Nothing is Wrong

Why I don't like deontic categories.

There are two rival ways of conceptualizing what to do. The more traditional approach uses deontic categories: Actions are right, wrong, permissible, obligatory, etc., and you should do the right things, not do the wrong things, so on and so forth. The other approach uses scalar language: Actions are better or worse, or you have more or less reason to perform different actions, and that’s that.

I’m very partial to the scalar option. Here I’ll try and give a short motivation for why.

Sidenote: The reason this post exists is that PhD-holder Lorenzo Elijah, PhD of First Principles fame asked me some questions about why I prefer the scalar option, in preparation for a post of his arguing for the deontic view. I got a little carried away with my reply, and so decided I might as well turn it into a post for the drama. So subscribe to his blog to see his response when it eventually comes:

I.

Before going on, we should be precise about exactly what the dispute is. We have these two types of categories—deontic and scalar—what exactly are we arguing about regarding them?

What we’re arguing about are which are fundamental, or what can exhaustively answer questions regarding what to do. There are basically 3 options:¹

1. Deontic categories are exhaustive.

2. Scalar categories are exhaustive.

3. Both categories are legitimate, but neither is exhaustive.

By “exhaustive” I mean something like that category being all there is to say about what to do. If deontic categories are exhaustive, then once you have said everything there is to say about what is right, wrong, obligatory, etc., there’s nothing more to be said—there isn’t any further question as to whether you have such and such strong reason to do X, Y, and Z. There might be other ways of spelling this out, such as what reduces to what or whatever, but I feel like this is good enough.

Notice what this doesn’t commit you to: It doesn’t commit you to being an eliminativist about the other category. Contrary to my outrageously provocative clickbait title, I think it makes a good deal of sense to talk about things as being “right” or “wrong” in many ordinary circumstances. I can say that it’s true that you ought to do X, and false that it’s right to do Y. All I have to insist on is that insofar as this talk will pertain to what to do, it’ll have to boil down to being some sort of roundabout way of talking about what you have more or less reason to do.

And—I might add—there aren’t any “deep” facts about how to map deontic categories onto scalar categories. Just like how there isn’t any “deep” fact about whether a man is going around a squirrel, or both are going around the tree—you can’t be correct or incorrect in your choice.²

As you can probably guess, my position is then (2): Scalar categories are exhaustive.

II.

I’ll argue in a bit of a roundabout way. In epistemology there’s a very similar dispute. We have the category of belief, which is discrete (you either believe P, disbelieve P, or withhold judgement), and the category of credence, which is basically a continuous degree of belief.

We can plot out the same 3 options for how these relate. Now, in this case I think it’s very obvious that credences are exhaustive! Belief-talk is just rough contextually dependent approximation of certain credence-talk.

We might think that an epistemology would have to give us a clear answer as to what to believe, what to disbelieve, and what to withhold judgement about. But once you actually start thinking about it a little more carefully, this starts to seem implausible. If your theory of epistemology can tell me everything about what my credences should be, and how they should change over time, I’d be satisfied.

You might want to ask “well, it’s fine that my credences should be so and so, but what should I believe?” To this I would say that that’s a purely pragmatic matter of how we want to convert credence-talk into belief-talk. Once we’ve agreed on everything about credences, there’s just nothing left to say! You can choose to carve the credences into beliefs, but if we agree on the credences and disagree about which beliefs to have, we’re not actually in any substantive disagreement! We’re just talking past each other.

III.

Similarly, I think it’s very obvious that we definitely have degrees of reason to do different things. I have strong reason not to punch someone in the face, but I have even stronger reason not to stab someone. I have reason to donate $100 to charity, but I have even stronger reason to donate $200.

So let’s suppose we fix all the facts about how strong reason I have to take various actions. Having done this, you might ask, “okay fine, but what would be right, what would be wrong, what would be obligatory, etc.” This strikes me as a very similar question to the one about beliefs and credences, and my answer will be the same: You can choose to talk about right and wrong as discrete categories, and it might even be quite useful to do so. But what’s really fundamental are degrees of reason, and once you’ve fixed those, there is no substantive question left.

So being good and austere theorizers, we should throw away dead weight. It’s very obvious to me that what to do comes in degrees. And you can’t really express these degrees with the discrete language of deontic categories, so I think option (1) (that deontic categories are exhaustive) is clearly wrong. Additionally, deontic categories add nothing—they don’t pay rent. Hence we should drop (3), unless we have some very good reason to think that there’s some job that degrees of reason can’t do by themselves.

IV.

You might object: When I decide what to do, I think about what’s right or wrong, not about degrees of reason—my moral judgements pertain to deontic categories, not scalar ones. Similarly you object: When I decide what to assent to, I think about what to believe, not about credences—my epistemic judgements pertain to beliefs, not credences. In fact I often think about wrongness and beliefs, but I only started thinking about degrees of reason and credences once I made the horrible mistake of opening my first philosophy textbook, sucking me into this miserable debate.

Fine, fine. I agree. I just disagree that this shows that scalars and credences are not exhaustive. To keep things confusing, let me introduce another analogy:

What do you think of the relation between chairs and arrangements of matter? I think facts about arrangements of matter are exhaustive, and I don’t think there are any “deep” chair-facts beyond the arrangement-of-matter facts.

For any particular chair, you can identify it with some set of matter. Still, I don’t think there is any clean identity between arrangement-of-matter-facts and chair-facts, no matter how complicated, since our concept and talk of chairs is simply too vague and fuzzy to be able to get this kind of treatment. This is not because chairs are distinct irreducible metaphysical entities, but because chairs are useful but vague ways of thinking.

