Skinner, B. F. (1957). A second type of superstition in the pigeon

Prévia do material em texto

University of Illinois Press is collaborating with JSTOR to digitize, preserve and extend access to The American Journal of
Psychology.
http://www.jstor.org
A Second Type of Superstition in the Pigeon 
Author(s): W. H. Morse and B. F. Skinner 
Source: The American Journal of Psychology, Vol. 70, No. 2 (Jun., 1957), pp. 308-311
Published by: University of Illinois Press
Stable URL: http://www.jstor.org/stable/1419345
Accessed: 02-04-2015 03:38 UTC
 REFERENCES
Linked references are available on JSTOR for this article: 
 http://www.jstor.org/stable/1419345?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at 
 http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
This content downloaded from 130.239.20.174 on Thu, 02 Apr 2015 03:38:08 UTC
All use subject to JSTOR Terms and Conditions
NOTES AND DISCUSSIONS 
A SECOND TYPE OF SUPERSTITION IN THE PIGEON 
When food is given to a hungry organism, any behavior in progress at 
the moment must be assumed to be reinforced by this event. When small 
amounts of food are repeatedly given, a 'superstitious ritual' may be set 
up. This is due not only to the fact that a reinforcing stimulus strengthens 
any behavior it may happen to follow, even though a contingency has not 
been explicitly arranged, but also to the fact that the change in behavior 
resulting from one accidental contingency makes similar accidents more 
probable. In an earlier experiment the automatic operation of a food- 
magazine every 15 sec. was found to induce hungry pigeons to engage in 
such ritualistic behavior as bowing, scraping, turning, and dancing.' In 
some cases the behavior was stable, in others the topography slowly 
changed; but in all cases superstitious effects survived indefinitely. Similar 
effects from the adventitious reinforcement arising from the presentation 
of discriminative stimuli have recently been observed.2 Such effects must 
always be allowed for in designing experiments on complex behavior. 
Accidental, but nevertheless effective, relationships may arise in the 
sensory control of operant behavior. For example, a stimulus present when 
a response is reinforced may acquire discriminative control over the re- 
sponse even though its presence at reinforcement is adventitious. Suppose, 
for example, that an organism is responding at a moderate rate on a 
variable-interval schedule of reinforcement, and let an incidental stimulus 
(A) occasionally appear for a brief period. Even though there is no 
explicit temporal relation between the appearance of A and the program 
of reinforcement, a response will occasionally be reinforced in the pres- 
ence of A. For a brief period the frequency of such reinforcement may 
be appreciably greater than in the absence of A. An organism which is 
sensitive to slight differences in rate of reinforcement will form a dis- 
crimination; its rate of responding in the presence of A will become 
greater than in the absence of A. This might be called a positive sensory 
superstition. If, on the other hand, reinforcements happen to occur rela- 
tively infrequently in the presence of A, a discrimination will develop 
'B. F. Skinner, 'Superstition' in the pigeon, J. exp. Psychol., 38, 1948, 168-172. 
2 W. H. Morse, An analysis of responding in the presence of a stimulus correlated 
with periods of non-reinforcement, Unpublished doctoral dissertation, Harvard Uni- 
versity, 1955. 
308 
This content downloaded from 130.239.20.174 on Thu, 02 Apr 2015 03:38:08 UTC
All use subject to JSTOR Terms and Conditions
NOTES AND DISCUSSIONS 
in the opposite direction, as the result of which the rate in the presence 
of A will be relatively low-a sort of negative sensory superstition. 
When an accidental contingency has produced a higher or lower rate of 
responding in the presence of an incidental stimulus, a second effect fol- 
lows. If the rate has fallen in the presence of A (because reinforcements 
have been relatively infrequent), responses will be even less likely to be 
reinforced in the presence of A. In the limiting case no responses will be 
made in the presence of A, and no response, of course, reinforced. More- 
over, reinforcements which are made available during A are not obtained 
because responses are not made. The first response following the with- 
drawal of A is then reinforced, and the discrimination is further strength- 
ened. Similarly, when the rate is increased during A because of favorable 
accidental reinforcements, all reinforcements set up during A are likely to 
be obtained, and if the preceding condition commands a relatively low 
rate, some reinforcements set up at that time may actually be obtained 
after A has appeared, to strengthen the discrimination. 
