Response to Critics

Phillip Stratton-Lake

Both TH and JD raise many interesting and important issues. To do justice to all of these issues in thirty minutes is just impossible. So I do not pretend that the responses I offer here are fully adequate – they couldn’t be. They merely indicate the direction in which a full response might go, some (I must admit) more promising than others. But even this more modest aim has meant that I cannot address every point that my critics raise. I have, however, tried to focus on what I take to be the more important issues that can be addressed in the time available to me.

 

JD’s comments

 

1. JD’s first point is a request for clarification. He notes a certain ambiguity in what it is I say needs explaining. He notes that sometimes I claim that it is the possibility of a normative fact - the moral must – that needs explaining. At other times I claim it is the possibility of a phenomenological fact – our experience of the moral must – that needs explaining. I have to admit that in the book I did switch from the normative to phenomenological in this way. But it is really the phenomenological fact that I start from. Like Kant, I assume that we can learn about the object of this experience, the moral must itself, by understanding our experience of it and how this experience is possible. I claim that Kant is best understood as maintaining that the Moral Law explains the possibility of this experience. Given the validity of some form of transcendental idealism, if the moral law explains the possibility of this experience it also explains the possibility of the object of this experience. But my primary interest is in explaining the possibility of the experience of the moral must, the experience of being bound to do some act in a way that is independent of our desires and concerns. As I understand him, Kant argues that it is the moral law that explains this experience. When we experience some act as necessary in the circumstances we take ourselves to stand under an absolutely universal law. The strict universality of this law must, in turn, be traced back to the moral law itself. What I wanted to emphasise is that this law is not to be taken as a very abstract ground of duty. It should, rather, be understood as having a transcendental, not a justificatory role.

 

2. JD, asks, how the moral law can explain the possibility of our experience of particular imperatives if the moral law is itself experienced as an imperative. It seems that my account simply begs the question at issue, or at least leaves one such experience unexplained.

 

The first point I would like to make in response to this is that the moral law is not intrinsically imperatival. In itself it is the mere form of law as such, which for Kant is the feature that all laws, qua laws, possess – namely, mere universality. It is for this reason that he calls it the mere form of law as such. In itself, then, the moral law is not an imperative. On Kant’s view, perfectly rational beings act from the same fundamental principle as we do; though it does not appear to them in the form it appears to us. It appears to us as an imperative because of our finitude, because we do not necessarily act in accordance with law-like maxims.

 

The second, and more important, point is that we do not have an experience of being bound by the moral law that is distinct from our experience of being bound to do some particular act. Kant need not be understood as claiming that we first experience the moral law as an imperative and then have a distinct experience of being required to do particular acts in particular contexts.[1] If we did, both experiences of moral necessity would have to be explained. Kant may, however, be understood as claiming that we only experience the imperatival character of the moral law in our experience of being bound to do some particular act. This experience is explained with reference to our finite nature and our taking ourselves to stand under an unconditional law which, qua unconditional, must be subsumed under the mere form of universality as such, i.e. the moral law. But the imperatival character of this law is only felt in our feeling constrained to do some specific action, and is not separable from this. Understood in this way, there is not a separate question of how we are to experience the moral law as an imperative.

 

3. Isn’t the Kantian explanation of the experience of moral necessity I outline refuted by particularists who deny any important role for universal principles? Many people do not think that they are subject to Kant’s moral law: JD confesses to being such a person. But such people are still able to experience moral requirements. JD notes that I do not think transcendental claims can be refuted in this way, but insists that if the ‘can’ that is at issue is non-normative, then it is falsified by particularists. Particularists do not see themselves as subsuming particular acts under strictly universal principles. If they did they wouldn’t be particularists. Isn’t it presumptuous to insist that particularists’s self-understanding is mistaken?

 

Well, this may sound presumptuous, but I think that transcendental principles cannot be refuted in this way. This is not because the relevant ‘can’ is normative, but because the claim that is being made is transcendental rather than psychological. I do not think I refute what Kant says in the first Critique by claiming, on the basis of introspection, that I am not aware of subsuming my intuitions under the categories. I suspect none of us are aware of doing this. If introspection could refute what Kant says in the first Critique, then no complex arguments are needed against Kant’s view. We need only note that no one has ever caught himself or herself applying the categories. We could then conclude that since these people have experience without applying the categories, the categories cannot be necessary for experience.

