Atheism and Moral Subjectivity

Talk given to the Atheist Society on 13 May 2014 by Jonathan Rutherford

Tonight I want to reflect on the way in which meta-ethics has been discussed in the context of the recent God debates. I want to suggest that the so-called ‘new-atheists’, have tried to defend an unconvincing form of moral realism. This has exposed atheists to justified criticism from theists. My argument tonight is that we would be on firmer ground if we began the, admittedly very difficult, task of fleshing out and explaining a convincing and palatable form of moral subjectivism.

Now I have just used some big philosophical words there, so let me explain. What do I mean by meta-ethics? Most of our daily moral debates are not about meta-ethics. They are about normative ethics: that is, they are about how we ought to behave in any given situation. Meta-ethics, on the other hand, takes a step back and asks more fundamental questions about the nature of the moral enterprise. It enquires into the meaning of moral terms. When we say honesty is ‘right’, for example, are we making a factual statement about the nature of reality or are we simply expressing our own approval for honest behaviour? When we say somebody ‘ought’ to always refrain from being cruel to others, are we referring to some kind of binding moral duty that exists irrespective of what that person happens to desire?

There are many important sub-theories and complexities within meta-ethics but, for the purposes of tonight’s talk, we can confine ourselves to defining the two major positions. The first and by far the most popular is moral realism. According to this view, terms like ‘right’ "wrong’ ‘good’ and ‘bad’, refer to moral facts or laws, which exist independently of what anyone happens to desire or prefer. In other words, there are some actions that are simply wrong (or right) period and, in principle it is possible to attain sound knowledge of these moral realities. An act of cruelty, for example, would be ‘wrong’ even if everyone in the world approved of this type of behaviour. Cruelty, according to this view, is in fact wrong.

The second view, which is the minority view, but the one I hold, is known as moral subjectivism. This view rejects the existence of ‘objective’ moral facts or laws ‘out there.’ The wood in this lectern has certain objective chemical properties that can be studied and analyzed. These properties, we assume, exist independently of what anyone thinks about it. Somebody could think this lecturn was made of fairy dust, but they would be wrong about that and science could demonstrate this to be the case. But, for the subjectivist, the same is not true for questions of human moral value. What, asks the subjectivist, is this mysterious ‘wrongness’ quality in an act of cruelty or murder?! How could you possibly begin to establish that moral acts have such qualities? We cannot detect this moral quality via our five senses nor can we infer that such a quality exist from the effects certain behaviour have on others. Why, then should we believe, that certain behaviours or actions have any moral quality at all?

What this means is that for the subjectivist the only the only stuff out of which a moral system can be built is natural things like human desires, preferences and the consequences of actions. Nothing else counts, because nothing else exists. If you think something else exists that may impinge on morality – some metaphysical category like moral truth, for example – please demonstrate how you know that!

It helps in getting handle on Subjectivism to recognise that value is something which only exists in minds.  Organisms value things but the things do not contain any value. All humans desire or dislike some behaviours more than others. Most people have a strong dislike for cruelty. But an act of cruelty, according to the subjectivist, has no inherent negative quality of ‘badness’. An act of cruelty is, in fact, valueless. For this reason, the philosopher Stephen Law describes subjectivism as the ‘spectacles view of morality’. The value that we think is objectively part of the world ‘out there’ is actually added by the emotional spectacles through which we are looking.

Long ago the philosopher David Hume summarised this view when he said:

Take any action allowed to be vicious. Willful murder for instance. Examine it in all its lights and see if you can find that matter of fact, or real existences which you call vice….You can never find it, till you turn your reflection into your own breast and find a sentiment of disapprobation which arises in you towards this action.

