Sic et Non self deconstructs

[_huckelberry]

Moving on from sentient beings to the criteria of self awareness as a requirement for animal consciousness, the list becomes considerably shorter. Animals that pass the mirror recognition test for self awareness include apes, monkeys, elephants, bottle nose dolphins and magpies. (Physics Guy, please note that all are wild.)

According to the literature, animals that pass the mirror test for self awareness exhibit the following behaviors:

- social response,
- physical mirror inspection,
- repetitive mirror testing behavior, and
- the mark test, which involves the animals (except for dolphins. elephants and magpies) spontaneously touching a mark on their body which would have been difficult to see without the mirror.

The magpie certainly seems to be an outlier. Many bird species exhibit planning and the ability to solve new problems, especially related to release from captivity or obtaining food. Crows and parrots are a good examples. Why is it that magpies are the only ones (to date) to pass the mirror test for self awareness?

Why do dogs, some breeds of which are clearly quite intelligent, and who we humans anthropomorphize to a great extent, fail to pass the mirror test?

Please pardon if I have no genuine science knowledge for this subject but think it is an interesting question. I will float a notion. If one considers human seeing you will find that we do not see in direct reporduction like a camera but have an inbetween stage where we form ideas of what we see based upon visual clues. Our eys will continue to feed us information which we use to clarify our field of vision. I would suspect that different animals would proceed with this process differently because of different patterns of idea formation. For dogs smell forms such a large portion of their understanding that they may give mirrors little attention because a mirror lacks the appropriate smell information.

I have dogs and have watched and wondered about how they experience and think. They seem to have limited interest in television, images of dogs, mirrors and the like. If a dog is on television there is not much interest unless it barks, something that enters the area of dogs interest. Dogs communicate, ask questions and hope from responses from me. They play and are full of curiosity. Obviously they are as conscious as you or I but that does not mean it is the same as human of course. Dogs think dog thought and have dog concerns. I am not quite able to imagine what thoughts would be like if smell was as big a part of awareness as it is for a dog.

[_Physics Guy]

I don't understand the difference between reductive and non-reductive materialism. Since like most people I like analogies to things I understand, I like physics analogies. So I think of the relationship between neurobiology and psychology as like that between molecular dynamics and fluid mechanics.

Molecular dynamics is describing the motion of zillions of molecules flying around—or, if they're too densely packed to fly very far, jostling around—and exerting forces on each other. As one might expect, it gets harder to understand what's going on the more molecules there are. What you might not expect is how fast it gets harder. The largest number of molecules for which we can exactly understand the interacting motion is, um, two. So molecular dynamics depends on a lot of approximations. Still in principle it is trying to talk about the zillions of molecules.

Fluid mechanics describes gases and liquids as if they were continuous "stuff". It never even mentions any molecules moving. It talks about pressure and density and flow velocity and viscosity and things like that, none of which refer to molecules in any way. Fluid mechanics has its own profoundly challenging problems but it certainly works. We build airplanes with fluid mechanics and they don't fall out of the sky.

We can derive fluid mechanics from molecular dynamics. All the concepts of fluid mechanics can be translated into molecular terms; the density of the gas is just the average number of molecules per unit volume times the mass of one molecule. The fluid velocity at a point in space is the average velocity of all the molecules within a small volume of space around that point; and so on. In fluid mechanics, the various properties of different kinds of fluid are all just empirical facts that no-one tries to explain, but in molecular dynamics one can compute things like the different viscosities of oil and water, from the properties and interactions of the respective molecules. The calculations are difficult and not always completely rigorous but we can pretty much connect the dots to get from molecules to fluid mechanics.

So has fluid mechanics been reduced to molecular dynamics? In one way I'd say Yes, absolutely. If this isn't reduction then nothing is. It's perfectly true that what makes airplanes fly is "nothing but" the motion and interaction of zillions of molecules. Pressure and viscosity are not additional things which have to be postulated alongside the molecules; they are simply high-level statements about the molecules.

Even if fluid mechanics can in principle be conceptually eliminated, however, it is never going to go out of style. It's so much easier to use than molecular dynamics and yet it works really well. If aeronautical engineers tried to abandon fluid mechanics in favor of designing new planes from pure molecular dynamics, they'd get fired.

Furthermore there is nothing wrong with fluid mechanics unless you really care about very fine details. Pressure and viscosity are not illusions. They're perfectly real, and molecular dynamics does not deny or even change that. All that molecular dynamics does is explain exactly how molecular motion and interaction make pressure and viscosity work.

