Dark matter and dark energy: what is observation, what is postulate

Why this article exists

Volume 2 of the MRIPR cosmology audit. Volume 1 addressed redshift. This volume addresses the two postulated entities that, together, account for ~95% of the mass-energy content claimed by the standard cosmological model: dark matter and dark energy.

Popular, pedagogical, and much professional usage presents dark matter and dark energy as discovered entities — substances whose existence has been established and whose properties are now being measured. This usage is incorrect. Dark matter and dark energy are postulated entities, introduced in specific historical moments to close specific residuals between framework predictions and observations. Their existence, quantity, and properties are not measured; they are derived from the requirement that a specified framework continue to describe the observations.

This article, in parallel with Volume 1, separates the observations from the postulates. Each statement is labelled for what it is. The same labels as Volume 1 are used, plus one additional label specific to this volume:

• Direct observation (Layer 0). Registered by the instrument. No interpretive content.
• Laboratory-backed measurement. Secured by terrestrial experiment under controlled conditions.
• Empirically transferred assumption. Supported by internal consistencies of the data, not independently verified at the target regime.
• Postulate. An interpretive commitment not contained in the observation.
• Framework-conditional output. A quantity whose value depends on accepting a specified stack of postulates.
• Residual-defined postulate. A postulate whose properties are specified by whatever is required to close the gap between a framework’s prediction and the observation it failed to reproduce.

The last category — residual-defined postulate — is the central category for this volume. It is the category to which dark matter and dark energy belong. It is distinct from “entity” in any empirical sense, because an entity’s properties are measured; a residual-defined postulate’s properties are inferred from what must be true for the framework to continue working.

The observations that preceded the postulate

1.1 Zwicky, Coma cluster, 1933

Primary source: Zwicky, F. (1933). “Die Rotverschiebung von extragalaktischen Nebeln.” Helvetica Physica Acta 6: 110–127.

What the instruments registered:

• Photon counts at wavelength bins at angular positions on the sky, for individual galaxies in the direction of the Coma cluster.
• Apparent brightnesses (fluxes) of those galaxies.
• Angular separations between galaxies on the sky.

What was computed from the observations:

• Wavelength ratios z for each galaxy with identified spectral features. (Layer 2, conditional on atomic-physics constancy.)
• Cluster membership inferred from angular proximity on the sky and similarity of z values.

What was assumed in Zwicky’s analysis:

1. z interpreted as line-of-sight velocity (Doppler postulate). See Volume 1, §4.
2. The Coma cluster is a gravitationally bound, dynamically relaxed system. Assumption about the dynamical state of the system.
3. The virial theorem applies to this system. A Newtonian-mechanical result: for a bound system in equilibrium, 2⟨T⟩ + ⟨V⟩ = 0. Using it requires Newtonian mechanics and gravity to hold at cluster scales (~Mpc).
4. Newtonian gravity applies at cluster scales. Extrapolation of ~12 orders of magnitude from the regimes (solar system, ~AU) where Newton had been tested to the target regime (cluster, ~Mpc).
5. Luminosity traces gravitating mass via a mass-to-light ratio inferred from stellar populations studied at much smaller scales (within the Milky Way and its neighbourhood).

What Zwicky found: Applying the virial theorem under assumptions (1)–(5) yielded a required gravitating mass ~400× greater than the luminous mass implied by assumption (5). Later revisions changed the factor but not the qualitative discrepancy.

What Zwicky did: He postulated the existence of dunkle Materie (“dark matter”) — non-luminous matter sufficient to close the gap between assumptions (1)–(5) and the observations.

Epistemic status of the postulate (as introduced):

• Not derived from a direct observation of dark matter. No direct observation of any dark-matter entity was involved.
• Defined by the residual. The quantity of dark matter postulated was exactly the quantity required to make assumptions (1)–(5) consistent with the observations.
• Conditional on five separate assumptions. If any of (1)–(5) fails, the dark-matter postulate is not required by Zwicky’s argument.

