External Reasons: A Natural Law Response to Williams

In my last post I criticized Bernard Williams' rather Humean argument in "Internal and External Reasons" for arguing that external reasons can't explain action. As a follow up, I wanted to post my paper I wrote for the meta-ethics class I'm taking where I criticize further Williams' argument against externalism, and at the same time build up a natural law account of reasons and practical rationality in opposition to Williams. Here it is!

                      In “Internal and External Reasons,” Bernard Williams presents an argument for thinking that external reasons do not exist, and thus all external reasons statements are false.[1] In this paper I will do three things. First, I will explain what Williams understands internal reasons and external reasons to be. I will then explain Williams’ argument against external reasons. Finally, I will attempt to give some defense of external reasons by critiquing Williams’ argument.

                      The general form of a reasons statement is “A has reason to do F in circumstances C.” Williams aims to show that statements of this form are only ever true on an internal interpretation. While Williams does not seem to give a definition of internal reasons statements, he does lay out what seems to be a necessary condition on internal reasons statements. He says that any internal interpretation of a reasons statement must “display a relativity of the reason statement to the agent’s subjective motivational set,” which we shall call “S.”[2] Roughly then, internal reasons for an agent are dependent on what is in the agent’s subjective motivational set.[3]

                      Williams also lays down as a necessary condition on internal reasons that they can be discovered by deliberative/practical reasoning.[4] While Williams does not explicitly define what deliberative or practical reasoning is, he specifies his conception of deliberative/practical reasoning via example. In particular, he says that practical reasoning includes means/end reasoning about the most preferable way of satisfying a desire, temporal ordering of when to satisfy which desires, determining which desires one is most interested in satisfying, and determining what would constitute satisfaction of one’s desires. So, this condition amounts to saying if A has an internal reason to F, A must be able to motivate herself to F by a process of reasoning of this sort (from S).[5]

                      It is clear then that Williams is working with an idea of internal reasons that ties them closely to an agent’s current subjective motivational set. It is also clear that he is working with a “thin” notion of practical rationality. Since Williams doesn’t explicitly define what he takes rational deliberation to be, it is difficult to precisely state what this “thin-ness” amounts to. However, roughly speaking, Williams’ account of rationality is “thin” insofar as, on his view, a decision will count as rational to the extent that it could be concluded to by a process of deliberation starting from one’s desires and satisfying certain (relatively weak) formal constraints.

Bernard Williams on Internal Reasons (Contra Natural Law)

Here is a short precis I wrote for my meta-ethics class about Bernard Williams' argument for internalism about reasons. Internalism about reasons is, roughly, the Humean-ish idea that an agent has a reason to act a certain way only if it is somehow grounded in the desires of the agent. This is directly contrary to the natural law view, according to which what is reasonable to do or not to do is independent of the desires of the agent, and thus reasons are independent of an agent's desires:

In “Internal and External Reasons,” Bernard Williams presents an argument for thinking there is a problem with external reasons statements.[1]

1.    If something R can be a reason for action, then R can be a reason for someone A’s acting on a particular occasion O. [Premise]

2.    If R can be a reason for A’s acting on a particular occasion O, then R can figure in an explanation of A’s acting in O. [Premise]

3.    So, if R can be a reason for action, then R can figure in an explanation of A’s acting in O. [1,2 HS]

4.    If R can figure in an explanation of A’s acting in O, then R cannot be an external reason. [Premise]

5.    So, if R can be reason for action, then R cannot be an external reason. [3,4 HS]

6.    If R cannot be an external reason, then R is an internal reason. [Premise]

7.    So, if R can be a reason for action, then R is an internal reason. [5,6 HS][2]

Hence, the argument establishes that the only coherent notion of a reason for action is an internal reason for action, i.e., one which is relative to an agent’s subjective motivational set S. For now, let us grant premises 1 and 6. I’d like to think about premises 2 and 4 a bit. My point will be that, depending the sense of ‘explanation’ here, either one or the other premise will be implausible.

                      First, let us consider the interpretation of explanation where an explanation of an action is something that (directly) motivates someone to do that action. In this sense, maybe we should accept 4. For maybe it is only one’s desires, or at least something closely bound up with one’s desires, that could explain an action in the sense of being able to directly motivate it.[3]

                      However, in this case, I think that premise 2 is implausible; for I don’t think that a reason for acting in an occasion has to be able to motivate you. Rather, a reason for performing P in circumstance C is some state of affairs or consideration that could make performing P in C practically intelligible.[4] And something can do this without being able to motivate. Let me explain.

                      If you found me in a chair stabbing myself in the leg you would ask me why I’m doing what I’m doing. If I said to you, “I just desire to,” this wouldn’t really answer your question. There is still a clear sense in which what I’m doing is unintelligible, and wouldn’t make sense even if I obeyed some bizarre psychological laws which led me to “just desire” to stab myself in the leg. If however I said, “I am scared there is a tiny alien in my leg, and I need to get it out so it doesn’t kill me,” I will at least have described some state of affairs where, when my action is seen in light of this, my action does at least become intelligible (even if, in all likelihood, it is still utterly ridiculous and unreasonable). Such a state of affairs is a reason for acting a certain way.

