Louder Than Words? Why Political Ethnography Cannot Fix the Flaws of Normative Behaviorism

Abstract

This article engages with Jonathan Floyd’s normative behaviorism, and Simon Stevens’ recent criticism of this approach. Its objectives are two: First, the article challenges normative behaviorism’s reliance on crime and insurrection rates as measures of political discontent. I raise three distinct objections against using crime and insurrection rates for this purpose, and argue that these undermine normative behaviorism’s justificatory strategy. Second, the article rebuts Stevens’ suggestion that the flaws of normative behaviorism can be fixed by adding political ethnography to the picture. Rather than remedying the flaws, I argue that supplementing normative behaviorism with an ethnographic sensibility will deprive it of the very features that rendered it attractive in the first place. I conclude that if fixing the flaws of normative behaviorism requires us to transform it beyond recognition, we should rather reject it in the first place.

Over the past 15 years, political theory has experienced a “methodological turn” as scholars in the field have come to question predominant methodological assumptions (Erman and Möller, 2015, 2018; Rossi and Sleat, 2014). At the core of the resulting debate are disagreements between the hitherto dominant “idealist” theoretical outlook and a range of “realist” approaches challenging this. Though realist accounts vary internally, most of them share a concern that “idealist” political theory fails to acknowledge core features of politics, such as disagreement and power, and an ambition to uncover a distinctively political normativity, separate from morality (Horton, 2010; Jubb, 2019; Philp, 2012). While idealist theoretical approaches typically focus on values such as justice and equality, realists prefer “thinner” ends like stability, security, and legitimacy (Galston, 2010; Geuss 2008; Sleat, 2014).

A recent contribution to the realist corpus is Jonathan Floyd’s normative behaviorism. As the name indicates, a central feature of this view is its focus on behavior. If dominant theoretical approaches ground justification of political systems in what people think about them (or the principles and values they embody), normative behaviorism holds that acts, not thoughts, reveal our true political commitments. In order to assess the legitimacy of a political order, we therefore need to ask how people act in response to this order. More specifically, Floyd suggests that we should use crime and insurrection rates as a measure of how well a political order satisfies people’s political preferences (Floyd, 2017, 2020, 2023a).

This article has two aims: First, it adds to recent criticisms of normative behaviorism, distinguishing between three objections that can be raised against its use of crime and insurrection rates as measures of political discontent. Second, it rebuts the proposal—recently advanced by Simon Stevens—that ethnographic sensibility could fix the flaws of normative behaviorism. Instead of filling in the gaps, I will argue that turning to ethnography will deprive normative behaviorism of the very features that rendered it attractive in the first place.

The article is structured as follows. The first two sections provide a brief account of Floyd’s normative behaviorism, and explain the role of crime and insurrection as measures of political discontent within this approach. The subsequent section presents three objections against normative behaviorism and argues that its reliance on crime and discontent falls back on an idealized account of agents’ abilities and motives. The fourth section discusses Stevens’ proposal to supplement normative behaviorism with an ethnographic sensibility, highlighting a set of problems that this proposal gives rise to. Drawing on this discussion, the following section argues that adding ethnographic sensibility deprives normative behaviorism of the very features that rendered it attractive in the first place. The final section summarizes my conclusions.

Floyd’s Normative Behaviorism

The novelty of Floyd’s approach consists in its attempt to ground justification on behavioral patterns rather than thought experiments or hypothetical consent. Normative behaviorism shifts the focus of political theory away from words and thoughts toward acts. Instead of assuming that a political order is justified if regulated by principles that correspond to people’s moral intuitions, or that they would agree to under some specific set of counterfactual circumstances, normative behaviorism takes an order to be justified if people respond to it in specific ways (Floyd, 2020: 139). Normative behaviorism aligns with most realist approaches in that it grounds its normative judgments in de facto legitimacy. What grants political institutions legitimacy is that real people actually find them legitimate. Normative behaviorism differs, however, from other realist accounts in that it does not take at face value what people say about legitimacy. Rather than words or thoughts, our acts reveal our true moral commitments (Floyd, 2020: 145).

Not any acts serve this function, however. Normative behaviorism takes the frequency of crime and insurrection in society as a measure of the legitimacy of its political institutions. According to Floyd, these acts express, more than others, a rejection of the political order. Crime and insurrection are reliable markers of political discontent because they involve risk. Since they typically come at a high risk to the agent, we shall expect people to engage in these acts only if they find their current lives unbearable, due to a lack of various goods (Floyd, 2020: 140). Lower crime and insurrection rates thus indicate higher satisfaction with political preferences, which Floyd takes to ground the legitimacy of political institutions.

Two things are important to note. The first is that crime and insurrection rates are not snapshot measures but can only be estimated once a political system has been up and running for a longer period of time. Floyd does not tell us for how long a system must be tested before we can assess its performance, but the general idea is that while coercion or manipulation can keep a system running for some time, resistance will eventually manifest itself through increasing rates of crime and insurrection (Floyd, 2020: 145–146).

