You Can′t Always Forget What You Want: Directed Forgetting of Related Words

Two experiments investigated the influence of directed forgetting on semantically related words versus semantically unrelated words. Subjects first studied a list in which a REMEMBER or FORGET instruction followed each word (e.g., crab … REMEMBER … leg … FORGET); they were tested on all of the words, regardless of study instruction. Recall and recognition-scored both for individual words and for word pairs-revealed a consistent pattern: Retention of FORGET words was markedly better when they followed related REMEMBER words, as opposed to when they followed related FORGET words or unrelated words. Thus, a pre-existing semantic relation was able to override an instruction to forget. This helps to explain why studies using unrelated word lists generally show substantial cued forgetting, whereas those using related material do not. The findings also place a clear constraint on the domain and explanation of directed forgetting.