An observational test for the anthropic origin of the cosmological constant

The existence of multiple regions of space beyond the observable Universe ( within the so-called multiverse) where the vacuum energy density takes different values has been postulated as an explanation for the low non-zero value observed for it in our Universe. It is often argued that our existence pre-selects regions where the cosmological constant is sufficiently small to allow galaxies like the Milky Way to form and intelligent life to emerge. At first glance, it would seem necessary to visit regions of space far beyond our current horizon in order to critically examine the validity of this anthropic argument. However, here we propose a simple empirical test for it within the boundaries of the observable Universe. We make use of the fact that dwarf galaxies formed in our Universe at redshifts as high as z similar to 10 when the mean matter density was larger by a factor of similar to 10(3) than today. Existing technology enables us to check whether planets form in nearby dwarf galaxies and globular clusters by searching for microlensing or transit events of background stars. The oldest of these nearby systems may have formed at z similar to 10. Direct observations of dwarf galaxies at redshifts z similar to 10 can be used to characterize their size, mass, metallicity, and star formation history, and identify the nearby systems that descended from them. If planets are as common per stellar mass in these descendents as they are in the Milky Way galaxy, then the anthropic argument would be weakened considerably since planets could have formed in our Universe even if the cosmological constant, rho(V), was three orders of magnitude larger than observed. For a. at probability distribution at the relevant rho(V) values ( which represent infinitesimal deviations from zero in Planck units), this would imply that the probability for us to reside in a region where rho(V) obtains its observed value is lower than similar to 10(-3). A precise version of the anthropic argument could then be ruled out at a confidence level of similar to 99.9%, which constitutes a satisfactory measure of a good experimental test.