Wednesday, August 13, 2014

Freeman-Stapp continued

This is an attempt to get around what I see as  the absurdity of mice making decisions, Schrödinger’s cat collapsing the wave-function etc all of which are implicit in the Freeman-Stapp dialog below

Consider this; Gödel, Von Neumann  Schrödinger and other Platonists/voluntarists/non-dualists were contemporaries from Mittel Europ. So let’s try and weave together a story, one that brings Stapp’s work into dialog with Penrose

A hypothesis; “Only systems of formal power >= standard arithmetic, used intentionally, can initiate process 1/ Penrose’s R”.  This mixes Gödel, Von Neumann  Schrödinger and may be a mite too courageous?

So the wheeler participatory universe needs symbolic humans., as does any process 1 event. What we’ve gained is links with computational complexity, the ; Gödel theorems, and now we can make our way back to cog sci

Sean O Nuallain


It  is important that any new discussion be well-informed. Thus, this is an incomplete list of the subjects required for mastery of Cognitive Science

  1. Cognitive psychology
  2. Philosophy; of mind, of language, epistemology
  3. Algorithms including formal language theory
  4. Linguistics; generative, functional, computational
  5. Neuroscience; microscopic (including neurobiology), mesoscopic, macroscopic
  6. Math; set theory, linear systems, chaoplexity
  7. AI
  8. Anthropology
  9. Physics; classical and  the observer in 20th century physics;

Briefly, we can perhaps progress  with formal languages,  the related topic of computability, and intentionality;

Without these concepts, you are stuck with the bare observer and process 1. With them, and their equivalents is set theory, we have the equipment to talk about how a mind can represent and act on the world in a way transcending sensorimotor loops

PPS A short tutorial

One central problem that keeps emerging is that we have QM talking about “decision”, “determination” (Unruh) while cog sci does the same thing with the same vocabulary but about different phenomena

Physics has a very well-attested methodology. So has cog sci, particularly if we introduce concepts from computation/formal language theory as I think we should.

For the record, I think the schema I’ve outlined works; in fact it is inter alia a via media between you and Penrose. If our system F is of formal power >= standard arithmetic we get a Gödel sentence G(F). Add it to F

F+ G(F).  = G’ 

This will have a Gödel sentence G’(F).

While “being” F, humans can “see” the truth of G’(F). (Penrose following Gödel)

An intentional system of formal power >= standard arithmetic therefore is a game-changer



In answer to your question, no, I don't reject causality. None of us
could live without it. And I believe as you do that "QM/QFT are
theories that preserve causality." I believe that my theory will
explain why we both believe that, when we learn how brains construct
QM/QFT and causality, and use them both so effectively. Learning that
is a worthy goal to strive for. Whether what we find is something that
we discover already existing, or something we create, or something our
machines create that we cannot understand, are matters for conjecture.


On Aug 12, 2014, at 5:37 PM, Stanley Klein wrote:

 Yes, I'm quite familiar with the Hilbert transform. And I'm pleased that  I raise it because I believe that for Henry the quantum Zeno effect that can take place with observations is causal.

    I should add that I'm in agreement with Henry about the nature of observations, though we may differ on what constitutes being an 'observer'.

 You said: "
What I hope for from this chat room is breaking away from the prevailing neural network Schema of Newtonian thinking and causality in point processes to a new paradigm that uses QM/QFT to explain how brains work"

 Does that mean that you reject causality? I hope not since I believe QM/QFT are theories that preserve causality.

 On Tue, Aug 12, 2014 at 3:22 PM, Walter J Freeman < wrote:

 As you know, the Hilbert transform is a linear operator, that is useful for EEG/ECoG/LFP measurements of frequency modulation, and for the instantaneous amplitude phase, giving high resolution for measurements of amplitude and phase modulation. It has nothing to do with causality. I wonder why you raise that vexatious issue in this context?

 In my opinion causation is a property of mind, not matter. It is a by-product of the action-perception cycle. We act and sense the act as 'cause' and then perceive the change in sense input as 'effect'. Hilary Putnam (quoted in my 1995 "Societies of Brains") concluded that this elementary dynamic emerges long before the capacity for language. It is the foundation of our epistemology and social organization. It forms so early in infancy that it is heretical to question it. It is an attribute of primitive thinking that is essential for engineering, medicine and society, but is unnecessary and often misleading and even stultifying in science. The list is long of writers who proposed synonyms. Hermann Haken created the oxymoron 'circular causality' to take the place of 'agency' in describing collectives scientifically typically that usage is only cosmetic.

