Breaking Mind-Matter Symmetries
Exploring some of the Philosophical Landscape around AI Development
Intro
Welcome back to the next installment of consciousness inquiry here at CogitoAlgo, where we'll be diving into the mind-body problem, Bernardo Kastrup's insights, and their implications for AI development. I'm delighted you could join me on this journey into a subject that has fascinated me for a long time.
During my undergrad program, I recall taking my first Philosophy course and being struck by how little we know about reality. Around the same time, The Matrix had just been released, and after watching that movie, along with grappling with philosophical themes in school, I never could quite shake this passion for contemplating the nature of reality.
Fast forward to the present day, with the emergence of numerous technologies that were merely science fiction just a generation ago. It's essential to remember that we are witnessing history in the making. As Max Tegmark recently stated on the Lex Fridman podcast1, “this is the most important fork humanity has ever reached in its [history]”. It's an awe-inspiring time to be alive, especially for those working in fields directly impacted by discoveries in the world of AI, but also for the rest of us. What's being uncovered now and what is yet to come in the very near future has far-reaching implications for our civilization, making it difficult to compare this to any other time.
In other words, we are currently witnessing astonishing developments that would have amazed and unsettled our ancestors. The fact that you are alive and observing this now should evoke a mix of awe and unease, but it's also something we can't afford to miss. For better or worse, there are crucial insights here for humanity to comprehend about itself. As we speak, long-held assumptions and theories about the nature of the world and the role of conscious humans within it are being both confirmed and spectacularly dismantled.
Since this publication focuses on the nature of consciousness, particularly as it relates to advancements in AI, some of these mysteries are on the verge of significant breakthroughs, while others are brewing in the background, with imminent discoveries just around the corner. That's why I'd like to dedicate the next few essays to laying the foundation for what we'll discuss in the future. Eventually, (in the very near future, I believe) we'll be examining current events and their implications for all that we're exploring here. So, my hope is that we can establish some groundwork to ensure later topics are more easily understood. It would be challenging to dive straight into discussions about consciousness, sentience, agency, the singularity, and more without first clarifying a few things.
A seemingly perennial mystery
That being said, one of the most enduring mysteries of the last few centuries is undoubtedly the mind/body problem. As briefly mentioned in the previous two essays, this problem can be defined as follows:
The mind-body problem is a philosophical question about the relationship between our mental experiences (thoughts, emotions, consciousness) and our physical bodies. It asks how the mind, which appears non-material, can interact with and influence the body, which is made of matter. This problem raises questions about the nature of reality, the existence of a soul, and the possibility of free will.2
This problem has existed in some form for a long time, but it was our old friend and muse for this Substack's icon, René Descartes, who probably contributed most to advancing this line of thought in our modern world.
Without getting too far in the weeds, the backdrop for this philosophical issue and many others emerged from what we now call the European Enlightenment. Most of you will likely be familiar with this time period. This cultural and intellectual movement primarily took place during the 17th and 18th centuries. Thinkers like Immanuel Kant, David Hume, and John Locke dominated the landscape at the time3, proposing ideas that generally marked a shift away from traditional thought processes involving the church and authority toward reason, empirical measurement, and scientific advancement using logic and other approaches. Much of this intellectual explosion aimed to break the stranglehold that a collusion between the nation-state and the Roman Catholic Church had established over nearly all areas of life and thought during the preceding centuries.
Part of the goal of breaking this grip on the mind and its faculties through scientific methods of inquiry was to encourage individuals to choose their own path by making sound measurements, forming arguments, and forging paths that would differ significantly from what the "authorities" at the time would have agreed with. Of course, this is a vast topic and would be more relevant on another Substack altogether, but here I'm merely mentioning this as the cultural backdrop for what we are about to discuss.
Enlightenment thinkers and “the Turn”
So, we had René Descartes in the mix back then, writing his "Meditations on First Philosophy," published in 1641.4 In this work, he aimed to explore the depths of reality and identify some foundational principles upon which he could base the rest of his system of thought. He proposed two distinct principles: the "thinking thing" or "res cogitans" and the "extended thing" or "res extensa."5 These ideas were very much in vogue back then, and many philosophers discussed those things that were "extended" as having some physical presence, in contrast to purely mental contents.
Later on, Immanuel Kant would come along and discuss the nature of reality too, and he would contemplate the concept of extension, as well as its relation to space and time. Kant's theories built upon Descartes' and others, and his work "Critique of Pure Reason"6 had such a profound impact on the philosophical world that we still refer to Kant's "Copernican Revolution"7 or the Kantian Turn. This was the moment in time when we in the West began to seriously question whether the objective world was indeed out there in reality or if much of it was somehow in the mind.
Alright, that's enough background for this present essay. My goal here is merely to establish the mind/body problem sufficiently to introduce a paper by one of my favorite contemporary philosophers, Bernardo Kastrup. If you've read any of my previous essays, you'll have already seen his name. And I'll likely mention him many more times in the future. His work in the field of ontology is, in my opinion (and lots of others much smarter than me), extremely important and particularly relevant for our current focus on consciousness.
Bernardo Kastrup and his work
Kastrup has numerous academic papers published in various journals, but he has thankfully compiled many of them into an excellent volume titled The Idea of the World.8 It's a fantastic read if you ever have the time. Of course, it requires some deep thought, but it's well worth it. My initial background is in computer science, but I also have a graduate degree (which is just enough for me to realize what I don't know) in a field where it's common to encounter issues related to consciousness. As such, Kastrup's arguments for an updated ontology of idealism are, in my view, important for our present world, especially concerning new artificial minds that may be capable of sentience.
In the volume mentioned above, Kastrup presents around ten of his academic papers related to his analytic idealism9, an ontology that essentially posits that all of reality exists within consciousness. There are some very fine distinctions here that set his work apart from others who propose ideas like cosmopsychism and panpsychism, among others, as mentioned in the previous essay.
For our current discussion, I'd like to briefly review Kastrup's first paper in the volume, which presents his project in a quite incisive manner. What I admire about Bernardo is that he's almost like an intellectual ninja. He boldly confronts some of the most significant problems that have plagued philosophy for centuries, directly targeting the weak spots in these arguments like a martial arts master. It's truly quite spectacular, in my opinion, but you'll have to read his work yourself when you get the chance.
The paper we’ll be examining
The paper we'll be discussing today is titled: Conflating abstraction with empirical observation: The false mind-matter dichotomy10. Now, before we delve into this, I'd like to say a few things just to set the mood. First, these are challenging subjects to grasp. So, if you find yourself re-reading a few times, there's nothing wrong with that! Second, I'm not a (professional) philosopher myself. As mentioned earlier, my background is in another field, so there are undoubtedly things that I might get wrong here. If you're more knowledgeable about these subjects and I've misunderstood something here, please let me know in the comments.
Alright, enough of the preamble; let's dive into the core of our discussion. As mentioned above, the mind/body duality problem has been around for a long time and is quite pervasive. Kastrup begins by defining the issue, much like we have above, but in greater detail. He mentions the issue originated with Descartes and Kant, but there have been attempts to soften this duality, so to speak, with more nuanced versions of these ideas. Some efforts have been made by thinkers such as Alfred North Whitehead, Henri Bergson, and, to name a more contemporary philosopher, David Chalmers.
Most of these attempts, however, aim to make the mind/matter split less tangible and more of a "property" issue. This is called "property dualism" as opposed to the mostly historically discredited "substance dualism"11 à la Descartes and any that followed with his line of reasoning. But Kastrup believes there's a better way through this "problem" by dismantling it at the root. And that root, he contends, lies with what he calls an "epistemic symmetry" that should not have been there from the beginning.
Epistemic boundaries
You might have heard of the term "epistemic." It's essentially a way of speaking about our ability to acquire knowledge. How do we know the things that we know, and where does this knowledge come from? There's a whole field dedicated to this called "Epistemology,"12 and within our context today, Kastrup is using this term to indicate that a false correlation has crept through society and into nearly all the crevices of our thought processes. Even and especially in places like our philosophical discussion, where we should have been much more rigorous in defining our terms in the first place.
At the outset of the paper, Kastrup defines what he means by the term "mind." It would be in our best interest to adhere to this definition as well. He notes the following:
I use the word "mind" in the sense of phenomenal consciousness. Following Nagel's original definition of the latter (1974) — which has since been further popularized by Chalmers (1996, 2003) — I stipulate that, if there is anything it is like to be a certain entity, then the entity is minded. (Kastrup 2019, p. 23)
Alright, so that's what he means by consciousness in this paper. It's not necessarily some notion of metacognition13 or being aware of what is known in the mind. His notion here is focused on "phenomenality,"14 a branch of the philosophy of mind that deals with perception. This is a common definition of consciousness, though there are others, but it's well understood that for something to be conscious, there must be a "concrete" sense of what it's like to be that thing.
For example, you've likely noticed that when you wake up from a dream, there's something that being awake feels like. It's hard to nail down exactly, but you know it when you feel it. There's just something that it's like to be you. There's also something that it's like to be me. These are probably similar feelings, but we can't directly compare them because it's so unique to each of us who are conscious. Our pets (and of course other animals) likely have these same feelings. There's probably a concrete sense of what it's like for my dog to just be himself.
How physicalism and idealism fit
With that said, Kastrup moves into mentioning both physicalism and idealism. He'll be making references to these ontologies, but he won't be putting forth an argument for the primacy of either one quite yet. This current paper is aimed directly at identifying why the epistemic symmetry between mind and matter should be broken. He goes on to mention the "hard problem of consciousness," coined by Chalmers:
By postulating a material world outside mind and obeying the laws of physics, physicalism can accommodate the patterns and regularities of perceptual experience. But it fails to accommodate experience itself. This is called the 'hard problem of consciousness' and there is now a vast literature on it (e.g., Levine 1983, Rosenberg 2004: 13-30, Strawson et al. 2006 2-30). In a nutshell, the qualities of experience are irreducible to the parameters of material arrangements—whatever the arrangement is—in the sense that it is impossible even in principle to deduce those qualities from these parameters (Chalmers 2003) (Kastrup 2019, p. 24).
We touched on this topic in the last essay, but in other words, we still can't figure out how the feeling of eating a perfectly grilled steak on a patio by the ocean is related to the neurons firing in the brain. We don't know where they are, what they mean, and how they are related to the sensations that make our lives largely worth living.
At this point in the paper, Kastrup explores one way that others have tried to deal with this issue, by proposing a third way between mind and matter—information. While we won't get into this here, there are those who work in a field called pancomputationalism where they believe that the basic facts of the universe and all reality are information and the processing of that information. Kastrup offers a scathing critique for those who follow this path, and we won't discuss that here because it would take us too far off course. Suffice it to say, he's not impressed by this option, and so we won't get into that here, but in a future essay I’ll certainly examine some of these ideas.
Levels of abstraction
According to Bernardo, the whole thing essentially comes down to what he calls "levels of explanatory abstraction." Allow me to elaborate on this a bit further. On this view, Kastrup notes that the things we know for certain fall into the category of that which is directly experienced. For example, if someone asked you about your inner feelings at the moment, you'd probably be able to introspect for a moment and say that you're either happy, sad, melancholy, or some similar feeling. If you're really in touch with yourself, you could probably provide a much more detailed map of what's going on inside you right now.
That inner experience is an example of what Kastrup would call a first-order experience. He mentions that you know the reality of the chair you're sitting on much better than you know the reality of the subatomic particles that are supposedly everywhere. Kastrup notes that as we looked around our world, we decided to first postulate the existence of an external world, and also that this world has certain properties and laws. But the problem occurred when we decided to take those laws and properties, which admittedly work very well in science, and then forget that we essentially invented the existence of these rules and properties to explain our firsthand experience.
A little thought experiment
In other words, it would be like floating in one of those sensory deprivation chambers and assuming that the chamber was your only reality. Anything beyond that couldn't be directly known. Now, at the present moment in the chamber, all you know is the fact that you feel alive. You might have certain inner emotions and various sensations in your body, but beyond that, you don't really know anything at all. But let's suppose that you start to sense certain regular vibrations or sounds coming from somewhere outside yourself. You begin to think that something might exist outside the bounds of your little chamber.
Let's imagine that you're in the chamber for a really long time. Eventually, you might come up with all kinds of theories about what's outside of your present chamber. These ideas might also be really useful. Your theory about what's out there seems to correspond with reality. You've decided that there's some kind of carnival happening (I don't know—just an example), and right on schedule, you hear a certain noise. You figure that it's some kind of large thing moving by. You come up with something that looks like an elephant in your mind. You get so used to imagining these things as the way it really is outside the chamber that you haven't ever actually seen outside the chamber.
This reality begins to overtake everything that you do, and it's very effective. Your ideas seem to get things done, and you can create other theories as well that fit in nicely. Now, imagine that you can communicate (somehow) with others in different chambers. You all eventually agree on some basic theories to explain things that you've all experienced together, albeit in your separate chambers. So you begin to say that the world outside the chamber is the opposite of your inner chamber-world. This idea begins to creep into everything that you and the other chamber-dwellers talk about.
And as time goes by, all the chamber people have forgotten that the outer world is not as epistemically primary as the inner chamber world. Remember, you've never been outside the chamber. Anything that you've come up with to describe a world beyond the chamber is just that, made up. Yeah it might be accurate at some level. It might have enabled you to build lots of things inside the chamber. And others find it helpful to build and explain stuff as well. But these things outside of the chamber are what Kastrup is calling "explanatory abstractions".
More context on abstractions
The term "abstraction" is used in many different fields, but here its usage means roughly that we take complex ideas and simplify them into other, more easily understood ideas. A good example of an abstraction in this sense is the idea of a car's engine. The fact that we can conceive of a car's engine as a single concept is the abstraction. In reality, there are thousands of moving parts, pistons, compression ratios, spark plugs, wires connecting all kinds of subsystems, cooling systems, computer monitoring systems for these systems, and more. An engine as a whole is a marvel of complexity, and to view it this way would make it all but impossible to bring up casually in a conversation.
So, we abstract things into more manageable chunks to make it easier to move them around in our conversations and brainstorming sessions. Kastrup makes this same argument about the so-called symmetry between "mind" and "matter" (or body, which is, of course, matter). He notes that there is actually an explanatory asymmetry to these concepts or the dichotomy of these two ideas. Furthermore, he says that if we postulate some particular dichotomy between two things, those two things must exist within the same level of abstraction.
For example, Kastrup notes that the first level would be the idea of an external world. Remember, we don't know anything truly beyond our own perceptions of anything but our bodies and our subjective experiences. Anything outside of that is pure speculation. So, the external world is an abstraction away from what he calls the "screen of perception," or, in other words, the things that we can directly perceive. The external world is a first-level abstraction, and there are many other levels that could be used beyond this. He gives the example that if we postulate some properties in this external world, this is a second-level abstraction. Beyond that, we could come up with many more levels of abstraction, trying to explain the interaction of properties or other concepts.
Keep in mind (see what I did there) that all we really know, epistemically speaking, are those experiences that we perceive on the screen of perception. Think of your inner life and your outer world as being the boundaries of your body. If you see a measurement of something on a tool like a thermometer, you are still experiencing this on the screen of perception. Your own eyes are seeing the thermometer readout, so it's coming through your eyes and being processed as an experience of that tool. Now, if somebody told you about a thermometer that they saw the readout for, that would be another level of abstraction. It sounds perfectly believable based on what you know about the external world, but you didn't experience this directly yourself.
As such, we've never directly experienced quarks or Higgs bosons, or the spin of an electron, to name a few other abstractions. I don't directly see the lift on the wing of a plane when I fly somewhere, but knowing that there's a low-pressure center above the wing based on the shape and the speed of the airflow helps to understand why the plane is in the sky. Otherwise, it's not something that can be directly known; the idea of lift is a scientific theory.
Kastrup goes in for the take down
Alright, so I think we've probably grasped the idea of unequal, or asymmetrical explanatory abstractions by now. Let's move into how Kastrup uses this notion to tackle the mind/body problem. Here's Kastrup himself:
The notion that idealism and mainstream physicalism are mirror images of each other arises from a failure to grasp this point. Lucid contemplation of these ontologies shows that idealism attempts to reduce an explanatory abstraction (physically objective matter) to that which articulates and hosts the abstraction in the first place (mind). This is prima facie eminently reasonable. Mainstream physicalism, in turn, attempts to reduce mind to mind's own explanatory abstractions, an obvious paradox that constitutes the crux of the 'hard problem' (Kastrup 2019, 35).
Got that? If not, go back and read it a few times. Now for the zinger. This is where Kastrup delivers the decisive blow. Pay attention, folks; this is philosophical ninja stuff:
There would be no 'hard problem' if one did not conflate explanatory abstractions with concrete ontological primitives; if one did not attempt to paradoxically reduce mind to abstractions of mind. The 'hard problem' is not an empirical fact but the salient result of internal contradictions in a logico-conceptual schema; contradictions that I hope to have helped make explicit with the present paper (Kastrup 2019, 35).
I hope you were paying attention. Did you see what he did there? He basically dismantled about 300 years of philosophical error. It's a problem with our "logico-conceptual schema." Well, thank you, Mr. Kastrup; you have indeed made those contradictions quite explicit. I don't know about you, but this is the equivalent of watching a skilled fighter take down a huge behemoth in a few swift moves. It's really quite phenomenal.
Something it’s like to be a conclusion
In conclusion, we've taken a quick look at how we've arrived at a point in time when we need to start rethinking our ontological notions of either some version of physicalism or idealism, and how they might enable us to better understand some of the biggest breakthroughs of our time. In particular, as various organizations working on new AI systems begin to advance further into larger models with increasing complexity and newer architectures, it will be crucial to have a way of gaining an intuition of what's going on beneath the surface. We'll eventually, no doubt, need to ask ourselves if these models are conscious and by what criteria.
In addition, we'll need to have a firm grasp on what we mean when we say consciousness. Are we just talking about metacognition here, or do we mean, like Kastrup suggests above, that there is "something that it's like to be" that system, however that is defined? These ideas and more will certainly be part of mainstream discourse soon, and it's good for us to get a handle on them now so that we can understand the discussion. In this essay, we've also examined how Kastrup uses various means, especially the notion of "asymmetrical explanatory abstraction," to break apart the conflation of "a rational-linguistic construction with reality itself."15
And with that, I'll leave you, fellow readers and thinkers. Until we meet again, I hope this has been as helpful for you as it has been for me. Remember, we're laying some of the groundwork we'll need for later discussions. In the meantime, get out there and enjoy your conscious experience; it might just be what defines you.
https://lexfridman.com/max-tegmark-3/
Summarized by GPT-4
https://plato.stanford.edu/entries/enlightenment/
https://www.sparknotes.com/philosophy/meditations/section2/
https://press.rebus.community/intro-to-phil-of-mind/chapter/substance-dualism-in-descartes-2/
https://www.sparknotes.com/philosophy/kant/section1/
https://plato.stanford.edu/entries/kant/#KanCopRev
https://www.amazon.com/Idea-World-Multi-Disciplinary-Argument-Reality-ebook/dp/B07PGQPV3R
https://www.essentiafoundation.org/analytic-idealism-course/
https://www.academia.edu/37049147/Conflating_Abstraction_with_Empirical_Observation_The_False_Mind_Matter_Dichotomy
https://press.rebus.community/intro-to-phil-of-mind/chapter/property-dualism/
https://plato.stanford.edu/entries/epistemology/
https://ctl.columbia.edu/resources-and-technology/resources/metacognition/
https://plato.stanford.edu/entries/phenomenology/
Kastrup 2019, p. 37