In 2022, Merriam-Webster declared ‘gaslighting’ its word of the year. Since then, the term has been everywhere: in media, in conversations about relationships, at work, and even in politics. Gaslighting has become shorthand for a form of manipulation. This is where someone systematically makes another person doubt their own perception or memory. But how well do we actually understand this phenomenon from a scientific perspective?
A recent article in Personality and Social Psychology Review (Klein, Wood & Bartz, 2025) tries to provide some answers. What makes this paper stand out is that it moves beyond anecdotes and loose descriptions. It offers a theoretical framework that situates gaslighting within broader insights from psychology and neuroscience.
From film to science
The term gaslighting originates from the classic 1944 film Gaslight, in which a husband subtly manipulates his wife into questioning her sanity: lamps mysteriously dim, objects disappear, and memories are denied. Since the 1960s and 1970s, the term entered psychiatric and psychological discourse, but often in a rather vague, psychodynamic way. For decades, it remained more of a cultural than a scientific concept. Only in the last decade has there been a surge of academic interest. These studies encompass romantic relationships, parent–child dynamics, workplaces, and institutional settings.
A theoretical model
The authors argue that gaslighting can be explained by normal learning and social mechanisms operating under atypical conditions. Their proposal draws on prediction error minimisation (PEM). This is the idea that our brains constantly generate predictions about the world and adjust when these predictions are violated. Typically, this process helps us to learn and function effectively. However, in close relationships, where trust is paramount, a partner holds a privileged position. If they repeatedly provide alternative explanations (“You’re imagining things”, “You’re too emotional”), we start distrusting our own perception and give more weight to theirs.
Gaslighting, in this view, does not happen because the victim is “weak”, but because close relationships are built on epistemic trust. We naturally assume that a loved one or authority figure helps us make sense of reality. That very mechanism can be exploited.
From doubt to loss of agency
The model describes gaslighting as a cycle. It often starts with small discrepancies – a misplaced object, a contested memory. The gaslighter then supplies the explanation: “You’re forgetful”, “You’re misinterpreting things”. Over time, the target begins to lose confidence in their own judgement, while the partner’s version of reality gains more weight. Slowly, this undermines agency – the sense of having control and mastery over one’s life.
The framework also sheds light on why people do not easily “wake up” in gaslighting relationships. Once strong trust has been established, contradictory evidence is given less weight than confirming cues. It feels easier to believe that you have made a mistake than to admit that your partner is untrustworthy or manipulative.
Recovery and open questions
The authors also consider recovery. Often, a catalytic event or an outsider is needed – someone who can act as a trusted alternative and affirm that your perceptions are valid. Activities that reconnect people with their own embodied experience, such as physical activity, art, or mindfulness, may also help restore self-trust.
Still, we should see this paper for what it is: a theoretical proposal. It offers an elegant framework that links gaslighting to concepts such as attachment, self-verification, and shared reality. But empirical evidence remains limited. For now, the model primarily serves as a roadmap for future studies.
Why it matters
Gaslighting is not just a buzzword. It touches on fundamental questions: how do we learn who we are? How dependent are we on others to understand reality? And how can that dependence be abused? By connecting gaslighting to normal cognitive and social processes, this new framework shows it is not an exotic or rare phenomenon. Instead, it’s a distortion of everyday mechanisms of trust and learning. Precisely because of that, it deserves our attention – both in research and in public debate.
P.S. For my latest album so far from my band Blue and Broke, I wrote a song about this phenomenon:
This is a live-version: