Since the 1980s, task-based language teaching (TBLT) has enjoyed considerable interest from researchers of second language acquisition (SLA), resulting in a growing body of empirical evidence to support how and to what extent this approach can promote language learning. Although transferability and generalizability are critical assumptions for TBLT, there is little empirical evidence that task-related language abilities are indeed transferable. The current study was conducted to address this need for empirical research on generalizability and transfer critical for the planning of teaching and assessment of learning by specifically investigating whether or not transfer occurs between two similar pedagogic tasks. Fifty-three randomly assigned low-level adult second language learners were trained in a computer lab to complete one of two pedagogic tasks or no task, after which all participants were tested on two transfer tasks. Although the results of a MANCOVA did not provide statistically significant evidence of transfer, a post hoc analysis on a subset of the lowest proficiency learners suggests that task-related language abilities are transferable. Directions for future research and implications for teaching and assessment are discussed in light of the present findings.