DURHAM, N.C. -- For over two thousand years, musicians and
scientists have puzzled over why some combinations of musical
tones played together sound more harmonious than others. Now,
Duke University Medical Center perception scientists David
Schwartz, Ph.D., Catherine Howe, M.D., and Dale Purves, M.D.,
have presented evidence that variation in the relative
harmoniousness, or "consonance," of different tone combinations
arises from people's exposure to the acoustical characteristics
of speech sounds. Schwartz and Howe are postdoctoral fellows,
and Purves is Director of the Center for Cognitive Neuroscience
and the George B. Geller Professor of Neurobiology.

The researchers said that their findings, reported in the
Aug. 6, 2003, issue of the Journal of Neuroscience, constitute
an important advance in understanding the biological basis of
music perception. The work also extends to hearing the
theoretical framework about brain organization that Purves and
his colleagues developed in earlier work on visual
perception.

Those studies of vision led to the idea that evolution -- as
well as individual experience during development -- created a
visual system in which perceptions are determined by what a
given visual stimulus has typically signified in the past,
rather than simply representing to an observer what is
presently "out there." That work is summarized in a new book
titled Why We See What We Do (Sinauer Associates,
2003).

In their Journal of Neuroscience paper, the neurobiologists
present a statistical analysis of the patterns of frequency and
amplitude in human speech sounds, based on a collection of
recorded utterances spoken by more than 500 people. They found
that the points at which sound energy is concentrated in the
speech spectrum predict the chromatic scale -- the scale
represented by the keys on a piano keyboard. Moreover, the
difference in the amount of sound energy concentrated at these
points predicts the relative consonance of different chromatic
scale tone combinations.

These results suggest that certain pairs of tones sound more
harmonious than others because they are physically similar to
the patterns of sound energy most familiar to human listeners
from their exposure to speech, said the researchers.

In deciding to analyze speech as a natural basis for tone
perception, the researchers were faced with a very different
challenge from that of exploring visual perception. In the work
on vision, Purves and his colleagues concentrated on analyzing
the perception of visual illusions.

"After studying the research literature on psychoacoustics,
we discovered several phenomena in tone perception that,
despite having been investigated for decades, remained
unexplained," said Schwartz. "Our general framework is that the
way to understand why somebody perceives anything the way they
do -- whether the stimulus is light or sound-- is to consider
the possible real world events that could have given rise to
that particular stimulus.

"This work on music perception represents a natural
extension of the work on visual perception," said Howe.
"Hearing presents many of the same challenges as vision, in
that the physical world cannot be known directly; we only know
about objects in the environment because of the energy
associated with them, such as light waves or sound waves.

"Determining the actual state of the environment on the
basis of this indirect information available to our senses is a
real challenge. The solution we have evolved is evidently to
respond to ambiguous optical and acoustical stimuli by taking
account of the statistical relationship between stimuli and
their sources. That seems to be the reason we hear music the
way we do."