Issues with Searle’s “Chinese Room”
John Searle proposes the Chinese Room test and then says that because the test is insufficient to determine a machine’s consciousness, machines can not be consciousness. I have a number of issues with his thesis:
Searle is correct in that his test can not determine consciousness, but he is wrong in concluding that a conscious machine can not exist because his test is impotent. Searle is only proving the inadequacy of his test, it is a straw man argument. He is not proving the inadequacy of all tests or of the feasibility of a conscious machine.
Let’s say we defined consciousness as the process of observing ones immediately previous thoughts in a symbolically meaningful way. With this simple definition a test can be created that would allow us to classify agents as having consciousness. Additionally this definition allows us to simply refute Searle’s test since it is clearly insufficient given this definition. What I mean by that is that per the definition I/O is not enough test the consciousness of an agent, the agent’s internal processes must be examined possibly using something along the lines of advanced fMRI, code inspection etc..
Searle’s argument implicitly assumes that if something is not testable it is not true, this assumption is false. One could easily create the reverse Chinese room argument where there is are 2 Chinese speakers, on on each side of the wall. Since the person outside of the room(having knowledge of the “Chinese room problem”) can not determine if the agent inside the room is machine or human he just as well could conclude that the agent is a machine and does not have consciousness. IE both conclusions are equal probable as Occom’s razor does not apply to this situation. Searle’s conclusion is biased in his favor and is inconsistent with itself.