Welcome to what we expect will be a very interesting and productive discussion of Paulina Sliwa‘s “Moral Understanding as Knowing Right from Wrong.” The paper is published in the most recent edition of Ethics and is available through open access here. Kieran Setiya has kindly agreed to contribute a critical précis, and it appears immediately below. Please join in the discussion!
Précis by Kieran Setiya:
Is there a phenomenon of moral understanding that is importantly distinct from moral knowledge? Non-reductionists say yes. In this impressively clear, creative, and perceptive essay, Paulina Sliwa argues that the answer is no. Moral understanding is knowing right from wrong.
Sliwa’s primary interlocutor is Alison Hills, who has made a powerful case for non-reductionism. I won’t defend Hills’ view specifically, but in this summary, I will explain why I am not convinced by Sliwa’s arguments. In doing so, I will ignore many points on which I am persuaded, as when she protests against over-intellectualized accounts of moral understanding in the final sections of her paper.
Sliwa begins with a helpful distinction between particular instances of moral understanding, as when someone understands why p – p being a moral proposition – and what she calls ‘the capacity of moral understanding’, which is the capacity to understand why p, for some range or category of moral propositions. I will focus on particular instances of moral understanding. And I will begin by conceding that knowledge is necessary for understanding. As Sliwa argues in section 3, you can’t understand why p without knowing why p.
In section 4, which is the heart of her discussion, Sliwa turns to the question of sufficiency. Dismissing the simple view on which an agent understands why p if and only if she knows why p, she proposes:
Moral Understanding as Sufficient Moral Knowledge: An agent understands why p if and only if she has a sufficient amount of knowledge why p.
You may ask: what does she mean by ‘amount of knowledge’? We will get to that. You may also ask: what counts as ‘sufficient’? Answer: this is context-sensitive, but in general, you need to know more for us to say that you understand why p than you do for us to make a bare attribution of knowledge why.
The principal challenge to Moral Understanding as Sufficient Moral Knowledge is moral knowledge acquired by testimony. In Hills’ example, Mary has thought a lot about moral vegetarianism. She tells Eleanor that it is wrong to eat meat because modern animal farming is cruel. Eleanor accepts what Mary says and comes to know both that and why it is wrong to eat meat. If she takes this on trust, however, there is no reason to suppose that Eleanor understands why eating meat is wrong.
Sliwa does not object to this description of the case. (Some doubt that moral knowledge can be conveyed by testimony. Sliwa does not seem to be among them, and I will follow suit.) She argues instead that the description is consistent with her view. On the natural way of imagining the case, Eleanor knows much less than Mary about why it is wrong to eat meat, about why it is cruel, how cruel it is, how much suffering is involved, how animals are kept and slaughtered, and so on. That is why Mary counts as understanding why but Eleanor doesn’t.
Sliwa’s treatment of this case is quite persuasive. It brings out how she thinks about degrees of knowledge. On her general conception of propositional knowledge, to know that p is ‘to discriminate among various ways the world might be and to correctly locate the actual world on the right side of the divide’. How much you know is a function of how finely you discriminate possibilities. What does that mean? In the particular case, what matters is that Mary can discriminate between various ways in which modern animal farming might be cruel and identify which ones obtain. Eleanor can’t. Her knowledge of the cruelty of modern animal farming is coarse-grained.
This makes sense to me, but I wonder how far it goes. To begin with a clarification: I take it that what matters is not merely the extent to which one can discriminate possibilities but the extent to which one can make discriminations that have moral significance. The fact that Mary knows which day of the week the animals are slaughtered or the make of the machine in a slaughterhouse: this is irrelevant, since it makes no moral difference. In contrast, it makes a moral difference, at least potentially, how much pain animals suffer on factory farms, how rich their psychology is, and how easy it would be to reduce their suffering. Mary’s knowledge of these facts is relevant to her moral understanding.
If what matters is the extent to which one can make discriminations that have moral significance, there is a question of how to generalize Sliwa’s response to Hills. Consider a different case of moral knowledge by testimony.
A philosopher of science who knows little about ethics has a meeting with Frances Kamm. Kamm explains that, in a certain complicated trolley case, it is wrong to push the switch that will cause one death, preventing five. The philosopher of science asks why. Kamm cites the Principle of Quintuple Effect, which states that it is wrong to kill one innocent person to save five if and only if … and here she gives an intricate description of causal and other relations that obtain between actions, omissions, deaths, threats, means, side-effects, intentions, beliefs, etc. The philosopher of science comes to know by testimony that it is wrong to push the switch in the complicated case, and he knows exactly why. He can discriminate worlds in which Kamm’s intricate description holds from ones in which it doesn’t as well as anyone ever could. But if he takes all this on trust, can’t we protest that he does not understand why it is wrong to push the button? If so, it is hard to see how Sliwa’s strategy could help.
There are several things Sliwa could say at this point. Perhaps the strategy that worked for the original case of Mary and Eleanor has limitations. She need not contend that it is fully general. All that she is committed to is that those who lack understanding lack relevant moral knowledge. For instance, she might argue, the philosopher of science does not know why the property captured by Kamm’s intricate description is morally significant.
