There’s been a fair bit of blogosphere buzz about Sam Harris‘s recent TED talk, entitled “Science can answer moral questions.” I didn’t expect to agree much with Harris, given my usual objections to empiricist scientism and related attempts to exalt “science” against “religion.” And I think there are indeed a number of problems with Harris’s view. And yet there’s quite a lot that Harris gets right – at least as much, I think, as most of his critics.

The most widely read response to Harris (and the one that Harris himself responded to at length) is one by Sean Carroll. I find the Harris-Carroll debate instructive because both seem to miss the most important point; and that, in turn, would seem to be because both fall prey to an unfortunate empiricism.

At the heart of the debate is the supposed dichotomy between “facts” and “values,” or “is” and “ought.” (I would rather say “should” than “ought,” because “ought” sounds increasingly rare and archaic in contemporary North American English, but that’s a quibble.) Harris insists that values are a kind of fact, even objective fact, so that “should” or “ought” statements have a meaning grounded in reality, not entirely relative to or dependent upon the subjects making the claim. “Should” statements, on this view, are a kind of “is” statement. In this, I think, Harris is entirely right.

Where Harris slips up is in missing the elision of “fact” with “empirical fact.” It’s this point that lends plausibility to Carroll’s criticism: Carroll is right to reply that we get “off on the wrong foot by insisting that values are simply a particular version of empirical facts.” Harris’s reply, however, misses this elision, not challenging it, and that’s why he’s vulnerable to Carroll’s counter-claim: “there exist real moral questions that no amount of empirical research alone will help us solve.” (his emphasis)

On Harris’s example of corporal punishment, for example, let us assume that Harris is right that corporal punishment negatively affects the well-being of children and of society in general. Does that give us sufficient reason to say that corporal punishment is wrong? Not if we buy Kant’s theory of punishment, according to which punishment is an obligation owed to those punished, irrespective of its consequences. For Kant, as is well known, does not take the well-being of conscious creatures as the primary measure of goodness or rightness. Is Kant wrong? I think he is; but I also think there’s something wrong with a viewpoint that takes happiness, or even more broadly defined consequences like “well-being”, as the sole standard for ethics. I think Harris is right to say well-being should be a standard by which we judge actions, but as far as I can tell he’s got no ground whatever to say it should be the standard.

But to get back to Carroll, the next question to ask here is: just what kind of question am I arguing with Harris about here? Harris, I think, is right to say that they are questions of fact. And to some extent even of objective fact: claims about good and bad do not depend entirely or even primarily on the subject making those claims. Even Kant would agree: lying is wrong whether or not you think it’s wrong, whether or not you want it to be. It’s just that, contra Harris, it’s not an empirical fact; establishing it relies on procedures of dialectical and demonstrative argument that can, but do not necessarily, involve reference to empirical states.

For Kant, knowledge of moral principles is surer than knowledge of the empirical world, because empirical facts change, but moral principles – like mathematical principles – are derived from a priori principles which are true no matter what happens to the physical world. We can imagine ourselves waking up in The Matrix and finding that the laws of physics in this new reality are completely different from what we thought they were. We cannot really imagine 2+2 being 5, even in the Matrix. That’s why Plato looked to mathematics, not empirical science, as a source of certainty; Kant saw moral truths as being like mathematical truths.

Now is Kant right about that? Not wholly. He is right to move the question beyond the realm of the entirely empirical; some ethical claims, especially those at the foundations, must involve the a priori. In his counterargument, Carroll starts to show that he gets this point: “The crucial point is that the difference between sets of incompatible moral assumptions is not analogous to the difference between believing in the Big Bang vs. believing in the Steady State model; but it is analogous to believing in science vs. being a radical epistemological skeptic who claims not to trust their sense data.” Indeed. What Carroll doesn’t get here, though, is that the disagreement between the scientist and the skeptic is itself a disagreement about facts, about the way that the universe is. It can in principle be resolved through argument, just as Carroll tries to resolve his own debate with Harris through argument, while still acknowledging that the debate does not rest on empirical evidence.

And so, while the analogy stands up very well, what doesn’t stand up is the way Carroll resolves the analogy: “In the cosmological-models case, we trust that we agree on the underlying norms of science and together we form a functioning community; in the epistemological case, we don’t agree on the underlying assumptions, and we have to hope to agree to disagree and work out social structures that let us live together in peace.” The assumption here seems to be that scientists can reach agreement because they share underlying assumptions, but that no agreement can be reached with those who don’t share those underlying assumptions. But if that’s so, science is wrong – or at least it’s no more right than Christianity, the Taliban, or any other belief system that Carroll might otherwise wish to condemn. Because of course the Taliban agree on underlying norms and form a functioning community – much more so, I dare say, than scientists do. The hard part, and the place where the norms of ethics are to be established, is arguing across the boundaries of those communities, finding truth between people whose assumptions are radically different. This is exactly what advocates of science like Carroll need to do, not just on questions of ethics, but on the value of science itself. For Carroll – unlike Harris – is saying here that science, like ethics, is itself true only relative to the assumptions of the scientific community. But the whole point of science is to do better than that – to say something about how the physical universe actually works, not just about how we think it works. (In other posts Carroll speaks of wanting to “convert” Catholics to atheism or naturalism or skepticism, which suggests that he does indeed think science’s views are not just different but better; for him to really claim that his views were simply equivalent to Christianity would, I think, be disingenuous.) Ethics is much the same here. Science and ethics both try to establish matters of fact; both rest on assumptions that are always disputed. But we do ourselves no favours in either arena by throwing up our hands and saying there is no truth that crosses communities.