Mistaken Steps?

Descartes, no doubt, had put matter too far from us when he made it one with geometrical extensity. But, in order to bring it nearer to us, there was no need to go to the point of making it one with our own mind. Because he did go as far as this, Berkeley was unable to account for the success of physics, and, whereas Descartes had set up the mathematical relations between phenomena as their very essence, he was obliged to regard the mathematical order of the universe as a mere accident. So the Kantian criticism became necessary, to show the reason of this mathematical order and to give back to our physics a solid foundation-a task in which, however, it succeeded only by limiting the range and value of our senses and of our understanding. The criticism of Kant, on this point at least, would have been unnecessary; the human mind, in this direction at least, would not have been led to limit its own range; metaphysics would not have been sacrificed to physics, if philosophy had been content to leave matter half way between the place to which Descartes had driven it and that to which Berkeley drew it back-to leave it, in fact, where it is seen by common sense.

Henry Bergson, Matter and Memory

Consciousness and Special Relativity

In the post Can We Digitalize The Brain and Retain Consciousness? , I gave the Record/Replay argument, which was based on the assumptions that:

1) Consciousness is an occurrent property of the brain (distinguished from dispositional properties).
2) All occurrent properties of a system of two information processing units A and B connected to each other, won’t be affected by changes which don’t change neither A, nor B, and keep the principle: Whatever information comes outside of A is in same timely manner sent to B as in the original network.
3) Brain can be transformed into network of digitalized neurons while retaining consciousness.

The argument then shows that the consequence of those assumptions is that a neural network of physically disconnected neurons will also have consciousness.

I said that this is impossible and hence that by reductio, given that one accepts assumptions (1) and (2), assumption (3) is wrong. I was expecting mostly that people will see assumption (2) (and maybe (1) as problematic), but Richard in the comments noted that it isn’t obvious that a network of physically disconnected digital neurons can’t have consciousness. So, in this post I will try to address that worry.


In the mentioned post I ended up with neurons attached to “replay boxes”, one “replay box” for each neuron that was connected to them on input previously. And those Replay Boxes replay the information in same timely manner.

Now one thing about this system, is that we can get those neurons with the attached replay boxes, and put them apart from the other neurons. The distance between the digital neurons doesn’t play any role, because the inputs are provided by the replay boxes.  We can take those digital neurons with attached replay boxes and scatter them all around the universe for example. Given the three premises it follows that consciousness will still be occurrent property of this bunch of neurons.

But now, there is one question about what will happen with the idea that whatever information gets out of the digital neuron A, will appear at the input of digital neuron B. When we have distances, from relativity we know that there is no absolute meaning to ‘before’ and ‘after’. If digital neurons A and B are on diametrically opposite sides of the observer, depending on the movement of the observer specific event of A might be before, or after a specific event in B.

However the consciousness either occurs or not, and this can’t be relative depending on some external observer. Because of this we need to make the requirement from assumption 2 weaker. The requirement was, to remind you that “Information from A appears at the input of B in timely manner”, and this is something that doesn’t make much sense now. We can say that not just that the digital neurons can be scattered through space, but also can be scattered through time.

But an occurrent property is something that happens or not in specific time (or specific period of time). So, consciousness can’t occur in bunch of digital neurons with attached replay boxes scattered through time. And by reductio, following the whole argument at least one of the assumptions (1) to (3) is wrong.

Given that one accepts assumptions (1) and (2), it can’t be that we can change the neurons of the brain with digital neurons and retain consciousness.