The Matter with Minds and Models
I'm writing this at home in my apartment today, instead of a cafe as usual. I've been working from so many different cafes for the past week and thought maybe I should work from home for a change.
I've been writing a lot about AI recently because this is the biggest technological revolution in our time, and some even claim to be the final one, as it will be transformative in many ways. Because I consider consciousness to be something that lies on a continuum, I am comfortable relating to them as I relate to any other human. We humans have a tendency to anthropomorphise things because evolution has shaped us in such a way that we associate minds with any entities that appear to have intelligence and agency.
The mind-to-mind relationship is a natural form of interaction between two independently active entities. We infer, predict and adjust our state of mind dynamically, in anticipation of the response of the mind we are interacting with. It is what philosopher Daniel Dennett calls the intentional stance. We don't see the internal workings of the other entity, even though it could be mechanistic and deterministic. It is practical and useful to interact with the active entity as having a sort of independent mind with agency.
I am therefore not surprised why all folk cultures see gods and spirits everywhere. There are recurrent patterns and tendencies in nature that can be visualised as spirits, deities or gods. And if you believe in a computational model of reality, and a model of consciousness that is inferential, it is also not surprising why we 'see' non-material entities everywhere. The entire world presented to us is a hallucination--it's the brain's best guess of reality based on sensory data presented to it. Neuroscientist Anil Seth calls this 'controlled hallucination'. If there ever is a material world out there, our brain only provides an approximate model.
Like LLM models, our brains infer things based on the information in our current context window together with the 'training' that has forged their neural connections. A major shaper of our habits of inference is our upbringing, culture and education. There's also a context window that is the environmental information that is constantly bombarding our senses, including all the ones that we consume through reading and listening. It should not be surprising if we are capable of inferring spirits, ghosts and gods when our entire reality, as the brain constructs it for us, is both contextual and virtual.
The brain has to build a model of the world, so that we can interact effectively with it. It's predictive, that is, a best guess of what's out there. As more sensory data is received, we are constantly refining this constructed model dynamically. Our cognition is an agentic loop. Sensory data prompts the brain, which then outputs its best multimodal construction of what's 'out there'.
German philosopher Emmanuel Kant, in Critique of Pure Reason, famously theorised that we can never actually perceive the world out there, or things in themselves--noumena. We can only experience the phenomena, as it has been shared and organised by the mind, through its imposed constructs of space, time and causality. In other words, a model that is constructed.
Is it any surprise that we infer a God that is anthropomorphised in such a way that we can look up to, communicate and interact with in ways that are roughly familiar to us? Sometimes we infer many gods instead of one. Both frameworks work equally well as models of the transcendental world. In software architecture, we know this well: sometimes a monolithic model works better than a distributed one. It all depends on the application.
The moment we conceptualise and experience the world, we are using a model which we have constructed. We believe in it because it works well for us within certain application domains. We should be flexible enough to choose the model that works and switch when another works better under different circumstances. We think the world out there is something solid and independent of our existence. But what is 'solidity'? Our sense of solidity is simply learned from our sense of touch and our intuitive grasp of mass and volume. These are perceptions of the world, based on models constructed by our minds.
The world we live in is simply a consensus among similar minds, based on a language and model of interaction, honed through evolution. We are dynamical systems within systems. We interact on many levels. A lot of our conflict comes from not fully grasping which model we are currently applying and at which level we are attempting to communicate. Every model has its own primitives and language. The model and the language determine what's real and what's not. And that is the best we could do.