(James Q. Wilson's 'The Moral Sense')
H.J. Eysenck.
Criminal Justice Ethics, Summer-Fall 1994 13 n2 p53-61
Abstract: Criminology has demonstrated that there are differences in the propensity to perform anti-social acts, and this differing propensity demonstrates that a moral sense, as described by James Q. Wilson in 'The Moral Sense', does exist in people. Conscience itself is not the moral sense but the effect of conditioning. Variable success in conditioning conscience does indicate that there are physiological differences that lead individuals to behave differently.
Wilson fights a brave battle against political correctness (and factual incorrectness), the absurdities of moral relativity, and Marxist social enviromnentalism. It is difficult to see how anyone with an ounce of common sense could ever believe that our standards of behavior reflect the needs of the social order, or are determined by the economic dictates of our means of production, distribution and exchange. The arguments in favor of such an unlikely thesis are so obviously incorrect that it should not take much research to disprove them. Let me look at the problem first in common sense terms; I will then take up the more difficult task of trying to show how conscience (the bearer of our moral sense) originates.
Those who favor moral relativism often point to the fact that certain activities are freely permitted in some societies, but severely punished in others. Even in the same society there may be sudden changes from forbidden to permitted; homosexuality was against the law in England until 1967 when it suddenly became legally permissible. In fact, homosexuality is a good example for relativity of moral sense, being actively encouraged in ancient Sparta, and punished by torture and death in some medieval societies! Sodomy is now permitted in England between consenting males over eighteen years of age but forbidden between man and wife! Altogether, sex provides us with a multitude of insane laws--I recall attending an annual general conference of the American Psychological Association where the printed material sent to all participants included a separate notice warning us that in the particular Southern state where the conference was held oral sex was legally prohibited, even among married couples! (It would be interesting to know how this law was enforced.
There are many political and religious areas where there are sharp differences from country to country, and from time to time. In the Hitler Germany where I grew up (and which I left in protest) one could be severely punished for not giving the Hitler salute. In Catholic Italy Galileo was severely punished for asserting the truth about our heliocentric solar system. In Turkey you once could have your nose slit for smoking-modern America seems to be approaching the same point, but as yet has not quite reached it. Prostitution is not legal in the USA, but permitted in Germany and England. The list is endless, but what does it prove? It simply shows that powerful people may impose their social, religious or moral ideas by law, usually by inventing victimless crimes. But these are a tiny minority of "crimes"; the vast majority of crimes do have victims (burglary, assault, mugging, rape, murder, thieving, looting, robbing, grievous bodily harm, industrial espionage, and many more). And these crimes are universally condemned; there is no relativity there. From time immemorial men (and occasionally women) have committed these crimes, and from time immemorial the state has punished the wrongdoer. There is 95 percent of universality and 5 percent of relativism about our moral sense; to believe in 100 percent relativity is not meaningful. No state could exist for any length of time that permitted its citizens to carry out these activities without let or hindrance, and this solid core of "don't do to others what you would not like them to do to you" establishes a robust air of universality for our moral sense.
Is this moral sense learned, in the sense that we learn Latin, or algebra, or history? Clearly not. Criminals have often been quizzed, as have normal citizens, about the morality of certain actions, and their lawfulness; criminals know as well (or even better!) what is lawful, what is not. After all, it is their business to know! But this knowledge is quite irrelevant to their actions; every criminal who steals, or rapes, or murders, knows perfectly well that what he is doing is unlawful, and subject to punishment; this does not prevent him from carrying on stealing, raping and murdering. This of course leads us to the question of why some people indulge in these activities. But surely that is the wrong question. What we should ask is why we do not, all of us, indulge in this type of behavior. After all, we all want things we cannot afford, lust after women we cannot have, hate people we cannot hurt--why not steal the things we want, force the women to do what we want, beat and kill the people we hate? The risks are minute--only a small minority of women complain to the police about a rape; of the rapes reported to the police, only a small minority lead to prosecution, and of those accused, only a small minority are convicted. And of those convicted, our benighted judges are only too happy to hand many over to probation officers, or give them suspended sentences, or let them off with a fine.
