The 'Participatory Horizon': Causal Limits and the CMB Low Power Anomaly

The ‘Participatory Horizon’: Causal Limits and the CMB Low Power Anomaly Gregory O’Grady December 21, 2025 Note to readers: This manuscript is shared as an exploratory preprint within an early-stage research programme. It presents a thought experiment supported by phenomenological modelling. The analysis and interpretations have not yet been independently reproduced or peer reviewed. Parts of the computational workflow and manuscript preparation used AI-assisted tools under human oversight. Abstract Several anomalies in the Cosmic Microwave Background (CMB) appear concentrated at the largest angular scales, most notably an anomalous suppression of temperature correlations for separations θ≳60◦. Curiously, the scale at which these correlations vanish corresponds to a comoving distance comparable to the radius of the observable universe. While conventional explanations invoke statistical flukes or tuned primordial initial conditions, we explore an alternative hypothesis motivated by Wheeler’s participatory view of quantum measurement. We propose that classical cosmological records are instantiated only to the extent that correlations can be resolved and irreversibly registered within an observer’s causal domain. In this stance, the CMB is treated as a relational reference frame rather than a fixed fossil record, with correlations on scales comparable to the participatory horizon remaining under-instantiated as classical records, effectively filtering their contribution to the realised sky. We explore this concept phenomenologically by applying a smooth horizon-scale infrared suppression to a fiducial ΛCDM spectrum while leaving the background cosmology unchanged. Fitting to Planck TT bandpowers over 2 ≤ℓ≤30 yields a preferred cutoff scale kcut =α(π/η0) with α≃0.69 (for a representative sharpness p= 6), corresponding to kcut ≃1.53 ×10−4Mpc−1. This result confirms a preference for a suppression scale anchored to the observer’s present horizon. The fit is improved by ∆χ2≃ −5.7 and the large-angle statistic S1/2is reduced by ∼85% relative to fiducial ΛCDM, while the acoustic peak structure at ℓ≳50 remains unchanged. The model makes testable, falsifiable predictions for the low-ℓ EE-mode spectrum and temperature-polarisation cross-correlations. These exploratory results suggest that the low power anomaly might be feasibly reframed as a dynamic participatory effect, providing a quantitatively viable overlay to the ΛCDM model. 1 1 Introduction The standard ΛCDM model provides a robust account of the Cosmic Microwave Background (CMB) and the large-scale distribution of matter [1,2]. However, a cluster of anomalies have persisted at the largest angular scales. The most prominent of these is a deficit of large-angle temperature correlations compared to concordance predictions, referred to as the low power anomaly. In real space, the two-point correlation function C(θ) of the CMB is observed to be anomalously suppressed, effectively vanishing for separations θ≳60◦. This behaviour is captured in the S1/2statistic, which integrates the squared correlation function over these large angular separations [3,4,5]. Simulated CMB skies drawn from the standard ΛCDM ensemble typically exhibit substantially higher correlation on these angular scales, with the probability of obtaining an S1/2value as low as that measured by Planck often estimated to be of order 10−3or lower [3,6]. While this is frequently treated as a statistical fluctuation in the presence of large cosmic variance, the persistence of the feature across sequential mission releases leaves open the possibility that it reflects a real breakdown of standard assumptions at the largest observable scales. Notably, the angular scale at which C(θ) crosses zero corresponds, when projected to the present epoch in comoving units, to a physical distance comparable to the radius of the observable universe. Beyond the power spectrum, other large-scale features remain debated, most notably the so-called ‘Axis of Evil’, an unexplained alignment of the quadrupole and octopole phases with one another and with the geometry of the Solar System [7,8,9,5]. Tensions have also emerged between the CMB kinematic dipole and the cosmic matter rest frame inferred from radio galaxy counts and related tracers [10,11]. Individually, each may be attributable to chance, residual systematics, or a posteriori