Cooperative Problem Solving in Rats
Content
COOPERATIVE PROBLEM SOLVING IN RATS1 WILLIAM J. DANIEL The Psychological Laboratory of the University of North Carolina Received May 20,1942 INTRODUCTION
Several experiments (2-5) have been presented which have more or less successfully demonstrated cooperative behavior in the higher apes. A few experiments (1, 6, 7), observational in character, have indicated this behavior in children. Only one experiment, that of Wolfle and Wolfle (9) has attempted to study cooperative behavior genetically by comparing the behavior of apes and children in nearly identical experimental situations. It has generally been believed that a study of cooperative behavior in animals as far down the evolutionary scale as the rat is rather fruitless. Only one such experiment (8) has come to the writer's attention. One of the three experiments which constitutes that monograph was designed to test for cooperative behavior in the rat. This experiment was negative; and aside from the films of Mowrer, no other attempt to obtain cooperative behavior in rats has been reported. The experiment reported here represents an apparently successful attempt at obtaining cooperation and one which relies primarily on quantitative data. PROBLEM
In the experiment described below we wanted to know if it is possible to arrange an experimental situation in such a manner that two animals can assist one another in obtaining food and at the same time escape electric shock. The experimental situation consisted of a grid box with an electrically insulated platform at one end which, when a rat stepped on it, would remove the charge from the grid. There was also a food crock flush with the grid and beyond the reach of a rat on the platform. This situation is represented schematically in figure 1. Our problem is concerned with the behavior of two rats in this situation. Will one rat go to the platform and remain on it, thus enabling the other rat to feed? Will the feeding rat leave the food crock and go to the platform, enabling the rat to leave the platform and feed? finally, will they exchange positions in such a manner that both are adequately fed and both escape or minimize shock? In short, will cooperative behavior be obtained when two rats are put into a double motive situation if the satisfaction of both of these motives is contingent upon the behavior of both animals? 1 A report of part of this work was presented at the Chicago meetings of the American Psychological Association in Evanston, 111., Sept. 1941. This study represents the substance of a thesis submitted to the faculty of the University of North Carolina in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the department of Psychology and was done under the direction of Dr. A. G. Bayroff of that department. 361
362
WILLIAM J. DANIEL ANIMALS
Heterozygous albino rats were used, ten males and two females ranging from 90 to 107 days of age at the start of the experiment. APPARATUS
The experimental situation {fig. 1)
The experimental situation consisted of a paraffined wood cage 22|* long x 12" wide x 4\" high with a grid floor and a glass top. In the center of the cage a 1
Pivot 4
Grid floor 1/8" brass rods $» apart
'
Wall atop
4 2"
Platform
y \ «
•*
12"
*.
Fig* 1. Diagram of apparatus* F s Food crock Scale - i Inch « 1 Inch
food crock, flush with the grid, was placed 8" from the edge of the platform thus making it impossible for a rat to feed from the platform. This grid cage was mounted on a set of stilts thus facilitating the replenishing and replacing of the food-crock. A small wooden wall stop was mounted over the platform at the end of the cage forcing all rats to remain beyond its center of gravity and making it impossible for a rat to administer shock without leaving the platform. Thetiltingof the platform also completed a light circuit so that a 40 watt bulb Sashed whenever a rat received a shock. This facilitated an objective counting of the number of shocks administered by each rat.