Both types of 'sensory superstition' have been demonstrated experi- 
mentally in the pigeon. The apparatus consisted of the usual experimental 
space 13 X 22 X 16 in. A small translucent plastic plate was mounted 
behind a 1-in. circular opening at head height on one wall. It was lighted 
from behind by an orange 6-w. bulb. The pigeon pecked this disk to 
operate the controlling circuit. Food was presented for reinforcement in 
an opening below this key. Water was available in the space. 
Three pigeons, two of which had previously been reinforced on other 
schedules, were placed on a variable-interval schedule of reinforcement 
with a mean interval of 30 min. The shortest interval between reinforce- 
ments was 1 min., the longest 59 min. Daily experimental sessions varied 
between 6 and 20 hr. in length. Body weight was maintained at approxi- 
mately 80% of the ad lib weight, and a reinforcement consisted of access 
to mixed grain for 5 sec. The resulting performance was at a low mean 
rate of responding with some local irregularity. Against this base-line, an 
incidental stimulus consisting of a blue light projected on the key instead 
of orange was introduced for 4 min. once per hour. The schedule of 
occurrence of this stimulus was independent of the programming schedule. 
The rate of responding was recorded continuously in the usual cumulative 
curve. Brief downward movements of the pen marked reinforcements, and 
the pen remained down during the 4-min. period of the incidental stimu- 
lus, thus slightly displacing the record made in the presence of the stimulus 
without changing its slope. 
Segments of records showing superstitious differences in rate under the 
309 
This content downloaded from 130.239.20.174 on Thu, 02 Apr 2015 03:38:08 UTC
All use subject to JSTOR Terms and Conditions
NOTES AND DISCUSSIONS 
control of the incidental stimulus are shown in Fig. 1. Except for the 
portions a, b, c, and d, the curves are characteristic of the base-lines ob- 
tained on the schedule described above. Curves A and B are for a pigeon 
which showed sustained periods of negative superstition. The overall rate 
generated by the schedule is relatively high (of the order of 0.5 responses 
{ R s/SEC. 
w / / 
10 MINUTES 
D 
FIG. 1. SUPERSTITIOUS DISCRIMINATIVE CONTROL OF 
RESPONDING BY AN INCIDENTAL STIMULUS 
The stimulus is presented for 4 min. once per hour, and is marked by the downward 
displacement of the recording pen (a, b, c, and d). Curves A, and B are segments 
of cumulative response-curves on a variable-interval schedule (mean interval = 30 
min.) for a pigeon witha lower rate when the incidental stimulus is present than 
when it is absent-a 'negative superstition.' Curves C and D are segments for 
another pigeon on the same schedule with a higher rate in the presence of the 
incidental stimulus-a 'positive superstition.' 
per sec.). Whenever the incidental stimulus appears, the rate drops to a 
low value or to zero (a and b). Records C and D are for another pigeon 
showing a positive superstition. The base rate is relatively low, and the 
rate in the presence of an incidental stimulus at c and d is clearly of a 
much higher order. 
The direction of the superstition is not necessarily stable. In a long 
experimental session a positive superstition may give way to a negative 
form, or vice versa. Such changes are usually easily explained in terms of 
adventitious reinforcement or failure to receive reinforcement in the pres- 
ence of the incidental stimulus. All three birds showed periods of both 
positive and negative superstition. 
There are several arbitrary features of such an experiment. In the 
present case the incidental stimulus was present 1/15 of the time during 
310 
This content downloaded from 130.239.20.174 on Thu, 02 Apr 2015 03:38:08 UTC
All use subject to JSTOR Terms and Conditions
NOTES AND DISCUSSIONS NOTES AND DISCUSSIONS 
a session. A relatively shorter period would be less likely to receive rein- 
forcements on a given schedule, and might be expected to produce negative 
superstition more frequently. At the other extreme, an incidental stimulus 
which occupied half the experimental session would presumably share so 
nearly equally in the reinforcements that there would be no substantial 
separation of rates. The schedule and the performance generated are also 
relevant in determining the frequency of adventitious reinforcement. 
Finally, the nature and intensity of the incidental stimulus also may have 
their effect. 
Pending an investigation of these parameters, it may at least be said 
that incidental stimuli adventitiously related to reinforcement may acquire 
marked discriminative functions.3 
Harvard University W. H. MORSE 
B. F. SKINNER 
SOME EFFECTS OF INTERMITTENT SILENCE 
Imagine that a monkey hits the keys of a typewriter at random, subject 
only to these constraints: (1) he must hit the space bar with a probability 
of p(*) and all the other keys with a probability of p(L) = 1 - P(*), 
and (2) he must never hit the space bar twice in a row. I wish to examine 
the monkey's output, not because it is interesting, but because it will have 
some of the statistical properties considered interesting when humans, 
rather than monkeys, hit the keys. 
In the monkey's output we will find runs of i letters in a row, where 
i = 1, 2, 3, . . ., separated by single spaces. We will expect to find runs 
of length i with a probability 
Pi = p(*)p(L)-'1, i = 1, 2, . ., [1] 
so that the probability of a word of length i will decrease exponentially as 
i increases. 
Now suppose that there are A different keys on the typewriter, exclud- 
ing the space bar. Then the number of different possible words of length i 
must be Ai. Since the probability of any particular word must be P, divided 
by the Ai different words of the same length, it must be equal to Pi/Ai 
= p(*)p(L)i-lA-i, on the assumption that all sequences of letters are 
equally probable. If we represent this quantity, the probability of a word 
of length i, by the symbol, p (w, i), we can write it as: 
p(w, i) =- [(*)/p(L) e-it?g A-log p(L)] [2] 
3This work was done under a grant from the National Science Foundation. 
a session. A relatively shorter period would be less likely to receive rein- 
forcements on a given schedule, and might be expected to produce negative 
superstition more frequently. At the other extreme, an incidental stimulus 
which occupied half the experimental session would presumably share so 
nearly equally in the reinforcements that there would be no substantial 
separation of rates. The schedule and the performance generated are also 
relevant in determining the frequency of adventitious reinforcement. 
Finally, the nature and intensity of the incidental stimulus also may have 
their effect. 
Pending an investigation of these parameters, it may at least be said 
that incidental stimuli adventitiously related to reinforcement may acquire 
marked discriminative functions.3 
Harvard University W. H. MORSE 
B. F. SKINNER 
SOME EFFECTS OF INTERMITTENT SILENCE 
Imagine that a monkey hits the keys of a typewriter at random, subject 
only to these constraints: (1) he must hit the space bar with a probability 
of p(*) and all the other keys with a probability of p(L) = 1 - P(*), 
and (2) he must never hit the space bar twice in a row. I wish to examine 
the monkey's output, not because it is interesting, but because it will have 
some of the statistical properties considered interesting when humans, 
rather than monkeys, hit the keys. 
In the monkey's output we will find runs of i letters in a row, where 
i = 1, 2, 3, . . ., separated by single spaces. We will expect to find runs 
of length i with a probability 
Pi = p(*)p(L)-'1, i = 1, 2, . ., [1] 
so that the probability of a word of length i will decrease exponentially as 
i increases. 
