Ph · Physics

What Are Retrocausal Interpretations of Quantum Mechanics?

Published By the 9to5Science editorial team

At a glance

  1. 1. Define the idea

    A future measurement setting may help determine earlier physical conditions in a quantum model.

  2. 2. Separate cause from signal

    Retrocausal influence does not mean an experimenter can send a controllable message into the past.

  3. 3. Connect to Bell’s theorem

    Some models relax the assumption that hidden variables are independent of later measurement settings.

  4. 4. Check the evidence

    Bell-test results constrain theories, but do not by themselves establish retrocausality.

Advertisement

A retrocausal interpretation of quantum mechanics allows a later measurement condition to help determine an earlier physical state or hidden variable. The name sounds like time travel, but the proposal is usually about how to describe correlations within a physical theory—not about sending a message to yesterday.

There is no single retrocausal interpretation. Different proposals add different physical pictures to the mathematics of quantum mechanics. Here is what the shared idea means, how it relates to Bell’s theorem, and what experiments do—and do not—tell us.

What “retrocausal” means

In an ordinary forward-causal picture, an earlier physical condition affects later events. A retrocausal model also allows a later boundary condition, such as a detector’s measurement setting, to constrain or influence an earlier variable. One way to express this is that a hidden variable λ\lambda at the source can depend on settings chosen later, represented by aa and bb:

P(λ∣a,b)≠P(λ).P(\lambda\mid a,b)\ne P(\lambda).

This does not necessarily mean that a controllable force travels backward through time. Some models instead treat the entire experiment as a boundary-value problem: conditions at both the beginning and end help determine a consistent history. The mathematical structure may be time-symmetric even if the everyday experience of cause and effect is not.

Why the idea comes up in Bell experiments

Bell’s theorem shows that no theory satisfying a set of assumptions can reproduce all the predictions of quantum mechanics. In common Bell-test analyses, an important assumption is measurement independence: the hidden state emitted by a source is statistically independent of the later choices of measurement settings. In symbols, this is P(λ∣a,b)=P(λ)P(\lambda\mid a,b)=P(\lambda).

A retrocausal model can reject that assumption. The later settings may help determine the source’s earlier hidden state, while physical influences within the model remain local in spacetime. This is one possible strategy for accounting for Bell correlations; it is not the only one. Other approaches reject or revise different assumptions, and a retrocausal account must still reproduce the observed quantum statistics without enabling faster-than-light or backward-in-time signaling.

A concrete Bell-test example

For a pair of spin-1/2 particles in a singlet state, quantum mechanics predicts that measurements along directions separated by angle θ\theta have correlation

E(θ)=−cos⁡θ.E(\theta)=-\cos\theta.

Consider four measurement directions in one plane: Alice chooses a0=0∘a_0=0^\circ or a1=90∘a_1=90^\circ, and Bob chooses b0=45∘b_0=45^\circ or b1=−45∘b_1=-45^\circ. The predicted correlations are

E(a0,b0)=−cos⁡45∘=−12,E(a0,b1)=−cos⁡45∘=−12,E(a1,b0)=−cos⁡45∘=−12,E(a1,b1)=−cos⁡135∘=+12.\begin{aligned} E(a_0,b_0)&=-\cos 45^\circ=-\frac{1}{\sqrt{2}},\\ E(a_0,b_1)&=-\cos 45^\circ=-\frac{1}{\sqrt{2}},\\ E(a_1,b_0)&=-\cos 45^\circ=-\frac{1}{\sqrt{2}},\\ E(a_1,b_1)&=-\cos 135^\circ=+\frac{1}{\sqrt{2}}. \end{aligned}

The CHSH combination therefore has magnitude

∣E(a0,b0)+E(a0,b1)+E(a1,b0)−E(a1,b1)∣=22.|E(a_0,b_0)+E(a_0,b_1)+E(a_1,b_0)-E(a_1,b_1)|=2\sqrt{2}.

This exceeds the Bell-inequality limit of 22 for theories satisfying the relevant assumptions. A retrocausal account may explain the correlations by allowing the later choices of aa and bb to be related to earlier source variables. But the calculation itself does not demonstrate retrocausality: it is a quantum prediction, and interpretations differ over what physical story, if any, to attach to it.

Examples of retrocausal approaches

The transactional interpretation, associated with John Cramer, describes quantum processes using offer waves and confirmation waves, with a transaction linking emission and absorption. Its use of advanced, time-reversed solutions gives it a retrocausal flavor, though its terminology and ontology are specific to that interpretation.

The two-state-vector formalism uses both a state prepared in the past and a state associated with a later measurement to describe a system between them. It is a time-symmetric formalism, but using it does not by itself commit someone to the claim that the future physically causes the past. The distinction between a useful mathematical description and an interpretation of physical causation matters.

Other researchers have developed retrocausal models to explore whether quantum correlations can be explained with spacetime-local interactions. These models are not all equivalent, and “retrocausal” does not name one agreed-upon theory.

Retrocausality is not backward signaling

A theory can allow later settings to constrain earlier variables yet prevent an experimenter from using those variables to transmit a chosen message into the past. The observable probabilities must still obey the no-signaling conditions: changing one observer’s measurement choice cannot change the other observer’s locally observed statistics in a way that communicates information.

This is a key distinction. Correlations between distant outcomes are not the same as a controllable signal, and a model’s hidden-variable story is not directly observable just because it describes those correlations.

Common mistakes

  • Treating Bell’s theorem as proof of retrocausality. Bell tests rule out broad classes of models under stated assumptions. They do not select retrocausality over every alternative.
  • Equating time symmetry with backward causation. Equations or formalisms that treat past and future symmetrically need not assert that the future causes the past.
  • Assuming an experiment can send a message backward in time. Retrocausal interpretations generally do not provide a usable signaling protocol.
  • Assuming all retrocausal models make the same predictions. Some are interpretations of standard quantum mechanics; others propose specific models. Their empirical claims need to be assessed individually.

What experiments can establish

Bell experiments test statistical predictions and constrain theories that try to explain them. If a retrocausal model makes exactly the same predictions as standard quantum mechanics, those experiments cannot distinguish it from other interpretations. A model that predicts a measurable deviation can be tested, but the deviation—not the label “retrocausal”—is what an experiment would assess.

For now, retrocausality is one proposed way to think about quantum foundations, especially the assumptions behind Bell’s theorem. It is not an experimentally established mechanism for sending influence or information into the past.

References

  1. Colloquium: Bell’s theorem and locally mediated reformulations of quantum mechanics — William C. Davidon and William G. Unruh?
  2. The Transactional Interpretation of Quantum Mechanics — John G. Cramer
  3. Time’s Arrow and Archimedes’ Point — Huw Price

Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.