> Anything that can perceive itself and can perceive itself perceiving will exhibit all the hallmarks of what we call consciousness. It's a recognizable behavior pattern with a name.
In my opinion this just delays the question.
The question now becomes whether machines can perceive themselves (or really anything at all) or whether they can just mechanically and symbolically represent the perception of objects, including themselves.
If the answer is yes, then that implies that any sufficiently large representation of computation has a consciousness and therefore that there is nothing special about it.
If it is no, then where's the relevant difference to a human brain?
You're arguing that consciousness is to be understood as a consciousness of the self (and its environment).
That implies that there can be intelligence that is not conscious, right?
About the missing sensation of the field of rocks; How about adding some sensors to the system, so depending on measurements like temperature, one could move a certain designated stone, which could then be taken as input by the rules moving all the other stones.
Would that change alone yield consciousness?
Also, how many different types of such sensations are needed to produce consciousness? Humans sense quite a few different types of such input data, but we can't sense everything there is to be sensed.
In your last two paragraphs, I believe you state that consciousness is just the (mathematical) reflexivity of "thoughts" (i.e. representational processes concerning objects).
That's an interesting thought I've pondered as well, but doesn't that just move the problem to whether machines can produce such thoughts?
Certainly they can produce representations of thoughts, but do they really think them?
First of all, thanks for mentioning those other examples to me, I haven't heard of them before and will definitely engage with them further.
You're right that I can't just dismiss the possibility of the stones or the paper developing a consciousness.
However, this fallacy wasn't unnoticed by me, I just don't see a way to approach this question in a more scientific way than by using my intuition.
So you're proposing that there is actually no difference between human brains and computers in respect to the ability to be conscious, did I get that right?
I don't understand your argument about the stones "buzzing and vibrating and bouncing around [...] in highly nonlinear ways" though.
That doesn't resemble the computations made by a discrete computer anymore, does it?
Or are you hinting at precisely those differences between computers and brains (i.e. mathematically discrete vs. mathematically continuous)?
Also, why do you think parallel information processing matters? Does it really matter in a computer?
Could you elaborate on your argument that scale (in terms of calculation speed; in terms of memory it surely matters) does indeed matter?
I don't think it's any more plausible for a huge pool of fast-moving "magic sand" to be conscious (excuse me for once again committing a scientific fallacy here, I'm open for suggestions on methodology).
You should really take a look at the paper "Computation and Pattern Formation by Swarm Networks with Brownian Motion" or, more generally, any of the papers by Teijiro Isokawa and Ferdinand Peper. What they present is really close to (a theoretical version of) the microbots you're talking about.
I even wrote a report and presentation on those papers for university, but sadly they're in German.
In my opinion this just delays the question. The question now becomes whether machines can perceive themselves (or really anything at all) or whether they can just mechanically and symbolically represent the perception of objects, including themselves. If the answer is yes, then that implies that any sufficiently large representation of computation has a consciousness and therefore that there is nothing special about it. If it is no, then where's the relevant difference to a human brain?