaret Sharpe's contribution to the festschrift for Isidore Dyen (Sharpe 1996) where the following passage caught my eye:

"With astonishingly good hearing, mentally very alert, he often comments on phonetic detail of all sorts of accents--on the phone, radio and TV, though he sometimes misses the wood for the tree--he does not respond to the voice quality difference [my emphasis--GWG] that can so clearly differentiate English and Australian accents. With his attention to detail in a number of areas there is tendency to tunnel vision."(Sharpe 1996: xix).

The reference to "voice quality difference" immediately caught my attention because for years I'd been convinced that there was indeed a forest that phonology had persistently been missing in its focus on sound "segments". What I had in mind is illustrated by the kind of "gear-shifting", as I described it--that is, the kind of shift in the "set" of the speech organs--that's required for switching from one language to another. I guessed that what Margaret was calling "voice quality" was very much like what I was thinking of as "set of the speech organs". I eventually managed to reach her by e-mail, and she confirmed that it was and referred me a paper (Sharpe 1970) which she had published some years earlier. She commented (the paper being nearly 30 years old), "I hope to goodness others have taken it further since then!". Maybe someone has, but I'm not aware of it (although there have certainly been advances in articulatory and instrumental phonetics that seemingly would permit it to be taken further).

Her 1970 paper is far from a how-to-do-it prescription for describing the voice quality of a language or dialect, but it did provide a better general survey of what kinds of things are involved than anything else I was aware of.

She groups the articulatory factors influencing voice quality into two categories, "phonatory setting" and "resonant setting". Phonatory setting is "that configuration of the larynx responsible for generating a sound" (1970: 118); resonant setting is "that configuration of the resonant chambers and articulators responsible for producing a particular resonance" (ibid.).

She makes particular reference to two papers in a festschrift for Daniel Jones (Abercrombie et al 1964) each of which is devoted to one of her categories (or at least to a very similarly defined one). The first, Catford 1964, is on what its author calls "phonation types", the second, Honikman 1964, on what are called "articulatory settings"

Catford discusses the various locations of phonatory stricture, and the types of laryngeal modifications of phonation. However, his principal concern is with the contrastive use of phonation types within the same language rather than with different languages or dialects being characterized by different types. Some of the types he discusses include: breath, voice, voiced creak, whispery voice, ligamental voice [phonation actively restricted to the ligamental glottis], and lowered larynx voice.

Honikman, on the other hand, is concerned with differences between languages, and she presents her own analysis of the differences between English and French articulatory settings. She describes articulatory setting (1964: 73) as "the gross oral posture and mechanics, both external and internal, requisite as a framework for the comfortable, economic, and fluent merging and integrating of the isolated sounds into that harmonious, cognizable whole which constitutes the established pronunciation of a language."

Sharpe (1970: 118-9) comments: "Some configurations and movements of the vocal apparatus are more obviously related to fluent merging of the isolated sounds than others; if required a further distinction within resonant settings can be made between articulatory setting in Honikman's sense, and modifying setting. This latter, modifying setting, would include degree of tension of palate and degree of opening of the nasopharynx (both affecting perceived nasality), degree of tension of the fauces (faucalisation gives a penetrating, harsh quality to the voice, sometimes dubbed a nasal twang), and position of larynx and tongue root (which affect the frequency of the vowel formant bands, and therefore the 'thinness' or 'heaviness' of the voice)." [Note: the "palate" whose tension is spoken of here is the soft palate.]

These papers obviously don't add up to a finished theory. The most fundamental distinction to be made seems to be that between phonation and "configuration of the resonant chambers and articulators", but there are still problems in determining a precise boundary between these two. There are also uncertainties about the placement of Sharpe's secondary boundary between articulatory setting and modifying setting.

There are, nonetheless, a number of very suggestive observations. Sharpe offers a number of observations and anecdotes about voice quality differences among different dialects of English. She also comments on such differences among Aboriginal Australian languages, and adds (1970: 116-17) "Here it should be noted that Australian Aborigines seem very conscious of voice quality. Different linguists have been criticised as having a voice 'too tender', or 'not heavy enough', or 'needing to be more light', etc. The linguist who is unaware of voice quality is often at a loss to interpret these comments."

However, the nearest approach to a systematic description of what I'd so vaguely conceptualized as the "set of the speech organs" (or of a very significant portion of it) is Honikman's (1964:81) description of the differences in the "'set' of the organs for utterance of " French and English (the only languages that had been "investigated in any detail" from that perspective). She describes the contrasting set of: Jaws, Lips, State of oral cavity, main consonant articulation, and tongue anchorage, tip, body, and underside for these two languages. The contrasts she identifies are the following:

English: Loosely closed (not clenched)

French: Slightly open

English: Neutral; moderately active

French: Rounded; vigorously active in spreading and rounding

English: Relaxed

French: Cheeks contracted

English: Tip--alveolar

French: Blade--dental

English: To roof laterally

French: To roof centrally

English: Tapered

French: Untapered

English: Slightly concave to roof

French: Convex to roof

English: Concave to roof

French: Neutral

Honigman reports that she has achieved excellent results in teaching English to French-speakers and vice versa using this analysis together with exercises designed to achieve the right set of the different organs.