COOPERATIVE PROBLEM SOLVING IK BATS
363
The rats were dropped onto the grid through a small glass door on the top of the apparatus. Directly beneath this door a small entrance alley 6" long and 4 | " wide served to orient the animals in the proper direction, that is, facing directly towards the food-crock and platform. The shocking circuit
The grid was wired in series with a high resistance shocking circuit and the platform automatically shorted out the grid when a rat stepped on it. The essential problem here was to apply an electrical stimulus to the rats, the physical constancy of which we could be reasonably assured. Our circuit was of such a high external resistance that the added resistance of one or two rats gave the same meter reading as when a copper wire was placed across the grid. The transformer of this shocking circuit applied 3,750 volts to the rectifier tube and the current at the shock grid terminals could be varied from 100 microamps to 5 milliamps. The average shock intensity of 250 microamps required a circuit resistance of 3,400,000 ohms. PROCEDURE
The preliminary training
The aims of the preliminary training were three-fold: 1. To train the rats to feed in the experimental situation. 2. To train the rats to go to the platform when the grid was electrified. 3. To develop this discrimination, basic to the solution of the problem to be presented in the social situation, to the point at which the rats immediately made the response appropriate to the situation when the situations were varied in an irregular order. The following schedule was maintained: 1. The rats were unfed for 24 hours. 2. One rat was placed in the grid cage alone with the shock off and the foodcrock in place. It remained there for two 450-second trials, and was weighed before and after each day's trials. This procedure was continued through the 8th day for each rat. 3. At the end of the 8th day's run the sated rat was put in the grid cage with the grid electrified at 100 microamps. It was not removed until it had reached the platform and remained on it for 30 seconds. This procedure was repeated for 20 trials on this day. 4. From the 9th through the 13th day the rat was run for 20 irregularly mixed trials with electrified and non-electrified grid. On the shock trials the rat had to learn to go to the platform and to remain on it for 30 seconds. The time it took the rat to make the appropriate response, i.e., going to the platform, was recorded. On the food trials the grid was not electrified and the rat was left in the apparatus for 100 seconds. If by the end of this time it had not made the appropriate response (feeding) it was removed from the apparatus and given the next trial. The time it took the rat to. commence feeding was recorded and if
364
WILLIAM J. DANIEL
the rat did feed it was allowed to do so for 30 seconds so long as it started sometime within this 100 second interval. By the end of the 13th day of the preliminary training the rats had mastered this discrimination. When dropped on a cold grid the rat immediately went to the food-crock and fed; when dropped on a hot grid the rat immediately went to the platform and remained on it for 30 seconds. It made this discrimination in less than a second or before the experimenter could get to his stop watch to start timing the rat. Thus at the conclusion of the preliminary training each animal had learned to escape from shock or to feed in the apparatus depending upon the situation and it had learned this individually and in isolation. The experimental trials
At the end of the preliminary training the rats were divided into pairs of as nearly equal weight as possible. In the experimental trials two rats were put into the cage with the grid electrified and the food-crock in place. They remained in the experimental cage for one trial of 120 seconds duration. They were run 12 trials a day, a total of 1440 seconds, which, on the basis of preliminary experimentation, was adequate for the hunger satiation of both animals. The trial was timed by an electric stopclock and the individual feeding times by a manually operated stop-watch. Throughout the experimental trials the apparatus operated automatically. With one or both rats on the platform the shock was off. With the grid not charged, a rat could feed at the food-crock. Thus at least one rat had always to be on the platform if the other was to get to the food-crock. Occasionally both rats would leave the platform and attempt to feed and take shock simultaneously. If this behavior persisted for 5 consecutive times the shock was increased 50 microamps, and this "double feeding" stopped. The rats were fed pulverized purina dog chow mixed with water in the ratio of 5:6 respectively. At no time did the rats receive food other than that obtained in the experimental situation. This procedure was continued for 40 days at which time it appeared that the rats were doing as well as they ever would. RESULTS
The most significant fact in the data is that the rats exchanged positions from food-crock to platform and from platform to food-crock. Many of these exchanges were accompanied by shock and many shocks were administered in between these exchanges. It will be remembered that whenever there was any shock both rats received it, but it was administered only by one rat (the platform rat) stepping off the platform and thus electrifying the entire grid. As the experiment progressed more and more of the position shifts were accomplished without shock. Also fewer and fewer shocks were administered which did not result in an alternation. This data for 6 pairs of rats are presented in table 1. Since we are most interested in the final stage of this behavior the data are given