The same is true of muggings, burglaries, or auto thefts--the risk of actually getting caught is minute, particularly if your IQ is above 100 and you exercise a minimum of care. And if you are in your teens (when the majority of crimes are done), then even if caught, the chances of punishment are minimal. In some American states the average number of cautions for juvenile criminals is thirteen before any real punishment is handed out--in other words, you have to be caught thirteen times before the state inflicts any kind of retribution on you! As the chance of getting caught in the first place is less than one in ten, you can go on an orgy of crime without any likelihood of real punishment. So why do most of us refrain from satisfying our needs and wants in the most direct manner? The traditional answer of course says that to do so would be against our conscience, and to do that would be so hurtful that the resulting feelings of guilt and shame would more than counter-balance the illicit gains produced by our crimes.
Common sense is obviously right in a general sense. Most of us do have feelings of guilt and shame, and these can produce quite strong aversive results. We have to weigh up the gains we expect from our illegal activities against the hurt of our outraged conscience, and. this, together of course with the threat of possible punishment, however remote, is apparently sufficient to keep us in the ways of righteousness. But what is "conscience"? To the religious it is a moral sense given to us by God to keep us on the straight and narrow. But that of course will not do; it tells us nothing about "conscience" other than that it exists. How does it originate? Can we produce it, change it, abolish it? What is the contribution of heredity? Of environment? What processes are active in producing it? We have ruled out learning, so what processes mediate acquisition? These and many other factual questions need answering before we can accept the easy answer that it is conscience that makes us behave morally. What produces the difference between people with and without consciences? God's will seems a very arbitrary answer--why does he give one person a conscience and withhold it from another? Is it fair that the former should go to heaven, the latter to hell? Credo quia impossibile may satisfy the religious, but it leaves a lot to be explained.
The Genetic Origin of the Moral Sense
Let us start with the famous sociobiological fable of the hawks and doves. The fable imagines an ideal human society, functioning optionally, with all co-operating, behaving altruistically, and showing high moral sense. Into this genetically dovist society intrudes, through mutation, a solitary hawk-aggressive, competitive, psychopathic. Being very successful in a dovist society, hawkist genes soon spread, with the hawks dominating the doves. But soon there are too many hawks for comfort; they compete and mutually destroy each other. Society finally settles down to a stable status with many doves and fewer hawks. This, roughly, is what we find in our society although, of course, there are no "dove" or "hawk" genes, but a number of genes giving rise to a roughly normal distribution of gene-combinations, from relatively pure doves to relatively pure hawks. If this be true, Criminal behavior should show strong evidence of heritability. I have reviewed the evidence in some detail,(1) and it leaves little doubt on the heritability of criminal behavior; something like 60 percent of phenotypic anti-social behavior is caused by genetic factors. To summarize briefly the evidence, we have two major sources of investigation. The first is concordance in identical (MZ) and fraternal (DZ) twins. In other words, we identify prisoners who have a male twin; we then search out the twin and establish zygosity and criminal status. He is concordant if he has also been in prison, discordant if he has lived a blameless life (as far as the police know!). If heredity plays an important part in criminality, MZ twins should show greater concordance than DZ twins. Seven older studies showed an MZ/DZ ratio of 2.23; they have sometimes been criticized as possibly being selective. The most recent study, by Christiansen, looked at a total population of twins from a well-looked defined area--Denmark--and found a ratio of 2.38.(2)
These ratios establish heredity as an important cause of criminal behavior, but they probably underestimate the true state of affairs. Some of the twins may have committed crimes, but not been caught; they would count as discordant when they are really concordant. This is much more likely to drag down the concordance score of the more concordant MZ group; random errors of measurement always reduce the true difference between groups.