                      But if this is a correct account of what reasons for action are, then 2 seems false. For it might be that a state of affairs S could make doing P intelligible, and thus counts as a reason for doing P, but my subjective motivational set makes me completely unable to even take S into consideration, and thus makes S unable to motivate me. This happens all the time. For instance, suppose I have no desire or motivation to quit smoking. However, the fact that smoking greatly increases my chances of dying certainly could make the action of quitting intelligible (i.e., if someone appealed to this fact when justifying their choosing to quit). So the fact that smoking greatly increases my chances of dying is a reason to quit. Yet, again, it might be that it couldn’t motivate me on any occasion.

                      All of this is to say that, on the interpretation of ‘explain’ where it is roughly equivalent to ‘motivate’, premise 2 seems implausible. On the other hand, if ‘explain’ means ‘make intelligible’, then premise 2 seems true, but premise 4 seems false, at least if reasons for action are, as I’ve explained, states of affairs which make acting certain ways practically intelligible.

---

[1] The “problem” is either that they are false, incoherent, or are misleading (and presumably should instead be rephrased as internal reasons statements). See the last paragraph of p. 297.

[2] This is my interpretation of Williams’ argument at the bottom left paragraph of p. 295. I may be misinterpreting Williams, but this seemed to me the best way to make his argument valid.

[3] I am assuming the sense of ‘can’ here means something along the lines of ‘is physically possible’.

[4] This account is largely adapted from G.E.M. Anscombe’s book ‘Intention’, sec. 37.

New Natural Law and Deriving an 'Ought' from an 'Is'

Many people have said that we cannot derive an 'ought' statement from an 'is' statement. In one sense this is trivially true, in another it is straightforwardly a mistake in logic. I'll discuss this in relation to new natural law theory and the grounding of morality in metaphysics.

First the sense in which it is true. Obviously if the premises of your argument contain no 'ought' statements then you can't logically derive an 'ought' from them. At best you can conclude from the premises that some ought statement is plausible. However, I think most people who want to say we can derive an ought from an is would all agree with this rather trivial observation.

Now some discussion of natural law: On new natural law theory we immediately perceive certain states of affairs as to be pursued, and from this we determine what we ought to do; the states of affairs to be pursued are instances of the 'basic goods', which are immediately recognizable aspects of human flourishing. Now, new natural law theorists say we cannot derive an ought from an is. However, they also say that new natural law theory does not carry any commitments one way or the other as to what, metaphysically speaking, human flourishing consists in or is grounded in. After all, everyone can know the natural law in their hearts, and you do not need to be a metaphysician to gain insight into what's right and wrong. But by the same token on new natural law theory, it should be consistent with the theory to say the human flourishing is grounded or consists in something like, say, the perfection of the human form. New natural law should not rule this out.

What follows from this? Well, if this type of theory is possible on new natural law, then supposing it is true, every aspect of human flourishing will at the very least coincide with some aspect of the perfection of the human form. The details of this perfection aren't that important, since the main point of this post is supposed to be logical. So just take some arbitrary aspect of human flourishing A and some arbitrary aspect of the perfection of the human form P. A and P coincide just means that if one exists then so does the other. Now we can launch an argument:

(1) If P and A coincide then P is an aspect of the perfection of the human form if and only if A is an aspect of human flourishing. [by def.]
(2) P and A coincide [prem]
It follows logically that
(3) P is an aspect of the perfection of the human form if and only if A is an aspect of human flourishing. [from 1 and 2]
From this we can infer by the rules of logic that
(4) If P is an aspect of the perfection of the human form, then A is an aspect of human flourishing. [from 3]
(5) P is an aspect of the perfection of the human form. [prem]
Again by the rules of logic:
(6) Therefore, A is an aspect of human flourishing. [from 4 and 5]
(7) If A is an aspect of human flourishing, then instances of A are states of affairs that ought to be pursued. [by NNL]
(8) Therefore, instances of A are states of affairs that ought to be pursued. [by 6 and 7]

(1) just follows from the definition of 'coincide'. (7) is just what new natural law says. Given the premises the rest follows by the incontestable rules of logic. Now, one might question premise (2) and (5). However, it is at least possible for them to be true given new natural law theory. And if they are true then this argument shows one can derive an ought from an is.

Basically, to say we can derive some statement Q from some set S of premises {P1, P2, ... , Pn} just means that there is a proof of Q from S. This is Logic 101 stuff. Given all the premises and definition in our argument we can derive an ought, namely (8). Our premises, (2) and (5) are paradigm 'is' statements. This is why it is false to say we cannot derive an 'ought' from an 'is' statement.

Here's the upshot for natural law theory: If NNL is true, then it is possible for (2) and (5) to be true. If (2) and (5) are true there's a perfectly good sense in which we can derive an 'ought' statement from a set of 'is' statements. Hence, even if NNL is true, it is possible to derive an 'ought' from an 'is' in the precise logical sense of 'derive' above. The only other possibility is to either say aspects of human flourishing can't coincide with anything, which basically means NNL is inconsistent with certain metaphysical theories of flourishing (viz. all of them), or else we are using some other sense of 'derived' in saying an 'ought' can never be derived from an 'is'. As a person sympathetic to NNL myself I think we should go with the latter (or possibly drop the use of the phrase altogether).