But how generalizable are the results she's achieved with these two languages? Do her parameters (jaws, lips, anchorage of tongue, etc.) refer to binary oppositions or are they scalar? What other parameters besides hers need to be recognized? In fact, to what extent can a single voice quality (or some portion thereof such as articulatory setting) actually be specified for a given language or dialect (or individual speaker)?

I won't try to guess what the answer to these questions might turn out to be. Unfortunately, I don't know anything further about Honikman or her work, but it's interesting to speculate about what might have happened if she had been an active grant-getter able to attract and support a significant number of top-flight students.

What I have in mind is an active research team able to work in full confidence that what they were doing was accepted in the profession as "normal science" (or, since linguistics isn't exactly a science in Kuhn's sense, as "normal linguistics" or whatever the linguistic equivalent should be called), and making a serious effort to analyze the articulatory settings of a variety of languages using Honikman's analysis of English and French as their model (their "paradigm" in Kuhn's [main] sense). How likely is it that such an effort wouldn't eventually have succeeded in producing descriptions in the same general form (sometimes, one would expect, with dialectal and/or stylistic alternates) for a substantial sample of languages? No doubt the analytical scheme would have evolved along the way, with changes in the parameters to be observed and with considerable discussion before anything approaching a consensus on the possible values on each parameter had emerged.

But if it did turn out that the contrasts in voice quality of different languages could be specified in great detail, what would the value of these descriptions be?

Well, for one thing phonology, even as it's constituted today, would certainly find it interesting if someone could show that any significant proportion of the segmental differences between languages could be predicted from the differences in overall voice quality of the respective languages. However, I think there would be much to be gained beyond the potential contributions to segmental phonology, and I suspect that if the kind of results that I've just been imagining had actually been achieved at the time when Honikman was active, phonology in general would have evolved rather differently.

To begin with, as I mentioned above I've never been satisfied that we had a very adequate understanding of the "gear-shifting" experience in switching between languages.

I've also long believed that better understanding of what we're calling "voice quality" and particularly "articulatory setting" might provide a big step toward understanding some of the phenomena that fit under Sapir's concept "drift". That is, I've long thought that this kind of set of the speech organs might go a considerable way toward accounting for the pattern of the changes that a particular language tends to undergo. (It might be worth emphasizing here that, although drift has perhaps attracted interest most particularly because of the tendency toward "parallelism" in the drift of related languages--something that Sapir also pointed out--it's a very much more general phenomenon that has nothing to do with whether or not there happens to be a related language that is drifting in parallel. In other words, I'm proposing that it might go a long way toward explaining sound change.)

And there's another quite different matter. I've been convinced for a long time that the phonological convergence that results from so-called "language contact" is phonetic--i.e., a matter of pronunciation rather than of, say, the system of contrasts--and that what drives and guides it is a pressure on the speakers to free themselves of the burden of "gear-shifting"--in other words, that it's essentially a harmonizing of the set of the speech organs.

However, I'd like to speculate about further possibilities: specifically, about what might have happened if the kind of descriptions of voice quality that we're imagining had been available before the notion of contrast (or opposition), or even that of segment, had become part of phonology.

My point is that linguistics arose out of a particular intellectual environment, and that it's hard to tell what part of its present assumptions are simply a legacy of that environment. The environment in question was one in which the prevailing attitude toward language was an overwhelmingly prescriptive one--so much so that linguists have been at great pains to emphasize the contrast between this prescriptive attitude and their (linguists') present non-evaluative "descriptive" posture. Still, at the time when linguistics was struggling to establish itself as a separate discipline, all of the available assumptions about what language was like and how it worked were those that had been developed in the prescriptivist tradition. Therefore, it was inevitable that linguistics would begin life with assumptions about the nature of language that, except for certain specific points, were identical with those of that tradition.

The prescriptive tradition conceived of any language deserving of the name (i.e., the kind of language that its prescriptions were designed to produce) as constituting a system (that is, something sufficiently standardized to be represented by a single set of grammatical rules) whose primary function was expository prose--in particular the expression of propositions. Contemporary linguistic theory has departed from that tradition to the extent of making the democratic assumption that all humans speak proper languages. However, the way it's reached this democratic assumption has been, not to challenge the prescriptivists' idea of what a proper language is, but to take the validity of that idea for granted and, accordingly, to treat the speech of any and every community as if it were essentially an instrument for encoding propositions.