The second thing to note is that normative behaviorism offers only comparative judgments about the legitimacy of political institutions. As long as some instances of crime and insurrection remain in society, there may always be an alternative set of institutions that outperforms even the best institutions tested so far (Floyd, 2020: 146, 2023a: 368). The comparative aspect relieves normative behaviorism from what would otherwise have been a debilitating bias toward the status quo. Still, we may think of normative behaviorism as the conservative sibling in the realist family, since we only have epistemic access to the virtues of political orders that have been tested in the real world. Whatever institutional arrangements exist, they have at least the advantage over any untested set of institutions that they have proven minimally workable for securing social peace. This points toward a reformist agenda where we proceed gradually and on an experimental basis, testing the workings of institutions in a smaller scale before implementing them in society at large (Floyd, 2023a: 369).

Crime and Insurrection as Measures of Discontent

According to Floyd, normative behaviorism has several advantages over idealist political theory as well as over competing realist approaches. Normative behaviorism, he argues, overcomes several disagreements that prevent progress in the field of political theory. It allows us to rank different political orders in a way that avoids problems of incommensurability and to offer clear institutional prescriptions rather than merely pointing to values that political institutions should embody (Floyd, 2020: 145, 147). One of the most tempting promises of normative behaviorism is that it offers a way to overcome the risk that ideological delusion leads us to incorporate a bias toward the status quo when elaborating our normative theories. This corresponds to two problems Floyd claims to have solved: The idealization problem and the legitimacy problem.

Idealization is a problem for mainstream normative approaches that involve the use of counterfactual assumptions, hypothetical agreements, and thought experiments. Examples of idealizations taken from John Rawls’s highly influential theory of justice are the assumptions that agents are rational and self-interested, that society is a closed scheme of social cooperation, and that moral demands will be strictly complied with (Rawls, 1999: 7, 12, 125, 215). Among the various objections that could be raised against idealizations in normative theory is the view that this will taint political theory with an ideological bias toward the status quo. When replacing facts with an idealized picture of reality, theorists give free leeway to their own cognitive biases and reproduce ideologically deluded views of agents and society (Mills, 2005: 172, Valentini, 2009: 348–349). Normative behaviorism, Floyd argues, overcomes the problem of idealization by avoiding idealizations altogether, taking people and their motives as they are. Rather than imposing normative standards from above, normative behaviorism takes as its normative standard whatever values and commitments people happen to hold (Floyd, 2020: 142).

The legitimacy problem arises precisely due to the ambition to start from people’s actual commitments. It is therefore at the core of all realist approaches. If we equate sociological and normative legitimacy—that is, if we assume that whatever people take to be legitimate actually is legitimate—how do we distinguish genuine acceptance of a political order from acceptance that is due to, for example, coercion or manipulation? The problem is captured by what Bernard Williams calls the critical theory principle, which states that acceptance must not be a product of the power that it seeks to legitimize (Williams 2005: 6). Floyd argues that the focus on acts rather than words solves the legitimacy problem. If people claim that they find their political order legitimate, yet “find themselves turning to crime and insurrection,” this is a reason for us to dismiss their proclaimed approval (Floyd, 2020: 145–146). In this sense, “acts speak louder than words” (Floyd, 2020: 140). A closer look at Floyd’s argument reveals how the focus on acts rather than words allows normative behaviorism to meet the critical theory principle.

The view that acts speak louder than words takes on a twofold meaning in Floyd’s theory. First, acts set the threshold for when dissatisfaction is severe enough to challenge the legitimacy of a political order (Floyd, 2016a: 176). Words that are not accompanied by acts of the relevant kind will be dismissed. As such, acts add weight to the discontent people express in words. Second, acts speak louder than words in the sense that when acts and words go apart, acts win (Floyd, 2016b: 160–161). It is in this latter sense that the focus on acts instead of words responds to the critical theory principle. Discrepancies between words and acts are often cast as a sign of ideological power at work (Gramsci, 1971: 326–327; Lukes, 2005: 51–52). To say one thing and do another—in good faith, that is—reveals that we are subject to delusion or manipulation of some sort. In particular, to say that we accept a social or political order while acting contrary to the norms or laws that uphold it reveals a tacit (conscious or unconscious) discontent with this order. To privilege acts when such discrepancies occur is normative behaviorism’s solution to the legitimacy problem.

Floyd’s theory proceeds from acts of crime and insurrection to legitimacy in two steps. The first step aligns preference satisfaction with the legitimacy of political institutions. When asking whether a political order is legitimate, what we ask is whether it satisfies the political preferences of the people it is meant to govern. This has implications that many will find hard to digest. For example, if people turned out to prefer a society based on strict social hierarchies, racial segregation or female subordination, normative behaviorism would consider political orders that institutionalized these features legitimate (see Erman and Möller, 2024: 13). Yet, Floyd clearly states that he is willing to bite this bullet, accepting that his account favors any order that produces less crime and insurrection than the alternatives (Floyd, 2023a: 361; 2023b: 492). For the sake of argument, I will therefore grant that whatever political order satisfies people’s political preferences better than the feasible alternatives is legitimate. Instead, I will focus on the second step, linking crime and insurrection to political discontent. According to Floyd, crime and insurrection rates tell us to what extent people are satisfied with the political order. In the following section, I explore three objections to this idea.

