DRMacIver's Notebook
Reason as a moral mechanism
Reason as a moral mechanism
Another draft bankruptcy post. This ended up being another overambitious philosophy piece of the sort I haven’t really figured a time or place to write.
I’ve since learned that these go better in a google doc. Maybe they’d go better yet if I expanded my notebook site to have an essays section now that I’ve got sidenotes and full pandoc support for references and such.
Attention conservation notice: This piece is quite inside baseball. I think it’s of interest to a general audience in where it ends up, which is why it’s on the newsletter rather than the notebookEditor’s note: lol never mind., but it starts by arguing about the logical underpinnings of Kantian ethics. You should not however require much philosophy background to read it.
I’m currently doing a read through of Velleman’s “Self to Self”. You can follow along with my increasingly frustrated noises I make as I try to decode it on my YouTube channel (Editor’s note: I also abandoned this series and project), but be warned the production quality is approximately zero these are very much notes I am taking in my reading.
I’m mostly still stuck on chapter 1, which is his introduction to Kantian ethics. I find the content a frustrating mix of tantalisingly close to pointing at something important I’m missing and just wildly wrong. This post is an attempt to explain some of my frustrations with it, try to make sense of what it might mean, and reconstruct something that I think actually works and is pointing roughly in the same direction, but definitely isn’t what Kant or Velleman mean.
Why should I act according to reasons?
One of the core premises of Velleman’s account of Kantian ethics is that we have a basic obligation to “act according to reasons”. He justifies this as such:
Where we previously asked “Why should I act on my desire?” let us now ask “Why should I act for reasons?” Shouldn’t this question open up a route of escape from all requirements?
As soon as we ask why we should act for reasons, however, we can hear something odd in our question. To ask “Why should I?” is to demand a reason; and so to ask “Why should I act for reasons?” is to demand a reason for acting for reasons. This demand implicitly concedes the very authority that it purports to question – namely, the authority of reasons. Why would we demand a reason if we didn’t envision acting for it? If we really didn’t feel required to act for reasons, then a reason for doing so certainly wouldn’t help. So there is something self-defeating about asking for a reason to act for reasons.
If you’re like me, your response to this is something along the lines of “What? No.” or possibly “What on earth are you talking about?”
This argument is too clever by half and relies in a fundamental way on an ambiguity in the term “reason” that, once you resolve, causes the argument to dissolve into uselessness.
The ambiguity is this: When you ask “Why should I do the thing?” the answer to this is an explanation of why you should do the thing, which may or may not itself be present prior to your doing the thing.
An answer to the question “Why should I act according to reasons?” is not necessarily providing the sort of reason that compels action, because the answer is intrinsically descriptive.
The core problem in all of this is that it is attempting to weasel out of the normative content of a “should” resting on you actually caring about something by trying to figure something that you logically must care about. But you can just not do that - I don’t need a logical argument to care about not being hurt for example, I can do things to avoid pain without needing a compelling logical justification for this. “Because if you don’t the reason fairy will appear and hit you with a big stick” is a perfectly logically consistent answer to “Why should I act according to reasons?” that does not have any circularity.
It grounds out the authority of reasons in the authority of pain, but that’s fine if you’re willing to accept an ethics for human beings who don’t like pain rather than an ethics for perfectly spherical ethicists in a vacuum. It is only circular if you think “Why should I act to minimise pain?” roots the authority of pain in the authority of reasons, but that’s not what that question means - it is asking for an explanation, not delegating authority, because the avoidance of pain already has authority over you.
The rest of the chapter tacitly acknowledges all of this by basically trying to work through a set of reasons for acting for reasons, it just can’t say out loud because it’s built on this premise that reasons are an inescapableIn general the idea of escapability of authority in Velleman’s account of Kantian ethics seems (Editor’s note: What does it seem?? I apparently left this sentence hanging) source of authority.
This is a shame, because I think the actual source of the authority of reasons is much more interesting.
What does any of this even mean?
I think Velleman considers the meaning of “act according to reasons” self-explanatory, but I think it’s not just not self-explanatory but borders on meaningless given the ambiguity of the word “reasons”. If a “reason” is a thing that causes something, then the obligation is tautologically satisfied - all your actions are caused by something. If a “reason” is a verbal explanation of why things happened as they did, then it will generally come after the fact rather than prior to it - almost everything we do well, we do without a significant amount of verbal reasoning prior to it.For an elaboration on this last point, I recommend “The Enigma of Reason” by Dan Sperber and Hugo Mercier, and “Sources of Power” by Gary Klein
Throughout, I will use “reason” to mean something like “verbal causal explanation” - i.e. something you would say in answer to a “Why?” question. I will use actual cause to refer to something like… the state of the world that lead to a particular outcome, a set of conditions which would lead to the same outcome reliably. A reason is a necessarily incomplete description of the actual cause.I’m using bold here to indicate that I intend these to be precise terms which are compatible with but more specific than colloquial meanings. When I use a bolded term without bolding I will usually be referring to someone else using the colloquial word. Or I forgot.
Importantly, reasons are not normally present in our mind at the point of action (they can be, but not necessarily). So when you are acting “according” to a reason, in that the reason accurately describes the actual cause, that reason is not itself part of the actual cause. This is what I mean by the argument resting on an ambiguity: If you are asking for a reason to act in accordance with reasons, the fact that you can answer the question does not create a circularity because your answer need not be part of the actual cause for acting for reasons.
One easy way to see this is that often the actual cause exists prior to it being possible to give an explanation for it. This is especially true for causes that are in some sense physical. e.g. if you ask me why I’m not working my best today and I have a low grade virus that is sapping my energy, that is part of the actual cause regardless of whether I know I have a virus, or even have the concept of virus. Both “I have a virus” and “I don’t know, I’m just not feeling it today” count as reasons. I could even answer “I’m lazy”, which would arguably count as an incorrect reason.
Additionally, there is often far more reason than you are aware of in the moment, so you could not have thought of all the reasons in the time you had prior to action. Suppose you see someone who has injured themself. You help them get to a safe place, and then call the emergency services. Why did you do this? Because they were hurt. Why is “because they were hurt” a sufficient reason? Because it’s morally the right thing to do. Why must you do the morally right thing? Because I want to, and will feel guilty if I don’t. Why would you feel guilty if you didn’t? Because I was raised proper… etc.
There is a long chain of verbal reasons you can give for this and almost none of them went through your head when you helped. You might have verbalised something like “Oh they’re injured, I should help them”. Maybe you considered a couple arguments and counter-arguments. But there was not a long chain of verbal reasoning for why you should help them, the helping comes with an immediacy of its own as the thing for you to do.
The actual cause is thus not the reasons you give for it, because those reasons were not present at the point of action. The actual cause is necessarily something like “because I am the kind of person who in these kinds of circumstances performs these kinds of actions”, with the reasons attempting to define some boundaries on all of those “kind of”s.
This should, I think, still be considered acting according to reasons, in that “I helped them because they were hurt” is a reason that accurately describes your actions.
This is, however, now a very weak requirement. All it demands is that your actions be in some way comprehensible. “I helped him because his shirt was yellow”, or “I helped him because I felt like it” are both also potential reasons. This weakness is not a failing of the Kantian account, which is in many ways all about deciding what counts as a good reason (although I think it is incorrect in how it goes about doing this), it’s just important to point out because we’re also going to need to strengthen it.
Two directions that I think are reasonable to strengthen it are as follows:
You should be able to provide reasons for your actions to peers (that it’s to peers will be significant later) who want them.
Those reasons should remain relatively stable, in the sense that if they would apply in future then they should contain to describe your actions, or you should be able to provide a reason why they don’t.
What does “should” mean here? i.e. why should you act according to reasons in this sense? Kant and Velleman want this to be through logical necessity, but I don’t think that’s valid, and I think the weaker form that I am aiming at here is in many ways better. It allows for more empirically accurate descriptions of how ethics works, and in many cases by being weaker allows for stronger norms.
Anyway, my answer is that you should because life will go better for you if you do.