1.2 Babcock, Andromeda, 1939

Primary source: Babcock, H. W. (1939). “The rotation of the Andromeda Nebula.” Lick Observatory Bulletin 19: 41.

What the instruments registered:

• Photon counts at wavelength bins at angular positions along the major axis of the Andromeda Nebula (M31).
• Wavelengths of emission and absorption features at each angular position, identified with atomic transitions.

What Babcock observed: The wavelengths of spectral features varied systematically across the major axis of M31, in a pattern consistent with the outer regions not showing the wavelength decrease expected under the (then-standard) interpretive stack.

What Babcock did: He did not postulate dark matter. He attributed the anomaly to high mass-to-light ratio in the outer disk and to absorption effects. The observation existed decades before the postulate was formalized.

Direct observation, consistent with the assumption that Newtonian gravity at galactic scales is problematic, but not interpreted as evidence for a non-luminous entity.

1.3 Rubin, Ford, Thonnard — M31 (1970), 21-galaxy survey (1978, 1980)

Primary sources:

• Rubin, V. C. & Ford, W. K. (1970). “Rotation of the Andromeda Nebula from a spectroscopic survey of emission regions.” Astrophysical Journal 159: 379.
• Rubin, V. C., Ford, W. K. & Thonnard, N. (1978). “Extended rotation curves of high-luminosity spiral galaxies. IV.” Astrophysical Journal Letters 225: L107.
• Rubin, V. C., Ford, W. K. & Thonnard, N. (1980). “Rotational properties of 21 Sc galaxies with a large range of luminosities and radii.” Astrophysical Journal 238: 471.

What the instruments registered:

• Photon counts at wavelength bins at measured angular positions along the major axes of M31 (1970) and 21 additional spiral galaxies (1980), out to radii limited by surface brightness of the tracers.
• Wavelengths of Hα and [NII] emission lines from HII regions at each angular position.

The pattern in the observation: The wavelength ratios of the emission-line gas did not decrease with angular distance from the galactic centre in the manner that would follow if (i) the gas is on circular orbits, (ii) Newtonian dynamics governs at galactic scales, and (iii) the visible light traces the gravitating mass.

What the observation IS, without interpretation: A set of wavelength measurements at angular positions, whose pattern, under those three assumptions, did not match the prediction.

What the observation IS NOT:

• Not “flat rotation curves.” (“Rotation curve” is an interpretation requiring Doppler + circular orbits + inclination + distance estimate.)
• Not “missing mass.” (“Missing mass” is an inference requiring Newtonian gravity at galactic scales and the mass-tracer assumption.)
• Not “evidence for dark matter.” (A conclusion requiring the prior acceptance of all the above plus the choice to resolve the mismatch by adding unseen matter rather than by modifying the framework.)

The wavelength measurements are direct observation. Everything further is layered on top.

1.4 The stack of assumptions between observation and “missing mass”

To convert the raw wavelength-at-angular-position data of Rubin, Ford, and Thonnard into a statement of the form “galaxies have missing mass,” the following must be accepted:

Each layer above 1b is an additional commitment. If any of layers 2–7 fails, the inference at layer 8 is not forced, and the conclusion at layer 9 does not follow.

The postulate: dark-matter halos (1974)

2.1 The three simultaneous papers

Primary sources:

• Ostriker, J. P. & Peebles, P. J. E. (1973). “A numerical study of the stability of flattened galaxies: or, can cold galaxies survive?” Astrophysical Journal 186: 467.
• Ostriker, J. P., Peebles, P. J. E. & Yahil, A. (1974). “The size and mass of galaxies, and the mass of the universe.” Astrophysical Journal Letters 193: L1.
• Einasto, J., Kaasik, A. & Saar, E. (1974). “Dynamic evidence on massive coronas of galaxies.” Nature 250: 309.

In 1973–74, three nearly simultaneous papers argued that the resolution of the galactic rotation-curve residual was massive, extended, non-luminous halos surrounding spiral galaxies. This is the moment at which the dark-matter-halo postulate was formalised as the consensus resolution.

