Comparative judgment: Tests of two theories using the Baldwin figure.

The filled-unfilled space illusion was investigated as a possible explanation for the asymmetry (difference in strength) between tails-out and tails-in versions of the Miiller-Lyer illusion. Previous work has shown that removal ofthe horizontal shaft makes both configurations look shorter, but affects them unequally and only partially reduces the asymmetry. In two experiments, illusion strength was varied by use of either different tail lengths or single-vs. double-ended figures. Subjects used method of adjustment with stimuli presented by computer. When the Miiller-Lyer was sufficiently weak, shaft removal had equal effects on both tails-in and tails-out configurations, but less effect on plain lines. This produced a reversal of the usual asymmetry, that is, tails-in was stronger than tails-out. It is concluded that the asymmetry is caused by the filledunfilled illusion. This may require reassessment of the view that the two Miiller-Lyer versions are different illusions.As early as 1909, Lewis argued that the Miiller-Lyer illusion could not be explained by any single principle. This view is now supported by a variety of studies, which have found differences between the tails-out and tails-in versions of the illusion. Of these, the best known difference is an asymmetry of strength. When tested separately, the tails-out (long) version is almost always stronger than the tails-in (short) version. In addition, the tails-out and tails-in versions respond differently to changes in such variables as the angle between shaft and tail, distance between tail tips (Sekuler & Erlebacher, 1971), and viewing duration (Erlebacher &Sekuler, 1974;Pollack, 1964). Finally, the two versions also respond differently to amputation of one or more of the tails (Day & Dickinson, 1976;Greist-Bousquet & Schiffman, 1981). Sekuler and Erlebacher noted that such differences are not easily handled by any single theory. They recommended that the two versions of the Millier-Lyer be treated as separate illusions.A slightly different strategy is to analyze the multiple causes of the illusion in such a way that factors responsible for the differences between versions can be separated from those that contribute similarly to both versions. The present study pursues this strategy by searching for the factor or factors that cause the asymmetry of strength.The most promising lead comes from Cooper and Runyon (1970), who reported that elimination of the horizontal shaft connecting the two sets of tails causes a I thank David Dexter and the Alma College Computer Center for technical support, Douglas Boyne and Steve Neff for help with data collection, and Carol Slater for a critique of the writing. Requests for reprints may be sent to Walter Beagley, Department of Psychology, Alma College, Alma, MI48801. 45 reduction in the strength of the tails-out illusion and an increase in the strength of the tails-in version. Although they did not include a plain-line control from which to measure asymmetry, the direction of change they reported should produce a reduc...

Section: Discussionmentioning

confidence: 98%

Interaction space of Müller-Lyer with filled-unfilled illusion: An explanation of Müller-Lyer asymmetry

Beagley

1985

“…Models predicting that illusion must be ruled out as explanations for the present results. Such models include those postulating comparative judgment based on information integration (Clavadetscher & Anderson, 1977) or contextual effects of frame size (Brigell, Uhlarik, & Goldhorn, 1977).…”

Section: Resultsmentioning

confidence: 99%

The major-axis effect: Axes of bilateral symmetry or loci of neural interactions?

Hartley

1978

Distortions in perceptual judgments have been attributed to the axes of bilateral symmetry of surrounding figures. In the figures used in prior investigations, the axes of bilateral symmetry have been confounded with loci where heightened neural interaction might be expected to occur. In the present experiment, subjects reproduced dots surrounded by figures for which the axes of symmetry and loci of hypothesized neural interactions differed. The observed displacements supported the hypothesis that distortions were due to spreading neural effects of the surrounding figures rather than to the effects of axes of bilateral symmetry. need not be due to lateral inhibition, nor need they be due to cortical rather than retinal interaction. It is only necessary to assume that real lines give rise to neural effects that spread away from their borders.The bilateral symmetry hypothesis predicts that the apparent location of a point within the boundaries of the figures should be displaced toward a real line, a figural contour, or an implicit line, an axis of bilateral symmetry, as shown in Figure 2, whichever is closer to the point. The neural interaction hypothesis predicts that points should only be affected by surrounding physically present lines. Whether the locations are displaced toward the lines or away from them depends on whether excitatory or inhibitory effects are posited. The bilateral symmetry hypothesis predicts that points falling on a virtual contour will show no displacement whatever. The neural interaction hypothesis predicts that, even after displacement, points lying on a line of heightened neural interaction will still lie on that line. For example, points along the diagonal of the square may appear closer to or farther from the corner of the square, but they will remain equidistant from the borders of the square.

“…From a perceptual standpoint, indeed, these tasks seem somewhat artificial. Tasks that involve more perceptual kinds of integration, such as bisection and various illusions (Anderson, 1974a;Clavadetscher & Anderson, 1977), deserve more attention, both as a basis for measurement and for inherent perceptual interest.…”

Section: Averaging and Differencingmentioning

confidence: 99%

Bisection of loudness

Carterette

Anderson

1979

Self Cite

In a loudness bisection task, subjects varied one sound to lie halfway between two given sounds in terms of loudness. The two given sounds were varied from 30 to 90 dB in a 4 by 9 factorial design. Functional measurement methods based on monotone analysis provided good support for the bisection model, and yielded a loudness scale with an exponent of about .3, except for a falloff at lower intensities. Two other tasks, judging average loudness and difference in loudness of the two given sounds, yielded mixed results. In Experiment 2, in particular, the differencing judgments were not additive, even under monotone transformation. These analyses also indicated that previous applications of monotone analysis have typically lacked adequate power to allow any conclusion about the operative model. Overall, the present bisection scale agrees with Garner's lambda scale, and the present theoretical approach agrees with that of Garner in its emphasis on algebraic models as a foundation for psychological measurement.The quality of loudness has good reasons for being a primary focus of psychophysical measurement. It has enormous dynamic range, in contrast to skin temperature or pressure. It is an energetic, sensory dimension, in contrast to length. It has nearly the ecological importance of brightness, but not nearly the complications from adaptation. And it is experimentally flexible and easy to control. For these and other reasons, loudness has become a primary test case for theories of psychophysical measurement.The method of bisection is attractive because it rests on sensory comparison and avoids artificial response language. If the subject's bisection response was indeed at the sensory midpoint, then a simple procedure of successive bisection would yield a linear (equal-interval) scale of sensation. Two arbitrary initial stimuli would be called 0 and 100. Their bisector would then have the value 50, two further bisections would yield the 25 and 75 points, four further bisections would yield the 12.5, 37.5, 62.5, This work was supported by National Institute of Mental Health Grant MH-07809 to the first author, and by National Science Foundation Grant BMS 74-19124 to the second author. A preliminary report of this work was presented at the XXI International Congress of Psychology, Paris, July 1976. We wish to thank Alan Barnebey for help with apparatus and computing, Cheryl Smart for programming and managing the experiments, Clifford Butzin and Anthony Harris for advice and assistance in data analysis, and Edward Alf for developing the HP-67 program used for fitting power functions. Requests for reprints may be sent to Edward Carterette, Psychology, University of California. Los Angeles, Los Angeles, California 90024, or to Norman Anderson, Psychology C-009, University of California, San Diego, La Jolla, California 92093. and 87.5 points, and so on, to as fine a division as desired. Granted that the bisection response is at the sensory midpoint, these numbers would constitute a linear scale of loudness.But the method of ...