Can There Be a Science of Morality?

Published by timdean on

Can we have a science of morality? This question has been thrown around quite a bit of late, especially fuelled by the spirited ejaculations of one Sam Harris. Harris firmly believes there are no barriers to a science of human values, but I fear things aren’t that simple, and I’m not alone in this concern.

Sam Harris

While a ‘science of morality’ is a laudable notion in a loose sense, such a science would, by necessity, look nothing like what Harris has in mind. Harris is seeking not only a science of morality, but a science of human values. He wants a “universal conception of human values” that can be checked, verified and proven using the tools of empirical science.

But that’s just not going to work. Science doesn’t do that kind of thing. At least not without assistance from other disciplines, like philosophy. And if we try to force science alone into providing us with values, there is no shortage of traps that will inevitably spring up.

Naturalistic fallacies

It’s worth remembering that Harris isn’t the first one to call for a science of morality. Herbert Spencer and a number of other naturalistic ethicists around the end of the 19th century thought they could finally use scientific tools to discover human values.

Spencer happened to believe that evolution, in particular, was the tool of choice. He held that evolution was intrinsically progressive, and he equated ‘more evolved’ with ‘more valued’. And we homo sapiens were the most evolved organisms around so, thankfully, we found ourselves at the top of the value tree.

There are a few clangin’ great problems with this view – such as that there’s no evidence from within science that evolution is intrinsically progressive in Spencer’s sense, nor that such progress (if it existed) could be equated with value. But the chief clanger was flagged by G.E. Moore in his book Principia Ethica.

You’ve probably heard of this issue under the oft used, and oft abused, moniker of the ‘naturalistic fallacy’. Despite its common usage, the naturalistic fallacy doesn’t simply state that you can’t derive an ought from an is. Instead, it suggests that ‘the good’ – which Moore sees as the object of moral enquiry – is a simple (rather than complex) and undefinable idea.

Like the colour ‘yellow’ can’t be broken down or explained in any other way – particularly not by simply pointing to a yellow thing and identifying the colour yellow directly with the yellow object. Yellow is a property of the object, and as such, defies any simple identification with any other natural properties. As such, identifying ‘the good’ with evolutionary progress, or another natural property like happiness, as Harris does, is spurious.

Then there’s the other naturalistic fallacy, which is more often referred to as Hume’s Law, or the is-ought fallacy. This says that no amount of is facts alone will tell you how something ought to be. Thus I can describe, in minute detail, the factual circumstances surrounding some killing, but without adding a value in there somewhere – such as ‘killing an innocent person is wrong’ – can we derive a moral judgement from the situation.

The question is: where do these foundational values come from? If science only trades in facts, then that can’t be the source of values. They must come from somewhere else, but where? And there you have one of the great driving problems of the last century or so of moral philosophy.

Likewise, it’s a problem for Harris, whether he chooses to acknowledge it or not.

Of metaethics and men

Harris has stated that the is-ought problem isn’t such a big barrier, and we can undertake empirical investigation into how much suffering or happiness an organism experiences – and that will lead us to derive new values that can shape our behaviour. But all he’s doing is smuggling in his prior values – i.e. that morality is about the well-being of conscious beings, as measured by happiness or suffering. But he doesn’t justify that foundational value. And that’s why he’s been the target of so much criticism by incredulous philosophers.

Yet Harris is outwardly dismissive of philosophy and metaethics:

Many of my critics fault me for not engaging more directly with the academic literature on moral philosophy. There are two reasons why I haven’t done this: First, while I have read a fair amount of this literature, I did not arrive at my position on the relationship between human values and the rest of human knowledge by reading the work of moral philosophers; I came to it by considering the logical implications of our making continued progress in the sciences of mind. Second, I am convinced that every appearance of terms like “metaethics,” “deontology,” “noncognitivism,” “anti-realism,” “emotivism,” and the like, directly increases the amount of boredom in the universe.

