Why, then, don’t more people realize that many worlds is correct?
I am going to try and provide short answer, as I see it. (Fighting urge to write about different levels of “physical reality”.)
Many Words is an Interpretation. An interpretation should translate from mathematical formalism towards practical algorithms, but MWI does not go all the way. Namely, it does not specify the quantum state an Agent should use for computation. One possible state agrees with “Schroedinger’s experiment was definitely set up and started”, another state implies “cat definitely turned out to be alive”, but those certainties cannot occur simultaneously.
Bayesian inference in non-quantum physics also changes (probabilistic) state, but we can interpret it as a mere change of our beliefs, and not a change in the physical system. But in quantum mechanics, upon observation, the “objective” state fitting our knowledge changes. MWI says “fitting our knowledge” is not a good criterion of choosing quantum state to compute with (because no state can be fitting enough, as example with Shroedinger’s cat shows) and we should compute with superposition of Agents. MWI may be more “objectively correct”, but it does not seem to be more “practical” than Copenhagen interpretation. So physicists do like to cautiously agree with MWI, then wave hands, proclaim “Decoherence!” and at the end use Copenhagen interpretation as before.
Introductory books emphasize experiments, and experimental results do not come in form of superpositioned bits. So before student gets familiar enough with mathematical formalism to think about detectors in superposition, Copenhagen is already occupying slot for Interpretation.
I am going to try and provide short answer, as I see it. (Fighting urge to write about different levels of “physical reality”.)
Many Words is an Interpretation. An interpretation should translate from mathematical formalism towards practical algorithms, but MWI does not go all the way. Namely, it does not specify the quantum state an Agent should use for computation. One possible state agrees with “Schroedinger’s experiment was definitely set up and started”, another state implies “cat definitely turned out to be alive”, but those certainties cannot occur simultaneously.
Bayesian inference in non-quantum physics also changes (probabilistic) state, but we can interpret it as a mere change of our beliefs, and not a change in the physical system. But in quantum mechanics, upon observation, the “objective” state fitting our knowledge changes. MWI says “fitting our knowledge” is not a good criterion of choosing quantum state to compute with (because no state can be fitting enough, as example with Shroedinger’s cat shows) and we should compute with superposition of Agents. MWI may be more “objectively correct”, but it does not seem to be more “practical” than Copenhagen interpretation. So physicists do like to cautiously agree with MWI, then wave hands, proclaim “Decoherence!” and at the end use Copenhagen interpretation as before.
Introductory books emphasize experiments, and experimental results do not come in form of superpositioned bits. So before student gets familiar enough with mathematical formalism to think about detectors in superposition, Copenhagen is already occupying slot for Interpretation.