What Is Classical Reality

Classical reality in quantum foundations usually means a collection of effective features: stable records, approximately compatible observables, robust trajectories and agreement among observers who inspect the same environment. These features need not appear together or become exact at a single universal scale. Asking which one a calculation explains is more useful than asking whether an object is simply quantum or classical.

Different criteria for classical behavior

A macroscopic body can have a sharply localized center of mass while its internal degrees of freedom require quantum statistics. A laser field can have nearly classical amplitude dynamics while serving as a quantum phase reference. A superconducting circuit can be large in particle number yet support coherent superpositions of selected collective variables.

Thus size alone is not a definition of classicality. Specify the variables, accessible observables, uncertainty scale, environment and observation time. A classical approximation is a claim about predictions within that restricted regime.

Coarse records can become compatible

Imagine a position detector that resolves only bins of width $\Delta x$. Its alternatives are spectral projectors onto disjoint intervals and hence commute. A classical probability distribution over those recorded bins is then meaningful for that measurement. Describing motion through a sequence of bins additionally requires control of temporal disturbance and interference between alternative histories.

For a wave packet with center $x_c$ and spread $\sigma_x\ll\Delta x$, most probability can lie in one bin away from its boundaries. This is an approximate definite record at the chosen resolution. It does not imply that momentum is simultaneously sharp or that finer coherent measurements are forbidden.

A quantitative bound makes the needed distance explicit. Let $x_c=\langle X\rangle$ lie inside the reported bin $C=[a,b]$, define $d=\min(x_c-a,b-x_c)>0$, and assume the position distribution has finite variance $\sigma_x^2$. Outside the bin, $|X-x_c|\geq d$, so averaging the inequality $\mathbf 1_{X\notin C}\leq (X-x_c)^2/d^2$ gives

$$p(X\notin C)\leq\min\left(1,\frac{\sigma_x^2}{d^2}\right).$$

This is Chebyshev’s bound applied to the Born distribution for the specified position measurement. It requires no Gaussian shape. A small width compared with the bin size becomes a useful error bound only when the mean is also far enough from either edge.

The classicality chapter separates decoherence, coarse-graining, trajectories and records. The Ehrenfest/WKB bridge explains when those approximations are reliable.

Shared records and objectivity

Several observers can agree about a pointer because its information has been redundantly copied into accessible environmental fragments. Their agreement is a physical correlation, not merely a convention. Which information is robustly available depends on the interaction and noise; arbitrary incompatible state information cannot be broadcast to everyone.

Quantum Darwinism investigates this redundancy. It must be distinguished from a unique-outcome postulate: many correlated records can exist within each branch of a global unitary state. Agreement among records does not by itself decide which ontology describes that global state.

Classical limits can fail nonuniformly

A short-time narrow-packet approximation can fail after spreading, at caustics, during tunneling or in chaotic dynamics. A decoherence approximation can fail in carefully isolated systems or when environmental information is coherently erased. A convincing classical limit states the controlling parameters and which errors remain small over the time interval of interest.

This approach explains why ordinary macroscopic predictions work while leaving room for controlled mesoscopic interference. It does not require a single fundamental boundary where all quantum effects disappear.

Exercise: a detector-resolution claim

A packet has $\sigma_x=1\,\mu\mathrm m$ and a detector has $\Delta x=1\,\mathrm{mm}$. Is its recorded position necessarily definite? What additional geometric fact matters? If its mean is at the midpoint of a bin, bound the probability of recording a different bin without assuming a Gaussian shape.

Solution

The small ratio $\sigma_x/\Delta x=10^{-3}$ supports concentration in one bin only if the packet center lies sufficiently far from a bin boundary. A packet centered on a boundary can split its probability between two bins despite its small width. Resolution, state width and placement all enter the record claim.

At the midpoint, $d=\Delta x/2=500\,\mu\mathrm m$. Hence

$$p(X\notin C)\leq\left(\frac{1}{500}\right)^2=4\times10^{-6}.$$

This is a bound on the specified measurement’s probabilities. At a bin boundary, $d=0$ and the bound supplies no concentration guarantee.

References

• Schlosshauer, Maximilian, Decoherence and the Quantum-to-Classical Transition, Springer (2007), doi:10.1007/978-3-540-35775-9.
• Zurek, Wojciech H., “Quantum Darwinism,” Nature Physics 5, 181–188 (2009), doi:10.1038/nphys1202.