Now suppose that there are A different keys on the typewriter, exclud- 
ing the space bar. Then the number of different possible words of length i 
must be Ai. Since the probability of any particular word must be P, divided 
by the Ai different words of the same length, it must be equal to Pi/Ai 
= p(*)p(L)i-lA-i, on the assumption that all sequences of letters are 
equally probable. If we represent this quantity, the probability of a word 
of length i, by the symbol, p (w, i), we can write it as: 
p(w, i) =- [(*)/p(L) e-it?g A-log p(L)] [2] 
3This work was done under a grant from the National Science Foundation. 
311 311 
This content downloaded from 130.239.20.174 on Thu, 02 Apr 2015 03:38:08 UTC
All use subject to JSTOR Terms and Conditions
	Article Contents
	p. 308
	p. 309
	p. 310
	p. 311
	Issue Table of Contents
	The American Journal of Psychology, Vol. 70, No. 2 (Jun., 1957), pp. 161-332
	Probability, Learning, the Statistical Structure of Concepts, and the Measurement of Meaning [pp. 161-173]
	Probability-Learning with Two and Three Choices [pp. 174-185]
	The Role of Repetition in Associative Learning [pp. 186-193]
	Alternative Measures for the Discrimination of Shift in Reinforcement-Ratio [pp. 194-202]
	The Tartini Pitches Created by the Primary Number Seven [pp. 203-210]
	The Influence of Size and Shape on the Discrimination of Visual Intensity [pp. 211-218]
	The Effect of "Right" in a Modified Thorndikian Situation [pp. 219-226]
	An Analysis of Individual Differences in Olfactory Thresholds [pp. 227-232]
	The Generality of the Norms of Word-Associations [pp. 233-237]
	A Descriptive Analysis of the Non-Randomness of Serial Threshold-Data [pp. 238-242]
	Characteristics of Mental Impairment in Hypoxia [pp. 243-247]
	Memory for Visual Figures by the Method of Identical Stimuli [pp. 248-252]
	Incidental and Intentional Memory for Lifted-Weights [pp. 253-257]
	The Serial-Position Curve as a Function of Organization [pp. 258-262]
	Criteria in Conceptual Transposition [pp. 263-267]
	On the Validity of the Point-Assignment Procedure in the Constant-Sum Method [pp. 268-271]
	Restriction of Range in the Judgment of Single Stimuli [pp. 272-275]
	The Effect of Time on Size-Constancy [pp. 276-279]
	The Frequency Procedure for Analyzing the Data Obtained by the Method of Limits [pp. 280-282]
	Intersensory Transfer of Verbal Material [pp. 283-285]
	Manifest Anxiety, Discrimination, and Transposition [pp. 286-288]
	Verbal Compared with Manipulative Solution of an Apparatus-Problem [pp. 289-290]
	Apparatus
	Apparatus for the Study of Visual TranslatoryMotion [pp. 291-294]
	Continuous Recording of the Fluid-Intake of Small Animals [pp. 295-298]
	A Continuous Millisecond Control for the Gerbrands Tachistoscope [pp. 299-302]
	A Lever-Pressing Device for Fluid Rewards [pp. 303-305]
	An Inexpensive Variable-Pattern Maze [pp. 306-307]
	Notes and Discussions
	A Second Type of Superstition in the Pigeon [pp. 308-311]
	Some Effects of Intermittent Silence [pp. 311-314]
	The Use of Foreign Languages by Psychologists, Chemists, and Physicists [pp. 314-316]
	Word-Length as a Factor in Differential Recognition [pp. 316-318]
	An Interocular Color-Effect [pp. 318-319]
	A Group-Method for the Study of Figural After-Effects [pp. 319-320]
	Fifty-Third Annual Meeting of the Society of Experimental Psychologists [p. 320]
	Third Annual Meeting of the Southeastern Psychological Association [p. 321]
	Book Reviews
	Review: untitled [pp. 322-323]
	Review: untitled [pp. 323-325]
	Review: untitled [pp. 325-327]
	Review: untitled [pp. 327-329]
	Review: untitled [pp. 329-330]
	Review: untitled [pp. 330-331]
	Review: untitled [pp. 331-332]
	Review: untitled [p. 332]