Flickers of Freedom

7. Blame, the objective attitude (again), and whether Donald Trump is a psychopath

 

Blame and the Objective Attitude (OA)

So far I have been speaking mainly about responsible agency, which I have described as a cluster of capacities constrained or partially constituted by rationality, which in turn is facilitated by conditions of social justice, cooperative interdependence, and diversity. Responsibility is thus a relational property, which is enabled and enhanced by these group dynamics, and impaired by factors that hinder them. 

I have said less about praise and blame (as opposed to responsibility qua capacity). I want to clarify here that if someone exhibits rationality deficits (which are responsibility deficits on my view), it does not automatically follow that the person is not blameworthy for his overt behaviour. If this followed, then people would be blameworthy for very little, since many (if not all) moral infractions stem from rationality deficits. However, I reject this deflationary conclusion. The reason is that responsibility deficits in a person’s moral cognition merely dispose the person to problematic behaviour, they do not necessitate that behaviour. Implicit gender bias, for example, is a responsibility deficit on my view, since it hinders rational thinking and inclines one toward overt sexist behaviour, but it does not necessitate this behaviour. The overt manifestation of the implicit state is blameable, even if the possession of the state is not. This is true, on my view, even if the action was not recognized by the speaker as morally problematic. On some versions of character theory (e.g. Angela Smith’s), the person who acts on an implicit state is blameable insofar as the action expressed part of his evaluative agency; and on the most natural interpretation of the control view, the person is blameable in case he failed to take reasonable steps to remediate or suppress the expression of his implicit bias (see Jules Holroyd 2012). Both views allow us to blame people for expressions of implicit states.

But what if the person can’t control the expression of an implicit state because he doesn’t recognize that it’s wrong? This is where the views seem to come apart. Character theory (as described above) grants that a person can be blamed for uncontrollable behaviour, while control theory requires that the agent have control over his action. Fischer and Ravizza, for example, say that responsibility depends on “conditions at the time of action,” together with considerations about the agent’s personal history (2000). I’m inclined to think that both models get something wrong. Re. character theory, I’m not sure why we would want to blame people who are incapable of responding appropriately (i.e., with recognition, not just in a mechanical way) under any conceivable circumstance. We don’t blame cats and dogs (I presume), and non-responsive human beings are seemingly morally equivalent to these species, in which case blaming them is a category mistake. Re. control theory, I’m not sure why we would blame people only in response to their capacities ‘at the time of action’ and prior to that time. It’s common for philosophers to say things like, ‘S is blameable for p at t if S had control over p at t—which suggests that S isn’t blameable for p at t+n, regardless of how S’s cognitive architecture evolved in the interim. But given that our motivational structures are relatively domain specific (e.g., Thach et al. 1992), situation sensitive (e.g., Doris 2002), and subject to ego depletion (e.g., Hagger et al. 2010), our ability to effectively exercise control over our behaviour may differ from one context to another, and from one moment to the next. If so, then whether someone is blameable the control-theory sense may also be contextual.

To bring this into relief, suppose that Smith made an insulting remark to me yesterday in a fit of uncontrollable rage, after being trounced in a tennis match by his colleague and storming off the court. While he managed to repress his urge to batter his colleague with his racket, he couldn’t help making an insulting remark to me later. The urge was utterly uncontrollable (imagine) given his motivational hierarchy and the circumstances. He was utterly unrepentant at the time (also an uncontrollable state). Does this mean that today, when he’s in a calm and collected state, I can’t demand an apology from him? It seems as if I’m entitled to hold Smith accountable whenever he has the motivational capacity to respond appropriately—whether today or tomorrow or next week—irrespective of his motivational psychology at the time of action. This isn’t an endorsement of character theory, but it’s something of a departure from control theory as it’s typically construed, since it denies that blame rests on considerations about the time of action and prior times only.

This example is meant to suggest that our reactive attitudes should do something—they should serve a function. (See Vargas’ ‘moral influence account’ for a similar view [2013]). But these attitudes can serve a function whether they track a person’s motives at the time of action, or her motives at the present time. It’s feasible to think that we sometimes blame people to alert them to the fact that they have a domain-specific rationality deficit that is sensitive to triggering contexts, and they should either avoid those contexts in future, or develop new cognitive strategies to deal with them. In this case, we’re holding the person responsible as a means of encouraging better behaviour in the future, by bringing to light intelligible moral considerations. Our blaming attitude serves a purpose if it is intelligible to the other person.

How can we know when someone is blameable? We can’t. We can suspect, infer, or hypothesize that someone is blameable, but we can’t know that someone is, for various reasons that I’ve discussed here and in previous posts. These include: (1) we don’t have direct access to another person’s motivational states, so every appraisal of a person’s quality of will is necessarily an inference; (2) these inferences are affected by epistemic injustice in a way that undermines the warrant of our moral claims; and (3) given that motivational states are context-sensitive, we can only infer that our blaming attitude will be intelligible in any given context. That said, we can, and often do, make fairly accurate predictions about people’s behaviour across a fairly broad range of contexts, especially if we know the person well. What we need to be wary of is blaming people hastily or over-confidently, though we should not suspend blame when there is sufficient evidence. We ought to blame people if we have good reason to think that they are responsive, could be responsive under different circumstances, or could use alternate cognitive strategies to respond better in future.

If someone is incapable of responding to blame, this is when we ought to take the objective attitude (OA) on Strawson’s view (see the last post). When we take this attitude, we don’t engage with the other person as a moral agent, but instead take steps or enact and enforce policies to control and/or alienate the person. Some people might be globally non-responsive, i.e., morally incapable across every conceivable circumstance: these people are psychopaths. They completely lack emotional empathy, which enables us to care about other people (see Hare 1999, 2002). Others have merely local responsibility deficits. 

To illustrate how my account pans out, let’s do a case study. This case study will also hopefully shed light on how we should think about the current political situation from a moral perspective.

Donald Trump: A Case Study 

How should we morally appraise Donald Trump, the current President of the United States? First, Trump could (in theory) be a good example of how rationality deficits do not automatically translate into non-responsibility. Trump has many rationality deficits, but this does not necessarily entail that he is not blameable for his overt behaviour. On the other hand, Trump has not shown signs of being capable of responding appropriately to blame, and he seems incapable of apologizing or showing remorse. 

Let's home in on a specific example of overt behaviour as our point of analysis. In October of 2016, the media released a video of Trump bragging about sexually assaulting women (in words not worth repeating). The public was justifiably outraged. Trump responded by releasing a so-called apology, which wasn’t an apology so much as a deflection of blame. In his ‘apology’ video, he denies that his remarks reflect his real self, insinuates that he shouldn’t have to apologize in any case because the public’s response is an overreaction that only serves to deflect attention from the ‘real’ issues, and tries to downplay his remarks by describing them as ‘locker room talk,’ insinuating that there is a context in which it is appropriate to demean women.

In other words, instead of taking responsibility, he tried to turn the tables by accusing his critics of being overly sensitive and obscurantist. Now, even if he had offered a satisfactory verbal apology—which he certainly didn’t—it wouldn’t have passed as a proper instance of ‘taking responsibility’ unless he had also taken steps to atone for the damage done to the epistemic and political climate by his statements. He would have had to, e.g., volunteer at women’s safe houses, donate to women’s causes, and so on. But instead of doing any of these things, he made matters worse by offering a sad pretence of an apology, and then, as if to add insult to injury, proceeded to introduce executive orders that harm women, such as the order barring the U.S. from giving foreign aid to NGOs that provide abortion services or even discuss the possibility of abortion with clients.

This shows that Trump has heaping tons of rationality deficits (e.g., implicit and explicit gender bias), and he also hasn’t shown any sign of moral recognition, or of being capable of responding appropriately to blame. Instead of taking responsibility, he seems intent on blaming his critics and escalating his attacks against women's rights and interests.

In light of the evidence, I think that there are two viable ways of construing the situation. (1) We can see Trump as fairly morally impaired, but potentially capable of responding to blame under a limited range of circumstances, in which case blame is appropriate in limited contexts, but OA is the appropriate default response; or (2) we can see Trump as globally morally impaired and susceptible to OA across all contexts. To recapitulate from the last post, OA entails using policy, legislation, and exclusion to control and regulate a person’s behaviour, rather than trying to engage with the person as a competent moral interlocutor. OA is justified only toward people with severe cognitive deficits—though we tend to deploy it in perverse ways in light of pervasive stereotypes, as I pointed out earlier. However, if a person is genuinely morally impaired—either in local but irremediable ways, or globally—the situation warrants OA.

Brian Leiter recently appraised Trump as either a sociopath or a narcissist, but favoured the narcissist diagnosis because Trump doesn’t try to fly under the radar—he "craves attention and adulation" (2016). Both personality disorders cause substantial moral deficits, and both are relatively resistant to treatment, because they cause antisocial behaviour and do not typically cause psychological distress to the agent. While narcissists can respond to prudential considerations, their ability to respond to moral claims is very limited, and sociopaths similarly lack normal moral responsiveness. In both cases, the person might be a better target for OA than useless attempts at blame and moral reasoning, which (likely) will not be intelligible to him. 

