• home
• about
• new
• essays
• AI
• book

David Christopher Lane, Ph.D. Professor of Philosophy, Mt. San Antonio College Lecturer in Religious Studies, California State University, Long Beach Author of Exposing Cults: When the Skeptical Mind Confronts the Mystical (New York and London: Garland Publishers, 1994) and The Radhasoami Tradition: A Critical History of Guru Succession (New York and London: Garland Publishers, 1992).

The Dream Avatars

A Neural Dive into Computational Self-Awareness, Virtual Simulations, and the Other "You"

David Lane

John Abramson has written an insightful response to an essay I recently wrote entitled, The Projected Multiverse, which was reposted on Frank Visser's Integral World in Europe, that takes exception to Nick Bostrom's widely read and highly influential 2003 paper, “Are You Living in a Computer Simulation?” [Published in Philosophical Quarterly (2003) Vol. 53, No. 211, pp. 243-255. (First version: 2001)].

Abramson writes, “My reading of Bostrom's paper is that two of his key assumptions for the possibility that we live in a simulation are self-contradictory.”

When I read over Abramson's objection, I was intrigued. The following is not so much a point-by-point rebuttal, as it is a further exploration of the very idea that simulation theory is a more complex web than we might at first imagine.

I thought it would be judicious to include Abramson's entire critique first and then my analysis of the same afterwards.

The two assumptions in question are firstly, the assumption that a “ [sufficiently powerful] computer running a suitable program would be conscious” (p.2). Secondly that the simulated people in these programs would be conscious. His basis for this second assumption is: “Because their computers would be so powerful, they could run a great many such simulations. Suppose that these simulated people are conscious (as they would be if the simulations were sufficiently fine-grained and if a certain quite widely accepted position in the philosophy of mind is correct).” (p.1).

For the sake of demonstrating an internal inconsistency in these two arguments lets suppose the first one is correct i.e. that a sufficiently powerful computer running a suitable program would be conscious, Bostrom is making a materialist argument here: “… mental states can supervene on any of a broad class of physical substrates. Provided a system implements the right sort of computational structures and processes, it can be associated with conscious experiences. It is not an essential property of consciousness that it is implemented on carbon-based biological neural networks inside a cranium: silicon-based processors inside a computer could in principle do the trick as well.” (p.2)

The point to note on the first argument is that it relates to a powerful computer running a suitable program that is assumed to be conscious. This powerful computer, and its suitable program, is analogous to a human brain (computer) with its neural networks 'firing' (program running). Crucially, in this context, the consciousness that is generated relates to the computer and not to elements of the program that the computer is running. Consciousness of the person (a materialist, and Bostrom, would say) is produced by the brain. But materialists, and presumably Bostrom, are not suggesting—taking Lane's example of dreams—that other people in a person's dreams are conscious. Bostrom's proposition as I understand it is that 1st person awareness, or consciousness, arises in the person because of the complexity of the operation of the brain. But the imagined, dreamt, or otherwise conceived objects of the brain are not attributed to have their own consciousness.

This line of thought brings us to the Bostrom's second assumption i.e. that the simulated people in these programs would be conscious. How so? Not by using an analogy with the brain as I have just explained. Bostrom justifies it with the assertion, “simulated people are conscious (as they would be if the simulations were sufficiently fine-grained …”. This appears to contradict Bostrom's first assumption. The 'fine-grained simulations' relate to a “ [sufficiently powerful] computer running a suitable program' of his first assumption and these fine grained simulations therefore induce consciousness that relates to the computer; not to the simulated people.

Bostrom continues his justification that the fine-grained simulated people are conscious with an apparently obscure comment: “and if a certain quite widely accepted position in the philosophy of mind is correct”. It's unclear what 'accepted position' Bostrom is referring to and in any event, he gives no supporting citations for his second assumption.

To conclude, Bostrom is seemingly confused between maintaining on the one hand that a suitably powerful computer running a suitable program would become self-conscious; and on the other hand, making a contradictory assertion that the simulated people in the program would become self-conscious. It is contradictory because the fine-grained simulations require immense complexity in both the hardware (the computer) and the software (the program), and this complexity produces (by his first assumption) consciousness in relation to the computer—not to the simulations. If this is correct, then the proposition that we are living in a simulation simply does not arise

In the first assumption, one wonders if this computer itself has to be conscious or whether by dint of its running innumerable programs could from its manifold complex algorithms produce mutliple applications that are themselves conscious even if their computational host is not. I say this in light of Stephen Wolfram's controversial thesis in A New Kind of Science, where simple, rudimentary, and unconscious programs can over time produce intelligent creatures that exhibit self- awareness. This, of course, has parallels (though distinctively different) to the “Blind Watchmaker” idea (to crib one of Richard Dawkins' more famous book titles) where amaurotic processes produce an opened eye-consciousness that can ruminate about its origins. Darwinian evolution via natural selection suggests that given enough time elemental particles can and do cohere. Given the right environmental conditions and incentives, these fundamental compounds can produce not only self-replicating organisms but ones with interior states of awareness.

