The Relational Determination of Length Illusions and Length Aftereffects

Abstract

When two lines were presented simultaneously, in an illusion paradigm, the judged length of one of the lines (the focal line) assimilated toward the length of the contextual line. When the lines were presented sequentially, in an aftereffect paradigm, the apparent length of the focal line was displaced away from the length of a previously presented contextual line (i.e. contrast). The largest effects obtained at inverse contextual-to-focal length ratios: 1.67:1 for contextual lines longer than the focal line and 0.60:1 for contextual lines shorter than the focal line. Overestimation of focal length was larger than underestimation for both illusions and aftereffects despite the fact that identical stimuli produced effects in opposite directions in the two paradigms. Results were discussed in terms of a neural length-coding model and were related to visual size constancy.

Orientation specificity of length assimilation and contrast

“…The mean of the 36 responses to this control line (2 responses from each of 18 observers) was 66.5 mm. This underestimation of objective test-line length has been consistently reported in studies that used the graded-series comparison scale (Brigell & Uhlarik, 1979;Jordan & Schiano, 1986;Jordan & Uhlarik, 1985. Additionally, Coren and Girgus (1978, p. 179) pointed out that this psychophysical "error of the standard" is often reported when observers attempt to match a standard length to a variable set of lines.…”

“…Ifassimilation is only produced by contextual stimuli within the attentive field, then progressively reduced assimilation would be expected in the 0°and 12°c onditions relative to Experiment 1. This should occur because the part of the contextual line remaining in the attentive field is similar in length to the test line, and lines of similar length do not affect one another (Brigell & Uhlarik, 1979). However, since the 90°contextual line is within the attentive field, and since the confound with intersection has been removed, we now have a better test of the ability of the 90°contextual line, and, to a lesser extent, the 45°line, to produce simultaneous length assimilation.…”

“…Brigell and Uhlarik (1979) argued that the framing ratio determined the magnitude of the parallel-lines illusion. However, the framing ratios were identical for the 0°conditions of Experiments 1 and 2, yet the illusion was present in one case (Experiment 1) but not in the other (Experiment 2).…”

“…Brigell and Uhlarik (1979) first reported that a 6O-sec inspection of a contextual line produced length contrast in a subsequently judged test line. Additionally, Jordan and Uhlarik (1985) reported that this length aftereffect could be induced by only 5 sec of prior inspection of a contextual line.…”

“…For example, at a viewing distance of 40 em, a 9-cm horizontal contextual line will produce overestimation of a parallel 6-cm test line located 5 mm below it. However, length contrast has been demonstrated in the array under conditions of simultaneous viewing with a much larger spatial separation (100 mm) of the same contextual and test lines (Jordan & Schiano, 1986) and under conditions of sequential presentation of contextual and test lines within the aftereffect paradigm (Brigell & Uhlarik, 1979;Jordan & Uhlarik, 1985). Jordan and Schiano (1986) proposed revisions of descriptive models of length distortion that account for the reversals from length assimilation to length contrast outlined above.…”