Quote:
| at least have the ability to be programmed to "imitate" speech |
If I add to an abacus a list of instructions for how to use it (which is precisely what a computer programme is) and the correct input/output devices (conceptually possible), it would be perfectly capable of doing that. If it had a limit it would be performing the calculationsfast enough -- but this is theoretically possible, even if not practically possible.
The point I am making is this: a computer is merely a computational device. All it can do is computational tasks. Take a Turing machine and add in the basic elements of a von Neuman architecture (memory and I/O), and you have a conceptual framework for what a computer actually is. If that is not the case, why don't you tell me what a computer is, if not a computational machine.
The fact that a computer has a lot more (and a lot faster) ways of computing does not make it anything but a bigger, faster computing machine. No-one who does AI would argue other-wise. They would debate about whether cognition is fundementally and essentially computational; whether it can be successfully modelled on a computational device; and so on.
Are you familiar with Searle's Chinese Room argument?