Here's my latest trick for this one. It vaguely and metaphorically
resembles a mathematical generalization over possible values.
Aristotle distinguished between extrinsic values and intrinsic values. He
thought intrinsic values were better because extrinsic values "partake of
drudgery". What bothered me about this is all the "higher values" like
truth and justice usually turn out to be generalizably extrinsic. Living
according to the "higher values" turns out to be useful in getting all
sorts of petty little personal preferences accomplished, and flouting the
"higher values" tends to screw you over.
And even though the quest for increasingly excellent scientific theories
turns out to be extremely generalizably extrinsic, those who contribute to
that quest hardly think of their activity as "drudgery". Usually it's
something that they come to value intrinsically as well, especially the
best of them.
Anyway (sorry about the laborious background): let's say there's something
you want, x, doesn't matter what it is. Now in general, for any possible
goal x, a better understanding of your environment will help you attain it.
Harmonious (and profitable!) interactions with your fellows will usually
help you attain it. (Minsky discusses this kind of thing in the "Functional
Autonomy" section of SOM.) There is a small nonpermissible subset of
possible goals x for which this is not true (e. g. a desire to go around
randomly chopping off people's heads), and cooking up a demarcation of that
nonpermissible subset is a matter for further research. (This is the
problem of "natural law"... the marked-off subset of non-permissible goals
include things like murder and plunder.)
But there are many extrinsic subgoals that generalize very nicely across an
extremely broad set of permissible end-goals (the intrinsic ones). So if I
whimsically decide to constrain my goals to accord to natural law (I decide
I'm not interested in killing and plundering), then no matter what my
idiosyncratic goals are within the remaining goal-universe, I can serve
those ends indirectly by working towards increasing my understanding of the
world and increasing the harmony of my relationships with others (take
harmony in a strictly market sense, *if* you like, with profit as its
measure... there are other interpretations).
So, given the "natural law" constraints, I don't even have to know much
about the particular identities of my idiosyncratic tastes and intrinsic
goals to know how I can serve them in the long run, to know what I ought to
do. (This is the theory, but see my remark at the end of this post.)
Why does the "natural law" mark off the non-permissible goals? Because
those goals are generalizably disextrinsic... they tend to work against the
attainment of any of the much larger set of possible goals not forbidden by
the "natural law". But the problem of making the demarcation precise is
still a hard one, though some good work has been done recently (e. g.
Gauthier).
The demarcation problem is also, for me, closely related to the foundations
of libertarian political theory. If there's going to be a government, it
should only prohibit a minimal set of "bad" things, and mandate practically
nothing. The demarcation of law (or what the law should ideally be) from
morality, and of morality from taste, is very important to me for this
reason. Law should be a lot less pluralistic than morality, and morality
should be a lot less pluralistic than taste. But all of these things are
values.
Making all this precise? Well, maybe I'll expand this into a thesis
someday, if I don't get sidetracked by some other interesting problem.
And before anyone leaps in, I already know that in theorizing about human
values, more than in any other theoretical domain, we must be very careful
not to confuse the map with the territory. These are just theories, or
rather analogies that point in a direction in which a theory might be
developed. I'll be happy if I can someday improve on Kant and Gauthier.
There will always be room for further refinements.
Eric Watt Forste <arkuat@pobox.com> http://www.c2.org/~arkuat/