365
COOPERATIVE PBOBLEM SOLVING IN BATS
in terms of the mean performances for the last 5 days of the experiment as compared with the mean performances for the first 5 days of the experiment. Notice that with the exception of pairs 1-2 and 9-10 the critical ratios indicate that there is a marked and statistically significant decrease in the number of shocks not resulting in an alternation. The shock seems to have been quite effectively reduced. The nature of the alternations is also important. Let us call one of these rats A and the other B. Now if rat A is feeding at the food-crock he may return to the platform and then again return immediately to the food-crock. We shall call this exchange in position an "individual" alternation since it is accomplished only by one rat. When rat B exhibits this behavior we shall also call this exchange in position an individual alternation. When rat A is at the food-crock and returns to the platform and rat B comes off the platform and goes to the foodTABLE 1
1-2 3-4 6-6 7-8 9-10 11-12
MEAN SHUTS JOK LAST 5 DAYS
PEB CENT OF TOTAL SHUTS WITHOUT SHOCK
18 15 92 14 25 95
14 68 89 92 89 93
C.I
-1.33 4.64 8.92 6.04 1.29 8.89
TABLE 2 PAHS
Total number of alternations Percentage which the mutual alternations are of the total
1-2
3-4
5-6
7-8
9-10
11-12
947
863
3748
957
1201
3271
97
94
97
94
95
99
crock we refer to this kind of a shift as a "mutual" exchange in position. The question is, then, what percentage of the total exchanges in position is mutual and what percentage is individual? These data are presented in table 2. These data support the conclusion that for the entire group the rats alternate in a mutual manner in at least 94 per cent of the total alternations. We can say then that they are "taking turns" 94 to 99 per cent of the time. Our next question is, how well do these rats get fed while they are eliminating shock and exchanging positions in the experimental situation? Table 3 gives the mean weights for the first 5 days and the last 5 days of the experiment. It is clear from this table that every rat gained weight during the experiment and these gains ranged from 24 to 140 grams. This, along with the fact of their general healthy and vigorous appearance, further support the conclusion that rats were adequately fed throughout the experiment.
366
WILLIAM J. DANIEL
Another factor of importance is the extent to which the animals use the total available time in the apparatus. Means of this data for the entire experiment are given in table 4. Since they were run for 12 two-minute trials a day the total available feeding time for each pair of rats is 1440 seconds a day. Occasionally both rats would go to the platform and remain on it together. This is considered time wasted in as much as it is time during which food was available for one or the other animal but was taken by neither. We can see that the rats used practically all of the available feeding time. This speaks well for our final choice of time interval and also indicates that the rats were actively working on TABLE 3 SAT
1 2 3 4 5 6
7 8 9 10 11 12
ORAHS EATEN P E S DAV
21 21 25 . 20 25 27 27 23 30 21 32 31
WEIGHT BEFORE I H E EXPERIMENT
WKXGHT AFT£R THE EXPERIMENT
WEIGHT GADiZD
96 113 109 121 118 126 127 140 121 158 182 161
153 253 144 179 205 189 170 211 160 230 250 231
57 140 35 58 87 24 43 71 39 72 68 70
TABLE 4
the problem set by the experimental situation practically all of the time that they were in the apparatus. DISCUSSION AND INTERPRETATION
First of all let us re-emphasize the fact that this was a double motive situation. Our original intention was to arrange these motives in an experimental situation in such a manner that neither of them could be satisfied without the co-ordinated efforts of both animals. Rather than put the organism into an experimental situation and observe if it exhibits "cooperative behavior" we attempt to put the animal through a pro-
COOPERATIVE PROBLEM SOLVING IN RATS
367
cedure which will train it to be cooperative. Next we put it into a situation which is a cooperative one and quantify the extent to which it exhibits the behavior. In short, we train the behavior into rather than draw it out of, the animal. At the end of the preliminary training the rats have mastered a discrimination basic to the satisfactory solution of the problem presented in the test situation. First they shocked each other a great deal as they exchanged positions from foodcrock to platform. They received many shocks and they did not get as much food as they did later. But very soon they shifted more frequently and received more food. They shocked each other less and less in between the shifts in position. They accomplished a greater and greater number of shifts without getting any shock. In doing this they fed more and escaped more shock. They satisfied both of the motives in the situation. There are, finally, several observations that throw light on the nature of the behavior of the animals. As the experiment progressed the rats directed their behavior more and more towards each other rather than towards the food-crock or the platform. The rat on the platform would reach off holding the platform down with only one foot, and nudge the feeding animal. It would sometimes crawl up on the latter's back and paw it. This frequently resulted in the feeding rat's return to the platform. Sometimes it would hoist the feeding animal up on its shoulders. It might even bite and pull on the feeding rat's tail. These are overt responses directed towards the other animal. It would seem, then, that the platform rat's leaving the platform is conditional and dependent upon the movements of the other animal. The rat keeps at least one foot on the platform until the other animal is on the platform, or until the feeding rat goes by the platform rat towards the platform, or at any rate until the feeding animal has completed its return. In short, the rat has apparently learned that it or the other rat must be on the platform if it is to escape shock. The feeding rat, which has returned to the platform, remains there as the platform rat leaves. Then it exhibits the behavior typical of the platform rat. They "take turns" and thus they both get adequately fed and they both eliminate shock. In all the other animal experiments on cooperation both animals do the same thing; i.e., they pull ropes, punch stimulus cards, operate levers etc. In this experiment the cooperative aspect of the situation rests on the animals' doing distinctly different things; i.e., one feeds and the other turns off the shock. They synchronize their activity on two different tasks. The products of this solution are mutually shared. There is no simultaneous sharing of the goal achieved and thus there is no chance for competition. There is also an element of inhibition in the shape of delaying of responses since one animal waits on the platform while the other eats and they take turns doing this. Since the animals do respond in the experimental situation in a manner consistent with the preliminary training, since they do distinctly different things in the cooperation testing situation, and since some of their behavior has the characteristics of synchronization, restraint and differentiation of response, it would seem that their behavior may fairly be called cooperative.
368
WILLIAM J. DANIEL SUMMARY AND CONCLUSIONS
To investigate the development of cooperative behavior in rats 6 pairs of rats were put into a double motive problem situation (feeding and avoiding shock) requiring the co-ordinated efforts of both animals for its adequate solution. Each rat was individually trained to feed when the gridfloorwas not electrified, and when it was charged to go to a platform which shorted out the grid floor when a rat stepped on it. The rats were then paired, and the problem was to discover if cooperative behavior would be obtained when two rats were put into a double motive situation in which the satisfaction of both of these motives is contingent upon the behavior of both animals. One rat of a pair had to run to a platform which shorted out the electrified floor grid of a feeding box in order that a second rat might feed. From the data obtained in this situation we might draw the following conclusions: 1. The rats learned to exchange positions in this situation and at the same time allow sufficient feeding time for each rat to become adequately fed in the course of the experimental session. 2. They showed marked improvement in alternating without shock and in eliminating the shocks which did not result in an alternation. 3. They learned to take turns at the food-crock and platform so that by the end of the experiment they spent almost all of the available time in the apparatus working on the problem and very little time together on the platform. 4. And finally, in this situation, cooperative behavior has been apparently established. In a food-shock situation both animals exchange positions so that both are adequately fed. Furthermore, they exchange positions with sufficient care and speed that they avoid shock. They satisfy both conditions of the experiment in a situation in which the satisfaction of both conditions was contingent upon the behavior of both animals. REFERENCES (1) BERNE, E. VAN C.: An experimental investigation of social behavior patterns in young children. Univ. la. Stud. Child Welf., 1930,4,61 pp. (2) CRAWFORD, M. P., AND NISSEN, H. W.: Gestures used by chimpanzees in cooperative problem solving. (Silent film.) New York: Instructional Films, Inc., 30 Rockefeller Plaza, 1937. (3) CBAWFOBD, M. P.: Cooperative behavior in chimpanzee. Psychol. Bull., 1835,32,714. (4) CRAWFORD, M.'P.: Cooperative solution by chimpanzees of a problem requiring serial responses to color cues. Psychol. Bull., 1938,36,70S. (5) CRAWFORD, M. P.: Further study of cooperative behavior in chimpanzee. Psychol. Bull., 1936,33, 809. (6) LEWIN, K., AND LIPPITT, R.: An experimental approach to the study of autocracy and democracy: a preliminary note. Sociometry, 1938,1, 292-300. (7) MOORE, E. S.: The development of mental health in a group of young children: An analysis of factors in purposeful activity. Univ. la. Stud. Child Welf., 1931,4,128 pp. (8) WINSLOW, C. N.: A study of experimentally induced competitive behavior in the white rat. Comp. Psychol. Monogr., 1940,15, 35 pp. (9) WOLJXE, D. L., AND WoLPtE, H. M.: The development of cooperative behavior in monkeys and young children. J. Genet. Psychol., 1939,65,137-75.
Several experiments (2-5) have been presented which have more or less successfully demonstrated cooperative behavior in the higher apes. A few experiments (1, 6, 7), observational in character, have indicated this behavior in children. Only one experiment, that of Wolfle and Wolfle (9) has attempted to study cooperative behavior genetically by comparing the behavior of apes and children in nearly identical experimental situations. It has generally been believed that a study of cooperative behavior in animals as far down the evolutionary scale as the rat is rather fruitless. Only one such experiment (8) has come to the writer's attention. One of the three experiments which constitutes that monograph was designed to test for cooperative behavior in the rat. This experiment was negative; and aside from the films of Mowrer, no other attempt to obtain cooperative behavior in rats has been reported. The experiment reported here represents an apparently successful attempt at obtaining cooperation and one which relies primarily on quantitative data. PROBLEM
In the experiment described below we wanted to know if it is possible to arrange an experimental situation in such a manner that two animals can assist one another in obtaining food and at the same time escape electric shock. The experimental situation consisted of a grid box with an electrically insulated platform at one end which, when a rat stepped on it, would remove the charge from the grid. There was also a food crock flush with the grid and beyond the reach of a rat on the platform. This situation is represented schematically in figure 1. Our problem is concerned with the behavior of two rats in this situation. Will one rat go to the platform and remain on it, thus enabling the other rat to feed? Will the feeding rat leave the food crock and go to the platform, enabling the rat to leave the platform and feed? finally, will they exchange positions in such a manner that both are adequately fed and both escape or minimize shock? In short, will cooperative behavior be obtained when two rats are put into a double motive situation if the satisfaction of both of these motives is contingent upon the behavior of both animals? 1 A report of part of this work was presented at the Chicago meetings of the American Psychological Association in Evanston, 111., Sept. 1941. This study represents the substance of a thesis submitted to the faculty of the University of North Carolina in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the department of Psychology and was done under the direction of Dr. A. G. Bayroff of that department. 361
362
WILLIAM J. DANIEL ANIMALS
Heterozygous albino rats were used, ten males and two females ranging from 90 to 107 days of age at the start of the experiment. APPARATUS
The experimental situation {fig. 1)
The experimental situation consisted of a paraffined wood cage 22|* long x 12" wide x 4\" high with a grid floor and a glass top. In the center of the cage a 1
Pivot 4
Grid floor 1/8" brass rods $» apart
'
Wall atop
4 2"
Platform
y \ «
•*
12"
*.
Fig* 1. Diagram of apparatus* F s Food crock Scale - i Inch « 1 Inch
food crock, flush with the grid, was placed 8" from the edge of the platform thus making it impossible for a rat to feed from the platform. This grid cage was mounted on a set of stilts thus facilitating the replenishing and replacing of the food-crock. A small wooden wall stop was mounted over the platform at the end of the cage forcing all rats to remain beyond its center of gravity and making it impossible for a rat to administer shock without leaving the platform. Thetiltingof the platform also completed a light circuit so that a 40 watt bulb Sashed whenever a rat received a shock. This facilitated an objective counting of the number of shocks administered by each rat.
