Moral blame and rational criticism

A central issue in practical philosophy concerns the relation between moral blameworthiness and normative reasons. As there has been little of direct exchange between the debate on reasons and the debate on blameworthiness, this topic has not received the attention it deserves. In this paper, I consider two notions about blameworthiness and reasons that are fundamental in respective field. The two notions might seem incontrovertible when considered individually, but I argue that they together entail claims that are highly contentious. In particular, I maintain that they entail unreasonable and contradictory claims since the practices of moral blame and rational criticism diverge with regard to three dimensions: justification, response, and function. Thus, we need to give up one of the principal notions. The solutions to this puzzle suggest that the connection between reasons and rationality is weaker than standardly presumed in metaethics. notions about blameworthiness and reasons.

incontrovertible when considered individually, but I argue that they together entail claims that are highly contentious. In particular, I maintain that they entail unreasonable and contradictory claims since the practices of moral blame and rational criticism diverge with regard to three dimensions: justification, response, and function. Thus, we need to give up one of the principal notions. The solutions to this puzzle suggest that the connection between reasons and rationality is weaker than standardly presumed in metaethics.

| INTRODUCTION
A key topic in practical philosophy concerns the relation between moral blameworthiness and normative reasons.
Bernard Williams famously observed that internalism about reasons appears to have difficulties to account for our practice of blaming, and his response to this worry has generated important contributions. 1 In spite of this, there has been little of direct exchange between the debate on reasons and the debate on blameworthiness, and the topic has not received due attention. In this paper, I consider two notions about blameworthiness and reasons that are fundamental in respective field. Individually, the two notions might seem incontrovertible, but I argue that they together entail claims that are highly dubious. In particular, I maintain that they entail implausible claims because the practices of moral blame and rational criticism differ with respect to three dimensions: justification, response, and function.
Thus, we reach a puzzle: We need to give up one of the principal notions about blameworthiness and reasons.
The responses to the puzzle imply that the connection between reasons and rationality is weaker than what usually is presumed in metaethics.
In the next section, I consider the two notions. According to a principal assumption in the literature on blameworthiness, an agent is morally blameworthy only if she has sufficiently strong reason to act otherwise than she did.
According to a principal assumption in metaethics, if an agent performs an action she has sufficiently strong reason not to do, she is irrational in a certain respect. The two notions entail that an agent who is morally blameworthy fails with respect to both morality and rationality. In Section 3, I spell out the practices of moral blame and rational criticism along three dimensions: justification, response, and function. It is found that the mentioned claim, together with certain uncontroversial assumptions, entails that an agent is morally blameworthy only if it is justified to direct both moral blame and rational criticism at her. In Section 4, which is the main section of the paper, I maintain that this latter contention implies a number of further claims that are unreasonable or even contradictory in view of the fact that morally blaming and rationally criticizing agents diverge with regard to the three dimensions. In Section 5, I spell out the mentioned puzzle and respond to objections. The paper is focused on laying out the puzzle rather than solving it, since the latter would require a thorough discussion. However, in Section 6, I briefly consider two solutions. The first maintains that there is no significant conceptual connection between reasons and rationality. Interestingly, on this alternative, my arguments provide unexpected support to a radical position in the recent debate on reasons and rationality. The second maintains that different reasons have different connections to rationality.

| BLAMEWORTHINESS, NORMATIVE REASONS, AND RATIONALITY
The first fundamental notion concerns the connection between the concept of moral blameworthiness and the concept of normative reasons: (1) An agent is morally blameworthy for φing only if she has a sufficiently strong normative reason not to φ and is fully informed.
The claim in (1) is generally presumed in the philosophical literature on blameworthiness. 2 To illustrate, Anna is morally blameworthy for neglecting to help a friend in financial trouble only if she has reason to help him and is aware of the relevant facts. If she comes to know that he is a millionaire, she would not be blameworthy in refusing to help him. It might be controversial how strong the reason entailed by blameworthiness should be assumed to be. According to an influential view, an agent is blameworthy for φing only if she has an all-things-considered strongest reason not to φ. 3 However, this view might be regarded as problematic on several grounds. 4 Therefore, I will let the formulation "a sufficiently strong normative reason" to be neutral about how strong the reason in question is. It is compatible with various possibilities, stretching from an all-things-considered strongest reason not to φ, to a pro tanto reason not to φ.
The second fundamental notion concerns the connection between the concept of normative reasons and the concept of practical rationality. In its most general formulation, it states that an agent has a reason to φ only if she would be motivated to φ if she were rational. It is difficult to overstate the impact of this type of claim in metaethics.
It functions as a guiding idea in theorizing about normative reasons and normative judgments, in particular moral reasons and moral judgments. 5 In the present context, it suffices to consider a weak version of the main idea: (2) If an agent has a sufficiently strong normative reason not to φ, then she would be overall most motivated not to φ if she were practically rational and fully informed.
There are presumably a number of explanations of the pervasiveness of claims like (2). To start with, intuitively there is a close connection between reasons and rationality that is reflected in ordinary thought and talk. Indeed, in metaethics, it is common to move seamlessly from assertions about an agent's reasons to assertions about her rationality and vice versa. Moreover, the presumption seems to be supported by reflection on certain types of reasons.
