The power to tolerate


Eduardo Fuentes Caro[1]


Abstract: Toleration is a privilege of the powerful. A person can tolerate only if she has the power to interfere with what she objects to. In this paper, I revise the circumstances in which someone has the power to tolerate. I argue that the answer must incorporate both metaphysical and practical considerations. Specifically, I argue that someone has the power to tolerate when her properties ground that power in the practical situation in which the person finds themselves, characterized by her practical interests. In doing so, I criticize Glen Newey’s analysis of this power.


Keywords: Toleration. Power. Practical possibility.



There are at least two features of toleration in the literature that are controversial from a normative point of view. First, toleration involves objections. “To tolerate is to insult,” as Goethe (1998, p. 87) said. One can only tolerate that which one regards as bad, wrong, evil, or the like. Some, like Heyd (2008), have argued that this precludes toleration from being a proper part of democratic politics. Second, toleration involves power. Toleration is a privilege of the powerful. (Svensson, 2011). Someone can tolerate only if she has the power to interfere with what she objects to. This means that toleration implies relations of power of one over another (of course, nothing precludes the latter from having power over the former in other circumstances). Many authors have claimed that this means that there is no place in a democratic society for such a non-egalitarian concept.[2]

This second feature is the theme of this paper. We can distinguish between two problems: normative and conceptual. The first is concerned with the rightness of tolerating others, given that it entails power relations. Concretely, the normative problem of political toleration is whether a democratic society should allow people to interfere with the lives of others. The conceptual problem, on the other hand, is concerned with the conditions under which someone has the power to tolerate the practices, beliefs, etc. of another person. I will focus on the latter and not dwell on possible normative difficulties as such.

This paper is organized as follows. First, I present Newey’s answer to the conceptual problem. Second, I argue that Newey’s account fails because it rests on a misleading counterfactual analysis of power. It amounts to a version of the “simple conditional analysis” of dispositions that has been, I will argue, successfully refuted by several philosophers working on the metaphysics of dispositions. Although this metaphysical discussion might seem off-topic at first, I argue that it is necessary to flesh out our judgments about alleged cases of toleration. Third, I will offer an alternative account. Unlike Newey’s account, mine is based on the metaphysical literature on dispositions (or “powers,” as I will call them), and argues that someone has the power to tolerate when her properties, given the practical circumstances she finds herself in, ground that power. Notably, this means that attributions of power are partially dependent on the interests of agents because a combination of metaphysical powers and practical interests constitute practical situations. Importantly, I will place the alternative account in the context of Newey’s distinction between holistic and quasi-naturalistic approaches to toleration, and I will explain why my proposal supports the latter. Finally, I end with some brief remarks on the role of empirical investigation in philosophical analysis of toleration.


1 Newey’s theory

It is common to find at least three components of toleration in the literature (Cohen, 2014). First, there is the objection component: one can only tolerate what one objects to. Second is the acceptance component: one must accept what one objects to in a principled way, that is, in virtue of reasons. Finally, there is the power component: one can only tolerate if one has the power to interfere with the objected practice.[3] Most of the work that has been done on the first two components is conceptual, probably because the power to interfere seems to be self-explanatory. It might appear that a mere empirical review would be required to determine whether or not an agent can interfere with a practice; however, as the subsequent discussion will show, that is far from true.

Of all the power component analyses, Newey’s (2013, 2017) is, without a doubt, the most developed.[4] In summary, he argues that S demonstrates the power to tolerate when she displays the right attitude to power: she would not prevent the disapproved of practice out of respect for moral principles, particularly out of respect for other people’s freedom. There are courses of action that, although broadly within our power, are inaccessible to us because they are not real options in deliberation; they are not narrowly within our power because they go against our ethical dispositions. In the third section, I show my support for the focus on actions not open to deliberation but eschew the role Newey attributes to moral principles as such, particularly to the principle of respect for freedom. More to the point, Newey judges the distinction between pragmatic costs and moral principles as crucial to his analysis. He argues that the former, but not the latter, voids toleration. I try to show that this distinction unnecessarily complicates our understanding of toleration.

The broad lines of Newey’s analysis are described here. At the most basic level, the power condition, as Newey calls it, requires that the tolerator must have the power to prevent (or at least interfere with) the disapproved of practice. More precisely, Newey thinks that the tolerator needs to believe she has the power to prevent it. Newey therefore agrees with many philosophers working on toleration: what matters is not so much the actual possession of power but the belief that one possesses it (COHEN, 2004, p. 93; FORST, 2013, p. 25; MCKINNON, 2005, p. 14). The overriding thought is that agents lacking power can be tolerant if they believe they have that power; that is, we are more interested in their attitude toward power than in the power itself. Thus, Newey thinks that not all limitations of power breach the power condition.

