Event Detail

Today’s dominant view of perception begins with the fast, parallel detection of “basic” features or parts followed by a slow, attention-demanding integration of those features into objects. Gestalts present a challenge for this approach because they entail wholes that behave as basic, in that they are perceived quickly ­ or even more quickly - than the parts and features from which they emerge. This talk will review what we know about Gestalts and identify methods by which they can be diagnosed, including Garner Interference, redundancy losses, configural superiority effects, false pop-out, and negative search slopes. I will describe a way to define Gestalts in terms of emergent features and suggest a new and possibly more principled approach that builds Gestalts from the ground up and measures them on a uniform scale.