2.2 The content of the postulate

What is postulated:

• A spatially extended, roughly spherical distribution of non-luminous matter surrounding each spiral galaxy.
• This matter does not emit or absorb light at detectable levels across any observed electromagnetic band.
• The matter’s gravitational behaviour is governed by Newtonian gravity at galactic scales (i.e., the same framework whose prediction was being rescued).
• The matter’s total quantity, spatial distribution, and dynamics are chosen such that, when added to the visible matter, the composite system reproduces the observed rotation curve under the same framework whose prediction was failing.

What is assumed about the postulated matter:

• That it behaves gravitationally the way Newtonian gravity prescribes. (The assumption whose failure prompted the postulate, now reapplied to the postulate itself.)
• That it clusters in stable halos on the galactic timescale.
• That it is non-relativistic (“cold”) at galactic scales.
• That it interacts with ordinary matter only gravitationally, or at most very weakly (to avoid detection in non-gravitational experiments).

Properties NOT specified by the postulate itself:

• The particle nature of the matter (if it is particles).
• The particle mass.
• Any non-gravitational interaction cross-section.
• Any direct signature.

These properties were deferred to later model-building programmes (WIMPs, axions, sterile neutrinos, primordial black holes, etc.), each of which adds further assumptions to the halo postulate.

2.3 The logical alternative

The rotation-curve residual could have been resolved in one of three ways:

1. Return to the raw observations and re-audit the interpretive layers (Doppler, circular orbits, inclination, distance, Newton at galactic scales, mass tracer).
2. Modify the framework at the scale where it failed — that is, modify Newtonian gravity at galactic scales.
3. Preserve the framework and postulate a new entity to absorb the discrepancy.

The 1974 papers chose (3). This was not a forced choice. It was a methodological choice, made in the absence of a compelling proposal for (2), and with (1) not pursued as a mainstream research programme.

The dark-matter-halo postulate is the result of a methodological choice among three logically available options, not the result of a direct observation forcing one option over the others.

The alternative that was also raised: MOND (1983)

Primary source: Milgrom, M. (1983). “A modification of the Newtonian dynamics as a possible alternative to the hidden mass hypothesis.” Astrophysical Journal 270: 365–370 (with companion papers in the same volume).

3.1 The content of MOND

Milgrom proposed: below a characteristic acceleration scale a₀ ≈ 1.2 × 10⁻¹⁰ m/s², Newton’s second law (or equivalently the gravitational force law) is modified. Above a₀, Newtonian gravity applies as usual.

What is assumed in MOND:

1. The Doppler + circular-orbits + inclination + distance-ladder stack (layers 2–5 of §1.4). MOND does not audit these; it accepts them.
2. Newtonian gravity applies above a₀.
3. A specific functional modification applies below a₀, parameterised by the single constant a₀.

What a₀ is: One number, fitted to galactic data. Its value is tuned to reproduce galactic rotation curves and the baryonic Tully-Fisher relation.

3.2 MOND’s successes and failures

Successes:

• Predicts the baryonic Tully-Fisher relation (M_baryon ∝ V_flat⁴) as a direct consequence of the functional form.
• Reproduces the observed radial acceleration relation (RAR) in galaxies.
• Requires no galaxy-specific parameters beyond the single global constant a₀.

Failures:

• Does not account for cluster-scale dynamics without additional (often dark-matter-like) components.
• Does not account for CMB structure.
• Does not account for large-scale structure formation in the standard cosmological timescale.

3.3 MOND’s epistemic status, compared to dark matter

MOND is structurally the mirror image of dark matter at galactic scale: one fitted parameter modifying the framework, versus many fitted parameters per galaxy modifying the matter content. Neither is a direct observation of anything. Both are residual-absorbing postulates, introduced to close the rotation-curve discrepancy.

The observation does not select between MOND and dark-matter halos. Each accepts layers 2–5 of the interpretive stack and chooses different modifications thereafter.

The “evidence for dark matter” since 1974

Every subsequent line of evidence adduced for dark matter’s existence is conditional on the framework whose failure motivated the postulate. Each is a framework-internal coherence check, not an independent observation of the postulated entity.

