(p118, Seven Sins of Evolutionary Psychology).
Wednesday, December 01, 2004
The Problem of Incest: A Debate with Dr. Sandra Leiberman
Arguably, the most fixed and inexplicable behavioral artifact of human nature is the near universal predilection towards incest aversion. The fact that we are disinclined to be sexually drawn to kith and kin certainly doesn't stand to reason, namely the rationalistic mental calculus that informs our common sense. Lately, evolutionary psychology has adduced it all to an inborn neural circuit, activated by growing up in close proximity, that allows us to instinctively detect kin, and thus to avoid sexual relations with them. But are we like turtles and finches, and possess an imprinting mechanism, or is there another process at work, at once more subtle and plainly in sight? That mechanism is learning, or how experience shapes behavior. However, what common sense informs us learning is like and what the modern science of learning tells us are quite different things, and engage at once not common sense but insights from our animal cousins and the quirky mechanics of our own brains. So how do these arguments shape up? I offer two different theories on the matter, an explanation based on learning theory, and an explanation based on evolutionary psychology. Which one is right? That I leave to the reader to decide. I do offer however a debate on the issue between me, the modest author of the learning theory argument, with Dr. Sandra Leiberman, the primary author and exponent of the evolutionary psychology or EP position.
Marr: I enjoyed your article on incest. Your data, methods, and conclusions are very clear, and I concur in general with your conclusions. My take however on operative and neural explanation of incest avoidance course is different, but as I am sure you agree, an explanation of incest must be based on neuroscience.
Leiberman: No, I would suggest that the neurophysiological instantiation of programs for incest avoidance is ONE level of explanation required for a complete understanding of incest avoidance behaviors. It is not necessary for the explanation to be based on neuroscience. Other levels of explanation include considerations of the adaptive problems our ancestors faced and the types of cognitive programs expected to exist to solve each problem. Any Tooby and Cosmides article should spell this out loud and clear.
Marr: For issues regarding behavior and motivation, a hybrid discipline has recently arisen called ‘affective neuroscience’. Practiced and espoused by psychologists such as Jaak Panksepp, Gerald Edelman, and Kent Berridge, affective neuroscience mixes and matches a behavioral language such as response rate, affect, etc. with neurological language such as dopamine, nuclei etc. to arrive at explanations for behavior that require behavior to be informed by the data language of neuroscience, but not necessarily to be reduced to neuroscience. Thus a statement that behavior need not be explained by neuroscience is a bit disingenuous, since neuroscience or the metaphors of neuroscience are always part of the equation, much like our common mixing and matching of metaphors of pain and suffering with biological metaphors of viruses and infection allows us to understand our bodies and how to care for them.
In the past ten years, affective neuroscience has demonstrated that human ‘drives’ for such things as food, sex, etc. are not singular processes, but may be bifurcated into processes that are different psychologically (i.e. behaviorally) and physiologically. Thus ‘wanting’ and ‘liking’, anticipation and consumption are two very different things that correspond to different behavioral laws and different physiological processes. For the sex drive in particular, sexual anticipation is different from sexual attraction, and the hypothesis necessarily follows that sexual anticipation is amenable to the same behavioral principles such as ‘blocking’ that have been demonstrated to inhibit similar drive like behavior such as approaching food, avoiding shock, etc.
Regarding incest, the question is whether proximity or predictability is the key to sexual aversion. That is, if sexual predictability is experimentally altered by changing expectations (as in Oedipus’s predicament) or incidental expectations (as in a child growing up and not historically expecting sexual relations with kin), will the 'habit' of sustained expectations block sexual anticipation if those expectations are logically removed in the future? For example, would Oedipus see his wife again in the same light if it were revealed a few weeks later that she was not indeed his mother? (I may add that proximity or associative learning was a key principle in early Pavlovian models of learning, but this has since been superceded by present day interpretations of Pavlovian conditioning that see predictability as key, not proximity. For example, a bell (CS) is associated with food (US), and an animal subsequently salivates upon the solitary introduction of the bell. Present day learning theory attributes this behavior to the value of the CS as a predictor of food rather than due to its mere association with food. It is thus ironic that EP conceptualizations of incest avoidance as due to a pairing of stimulus events --brother and sister in pre-adolescence- entail de-facto Pavlovian mechanisms that have been long been discredited in learning theory!)
