It always stuck in my head that in Snow Crash, when Uncle Enzo is trying to get the drop on Raven at the end, he opens his mouth to hear better.
That’s my random association from this Sci Am news story:
In the study, a specially designed robotic device stretched the mouths of volunteers slightly up, down or backward while they listened to a computer-generated continuum of speech verbalizations that sounded like “head” or “had,” or something in between. When the subjects’ mouths were stretched upward, closer to the position needed to say “head,” they were more likely to hear the sounds as “head,” especially with the more ambiguous output. If the subjects’ mouths were stretched downward, as if to say “had,” they were more likely to hear “had,” even when the sounds being generated were closer to “head.” Stretching subjects’ mouths backward had no effect, implying a position-specific response. Moreover, the timing of the stretch had to match that of the sounds exactly to get an effect: the stretch altered speech perception only when it mimicked realistic vocalizations.
Ostry and his colleagues hope to help answer this question with follow-up work that inverts the experiment: instead of hearing a continuum of sound, subjects will endure a continuum of stretches to see if auditory input can influence what they feel.
And that reminds me of Alexander technique.