"Only 5% of the matter and energy in the universe is known; 25% of an unknown dark matter (6 times more abundant than ordinary matter) and 70% of a mysterious dark energy are missing from the inventory." These proportions are robust and relate to average values taken from a variety of objects and locations. We propose a simple solution to these challenges by considering the universe as a refractive medium. To this end, if the current value c0 of the "speed" of light is correct locally, for example in the "vacuum" between the Sun and Earth, its value on the cosmological scale (a few light-years to billions of light-years), non-empty, is postulated to be cc = c0 / α, with α a factor greater than 1. We extend an earlier work where the previous proposition was stated in detail, and answer the remaining open question of explaining the reasons for the refractivity. We can now say: it is gravitational in origin. Using the Schwartzchild metric, we can calculate an optical index nc equivalent to a mass density filling the universe. Remarkably, we obtain a range of indices covering the value 2.4 for a set (density, to be counted in 10^(-27) kg/cubic meter; "radius" of the universe to be counted in tens of billions of light-years) in a range accepted today (e.g. nc = 2,4 for universe density = 6,4 10^(-27) kg/cubic meter, and universe radius = 33.10^9 l.y.). The α factor (reflected by the optical index nc) alone opens the discussion to a variety of problems, disconnected a priori from each other. The value α ≈ 2.4 accounts for the observed velocity differences for celestial objects; its square α^2 ≈ 6 gives the ratio of dark matter to baryonic (or ordinary) matter, its power 4, i.e. α^4 ≈ 36, the ratio of dark energy to ordinary matter. These powers are derived from physical reasoning using Newton's laws and Einstein's equations. Dark matter and dark energy are the names of the corrections to compensate for the error made in keeping for the speed of light at cosmological scales its "usual" value. Our proposal is motivated by a critical analysis of the way current physics treats the value of the speed of light c, pointing out that we only know v/c type velocity ratios. These results are consistent with the so-called relational epistemology, according to which we do not know things in themselves, but only by the links between them. One consequence is that the universe could be a factor α older than the value announced today.