An anatomy of moral improvement

We may distinguish three relevant kinds of improvement:

1. Motivational improvement: better motives, character traits and overall motivation to do what is right.

2. Improved insight: better understanding—accessible when decisions are needed—of what is right.

3. Behavioural improvement: greater conformity to appropriate moral norms and therefore a higher frequency of right action.

This classification may be oversimplified. For one thing, ‘internalists’ in moral philosophy maintain that it is impossible to make a judgment about what is right without having at least some motivation to comply with the judgment, suggesting that moral insight and moral motivation are not entirely separable. (For the record, I reject internalism.) Moreover, some philosophers and psychologists believe that emotions play a role in informing us about what's right, again suggesting that insight and motivation are not completely separable.24 This may be so. Despite possible oversimplification, I believe the suggested anatomy of moral improvement will facilitate discussion of the targets of MB and of what we value in moral behaviour.

Motivational improvement and improved insight are conducive to behavioural improvement. Other things being equal, either type of improvement will tend to bring about better behaviour. If someone's moral motivation remains constant as she gains moral insight, she will become more likely to do what is right (assuming she has any motivation to do so) because she's more likely to know what is right and therefore what to do. Conversely, if someone's moral insight remains constant as his moral motivation improves, he will become more likely to do what is right (assuming he has any moral insight at all) because he will be more motivated to do it. A fortiori, the conjunction of motivational improvement and improved insight will be conducive to behavioural improvement. Let us now consider each kind of improvement in greater depth.

Improved insight is highly cognitive and may come about through cognitive enhancement. Knowing what is right, after all, is a kind of knowledge. Here one is reminded of Aristotle's distinction between (1) ‘moral virtues’ (by our modern lights, a combination of moral and prudential virtues) such as courage, temperance, and justice and (2) intellectual virtues such as theoretical wisdom and practical wisdom (Bks. II-IV)25. Moral insight as construed here is closely related to practical wisdom as Aristotle understood it: an intellectual strength that is indispensible for making good choices.

By contrast, motivational improvement is commonly regarded as more conative (pertaining to caring) or affective (pertaining to emotions). This sort of improvement seems more clearly a type of moral improvement than improved insight is. We tend to esteem a high degree of moral motivation. Kant even thought that the good will—the will to do everything in one's power to do what is right—was the only thing in the world that was unconditionally valuable (chap. 1)26. (But Kant also thought that the good will was guided by reason and not by affect or emotion, that reason could generate motivation on its own.) Someone who acts with a perfectly good will in this very demanding sense leaves no room for motivational improvement. Often we act with a relatively good will. Some theorists, however, construe motivational excellence less in terms of a good will than in terms of particular moral virtues such as sympathy and fairness.

Regarding the third kind of moral improvement, I make the following assumption: Behavioural improvement is highly desirable in the interest of making the world a better place and securing better lives for human beings and other sentient beings. It follows that motivational improvement and improved insight, given their conduciveness to behavioural improvement, are at least instrumentally valuable. One or both may also be intrinsically valuable.

Note, finally, that the concept of moral improvement can be understood to apply to an already-existing individual or to choices about which individuals will come into being and be brought to term. Thus, we might characterise the selection of an embryo possessing a gene coding for a disposition to fairness rather than an embryo lacking this gene, or of an embryo lacking a gene associated with antisocial personality disorder rather than an embryo possessing this gene, as an attempt at moral improvement. Here the intended improvement is applied not to a given individual, but to a choice among potential persons or ‘across possible persons.’ The improvement is transindividual rather than intraindividual. A large number of choices of this kind might be regarded as an improvement in some relevant population.

Addressing a challenge about the content of moral improvement

Before we consider the desirability of MB, let us address a sceptical challenge that stems from moral pluralism. People have different views about what a moral life would look like. Some think that morality requires behaviour in accordance with politically conservative values: a premium on individual responsibility and avoidance of ‘free lunches,’ adherence to time-honoured social expectations and moral rules, strong partiality towards members of one's own community, acceptance of a government that mostly leaves free enterprise and family life to their own devices, and possibly religion. Others think that morality requires politically progressive values: compassion for those whose basic needs are not met, acceptance of a government that helps to meet those needs, an impartial and inclusive view of humanity (and maybe animals), rejection of traditions and mores that seem regressive, and a fully voluntary relationship to religion. Among ethical theorists, some are deontologists with significant non-consequentialist commitments, some are consequentialist, and some turn their attention away from right action towards virtue, politics or gender relations. Nor are these differences purely theoretical, for they often translate into different normative positions on such matters as abortion, the death penalty, affirmative action, drug use, euthanasia, warfare, animal usage, and world hunger.

Amid this diversity of views about what morality requires, how can we determine what sort of change should count as a moral improvement? It is not enough to speak of greater conformity with appropriate moral norms, or the insight or motivation that conduces towards such conformity, if we have no idea which moral norms are appropriate. Without knowing the criteria of right (or morally best) action, according to the present challenge, we can't know what counts as greater insight into what is right or what motivational factors would help us do what's right.

This is a significant challenge to any enthusiast of moral enhancement—including traditional moral enhancement. Thus, parents who believe in providing moral instruction to their children face this challenge. One should not inculcate moral values that are wrong, so how can a parent be sure that she or he is justified in providing a particular type of moral instruction? Also facing this challenge are public school teachers who attempt to inculcate in students certain moral virtues such as civility, respect for differences and concern for the poor.

When public schools inculcate values appropriately, they stick to values that are not really debatable among people who genuinely care about morality. So, in a way, does criminal law insofar as it punishes murder, theft and rape but does not punish a failure to contribute to charity or a failure to work hard and cultivate self-reliance. This suggests a way of determining what should count as an enhancement in considering what MB research the state should support and what kinds of MB, if any, the state might encourage or possibly even require: Stick to improvements that represent points of overlapping consensus among competing, reasonable moral perspectives. What do I mean by ‘reasonable’ here? It is very difficult to provide a satisfying, non-question-begging answer, but let me venture a brief reply—while noting, again, that the challenge of determining what counts as a moral improvement confronts traditional means of moral enhancement no less than it confronts MB.

Leading contemporary progressive and conservative visions—ranging from socialism to welfare-state capitalism to moral conservatism to libertarianism—count as reasonable whereas neo-Nazism, apartheid and the Taliban's worldview (at least as regards women's status) do not. Consequentialist, deontological, virtue-based and feminist views that accord persons some sort of moral equality qualify as reasonable; Nietzschean elitism according to which only the most powerful and creative are worthy does not. As far as justified state policy goes, the idea is to locate points of overlapping consensus among reasonable visions.

Now, it is generally accepted that parents enjoy greater latitude than the state in raising their children as they see fit, a prerogative grounded in the importance ascribed to family life and the presumptive freedom from interference that it requires in order to thrive. Consistent with this value, it might make sense to permit parents to adopt more debatable visions of morality—among reasonable alternatives—and therefore of what counts as moral enhancement, and to select forms of MB accordingly. This would be a biological extension of the current parental prerogative to provide moral instruction to their children as parents see fit. Competent adults deciding for themselves could also be accorded wider latitude to determine what should count as a moral enhancement.

Public policy decisions regarding MB should be acceptable from any reasonable view of morality. I suggest that this area of overlapping consensus is fairly broad. That, in large part, is because we can agree on many things that can go wrong with our moral agency: our motives, our insight or our consequent behaviour.^x Thus, our public policies might support research into and possibly—if and when some forms of MB are demonstrably safe and effective and the state is prepared to make them universally available—encourage or even require the use of certain MBs that help to reduce or eliminate any of the following moral defects.