One dramatic implication of subjectivism is that, the only ‘ought’ statements that make sense are instrumental ‘oughts’ as opposed to categorical ‘oughts’’. One can only say, for example, you ought not to be ‘cruel,’ if ‘you prefer to be kind’ or if you don’t want to incur punishment or rebuke from your community.’ But if you find somebody who likes being cruel, and/or doesn’t care about any potential disapproval or punishment, or any other consequences that might flow from being cruel, then it would be logically incorrect to tell that they ought not to be cruel’ because this is not connected to anything they happen to care about.

Now to be clear, I personally detest cruelty. I strongly disapprove of that type of behaviour. If confronted with a cruel person I would try to alert them to the needless suffering they would inflict on their victims. I might also point out that if we allowed cruelty to become widespread, social order would be seriously threatened. Or perhaps I would seek to persuade them that they would be happier if they learnt to deal more constructively with their negative emotions. But ultimately, a moral subjectivist, like me, simply cannot categorically say ‘you ought not to be cruel because it is wrong.’ Such a statement has no validity because cruelty does not have any inherent value. You only ought not be cruel, if you dislike that type of behaviour, or dislike the consequences that cruelty might bring about.

For the moral realist this seems totally unacceptable. They would say that even if everyone in the entire world enjoyed being cruel, cruelty would still be wrong! But, again, the problem for them is to demonstrate to us sceptics how they know this to be the case? In my view, nobody has ever done so.

I guess you will now be able to see why subjectivism has been the least popular meta-ethical theory!

Moral Realism in the God Debate:

The unpopularity of subjectivism extends to atheists. Many atheists and secularist ethicists in general, claim to reject moral realism and embrace some variety of subjectivism. But I find that, in practice, they still cling to the notion, whether explicitly or implicitly, that moral truth exists and can be discovered by reason and science. I recently, for example, listened to an interview with renowned moral philosopher Peter Singer. He argues for a form of normative ethics known as ‘preference utilitarianism.’ So for him, morality has a lot to do with desires and the consequences of action. One would think he might, therefore hold a meta-ethical position in line with subjectivism. But, in the interview, he clearly proclaimed his conviction that some moral views are more correct than others, and moral truth merely awaits our discovery. He gave no convincing empirical or logical reasons for this, however.

The weakness inherent in atheist attempts to justify moral realism, was brought home to me very clearly when listening to the Christian podcast 'unbelievable.' I highly recommend this podcast by the way. Despite being run by Christians it is very fair and balanced and features leading theists and atheists discussing a variety of subjects. On this occasion there was a lively discussion on meta-ethics.

The atheist could certainly give a compelling explanation for why humans, as a social-species, have developed complex moral systems. As anthropologist David Eller points out, social living depends on a degree of ‘social regularity’ which requires all members of the group to share common expectations about the way oneself and others should be treated. Evolution also blessed humans, rather uniquely, with several other abilities relevant to moral development. We have, for example, advanced capacities to empathise with others - literally ‘feel what others feel.’ Our brain power enables us to inhibit our own behaviour as well as remember previous interactions (and therefore learn from them). Our advanced language systems allow us to accurately communicate our desires and pass on moral learning from generation to generation. This, as Peter Singer has argued, provided us with the material basis for complex moral systems. These could then be developed and enhanced, albeit in diverse ways across cultures, via cultural evolution and social-psychology.

But the Christian rightly pushed the atheist to explain how, given this naturalistic understanding, he could justify believing specifically in moral truth? Why were some moral positions true or false, irrespective of human preference? How had this mysterious moral realm evolved over time? And how did the atheist know that his moral system happened to be the correct one? The atheist was, frankly, stumbling for an answer, and contradicting himself at every turn. ‘Yes,’ he agreed, slavery, sexism, and the concentration camps were absolutely wrong for all time, despite the fact that some cultures and individuals have, at times, approved of them. But no he was not a ‘moral realist’ but rather a ‘quasi realist’...whatever that was supposed to mean!