Three Objections Against Crime and Insurrection as Measures of Discontent

Normative behaviorism tells us to evaluate political institutions according to how much crime and insurrection they produce. While seditionary acts disturb social peace, stability is not what renders a political order legitimate on this account. Instead, crime and insurrection are treated as markers of political discontent. Low rates of crime and insurrection inform us that people’s preference satisfaction is high, while high rates indicate the opposite. Previous criticisms of normative behaviorism have raised concerns regarding the accessibility and reliability of data necessary for determining crime and insurrection rates (Baderin 2023: 478, Dowding, 2023: 448, Miller, 2023: 442). Though clearly relevant for the application of normative behaviorism, these are practical hurdles that do not undermine the principled force of its justificatory framework. For the sake of argument, I will here set problems of data collection and data analysis aside, focusing instead on three objections that target the focus on crime and insurrection as a justificatory strategy.

Circular Evaluation

The first objection states that the justificatory model based on crime and insurrection allows each political order to set the standards for its own evaluation and thus allows for a form of circular evaluation. According to normative behaviorism, the legitimacy of a political order depends on the crime and insurrection rates it produces. Yet, what counts as crime and insurrection will itself depend on the political order. In this sense, the political system sets the standards against which it is to be evaluated, standards that will vary from one society to another. As Dowding (2023) observes, liberal democracies would display considerably higher levels of sedition than autocracies if peaceful protests against the government were to count as sedition. Yet, in liberal states, peaceful protest is usually considered a sign of a vital democracy, whereas it would count as sedition in an authoritarian state (Dowding 2023: 448). Egalitarian liberal democracy happens to be the political order Floyd takes to be sanctioned by normative behaviorism (but see Miller, 2023: 145). Now, if the lower rates of crime and insurrection displayed in democracies is an effect of their narrower account of crime and insurrection, what seemed like a comparative advantage rather reflects the different evaluative standards at work.

Floyd touches upon this issue when outlining his argument. Rather than a problem, however, he takes democracy’s “malleability under pressure” to constitute a virtue that “explains it resilience” (Floyd, 2020: 142–143). Here, Floyd seems to encourage us to evaluate political orders from an internal viewpoint, holding each system up against its own standards of crime and insurrection. This makes sense when we take into consideration that normative behaviorism focuses on the legitimacy of political orders rather than specific policies or regimes. What matters for the legitimacy of the political order is the extent to which people subjected to this order are willing to play by its rules. The fact that those rules—and, hence, standards of crime and insurrection—vary is thus not a problem for assessment but precisely what lies the ground for it. Because democracies allow peaceful protest, they are capable of harnessing dissent on substantive issues more effectively than competing political orders, thus avoiding that people turn to crime and insurrection.

Note, however, that this solution appears, unwittingly, to settle the debate about the best political order in favor of anarchy. And it does so without taking any empirical facts into consideration. In the anarchic society, there is no law and there will consequently be no crime. Furthermore, there is no government to revolt against, and hence, no insurrection can occur. If our comparison of political orders is based on crime and insurrection as defined in each system, then, anarchy will—per definition—outperform every alternative political order. Normative behaviorists may respond to this challenge by insisting that their account assesses the legitimacy of political institutions, and that anarchy, since it lacks such institutions, falls outside its scope. Even if valid, however, this response would show that normative behaviorism is silent on what may be considered the most central issue when it comes to political legitimacy—the legitimacy of government as such (Simmons, 1979; Wolff, 1970).

Unequal Impact

The second objection states that measuring political discontent through crime and insurrection rates grants unequal consideration to different groups in society. Some groups (e.g. women and the elderly) are less likely than others (notably young men) to engage in crime and insurrection. As a result, normative behaviorism will err both on the side of the positive and the negative when estimating political preference satisfaction. While the discontent of some will not tell against the legitimacy of a political order (since it does not translate into seditious acts), even trivial discontent—of the kind normative behaviorism urges us to disregard—of others will affect our assessments (since these others turn easily to crime and insurrection).

In a recent article, Simon Stevens (2023) elaborates on this problem. He argues that normative behaviorism fails to capture political discontent within groups that, precisely due to the prevailing political order, lack the capacity to engage in crime and insurrection. Their incapacity may be due to different factors, including geographical distance, physical hindrances, psychological constraints, or risk-aversion due to particular responsibilities (Stevens 2023: 6, 9). By failing to acknowledge the discontent of groups incapable of turning to crime or insurrection, normative behaviorism will produce an overly optimistic view of the legitimacy of political institutions. The resulting bias is all the more troubling since (a) the inability is—directly or indirectly—caused by the existing political order, and (b) it silences the discontent of the most disempowered (Stevens 2023: 12). The apparent legitimacy of the political order will thus be an effect of the workings of its power, violating the aforementioned critical theory principle.

