Tuesday, July 28, 2009


Rationality, like/possess it or not, plays a major role in determining the kind of life that we all lead. Rationality is that faculty of mind that we use to make decisions about day to day choices, what beliefs to hold and which world views to adopt (at least this is true for anyone that does not subscribe to radical cognitive-determinism). Alongside emotional and random brain processes, it is responsible for many of the decisions we make and thus shapes our life paths. As far as decision-making faculties go, rationality is a goodun': via processes of inference and deduction, our rationality helps us to make the best possible decisions with the (often limited) information that we possess.

Or at least that is what we might assume.

The prisoners' dilemma is a hypothetical game played between two people, each of whom seek the best possible outcome for themselves via rational decision-making. Imagine that two acquaintances are caught by the police trying to smuggle drugs into the country. They are placed in separate interrogation chambers and each told that when it comes to their trial they will be given the option of confessing or remaining silent. However, the resulting prison sentence will depend not just on their own decision but on that of their acquaintance. If both players remain silent there will not be enough evidence to charge them with drug smuggling and each will be convicted of the lesser crime of weapon possession and will land 1 year in prison. If both players confess they will be convicted of drug smuggling and each will land 5 years in prison. However, if one player confesses and the other does not, the player that confesses will be released with no sentence for turning evidence against his acquaintance, while the player that remains silent will go to prison for 10 years. The game is represented below.

Knowing that his acquaintance will have to make a decision at the same time as himself, what does the player's rationality tell him is the best decision to make? Notice that regardless of what the other player chooses, it seems that the best choice for a player is to confess. If player 2 is definately going to remain silent, then player 1 ought to confess because he will get to go free, as opposed to spending 1 year in prison. If player 2 is definately going to confess then player 1 ought to confess also because this lands him only 5 years in prison compared to the 10 years he would get if he remained silent. Thus, rationality tells each player that the best option is to confess.

Yet if we look at the game again, we see that the best outcome for the players does not occur if both confess. The best outcome for both players is if they both remain silent (they receive 1 year each as opposed to 5 years each). If rationality tells us what is best for us, how can it be so wrong in this situation?

There are many answers to this: rationality only works in a world in which there is perfect information (we don't have that here, neither player knows what the other is going to choose); rationality is not in fact the best guide for our actions; the most rational option is actually to remain silent, not to confess. Each of these has its merits (and flaws), but each has been touted before. I propose that to exercise our rationality properly, we must consult the 'moral law'.

Immanuel Kant, often considered the most important firgure in enlightenment rationalism (though he expounded empiricist tennets also), suggests that our rationality naturally brings us to the moral conclusion that we should always "Act according to a maxim which can be adopted at the same time as a universal law" (Kant, Groundwork for the Metaphysics of Morals). That is, we should always act in such a way that if anyone else were in our position, it would be best for them to act in this way also. What does this imply for the prisoners' dilemma? Well, we must ask ourselves, what decision is the one that would be most rational to make if everyone else were to make that decision also? Obviously, the answer is to remain silent: if all players were to make this decision also, this would lead to the best outcome for everyone. If all players remain silent, all players receive 1 year in prison as opposed to 5. (Of course, this outcome hinges on the assumption that all people arrive at Kant's morality via their rationality. We might not think this to be the case in the real world, but in the prisoners' dilemma we assume that each player reasons similarly.)

As a product of truly rational thinking Kant's categorical imperative (also referred to as "The Golden rule" [do unto others as you would have them do unto you], though this is not how Kant in fact preferred his thoughts to be interpreted), shows us why it is not in fact irrational to be rational in the prisoners' dilemma. One would only think that this was the case if they did not realise that moral thinking is in fact a part of rational thinking. But one can hardly reject the result that morality is a natural conclusion of rationality, for it does indeed produce the best outcome for each individual.

Why are such thoughts at all important or relevant to modern life? I hear you ask. Morality is so last century, I hear you sneer. I'm all about me, man.

The fact of the matter is, examples - and well reasoned solutions - like the prisoners' dilemma show that even if we are truly only concerned with ourselves (the selfish bastards that we are), it pays to be moral. Not only does society at large do better with a system of morality, individuals themselves do better than they otherwise would by following a rational system of moral conduct.

Hence, next time you're thinking about throwing shoes at your friend's wall, vomiting in someone else's bed or shitting in his or her shower, consider whether this is really in your best interests. If you do not heed the rational call of morality, you might find that your life is "solitary, poor, nasty, brutish and short" (Hobbes, Leviathan).

- J. A. Graham


  1. I think if you are trying to act in a moral way then you should confess to your crime regardless of the payoff you will receive.

    As a practical matter, how do you decide what should be a universal law? Big general cases like "Is stealing wrong?" are easy but what about "Is invading a country to remove a tyrant wrong?".