I’m personally more inclined to think that Trump is a psychopath, which overlaps with sociopathy and narcissism but entails a complete lack of emotional empathy. Psychopaths, however, have cognitive empathy—the ability to read others’ emotional states, and this can make them more cunning predators. Trump appears to be a living incarnation of Robert Hare’s profile of a psychopath from his book, ‘Snakes in Suits’ (2002), where he says that psychopaths comprise a disproportionate number of CEOs as well as incarcerated violent offenders—though psychopaths prefer to operate on "the shady side of the law" whenever possible (Hare 1995: 25). Although psychopaths comprise a small percentage of the human population (maybe as low as 1%), they cause a disproportionate amount of harm, in part because they are protected and even rewarded in unjust political systems that prize confidence and glib charm over moral character, and systematically fail to protect and defend the psychopath’s primary targets: vulnerable groups. This environment exacerbates the psychopath’s non-responsiveness and anti-sociality by rewarding and reinforcing it (especially, perhaps, in locker rooms?). Not all psychopaths are actively anti-social (e.g., James Fallon, the American neuroscientists, is a benign psychopath by his own reckoning); but those who do exhibit anti-social traits appear to be totally untreatable. For example, violent psychopaths who are placed in forensic treatment programs tend to come out either no different or more violent than before (Rice, Harris, & Cormier 1992). Psychopaths often use what they learn in treatment to harm people more effectively; and they tend to disrupt treatment sessions, spoiling them for more treatable patients (Hare 2002). For this reason, Hare believes that criminal psychopaths are the one group of prison inmates that should not be offered treatment, and should be incarcerated for longer than other groups.

If Trump is a psychopath, then he’s not blameable for any of his behaviours, including failing to remediate his psychopathic traits, given that they are situation-invariant. Hare says that the only reasonable way to deal with a malignant psychopath is to avoid the person at all costs—which is essentially an injunction to take the OA, rather than trying to engage with the person as if he were a moral agent. This is likely what P.F. Strawson would advise as well, and what he had in mind when describing OA. Instead of trying to reason with a dangerous psychopath, the reasonable course of action is to protect yourself, alert others, rally confederates against the person, and contain the person’s anti-social behaviour. These are reactions external to the reactive attitudes, which also function to protect the moral community. Unfortunately, whether Trump is a psychopath or merely a malignant narcissist as Leiter suggests, we may be forced to interact with him (indirectly at least) for four more years—in which case we can only hope that the formal mechanisms of American democracy can contain his recalcitrant behaviour.

6. Strawson, objectification & ‘the objective attitude’ (OA)

Earlier I said that it is objectifying to unfairly blame or praise people on the basis of identity prejudice as opposed to the person’s quality of will. When we blame people on the basis of negative social stereotypes, we don’t accurately gauge the person’s quality of will: we judge the person to be less trustworthy or morally worthy than she really is. This can be seen as a kind of objectification. In Strawson’s language, it can be viewed as an instance of “taking the objective attitude” toward the person, and treating her as an object of “policy, treatment,” or “control,” as opposed to an agent in her own right (1963). OA, as Strawson describes it, represents a complete suspension of the reactive attitudes, and a refusal to treat the other person as a member “of the moral community” (1963). This attitude is justified by perceived cognitive deficits that prevent the person from responding competently to moral address. (He admits, however, that this attitude can come in degrees, so we can either fully exclude, or merely marginalize, someone in our community practices).

If I am right, then in a climate of Epistemic Injustice (EI) we don’t just take OA toward people with severe cognitive deficits; we also sometimes take OA (whether consciously or not) toward neurotypical people, in case we judge them on the basis of non-will qualities such as implicit bias or explicit prejudice. When we do this, we treat the person as an object of prevailing systems of control and oppression. This isn’t exactly what Strawson meant by OA, of course, but these biased attitudes share OA’s primary functions as described by him, i.e., full or partial exclusion from the moral community, control, and enactment of policies that enforce these exclusions—though these biased attitudes are completely lacking in justification. Hence, they are perverse expressions of OA, which function to exclude and control competent moral interlocutors, and enforce policies that buoy oppressive power structures.

To illustrate with a concrete example, when someone is ‘slut-shamed’ or demeaned on the basis of sexist norms, the person is subjected to patriarchal structures of control that function to regulate and suppress the person’s will (especially her bodily autonomy), while simultaneously solidifying the patriarchal order. The slut-shaming attitude, which might superficially resemble an instance of blame (in its form and semantic content), is actually a perverse instance of OA, viz., an attitude that functions to control the target agent and push her to the margins of the moral community, denying her equal standing with men and women who abide by puritanical norms. This attitude matches what Strawson takes the primary functions of OA to be—i.e., to exclude, marginalize, and control people—but it has no justification in the agent’s cognitive structures (which is what makes it perverse). The slut-shaming attitude is assumed, not on the basis of the target’s quality of will, but on the basis of misconceptions about the person’s moral obligations, or else explicit and conscientious attempts to deny women the same privileges enjoyed by men.

I should note here that justified deployments of OA, on Strawson’s view, apply only to severely cognitively impaired or underdeveloped individuals, who are not capable of participating in the moral community: for example, very young children, and people in a coma. (These are moral patients, not moral agents). However, in a climate such as ours, it’s a reasonable conjecture that many people with morally-irrelevant cognitive deficits, or localized albeit morally-relevant cognitive deficits, are liable to be misperceived as apt targets of OA, even though they do have responsibility-constitutive capacities. Indeed, they may have higher net responsibility, across all of their responsibility-constitutive capacities, than the average person; but due to pervasive disablist stereotypes, together with zero-sum thinking about responsibility, they are liable to be mislabelled as exempt from praise and blame. In reality, very few individuals fall under the blanket exclusion category. While many people have localized deficits that may warrant specific accommodations, they are not by that token exempt from any moral appraisal.

If this is right, then in a climate high in EI we’re susceptible to perversely objectifying disadvantaged social groups. I think that this is a sensible way of thinking about objectification, but I also think that it’s an unorthodox usage within the philosophical tradition, in which many theorists (perhaps especially retributivists) have been concerned with a different source of objectification—the kind that allegedly takes place when we unjustifiable excuse someone by appealing to sociological factors to explain the person’s behaviour. The ‘sociological stance’ is supposed to be objectifying and disrespectful because it reduces people to mere nodes in the causal nexus, not uncaused causes, capable of forging our own paths.

On the opposite end of the spectrum, some people have no problem with the idea that we are causally determined. Marx, for instance, held that we’re all determined by the material conditions of our society, and he didn’t see this materialist analysis as disrespectful in the least. But Marx regarded capitalism as a special causal vector, which is distinctively objectifying, because it alienates workers from the objects of their labour, preventing them from either identifying with their labour as a source of value, or cultivating other value-conferring dimensions of their agency. In other words, capitalism reduces people to mere cogs in the capitalist machine.

I think that the truth lies somewhere between these extremes. It’s true that we can disrespect people by underestimating their agency, and this is what happens when society depicts people with morally-irrelevant and localized disabilities as exempt from praise and blame. Yet the reason this attitude is objectifying is not that it takes a sociological stance toward the person, but rather, that it perceives the person through a biased lens. What we ought to do is be better sociologists and take proper account of the role that social structures play in both our material opportunities and our moral perceptions. These factors can constrain our choices or influence our perceptions in morally-relevant ways—ways that can excuse us or inculpate us depending on the particularities of the case. We need to be conscientious of these factors.

On the other side, Marx was right that capitalism is a salient source of objectification, insofar as it causes worker alienation. We can add to this assessment that capitalism gives rise to classist ideologies that oppress the poor, denying them responsibility-enhancing resources and social roles, and excluding them from discursive spaces. That is, capitalism impairs relational responsibility. Capitalism, however, isn’t the only source of perverse objectification in our lives—and it may not even be the primary source. Other sources of systemic objectification include patriarchy, racism, heterosexism, and transphobia (to name just a few). How we are affected by these social vectors will depend on our intersectionalities and life experiences, but they are all objectifying by their very nature, because they all rest on false hierarchies that elevate some people above others for no good (rational) reason.

What I am proposing here is that when we take OA toward another person on the basis of any of these oppressive power structures, we perversely objectify the person. This is true whether we neglect the role of social factors in the person’s life, or we neglect the expectable role of bias in our perceptions of people’s qualities of will.

