> Ethics is only normatively useful if it helps you make a decision, and
> decisions have deadlines. So imagine some very important decision you
> will make in a thousand years. There is some fact of the matter about
> what your ethical opinions will be at the time you have to make that
> decision. Anything that helps you learn about that fact helps you in
> making your decision. And knowing enough about those facts tells you
> everything useful that you could know. So the only question is how to
> best arrive at estimates of those facts.
I'm not sure in what sense it would be helpful. Suppose the situation was
a hard situation, a case in which things which we value highly were
weighed against one another in a difficult to resolve manner. (Any of the
lifeboat, trolley car, or harming+helping examples will do here.)
To give this example a little more bite, suppose that you found yourself
in one of the classical ethical dilemmas. For the sake of argument,
suppose you have to decide whether to allow a trolley car to continue on
its way and kill half a dozen strangers, or whether to redirect the train
down another track, where it will kill just one person whom you know very
well and love dearly. (Or whatever.)
Now suppose the ethical computer tells you that you will choose to allow
the trolley car to strike the strangers. (Or whatever.)
Will this give you any less consternation when the situation comes up?
Will you be able to sleep better at night knowing that you did what the
process told me you'd do?
Or maybe that wasn't the sort of help you had in mind. Maybe you mean
that you simply wouldn't have to think about the problem as much, because
you'd be able to offload some of the processing to the computer? But
isn't the computer just predicting the thoughts YOU'LL have to go through
in order to make this decision? If so, don't you have to think all of
those thoughts through anyway?
How exactly has your decision been made easier? Can you actually imagine
yourself doing what the computer said *because* the computer told you that
you'd do it that way? If so, why would you do that? If not, what exactly
has the computer told you that was helpful?
---Believe it or not, this is actually my computational argument written dramatically... though it would make the stronger claim that you COULDN'T make the decision *because* the computer had told you to do X, and that, therefore, the computation couldn't be done if that was how you'd intended to use it.
For suppose you've decided to do whatever the computer tells you to do. The computer notices that if it tells you to do A, you'll do A. But if it tells you to do B, you'll do B. But it will only tell you to do one thing, and you will only make one of those choices. But to know which it is, the computer has to know whether it will tell you to do A, or whether the computer will tell you to do B. But the computer can't know that, because that's what it's trying to calculate. So this computation can't be done.
-Dan
-unless you love someone- -nothing else makes any sense- e.e. cummings
This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:06:17 MDT