At the moment I’m writing a book on the integration of evolutionary psychology and behaviorism in personal relationships that is tenatively called “How to Train Your Boyfriend”. Just today a video came out where I discuss my thoughts on the intersection of behaviorism and evolutionary psychology for the comedy podcast, TRIGGERnometry. But, I’ve been thinking about how we’re intutive beahaviorists for a long time.
I wrote this paper for an evolutionary psychology seminar run by David Buss back in 2007 when I was working towards my PhD. Around that time David had published many papers and The Murderer Next Door, a book that outlined his thesis that murder is an adaptation. This was contrary to Martin Daly and Margot Wilson’s thesis that murder is a byproduct as outlined in their masterful book, Homicide. David’s papers both on homicide and on conflict between the sexes had a big influence on my thinking here.
A couple of months before I wrote this paper I had only ever heard Skinner strongly criticized by evolutionary psychologists and social psychologists, who made up much of the department at UT Austin. When I picked up Skinner’s Beyond Freedom and Dignity, I expected that I’d write a strong critique. It ended up being one of my favorite books. Reading it was so rewarding that it’s probably caused me to read more stuff I thought I would disagree with over the years. Meta. I assigned the first chapter, A Technology of Behavior (you can read it here), to over a thousand undergraduates in the UK over the years, with pretty polarized responses.
I don’t agree with everything in this essay anymore, of course I hardly knew anything about behaviorism at that time and had only read Skinner on behaviorism. By 2007 behaviorism had dealt with many of the critiques I lay out here, for example, equipotentiality. But this is a good insight into the origins of the book and other projects I’m working on now. I’ve edited it a bit and added commas, which I’m terrible at using now, and was even worse about using then. I’ve added a few links to clarify some concepts and save the reader a google.
Content warning: Child abuse, Jealousy, Domestic Abuse, Murder
The Evolutionary Psychology of Social Shaping
Behaviorism and evolutionary psychology may seem most antithetical to one another in terms of their ramifications for an understanding of human nature, and the premises that they begin with. Behaviorists implicitly posit a scant number of general purpose learning mechanisms that govern all behavior regardless of the adaptive landscape of the species, whereas evolutionary psychologists posit a multitude of domain specific learning programs with varying input flexibility depending on the adaptive landscape of the particular organism or species. While evolutionary psychologists and others have approached learning from the perspective of classical and operant conditioning, the behaviorist study of human behavior can be evolutionized further. Specifically this paper addresses the nature of primary reinforcement and punishment from an evolutionary perspective and the way learning schedules are established by individuals in social interactions.
Behaviorism had its heyday throughout most of the recent history of psychology. Later, radical behaviorism was extensively criticized because of various implicit assumptions it makes about equipotentiality, and its rebuking of species specific adaptations in response to environmental stimuli. Behaviorist principles are still incredibly useful when understanding behavior and learning. While it is patently obvious that some behaviors are reinforcing and punishing, most psychologists do not address this facet of interactions. For instance it is understood that a man who is jealously enraged by his mate’s sexual advances towards an intrasexual rival may verbally or physically abuse her, but this is rarely if ever directly labeled “punishment”: a consequence that can reduce the frequency of the behavior it follows. Explicitly using the terminology of the psychology of learning can help elucidate the framework of evolutionary psychology further as well as integrate our language more with the language of mainstream psychology. I will begin by discussing the nature of primary reinforcers in human beings and move on to the thesis that human beings have an evolved psychology to behaviorally condition one another.
