Perception

There are many musical sequences in which the rhythm is evident from the mere durations of the notes. A simple theory is proposed of how a listener may infer the rhythm of such a sequence by comparing the note lengths. It is assumed that the listener forms an idea of the rhythm as the sequence unfolds, constructing and eliminating metrical hypotheses in the light of what he hears. The theory, which differs in some important respects from earlier proposals, has been implemented as a computer program. The program has been tested on a wide variety of musical examples, and its successes and failures are discussed in detail.

It is shown, by elementary mathematical reasoning, that visual contours can only be concave at their endpoints. This simple natural fact is contrasted with general mannerisms of draftsmen: in the great majority of cases contours are drawn to have convex ends. It is argued that this results from our general concept of solid shapes: a general shape is conceived of as a conglomerate of convex (‘ovoid’) elementary shapes, these shapes act as ‘figure’ and the way they are glued together is treated as the—relatively unimportant—‘ground’. The hypothesis is supported through citations from academic-art literature. An attempt is made to give a geometrical specification of just what draftsmen draw if they disregard the contour.

Previous investigators have reported some difficulty in obtaining evidence for adaptation to optically reduced size. A study is reported in which a ballistic target-pointing task was used to examine such adaptation. In the first experiment size was reduced either by means of negative spherical lenses, or cylindrical lenses which reduced image size in only one dimension. Both lens types produced a negative aftereffect of adaptation, which increased as a function of lens strength. There was also a tendency for cylindrical lenses to give larger aftereffects. In a second experiment, in which a modified adaptation procedure was used, adaptation to cylindrical distortion was found to be significantly greater than adaptation to equivalent spherical distortion. These data are interpreted with respect to the relative amounts of distance distortion induced by viewing through the different lens types. The third experiment employed a cylindrical lens and a mental rotation procedure to determine whether adaptation in the first two experiments may have been visual or visuomotor in nature. Results indicated that adaptation was not purely visual. It is concluded that it is possible to demonstrate adaptation to size distortion by using an appropriate procedure.

Subjects observed a briefly exposed square and set a variable square to match its size. In experiment 1 the square was black or white and was preceded or followed by a flash of light. The preceding adaptation field was complete darkness or an illuminated field. Marked underestimation of size occurred when flash followed the square in all conditions, compared with trials with no flash. Preceding flash produced a small reversal of this effect, the extent of which depended on whether the square was black or white. Absence of an illuminated adaptation field reduced estimates in all conditions except those where flash preceded the square. In experiment 2, a black square was used, and the stimulus duration and the interval between stimulus and flash were varied. No significant effects were found with preceding flash. When flash followed the square, apparent size increased linearly over the first 300 ms with the logarithm of the interval between onset of the square and onset of the flash, then levelled off. Experiment 3 demonstrated that the reduction in apparent size occurred with figure and interfering flash delivered to the same eye but not with dichoptic interference. The results are explained by reference to a suggestion of Georgeson that the optimum frequency of spatially tuned receptors decreases in impoverished viewing conditions. Hence the presented figures trigger receptors which normally are responsive to smaller figures, and the presented figures are therefore categorized as being smaller than they really are.

The problem of the origins of the perception of causality in infancy has received relatively little attention in the literature despite its obvious importance. Two experiments with infants 4 1/2 and 8 months old are reported which seek to investigate sensitivity to spatiotemporal continuity in simple causal events with a differential dishabituation-of-looking technique. In the first experiment inanimate events of the familiar ‘billiard-ball launching’ type were used, while in the second animate events involving a hand/object pick-up were presented. The results suggest that both age groups of infants were sensitive to certain changes in spatiotemporal continuity in both types of event, although in the case of the inanimate stimuli the younger infants reacted less positively. It is suggested that infants in the first year of life are sensitive to certain spatiotemporal event configurations and that this sensitivity could be regarded as at least a required component of a perception of causality.

Induced motion is not limited to continuous motions presented on a frontoparallel plane. Experiments were conducted to investigate several varieties of induced motion to which theories of induced motion must apply. The observer indicated the perceived path of motion of a vertically moving test point to which induced motion at right angles to the physical motion was added by the motion of two inducing points. In experiment 1 all motions (both apparently and physically) were in a frontoparallel plane. It was found that discrete displacement as well as continuous motion of the test and inducing points produced substantial amounts of induction. In experiment 2 the inducing points were continuously moved in stereoscopic distance rather than remaining in an apparent frontoparallel plane. A large amount of apparent motion in depth was found in the vertically moving test point and was interpreted as an induced motion in depth. In experiment 3 an alternative interpretation of the phenomenon of experiment 2, in terms of an apparent vergence for the two images of the test point, was investigated and found to be unlikely. In experiment 4, with all the points moving continuously in a frontoparallel plane, eye motions as well as induced motions were measured, with the observer fixating either the test point or an inducing point. Substantial amounts of induction were obtained under both conditions of fixation. The consequences of these findings for theories of induced motion are discussed.

Two cartoon animation sequences were presented to subjects to test whether the human visual system prefers to construct the mental representation of movement by transforming already existing forms. The results support the view that the sensory system has developed economical methods for rapidly picking up the structure of any situation and prefers to use revision rather than reconstruction to account for a change in sensory stimulation.

The contribution of eye dominance to the large individual differences found in the speed of stereoscopic fusion has been examined. The degree of sensory dominance was measured and correlated with the time taken to perceive depth in three random-dot stereograms of varying complexity. A significant positive correlation was found with the simple stereogram and nonsignificant positive correlations were obtained with the two more complex stereograms. The results indicate that a high degree of eye dominance can significantly interfere with the speed of fusion of simple stereograms.

Two rhesus monkeys viewed black/white photographic slides depicting rhesus, human, chimpanzee, and schematic faces with direct gazes. Eye-track apparatus was used to assign visual fixations to one of four facial regions: the two eyes, nose, or mouth. Results showed that the eyes of stimulus faces received a disproportionate number of fixations from both observers across all stimulus face types. Stimulus faces depicting rhesus and human facial gestures shifted scan patterns somewhat, but did not disrupt the preoccupation with eyes. When the features of schematic faces were rearranged into non-facelike configurations, fixations directed to schematic stimuli were typically reduced in number.

Factors which govern the optimum perception of odours have not yet been defined. This has hindered the development of standard methods and instruments for measuring olfactory responses. As an initial step towards defining these conditions, inhalation rates and volumes, number of sniffs, and sniff duration were measured for twenty-three humans in odour-threshold and odour-intensity tests with pentyl acetate, 1-butanol, and diethylamine. Measurements were made with the aid of a hot-wire anemometer concealed within the outlet of an air-dilution olfactometer. Individuals varied markedly in their sampling techniques but maintained their characteristic sniffing patterns with different odours and olfactory tasks. Only three parameters consistently varied with odour concentration: total volume of odour sampled, total sampling time, and number of sniffs. Maximum inhalation rate was remarkably stable and was independent of the type, concentration, and pleasantness of odour. Values recorded for sniff volumes and inhalation rates indicate that most olfactometers in use do not accommodate human inhalation requirements during a sniffing episode. The many common characteristics in the varied sampling techniques of different subjects suggest that the techniques are close to those providing optimum odour perception. Whether these are inherited or developed through habit or experience is not known.