THE INTEGRITY PAPERS Genre Group ceptualinstitute.com
In June, 1997, I had the pleasure of meeting
Michel Cabanac at the Assoc. for the Scientific Study of Consciousness conference at Pomona CA. I reprint for you here his 1991 paper (with minor graph and text omissions due to technical email problems). I find his work to be of particular importance in regard to Integrity concepts because his research clearly exposes the impact of condition changes which absolutely alter gödel information-sets involved in evaluative decision making ... how a given system weighs plural inputs in order to arrive at appropriate subsequent behaviors. Cabanac discusses the effect in terms of pleasure. He has experimentally proven that there are no absolutes in system health and maintenance, and that we (as representatives of the class of how living systems operate) constantly monitor the best-states and most comfortable states of several aspects of anatomy and experience at the same time. What one sub-system (such as taste) "normally" or "conditionally" regards as pleasurable or good, can, in the holistic schema of body-state, be made to be "not-good", if for example the body discomfortingly injests too much sugar-water, and puts strain and pressure on the digestive system or soft tissue in general. Cabanac's work is extremely significant.
PLEASURE: THE COMMON CURRENCY
Michel CABANAC
UNIVERSITÉ LAVAL, Department of Physiology QUÉBEC, CANADA, G1K 7P4
email: michel.cabanac@phs.ulaval.ca
ABSTRACT
At present as physiologists studying various homeostatic behaviors, such as thermoregulatory behavior and food and fluid intake, we have no common currency that allows us to equate the strength of the motivational drive that accompanies each regulatory need, in terms of how an animal or a person will choose to satisfy his needs when there is a conflict between two or more of them. Yet the behaving organism must rank his priorities and needs a common currency to achieve the ranking (McFarland & Sibly, 1975). A theory is proposed here according to which pleasure is this common currency. The perception of pleasure, as measured operationally and quantitatively by choice behavior (in the case of animals) , or by the rating of the intensity of pleasure or displeasure (in the case of humans) can serve as such a common cvarious motivations would thus be accomplished by simple maximization of pleasure. In what follows, the scientific work arising recently on this subject, will be reviewed briefly and our recent experimental findings will be presented. This will serve as the support for the theoretical position formulated in this essay.
INTRODUCTION
One shortcoming of the theories of the optimization of behavior proposed by ethologists and behavioral ecologists is that the mechanism, by which behavior is optimized, is never mentioned. In other words, they do not explain how the subject "decides". The main reason for this shortcoming is that there exists some uncertainty whether their predictive models concerned with proximate causation of behavior are functional or mechanistic (Fantino & Abarca, 1985, 1987; Wynne, 1987). Most often they work at a level of ultimate causation, emphasizing the evolution of traits that contribute to lifetime reproductive success. This is reasonable when the object studied is animal behavior. With humans, however, it is possible to obtain verbal reports and to analyze in cognitive terms the proximate physiological and psychological mechanism and, thus, to go further than the mere measurement of behavior, especially when the behavior studied is serving a physiological regulation. First, it is possible to identify, in the subject's body or mind, the signals aroused which are taken into account to produce a behavior with consequences for the physiological state. Secondly, one may explore, quantitatively, the relationship between the motivation thus analyzed, the behavior displayed by the subject, and the environmental parameters. This general method has produced results showing experimentally the paramount importance of pleasure in the determination of behavior and has thus confirmed the intuitive conclusions of philosophers, based on self-introspection and common-sense (Cabanac 1971,1979). It should be pointed out that, in this case, pleasure was not inferred from the behavioral measurements; it was reported by the subject who experienced it. With this method pleasure is no longer a circular description of the causality of behavior, but instead is an object of study in its own right.
The thesis presented here is that the maximization of pleasure, and the minimization of displeasure, not only leads to useful behavior, but is also the answer to motivational conflicts. It is hypothesized that pleasure serves as the common currency in the tradeoffs between clashing motivations. The displeasure of frustrating one motivation being accepted for the sake of a larger pleasure obtained in satisfying another one. This thesis will be presented step by step starting with the role of pleasure in situations involving only one motivation for a behavior serving only one physiological aim. The second step will be devoted to situations where two motivations clash for behavior that serves physiological aims. In the last step motivations not directly serving physiological aims will be introduced in the experimental conflicts of motivations; thus allowing generalization of the pleasure theory according to which unpleasant and even noxious behaviors may be accomplished because they are traded off for pleasurable rewards.
DEVELOPMENT
Sensory pleasure, one sensory modality at a time
In the commerce of a subject with stimuli, it has been shown experimentally that the wisdom of the body leads the organism to seek pleasure and avoid displeasure, and thus achieve behaviors which are beneficial to the subject's physiology (Cabanac, 1971). Relations exist between pleasure and usefulness and between displeasure and harm or danger. For example, when subjects are invited to report verbally, the pleasure aroused by a skin thermal stimulus can be predicted knowing deep body temperature (Cabanac et al., 1972; Attia, 1984). A hypothermic subject will report pleasure when stimulated with moderate heat, and displeasure with cold. The opposite takes place in a hyperthermic subject. Pleasure is actually observable only in transient states, when the stimulus helps the subject to return to normothermia. As soon as the subject returns to normothermia, all stimuli lose their strong pleasure component and tend to become indifferent. Sensory pleasure and displeasure thus appear especially suited to being a good guide for thermoregulatory behavior.
The case of pleasure aroused by eating shows an identical pattern. A given alimentary flavor is described as pleasant during hunger and becomes unpleasant or indifferent during satiety. Measurement of human ingestive behavior confirms the above relationship of behavior with pleasure; it has been repeatedly demonstrated in the case of food intake (Fantino, 1984), that human subjects tend to consume foods that they report to be pleasant and to avoid those foods that they report to be unpleasant. Pleasure also shows a quantitative influence: the amount of pleasurable food eaten is a function of alimentary restrictions and increases after dieting. The result is that pleasure scales can be used to judge the acceptability of food.
Thus, in the cases of temperature and taste, the affective dimension of sensation depends directly on the biological usefulness of the stimulus to the subject. This was already noticed by Aristotle (quoted by Pfaffmann, 1982). The word "alliesthesia" was coined to describe the fact that the affective dimension of sensation is contingent, and to underline the importance of this contingency in relation to behavior (Cabanac, 1971): A given stimulus will arouse either pleasure or displeasure according to the internal state of the stimulated subject. The seeking of pleasure and the avoidance of displeasure lead to behaviors with useful homeostatic consequences. Garcia has shown how past history, such as illness induced in association with the taste of an ingested substance, can "stamp in" a change in the affective quality of that taste (Garcia et al., 1985). The behavior of subjects instructed to seek their most pleasurable skin temperature could be described and predicted from their body temperatures and the equations describing their behavior were practically the same as those describing autonomic responses such as shivering and sweating (Cabanac et al. 1972; Bleichert et al. 1973; Marks & Gonzalez, 1974; Attia & Engel, 1981).
It is possible therefore from verbal reports to dissociate pleasure from behavior and to show thus that the seeking of sensory pleasure and the avoidance of sensory displeasure lead to behaviors with beneficial homeostatic consequences. Pleasure therefore indicates a useful stimulus and simultaneously motivates the subject to approach the stimulus. Pleasure serves both to reward behavior and to provide the motivation for eliciting behavior that optimizes physiological processes. One great advantage of this mechanism is that it does not take rationality 1 nor a high level of cognition to produce a behavior adapted to biological goals. Indeed, conditionned food aversion can be induced during sleep and under anesthesia (Garcia, 1990). As soon as a stimulus is discriminated, the affective dimension of the sensation aroused tells the subject, animal or human alike, that the stimulus should be sought or avoided.
Maximization of bi-dimensional pleasure: The hypothesis
In daily life a stimulus rarely comes alone; a motivation is seldom solitary. On the contrary, a subject must permanently rank simultaneous motivations. One basic postulate of Ethology is that behavior tends to satisfy the most urgent need of the behaving subject (Tinbergen, 1950; Baerends, 1956). Again here, nothing is said on how a subject ranks conflicting needs and motivations in a given order. Since pleasure led to optimization of behavior in the experiments arranged in such a way as to present only one motivation at a time, it is hypothesized here that maximization of pleasure might lead to the solution of conflicts of motivations, also.
It has been shown in the case of gustatory sensation that affective processes do sum algebraically in rats (Young & Christensen, 1962). One may expect that algebraic summation of pleasure and displeasure occurs also, not only within one sensory modality such as taste, but across different modalities of perception and experience. The resultant behavior would thus represent the tendency to maximize the algebraic sum of pleasures and displeasures. Such a process where two yoked behaviors increase total pleasure would be analogous to the chemical process of an improbable endothermic reaction that can reach completion because it is yoked to an exothermic one.
