Screening and Environmental Selection

Screening makes the observable strength of an additional gravitational field depend on an object and its environment. A laboratory source, Solar-System body, dwarf galaxy, and neutron star can therefore realize different effective charges even when the microscopic action is the same. Cross-regime synthesis must classify those states before looking at the gravity residuals; otherwise environmental selection can be tuned to manufacture agreement or tension.

Required background. Probability and Inference for Gravity supplies probability and statistical inference, latent-class models, selection effects, and uncertainty propagation.

Helpful background. Weak Metric Perturbations and Gauge Freedom supplies weak-field gravity and potentials. The Tests of Gravity overview supplies the regime map. Gravitational-Wave Astronomy supplies gravitational-wave data analysis and inference for compact sources.

Effective charge is a solved quantity

For a conformally coupled scalar, write the matter coupling as $A(\phi)$ and define $\beta=M_{\rm Pl}\,d\ln A/d\phi$. In physical spatial coordinates, the quasistatic equation for nonrelativistic matter with conserved Einstein-frame density $\rho$ contains the source $\rho A_{,\phi}$. In the weak-coupling regime $A\simeq1$, it reduces to

$$\nabla^2\phi\simeq V_{,\phi}(\phi)+\frac{\beta(\phi)}{M_{\rm Pl}}\rho.$$

The reduced Planck mass is defined by $M_{\rm Pl}^{-2}=8\pi G$. The solution depends on the density profile, object size, material or halo boundary, and asymptotic field. A useful effective force between well-separated bodies may be written

$$F_\phi\simeq 2\beta^2 Q_1Q_2 F_N(1+mr)e^{-mr},$$

where $Q_A$ are dimensionless effective charges, $\beta$ is the coupling evaluated in the common ambient field, and the displayed $(1+mr)e^{-mr}$ factor is the point-source force response of a mediator with ambient mass $m$. In an unscreened linear limit $Q_A$ can approach unity; in a thin-shell regime it can be much smaller. The charge and the applicability of the point-source form are derived from the nonlinear boundary-value problem, not assigned from the experiment’s residual.

Different mechanisms use different diagnostics. Chameleon and symmetron screening respond to density, size, potential depth, and ambient field. Vainshtein screening is controlled by derivative nonlinearities and source geometry. Kinetic screening has its own gradient and hyperbolicity conditions. A universal threshold in one proxy, such as Newtonian potential, can be a useful classifier inside a calibrated theory but is not a law spanning all mechanisms Burrage and Sakstein 2018, §§2–5.

Strongly self-gravitating bodies add sensitivities, spontaneous scalarization, equation-of-state dependence, and possibly disconnected solution branches. A weak-field “screened” label cannot be copied to a neutron star without solving its relativistic structure and checking stability.

Classify environments before measuring the outcome

Environmental inputs include local and external density, gravitational potential and acceleration, halo membership, distance to massive neighbors, formation history, baryonic mass distribution, and cosmological boundary values. These are inferred from imperfect surveys and simulations. Their uncertainty should produce a probability of each screening state, not a deterministic label chosen at a convenient threshold.

Selection must be blind to the tested residual. For example, choosing galaxies that look anomalous and then calling them “unscreened” because of a noisy environmental proxy biases the effect upward. A defensible design fixes the theory, reconstructs the environment from independent data, sets selection and quality cuts, predicts $Q_A$, and only then examines the gravity observable.

The likelihood can retain a latent screening state $z_j$,

$$p(d_j\mid\alpha,\eta_j,\mathcal E_j) =\sum_{z_j}p(d_j\mid\alpha,z_j,\eta_j,\mathcal E_j) p(z_j\mid\mathcal E_j,\alpha,\eta_j),$$

where $\alpha$ denotes common theory parameters and $\mathcal E_j$ is the measured environment. Nuisance parameters $\eta_j$ must also be integrated out or fitted; they do not disappear by summing over $z_j$. The sum permits more than two discrete charge classes, but a continuously varying partial charge requires an integral over that charge or sufficiently resolved quadrature. If the environment estimate and the gravity observable reuse data, replace this product by their joint sampling model.