I also wouldn’t want to be an eliminativist about chairs. And I would grant that chair-talk and chair-beliefs are very useful, and perhaps indispensable. In fact, they might seem primary in an important sense. I encounter things as chairs before I encounter them as arrangements of matter (Heidegger cheering from his grave). In questions of sitting, what is important is whether something is a chair, rather than the exact arrangement of its matter. [Insert response to whatever other parody argument you might find persuasive.]

I hope everyone agrees that arrangements of matter are exhaustive in this case. But I also hope you can see that judgements about chairs might seem primary in the same way that judgements about wrongness or beliefs might. In this case, however, it’s perfectly obvious why this doesn’t show anything.

V.

What of the relation between the scalar vs. deontic debate, and questions in normative ethics? There’s an obvious affinity between utilitarianism and the scalar view. Utilitarianism evaluates actions by how good consequences they produce. So since goodness of consequences comes in degrees, it seems like you should say that answers to what to do should as well.

This is why I don’t like maximizing consequentialism (the view that it’s obligatory to do the optimal thing): The verdict that anything less than what’s strictly optimal is wrong is clearly faulty to me.

Not only that, but it’s not even clear why in the world one would think that as a consequentialist! There’s no sharp difference between doing what’s optimal, and doing what’s very very very close to being optimal—there’s no reason to draw a sharp boundary between these two.

So since I also think that scalar talk is perfectly adequate—in fact more adequate than deontic talk—at capturing normativity, I see no reason to be a maximizing consequentialist. Besides, when I put consequentialist judgements in talk of what I have more or less reason to do, rather than what is right or wrong, it seems a lot more plausible. For instance compare:

It’s wrong to not sell all your belongings and donate the proceeds to charity—it’s wrong to keep an extra pair of shoes for yourself.

It’s as wrong to kill someone as to let someone die.

You have strong (moral) reason to sell all your belongings and donate the proceeds to charity—you have less strong reason to sell everything except an extra pair of shoes.

You have as much reason to not kill someone as to not let someone die.

Not only do I think the latter are more accurate, but they also just strike me as a lot more plausible. However, it should also be clear why talk of right and wrong is often useful: Talking about strengths of reasons quickly becomes a tangled mess.

VI.

But you might think that deontologists have good reason to think that deontic categories have a shot at also being fundamental in an important sense. After all, they’ll think that there are some sharp boundaries in what to do. For instance, there is a clear difference between doing and allowing harm. But I’m not actually sure how far this gets you—at least if you’re not an absolutist deontologist.

Let me introduce some machinery from Parfit’s Reasons and Persons. In it, he talks about theory-given aims. Call a theory T. Your T-given aims are then just whatever the theory tells you to do. For consequentialism your T-given aims are to promote the good.

We can spell out what deontology tells you to do by asking what your T-given aims are, and grading how well you follow these on some scale. Doing harm will count very strongly in the negative, and allowing harm will not count as much—and doing good will count a little positively, but not as much as doing harm counts negatively, etc. Never mind whether this is a plausible way of representing the theory, or one that appropriately captures what the deontologist takes to be normatively fundamental or whatever—all we need is that it’s a way of representing the duties given by deontology.

Unless you’re an absolutist deontologist, there will be some amount of doing good that can outweigh cases of doing harm. Saving a billion people might outweigh killing a single person. If that’s right, then doing harm can only count a finite amount against your T-given aims, and the more good the doing harm causes as a side effect, the less the action will count against your T-given aims. That means that the wrongness of even a paradigm “wrong act” like killing a person will come in degrees, and can ultimately turn out to be barely wrong, or even slightly positive, depending on external circumstances.

The point of this exercise was to show that even though we might initially think that there are clearly right and wrong actions for a deontologist, this is not the case! Unless they’re an absolutist deontologist, everything can be put in degrees of wrongness—and talking about degrees of wrongness captures normative reality more precisely than simple talk of “right” and “wrong” does.

So even if you’re not a utilitarian, I think you should be a scalar.³

VII.

Now, this conclusion is a little uncomfortable. It would be very nice to be able to just say definitively what actions you should choose and what you shouldn’t, and leave it at that. But I just think that’s sadly not a plausible picture. Rather you just have more reason to choose some courses of action, and we can’t tell you what you “should” do, or what would be “wrong” to do *per se*. All we can tell you is that it’d be better if you did this than that.

But as I argue here, this is just a result of any theory that cares about beneficence, whether utilitarian or not. Once you’ve accepted this disconcerting consequence in one case, I just think it’s not that much worse to accept it across the board.

Anyways, that’s my attempt to make the case for being scalar. It’s maybe not so much an argument for the unconvinced, as a report of the considerations that convince me. I haven’t read what Lorenzo will write yet—no matchfixing—so I look forward to seeing what he’ll say.

1. I’m here assuming that deontic and scalar categories are the only options, and that you think at least one is legitimate.

2. Once you have fixed how to do the mapping, you can of course be right or wrong. Say I’ve chosen to say that actions which you have reasons above level X against doing count as “wrong” (not that the mapping is ever as explicit or intentional as this in the real world). In that case I’d be wrong if I said of an action which I have reasons of levels lower than X that it’s wrong.

3. Well, I don’t know if it’s possible for you to be a mathematical object. But you can at least try!