• Antisocial personality disorder, a severe failure of motivation

• Specific forms of evil such as sadism and intrinsic delight in cheating others, another severe failure of motivation

• Lesser forms of moral cynicism that make one more likely than a good person to be corrupted, to cheat on taxes, not to bother to contribute what one agrees is one's fair share, etc—a more ordinary failure of motivation

• Defective empathy as found in persons with narcissistic personality disorder and in others who are very self-absorbed—a failure of insight

• Significant prejudice against the interests of those outside one's group of identification, a failure of insight or motivation

• An inability to focus on unpleasant realities (eg, starving children, the abuse of women, the worst conditions of factory farms) that all reasonable people can agree are morally problematic—a failure of insight

• Weak will or susceptibility to temptation, a failure of motivation

• Impulsivity in relation to violence, a failure of motivation

• Unwillingness to find common ground when failure to compromise is disadvantageous to all, a failure of motivation

• Inability to find creative solutions to difficult problems involving competing interests and values, a failure of insight

• Inability to grasp subtle, complicated details that are of undeniable moral relevance (eg, the ways in which affluent persons benefit economically from the legacies of colonialism and slavery and from current injustices such as treaties with dictators or strongmen who disserve their country-people^xi), a failure of insight

All reasonable persons can agree that the items on this list represent moral defects. Disagreements on the content of moral improvement do not include (reasonable) disagreement on these matters. Thus, public policy could responsibly consider supporting MB that addresses the aforementioned defects.

Meanwhile, competent adults and parents making decisions for children could go somewhat further in the direction of a specific (reasonable) moral vision in selecting forms of MB. For example, progressive individuals might want to select for themselves or their children greater compassion for the needy than some libertarians would accept as optimal. Conservatives might want to select an extraordinary sense of individual responsibility. Both might want to choose MBs that enhance one's disposition to fairness, but they might have different understandings of what fairness requires in some domains (eg, economic justice) even as they agree in other domains (eg, adhering to the terms of an explicit contract that was freely entered into). Such selection of MBs in accordance with particular visions of morality is appropriate so long as the vision of morality is reasonable.

These remarks, I suggest, constitute an adequate general reply to the challenge of moral pluralism.

Setting up the questions

Suppose, somewhat fancifully, that research into particular types of MB demonstrated that they were quite safe and effective. That would eliminate some leading reasons to doubt the value of at least some types of MB. Suppose also that the relevant types of MB were not very expensive and the state made them available to anyone who could not afford them. That would address concerns about distributive justice. Let's also suppose that efforts to improve the efficacy of traditional means of moral enhancement have proved disappointing. Suppose, further, that there are no realistic grounds to fear long term effects of these forms of MB on society (eg, the emergence of a dangerous gulf between the morally enhanced and the unenhanced). I make all these idealising assumptions in order to isolate our question: Would MB, against the background of those assumptions, be morally desirable just in terms of what it involves? In effect, the question is whether there is anything inherently wrong with MB. My thesis is that there is nothing inherently wrong with MB and that, against the background of our idealising assumptions, the relevant types of MB would be desirable. I also contend that in the real world right now, in which those idealising assumptions are not true, the state should support research into MB and allow private companies (properly regulated) to do the same. We should at least learn more about the potential and the risks of MB.

Many scholars doubt what I assert: that there is nothing inherently wrong with MB. Some doubt this on the basis of a conviction that there is something inherently wrong with biomedical enhancement technologies in general. Chief among their objections are the charges that (1) biomedical enhancement is unnatural, (2) use of biomedical enhancements evinces an insufficient appreciation for human ‘giftedness’, and (3) biomedical enhancements pose a threat to personal identity. Elsewhere I have attempted to neutralise these objections.^xii Here I will address a set of concerns that are directed at MB in particular and appeal to the nature and value of human freedom.

Articulating these concerns, John Harris writes: The space between knowing the good and doing the good is a region entirely inhabited by freedom. Knowledge of the good is sufficiency to have stood, but freedom to fall is all. Without the freedom to fall, good cannot be a choice; and freedom disappears and along with it virtue. (p. 104)7

Importantly, this statement pertains to motivational improvement but not improved insight (which assists in ‘knowing the good’). Because Harris is not addressing improved insight, and because it is quite implausible to hold that the latter poses a threat to freedom, I will construe Harris’ argument and similar arguments as directed entirely at motivation-based MB—though I will hereafter omit the qualifier, ‘motivation-based.’ (Certainly, these arguments do not apply to embryo selection, which cannot alter—and therefore cannot affect the freedom of—particular individuals).

