Folks,

I just read an blog post by Martine Rothblatt arguing that Consciousness can be created in software. I've always assumed that if you believed in qualia, that you would disagree with that assertion. But now that I better understand the Consciousness arises from any equivalent functionality camp, I wonder if my assumption has been mistaken.

Those of us in the Nature has ineffable phenomenal properties camp surely believe this to be false, but what do all you arises from anything folk believe? Do all of you unanimously agree that software can or cannot do it?

Thanks,

Brent

Hi Brent

I disagree fundamentally with Martine's proposition. There is no consciousness by virtue of software, whether as a virtual machine or a relational database system or a network of artifical neurons with myriad weighted connections or any other sophisticated programmed artefact. Yes you can simulate some (perhaps, eventually, all?) aspects of what the brain does on a computer but this is empty: the input-output function is the same but the essential functionality is missing. It flies like a duck, quacks like a duck etc. but it isn't a duck! This is not because of some lack of spirit inhabiting the machine but because the mechanism - computation - cannot support consciousness.

Information system theory has nothing to say about the hard problem though some of the ideas are relevant to how the brain represents the world if only to tell us how it doesn't do it!

"Subjectivity is simply each person’s unique way of connecting the higher-order neuron patterns that come after the sensory neurons." seems to me softwarebabble - what does it mean? Martine confuses artificial neural networks with the real thing: "The strengths of neuronal couplings can be replicated with weighted strengths for software couplings in relational databases."...

There again, at least some cognitive issues are addressed by her approach whereas claiming that nature has ineffable phenomenal properties is avoiding explanation completely ;)

cheers

Richard

I agree with Richard here totally.

john

Thanks for the responses Rich and John.

So perhaps we are all unanimous in the idea that "Softwqare can do qualia" is not the same thing as "Consciousness can arise from any functional equivalent"?

If that is true, then would anyone object to adding a paragraph like the following to this Representational and Real camp statement?:

    "Many people believe that today's abstract computers, based on
    software, are capable of producing qualia.  We believe this to be
    an absurd and obviously mistaken idea.  Today's abstract
    computers are designed so that it doesn't matter what the
    information, or the ones (red) and zeros (green), are
    represented by - by design.  while our brain, what we represent
    700nm (red) vs 500nm (green) light is represented by, what
    these representations are phenomenally like and how they are
    different is of fundamental importance."

And If everyone in this camp agrees to such an addition, I have yet another question I'm struggling with.

How is Chalmers "Principle of organizational Invariance", or the idea that "Qualia can arise from any functional equivalent system", any different than the idea that software can produce qualia? I can't see any difference.

Brent Allsop

"Software can do qualia" *is* the same as "qualia can arise from any functional equivalent". This follows from the Church-Turing thesis, that any Turing equivalent machine can carry out any computable function; and software running on a general purpose computer is obviously carrying out computable functions.

There is a theoretical problem if the brain is not Turing-emulable. That would mean that a neuron or collection of neurons could behave in ways that no computer could ever predict or model, no matter how complete the initial information. This is what Penrose and Hameroff have proposed: that brains function using quantum gravity effects, which involves new and non-computable physics, and therefore even simple one-celled organisms behave in ways no computer could ever adequately simulate. This is not a theory accepted by most scientists, but it is an example of how far you have to go in order to consistently maintain that consciousness is not computable.

Hi Stathis,

Thanks for speaking up. I figured at least some people in this most popular camp at this level felt this way.

I'd really like to know who and how many people think on each side of this particular "software can do qualia" issue and why, but I'm not seeing any easy way to do that in this topic given the current camp structure.

So I think the best thing to do would be to create a new survey topic on just this specific issue.

What does every one think of the name "Software can do Qualia?" for a new topic?

And in response to your theoretical problem if the brain is not Turing-emulable - I believe computation or computability, Turring-emulability, and all that are only about behavior / ability to model, abstractly represent, and so on. What something is phenomenally like is entirely and categorically different than behavior or abstract representations. Turing-emulable has nothing to do with what something is phenomenally like.

Yes, the intelligent behavior of qualia can be modeled in a way such that perfect predictions about any such behavior are possible.

But, just because something can model, or behave like, or abstractly represent something else, in no way implies that it is fundamentally and phenomenally like the original.

Brent

A survey topic would be good. Are there other examples of that on Canonizer that you could point to?

As for whether an emulation would *necessarily* emulate the phenomenal properties, that is the whole point of the discussions around this topic. Functionalists believe that it would, essentialists believe that it wouldn't. Here are three positions:

A. The behaviour of the brain can be copied, and necessarily that would also copy the consciousness of the brain (functionalism, eg. David Chalmers, who as discussed has an argument purporting to prove this which I have not yet seen refuted).

B. The behaviour of the brain can be copied but that would not necessarily result in the consciousness being copied (eg. John Searle).

C. The behaviour of the brain cannot be copied and therefore neither can the consciousness (eg. Roger Penrose).

Hi Stathis,

Yes, all topics at canonizer.com are basically surveys. Their goal is to concisely and quantitatively represent what all participants believe on that topic. (It's not to determine truth, as many people falsely assume - just to document or survey what people think is truth.)

I've just created a new Software = Qualia? topic including a starting statement for the No camp which I'm in. As usual, I'm very willing and motivated to change things in order to win more people to this camp.

As far as Chalmers 'purported proof', along the lines of replacing neurons with silicon, I'm in the camp that this idea is based on a Transmigration Fallacy. So obviously I don't think it is a proof at all.

If you think there are problems with thinking there is a transmigration fallacy with this purported proof, I'd love to get this canonized in that topic on Transmigration and find out how many people would agree with you.

Brent

OK, I hadn't seen a simple YES/NO topic.

As for the "transmigration fallacy", I think what you are basically saying is that the behaviour of neurons is not Turing emulable. If some of the neurons in your visual cortex were replaced with electronic analogues, you would say that part of the strawberry has a weird greyish colour, not the red you are used to. That would mean electronic analogue is unable to exactly copy the behaviour of the replaced neurons; for if it could do so, you would compare the strawberry you see with the strawberry you saw before the replacement and declare it to be exactly the same.

The only other possibility is that the strawberry *looks* different but you don't *notice* anything different about it. But the implications of being consistently deluded about your qualia would be pretty weird. It would mean that strawberries could look completely different to you every Tuesday, but there is no objective or subjective way for you or anyone else to know this, so no reason to worry about it. And if you don't worry about strawberries changing colour without you knowing it on Tuesdays, why should you worry about them changing colour without you knowing it as a result of neural replacement?

There is a distinction to be made between electronic analogues and software analogues.

Consider an oscillator: this can be directly implemented using analogue electronics or you could drive a DAC from the output of a suitable program so that the periodically increasing and decreasing numbers produced by the computer are converted into a rising and falling waveform. The electronic waveform of each device is the same but only the analogue device actually oscillates.

I think this distinction is crucial and is at the heart of my objection to computational functionalism. IMO replacing neurons with digital devices runnng programs removes consciousness whereas replacing them with functionally equivalent analogue devices preserves qualia.