Political realism and ideology critique have, as of late, sparked renewed interest within Anglophone political philosophy. Contemporary political philosophy, realists complain, has become excessively grounded in moral philosophy—a branch, some say, of applied ethics (Bernard Williams 2005; Geuss 2008; Rossi and Sleat 2014; Sleat 2018). The latter, in turn, contend that political philosophers should spend greater effort analyzing how ideologies legitimize and help sustain unjust social arrangements (Shelby 2003; Haslanger 2017; Lafont 2024; Celikates et al. 2025). Recently, a growing number of philosophers—realist ideology critics—have attempted to combine both views, claiming that while political philosophers should indeed be interested in ideology critique, they should, in doing so, eschew moral considerations, emphasizing instead ideologies' epistemic flaws—that is, drawing on epistemic, rather than on moral, normativity (Prinz and Rossi 2017; Rossi and Argenton 2021; Aytac and Rossi 2022; Clark 2024).1 The core of realist ideology critique is fairly straightforward. Powerful agents and institutions usually appeal to, and rely on, legitimation stories—ideologies, in a purely descriptive sense. These justifying ideologies might be transmitted and disseminated in different ways. For instance, because they have survived an ideal deliberative procedure, where all viewpoints are adequately represented, and participants abide by stringent epistemic norms. Or, quite often, by dint of the very power of those whose power is being justified in the first place—who, in highly hierarchical settings, are powerful enough to disseminate beliefs supporting the legitimacy of their own power and status, while simultaneously preventing others from effectively contesting such legitimation stories. In the latter kind of cases, which constitute the main focus of realist ideology critique, the relevant legitimating ideologies are viciously circular: justifications of power differentials, that is, become widespread precisely in virtue of those power differentials. Drawing on extensive historical, anthropological, and sociological evidence, Aytac and Rossi (2022) and Rossi and Argenton (2021) argue that such an account can explain folk intuitions about the legitimacy of private property, as well as the persistence of neopatriarchical social arrangements in the MENA region (Middle East and North Africa). Realist ideology critics argue that we have compelling reasons to be cautious about, distrust, or perhaps even abandon legitimation stories that are viciously circular, as they are epistemically unreliable. In other words, given how we acquired them, we should not expect those beliefs to reliably (i.e., non-accidentally) track the truth. Instead, we should expect them to track the motives and interests of those with the power to disseminate those beliefs without significant contestation. Realist ideology critique, as we will see, admits of radical and moderate declinations, but the above (stylized) characterization should suffice for now. Neither the new political realists—nor, more specifically, realist ideology critics—have given much thought to animals, however. This is surprising for a number of reasons. First, humans exert significant power over animals—whom we routinely kill for food and clothes; to test cosmetics, vaccines, and drugs; to advance scientific knowledge; or, quite often, in the name of tradition. These exercises of power, typically authorized or even actively supported by political institutions, rest to a large extent on legitimation stories involving the superiority and/or specialness of humans. Among those working within the "critical animal studies" tradition—alongside others working on ideology more generally (cf. Mills 2019)—it is common to treat these anthropocentric or speciesist systems of belief as ideologies (cf. Nibert 2002). Second, a growing body of social-scientific literature suggests that those stories are indeed epistemically suspect, being often the result of status quo bias, motivated reasoning, and other unreliable belief-formation processes. Third, and crucially, those legitimating stories can be easily disseminated because those with a stronger interest in challenging them (animals themselves) cannot, for obvious reasons, do so. At first glance, then, speciesist and anthropocentric social arrangements would appear to be every bit as viciously circular as those mentioned above, and thus ripe for the kind of analysis realist ideology critics favor. In this paper, I defend two main theses. First, I argue that realist ideology critique can be fruitfully used to condemn, on epistemic grounds, the exploitation of nonhuman animals—and, more generally, the social arrangements that underpin such exploitation. Second, I also contend that thinking about how to apply realist ideology critique to animals invites us (and suggests ways) to refine and expand the realist ideology critic's toolkit.2 The argument unfolds as follows. I begin, first, by reviewing a model of realist ideology critique developed, in a series of recent papers, by Enzo Rossi, Uğur Aytac, and Carlo Argenton. The Rossi–Aytac–Argenton model, as I shall call it, pinpoints two ways in which power asymmetries can render ideology formation and transmission mechanisms defective. Under conditions of strong social stratification, powerful agents are liable to motivated reasoning, which likely pervades their attempts to justify their own power. Such legitimation stories, moreover, can be easily transmitted since these agents can successfully obstruct significant, widespread, and coordinated contestation. The Rossi–Aytac–Argenton model, I argue, partially applies to speciesist and anthropocentrist social arrangements, insofar as those arrangements depend on legitimation stories that are (i) at least in many cases, traceable to motivated reasoning, and (ii) easy to disseminate, given the power asymmetries existing between humans and the other animals. There is an important difference, however, between the case of adult, able-minded humans—whose contestatory capacities are forcibly hindered by hierarchical social arrangements—and the case of nonhuman animals—who lack the relevant contestatory capacities to begin with, independently of how society is structured. The Rossi–Aytac–Argenton model, I suggest, can be fine-tuned to accommodate this distinction. In the last part of the paper, I argue, more tentatively, that even if some of our beliefs about animals' psychological capacities were not the result of social forces, but hardwired in our evolved psychology, realist ideology critics should nevertheless be concerned with them—a claim I defend by appealing to what I take to be realist ideology critique's main tenets and desiderata. Although this paper discusses a version of political realism, it does not merely address political realists. Others, too, have reason to care about the arguments presented here—especially those working in animal ethics. First, because there is a non-zero chance that political realism might be the correct position. Since almost all normative theorizing on animals relies on moral considerations (in fact, we label the whole discipline animal ethics), the truth of political realism might seem potentially devastating. If what I argue in this paper is correct, however, that might not be the case, for the exploitation of animals would be objectionable on non-moral (i.e., epistemic) grounds too. Under conditions of normative uncertainty, it seems reasonable to explore all the options available, so as to forestall risks. Second, even if we believe that normative political philosophy cannot do without appealing to moral considerations, we can (and should) recognize that particular moral judgments are fallible and quite often liable to (local) moral debunking explanations.3 Thus, we can strengthen our normative conclusions if we appeal both to moral and non-moral normativity. Realist ideology critique offers animal ethicists valuable tools to broaden their argumentative portfolio, strengthening it against potential debunking objections.4 This paper is structured as follows. In Section 1.1, I define political realism and realist ideology critique, and provide a more developed characterization of the Rossi-Aytac-Argenton model. Section 1.2 shows how the Rossi-Aytac-Argenton model can be fruitfully applied to at least some of our beliefs about the moral status and psychological capacities of animals (I focus, to make the argument manageable, on beliefs legitimating meat-eating). In Section 1.3, I turn to an evolutionary account of speciesism, as formulated by some in the literature, to suggest particular ways in which realist ideology can fruitfully target extra-political mechanisms of ideology formation and transmission. Political realism is a broad church, unified by a dissatisfaction with the way in which many contemporary political philosophers are perceived to rely on moral considerations. This complaint is, of course, fleshed out in different ways. For some, it seems to be a problem that political philosophers rely on moral normativity at all (Rossi 2019; Aytac and Rossi 2022, 1216; Kreutz and Rossi 2022) and defend instead that normative political philosophy should instead tap into non-moral sources of normativity—such as, for instance, instrumental and/or epistemic normativity (Burelli and Destri 2021; Aytac and Rossi 2022). Others, in turn, object only to particular ways of grounding political normativity in moral values (Hall and Sleat 2017; Jubb 2019; Sleat 2021). Bernard Williams (2005, 1–2), for instance, famously railed against two models of political philosophy: an enactment model, which sees political institutions as an instrument to realize more fundamental, independently identified moral values, and a structural model, which relies on such values not to provide goals for political action, but side-constraints on its operation. Realist