STIMULUS CONTROL OF BEHAVIOR Chapter 10 Stimulus Control

STIMULUS CONTROL OF BEHAVIOR Chapter 10

Stimulus Control of Behavior 2 Ø Generalization l Ø Responding in the same manner to similar stimuli. Discrimination Learning l Responding in different ways to different stimuli.

The Generalization Process 3 Ø Ø Ø Stimulus generalization/discrimination occurs frequently in the real world Sometimes it is undesirable Ø Black Americans hold negative stereotypes of their own group (e. g. , racial prejudice). Ø Children over generalize rules of language (e. g. , Christine “goed” to the hockey game) Sometimes it is undesirable Ø Republican talk show hosts are idiots. Ø It “snowed” today.

Generalization Gradient 4 A visual representation of the response strength produced by stimuli of varying degrees of similarity to the training stimulus. Ø Excitatory generalization gradients l l A graph showing the level of generalization from an excitatory conditioned stimulus (S+) to other stimuli. Trained on S+, and then tested in extinction on S+ and other values. Many studies of these gradients employ pigeons, which have excellent color vision.

Gutman and Kalish (1956) 5 Group 530 Group 550 Group 580 Group 600

Inhibitory Generalization Gradients 7 Ø Weisman and Palmer (1969) illustrate the inhibitory-conditioning generalization gradient. l Green disk (S+) rewarded with food on a VI-1 min. schedule. l White vertical line (S-) on green disk no food. l Tested with a series of white lines ranging from 0 -90 degrees off the vertical on green disk

Results from 5 pigeons 8 Smartest Pigeon Dumbest Pigeon

Hull-Spence Theory 9 According to the Hull-Spence view, discrimination learning develops in 3 stages. First, conditioned excitation develops to the S+ as a result of reinforcement. Second, nonreinforcement in the presence of the Sresults in the development of conditioned inhibition to the S-. Finally, the excitation and inhibition generalize to other stimuli.

Predictions 10 Ø Ø The Hull-Spence model predicts a steeper generalization gradient with discrimination training than with nondiscrimination training. The maximum response occurs not to the S+, but rather to a stimulus other than the S+, and in the stimulus direction opposite that of the S+.

The Peak Shift Phenomenon 12 Hanson (1959) reported responding might be shifted from S+. The greatest response for discrimination-training subjects was not to the S+ but to a stimulus away from S+ opposite the direction of the S His results support Spence-Hull

Peak shift 13 The shift in the maximum response, which occurs to a stimulus other than S+ and in the stimulus direction opposite that of the S- In contrast, pigeons receiving nondiscrimination training responded maximally to the S+ The overall level of response was higher with discrimination training than with nondiscrimination training, which the Hull—Spence model did not predict.

14

15 Is Generalization Passive or Active? Ø Both! Ø Passive generalization: Perceptual confusion does happens because the stimuli can’t be distinguished. Ø Active generalization: In other situations, animals or people may generalize to stimuli that are very different to the conditioning stimulus

Passive Generalization Process 16 Ø Lashley-Wade theory l Animal is unable to distinguish between the generalization test stimulus and the trained stimulus. Thus, failure to discriminate between training and test stimuli is responsible for stimulus generalization.

Lashley and Wade Predictions 17 1. Perceptual experience should influence the amount of stimulus generalization. 2. Discrimination training results in generalization only to stimuli very similar to the conditioning stimulus 3. Generalization to stimuli dissimilar to the training stimulus should occur when nondifferential reinforcement training is used. 4. Preexposure to a common element may reduce generalization

Ducklings raised in mono-light 18 Monochromatic-Light (Evidence for #2) Normal-Light (Evidence for #1)

Jenkins and Harrison (1960) 19 Nondifferential (Tone always “On”) Differential (Tone, VI, No Tone: EXT (Evidence for #3)

20 Mackintosh, Kaye, Bennett (1991) Grp Preexposed: Lemon Drink Grp Novel: No Lemon Drink Test Lemon + Saline Lemon + Sucrose Conditioning Lemon + Saline followed by Li. Cl illness

21 Example of Active Generalization Acquired Equivalence: Generalization between physically different stimuli trained alike Group 1 Noise No Food Tone Food Click Food Group 2 Noise No Food Tone No Food Click Food Testing Click Shock Generalization Test: Tone (more fear) vs. Noise Testing Click Shock Generalization Test: Tone and Noise (equal fear)