Section 1: Introduction

Permissivism can be characterized as the view that there are cases where agents, given their circumstances, have multiple rational options. Permissivism has both practical and epistemic versions which share structural similarities but are tackled in strikingly different ways. Consider the epistemic side first. Epistemic permissivism comes in two main varieties: intrapersonal and interpersonal.Footnote ¹ Intrapersonal epistemic permissivism maintains that given some evidence E relative to proposition p, a single agent has available to them more than one rational doxastic attitude towards p. For example, the evidence that Smith is guilty may permit a range of doxastic attitudes any one of which is permissible for me to have. In contrast, interpersonal epistemic permissivism maintains that given E relative to p, multiple agents have available to them more than one rational doxastic attitude toward p. For example, you and I may have different Jamesian goals (you value avoiding error while I value believing truths) which make it the case that it is rational for me to believe that Smith is guilty while it is rational for you to suspend judgment on whether he is guilty.Footnote ² Interpersonal epistemic permissivism is more plausible than the former, but an intrapersonal one can support an interpersonal one. Since support for intrapersonal epistemic permissivism is rare even among permissivists, the kind of permissivism that is popular is what I call a pure interpersonal epistemic permissivism: an interpersonal permissivism that is compatible with the denial of its intrapersonal sibling.

Practical permissivism similarly maintains that there are cases where agents, given their practical circumstances, have multiple rational act alternatives. This permissivism admits to both intrapersonal and interpersonal varieties like its epistemic cousin, but discussions of practical permissivism focus on intrapersonal cases.Footnote ³ Such cases abound: I find myself with extra money and am rationally permitted either to give it to charity or spend it on a gift for my spouse; I drive home from work and am rationally permitted to take the long, scenic way or the short, less-scenic way. This intrapersonal focus remains even in literature that discusses permissibility in both epistemic and practical rationality (see Schroeder (Reference Schroeder2015), Brunero (Reference Brunero2022), Tucker (Reference Tucker2025), Li and Saad (Reference Li and Saad2023), Berker (Reference Berker2018)).

In the epistemic literature, most of the action is on the interpersonal side, but in the practical literature the action is on the intrapersonal side. This is odd given the structural similarities between epistemic and practical rationality. We might expect that a pure interpersonal practical permissivism (the practical analog of the popular epistemic permissivism) would have received some attention. If such a position is plausible, we need an argument for it. That task is on the practical side, but trouble may rest on the epistemic side. Paul Forrester (Reference Forrester2024) recently argues that epistemic permissivists face great explanatory burdens: (1) they must explain why some cases are permissive while others are not and (2) in permissive cases, they must explain why some options are permissible and others are not. Appreciating the similarities between permissivisms may illuminate ways to meet Forrester’s challenges.

I argue that reasons holism provides an argument for a pure interpersonal practical permissivism, thereby motivating the underappreciated position. Furthermore, if epistemic permissivists adopt reasons holism about epistemic rationality, then they can meet Forrester’s challenges. According to holism, there are contextual factors distinct from the grounds of our reasons that modify our reasons’ weights.Footnote ⁴ So, distinct agents may have the same reasons (in virtue of having the same grounds of their reasons) but the weights of these reasons may differ due to these modifiers. Appealing to these contextual factors can motivate a pure interpersonal practical permissivism, but the epistemic permissivist can tell a similar story: distinct agents can have the same epistemic reasons (in virtue of having the same evidence) but the weights of these reasons differ in virtue of the agents having different epistemic modifiers. This explanation also furthers the discussion on holism since there has so far been no literature on the existence of epistemic modifiers.

We need to do quite a lot to bring out the above results. First, we need to take a trip through epistemic permissivism, then turn to the practical side. Section 2 focuses on epistemic permissivism and notes Forrester’s explanatory challenges. Section 3 unpacks the relevant landscape of practical permissivisms and identifies the unnoticed pure interpersonal practical permissivism. Section 4 discusses reasons holism and how it motivates this practical permissivism. Section 5 returns to epistemic permissivism, showing how if we understand it along holist lines, then we can answer Forrester’s challenges.

Section 2: Permissivisms in Epistemic Rationality

As noted above, we can understand permissivism as the view that agents, given their circumstances, have multiple rational options. Permissivism is a view about how restrictive rationality is regarding how one responds to their circumstances, and so it has epistemic and practical versions residing in epistemic and practical rationality respectively. To draw the conclusions that I think are important, we need to first sketch rationality in general.

