1. How we go wrong in our moral judgments

In order to act well, we need to know what the appropriate response is within a given situation. Should we honestly tell someone that their painting isn't very good, or should we support their efforts by mentioning the positive aspects of their work and silently passing over the deficits? From Aristotle onwards, virtue ethical accounts have stressed the importance of this skill, termed practical wisdom or phronesis, to moral character and action. Character traits like honesty, kindness and courage are thought to be developed and become stable as a result of the process of habituation and learning (Aristotle, Reference Aristotle and Irwin1985). Persons must learn – typically, through the normal processes of moral development and via trial and error – how to discern which actions are morally appropriate or required in any given situation. As an end result of this learning process, a stable disposition to be, for example, honest may be built up as a result of making appropriately honest choices in different contexts over time, where practical reason guides our assessment of how much information we must divulge to be appropriately honest. A virtuous trait such as honesty becomes habituated when there is a stable disposition set – in this case, where the disposition to be honest is so well-established that effortful use of practical reason to determine the honest action becomes less necessary. Developing virtuous character traits can therefore be understood as similar to developing expertise in some area (e.g., when a long-term car mechanic can very quickly assess problems and solutions regarding car engines), and practical reason is crucial to honing this expertise (Annas, 2003).Footnote ³

It is hardly surprising or controversial that moral character and action requires development, learning, and reasoning. Moral judgment and practical wisdom are complex abilities, and many different psychological features will be involved here. Practical reasoning would seem to include cognitive capacities such as top-down attentional focus, as we must be able to attend to the salient features of a situation to notice a moral problem. For example, only once a lecturer notices that a student's joke is off-colour because it is sexist or racist can she indicate to the student that the joke is inappropriate. Practical reason also requires self-control, as we must set aside impulses to gratify ourselves and instead focus on the features of a moral problem. In addition, in order to make good moral decisions we need to be able to understand the possible outcomes of an action – to return to our example, whether a certain person will benefit from being told the truth and to what extent they may be harmed by having their feelings hurt. In their introduction to Virtue Ethics, Rosalind Hursthouse and Glen Pettigrove argue that in deciding whether to tell the truth, the virtuous agent needs practical wisdom to know when to speak and when to be silent. ‘It is part of practical wisdom to know how to secure real benefits effectively; those who have practical wisdom will not make the mistake of concealing the hurtful truth from the person who really needs to know it in the belief that they are benefiting him.’ (Hursthouse and Pettigrove, Reference Hursthouse and Pettigrove2018, Section 1.2.)

The traits associated with a condition such as autism may impact a person's ability to be virtuous in at least two ways. First, they might affect the development and exercise of a particular character trait, like kindness. For example, problems in gauging others’ thoughts and emotions may affect the development of the trait of kindness due to a difficulty in assessing a person's psychological needs. A well-meaning person who wants to be kind or compassionate needs the practical wisdom to gauge others’ beliefs, feelings and emotions. In a discussion of autism and moral agency, Jeanette Kennett (Reference Kennett2002) presents the case of ‘an autistic teenager with a passion for the piano and perfect pitch [who] suggested a constitutional amendment “to require that every home have a piano with 88 keys, and to require that the piano be kept in tune”’ (p. 351). As an act of kindness, this is well-meant but lacks practical wisdom in that it fails to detect what will actually benefit others. Consequently, some authors argue that problems autistic people have gauging other people's thoughts and emotions constitute a problem for their moral agency (Stout, Reference Stout2016; Shoemaker, Reference Shoemaker2015). However, it should be pointed out that the problem with gauging other people's emotions is a mutual one. Neurotypical individuals often struggle to empathize with autistic individuals and to gauge or understand their emotional reactions. This has become known as the double empathy problem (Milton, Reference Milton2012). Indeed, autistic scholar Robert Chapman (Reference Chapman2020) notes that neurotypicals are often poorly equipped to empathize with autistics because in an ableist society they usually don't have to think about autistics’ needs; whereas autistics work hard to empathize with neurotypicals due to the obvious benefits of doing so.

What the piano example shows is the necessity of taking into account the differences between individuals – in terms of their needs and preferences – in deciding what the morally best action is in a situation. While many responsibility theorists focus on how autistic persons might struggle with this due to difficulties in accessing others’ mental states, a second problem may arise when autistic individuals employ moral rules in an overly rigid fashion. It is plausible that in some cases, difficulties with gauging what other people are thinking is what leads to rigidity and a strong reliance on explicit rules. However, this is unlikely to be the whole story, as rigidity and need for sameness in autistic persons also apply in the non-social domain. For our current purposes, we can think of moral reasoning in autistic individuals as more rule focused and less sensitive to context than that of neurotypicals. To illustrate, consider the case of a young autistic man confronted with the scenario where an unemployed woman with several young children steals a small amount of food in a store. He insists that what she is doing is illegal and that she should therefore be arrested (example from Keel, 1993 cited in McGeer, 2008, p. 240). This is a case where the context of the woman is not taken into account in coming to a moral judgment.