And finally, fluid mechanics remains logical independent of molecular dynamics. In reality, gases and liquids are made out of molecules, and all the concepts of fluid mechanics can be expressed in molecular terms. They don't have to be so expressed, however, if all you want to do is fluid mechanics (for example, to design a plane). Fluid mechanics is a self-consistent subject that was formulated before we even believed in molecules.

The most important aspect of logically independence is that molecular dynamics hardly ever adds anything interesting to the fluid mechanical explanation of any phenomena. Molecular dynamics can tell you a story about why the viscosity of water is what it is; fluid mechanics has no story for that. But if you have a fluid mechanical explanation of how a tornado works then all that molecular dynamics can do is tell you the same story again in much more cumbersome language. In that sense the reduction cuts both ways. Fluid mechanics is a brilliantly concise and yet accurate summary of molecular dynamics, which is why in practical applications it is fluid mechanics that eliminates molecular dynamics instead of the other way round.

This relationship between molecular dynamics and fluid mechanics is my model for reduction in general. If and when we understand the neural mechanisms of consciousness I think that the relationship between neurobiology and psychology will be essentially similar. We'll be able to tell stories about why psychological facts are what they are and this will be great. It won't make the facts false. We'll still describe mental phenomena in psychological terms, for good reason.

We won't ever discover that love isn't really a thing because everything is just atoms. We'll discover how large groups of atoms can love.

[_Gadianton]

I've cleared out some stuff from my post above that I think was problematic.

So has fluid mechanics been reduced to molecular dynamics? In one way I'd say Yes, absolutely. If this isn't reduction then nothing is.

I agree. The question is if the experience of pain or taste of steak is reducible in the same way.

The point of functionalism is:

pain = c-fibers firing

pain = some pattern of silicon transistors or some alien physiology that isn't neurons.

If people could be simulated with experiences within a computer, then the experiences don't logically reduce to physiology of neurons, but something that is in common with other very different hardware that produces the same thing.

[Guest]

I have no opinion on any of Arc's questions about consciousness because I don't know what consciousness is. My working assumption is that consciousness is a physical process, kind of in the same category as combustion or nuclear fusion tbough of course not much like them. I just don't know what the process is in the case of consciousness.

Last time I followed the field, no-one did. A couple of people maybe thought they did but they convinced no-one else. The fact that Arc can ask for my opinion about what kinds of organisms might exhibit consciousness is a sign to me that we're not much farther yet. Nobody asks for anyone's view on whether only red giant stars have fusion or whether gas giant planets do, too.

The last time I followed neuroscience at all was 25 years ago, but I've been assuming that if in the meantime we had discovered the physical mechanism of consciousness then I would have heard of it anyway because it would have been huge news. Did I miss it? If so, what's the scoop?

There have been recent advances in the understanding of consciousness. However, advances in consciousness research are not like many of those in physics or astronomy, where the shadow of a distant black hole is obtained by an international effort using data from eight synchronized radio telescopes in different countries, and where the breakthrough can be summarized in a single image and a paragraph of text.

Advances in the understanding of consciousness come from careful experiment and observation by simulating and manipulating the brain and noting cause and effect relationships. Understanding has been gained by extensive observation and testing after dosing with drugs, brain injuries, magnetic stimulation to specific brain regions, and certain medically indicated surgical procedures. The more we explore the brain, forming and testing hypotheses as we go, the more we understand.

To get an idea of what is currently being taught about consciousness in colleges and universities, I had a look at the Neuroscience of Consciousness course description and reading list for PSB 4934 at the University of Florida for 2018. Here is the required reading list:

Dehaene, Stanislas (2014). Consciousness and the brain: Deciphering how the brain codes our thoughts. Penguin.

Nunez, Paul L (2016). The New Science of Consciousness: Exploring the Complexity of Brain, Mind, and Self. Prometheus Books.

That's it. Neither one is really a textbook. Both are on my Kindle reader. The first is much more informative than the second.

The first book listed, Dehaene, covers work on understanding the specific brain functions that support perception and memory, both at the conscious and subconscious levels. He defines consciousness as simply "brain-wide information sharing."

When it comes to what David Chalmers has called the “hard problem” of consciousness, Dehaene responds that "mental experience is a pre-scientific concept that will disappear as we better understand the connections in the brain." So much for that.

