In experiments measuring reaction time (RT) to a stimulus masked by metacontrast (Fehrer and Biederman, 1962; Fehrer and Raab, 1962; Schiller and Smith, 1966; Harrison and Fox, 1966), it was reported that subjects could perform a voluntary response to “an event whose presence is not even suspected by the reacting S,” (Fehrer and Biederman, 1962, p. 130). This paradoxical conclusion was arrived at on the basis of the fact that S’s RTs to the presentation of a test stimulus (TS), followed in 75 msec or less by a pair of flanking masking stimuli (MS), were much too short to be attributed to the onset of the MS, although Ss were unable to report any perception of the (TS). In fact, the RTs to the TS followed by the MS were faster, in some cases, than the RTs to the TS alone, even though faster, in some cases, absent.” the TS in the former case was reported as “phenomenally absent.”
Ostensively, the conclusions drawn from these experiments seem to support the concept of “subliminal perception.” Such at least is implicit in Kahneman’s evaluation of these reaction time experiments:
However, the more radical conclusion appears warranted that subjective visual experience is not causally involved in triggering a simple response, since the two occur at about the same time. Indeed, a response can occur when the stimulus that triggered it is not represented at all in experience. (Kahneman, 1968, p. 421) {5}
A careful analysis of the above experiments will show that the interpretation of RT to masked stimuli in terms of subliminal perception or in terms of voluntary reactions to phenomenally absent stimuli is not logically warranted by the data.
Fehrer and Raab (1962), citing numerous earlier meta- contrast experiments, state that the TS is “phenomenally absent” at stimulus onset asynchronies (SOA) or 70 to 100 msec but that TS is “not, however, without effect on perception since under these conditions there is an awareness of movement involving the second stimulus,” (p. 143). What is meant by describing the TS as phenomenally absent is that the S is not aware of the TS as a separate, distinct event, i.e., the S is not aware of the TS as adifferentiable or discriminable from other things” which is Efron’s criterion of what is meant by perceiving something (Efron, 1967). But the S reports that he perceives movement of the masking stimuli from the center out-ward (Fehrer and Raab, 1962; Kahneman, 1968), which means that the S perceived something move from the TS area out toward the flanks, which means that the TS was perceived by the S. The fact that Ss attribute the apparent movement to the MS does not warrant the conclusion that the TS was phenomenally absent.
To illustrate this point, consider a S sitting in a movie house watching an animated cartoon of a balloon “moving” across the screen. If asked what he perceived, the S would undoubtedly report that he saw “a balloon moving across the screen.” If the experimenter were then to ask, “Which balloon image appeared to move?” the S would probably look at the {6} experimenter askance and reply, “What do you mean, ‘which balloon’? There was only one balloon; the one that just moved off the screen traveling from left to right!” Certainly it would be absurd for the experimenter to conclude that the earlier frames of the balloon film were “phenomenally absent” merely because the S reported seeing only one balloon.
The conclusion that the TS in the metacontrast experiment is phenomenally absent involves the same error and is based on the fact that the TS is not perceived as a separate event. What is perceived is a single event of movement attributed to the MS. It is necessary, therefore, to distinguish, as Efron (1967) does, between the objective event and the perceived event, i.e., the experience of the objective event. In the metacontrast experiment of Fehrer and Raab (1962), employing an array of three identical luminous squares, a center square serving as the TS flanked on either side by two other squares serving as the us, two objective events (viz. the TS and us) are perceptually integrated into a single experience of apparent movement, which differs phenomenally from either of the perceptions produced by presenting the TS or the us alone.
It is a well-known fact (Kahneman, 1967, 1968; Breitmeyer et al., 1974; Stoper and Banffy, 1977) that those situations producing maximal metacontrast masking are also maximally effective in producing beta movement. When, for in- stance, the TS is substantially different in form from the MS, masking and no movement is observed (Uttal, 1970c). {7} Masking and movement are maximized when the TS and us have common boundaries occurring over the same extent. On the other hand, if the boundaries of the MS significantly extend beyond those of the TS, the masking and movement effects are greatly reduced or eliminated (Werner, 1935; Fehrer, 1966). TS and IS stimuli of similar form, intensity and duration are most effective for producing metacontrast masking and beta movement. Using such stimuli, one obtains a U-shaped type B masking function (Kahneman, 1967, 1968) with maximal masking occurring at SOA between 75 and 120 msec, This function is virtually identical to that obtained by plotting apparent motion against SOA, in which case apparent movement is also most effectively obtained at SOA of 75 to 120 msec (Kahneman, 1967). Furthermore, essentially identical functions are obtained for all TS and MS exposure durations between 25 and 125 msec. This means that, for brief stimuli, stimulus onset asynchrony (SOA) and not interstimulus interval (ISI) is the important temporal factor influencing metacontrast and apparent motion.