The theist, on the other hand, could give a straightforward reason for why moral facts or laws exist: they derive, of course, from the will of God. Now, as most of you will be aware, this ‘divine command’ view of ethics is not without its own problems. Theists are obliged to explain why different cultures and individuals have perceived God’s moral law in such diverse and contradictory ways. Also, as Plato made clear in his famous dilemma, if morality is the product simply of God’s command, then morality reduces to God’s arbitrary whim. If, for example, God decreed that cruelty was 'right' then it would be! In this sense, the divine command view of ethics, as many have noted, is ironically a form of subjectivism: morality derives from what God approves, even if he approves of things that we humans might detest.

Despite these problems, the theist, unlike the atheist, did at least have some basis for believing in an objective morality for humans. Moral truth for humans is derived from God’s approval or loving character. The atheist, by contrast, had no basis whatever to ground his belief in moral truth, in the sense of a realm outside of varying human desires or preference.

For this reason, the subject of meta-ethics, I submit, creates real dilemmas for atheists. If they claim ‘moral facts’ exist they have a metaphysical problem not dissimilar to the problem of the theist. Atheists have rightly required of theists that they provide us with convincing evidence to substantiate belief in a supreme being. Given no persuasive evidence has been put forwarded – and our naturalistic understanding of the universe has grown with the progress of science – we reject belief in gods, angels and other unsubstantiated spiritual entities.

But, as we have seen, in the realm of morality, theists have thrown back the challenge in our face. What reason or evidence do you have, they ask, for your belief in moral facts? Doesn’t this require that you posit the existence of a metaphysical realm, perhaps a spiritual realm beyond our senses, in which to ground your belief in moral truth? And if not, how do you justify your belief in moral facts, given you cannot provide us with substantive evidence? To be consistent, they say, shouldn’t you reject such a belief? And doesn’t that lead straight towards the dark and dangerous territory of moral nihilism.

As I will explain further below, I think there is a satisfying and convincing way out of this dilemma for atheists. But, theists are correct to point out the tension here. As a subjectivist and also an atheist, I would put the tension this way: If atheists are prepared to expose the God Delusion, why are they not also prepared to expose, what the moral philosopher Ted Trainer calls, the ‘moral law’ delusion?

Atheist attempts to detect moral facts through science

So how have the new-atheists attempted to solve this dilemma? Most, if not all of them have answered by arguing that science and morality are perfectly compatible. So far so good, and I agree with them – more on this below. The trouble is that they think science can help us establish the existence of moral facts. This is where I disagree. A classic example is the leading U.S atheist Richard Carrier. In the book End of Christianity (which I highly recommend by the way) Carrier argues boldly that ‘moral facts naturally exist and science can find them.’ Sam Harris has argued along similar lines, in a book called ‘the moral landscape: how science can determine human values’.

Their attempts, it should be noted, have differed from the traditional arguments for moral realism. Perhaps the most famous traditional argument was made by G.E Moore. Moore claimed that we can detect moral facts through intuition. But this was never convincing. The mere fact that humans intuit, often very strongly, that some acts have a moral quality, does not mean they actually have such a quality. As Trainer argues ‘it is not a very satisfactory argument to say that because I have a powerful feeling that X is wrong, there must be Morally Wrong things and X is one of them. Especially when there may well be another society somewhere in which people feel or sense that X is Morally good.’ The whole history of science, I would add, acts as a cautionary tale against trusting our intuitions. Clearly they can get things very wrong!

The approach taken by the new-atheists, however, is different and seems initially more promising. As I said, they appeal directly to science arguing that moral truths can be discovered by empirical scientific investigation. Have they provided us subjectivist sceptics with a convincing case? I am most familiar with Richard Carrier’s argument, so will I will focus mainly on that.

Carrier tries to ground the existence of moral facts in natural desires shared by all humans. We all have similar biological desires for sex, food, shelter but also for love, companionship, belonging, and, ultimately, happiness. He points out, rightly, that there is now a science of happiness, which enables to understand the broad parameters for human well-being.