These are hypotheses that do not have to engage evolutionary psychology, but in a general discussion must do so because in my opinion the teleologically informed principles of EP block the consideration of alternative hypotheses such as these. For example, it was because of the teleological principles of a God creating the universe and man that stopped a serious consideration of Copernicus’ heliocentric model of the universe, and for Darwinian evolution, it impedes consideration of evolution to this day.
Marr: Blank slate explanations for behavior can be attributed to Skinnerian methodological behaviorism, but must not be extended to behaviorism itself, which has long recognized the influence of innate or nativistic influences on behavior. Indeed, the data language of Pavlovian conditioning is commonly used in conjunction with neurological data to provide explanations of behavior that are far removed from the Skinnerian point of view. A case in point of course is the concept of blocking that I briefly discussed in the small article I sent you. The web site of Kent Berridge is an excellent starting point for knowledge of this bio-behavioristic perspective. If you posit an evolved neural architecture or module to explain incest, you must have a program to test that hypothesis. That of course involves neuroscience.
Leiberman: Not necessarily. It is possible to generate TESTABLE HYPOTHESES regarding the design of an information processing circuit for kin detection and sexual aversions. You do not need to rely on neuroscience. It will be important to look for neural circuitry in the brain to have a complete understanding of how this system works, but as I said above, a neuroscience level of explanation is compatible with other cognitive explanations of behavior.
Marr: If you are going to design a testable hypothesis regarding the design of an information processing circuit for kin detection, you must rely on neuroscience, and would if you were testing the design of an information processing circuit of silicon or biological cells. For example, if I were to deduce the circuitry that is embedded in the microchip of my computer, the inputs (keystrokes) and outputs (readouts) tell me nothing about its design, or whether the design is digital, analog, parallel, sequential, built on silicon, or built on neurons. To find out how computer circuits work, I therefore have to constrain my theorizing by actual facts about how microchips are built. Many cognitive explanations for behavior do indeed ignore the workings of the human brain, and that is a failure noted by many neuroscientists. In the words of the distinguished neuroscientist Jaak Panksepp, "it is a mistake to believe something is biologically real simply because one can computationally simulate the shadow of an end result…. it is certainly premature and unwise, for any science of mind to neglect the brain, as is still too common in evolutionary psychology, and most of psychology, in general."
(p118, Seven Sins of Evolutionary Psychology).
(p118, Seven Sins of Evolutionary Psychology).
Marr: Neuroscientific perspectives on evolutionary psychology however take a very different route than that espoused by > EP practitioners. I strongly recommend you peruse Jaak Panksepp's article on the ‘7 sins of evolutionary psychology’ to note ‘corrective’ suggestions to EP that I espouse as well.
Leiberman: I am sure there is misunderstanding here too.
Marr: When Panksepp wrote his article on the Seven Sins of Evolutionary Psychology, he invited distinguished evolutionary psychologists such as Steven Pinker, David Buss, Leda Cosmides, and John Tooby to respond for a follow up article that appeared in the Journal Evolution and Cognition entitled Seven Sins for Seven Sinners. They refused comment. So much for resolving misunderstanding if they cannot contribute to a distinguished journalistic forum to achieve this end.
Marr: Your data and conclusions fit almost exactly (in my opinion) with a Pavlovian explanation that in turn maps to neurological events that have been well observed in the current literature. The difference is that incest aversion is due to the recruitment of many different morphological structures in the brain, from the neocortex to the hippocampus. Some of these structures evolved from hunter-gatherer days, and some are far more ancient (particularly the dopamine system). It is my strong belief that incest avoidance is a spandrel from their integrated activity and is not an adaptation. This I am confident can be demonstrated theoretically and empirically.