For instance, if an agent has a reason to leave the building because it is on fire and she has access to all relevant facts, but does not leave, this is taken to indicate that she fails in rationality. Further, the presumption gets support from various theoretical considerations. For instance, it might be maintained that the difference between "real" normative reasons-like prudential and moral reasons-and conventional reasons-like etiquette reasons-is that while the former has a conceptually necessary connection to rationality, the latter does not.
The fact that claims like (2) standardly are assumed in metaethics is even more evident in light of the fact that it is compatible with a variety of different views on reasons and rationality. First, it does not entail that the former can be reduced to the latter or that the latter is more fundamental than the former. It is consequently compatible with the contention that reasons are the primary normative concept. Indeed, it is implied by the view that rationality should be understood in terms of responding to reasons which have become influential in the recent debate on rationality. 6 Second, it does not presume any particular view of rationality, even if this concept should be understood independently of reasons. For example, it is compatible with procedural and substantive views of rationality and various instances of them. Third, as formulated here, it is compatible with a spectrum of views about the connection between the strength of reason and rational motivation. It is commonly assumed that if an agent has an all-thingsconsidered strongest reason to φ, then she must be overall most motivated to φ, given that she is rational and fully informed. However, there might be grounds to be skeptical about this type of reason. 7 Therefore, I will let the formulation "a sufficiently strong normative reason" to be neutral as to whether this is an all-things-considered strongest reason or something weaker. Lastly, it is concerned with an agent's rationality on the condition that she is fully informed about all relevant nonnormative facts. In the cases to be considered below, it is thus assumed that the agents are aware of all such facts.
It is now possible to see that from (1) and (2) follow a claim about the connection between the concept of blameworthiness and the concept of rationality: An agent is morally blameworthy for φing only if she would be overall most motivated not to φ if she were practically rational and fully informed. Assume that an agent would be overall most motivated not to φ if she were rational. However, assume that she φs in which case she is overall most motivated to φ. It then follows that she fails in rationality by being overall most motivated to φ. Hence, from (1) and (2) follows: (3) An agent is morally blameworthy for φing only if she would fail in practical rationality by being overall most motivated to φ.
As we have seen, (3) follows from two notions that are seen as uncontroversial and function as background assumptions in practical philosophy. In spite of this, (3) itself presumably does not strike us as uncontentious or trivial. Intuitively, it seems that an agent might be morally blameworthy for, say, stealing without necessarily failing in rationality by being most motivated to steal.

| NORMATIVE CRITICISM: MORAL BLAME AND RATIONAL CRITICISM
In the last section, it was found that two widely accepted notions in practical philosophy entail that there is a noteworthy connection between moral blameworthiness and rationality. In this section, we will see that these notions mean that there is an even more remarkable connection between blameworthiness and two types of normative criticism: moral blame and rational criticism.
It should be uncontroversial to maintain that an agent failing with regard to a certain normative standard entails that she is subject to normative criticism on that standard. In the present context, the standards are morality and rationality, and the two types of criticism blame and rational criticism.
As regards morality, failure to meet the normative standard justifies moral blame: (4) If an agent is morally blameworthy with respect to φing, it is justified to direct moral blame at her with respect to φing. 8,9,10 Different types of normative criticism share a certain basic structure. In particular, both moral blame and rational criticism entail a negative judgment that is purported to be justified on a certain ground, the judgment is accompanied by certain responses, and the expression of it fulfils a certain function.
If an agent A morally blames another agent B for φing, she makes a negative judgment about B for φing on the ground that B fails morally by φing. As often is emphasized, such judgments are closely associated with reactive attitudes such as resentment, indignation, or guilt. If A expresses her judgment, she might use particular words stretching from weaker to stronger depending on the nature of her criticism, such as "unkind," "inconsiderate," "immoral," "bad," "mean," "cruel," "evil," and so on. Moreover, A expressing blame to B typically has the function of modifying B's behavior in some respect. To illustrate, if A blames B for having lied to her, A makes a negative judgment about B on this ground and is likely to feel resentment toward B. Agent A might express her judgment to B by, for example, saying that B has betrayed her confidence. Moreover, A expressing her blame, and the support she provides for it, might have the function of making B less inclined to perform such actions.
As regards rationality, failure to meet the normative standard justifies rational criticism: (5) If an agent fails in practical rationality with respect to φing, it is justified to direct rational criticism at her with respect to φing.
In an ordinary and nontechnical sense of "rational," the same structure applies to rational criticism. If an agent A directs rational criticism at another agent B for φing, she makes a negative judgment about B for φing on the ground that B fails in rationality by φing. Moreover, the judgment might be accompanied by a negative attitude. Agent A might use certain words to express her judgment, such as "silly," "foolish," "bad idea," "unreasonable," "unintelligent," "irrational," "stupid," "idiotic," and so on. Derek Parfit observes: When we call some act 'rational', using this word in its ordinary, non-technical, sense, we express the kind of praise or approval that we can also express with words like 'sensible', 'reasonable', 'intelligent', and 'smart'. We use the word 'irrational' to express the kind of criticism that we express with words like 'senseless ', 'stupid', 'idiotic', and 'crazy'. 11 Moreover, A expressing rational criticism to B might have a certain function in relation to B's behavior. To illustrate, A might criticize B for his heavy smoking and consequently judge that B fails with respect to rationality in this regard. Agent A might express her judgment by saying that B is crazy in keeping on smoking. Moreover, A's expressing her criticism, and the support she provides for it, might have the function of getting B reconsider his smoking habit.
Thus, there are certain structural similarities between the practices of blame and rational criticism. The same structure occurs as regards moral and rational praise.
What is noteworthy is that from (3), (4), and (5) follows: (6) An agent is morally blameworthy for φing only if it is justified to direct moral blame and rational criticism at her for being overall most motivated to φ. 12 In my view, (6) should strike us as surprising. The fact that an agent is morally blameworthy entails that it is justified to morally blame her. However, why should this fact necessarily mean that it also is justified to direct rational criticism at her? At the same time, (6) follows from notions that appear quite uncontroversial and are regarded as platitudes in practical philosophy.

| MORAL BLAME AND RATIONAL CRITICISM: JUSTIFICATION, RESPONSE, AND FUNCTION
In what follows, I will argue that we should take our first impression of (6) at face value: The claim is implausible and should be denied. More precisely, (6) is mistaken because moral blame and rational criticism diverge along three dimensions: justification, response, and function.

| The argument from different justification
In this section, I argue that (6) provides a mistaken conception of the relation between the justification of moral blame and the justification of rational criticism.
We might start with observing that (6) faces an initial problem in that there seems to be cases where an agent is morally blameworthy, but it does not necessarily follow that it is justified to direct rational criticism at her. Thus, it seems that moral blame and rational criticism need not be connected in the manner entailed by (6). Consider: Beatrix. Beatrix is a computer technician who manipulates the computer system of the tax authority, thereby transferring a considerable amount of the taxpayers' money to her secret bank account. In doing this, she informs herself about all the relevant facts about the system, carefully evaluates the various options open to her, calculates the risk of being caught, etc. As a result, she has come up with a manner of manipulating the system ensuring that no one ever will find out.
First, it is plausible to assume that on a commonsense understanding of morality, Beatrix is morally blameworthy for manipulating the computer system, since it fulfils everyday criteria of such actions. 13 As a consequence, it follows that it is justified to morally blame her for the action. Thus, it seems accurate to claim that she was, say, "immoral," "greedy," or "bad" in manipulating the system. By contrast, it seems less reasonable to think that, on a common sense understanding of morality, it necessarily follows that it is justified to direct rational criticism at her. Thus, it seems more doubtful that it would be accurate to claim that she was, say, "unintelligent," "not that smart," or "irrational" in manipulating the system. Second, it is plausible to assume that on most normative theories in ethics, Beatrix would be blameworthy for her action, in which case it follows that it is justified to blame her, although they disagree about the grounds for blaming her. However, it does not seem to be the case that these theories, in general, imply that it would be justified to direct rational criticism at her. 14 There are reasons to think that (6) faces a more serious problem because it means that what would justify rational criticism of an agent needs to be identical to what justifies moral blame of her. According to (6), if an agent is morally blameworthy for performing an action, it is justified to direct both moral blame and rational criticism at her. This claim raises the following question: What is it that justifies blame and rational criticism in such cases? The claims that lead up to (6) entail a particular answer to this question: What would justify rational criticism must be the same as what justifies moral blame. To see this, let us consider the claims that lead up to (6). First, an agent is morally blameworthy for performing an action only if she has a sufficiently strong reason not to perform it. It is her failure to abide by this reason that justifies directing moral blame at her. (From (1) and (4)). Second, if an agent has a sufficiently strong reason not to perform an action, she would be overall most motivated not to do it if she were rational and fully informed. It is her failure to abide by this reason that justifies directing rational criticism at her. (From (2) and (5)). It follows: If it is justified to morally blame an agent for performing an action because she has a sufficiently strong reason not to do it, it would be justified to rationally criticize her for performing the action because she has that very reason not to perform it.
However, it seems implausible to assume that the reason which would make it justified to direct rational criticism at an agent for performing an action needs to be the same reason which makes it justified to morally blame her for performing that action. To illustrate, assume that the reason not to manipulate the computer system is that doing so would have overall bad consequences. Thus, it is the fact that Beatrix's action has this feature which makes it justified to direct moral blame at her. However, it appears farfetched to maintain that it needs to be the fact that manipulating the computer system has overall bad consequences which would make it justified to direct rational criticism at her. The type of reason that typically makes it justified to morally blame an agent does not typically seem to be the type of reason that makes it justified to rationally criticize her. It should be stressed that this argument is consistent with it being justified to direct both blame and rational criticism at an agent for performing a blameworthy action and that the reason for doing so might be the same. The point is that it is a much stronger and implausible claim that the reasons need to be identical in the indicated manner. It should further be noted that it does not matter in what the reason in question consists. In the example, I referred to an act having certain consequences, but the same result is reached in case we refer to reasons suggested by alternative views.
Importantly, (6) faces an even graver difficulty in that it implies contradictory judgments about an agent's rationality in relation to a certain action. The source of this difficulty is that there are cases where it seems justified to morally blame an agent for an action and, at the same time, justified to rationally praise her for that very action. To illustrate, return to Beatrix. On the one hand, it is justified to morally blame Beatrix for manipulating the computer system of the tax authority. On the other hand, it also seems justified to rationally praise her for the action. In carrying out the action, Beatrix carefully informs herself about all relevant facts about the system, meticulously considers the various options, and so on. It seems plausible to say that in performing this action, she was "smart," "clever," "intelligent," and so on. According to (6), if it is justified to morally blame an agent for performing an action, it is also justified to direct rational criticism at her for performing it. Thus, this view entails that it is both justified to rationally criticize Beatrix and justified to rationally praise her for manipulating the system. However, it is paradoxical to maintain that an agent can be both rationally criticized and rationally praised for the very same action. Accordingly, on this view, it would be correct to claim that Beatrix is both "stupid" and "smart" in performing the action in question. It seems implausible to maintain that it would be correct to apply these characteristics to a person in relation to the very same action. Moreover, even if it would be possible to come up with some occasional cases where this verdict seems reasonable, we should resist a view which entails that it needs to be correct in standard cases like the one above. Alternatively, an advocate of (6) would need to maintain that Beatrix is not rationally praiseworthy, but this option seems difficult to sustain.
It might be tempting to respond to this argument by maintaining that it is justified to rationally criticize and praise an agent for one and the same action, but what justify the criticism and what justify the praise are different features of the action. For instance, what justifies rational criticism of Beatrix for manipulating the computer system might be that it has overall bad consequences, while what justifies rational praise of her might be some other feature of the action. This line of defence does not succeed however. First, it is at least doubtful that it is justified to rationally criticize Beatrix for her action with respect to any of its features. It is not obvious that it needs to have any feature which would justify calling her "stupid," "irrational," "unintelligent," or the like. Second, and more importantly, we can see that this line of defence does not meet the objection by returning to what justifies moral blame and rational criticism. According to the view under consideration, if it is justified to blame an agent for an action because she has a sufficiently strong reason not to perform it, it is justified to rationally criticize her for performing the action because she has that very reason not to do it. As a result, the content of this reason puts restrictions on what features of an action that can justify rational praise of an agent. Assume that the reason not to manipulate the computer system is that doing so has overall bad consequences. According to the view under discussion, it is then justified to rationally criticize Beatrix for performing the action because it has these consequences. In that case, it can hardly be justified to rationally praise her for the action because it has any feature that makes the action have these consequences, for example being carried out with skillful manipulation of the system. However, this puts implausible restrictions on what features of the action that can justify rational praise of Beatrix. Indeed, it is presumably precisely the features of the actions that made it have these consequences that justify rational praise of her.
To sum up, (6) suffers from three difficulties as regards the connection between justification of moral blame and rational criticism. First, it entails that from the fact that an agent is morally blameworthy, it necessarily follows that it is justified to rationally criticize her. Second, it entails that what would justify rational criticism needs to be the same as what justifies blame. Third, it entails contradictory judgments as regards an agent being rationally criticizable and rationally praiseworthy.

| The argument from different responses
In this section, I argue that (6) presents an implausible conception of the relation between the responses associated with moral blame and the responses associated with rational criticism.
If agent A morally blames agent B for φing, she makes a negative judgment about B for φing on the ground that B fails morally by φing. The judgment is closely associated with reactive attitudes toward B, such as resentment or indignation. A number of authors argue that such judgments partly or wholly consist in reactive attitudes. 15 In this paper, I will only make the weaker and uncontroversial assumption that they typically are accompanied by reactive attitudes.
In a corresponding manner, if A directs rational criticism at B for φing, she makes a negative judgment about B for φing on the ground that B fails in rationality. Also in this case, A's negative judgment might be accompanied by a negative attitude, such as contempt for B's poor use of her intellectual abilities, irritation at B, or concern for B's future.
According to (6), an agent is morally blameworthy for φing only if it would be justified to direct both moral blame and rational criticism at her. In conjunction with what was said above about the responses associated with blame and rational criticism, (6) entails the following: If agent A directs moral blame at agent B for φing, A typically has both reactive attitudes associated with moral blame at B and makes the kind of judgment associated with rational criticism about B. 16 We might start with observing that (6) faces an initial problem in that responses associated with moral blame do not seem to be systematically accompanied by responses associated with rational criticism in this manner. Consider: Betty and Anne. Betty is a criminal who regularly breaks into people's home, threatens them and sometimes beats them in order to steal their belongings. One night, Betty breaks into Anne's house and beats her so brutally that she will suffer severe physical and mental damages for life.
