In a tweet directed at David Deutsch, and some other people, a tweeter asks:
What would a decision-making agent base its decisions on if not entrenched goals?
I am going to take it that a goal means some specific fixed objective. A goal involves taking some specific kind of action. ‘Choose the best thing to do today’ is not a goal since it makes no specific claim about what you should do. ‘Go to see a film at a cinema today’ is a goal since it makes a specific claim about what you should do.
The answer to this question is that a rational person making a decision will not base his decision on anything. From a conventional point of view, this sounds ridiculous, but that conventional point of view is wrong.
Making a decision involves creating knowledge about what to do next. As such, before you can understand decision making you have to understand epistemology (the theory of knowledge) more generally. I’m not going to explain the whole of epistemology, but I will outline epistemology, explain how it is relevant and point you to where you can learn more.
Philosophers often say that knowledge is justified true belief. Justification is a process that allegedly shows an idea is good or true or something like that. People who believe in justification might hedge a bit and say it shows an idea is probably a good idea, or probably better than the alternatives. This sounds superficially like a reasonable position: who would want to act on an idea that hasn’t been shown to be true or good? The idea that you need to have a goal to make a decision assumes that it is possible and necessary to justify your decision.
The apparent reasonableness of this position is spoiled by the fact that justification is impossible and unnecessary. The correct alternative is to focus on solving problems, not on justifying your decisions. Saying you’re going to solve problems puts you in a different position from pursuing a goal. There is no fixed standard by which you judge every decision. Rather, you look for problems with your current options and try to solve the problems. If some goal you thought was good turns out to be problematic, you can discard or modify it and you should.
Problems with justification
To understand the problems with justification, you have to understand something about how arguments work. Some arguments are informal and are not really candidates to prove anything. The fact that people are prepared to make informal arguments, and sometimes to take such arguments seriously, are problems for the idea of justification since such arguments aren’t justified. But even formal arguments don’t allow justification. Any formal argument starts with some assumptions and rules for getting conclusions from those assumptions. If the premises are true, and the rules reflect those that hold in reality, then the conclusion is true.
An example of a formal argument. If I am in the House of Commons and the House of Commons is in London, then I am in London. I am in the House of Commons, so I am in London. The rule being used is that if place A is contained in place B and object C is in place A, then it is also in place B. The assumptions are that I am in the House of Commons and the House of Commons is in London. Now, to show that the conclusion is true you have two options:
- Say that the rules and the assumptions are correct by fiat.
- Show that the assumptions and rules are correct.
Option 1 has the problem that if you admit it, then anyone can claim to prove anything by saying the assumptions are true by fiat. You can say the Earth was created 6000 years ago by saying the Bible is true by fiat. But surely we can’t just say the Bible is true by fiat because it makes ridiculous claims about some guy turning water into wine and stuff like that, but then you’ve adopted option 2. But you can see whether I’m in the House of Commons but you can’t see a lot of the stuff in the Bible because it’s abstract. There are two problems with this. First, to properly understand whether you can see me in the House of Commons you have to understand the physics of eyes and that involves abstractions. Second, the rule itself is abstraction you can’t see. So if you reject anything you can’t see you must reject the rule and the argument falls apart.
Option 2 leads to a different problem. If you’re going to show the rules and assumptions are correct you need to make another argument that justifies them. And that argument will have assumptions and rules that have to be justified. And then you have to make more arguments to justify the rules and assumptions of your new argument. And you have to keep repeating this process indefinitely, so you can never actually justify anything.
Saying that justification can make do with showing your conclusion is probably correct doesn’t solve this problem. Your conclusion is only probably correct if the assumptions and rules are probably correct, which leads to the same problem. In addition, there is no such thing as a theory that is probably correct. Your ideas are either right or wrong. And probabilities of events not of theories and can only be obtained from theories such as quantum mechanics that are themselves either right or wrong. There are other problems with assigning probabilities to theories, see The Beginning of Infinity by David Deutsch Chapter 13.
When a person thinks he has justified his ideas, in reality he has made assumptions and used rules that he has not justified. Those rules and assumptions could be wrong. Any viable epistemology has to take account of the fact that any idea you hold could be wrong. That includes ideas that you think are certainly correct. People often think an idea is obviously correct when it turns out to be flawed on closer inspection, such as the idea that justification is necessary and desirable.
The alternative to justification and goals
So if you don’t justify your ideas, including your decisions, how do you create knowledge rationally? You start with a problem. You guess solutions to the problem. You criticise the guessed solutions until only one is left and you don’t know of any criticisms despite looking for them. The surviving idea is the solution to that problem. You then move on to a new problem. This solution was first pointed out by Karl Popper, David Deutsch and Elliot Temple: it’s called critical rationalism.
For more details on critical rationalism, see Objective Knowledge Chapter 1 by Popper, Realism and the Aim of Science Chapter I by Popper, ‘one the sources of knowledge and of ignorance’ inConjectures and Refutations by Popper, chapters 3 and 7 of The Fabric of Reality by David Deutsch, most of The Beginning of Infinity by David Deutsch and Critical preferences and strong arguments by Elliot Temple.
So how do you apply this to making decisions? Your decision making has to start with a problem you’re trying to solve. You might be trying to decide what to have for breakfast. You then look for solutions to the problem. You could have cereal, or boiled eggs or whatever. Then you look for problems with the options. You might not have enough time to make and eat boiled eggs, so you pick cereal. So then you’ve solved the problem by picking cereal.
But your breakfast decision could go very differently. You might find that when you wake up you’re not hungry. So then you might think eating is pointless and you decide not to have breakfast at all. So you had a goal when you started the problem: the goal of having breakfast. And you ditched that goal because you had a criticism of it. If you had looked on having breakfast as a goal you must fulfil, you would have missed the option of not eating breakfast. So thinking of decision making in terms of goals is an obstacle to making rational decisions.
The misconception that you need to have goals is one of many misconceptions that can get in the way of making rational decisions. And you can’t expect to get rid of all your misconceptions without rational discussion, so you may want to read Fallible Ideas and contribute to the e-mail list linked on that page.