Hey, I’m sympathetic to the idea that doing metaethics is a dreadful way to spend your time. But even if you think that most metaethics is garbage (as I happen to), you can’t just ignore it and move on. You at the very least need to state why you think it’s garbage, and why it doesn’t impact your argument.

Sidestepping your opponents has never been a well respected argumentative technique, at least, not in respectable circles.

And, sadly, Harris’s half-arsed engagement with philosophy has only done his cause a disservice. Because there is an important role for science in moral philosophy – but science alone doesn’t a moral philosophy make.

Rebel without a clue

Ultimately, I think Harris is rebelling against two long-standing traditions that have deeply (and often detrimentally) influenced moral discourse. The first is moral rationalism and non-naturalism. This is the kind of moral philosophy that stems from the likes of Plato and Kant, and maintained by G.E. Moore. All of these thinkers held that morality was a subject seeking truth, but that truth couldn’t be discovered by looking around the world. Thus they held a disdain for empirical research. Kant even dismissed doing moral psychology – or ‘moral anthropology’ as he called it – as a dreadfully mistaken way of understanding how we ought to behave.

The other is moral relativism – but not relativism of the philosophical brand, as advocated by the likes of Westermarck, Wong or Harman. Harris is rebelling against post-modernist moral relativism. This is the brand of relativism that is based on a worldview that suggests there is no objective reality that is knowable, so everything is ultimately subjective.

It says that we subjectively project our values on to the world, and our subjectivity is in turn shaped by our enculturation – which is largely influenced by existing power structures in the world and what others want us to believe. As such, oppression of the weak by the strong involves the strong (i.e. government, corporations, mainstream media, etc) shaping the worldview of others to have them behave in ways that maintain the power of the strong. Nefarious stuff.

According to this pomo perspective, our values are intricately linked to our particular worldview, as shaped by our history. But there’s no privilaged worldview, no objective worldview. As such, it’s impossible to justify criticism against another culture’s values because we don’t have any objective bar against which to measure it. And any criticism of the ‘weak’ by us (i.e. the Western mainstream oppressive imperialists) risks further perpetuating the oppression of the weak by the strong.

That view is, sadly, deeply flawed. And Harris isn’t the only one who thinks so. But it’s sad that in rebelling against both non-naturalism and mad-dog pomo relativism he neither understands what he’s rebelling against, nor wishes to engage with literature that might help him rebel effectively.

Towards a science of morality

But rebel one can, and science is a crucial tool in that rebellion – but it’s only one tool, not the entire toolbox.

In sidestepping philosophy, Harris is actually missing out on a route that might not only aid his programme of naturalising ethics (in the sense of grounding it in the natural world – the world described by science), but might improve his position. That route is a new kind of anti-realist error theory developed by Joshua Greene.

Greene, in his 2002 dissertation, argues that morality isn’t about truth – not of the mystically binding kind sought by many philosophers – but that this doesn’t doom ethics to rampant relativism or bleak nihilism. Instead he admits that there are no fundamental, binding, metaphysical oughts – so the thing the realists of the naturalist or non-naturalist kind are looking for just don’t exist. But that there are facts about being nice to other people.

The key is to realise that seeking metaphysically necessary oughts is folly, so we need to construct morality on other grounds, and use this as the source of our foundational values. I’ve argued before that morality is best understood as a mechanism that facilitates prosocial and cooperative behaviour, because such behaviour can help us advance our interests. I believe this is where morality came from, and our moral sense evolved over millennia in order to help solve the problems of fostering prosocial and cooperative behaviour. This is an empirical fact, not a metaphysically necessary fact.

But, I think that enough people will agree that they want to advance their own interests, and that being social and cooperative is the best way to do it (‘I won’t interfere with your interests if you won’t interfere with mine’ – classic social contract theory, a la David Gauthier). As such, we can build a moral framework that allows us to promote our own interests by protecting the interests of others.

That’s where you get your foundational values. They’re contingent, empirical and debatable – and forever will they be that way. This stands in stark contrast to Harris, who wants to use science to discover values by looking in the world, and hold them to be true with the same conviction that we know the charge of an electron.

