PERCEPTION 



1629 



1.5 msec. It has often been suggested that the neural 

 basis for response to the minute time differences 

 involved in lateralizing clicks must be some spatial 

 separation in the central nervous system of the 

 impulses corresponding to the two asynchronous 

 stimuli (172, 240, 380). 



It is obvious that these stereophonic phenomena 

 are analogous in several ways to the experiments on 

 binocular stereopsis. Both sets of experiments suggest 

 mechanisms capable of a degree of accuracy which 

 is difficult to reconcile with the supposed imprecision 

 of central nervous connections. Moreover, for visual 

 and auditory space perception, actual achievements 

 exceed what is possible under the restricted con- 

 ditions of a motionless observer. A stationary listener 

 can judge sounds as emanating in the median plane, 

 or to the left or right of it, but his sensory information 

 is ambiguous in other respects, e.g. as to whether the 

 sound is in front or in back of him. In more natural 

 settings, where head movements are not restricted, 

 such confusions do not arise. As Wallach (517) has 

 shown, normal auditory localization utilizes the 

 sequential changes in binaural stimulation which 

 occur systematically as the observer moves his head 

 through space. The ubiquity of (his form of auditory 

 motion parallax can be illustrated by the way in 

 which it fails when the observer is given misin- 

 formation regarding the relationship between stimulus 

 motion and body movement. 



AUDITORY LOCALIZATION DURING HEAD AND BODY 



movements. Assume that the observer can rotate his 

 head in a holder to the left or right while listening to 

 a sound directly in front of him. If the sound source 

 (in a dark room) is made to move with every head 

 movement (so that it keeps its relative position to 

 the head), the listener perceives it directly above his 

 head (and not in front of him); it is from this over- 

 head position that the cars would normally continue 

 to receive invariant stimulation, in spite of varying 

 rotation of the head to the left or right (517). If the 

 listener is permitted to tilt his head (and not only to 

 rotate it), the mislocalization disappears. An identical 

 mislocalizaton can be obtained by keeping the 

 observer motionless inside a rotatina; striped drum. 

 The sound source, which is stationary, is again 

 directly in front of the observer (but invisible to him 

 because it is outside the drum). The striped drum 

 is rotated at sufficient speeds to induce illusory 

 motion of the observer's body in the direction op- 

 posite to that of the vertical black stripes. At this 



point, the sound is again perceived as directly over- 

 head (517). 



These simple experiments underscore the crucial 

 role of normal patterning in the relation between 

 movements of the observer and the corresponding 

 relative motion of stimuli for the proper maintenance 

 of spatial organization. 



Interaction Between Posture and Distance 

 Receptors in Spatial Localization 



It is also obvious that localization of external events 

 usually involves concurrent perception of one's own 

 posture. This information enters into the 'constancies' 

 of perceived direction in space. 



EFFECTS OF BODY TILT: THE AUBERT PHENOMENON 



and its variants. When we tilt our heads, vertical 

 lines in the environment remain vertical as if the 

 signals generated by the postural change were able 

 to counteract precisely the inclination of lines on 

 our retina. The precision is diminished when a 

 tilted observer attempts to set a luminous line to the 

 vertical in an otherwise dark room [Aubert's experi- 

 ment in i860 (13)]. With moderate lateral body 

 tilts (up to 30 '), the settings are close to the true 

 vertical but deviate from it in the normal adult by 

 several degrees in a direction opposite to the tilt of 

 his body, i.e. there is overcompensation or over- 

 constancy [the T'.-phcnomenon' of Miillci , , | 1 

 With larger body tilts, the correction lor abnormal 

 posture undershoots instead, and the line is set so 

 that it falls short of the true vertical in the direction 

 of the body tilt [the original Auberl phenomenon 

 (13), or 'A-phenomenon 1 of Muller I ;-,) Quanti- 

 tative information on the effects ol bod) tills beyond 

 30 ° is scarce and insufficient to permit an analvsis 

 ol the underlying mechanisms (in terms ol feedback- 

 control theory), as has been accomplished so con- 

 vincingly by Mittelstaedt (349) for visuopostural 

 interaction in the praying mantis and by von Hoist 

 (503) lor fish orienting themselves in abnormal 

 gravitational fields. 



For moderate tilts (30 and less) there is consider- 

 ably more information. The effects here are not 

 limited to vision, but exist analogously in the tactile 

 and auditory spheres. A tilted subject, blindfolded, 

 adjusts a palpated rod to the vertical in such a way 

 that he overcorrects by a few degrees, although 

 congenitally blind subjects do less so (47). Again, 

 blindfolded subjects, tilted to 30 to their right or 

 left, adjust a single (ambient) sound to the apparent 



