Autor: A. Jaster | Datum: 2026/06/12 | DOI: 10.5281/zenodo.20667781
We develop a purely cosmological model in which the cosmological constant is replaced by a causal --- that is, retarded-source --- \emph{constraint pressure} built from the retarded matter distribution, resting on two axioms of emergent causal time: time couples to matter, and causality as a construction principle (all local sources are built retarded, on the past lightcone); the homogeneous background closes as a causal system, while full perturbative causal closure remains open. For the late, classical, homogeneous background no quantisation or causal smearing is required, and we work without it. The minimal realisation --- a bare, volume-filling retarded bulk measure --- introduces \emph{no additional equation-of-state parameter}: the \emph{shape} of $w(z)$ is fixed, with no crossing of $w=-1$, deepening toward $-4/3$ at high redshift and relaxing to $-1$ in the de~Sitter future; numerically $w_0\simeq-1.20$ at a fiducial $\Om=0.30$ and $w_0\simeq-1.19$ at the model's own best-fit $\Om=0.265$. Throughout we distinguish the fiducial $\Om=0.30$ (used to isolate the dark-energy effect) from the fit-preferred $\Om=0.265$ (used for all data-facing statements). Two structural problems are addressed: the vacuum-energy fine-tuning is sidestepped because the constraint couples only to real matter, and the coincidence is structurally mitigated: the \emph{time dependence} $\rhoC/\rho_m\propto a^4$ follows from the constraint structure, while the present ratio is calibrated through $\muB$. Energy conservation requires a non-local companion $T^{\rm nl}_{\mu\nu}$; its local sources are retarded, while strict causality of the full sector is \emph{expected} from an in-in construction that has not yet been carried out explicitly. Within the assumed companion limits its effect on the background is bounded: the companion can deepen $w_0$ by at most $\approx0.04$. A Dirac analysis establishes kinematic ghost-freeness --- the constraint sector adds no propagating degree of freedom --- while the deeper field-theoretic health is left open. Linear perturbations are lightcone-suppressed, so the constraint sector is smooth on structure scales: to leading order in this suppression $G_{\rm eff}=G$ and the gravitational slip $\eta_{\rm grav}=\Psi/\Phi=1$ are predicted, and dark matter is \emph{not} eliminated. At fit-consistent parameters the growth signature is a \emph{pattern}, not a uniform boost: $f\sigma_8$ enhanced by $\sim6\%$ at $z\gtrsim0.5$ but slightly suppressed at $z=0$, with $S_8\approx0.79$ --- comparable to $\Lambda$CDM and \emph{not} a second tension, though this relief itself rests on the low $\Om$ preferred by the otherwise poor BAO fit. Confronted with DESI DR2 baryon-acoustic-oscillation data plus Planck distance priors, the model is strongly disfavoured relative to $\Lambda$CDM, $\Delta\chi^2\simeq28.5$ at equal parameter count ($\Delta{\rm AIC}=\Delta{\rm BIC}=\Delta\chi^2$), the conflict residing in the BAO distances --- the committed kernel is thus likely already strongly disfavoured by current geometry --- while it remains compatible with the CMB within the compressed-prior approximation and drives $H_0$ upward. The single amplitude $\muB$ may acquire a physical value, and the background a falsifiable modification, through time-bubble nucleation. The model is sharply falsifiable on independent fronts --- expansion, growth pattern, and gravitational slip --- which is the point of a fixed-shape prediction.