↪RogueAI
Ah. I'm sure it will happen.


I agree that machine consciousness is metaphysically possible, but I'm far from sure it will happen. Evolution has had around 500 million adapting neuron 'design' and the way neurons interconnect, to result in us. It's not obvious to me that human ingenuity can achieve a sufficiently similar result. (More likely seeming to me, is that humans could develop a nonconscious AI that then proceeds to design a conscious AI, but that is scary^2 and I tend to doubt that humanity is that stupid.)

I'm also inclined to raise an ethical objection to an attempt to create a conscious AI, in that I think it is almost a certainty that initial attempts at a conscious AI which did result in consciousness, would result in an 'insane' consciousness.

The hard problem has nothing to do with whether consciousness resides in the brain. Its about creating an objective measurement for subjective experience.

Ok, I would say that from my physicalist perspective I would expect reaching such a goal to be impossible. However, I don't see such a goal being unreachable as posing a logical problem for my perspective.

Anyway, I am undoubtedly most accustomed to encountering "the hard problem" being brought up, as an attempt to disprove reductive physicalism (often accompanied by suggestions that panpsychism is more reasonable), and I haven't been reading here long enough to have good intuitions* about where individual people are coming from. So I hope you will forgive me projecting my biases while I get to know you all.

My neural networks are insufficiently trained. ;-)

Its not that conscious experience can't be explained in terms of physical interactions, its the qualitative experience itself. Qualitative experience is subjective. Therefore it cannot be objectively captured in a physical model.


Right. Knowing all the details of what is physically going on in a system (brain) is a different matter from having the experiences resulting from the processes which are occurring in that system.

But why should we find that even surprising on physicalism, let alone a hard problem? We don't say to ourselves, "I have full knowledge of how that car works, so why don't I find myself running down the road at 55 MPH from time to time?"

They can say that, but if a loved one dies, we can ask them why they are crying over chemicals, and we'll see how much they really believe in this stuff.


My answer to you asking the question* would be that it is not the chemicals, but the loss of the chemicals being arranged grandma-ishly that I am mourning, because I really liked the effect of the chemicals being arranged grandma-ishly.

Neurons being interconnected as they are in a living brain, and the interactive processes occurring as they do between the neurons of a living brain, results in events that don't happen when those interactive processes within that network of neurons is no longer occurring.

It isn't surprising that you don't get physicalism, if you haven't considered the subject charitably. However, I suspect you will have difficulty providing objections that would be found perplexing by people who have held a physicalist perspective for a long time, unless or until you do extend more charity on the matter.

*Given that I am posting on a philosophy forum and not currently deeply mourning someone recently deceased. In the actual case where I was asked in the midst of crying about the loss of a loved one, my answer would undoubtedly be more like, "Jeesh, why are you being such an asshole?"

"The hard problem is that we cannot ourselves know what it is like for another being to experience that qualitative experience."


I wasn't under the impression that the hard problem was specifically about our inability to experience the experiences of another, but rather a question of how can conscious experience be explained in terms of physical interactions at all.

Anyway, I don't see the inability to experience the experiences of another as providing any difficulty for physicalism. Supposing my experiences arise from the physical operation of my brain, it seems unsurprising to me that the only experiences available to me to have are those that arise from the physical operation of my brain, and not those of another brain.

So could you elaborate on why you think that an inability to experience the experiences of another would pose a hard problem for physicalism?

If you take an airtight bag - or, rather, a hydrogentight bag - fill it with hydrogen and oxygen, and shake the hell out of it, will you maker water?


I don't know about shaking the hell *out* of it. But shaking the hell *into* it makes water.

I just applied a chunk of burning sulfur to my big bag of hydrogen and oxygen... ...and hell!!!

Isn't that like being a bit pregnant? Sorry - couldn't help it... :wink:
2 minutes ago


Well... There have been many times when it has been a bit like there was something big inside me that I wanted to push out, but the linear structure of language provided a painfully narrow orifice through which to do so. So maybe so. ;)

Perhaps the babbling you’re hearing is a result of your tone deafness to unfamiliar paths of thinking.


Nope that wouldn't be it. I'm autistic, and a bit savantish, and quite familiar with unfamiliar paths of thinking. Paths of thinking that don't involve words among them. ;-)

"In order to go beyond a way of thinking, you first have to demonstrate a proper understanding of it."


This seems to me a prima facie false statement. Do you have an argument for it?

Do I need a "proper understanding" of ancient Greek cosmology in order to go beyond it? What does "proper understanding" even mean in such a case? Or to ask it another way, did Einstein need a proper understanding of luminiferous aether to go beyond it?

