The argument

Here is the argument reduced to three lines:

1. 1. Normative reasons exist.

2. 2. Normative reasons are the actual favourings a single mind is having about us (all of us who have ever and will ever exist) doing and believing things.

3. 3. So, a single mind that is favouring all of us doing and believing things exists.

Premise 1

Normative reasons are reasons to do and believe things. Nearly all of us get the impression they exist. It appears to me I have reason to go and get a drink because I am thirsty; it appears to me I have reason to believe there is a computer monitor in front of me because I am currently getting the strong visual impression of one; it appears to me there is reason to believe that inferences of the form ‘if P then Q; P; therefore Q’ are valid, and so on. So note: all arguments – including any one might deploy to call into question the existence of normative reasons - presuppose that at least some appearances of normative reasons are accurate.

What I have just said does not establish that premise 1 is definitely true. Normative reasons appear to exist, but appearances can be deceptive and so perhaps in reality they do not. The point, however, is that any attempt rationally to challenge the appearances – so any attempt to provide a justification for the belief that normative reasons do not exist – would have to appeal to some apparent reasons and so would be self-stultifying. Premise 1 may be false, but there is no reasonable basis for thinking it is.

Normative and metanormative theories

Before moving on to premise 2 it will be helpful to distinguish between a normative theory and a metanormative theory. A normative theory attempts to answer this question: ‘what do I have reason to do and/or believe?’ A metanormative theory attempts to answer this question: ‘what is a reason to do and/or believe something?’ So, ‘you have reason to do whatever will further your goals’ is a normative theory of reason. The view I am arguing for here is a metanormative theory. It is a theory about what reasons are made of rather than where they are located.

Here is another metanormative theory: that normative reasons are facts about what doing or believing something will achieve. This theory identifies reasons with facts. So, to the question ‘what is a reason?’ comes the answer ‘a certain sort of fact – a fact about what doing or believing something will achieve’.

There is already reason to think this kind of metanormative theory is false (shortly we will see there is an additional reason). We are justified in believing normative reasons exist because they appear to. Appearances may not match the facts. Thus it is possible there are no normative reasons in reality. Yet there will always be facts about what doing or believing something will achieve. So, if reasons are facts about what doing or believing something will achieve, then normative reasons definitely exist and holding otherwise would involve one in a contradiction (one would be saying it is a fact there are no facts). However, denying that normative reasons exist does not involve one in a contradiction. One cannot defend the claim that normative reasons do not exist, but it is not self-contradictory to assert it (they may not exist, after all). Thus normative reasons cannot be facts about what doing or believing something will achieve (or facts about anything else, for that matter).

The view that normative reasons are facts can appear plausible because it is easy to confuse it with a plausible normative theory. For instance, identify reasons to believe things with the fact that doing so will mean one has true beliefs, and it follows that we have reason to believe something if it is true. And that sounds correct: intuitively we do have reason to believe what is true. But the view that we have reason to believe what is true is a normative theory, not a metanormative theory.

Premise 2

Here is an argument for premise 2:

1. a. Normative reasons are favourings of us doing and believing things.

2. b. Favourings are essentially mental.

3. c. Therefore, normative reasons are the favourings some mind/s is/are having of us doing and believing things.

4. d. Normative reasons have an external single source.

5. e. Therefore, normative reasons are the favourings an external single mind is having of all of us doing and believing things.

6. f. Favouring relationships require the items being related to be actual.

7. g. Therefore, normative reasons are the favourings some actual, external, single mind is engaging in of all agents doing and believing things.

Regarding (a): it is a conceptual truth about normative reasons that they are favourings of us doing and believing things.

If that is correct, we have another reason to think normative reasons cannot be identified with facts. The fact the sun is up is just a fact about the sun's location. It is not a favouring of the sun being up or a favouring of believing the sun is up. The fact that doing X will fulfil your goals is just a fact about what doing X will achieve. It is not a favouring of you Xing. So facts lack normativity.

True, we sometimes say things like ‘that it will serve your ends is a reason for you to do it’. That sounds as if we are saying the fact it will serve your ends is one and the same as the reason to do it. But that's hardly compelling evidence. We also say ‘the ice cream is delicious’. We do not mean the ice cream and deliciousness are one and the same. We mean the ice cream has deliciousness as a feature. Likewise, when we say that the fact something will serve your ends is a reason for you do to it, we do not mean that the fact and the reason are one and the same. We mean the fact is giving rise to there being a reason.