The practical upshot is that if we wish to be responsible moral interlocutors—responsible ‘adressors’ and ‘accounters,’ as McKenna puts it—we need to take these sociological considerations into account when appraising a person’s quality of will. That is, we need to exercise due diligence as moral conversants. Yet, as we saw earlier, even due diligence can’t insulate us against the warrant-undermining effects of EI. Thus, we often have cause to blame and praise people with an attitude of epistemic modesty and caution, being inquisitive rather than assertive—especially if a person is being accused of a stereotypical infraction (where bias is likely to play a role, and the response is likely to be punitive). In this connection, McKenna says that “moral inquiry could be added to the [three-stage conversational] model” (2013: 99), though he doesn’t officially include it. But if EI affects our recognitional capacities in such a way as to systematically undermine the epistemic warrant of our moral attitudes, then moral inquiry is quite indispensable. Specifically, we should, at least initially, approach the target in the interrogative mood as opposed to the declarative mood, inquiring as to whether our intuitive impression of the situation is right; and we should be sensitive to proffered justifications and explanations that might contradict our preconceived notions. This inquisitiveness and sensitivity is demanded by epistemic justice. In practice, this means that if we suspect someone of morally appraisable behaviour—especially if that behaviour is stereotypical in the context—we should (1) question, interrogate, accuse, or ask for an explanation, (2) consider the person’s response, and (3) express the fitting reactive attitude in light of all relevant information, in that order.

This might seem obvious—and indeed, we often do preface blame and praise with moral inquiry—but placing this stage at the heart of moral conversation is a proposed revision to MacKenna’s view, which situates it at the periphery. 

Now, earlier I said that the notion of objectification that I am proposing here does not fit with the sense of ‘objectification' that interested many classic philosophers. But my sense of the word does cohere with a familiar equivalent in ethical theory: the sense defended by Kant. Kant described objectification as treating someone as a mere means, not an end in herself. He had in mind things like murdering someone, lying, and (weirdly) masturbating. But aside from some bizarre examples, the idea that objectification is an action or practice that uses someone as an inert prop in someone else’s projects is very intuitively compelling. And this is precisely what perverse expressions of OA do: they treat the target agent as a mere means of sustaining oppressive social hierarchies, and they exclude, marginalize, and/or control the target agent. This is the opposite of treating someone as an end in herself. To treat someone as an end, we would have to evaluate her quality of will fairly in light of relevant social considerations—and doing this requires epistemic virtue.

These instances of OA, once again, are not explicitly discussed by Strawson, but they can be seen as extensions of his notion of OA, since they serve the same essential functions as that attitude, though they lack its characteristic justification. Strawson was not in a position to address these applications because he saw moral responsibility as an intimate interpersonal exchange, and did not look farther afield at the systemic factors that inform our more local interactions.

5. Relational responsibility and (oppressive) gender roles.

What I have said so far implies that responsibility is relational. Why? Because we’re all implicated in joint rational projects (except for total social isolates), and rationality deficits in one social group tend to precipitate correlative, but asymmetrical, rationality deficits in other social groups. For example, implicit bias in privileged group members tends to induce stereotype threat in the least-well-off, as well as undermining the social roles realistically available to members of these groups.

To illustrate this in more concrete terms, we can look at how traditional gender roles impaired responsible agency in men and women (and trans folks, who were virtually erased from public discourse), by restricting the range of social roles available to each group. (I’m simplifying by talking as if there were three genders, though I believe that gender lies on a spectrum; however, since it’s not feasible to discuss every possible gender intersection). The traditional division of labour along binary gendered lines prevented each gender from achieving robust responsibility, but it also arguably gave rise, in certain cases, to group-specific cognitive deficits that impaired moral cognition in the affected group. For example, men’s almost total exclusion from care-taking roles may have negatively affected their emotional and social cognition, impairing an important component of responsible agency.

As Sharyl Sandberg has famously pronounced (2013), women are to this day expected to ‘lean in’ and work harder for the same credit and compensation as men for comparable work, because women also typically (still) do a majority of the domestic labour. Women also face such hurdles as workplace discrimination.

Naomi Wolf raised similar considerations in ‘The Beauty Myth,’ published in 1990 (republished in 2002). There, Wolf argues that (salaried) working women are responsible for ‘three shifts’: (1) paid labour, (2) domestic labour, and (3) beauty labour. Wolf continues that if women were emancipated from their secondary and tertiary ‘extra’ responsibilities, they would quickly outpace men in their academic and professional achievements, since women have been acculturated into working harder for less recognition and compensation. Now we’re (finally) starting to see Wolf’s prediction come to fruition, with women completing college at higher rates than men (Conger & Long 2010)—though they still earn less on the dollar due to resilient patriarchal norms. The reason for women’s higher academic achievement in the last 2.5 decades is largely, as Wolf predicted, because men are doing more domestic labour, creating a more equitable split of domestic duties. This allows women to focus more on their salaried careers. (One would also hope that women would be under less pressure to conform to Western beauty standards, but instead it appears that men are under more pressure to conform to a constructed masculine aesthetic. Nonetheless, this has had an equalizing effect on the paid workforce, levelling the playing field, regardless of whether it has been ultimately emancipatory).

Wolf’s example of women’s extra shifts helps to illustrate how responsibility is relational, in the sense that it depends upon everyone doing their fair share. When women are confined to the domestic sphere by artificially-imposed social structures and social expectations, they are not able to take on social roles that confer role responsibilities—responsibilities that women might want to cultivate, and that are valued by society (as evidenced by the higher remuneration typically attached to historically male positions). Until fairly recently in our history, women were only allowed to cultivate responsibilities related to the domestic sphere, providing a relatively ‘thin’ swathe of responsibility-constitutive capacities available to them.

Furthermore, women’s systematic exclusion from public life sustained false and mythical ideologies about such things as gender (binary), sex (equivalent to gender), the self (atomistic), and many other topics that have been criticized by contemporary feminist scholars. Women’s exclusion from public life thus impoverished the public discourse, and propped up oppressive falsehoods.   

Men, of course, were also harmed by patriarchy in many ways, including in their responsible agency (though of course they were socially privileged). Yet men were limited in the range of viable social positions and corresponding role responsibilities available to them under the binary gender paradigm and its associated social possibilities. By being barred (or exempting themselves) from the domestic sphere, men were also harmed in their capacity for care-taking, and perhaps in their ability to develop robust social and emotional capacities. It seems that these capacities are implicated in moral responsibility, since deficits in emotional empathy, for example, result in psychopathic traits, which are taken to be responsibility-impairing by many (see, e.g., Nelkin 2011). If so, then complete lack of acquaintance with care-taking may have negative affects on responsible agency.

In addition to being largely exempt from care-taking roles, most men grew up in a culture of toxic masculinity—a climate that normalized (even glamorized) rape, sexual assault, and other misogynistic aggressions and microaggressions. (Don’t forget that spousal rape was not a recognized legal category in all 50 states until 1992, and a vast majority of rapes still go unreported). It’s a reasonable conjecture that the lack of exposure to caretaking roles for the average mid-20^th-century cisgendered man, combined with saturation in a climate of toxic masculinity, may have had responsibility-impairing effects on many members of this social group—particularly those who were never exposed to alternate discourses. 

While things are improving in many respects, Donald Trump’s recent boast about sexually harassing women perfectly exemplifies how this attitude is a salient part of American culture. And Trump won the Presidency, notwithstanding his attitude towards women. One can only hope that this is the calm before the (feminist) storm.

I haven’t said anything about trans people, but it should go without saying that they have been disproportionately disadvantaged in many ways by the classic gender binary, including being practically erased from public discourse. Trans folks were thus very limited in the range of viable social roles that they could perform as their authentic selves; and they are still limited by negative attitudes and institutional discrimination to this day. When people consistently play an artificial role, this can cause significant psychological distress. Researchers recently found that when employees pretend to be happy at work, they suffer from increased insomnia, anxiety, depression, emotional exhaustion, and family conflict (Wagner et al. 2013). Imagine wearing a social mask all the time! Psychological distress is one of the paradigm factors identified by philosophers as impairing responsibility. So, the psychological cost of the gender binary on trans people has been (and still is) enormous, and the collateral effect on responsible agency can be very negative. (Though recall that responsible agency can be enhanced by privileged insight!)

Moreover, since trans people are still marginalized within mainstream discourse, society is losing out on valuable sources of enlightenment and objective knowledge—knowledge that could enhance our collective responsibility. The recent Women’s March—also colloquially called the ‘pussy march’ in response to Trump’s sexist remarks—was highly commendable (I was there); but some trans women pointed out that symbolically associating women’s rights with a part of cisgender female anatomy is both marginalizing to some trans women (non-binary, pre-operative, etc.), and reminiscent of the reductionist/biological view of gender that poststructural feminists have rightly militated against. Although the pussy rhetoric was clever and possibly galvanizing, it’s worth listening to these voices from the margins, and ensuring that we don’t lapse back into an essentialist sex-gender conflation. If we exclude trans voices, in other words, we (to that extent) impoverish feminist discourse and exclude valuable allies, limiting the movement’s social efficacy and justificatory backing. Only by including every valid perspective can we achieve the robust rationality that supports relational responsibility as I understand it.

(I am indebted to Rachel Williams at Washington University in St. Louis for brining some of these considerations to my attention. You can read her blog, transphilosopher, here). 