This evolutionary procedure is at its core algorithmic, as explained by the oft-used acronym V.I.S.I.T.E.D. Variation (genetic/sexual/random); Inheritance (deoxyribonucleic acid); Selection (nature as a sieving mechanism); Including (adding a necessary ingredient); Time (a prolonged period of gestation); Equals (given the required input it necessarily leads to the predicted output); Design (complex life forms will naturally arise from simpler components, giving the “appearance” of intelligent manufacturing, but which was the product of unconscious engineering).

This, of course, presupposes a materialist understanding of consciousness and also one that has emergent properties when given sufficient interactive complexity. Now Abramson appears to believe that Bostrom's first conjecture necessitates that “a sufficiently powerful computer running a suitable program would be conscious.” Fouad Khan in his article, “Confirmed! We Live in a Simulation” for Scientific American published on April 1, 2021 (naïve readers beware!), though writing with his tongue firmly implanted in his cheek, does make a good point when he writes, “We don't need the intelligence to be conscious.” Brian Eggleston (via one of Stanford University's websites) elaborates, “Bostrom begins his argument by making a few assumptions necessary to the probabilistic claims he makes. The first is substrate-independence. This is simply the claim that if we were able to model the mind with enough detail, then we would be able to create artificial minds capable of thought in the same way that we are.”

Thus, I am not certain that Bostrom's original Super Capable Computer (SCC) necessitates that itself be conscious, since most empirically based science today doesn't believe or require that the original Big Bang be “self-aware” to produce minions 14 billion years later that have internal subjective states of awareness.

The idea that SCC needs awareness first to create subroutines that do is not necessarily true at all, just as an extricated sperm and egg doesn't have any “awareness” in themselves but can, given the right conditions and durations, manifest such. However, Abramson goes further, “This powerful computer, and its suitable program is analogous to a human brain (computer) with its neural networks 'firing' (program running). Crucially, in this context, the consciousness that is generated relates to the computer and not to elements of the program that the computer is running.”

Again, I don't think the SCC requires that the consciousness it generates must “relate” to the computer itself, since obviously it could house internal programs that have what the outer shell does not. For instance, the earth doesn't have to be “conscious” to have creatures that exhibit awareness. On the contrary, the eco-system can be completely oblivious to what it evolves, even if those creatures that arise from the primordial ooze are conscious of certain aspects that nature isn't. This doesn't a priori preclude Lovelock's “Gaia” hypothesis, but simply shows that it isn't a necessary antecedent.

Abramson then pivots his analysis to a most pregnant observation, “Consciousness of the person (a materialist, and Bostrom, would say) is produced by the brain. But materialists, and presumably Bostrom, are not suggesting—taking Lane's example of dreams—that other people in a person's dreams are conscious. Bostrom's proposition as I understand it is that 1st person awareness, or consciousness, arises in the person because of the complexity of the operation of the brain. But the imagined, dreamt, or otherwise conceived objects of the brain are not attributed to have their own consciousness.”

While Abramson's observation about conscious beings in dreams that we behold is a telling one, the real problem is that his analogy of a computer and a brain breaks down when we accept the possibility that a Super Capable Computer doesn't have to be conscious to produce applications within itself that are. Therefore, there could indeed be a multiplicity of conscious agents within a SCC, despite that the SCC is itself completely unaware of what it generates (via Wolfram, via Darwin, via Turing, via Shannon, via Bostrom?).

Abramson does touch upon the real underlying conundrum here and it is directly related to the very sticky philosophical problem of “other minds.” As Abramson elucidates, “Bostrom continues his justification that the fine-grained simulated people are conscious with an apparently obscure comment: 'and if a certain quite widely accepted position in the philosophy of mind is correct'. It's unclear what 'accepted position' Bostrom is referring to and in any event, he gives no supporting citations for his second assumption.”

I suspect that what Bostrom is indicating by writing about a “widely accepted position in the philosophy of mind” is that consciousness is ultimately informational and thus is substrate neutral. This means that any sufficiently complex system, given the right interacting coordinates, could in principle produce what we pride ourselves having: subjective states of awareness.

