Anil Seth Finds Consciousness in Life’s Push Against Entropy
quantamagazine.org
quantamagazine.org
I'm not sure I agree though with the connection between life and consciousness in the article. I think it's very possible to have things that are alive in that entropic-displacing sense that need not be conscious or intelligent at all, say fairly trivial, self-replicating machines.
I am not a physicist but the Hawking model that black holes destroy information always struck me as an asymmetry. Apparently that asymmetry also motivated some researchers who know more than me.
Is there actually evidence for the conservation of information? I'm not even sure what information refers to within this context, but the writeup on the paper I read used that term.
I have toyed with the idea that the anthropic principle doesn't just rely on chance events on really long time scales. If indeed information is indestructible, the universe could be in a birth-death cycle of refinement, where it orders information in ways to increase self-awareness each cycle.
https://en.wikipedia.org/wiki/How_many_angels_can_dance_on_t...
If you know of JMG one statement of his was roughly this: Every school-child (of that time period) used to know that an 'eternal being' was outside of time, not a being [within time] that would "live forever."
That is, 'eternal' apparently had a precise sense [then] which most are now ignorant of. Again, I could be wrong.
It is hard for me to look down on curiousity about "what could be outside of time" when it remains a popular science question -- i.e., where did the big bang happen?, etc
People often make this rhetorical mistake when speaking about the second law. Entropy is increasing full steam ahead, and what looks like a small bit of order is revealed to be hiding ‘the man behind the curtain’ i.e. a massive increase in entropy through the expulsion of heat, so more things in the universe can enter higher entropy states in vibrational modes, rotational modes, etc.
Life exists because it increases the entropy in the universe
This is an concise idea I sit on quite often.
Is this a commonly known idea? Among scientists/engineers/philosopher types? Would appreciate other thoughts.
We can predict the final total maximum entropy of the visible universe, we are about half way there. Life produces entropy, but not as ’efficiently’ as exploding stars.
Life causes entropy, but do you literally mean it exists because it increases entropy? I don’t see the connection, only that life causes entropy. Can you explain?
Of course local entropy can decrease, hopefully in a way that is beneficial to our little corner of the universe :)
Would recommend Schrodingers “what is life” if interested further in life/entropy
Wow, that also nicely explains how colonisation & impoverishment of some human groups is necessary in the current framework, so that high HDI can be achieved in other human groups (which then explains various historical facts).
This makes space exploration and attempts of automated resource extraction from celestial objects a noble goal, something that can achieve world peace (assuming the results are distributed).
If the number of states labelled ordered is substantially smaller than the number of states labelled disordered and a system moves essentially randomly between states, it is far more likely that it will spend most of its time in a disordered state.
Entropy is just the universe settling into a state that cannot be meaningfully differentiated from most other states; there are far more of these than differentiable states, and so it's overwhelmingly likely that the universe ends up there.
2. got any other ideas? (remember, we're talking about large scale, statistical systems)
I don't know if other universes with different maths are possible. Greg Egan's books take my to the limit of my conceptual capacity, but I do think there are limits to what it is possible to know from within this universe, because everything needs to be built up from axioms from this universe. I agree with your explanation, right now while I'm sitting here in this universe, but if the celestial teapot turns up this afternoon and takes us both to perpetual motion machine universe, we might start to see things differently
His point is that living organisms are, temporarily, islands of relative order in a universe that is generally less so, which is both a fact and consonant with the 2nd. law. As all living organisms, not just conscious ones, share this feature, it is hardly a distinguishing feature, let alone an explanation, of consciousness.
This has led some people to redefine consciousness to encompass all such systems, whether living or not - they all, supposedly, have a certain amount of of it, according to the proponents of panpsychism and its more quantitative offspring such as Integrated Information Theory (IIT).
To me (and I am a materialist in matters of the mind), this seems to be an attempt to force consciousness into the woefully inadequate explanations we currently have on hand. The sort of consciousness we are most interested in, and are most challenged by, is the sort of self-aware, language-creating, theory-of-mind holding consciousness that is most fully developed in our own species (and probably in some of our extinct predecessors), not the sort of consciousness that might be attributed to a paramecium or a thermostatically controlled building. It is not useful to lump together such a diverse range of phenomena, and we are only fooling ourselves if we claim that anodyne statements about preserving order against a chaotic universe are, in any useful sense, an explanation of consciousness.
For a much more thorough critique of IIT specifically, see Scott Aaronson:
Boy, that is not at all what I got from Klara and the Sun.
Anyway, the humans in Klara and the Sun don’t even question that aspect of AFs. It’s more of a biological snobbery and that the AFs are short lived like Blade Runner replicants, and replaced by newer models.
The funny part is every analogy we make is just something in this universe, so all we do is move the mysticism to a lowlier object.