statistics [12,13,14]. Collectively, however, they motivate ongoing enquiries into explanations that may be physically plausible, with two potential clues being that they predominantly involve the largest angular scales, and might hint at a non-trivial relationship to the observer’s specific context. Proposed resolutions to the low power anomaly typically modify the dynamics or initial conditions of the early universe. Leading scenarios include pre-inflationary phases, departures from slow-roll inflation, or tuned initial conditions that selectively suppress the longest wavelength modes [15,16,17]. Alternatively, an infrared cutoff may arise from non-trivial spatial topology or compact spatial sections, where the allowed wavelengths are restricted [18]. Phenomenological templates employing smooth exponential or hyperbolic tangent cutoffs have also been explored [e.g. 17]. While several constructions improve one or more low-ℓdiagnostics, many require tuning or additional non-standard ingredients, and no single explanation has achieved broad consensus. This paper explores an alternative line of enquiry motivated by a longstanding strand of quantum thought in which the observer and the act of measurement are not treated as external to the description, but as central. John Archibald Wheeler championed this perspective through his ‘Participatory Universe’. Although his proposal remained incomplete, Wheeler repeatedly used ‘visual thinking’ and gedankenexperiments to press the idea that the quantum principle might extend to cosmic scales (Figure 1). Three of Wheeler’s concepts are most relevant here. First, he conceptualised the universe as a kind of ‘self-excited circuit’, in which acts of observation, including those performed 2 Figure 1: Visual thinking diagrams used by Wheeler to illustrate the participatory role of the observer. (A) Wheeler’s ‘U’ diagram: posing the universe as a kind of self-excited circuit, intended to ‘inspire thought’ regarding how observation imparts tangible reality to the early universe (Modified from [19]). (B) The Great Smoky Dragon: a metaphor for the quantum phenomenon between emission and detection. The tail (entry) and mouth (detection) are sharp, but “about what the dragon does or looks like in between we have no right to speak” (modified from [20]). In the present framework, the uninstantiated largescale CMB is treated as such a superposition. (C) Smashing the Glass: the classical view (left) assumes the observer watches the universe safely behind a slab of plate glass. Wheeler’s quantum-first view (right) reminds us that to observe even a photon we must ‘smash the glass’ and install equipment, making the observer a participator with an inescapable effect on the resulting measured history (modified from [21]). on the earliest light, play a necessary part in bringing that universe into being [21,22] (Figure 1A). Second, he distinguished decoherence from a registered outcome, insisting that “no elementary phenomenon is a phenomenon until it is a registered phenomenon. ..brought to a close by an irreversible act of amplification” [21,22]. This was captured in his metaphor of the unmeasured past being a ‘smoky’ superposition, which he labeled a ‘Great Smoky Dragon’ (Figure 1B). Third, motivated by his delayed-choice experiments, he insisted that measurement choices constrain what can be meaningfully said about the past (Figure 1C) [23]. In this view the observer is not separated from reality by a ‘plate of glass’ but is critically involved in defining what is measurable. A recent conceptual framework termed the ‘Measureverse’ (MV) proposes to extend and operationalise this Wheelerian stance in the context of current cosmological anomalies [24]. The core move is to treat measurements performed today not as retrocausal influences but as late-time boundary conditions that actualise specific histories from a prior quantum superposition. This stance motivates a new hypothesis regarding the CMB: instead of a fixed fossil record, it is treated as a ‘relational reference frame’ coupled to any participating detector. The present paper advances this line of enquiry toward an initial quantitative realisation. The central hypothesis is that the local causal horizon acts as an intrinsic information aperture for the observer, limiting the resolution at which long-wavelength modes can be instantiated as distinct classical records. Consequently, suppression is anticipated for the very largest scale correlations (as per the low power anomaly), while the small-scale acoustic structure, which lies well within the horizon and is robustly classical, remains essentially 3 unchanged. We develop this idea phenomenologically by retaining the standard ΛCDM background and encoding the participatory horizon as a smooth, horizon-anchored record response acting on the recorded sky. The resulting modified spectrum is propagated through a standard numerical pipeline to obtain the angular power spectrum Cℓ, the real-space correlation function C(θ), and the associated S1/2statistic. The aim is exploratory, being to evaluate whether an initially minimal, physically anchored participatory horizon model can move the large-angle correlations of the observed sky toward the data without sacrificing the established success of ΛCDM at smaller scales. 2 Conceptual framework: participatory horizon and record instantiation As above, the Measureverse proposal [24] takes seriously Wheeler’s insistence that quantum phenomena are only closed into classical reality by irreversible registration, including light from the earliest universe. In cosmology this implies a shift in emphasis regarding the CMB. Rather than treating the CMB as a fully classical, observer-independent record that simply exists everywhere, the question explored here is what subset of degrees of freedom can be irreversibly instantiated as records for a given observer. 2.1 Observer as irreversible record Wheeler used an intentionally minimal definition of an ‘observer’, maintained here, which was any physical process capable of producing an irreversible record accessible within a causal domain. Conscious human observers are therefore explicitly not required, and in the context of CMB observations, the detector is the observer. 2.2 The participatory horizon as an information aperture For any worldline, the causal diamond defines the region from which signals can reach the observer and within which records can be assembled. The present hypothesis is that this causal domain functions as an information aperture, and only degrees of freedom that can be resolved and redundantly encoded within it can enter the observer’s realised cosmological history as definite records. Long-wavelength modes that vary too slowly across the causal domain, or whose correlations require regions that are causally inaccessible, are underdetermined by the available records and therefore remain effectively uninstantiated in the observer’s realised sky. Figure 2contrasts the standard passive stance with the participatory horizon stance used in the phenomenological model below. 2.3 Decoherence versus record instantiation A standard objection is that inflationary squeezing and environmental decoherence are commonly assumed to classicalise super-horizon modes long before recombination, fixing the 4 (a) Passive view Observer Last scattering surface 0 rec Classical trajectories (b) Participatory view Detector / causal aperture RH (participatory horizon) Underdetermined superposition Record instantiation at the aperture Long-wavelength correlations weakly instantiated 0123 k / k cut 0.0 0.2 0.4 0.6 0.8 1.0 h 2 p =6 (c) Participatory horizon filter Figure 2: Passive and participatory views of the CMB on a conformal spacetime diagram. (a) Passive view: classical photon trajectories propagate from a fixed last scattering surface to the observer. (b) Participatory view: the interior of the light cone is treated as underdetermined until irreversible record instantiation at the detector (causal aperture), with the comoving horizon radius RH≡η0acting as a participatory horizon that limits the effective resolution of the longest-wavelength correlations. (c) Participatory horizon filter: the smooth, horizon-anchored suppression used in this paper, shown as a representative h2(x) transition (here plotted against x≡k/kcut; in the numerical implementation we apply the multipole analogue h2(ℓ/ℓcut) with ℓcut ≃kcutη0). large-scale pattern independent of any late-time observer. The present framework adopts a stricter Bohr and Wheeler inspired criterion. Decoherence is necessary to define robust pointer states,[25] but is not, by itself, accepted here as selecting a single realised history unless an outcome is irreversibly registered into records accessible within the observer’s causal domain. Information that is redundantly encoded only in regions that never enter the observer’s horizon is therefore, by the current proposition, not part of that observer’s physical record. In operational terms, it remains unresolved for that observer. 