In this paper it is championed that a two stages hypothesis for the evolution of language must take into account a qualified approach to heterochrony and the available information from the archaeological record. As it seems, a protracted childhood and youth was already at work in Homo erectus, but early postnatal brain growth was only available to Homo sapiens. According to these facts, the term hysteron proteron is given here to the reversal that sets off linguistic capacity during the first (...) three years of life, upon the ancient abilities for social, natural and technological knowledge, which were already established during the last 2 Mya BP. The unfolding of linguistic capacity creates the typical effect that ontogeny reverses phylogeny: since late postnatal brain growth, quite recent in evolution, launches in ontogeny before the phylogenetically old processes of playing and learning take place during childhood and youth. Full symbolic abilities may be explained as emerging from older semiotic competences based on iconicity and indexicality, which would have presumably sustained human life for millennia. Furthermore, it is argued that the formation of linguistic categories follows from that overlaying of extended connectivity linked to early postnatal brain growth, inasmuch as it entangles and transforms previous iconic and referential information: i.e. the ancient semiotic capacities for proprioception, embodiment, extended mimesis and pragmatic knowledge, shaping a sort of self-organization process that provides the formal and semantic properties we find in articulated languages. (shrink)