No offense intended, but your statement strikes me as something a member of a priesthood might say, in an attempt to cow anyone who might suggest it might be reasonable to dismiss the priesthood's theobabble.


"An interesting article thanks! I wonder how long protections such as:

The researchers addressed questions about the potential misuse of the technology. Decoding worked only with cooperative participants who had participated willingly in training the decoder. If the decoder had not been trained, results were unintelligible, and if participants on whom the decoder had been trained later resisted or thought other thoughts, results were also unusable.

will last!"

Speaking as an electrical engineer, with a long time interest in the uniqueness of brains, I'm quite confident that without the AI having been trained on a specific individual's brain, the AI would not be able to decode that individual's linguistic thought.

However, in cases where a more advanced AI:
1) had been trained to decode a specific individual's thoughts
2) had the ability to provide input to the individual (spoken word, images, video, etc.) in an attempt to 'interrogate' the individual's thinking
...I wouldn't be too confident that the individual would be able to thoroughly prevent the AI from learning something that the individual didn't want known.

Recent AI testing has been done...


Even more recent...

https://www.eedesignit.com/oh-no-ai-now-reads-minds/

I think consciousness is casual. But I'm hoping someone who agrees that it is, indeed, nothing but physics, but also thinks it is causal, can explain how they believe both things, since they appear to contradict each other. Because, otherwise, I'm looking at panprotopsychism. Which is an awkward ideas. Even if true, it doesn't seem to be anything about which we can do more than speculate.


I highly recommend Peter Tse's The Neural Basis of Free Will: Criterial Causation.

https://www.amazon.com/Neural-Basis-Free-Will-Criterial/dp/0262528312

Confession: I've read the sciency first half of the book and haven't finished the more philosophical second half. So I can't say whether I agree with Tse on libertarian free will or not. However I do agree with him that what he calls criterial causation is what should be under consideration in order to have a scientifically informed discussion of the subject.

Exactly, because Hume's method is to portray reason as infallible, then demonstrate the fallibility of our assumptions about causation, and induction in general, and conclude therefore that these types of reasoning are not properly called "reason". That is a problem, because it leaves these processes without any category, no means to understand them, therefore no means to address and rectify their problems. Instead, we ought to class them as forms of reasoning which are more fallible than some others, therefore these forms of reasoning have issues which need to be addressed.

Ah, wonderfully insightful I think.

I'm not well informed enough about Hume's argument to assess the accuracy of your interpretation of Hume's thinking. But having put a fair bit of consideration into how minds might emerge from brains with the benefits of modern scientific findings, I think Hume having an overly simplistic view of reasoning was unavoidable. If only we had a time machine and could go back and talk to him.

"What goes on in a computer doing a calculation is, no doubt, entirely governed by physics. But it is also governed by mathematics - that's why we call it a calculation. Of course, humans have organized the computer to ensure that's the case. So the basis of the physical processes in a computer is mathematics and the basis of that situation is that humans have arranged it..."


Perhaps I am interpreting you overly literally, but as an electrical engineer I would put it differently.

I would say that a computer is constructed such that, in a (weakly) emergent sense, the computer behaves as if it were governed by mathematics/software. However, it would be suggesting overdetermination to claim that the behavior of the computer is governed by mathematics as well as physics. (I'm not sure what "governed by mathematics" would mean.)

I can't speak for what others are thinking when they say that "a computer is performing a calculation", but what I am doing in that case is taking pragmatic advantage of speaking simplistically in terms of the emergent properties a computer was designed to have.

"...Yet the basis of human activity is physics. But physics left to itself does not produce computers."


I'd say physics left to itself produced stars, which produced the elements of which the Earth is composed. Physics occurring on the Earth through evolution produced brains, and brains can reasonably be considered computers. (Though not digital computers.) The operation of brains is still physics and resulted in the production of digital computers. So in a roundabout way physics left to itself did produce digital computers. We just don't tend to think of ourselves as being aspects of "physics left to itself".

Thoughts?

If consciousness does not arise from the physical properties we know, and it does not arise from something like panprotopsychism (and I'm sure many here do not believe it does), then what?


I'm virtually certain consciousness arises from the sort of information processing which neural networks are good at. Though I'm not going to go into detail about why I am so certain.

Not to say there aren't a lot of unknown details to how consciousness arises, but doesn't information processing seem likely to be the substrate on which consciousness is built?

I've never understood why so many philosophers seem credulous towards panpsychism.

FYI, your counterargument commits the fallacy of denying the antecedent.