The epistemic aspect of the surprise exam puzzle

In the paper, the puzzle known as the ‘surprise exam paradox' is examined. The puzzle is generated by the following story: A student was told that the following week, from Monday to Friday, she was going to get an exam and that a day before the exam she would not know that the next day would be the exam day. The student observed that if the exam was going to take place on Friday she would know that Thursday night, so she concluded that Friday was not suitable for the exam. By analogous reasons she also excluded all days from Monday to Thursday and came to believe that there would be no exam at all the following week. But on Monday the exam took place, and since that was a surprise for the student, we can observe that the exam was held in perfect accordance with the initial announcement. The puzzle is generated by the fact that the student's reasoning seems to be quite sound whereas its conclusion turns out to be wrong. The author proposes a solution to the puzzle close to the one offered by Quine but differing from it in an important respect. The author agrees with Quine in the point that the student should not exclude any day from the period suitable for the exam. The author's solution differs from Quine's: he explains this by showing that the student's reasoning affected her epistemic state, and that the student should have taken this into account (but failed to do it). Thus, in the author's solution, the dynamics of the student's epistemic states plays a crucial role. The author demonstrates that the seeming paradox is generated by the following circumstances: 1. The student's reasoning was dependent on her epistemic states. 2. In the course of the reasoning, her epistemic state was substantially changed. 3. This change nullified the justification of the conclusion she came to. 4. In the traditional version of the story, the third point was not noticed by the student, which brought about a wrong expectation on her side. That point seems not to be noticed by the authors of some solutions to the puzzle either. The author of the paper also shows that the puzzle in question differs from logical paradoxes in the fact that a non-logical factor - the causal link between reasoning performed by the agent and her epistemic states - plays an essential role in it.

Keywords

Authors

References