Worrying at the Problem

Several (long) attempts to approach the issue:

1. After two and a half millennia, we are still without an answer to Thrasymachus, or at least so it seems. The ability to ask the question "Why should I be moral?" still exercises us to formulate an answeer. For it seems that any moral theory, to be adequate to ground the normativity of its judgments, must be such that it can give a motivating reason to everyone for acting in ways prescribed by the theory, on the grounds given by the theory. We could call this the motivation test (MT). The test merely sets a minimum level of justification for action that a moral theory must provide. The theory gives grounds of some sort justifying moral behavior. If someone understands these grounds, and if that person takes them to be reasons for action that do justify moral behavior, but do not take them as (at least some of) their own reasons for action, the theory fails the MT.

That, at least, is the sort of story one might tell; a Kantian moral theorist will surely like this idea, since the categorical imperative's own justification seems to contain a claim similar to the MT. A rational creature, as such, if it does understand the justification of the categorical imperative, cannot help but take some sort of universalizing principle as a reason to act in one way rather than another. The creature, understanding, thinks to itself I am a rational creature and This principle applies to the actions of rational creatures as such and then This principle applies to my (thinking about my) actions and I am doing something wrong if I fail to act in accordance with the principle. The rational creature may not thereby actually act in accordance with the moral imperative, but the creature, in understanding the theory, on a view like this, must thereby take it as justifying at least some reasons for action.

Some modern theorists working in a deontological vein have suggested that a Humean moral theory cannot meet such a test. One way to crudely characterize such criticism is to say that it takes the idea that moral utterances are expressive of attitudes toward (proposed) moral behaviors and shows that if such is the case no justification for any particular reasons will come about from a general moral theory since attitudes are merely subjective and without any necessary persuasive force. In his argument in the Treatise for why we necessarily need moral rules, Hume imagines a world full of human beings but in which there are no moral rules. He concludes that one could not have a coherent self-identity, since there could be no "relations" motivating prudential behavior, because the future self–like other people in such a world–must be connected to by the very relations that Hume says are necessary and necessarily for moral behavior in order to be cared about (in order for the future self to be me in a more than purely intellectual understanding), and the world without moral rules cannot have such relations in it. So, the claim goes, Hume's moral theory justifies acting in some ways rather than others, even after the theory has been explained, for those who understand the theory, because they must relate in ways that entail moral rules being taken to exist for themselves on pain of not being able to have a coherent self-identity. So they will say These are rules that apply to me and are reasons for me to act morally even if they do not actually end up acting on those reasons.

What these accounts suppose is that the MT can only be met on grounds that convince a "rational" being that the justification given by the moral theory actually gives that being reason to act in so far as it actually is a "rational" being. But what then of the immoralist, the egoist who holds no allegiance to moral rules as such? The egoist will claim that their actions and reasons are "rational," in some sense, and so if they do not take moral reasons to be any of their own reasons, they are going to provide thereby a counterexample to any such theory's meeting the MT. I want to agree that the presupposition that a theory's passing the test on pain of the being's "irrational" reasons leads to a failure to actually pass the MT, but I do not want to do so in such a way that allows the egoist off the hook for immoral behavior.

2. Propose that the test, as conceived, requires to narrow a view of what counts as having a reason to act, such that those to whom the question of whether or not to act morally applies can deny of themselves that they necessarily actually have reason to act morally even while understanding the theory justifying said reason. This means, basically, that the MT fails as a test of minimal adequacy for a proposed moral theory, because it gets wrong how moral reasons apply to agents. A first indication how moral reasons do apply is this: observe first the difference between saying "Barbara acted morally" and saying "That lion acted morally." It is not clear what we could mean by the second statement. Now suppose we have an intelligent Martian of whom, having flown to earth and taken up residence, we say "That Martian acted morally." I propose that this third statement is as difficult to understand as the second, because moral rules do no attach to us qua rational being, but rather qua human being.

