X-Message-Number: 11679
Date: Wed, 05 May 1999 01:02:14 -0700
From: Mike Perry <>
Subject: Re: Sub-Emulations, etc.

Bob Ettinger, #11671, writes

>Mike Perry writes:
>>The emulation does not stop the real world, and similarly, the sub-emulation 
>>(and any descendant emulations) would not stop the emulated world, if 
>>reasonably programmed.
>I don't want to spend much time on this, but I don't think that analysis is 
>correct. An emulation in a computer of course will not stop or slow down the 
>real world, because the computer is only a tiny part of the real world and 
>has virtually no interaction with it. But the emulation, and any 
>sub-emulations, all rely on physical events (data processing) in the one real 
>computer. Therefore, if we have a cascade of sub-emulations, all of those 
>resulting virtual computers must rely upon, and tend to overload or slow 
>down, the one real computer.

But there is no reason this "overload" would be different in character from
the demands the real computer places on the real world, which can themselves
be quite small, as you suggest. On the other hand, it is possible that real
people would decide to enlarge the real computer and cause it to consume
more resources, and something like this could happen in the emulated world
too. But the emulated world would go on, pace unhindered, just as the real
world would go on, pace unhindered, even if the real computer was bigger.
The *type of activity* in either world would would alter, but not the pace
of events at an underlying level. 
>"reasonably programmed"--what does this mean? The programmer, as far as I can 
>see, has no wiggle-room.

You could, for instance, write your emulaton program so it will halt
execution if certain register has the number "13," or if some other
arbitrary condition holds, but there would be no particular reason to do so.

> He either programs an emulation or he doesn't. If it 
>is a true emulation, it rigorously follows the script of what the original is 
>doing or would do.

But there is more than one way to do this, and the overall program, in
addition to just doing the emulation, could on the side be doing various
analyses, and could, for example, interrupt and restart execution of the
emulation proper from time to time.

>Of course, as Donaldson again recently noted, a true 
>emulation is impossible as long as the programmer does not completely 
>understand the original, and no one does have such understanding or is likely 
>to in the foreseeable future, since new facts and laws of nature are 
>constantly  emerging.

Well, there is always the possibility that we ourselves are in an
"emulation" that isn't really complete, relative to a world outside, but
that we will still be able to discover new scientific laws, which our
Programmer will see fit to make discoverable to us by suitably interrupting
and modifying the program.
>As for an interactive emulation, again, if it is interactive then it is not a 
>true emulation--i.e., its behavior no longer mirrors what the original is 
>doing or would have done. 
This would be true for emulating a whole universe, which by definition must
be self-contained. But (in addition to this possibility, or some other
self-contained unit) I was also thinking of the idea of expressing a human
being in a non-protoplasmic form, i.e. "emulating" in that sense, even
though contact with an outside would be allowed. (Is there a better term I
should use?)

It's worth remarking here that, despite all the foregoing, I don't think
it's likely we are really in an emulation, or something like it, but I do
think it is at least possible. I brought up this possibility earlier as part
of an argument for (what amounts to) strong AI. My reasons for thinking we
are, nevertheless, "real" not emulated are that, as Bob notes, an emulating
computer would (very likely, I think) only be a very small part of the world
of which it is part. In addition, if the emulation really goes down to a
deep level, it should run at a much slower pace than in the world
"above"--which would further rarefy the emulated events relative to the
"real" ones.

Mike Perry

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=11679