But it may not be clear what this has to do with phonology. Well, in the first place phonology would inevitably be concerned if only because of the centrality of written language in the prescriptivist tradition. And writing is no doubt an indispensable prerequisite for an expository prose tradition of any consequence. In any case, the whole prescriptivist tradition was focused on written language.

Although written language was recognized as having a spoken counterpart, the written form was accorded unquestioned priority. The spelling was (and still is for the general population in the Western cultural tradition) seen as the point of departure. Each letter characteristically had its own pronunciation (although in atypical--and mostly unnoticed--cases there were different pronunciations in different environments), and these pronunciations typically defined letter-sized segments of sound. This view has been so deep-rooted that even today when one speaks of "a sound", it is ordinarily assumed that one is speaking of such a letter-sized segment.

At the same time, the emerging science of linguistics did have to take cognizance of unwritten languages (included among which, of course, were the Indo-European ancestors) and to make the conversion from a visual to an auditory medium. Although these unwritten languages provided no letters at all to be used as a starting point, linguists were able to proceed as before on the assumption that the pronunciation of these languages was based on the same kind of "sounds" (i.e., letter-sized segments of sound) as those that they attributed to the spoken version of their own languages. These "sounds" were assumed to be, as it were, already up and running, ready to have letters assigned to them.

The fact that these "sounds" (these sound "segments") were taken to be those of the alphabetic writing tradition in preference to any other deserves some attention. In fact, Western histories of writing have frequently represented the decision (of the alphabetic tradition) that vowels (but not pitches, etc.) are segments on a par with consonants as the culminating step in our quest to identify the correct segments of sound in human language.

I think the foregoing is a fairly accurate characterization of the intellectual landscape in which phonetics and phonology emerged. They began with the assumption that such segments of sound--such "sounds"--essentially constituted their subject matter. The questions they asked were such as: how these sounds were produced and what their acoustic properties were--or for a particular language: what sounds were used in it, what variations occurred within the same sound, what sequential combinations occurred. In sum, the questions presupposed a central role for letter-like "sound segments".

But of course questions that contain a presupposition don't allow for any answer that would call the presupposition itself into question. And that's why it seems interesting to speculate on what might have happened if the scientific investigation of the pronunciation of language had begun with the study of articulatory settings under conditions in which there were no prior assumptions about segmentation of the speech stream.

Abercrombie, David, D. B. Fry, P.A.D MacCarthy, N.C. Scott, and J.L.M. Trim (eds.). 1964. In honour of Daniel Jones: Papers contributed on the occasion of his eightieth birthday 12 September 1961. London: Longmans. (Back up)

Catford, J. C. 1964. Phonation types: the classification of some laryngeal components of speech production. In Abercrombie et all (eds.), pp. 26-37. (Back up)

Fillmore, Charles J. 1984. Lexical semantics and text semantics. In James E. Copeland, ed., New directions in linguistics and semiotics. Houston TX: Rice University Studies,

Honikman, Beatrice. 1964. Articulatory settings. In Abercrombie et al (eds.), pp. 73-84. (Back up)

Sharpe, Margaret C. 1970. Voice quality: A suggested framework for description and some observations. In S. A. Wurm and D. C. Laycock (eds.), Pacific Linguistic Studies in Honour of Arthur Capell. Pacific Linguistics Series C, no. 13, pp. 115-32. (Back up)

Sharpe, Margaret. 1996. Isidore Dyen (in his eighties): A personal note. In Bernd Nothofer (ed.). Reconstruction, classification, description: Festschrift in honor of Isidore Dyen. Hamburg: Abera Verlag, xviii-xxi. (Back up)

Addendum (30 May 2002)

I said above that if anyone had taken the study of voice quality further than Sharpe 1970, I wasn't aware of it. I've just recently been made aware that there is more--and more recent--work than I knew of, and I want to acknowledge that work. As I do so, however, I'm now aware that I'm only scratching the surface of what's been done.

I only chanced on the sources below when I saw a reference in a review to Stuart-Smith's 1999 paper. It was that paper that led me to the work of John Laver--his very substantial bibliography of 1979 and his 1980 book. The very large 1994 book on phonetics also includes a substantial treatment of voice quality.

REFERENCES

Laver, John. 1979. Voice Quality: A classified research bibliography. Amsterdam: John Benjamins B.V.

Laver, John. 1980. The phonetic description of voice quality. Cambridge: Cambridge University Press.

Laver, John. 1994. Principles of phonetics. Cambridge: Cambridge University Press.

Stuart-Smith, Jane. 1999. Glasgow: Accent and voice quality. In Paul Foulkes and Gerard J. Docherty (eds.). Urban Voices: Accent studies in the British Isles. London: Arnold, pp. 203-222.

First put on the Web 7 September 1999
Last changed 12 June 2007