Consider the following case:

Martha has it within her power to stop her daughter Lois from going out one evening, but fails to do so. However, Martha fails to do so only because the costs of prevention are unacceptably high to her: Lois has convincingly threatened that she will hurl a chair through the window unless she gets her own way. Martha decides not to use the parental power she has to thwart Lois’s plans. Here, although Martha fails to use her power to stop Lois from going out, her failure to use her power results only from the fact that to do so would inflict unacceptable costs on her, Martha. (NEWEY, 2017, p. 429).


On the face of it, it is implausible to claim that Martha tolerates her daughter’s going out. Lois has somehow limited Martha’s power, but how? Martha still has the capacity to prevent Lois from going out, it just happens that she is not willing to bear the costs. It seems that sheer power does not suffice for toleration, however, because Lois’s threat has not changed Martha’s power. The problem appears to be Martha’s attitude toward her power, because were the cost lower she would otherwise exercise it. Thus, Newey proposed that the power condition must take costs into account. After a few proposals, he arrives at the following formulation:

Power Condition 3 (PC3): S can only tolerate a disapproved-of action or state of affairs p if (i) S believes that it is within her power to prevent (etc.) p, and (ii) it is not the case that S would seek to prevent p, if S believed she had the power to do so even if, in S’s judgement, prevention did not incur undue costs. (NEWEY, 2017, p. 431).[5]


S does not tolerate if she shows the wrong attitude and is unwilling to bear the costs. PC3 has the consequence that usual ways of describing a modus vivendi are flawed, since they describe it as an instance of toleration. In a modus vivendi, parties to a conflict do not accept peace because they find it itself valuable; they would rather see their ideas prevail, yet they realize that they are unable to impose them upon others. Take the Belfast Agreement, for example, which represented a stalemate of sorts and was designed to end violence in Northern Ireland. It is naive to think that all signatories believed that they got everything that they wanted or that they would not have demanded more in their ideal world. It is all too obvious that Sinn Fein, to name just one signatory, would have pushed for complete secession from the UK if the costs of doing so had not been unduly high. Hence, the Belfast Agreement modus vivendi and others like it are not examples of toleration.

However, Newey argues that, there is a difference between judging not signing the Belfast Agreement as too costly because of the pragmatic consequences and as too costly because it would breach a moral principle. He does not want moral principles to limit the agents’ power in the same manner that pragmatic costs do. If Martha were committed to respecting autonomy, she would regard her preventing Lois from going out as too costly in terms of Lois’s loss of autonomy. To Newey, this sounds very different from judging prevention as too costly because of the psychological discomfort of an argument with her daughter.

Consequently, PC3 must be read as referring only to pragmatic costs. Nevertheless, one may object that both moral principles and pragmatic costs have the same result–the inability to prevent a practice. By way of reply, Newey appeals to idealization. If, in the ideal world, she does not interfere with the objected practice, then she displays toleration, and if, on the other hand, she does interfere with the objected practice, then her noninterference in the actual world is due to pragmatic costs. The reason is that in S’s ideal world, pragmatic costs do not exist, but moral principles remain in place.

However, an obvious difficulty arises. If the world is so ideal as not to include costs,[6] then it would not contain any practice that S could object to. The consequence is that in an ideal world, toleration cannot even exist. Therefore, the idealized distinction between cost and principles comes at the expense of toleration.

To rescue toleration, we need a reason why even ideal worlds would be sub-optimal–that is, why ideal worlds would include disapproved of practices. Why would S choose to actualize a world that is not completely ideal? Newey sees this question as analogous to the theological problem of evil: Why would a perfect God choose to actualize a world riddled with evil? Following strategies designed to deal with the latter question, Newey concludes that a commitment to the value of freedom offers the best answer. Since being free entails people behaving in ways not approved of by the tolerator, the ideal world must include that freedom as part of its structure. Thus, in her ideal world, the tolerator would allow disapproved of practices because respecting people’s freedom is part of what makes that world ideal.

Consequently, there are courses of action that we have the power to take but cannot be entertained in deliberation because they go against our principles: even in our ideal world, we would not take them. We “cannot” take them, not because we lack the power to, but because, being against our principles, they are not real options for us in our process of deliberation. Specifically, the principle of respect for freedom seems to be the one principle that can limit options without breaching the power condition.

Wrapping up, what does it take for S to have the power to tolerate? Curiously, she does not actually need to have the power to interfere with the relevant practice. She only needs, according to PC3, to believe that she has the power and display the right attitude; that is, she would not interfere with the practice in her ideal world. It seems that power, as a real feature of the world, has vanished from the analysis.[7] In the next section, I will press on with that point.


