
We readily imagine memory as a recording stored in the brain, with remembering simply playing the original. Research on reconsolidation offers a more dynamic picture. When some long-term memories are reactivated by a cue, they temporarily become more open to change. As they stabilise again, new information may be incorporated, while existing details may be strengthened, weakened or reorganised.
This helps explain an everyday experience. After telling the same story many times, what comes most readily to mind may include not only the original event but also other people's questions, photographs seen later and the narrative repeatedly used to organise it. A memory that feels increasingly fluent is not necessarily becoming closer to an “original version”; it may simply have grown more coherent through repeated reconstruction.
Yet the claim that every act of remembering rewrites a memory is too strong. Whether reactivation produces change depends on boundary conditions, including the memory's strength and age, the kind of reminder, and whether new learning or interference follows retrieval. A meta-analysis of human episodic memory supports greater susceptibility to interference after reactivation, but effect sizes and experimental conditions vary.
My judgement is that we need not distrust every memory, but neither should we treat memory as courtroom video. It is better understood as a reconstruction that preserves the past while serving the present. In an important dispute, personal certainty can be evidence, but it cannot automatically replace contemporaneous records, independent evidence and comparison across accounts. Remembering vividly and being completely accurate are different judgements.
https://pubmed.ncbi.nlm.nih.gov/28025069/
https://pubmed.ncbi.nlm.nih.gov/32161213/
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.