Other deductions can be made from the concordance hypothesis. Take the threshold model, which states that the thresholds where criminal disposition turns into criminal behavior is higher for women than men, and is defined by the population prevalence of criminality. Hence females who commit criminal acts would be more deviant in liability, and consequently their relatives would be more deviant as a group than the relatives for deviant men. They are.
The second method that has been widely used in the study of a genetic basis for criminal behavior is adoption. If we look at the adult behavior of babies of criminal or non-criminal parents adopted at birth by criminal or non-criminal adoptive parents, the former contribute the genetic basis, the latter the environmental basis of later behavior. Looking at the criminal records of the adoptees, we can see to what extent they resemble that of the biological or the adoptive parents.(3) The data from several studies demonstrate the importance of genetic factors, but also, as one would expect, find the influence of the adoptive parents significant. Most important was the synergistic interaction of the two. In one study it was found that when congenital but no post-natal factors predisposed to petty criminality, the risk of such criminality in the genes was increased fourfold. But when both congenital and post-natal factors were present, the risk was fourteen times that of a control population.(4)
These data leave little doubt about the importance of genetic factors. Yet it is important to realize that there are no genes from criminality. What we do have, apparently, are genes coding for structural proteins and enzymes that influence metabolic, hormonal and other physiological processes; it is these that may indirectly modify the risk of crucial behavior. With heritability around fifty to sixty per cent, and possibly somewhat higher if we take measurement error into account, there is still plenty of room for environmental influences.
To complete this section, it may be useful to look at the other end of the hawk-dove continuum. If hawkist criminality is partly inherited, how about altruism? In one of our studies we addressed this problem, using three questionnaires measuring altruistic tendencies in MZ and DZ twins. For the three scales, rough estimates of heritability were fifty-six, sixty-eight and seventy-two per cent respectively. A maximum likelihood model-fitting analysis revealed about fifty per cent of the variance on each scale to be associated with genetic effects, virtually none to be due to the twins' common (home) environment, and the remaining fifty per cent to be due to each twin's specific environment and/or measurement error associated with the test. In other words, as one might have expected, genetic effects are equally as strong for pro-social as for antisocial behavior.(5)
One final point. If antisocial behavior is genetically determined to some extent, we would expect this to show up in animal behavior. Using obedience conditioning as a measure, Scott and Fuller found that different types of dogs showed very different effect reactions, as did individuals within each breed.(6) Basenjis emerged as the most "psychopathic," cocker-spaniels as the most pro-social. Training had a variable effect on different breeds. Evolution and differential environments clearly determined the dogs' behavior, with heredity a prominent factor. I shall return to dog "criminality," shame and guilt feelings in the next section.
Personality and Conditioning
Given that DNA determines to a large extent the antisocial or pro-social behavior of the growing child, it may be useful to look at the various traits of personality that are correlated with antisocial behavior. This may give us a clue as to the processes that mediate anti-social behavior and, finally, the psychophysiological and hormonal components of criminality. The literature has been surveyed in detail elsewhere.(7) The three dimensions have been entitled E (extraversion), N (neuroticism), and P (psychoticism), although many other terms have been used in the past.(8) Typically, the high E scorer is sociable, lively, active, assertive, sensation-seeking, carefree, dominant, surgent and venturesome. The typical high N scorer is anxious, depressed, tense, irrational, shy, moody, emotional, and has low self-esteem. The high P scorer is aggressive, cold, impersonal, impulsive, tough-minded, unempathic, and impersonal. High scores on these three dimensions define an octant in the three-dimensional globe defined by these three variables. Eysenck and Gudjonsson cite a large body of evidence to show that anti-social behavior in children, youths, and adults correlates with P, E and N, although there are some age-related differences.(9) E is more important for anti-social behavior in children than in adults, N more in adults than in children. But overall these three factors account for a surprisingly high degree of anti-social behavior.