2 Actual power

I mentioned that toleration is a privilege of the powerful. Newey’s account changes that, turning it into something closer to a privilege of the virtuous. Significantly, it makes the very concept of toleration more palatable than (one may argue) it should be. When philosophers like Brown (2009) critique toleration, they aim for the power relations that lie at its core. They object to toleration because accepting it entails that some agents in the actual world have the power to interfere with the lives of others. Newey’s idealized analysis ends up attributing the power to tolerate to how agents would behave in their merely possible ideal world. Power’s “[…] significance lies in the underlying attitude of the putative tolerator, which can be brought out by examining the counterfactuals which apply.” (NEWEY, 2017, 433). His solution to the conceptual problem, in my opinion, too easily eliminates the normative one: if you believe that you have the power and tolerate in your ideal world, then you have the power to tolerate regardless of how weak you are in the real world.

Let us be clear. There are no knock-down arguments against the “possibilist” approach as such. Someone may prefer to analyze toleration in terms of what agents believe and ideally would do so because of their interest in how much agents resemble virtuous agents.[8] Both Jubb (2019) and Swift (2008) convincingly show that the analysis of political concepts depends on what we want to achieve. I opt for an “actualist” approach because I want an analysis that will be useful to political agents when they confront political conflicts. More importantly, for the reasons to be detailed below, I want to isolate the power of agents from their beliefs and attitudes toward it. Nevertheless, there is a decisive argument against Newey’s version of the possibilist approach, so there is a strong reason to reject that approach. Unless, of course, a better version is found in the future.

His proposal fails because it relies on fallacious inference. He thinks that from the proposition ‘Martha has the power to F’, it follows the proposition ‘if Martha were subject to stimulus C, then Martha would F’, where the stimulus C is Martha’s finding out about Louis’s plans. His account is an instance of the simple conditional analysis (SCA).[9] Before delving into the details, I will show why he is committed to the inference and why it is prima facie problematic. After that, I will draw on somewhat technical arguments against SCA made by several philosophers working on the metaphysics of power.

Newey (2017, 433) claims that “[…] the counterfactual sifts correct and mistaken judgments by asking how agents would act in similar circumstances, adjusted to take account of their relevant capacities.” In other words, if Martha (in her ideal world) would not thwart Lois’s plans, then we could conclude that the judgment that ‘Martha has the power to thwart Lois’s plans’ is correct. Counterfactual idealization is underpinned, as already seen, by Newey’s belief that what matters is the attitude toward (believed) power instead of actual power. Counterfactuals, Newey holds, are well suited to analyzing the dispositions (attitudes) of agents.

Rainer Forst, among others, shares the shift from actual to believed power. Forst (2013, p. 25-26) says that […] a minority which is not equipped with such power can also adopt an attitude of tolerance and be of the (uncoerced) conviction that, were it to have sufficient means of power at its disposal, it would not use them to the disadvantage of others.” It follows that [...] to conclude that the tolerating party must be in a position of power from which it could effectively prevent the practices in question is unfounded. (FORST, 2013, p. 25). It seems clear that once we opt for believed power, counterfactual idealization comes naturally.

This counterfactual analysis is prima facie problematic for two reasons. First, it relies on the importance of attitudes toward power. While it is true that some prefer this “possibilist” approach, for real-life political agents, mere attitudes do not matter much. Political agents normally care about what their opponents can do and what they themselves can do. They may care about beliefs, but those cases are exceptions. Take a conflict between the State and a political group. The latter wants to know what it is in the power of the State to do because it must prepare for the ways that, say, the State may respond to (punish) its actions. For the group, what the State believes about its own power is only relevant inasmuch as it relates to the power the State in fact yields. Second, and more importantly, it rules out those agents whose disposition to prevent is not yet settled from being potential tolerators. Otherwise said, there may be agents regarding which there is no fact of the matter as to whether they would prevent a disapproved of practice or not. For example, we would like to know whether contemporary States have the power to tolerate without assuming that their dispositions are settled.[10] After all, it should be possible for States to have the power to tolerate and actually to tolerate if they were to accept certain reasons; this possibility should not mean that States would thereby change their power.[11]

Moreover, there may be agents that have the power to tolerate yet are not disposed to interfere. That is, if we understand that being disposed to interfere with p means would prevent p from happening’. As I will argue, an agent can have the power to interfere while still being not disposed – in that sense – to interfere. The gist of the problem is that the possession of relevant power cannot be taken directly from the counterfactuals that are true of the agent.

We can formulate the latter difficulty more rigorously. Newey’s account is an instance of simple conditional analysis (SCA). According to SCA, an object x is disposed to F if and only if, if x were subject to stimulus C, then x would F. Among philosophers working on power, this analysis has fallen out of favor. It has since been replaced with a refined version, famously defended by David Lewis (1997). However, regardless of how refined the analysis may be, the same problems befall any conditional analysis. Due to space limitations, I will only touch on the conditional analysis and the main objections against it to show why Newey’s account fails.