According to Harris, MB would eliminate the freedom to choose to do wrong—to ‘fall’—and therefore the freedom to choose rightly. With freedom, he thinks, goes virtue. His position assumes that virtue is intrinsically, not just instrumentally, valuable. Perhaps that is correct.

What, more precisely, is the objection? Harris’ discussion seems to suggest that MB would entirely eliminate freedom. This seems exaggerated, so let's construe the objection as asserting that MB would eliminate some significant measure of our freedom and thereby of virtue—and that this loss would be unacceptable.

How valuable is freedom? Harris asserts that ‘sufficiency to stand is worthless, literally morally bankrupt, without freedom to fall. I, like so many others, would not wish to sacrifice freedom for survival’ (p 110).7 He holds that moral instruction and socialisation are the best means of moral improvement. These traditional means of moral enhancement, the thinking goes, do not significantly diminish an agent's freedom whereas MB would do so. This assumes, plausibly, that the traditional means can improve motivation and not merely insight.

Harris’ discussion provokes two questions. (1) To what extent would MB threaten our freedom, including the freedom to act wrongly? Answering this question requires an exploration of the nature of freedom. (2) How valuable is freedom, including the freedom to do wrong? This question prompts us to consider what we value, or should value, in moral behaviour.

On the nature of freedom

What is human freedom? According to one conception, each human agent is an ‘unmoved mover’ who is capable, in acting freely, of injecting causality into the natural order of causes and events without her action being determined by that causal order. As Chisholm puts it, ‘(i)n doing what we do, we cause certain events to happen, and nothing—or no one—causes us to cause those events to happen.’29 ^xiii A free agent can sometimes exercise her will in such a way that is independent of causal influences that bear upon her. She may be hungry and want to eat, for example, but may decide to fast for some reason that she takes to be compelling—perhaps a belief that she should save the only available food for a hungrier person. Importantly, according to this view, if she freely chooses not to eat, her decision is not causally determined by psychological forces such as a desire to appear admirable, a need to shore up her self-image or even an ambition to live up to a moral ideal. On this view, reasons to act in a particular way and causes of behaviour are distinct sorts of things. The free agent sometimes manages, in effect, to step out of the causal order and act on the basis of a reason that does not cause her choice and action. She is thus an ‘unmoved mover’.

This classical conception of human freedom accommodates the widely held intuition that we cannot be free if anything causes, or determines, us to choose and act as we do. It therefore embraces ‘incompatiblism’—the thesis that freedom and universal causal determinism are incompatible—plus the belief that freedom is a reality. (Other incompatibilists are ‘hard determinists’, holding that there is no such thing as genuine freedom). Many philosophers have doubted that this view is even coherent.^xiv After all, if the agent's choice is not causally determined by anything, as the present view requires of free action, it is debatable whether the choice is more than a random occurrence. Put another way, it is unclear whether the action can be properly said to belong to the agent, to be his, as an action must be in order to manifest his freedom. But we can set such doubts aside. For if freedom involves such radical independence from the causal order, then MB, as part of the causal order, can't touch it. In that case, MB can't possibly constitute a threat to freedom (p 5).31

Suppose instead that we are free in a sense that is compatible with universal causal determinism (as ‘compatibilists’ claim). In this more Humean spirit, we might say, roughly, that we are free when we determine our actions through our own will, or when we are able to do what we want.32 I abstain from eating after dinner, say, not because someone or something is blocking my access to food, nor because anorexia dominates my eating choices, but because I want to maintain a healthful diet and see the avoidance of after-dinner snacks as conducive to this goal. I want to abstain and, in the absence of external or internal constraints, I do what I want. Unless there is a very special story, such as one about hypnosis, it appears that we may judge that I act freely in this case. Is there any reason to think that MB would threaten freedom under this conception?