COOPERATIVE PROBLEM SOLVING IK BATS
363
The rats were dropped onto the grid through a small glass door on the top of the apparatus. Directly beneath this door a small entrance alley 6" long and 4 | " wide served to orient the animals in the proper direction, that is, facing directly towards the food-crock and platform. The shocking circuit
The grid was wired in series with a high resistance shocking circuit and the platform automatically shorted out the grid when a rat stepped on it. The essential problem here was to apply an electrical stimulus to the rats, the physical constancy of which we could be reasonably assured. Our circuit was of such a high external resistance that the added resistance of one or two rats gave the same meter reading as when a copper wire was placed across the grid. The transformer of this shocking circuit applied 3,750 volts to the rectifier tube and the current at the shock grid terminals could be varied from 100 microamps to 5 milliamps. The average shock intensity of 250 microamps required a circuit resistance of 3,400,000 ohms. PROCEDURE
The preliminary training
The aims of the preliminary training were three-fold: 1. To train the rats to feed in the experimental situation. 2. To train the rats to go to the platform when the grid was electrified. 3. To develop this discrimination, basic to the solution of the problem to be presented in the social situation, to the point at which the rats immediately made the response appropriate to the situation when the situations were varied in an irregular order. The following schedule was maintained: 1. The rats were unfed for 24 hours. 2. One rat was placed in the grid cage alone with the shock off and the foodcrock in place. It remained there for two 450-second trials, and was weighed before and after each day's trials. This procedure was continued through the 8th day for each rat. 3. At the end of the 8th day's run the sated rat was put in the grid cage with the grid electrified at 100 microamps. It was not removed until it had reached the platform and remained on it for 30 seconds. This procedure was repeated for 20 trials on this day. 4. From the 9th through the 13th day the rat was run for 20 irregularly mixed trials with electrified and non-electrified grid. On the shock trials the rat had to learn to go to the platform and to remain on it for 30 seconds. The time it took the rat to make the appropriate response, i.e., going to the platform, was recorded. On the food trials the grid was not electrified and the rat was left in the apparatus for 100 seconds. If by the end of this time it had not made the appropriate response (feeding) it was removed from the apparatus and given the next trial. The time it took the rat to. commence feeding was recorded and if
364
WILLIAM J. DANIEL
the rat did feed it was allowed to do so for 30 seconds so long as it started sometime within this 100 second interval. By the end of the 13th day of the preliminary training the rats had mastered this discrimination. When dropped on a cold grid the rat immediately went to the food-crock and fed; when dropped on a hot grid the rat immediately went to the platform and remained on it for 30 seconds. It made this discrimination in less than a second or before the experimenter could get to his stop watch to start timing the rat. Thus at the conclusion of the preliminary training each animal had learned to escape from shock or to feed in the apparatus depending upon the situation and it had learned this individually and in isolation. The experimental trials
At the end of the preliminary training the rats were divided into pairs of as nearly equal weight as possible. In the experimental trials two rats were put into the cage with the grid electrified and the food-crock in place. They remained in the experimental cage for one trial of 120 seconds duration. They were run 12 trials a day, a total of 1440 seconds, which, on the basis of preliminary experimentation, was adequate for the hunger satiation of both animals. The trial was timed by an electric stopclock and the individual feeding times by a manually operated stop-watch. Throughout the experimental trials the apparatus operated automatically. With one or both rats on the platform the shock was off. With the grid not charged, a rat could feed at the food-crock. Thus at least one rat had always to be on the platform if the other was to get to the food-crock. Occasionally both rats would leave the platform and attempt to feed and take shock simultaneously. If this behavior persisted for 5 consecutive times the shock was increased 50 microamps, and this "double feeding" stopped. The rats were fed pulverized purina dog chow mixed with water in the ratio of 5:6 respectively. At no time did the rats receive food other than that obtained in the experimental situation. This procedure was continued for 40 days at which time it appeared that the rats were doing as well as they ever would. RESULTS
The most significant fact in the data is that the rats exchanged positions from food-crock to platform and from platform to food-crock. Many of these exchanges were accompanied by shock and many shocks were administered in between these exchanges. It will be remembered that whenever there was any shock both rats received it, but it was administered only by one rat (the platform rat) stepping off the platform and thus electrifying the entire grid. As the experiment progressed more and more of the position shifts were accomplished without shock. Also fewer and fewer shocks were administered which did not result in an alternation. This data for 6 pairs of rats are presented in table 1. Since we are most interested in the final stage of this behavior the data are given
365
COOPERATIVE PBOBLEM SOLVING IN BATS
in terms of the mean performances for the last 5 days of the experiment as compared with the mean performances for the first 5 days of the experiment. Notice that with the exception of pairs 1-2 and 9-10 the critical ratios indicate that there is a marked and statistically significant decrease in the number of shocks not resulting in an alternation. The shock seems to have been quite effectively reduced. The nature of the alternations is also important. Let us call one of these rats A and the other B. Now if rat A is feeding at the food-crock he may return to the platform and then again return immediately to the food-crock. We shall call this exchange in position an "individual" alternation since it is accomplished only by one rat. When rat B exhibits this behavior we shall also call this exchange in position an individual alternation. When rat A is at the food-crock and returns to the platform and rat B comes off the platform and goes to the foodTABLE 1