4.1 Gravitational lensing

What is observed: Photon arrival positions from distant sources (quasars, galaxies) showing systematic distortions consistent with light bending around intervening mass concentrations. Direct observation at Layer 0.

What is inferred: The amount of mass required to produce the observed lensing, under general relativity applied at the relevant scales.

What is concluded: The inferred mass exceeds the visible mass.

Assumptions required to reach the conclusion:

1. Photon paths in a gravitational field are described by GR.
2. GR applies at the scales of the lens (kpc to Mpc).
3. The mass-tracer assumption (visible light traces mass).
4. The geometry of the lens system is as assumed in the model fit.

The lensing does not observe dark matter; it observes photon distortions, which are interpreted as implying mass under a framework whose application at the relevant scales is part of what is being preserved.

4.2 The Bullet Cluster (2006)

Primary source: Clowe, D. et al. (2006). “A direct empirical proof of the existence of dark matter.” Astrophysical Journal Letters 648: L109.

What is observed: X-ray emission from hot gas in two colliding clusters (direct observation, with assumptions about X-ray emission mechanisms and thermodynamics), plus gravitational-lensing signal from the same system.

What is claimed: The lensing-signal peaks are spatially offset from the X-ray gas peaks. Under the framework, this implies that the dominant mass component is collisionless and separated from the baryonic gas during the collision.

Assumptions required for the interpretation:

1. All assumptions in §4.1 (GR at cluster scales, mass-tracer, etc.).
2. X-ray emission traces most of the baryonic mass in the cluster.
3. The interpretation of the lensing map as a mass map requires the framework to apply.
4. The spatial offset implies separation of components with different collisional properties.

Widely cited as “direct proof” of dark matter. The citation overreaches. The lensing peaks being offset from the X-ray peaks is consistent with several interpretations including the dark-matter one; the choice among them is made by which framework one has already committed to applying at cluster scales.

4.3 CMB acoustic peaks

Primary source: Planck Collaboration (2020). “Planck 2018 results. VI. Cosmological parameters.” Astronomy & Astrophysics 641: A6.

What is observed: Temperature variations in the cosmic microwave background sky at angular scales from ~degree down to arcminute. Direct observation at Layer 0.

What is claimed: The angular power spectrum of these variations has features (acoustic peaks) whose relative heights and positions require a specific fraction of non-baryonic “cold dark matter” in the early universe.

Assumptions required for the interpretation:

1. The CMB is the redshifted remnant of a hot, dense early phase. Requires the full expansion-postulate stack (V1, §5).
2. FLRW metric applies. 20-item assumption stack (V1, §6).
3. Photon-baryon fluid dynamics in the early universe.
4. Specific equations of state for each assumed matter component.
5. Fixed values of fundamental constants across the relevant epoch.
6. The power spectrum is fit using a six-parameter ΛCDM model, including Ω_c (cold dark matter) and Ω_Λ (dark energy) as two of the parameters.

“The CMB shows dark matter is 27% of the universe” is the output of the ΛCDM fit to the power spectrum. Different frameworks applied to the same power spectrum would yield different parameter extractions.

4.4 Large-scale structure formation

What is observed: Galaxy positions, redshifts, and clustering statistics (direct observation at Layer 0, after the redshift interpretive stack).

What is claimed: The observed pattern of galaxy clustering requires seed perturbations plus enhanced gravitational growth that cannot be supplied by baryons alone in the available cosmic time.

Assumptions required: the full ΛCDM framework, plus specific assumptions about initial perturbations (usually from inflation), plus N-body simulation assumptions about how structure forms gravitationally from those initial conditions.

4.5 Direct-detection experiments

Experimental programmes: XENON, LUX, LZ, DAMA/LIBRA, CDMS, CRESST, ADMX, and numerous others.

What the experiments do: Operate detectors in deep underground laboratories designed to register interactions of hypothetical dark-matter particles with ordinary matter, at cross-sections and mass ranges predicted by specific particle-physics models (WIMPs, axions, etc.).