ideology critique, the main focus of this article, is largely animated by the worry that many (perhaps all) of the moral values that we use to vindicate or criticize social arrangements have become widespread not through the uncoerced force of the better argument, as Habermas would put it, but through other, less reputable and often very much coercive means. This worry is nicely summed up in Raymond Geuss' (2010, 42) often-cited contention that "ethics is usually dead politics" (cf. Bernard Williams 2002, 225–232; Bernard Williams 2005, 6; Newey 2010, 461–462; Rossi and Sleat 2014, 691–692; Aytac and Rossi 2022, 1217; Cross 2022, 1112; Clark 2024). Some realist ideology critics, of course, will claim that we have reasons to avoid moral considerations besides reducing exposure to moral debunking explanations. However, again, I am trying to be as inclusive and noncommittal as possible, thus pinpointing a concern that seems common to most, if not all, realists—and that, in addition, should also be compelling to those who harbor doubts about political realism.8 As noted earlier, Enzo Rossi, Ugur Aytac, and Carlo Argenton have suggested a model of epistemically flawed ideology transmission. To be clear, the Rossi-Aytac-Argenton model, as I will call it, is one possible way to articulate the realist ideology critique research program. But, since it is a very detailed articulation, I will rely on it as an illustration.9 This model begins with a small number of agents who, in the context of a hierarchical social structure, are empowered to disseminate beliefs about that structure's legitimacy (i.e., ideologies) that meet with little (or no) effective contestation. For the purposes of this model, it does not seem particularly relevant how these agents became powerful: this might be the result of conquest, war, natural catastrophe, or whatnot. What matters, essentially, is that, due to motivated reasoning, the powerful are likely to rationalize their own advantage as legitimate, natural, and justified (Aytac and Rossi 2022, 1019–1020; Kreutz 2024)10 and that, given that society's hierarchical structure, those in the lower rungs of the social scale will be unable to effectively challenge these rationalizations (Aytac and Rossi 2022, 1020). Under these conditions, ideology-transmission processes are epistemically unreliable. Not because the relevant beliefs must be false. After all, the powerful might have gotten it just right. The problem, rather, is that those beliefs have a fair chance of being the product of bias—in this case, self-serving, motivated reasoning—and "epistemic checks and balances" are either absent or weak. To put it in the form of a slogan, under such circumstances, "legitimation stories track power, rather than truth." The Rossi-Aytac-Argenton model is, of course, idealized. Even in highly hierarchical, oppressive societies, there are typically some contestatory mechanisms in place (Scott 1985). And some have argued that ideologies are quite often sustained due to strategic reasons because they constitute an equilibrium solution to a collective action problem (Sankaran 2020) or because endorsing the relevant beliefs sends prudentially rational signals (Mercier 2022).11 But to the extent that (i) ideas about social, political, or religious legitimacy matter, (ii) the power to disseminate them is unequal, and (iii) the capacity to contest them is restricted,12 the Rossi-Aytac-Argenton model—understood as one specific way to operationalize the realist ideology critique research programme—can helpfully isolate aspects of belief formation and transmission processes that negatively impinge on those ideas' epistemic status. In this section, I argue that the Rossi-Aytac-Argenton model of realist ideology critique partly applies to animals.13 As formulated above, this model posits that, in hierarchical social orders, those in the higher rungs of the social scale (i) are likely to rationalize their own advantage due to motivated reasoning, and (ii) are capable of effectively obstructing significant, widespread, and coordinated contestation by those in the lower rungs of the social structure. The model applies in part, I will contend, because there is a social hierarchy between humans and the other animals, in which those in the higher rungs of the social scale—namely, humans—are liable to motivated reasoning and can effectively shield their self-legitimating stories from criticism by those most affected by them—that is, animals themselves. Now, there is, of course, a crucial difference: animals, unlike humans, are unable to contest speciesist and anthropocentrist legitimation stories, not because they are in the lower rungs of the social order, but because they lack the relevant contestatory capacities. Their disempowerment, then, is not enforced but constitutive. The Rossi-Aytac-Argenton model, however, can be fine-tuned to register this distinction—a move that, I will contend, is not ad hoc but serves the