A 3 STATE MARKOV MODEL FOR DISCRIMINATION LEARNING

Previous treatments of discrimination learning have been ineffective in predicting both marginal responding and reinforcement effects in two-stimulus probability learning. A three state Markov model is proposed which is designed to account for these results. The basic assumption of the theory is that the subject is reinforced for behaving appropriately to a distinctive cue rather than for making a specific motor response. Empirical results that include variations of π, β, and cue similarity are accounted for by the model. © 1969.