Out cycling in the rain I started daydreaming about a theory of the development of self-awareness.
USUALY, when I have what I think is a 'clever' idea, some or many other people have had it first or its been disproven somehow or just doesn't work that well.
Whatever, put it in the comments if you like so long as you are not writing stuff about octopi.
INTERLOCKING NETWORKS THEORY OF CONSIOUSNESS
Particular problems create the potential for highly complex networks to deal with them. If the solution to the problem has a high enough payoff then probably a network will grow around dealing with it.
Many of these networks create, almost by the spread of their functions a nascent, silent 'I', of varying depth, continuity and complexity.
Over time these networks interpenetrate, interacting with and pushing against each other and as a result of this interaction, this nascent, potential 'I', really more a potential informational space between networks, deepens in capacity.
Like branches from different trees reaching towards each other, ultimately creating a place where many and most branches meet and interact. This spot in particular ultimately, slowly growing an identity itself as the complexity of so many different networks, approaching and intermingling from many different directions, shapes this space, and then finally (or just currently) the sheer need to both solve the problems the networks are designed for individually, but the meta-governance need for management of the interactions of the networks and this weight or need being so great, and/or so fruitful if resolved, that an evolutionary system will invest the huge amounts of resources needed to create it.
"I" IS COSTLY
Self awareness at any level is evolutionarily expensive, like a boutique tool or mechanism. It has to be worth the expense. It has to do something very useful and if not, just use instinct or non-self-aware intelligence.
At every juncture, every single increase in the complexity of these nascent 'I's is going to cost proportionally more than the one before, and if we adopt a human centric model (for the sake of utility in explanation, I will go into more detail at the end about what happens beyond this conception), then to "get to the goal" that full house or high level flush you need to advance not just in one suite but in many and each separate advance has to pay off big for the amount of energy/complexity invested in it, which is why few things are self-aware, or not very self-aware.
TERRAIN - FISH HAVE NO "I"
Little terrain in the Sea. Few complex escape/pursuit choices. Everything is down to pure sensory power, escape velocity. Infinite opportunity of movement and almost NO "objects", or "terrain", little to hide behind, go around, climb, get under etc, means few complex choices and little to no benefit in being able to predict the decision pathways of either predators or prey.
The most self-aware creatures in the sea look to be mammals who are basically importing a complexity gained on the surface and returning it to the ocean, apparently making use of the, in particular, advanced social organisation systems they gained on the surface.
But if they had stayed in the sea, would they have anything similar?
Two points, fish schools and octopi.
What is the difference between the awareness of a school of fish, acting sometimes in a very highly complex pattern, and school or whatever of dolphins or Orca, which exhibit les hyper-complexity of form in the general gestalt but we would assume, more inter-communication between 'selves'.
(I suppose a good question here is; "Why don't the creatures of coral reefs have more complex sense of self?)
Secondly Octopi and that whole group often seem pretty smart, what the hell to think about them I don't know. Some who have worked with octopi say the smartness is inconsistent and overrated, strong problem-solving capacity but intermittently employed and often forgotten between events.
(Maybe because the Octopi has no 'self' to remember for? Selfhood as continuity and retention of complex problem solutions over time? The 'I' of memory?)
Please don't comment repeating common octopi stories about them escaping ang getting across the room or whatever unless you are one of the very few people with a non-pop-culture and representative overview of their relative intelligence and capacities. Like if you work in a lab with octopi or have studied octopi then you can comment about the octopi but if you have maybe read a reddit thread about them or seen a documentary once please don't.
FLATLAND
Complex three-dimensional environment; now there are not just directions but terrain, paths, and highly complex decisions to make. Not just how fast do I run or do I swim up or down but, do I go over this rock or around it, up this tree or past it, hide in the grass, moving slowly, or run along the bare ground fast, now, being able to predict, not just how fast something will attack you or flee from you, but the kind of decisions it will make, becomes useful and if the environment is complex enough and bountiful enough then this can become an arms race.