main goals (and respects the main tenets) of realist ideology critique, as developed in the previous section. Humans regularly kill animals for food, medical experimentation, clothing, military purposes, or following the demands of culture and tradition. Of course, humans can do all of this because millennia of biological and cultural evolution have given us the power to dominate the other animals. We can cooperate at a very large scale and learn from others, create complex tools, make long-term, multi-layered plans, and accumulate, transmit, and revise a formidable storage of knowledge across generations (Sterelny 2012; Henrich 2016). But, most crucially, humans can do all of that because we largely believe it is legitimate (Mathieu and Ritchie 2022). Humans are, by and large, speciesists: we believe that mere species membership is morally relevant, and that membership in the human species is especially relevant (Caviola et al. 2019). Being not only distinct, but superior, we can subject the other animals to our power through practices we deem natural, normal, necessary, and nice (Joy 2009). These are not peripheral beliefs. Rather, they are deeply pervasive and robust, entrenched in the law, in religious texts, moral codes, and everyday language (Waldau 2003; Schaffner 2011; Swartz and Mishler 2022; Leach et al. 2023; Faria and Almirón 2024). This hierarchy, crucially, is upheld by social, economic, and political institutions. Although, for instance, most countries have passed animal welfare laws regulating the use of animals for food, no country has challenged the exploitation and slaughter of animals for food as such. States protect the property rights of animal-agriculture business owners, recognize as legitimate their business interests, usually subsidize their operation, and in some cases pass laws imposing strong penalties to those who trespass into factory farms or disseminate footage recorded within. As Will Potter (2018, 877) notes, "new … forms of animal rights social activism … have been met with an increasingly aggressive backlash, including new terrorism laws, widespread surveillance, experimental prisons, and legislation explicitly criminalizing journalists and whistleblowers."14 Our societies, then, are arranged on the basis of the principle that animals matter much less than humans, and can therefore be harmed and killed in a myriad of ways. Such legitimation stories are (at least partly) transmitted socially, and are reflected in the behavior of social, economic, and political institutions. Recently, a growing body of empirical literature has converged on the finding that humans' attitudes towards animals' moral status and minds are strongly vulnerable to motivated reasoning. In recent years, for instance, psychologists have explored the so-called "meat paradox": the fact that, although most people report being concerned with animal suffering, they nonetheless consume meat—which is, of course, an important source of animal suffering. This creates a psychologically discomforting cognitive dissonance, which people often seek to reduce by downgrading animals' cognitive capacities and mental status. For example, Bastian et al. (2011) found that, when participants in several studies were reminded of the link between meat-eating and animal suffering, they tended to doubt animal minds. More recently, Ioannidou et al. (2024) have reported similar results when individuals are made aware of the connection between animal suffering and the dairy industry: that is, lower degrees of mind attribution and moral concern for dairy cows. And, more generally, François Jaquet (2021) has argued that the same mechanism might explain the belief that human interests matter more than the comparable interests of animals: meat eaters, he suggests, seem to "resolve" the meat paradox by reducing the moral worth of animal interests tout court. Because these reactions seem prompted by a desire to accommodate one's beliefs to one's antecedent behaviors, they are plausibly instances of motivated reasoning (see also Bilewics et al. 2011; Rothgerber 2013; Piazza et al. 2015). Speciesism and anthropocentrism shape, to an important extent, our social arrangements. Presumably, animals have the highest stake in these legitimation stories being contested. Unfortunately, they cannot do any contesting themselves. As anticipated above, this is a crucial point. In hierarchical societies, individuals can be disempowered in two ways. On the one hand, because of their position in the social pecking order, they may be effectively prevented from exercising or developing the contestatory capacities which, in a fundamental sense, they can potentially exercise. I call this enforced disempowerment. On the other hand, an individual may be disempowered because they cannot exercise (not even potentially) the relevant contestatory capacities in the first place. These are cases of constitutive disempowerment, which occur when, no matter how their society