[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Motor theory of absolute pitch

Rebecca and others have suggested that absolute pitch may rely to
some degree on muscle memory.  I have two comments.

First, I believe that absolute pitch is of interest because it
putatively involves some sort of unusual/special memory ability.  It
appears that possessors have stable long-term representations of
musical pitch, and they are able to categorize or label these
representations with linguistic labels.  If AP is subserved by
muscle-memory, that doesn't make it any less interesting it seems to
me -- we are simply specifying the type of memory that is involved,
but it is still a feat of memory.

Second, and more relevant, the late Dixon Ward and Ed Burns conducted
a study that addresses this issue head on.  Ward and Burns (1978)
denied auditory feedback to trained singers who possessed absolute
pitch (forcing them to rely solely on muscle memory); the singers
erred by as much as a minor third, or three semitones.   Thus muscle
memory was only enough to get them in the ball park, and did not
account for their AP ability.

Ward, W. D. & Burns, E. M. (1978) . Singing without auditory
feedback. Journal of Research in Singing and Applied Vocal Pedagogy,
1(2), 24-44

Dan Levitin

That sounds like feedback control via the auditory system. You might
look at how continuous-wave (CW) bats do it. Note also that CW bats
use doppler shift to detect insect wingbeats and possibly to
declutter the scene. Schuller, et al., 1974, in "Response to
frequency shifted artificial echoes in the bat _Rhinolophus_
_ferrumequinum_," J. Comp. Physiol., 89:275-286, reports that R.
ferrumequinum maintains its CF with an observed accuracy of +/- 50
Hz. This bat's resting frequency is about 82.5 KHz, so that
corresponds to about 0.06 percent variation. For a human singer at
440 Hz, a similar accuracy would be +/- 0.2-0.3 Hz. Hyperacuity
probably plays a role.
Harry Erwin, PhD, Senior Lecturer of Computing, University of
Sunderland. Computational neuroscientist modeling bat bioacoustics
and behavior. <http://world.std.com/~herwin>