While my argument does not require it, I assume a substantive (rather than structural) view of rationality for the sake of instruction. That is, rationality concerns how we respond to reasons.Footnote ⁵ We limit our focus to normative reasons as opposed to motivating or explanatory reasons. The former concerns what justifies actions or attitudes while the latter concerns what motivates agents to have attitudes or perform actions. Adopting a common metaphor of weight in reasons, some reasons outweigh others, and rationality recommends that agents abide by their weightiest (or sufficiently weighty) reasons. A sufficient reason is a reason that is weighty enough to make the relevant action or doxastic attitude permissible. This might be because the reason is not outweighed, or it is weighty enough with a certain kind of weight (a justifying weight). So, agents are rational insofar as their actions and attitudes abide by their sufficient reasons. One’s set of sufficient reasons depends on their set of reasons and their respective weights. The weight of a reason is largely determined by those facts which are the grounds of it, but there may be other factors (discussed in section 4) that influence the weight of reasons. Practical and epistemic rationality will naturally differ in what kinds of considerations get to be grounds of one’s epistemic/practical reasons, but we can call this category “one’s circumstances.” So far, we can maintain that rationality roughly concerns the following three-step process:

The above should be relatively uncontroversial since it is standard to consider how reasons weigh against others, and it is standard to consider what considerations are the grounds of these reasons.

The plausibility of epistemic permissivism relies on the strictness of epistemic rationality. Epistemic rationality is largely concerned with how agents respond to their evidence and form doxastic attitudes.Footnote ⁶ We can put these concerns into the general three-part process described above to get a better picture of what epistemic rationality looks like in a reasons-focused form. One may hesitate in this endeavor since it may be unclear how exactly evidence (a central player in epistemic rationality) figures in the relevant three-step process. I suggest, following Errol Lord (Reference Lord, Lasonen-Aarnio and Littlejohn2023, 409), that we can consider evidence as a ground of one’s epistemic reasons. As he puts it, “some consideration being evidence for p is sufficient for that consideration to be a normative epistemic reason to believe p.” Here is the idea: we consider our doxastic attitudes as justified by our evidence, and normative reasons are those things that justify our actions and attitudes. Intuitively, at least some of our epistemic reasons seem to be grounded in our evidence and it is important to distinguish between the reason itself and its grounds. Distinguishing in this way helps us see what normative roles different factors play (this will be important for later). Accepting that one’s evidence grounds epistemic reasons is compatible with most conceptions of evidence, so this move is relatively ecumenical. Even so, I will largely discuss evidence in propositional terms for the sake of simplicity. I also acknowledge that there is debate about whether evidence always grounds epistemic reasons or whether there can be non-evidential grounds for epistemic reasons (See Schroeder (Reference Schroeder2015), Brunero (Reference Brunero2022), Lord (Reference Lord, Lasonen-Aarnio and Littlejohn2023)). To avoid distraction, I will limit talk of the grounds of epistemic reasons to cases where those grounds are one’s evidence.

Epistemic reasons favor having certain doxastic attitudes, and each epistemic reason has weight. As with rationality more broadly, agents are epistemically rational insofar as they abide by their sufficient epistemic reasons (those sufficiently weighty epistemic reasons). Following the above, we have a three-part process that epistemic rationality concerns:

It is conceptually possible that there is more than one sufficient epistemic reason at the end of this process. This is where we turn to the epistemic permissivist. Employing the above three steps, the epistemic permissivist maintains that something in (ERi)-(ERiii) makes it the case that at least sometimes there are multiple rationally permissible doxastic attitudes towards some proposition. There are two main kinds of epistemic permissivism since the factors constraining the rational options of a single agent may differ from those that constrain the rational options of multiple agents. In the previous section, we noted the distinction between intrapersonal and interpersonal epistemic permissivism. Here is how we can understand this distinction for the remainder of the article, understanding E as total evidence:

So long as the agents involved are sufficiently similar, cases of intrapersonal epistemic permissivism will also be cases of interpersonal epistemic permissivism. After all, one way for distinct agents to have multiple rational options is for each individual to have multiple rational options. Things do not necessarily go the other way. A case of interpersonal epistemic permissivism need not also be a case of intrapersonal epistemic permissivism. Two agents might each have one rational doxastic attitude to have towards p, but they differ regarding which doxastic attitude is rational for them to have. This possibility, however, motivates a third kind of position: an interpersonal epistemic permissivism that does not rely on an intrapersonal one. On this view, distinct agents have multiple rational options not in virtue of each agent individually having multiple rational options. So, the agents would have different sets of rational options (put another way: they would have different sets of sufficient epistemic reasons). For example, it might be the case that you and I have the same evidence relative to the proposition that we have free will, but we have different (rational) ways of responding to our evidence. You are more inclined to avoid error whereas I am more inclined to risk error for the sake of having a true belief (see Kelly (Reference Kelly, Steup and Turri2013)). So, we might have different sets of rational options: you are rational only to suspend judgment or disbelieve that we have free will while I am rational to suspend judgment or to believe that we have free will. While we have different rational options, this is not explained by both of us having the same, non-singleton set of options. Here is how I characterize this position:

A pure interpersonal epistemic permissivism is likely popular among epistemic permissivists. At least, in light of the fact that so few permissivists defend intrapersonal epistemic permissivism (and some explicitly deny itFootnote ⁷) interpersonal epistemic permissivists plausibly endorse pure interpersonal epistemic permissivism. A pure interpersonal epistemic permissivism, to be clear, does not rely on the denial of intrapersonal epistemic permissivism. It might very well be the case that two agents are in an intrapersonal epistemically permissive case where each has multiple rational options but each has a different set of rational options. Such a case would be a case of intrapersonal epistemic permissivism and a case of pure interpersonal epistemic permissivism. A pure interpersonal epistemic permissivism is simply an interpersonal epistemic permissivism with the added restriction that the sets of rational options be different between agents rather than merely requiring a plurality of rational options between agents.