  2. Aww Thanks James. You're a geek, but that was still pleasant. Maybe i shouldn't say anything else but I'm going to, cause this blog needs more illiterate people trying to make points...

    Hey Giles

    you wouldn't confess to drug smuggling on moral grounds because drugs are cool.


    If a person really was a tyrant, and if invading the country to capture the tyrant didn't mean killing the innocent then isn't that as obvious as the stealing example. I mean, sure, it might not be possible to achieve, but I don't think proving what's possible is relevant. If the tyrant wishes to harm you then doesn't "do unto others as you would have them do unto you" apply inversely with equal merit? I mean, I know that's not what it's saying but fuck it, I have that moral stand point that no tyrant should be able to harm me.

    Also, who said stealing was "easy" to morally slot into universal law? Add any vairables and everything gets confusing. For example, who are you stealing from? Why are you stealing?

    See now everything is fucked up? Wasn't Robin Hood kind of a good guy???

    I mean, I don't know if there are any good arguments that cock up what James is saying, he probably studied them if there are anyway, but whatever, you seem to be off the mark with your examples.

  3. I think you're onto a good point, Adam. In simple cases, it seems like what's morally right is easy to figure out. But it seems like to get to those cases you'd need alot of information that we generally don't have (i.e. can we really invade a country without harming anyone else?).

    Giles, while you might be right that confessing to your crimes is morally right, the example tells us no information about whether the prisoners actually committed a crime; all we know is that they have been arrested and told that they will be put to trial. Also, Adam has a good point in that perhaps we think that drug-smuggling, though illegal, is not wrong. I, for one, do not think that it is wrong to disobey an unjust law (e.g. disobeying laws against homosexuality...not that I have done this, I just think such a law would be unjust).

    As for determining, practically, a universal law, Kant has done it for you: act in such a way that your actions could be decreed to be a universal law. That is, act so that everything would work out if everyone else acted that way also. For example, if I'm wondering whether or not I should steal this ipod, I must think "what would happen if everyone did the same thing?" and I would find that it would not be such a good outcome if everyone were to steal everyone else's things (think Hobbes' state of nature). Not only would society be worse off if stealing was to be universally approved off, but I, myself, would be worse off. Rationally, stealing is not something that should be done.

    The prisoners' dilemma points this out nicely, I think, because it shows that although I might think I can get away with some action and have it benefit me, the fact that everyone else is trying to do the same prevents me from getting the outcome I desire. Acting morally, then, procures a much better outcome for everyone.

  4. Actually, sorry Giles. In the post I actually did mention that the prisoners are "caught" trying to smuggle drugs. Ah well, just imagine that perhaps that is not in fact what they were doing. Anyway, why they are in that situation isn't important, its how they deal with it that is.

    If you don't like the example, we can reformulate it easily: think of two countries trying to decide whether or not they should goto war. If they both remain peaceful they are happy, though their neighbour still bothers them. If they both go to war they blast the fuck out of each other and everyone is messed up, while if one attacks and the other remains peaceful, the one that attacks receives the "spoils of war" while the one that remains peaceful gets mega-owned.

  5. I have always been a fan of Kant, especially as an alternative to Utilitarianism. The problem I have always found with him though is that, while the "golden rule" is a pretty good heuristic for determining right from wrong, and doing so by appeal to rational self interest (as to something arbitrary like 'the general good' as is the case of Utilitarianism) it starts to look pretty shitty when we consider the case of those who do not have the potential to harm us.

    The prisoner's dilemma is predicated on the possibility of either party in some way 'hurting' the other (as indeed is Hobbes analysis of the state of nature). It maintains that if both parties are equally capable of hurting one another, then it is in the best interests of both to just chill the fuck out. Where the problem arises, is when one party (or prisoner) simply cannot hurt the other. In such a case, the one who can act with impunity is 'rationally justified' in confessing. This is not a trivial point. In real life, it means we needn't consider treating those who cannot hurt us with respect (such as when Russia invaded Georgia).

    The problem gets worse if you look at, say, homeless people or orphans in third world countries. In many cases their governments simply don't care enough to bring to trial people who murder these social non-entities. Thus if you get some sort of pleasure out of it, then on a Kantian analysis there is no reason not to do some murders/rapes.

    And yet for some reason we intuitively feel like morality should still apply to these people. If Kantianism is true, and I'm by no means saying that it isn't, at the very least we have to accept that it doesn't accord with our basic moral impulses, which I take to be something of a downside.

  6. While I like your deconstruction of the Kantian position,I think you misinterpret the categorical imperative. It asks you to ponder what would happen it it were universally permissible for another person in your position to do exactly what you are contemplating doing now. What I think you have some 'beef' with is that if, say, I were to consider giving some kid the beat down, I might think to myself 'well, hey, if kids are gonna try giving me the beats, that's all good g, 'cuz I can take 'em with one hand, yo'. I think you are saying that when we consider an equal in our place, we are worried that if they did what we are contemplating, things just wouldn't be so good for us, but when we consider the non-equal, we can get away with doing what we want, because if they tried it, no harm would come of it.

