pith. machine review for the scientific record. sign in

arxiv: 2503.19898 · v2 · submitted 2025-03-25 · 🌌 astro-ph.CO · gr-qc

Recognition: 4 theorem links

· Lean Theorem

Non-minimally coupled gravity constraints from DESI DR2 data

Jiaming Pan , Gen Ye
Authors on Pith no claims yet

Pith reviewed 2026-05-06 20:58 UTC · model claude-opus-4-7

classification 🌌 astro-ph.CO gr-qc
keywords <parameter name="0">dark energy
0
0 comments X

The pith

Combined BAO, supernova, and CMB data show ~3σ evidence that gravity is non-minimally coupled to matter at late times.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors take the recent BAO data, which appears to favour a dark energy whose equation of state crosses the "phantom divide" w = -1, and ask what kind of physics could produce that crossing without running into the theoretical instabilities that doom a single canonical scalar field. Their answer is non-minimal coupling between gravity and matter, encoded in the effective field theory of dark energy by a single function Ω^EFT(a) multiplying the Ricci scalar. Combining BAO with supernovae and CMB data, they reconstruct this function in two ways — bin-by-bin in redshift, and as a Taylor series in the dark-energy fraction Ω_DE(a) — and find that Ω^EFT is non-zero at roughly the 3σ level, with a bump near z ≈ 0.6 that a linear ansatz cannot capture. The takeaway they push is that the data already prefer a modified-gravity ingredient, and that the popular one-parameter form Ω^EFT ∝ Ω_DE is too rigid to see it cleanly.

Core claim

Using DESI DR2 BAO together with DESY5 supernovae and Planck CMB data, the authors claim the first ~3σ detection that the effective-field-theory function describing non-minimal coupling of gravity to matter, Ω^EFT(a), is non-zero. Both a six-node non-parametric reconstruction over 0 < z < 1 and a Taylor expansion Ω^EFT(a) = Σ c_i Ω_DE^i(a) up to quadratic order point to the same conclusion, with the quadratic term needed to capture a peak near z ≈ 0.6 that a single-coefficient linear form misses.

What carries the argument

The effective field theory of dark energy: a quadratic action for cosmological perturbations whose only free time-dependent coefficient kept here is Ω^EFT(τ), the function multiplying the Ricci scalar. Setting Ω^EFT ≠ 0 is equivalent to non-minimal coupling and covers the Horndeski class of scalar-tensor theories; combined with a CPL w_0–w_a background, it lets the authors test whether a stable phantom-crossing cosmology requires modified gravity at the level of the data.

If this is right

  • <parameter name="0">The apparent phantom crossing seen in BAO data can be reproduced inside a stable scalar-tensor theory
  • removing the theoretical objection that a single canonical scalar cannot cross w = -1.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • <parameter name="0">The Δχ² improvement over ΛCDM (~22) is almost entirely absorbed by the w_0–w_a background fit
  • the non-minimal-coupling extension barely improves χ² further
  • so the 3σ evidence is being driven by the stability requirement on perturbations rather than by direct goodness-of-fit — a subtle inferential structure that a reader should weigh carefully.

Load-bearing premise

The whole signal rests on imposing that the dark-energy theory be free of perturbative instabilities while reproducing the apparent phantom-divide crossing seen in BAO+SN+CMB; if that crossing is instead an artefact of dataset tensions or supernova systematics, the case for non-minimal coupling weakens accordingly.

