Adult heritage re-learners of their childhood language hold selective benefits over other L2/L3 learners as to phonetics/phonology, but miss a global gain at re-learning their L1's prestige variety. Morphosyntactic transfer only takes place from the dominant language. Possible clarifications for this asymmetry bear consequences for L3/Ln learning models: the Cumulative Enhancement Model, the Typological Proximity Model, and the L2 Status Factor Model. The here debated data mostly comply with the latter model, but they also point out the interest of the typological distance between learned languages.