Human species specific primary reinforcers and primary punishers
Animals have different primary reinforcers depending on the cues in their environment and which cues are most likely to be statistically associated with adaptive outcomes. Primary reinforcers are defined as unconditioned stimuli that are inherently reinforcing without classical or operant conditioning involved. Food, air, sex and sleep are examples of primary reinforcers, and these are well preserved among many organisms. From an evolutionary perspective, there must be many many more primary reinforcers than just the handful put forward by behaviorism. Because evolved psychological mechanisms solve adaptive problems with biologically prepared inputs, there are primary reinforcers that exist for the human species that do not exist for other species. We can also predict that there will be primary reinforcers associated with most psychological mechanisms that function to help motivate organisms the adaptive goals of those mechanisms. As a generalization, we can expect that any cue indicating that a relevant adaptive problem has been solved would be a primary reinforcer. One example in most contexts would be a smile. While behaviorists would argue that a smile is a secondary reinforcer, because it has been classically conditioned by being paired with food or another primary reinforcer, an evolutionary learning perspective would differ. Experimental evidence shows that infants that are 2-7 months old can be conditioned to prefer some sounds over others using smiles as reinforcement (Routh, 1969). For infants, the major adaptive problem they face is having a caretaker whose parental mechanisms are engaged towards him or her adequately for survival. Therefore a smile by the caretaker towards the infant is a cue that a relevant adaptive problem is at least partially being resolved by the behavior that elicits the smile.
Another prediction derived from evolutionary psychology is that primary reinforcers differ over the lifespan and that different primary reinforcers would have different salience depending on the adaptive problems that were recurrently faced by our ancestors during particular life stages. For example pictures of nude postpubescent females would be a strong primary reinforcer for males after puberty, but probably would not be a reinforcer for males that had not reached puberty. Researchers have found that male rhesus macaques will “pay” (give up part of a juice reward) to view the perineum region of female macaques or the faces of high status males but will not “pay” to view the faces of low status conspecifics (Deaner, Khera, Platt, 2005) (See Anderson, 1998 on the history of studies examining the operantly reinforcing properties of social stimuli in nonhuman primates). This study is interesting because it shows that social information in and of itself can act as reinforcement. Evolutionary psychology would predict that these same stimuli would not be reinforcing for female monkeys or monkeys who have not reached sexual maturity. Therefore primary reinforcers also do not display equipotentiality.
Primary reinforcers do not always feel pleasurable and may even have “negative” emotional affect associated with them. Many emotions may have negative emotional valence to them, making an individual feel anxious or upset, but these emotions may still cause motivation towards an act in a certain domain or perpetuate a behavior upsetting to the actor. While these emotions may feel negative to the individual they have evolved to help individuals solve certain adaptive problems. When a jealous person goes through their mate’s belongings and finds a phone number or other evidence of infidelity that is a primary reinforcer to continue investigative behavior even though the finding itself is unpleasant. In a fight a hard hit from one’s rival may be painful but can serve to motivate and perpetuate the desire to fight and increase proximity to the source of the pain. Both of these examples are exactly the opposite of what radical behaviorism would predict. All stimuli have context and domain specific effects on the organism depending on the adaptive problem they face.
Just as there are primary reinforcers there are also primary punishers. Shocks, loud noises, pain, and other such stimuli are almost always primary punishers, aversive stimuli that do not have to be paired with other aversive stimuli to reduce the frequency of behaviors that they follow or to cause aversion to stimuli they are paired with. Just as humans have their own suite of primary reinforcers they also have their own species specific primary punishers. Exclusion, embarrassment, reputational damage, shame, and the realization of inequity in a social exchange could all be examples of social primary punishers or consequences that reduce the prevalence of the action they follow (operant conditioning) or make the individual more averse to the stimuli they are paired with (classical conditioning).
Social primary reinforcers such as cues to sexual access, affection, grooming, smiling, friendly eye contact, cues of elevation in a status hierarchy and social primary punishments like those mentioned above are important for the thesis that humans have evolved to be shapers of the behavior of other humans. Evolutionary psychologists have investigated the evolved cognitive structure of humans in regard to adaptive preferences and aversions. This may be important to understand how we implicitly understand how other humans learn in their social interactions.
