RE: computer pain

From: Jef Allbright <jef.domain.name.hidden>
Date: Fri, 22 Dec 2006 12:59:25 -0800

Immediately upon hitting Send on the previous post, I noticed that I had
failed to address a remaining point, below.

> Brent Meeker wrote:
> >
> > Stathis Papaioannou wrote:
> >>
> >> Jef Allbright writes:
>
> <snip>
>
> >>> Further, from this theory of metaethics we can derive a practical
> >>> >>> system of social decision-making based on (1) increasing
> >>> fine-grained knowledge of shared values, and (2) application of
> >>> increasingly effective principles, selected with regard
> to models of
> >>> probable outcomes in a Rawlsian mode of broad rather than narrow
> >>> self-interest.
> >>
> >> This is really quite a good proposal for building better
> societies,
> >> and one that I would go along with, but meta-ethical
> problems arise
> >> if someone simply rejects that shared values are important (eg.
> >> believes that the values of the strong outweigh those of the weak),
> >
> > Historically this problem has been dealt with by those who think
> > shared values are important ganging up on those who don't.
> >
> >> and ethical
> >> problems arise when it is time to decide what exactly these shared
> >> values are and how they should best be promoted.
> >
> > Aye, there's the rub.
>
> Because any decision-making is done within a limited context,
> but the consequences arise within a necessarily larger
> (future) context, we can never be sure of the exact
> consequences of our decisions. Therefore, we should strive
> for decision-making that is increasingly
> *right-in-principle*, given our best knowledge of the
> situation at the time. Higher-quality principles can be
> recognized by their greater scope of applicability and
> subtlety (more powerful but relatively fewer side-effects).
>

It's an interesting question as to how we might best know our
fine-grained human values across an entire population, given that we can
hardly begin to express them ourselves, let alone their complex internal
and external relationships and dependencies. There's also the question
of sufficient motivation, since very few of us would want to spend a
great deal of time answering (and later updating) questionnaires.

The best (possibly) workable idea I have is to use story-telling. It
might be done in the form of a game of collaborative story-telling where
people would contribute short scenarios where the actions and
interactions of the characters would encode systems of values. Then,
software could analyze the text, extract significant features into a
high-dimensional array of vectors, and from there, principle component
analysis, clustering, rankings of association and similarity could be
done mathematically via unsupervised software with the higher level
information available for visualization. This idea needs more fleshing
out and it might be possible to perform limited validation of the
concept using the existing (and growing) corpus of fictional literature
available in digital form.

- Jef


--~--~---------~--~----~------------~-------~--~----~
 You received this message because you are subscribed to the Google Groups "Everything List" group.
To post to this group, send email to everything-list.domain.name.hidden
To unsubscribe from this group, send email to everything-list-unsubscribe.domain.name.hidden
For more options, visit this group at http://groups.google.com/group/everything-list?hl=en
-~----------~----~----~----~------~----~------~--~---
Received on Fri Dec 22 2006 - 16:09:37 PST

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:12 PST