For Wittgenstein....the question 'Can a computer think?' is not a question about computers. It's a question about our PERCEPTION of the machine....not about the machine itself. This is reflected by Garland in that Nathan's 'test' is revealed to be more about Caleb than it is Aja. In other words....Nathan deceptively presents the test as an effort to give a characterization to Aja. 'Can she think?'. But what it's really about when we find out his true intentions.....is 'will Caleb believe that Aja thinks?'.

And that is the essence of the turing test in the first place. The criteria is in our PERCEPTION....not something inherent in the AI. The turing test doesn't necessarily claim that a computer CAN ACTUALLY THINK. It's based on the behaviourist/pragmatic spirit that if we perceive it to be thinking.....that is the SAME THING as it actually thinking. In essence, our perception CAN'T BE WRONG in this regard. Our perception actually determines or drives the truth of the statement. There is 'no truth' that exists outside our perception. This is a subtlety of Turing's philosophy that is not acknowledged in other sci-fi films that have referenced him. ie Blade Runner. And it's grounded in pragmatism/behaviourism/post-modernism as a whole.

It's also a good way to sum up pragmatism which is often misunderstood. In the pragmatic method....when we ask 'Can a computer think?'.....we don't attempt to determine whether a computer is thinking or not. We consider the CONSEQUENCES of us BELIEVING if a computer can think. If those consequences are favourable....than we will accept it as true.

This is where Wittgenstein broke away from pragmatism. I'm pretty sure Wittgenstein would say that there ARE no consequences in believing that a computer can think. We can say it does...we can say it doesn't and it's not going to have practical consequences either way. Therefor, it's a metaphysical (or philosophical) question and thereby meaningless....purely semantical. 'Philosophy is when language does nothing'. So for Wittgenstein...the question 'Can a machine think' is about how we use the word 'think'. Not about the machine itself. The reason the question puzzles us is because of the inherent limits and restrictions of language. In this specific case...we're trying to use the word 'think' in a way that is foreign to the way we are accustomed to using it. Which is why he called his work 'the end of philosophy'.