Consciousness and Experience

by Dave Schuler on May 8, 2014

There’s an interesting post at The Physics arXiv Blog on the work of neuroscientist Giulio Tononi in coming up with a rigorous characterization of consciousness. Here’s the nugget from the post:

The central part of their new work is to describe the mathematical properties of a system that can store integrated information in this way but without it leaking away. And this leads them to their central proof. “The implications of this proof are that we have to abandon either the idea that people enjoy genuinely [integrated] consciousness or that brain processes can be modelled computationally,” say Maguire and co.

Since Tononi’s main assumption is that consciousness is the experience of integrated information, it is the second idea that must be abandoned: brain processes cannot be modelled computationally.

One of the virtues of rigorous description is that it defines the limits of the possible. They become tautological. If Tononi is right, machine consciousness is not possible. That includes computer programs to implement machine consciousness.

{ 5 comments… read them below or add one }

Modulo Myself May 8, 2014 at 12:22 pm

Very interesting. It sort-of brings to mind Schrodinger’s What Is Life? , where he hypothesizes a DNA-like structure that is small enough to withstand entropy in order to facilitate reproduction of the information.

At the same time, not being a neurologist, integrated consciousness sounds very puzzling. I can’t control my heart rate, for example, but I can feel my heart beating, or the pain of a heart attack. Similarly, chocolate smells like chocolate to me, but my ability to know that I am smelling chocolate seems to work outside of my actual consciousness. Yet if I don’t know what a smell is, I might search for it intentionally, but if I do know, I don’t have any awareness of the search.

Wittgenstein was really interesting in counting because of this, I think. We don’t have have to search for the idea of counting by evens when we do it. (Though when a child who is learning to count, the child is clearly searching and struggling with counting.) And yet we know how to count by evens, as if we were accessing an overall rule in our brains. But are we?

What this means, I haven’t a clue.

... May 8, 2014 at 2:26 pm

If Tononi is right, machine consciousness is not possible. That includes computer programs to implement machine consciousness.

That’s a relief, actually.

michael reynolds May 8, 2014 at 3:27 pm

Modulo:

Similarly, chocolate smells like chocolate to me, but my ability to know that I am smelling chocolate seems to work outside of my actual consciousness. Yet if I don’t know what a smell is, I might search for it intentionally, but if I do know, I don’t have any awareness of the search.

We should spark a bowl and really work this through.

Ben Wolf May 8, 2014 at 4:55 pm

Ray Kurzweil is sobbing in a corner over this. Someone be a dear and let him know he won’t, in fact, live long enough to live forever.

mike shupp May 10, 2014 at 4:48 am

“consciousness is the experience of integrated information”

Perhaps this is wrong? What we refer to as our consciousness is but partial and very poorly integrated information, which we mistakenly accept as complete and well integrated. It seems quite plausible that we might someday construct complex devices which emulate human behavior by forming an incomplete view of nature and treating it as complete.

Whether machine behavior is really consciousness is another matter, of course.

Leave a Comment

Previous post:

Next post: