Theoretical Mind Experts,

I just returned from the Singularity Summit conference in NYC last weekend where David Chalmers was one of the presenters. It was very inspiring to have a chance to meet him in person, and have the opportunity during down time to talk with him. As many of you know, he has been included in various e-mail solicitations where some of you found out about and started participating in this survey process. But at the time, Chalmers wasn't interested.

However, when I first went up to meet Chalmers after his presentation, he recognized me. He said I would be interested to know that he was working on a survey of his own to measure for scientific consensus. From what he described, his survey will be a traditional 'snapshot' survey where he writes the questions and the possible answers, and various experts at top ranked philosophical institutions are asked to take the survey. I think this is great that he is also realizing how important measuring for scientific consensus is in a still controversial field like this one. And I'm hoping whatever survey he comes up with, we can get it all 'canonized' so we can extend it, and find out what more than just the set of experts he chooses to have participate - in comparison to his results. I bet the results of any such survey will be very surprising and have profound effects on this field, giving the idea of the importance of measuring for consensus a big boost.

He also seemed interested when I pointed out how much consensus there was in our small survey here for his 'consciousness arises from anything functionally equivalent' camp, though he wasn't very impressed with the still small sample size. When I told him I was in a different camp, he was gracious enough to spend some time asking what it was I do believe, and discussing this with me. This gave me new insights in how I might rewrite the Transmigration Fallacy camp statement so it is, I hope, much easier to understand and comprehensively covers the important issues in a more clear way.

I of course would love to hear any thoughts on this, especially from people that are in this camp, and agree with this position. Also, I make reference to the Smithies Carr camp, and how it includes a critical mechanism to resolve the binding problem - which is lacking in the transmigration argument. I'd love to here if you guys think this makes sense.

I'm looking forward to sending a reference to this argument for review to see what David Chalmers thinks about it.

Thanks,

Brent Allsop

Brent,

It's good that you got David Chalmers interested in Canonizer. It does look like you are gathering a collection of experts on (at least) the consciousness topics, although the total sample size is not large. Do you think it would be better to keep it a small club or do you think it would be better to increase the numbers with larger numbers of people who have only a casual interest in the topics?

As for the "Transmigration Fallacy" topic, what you are effectively saying is that the physical behaviour of neurons cannot be emulated by computer circuits. That may be true, but there is no good reason to think that it's true.

Stathis says

"As for the "Transmigration Fallacy" topic, what you are effectively saying is that the physical behaviour of neurons cannot be emulated by computer circuits. That may be true, but there is no good reason to think that it's true."

Neurons are physico-chemical, not just physical, machines—is that important? For example redox mechanisms are important in their function which would not be emulated in any ordinary sense by computer circuits. Secondly I do not think the basic question is whether computer circuits can emulate to the full what is going on in the brain-consciousness system. The basic question is whether the events in the brain are identical, or not, with what goes on in a person's consciousness. This is a separate question to whether computers can emulate what is going on in the neuronal circuits.

john locke says:

"Neurons are physico-chemical, not just physical, machines—is that important? For example redox mechanisms are important in their function which would not be emulated in any ordinary sense by computer circuits. Secondly I do not think the basic question is whether computer circuits can emulate to the full what is going on in the brain-consciousness system. The basic question is whether the events in the brain are identical, or not, with what goes on in a person's consciousness. This is a separate question to whether computers can emulate what is going on in the neuronal circuits."

It is generally accepted that chemistry is computable, meaning that you can write an algorithm which will allow you to model a particular chemical reaction and predict what it will do. For example, given an amino acid sequence, it should be possible to calculate its 3D structure under certain pH and temperature conditions, calculate its conformational change when exposed to particular neurotransmitters, and calculate how this conformational change would alter its permeability to sodium and potassium ions if the protein molecule is embedded in a phospholipid bilayer membrane. In this way, the behaviour of a whole neuron or a whole brain could be modelled by the appropriate computer program. Computationally very difficult, no doubt, but the field of computational neuroscience is devoted to solving such problems.

Now, it is possible that there will be some aspect of nature that is just not computable, and that neurons rely on this and the behaviour of brains is therefore not computable. For example, it may be that neurons rely crucially on infinite precision real numbers to function, or it may be, as in the Smythies-Carr hypothesis, that the operation of neurons includes interactions with branes in other dimensions, the physics of which is not computable. In these cases, perhaps brain activity and therefore consciousness cannot be modelled by a computer program. But you have to ask, what reason is there to believe this, and why would evolution not just use straightforward chemistry to build brains, as seems to be the case, but rather go to such trouble to confound us all?

As for the question of whether the physical activity in the brain is identical to consciousness, I'm still not sure why you consider it so important. I'll say the physical activity in the brain is not identical to consciousness, but is sufficient to give rise to consciousness. Is that acceptable?

I agree with what John Says:

    "The basic question is whether the events in the brain are
    identical, or not, with what goes on in a person's consciousness.
    This is a separate question to whether computers can emulate
    what is going on in the neuronal circuits." 

All of this can emulate and model the other. But to emulate, or behave like something else isn't all that is important, and misses some of what something is fundamentally and phenomenally like.

The surface of the strawberry has a behavioral property (behavioral or traditional red), such that it reflects 700 nm light. Though the light can 'model' this property in much detail, it is not fundamentally, and likely also not phenomenally like the surface of the strawberry. This abstracting process repeats itself thorough the entire chain of cause and effect perception process. Again, each downstream representation models the previous, and the initial behavioral property that started everything - but non of them are fundamentally like any of the others. And, though each subsequent level can model any property, whether phenomenal or not, the only way to know what the original is fundamentally like, is to ground such representations with the original (effing the ineffable) so you can map the perfectly complete representations back to what they are phenomenally like. The only representation that we truly know what it is like, is the final one. The phenomenal red. Phenomenal red is a property of the final result of the perception process or our knowledge of the behavioral red that started everything. All of this perfectly models the original - but till you map it back, you don't know what the original is really like.

