From: Wei Dai <[EMAIL PROTECTED]> To: Jesse Mazer <[EMAIL PROTECTED]> CC: [EMAIL PROTECTED] Subject: Re: Subjective measure? How does that work? Date: Sun, 25 Jan 2004 03:09:08 -0500
On Sat, Jan 24, 2004 at 11:49:09PM -0500, Jesse Mazer wrote:
> But measures aren't just about making decisions about what to *do*, the main
> argument for a single objective measure is that such a measure could make
> predictions about what we *see*, like why we see regular laws of physics and
> never see any "white rabbits". Although Bob can decide that only universes
> where gravity is repulsive matter to him in terms of his decision-making (so
> that he'd be happy to bet his life's savings that a dropped ball would fall
> up), he'll have to agree with Alice on what is actually observed to happen
> when a particular ball is dropped.
Well, when the ball is dropped, in one universe it falls down, and Bob has to agree with Alice, and in another universe it up, and Alice has to agree with Bob. Alice thinks the second universe is less important than the first, but Bob thinks it's more important. How do you break this symmetry?
By looking at my actual experiences, from a first-person point of view. If I drop the ball over and over again and consistently have the experience of seeing it drop, that suggests there's some kind of objective measure, independent of my own preferences, that assigns greater probability to versions of me experiencing the ball drop than to versions seeing it fly upwards.
An objective measure would seem to be the only way to capture the notion that there is something about my experiences that is externally given, and not just a matter of my own arbitrary choices. Even if I choose a weird measure for the purposes of making decisions, like one that tells me I should bet my life savings that I will be able to fly, I don't think that would change the probability that my next experience will actually be that of flying, which I assume will stay very low regardless of what decision-theory measure I feel like choosing. Again, there seems to be an external reality guiding the probabilities of different experiences which I have no control over. Do you disagree? Do you think that by choosing a different measure, you could change the actual first-person probabilities of different experiences? Or do you reject the idea of continuity of consciousness and "first-person probabilities" in the first place?
> Without an objective measure, I don't
> think there's any way to explain why we consistently see outcomes that obey
> the known laws of physics (like why we always see dropped balls fall towards
> the earth).
What good are the explanations provided by an objective measure, if I choose to use a different subjective measure for making decisions? How do these explanations help me in any way?
An objective measure would tell you the probability that you will actually have a particular experience in the future, even if that knowledge would have no influence on your decisions. But for most people, the probability of actually having future various future experiences probably *would* influence their decisions, no? For example, I am more likely to take a gamble that has a very low probability of leading to my experiencing pain in the future than one that has a high probability of leading to a pain-experience.
Learn how to choose, serve, and enjoy wine at Wine @ MSN. http://wine.msn.com/