Epistemic permissivism enjoys some support and its fair share of objections. I turn my attention to a recent objection from Forrester who argues that permissivists face serious explanatory burdens. He argues that permissivists must explain why some cases are interpersonally permissive while others are not (assuming that the permissivist denies that all cases are permissive). Here’s how Forrester (Reference Forrester2024: 1271) puts it: “Permissivists must explain why some bodies of evidence are such that they rationalize more than one doxastic attitude toward a proposition, and others are not.” Additionally, permissivists must explain why, in permissive cases where not every attitude is permitted, some attitudes are permitted and others are not. In his (2024: 1271) words, “The permissivist must explain the scope of epistemic permission not only across the cases, but also within each case.” While I do not pretend to have ready-at-hand the details to discharge these burdens in every possible permissive case, I think we have the tools for a general response available to the permissivist by appealing to holism.Footnote ⁸ Before we do that, however, I turn to practical rationality.

Section 3: Permissivisms in Practical Rationality

Practical rationality is famously permissive—we have an abundance of cases where agents have more than one rational act alternative.Footnote ⁹ Most obviously, many non-moral cases exhibit this. I drive home and have at least two rational act alternatives: take the long, scenic way home or take the short way home. Even when we look at quintessentially moral cases, we find a plethora of permissive cases. If I and five strangers are stranded and a lifeboat approaches unable to save us all, I have multiple rational act alternatives: save myself or direct the lifeboat to the strangers, saving them (see Parfit (Reference Parfit2011: 186-7)). If I find myself with extra money, it seems I have multiple act alternatives: give it to charity or spend it on a gift for my spouse. Cases like these litter the ethical literature on permissions. Indeed, Joseph Raz (Reference Raz1999: 100) calls the basic belief the commonsense thought that in the majority of choice situations, agents have multiple rational act alternatives. He rightly maintains that its intuitive power is strong enough that we need not defend it. Instead, we need to only explain it.

Working with Raz’s basic belief first, there are multiple explanations on offer. Raz (Reference Raz1999: 102-104) himself maintains that in many choice situations, agents encounter incommensurable reasons. When reasons are incommensurable, it is a mistake to weigh them against each other. Since these reasons cannot properly be weighed against each other, the actions they favor remain rationally available, providing the agent with multiple rational act alternatives. For example, on Raz’s own view, if impartial reasons are incommensurate with self-interested reasons, then we can explain the rational options in the above lifeboat case: the reason to save myself cannot be weighed against the reason to save the others, so I am rational in choosing to do either. Alternatively, Joshua Gert (Reference Gert2004a) maintains that agents have multiple rational act alternatives when there are reasons that have more justifying strength than opposing reasons with requiring strength. A justifying strength generates rational permissions while a requiring strength generates rational prohibitions. In many cases, reasons have merely justifying strength but not requiring strength. In these cases, the agent is rational to abide by such reasons but not required to do so. In our lifeboat case, for example, there are altruistic reasons to save the other survivors, but altruistic reasons do not have requiring strength. Self-interested reasons likewise only have justifying strength, making it the case that I have the two rational act alternatives described in the case (see also, Tucker (Reference Tucker2022c)). Douglas Portmore (Reference Portmore2012) provides a different explanation, maintaining that in most choice situations, agents have imperfect reasons: reasons that favor achieving some worthy end but leaves it up to the discretion of the agent how to achieve it. We have imperfect reasons in cases where we have multiple ways of achieving the same end, and Portmore claims that most choice situations are such cases: we have multiple ways of achieving a reasonably choice-worthy future. Each option favored by an imperfect reason is a rational act alternative, so we can explain Raz’s basic belief. I do not plan on defending any of the above views; presenting them is merely instructive for showing how the permissiveness of practical rationality has been presented in the extant literature.

We also find permissiveness in practical rationality in the growing literature on hard choices. These are choices between options where one option is better than another in some respect, the other is better in a different respect, but the options are not equally good. Ruth Chang (Reference Chang2002, Reference Chang2005, Reference Chang2017) maintains that such choices arise when the options are on a par. Two options are on a par when they are comparable, but it is not the case that one is better, worse, or equally as good as the other. When an agent is stuck in a hard choice, says Chang, the choice is permissive. Others (Broome (Reference Broome and Chang1997), Constantinescu (Reference Constantinescu2016), Dorr et al. (Reference Dorr, Nebel and Zuehl2023), Williams (Reference Williams2016)) claim that hard choices arise from indeterminate values or reasons. On these views, rationality recommends that the agent chooses the best option but, due to the indeterminacy of the values or reasons, it is indeterminate which option is best. Still others (Gert (Reference Gert2004b) and Tucker (Reference Tucker2022b)) appeal to Gert’s distinction between justifying and requiring strengths of reasons to explain hard choices, accounting for the permissiveness of the choice in a similar way as we saw above.

