We revisit our argument that shows that parton distribution Functions (PDFs) in the MSbar{ scheme are non-negative in the perturbative region, with the main goals of elucidating its domain of validity and clarifying its theoretical underpinnings. We specifically discuss recent results proving that PDFs can turn negative at sufficiently low scale, we clarify quantitatively various aspects of our derivation of positivity in the perturbative region, and we provide an estimate for the scale above which PDF positivity holds.
Comment: 19 pages, 1 figure. Final version, to be published in EPJC. Sections 3-4 merged and streamlined; introduction and conclusion improved and clarified; typo in Eq.(45) corrected