Psychology and Phronesis: the import of findings in psychology, social cognition and neuroscience for moral wisdom

By Marnina Norys

Note: This paper was presented to the Canadian Philosophic Association Congress 2010. McGill University (Montreal, QB), June 2, 2010.

Undoubtedly, many still recall the 1972 image of a young Vietnamese girl running down a road, her clothes burned off by napalm. Because the Vietnam War was the first war to be extensively televised, Americans were regularly exposed to disturbing images including ones of the point-blank execution of a Viet Cong terrorist, stacks of children’s corpses and bloody battles in which parents saw their own boys killed. Such coverage moreover, may have turned the tide of public opinion against the war. By the end of the conflict, a previously popular effort was now being called “immoral” by 65 percent of Americans. According to some, however, Vietnam was no more atrocious or barbaric than previous American conflicts that saw strong support. Thus, there is reason to believe that graphic depictions of the war worked to change some people’s attitudes.[1]

Some might argue that this moral turnaround that occurred during the Vietnam War was achieved because people attended to the horrible reality that lay before their eyes, rather than having quietly reflected on the import of their moral principles. Do we really need a moral principle to know that a stack of children’s corpses is a horrible consequence that we should aim to avoid? Can we not just see, sometimes, that what we are looking at is morally objectionable?

Moral perception, or the ability to recognise certain situations as calling for moral deliberation, has received a certain amount of attention in philosophical literature. One reason for an upsurge of interest in moral perception is due to the work of moral particularists such as Jonathan Dancy and Margaret Little, who argue that good moral deliberation is not adequately described in terms of appeals to exceptionless moral principles, or even the weighing of a plurality of moral principles. What is required, they argue, is to be sensitive to particular contexts when judging the rightness or wrongness of given acts. Moral perception, then, has been posited as a sort of sensitive faculty that we employ in making such determinations. In this paper I will describe moral perception and the role that such a faculty plays in morality. Because moral acts are undetermined by physical descriptions, however, I shall argue that certain inferential biases stand to distort moral perception. In the end I shall suggest a way of countering such biases that is suggested by recent research on cognition.

My purpose here, then, is neither to defend nor deny the utility of moral principles. Regardless of one’s stand on the debate between particularists and principlists, there are interesting things to say about moral perception and how it can go wrong. Even proponents of principle-based theories recognise a gap between moral principles and those situations falling under particular principles. For example, Onora O’Neill, a Kantian, admits that moral principles and concepts are to some extent indeterminate since their applicability is not dictated by explicit rules.[2] While many philosophers tend to focus on constructing and justifying sets of moral principles, when it comes to the moral judgements we make in our day-to-day lives a great deal of work done at the ground floor, as it were, or at the level of perception. At least sometimes, moreover, it does seem that we can witness an action and just see that it is wrong before we have any clear sense of the moral principle that may have been violated.

According to Little what we ‘see’ in such cases can be likened to an immediate perception of colour, while the exercise of wisdom and discernment are analogous to well-functioning faculty of sense:

Someone who sees that a rose is yellow … is hardly appealing (even tacitly) to a general principle to license some inference; she instead is justified in believing the rose to be yellow because her faculty of sight is in good working condition. So, too, we can be justified in our moral conclusions, not just by subsuming the situation under some general laws, but by taking in the situation, exercising discernment and wisdom—Aristotle’s phronesis[3]—and seeing that it’s cruel[4]

We might examine a case put forth by Gilbert Harman for a concrete example involving moral perception:

If you round a corner and see a group of young hoodlums poor gasoline on a cat and ignite it, you do not need to conclude that what they are doing is wrong; you do not need to figure anything out; you can see that it is wrong.[5]

It certainly seems that our reactions to Harman’s description are non-inferential, meanwhile there is no pressing need to justify our judgement that what the kids are doing is wrong. According to Little, what we might be seeing directly in this instance is a clear instantiation of the concept ‘cruel.’ If we are not making an inference, moreover, then, according to Little, we have mastered the concept. Such mastery, on her account is the product of experience with instantiations of the concept rather than some explicit (or even implicit) understanding of a set of necessary and sufficient conditions for that instantiation. If we did link the moral and non-moral in this fashion, then it would seem that in moral deliberation we are first aware of certain non-moral properties, such as physical descriptions of behaviour, that are linked in a law-like fashion to moral concepts. Instead, on Little’s view, it seems that our first act of awareness is of a situation with a distinct moral shape. What Little sees happening here could be similar to what occurs when we see a painting. On seeing a painting of a mountain, we do not initially see various splatters of oil on a canvas, and from patterns in the globs of paint, infer that a mountain is being depicted. Rather, our first act of awareness seems to an awareness of mountains.

Michael Depaul provides an example that might give a sense of how one might gain mastery of a moral concept by learning to recognise instantiations. According to Depaul he once had great difficulty telling whether a horse had a right or a left lead. A lead is the leg that a horse keeps forward while it canters. Although Depaul maintains that he had a theoretical understanding of the concept, he simply could not see a horse’s lead. “Along thunders the beast, legs going every which way and it was just hard to sort it all out,” he explains. With practice, however, Depaul eventually became as proficient at spotting a horse’s lead as he was at seeing the difference between a trot and a canter (Depaul, 1988, 562).

It seems however, that there is a difference between a moral concept and a concept such as a horse’s lead. This is because with the latter type of concept a physical description of an action exhausts the concept. The concept cruel, however, seems to involve more than a description of directly observable features that constitute a given act. Setting fire to a cat, for example, is not necessarily a cruel action. Take an animal suffering from painful and highly contagious flesh-eating disease. Of course there are no available means for rendering the animal unconscious or otherwise anaesthetised[6] and the only possible chance for its survival is to burn the pathogen off the animal’s body for thirty seconds. Otherwise the animal will suffer through days of agony while posing great risk to other animals and people. Under such circumstances it seems counter-intuitive to maintain that the one who strikes the match and throws it on the cat has performed a cruel deed.[7]

From this perhaps one might want to say that cruelty involves harming an unwilling being unnecessarily. For in the example noted above, there were good reasons for setting fire to the animal, including a concern for the animal’s overall welfare. While this analysis, I will suggest, is too broad and only approaches a full explication of ‘cruel,’ it is worth noting that even at this stage the concept already requires we know something that we can not observe directly. This is because on this account to see that someone is behaving cruelly we must know that they lack any good reason for causing harm to another. Reasons, however, are not the sort of things that can be observed directly, but must be inferred. If we look at Harman’s case we can see that given the description and our background knowledge it is highly improbable that the kids have any good reason for setting fire to cat. There is a limited range of possible reasons that a group of juvenile trouble-makers could have for such a horrifying act, and it is highly unlikely that any of these include a desire for the animal’s welfare, or anyone else’s.

If causing unnecessary harm to another were a complete analysis of the concept cruel, however, then it is difficult to say why Typhoid Mary is not typically characterised as behaving cruelly when she prepared and served meals to her employers. Given that she worked as a cook and was an asymptotic carrier of typhoid, it would appear that her actions regularly caused unnecessary harm to the families she served. Perhaps one should like to argue that because the action of preparing and serving a meal is characteristically a kind act, it is difficult to see it as cruel. However, if we change but one non-observable feature in this scenario, such acts take on a cruel flavour. This can be achieved by specifying that Mary fully understands that her actions will likely cause severe headaches, stomach cramps, vomiting and possibly death in her employers and their children.

This observation, however,  should not be taken to suggest that full knowledge of the consequences are required for an act to be cruel. In Harman’s case it could very well be that the young hoodlums are entirely lacking in empathy and have not even considered the level of suffering they are inflicting for their own perverse pleasure. Under such conditions I still want to say that the boys are behaving cruelly. Largely, I expect, this is because the boys should know that their actions will cause great suffering. If we take a trip to Twin Earth this last claim will become clearer. Although there are no live cats, Twin Earthlings have built millions of robotic pest control devices that look exactly like cats. Moreover, these robots have special power cells that emit heat for hours once the machine is soaked in gasoline and ignited. To stay warm on those cold Twin Earthian nights, then, young homeless hoodlums often burn robotic cats. Unfortunately, Fluffy the cat, who has accompanied us on our travels, is mistaken for one of these robots, caught by the boys, and, to their collective horror, burned alive.

What this more fanciful case has in common with the story of Typhoid Mary, is that in both scenarios, the agents have no reasonable basis for inferring that their actions will harm another. In the real world, however, it is very difficult to comprehend how any normal person could fail to appreciate the consequences that will flow from the hoodlums’ actions as described by Harman.[8] My suggestion then, is that to perform a cruel action, one must have sufficient reason to infer that one’s actions are likely to cause unnecessary harm to another.[9] If this is correct, moreover, then cruelty is importantly different from other act descriptions such as a horse’s lead, a pirouette. This is because the former concept goes beyond a physical description and is grounded by assumptions about the actor’s epistemic situation and reasons for acting. Changing either of these features, moreover, can also change the way we see a given action. Therefore, unlike other actions, understanding a moral concept and visually perceiving an action are not always enough to know that the concept applies. Moral judgements are underdetermined by visual evidence. Seeing that someone lacks a good reason for an action that causes suffering, for example, is not something we can know just from looking, it is something that must be inferred, or grounded on further assumptions.

It is this latter mentalistic aspect of moral concepts that interests me most. Partly this is because other moral concepts are not so dissimilar from cruelty in that they too involve presuppositions about people’s intentional states. Acts that can be called brave, kind or deceitful are all forms of deliberate behaviour. A person will probably not be taken to have acted bravely if we discover that her actions stemmed from the delusional belief that she is impervious to harm, nor can an act be perceived as kind if the actor did not mean to benefit another through her actions. Meanwhile, someone who has mislead us has not been deceitful if this occurs because she comes from a place where nodding is a sign of negation. In fact, it may be that on closer inspection, we might be hard-pressed to suggest a moral concept that we can apply without making at least some assumptions about people’s mental states. This mentalistic dimension of moral concepts is important because they allow for moral vision to be distorted by certain inferential biases. In other words, we may be systematically duped into seeing instantiations of concepts that are not, in fact, there to be seen at all.

Thus far I have been suggesting that Harman’s case seem to involve an instance of clear moral perception. Whether those features I see as constitutive of the concept cruel actually underlie people’s judgements about Harman’s case, however, is open to debate. Some may hold that an implicit recognition of an underlying principle is what causes them to respond to Harman’s case as they do, while others may hold that their judgement was most influenced by empathy for the animal. Few people, I imagine, would want to accept the possibility that an inferential bias might have contributed the strength of their judgement. Yet it may be that in describing his subjects as “young hoodlums” Harman has actually primed his reader to issue the predicate ‘cruel’—everyone knows that kids can be cruel, do they not?

Certain empirical evidence, moreover, supports this suggestion. Many researchers have suggested that people regularly make spontaneous trait inferences (STI’s) based on observations of other people’s behaviour and that this may occur even if we are not aiming to characterise others.[10] Stereotypes, moreover, appear to activate certain traits to be made available to cognition and, interestingly, may even inhibit the inference of traits that run counter to the stereotype.[11] For instance, according to researchers, the description ‘professor’ will tend to activate the trait ‘clever,’ so that when we hear of the professor solving the mystery half-way through the novel, most subjects readily describe the professor as clever. On hearing of a garbage man who solves the mystery early, however, it is less likely that the term ‘clever’ will be inferred, and may even be suppressed. Harman, in fact, uses two loaded terms in his case and it is possible that the presence of stereotypical terms such as ‘young’ and ‘hoodlums,’ cause readers to infer the stereotypic-consistent trait ‘cruel.’ That we are unaware of such processes, moreover, could contribute to the sense that we just see that what the kids are doing is wrong and that no further inferences are necessary.

The suggestion that the description itself helps to inhibit further inferences, moreover, is made more plausible if we look at what happens when we substitute the description ‘hazard control workers’ for ‘young hoodlums.’ On hearing of a group of hazard control workers lighting a cat on fire, many people might at least think to ask whether the cat in the scenario is alive. While this detail is of critical importance, it is merely presumed, not stated in Harman’s case. Yet if the boys were, in fact, setting fire to a cat that was obviously dead, then it is less obvious that their action is immoral.

This observation, however, by no means provides a decisive argument about the factors determining the judgements drawn from Harman’s case. The problem is, we probably do not know what caused us to respond as we did.  Support for this claim can be found in a growing body of research following in the footsteps of Nesbitt and Wilson’s 1977 pantyhose experiment.” [12]  In this experiment, subjects were asked to assess four identical pairs of pantyhose and say which pair they preferred. Overall, people showed a tendency to prefer the right-most pair although experimenters did switch positions of the samples between trials. Subjects also referred to qualities of the pantyhose to explain their choices and showed surprise if the experimenter suggested that placement might be influencing them. Nevertheless, in a culture that reads from left to right, it is probably the case that the right-most pair was examined last.[13]

The point that is emerging from this sort of research is that our own minds are not as transparent as many philosophers might like to believe. It is probably true that the judgement that we are having a certain thought remains incorrigible, but the judgement as to why we are having that thought probably is not. What is even more discomforting is evidence that suggests that when we do not understand our own judgements, we may fabricate explanations, albeit plausible ones given what we do know. A more dramatic example of this comes from the case studies of neuroscientist Michael Gazzinaga, who has worked extensively with split-brain patients. In such patients the corpus callosum is severed, often as a means of treating severe epilepsy. Because of this the flow of information between the left and right hemispheres of the brain is interrupted. When Gazzinaga showed pictures that were visible only to the right brain, the side that is typically not involved in linguistic processing, the left brain, where speech centres are located, explained subsequent behavioural responses via confabulations that were consistent with the left brain’s sphere of knowledge.

In one such instance the left sphere was shown a picture of a chicken claw and the right sphere was shown a shovel. As a result, the patient’s right hand pointed to a picture of a chicken claw, while his left hand pointed to a picture of a shovel. When asked to explain this behaviour, the subject, P.S., said, “oh that’s simple, The chicken claw goes with the chicken and you need a shovel to clean out the chicken shed.”[14]  After observing many similar responses from his patients Gazzinaga came to posit a “left brain interpreter” or a part of the brain that aims to develop “theories that assimilate perceived information [including the actions of it own body] into a comprehensible whole.”[15] In Gazzinaga’s view, moreover, it is unlikely that his research is only significant for split-brain patients. According to him, when it comes to knowing our own minds the only real difference between those with split-brains and most others is that normal people possess only one conscious system they do not understand rather than two.[16]

Given then, that we stand to be mistaken about our own responses, and may even be telling plausible stories about our own inferences, we have all the more reason to be wary of inferential biases that can affect the way we perceive a situation. As John McDowell has argued, “ethical reality is difficult to see clearly. If we are aware of how, for instance, selfish fantasy distorts our vision, we shall not be inclined to be confident that we have got things right.”[17] Distortions in moral vision, however, need not stem from self-interest. As noted, spontaneous trait inferences could conceivably alter the way we view a situation by influencing the traits we suppose give rise to perceived behaviour. Take the wealthy heiress who, without a glance, passes by her housecleaner in the street. It could very well be that people would be more inclined to see this individual’s behaviour as an instance of ‘snubbing’ but will be more reluctant to level a similar charge against a computer nerd behaving in exactly the same manner.

Further information on sources of possible bias come from the significant body of research concerning the attribution of a host of positive traits to individuals deemed attractive.[18] In fact, various studies have shown that good-looking people are typically seen as more honest than homelier people.[19] People also respond more favourably to members of the opposite sex displaying dilated pupils without knowing what exactly has determined their preference.[20] Other experiments, meanwhile, have led researchers to suggest that minimal conditions, such as the arbitrary division of children into teams at a summer camp, can result in the formation of concepts of an in-group and out-groups. In-groups consist of people one readily identifies with, while out-groups represent the ‘others’ as it were. Other research suggests that people are less willing to attribute distinctively human emotions such as love, hope, resentment or contempt, for instance, to members of an out-group.[21]  These human emotions are understood in contradistinction to universal emotions that are applicable to humans and non-humans alike, such as fear, joy and anger.

Since having good reasons for a given action may sometimes mitigate judgements of moral impropriety, it is worrisome to think that judgements about a person’s sincerity might be influenced by superficial features such as physical attractiveness. Who is to say, moreover, how our perception of people’s actions might be influenced by our conception of in-groups and out-groups. We might imagine such factors at play in judging a particular form of punishment to be an instance of “tough love” rather than cruelty. More generally, the interconnected web of beliefs and desires that underlie behaviour is often complicated and it can be difficult to say how various elements of a person’s character will come together to produce a given action. Cognitive processes, some of which may serve us well in times calling for swift critical judgements, can, and probably do on occasion, effectively obscure and oversimplify what should be a process of subtle and complex discernment.

A final source of bias is especially problematic for those who emphasis the value of moral perception due to the efficacy of experience in changing people’s moral outlook. Little, for one, writes that “in real life, moral views are changed by experience and art as much as by argument: someone’s sexist views about women change by fighting alongside them in battle, their views of unfettered capitalism are undermined by working as a night janitor, and their views about eating animals shift by reading a poem”[22] Sometimes it does seem that vivid information, such as, perhaps, the television coverage of the Vietnam War, can have a greater influence on our judgements than theorising about abstract moral principles. Another shift we might imagine occurring is seeing the rightness of social assistance programs after spending one afternoon with a struggling single mother on welfare. Endless arguments about our obligations to those in need or showing how welfare provision raises utility, by comparison, may represent relatively ineffectual forms of persuasion.

