First, to cite the (Meiklejohn) version of the argument:
If understanding in general be defined as the faculty of laws or rules, the faculty of judgement may be termed the faculty of subsumption under these rules; that is, of distinguishing whether this or that does or does not stand under a given rule (casus datae legis). General logic contains no directions or precepts for the faculty of judgement, nor can it contain any such. For as it makes abstraction of all content of cognition, no duty is left for it, except that of exposing analytically the mere form of cognition in conceptions, judgements, and conclusions, and of thereby establishing formal rules for all exercise of the understanding. Now if this logic wished to give some general direction how we should subsume under these rules, that is, how we should distinguish whether this or that did or did not stand under them, this again could not be done otherwise than by means of a rule. But this rule, precisely because it is a rule, requires for itself direction from the faculty of judgement. Thus, it is evident that the understanding is capable of being instructed by rules, but that the judgement is a peculiar talent, which does not, and cannot require tuition, but only exercise. This faculty is therefore the specific quality of the so-called mother wit, the want of which no scholastic discipline can compensate. [emphasis added]
It seems that if one formalized the emphasized passage in modern logical parlance, one might end up with something very much like a "diagonal argument" regarding the "faculty" of judgment, or then (in modern terms) we would be "diagonalizing" over some relevant propositional function (c.f. Kant's own functionalism about logic in general, then). Is Kant anticipating our recognition of the difference between lower- and higher-order logic, or perhaps the object-/meta-language distinction, or any of these sorts of things? Even if he is, was he himself anticipated somewhere in scholasticism or antiquity, i.e. is he citing an argument he read about somewhere else, or at least might he be citing something else prior to his own analysis?