🔬 physics

Does DESI Provide Evidence for Dynamical Dark Energy?

This paper argues that while recent DESI results favor a dynamical dark energy model over $\Lambda$ CDM within the standard fluid framework (claim D2), they do not constitute evidence for a physical dynamical dark energy component (claim D3) because the same observational data can be equally explained by modified gravity theories, revealing a fundamental underdetermination between matter-sector and gravitational-sector interpretations.

Original authors: Nicola Bamonti

Published 2026-08-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Nicola Bamonti

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

For a quarter of a century, the most successful story we have told about the universe has been one of simple, steady expansion. Astronomers have long believed that the cosmos is filled with ordinary matter, invisible dark matter, and a mysterious force called dark energy that pushes galaxies apart. In the standard version of this story, dark energy is a constant, unchanging pressure built into the fabric of space itself, a cosmic background hum that has remained the same since the beginning of time. This model, known as the cosmological constant, has fit almost every observation we have made for decades. However, a new set of measurements from a massive telescope survey called DESI has shaken this quiet certainty. The data suggests that the push of dark energy might not be constant at all, but rather a force that changes over time, growing stronger or weaker as the universe ages. If true, this would mean the universe is not governed by a simple, static rule, but by a dynamic, evolving process that we do not yet understand.

A new paper by philosopher Nicola Bamonti examines the recent headlines surrounding these DESI results and argues that the scientific community may be jumping to a conclusion that the data does not yet support. The announcement that "dark energy is dynamical" is a bold claim about the nature of reality, suggesting that the universe contains a new, shifting substance. Bamonti's analysis shows that while the DESI measurements are real and significant, they only prove that the universe is expanding in a way that differs from the old, constant model. They do not prove that this difference is caused by a new, changing substance. The data could just as easily be explained by a change in the laws of gravity themselves, rather than a change in the stuff filling the universe.

The DESI survey works by mapping the positions of millions of galaxies to measure how the universe has stretched over billions of years. When these measurements are combined with other data from the early universe and exploding stars, the results point toward a specific mathematical pattern where the strength of dark energy changes. In the language of the standard model, this pattern looks like a fluid that is evolving. The statistical confidence in this pattern is strong, reaching a level that scientists call a three-to-four sigma preference, which is usually enough to suggest a new discovery. However, Bamonti points out that this statistical preference is calculated within a very specific framework that assumes the laws of gravity are fixed and that any change must come from a new type of fluid. The calculation essentially asks, "If gravity is fixed, what kind of fluid fits the data?" and the answer is "a changing one." But the calculation does not ask, "What if gravity itself is changing?"

The paper demonstrates that the same expansion history measured by DESI can be generated perfectly well by theories where gravity is not fixed, but instead behaves differently at different scales. In these theories, there is no mysterious, changing fluid at all. Instead, the apparent change in dark energy is an illusion created by the way gravity itself works. The author uses a mathematical tool known as a "designer construction" to show that for any observed pattern of expansion, one can construct a theory of modified gravity that reproduces that pattern exactly. This means that the background data alone cannot tell the difference between a universe filled with a changing substance and a universe where the rules of gravity are different. The two explanations are indistinguishable when looking only at the overall expansion rate.

This ambiguity becomes even more striking when looking at a specific feature of the DESI data: the "phantom crossing." The data suggests that the strength of dark energy was once weaker than a specific threshold and has since become stronger, crossing a line that separates two different types of behavior. If we assume the standard model of fixed gravity, crossing this line requires the existence of exotic, unstable matter that violates fundamental physical laws. It is a theoretical nightmare. However, in the alternative view where gravity is modified, this same crossing happens naturally and without any instability or exotic ingredients. The feature that makes the "changing fluid" story so exciting to some scientists actually makes it much more difficult to believe from a theoretical standpoint, while the "changing gravity" story handles it with ease.

The paper concludes that the current reporting of the DESI results is misleading because it presents a mathematical preference for a changing fluid as a discovery of a new physical substance. The data establishes that the universe is expanding differently than the old constant model predicted, but it does not establish what is causing that difference. The choice between a new substance and a new law of gravity is not something the current measurements can decide. The author argues that the scientific community should be careful not to declare a victory for one explanation before the evidence is ready. To settle the question, scientists need to look beyond the overall expansion rate and study how galaxies clump together, how light bends around massive objects, and how gravitational waves travel. These other channels of observation are currently being explored, but until they provide a clear answer, the true nature of the universe's acceleration remains an open mystery, waiting for a verdict that the current data cannot yet deliver.

Technical Summary: "Does DESI Provide Evidence for Dynamical Dark Energy?"

Problem Statement
Recent results from the Dark Energy Spectroscopic Instrument (DESI) collaboration, particularly when combined with Cosmic Microwave Background (CMB) and Type Ia supernova (SN Ia) data, have been widely interpreted as strong evidence for dynamical dark energy. Specifically, the data show a preference (at roughly 3.1σ to 4.2σ significance) for a Chevallier–Polarski–Linder (CPL) equation of state $w(a) = w_0 + w_a(1-a)$ where $(w_0, w_a) \neq (-1, 0)$ , implying an evolving dark energy component that crosses the phantom divide ( $w = -1$ ).

