The MOND acceleration scale a0 ~ 1. 2e-10 m/s², below which galaxy dynamics depart from Newtonian gravity, numerically coincides with the cosmological quantity c*sqrt (Lambda). This paper separates, and proves, three distinct claims about that coincidence. (1) A reframing (modest, not original): the scale is written as a horizon surface gravity, a0 = c²*sqrt (Lambda/32pi) = (c/2) *sqrt (G*rhoLambda) = c*HLambda/Z with Z = sqrt (32pi/3) ~ 5. 789 -- algebraic identities checkable in minutes. (2) A mechanism for the functional form (established physics): the deep-MOND square-root law, the Radial Acceleration Relation, and the Baryonic Tully-Fisher Relation follow from treating inertia as a response to the de Sitter-Unruh temperature of the cosmic horizon (Milgrom 1999; Deser and Levin 1997). (3) Rigorous limits (the new content): the O (1) coefficient Z is NOT derived; the temperature route that yields the shape predicts a scale ~12x too large and, as a covariant action, the WRONG SIGN (anti-MOND) -- stated as a passivity theorem; and NO covariant modified-inertia completion exists in three exhaustive cases -- local (an Ostrogradsky ghost), field-theoretic (a Cassini-violating metric coupling), and nonlocal (the passivity -> anti-MOND sign theorem: for any causal, ghost-free, unitary cosmological source the low-acceleration inertia shift deltaₘ = 2*Integral (rho/omega²) >= 0, raising inertia, so the MOND sign requires a negative-norm ghost). The defensible conclusion is that a0 ~ c*sqrt (Lambda) is a FORCED SCALE, while the MOND sign and the precise normalization are POSTULATED, not derived. This is a suggestive geometric/thermodynamic reframing of the MOND constant with sharply proven boundaries -- NOT a complete theory of MOND, and NOT a theory of everything. The one observationally distinctive consequence -- a declining scale, a0 (z=3) ~ 0. 74*a0 (0), tracking the dark-energy density -- is falsifiable this decade. This manuscript intentionally claims LESS than, and corrects, an earlier over-stated description of the same material by the author. All symbolic and numerical checks are reproducible from the public repository.
Carl P. Zimmerman (Sat,) studied this question.