I don't see the word "consciousness" in the article. I thought that was the thing to figure out to understand the emergence of the mind.
Consciousness, on the other hand, is (even) less well-defined and is usually considered to be subjective. Being subjective, it tends to resist all of the usual objective approaches of description and analysis. Hard problem and all that.
Similarly to how computer is computer and half sized computer is half of its bigger friend – you can keep halving it until there is no "computer" left in it.
Or pencil – you have pencil that you call pencil; what about pencil half size of it? and so on until you hit single atom. You had pencil, now you don't, where on this line there was pencil and then there wasn't?
Edit: you can’t measure “pencilness,” but you can’t help but know whether or not you’re in pain.
These are all proxy measures, but all of science is done by proxy.
What is mind emerging into? When a video game experience emerges from the combination of processing, display, sound, and controller input, it emerges into a level of organization that a mind can participate in. It emerges into a system of organization emanating downward from the mind experiencing it. It can't just "emerge" into existence on its own. If a game falls in the woods, its not a game.
If you call the mind an emergent phenomenon but can't describe the context into which it emerges, you've added nothing to our understanding.
The emergent experience exists at the level into which it emerges. It's constructed at a lower level of organization, but not decomposable to them in a way that's meaningful without their recomposition—that's what makes the phenomenon emergent. The qualitative experience of a film does not meaningfully break down to the bits in the video stream, the compressed sound waves carrying the dialog, the photons hitting your eyes' rods and cones, or the biochemical signals in your brain.
The mind is not in the brain, but on the brain.
If a car is hurtling towards you, you don't perceive its handle. But if you're trying to go somewhere, you have to open the door. "Threat", "vehicle", or "handle" aren't just convenient framing device, but an accurate depiction of the object within your perceptual/motivation systems based on the current level of organization and analysis you're participating in.
We choose to see the mind because we are minds. Consciousness is. There is something which it is like to be. Denying our emergent experience of it, or reducing it to a "convenient framing device" tosses out the most fundamental empirical experience we have: to exist.
A video game is relatively easy, at least seemingly, to reduce down to its underlying principles. The content dissolves the more closely I look at the game. The issue here isn’t whether the game still exists (it does, in the place I’m no longer looking), the issue is in seeing why the game arises from its component parts, and not something else. Easy-ish for the game, it follows directly from what we know about physics and such, but hard for the mind. Why do neurons together produce pain that exists, rather than pain as a purpose-driven internal signal to help organize the escape from a predator? Emergence doesn't tell us why one or the other, just that whatever it is must emerge from constituent parts.
The game is certainly easier than the mind—I like it as an example because most of us have a hands-on knowledge of what the qualitative experience of "playing a game" is like. But the game still only emerges because the game developer, computer manufacturer, and player jointly give it a emanating system into which it can emerge. On its own, the raw game data doesn't really mean anything at all—if the bitstream of Diablo IV washed up on the beach, there's nowhere in that data encoding the experience of killing Diablo for the first time. One wouldn't even recognize it as something that could be decoded into such an experience [2].
I agree with you that the "why" is tough. Why have a conscious experience? Why have a sense of self at all? Why experience emotions rather than have them be—like you described—a purpose-driven internal signal? And then you get into theories like Internal Family Systems, which has empirical support at least within a prescriptive context if not necessarily a descriptive one [3].
The whole thing is a mess. A great, big, beautiful mess.
[1: https://www.amazon.com/Case-Against-Reality-Evolution-Truth/...] [2: https://benjamincongdon.me/blog/2021/02/21/Three-Layers-of-I...]
Consciousness is the emergent, graduated phenomenon of an information processing system when that information processing system has achieved significant complexity to model itself with relation to the various systemic inputs.
It's not binary, it's a gradient. I have more developed consciousness than my dog, which has more developed consciousness than a rat, and then a fish, then an insect, etc.
Somewhat disturbingly it also goes the other way, AIs may achieve a more profound conscious experience than humans - same for aliens. What does it mean for inferior forms of consciousness that have always placed themselves on a pedestal in relation to the rest of the animal kingdom simply because they have the most developed consciousness?
What you're describing is what's been proposed by Giulio Tonini as "Integrated Information Theory" (IIT) [1]. I quite like the framework and the math behind it is beautiful. Unfortunately, it hasn't been supported well empirically.
Re: AI, IIT actually gives basis for AI not being conscious. Not to mention that all conscious systems we can currently observe are dynamic/continuous, not discrete. The difference there is qualitative—there's no reason to assume that because a dynamic system is conscious that a discrete system approximating it is conscious too.
The article you cite [0] is labelled as "opinion". The standards for peer review of opinion articles in scientific journals are a lot lower than those for ordinary research articles. While precise standards vary from journal to journal, for opinion articles peer reviewers often see their role as simply excluding egregious misinformation and blatant errors, as compared to research articles where their role is to make sure the article is presenting high quality evidence in support of its conclusions. [1]
[0] https://www.nature.com/articles/nrn.2016.44
[1] https://ecologyisnotadirtyword.com/2021/02/24/lets-talk-abou...
https://journals.plos.org/ploscompbiol/article?id=10.1371/jo...
Scott Aaronson’s attempted refutation of IIT - https://scottaaronson.blog/?p=1799 - I think is better in that he actually tries to relate IIT to some of the philosophical literature (e.g. his distinction between Chalmers’ “Hard Problem” and the distinct “Pretty Hard Problem” which he sees IIT as trying to address)
I think it is a pity that Aaronson has never (to my knowledge) published his criticisms of IIT in a more formal setting-and I don’t know if Tononi has responded to them anywhere. I think Aaronson is probably right - that IIT fails as a mathematical model of what we intuitively consider conscious, since even though it excludes many common electronic devices we wouldn’t “conscious”, it is possible to mathematically construct an algorithm, capable of being physically implemented in electronics, which would be conscious per IIT but not per our intuition. And even if Tononi patches his mathematics to solve a particular case of that problem, someone with Aaronson’s skillset may just be able to construct another.
Tononi might then argue that if there is no mathematical model of our intuitions about consciousness lacking in special pleading, that’s a sign our intuitions are flawed. Okay, but then if we accept our intuitions can be flawed in some cases, why not in more cases? One could decide the intuition of consciousness is completely erroneous and become an eliminativist about it. Or, if IIT forces you to accept (contrary to our intuitions) certain (special cases of) simple electronic devices or computer systems as just as conscious as humans, why not violate those intuitions further and insist on that for even more cases?
My personal belief is that consciousness requires dynamic continuity. I don't think an algorithmic system is conscious because it's "cognition" is discrete and the information isn't integrated across frames. I don't have a "why that works"—its just a gut belief.
Stanislas Dehaene's book Consciousness and the Brain does a great job of describing this, though it's 10 years old now.
consciousness = meta-attention