An experiment tested the hypothesis that cognitive change resulting from information inputs can be represented as linear motion of concepts in multidimensional space. The theoretical background is reviewed and the mathematical derivation of the hypothesis is given. A set of 15 nations was scaled using Woelfel's Galileo system of multidimensional scaling. Experimental messages were introduced and the posttest interconcept distances compared with those predicted by theory. The crucial partial correlations were low, a failure to confirm the hypothesis. Secondary analyses suggested that the failure may have resulted from inadequate control of message content and failure to ensure that the concepts scaled compose a cognitive domain. The theory made better predictions for a subset of the concepts that might be a domain.