for this—if you ask most non-musicians to sing an ‘‘E-flat,’’ they will not
understand. As a term project when I was a student in the Stanford CCRMA
psychoacoustics class, I designed a test to determine whether non-musicians
could demonstrate AP capabilities. The first test was to determine if non-
musicians had an ability to remember pitches over a long period of time—even
if they hadn’t learned the fancy labels that musicians use. These subjects were
given tuning forks, and they were asked to carry the forks around with them
for a week, bang them every so often, and to try to memorize the pitch that the
forks put out. After a week the tuning fork was taken away, and a week later
the subjects were tested on their memory for the tone. Some of them were
asked to sing it, and others had to pick it out from three notes played to them.
The distribution of the subjects’ productions is shown in figure 13.2. Notice that
the modal response was perfect memory for the tone, and those who made
errors were usually off by only a small amount.
Perhaps, then, absolute musical pitch is an attribute of sound that is encoded
in long-term memory. In spite of all the interference—the daily bombardment
by different sounds and noises—the subjects were able to keep the pitch of the
tuning fork in their heads with great accuracy. A harder test would be to study
non-musicians’ memory for pitch when that pitch is embedded in a melody.
Because melodies are transposition-invariant, the actual pitch information may
be discarded once a melody is learned. On the other hand, if somebody hears a
melody many times in the same key, we might expect that repeated playings
would strengthen the memory trace for the specific pitches.
To test whether people can reproduce the absolute pitch of tones embedded
in melodies, I asked subjects to come into the laboratory and sing their favorite
rock ‘n’ roll song from memory (Levitin 1994). The premise was that if they had
memorized the actual pitches of the songs, they would reproduce them. It
would then be easy to compare the tones they sang with the tones on the orig-
inal compact disc (CD) version. Rock songs are especially suited to this task
because people typically hear them in only one version, and they hear this over
and over and over again. Contrast this with ‘‘Happy Birthday’’ or the national
Figure 13.2
Results of pitch memory in non-musicians. The subjects were asked to retain the pitch of a tuning
fork in memory for one week.
304 Daniel J. Levitin