Mackie (Ethics: inventing right and wrong, Penguin Books, London, 1977) famously argued for a moral error theory on the basis that objective moral values, if they existed, would be very queer entities. Unfortunately, his argument is very brief and it is not totally obvious from what he says exactly where the queerness of moral values is supposed to lie. In this paper I will firstly show why a typical interpretation of Mackie is problematic and secondly offer a new interpretation. I will argue that, whether or not we have reason to live in the morally correct way, what seems queer about moral properties is that there is a morally correct way in which to live in the first place. This interpretation makes sense of Mackie’s claim that theism might be able to solve the queerness problem; the notion of an objectively correct way to live may make sense if theism is true, but not otherwise.
Mackie (1977) famously argued for a moral error theory on the basis that objective moral values, if they existed, would be very queer entities. Unfortunately, his argument is very brief and it is not totally obvious from what he says exactly where the queerness of moral values is supposed to lie. Several philosophers have attempted to clarify where the queerness is supposed to lie, but in this paper I want firstly to show why a typical interpretation of Mackie is problematic and secondly to offer a better interpretation, which is more consistent with Mackie’s overall view.
I will begin by outlining Mackie’s original argument in Sect. 2. Then in Sect. 3, I will look at Olson’s (2014) interpretation of it. In Sect. 4, I will discuss Olson’s attempt to extend the argument to support not just moral error theory but also epistemic error theory. I will argue that Olson’s argument fails; however, thinking about why it fails helps to show where the apparent queerness might really lie. In Sect. 5, I will argue that if there are objective moral properties then there must be an objectively correct or prescribed way to live, and that it is this which, on the face of it, seems queer, irrespective of whether or not we have a normative reason to live in the correct or prescribed way. This interpretation makes sense of Mackie’s claim that theism might be able to solve the queerness problem; the notion of an objectively correct or prescribed way to live may be plausible if theism is true, but not otherwise.
Mackie’s Queerness Argument
Mackie’s moral error theory is encapsulated in the opening sentence of his 1977 book Ethics: inventing right and wrong: ‘There are no objective values’ (1977, p. 15). He gives two main arguments for this claim, which he calls the ‘argument from relativity’ and the ‘argument from queerness’ (pp. 36–42). I am interested here only in the argument from queerness, which Mackie sub-divides into two parts, ‘one metaphysical, the other epistemological’ (p. 38). Firstly, there is ‘the metaphysical peculiarity of the supposed objective values, in that they would have to be intrinsically action-guiding and motivating’; related to this is ‘the problem how such values could be consequential or supervenient upon natural features’ of the world (p. 49). Secondly, there is the epistemological problem that, in order to have moral knowledge (whether of the values themselves or of the supervenience relation), ‘we would need some special faculty of perception or intuition, utterly different from our ordinary ways of knowing everything else’ (p. 38).
Sturgeon (2006) thinks that Mackie rejects moral properties on the grounds that they ‘could not be natural’ (p. 93), in that they do not ‘fit into a scientifically based, naturalistic view of the world’ (p. 92). He says that Mackie ‘argues that a naturalistic view of the world has no room for facts of this sort’—that is, facts that are objectively prescriptive (p. 110). No doubt there is some truth in Sturgeon’s claim. Dreier agrees that at least part of Mackie’s argument is that moral properties are ‘causally isolated’ (2006, p. 249). However, if this is what the queerness argument amounts to then presumably all putative non-natural properties or facts, such as mathematical facts, would be queer; and indeed, in talking about ‘companions in guilt’, Mackie (1977, p. 39) does discuss the possibility that our knowledge of numbers might be just as difficult to explain as our knowledge of moral facts, and concludes that the only way to show that mathematics can escape the queerness problem is ‘to show how, on empiricist foundations, we can construct an account of the ideas and beliefs and knowledge that we have’ of mathematics—something he is confident we are able to do. As Dreier points out, ‘mathematical facts enter into scientific explanations all the time, so our reasons for believing in mathematical facts are reasonably secure’ (2006, p. 249).
Given that Mackie is here discussing our knowledge of mathematical facts, he is clearly talking here about the epistemological, rather than the metaphysical, aspect of the queerness argument—that is, the problem of how we can gain moral knowledge. There is, as Dreier points out (ibid) ‘another queer feature of moral properties, what Mackie calls their intrinsic action-guidingness or “to-be-doneness”’. This is the metaphysical problem; and this does not apply to mathematical facts, and has nothing to do with whether moral properties are understood to be natural or non-natural properties. It is not non-natural entities per se that are queer in this sense, but properties, whether natural or non-natural, which are objectively prescriptive.
Hence Hattiangadi (2007, pp. 49–50) is surely correct to argue that putative moral (or normative) properties or facts are not merely queer (in this sense) in comparison to natural properties or facts, but also in comparison with other putative non-natural properties or facts, ‘such as the facts of mathematics’. She points out that ‘if we compare mathematical facts with putative moral facts, we still find a substantive difference’, and goes on to explain that
even if we postulate objective, sui generis, mathematical facts, we are not postulating facts that are categorically prescriptive. In contrast, if we suppose that there are normative facts, we are postulating sui generis, objective facts that are also categorically prescriptive. Putative normative facts would be utterly unlike any other facts with which we are acquainted, including putative non-natural facts about mathematics. Thus, to postulate sui generis normative facts or properties is to postulate entities that are inherently queer.
It is this metaphysical aspect of the queerness argument, rather than the empirical aspect, with which I am concerned in this paper; and what it means to say that putative moral properties are queer in this sense has nothing to do with whether or not they fit in with a naturalistic view of the world but rather to do with the fact that such properties would be intrinsically action-guiding.
Mackie says the following about putative objective values (1977, p. 40):
Plato’s Forms give a dramatic picture of what objective values would have to be. The Form of the Good is such that knowledge of it provides the knower with both a direction and an overriding motive; something’s being good both tells the person who knows this to pursue it and makes him pursue it. An objective good would be sought by anyone who was acquainted with it, not because of any contingent fact that this person, or every person, is so constituted that he desires this end, but just because the end has to-be-pursuedness somehow built into it. Similarly, if there were objective principles of right and wrong, any wrong (possible) course of action would have not-to-be-doneness somehow built into it.
Olson (2014, p. 104), and others such as Brink (1984, pp. 113–114) and Joyce (2001, p. 17), further divide this argument into two distinct arguments: there is an argument about normative reasons and an argument about motivation. Such writers clearly interpret what Mackie says about ‘direction’ and ‘to-be-pursuedness’ in terms of normative reasons. Certainly Mackie does seem to distinguish between having a direction and having a motive, because he refers to ‘both a direction and an overriding motive’ (emphasis mine). However, whether Olson and others are correct to interpret having a direction as having a normative reason is something to which I shall return shortly.
Note that Mackie does not claim that moral beliefs must provide motive or direction, and with good reason, since error theorists agree with moral realists that we have moral beliefs, and there is no reason to suppose that realists, any more than error theorists, must insist that such beliefs provide either motivation or a direction.
He does, however, say that moral knowledge provides both a direction and a motive, and he also says that an objective good would be sought by anybody acquainted with it, clearly implying that such acquaintance would be intrinsically motivating. If moral belief does not provide motivation, then it is difficult to see how moral knowledge would do so. Suppose (ignoring Gettier-type problems) we were to define knowledge as justified true belief: presumably it cannot be the fact that the belief is justified which provides the motivation, since even an error theorist can accept that our moral beliefs can be justified, though false, but on the other hand it is difficult to see how the motivation could be provided by the fact that the belief is true. Given what Mackie says about being acquainted with an objective good, then, it is probably best to interpret him, as Olson does, as saying that it is knowledge by acquaintance which is motivating.Footnote 1 Moral acquaintance could be compared to acquaintance with a beautiful painting. Knowing (perhaps from the testimony of experts) that a particular painting is beautiful is different from perceiving its beauty for yourself; perceiving the beauty of a painting might captivate us in a way which mere knowledge of the fact that it is beautiful would not, and similarly, perceiving moral goodness might (according to Mackie) motivate us in a way that knowing that something is morally good would not.Footnote 2
As well as saying that moral knowledge provides a direction and a motive, Mackie also seems to imply that the mere fact that something is good, regardless of whether or not we know that it is good, provides a direction. This is implied in his claim that ‘if there were objective principles of right and wrong, any wrong (possible) course of action would have not-to-be-doneness somehow built into it’. Perhaps, then, the most plausible way of interpreting Mackie’s claims would be to say that moral values themselves would provide a direction and that acquaintance with these values would be intrinsically motivating.
Mackie bases his claims about motivation on Plato’s claim in the Republic that rulers, having been educated to be acquainted with the Form of the Good, can be entrusted with absolute power because they can be trusted to be motivated to pursue the good (1977, pp. 23–24). Of course, Plato could be mistaken; and it may be that what Mackie says about putative objective values (or knowledge of, or acquaintance with, them) providing a direction, or having prescriptive force,Footnote 3 is more plausible than what he says about them providing a motive.
This is certainly what Olson (2014, pp. 103–138), Garner (1990, pp. 142–143) and Joyce (2001, pp. 30–31) think. They all think that the real queerness of putative objective moral values is that such values (or knowledge of, or acquaintance with, them) would provide us with a direction, and they understand ‘provides a direction’ as Mackie’s way of saying ‘provides, or at least entails, a reason for acting’. In what follows, I will limit the discussion to the claim that objective values would provide a direction, rather than that they would provide a motive.
Why should the fact that moral values provide a direction make such values queer? What does it mean for a moral value to provide a ‘direction’, or for some end to have ‘to-be-pursuedness’ somehow built into it?
Olson suggests that we think of moral queerness in terms of ‘irreducible normativity’. According to Olson (2014, pp. 117–118, emphasis in original),
the best articulation [of the queerness argument] is that moral facts are queer in that they are or entail facts that count in favour of or require certain courses of behaviour, where the favouring relation is irreducibly normative. Remember that for a fact to be an irreducibly normative reason is for that fact to count in favour of some course of behaviour where the favouring relation is irreducibly normative. The fact that is the reason need not be irreducibly normative. Strictly speaking, then, it is the irreducibly normative favouring relation, or reason relation, that is queer.Footnote 4
In order to understand what it means to be irreducibly normative, it is worth looking first at what it means to be reducibly normative. A reason (or strictly speaking, a reason relation) is reducibly normative, according to Olson (pp. 120–122), if it is constitutively dependent upon either some desire, interest or end which an agent has, or some code, rule or norm of some role or activity which the agent is engaged in.
To say that a reason is constitutively dependent on a desire, for example, is to say that the reason can be understood in terms of its dependence on the desire. The fact that the coffee shop at the top of my road sells really good coffee is a reason for me to go to that coffee shop. This reason is not merely dependent on my desire to have a really good cup of coffee, but can also be understood in terms of its dependence on this desire, and is therefore constitutively dependent on this desire. A reason can also be non-constitutively dependent on a desire. Suppose that you also desire a really good cup of coffee. I may well have a reason to buy you a cup of coffee, which is dependent on your desire. Nevertheless, although this reason is dependent on your desire, it cannot be understood in terms of its dependence on your desire; it can, however, be understood in terms of its dependence on some desire that I have, such as my desire to give you whatever you want. It is therefore non-constitutively dependent on your desire and constitutively dependent on my desire.Footnote 5
Reducible normativity need not be reducible to desires; according to Olson, it can be reducible to codes, rules or norms. For example, a soldier has a reason to comply with a general’s order, irrespective of the soldier’s desires, ‘since complying with the orders of those of superior military rank is part of the role of being a soldier’ (p. 121). Alternatively, I might have a reason to perform some action because this action is prescribed by the rules of a board game that I am playing. If I am playing chess, then I have a reason to move a piece if I’ve touched it, because this is one of the rules of chess. My reason for moving the piece is that I touched it; what explains why this is a reason to move the piece is the rules of chess. The reason can be understood in terms of its dependence on the rules of chess and is therefore constitutively dependent on the rules of chess.