is arranged, an individual lacks the relevant psychological (cognitive and metacognitive) capacities required to contest legitimation stories. Both cases are, in a crucial respect, different. In cases of enforced disempowerment, individuals are disempowered by the very social hierarchy they may otherwise seek to contest. In such cases, the social hierarchy plays a causal role, insofar as it explains why some are unable to contest legitimation stories to begin with. In cases of constitutive disempowerment, however, the social hierarchy makes no difference to disempowerment—it all hinges on one group's capacity to shape social arrangements, and on the other group's corresponding inability. The Rossi-Aytac-Argenton model, as formulated above, focuses on cases of enforced disempowerment. Animals, however, are constitutively disempowered. This is an important challenge, to which I shall respond in two ways. First, I will note that the disempowerment of animals is, to some extent, enforced. Second, I will argue that constitutive disempowerment, even if distinct, should nevertheless be of interest to realist ideology critics—and those working with the Rossi-Aytac-Argenton model in particular. First of all, animals' disempowerment is, to some extent, enforced. On the one hand, animals do often resist (even violently) practices that harm them—for instance, complaining when confined, trying to escape from factory farms and slaughterhouses, or attacking their keepers (Colling 2021). These behaviors might be understood as expressions of dissent towards those practices. Now, whatever rudimentary forms of contestation animals are capable of, however, humans can effectively (and routinely do) crush them. Animals trying to escape from captivity, for instance, will more likely than not be shot with a tranquilizer gun—or, at worst, with real ammunition. On the other hand, and whether or not animals can, in any meaningful sense, directly challenge speciesism and/or anthropocentrism, other humans can do so on their behalf, acting as their "discursive representatives" (Dryzek and Niemeyer 2018). But, again, this kind of indirect, surrogate empowerment is severely constrained. Vegetarians, vegans, and animal activists face a host of pressures, both formal and informal—ranging from, as noted above, quite stiff penalties for acts of civil and uncivil disobedience to hostile environments in which they are the object of nasty remarks or ridicule (vegans, especially, regularly report feeling discriminated against; see Cole and Morgan 2011, MacInnis and Hodson 2017, Horta 2017). So, even if the Rossi-Aytac-Argenton model (and/or realist ideology critique in general) only applied to cases of enforced disempowerment, it would still apply—to some extent—to our treatment of animals. Still, it seems hard to deny that animals' disempowerment is, to a crucial extent, constitutive. No matter how attentive we are to their interests, preferences, and behaviors, they cannot—for biological reasons—empower themselves in the way that large-scale contestation in complex societies requires. Animals' ability to (directly or indirectly) contest speciesism and anthropocentrism might not be fully insensitive to social hierarchy, but it is much less sensitive than humans' corresponding abilities. In the remainder of this subsection, I will defend expanding the Rossi-Aytac-Argenton model, so that it covers both cases of enforced and constitutive disempowerment. As pointed out above, the Rossi-Aytac-Argenton model fundamentally seeks to criticize, from an epistemic standpoint, the legitimation stories underpinning hierarchical social arrangements—empirically unveiling the circumstances under which, in the words of the above slogan, "legitimation stories track power, rather than truth." Now, when we have a social hierarchy, legitimized on the basis of beliefs that are liable to motivated reasoning, and which cannot be contested effectively by those on the low end of the social hierarchy, whether the disempowerment is enforced or constitutive seems immaterial. In both cases, we will have compelling reasons to believe that the relevant legitimation stories track power, rather than truth. In cases of enforced disempowerment, of course, we will have a distinctive story, wherein the social hierarchy itself plays a causal role. In cases of constitutive disempowerment, we will "only" have (inevitable) power asymmetries casting doubt on the foundations of a pervasive social hierarchy. But, again, if the main goal of realist ideology critique in general (and the Rossi-Aytac-Argenton model in particular) is to criticize social arrangements by appealing to normative standards that are, comparatively speaking, less immune to debunking explanations than moral standards allegedly are, then there seems to be no principled reason to oppose extending the model. As Rossi (2024, 290) puts it, the "rough idea is that people and social groups who enjoy a distribution of power particularly