Assume that there is no non-natural quality of "goodness" attaching to things or events. Goodness is then predicated of things or events (or states of affairs) relative to some concept supplying content in which we can cash out the term "good." Nowadays this is a commonplace; I do however want at the outset to reject any notions of subjectivity about goodness. Thus it is not my claim that utterances about the moral worthiness of some thing or event are solely (or even primarily) mere reports of some internal attitude towards them, which is not to say that some utterances of the same grammatical structure as those about moral worthiness are not used in exactly this manner.

I wish to go a step further in narrowing how moral evaluations function, however, in order to specify the sorts of utterances under discussion here. In distinguishing 'good' as an attributive adjective, I want to say that moral evaluations–where 'good' may be substituted as 'well' or 'fine' or even 'moral' (as I have done above), and is opposed to wrong or evil or any number of negative terms–can be cashed out to some other description without which the content of the 'goodness' is merely attitudinal. So, following P.T. Geach, to say that some x is good, there is some respect in which it is evaluated, whether it be as beachball or as dog or as money-gathering device or as moral [Good And Evil ?Ch. VII?]. The second description, in practice, is often implicit in conversational context. It shows up in statements like That's a good job! (for someone like you) and What a good day! (for flying a kit in high winds) and She sure is a good person! (to tell secrets to) and so forth.

So we would expect moral goodness to be cashed out in terms of morality; but since in morality we are concerned with goodness or badness it looks like we are merely required to assume some moral system, and then move forward. This is hardly conducive to maintaining any sort of objectivity in ethics. Now consider the statement, nowadays often heard with a tone of moral approval, "She's a good person." The terminology of virtue has long since passed out of fashion, but one can, I think, make out the echo of something like "She's a virtuous person" if such a statement as the previous one carries moral approbation. The short version of the claim, then, is that as 'virtue' and 'person' are still to be specified, we will find that they are finally cashed out in terms of the concept of a human being as such, not in terms of rational actors or pure practical reasoners or the like; in truth it seems to me that all moral "goodness" must be understood as goodness of a human being per se.

3. In order to get such a claim off the ground, we may need to return to the kind of theory (here focussing mainly on Kantian or neo-Kantian theories) that understands morality in terms of a priori principles for rational behavior as such. At first it looks as though the idea I have proposed and such theories are speaking past one another. After all, the theory put forward earlier discusses morality as necessarily containing some empirical element in order to be understood, while the Kantian theorist wants to say that we must abstract from all that is empirical in order to understand morality as such. What I want to say is that the Kantian theorist, in abstracting to pure practical reason, loses her grip on what 'moral' means, and the word becomes merely a label for the sort of behavior that (canonically) conforms to the categorical imperative. Or, more generally, what is going on is finding some standard of behavior based on a priori reasoning about that by which certain sorts of agents qua their type of practical reasoner are bound as such to be motivated by (or at least generates strong prima facie reasons for action).

Against such views, I think it is necessary to point out that, when abstracting from empirical elements usually present in understanding certain concepts, one runs the risk of talking about a concept not properly subsumed under the label one wants to apply to it. To discover the form of practical reasoning that constrains action to what is moral, the Kantian abstracts from all content having to do with actual reasoning about action that human beings do. Then–the claim seems to go–the next move is to understand any being capable of rational deliberation about action (capable, that is, of practical reason) can only be described as acting morally well when following the dictates of pure practical reason understood by the being as such. Such a view will then take it that the only reasons properly so called for action for a practical reasoner are reasons at least conforming to the a priori rules of reasoning about action. Such a view will contrast reasons for action with things like mere desires and whims.

We need not rehash all of the various problems found in these accounts, as a large industry of other philosophers is currently churning out new ways to find problems in them. What I would really like to point out is that the view demands that only that action which, conforming to the strictures of an a priori conception of practical reason, is found to be based on a universalizable reason for action can be called morally good. But it is simply not clear that what this view demands is anything like what morality is actually like. I am not trying to refute an a priori claim with an empirical one; I am trying to point out that what is supposed to be a priori true about a concept labeled with 'moral' is true (if it is) about something unintelligible as moral. On the view described, a morally good reason for action, with our understanding of how 'good' works attributively, is merely good as pure practical reason, which in turn is good as universalizable (but instrumental) reason for action. If this account is right (hardly likely to be granted, but let us move forward), what we see is that morally good reasons for action are reasons that are good qua universalizable (and of course on the Kantian view conforming to the other formulations of the categorical imperative as well) and which, as such, have not necessarily (though often accidentally) to do with reasons for doing good.