According to Olson, then, ‘to say that some behaviour is correct or incorrect according to some norm is to say something reducibly normative’ (p. 120); he continues: ‘Reducible reasons are reducible to facts about what promotes desire satisfaction, or to correctness norms that may or may not be conventional’ (p. 121).
What, then, is irreducible normativity? To say that some behaviour is correct or incorrect irrespective of any desire, interest or end that I might have, or any code, rule or norm is to say something irreducibly normative. Something is irreducibly normative if its normativity is not constitutively dependent upon desires, goals, interests, conventions, ‘engagement in rule-governed activities’ (p. 122), etc. For example, if it is morally wrong to eat meat then we would all have a reason not to eat meat, but this reason would not be constitutively dependent on any desire that we might have, or on any norm (pp. 118–119). Olson insists that moral reasons would be irreducibly normative, but that irreducibly normative reasons would be queer. Because irreducible normativity would be queer, we cannot have irreducible normativity, and hence we cannot have moral reasons, and hence we cannot have objective moral properties.
The argument for moral error theory could then go something like this:
Premise 1: If there are moral properties then there is irreducible normativity.
Premise 2: There is no irreducible normativity.
Conclusion: There are no moral properties.
I think this is a fair interpretation of Olson’s argument. Is it also a fair interpretation of Mackie’s argument? It certainly seems to be the common interpretation of Mackie’s argument regarding the prescriptivity of moral properties and the fact that they would provide us with a direction; and undoubtedly this is at least part of what Mackie thinks is queer about moral properties. This seems clear from the following passage (1982, p. 115, emphasis mine):
Objective wrongness, if there is such a thing, is intrinsically prescriptive or action-guiding, it in itself gives or constitutes reason for not doing the wrong action, and this holds also for some, if not all, other moral features. To say that they are intrinsically action-guiding is to say that the reasons that they give for doing or for not doing something are independent of that agent’s desires or purposes.
So it does look as if this interpretation is at least part of what Mackie means about having a ‘direction’.
However, even if Mackie’s talk of having a ‘direction’ includes the idea of having a ‘normative reason’, there are, I think, problems with interpreting ‘providing a direction’ solely in terms of irreducible normativity, where this refers to reasons which are not reducible to some desire, interest or end or some code, rule or norm. In what follows I will discuss two reasons why we should not interpret Mackie in this way. Firstly, as we will see in the next section, Olson argues that putative epistemic reasons would also be irreducible in this sense, and although Olson concludes that we should extend the queerness argument to the epistemic domain, this is something which Mackie would be unlikely to want to do. But if we are to accept epistemic reasons then there must be reasons which are not reducible to desires, interests, norms, etc. (though they might still be reducible to something else). Hence this cannot be where the queerness lies.
Secondly, there is an aspect of Mackie’s discussion which has been overlooked in the literature and which, I believe, throws some light on where exactly the queerness of putative moral properties might lie. This is his discussion of how theism, if true, might resolve the problem of queerness. If ‘providing a direction’ is understood simply in terms of ‘providing a normative reason’ then it is not all that clear how theism would resolve the queerness problem. I will argue in Sect. 5, however, that there is another way in which we could interpret ‘providing a direction’, such that theism might indeed be able to resolve the queerness problem; and so Mackie’s claim that theism might resolve the queerness problem suggests that this interpretation may well be (at least partially) what Mackie had in mind.
Moral Error Theory and Epistemic Error Theory
If the queerness of moral properties lies in the fact that they provide us with irreducibly normative reasons then a moral realist might point to a ‘companion in guilt’. Cuneo (2007, Chapters 3 and 4), for example, has argued that if the queerness argument is successful against moral realism then it must also be successful against epistemic realism. One might have thought that epistemic facts, no less than moral facts, entail irreducible normativity; after all, if I believe proposition ‘p’ and I also believe that ‘if p then q’ then it seems to make sense to say that I have reason to believe proposition ‘q’, but it is not obvious that this reason is reducible to any desire that I might have, or to any norm, etc. It is difficult, therefore, to see how we can reject irreducible normativity without abandoning epistemic facts.
Olson (2014, pp. 155–172) agrees that epistemic facts would be analogous to moral facts in the way that Cuneo suggests, but his solution is to embrace not just moral error theory but also epistemic error theory. So let’s have a look in a bit more detail at the alleged analogy between the moral domain and the epistemic domain. Although I think that Olson’s argument in favour of embracing both moral and epistemic error theory fails, looking at his argument will help us to see more clearly precisely where the apparent queerness of objective moral values really lies.
If we have reason to believe proposition ‘q’ on the basis of believing the propositions ‘p’ and ‘if p then q’, then this reason looks like an irreducible reason. Olson agrees that epistemic reasons would be irreducibly normative, but his solution is simple: he denies that we have epistemic reasons. There is, he agrees, a sense in which I could have a reducible reason to believe proposition ‘q’ on the basis of believing the propositions ‘p’ and ‘if p then q’, just as there is a sense in which I could have a reducible reason not to move my rook diagonally: if I am playing chess then I can have a reason not to move my rook diagonally which is reducible to the fact that moving my rook diagonally is contrary to the rules of chess, and similarly if I am engaged in reasoning then I can have a reason to believe proposition ‘q’ which is reducible to the fact that complying with modus ponens is demanded by the rules of correct reasoning. However, the rules of chess only apply to those who are engaged in playing chess, and similarly, Olson says, the ‘modus ponens rule is an example of a rule that tells agents what there is reason to do qua (occupying the role of) reasoners, or qua engaging in the activity of reasoning’ (p. 137). It is true that I cannot reason correctly unless I comply with modus ponens, but it is equally true that I cannot play chess correctly unless I comply with the rule which says I cannot move my rook diagonally. If Olson is right, then it seems to be only those who are playing chess who have any reason to comply with the rules of chess, and we need not have any reason to play chess; and similarly, it is only those who are engaged in reasoning who have any reason to comply with modus ponens, and we need not have any reason to engage in reasoning. We have no irreducibly normative reason to comply with either the rules of chess or the rules of reasoning.Footnote 6
Olson concludes, then, that error theory is true regarding not just moral reasons but also epistemic reasons. The concept of an epistemic reason is one of irreducible normativity, and since there is no irreducible normativity it follows that there are no epistemic reasons. Any reason we may have to believe what we take to be evidently true must be contingent on some desire, interest or end that we have, or on some code, rule or norm. Olson compares the two types of error theory as follows (p. 164):
Compare: Moral error theory holds that it is a non-normative claim that some act is an act of torture but that it is a normative claim that there are moral reasons not to torture. The latter kind of claim is never true. Epistemic error theory holds that it is a non-normative claim that a proposition, q, is evidence that a distinct proposition, p, is true but that it is a normative claim that q is an epistemic reason to believe p. The latter kind of claim is never true.
