[ExI] The symbol grounding problem in strong AI

[Gordon Swobe]

--- On Mon, 12/28/09, Stathis Papaioannou <stathisp at gmail.com> wrote:

> You claim both that the physics of neurons is computable

Yes.

> AND that it is impossible to make program-driven neurons that behave
> like natural neurons, which is a contradiction. 

No you misunderstood me, and I should have made myself more clear. I meant that your artificial neurons in your experiment would not act as would have the natural neurons that they replaced -- not that they would act in a manner uncharacteristic of neurons.

> Even Searle agrees that you can  make artificial neurons that behave 
> like natural neurons, 

As do I. That was the #2 point in my post to you yesterday. I quote myself here:

"2) I believe we can in principle create neurons "based on" those computer blueprints, just as we can make anything from blueprints, and that those manufactured neurons will behave exactly like natural neurons."

To my way of thinking, your cyborg-like thought experiment takes a single snapshot of a process. You want to focus on that single snapshot but I look at the entire process. In that process you have me changing into a computer simulation. While the circumstance pictured in that single snapshot seems odd to me subjectively, to you as the objective observer everything would seem quite normal. 

At the end of that process I no longer exist as an intentional entity. Although that simulation of me exhibits all the objective characteristics of a person with intentionality, my simulated consciousness no longer has a first-person ontology. 

I once believed as do many here that a computer simulation of me would equal "me". I now understand that such a simulation would no more equal me than would a photograph of me. The simulation of me differs from the photograph of me only in the accuracy of the caricature. Rather silly of me to have thought otherwise!

-gts