The last move, however, looks like a magic trick on my part: at one moment we are allowed a way to justify the normative force of ethical judgments about what to do, and at the next anything recognizable as morality has vanished in a puff of philosophical smoke. Where has it got to? It seems as though rationality dropped out, instead of the actual principle, and so we should not be surprised, if we look at how the trick is done, that the moral content also dropped out. Perhaps the description that allows moral content to remain under this conception of "moral" when giving evaluations of reasons for action as good or bad on a view like the one under discussion is that of something's being "rational." Certainly this looks promising; for if we can cash out moral goodness in terms of rationality, then we have a sort of non-subjective justification grounding a theory that makes a necessary claim on the reasons actually had for action by the beings to which it applies. Then we can meet the MT while keeping the content of the concept of morality in plain view.

4. Someone might object that it is not clear that this kind of account can really escape what has been called Prichard's Fork: the thought that a moral theory that tries to ground morality on what is not moral merely endorses actions that happen to be moral, or else assumes some moral framework that ends up being otiose to explanation of the normative force of morality. This looks like another, stronger way to set up the MT: demand that explanation not assume a prior normative framework, and then allow only theories that can persuade someone not having the intended framework to adopt it. Thus we return to the accounts given at the outset, of showing the attempts of two sorts of theories to explain why the normative force of morality really applies even to someone who is able to actively reject it (however spurious the grounds may be).

Try another tack. We ask someone for the general reasons that they have for acting. They will tell us that they do what they do in order to bring about some state of affairs, either in the performance of the action or in some separate thing it results in. The will tell us that they "want to be happy" or that they are trying "to do the right thing" or "get ahead" or "live a good life" or something similar. What all of these statements have behind them is a conception of some normative standard applying to the behaviors they undertake, and thereby to their practical reasoning. The egoist might say something about that their desires should be fulfilled, or that things in the world ought to be as they want them to be. And whether or not the conception for the egoist of how the world ought to be changes, the general reason for action–achieving some state of affairs, say–remains, as does the content of the normative standard–for instance bringing about whatever it is that one presently wants to be the case.

Imagine another, very odd sort of egoist, who is interested only in achieving their personal aims which have only accidental connection to any particular moral system of beliefs. This peculiar egoist desires to act only morally well (their payoff for this need not be specified–whether anyone is actually like this is questionable but the question I think need not be answered in order to make the point) according to, for the example, a set of religious rules. Their wanting is such that in order to achieve what they think ought to be the case they have only to act according to the rules. But the force behind their reason for action is assumed, and it is the same force as for all egoists, that is, satisfaction of their own wants (we are avoiding talk of desires strictly speaking) whether or not people like it. Only such things as they want to bring about provide the egoist with reasons for action–for them, no other facts or claims count as possible reasons to them. And the egoist will think it is surely quite odd to claim that something could be a (prima facie) reason for action for someone without its also being a reason to someone.

This (and the usual) egoist will claim that their practical deliberation is, in at least one sense, rational. He can tell a story about the things that can be taken as reasons, and about how one decides between them, and about what standards one is held to in deliberation of what to aim at and how to go about getting to the target. And the story is initially plausible at least as a psychological account, whether or not there are any actual people like this, since it is a time-honored philosophical problem. Our example egoist happens to have the aim of acting in accordance with some very old religious rules, and, keeping the aim fixed, everything else in their deliberative process seems to be working. But we want to object to what this person does, and the egoist generally, because they misconceive what the right sorts of reason for action are. The egoist will make this objection (especially to the deontologist): Why in the world would I want to be rational (when that means giving up these other reasons for acting I now have)? It is not immediately clear what can be said to such an objection.