What would settle it

A future analysis that uses late-time perturbation probes — weak lensing and full-shape galaxy clustering from Euclid or LSST — to directly measure the growth and lensing signatures predicted by the reconstructed Ω^EFT(a). If those probes find growth and lensing consistent with GR while BAO+SN+CMB still prefer phantom-like w(z), the non-minimal-coupling explanation is ruled out, and the apparent phantom crossing must come from something else (multi-field DE, dark-sector interactions, or systematics).

read the original abstract

It has been observed that the hint about dynamical dark energy in the DESI BAO observation might point to non-minimally coupled gravity. We report the first $3\sigma$ evidence for non-minimal coupling in a model-agnostic effective field theory (EFT) approach. In a non-parametric reconstruction approach, we detect a clear departure from the General Relativity expectation of non-minimal coupling based on a combined analysis of DESI DR2 BAO together with CMB and Type Ia supernova data. Also, it is found that current data can constrain up to the quadratic order $(n=2)$ if the EFT function representing non-minimal coupling is Taylor expanded as a general function of the dark energy fraction $\Omega_{\rm DE}$, i.e. $\Omega^{\rm EFT}(a)=\sum_{i=0}^{n} c^{\rm EFT}_i \Omega^i_{\rm DE}(a)$. Our findings constitute a detection of modified-gravity effects and call for a more flexible parametrization of the EFT functions than commonly used ones in literature.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

6 major / 8 minor

Summary. The authors analyze DESI DR2 BAO together with DESY5 SNIa and Planck CMB data within the EFT of dark energy framework, holding a CPL w0wa background fixed and switching on only the non-minimal coupling function Ω^EFT (with γ_{1,2,3}=0). They perform (i) a non-parametric reconstruction of Ω^EFT(a) at six nodes in 0.5≤a≤1, and (ii) a parametric reconstruction Ω^EFT(a)=Σ c_i^EFT Ω_DE^i(a) up to n=2. The principal claim is a ~3σ departure from the GR value Ω^EFT=0 (driven, via PCA, by the leading mode), with c_0^EFT>0 at >2σ in both n=1 and n=2 fits, and a preference for the n=2 form to capture a bump near z≈0.6 seen in the non-parametric reconstruction. They argue commonly used parametrizations of the form f^EFT∝Ω_DE are too restrictive.

Significance. If robust, an NMC signal in late-time data would be a meaningful step beyond the now-familiar w0wa fits to DESI, since it ties phantom-crossing phenomenology to a concrete theoretical structure (Horndeski-class scalar-tensor) and motivates richer parametrizations of EFT functions. The paper has real methodological strengths: it uses the public EFTCAMB pipeline with explicit stability checks, performs a non-parametric node reconstruction in addition to a parametric expansion, applies PCA to quantify the deviation, checks robustness against swapping DESY5↔Pantheon+ and CamSpec↔HiLLiPoP, and reports MAP locations to address projection effects. The proposed Σ c_i Ω_DE^i expansion is a useful and natural generalization. The main caveat — which the paper itself acknowledges in its conclusion — is that the signal is largely driven by the requirement of theoretical stability for a phantom-crossing CPL background within a single-field Horndeski EFT, rather than by a direct data preference for Ω^EFT≠0 over w0waCDM+PPF.