The extended phenotype
In “The Extended Phenotype”, Dawkins advances the hypothesis that those things that are proximal to an organism, their physiological characteristics, behavior patterns and individual differences are not the only things that make up their phenotype. Phenotype also consists of an organism’s effect on the environment and even its effects on other organisms. In a famous example, the “Bruce effect” is a pregnancy disruption that occurs when a female mouse smells an unfamiliar male mouse’s pheromones. In this case her pregnancy is immediately terminated and she becomes sexually receptive and fertile much more quickly than if she had carried her litter to term. Dawkins argues that one does not have to view the abortion and early receptivity of the female mouse as an adaptation on her part but can also view this as an adaptation on the part of the male mouse. Manipulation of the external environment, including the manipulation of other organisms and conspecifics is constantly under just as intense selection pressure as those phenotypic characteristics that enable the organism to adapt to its immediate environment. Some of the physical and social environment represents an organism’s phenotype and humans are an example par excellence of a species that shapes its social environment in myriad ways. More novel is the idea that social behavior can be looked at from a behaviorist perspective when one considers the extended phenotype that is an individual’s effect on the behavior of others.
Social interaction as shaping behavior
Social interaction takes a variety of forms. There is cooperation, competition and socializing but interaction has more beneficial utility than communicating vital information to conspecifics, winning mates and besting intrasexual rivals. From an evolutionary social learning perspective part of the reason we interact with others is to establish reinforcement and punishment schedules for their behavior and monitor their actions in terms of relevance to our adaptive goals. Some theorists have claimed that all communication is really an instance of one organism manipulating another (Dawkins & Krebs 1978). To take this further I would claim that consciously or unconsciously, reciprocal training is always occurring between interacting members of a social group. Social interaction and every use of language can be seen as a way to adaptively expand your phenotype into the behavior of another conspecific.
Above I mentioned that organisms all have primary reinforcers that feed into their evolved psychological mechanisms motivating them towards adaptive goals. Just as these primary reinforcers or primary punishers have intrinsically rewarding or punishing properties respectively so also do organisms interact with other organisms in the language of their primary reinforcers or primary punishers. Part of every organism’s “theory of mind” of other organisms is an innate understanding of the contingencies they can establish in order to manipulate the behavior of other organisms towards their adaptive goals.
The puzzle of inflicted costs
Evolutionary psychologists make frequent mention of individuals inflicting costs on one another without explicitly stating why individuals would do this. Although tactics of manipulation have been studied in romantic relationships (Buss, Gomes, Higgins, Lauterbach 1987) the intended effects of both conscious and unconscious manipulation have not been approached from an evolutionary behaviorist perspective.
One area in which the idea of shaping behavior could be leveraged is in the debate between homicide byproduct theorists, (Daly & Wilson, 1988) and homicide adaptation theorists (Duntley & Buss 2011). Step parents are far more likely to abuse step children than genetically related parents and being a step child is one of the major predictors of child abuse. This has been termed “the Cinderella effect” and Daly and Wilson argue that the reason step parents abuse step children is because the evolved parental psychological mechanisms that enable the generous care of children in biological parents are not fully activated in step parents, if they are activated at all. On the other hand, Buss and Duntley argue that step parents have adaptations to directly inflict costs on and kill step children. But why is it that step parent abuse towards step children is more common than homicide? Why are there are adaptations for “inflicting costs” as well as killing? Why would stepparents not more efficiently cut off the flow of resources to non-genetic children via homicide but instead engage in a protracted period of inflicting costs? From a purely resource safeguarding standpoint, if a man wants to best use parental resources to his genetic offspring the most effective strategy is for him to kill these non-genetic offspring. A well-known example of this behavior is that male lions quickly kill all the cubs in a pride when they are newly dominant (video of the behavior here). In humans, a period of inflicting costs is actually much more costly to the male because any abuse above and beyond negligence will cost him time and energy as well as leave him a child that is potentially injured due the abuse and will therefore require extended maternal time and energy and potentially more resources. From the perspective of evolved psychology for shaping the behavior of others the question of the adaptive value of abuse can be answered. From this perspective, abuse is one way to deter step children from soliciting parental resources by using primary punishers such as pain, bodily injury, or humiliation. Another way of reducing the incidence of costly solicitous behaviors in an unrelated child, is to extinguish the behavior. Many (but not all) behaviors undergo extinction when they are no longer followed by a positive consequence. From this perspective, neglect can be seen as a way of not responding to the demands of unrelated children and further decreasing their depleting of resources that can be channeled to other adaptive opportunities.
