Cicurel, Ronald and Miguel A. L. Nicolelis. 2015. The Relativistic Brain: How it Works and Why it Cannot Be Simulated by a Turing Machine. Kios Press, Natal, Montreux, Durham, São Paulo.
There are too many fascinating points in this discussion to fully describe.
But what is notable is their scepticism of the more grandiose aims of the EU Human Brain Project (HBP) and Obama’s White House BRAIN Initiative.
The most important issue is this: Nicolelis and Cicurel propose a hypothesis called Relativistic Theory of the Mind, which is very similar to the Conscious Electromagnetic Information (CEMI) Theory of Johnjoe McFadden (McFadden 2002a, 2002b).
In essence, in their book and in the video, Nicolelis and Cicurel argue as follows:
(1) there are two types of information as follows:BIBLIOGRAPHY(1) Shannon information, which is purely syntactic and devoid of semantics, andTuring machines can only deal with Type (1) information, not Type (2). The human brain and mind can deal with both types of information.
(2) Type II or Gödelian information, which can convey semantics (Cicurel and Nicolelis 2015: Kindle Edition Appendix 1).
(2) If we take as our scientific hypothesis that our mind is an emergent property from brain activity, then, if it is not just chemical activity or neural firing, presumably it must be some emergent property deploying one of the four forces: the weak nuclear force, strong nuclear force, electromagnetic force, or gravitation force. The electromagnetic force is the most probable of these.
(3) The aggregate of our internal brain neuronal electromagnetic fields, produced by neurons and nerve bundles or white matter, are the fundamental cause of consciousness. The neurons and diencephalic white matter produce small electromagnetic fields and so the white matter bundles of the brain are a type of “biological coil.”
These complex patterns of neuronal electromagnetic fields have to be considered as a global/aggregate phenomenon in the brain, and they seem to be capable of analogue-like information processing and even of modifying the neural networks and interacting with them.
Nicolelis and Cicurel propose that the brain has both a type of digital and analogue information processing: the frequency spikes of neurons function in a digital-like manner. But the aggregate masses of neurons continuously create dynamic neuronal electromagnetic fields that create a type of analogue information processing.
The human brain that produces the mind is an integrated system and in that respect resembles an analogue computer (where there is no separation between hardware and software), not a digital computer. Information in the brain is embedded in matter, and so cannot be extracted. In other words, information in human brains is substrate-dependent.
The emergent properties from neuronal electromagnetic fields and their interactions are mental life, and they are analogue and non-computable.
The brain creates an internal representation of the human body and the external world, in which this model is constantly updated.
A digital computer or Turing machine cannot reproduce or create a mind like ours. And this is before we get to a biological explanation of emotions, etc., which manifestly computers cannot experience.
(4) if there is any kind of metaphor with computers, our brain/mind is not a Turing machine, but a much more powerful computer called a “hypercomputer” (Cicurel and Nicolelis 2015: Kindle Edition Appendix 1).
(5) All our higher-order mental traits like pain, tactile, auditory and visual sensation, consciousness, dreaming, etc. are a type of analogue brain processing (Cicurel and Nicolelis 2015: Kindle Edition Appendix 1).
(6) The Relativistic Theory of the Mind predicts that(1) in deep sleep or comatose states or during anaesthesia, the white matter “biological coils” do not function.(7) Turing machines and digital computers can never attain consciousness and the whole Singularity hypothesis (as argued by Ray Kurzweil) that human minds can be reproduced in digital computers is completely false and a cult-like delusion.
(2) in terms of evolution, brains of higher and higher complexity required more and more white matter, and so this should be reflected in the gradation of animal brains.
Cicurel, Ronald and Miguel A. L. Nicolelis. 2015. The Relativistic Brain: How it Works and Why it Cannot Be Simulated by a Turing Machine. Kios Press, Natal, Montreux, Durham, São Paulo. Kindle Edition.
McFadden, J. 2002a. “The Conscious Electromagnetic Information (Cemi) Field Theory: The Hard Problem Made Easy?” Journal of Consciousness Studies 9.8: 45–60.
McFadden, J. 2002b. “Synchronous Firing and its Influence on the Brain’s Electromagnetic Field: Evidence for an Electromagnetic Field Theory of Consciousness,” Journal of Consciousness Studies 9.4: 23–50.
Seems plausible that the computer system dealing with type 2 information already exists. Deep belief networks, what google uses for its image search, are applicable to certain symantic problems. They use an algorthm which sweeps over the whole network to speed up training. And arguably produce dream like images as well.
ReplyDeleteOf course the information in any neural network is substrate dependent.