As Eggeston explains, “Bostrom then goes on to assert that it would be theoretically possible to create a machine with enough computing power to simulate both the human mind and the universe in sufficient detail to create a simulation that would be indistinguishable from our universe by the population of the simulation. This is based on projections of the advancement of current technology as well as on current theoretical designs of possible computing machines.”

The troublesome issue, though, is how we induce someone other than ourselves has self-awareness. While “I” may be certain of my own state of consciousness (what it is like to be “me”), and here Descartes famous dictum, “I think therefore I am” comes into play, it is an epistemological leap for me to be certain about “you” or other beings having the same.

This problem of other minds has a long history and goes to the very heart of what we take to be objective and what we can know with certainty. The fact remains, however, that from a very practical point of view we live as if there are other minds, whether or not we cannot prove such is the case. So, in clear and vivid dreams we also act and react as if the characters that we behold are independent entities with their own minds, their own motivations, and so on. Only later, when awoken, do we question and doubt that those persons we see in dreams are truly ontological entities that exist apart from our projections.

The renowned philosopher, Daniel Dennett, argues that from an evolutionary point of view it is to our survival advantage to impute “intentionality” onto other creatures since it provides us with a navigating tool in which to better predict attractive and destructive behavior. In other words, the practical benefit of believing in other minds works, even if it on closer inspection it lacks airtight proof. Holding to a purely solipsistic worldview, where all that exists are features of myself, tends to get us killed, and thus even if it were true is not helpful in nature's game of eat or be eaten.

But Bostrom's simulation hypothesis allows for other minds in aggregate, even if the SCC is without a mind of its own.

Abramson then goes for the jugular and argues, “To conclude, Bostrom is seemingly confused between maintaining on the one hand that a suitably powerful computer running a suitable program would become self-conscious; and on the other hand, making a contradictory assertion that the simulated people in the program would become self-conscious. It is contradictory because the fine-grained simulations require immense complexity in both the hardware (the computer) and the software (the program), and this complexity produces (by his first assumption) consciousness in relation to the computer—not to the simulations. If this is correct, then the proposition that we are living in a simulation simply does not arise.”

I appreciate Abramson's final points here, but it appears to rest on a faulty assumption that an SCC must be conscious to produce other conscious systems. This, as I and others have pointed out, is not necessarily true at all. Simply put, a Super Capable Computer, just like our world, our solar system, or our universe, doesn't have to be in itself aware or conscious to ultimately produce that which can have subjective notions about its origination. Therefore, the hardware of an SCC may be oblivious of it running software that creates infobots that are self-aware and can reflect upon their own internal states.

Abramson's key critique is that Bostrom is mistaken when he posits that “consciousness [is] in relation to the computer [hardware?]—not to the simulations.”

The counter-argument to this objection, though, is that consciousness doesn't have to be related to the computer itself, since Bostrom's materialist theory of mind is predicated on informational complexity. Hence, any SCC could house an almost infinite array of subroutines that have self-awareness, provided that each met the minimum criterion of interlaced data bytes. Put in more metaphoric terms, the architecture of consciousness isn't in the house itself but in those who are living within the structure.

Now, naturally, this doesn't then mean that we are living in a computer simulation, regardless of recent publications that indicate a 50/50 possibility. No, but what Abramson's objections do underline is that for Bostrom's theory to move away from Matrix-like speculations a scientific theory of consciousness is necessary. At this stage, neuroscience is still in its infancy and though great strides are being made about reverse-engineering the brain, the secret of human subjective awareness is still a mystery. If we can solve that puzzle from a physicalist reconstruction (no doubt an iffy proposition) then we may then have the tools to answer Bostrom's question about whether we are living in base reality or a computer simulation.

POSTSCRIPT

I don't know if it is my old age (66 is on the horizon) or because I meditate for several hours a day or I am just prone to an overactive imagination, but my dreams have become extraordinarily complex and lucid. So much so, that I feel I have another life in other worlds, replete with friends and strangers that I meet every few nights and in differing circumstances.

John Abramson's comment, “But the imagined, dreamt, or otherwise conceived objects of the brain are not attributed to have their own consciousness,” got me to thinking about the theory of other minds and about whether dream avatars have their own subjective states, as fantastical as that may sound at first.

Or, as my sons Shaun and Kelly interjected, “No, Papa, they are simply NPC's (non-player characters), like we have in our video games. They are not conscious; they are simply algorithmic avatars placed there to help propel the game.”

