For an example of the difficulties involved, consider John Searle's celebrated Chinese Room thought experiment, which purports to show— contrary to Alan Turing's claims—that input-output characteristics alone are insufficient to determine the existence of a conscious mind. Intuitively this conclusion seems right: a sleeping person, immobile and inattentive, might nevertheless be experiencing vivid dreams. Conversely, I can drive a familiar route without forming any conscious record of the journey that took me to my destination. But the Chinese Room does nothing to prove this thesis, for it is a thought experiment and the trouble with thought experiments is that the researcher chooses not only the experimental conditions but also the results. This makes them useful for testing the internal consistency of ideas but almost useless for probing mysterious, apparently emergent phenomena like consciousness. (To see this, carry out the same thought experiment on the 3kg lump of electrophysiological goo called the human brain and if you're being consistent then you'll get the same result: there appears to be no conscious understanding anywhere inside.)