Tuesday, February 21, 2012

Aristotle on akrasia and Moore's 'open question' argument

To: Pat F.
From: Geoffrey Klempner
Subject: Aristotle on akrasia and Moore's 'open question' argument
Date: 26 February 2007 14:26

Dear Patrick,

Thank you for your email of 18 February, with your two timed University of London essays, in response to the questions:

'How in Aristotle's view can thought cause action? How on occasion may it fail to cause (correct or expected) action?'

'What force if any is there in Moore's 'open question' argument?'


The reason why Aristotle's discussion of akrasia is so important, is that he wants to salvage as much as possible from the Socratic doctrine that 'virtue is knowledge'. In the end, his view is not that far from Socrates.

The first question we have to answer is how can thought, any thought, cause action? Surely, thinking, 'such and such is the case' is one thing, doing is another.

Consider this example (which will be a bit more plausible when we come to consider akrasia):

1. Eating at least one portion of green veg a day is good for my health, and I want to be healthy.

2. I haven't had my green veg today, there is a portion of green veg on my plate, and I won't be offered any more food until tomorrow.

Therefore... what?

Aristotle wants to say that by contrast with a Universal syllogism, the conclusion of this 'practical' syllogism is an action. If I don't eat the green veg that is on my plate, I am acting inconsistently with the two premisses.

Well, I can think, 'I must eat the green veg.' But why should that thought lead to action? One answer is that, bearing in mind the meaning of 'I must...', to say, 'I must do X' and yet not do X is irrational.

OK, so what's wrong with being irrational? why is it necessary that any of our actions agree with our thoughts? Imagine someone whose actions do not agree with his thoughts. He doesn't do anything that he 'thinks' or 'decides' to do, but instead does a whole load of other things. The we would say that we are simply not dealing with an individual that one could recognize as an agent. It is an essential part of what it is to be an agent that one successfully executes practical syllogisms. The action component is not an extra bit added on, but a necessary part of the package.

I think that at the least what the examiner is looking for in the first part of the question is acknowledgement of this problem - the problem why any thought should issue in action. Aristotle's concept of a practical syllogism embodies the insight I have described above.

Because Aristotle is impressed with the core content of Socrates' view, the last thing he wants to do is allow that an agent can ever be faced with the choice between doing what the practical syllogism says, and simply 'acting out of passion'.

That's why Aristotle goes to some lengths to explain how we can 'know' something but not really KNOW it. For example, I 'know' that I should eat this green veg. My doctor's words to me last week are ringing in my ears ('serious iron deficiency, blah blah...') but faced with the unpalatable greens on my plate, I allow myself to dispute the very proposition I claim to know ('missing veg for one day won't hurt me' etc. etc.).

That is why I would question the second of the two alternatives that you offer at the end of your essay. Overriding passion can prevent a practical syllogism from taking place - as when a person acts out of 'impetuosity' - but when the agent does go through a practical syllogism, but fails to act, the explanation has to be in terms of a failure of his cognitive rather than affective capacity.

Hume famously remarked that reason is a 'slave of the passions', but by saying this he didn't escape the problem of weakness of will as such, as the greens example demonstrates. However, the problem of weakness of will is especially acute for anyone, like Aristotle, who holds that moral actions are dictated by knowledge and reason alone, and do not require an additional 'desire to be moral'.


I did struggle with this essay, although I won't chop your head off.

Perhaps your idea here is to question the validity of the analytic/ synthetic distinction as Quine does in 'Two Dogmas of Empiricism'. Let's see how this would work. All propositions are synthetic. 'X is good if and only if X increases pleasure' might be sufficiently embedded, in a given web of belief, to behave in every way as if it was an analytic proposition, the very last thing we would give up in the face of recalcitrant experience (like the laws of logic, which for the same reason are not immutable but ultimately capable of being revised - e.g. quantum logic).

In a world where everyone is a hedonist, the good IS pleasure. We don't ask whether pleasure is good because our only notion of good is what promotes pleasure, and no alternative has ever occurred to us.

Imagine a Nietzschean individual comes along and says, 'there is something is ultimately preferable to pleasure, which is overcoming oneself in the struggle to create values'. Maybe he is a lone voice crying in the wilderness, but if so this is only because his audience are too dim-witted to grasp his arguments. The fact that the good=pleasure equation is not questioned does not necessarily imply that it is beyond question.

What about the idea of 'simple concepts'? There is a way to define yellow - arguably the only way to define it - which far from showing that it is 'capable of decomposition ad infinitum' demonstrates why it is 'simple' in Moore's sense.

'An object is yellow if and only if it appears yellow to normal perceivers in normal conditions.'

This crude definition has to go through a lot of fancy refinements until it can be made reasonably secure against objections. But the point is that the term 'yellow' will always occur on both sides of the biconditional. You can't get rid of it. As soon as you start talking about wavelengths of light, you are changing the subject.

The term is 'good' is simple for a different reason. Anything can be good. There is nothing that good things have in common apart from their being good. However, the point of the term 'good' is not to describe an indescribable, in explicable quality which a thing either has or hasn't, but rather to state, to agents who are considering the possibility of different courses of action, 'this is to be preferred', or 'do this'.

In making his claim about 'Good', Moore is talking about the problem first given vivid expression by Hume - the gap between 'is' and 'ought'. This is pleasurable, but is it good? ought I to pursue it?

So let's now look at your water example, which could be interesting.

On twin earth water is XYZ. Therefore, intuitively, a la Putnam, what twin earthers drink, bathe in etc. is not water. It would be pointless to argue whether water IS H20 or XYZ because in referring to 'water' you are not referring to the same thing.

What about 'Good is what leads to happiness'? Let's say this is believed on earth while on twin earth the view is that good is 'overcoming oneself in the struggle to create values'.

Well, it's one thing to say that this is 'believed', but what could make it actually TRUE? In other words, how can this be seen as anything other than a philosophical disagreement about what things are good?

I assume that when, e.g. Nietzsche takes issue with utilitarians, both sides are concerned to persuade us to take a particular courses of action, make particular choices, prefer particular things. I certainly would not wish to rule out the possibility that an argument for a given moral theory could be valid, i.e. in the Aristotelian sense leading from thought to necessary action.

Moore is surely right that it can never be simply a natural 'fact' that such-and-such is good. But I would argue that it is fully consistent to hold this, while also allowing that there can be philosophical considerations which, when followed through, lead to the indisputable conclusion that actions of kind X, and only X, are 'good'.

All the best,