The guiding principle of SCA, or any conditional analysis, is that powers can be understood without loss by reference to a conditional, where the antecedent and consequent of the conditional present a stimulus to x and an action by x, respectively. The qualification means that once the conditional is in place, there is nothing more to say about the power. Thus, to know the conditional is to know all that matters about power.[12] Yet, a common objection is that conditionals cannot account for our intuitions if we pay attention to the fact that between the stimulus and the manifestation many things may occur.

A power to F can be interrupted so that its manifestation does not occur. In Martin’s (1994) terms, powers may be finkish. In Martha’s case, something might interfere with her power to prevent Lois from going out. She could suffer a stroke just after deciding to thwart Lois’s plans. Hence, from Martha has the power to prevent Lois from going out’, it does not follow that if Martha were subject to stimulus C, then Martha would prevent Lois from going out’.

Let us see with more precision how this applies to Martha’s case. The stimulus is Martha finding out about Lois’s plans. Under normal circumstances, this would trigger Martha’s manifestation of her parental power. However, between her finding out and the manifestation, Lois introduces the threat of hurling a chair through the window. The threat gets in the way, acting as a fink. The important point is that it is not necessary for there to be an actual fink, as the mere possibility of its existence suffices for us to reject the conditional analysis. The fact that Lois could execute that threat is enough to conclude that having parental power does not entail any conditional.

Finks work by removing power after the stimulus takes place. Strictly speaking, Martha had power until Lois made her threat. Because Newey believes that Martha still has the relevant power, he must introduce the distinction between costs and principles to explain conflicting intuitions.

Besides finks, there are antidotes (BIRD, 1998).[13] They work by changing the subject that receives the manifestation of power. Lois could ask her grandmother to give her permission to go out, preempting Martha’s use of her parental power. Martha does not lose the power because nothing has changed in her; the issue is that Lois became immune to it. If by parental power we mean Martha’s capacity to command her daughter, then Martha has not lost it; however, if by power we mean the capacity to ensure that her commands will be followed no matter what, then she never had it. That is what antidotes show. It is unwise to deny that an agent has some power just because the patient could become immune to it, because that would mean that nothing has any power whatsoever.

This may seem misguided. After all, what does it mean to say that Martha has parental power other than to say that some conditionals are true of her? Sure, from Martha has the power to thwart Lois’s plans’, it follows if Martha were subject to stimulus C, then she would thwart Lois’s plans unless prevented from doing so’. In other words, it is obvious that if Martha had that power, then there would be circumstances under which she would exercise it. If these circumstances did not exist, then it would be impossible for her to manifest her power. However, a power that is impossible to manifest is no power at all; it is an unactualized possibility that cannot ever be actualized, which is tantamount to nothingness. The key to understanding the problem lies in the word prevent. As Oderberg (2007, p. 137) noted, […] prevention essentially involves interference with a thing’s doing what it is capable of doing. Thus, to know that Martha’s thwarting of Lois’s plans can be prevented, we must first know that she has the power to thwart her plans. Otherwise, if we did not know that Martha had that power, we would not know that she could be prevented from manifesting it. Therefore, conditionals are only informative insofar as they clarify what a power entails, not in analyzing that power from scratch. Newey tries to derive all his understanding of power from conditional analysis, which leads him to idealization and the distinction between costs and principles, but for that analysis to make sense, it presupposes that we already know what it means for an agent to have the relevant power.

Looking back, although Newey wanted to analyze the power condition, he ended up relying on intuitions about idealized counterfactuals. I have argued that his proposal, as for any conditional analysis of power, fails because it is unintelligible without previous grasp of the power to be analyzed. Politically, this is important because it allows us to differentiate between the power that agents have and the conditions that can interfere with that power. We often face situations where an agent would not do F despite having the power because there may be finks or antidotes operating. For instance, the State might not prevent a riot because an international organization would remove one of its intrinsic properties (required by the power) were the State to act. Does that mean that the State lacks the power? No, for the international organization has not removed anything yet. It means that the power does not entail a conditional, that is a significant item of knowledge for political actors. Arguably, political agents do not simply want to know possible outcomes but what the agents involved are capable of. Political life is dynamic; hence, it is not enough to know what would happen in any given situation; we need to know what would happen in other situations, too. By knowing the power of agents, we can evaluate conditionals in that situation and others, and not the other way around.

Where to look if not to conditionals? From what I have argued, it follows that we must investigate agents themselves and their powers. I turn now to those.


