Synthese

, Volume 190, Issue 17, pp 3865–3888

Acting for reasons, apt action, and knowledge

Authors

    • Philosophisches InstitutUniversität des Saarlandes
Article

DOI: 10.1007/s11229-012-0230-8

Cite this article as:
Mantel, S. Synthese (2013) 190: 3865. doi:10.1007/s11229-012-0230-8
  • 257 Views

Abstract

I argue for the view that there are important similarities between knowledge and acting for a normative reason. I interpret acting for a normative reason in terms of Sosa’s notion of an apt performance. Actions that are done for a normative reason are normatively apt actions. They are in accordance with a normative reason because of a competence to act in accordance with normative reasons. I argue that, if Sosa’s account of knowledge as apt belief is correct, this means that acting for a normative reason is in many respects similar to knowledge. In order to strengthen Sosa’s account of knowledge, I propose to supplement it with an appeal to sub-competences. This clarifies how this account can deal with certain Gettier cases, and it helps to understand how exactly acting for a normative reason is similar to apt belief.

Keywords

Acting for reasonsActing for a normative reasonGettier problemKnowledgeApt beliefNormative reasonsExplanationCompetence Virtue epistemology Apt action

1 Introduction

In this paper I argue for the view that there are important similarities between knowledge and acting for a normative reason. These similarities are easiest to see when one first considers belief and action and then turns to knowledge and acting for a normative reason. Both beliefs and actions have standards of success: beliefs are assessable in terms of being true or false, and actions are assessable in terms of being favored or disfavored by normative reasons. With this structure in place, knowledge can be compared with acting for a normative reason. In both cases, a standard of success is met: knowledge involves true belief and acting for a normative reason involves action that is favored by a normative reason. Furthermore, knowledge involves justification and some further component that guards against Gettier cases. On my view, acting for a normative reason involves a component that is analogous to justification and a further component that guards actions against being favored by a normative reason by mere luck, much like an anti-Gettier condition guards beliefs against being true by mere luck.

The topic of this paper is part of the project to explore how theoretical and practical philosophy can learn from each other. I focus on an account of knowledge that is already part of this project, namely, Ernest Sosa’s Virtue Epistemology. Sosa draws on practical considerations in developing his account of knowledge. He applies some terms we use to talk about performances, such as ‘success’, ‘credit’, and ‘competence’, to the domain of belief. His analysis of knowledge as belief that is true because of competence relies on viewing belief as a performance. Accordingly, he describes and illuminates knowledge by comparing it to action.

My aim is to describe the similarities between acting for a normative reason and knowledge by reference to Sosa’s account of knowledge. Apart from some sketchy remarks in Sect. 5, I will have to leave it open as to whether some similarities between knowledge and acting for a normative reason could equally be stated without a virtue-theoretic account of knowledge. My view is that there are general similarities between these phenomena, and only this general character of the similarities explains why one can give a rough description of them without reference to any specific account of knowledge. This is what I outlined in the first paragraph of this introduction. Therefore, I would not be surprised if some similarities between knowledge and acting for a normative reason could also be stated by not using Sosa’s account of knowledge but another one.1

The paper consists of five sections and a conclusion. In Sect. 2 I point out exactly what notion of acting for a normative reason I am concerned with. In Sect. 3 I present Sosa’s account of apt action and his analysis of knowledge as apt belief. I propose to illuminate aptness by reference to sub-competences and point out how this helps to deal with certain Gettier cases. In Sect. 4 I point out a kind of apt action that differs from the one that Sosa focuses on. While Sosa focuses on what I call ‘intention apt action’, I will concentrate on what I call ‘normatively apt action’. In the fifth section I argue that acting for a normative reason can be understood as normatively apt action.

2 Different notions of acting for a normative reason

The notion ‘acting for a normative reason’ needs clarification with respect to the expression ‘normative reason’ and the expression ‘acting for’. A normative reason for an action is a fact (or, more generally, an entity) that favors an action in an objective sense (e.g., Scanlon 1998, pp. 17–18; Smith 1994, pp. 95–101). To say that it favors an action means that it determines (possibly in combination with other reasons), pro tanto or flat out, what someone ought to do (but not necessarily what it would be rational to do). An agent ought to perform an action if and only if it is favored by the balance of normative reasons. By contrast, if an action is favored by a specific pro tanto reason, sometimes one ought not to perform it because contrary reasons outweigh that reason. Throughout the text, when I speak of reasons I refer to normative reasons. I am not concerned with motivating reasons, although, in a sense, normative reasons motivate when one acts for them.

When philosophers say, “This person acts for the normative reason R,” or, what is meant to express more or less the same idea, “This person acts for the good reason R,” what they refer to might be different phenomena, depending on which philosophical question they are trying to answer.2 Let me make it clear how I use this philosophical expression by distinguishing the notion I use from two different possible notions. By ‘acting for a normative reason’ I mean acting in accordance with a (real) normative reason such that the accordance is due to the agent (as contrasted with being due to, e.g., luck).3 This is a rather strong factive notion. One could think of two weaker notions.

First, philosophers might be interested in the question of how acting for a certain consideration bears on the agent’s character in a normatively significant way. Presumably, they would then use the expression ‘acting for a normative reason’ in a subjective, non-factive sense and mean that a person acts for a consideration that, from her point of view, favors the action.4 Arguably, this sense is the most important for that question.5 But what favors the action from the agent’s point of view need not really do so. The agent can be mistaken. She can act for a normative reason in this subjective sense even if there really is no normative reason that favors the action.6 That is why this subjective sense of the expression ‘to act for a normative reason’ is not factive.

Let us further assume that acting for a consideration that, from the agent’s perspective, favors the action can be thought of as exercising a disposition of character—roughly, a disposition to be, to an appropriate extent, motivated by such considerations. In the following, I will refer to such dispositions as ‘dispositions to appropriate motivation’.7 Figure 1 illustrates the subjective sense of the expression ‘to act for a normative reason’, where ‘R for A8 stands for some consideration that, from the agent’s perspective, favors the action A.
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig1_HTML.gif
Fig. 1

The subjective sense

Second, there is also an objective sense of the expression ‘acting for a normative reason’ in which it is a necessary condition of acting for the normative reason R that the agent acts in accordance with the (actual) normative reason R for performing the action. In this objective sense, the expression ‘acting for a normative reason’ is factive in at least two respects: the agent acts on a true consideration and the consideration actually favors the action.9

This objective notion of acting for a normative reason can be made more precise by determining how the normative reason and the agent’s belief that represents it are related. The objective notion may be understood as a conjunction: whenever there is a consideration that, from the agent’s perspective, favors the action, and the agent is properly motivated by the belief that represents this consideration, and there really is this normative reason, then the agent acts for this reason in the objective sense (even if the normative reason and the agent’s belief only happen to coincide). This is the second sense in which the expression ‘acting for a normative reason’ might be used. I call it the ‘conjunctive version of the objective sense’ (see Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig2_HTML.gif
Fig. 2

