An experiment by Bregman and Rudnicky (1975) asked whether streams are created by attention or by a pre-attentive mechanism. The strategy of the experiment was to cause a grouping mechanism to capture material away from attention, thereby showing that the grouping mechanism was not part of attention.
This demonstration is very similar. The figure shows a number of tones, symbolized as letters. Those called A and B are to be judged for their order -- ascending or descending in pitch. The order shown in the figure is ascending (AB), but in the Bregman-Rudnicky experiment, it could be either ascending (AB) or descending (BA).
When the AB or BA pair is played in isolation, it is easy to judge the order of A and B. This is illustrated in the first part of the demonstration which plays the pair first in the order BA, as a standard, then in the order AB (repeating this comparison twice).
In the second part of the demonstration, we make this order discrimination very difficult by surrounding the AB or BA pair by two instances of tone X (1460 Hz), to generate the four-tone sequences XABX, or XBAX. Despite the fact that we are being asked to discriminate the same AB pair in the two- and four-tone sequences, the addition of the two bracketing tones makes the task very difficult. This is probably because the four tones form a higher-order unit in which the A and B tones lose their prominence, in favor of the X tones, which fall at the perceptually prominent beginning and end positions. In the second part of the demonstration, a two-tone standard (AB) is followed by a four-tone test sequence XABX. Although AB is in the same order in the two- and four-tone sequences, it is very hard to judge that this is so.
In the third part, we show that, paradoxically, we can restore some of the original salience of A and B by embedding the four-tone XABX (or XBAX) sequence in an even longer one, in which there are a sequence of C (captor) tones, which fall at the same frequency as the X tones, and have the same intertone spacing as the X tones, C--C--C--C--XABX--C--C (in which each "-" represents a silence of the same duration as A or B). The X's are captured into an isochronous single-frequency stream with the C's, rejecting A and B from this stream. This partially releases A and B from interference by the X's. Accordingly, in the final part of the demonstration, when a two-tone standard, BA, is followed by the full ten-tone sequence that contains the C's, it is fairly easy to tell that the AB order is different in the long sequence.
This demonstration illustrates the fact that there is a competition in the grouping of tones, so that adding more tones to a sequence can change the perceptual organization. In the present demonstration, this change in grouping changes the isolation of some of the tones, and hence their perceptibility.
Next, the two-tone target is bracketed by a pair of tones of identical frequencies. Again, the pattern is played twice. Judge whether the order of the middle two tones is the same as in the standard.
Finally, we add a sequence of tones of the same frequency and timing as the flanking tones. This pattern is presented twice. Judge whether the decision about the order of the target tones is easier or harder than in the previous case.