2.4 Minimal phenomenological requirement A participatory horizon effect should therefore appear as a selective suppression of the longest-wavelength correlations, with a characteristic scale tied to the horizon. The simplest phenomenological implementation, as explored here, is an isotropic, smooth infrared filter that transitions from strong suppression at the very lowest multipoles to negligible effect by ℓvalues that correspond to sub-horizon acoustic structure. In this minimal implementation, the filter is characterised by a dimensionless scale parameter αthat sets the effective cutoff relative to the fundamental horizon mode kfund ∼π/η0, and a sharpness parameter pthat controls how quickly the transition occurs. 2.5 Scope The current model is used as a quantitative feasibility test, not as a completed microphysical theory of quantum measurement in cosmology. In addition, the initial implementation is 5 deliberately isotropic. Later work plans to evaluate how a relativistically moving causal diamond can become slightly anisotropic, providing a natural route to explore alignments such as the Axis of Evil. 3 Phenomenological model and numerical methods 3.1 Baseline spectrum and horizon scale We take a fiducial ΛCDM background consistent with Planck 2018 parameters [2] and compute the corresponding unlensed CMB spectra using the camb Boltzmann pipeline [26]. We denote the conformal age of the universe by η0. The fundamental horizon scale is defined as: kfund ≡π η0 ,(1) which provides the physically motivated reference scale for an infrared suppression anchored to the present horizon. We work with the conventional bandpower quantity Dℓ≡ℓ(ℓ+1) 2πCℓ for direct comparison to Planck TT data. 3.2 Participatory Horizon response We model the impact of incomplete horizon-scale records through a smooth response function, representing the ‘participatory horizon’, which treats the local causal diamond as an intrinsic measurement boundary. Under this hypothesis, modes larger than this scale contribute only weakly to the realised sky because the corresponding macroscopic record is unresolved. We implement this as a smooth, scale-dependent suppression characterized by a dimensionless scale parameter αand sharpness parameter p: kcut ≡α kfund =απ η0 .(2) The functional form adopted for the record response is a smooth transition: hk(k;α, p)≡1−exp−k kcut p, kcut ≡απ η0 .(3) The present implementation applies the corresponding multipole analogue hℓ(ℓ;α, p)≡1−exp−ℓ ℓcut p, ℓcut ≡kcutη0=απ , (4) and propagates the suppression to the angular spectra as DMV ℓ(α, p)=DΛCDM ℓh2 ℓ(ℓ;α, p),(5) applied for XY ∈ {TT, EE, TE}. Rather than fixing p, we treat it as a robustness hyperparameter and scan representative values (p= 4,6,8). 6 3.3 Low-ℓdiagnostic fit The fit parameter αis obtained by minimizing a diagnostic low-ℓstatistic over the range 2≤ℓ≤30: χ2(α) = 30 X ℓ=2 DPlanck ℓ−DMV ℓ(α, p) σℓ2 ,(6) using the published Planck bandpower uncertainties in diagonal form. This serves as a controlled diagnostic to identify the best-fit suppression amplitude within a restricted range of multipoles, with a full likelihood analysis reserved for future work. 3.4 Angular correlation function and S1/2 To connect to real-space anomalies, we compute the angular correlation function via a standard Legendre expansion: C(θ) = ℓmax X ℓ=2 2ℓ+ 1 4πCℓPℓ(cos θ),(7) and evaluate the large-angle S1/2statistic: S1/2≡Z180◦ 60◦ [C(θ)]2sin θ dθ . (8) We also construct a Planck-derived reference value by applying the same estimator to an interpolated reconstruction of the TT bandpowers. 3.5 Polarization as a prediction Because the participatory horizon is defined at the level of the causal structure, the response should generically affect all CMB fields. We therefore propagate the fitted (α, p) parameters to the fiducial EE and TE spectra. These resulting low-ℓcurves serve as a falsifiable prediction for future polarization tests. 4 Results 4.1 Best-fit suppression scale and robustness Fitting the participatory horizon filter to Planck TT bandpowers over 2 ≤ℓ≤30 yields a preferred cutoff scale close to the horizon fundamental mode. For a representative sharpness choice p= 6, the best-fit parameter is αbest ≃0.69 ±0.19 ,(9) corresponding to a cutoff kcut =απ η0 ≃1.53 ×10−4Mpc−1, ℓcut ≃απ ≃2.17 .