 

This is not of course to say that particularists are mistaken about the form of their moral deliberation. I am not insisting that universal moral principles must figure in one’s practical deliberation if one is to conclude that one must act in a certain way. The distinction between the transcendental and the justificatory role of moral principles enables us to claim some essential role for moral principles without insisting that they must, or should, figure in our deliberation. On the view I favour, the moral law and the more specific, intermediate moral principles function solely as transcendental conditions of some consideration necessitating some act. They do not also function as considerations that necessitate this act, or as part of some more complex consideration that necessitates this act. As such, there is no presumption that they should figure in our deliberation about what to do.

 

Kant may be wrong in maintaining that we need to synthesise our experience by means of the categories, but I do not think that he can be shown to be wrong by introspection. Similarly, if the moral law is to be understood as a transcendental principle, explaining the possibility of moral experience, I do not think it can be refuted by introspection.

 

4. JD suggests that the possibility of the moral must might be explained in some other way, a way with which the particularist would be more comfortable. One possibility JD suggests is that the possibility of fully particular laws that explain particular musts, which do not need to be explained by universal laws. This proposal accepts the assumption that the relevant sort of law is one that states what must be done, but pushes the possibility that this law could be fully particular. It could be fully particular in the sense that the thing that must be done is a particular action and the law only requires this particular action in this particular set of circumstances. According to this proposal, there is no universal claim to the effect that this type of act must always be done in this type of circumstances. Call this the ‘particular law’ account of moral necessity.

 

Another possibility JD suggests is ‘to try to appeal simply to the differences between different types of reasons’. The type of reasons he has in mind are those that are categorical n the sense that they independent of the agent’s purposes or concerns. The idea is that the balance of categorical reasons can explain the categorical musts they give rise to. Call this the ‘balance of reasons’ account of moral necessity.

 

In my book I do not offer arguments for the view that particularists cannot explain moral necessity. I simply outline a Kantian explanation of this. Whether some particularist alternative is better will depend on the details of their alternative account, and this is not the place to push this matter. I shall, however, briefly mention some problems the particular law and balance of reasons accounts of moral necessity may run into.

 

The main problem I have with the particular law account of moral necessity is that there does not seem to be enough of a gap between the fully particular law and the particular act it requires. What is the difference between a law that states that I must do some particular act A in a particular set of circumstances C and the fact that I must do A in C?  There must be a difference if the particular must is to be explained by the particular law, but it is unclear what this difference is.

 

One might also wonder what conception of law is at work here. The Humean, regularity theory of laws has trouble even with laws that are instantiated only once, and whereas Armstrong’s conception of laws[2] as relations between universals allows for there to be just one instance of a law in a possible world, this falls a long way short of JD’s fully particular law. As JD understands it, a fully particular law not only might be instantiated only once, but could only be instantiated once. It would seem that any law cannot be fully particular in the way JD suggests, but would have to pick out some sort of regularity. If this is right, then a particular moral must could not be explained with reference to a fully particular law.

 

Finally, it is not clear to me that we make sense of a law that requires that some particular (token) act be done prior to the doing of that particular act? It seems to me that prior to the doing of the token act we can only talk of act types.

 

I have to say that I am more inclined to the balance of reasons account of moral necessity than I was when I wrote the book, but I still have some doubts about it. There is, I think no doubt that the balance of reasons fully explains the oughts they give rise to. The only issue, then, is whether musts are different from oughts. I think there are a number of ways in which they are different, and these are intimated in the way in which we use ‘must’ instead of ‘ought’. In particular, in thinking that I must do something I think not only that the balance of reasons favours this act, but also think of myself as unable to do anything else. As Bernard Williams has pointed out, this feature is particularly conspicuous when we apply these terms to the past. There is nothing odd about saying ‘I ought to have Φed, but I didn’t’. It would, however, be odd to say ‘I had to Φ, but I didn’t’.