If we eventually do find out that materials cause consciousness, then material will become mystical. Really, we cannot win.
This why AI will never happen in the generalized sense. Because as long as we keep thinking we can engineer a mind out of matter, we'll keep chasing our tails.
Matter is a by-product of the universal mind. And we all experience our lives as singular nodes of that mind.
This has been known for millennia and is the basis for the world religions (which are mostly dumbed down co-optings of this).
Also, Bitcoin fixes this.
Why not come to the concrete level and realise that without consciousness you would not get something to eat and quickly die. So it has to do with life. We're complex self-replicators in an environment with scarce resources. We rely on consciousness to continue to exist. It feels like something to be you because you got a lot to win or lose, you're in a game of life and need to choose your actions.
For the latest in consciousness research - look at: https://arxiv.org/abs/2109.12798 Here (p14) they make artificial consciousness testable by making mental models of action, and the conscious 'thought' is the difference between what the action that was expected vs the action that occurred.
Breaking physics is not how biology works.
Consciousness implies awareness and perception of the world (and not necessarily self-awareness and ability of introspection).
And life is a result of selection what works and fits by trial and error.
Cells are mechanisms and aggregates of other mechanisms. The more correct word would be a process. Biological processes, which is a biochemical processes.
> We see the world not as it is, but as it’s useful for us to do so.
> A computer that plays chess is actually playing chess. But a computer simulation of a weather system does not generate actual weather. Weather is substrate-dependent.
> This is why I tend toward the substrate-dependent view. This imperative for self-organization and self-preservation in living systems goes all the way down: Every cell within a body maintains its own existence just as the body as a whole does. What’s more, unlike in a computer where you have this sharp distinction between hardware and software — between substrate and what “runs on” that substrate — in life, there isn’t such a sharp divide. Where does the mind-ware stop and the wetware start? There isn’t a clear answer. These, for me, are positive reasons to think that the substrate matters; a system that instantiates conscious experiences might have to be a system that cares about its persistence all the way down into its mechanisms, without some arbitrary cutoff. No, I can’t demonstrate that for certain. But it’s one interesting way in which living systems are different from computers, and it’s a way which helps me understand consciousness as it’s expressed in living systems.
> I think the situation we’re much more likely to find ourselves in is living in a world where artificial systems can give the extremely compelling impression that they are conscious, even when they are not. Or where we just have no way of knowing, but the systems will strongly try to convince us that they are.
That's true, but it doesn't mean that the simulation doesn't have desired effects in the real world. The simulation will not make you wet, but it will keep you dry if you use its predictions and take your umbrella with you.
If consciousness is related to assessment of predictions and their effect on the organism, then those predictions can be assessed by living "through" them in the real or simulated world and there are significant advantages in doing this in simulation.
While scientific reductionism has its limitations (e.g. trying to fully describe the complexity of a single neuron in c.elegans is sort of a rabbit hole) you still have to understand at least some of the basic mechanisms, the "LEGO blocks" of cognition, before talking about such higher level of abstraction, especially when it's so ill-defined.
I think Feynman' definition of "cargo cult science" is appropriate here, where we are trying to explain our perception of reality via superficial "neural correlates", or attaching to it these arbitrary "complexity theories", etc., without understanding the fundamental driving factors, mechanics and constraints of perception, the "why" and "how" first.
I’m quite certain Demis Hassabis uses the term regularly, Nature should unpublish him.
Also: https://www.journals.elsevier.com/artificial-intelligence
We are biological machines. Sure, we have an awareness of ourselves, but so do computers in some contexts; for example you could argue a certain types of self-modifying or monitoring code has an awareness of itself. It's already been proven that animals have an awareness of themselves when looking in a mirror, such as Dolphins, Chimpanzees, etc. Humans may be a bit further along in our cognitive development but there is no obvious point where you can say "Humans are conscious" because we're all somewhere on a sliding scale of intelligence and self-awareness.
Everytime you wake up in the morning.
Some work detecting it as well for medical reasons https://www.science.org/doi/10.1126/scitranslmed.3006294
Having a representation of itself doesn't mean it has an awareness of itself.
Maybe or maybe not. This is the nub of the debate. I would argue that answering "yes" to this is a partial endorsement of pan-psychism. If the ability to experience qualia is property of certain algorithms or types of information flow then it's a fundemental property of the universe.
Thought experiments about p-zombies and mind simulation are an interesting litmus test to separate different points of view on this.
(sorry) This would mean, at a minimum, that every biological entity down to single cell organisms are conscious. They could be, I don't know.
In your lingo, something can know what it looks like without taking its aspects into consideration.
I think the inverse is true, though: something that cannot perceive a representation of itself cannot be self-aware.
