A Metaphysics of Dehumanization

3.1 Constitutive Dehumanization

Recall that social kinds are constructed and maintained in part through collective recognition of both constitutive and deontological rules. For the human, the constitutive rule in effect globally and in many local societies is that all and only members of the species Homo sapiens (bar fetuses) qualify as human. Constitutive dehumanization refers to acts and practices that aim to change this rule. This is typically done through public avowals that certain kinds of people are not really human. For instance, when Donald Trump said of asylum seekers “You wouldn’t believe how bad these people are. These aren’t people, these are animals” (cited in Davis 2018), this was a prime example of constitutive dehumanization.

A useful way to think about constitutive dehumanization is as a form of metalinguistic negotiation. When people are engaged in metalinguistic negotiation, terms are used in order to change what they mean:

When a young man in the year 2000 said his biological father’s life partner is his father, he is not to be seen as having uttered a literal falsehood (given what was then the current legal definition of ‘parent’). Instead, he was pragmatically pressing for a change [in] our usage, by drawing attention to the commonalities between this relationship and those that are recognized with the title and given the relevant social, legal, and economic benefits. (Thomasson 2017, 14)

Conversely, when Donald Trump denied the humanity of asylum seekers, he was not making a statement that was intended to be taken literally, but was rather pragmatically pressing for a change in who the term “human” applies to.

While Thomasson is concerned here with negotiations over what terms mean, her point also has significant metaphysical implications. When it comes to social kinds, how we talk about them is a significant part of how we construct and maintain them. When we say “he is not an animal,” we are reinforcing norms concerning the boundaries of the human.¹⁴ Conversely, then, when someone declares of a whole class of people that “they are not human,” the point is not just to alter the meaning and usage of the term human, but also to challenge the standing of certain people as human — it is an attempt to change the constitutive rules for who counts as human.

3.2 Corrosive Dehumanization

While constitutive dehumanization is the most easily recognizable form of dehumanization, corrosive dehumanization is the more pervasive. Whereas constitutive dehumanization directly targets the constitutive rules maintaining the boundaries of the social kind human, corrosive dehumanization seeks to change the constitutive rules indirectly, by challenging to whom the deontological rules and norms of treatment apply. This point echoes Searle’s observation about the durability of institutions:

Cars and shirts wear out as we use them but constant use renews and strengthens institutions such as marriage, property, and universities. […] [E]ach use of the institution is a renewed expression of the commitment of the users to the institution. Individual dollar bills wear out. But the institution of paper currency is reinforced by its continual use. (Searle 1995, 57)

Social kinds are maintained through use. Conversely, social kinds are eroded through failure to treat their members in the appropriate ways.

Crucially, though, a social kind may not stand or fall as a whole, but can instead splinter through failure to use only certain elements of it. For instance, paper currency as a whole could survive and thrive in a particular community, even while widespread refusal to accept $1 bills leads those notes to lose their power as money. Analogously, widespread refusal to accept that certain human beings have human rights, or ought to have their death marked in some way, can lead to those people losing their power as humans, and eventually their membership in the human kind.

Corrosive dehumanization can be more or less explicit, but it always constitutes a denial that certain individuals possess the powers associated with membership in the human kind, and/or that they ought to be treated as humans. An example at the more explicit end of corrosive dehumanization is the Bush government’s stance on detainees at Guantanamo Bay: one of their key arguments in court was that these individuals did not possess any human rights whatsoever.¹⁵ Since one of the central deontological rules of the human, at least in liberal democratic societies, is that all humans possess human rights, such statements function to erode membership in the human for those accused of terrorism.

Another example of explicit corrosive dehumanization lies closer to our philosophical home. Consider the norm that human life is grievable¹⁶ — that part of what it is to be human is to be such that one’s death is a significant event, one that ought to be marked in some way. Then consider the way that some philosophers talk about individuals with significant cognitive impairments:

[O]ur traditional beliefs about the special sanctity of the lives of severely retarded human beings will have to yield. […] [A]llowing severely retarded [sic] human beings to die, and perhaps even killing them, are […] somewhat less serious matters than we have believed. (McMahan 2002, 230)

Such statements are explicit attempts to shift norms about how we respond to the deaths of some people. In contemporary liberal democracies, where the way in which we respond to death is part of how we recognize someone as human, disputing these norms functions to remove certain categories of people from the human kind.