major comments (6)
  1. [Results / Table II] The headline '3σ evidence for non-minimal coupling' is in tension with the goodness-of-fit ledger reported by the authors themselves. Table II gives Δχ²(ΛCDM) = −20.5 (w0waCDM+PPF), −20.6 (n=1) and −22.3 (n=2). Adding two extra free parameters at n=2 improves χ² by only ≈1.8 over w0waCDM+PPF, and n=1 is essentially indistinguishable. The conclusion acknowledges this ('inclusion of Ω^EFT does not significantly improve fit'). The manuscript should reconcile this with the abstract's framing: report explicitly an information-criterion comparison (AIC/BIC or Bayes factor) of Ω^EFT models versus w0waCDM+PPF, and clarify in the abstract that the 3σ statement is a posterior shift away from Ω^EFT=0 conditional on the stability prior, not a data preference for MG over the PPF baseline.
  2. [Methodology / stability prior] The authors state in the conclusion that 'the signal for non-minimal coupling … arises from requiring a theoretically stable phantom crossing cosmology … within the EFT framework.' This makes the stability conditions a load-bearing piece of the inference. The paper should (i) display, for representative MCMC samples, which stability conditions (no-ghost, no-gradient, tachyon) are active and where in (w0,wa) space they cut, and (ii) quantify the prior volume effect by repeating the analysis with stability checks disabled or relaxed (e.g. allowing γ_{2,3}≠0 to absorb instabilities, as in standard EFTofDE practice). Without this, the 3σ cannot be cleanly attributed to the data.
  3. [Background choice (CPL)] Because the NMC signal is sourced by the phantom-crossing region of the CPL background, the result is sensitive to the assumption that w(a)=w0+wa(1−a) is the correct background. The paper should test at least one alternative: e.g. a thawing parametrization, a piecewise/binned w(z), or a background that does not cross −1. If c_0^EFT>0 survives only when the background actually crosses the phantom divide, that is important context for the abstract's 'detection' language.
  4. [Single-field assumption / γ_{1,2,3}=0] Setting γ_{1,2,3}=0 is justified by citing [28] as 'γ_{1,2,3} has little effect on stabilizing the theory,' but this is precisely the regime where stability pressure on Ω^EFT is maximal. Multi-field, interacting-DE, or non-Horndeski realizations (cited in the introduction as alternative routes to phantom crossing) need not produce Ω^EFT≠0. A short discussion — or, ideally, a fit allowing γ_2 or γ_3 nonzero — would clarify whether the NMC preference survives once the EFT is given more freedom to stabilize phantom crossing without modifying the conformal coupling.
  5. [Non-parametric reconstruction (Fig. 1) and PCA] The 'first principal component … deviates from GR at 2.9σ' statement should be accompanied by (a) the explicit shape of that mode, (b) the eigenvalue spectrum of the six nodes, and (c) the prior-induced correlation of the smoothing/extrapolation choice for a<0.5 (where Ω^EFT is set to the mean of the six nodes). This extrapolation is non-trivial because Ω_DE→0 at high z and any constant offset propagates through the ISW. Please show how the significance changes if the a<0.5 region is instead held to Ω^EFT=0 (i.e. enforcing GR in the matter era explicitly) rather than to the node mean.
  6. [Fig. 4 / perturbation discriminants] Fig. 4 shows that the predicted CMB TT residuals between w0waCDM+PPF, n=1, and n=2 are sub-percent and concentrated at low ℓ where cosmic variance dominates. The authors correctly note this, but the manuscript would be stronger if it provided a forecast (or at least a qualitative estimate) of the sensitivity required from weak lensing / RSD to break the degeneracy between PPF and Ω^EFT extensions. Otherwise the reader is left with a claim that is simultaneously '3σ evidence' and 'requires future perturbation data to test,' which deserves explicit reconciliation.
minor comments (8)
  1. [Abstract] The phrase 'first 3σ evidence for non-minimal coupling' would be more accurate as 'first 3σ posterior preference for Ω^EFT>0 within an EFTofDE analysis of DESI DR2+CMB+SN under a CPL background and stability priors.' Consider rewording to avoid overstating the data-only content.
  2. [Eq. (3)] Please state the convergence/regularity assumptions for the Taylor series in Ω_DE. Footnote [93] mentions n=3 is unconstrained; it would help to quote the n=3 posterior widths to substantiate this and to motivate truncation at n=2.
  3. [Table I] The prior on Ω^EFT_i is U[−1,1] and on c_i^EFT is U[−5,5]. Given that the posteriors on c_0^EFT are O(0.01), it would be useful to verify that the 2σ/3σ statements are insensitive to prior width (e.g. comparing to U[−1,1] on c_i).
  4. [Fig. 1] The caption says 'blue points with 3σ error bars' — please clarify whether these are the marginalized 1D 3σ ranges of the Ω^EFT_i node parameters, or projected from the joint posterior. Also indicate the smoothing/interpolation scheme between nodes (linear? cubic spline?), as this affects the PCA modes.
  5. [Datasets] The robustness check ('switching … to Pantheon+ … or … LoLLiPoP/HiLLiPoP … has only marginal impact') would be more credible with a supplementary table giving the c_0^EFT posteriors under each combination, rather than a one-line statement.
  6. [Fig. 2] The analytic estimate c_1^EFT/c_2^EFT ≈ −0.8 is a nice diagnostic, but the dotted line in Fig. 2 is hard to see; consider thickening or relabeling.
  7. [References] Several recent works on stability of phantom-crossing EFTofDE (e.g. discussions of operator γ_2 absorbing crossings) are cited only via [28]; an additional pointer to the systematic stability literature in the EFTofDE α-basis would help readers assess the stability-prior dependence.
  8. [Typography] 'refereed as the EFT functions' → 'referred to as'; 'we perform a the non-parametric reconstruction' → drop 'a'; 'the smallest uncertainty—deviates' is fine but appears twice with slightly different wording.

Simulated Author's Rebuttal

6 responses · 2 unresolved

We thank the referee for a careful and constructive report. The central methodological point — that our 3σ statement is a posterior departure of Ω^EFT from zero within a theoretically stable single-field Horndeski EFT supporting a CPL background, rather than a Δχ² preference of NMC over w0waCDM+PPF — is well taken, and we agree it is under-emphasized in the abstract relative to the conclusion. We will revise the framing accordingly and add the diagnostics the referee requests: (i) AIC/BIC and Bayes-factor entries to Table II; (ii) explicit visualization of which EFTCAMB stability conditions are active in the phantom-crossing region, plus chains with γ_3 free and with stability relaxed; (iii) tests with alternative backgrounds, including a binned w(z) and a thawing-style prior that excludes phantom crossing; (iv) the PCA eigenvalue spectrum and mode shapes, and a test with Ω^EFT=0 enforced in the matter era; (v) a short forecast of the WL/RSD sensitivity required to discriminate PPF from the NMC extensions at the perturbation level. We believe these additions strengthen rather than weaken the paper's message: the NMC preference is conditional on a theoretically consistent embedding of the phantom-crossing background indicated by DESI DR2, and this conditionality is itself the physically interesting result.

read point-by-point responses

  1. Referee: Headline '3σ evidence' in tension with Table II Δχ² ledger; n=1 indistinguishable from w0waCDM+PPF, n=2 improves only by ~1.8 with two extra parameters. Request explicit AIC/BIC/Bayes factor and clarification in the abstract that the 3σ is a posterior shift conditional on stability, not a data preference over PPF.

    Authors: We agree this needs to be stated more carefully. The 3σ statement is a posterior-level departure of Ω^EFT from zero under the EFTofDE stability prior, not a Δχ² preference of NMC over w0waCDM+PPF — a point we already make in the conclusion but which is under-emphasized in the abstract. In the revised version we will (i) add an AIC and BIC comparison alongside Δχ² in Table II, and a Bayes-factor estimate from the MCMC for n=1 and n=2 vs. w0waCDM+PPF; (ii) reword the abstract and the headline of the Results section to state explicitly that the 3σ is a posterior shift away from Ω^EFT=0 within a theoretically stable single-field Horndeski EFT supporting the CPL background, and not a frequentist goodness-of-fit improvement over PPF. We will retain the language of 'evidence for NMC' but qualify it by this conditional structure, which is in fact the central methodological point of the paper: PPF and the NMC EFT fit the data comparably well at the background level, but only the latter is a theoretically consistent embedding of phantom crossing. revision: yes

  2. Referee: Stability prior is load-bearing. Show which stability conditions (no-ghost/gradient/tachyon) are active and where they cut in (w0,wa); quantify prior-volume effect by repeating with stability disabled or with γ_{2,3}≠0 absorbing instabilities.