 

If there is something to the idea that ‘ought’ is distinct from ‘must’, and if ‘ought’ can be understood solely in terms of the balance of reasons, then we cannot understand ‘must’ in this way. There seems to be more to the moral must (and not just the moral one) than that the balance of reasons come down a certain way, or even that they come done emphatically a certain way. A necessitating reason is not simply an undefeated reason, not even an undefeated categorical reason. I think the fact that I could save my life by jumping off the burning building is a reason for me to jump if my future life would be worth living. I also think I have this reason even if I do not, at present care about my future life, so this reason looks categorical. But although I think this undefeated reason means that I ought to jump, it is not clear to me that it implies that I must jump. I am not, therefore, wholly convinced that one can understand the moral must with the idea of an undefeated categorical reason.

 

It might be that a necessitating reason is one that not only favours a certain course of action but also silences (McDowell) or excludes (Raz) other competing considerations. This seems to me to be the right way to go in understanding moral necessity in terms of reasons. But however much such an account helps us to understand the nature of moral necessity, it still seems to leave us with a possibility question. How is it possible not only for some consideration to favour an action, but also to exclude other competing reasons from consideration? Why do some considerations exclude as well as favour, whereas others do not? It seems to me that the prospects for the balance of reasons account will depend on its ability to answer these questions, or at least to show why it does not have to answer them.

 

5. JD notes that the standard objection to particularism is to insist that moral reasons obtain if and only if certain principles obtain. This biconditional is often defended by the claim that a moral reason for X just is a feature that would make X a duty if it were the only morally relevant feature. ‘Every reason-giving feature is such that if alone it is a duty-making feature’ (7). JD thinks that there is a decisive refutation of this argument. Whether my defence of universal principles is different from this, and so not vulnerable to the same refutation, depends, on the details of my view. He claims there are two ways in which I propose the relation between the moral law, intermediate moral principles and required acts. On one interpretation my view is distinctive, but, he claims, can be objected to on other grounds. On the other interpretation my view is not distinctive, and so is vulnerable to the same objection as the standard (subjunctive conditional) view.

 

On the first interpretation of my view the necessity of specific acts is explained with reference to intermediate moral principles, the possibility of which is explained in turn by the moral law. On this view, the moral law does not stand in a direct enabling relation to the deontic feature of particular acts, but stands in this relation to the intermediate moral principles. These intermediate principles then stand in the same enabling relation to the relevant feature of particular acts that fall under them (7). Call this the possibility view.

 

On the second view, the moral law stands in a direct, enabling, relation to particular acts, and the intermediate moral principles merely determine to which acts the moral law is to be related. Here the intermediate moral principles stand in a different relation to the particular acts and the moral law than the moral law stands to the particular acts. The moral law stands in an enabling relation to the particular acts, whereas the intermediate principles stand in what might be called an act-specifying relation. On this view, JD claims, there is only one possibility question. Questions about which action is required are not possibility questions (7). Call this the mixed view. JD thinks the possibility view is distinctive, whereas the mixed view looks more like the standard defence of moral principles (7).

 

In the book I made a number of claims that made it look as though I sometimes advocated a mixed view. But although I sometimes state that the intermediate moral principles have an act-specifying role, I did not mean to imply by this that they did not also have an enabling role, and I do not see why I cannot regard particular moral principles as playing both roles. The sort of view I had in mind was that moral necessity gets to particular acts from the moral law via more specific moral laws, and that the specific moral laws also determine which particular acts this necessity finds its way to. By itself the moral law explains the possibility of some act being required. More specific principles are needed to explain the possibility of some particular act’s being required in some specific set of circumstances. These intermediate principles are needed not only in their act-specifying role, but also because their having this role enables them to complete the explanation of the possibility of the imperatival character of the specific acts they specify. I am not sure that the act-specifying role commits me to the subjunctive conditional claim, but I do not see why it forces me to abandon the enabling role for these intermediary principles. As I see it, the act-specifying role of these principles is part of the enabling story.

 

The question is, though, whether I can defend the possibility view given the way in which I understand the intermediate principles. JD claims that because Rossian principles are not absolute in the traditional sense, they are not well suited to explain practical necessity.