It thus seems that virtue ethicists’ justified emphasis on cognitive flexibility and assessing situations case by case entails a moral weakness associated with autistic traits such as rigidity and need for sameness, as well as difficulty in reading social situations and predicting others’ mental states. To be good moral actors, we do need the ability to pick out the morally salient aspects of new situations, weigh competing moral values against each other, choose amongst moral options, etc. Of course, neurotypicals also frequently fail to exercise practical reason in an appropriate way, so the difference is better conceptualized as one of degree. As Kenneth Richman argues, when neurotypicals are tired, stressed, or even lonely they may ‘…simply not register facts or fail to perceive options even when they are paying appropriate attention’ (Richman, Reference Richman2018, p. 29). That is, neurotypicals also fail to pay attention to a moral problem or to reason regarding others’ emotional needs. In this way, Richman argues, ‘…autists are not different in kind from neurotypical people, just subject to more of the common sorts of moral frailties’ (Richman, Reference Richman2018, p. 29).Footnote ⁴

It should again be emphasized that virtue ethics construes practical wisdom as a goal we strive for that many do not achieve, rather than as the default mode of moral decision-making. Most of us fail to exhibit practical wisdom much of the time; and almost none of us are moral experts across the realms of different moral problems. However, this focus on the importance of flexible moral reasoning fosters an overly narrow and deficit-oriented view of the way in which autism can impact moral reasoning. Rather than emphasising the weaknesses, we should note that autistic individuals’ style of reasoning has costs and benefits in terms of moral decision-making and action.Footnote ⁵ The costs and benefits of moral reasoning and decision-making styles in autistics are different than those of neurotypicals – who also suffer from certain deficits in practical reasoning.

A narrow focus on scenarios where we need a good understanding of others’ inner lives to make the right decision distracts from the fact that all of us face a whole host of different kinds of challenges to moral decision making. Many challenges we face in acting morally are not those where the main difficulty is taking notice of others mental states or figuring out the right thing to do. Rather, neurotypicals often face the moral challenge of keeping the right course of action in mind when there are powerful countervailing factors that make it hard to follow. For example, if we look at the problem of climate change, the moral (and prudential) imperative to prevent global temperatures exceeding 1.5 degrees of pre-industrial levels is clear. And yet people who want to stop climate change still fly, eat meat, and buy new boilers and gas-guzzling SUVs. With the exception of the SUVs, we authors would include ourselves in the list of environmental akratics. Of course, weakness of will is not the only possible explanation for acting inconsistently with our professed values. People may also not care as much as they claim to, or not fully grasp the gravity of the moral situation, for example because they are unrealistically optimistic about likely outcomes. We take these explanations for moral weakness to be compatible and mutually reinforcing.

The moral weaknesses that many neurotypicals display are arguably aided by sensitivity to context and by a strong attunement to one's social environment, the very capacities often posited as necessary to practical wisdom. Furthermore, motivational and contextual factors don't just deter us from doing the right thing, they also affect our perception of what the right thing to do is. How so? Aristotle and neo-Aristotelians stress the importance of habituation and learning from moral exemplars, people who we admire and respect as moral role models. This means that we can be influenced by moral slippage within our moral community, because we calibrate our moral compass against others in our community (Merritt, Reference Merritt2009). Furthermore, the human tendency to self-justify morally problematic behaviour has more scope for coming up with creative self-justifications when we don't rely on explicit moral rules and take the context of a specific moral decision as a reason to justify our behaviour. We may for example say that of course flying is wrong, but in this particular situation, when we are visiting relatives abroad after a long time, it is perfectly justified.

A lack of cognitive flexibility may serve to keep certain moral reasons, such as environmental dangers, in focus in a way that is uncomfortable in light of societal norms but more conducive to consistent action. An anecdotal example is that of an autistic boy who cares a lot about the environment and is seriously upset by the fact that his family will happily consume chocolate which contains palm oil, even though palm oil farming is a major contribution to deforestation. Unlike his family, the boy cannot put this concern to one side when it comes to enjoying the chocolate. This action provides an example of the consistency but also rigidity often associated with autism. One might argue that part of situational flexibility and good moral agency is to know when to make an exception. But, on the flip side, we know that many of us are far too easily swayed to make exceptions. The boy is exhibiting a consistency of reasoning which is just as important to good moral judgment and decision making as situational flexibility is.

