Saturday, September 16, 2006

Inside the Morality Room

Suppose I design a Morality Room to test your feelings about the morality of certain actions. In the room, I screen a movie showing an act, the consequences of that act, and the consequences of not acting. You then press the "Good" or "Evil" button to indicate whether you feel the act was good or evil.

So, when I show you a movie of your foiling a handbag-snatching attempt, and this leading to the thief's immediate arrest (instead of his striking again), you press the "Good" button.

[We can imagine variations on this theme, with movies showing multiple actions and scoring them each relative to one another. For now, the simple version will suffice.]

The Morality Room is a useful tool for illustrating how we can be confused when we consider counterfactuals.

The classic one is "how would you feel if you were aborted and never existed?" The movie shows you the alternatives, namely life as you know it, then life without you. By now, you have probably noticed the problem. You cannot objectively answer whether your non-existence was good, when you are the one existing to provide the answer. Your answer presumes your existence.

Another question is "would it be okay if you were murdered, assuming that you did not suffer, and the memories of others were erased to ensure you did not suffer?" Or, equivalently, "would it be okay if you ceased to exist if no one suffered directly on account of your non-existence?"

Again, this question cannot be answered by the room. You, the judge, will suffer when you see the movie, so you must answer that the act is evil.

What can we take from this? Well, I'm not saying that we ought not think that "disappearing" people, even painlessly, is not evil. It feels pretty evil to me. Instead, what we learn is that we cannot rationally reach the conclusion that an act is objectively good or evil when we are hypothetically indifferent to the act. We cannot objectively say that murder is wrong even when it happens to someone we don't know, when there is no suffering, and when we are unconscious of the event. It's too late, because we are already in the room and we've seen the movie. We can't be unconscious of the event and still answer the question.

This isn't to say that subjective morality isn't perfectly adequate. It is.
P.S. I can't have been the first person to think of this thought experiment. Anyone know of a reference to something more original?

10 comments:

Colin Caret said...

doc: "The classic one is "how would you feel if you were aborted and never existed?" The movie shows you the alternatives, namely life as you know it, then life without you. By now, you have probably noticed the problem. You cannot objectively answer whether your non-existence was good, when you are the one existing to provide the answer. Your answer presumes your existence.

I'm not sure how to read this conditional. There are at least two ways of doing so. On one reading, it is perfectly innocuous. Read it this way: "Do you feel the world would have been better or worse if you had never existed?" This way of reading the question is perfectly sensible because it only assumes your actual existence. It doesn't assume your existence in the possible alternative history where you were not born.

Now, we could also read it this way: "If you had never existed, how would have felt about it?" Of course, this makes no sense because it is asking you (actual you, reading the question) to speculate about the features of some non-existent object in some alternative history. Clearly non-existent objects have no features since they don't even exist, so it makes no sense to speculate in this vein.

Okay, now that we have clarified that, the question is: what does this tell us about morality? So far as I can figure, it tells us nothing at all about morality. All it tells us is something about the logic of existence, property instantiation, and the way that language and thought experiments can conspire to create paradoxical statements or commands. Maybe you can expand on how this is connected to morality.

Doctor Logic said...

Hi Colin,

I certainly wanted to dispell nonsensical questions like those of your second reading.

The relavance to morality has to do with subjectivity. Many people argue that moral claims are beliefs about something objective and not merely feelings or preferences. In making this case, they try to imagine scenarios in which the judge is supposed to isolate himself from personal feelings about the moral question.

In the typical case, they suppose that a person has been subject to an act that painlessly ends his life, but no one suffers during or after the event (e.g., due to memory erasures of people who knew him).

According to the assumptions of the case, there was no suffering. Since almost everyone agrees that such acts are generally immoral, they conclude that morality is fixed independently of our emotions or feelings. That is, moral claims are beliefs about something other than how we feel about things.

The room demonstrates that this is false because the room violates the assumptions of the case under study. We, the judges, are conscious of everything that happened, and we would suffer while viewing/imagining the acts. Thus, moral beliefs are beliefs about our own emotional reactions, and not about some external moral reality.

Colin Caret said...

I'm afraid I'm not getting it. Let me try to formulate what you are saying in my words and you tell me if it comes out right. You seem to be saying that the reason most people think of morality as an objective discourse is because they reach moral judgments by way of thought experiments where they assume the role of an impartial observer. You claim that your thought experiment violates this standard reaction in some way by demonstrating the subjectivity of a particular moral judgment.

First, I don't get the last part. Why does considering this thought experiment demonstrate that moral judgment is subjective? I am completely missing that connection.

Second, and more importantly, I think your explanation for people's belief in objective morality is mistaken to begin with (if I have understood your explanation correctly, which might not be the case!). I think that people believe morality is objective because of the form of the discourse and the related features of correct use of the discourse. For example, when we state moral judgments they characteristically have the same structure as any other objective claims. I say "it is wrong to kill" and I appear to predicating a property to an act type. That seems to be of the exact same kind of phenomena as what I describe when I say "my car is blue". Since the latter is objective, so is the former. Also, when we assert moral judgments we do so categorically, that is, in such a way that they range over everyone. So my claim that killing is wrong implies that it is wrong for all people, regardless of their personal feelings. These features of moral discourse give every impression that the subject matter of morality is objective.

Nevin ":-)" said...

The classic one is "how would you feel if you were aborted and never existed?"