    Authors: We agree this is the most important diagnostic to add and we will include it. In revision we will: (i) provide a figure showing, for a representative subset of MCMC samples, which of the no-ghost, no-gradient, and tachyonic conditions in EFTCAMB are saturated, projected onto the (w0,wa) and (w0,wa,c_0^EFT) planes — this directly visualises the cut imposed by stability on the phantom-crossing region; (ii) rerun the n=1 and n=2 chains with the EFTCAMB stability flags disabled, to isolate the prior-volume contribution to the Ω^EFT≠0 preference; (iii) rerun with γ_2 or γ_3 free, as the referee suggests, to test whether opening additional stabilising directions in the EFT relaxes the pressure on Ω^EFT. We expect, on the basis of [28], that γ_{2,3} alone cannot stabilise phantom crossing and therefore the c_0^EFT>0 preference will largely persist, but the referee is right that this must be demonstrated rather than asserted. The outcome of these tests will be reported in a new appendix and reflected in the abstract framing. revision: yes

  3. Referee: Result is sensitive to assuming CPL. Test at least one alternative background (thawing, binned w(z), or non-phantom-crossing) to see if c_0^EFT>0 survives only when the background actually crosses −1.

    Authors: This is a fair request and is consistent with our own statement that NMC is triggered by the requirement of stable phantom crossing. We will add two background-variation tests in revision: (a) a binned/piecewise w(z) reconstruction with the same six-node spacing as our Ω^EFT reconstruction, jointly sampled with the NMC nodes; and (b) a non-phantom-crossing case enforced by restricting the prior to w0+wa≥−1 (thawing-like), to check whether c_0^EFT is consistent with zero when crossing is forbidden. We will report both posteriors. We expect — and will state explicitly — that the NMC preference is contingent on the data's preference for phantom crossing; if the phantom crossing region is excluded by prior, the NMC signal should weaken substantially. We will revise the abstract to make this conditional structure explicit rather than calling the result a model-independent 'detection'. revision: yes

  4. Referee: γ_{1,2,3}=0 assumption is questionable in precisely the regime where stability pressure on Ω^EFT is largest. Multi-field, dark-sector interactions, or non-Horndeski theories need not produce Ω^EFT≠0. Suggest fitting with γ_2 or γ_3 free, or at least discuss.

    Authors: We agree, and this is closely tied to the stability-prior diagnostic above. The reference to [28] establishes that γ_{1,2,3} cannot by themselves stabilise phantom crossing within Horndeski, but the referee correctly points out that this does not imply they are irrelevant when Ω^EFT is also free. In revision we will run an extended chain with γ_3 free (Hu-Sawicki-like braiding) in addition to Ω^EFT, and report the joint posterior. We will also add a paragraph in the Discussion explicitly noting that NMC is one of several theoretical routes to non-pathological phantom crossing — multi-field DE, dark-sector interactions, beyond-Horndeski — and that our 3σ statement should be understood as evidence within the single-field Horndeski EFT class, not as a generic detection of MG. The introduction already cites these alternative classes; we will tie this caveat explicitly to the headline claim. revision: yes

  5. Referee: PCA claim 'first PC deviates from GR at 2.9σ' needs (a) explicit mode shape, (b) eigenvalue spectrum, (c) sensitivity to the a<0.5 extrapolation (currently set to the mean of the six nodes); test enforcing Ω^EFT=0 in matter era instead.

    Authors: We will expand the PCA presentation accordingly. In revision: (i) we will add a figure showing the eigenvectors of all six modes and tabulate the eigenvalue spectrum, so the reader can see which features in a∈[0.5,1] dominate the leading mode and how rapidly the constraining power degrades for sub-leading modes; (ii) we will rerun the non-parametric analysis with two alternative high-z extensions of Ω^EFT(a<0.5): (a) Ω^EFT=0 enforced in the matter era (as the referee suggests), and (b) a smooth taper Ω^EFT(a)∝Ω_DE(a)/Ω_DE(a=0.5) so that NMC tracks the DE fraction at high z. We will report the PCA significance under each extrapolation choice. The current choice of extrapolation to the node mean was made because Ω_DE→0 at high z makes the late ISW contribution from a constant Ω^EFT small in any case, but the referee is right that this should be demonstrated quantitatively rather than asserted, and we will include a comparison plot of the resulting low-ℓ TT residuals. revision: yes

  6. Referee: Fig. 4 shows sub-percent CMB TT residuals at cosmic-variance-dominated low ℓ. Provide forecast or estimate of the WL/RSD sensitivity needed to break PPF vs Ω^EFT degeneracy, otherwise the '3σ evidence' and 'needs future data' statements are in tension.

    Authors: We agree this reconciliation is needed for the reader. We will add a short forecast section computing the linear growth signature f(z)σ_8(z) and the lensing convergence κκ spectrum for the n=1 and n=2 MAP models versus w0waCDM+PPF over 0<z<2, and compare to forecast error bars from DESI full-shape, Euclid weak lensing, and LSST Y10. Even an order-of-magnitude estimate of the redshift range and precision at which PPF and Ω^EFT diverge in growth will sharpen the message. We will also revise the closing paragraph to state explicitly that the 3σ here is a constraint on the model space within EFTofDE given current background+CMB data, and that a direct data-level discrimination of NMC from PPF requires the perturbation observables we forecast — these two statements are complementary, not contradictory, and we will phrase them as such. revision: yes

standing simulated objections not resolved
  • A fully model-independent Bayes-factor comparison between the NMC EFT and w0waCDM+PPF requires a well-defined prior on the Ω^EFT node values and Taylor coefficients; our priors (Table I) are uniform but their ranges are chosen on physical/stability grounds rather than from a first-principles theory measure, so the resulting Bayes factor will inherit some prior-range dependence which we will report transparently but cannot eliminate.
  • Demonstrating that the NMC signal is genuinely 'data-driven' rather than 'stability-prior-driven' is intrinsically limited: with stability checks disabled, the EFT samples include ghosts/gradient instabilities and the resulting posterior is not physically interpretable. We can report it as a diagnostic of prior volume but it is not, by itself, a meaningful alternative inference.

Circularity Check

4 steps flagged

"3σ NMC detection" largely renames a phantom-crossing CPL background as modified gravity through a self-citation-justified stability prior; data Δχ² is essentially flat between w0waCDM+PPF and the NMC extensions.

specific steps
  1. ansatz smuggled in via citation [Methodology, paragraph after Eq. (2)]
    "We therefore focus on the non-minimal coupling EFT function Ω^EFT and set γ_{1,2,3}=0. As shown in [28], this is justified because γ_{1,2,3} has little effect on stabilizing the theory."

    Reference [28] is Ye, Martinelli, Hu, Silvestri (2024) — same author Ye. Setting γ_{1,2,3}=0 is the assumption that channels all stability pressure into Ω^EFT, which then drives the claimed detection. The justification is a self-citation to a paper using the same framework, not an external or model-independent result. The choice is load-bearing: relaxing γ_{1,2,3}≠0 could absorb the stability requirement without forcing Ω^EFT>0.

  2. self citation load bearing [Results, after Fig. 1 discussion]
    "This is also consistent with the finding of [103] in that stability generally prefers Ω^EFT>0."

    The interpretation that the reconstruction's preference for Ω^EFT>0 is meaningful is supported by citing a review stating that stability prefers Ω^EFT>0 in this EFT class. That is the very prior being imposed, not independent corroboration. The 3σ shape constraint then reduces to: data prefer phantom crossing, and the imposed stability prior in single-field Horndeski pushes Ω^EFT>0.

  3. renaming known result [Conclusion]
    "the signal for non-minimal coupling of gravity arises from requiring a theoretically stable phantom crossing cosmology, as indicated by the data, within the EFT framework."

