The moral formula: How facts inform our ethics

Sam Harris may not be a household name in Britain, but in America he is right up there with Richard Dawkins, Christopher Hitchens and Daniel Dennett as one of the new atheism's "four horsemen".

He burst on to the public science scene in 2004 with his anti-religious polemic, The End of Faith, which he wrote while still a graduate student in neuroscience. It went on to sell half a million copies, and he has been a major public figure ever since.

Having used science to attempt deicide, Harris now threatens to do the same for moral philosophy. In his new book, The Moral Landscape, Harris sets out to convince us that science can not only help us to understand human values, but determine them.

Richard Dawkins has said of the book that "moral philosophers will find their world turned upside down". However, when the book was released in America last year, they argued it was Harris, a Stanford philosophy graduate, who had got things the wrong way up. I got the opportunity to put some of their criticisms to Harris when I met him in California.

Julian Baggini: Why do we need science to tell us what is right and wrong?

Sam Harris: I'm using the term "science" very broadly, in terms of our best efforts to make valid truth claims based on evidence and clear reasoning. So history, for my purposes, is a science in the sense that we can make true or false claims about historical fact.

What I'm arguing is that morality, questions of good and evil, right and wrong, because they relate to questions of human and animal well-being, also entail truth claims about our world, human nature or the prospects of human happiness that fall within the purview of science. Otherwise, we're just left to argue over preferences: things are wrong because we don't like it or a majority of people don't like it.

Your case rests on the claim that morality must relate to the well-being of conscious creatures. What is your argument for that?

I call it the "worst possible misery for everyone" argument. A universe in which every conscious creature suffers as much as it possibly can, for as long as it can, is bad, and anything else is better than that, by definition. Given that conscious experience arises based on the laws of nature in some way, movement out of that space is constrained by the laws of nature. There are going to be ways to avoid the worst possible misery for everyone and there are going to be ways to think you're avoiding it and fail.

The moment you grant me that, then it seems I have everything I need to say that there is a continuum of experience in which we have the worst possible misery for everyone on the one hand and then all of these other degrees of well-being on the other. I call this the moral landscape, with various peaks and values.

How do you get from that to the claim that science can actually determine human values?

All of this is a domain of discovery. If we understand the dynamics of mental life in real detail – to speak particularly of our case, if we understand the human brain in real detail – then we will know how various experiences and ways of living with one another – thoughts, intentions and behaviours – affect human life.

How much should we value freedom of speech? How much should we value compassion? How should we raise children to think about the suffering of other human beings? All of this can only be understood in the context of science in the end.

Then science will be answering questions about, for example, whether we should we be privileging obedience to parental authority over free expression and if so, exactly how much and when and what are the consequences of getting the balance wrong. All of these are neurological questions in the end.

But it's puzzling how science could tell us, for example, how to prioritise between rights of free speech and privacy?

There are probably some trade-offs where there isn't an important difference. So privileging free speech to some degree and privileging privacy to another degree leads you to different circumstances, but perhaps they are not importantly different. If you and I and everyone affected by those changes could live out both lives, and have our brains scanned all the while, and have every marker of our inner lives analysed, we would come out saying they were a little different, but we don't know which we like better. That is an intelligible prospect and that is why the moral landscape has many peaks and valleys that are different but equivalent in terms of well-being.

Isn't well-being too ill-defined to be scientifically tractable? Take the classic thought experiment of whether a person who lives a normal life with ups and downs is better or worse off than someone who takes a happiness pill. There doesn't seem to be a factual answer as to what's better, discoverable by examining fMRI scans, for instance.

I think we can have a rational discussion about how much we want our states of consciousness, our emotional lives, to track the reality of our lives. We definitely want it to track it for the most part because otherwise, if we're just taking this perfect narcotic each day, it's not a sustainable situation. You're just lying on the couch in bliss, but your relationships have dissolved, you've lost your job, and your children have starved to death. It's materially unsustainable if nothing else. But your love for the people in your life, which you value and which is major component of well-being – your connections to others, your ability to function in the world – all of this is predicated on your states of consciousness tracking the actual reality of your life in the world.

Surely there are going to be cases where you don't get that clear fact that tells you it's a failure. So for example, we could have a world where machines could do most of our work and a significant proportion of the population could, sustainably, if they chose to do so, take the happiness pill and go on to live delusional, happy lives. I don't know what fact about the world would tell us that was wrong.

You can concoct odd situations in which, if you're really conserving everything we could plausibly mean by well-being and maximising it, then I would have to say that was good by my definition, But the worry is that some really important things are left out by that account. But I am arguing that if they really are important, they must translate into some sort of well-being. So, for instance, these stoned people don't know anything about science, they're ignoramuses. They are forsaking all of the joys of learning about the universe. If that has a cash value in curiosity and in the fulfilment that comes from being able to teach your children things, we're still talking the talk of human well-being.

Can you think of any situation where there are significant disagreements about what well-being is where science – a discovery or new fact – would tell us who was right? Because if you're going to claim that science determines human values, it means that the arguments we currently think of as philosophical disputes about ethics will be able to be settled scientifically in the future.

I think that's clearly the case. One way to see that is by analogy with health. When I talk about morality I'm really talking about psychological health, and the health of societies. Can science tell us about psychological health? If the sciences of mind are, in fact, sciences, and they are, in fact, of mind, then one would hope so, at some point.

Take a truly fraught, value-laden question such as "How should parents raise children?" Clearly, if we understand anything about child development, healthy emotional lives, healthy cognition and what it means to equip children to become high-functioning adults , enjoying all the fruits of civil society peacefully and collaboratively, we are talking about how you should raise children, and there are scientific truths there waiting to be discovered.

Do we know it all in detail? No, but we certainly know enough to know that many people are raising their children badly. Beating children – subjecting them to pain, violence and humiliation – is just not a good way to discipline them. I think this is now well established in psychology.

On your view, is moral philosophy going to go the way of alchemy and be superseded by science?

Yes and no. I view philosophy as essentially the womb of the sciences. Whenever a question is not experimentally tractable, not quantifiable, then it's squarely in the domain of philosophy. The frontier between philosophy and science is never clear. But the moment you start actually talking about data and neurophysiology it would seem you're playing more the language game of neuroscience than philosophy.