3 could have, would have, and the agent’s practical situation

Newey (2000) makes an important distinction between holistic and quasi-naturalistic approaches to toleration. The former fleshes out the components of toleration using a normative theory independent of the agent. John Horton (1996, p. 42) follows this approach when he claims that […] there is no reason why we must accept people’s own perspective on the validity of their objections to something. Mutatis mutandis, the holistic approach, does not attribute power according to the agent’s perspective. Philosophers with the right theory make the judgments on the power to prevent. To illustrate, recall that Newey’s analysis of power makes use of the putative tolerator’s ideal world. A holistic approach could define an ideal world for a person according to principles not held by her, whereby that ideal world would differ from the one she herself would choose. Therefore, the philosopher’s judgment and her own judgment may disagree as to whether she has the power to prevent. Note that Newey is aware of this possibility, for he cautions us not to idealize too much and, in particular, rejects idealizing our basic principles away. (NEWEY, 2017, p. 438).

On the other hand, a quasi-naturalistic approach does not refer to normative criteria independent of the agent. It assumes her normative point of view. With this approach, agents qua agents make the judgments about power. The analysis I will present below can be read as part of a larger defense of this approach because it shows how a plausible analysis of one of the components of toleration can be made by paying close attention to the normative framework of political agents.

In the most general sense, S has the power to F if and only if S can bring about F. For instance, S has the power to walk if and only if S can bring (herself) to walk. Unlike what the conditional analysis implies, having power is not a fact about how the agent behaves in some possible world, but about the agent itself. The power to F is a property of agents, but what kind of property is it? Borrowing from the recent burgeoning literature (see the references in CHOI; FARA, 2018), I take it to be a dispositional property whose manifestation is or constitutes F. That is, the power to thwart Lois’s plans is a dispositional property whose manifestation is, or constitutes, Lois’s plans being thwarted.

For those with Humean inclinations, to say that Martha has that power because she has a dispositional property may sound like tautology. It is reminiscent of Molière’s famous virtus dormitiva. Yet, there is nothing mysterious about it. I am not claiming that Martha has a property whose entire nature consists in giving her that power. Although fundamental properties could be, as Molnar (2003) calls them, pure powers–being nothing but powers–the agential powers I am interested in are not. When we say that someone has the power to jump or hail a taxi, we do not imply that the power admits no further grounding; it is quite the opposite, because we can easily explain what features account for it. There is something about legs that allows people to jump. The same goes for agents like the State. It has the power to shut down businesses because it commands the police, and thus, if it comes to it, it can apply physical force over business owners. Of course, to fully understand the dynamic, we would need to explain why the State has the power to command the police, among many other things. There is a pattern here: we explain why an agent has a power by identifying the properties grounding it. I remain noncommittal with respect to the nature of grounding and leave open the possibility that, as Heil (2003) argued, the dispositional property is, in fact, an aspect of the categorical properties that ground it. Be that as it may, what matters is that we can conclude that an agent has some power by looking at the properties grounding it.

Which of her categorical properties grounds her agential powers? First, there are her physical properties. She needs to have certain physical features to issue and enforce her commands. For example, she needs to have a voice, hands to signal, or any other mechanism to express herself. Second, there are her mental properties.[14] She must have a mental configuration that allows her to form relevant beliefs and intentions. Plausibly, that entails having certain beliefs and desires. If she lacked some basic beliefs, she would be unable to form the belief that Lois was planning to go out. Likewise, if she lacked the basic desire to act on what she believed, she could not form the intention to thwart Lois’s plans. As it turns out, these properties are common to most people. Things change when we investigate her relational properties. Being Lois’s mother is a relational property that grounds her parental authority. Nobody else has that relation to Lois, and it is one property that cannot be taken away from Martha. Undoubtedly, there are cases where being a mother fails to ground parental authority because other properties (intrinsic or relational) change. In any case, we can streamline our exposition if we assume that ceteris paribus Martha’s relation to Lois grounds parental authority.

With these distinctions in hand, we can now revisit Martha’s situation. Lois’s threat does not seem to remove any of the above properties as it affects neither her physical properties nor her relational property; however, the question remains whether it affects her mental properties. Note that even if it does, it does not follow that Martha did not have the relevant power, as I remarked earlier. Newey’s conditional analysis obscures the fact that powers can be removed by a fink and then regained after the fink disappears (or mutatis mutandis by antidotes). He should have concluded that when we ask in the abstract whether Martha has the power, the answer is affirmative, but when asked after Martha receives Lois’s threat, the answer is negative.

However, on closer inspection, we see that Martha does not lose her power. The threat does not remove the mental properties necessary to form the intention to prevent Lois from going out. The assertion that the costs are unacceptably high is ambiguous: either she cannot accept them or she would not accept them. Only in the former case does she lose her power. This follows from the rejection of the conditional analysis, for we have rejected the inference from ‘she would not F’ to ‘she lacks the power to F’. Although I raised the rejection by means of finks and antidotes, the general lesson to be drawn is that would notdoes not equate to cannot. Hence, it is possible that, under no circumstances, Martha would accept those costs even though she could. Why would she not? Well, because she does not want to. In the accepted possible worlds parlance, she does not accept the costs in close worlds, although she does so in farther worlds. Following the familiar suggestion by Lewis (1973), she would F is restricted to close worlds. Obviously, if we change our selection function and include all possible worlds, no matter how distant, there are circumstances where she would F if she could F. We need to keep in mind that when we say that Martha would not prevent Lois’s plans from occurring because the costs are too high, we are not saying that in no circumstance whatsoever would she not prevent them. To put it bluntly, it depends on whether we are thinking in theoretically absolute terms or in more down-to-earth ones.