Practical permissivism maintains that, given some circumstances, there are multiple rational act alternatives. An act alternative is rational when there is a sufficient practical reason (or reasons) for it. These sufficient reasons, of course, depend on one’s set of practical reasons, and one’s practical reasons are grounded in action-relevant facts of one’s circumstances. For the sake of this discussion, let’s say that these facts ground one’s practical reasons.Footnote ¹⁰ As with rationality more broadly, agents are practically rational when they abide by sufficient practical reasons. Which reasons are one’s sufficient practical reasons depends on the weights of one’s practical reasons. Drawing from the above three-part process that rationality concerns, we can understand the process that practical rationality concerns in the following way:

A practical permissivist maintains that we get multiple rational act alternatives due to one of the three parts of this process or the steps between them. For example, the weighing of the different reasons in (PRii) may get us multiple sufficient reasons in (PRiii) when the weighing results in multiple, equally weighty reasons.

What kind of permissivism does the above support? In light of the discussion of the previous section, the above approaches to the plurality of rational options in practical rationality support this:

If we accept this intrapersonal practical permissivism, then we can also motivate its interpersonal cousin:

When there are multiple agents who are in a relevantly similar case of an intrapersonal practical permissivism, then it is possible that this will also be a case of interpersonal practical permissivism. If, given my practical situation, I have a sufficient reason to take a nap and a sufficient reason to work then you (who are just as tired and busy as me) would also have a sufficient reason to nap and a sufficient reason to work.

It should also be clear that the above motivations do not support the practical analog of pure interpersonal epistemic permissivism, the more popular kind of epistemic permissivism. That is, just as with the epistemic discussion, there is an interpersonal practical permissivist position where the fact that agents have multiple rational act alternatives does not rely on the truth of an intrapersonal practical permissivism. Here is such a view:

The above motivations support various kinds of intrapersonal practical permissivism, but a pure interpersonal practical permissivism has gone unnoticed in the literature on rational act alternatives due to the near exclusive focus on intrapersonal cases. If, as we are told, practical rationality is more permissive than epistemic rationality, then we should expect that this position is plausible.

It is worth noting that even the literature addressing the permissiveness of both epistemic and practical rationality exhibits an intrapersonal focus. Han Li and Bradford Saad (Reference Li and Saad2023), for example, explore epistemic and practical permissivism, but only explain why there is a multitude of intrapersonal practical permissive cases and a dearth of epistemic ones. In a similar vein, many are concerned with the differences between how epistemic and practical reasons weigh. Mark Schroeder (Reference Schroeder2015), John Brunero (Reference Brunero2022), and Chris Tucker (Reference Tucker2025) all provide independent explanations for why equally weighty practical reasons result in an intrapersonally permissive case while equally weighty epistemic reasons do not (at least not obviously) (see also Berker (Reference Berker2018)). The point here is not that intrapersonally permissive (or impermissive) cases are uninteresting. Instead, the point here is that the literature on practically rational options focuses only on intrapersonal cases while the literature on epistemically rational options considers both intra- and interpersonal cases. Perhaps the focus on intrapersonal practical permissivism is because, as noted above, cases of it will also be cases of interpersonal practical permissivism so long as the agents involved are sufficiently similar. So, all the action is on the intrapersonal side. Even so, the near exclusive focus on intrapersonal practical permissivism seems to have led to an underappreciation of a permissivist view that does not rely on it.

To be clear, a pure interpersonal practical permissivism is compatible with accepting an intrapersonal practical permissivism. The former merely maintains that two agents can have different sets of sufficient reasons, not that these sets need to be singleton sets. It is possible that each agent has multiple rational act alternatives, but these are different act alternatives than what other agents in the same circumstances have. The next section motivates pure interpersonal practical permissivism. Once we see its plausibility, we can better appreciate just how permissive practical rationality is and what the epistemologist can learn from it.

Section 4: Holism and Pure Interpersonal Practical Permissivism

A pure interpersonal practical permissivism is plausible once we identify the normative roles of facts that may differ between agents. We see the relevant normative roles in reasons holism which enjoys a number of defenders (see Dancy (Reference Dancy2004), Bader (Reference Bader, Lord and Maguire2016), Lord (Reference Lord2016), Dorsey (Reference Dorsey2016), Tucker (Reference Tucker2022a), Cullity (Reference Cullity and Star2018)). According to holism, a reason’s weight is sensitive to context. We focus on a reason’s weight because it is the facet through which we understand how reasons relate to actions, attitudes, and other reasons. All reasons are grounded in considerations that have normative weight; if some consideration does not have weight or loses its weight, then it does not or no longer grounds a reason. The weight of a reason consists of two things: (1) polarity and (2) magnitude. Polarity indicates whether the reason favors or disfavors some action or attitude. Magnitude determines how much the reason favors or disfavors, how weighty the reason is. To put it in terms of a well-worn metaphor of a balance scale: polarity determines which side of the scale the reason ends up on while the magnitude determines how much that side descends.

