X-Message-Number: 10174 Date: Mon, 03 Aug 1998 22:57:48 -0700 From: Brian Manning Delaney <> Subject: Re: What you OUGHT to do References: <> (This is long: ~11KB or so.) Robert Ettinger wrote (in a Cryonet msg. not yet sent out to the List): >I posted the following today to Cryonet. My >guess is that there is more potential benefit >than trouble-making in posting to both groups. >Thanks to Dr. Delaney for his interest and his >apparent recognition that "what we ought to do" >is from some standpoints the most important of >all questions. I like the sound of "Dr. Delaney" enough that I'm impelled to wrap up my dissertation quickly! Thanks for the extra motivation! But, for the moment, just Brian (or Delaney, or my full name) is more accurate. :) >Brian Manning Delaney (#10170) wrote: >>Brook Norton <> wrote: >>>I'll restate that the underlying assertion is (borrowing >>>some from Ettinger) ** The only rational >>>approach for anyone is to try to maximize >>>personal happiness over future time, >>>appropriately weighted.** >>Hi Brook. I think you are wrong, or saying >>something empty. >>If you really believe that at "the most basic >>level, ... the brain is hardwired to always >>choose to increase happiness," then what you >>mean by happiness is simply what we choose. Thus, >>you're saying, at bottom: the only rational >>approach for anyone is to choose what we choose. [....] >First, it is not exactly true that the brain is >hard-wired to choose to increase happiness. >The brain, in Lorentz' metaphor, is a disorderly >parliament of instincts (and habits and >preferences etc.). "Choices" can arise in >various ways, not all of them the result of >balanced appraisal or cool calculation or >anything similar. (See my [cryonet] post >yesterday.) Nevertheless, Brook Norton is >basically correct, that our most basic value is >to maximize personal happiness (satisfaction, >feel-good, whatever you want to call it). I don't think I understand the significance of the difference between what is "not exactly true" and what is "basically correct." It appears to hinge on what you say below, which I'm also not sure I follow (though I suspect I follow, and just disagree). >It seems superficially reasonable to object, as >Dr. Delaney does, that the statement is circular >and meaningless--that saying we always choose to >increase satisfaction is the same as saying that >what we choose is what we mean by "satisfaction." >One way to understand the error is simply to >compare alleged or chosen criteria of value, and >ask "Why?" >For example, suppose someone says his highest >value is to improve the lot of humankind, >regardless of his own fate. We simply ask, "Why?" >It will develop that this is just what he wants. >That is his value because that is his value. >THIS is circular. Someone like Brook, on the >other hand, will say I want to feel good because >that is the way I am made, at the most basic >biological level. Feeling good is an end, not a >means. "Helping humanity" is a means, not an end. But as a response to the question of what to do, Brook's answer is still, as I said, ultimately circular, or, empty and therefore unhelpful. You've shown that the "means" answer isn't necessarily circular, but in such a way that you've shown it's empty, that is, nothing about the criteria for "satisfaction" applied to means helps answer the question of what to do (even if your analysis might say something about what our ultimate end unavoidably is). That was my point. In order to work out some of these differences, we may require a conversation that's far too long to have here, but I'll make another brief attempt. (I suspect a discussion of determinism is part of what's necessary -- but I'll skip that here.) The original statement was: "The only rational approach for anyone is to try to maximize personal happiness over future time, appropriately weighted." My contention is that this is deeply wrong (because empty or ultimately circular). The are a number of things that have to be dealt with for Brook's statement to work, among them: 1. That a rational approach is better than a non-rational or irrational approach needs to be shown. 2. That there are no other better rational approaches -- that is, that maximizing personal happiness is the most rational approach -- needs to be shown. This of course requires #3: 3. Personal happiness needs to be defined. 4. "Maximizing" needs to be defined. (Not one, but all of these have to be dealt with, along with other smaller problems.) Proving #1 is probably easy (feel free to ignore it, if you like -- though see the condition in next paragraph). Still, "rational" needs to be defined in a non-circular way. (That is, it won't do to say, "rational is what makes sense," or "rational is what reasonable people would agree on," etc.) You appear to want to demonstrate #2, and, at least to some extent, #3, largely by means of an understanding of the brain. This is an understandable approach. However, I think it's futile. I've suggested one problem already: than any hard-wired "engine of action" (even understood as Lorentz does) can't be called "rational" in a way that permits us to say that action X is more rational than action Y, unless rational is being defined in a way _radically_ different from everyday usage. The brain, after all, does irrational things. If a radical redefinition is needed, such a redefinition needs a warrant (and answering #1 will no longer be so easy, but will become more important). It's also difficult to see how happiness can be defined by means of an understanding of the brain. Part of my objection has to do with the difference between the mind and the brain. This is a long story, but one way of summing it up is this. The Churchlands, whose work many people reading this know (Paul and Patricia Churchland make claims