Stathis Papaioannou wrote:

George Levy wrote:

[quoting Stathis]
You would also need to know the electrical potential at every point of every cell membrane; the ionic gradients (Na, K, Ca, pH and others) across every cell membrane, including intracellular membranes; the type, position and conformation of every receptor, ion channel and other proteins; the intracellular and local extracellular concentrations of every neurotransmitter; the workings of the cellular transport, synthetic and repair mechanisms for each neuron and probably also for each supporting glial cell; the intracellular and extracellular concentration of other small molecules such as glucose, O2, CO2; how all of this is changing with respect to time; and probably thousands of other paramemters, many of which would currently be unknown.

This is unfair. According to this strict standard you are not the same person today as you were yersterday. In fact even an automotive transportation method would violate the above standard. We can't expect a Star-Trek tranporter to have more "High Fidelity" than a car. The question is how much can we relax the standard until the person at the output is "not the same" as the person at the input. In a brain substitution experiment, when should the patient say "yes doctor" or "no doctor"?

The "high standard" I have described does not go nearly as far as copying the exact quantum state of every atom. It is merely aknowledging the fact that information in brains is not stored in the anatomical arrangement of neurons, any more than data on a computer is stored in the computer's circuit diagram.

If you scan the anatomical arrangement of synapses *and* the concentration of all the relevant proteins at the synapses, you probably would have enough to run a simulation that would act like a continuation of the original person. The upload might find he'd lost his short-term memories of what happened immediately before he died and his brain was frozen (just as we often do when we regain consciousness after being suddenly knocked unconscious by an accident), but as I understand it long-term memories are stored in terms of the pattern of synaptic connections and the neurotransmitters at each synapse, and as long as the simulated neurons behave closely enough to how the original neurons behaved, shouldn't the upload behave like the original person in terms of personality, thought processes, emotions, preferences and so forth?


Reply via email to