The relationship between metacontrast masking and apparent motion was repeatedly demonstrated by Werner (1935) in his comprehensive study, although Werner tended to de-emphasize the importance of this relationship. Almost all of Warner’s stimuli were black figures against a white background, arranged so that the outer borders of the T8 were partially or totally congruent with the inner borders of the MS. Describing his famous ring-disk sequence, Werner reported (1) that the {8} optimal SOA range for masking the disk by a subsequently presented ring fell within the range of apparent movement, and (2) that only those elements of the TS will undergo apparent movement and be “masked” which possess a strong geometrical and close spatial relationship to the MS.
Two experiments taken from Werner’s study will serve to illustrate this latter point.
Experiment 24: The TS consisted of a black ring super- imposed over a black disk. The inner border of the ring was congruent with the outer border of the disk, and the disk was placed eccentric to the ring. The MS was a second ring, identical in dimensions to the ring portion of the TS (with the exception, of course, that the TS ring appeared interrupted by the overlapping region of the eccentric disk). The MS was presented so that its center was coincident with that of the black disk component of the TS. When the TS and MS were alternately presented, the disk never disappeared and the ring was seen to undergo motion from its TS to its MS position (see Werner’s Fig. 28).
Experiment 25: TS consisted of a ring (5) whose inner diameter was equal to the outer diameter of a second ring (b) serving as the MS. The TS also contained a disk concentric to its surrounding ring 3 but whose diameter was slightly less than the inner diameter of ring 3 (see Warner’s Fig. 29). When TS and US were alternately presented, the resulting perceptual effect was described by Werner as follows:
(1) The black disk always remains visible (even {9} at a speed greater than the optimum). (2) Moreover there can be seen a movement of ring a as it shrinks towards the size of ring 3. (Werner, 1935, p. 60)
Thus, Werner was able to demonstrate, by using different variations of ring-disk stimuli, that successive rings can be made to interact so as to produce beta or radial motion without affecting the perception of a disk presented simultaneously with the first ring and concentric to one or both of the rings. When, on the other hand, a single ring is made to follow a concentric disk, the former may be seen to undergo radial expansion, rendering the disk, qua disk, “invisible”.
The observation that radial expansion motion occurs under conditions producing maximal metacontrast was also ob- served by Schiller and Smith (1966), using the Werner ring-disk display but employing luminous rather than black stimuli. This arrangement was employed intentionally to relate the experiments of Fehrer and Raab (1962) to those of Werner (1935). By using a TS duration of only 5 msec, Schiller and Smith were able to maintain a negligible ISI-SOA difference. They obtained a U-shaped function showing a maximum masking effect at ISI between 60 and 70 msec (corresponding to SOA between 65 and 75 msec). Discussing their S’s experiential reports under stimulus conditions producing maximal masking, Schiller and Smith write:
We believe that the processes which underlie meta- contrast are in part analogous to those which are involved in apparent movement. At certain ISIS (those at which metacontrast is maximal) the disk seems to expand and “turn into” the ring. The perception of this double change–both in position (from inner disk to outer ring), and in configuration, {10} leads to the inference that the disk is no longer present. (Schiller and Smith, 1966, p. 38) (emphasis added)
The results of all of the experiments discussed thus far support the conclusion that the perception of apparent movement requires that successive stimuli occur within a certain SOA range and possess a certain degree of similarity. When these requirements are not met, no apparent movement will be perceived. This may be explained by the assumption that the visual experience of apparent movement proceeds as a consequence of the neural integration of information from two distinct objective events; anything which interferes with this integration will necessarily interfere with the perception of movement.
Latest revision: April 8, 2014 5:37 pm
bioperipatetic 