4. The reactive attitudes

Another prominent locus of responsibility discourse is Strawson’s theory of the reactive attitudes. Strawson is one of the most influential contemporary responsibility theorists, being a central catalyst for mainstream compatibilism. On Strawson’s view, responsibility is an interpersonal practice in which we deploy (feel or express) the reactive attitudes, consisting of approbation, praise, disapprobation, resentment, indignation, and the like (1963). Michael McKenna defends a more specific, situated version of Strawsonianism, on which the reactive attitudes are part of a conversational practice with three paradigmatic stages: (1) a moral contribution, in which someone perform a morally-relevant action, (2) a moral address, in which someone deploys the reactive attitudes, and (3) a moral accounting, in which the addressee responds, explaining or accounting for her actions (2013). McKenna’s view situates Strawson’s theory within a two-person conversational exchange. The stages of this exchange often include the familiar speech acts of: accusing, blaming or praising someone; accepting responsibility, apologizing, or offering amends; and offering an explanation, excuse or justification. 

This is a helpful elaboration of Strawson’s model, as it gives responsibility a concrete form and helps us to understand what Strawson meant when he described responsibility—somewhat ambiguously—as an ‘interpersonal practice.’ McKenna’s explanation is a step in the right direction, I think, in that it clarifies the context in which responsibility takes place (a two-person dialogue), and the some of the discursive forms that our reactive attitudes and responses often take.

But this is only part of the context, surely. McKenna’s account describes a two-person exchange, but says nothing about broader social conditions in which the exchange takes place. It’s reasonable to think that EI could affect our ability to participate in our roles as moral addressors and moral accounters in a rational way—a way that enhances, or at least preserves, the responsible agency of the conversational partners. We might, for example, deploy the reactive attitudes on the basis of implicit or explicit bias in a climate of EI, or respond to blame in biased ways. If so, then our epistemic climate impairs the proper functioning of our moral conversations—it prevents them from proceeding in a fair and constructive way. 

We can bring this into relief by homing in on the second and third stages of McKenna’s conversational model, and situating them against a background of EI.

1. In the moral address stage, addressors might unfairly underestimate the moral standing of disadvantaged group members due to implicit bias. They might also deploy the reactive attitude in discriminatory and punishing ways, if they are explicitly biased. In either case, there will be a tendency to excessively blame and criticize disadvantaged group members compared to other people. There will also be a tendency to reject or discount the moral accounts of members of disadvantaged groups, since EI tends to result in ‘testimonial injustice’—the unfair discrediting of people’s testimony on the basis of the demographic attributes and social scripts about those attributes. Testimonial injustice includes biased responses to the testimony constitutive of a moral accounts—accounts that testify to one’s innocence, that respond defensively to unfair blame and censure, and so on. Testimonial injustice can prevent the moral addressor from responding appropriately to the moral account of disadvantaged group members. It can also operate in the inverse way, prompting the person to overestimate his own credibility and moral worth.

2. In the moral accounting stage, stereotype threat and low self-regard might inhibit the performance of members of disadvantaged groups, in their role as moral accounters; for example, it might prevent them from providing confident rebuttals to misplaced blame, or assertively defending their moral standing in the face of unfair and biased recriminations. (This is obviously not a necessary effect of EI, but it is something to watch for if we want to be responsible moral accounters). In addition, stereotyped individuals may not receive proper uptake (or due credit) from others for their moral accounts due to EI effects. By contrast, privileged group members are more likely to have high epistemic confidence, enabling them to defend their moral standing more effectively—perhaps too effectively: given too much epistemic confidence, one might be tempted to deny apt blame and censure.

These considerations show how EI can affect our moral conversations in biasing and responsibility-impairing ways. The cognitive effects of EI can place negatively stereotyped groups in a position of relative disadvantage in moral conversation. This is a kind of ‘moral-conversational disenfranchisement, which we should be cognizant of if we want to be responsible (rational) members of moral conversation.  

In addition to these direct effects, EI might undermine our moral conversations in more insidious ways. Specifically, it might undermine the ‘epistemic warrant’—the degree of justification or authority—that every single one of our moral claims has, including our reactive attitudes and moral accounts, viz., the central constituents of moral conversation. Why is this? We can explain this phenomenon by reference to standpoint epistemology (see, e.g., Harding 2015, Longino 2001, Potter 1993), which holds that our knowledge claims are epistemically underwritten not only by our individual beliefs and attitudes, but also by the background assumptions shared by our culture—the total network of beliefs and attitudes held by everyone with whom we share a form of life. This means that flaws in our epistemic environment—such as pervasive negative stereotypes and social scripts—undermine the epistemic warrant of our knowledge claims, including (but not restricted to) our moral knowledge claims, i.e., our claims about what is morally the case. Hence, when we claim that S is blameworthy or S is excused from responsibility or what have you, we must treat this is a mere moral belief or opinion, not a piece of moral knowledge. We must, in other words, be modest in our moral claims, and amenable to revising those claims in light of new information. One way of increasing the justification of our moral claims is to be vigilant about the role of cognitive distortions in our conversations—but this won’t give us moral knowledge. To be truly confident, we need to fix the flaws in our collective discourse. We need to reject the harmful stereotypes, subtle biases, and exclusions that we encounter on a regular basis. 

Notably, there are ways of indirectly controlling and remediating implicit biases, which have received extensive research—methods such as counter-stereotypical exposure and implementation intentions (see Kelly et al. 2010). Privileged people are arguably blameworthy for failing to use these methods to improve their moral profile (Holroyd 2012). But on the other hand, these methods are of limited immediate efficacy and questionable long-term efficacy. In a climate of epistemic injustice, implicit bias is practically inevitable, even for the most reflectively scrupulous among us. It is a dangerous mistake to think that meticulous self-examination will inoculate us against implicit bias. While we out to try to be critically reflective and vigilant about the role of implicit biases and other cognitive distortions in our moral reasoning, this doesn’t exhaust our moral responsibilities. We also need to take substantial measures to improve the quality of our public discourse.

Biased responsibility practices are not only unfair to the disadvantaged parties, but harmful to the entire moral community, since discriminatory treatment tends to alienate disenfranchised groups—those who are not getting the moral credit they deserve—and alienation in turn tends to create subcultures of disaffected individuals who reject the prevailing social contract. This is one of the reasons that EI undermines our relational responsibility—it deters some people from participating in moral conversations at all, or participating in good faith; and it also excludes the people who have the most epistemic insight and authority vis-à-vis our flawed social discourses and institution—the people we most need to contribute. When we exclude people from moral conversation, we undermine our moral culture.

Interestingly, ethical theorists tend to suppose that most people are already and at all times recognized members of the moral ingroup, or ‘moral community.’ This includes Strawson, who takes for granted that most people are subject to the ‘participant stance,’ with only a few ‘deranged’ people falling outside of the moral community as we know it. Similarly, Scanlon says that morality includes all “suitably motivated” individuals (1998: 189), meaning those willing to cooperate in the social order; and he seems to think—though I could be wrong—that most people are interested in cooperation, and are also given a seat at the table and treated as equal contributors. There is, in any case, no discussion about how some people’s speech counts for more than others in our public discourse and public understandings. Rawls took a completely different tack, assuming that only landowning white males were fit to be contractors in the original position; and Kant similarly included only White males in the sphere of moral reasoning. In the real world, our moral-conversational practices may superficially purport to be open to everyone, and grounded on fair and impartial principles, but that’s not how they work. In reality, these practices resemble Kant’s and Rawls’ Eurocentric worldview more than we might like to admit, inasmuch as landowning White males still contribute disproportionally to public discourse, and receive more credit, trust, and respect than they deserve (in general). This is because EI functions in such a way as to favour historically privileged groups—those who were previously regarded as the only moral agents, or at least the paradigm case. This radical ignorance is a predictable effect of a high degree of EI. As Fricker points out, EI creates disparities in our social interactions, representing some people as moral credible and socially valuable than others; since moral conversation is a social interaction, it is not immune from the structural inequalities that beset our broader social interactions. 

To eliminate stratification in the moral community, we need to work to eliminate the biases and unjustified assumptions that (still) play a role in our responsibility system.

Before closing, I’ll anticipate an upcoming discussion by noting that, on Strawson’s view, when we unfairly morally appraise someone, we (technically) objectify the person by failing to respond to the person’s quality of will. Strawson says that the reactive attitudes must respond to the target’s quality of will. We can suspend or modify these attitudes if the person is either temporarily or permanently cognitive impaired—for instance, cases of hypnosis or severe cognitive disability. If I am right that we often (intentionally or unintentionally) morally appraise people on the basis of extraneous factors, such as EI or the cognitive effects of EI, as opposed to homing in on the person’s actual quality of will, it appears that we often modify the reactive attitudes for no good reason. In some cases, we might illegimately ‘take the objective attitude’ toward the person, although the person is not, in fact, morally incapacitated—we are just treating the person as if he were. In other cases, we might illicitly blame someone, modifying the reactive attitudes in a punitive way.

By thinking about responsibility against the backdrop of EI, we pinpoint characteristic ways in which our responsibly attributions can, and often do, go wrong.

I will expand on these points in post #6.