The truth is that there is no truth of the matter to be discovered concerning fundamental moral values, only agreement to be made between individuals.

However, once you have this anti-realist, contractarian framework, you use facts to fill in the gaps. Once you agree that killing an innocent person is wrong, you need a lot of facts to determine whether, in a particular situation, an innocent person has been killed. This is where science is crucial.

Science can also play a role in exploring the problems that morality is trying to solve, i.e. how to enable a large unrelated group of individuals to live and work cooperatively together without chaos ensuing. It can help us understand that there is more than one strategy to promote our values. Science can also help us understand our moral psychology and how we actually make moral judgements – in the hope that we can become better at making such judgements.

But science can’t tell us the fundamental values by itself. That’s Harris’s error.

Science and morality are closely linked, but science is but one discipline that will help us better understand morality, it is not the only discipline we need to understand morality.

Even in light of Harris’s errors, I applaud his efforts to escape from non-naturalism and pomo relativism. Like him, I’m optimistic about the future of morality, and a closer integration of science with philosophy. But while philosophers have taken centuries to wake up to the need for empirical investigation, it seems scientists are likewise slow to realise the benefits of good old fashioned philosophy. The closer these two streams get, the better we’ll all be.


James Gray · 21st October 2010 at 5:06 pm

I am sympathetic to many of your points. However, I have some points I would like to add:

First, it’s a bit vague what Harris means by “science.” He might think that moral philosophy is “relevant” and he might merely not discuss it because he is talking to the “masses” rather than philosophers. Additionally, no one knows what “philosophy” is, which partially justifies his use of the word.

Sam Harris’ views on morality would make a lot more sense if he didn’t use the term ‘science’ to mean, apparently, ‘rational inquiry’.” —

You suggest that Harris thinks we can know moral values from “empirical” information, but I’m not sure that is what he thinks. Is there a quote you can give to support that?

Second, it’s not clear that Harris is a moral realist or even cares about meta-ethics at all. Normative ethics could be the domain he is most interested in. To say that something is “valuable” doesn’t necessarily mean it’s “intrinsically good.” To say that something is right or valuable could have some irrealist meaning.

Third, a large portion of philosophers are now “epistemological naturalists” and see philosophy itself as a kind of science. If you are a naturalist, then “science” is the method used to find truth — and science can find moral truths and values. That doesn’t mean that moral realism is true. It just means that we can find out right and wrong in terms of “well being” of some sort — and we can even describe “well being” using science.

Tim Dean · 21st October 2010 at 11:46 pm

Hi James. If Harris is talking about reason instead of empirical science, then he’s doing a poor job of articulating that. He uses the term science. He’s a scientist. I presume he knows what science is. And it’s not as broad as ‘rational enquiry’. And if all he was saying was that morality should be investigated using rational enquiry, well, I’d agree.

As for deriving values from empirical investigation – he’s talking about using science to learn about values, so I take that to mean he wants to use the tools of science to discover moral values like you use them to discover the charge of an electron. I find that problematic.

That said, I also think he smuggles in his own preconceptions about what’s valuable – i.e. happiness or wellbeing. If you agree that his (loose) definition of happiness is the cardinal value, then sure, you can use science to discover the best means to achieve it. But he’s inexcusably imprecise on how to justify his claim that happiness or wellbeing are the point of morality, or even what these things means.

I take him as being a moral naturalist, like the utilitarians of old, but perhaps I’m wrong. His refusal to talk philosophically means it’s hard to see precisely where he stands.

And he can talk normative ethics all he wants, but if he does so, he can’t just ignore challenges to justify his metaethical stance – even if it’s to provide some reason for why he doesn’t accept the current metaethical discourse.