Figure 1 gives a theoretical example of such a process where a behavior unlikely to occur spontaneously because it generates displeasure, will be achieved nevertheless because it is yoked to another behavior that generates pleasure:
If behavior l modifies the affective experience of the behaving subject from state a to state A, with A being larger than a,
then, behavior l tending to increase the subject's pleasure can be produced alone.
If behavior 2 modifies the affective experience of the behaving subject from state B to state b, with B being larger than b,
then behavior 2 tending to diminish the subject's pleasure will not be produced alone.
However, if the algebraic sum A + b is larger than the algebraic sum a + B,
then, behavior l and behavior 2 can be produced simultaneously since total pleasure is also increased.
To verify the hypothesis, it must be demonstrated that subjects placed in situations of conflict tend to maximize the algebraic sum of their pleasure. This was verified, first, in situations where pleasure and displeasure were aroused by relatively simple stimuli and the affective experience was sensory in origin.
Maximization of bi-dimensional sensory pleasure: The evidence, human behavior.
In a series of experiments subjects were placed in bi-dimensional sensory situations. The experiments consisted in comparing the ratings given by a subject for the pleasure aroused by the stimuli, with the actual behavior choice displayed by the same subject. The stimuli were presented in x,y matrices. x and y extended, in discrete steps, over a range that the subjects might encounter in natural life. On a first session, the subject was invited to rate his pleasure in one dimension, say x, of the matrix of stimuli. On a second session, the subject was invited to rate his pleasure in the second dimension, say y. In two additional sessions the subject was invited to select his preferred stimulus when he could manipulate one dimension of the matrix, then the other. For example, x was imposed in a session and the subject could manipulate y; then y was imposed in the last session and the subject could manipulate x.
This general method allowed the experimenter to obtain the subject's x,y map of pleasure/displeasure on the first two sessions, then to record the subject's preference behavior in the same bi-dimensional space on the last two sessions, and finally to compare the map with the behavior.
A first experiment combined basic taste stimuli (Cabanac & Ferber, 1987). Simple tastes were selected, for the stimuli are easy to produce and combine, and they are less prone to be associated with cultural factors. Not that preference induced by cultural factors would be less valid as a test of the hypothesis, but would add inter-subject variability. A combination of sweet gustatory stimuli (x) was thus pitted against sourness (y). In a matrix, 25 gustatory stimuli combined 5 suc concentrations (0.l5 to 2.35 mole/l) in small cups with 5 sournesses (pH 1.8 to 5.7). Four sessions took place on four different days at the same time of day. Throughout the first session, the 25 stimuli were tasted for 5 sec each and the subject was asked to give a magnitude estimate of his pleasure or displeasure in response to the x dimension of the matrix (e.g. sweetness) regardless of the y dimension. Similarly, throughout the second session, the same stimuli were tasted again and the subject rated his pleasure or displeasure in response to the y dimension of the matrix (e.g. sourness) regardless of the x dimension. For the third session the subject was invited to mix samples himself, this to allow a behavioral choice in one dimension of the matrix, the other being imposed (e.g. he could mix suc and water ad libitum for each of five imposed sournesses). For the last session, the subject could adjust the previously imposed dimension, the other previously adjustable dimension now being imposed (e.g. he could modify sourness for five imposed suc concentrations).
The results of the first two sessions showed that pleasure/displeasure resulted from a two-dimensional combination of the taste modalities offered. This confirms the algebraic summation of suc and sodium chloride palatability obtained in rats by Young and Christensen (1962). Similar bi-dimensional maps of pleasantness have been obtained with humans when combining various fat and sugar concentrations in taste solutions (Drewnowski & Greenwood, 1983; Drewnowski et al. 1987). In addition, the results of the last two sessions (Figure 2) showed that the subjects tended to maximize their sensory pleasure. The operant choices obtained coincided with the ratings obtained in the first two sessions. Whether a dimension of the matrix was imposed or operantly adjustable made no difference on the results; the subjects tended to maximize sensory pleasure in the imposed dimension as well as in the adjustable dimension. The subjects therefore maximized their pleasure in the bi-dimensional space offered when two dimensions were independent. This means that they tended to maximize the algebraic sum of pleasure aroused by the combination of stimuli, and that maximizing the pleasure of one sensory dimension sometimes entailed increasing the displeasure of another sensation.
This interesting result was verified in treadmill vs ambient temperature experiments where a clear-cut cost was involved, either fatigue or cold discomfort. The results of this experiment too showed that the subjects' behavior tended to place them in pleasurable areas of the offered space. In that experiment the subjects were again placed in a bi-dimensional (x, y) sensory space, but they had to make a tradeoff. The perception aroused by a thermal environment was pitted against that aroused by walking on a treadmill (Cabanac & LeBlanc, 1983). They could improve thermal comfort (x) at the cost of fatigue (y). Dressed in swim suits and tennis shoes, the subjects walked at 3 km.h-1 on a treadmill placed in a climatic chamber. In a first series of measurements, the treadmill was varied from 0 to 24 percent in slope, and this condition combined with the ambient temperature from 25(C to 5(C in a 25 node matrix. Both treadmill slope and ambient temperature were imposed on the subjects who gave separate ratings of the pleasure or displeasure evoked by ambient temperature and by exercise. Actual ratings of pleasure/displeasure of x and y were obtained from the nodes of the matrix, i. e. at ambient temperatures 5, l0, l5, 20 and 25(C, combining with 0, 6, l2, l8, and 24 percent slopes. The ratings of x and y were then algebraically summed. Figure 3 gives, as iso-hedonic lines interpolated between the nodes, the algebraic sum of the two ratings. The figure is a bi-dimensional map of pleasure/displeasure in the bi-dimensional sensory situation explored (exercise vs ambient temperature). In a second series of measurements one variable was imposed by the experimenter, either slope of the treadmill or ambient temperature, and the subject could manipulate the other variable. The results, indicated on figure 3 as dots, showed that the subjects adjusted reciprocally exercise intensity, and ambient temperature. When a steep slope was imposed, they selected a low ambient temperature, and when walking at level was imposed, they selected a lukewarm ambient temperature. When the subjects were allowed to adjust treadmill slope and various ambient temperature were imposed, they selected steep slopes at low ambient temperature and no slopes at high ambient temperature. Figure 3 superimposes the subjects' behavioral choice (dots) with the map of pleasure/displeasure (iso-clines). It is quite striking that the dots showing the finally selected experimental conditions in quasi-steady states at the end of one-hour sessions are located in the white areas indicative of bi-dimensional pleasure. The subject's operant behavior was therefore guided by the tendency to minimize displeasure (or maximize pleasure), in a bi-dimensional space.
Similar results were obtained in another experiment which consisted of recording the subjects' sensations aroused in the chest and the lower limbs by simply walking on a treadmill. The subjects' behavior paralleled their algebraic sum of bodily discomfort (Figure 4). In subjects walking on a treadmill, perceptions are aroused in the chest (variable x) and in the lower limbs (variable y) and an experiment was set to study how the perceptions combine (Cabanac, 1985). A treadmill was geared to five slopes (x) and five speeds (y) in combination, and subjects were invited to walk on it. At minute 16, the subjects gave estimates of displeasure for sensations aroused in their chest and another one for the displeasure aroused in their lower limbs. Figure 4 gives the results of these ratings as iso-hedonic lines. The lines were obtained for each sensation explored (top box: displeasure in the lower limbs, middle box: displeasure in the chest) and for the algebraic sum of the two ratings (bottom box). These estimates were then compared to the subjects' actual behavior obtained in other sessions. The subjects were given the task of climbing up 300 m on the treadmill at varying speeds and slopes. When speed was imposed they could adjust slope, and conversely. The dots on figure 4 show the actual behavior. The dots fall along the lines although they did not generate them since they were obtained on different sessions. The pattern of conditions selected behaviorally (dots) was strikingly adapted to that of the algebraic sum of ratings of displeasure in the chest, plus displeasure in the lower limbs (iso-clines).
Thus, in the three situations explored (sweet vs sour, temperature vs fatigue, and chest vs legs) the subjects' behavior were repeatedly coherent: in the bi-dimensional sensory situations imposed by the experimenters, the subjects described maps of bi-dimensional pleasure in sessions where their pleasure was explored, and tended to move to the areas of maximal pleasure in these maps, in sessions where their behavior was explored. Figure 3 illustrates the coincidence of behavior choice with ratings of pleasure in the case of conflict between thermal environment and temperature, and Figure 4 when conflicting variables were slope and speed of treadmill.
The condition as to whether subjects in conflict situations tend to maximize their sensory pleasure is therefore satisfied. In the three experiments the subjects tended to maximize the algebraic sum of their sensory pleasure or to minimize their displeasure as in the third experiment described above. As a corollary of this observation, it can be stated that, in a situation of conflict of motivations, one can predict the future choice of the subject from the algebraic sum of affective ratings of pleasure and displeasure, given by the subject, to the conflicting motivations.
