hal@finney.org wrote:
> if the determining factor
> of whether a physical system is conscious depends on the K complexity
> of certain aspects, then consciousness is a noncomputable predicate.
> This is not as bad as Eliezer's making consciousness rely on uncomputable
> physics, but it is still a disappointing result.
Sheesh, folks, it's hard to imagine why anyone would've defined a formal complexity definition that can't be computed, this is getting *real* obscure! Do you mean that actually crunching the numbers on this Kolmogorov whatever-it-is would require a solution of some particular unsolved number theory problem, or something like that?
>
> The second problem is that K complexity is not uniquely defined. For any
> proposed string, you can find a universal Turing machine which outputs
> that string using a very small program. You can "cook the books" and
> find a UTM which will cause any chosen string or finite set of strings
> to have a low K complexity by the standards of that UTM.
> If K complexity is to determine whether something is conscious or not,
> which is an objective fact, then we can't have it be variable based on
> whatever UTM we choose.
This last seems a lot more understandable to me, as it sounds like the same kind of complication that they have in calculating the entropy of molecular machines, as I mentioned in the related "Computability of Consciousness" thread back in March. The calculated entropy of a molecular machine depends on the state of knowledge of the observer doing the calculation, and this is what's being said for this K-complexity thing as well.
Now there is something here that's almost a little embarrassing to me, as I seem to have fallen, earlier, into a trap that the assumptions being made here led me to. As I recall, where possible K-related consciousness measures were concerned, I suggested a rather mystical approach to trying to solve this uniquely, namely that there might be a final, unified observer, in effect a Good Observer Deity, a G.O.D. above it all, to give an absolute definition to these observer dependent quantities! Unfortunately for this grand notion, I now have strong second thoughts about why any such special observer would have the effect being looked for here, the effect of making computations truly real (or of "instantiating" them, as Eliezer Yudkowsky talks about it).
Really when you think about it, doesn't it seem quite likely that most of this discussion, this whole perceived need for consciousness "qualia" to be "instantiated" is entirely beside the point? For instance, Hal talked earlier about a chain of logic that led him to wonder if a whole bunch of scattered nerve cells that just happen to exist in a whole lot of unrelated people could nevertheless form up an independent "brain" of their own, by just happening to proceed through the same firing states as would occur in some hypothetical real brain? Whatever the value of the intermediate ideas that led to this scenario may be, doesn't it get totally indefensible if you're counting on random unrelated bits to do calculations of some kind?
Look, what if a computer engineer came to you and told you that "you don't need no computer, just think of all the random unrelated bits in the world that might happen to be arranged the same as any software you want", what would you think of a computer engineer who told you *that*? A *real* calculation is a particular, local, physical procedure, with results that should be accessible and useful, at least in principle, to other real, physical, local processes, isn't that much obvious? Hal's "shadow brain" wouldn't be accessible to anyone in the real world of scientific instruments, etc., unless someone accessed it by running it in a computer simulation, it which case, clearly, it'd be the simulation itself, running on an actual, real, computer, that would be the candidate for consciousness. Why does anyone think they need some brand new, magical, sort of physics for consciousness calculations when calculation is, itself, *already* a physical thing?
Going back to my earlier mistake, the one about a special observer instantiating everyone's consciousness, I actually think I may be able to discern the underlying problem here. Basically, we humans are just so doggoned sure, aren't we, that our particular kind of perceptual information processing experience is unique, uniquely real, a more basic "feel" of things than our part of the universe ever had before we came along? As in, the world would hardly be *out* there if I didn't see it? To put it another way, the soul of me, the little god "in here", is essential to helping the world be just what it is -- at least that's the feeling that I think we tend to have of ourselves. Someone may try to unify this feeling out to make a Great Soul helping all us little souls, but it's still the same idea, I think, that *our* sort of perception is the center of reality, and only what *we* perceive is really there in the end.
The true reality, surely, is that even a rock is just as real as you or I, albeit not as computationally organized. Going from a rock, to say, a horse, the horse certainly has computational brain events going on that are important to the horse, so who is to say that the horse's brain coded perceptions are unreal or in need of some special "instantiation" before those perceptions can become real? I mean, not only is a horse's brain real, and not some kind of shadow brain, the perceptual calculations going on in there are real too, darn it! In our own, human case, we've apparently got special language handling and symbol handling circuitry that a horse doesn't have, not that it always saves us from goofily egocentric mistakes . . .
David Blenkinsop <blenl@sk.sympatico.ca>