Parameterized Frameworks and Null Tests

Scrollable formula. Use Left and Right arrow keys to reveal hidden content.

Scrollable table. Use Left and Right arrow keys to reveal additional columns.

Scrollable figure. Use Left and Right arrow keys to inspect hidden content.

Parameterized frameworks expose departures from general relativity in coordinates chosen to match an observational regime. PPN coefficients organize weak, slow, quasistatic gravity; PPK parameters overdetermine relativistic timing; waveform and ringdown coefficients organize radiative strong-field signals; polarization and propagation variables classify how waves travel and couple to detectors; residual tests look for failures outside a chosen basis. The organizing principle is useful precisely because it is narrower than a complete theory.

Helpful background. Weak Metric Perturbations and Gauge Freedom supplies weak-field gravity and radiative perturbations. Probability and Inference for Gravity supplies probability and statistical inference. Relativistic Two-Body Dynamics supplies conservative dynamics, radiation reaction, and compact-binary source modeling.

Enter this chapter

A parameterized test begins with a declared GR point $\boldsymbol\delta=0$ and a forward model

d=s(\theta_{\rm source},\boldsymbol\delta,\eta)+n.

The deformation coordinates $\boldsymbol\delta$ should be observable in the available data, reduce cleanly to GR, and stay within the domain of the underlying approximation. The nuisance vector $\eta$ carries calibration, environment, selection, and model discrepancy. A complete alternative theory supplies a map $\boldsymbol\delta=F(\alpha)$ , along with correlations, stability conditions, and a validity range. Without that map, a coefficient interval is a null-test result rather than a theory exclusion.

Current-release statements in this chapter use the evidence cutoff 2026-08-19. Stable definitions and derivations do not depend on that date; catalog identities, analysis choices, and frontier assessments do.

Six frameworks in increasing dynamical reach

Parameterized Post-Newtonian Gravity standardizes the weak-field metric and connects its coefficients to light propagation, orbital motion, preferred-frame effects, and conservation laws.
Parameterized Post-Keplerian Timing fits relativistic timing effects phenomenologically, then tests whether several mass-dependent curves intersect under a specified theory.
Parameterized Inspiral Waveforms deforms frequency-dependent phase and amplitude coefficients while tracking mass, spin, waveform, and selection degeneracies.
Parameterized Ringdown Spectra shifts quasinormal frequencies and damping times, with special attention to start time, mode identification, and what present spectroscopy does not yet establish.
Polarization and Propagation Frameworks separates extra radiative modes, dispersion, speed, damping, and birefringence from source generation and detector response.
Residual and Consistency Tests looks beyond a fixed coefficient basis using calibrated residual structure and covariant comparisons across data subsets.

The order moves from controlled weak-field expansions toward more agnostic diagnostics. It is not a ranking of experimental strength. A narrow, theory-connected coefficient can yield a sharper physical conclusion than a flexible residual search, while a residual search can discover a failure no chosen coefficient anticipated.

What each framework holds fixed

Framework	Deformation object	Principal fixed structure	Characteristic failure boundary
PPN	Weak-field metric coefficients	Slow-motion near-zone potentials and gauge	Screening, strong self-gravity, radiation, cosmology
PPK	Relativistic timing parameters	Timing formula and binary kinematics	Theory-dependent masses, body charges, external acceleration
Inspiral	PN or generic phase/amplitude terms	Baseline waveform and detector likelihood	Degeneracy, missing source physics, merger extrapolation
Ringdown	Complex mode frequencies	Damped-mode basis and remnant mapping	Start time, nonlinear merger, weak mode identification
Polarization/propagation	Mode content and travel law	Source waveform or phenomenological emission	Network rank, intrinsic delay, cosmology, calibration
Residual/consistency	Unexplained structure or parameter difference	Split rule, signal family, noise model	Trials, shared covariance, flexible residual basis

This table is a compact antidote to a common mistake: calling a test “model independent.” Every row fixes a signal representation and a likelihood. The useful question is not whether assumptions exist, but whether they are explicit, testable, and broad enough for the scientific claim.

From bottom-up deformations to top-down theories

A bottom-up framework varies observable coefficients with minimal theory commitment. It is well suited to anomaly detection and comparisons across analyses. A top-down test starts from an action, derives source solutions and propagation, and predicts a correlated location in the framework. The two approaches should meet in the middle.

The mapping may fail to be one-to-one. One theory can occupy different framework points in different environments, and several theories can share a leading coefficient. It may also be singular near the GR limit or restricted by stability. Transforming published one-dimensional posteriors coefficient by coefficient then loses the joint geometry. Whenever possible, evaluate the original likelihood with the complete correlated map.

No single framework spans all scales. PPN is not a merger model; a waveform-dispersion coefficient is not a cosmological perturbation theory; a ringdown shift is not a remnant solution. Cross-regime synthesis belongs only after each local mapping and its covariance are controlled.

Review the chapter

Suppose a theory predicts a PPN shift, a dipole-radiation inspiral term, and a dispersive propagation term, all determined by one coupling $\alpha$ . Explain why multiplying three published one-dimensional limits is generally unsafe and outline the correct analysis.

Solution

Each published limit may use a different prior measure, fix other deformation coefficients, and marginalize shared source or calibration information differently. The theory mapping may also depend on environment, compact-body sensitivities, and redshift, so the same $\alpha$ does not enter as three independent constant shifts. Therefore derive all three observables from the stable theory branch, build or obtain the joint likelihoods with selections and shared nuisance parameters represented once, and evaluate them as functions of $\alpha$ . Comparing the resulting correlated prediction with independent observables provides a stronger check than multiplying marginalized curves.