This result is not surprising if one considers that, at each instant, all motivations are ranked to satisfy only the most urgent. There must exist a common currency to actuate the behavioral final common path (McFarland & Sibly, 1975) and the results of the above experiments show that sensory pleasure fulfilled the conditions required of a common motivational currency, at least in the case of the behaviors selected which have clear physiological implications. Pleasure permitted ranking thermal comfort above fatigue, or conversely. We shall return to this point after having examined the case of optimal behavior.
Optimal behavior
The word optimality applied to behavior can be ambiguous because it bears somewhat different meanings when used by Ethologists, Economists, or Physiologists (Lea et al., 1987). Ethologists differentiate between goal and cost. Economists differentiate between utility and cost. The goal of a subject, as well as utility, is some entity that an optimal behavior will tend to maximize, which may appear tautological to the physiologist. The cost is a characteristic of the environment that the optimal behavior will tend to minimize. All would agree that an optimal behavior gives the maximal net benefit (or fitness) to the behaving individual. Specialists diverge in their definition of benefit (or fitness). The benefit can be defined in terms of reproductive efficacity (Krebs & Davies, 1981) as well as financial profit and physiological function (McFarland, 1985). We are concerned here with this last aspect: physiological benefit.
To the Physiologist a behavior is optimal when it leads to homeostasis. Optimal behavior could be recognized easily when subjects instructed to seek their most pleasurable skin temperature selected stimuli which, after data analysis, could be described by mathematical models identical to the models describing the autonomic responses (Cabanac et al. 1972; Bleichert et al., 1973; Marks & Gonzalez, 1974). One may wonder whether optimization, as seen from the physiological point of view, was also achieved in the three experiments above where subjects maximized the algebraic sum of two modalities of sensory pleasure. As illustrated below, physiological criteria of optimization will show us that maximization of pleasure was the key to optimal behavior in the experimental conflicts of motivations studied.
The preferred taste is likely to be related to the chemical properties of sour and of sweet stimuli. It remains difficult to tell what was optimal behavior from the point of view of physiology in the case of the gustatory experiment above, where subjects maximized their bi-dimensional taste pleasure. However, it was possible to find such physiological criteria when exercise was involved.
Figure 5 compares the subject's behavioral choice (dots) at the end of a session with the theoretical time (lines) necessary to climb 300 m as directed: the dots were obtained with the subject manipulating speed or slope. It can be seen that the subject's choice paralleled the iso-chronic curves, that is, the subject tended to walk at a constant external power, which makes sense from the point of view of physiology, in the various combinations of slopes and speeds.
Figure 6 analyzes usefulness in the conflict temperature vs exercise as a plot against ambient temperature of the external work produced by a subject at the end of the one-hour sessions. The figure also combines sessions when the subject could manipulate ambient temperature while exercise was imposed, and those when he could manipulate the treadmill slope while ambient temperature was imposed. Obviously, heat production by walking on the treadmill, which can be estimated to be approximately 3 times the external work, was inversely proportional to ambient temperature. Behavioral heat production was thus proportional to the need to equilibrate heat loss and was therefore optimal, as seen from the point of view of temperature regulation. Taken separately from verbal reports of pleasure/displeasure, these results would not be different from animal observation (Krebs & Davies, 1981) and experiments (Collier & Rovee-Collier, 1981) which show a good adaptation of animal behavior to the physiological need. This has been repeatedly demonstrated and would not need further demonstration. However, I present here a different point of view, since the variations of the behavior displayed by the subjects could be compared with the variations of their sensory pleasure, as judged from their ratings obtained in separate sessions. Behavior and pleasure followed identical patterns.
Thus, the trend to maximize sensory pleasure found a physiological finality in physiological regulation both with the experiment slope vs speed, and with the experiment ambient temperature vs exercise. Pleasure coincided in both experiments with a clear-cut adaptive physiological aim. This, more than mere coincidence, suggests strongly that pleasure was the key to optimal behavior, and that maximizing pleasure leads to optimal physiological performance.
The common currency
The experiments reported above were limited to sensory pleasure and to conflicts of motivations with clear physiological implications. One may question whether it is possible to extend the conclusions to other domains than biology. The notion of behavioral common path is especially enlightening to answer that question. Paraphrasing Sherrington's image of the motoneuron final common path of all motor responses, McFarland & Sibly (1975) pointed out that behavior is also a final common path on which all motivations converge. This image incorporates all motivations into a unique category since behavior must satisfy not only physiological needs but also social, moral, :sthetical, playful motivations. Indeed, it is often the case that behaviors are mutually exclusive; one cannot work and sleep at the same time. Therefore, the brain, responsible for the behavioral response, must rank priorities and determine tradeoffs in the decisions concerned with allocating time among competing behaviors. It can be expected that the brain operates this ranking by using a common currency (McFarland & Sibly, 1975; McNamara & Houston, 1986). The metaphor of the common currency was used by McFarland and Sibly (1975) because:
"the necessity for comparing the merits of different courses of action [implies] that there must be some trade-off mechanism built into the motivational control system. Since the trade-off process must take into account all relevant motivational variables, it is clear that the mechanism responsible must be located at a point of convergence in the motivational organization" (McFarland & Sibly, 1975).
The results of the above experiments show that sensory pleasure fulfilled the conditions required of a common motivational currency, at least in the case of the behaviors selected which have clear physiological implications. Pleasure permitted ranking thermal comfort above fatigue, or conversely. If it can be accepted that, at each instant, a subject responds in the realm of physiology to the motivation that will provide the greatest additional sensory pleasure for any given cost, because there is a need for a common currency, we may conclude that pleasure is also the common currency for non-physiological motivations. In the experiments described below cold discomfort or the relief from pain, was pitted against money or teh pleasure of a videogame. In these conflicts the subjects used pleasure as the common currency in the respective trade-offs.
Money and playful behavior
Money was pitted against cold discomfort in a climatic chamber (Johnson & Cabanac, 1982), and against pain from isometric contraction in the thighs (Cabanac, 1986). In these experiments, human volunteers could earn money against duration of exposure to these unpleasant sensations. Thus, the longer they stood cold discomfort, or pain, the more money they earned. In several sessions, the rate of monetary reward was varied. In both cases it was found that the discomfort or the painful sensation, as described by the ratings given by the subjects, increased linearly as a function of time. In both cases it was found that the subjects tolerated higher discomfort, or pain, for a longer time when the monetary reward was higher. This finding is in conformity with common-sense expectation. In addition however, it experimentally measures and demonstrates that an unpleasant sensation was reliably and quantitatively matched against money, a non-physiological motivation. The relation between reward and duration of tolerated displeasure was logarithmic. It can be assumed that the subjects decided to end a session just at the instant when the displeasure of the sensation became greater than the pleasure of the monetary anticipated reward.
This result shows that other motivated behaviors may be comprehended by the hypothesis, and that decisions involving non-physiological motivations can be made by following the tendency to maximize pleasure. The last experiment to be discussed here pitted a hedonoid motivation, thermal discomfort, against a hormoid motivation, the pleasure of playing a video-game (Cabanac, 1989). The general principle was the same as above and consisted in exposing each subject to three sessions: in one session (play) the subject played a video-game for one hour, and every five minutes was requested to rate on a magnitude-estimation scale, the pleasure of playing the video-game. In the second session (cold) the ambient temperature was reduced progressively from 25(C to 7(C over one hour, and every five minutes the subject was requested to rate the pleasure or the displeasure aroused by ambient temperature. No anchor (landmark) was given for the ratings; the only instruction received by the subjects was to use positive figures for pleasure and negative figures for displeasure. In the third session (conflict) both video-game and cold ambient temperature were presented simultaneously and the only instruction received by the subjects was that they could interrupt the session whenever they wanted to. The pertinent result of the experiment was the duration of the subjects' stay in the climatic chamber in the conflict situation. Figure 7 gives two examples of results from the subjects with the shortest (left) and longest (right) durations in conflict sessions. This figure shows the ratings given to the pleasure of playing and to the displeasure of cold environment obtained in different sessions. This figure gives the time-course of positive ratings for the pleasure of playing and negative ratings for the displeasure and cold discomfort. The duration of conflict session of each subject is shown by the arrows. None of the monitored autonomic variables (heart rate, arterial blood pressure, body-core and various skin temperatures) correlated with the ratings of cold discomfort in such a way as to predict the duration of the conflict sessions. These variables may therefore be rejected as predictors of duration tolerated in the cold environment during the conflict session. The hypothesis under test states that at each instant a subject will tend to maximize the algebraic sum of his pleasures. To hold true the hypothesis under test would imply that the subject stayed in the climatic chamber during the conflict sessions as long as the pleasure of playing the video-game was stronger than the displeasure aroused by the decreasing ambient temperature. It can be expected that a subject would end the conflict session when the sum of the rating obtained in the cold session plus that obtained in the play session equalled zero. This can be verified first by the two examples of Figure 7. For the subject shown on the left the arrow shows that the actual duration of her conflict session was about six minutes longer than the time when both curves obtained in other sessions crossed each other. The subject shown on the right was the only one of this kind: the ratings given by her did not cross each other during the 60 min sessions of play and cold. This was confirmed by her long stay in conflict session; it can be seen that she stayed 74 minutes during the conflict session, i.e. about 9 min beyond the extrapolated time of crossing of the two curves. Thus, both subjects shown in Figure 7 stayed somewhat longer than the predicted time, but the basic hypothesis remains valid since the error on actual duration was + 6 min and + 9 min out of a span of theoretical duration extending from 5 min (left) to more than 60 min (right). It should be pointed out that the examples picked out for Figure 7 were not "the best" from the group of subjects in terms of matching theoretical and actual duration, but simply the shortest and longest durations in conflict sessions. Figure 8 shows the results of the group of 12 subjects. Actual duration is plotted against theoretical duration. Theoretical duration was that predicted through two different methods.