    The authors concede the signal is not an independent data preference for MG but a consequence of imposing single-field EFT stability on an assumed CPL phantom-crossing background. Combined with Table II (Δχ² differences of 0.1 and 1.7 between w0waCDM+PPF and the NMC extensions), this means the headline '3σ detection of non-minimal coupling' largely renames the previously reported phantom-crossing preference under a stability prior. The novel content is the relabeling, not a new data-driven detection.

  4. fitted input called prediction [Table II and Conclusion]
    "w0waCDM + PPF: Δχ² = -20.5; w0waCDM + Ω^EFT(n=1): -20.6; w0waCDM + Ω^EFT(n=2): -22.3."

    The 3σ posterior preference for Ω^EFT≠0 is reported alongside near-identical χ² to the NMC-free w0waCDM+PPF baseline. The posterior shift is therefore driven primarily by the prior structure (theoretical stability in the chosen EFT sector) rather than by the data's likelihood preference. Presenting this as a '3σ detection' conflates the prior-driven posterior with a data-driven measurement.

full rationale

The paper's central claim — first 3σ evidence for non-minimal coupling — does not derive from a direct data preference for Ω^EFT≠0. The authors' own Table II shows Δχ² = −20.5 for w0waCDM+PPF vs. −20.6 (n=1) and −22.3 (n=2), i.e. adding NMC parameters improves the fit by ~0.1–1.7 in χ². The Conclusion concedes this explicitly: "inclusion of the non-minimal coupling function Ω^EFT does not significantly improve fit to SNIa, BAO and CMB." The same paragraph states the signal "arises from requiring a theoretically stable phantom crossing cosmology … within the EFT framework." Thus the chain is: data → phantom crossing in CPL → impose single-field Horndeski stability with γ_{1,2,3}=0 → Ω^EFT>0 forced. The 3σ is a 3σ statement about phantom crossing convolved with a chosen stability prior, partially relabeled as MG detection. Two specific circularity-adjacent moves are load-bearing: (i) the choice to set γ_{1,2,3}=0 is justified by self-citation to Ye et al. [28], an author of the present paper, with the bald assertion that "this is justified because γ_{1,2,3} has little effect on stabilizing the theory" — the cited result itself is not machine-checked or externally falsifiable independent of the same EFT framework; (ii) [103] (Frusciante & Perenon) is cited to assert "stability generally prefers Ω^EFT>0," which is precisely the prior generating the apparent detection. The PCA-based 2.9σ statement on the first principal component is computed on the same posteriors that incorporate the stability prior, so it inherits the same renaming. None of the steps is strictly self-definitional (no parameter is fit and then renamed as a prediction), so this is not a 6+ case. But the headline claim is materially weaker than presented: phantom crossing in CPL plus a self-cited choice of EFT sector is being relabeled as a "detection of modified-gravity effects." Score 4: load-bearing self-citations and an ansatz (γ=0, single-field Horndeski) doing the work, with the data themselves agnostic per the paper's own Δχ².

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Model omitted the axiom ledger; defaulted for pipeline continuity.

pith-pipeline@v0.9.0 · 9553 in / 7186 out tokens · 109185 ms · 2026-05-06T20:58:18.654542+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Late-time reconstruction of non-minimally coupled gravity with a smoothness prior

    astro-ph.CO 2026-05 unverdicted novelty 7.0

    Non-parametric reconstruction of non-minimally coupled gravity with a smoothness prior on CMB, DESI BAO, supernovae, and DES data yields a 2.8σ hint for coupling and a preference for phantom divide crossing stabilized...

  2. Post-Newtonian Constraints on Scalar-Tensor Gravity

    gr-qc 2026-04 unverdicted novelty 6.0

    Unified post-Newtonian analysis reveals that Palatini scalar-tensor theories often face weaker Solar System bounds than metric versions due to stronger Yukawa suppression, with Palatini f(R) reproducing GR limits for ...