Zero Degrees of Empathy and Evil

In this posting I want to set out some thoughts about evil and empathy rather than pursuing a specific argument. My starting point is Simon Baron-Cohen’s book (2011, Zero Degrees of Empathy, Allen Lane.) Baron-Cohen defines empathy as follows,

“Empathy is our ability to identify what someone else is thinking or feeling, and to respond to their thoughts and feelings with an appropriate emotion.” (Page 11)

However it might be questioned whether a lack of empathy is really directly connected to evil. Consider the Milgram experiments. These experiments involved the participants in giving what they believed to be very large electric shocks to a learner. In practice there were no real shocks and the learner was really an actor who faked his response. This experiment raises two questions. Firstly were the participants doing something evil? It seems clear to me they were. Secondly did these participants lack empathy? There were 40 volunteer participants in the trial all male whose ages ranged between 20 and 50 and whose jobs ranged from unskilled to professional. Let it be accepted that sociopaths who lack all empathy can commit evil acts. However it seems improbable to me that Milgram could have picked 40 sociopaths for his experiment and that at least some of the participants must have had some degree of empathy. It follows evil acts can be committed by both people who lack and possess empathy.

Baron-Cohen suggests that the extremes of evil are usually relegated to the unanalysable, see page 100. The question immediately arises why do we want to analyse the term evil? Perhaps analysing the term evil might be useful in some contexts? Consider the case of Josef Fritzl who imprisoned and raped his daughter Elisabeth. It seems to me the term evil is useful in this context by simply expressing our moral disgust at Fritzl’s actions. To re-describe Fritzl’s actions as lacking in empathy rather than evil would seem to be much less effective way of expressing our disgust. The term evil simply doesn’t need analysing in this context. However the term evil may have other uses. Baron-Cohen further suggests we can replace the terms ‘evil’ and ‘cruelty’ with the term ‘empathy’ in relation to those with zero negative empathy, see page 65. I will assume here by ‘empathy’ he means ‘lack of empathy’. It seems obvious that simply labelling people like Fritzl simply as evil is not useful in the context of explaining their actions. This context includes both the causes of evil and the way it spreads. Indeed in this context such simple labelling prevents explanation. In the rest of this posting I want to examine the causes of evil and the reasons for its spread. In order to do so I want to differentiate between two categories of people who commit evil acts. Firstly there are people who do evil acts but who do not conceive the acts. They simply propagate these acts. Secondly some people cause evil, originate the evil acts that they and others propagate.

I will consider those who cause, originate, evil first. It seems to me there are three main reasons why people originate evil. Firstly they commit an evil act when pursuing some other goal; for instance the CEO of some company, who knowingly lets his company pollute the environment near to a factory, hereby causing birth defects, in order to maximise profits. Secondly some people simply enjoy the infliction of suffering. Lastly evil is committed to further some greater cause; for instance the purification of the Germanic race or a jihad against the infidels. In this last case the perpetrators of evil may actually see themselves as acting for the best especially if they understand morality in a utilitarian way. None of the above reasons for the origin of evil depend directly on the perpetrators of evil having zero degrees of empathy. The CEO may love his wife and children. The torturer who enjoys suffering may actually need some partial form of empathy with his victim in order to increase his pleasure. He needs the ability to identify what his victim is thinking or feeling even if he lacks an appropriate response. And lastly a Jihadist may have started his Jihad because of his empathy for fellow Moslems. In the light of the above it would appear there is nothing to be gained by simply re-describing people who originate evil as people possessing near zero degrees of empathy rather than evil people. Nonetheless it is true that some people who originate evil acts do have a zero degree of empathy. Does having a zero degree explain the origin of evil in these cases? It would appear not. For as Baron-Cohen points out people with Asperger Syndrome and mild autism often develop a moral code through systemizing rather than empathy and as a result would appear to be no more likely to originate evil than anyone else, see page 84. It would appear having a zero degree of empathy is not useful in explaining why some people originate evil.

Nonetheless Baron-Cohen is correct to connect evil and empathy. A lack of empathy may not originate evil but if someone has sufficient empathy his empathy might act like a vaccine preventing him from originating evil. Sufficient empathy may be very important in controlling what we can think of as permissible. Indeed I would suggest that in the Western World the re-description of the civilian victims of bombing as collateral damage might be seen as an unconscious, at least I hope unconscious, attempt to limit our empathy. After all we describe famine victims as victims and not as the collateral damage caused by drought. It is an interesting question as to what proportion of evil doers are capable of originating evil. Perhaps this question is open to empirical research.

I now want to consider the possible causes for the propagation of evil? I would suggest three main interconnected causes. Firstly it seems clear from the Milgram experiment that respect for authority is a factor. In this experiment the experimenter told the teacher either it was essential to continue, the experiment required that he continue or that he had no choice but to go on. The experimenter would seem to be a figure of some authority. The second cause is linked to the first. Most people seem to find some sort of reassurance that what they are doing the right thing if others behave in a similar fashion. The Herd Instinct seems to confer some sort of legitimacy to their actions. The Herd Instinct means people tend to act as others do and their individuality declines. Lastly if the number of people contributing to an evil act is large enough individuals may not see themselves as really responsible for the act. The trouble is no one else may see themselves as responsible for the act. I argued in my posting of 24/05/11 that in such situations responsibility becomes smeared, see also Parfit (1984, Reasons and Persons, Oxford, page 80).

What then might be done to combat the propagation of evil? I argued above a lack of empathy does not cause someone to propagate evil. In addition I suggested if someone has sufficient empathy his empathy might act like a vaccine preventing him from originating evil. I believe this suggestion also applies to the propagation of evil. Intuitively it seems plausible that increasing someone’s empathy might help him resist both authority and herd pressure in his dealings with others. I am however doubtful that increasing someone’s empathy will make him more aware of his responsibilities. Empathy might be increased in at least three ways. Firstly, John Bowlby’s attachment theory shows that children who experience stable loving parenting are likely to grow up well adjusted to the world. This good adjustment may be partly due to an increased capacity to feel empathy, see my posting of 30/03/09. Attachment theory predicts trans-generational effects so it follows that society should take active steps to enable parents, especially mothers, who experienced attachment problems as children become good parents. Someone might object I am encouraging the nanny state and that parenting is a purely personal matter. Let it be accepted bad parenting harms children. Mill argued “that the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others.” (1859, On Liberty, quote from Penguin Books 1974 page 72.) In the light of the above it would appear that my objector must either accept that the state should be concerned with good parenting or that the state does not have a duty to prevent harm to others. In the real world we of course accept that the state has a duty to prevent harm to children by our acceptance that the state has a duty to ensure children’s education. What steps can the state take to aid good parenting? One way in the UK might be to use the Sure Start scheme. Sure Start was originally conceived as means of teaching poor parents to parent. It has now become popular with parents because it gives them free day care and other benefits. Perhaps it is time Sure Start returned to its original goal of enabling ‘poor parents’ become better parents when ‘poor parent’ simply means one lacking in parenting skills. However not all ‘poor’ parents will use the scheme. Perhaps poor parents might be encouraged to use the scheme by nurses at pre natal checks. Another way of encouraging good parenting might be a campaign similar to the one designed to encourage healthy eating by consuming five portions of fruit or vegetables a day as suggested by Chris Paterson for Centre Forum, see http://www.centreforum.org/assets/pubs/parenting-matters.pdf .

Secondly it is possible that individual empathy flourishes better in less individualistic and more caring societies. Robert Cialdini’s work has shown if we give someone something he is likely to reciprocate, see http://en.wikipedia.org/wiki/Robert_Cialdini. Cialdini’s work is based on giving and not on expressing an emotion. However it seems plausible that the expression of an emotion leads others to reciprocate. Anger leads to anger and perhaps empathy to empathy. It is possible that a more empathic society might increase individual empathy. Someone might object that the above is too simplistic. He might point out there are angry people, empathic people but there are no angry or empathic societies. I accept my objectors point. I would however suggest societies in which people feel at ease are societies which contain less angry and more empathic individuals. I would further suggest people might feel more at ease in caring, giving societies. Lastly in a more speculative vein in the future we might be able to enhance our natural empathy by biomedical means. For instance Paul Zak has shown oxytocin delivered as a nasal spray increases our natural empathy, see http://www.luiss.it/esa2007/programme/papers/3.pdf .

In the light of the above it might appear that we should embrace all three of these means of increasing our natural empathy in order to combat the spread of evil. However I would urge caution. Firstly caution is needed with all biomedical enhancements. Secondly the above appearance depends on the assumption that increased empathy will reduce evil. However we should consider exactly how empathy might be increased. Empathy might be increased in two ways. Firstly empathy might be increased by increasing the empathy we feel towards those people who we already feel empathic concern for. Secondly it might be increased by increasing the range of people we feel empathic concern for; it might increase our domain of empathic concern. I would suggest simply increasing empathy, whilst desirable, is not always useful in combating the spread of evil. For instance if the empathy Germans felt for Germans under the third Reich had been increased it is extremely doubtful whether this increase would have prevented the Holocaust. It follows if we are serious about using increased empathy to combat the spread of evil that we must pay attention to increasing the domain of empathy in addition to increasing empathy in general.

The second cause of the propagation of evil I mentioned above was the herd instinct. For instance when crossing the road a Leeds University study discovered pedestrians are likely to act as a herd, blindly following other pedestrians, see http://www.physorg.com/news/2010-11-pedestrians-herd-instinct-road.html . It therefore seems quite plausible someone might blindly follow others, the herd, in propagating evil. One way of combating the herd instinct is to encourage individuality. It might then be suggested encouraging empathy damages individuality. Accepting this suggestion gives us a further reason to be cautious about simply increasing our empathy if we are seriously thinking about combating evil. However I would reject such a suggestion. I don’t believe building a more caring society must inevitably damage individuality. Firstly if we attempt to increase empathy by encouraging good parenting it seems likely any increase in empathy will not damage individuality. In what follows I will assume individuality means the capacity to act autonomously. Bowlby’s attachment theory predicts children with firm attachment will be more confident in exploring relationships with others. This confidence should increase rather than decrease their capacity to act autonomously. Moreover this confidence to explore their relationships with others should expand the domain of their empathic concern. The second way in which we might encourage empathy I listed above was to foster more caring and giving societies. I see no reason why increasing empathy by this means should damage individuals capacity for autonomy provided this is done from a position of epistemic humility. I believe truly beneficent care must involve adopting a position of epistemic humility, see my posting of 19/06/08. Building a more caring society from such a position means we must give priority to respecting the autonomy of others and provided we do so we should not damage individuality. Unfortunately even if we increase empathy by building a more caring society does not automatically follow that we increase the domain of our empathic concern, which is what I have suggested, is really needed to combat evil. It seems to me that the way our society is structured may be much more likely to increase empathy than by increasing how much it cares. Lastly it seems it may be possible to enhance our natural empathy by biomedical means. I would argue such enhancement should not be attempted if it damages our capacity for autonomy. In addition there appears to be some evidence from Zak’s experiments that such enhancement may only increase our empathic concern for those we already feel some empathy for, rather than expanding the domain of our empathic concern.

I have considered using increased empathy like a vaccine to combat the spread of evil caused by either deference to authority or an inability to resist the herd instinct. However I have suggested increasing empathy may not increase our sense of responsibility. In addition men as a whole feel less empathy than women. Lastly there are some people are incapable of feeling any empathy at all. In the light of these facts are there then any other means of combating the spread of evil? It is possible that increasing individuality might help in combating the spread of evil. I would suggest seeing ourselves as individuals rather than part of a herd increases our ideas of responsibility. Individuality can only flourish in a tolerant society and it might appear that a tolerant society is simply one accepting Mill’s dictum “that the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others.” However as with the domain of empathic concern mentioned above there might be a problem with the domain of ‘civilised community’. For instance a large number of whites in apartheid South Africa did not regard non whites as part of their ‘civilised community’ and as a result many evil acts were tolerated. It follows simply accepting Mill’s dictum is not enough to ensure toleration or prevent evil. Perhaps however if we accept his dictum should apply to all autonomous creatures we might live in a more tolerant society which encourages individuality hence lessening the propagation of evil.

In conclusion I would suggest the methods discussed above can control the spread of evil acts to some degree. Unfortunately I believe these methods will do less to prevent the occurrence of evil acts such as those committed recently in Norway by Anders Behring Breivik.

Knobe, Erler and our ‘True Self’

In this posting I want to consider the comments by Alexandre Erler in Practical Ethics on an article by Joshua Knobe in the New York Times concerning personal identity see http://blog.practicalethics.ox.ac.uk/2011/06/what-is-my-true-self/#more-1548. Knobe uses the example of the evangelical preacher Mark Pierpont who encourages homosexuals to seek a “cure” for their sexual orientation. Pierpont himself was a homosexual and continues to battle his continuing homosexual urges. Knobe uses him as an example to question what is meant by someone’s ‘true self’. He firstly presents two common concepts of self. The first concept is an unreflective one in which our true self is determined by our nature, our urges. The second concept, as advocated by such as David DeGrazia, holds that our true self is defined by our commitments, values, and endorsements. Knobe suggests both of these concepts are challenged by the case of Pierpont. He further suggests that people regard the traits they value in someone as part of that person’s true self. In this posting I want to question whether the concept of someone’s true self as opposed to simply self is a useful concept.

To whom might the concept of a true self be useful? Firstly it might be useful to someone deciding what to do. She might ask what her true self would do in this situation when making some important decision. She might ask herself if she is acting authentically; to her own self being true. Secondly another may question what, is someone else’s true self, before ascribing praise or blame and predicting what that person will do. Intuitively it seems to me these two uses of the idea of true self are distinct and that a different concept might be useful in each case. I will deal with the second case first.

Knobe suggests that others regard the traits they value in someone as reflecting her ‘true self’. I would suggest any such concept of ‘true self’ should not help them ascribe praise or blame and it is not very useful in predicting the actions of others. Praise and blame seem to be more naturally connected to someone’s autonomy rather than her ‘true self’. Of course I accept that someone’s ‘true self’ may be connected to her autonomy. Nonetheless I believe the ascription praise and blame depends directly on someone’s autonomy and that there is no need to involve the concept of ‘true self’. It is plausible that understanding a person’s values might be useful in predicting her actions. However if our concept of ‘true self’ is limited to the values we approve of then this concept will be useful in predicting what we think someone should do rather than what she actually will do. For these reasons Knobe’s suggested concept of someone’s ‘true self’ does not seem to me to be a useful one when applied to others. Erler suggests that our ‘true self’ might be a composite of the two concepts Knobe introduced initially; our natural idea of a ‘true self ‘and an idea of a ‘true self’ being defined by our values and commitments. Is then Erler’s suggested composite concept any more useful than that of Knobe in this context? Once again I would suggest the ascription of praise and blame should be concerned with the idea of someone’s autonomy rather than her ‘true self’. Erler’s suggested concept may well be more useful in predicting someone’s actions but I wonder whether there is any useful difference in this particular context between the concepts of self and ‘true self’.

Is then any concept of a ‘true self’ useful to someone making a decision? Most ordinary decisions are largely unreflective and we just act without questioning our motives too much. However for some big decisions such as whether to pursue a particular career or start a family much more reflection is usually involved. Perhaps in these cases we might question what our ‘true self’ would do; what is the authentic thing to do. It seems clear to me if someone accepts her ‘true self’ is defined only by the values others value in her and she acts in accordance with this concept then she isn’t acting authentically. Knobe’s suggestion about true self is not useful in this context either. Let us return to the second concept of ‘true self’ initially introduced by Knobe. I would argue our deeply held commitments, values, and endorsements are what we ‘care about’. According to Harry Frankfurt if someone ‘cares about’ something she identifies herself with what she cares about and makes herself vulnerable to losses and susceptible to benefits depending on whether what she cares about is diminished or enhanced, see (1988, The Importance of What We Care About, Cambridge University Press, page 83.) Prima facie it might be assumed this concept of ‘true self seems to be a useful concept for someone to employ when deciding how to make some big decision and there is no need for Erler’s composite definition. Prior to making that decision they ask themselves what they really ‘care about’. It would further seem when Pierpont encourages homosexuals to seek a “cure” for their sexual orientation he is acting in accordance with this concept of his ‘true self’. He is acting authentically.

I believe it is certainly true people act with respect to what they ‘care about’. This concept of ‘true self’ determines their actions. However even if it is accepted that one’s ‘true self’ determines one’s actions it does not automatically follow that one’s true self is useful in deciding what to do. What someone intends to do and what he ‘cares about’ need not be identical. Frankfurt argues someone might be unable to carry out her intentions. She might discover, when the chips are down, that she simply cannot bring herself to pursue the course of action which she has decided upon (page 85). For instance a single mother might intend to have her baby adopted. She might believe he would have a better life if this was done. She might further believe this is what she ‘cares about’. However when the time comes for the adoption she finds she cannot go through with the adoption; it is not what she really ‘cares about’. The reason for the mother’s false belief is the way Frankfurt links ‘caring about’ to wholeheartedness and satisfaction. Frankfurt defines satisfaction as an absence of restlessness or resistance. A satisfied person may be willing to accept a change in her condition, but she has no active interest in bringing about a change (1999, Necessity, Volition, and Love, Cambridge University Press, page 103). Our mother may not be able to accurately predict what will satisfy her prior to her actually having to hand over her child for adoption. It follows the concept of our true self defined by our deeply held commitments, values, and endorsements may not be as useful in making big decisions as I assumed above. It might of course be that I am wrong to equate our deeply held commitments, values, and endorsements with Frankfurt’s ideas on ‘caring about’ or that I have misinterpreted Frankfurt’s ideas on wholeheartedness and satisfaction. Nevertheless it seems to me in making a big decision someone should consider what she ‘cares about’ in conjunction with her nature and urges. It follows Erler’s composite idea of true self’, may be more useful than a concept based purely on ‘caring about’ when making major decisions.

Sexually Coercive Offers

James Rocha asks what is wrong with the following coercive offer.

“Hal is Vera’s supervisor at food services company which is expanding into the global market. The company decides to staff its international offices with workers from the US. Hal must send one of his employees to the new Paris or Bucharest office. Vera, while happy to accept a new foreign assignment with much higher pay, would much prefer Paris. Unfortunately the company has randomly assigned her to Bucharest. Hal, knowing the contents and strength of Vera’s preferences, offers to change her to Paris in exchange for sex. If Vera refuses, she will simply be assigned to Bucharest, which has the benefit not only of higher pay, but gets her away from Hal” (2011, The Sexual Harassment Coercive Offer, Journal of Applied Philosophy, 28(2)).

Rocha connects the wrongness of Hal’s action to his disrespect for Vera’s autonomy. He states it is possible to respect an agent’s autonomy whilst changing her actions to a more preferable autonomous action, page 206. I have argued in this blog that autonomy is not simply about choices. Autonomy concerns what the agent cares about or values. I have also argued respecting autonomy means simply respecting the choices autonomous agents make about provided these choices do not harm others. Rocha argues Hal disrespects Vera’s autonomy by inserting influence over her sexuality standards, page 210. Caring about something means you identify yourself with that thing and that this identification must have persistence. It seems to me provided Vera cares about her sexual standards Hal’s offer is unlikely to influence these standards. It follows provided Hal accepts any decision Vera makes he does not disrupt her autonomous decision making.

Nevertheless there seems to be something morally wrong about Hal’s offer. Before continuing to consider disrespecting autonomy I will briefly point out two of these. Firstly Hal seems to have no natural sympathy or empathy for Vera. Slote defines a morally wrong action as one that reflects or exhibits or expresses an absence (or lack) of a fully developed empathic concern for (caring about) others on behalf of the agent. If Hal felt empathic concern for Vera he might offer her the Paris posting unconditionally. Secondly a virtue ethicist might point out Hal’s proposed offer is simply not one a virtuous man would make.

Rocha argues what is disrespectful in connection with autonomy in Hal’s offer is that it seeks to alter Vera’s ends. I would agree with Rocha that Hal’s offer is disrespectful to Vera’s autonomy but would argue this disrespect is not primarily connected with Hal seeking to alter Vera’s ends; it is connected to Hal failing see Vera as an end in herself. Hal sees Vera primarily as a means to his own sexual gratification. We respect autonomy because it has value. Autonomy has both instrumental and intrinsic value. According to Dworkin,

“there is a value connected with being self-determining that is not a mater of either of bringing about good results or the pleasure of the process itself. This is the intrinsic desirability of exercising the capacity for self-determination. We desire to be recognized as the kind of creature capable of determining our own destiny.” (1988, The Theory and Practice of Autonomy. Cambridge University Press page 112)

I have argued Hal’s offer is unlikely to disrupt Vera’s decision making. It follows in practice Hal’s offer is unlikely to disrupt the instrumental value of Vera’s autonomy. Perhaps then Hal’s offer does not disrespect the instrumental value of Vera’s autonomy. However Hal by making his offer fails to see Vera as the kind of creature who can fully determine her own future. Hal’s offer means he disrespects the intrinsic value of Vera’s autonomy and as a consequence fails to truly respect her as a person. I would suggest we value being truly respected as a person more than we desire good options to choose from. My suggestion is open to empirical investigation. However provided my suggestion is correct then the real harm Hal’s offer does to Vera’s autonomy is that he fails to respect its intrinsic value.

Roboethics and Autonomy

In this posting I want to examine roboethics and responsibility. I will do this by considering autonomy. The idea of roboethics seems to have originated with Asimov’s three laws of robotics which he proposed in 1942. In this posting I will consider a revised draft of these laws based on an EPSRC/AHRC initiative in roboethics, see Alan Winfield . These revised Laws are,

1.      Robots are multi-use tools. Robots should not be designed solely or primarily to kill or harm humans, except in the interests of national security.

2.      Humans, not robots, are responsible agents. Robots should be designed & operated as far as is practicable to comply with existing laws and fundamental rights & freedoms, including privacy.

3.      Robots are products. They should be designed using processes which assure their safety and security.

4.      Robots are manufactured artefacts. They should not be designed in a deceptive way to exploit vulnerable users; instead their machine nature should be transparent.

5.      The person with legal responsibility for a robot should be attributed, (with responsibility?)

I agree that humans rather than robots are responsible agents. I also believe robots such as those working on car production lines are just machines like washing machines. However I am going to argue that advances in robotics are likely to smear the attribution of responsibility for events that happen due to a robot’s actions.

Let it be assumed all possible robots are used in a similar way to those working on car production lines. If a machine cause’s damage due to a design fault then the manufacturer is at fault, for instance Toyota admitted responsibility for faults with some of its cars. If however a machine causes damage due to way it is operated then the user is responsible. A driver of a Toyota car who crashes due to excessive speed is solely responsible for the crash provided there are no mechanical faults or other extenuating factors. The legal responsibility attached to the use of robots should be the same as the legal responsibility attached to the use of any other machine such as cars, washing machines or lawn mowers. I would suggest accepting my assumption means accepting there is no real need for roboethics. Roboethics can be safely confined to the realm of sci-fi and is of no practical importance.

However I believe my assumption is false. Robots are indeed instruments we use to obtain our goals as is any machine. But the operations of a robot need not be the same as any other machine. If I use a knife or washing machine the operation of knife or washing machine is completely determined by me. The goals of a robot are completely determined but the way it operates need not be. For instance a driverless train may be classed as a robot if it has no human on board even to act as a controller in an emergency such as those on the Copenhagen Metro. A computer program that automatically trades on the stock market might also be classed as a (ro)bot provided we think the terms instrument and machine are interchangeable in some contexts. If it is accepted that (ro)bots can operate in ways that are not completely determined by the designer or operator then there is indeed a need for roboethics.

Let us consider a (ro)bot that independently trades in stocks for a firm of stockbrokers. Let it be assumed for the sake of simplicity that the goal set by the roboticist who designed this machine is only to maximise profits for the stockholders. Let it be further assumed this machine buys stocks in one company for only some of its stockholders inflating the share price. It then sells the shares other stockholders hold in this company at this inflated price. This sale causes the price to return to its former level. Let it be assumed such a strategy maximises profits for stockholders in general, but that some stockholders gain at the expense of others. The actions the machine takes are unethical. Clearly this machine contravenes the 4^th law of robotics as set out above. This example raises two questions. Firstly what can be done to prevent the contravention occurring again and who is responsible for this.

It seems to me there are two possible ways of ensuring the actions a machine takes are ethical. Firstly the goals of the machine must be adequately specified in order to prevent it producing unethical actions. Secondly if the machine can operate in a way not completely determined by the designer or operator then the machine must have some inbuilt standards against which it can check the morality of any its proposed actions. Let us assume goals of a machine can adequately specified so as to prevent it producing immoral actions. Let us consider this assumption in conjunction with my example of a stock trading machine. The 2^nd law of robotics states that humans rather than robots are responsible agents. In my example I would worry that even though humans are responsible for the machines goals it is by no means clear who should be held accountable for any failings, responsibility becomes smeared. In my example the machine’s designer is unlikely to be fully aware of all the possibilities of stock trading whilst the machine’s owners may be unaware of how to build goals into a machine. Someone might object my worry is illusory and that the stockbrokers involved must be able to fully set out the machine’s goals to the designer. However I have other important worries. Firstly I worry whether, as the tasks machines become able to do become more complex, it is possible to completely specify the goals of a machine so none of its actions can be considered as immoral. Secondly I worry that the ends a designer sets a machine means the means it takes to achieve these ends is always justified. Because of these worries I would suggest instead of attempting to prevent a machine producing immoral actions by adequately framing the machine’s goals that it would be preferable to have a machine with some inbuilt standards against which it can check the morality of doing any its proposed actions.

Someone might object if my suggestion is adopted that the 2^nd law of robotics is contravened. She might point out I am proposing giving autonomy, albeit limited, to robots and with that autonomy comes responsibility. I will now argue my suggestion means giving some form of self-government to robots but it does not mean giving them even limited autonomy. Self-government in this context simply means having the capacity to independently decide on and follow some course of action. A child can be self-governing in a domain specified by her parents. A robot can also be self governing in some limited domain specified by its operators, for example a planetary rover. However self-government as defined above does not mean being autonomous. I have argued repeatedly in this blog autonomy requires that an agent must ‘care about’ or love something. It seems extremely doubtful to me whether robots will ever ‘care about’ or love anything. It is certain present robots day cannot care. It follows it must be doubtful whether robots can be autonomous. It further follows even if a robot has some limited self government it cannot be held responsible for its actions if responsibility requires autonomy. It still further follows that my suggestion, that (ro)bots have inbuilt standards against which they can check the morality of doing any some proposed actions, does not contravene the 2^nd law of robotics.

Nevertheless my suggestion does have worrying implications for responsibility. Only humans can be held responsible for the actions of (ro)bots but adopting my suggestion would mean this responsibility becomes smeared. At the present the Military is showing great interest in military robots, see (1). Let it be assumed such a robot kills a large number of non-combatants in error. Let it be further assumed this machine had inbuilt into its software a set of moral standards. The outcry would be enormous and there would be demands that someone should be held responsible. Clearly as I have argued the robot cannot be held responsible. Who then should be responsible, the operator, the software designer or the ethicist who drew up the moral standards? I would argue no one is clearly responsible as responsibility has become smeared between a large numbers of people due to the way the machine operates. Because of this smearing of responsibility it might be argued no (ro)bots should ever be made which operate with an inbuilt set of moral standards. However I would argue this smearing of responsibility is not a problem peculiar to robotics but is an inevitable consequence of the complexity of modern life. Let us assume a pilot in a plane over Libya fires a missile killing a large number of non-combatants in error. Once again who is responsible, the pilot, his commanders, or the systems designers? Once again responsibility appears to become smeared. If it is accepted this smearing of responsibility is inevitable due to the complexity of modern life rather than being peculiar to robotics then I see no reason why (ro)bots should not be built with an inbuilt set of moral standards provided this is done with great care. I would further suggest that both Philosophy and the Law need to seriously consider the consequences of this smearing of responsibility.

1. Wendell Wallach, Colin Allen, 2009, Moral Machines, Oxford University Press, page 20.

Medical Ethics and Thoroughgoing Autonomy

In a posting on Practical Ethics Charles Foster defends dignity against “the allegation that dignity is hopelessly amorphous; feel-good philosophical window-dressing; the name we give to whatever principle gives us the answer to a bioethical conundrum that we think is right”. He suggests this allegation usually comes from “the thoroughgoing autonomists – people who think that autonomy is the only principle we need”. He further suggests “there aren’t many of them in academic ethics, but there are lots of them in the ranks of the professional guideline drafters”, see http://blog.practicalethics.ox.ac.uk/2011/03/autonomy-amorphous-or-just-impossible/#more-1320 . Foster believes “we need urgently to disown the monolithic fundamentalism of pop-ethics, and embrace a truly liberal pluralism that listens respectfully to the voices of many principles. Proper pluralism isn’t incoherence.” In this posting I want to defend a limited thoroughgoing autonomist’s position.

Before making my defence I must make clear the position I am defending. I do not believe autonomy is the only principle relevant to medical ethics. Even if some form of medical ethics could be based solely on patient autonomy, which I doubt, the ethos of medicine depends on the notion of caring. I accept beneficence is of major importance to medical ethics. Foster characterizes the form of autonomy he is attacking as an icy, unattractive, Millian, absolutist version of autonomy. I do not believe such a form of autonomy should form the basis of most medical ethics. Firstly I will argue a more user friendly concept of autonomy should be applied in medical ethics. I will then argue this concept of autonomy is not subject to the same problems as a more absolutist version. Lastly I will defend the position that respecting the autonomous wishes of a patient should always be given precedence over acting beneficently towards him.

I argued in my posting of 03/09/10 that an autonomous decision is simply a decision with which the agent identifies himself and “cares about”; in this context I am using “cares about” in the same way as Harry Frankfurt, see for instance (1999, Necessity, Volition, and Love, Cambridge University Press). An autonomous decision is one the agent is wholehearted about. Frankfurt argues a wholehearted decision is one with which the agent is satisfied with. He defines satisfaction as an absence of restlessness or any desire for change. If we accept this concept of autonomy then it has consequences for the autonomous decisions patients make. I argued in my posting of 01/07/08 that this form of autonomy is closely connected to the idea of satisficing. It follows patients can make autonomous decisions which are sub-optimal. But don’t some patients such as Jehovah’s Witnesses already make sub-optimal decisions in practice? Moreover a patient may also make an autonomous decision simply to trust his medical team to do what they believe is best for him. In reality I would suggest this is what most patients implicitly do when making a consent decision. This after all is how many of us explicitly make decisions outside a medical context. We simply trust lawyers or financial advisors for instance, without others questioning our autonomy. Are doctors less trustworthy than lawyers or financial advisors?

Foster believes a Millian absolutist version of autonomy causes problems for medical ethics. I agree with Foster. But does a concept of autonomy based on “caring about” cause the same problems? In what follows I will argue it does not. Firstly Foster seems to assume autonomy requires that agents are able to give an unequivocal answer to the question what do you want? He then suggests it is unusual to meet someone who is so well integrated as to be able to do so. I agree many people have difficulty giving an unequivocal answer to the question, what do you want? But how is this difficulty connected to medical ethics? It seems to me Foster must believe that informed consent requires a patient is able to give an unequivocal answer to what he wants. In medical practice a patient is seldom, if ever, asked what treatment he wants. Rather the question usually posed is simply this; “we believe this treatment is in your best interests do you give your consent?” The validity of this consent is of course dependent on the patient being given adequate information about the relevant details of the proposed treatment. If we conceive autonomy as “caring about” linked to satisfaction then autonomous decisions are also not linked to someone being able to give an unequivocal answer to the question what do you want? Autonomy is simply linked to an absence of any desire on the part of a patient to change his decision. Unequivocal answers are only required when a patient doesn’t want some treatment. The above suggests if autonomy is conceived as “caring about” that Foster’s worry about medical practice and respecting autonomy being incompatible because patients cannot always give unequivocal answers when giving informed consent is not justified.

Secondly Foster worries that the giving informed consent in clinical practice is linked to the giving of informed consent in clinical trials. Foster states informed consent requires that a patient in a consultation with a surgeon about his osteoarthritic hip talks in much the same way as a subject in a clinical trial would talk with the trial’s co-ordinator. I once again agree. However I believe if an autonomous decision depends on an agent’s satisfaction with this decision there is no good reason based respect for autonomy why the above linkage should not be broken. For instance I suggested above a patient may make an autonomous decision simply to trust his medical team to do what they believe is in his best interests because doctors are no less trustworthy than lawyers or financial advisors. Different agents may need different amounts of information to make a decision that satisfies them. Moreover an agent might need different amounts of information in order to satisfy him make a decision in different situations. For instance if I am going to have my blood pressure taken all I need to know is I am going to have my blood pressure taken. If I am going to have an operation for my osteoarthritic hip I may need information about the benefits and risks involved. I may only need to understand the risks involved in very broad terms as the pain in my hip means I will discount these to some degree. If however I am consenting to take part in a clinical trial I need to be better informed about any risks involved as I have no factors which will make me discount these risks. In the light of the above I see no reason based on respect for autonomy why the information needed to give consent to treatment should comparable to the information needed to consent to take part in a clinical trial.

Lastly Foster mentions an important paper by Agledahl, Forde and Wifstad (Journal of Medical Ethics 2011; 37) see http://jme.bmj.com/content/37/4/212.full?sid=7a5d7b0f-c5e1-4291-838e-b0d9414fc1d2 . Agledahl, Forde and Wifstad state “patients' right to autonomous choice is upheld as an ideal although the options of both the patients and the doctors are very limited” and then they rightly point out that “in the healthcare setting, choices are often neither explicit nor available.” The implication of the above seems to be that the authors believe a lack of choice means concern for patient autonomy is basically a sham. Let it be accepted all competent patients have some choice. All competent patients can consent or refuse to consent to treatment. I would suggest if autonomy is based on “caring about” linked to satisfaction that in a clinical setting concern for autonomy is not a sham. It is not a sham because a doctor does not have to offer a patient an array of options out of concern for his autonomy. One option is all that is needed. Indeed accepting a concept of autonomy based on “caring about” might mean in a clinical setting too many choices actually erode a patient’s autonomy. Frankfurt argues,
‘For if the restrictions on the choices that a person in a position to make are relaxed too far, he may become, to a greater or lesser degree, disorientated with respect to where his interests and preferences lie.’ (1999, page 109).
In the light of the above I would suggest the fact patients have limited options does not mean it is impossible to respect patient autonomy in practice.

It might be argued the concept of autonomy I have outlined above is an amorphous concept offering little practical guidance. If this is so it might be asked how I am going to defend a thoroughgoing autonomist’s position. Foster argues we need to embrace a truly liberal pluralism that listens respectfully to the voices of many principles. It seems to the soundness of his position depends on what he means by pluralism. I believe if pluralism means some sort of competition between different moral goods the result would be incoherence. I would suggest the only way to avoid this incoherence is to give priority to some moral goods. This prioritization does not imply we must be able to weigh moral goods but that we must be able to rank them. I will now argue this prioritization means giving precedence to respecting autonomy over acting beneficently. If I am going to act beneficently towards someone I must care for him. The basis of my care may be sympathy or empathy. If my beneficence is based on sympathy it seems clear to me that I may act in what I conceive to be his best interests and override his autonomy. If however my beneficence is based on empathy this option is not open to me. If I feel empathy for someone I must focus on what he cares about rather than what I think might be in his best interests. It follows if I want to act beneficently towards someone, and my beneficence is based on empathic concern rather than sympathy and I believe his best interests clash with his autonomy, I should nevertheless give precedence to respecting his autonomy over acting in these interests. Beneficence based on empathy automatically gives precedence to respecting autonomous decisions. It follows if beneficence is based on empathy it is possible to defend a thoroughgoing autonomist’s position. Agledahl, Forde and Wifstad seem to partially support my position because they believe “the right to refuse treatment is fundamental and important”. If autonomy is based on “caring about” and beneficent care based on empathy then a limited thoroughgoing autonomist’s position means doctors need not concern themselves to much with providing choices but must respect all autonomous decisions. However an objector might point out I have provided no reason as to why beneficence should be based on empathy rather than sympathy. We can act beneficently towards animals. Clearly such beneficence is based on sympathy. I would suggest we cannot feel empathy for animals due to epistemic ignorance. I would further suggest we can act more beneficently towards people if our concern is empathic rather than sympathetic. It follows good beneficent care is based on empathic concern. Accepting my suggestions means medical ethics should be prepared to accept a limited thoroughgoing autonomist’s position. Indeed I would argue such a position concurs very well with the practice of medicine with the exception that accepting such a position would mean accepting a fully autonomous decisions by a patient simply to trust his clinicians as to which is the best course of treatment for him.

Brooks, Brudner and Justice

In my previous posting I suggested that any society which insists on compulsory beneficence or the compulsory acceptance of beneficent care is not a truly flourishing society. I further suggested that the only requirement necessary for a flourishing society is that each should be free to do as seems best for herself provided by so doing she does not harm others. This might be classed as the Millian position. In this posting I want to examine what this suggestion means for the basis of law. My examination will consider the position of Alan Brudner, (2009, Punishment and Freedom: A Liberal Theory of Penal Justice, Oxford) and a critique of that position by Thom Brooks (2011, Autonomy, Freedom, and Punishment, Jurisprudence, Forthcoming available at SSRN: http://ssrn.com/abstract=1791538 ).

Brudner believes we should punish others when they damage or threaten to damage our autonomy rather than because of any harm they cause us. Of course if we damage or threaten someone’s autonomy we harm them. In the light of the above I would suggest Brudner’s position might be described as follows. We should punish others only because of the harm they do to our autonomy and not because of any other harm. It is this re-described position I will consider in this posting even though it is by no means certain Brudner would entirely agree with my re-description. I have suggested the only requirement necessary for a truly flourishing society is that each of us should be free to do as she sees best provided by so doing she does not harm others. In the light of this suggestion it might appear to follow that I believe the basis of the law should be harm to others rather than restricted to harm an individual’s autonomy. This appearance is deceptive. Any harm to an autonomous agent always also harms her autonomy. If I harm an autonomous agent I do something to her which she would not do to herself. My actions prevent the agent from making a choice she would identify herself with or which satisfies her. My actions damage her autonomy. It follows all actions that harm autonomous agents also do harm to their autonomy.

Brooks believes Brudner’s position is unconvincing in some contexts. He points out “that there may be many actions we may want criminalized (e.g., traffic offences) that would not be clearly included” in any criminalization as defined by Brudner. Brooks also points out suicide, which ends the victim’s autonomy, and moderate alcohol use, which temporarily affects the user’s autonomy, both of which we would intuitively not want to be criminalized, might well be criminalized provided criminalization is based on harm to an agent’s autonomy. Someone might argue some traffic offences such as speeding have the potential to damage the autonomy of another and hence their criminalization can be justified on Brudner’s account. She might then proceed to point out it is possible to respect autonomy in different ways. Firstly it is possible to respect someone’s autonomous decisions and secondly to protect her capacity to make autonomous decisions. Let us assume a patient suffers from some incurable disease and makes an autonomous decision to commit suicide. If criminalization is justified by damage done by not respecting autonomous decisions rather than harm done to an agent’s capacity to make autonomous decisions then suicide and the moderate use of alcohol should not be criminalized. If criminalization is justified by damage to an agent’s capacity to make autonomous decisions then these actions should be prohibited. How do we choose between respecting autonomous decisions and protecting someone’s capacity to make these decisions? I have argued in previous postings respect for autonomy is intimately tied to respect for persons and it is impossible to respect persons if one doesn’t respect their decisions, see for instance http://woolerscottus.blogspot.com/2008_08_01_archive.html . If my arguments are accepted then it follows that whilst we must respect someone’s autonomous decisions and protect her capacity to make these decisions that when these two ways of respecting autonomy are incompatible we should give precedence to respecting autonomous decisions. It further follows Brudner’s justification of criminalization may be able to account for the cases Brooks raises.

Brooks states that he strongly agrees with Brudner’s focus on offering a political, not a moral, theory of punishment. I’m by no means sure this focus on a purely political basis is possible. It seems to me if we accept Brudner’s position, that the legality of some action depends on whether it harms or threatens to harm our autonomy, we must also implicitly accept that the domain of legal responsibility is the same as the domain of legal concern. Accepting the above means we must accept only all autonomous agents can be held legally responsible and it also means we must accept the domain of legal concern only extends as far as these agents. Accepting the domain of legal concern only extends as far as autonomous agents seems problematic. Clearly children are not fully autonomous agents and equally clearly a child is of legal concern even though she cannot be held to be legally responsible for her actions. In the case of children this problem might be overcome in the same way as traffic offences above. Brudner’s position might be slightly modified so that the legality of some action depends on whether it harms or threatens to harm our potential autonomy. However even if this modification is accepted a problem remains. Let us assume that animals cannot be autonomous. Clearly if a farmer or pet owner lets her animals starve to death the fate of these animals should be of legal concern. Even more clearly severely handicapped children and old people suffering from dementia should be of legal concern even though these people will never be autonomous.

In the light of the above it seems to me that the legality of an action cannot be based solely on political concerns about autonomy. Let it be accepted the only legal limitations a truly flourishing political society may impose on an agent’s autonomous decisions must be to limit the harm her decisions do to others. However what counts as harm is based sometimes depends on moral concerns completely unconnected to concerns about autonomy. Depending on the moral interpretation of what counts as harm the constraint of harm on the unbridled exercise of autonomy may vary greatly.