Carrier, being a smart man, knows that not all humans will act rationally to maximize wellbeing. Humans can be very irrational. Smokers, for example, seem to put their long term health at risk for the sake of short term pleasure or addiction. So Carrier qualifies himself. All humans, when rationally and fully informed, will have similar desires for a good, happy, healthy life. For Carrier we ‘ought’ to act to satisfy our deepest desires, when fully rationally informed about how best to satisfy those desires. This is a moral imperative that science can discover.

Carrier makes a further qualification: differing environmental conditions, may change how we ought to act. For example, most of the time humans want to live in societies where they are free from the risk of being killed. In some situations, however, such as war, we find it necessary to kill or harm others in self-defence, self-preservation, or the greater good. According to Carrier, this does not undermine the existence of moral facts because ‘any system of true moral facts will already include the fact that if we were forced into the same conditions, we would be compelled by the same imperatives that then obtain’ (Carrier, 2011, p353). In other words, fundamental human desires are always the same, across cultures; it’s just that environmental conditions may change how we act to achieve those desires.

Carrier summarises his argument as follows:

‘as human beings share all the same primary biological desires (which are not limited to the so-called based desires for, say food and sex but include, as science has demonstrated, desires for love and companionship and joy and fulfilment and more, ordered in similar hierarchies of ultimate and instrumental necessities), and only such desires can ever rationally entail (in conjunction with knowledge) an informed conclusion about we most want, it follows that we will all (when rational and equally informed) desire most the very same thing (when in the same circumstances) which logically entails that the same moral facts will be true for us all. Therefore, universal moral facts must necessarily exist.’ (Carrier, 2011, p356)

Carrier’s argument is, in fact, very close to the kind of subjectivism I outlined earlier. It focuses entirely on natural things that science can indeed detect and evaluate, like human desires, preferences and the consequences of actions. He thinks that when humans are fully aware of all the facts, and reasoning rationally about their desires/preferences, they we will all share similar desires or preferences for how to live and behave. This, he argues, is the basis for some kind of ‘objective’ scientifically based moral system. Humans can indeed be wrong about what they ‘ought’ to do, because they can act in ways that do not achieve their most fundamental and deepest desires for a good life.

The argument, however, is wrong for two reasons.

Firstly, there is a false premise: it is simply not true, that all humans share the same or even similar values, even when they are reasoning rationally, are fully informed, and are operating under similar environmental conditions. Even when all these conditions hold humans can vary immensely in the relative value they assign to both their own long-term welfare and the welfare of other sentient beings.

Take for example, a ‘rational robber.’ This robber is very skilful and well practiced. He believes, correctly, that he will not get caught and will therefore escape punishment. This particular robber happens to place a much higher value on the possessions he will attain, than any misery he might inflict on the victims. It follows that the most rational thing for the robber to do, in light of his own value hierarchy, is to steal! Carrier provides us with any reason to think the robber is doing the wrong thing. The robber is acting on his strongest desire, and he is acting rationally. The robber of course is different from most people. Most people will be distressed or concerned, at least to some extent, about the misery of the victim. So here, contrary to Carrier, we have a clear case of variation in human moral values.

In reply, Carrier will say that the robber is not really being rational. Stealing is not really what the robber wants. If the robber was rationally informed he would know that stealing is likely to lead to a life of crime, which will not lead, in the long term, to a fulfilling life. Maybe: maybe not. But this is beside my point. As I said, the robber has fundamentally different values from most people. The robber does not, for whatever reason, care much about the misery of his victims. Carrier has just introduced another reason for why the robber might not steal – namely to achieve a more satisfying life for himself in the long term. But this does nothing to alter the fact that the robber has very different desires from most people: he does not place much value on the welfare of others.

