Observer-Moments vs Observers

From: Hal Finney <hal.domain.name.hidden>
Date: Wed, 15 Jun 2005 10:07:23 -0700 (PDT)

Someone forwarded the following Time Magazine article to another list:

> Time: Special Body and Mind Issue: The Science of Happiness
> January 17, 2005
>
> The New Science of Happiness;
> What makes the human heart sing? Researchers are taking a close look.
> What they've found may surprise you
>
> Claudia Wallis, With reporting by Elizabeth Coady/ Champaign-Urbana; Dan
> Cray/ San Francisco; Alice Park/New York; Jeffrey Ressner/ Los Angeles

It is kind of interesting, but I particularly noted a segment which can
be thought of as addressing some of the issues of thinking of the needs
of observer-moments vs those of observers:

> Just last month, a team led by Nobel-prizewinning psychologist Daniel
> Kahneman of Princeton University unveiled a new tool for sizing up
> happiness: the day-reconstruction method. Participants fill out a long
> diary and questionnaire detailing everything they did on the previous
> day and whom they were with at the time and rating a range of feelings
> during each episode (happy, impatient, depressed, worried, tired, etc.)
> on a seven-point scale. The method was tested on a group of 900 women in
> Texas with some surprising results. It turned out that the five most
> positive activities for these women were (in descending order) sex,
> socializing, relaxing, praying or meditating, and eating. Exercising and
> watching TV were not far behind. But way down the list was "taking care
> of my children," which ranked below cooking and only slightly above
> housework.
>
> That may seem surprising, given that people frequently cite their
> children as their biggest source of delight--which was a finding of a
> TIME poll on happiness conducted last month. When asked, "What one thing
> in life has brought you the greatest happiness?", 35% said it was their
> children or grandchildren or both. (Spouse was far behind at just 9%,
> and religion a runner-up at 17%.) The discrepancy with the study of
> Texas women points up one of the key debates in happiness research:
> Which kind of information is more meaningful--global reports of
> well-being ("My life is happy, and my children are my greatest joy") or
> more specific data on enjoyment of day-to-day experiences ("What a
> night! The kids were such a pain!")? The two are very different, and
> studies show they do not correlate well. Our overall happiness is not
> merely the sum of our happy moments minus the sum of our angry or sad
> ones.

The point here is that the actual OMs you are experiencing may be
unhappy, but you may have a sense as a person, an observer, that you
are in fact happy. Now, in this particular case, I think it is more a
matter of social pressure to say that you love your children and they
bring you joy. Everyone knows you are supposed to say this and think
it, and so people do say this and maybe even convince themselves that it
is true. Someone who said, no, my children are a negative in my life,
they are a source of stress and a total pain to deal with, would be
thought something of a monster. No one wants that for themself.

But I think the overall point about OMs versus memories as an observer
still applies, and the article goes into it further:

> This is true whether you are looking at how satisfied you are with your
> life in general or with something more specific, such as your kids, your
> car, your job or your vacation. Kahneman likes to distinguish between
> the experiencing self and the remembering self. His studies show that
> what you remember of an experience is particularly influenced by the
> emotional high and low points and by how it ends. So, if you were to
> randomly beep someone on vacation in Italy, you might catch that person
> waiting furiously for a slow-moving waiter to take an order or grousing
> about the high cost of the pottery. But if you ask when it's over, "How
> was the vacation in Italy?", the average person remembers the peak
> moments and how he or she felt at the end of the trip.
>
> The power of endings has been demonstrated in some remarkable
> experiments by Kahneman. One such study involved people undergoing a
> colonoscopy, an uncomfortable procedure in which a flexible scope is
> moved through the colon. While a control group had the standard
> procedure, half the subjects endured an extra 60 seconds during which
> the scope was held stationary; movement of the scope is typically the
> source of the discomfort. It turned out that members of the group that
> had the somewhat longer procedure with a benign ending found it less
> unpleasant than the control group, and they were more willing to have a
> repeat colonoscopy.
>
> Asking people how happy they are, Kahneman contends, "is very much like
> asking them about the colonoscopy after it's over. There's a lot that
> escapes them." Kahneman therefore believes that social scientists
> studying happiness should pay careful attention to people's actual
> experiences rather than just survey their reflections. That, he feels,
> is especially relevant if research is to inform quality-of-life policies
> like how much money our society should devote to parks and recreation or
> how much should be invested in improving workers' commutes. "You cannot
> ignore how people spend their time," he says, "when thinking about
> well-being."
>
> Seligman, in contrast, puts the emphasis on the remembering self. "I
> think we are our memories more than we are the sum total of our
> experiences," he says. For him, studying moment-to-moment experiences
> puts too much emphasis on transient pleasures and displeasures.
> Happiness goes deeper than that, he argues in his 2002 book Authentic
> Happiness. As a result of his research, he finds three components of
> happiness: pleasure ("the smiley-face piece"), engagement (the depth of
> involvement with one's family, work, romance and hobbies) and meaning
> (using personal strengths to serve some larger end). Of those three
> roads to a happy, satisfied life, pleasure is the least consequential,
> he insists: "This is newsworthy because so many Americans build their
> lives around pursuing pleasure. It turns out that engagement and meaning
> are much more important."


So here the two philosophies are expressed more plainly. Do we
care about the OMs, trying to make each one as pleasant as possible?
Or do we go for how happy people *think* they are, how happy they claim
to be, based on their memories of how things went? This leads to the
paradoxical situation where someone who has a lot of unhappy OMs but does
not remember them well (like in the colonoscopy experiment described
above - talk about "more torture"!) is considered to be happy, by the
observer based perspective.

I do have a new idea along these lines. It is something of a middle
ground between the two. It starts of course with the principle that
the measure of an OM is 1/2^KC where KC is its Kolmogorov complexity.
This should be unproblematic for those who adopt this general approach
to understanding the Multiverse, the Schmidhuberian/Tegmarkian method.
The measure of ANYTHING is 1/2^KC and OMs are no exception. This is the
basic principle which I elaborated on in my Observer Moment Measure from
Universe Measure posting.

Okay, so the new idea is simply that the KC of an OM depends not just on
how hard it is to instantiate it (ie how big a program it takes to create
it in a universe) the first time, but on also how big an "echo" it makes
in the universe. This is the same basic idea I described earlier which
causes bigger and slower instantiations of OMs to have higher measure.
But this applies as well to the impact an OM makes on the universe.
An OM with a bigger impact adds somewhat to its measure, because such
an OM can be output by a shorter program (as I outlined earlier).

The most straightforward case would be an OM which is remembered. It has
echoes down through time. Such an OM has higher measure by virtue of
its existence not only the first time, but as an echo in subsequent OMs.
A more typical OM which is immediately forgotten, in contrast, has no
such echo and will have a lower measure.

This allows us to synthesize the two approaches advocated above by the
researchers Kahneman and Seligman. With Kahneman, we will seek to make
OMs happy. But with Seligman, we will pay more attention to those OMs
which get remembered. Those are the "peak moments" which he argues make
a bigger difference in terms of "real" happiness. This allows us to
make a tradeoff where we let typical OMs be less happy in exchange for
"peak" OMs being happier.

The main question is how much extra measure an OM should get by virtue
of being remembered. How much smaller can a program be to output such
an OM? I don't know if we understand consciousness well enough to
begin to answer this question. But in principle it will be calculable
once we have a good theory of consciousness and brain function. Then
we will know how to properly balance the two perspectives of OM vs
observer.

Hal Finney
Received on Wed Jun 15 2005 - 13:57:49 PDT