Lieberman: As for the brain regions, this all sounds fine to me. However, I am not going to argue whether incest avoidance is an adaptation -- there is clear evidence from humans, non-human animals, plants, and other more simple creatures that avoidance of sexual reproduction with individuals having a high probability of sharing genes by virtue of common descent would have been advantageous. The fact that an incest avoidance system might use some of the same neuro-chemicals to regulate attraction that are used for other purposes does not mean that it is not an adaptation. Again, you are confusing levels of explanation. The same building blocks can be used to have extremely different effects computationally.
Marr: Incest of course has deleterious effects, and I agree with you that there are sound logical grounds for its avoidance. But you cannot impute mechanism from logical necessity. In the natural world incest avoidance is also due to the unintended consequences of behavioral tendencies that occur for reasons that have an entirely different logic. For example, animals usually disperse into the environment shortly after birth, thus making it practically impossible for kin to interbreed. But, as any farmer will tell you, keep a litter of bunnies/pigs/puppies etc. together, and incest will occur as a rule. So although dispersion occurs for certain evolutionary reasons, the unintended practical consequence or spandrel of incest avoidance follows because animals are dispersed into a larger group.
I also fail to see how levels of explanation are confused, since affective neuroscience uses converging methodologies, and does not conflate them. That is, I can use neurochemistry (e.g. dopamine systems and the behavioral processes that they influence) to inform avoidance behavior just as I use neurochemistry (adrenaline) to inform why other primary behaviors such as anger and fear. That an incest avoidance system may use neurochemicals presupposes that you have a neurally based theory for such a system to begin with. EP does not have this, and is thus unconstrained by any neurological facts.
Marr: You use information processing metaphors to explain incest avoidance, but much of the brain works in an analog rather than digital manner, and these metaphors simply cannot apply. Blocking of course is a case in point.
Lieberman: Last I checked, the brain is a computational device that processes information. This is a level of explanation that allows one to think through the kinds of information required to solve a particular problem.
Marr: If the last time you checked the brain was a computational device, may I suggest you check again. According to modern neuroscience, the brain is far removed from the computational picture so commonly portrayed by EP. To pursue this argument let me quote the distinguished neuroscientist Jaak Panksepp from his article "The Seven Sins of Evolutionary Psychology" .
"Although there is increasing talk of neural circuits for cerebral modules, especially since evolutionary psychology became a compelling view in cognitive neuroscience, none of the proposed sociobiological modules have coalesced with establish neural realities. … Instead of a solid confrontation with the brain, there is abundant talk of computational representational views that ignore the fact that many neuroscientists are not convinced that such information processing metaphors provide much that resembles an accurate perspective of how the brain creates meanings. Perhaps the higher cortical systems are ‘computational’ by some stretch of the digital information processing metaphor, but the subcortical reaches that mediate emotions and motivations are not. ….We should never forget that the capacity to simulate certain brain functions in a digital processor does not mean those computations reflect physiological realities…. Evolutionary psychology, as most other forms of cognitive psychology, has been inspired by computer science rather than the molecular biology/neuroscience revolutions of the past three decades. Indeed, at times, it seems that practitioners of evolutionary psychology have an active aversion to organic perspectives. They talk about the brain simply as a modular computational device. Although that view has also been pushed forward by many cognitive neuroscientists, an equally fundamental alternative is rarely discussed: mental life is fundamentally organic."
Marr: Ultimately, an explanation for incest is an empirical question, not a theoretical one. Presently, we have the tools and the data to arrive at an explanation for incest aversion that will demonstrate how evolved structures account for the behavior, but this explanation will be constrained by neuroscience, a constraint that EP has scarcely acknowledged and has never tested.
The brain is not a computer, sexual anticipation as a ‘behavior’ is not controlled for experimentally in any research on incest (thus your work is incomplete and conclusions premature), and the metaphors and evidence of neuroscience is a critical and integral part of the psychological enterprise, and are wrongly not considered by EP and much of cognitive science (that is, contrary to what you say, cognitive science is not compatible (or integrated) with neuroscience because generally it’s not informed or constrained by neuroscience to begin with!). As for incest, my hypothesis is clear. It must be tested empirically rather than be dismissed or even confirmed logically. That must be the enduring challenge of your work as a psychologist and a scientist.