Or, to take another example, think about charitable giving. As Peter Singer argues, in affluent societies such as ours most of us have lots of surplus income beyond that needed to meet our basic material needs. There are also many reputable organizations and groups that are, in one way or another, working to alleviate or overcome conditions of appalling poverty or offer relief in the wake of natural disasters. We could therefore easily give some of our surplus money to such efforts. But despite most people being aware of appalling suffering across the world, and their power to do at least something to alleviate it, we typically find a great variety of response. Some give some of their money to humanitarian causes, some a lot. Probably the vast majority spend the bulk of their surplus income on utter trivialities when viewed in light of the vast ocean of urgent and serious human need. There is, in other words, a spectrum of responses from serious concern and financial commitment, to utter disregard, and everything in between. Carrier might say that whether or not we give to distant others is not a ‘moral’ issue. But why not? It seems clear that we have the power to alleviate suffering and we either choose to do so or not, depending on the degree to which we care about the suffering of the distant other. So again, here we have a clear case of differences in human values, even when people are more or less fully informed about the facts of the situation and consequences of their actions.

A final example: the debate between vegetarians and meat-eaters. When vegetarians point out the needless suffering inflicted on animals throughout the factory farming process, how do meat eaters typically respond? Very commonly by saying that they simply care more about satisfying their palate with some delicious chicken or lamb dish, than any potential misery inflicted on sentient creatures. The point here is not to say that the vegetarians are morally right. As a subjectivist no such position is available to me. Though neither, of course, are they wrong. The point is, here we have another clear case, contrary to what Carrier asserts, of where humans hold different values, even when fully rationally informed.

This false premise crucially undermines Carrier’s argument. His whole case rests on the idea that, when fully rationally informed, humans will share very much the same kinds of desires about what they want and therefore how to behave in order to achieve their desires. But, as we have seen, this is simply not the case. Humans have different value hierarchies: they approve or detest of different things, and with different degrees of strength. And this variation occurs both at the individual level, and across cultures.

Here is the second problem with Carrier’s argument: Even if humans held identical value hierarchies, it would not show that moral facts exist! At least not moral ‘facts’ in the way the term ‘facts’ is usually understood.

It’s true that if we all wanted the same things, there would be widespread agreement on how we ‘ought’ to behave, particularly if we were all well informed and all agreed on the likely consequences of certain actions. That is to say, there would be inter-subjective agreement. In many realms of morality, this is a close approximation of what exists today. The vast majority of people, for example, want to live in a society where they feel safe, so they rationally approve of strong rules and laws sanctioning murder.

But, intersubjective agreement is not the same as moral facts. When we talk about the facts of evolution, for example, we are saying these are realities about how the world works irrespective of what the creationist thinks. Carrier provides us with no clear reasons for thinking that science can detect moral facts in this sense. He has only shown that if we happen to desire certain things then certain courses of action might be instrumentally effective in achieving those ends.

And here is the fundamental point. What possible rational or scientific justification can Carrier give for the ends that we human seek? All of you here in this room have a value hierarchy. There will be whether you not or not some things that you are fundamentally for or against. At the top of your value hierarchy will be those things that are most important to you, not as means to other ends, but as ‘ultimate ends’ which you seek. It might be popularity, or wealth, or something less self-interested like the advancement of your nation or perhaps even the maximisation of happiness for all. But, none of us can give any rational defence of the ‘ultimate ends’ we seek. The pacifist can give no more rational defence for why they abhor violence, than the sadist can for why they enjoy inflicting pain. No human end is more ‘rational’ than any other. Humans simply desire things. As David Hume said: reason is the slave of the passions.

To summarise, I have been arguing that Carrier’s attempt to ground moral truth in science is unconvincing. But, for all that, Carrier does actually come close to explaining the nature of morality in terms that would be acceptable to a subjectivist, like me. Carrier is completely right that moral enterprise can be reduced entirely to an investigation of natural things like desires and the consequences of action. Where Carrier errs is by adding a totally unsubstantiated metaphysical category – namely ‘moral facts,’ or moral truth – on top of this entirely naturalistic framework.