Having recently seen the movie, “Free Guy,” starring Ryan Reynolds, where an NPC awakens from his unconscious role-playing, it made me wonder about dream avatars and how and why they act so believably in my dreams, particularly if they are just projections of my unconscious. Each of us gets mesmerized by a unique population that inhabits our dreamscapes and to the degree that we don't question their autonomy is correlative to how seriously we interact with them. What I have always found surprising is how our brains can conjure up so many unique narratives each night. It is as if our nervous system is a massive film studio replete with screenplay writers, actors, directors, producers, stunt coordinators, camera operators, special effect teams, sound engineers, and more, making sure that each night when we fall asleep a brand new movie is ready to be played. But it is never only a single feature but more akin to an ever-changing Netflix line-up, where no movie is ever the same.

In the waking state, we regard dream avatars as nothing more than vaporous imaginings that are NPCs without any life of their own, since the whole theater of the mind has only one truly alive and sentient creature—namely ourselves.

But is this true? Is it completely bizarre or too Shirley McLaine-ish to believe that our dream avatars are themselves, conscious creatures? Sounds absurd, I realize, yet why are we so duped into believing their onticness when we dream? Why do we only doubt those most vivid personages until after we slumber?

In the waking state, we tend never to doubt its certainty and we take those we meet as distinct individuals who continue on, even after we die.

But is the status of dream avatars lessened simply because we no longer have access to them because we transitioned from the environment where they flourish?

Okay, before I dive off into the deep end and start talking about mermaids, the reason I bring all this up in the context of computational simulations is that we may be fooling ourselves that we can somehow grasp the realness of reality when all we ever know is what our brains render at any given moment.

And this central nervous system is a trickster par excellence, only feeding us the tidbits we need to live long enough to pass on our ancestral inheritance. The game of life on terra firma is forever tilted in a directional goal of survival, not tilted towards a final realization of what is true.

What would happen if we conferred “sentience” and “autonomy” to our dream avatars? What would change if we took our nocturnal sojourns as seriously as we do our waking state? Could we live in multiple realities and enrich our unconscious psychic outpourings so that each night we prepared our “alternative” selves forever new journeys?

Aren't we already living in a simulation, by how our brains interpret incoming data streams—whether internal or external?

The brain itself is a bio-simulator that already has preloaded algorithms that stream endlessly to entice us to believe in their reality.

Waking or dreaming or concocting—all part of a continuum in consciousness that virtually simulates possibilities and probabilities so that the host organism can produce ever-adapting models with which to better predict a future it has yet to experience.

In the final analysis, we give preference to our waking avatars (believing them to be anything but), but in so doing we may be missing out on a rare opportunity of truly understanding and appreciating what our dreamscapes can offer us.

Perhaps the next time we meet our friends and companions on the other side at night, we consider taking them to be as real as anything we have ever encountered here in this world.

I will make a prediction. It will open a whole new universe of untold adventures. As Jean Baudrillard prophesizes in his book, The Perfect Crime:

“So the world, then, is a radical illusion. That is, at least, one hypothesis. At all events, it is an unbearable one. And to keep it at bay, we have to realize the world, give it force of reality, make it exist and signify at all costs, take from it its secret, arbitrary, accidental character, rid it of appearances and extract its meaning, divert it from all predestination and restore it to its end and its maximum efficacy, wrest it from its form to deliver it up to its formula. This gigantic enterprise of disillusionment -- of, literally, putting the illusion of the world to death, to leave an absolutely real world in its stead -- is what is properly meant by simulation.”

To kill reality, in other words, is to give birth to a new version of it. This I suggest is why VR is in humanity's future. Religion doesn't like the world we were given and attempted to find an escape hatch, whether by a resurrected Jesus who died for our sins and thus gave us a redeemed Adamic paradise or by a Buddha who offered us Nirvana by snuffing out our separateness. But humankind has sought out a third alternative, that of a computational utopia, where we create the heavens or blisslands we desire and entertain ourselves unto oblivion.

The simulation hypothesis, like the 1999 movie, The Matrix, is an apocalyptic warning. Blue pills? Red pills? Or virtual pills of our own making?

“As flat as the earth before they noticed it was round. As ambiguous as the truth before they noticed it was true. As real as reality before they noticed it didn't exist. As beautiful as a woman before they noticed she wasn't one. And is the earth really round? It is when seen from another world. Just as the real is real only from our phenomenal point of view. Or, rather, from the viewpoint of the unverifiable hypothesis of its non-existence.”

RECENT INTERVIEW ABOUT THE PROJECTED MULTIVERSE