Related articles

4. The reactive attitudes - Flickers of Freedom

3. Responsibility & epistemic injustice

There is a prodigious literature on responsibility, most of which emphasizes the internal properties of persons constitutive of responsible agency. We can call this approach ‘internalism,’ in contrast to ‘externalism’ (see previous post). Internalism is reflected, for example, in debates about whether character or control is the proper locus of responsible agency. These debates tend to focus on embodied internal properties (reflective states, values, virtues), and sometimes a limited range of exogenous factors. The most salient example of the latter are direct neural interventions, like nefarious neurosurgery, counterfactual brain implants, and brainwashing. While I don’t want to deny that these are interesting and constructive discussions, there’s a broader range of situational factors that we can explore—factors that operate on the agent’s responsibility-constitutive neural states less directly, but might have equally significant effects on these states. 

One of the problems with focusing on direct neural interventions in discussions of responsibility is that these interventions are actually fairly useless, given the current state of cognitive neuroscience. The range of realistic moral neuro-interventions is (at this time) very limited. We cannot do personality-altering neurosurgery, for example. We also cannot very effectively enhance a person’s moral psychology with pharmaceuticals. In spite of some transhumanitsts’ optimistic claims about ‘moral neural enhancements’ (e.g., Paul Zak), existing examples of these interventions, like oxytocin nasal spray and transcranial magnetic stimulation, are not very effective at inducing moral traits (properly understood). Oxytocin boosters, for example, are ambivalent in their effects, producing both positive effects (trust) and negative effects (outgroup bias); are relatively weak at modifying the targeted personality structures; and are sensitive to a person’s background psychology, meaning that they sometimes produce trust but sometimes don’t. Some people are categorical distrusters, resistant to the effects of oxytocin. These considerations are meant to show that we cannot simply enhance ‘moral agency’ by intervening in neural states with pharmaceuticals and other interventions. What we can do, given sufficient political will and organization, is address social and situational factors that hinder the acquisition of responsible agency.

Some philosophers have begun to pursue this line of research (externalism), focusing on the moral landscape. Notably, Manual Vargas discusses the ‘moral ecology,’ or set of moral considerations in a given culture that underpin responsibility (2013), and Susan Hurley writes about the importance of the ‘public ecology’ (2011). These approaches mark a methodological shift away from internalism, toward a more sociological and anthropological analysis. Hurley and Vargas are both interested in (i) how public policies and practices inform our collective uptake of responsibility, and (ii) how we can best intervene to improve these structures. They are both, as far as I can tell, consequentialists and pragmatists, interested in altering the moral landscape in positive ways. Their work addresses such ‘ecological factors’ as psychiatric discourse, advertising, and political campaigns, and how these things either scaffold or sabotage our responsible agency.  

I wish to expand on this discourse by focusing on another aspect of the moral ecology: namely, epistemic injustice (EI). EI, as Miranda Fricker describes it (2007), refers to an epistemic climate in which pernicious social stereotypes and social schemes give rise to systemic discrimination against historically disadvantaged social groups—groups like women, African Americans, and the queer community. One of the vehicles for EI identified by Fricker is ‘testimonial injustice,’ which occurs when someone fails to give another person the credit or trust that he deserves. For example, if a police officer distrusts someone’s testimony because the person is African American (as often happens), he is enacting testimonial injustice (Fricker 2007: 1).

While Fricker does not describe the cognitive mechanisms through which EI operates, good candidates seem to be: negative implicit biases, which are relatively automatic and unconscious prejudices toward vulnerable social groups (e.g., implicit racial bias, implicit gender bias, implicit disability bias); negative explicit biases, which are conscious prejudices toward these social groups; stereotype threat, i.e., the tendency to conform to negative stereotypes about one’s social group, impairing performance on certain tasks, though not competency; and low self-regard, which is a negative (perhaps conscious) self-conception that similarly impairs task performance. I see these mechanisms, which flourish in a climate of EI such as ours, as responsibility-impairing, insofar as they impair rationality—the ability to grasp and appreciate another person’s epistemic and moral character, to give others the credit they deserve, to respond appropriately to other people’s moral testimony, etc. As I will argue, these cognitive distortions can interfere with our shared moral practice, or ‘moral conversation’ as McKenna calls it (2013), by preventing us from holding people responsible in a rational way—a way that responds to their epistemic and moral attributes—and taking responsibility for ourselves in a rational way.

When we discredit or disvalue others due to the cognitive effects of EI, we exhibit and entrench responsibility-deficits in our own moral psychology, and in the same stroke, we harm the target’s responsibility status, either directly—by limiting the person’s social role options—or indirectly—by lowering the quality of public knowledge. When we harm public discourse, we deprive ourselves of epistemic justification for our reactive attitudes, making our moral attitudes, at best, weakly justified. We can never, that is, hold someone responsible with any confidence.

I’ll expand on the first part of this claim first: our responsibility deficits harm other people’s responsibility status.

Suppose that a hiring committee is seeking the most qualified candidate for an opening in the company, and they make their selection on the basis of implicit racial bias rather than merit, as per the famous Bertrand & Mullainathan study (2014). They select the best applicant out of the candidates with stereotypical Anglo-Saxon name, rather than the best applicant, full stop. The committee members in this scenario are acting irrationally by their own lights, since they are explicitly committed to hiring the best candidate (full stop), and have (as per the study) even advertised a commitment to employment equity. Since they are acting against their rational values and commitments, as well as on the basis of a relatively uncontrollable, automatic, and unconscious cognitive state (implicit bias), they are responsibility-impaired as members of the hiring committee. At the same time, by choosing an Anglo-Saxon candidate, they are preventing a more qualified candidate, from a socially disadvantaged group, from achieving her rational goals. If she cannot find equivalent work elsewhere, she is responsibility-impaired by virtue of not being able to attain a role responsibility that she finds valuable, and that society values.

(This is what happens when there is a ‘glass ceiling’, or social and material obstacles, barring certain social groups from obtaining jobs with certain role responsibilities; the person’s range of responsibilities are artificially circumscribed, and thus her responsibility-constitutive capacities are not as robust or expansive as they could be).

Now suppose that the hiring committee is explicitly biased against stereotyped social groups. Are they rationality impaired? Yes, but not in their instrumental rationality (their means-end reasoning). In this case, they are impaired in their substantive rationality—they cannot (due to their cognitive architecture) act in a manner that serves their substantive interests, both as business people and as moral agents. In our society, companies benefit from hiring the best candidate for the job from a range of demographic backgrounds, and society benefits from demographic diversity within companies (combined with positive diversity mindset ad other positive epistemic conditions). Thus, the company, and the broader society, would be served by a moral objective (less racist) hiring choice. Moreover, the hiring committee, as moral agents, would be better moral agents if they remediated their racist explicit beliefs and values; thus, they would be more morally responsible if les racist. For these reasons, explicit racism also impairs the agent’s responsibility. 

Job applicants from stereotyped social groups may also be responsibility-impaired by EI, if they exhibit stereotype threat—as tends to happen in this type of climate. Supposing that job applicants for a certain job were subject to stereotype threat, they would be responsibility-impaired in the sense that they would be prevented from exhibiting their competency on relevant tasks, such as the interview, and thus they would be less likely to get the job. Manifestations of stereotype threat tend to undermine one’s ability to demonstrate one’s character profile in public contexts, and to control one’s performance on relevant tasks, so in these regards this cognitive state is responsibly-impairing on the traditional models. It can also be responsibility-impairing on my view if it prevents one from obtaining role responsibilities that one is capable of exceling in, only because of stereotype threat effects.

On the other hand, the hiring committee might underrate the applicant’s interview performance, if they harbour implicit biases toward the person’s social group, as often happens in climates of EI. (This is why applications are becoming increasingly anonymized, at least in responsible institutions that are committed to real employment equity). Bias against certain social groups can directly impair the responsibility prospects of those groups by barring them from valuable role responsibilities, and it can also indirectly harm their responsibility status by perpetuating harmful stereotypes—stereotypes that induce implicit bias, stereotype threat, and other cognitive states that typically manifest in responsibility-harming ways.

This discussion illustrates how EI tends to undermine responsible agency in members of different social groups in (characteristically) differential patterns. Namely, EI undermines the responsibility of those in positions of privilege by making them susceptible to negative implicit and explicit biases, and it undermines the responsibility of members of disadvantaged social groups by making them susceptible to stereotype threat and low self-regard (generally speaking). EI-based cognitive distortions also undermine our collective responsibility by tainting the public discourse, perpetuating pernicious stereotypes that reinforce EI. In a climate of EI, we cannot be confident that we are gauging people’s responsibility—particularly, that we are blaming and praising people—rationally, given that rationality-impairing cognitive distortions are pervasive and inescapable. So, we can only tentatively, or weakly, hold people responsible—at least until we repair the flaws in our epistemic climate. An attitude of moral uncertainty and epistemic humility is presently forced on everyone—particularly privileged individuals—by EI.

