CHAPTER 8 Notes 7/11/2012 5:59:00 PM
Measuring Stimulus Control
Demonstrating that an organism responds differently to one stimulus than to
another is necessary for concluding that behavior is under stimulus control.
Imagine a simple experiment. When a key-light is illuminated with red light
to the key are reinforced, but when the key is illuminated with green light
are not reinforced. After several training sessions, pigeons will display high
of pecking when the red stimulus is present and near-zero rates when the
stimulus is presented. Pigeons thus show differential responding –
one way to one stimulus and a different way to a different stimulus.
Consider an experiment involving a more complex stimulus. In this
pigeons were reinforced on a VI schedule for pecking a key that was
with a white triangle on a red background. Note that such a stimulus has two
prominent elements: red color and a white triangle. (A stimulus with
elements is called a compound stimulus – in reality, every stimulus is a
compound stimulus). Reynolds (1961) wondered which of these elements
controlled responding. To answer this question, he introduced occasional
in which only the white triangle was projected on the key. On other test
only the red color was projected on the key. He found that some pigeons
responded at high rates when the red background was presented and at low
rates when the white triangle was presented. Other pigeons showed the
opposite pattern. Thus, for some birds pecking was strongly controlled by
red background; for others pecking was controlled by the white triangle. Differential responding indicates stimulus control. For an organism’s
be under stimulus control requires that the organism is able to discriminate
among different stimuli present.
How different must two stimuli be in order to be identified as different
One way to answer this question is to train a response to a particular
and then confront the subject with additional examples of the stimulus but
features differ to varying degrees. A classic example of this approach was
provided by Guttman and Kalish (1956). Pigeons were trained to respond to
key illuminated with light measuring 580 nanometres (nm). This wavelength
appears yellow-orange. After training, subjects were then presented with a
variety of stimuli of different wavelengths, ranging from 520 nm (greenish)
nm (reddish), and responses were not reinforced.
Maximal responding occurred at 580 nm – the training stimulus.
nearly as much responding occurred to wavelengths of 570 nm and 590 nm.
Responding thus generalized to these wavelengths. As the wavelength
more different from 580 nm, responding diminished. The graph shown above
called a stimulus generalization gradient. Such gradients provide
information about the nature of stimulus control. A fairly flat gradient, for
instance, indicates that the stimulus being varied does not control behavior.
flat gradient is obtained when equal responding occurs to all stimuli
steep gradient, on the other hand, indicates that the stimulus being varied
does control behavior.
What Factors Influence Stimulus Control?
It is a trivial point that an organism’s sensory capacity will influence which
or stimulus elements it perceives and which stimuli come to control
Bats and noctuid moths detect ultrasound, for example, and therefore can
under the control of ultrasonic stimuli. Humans do not hear ultrasound and
obviously our behavior cannot be controlled by such stimuli.
When we are hungry we are more sensitive to food-related cues. Do you
differently walking through the food-court of a mall with a full stomach than
do with an empty stomach? Yes. Yes, you do.
Behavior is more likely to come under stimulus control by an intense
than by a weak stimulus. This phenomenon is known as blocking. Blocking
was demonstrated by Kamin – remember him? Kamin investigated
fear using the conditioned suppression procedures. Rats were trained to
lever in an automated chamber and were occasionally presented with a
signaling that a shock would be delivered. For one group of rats, the
was a soft tone; for a second group, the stimulus was a compound of the
tone plus a bright light. After several sessions, both groups of rats were
the chamber with only the soft tone presented. Kamin observed conditioned suppression (withholding of lever pressing) by the group trained with the
alone. The group trained with the soft tone plus the bright light failed to
conditioned suppression. For this group, the presence of the light during
interfered with or “blocked” conditioned fear to the soft tone.
It should be pointed out, however, that there is another way to interpret this
finding. What if subjects did not perceive the tone and the light as separate
elements? What if instead, they perceived the tone and the light together as
stimulus configuration? If this was the case, the failure to respond on test
shown by the overshadowing group may have occurred because of a failure
generalize responding to the test stimulus, which was different than the
configural stimulus to which they had been trained. In contrast, test and
trials were identical for the control group (both involved only the soft tone).
the failure to demonstrate conditioned suppression in the overshadowing
might reflect a failure in generalizing responding to the test stimulus. This
purported failure would be called “generalization decre