Your boss may be gaslighting you if they repeatedly deny or rewrite things you saw and heard, and you leave those conversations doubting your own memory. One disagreement about what was said is not gaslighting. A repeated pattern of denial, from someone with more power than you, that leaves you checking your own sanity, is consistent with it.
The word gaslighting is used loosely, so this post sets a higher bar. It gives you ten composite examples, a test that separates gaslighting from an ordinary bad memory, and steps you can take this week. It cannot tell you what your boss intends, and it will not try.
What gaslighting is, and what it is not
The APA Dictionary of Psychology defines to gaslight as to “manipulate another person into doubting their perceptions, experiences, or understanding of events.” It adds that the word is usually considered a colloquialism, though it occasionally appears in clinical literature. That matters here. Nobody can diagnose your boss from a distance, and this post does not.
A 2026 conceptual analysis of workplace gaslighting in Frontiers in Psychology by Popat and Pandey adds two useful ideas. First, the behavior itself has two necessary ingredients: trivializing your concerns, and a capacity to cause you distress. Second, the deeper effect, where you start to doubt your own reality, depends on a perceived power imbalance. In other words, the same words from a peer and from the person who controls your review do not carry the same weight. The authors also note that intent is hard to prove, so they treat it as contingent, not required.
Sociologist Paige Sweet argued in a 2019 American Sociological Review article that gaslighting is rooted in social inequality and works through power-laden relationships. Her case study was domestic violence, not offices, so applying it to work is an analogy. It is a useful one, because a boss is one of the few people at work who hold real power over your income, your reputation and your record.
Two things are not gaslighting. Honest disagreement is not, because two people can leave the same meeting with different memories. Blunt feedback is not either, because it can sting without making you doubt what happened.
Ten examples, as composite scenarios
These ten short scenes are all (composite scenarios): none is a real person or company. Any one of them alone proves nothing. The pattern is what counts.
- The vanished instruction. “I never asked you to do it that way.” You have the task in your notes. The denial comes flat and calm, and then the blame follows you.
- The disappearing praise. After a good project, you hear “great work.” Two months later, in your review: “I don’t recall ever saying that.” Nothing was written down.
- The rewritten priority. Last week the client deck was the top priority. Now: “That was never the priority, we talked about this.” This is the memory version of moving the goalposts.
- The joke that wasn’t. A cutting remark in a meeting, and when you raise it: “I was joking, you’re too sensitive.” The event is reclassified, and your reaction becomes the problem.
- The two-channel boss. In a call, an instruction. In the email thread, nothing. Later: “I don’t remember that call.” The verbal version leaves no trace.
- The worry turned into a symptom. You report an unrealistic workload. The reply is: “Are you okay? You seem stressed lately.” Your work problem quietly becomes a comment on your mental state.
- The unnamed chorus. “Everyone has noticed this.” Nobody is named, and nobody is quoted. A manufactured sense of agreement is doing the work, which the guide covers as group gaslighting.
- The edited notes. The meeting summary arrives and leaves out your objection. When you mention it, you are told that is not what happened. A summary written after the fact becomes the official memory, a pattern the guide calls manufacturing the record.
- The reversed complaint. You raise a missed handoff. By the end of the conversation you are the one explaining your tone. If this happens after your calm, specific claim, it resembles the sequence known as DARVO, a term from research on interpersonal violence that the post on DARVO at work applies by analogy.
- The memory verdict. “You always forget things.” It is said in front of others, and soon colleagues double-check what you say. A general claim about your character has replaced an argument about one event.
Example 6 shows the basic move in one picture. The subject of the conversation changes while you are still in it.
Why a confident denial works on memory
Memory is not a recording. In a classic 1974 study, Elizabeth Loftus and John Palmer showed people a film of a car accident. A week later, 32 percent of those who had been asked about cars that “smashed” said they saw broken glass, against 14 percent of those asked about cars that “hit.” There was no broken glass in the film. Loftus’s 2005 review in Learning & Memory summarizes three decades of similar work: information that arrives after an event can change how people remember it.
A boss saying “I never said that” is exactly this kind of after-the-fact information, delivered with authority and confidence. That is why doubt is a normal reaction, not a sign that something is wrong with you.
It also means your own memory can be bent, and that is the fair part of the test below. It protects you from a boss’s revision, and it protects your boss from yours.
The paper-trail test
Ask these three questions about any incident.