Floyd is well aware of the problem of unequal impact. He offers no less than four different replies to it, only to immediately reject two of them. Of the two remaining, the first falls back on his conclusion that normative behaviorism ends up recommending a liberal democratic political order. Though the discontent of groups that do not engage in crime and insurrection will not be acknowledged, a democratic political order contains the elements necessary for improving their situation (Floyd, 2020: 142–143). Similarly, he argues that despite the fact that normative behaviorism grants undue consideration to the discontent of some hot-headed groups, a democratic order awards no pay-off to their behavior (Floyd, 2020: 155–156). But this reply is unpersuasive, since it falls back on the substantive outcome of normative behaviorism. The fact (if it is a fact) that democracy caters to the needs of vulnerable groups and neglects hot-headed protesters’ demands does nothing to save normative behaviorism’s justificatory strategy. Normative behaviorism was supposed to provide a justification for democracy, not the other way around.

Floyd’s second reply states that the mere presence of voiceless groups may inspire others to engage in seditious acts (acting on behalf of, or in solidarity with, the disempowered). This, he argues, will matter since, on normative behaviorism’s comparative approach, even a small decrease in crime and insurrection will count in favor of a political order (Floyd, 2020: 143). I take him to suggest that even if the protests are not proportionate to the discontent, a society with silent disempowered groups will come out as comparatively worse than a society without such groups, due to others acting on behalf of the disempowered. What this reply fails to acknowledge is that once we recognize that crime and insurrection rates are not proportionate to the discontent experienced in the population, this casts serious doubts over normative behaviorism’s comparative strategy. If crime and insurrection rates are not proportionate to discontent, we cannot tell whether a decrease in crime and insurrection following a shift from one political order to another actually corresponds to a decrease in discontent overall. For all we know, any decrease in open discontent reflected by lower crime and insurrection rates may be outweighed by a larger increase in silent discontent.

Irrelevance

The third objection challenges the extent to which criminal acts are reliable markers of political discontent in the first place. Floyd argues that while insurrection is a direct expression of discontent with the political order, crime expresses such discontent indirectly, by rejecting the various ways of life that the order facilitates (Floyd, 2016a: 167). Yet, while this is arguably true for some criminal acts, it is far from obvious that all instances of crime express discontent of this kind. It is, for example, not clear that we should understand the criminal acts of, for example, speeders, swindlers or rapists as rejections of the political order. Rather than political discontent, we may think that crimes like these are indicators of an agent’s negligence, egoism, or malevolence. This calls for an inquiry of the link between crime and political discontent stipulated by normative behaviorism. What reason do we have to think that criminal acts express political discontent?

One possibility would be to point to the fact that all crimes involve breaking laws. Laws, in turn, are the result of a political order and also what upholds it. If crime expresses discontent with laws, then, we may take this as a sign of political discontent. There are, however, several problems with this proposal. First, criminal acts do not always signal discontent with the broken laws. Stealing need not entail a rejection of property rights, just as speeding need not express a wish for unregulated highways. The typical criminal does not wish to annul the law but rather exempt herself from it; when she falls victim of theft, the thief will call for justice. Indeed, we often think of criminal acts that are clear expressions of discontent with a specific law in terms other than crime, such as “civil disobedience” or “political violence.”

Second, it is not clear that discontent with a specific law translates into discontent with the political order taken as a whole. Surely, we may find a specific law silly or groundless and therefore decide to break it without rejecting the political order that has issued it. Even in cases where crime does express discontent with a specific law, then, it is an open question whether it also expresses discontent with the political order.

Perhaps criminal acts express political discontent even if they do not express discontent with specific laws. Philosophical accounts of legitimacy often assume that the legitimacy of a political order provides reasons for us to obey the law independently of its content. If people’s political views were fully consistent and their reasoning flawless, we should thus expect them to obey any law issued by what they perceived of as a legitimate political order. Conversely, we should expect every breach of the law to express a form of disrespect for the political order from which it originates. In reality, however, epistemic conditions including fallible reasoning, imperfect information and cognitive bias hinder these expectations from arising. People may fail to recognize the connection between legitimacy and law obedience, or between the political order and the laws that uphold it. Or concerns for the duty to obey the law may be overridden by other motives, including self-seeking ones. Under such conditions, whether people decide to engage in crime or not may depend more on the prospects of getting caught and the cost of punishment than on how they perceive of the political order (see Dowding, 2023: 448). Insofar as crime expresses a dismissal of government institutions, then, this may concern not the legitimacy of the legislating body, but the efficiency of law enforcement agencies.

A different attempt to establish a link between crime and political discontent stresses the risk involved in criminal acts. Since people are generally risk-averse, we must assume that people will engage in criminal activity only in response to an unbearable social situation for which the political order is responsible. That crime occurs only as a response to unbearable conditions is a point Floyd repeatedly returns to (Floyd, 2020: 140; 2023a: 359). His view captures the well-established fact that social conditions of, for example, deprivation and humiliation put people at risk of committing crime. Even if people do not explicitly blame the political order for their situation, we may therefore argue that the mere fact that they find themselves turning to crime suggests that this order has failed to satisfy some basic need of theirs. Yet, though a plausible explanation for some types of crime—a poor person stealing bread is perhaps the most clear-cut example—this seems clearly implausible when it comes to other criminal acts, including, but not restricted to, “white-collar crimes” (Baderin, 2023: 477). Though exceptions may occur, swindle, rape, and speeding are not typically responses to an unbearable social situation, but rather reflect the agents’ greed, ruthlessness, or lack of impulse control. In such cases, as well as those where law-breaking responds to an emergency situation (speeding to get a wounded friend to the hospital) or is due to mere ignorance (trespassing because failing to see the “No trespassing” sign), it would be a mistake to equate crime with political discontent.