(10) 7 Table 1: Robustness scan over the sharpness parameter p. The diagnostic χ2is evaluated for 2 ≤ℓ≤30 using Planck TT bandpowers with diagonal errors. p αbest ℓcut =απ ∆χ2SMV 1/2[µK4] 4 0.710 ±0.235 2.23 −5.69 4065 6 0.690 ±0.190 2.17 −5.70 5156 8 0.670 ±0.175 2.11 −5.69 5164 Figure 3: Temperature power spectrum. Left: full range of multipoles with Planck TT bandpowers. Right: zoom on the lowest multipoles (fit range 2 ≤ℓ≤30). The MV curve applies the participatory horizon filter to the fiducial ΛCDM spectrum for the representative case p= 6 and best-fit α≃0.69. The Planck points are shown with diagonal bandpower uncertainties, consistent with the diagnostic χ2definition used in the fit. The baseline diagnostic fit for the unsuppressed fiducial ΛCDM spectrum is χ2 base = 27.54 for 29 multipoles. Introducing the single suppression parameter yields χ2 best ≃21.85, an improvement of ∆χ2≃ −5.7. A scan over the sharpness parameter pshows that the improvement is stable. ∆χ2is essentially unchanged while αshifts mildly to keep ℓcut ∼2.1–2.2. Table 1summarises these results. The weak dependence on pindicates that the diagnostic improvement is driven primarily by the location of the transition scale rather than detailed tuning of the filter shape. We note that S1/2is especially sensitive to the detailed suppression of the quadrupole and octopole, so modest shifts in the transition shape at fixed ∆χ2can produce noticeable changes in S1/2across the pscan. 4.2 Temperature spectrum Figure 3shows the full TT spectrum and a zoom of the lowest multipoles for the representative choice p= 6 with α≃0.69. The modification is confined to the very lowest multipoles and leaves the acoustic peak structure at ℓ≳50 essentially unchanged. In the low-ℓpanel we show the Planck bandpowers with their (diagonal) uncertainties for reference, using the same bandpower errors as in the diagnostic χ2fit. 8 Figure 4: Low-ℓdiagnostics as a function of the participatory horizon scale parameter αfor the representative choice p= 6. Left: diagnostic χ2(α) evaluated over 2 ≤ℓ≤30. Right: large-angle statistic S1/2(α) computed from the corresponding correlation function C(θ). The preferred region that improves the low-ℓfit also drives a strong reduction of large-angle correlations. 4.3 Low-ℓdiagnostics as a function of α The suppression scale is controlled by α, and the two key low-ℓdiagnostics respond differently as αvaries. Figure 4shows the diagnostic χ2(α) over 2 ≤ℓ≤30 and the corresponding S1/2(α) for the representative sharpness choice p= 6. The χ2(α) curve exhibits a broad minimum near α≃0.69. The quoted uncertainty is obtained from the standard diagnostic criterion ∆χ2= 1 around the minimum. The S1/2(α) curve falls rapidly once the filter begins to suppress the lowest multipoles, and then flattens as the suppression saturates for ℓ≳ℓcut. The key point is that the range of αthat improves the low-ℓfit also produces a substantial reduction of large-angle power, consistent with the notion that both signatures are driven by the same horizon-linked suppression of the longest wavelength modes. 4.4 Angular correlation function and S1/2 Figure 5shows the corresponding real-space correlation function C(θ). The fiducial ΛCDM curve exhibits large positive correlations over θ≳60◦, while the participatory horizon filter strongly suppresses these large-angle correlations. For p= 6 and α≃0.69, the large-angle statistic is reduced from SΛCDM 1/2≃3.46 ×104µK4to SMV 1/2≃5.16 ×103µK4, a reduction of ∼85% relative to fiducial ΛCDM. A Planck-derived estimate computed from the same TT bandpower file yields SPlanck 1/2≃6.7×103µK4, so the suppression moves the model into the observed low-correlation regime. 4.5 Low-ℓpolarisation predictions Because the filter suppresses the longest wavelength modes rather than introducing a temperatureonly template, the same large-scale suppression must propagate to polarisation. Figures 6 and 7show the predicted low-ℓEE and TE spectra for the representative case. This provides a direct and falsifiable observational discriminator for the current hypothesis. If future measurements recover low-ℓpolarisation consistent with unfiltered ΛCDM, the participatory horizon interpretation of the temperature anomaly would then be strongly disfavoured. 9 [32] John Archibald Wheeler. Personal journals 1980–2006. American Philosophical Society, John Archibald Wheeler Papers, 2006. Manuscript collection. 16