The distinction between calibration and prediction sits at the heart of how this study should be read, and it is worth unpacking because it is a discipline that much of the modified-gravity literature has historically lacked. When a phenomenological model is fitted to a sample and then evaluated on the same sample, the resulting goodness-of-fit statistics carry an optimistic bias whose size grows with the number of adjustable parameters and shrinks with the number of independent data points. The authors address this by keeping the galaxy-side predictor entirely locked: the stellar mass-to-light ratios are fixed in advance, no per-galaxy halo or response amplitude is tuned, and the acceleration scale is set by cosmological quantities rather than fitted to the disk data. This is why the comparison with MOND is meaningful at all. Both prescriptions are evaluated under the same baryonic inputs and the same frozen assumptions, so the paired root-mean-square residuals measure the relative merit of two fixed formulas rather than the relative merit of two fitting procedures. The absence of a statistically significant difference between the two is therefore an honest null result, not a failure of execution, and it places Σ-gravity in the same empirical territory that MOND has occupied for decades: a one-line algebraic rule that captures a large fraction of disk-galaxy dynamics without per-system freedom.

The ablation of the bounded factor deserves particular attention because it isolates the one genuinely novel ingredient of the proposal. Removing the bounded factor and retaining only the acceleration-dependent response degrades the mean velocity residual by roughly half a kilometre per second, a small but non-zero margin whose bootstrap interval excludes zero. In galaxy-scale fitting, where typical observational uncertainties and mass-to-light systematics are of comparable or larger size, an improvement of this magnitude is best described as a measurable but modest contribution. The authors are explicit that this improvement cannot be taken as evidence for the coherence hypothesis itself, because the factor is constructed endogenously from the model-predicted velocities rather than from any independently measured property of the stellar or gas motions. This is a subtle but important epistemic point: a term that is derived from the model’s own output will, by construction, correlate with the model’s successes, and only an external, independently measured variable can break that circularity. The proposed resolution, testing whether independently measured phase-space order predicts the response on held-out systems, is precisely the kind of falsifiable follow-up that separates a motivating hypothesis from a fitted fudge factor.

The cluster result is arguably the most instructive part of the paper precisely because it is a negative one. Strong gravitational lensing probes the projected mass distribution of clusters at radii and in regimes where disk-galaxy intuition does not automatically transfer, and the baryonic mass budget of a cluster is dominated by hot intracluster gas rather than stars, which changes the character of any baryon-based prescription. Calibrating a single effective amplitude on forty-two strong-lensing systems and then applying it without refitting to seventeen CLASH clusters yields a median predicted-to-observed ratio well above unity, meaning the prescription systematically underpredicts the lensing mass on average, and the discrepancy varies with radius rather than behaving as a simple overall offset. A radius-dependent bias of this kind cannot be repaired by rescaling the amplitude; it indicates that the functional form itself, including the assumption of a single nominal path length for every cluster, is missing physics that differs from the disk case. The authors’ decision to present this as a falsifiable limitation rather than to introduce cluster-specific parameters is methodologically commendable, because adding per-cluster freedom would have converted a genuine transfer failure into an apparent success while telling us nothing new about nature.

The auxiliary QUMOND construction provides a theoretical bridge that clarifies what kind of underlying structure the empirical response could have. QUMOND, the quasi-linear formulation of modified Newtonian dynamics, replaces the Poisson equation with a modified one in which the Newtonian potential feeds back through an algebraic function of its own gradient magnitude. The finding that an action of this type can reproduce the Σ-gravity response when the parameter B is prescribed as a spatially constant quantity, but not when it is generated endogenously from the galaxy model, tells us where the theoretical difficulty lies. Constant-B limits are the easy case for any modified-Poisson framework, since they reduce to a fixed algebraic modification of the field equation. A spatially and temporally varying B that depends on the dynamical state of the baryons is a far stronger requirement, and no existing covariant formulation supplies it without additional structure. This gap is not a fatal objection at the phenomenological stage, but it defines the boundary between what has been demonstrated and what remains aspirational.

