Cognition and Consciousness - Quantum Physics

Lecture 2 for the class on Cognition and Consciousness, March 12, 1997

Jussi Karlgren


Literature

David Bohm

The original connection between quantum physics and thought is made by David Bohm in 1951.

Roger Penrose

Roger Penrose is a renowned cosmologist. His book The Emperor's New Mind is 460 pages. On page 381 he gets into consciousness.

Penrose proves to himself that our consciousness is non-algorithmic, and that we seem - to our conscious selves - able to make decisions in a flash.

This to him is strange, and he speculates why. He finds that this could be explained by quantum physical thought processes that proceed in sub-graviton parallellism until they reach graviton level, when they collapse and produce a concsious thought.

Danah Zohar (with I. N. Marshall)

Danah Zohar is a physics and philosophy graduate and a psychotherapist. Her book the Quantum self is 260 pages. In chapter 6, page 75, she gets into consciousness.

Her view is strongly grounded in biology. She proposes a physical explanation of biological functions: she believes micro-wave transmissions that "align" cells to each other over distance explain how non-local phenomena may occur on a superatomic level.


Quantum physics

Quantum physics is a tool for studying subatomic particles. Very small particles at very high velocities behave differently from billiard balls and solar system planets; there are some non-intuitive effects of trying to observe and pinpoint features of individual particles. Both Penrose and Zohar note that where quantum physics seems mystical, it is because it is not complete, stable, or a finished theory. Since quantum theory cannot explain the collapse of wave functions adequately we should not try to use to explain more complex phenomena either. We need better tools.

Quantum physics and the brain

Edelman's straw man:

Obviously, the brain is composed of particles obeying quantum laws.

A notable case is that the retina accepts photons, which are small enough to behave strangely in terms of classical physics.

Quantum physics is strange. So is consciousness. Maybe there is something in common between the two.

He goes on to claim that mind without biology is a non-starter. But neither Penrose nor Zohar claim mind without biology! They try to explain biology.


Quantum physics and indeterminism

Some processes in quantum physics are indeterminate until measurements are made on them. Several different outcomes are possible, and the result needs not be determined before the outcome concerns particles on a larger scale.

The indeterminism is commonly modeled in a wave function - which is a combination function of possible outcomes, and determining the outcome is commonly termed "collapse of a wave function".


Superimposed states - the cat example

Picture from Zohar, page 40.

Picture from Zohar, page 41.

The cat: "The consciousness of the observer has killed it!" Zohar points out that this is silly. If there are paradoxes in the application of quantum theory, this has to do with quantum theory, not with consciousness. Mend the theory, she says.


Superimposed states - wave vs particle

Picture from Zohar, page 45.

Quantum physics and non-local correspondences

Examples:

Crystals - picture from Penrose, page 436.

Magnets - picture from Zohar, page 83.

Nervous systems as electric circuitry - picture from Penrose, page 390.

Twin particles - picture from Penrose, page 282.


The Brain as a Computer

The most important common thread in many of the searches for consciousness is scepticism towards the claim that human thought is algorithmic.

This scepticism is entirely based on introspection and awareness of conscious thought, and thus cannot be taken for more than a feeling. And as can be seen from most materials on the subject, the arguments range from "it is obvious" to "this is what once happened to a colleague of mine".

Typically, formal axiomatic systems such as formalizations of Turing machines assume that an algorithm operates on data expressed using symbols from a finite alphabet and that the data processed can be infinitely large.

The critique against algorithmic models of thought generally focuses on one obvious oversimplification in axiomatizing information processing.

Can you see which one?


Why the brain is no Turing Machine

So, typically, formal axiomatic systems assume that an algorithm operates on data expressed using symbols from a finite alphabet and that the data processed can be infinitely large.

It is patently obvious that the alphabet human information processors make us of is not finite. We coin new words at every turn of a clause, and we use old words in novel ways at the least opportunity.

Besides, it is not clear what the words mean, even when they are old and well worn.

Various forms of this critique are known as "the symbol grounding problem" or the "chinese room".


Again, Why the brain is no Turing Machine

Now, if, formal axiomatic systems assume that an algorithm operates on data expressed using symbols from a finite alphabet and that the data processed can be infinitely large, and we know from common sense and patient introspection and observation that the alphabet is infinite - what else is wrong with the assumption?

Human information processing never operates on infinitely large data. Sentences are never infinitely long. Clauses are never infinitely complex. In fact, it is theoretically possible to list all clauses that a human listener or reader can understand. (cf. Boman and Karlgren, 1996).


Predicate Logic as a Model for Thinking

If we try to formalize mutual knowledge - which is necessary for modeling dialog - we end up with a problem of infinite regress. (Smith, 1982).

This is ridiculous.


Turing Machines are Turing Machines and brains brains.

The preceding two examples show how properties of formal models can be used to prove things about original entities the models are based on. A simpler example, founded in formal linguistics, shows how the "problem" of intraposed subclauses goes away, when coded into the formal apparatus of the formalism rather than left to the grammar. (Joshi, 1994).

Thus, we can say that no, the brain is no Turing machine - not because we need to prove its non-algorithmic nature, but simpler, because the formalism used is simply not appropriate for modeling the brain's behavior.

The algorithmic nature of thought is not proven, nor is it disproven by virtue of Turing excercises.


Consciousness - Why?

Quite surprisingly, at second reading, one finds that neither Penrose nor Zohar treat consciousness in any organized way. Indeed, Penrose explicitly declines to define it (p. 406).

Penrose's main point is that our thinking is non-algorithmic.

Penrose says that consciousness as a side effect of running an algorithm is not possible. (pp. 405-7)

Penrose also believes that pure incomputability is not strong enough as an explanation for the perceived non-algorithmic nature of thought (pp 431-432): "However, it is very hard to believe that this kind of classical uncertainty can be what allows us our (illusion of?) free will."

But there is ample evidence that consciousness is a desirable mechanism in itself, and that conciousness is a tool, developed for a purpose. Neither of the authors address this; they see consciousness as a side effect of intuition, awareness, or of existence. Penrose discusses at length why natural selection would have developed consciousness if it weren't for some use (pp 408-9) but never proposes any use for consciousness apart as being a contrast to intuitive thinking.

If we see consciousness not as a side effect of algorithm operation nor of awareness or existence, but as a tool developed for a purpose, such as would Humphries (1980).


References