 

‘The moral law is just the right sort of explanans, since it is, as one might put it, in the requiring business. But Rossian principles are not - and rightly not’ (8).

 

Given JD’s earlier criticism this is a surprising objection. For here he seems to be insisting that the imperatival character of required acts be explained by requiring principles, whereas he earlier objected to an explanation of this form as question-begging. Together these two points make it impossible for the problem of moral necessity to be solved. For, he claims, the only hope of solving it is with reference to a principle that itself necessitates, but he also objects that explaining moral necessity in this way merely relocates the problem at a higher level. If, therefore, the problem is solvable, one horn of the dilemma must be graspable.

 

I have tried to respond to the first by arguing that the experience of the categorical imperative is not a distinct phenomenon from the experience of the necessity of some particular act. Consequently, it does not stand in need of a separate explanation. I think one could also deny the second horn of the dilemma. I do not see why all of the conditions that explain the possibility of moral necessity must be necessitating principles. If ought implies can, then my ability to X enables X to be an act I ought to do. But although my ability makes this deontic feature of my act possible, ability is not itself deontic. Not everything that enables an ought must itself be an ought. It may also be that not everything that enables a must need be a must. As I understand him Kant’s aim is to explain moral necessity by the kind of universality that is embedded in universally valid principles, and Ross-style principles are as universally valid as more traditional (verdictive) deontic principles. The basic principles of prima facie duty specify which features give us moral reason to act in certain ways, and claims that these features give us these reasons in every possible case. The presence of such reasons does not imply that we have a moral requirement, for we may have moral reason not to do what we morally must do. So the reasons that fall under principles of prima facie duty are not in the requiring business in the sense that their presence implies a requirement. But they are in the requiring business in the very real sense that they are apt to give rise to moral requirements. It is for this reason that if undefeated they succeed in giving rise to a moral requirement.[3] Not all reasons are like this. The fact that something will be pleasant, interesting, or will enable me to live a good life, often gives me reason to pursue it, but is not such that if undefeated it will give rise to a moral must (or even a non-moral must). There is then something about moral reasons, the sort of reasons picked out by prima facie duties, that is different from other reasons. And this something seems to be that they are in the requiring business in this weaker sense.

 

How is it that these reasons have the distinctive feature of being apt to require? It is that they, and not others, fall under principles of prima facie duty that are strictly universal, and that inherit this strict universality by falling under the mere form of universality as such. It is through such principles, I argue, that the imperatival character of the moral law finds its way down to some specific act.

 

Tom Hill’s Comments.

TH agrees with me that Kant’s position is not tenable unless significantly qualified and supplemented. He offers an interpretation of Kant which, he claims, is less radical than my own, and maintains that this interpretation deals with the objections I raise against a non-modified Kant.

 

I’m not sure that I should accept that his interpretation leaves more of Kant’s theory in place, as it still seems to hold on to the justificatory conception of the moral law. What it seems to leave in place is the standard understanding of the role of the moral law in its justificatory role. In doing this I think it fails to capture what is most distinctive in Kant’s view – the moral law in its transcendental role. I think that what is distinctive of Kant’s ethics is not a new answer to the old question, the question ‘Why should we do what morality requires?’. As I understand him Kant’s theory is an answer to the distinctive and novel question, ‘How is it possible for us to be morally required?’ But I do not want to labour this point.

 

TH begins by expressing some doubts about the argument and conclusions of Groundwork I. The important point for what is to follow is his claim that the Formula of Universal Law (FUL) cannot serve as a criterion of right acts by itself. This criterion must, he claims, be constructed from the three formulations Kant offers. This is important because, if correct it undermines one interpretation I consider and reject, according to which acting from duty is understood as acting from respect for the formula of universal law (FUL).

 

TH doesn’t tell us why FUL cannot serve as a criterion by itself, or what the criterion might look like if we include the other two. But despite the fact that the formula of humanity as an end in itself (FEI) and of the Kingdom of Ends (FKE) are more attractive principles, Kant does seem to think of FUL as the more reliable and fundamental principle. Furthermore, some at least of what he says implies that he thought that this formulation does not need supplementation by the others.

