Skip to main content
Have a personal or library account? Click to login
Forget For Now, but Remember Later: Can People Selectively Remove Information From Working Memory While Keeping it in Long-Term Memory? Cover

Forget For Now, but Remember Later: Can People Selectively Remove Information From Working Memory While Keeping it in Long-Term Memory?

Open Access
|Apr 2026

Abstract

When instructed to remove a just-encoded item from working memory, people can do that very effectively. Here we investigate the side effects for episodic long-term memory of removing an item from working memory. Participants encoded lists of words into working memory, and each word was followed by a cue to remember or to forget that word. In a subsequent test of long-term memory, words to be maintained in working memory were remembered better than words to be removed. This was the case regardless of whether participants expected the long-term-memory test. In the final experiment we cued participants after each word to either remove it from working memory while keeping it in long-term memory, or to maintain it in working memory but forget it for the upcoming long-term memory test. Participants could not selectively remember information for one kind of memory test but not the other. Instead, they compromised, removing words less effectively from working memory than in the preceding experiments, and yet failing to remember those words better in the long-term memory test than the words they were told to forget for the long term. People cannot remove information from working memory and maintain it in long-term memory, or vice versa.

DOI: https://doi.org/10.5334/joc.497 | Journal eISSN: 2514-4820
Language: English
Submitted on: Sep 19, 2025
Accepted on: Mar 25, 2026
Published on: Apr 7, 2026
Published by: Ubiquity Press
In partnership with: Paradigm Publishing Services
Publication frequency: 1 issue per year

© 2026 Hannah Dames, Vencislav Popov, Klaus Oberauer, published by Ubiquity Press
This work is licensed under the Creative Commons Attribution 4.0 License.