The Humean scholar seems to have an answer here. As in the account given earlier, the one who is without the sorts of relations that entail moral rules (in some sense) to be in the world of that one is unable to properly relate to their future self, and so unable to have a conception of themselves as identical through time. The threat, of not being able to have a coherent understanding of oneself, is the lynchpin that convinces the egoist–or the knave–that they have at least a reason to act morally. Understanding the theory, they see that such relations are necessary for self identity, but are also the source of other-regarding relations which generate the attitudes necessary to motivate action (but not hydraulically). So the egoist recognizes that there ought to be some attitudes of the same sort as moral rules entail “in” their various reasons for action.

Let us suppose that the egoist, however, sees no such reasons. He is a free-rider. We are certainly going to regard this person as a sociopath, whatever else we may say, since they are entirely unmoved by any concern for others whatsoever. We will condemn them for this, and condemn them morally for it if we find them acting badly. But certainly this person does not suddenly lose a coherent sense of their personal identity. They know who they are and are able to project this into the future, and even recognizing that their desires and wants and reasons may change this does not prevent them from (successfully) identifying with their future self, or even from doing well by that future self (even the hedonist need not be so dissolute as to be unable to plan for the future).

Still, this egoist is certainly able to navigate moral rules well enough, and can even explain why other people have them (for they understand the theory justifying their normative power). How can we who feel the force of morality understand this? Obviously we want to condemn the individual for being irrational, for actually failing to understand the theory–either a deontological or Humean one. But such condemnation will not get us far; it is not fear that gives us reason to act well, and so we should not expect fear to do any better for the egoist; it is not clear that threats are in any case a thing that can motivate moral action as such.

5. I want to pose the problem this way: since (1) we cannot convince the egoist that they have reason to be moral without their demanding a reason to be rational–on the accounts discussed so far–and (2) it is not clear that any individual must actually take supposedly rational grounds for reasons to be moral (reasons to have certain reasons) to be convincing, perhaps we ought to consider framing reasons to be moral as reasons to do what the egoist was going to do anyway. I would suggest that we can make the argument along the lines of the idea that Aristotle takes to be a commonplace, namely the idea that it is everyone’s aim to be “happy” in some sense. However the notion so formulated is clearly inadequate for the work it seems to need to do. A more robust (but also a more vague) notion is that of “doing well.”

The very vagueness of the notion is suggestive. The person who aims at fulfilling the dictates of religious strictures is, in a sense, trying to do well. The person who aims at getting ahead is, in the same sense, trying to do well. The person who is aiming to satisfy their present desires is also trying to do well; they will characterize doing well as getting what you want, of course, but this does not, I think, take away from the fact that they can recognize themselves as having an aim like that of others. The psychological hedonist will attempt to pin his motives on others as well, but barring such projections we can still recognize doing well as a nonempty way of understanding the common thread in these disparate aims. What is gone after, as it were, is gone after as in some sense good (cf. Aquinas).

Suppose we buy that account of reasons for action. In order to get the claim off the ground, we have to assume that everyone who acts on reasons derives their reasons from some preexisting notion that carries not only motivational force but normative force (the normative force could be attached to any feature of the notion that allows choice between competing or contradictory reasons). That is to say, we have to claim that everyone thinks they ought to do well; but of course this notion is constituitive of the thought that what is gone after is gone after as in some sense good.

As mentioned earlier, however, it is not the case that we can leave “good” unspecified here. Otherwise we are going to be getting on only with a begged question dogging our heels. What I want to say is first, negatively, that no purely “rational” feature of human thought can fulfill the role demanded to specify what good is here. Second, more positively, I want to say that the notion of a life as lived by a human is enough to fulfill the role; but the specification of a human life will, unfortunately, have to remain vague because because it is based on an accidentally generated concept rather than some a priori understanding of practical reason.

About the first claim. As mentioned earlier, it is not clear that rationality in a deontological theory derived from Kant can really fit the bill in terms of providing a satisfactory specification of the sense of “good” we have in mind. Nor is it clear that the Humean theory can provide that specification either because rational understanding of why there are moral rules does not actually necessarily result in reasons for the knave to act (setting aside whether these reasons actually result in action, another murky area in Humean moral theory).

