Psychological Bulletin A HISTORY OF INTROSPECTION

ed from and the physicist has resort to measurement and other physical technics. Titchener held to this distinction by point of view all his life (85, pp. 259-268). It was Kiilpe who split Wundt's psychological atom, analyzing sensation into its four inseparable but independently variable attributes: quality, intensity, extensity, and duration (41, pp. 30-38). Titchener later held to this view which served to tighten rather than to loosen the constraints of atomism upon introspective psychology (6, pp. 17-35). One of the most thorough discussions of introspection was provided by the erudite G. E. Muller in 1911 (55, pp. 61-176). Muller was more liberal than Wundt and left room for all the indirect and retrospective forms of introspection. Being primarily interested in the application of introspection to memory, he distinguished, for instance, between the present recall of the past apperception of a past event and the present apperception of the present recall of a past event, an important distinction, since present apperception can be interrogated as to detail whereas past apperception has become fixed and no longer subject to exploration. It was Titchener who placed the greatest constraints upon introspection by his requirement that the description of consciousness should exclude statements of meaning. At first Titchener had perception in mind and called the report of meanings the stimulus-error, insisting that trained observers by taking the psychological point of view would describe consciousness ("dependent experience") and attempt no statements about the stimulus-objects ("independent experience" as given by the point of view of physics) (5; 79, pp. 202 f.). After Kiilpe had claimed to find imageless (nonsensory) thoughts in the consciousnesses of judgment, action, and other thought processes, Titchener broadened his criticism to an objection against the inclusion of any meanings at all in the data of introspection (80). He was arguing that straight description (Beschreibung, cognitio rei) would yield the kind of sensory contents that had become standard in classical introspection, and that inferences about conscious data (Kundgabe, cognitio circa rem) are meanings which do not exist as do the observed sensory processes (81,