However, there is an important disanalogy between the moral case and the epistemic case as just outlined in the quote from Olson. In the epistemic case, p and q are both included in both the non-normative claim and the normative claim, and they are related to one another in a particular way in the non-normative claim such that one is evidence for the other; there is no analogous relationship in the moral case. The epistemic case could be rewritten as:
Even though the truth of q provides evidence for the truth of p, it is never the case that the truth of q provides an epistemic reason to believe p.
The closest analogy in the moral case would, I think, be something like:
Even though the fact that φ is an act of torture means that φ is morally wrong, it is never true that the fact that φ is an act of torture means that there is a moral reason to avoid φ-ing.
But this would not amount to a moral error theory, since it does not deny that torture is morally wrong; it only denies that there is a moral reason to avoid φ-ing. In order to be a moral error theorist, it is necessary to deny that any action is morally right or morally wrong.
Let’s look at it this way. Suppose that I have a standing desire to believe all propositions which are evidently true. If I know that q is true, and I know that q is evidence that p is true, then given my desire to believe all propositions which are evidently true, it looks as if I have a reason to believe that p is true. I presume that Olson would accept this: we are not now talking about an irreducible reason, but a reason which is reducible to my desire to believe all propositions which are evidently true. But now suppose that I have a standing desire to avoid all wrong actions. It does not follow, simply from the fact that some act is an act of torture, that I have a reason to avoid this act, even given my desire to avoid all wrong actions. If I have a reason to avoid an act of torture which is reducible to my desire to avoid all wrong actions, then this would imply that acts of torture are morally wrong—which would entail the existence of objective moral values, the very thing which is being denied.
Again, suppose we say that I have a reason to believe q on the basis of p because I am engaged in reasoning and believing q on the basis of p is demanded by the rules of correct reasoning; we could not legitimately say, by analogy, that I have reason not to torture somebody because I am engaged in living a morally correct lifestyle, and torturing people is prohibited by a morally correct lifestyle, because again this would presuppose that some actions are morally good or morally bad, the very thing which moral error theory denies.
This shows that Olson’s analogy between moral error theory and epistemic error theory is flawed. Moral error theory denies the existence of objective moral values, not merely the existence of moral reasons.
Of course, the proposition that there are no irreducible moral reasons would entail the proposition that there are no objective moral values on the assumption that objective moral values would entail irreducible moral reasons, as the following argument shows:
Premise 1: If there are objective moral values then there are irreducible moral reasons (that is, irreducible reasons to live in according with these values).
Premise 2: There are no irreducible moral reasons.
Conclusion: Therefore there are no objective moral values.
But some moral realists, such as Brink (1984, pp. 114–115), deny the first premise of this argument. Olson recognises this, but thinks that such realists are mistaken; Olson thinks that the first premise of this argument is a conceptual claim (2014, pp. 123–124).Footnote 7 But if Olson wants to insist that the first premise is a conceptual claim, then a moral realist might respond by insisting that the first premise of the following argument is also a conceptual claim:
Premise 1: If there are propositions whose truth is evident to us then there are irreducible epistemic reasons (that is, irreducible reasons to believe these propositions).
Premise 2: There are no irreducible epistemic reasons.
Conclusion: Therefore there are no propositions whose truth is evident to us.
Whether or not the first premise of this second argument is a conceptual claim, it certainly looks like a candidate for being one, and certainly some would say it is a conceptual claim. Heathwood (2009), for example, thinks that there would be something puzzling in the sentence ‘I see that it’s quite likely to be true that there is a table in front of me, but, still, I don’t think it’s reasonable for me to believe that there is a table in front of me’; such a sentence would ‘have an air of incoherence about it’, and if I were to make such an assertion, it would seem ‘to be grounds for thinking that I do not really understand what I am saying’ (p. 90).Footnote 8 What Heathwood is claiming is not merely that if it is evident to me that some proposition is true then I have reason to believe that proposition, but, in addition, that this is a conceptual claim.Footnote 9 Mackie himself would almost certainly have agreed that this is a conceptual claim; as Cuneo (2007, p. 122) points out, his belief in the power of reason and argument would have led him to reject epistemic error theory.
It certainly seems to me to be at least as incoherent to say ‘I have no reason to believe this proposition even though it is evidently true’ as to say ‘I have no reason to avoid this action even though it is morally wrong’, if not more so. However, my concern here is not to argue that either of these two statements is incoherent, but rather to argue that if Olson insists that the second statement is incoherent then he should accept that the first statement is also incoherent. But if the first statement is incoherent, then, given that some propositions are evidently true, we must have irreducible reasons, and therefore irreducible reasons cannot be queer. If, then, there is something queer about moral properties, the queerness cannot lie in the fact that they entail irreducible reasons; the queerness must lie elsewhere.
