Stephen P. King (stephenk1@home.com)
Fri, 01 Oct 1999 23:50:21 -0400
Hi Ben,
Ben Goertzel wrote:
>
> hi,
>
> > my idea that Local Systems interact
> > via bisimulation:
>
> I'm trying to get at the intuition underlying this idea...
>
> When A and B interact, this is somehow because A is simulating B and B is
> simulating A
>
> Now, when you and I interact, are we simulating each other?
Yes, this is subtle, but when we consider that what each of us
experiences in any interaction is a subjective experience, we can see
that what is happening is that we react only to our subjective
simulation of each other.
> In a sense we are: I have a model of you in my mind, you have a model of me
> in your mind.
Yes, we see that your model of me is not "me" and my model of you is
not "you", in the sense of identity. What is the key difference between
the model and the modeled object? I see this as a duality in that your
model of me is an information representation of a material "me", and my
model of you is an information representation of a material "you". So
the model is information and the modeled object is matter.
> We are simulating each other as part of the process of intelligent
> conversation. These are very rough and partial simulations.
Yes, we approximate since we do not have unlimited time or resources to
be exact. ;-)
> But this is because we are intelligent systems... involved in what Buber
> would call an I-You interaction,
> in which the other participant is treated as a mind/universe in itself
> rather than as a rigid object, an It,
> as in an I-It interaction...
Umm, so in the I-It interaction we infer intelligence
(self-simulation!) and in the I-It interaction we infer none?
> When one LS A interacts with another LS B, the particles in B are linked
> with the particles in A via
> quantum nonlocality. Thus in a sense B does contain an image of A in
> itself: its own state can only be
> fully understood by including an understanding of A's state. Quantum
> systems interacting therefore seem to
> use this I-You mentality, of approximate bisimulation.
Umm, I don't think this is right, since the "linking via quantum
nonlocality" is what is involved in the composition of an individual LS.
The last statement, I do agree with in a sense. It is in the internal
"scattering" behavior of the LSs A and B, that the images are given.
This is what Lance and I call "clocking".
> It's not really bisimulation though: It's not that A is being a simulation
> of B, or B is being a simulation
> of A; it's that each of them, in some sense, is ~containing~ a simulation of
> the other. And an approximate,
> lossy simulation at that.
Yes, the "interaction" is within the image, as Hitoshi said in [time
846]:
"...I propose that the place where the "phenomena" occur is not outside,
which
physics takes for granted, but it is inside our mind where the problems
of
physics appear; the interaction between QM and GR is not outside but
inside us
and thus something must have been overlooked by the conventional
physical
thinking/framework."
> Hitoshi's theory, however, treats the interaction of LS's classically --
> this is how GR comes out.
Yes, but we must remember that the classical world is plural, each LS
has its own: it is the realm of what it simulates. I am identifying an
internal image with an external behavior, e.g. the subjective "view"
(that the LS "experiences") is identified with some external "classical"
behavior via the mechanism of residuation as explained by Pratt in his
papers.
"Can a body meet a body? Only indirectly. All direct interaction in our
account of Cartesian dualism is between mind and body. Any hypothesized
interaction of two events is an inference from respective interactions
between each of those events and all possible states of the mind.
Dually, any claimed interaction of two states is inferred from their
respective interactions with all possible events of the body.
The general nature of these inferences depends on the set K of values
that events can impress on states. The simplest nontrivial case is K = 2
= [0, 1], permitting the simple recording of respectively nonoccurrence
or occurrence of a given event in a given state. In this case bodybody
and mindmind interactions are computed via a process called
residuation. Specifically, event a necessarily precedes event b when
every state x witnessing the occurrence of b also witnesses a. This
inferred relationship is calculated formally by left residuation, which
we describe in detail later. The dual calculation, right residuation,
permits a transition from state x to state y when every event a
impressing itself on x does so also on y. That is, any transition is
permitted just so long as it forgets no event. These simpleminded
criteria are the appropriate ones for the small set K = 2.
For K = 3 more complex rules for inferring necessary precedence and
possible transition obtain, including the possibility of forgetting (to
be written up). At K = 8 we have groups and semigroups, the latter
embedding all abstract category theory [PT80]."
http://boole.stanford.edu/chuguide.html#ratmech
> The bisimulative, I-You aspects of the interaction occur only when you
> consider the two LS's as fusing into
> a larger, inclusive LS. These disappear when you consider the interaction
> as classical. classical interactions
> are I-It and do not involve this kind of inclusion.
I don't think so, the whole point of LS theory is that if we "fuse" two
LSs into one, the classical behavior that is observed (bisimulated) by
the two of each other when separate, disappears. This situation would
haven if our brains were some how linked, our minds would also merge
such that our separate identities would merge. The cutting of the corpus
callusum (?sp) of the brains of severe epileptics has the reverse
effect!
But, thinking about your wording, I can see that it is my hypothesis of
matter/information duality that is hindering effective bisimulation
between us... ;") I see the I-You and I-It as differing in their
hierarchical status. I-You interactions seems to involve bisimulating a
simulation of behavior while I-It interactions seems to be just a
bisimulation of behavior.
Perhaps I am missing your point and you are saying the same thing in
different words... ;")
Later,
Stephen
This archive was generated by hypermail 2.0b3 on Sun Oct 17 1999 - 22:40:46 JST