Is this what Sliwa has in mind? It has a certain plausibility, but it has limitations, too. For one thing, some philosophers will deny that there is anything more to know about why the property in question matters in the way it does. We may have reached moral bedrock. What is more, I doubt that Sliwa can avoid the basic dilemma. If someone fails to understand why p because they fail to discriminate possibilities finely enough, or for some other reason alleged to be consistent with Sliwa’s view, we can ask: is there a proposition, q, such that learning q, however difficult that might be, would give them what they are missing? If not, moral understanding cannot be reduced to moral knowledge. If so, we can stipulate that they learn q by testimony, taking it on trust, and then I am inclined to agree with Hills, that they still lack moral understanding. No amount of testimonial knowledge can provide it.
Sliwa may protest that this rather abstract dilemma begs the question. One way to make progress is to look at an alternative to Sliwa’s view, which she goes on to consider:
Moral Understanding as Inferential Knowledge: An agent understands why p if and only if she knows why p by inference from its grounds, e.g. by inferring why an act is wrong from its wrong-making features.
Sliwa makes two objections. First, inference of this kind is not sufficient for moral understanding. Consider Yasmin, who infers that eating meat is wrong from a vague memory of a documentary she heard on the radio, which mentioned that modern animal farming is cruel. Although Yasmin knows why it is wrong to eat meat by inference from a fact that makes it wrong, she does not understand why it is wrong to eat meat in that way that Mary does.
This description of the case seems right, but it is consistent with Moral Understanding as Inferential Knowledge. That is because, while Yasmin may know that it is wrong to eat meat by inference from one of its grounds, her inference is not sensitive to the whole array of facts that bear on the wrongness of eating meat: how much pain animals suffer on factory farms, how rich their psychology is, and how easy it would be to reduce their suffering. These facts make a difference to how strong her inference is and how strong its conclusion should be. Like Moral Understanding as Sufficient Moral Knowledge, Moral Understanding as Inferential Knowledge suggests that moral understanding comes by degree, depending, for instance, on how far one’s inference takes in the wrong-making features of an act.
Moral Understanding as Inferential Knowledge derives support from a variation on Kamm and the philosopher of science. Here we stipulate that the grounds of the moral conclusion are fully specified: Kamm’s biconditional is as detailed as can be. If the philosopher of science is able to infer directly from Kamm’s description to her verdict on the case, without further testimony, I would think he does understand why it is wrong to push the button, though he has no additional knowledge why.
Sliwa’s second objection is that inference of the relevant kind is not necessary for moral understanding. She describes Samir, who knows which factors are relevant to a moral question but needs help in weighing them. He takes advice from a trusted friend, who tells him which factor is more significant, and comes to know what he should do and why. Sliwa claims that Samir has moral understanding, despite the fact that it is not based on his own moral inference.
I think the case is under-described. It indicates a subtle but vital distinction between epistemic achievements that are enabled by testimony, on the one hand, and second-hand knowledge, knowledge that rests on and is justified by testimony, on the other.
In one version of the case, Samir learns to weigh the factors properly himself: he learns to make the right inference. It doesn’t matter that he learns this by listening to his advisor’s testimony. The causal origins of his knowledge are irrelevant to its status as inferential, in the sense that matters to Moral Understanding as Inferential Knowledge. That is why he has moral understanding. In another version of the case, Samir is still reliant on testimony. His knowledge rests on and is justified by (or through) his advisor’s knowledge. Then Samir lacks moral understanding, as Moral Understanding as Inferential Knowledge predicts.
I don’t mean to defend Moral Understanding as Inferential Knowledge. The distinction that matters is, I believe, more general that the distinction between inferential and testimonial knowledge. It is the distinction between first-hand and second-hand knowledge, knowing for oneself, and knowing by trusting another. Without this distinction, we cannot make sense of a central task of moral education, which is teaching others to know things for themselves.
Philosophers sometimes write as if explicit moral education were a matter of giving moral testimony. I think that is profoundly wrong. It is true that explicit moral education, as when I talk to my son about racism or sexism, involves a lot of testimony. Much of what I say about the non-moral facts he can only take on trust. But if, at the end of the day, his moral knowledge is merely testimonial, I have failed. At best, the process of moral education is radically incomplete. What I want is for him is to see for himself what is wrong with racism and sexism, to make the inferences himself, or where the knowledge in question is non-inferential – as perhaps, when he comes to know the axiom that all human beings have fundamental rights – I do not want his knowledge to be testimonial, but his own. He should not believe that human beings have rights on the basis of his trust in me, but on his own authority.
Let me step back from this debate to ask a wider question. I have been suggesting that, while she is right to focus on depth of knowledge in explaining moral understanding, Sliwa omits a further aspect: the contrast between knowing first- and second-hand. There are two issues here. One is whether moral understanding requires first-hand knowledge. The second is whether, even if it doesn’t, we should admit the existence and importance of first-hand knowledge as a topic for moral philosophy.
These issues raise a cluster of related questions for Sliwa, with which I will close. I am interested in Sliwa’s arguments for reductionism, but I am also interested in why it matters to her that reductionism is true. What is at stake in her refusal to countenance moral understanding as a phenomenon distinct from knowing right from wrong? Does she admit the existence of first-hand knowledge as a distinctive phenomenon, but deny that it is required for moral understanding? Or does she reject its existence altogether? If she rejects its existence, how does she think about the enterprise of moral education, as in the difficult conversations I have with my son?
(Thanks to the SHAPE reading group at MIT for a wonderful discussion of Sliwa’s essay, on which I have extensively drawn, to Brad Skow for helpful comments, and to Paulina Sliwa for writing such a lucid and provocative piece.)1