2. Responsibility, rationality, and collaboration

Much of the philosophical literature on responsibility revolves around conceptual issues: what is responsibility? Is it a capacity, a suite of capacities, a trait of persons, a privileged subset of a person’s mental states? I don’t hope to resolve these debates here—maybe they’re irresolvable, if there turns out to be insufficient evidence on either side. What I do hope to accomplish here is to shift the focus of the debate in a different direction— away from a focus on the internal properties of persons, which are supposed to be constitutive of responsible agency, toward a focus on the environmental, public, or ‘ecological’ factors that cause, support, and arguably co-constitutive those capacities—a view that I call ‘methodological externalism,’ as opposed to ‘methodological internalism.’ I use these terms to distinguish two separable, but not entirely distinct, approaches. One focuses more on the internal states of agents—things like the structure of an agent’s desires or the robustness of an agent’s values—and the other focuses more on the social conditions in which an agent comes to have and exercise those capacities or structures.

Before proceeding, I should say something about what I take responsibility to be. We can demarcate the literature into two camps: the character view, on which responsibility is a matter of the agent’s deep self or most robust character traits, and the control view, on which responsibility is a matter of an agent’s control mechanism or capacity to respond to reasons. This is a popular way of systematizing the literature (see, e.g., Levy 2014). Rather than decide between these views, I’m going to assume that they both get something right: responsibility has something to do with character and something to do with control. But I will also focus on what I take to be a common thread between these views: they both tie responsibility to rationality. As Susan Hurley argues (2013), both models of responsibility—perhaps all models—presuppose that responsible agency consists of or is constrained by rational agency. If an agent is acting either against her own instrumental reasons (reasons she takes herself to have), or against her own substantive reasons (reasons she objectively ought to care about), then she is, to that extent and in that capacity, responsibility-impaired. For example, if someone is shopping compulsively and spending inordinate amounts of money on commercial products, he is rationality impaired and, by the same token, responsibility impaired. Put differently, the compulsive shopper is acting ‘out of character’—his compulsion is not part of his deep self; and he lacks control over his compulsive urges. His compulsion is ‘shallow’ and relatively uncontrollable, and this is in part because it is a rational defect.

I should also clarify out the outset that responsibility, as I understand it, is not simply a capacity or a mental state—although it is typically described in singular terms, which can be misleading. (In particular, it can lead one to assume that responsibility is a zero-sum proposition, rather than a spectrum concept, consisting of multiple parallel, largely independent mechanisms, rather than a monolith). I take responsibility to be “a suite of capacities,” as Vargas puts it (2015: 201), or a cluster of entangled mental states, as Sher describes it (2009). This leaves open the possibility that a deficit in one responsibility-constitutive capacity/state could leave intact or even enhance other, parallel responsibility-constitutive capacities/states. To give a simple example, having a mental health disability could impair one’s ability to function in certain areas of one’s life, while enhancing one’s artistic capacities—as often happens. This seems to have been the case for Van Gogh, who was diagnosed with schizophrenia and was hospitalized several times (Schindeler 1935), and yet managed to produce esteemed expressionistic artworks until his death. It seems perfectly coherent to say that Van Gogh was praiseworthy (and thus responsibility-apt) for his artistic accomplishments, even if they were completed while he was in a psychotic state that impaired some of his basic functional abilities. 

In a similar vein, a member of a disadvantaged social group could be impaired in her ability to secure a job due to epistemic insecurity or material obstacles, and yet be responsibility-enhanced in terms of her ability to understand and appreciate the social factors that undermine her own agential capacities. This is what is called ‘epistemic privilege’ by feminist social epistemologists, and it refers to special epistemic authority on certain matters, as conferred by one’s social position. Often, social oppression engenders epistemic privilege vis-à-vis the social structures and mechanisms that support and sustain a privileged class, while oppressing historically disadvantaged groups. This might be especially salient to the victims-survivors of social oppression. In this case, the individual might be impaired in certain facets of her responsible agency by her context, while responsibility-enhanced by that same context in terms of her ability to grasp, appreciate, and critique dominant social paradigms. In this way, responsibility—conceived of as a set of dissociable capacities—can be something of a double-edged sword. 

One of the implications of seeing responsibility as tied to rationality, I will argue in what follows, is that responsibility turns out to be relational, just because rationality is relational. More specifically, our rational interests are interpersonal interests—they are served by collaboration in pursuit of our joint projects. To give an example, if I’m interested in being a responsible philosopher, I will read as much relevant literature as possible. This includes literature by underrepresented groups. If I ignore this literature, I’m acting in such a way as to undermine my own rational self-interests, particularly my interest in producing objective research. At the same time, I’m undermining the career prospects and visibility of members of underrepresented groups, i.e., I am undermining their rational interests. Hence, when I undercut my own rational interests, I undercut those of other people by the same stroke. By acting irrationally, I am undermining my deep-seated values, and failing to respond to objective reasons that I ought to care about—that is, I am responsibility-impaired on both standard views. This is because, when I undermine my own rational interests, I thereby undermine my prospects for responsibility. And I undermine the possibility of ideal rationality, and responsibility, in the entire group.

Now, this example is about ‘philosophical agency’—arguably a kind of epistemic agency; but epistemic agency is also, I think, moral, in that it has robust moral implications. If you manifest implicit biases (by acting in a discriminatory way), you are also, by the same token, acting in a morally objectionable way. Epistemic flaws often have moral consequences, and are in this sense morally significant. Thus, while it is possible to conceptually distinguish epistemic reasoning and moral reasoning, they are, in fact, interrelated: moral reasoning is undermined by epistemic deficits such as implicit bias, and are served by (what Miranda Fricker calls) ‘epistemic virtue’—sensitivity to a person’s character. Unfairly discrediting someone on the basis of identity prejudice is an epistemic error, but it is also a moral infraction.

I should also clarify that moral responsibility, as I understand it, is related to our various social roles. Being a responsible philosopher, teacher, friend, co-worker, citizen, parent, teacher, etc., is part of what it means to be a morally responsible person. Impairments in any of these ‘role responsibilities’ thus constitute impairments in moral responsibility, or some facet(s) of morally-responsible agency. It’s hard to even fathom what it would mean to be a responsible moral agent independent of these roles, since being a moral agent just means being an agent embedded in a social network with various social roles—roles that generate responsibilities. Impairments in morally-relevant ‘role responsibilities’—those that have significant consequences for others—thus equate to impairments in moral responsibility.  

Now, I’ve said that responsibility is tied to rationality, and rationality is relational. This is, in part, because rationality is a cooperative enterprise. Or, as John Doris puts it, optimal human reasoning is facilitated by cooperation (2015). This is perhaps most obvious when we think of global projects like preventing catastrophic climate change. No one person can make a substantive difference to the state of the global climate—which prompts Walter Sinnott-Armstrong to provocatively declare that ‘no one’ is responsible for climate change (2010). But if no one is responsible for climate change, that can only be because everyone is; and it is only by cooperating (e.g., introducing laws and regulations and social practises) that we can make an appreciable difference. If there is a cataclysmic climate disaster (as there is forecast to be if global temperatures increase by 2 degrees C compared to pre-industrial levels—1.2 degrees C more than the current temperature), we will all be impaired in our responsible agency, since we will either be killed, or geographically displaced, or otherwise encumbered by burdens that we didn’t previously face, which potentially undermine our ability to perform our familiar role responsibilities—as citizens, parents, workers, etc. The environmental case brings into relief how our rational interests are intertwined: we need to cooperate on large-scale, international environmental projects to ensure the continued survival of the species. If we fail in this particular joint project, we will all be responsibility-impaired at best, and dead at worst. So, this joint rational project is, at the same time, a responsibility-preserving project.

This is a big-picture example, but small-picture scenarios also illustrate the relational nature of responsibility. For example, without healthy interpersonal relationships, we cannot develop the capacity to function in any role responsibility whatsoever—we become quite anti-social. This is demonstrated, for example, by Harry Harlow’s inhumane experiments on monkeys (hopefully never to be repeated): Harlow showed that when monkeys are isolated from all social contact, they become anti-social and psychotic (Blum 2002). When re-introduced to other monkeys after a year, the social isolates were patently disturbed, incapable of playing, and two of them starved themselves to death. Similar psychological disturbances can be observed in prison inmates placed in solitary confinement—which ironically makes them more antisocial and more likely to recidivate (Grassian 1983). Attachment theory builds on this work, showing that children whose primary care-givers are unresponsive and insensitive tend to develop ‘insecure attachment’ (anxious, avoidant, or disorganized), which impairs their ability to have healthy adult relationships (Bowlby 1982, Ainsworth et al. 1978). Individuals with anti-social personalities and severe social deficits will not be particularly effective at taking on and adequately fulfilling their role responsibilities—in other words, they will be very substantially responsibility-impaired. Their social deficits may incite them to harm members of the general community, either be being freeloaders or positively harming others. A salient example of an antisocial agent is the psychopath, who suffers from deficits in cognitive empathy, making him extremely dangerous. Canadian psychologist Robert Hare says that psychopaths comprise only 1% of the human population but commit half of all violent crimes (2007: 18). These crimes can drive innocent victims into lives of self-harm, criminality, and mental illness (e.g. PTSD). That is, psychopaths’ harmful actions can deprive others of responsible agency. While some psychopaths are (according to Hare) untreatable, others (‘secondary psychopaths’) may respond to treatment; and many violent offenders are non-psychopathic, and more responsive to treatment. We have an interest as a society in rehabilitating offenders’ responsibility status (if possible), not only for their own sake, but to protect the responsibility-constitutive capacities of others. And we have a responsibility to incarcerate violent offenders who do not respond to treatment—those constitutionally incapable of responsibility—to protect the responsible agency—and other dimensions of agency and wellbeing—of others. 

