Relational classification in behavior analysis and ecological psychology


Publication date


Series/Report no



Høgskolen i Oslo og Akershus. Institutt for atferdsvitenskap

Document type


Master i læring i komplekse systemer


The thesis discusses relational classification in behavior analysis and ecological psychology. First, relational classification within behavior analysis is discussed in a theoretical article. The article aims to technically define the distinction between structure and function. The definition endeavors to be (a) in line with previous usage, (b) pay homage to the idea that different yet continuous functional levels can be studied in their own right, and (c) be empirically accountable. The philosophical concept of multiple realizability is used as a theoretical reference point. It is argued that structure-function relationships can be explicated as operationalized multiple realizability. Three criteria for structure-function relationships are distilled. First, structural variables s1…sn must converge on the same relational variable(s) C1…Cn. Second, the structural variables s1…sn must diverge from each other by criteria D1…Dn. Third, the structural variables s1…sn must be relevant for instantiating the relational variables(s) C1…Cn. This definition is tested out in an interdisciplinary and dual purposed experiment, conducted at the University of Cincinnati. The study aims to exemplify structure-function relations as data and introduce the concept of dynamic agent-environment fit from ecological psychology. Dimensionless ratios, or “pi-numbers”, are used to investigate time-based avoidance behavior in a ball transportation task. Plastic play-pen balls were delivered 8 meters away from a wooden box on a progressive and non-contingent interval schedule. Participants were asked to transport the delivered balls into the box. The balls had to be transported rapidly, so that new balls would not stack at the delivery location. The delivery intervals would change gradually from 2-12 or 12-2 seconds. Participants learned to adapt to these changing conditions and time their transportation attempts to avoid stacking. The temporal properties of this avoidance behavior, when considered in proportion to the delivery interval, proved to be an invariant ratio regardless of absolute time. Participants spent 80% of the delivery interval to move towards the box and back. In addition, they moved faster towards the box than back to the delivery location. This invariance is revealed when the data is described using pi-numbers. These pi-numbers were used as a reference point to evaluate structure-function relations of the time-based avoidance behavior. The velocity at which participants moved is cautiously put forth as a candidate for structure in relation to the pi-numbers as function.




Permanent URL (for citation purposes)