It would seem not. Consider examples. As a result of MB, I might become (1) stronger-willed and therefore less vulnerable to weakness of will, (2) less prone to acting on violent impulses, (3) more prone to act altruistically or (4) more inclined to act fairly. Let us assume, reasonably, that the effects of MB are not so overpowering as to cause what amounts to temporary paralysis (say, when I feel a violent impulse) or involuntary muscle movements (say, that cause me to help someone). The agent, that is, behaves as an agent: performing intentional actions that are under voluntary control. It is compatible with performing such actions that one is influenced by any number of external or internal influences. Otherwise, since we always act under influences, we would never act freely on the present conception of freedom! So suppose that I render help to someone in need in a case where it is inconvenient to do so yet wrong not to help; and MB made me more inclined to help in such a case. The ‘boost’ provided by MB does not rob my behaviour of freedom any more than the caffeine in my tea robs me of any personal credit for writing this paper. Indeed, my choice to drink tea to enhance my ability to think and write is itself an expression of my agency. The same may be said of moral behaviour that is supported by motivations that are in turn supported by MB. I help the person in need because I want to help that fellow, or because I think I ought to, not because I am forced to do so. Moreover, I sought the motivational assist that I received from MB. There is no threat to freedom here.

‘But wait’, one might interject. ‘You are free in helping that fellow only if you could have not helped him. One acts freely only if one could have acted otherwise.^xv Because of MB's influence, this condition might not have been satisfied here’. For several reasons, I disagree.

First, suppose scepticism were warranted about whether I was free—more plausibly, sufficiently free—on the grounds that MB made me less likely to choose wrongly. If so, then parallel scepticism would be warranted about whether one acts sufficiently freely under the influence of such traditional means of moral enhancement as moral instruction or appropriate socialisation. One might reply that the two cases are not parallel: Traditional means of moral enhancement appeal to reasons for moral conduct whereas MB simply exerts causal influence on my motivations. But this oversimplifies matters. Traditional means often do exert psychological and therefore causal influence, yet unless truly excessive, this influence is compatible with the agent's acting freely. In some cases, it is true, there may be a difference: A relatively unpressured moral education may appeal to reasons, with little or no psychological pressure, whereas a form of MB may directly (causally) affect one's motivation. But, as we have seen, acting freely is compatible with acting under a degree of influence. Moreover, in the case of competent adults deciding to take MB, they choose the means to improved motivation—just as they might choose to go to a consciousness-raising group that applies a certain amount of psychological pressure. Even if some significant causal work is done in either case by something outside of one's agency, one is not a passive subject of this influence but instead actively welcomes it.

On the compatibilist conception of freedom, acting freely requires not perfect independence from causal influence—which is impossible on this conception—but the right sort of relationship between one's preferences and action as well as the absence of certain sorts of freedom-subverting conditions. Elsewhere, I have defended the following analysis of free action (which I called ‘autonomous action’): ‘A autonomously performs intentional action X if and only if (1) A does X because she prefers to do X, (2) A has this preference because she (at least dispositionally) identifies with and prefers to have it, and (3) this identification has not resulted primarily from influences that A would, on careful reflection, consider alienating’. (p 102).28 I maintain that any action that meets these conditions is genuinely free. Yet there is no reason to suppose that the moral behaviour of those who have undergone MB would necessarily or routinely fail these criteria. In the case in which I render assistance where doing so is inconvenient but morally required, I would not consider the influence of MB on my motivation to be alienating. I welcomed this influence.

What about those whose moral behaviour is motivationally influenced by MB that was ‘imposed’ on them by their parents? They, too, have no special reason to consider the influence alienating. After all, it is helping them—without forcing them—to act as they should. This, like moral instruction that comes with a certain amount of psychological pressure, is within the bounds of appropriate parenting (notwithstanding the fact that some people don't want to do what they should). Because the children are not choosing the interventions for themselves, there is a greater possibility of alienation than in the case of consenting adults, but no more than in the case of many unremarkable instances of parents’ ‘imposing’ traditional means of moral enhancement on their children.

