Proper Phonetic Pedagogy:
Field Notes from a Laboratory Science

by Jacob Bartholomew Pine
Professor of Acoustic and Instrumental Phonetics
New University of Southern North Carolina

As Introductory Graduate Phonetics is my favorite class to teach, I was overjoyed to receive this journal’s summons to share the fruit of my many years of experience training the minds of the next generation of linguists. My first piece of advice is to pay close attention to the students in attendance. For example, Introductory Undergraduate Phonetics is a far different beast to the graduate course; in that course I have found it useful to restrict myself to slinkys, pipes, and violins until after the refund deadline, whereas with graduate students I can start with the differential equations on the first day.

On a related matter, the instructor must not allow herself to forget that her students are in graduate school. They can take it. Indeed, they have to take it. A student who goes into linguistics expecting the usual humanities sidestep shuffle away from rigor is a soft-shelled thing who must be disabused of such cowardice.

Ideally, an introductory phonetics class should consist of at least four semesters. However, if you are restricted to a one-semester course, as many phoneticians are, a judicious condensation of the necessary subject matter is in order. I have found the following program the most conducive to laying a proper foundation for more advanced work.

First, of course, one must start with the equations of continuity and the various conservation equations. I have found that the Lagrangian treatment, while less intuitive, is more conducive to a quick derivation, after which the usual approximations should be applied and at the very least the Webster horn equation should be derived by the end of the first hour, as well as a close analysis of the conditions under which the approximations behind the equation are valid. Then and only then may one proceed to the brutally simplistic treatment with two narrow tubes with various apertures, which, while useful for building an intuitive understanding suited to the capacities of middle-school students, is far too primitive for the modern scientist of language.

In the second week, students turn to the anatomy underlying phonetics. Ideally the university should have a medical school with the odd cadaver; otherwise there’s really no point, is there? Indeed, I dare say the instructor should insist upon several cadavers, ideally of subjects who were recorded before their demise. After one class period devoted to a detailed study of the gross anatomy of the vocal tract, it is then time for the students to start taking measurements: mechanical inductances, capacitances, and resistances, muscle densities and tensions, and the maximum and minimum curvatures of the surface of the vocal tract at 0.1 mm intervals between the lips and the pharynx.

The next class period should then be devoted to constructing a finite-element model of the vocal tract using detailed mechanical analyses of the muscles of the vocal tract, all sound-reflecting surfaces, and the acoustic equations from the preceding lessons. Students never get the right answer on the first try, so the next three classes should be devoted to (1) a derivation via differential geometry of a coordinate system ideally suited to describing the movements of the articulators, (2) a close analysis of the consequences of this analysis using tensor calculus, (3) a set of biochemical lab sessions measuring the effects of formaldehyde, rigor mortis, and post-mortem deliquescence upon the mechanical properties of the tissues of the vocal tract, and (4) a lab section in which the necessary properties of a living vocal tract are measured on a class volunteer. I have found this is best combined with the next section of the class, the use of electromyelography, static palatography, and other primitive techniques, before advancing to the use of X-ray films for a true look inside the fully functional vocal tract—here the students finally see the value in the work they did deriving an appropriate coordinate system!

The next section of the class treats the neural underpinnings of vocalization in as much rigor as access to anatomical subjects allows. Again, ideally your university should have a school of neurology and neurochemistry, but failing that you can still obtain the necessary equipment to perform basic neurochemical and neuroelectrical measurements in your own lab. Remember, the motion of the vocal tract is governed by 1400 neural signals per second—you needn’t fear running out of interesting things to measure, so don’t skimp! The more adept the students become at vocal neurology, the easier it is for them to proceed to the next section, producing a neural-net model of the processing of underlying segments yielding the corresponding neural triggers they just spent the last month measuring, which can then be fed into the mechanical model of the vocal tract they created for their midterm, which then allows a full understanding of the phonetics-phonology interface, one far richer and deeper than the bloodless reified abstractions touted by phonologists. At this point, ideally, they can then compare the outputs of their model with the recordings of the cadaver made before decease.

The last section of the class should be devoted to auditory phonetics. The theoretical section, of course, should be taught like a mechanical engineering class with strong acoustics and regional anatomy lab components through which the students are led to build a finite-element model of the inner ear that outputs the appropriate neural signals.

Time permitting, you can then turn to ancillary matters like narrow transcription and the IPA, but don’t worry if you run out of time. After all, you’re doing science, for heaven’s sake, not engineering.

	Conference Report: Educational Technology in the Linguistics Classroom
	Teaching Historical Linguistics: A Program of Much-Needed Reform—Lea Kim Shopmont
	SpecGram Vol CLXXX, No 2 Contents