Paul C. W. Davies, João Magueijo

Although cosmic acceleration is well established, its physical origin remains contentious. Most explanations invoke either a non-zero vacuum energy, i.e. a cosmological constant, or new fields. We propose instead a mechanism arising purely from quantum mechanics, without additional constants or local sources. The key point is that quantum theory permits both initial and final boundary conditions on a state, here the wave function of the universe. Pre- and post-selected systems are familiar in laboratory quantum mechanics; we extend the idea to minisuperspace quantum cosmology. As a warm-up, we show that a free non-relativistic particle, initially in a semiclassical wave packet and conditioned on a final quantum state, can have an intermediate peak which accelerates even though the Hamiltonian is free. A semiclassical observer would infer a contrived classical force. We then implement the analogous construction in quantum cosmology using connection variables and unimodular time. A forward semiclassical packet, taken for simplicity to describe pure radiation, is post-selected by a normalizable Chern-Simons soliton. The resulting two-boundary amplitude has a peak which leaves the radiation trajectory and enters an accelerating regime, while the forward Hamiltonian has $Λ=0$. A classical model can mimic this trajectory only by introducing a contrived effective component: near the transition it resembles $w\simeq -1$, but when extrapolated it evolves towards strongly phantom behaviour, $w<-1$. The acceleration is therefore more naturally interpreted as a quantum boundary-condition effect than as a local classical source. We also discuss why the Chern-Simons soliton is a clean final state, the small overlap between initial and final states, and possible tell-tale signatures.