Results after four decades (as of 2025):

• No confirmed direct detection of any dark-matter particle.
• Upper limits on interaction cross-sections have been tightened by many orders of magnitude.
• The parameter space favoured by the “natural” WIMP scenario is largely excluded.
• The DAMA/LIBRA annual-modulation signal remains, but no other experiment has confirmed it.

Four decades of direct-detection experiments have not produced a direct observation of any dark-matter entity. The postulate’s empirical status has not changed since 1974.

Summary: the epistemic status of dark matter

Dark matter is a residual-defined postulate, introduced in 1974 to preserve the Newtonian / GR framework at scales where its application had not been independently verified. In the half-century since, no direct observation of any dark-matter entity has been made.

The observation that preceded the postulate

6.1 Type Ia supernova distance–redshift measurements (1998–1999)

Primary sources:

• Riess, A. G. et al. (1998). “Observational evidence from supernovae for an accelerating universe and a cosmological constant.” Astronomical Journal 116: 1009–1038.
• Perlmutter, S. et al. (1999). “Measurements of Ω and Λ from 42 high-redshift supernovae.” Astrophysical Journal 517: 565–586.

What the instruments registered:

• Photon counts at wavelength bins at angular positions at times, for Type Ia supernova events.
• The resulting light curves (brightness as a function of time) and spectra for each event.
• Identification of each event as a Type Ia based on spectral features.

What was computed:

• The redshift z for each host galaxy (V1).
• The apparent peak brightness of each supernova.
• The apparent “distance modulus” μ, inferred from the apparent peak brightness under the assumption that Type Ia supernovae are standardisable candles with a known absolute magnitude after light-curve shape corrections.

Assumptions required to convert observations into the dataset:

1. Type Ia supernovae have an absolute magnitude that is the same (after standardisation) at z = 0.5 or 1 as at z ≈ 0 where their absolute magnitudes are calibrated. Empirically transferred assumption.
2. The light-curve shape corrections (stretch, colour) apply across cosmic time. Assumption.
3. No systematic evolution of Type Ia progenitor populations with redshift affects the inferred absolute magnitude. Assumption.
4. Redshift as interpreted per V1 (Doppler + expansion + FLRW + distance ladder). Multi-layer postulate stack.
5. The ΛCDM-family models (with specified content Ω_m, Ω_Λ, etc.) are the correct framework against which to compare the data. Framework commitment.

What was found: The distance-modulus-versus-redshift relation for Type Ia supernovae at z ≈ 0.3–0.8 deviated from the prediction of a matter-only ΛCDM model (Ω_Λ = 0). Under the framework, the deviation could be absorbed by allowing Ω_Λ > 0 — that is, by introducing a positive cosmological constant or equivalent dark-energy component.

The observation is a specific relation between Type Ia apparent brightnesses and their host-galaxy redshifts. Everything further is layered on top.

The postulate: dark energy (1998–1999)

7.1 The content of the postulate

What is postulated: A component of the universe’s energy content with an equation of state p ≈ −ρ (or p = wρ with w ≈ −1), whose density remains approximately constant or evolves slowly as the universe expands, and which dominates the universe’s dynamics at late times, producing an accelerating expansion.

Two common realisations:

• Cosmological constant Λ: a constant term in the Einstein field equations, equivalent to an energy density of the vacuum. Mathematically permitted by GR since Einstein introduced Λ in 1917.
• Dynamical dark energy: a scalar field or other dynamical component whose equation of state may deviate from w = −1 and may evolve with time (quintessence, phantom energy, and similar proposals).

What is assumed in the postulate:

1. The Doppler + expansion + FLRW stack (V1, §§4–6).
2. Type Ia supernovae are standardisable candles across cosmic time.
3. The distance-ladder calibrations apply.
4. GR applies at cosmological scales.
5. The Einstein field equations with a Λ term or equivalent dynamical field are the correct description of the universe’s large-scale dynamics.
6. The observed deviation from the matter-only prediction is due to (a) an actual acceleration of cosmic expansion, caused by (b) a specific form of energy content with the required equation of state.

