From: Hal Finney (hal@finney.org)
Date: Wed Jun 19 2002 - 23:25:46 MDT
Lee writes:
> Hal writes
>
> > Lee writes:
> > > We all know the basic difference, apparent every day of our lives,
> > > between static representations and dynamically changing ones over
> > > time. We never suppose a frozen state, or even a sequence of frozen
> > > states, to be conscious or to be experiencing anything. So where does
> > > this bizarre notion that appears to afflict you and Greg Egan that
> > > a large enough number of separate patterns---say separate integers---
> > > take on any sort of dynamism? For me, since no information flows
> > > between dead and passive structures, they're not alive.
> >
> > Actually I was not claiming this, and I don't think my argument depends
> > on it. My claim was only that computer programs and even executions of
> > programs exist in the same abstract sense that numbers and geometric
> > shape exists.
>
> First, your argument as to what? (your first sentence)?
I mean the argument in the message you were replying to, which was
specifically about how computer programs might exist in a Platonic sense,
how the conscious beings they create might be just as "alive" as we
are, and then how such a worldview still allows for human action to have
meaning. None of this relied on believing that static data would contain
conscious beings, just abstract computer programs, which are not static.
> > So, here is a structure which is (in a sense) static and yet it does
> > contain conscious entities. So this might give us some reason to think
> > that static entities can contain consciousness if they have the right
> > kind of internal structure.
>
> Leaving alone higher dimensionality, we could wonder the same
> thing about a three dimensional object. Consider a rock in
> my hand. By *forcing* interpretations, it is possible to find
> any pattern in there at all, and to find sequences of any
> patterns we choose, e.g., conscious entities. One possible
> escape (showing *my* prejudice!) is to point out that forced
> interpretations require too much input from the forcer, and
> that the real information starts to reside in the person
> holding the rock (i.e. the person forcing such interpretation)
> and very little in the rock itself.
Yes, I am familiar with some of the literature on this issue, and I
agree that your response is a good one.
> What I want to know is "How explicit is the cause and effect
> that connects the gels?". Notions of causality aren't so
> easy to deal with themselves, so far as I know. I wonder if
> philosophers have made any headway in the past couple of
> decades. Anyway, the "causality" linking the gels together
> isn't apparent to me! No more than a single 1-dimensional
> slice of Wolfram's CA is causally connected to the next
> 1-dimensional slice. Again, to be meaningful to me, they
> have to actually execute.
I agree that this is a difficult issue. I have a couple of thoughts
on it.
One idea is to think about a mathematical proof as an abstract object.
Proofs have a kind of causality to them as well. Each step has to
rely on specific earlier steps, on axioms, and on rules of inference.
In principle each step can be fully annotated so that the proof is
mechanically checkable, as is done sometimes in computer-generated proofs.
Can you imagine an abstract proof as existing in a timeless sense,
and yet somehow the abstraction captures the notion of causality which
must exist for the proof to be valid? If so, perhaps your stack of gels
could be thought of in the same sense, not as simply a set of disconnected
layers, but conceptually joined with the rules that lead from one layer to
the next. Perhaps these could be thought of as virtual annotations which
would represent the laws that make two consecutive layers consistent.
Or in a physical model, you could even imagine a set of mechanical gears
and pegs which would connect each pair of layers and which implement
the constraints of the CA algorithm, such that only mutually consistent
pairs of layers are mechanically possible. The existence of these gears
would physically enforce the type of causality which is necessary for
a legal computation to take place.
These kinds of models suggest that causality can exist in a timeless
sense. So perhaps it is not so impossible to imagine consciousness
existing in the same way.
> To me, the whole purpose of philosophy is prescription. I want
> to know what to do in various circumstances. For example, how
> should I value my life if I have to choose between living here
> and uploading? If some program, say a friend of mine, has
> undergone a lot of pain, does re-running good runs he had help
> him at all, or compensate for his unpleasantness? Should I
> vigorously oppose the execution of programs which experience
> pain, on the grounds that it doesn't matter whether they've
> been run before?
To be sure, these may become important questions some day. But for
now they are entirely hypothetical! So I don't think there is much
urgency to make decisions on these topics right away. At this point I
just play with these various thought experiments and listen to different
people's ideas. Hopefully by the time issues of re-running, uploading
etc. become relevant, we will have made some progress towards better
understanding of these philosophical conundrums.
Hal
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:14:55 MST