Moving on to (b). This seems like a conceptual truth as well. I favour things. I can command, advise, recommend, commend and so on. I favour eating a biscuit right now. I favour you getting me one. I command you to get me one. By contrast, a biscuit cannot favour me eating it. An open door cannot recommend I close it. A chair cannot bid me sit in it. Favourings are attitudes, and attitudes are essentially mental, that is to say they can exist in minds and nowhere else. The idea of an extra-mental favouring seems, well, unintelligible. Thus, we arrive at (c): reasons must be made of the favourings some mind or minds is/are engaging in of our acts and beliefs.

It might be objected that I am taking things too literally. Perhaps talk of favouring should be taken metaphorically. Philosophers become poets when they talk about normative reasons, it seems. Normative reasons are like favourings, or favouring-esque, but not actual favourings.

I think that is false. And even if it is not false, something ‘favouring-like’ is going to have to be mental as well. Why do I think it is false? Well, because in the literature everyone uses terms like ‘favour’, ‘commend’, ‘advise’, ‘direct’ to capture what they mean by normativity. Why shouldn't they be taken literally? Personally I do not begin to understand how something could be like a favouring, without actually being one, just as I do not know how something could be like a sound without being a sound. Anyway, if terms like ‘favour’ are metaphors, let the philosophers who use them stop it and say what they really mean instead.

Even if I am mistaken and we should understand normative reasons to be favouring-like, the simple fact is only minds can favour or do anything remotely like favour. Only minds can favour or do anything remotely like favour. The idea of an extra-mental favouring-type thing is as unintelligible as an extra-mental favouring. So, I think there is no reasonable way of resisting (c).

Obviously, if you can make sense of and tolerate the idea of extra-mental favourings then you have a way out. But I do not think you are very reasonable if you take it. For now you have avoided my conclusion at the cost of positing a wholly new type of thing or relation: an extra mental favouring-type thing. My argument concludes with a mind. Granted, it is quite a strange mind. But it is only strange in the way that, say, the mind of a psychopath is a strange version of something we're already familiar with. It is a strange version of the same kind of thing. Whereas, an extra-mental favouring is strange in a different way; it is a new kind of thing. So, even if such things are intelligible (and that is a big if) positing such things to avoid my conclusion is unreasonable.

(c) leaves open the possibility that normative reasons are our own favourings of acts and beliefs. However, the evidence is that they are not. For instance, if I favour you doing something does it start to appear that you have reason to do it? Conduct the experiment. Favour someone doing something and see if that person subsequently gets the distinct impression that he has reason to do it. I bet he won't. So his – and most people's – impressions of reasons to do and believe things do not seem to be impressions of what we favour each other (and ourselves) doing and believing.

Plus, how plausible is it that we are wondering what we favour ourselves (or others) doing or believing when we wonder what we (or others) have reason to do or believe? That does not seem to be what I am doing. I seem to be trying to make myself sensitive to something outside, to something not me. Underscoring this is the fact that I sometimes favour myself doing something yet it appears reason favours me doing otherwise. Moral reasons are like this. They are, to use the jargon, ‘categorical’. That is to say, they appear to apply to us irrespective of our desires. For instance, it appears to me that I have reason not to hurt others for fun even if I really want to. Similarly, reasons to believe things appear to be categorical as well. It is a fact that the world is roughly spherical and this fact seems to give rise to there being a reason to believe that the world is roughly spherical irrespective of whether I want to believe it. You cannot have reason to do or believe something irrespective of your attitudes if reasons are some of your attitudes. So the appearance of categorical normative reasons is evidence that reasons are not composed of our favourings.

That reasons have a single external source is supported by the fact that if I say ‘Tim has reason to X’ and you say ‘Tim has reason to X’ we are clearly agreeing with each other. Yet if reasons are our own favourings then my claim that ‘Tim has reason to X’ is true just if I favour Tim Xing, and your claim that ‘Tim has reason to X’ is true just if you favour Tim Xing. And that means that we're not agreeing, but talking past each other. This point can be made about what all reasons; reasons anyone at any time has had (or will have). The only way that could make sense is if normative reasons (all the normative reasons there have ever been) have a single source across us all. Thus, it appears we have reason to believe that reasons have a single source.