Properties NOT specified by the postulate itself:

• The physical origin of the energy (vacuum, scalar field, modification of gravity, something else).
• Why its density has the observed value (the “cosmological-constant problem” — the discrepancy between the observed energy density and any naturally predicted value from particle physics, which spans ~120 orders of magnitude).
• Whether it is truly constant or evolves.

7.2 The cosmological-constant problem

The predicted value of the vacuum energy density from quantum field theory, computed naively, exceeds the observed cosmological-constant value by a factor of ~10¹²⁰. No resolution of this discrepancy that is both empirically adequate and theoretically motivated has been established. Proposals (anthropic selection in a multiverse, supersymmetric cancellations, modifications of QFT in curved spacetime, etc.) are not observationally supported.

The most widely cited dark-energy candidate (the cosmological constant) has no empirically viable derivation from underlying physics. It is a number fitted to the observations, without an independent theoretical prediction of its value.

7.3 The logical alternatives to dark energy

The supernova-distance-modulus residual could have been resolved in multiple ways:

1. Return to the observations and re-audit the interpretive stack: standardisability of Type Ia across cosmic time; intervening physics (dust, lensing, metallicity); the underlying redshift-interpretation stack itself (V1).
2. Modify GR at cosmological scales (f(R) gravity, DGP braneworld, MOND-cosmological extensions, TeVeS, etc.).
3. Postulate a new component (dark energy) to absorb the residual within GR + FLRW.
4. Consider inhomogeneous cosmological models in which the apparent acceleration is an artefact of inhomogeneity rather than a real dynamical effect (Lemaître-Tolman-Bondi, Szekeres).

Option (3) was chosen. (2) and (4) have been pursued as minority programmes. (1) has been pursued only partially, primarily regarding Type Ia systematics, with most attention focused on keeping Type Ia as standardisable.

The dark-energy postulate is the result of a methodological choice among available options, not the result of a direct observation forcing one option over the others.

The “evidence for dark energy” since 1998

8.1 Additional supernova surveys

Larger Type Ia samples at higher redshifts have refined the parameters of dark energy within ΛCDM, without changing the postulate’s epistemic status: the data are fit by allowing Ω_Λ to take a specific value, and that value is reported.

8.2 Baryon acoustic oscillations (BAO)

What is observed: A characteristic angular scale in galaxy clustering corresponding to a spatial scale that, under the framework, is the sound horizon at the epoch of photon decoupling (~150 Mpc comoving).

What is claimed: The measured angular BAO scale as a function of z constrains the cosmic expansion history, supporting a dark-energy component.

Assumptions required:

1. The BAO feature’s physical scale is known from the CMB sound-horizon calculation, which requires the full ΛCDM framework at early times.
2. The distance–redshift relation at the relevant z is compared to the framework’s prediction.
3. Galaxy positions and redshifts are interpreted per V1.

8.3 CMB power-spectrum fit

The CMB fit (§4.3) includes Ω_Λ as one of its parameters. Its extracted value is consistent with the supernova-derived value, under the same framework.

8.4 The Hubble tension as a dark-energy stress point

The Hubble tension (V1, §7.3) is a ~5σ disagreement between H₀ values extracted from the CMB (assuming ΛCDM including dark energy) and from the local distance ladder. Various proposed resolutions invoke modifications to dark energy (early dark energy, evolving w, etc.) as framework patches.

A framework-internal tension that the dark-energy postulate does not resolve, and whose proposed resolutions further extend the postulate.

8.5 Recent DESI results (2024–2025)

Recent baryon-acoustic-oscillation results from the Dark Energy Spectroscopic Instrument (DESI), combined with supernova and CMB data, have been reported as showing ~2–3σ evidence for an evolving dark-energy equation of state (w ≠ −1 and evolving with z), as opposed to a pure cosmological constant.