I’m normally pretty patient with all manner of thinkers and ideas. But I do get impatient with Harris. To come blustering into moral philosophy – which, IMO, is going through a bit of an empirical renaissance at the moment care of Haidt, Greene etc anyway – and start dissing philosophy entirely and claiming it’s time for something new, then to offer up something that is known to be philosophically problematic, well, it’s irresponsible. And just not becoming of a respectable scholar or intellectual. So I get a bit more impatient than perhaps I should.

Now, this ‘epistemological naturalism’ is new to me. For mine, I see science as a kind of (or daughter discipline of) philosophy – one bounded by the limits of empiricism – not vice versa. Philosophy should defer to science when the question at hand can best be answered by empirical means. But if that question can’t be answered by looking in the world, then all we have beyond that is reason, and that’s philosophy. But that might just be my crazy definition of philosophy (which I’ll have to write about some time…).

Paul · 22nd October 2010 at 5:31 am

Is there any reason not to quickly turn the claim that morality is a mechanism that evolved promotes prosocial and cooperative behavior into a truth about fundamental moral values? I’ve actually seen Steven Pinker, on a bloggingheads t.v. episode, Platonize this view by saying that morality, always and everywhere, will consist of the evolved adaptive responses of organisms to social living. The real kicker was that he formulated this view as an iteration of a mathematical truth, that of the prisoner’s dilemma.

Anyway, defending Harris, if it is empirically demonstrable that morality is just an empirical fact, then why drag philosophy in to it? Science can demonstrate what it is that our minds actually do value. It strikes me as a little bit odd to take the analysis any further than that. If science can reliably demonstrate what it is that our minds actually do value, than why should we question, at all, using those values as our foundational values? As well, I share Plato’s suspicion that it isn’t even possible for most human beings to rationally question the values that their minds unconsciously hold under current conditions, and moreover I think that such a view is empirically verifiable, or at least as verifiable as many social science hypothesis. So what good then is moral philosophy? Nietzsche, Strauss, and Plato suggested that empirically speaking, moral philosophy has had little direct impact on the masses, but has had a profound impact on the powerful, and has often served as the motivation for like minded elites to join efforts in trying to change the political institutions in order to change the broader culture and society. The enlightenment is one example of this; Marxism, and nationalism, influenced by Romanticism, might be another. Would it even be a good thing(i.e. would we affirm it to be a good thing if we actually experienced both and choose to continue to engage in it) if we were able to interrogate our values in the way that Plato suggests is characteristic of a philosopher?

Not that I’ve actually read Harris; but my hunch is that Harris is probably thinking those sorts of things to himself. Even if his argumentation is lazy, I think he may be closer to the truth. So long as it is very difficult to change people’s values, we may as well take them as given, unless we ourselves are true outliers and have different values. I would suggest that philosophers, in their ability to question and flout communal norms, are such outliers. But then the discussion of what is moral is only really relevant to people with the inclination to change values, both their own and others, and to what extent they should give consideration to others who are unlike themselves.

James Gray · 22nd October 2010 at 5:47 am


Here is something Harris said about science:

Some of my critics got off the train before it even left the station, by defining “science” in exceedingly narrow terms. Many think that science is synonymous with mathematical modeling, or with immediate access to experimental data. However, this is to mistake science for a few of its tools. Science simply represents our best effort to understand what is going on in this universe, and the boundary between it and the rest of rational thought cannot always be drawn.

Here is what he said about there being right and wrong answers about morality:

When I speak of there being right and wrong answers to questions of morality, I am saying that there are facts about human and animal wellbeing that we can, in principle, know—simply because wellbeing (and states of consciousness altogether) must lawfully relate to states of the brain and to states of the world.


I don’t know if he would allow “well being” to include deontological theories of philosophy or not. It sounds consequentialist. Perhaps his main interest is values rather than “right and wrong.” If so, what he says could be deontological.

Harris does not necessarily say we can know everything empirically. Here is what he said about knowing what has value:

In fact, I believe that we can know, through reason alone, that consciousness is the only intelligible domain of value.

Therefore, he thinks that “reason alone” plays a role that observation doesn’t in understanding morality.

