An Essay on Consciousness (I)

What can be known?

What is consciousness? Finding an explanation should begin with the definition of the meaning of the terms "consciousness" and "explanation".

The first part of the definition of consciousness is the indexical demarcation, which means that the term consciousness has to be defined in a way that actually captures the phenomenon that we mean when talking about consciousness in the first place. Otherwise, one can just define something in its likeness, explain that thing, and pretend that consciousness has been explained, while in fact, it hasn’t. What I want to get to in this text is the proper explanation of the indexical meaning of consciousness, which is "the feeling of what it’s like". That is what has to be explained in the end. No more, no less. Another aspect often subsumed under the term consciousness is self-awareness/sentience, which will be addressed as well. However, the puzzling thing to most people is the feeling of what it’s like, the qualitative aspect of experience. How can this arise in a purely mechanical universe? How can this arise without being already there in some form or another? To many people, this seems like getting something from nothing, which should not be allowed in any sound proposal.

My goal here is to really address the feeling of what it’s like. This text is expressing my understanding of consciousness at which I arrived over the past few years through the works of Joscha Bach, Thomas Metzinger, Michael Graziano, Frances Egan, Stephen Wolfram, Nigel Cutland, Immanuel Kant, Marcus Müller, some of my own thoughts, and some self-observation as well as some meditation. I do not claim originality nor correctness.

In order to follow through, one has to be willing to give it a try. The matter is very counterintuitive at first until it looks trivial. And it is hard, in my experience, to see through all of it if one doesn’t want to. This doesn’t mean that this take is the correct one, but it means that if one wants to understand this take well enough in order to judge its potential correctness, one has to follow through with it at first for a longer period than might be the case with other ideas that are to be judged. It should be mentioned that with giving it a try, I do not mean accepting something that one does not understand indefinitely. This would defeat the purpose of the text. However, it means being patient and staying in unfamiliar metaphysical territory for a while until the counterintuitive will become intuitively graspable. I think that this is the main obstacle for most people when it comes to developing an understanding of consciousness.

Consciousness:

Indexical demarcation: The feeling of what it’s like, the qualitative aspect of experience.

Approximate definition: Self-reflexive observation.

Explanation: There are two aspects to this. One: how does something work? Two: How did it come into existence? As in, "how does a watch work and why is there a watch?" To show how something works, one has to show how it can be built from the axioms of a descriptive framework. The ultimate axioms that can be found are the atoms of our epistemology, i.e that which is directly given. Conscious experience itself is directly given but what are the smallest units? If one wants to really explain something, one has to show how (at least in principle) it can be constructed from what is directly given at the lowest level. This is where consciousness becomes a problem in the eyes of many. As I mentioned before, the feeling of what it’s like seems to be 'inconstructable'. I will argue that this is mostly because of a fault in our epistemology (though not exclusively). We have to get the atoms of our epistemology straight before building anything with them.¹ What are those? The term "atoms" (and "build" for that matter) might be a little bit misleading because it is usually associated with a notion of the material/matter. But in the context of epistemology, i.e. the study of what can be known, atom simply means the smallest unit of knowledge.

What can be known?

The question implies that there is something that can know.² Generally, this implies human minds since we are the ones doing philosophy in the first place and, in that, epistemology. A more general concept of something that can know is the notion of an observer.³ This notion applies to the mind. In whatever way the exact definition of mind or observer is framed – in the end, minds seem to (at the very least) fulfill the minimally sufficient conditions for being an observer (all minds are observers, not all observers are minds). Minds or, more precisely, human minds do in fact observe: The "world" around them, their feelings and emotions, or their more abstract thoughts. All of these are observations (including the abstract thoughts and emotions).

In fact, all a given mind has at its disposal are different sets of observations generated by perception, cognition or empirical investigation with various measurement devices such as telescopes and microscopes, or particle accelerators. All of these generators provide a (very large) set of observations.

Intermediate answer – what can be known?: Observations.

What do observations consist of? It can be put in different ways but in the end, all observations consist of phenomena/features – which can be reduced to discernible/perceptible differences, i.e. information.

