Toward a Unified Science of Mind-Brain-Behavior
The information revolution and cybernetics movement of the mid-20th century significantly influenced academic approaches not only to physics but also to the study of human mind and consciousness, brain and body, and human and animal behavior. During this period, the relationship between information and mathematical formalisms began to draw attention, leading to significant advancements in information theory and statistical mechanics. These movements aimed to formalize the study of human nature and life, but due to technological limitations and philosophical challenges at the time, the latter half of the 20th century saw a fragmentation into numerous fields, including psychology, psychiatry, neuroscience, economics, anthropology, artificial intelligence, and robotics.
Looking at the history of physics, it becomes evident that developing a strong theory capable of explaining, controlling, and predicting new phenomena requires the evolution of principles, mechanics, and dynamics that address the questions of ‘Why’, ‘How’, and ‘What’. Currently, there is confusion among the various dynamics, mechanics, and principles attempting to partially explain mind-brain/body-behavior relationships.
In this context, a significant development is occurring in the form of normative theoretical approaches that seek to integrate and reduce these various academic frameworks. Notable examples include Karl Friston’s Free Energy Principle and Active Inference, as well as Samuel Gershman’s Inductive Bias Principle and Approximation Principle. These theories focus on explaining biological systems, as well as psychological and behavioral phenomena. However, complex phenomena such as consciousness, emotions, and hypnosis remain elusive to modern science. These phenomena encompass biological complexity and evolutionary characteristics, making them difficult to fully explain through simple physical or information-theoretical approaches. Thus, a multidisciplinary and integrative approach is required to understand these aspects.
While these approaches emphasize the search for ‘Why’, modern AI research is more focused on ‘How’—that is, engineering applications. Concepts introduced by researchers like Jurgen Schmidhuber’s Self-supervised Learning and World Models, Geoffrey Hinton’s Boltzmann Machine, and David Rumelhart’s Helmholtz Machine have been crucial in developing AI that mimics human intelligence. Since information theory is not merely a subset of physics but holds a unique status, it may require the introduction of irreducible additional principles, allowing for the formation of a diverse set of principles.
The language of modern AI is fundamentally based on various mathematical formalisms within the foundation of information theory, such as information geometry and optimal transport. Just as E. T. Jaynes reinterpreted statistical mechanics from the perspective of information theory, thereby proposing an integrated relationship between the two fields, new insights can be gained at the intersection of information theory and physics. This integrative approach enables information theory and physical principles to complement each other, playing a crucial role in explaining mind-brain-behavior systems.
In conclusion, understanding mind-brain-behavior systems requires a unified approach to information theory and physical principles. Information-theoretic principles help explain how physical systems process, predict, and regulate information, while physical laws provide a foundational understanding of how these systems actually function. By exploring the intersection of information theory and physics, new research and discoveries can be fostered. This essay provides important insights toward building an integrated scientific system for mind-brain-behavior, laying the groundwork for a better understanding of complex biological systems.
Summary
The information revolution and cybernetics movement of the mid-20th century had a profound impact on the study of human mind and behavior, leading to the development of information theory and statistical mechanics. This period saw a fragmentation into various academic fields. The development of principles, mechanics, and dynamics in physics is necessary to address questions of ‘Why’, ‘How’, and ‘What’. Normative theoretical approaches, such as Karl Friston’s Free Energy Principle and Samuel Gershman’s Inductive Bias Principle, have developed to integrate various fields. However, complex phenomena like consciousness and emotions remain difficult to explain with simple physical approaches, requiring a multidimensional approach. Modern AI research focuses on ‘How’, with significant contributions from scholars like Jurgen Schmidhuber and Geoffrey Hinton. The intersection of information theory and physics offers new insights, playing a crucial role in understanding mind-brain-behavior systems. Integrating information theory with physical principles provides a foundation for understanding complex biological systems.
Comments