In response to this, we may suggest that even if crime sometimes is an effect of moral flaws, the very fact that people are morally corrupted shows that something is wrong with the political order. Prima facie, it seems plausible to think that a political order that fosters the moral character of its citizens is more legitimate than one that fails in this respect. A political order that produces many rapists is arguably worse than one that endows its citizens with a deeply felt respect for others’ bodily integrity. There are, however, several reasons why normative behaviorism should avoid this argumentative route. The first is that Floyd implicitly rejects it. In his discussion of tax evaders, he clearly distinguishes between acts that express political discontent and purely self-seeking acts. Tax evasion, he argues, cannot replace crime and insurrection as markers of political discontent, because rather than political, tax evasion is “a self-serving act, focused on wealth-maximization, and without regard for the regime one might undermine through lack of support” (Floyd, 2020: 155). Bracketing the (for Floyd) potentially troubling fact that tax evasion, too, is unlawful and hence already included in the broader category of crime, this suggests that normative behaviorism does not take people’s character flaws to cast any shadow over the political order.

Another reason why normative behaviorism should avoid this line of reasoning is that the connection it establishes between crime rates and the legitimacy of political orders appears to sidestep discontent altogether. While normative behaviorism cannot depend on political discontent being fully conscious and explicit—if it did, it could not solve the legitimacy problem—we must at least be able to offer a convincing reconstruction of criminal acts as expressions of political discontent. While a poor person who steals a loaf of bread to satiate her hunger is clearly driven by self-interested motives, it is easy to see how this motive translates into a political preference for an order where everyone has their basic needs provided for. In contrast, if we take people’s character flaws to directly challenge the legitimacy of the political order, discontent has no role to play in our theory. This would deprive normative behaviorism of its advantage over idealist theory, since it would no longer take people’s actual moral commitments as its normative standard.

This, in turn, points to the last reason why normative behaviorism should avoid faulting the political order for people’s character flaws, namely that doing so undermines normative behaviorism’s justificatory strategy. According to normative behaviorism, a political order is legitimate if it corresponds to people’s actual motivations. Yet, if we take motivations to be a product of the political order, they cannot also be what legitimizes this order. This would violate the critical theory principle, according to which acceptance must not be an effect of power, and thus deprive normative behaviorism of its advantage over other realist approaches.

The objections outlined in this section cast doubt over normative behaviorism’s choice of crime and insurrection as measures of political discontent. I have argued that this choice commits normative behaviorists to anarchism since this political order rules out the very possibility of crime and insurrection. But even if we think that normative behaviorism can escape this conclusion, the objections from unequal impact and irrelevance withstand. In some cases, focusing on crime and insurrection rates will lead us to underestimate existing discontent, ignoring the discontent of groups incapable of engaging in crime and insurrection. In other cases, it will lead us to overestimate levels of discontent, granting excessive attention to the discontent of “hot-headed” groups, and/or ascribing political motives to criminal acts that lack such motives. Either way, normative behaviorism will fail to identify the actual levels of discontent, and hence fail—by its own standards—to determine the legitimacy of political orders.

Interesting to note is that these flaws follow from idealizations of agents’ abilities and motives, of precisely the kind that normative behaviorism claims to avoid. Normative behaviorism attributes (counterfactually) to all agents an equal capacity to engage in crime and insurrection (Stevens, 2023: 12). Furthermore, it assumes (also counterfactually) that agents engage in crime and insurrection only as a response to an unbearable life situation. Once we lift these idealizations, we come to doubt that crime and insurrection rates constitute reliable measures of political discontent.

Ethnography as Remedy

I have argued that crime and insurrection are unreliable measures of political discontent and thus risk to produce flawed judgments of the legitimacy of political orders, even by normative behaviorism’s own standards. To remedy this, we may try to expand the set of acts that serve as measures of discontent in normative behaviorism. Apart from crime and insurrection, we may, for instance, include migration rates, tax evasion, and various forms of antisocial behavior in the relevant set. There is arguably much to say for this proposal. Yet, Stevens (2023) argues that such amendments will not suffice. Instead, he proposes that we supplement normative behaviorism with an ethnographic approach. By developing an “ethnographic sensibility,” closely engaging with the reasons and actions of vulnerable groups, theorists will be able to uncover discontent that comes in shapes other than crime and insurrection (Stevens 2023: 14–15).

