[DSRP Evidence](https://dsrpevidence.org/)

# Frauchiger-Renner and Wigner's friend

Measurement & self-reference · 

## Why it looked like it would break the theory

Four agents run a chain of valid inferences and arrive at contradictory conclusions about the same outcome in the same run. The contradiction is not left standing between two people: it ends up inside a single agent's record, which holds both answers for the same variable at the same moment. That appears to require a fifth rule for contradictory perspectives.

## Why it failed

**The counterexample failed** because the scenario is nested and the calculus already represents nesting. Each laboratory is a whole and the agent inside it is a part. An outer agent's view takes in the whole laboratory, inner agent included — so the outer agent is not a second observer standing beside the inner one, but a point whose view contains a system the inner one belongs to. A perspective may contain another perspective: P ∈ P. Every conclusion in the chain is a view taken from a point. The step that produces the contradiction carries a conclusion forward without the point it was taken from, and a view with no point is a Perspective with one pole missing, which Equality forbids. Representing all of this — including the contradiction — takes identities, wholes with parts, relationships among them, and points with views. No fifth pattern, ninth element, or fifth dynamic is required. Whether the argument's conclusion stands, and which of its assumptions ought to be abandoned, are questions about the information. The calculus does not settle them and does not need to.

Settled by [R12](https://dsrpevidence.org/resolution/12). Not yet published.

[Read it](https://doi.org/10.1038/s41467-018-05739-8)

[All cases](https://dsrpevidence.org/counterexamples)