Thursday, May 26, 2011

Rewards and Punishments

Before I took a brief detour to discuss the rapture, I was posting about the fact that you are an intentional agent in the world with desires motivating you to realize that which you desire. You are surrounded by other intentional agents. However, the fact that you desire that P provides them with no motivation to realize P or to refrain from realizing not-P.

So, what can you do to get these other agents to realize P or, at least, refrain from realizing not-P?

I have looked at three options so far. Given an intentional agent in the community with a desire that Q, you might have an opportunity to realize P with any of the following:

(1) Bargaining: "If you act so as to realize P, then I will act so as to realize Q."

(2) Threats: "Unless you act so as to realize P, I will act so as to realize not-Q."

(3) Belief Modification: "The best way for you to realize Q is via Action A (which will realize P).

There are versions of each of these for getting other agents to refrain from realizing not-P.

In this post, I would like to discuss a fourth option.

Desire modification.

I want to begin by introducing another fact about those other intentional agents that you find yourself living with. For the most part, they each have a reward-learning system. The way this works is that, when an agent performs an act that produces a state called a "reward", their desire to perform that type of action gets stronger. And if an action produces a state of "punishment", an aversion to performing such an act grows stronger.

Furthermore, those other intentional agents have mirror neurons. This means that if Agent A experiences a reward or punishment, and agents B, C, and D witness it, then they will experience something very similar to the same state that A has experienced, with these same effects.

So, when A performs a type of action that would help to realize P, and you reward him in the presence of B, C, and D, then all four agents will likely acquire a slightly stronger desire to perform that type of act - contributing to the greater realization of P in the future.

The same is true if you punish A for acts that tend to realize not-P in the presence of B, C, and D.

In fact, you don't even need to have a real agent A performing a real action resulting in a real reward or punishment to have this effect on B, C, and D. Agent A might be a fictitious character, enduring fictitious rewards and punishments, while the community identifies him as somebody who would be worthy of rewards or punishments in the real world. The effect will still be to trigger the reward-learning system in the audience so as to promote some desires and inhibit others.

Using these tools, you have the ability to cause that intentional agent with the desire that Q to acquire a desire that R which, in turn, will help to realize P. Perhaps you can cause him to have a desire that P.

In my post on bargaining, I mentioned that bargains, where one person would complete their terms before the other, are doomed to failure. In these cases, you would be wise to seek out those with an aversion to breaking promises. And you have reason to acquire this property yourself so that others (with useful bargains to present) will have reason to seek you out.

Now, you have a way of promoting this aversion to breaking promises. By publicly rewarding and praising those who keep promises - and punishing and condemning those who do not, one can strengthen the desire to keep promises in the community at large. Stories in which the heroes keep promises even at great cost and villains break promises would also be useful.

The same methods can be used to promote an aversion to making threats and an aversion to lying.

Of course, you are not the only one in the community who has reason to trigger the reward-learning system to promote aversions to promise-breaking, threats, and lies. You should be able to convince a great many others that they have many and strong reasons to join you in this project.

Remember, false beliefs can seriously muck up this project.

If people get it into their head that eating with the left hand will offend the gods, who will punish the people with floods and famine, they might draw the false conclusion that they have reason to condemn - and to promote an aversion to - left-handed eating.

A fish-vendor might have difficulty convincing people to eat more fish, until he circulates a story about some divine power that will bestow blessings on the community whose people who eat fish on Friday - causing people to think that they have reason to praise those actions and punish non-compliance.

People, victims of foolish notions like prayer will bring rain or prevent terrorist attacks, or who come to think that God directs the course of hurricanes to punish the acceptance of homosexuality, might think that they have reason to direct the reward-learning system in directions that there is no real-world reason to travel.

It's rewards and punishments are unjustified.

Of course, the same is true of non-religious systems that are grounded on false premises - such as act-utilitarianism, Ayn Rand's Objectivism, intrinsic value theories, and any and all forms of social contract theory.

To avoid these unjustified rewards and punishments and the misdirection of our learned sentiments, we have reason to surround ourselves with people who have some aversion to making these mistakes.

People who have this aversion will be motivated to think twice about who they reward and who they punish. They will seek to double-check their work for possible errors. In the realm of punishment, they will want to presume innocence and will need to have guilt proved beyond a reasonable doubt. The greater the punishment, the greater the strength of these presumptions.

Just as we have reason to promote in others an aversion to unjustified rewards and punishments, they have reason to promote this aversion in us.

Now, we have a way to promote this aversion to unjustified rewards and punishments. We do this by praising those who seek solid ground for their rewards and punishments, while we condemn those who assign praise and condemnation recklessly.

3 comments:

Bradley said...

I think it would be better to think of praise and condemnation as two ways of promoting an aversion or affinity to something. This would help us to understand that we can use one without using the other. In general I would support using praise to encourage people to do things that are helpful to others, while we should condemn in order to discourage actions that are harmful.

David DeShetler said...

I'd enjoy an elaboration on your comments regarding act-utilitarianism. I've been studying it recently and would love to have your thoughts on it in a post or email if you have the time and inclination.

hippiasrex at gmail dot com

kvministries said...

Now, this is what I call a superb article. I wonder where you get so much resources from. Whatever, you have real power of words.