It is plausible to think that Anne feels moral resentment and indignation at Betty for what she has done to her. By contrast, it is much less evident that Anne judges that Betty is irrational, not that smart, stupid, or the like, in performing the action. Thus, there seem to be reasons to be skeptical to the conception (6) provides of the connection between responses of moral blame and judgments of rational criticism. This skepticism is confirmed by the findings of the last section. According to (6), if Anne is justified to feel resentment at Betty, Anne is also justified to judge that Betty fails in rationality. Moreover, what would justify Anne's negative judgment about Betty's rationality is the same as what justifies Anne's resentment at Betty. However, it appears that while Anne is justified to feel resentment at Betty, it is more doubtful that she is justified to judge that Betty fails in rationality. Moreover, it seems even less evident that Anne's grounds for a negative judgment about Betty's rationality would be the same as her grounds for feeling resentment at Betty.
There are reasons to think that (6) faces a more severe problem in that the responses associated with moral blame and the responses associated with rational criticism might pull in different directions. In the last section, we found that, on (6), what would justify rational criticism needs to be the same as what justifies blame. More precisely: If it is justified for A to morally blame B for φing because B has a sufficiently strong reason not to φ, it is justified for A to rationally criticize B for φing because B has that very reason not to φ. In this section, we have seen that (6) means the following: If A morally blames B for φing, she typically has both the reactive attitudes associated with blame of B and makes the kind of judgment associated with rational criticism about B. As a consequence, (6) predicts that A's responses with respect to her rational criticism of B correspond to her responses with respect to her moral blame of B. More precisely, since A is justified to morally blame B for φing because B has a sufficiently strong reason not to φ and A is justified to direct rational criticism at B for φing because B has that very reason not to φ, A's judgment about B's rationality should match her moral reactive attitudes toward B. That is, A having a low level of rational criticism of B should correspond to A having a low level of moral resentment toward B. Conversely, A having a high level of rational criticism of B should correspond to A having a high level of moral resentment toward B. The problem is that the reactive attitudes related to moral blame and the judgments related to rational criticism might diverge in at least two manners.
First, if A judges that B excels with respect to rationality in φing, A would presumably not be less inclined to have moral reactive attitudes toward B for φing. To see this, let us complement the scenario above: After a period of time, Anne learns that Betty has planned the break-in and what to do with Anne meticulously. Betty informed herself about all the relevant facts about Anne's house, her lack of physical strength, carefully evaluated the most efficient way to beat her in order to lessen her resistance, calculated the risk of being caught, etc.
How would Anne react when she receives this information? It seems plausible to hypothesize that Anne's feeling of resentment toward Betty will not decrease, but stay the same, or even increase, as she comes to know that Betty did not fail in rationality but rather excelled in it when she planned the break-in. Indeed, it might even be the case that although Anne feels deep resentment toward Betty, she also judges that Betty actually was quite clever in her careful preparations. However, this runs contrary to (6) which predicts that Anne's responses associated with moral blame and rational criticism should correspond with one another. More precisely, (6) predicts that after judging that Betty excelled in rationality, Anne would feel less resentment toward Betty. But this does not seem to be the case: Anne would not feel less, but perhaps even more, resentment toward Betty after acquiring this information.
Second, if A judges that B fails with respect to rationality in φing, A would presumably not be more inclined to have moral reactive attitudes toward B for φing. We might see this by adding the following to the original scenario: After a period of time, Anne learns that Betty's planning of the break-in was very sloppy. Betty did not inform herself about the relevant facts about Anne's house or her physical abilities, did not consider what would be the best way to lessen her resistance, did not calculate the risk of being caught, etc. Indeed, because of her bad planning, there was a serious risk that Betty could have failed completely and been caught.
In this scenario, we get the reverse result compared to the former case. It seems plausible to hypothesize that Anne's feeling of resentment toward Betty will not increase, but stay the same, or even decrease, as she is informed that Betty failed in rationality when planning the break-in. Indeed, it might even be the case that although Anne's feeling of resentment stays the same or decreases, she now looks with some disdain at Betty for not having planned the inbreak more cautiously. Once again, this runs contrary to (6) according to which it should be expected that Anne's responses associated with moral blame and rational criticism correspond with one another. That is, (6) predicts that after judging that Betty failed in rationality, Anne would feel more resentment toward Betty.
It might be speculated why the responses associated with moral blame and rational criticism have a tendency to diverge. An underlying explanation is presumably that if we judge that an agent excelled with respect to rationality in performing a blameworthy action, this tells in disfavor of her moral character, thus making us more prone to feel resentment toward her. Conversely, if we judge that an agent failed with respect to rationality in performing a blameworthy action, this tells against her moral character to a lower extent, thus making us less inclined to feel resentment toward her than otherwise would have been the case.
