It follows that in proportion as we ascend in generalisation we rise, not into vacuity, as has been said, but into the simple—as also, it must be confessed, into the approximate. The relatively empty concepts (there are none that are absolutely void of content) are the product of a discontinuous generalisation which prevents descent without interruption or omission into the concrete. Of course these are chiefly encountered in the world of pure speculation. They are names representing a knowledge that is incomplete, partial, inadequate, or ill-organised; they correspond not to elimination of what is useless, but to deficit of what is necessary. Having no possible contact with reality, they float in an unreal atmosphere, and are material for a fragile and quickly crumbling architecture. The aim of thinking by concepts is to substitute for complex states, simpler conditions that may be turned and re-turned in every possible sense, in order the better to discover their relations: whereas here, by the nature of things, the unconscious activity, the labor that operates silently in the lower strata, is applied to a soil that is full of faults and fissures, and can but project a false light into consciousness. It has frequently been stated that symbolic thought is thinking by substitution. This formula is admissible only when we recognise that the substitute supposes, nay expects, the _actual_ existence of that for which it is substituted. Substitution is valid in consciousness, but not for the total operation. To sum up in a word: the psychology of abstraction and generalisation, is in a great measure the _psychology of the unconscious_. We have merely studied general ideas in so far as they have an assignable origin in experience, and do not transcend its limits. Are there, as some maintain, notions anterior to any sensory intuition—that can by no means nor effort be derived from empirical data? It is not our part to discuss this question. The thesis—whether or no it be legitimate—is a contention in favor of innate ideas, and in whatever fashion it is conceived (_à priori_ forms, hereditary disposition, cerebral conformation), it is the problem of the ultimate constitution of human intelligence, which we have rigorously eliminated from our present subject. FOOTNOTES [1] _La parole_ is here, and subsequently, translated by _speech_; _le mot_ by _words_, or _language_,—_verbal_ language being throughout understood.—_Trans._ [2] Schmidkunz, _Ueber die Abstraction_. Halle: Stricker, 1889. This little work of forty-three pages contains a good historical and theoretical exposition of the question.