Many continental growth models have been proposed over the years to explain geological and geochemical data. Amongst these data, the evolution of the 87Sr/86Sr of marine carbonates has been used as an argument in favour of delayed continental growth models and of a Neoarchean pulse in continental growth. This interpretation requires that continental freeboard and continental hypsometry have remained constant throughout Earth's history. However, recent studies suggest that Archean sea levels were higher, and Archean relief lower, than present-day ones. To assess the validity of the evolution of the 87Sr/86Sr of marine carbonates as a proxy for continental growth, we have developed a model that evaluates the co-evolution of mantle temperature, continental hypsometry, sea level, ridge depth, emerged area of continental crust and the 87Sr/86Sr of ocean water as a function of continental growth. We show that Archean sea levels were between ∼500 m and ∼1800 m higher than present-day ones, that Archean mid-oceanic ridges were between ∼700 m and ∼1900 m shallower than present-day ones, and that the Archean emerged land area was less than ∼4% of Earth's area. Importantly, the evolution of the area of emerged land, contrary to that of sea level and ridge depth, barely depends on continental growth models. This suggests that the evolution of surface geochemical proxies for felsic lithologies does not constrain continental growth. In particular, the evolution of the 87Sr/86Sr of ocean water predicted for an early continental growth model is in broad agreement with the 87Sr/86Sr data on marine carbonates when changes in continental freeboard and continental hypsometry are taken into account. We propose that the Neoarchean shift in the 87Sr/86Sr of marine carbonates recorded the emergence of the continents rather than a pulse in continental growth. Since the evolution of other geochemical indicators for felsic crust used as proxies for continental growth is equally well explained by continental emergence, we suggest that there could be no need for delayed continental growth models.