Moving toward less explicit forms of corrosive dehumanization, we can look to forms of hate speech. In her paper “Genocidal Language Games,” Lynne Tirrell (2012) examines the role of dehumanizing language in inciting genocidal violence in Rwanda. As she explains, how we speak about people licenses changes in how we treat them, owing to the inferential role of speech acts. To call someone an inyenzi (cockroach) carries with it a series of inferences about how that person ought to be treated:

What are some of the inferences we can make about calling A ‘cockroach’? Common inferences include that cockroaches are pests, dirty, ubiquitous, multiply rapidly, are hard to kill, ought to be killed, show emergent tendencies when in groups, are resilient, carry diseases, can go long periods without food or water, tend to only emerge at night when they are hard to see. (200)

Those using and hearing these terms need not interpret them literally — they need not believe that Tutsis actually were, biologically speaking, cockroaches. Rather, the persistent and repeated use of such terms leads people to believe that Tutsis ought to be treated as cockroaches.

Shifting from a linguistic to a metaphysical key, these kinds of inferential moves challenge the standing of Tutsis to be treated as human, because they construct a different, and conflicting, set of norms to apply to them. As these norms get uptake, they make it much more difficult to simultaneously maintain recognition that Tutsis are human. Constructing Tutsis as cockroaches involves deconstructing them as human.

While incitement to genocide makes vivid how dangerous corrosive dehumanization can be, the phenomenon can be much subtler. At its most implicit, corrosive dehumanization occurs simply through failure to treat someone in accordance with the rules for how humans ought to be treated. Whether by violating an individual’s human rights, enacting institutional systems that treat them as a number rather than a person, responding to their death as if it were insignificant, or caging them as if they were a non-human animal, such actions communicate that the person subject to them is not worthy of being treated as human, and hence is not really a member of the human kind. Especially when such treatment is regularly inflicted on certain kinds of people, it chips away at their membership in the human, because they can gradually cease to be seen as a member of the group to whom such norms apply. As Andrea Dworkin puts it, “those who can be used as if they are not fully human are no longer fully human in social terms” (Dworkin 1985, 15).

It is important to reiterate that the phenomenon of corrosive dehumanization is orthogonal to the attitudes of the person or persons responsible for the action, because what makes an act or practice a case of corrosive dehumanization is the effect it has on the collective recognition of the target as human. This focus on effects, rather than attitudes, helps to draw attention to the possibility of dehumanization through neglect or indifference, rather than hate or fear. A case in point here is the treatment of the elderly, especially those in institutional care facilities. Consider this exchange about an elderly patient with advanced dementia at a hospital in South Carolina: “The intern’s instructions to me were roughly this: ‘Think of it this way. She’s a plant; you’re the gardener; your job is to make sure she is watered’” (Elliott 2003, 18). Such attitudes are far from isolated: for instance, a recent Royal Commission into Aged Care in Australia found, among other problems, that care homes are routinely resorting to the use of physical or chemical restraints to keep elderly residents — especially those with dementia — immobile and thus out of harm’s way.

Residents in aged care homes, their families, and those who care for them commonly turn to the language of dehumanization to describe their experiences. One resident’s submission to the Royal Commission explained how it felt to be left waiting for hours for the assistance necessary to perform basic tasks, such as going to the toilet:

When neglected like that, I feel I have been dehumanised: left as a carcass in an aged care abattoir; ready to be processed like a slab of meat in a sausage processing factory at some future time. (Royal Commission into Aged Care Quality and Safety 2021, 613, Vol 4b)

Likewise, a journalist reporting on aged care describes what she found as follows:

“The staff call them the Os and the Qs,” says a seasoned nursing home visitor, describing residents with their mouths hanging open and those with their tongues hanging out. The staff mean no disrespect, but reducing someone to a letter of the alphabet is just one example of the unconscious dehumanising that happens often in the treatment of the elderly. (Baum 2018)

What we see with aged care is a situation in which a particular subset of the community is regularly treated in ways that violate accepted norms for how humans ought to be treated. In the absence of vocal and sustained pushback, the broader community’s recognition of the elderly as people who ought to be treated in certain ways — our recognition that the elderly are human — starts to erode.

We have seen that corrosive dehumanization covers a wide spectrum of acts and practices, all of which function to erode their target’s membership in the human kind. At one end, we find explicit calls for certain kinds of people to be denied the powers attached to the human, and/or explicit denials that particular norms ought to be extended to certain kinds of people. Shifting to a slightly less explicit key, we find language games that describe people in ways that imply it is appropriate to treat them in certain ways — ways that are incompatible with the norms of how humans ought to be treated. Finally, we come to the most implicit, but most pervasive, form of dehumanization: the routine failure to treat certain people in accordance with the powers or norms appropriate to their membership in the human kind.

To be clear, this spectrum does not reflect the level of corrosiveness of different forms of dehumanization — I am certainly not suggesting that philosophical discussions of disability are more dehumanizing than incitements to genocide! How corrosive a particular act or practice will be depends not on how explicit it is, but rather on the vulnerability of the targeted group, the centrality of the right or norm being violated to our shared understanding of what it is to be human, and, most importantly, the power and influence of those engaged in the mistreatment. Not every member of the community is equally positioned to contest social norms, and not everyone’s failure to extend the appropriate norms to some will affect how others see them. The actions of some members of the community can thus have a disproportionate effect on collective recognition (Bicchieri 2017, esp. Ch 5). Those acting in the name of the state, in particular, wield immense power to shape social norms.

3.3 Hermeneutical Dehumanization

The final form of dehumanization to consider is hermeneutical dehumanization. Unlike constitutive and corrosive dehumanization, hermeneutical dehumanization focuses not on the constitutive or deontological rules attached to the kind but rather on the social imaginary that shapes people’s explicit and implicit beliefs about the human. Hermeneutical dehumanization occurs when the social imaginary forecloses certain ways of being human. In other words, the social imaginary accommodates only a limited number of prototypes of the human, which are non-exhaustive of the identities people in fact hold.

Hermeneutical dehumanization offers a way to understand a particular kind of harm experienced by gender diverse people. If, in the social imaginary of a particular community, to be human one must be either a man or a woman, then those who are not easily read as exclusively either men or women — or who explicitly reject those categories — will not be seen as human. And, I posit, many of us do in fact live in communities with such a social imaginary. Consider, for instance, the question that is invariably asked upon the birth of a child: “boy or girl?” Such a question takes for granted that every child will be one or the other, and that the answer can be read off the child’s body at birth. The options of boy or girl thus exhaust the imagined possibilities for a human child.

Such tendencies to conceive of the human in binary gender terms are sustained by the social environment. As Sally Haslanger (2012) points out, the social imaginary (or, in her words, narratives and symbols), functions in feedback loops with other factors, such as institutional roles. We can see that in play here; there are a wide variety of institutions that construct and maintain the human, including registries of births and deaths, national censuses, schools, hospitals, welfare departments, and licensing bodies. Until very recently — and in all too many places still — such institutions required those they interact with to be registered as either male or female. Our daily interactions with these institutions, and our experiences of their categorizations of us, function to sediment the gender binary in the social imaginary of the human, thus excluding gender diverse people.

When a society lacks the resources to straightforwardly imagine some of its members as human, it is unsurprising to find it dehumanizing them in other ways, and this pattern is tragically evident in the injustices and indignities faced by trans and gender diverse people in many societies. However, hermeneutical dehumanization should not just be evaluated in terms of its propensity to trigger other, more visible, forms of dehumanization: how the social imaginary represents people has a profound effect on who we take ourselves to be and our sense of self-worth. Hermeneutical dehumanization draws attention to the harms that are inflicted when the social imaginary fails to expand to make room for all ways of being human.

4.1 Mikkola on Dehumanization

Mikkola’s account of dehumanization is developed in the context of a broader theory of social justice. More precisely, for Mikkola dehumanization provides the normative framework for explaining the wrong of social injustice: as she puts it, “Dehumanization is not another form of injustice — rather, it is that which makes forms of social injustice unjust” (Mikkola 2016, 2). To establish whether or not an act or treatment constitutes social injustice, then, we need to identify its wrong-making feature — and that wrong-making feature is what Mikkola calls dehumanization. An act or treatment qualifies as dehumanization, for Mikkola, “only if it is an indefensible setback to some of our legitimate human interests, where this setback constitutes a moral injury” (8).

The final piece of the dehumanization puzzle concerns what it means for a setback to be indefensible. Here, the fact that Mikkola is aiming for a theory of social injustice becomes crucial, as it leads her to build in a requirement that the act or treatment target a social identity, such as race or gender:

Delimiting indefensible interest violations is not just a matter of focusing on harms experienced. Whether some interest violation is indefensible or not trades also on whether the violation involves some delineation or ‘flavor’ of injustice that is associated with social identity groupings (like racism). (224)

Putting the pieces of her account together, an act qualifies as dehumanization if and only if it sets back the target’s legitimate interests, and is to be explained by the target’s social identity.

I take Mikkola to have provided a compelling theory of social injustice. Where we part ways, and where I think my account is preferable, lies specifically in what each of us picks out with the term “dehumanization.” Because of its role in her larger theory of social injustice, Mikkola’s account of dehumanization is both too broad and too narrow: it classifies as dehumanization phenomena that I think are better understood as non-dehumanizing forms of social injustice; and it fails to classify as dehumanization phenomena that I think ought to be understood as dehumanization in the absence of social injustice.

In terms of the former, Mikkola argues that prejudicial discrimination that sets back the targets’ interests always counts as social injustice. Examples she gives include “social mechanisms like women being left out of professional networks or being ignored in meetings and thus stymied in careers in certain male-dominated fields” (194) or a decision not to hire a single mother on prejudicial grounds (209). While I agree that these qualify as examples of social injustice, for Mikkola they must also qualify as dehumanization — that is what makes them wrongful, and hence instances of injustice. On my account, by contrast, there are myriad forms of social injustice that fall short of dehumanization, because the latter requires the violation of powers or norms attached to us qua humans.

Conversely, though, my account allows for the possibility that an individual can be dehumanized even when the treatment is not motivated or explained by their social identity. This will admittedly be rare — I accept that most examples of dehumanization do involve despised minorities — but it is far from inconceivable. Consider, for example, the brutal murder and dismemberment of the journalist Jamal Khasshogi by the Saudi government. Such action surely qualifies as dehumanization, especially given the treatment of the corpse; however, it would not be accommodated by Mikkola’s account. On my account, by contrast, it would constitute an instance of corrosive dehumanization: a powerful agent is treating an individual in such a way as to publicly communicate that the norms appropriate to the human do not apply to them.

In allowing social injustice and dehumanization to crosscut one another, my account makes room for the concept of dehumanization to do important diagnostic and normative work over and above what can be done by the concept of social injustice. This gives it distinct advantages over a theory, such as Mikkola’s, which takes dehumanization to be co-extensive with social injustice.

4.2 Smith’s Account of Dehumanization

Let us turn now to Smith. The central idea in Smith’s (2020) account is that dehumanization equates to taking another human being to have a sub-human essence: it is a matter of seeing someone as other than human. As he puts it, “dehumanization [is] a kind of attitude — a way of thinking about others. To dehumanize another person is to conceive of them as a subhuman creature” (2020, 19, emphasis added).¹⁷ Smith appeals here to the notion of psychological essentialism, which is the folk tendency to attribute some kind of mysterious essence to natural kinds that explains why they are the way they are and which is “unchangeable and irrepressible” (2020, 68).¹⁸ For those in the grip of psychological essentialism, “what makes a cat a cat is its possession of an inner cattiness — the cat essence” (2020, 67). Likewise, “[r]aces are supposed to be natural human kinds with distinct racial essences” (ibid), such that even someone who displayed no outward indication of being a particular race could nonetheless be one. Applying this theory of folk essentialism to the human, Smith explains

When we dehumanize people, we think of them as apparently human on the outside, but essentially subhuman on the inside. From the dehumanizer’s perspective, dehumanized people are subhumans passing as humans, because their humanness is only skin deep. Furthermore, dehumanized people are thought to be irredeemably sub-human. It is their permanent condition. Although these subhumans may be very good at mimicking true human beings, this is merely a façade and they are always on the verge of reverting to type. (2020, 69)

In sum, for Smith, those who are dehumanized are perceived as lacking a human essence and as possessing instead a subhuman essence.

