some principle of grouping were able to assign the F tones to a different per-
ceptual stream, the order of A and B might become audible again. With this in
mind, they introduced yet another group of tones, labeled C (for ‘‘captors’’) in
figure 9.7. They varied the frequency of these C tones. When they were very
low, much lower than the frequency of the F tones, the F tones grouped with
the AB tones and the order of A and B was unclear to the listeners. However,
when the C tones were brought up close to the frequency of the F tones, they
captured them into a stream, CCCFFCC. One reason for this capturing is that
tones tend to group perceptually with those that are nearest to them in fre-
quency; a second is that the F tones were spaced so that they fell into a regular
rhythmic pattern with the C tones. When the capturing occurred, the order of
AB was heard more clearly because they were now in their own auditory
stream that was separate from the CCCFCC stream. The belongingness of the F
tones had been altered, and the perceived auditory forms were changed.
Scene analysis, as I have described it, involves putting evidence together into
a structure. Demonstrations of the perceptual systems acting in this way are
seen in certain kinds of illusions where it appears that the correct features of
the sensory input have been detected but have not been put together correctly.
Two examples will make this clearer.
The first is in vision. Treisman and Schmidt carried out an experiment in
which a row of symbols was flashed briefly in a tachistoscope.^4 There were
three colored letters flanked by two black digits. The viewers were asked to
first report what the digits were and then to report on the letters. Their reports
of the digits were generally correct, but the properties of the letters were often
scrambled. A subject might report a red O and a green X, when actually a green
O and a red X had been presented. These combinations of features often
seemed to the viewers to be their actual experiences rather than merely guesses
based on partially registered features of the display. The experimenters argued
that this showed that the human mind cannot consciously experience disem-
bodied features and must assign them to perceived objects. That is, the mind
obeys the principle of belongingness.
The second example comes from audition. In 1974, Diana Deutsch reported
an interesting illusion that could be created when tones were sent to both ears
of a listener over headphones. The listener was presented with a continuously
Figure 9.7
A tone sequence of the type used by Bregman and Rudnicky (1975).
224 Albert S. Bregman