about the mind based on scientific understandings of the brain), are very, very smart people, and do great work. But they aren't philosophers, in my book. They're cognitive scientists. The question of the definition of happiness -- especially as something claimed as relevant to the question of what to do -- is not a scientific question, however much science comes into play in helping us do the right thing, and however much science can help us understand certain mechanisms by which happiness levels change. (There's a much longer story here -- where isn't there in this debate, actually?!) About #4: >Dr. Delaney also appears to mistake the nature >of probability calculations about the future. In >order to reach a rational decision, it is NOT >necessary to calculate every consequence of >every possible choice out to infinity. Decision >theory, rather, is precisely the science of >making choices in the face of uncertainty and >limited information. I believe I appreciate the nature of probability calculations about the future, I just have a different understanding of whether probability calculations are relevant in the way you think they are. Consider the following two statements: S1. The right thing to do is to make choices that maximize happiness. S2. The right thing to do is to make choices that we assess as having the best chance of maximizing happiness. I take it A) you think the only relevant claim at hand is #S2, and B) you think it can be answered using probabilistic reasoning. (Correct me if I'm wrong.) To take (B) first: Even if the only relevant question is the second one, it's not clear that probabilistic reasoning helps. I can, to be sure, come up with a mathematically robust likelihood of my having an overall winning record at poker, played against a one particular person whom I know (and have played before), for a limited time. There are the odds of certain hands being dealt, the odds of the relative skill levels not changing, etc. All these things can be factored in. I can even, I think, come up with a fairly good estimate of the probability of my having an overall winning record if we played forever. Sure, my partner might take advanced smart drugs before I, etc., etc., but I can factor in the probabilities of such eventualities. It's not clear to me that we could _prove_ that a given calculation of probability of an outcome over the course of infinity is calculable, but I'm happy to grant that it is indeed calculable. The possibility of calculating my happiness, under the condition either 1) that I live forever, or, more likely, 2) that my current happiness is contingent upon my current assessment of future events (my children's well being, my grandchildren's well being, the continuation of human life beyond the time the sun goes nova, etc.), is an entirely different matter. The system is not closed temporally (and probably not spatially). The poker game isn't closed temporally either, but one might argue it's like the question of calculating the probability that flips of a coin will average (under fixed conditions) to the same number of heads as tails over time: the amount of time is irrelevant, we know. The calculation of future happiness, however, is not like the flipping of a coin (or rather IS, in a different sense...). There are a few arguments I could make to support this. The simplest is just to claim that the system is not closed spatially, as the flipping of a coin is (though even that isn't spatially closed, one could argue, but that's not relevant). Take this as a claim that we can't prove there aren't an infinite number of alternative universes, if you're inclined. There are other ways to show the system isn't spatially closed, but this is getting too long. In any event, my response to (B) is a very minor point, compared to the following. And now (A). I argue that #S2 turns out to be incoherent, or useless, because of its relation to #S1. I'll have to make some assumptions about what you're thinking (as I've already done!), which I'll happily see corrected (though I think the assumptions actually follow from what you've said). The main problem is that #S1 and #S2 contradict each other, and it appears that you've claimed both (or made claims that imply both): #S1 here: > [...] that our most basic value is >to maximize personal happiness (satisfaction, >feel-good, whatever you want to call it). and #S2 here: >Decision theory, rather, is precisely the science of >making choices in the face of uncertainty and >limited information. With #S2, you admit the possibility that someone could make THE correct probability assessment (I'm granting for the moment that which I argued above might not be possible), and yet it could turn out that a different course of actions could have resulted in a happier (or more satisfied, or whatever) life -- possibly MUCH happier. So then, by S2#, the person did the right thing, but by #S1 they didn't do the right thing, and perhaps even did the WORST thing. To get out of this problem, you either have to eliminate #S1 or #S2 (it seems to me). If you eliminate #S1, #S2 becomes incoherent, for obvious reasons: the right thing to do becomes the making of choices that we assess as having the best chance of doing something the complete success of which we've eliminated as our goal. If you eliminate #S2, then my earlier claims about the problems of calculating something into infinity become fatal, for we no longer have the goal of best assessment, but of best actuality, for which decision theory will not help us. This was too long. Sorry. I will probably not have time to add much more, but any responses will be greatly appreciated. >Yes, there is much, much more. Yes -- and hardly less so now, alas! Best, Brian. -- Brian Manning Delaney <> Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=10174