My sceptical interlocutor might insist that in some cases of MB an agent would lack ‘freedom to fall.’ This may be true. But it is also true of many everyday behaviours whose moral value we do not question. For example, I am psychologically incapable of stabbing or shooting a person I perceive to be innocent and non-threatening. In the presence of a handy weapon and a potential victim, I rightly abstain from immoral assault—and it seems that I could not act otherwise and commit the assault. But my inability here is not due to squeamishness, for I could probably stab or shoot someone who was trying to kill a loved one or an innocent person if no less violent means of protection were available. My inability to stab or shoot a loved one or innocent person is due to my stable values and preferences. I freely abstain from immoral assault even though I cannot act otherwise. I meet the conditions of autonomy stated above. The same may well be true, depending on details, where I render assistance for moral reasons and am psychologically incapable of acting otherwise. Ability to act otherwise, or freedom to fall, is not a necessary condition of free action.33 ,34

Nevertheless, some conceivable forms of MB would pose a threat to freedom. Imagine a computer chip that could be implanted in someone's brain such that whenever the agent decided to perform a certain kind of immoral action, he would change his mind.^xvi The implant would reliably, automatically redirect his decision-making. It seems plausible to assert that this sort of MB would make an agent unfree with respect to certain actions (even if the agent welcomed the MB), and possibly in motivation-based ways—say, if it made the agent suddenly lose the desire to perform the action in question. But most of the types of MB under consideration, including the examples enumerated earlier, do not have this apparently freedom-robbing quality.

There is nothing about MB in general that poses a significant, across-the-board threat to freedom when it acts on an agent's motivations. Nor, as explained earlier, does MB threaten freedom when it acts on an agent's moral insight. Even more obviously, MB poses no threat to freedom in cases involving embryo selection. Except in extreme cases involving motivational improvement, MB does not pose a significant threat to freedom.

On the value of freedom

Suppose matters were otherwise and MB did pose a significant, across-the-board threat to freedom. What would follow about the desirability of MB? Presumably, freedom has significant value. It doesn't follow that its value overrides everything else that matters.

One can choose and act more or less freely, depending on the extent and source of influences and pressures that bear on one. In some cases, they might be a bit alienating but not so alienating as to make our choice forced or unduly pressured. That is, we act freely—we meet some appropriate threshold to judge that our action is free (enough)—but our degree of freedom is far short of what characterises our actions when we act most freely. I see no reason to preclude tradeoffs between the degree of freedom with which we act and other values.

Suppose, contrary to our discussion in the previous subsection, that MB consistently had the effect of substantially reducing our freedom of action. Imagine that there was a reasonable way of quantifying this effect and that, in comparison with the average degree of freedom with which we act, MB reduced our freedom by 25%. Imagine further that, as a result of MB, there were no more wars or starvation and everyone in the world had access to the basic necessities of life. I, for one, would accept this reduction in freedom across the board if it were necessary for such great improvement in people's moral behaviour and such welcome consequences.

Alternatively, suppose that while most forms of MB pose no significant threat to freedom, a particular one does. A brain-implanted computer chip reliably causes the agent to change his mind whenever he forms an intention to rape someone or molest a child. It robs the agent of 100% of his freedom with respect to rape and child molestation without diminishing his freedom to perform other types of action and without producing any deleterious side effects. In my view, the elimination of freedom with respect to these types of action is no great loss. Indeed, in terms of freedom alone, the trade-off might be worthwhile to the agent in question, by enabling him to avoid the loss of freedom imposed by incarceration.

If you agree with the assertions made in either or both of the two previous paragraphs, then you accept the proposition that maximal freedom of action does not automatically deserve priority over other values. Thus, even if MB significantly reduced our freedom, it would not straightforwardly follow that MB was undesirable. It might yet be desirable in view of its good results. This brings us to a fundamental question.