Results and limitations through 2026-08-19

At the 2026-08-19 cutoff, screening remains a theoretically developed way to reconcile scale-dependent departures with stringent local tests, but its cross-regime observational assessment is mechanism- and environment-dependent. No single screening proxy has established a universal empirical classification across laboratories, galaxies, cosmology, and compact bodies.

Evidence or result	Status at the cutoff	Limitation or uncertainty boundary
Nonlinear field equations yield screened and unscreened analytic limits	Well established for specified actions and symmetric configurations	Real geometry, boundary data, branch stability, and effective cutoff can change the response
Laboratory and Solar-System null tests restrict unscreened or partially screened forces	Strong within named apparatus and local models	Chamber walls, ambient field, material response, and extrapolation to astrophysical environments matter
Galactic environment comparisons offer large lever arms	Potentially discriminating matched samples exist	Baryonic mass, halo reconstruction, distance, neighbors, and outcome-dependent selection remain limiting
Compact binaries test strong-field charges and radiation	Distinct regime with powerful timing and wave observables	Equation of state, scalarization, cosmological boundary value, and kinematic corrections enter the map

One competing position interprets a residual correlated with an “unscreened” proxy as evidence for a fifth force. The conservative position attributes it to baryonic modeling, external-field reconstruction, catalog selection, or a proxy that does not solve the relevant theory. Negative evidence includes disappearance under a predeclared alternative environment map, absence of the predicted composition or distance scaling, and failure of the same theory parameters in a second regime.

A discriminating matched-environment test

An adversarial design constructs matched pairs that agree in observable nuisance properties but differ in independently reconstructed screening probability. The field profile is solved for every object using the same action and boundary model. The analysis predicts both a primary residual and a negative-control observable that screening should not affect.

The independent check is a simulated population with known environments and solved charges. Apply the full survey mask and classification pipeline, then verify coverage for common theory parameters and quantify bias from misclassification. Vary halo finder, baryonic model, map resolution, and threshold before unblinding. A result that survives only one environment catalog or a threshold selected after seeing residuals is not evidence for screening.

Cross-regime consistency requires the same microscopic parameters to predict laboratory thin shells, Solar-System response, galaxy charges, and compact-body sensitivities within their respective domains. Freire and Wex show why strong-field pulsar charges and damping cannot be inferred from weak-field coefficients alone Freire and Wex 2024, §§4–5.

The unresolved issue is model-dependent: environmental observations can be extremely informative, but uncertainty in nonlinear matching and astrophysical reconstruction can dominate nominal statistical precision. More objects do not fix a biased classifier.

What could change the conclusion. The interpretation could change when a new environment catalog or simulation materially changes screening classifications, when a stable nonlinear solution opens or closes a theory window, or when a blinded matched-sample study passes the solved-profile and misclassification checks above.

Check your understanding

Why is it unsafe to combine two experiments by assuming $Q=0$ for every deep-potential object and $Q=1$ for every shallow-potential object?

Solution

That binary rule assumes a particular mechanism, threshold, geometry, and boundary value. Real objects can be partially screened, an external environment can screen a shallow object, and different mechanisms do not share one potential criterion. Therefore the rule can misclassify systems and bias the common coupling, especially if the threshold was chosen after viewing residuals. A valid analysis solves or emulates the specified field equation, propagates uncertain environmental inputs into a distribution for $Q$, and tests the classifier on simulations. Analytic screened and unscreened limits provide independent checks on the numerical map.

References

• Burrage, C., and Sakstein, J. (2018), “Tests of chameleon gravity,” Living Reviews in Relativity 21, 1. Burrage and Sakstein 2018, §§2–5.
• Freire, P. C. C., and Wex, N. (2024), “Gravity experiments with radio pulsars,” Living Reviews in Relativity 27, 5. Freire and Wex 2024, §§4–5.

Further reading

• Will, C. M. (2014), “The confrontation between general relativity and experiment,” Living Reviews in Relativity 17, 4. Will 2014, §§3–4.