1-2 3-4 6-6 7-8 9-10 11-12
MEAN SHUTS JOK LAST 5 DAYS
PEB CENT OF TOTAL SHUTS WITHOUT SHOCK
18 15 92 14 25 95
14 68 89 92 89 93
C.I
-1.33 4.64 8.92 6.04 1.29 8.89
TABLE 2 PAHS
Total number of alternations Percentage which the mutual alternations are of the total
1-2
3-4
5-6
7-8
9-10
11-12
947
863
3748
957
1201
3271
97
94
97
94
95
99
crock we refer to this kind of a shift as a "mutual" exchange in position. The question is, then, what percentage of the total exchanges in position is mutual and what percentage is individual? These data are presented in table 2. These data support the conclusion that for the entire group the rats alternate in a mutual manner in at least 94 per cent of the total alternations. We can say then that they are "taking turns" 94 to 99 per cent of the time. Our next question is, how well do these rats get fed while they are eliminating shock and exchanging positions in the experimental situation? Table 3 gives the mean weights for the first 5 days and the last 5 days of the experiment. It is clear from this table that every rat gained weight during the experiment and these gains ranged from 24 to 140 grams. This, along with the fact of their general healthy and vigorous appearance, further support the conclusion that rats were adequately fed throughout the experiment.
366
WILLIAM J. DANIEL
Another factor of importance is the extent to which the animals use the total available time in the apparatus. Means of this data for the entire experiment are given in table 4. Since they were run for 12 two-minute trials a day the total available feeding time for each pair of rats is 1440 seconds a day. Occasionally both rats would go to the platform and remain on it together. This is considered time wasted in as much as it is time during which food was available for one or the other animal but was taken by neither. We can see that the rats used practically all of the available feeding time. This speaks well for our final choice of time interval and also indicates that the rats were actively working on TABLE 3 SAT
1 2 3 4 5 6
7 8 9 10 11 12
ORAHS EATEN P E S DAV
21 21 25 . 20 25 27 27 23 30 21 32 31
WEIGHT BEFORE I H E EXPERIMENT
WKXGHT AFT£R THE EXPERIMENT
WEIGHT GADiZD
96 113 109 121 118 126 127 140 121 158 182 161
153 253 144 179 205 189 170 211 160 230 250 231
57 140 35 58 87 24 43 71 39 72 68 70
TABLE 4
the problem set by the experimental situation practically all of the time that they were in the apparatus. DISCUSSION AND INTERPRETATION
First of all let us re-emphasize the fact that this was a double motive situation. Our original intention was to arrange these motives in an experimental situation in such a manner that neither of them could be satisfied without the co-ordinated efforts of both animals. Rather than put the organism into an experimental situation and observe if it exhibits "cooperative behavior" we attempt to put the animal through a pro-
COOPERATIVE PROBLEM SOLVING IN RATS
367
cedure which will train it to be cooperative. Next we put it into a situation which is a cooperative one and quantify the extent to which it exhibits the behavior. In short, we train the behavior into rather than draw it out of, the animal. At the end of the preliminary training the rats have mastered a discrimination basic to the satisfactory solution of the problem presented in the test situation. First they shocked each other a great deal as they exchanged positions from foodcrock to platform. They received many shocks and they did not get as much food as they did later. But very soon they shifted more frequently and received more food. They shocked each other less and less in between the shifts in position. They accomplished a greater and greater number of shifts without getting any shock. In doing this they fed more and escaped more shock. They satisfied both of the motives in the situation. There are, finally, several observations that throw light on the nature of the behavior of the animals. As the experiment progressed the rats directed their behavior more and more towards each other rather than towards the food-crock or the platform. The rat on the platform would reach off holding the platform down with only one foot, and nudge the feeding animal. It would sometimes crawl up on the latter's back and paw it. This frequently resulted in the feeding rat's return to the platform. Sometimes it would hoist the feeding animal up on its shoulders. It might even bite and pull on the feeding rat's tail. These are overt responses directed towards the other animal. It would seem, then, that the platform rat's leaving the platform is conditional and dependent upon the movements of the other animal. The rat keeps at least one foot on the platform until the other animal is on the platform, or until the feeding rat goes by the platform rat towards the platform, or at any rate until the feeding animal has completed its return. In short, the rat has apparently learned that it or the other rat must be on the platform if it is to escape shock. The feeding rat, which has returned to the platform, remains there as the platform rat leaves. Then it exhibits the behavior typical of the platform rat. They "take turns" and thus they both get adequately fed and they both eliminate shock. In all the other animal experiments on cooperation both animals do the same thing; i.e., they pull ropes, punch stimulus cards, operate levers etc. In this experiment the cooperative aspect of the situation rests on the animals' doing distinctly different things; i.e., one feeds and the other turns off the shock. They synchronize their activity on two different tasks. The products of this solution are mutually shared. There is no simultaneous sharing of the goal achieved and thus there is no chance for competition. There is also an element of inhibition in the shape of delaying of responses since one animal waits on the platform while the other eats and they take turns doing this. Since the animals do respond in the experimental situation in a manner consistent with the preliminary training, since they do distinctly different things in the cooperation testing situation, and since some of their behavior has the characteristics of synchronization, restraint and differentiation of response, it would seem that their behavior may fairly be called cooperative.
