208 ARTHUR O. LOVEJOT : ing and making his own Eberhard's classification of the ways in which predicates may be related to subjects a priori Kant even proposes an emendation of the technical phraseology, which has the effect of bringing out all the more clearly the principle that, in attributa, we have predicates which do not belong to the definition of a concept, and yet cannot be negated of it without destroying its logical possibility. The passage runs as follows : " A predicate which is ascribed to any subject by an a priori judgment is thereby declared ne- cessarily to belong to that subject, or to be inseparable from it. Such predicates are said to belong to the essence, or the inner possibility, of the concept (ad internam jjossibilitatem pertinentia). Consequently all propositions that profess a priori validity must have predicates of this character. All other predicates (those, namely, which are separable from the concept without destroying it) are called extra-essential marks of the subject (ausserwesentliche Merkmale, extraessentialia). Now predicates of the first sort may belong to the essence in two ways, either as constitutive elements of it (Bestandstiicke, constitutiva), or as consequences of the essence that have in it their sufficient ground (ut rationatd). The former are called essential elements (itesentliche Stiivke, essentialia), and, as such, contain no predicate that could be deduced from any other predicate contained in the same concept ; and in their totality they make up the logical essence. The latter are called properties (Eigenschaften, attributa)." Kant, then, here grants that there may be predicates, technically to be named attributa rationata, which form no part of the essentialia of a concept, as it is defined, but yet are indispensable to the interna pos- sibilitas of it. This is merely another way of saying which is the substance of the Wolffian theory that not only tauto- logical judgments but also judgments per attributa, are legiti- mate sources of a priori knowledge. And the whole realm of metaphysics, according to Wolff and (in his real intention) according to Leibniz, lay within the limits of this last- mentioned class of judgments, the possibility, validity and potential fruitfulness of which Kant now 7 admits. 1 1 Paulsen has summarised this rather important part of the Reply to Eberhard in a footnote in his volume on Kant in the Kbissik-r <(tr Philo- sophie (translation of Creighton and Lefevre, p. 143). Since the book is widely read, it is perhaps worth while to point out that here, as in several other instances, Paulsen goes somewhat seriously astray in his exposition. For he represents Kant as classifying predicables of the sort called attributa or rationata, among the " extraessrntitth'a, which can be separated from the concept without affecting its nature". But Kant plainly and repeatedly classifies attributa as one of the two kinds of predicates that are ad internam possibihtatem pertinentia, inseparable from
