Participatory horizons and the cosmic dipole anomaly

The observed dipole in extragalactic source number counts is approximately twice as large as the kinematic expectation derived from the CMB rest frame, a discrepancy that has persisted across radio, infrared, and mid-infrared catalogues and remained contested despite extensive investigation of local structure, masking, and survey-systematic effects. Most discussions treat the excess as evidence for an unknown source-side effect, a violation of the cosmological principle, or a statistical fluctuation. Here, we explore an observer-relative alternative motivated by Wheeler's participatory view of measurement. Building on companion analyses in which the observer's causal diamond was modelled as an information aperture (isotropic for the low-power anomaly, quadrupolar for the Axis of Evil) we extend the participatory-horizon framework to the dipolar ($L = 1$) sector of the recording kernel. In the additive formulation, the kernel's first Legendre mode couples to the dominant source-count monopole via Gaunt integrals, injecting a constant excess $\Delta D_\parallel = \mathcal{A}\beta$ into the observed dipole parallel to the observer's velocity $\hat{\beta}$, where $\mathcal{A}$ is an aperture amplification factor. Applied to the CatWISE2020 photometric quasar catalogue (${\sim}\,1.36$ million sources over ${\sim}\,47\%$ of the sky), the pipeline yields $\mathcal{A} = 5.3 \pm 1.3$ at the fiducial threshold ($W1 \leq 16.4$\,mag), a $4.0\sigma$ excess above the kinematic-only model. The corresponding dipole gain is $G = D_\parallel / D_{\mathrm{kin}} = 1.86 \pm 0.21$; within the additive framework the full $\hat{\beta}$-projected excess is accounted for, reducing the ${\sim}\,4\sigma$ tension along the velocity axis to a single fitted parameter. A cumulative magnitude-threshold scan finds $\mathcal{A}$ stable over the high-purity range ($W1 \leq 16.1$), with a mean of ${\simeq}\,6.7$, declining only where catalogue contamination is expected to dominate the sample; in contrast to the monotonic dilution predicted by local-structure models. A 30-configuration robustness grid confirms insensitivity to neighbour-buffer depth and ecliptic model, with the Galactic latitude cut as the dominant systematic. The transverse component $D_\perp$ is small at bright thresholds and grows with catalogue depth, compatible with an on-axis additive excess from an axisymmetric kernel and a separate threshold-dependent contaminant. The framework makes a parameter-free cross-survey prediction that the absolute excess $\Delta D_\parallel$ should be common to all source populations, so that surveys with different Ellis--Baldwin prefactors should exhibit predictably different fractional excesses. Spectroscopic redshift slices from DESI and forthcoming surveys can test this directly, since a constant $\mathcal{A}$ across redshift bins would be difficult to explain by any source-side mechanism. Together with the companion analyses of the CMB power deficit ($L=0$) and the Axis-of-Evil alignment ($L=2$), these results extend the participatory recording kernel to a third anomaly channel, the source-count dipole tension, conventionally discussed outside the CMB large-angle literature. A single kernel $W(\mu)=1+\epsilon_1 P_1(\mu)+\epsilon_2 P_2(\mu)$, applied to a standard $\Lambda$CDM background, can feasibly accommodate all three anomalies through successive Legendre sectors; several significant open problems are discussed.