To some readers, the claim that all phenomena are necessarily composed of discernible differences is not apparent. But I mean this in a very literal way, both conceptually and experientially. Think of it like this: a phenomenon can only consist of what is discernible/perceptible, otherwise it would not be part of the phenomenon. The structure of the phenomenon is defined by what is discernible about it. The phenomenon is exactly the appearance and nothing beyond it. All of this implies finiteness and discreteness since everything that is built from discernibilities has to have some smallest unit (there always has to be something which is discerned from something else).⁴

So all phenomena are discrete, finite, and definite structures.⁵⁶⁷ If they weren’t, they wouldn’t be in a specific way, and there would be nothing about them that would make them this appearance and not another or none.⁸ But we have experiences; all of them are in a specific way, otherwise we couldn’t be puzzled about some of them. The structure of experience consists of the perceptible.

Answer – what can be known?: Information.

This would be my personal Cogito: There are discernible differences. That can’t be refuted.

This epistemic position is usually referred to as epistemological computationalism or functional constructivism. Information, i.e discernible differences, is usually measured in bits. At first glance, this seems insufficient. In our materialist culture, we encounter the concept of information mostly in the context of information about something, where the information is a layer of description of something real like a material object or maybe an event. But this is not so. Information is epistemically more fundamental than matter or energy. Matter and energy can never be observed. All observations consist of discernible differences, i.e. information (Bach, Verdicio 2018).

This goes for all scientific measurement devices – they reveal information – as well as for all our sensory organs – they reveal information – as well as for our inner experience – it consists of information – and for all abstract concepts and theories, which are built from information.⁹

If this still sounds off to you, replace the four occurrences of "information" in the previous sentence with "appearances" or "phenomena" and read it again. Then ask yourself again what the smallest unit of an appearance or phenomenon is. Disclaimer: Whenever I use the terms "appearance" and "phenomenon", I do not mean "illusion", which some readers seem to associate automatically.  Appearances clearly exist. They are what they are. Illusion usually refers to the instance in which something appears as something which it is not. But the appearance itself is never an illusion.¹⁰ The interpretation of the appearance may be one. If I am wearing a VR-headset and perceive a person standing in front of me then this might be an illusion. But the pattern that I interpret as  being a person is still there. It is what it is. The appearance is simply the pattern that remains after all interpretations have been removed.

How do we arrive at notions of matter and energy then? If all that can be known reduces to information, then everything that can be soundly thought of as existing – real/abstractly or in any other way – must be constructible from information. Given a set of appearances/discernible differences an observer may find regularities in the way the discernible differences are correlated. The regularities can then be framed as entities – to be used as predictive models of further regularities in the correlation of the observables.

An example from our everyday experience is movies. Let us suppose an old-school analog video projector and we are watching a movie on a big screen. In this movie there is a billiard scene. If we assume that this scene consists of people hanging out in a bar, making billiard balls hit each other, we can understand what is happening on the big screen. We can even do this in a predictive way: Even though it is a movie, the billiard balls will most likely not get up from the table to order a drink at the bar. "The billiard ball only moves because it got hit by another one", we think to ourselves. In this way, we are watching the movie scene and we unconsciously constructed billiard balls, people, causality, and momentum – a predictive set of made-up modeling entities – to understand the dynamics on the big screen. And it works!¹¹ But at the same time, we obviously know that all these things do not exist on the big screen. There are no billiard balls, no people, and no causality. All these dynamics are byproducts of a video projector going through a series of slides. On the big screen, there is only the light of the projector in varying intensities – a dance of discernible differences.¹²

