The Markov Test

The Turing Test is used to determine whether a program exhibits human behavior. I’d like to propose that we use a Markov Test to determine whether a human being exhibits interesting behavior:

  1. Train a Markov chain text generator on a sufficiently large corpus of the person’s speech and writing.

  2. Put samples of the person’s actual utterances and equal-sized samples of the generator’s output in front of a human judge.

  3. If the human being can’t tell the difference between the two, classify the person as “uninteresting”.

I’m pretty sure that some prominent people would fail to pass this test…