Isn't that just the plot of It's a Wonderful Life?

Every time a bell rings an angel gets his wings...

Doctor Logic said...

Colin,

I agree with your second point about the psychological reason for beliefs in objective morality being due to the features of moral discourse.

However, I'm not really looking for a theory to explain why we think morality is objective when it isn't. I'm looking for a test of objectivity of the form "X is objective if we can find a way to meet conditions Y on X." Specifically, I regard the truth of X as objective if X is true whether I feel it ought to be true or not. For example, given the axioms of arithmetic, 2 + 2 = 4, whether I like it or not. Yet, when I say that "french fries taste good," I am saying that I like the taste of french fries, or possibly the subjective statement that I think everyone ought to find them tasty. The taste of french fries is about me, not purely about the french fries.

For morality, objectivity would require that we have a way to test a moral claim without reference to how we think things ought to be. Of course, this doesn't exist.

Typically, the moral realists I debate with (mostly theists) will respond by trying to isolate beliefs about moral truths from personal moral feelings. For example, they will argue that, if I believe that murder is wrong even when I hypothetically don't know about the crime or the victim, then my moral belief is somehow not about how I feel about murder. My thought experiment demonstrates that in order to answer the question, I effectively have to become aware of the crime.

So, the thought experiment doesn't demonstrate that morality is subjective per se. It defeats an argument that morality is separable from personal taste.

Doctor Logic said...

Hi Nevin,

I've always disliked that movie! Now, I know why. :)

Ken Brown said...

doctor(logic),

You, the judge, will suffer when you see the movie, so you must answer that the act is evil.

You are putting the cart before the horse. On your view, the problem with murder is not that a unique human person has been irretrievably lost from the world, but merely and only that the loss makes you feel bad. But the loss does not matter becase you feel bad about it; rather, you feel bad about it because the loss matters.

To deny this is to make personhood itself an arbitrary criteria. If it is the feeling itself that makes it wrong (to you), then there is no reason why personhood should be the basis for rights (as you claim), it could just as easily be IQ, or athletic ability, or race, or being named Dave, or anything else that one might happen to care about. None of these would be any more nor less an arbitrary basis for rights, than any other.

Doctor Logic said...

Ken,

But the loss does not matter becase you feel bad about it; rather, you feel bad about it because the loss matters.

I realize that this is your position, but you have no evidence to substantiate it. You only claim some externality is evil by how you feel about it, but that means that it is no different from any other subjective opinion. It can only be an objective statement about our subjective taste.

To deny this is to make personhood itself an arbitrary criteria.

This is not quite true. It would not be so much arbitrary as subjective. I don't arbitrarily prefer Coca-Cola over Pepsi. One definitely tastes better than the other.

However, rules about personhood are only generalizations and approximations to my moral opinions. For example, I think it might be okay to kill Osama bin Laden despite the fact that he is a person. However, as a generality, I feel it is wrong to kill persons (all other factors being equal).

...it could just as easily be IQ, or athletic ability, or race, or being named Dave, or anything else that one might happen to care about. None of these would be any more nor less an arbitrary basis for rights, than any other.

Again, not quite arbitrary. I don't happen to feel that Daves should have special moral status above or below Burts or Sarahs. Unless Dave is my brother, in which case I might prefer that he have a higher moral status.

Again, your claim is far stronger than mine. You claim that, though we only know right and wrong by our feelings, that right and wrong are not subjective. Well, you need to find evidence for this. I'm not saying it could never ever be found (though I'm close to that opinion), just that not a sliver of evidence has yet been found.

Ken Brown said...

doctor(logic),

Frankly, I just don't understand what you mean when you say we lack evidence of moral objectivity. Our feelings are evidence of moral objectivity, just as our other senses are evidence of physical objectivity. You don't reject the external world because our only way of seeing it is through our eyes!

All I can figure is that you think the experience of one sense must be corroborated by the experience of another sense, for us to accept it. So if I see something I can also feel, it is more likely real than something that I can only see. Since we can't see or hear morality (we can only feel it) it must not be real.

Is that your claim? Because if so, it seems open to simple counter-example. We can only see the stars (we can't feel them or hear them or taste them), but surely they are real.

The only other possibility I can figure, is that since we can build machines that record light, therefore sight is real. Since we haven't built machines that detect moral feelings, morality must not be real.

Is this your claim? Again, it seems easily defeated. What kind of sense is our moral sense? Is it not mental? So if we were going to build a machine to detect morality, wouldn't it have to be a mental machine - an AI? Perhaps you think morally reasoning computers are an impossibility? Otherwise, you would have to admit that morality is, in some sense, objective.

Doctor Logic said...

Ken,

I wonder, based on your last response, whether anything is subjective by your definition. Taste in food, color and music meet your definition, too. Are you a realist about everything?

Surely, objectivity requires that something be discernably true or false independent of our personal tastes. This is true of location, temperature, chemistry and so on. The way we know that these things are independent of taste is that we establish tests for truth about these things, and conduct those tests in such a way as to isolate the results from personal taste. Implicit in this is prediction because we need to predict what test results would be appropriate to the thesis in advance.

Creating an AI doesn't test for objectivity beyond personal tastes because we've simply created another person with personal tastes.

So how do we test whether moral taste is measuring something objective? How would we test whether murder was wrong, independent of our moral distaste for it? We cannot do so.