Theory of Consciousness (ToC): A More Scientific Approach

The theory of the mind is also known as ToC or ToM. The term mind is frequently used in philosophy of psychology, whereas in the scientific field, consciousness seems to be the preferred term—perhaps because mind appears unscientific. In this essay, I will use the term ToC. As mentioned in previous essays, the current mainstream perspective in the scientific community is a biological (especially neuroscientific) and physical view. Given the subject matter, it appears that research is more active in the biological domain than in physics. Here, I will explain the various theories that exist (selecting only those that align with physical phenomena among neuroscientific theories). To date, the most prominent theories are the Integrated Information Theory (which sees the workings of human consciousness akin to an input-output maximized optimization process) and the Computational Theory of Mind. In early research on cognitive science, studies on human consciousness focused primarily on the brain and the states of neural networks that constitute it (Neural Correlates of Consciousness). This was unsurprising, given that, even today, we assume that human intellectual activity occurs in the brain and it indeed seems to be the case. Although we cannot pinpoint the exact location or mechanism that triggers the system of intelligence, we can at least be certain that it is closely related to the brain. Although NCC is criticized in philosophical circles, it is treated almost as fact within the scientific community, making it a widely supported assumption and thus empowering epiphenomenalism. There have also been studies that focused on specific areas, such as the prefrontal and parietal lobes, observing brain activity or examining the spatial connection patterns of neural networks algebraically as if the components of the brain were individual nodes. The most impactful outcomes of CTM were probably Bayesian & Kolmogorov algorithms and artificial neural networks (NN), as they offered a wealth of practical applications. Moreover, the strong internal consistency of syntactic mapping, preservation of logical compounds, and relativization of quantifiers in physics made extending this framework relatively easy. Effective Field Theories could also be considered an example of this. Fortunately, a neighbor once posted an article about NN-QFT, which I found interesting at the time. However, the issue with the NCC-based framework mentioned above is that its limitations are far too evident. IIT does not yield outcomes that could be described as consciousness. With the advent of GPT, these limitations have already been exposed… unfortunately. Of course, if the sole purpose is technological development, then there’s no problem; we could simply treat artificial intelligence as a tool. Over time, the technology will become more sophisticated, and it will likely “almost” converge to human intellectual performance. (Just three days ago, DeepMind unveiled AlphaGeometry, capable of solving geometry problems at the level of International Mathematical Olympiad winners.) In summary, IIT-based theories of consciousness serve as an excellent tool for enhancing research efficiency. However, they have failed in the aspect of attempting to artificially implement intelligence. It remains to be seen how meaningful machine learning will be in cognitive science in the future. It might turn out to be a temporary trend, much like the DFT or ab initio automation sections that shone for a while and then faded away. It is true that IIT-based ToC is nearing its limitations. Therefore, we could potentially attempt a new approach that has yet to be tried. Of course, there is no guarantee that attempting a new theory will result in anything superior to existing ones. Cognitive science is just taking its first steps—much like the situation when quantum mechanics emerged 100 years ago. At that time, physicist Wolfgang Pauli made this humorous remark: “Until now, no one has known any hypothesis better. Therefore, those who are not yet familiar with the completed state of science may find it easier to determine the direction of their exploration. Of course, ignorance does not guarantee success.” From the outset, there have been numerous theories that attempt to explain consciousness besides IIT, such as the Quantum Integrated Information collapse model, Global Workspace Theory, Higher-order Theories of Consciousness, Electromagnetic Theory of Consciousness, and the QFT of Brain States. There are many theories, but not all are practically useful. Some theories are unverified, some assumptions are virtually impossible to validate, some are overly abstract and hard to apply, while others are simply personal opinions of scholars (e.g., Orch-OR, Consciousness Field Theory, p-Adic QM (haven’t looked into this yet), Topological Geometrodynamics (what is this?); etc.). The Temporo-spatial Theory of Consciousness is also an interesting theory. But why is Category Theory so frequently used among these countless theories? After doing some research, it seems that to approach consciousness from a physical standpoint, one must define it physically. A model must satisfy the following characteristics of consciousness: 1) self-organization, 2) self-similarity, 3) complementarity (duality), 4) complexity, 5) causality, and 6) non-locality. To achieve this, one must inevitably draw upon knowledge of topology or formal science from physics. Concepts like sheaf, type, and category are excellent for bridging different disciplines, which explains their use here (computational trinitarianism). By the way, this somehow seems to inherit Hilbert’s formalist program—isn’t this strange? The theory attempting to explain consciousness using Grothendieck topology belongs to mathematical formalism, and examples might be the Wolfram Physics Project and Categorical Quantum Mechanics. The former even explains synthetic theory of homotopy type, which complements the limitations of classical logic and existing ZFC axiom-based set theory. I’m writing like I’m collecting stamps in a haphazard fashion here, but is there a way to look at ToC in a more naive manner? In my opinion, if we organize the current approaches of scholars, consciousness could be largely divided into 1) analogy and 2) abstraction. One of the scholars who laid the foundation for this was Douglas Hofstadter. Incidentally, in Surfaces and Essences, Hofstadter mentioned syntactic mapping in physics—something I just mentioned—as an example of analogy (The Ubiquity of Analogies in Physics… although I can’t seem to find any resources on it even by searching. It might be worth checking out. It seems that theories of formalism like this are hotly debated in philosophy of physics. Homotopy Type Theory/CT, used in string theory, appears to be a particular target of such criticism, and if that’s the case, TTC may soon face similar criticism… Online discussions illustrate this viewpoint quite well.)