A subjectivist account of morality:

At this point, I imagine, many of you, if you don’t disagree entirely, would be horrified and repelled subjectivism. Doesn’t it lead, you may ask, to complete nihilism? Certainly there is a nihilistic element to subjectivism – after all we reject the existence of moral facts – but what I want to now show is that this is not the whole story. A subjectivist understanding of morality gives us everything we need to construct workable and attractive moral systems.

Even without an objective moral realm it remains the case that humans have shared desires to live together well. It is also the case that, to a greater or lesser extent, all of us places some value on the welfare of others. That is, most of us have the capacity to empathise. Given this, most of us want to live in societies in which there are certain standards of behaviour that secure not only our own welfare but also the welfare of others. This provides the basis for moral and ethical codes.

The moral enterprise is in no way dissimilar to the implementation of road rules. Most of us want to be safe when we are driving. It is therefore instrumentally effective for us to design road rules like ‘drive on the left at all times.’ Sensible road rules like this are likely to minimize road accidents. But as with road rules so to for moral rules such as ‘do not kill’ or ‘do not lie’. In most circumstances these are wise guidelines for individual behaviour that help us achieve our mutually shared desires for self-preservation, security, and cohesive orderly societies.

In short, morality and moral system comes down to what you and I individually and collectively want. Everything depends on the kind of society we would like to live in. Subjectivism implies that there are no truths or laws to which we must submit, or which can help guide us. We are radically free to construct our own rules and regulations and moral codes. Certainly we would be doing nothing factually ‘wrong’ if we decided to allow free reign for murderers every Monday. But actions are not without consequences and so would we really want to that? If we are wise – and if we care about others – we will choose to construct social codes and systems that are peaceful, fair and sustainable. This will work out best for all of us.

Matthew Ferguson points out that the implications of rejecting moral realism may not be as dire as it may first at seems. As he says:

The real difference between moral realism and moral anti-realism is ultimately conceptual, not consequential. If moral anti-realism were true, it would be something we discovered about how the world already is and operates. It would not entail that we should or will behave any differently. We would just discover that the ways in which we have conventionally understood morality is mistaken. Rather than being based on external, factual truths, we would discover instead that what we have traditionally construed as moral truth is actually our set of preferences for how we ideally want the world to be. And, if we should discover that is the case, we can still want the world to be that way and keep working towards it. Nothing really changes except our understanding of what is and has already been the forces in play. If moral anti-realism is the case, then we have still effected moral change, built democracies, instituted human rights, and achieved many of the other things we consider "good," and we can continue to go on doing it.

Should Atheist come clean on the subjectivity of morality?

To conclude, tonight I have argued that the widespread belief in moral truth creates real dilemmas for atheists. If atheists argue that ‘moral facts’ exist they have a metaphysical problem not dissimilar to the problem of the theist who believes in God. But, as we have seen, in the realm of morality, theists have thrown back the challenge in face. On what grounds do atheists possibly have for belief in moral realism? As I have attempted to show the best attempts of new atheists have thus far been unconvincing.

On the other hand subjectivism is very counter-intuitive. If atheists started exposing the widespread moral law delusion would they begin to lose credibility among the majority of people who feel quite certain moral facts exist? Would we run the risk of being branded as a bunch of nihilists, out to destroy the moral fabric of society? As I briefly suggested, subjectivism is perfectly compatible with viable and attractive moral systems, understood purely as a human creations designed to meet common human desires. But, given the moral law delusion is so deeply embedded in our culture it probably would be a very risky business to adopt this position explicitly. What should we do? I will leave you to ponder!


David Eller, ‘Christianity does not provide the basis for morality,’ in ‘The Christian Delusion: Why faith fails,’ John Loftus (ed), Prometheus Books, 201o.

Fredrick Trainer, ‘The Nature of Morality: An introduction to the Subjectivist Perspective,’ School of Education, University of NSW, 1985

Richard Carrier, Moral facts naturally exist and (and science could find them), in ‘The End of Christianity’, John Loftus (ed), Prometheus Books, 2010