Matching Concept

Simultaneous Matching to Sample (SMS) Task

three projectors - sample stimulus comes up on top projector - animals makes a response - delay period with no stimuli - two bottom projectors produce stimuli - animal must select same stimulus

Behavioral Rule

stimulus rule: animal forms a set of rules (if circle on top, then respond circle bottom)

configuration rule: treats all three projectors as one stimulus, learns the configurations

Cognitive Rule

a concept of same/different

independent of the stimuli

animal understands it must respond to the stimulus on the bottom that matches the stimulus on top

Transfer Test

give animal new stimuli in an SMS task

if it does well, it is using a cognitive rule, if not it’s using behavioral rule

Orthogonality

new stimuli must be very different from the first stimuli otherwise you get stimulus generalization

Savings — Evidence of a Cognitive Rule

when transferring, the number of trials it takes to learn the new stimuli needs to be less than the number of trials it took to learn the old stimuli to show evidence of a cognitive rule

if you train for 500 trials and test for 500 trials you are not testing for a matching concept but instead teaching a new task

Matching Concept in Chimps

Oden, Thompson, and Premack

trained to 83% correct with nondifferential reward (got reward irrespective of getting it correct or not)

method: two stimuli (cup and lock) - presented chimps with one stimulus, took it away, then brought back both - tested on one of three transfer tasks (novel object, fabric swatches, food items)

outcome: significant performance on novel object and fabric transfer, evidence of a matching concept

not all monkeys got the food transfer because they were likely just reaching for food and not thinking of the task

Matching Conce