Tuesday, May 17, 2011

An Analysis of Threats

You are an intentional agent in the world. You have desires that motivate you to realize states of affairs in which the things that you desire have been realized.

You are surrounded by other intentional agents with their own desires (i.e, a “desire that Q”).

Furthermore, your “desire that P” does not, in itself, motivate anybody but you to realize states in which P is true.

How can you get others to act so as to realize states in which P is true or, at least, refrain from acting so as to realize states in which P is false?

I have already discussed one option.

Bargaining. “If you act so as to realize P, then I will act so as to realize Q.” You give the other person an instrumental reason to realize P – as a means to realizing Q. One problem with bargaining, however, is that as soon as one participant completes their part of the bargain, there is no more motivation for the other participant to complete their part. It loses its instrumental value. Some other motivation is needed – such as reputation or an aversion to breaking promises. Dealing with a person who lacks further motivation means it will be foolish to be the first to meet the terms of one’s agreement.

Today, I want to look at another option.

Threats. You find somebody with a desire that Q and say, "Unless you help to realize P, I will act so as to realize not-Q."

For example, "If you help me to rob this bank, or I will cause your child a great deal of pain.”

At first glance, threats are taken to be the opposite of bargains. However, a quick second glance shows us that they have a lot in common. Every bargain contains an implicit threat - "If you do not act so as to realize P, then I will not act so as to realize Q." Every threat can be expressed as a bargain - the proverbial "deal you can't refuse."

One of the problems with threats is that other agents also have motivation to make threats. You may find yourself faced with other intentional agents giving you the choice, “Either you act so as to realize Q, or I will act so as to realize not-P.” Such as, “Your money or your life,” or “The penalty for offending God or the King is death.”

The fact of the matter is – like it or not – threats have, and will continue to have, instrumental value. They will always be a means for people with a desire that Q, confronted with other agent with a desire that P, to get those others to realize states in which Q is true. This is not going to change.

So, let us assume that somebody is threatening you. If you do not act to realize Q, then he will act to realize not-P – where P is something that you want. You want your child to be free from pain, so the threatening agent says, “If you do not give me the money in your bank account, I will realize a state in which your child is not free from pain.”

One thing for you to look for is whether the agent has any reason to realize not-P other than your decision to help realize Q. Until you realize Q, then his restraint from realizing not-P has instrumental value – to motivate you to realize Q. Once you realize Q, then he loses any motivation to restrain from realizing not-P, and his other desires will dominate his action. If those other desires motivate him to realize not-P, then your realizing Q was for nothing.

Consider, for example, a bargain with your kidnapper. “If you give us $250,000, then you will get your child back.” You give them $250,000. Q, now, has been realized. They now have no more incentive to keep your child alive. It does not do them any good to do so – unless, somewhere, they have some other desire motivating them not to kill your child.

Just like with bargains, one possible motivation is reputation. If he wants to make useful threats in the future – to kidnap other people and collect ransom from them - it would be useful (have instrumental value to him) to be known as somebody who does not realize not-P when those he threatens realize Q.

However, the instrumental power of reputation requires that the agent wants to make future threats. If this is “the one big haul” through which the agent will be “set for life”, then it would be foolish to expect the threatening agent to keep their side of the bargain based on reputation. The same is true if the threat is made in secret, so that the threatening agent’s reputation cannot be affected.

Another possible source of motivation is an aversion to breaking promises. Here, threats are just like bargains. Whereas the instrumental value of the threat is what motivates the agent to make the threat, the aversion to breaking promises will motivate the agent to live up to their end of the bargain even after the person threatened has done what is demanded. If you can reliably determine if others have an aversion to breaking promises, you can reliably determine if the person threatening you will keep their end of the bargain after you have kept yours.

I have mentioned that the threat – or the restraint from doing that which was threatened - loses its instrumental value the instant you realize Q. It also loses its instrumental value the instant that you make the realization of Q impossible. If the agent threatens to kill your child unless you turn over the key to the vault, turning the key over to the threatening agent, or destroying the key, both eliminate the instrumental value of not killing your child. In many cases, the best option is neither to realize Q, nor to render Q impossible, but to stall and negotiate.

The last claim I want to make about threats in this post is to point out that, all things being equal, you have reason to surround yourself with people who have an aversion to making threats; or, at least, an aversion to threatening you. A person with an aversion to issuing threats will not come to you and say, “either you act so as to realize Q, I will act to realize not-P,” even when it would otherwise benefit them to do so. This is true in the same sense that a person with an aversion to pain will avoid states of affairs where he is in pain, even where he would otherwise benefit.

But, let's be honest, there will always be some people who deal in threat-making. Like I said earlier in this posting, threats will continue to have instrumental value. There will always be some motivation to generate threats. So, instead of the pipe dream of pursuing a universal aversion to making threats, perhaps an aversion to threatening non-threatening individuals will be more useful. There are a lot of details to work out as to exactly what this would look like. However, the general idea seems to make some sense.

And, as with bargains, people who can reliably detect whether others have this aversion to threatening non-threatening individuals will have reason to welcome those who have this property, and exclude those who do not. To obtain the benefits of belonging to such a community, it would be useful to acquire the property of having an aversion to threatening non-threatening individuals yourself.

No comments: