Praca ta została dofinansowana przez Narodowe Centrum Nauki (NCN) w ramach grantu badawczego OPUS 23, nr 2022/45/B/HS1/02700

This paper discusses perspectivism and relationalism in the two versions of Relational Quantum Mechanics (RQM): that initiated by Rovelli (Int. J. Theor. Phys.,35(8), 1637–1678 1996) and the other, by Adlam and Rovelli (Philos. Phys., 1(1) 2023). To this end, we offer a substantial discussion of this interpretation. We investigate the issue of the disagreement of results between different observers; distinguish four senses of agreement: strong, weak, very weak, and perspectival; and argue that the old version guarantees only the very weak and perspectival agreement, whereas the new version guarantees the weak agreement as well. The problem that individuals in RQM need to be propertyless when they do not interact is also investigated. Concerning the issue of probabilities in RQM, we argue that to express probabilities involving events relative to different systems, a new kind of probability spaces needs to be devised. The final problem that we identify concerns underspecified dynamics; we argue that some additional postulates are indispensable to make RQM’s dynamics fully specified. The last two points lead to the conclusion that, contrary to declarations of RQM proponents, this interpretation requires some modifications of quantum formalism and not merely conceptual changes. Finally, concerning the problem of testimony, while the new version of RQM was devised to solve it, we argue that in the old version, this problem can be alleviated to the extent that perspectivism permits.