Before looking into what conducting political theory with an ethnographic sensibility entails, we should attend to an apparent ambiguity in Stevens’ account. Outlining his proposal, Stevens writes both that political ethnography should “complement” normative behaviorism, and that it will be used to “tweak” it (Stevens, 2023: 16). Strictly understood, these may be held to constitute distinct, and even opposing, aims. While tweaking involves making changes to the original theory in order to improve it, complementing normative behaviorism means to add additional methods to fill the gaps left by normative behaviorism. Given this, we may consider Stevens’ appeal for an ethnographic sensibility not so much an attempt to fix normative behaviorism as a call for methodological pluralism. ¹

Methodological pluralism makes sense as long as the different methods respond to different questions, covering separate areas of the normative terrain. Indeed, this is arguably what “filling the gaps” left by a theory means. Yet, rather than responding to different questions, ethnographic sensibility provides different answers to the question asked by normative behaviorism—how much discontent does a political order produce? Stevens’ proposal “inserts” ethnographic sensibility into normative behaviorism’s justificatory apparatus in order to improve the validity of its results. Despite his claim that ethnography will “complement” rather than amend normative behaviorism, this suggests that ethnographic sensibility, too, aims to fix this methodological approach. At least, this is what Stevens should strive for, given his conclusion that normative behaviorism produces flawed assessments of political discontent. For this reason, I will here treat ethnographic sensibility as a remedy to normative behaviorism rather than a distinct method operating parallel to it.

Stevens’ proposal draws on work from the recent “ethnographic turn” in political theory, calling for a closer engagement with empirical facts, advancing a contextually situated, bottom-up approach to political theory (Herzog and Zacka, 2019; Longo and Zacka, 2019; Perez, 2024). Janosch Prinz has offered a convincing explanation for why ethnographic sensibility is particularly suited to fill the gaps of realist political theory. While attentive to people’s actual commitments, he argues that ethnography also provides the tools for uncovering ideological delusion and false consciousness:

More specifically, an ethnographic sensibility is well-suited to address the realist predicament, because it combines two levels of interpretation. On one level it seeks to reconstruct people’s understanding of what they are doing and who they are. /. . ./ At a different level, ethnography seeks to interpret the larger ideational and material power relations that affect people’s values and practices (and come in the form of ideologies). /. . ./ Exploring the misalignments, tensions, and gaps between these levels provides a starting point for criticism (Prinz, 2020: 67).

According to Stevens, ethnographic sensibility points toward a way to correct the flaws of normative behaviorism. Engaging closely with the acts and self-understanding of vulnerable groups will allow us to identify political discontent that does not take the shape of crime and insurrection. Some groups will, however, have difficulty articulating discontent or even lack voice altogether (e.g. non-human species). In such cases, Stevens suggests we employ tools from so-called “moral sentimentalism,” a method that tells us to adopt an empathic approach, seeking to “feel others pain.” But rather than an “empathy of feeling,” Stevens suggests we develop an “empathy of thought,” trying to imagine how our objects of study might think about political institutions. This accords a place to fictional narrative in the otherwise empirically informed ethnographic approach (Stevens, 2023: 17–18). Since fictional narratives raise problems that do not necessarily apply to political ethnography at large, however, my coming discussion will mainly focus on ethnographic inquiry.

As a remedy to the problem Stevens identifies—that normative behaviorism ignores the discontent of groups incapable of engaging in crime and insurrection and hence produces overly optimistic verdicts of the legitimacy of political orders—the turn toward ethnography seems initially promising. Where crime and insurrection fail to recognize discontent, ethnographic inquiry may prove a useful tool. It is less clear, however, that this strategy can help to solve the opposite problem that normative behaviorism grants too much attention to the discontent of “hot-headed” groups and therefore produces overly pessimistic judgments of political orders’ legitimacy. In particular when adopting an empathic approach, ethnography seems more apt for giving voice to the voiceless than to silence overly loud voices.

This points to three distinct problems that normative behaviorism with an ethnographic sensibility must solve. The first is to determine which groups should be granted ethnographic attention. Stevens’ concern is with groups who are incapable of engaging in crime and insurrection. Yet, rather than revealing an incapacity, some of the cases he discusses suggest that these acts impose excessive costs on the agent. Being a single parent, for example, does not render crime and insurrection more difficult, only costlier (since going to jail would leave the child an orphan). This is potentially troubling, since the reason why privileged groups do not commit crime and insurrection (to defend or increase their privileges) could also be cashed out in terms of costs. Members of a former aristocracy may have too much to lose to engage in crime and insurrection seeking to restore the ancient order. If asked, they might even confirm their commitment to liberal ideals of equality and meritocracy. Still, a closer look into their reasons and actions—for example, privileging members of the aristocracy as employees or business partners, or socializing mainly in exclusive clubs and societies—will reveal their discontent with the existing political order. To know which groups are unable to participate in crime and insurrection, we would need a clear account of which costs could be tolerated and not.

The second problem is that of determining when members of a group should be taken to express political discontent. This holds even if we constrain ethnographic attention to groups strictly incapable of engaging in crime and insurrection. Clearly, the mere fact that someone is unable to engage in crime and insurrection does not tell us whether they would wish to do so. Which acts, then, should we take to express people’s political preferences? Given the complexity of human reasons, the attentive ethnographer is likely to collect a multitude of observations for any given group or agent. While some of an agent’s reasons and actions may seem to express political discontent, others may point in the opposite direction.

