Taking Stump's (2016) PFM2 C(ontent)/F(orm)/R(ealized) paradigm distinction I argue that the F/R-paradigm features are conceptually different from C-paradigm features. C-paradigm features interface with syntax/semantics, hence are 'interpretable'. F-paradigm features, by contrast, induce purely formal (morphomic) partitionings (cf. Boyé & Schalchli 2019), even for canonical systems (one:one Content-Form correspondence), a reflection of the true autonomy of inflectional morphology, 'morphology-by-itself '. The C-paradigm features are a subset of Sadler & Spencer's (2001) 's(yntactic)-features'. Canonical Content-Form correspondence is achieved by typing features as m- and s-features.