Jean-louis Dessalles

DIG Team

→    Home page

I am very much impressed by three theories, in radically different domains (anthropology and cognition). I am ready to bet that these theories will be remembered in a far future.

His theory is one of the rare successful attempts to solve the many apparent contradictions between anthropological universals and what we expect from evolution through natural selection. Knight expands Turke’s explanation of sexual specialisation (with men leaving women alone when hunting). Woman synchronicity becomes central in explaining various universals like food restrictions, taboos, costly rituals, and myths.
Knight is aware of many constraints (esp. evolutionary and ethological constraints) that are merely ignored by most scholars in anthropology. His great achievement is to put logic in what, otherwise, looks like a vast mess of anecdotal anthropological facts.
Knight’s logic is impressive! Just read his Blood relations – Menstruations and the origin of culture (Yale University Press, 1991). Read the New Scientist’s review by Kate Douglas. Leyton’s model is a remarkable synthesis of many previous theoretical advances like Gestalt theory. Leyton’s basic idea is that perception works through structure transfer. Mathematically, transfer results from elements of a group acting on another group. You perceived your tiled floor, not as a set of disconnected tiles, but as one repeated tile. This is why you find structure in the floor. One tile is transferred through the group of 2D-integer translations. The world appears to us as a nested construction in which each step consists of a maximal transfer of structure.
Leyton construction captures a fundamental property of human cognition. Our brain is not any universal Turing machine. We are bound to see the world through nested group operations.
Leyton’s work, though mathematically intricate, is insightful. I am impressed by its explanation power. Just read his Symmetry, Causality and Mind (MIT Press, 1992) and A generative theory of shape (Springer Verlag 2001). Peter is well-known for his book Conceptual Spaces (2000), in which he explained for the first time in a clear way (1) that meanings are similar in nature to perceptions and (2) that they have certain geometrical properties like convexity. His account, once stated so clearly, seems evident, as it solves many problems and paradoxes in which traditional semantic theories get tangled up (especially all theories that postulate a ‘language of thought’, including many theories in cognitive linguistics).

Peter’s rencent book, The geometry of meaning, is a fantastic follow-up. The Publisher asked me how I liked the book and printed by appraisal on the back cover:

"Peter Gärdenfors is creating a new science of meaning. The recent ideas, expressed so clearly in The Geometry of Meaning, make his achievements even more impressive. The book leaves us with the impression that semantics may be a tractable problem after all."

Chaitin applied Algorithmic Information and MDL to epistemology. The idea is that best theories are ideally obtained by minimizing the sum of their desciption length (in bits) and the desciption length of exception (again in bits).
You can download the book he wrote for a large audience for free:

Chaitin, G. J. (2005). Meta Math! The quest for Omega. Vintage Books, ed. 2006.

The fact that Algorithmic Information might be applied beyond the strict limits of pure maths inspired me when I developed Simplicity Theory, and it motivated me to create the MOOC Understanding AI through Algorithmic Information Theory. ... where once self-replicating entities, thus appearing as an ancestral life form previous to cells.

Root-Bernstein, M. & Root-Bernstein, R. (2015). The ribosome as a missing link in the evolution of life. Journal of Theoretical Biology, 367, 130-158.

I find this theory illuminating and I am only surprised that it is not better known. ... and the solution to the binding problem both result from neurone synchronization was first proposed by Christoph von der Malsburg in 1981.

von der Malsburg, C. (1981). The correlation theory of brain function. Göttingen: Internal Report 81-2, Abteilung für Neurobiologie, MPI für Biophysicalische Chemie.

This hypothesis has then been popularized by Francis Crick (yes, that one) and Cristof Koch:

Crick, F. & Koch, C. (1990). Towards a neurobiological theory of consciousness. Seminars in the neurosciences, 2 (), 263-275.

and had some influence in neuroscience (see for instance works of Wolf Singer). However, probably due to the difficulty of recording many cells simulatenously, the neuroscience part has not yet been developed at the level it derserves. Similarly, the modeling part initiated by von der Malsburg has somewhat stalled (as far as I can see) after a good start (see for instance (Shastri & Ajjanagadde, 1993)). I still expect a revolution in this domain.
    

→    Home page