To illustrate this problem, we may return to Floyd’s discussion of tax evaders. Rejecting the proposal that crime and insurrection should be replaced by tax evasion as a measure of political discontent, Floyd points to the fact that tax evaders typically choose to live in high-tax societies, benefiting from the public goods they deliver. In this case, he argues, the choice of residency reveals that tax evasion is not actually an expression of political discontent that challenges the legitimacy of the political order. When looking at people’s behavior, he concludes, “some acts speak louder than others” (Floyd, 2020: 156).

Here, Floyd employs something similar to the ethnographic method Prinz describes, identifying a discrepancy between what tax evaders say they prefer and the preferences revealed through their acts. Yet, to simply point to a contradiction between two acts will not suffice: We must also be able to tell which of the two acts speaks louder than the other. Floyd takes the fact that tax evaders choose to stay in high-tax countries to reveal their true preference for high-tax political orders. He concludes that tax evasion does not express political discontent, but only the self-seeking desire for wealth maximization (Floyd, 2020: 155). But this arguably sets the threshold for political discontent too high. By the same logic, the battered woman who returns home to her violent husband upon leaving the hospital would have justified the violence she is subjected to. There could be many reasons—cultural attachment, family bonds, language barriers, etc.—for staying in a country despite rejecting its political order. Some of these reasons could arguably even, using Prinz’s terminology, be seen as the effects of “ideational and material power relations” that keep tax evaders in their countries of origin. Of course, it could also be that Floyd is right and that tax evaders do not display political discontent of the relevant kind. My point is that this cannot be settled by observation alone. We also need a background theory of e.g. power or interest that could help us determine which of two conflicting preferences reveals our true commitment. But if this is correct, it will be this theory—not our empirical observations—that does the normative work in normative behaviorism with an ethnographic sensibility.

The third problem is to set the threshold for how much discontent members of the relevant groups must express in order for this to challenge legitimacy. Given that no political order satisfies all of its citizens’ preferences, an attentive ethnographer should be able to detect at least some degree of (conscious or unconscious) discontent in virtually every group under consideration. I might be overall happy with the security and freedom offered by my social democratic political order, but regret that it does not sufficiently reward my particular talents, or that it allows too little leeway for entrepreneurship or eccentric lifestyles. As long as my discontent does not push me to commit crimes or insurrection, however, normative behaviorism tells us that it is too trivial to challenge the legitimacy of the political order. As Floyd puts it: “People always complain. The acid test is when they move themselves to action” (Floyd, 2020: 156, emphasis removed). Ethnographic sensibility, however, offers no similar threshold for when discontent is sincere enough to challenge legitimacy.

One may respond that this problem is a minor one, given that normative behaviorism offers only comparative judgments of political orders. Expressions of trivial discontent will only count for so much, and hence my social democratic order will still perform better than those met by stronger rejections. But this reply ignores the fact that it is the focus on crime and insurrection that allows normative behaviorism to quantify discontent in a way that allows for inter-societal comparison. Once we add a qualitative dimension, comparative assessment will become immensely more complicated if not outright impossible.

Why Ethnographic Sensibility Cannot Fix the Flaws of Normative Behaviorism

The preceding section highlighted problems that arise when we try to supplement normative behaviorism with an ethnographic sensibility. Defenders of ethnographic political theory may respond that my worries only demonstrate the relevance of this approach: What is needed to solve these problems is more, not less, ethnographic interpretation. For example, political ethnographers may argue that careful ethnographic inquiry might provide well-grounded accounts of what kind of behavior reveals political discontent, and what kind of discontent poses a challenge to legitimacy in specific contextual settings. In this section, I will argue that even if true, this will deprive normative behaviorism of the very features that rendered it attractive in the first place. Below, I list six potential merits of normative behaviorism, and show that adding ethnographic sensibility prevents their realization.

• 1. Simplicity. Normative behaviorism offered a simple and straightforward way to measure the legitimacy of political orders. Though simplicity can never be the sole virtue of a normative theory, this arguably gave normative behaviorism a comparative advantage over competing methods. Adding ethnographic sensibility renders the work of normative theorists—with regard to both data selection and analysis—immensely more complicated, depriving normative behaviorism of this advantage.

• 2. Ranking. Normative behaviorism promised to offer a ranking of political orders based on their legitimacy. Adding ethnographic sensibility impedes the fulfillment of this promise. Once we allow for different forms of discontent to enter into our studies, it is no longer clear that these could be weighed against each other in a way that allows for inter-societal comparison. How shall we, for instance, rank a society with high crime and insurrection rates but next to no silent discontent, compared to one with low crime and insurrection rates but where silent discontent is widespread? Due to the fine-grained and context-sensitive nature of ethnographic work, ethnographers are typically reluctant toward both quantification and inter-societal comparison. Since normative behaviorism offers only comparative judgments of political legitimacy, adding ethnographic sensibility will deprive this approach of its normative cut.

• 3. Institutional recommendations. That normative behaviorism with an ethnographic sensibility cannot deliver comparative assessments of political orders also means that it cannot offer institutional recommendations. This is because normative behaviorism based its recommendations of political institutions on the outcome of their implementation, compared to other institutional orders. If we cannot compare different political orders, we thus cannot tell which political institutions are the best.