i) graphic: theoretical duration obtained from the simple time-course plot, as the time when positive rating was equal to negative rating as in Figure 7.
ii) algebraic: theoretical duration obtained from the family of two equations fitted to the time courses of both ratings when their sum was equal to zero.
Both methods gave similar results which are plotted on Figure 8 for the 12 subjects. The respective regression lines were: algebraic, y= 0.993 x + 5.29 with r= 0.977, and graphic, y= 1.093 x + 2.91 with r= 0.976. The coefficient of correlation as well as the slope was close to 1 whatever the method to obtain theoretical duration. From the ratings of pleasure and displeasure, it was therefore possible to predict the actual duration of the session in situation of conflict. The intersection of the regression line with y axis in Figure 8 was y = 5.3 min and would be 2.9 min with graphic theoretical duration. This indicates that the subjects did not stop the session at the time when both motivations, to stay in the climatic chamber and to leave it, became equal but only some minutes later. Such a delay, necessary for the subjects to make up their mind, would tend to confirm the hypothesis that the subjects made their decision from the algebraic sum of pleasure/displeasure aroused in both dimensions, for, at the time of equal magnitude for both motivations, there was no reason for a subject to remain in the climatic chamber, but there was none either to leave it. The few minutes of overstay can be understood as the time for cold discomfort to become stronger than the pleasure of playing.
This result validates the hypothesis tested. It is therefore permitted to conclude that pleasure was the common currency of the motivations here pitted against each other. This extends to playful behavior the conclusions made previously on behaviors with physiological purposes and on learned reward such as money. If maximization of the algebraic sum of pleasure as consequences of behaviors as different as thermoregulatory behavior and playing a game was the key to the solution of the conflict, it appears legitimate to extend it to other motivations as a general law for optimization of behavior and for solving conflicts.
DISCUSSION OF THE HYPOTHESIS
Pleasure and behavior
The relation of pleasure to behavior was regarded as obvious by the Greek philosophers Aristotle (284-322, B.C.) and Epicurus (241-170, B.C.,see Conche, 1977):
"Life and pleasure, as we can see now, are not separable; for without behavior there is no pleasure, and pleasure improves behavior." (Aristotle)
After the Greeks, many philosophers and thinkers such as St Augustine (354-430, see Lamarre, 1986), St Thomas Aquinas (1224-1274), Montaigne (1533-1592, ses Saulnier, 1988), Gassendi (1592-1655, see Bloch, 1971), or Sulzer (1751) recognized in the affective experience a great role, if not the essential role, as a motivation. Bentham (1742-1832, see Bowring,1838) based his "greatest happiness principle" on pleasure, "the spring of action". Kant (1788, see Picavet, 1983) and Mill (1863) were more concise but equally clear on this point. For Freud (1920) the "pleasure principle" determines the aim of life. An important analysis of the role of the affective process in behavior has been carried out by Duncker (1940-1):
"A search for the ultimate motives of human conduct cannot disregard pleasure which many eminent minds have considered to be the fundamental motive, or at least an important one. Others, to be sure, have held that pleasure is the outcome rather than the motive or goal of human striving... There cannot be the slightest doubt that many human strivings bear some kind of reference to pleasure, and likewise that many pleasures bear some reference to striving".
It is therefore generally assumed that a motivated behavior is oriented by the incentives received by the subjects (Killeen 1962; Nutin, 1975; Toates, 1986).
"From every point of view the affective processes must be regarded as motivational in nature". (Young, 1959, 1961).
Yet pleasure has never been popular in history for moral reasons. In addition the excesses of psychoanalysis have led to the rejection of all mentalistic explanations of behavior. As a result pleasure is shunned by most recent textbooks, a drawback of behaviorism in modern literature. According to the part they attribute to pleasure, Duncker (1940-1) has sorted philosophers into two schools: "Hedonists" for whom pleasure is the fundamental motive, as opposed to "Hormists" for whom pleasure is the outcome. "Hormism" was coined from McDougall's "hormic force" (1923) (a force that urges us to strive). One can easily recognize in hedonism and hormism the ancient opposition of the philosophers of the Garden, the Epicurists, and those of the Portico, the Stoicists, or with Pl_ (1982) the morals of pleasure as opposed to the morals of duty. This opposition has lasted through the centuries. It is interesting to notice that this dual way of looking at pleasure can be found as an internal fracture within most societies, families of thought, and churches. Eventually duty always wins its struggle against pleasure.
One good reason for the rejection of pleasure from the realm of science was the lack of experimental evidence and the fact that the philosophers conclusions were based on their own introspection only. This rejection may be based on earlier approaches to these phenomena but,
"as new techniques are developed, our ideas often have to be revised to encompass the new information obtained" (Teitelbaum, 1964).
The analysis of the causes for the dominant rejection of pleasure is out of place here. Interested readers will find it together with a historical review of epicurism in Pl_ (1982). Additional documentation on the relation of pleasure to behavior will be found in Toates (1986) and in Lea et al.(1987).
Animal behavior
"Although feelings of pleasantness and unpleasantness are known directly only in human experience, the facts of animal behavior make it necessary to postulate that affective processes have an objective existence". (Young, 1959).
There is evidence that animals, too, tend to maximize pleasure in multidimensional sensory situations, i. e. when a cost is involved. Wild animals make decisions that permit their survival. They are capable of coping with nutritional stress by substituting various responses (King & Murphy, 1985); the diversity and potency of such compensatory mechanisms have often been underrated. These choice situations have been studied experimentally in the laboratory: It is likely that the behavior displayed by pigs which preferred to feed than to thermoregulate (Ingram & Legge, 1970), steers which preferred to thermoregulate than to feed (Malechek & Smith, 1976), and sheep which preferred to drink than to feed (Squires & Wilson, 1971), are examples in farm conditions of animals which expressed a clear-cut preference of one stimulus over another one. It may be accepted that this preference optimized, from the point of view of cost, their behavioral response to stressful environments. Recently Stricker and Verbalis (1988) have shown how rats alternate drinking pure water and salted water in response to their need for water and for salt. Behavior thus is alternately serving both motivations. Because humans in an analogous situation maximized their pleasure (Figures 3 and 4), one may wish to find examples where animals adapted their behavior to physiological need by maximizing pleasure in a situation of conflict.
The past 20 years of research in operant conditioning have been dominated by the "matching law" which in its most general form states that the distribution of activities matches their relative "value". There have been numerous demonstrations of quantitative tradeoffs between different aspects of reward (e.g. frequency vs intensity Baldwin, 1968; Carlisle 1969, 1970; Collier & Rovee-Collier, 1981; Dawkins, 1990), and some showing that electric shock can be scaled in terms of "negative food units" (Farley & Fantino, 1978; Farley, 1980). There has been a heated debate regarding whether animals actually do maximize "value" (Green, Rachlin & Hanson, 1982; Rachlin, Battalio, Kagel & Green, 1981) or simply respond to various schedules of reinforcements in such a way as to match reinforcement ratios (Heyman & Herrnstein, 1986). The controversy seems to point to the matching law as a better description of animal behavior when only alimentary intake is taken into account. Yet, animal pleasure should not be reduced to the pleasure of feeding. Herrnstein (1977) pointed out that motor activity itself, such as hunting and prey capturing in predators, might be reinforcing; this hypothesis was also defended by Toates and Jensen (1991). To further explore pleasure in animals it is therefore necessary to find situations close to natural conditions where little constraint is involved, as opposed to situations where, say, animals maintained at 80% of their free-feeding weight must choose between starving and feeding plus receiving an electric shock.
