A classical question in phonetics is whether 'articulatory ease' should be invoked as an explanatory principle. This is a controversial issue that is often avoided, since 'ease' is notoriously difficult to define. On the other hand, on most phoneticians' reading of the evidence, articulatory ease is indeed likely to underlie numerous phonetic and phonological facts.  So the problem remains. What is to be done?
    This paper begins in experimental biology. For a large number of species, walking, running and other locomotory behaviors tend to occur at speeds that optimize metabolic energy costs. Are speech movements shaped by cost functions similar to those for locomotion? Or do other principles apply?
    We decided to begin investigating those questions using a simplified biomechanical model of jaw movement. Our results are qualitatively similar to the measurements reported in the locomotion literature. A U-shaped curve was found for energy consumption as a function of speed of jaw movement.
    Accordingly, moving the jaw either too fast or too slowly entails larger energy costs. Significantly, the energetically optimal speed produced a jaw movement reminiscent both of the open-close alternation of 'canonical babbling' and the 'syllabic' organization of speech. We take this as preliminary support of the idea that speech movements may be shaped by the same principles that govern other motor behaviors.
    However, a caveat is in order. A closer look at basic muscle physiology indicates that mechanical work is only part of the story. The key factor in physiological energy consumption is not only whether a muscle shortens (=does work) or not, but whether it contracts or not. All muscle contractions, whether isotonic or isometric, cost metabolic energy. This energy comes from the oxidization of foodstuffs, in other words, from a process using oxygen.
    In experimental biology the energy cost of a given motor behavior is estimated by measuring the subject's oxygen consumption. We are currently looking into various ways of reliably making such measurements for speech.
    Since energy costs are much smaller in speech than in whole body movements, their measurement poses several difficult problems. At the meeting we will review some of the options and give a preliminary progress report on their feasibility.