in some form of mental computation or thinking. The capacity to perform simulations
or thought experiments
would then be the unique feature distinguishing human beings from other animals, in the words of Descartes expressed as
- I simulate, therefore I exist.
Turning this argument around we may ask if a mechanical robot with the capacity of performing simulations of itself and the reality around would have a soul? Let’s see first what a robot you can buy today can do
Let’s next see what capacity to think
this robot could have. Suppose the robot has a computer inside with software in the form of a digital model of the robot itself and the world around allowing simulation of the action of the robot. This would give the robot the capability to test alternative actions before actually trying to perform them. For instance, it would allow the robot to decide if it would be safe to cross a road with traffic, or if engaging in a fight with another robot would be advisable, and if so what actions to take. This software could be similar to the software of computer chess
evaluating sequences of moves and choosing the best
according to some criterion.
Such a robot would beat you in chess and would also have other capabilities:
Would you say that such a robot is intelligent and that it has a soul? Maybe intelligent, but maybe not soulful because it would lack feelings in the same way as computer chess lacks feelings (as far as we can tell).
The Role of Computational Mathematics
To simulate physical reality such as the motion of a robot, the language of mathematics can be used to construct digital computational software capable of modeling physics seen as analog computation.
Computation is here seen as a sequential process taking input to output, or from cause to effect, which in real physics is performed in analog form through the action of physical forces, and which in computational simulation is perfomed by digital processing.
The Game Playing Robot
The world champion chess robots Deep Fritz
and Deep Rybka 3 Aquarium
is likely to beat any human player, and similar robots can be expected to outperform humans in all forms intellectual games based on given rules, like bridge and even poker.
Further, simulating robots would have distinctive advantages in all forms of sports because an optimal control by simulation would possible. The optimal motion in high-jump under given restraints on physics would be possible to first compute and then realize, and no jump could be better than an optimal jump!
The Feeling Robot
The emotional life of Deep Fritz is probably rather dull, and the question is what need or use of feelings robots might have? Would it be helpful for the interaction of many robots towards realizing some common goal, or prior to that for formulating some common goal? Could happiness and reward be quantified e.g. in terms of dollars as a way of stimulating action instead of lethargy? Or would robots be better robots without feelings, with only the capability to simulate and optimize?