Why a Truly Flourishing Society must give Priority to respecting Autonomous Decisions over Beneficent Care

In my previous posting I concluded that the only belief all must share in a flourishing society is that each of us should be free to do as seems best to himself provided by so doing he does not harm others. Some might find such a conclusion absurd. Surely they would argue a truly flourishing society must foster compassion and aid it’s less fortunate members. Intuitively if someone is about to commit suicide we should rush to stop him. However if my conclusion is accepted it casts doubts on this intuition. Prima facie my detractors appear to be right. Nonetheless in this posting I want to defend my conclusion.

Before defending my position I must make this position clearer. Firstly this freedom should be restricted to sane adults. I believe any adult should be presumed to be sane unless he can be shown to be insane. In addition I believe the grounds of insanity cannot be based on the outcome of someone’s decision. The fact we believe someone to have made a bad or irrational decision is not grounds for even ascribing temporary insanity to him. I do accept that the enormity of some decisions might render someone temporarily insane due to some connected factors; fear for instance. Nevertheless I believe any decision about someone’s sanity must be based on such factors and not the actual outcome of his decision. In the light of the above how should we react if we see someone about to commit suicide? In such a situation speed is usually essential and we do not have time to assess whether there are any factors that render our victim temporarily insane. For this reason we should usually act to save him. However if we not under the pressure of time and are aware that our potential suicide has no factors that would render him insane then we should not try to save him even if he would have been grateful later if we had. For instance I would argue we should not attempt to stop someone who suffers, from a chronic but not necessarily a terminal illness, from committing suicide. The above does not mean we should not care about the victim, I do. However I believe the freedom of a sane adult to do as he seems fit must be given priority to acting beneficently towards others.

My reason for believing we must accept all the decisions of most adults, even if we feel these decisions may be harmful to them when these decisions won’t harm others, is to respect autonomy. In what follows a freely made decision refers to a freely made decision that may be harmful to the agent and his decision won’t harm others. I would argue if we value autonomy we should accept the freely made decisions of others in order protect this autonomy. My objector might immediately raise two objections to the above. My objector might agree with me that we should protect our autonomy and that this might be achieved by accepting all our freely made decisions but he might then suggest this could be done more effectively by accepting only autonomous decisions. I believe his suggestion to be impractical. In everyday life it is simple to recognise a freely made decision. It is not simple to recognise an autonomous decision. Let us assume we respect all freely made decisions which will not harm the agent or others. But let us also assume in a situation which we think the agent’s decision may harm him we will only respect his decision if it is autonomous. In this situation if we are to act in caring way, be sympathetic or compassionate towards someone must we first go through something like the informed consent process to protect his autonomy? Caring is a disposition people have. This disposition is natural though I believe it can also be cultivated. The basis of caring is not a cognitive activity and applying an informed consent process in a caring context seems inappropriate. However my objector might raise a much more serious objection. He might agree autonomy should be protected but argue that I am wrong always to give priority to respecting autonomous decisions over caring. He might point out some autonomous decisions are trivial compared to someone’s real needs. Surely in such situations he might argue priority should be given to caring for someone over respecting his autonomous decision.

In order to answer this second objection we must ask why we value autonomy. It seems to me we should respect all autonomous decisions because we value respecting people. People are distinct persons. I would suggest if we fail give precedence to respecting the autonomous decisions of others over acting beneficently towards them that we fail to respect them as the kind of creatures who can make their own decisions, respect them as persons. My objector might argue that failing to respect all autonomous decisions whilst respecting most autonomous decisions does not mean we fail to respect others as the kind of creatures who can make their own decisions. He might further argue that those decisions, not respected for beneficent reasons, do usually flow from the agent’s real self. In reply to his second argument I would simply follow Berlin by pointing out what is important to us is not some idealised real self but my empirical or actual self (1969, Four essays on liberty, The Clarendon Press page 132). I would point out that failing to respect all autonomous decisions whilst respecting most autonomous decisions means we only respect other people as the kind of creatures who can make most, or even only some, of their own decisions. I’m by no means sure this is showing true respect for persons. Nevertheless I am prepared to concede for the sake of argument that it remains possible that respecting a person might mean respecting most of his decisions and acting a caring way towards him rather than respecting all the decisions he makes.

I will now consider a second reason why we should respect autonomous decisions. Decisions made at random or which are coerced in some way are not moral decisions. For all moral decisions the agent must feel the decision is his own. He cannot be detached or even semi-detached from his decision. It follows it is basic to the idea of morality that a moral agent identifies with his decision. I have argued that any decision an agent cares about, identifies with or is satisfied with is an autonomous decision, see some of my previous postings or Frankfurt (1988, The Importance of What We Care About, Cambridge University Press). It follows if we fail to respect someone’s autonomous decision that we narrow the domain of moral agents. Of course morality is not just concerned with moral agents. Children and animals are of moral concern even though animals and young children are not moral agents. Kant believed the heart of morality depends upon the ability of agents to make autonomous choices. I believe any narrowing of the domain of moral agency damages morality in general. Any system of morality in which the number of creatures of moral concern remains constant but in which the number of moral agents decreases is a weakened system. Any system of morality in which moral agents are only part-time agents, because some of their freely made decisions are constrained, is a damaged system. In the light of the above I would suggest any system of morality which gives precedence to acting beneficently over respecting all autonomous decisions makes moral agents only part-time agents and as a result is a damaged system.

Let it be accepted any flourishing society must be a moral society. I have argued any moral society must be based on respect for autonomy; we cannot have a truly flourishing society without universal acceptance of this fact. Unfortunately some people in a flourishing society might not accept the need to care for others. Equally unfortunately some people might reject beneficent care. We may deplore these facts but nonetheless if we want to live in a flourishing society we must be prepared to accept them. It follows caring about others cannot be a necessary requirement of a truly flourishing society, caring is of course desirable. Indeed a society that insists on compulsory caring for others or the compulsory acceptance of beneficent care is not a truly flourishing society. It further follows that the only requirement necessary for a flourishing society is that each should be free to do as seems best to himself provided by so doing he does not harm others.