 

Nonetheless, I did not consider the position TH proposes, so it is worth considering whether a view which assumes this conception of the basic principle of morality can deal with the objections I raise. At the centre of TH’s interpretation of Kant is his attempt to understand a reason in terms of a rationale that includes the basic moral principle constructed out of the three formulations of the moral law. I shall, therefore, focus on this.

 

 

 

Rational Principles and good reasons for action.

TH notes that Kant does not talk of reasons to act (something that should strike as quite remarkable!), so we need to translate his language into the language of reasons. According to TH the rationale for a morally required action will include

 

(1) The basic moral principle (the Categorical Imperative, constructed out of the three formulations Kant offers in the Groundwork)

(2) Concrete particular facts pertaining to the specific situation in which the action is called for – perhaps the fact that I promised, or the fact that someone is in distress.

(3) A practical verdict – I ought to X.

 

TH states that

 

‘a complete account of the reason for doing X would be expressed in the whole rationale, and this, presumably, is the full explanation of the motivation of a good will when acting in a morally worthy way. Colloquially we might pick out aspects of the rationale as ‘the reason’, depending on what is salient in the context…however, these partial answers presuppose that the rest of the rationale can be filled in ‘ (8-9).

 

I think it would be helpful in understanding this claim to distinguish the reason and a description of this reason. The description need not be complete, and typically isn’t. It may merely pick out some aspect of the reason. The reason itself, however, must, I think, be complete. A reason could not be an aspect of itself, though a perfectly adequate (though incomplete) description could just pick out this aspect. In what follows when I talk of the reason for some action, I mean the complete reason. The same goes for the rationale and our description of it. We have, then, four notions in play:

            (1) The (complete) reason.

            (2) A description of the reason (that need not be complete).

            (3) The (complete) rationale

            (4) A description of the rationale (which need not be complete).

 

I do not wish to deny that the reason we have to act in certain ways may sometimes be complex, and that when it is we do not need a complete description of it in order to make ourselves understood. When we refer to the complex reason with a partial description of it, we may rely on numerous context-dependent assumptions that need not be made explicit. The same applies to the rationale. So I do not wish to argue with TH about that. What I would like to know is what precisely the relation is between the reason we have to X and the rationale for X.[4]

 

According to TH, the rationale for X is a practical syllogism that has some formulation of the moral law as its first premise, “that the agent ought to X” as its conclusion, and concrete facts, such as the fact that I have made a promise, or that he needs help, taking us from the first premise to this conclusion. 

 

What, then is the relation between this syllogism and the reason we have to X? TH seems to identify the (complete) reason with the (complete) rationale, though it is, he thinks, perfectly OK to pick out this reason with a partial description of it. This understanding of a reason is in conflict with my own, for I think that reasons, complete reasons, can be simple concrete particular facts, such as that he is in distress. TH’s view rules this out. For him, such concrete facts are only ever aspects of reasons.

 

But we can allow that there is an essential relation between the reason to X and the rationale for X without identifying the two, and many of the things TH says do not force him to identify the reason with the rationale. Much of what he says is compatible with a different view – one that is not in conflict with my own.

 

TH claim that the rationale for X is ‘the most complete explanation of the reason’ to X (8). This may be a true account of the relation between the rationale and the reason, but it does not force us to identify the two. Typically if one thing explains another then these two things are different, as when gravity explains the movement of the planets, or when someone’s motive explains their action. It may, therefore, be that the rationale for X tells us why some particular concrete consideration is a reason to X, where this consideration is the complete reason to X. and is not to be identified with the explanans.

 

Elsewhere TH says that the rationale is “the full explanation of the motive of a good will” (p.8). On the assumption that reasons (in the standard normative sense) are not motives, this is a different claim from his account of the rationale as that which explains the reason. But given that a good will will tend to be motivated to do what she should by the normative reason why she should so act, this claim can be seen as drawing out the consequences of the symmetry thesis. It need not be seen as proposing an alternative account of the relation of the rationale and the reason.

