Somebody you haven’t seen in years brings up a kind thing you once did for them. You shrug and say, “Well, of course. That’s what friends do.” Now imagine the same person brings up a time you let them down, or told them something that wasn’t true. This time the words that come out are different: “I was a different person back then.” Or, “That’s not who I am.”
In both cases you remember the event. You are not denying anything. What changes is how close you let the memory sit to you. The good deed stays right next to you. The bad one gets pushed back in time and handed to an earlier version of yourself. A series of studies from Felipe De Brigard’s lab at Duke University (De Brigard was named a MacArthur Fellow in 2026 for his work on memory) suggests this is not a quirk of a few people. It is a regular feature of how we remember our own moral history.
This is not quite the self-serving bias
If you have taken an intro psych class, the reunion scenario probably sounds like the self-serving bias: I take credit for my successes and blame my failures on the situation. That bias is about explaining outcomes. The pattern in these studies is a cousin of it, but the move is different. When you say “I was a different person then,” you are not blaming the situation. You keep the blame and assign it to a past self who conveniently no longer exists.
The idea has roots in what Anne Wilson and Michael Ross (2001) called temporal self-appraisal theory. People tend to rate their past selves more harshly than their present selves, and the further back they look, the worse the past self appears. Wilson and Ross argued this is motivated: believing you have improved feels good, so you arrange your past to tell a story of improvement. The Duke group asked what happens when the memories in question are moral ones.
Study one: your own lies look worse the older they are
In a 2017 paper in the Journal of Experimental Psychology: General, Matthew Stanley, Paul Henne, Vijeth Iyengar, Walter Sinnott-Armstrong, and Felipe De Brigard asked people to recall real events, not hypothetical vignettes. Forty-eight adults spent thirty minutes writing down specific times they had lied to someone in the past ten years, and thirty more minutes writing down times someone had lied to them. For each memory they rated how morally wrong the act was, how negative and intense the emotion was, and how much they felt like the same person now as they were then.
Three findings came out of this. People rated their own lies as less wrong and less negative than lies told to them (on a seven-point scale where lower means more wrong, their own lies averaged about 3.7 and others’ lies about 2.9). Their own lies were rated as more wrong the further back in time they occurred, while other people’s lies did not change with time at all. And the more people felt they had changed since the event, the more harshly they judged their own past lie. Again, no such pattern for lies told to them.
A second version with memories of emotionally hurting someone showed the same shape, though the difference in how wrong people rated their own versus others’ behavior was borderline until the researchers controlled for how well each event was remembered. A third study with 100 online participants asked people directly for lies from a period when they felt like a very different person and from a period when they felt like the same person. Lies from the “different me” era were rated as more wrong about 72 percent of the time in paired comparisons (75 percent for emotional harm), where chance would be 50.
Study two: remembering a wrong act makes you feel changed
A 2019 paper by Stanley, Henne, and De Brigard in Memory and Cognition went at the question from the other direction. Instead of asking whether distance changes your judgment of the act, they asked whether the act changes your sense of distance. Across four online studies, participants recalled morally right and morally wrong things they had done, then answered three questions: do you feel like a different person now, do you feel like the same person, and how much have you changed.
After recalling a wrong act, people felt like a different, changed person. After recalling a right act, they felt like the same person. The gap was about a point and a half on a seven-point scale, and it held after controlling for when the event actually happened, measured both as a time bin and as a count of months. This was not about the calendar. It was about the content.
The pattern held across four moral domains: honesty, helping, fairness, and loyalty. And in the third study, with 238 participants in a between-subjects design, the researchers added a condition where people recalled good and bad acts they had watched other people commit. The effect vanished. People did not think a stranger had changed after doing something wrong. They only thought that about themselves. That contrast is what turns an interesting pattern into evidence for a self-protective bias.
A fourth study with 804 participants asked a narrower question: after recalling a time you were dishonest, do you think you are more honest now? Yes, compared with people who recalled an honest act. The same held for fairness and loyalty. For helping and harming, the effect did not survive the time control, so three of the four domains held up.
Reasons to hold this loosely
Everything here is self-report. The authors raise an alternative reading themselves: saying “I’m a different person now” after recalling a lie may be a way of disavowing the act rather than a perception of change. They note it is hard to tell which comes first, and that question remains open.
The participants were mostly American adults recruited through Amazon Mechanical Turk. The motive to see oneself positively is well documented in Western cultures; whether this exact pattern appears in cultures that understand the self differently is not yet known.
On the other side of the ledger, the researchers did the work to rule out the obvious alternatives. They moved from a within-subjects design, where participants might guess the hypothesis, to a between-subjects one. They controlled for objective time. They added the other-person condition. And they reported the helping-harming result as a miss rather than leaving it out.
One question nobody has tested yet: if you push a bad deed onto “the old me,” are you less likely to make amends for it? The 2019 paper closes by noting that people who see themselves as morally upstanding still lie and cheat with surprising frequency, and that helping people see their moral selves accurately is a goal for future work.
What to do with this
Sometimes you really have changed. But the next time you catch yourself saying “that’s not who I am” about something you did, it may be worth pausing on the fact that you did it. Keeping that memory on the shelf next to you, instead of shipping it off to a former self, is harder. That difficulty may be exactly why memory makes the other option so easy.
Psychology Terms in This Article
Self-serving bias. The tendency to attribute one’s successes to internal causes such as ability and effort, and one’s failures to external causes such as bad luck or an unfair situation. The reunion scenario in this post looks like the self-serving bias at first, but the Duke studies describe a different move: keeping responsibility for a bad act while assigning it to a past self.
Temporal self-appraisal theory. Wilson and Ross’s proposal that people evaluate their past selves more harshly than their present selves, and feel psychologically closer to flattering past events than to unflattering ones, in order to maintain a positive view of the current self. Both Stanley papers build directly on this theory by asking whether it applies to memories of moral and immoral behavior.
Autobiographical memory. Memory for events from one’s own life, which research has consistently shown to be reconstructive rather than a literal record. The studies asked participants for real autobiographical memories of lying, harming, helping, etc., rather than the hypothetical vignettes common in moral psychology.
Within-subjects and between-subjects designs. In a within-subjects design every participant experiences every condition; in a between-subjects design each participant is in only one condition. The first 2019 studies were within-subjects (everyone recalled both right and wrong acts); Study 3 switched to between-subjects to rule out participants guessing the hypothesis.
Demand characteristics. Cues in an experiment that lead participants to guess what the researcher expects and adjust their behavior to match. Asking the same person for both good and bad memories and then asking about change could tip off the hypothesis; the between-subjects replication addressed this.
References
- Stanley, M. L., Henne, P., & De Brigard, F. (2019). Remembering moral and immoral actions in constructing the self. Memory & Cognition, 47(3), 441–454.
- Stanley, M. L., Henne, P., Iyengar, V., Sinnott-Armstrong, W., & De Brigard, F. (2017). I’m not the person I used to be: The self and autobiographical memories of immoral actions. Journal of Experimental Psychology: General, 146(6), 884–895.
- Wilson, A. E., & Ross, M. (2001). From chump to champ: People’s appraisals of their earlier and present selves. Journal of Personality and Social Psychology, 80(4), 572–584.