skewed in their favor … cannot be seen as epistemically reliable producers and reproducers of legitimation stories about their social order." Humans, in our world, benefit from a distribution of power particularly skewed in our favor. That this is so, to an important extent, because of animals' constitutive disempowerment, is surely important. But that does not seem to alter the underlying point: given that, for whatever reasons, the social order is skewed in our favor, we are not in the best position, epistemically speaking. Let me approach this issue from a different angle. In a recent paper, Rossi (2024, 286–287) describes his model as an attempt to address what Bernard Williams (2005) famously called the "first political question:" namely, the problem of securing "order, protection, safety, trust, and the conditions of cooperation." This is, in Williams's view, the first question because it is only when a political order has secured a minimum of order, protection, etc. that we can hope to tackle other questions—say, questions about income redistribution and the like. The basic idea is that the first political question has an important epistemic dimension: if our deliberations are clouded by ideological distortions, the argument goes, our responses to the first political question will be, inevitably, suspect. The Rossi–Aytac–Argenton model offers tools to diagnose such distortions so that we can achieve a good epistemic position from which to address the first political question. So far, so good. But here it is important to note that when we address the first political question, we must, inevitably, address a closely related question: "Whose good is relevant when assessing questions of order, protection, safety, trust, and cooperation?" Williams's very abstract formulation gives us no indication. It may be that the answer is: "only the good of humans." Or, alternatively, "the good of all sentient beings." Or, of course, perhaps something in between. At any rate, responding to the first political question requires thinking about the political standing of animals, and thinking clearly about the political standing of animals—that is, from an appropriate epistemic position—requires being able to identify when (and why) our thoughts on the matter are clouded by ideological distortions. An expanded Rossi–Aytac–Argenton model seems well suited for this task. I conclude, then, that the original Rossi-Aytac-Argenton model of flawed ideology formation and transmission can partly apply to our treatment of animals, and that an expanded version of the model can apply more generally. This expansion, I have argued, is not ad hoc. Rather, it can be motivated by appealing to the main tenets and goals of realist ideology critique in general—and the Rossi-Aytac-Argenton model in particular. In the next section, I will argue that, drawing on these same desiderata, we also have reason to pay attention to the ways in which legitimation stories may be formed through epistemically deviant belief formation mechanisms even if those mechanisms are not necessarily social or political—such as, for instance, adaptive beliefs favored by natural selection. Before proceeding any further, however, I will address two potential objections to the argument presented in this section. One first problem, an objector may claim, is that the argument defended above overgeneralizes.15 Because animals' disempowerment is, to a large extent, constitutive, there is no way to arrange our society in which they would enjoy an equal capacity to contest any legitimation stories. Suppose, for instance, that we lived in a society effectively arranged on the principle that humans and animals' comparable interests must be weighted equally, as Peter Singer (1975) has famously championed, or in a "zoopolis" in which, as Donaldson and Kymlicka (2011) have defended, domestic animals are considered our co-citizens, communities of wild animals are granted sovereignty rights, and "liminal" animals—non-domesticated animals living in our cities, such as rats or wild boars—enjoy some kind of "denizenship" status. Even in those societies, animals would remain, by and large, unable to contest the relevant legitimation stories—and so, it would seem, the relevant stories would remain every bit as suspect as the speciesist and anthropocentrist narratives criticized in this section. My response to this first objection is three-pronged. First, I do believe that, given the inevitable (constitutive) power asymmetries between humans and the other animals, our discourses about animals are always going to be, to some extent, suspect. As animal ethicists are usually happy to concede, we cannot escape epistemic anthropocentrism: everything we claim about animals, we infer it from our (human) point of view. We cannot inhabit the subjective viewpoint of a lion or a bald eagle—nor, in the often-repeated phrase, can we experience what it is like to
Pablo Magaña (Mon,) studied this question.