interesting, but I have a take on the Monty Hall problem.

The assumption is that Monty knows behind which door the car is and purposely opens the other door. This reveals a bit of information to you that makes the odds better if you switch.

But if Monty does not know which door contains the car he would just have to open a door at random. In this case the 50/50 estimate for the two remaining doors is correct. (unless Monty opens the door containing the car and then its over)

So it seems that humans are (subconsciously) guessing that Monty does not know which door contains the car.

Suppose instead that the procedure goes as usual except that before Monty opens the door he very obviously has a good look behind the wall so it is clear that the contestant knows that Monty knows.

I believe that in this case the contestants are more likely to switch and this shows that the behaviour of the contestant is influenced by a mental model of what Monty knows, this mental model requires priming, and the artificial game show situation does not do that.