For example, when an optometrist or neurologist talks about “sight” they’re talking about the biochemical mechanisms that allow us to see. When I read Plato say “I see therefor I am” it was clear to me he was describing the metaphysical act of personal perception. He was describing the “observer” that was present in the moment, perceiving the biochemical process, not sight itself.
Consciousness is something you likely have. But we (humans) don’t know what it is. To my knowledge, we don’t even know where to start looking for it. Consciousness belongs to the mystics today, by default, because science can not claim it. There isn’t something we can hold and point to saying “this, this is what the mystics call a soul.”
Science doesn't yet have an answer, but that doesn't grant the mystics any credibility. There is no 'by default' when we don't know. I'm reminded of a Dawkins quote:
Lecturer: Scientists answer questions of 'how'. If your question is 'why', I refer you to the theologians.
Dawkins: Why the theologians? Why not the gardeners?
(Trivia: If I recall correctly, Dawkins didn't actually give this response, he thought it up far too late.)
> There isn’t something we can hold and point to saying “this, this is what the mystics call a soul.”
That isn't a game-over, though.
The same goes for information processing, but we're fairly confident in what we call a 'computer'.
You can't hold a center-of-gravity, but we're able to reason about those pretty clearly.
> Dawkins: Why the theologians? Why not the gardeners?
I’ve never heard this. My initial reaction is that “god” exists in the unknown. Asking “why go to the mystics for mystical stuff” is a tautology. If you go to the gardener with a question unanswered by science, and they have an answer, they’re a mystic (or have made a scientific breakthrough).
> You can't hold a center-of-gravity, but we're able to reason about those pretty clearly.
Consciousness is not understood. Full stop. We don’t know what manifests it. We don’t know where it comes from. We don’t know where it goes.
We’ve scratched the surface of the biological mechanisms that give rise to the phenomenon we experience. If you look at me and see a biological computer, I think your statement holds. But when I look at me, in a mirror, I see a biological vessel I occupy; and I don’t know what “I” is.
Trying to oversell our understanding of the universe and our place in it does a disservice to science and the work left to be done.
Everything at one time escaped basic explanation from parts (and there always will be, you can't really explain why the standard model, or whatever more fundamental system which might surpass it)
Consciousness is on the edge of being understood quite a lot better as neurobiology is progressing quite quickly.
What qualia are made up of are not easy to be accurate about. My favourite thought experiment that interrogates this question is,
Of course it isn't. We have a very poor understanding of brain states. So all we can say is that Mary will learn something new today. But it's not reasonable to extend this to an assumption that a physicalist explanation of qualia is impossible in principle and will become available in the future.
As it happens I'm not a physicalist, and I suspect - but can't prove - that physicalism won't solve this problem.
But I also don't like poor arguments, and I think the Knowledge Argument is not a good one.
The problem is more fundamental. Qualia are definitively subjective and the only way to prove that physicalism explains them is to somehow make them objective - with some kind of qualia-ometer. Or consciousness-ometer. Or something similar.
That doesn't mean finding correlates - neural states, chemical processes, quantum uncertainty, whatever. It means being able to measure experience itself.
Without that you can build simulacra that show all the correlates, and possibly behave as if they're conscious.
But all you've done is built a robot. You can't prove it actually has experience - including self-awareness - unless you can measure experience with objective instrumentation.
This is a nice paradox, because it requires science to measure subjectivity itself.
It may or may not be possible. But clearly it's not possible now, and is unlikely to become possible any time soon.
We build things all the time based on the models of the universe we've got. Your argument strikes me as more or less saying, "sure, we built that jet engine, but unless we can measure what 'being a jet engine' is we've just not really built the complete, real thing."
If we build a thing that appears to have "experiences" indistinguishably from how humans appear to, then that is it. That is, the notion of "experience," as you describe it, is more or less meaningless in the context of the very physical world we occupy.
“I think that consciousness is the way information feels when being processed in certain complex ways.”
Regardless of the breadth of the spectrum of definitions, at the very bottom you find phenomenal consciousness. The given, undeniable fact that "it feels like something". We don't have a plausible avenue, not even in principle, to even start addressing this fact of existence, to the point that the most rational stance is to assume that it is fundamental in nature.
IMO Biology explains consciousness fine. Reinforcement of sensory data given repeat exposure leaves agency memory.
I think “higher abstraction” is such a flimsy term as to be meaningless.
Nothing is more or less abstract. It’s just the abstract it needs to be to “exist”. Humans are the abstract they need to be to exist on Earth.
This idea there’s layers of abstraction and meaning beyond what we can know literally is just a neural implant from years of having tribal elders dictating we fetishize a small set of bad ideas ad naseum, we conjure alternatives in order to escape.