The conjunctive version of the objective sense

Now I turn to the third sense of the expression ‘acting for a normative reason’, which is the sense I will use. It is also objective, but compared to the conjunctive version of the objective sense it posits a stronger connection between the reasons-belief and the reason. The fact that there is a normative reason from the agent’s point of view needs to hold because there really is a normative reason that favors the action and the agent exercised a disposition to represent a reason when there really is one.10 I will call such dispositions ‘dispositions to appropriate belief’. This is the explanatory version of the objective sense, which is illustrated by Fig. 3.
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig3_HTML.gif
Fig. 3

The explanatory version of the objective sense

The two versions of the objective sense of the expression ‘acting for a normative reason’ provide answers for different philosophical questions. The conjunctive version is important to the following question: “When is it true that, first, the consideration for which the agent acts reflects well on the agent’s character and, second, the agent performs an action that is in accordance with a normative reason?”. This is really a conjunction of two questions, and one may wonder whether it is especially illuminating to ask such a conjunction of questions rather than asking the questions it involves separately. I am agnostic as to whether the conjunctive notion is an important philosophical notion.11 For this paper all that matters is to see how it differs from the notion I am going to be concerned with, namely, the explanatory notion.

The explanatory notion can be understood as helpful for answering the more demanding question: “When is it due to the agent (and her competence) that she acts in accordance with a normative reason?”. The idea behind this question is that, in certain circumstances, performing an action which is favored by a reason is something that the agent achieves by exercising certain competences (where manifesting the disposition to have true beliefs about reasons and the disposition to act accordingly constitutes exercising the corresponding competences).12 In such cases there is a connection between the action’s being favored and the agent’s performing the action that makes it possible to view the fact that the action is in accordance with a normative reason as the agent’s achievement. The other two notions of acting for a normative reason do not answer the question when it is the agent’s achievement that he performs the action which is in accordance with the normative reason.

Note how important this question is. Suppose there are normative reasons that favor actions. Maybe the actions we perform are favored by those reasons, but maybe not. Certainly one should hope that we could achieve performance of those actions which are favored and not those which are not. So one should hope that we were competent in performing those actions which are favored by reasons, and that we exercised this competence successfully. The explanatory notion is an analysis of the expression ‘acting for a normative reason’ as referring to the successful joint exercise of two competences: the competence to have true beliefs about reasons and the competence to be motivated by those beliefs accordingly. These competences together form the competence to act in accordance with reasons, and acting for a normative reason is the successful exercise of that competence.13

These examples of how the philosophical notion ‘acting for a normative reason’ might be used are supposed to show that, in principle, different senses of that notion might apply more or less well to different philosophical questions. Maybe all these senses of the philosophical notion are useful for different philosophical purposes and are therefore legitimate.

In the following, I am going to focus on the explanatory version of the objective notion (short: the explanatory notion) of acting for a normative reason. This notion is special in at least one respect, namely, in that it is the strongest one of the three notions and thus construes acting for a normative reason in the most demanding and encompassing way. My hope is that this notion will lead to seeing a very abstract ideal relationship between actions and normative reasons, one that encompasses and integrates the different, more fine-grained, concerns of narrower notions of acting for normative reasons. As I have indicated above, to understand this abstract ideal relationship between actions and normative reasons is necessary in order to understand how normative reasons can have the appropriate impact on our actions.

I will argue that the explanatory notion is best understood as normatively apt action, and that it is strikingly similar to Ernest Sosa’s conception of knowledge, but first I will critically discuss Ernest Sosa’s views about apt action and knowledge.

3 Sosa on apt action and knowledge

Ernest Sosa develops a theory of knowledge by comparing knowledge to apt action. Apt action is action that is accurate because adroit, in other words, it succeeds (is “accurate”) because of an exercise of competence by the agent (“because adroit”). The action’s success is to be attributed to the agent because it is produced by the agent’s competence.14

A behavior succeeds if it attains its aim, where, in principle, that aim need not be intentional (as when a heartbeat aims at pumping blood, Sosa 2007, p. 23). But Sosa’s standard example is not that of mere behavior but that of intentional action. An intentional action succeeds if it attains what the agent intends. I will call this kind of aptness ‘intention aptness’ (see Fig. 4). An example is the shot of an archer. “A shot succeeds if it is aimed intentionally to hit a target and does so” (Sosa 2007, p. 23). The shot is apt when the archer hits the target because the shot manifests skill in shooting, i.e., the competence to hit targets he intends to hit (see Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig4_HTML.gif
Fig. 4

Intention apt action

https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig5_HTML.gif
Fig. 5

The example of the archer

The theory of apt action is supposed to distinguish cases in which success is due to an exercise of competence from cases of success without competence, where the success is produced by mere luck. But it is supposed to do even more: the important idea is that it can distinguish two cases of success with competence, namely one in which success is explained by an exercise of competence from one in which success is merely accompanied by an exercise of competence but not explained by it—as when the competence is exercised but fails and a lucky coincidence produces success anyway. An example is a competent archer who aims at a target and shoots, thereby exercising competence, but finds himself in circumstances that render his competence useless.

Take a shot that in normal conditions would have hit the bull’s-eye. The wind may be abnormally strong, and just strong enough to divert the arrow so that, in conditions thereafter normal, it would miss the target altogether. However, shifting winds may next guide it gently to the bull’s-eye after all. The shot is then accurate and adroit, but not accurate because adroit [...]. So it is not apt, and not creditable to the archer (Sosa 2007, p. 22).

It seems to be important for this example to view the wind as having such a big impact that the archer’s competence does not stand in an explanatory relation to the success.15 The winds blow unforeseeably and under these conditions it would have been just as likely that the arrow of an incompetent archer had been blown to the bull’s-eye (for example, if the incompetent archer had shot a second later such that his shot had only been hit by the second gust of wind, and that gust of wind would have corrected the shot’s direction), although the incompetent archer’s shot would under normal conditions probably have missed the target.

Thus there are two substantial ideas here. One is that competences are explanatory (in the relevant sense) for success only when they have a tendency to produce success that is reliable enough.16 The second is that they have that tendency only under certain conditions. Call the conditions under which a competence reliably tends to produce success the ‘competence conditions’ of that competence. When there are no competence conditions the competence may still be there and may be exercised—but since it is deprived of its tendency to produce success by these other conditions it no longer stands in the relevant explanatory relation to the action’s success.17

It seems that the distinction between success because of competence and success merely accompanied by competence is made by the distinction between a successful exercise of competence under competence conditions and a successful exercise of competence under conditions other than competence conditions. In the second case the success is attributed to the circumstances and not to the agent.18 The reason is that success is attributed to the exercised competence, and accordingly to the agent, only in conditions in which the competence reliably produces success.