This explains how social deficits impair rationality. On the other side, socializing enhances rationality—and hence responsibility. For instance, collaborating with others on reasoning tasks tend to improve problem-solving outcomes (e.g., Laughlin 1965; Davis 1969; Maier 1970; Hill 1982; Shwartz 1995—cited in Doris 2015). Collaboration, in other words, enables us to be optimally rational. In turn, collaborating helps us to be optimally responsible, since it helps us to reason effectively and objectively (in a non-biased way) about problems that threaten our responsible agency. For example, if I am trying to decide how to address a student with putative behaviour problems in such a way that I am able to serve the rest of the class and the students’ interests at the same time—which is not an easy matter—it might serve me to consult with the student, the university’s counselling and disability services centre, and other relevant people, before coming to a decision about how to address the issue. Collaborating with the right people can help me come to a responsible decision in the matter—a decision that enables me to fulfil my role-responsibility as a teacher, in a way that also enhances the student’s responsible agency and preserves the baseline rate of responsible agency in the class.

Collaboration is thus good for responsibility. But not just any kind of collaboration is helpful, and this is important to note. First of all, we should collaborate with the right people—people who are implicated in the problem or have relevant expertise. But aside from this matter, we also want the collaborators to have the right demographic status and the right mindset—a point that should not be overlooked. Just saying that ‘collaboration enhances rationality’ is a misnomer, since some groups are incapable of cooperating and respecting each other’s contributions. We now know, on the basis of multiple research project on workgroup problem-solving, that demographic diversity within the workgroup’s composition improves the group’s performance on cognitive tasks. Early research, however, found exactly the opposite: that diversity decreased reasoning efficacy. This is because the early research failed to control for intervening process variables, such as the attitudes or ‘mindsets’ held by the group members. More modern studies, supported by comprehensive literature reviews, find that diverse group composition predictably improves reasoning outcomes, but only if certain group dynamics are in place—factors such as low ‘subcategorization’ (e.g., implicit bias), high cooperative interdependence, and high ‘diversity mindset’ (i.e., an appreciation of the value of diversity) (Kippenberg & Schippers 2007). This research lends support to the claim that collaborating can serve our responsible agency, as per the teaching example above, but we should be mindful to collaborate with the right people under the right epistemic conditions—conditions of receptivity to everyone’s epistemic authority. 

To conclude this section: I submit that our rational goals are social goals, and thus relational goals. And responsibility is a function of effectively pursuing these rational-social goals—hence responsibility is relational. We cannot be ideally responsible, across the full panoply of our role responsibilities, if we shun others, discriminate against others, find ourselves (due to bad luck or by choice) in conditions of social isolation, deprivation, and social injustice, or if we destroy the planet or decimate the earth’s resources. While cataclysmic climate change will undermine our responsible agency in deep, potentially irreparable ways, more intimate harms, like discrimination, can also hamper our prospects for responsible agency.

One of the main implications of seeing responsibility as relational is that we all have an interest in distributing resources and care in such a way as to enhance the responsibility of every member of society, perhaps especially the least well-off, because responsibility-deficits do not remain isolated and cloistered: they spread around. There are no ‘gated communities’ for responsible agency: we’re all in the same boat.

One potential objection to this view is that a total recluse could be responsible, in spite of being completely socially isolated. I think that this is impossible: there is no responsible social isolate. This is because the social isolate is not capable of taking on social roles constitutive of responsibility; he is capable, at best, of a very limited range of capacities: perhaps obtaining food and shelter and feeding himself, but not much else. This person cannot obtain responsibility as I envision it. He is far from being able to acquire what I have been calling ‘ideal responsibility,’ viz., a range of context-sensitive social capacities that respond reliably to environmental cues. Moreover, I suspect that the complete social isolate is a fiction—though one that we are familiar with from the myth of the ‘state of nature’ and Romulus and Remus raised-by-wolves myths. In reality, humans evolved to live in societies. Hence, every real human being depends on community, and the products of joint efforts, for survival. Underestimating this fact can be deadly. The naïve and unfortunate John McCandless trekked into Alaska with nothing but a bag of rice, and summarily died of food poisoning (as described in ‘Into the Wild’ by Krakauer: 1993). Every successful ‘survivalist’ depends on human agricultural and manufactured products to survive in isolation; thus, even if they remove themselves from human society, they depend on the proceeds of this society for survival in the wild. But these wilful social isolates are not, on my view, as robustly responsible as members of society.

Now, one might alternatively object that philosophers already (tacitly?) take responsibility to be relational, so I’m not making a substantive point. But if they do, then why do so few responsibility theorists talk about relationships and social structures, and their impact on responsibility—aside from the many ‘nefarious neuroscientist’ examples?

In what follows, I intend to shed a spotlight on the role of the social in responsible agency. I will focus in particular on epistemic justice (and lack thereof).

Relational Responsibility: Introduction

 

I’m going to post a series of blog entries that argue that responsibility is relational. What does this mean? It means, first and foremost, that responsible agency depends upon responsibility-conducive social relationships and structures. These structures include conditions of epistemic justice—conditions free from pernicious stereotypes. Epistemic injustice (EI) impairs the capacities implicated in responsible agency (responsibility-constitutive capacities) in people living in these epistemic conditions. It undermines these capacities in privileged and disadvantaged groups alike, albeit (often) in different ways. In privileged group members, EI tends to induce implicit and explicit biases that prevent people from making rational responsibility attributions; and in disadvantaged group members, it tends to induce stereotype threat and to impose material disadvantages that prevent people from exercising responsibility to the fullest possible extent, and from effectively defending themselves against misplaced blame and censure. EI also undermines the confidence with which we can hold people responsible, since it contaminates the epistemic background against which we make responsibility attributions. This means that our reactive attitudes can only be, at best, weakly justified. In these ways, EI undermines our responsible agency, and creates inequalities our shared set of responsibility practices. 

A quick note: I’ve been working on relational responsibility for some time now, but the ideas in these posts were to some extent influenced by talks given at the ‘Extended Responsibility’ workshop that I organized at the Practical Justice Initiative at the University of New South Wales in November of 2016. The speakers were Alicia Coram, Paul Formosa, Katrina Hutchsion, Catriona Mackenzie, and Gualtiero Piccinini. Catriona Mackenzie is working on a relational account of responsibility, and Katrina Hutchison is writing about oppression and how it affects our responsibility practices, and these ideas happen to intersect with some of my own.

 

Help from the Hive Mind

I am preparing for a new course this semester on philosophy and cognitive science. In the third and final part of the class, we are going to read and discuss Moral Machines: Teaching Robots Right from Wrong by Wendell Wallach and Colin Allen (OUP 2009). Needless to say, I am looking forward to that stretch of the course. In the meantime, reading through some of the recent work in robot ethics got me thinking about compatibilist (and semi-compatibilist) accounts of moral desert that appeal to something along the lines of Fischer and Ravizza's reasons responsive mechanisms as a way of grounding moral responsibility even in the face of determinism (and even, perhaps, in the absence of free will).

Wallach and Allen talk about machines that have the capacity to run "ethical subroutines." These subroutines are sensitive to morally salient features of various situations and circumstances--e.g., people's emotions as expressed via their facial expressions. As such, these routines could enable the robot to navigate our moral landscape by making computational decisions that are informed by moral principles--e.g., never harm humans unless harm is unavoidable, in which case, minimize harm--and that are sensitive to morally relevant features of the world. These are so-called "moral machines" (see some nice popular press stuff on related issues in robot ethics here, here, and here)--a morality which Wallach and Allen plausibly claim comes in degrees.

I  have two questions for the readers:

First, are these kinds of moral machines--with their ethical subroutines--morally responsible if we adopt a reasons responsive view of agency and desert? Or would we just say that the programmer is responsible for the subroutine?  If the latter, how do we define "reasons responsive mechanism" in a way that can't be boiled down computationally into something like a mechanism that runs an ethical subroutine? Second, can people point me in the direction of any recent work that addresses these sorts of issues at the intersection of machine ethics and moral agency? See below the fold for some papers that will now be on the reading list!

