Selective Operations, Postselection, and State Update

Postselection replaces a deterministic channel by a subnormalized branch and a classical record. It may reveal strong conditional correlations, but it cannot be used for signaling or resource claims unless the record is accessible, its communication is included, and the success probability remains in the accounting.

Required background. Local measurement instruments defines selective and nonselective operations.

Helpful background. Hypothesis testing and asymptotics supplies operational error criteria for comparing rare conditional ensembles.

Subnormalized branches

For outcome $x$, let $\mathcal I_x$ be completely positive and trace nonincreasing. Then

$$p_x=\operatorname{tr}\mathcal I_x(\rho), \qquad \rho_x=\frac{\mathcal I_x(\rho)}{p_x},$$

when $p_x>0$. The nonselective channel is $\mathcal E=\sum_x\mathcal I_x$. For a remote observer $B$ who has not received $x$, the relevant state is the unconditional marginal of $\mathcal E(\rho)$, not $\rho_x$.

This distinction is particularly important when $A$ and $B$ are spacelike. Local conditioning at $A$ changes the mathematical state used to predict the subensemble labeled by $x$, but it does not make that label available at $B$. Only after an ordinary classical message reaches the joint future can the parties sort and use the conditional ensemble.

Selective operations add a record branch to the protocol. The availability and transport of that record are part of the causal circuit, not an afterthought. The diagram is schematic.

A two-region calculation

Let $\mathcal I_x^A$ be a local instrument in region $A$ and let $E_y^B$ be an effect in spacelike region $B$. The joint probability is

$$p(x,y)=\operatorname{tr}\!\left[ E_y^B\,\mathcal I_x^A(\rho) \right].$$

The locally conditioned distribution is

$$p(y\mid x)=\frac{p(x,y)}{p(x)},$$

which can depend on $x$ because the initial field state is correlated. The distribution accessible at $B$ before communication is

$$p_B(y)=\sum_xp(x,y) =\operatorname{tr}\!\left[E_y^B\,\mathcal E_A(\rho)\right].$$

For a properly localized spacelike intervention, $p_B(y)$ is unchanged when $A$ changes its local trace-preserving instrument. Conditional dependence and no signaling are therefore compatible. The causal composition theorem for probe-induced instruments is given by Fewster and Verch 2020, §§ 3–5; Bostelmann, Fewster, and Ruep 2021, §§ II–IV explain why an idealized update that violates this conclusion cannot be assigned the advertised localized apparatus.

Vanishing success probability

Suppose a postselected quantity $G_x$ grows as $p_x\to0$. The operational per-trial yield is not $G_x$ alone. Depending on the task it is bounded by a quantity such as $p_xG_x$, a rate that includes repetitions, or a confidence interval whose sample size scales with the number of successes. Claims of “arbitrarily large amplification” must specify which averaged cost or risk remains finite.

Postselection also changes fair comparisons. If one implementation discards failures while another reports all trials, compare them at equal total preparations, energy, duration, and classical communication. Heralding can be useful, but it is a resource trade, not a free deterministic channel.

Postselection bias enters the inference branch. A remote conditional change becomes operational only after the outcome record is communicated through the causal future. The map is schematic.

Check your understanding

Can a spacelike observer infer which measurement setting was used at $A$ by conditioning on $A$‘s outcome label if that label has not arrived?

Solution

No. Conditioning on an unavailable variable is an analysis performed by an observer who possesses the joint record. The local observer at $B$ must average over $A$‘s outcomes, and a localized trace-preserving intervention leaves that marginal unchanged. After classical communication, the conditional statistics can be studied in the joint future without violating causality.

References

• Bostelmann, H., Fewster, C. J., and Ruep, M. H. (2021). “Impossible Measurements Require Impossible Apparatus.” Physical Review D 103, 025017. DOI. Open PDF.
• Fewster, C. J., and Verch, R. (2020). “Quantum Fields and Local Measurements.” Communications in Mathematical Physics 378, 851–889. DOI. Open PDF.