Behaving in novel ways is essential to the development of the types of complex performances described by the term creativity, problem solving, and perseverance. Some research suggests that response variability is an operant and a critical component of novel behavior. However, other account of novel behavior may be more parsimonious. Topographical variability has rarely been examined, nor has operant variability with organisms with baselines featuring stereotypic responding. This study examined the effects of a variability-specifying contingency on the cumulative novel responses of undergraduate students. Using the PORTL apparatus, participants interacted with a ball with a single hand. When the variability-specifying contingency was in effect, novel topographies were reinforced. When a reinforce every response condition was implemented, the participants did not emit any novel responses. When variability-specifying contingencies were in effect, novel responses were rarely followed by subsequent novel responses. They were mostly followed by repeated emission of the same topography, or by other previously emitted topographies. Novel responding did not persist long, although the variability-specifying contingency remained in effect and the potential for novel responding was great. The variability-specifying contingency often resulted in stereotypic response chains. Each of these findings call into the question the assertion that variability is an operant and suggests other possible explanations for the observed novelty.