In a bigger and more extensive way, all of this happens as well when we try to understand the world.¹³ In this case, the big screen is the set of all observations (a finite vector of bits). To understand and predict the dance of discernible differences in this bit vector, we use a set of made-up modeling entities once again. A possible predictive set of constructed entities consists of matter/energy and physical laws. This set of constructed modeling entities is captured in the idea of the physical universe. The physical universe is a possible theory that encodes the observed data (Bach, Verdicchio 2012, S. 17). Among all the other discovered possible theories, it has been so successful that it stirs metaphysical confusion in philosophy, leading to sets of beliefs that assume a metaphysical primacy of the entities making up the physical universe while treating all other discovered possible theories as being theories about the physical universe. The confusion disappears by locating the physical universe and all other discovered possible theories in the same domain – as valid encodings of the observations – and, in that, ascribing epistemological primacy to the observations themselves, i.e., to discernible differences. This makes the notion of an ultimate physical substrate unproductive since it is empirically unobservable and conceptually unnecessary. In that, even if one starts out from a physicalist stance, one will arrive at epistemological computationalism just by excluding superfluous concepts from the physicalist approach (Bach, Verdicchio 2012, S. 17). Note the term superfluous. Epistemological computationalism is not derived by adding crude concepts to existing philosophy and physics in order to make something mystical work but rather by getting rid of crude entities that have no foundation. It is the result of trying to fix the foundation of epistemology. One way to think about this is is that all other metaphysics also have information as a part of them.

All existing theories need at least discernible differences AND some more on top of that. But this "some more" is always just assumed, never given.¹⁴

Disclaimer: While solipsism is a possibility that can’t be ruled out, it is not the hypothesis of this text.¹⁵ The content of the previous pages is not supposed to suggest that the universe doesn’t exist and "it’s all just in our heads". What I mean is that the universe is not directly given nor should it be assumed as an axiom but that it is rather a rational inference that results from investigating the presence and the dynamics of all observations as well as their apparent regularities. More on that later.

Appearances, Sensations – Introspection

Percepts: Whenever I speak of perception or percepts, I am referring to our sensory states of experience. Right now you are in a specific state. There is visual structure, auditory and tactile structure, emotional structure and also abstract structure like intuition, meaning, association etc. The entirety of this is your current perceptual state and specific aspects of that – like your current stress level or your current feeling of pressure towards the ground, your sense of touch – are percepts. Your current perceptual state encompasses every aspect of your experience except your inner monologue in which you reason and think explicitly.

Percepts can, in some sense, be detached from what you think explicitly. With OCD, it might be the case that you know that you’re most likely not in danger, but the percept of being in possible danger is still so strong that it forces your hand regardless.¹⁶

Let us investigate a little further the claim that all experience reduces to information. Which means that what you see, touch, feel or think is made entirely from perceptible differences. Vision: Everything you see is perceptible difference. Look around: There is some dynamically evolving space of relations that vary on different dimensions.¹⁷ Shapes change and evolve in 2 and 3 dimensions (it is the same as a phase-space or a coordinate system in math-class), colors span stationary dimensions (As we manipulate them in photo editing softwares¹⁸), and relations between these different spatial relations evolve on a time dimension.¹⁹ But in the end, all you get is a sequence of states, i.e., collections of perceptible differences.

The notions of people, object permanence/identity are not directly given to you.²⁰ These are your default interpretations of this evolving vector of shape-relations and color-blobs that is your visual experience.²¹ Hearing is the same game. Except the patterns here are changes in pitch and loudness etc. All  these sensations are patterns nevertheless. They all have a distinct structure (that is why mostly you can distinguish between vision and sound). You know that "this is a sound" because it is a different kind of observation from the rest. Touch is similar. Although people often perceive it as more real (perhaps because it requires a more direct interaction-type within our world model and because it is currently harder to recreate it by technological means) But think again. What do you feel when touching something? It is a dynamically evolving pattern consisting of dimensions of differences like pressure and texture. You never feel an object; you only have a sensation of tingling and pressure distributions that make you infer the notion of a material object. Smell and proprioception can be described in the similar framework. Emotions are also geometrical modulations of cognition – combinations of pressure, expansion, tingling, movement that have specific shapes and places in your body-map (which is in itself an interpretation). An emotion is an appearance as well – a pattern of information. Once again, it will be different from all the other patterns (otherwise, you wouldn’t be able to tell whether you’re in a certain emotional state or not, and you couldn’t even talk about it nor experience it). And on top of all that you also have abstract ideas that have certain relationships, which are structures as well.