Knowledge is similar to apt action in that it is what Sosa calls “truth apt belief” (see Fig. 6): A belief is knowledge if and only if it is “accurate because adroit”.19 In a case of knowledge, a believer succeeds (believes the truth) because he exercises epistemic competence.20 The belief’s truth is to be attributed to the believer because it is explained by the believer’s exercising (epistemic) competence.21
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig6_HTML.gif
Fig. 6

Truth apt belief

This analysis of knowledge is supposed to accommodate the view that for a belief to count as knowledge it is necessary that it be a justified true belief.22 The belief is successful in that it is true, and it is justified in that it is the result of the exercise of a competence to attain true belief. Most importantly, this analysis of knowledge is supposed to deal with the challenge posed by Gettier cases (Gettier 1963). This is the challenge to show what else is necessary for a belief to be knowledge, apart from its being a justified true belief, such that these necessary conditions together are sufficient. Gettier cases are construed to show that for a belief to be knowledge it is not sufficient for the belief to be a justified true belief. Gettier cases thereby press philosophers to give a new definition of knowledge, namely one that states sufficient conditions for knowledge such that Gettier cases are ruled out.23

The analysis of knowledge as truth apt belief requires more than justified true belief; it requires the belief to be true because of the exercise of competence, as opposed to requiring the belief to be true and to be formed by an exercise of competence.24 As in the practical case, there is a distinction between success because of competence and success merely accompanied by competence. We can map this idea onto the slogan of justified true belief because the justification of a belief is its being formed by an exercise of competence. This means that truth apt belief is more than justified true belief, namely belief that is not only true and justified, but that is true because justified. Sosa applies this idea to a typical Gettier case as follows.25

Smith has strong evidence that Jones owns a Ford (e.g., he can remember many situations in which he saw him driving a Ford). He is justified in believing that Jones owns a Ford, although the belief happens to be false (the Ford that Jones drives is a rented car). Next, he infers that either Jones owns a Ford or Brown is in Barcelona. Because this inference is in accordance with the rules of logic, the justification is transmitted. The belief that either Jones owns a Ford or Brown is in Barcelona happens to be true because, although Smith has no evidence concerning Brown’s whereabouts (and although Smith might not even believe that Brown is in Barcelona), he happens to be in Barcelona. Therefore this belief is a justified true belief, but intuitively it is not a piece of knowledge.

Sosa deals with this case by stressing that one must not conflate the fact that the exercise of a competence (partly) explains the believer’s forming the belief (that happens to be true), with the fact that the exercise of a competence explains the success, i.e., the truth of the belief.

The reasoning [\(\ldots \)] does of course help explain why the believer has the belief, but it does not in the slightest help explain its correctness. In order to do so, it would have to be a factor that, either singly or in combination with other factors, accounts for how the belief is true rather than false. This means that it must help establish a connection between how the believer believes on that matter, and the truth of the matter (Sosa 2007, p. 96).

The problem with this idea is that it is still not clear what it means to say that a competence explains success—as opposed to merely explaining the formation of the belief. If we do not know what this means, it is unclear how Sosa’s account of knowledge could be satisfactory. Gettier cases are important because our case specific intuitions about these cases are that the person in question lacks knowledge. An account of knowledge is supposed to tell us which feature it is about these cases that makes it true that the person in question lacks knowledge. That feature is supposed to be described in terms that are clear to us. We must be able to make judgments about whether that feature is present in these cases or not—judgments that can either fit with or not fit with our intuitions about whether the person lacks knowledge, and thereby either vindicate or disprove that account of knowledge. But if the feature is that a competence explains success, we do not seem to have any clear idea of whether in a Gettier case a competence explains success (i.e., truth), as opposed to merely explaining the formation of the belief. If that is the case, it might happen that we simply say that a competence explains success in all cases in which we have the intuition that the person has knowledge. That would render the account of knowledge insignificant. An account of knowledge that explains the lack of knowledge by a certain feature is insignificant, at least with respect to certain cases, if in these cases it simply tells us to assume that this feature is present whenever we intuitively judge that the person in question lacks knowledge.

Of course Sosa does not intend the feature that a competence merely explains the formation of a belief (and not its success) to work in that way—but the danger is that if we do not have any sufficiently clear independent grasp of the notion of success explained by competence, it is not clear that it does not. The analysis of knowledge as belief that is true because of competence should therefore be supplemented by a statement of necessary and sufficient conditions for a belief to be true because of competence. Only then can we be sure that it really is significant. Here is my proposal. It relies on the importance of competence conditions and on the role of sub-competences.26

In discussing the archer in abnormal wind, I proposed to think of success because of competence as the successful exercise of a competence in competence conditions. In the Gettier example the solution is not that simple, as there seems to be a successful exercise of a competence in competence conditions and yet no success because of competence: epistemic competence is exercised, it produces true belief, and the conditions might be normal, in a sense. The problem is not that the conditions are abnormal, but that a false belief is involved.27 Here is where we need to focus on sub-competences.

Smith forms his first belief by using an epistemic sub-competence. Even though it may be exercised under competence conditions, this sub-competence produces a false belief. But given this kind of mistaken input, the overall-competence is not reliable enough anymore. Sosa argues for the need of a connection of “how the believer believes on the matter” and “how the belief is true” (Sosa 2007, p. 96). This connection is, perhaps, met if we require all outputs of sub-competences to be not only produced under sub-competence conditions but also to be of such a kind that the overall-competence works reliably enough when operating with input of that kind. Usually, that means that the output of the sub-competences must be successful and apt, but there are cases where the overall-competence works reliable enough with false belief which happens to be close enough to the truth, and in these cases less than success is required. Instances of so-called knowledge from falsehood (Warfield 2005) can be seen as examples of such cases.28 When the exercise of sub-competences is part of the exercise of an overall-competence, the exercise of the overall-competence cannot explain success if the exercise of the sub-competences do not explain sub-success (or something close enough to sub-success). Aptness of an output of an overall-competence is a matter of aptness (or something close enough to aptness) of the outputs of the sub-competences. In the following I will omit the reminder that the required kind of output of the sub-competence may sometimes be only close enough to apt output.

Demanding aptness of the output of a sub-competence might seem to lead into a regress. However, it seems plausible that any exercise of a competence consists of a finite number of instances of exercising sub-competences such that there are finally competences that do not involve other competences. With respect to these it is unproblematic to define aptness as successful exercise in competence conditions. It is only the output of complex competences that can fail to be apt even though it is successful and formed under competence conditions. Therefore, the aptness of the output of an overall-competence involves the aptness of the output of the sub-competences, as demonstrated by Fig. 7. The aptness of output of the sub-competences is defined as success under competence conditions.
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig7_HTML.gif
Fig. 7

Aptness of an output of an overall-competence, consisting of sub-competences

This interpretation of success because of competence seems to explain why Smith does not know that Jones owns a Ford or Brown is in Barcelona. The belief formed by exercising the sub-competence to infer from evidence whether Jones owns a Ford was not apt (and not even close enough to being apt), therefore the belief formed by the overall-competence cannot be apt. The proposal then yields the desired result.

We arrive thus at the following necessary and sufficient conditions of aptness, where aptness is success because of, or explained by, competence.

Necessary and sufficient conditions of aptness of an output:

An output is apt if and only if the success of the output is explained by an overall-competence, and the success of an output is explained by an overall-competence if and only if, for every exercise of a sub-competence involved in the exercise of the overall-competence, the exercise of the sub-competence takes place under sub-competence conditions and yields a successful output (or an output close enough to success such that the overall-competence works reliably enough when operating with input of that kind).

There are alternatives to this characterization of aptness. One alternative is Greco’s explanatory salience contextualism (which is different from his contextualism about competences, see fn. 25). According to Greco’s explanatory salience contextualism, an output is apt if and only if it is successful and the exercise of competence is the salient part of the entire set of causal conditions that would figure in a complete causal explanation (Greco 2010, p. 74, 2003, p. 131). It is not entirely clear how salience is determined, but at least there are certain criteria, for example, that the most unusual part of a causal story often is the salient part. In some cases Greco’s proposal yields results similar to mine. For example, in Gettier cases in which there are unusual conditions the competence will not be the salient feature. Instead, the unusual conditions are the salient part of the explanation. However, Greco admits that explanatory salience is not fully understood (cf. Greco 2010, p. 75). And one might worry that aptness should not be a matter of whether the conditions are usual. On my view, aptness is a matter of the reliability of the competence and its sub-competences in the circumstances. The conditions are competence conditions if the competence is reliable enough in these conditions. Thus, the conditions may be competence conditions even if the conditions are unusual (in some sense) as long as they do not deprive the competence of its tendency to produce success. What is important is whether the conditions are such that the competence is reliable enough in producing success—and this may or may not coincide with the conditions being usual. Thus it seems to me that my account is superior to Greco’s in two respects: First, it does not rely on the somewhat unclear notion of explanatory salience, and second, it allows that aptness is a matter of reliability instead of a matter of what is usual.

A second alternative to my account of aptness is Pritchard’s Anti-Luck Virtue Epistemology (Pritchard 2010, pp. 54–65; Pritchard forthcoming). Pritchard argues that Virtue Epistemology cannot deal with certain Gettier cases (such as the barn façade case) unless it is supplemented with a separate safety requirement. However, there are resources for handling the barn façade case without a separate safety requirement (see fn. 25). This is why I prefer Sosa’s account with the specification of aptness that I have presented. Pritchard offers further and rather general considerations for why a separate safety requirement is needed (having to do with Craig’s notion of a reliable informant). Basically, he thinks that what he calls the ‘ability intuition’ and the ‘anti-luck intuition’ are two independent intuitions that pull into two different directions. A full-scale discussion of these general considerations, however, goes beyond the scope of this paper. Suffice it to say that the right conception of competences may make it questionable whether there really is such a tension between two different intuitions and corresponding requirements as Pritchard claims.

If my argument in this section is correct, Sosa’s analysis of knowledge can be defended by showing precisely how he can deal with Gettier cases involving false belief. His analysis of aptness as success explained by competence can be explicated independently of our case specific intuitions about knowledge.

4 Normatively apt action and truth apt belief

The comparison of knowledge with apt action is illuminating, but the specific kind of aptness of action that Sosa focuses on, intention aptness, is less parallel to the case of knowledge than another, more important kind of aptness, or so I will argue. At any rate, it is another kind of aptness that can help illuminate the notion of acting for a normative reason.

This other kind of aptness is normative aptness. A normatively apt action is apt with respect to one or more normative reasons there are to perform the action. We can distinguish a weaker and a stronger kind of normative aptness. An action is normatively apt in the weak sense if and only if the action is in accordance with some normative reason (success/accurateness) because of the agent’s exercise of the competence to act in accordance with a normative reason (competence/adroitness). An action is normatively apt in the strong sense if and only if the action is in accordance with the balance of normative reasons because of the agent’s exercise of the competence to act in accordance with the balance of normative reasons. I will concentrate on the weak sense because it is the more basic sense and also because it obviously stands in a closer relation to the notion that interests me, namely, the notion of acting for a normative reason (although, hopefully, the notion of acting for the balance of normative reasons can be derived from the notion of acting for a normative reason). The weaker notion is illustrated by Fig. 8.
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig8_HTML.gif
Fig. 8

Normatively apt action

Normatively apt action is not the same as intention apt action (compare Fig. 4). An action need not be normatively apt (not even in the weak sense) when it is intention apt. It is intention apt when it attains the agent’s aim (i.e., what he intends to do) because of the agent’s exercise of skill. But even if the agent attains his aim in this way, the action is not normatively apt if there is no normative reason that favors the action. Consider shooting a child on a playground:29 The shot may be intention apt if the agent aims at shooting the child and he succeeds in shooting the child by exercising a competence to hit his target, but it is not a normatively apt action. It cannot be normatively apt in the strong sense because it certainly is not normatively accurate in that sense: it is not in accordance with the balance of reasons. And, assuming there are no pro tanto reasons for shooting the child, the shot is also not normatively accurate in the weak sense: there may be no normative reason at all for shooting the child. It is not a success in the weak sense that something is achieved that is favored by at least one normative reason and thus at least has something to be said for it.30 Since aptness involves accurateness, but the action is not normatively accurate in either of these senses, it cannot be normatively apt.

I will now point out similarities between normatively apt action (in the weak sense) and truth apt belief. For a schematic comparison of these kinds of aptness, see Table 1.
Table 1

Similarities between normatively apt action and truth apt belief

 

Standard of success

Constituent of standard

Competence

Success by competence

Theoretical domain

Truth accurateness: the belief is true

fact F, where F is the truthmaker of the belief

Truth competence: competence to form true beliefs

Truth apt belief, i.e., knowledge

Practical domain

Normative accurateness: the action is favored by a normative reason\(^\mathrm{a}\)

Fact F\(^{\prime }\), where F\(^{\prime }\) favors the action, i.e., F\(^{\prime }\) is the ‘favorer’

Normative competence: competence to perform actions that are favored by normative reasons\(^\mathrm{b}\)

Normatively apt action, i.e., acting for a normative reason

\(^\mathrm{a}\) This table concerns the weak sense of normatively apt action. On the strong sense, by contrast, the success of action is being in accordance with the balance of normative reasons, i.e., being right. Arguably, the strong sense of normatively apt action is at least as much similar to knowledge as the weak sense, but I concentrate on the weak sense because my topic is acting for a normative reason, not acting for the balance of reasons.

\(^\mathrm{b}\) More precisely, the normative competence is the competence to represent whatever normative reasons there are for the agent to perform certain actions and to respond to such representations of reasons by being motivated to perform the actions which the represented reasons favor. Figure 10 is an illustration of this structure of the normative competence

Normatively apt action is more similar to truth apt belief than intention apt action because it is independent of the agent’s intention.31 Therefore, normative accurateness is a better comparison to epistemic accurateness than intention accurateness. With respect to action, maybe some people will feel uneasy about applying a standard of success that is intention-independent. If the agent does not intend to act in accordance with a normative reason, then it is, perhaps, challenging to explain how the normative standard can be applied to him. It might seem odd to say of an action that it fails with respect to a standard that the subject does not intend to meet. This worry can equally be formulated for the epistemic realm, and it can be answered in both these domains alike. Anyone who accepts the standard of truth for belief accepts that there are intention-independent standards, so there could be such a standard for actions as well. Success in believing is determined by truth whether or not the believer intends to believe what is true. He may intend to believe whatever will be most comforting, and he may do so very competently and thereby succeed in attaining what he intends. Such success by competence (i.e., intention aptness) will not render the belief a piece of knowledge. Intention aptness is, as such, irrelevant for knowledge. The standard of knowledge is fixed independently of what the believer intends to believe. It is truth, even if the believer is not interested in truth, and even if the believer intends instead to believe what is most comforting.

With respect to actions there seem to be two distinct kinds of success, two standards by which an action can be assessed. One is whether the agent attains what he intends, and the other is whether the agent attains what is favored by a normative reason (or even the balance of reasons). The second seems to be the better comparison to the epistemic standard of truth in that it is also independent of the subject’s intention. This is the criterion for practical success that is needed in order to understand acting for a normative reason.

A second worry might be that the idea of normatively apt action presupposes a wrong picture of action or practical reason, namely that it aims at the good, or, more precisely, at doing what is favored by normative reasons.32 Even if the subject need not intend to do what is favored by normative reasons, there needs to be some sub-personal aim in order to apply this standard. The action, the motivational states involved, or the rational capacities involved must aim at meeting the standard, or so the objection goes. This being said, the view that we can assess an action in terms of whether it is normatively accurate and normatively apt does not presuppose that practical reason has the constitutive aim to deliver action that is favored by normative reasons, or that practical reason aims at the good. Sosa’s account of knowledge as truth apt belief does not rely on the idea that belief constitutively aims at truth. Similarly, the corresponding account of normatively apt action does not rely on a constitutive aim in the practical realm.

At most, one could make out a constitutive aim of apt action and, accordingly, of apt belief. One might argue that competences, whether intentional or not, can always be described as having a corresponding aim.33 Such an aim is not to be conflated with an aim of the agent in the sense of an intention of the agent. Rather, it is an aim of a competence. The competence to form true beliefs aims at truth, or true beliefs, and the competence to perform actions that are favored by reasons aims at actions that are favored by reasons, or the good. If this is so, truth-competent belief aims at truth constitutively, and normatively competent action aims at the good constitutively. That implies that truth apt belief and normatively apt action aim at truth and the good respectively, because aptness involves competence. That is just analytically so, however, and it should be distinguished from the more demanding and more contested idea that intentional action as such, whether or not normatively competent, aims at the good. The aim that is implied by the normative competence need not be present where that competence is lacking, and where normative competence is lacking there may still be action.

A third worry about the similarity between normatively apt action and knowledge concerns not independence of intention but independence of who the agent is. The standards of action and belief are both independent of the intention of the subject, but the standard of action is often dependent on a subject in a way that the standard of belief need not be. Normative reasons often favor one particular person’s action depending on who that person is and in which situation he finds himself. Truthmakers often make one particular person’s belief true independently of who that person is and in which situation he finds himself. This is certainly a difference in the standards of action and belief, but it does not undermine the view that normatively apt action is the right comparison to knowledge. I do not claim that action and belief have exactly the same standard or that their standards may not differ from each other in important respects. Surely there will be important differences between the standards in the theoretical and the practical domain. Therefore it poses no problem to my view that the standard of action is often dependent on who the agent is while the standard of belief is often not dependent on who the believer is. My claim is that, while the standards may differ, the structure of knowledge (i.e., belief that meets its standard by exercise of competence) is similar to the structure of acting for a normative reason (i.e., action that meets its standard by exercise of competence).

Furthermore, I want to compare knowledge with the most analogous practical phenomenon I can find. There are two practical phenomena of apt action, namely, intention apt action and normatively apt action. When it comes to deciding which of these is the better comparison to knowledge it makes sense to prefer normatively apt action because its standard is more similar to truth than that of intention apt action, as far as independence of intention is concerned. With respect to dependence on who the agent is, the standards of normative aptness and intention aptness are both unlike the standard of truth aptness. It obviously depends on who the agent is whether the action is in accordance with the agent’s intention. Therefore, normative aptness is the better practical comparison with knowledge than intention aptness despite the differences between the standards involved.

However, there is another disanalogy between the standard of normative aptness and the standard of knowledge. Normative reasons are often weighed against each other whereas truth is not. This would not be troublesome (as I just said, the standards of the practical and the theoretical domain need not be similar in every respect) if there was not another theoretical standard, namely, being in accordance with evidence or reasons for belief, that allows weighing. So it might look as if the best theoretical comparison to normatively apt action was not knowledge but rather that of belief held for a (theoretical) reason or evidence. However, this impression is mistaken. Although theoretical evidence can be weighed like normative reasons, there are important differences between normative reasons and evidence or epistemic reasons. Evidence is an indicator of truth, something that makes truth probable without necessarily being partly or totally constitutive of truth. The balance of evidence can point towards the truth of a belief although the belief is false. This is not so in the case of normative reasons. Normative reasons are easily conflated with evidence that the act is right because they can be weighed like evidence, but the difference is that they do not only indicate rightness; they are partly constitutive of rightness. The balance of normative reasons constitutes that the action is right and it cannot do so if the action is in fact wrong. By contrast, the balance of evidence can indicate the rightness of an action although the action is wrong.34 Therefore, if truth of belief and rightness of action are good comparisons, evidence and normative reasons are not because normative reasons are constituents of rightness, not only indicators of rightness. Normative reasons (at least partly) constitute rightness and truthmakers constitute truth, even though normative reasons can often be weighed against each other, whereas truthmakers cannot.

Let me briefly sketch one last consideration in favor of comparing truth apt belief to normatively apt action. The analysis of knowledge as truth apt belief is known as a version of ‘Virtue Epistemology’.35 Virtues are often viewed as dispositions to respond appropriately to normative reasons.36 For example, virtuous actions are not seen merely as actions that attain what the agent intends, but rather as actions that aim at the right, and, if they are successful, attain the right, where what is right is what is favored by (the balance of) normative reasons. So there is an important connection between virtuous actions and actions that are favored by reasons.37 Therefore, the very name of Sosa’s epistemic theory creates a strong pull towards connecting it with normative reasons. In as much as we think of epistemic competence as virtuous, we tend to compare it, in the practical realm, to genuinely normative competences, and not to the competence to attain what is intended. Accordingly, even Sosa occasionally happens to draw the comparison between correct belief and right action rather than intention accurate action.38

One might object that the connection between knowledge and virtue is not as tight as Virtue Ethicists would have it and that this generates a disanalogy with normatively apt action. There are cases of trivial knowledge, e.g., knowledge of how many blades of grass grow in a particular garden. To acquire this kind of knowledge is a waste of time and thus does not seem to indicate epistemic virtue but rather epistemic vice. Normatively apt action, by contrast, is always a sign of virtue. My response to this objection is to deny the premise that to acquire trivial knowledge does not involve epistemic virtue. It always involves the capacity to form a belief in accordance with the facts. At most one could say that, although this is an epistemic virtue, there are other intellectual virtues which might pull the subject away from acquiring trivial knowledge, such as the virtue to save one’s cognitive resources for important topics (cf., e.g., Sosa 2007, p. 89). However, the epistemic virtue to form a belief in accordance with the facts is always involved in knowledge. Therefore it seems right that a final similarity between normatively apt action and knowledge is that both involve virtue (of the specified kind).

5 Normatively apt action and acting for a normative reason

If the arguments in the previous section are right, normatively apt action parallels knowledge (understood as truth apt belief) as demonstrated in Table 1. Both normatively apt action and knowledge are defined by being successful because of competence. The standard of success, which is (the balance of) normative reasons in the practical case and truth in the epistemic case, is independent of what the subject intends.

Now let us try to connect the idea of normative aptness to the explanatory notion of acting for a normative reason. I will try to show that the explanatory notion can be interpreted as the notion of normatively apt action, but in order to do so I have to show that aptness, or success because of competence, fits the characterization that I have given of this notion in section two.

The explanatory notion of acting for a normative reason was spelled out as an explanation of why an action is performed, not of why an action is in accordance with a normative reason. In Sosa’s terms, it is not an explanation of success, but of the action taking place. This poses a problem. In order to show that the explanatory notion refers to normatively apt action we need to show that it is of the right kind to explain success. Put differently, we need to show that it meets the necessary and sufficient conditions of aptness. In order to do that we need a bridge-principle that tells us how to move from aptness, which is an explanation of the success of an output, to an explanation of an output being produced, and vice versa. I will try to develop such a bridge principle by focusing on the case of knowledge, once more, and then apply that principle to the case of acting for a normative reason.

In comparing knowledge with justified true belief Sosa argues that knowledge is distinctive in that success (here: truth) is explained by competence. It is not sufficient for knowledge that the formation of belief is explained by competence, even if that belief is true (Sosa 2007, p. 96). I agree, but this does not exclude the possibility of finding some other explanation of the formation of the belief that corresponds to an explanation of the belief’s truth in terms of competence. After all, even in cases of truth apt belief there must be explanations of the formation of the belief. It is not as if the formation of the belief was mysterious. These explanations, that surely must exist, might be of a special kind—a kind that applies if and only if it is correct to explain the belief’s truth by the competence.

My proposal for this kind of explanation of the formation of a belief is inspired by my explication of aptness (see Sect. 3) and turns it into an explanation of the formation of the belief. It reads as follows: In a case of knowledge, a subject forms the belief that F because a) F is the case and b) the subject exercises a competence to find out whether F is the case, and c) it does so under conditions in which the competence is reliable (competence conditions), and if sub-competences are involved, their outputs are successes (or close enough to successes) and they are also formed because of a), b) and c) with respect to the sub-competence. For simplicity, let me give an example with a competence that does not involve sub-competences.

Suppose a person forms the belief that there is a red flower in front of her because she exercises her visual perception, the light is normal, and there is a red flower in front of her. The proposal is that we can say that she has a true belief because of competence. A general bridge-principle (one that is not only concerned with knowledge but with aptness generally) could then be stated as follows.

Bridge-principle:

A success of an output is explained-1 by a competence if and only if 1) the formation of the output is explained-2 by the following three facts: a) by the fact that constitutes that the output meets the standard of success, b) by the fact that the subject exercised the competence, and c) by the fact that this was done in competence conditions, and 2) all outputs of the exercise of sub-competences that are involved in the exercise of this competence are successes (or close enough to success) and their success is explained-1 by the sub-competences.39

There are two different kinds of explanatory relations involved. The second explanatory relation (explain-2) concerns the formation of the belief or the performance of the action. It is a causal explanation insofar as it involves a causal element: the exercise of the competence causes the formation of the belief or the performance of the action. But the first explanatory relation that concerns the success (explain-1) is certainly not causal because the exercise of competence does not cause the belief to be true or the action to be in accordance with a reason. What is caused is the formation of the belief, not its truth. When it comes to an explanation of the truth of the belief, or of the success of an output generally, a specific type of non-causal explanation by appeal to competence is at play. The competence is exercised in conditions under which it produces successful output reliably enough. Without exercise of the competence, success would have been far less likely. Therefore the success of the output is attributed to the competence. This is a non-causal explanation because what is caused, strictly speaking, is the output, not its success. To explain success by a competence (if it is exercised in competence conditions) is to indicate that the competence, in competence conditions, usually delivers successful output. In that sense, the fact that a certain type of action is favored by a normative reason, together with the fact that the agent is in certain conditions and exercises a competence to act in accordance with normative reasons in those conditions, can explain that the agent acts in accordance with that reason.

If we apply the bridge-principle to the practical domain, it is just what we need in order to move from the explanatory notion of acting for a normative reason to Sosa’s description of normatively apt action. We only need to supplement the original statement of the explanatory notion with a reference to competence conditions. Then, the bridge principle allows us to move from the explanation of the action that characterizes the explanatory notion of acting for a reason (Fig. 9) to the explanation of success that characterizes normative aptness (Fig. 10). Figure 10 shows that acting for a normative reason, on the normative notion, is of exactly the form of aptness of an output of an overall-competence (compare Fig. 7). This shows how the explanatory notion of acting for a normative reason can be interpreted as referring to normatively apt action.
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig9_HTML.gif
Fig. 9

The explanatory version of the objective sense (with competence conditions)

https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig10_HTML.gif
Fig. 10

Normatively apt action (with sub-competences)

Two additional interesting points remain which should be mentioned. The first is that the outdated, pre-Gettier analysis of knowledge as justified true belief is strikingly similar to the conjunctive notion of acting for a normative reason. To see this, compare Fig. 11 with Fig. 2. The analysis of knowledge as justified true belief is similar to the conjunctive notion of acting for a normative reason in that the formation of the belief (or the action) is explained by the exercise of competence in such a way that the bridge principle does not allow explaining the success by competence. The bridge principle states that, if the success is explained by competence, what constitutes the standard of success (truth/normative reason) needs to play a role in explaining the formation of the belief (or the action), which it does not.
https://static-content.springer.com/image/art%3A10.1007%2Fs11229-012-0230-8/MediaObjects/11229_2012_230_Fig11_HTML.gif
Fig. 11

Knowledge as justified true belief

The fact that the conjunctive notion of acting for a normative reason, according to the bridge principle, does not correspond to the analysis of knowledge as apt belief but rather to the analysis of knowledge as justified true belief suggests that the analogy between the phenomenon of acting for a reason to some analysis of knowledge seems to have application at a more general level. Maybe, if we took into account even more notions of acting for a reason and even more analyses of knowledge, we would find even more analogies between certain items among these. With respect to the two analyses of knowledge that I have discussed, the analysis of knowledge as justified true belief is inferior to the analysis of knowledge as truth apt belief in that it fails in the Gettier cases. This might, by analogy, point to the inferiority of the conjunctive notion of acting for a reason as compared to the explanatory notion of acting for a reason.

The second point is about the internal structure of the explanatory notion of acting for a reason. What is interesting here is that acting for a reason involves truth apt belief.40 The competence to act in accordance with a normative reason is composed of two competences: the competence to form a true belief about a normative reason (i.e., a true descriptive or normative representation of the reason) and the competence to be motivated appropriately by that belief. The first competence is an epistemic one. The belief it renders is a truth apt belief when it is true because it is formed by the competence. The belief must be truth apt for the action to be normatively apt. The reason for this is that the overall-competence that involves this sub-competence can only explain the success when its sub-competences explain their success. So only when the sub-competence to form a true belief about a reason explains the belief’s truth (when the belief is truth apt) can the action be normatively apt.

However, it is important that the similarity between normatively apt action and truth apt belief is not simply that one involves the other, but that both kinds of aptness, normative aptness and truth aptness, share common features—features that are characteristic of aptness as such. It is only a further observation that normatively apt action, being a complex form of aptness, involves truth aptness as a component. Similarly, truth aptness of a belief can be complex, for example when the belief is inferred from premises. In this case, the truth aptness of this belief involves the truth aptness of other beliefs as a component, much like normatively apt action involves truth aptness as a component. For example, imagine a person forming a belief by a similar inference as in the Gettier case mentioned above without being Gettiered.

6 Conclusion

If there are normative reasons, i.e., entities that favor certain actions, it is important that we can act in accordance with those reasons because of a competence to act in accordance with reasons. If it was never due to our competence that we act in accordance with reasons, there might still be reasons that favor our actions, but they would not be practical for us in the sense that our actions could not be guided by them. Our actions would at most coincide with what those reasons favor. This shows that for reasons to be appropriately practical we must assume that our actions can stand to reasons in a non-accidental relation which I refer to by the explanatory notion of ‘acting for a reason’. I have tried to illuminate this relation by comparing it to the non-accidental relation in which our beliefs stand to truth when they are knowledge. The comparison between acting for a normative reason and knowledge prepares the scene for looking for a similar answer to the question of the value of knowledge (as contrasted with merely true belief) and the question of the value of acting for a reason (as contrasted with acting merely in accordance with a reason). However, this is a task for another paper.

Footnotes
1

Williamson, for example, makes a very insightful comparison of knowledge and successful action, calling knowledge the “reverse counterpart” of action (Williamson 2000, p. 2). An important difference is that he is concerned with action that successfully satisfies desire, not with action that is favored by a normative reason. However, it is notable that Williamson’s account and mine nevertheless point in the same direction despite the fact that they rest on very different accounts of knowledge. Williamson does not even share the view that knowledge is justified true belief with an anti-Gettier component. The similarities between action and knowledge seem to be visible from very different perspectives.

 
2

Note that I am here concerned with how philosophers might want to use this technical phrase for different philosophical purposes. I am not directly concerned with how the colloquial expression ‘doing something for a (good) reason’ is used in everyday talk. That expression may have only one usage.

 
3

The distinction between acting for a reason and acting merely in accordance with a reason can be traced back to Kant’s distinction between acting “aus Pflicht” and “pflichtgemäß”, Kant (1785/6, B, pp. 397–400).

 
4

Distinguish two senses in which a consideration can be said to favor the action from the agent’s perspective, either of which might fit the subjective notion. The agent might represent something as a normative reason by a normative belief. Or the agent might represent something as a descriptive fact by, e.g., a descriptive belief, where the content is such that there would be a reason if the representation was true. The descriptive representation might as well be a non-conceptual representation, e.g., a perception.

 
5

The idea that acting for a reason in this subjective sense reflects well on the agent’s character is found, e.g., in Setiya (2007), Markovits (2010), and Hurka (2001), and, of course, Kant (1785/6), where it is put in terms of the “moral worth of an action”. It is controversial whether it reflects better on the agent’s character to act for an explicitly normative consideration or for a descriptive consideration that is, de facto, normative. Compare, e.g., Arpaly (2003) and Williams (1981).

 
6

The subject might have a false non-normative belief, a false normative belief, or both.

 
7

There are different dispositions that are dispositions to appropriate motivation. There needs to be an exercise of one such disposition, but it need not be the same in every instance of acting for a normative reason. For example, normative and descriptive representations will require different dispositions.

 
8

This symbol is supposed to stand for either a descriptive representation of an apparent fact or a representation of something as a reason, i.e., a normative representation.

 
9

Some authors propose a weaker sense in which acting for a reason is factive by saying that such action is performed for a true consideration or a fact, whether or not that fact favors the action (e.g., Unger 1975; Hyman 1999). My notion is stronger; it implies that there is a normative reason, not only a fact.

 
10

It is possible to think of different, more detailed, descriptions of dispositions to represent a reason when there really is one. For descriptive representations, usual ways of belief formation, e.g., via sense perception, can be filled in. For an account of how to form a normative representation of a reason, a special epistemology for normative entities needs to be filled in.

 
11

I explore this notion for systematic reasons and in order to contrast it with the explanatory notion. I will not rely on the appropriateness of the conjunctive notion in the following.

 
12

It is not easy to say how exactly competences and dispositions are related generally. I am inclined to think that every competence C to do X that I have under conditions C’ just is a certain disposition D (or a set of dispositions) to do X under conditions D’. What is important, if we are to identify competence C with disposition D, is that the conditions C’ under which someone has the competence sometimes are different ones than the conditions D’ under which he is disposed to do X. Therefore it is no counterexample to the view that competences are dispositions that I have the competence to cut my hair under conditions where I have a pair of scissors and control over my arms even if I am not disposed to cut my hair under these conditions. The competence to cut my hair, which I have under conditions where I have a pair of scissors and control over my arms, will plausibly be the disposition to cut my hair if I intend to, whereas I have the competence to cut my hair even if I do not intend to cut it.

 
13

The competence to act in accordance with normative reasons involves being influenced not just by any kind of fact but, more specifically, by normative reasons qua their normative character. This is why I prefer a more demanding account of acting for a normative reason than, e.g., Alvarez (2010), Hyman (1999), and Unger (1975). This is why my account implies that something actually favors the action.

 
14

A similar view is put forward in Greco (2010).

 
15

The exercise of the competence affects the arrow’s initial speed and direction, and these cause, together with the winds, the arrow to hit where it does. Consequently, in a sense, the exercise of the competence does enter into a certain kind of explanation of the result of the shot, namely a causal one. But it does not do so qua competence, and the explanation of a success in the form described in Fig. 5 (which is explanation by competence qua competence) is thus prevented.

 
16

The expression ‘reliable enough’ indicates that the ratio of success that counts as reliable is dependent on the competence in question and need not always be considerably high. For example, certain kinds of very difficult sports might require only that a person has success in twenty percent of her performances for that person to be very competent in that sport.

