Dark Energy Is a Rate, Not a Substance: Time Runs Faster in Emptier Space

Abstract

Black holes mark the slow end of the clock. Voids mark the fast end. The Bulmash Record-Carrier Expansion Law puts that axis on the sky—fixed before the likelihood is opened, with no continuous shape left to tune—and is confronted with DESI DR2 and Pantheon+. Dedicated to the memory of my father, Professor Samuel Barak Bulmash, Professor of Finance at the University of South Florida—a devoted teacher and researcher whose intellectual curiosity, independence, and determination to pursue difficult questions continue to inspire this work. Abstract Dark energy is not a substance. It is a rate. Time runs faster in emptier space. Everyone knows clocks crawl near a black hole. The cosmological extension to differential clock rates between walls and voids is not new: it has been developed in inhomogeneous and timescape cosmology. The distinct step tested here is to fix that direction to one record-carrier response law before any likelihood is evaluated. In the record-time construction, the local clock rate is d tau/dt = sqrt(1-rho_hat), with rho_hat the committed fraction of local record capacity. Under the declared compactness reading rho_hat = r_s/r, this is exactly the exterior Schwarzschild factor—frozen at the horizon and standard in the weak field. It is set by gravitational depth, not by matter physically sitting beside the clock; the exterior may itself be vacuum. The same bounded clock axis, read where compactness is low, runs toward its opposite end: emptier regions keep more record capacity unwritten and tick faster relative to denser regions. Black holes approach the slow, saturated endpoint. Cosmic voids occupy the fast, open-capacity side. The claim tested here is a specific record-carrier realization of that axis—not a second invisible fluid. For billions of years, matter has gathered into walls and clusters while voids have taken an increasing share of cosmic volume. More of the universe has therefore moved onto the fast side of the clock field. Distance and redshift measurements do not discover a new substance filling space; they register a rising cosmic-average clock rate as acceleration. What the field has called dark energy is tested here as the expansion imprint of fast time in underdense space. The slow-wall/fast-void direction is established prior work, particularly in Buchert backreaction and Wiltshire's timescape cosmology. Recent timescape analyses reconstruct the Pantheon+ light-curve likelihood and report evidence over flat LambdaCDM. This paper tests a different object: a frozen analytic record-carrier response, jointly confronted with the full DESI DR2 BAO covariance and the released Pantheon+ distance covariance, with a non-nested calibration and fixed jerk/snap falsifiers. See Seifert et al. (2025) and the related-work discussion in the manuscript. The Bulmash Record-Carrier Expansion Law makes that claim quantitative and falsifiable. With load u = sqrt(2) ln a and fast-side fraction F(a) = a^(sqrt(2))/(1+a^(sqrt(2))), the normalized binary-entropy deficit and orientation are Delta = 1-H_2(F)/ln 2 and chi = 2F-1. The resulting dark-sector history is fixed: R_B(a) = 1 + 4 Delta + 8 sqrt(2) Delta(1-Delta) chi, R_B(1) = 1. It enters the background expansion through E_B^2(a) = Omega_r a^(-4) + Omega_m a^(-3) + Omega_de R_B(a). This is a declared record-carrier construction implementing the rate picture at background level—not a theorem that vacuum general relativity alone forces this response law. Within that construction, the even amplitude A = 4 is the midpoint Fisher stiffness, the load slope p = sqrt(2) is fixed by the binary carrier, and B = 8 sqrt(2) follows from one declared relative normalization c_3/c_2 = 2, locked before any likelihood is opened. No continuous coefficient in R_B(a) is fitted to the sky. The shared matter density Omega_m, acoustic scale, and supernova intercept remain ordinary cosmological and calibration quantities and are treated identically for this history and for flat LambdaCDM. The rate claim concerns physical generation, not cosmological bookkeeping. The background integrals still use an effective dark-sector density, and any phantom crossing in that effective description is not a fundamental phantom substance or scalar field. Where a phenomenological fluid parameterization fits an equation-of-state history, this construction derives one rate history. The missing shape parameter is not a modelling convenience. It is the mathematical fingerprint of the category change—from an independently evolving substance to a derived rate of time. Using the full released covariances, the fixed curve is confronted with the public 13-component DESI DR2 baryon-acoustic-oscillation likelihood and Pantheon+. At equal cosmological parameter count, Delta chi^2 = -4.727 relative to flat LambdaCDM. A frozen 200,000-realization non-nested Monte Carlo gives p_null = 0.01444 and Z = 2.185 sigma. Allowing each null realization its best of three predeclared carrier branches gives p_null,branch = 0.02025 and Z_branch = 2.049 sigma. This is a calibrated preference, not a detection, and that evidence grade applies to this distance-data comparison alone. BAO geometry drives the separation. Pantheon+ is compatible with both histories and supplies no independent shape detection. The preference survives every single-observable BAO deletion. The curve does not receive a second chance in its derivatives. It predicts j_0 = 1.9883 and s_0 = +8.9954, against j_0 = 1.0002 and s_0 = -0.4187 for flat LambdaCDM at the same matter fraction—nearly double the jerk, with the snap reversed in sign and more than an order of magnitude larger in magnitude. Nothing in the response can be retuned to hide that difference. Future cosmography will not be asked to confirm this law. It will decide it. If the predicted jerk and snap are absent, the fast-time expansion law fails.