One of the important advantages of Smith’s account over rival psychological approaches is its ability to deal with a common objection. Standardly, psychological approaches define dehumanization in terms of failing to perceive the target as human, where this is understood in terms of them lacking particular cognitive capacities taken to define humanity. For instance, according to Nick Haslam, one of the most prominent psychologists working on dehumanization, “[a]ny understanding of dehumanization must proceed from a clear sense of what is being denied to the other, namely humanness” (Haslam 2006, 255). Haslam goes on to define humanness in terms of both characteristics unique to humans, such as cognitive sophistication and culture, and characteristics reflective of human nature, which include more capacious capacities such as agency and emotional responsiveness. As a number of critics (Smith 2014; Manne 2016; Appiah 2008) have pointed out, however, if we define dehumanization in terms of failing to see others as possessing these types of cognitive capacities, many paradigmatic instances will fail to count as dehumanization. Appiah puts the point especially well:

The persecutors may liken the objects of their enmity to cockroaches or germs, but they acknowledge their victim’s humanity in the very act of humiliating, stigmatizing, reviling, and torturing them. Such treatment — and the voluble justifications the persecutors invariably offer for such treatments — is reserved for creatures we recognize to have intentions and desires and projects. (2008, 144)

Smith avoids this problem, because the attribution of a sub-human essence is fully compatible with seeing the other as possessing any — or even all — of the cognitive capacities typically associated with human beings.

While Smith avoids the most common objection to psychological approaches, there are other limitations to his account. The first of these mirrors an objection Kate Manne has raised against what she calls the naïve approach to misogyny, which construes misogyny in terms of the hateful attitude individuals have toward women. As she points out:

[W]hat lies behind an individual agent’s attitudes, as a matter of deep or ultimate psychological explanation, is frequently inscrutable. So the naïve conception would threaten to make misogyny very difficult to diagnose, short of being the agent’s therapist (and sometimes not even that would be sufficient). This would threaten to make misogyny epistemically inaccessible to women, in particular. That is, it would threaten to deprive women of the wherewithal to acquire knowledge and justified beliefs about the manifestations of misogyny that they may encounter, and to go on to make warranted assertions on that basis. So in effect, this notion of misogyny would be silencing for its victims. (Manne 2017, 44)

The same is true, mutatis mutandis, for dehumanization. For an individual being subject to degrading torture, why should it matter whether the torturer sincerely believes her to have a subhuman essence or is instead just callously pursuing his own political ends? And, more importantly, why should her claim to have been dehumanized be contingent on what was happening in her torturer’s head?

Alongside these epistemic and contingency worries, Smith’s commitment to defining dehumanization in terms of the attitudes of perpetrators also renders his account unsuitable for analyzing cases that are systemic, rather than interpersonal. For Smith, the result of attributing a sub-human essence to individuals who have the outward trappings of humans is a sense of contamination:

When people dehumanize others, they think of them as both human and subhuman at the same time, and as violating the categorical distinctions that underpin the natural and social order. That’s why dehumanized people are seen as harbingers of disorder, pollution, and disease. And even though these people are almost always marginalized and vulnerable, they’re depicted and treated as though they are profoundly threatening — thus justifying the violence against them. (Smith 2020, 155)

While such phenomena are undoubtedly central to cases such as lynching and genocide, it is far less clear that they are relevant to other cases that we can have good reasons for classifying as dehumanizing.

In the previous section, I suggested that the kind of neglect commonly seen in aged care facilities constitutes dehumanization. However, such neglect is not due to those working with the elderly, nor those in positions of power, attributing any particular kind of essence to them. Contra Smith (2020), dehumanization is occurring here without propaganda encouraging us to see the elderly as threatening or contaminating. Rather, their treatment is the predictable result of a series of economic decisions — in particular, the privatization and deregulation of care homes, and the increasing prevalence of casual and insecure work in the sector — coupled with indifference and/or incompetence on the part of anyone in a position to remedy it.

