[ Abstract ] - [Bibliografia] - [Tematiche]

Kanizsa, G., Minguzzi, G. F. (1986). An anomalous brightness differentiation. Perception, 15, 223-226.

An anomalous brightness differentiation


Gaetano Kanizsa
Istituto di Psicologia, Università di Trieste
Gian Franco Minguzzi
Istituto di Filosofia, Università di Bologna,

Received 14th March 1986, in revised form 26 May 1986
1 Introduction
When a white homogeneous surface surrounded by a black area is divided by a thin black line, the two resulting regions can appear slightly different in brightness (see figure 1). Since the conditions of brightness contrast are identical for the two regions, this unexpected effect is not easily explained. In fact, it cannot be accounted for by any simple physiological mechanism such as lateral inhibition or frequency filtering. Furthermore, it does not seem obvious to invoke organisational factors, like figural belongingness or figure-ground articulation (Kanizsa 1979; Minguzzi 1981; Kanizsa and Minguzzi 1984).
Informal observations indicated that some subjects did not notice any difference in brightness, while others were confident about its existence. Therefore, the first aim of this research was to obtain a quantitative measure of the occurrence of the effect in an experimental setting. A secondary aim of this paper is to present a hypothetical account of the phenomenon.


Figure 1. When a region of homogeneous luminance is divided by a thin line an anomalous difference in brightness may appear (Kanizsa 1979). The phenomenon is enhanced when a tissue-paper sheet (Florkontrast) is used.
2 Method
2.1 Stimuli
The experimental displays consisted of a white (Munsell number 9) or gray (Munsell number 6) ring surrounded by a white (Munsell number 9), gray (Munsell number 7.5), or black (Munsell number 3) ground. The ground was a 12 cm x 12 cm cardboard square. The outside diameter of the ring was 33 mm; the inside diameter was 22 mm.
Six rings, three white and three gray, were each divided into two equal parts by two thin black lines (0.5 mm wide). Six rings, three white and three gray, were each divided into two unequal parts (270° and 90°). The remaining six rings were undivided. The white rings on white ground were outlined by black lines (0.5 mm wide).
The eighteen experimental conditions, shown in table 1, were obtained by combìning two ring colours (white, gray) with three ground colours (white, gray, black) and three ring patterns.
2.2 Subjects and procedure
Thirty-six volunteers, undergraduates or graduates of the University of Trieste, were employed as subjects. In order to explain the task, each subject was presented with a Benussi - Koffka (B - K) ring, ie a black and white square with a gray ring in the middle divided by a thin black line continuing the black/white edge (MacKavey 1969; Anderson et al 1975). All subjects reported a difference in brightness between the two halves of the B - K ring.
The experimental patterns were presented in a random order for 2 s by means of a Gerbrands tachistoscope. The luminance of the white ground was 4 cd m¯². The patterns divided into two equal parts were presented with the dividing line horizontal; the patterns divided into unequal parts were presented with the smaller sector at the top. Subjects were asked: (i) to report any brightness difference of the kind observed in the B - K ring, and (ii) to point out which area of the ring, if any, was brighter.

3 Results
No brightness differences were noticed by seven out of the thirty-six subjects. The remaining twenty-nine subjects noticed a difference with a frequency that varied from a minimum of 4 (one subject) to a maximum of 13 (two subjects). Table 1 gives the overall frequencies of dishomogeneity judgements, and the relative frequencies of topbrighter and bottom-brighter judgements.
Table 1. Distribution of dishomogeneity judgements and the relative frequency of top-brighter and bottom-brighter judgements in the eighteen experimental conditions.
Ring/ground
colours		 Ring
pattern		 Brightness
difference		 Top
brighter		 Bottom
brighter
White/whiteundivided
equal
unequal		 0
8
13		 0
6
11		0
2
11
White/grayundivided
equal
unequal		 2
17
21		 2
13
16		 0
4
5
White/blackundivided
equal
unequal		 4
20
22		 2
11
15		 2
9
7
Gray/whiteundivided
equal
unequal		 0
18
26		 0
13
24		 0
5
2
Gray/grayundivided
equal
unequal		 0
21
27		 0
13
25		 0
8
2
Gray/blackundivided
equal
unequal		 2
21
9		 2
9
15		 0
12
4

Inspection of the data suggests that the ring/ground colour relationship does not seem to exert a noticeable influence. Clearly, dishomogeneity judgements occur almost exclusively with the divided rings. In fact, only 8 dishomogeneity judgements out of 241 were obtained with the undivided rings (3.3%). There were, however, many dishomogencity judgements with the two types of divided rings: for the rings divided into two equal parts, 105/216 (43.6%), and for the rings divided into unequal parts, 128/216 (53.1%). The distribution of top-brighter and bottom-brighter judgements is shown in table 2. The data indicate that this illusory brightness difference is influenced by both the relative size and the spatial position of the sectors X² = 12.91, p < 0.001).
Table 2. Distribution of top-brighter and bottom-brighter judgements in the ring patterns with equal and with unequal sectors.

Top
brighter		Bottom
brighter		Total
Equal sectors 65 40 105
Unequal sectors10622128
Total171 62



4 Discussion
With regard to the first aim of our research, the data support the conclusion that, under the conditions described, an anomalous brightness differentiatìon can be reliably obtained and systematically studied. The phenomenon is weak and this fact may explain why for a long time it escaped the attention of colour scientists. However, we deem it is worthy of study because its very existence challenges current theories of brightness perception. This phenomenon ìs not only incompatible with a first-order formulatìon of the constancy hypothesis, because unequal perceived brightnesses correspond to equal luminances, but also with the notion of brightness as a function of luminance ratio, because contrast conditions are even.
We can only be very speculative about the explanation of the effect. Our research, which was exploratory in nature, does not support any definite conclusion. Apart from its very existence, several aspects of the phenomenon require an explanation: (i) Its dependence on spatial position. The advantage of top-brighter over bottom-brighter judgements suggests the existence of a spatial anisotropy for brightness, analogous to that for size. The anisotropy could directly reflect a retinal functional difference between upper and lower hemiretinal systems, evidenced by neurophysiological research (Lehmann and Skrandies 1979). (ii) The relative size of the two regions divided by the thin line seems to be relevant. This effect may indicate that figure/ground organisation is involved. (iii) Finally, because the effect looks like a kind of brightness contrast it is rather surprising that the ring/ground luminance ratio is not influential.
Acknowledgement. This work was supported by CNR grant 83.02187.04.
References
Anderson N., Pine S, Rosenfeld A., 1975 "Derived scales for degree of simultaneous contrast in six Benussi ring figures Bulletin of thePsychonomic Society, 6, 289-292.

Kanizsa G., 1979 Organization in Vision (New York: Praeger).

Kanizsa G., Minguzzi G F, 1984 "Sulla differenziazione anomala del contrasto di chiarezza" in Fenomenologia Sperimentale della Visione ed. G Kanizsa (Milan: F. Angeli) pp 73 - 96.

Lehmann D., Skrandies W., 1979 "Multichannel evoked potential fields show different properties of human upper and lower hemiretina systems" Experimental Brain Research, 35, 151 - 159.

MacKavey W., 1969 "Spatial brightness changes in Koffka's ring" Journal of Experimental Psychology 82 405 - 409.

Minguzzi G. F., 1981 "Una ricerca sulla differenziazione anomala del contrasto di chiarezza" Reports from the Institute of Psychology, University of Trieste.