Interpretations of Quantum Mechanics · Chapter 08
Who or what is a quantum state a state for?
Starting point: Chapters 1–7; familiarity with density operators and linear algebra.
Anchor example: A friend records a qubit while another system describes the combined laboratory.
Method: Calculate a concrete experiment, identify the assumptions, then compare what different accounts explain.
Watch for: Relational quantum mechanics and QBism differ about events, agents, and the role of probabilities.
This lesson includes explanations, worked examples, and problems with solutions. Advanced results whose proofs are outside the chapter are identified explicitly.
Who or what is a quantum state a state for?
A friend interacts with a qubit and records zero or one. An outside system can still use \((|0,F_0\rangle+|1,F_1\rangle)/\sqrt2\) for the combined laboratory if the interaction remains coherent. Relational quantum mechanics treats values and events as relative to interacting systems, rather than demanding one observer-independent quantum state that simultaneously describes all facts in the same way.
“Observer” here need not mean a conscious person. A physical system can participate in an interaction relative to which another system has a value. That differs from a claim that minds create matter. The conceptual work lies in specifying relations between events and in explaining comparisons of records, not in assigning a privileged mental role to a detector.
If the outside system later reads the friend's Z record, the interaction correlates their records. Conditional on the friend's zero record, an ideal subsequent Z readout agrees with it. If instead the outside system coherently measures an incompatible observable of the whole laboratory, the experiment is different.
The formal state calculation does not by itself settle the philosophical status of relative facts. It does show why apparently inconsistent descriptions should be indexed by the systems and interactions involved before being combined. “Relative” must constrain an account, not merely excuse any contradiction.
Worked check 8.1. If W reads the friend’s Z record ideally, what is the probability of mismatching records?
Zero in the ideal model: the joint state contains only matching |F₀,W₀⟩ and |F₁,W₁⟩ components.
QBism assigns a different role to the quantum state: it expresses an agent's personal probabilities for the consequences of that agent's actions on the world. Even a probability-one assignment is a commitment in the agent's expectations, not automatically an assertion that the outcome was a pre-existing property. The Born rule imposes a specifically quantum constraint on those commitments.
For a qubit, an agent's state can be written \(\rho=(I+\mathbf r\cdot\boldsymbol\sigma)/2\), with |r|≤1. If the intended measurement is spin along a unit direction n,
\[p(+|\mathbf n)=\frac{1+\mathbf r\cdot\mathbf n}{2}.\]These probabilities are not arbitrary. Positivity and the Born rule relate expectations for different possible actions. For example, assigning probability one to both X+ and Z+ would require rₓ=r_z=1, violating |r|≤1. Personal probability does not mean any collection of numbers is acceptable.
This perspective treats an outcome as an experience for an agent and an action as an engagement with an external world. It is not simply a hidden-variable model in which a state represents ignorance about an underlying λ. Its distinctive explanatory commitments should be assessed on their own terms, including how it accounts for intersubjective scientific practice.
Worked check 8.2. Can a valid qubit state assign p(X+)=p(Z+)=0.9?
Those assignments require rₓ=r_z=0.8, so |r|²≥1.28>1. No positive qubit density operator satisfies both.
Both relational and QBist approaches resist treating every state assignment as an observer-independent physical wave. That common feature should not erase their differences. Relational mechanics concerns physical interactions and relative events between systems. QBism places agents, actions, and personal probability at the center of its account. An inanimate subsystem as relational observer is not automatically a QBist agent.
Imagine two people who assign |0⟩ and |+⟩ to a qubit. Their Z predictions are one and one half. This is not an algebraic contradiction: the assignments may reflect different preparation information or different conditions. If they claim identical preparation information and agree on the same operational procedure, however, their differing assignments become a substantive question to resolve.
A record comparison is itself a physical interaction. For an ideal correlated state \(\sum_k c_k|k,F_k\rangle\), an external Z readout can append |Wₖ⟩, giving \(\sum_k c_k|k,F_k,W_k\rangle\). There is no term with mismatched F and W records. This calculation establishes agreement for that readout, while leaving the account of global or relative actuality to the interpretation.
Be equally careful about locality claims. A reinterpretation of what probabilities mean does not change the numerical CHSH value. Distinguish a philosophical account of locality from the precise factorization condition proved impossible for the stated ontological model.
Worked check 8.3. Can different quantum-state assignments be compared without specifying the preparation information?
Not reliably. First identify whose probabilities or relations are involved, what information is conditioned on, and which future action is being predicted.
These exercises draw on the whole chapter. Attempt each question before opening a hint, explain your reasoning, then compare it with the worked solution.
Does relational observer mean conscious mind?
Use the physical-interaction definition.
No. A system participating in an interaction can serve as a relational reference; consciousness is not required by that usage.
Write the outside state after an ideal friend measurement of |+⟩.
Apply linearity to the record coupling.
It is (|0,F₀⟩+|1,F₁⟩)/√2 if the full laboratory remains coherent.
Why is reading the friend’s record different from measuring interference of the laboratory?
Compare the measured projectors.
The record measurement distinguishes |0,F₀⟩ and |1,F₁⟩. An interference measurement projects onto their superpositions; it is an incompatible procedure.
Find p(X+) and p(Z+) for r=(0.6,0,0.8).
Check the Bloch length and use the formula.
The state is pure because |r|=1. The probabilities are 0.8 and 0.9.
If p(Z+)=1, what is p(X+) for a valid qubit state?
Use positivity.
r_z=1 forces r_x=r_y=0, so p(X+)=1/2.
Why is QBism not merely unrestricted subjective guessing?
Identify the constraints on assignments.
Probabilities must satisfy coherence conditions and quantum constraints such as positivity and the Born rule linking different actions.
Give one important distinction between QBism and relational quantum mechanics.
Compare the central entities in each account.
QBism centers an agent’s actions, experiences, and personal probabilities; relational mechanics centers physical events relative to interacting systems.
Does personal probability automatically imply a ψ-epistemic ontological model?
Compare probability commitments with distributions over λ.
No. A ψ-epistemic ontological model presupposes an underlying state space and overlapping preparation distributions. QBism need not adopt that model structure.
What can the ideal agreement calculation establish?
Read the joint record amplitudes.
It establishes zero probability of mismatched records for the specified comparison. It does not by itself establish an observer-independent global event ontology.
How should a claim that an interpretation is local be assessed?
Separate operational and ontological meanings.
Ask whether it means no controllable signalling, Bell-local factorization, or a different account of physical events. The first does not imply the second, and CHSH statistics remain unchanged.