So when I tell you that everything you perceive is changes in information, I’m not just talking about the visual sensations, auditive patterns or the tingling inside your body-map but also about this multidimensional abstract space in which your language of thought plays out. Which seems very vage and very fuzzy, even to yourself. And if you try to reason about it, you have to combine this multimodal array of functions – of differences that can change – and sometimes turn it into some natural language in which you talk to yourself internally. If you’re really good you can start to explain those patterns you experience and construct to other people as well. We usually marvel at people who are very good at this, like authors or comedians. Because we all feel and experience so much – this language-of-thought space is so intricate and interrelated and it is so hard to convey and share its evolving contents satisfactory with other minds.

Again: information is the basic component of our epistemology. And it therefore limits the possible metaphysics we can use to talk about a universe. There is nothing that we could possibly know that isn’t information. All the other levels of description or construction – like quantum mechanics, particles, molecules or social structures supervene on the notion of information. Those are all different ways to talk about information. And often, if you are deeply rooted within one scientific domain, you forget this and you think the entities that you are dealing with on a daily basis are the real stuff. "It’s all chemistry" or "no, it’s all atoms". But these are just different descriptive layers. You have different ontologies supervening on top of each other. If we want to understand minds or systems that are conscious and make self-models, we have to know what the building blocks of the models are at which they arrive. Some of these models are: The self, free will, the body map, the world model, the idea of the universe or personal identity – all these concepts and modes of experience that we face. If we want to understand systems that arrive at these models – and the human mind seems to be within the class of systems that do – we have a self model, most of us have the intuition or the conception of free will, most of us have the intuition that there is a delineation between our self model and the universe, that there is something out there and something in here – we need to start at the basic atoms of epistemology or else we will get lost in impossibilities. Since epistemological computationalism is basically discrete functionalism, the notion of functionalism and function will be explored in part 2 – from my perspective.²²

Literature and useful References

Bach, J. (2009): Principles of Synthetic Intelligence: Psi: An Architecture of Motivated Cognition. Oxford University Press, Oxford Series on Cognitive Models and Architectures.

Bach, J. und Verdiccio, M. (2012): „What kind of machine is the mind?“, Turing-100, S. 16-19

Chaitin, G. (2006). Epistemology as information theory: from Leibnitz to Omega., Collapse, vol. 1, pp. 27-51.

Hildebrandt, F., Glauer, R. und Moore, R. (2023): Rethinking how children individuate objects: spatial indexicals in early development.

Kant, Immanuel (1787) KdrV

Wolfram, S. (2020): Combinators and the story of computation.https://writings.stephenwolfram.com/2020/12/combinators-and-the-story-of-computation/

Wolfram, S. (2023): Observer Theory: https://writings.stephenwolfram.com/2023/12/observer-theory/

1. In his introduction to the Critique of Pure Reason, Kant addresses the fact that this is often neglected: "But it is a common fate of human reason to finish its building in speculation and as early as possible, and then only to examine whether the foundations for it have been well laid. But then all sorts of embellishments are sought in order to console us with the prowess of the foundations, or even to reject such a late and dangerous examination altogether." (KdrV A 9-16)

2. Note that knowledge (in this case) is not used in the classical way as Justified True Believe or similar constrained concepts tied to specific methodologies. It is used in more general sense here as "what is directly given to any kind of possible observer?"

3. See Wolfram, S. (2023): Observer Theory: https://writings.stephenwolfram.com/2023/12/observer-theory/

4. The infinite is an empty concept since the thing it claims to express can never be observed – by definition – there can never be an observation that starts and then at some point becomes infinite. Empirically, there is nothing that can allow us to assume the existence of an infinity. Abstractly, there is nothing that allows us this either. And if we assume it nevertheless, we get contradictions everywhere. All the antinomies like the liars paradox, Russel’s antinomy, Zeno’s paradox, or Gödel’s incompleteness theorems are different examples in which the notion of the infinite (in its various forms: continuum, infinity and statelessness) gives us contradictions and therefore renders the descriptive systems that allow for them fundamentally meaningless.