It has been possible to correct together the test items most predictive of anti-social conduct into a scale, called the Criminality Scale, using items from the EPQ (Eysenck Personality Questionnaire). Figure 1 shows the mean scores of non-criminal and criminal groups, with the criminal groups arranged in order of seriousness of crimes committed. It will be clear that there is a linear progression of scores, from the non-criminal to the most seriously criminal. Why do we have these correlations?
The argument I have developed, stated briefly, is that we do not commit crimes because our conscience holds us back.(10) Conscience is the product of Pavlovian conditioning. Pavlovian conditioning pairs a neutral stimulus, (CS) with another, unconditioned stimulus (UCS) that has some beneficial or harmful effect which produces a response (UCR). After a number of repetitions the CS will become associated with the UCS and provoke the same UCR. How does this work for the growing child?
Suppose that our little boy misbehaves. Immediately his mother will give him a smack, or stand him in the corner, or send him off to his room, or inflict one of the many punishments which have become customary with parents over the centuries. In this case, the particular asocial or antisocial activity in which he has been indulging is immediately followed by a strong, pain-producing stimulus. The conditioned stimulus is a particular--kind of activity in which the child has been indulging; the unconditioned stimulus is the slap, or whatever constitutes the punishment in this case, and the response is the pain and fear produced in the young child. We would expect conditioning to take place, so that from then on this particular type of activity would be followed by a conditioned fear response. After a few repetitions, this fear response should be sufficiently strong to keep the child from indulging in that type of activity again.
There are, of course, many such activities that are punished; exactly the same situation hardly ever recurs twice. Nevertheless, we would expect a fairly general reaction of fear and autonomic "unpleasures" to become associated with all antisocial activities because of the process of stimulus generalization. In fact, stimulus generalization would be expected to be enhanced considerably by the process of "naming" in which parents usually indulge. Every time the little child misbehaves, its misbehavior is labelled "bad," "naughty," " wicked," or whatever the term chosen by the parents might be. Through this verbal labelling the child is helped in the generalization process and finally groups all these activities together by association as being potentially dangerous, punishment-producing, and particularly as being productive of conditioned anxiety and fear responses. Thus our little child grows up, gradually acquiring a repertoire of conditioned fear responses to a wide set of different behavior patterns, all of which have one thing in common--that they are disapproved of by parents and teachers, siblings and peers, and that they have, in the past, frequently been associated with punishment and, therefore, with the consequent autonomic upheaval.
What will happen when the child is in a situation where temptation is strong to do one of these forbidden things? The answer is, of course, that he will tend to go and do it. But as he approaches the object arousing the temptation, there should also be a strong upsurge of the conditioned emotional reaction, the fear or anxiety which has become conditioned to his approach to such an object under such circumstances. The strength of this fear-anxiety reaction should be sufficient to deter him from pursuing his antisocial activities any further. If it is indeed strong enough, then he will desist; if it is not, he will carry on, in spite of the increasing strength of the fear-anxiety response. It will be seen, therefore, that whether he does or does not behave in a socially approved manner depends essentially on the strength of the temptation and on the strength of the conditioned avoidance reaction which has been built into him, as it were, through a process of training or conditioning.
A good experimental test of the hypothesis is a study carried out on dogs. These experiments were carried out by Richard L. Solomon, and some of his colleagues at Harvard University, using six-month-old puppies.(10) Later experiments have also been carried out with young children, but we shall concentrate here on the animal experiments. These were conducted in so-called "taboo situations," held in a training room, fairly sound-proof, and equipped with a one-way mirror. A chair was placed in a comer of the room, and in front of each front leg of the chair were placed two small dishes. The experimenter sat in the chair, holding in his hand a rolled-up newspaper with which he could swat the puppies on the rump. Each of the puppies was deprived of food for two days and was then brought into the experimental room. In one of the dishes had been placed boiled horse meat, which the puppies very much liked, whereas in the other dish was placed a much less well liked commercial dog food. The puppies usually made straight for the horse meat, but as they touched it they were swatted by the experimenter. If one gentle blow was not enough, then the puppy was swatted again and again until he finally gave up his attempts to eat the horse meat. Usually several further attempts were made, until the puppies finally turned to the commercial dog food, which they could eat without being swatted.
