Decrement of the Brentano Müller-Lyer Illusion as a Function of Inspection Time

Two experiments are reported in which the decline or decrement in the magnitude of the Brentano Müller-Lyer illusion was measured. Observers made a pre-test judgment and, after a variable intervening time period, a post-test judgment of illusion magnitude. In experiment 1, the intervening time periods were 1, 2, and 3 min during which time the independent groups of observers allocated to each of the three time periods either systematically scanned the Brentano figure (inspection conditions) or waited until the intervening period had elapsed (no-inspection conditions). Experiment 2, which included an additional 5 min intervening time period, evaluated a response-bias explanation for the results of the inspection conditions of experiment 1. Taken together, the findings of the two experiments indicate that sheer inspection of the Brentano figure produces illusion decrement. However, illusion decrement was independent of the duration of the inspection period, with equivalent amounts of decrement occurring across the range of viewing times examined in the two experiments. The pattern of these results suggests that theories of Müller-Lyer decrement must incorporate a factor attributable to, or correlated with, inspection time, whose effect in reducing illusion magnitude is confined mainly to the first 1 or 2 min of active visual inspection of the Brentano illusion figure.

Illusion Decrement as a Function of Duration of Inspection and Figure Type

Illusion decrement for the Müller-Lyer and Horizontal-Vertical illusions was examined. The experiment consisted of an initial adjustment of an illusion followed by 20 test trials, each with an intervening 60-sec. intertrial interval during which a comparator line and a standard line set to equality were visually inspected for 0, 20, 40, or 60 sec. After each intertrial interval the length of the comparator line was reset by the experimenter to either 0 or 90 cm, and subjects then adjusted its length to perceived equality with the standard line (42 cm). Illusion decrement was inversely related to the duration of inspection for each illusion, with significant reductions in magnitude of illusion observed for all groups. These results support prior demonstrations that perceptual learning mechanisms are operative during brief periods of visual inspection, especially when these periods are followed by the opportunity to make repeated adjustments.

The Horizontal Vertical Illusion: Evidence for Strategic Factors in Feedback-Induced Illusion Decrement

This study assessed whether feedback would improve performance on a horizontal vertical illusion task and whether such improvement would transfer to different forms of the illusion. Subjects shortened the extended vertical line of small, medium, and large inverted-T figures to be equal in length to the horizontal line before and after being shown a correctly adjusted medium-sized figure (visual feedback). Next, they were tested on three alternate forms of the illusion; a production task (drawing 1-in. lines in the horizontal and vertical planes), an adjustment task using L figures, and a task requiring them to choose which of several sailboat drawings had mast height equal to hull length. Prior to feedback, vertical lines were adjusted shorter than the horizontal lines on each size inverted-T figure. After feedback on the medium-size figure, experimental subjects were more accurate than the controls on each size inverted-T figure. The results suggest that transfer of illusion decrement was also obtained for the boat-selection and, to a lesser extent, the L-figure adjustment tasks but not to the line-production task. These findings are consistent with the notion that improvement after feedback is not due to structural changes in visual processing or to simple feedback-induced compensation in performance but involves some strategic or cognitive change in judging line length.

Visual Inspection Alone Produces a Decrement in the Horizontal-Vertical Illusion

Subjects inspected either inverted-T or L figures for 5 min. and then shortened the extended vertical lines of inverted-T figures in an attempt to make the vertical line equal in length to the horizontal line. Those who inspected inverted-T figures were more accurate on initial trials than controls (who did not inspect figures), those who inspected L figures, or those who inspected inverted-T figures and made adjustments on the inspected figures. The results indicate that visual inspection alone can produce a decrease in the Horizontal-Vertical illusion and may account, in part, for changes in strategic factors which have been hypothesized to underlie illusion decrement.

Decrement and the Illusions of the Mueller-Lyer Figure

Decrement, a time-related decrease in the magnitude of the Mueller-Lyer illusion, was measured separately for the wings-out and the wings-in variants of the Mueller-Lyer figure. There were significant reductions of wings-out illusion magnitude during the decrement period. Observers viewing the wings-in segment showed a non-significant decrement pattern. Analyses of individual decrement patterns showed that illusion magnitude did not decrease for a number of observers even when there were significant time-related trends at the group level. Data for 80 observers imply that the mechanisms of perceptual learning proposed by previous models of Mueller-Lyer illusion decrement are not sufficient explanations of the decrement process.

Illusion Decrement and Transfer of Illusion Decrement in Müller-Lyer Figures

In the work reported in the literature the reduction or decrement in the magnitude of the Müller-Lyer illusion with continued inspection has been typically investigated with the use of the composite illusion form. Three experiments are reported in which the illusion decrement was separately examined in the underestimated (wings-in) and the overestimated (wings-out) forms of the Müller-Lyer illusion, with particular attention paid to the transfer of illusion decrement between the two forms. Decrement occurred in both forms of the Müller-Lyer illusion, although there was considerable intersubject variability in decrement effects, and nonuniform rates of decrement across the inspection period. In none of the experiments did transfer of illusion decrement between the two forms occur. It is argued that the attentional/differentiation hypothesis of illusion decrement provides a plausible account of the present findings as well as of those found with the composite Müller-Lyer figure.

The Method of Reproduction and Mueller-Lyer Changes

In a three-variable study of Mueller-Lyer decrement, each of eight independent groups of 10 subjects responded 30 times to one of the two Mueller-Lyer components by the Method of Reproduction. The mean of the initial Out response was significantly greater absolutely than that of the In response. All three interactions were significant. The In figure demonstrated illusion decrement across the 30 trials, but the Out figure demonstrated illusion increment.