368
WILLIAM J. DANIEL SUMMARY AND CONCLUSIONS
To investigate the development of cooperative behavior in rats 6 pairs of rats were put into a double motive problem situation (feeding and avoiding shock) requiring the co-ordinated efforts of both animals for its adequate solution. Each rat was individually trained to feed when the gridfloorwas not electrified, and when it was charged to go to a platform which shorted out the grid floor when a rat stepped on it. The rats were then paired, and the problem was to discover if cooperative behavior would be obtained when two rats were put into a double motive situation in which the satisfaction of both of these motives is contingent upon the behavior of both animals. One rat of a pair had to run to a platform which shorted out the electrified floor grid of a feeding box in order that a second rat might feed. From the data obtained in this situation we might draw the following conclusions: 1. The rats learned to exchange positions in this situation and at the same time allow sufficient feeding time for each rat to become adequately fed in the course of the experimental session. 2. They showed marked improvement in alternating without shock and in eliminating the shocks which did not result in an alternation. 3. They learned to take turns at the food-crock and platform so that by the end of the experiment they spent almost all of the available time in the apparatus working on the problem and very little time together on the platform. 4. And finally, in this situation, cooperative behavior has been apparently established. In a food-shock situation both animals exchange positions so that both are adequately fed. Furthermore, they exchange positions with sufficient care and speed that they avoid shock. They satisfy both conditions of the experiment in a situation in which the satisfaction of both conditions was contingent upon the behavior of both animals. REFERENCES (1) BERNE, E. VAN C.: An experimental investigation of social behavior patterns in young children. Univ. la. Stud. Child Welf., 1930,4,61 pp. (2) CRAWFORD, M. P., AND NISSEN, H. W.: Gestures used by chimpanzees in cooperative problem solving. (Silent film.) New York: Instructional Films, Inc., 30 Rockefeller Plaza, 1937. (3) CBAWFOBD, M. P.: Cooperative behavior in chimpanzee. Psychol. Bull., 1835,32,714. (4) CRAWFORD, M.'P.: Cooperative solution by chimpanzees of a problem requiring serial responses to color cues. Psychol. Bull., 1938,36,70S. (5) CRAWFORD, M. P.: Further study of cooperative behavior in chimpanzee. Psychol. Bull., 1936,33, 809. (6) LEWIN, K., AND LIPPITT, R.: An experimental approach to the study of autocracy and democracy: a preliminary note. Sociometry, 1938,1, 292-300. (7) MOORE, E. S.: The development of mental health in a group of young children: An analysis of factors in purposeful activity. Univ. la. Stud. Child Welf., 1931,4,128 pp. (8) WINSLOW, C. N.: A study of experimentally induced competitive behavior in the white rat. Comp. Psychol. Monogr., 1940,15, 35 pp. (9) WOLJXE, D. L., AND WoLPtE, H. M.: The development of cooperative behavior in monkeys and young children. J. Genet. Psychol., 1939,65,137-75.