 Witness the progression of our explanations of the motion of the sun from Phaeton's chariot to Ptolemaic spheres to Copernican geometry to Newton's gravitation to Einstein's curved space - from agency to relativity in successively more sophisticated schemata.
​​What I hope for from this chat room is breaking away from the prevailing neural network Schema of Newtonian thinking and causality in point processes to a new paradigm that uses QM/QFT to explain how brains work, and thereby helps to create a unified view that transcends some of the limits, confusions, paradoxes and uncertainties we now experience. I see the findings in the past 50 years of brain dynamics as the greatest advance in 2500 years. Yet most neuroscientists don't yet see the significance of these new data. We need informed commentary from physicists, who have their own limitations. This is why I am overjoyed that Henry understands the significance of the inverted phase cone. I feel like Robinson Crusoe when he perceived a human footprint in the sand.

 At your request I attach a reprint of my views on causality, based on my 1995 book.


 On Aug 12, 2014, at 1:19 AM, Stanley Klein wrote:

 Does your Hilbert transform maintain causality?
 Can it replace Granger causality for looking at how signals move around in the brain?
 Even Granger causality is challenging to implement as I recall.
 One approach would be to send us the link to your article that discusses the causality issues.

 On Mon, Aug 11, 2014 at 8:56 PM, Henry Stapp < wrote:
 This certainly needs to be part of this thread. It seems to me to lend (strong?) support to the idea

 that Walter’s ideas fit into QM in the way that I had been hoping/pushing for!

 From: Walter J Freeman []
 Sent: Sunday, August 10, 2014 7:04 PM
 To: Henry Stapp
 Subject: Re: The inverted cones [continuing]

 Dear Henry,

 Your understanding of our "inverted cone" is correct. Your "melting ice cube" is an apt metaphor. We first found the cones in the olfactory bulb in 1987, and found them in the visual, auditory, and somatic ECoG in 2002

 We have substantial experimental evidence bearing on whether or not phase cones are NCC.

 The best documented data are wave packets that typically formed three times in the interval between conditioned stimulus (CS) and response (CR). Our subjects were trained to discriminate between a reinforced CS+ and an unreinforced CS-. The wave packets had a steady-state carrier frequency  in time and nearly so in space. The Hilbert Transform of the 64 records of the electrocorticogram (ECoG) in the beta or gamma range gave the analytic amplitude and phase. Each multichannel time series gave for the wave packet a 1x64 feature vector.

 The analytic amplitude vector gave a spatial pattern of amplitude modulation (AM) that specified a point in 64-space. Multiple vectors from 20 trials of CS+ and 20 trials of CS- gave two clusters of points that were linearly separable and categorizable. Our subjects included cats, rabbits and human volunteers (scalp EEG).

 We believe that our subjects were conscious, but that is circumstantial. If they were conscious of the CS that they were categorizing and made the correct choice of CR+/CR-, than our categorized AM patterns were NCC.

 Each wave packet with its categorizable AM pattern was accompanied by a spatial pattern of phase modulation (PM) that specified a 1x64 phase feature vector. The two clusters of points in 64-space were not categorizable with respect to CS+/CS-. The PM pattern told us when and where each wave packet formed, how large it was in area, and how long it lasted. None of these aspects bore on consciousness directly. Moreover, many phase cones occurred without any categorizable AM pattern, CS or CR. They manifested neural avalanches in the maintenance of cortical criticality. We conclude that the PM is not a NCC because it is not correlated with our marker for consciousness, which is a correct CR in the context of a set of trials.

 To answer your questions, the wave packets occur often enough to be candidates for NCC, but Sean is right, that the phase transition by which wave packets produce memories was not what von Neumann explicitly had in mind, and is,  as he says "entirely different". Whether it "evades" or "solves" the "hard problem" I cannot say, but I do think that, as you say, our "paths  might connect at this point".


 On Aug 8, 2014, at 11:22 PM, Henry Stapp wrote:

 Dear Walter,

 I just want to stress that I am really interested in those (inverted) cones that, instead

 of spreading out from sparsely distributed sites on the cortex, like the transformation

 of sub-freezing-temperature water to ice,   rather contract from the outside in, like an

 ice cube melting from the outside boundary in.

 Is my understand of these two kinds of cones, which you say are roughly equal in number,

 is correct (I am not sure it is) then you seem to have very potent evidence that your phase

 transition idea is really correct.

 Do you have evidence of correlations between these events and conscious events?

 Sean suggests that these brain events are not connected to conscious events, but to an entirely

 different kind of brain event, thus evading the “hard problem” that would ensue if your events

 were empirically correlated to conscious experiences.

 I had been hoping that our two paths might connect at this point: that perhaps your brain events

 could be the brain correlates of the conscious events that play so big a role in Standard

 (Copenhagen- vN) QM.

 Or is such a correlation ruled out by the infrequency of your cone-related events?



 Walter J  Freeman MD, Professor of the Graduate School

 Department of Molecular & Cell Biology

 Division of Neurobiology, Donner 101

 University of California at Berkeley

 Berkeley CA 94720-3206 USA

 Tel 510-642-4220  Fax 510-642-4146

No comments:

Post a Comment