THE EMPTY SPACE BETWEEN PREDATOR AND PREY
If you only have one job then "I" will probably cost too much, if you are only prey and never predate, or if you are only predator but never prey, even though it might help to develop highly complex theories or models of something else’s mind (the thing chasing me will do this" or "the thing I want to chase will act like this), "I" will still cost a lot.
The self is silent, or quiet, as all it needs to do is act or react to that one particular type of situation, the only axis of combination I Can think of is between the system which is modelling the target, that moving your own body around so in the interstice between those two systems is the vague shadow of a protoplasmic "I"
But..
When you are BOTH predator and prey, when at any time, any particular action might give you access to an active, living and reacting prey and/or make you vulnerable to an active, living and reacting predator, THEN - the interstice between these two powerful and complex systems, one designed to provide you opportunity, the other to save you from destruction, both modelling the behaviour of complex organisms that are not you, now the shadow of the "I" between these two systems becomes thick and almost tangible. As if there were two complex networks whose branches crossed and whose reward/punishment systems interrelated, or like two lamps shining on a dull, matte black space, triangulating its shape and form.
The creation of the living concept of "prey" and that of "predator" at the same time, nearly necessitates an almost explicit but still silent construction of "I" even if only by absence, or the locus and continuation of the reward/punishment systems.
SOCIETY
Theory of mind of theory of mind of theory of mind, and a bottlenecking of a species which can be successful, but only as a group.
Fuck do I really need to go deep into this?
Its immensely complex and I feel like for an educated audience there is little I could say about social organisation being very very important for the success of humans as a species, and that these social organisation issues become more and more and more complex yet can pay of more and more and more fruitfully as the size and complexity of the society they allow grows.
Probably this is the most important one as the 'Social I' has the greatest amount of triangulation between different conceptions of 'I', the most layers of potential trust or distrust, of revealed and unrevealed knowledge.
And the social 'I' strongly advances a highly complex theory of mind which, the better it works and the more complex the society it works in, the more resources you get and problems you avoid.
Once this complex theory of mind starts to exist, as a first element of the process of colonisation/synthesis, it can be turned upon the object doing the analysing, creating the formal abstract concept of self, the last 'I' achieved (so far) but the cornerstone of the colonisation/synthesis process described below.
EMERGANCE AND COLONISATION
At some point, some system or combination of systems becomes so necessary, so complex and so dominant, that it can, or has to, invade or "take over" or at least become fully aware of, a whole bunch of other networks.
Probably this involves the complex theory of mind network, driven by complex and competing social needs and provoked both through language and interaction, turning on 'itself', analysing itself as if it were another person.
(If people don't ever have a spoken language, or words by either sign or symbol, do they have an 'I'? I would say yes but I suspect a very different, by our standards a duller, less-focused version.)
But what exactly is 'itself'?
Well snap, that concept must be created and formalised and there is your spark.
Now comes a vast interrogation, maybe consumption and colonisation, maybe just awareness, of all those other nascent networks; the basic bodily continuation system that even fish have, the advanced tactical systems for navigating complex ground based environments, the long-term environmental prediction stuff, the predator/prey tactical and strategic theory of mind system, even the basic bodily awareness/positioning system, all are consumed, or assumed, by this core collection of advanced theory of mind complex social systems which themselves centralise and "grow" in power as they comprehend and absorb more systems
NOT THAT SELF-AWARE
As humans we know that though we strongly intuit we are self-aware we have also learned that we are not quite as self-aware as we think.
We can fiddle with our pain feedback system to an extent, sometimes, for surface pain, but in all cases deep body pain seems to totally go above and around our awareness. Some rapid decision paths go in advance of our consciousness. Decisions and reactions can take place before and after we think they do or did, our self-interrogation for complex decisions and events - we know is fudgy and often our left hemisphere, , well basically it will tell us whatever answer comes to hand, and if it doesn't really know it will still give us whatever answer comes to hand. We can meditate to become super duper self-aware and highly lucid about our intuitions and response, but if we do this a LOT, like a LOT A LOT, then in many cases we seem to lose actual useful life abilities like making difficult decisions in complex environments very fast making sometimes necessary aggressive actions to defend out lives, those of others or vital principals, becoming very empathic but almost passionless, so our empathy has little practical effect and no longer investigating or working in the world, and so no longer discovering unknown physical things.