Holism contends that (regarding reasons) considerations can play three normative roles. The first is the ground of the reason. The ground of a reason is that fact that has the normative weight (with its polarity and magnitude). This weight may be sensitive to context, but if we abstract away contextual factors, we see the default weight of the reason. That is, the default weight of the reason is the weight it has in virtue of the ground alone. When we take into consideration the influences of contextual factors, we see the embedded weight of the reason (the weight it has when embedded in a particular context). The embedded weight might differ from the default weight (the polarity or magnitude might change due to context), but it might not. Importantly, when we are determining those sufficient reasons that rationality requires an agent abide by, we look at the embedded weight of reasons. To do otherwise would be to ignore normatively relevant context.

The two other normative roles that considerations can play are those contextual factors that influence the weights of reasons. Our second normative role is the condition: a contextual factor that determines whether some consideration has weight at all. These come in two forms: enablers are considerations that make it the case that some fact has weight; disablers make it the case that some fact does not have weight. For example, the fact that it is raining outside is not usually a reason to go for a drive. The rain, though, will rinse off the mud on my car. So, here, the fact that it is raining is a reason to go for a drive and is enabled by the fact that my car is dirty. Without an enabler, the default weight would be 0, but the embedded weight of the reason (with the enabler) would be >0. In contrast, disablers do the opposite. The fact that you promised to Φ is usually a reason for you to Φ, but if your promise was made under duress, then you are no longer bound to this promise. The fact that you promised under duress disables the reason grounded in the promise (see Dancy (Reference Dancy2004: 38-39)). In this case, the default weight of the reason would be >0, but the embedded weight would be 0.

The last normative role is another contextual factor: the modifier. Modifiers are multipliers of the magnitudes of weights and come in two varieties: intensifiers are multipliers >1, and attenuators are multipliers <1. Modifiers operate differently from conditions. For example, enablers make it the case that some consideration has normative weight, but an intensifier increases this weight’s magnitude. If the consideration is not a ground of a reason, then there is no weight for the intensifier to modify. Likewise, if some reason is disabled, it is no longer a reason, and there is no magnitude to decrease, nullifying any potential influence an attenuator would have. To illustrate modification, we can adapt an example from Dancy (Reference Dancy2004: 42). Consider the fact that Ana needs help. This fact is a ground for the reason to help Ana. The fact that Betty is the only other person in the room is itself not a reason to help Ana since Betty’s location is not tied to Ana’s needing help. The fact that Betty is the only other person in the room, however, intensifies the reason for Betty to help Ana—Betty has a weightier reason to help Ana than someone outside of the room. Consider a variation of this case where it is Ana’s own fault that she needs help. The fact that Ana’s needing help is her own fault seems to attenuate the reason to help her—Betty would have a less weighty reason to help. Notably, while some facts that are modifiers are not grounds of reasons (like the aforementioned intensifier), some may be. The fact that Ana’s needing help is her own fault may be an attenuator and a ground of the reason to scold her (though not a ground of a reason to refrain from helping her).

Assuming that holism is plausible, we can easily motivate a pure interpersonal practical permissivism. Recall that this permissivism maintains that there are cases where agents have different sets of sufficient reasons even though they share the same considerations that ground their practical reasons. The practical permissivist can claim that there are contextual factors that, when differing between agents, make it the case that each agent has a different set of sufficient reasons even when in the same circumstances. In other words, the practical permissivist can maintain that when agents are in the same circumstances, they would have the same sets of sufficient reasons according to the default weights of reasons, but there are cases where they differ regarding the embedded weights of reasons. This is because one’s circumstances (as I have been using this term) concern the grounds of one’s reasons and, as a result, determine the default weights of these reasons.

Though Holists differ in the particular agent-relative facts to which they appeal, the general picture across views is largely the same: the fact that an agent takes something to be especially important to them is an intensifier. Lord (Reference Lord2016), for example, appeals to the fact that some agent cares about some X to explain permissible partiality regarding X. The idea is that what permits me to rescue my brother from a pond when he and a stranger are drowning, and I cannot save both is explained by the fact that I care about my brother. Both my brother and the stranger have equal value in virtue of being persons, and these values plausibly ground reasons to rescue each one. The fact that I care about my brother intensifies the reason for me to rescue him, but it does not intensify the reason to rescue the stranger. If I can only rescue one person, the reason for me to rescue my brother outweighs (according to its embedded weight) the reason to rescue the stranger. In this way, Lord explains permissible partiality by appealing to agent-relative modifiers.

