On Tue, Aug 16, 2011 at 11:09 PM, Craig Weinberg <whatsons...@gmail.com> wrote: > On Aug 16, 1:49 am, Stathis Papaioannou <stath...@gmail.com> wrote: > >> The I/O interface could involve neurotransmitters which are >> synthesised and released when the artificial neuron sees the >> appropriate voltage, and an enzyme which mops up the released >> neurotransmitter. > > Right. Not really an emulation of the function of a neuron, just an > emulation of how that function is controlled. An automated kitchen > which produces real meals from real groceries as opposed to artificial > meals from quantitative theoretical groceries.
A computer needs I/O devices such as keyboards and screens if it is to interact with its environment. >> The internals could be completely different: a >> computer modelling the biochemistry of a neuron and controlling a >> chemical factory. > > An automated kitchen with artificial cooks making real meals from real > groceries. > >>Another way to handle the interface would be not to >> use neurotransmitters but to directly connect to the neighboring >> neurons and control their membrane potentials, mimicking the >> depolarisation/depolarisation cycle. > > A pharmacy which distributes appetite suppressants as meal > substitutes. > >> If done that way there may have >> to be further signals to the postsynaptic neuron as the >> neurotransmitter may have an effect on regulating the number of >> surface receptors and hence to sensitivity to future stimulation. It >> would be difficult getting it exactly right, but there is no problem >> *in principle*. > > Each of those theoretical options would be orders of magnitude more > difficult to achieve and equally tremendously likely to fail compared > with replication. Why? It could be more robust. The biological brain eventually fails after a century or so in use. > There is still no evidence to suggest that the final > product would be experienced by the neuron or the brain made of such > neurons as something like human awareness. If it doesn't then that would allow for the possibility of partial zombies, which means you could be one now and not know it, which means being a partial zombie (if coherent) is not distinguishable from full consciousness and not a problem. >> > "While we are considering numbers, it is worth noting that there are >> > as many as 50 times more glia than neurons in our CNS! Glia (or glial >> > cells) are the cells that provide support to the neurons. In much the >> > same way that the foundation, framework, walls, and roof of a house >> > prove the structure through which run various electric, cable, and >> > telephone lines, along with various pipes for water and waste, not >> > only do glia provide the structural framework that allows networks of >> > neurons to remain connected, they also attend to the brain's various >> > house keeping functions (such as removing debris after neuronal >> > death). " >> >> > If your replacement neurons don't die, then you're changing the >> > relationship of them to 98% of the cells in the brain and the I/O is >> > different to the ecosystem overall. If you could manage to engineer a >> > replacement component which satisfies all of those electrical, >> > biological, and chemical roles, but still somehow manages to be, in >> > some significant way 'not a living cell' then there is still the >> > matter of whether or not the cell body itself is the thing that >> > actually experiences the various inputs and determines the outputs >> > according to uncomputable awareness-based algorithms or whether >> > experience somehow arises metaphysically through the aggregate of >> > unexperienced mechanical I/Os which can be replicated >> > deterministically. >> >> If you are proposing that the cell does a computation that a Turing >> machine cannot do then that would be an argument against >> computationalism (although not against functionalism). However, there >> is no evidence that there are non-computable processes in the brain. > > I'm not talking about computation, I'm talking about the implications > of biological realities of the nervous system. It does you no good to > compute the amount of plaque building up in your arteries if your > computation has no way to remove it. You would have to have the > machine itself pretend to die and then fool the glia into thinking > they already cleaned it up. It's just dumber and dumber. But if it is possible to compute when a neuron will fire that would allow the creation of an artificial neural network which would drive other neurons and muscles in the same way as a biological neural network would. It would be like any other artificial body part which a person might have and not notice. >> > If the former case is true, the replacement cell body may not be able >> > to produce the organic sense required to modulate the functions of the >> > cell in it's native improvisational mode so that it will neither fool >> > surrounding tissues nor perform the critical experiential function in >> > between inputs and outputs which would form the meat of perception and >> > awareness. If the latter case is true, there is no way to tell whether >> > the metaphysical requirements form instantiating high level awareness >> > could be satisfied by the design of the replacement. The exact >> > mechanism by which dumb I/Os are translated into nonphysical emergent >> > properties would have to be fully understood in order to accomplish >> > substitution by engineering. >> >> Do you understand this: >> >> (a) everything in the universe follows physical laws; >> (b) these physical laws are computable > > I understand that you believe that, but I think the worldview that > understanding arises from is obsolete. > Because we have qualitative experiences which are not reducible to > computation, and that is an undeniable fact with epistemological > validity equal to or exceeding that of physics, either: > > (a) Not all physical laws are completely computable or > (b) our qualitative experiences are not physical or > (c) The terms 'computable' and 'physical' are meaningless because they > include everything. > > I choose (a). Obviously our experiences of qualia like yellow and pain > are not meaningfully computable, and would have no conceivable place > in a cosmos that was purely computational. I'm not asking if qualia are computable, only if the observable behaviour of matter is computable. That means it would be possible to model on a computer what the behaviour of a collection of matter will be over time, given initial conditions. For a billiard ball that would be easy, for a weather system or human more difficult, but possible. >> So you are saying either that cells disobey the laws of physics or >> that there are certain laws of physics that are non-computable, but >> that only you know about them. > > I suspect that fully half of the laws of physics are unknown and non- > computable. They may be understandable though, through metaphor and > direct experience. These consequences of these laws are known by > everything that has awareness, which may be every physical phenomenon > to some extent or another. I'm not saying that I 'know about them' or > that 'Only I know about them' at all, I'm only pointing to their > existence as the possible solution to the mind-body problem, quantum > uncertainty, cosmology, and the crisis of post-modernity. Strictly speaking, quantum randomness is uncomputable, but a random process can be modelled by a pseudorandom process, or else a true source of randomness such as radioactive decay can be used. Also, real numbers are not in general computable, so if the world is continuous rather than discrete it would not be computable; however, every physical process will have a level of engineering tolerance, so infinite precision arithmetic would not be required and for practical purposes the world would still be computable. Any other ideas as to what could be uncomputable? -- Stathis Papaioannou -- You received this message because you are subscribed to the Google Groups "Everything List" group. To post to this group, send email to everything-list@googlegroups.com. To unsubscribe from this group, send email to everything-list+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/everything-list?hl=en.
