Previous empirical studies of divergent thinking (DT) have measured originality by identifying ideas that are unusual or remote but not necessarily divergent. The present investigation used the same kind of open-ended tasks as previous investigations but operationalized scores to capture truly and literally divergent ideas. To this end, 13 dimensions were identified and used to categorize ideas from six DT tasks. These 13 categories represented a kind of ideational hyperspace and insured that actual divergence of thought was captured. Interitem and interrater correlations indicated that 11 of the 13 hyperspace categories were reliable. Furthermore, the tendency to give a large number of ideas within each category was positively associated with originality and fluency. When fluency was statistically controlled, several of the categories (i.e., impractical, synthetic, breadth, non-natural, infeasible, playful, and remote) were positively related with originality, whereas the complementary categories (e.g., practical, nonsynthetic, depth) were negatively related to the originality of both verbal and figural DT. A composite Literal Divergent Thinking (LiDT) Index, calculated by taking all categories into account, was positively associated with attitudes about originality, even after fluency was statistically controlled. The present effort was the first investigation of LiDT and both limitations and direction for future research are explored. Given the reliability of the new LiDT scores, future research should be conducted to determine whether or not LiDT will predict creative thinking more accurately than the traditional indices of DT.