Giulio Tononi holds the Distinguished Chair in Consciousness Science, at the University of Wisconsin, and his work should be required reading in any such course. Along the same basic lines as Duhaene, Tononi and colleagues have developed the integrated information theory (IIT) to explain consciousness.

IIT starts by recognizing the reality of consciousness and then looking for the kinds of physical substrates that could support it, or from which it could emerge. Being a physicist (If I recall correctly) by training Tononi and colleagues couch IIT in terms of axioms and postulates. https://en.wikipedia.org/wiki/Integrated_information_theory
Using IIT principles applied to the brain's responses to transcranial magnetic stimulation (TMS), the group has been able to assess level consciousness using a perturbational complexity index (PCI) in-

"--healthy subjects during wakefulness, dreaming, nonrapid eye movement sleep, and different levels of sedation induced by anesthetic agents (midazolam, xenon, and propofol), as well as in patients who had emerged from coma (vegetative state, minimally conscious state, and locked-in syndrome). PCI reliably discriminated the level of consciousness in single individuals during wakefulness, sleep, and anesthesia, as well as in patients who had emerged from coma and recovered a minimal level of consciousness."

That is, they could stimulate specific regions in the brain using magnetic coil sets placed on the skull, and then watch the brain's regional electrocortical (EEG type) response. They could see, with some precision, which centers in the brains of various non-responsive subjects reacted to the TMS pulse when originating from specific locations. There were readily identifiable differences among non responsive patients, allowing classification of their treatment or deficit. Looking at the data it was clear that responses from awake healthy subjects involved more brain regions than reported from those who were non-responsive. These resulting data support the idea that consciousness is indeed "brainwide information sharing."

Tononi "proposes an identity between phenomenological properties of experience and causal properties of physical systems: The conceptual structure specified by a complex of elements in a state is identical to its experience."

The question upthread about where along the phylogenetic scale does consciousness emerge is one in which Tononi was interested. IIT would say that there are various levels of consciousness along the phylogenetic scale, just as there are various levels of intelligence, and that the two show a positive correlation.

The question that naturally arises in the context of this thread is, "where along the phylogenetic scale would religionists say that the soul first appears?" Any possible answer to that question renders the material soul highly unlikely as the source of consciousness.

My personal interest in consciousness has not much to do with the theory of mind. Rather it is in what can be learned about consciousness that will help maintain, and hopefully improve, brain function and repair the brain when needed, in order to make conscious life less stressful, less painful, more enjoyable and of greater value to the brain's owner.

[_Physics Guy]

Thanks, Arc. That's the kind of answer I wanted.

I'm not immediately impressed by "brain-wide information sharing" as a definition of consciousness, but it does seem like a worthwhile preliminary target. I bet that Volvox colonies could probably be said to have colony-wide information sharing in the sense that the behaviors of all the individual cells are correlated, but I also agree that usefully more sophisticated information sharing is probably a basic task for more advanced brains.

It's not clear to me that the information sharing of a Volvox colony would be anything that the colony had to work to achieve. The individual cells could just be reacting to their own local environments and the coherent colony behavior would simply reflect the fact that all the colony cells were in the same environment. Nonetheless the individual cells probably have evolved behaviors that are based on the nearby presence of other colony cells, and so the collective behavior of the colony isn't just a coincidence. The principle might scale up a long way. So thinking about information sharing might well be a good place to start.

Just getting neurons on both sides of our skulls to share information, in any sense, is not a trivial task. It only seems trivial if you take a crude homunculus view of consciousness. I don't think information sharing automtically means you have consciousness but it's not something we can just take for granted.

Advances in the understanding of consciousness come from careful experiment and observation by simulating and manipulating the brain and noting cause and effect relationships. Understanding has been gained by extensive observation and testing before and after certain medically indicated surgical procedures. The more we explore the brain, forming and testing hypotheses as we go, the more we understand.

Not all physics is Big Science breakthroughs, either. But I'm not dismissive of neuroscience at all. If we somehow had to choose what to fund, I'd shut down CERN in a heartbeat to keep the lights on for neuroscience. CERN has already found its Higgs boson anyway.

I just think that understanding consciousness is going to be a lot harder than understanding particle physics and we're a long way from the goal. The standards of science are still success and not effort. So I feel that humility is the scientifically appropriate attitude for neuroscience so far.

