Academic journal article International Journal of Psychology and Psychological Therapy

Shifting Differentiation and Its Implications for the Response Concept1

Academic journal article International Journal of Psychology and Psychological Therapy

Shifting Differentiation and Its Implications for the Response Concept1

Article excerpt

As Galbicka (1988) emphasized, differential reinforcement is central to opérant conditioning research: «Response differentiation is opérant conditioning, and vice versa» (p. 343). In differential reinforcement, responses that meet the experimenter's criterion have the scheduled effect (e.g., access to food) and increase in relative frequency as a result. Other responses, which do not meet the criterion, decrease in frequency relative to the reinforced responses. Differentiation is this change in relative frequencies.

An experimenter can repeat differential reinforcement across two or more related criteria. For instance, in successive differentiation, successive criteria specify increasingly narrow subclasses of the original class of behavior (e.g., Herrick, 1964). As a second example, in shaping, successive criteria specify classes that have some overlap with each other but increasingly less overlap with the original class (and eventually perhaps no overlap). Examples of research on shaping have included training animals to move to a location not previously occupied by them (Pear & Legris, 1987), to deposit ball bearings down a hole (Midgley, Lea, & Kirby, 1989), to jump onto a platform and extend their noses downward over the edge until they lost their balance and fell (Rasey & Iversen, 1993), and to press a bar (Stokes & Balsam, 1991).

An experiment by Eckerman, Heinz, Stern, and Kowlowitz (1980) suggests a third type of repeated differential reinforcement. Pecks by pigeons on the two rightmost keys of a row of 20 keys produced food. After many sessions, the experimenter changed the keys at which pecks produced food. The food-effective keys were shifted from the right edge of the row across various intermediate keys to the left edge, then back to the right edge via various intermediate keys, and so on. At each repetition of differential reinforcement, the new food-effective keys were one, two, or three keys from the previous location. This change in the reinforced location followed a varying number of reinforcers for pecks at the currently-reinforced location.

Eckerman et al. (1980) interpreted their experiment as concerning shaping. However, Midgley et al. (1989) doubted this interpretation on three grounds. First, after the first horizontal sweep across the keys, successive locations were increasingly likely to have a previous reinforcement history. In contrast, with shaping, newly shaped acts have no such history. second, in Eckerman et al. (1980), locations were under continuous reinforcement or extinction. In shaping, intermittent reinforcement commonly precedes extinction. Third, no location was a precursor to, or a component of, an earlier location. In shaping, earlier reinforced acts are precursors to, or constituents of, each newly-reinforced act (e.g., rearing anywhere as a precursor to rearing over the bar).

The experiment by Eckerman et al. (1980) might not demonstrate shaping but nonetheless does demonstrate a type of repeated differential reinforcement. This type of repeated differential reinforcement might be characterized generically as follows. First, there are n classes (of locations, forces, latencies, etc.). second, during any one part of an experiment, only one of these classes is reinforced. Third, the order in which the classes are reinforced follows a pattern specified by the experimenter (e.g., a threekey shift after 100 reinforcers at the currently-reinforced location, then a two-key shift after 100 reinforcers, then a three key shift followed by a two key shift, and so on.). For convenience in the following discussion, differential reinforcement with these characteristics is referred to as shifting differential reinforcement.

The present experiment extended shifting differential reinforcement to a simple game-like task with human participants. The experiment was conducted to describe what participants do under shifting differential reinforcement. Data analysis in Eckerman et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.