Perhaps someone with more insight can explain how this isn't a violation of the Church-Turing thesis?
Fundamentally, we can simulate neuromorphic computer (or even a quantum computer, or even the entire human brain) on Von Neumann machine (with performance degradation). According to our current understanding of physics, the behavior of these systems should be identical in all respects except the simulation will be slower. However integrated information theory says that a neuromorphic computer may be highly conscious and the brain is most definitely highly conscious, but their simulation isn't.
What's still unclear to me is whether this is indeed just metaphysics (i.e., the simulation and the real thing are absolutely identical and but we should still treat them differently from ethical standpoint), or is it a hypothesis about our physics (i.e., the brain somehow fundamentally cannot be simulated on a Von Neumann machine) as well as computational theory (i.e., neuromorphic or quantum computer cannot be simulated either).
That statement can be true without being metaphysical. You can't treat a human as a black box, and then use inputs and outputs to definitively draw conclusions about what's going on inside the black box.
You could potentially have two black boxes that give identical outputs for the same inputs, yet have completely different mechanisms inside for arriving at those outputs. For example, say you have someone who's very good at arithmetic able to multiply two numbers. You put him inside one black box and a pocket calculator inside another. You give each box two numbers and they both output the product. Both black boxes will give you identical outputs for any given input. You know the box with the person inside is conscious, but this is not enough information to conclude the other black box is conscious.
It's quite possible that there is a way to simulate human behavior through some other non-conscious mechanism. And just because we can't currently prove whether something is conscious doesn't mean it's something metaphysical we'll never be able to prove.
For example, if two people tell you they're experiencing some pain, but one is lying about it, a few hundred years ago you'd be unable to prove it. That didn't mean pain is something metaphysical. Today, we know it might be possible to put both people in an MRI and prove whether they're actually feeling pain
It's not proven that consciousness can be simulated by a Turing machine. There's a possibility consciousness requires some hypercomputation to simulate
The metric depends on how things are connected inside. There are a few versions of it, but generally systems with more direct wires between components are more conscious.
I guess if you don’t think consciousness is measurable through behavior, you have to cook up some metric that depends on internal organization.
Sure it is; brain imaging, which is how IIT tests for consciousness according to the article, is looking at phenomenology.
First, while that's true now, there is no guarantee that it will always be true as technology advances.
Second, neural activity is the physical basis for conscious experience, which, to a physicalist like me, means it is conscious experience, when the neural activity has the right properties (according to IIT, those properties are what "integrated information" is trying to capture). So I don't accept the distinction you are making between "the conscious experience itself" and neural activity.
True, but what does that mean? We get to watch someone else's dream like a movie because neuroscientists will have figured out how to translate the neural activity into a format that can be outputted as video and audio.
That won't work for every sensation. You can't feel that you're having the same dream as being that person's body.
> Second, neural activity is the physical basis for conscious experience, which, to a physicalist like me, means it is conscious experience,
This is similar to an identity theory of mind, with a focus on information. The problem is understanding how integrated information produces smells and pains. Saying the experience is identical is making a claim that seems mysterious. Why would some forms of information be conscious? Did God set that up?
It sounds similar to Chalmers' position on information, except that Chalmers says there is a natural link between rich information and being conscious that is an additional law in addition to the physical ones.
I will apply Occam's razor to that additional law until there is evidence for it.
More generally, there is a straightforward physical possibility for explaining the subjective and personal nature of qualia, and other aspects of consciousness: our brains are not connected in such a way that language processing is capable of initiating all of the state changes that other sensory inputs can, and so we cannot communicate those state changes directly. This not only accords with our actual experiences, but it is probably just as well, or else we might find it (even) hard(er) to distinguish reality from imagination (and by "reality", I mean concrete, "that bear just noticed me", reality.)
One may, of course, take one's own Occam's razor to this idea, if you think physical constraints on information flow in a brain are less pausible than Chalmer's extra-physical link.
Fair enough, but Hume already did that for laws (causality). We don't observe laws, just constant conjunction. B always follows A, so we say law C is the cause of that. Or description, if one doesn't like the implication.
Adopting a Humean approach, there is already evidence that consciousness is conjoined to certain neural activity. But, that can work just as well for IIT as it does Chalmers.
And I have no idea whether causal laws exist or what consciousness is. I guess IIT is as good an approach as any.
If one takes this way to leave the discussion, then there would be more than just a hint of motte-and-bailey equivocation if one were to then re-enter it with a specific claim, such as that Chalmers or Searle or Koch have a plausible argument that consiousness cannot possibly be this or that.
As it happens, I very much doubt that IIT is on the right track, but at least it is (or was, initially) an attempt to find out what consciousness is, rather than what it is not.
What I meant was that future brain imaging technicians might be able to record your brain waves while you sleep and then, when you wake up, correctly tell you what you dreamed about.
However, it is also possible that future brain imaging technicians might be able to record your brain waves while you sleep and then use that data to construct a virtual reality experience that is so convincing that it could make someone else directly experience your dream.
> You can't feel that you're having the same dream as being that person's body.
You have no basis for making this claim except that we currently don't know how to do it. That is a very weak basis for such a claim.
> The problem is understanding how integrated information produces smells and pains.
Bear in mind, I'm not saying I myself believe IIT, I'm just trying to describe what it says. I don't know that "integrated information" is the right way to describe what it is about the neural activity in our brains that produces smells and pains.
> Why would some forms of information be conscious?
If I generalize this to "why would some forms of neural activity be conscious?", the answer is that consciousness has survival value, so neural activity that can produce conscious experiences evolved.
If it can simulate conscious behavior well enough to fool conscious observers (like us), how do you know it wouldn't be conscious?
I don't fully understand it myself, but based on the axioms in that theory, you can calculate the potential consciousness of a system.