Epistemic status: If this result firms up, it indicates that even within the framework, the simplest form of dark energy (a cosmological constant) is disfavoured. The response within the framework is to allow w to evolve — i.e., to introduce an additional layer of postulate on top of the dark-energy postulate. The move is structurally identical to earlier moves: observation disagrees with the simplest form of the postulate, a more complex form of the postulate is introduced.

Summary: the epistemic status of dark energy

Dark energy is a residual-defined postulate, introduced in 1998 to preserve the ΛCDM framework against the supernova-distance-modulus observation. Its physical nature has not been established; its value has no derived theoretical prediction; its existence is inferred, not observed.

Dark matter and dark energy as instances of one procedure

Dark matter (1974) and dark energy (1998) are instances of the same procedural pattern. Each follows the same steps:

1. Direct observation is made at Layer 0.
2. An interpretive stack is applied to convert the observation into a framework-internal quantity (rotation curves, distance moduli).
3. The framework’s prediction fails under the interpretive stack.
4. A new entity is postulated with properties defined by the residual between prediction and observation.
5. The new entity has no direct empirical access — no direct detection, no independent measurement, no theoretical prediction of its properties from underlying physics.
6. The new entity’s existence is subsequently “confirmed” by additional framework-conditional coherence checks, each of which operates inside the same framework whose preservation motivated the original postulate.
7. Observations that strain the postulate are absorbed by further extensions (dark-matter subclasses, evolving dark energy, etc.).

This is the procedural pattern. It is the pattern by which a framework preserves itself against counter-observations by postulating residual-absorbing entities rather than modifying the framework or re-auditing the interpretive stack.

The joint claim of the standard cosmological model

The standard model (ΛCDM) claims that the universe’s mass-energy content is approximately:

• ~5 % ordinary baryonic matter
• ~27 % dark matter
• ~68 % dark energy

The ~95 % non-baryonic content — nineteen-twentieths of the claimed universe — is made up of two residual-defined postulates. The direct empirical support for each postulate’s existence (as distinct from framework-conditional coherence) is, as catalogued above:

• Dark matter: zero direct detections after four decades of dedicated experimental search.
• Dark energy: zero direct detections, no derived theoretical prediction, value fitted to supernova observations within the framework.

The standard cosmological model’s inventory of the universe is ~95 % residual-defined postulate. The inventory is framework-conditional. Outside the framework, the postulates do not necessarily exist.

What is empirically backed, and what is not

Empirically backed

• Photon counts at wavelengths at angular positions at times, from instruments pointed at galaxies, clusters, supernovae, and the CMB. (Layer 0.)
• The ratio z = (λ_obs − λ_lab) / λ_lab for identified transitions. (Layer 2, conditional on atomic-physics constancy.)
• The observation that galactic rotation-curve tracers do not show the wavelength pattern expected under Newton + circular orbits + mass-tracer.
• The observation that Type Ia supernovae at z ≈ 0.3–0.8 have apparent brightnesses deviating from matter-only ΛCDM predictions.
• The observation that the CMB has a power spectrum with specific features at specific angular scales.
• The observation of gravitational-lensing distortions in images of background sources.

Not empirically backed, but postulated

• That galaxies are surrounded by dark-matter halos. Residual-defined postulate.
• That dark matter is ~27 % of the universe’s content. Framework-conditional extraction.
• That dark-matter particles exist. Unobserved after four decades of direct search.
• That the universe’s expansion is accelerating. Postulate, conditional on SN standardisation and expansion framework.
• That dark energy is ~68 % of the universe’s content. Framework-conditional extraction.
• That dark energy is a cosmological constant with the observed value. Postulate without independent theoretical derivation.
• That dark matter and dark energy together constitute ~95 % of what exists. Framework-conditional inventory.

The test, applied

A quantity is empirically backed if its value can be determined from the observation without passing through an interpretive layer. By this criterion:

• Galaxy rotation tracers’ wavelength patterns: empirically backed.
• “Galactic dark-matter halo”: not empirically backed; residual-defined postulate.
• Supernova apparent brightnesses: empirically backed.
• “Accelerating expansion of the universe”: not empirically backed; postulate conditional on multiple framework layers.
• “Dark-energy density = 0.68 × critical density”: not empirically backed; framework-conditional extraction.