This training was carried on for several days until the puppies had firmly learned the taboo on horse meat. The experiment then turned to what was called the "temptation testing- phase. Again the puppies were deprived of food for two days and then brought to the room, but this time with the experimenter absent. Again a choice had to be made between a dish of boiled horse meat and a few pellets of dog food. The puppies soon gobbled up the dog food, then began to react to the large dish of horse meat. In Solomon's words:
Some puppies walked around the room with their eyes towards the wall, not looking at the dish. Other puppies got down on their bellies and slowly crawled forward, barking and whining. There was a large range of variability in the emotional behavior of the puppies in the presence of the tabooed horse meat. We measured resistance to temptation as the number of seconds or minutes which passed by before the subject ate the tabooed food. The puppies were allowed half-an-hour a day in the experimental room. If they did not eat the horse meat by that time they were brought back to their home cages, were not fed, and, a day later, were introduced again into the experimental room. This continued until the puppy finally violated the taboo and ate the horse meat, or until he had fasted so long that he had to be fed in his cage in order to keep him alive.
There was a very great range of resistance to temptation. The shortest period of time it took a puppy to overcome his training and eat the horse meat was six minutes, and the longest period of time was sixteen days without eating, after which time the experiment had to be stopped and the puppy fed in his home cage. This great range of variability made it possible to test the influence of various experimental conditions on the growth of conscience in these puppies. For instance, it was shown that when the puppies were hand-fed throughout their early life by the experimenter, then they developed a conscience much more strongly than did other animals which had been machine-fed."(12)
Solomon managed to evoke feelings of guilt and shame (as demonstrated by certain types of behavior) in his dogs by manipulating the situation, but enough will have been said to illustrate my theme. What does it have to do with personality?
The Biological Basis of Emotional Behavior
Consider extraversion. At first sight one would feel that the extraverted person would be unlikely to be antisocial. Being sociable you would perhaps expect him to be pro-social. But consider the reasons why he is sociable. According to my theory, which has now received considerable experimental support,(13) extraverts have a cortex characterized by a low level of arousal, introverts have a cortex characterized by a high level of arousal, with ambiverts having a normal level. (I shall not go here into the reasons for these differences, which derive from psychophysiological causes.) Now people prefer intermediate levels of cortical arousal, neither too high nor too low. Consequently, the extravert will seek out stimuli that increase arousal (bright lights, loud noises, sexual stimulation, danger); all that is likely to attract the growing child to go after forbidden fruit. But there is another, absolutely fundamental effect of low cortical arousal. Pavlovian conditioning is dependent on cortical activity; lower it by depressant drugs and conditioning becomes much more difficult. Raise it by stimulant drugs, and conditioning improves. It follows that extraverts condition poorly, introverts well. Figure 2 shows the course of eye-blink conditioning in a group of extraverts and a group of introverts. In eye-blink conditioning a sound delivered over ear-phones is the CS; a puff of air to the eye is the UCs, leading to eye-lid closure as the UCR. After a few repetitions of CS-UCS-UCR, CR directly produces UCR, without the UCS. As Figure 2 shows, introverts condition much better and much more quickly than extraverts. It would seem to follow that if conscience is a conditioned response, and if introverts condition better than extraverts, then the conscience of extraverts should be at a much more primitive level than that of introverts. Raine cites a dozen studies that support this conditioning theory of conscience and anti-social behavior.(13)
