Thursday, May 19, 2011

Beliefs and Lies

You are an intentional agent in the world. You have desires that motivate you to act so as to create or preserve states of affairs in which the things that you desire are realized.

You are surrounded by other intentional agents with their own desires.

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 two options.

(1) 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.

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

In this post, I want to introduce another fact about intentional agents and discuss another way in which you might get them to realize P.

This fact us that, while desires motivate agents to realize states in which those desires are objectively fulfilled, their actions are mediated by their beliefs. If an agent is thirsty and believes that a pitcher contains clean, cool water, he will drink the water - perhaps discovering after the fact that his belief was mistaken.

So, as an intentional agent with a desire that P, one way you can get another intentional agent with a desire that Q to act to realize P is by altering his beliefs. If, given his beliefs, an act that realizes P will realize Q, he now has a motivating reason to perform the act that will realize P on his way to realizing Q.

This phrase, "reason to act so as to realize P" is intentionally ambiguous. He may be caused to intentionally realize P as a means to realizing Q. Or he may be caused to act in ways that realize P as an unintended side effect or realizing Q. Your desire that P is only concerned with the realization of P, not with how it is done (unless that is a part of P).

So, if you can convince another agent with a desire that Q to believe T, he will act so as to realize P. If you convince him to believe not-T, he will not act so as to realize P. Given these facts, in this simple model, your only motivation is to convince that agent to believe T. That is the only option that will realize P.

Is T true?

You have no motivation to even ask, let alone answer, that question in this simple model. The other agent's motivation to act depends only on believing T, not on whether T is true.

What is true of you in this case is true of every other agent out there giving you information. If convincing you to believe V will cause you to act so as to realize Q, then that other agent has a motivating reason to cause you to believe V. He has no reason at all to refrain from convincing you to believe V based on the fact that V is false - not unless this is built into Q or that agent has some other motivating reason to refrain from convincing others of falsehoods.

That other agent may have no motivating reason to consider the truth value of V in getting you to believe V, but you do. Your desire that P gives you a motivating reason to realize P. You use your beliefs to choose those actions (and inactions) most likely to realize P. You will act as if your beliefs are true. When they are not true, this will likely have an adverse effect on your ability to predict accurately. You might choose the action that realizes not-P.

You are thirsty. There is a glass of what you believe is clean, cool water free for the taking in the server tray. You drink from the glass. You are mistaken; it is not clean water. You end up being violently ill. If you had known that in advance, you would have never drank from the glass. Your lack of true and relevant beliefs would cause you to act in a way that you would not have acted if your relevant beliefs were true and complete. And, of the two actions, the one grounded in false beliefs was mistaken.

Your choice of actions that (you predict) will realize P will almost always (though, in important cases, not always) depend on having true beliefs. So, while that hypothetical other agent only has motivating reason to convince you of what will cause you to realize Q - whether true or false, you have a motivating reason to be convinced only of that which is true.

And while you only have a motivating reason to convince others of that which will cause them to realize P, they have motivating reasons to be convinced only of that which is true.

Now, while other people are motivated to tell you that which will help to objectively satisfy their desire that Q, what if their desires includes a particularly strong aversion to making false claims?

If such a person existed, that person would refrain from providing you with false information - or, at least, with information he thought to be false - even when he would otherwise benefit. Perhaps, when put up against something like his own aversion to death or the well-being of his child, he may be motivated to lie. However, where his aversion to lying is strong, it would take something like this to get him to lie.

Even here, a particularly strong aversion to lying would serve as a particularly strong motivation to find some other way - any way - to accomplish the same end without the lie. Here, too, the motivation is the same as that a person with a strong aversion to pain would have to find some option that promised not to involve pain before reluctantly settling on an option that does.

The instrumental value of identifying those with an aversion to lying would motivate agents to adopt methods that reliably identify agents as having or lacking this aversion to making false claims. This might include working with others to identify those who lack this aversion - perhaps identifying them as "liars".

And, if methods existed (e.g., by using the reward-learning system) to promote or strengthen this aversion to lying, your motivating reasons to acquire accurate information suggests that you employ these methods - and negotiate with others having the same interest - to promote this aversion to lying. A social institution for encouraging this aversion to lying, identifying those who do not have it, and labeling them publicly, could well be mutually beneficial.


Martin Freedman said...

You are delivering an excellent reboot of desirism.

One point I would taboo the subjective/objective satisfaction categories and I suggest, instead:

subjective satisfaction -> psychological satisfaction


objective satisfaction -> material satisfaction

Peter Hurford said...

Mr. Fyfe and "faithlessgod", I have a question about desirism that is unrelated to this post (but posting it in the most recent post should catch your attention):

Imagine there is this sadist, and the only way to stop her from kidnapping and torturing people is to lock her in jail forever (a realistic situation for a sadist who is unresponsive to other forms of persuasion).

Is this desire to incarcerate a good desire? While it does fulfill everyone's desire to not be tortured, it thwarts the sadist's desire to not be incarcerated. Therefore, it is a desire that fulfills some desires and thwarts some desires.

How would we evaluate desires like this from a desirism standpoint? What about other desires that both thwart some and fulfill some?

Is desirism really a theory that we ought to have desires that fulfill more desires than they thwart, or really only about fulfillment?

Martin Freedman said...

Hi Peter

There are quite a few posts that Alonzo has written in the past that covers your question. (Unfortunately he did not take advantage of the blogger label system which would make it easier to discover those relevant posts).

