Purchase this article with an account.
Mary C. Portillo, James R. Pomerantz; Evaluating grouping via emergent features: A systematic approach. Journal of Vision 2005;5(8):350. doi: https://doi.org/10.1167/5.8.350.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
When basic elements combine and Gestalt grouping occurs, emergent features (EFs) arise. We identify the presence of EFs through configural superiority effects (CSEs) in an odd-quadrant task in which RT and accuracy measures are obtained for locating which of four stimuli differs from the other three. Our experiments introduce a systematic method in which EFs can be created from the ground up and in a hierarchical fashion. This allows for direct comparisons of grouping strength among EFs, something that up until now has been impossible. Thus we may be able to determine whether grouping based on certain EFs is stronger than that based on others and also whether the effects of multiple EFs are additive. We start with the simplest stimulus, namely a single dot. Making the identical change in the position of this single dot can yield sharply different EFs depending on the configuration of other contextual dots: with a one-dot context, changing the position of the target dot alters the proximity or angle between the two dots of the target-context pair. Similarly, with a two-dot context, the same change in the target dot position alters the linearity or symmetry among the three dots of the target-context triplet. (Note that asymmetric and non-linear configurations are impossible with two-dot patterns; they can emerge only with three). With a three-dot context, the same procedure yields the EF of surroundedness. Replacing pairs of dots with line segments connecting them, and then using these lines as primitives in place of the dots, yields the EFs of parallelism and collinearity. Thus, locating the odd quadrant becomes easier when that quadrant differs from the others on the basis of an EF such as proximity or orientation (defined by the distance or angle between two dots) than when it differs only in the position of a single dot. We demonstrate CSEs for the first four EFs and provide pilot data for some higher order ones.
This PDF is available to Subscribers Only