Point being; we are probably self-aware and probably the most self-aware thing on the planet. Though its true to say we are never as self-aware as we think we are. Curious that self-awareness has trouble with its own limits.
Being self-aware is probably good, useful and necessary. For the sheer amount of energy our bodies dump into our brains and that our culture dumps into self-awareness.. if it is not useful, or isn't really happening, then life on this planet is a kind of crazy joke. Which if you are Peter Watts or like a mid 2010s atheist sure it could be so, but the idea of a species becoming hyper-successful through the (INCREADIBLY EXPENSIVE) delusion of self-awareness is way more stupid than the idea of a species becoming hyper successful via self-awareness.
Alien Commentator; "I, or the epiphenomenon labelled 'I' see that the humas have dumped a super Quintilian Jules into the 'self-awareness' project, how do you think that is going for them?"
Other Alien - "Actually they haven't but they think they have and its actually working great."
FEW POINTS;
WAKING UP MAY BE WAKING UP
A bit like the development of foetuses from various species mimics or leaves a strange combined map of the unified origins of life, very likely the process of waking up each morning, is, in abstract, and highly compressed in time, probably a general sketch map of the process of humanity becoming self-aware.
If we could read in detail the way in which movement, awareness, memory, self-hood and self-awareness come "on-line" or mutually integrate when a human being wakes up and goes about their day we would probably have an 'embryo' of consciousness if you will.
MAN IS INEXTRICABLY LINKED TO EMBODIED AND SOCIAL LIFE
Self-awareness as we understand it probably isn't like a programme or a line or code or a quality that can be imported even into a complex system, a bit like the Ted Chiang story "The Life Cycle of Software Objects", those systems can probably produce something but its probably not going to be self-awareness as we understand it and most likely will probably be Bostromian highly-intelligent non-self-aware systems like super chat-bots of the kind encountered in Peter Watts Blindsight.
Of course its possible they might become self-aware in ways we don't understand and can't perceive, but, that being the case, what could it mean to us? and what difference would there be to us between a really-strange AI self aware in a way we don't comprehend and a superinteligent AI that isn't self-aware?
Making something we understand as self-aware, enough like us that we can communicate with it, but different enough for that to be interesting, will likely be more like that Ted Chaing story, or even more like Commander Data, or even Johnny 5, than we probably assume from super-brain sci-fi movies, more human, comfortable and social.
Interacting with a complex society in a complex 3d ground based environment with something like human social desires and - since we are the only things we know of that can do that or provide that environment, it will likely happen in "real time" and either complex virtual worlds or simply in real space.
If we want to create a complex self-aware synthetic intelligence then it is probably going to want to care about going to parties, getting a job, having friends etc. It may need to ultimately fear poverty and social estrangement for these are powerful and subtle reward/punishment systems in our own minds which help to create the triangulated shadow of "iI"
THE PROCESS IS SLOW AND ONGOING
This means that self-awareness, rather than being the super-explody bit of a hyper-brain sci fi film is something integrated into our society and into the complexity of our society, and has been developing, slowly, invisibly slowly, and is developing and will continue to develop with our mutual society and all the complex things it can do, and ways it can be.
It’s not really a deck of cards with humanity having the trump hand, but perhaps an, if not infinite, than perhaps ever-expanding deck of cards where you get the royal flush but then they have invented or evolved a new card above the king so new hands are possible at every level of the game.
Strange new forms of self-awareness.
very interesting views, I'd like to re-read it when i have more relaxed time. For now I'd just like to point an observation: fish life or even cell life seems only simple to us because we compare it with ours. While we can have more rationality than fish in a linear spectre that measures rationality, they have an amount of rationality or self-awareness. It might be a residual or just a potential amount, but yet there is. All in nature is made with the same pattern. But they have the right amount of it: a fish who developed more self-awareness or "elevated" interests would probably find no practical way to express or channel those interests, but on the other hand they might take him away from more immediate pursuits or even cause them a crippling depression and lose the will to live.