- Is there a dated record that shows what happened, such as an email, a chat message, a calendar invite or a document history?
- When the record is put in front of the other person, do they engage with it, or does the conversation move to your attitude, your memory in general, or your motives?
- Does this happen across different topics, in one direction, and mostly with you?
A genuine memory gap looks like this: “You may be right, let me check.” It updates when shown a record, and it goes both ways. The person also does not hang a verdict on you afterward.
A gaslighting pattern looks like this: the record is dismissed, minimized or reinterpreted. The denial shifts from “that didn’t happen” to “you always do this,” and the outcome leaves you more uncertain, not less. One yes to the third question means little. If it repeats over weeks or months, across several topics, it is worth taking seriously. This is a rule of thumb from this site, not a validated scale.
If there is no record, that is not proof of anything either. It only means you are in the position most people are in, which is why the next section starts with writing things down.
Two cautions keep the test honest. First, do not run it after every awkward exchange. A manager juggling ten projects will sometimes forget a conversation, and one forgotten conversation says nothing about character. Second, do not treat your own certainty as a record. Your feeling of being sure is real, but it is not evidence to anyone else. The record, and the pattern over time, are what carry weight with a colleague, a mentor or HR.
Why the power gap matters
Trusting your memory is easier when you have a colleague to check it with. A boss can make that harder by being the person who controls what counts as the official version.
The research on abusive supervision shows why this is more than an annoyance. In a 2000 study in the Academy of Management Journal, Bennett Tepper found that employees who saw their supervisors as more abusive were more likely to quit. Those who stayed reported lower job and life satisfaction and more psychological distress. Gaslighting is not named in that study, but the 15 items on the abusive supervision scale used in that line of research include “lies to me” and “breaks promises he/she makes,” as reproduced in a 2023 paper by Tziner and colleagues.
Feeling worn down by this is not a personal weakness. It is consistent with what the abusive supervision research reports.
What to do this week
- Write down what was said within the hour: date, time, who was present, and the exact words in quotation marks. Keep it somewhere your employer cannot access, such as your personal phone. The guide's [document contemporaneously](/guide/document-contemporaneously/) entry gives the format.
- After important conversations, send a neutral recap: "To confirm what we agreed: [two or three points]. Tell me if anything here is different."
- Ask for new instructions and criteria in writing, and keep the request neutral, not accusing.
- Talk to one person you trust who does not work on your team. A second read restores what isolation removes.
- Notice your own signal. Repeated self-doubt after conversations with one person is data, and [getting your own judgment back](/guide/getting-your-own-judgment-back/) is a slower project you can start now.
For exact wording, the sibling post what to say when gaslit at work offers nine calm scripts, and how to document workplace bullying covers the record in more detail.
Try not to use the word “gaslighting” with your boss. It invites an argument about the label. Stick to one specific, dated fact at a time.
If you are worried about retaliation, or your health is suffering, treat that as a reason to seek outside support, not something to push through alone. You can protect yourself without proving anything about anyone else.
Sources
Primary research and original sources
- Popat, N., & Pandey, J. (2026), “Workplace gaslighting: a construct for organizational research,” Frontiers in Psychology, 17, a conceptual analysis of workplace gaslighting, its necessary attributes and its dependence on perceived power imbalance.
- Sweet, P. L. (2019), “The Sociology of Gaslighting,” American Sociological Review, 84(5), 851-875, the argument that gaslighting is rooted in social inequality and power (case study: domestic violence).
- Loftus, E. F., & Palmer, J. C. (1974), “Reconstruction of automobile destruction,” Journal of Verbal Learning and Verbal Behavior, 13(5), 585-589, the “smashed” versus “hit” experiment and the broken-glass question.
- Loftus, E. F. (2005), “Planting misinformation in the human mind: a 30-year investigation of the malleability of memory,” Learning & Memory, 12(4), 361-366, a review of the misinformation effect.
- Tepper, B. J. (2000), “Consequences of abusive supervision,” Academy of Management Journal, 43(2), 178-190, the outcomes of perceived abusive supervision, including quitting and distress.
- Tziner, A., et al. (2023), “Insights into abusive workplace behavior,” Frontiers in Psychology, 14, an open-access paper that reproduces the 15 items of Tepper’s abusive supervision questionnaire.
Clinical and professional references
- APA Dictionary of Psychology, “gaslight”, the definition and the note that the term is usually a colloquialism.