I am not a huge fan of Harris, but I just want to have a clear understanding. I don’t think he should insult philosophy. Do you have any other quotations where he actually insults philosophy? Like I said, he might be talking more about how “the masses” might view moral philosophy rather than himself (where you quoted him).

The fact that Harris avoids the traditional philosophical language makes it difficult to know exactly what he thinks or wants to prove.

Mark Sloan · 22nd October 2010 at 7:52 am

Tim, I challenge your assertion: “… there is no truth of the matter to be discovered concerning fundamental moral values….”

I understand we agree that searching for ‘magic oughts’ (justification beyond rational choice for accepting the burdens of any morality) in moral philosophy is probably a fool’s errand and searching for them in science is nonsense. So rational choice (the choice expected to best meet needs and preferences) is the only secular justification, as far I know, for accepting the burdens of any morality. This idea supports your position that moral norms adopted by secular, rational groups will only be matters of agreement (social contract kinds of morality).

Consistent with the above, we could also agree that “… there is no truth of the matter to be discovered concerning WHAT fundamental moral values OUGHT TO BE….”.

But science is fully capable of teasing out the underlying necessary characteristics, if there are any, of “…WHAT fundamental moral values ARE…”. Note we are talking about discovering the underlying necessary characteristics of what fundamental moral values ‘are’; there is no need to do anything equivalent to defining ‘good’.

Whether such characteristics exist is an empirical question. As I have said previously, I think they do. More than that, these characteristics imply a specific universal morality (a cross species, beginning of time to end of time type universal morality) that I expect to be a more rational secular choice for groups than any alternative presently available to people.

Paul · 23rd October 2010 at 1:56 pm

Mr. Sloan- why do you think that a universal, cross species morality would be “a more rational secular choice for groups” than the evolved morality that people already have? If fundamental moral values are only the values necessary to more or less solve the problems of living as a social animal with a modicum of intelligence and memory, wouldn’t we as a species already exhibit one such instantiation of a universal morality?

Mark Sloan · 24th October 2010 at 3:03 am

Paul, that expectation is based on the characteristics of this definition of morality, which is something like: “Moral behaviors are behaviors that increase, on average, the synergistic benefits of cooperation in groups and are unselfish at least in the short term”. These characteristics are:

1) Moral behavior is defined not as an obligation, but as a means of producing benefits – it is, on average, in your long term self interests to act morally.

2) It reveals the underlying, necessary elements of all cultural moral standards and our own biologically based/culturally modified moral intuitions. Therefore, there should be less dissonance with our moral intuitions than any alternative morality.

3) It the morality implied by what science has determined (or could determine) that moral behavior ‘is’. It is a unitary definition of morality that is objective in the same sense the rest of science is objective. Alternative moralities are hugely variable in intent and content and have been and I expect will be the subject of constant disagreement.

4) People’s abilities to predict which action will actually be in their best interests are often poor. It is usually impossible to know all relevant data needed to make such a prediction. If we had that relevant data, we lack the brain power usually needed to make such a prediction. Finally, in the heat and confusion of the moment we are often unduly affected by our “base animal instincts” such as greed and therefore are often mistaken in our judgments about what action will most likely be in our best interests. Therefore, it will usually be in our best interests to rely on this “wisdom of the ages” rather than our own “wisdom of the moment” in choosing how to act. We will then usually be making the rational choice (the choice expected to best meet our needs and preferences) if we choose to conform to this morality even though we may expect, in the heat of the moment, that doing so will not be in our best interest.

These four characteristics seem to me to make the proposed science based morality “a more rational secular choice for groups” than any alternative.


Young links | Evolving Thoughts · 21st October 2010 at 5:19 pm

[…] Does God’s omnipotence extend to vision? The study of society last part (Robert Paul Wolff). Can There Be a Science of Morality? [Nope] Little ‘Value’ in New Harris Book. At the Guardian: Socrates – a man for our times. […]

Leave a Reply

Avatar placeholder

Your email address will not be published.