Activity in the left inferior prefrontal cortex (LIPC) is often thought to reflect processes that support episodic encoding. Functional magnetic resonance imaging (fMRI) was used to test whether processes subserved by LIPC could be negatively related to subsequent memory performance. Specifically, the current experiment explicitly tested the hypothesis that LIPC processing would positively impact encoding when primarily focused towards specific target items (item-level processing), whereas it would negatively impact encoding when primarily focused on the retrieval and instantiation of current task instructions (task-level processing). Two methods were used to identify regions that were sensitive to the two types of processes: a block-level manipulation of encoding task that influenced subsequent memory, and a back-sort procedure. LIPC was sensitive to item- and task-level processing, but not in a way that always facilitates encoding. LIPC was more active for subsequently remembered words than subsequently forgotten words, but it was also more active in a task that emphasized task-level processing relative to a task that emphasized item-level processing, although this former condition led to poorer subsequent memory performance. This pattern indicates that processes subserved by LIPC are not always positively correlated with episodic encoding. Rather, LIPC processes can support both the controlled semantic processing of items and the controlled retrieval of relevant semantic task context. When devoted to the latter, the diversion of LIPC processes to the task level can have a negative consequence for item-level analysis and encoding.