ReplyDeleteHom many times a fish or an anemona dies because they wanted to pursuit thought cycles that were not the most optimal to their survival is something that we cannot possibly know.
This is very interesting stuff. I’d like to note that this is a theory about the contents of consciousness and the or evolutionary origin, not, it seems to me, an explanatory theory about the existence of phenomenal consciousness, ie the “hard problem” of consciousness. Which is to say: there’s nothing about this theory that would be inconsistent with a universe in which phenomenal consciousness didn’t exist, and in which humans operated exactly as they do, with neural mechanisms presenting these sort of “self-awareness” feedback loops, but wherein there is nothing that it is like to be them.
ReplyDeleteI don’t say this to detract from the theory at all, or to be a pedant — the questions surrounding the contents of consciousness are by no means “easy problems” in the literal sense. I’m just dismayed by how often readers seem to mistake theories about the origins or contents of these self-awareness feedback systems for an explanation of why phenomenal consciousness exists.
Could you describe the difference between "phenomenal consciousness" and what I talk about above, in a way I can understand? No shade I just googled it and got what seemed like an insane word salad I did not comprehend.
DeleteFrom my understanding (and I'm actually doubtless that Picador has a much more solid understanding than I do), this point is essentially that eternally sticky question of "okay so if this is how consciousness arises how do we ACTCHULLY know it's a thing" because all of these processes could be present in a system that is only performing the steps of consciousness without any understanding of what it is like to be performing those steps.
DeleteKinda like the Chinese Room deal; if I've got a book that I don't understand, but which allows me to match correct outputs (that I don't understand) to inputs (that I don't understand), then to the outside observer providing those inputs, I'm understanding and communicating with them, but in reality, I don't know what's going on at all.
I didn't track with the fish part. I don't buy that an aquatic environment = "few complex escape/pursuit choices" because the seas have terrain too. There is the sea floor obviously, as ell as the shifting terrain of currents/bands of different temperature water flowing in different directions, shafts of light from the surface creating a volumetric terrain of shadows that can be moved through for cover...
ReplyDeleteIf the only "meta" for survival in the sea was "pure sensory power, escape velocity" as you say, then we wouldn't see so many species which rely on a strategy other than "swim directly away from my pursuer through open water". A lot of sea life relies on very terrain-based survival strategies, like living inside anemone which the prey species is immune to, hanging out near holes in coral or rock which they can fit through but the pursuer cannot, burying themself in the ground, releasing a cloud of ink or bubbles to break line of sight and then hiding...
What I'm saying is that nothing about the surface says to me "you must necessarily be more complex/self aware to survive here vs the ocean". If anything the ocean seems more demanding since there are a bunch of extra variables to account for and adapt strategies around (being able to survive higher pressures than potential pursuers so you can escape to depths they cannot for example). The scope of sea life is vast as I'd expect from such an intricate and demanding environment, i.e., not all sea life looks the same or shares a similar body plan.
The "pure sensory power, escape velocity" dominance seems a better fit for birds to be honest. Birds and flying insects are pretty much the only creatures (besides bats!) spending significant amounts of their lives in a "terrain-less" aerial environment, and predictable we see only a handful of body plans among them since everyone picks from the small pool of effective strategies, which usually boil down to "see good and fly fast".
I mean a fair amount of that is from coral reefs which I did bring up a little. The spray ink thing seems like its based totally on the creatures body plan and wouldn't involve enhanced planning etc.
DeleteI am cheating somewhat but can you think of any complex pre/predator matrices which are *not* based around coral reefs?
The issue with that last concept is that there are multiple clades of birds (corvids and psittacines) with highly developed problem-solving abilities and social structures. Some species of crows have even been observed using simple tools.
DeletePossibly irrelevant to the topic, but it's interesting to note that almost all the cephalopod lineages that survived past the K-T boundary were the ones that weren't both sheltered and burdened by large shells. Faster-moving, more able to swim down near the seafloor. So, that problem-solving ability might have arisen in response to not being able to simply hide inside a shell when problems arose.
ReplyDelete