Is it possible to obtain such empirical evidence showing that animals will seek sensory pleasure and succeed in trading off some amount of displeasure for it? In the obstruction method (Warden, 1931) in contrast to the operant conditioning method, the strength of a motivation is measured not as an operant or motor response, but rather as the decision made by an animal to overcome a resistance to obtain a reward. Such a situation can be explored in the laboratory under conditions close to nature. Rats were trained for several weeks to feed each day from 10 in the morning to 12 noon, i.e. 2 h/day. Over the same period, they learned in a zigzag maze that once a week, additional highly palatable food was available to them during the regular feeding session, but at 16 m from their shelter. On the day of an experimental session their shelter was kept warm but the maze and the bait were in a very cold environment of -15(C in turbulent air. Although regular laboratory chow was available ad libitum in their warm shelter, rats invariably ran for short meals to the cold feeder to obtain the highly palatable food and rushed back to their warm shelter between these meals (Cabanac & Johnson, 1983) (Fig. 9). With such foods, the animals took as much as half the nutrient intake of that day in an environment potentially lethal to them if they were to stay in it. For less palatable foods, the rats went only once or twice to the feeder over the two-hour session, and stayed there for a shorter time. Although they did not become hypothermic, the rats gave signs that the environment was unpleasant: They rubbed their ears and areas of naked skin, and their ear pinnae and tip of the tail sometimes were necd from frostbite. Thus, the animals faced the unpleasant or even painful cold not out of necessity, since regular food was provided at no cost in their warm shelter, but for the pleasure of ingesting a palatable bait. This also shows that there existed a quantitative matching of the pleasure of food with the displeasure of enduring cold. Such a result can be interpreted, in the light of the experiments on humans, by stating that the rats exercised their freedom to increase their algebraic sum: alimentary pleasure minus cold displeasure.
Maximization of pleasure may thus be the link between physiology and behavior and give the key to the problem of physiological optimization without the implication of the animal's knowledge and rationality about its physiological state. A working man does not have to know his body temperature, blood oxygen and glucose, muscle glycogen and lactic acid, etc to take an occasional but necessary break. He just has to "listen to" his sensations and maximize the algebraic sum of pleasures, a summation that is hardly conscious. By using the same mechanism animals also can make the right decisions. The seeking of pleasure is sufficient to actuate an appropriate behavior. This central place of pleasure in animal behavior is now being postulated when interpreting experimental results (Waldbillig & O'Callaghan, 1980; Baldwin, 1985; Mehiel & Bolles, 1988; B_dard & Weingarten 1989) and field observation (Hladik, 1977), or plainly recognized from theoretical considerations (Bindra, 1978; Lester, 1984; Toates, 1986).
Utility
"A general principle, subscribed to by both economists and ethologists, is that in the process of decision-making, something is maximized" (McFarland & Houston, 1981).
There is a concern for bridging the gaps between Economics and Biology (McFarland, 1977; Campbell, 1986; Lucas, 1986; Schwartz, 1986; Lea et al., 1987). McFarland (1977) has applied the concept of optimality to the mechanisms of decision making and brought together biological, psychological and economic approaches.
The notion of utility used in economics (Marshall, 1890; Samuelson & Scott, 1975) is similar to the notion of pleasure as used in this article.
"Those who know anything about the matter are aware that every writer, from Epicurus to Bentham, who maintained the theory of utility, meant by it, not something contradistinguished from pleasure, but pleasure itself, together with exemption from pain;..." (Mill, 1863).
Following Mill, I would suggest that pleasure, as measured in our experiments, corresponds to the Economists' concept of utility. The word utility is widely accepted by Economists; however this word is sometimes ill-understood by the laity, which has difficulty in accepting that utility may refer even to a detrimental or noxious object, such as a cigarette or a drug. On the other hand, the word pleasure is unequivocal, at least in the common language.
The concept of utility has been kept aloof by those who shun mentalistic explanations in much the same way as radical behaviorists have ignored pleasure. Utility has been measured experimentally (Mosteller & Nogee, 1951; Galanter, 1962) yet is seen as a circular explanation of whatever motivations economic behavior reveals to us (Lea et al., 1987).
"The utility concept purports to look behind the Veil of Money but utility cannot be measured, while money values can, and economists have a bias in favour of the measurable..." (Robinson, 1962).
If utility is pleasure, then the experiments reported in the previous pages opened the black-box circularity of the concept of utility, since it should be remembered that the subjects were asked to rate their own affective experience. These studies thus confirm the fundamental assumptions and experiments made in economics that individuals are utility maximizers. The iso-hedonic lines presented in figures 3 and 4 are homologous to the iso-utility functions of McFarland and Houston (1981). In these figures the zero iso-hedonic line is also homologous to the indifference curve predicted theoretically in microeconomics (Asimakopulos, 1978; Lea et al., 1987). The logarithmic tradeoff of thermal discomfort and pain, for money, displayed by the subjects parallels strongly the economic law describing the "diminishing marginal rate of substitution" (Asimakopulos, 1978; Lipsey et al, 1979; Lea et al., 1987).
Apparent contradictions
Complex situations with more than one stimulus at a time, seem frequently not to follow the rule:
pleasant = useful,
and are often taken into consideration to refute that rule. Five cases can be listed where the search for pleasure does not seem to result in useful behavior.
l) Food habits. People as well as animals tend to eat what they like and this may result in intake of the one preferred food to the exclusion of others and, in turn, in deficient diets (Galef, 1991). Young (1959) has shown in rats that food preferences are established in accordance with the animal's needs. The divergence from the law of usefulness is therefore only apparent and limited to the case, unlikely to occur in natural situations, where a familiar food is artificially made deficient in a given vitamin or aminoacid.
2) Artificial sweeteners are a similar case. The sweet sensation has been selected in phylogeny as pleasant because carbohydrates are both sweet and a source of energy. The fact that chemists have been able to uncouple sweetness artificially from energy content does not invalidate the natural relationship of pleasure, and sweetness, with usefulness.
3) A group of monkeys is first trained to respond in order to postpone the delivery of electric shock. Later the contingency is changed so that, rather than postponing shock, responding actually serves to trigger shock. Paradoxically, it is found that, rather than the response extinguishing, the animals accelerate their responding, an effect termed the 'McKearney' effect (McKearney, 1969). This situation resembles that of point 1) above. Pleasure is a signe of usefulness when a new behavioral response is being estyablished. The McKearney effect seems to be the result of a regression in an animal which cannot learn a new response, fot an unknown reason.
4) The case of drug addiction is more complex. The fact that addicts seek and consume neuroactive drugs able to actuate nervous paths leading eventually to pleasure, but without any obvious beneficial result for the subject, could be understood as a case of maladjustment of pleasure with usefulness. Yet it may be recognized that such a situation confirms that pleasure is a powerful drive of behavior, and therefore that only the relationship of pleasure with eventual usefulness is to be discussed.
One should distinguish the first trial with the drug from the later stage of established addiction. On first contact with the drug, curiosity rather than pleasure is likely to be the main motivation. Although curiosity in that case turns harmful for the subject, nobody would deny that curiosity has a formidable survival advantage and can be regarded as useful. In long addicted subjects, the withdrawal syndrome may be so noxious that the pleasure brought by taking the drug can be considered useful in the short term, by removing the physical discomfort. Finally, the consumption of drug is likely to be the outcome of several clashing motivations (point 5, below).
5) The fifth case is when useful is unpleasant.
"People often behave in ways that do not seem pleasurable as measured by other than behavioral measures. People go on diets, put themselves through unpleasant courses of study, read unpleasant books and articles, sacrifice their own pleasure for the good of relatives, friends, compatriots, etc." (Anonymous referee).
All the examples listed in this quotation enter into the general category of conflicting motivations that has been studied above.
Pleasure at the crossroads
"How can you compare the pleasure of cheese and beer with the pleasure of seeing a good Hamlet?" (Ladd, 1894).
The question raised at the beginning of this essay was to understand how subjects make their decisions. At the end of the essay one might ask whether anything is new from what was known to Epicurus, Gassendi, Bentham, and Mill. The experimental approach summarized in the above pages answers this question and provides two advantages. First, reproducibility and the sharing of evidence means that we can extend the discussion beyond the domain of morals to that of science. We can apply rigor to an area of knowledge that was previously based only on intuition and common-sense. Second, the experiments are at the crossroads of Biology, Psychology, and Economics and contribute to bring these sciences together.
CONCLUSIONS
As a general conclusion, it is proposed that animals and humans rank priorities in choice situations and thus optimize their behavior by the amount of pleasure aroused by this behavior. As an economist wrote:
"Pleasures and pains represent the sole genuine basis for understanding human motives" (Jevons, 1871).