There have been attempts to fix these antinomies by thinkers like Priest or Russel, but these attempts are like epicycle theories which try to constrain movements in a faulty logical space so that antinomies can’t be constructed anymore. But these antinomies are not interesting mechanisms of powerful abstract systems that just have to be addressed adequately but rather emergent artifacts of an impossible substrate – that  is the "eternal moment" (the term is supposed to already show the contradictory nature of it) of classical mathematics.

This faulty foundation can "do" astonishing things. In set theory, one can add a property phi to all elements of a set with infinite cardinality in one step (here set theory hides an infinite amount of operations in a single step, which makes the step equivalent to the eternal moment of mathematics since, by this logic, everything could be hidden in it). Pi is – in the eternal moment of mathematics – a value that might not have been computed by us and can’t ever be computed entirely by its own definition but nevertheless simply "exists". Real numbers are a regular abstract substrate with an existing infinite resolution, and self-referencing sentences realize the self-reference instantaneously and by that bring a straight forward contradiction into the eternal moment of classical mathematics. All of this makes no sense.

If we only assume what can be observed/constructed/proven, if we only assume what has actually been done, computed, or thought in mathematics and philosophy and all other domains (never has anyone ever actually dealt with the infinite. At most, finite and discrete symbols were manipulated as referents to untenable concepts) then we realize that impossibilities are just what they are: impossible. And finiteness and discreteness is the precondition for anything that can make any sense. So whatever we claim the universe or consciousness is build from, it has to be expressible in a language that doesn’t contradict itself. It has to be a constructive language.

5. If one doesn’t understand this point introspectively or understands it but doesn’t agree, it can also be looked at on another – and to most more familiar – physical level of description: If the brain is some system that creates non-discrete, non-finite thought (whatever that is supposed to be) or experience, it would mean that it is some system that is able to create infinite amounts of experience in a finite amount of time, which consumes infinite amounts of energy. This would pose big challenges to a lot of theories and lead into conceptual contradictions.

6. They might rapidly change and evolve constantly, but at any given point they will be in a specific way.

7. As a side note: one thing I still struggle to fully understand is the fact that some experiences seem so damn continuous. And even if they aren’t really and just seem like it, the bare appearance of continuity would still need some kind of continuity. I somehow don’t believe that there really is the appearance of continuity. Maybe there is just a fuzzy, discrete representation of experiencing continuity or something like that, but this is definitely a point in my current understanding that I can’t back up to the core.

8. Something that is not in any way makes no sense. It can’t exist, and it can surely never be experienced nor talked/thought about.

9. A common intellectual reflex is to regard this as being the result of some present-day-technology bias. Assigning epistemic primacy to information in the information age of the 21th century seems to invite this kind of criticism. In addition to that the notion of computation will play a crucial role in understanding consciousness (and anything really). Is this simply a bias of our time? After all, in the the late 18th and early 19th century, after the discovery of electromagnetism, natural philosophers where suddenly explaining virtually anything with electricity (see Lamarcks’ chain of being as an example). Similar attempts happened in the steam age, where thinkers of all kinds looked at complex systems in terms of wheels and levers, trying to explain the world and everything in a materialist-mechanical fashion. These types of discovery, however, where always referring to specific technologies or objects or entities of specific theories like electric current (twitching frog legs are extremely convincing). Computation and information as concepts are also linked to specific material structures such as personal computers, transistors and digital media but the concepts themselves are in fact abstract. Computation was simply the discovery of a class of universal systems of description and construction (see Wolfram 2021) and an epistemic commitment to information is in some sense a Neo-Pythagorean stance (Chaitin 2006). Information and computation are just the latest terms for concepts of description and expression that are the result of the previous millennia of science and philosophy. In that, computation and information are very different from steam engines or electricity and do not lend themselves easily to the accusation of a present-day-thinking bias.

10. Nur der Schein trügt nicht – only appearances are not deceiving.

11. Sometimes it doesn’t work. If there is a malfunction on one of the slides or in the video projector, we will get artifacts on the big screen that are not explainable by the dynamics of our made-up entities, and we have to take a step back and look at the video projector itself to understand them.

