Wigner's friend

Wigner's friend is a thought experiment, proposed by Eugene Wigner, that sharpens the measurement problem by adding a second observer. Imagine a friend inside a sealed laboratory who measures a quantum system and records a definite result — say, spin up. From the friend's point of view, the measurement is over and the answer is settled. But Wigner, waiting outside and treating the whole sealed lab quantum-mechanically, should describe the friend and the system together as one big entangled superposition of 'friend saw up' and 'friend saw down'.

The puzzle is that two honest observers seem to give incompatible accounts of the same situation. For the friend a definite outcome has happened; for Wigner no outcome has happened yet, only a superposition awaiting his own measurement when he opens the lab. Whose description is correct? Does the friend's observation count as a real measurement that collapses the state, or is the friend just another quantum system that only acquires a definite result when Wigner finally looks?

Wigner's friend dramatizes the awkward question of where, if anywhere, a measurement truly takes place, and whether 'observation' has any privileged or even consciousness-related status. Recent variants have turned the puzzle into testable theorems showing that certain comfortable assumptions — that observers always agree on facts, that measurements yield observer-independent outcomes — cannot all hold together. The scenario does not have a settled resolution; it is a clean lens through which the deepest interpretive disagreements come into focus.