Abstract representations can model anything, including random behavior. John, would you agree with the idea that even the multidimensional 'branes' and everything in the Smythies-Carr hypothesis could be modeled to any level of accuracy desired. But, again, the model is not phenomenally like the original.

Phenomenal properties can be modeled. But you must know what the original is like, and be able to map the downstream abstracted representations of such back to the original, to fully understand the full phenomenal meaning of what it is they are representing.

Brent Allsop

Stathis asked:

    "Do you think it would be better to keep it a small club or do you
    think it would be better to increase the numbers with larger
    numbers of people who have only a casual interest in the
    topics?"

Chalmers' bibliography of publications on this topic just blew by the 20,000 mark and is surely growing exponentially. Especially the ones that describe 'camps' that are radically different than my own, even if I spent a year on one publication, I'd still likely not fully understand it. And I'm certainly no Chalmers, and have a real job so can't spend all day every day trying to do so. And I bet even Chalmers doesn't have a clue about much of it.

I believe most of it is junk, but nobody can see the signal (I believe where there is the most consensus) from the noise. I also believe there is much more consensus than anyone realizes - it's just that nobody has attempted to measure such yet.

And also, everyone spends all their time talking about much less important things than what they agree on, and once they agree, they stop talking. You can see this very clearly in what we have so far - the consensus is what is contained in the Representational and Real camp. All the stuff that we disagree on, represented by the sub structure, is much less important than what is contained up there.

So many people jump to the mistaken assumption that the goal of canonizer.com, is to determine 'the truth' via a voting process. This is not it at all. The goal is to concisely state all the camps that I don't fully understand, and quantitatively measure who and how many people are in each camp, and finally have a useful survey and understanding of what everyone else believes outside of what I believe and why.

The hope is, that we can use methodoligies, like the ones we've started to collect here, to achieve some kind of concise and quantitative survey of the 20K publications, and to better know and communicate to each other what everyone believes on all this.

And once we achieve this, the nuts and bolts researchers will finally realize what they are looking for (qualia), where to look for it (not on the surface of the strawberry), and how to look for it (effing the ineffable). And I believe once they know all this, they'll simply point and say - there it is - it was there all along, it's just that nobody ever successfully communicated to me what, where, and how I should be looking for it.

And then, finally, all the other masses of hurdling mistaken sheep in the wrong camps, all over the world, will finally convert to THE ONE true camp (may not be a qualophile camp after all), with the sheep dogs of demonstrable scientific results nipping at their heals. The goal of canonizer.com is to rigorously measure this process, and hopefully by significantly increasing our ability to communicate, to dramatically speed it up.

Brent Allsop

Brent,

You haven't answered the question (that I recall) of what it would be like if half your brain were replaced with a functional equivalent which lacked qualia. By definition, you would behave in exactly the same way, but half your qualia would be gone. Is it possible that you wouldn't notice this, or would you notice it but be unable to say anything or change your behaviour in any way, as if your body were being manipulated like a puppet by an outside force?

The questions about emulating neurons, chemistry, etc., are all enlightening as are ideas of computational neurobiology. However, I think it is still important to focus in on some of the ways that cellular activity differs from our normal computer equipment.

What I refer to, of course, is in the 10^20 water molecule per second generation rate (throughout the body) that occurs as a regular output of the life-supporting respiration reaction. Each molecule has a rather simple ++-- tetrahedral-like structure and thus as each is formed it can orient in one of six (perhaps twelve -- a small number) of ways, where sequences of molecules generated at one site provide an on-going ~6^n internal analog math.

Now, the interesting part here is this level of activity is internal to neurons and induces structural changes while at the same time is engaged through hydrogen bonding in making subtle shifts in protein foldings as well as influencing the environment in which various replications are occurring. That is, the wild blast of structural chemistry flux is intimately entangled with all sorts of delicate arrangements -- all through the structural chemistry.

Clearly, we might catch a glimpse of such wonderful complexity, or notice parts of it sequentially, but the details, they are a bit mind-boggling. Suffice it to say that adding this level of on-going activity provides a way to have repetitive internal structural adjustments emerging from within that can "open and close switches" if one needs to think in such terms - ~separate from the down-gradient (energy-wise) neural network sparking.

This sort of more robust, multi-state order thing is where "neuro-computation" is substantially different from a Univac or x86 type computer ...and, in other respects also is quite similar.

Think about it.

Ralph

Stathis,

Wow, it sure is hard to communicate on this topic. From my perspective, I've clearly answered this question in the recent rewrite of the transmigration fallacy camp.

The group of neurons or the 'isomporph', is closer to one neuron, or whatever it is that represents something like one voxel, because I believe it is easier to understand at this level. But the same argument, and answer, applies if you scale this isomorph up to include an entire brain hemisphere. Does it not?

I'm really struggling, trying to better understand how to communicate, what it is I'm trying to say. Thanks for your continued effort to communicate back, what it is you understand (what I'm missing?), and where I'm still failing.

Brent

Ralph,

You suggested:

    "Think about it"

Everything you are talking about here is only about behavioral properties. I don't yet know much about the 'wonderfully complex' 'tetrahedral-like structures' of all this behavior. But, first, for me at least, understanding the what, why, how, and where of phenomenal properties is much more important. So this is all I find I have time to think about and be interested in at the moment.

Brent Allsop