With these insights in hand, we turn to the moral strengths associated with autistic traits and survey the growing body of evidence that shows that autistic traits are associated with enhanced rationality in the form of more consistent and less biased reasoning.

2. Cognitive biases and enhanced rationality in autism

Although autistic persons may be less flexible on average, there is a raft of evidence that they are also more rational and less prone to cognitive biases than neurotypicals. Jeanette Kennett argues that rationality is core to autistic moral agency (2002). Kennett's approach is explicitly Kantian and therefore focused on universalizable rules for moral action. But even if we believe practical wisdom and sensitivity to context is necessary to making the right judgment in a given situation, there is a role for autistic traits. This is because autistic traits make individuals less prone to being influenced by irrelevant features of the context that tend to distort moral judgment. In this section, we look at how this plays out at the individual level. In the next section, we turn to benefits at the group level that result from having different cognitive styles.

Humans are famously subject to a large number of cognitive biases. How information is framed influences the way we evaluate it: for example, hearing that a new treatment will prevent 60% of deaths makes people evaluate it more positively than hearing that it cannot prevent 40% of deaths (Kahneman, Reference Kahneman2011). This phenomenon is called the framing effect. People are also unrealistically optimistic; they think that bad events are less likely to happen to them than to others or than is objectively likely (Shepperd et al., Reference Shepperd, Klein, Waters and Weinstein2013; Jefferson, Bortolotti, and Kuzmanovic, Reference Jefferson, Bortolotti and Kuzmanovic2017). These are just two examples; there are many more biases that human cognition is prone to. We cannot review all the forms of irrationality reflected in human psychology here. Let's instead propose a general characterization. Human beings have cognitive biases, which are patterns of thinking where information processing isn't ideally epistemically rational; they are often not sensitive to available information in a way that would lead to the most accurate beliefs and decisions. This can mean certain pieces of information are given too much (or too little) weight, or that motivational factors affect what information is taken into account. It can also mean that one and the same piece of information is processed differently depending on how it is presented, as in the framing case.

Over the last 15 years, researchers have conducted studies that showed enhanced rationality in people with autistic traits. A recent review paper by Rozencrantz and colleagues (Reference Rozenkrantz, D'Mello and Gabrieli2021) compiles the findings on the topic from several studies. According to these studies, individuals with autistic traits are less prone to framing effects, the optimism bias, and other common biases. Furthermore, people with autistic traits rely less heavily on intuition and are more consistent in their choices (i.e., their choices are less influenced by irrelevant contextual factors than those of neurotypicals tend to be (Farmer, Baron-Cohen, and Skylark, Reference Farmer, Baron-Cohen and Skylark2017)).

Here are two examples:

In a 2017 study, Farmer and colleagues tested the susceptibility of people with lower and higher levels of autistic traits to being influenced by contextual information that normally (in the non-clinical population) biases choice in ways that are irrational. The choice situation in these kinds of experiments is as follows: Participants are asked to choose between two options, A and B. In these choice scenarios option A is better than option B in one respect, but option B better than option A in a different one. So, for example, part-time job A might pay $ 6.60 per hour, whereas job B pays $ 8.20 per hour. However, the commuting time for job A is only 20 minutes, whereas the commuting time for job B is 60 minutes. People will choose according to how they evaluate the trade-off.

So far, so good. But if a decoy option is introduced, which makes one option look more attractive by comparison to the decoy, then choice behaviour changes in irrational ways. So, if a third job, C is introduced, where the commuting time is 30 minutes and the salary is $ 6.20 per hour, this influences choices in such a way that option A is preferred over option B because option C makes A look more attractive. While A is clearly more attractive than C, the respective advantages of A and B compared to each other have not changed. Similarly, if a decoy job option is introduced where a job pays $ 7.80 per hour but has a commuting time of 70 minutes, this has the effect of making B more attractive to participants, and B is chosen over A more frequently, even though its respective advantages and disadvantages compared to A have not changed. So in these cases, information that is irrelevant to the choice at hand (that a third option is worse than one of the existing options) influences the perception of attractiveness and the subsequent choice.