It is also worth situating the acceleration scale used in the response function. The quantity g† is built from the speed of light and the Hubble constant, echoing a long-standing observation that the characteristic acceleration below which disk galaxies deviate from Newtonian expectations is numerically close to cH0 divided by a factor of order ten. Any successful theory of this regime must ultimately explain why a cosmological parameter appears in galaxy-scale dynamics, whether through a genuine physical coupling or through a coincidence of structure formation. The baryonic Tully–Fisher relation, which connects the flat asymptotic speed of a disk to the fourth root of its baryonic mass with remarkably small scatter, constrains any candidate response function in its deep-acceleration limit, and the square-root structure of the transition function in Σ-gravity is chosen to respect that constraint. The authors acknowledge that the precise shape of the transition between high and low acceleration is a design choice rather than a derivation, which is the correct posture for an empirical proposal at this stage.

For readers weighing these results against the broader landscape, several contextual points are useful. First, the standard cosmological model rests on evidence far beyond galaxy rotation curves, including the acoustic peak structure of the cosmic microwave background, the growth rate of large-scale structure, and the inferred mass-to-light ratios of galaxy clusters through independent channels. Any alternative phenomenology that aspires to more than galaxy-scale bookkeeping must eventually engage that full body of evidence, and the authors correctly decline to claim otherwise. Second, the residual mass discrepancy in galaxy clusters is a known and persistent challenge for MOND as well, typically requiring either additional unseen matter or cluster-specific modifications, so the transfer failure documented here is consistent with a difficulty that predates this proposal. Third, the exclusion of bulge-dominated radial measurements from the SPARC analysis, while defensible on the grounds that the disk prescription is not expected to describe bulge kinematics, narrows the tested domain in ways that should be kept in mind when comparing headline statistics across the literature.

The matched counter-rotation diagnostic and the MaNGA integral-field comparison serve as secondary checks that broaden the empirical footprint beyond the SPARC sample, and the structural sensitivity test on the assumed photometric scale lengths probes how much of the agreement depends on the input baryon models rather than the response function itself. That the scale-length test does not outperform simple mass-to-light assignment controls is a sobering reminder that in disk-galaxy phenomenology, uncertainties in the stellar population synthesis inputs can rival the differences between competing prescriptions. This is not a reason for despair but a calibration of expectations: distinguishing between closely related algebraic responses requires either tighter photometric constraints or samples selected to maximize the discriminating power of the comparison.

Looking forward, the most consequential test the paper identifies is whether an independently measured coherence variable, drawn from the observed phase-space structure of stellar and gas motions, can predict the response amplitude on galaxies not used in any calibration. If such a variable exists and carries predictive power beyond the acceleration alone, the coherence hypothesis graduates from motivation to measurement. If it does not, the bounded factor joins the long list of empirically motivated corrections that fit well but explain little. Either outcome would be informative, and the framing of the paper ensures that the community can adjudicate it with data rather than rhetoric. In the meantime, the study stands as a careful, self-limiting contribution: a fixed-form galaxy response whose aggregate performance is statistically indistinguishable from the established MOND prescription under identical assumptions, a cluster amplitude that demonstrably fails to transfer, and a clearly marked boundary between what has been tested and what remains hypothesis.

Subject of Research: Σ-Gravity: a coherence-motivated empirical response tested in galaxies and clusters

Article Title: Σ-Gravity: a coherence-motivated empirical response tested in galaxies and clusters

Article References: Speiser, L. (2026). Σ-Gravity: a coherence-motivated empirical response tested in galaxies and clusters. Frontiers in Astronomy and Space Sciences, 13, Article 1866133. https://doi.org/10.3389/fspas.2026.1866133

Image Credits: AI Generated

DOI: 10.3389/fspas.2026.1866133

Keywords: Gravity, coherence-motivated, empirical, response, tested, galaxies, clusters, scientific research