If, then, Mackie’s argument is best interpreted along the lines that moral properties are queer in that they entail irreducible normative reasons then Mackie’s argument seems to fail. But I think there is reason to think that when Mackie talks about moral properties providing us with a ‘direction’ he means more than that they provide us with irreducible reasons, even though he may have had this in mind too. In the next section, I will seek to show a better way of interpreting Mackie’s argument, indicating where the apparent queerness of moral properties might really lie.
The Real Queerness of Morality
Suppose it is the case that I have a reason to believe any proposition which, on the basis of the evidence before me, is probably true, and that this reason is not reducible to any desire, interest or end that I have, or any code, rule or norm. Could it be reducible to something else? Heathwood (2009) thinks that epistemic reasons are reducible to descriptive facts about the world, such as facts about probability.Footnote 10 He thinks it possible, in fact, that epistemic facts could be identical to descriptive facts (p. 85), and suggests an analytical equivalence, so that when we say that somebody has a reason to believe a particular proposition, all we mean is that, given that person’s evidence, that proposition is likely to be true (pp. 89–90; cf. Cuneo 2007, p. 121). If Heathwood is right then there seems nothing mysterious about epistemic reasons, given that there is nothing mysterious about the fact that some particular propositions are likely to be true. Could something analogous be the case in the moral domain?
As Heathwood points out (pp. 86–89), the Open Question Argument would seem to tell against such an analogy. Suppose that we were to say that my reason to help you is reducible to the fact that you desperately need help from me which I can give at no cost to myself; we might even say, for example, that the proposition ‘you desperately need help from me which I can give at no cost to myself’ entails the proposition ‘I have a reason to help you’, and that somebody who says ‘I have no reason to help you even though you desperately need help from me which I can give at no cost to myself’ is conceptually confused. The problem is that this just does not seem to be the case. It may well be the case that your pain is a reason for me to help you, but somebody who insists that your pain is a reason to celebrate would strike me as (at best) morally confused rather than conceptually confused. Here, then, is a disanalogy between the moral case and the epistemic case, since one does appear to be conceptually confused if one insists that what they accept as evidence for p is a reason to believe not-p.
Even if moral reasons do not reduce to descriptive facts about the world, this does not mean that there is nothing that moral reasons could reduce to. If epistemic reasons are reducible to the fact that the world is a certain way, or that the evidence points to it being a certain way, then I think the most plausible analogy in the moral domain would be to say that if there are moral reasons then they are reducible to the fact that the world or human nature should be or is meant to be a certain way. Things are such that if I go out in the rain without an umbrella I will get wet; things are meant to be such that if you are in pain and I can alleviate your pain at no cost to myself then I will alleviate your pain.Footnote 11
But now we can see where the apparent queerness of moral properties really lies: it lies in the fact that the world generally, or human nature specifically, is meant to be a certain way. It makes perfect sense to say that the world or human nature is a certain way, but there does seem to be something prima facie queer in the thought that the world or human nature is meant to be one way rather than another.Footnote 12 This means that if epistemic reasons are reducible to the way the world is and moral reasons are reducible to the way the world or human nature is meant to be, then epistemic reasons would be perfectly respectable but moral reasons would seem queer. Pace Olson, the problem seems unique to the moral domain.Footnote 13
I have just spoken as if the fact that the world or human nature is meant to be a certain way could potentially provide us with reasons to act in the appropriate way; but in fact, if there is something queer in the fact that the world or human nature is meant to be one way rather than another, then this fact would be queer whether or not it provides us with reasons. In other words, it is not that this fact is queer because it provides us with reasons, but simply that this fact is queer whether or not it provides us with such reasons. Brink, for example, thinks that moral facts need not provide reason for action; according to Brink, ‘at least on standard theories of reasons for action, whether recognition of these facts provides reason for action will depend upon contingent (even if deep) facts about the agent’s desires or interests’ (1984, p. 114). He does not deny that moral facts are requirements, however, but suggests that one can deny that it is ‘a condition of the application of moral requirement that it provide the agent to whom it applies with a reason for action. While moral requirements apply to us independently of our antecedent desires and interests, they give us reasons for action conditional on their promoting our interests or desires’ (1992, p. 7).Footnote 14 What I am proposing is that it is the fact that the universe requires things of us—in other words, is prescriptive—which seems queer, whether we see this as providing us with normative reasons or not.
Mackie clearly does not think that prescriptivity, in and of itself, entails reasons for action. He suggests that prescriptive imperatives expressed as ‘must’ (as opposed to ‘ought’) statements need not be accompanied by normative reasons; if we ask for a reason why we must do what is prescribed, the answer ‘because I say so’ seems perfectly possible, and (although Mackie doesn’t say this) the answer ‘you just must’ seems equally possible (1977, p. 76). These would be (subjective) prescriptive statements, or directions, which make requirements on us and direct us towards some action, even though there need be no reason to comply.
Of course, he does not have moral requirements in mind here. However, in discussing moral non-cognitivism, he appears to accept the possibility of prescriptivity even in the moral domain where there is no guarantee that all agents will have normative reason to live in accordance with the prescriptions. According to prescriptivism, for example, moral terms express ‘prescriptions or recommendations’ (p. 32), so that the sentence ‘stealing is wrong’ means something along the lines of ‘do not steal’; but in prescribing that you must not steal, I need not, at least if I adopt a broadly Humean theory of normative reasons, believe that you have a normative reason to refrain from stealing. I presumably would not retract my statement if it became clear to me that you had no normative reason to refrain from stealing.Footnote 15 So Mackie appears to accept the possibility of moral sentences being prescriptive without providing normative reasons.
Nevertheless, Mackie does think that ‘ought’ statements are accompanied by reasons. He suggests that ‘a ought to G’ is more or less equivalent to ‘there is a reason for a’s G-ing’ or ‘a has a reason to G’ (1977, pp. 73–4, 77).Footnote 16 It also seems natural that if we think some action is morally wrong then we will generally think that one ought to refrain from doing it.Footnote 17 But even Brink acknowledges that those who deny that there is necessarily a reason to do what is morally required can concede that there is a sense in which there might be such a reason; in this sense, he suggests that ‘for there to be a reason for me to do something is simply for there to exist the relevant sort of behavioural standard or norm’. But what is denied is that there is a ‘reason to behave in accordance with such a norm such that failure to behave in that way is ceteris paribus or pro tanto irrational’ (1992, p. 8). It is their denial that there is anything necessarily irrational in acting immorally which lies behind the claim by both Foot (1972, pp. 309–310) and Williams (1981, p. 110) that moral requirements need not provide us with reasons.
Even though Mackie does appear to believe that objective moral facts would provide us with reasons, it is not clear whether he is suggesting that ‘a ought to G’ is equivalent to ‘it would be irrational for a to not G’, in which case immorality would entail irrationality. Whatever his views on this, it certainly doesn’t follow that those who deny that requirements entail reasons must (as Mackie assumes they must) be using ‘words like “right” and “wrong” with no prescriptive force—to say, for example, “x is right and y is wrong, but of course it is entirely up to you whether you prefer what is right to what is wrong”.’ (Mackie 1980, pp. 54–55). Moral facts would still be prescriptive because they would require things of us, but prescriptivity need not in itself entail normative reasons.
I maintain then that Mackie’s argument is best understood as claiming that if there is objective prescriptivity then the world or human nature is meant to be one way rather than another. It would be as if a mindless universe was capable of making demands on us, or determining what our nature is meant to be like. This, I think, is primarily what it means to say that moral properties provide us with ‘a direction’. They direct us to be one way rather than another. Even if Mackie thinks that moral requirements entail normative reasons, the queerness problem is not resolved by simply denying this and insisting that there can be moral requirements without normative reasons. Even if Brink is right, and we don’t necessarily have reason to comply with moral requirements, the fact that we are somehow required to be a certain way in the first place seems queer. This means that the question of normative reasons is largely irrelevant to the queerness problem and I can therefore remain neutral on whether or not moral facts provide us with reasons for action.
I think the way I have described the queerness problem is faithful to the gist of Mackie’s argument, and indeed Mackie’s suggestion that theism might be able to resolve the queerness problem suggests that this interpretation is the best one. If the queerness problem is that we are meant to be one way rather than another then the apparent queerness seems to disappear if we suppose that theism is true, since there doesn’t seem anything queer about God creating us such as to be one way rather than another. Although Mackie himself did not think that theism could be defended, he thought that if it could be then this could provide a vindication of objective moral values.
Why might we think that theism would make objective moral values less queer? Mackie’s answer seems to be that, whilst it does not make sense to think of a mindless universe making demands on us, there is nothing queer about the universe making demands on us if there is a mind behind it. In The miracle of theism, Mackie suggests that there might be something to be said for a version of the moral argument for God’s existence, precisely because ‘intrinsically prescriptive objective values’ would be queer in the absence of God (1982, p. 115). He then explains why the existence of God might explain away the apparent queerness of such values, as follows (pp. 116–117):
We can understand a human thinker, either as an agent or as a critic, seeing things as to be done or not to be done, where this is a reflection or projection of his own purposiveness; hence if we are to explain an intrinsic to-be-done-ness or not-to-be-done-ness, which is not such a reflection or projection, it is natural to take this as an injection into reality made by a universal spirit, that is, something that has some analogue of human purposiveness.
The idea seems to be that there is something queer about an impersonal universe making demands on us but nothing queer about God, if he exists, making demands on us. As Garner (1990, p. 143) says, moral facts ‘are unusual in an unusual way—they demand’. Although Garner suggests in a footnote that it would be a mistake to think that theism could solve the queerness problem, he nevertheless continues: ‘It is hard to believe in objective prescriptivity because it is hard to make sense of a demand without a demander…. We know what it is for our friends, our job, and our projects to make demands on us, but we do not know what it is for reality to do so.’
In Ethics (1977, p. 45), Mackie acknowledges that his argument is not dissimilar to Anscombe’s argument that moral duty and obligation make little sense outside a divine law concept of ethics (Anscombe 1958). However, Mackie goes further than Anscombe. Anscombe thought that we should adopt an Aristotelian concept of ethics and abandon talk of moral obligation in favour of virtue ethics, but Mackie thinks that even this would be problematic without recourse to God. This is because such an Aristotelian concept of ethics seems to imply a goal or purpose for human life. But as Mackie points out (1977, pp. 46–47), the idea that human life has a purpose or goal can be understood in two ways: we could say that human beings, as a matter of fact, strive towards some goal, which is a descriptive statement, describing what human beings happen to be like but saying nothing about why it should be virtuous to pursue this purpose or goal; or we could say that there is some proper or right purpose or goal, which is a prescriptive statement which says that we ought to be striving towards this goal.Footnote 18 But if we take this second interpretation, so long as ‘ought’ is understood in an objective rather than subjective sense, then the queerness problem seems to arise again. Hence he concludes (1977, p. 48, italics his) that
To meet these difficulties, the objectivist may have recourse to the purposes of God: the true purpose of human life is fixed by what God intended (or, intends) men to do and to be. Actual human strivings and satisfactions have some relation to this true end because God made men for this end and made them such as to pursue it—but only some relation, because of the inevitable imperfection of created beings.
In Chapter 10 of Ethics, Mackie elaborates on this. There is nothing mysterious or queer in acknowledging that there may be a way of life which enables us to develop our natural capacities, or to find deepest satisfaction, but this is a purely descriptive fact about the world, and there is nothing objectively prescriptive about this; the fact that my life may go better for me if I live a certain way or cultivate certain character traits does not seem to entail any requirement for me to live this way or cultivate these character traits. However, he continues (pp. 230–231),
But, further, God might require men to live in this appropriate way, and might enjoin obedience to the related rules. This would add an objectively prescriptive element to what otherwise were hard, descriptive, truths, but in a quite non-mysterious way: these would be literally commands issued by an identifiable authority. ….
… The descriptive component of moral distinctions is logically independent of God’s will: God approves of this way of life because it is, in a purely descriptive sense, appropriate for men. But the prescriptive component of these distinctions is constituted by God’s will.
But why should commands issued by God be ‘objectively prescriptive’ in the appropriate sense? After all, Mackie himself points out that somebody who ‘uses the concept of objective moral value will suppose that there are requirements which simply are there, in the nature of things, without being the requirements of any person or body of persons, even God’ (1977, p. 59, emphasis mine). Does this sentence contradict what he says elsewhere in the same book and also in The miracle of theism?
I think that there are two things that could be said in response to this. Firstly, in this passage he is discussing a number of possibilities as to what we generally mean when we use moral language, and he refers to this as ‘a further possibility’—albeit one which, he thinks, ‘ethical uses are particularly likely to exemplify’. But even if he thinks this is in the region of what we mean in our pre-philosophical moral usage, it is not clear that he is insisting that this is precisely what we have to mean, and it certainly seems plausible to suppose that our (pre-philosophical) concept of objectivity might allow for objective moral values to derive from the requirements of God. Such requirements would still be objective in the sense that they would constitute some external standard against which our own subjective human values (whether the values of individuals, cultures or the whole human race) could, at least in principle, be measured. All human moral judgments would then be objectively right or wrong in accordance with whether they coincide with this objective standard.
But secondly, and more importantly, I do not think that Mackie is proposing that God’s requirements, however arbitrary they might be, would be morally relevant in and of themselves. In fact, he argues that God can only make an action obligatory by commanding it so long as we already have an obligation to obey his commands (1982, pp. 114–115). If moral properties were constituted solely by arbitrary commands of God, then moral properties would presumably not supervene on natural properties, which is something which Mackie thinks is an essential component of our concept of objective morality.
What is needed then, according to Mackie, is for God to create the supervenience relation between natural and moral properties. This ‘creation of supervenient value’, he says, is ‘quite different from the creation of obligation by command which has been rejected with good reason by Plato and his many followers’ (ibid., p. 115). Presumably Mackie’s idea is that the creation of the supervenience relation would ensure that an act is wrong directly in virtue of the natural features of the act, rather than in virtue of God’s command; the action would be wrong only indirectly in virtue of God’s purposes, since the supervenience relation would be ‘an injection into reality’ by God on the basis of his purposes. Moral requirements would therefore be part of the fabric of the world, ‘which simply are there, in the nature of things’ (1977, p. 59), even though they would have been injected into the nature of things by God.
Would the fact that there are moral requirements injected into reality by God entail that we have normative reason to comply with them? Well, if God has not only fixed the purpose of human life, but has also created us such that ‘actual human strivings and satisfaction have some relation to this’ purpose, then it would seem likely that, as a general rule, we will have relevant desires and may therefore have normative reason which are reducible to these desires. But if, as Mackie adds, there is ‘only some relation, because of the inevitable imperfection of created beings’, then we may not always have the relevant desires, and therefore may not always have the relevant reducible normative reasons. Any reason to act in accordance with moral demands in such a case would therefore presumably remain irreducibly normative in Olson’s sense.
If this is the case, then it looks as if theism can help to explain how it is that we are meant to be one way rather than another, even if it doesn’t necessarily provide us with the relevant normative reasons which are reducible to our desires, etc. Hence, since Mackie is proposing that theism could resolve the queerness problem, this seems to confirm my interpretation as to where the apparent queerness really lies. It must lie in the fact that we are meant to be one way rather than another, because it is this which Mackie suggests can be explained by theism.
In this paper, I have aimed to show where the apparent queerness of moral properties might lie. Objective moral values seem queer, not so much because they entail reasons which are not contingent on our desires or interests, but rather because they imply that the world or human nature is meant to be one way rather than another; there is an objectively correct way for things to be. Although there would be nothing queer in God intending things to be one way rather than another, it would be queer, if there is no God, if things just are meant to be one way rather than another. The challenge for moral realists, unless they are also theists, is to explain how it could be that the world generally, or human nature in particular, is meant to be one way rather than another. In claiming that objective moral values would be queer, Mackie is not necessarily insisting that they would be incoherent, so perhaps moral realists can live with this queerness; but if not, then the only two options would seem to be either to reject moral realism or to embrace theism.
Olson suggests that we should ‘interpret Mackie as claiming that it is knowledge by acquaintance of moral facts that entail motivation. The reason is that in order to be affected by the motivational pull of a moral fact one has to be in some sort of direct causal contact with this fact. It is not enough to have mere second-hand knowledge about it, such as knowledge by testimony’ (2014, p. 110).
Of course, Mackie denies the existence of objective beauty. He is specific that the type of objective values he rejects includes not just moral values but also ‘aesthetic ones, beauty and various kinds of artistic merit’ (1977, p. 15). My point is simply that one can easily imagine somebody being moved by perceiving beauty (whether objective or subjective) for oneself in a way in which they might not be moved by mere knowledge that something is beautiful.
For a specific reference to moral values having prescriptive force, see, for example, Mackie’s reference to ‘objective ethical prescriptivity’ (p. 48).
Suppose that the reason why I should help somebody is that they need help. Strictly speaking, this reason is not queer; there is nothing queer about the fact that they need help. Instead, what is queer, according to Olson, is that the fact that they need help favours me helping them. In other words, it is not the reason itself which is queer; what is queer is the fact that it is a reason.
Given Olson’s view that, strictly speaking, it is the reason relation, rather than the reason itself, which can be reducible to a desire, code, etc., I assume that he would say that, whilst my desire may explain why something is a reason for me, my desire would not itself be part of the reason (cf. Schroeder 2007, pp. 23–40). Mackie, on the other hand, suggests that my desire would at least partially constitute my reason (1977, p. 66). I remain neutral in what follows as to which of these views is preferable.
Of course, anybody sincerely concerned about whether or not they have reason to comply with the rules of correct reasoning will presumably be engaging in reasoning, and will, in virtue of that fact, have reason to comply with these rules. Nevertheless, it seems possible to legitimately question whether a third party, who is not currently engaged in reasoning and isn’t bothered about having true (or probably true) beliefs, has reason to comply with the rules of correct reasoning. The possibility of asking this sort of question about a third party is apparently overlooked by Joyce in his discussion of the inescapability of practical reasons (2001, pp. 49–51).