This, however, is to introduce an inferential bias that Nesbitt and Ross call the “vividness criterion.” It would appear that the impact of information on the mind is only imperfectly related to its true value as evidence. This is because more vivid information and first-hand experience often carry more epistemic weight then other ‘paler’ forms of evidence such as statistics, or, we might imagine, theoretical arguments. A hackneyed example of this is after seeing a televised plane crash many are prone to judge it safer to drive home for Christmas even though more fatalities occur on highways than in the air. Nesbitt and Ross provide us with another, somewhat more surprising example of the effect of vivid information on reasoning. In one experiment subjects read a case history of a long-term welfare mother living in squalor while raising a host of delinquent children. Alongside this description, participants also read accurate, albeit drier, statistics showing that the average stay on welfare was two years. After their exposure to both types of material subjects were more likely show more negative attitudes towards welfare programs than subjects exposed only to the statistics.[23]  The efficacy of vivid information in affecting attitudes thus stands to cut both ways. Just as a few graphic images might make people recognise the immorality of a war, so could a few beatings at the hands of white kids effectively entrench life-long racial prejudices in a young Chinese boy. This is because the strength of our reactions to first-hand experience is not a result of the representativeness of that experience, but is instead determined by the manner in which the information is transmitted.

In recognising how certain inferential biases stand to colour our perception we might try to guard against such influences. As we have seen, Little holds that good moral perception is largely characterised as what a wise and discerning agent would see. What is becoming clearer is that the morally astute individual is probably as careful in assessing his own responses as his is in characterising other people’s actions. Moreover, we may find that making comparisons across cases is an ideal way to ‘factor out’ responses that are irrelevant to our decisions. Interestingly, comparing cases could even change the way we see a particular case. Eldar Sharif, for instance, has noted that great discrepancies in responses can occur when various cases are evaluated in isolation.[24] For instance, in one experiment subjects were asked to evaluate the level of compensation due to a man injured in a robbery. Different subjects, however, were asked to assess two slightly different scenarios. In one, subjects were told that the man shot was shopping in his regular convenience store. In the other, subjects were informed that the man’s usual store was closed, and that he was shot in a store he rarely frequented. On average, people who read the second scenario in isolation suggested compensation packages worth $100,000 more than packages suggested by subjects who read the first scenario.[25] Of those who read both scenarios, however, 90 percent said that both men deserved equal levels of compensation. It would appear that by comparing the cases subjects were better able to pick out factors irrelevant in determining levels of compensation such that the two cases were seen as equivalent.

If comparing cases is effective in lessening bias in these circumstances, then the same should hold in issuing moral judgements. One way to counteract the effect of vivid information, in fact, could be to insure that contrary views are presented in an equally vivid format. For instance, a more representative scenario could be presented alongside the description of a single mother languishing on welfare before judgements about social assistance are elicited. Sharif’s research, moreover, also suggests that there is significant value in what Dancy calls “switching arguments,” where moral principles and other judgements are tested by altering the context of a given case. Switching, in one’s imagination, a member of an out-group for a member of an in-group, for example, may be a good way to test one’s response to a particular case. The possible utility of such techniques for improving moral perception, moreover, might cause one to better appreciate the notion that good moral judgement is a skill that comes with age and experience. As Aristotle writes regarding phronesis:

[O]ne should pay attention to the undemonstrated assertions and judgements of experienced and older people, or wise ones, no less than to demonstrations; because they have an eye, formed from experience, they see correctly.[26]

With age our store of cases for comparison should increase as past judgements accumulate thereby diminishing the possibility that our perceptions will be distorted by biases. For one thing, an active, alert and involved person should have a greater share of vivid experiences, which if sufficiently varied, will be more representative of reality over time. It is also possible that years of comparing situations might increasingly serve to filter out irrelevant factors that once shaped our perceptions, so that our way of seeing particular actions begin to converge. Perhaps it is just such a process, moreover, that allows a person to extract moral principles from cases she has encountered over the course of her life. I shall leave this final speculation, however, to be developed in a later work.