(d) appears true and (e) follows from (c) and (d). We can now move on to (f). A real favouring requires an actual connection between the items being related. In other words there needs to be some actual favouring going on. Given that what is being related are on the one side facts about the world, appearances, and our desires, and on the other side acts and beliefs, it would seem the god (the single mind that is doing the favouring) must be omniscient (know everything that can be known) and omnipercipient (see what things are like from each of our perspectives and be aware of the contents of our minds). Note: she is not omnipotent. For nothing suggests she is able to guarantee that we believe what she favours us believing, or do what she favours us doing. Nor does anything suggest she is able to guarantee that our impressions of what she favours us doing and believing are accurate. Indeed, it is often unclear what we have reason to do or believe and so we seem to have reason to believe her powers, though great, are severely limited (otherwise she would presumably have made sure we do and believe what she wants us to do and believe). The evidence, then, suggests a powerful cosmic mind, but not an omnipotent one.

As an alternative to a cosmic mind some philosophers propose that normative reasons could be the favourings of a hypothetical mind: a fully informed version of our own. They also propose that this idealized mind would be the same for all of us, and thus reasons would have a single source. So this idealized mind is, in effect, a hypothetical version of the actual mind I am arguing for.

Such views – known as ideal observer views – cannot be true. Not only is it highly dubious that a fully informed ‘me’ would be identical to a fully informed ‘you’, more importantly a hypothetical version of me (or you) does not actually exist and so can no more really favour me doing or believing something than he can write me a letter or make me a cup of tea. The favourings of imaginary versions of us are imaginary favourings. Normative reasons are real favourings.

(g) just follows from (e) and (f). (g) is premise 2 of the master argument. Premise 1 of the master argument is true beyond any reasonable doubt. Premise 2 seems somewhere close, for we have arrived at it via an argument all the premises of which are either conceptual truths or overwhelmingly well supported by the evidence. 3 follows from 1 and 2. It seems we have a strong rational basis for believing in a god, then.

Objections

It might be objected that if normative reasons are just the favourings a single, omniscient, omnipercipient mind is having of us doing and believing things, then such favourings have no special authority over us. Why should we do and believe as she favours? This criticism can be expressed as a question: ‘what are this curious god's favourings to me?’

In reply: the critic is either asking to be given a reason to do this god's biddings, or he is asking to be made to care about reason's biddings. If the former then the critic is assuming that reasons are not one and the same as this god's biddings. That is to just assume my view is false rather than demonstrate it to be. The critic is just refusing to acknowledge the arguments given in its favour.

If the critic is asking to be made to care about reason's biddings then this does not even qualify as a criticism. It is what it is: a request to be made to care about reason's biddings. And frankly there is no point trying to get such a person to recognise that he has reason to care about reason's biddings (no point whatever analysis one gives of such things). He is not disputing that there are such biddings, he is asking to be made to care about them. Finding more of them and pointing them out to him will not address his request.

It might be objected that there is no important difference between this god's biddings and the biddings of, say, my friend John.

False. John's favourings are incapable of rendering accurate the appearances of external favourings of the acts and beliefs of all of us who have ever existed. John is not omniscient, not omnipercipient, is not favouring the acts and beliefs of all agents. He is not in any way responsible for us getting the impressions of such things. If he is who he appears to be, his favourings are clearly incapable of constituting reasons.

It might be objected that my view must be false for nothing seems to prevent the god-like mind from coming to favour us doing terrible things, like hurting each other for fun. So even the most atrocious deeds could be right if the god started favouring us doing such things.

I would first want my critic to clarify something. It appears strongly to me that I am sat in a chair in front of a computer. It appears strongly to me that I (a mind) exist. And it appears strongly to me that I have reason not to hurt others for fun. My question is: is it possible for a really strong appearance to exist absent its object of awareness?

Assume the answer is ‘no’ (and I am not saying it is). Then by hypothesis it is not possible for there to exist a strong appearance of a reason not to hurt others for fun without there actually being a reason not to hurt others for fun corresponding to it. Thus, given what has been argued, it is not possible for there to exist a strong appearance of a reason not to hurt others for fun unless the god in 3 exists and is disfavouring us doing so.