Toleration is executive, as Newey (2000) argues. That is, it is freely chosen. It is also salient in very particular circumstances, namely when the world is not the one we would like it to be. Lest we forget, it is only when we are faced with suboptimal alternatives that toleration becomes relevant. Thus, there are always costs involved. In some sense, we would like never to have to tolerate that which we object to. Sometimes, the costs of prevention are bearable, and we decide not to tolerate, yet sometimes, they are not, and we decide to tolerate. Even so, unbearable costs do not imply that we will avoid preventing the practice; we must still choose what to do. Recall that an unacceptably high cost could simply mean that we would choose not to bear it (in close possible worlds), not that we could not.

Interestingly, here we find a confusion between the power and the acceptance components. The cost of prevention being unacceptably high could be a reason to accept the practice. For example, Martha would not thwart Lois’s plans because of this powerful reason: the cost does not constrain her power but her rationality.

Nonetheless, I have not discussed what happens when Martha truly cannot accept those high costs. In this case, she lacks the power to prevent and hence cannot tolerate. Her mental properties are changed by the belief that the costs are too high. She is no longer able to form the relevant intention. There may be psychological barriers that she cannot overcome. Obviously, we could ask if she could, in principle, overcome them even if she could not do it given the present circumstances. However, it is likely that Lois’s threat does not pose such an insurmountable psychological challenge. After all, Martha could start reading Stoic philosophy and stop caring about broken windows and other external goods. Whatever could be intended, Martha could also intend, given a suitable change in her mental composition. That is too abstract, however. By asking whether Martha has the relevant power, we ask whether her properties ground that power. That is, the properties she currently has, not the ones that she might have only in principle. Therefore, it is possible that Martha prevents Lois from going out does not signify that Martha has the power to prevent Lois from going out.

We also run the risk of being too concrete. It is true that, at a given point, Martha could lack the properties to ground the relevant power or have one interfering with her other properties.[15] Nevertheless, with a reasonable amount of effort and time, she could form the intention to take a course of action by modifying her mental properties to (re)gain power. If Martha needs to become a Stoic philosopher to become immune to Lois’s threat, then we are idealizing too much, but if she only needs to take a breath and count to ten, then she could become immune with reasonable effort. Note that I am assuming that if she does not pause, she cannot become immune and thus regain her power. From the discussion on finks and antidotes, we learned that the manifestation of power takes time. This has interesting consequences. Sentences attributing powers can refer to different propositions, depending on the timeframe we choose for the manifestation to occur. Consider the sentence “Egyptian pyramids are fragile,” which attributes to the pyramids fragility (the power to break easily under moderate pressure). Yet what ‘easily’ means depends on the timeframe we choose. If the timeframe between the stimulus and the effect was set at five seconds, then the pyramids could not be considered fragile since they would not break under those conditions. On the contrary, if the timeframe was set at five million years, they could be considered fragile, as they would break within that time under moderate pressure. Note that I am not saying that powers themselves depend on our conceptual decisions; they effect causal processes in the world. We can use concepts without time limits, and then anything that brings about any manifestation, regardless of how long it takes, is a power. However, limiting concepts is useful for highlighting real differences; that is, it is useful to take our sentences as expressing propositions involving more limited concepts. In talking about toleration, it is useful to distinguish between Martha having the power to change her mental properties rapidly and the power to do so slowly.

Faced with a practice to which we object, we usually have a timeframe that allows us to evaluate whether or not to tolerate it. Thus, we can take that span of time to consider whether the agent is able to acquire the power to prevent. Can the State prevent a protest? It may lack the required property at first (say, a parliamentary declaration) but be able to obtain it before the protest begins. If Martha can regain her power to thwart Lois’s plans after the threat has been issued, but before Lois goes out, then we can say that she has that power. This is a welcome result because it fits the dynamic nature of our practical lives. Seldom are we interested in what an agent can do at any one instant. Drawing from Gilabert (2017), let us distinguish between soft and hard constraints. The former can be overcome given enough time and resources, unlike the latter. In the circumstances of toleration, the time between the occurrence of the objected practice and our practical interests determine how much time is enough. If the objected practice has not yet occurred, the time of its occurrence determines the time limit. We can hardly say that an agent can prevent a practice that has not happened if she can only do so after it happens. It would go against our (usual) interests when talking about toleration. Thus, if the objected practice has not occurred, all constraints that cannot be overcome before it occurs are considered hard. On the other hand, if the objected practice is already happening, then the timeframe is almost entirely determined by our interests.

I have arrived at a somewhat paradoxical conclusion. I criticized Newey for failing to account for real power, only to end up claiming that having the power to prevent is relative to our interests. However, note that only our attributions of powers are relative to us, and even then, not completely so. Using an unrestricted concept of power, only those constraints that cannot be overcome in principle are classified as hard. If the costs were so high that Martha could not overcome them, even with infinite time and resources, then she would (unrestrictedly) lack power. By restricting the concept, we are paying attention to the fact that toleration is a practical concept; we need it to describe how agents act. Therefore, it is not strange for our practical interests to limit the concept of power. Furthermore, it is important to remark that, unlike Newey, I am not attributing power to where it does not exist but, instead, limiting existing power.

More precisely, what matters are the practical interests of agents qua agents. As philosophers, we can adopt a detached position and try to figure out the interests that agents should have according to our theories. This would be a holistic approach. Yet agents qua agents cannot afford to judge in a detached manner. They are already immersed in a practical conflict that cannot be modified at will. Martha must respond to her situation in a limited time and with the resources that she can gather at that time. She sees the situation in a particular way; therefore, her practical interests are molded by that. What she cares for depends on how she interprets her situation. To her, it might be important, say, to respond within a couple of minutes to Lois’s threat. Newey seems to think that Martha may fail to prevent Lois’s actions because of costs or principles. But Martha could come to believe that if she does not respond quickly, the situation would become absurd. That ship has sailed, she might think. One needs not turn into an existentialist to appreciate how some actions can stop making sense to us. Maybe Martha ought not to describe her situation like that, but she does. A quasi-naturalistic approach has no problems accommodating her description of her situation.

In summary, I have argued that someone has the power to prevent an objected practice from occurring when she has the properties that ground that power. I showed that even when she has them, that does not entail that she would exercise the power, so counterfactual analysis gives us no insights. We need to inquire into the agents’ current properties and how they ground their powers or fail to ground them. Additionally, I defended a quasi-naturalistic account of the power component that includes the agent’s practical interests alongside the metaphysical consideration of properties. Notably, I made no reference to the distinction between principles and costs. Whenever I say cost’, ‘or principle’ may be added to the phrase. As Newey himself acknowledges, in politics, it is usually difficult to differentiate between pragmatic costs and principles, even for agents themselves. Hence, it is best to discard the distinction altogether.


Concluding remarks

I strove for an analysis of the power component, which is both metaphysically grounded and rooted in the agent’s practical situation. Toleration involves agential powers, which we are free to manifest. A purely metaphysical description misses that point since it cannot account for the practical impossibility of manifesting a power that we objectively have. This is why objective powers must be restricted, depending on the practical situation of the agent.

I have made use of toy examples to be as precise as possible, but it must be recalled that real life is much messier. Therefore, it is often impossible to know an agent’s practical situation without doing fieldwork. This is particularly true when dealing with political toleration, because agents in conflicts tend to hide or disguise their real interests. The upshot is that a quasi-naturalistic approach, like the one that I have advanced, must be willing to go beyond philosophy into what the social sciences tell us about agents and their practical situations.


El poder de tolerar

Resumen: La tolerancia es un privilegio de los poderosos. Una persona puede tolerar sólo si tiene el poder para interferir con aquello que objeta. En este artículo analizaré las circunstancias en las que alguien tiene el poder para tolerar. Argumentaré que la respuesta debe incorporar consideraciones tanto prácticas como metafísicas. Específicamente, defenderé que alguien tiene el poder para tolerar cuando sus propiedades fundan tal poder en la situación práctica en que ella se encuentra, según es caracterizada por sus intereses prácticos. Al hacerlo criticaré el análisis del mencionado poder realizado por Glen Newey.


Palabras clave: Tolerancia. Poder. Posibilidad práctica.



BALINT, P. Acts of tolerance: A political and descriptive account. European Journal of Political Theory, v. 13, n. 3, p. 264–281, 2014.

BIRD, A. Dispositions and Antidotes. The Philosophical Quarterly, v. 48, n. 191, p. 227-234, 1998.

BROWN, W. Regulating Aversion: Tolerance in the Age of Identity and Empire. Princeton, NJ: Princeton University Press, 2009.

CHOI, S. Do Categorical Ascriptions Entail Counterfactual Conditionals? The Philosophical Quarterly, v. 55, n. 220, p. 495-503, 2005.

CHOI, S.; FARA, M. Dispositions. In: ZALTA, E. (ed.) The Stanford Encyclopedia of Philosophy. Stanford, CA: Stanford University, 2018.

COHEN, A. J. What Toleration Is. Ethics, v. 115, n. 1, p. 68-95, 2004.

COHEN, A. J. Toleration. Cambridge: Polity, 2014.