12. Discernible differences are usually described by mathematical functions. Functions are a mode of description. They describe the abstract shape of change. Math deals with all possible shapes of abstract change. Integers, for example, are an ongoing successor function, which can be given arbitrary shapes by using a couple of other functions which we call logical operators. Integers are, in some sense, an abstract synthesizer with number-towers as the substrate and the logical operators as the interface.

13. Movies (simulacra) and computer games (simulations) are controlled pattern generators – they produce discernible differences which can be predicted with the help of made-up modeling entities. From our perspective the universe is a controlled pattern generator as well.

14. That something „works so well“ is not an epistemic or ontological argument. I can easily pretend that traffic flow is liquid and then use fluid dynamics to successfully predict and direct traffic but I would never then claim, that traffic is ontologically a fluid.

15. In fact, I do think that solipsism is ruled out in the sense that the amount of data that would be needed to encode the base level of a universe which produces my conscious experience is less than encoding all my experience as it is.

16. Karl Friston once hypothesized that people with OCD have a far but counterfactually shallow horizon. I think that OCD doesn’t go along with a shallow counterfactual horizon but might in some parts be the consequence of one that is too broad and presented in a very high resolution (and in some cases highly superstitious) so that it cannot be intuitively or analytically collapsed by evolutionary priors or our limited abilities for abstract reasoning and thus has to be forcefully collapsed by the limbic system in order to keep the organism going in some way. If the counterfactual horizon is gone, so is the OCD.

17. For Kant, the space and the time dimensions are the relation-spaces that the mind uses to project perceptible differences (phenomena in his lingo) into. So the perceptible differences are already ordered in a way in which they appear as dimensions, which then are shaped in specific ways depending on the type of appearance that is to be constructed.

18. Hue, brightness, contrast. If there are just shapes and no colors, how do you add additional discernible difference? You need an additional dimension of representation. In this case you don’t change the place or the shape in the coordinate system but rather the appearance of that place. Why colors (and all other appearances) have that additional quality about them will be addressed further down in this text. But even with this additional quality colors are specific appearances, i.e, sets of discernible difference that are distinctively different form all the other appearance you experience. That’s what makes them colors, and that is why you can single them out experientially in the first place.

19. I currently think that time might be a dimension of relations which models the computational phenomenon of entropy, which any observer, depending on its relevant pockets of reducibility (see Wolfram 2023 Second Law), will encounter when dealing with some observation vector. There also seems to be a fixed relation between the spread of order/control and chaos/noise relative to any observer – a relation we know as light speed which is sometimes confusing to us – why is it exactly this value? – since we use miles per second to express the property order per chaos. I wonder if time, being the representation of some necessarily emergent property of computation, could be the reason why no computing model can be thought without some notion of this property, and maybe that’s why time doesn’t ever really leave the picture. It’s always there even if it’s just "state transitions". It seems like some condition of the possibility of structured thought. Kant describes time as an a priori condition of appearance itself (KdrV A34).

20. See, for example, Hildebrandt, Glauer, Moore, 2023

21. These default interpretations can be fooled. VR-Headsets are an example of this. They achieve the default interpretation of a 3-dimensional space with objects and people in it even though the user is aware of the fact that there is a 2-dimensional screen in front of their eyes on which there is nothing but blinking pixels. But the statistics of the pixel blinking are constructed in a way that makes our mind infer a space with objects and people. And then you take the headset off, look around, and say "Technology fooled me but this is the real space now!". No, the space you see is a possible representation of the statistical dynamics of your current sensory input. Nothing beyond that.

22. It seems to me that most definitions of functionalism (Stanford encyclopedia, Britannica, etc.) focus on the historical notion of how the term functionalism has been understood and used by well known academics. These stances seem to refer to a "function" in a more coarse-grained, everyday kind of sense (i.e., "what is the function of  'something '?" while there is always something "real" behind the function that "has" a function). I will work with the more general, formal notion of a function as a relation/mapping between states/values.