Cf. Mackie’s claim (1980, pp. 54–55) that it would be ‘linguistically odd to use words like “right” and “wrong” with no prescriptive force—to say, for example, “x is right and y is wrong, but of course it is entirely up to you whether you prefer what is right to what is wrong”.’ In an endnote (on p. 158) he quotes Richard Price (himself quoting William Adams) as saying that ‘to perceive an action to be right, is to see a reason for doing it in the action itself, abstracted from all other considerations whatsoever’.
It is clear that Heathwood is equating ‘it is reasonable for me to believe …’ with ‘I have a reason to believe …’.
See also Cuneo 2007, p. 192, footnote 12: ‘Unless the context indicates otherwise, I will use the terms “evidence” and “(epistemic) reason” more or less interchangeably. This use is, in part, based on what I take to be the linguistic data. It would be odd to say, for example, that S has evidence to believe p, but no reason to believe p. Likewise, it would be odd to say that S knows or is justified in believing p, but has no reason to believe p.’ Cuneo’s reference to ‘the linguistic data’ suggests that he, like Heathwood, is claiming that it is a conceptual truth that if I have evidence that a proposition is true then I have reason to believe that proposition. Earlier, Cuneo says (plausibly, in my view), ‘A statement’s being offered as evidential support for a conclusion… is just a matter of its being offered as a reason for accepting that conclusion. And, when all goes well, premises are reasons to accept a conclusion’ (p. 121).
Note though that they would still be irreducible in Olson’s sense—that is, they would not be reducible to our desires, interests or ends, nor to any code, norm or rule.
I am not, of course, concerned here with idea that the world is ‘meant’ to be one way rather than another in the sense of expectation, as in ‘It is meant to be nice weather at the weekend’.
Could the analogy between the moral domain and the epistemic domain be reasserted here? Could it be that our nature is meant to be such that we believe those things the truth of which we have evidence for? Perhaps; but the point is that we don’t have to say this, since we have a perfectly good, non-queer, explanation of our reasons for believing what is evidently the case: namely, my reason to believe p just is the likelihood, given my evidence, that p is true.
Although he doesn’t say specifically here that this is his own view, it seems consistent with his view in (1984).
It is worth noting that Sinclair (2016) has developed an expressivist theory of moral reason statements, according to which, when I say that certain considerations provide you with a reason to act in a certain way, I am expressing approval of those considerations motivating you to act accordingly. This, however, as Sinclair acknowledges, would involve rejecting the Humean theory of reasons.
Foot (1972, p. 309) thinks this is an implausible claim.
It is true that Mackie thinks that ‘ought’ is a ‘relatively weak modal auxiliary’, and says that anybody ‘who really means business uses “must” or “shall” rather than “ought” (or “should”) in his moral pronouncements’ (1977, p. 64). However, his point is not that they need not be accompanied by reasons, but rather that the force of the demand, or the sense of urgency, is better captured by a ‘must’ statement than by an ‘ought’ statement. As he points out, ‘must’ statements are also sometimes more appropriate in the non-moral domain.
He makes a similar point in his discussion of Butler (Mackie 1980, pp. 38–39).
Anscombe, G.E.M. 1958. Modern moral philosophy. Philosophy 33 (124): 1–19.
Brink, D. 1984. Moral realism and the sceptical arguments from disagreement and queerness. Australasian Journal of Philosophy 62: 111–125.
Brink, D. 1992. A puzzle about the rational authority of morality. In Philosophical perspectives, 6: ethics, ed. J.E. Tomberlin, 1–26. Atascadero: Ridgeview Publishing Company.
Cuneo, T. 2007. The normative web: an argument for moral realism. Oxford: Oxford University Press.
Dreier, J. 2006. Moral relativism and moral nihilism. In The Oxford handbook of ethical theory, ed. D. Copp, 240–264. Oxford: Oxford University Press.
Foot, P. 1972. Morality as a system of hypothetical imperatives. The Philosophical Review 81 (3): 305–316.
Garner, R. 1990. On the genuine queerness of moral properties and facts. Australasian Journal of Philosophy 68 (2): 137–146.
Garrett, A. 2012. Joseph Butler’s moral philosophy. In The Stanford encyclopedia of philosophy (Winter 2014 Edition), ed. E. N. Zalta. Stanford: Stanford University. http://plato.stanford.edu/archives/win2014/entries/butler-moral/.
Hattiangadi, A. 2007. Oughts and thoughts: rule-following and the normativity of content. Oxford: Clarendon Press.
Heathwood, C. 2009. Moral and epistemic open-question arguments. Philosophical Books 50 (2): 83–98.
Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press.
Mackie, J.L. 1977. Ethics: inventing right and wrong. London: Penguin Books.
Mackie, J.L. 1980. Hume’s moral theory. London: Routledge.
Mackie, J.L. 1982. The miracle of theism: arguments for and against the existence of God. Oxford: Clarendon Press.
Olson, J. 2014. Moral error theory: history, critique, defence. Oxford: Oxford University Press.
Schroeder, M. 2007. Slaves of the passions. Oxford: Oxford University Press.
Sinclair, N. 2016. Reasons, inescapability and persuasion. Philosophical Studies 173: 2823–2844.
Sturgeon, N. 2006. Ethical naturalism. In The Oxford handbook of ethical theory, ed. D. Copp, 91–121. Oxford: Oxford University Press.
Williams, B. 1981. Internal and external reasons. Moral luck, 101–113. Cambridge: Cambridge University Press.
I would like to thank Andrew Fisher and Neil Sinclair, as well as members of the Philosophy Post-Graduate Seminar Group at the University of Nottingham and two anonymous reviewers, for helpful comments on earlier versions of this paper.
Conflict of interest
I declare that there is no conflict of interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.