 

At one point TH states that my reason to X is a part of the rationale for X.[5]

This claim also need not be seen in conflict with the first account of the rationale as what explains the reason. The first account is an answer to the question: ‘Why is consideration C a reason to X?’ It answers this question with reference to the rationale for X in which C figures. This question is also answered by stating that my reason to X is a part of the rationale for X. According to this view, C is a reason to X because C can be fitted into (is a part of) the rationale for X. This answer may be seen as clarifying the first answer.

 

But the view that the reason for X must be able to be fitted in to the rationale for X falls a long way short of the view that the reason and the rationale are the same. Concrete particular facts, such as the fact that someone is in distress, or the fact that I have promised to X, do figure in the complete rationale for X so, according to my way of accommodating what TH says, these particular facts can be reasons. They need not be partial descriptions of more complex reasons, but may be complete reasons in themselves. When they are complete reasons to X, a complete description of them will not be a complete description of the rationale for X. But in such cases one does not need to give a full description of the rationale for X to give a full description of the reason to X, at least in many contexts.

 

In other contexts, one might have to provide more of the rationale than the reason itself, for if someone does not know the full rationale they may not grasp that C is a reason. But in describing the full rationale one is not attempting to provide a full description of the reason (that has already been given), but is helping the other person to see that this particular consideration is the reason to X.

 

How does this fit into my second-order interpretation and revision of Kant? My view is that the moral law is not itself a moral reason to act, but is the condition of the possibility of something else being a moral reason to act. This something else is some particular fact in the situation in which the agent finds herself. So this particular consideration would not be a moral reason (one that is apt to necessitate) unless it fell under a universal principle that in turn fell under the moral law. But the moral law and the intermediate principles are not themselves further reasons to act, or even parts of a complex reason to act. They are rather, things that must obtain in order for something else (the particular fact) to be a moral reason. This is quite consistent with my proposal for understanding the relation between the reason to X and the rationale for X.

 

TH seems to see things differently. He states that

 

‘They [the particular facts] are not free-standing and complete reasons by themselves, but they make sense as reasons under presupposed standards of rational choice that we typically take for granted’ (10)

 

My view is that particular facts sometimes are complete reasons in themselves. TH claim that they only make sense as reasons under presupposed standards of rational choice is not incompatible with my view. It may be that certain particular facts only strike us as reasons if we assume certain principles, but that doesn’t mean that we must think of these principles as part of our reason. One possibility I tried to highlight is that we can think of these principles simply as conditions of the possibility of these particular facts being the reasons they are. As such these principles would not be part of our reason for acting, though the reasons we have would only be reasons if these principles apply.

 

I still think the causal analogy can help us get a grip on this. Certain causal laws are not part of the reason why some event (B) happened; though it may be that another event (A) would only cause (B) to occur if this law is in place. The law is causally relevant, but not itself a cause or part of the cause.

 

What I am trying to draw attention to with reference to causation is something analogous in the normative realm. Consideration C may be a reason for me to X, and it may be that C would not be a reason for me to X unless principle P obtains. But that should not lead us too hastily to conclude that this principle is part of the reason for me to X any more than we should conclude that the causal law is part of the cause of an event.

 

One question that may remain is how we know which part of the rationale of some action is the reason, and which parts are merely conditions of this other part being a reason (without themselves being a part of the reason)? I do not have a quick answer to this question. But I do think it is important to be sensitive to the agent’s own conception of her reason. We may want to say that what she regards as a reason can only be a reason when embedded in a rationale that goes beyond it, but should resist the temptation automatically to enlarge the reason so that it must always fill up the same space as that taken by the complex rationale.

 

In general then I think we should be sensitive to the different ways in which the dependency of particular reasons on principles can be understood, and try to understand this dependency in ways that are faithful to the moral phenomenology. In my book I was contrasting my transcendental account of the role of the moral law with the justificatory account, according to which the moral law is thought of as a distinct reason for action. TH is right to point out that this is not the only alternative. One might think of the moral law as a part of the same reason of which some concrete fact is. This is surely the right way to go for someone who thinks of the moral law as in some way figuring in the justificatory story. But I am not one of those who think it should figure in the justificatory story. Rather, I think a distinctively Kantian approach is to understand the central role of the moral law as a necessary condition of the possibility of certain particular facts providing us with reasons of a certain sort.