The god's favourings could change, of course. But by hypothesis if the god came to favour us hurting others for fun then things would appear very differently to how they appear now. They would have to. If the god came to favour us hurting others for fun, it would stop appearing as if there were reasons not to hurt others for fun. It cannot be a fault in a theory that it respects how things appear. So if the answer is ‘no’ my view is unobjectionable in this respect.

What if the answer is ‘yes’? Well, now the critic accepts that it is possible for even very strong appearances, including the strong appearance that there is reason not to hurt others for fun, to be inaccurate in reality. The critic cannot consistently hold it against my view that it allows for such a possibility. And note: it is the mere possibility that my view allows. There is no implication that hurting others for fun is actually something we have reason to do (quite the reverse: it appears something we have categorical reason not to do, and there is no reason to doubt those appearances). So if the answer is ‘yes’, there seems nothing to object to either.

Perhaps the problem is that my view allows that although it is wrong today to hit Marjorie in the face, tomorrow it might be ok despite nothing else having changed.

First, it is worth reiterating that if normative reasons are the favourings of a god this is entirely compatible with it having been the case that we have always had reason to do and believe the same things in the same circumstances for all time (and that this will remain the case going into the future). My view simply opens up the possibility that the rational landscape might alter, it does not imply that it actually does. Nevertheless, some will think this is enough to justify the absurdity charge.

However, the view that the rational landscape cannot alter is pure dogma. There is no evidence to support it. Indeed, if anything, the evidence is in the other direction.

For instance, engage in any moral debate and it will not be long before someone makes the following (normally irrelevant) observation: at different times people have believed different things to be right and wrong. That is to say, at different times people have perceived there to be reasons to do different things in otherwise similar circumstances. If that is true (I am not saying it is) then on the assumption that our faculty of reason is fairly reliably tracking the reasons that there are (and that there is no reason to think the faculty of reason of those in the past was less reliable than ours is today) we have prima facie evidence the rational landscape has altered, don't we?

A critic might say that my view makes it ‘arbitrary’ what we have reason to do and believe. This is just to insist that my view makes it problematically variable what we have reason to do and believe. As such this criticism is not distinct from those already considered.

Another objection is to suggest that my view generates a regress. For surely the god must have reasons to believe and do things? Wouldn't that require that there be another god and another and so on?

No. Reasons are essentially favourings. They are not essentially external. They just appear that way to us. So, the accuracy condition of our appearances of normative reasons is a single, external favourer who is favouring us doing and believing things. However, from the god's perspective there are just things she favours us doing and believing. These are favourings that have a single source across all agents: her. So, there is no regress, so far as I can see. She is the accuracy condition of our appearances of normative reasons. She is also the accuracy condition of her own. It is only if you mistakenly think my argument assumes that normative reasons are essentially external that you get a regress. They are not. I have argued that normative reasons are external favourings because they appear external to us, rather than because they must be external.

A different objection is that given the astounding and/or unwanted nature of my conclusion it is always going to be reasonable to reject premise 2.

In reply: it is not at all reasonable to reject a premise of an argument on the basis of the astounding/unwanted nature of its conclusion. The truth may be astounding/unwanted. You should not assume it won't be.

Ockham's razor might be invoked: surely it is better to avoid postulating such an exotic creature if we can? True, but Ockham's razor enjoins us not to multiply entities beyond what is necessary. The argument I presented appears to demonstrate that it is necessary to posit a god-like mind. So absent some independent reason to think the argument is faulty you cannot use Ockham's razor. And note: if you avoid drawing my conclusion by positing extra mental favourings instead, then you cannot deploy Ockham's razor against my view for you have posited something extra as well (indeed, something wholly different from anything else).

It might be objected that it is sufficient that the conclusion of my argument is in tension with the scientific worldview to make it reasonable to reject its conclusion.

So far as I can see, nothing in my argument conflicts with any scientific claims (for instance there is nothing in my argument to suggest that the god in question created the physical universe; if there was space I'd argue that it is fairly obvious that she had nothing to do with it). However, even if there is some conflict with the scientific worldview, so much the worse for the scientific worldview. Whatever certainty one has about the truth of the scientific worldview rests on a prior certainty that normative reasons exist. So the credibility of the scientific worldview is mortgaged to how well it can accommodate normative reasons, not the other way around.