In a similar vein, Dale Dorsey (Reference Dorsey2016) appeals to a wide variety of what he calls “facts of self.” These are facts like that one has a certain character, that one has certain commitments, that one has certain projects, and that one has certain practical identities. According to Dorsey, these facts make a difference in what an agent’s sufficient reasons are by functioning as intensifiers. For example, Dorsey describes how aesthetic reasons are stronger for someone with the practical identity of an artist even though these reasons are reasons for everyone to advance beauty:

On Dorsey’s view, everyone has a reason to advance beauty, but artists have weightier reasons to do so. We explain this increased weight in terms of the fact that they have certain commitments (to being an artist) or the fact that they have certain practical identities (as artists) intensifying the reasons to advance beauty. Importantly, Dorsey seems to understand (at least some) facts of self in terms of what an agent takes to be especially important to them. In the above case, part of what it is to have a commitment to being an artist is to take beauty to be especially important to you.

Holists like Lord and Dorsey focus their attention on intensifiers, but we can also tell similar stories about attenuators. Consider again the fact that rescuing someone from drowning is a reason to do it. Suppose that you discover that the drowning person is your worst enemy. The fact that they are your enemy seems to attenuate the reason to save them. Indeed, if you had to choose between saving your enemy and saving a stranger, you would intuitively have weightier reason to save the stranger. The change in weights is not explained by an intensification of the reason to save the stranger but by the attenuation of the reason to save your enemy. Mirroring the above story of intensifiers, it is plausible that the fact that you take this person to be especially unimportant is an attenuator.

We can provide other holist accounts that focus on modifiers, but we currently have enough to motivate a pure interpersonal practical permissivism. The application is straightforward: there are cases where distinct agents are in the same circumstances (they share those considerations that ground their practical reasons), so they have the same set of practical reasons, but these agents have different sets of sufficient practical reasons in virtue of having different modifiers. Assuming that the holists noted above are correct, the modifier that makes the relevant difference is the fact that one agent takes something relevant to the situation to be especially important to them while the other agent does not. Since modifiers influence the weights of reasons, they may change which reasons are sufficient reasons for each agent. If so, then there are cases where agents are in the same circumstances but have different sets of sufficient reasons. Even though pure interpersonal practical permissivism has received little attention, we have compelling support for it from holism.

Section 5: Holism for the Epistemic Permissivist

We can now look back to the epistemic side of the discussion. Holism provides powerful tools to also explain how exactly a pure interpersonal epistemic permissivism works. To see this, we need to take a detour through two representative epistemic permissivist views.

Our first epistemic permissivist view is from Miriam Schoenfield (Reference Schoenfield2014) who argues that there is a plurality of rationally available epistemic standards and that agents can accept different epistemic standards, resulting in different sets of rational options between them. Schoenfield understands one’s epistemic standards as functions between one’s evidence E and the doxastic attitudes the agent forms in light of E. Importantly, according to Schoenfield (Reference Schoenfield2014: 199), one’s epistemic standards are separate from E. So, this view can motivate an epistemic permissivism along the following lines: there can be cases where agents consider the same E but differ in their epistemic standards. Due to the differing epistemic standards, the agents may differ in their doxastic attitudes (either due to a difference in epistemic reasons or merely a difference in sufficient epistemic reasons), and both may be rational in so differing. One might think that we need to appeal to our evidence to justify our epistemic standards, but Schoenfield (Reference Schoenfield2014: 202) maintains that this misunderstands how epistemic standards work. Trying to justify our standards by appeal to evidence would rely on another set of epistemic standards that help us draw conclusions (like that an epistemic standard is justified) from our evidence, and these epistemic standards would then require their own justification.

The second epistemic permissivist view to look at is from Anjan Chakravartty (Reference Chakravartty2017) who argues that there are multiple rationally adoptable epistemic stances which influence the rationally available doxastic attitudes for agents. A stance is one’s orientation towards some subject, and he argues that one’s epistemic stance significantly influences how one forms beliefs about scientific ontology. However, Chakravartty’s view has implications beyond such beliefs. Chakravartty’s idea is similar to Schoenfield’s: there is a plurality of rational epistemic stances that may inform an agent how to respond to their evidence. Agents may rationally differ in their doxastic attitudes because each agent has a different (but still rational) epistemic stance. More relevant for my purposes is that he (2017: 202) argues that one’s epistemic stance involves guidelines for how one should behave epistemically and form beliefs. Here’s Chakravartty:

An epistemic stance, then, is similar to what Schoenfield calls an epistemic standard. While they may disagree on the details, Schoenfield and Chakravartty maintain that there is some aspect of an agent’s epistemic perspective, independent of their evidence and the requirements of rationality, that informs agents what doxastic attitudes to form in light of their evidence. Just like Schoenfield’s position, the role played by Chakravartty’s stances may lead agents to have different epistemic reasons or may make it the case that agents merely have different sufficient epistemic reasons.

Schoenfield’s and Chakravartty’s approaches appeal to some third factor (separate from one’s evidence and the proposition it supposedly favors) and exemplify one of the main ways permissivists motivate their positions (see also Callahan (Reference Callahan2021)). Others (Titelbaum and Kopec (Reference Titelbaum, Kopec, Jackson and Jackson2019)) argue for similar positions by appealing to ways of reasoning, how much one values the Jamesian goals of not believing what is false and believing what is true (see Kelly (Reference Kelly, Steup and Turri2013)), or the stakes one has in the issue (Rubin (Reference Rubin2015)). The structures of such permissivist views are largely the same. Importantly, epistemic standards (or stances, ways of reasoning, or Jamesian goals) may differ between agents, establishing a permissivism reliant on such differences. Positions like Schoenfield’s and Chakravartty’s, then, almost exclusively motivate a pure interpersonal epistemic permissivism. There are, of course, those who would accept both intrapersonal and interpersonal epistemic permissivisms, but what explains how distinct agents have different rational options need not also explain how a single agent has multiple rational options.

The above permissivist positions provide compelling cases to support their conclusions. On our reasons-focused story, we need an explanation of what normative roles these standards and stances play. Consider the way epistemic standards and stances may influence one’s belief-formation processes: Schoenfield (Reference Schoenfield2019: 199) describes epistemic standards as functions from bodies of evidence to doxastic states which the agent takes to be truth conducive, and these functions might be understood as rules for forming beliefs or beliefs about how best to form other beliefs. Chakravartty provides examples of how one’s stances might make one hesitant to draw certain conclusions from their evidence. Taking E as the grounds of one’s epistemic reasons (the contents of (ERi)), we have two different roles that are compatible with Schoenfield’s and Chakravartty’s descriptions. First, it might be the case that the standards and stances play some role between steps (ERi) and (ERii) (between the grounds of one’s epistemic reasons and one’s set of epistemic reasons), leading different agents to have different epistemic reasons. Second, it might be the case that the standards and stances play some role between steps (ERii) and (ERiii) (between the set of one’s epistemic reasons and the set of one’s sufficient epistemic reasons), leading different agents to have different sets of sufficient epistemic reasons even with the same epistemic reasons.

Pure interpersonal epistemic permissivism can be explained via holism in the same way as a pure interpersonal practical permissivism. In short, epistemic standards and stances function as modifiers of epistemic reasons. There is some motivation for this explanation in the epistemic literature we see above. In describing what an epistemic stance is, Chakravartty surveys several accounts of stances, concluding that “[stances] may be described in terms of what, for a given agent, has value, or what one values most about the outputs of some scientific or other investigation into the ontology of the world.” (Chakravartty (Reference Chakravartty2017: 215). His idea, then, is that one’s values influence the kinds of conclusions they draw from their evidence. While Chakravartty does not provide an account of valuing, it is plausible to take valuing (especially Chakravartty’s use of “values most”) to be in the same neighborhood as taking something to be especially important. If this is right, then we have some motivation for the appeal to holism in the same way as in the practical realm: valuing something modifies the reasons relevantly related to it. At least, if we think that holists have correctly identified the normative role of taking something to be especially important in the practical realm, we should be inclined to think that we can extend this explanation to the epistemic realm.

If epistemic standards and stances function as modifiers, then they play a role in the step between (ERii) and (ERiii) (between the set of one’s epistemic reasons and the set of one’s sufficient epistemic reasons). The resulting explanation of pure epistemic permissivism would be similar to what we gave for pure interpersonal practical permissivism above. One may resist this thought, though, reminding us that evidential support is indeed a three-place relation between some evidence, the relevant proposition, and some background information (see Jackson and LaFore (Reference Jackson, LaFore, Lasonen-Aarnio and Littlejohn2023: 361-362)). It is standard to think that this background information consists of the relevant prior probabilities or one’s epistemic standards. So, the worry goes, we already have a role for epistemic standards (and maybe stances): they are part of evidential support and so are part of the grounds of one’s epistemic reasons. I suspect that standards and stances (at least of the kind permissivists appeal to) perform a different role than merely being part of the evidential support relation. To be clear, I accept that evidential support is indeed a three-place relation and the relevant priors are part of this relation. I am merely skeptical that the standards and stances to which permissivsts appeal play (only) this role.

Here is a short argument to justify my skepticism. If epistemic standards or stances play the abovementioned role in evidential support, then we can translate this into holist terms: the standards or stances would function as conditions. This is because what counts as evidence (considerations with evidential support) depends on one’s standards. Chakravartty even notes that one’s stance can inform one about what to ignore (and this sounds like a disabler) (see Chakravartty (Reference Chakravartty2017: 210-211)). As conditions, standards and stances would play a role between (ERi) and (ERii) (between the grounds of one’s epistemic reasons and one’s set of epistemic reasons), but the resulting position would hardly count as an epistemic permissivism. According to epistemic permissivism, the agents in the relevant case supposedly have the same evidence, but if what counts as evidence depends on one’s standards or stances then it is difficult to see how the agents would differ in these but would not also differ in their evidence. Conditions change what considerations function as grounds of one’s reasons—the agents would differ in their epistemic reasons but this difference would be due to a difference in the grounds (a difference in evidence). So, we would violate the requirement that agents have the same E. A conditions-focused explanation of these kinds of epistemic permissivism, then, is implausible. Furthermore, this way of understanding epistemic permissivism robs permissivism of an interest we intuitively think it has. After all, epistemic permissivism is initially interesting because it supposedly informs us about some underappreciated part of epistemic rationality. If all the permissivist does is appeal to conditions, then their story is ultimately uninteresting—we already know that differences in the grounds of reasons will result in different reasons. In light of these concerns, I suggest we look elsewhere to understand the role of standards and stances.

According to the present holist view, pure interpersonal epistemic permissivism is explained by distinct agents having different sets of sufficient epistemic reasons in virtue of the fact that they have different modifiers. In light of the above, the explanation is straightforward. Each agent shares their total evidence which grounds their epistemic reasons. The weights of these epistemic reasons are sensitive to the modifiers each agent has, resulting in a difference in the embedded weights of these reasons. In cases where agents differ with respect to their epistemic stances or standards, they have different modifiers which result in each having different sets of sufficient epistemic reasons. Since sufficient epistemic reasons identify the rationally available doxastic attitudes to have towards p, we have an explanation of pure interpersonal epistemic permissivism. In contrast to the conditions-focused explanation above, a modifier-focused explanation does not entail that different information gets to count as evidence for different agents and we see an interesting upshot about epistemic rationality: there are epistemic modifiers!

With holism in mind, we can answer Forrester’s challenges. Recall that he argues that (1) permissivists must explain why some cases are interpersonally permissive while others are not (assuming that the permissivist denies that all cases are permissive) and (2) permissivists must explain why, in permissive cases where not every attitude is permitted, some attitudes are permitted and others are not. One may think that permissivists like Schoenfield and Chakravartty have a ready explanation to satisfy these burdens: permissive cases are those where agents have different standards or stances, and the permitted attitudes are those favored by the set of one’s evidence and standards or stances. Forrester responds that such positions merely push the burden back one step: now we need to explain which standards or stances are permissible and which ones are not. Schoenfield and Chakravartty may have their own answers to this challenge, but I think appealing to holism allows us to see another route.

Suppose that standards and stances are modifiers. We can discharge the first explanatory burden at a general level. First, a case is a permissive one just in case there are epistemic modifiers that differ between the agents in the case. We should think that there are indeed epistemic modifiers in light of what I argue above, namely, that there are significant structural similarities between epistemic and practical rationality and if there are modifiers in the practical realm, then, plausibly, there are modifiers in the epistemic realm too. The epistemic permissivist can best be understood as contending that there are cases where agents differ between their modifiers in virtue of having different standards or stances. If the permissivist is right to think that standards and stances function as modifiers, then these modifiers will perform their relevant normative roles whenever they are present and relevant in the case. Accordingly, asking which modifiers are permissible in a case is mistaken since, when they are relevant, they are part of the context that the agent cannot rationally ignore. Modifiers, after all, are normative phenomena that partly determine the (embedded) weights of the agent’s reasons. Next, we may wonder when modifiers are relevant for a case. I find Dorsey’s way of putting the power of modifiers helpful: modifiers reflect the agent being more or less susceptible to certain values in a case. When an agent values explanatory power (or takes explanatory power to be especially important to them), for example, they are more susceptible to the epistemic value of explanatory power. This fact about modifiers helps us see that only those that concern the values of a case will influence the weights of the reasons in it. When an agent values something that does not concern an epistemic value in the case, there will be no modification.

Turning to the second explanatory burden, what explains the range of permissible attitudes? The short answer is that the range of permissible attitudes for an agent is determined by the embedded weights of their epistemic reasons. Each modifier reflects some values (or susceptibilities) of the agent and modifies according to the strength of these values. As noted above, this can also be understood in terms of what values the agent is more susceptible to. The range of permissible attitudes is determined by the values/susceptibilities of the agents that modification reflects. We can even put this in terms of weights and a balance scale: the range of permissible attitudes is determined by the embedded weights on a balance scale—the weightiest epistemic reasons favoring certain attitudes are sufficient epistemic reasons and the agent is rational when abiding by them. By putting the discussion in terms of the familiar language of reasons and weights, the explanatory problem Forrester posits evaporates.

Holism, then, helps the permissivist discharge explanatory burdens from Forrester at a general level. This discussion also advances the discussion on holism since there is no discussion of epistemic modifiers. When holism about epistemic reasons is discussed, it is regularly put in terms of conditions. Dancy (Reference Dancy, Hooker and Little2000), (2004: 49ff) famously motivates holism about practical reasons because he finds holism about epistemic reasons so plausible, but his description of holism in the epistemic realm focuses on enablers and disablers. That I see some object looking red is normally a reason to believe that it is red, but if I know that I just took a pill that makes blue objects look red then the reason is disabled—that it looks red is no longer a reason to believe that it is red. While holism regarding epistemic reasons is generally considered plausible, the role of epistemic modifiers has been underappreciated. In light of the above, we can have a new appreciation for what the permissivist shows us: there are agent-relative modifiers of epistemic reasons.