These experiments show that context affects our choices in ways that are irrational, because we lose sight of the respective advantages and disadvantages of the two best options. Farmer and colleagues found that this effect is reduced in people with an autism diagnosis compared to those without, and it is also reduced in those who score high on autistic traits (even if they don't have a diagnosis) compared to those who score low.

Similarly, Kuzmanovic and colleagues (Kuzmanovic, Rigoux, and Vogeley, Reference Kuzmanovic, Rigoux and Vogeley2019) compared optimistically biased belief updating of people with and without an autism diagnosis. In line with other studies that have found unrealistically positive expectations for oneself, they found biased belief updating that supports positive expectations for the future in neurotypicals. People were asked to estimate the likelihood of a number of adverse life events (experiencing dementia, getting burgled etc.) and were then given (fictional) population base rates. If these were lower than expected, meaning the bad event was on average less likely than participants had assumed, they updated their predictions accordingly, assuming the adverse life event was less likely to happen to them. However, they updated their predictions far less when they found out that the adverse event was on average more likely to occur than they had assumed (unwelcome news). When making predictions for similar others, this asymmetry in belief updating was not observed. By comparison, people with an autism diagnosis were less likely to update their beliefs in a way that ignored unwelcome news, making their belief updating more rational.

How are these studies relevant to moral decisions by autistic persons? The enhanced rationality associated with autistic traits is not specifically related to moral thinking, and the biases investigated are often highly specific. However, the research on enhanced rationality matters because moral decision-making depends on good practical reasoning skills, as virtue ethicists stress. There is much talk about the importance to reason flexibly and context sensitively, but obviously, practical wisdom requires us to use the context in the right way. It requires the individual not to be swayed by contextual factors that are irrelevant to the morality of an action, and it also requires them to not be overly influenced by motivational aspects such as the desirability of certain beliefs, be they moral or non-moral.

So, for example, when considering the optimism bias, we can see how this might be particularly harmful in making assessments of what action is needed to avert climate change. Of course, it isn't really possible to think that climate change will only happen to others.Footnote ⁶ But it is perfectly possible to rationalise the belief that climate change is less likely to affect oneself and family – to whom one has a heavy moral obligation – and to downplay one's moral obligation to those who are more likely to be affected. It may also be easy to decide that, because the problem of climate change is so large and difficult, one's own ways to contribute to solving the problem are too insignificant to be worth doing; and to note that one's actions seem in line with or endorsed by the actions of others. Here we can see how enhanced consistency and accuracy in evaluating scenarios with moral implications, as well as a lower dependence on contextual cues, can be desirable features in moral thinking and decision making. They prevent people from being distracted by the context, using the context for self-justification, or making decisions based heavily on wishful thinking in the ways we have described.

At this stage, one might object that autistic traits predispose individuals to a moral rigidity that is incompatible with true practical wisdom. Being too situationally flexible is not a good thing, as it comes at the price of consistency, but practical wisdom requires the right amount of flexibility. While many neurotypicals are overly morally flexible or swayed by irrelevant features of the context, autistic individuals err on the side of rigidity. We agree that most individuals, be they autistic or neurotypical, fall short of having fully developed practical wisdom – they fail to become moral experts in at least some moral realms. The boy who gets distraught at the fact that his friends and family are eating chocolate which contains palm oil may well be lacking a sense of perspective. But, so are his friends and family who only attend to the wider implications of their food consumption when this does not interfere with their pleasure. Real agents fall short of practical wisdom in many ways. Again, practical wisdom is an ideal rather than the normal state of adult humans. Some of us with more autistic traits fall short by not readily perceiving reasonable exceptions to moral rules, whereas others are driven by sensitivity to context to make so many exceptions that it is hard to still see that there is a rule.

However, it may still seem that there is a qualitative difference between autistic individuals and neurotypicals, because in principle, neurotypicals could overcome their biases and weakness of will, whereas this is not possible in the same way for autistic people. We note that the extent to which flexibility can be learned by autistic individuals and biases can be overcome by neurotypical ones is in the end an empirical question, so we won't be able to give a definite answer to this question from the armchair. Also, the relevant traits are a matter of degree, so there is unlikely to be a one size fits all generalisation that can be made of autistic and neurotypical people, respectively.