What became of anomalies, applied to the dark sector

The same relabelling move identified in Volume 1 operates throughout the dark sector:

• Direct-detection experiments find nothing after decades: called “null results” and “improved upper limits,” rather than falsification of the underlying postulate.
• Small-scale structure problems (cusp-vs-core, missing satellites, too-big-to-fail): called “small-scale challenges,” absorbed by adjunct hypotheses (warm dark matter, self-interacting dark matter, baryonic feedback).
• Cosmological-constant problem (~120-order discrepancy with naive QFT prediction): called “a puzzle,” not a falsification of the postulate’s theoretical basis.
• Hubble tension: framework-internal inconsistency absorbed by proposed dark-sector extensions (early dark energy, interacting dark sectors).
• DESI’s hint of evolving w: absorbed by upgrading from cosmological-constant to dynamical dark energy.
• JWST’s mature galaxies at z > 10: “puzzle,” absorbed by proposed accelerated early structure-formation mechanisms.

Observations do not have anomalies. Models have anomalies.

The minimal epistemic commitment

To speak about galactic rotation curves, cluster dynamics, and cosmic expansion without dark matter or dark energy as postulated entities, one can adopt the following minimal commitments:

1. Detector outputs exist. Photons at wavelengths at positions at times.
2. Laboratory physics is reliable. Atomic transitions and calibrations.
3. Atomic physics is approximately constant across the observed space and time.
4. The Layer 0 observations of galactic spectra, supernova light curves, CMB temperature maps, and gravitational-lensing images are what they are — data.

From these commitments, one can report:

• That certain wavelength patterns are observed in the spectra of galaxies and clusters.
• That Type Ia supernovae at certain z have certain apparent brightnesses.
• That the CMB has a specific angular power spectrum.
• That certain lensing distortions are observed around certain foreground mass concentrations.

Everything beyond this — “dark-matter halos,” “dark-energy density,” “27 % and 68 %” — requires additional postulates. None of these is contained in the observations.

The observations are the photon measurements. The rest is what the framework introduced to preserve itself.

Dark matter and dark energy are the two largest components of the standard cosmological model. Together, they constitute ~95 % of the claimed content of the universe. Neither has been directly observed. Neither has a derived theoretical prediction of its properties from underlying physics. Each was introduced at a specific historical moment to preserve a specific framework against a specific observational residual, and each has subsequently been “confirmed” only through framework-conditional coherence checks conducted inside the framework that motivated the original postulate.

This does not prove dark matter and dark energy do not exist. It demonstrates the status of the claims about them: they are residual-defined postulates, not empirically established entities. A reader who sees “dark matter is 27 % of the universe” or “the universe’s expansion is accelerating due to dark energy” and takes these as observations has been misinformed about their epistemic status.

The observations are: photon counts at wavelengths at angular positions at times. From those, one can compute redshifts, rotation-curve tracer velocities, supernova apparent brightnesses, CMB power spectra, and lensing-distortion patterns. Beyond that, the framework takes over, and the framework’s preservation is what defines the postulated entities.

If the framework is correct at all the scales at which it has been extrapolated, dark matter and dark energy may be real entities. If the framework is incorrect at any of those scales, the postulates may be absorbing structure that is not there, in the way that epicycles absorbed the residuals of geocentric prediction and negative-weight phlogiston absorbed the residuals of combustion chemistry. The historical record does not tell us which of these is the case. Only a methodology that strips the assumptions and returns to the direct observations can tell.

That methodology is what this article, and Volume 1, and the MRIPR audit programme, exist to apply.

This article is Volume 2 of the Mathematical Research Institute of Physical Reality’s ongoing audit of the assumption stack of modern cosmology. Every factual claim is traceable to the primary sources cited above. No claim depends on any framework-internal inference. Volume 1 (Redshift) is available separately.