• 4. Solving the idealization problem. Through avoiding idealizations altogether, normative behaviorism claimed to offer a solution to the idealization problem. I have argued that this ambition ultimately fails, since the use of crime and insurrection as measures of political discontent itself relies on such idealizations. To supplement normative behaviorism with an ethnographic sensibility will, however, add to this problem rather than solve it. To identify how the workings of power affect people’s behavior will require that we imagine a situation where power is absent, or where people are fully informed about its presence and effects. These are idealizations, counterfactually asking how people would behave had they not been subjected to power.

Complementing our ethnographic inquiry with fictional narratives will arguably make things even worse. Such narratives risk to be tainted not only with the ideological bias of the theorist, but also, insofar as the narratives are theory-driven, of the background theory at work. And since fiction—especially when conforming to aesthetic standards like “show, don’t tell”—is “thick,” offering a more complex picture than thought experiments and hypothetical examples (that are usually tailored to isolate a single aspect of the phenomenon under scrutiny) it will be even harder to disentangle the effects of such biases from the narrative’s truth-bearing elements. Hence, if we are skeptical toward the idealizations made in mainstream political theory—like most realists tend to be—we should equally reject the idealizations necessary to conduct political theory with an ethnographic sensibility.

• 5. Overcoming disagreement. Normative behaviorism set out to dispel disagreement in the field of political theory. It did so by offering a simple, conclusive, and non-idealized account of justification. As we have seen, ethnographic sensibility reinvigorates many of the disputes normative behaviorism had claimed to overcome. In addition, it gives rise to new ones, such as disagreements regarding the correct background theory of power or the correct interpretation of agents’ motives. Though normative conclusions are always subject to scholarly dispute, the inevitably subjective element at the core of interpretative methods should lead us to expect the disagreements surrounding ethnographic work to be particularly frequent and pervasive.

• 6. Normative independence. Normative behaviorism claimed to provide an independent method of justification, drawing normative conclusions from people’s responses to political orders rather than imposing moral standards from above. As we have seen, however, ethnographic inquiry will rely on other theories—first and foremost, background theories of power—for example, to determine which groups should be granted ethnographic attention in the first place, and when an act should be interpreted as expressing an agent’s true preferences. Such theories will often involve value judgments, for example, casting power as detrimental to freedom or the pursuit of objective interest. Insofar as this is so, normative behaviorism will no longer provide an independent standard of justification.

The above list is provisory and inconclusive. Still, it shows that normative behaviorism would have to pay a high cost for incorporating ethnographic sensibility. This should perhaps come as no surprise. Prinz proposes two distinctions that could help us map the family of realist theories. The first is that between revisionist and radical realist approaches. While revisionist realism searches for more realistic justifications for the existing liberal democratic order, radical approaches seek to challenge this order by uncovering the power relations at play. The second opposition is that between prescriptive and interpretative realist approaches. While prescriptive realism shares the action-guiding ambitions of moralist political theory, interpretative realism abandons prescription in favor of understanding and critique (Prinz, 2020: 68–69). For each dichotomy, normative behaviorism falls firmly into the first category. It is revisionist in that it defends liberal egalitarian democracy as the (to our knowledge) most legitimate political order. And it is prescriptive because it tells us to build liberal democratic institutions. In contrast, Prinz takes ethnography to constitute a tool for conducting realist political theory of the radical and interpretative kind (Prinz, 2020: 70). If this is correct, we should—given the opposing aims of these approaches—expect the attempt to supplement normative behaviorism with an ethnographic sensibility to be doomed to fail.

Of course, this does nothing to rebut political theory with an ethnographic sensibility as such. Its defenders may happily discard what others take to be the advantages of normative behaviorism. Stevens, for one, explicitly denies that political theory should seek to offer conclusive verdicts about political orders. According to him, all substantive issues that concern political structure or principles are subject to deep disagreement (Stevens, 2023: 17). Likewise, we may argue that understanding is more important than ranking, that disagreement contributes to scholarly progress, or that idealizations, if carefully applied, are useful tools for political theorists. At some point, however, the account we end up with will no longer be recognizable as normative behaviorism. Rather than improving this approach, we will have replaced it with a completely different one.

My argument also does not show that normative behaviorism is beyond rescue. Here, I have only considered the specific attempt to fill its gaps by supplementing this approach with an ethnographic sensibility. There may be other ways to remedy the flaws I have identified, including but not restricted to expanding the set of acts that serve as measures of political discontent and/or restricting the set of crimes considered relevant for this purpose. The success of such attempts will depend on whether they are able to identify reliable indicators of political discontent that do not fall back on e.g. controversial background theories or non-quantifiable measures.

Conclusions

This article has demonstrated that the problems that stem from normative behaviorism’s focus on crime and insurrection as measures of political discontent are more pervasive than previously acknowledged. I have outlined three distinct objections against normative behaviorism, and argued that each of these undermines the reliance on crime and insurrection as measures of discontent. Furthermore, I have argued that supplementing normative behaviorism with an ethnographic sensibility will not remedy these problems. To the contrary, adding political ethnography to normative behaviorism will deprive it of the very features that rendered this approach attractive in the first place. If fixing the flaws of normative behaviorism requires us to transform it beyond recognition, this constitutes a reason to reject it.