This paper argues that the inferential leap from these statistical preferences to the ontological claim that "dark energy is dynamical" is unjustified. The author distinguishes three claims often conflated in the literature:

(D1) Background Anomaly: The data disfavour the $\Lambda$ CDM expansion history $H(z)$ .
(D2) Parametric Preference: Within the restricted model class of General Relativity (GR) plus a smooth fluid with a CPL equation of state, the data prefer $(w_0, w_a) \neq (-1, 0)$ .
(D3) Ontological Claim: There exists a physical dark-energy component in the matter sector with an evolving equation of state.

The paper contends that while DESI establishes D2 (and supports D1), it does not establish D3. The background expansion history alone cannot distinguish between a dynamical fluid component and modified gravitational dynamics.

Methodology
The paper employs a philosophical and theoretical analysis of the inferential structure of cosmological data interpretation, utilizing specific results from modified gravity theory:

Inferential Reconstruction: The author traces the chain of inference from raw observables (BAO distances) to the headline claim, identifying where theoretical assumptions (FLRW geometry, GR, fluid continuity) are introduced.
The Designer Construction (Degeneracy Theorem): The paper utilizes the "designer" reconstruction method in $f(R)$ gravity. This method demonstrates that for any sufficiently regular target expansion history $H(z)$ , there exists a one-parameter family of $f(R)$ Lagrangians (with no dark energy component in the matter sector) that reproduces that history exactly.
Asymmetry Argument: The paper analyzes the specific qualitative feature of the DESI-preferred solution: the crossing of the phantom divide ( $w < -1$ ). It compares the theoretical cost of realizing this crossing in the fluid framework versus the gravitational framework.
Epistemological Analysis: The paper examines the reporting conventions of the DESI collaboration, specifically how the choice of the CPL parameterization as a headline statistic biases the interpretation toward a fluid ontology.

Key Contributions and Results

Exact Background Degeneracy: The paper establishes a "designer theorem" (Theorem 3.1) proving that for any admissible expansion history $H(z)$ , there exists a family of $f(R)$ theories that reproduce it exactly. Consequently, any observable determined solely by the homogeneous expansion history (BAO, SN Ia distances, CMB distance priors) is incapable of discriminating between GR + a fluid with $w(z)$ and a modified gravity theory mimicking that history. This degeneracy is exact, not merely approximate.
The Phantom Divide Asymmetry: The DESI-preferred region requires $w(z)$ $w (z)$ to cross $-1$.
- Fluid Reading: Realizing this crossing within GR with a single, minimally coupled component requires violating the null energy condition or introducing pathological structures (ghosts, multiple tuned fields).
- Gravitational Reading: In $f(R)$ and scalar-tensor theories, the effective equation of state $w_{eff}(z)$ crossing $-1$ is a generic, non-pathological feature of the scalaron evolution.
- Result: The feature driving the "dynamical dark energy" announcement is theoretically more natural under the modified-gravity interpretation (which the announcement ignores) than under the fluid interpretation it asserts.
Viability of Gravitational Candidates: The paper notes that while specific screened $f(R)$ models (like Hu-Sawicki) are too close to $\Lambda$ CDM to fit the DESI amplitude, non-minimally coupled theories fit the data as well as the fluid template. While their viability regarding screening mechanisms is an open question, they are not excluded. The burden of proof lies on demonstrating that all healthy gravitational accounts fail, which has not been done.
Reporting Bias: The paper argues that the use of $(w_0, w_a)$ as the primary headline statistic inherently frames the expansion history in the vocabulary of a fluid. This reporting convention invites the slide from D2 to D3, presenting a model-dependent reconstruction as an ontological discovery.

Significance and Claims
The paper does not claim that the DESI data are wrong or that modified gravity is definitively correct. Rather, it claims that the current data are underdetermined regarding the physical origin of cosmic acceleration.

Limitation of Background Data: The "evidence for dynamical dark energy" is an elliptical claim. The data provide evidence for a departure from the $\Lambda$ CDM expansion history, but the physical interpretation (new substance vs. new gravitational law) cannot be settled by background observables alone.
Breakability of Underdetermination: The underdetermination is exact at the level of the background but breakable at the level of perturbations. The paper identifies three channels capable of distinguishing the rivals:
1. Cosmological Perturbations: Measurements of growth rates, weak lensing, and the parameters $\mu$ and $\eta$ (which describe modifications to the Poisson equation and the ratio of metric potentials).
2. Gravitational Wave Cosmology: Standard sirens can detect differences between gravitational wave luminosity distance and electromagnetic luminosity distance, a signature of evolving effective gravitational coupling.
3. Local Tests: Solar System and laboratory fifth-force searches can constrain the screening mechanisms required for viable modified gravity models.
Conclusion on Reporting: The paper concludes that survey reporting should avoid ontological headlines based on background-only data. Instead, results should be framed in terms of shared quantities (like $H(z)$ ) or include constraints on gravitational response parameters ( $\mu, \Sigma$ ) alongside expansion parameters. Until the perturbative and local channels provide discriminating evidence, the physical origin of the apparent deviation remains undetermined.

In summary, the paper asserts that the DESI collaboration's announcement of "dynamical dark energy" oversteps the evidential support of the data, which currently only supports a framework-internal preference for a specific parametric form of the expansion history, a form that can be equally well generated by modified gravity.

More like this