Many theories assume that actively maintaining information in working memory (WM) predicts its retention in episodic memory (EM), as revealed by the beneficial effects of more WM time. Here, we examined whether affording more time for intentional WM maintenance does indeed drive EM. Participants either intentionally or incidentally encoded sequences of four words presented during trials of simple span (short time) and complex and slow span (long time). Long time intervals entailed a pause of equivalent duration between the words that presented an arithmetic problem to read aloud and solve (complex span) or a blank screen (slow span). To ensure similar encoding of the words across the intentional and incidental encoding groups, participants silently decided whether each word was a living or nonliving thing via keypress (i.e., an animacy judgment; Experiment 1) or read the words aloud while pressing the spacebar (Experiment 2). A surprise delayed recall test at the end of the experiment assessed EM. Longer time in WM, particularly during slow span, improved retrieval from EM for both intentional and incidental encoding groups relative to short time, but for different reasons: modeling of the data indicated that longer intentional encoding increased binding memory (i.e., retrieval of the items’ positions in the trial; Experiments 1 and 2), whereas longer elaborative but incidental encoding increased item memory (i.e., memory of items irrespective of their bindings; Experiment 1). This suggests that time spent actively keeping information in WM is special for EM because WM is a system that maintains bindings.
Active maintenance in working memory reinforces bindings for future retrieval from long-term episodic memory
