Can Intelligence Define Itself?
On the Inability of Intelligence to Define Itself Completely
Dear Thinkers,
It's been a while! Five months to be exact. After completing an intense period of travel, work and some unanticipated interruptions, I am happy to be back to writing again.
While life did manage to get in the way of writing, it didn’t prevent me from continuing to think. And I have been thinking about a lot of different things. Primary of those, was the concept of coherence. To be precise, the way that coherence is defined, the role of intelligence, and the process of conceptualisation.
While I wasn’t writing all these months, I was silently working on problems related to coherence and intelligence. I feel now, I may finally have a thing or two to share on this subject.
The goal with my writings on coherence and intelligence is to not only articulate the existing problems, but to also shine light on their roots. I realise now, that in the course of my research on coherence, I have diverged significantly from the existing ideas and concepts defined in the field.
I hope, in all humility, that this article, and the articles that I publish in the following weeks, bring forth a fresh perspective, clarity and insights that prove useful, not just in consciousness studies, but across the fields of neuroscience, ontology, epistemology, and of course, AI. I would like to start with intelligence. Specifically, the problem that arises, when we try to define intelligence. It requires intelligence to define intelligence. Ergo, can intelligence define itself in its entirety? Let’s discuss.
What is Understanding?
Have a look at this picture. What do you see?
What you see is a coffee mug—I hope. Why do you see a coffee mug and not, say, a couch or a lantern? Because you understand the concept of a coffee mug. What does it mean to understand a concept, or for that matter, what does it mean to conceptualise? It’s a peculiar word- understanding. We use it constantly, almost casually, yet rarely pause to examine what it actually implies.
Try asking someone what it means to understand, and you might get a circular reply such as, “To understand is… to understand.” It goes to show how deeply embedded and self-referential the act of understanding is. So what then, is understanding really?
To understand is to cohere. And to cohere is to conceptualise. When you say you understand a coffee mug, you’re really saying that the concept of a coffee mug fits together in your mind. But for that coherence to emerge, it must rest on other coherent ideas. For instance, you understand a coffee mug because you also understand the concepts of coffee, and mug. These micro-coherences assemble into the concept. And when they do, the result is that you see what coheres, i.e., a coffee mug.
From this perspective on understanding and coherence, three important points emerge.
Inherent intelligence within us interacts with inherent intelligibility in the world. The result is conceptualisation.
Conceptualisation gives rise to coherence, which we often name as “understanding.”
Coherence, especially of abstract concepts, is self-referential. We understand, but our understanding cannot fully explain itself from within. In other words, intelligence cannot fully define itself.
You may have some thoughts on the first two points. I have plenty of thoughts on them myself. But let’s keep that for later. In this article, let us focus on the claim made in the third point, i.e., intelligence cannot fully define itself. We shall discuss the other points in successive articles.
The Problem with Perception
The self-referentialism of coherence is deeply tied with the problem of perception itself. Something that we refer to in the field of consciousness studies as qualia. Simply put, we are able to perceive a coffee mug, because we are able to conceptualise it as such. But in order to conceptualise a coffee mug, one has to, first, conceptualise coffee, mug, and the implications of the presence of coffee within a mug. How do we combine different sensory data to conceptualise? That right there, is the very essence of consciousness.
One way to understand qualia in rudimentary terms is to consider consciousness as a smaller version of ourselves, sitting inside our brain, and creating the concepts from all the sensory data that it receives.
However, a logical question arises as to how this little version of ourself was able to conceptualise from the sensory data it received?
The only possible way to answer that would be to say that the little versions of ourselves have little versions of themselves inside their brains running the racket. And so begins an infinite regression, as shown in the illustration above.
This informal fallacy is more popularly known as the ‘Homunculus Argument’, wherein a homunculus (perfectly formed miniature version) of yourself is present in your brain doing the work of conceptualisation. In turn, that homunculus has a homunculus of their own inside their brain doing the work of conceptualising for them, going into an infinite regressive loop, i.e., Original Perceiver -> Homunculus 1 -> Homunculus 2 -> … -> Homunculus n.
The result, then, is that no matter how many iterations you use to explain the qualia, the very definition of the qualia itself remains incomplete.
The Limits of Self Definition
In 1931, Kurt Gödel, a Czech logician and mathematician, shook the very foundations of mathematics with what came to be known as the ‘incompleteness theorems’. In the simplest of terms possible, Gödel managed to bring forth a self-evident albeit shocking discovery that any sufficiently powerful formal system cannot fully capture its own structure or truth.
I will try to demonstrate the basic idea behind these theorems with a non-mathematical example, letting your intuition do the work for you.
Consider the following statement: “This statement cannot be proven”.
Now, tell me if this statement is true or false. There are two possibilities here:
If the statement is provable, then the statement is false.
If the statement is not provable, then the statement is true.
Now, if we assume that the statement is true, then that would imply that the statement is true but also cannot be proven using the statement itself. This is a big deal, because it represents a closed loop system that yields a result, and yet, cannot prove its own validity from within the system.
Gödel’s first incompleteness theorem states that any rich enough formal system (for instance arithmetics), will contain true statements that cannot be proven from within the system. Building on this, Gödel’s second incompleteness theorem states that any internally consistent system cannot demonstrate its own consistency from inside itself.
Notes
We have discussed intelligence, as far as this article is concerned, as an essential tool for achieving coherence. We’ve described coherence as the result of conceptualisation. But coherence, by its very nature, is self-referential—because the intelligence required to achieve it is a closed and internally consistent system. Ergo, intelligence cannot define itself in its completeness.
So what does that imply? A few things.
Intelligence cannot define itself completely because it is an internally consistent closed loop system, contained within a larger system that possesses its own inherent intelligibility.
Coherence is a game of two. You need an intelligent agent within a system that has inherent intelligibility of its own. You need intelligence to achieve coherence. But intelligence, in and of itself, is not enough to conceptualise or ‘create reality’. To say otherwise would collapse into an infinite regress, i.e., intelligence trying endlessly to validate itself.
The second half that is required for coherence, is the inherent intelligibility of the environment in which a certain intelligence functions. In essence, this can be understood as follows: we understand a coffee mug as a coffee mug because of the intelligence we possess as well as the inherent intelligibility of the system in which the coffee mug exists.
This understanding of intelligence, conceptualisation, and coherence, takes a significant step away from Idealism which suggests that reality is entirely created in the mind.
The experience of reality or qualia is a result of coherence that exists between an intelligent agent, functioning within a system that possesses inherent intelligence of its own.
I've never thought about intelligence defining itself completely. I can understand that I recognize the coffee mug with coffee in it in the painting because of my own past experiences of seeing and well, drinking coffee. If I had never seen a cup of coffee, I really don't know what I would think about the painting. I suppose I'd be like a child who hasn't seen so many things and must wonder greatly about it before moving on to the next activity. I feel like I'm in over my head here. Haha. Great post.
Brilliant!