[_Arc]

I'm not immediately impressed by "brain-wide information sharing" as a definition of consciousness,.."

Due in part, no doubt, to my poor quote selection. I edited in a paragraph with a more sophisticated description of IIT from Tononi.

Also, as a physicist, you might be interested in Tononi's mathematical formulation of ITT. It should be included in, or directly linked to, the URL in my post.

Also, if you enter Tononi and consciousness as a search terms in Google, it will return links to videos of his lectures on the subject. The one with the slide presentation is worth watching if you are interested.

[_Physics Guy]

The point of functionalism is:

pain = c-fibers firing

pain = some pattern of silicon transistors or some alien physiology that isn't neurons.

My problem with the apparently heated debate about eliminative versus non-eliminative reductionism is that I have a hard time seeing how there can be a debate when either one side or the other just sounds obviously stupid.

It is the intrinsic nature of summaries to be many-to-one. That's the whole point of summarizing: leaving out the less important information is leaving out information. So any useful higher-level effective theory is by definition going to be compatible with multiple lower-level theories. If neurons can collectively exhibit pain, then it's perfectly plausible that semiconductor chips can, as well.

It's not guaranteed. Perhaps pain is a phenomenon that somehow does depend on some microscopic features that aren't there in silicon. I find that unlikely but it could perhaps be true. Lots of different kinds of atoms can be solids or liquids but not every kind of element can make up a superconductor. Perhaps pain will turn out to be more like superconductivity than solidity.

Perhaps sensations like the taste of a steak are not so much like solidity as like the specific shape of an individual snowflake. So maybe nobody tastes things in quite the same way. I still don't see anything to debate, here. If we agree there's no additional substance in consciousness then what's left to dispute?

[_Physics Guy]

The conceptual structure specified by a complex of elements in a state is identical to its experience.

I'm not exactly sure what this means, but it sounds like the kind of thing I would end up saying myself if I had to state the obvious. By which I don't mean to knock Tononi. I'm saying that I think I agree with him. I just feel as though the statement itself is only the obvious starting point and the real work is to figure out what it actually means for any "complex of elements" in the brain to "specify" a "conceptual structure."

[_Arc]

The conceptual structure specified by a complex of elements in a state is identical to its experience.

I'm not exactly sure what this means, but it sounds like the kind of thing I would end up saying myself if I had to state the obvious. By which I don't mean to knock Tononi. I'm saying that I think I agree with him. I just feel as though the statement itself is only the obvious starting point and the real work is to figure out what it actually means for any "complex of elements" in the brain to "specify" a "conceptual structure."

In my library are a number of standard texts in the neurosciences. However, the author from whom I learned the most in a short time about brain function was Ray Kurzweil. His How to Create a Mind: The Secret of Human Thought Revealed is an oldie but a goodie. You may know this guy from his work on electronic keyboard instruments. Yep, that's him.

He has also done groundbreaking work on text-to-speech synthesis, and speech recognition technology. He approaches the subject of mind as one who understands neural networks and builds on that understanding, relating it directly to cellular organization and structure in the brain. Once I gained an appreciation for how this stuff works from a neural network standpoint to overlay on to the standard text subject matter, consciousness became more of an academic issue.

https://www.amazon.com/dp/B007V65UUG/ref=dp-kindle-redirect?_encoding=UTF8&btkr=1

[_Gadianton]

If neurons can collectively exhibit pain, then it's perfectly plausible that semiconductor chips can, as well.

As I was saying, more people nowadays, especially atheists, would see it that way, supporting my point. so thanks.

If we agree there's no additional substance in consciousness then what's left to dispute?

one thing, would be whether the taste of steak fully reduces to properties of neurons or alternatively, to some kind of functional role of neurons that other materials like transistors could fulfill. As you pointed out, it's not guaranteed one way or another. If something isn't settled, then that indicates people will debate it. Another would be whether the taste of steak reduces to a functional role at all or alternatively, violates a minimal standard of physicalism.

Another would be whether or not there really is a such thing as a sensation. Eliminativism refers to eliminating mind or some aspect of mind. So back to Dennett, his computational model of mind differs from functionalism in that he doesn't see qualia, or sensations, as stable ideas. Some say he's an eliminativist in this regard e.g., we actually don't have feelings. I don't think that's right but, anyway, some of this stuff is similar to behaviorism vs. cognitivism. there's more obviously, but that's a couple examples.