Another puzzle evolutionary psychologists have wrestled with is prosocial punishing (AKA altruistic or third party punishment). Why would anyone expend time or effort to punish an individual whose actions do not directly affect them? There have been many reasons posited for this including social displays, displays of dominance and even group selection. The simplest explanation may be that this prosocial punishing is shaping the behavior of the individual being punished in order to deter them from engaging in costly acts in the future. From this perspective one could predict that prosocial punishing is more likely to occur in domains in which one predicts they may find themselves or their kin with the individual perpetuating the undesirable behavior. For example if I am a man with a daughter I will be more likely to prosocially punish a man who sexually assaults another woman in my group than if I am a man with a son. In small hunter gatherer societies, deterring an undesirable behavior in an individual may have conditioned him or her not to perpetuate that behavior in the future. With a psychology that expects interactions at some point with every individual of the community prosocial punishment may be a preemptive means of avoiding behaviors that will be costly in the future.
Manipulation of Mates: Inflicting costs and doling out rewards
Of all the relationships humans have with one another romantic relationships often result in the most shared genetic fate. Because the behavior of one’s mate may have such a strong impact on one’s reproductive success, tactics of shaping behavior should be heavily employed in the context of romantic relationships. We should expect that whether consciously or unconsciously primary reinforcers and punishments should be used in mating contexts from the initial stages of courtship to managing parental duties and deterring the divergence of resources.
In Buss et. al’s 1987 paper, 6 major classes of manipulation tactics were found in romantic relationships: charm, silent treatment, coercion, reasoning, regression and debasement. Each of these can be viewed from a social shaping perspective. Charm, reasoning, and debasement can be viewed as examples of positive reinforcement. The “charm” tactic included acts of love and affection, compliments, gifts and promising a reciprocal favor in exchange for a behavior in the mate. The reasoning tactic included the item “I point out all the good things that will come from [the desired behavior]” which may represent a further extension of shaping via positive reinforcement. Finally debasement could be a tactic used if the couple is asymmetrical in mate value. Cognitions regarding having equal or better mate value than one’s partner may be primary reinforcers because they signal a reduction in likelihood of being abandoned and a increase in likelihood of sharing of resources. If one mate debases or lowers themselves in the eyes of the other mate, this could play on the primary reinforcers intrinsic to motivations in the psychology of human mating. Silent treatment, coercion and regression are forms of punishment. Coercion, including threats and criticism is simple punishment similar to other kinds of abuse. Silent treatment, like the neglect discussed above, represents “negative punishment” withholding positive consequences for the behavior as an adaptive tactic meant to extinguish the behavior. If the actor values social interaction with the mate taking away this reinforcer is a way of simultaneously punishing the behavior and no longer supplying positive reinforcement towards that behavior. Regression, including whining, sulking, crying and pouting parasitizes the domain of child rearing in order to meet its ends. Children engage in these tactics in order to elicit parental resources via manipulating parental mechanisms attuned to real displays of these emotions in emergency situations. Because of error management, parents err on the side of giving children what they want when crying or screaming because ignoring a desperately needed appeal is more costly reproductively than sometimes giving in to ersatz appeals. By using these same tactics a mate may be able to elicit the behavior desired.
More subtle and potentially unconscious forms of shaping behavior may also be employed. Pupil dilation signals interest and love and may act as primary reinforcers on the part of both romantic partners. Pupil dilation in response to statements of love and affection can reinforce the cognitions that stimulate these expressions and make them more frequent. Women use sex as a reinforcer and the withdrawal of sex as a punishment which also need not be conscious. Males may use resource displays, statements of affection and exclusivity including the derogation of other females in order to elicit sexual or other behavior desired from the female. A final example is that men and women may evoke jealousy directly (in a social encounter with a sexual rival of the mate) or indirectly (by speaking highly or commenting on the attractiveness of an intrasexual rival) in order to produce a punishment for unwanted behavior on the part of the mate.
Antagonistic coevolution and responses to shaping adaptations
Because individuals evolved the ability to shape others’ behavior a number of coevolved adaptations may have arisen in response to this. As discussed above positive reinforcement may be given to an individual in return for an act or behavior. Individuals will calibrate their expectations for positive and negative consequences for their actions based on previous consequences.
In a famous study in developmental psychology, children were divided into three groups. One group was given a reward for coloring with magic markers (an enjoyable activity) another was given a reward unexpectedly and the third group was given no reward (Lepper, Greene & Nisbett, 1973). The reward consisted of a ceremonious presentation of a “good player” award which simultaneously was a material and social reward. Later, they saw how long children played with markers when given free opportunity depending on if they had been rewarded for doing so in the past. They found that the children who were initially rewarded played with the markers the least while children that had been given an unexpected reward played with them the longest. Over the years this effect has come to be known as the “overjustification hypothesis” and individuals who are rewarded for actions they intrinsically enjoy reduce their frequency of these actions when given an expected extrinsic reward. An evolutionary psychology of learning could have a different rationale for this effect. If individuals have evolved in an environment where their behavior was being recurrently shaped by other humans using reinforcement then we may have an evolved psychology to “hold out” on performing rewarded actions unless this benefit is initially offered. Rewards that are given before an action is performed are more obviously given by other humans (environmental rewards rarely precede a behavior) whereas beneficial consequences that occur after an action is performed could be “payment” or a positive conditionality generated by the environment. One prediction generated by this hypothesis is that if a reward that follows a behavior is obviously generated by the environment that behavior will either stay the same or increase in frequency whereas if a positive consequence is obviously generated by a human potentially shaping one’s behavior then the behavior may be “withheld” unless the reward is seen as available or unless the individual has some evidence that they will be given a greater reward for the same action.
Other ramifications of the current thesis
Reinforcements and punishments that are socially provided by conspecifics need not be detrimental to one’s fitness or otherwise “manipulative” but may help people achieve shared adaptive goals better than without them. On the one hand, they may act as guideposts about socially appropriate behavior but on the other hand they could also disrupt behavior moving it towards the selfish ends of others and away from one’s adaptive goals. But do individuals have an evolved psychology of shaping their own behavior towards adaptive goals with reinforcement and punishment?
It is rumored that B.F. Skinner used operant conditioning techniques on himself to increase his productivity. As a reward for work completed he would take a break from working to listen to music or play ping pong, one of his favorite pastimes in graduate school. One of my best friends in college used a similar tactic with herself, smoking a cigarette each time she completed a page of a term paper. Consciousness itself may be the most comprehensive shaper of one’s own behavior. Consciousness affords us the ability to view ourselves as an outside actor thus enabling one to punish or reinforce one’s own behavior in light of the conscious plans and goals that one has for oneself. Internal emotional states act as positive and negative consequences for actions that one’s consciousness deems acceptable or unacceptable.
Anderson, J.R. (1998). Social stimuli and social rewards . J. Neurophysiol. 92, 3056–3068.
Buss, D.M. (2005). The murderer next door: Why the mind is designed to kill. New York: The Penguin Press.
Dawkins, R., & Krebs, J. R. (1978). Animal signals: information or manipulation. Behavioural ecology: An evolutionary approach, 2, 282-309.k
Daly, M., & Wilson, M. (1988). Homicide. Hawthorne, NY-. Aldine de Gruyter.
Deaner, Robert O., Amit V. Khera, and Michael Platt (2005) Monkeys Pay Per View: Adaptive Valuation of Social Images by Rhesus Macaques. Current Biology, Volume 15, Issue 6, Pages 543-548
Duntley, J. D., & Buss, D. M. (2011). Homicide adaptations. Aggression and Violent Behavior, 16(5), 399-410.
Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining children’s intrinsic interest with extrinsic rewards: A test of the “overjustification” hypothesis. Journal of Personality and Social Psychology, 28, 129-137.
Routh, D.K. (1969). Conditioning of vocal response differentiation in infants . Developmental Psychobiology, 1, 219–226 .