On Thu, Apr 15, 2010 at 11:41 PM, Brent Meeker <meeke...@dslextreme.com> wrote:
> On 4/15/2010 8:01 PM, rexallen...@gmail.com wrote:
>> Let's assume that our best scientific theories tell us something
>> true about the way the world *really* is, in an ontological sense.
>> And further, for simplicity, let's assume a deterministic
>> interpretation of those theories.
> Haven't you heard?  Almost all scientific theories are false; that's why we
> keep changing them.

First, I mean "scientific realism" in the sense described by this
Wikipedia article:


Which I'm not sure you are taking into account in your response.

Second, you say "almost" all scientific theories are false.  Which
scientific theories do you believe are not false in a scientific
realist sense?

And third, I set the bar somewhat lower than you imply.  A closer
reading of my first sentence shows that I am only assuming that our
best scientific theories tell us something true about the way the
world really is (ontologically)...not that they are true in every

This would be in opposition to a purely empirical, Kantian,
instrumentalist view that our scientific theories tell us about our
perceptions without necessarily revealing anything about what really


>> It seems to me that it would be a bit of a miracle if it turned
>> out that we lived in a universe whose initial state and causal
>> laws were such that they gave rise to conscious entities
>> whose beliefs about their universe were true beliefs.
> I think you need the concept of "approximately true", otherwise you will
> conclude there are no true beliefs whatsoever - in which case "true" loses
> all meaning.  The approximation can be both in scope and accuracy.  Then I
> think it might be possible to show that all conscious entities arising
> through deterministic evolution of the universe must have approximately true
> beliefs.

So given our current knowledge of the universe, it would seem that a
computer simulation of a human brain would be conscious in the same
way that I am conscious.

Some kinds of 2-D cellular automata are Turing complete and thus could
run such a simulation and also have a cache of data that could be fed
into the brain simulation in a way that the simulated brain would
interpret as sensory data from a surrounding environment.  No
simulation of the environment actually needs to be done, just
time-indexed lookup tables of equivalent data.

Going further, it seems possible that a very simple "physical"
universe could exist with the bare minimum of furniture (e.g., 1
spatial dimension, 1 time dimension, only 1 type of particle that has
2 states, etc.) necessary to implement such a cellular automaton, and
a single causal law that was the equivalent of Rule 110.

Given the right initial conditions, this cellular automaton would give
rise to a human consciousness whose beliefs about how his physical
universe really was would be false.  Only his beliefs about his
perceptions would be true...e.g. "I believe that I'm having the
experience of seeing 3 birds fly overhead" would be a true belief.
However, "I believe that three birds flew overhead" would be a false
belief...because there really are no birds in that universe (not even
simulated ones).  Also, since that universe only has 1 spatial
dimension, there were be no "overhead" either.

The birds and the extra two spatial dimensions would only exist in the
mind of the simulated brain.  They would only exist within his
perception, not external and independent of it.

SO...it seems to me that it is NOT possible to show that all conscious
entities arising through deterministic evolution of a universe must
have approximately true beliefs.  Unless you can show that the above
scenario is impossible.

A good XKCD comic that runs along similar lines!


>> Note that a similar argument can also be made if we choose
>> an indeterministic interpretation of our best scientific
>> theories.
> Except in a stochastic universe another form of "approximately true" is
> introduced: approximation in probability.  Note that even if a universe is
> deterministic, it may be strictly unpredictable because at any give time
> only a portion of the initial state can have affected us due to the finite
> speed of light.  So new and unpredictable information continually reaches
> us.  So this is operationally equivalent to inherent randomness.

An interesting point.  But I don't see that it changes any of the
conclusions I drew in my initial post...do you?

My central point is that if we are in a deterministic universe, then
for us to have *any* true understanding of this universe, that
understanding *must* have already been implicit and inevitable in the
universe's first instant.

Which doesn't seem probable if you were selecting a universe at random
from the list of conceivable universes.  The most common type of
conceivable universe would seem to be lifeless universes.  The next
most common would seem to be universes that gave rise to "life", but
where this life never arrives at any true understanding of the
universe that contains it.  And LEAST common would be universes that
contain life which does achieve some true understanding of the
universe that contains it.

Even if we are in an indeterministic universe, you don't get any
further.  There are still initial conditions and there are still laws
- the laws just have an intrinsically probabilistic aspect.

The laws are like the rules of a card game that includes a certain
amount of randomness...for instance, requiring occasional random
shuffling of the deck.  But the number of cards, the suits, the ranks,
and the rules themselves are not random...those aspects are

Similarly, in (non-relativistic) quantum mechanics using the
Schrodinger equation, the evolution of the wavefunction describing the
physical system is taken to be deterministic, with only the "collapse"
process introducing an indeterministic aspect.

But like with the card example, the random aspect doesn't change the
game.  No matter how randomly the deck is shuffled, you still only
ever have 52 cards, 4 suits, and 13 ranks.  The randomness is
constrained by the rules.

The same goes for algorithms that use randomness, for instance the
Randomized Quicksort.  No matter what pivots you randomly select, the
algorithm is still going to correctly sort your list.  The randomness
is constrained by the context.

And the same goes for our universe in the indeterministic case.  The
randomness only increases the probability of life discovering
something true about the underlying nature of the universe IF the
initial conditions and the non-random aspect of the causal laws allow
for this to be the case.  You'll never run a randomized quicksort and
as a result get a conscious entity that realizes that it's living in a
computer running a randomized quicksort.

You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To post to this group, send email to everything-l...@googlegroups.com.
To unsubscribe from this group, send email to 
For more options, visit this group at 

Reply via email to