    I think, however, that Kant meant us to assume that in every case of applying the law of universalization of actions we are not to consider the non-equal, but some unknown quantity. And since, as Hobbes noted, a group of men could overpower any given man, we have to assume that no man is any worse or better position than any other.

    This is precisely what the prisoners' dilemma implies, and it is exactly why Kant's categorical imperative works: we cannot assume that any action taken on our part could remove the threat of retaliation by any other (even if that 'other' is an amalgamation of several men).

    To demonstrate fully that the Kantian principle works (and does in fact accord with our basic moral impulses [though there is no reason why we should think that these impulses are in fact correct]) I take your example of contemplating doing some rapes.

    Picture, if you will, yourself in a position to rape a poor orphan girl while you are on holiday in India. You know that noone is around, that the girl has no family, and that noone could possibly track you down. To invoke the Kantian principle is to think "what if ALL other people could do what I am doing right now?". You could interpret this question in two ways: 1) what if all people were at liberty to rape whomever they felt like. In this case, clearly rape would be a bad thing because not only would everyone else be worse off for it, but you'd be nursing some nasty ass-sores for the rest of your life come rugby season. 2) what if all people were at liberty to rape ANYONE IN AN INFERIOR POSITION TO THEMSELVES SO THAT THERE IS NO THREAT OF RETALIATION? In this case, you might think that Kant would endorse the rape, however, there will always be someone superior to you (in strength, wits etc...here come your favourite Australian Rugby League team...) and if not one person, a group of people (...here come the New Zealand Police force...). Thus, on either interpretation, rape cannot be seen as a good thing to do.

  7. I think you are at least partly right, I did to some extent misinterpret the categorical imperative, at least insofar as Kant characterizes it. And I agree, what he is asking us to do is consider some "unknown quantity" or abstraction. His argument is largely based on a kind of "you wouldn't like it if it happened to you" kind of idea.

    What I was trying to suggest, was that this is not the same as being an actual motivational factor in real life, as opposed to in a thought experiment. Kant's argument, to me at least, seems to be that if we do something (e.g rape) it is only "right" or "just" or "rational" that we should expect the same thing to be done to us. This does not in actual fact mean it will be done to us (again, Russia is unlikely for various reasons to suffer a retaliatory attack for it's invasion of Georgia any time soon).

    This means that a completely self interested individual (such as Putin)has, in certain situations, no "actual" reason to refrain from harming others, even though they may have an ethical obligation. You are totally right, rape is not a good thing, even on a Kantian evaluation. What I wanted to suggest was that: "if", and only "if", there is no possible threat of retaliation whatsoever, we are rationally, if not morally, justified in pursuing a path detrimental to others.

  8. Ah, I see what you're saying now and I think you have a good point. However, Russia banks on the fact that Georgia will not band together with other states to defend itself. While this is probably the case for Georgia, it is not for other small/weak states. Think either of the world wars; small nations sided with larger nations and the costs to the protagonists were much higher than any benefits they got from going to war. That is, the costs of war make the prospects of starting one as rationally sub-optimal as they are morally abhorrent.

    It is for this reason that we think that a country is morally justified in going to war in aid/protection of another: in this way the global moral community can ensure that the costs of acting immorally (in terms of war) force nations to rethink their immoral behaviour.

    Nevertheless, one problem that I can see with the Russia vs. Georgia scenario is that we are dealing with nation states and not individuals. Though there seems to be a nice analogue between the two, I am not sure to what extent Kant's moral system can be extended to include group-actions (it might be that the categorical imperative works perfectly well at the state level, though I have not thought about it very much).

    Personally, I like the Kantian system because it prevents one from thinking overly selfishly and, if stated properly, I think it can be incorporated into otherwise seemingly selfish social positions such as libertarianism. What about you, Ben? What moral philosophy takes your fancy?

  9. Did Kant really think, with the whole do unto others as you would have them do unto you deal, that you require a 3rd party or a threat to keep you on the straight and narrow? If you need the idea of a real threat to stop you, you're not being moral, you're just being prudent.

    I always kinda interpreted it as how to be moral, not how to avoid punishment for being immoral. So I see a helpless orphan girl and by cricky what fun I could have! but then I put myself in her position to consider Kants point, and OMG! I empathise with her. I choose not to because it's not nice.

    Just saying cause the idea that you wont rape(or go to war) because there could be someone more adept in rape than you, who might just get cha in the bum, didn't seem like a moral stand point at all.

    And if Kant did mean that, fuck him, that creepy wanker. I don't want to hurt someone regardless of the threat to me. (To cover my bases and not pretend I'm a good person or something, alternatively, if i did want to hurt someone I'd be in such a state, pissed or otherwise, it would also be regardless of the threat to me)

  10. Lol, no Kant didn't think like that at all. He didn't use examples, and he didn't explain himself particularly well, he just noted that the categorical imperative/golden rule was the end result of rational thinking.

    What I was trying to show is that the prisoners' dilemma gives us a nice example of WHY it is rational to be moral. While we might think that it's nice not to do some rapes, this is more of an emotional response, it's not a rationalised response. My Kantian solution to the prisoners' dilemma, on the other hand, is a rational response because it shows you that being moral gets you to the best outcome.

    It's not so much about avoiding punishment, but more about gaining a better outcome than you otherwise could. If you think the prisoners' dilemma is a bit off (because its more about punishment than anything else), think about our reaction to global warming, for instance.

    Say I'm deciding whether it would be a good idea to reduce my carbon emissions and whatnot, while at the same time recognizing that other people are weighing up whether to make this decision (or not) also. We could use the Kantian framework to show why it would be both morally and rationally optimal to reduce our own carbon emissions. (Note, this only works if 1) you think global warming is a real threat 2) you think our reducing carbon emissions will actually reduce the threat, neither of which anyone necessarily has to agree with; I'm just using this as an example of a prisoners' dilemma where the story isn't about punishment, but about good/better outcomes).

  11. Personally
    I am on your side 100% when it comes to ethics. If you had to choose
    between one of the big 4 (Utilitiarian, Virtue, Kantian and Social
    Contract)then some kind of amalgamation of the last two (which you
    seem to have been kind of aiming for by invoking Hobbes, the Prisoner's
    Dilemma and Kant simultaneously) is probably the best we can do. I
    actually thought about what you were saying concerning nation states
    vs. individuals as I was writing before and I tend to agree. The fact
    that there are markedly uneven power relations in international affairs
    makes a Hobbesian social contract kinda more difficult to implement.
    Might be a different story if everyone had long rang nuclear missiles
    but I'm not really sure we want to go down that path! Your example
    (world wars) is possibly of limited significance if you look at some
    situations though. Say the members of the UN security council (or maybe
    the G20, or even just a list of the 100 most powerful military nations
    of the world) decided to band together and get their metaphorical rape
    on to some oil rich african nation and share the spoils, I don't think
    the risks are terribly high that they will suffer from reprisals. I
    mean, it is possible (terrorist attacks are a good example) but my
    point was kind of a hypothetical one, designed for a world in which you
    could "actually know" that there you would be safe. I think one big
    problem for Kantianism, and this is one which extends to pretty much
    all ethics (except perhaps social contract) is why we should consider
    morality binding. In Kant's original formulation, the only reason to
    obey the categorical imperative was that "reason" dictated it so. Fair
    enough. But in a post-modern world, where dickheads like Foucault start
    claiming that "reason" as a concept ought to be discarded as an
    oppressive colonial and patriarchal value used to subjugate others, you
    can see where it falls down: not every one recognizes it's authority.
    What you did in marrying Kantianism to Social contract theory I thought
    was fascinating, and actually a pretty novel idea. In my view though,
    on the question of bindingness, it is actually a purely social contract
    position, which I guess I was trying to point out. As I said, we only
    "actually" have reason to respect those who we would rationally have
    forged a social contract with.
    Because of this, I personally find social contract ethics a little
    abhorrent. I think it is better, when confronted with the problem of
    bindingness, to assume as a first principle that people do in fact want
    to be moral, and then use philosophy as a means to describe exactly
    what this morality consists of. Morality then becomes a contingent
    proposition: "if" you desire to act morally, "then" you should act in
    accordance with the dictates of (insert favorite moral theory here). (p.s sorry about the line spacing, my computer is being a dick)

  12. What I was trying to say is that it seems to me that the decision as to whether 'everything would work out if everyone else acted that way also' is not something that everyone will agree on. How do you actually define what 'everything working out' is? Is it minimizing 'suffering'? Maximizing 'freedom' or 'payoff'?

    I suppose I don't grasp how you are to decide whether you'd want something to be a universal law other than by some subjective weighing up of the consequences of everyone following it, which doesn't sound like a very universal principle.

  13. "I don't grasp how you are to decide whether you'd want something to be a universal law other than by some subjective weighing up of the consequences of everyone following it" - this is exactly what Kant was getting at. Everyone has to decide for themselves (based on their subjective principles) what could legitimately held up as a universal law (an objective principle). Of course, this willinevitably lead to problems in that people will get things wrong, and there will be disputes (as you rightly point out, is minimisiing or maximising the right way to go?). However, nothing in philosophy says that moral questions are easy to answer. I suppose all we can hope for is that we will get better at making moral decisions as we become more educated and gain experience.

  14. Oh yeah, the better outcome makes sense. Word G.