Carlos E. Perez writes thoughtful articles on AI with admirable clarity. I share his conviction that AI will be more effective when it employs a self-model. While revisiting his articles at Medium, I encountered the article titled “The False Equivalence of Categorization as Thought” [1] where he criticizes the overreliance on categorization in AI in particular and “substance metaphysics” in general.
Earlier methods of AI, especially the supervised machine learning, relied heavily on categorization because large training data and computational power were not available. In modern AI, other techniques such as similarity clustering, function approximation, extrapolation, and prediction are used.
Carlos E. Perez has a sharp opinion on categorization:
“Human cognition does not revolve around categorization, rather, it’s based on just-in-time conceptual blending. Douglas Hofstadter identified this as analogy-making.” [1]
“Generative systems avoid the categorization problem because they simply don’t attempt to do it. These language models exploit the statistical measures embedded in our language use. These systems discover the regularities found in human discourse.” [1]
I highly recommend all articles written by Carlos E. Perez. I agree with him that categorization does not represent full understanding. But it is clear that categorization is the first step. Science is based on analysis and categorization is an essential mental technique in analysis. Among the higher functions of the human mind the ability to categorize stands out.
Group Theory
Particle physics today heavily uses group theory which is based on categorization. Physicists classify and categorize the elementary particles in a system labeled by intrinsic properties such as spin and charge. Categorization is not physics but it is an important first step to develop insights.
Category Theory
Category theory, especially as championed by John Baez, definitely represents progress in analytical thinking.
Lectures on Applied Category Theory by John Baez
“We build scientific understanding by developing models, and category theory is the study of basic conceptual building blocks and how they cleanly fit together to make such models. Certain structures and conceptual frameworks show up again and again in our understanding of reality. No one would dispute that vector spaces are ubiquitous. But so are hierarchies, symmetries, actions of agents on objects, data models, global behavior emerging as the aggregate of local behavior, self-similarity, and the effect of methodological context.” – David Spivak
Clustering instead of categorization
Based on various similarity measures we can come up with clustering algorithms to discover clusters that do not have a definition. This is categorization without labels.
Categories used in Spiritual Philosophy and Metaphysics
In all traditions of spiritual philosophy and metaphysics there is heavy use of classification and categorization. I find categories discussed in these posts very helpful.
Substance versus process
I will resist the urge to write about “substance vs. process” ad nauseam.
Question: what is more fundamental? substance or process?
- Answer 1: substance.
- Answer 2: process.
- Answer 3: both substance and process, equally.
- Answer 4: neither substance nor process, it is a mystery, cannot be described.
- Answer 5: perhaps we need to think about the modalities of the interplay between substance and process.
References
[1] Carlos E. Perez, The False Equivalence of Categorization as Thought
Suresh Emre
Renaissance Universal 