Norming and consequences
One of the things we do in our groups of peers is we create norms of behaviour. What is, and isn’t, an acceptable to behave. We do this by imposing consequences for that behaviour - we might chastise each other for bad behaviour, withdraw future opportunities, etc. e.g. you might not get a promotion if you behave badly at work, or you might just get reprimanded the first time, in an extreme case you might get fired. Similarly if you behave like a dick at a party I invited you to, you’re not going to get invited back. Extreme cases might also include e.g. physical violence, but most norming tends to be at least slightly more subtle than this.
Evaluating reasons is part of how we do this. When someone does something, especially something we disapprove of, we will often ask why they did it, and this factors in to our evaluation of their behaviour.
e.g. suppose we’re going to the movies together. I’m waiting for you at the cinema and you never show, even by the time the movie was supposed to have ended let alone started. You’ve not sent me a text, you’re not responding to mine. By default, I’m likely to be pretty pissed off at you.There’s likely some cultural variation in this, especially if you’re merely late - I’m a stickler for punctuality, but know people from places where turning up an hour late is considered completely normal. Regardless though, I’m not making a universal claim that it is right to be pissed off about this, I’m saying that I would be. Feel free to substitute your own example of someone letting you down.
It is likely that when we finally make contact, my first question is going to be some variant on “What happened?” - I expect you to provide a reason for the fact that you didn’t show when we’d previously agreed to.
Consider how I might respond differently to the following three answers:
- “I don’t know, no reason really, I just didn’t feel like it.”
- “I got the day wrong, was caught up in work, and didn’t hear the notification and completely forgot.”
- “I was on my way here and there was someone in a terrible accident and I had to stop and help and keep them conscious while waiting for the ambulance.”
These would change my response significantly. (1) would probably make me more angry at you. (2) would probably reduce my anger a bit, with how sorry you seemed about it and how many times you’d made this sort of error error mattering more than the actual reason. (3) on the other hand my anger would probably instantly evaporate, because I would agree that under these circumstances that was completely the right thing to do. Indeed, if instead you’d turned up and told me that there was such an accident but you didn’t stop to help because it would make you late to the cinema, I’d probably be angry at you for an entire different reason - of course you shouldn’t prioritise going to the movies with me over helping someone.
Part of this judgement is that we now have more context for the action, and we can now ask not just how we feel about this particular outcome, but how we feel about the process that lead to it. i.e. if similar circumstances were to occur again, and you were to behave in the same way in them, how would I feel about that?
This is a very lightened version of Kant’s categorical imperative. Instead of asking whether you are acting according to reasons that constitute universal law, we are asking wether in our particular relational context your reasons are ones I would consider acceptable if they consistently described your behaviour. Your behaviour is judged not based on its consequences, but on what the generalisation of it (as described by your provided reasons) looks like. By requesting reasons, we can judge you (positively as well as negatively) not just by your one-off actions but by how you tend to behave - i.e. what sort of person you are.
Trusting our peers
Of course, I can only judge what sort of person you are based on your reasons for action if the reason you provide is one that you are likely to act in accord with in future.
I am a big fan of C. Thi Nguyen’s paper “Trust as an unquestioning attitude”. My light generalisation of it is that you trust that X to the degree that you do not really have to entertain the possibility of not X. e.g. you trust that the sun will come up tomorrow, because even if you think about whether or not it will the possibility that it will not seems neither plausible nor emotionally relevant. You trust that your partner is not secretly a serial killer because if you think about the possibility it doesn’t seem remotely plausible. You also trust to a lesser extent that your long-term business partner isn’t going to betray you, your friends don’t secretly hate you, etc. - it’s not that there is zero concern here, and you may worry about it a bit, but it’s not a major part of your day to day experience.
Importantly, you trust that something will or won’t happen. You might trust totally that someone won’t betray you, but that doesn’t necessarily mean that you trust them to fly a plane.
You can still trust someone in a more generalised sense, in that the better you know them the more you trust about them. When I first meet someone I can’t trust that they won’t be a dick to me, or that they’re reliable, or that they are fun to be around, but as I get to know them better I learn the ways in which I can trust them (and the ways I cannot).
Afterword
I actually still think that there’s something important here and I’d like to try to revisit it at some point, but I think this kind of dissecting of someone else’s arguments isn’t really for me.
In particular the role of trust and reason in constituting our moral behaviour is something that I’ve been thinking about since and would like to try to articulate more at some point.