About the second claim. The notion of self-identity is assumed. Such self-identification includes the notion of what sort of thing one is–and answer to the thing’s question to itself (if it cares to ask) What am I? The answer will include referenced to participation in some kind(s); we would expect any adult human to self-identify this way, for it would be decidedly odd if someone thought of themselves as say, only N.N. and not as any particular kind of thing. Even if this were the case we would probably expect if we can ask What is N.N. like? that the answer will be a descriptive list: N.N. is like this and that but not the other, etc. and so on, and there is only one of N.N.

6. Return to the discussion of where moral evaluations attach. It would be wrong, I think, to attach moral evaluations to the lion. It would be wrong, in addition, to attach moral evaluations to small children; what they do is “wrong” in a sense, but they cannot be expected to actually conform to moral rules as such for they are still in the process of developing into moral beings. The lion is not a proper object of moral evaluations for two reasons: (1) it is not conscious of reasons for action and so cannot choose between them–it just does not have the cognitive ability to be a moral or immoral agent; (2) it is not human, so its behavior is not properly described in the same terms as human behavior–its inhumanity is also the efficient cause of its non-consciousnes vis-à-vis moral reasons–but in terms of lion behavior, so it should seem at least initially gripping that right or wrong behavior in lions is for them right or long qua lion behavior. Thus the Martian, even if its cognitive process is very similar to ours, must be “morally” evaluated in terms of Martian moral behavior, whether or not that accidentally happens to be similar to human behavioral standards.

Now it may seem that I am smuggling in the sorts of notions that some virtue ethicists take as stock-in-trade, that is, notions of the natural kind as the keeper of some standard of conformity or disconformity permitting of evaluation, which have to do with the achievement of the thing as what it is. In some theories this could be explicitly connected with a proposed telos for the (each, every) kind. But let us look at what this importation amounts to. On the deontological view, if the Martian is “rational” in the same manner as humans are, then its behavior will be indentically constrained by the unity of pure practical reason to conform to the categorical imperative under the same conditions as humans are constrained. On the Humean view, if relations between Martians and others function in the same manner as in humans then their need for morality and its source and justification in sympathy will be the same as for humans and otherwise not. It is only if the Martians are, in fact, like us in certain ways that they can be expected to be held to the same standards of behavior, whether these be “moral” or otherwise. (We do not expect the Martians to enjoy bread, milk, or sushi except accidentally as they are biologically made up.) So it seems perfectly plausible to say that if the Martians are like us in the relevant respect they can be held to standards of behavior that are just like those that humans call moral behavior, but in the one case this will be Martian moral behavior and in the other human moral behavior, and these are just coincidentally similar.

Given the picture sketched above regarding self-identity, is it so implausible to expect that an adult human will self-identify as a human being, or as a person, or as an agent with desires or interests? Not at all. Moreover it seems to me that any specification of these kinds will finally cash out in terms of what it is to be a human being. For if one indentifies a person and then allows that there is more than one sort of person, there will be a specification (in the absence of actual Martians being known about) of what sort of person that accords with the general description of what a human person is like. Even if there were actual Martians each of whom was accidentally (due to its Martianhood) a person in the same description as humans are, we would not be led astray. The specification will merely say that the kin of person acts as a category within which human persons fall.

Now here it looks like there is an opening for the Kantian to re-enter the fray.

One can think in such a way as to question whether to be rational, and reject thinking rationally. This decision if carried through results in actually not being rational. But a similar denial of humanhood fails because one cannot choose whether or not to be human; this is merely empirical and fixed. Whatever else your are (e.g. in denial) you will still actually be human. And not wanting to be (merely) human is, in fact, intelligible as a want that some people have (e.g. cult leaders) even though they cannot escape their essential nature. But if rationality fails to occur in rejecting rationality if one starts out rational, then thereby denying it accomplishes the point of the denial.
*******************

Questions: What is "rationality", especially of the rational sort, and is that really subject to the pure intellect; what is it to be a "human being" as opposed to a "person"--and does it matter; can we ever justify a claim of the existence of a necessary connection between a rational judgment of what ought to be done and actually acting (let alone discover it)?

Not light stuff, no Betsy, not so light at all.