Anyway as I commented here, this latest series this is like a "reboot" of desirism. Part of this seems to be stating what others have criticised as "obvious". However the detailed expansion of these ideas is required as it shows that is all that is needed for solving the "problem of morality" as I put it. On the assumption that Alonzo either has or will or can give equivalent detailed expansions of the desires in your question I will just use the short-cut conclusions here but remember the same pattern of argument underlies these short-cuts.

People generally have many reasons to encourage an aversion to sadism or, equivalently, discourage an desire for sadism.

Similarly, people generally have many reasons to encourage an aversion to incarceration or, equivalently, discourage a desire for incarceration.

In order to promote a harmonious society people generally use the social forces of praise and condemnation, promises and threats, reward and punishment to encourage and discourage the relevant desires and aversions.

(An aside I find it interesting to note that I have used for a while in comments this 6 item phrase for social forces rather than the previous 4 item "praise and condemnation, reward and punishment" and I am pleased that Alonzo has explicitly developed and analysed the concepts of promises and threats in this recent series of posts, this needed to be done).

Such social reinforcement needs to be backed up by legal reinforcement - legal promises and threats - that are coherent and consistent with such social reinforcement.

For those whom such social and legal reinforcement fails there needs to be a system of legal rewards and punishment, which should also be coherent and consistent with such social reinforcement.

Still we have reason to promote a single standard of evaluation of transgressors and, also, a single standard of penalties for such transgressors.

So incarceration is "prima facie" desire thwarting but can and should be used - albeit reluctantly - for those whom social reinforcement fails, such as the sadist in your example.

However this does not justify retribution and treating them sadistically in turn, That would defeat the overall encouragement of an aversion to sadism.

The primary goal here is protection of others from the desire thwarting desires of the sadist.

Secondly, such incarceration should not be forever, we cannot assume that once a sadist always a sadist and should endeavour to rehabilitate them back into society.

(As to what are the appropriate means to achieve this, this is a matter of analysing the empirical records of the numerous different methods that have been tried in the past and present around the world and choosing the best available to use. Again we have many reasons to prefer using the best means of rehabilitation of, in this case, sadists, over inferior means or just retribution, - by doing so we can reduce the amount of sadists in society and in incarceration this way and so reduce the amount of desire thwarting desires (both sadism and incarceration of sadists) that deleteriously affect everyone (including sadists).)

If and only if this fails would it lead to their permanent incarceration but that still is no excuse to treat them cruelly and it is not an excuse to not try new methods of treatment as and when they become available and with their consent if needed, certainly if they wish to be rehabilitated..

Alonzo Fyfe said...


Imagine there is this sadist, and the only way to stop her from kidnapping and torturing people is to lock her in jail forever (a realistic situation for a sadist who is unresponsive to other forms of persuasion).

Is this desire to incarcerate a good desire?

The “desire to incarcerate” would not likely to be a good desire.

Now, we do have to take note of the distinction between “desires-as-ends” and “desires-as-means”. I can have an aversion-as-end to causing a child pain, but still subject a child to a painful medical procedure because it is a means to realize some stronger desire, such as a desire to secure the child’s health. The reluctance to cause the child pain would motivate me to look for some non-painful option, but would not necessarily block me from using the painful procedure if I had no choice.

A “desire-as-end” to incarcerate, it appears to me, would have the same status as the ‘desire-as-end’ to cause a child pain. It is something we would not want to do, but may be forced to do where no other option is available for obtaining a more desirable end.

This aversion to incarceration would motivate us to impose the least restrictive measures consistent with our desired end. That is to say, there is no need to make the patient suffer to any extent beyond what is strictly necessary to prevent others from harm, while at the same time offering treatment that would make even these restrictions unnecessary (if possible).

Promoting an aversion to something does not imply promoting an absolute reluctance to it on all possible circumstances. It implies giving that option weight as something to avoid – a weight that can be outweighed if other concerns are sufficiently great. It says, “look for options”. However, if other options are not available, and the need is sufficiently great, it says ‘Go ahead and, grudgingly, use the option that you do not like.’

Think of pain. To experience pain is not an absolute prohibition on that which is painful. It tells the agent to seek a non-painful option if possible. If no option is available, it says to check to see if the concern is important enough to outweigh the pain.

All things considered, I would prefer to be surrounded by people with an aversion to incarceration – an aversion that will motivate them to look for other options. However, the aversion would not be so strong that they would leave me vulnerable to those who would do harm, either by refusing to offer a deterrence, or refusing to confine those who are a threat in spite of the deterrence. It would be an aversion that motivates a presumption of innocence, an unwillingness to easily accept evidence against this presumption, but not an ability to ask when the evidence proves the agent to be a threat (and a sufficiently serious threat at that) beyond a reasonable doubt.

Peter Hurford said...

Thank you both very much for two comprehensive and well-worded responses. I believe both completely resolve the objection I made. I appreciate all the time you two are devoting to moral theory.

I think my confusion came from a confusion between desires-as-means and desires-as-ends, and that prima facie morally bad desires (desires that we ought to condemn because they prevent the objective satisfaction of other desires) can be morally good (in actuality, promote the objective satisfaction of other desires) because they are in support of good desires-as-ends.


As a somewhat related yet distinctly separate question, if I may -- in what sense does desirism not promote a utilitarian-style maximization of happiness, where happiness is defined as the quantities of desires fulfilled multiplied by the total strength of the desires among all people who hold that desire?