In response to the arguments in this section, it might be suggested that in cases where an agent A morally blames another agent B, considerations about B's rationality become less salient. For example, Anne is focused on blaming Betty because of what she has done to her, but this is compatible with Anne in addition judging that Betty fails in rationality. It is just that this judgment is less salient to Anne, perhaps because emotions associated with blame are in the foreground of her attention. 17 It might be granted that this reply provides an answer to the initial argument which observes that blame does not seem to be systematically accompanied by rational criticism. However, I doubt that it applies to the subsequent arguments. First, the basic idea in these arguments is that responses of blame and rational criticism pull in different directions in two distinct manners. As this seems to be the case even if Anne's judgment about Betty's rationality is salient to her, I doubt that the mentioned suggestion lessens the force of the arguments. Second, in the two scenarios, some time has gone since Anne was exposed to Betty's action. As a result, it is reasonable to think that Anne's emotional response associated with blame is less in the foreground, and it is not implausible to consider whether Anne holds a judgment about Betty's rationality.
To sum up, (6) suffers from two difficulties as regards the connection between responses pertaining to moral blame and rational criticism. First, it entails that responses associated with blame are systematically accompanied by responses associated with rational criticism. Second, it predicts that reactive attitudes and rational criticism correspond, although they in fact pull in different directions.

| The argument from different functions
In this section, I argue that (6) provides an erroneous conception of the relation between the function of expressing moral blame and the function of expressing rational criticism.
If agent A morally blames agent B for φing, A makes a negative judgment about B on the ground that B fails morally by φing, and A's judgment is typically accompanied by a reactive attitude toward B. Agent A might express her negative judgment to B by employing language, overt behavior, or facial expression. Moreover, agent A expressing a negative judgment to another agent B typically has a certain purpose. In the literature, it is characteristically taken to consist in A making B admit that she is blameworthy, take responsibility, and feel regret for φing. Most significantly, A expressing a negative judgment to B typically has the general function of regulating B's attitudes or behavior so as to make B improve morally in a certain respect. 18 An explanation of why A expressing blame to B can have this function is presumably that A's negative judgment typically is accompanied by a reactive attitude toward B for φing, since we can influence other agents by letting them know about our attitudes.
It is plausible to assume that rational criticism works in a corresponding manner. 19 If agent A directs rational criticism at another agent B for φing, A makes a negative judgment about B on the ground that B fails in rationality by φing, and the judgment might be accompanied by a negative attitude, such as contempt, irritation, frustration, or concern. In the event A employs language in expressing her negative judgment, she may use words like "foolish," "irrational," and "unintelligent." It seems reasonable to suggest that A expressing her negative judgment to B can have a certain purpose or point, although the connection is looser than in the moral case. It might consist in B realizing that she has failed in rationality and feel foolish about φing. More generally, A expressing her negative judgment may have the function of regulating B's attitudes or behavior so as to make B improve rationally and deliberate more thoroughly about a certain issue.
According to (6), an agent is morally blameworthy for φing only if it would be justified to direct both moral blame and rational criticism at her for being overall most motivated to φ. In the last section, we saw that this claim entails a certain connection between the responses associated with blame and rational criticism: If A morally blames B for φing, she typically has both the reactive attitudes associated with moral blame of B and makes the kind of judgment associated with rational criticism about B. In conjunction with the functions of expressing blame and rational criticism, (6) entails the following: If A expresses moral blame to B for φing, this typically has both the function associated with expressing moral blame and the function associated with expressing rational criticism.
There are reasons to think that (6) faces an initial problem in that the function of expressing moral blame does not seem to be accompanied by the function of expressing rational criticism in this manner. Consider: Ada and Bertie. Ada and Bertie are friends and use to help one another when needed. One day, Bertie asks Ada to lend him some money. Although Ada has financial problems herself, she fulfils Bertie's request. Bertie promises to repay Ada within 2 weeks. However, after 1 month, Ada has still not heard from Bertie. Ada calls him and expresses her blame: 'It's really rotten of you! You know that I'm in a difficult situation'.
It is plausible to assume that Ada expressing moral blame to Bertie has the typical function of such communicative acts. For example, it is likely to have the purpose of making him understand that he has acted in an unacceptable manner toward her, feel bad about it, and thereby get him to repay the loan. The claim in (6) implies that it further has the function of expressing rational criticism. However, this appears much less evident. Especially, it is not in the same manner likely that Ada expressing blame to Bertie has the purpose of making him think that he has failed in rationality and make him reason more carefully. Indeed, as far as Ada knows Bertie's refraining from repaying the loan might be based on considerable deliberation. It is worth noticing that the present argument is compatible with expressions of moral blame typically having the consequences associated with expressions of rational criticism. That is, it might well be that A expressing blame to B typically has the effect that B feels intellectually inferior and comes to think that she needs to deliberate more thoroughly. However, even if this might be a consequence of communicating blame, it is not the function such acts primarily are aimed to fulfil.
There are reasons to think that (6) is subject to a more serious problem in that the function of rational criticism does not seem to correspond with the function of moral blame in the manner suggested by this view. The claim entails that if A expresses moral blame to B for φing, this typically has both the function of expressing blame and the function of expressing rational criticism to B. As a consequence, it predicts that if A thinks that expressing rational criticism to B would not fulfil its function, she would typically abstain from expressing moral blame to B. However, this does not seem to be the case. To see this, we might add some information to the scenario: After a period of time, Ada learns that Bertie's failure to fulfil his promise to repay the loan is not a matter of mistake, misunderstanding, oversight, or the like. It turns out that Bertie informed himself about all the relevant facts and carefully considered from whom to borrow so as to postpone a repayment of the loan and minimize the risk of an unpleasant confrontation.
In view of this information, it is plausible to think that Ada recognizes that expressing rational criticism to Bertie would not fulfil its typical function. The explanation is that she has come to understand that he evinces typically traits of rationality in which case expressing rational criticism would be pointless. The claim in (6) predicts that Ada would abstain from expressing moral blame to Bertie with the purpose of this type of communicative acts. However, the fact that Ada now realizes that Bertie does not fail in rationality and that expressing rational criticism to him would not fulfil its function does not seem to mean that she would abstain from expressing moral blame to him. It is reasonable to assume that she thinks that expressing blame might fulfil its typical function. For example, she might plausibly believe that the utterance mentioned in the original scenario has some prospect of influencing him to take responsibility for his action, regret it, and repay the loan. After all, she has not received any information indicating that he is less responsive to expressions of moral blame than she thought before, although she now is aware that he is not responsive to expressions of rational criticism.
To sum up, (6) suffers from two difficulties as regards the connection between the functions of expressing moral blame and rational criticism. First, it entails that the function of expressing blame is systematically accompanied by the function of expressing rational criticism. Second, it predicts that an agent would refrain from expressing blame if she thinks that expressing rational criticism would not fulfil its typical function.

| A PUZZLE ABOUT BLAME AND REASONS
Let us take stock. We started with seeing that two fundamental notions in practical philosophy, (1) and (2), together entail a disputable claim about the connection between moral blameworthiness and rationality, (3). We then saw that (3), together with certain uncontroversial assumptions about normative criticism, (4) and (5), entail an even more problematic claim, (6), about the connection between moral blame and rational criticism. We finally saw that there are strong reasons to think that (6) is false, since moral blame and rational criticism diverge as regards justification, response, and purpose.
The arguments against (6) reveal a pattern concerning the different manners in which it is mistaken. First, it entails that what is the case as regards moral blame is the case as regards rational criticism with respect to justification, response, and function (beginning of 4.1, 4.2, and 4.3). Second, it entails that what would justify rational criticism and what justifies moral blame need to be the same (4.1). As a result, it further implies contradictory judgments as regards rational criticism and praiseworthiness (4.1). Third, it predicts correspondences between the responses in relation to moral blame and rational criticism, although they might pull in different directions (4.2). Likewise, it predicts correspondences between the functions of expressing moral blame and expressing rational criticism, although they might come apart (4.3).
However, it might be argued that my arguments against (6) are mistaken. There are particularly two objections on which I would like to comment.
The first objection is that the arguments overlook that an agent fails in rationality if she does not respond to reason. In the scenarios above, agent B appears to have a sufficiently strong normative reason not to perform an action but does so in spite of being fully informed. It might then be thought that she needs to be lacking in rationality and that rational criticism at her is justified, since she does not respond to a reason that is available to her.
First, the objection takes for granted that there is a connection between reasons and rationality. It presumes that if an agent has a sufficiently strong reason not perform an action, but does so despite being fully informed, she fails in rationality. Thus, it presumes (2) or some closely related claim. However, this is one of the principal notions that lead up to (6), in which case, it cannot be appealed to in defence of this claim. Second, my reasoning is compatible with the view that such a connection holds between certain normative reasons and rationality, although it does not hold between moral reasons and rationality. I will return to this possibility in the next section. Third, the objection would only constitute a limited problem for my arguments even if it were correct. It would challenge the initial type of arguments as regards justification and response at the beginning of 4.1 and 4.2. However, it would not challenge the more substantial arguments. It would not challenge the second type of argument about what justifies moral blame and rational criticism in 4.1. Moreover, it would not challenge the third type of arguments about correspondences between moral blame and rational criticism with respect to response and function in 4.2 and 4.3.
The second objection is that my arguments presume a too narrow understanding of practical rationality. In the scenarios above, agent B is concerned primarily with actions that would promote her own goals, and it was suggested that it need not be justified to direct rational criticism at her. It might then be thought that my arguments implicitly trade on an egoist understanding of this notion.
First, the arguments do not presume any egoist view of practical rationality according to which an agent is rational insofar as she serves her own interests and disregard other people's concerns. They make the minimal assumption that a requirement on an agent being rational is that she considers possible lines of actions and the outcomes of these actions in the light of her goals and the information to which she has access. However, it is uncontroversial that such a broad notion of instrumental rationality is a fundamental aspect of rationality. Second, the arguments do not presume that this is the only aspect of practical rationality. They are compatible with an agent being rational means that she needs to meet other and more demanding requirements. Third, it is important to stress that proponents of (6), by contrast, are committed to a substantial view about the connection between morality and rationality.
They are committed to the claim that if an agent is morally blameworthy, it necessarily follows that she fails in rationality and it is justified to direct rational criticism at her. However, it is difficult to uphold this claim in consideration of the arguments provided above. Finally, the second objection, like the first one, would only constitute a challenge to the initial type of arguments concerning justification and response (beginning of 4.1. and 4.2). However, it would not challenge the second and third type of arguments (4.1, 4.2, and 4.3).
Thus, we face a puzzle: While both (1) and (2) are fundamental notions in practical philosophy, they entail, together with certain uncontroversial assumptions, (6) which there are strong reasons to deny. It appears that we need to reject either of the notions leading up to this view: (1) or (2). In my view, it would presumably be difficult to give up (1) although this option should not be excluded. At least in standard cases of moral blameworthiness, it seems hard to deny that an agent is blameworthy only if she has a sufficiently strong reason not to act as she did. This is particularly the case in light of the connection between blameworthiness and blame in (4). If an agent does not have a sufficiently strong reason to avoid acting as she did, it can hardly be justified to blame her and direct reactive attitudes at her. In finding a solution to the puzzle, I think we need to consider rejecting (2).

| BRIEF OUTLINE OF TWO SOLUTIONS
Recall: (2) If an agent has a sufficiently strong normative reason not to φ, then she would be overall most motivated not to φ if she were practically rational and fully informed.
As far as I understand, there are basically two ways to reject (2) and thereby solve the puzzle.
The first way to reject (2) is to deny that there is any significant conceptual connection between reasons and rationality, in which case rationality might be thought to lack normativity. This alternative is radical in view of the fact that the two concepts are closely associated with one another by both philosophers and laypeople. However, it conforms to a recognized position in the intense and ongoing debate on the relation between reasons and rationality. As the considerations that might be advanced for this position are quite complex, I merely give a broad indication of them in the main text and provide some of the arguments in the footnotes. According to a common assumption in this debate, an agent is rational insofar as she adheres to certain rational requirements, where this is understood as having attitudes that are coherent with one another in particular manners. 20 It is then argued that an agent does not have any reasons to have coherent attitudes that conform to such rational requirements. 21,22 This general contention can be taken to support the view that there are no conceptual connections between reasons and rationality like (2). 23 As a result, (6) would not follow. Surprisingly, the considerations submitted in this paper about moral blame and rational criticism can thus be regarded as further and independent arguments for denying that there is any significant conceptual connection between reasons and rationality.
The second way to reject (2) is to argue that different types of reasons have different connections to rationality.
According to one version of this view, there is a distinction between rationally requiring reasons and rationally permissible reasons. The distinction is understood slightly differently by different authors. 24 However, a broad characterization is the following: An agent has a rationally requiring reason to φ insofar as she is rationally required to be motivated to φ. That is, she fails in rationality if she is not motivated at all to φ. An agent has a rationally permissible reason to φ insofar as she is rationally justified, but not rationally required, to be motivated to φ. That is, she need not fail in rationality even if she lacks any motivation to φ. We can now see that (2) is correct as regards rationally requiring reasons: If an agent has a sufficiently strong rationally requiring reason not to φ, then she would be overall most motivated not to φ if she were rational and fully informed. By contrast, the following claim is true for rationally permissible reasons: (2 0 ) If an agent has a sufficiently strong rationally permissible reason not to φ, she need not be overall most motivated not to φ even if she is rational and fully informed. Why would this be relevant in the present context? Consider (1): An agent is morally blameworthy for φing only if she has a sufficiently strong normative reason not to φ and is fully informed. Assume that moral reasons consist in rationally permissible reasons. Especially, suppose that the reason referred to in this claim is a rationally permissible reason. Importantly, (1) in conjunction with (2 0 ) does not entail (3). That is, an agent who is morally blameworthy for φing need not fail in rationality. As a result, when (1) and (2 0 ) are combined with the other relevant claims, (4) and (5), (6) does not follow. In particular, if an agent is morally blameworthy for φing, it is justified to morally blame her for being most motivated to φ, but it is not justified to direct rational criticism at her on that ground.
In this paper, I do not have space to discuss which of the mentioned alternatives that are most plausible. However, I would like to end by observing that both options face challenges. The first option runs contrary to the intuitive view that the concepts of reasons and rationality are necessarily connected. Moreover, it needs to provide an explanation of why agents should avoid being irrational given that rationality as such lacks normativity. 25 The second option complicates the relation between reasons and rationality in a way that have a number of significant consequences. Most importantly, it seems difficult to weigh rationally requiring and rationally permissible reasons against one another in such a manner that conclusions can be reached about what agents have all-things-considered strongest reasons to do.

ACKNOWLEDGMENTS
I am grateful to Andreas Brekke Carlsson, Carla Bagnoli, and Christel Fricke for valuable comments on earlier versions of this paper. Special thanks are due to an anonymous reviewer for EJP for perceptive and detailed comments that helped me improve the paper considerably.
7 For instance, we should allow for the possibility that different kinds of reasons cannot be weighed against one another in such a way that an agent has an all-things-considered strongest reason to perform a particular action. 8 It is important to distinguish between non-expressed and expressed blame and rational criticism. In using "blame" and "direct blame," I have non-expressed blame in mind unless I indicate otherwise. The same applies to "rationally criticize" and "direct rational criticism."