Plan Rationality

I have been discussing a planning agent’s self-governance, both at a time and over time. To begin this third and final post (though of course I’d be pleased to respond to further queries/challenges) I’d like to back up a bit. The planning theory behind these models of self-governance supposes that a planning agent’s practical thinking is guided by norms of synchronic plan consistency and means-end coherence, as well as by a diachronic norm of stability of intention. And it sees these as norms of practical rationality. But this can seem puzzling. First, we don’t think ordinary desires need to be consistent or means-end coherent. What is special about intentions and plans? Second, these norms may seem simply to favor forms of, in the words of Niko Kolodny, psychic tidiness. But why think such tidiness is a big deal?

It seems plausible that a general disposition to conform to these norms as an aspect of one’s planning agency will have significant benefits, especially given the important coordinating roles of our plans. But what we have learned from J.J.C. Smart is that an inference from a pragmatic defense of a general disposition of thought to a claim of irrationality in the particular case is fraught. So we need at least to supplement such a two-tier pragmatic strategy.

These questions in normative philosophy have implications for descriptive/explanatory philosophy of action. The thought (championed by Joseph Raz and Niko Kolodny, and anticipated by Hugh McCann in the 1980s) that these norms are a “myth” would tend to undermine our confidence that structures of planning – as understood within the planning theory – are basic for our descriptive/explanatory philosophy of action.

My response begins with two ideas: (i) Intentions and plans have distinctive coordinating, organizing roles in our individual and social lives. In part because of these roles, they help constitute where the agent stands, and so play an important role in self-governance (as discussed in my first post). (ii) The rationality norms at issue track conditions of a planning agent’s self-governance: plan inconsistency or means-end incoherence normally baffle a planning agent’s synchronic self-governance; and certain kinds of plan instability normally baffle a planning agent’s self-governance over time.

I argued for this claim about the synchronic norms in a 2009 essay. The idea that a diachronic norm of intention stability tracks conditions of diachronic self-governance is more difficult. Here I think we need to be prepared to adjust our formulation of the norm in the light of our account of a planning agent’s diachronic self-governance. Nevertheless, one idea that will emerge, given our approach to diachronic self-governance, is that the shuffling between non-comparable options discussed in the first post will be in violation of a diachronic norm that tracks conditions of diachronic self-governance.

But now we need to ask: How does the claim that these norms track conditions of self-governance – call this the tracking thesis -- show that they are norms of practical rationality?

Here again some might be tempted by talk of the constitutive aim of agency and try to argue that this aim is self-governance and that this explains why norms that track self-governance are norms of practical rationality. However, as I said earlier, this seems to overburden our descriptive and explanatory understanding of our agency. But then how does the tracking thesis explain why these are norms of plan rationality?

My tentative answer (sketched in a forthcoming paper and a forth-existing essay) is that it does this as part of an overall and reflectively stable understanding of these norms, an understanding that is available to a reflective planning agent. A planning agent who reflects on the basic norms that guide her plan-infused practical thinking would see the tracking thesis as part of the best explanation of those norms.

This explanation would involve the following ideas: (1) The tracking thesis shows that these norms do not merely track mere mental tidiness. (2) The tracking thesis articulates an overarching commonality across these norms, a commonality in light of which they make more sense. (3) The tracking thesis shows that if (a) one has a normative reason in favor of governing one’s own life, then (b) if one has the capacity for relevant self-governance, this reason transfers to a reason in favor of conformity to the norms in the particular case. (4) A conclusion I reached in my second post is that the end of diachronic self-governance is essential to the general exercise of a planning agent’s capacity for diachronic self-governance. So, suitably generalized, the tracking thesis supports the claim that the presence of the end of diachronic self-governance is itself enjoined by diachronic plan rationality. (5) Given this end, and the plausible assumption that diachronic self-governance is a good thing, a planning agent with the capacity for diachronic self-governance will have a normative reason in favor of diachronic self-governance, and so in favor of the involved synchronic self-governance. (Here I assume that the idea of a normative reason that is germane to these reflections of a planning agent will involve both a connection to her ends and the desirability of those ends.) (6) So given this end, (3)(a) is true. (7) In this way the tracking thesis supports (3)(b), and thereby the claim of a reason of self-governance to conform to the cited norms. (This is a response to John Broome’s and Niko Kolodny's question about whether there is reason to be rational.)

My idea, then, is that a planning agent who reflects on the basic norms that guide her plan-infused practical thinking would see the tracking thesis, together with (1)-(7), as part of the best explanation of these norms. It remains possible for her to reject this entire package. But then she would not be in a position to defend in this way central modes of practical thinking characteristic of her planning agency. In the absence of some other defense, she would then be under pressure to give up these modes of thinking -- though this is something she has pragmatic reasons not to do, and is also unlikely to be something she can do simply at will.   Short of this, however, she is in a position to see that her plan-infused practical thinking can be embedded within a framework that is available to her and that provides support for the application of the cited norms to particular cases, support that is over and above (but compatible with) pragmatic support for relevant general modes of thinking. So it will be reasonable – even if not strictly inescapable -- for her to retain her plan-infused practical thinking and its associated norms as part of this framework. So we need not worry that these norms are reflectively unstable in a way that would challenge the planning theory.

And – in closing – let me again thank Thomas Nadelhoffer for making it possible for me to try out these ideas on this blog.

Related articles

Self-governance Over Time - Flickers of Freedom

Doxastic Self-Control

Willpower and the End of Diachronic Self-Governance

In this second post I want to reflect on issues about willpower in light of my sketch, in my previous post, of an approach to a planning agent’s self governance over time. (In the background is Holton 2009.)

Suppose you know you will be tempted to drink a lot at tonight’s party. Since you now think that would not be a good thing to do, you now form the intention to stick with one drink at the party. You know, however, that at the party your judgment will temporarily shift and you will at least initially judge that it would be better to have many drinks – though you also know that if you did give into this temptation you would later regret that.

If you stick with your prior intention there will be an important continuity in intention; in contrast, if you give into temptation there will be a break in this intention continuity. This suggests that such willpower may be a form of diachronic self-governance. However, continuity of prior intention helps constitute a planning agent’s diachronic self-governance only given synchronic self-governance at times along the way. But if your judgment shifts at the time of the party it seems that sticking with your prior intention would not be a case of synchronic self-governance: given the judgment shift, where you stand at the time of the party seems to favor drinking more.

Is there an end of the agent’s that favors relevant intention continuity and thereby can help re-shift her standpoint at the time of the party back in favor of stopping with a single drink, as she had intended? This would open up the possibility that sticking with the prior intention is indeed a case of synchronic self-governance, and so also potentially a case of diachronic self-governance.

What end? We don’t want to appeal simply to an end of intention continuity. (An idea along these lines is in work of Jordon Howard Sobel and Wlodek Rabinowicz.) Such an appeal would face a diachronic version of a worry (from Niko Kolodny) that we are simply appealing to a kind of mental tidiness. A natural proposal, instead, is to appeal to the end of diachronic self-governance itself – where that is not merely a matter of mental tidiness. This end would potentially help support willpower in the face of temptation, since such willpower would involve a continuity of intention that is itself a potential element in diachronic self-governance. And given the support from this end, the agent’s standpoint at the time of the party might well shift back into favoring willpower in a way needed for synchronic self-governance (and so for diachronic self-governance).

Indeed – as Sarah Paul has emphasized in conversation -- problems of temptation, and related problems about procrastination, pervade our lives. This supports, by way of a kind of inference to the best explanation, the conjecture that the presence of this end of diachronic self-governance is a central element in ubiquitous cases of a planning agent’s diachronic self-governance.

This suggests that the end of diachronic self-governance is essential to the general exercise of the capacity for such self-governance. And since diachronic self-governance essentially involves synchronic self-governance at times along the way, it is plausible to infer that this end of diachronic self-governance will induce an end of synchronic self-governance. So an end in favor of synchronic self-governance will also be an element in the general exercise of the capacity for such self-governance.

Some may think that the end of self-governance is a kind of necessary, constitutive end of agency quite generally. But this seems to me to overburden (to borrow a term from Barwise and Perry) our descriptive and explanatory understanding of our agency: there are just too many cases of our agency that seem not to involve guidance by that aim. The connection to an end of self-governance instead goes by way of our model of self-governance over time. Further, it is interesting to note that to get to this result we needed to go beyond a focus on synchronic self-governance and turn to diachronic self-governance.

Would this model of a planning agent’s diachronic self-governance entail that sticking with one’s intention to drink the toxin, in Gregory Kavka’s case, is in the same way a candidate for diachronic self-governance? Not if we are careful in our formulation of the model of self-governance over time. In the temptation case one anticipates later regret at giving into the temptation. So the discontinuity of intention involved in giving into temptation is a discontinuity both with earlier resolution and with expected later assessment. In contrast, the discontinuity of intention involved in abandoning, when the chips are down, the prior intention to drink toxin is only a discontinuity with earlier intention – at least on the plausible assumption that one expects that would later be glad that one did not drink the toxin. We need to be careful to understand the intention-continuity aspects of diachronic self-governance in a way that tracks this difference.

Related articles

Self-governance Over Time - Flickers of Freedom

Doxastic Self-Control