FORST, R. Toleration in Conflict: Past and Present. Cambridge: Cambridge University Press, 2013.

GILABERT, P. Justice and Feasibility: A Dynamic Approach. In: VALLIER, K.; WEBER, M. E. (ed.). Political Utopias: Contemporary Debates. Oxford: Oxford University Press, 2017. p. 95-126.

GOETHE, J. W. V. Maxims and Reflections. London: Penguin Books, 1998.

HEIL, J. From an Ontological Point of View. Oxford: Clarendon, 2003.

HEYD, D. Is toleration a political virtue? In: WILLIAMS, M. S.; WALDRON, J. (ed.). Toleration and Its Limits. New York: New York University Press, 2008. p. 171-194.

HORTON, J. Toleration as a Virtue. In: HEYD, D. (ed.). Toleration: An Elusive Virtue. Cambridge: Cambridge University Press, 1996. p. 28-43.

JOHNSTON, M. How to Speak of the Colors. Philosophical Studies, v. 68, n. 3, p. 221-263, 1992.

JUBB, R. On What a Distinctively Political Normativity Is. Political Studies Review, v. 17, n. 4, p. 360-369, 2019.

KING, P. T. Toleration. London: Frank Cass, 1998.

LEWIS, D. Counterfactuals. Malden, MA: Blackwell, 1973.

LEWIS, D. Finkish Dispositions. The Philosophical Quarterly, v. 47, n. 187, p. 143-158, 1997.

MARTIN, C. B. Dispositions and Conditionals. The Philosophical Quarterly, v. 44, n. 174, p. 1-8, 1994.

MCKINNON, C. Toleration: A critical introduction. New York: Routledge, 2005.

MOLNAR, G. Powers: A Study in Metaphysics. Oxford: Oxford University Press, 2003.

NEWEY, G. Virtue, Reason and Toleration: The Place of Toleration in Ethical and Political Philosophy. Edinburgh: Edinburgh University Press, 2000.

NEWEY, G. Toleration in Political Conflict. Cambridge: Cambridge University Press, 2013.

NEWEY, G. Modus vivendi, Toleration and Power. Philosophia, v. 45, n. 2, p. 425-442, 2017.

ODERBERG, D. S. Real Essentialism. New York: Routledge, 2007.

PLANTINGA, A. The Nature of Necessity. Oxford: Clarendon,1974.

RYLE, G. The Concept of the Mind. London: Hutchinson, 1949.

SVENSSON, M. A Defensible Conception of Tolerance in Aquinas? The Thomist, v. 75, n. 2, p. 291-308, 2011.

SWIFT, A. The value of philosophy in nonideal circumstances. Social Theory and Practice, v. 34, n. 3, p. 363-387, 2008.


Recebido: 03/05/2021

Aceito: 31/01/2022


[1] Faculty of Humanities and Social Sciences, Universidad San Sebastián, Santiago – Chile. ORCID: E-mail:

[2] Among several authors, Brown (2009) is probably the one who has most thoroughly developed that critique.

[3] I will henceforth use “practice” to refer to the object of toleration whatever its nature. I follow Newey (2000) in this usage.

[4] King (1998, p. 22-24) offers another analysis.

[5] Instead of the original “A” and “P”, I write “S” and “p” for consistency with the rest of the paper.

[6] From now on, by “costs”, I mean “pragmatic costs” and by “principles”, I mean “moral principles”, unless explicitly noted.

[7] “[…] one could conclude that what is necessary for toleration is not to have the power of prevention, but not to be disposed to use the power if one believes one has it.” (NEWEY, 2017, p. 429).

[8] Indeed, this is what Newey (2017, p. 440) points out.

[9] See Ryle (1949) for an early example.

[10] An anonymus reviewer was puzzled by the claim that States have dispositions. Since they are entities of the material world, I see no problem with attributing them dispositions (or powers, etc.). It just means that they are “inclined” towards some effect. Insofar as it makes sense to talk about things being “inclined” to any effect at all, it makes sense to talk about States being disposed. Of course, I am assuming that States do exist in the material world, which could be (albeit implausibly) disputed. For a defense of the reality of inclinations of things, see Molnar (2003).

[11] “We do not know how those who do not yet have the opportunity will ultimately act in conditions of real power.” (BALINT, 2014, p. 268).

[12] Molnar (2003) convincingly argues that although there is only an indirect connection between reductionism and the conditional analysis, the latter is generally advanced as part of the reductionist program.

[13] There are also masks, which prevent the power from being exercised. See Johnston (1992).

[14] Of course, if reductive materialism is true, the latter supervenes on the former. Be that as it may, it is useful to distinguish between them.

[15] For example, the belief that Lois’s threat could easily materialize can ‘fink’ her other properties. In the literature on powers, this is called an ‘intrinsic fink’. See Choi (2005).