However, there are reasons to doubt that it really is more possible for neurotypicals to avoid cognitive biases than it is for autistics to become sensitive to environmental cues. There is evidence that autistic people can get better at taking relevant factors about context and what matters to others into account when making moral decisions. For example, Shulman et al (2012, p. 1375) noted that autistic persons can benefit from being explicitly taught the principles upon which behavior-governing rules are based, and from practicing transferring these principles from one situation to another. Autistic individuals describe a process of coming to grips with the moral rules and sensitivities in their environment by way of an expanding list of considerations that are relevant to the moral evaluation of a situation (Grandin, Reference Grandin1995).Footnote ⁷ So it's plausible that for autistic persons acquiring context sensitive moral judgment is possible, but that it will happen in different ways than it does for neurotypicals – it may in some cases require more explicit instruction and practice.

3. The benefits of psychological diversity in moral reasoning

In a society where people often fail in moral judgment and action, it is helpful for us to fail in different ways. It is worse for society if we are all too optimistic about morally relevant outcomes; or all find it difficult to flexibly take contextual information into account. It is also helpful for us to morally succeed in different ways: for example, it is better if some of us are quite good at assessing other people's needs, and some of us can offer intense focus on a moral problem. Different societal roles require different moral strengths and can be more or less tolerant of certain moral weaknesses. For example, if a person finds it hard to anticipate and understand the emotions of the average child, they are probably not well suited to being a primary school teacher.

The benefits of diversity in moral reasoning can be seen from other perspectives than through the lens of practical wisdom. In her paper ‘The Advantages of Moral Diversity,’ Amelie Oksenberg Rorty uses the differences in styles of moral reasoning related to different types of ethical theory as an example. Ethical theories tend to give rise to different approaches to solving moral problems, even though they often agree on which moral problems need to be solved. For example, deontologists may focus on the duty to follow moral rules, whereas consequentialists focus on picking actions based upon the likelihood that they will achieve better moral outcomes. According to Rorty, a society containing some acting as deontologists and others as consequentialists is better off because ‘it is safer and easier to be a full-blown consequentialist if you know that there are enough deontologists around to prevent you from doing something awful for the sake of a distant good’ (2009, p. 53). Similarly, it is safer to act for the sake of pure duty if there are utilitarians around to argue for the distribution of basic human goods (2009, p. 53).Footnote ⁸ In other words, focus on moral rules alone can lead to problematic outcomes (e.g. being honest in cases where it doesn't have good effects); but focus on outcomes alone can lead to violation of important moral rules (such as killing one to save five).

It is especially valuable that diverse moral reasoners are often drawn toward different types of moral projects and roles. Society at large has benefited from Greta Thunberg's focus on climate change – she has increased awareness of the problem, motivated many young people to get involved in the cause, and made it clear what sorts of actions might help us address the problem if we act collectively. This point is also made by Robert Chapman, who argues that while her autistic traits ‘are both beneficial and disabling for her as an individual, Thunberg's role as a climate activist is arguably a vital niche from the group perspective’ (Chapman, Reference Chapman2021, p. 7). Chapman claims that at the level of the group or society, having cognitive diversity is beneficial for the system as a whole, even if the individual traits can at times make life more difficult for the individual and be disabling. He introduces the example of individuals with autistic traits being over-represented in engineering and points out that having a subset of people with a cognitive profile that lends itself to working in engineering and technology is beneficial to society. Chapman stresses that looking at the system level effects of different cognitive styles allows us to acknowledge the benefits of cognitive diversity in ways that are overlooked when focusing on perceived deficits at the level of the individual.

Applied more specifically to the realm of the moral and the idea of practical wisdom, this means that, given human limitations, practical wisdom (or some group analogue of it) may be best achieved if we have a variety of moral reasoning styles that complement each other, especially if persons occupy moral roles well-suited to their moral strengths. Persons with different strengths and weaknesses in practical reasoning may determine that different actions are necessary to address the problem of climate change; and these different decisions and actions may at times serve as a correction to others’ errors in judgment. That is, a person who is sensitive to context may decide that given time and resources, it makes sense to fly to the UN climate change summit COP 26 to make the case for climate action; and a person skilled at focusing on the importance of consistent action on climate change may warn others that refusing to fly is an important – even necessary – way to address global warming. By having these different voices in the mix, we as a society are aided in giving due weight to morally relevant considerations at play in the context we find ourselves in.

From this perspective, rigidity of thought and an intense focus on an object of interest, both of which are features frequently found in autism, can be moral strengths. The world is full of complex moral challenges and keeping all of them in view can lead to distraction and possibly paralysis and a feeling of hopelessness. In many cases, a person will achieve more in addressing current moral challenges if they put a lot of energy into one specific problem and consider how to best approach that problem than if they spread efforts thin. This is particularly true if the moral problem one focuses on is something like climate change, which is complex and requires quite a lot of fact finding to evaluate in the first place.Footnote ⁹