Semi-crystalline polymers can be drawn up to very high draw ratios that involve a complete molecular rearrangement of the chain-folded lamellar morphology into a more or less chain-unfolded fibrillar microstructure. This metamorphosis is likely to take place through an intermediate state of high molecular disorder at a local scale. This is the reason why some authors talk of a strain-induced melting-recrystallization process. In contrast, several structural features occurring at moderate plastic strains are relevant to strictly crystallographic processes. A critical discussion of experimental findings and theoretical approaches is made for pointing out the strength or the deficiency of the different author argumentations. No doubt that both phenomena can occur: crystallographic processes are active at all strain levels whereas meltingrecrystallization is restricted to the post-yield stage accompanied with chain unfolding. Melting-recrystallization is rather a corollary of chain unfolding than a basic mechanism of the plastic deformation. Besides, it may concern only a part of the crystalline phase and can be inhibited under adequate deformation conditions.