The metaphysical account sketched above has the resources not only to deem such cases dehumanization, but also to explain what it is about these situations that sees people so often reach for the language of dehumanization in describing them. To be human, on my account, is to be a member of a social kind; and to be a member of a social kind is, among other things, to be collectively recognized as having certain powers and to be owed certain kinds of treatment. While it is highly uncommon for people to consciously reflect on the social construction of the human, we are nonetheless typically highly attuned to the powers, norms, and scripts through which it is constituted. When we hear of an elderly person being described as “a plant” or an “O or Q”, or when we see an elderly person rendered placid and immobile through chemical restraints to make her carer’s job a little bit easier, there is a striking sense that this is not how we treat a human being. This sense explains the propensity to describe such actions as dehumanizing, and the metaphysical approach provides a theoretical framework to vindicate such descriptions.

I have argued that my metaphysical approach is better equipped than psychological approaches, such as Smith’s, to identify the continuities between deliberate denigration and systemic neglect. A further advantage emerges when we look at the way Smith’s account of dehumanization deals with issues like misogyny and transphobia.

For Smith, dehumanization occurs via a process of racialization:

When people racialize another group of people, they have to do two things. First, they’ve got to attribute a racial essence to every member of the group — one that’s not shared by any other group. And second, this essence has to be seen as transmitted by descent, and when that group of people is dehumanized, their imagined racial essence becomes an imagined subhuman essence that’s transmitted by descent from one generation to the next. The logic of dehumanization is such that if your parents are subhuman, then you’re subhuman too. (Smith 2020, 180)

Importantly, as gender is not seen as being transmitted by descent, Smith denies that either misogyny or transphobia, no matter how egregious, count as dehumanization. My metaphysical account, by contrast, has the resources to identity the various ways in which both misogyny and transphobia can manifest as dehumanization. For reasons of space, I will focus here just on transphobia and, as the connection between hermeneutical dehumanization and transphobia was discussed above, just on corrosive and discursive dehumanization.

It is not difficult to find examples of transphobia that qualify as constitutive dehumanization. For instance, Janice Raymond infamously claimed that “transsexualism” ought to be “morally mandate[ed] out of existence” (cited in Williams 2020, 721). Moreover, one does not have to venture far into online spaces to see trans and gender diverse people described as “Frankenstein” (e.g., Wolfe 2022) or “it” (for discussion, see Richardson-Self 2019) and for trans women breastfeeding described as producing “Zombie milk” (Anon 2019). Such descriptions clearly posit trans and gender diverse people as other than human.

As with most targets of dehumanization, though, trans people are more commonly subject to corrosive dehumanization. From a willingness to violate their most basic bodily rights (especially of trans women of color) to the scoffing reaction that transgender days of remembrance elicit in some circles, many of the rights and norms constitutive of being human are denied to trans and gender diverse people. These rights and norm violations are further compounded by the indifference of the broader community. Responding vocally and forcefully can help maintain social norms despite their violation; failure to respond as if a violation has occurred, by contrast, signals to the community that these norms do not apply to these people. The strategies and imagery used to dehumanize trans and gender diverse people will undoubtedly be different from the strategies and imagery used to dehumanize those who have also been racialized in the ways Smith describes. Nonetheless, my metaphysical account allows us to see the continuities between them in their respective attempts to re-shape norms about who ought to be treated — and recognized — as fully human.

The metaphysical account of dehumanization I have sketched in this paper shifts attention away from the psychology of the perpetrator and toward the social structures within which the human is both constructed and contested as a social kind. The psychology of individuals will, of course, be relevant to these processes, especially when the contestation is deliberate. However, to define dehumanization in terms of psychology is to lose sight of the broader social patterns at play. When we see the human as a social kind, and hence see dehumanization as a range of practices that function to remove certain people from that kind, we are better positioned to identify the continuities between racism and transphobia, hate and indifference, and intentional actions and institutional inertia.

There are myriad ways in which we are held in the human and, likewise, myriad ways in which that membership can be threatened. As valuable as it is to explore the psychology behind the most violent of these, it is also valuable to have a framework that allows us to take a more birds-eye view of the broader phenomena of dehumanization.