 
17

For the importance of competence conditions, or, as Sosa calls them, “appropriate conditions”, see, e.g., Sosa (2007, pp. 28–33, 81–84).

 
18

Note that the success may in such a case have to be attributed to the circumstances and not to the agent even though the state that is successful is caused, in part, by the agent’s exercising the competence (and, for the other part, by the freaky conditions). This will come out in Sosa’s discussion of the Gettier example that I cite below.

 
19

In other contexts, the expression ‘truth apt belief’ refers to belief that can be either true or false. Sosa’s notion of aptness is a technical notion he stipulates and differs significantly from the notion of aptness that is used in those contexts. Throughout this article, I use only Sosa’s notion of aptness. By ‘truth apt belief’ I mean belief that is true because of epistemic competence, not belief that merely can be either true or false.

 
20

For an analysis of epistemic competences as dispositions, see Sosa (2011, pp. 80–85, 2009, p. 135).

 
21

For a criticism of the idea that every instance of knowledge implies credit, see Lackey (2007).

 
22

According to some epistemologists, e.g., Dretske, knowledge does not require justification. See Dretske (1981, 1971). Instead of justification, some other factor is postulated, such as “conclusive reasons” or “information basedness”. It is unclear, however, whether one could not count such a factor as amounting to justification.

 
23

For the view that this is not possible, see Williamson (2000).

 
24

One might wonder how, literally, the truth of a belief can be explained by the competence. Should one not rather say that the formation of a belief that is true is explained by the competence? However, it is crucial for Sosa’s theory that the success and not the mere formation of the belief is explained by the competence (see the Sosa quote below). Maybe worries about this kind of explanation can be overcome if the explanation of a belief’s truth is viewed as corresponding to a specific kind of explanation of the formation of the belief. In Sect. 5 I spell out a bridge principle that states such a correspondence.

 
25

I cannot discuss in full how Sosa deals with the Gettier problem in general, because he handles cases that do not involve false belief differently. One such example is the barn façade example, see Sosa (2007, p. 96, fn. 1, 2004, pp. 292–293). Sosa’s strategy is to ascribe knowledge in the barn façade example. Greco responds to the barn façade example in a more satisfactory way by claiming that it is dependent on context which competence is called for and whether the belief is apt with respect to that competence (Greco 2010, pp. 76-80). When context requires the competence to distinguish between barns and fake barns, Greco does not ascribe knowledge. For a discussion of the barn façade case as a challenge for Virtue Epistemology and another proposal how to react to this challenge, see Pritchard (2010).

 
26

An appeal to certain sub-competences can be found in a passage in Sosa’s work, namely, in his discussion of perception, see Sosa (2011, pp. 74–78). However, Sosa does not generally assign sub-competences the importance that I think they have.

 
27

A false belief as such does not imply that the exercise of competence did not take place under competence conditions. Even under competence conditions competences are only reliable, not infallible. Even under competence conditions they can produce false belief.

 
28

Why not simply demand that all outputs of sub-competences be successful and apt? To demand success of sub-competence is, in inferential reasoning, to exclude inferences from false beliefs. But Warfield argues that there is knowledge from falsehood. These are exactly the cases in which the output of the sub-competence, despite being false, is close enough to the truth to qualify as good enough input for the overall-competence to work reliably enough.

 
29

I borrow this example from Christian Piller’s inspiring talk at the conference “Knowledge, Virtue, Action” in 2010.

 
30

The idea that there need not be any normative reason that favors the action when the action is intention apt is challenged by the view that the mere fact that the action succeeds in attaining what an agent intends (i.e. that the action is intention accurate) does itself favor the action. This view yields that there is always at least one reason that favors an intention apt action, namely, the reason that it is intention accurate. I do not endorse this view because to me it seems implausible that there be at least one thing that favors that the archer intentionally shoots a child on a playground, namely that it is what he intends to do. But even if this view could be defended, this would not threaten my main point: that intention aptness is different from normative aptness. Even if intention aptness implied normative aptness in the weak sense with respect to an intention-reason of the agent, it would not imply normative aptness (in the weak sense) with respect to certain other reasons, such as the normative reason that the child do well. Normative aptness can accommodate standards of success that can be independent of the agent’s intention, whereas intention aptness cannot.

 
31

It is an interesting question whether normative reasons are totally independent of what an agent intends, or to what extent they depend upon intentions. I do not mean to exclude the possibility that sometimes an agent’s intention gives rise to a normative reason. All I want to say is that the standard is not to attain what the agent intends, as such, but to act in accordance with a normative reason—whether this reason relies on or is constituted by an intention is a different question. In what follows, this will be the sense I have in mind.

 
32

The idea that intentional action constitutively aims at the good is rejected, e.g., in Stocker (1979), Velleman (1992) and Setiya (2007).

 
33

For example, the competence to pump blood aims at pumping blood Sosa (2007, p. 23).

 
34

For a different view, see Kiesewetter (2011).

 
35

Another version of Virtue Epistemology is developed in Zagzebski (1996) and Fairweather (2001).

 
36

This view is made especially clear in Setiya (2007), but something close to it (namely that virtues are dispositions to respond to values) can be found also, e.g., in Hurka (2001, 2011).

 
37

A virtue ethicist may say that the virtuous action is favored by normative reasons because it is virtuous and not that it is virtuous because it is favored by normative reasons. I am neutral on that. But a virtue ethicist will surely see a closer connection between virtues and normative reasons than between virtues and attaining whatever one intends, as in the example of the archer intending to shoot a child at a playground.

 
38

“A virtuous performance, whether a correct belief due to intellectual virtue or a right action due to practical virtue, will involve both the agent’s constitution and his situation” (Sosa 2007, p. 81).

 
39

Requirement (2) could also be expressed as follows: (1) holds also for all outputs of exercise of sub-competences that are involved in the exercose of the competence.

 
40

For the view that acting for a reason involves knowledge of that reason, compare Unger (1975, pp. 200–211), Hyman (1999), and Hornsby (2008, pp. 250–252). My account of acting for a normative reason differs from their accounts in that they primarily focus on acting intentionally for a known fact, and not so much on whether that fact really favors the action and whether the agent exercises the competence to respond to this normative dimension in particular.

 

Acknowledgments

I especially want to thank Christoph Fehige, Heinz-Dieter Heckmann, Tim Henning, Frank Hofmann, Rob McGee, Oliver Petersen, Christian Piller, Eva Schmidt, Thomas Schmidt, Peter Schulte, David Schweikard, Michael Smith, Ernest Sosa, Ulla Wessels, and two anonymous referees for comments and inspiration. I also want to thank the audiences of talks I gave at the Universität des Saarlandes in 2011, at the XXI. German Congress of Philosophy in Munich in 2011, and at a meeting at the island Reichenau, Germany, in the same year for very illuminating discussions of this paper.

Copyright information

© Springer Science+Business Media Dordrecht 2012