Bentham's fourteen tables of the springs of action, and Duncker's four causes of pleasure can be compared to one another for the very good reason that present or expected pleasure is the common currency to past, present, and future actions. The "law of effect" would act, not as a learning process only, but also as a way for living organisms "to know" that they are producing an optimal behavior. At the same time pleasure renders unnecessary high levels of rationality in the process of decision making. Indeed it has been recognized that rationality plays only a modest part in the determination of behavior (Cooper, 1987). In situations of conflicting motivations, pleasure would serve as a common currency, with additive properties, for the ranking of priorities and the resulting tradeoffs. The great advantage of pleasure both as a motivation and as a key to optimization lies in its versatility. Pleasure renders unnecessary the multiplication of instinctual rigid stimulus-response programs and stimulus-bound reflexes (see Epstein, 1982) whose number would have to increase ad infinitum with the complexity of living organisms2. S-R programs could not provide the flexibility which characterizes most of the purposive, goal-directed behavior of more advanced animals. Pleasure opens an infinite register of new responses. Indeed it is far more simple to maximize pleasure than to accumulate within the CNS an infinite number of instinctual responses. Pleasure in this case is analogous to the multiplicative function programmed in a calculator. It is far more simple to have access to that function than to store the infinite number of possible multiplications of rational numbers.
The experiments described in this essay were deliberately simple, a motivation was pitted against another motivation and all other variables were kept constant, to allow clear conclusions. The results are clear enough to support the theorizing that optimization of behavior is reached by maximizing pleasure, and that pleasure is the common currency allowing tradeoffs between the two clashing motivations. This conclusion satisfies two fundamental postulates, thus making them laws: the first, in ethology, stating that animals rank their priorities so as to satisfy, at each instant, their most urgent motivation; the second, in economics, stating that humans tend to maximize utility. The spectrum of motivations explored so far, though not exhaustive, seems broad enough to extrapolate the theory to all motivations including expected and learnt pleasures. It remains to explore situations with more than two conflicting motivations. Experimental situations with more than two motivations at a time are very difficult to handle and to understand because of their complexity. Displacement reactions are likely to occur when subjects can maximize pleasure only by fleeing the situation. Apparent paradoxical behaviors, such as impulsive behavior (Ainslie, 1975), maladaptive behaviors (Wiepkema, 1985), absurd risk taking (Mosteller & Nogee, 1951), reversal of decisions (Tversky & Kahneman, 1981), and irrational (in the economist's acceptation of the word) behavior (Plott, 1986) might enter into the model if all the variables were under the observer's control, and the subjects' pleasure were measured rather than their behavior. Even drug addiction can be described as a behavior whereby subjects, in conflict situations, maximize pleasure or minimize displeasure, as much as they can or know. In that case addiction can even be considered useful if, thus, subjects suppress their anxieties and avoid other drastic suicidal conducts. In an enlightening article, supportive of the thesis defended here, Alexander (1990) sees drug addction not as a disease but rather as an adaptive behavior.
"Of course many satisfactions are not common pleasures, but belong to the development of a man's higher nature, or to use a good old word, to his beatification; and some may even partly result from self-abnegation" (Marshall, 1890)
The failure of the rational economic man to reflect human nature, as rightly denounced by Schwartz (1986), does not invalidate the hypothesis presented herein, but reminds us that
"man does not live by bread alone" (Deuteronomy 8:3).
Human liberty is often ill understood as the freedom to do everything. Actually it is to be understood as the freedom to choose one's own way to maximize pleasure. Among the motivations sorted by Sulzer (1751) as sensory, intellectual, and moral, the latter has always been considered by the philosophers as the most rewarding.
Acknowledgements: I wish to thank Philip Teitelbaum and Frederick Toates for critically reading, and improving, previous versions of this manuscript.
LEGEND FOR FIGURES
Figure 1: Mechanism by which a behavior (behavior 2) that produces displeasure can be chosen by a subject if another behavior (behavior l) that produces pleasure is simultaneously chosen. The necessary and sufficient condition for the behavior 2 to occur (action) is that the algebraic sum of affective experience (pleasure) of the yoked behaviors is positive (a+B<A+b). Capital letters B and A indicate larger pleasure than respective small letters a and b.
Figure 2: A subject tends to maximize pleasure in a bi-dimensional gustatory space. The four squares are identical combinations of sweetness (xs) and sourness (ys). A) reports the results of one session: the subject rated his pleasure, left box, from the sour component of sensation, and, right box, from the sweet component of sensation. The actual rating is not indicated but the zone of highest rating in each vector is indicated with a heavy black bar. These bars are parallel to the y axis in order to be compared with the subject's actual choice in five other sessions; the open circles show the actual choices obtained in these behavioral sessions. The subject could modify sourness but not sweetness which was fixed at a different level for each of these five sessions.
B) is a similar experiment as A) except that he subject could modify sweetness instead of sourness which was fixed at five different levels: the left box shows maximal pleasure of the sweet component of sensation and the right box shows the sour component. Maximal pleasure zones are indicated with heavy black bars. These bars are parallel to the x axis to be compared with the subject's actual choice in five other sessions, as shown with open circles.
This figure shows that in 8 cases out of ten the subject's behavioral choice (open circles) was on maximal pleasure zones (either in the imposed dimension, or in the dimension that the subject could manipulate, or in between). Thus the subject based his behavior on the pleasure aroused simultaneously in both dimensions (sourness and sweetness) (from Cabanac & Ferber, 1987).
Figure 3: Behavior tends to place a subject in pleasurable area of a motivational space. This figure superimposes results obtained in one subject who walked at a constant 3 km.h-1 on a treadmill over several sessions. The iso-clines were obtained in the first series of sessions as follows: A combination of treadmill slopes and ambient temperature (Ta) was imposed, and the subject rated separately his discomfort aroused by fatigue and by cold. The lines show the algebraic sum of ratings triggered by the various combinations of environmental temperature with intensity of exercise; positive ratings indicate a pleasurable sum, negative ratings indicates an unpleasant sum. The lines are therefore iso-hedonic lines.
The dots were obtained in the second series of sessions. They show the actual locations in the map sought by the subject in ten sessions. Either ambient temperature was imposed on the subject (5, l0, l5, 20, or 25(C) and he could select slope (H) i.e. intensity of exercise, or the slope of the treadmill was imposed (0, 6, 12, 18, or 24 percent slope) and the subject could choose environmental temperatures (J). For example, when Ta was imposed at 10(C, the subject requested to raise the treadmill up to 23.5 percent slope and the H is located in the white area (pleasure) beyond the + 0.5 pleasure incline. When the treadmill was imposed at 0 percent slope, the subject requested Ta to be raised to 26(C and the J is located in the white area (pleasure) (from Cabanac & LeBlanc, 1983).
Figure 4: Subjects sum algebraically their affective perceptions in chest and in lower limbs. Each of the three boxes superimposes the results obtained with one subject in two series of sessions. In the first series of sessions the subject walked in a combination of treadmill slopes and speeds and was requested to report his affective ratings that describe the displeasure aroused in his legs (top box), and in his chest (middle box). The algebraic sum of the two above (bottom box), was calculated by the experimenter. The subject's ratings are indicated as lines. These lines show the iso-hedonic thresholds for displeasure (-l) in the slope/speed space. The negative ratings indicate increasing displeasure with increasing figures. These lines are negative since the subject reported only displeasure and no pleasure. The dots show the operant choice made by the subject in the second series of sessions. The subject walked on the treadmill of which the slope (P) was imposed and he could adjust the speed, or the speed (J) of the treadmill was imposed and he could adjust the slope. The same dots are drawn in the three boxes to allow the comparing with the iso-hedonic lines. It can be seen that the behavior fitted best with the threshold of combined (sum) bi-dimensional displeasure (From Cabanac, 1985).
Figure 5: Example of spontaneous optimization of behavior. The eight dots show the performance in eight separate sessions of a subject asked to climb up 300 m in elevation on a treadmill. He could adjust either the slope of the treadmill while the speed was imposed, or vice versa. The solid lines indicate the theoretical iso-chronic profiles for durations to climb up 300 m. The durations (12, 15, 20, 30, 40, 60, and 120) in minutes are indicated above each iso-chronic line. It can be seen that the subject adjusted his environment in order to perform within an approximate constant 40 min time i.e. he chose a relatively constant rate of work output no matter what the condition of slope and speed, in the eight different sessions. The behavior was thus optimal as seen from a physiological point of view (from Cabanac, 1985).
Figure 6: Another example of optimization. Each dot shows the amount of external work produced by a walking subject at the end of a one hour session when he could adjust either ambient temperature while treadmill slope was imposed (H), or treadmill slope while ambient temperature was imposed (J) (constant speed). It can be seen that the amount of work the subject chose to produce, i.e. his heat production, was inversely proportional to the cold environment, thus optimal as seen from the point of view of temperature regulation. (From Cabanac and LeBlanc, 1983).
Figure 7: Matching cold discomfort and pleasure of a videogame. Results obtained on three separate sessions by the individual subjects who stayed the shortest (left) and longest (right) duration on the third session. The positive ratings of video-game (pleasure) were obtained in one session (F). The ratings related to the cold environment (G) started positive (pleasure) then turned negative (displeasure) after a few minutes; they were obtained in another session. On the conflict sessions, both video-game and cold environment were present and the arrows show the duration actually endured by each subject. (From Cabanac, 1989).
Figure 8: Theoretical behavior compared compared to actual behavior. Actual duration endured by the 12 subjects, considered the dependent variable, plotted against theoretical duration considered the independent variable. Theoretical duration was obtained through two different methods: graphic was the first crossing of the respective time courses of positive rating of the pleasure of playing, with negative rating of the displeasure of cold discomfort; algebraic was obtained in solving the family of 2 equations: y' = f (rating play) and y" = f (rating cold) with y' = y". Interrupted line gives the hypothetical case where actual duration would equal theoretical. Solid line gives the regression of the actual duration plotted against theoretical duration with algebraic method (y = 0.993 x + 5.3, r = 0.977). The regression of actual duration plotted against theoretical duration with graphic method is not represented. It can be seen that actual duration for the group of subjects was theoretical duration plus 5.3 min. (From Cabanac, 1989).
Figure 9: Rats trade cold discomfort for palatability. This figure shows the characteristics of rats' meals taken from a feeder (restaurant) at -15(C placed 16 m away from the rats' warm shelter. The rats had water and chow ad lib in their shelter and were offered either cafeteria diet, the most and the least favoured bait of each animal, or lab-chow at the feeder . Bars above columns indicate S.E.; x____x lines join columns when not significantly different. Ingested mass is the amount ingested at the feeder in addition to the chow ingested in the warm shelter (not indicated on the figure); Mean meal duration shows the time spent at the feeder in the biting-cold environment. Total duration at the feeder was obtained from direct recording. The columns on the left are higher than on the right; this shows that rats ventured repeatedly in the cold for high palatability- but not for low palatability-food (From Cabanac & Johnson, 1983).
REFERENCES
Ainslie, G. (1975). Specious Reward: A behavioral theory of impulsiveness and impulse control. Psychol. Bull. 82, 463-496.
Alexander, B. K. (1990). The empirical and theoretical bases for an adaptive model of addiction. J. Drug Issues 20, 37-65.
Asimakopulos, A. (1978). Microeconomics - Oxford University Press, Toronto.
Attia, M. (1984). Thermal pleasantness and temperature regulation in man. Neurosci. Biobehav.
Rev. 8, 335-343.
Attia, M. & Engel, P. (1981). Thermal alliesthesial response in man is independent of skin location
stimulated. Physiol. Behav. 27, 439-444.
Augustine, Saint. De trinitate. Fr. transl. J. M. Lamarre (1986) Magnard, Paris.
Baldwin, B. A. (1968). Behavioral thermoregulation in mice. Physiol. Behav.3, 40l-407.
Baldwin, B. A. (1985). Neural and hormonal mechanisms regulating food intake. Proc. Nutrition
Soc. 44, 303-311.
Baerends, G. P. (1956). Aufbau des tierischen Verhaltens. In: Handbuch der Zoologie,
W.Kukenthal & T. Krumback eds. DeGruyter Co., Berlin, Teil 10 Lfg 7.
Bedard, M. & Weingarten H. P. (1989). Post absorptive glucose decreases excitatory effects of
taste on ingestion. Am. J. Physiol. 256, R1142-R1147.
Bindra, D. (1978). How adaptative behavior is produced: a perceptual-motivational alternative to response-reinforcement. Behav. Brain Sci. 1, 41-91.
Bleichert, A., Behling, K., Scarperi, M. & Scarperi, S. (1973). Thermoregulatory behavior of man during rest and exercise. PflŸgers Arch. 338, 303-312.
Bloch, O. R. (1971). La philosophie de Gassendi. Martinus Nijhoff, La Haye.Bowring, J. (1838). The works of Jeremy Bentham published under the superintendance of his executor (1838-1843). Republished in 1962, Russell & Russell inc., New York.
Cabanac, M. (1971). Physiological role of pleasure. Science, 173, 1103-1107.
Cabanac, M. (1979). Sensory pleasure. Quart. Rev.Biol. 54, 1-29.
Cabanac, M. (1981). Physiological signals for thermal comfort. In: Bioengineering thermal
physiology and comfort. K. Cena & J. A. Clark, (Eds) Elsevier, Amsterdam, p. 181-192.
Cabanac, M. (1985). Optimisation du comportement par la minimisation du d_plaisir dans un
espace sensoriel deux dimensions. C. R. Seances Acad. Sci. (Paris) 301, III, 607-610.
Cabanac, M. (1986). Money versus pain: experimental study of a conflict in humans. J. Exp. Anal. Behav. 46, 37-44.
Cabanac, M. (1989). La maximisation du plaisir, reponse un conflit de motivations. C. R. Sances Acad. Sci. (Paris) 309, III, 397-402.
Cabanac, M. & Ferber, C. (1987). Pleasure and preference in a two-dimensional sensory space
Appetite 8, 15-28.
Cabanac, M. & Johnson, K. G. (1983). Analysis of a conflict between palatability and cold
exposure in rats. Physiol. Behav. 31, 249-253.
Cabanac, M. & LeBlanc, J. (1983). Physiological conflict in humans: fatigue vs cold discomfort.
Am. J. Physiol. 244, R621-R628.
Cabanac, M., Massonnet, B., & Belaiche, R. (1972). Preferred skin temperature as a function of
internal and mean skin temperature. J. Appl. Physiol. 33, 699-703.
Campbell, D.T. (1986). Rationality and utility from the standpoint of evolutionary biology.
J. Business 59, S355-S364.
Carlisle, H. J. (1969). Effect of fixed ratio thermal reinforcement on thermoregulatory behavior.
Physiol. Behav. 4, 23-29.
Carlisle, H. J. (1970). Intermittent heat as reinforcer for rats in the cold. Physiol.Behav. 5,861-866.
Collier, G. H. & Rovee-Collier, C. K. (1981). A comparative analysis of optimal foraging behavior:
laboratory stimulations. - In: Foraging Behavior. A. C. Kamil & T. D. Sargent (Eds.)
Conche, M. (1977). Epicure: lettres et maximes. - Editions de M_gare, Villers sur Mer.
Cooper, W. S. (1987). Decision theory as a branch of evolutionary theory: A biological derivation
of the Savage axioms. Psychol. Rev. 94, 395-411.
Dawkins, M. S. (1990). From an animal's point of view: Motivation, fitness, and animal welfare.
Behav. Brain Sci. 13, 1-9.
Drewnowski, A. & Greenwood, M. R. C. (1983). Cream and sugar: human preference for high-fat
foods. Physiol. Behav. 30, 629-633.
Drewnowski, A., Halmi, K. A., Pierce, M. A. Gibbs, J. & Smith, G. P. (1987). Taste and eating
disorders. Am. J. Clin. Nutr. 46, 442-450.
Duncker, K.(1940-1941). On pleasure, emotion, and striving. Philos.Phenomenol.Res.1, 391-430.
Epstein, A. N. (1982). Instinct and motivation as explanations for complex behavior ch. 2. In The
physiological mechanisms of motivation, D W. Pfaff (ed.) Springer Verlag, New York,
pp. 25-58.
Fantino, E. & Abarca, N. (1985). Choice, optimal foraging, and the delay-reduction hypothesis.
Behav. Brain Sci. 8, 315-330.
Fantino, E. & Abarca, N. (1987). Delay-reduction theory: Straddling the functional-mechanism
continuum. Behav. Brain Sci. 10, 317-318.
Fantino, M.(1984). Role of sensory input in control of food intake.J Auton.Nerv.Syst.10, 347-326.
Farley, J. (1990). Reinforcement and punishment effects in concurrent schedules: a test of two
models.J Exper. Anal. Behav. 33, 311-326.
Farley, J. & Fantino, E. (1978). The symmetrical law of effect and the matching relation in choice
behavior. J Exper. Anal. Behav. 29, 37-60.
Galanter, E. H. (1962). The direct measurement of utility and subjective probability. Am. J. Psychol.
75, 208-220.
Galef, B. J. Jr. (1991). A contrarian view of the wisdom of the body as it relates to dietary
self-selection. Psychol. Rev. 98, 218-223.
Garcia, J. (1990). Learning without memory. J. Cognitive Neurosci. 2, 287-305.
Garcia, J., Lasiter, P. S., Bermudez-Rattoni, F. & Deems, D. A. (1985). A general theory of
aversive learning. Ann. N. Y. Acad. Sci. 443, 8-21.
Green, L., Rachlin, H. & Hanson, J. (1983). Matching and maximizing with concurrent ratio-interval
schedules. J Exper. Anal. Behav. 40,.217-224.
Heyman, G. M. & Herrnstein, R. J. (1986). More on concurrent interval-ratio schedules: a
replication and review. J Exper. Anal. Behav. 46,.331-351.
Herrnstein, R.J.(1977). The evolution of behaviorism. Am. Psychol. 32,.593-603.
Hladik, C. M. (1977). A comparative study of the feeding strategies of two simpatric species of leaf
monkeys: Presbytis senex and Presbytis entellus. In: Primate Ecology, T. H.
Clutton-Brock (Ed.) Academic Press, New York, p. 323-353.
Ingram, D. L. & Legge, K. F. (1970). The thermoregulatory behavior of young pigs in a natural
environment. Physiol. Behav. 5, 981-987.
Jankelevitch, S. (1951) French translation of S. Freud's Beyond the pleasure principle (1920)
In, Essais de psychanalyse, Fr. Transl. S. Jankelevitch (1951),Payot, Paris.
Jevons, W. S. (1871). Theory of Political Economy, Mac Millan, London.
Johnson, K. G. & Cabanac, M. (1983). Human thermoregulatory behavior during a conflict
between cold discomfort and money. Physiol. Behav. 30, 145-150.
Kant, I. (1788). Kritik der praktischen Vernunft, Part I, Fr. transl. F. Picavet (1983), Presses
Universitaires de France, Paris.
Killeen, P. R. (1962). Incentive theory. In: Nebraska Symposium on Motivation, 1981. Response
Structure and Organization, D. J. Bernstein (Ed.). University of Nebraska Press, Lincoln,
p. 169-216.
King, J. R. & Murphy, M. E. (1985). Periods of nutritional stress in the annual cycles of
endotherms: Fact or fiction. Am. Zool. 25, 955-964.
Krebs, J. R. & Davies, N. B. (1981). An introduction to behavioral ecology. Blackwell Scientific
Publ., Oxford.
Ladd, G. T. (1894). Psychology, Des. and Explan (quoted by Titchener 1908).
Lea, S. E. G., Tarpy, R. M. & Webley,P. (1987). The individual in the economy. Cambridge
University Press, Cambridge.
Lester, N. P. (1984). The "feed drink" decision. Behaviour 89, 200-219.
Lipsey, R. G., Sparks, G. R. & Steiner, P. O. (1979). Economics. 3rd, ed. Harper and Row,
New York.
Lucas, R. E. Jr (1986). Adaptive behavior and economic theory. J. Business 59, S401-S426.
Malechek, J. C. & Smith, B. S. (1976). Behavior of range cows in response to winter weather.
J. Range Management 29, 9-12.
Marks, L. E. & Gonzalez, R. R. (1974). Skin temperature modifies the pleasantness of thermal
stimuli. Nature 244, 473-475.
Marshall, A. (1890) Principles of Economics. MacMillan, London.
McDougall, W. (1923). Outline of psychology. Scribner's, New York.
McFarland, D. J. (1977). Decision making in animals. Nature 269, 15-21.
McFarland, D. J. (1985). Animal behaviour. Oxford: Pitman.
McFarland, D. J. & Houston, A. (1981). Quantitative ethology, the state space approach.
Pitman, Boston.
McFarland, D. J. & Sibly, R. M. (1975). The behavioral final common path. Philos. Trans. Roy
Soc. (London) 270, Biol, 265-293.
McKearney, J. W. (1969). Fixed-interval schedules of electric shock presentation: extinction and
recovery of performance under different shock intensities and fixed interval durations. J
Exper. Anal. Behav. 12, 301-313.
McNamara, J. M. & Houston A. I. (1986). The common currency for behavioral decisions. Am.
Natural. 127, 358-378.
Medicus, G. (1987). Toward an Etho-Psychology: A phylogenetic tree of behavioral capabilities
proposed as a common basis for communication between current theories in psychology
and psychiatry. Ethol. Sociobiol. 8, 131S-150S.
Mehiel,R.&Bolles,R(1988).Hedonic shift learning based on calories.Bull.Psychon.Soc.26, 459-462.
Mill, J. S. (1863). Utilitarianism. In: Utilitarianism, Liberty, and Representative Government. New
York, Dutton & Co. 1951.
Montaigne, M. de. Essais, I. Edition conforme au texte de l'exemplaire de Bordeaux par Pierre
Villey. Sous la direction de V.L. Saulnier (1988), Presses Universitaires de France, Paris.
Mosteller, F. & Nogee, P. (1951). An experimental measurement of utility. - J. Politic. Econ. 59,
p. 371-404.
Nutin, J. (1975). La motivation, In: Traiete de psychologie experimentale V. (3 Edition), J. Piaget,
(Ed.) Presses Universitaires de France, Paris, p. 5-96.
Pflaffman, C. (1982). Taste: A model of incentive motivation. In: The Physiological Mechanisms of
Motivation. D. W. Pfaff. (ed.) Springer Verlag, New York, p. 61-97.
Pl_, A. (1982). Par devoir ou par plaisir. Editions du Cerf, Paris.
Plott, C. R. (1986). Rational choice in experimental markets. J. Business 59, S301-S327.
Rachlin, H., Battalio, R., Kagel, J. & Green, L. (1981). Maximization theory in behavioral
psychology. Behav. Brain Sci. 4, 371-388.
Robinson, J.. (1962). Economic Philosophy. C. A. Watts & Co. Ltd., London.
Samuelson, P. A. & Scott, A. (1975). Economics. 4th Can. Ed. McGraw Hill Ryerson, Toronto.
Schwartz, B.(1986). The battle for human nature. W.W. Norton & Co, New York.
Squires, V. R. & Wilson, A. D. (1971). Distance between food and water supply and its effect on
drinking frequency and food and water intake of Merino and Border Leicester sheep.
Austral. J. Agricult. 22, 283-290.
Stricker, E. M. & Verbalis, J. G. (1988). Hormones and behavior: the biology of thirst and sodium
appetite. Am. Scientist 76, 261-267.
Sulzer, M. (1751). Recherches sur l'origine des sentimens agreables et desagreables. Mm. Acad.
Roy. Sci. Belles Let. Berlin, 7, 57-100.
Teitelbaum, P. (1964). Appetite. Proc. Am. Philos. Soc. 108, 464-472.
Thomas Aquinas. (1224-1274) French translation of Summa theologiÏ, Paris: Cerf (1984-1986).
Tinbergen, N. (1950). The hierarchical organization of mechanisms underlying instinctive behavior.
Symp. Soc. Exper. Biol. 4, 305-312.
Titchener, E. B. (1908). Lectures on the elementary psychology of feeling and attention. Arno Press,
New York.
Toates, F. (1986). Motivational systems. Cambridge University Press, Cambridge.
Toates, F. & Jensen, P. (1991). Ethological and psychological models of motivation. Towards a
synthesis. In: From animals to animats. J. A. Meyer & S. Wilson eds. Cambridge
Massachussetts: MIT Press pp. 194-205.
Tversky, A. & Kahneman, D. (1981). The framing of decisions and the psychology of choice.
Science 211, 453-458.
Voilquin, J. (1965). French translation of Aristotle's Ethique de Nicomaque 10. Paris: Flammarion.
Waldbillig, R. J. & O'Callaghan, M. (1980). Hormones and hedonics, cholecystokinin and taste:
A possible behavioral mechanism of action. Physiol. Behav. 25, 25-30.
Warden, C. J. (1931). Animal motivation, Experimental studies on the Albino rat. Columbia
University Press, New York.
Wiepkema, P. R. (1985). Abnormal behaviours in farm animals: ethological implications.
Netherl. J. Zool. 35, 279-299.
Wynne, C. (1987). Mechanisms of optimal choice. Comment on Fantino & Abarca. Behav.
Brain Sci. 10, 316.
Young, P.T.(1959).Role of affective processes in learning and motivation.Psychol.Rev.66, 104-123.
Young, P. T. (1961). Motivation and emotion. New York: Wiley & sons.
Young, P. T. & Christensen, K. R. (1962). Algebraic summation of hedonic processes. J. Comp.
Physiol. Psychol. 55, 332-336.
Resulting
affective Action
experience
__________________________________________________________
Behavior 1 a ------> A yes
Behavior 2 B ------> b no
Behavior l
+ a + B-----> A + b yes
Behavior 2
__________________________________________________________
with a < A, and B > b
and with a + B < A + b
Figure l
1 Rational is understood, here and in the following, in its philosophical acceptation (i.e. reason), and not in its narrower economical sense.
2 At this stage it is not possible to answer the question of the phylogenetic origin of pleasure and of the phylogenetic correlation of pleasure with behavior. Medicus (1987) would probably place pleasure in his Column III.