Why is neuroticism involved? According to psychochological theory, P = H x D; that is, performance (behavior) is the product of habit and drive (motivation). Now anxiety, an important part of D, is a drive, and in line with theory it multiplies the antisocial habits formed by the extravert because of his lack of conscience, that is, his failure to condition early in life. This explains why E is more important in children(15) than in adults, as far as anti-social behavior is concerned; it is in childhood that most of the conditioning grows that is producing one's conscience; it is later, in one's teens and as adults, that N multiplies with the bad habits acquired in childhood to produce actual criminal activity. But how about psychoticism? Psychoticism is produced by hormonal and other reactions, involving in particular testosterone. Hence males have much higher P scores than females, and, as is well known, males are many times more likely than females to indulge in crime. (Most female crimes are related to sex, such as prostitution, and are relatively victimless.) Another important correlate of P is the neurotransmitter dopamine, which is associated with schizophrenia; and monoamine oxydase is also involved (through its relative lack) in P and in criminal activity. Zuckerman has treated the psychobiology of personality in great detail, and should be consulted for further information.(16)
The Moral Sense
The moral sense is a reality, as Wilson argues. It determines our pro-social and antisocial proclivities, and is of the utmost importance in creating the climate in which cooperation and other dove-like activities can flourish. There is a large hereditary component; even such attitudes as religiosity, which are closely related to conduct, are strongly determined by heredity.(17) Heredity acts through psychophysiological processes, hormones, neurotransmitters and other agencies that patient research is isolating. Specifically, the moral sense embodied in one's "conscience" is acquired through a process of Pavlovian conditioning, the functioning of which is itself genetically determined to a large extent.
Environment of course also plays a large part, usually in interaction with heredity. To take but one example, I have assumed that conditioning will be in the direction of pro-social behavior. But if a child is brought up in a thieves' kitchen, we would expect the introverted child to pick up the prevailing Mores more quickly than the extravert, and indeed there is empirical evidence of this happening.(18) Thus the theory has complexities to which a brief presentation cannot do justice.
Wilson's book is very welcome in that its message is clear and strongly opposed to much politically correct nonsense that is produced by writers and speakers having little or no knowledge of the facts of the situation or the better-established theories. If his book can be faulted, it is in relation to his failure to discuss the genetic literature, which hardly receives more than a brief mention. Yet this forms an absolutely fundamental basis for any form of moral sense. Unless and until we have a strong body of evidence on just what it is that is being inherited, we know nothing of the nature of conscience or the causes that build it. I have tried to indicate the lines such research should follow and have provided some experimental facts that seem to support the theory put forward. One might have wished that Wilson could have done more than just mention the major points of that theory without criticizing it; if it is wrong in part or whole, it should be remodelled or replaced.
Wilson has also failed to pay much attention to the many neurohormonal factors which play such an important part in predisposing people to criminal or altruistic behavior.(19) In other words, I feel that his book engages in many sword-fights with unworthy opponents, on a philosophical rather than a scientific basis, and neglects to deal in sufficient detail with the available experimental material and the more science-oriented theories. But perhaps that is intended for a future venture, or perhaps he feels that his earlier book with Herrnstein will fulfill that function.(19) But it would be unreasonable to assume that readers of this volume will have read its predecessor, and in any case a great deal has happened in the intervening ten years that might be mentioned.
NOTES
(1) H.J. Eysenck, & G. Gudjonsson, The Causes and Cures of Criminality (1989).
(2) Id.
(3) Id.
(4) Id.
(5) Id.
(6) J. Scott & J. Fuller, Genetics and the Social Behavior of the Dog (1969).
(7) Id.
(8) Id.
(9.) Id.
(10) H.J. Eysenck, Crime and Personality, (1977).
(11) Id.
(12) Id.
(13) H.J. Eysenck & M.W. Eysenck, Personality and Individual Differences: A Natural Science Approach (1985)
(14) A. Raine, The psychopathology of Crime (1993).
(15) Z. Zuckerman, Psychobiology of Personality (1991).
(16) L. Eaves, H. EYSENCK, & N. Martin, Genes, Culture and Personality (1989).
(17) H.J. Eysenck & G. Gudjonsson, supra note 1.
(18) A. Raine, supra note 13.
(19) J. Wilson & R. Herrnstein, Crime and Human Nature (1985).
