Behavioral dynamics of affordance transitions

Date of Completion

January 2009


Psychology, Experimental|Psychology, Cognitive




Whether an object affords grasping with one or two hands depends on its size relative to an individual's hand size. In experiments that manipulate whether object size increases or decreases in a trial, the transition between the two grasping modes may not occur at the same object-size/hand-size ratio. In four experiments these transitions were evaluated as the outcome of a self-organized dynamical process by means of the grasping-transition (GT) model (Frank, Richardson, Lopresti-Goodman, & Turvey, 2009). The perception of a behavior was indexed by selective action (e.g., grasping with two hands) or verbal classification (e.g., "two hands"). Parameter estimates derived from previously conducted experiments (Lopresti-Goodman, Richardson, Baron, Carello, & Marsh, 2009) suggested that the degree of positive hysteresis (increasing transition ratios greater than the decreasing transition ratios) depended on two parameters expressing (a) the interaction between the one- and two-hand attractors, and (b) the strength of the two-hand attractor. Experiment 1 evaluated the model's ability to account for manipulations of task difficulty on grasping transitions via manipulations of a cognitive task that was performed concurrently with perception indexed by selective action. The results indicated that increases in the model's parameters and positive hysteresis were proportional to task difficulty. In Experiment 2 the selective action and verbal indices were compared directly in a within-subject design. Whereas selective action yielded critical point transitions, verbal classification yielded negative hysteresis. In terms of the GT model, verbal classification yielded larger interaction parameter values and smaller two-hand attractor values than selective action. Additionally, the model exhibited the anomalous feature of a range of object-hand ratios for which both the one- and two-hand grasping modes were unstable. Experiment 3 sought, unsuccessfully, to reverse the negative hysteresis characterizing verbal classification by adding some of the intermediary non-grasping behaviors present in selective action. The model parameters for experiment 3 matched those for verbal classification in Experiment 2. Finally, Experiment 4 generalized the GT model and negative hysteresis to verbal classification of the affordance "sit-on-able". The general discussion focused on the strengths and weaknesses of the dynamical model and the challenge of negative hysteresis for future investigations. ^