You feel something. Before you have finished feeling it, you have already named it. Anxiety. Sadness. Anger. The word arrives, and the feeling stops. Not because the emotion has been resolved, but because the label has replaced it. You move on. You believe you have processed something. You have not. You have filed it.
Therapists call this affect labeling, and for nearly two decades it has been treated as one of the most reliable tools in emotional regulation. The idea traces back to a 2007 UCLA study that showed labeling an emotion reduces amygdala activation. The finding was clean, replicable, enormously useful. It entered DBT workbooks, CBT protocols, mindfulness apps, school counseling programs. Within a decade it had become one of psychology’s rare crossover truths, the kind of result that shows up in TED talks and therapy intake forms alike.
I taught it for years. I am no longer sure it is the whole story.
The trouble is not that the original finding was wrong. Affect labeling does reduce amygdala reactivity. The trouble is what we assumed that reduction meant. We took it to mean the emotion had been processed. It may instead mean the emotion had been terminated.
In 2021, Katharina Kircanski and colleagues tested what happens after the label is applied. Participants who named their emotional responses to aversive images subsequently performed worse on both cognitive reappraisal and mindful acceptance tasks, compared to those who had not labeled at all. The label did not prepare the ground for regulation. It foreclosed it. Once the word arrived, the downstream capacity to sit with or restructure the feeling had already been compromised (Kircanski et al., 2021, Affective Science, 2, 389-398).
This finding should have created more discomfort than it did.
Consider the clinical pattern. A client enters therapy unable to articulate what they feel. Over months, they learn the vocabulary: attachment anxiety, avoidant withdrawal, shame spirals, grief layered under anger. Their fluency improves. Their therapist notes progress. And yet something else happens in parallel: the time they spend inside the emotion gets shorter. The label arrives faster. The classification is more precise. And the felt duration of the experience contracts. They can describe their sadness with increasing sophistication, but they spend less and less time actually being sad.
This is not recovery. This is abbreviation.
James Pennebaker’s research on expressive writing showed that the therapeutic benefit of writing about emotional experiences came not from labeling alone but from the construction of meaning across sessions (Pennebaker, 1997, Psychological Science, 8(3), 162-166). Participants who simply named their feelings showed no significant health improvements. Those who built narrative structure around their experiences did. The implication is plain: labeling is only the entrance to processing, not the process itself. When labeling becomes the endpoint, the construction that produces actual change never begins.
Lisa Feldman Barrett’s constructionist account deepens the problem from another angle. In How Emotions Are Made (2017), Barrett argues that emotions are not triggered but constructed: the brain assembles an emotional experience from interoceptive signals, prior learning, and the conceptual categories available to it. Language is central to this construction. It provides the scaffolding around which otherwise diffuse physiological states coalesce into recognizable feelings. If that is correct, then the speed of labeling matters enormously. Rapid labeling does not describe an emotion that was already there. It constructs one and closes it in the same motion. The faster the concept arrives, the narrower the construction.
What does “I am anxious” do to the seventeen other things happening in the body at that moment? It folds them behind a single word. The relief, the anticipation, the faint guilt, the unnameable pressure beneath the sternum: all of it collapses into “anxiety” the instant the label lands. The word does not rescue the experience. The word replaces it.
In my own research, I have been tracking a version of this phenomenon across different registers. Analysis of 351,734 relationship narratives revealed a weaker correlation between linguistic precision and emotional stability than standard models predict; in certain conditions, the correlation reversed (Kim, 2026, under review, PLOS ONE). Individuals whose language was most precise about their emotional states were not the ones who showed the most stable trajectories. In some cases, precision appeared to correlate with premature closure rather than genuine resolution.
This led me to formalize what I now call Affective Suppression Fatigue (Kim, 2026, under review, Cognition and Emotion). The model proposes that emotion regulation operates along an intensity-dependent bifurcation: at low intensities, suppression produces numbing; at high intensities, it produces collapse. Naming can function as a form of suppression. When a client says “I am sad,” the layered affective tones behind that statement compress behind the label. The anxiety folded into it, the relief that recognition brings, the grief that the recognition simultaneously seals off. The label is not neutral. It acts as a sorting mechanism that privileges one reading and discards the rest.
The question then becomes: what happens to the discarded material?
It does not disappear. It persists below the threshold of articulation, accumulating what the model describes as suppressive debt. Over time, this debt expresses itself not as the return of specific emotions but as a generalized erosion of affective range. The person does not become overwhelmed. They become flat. They can still name what they feel. They simply feel less.
This connects to a broader structural concern I have described through the PESAM framework (Kim, 2026, under review, Acta Psychologica). PESAM models the self as a predictive system that continuously generates emotional forecasts. When labeling accelerates, the system begins to rely on prior labels rather than current interoceptive data. The emotional self becomes increasingly scripted. Interpretive authority shifts from the body’s present-tense signals to the mind’s retrospective categories. I call this process interpretive displacement: the gradual replacement of felt experience with conceptual shorthand.
The endpoint of interpretive displacement is not confusion. It is a false clarity. The person believes they understand what they feel. They can articulate it with precision. But the articulation has outpaced the experience. They are narrating an emotion they are no longer inside.
Krishnamurti identified something structurally identical in 1954, without the neuroscience. In The First and Last Freedom, he observed that naming creates a false sense of understanding that prevents further examination. “The moment you have given it a name, you have stopped looking at it.” The observation was dismissed by the empirical tradition for decades. Kircanski’s data suggests it deserved more attention.
Donald Winnicott’s concept of the false self maps onto this problem with unsettling precision. Winnicott described patients who had learned to perform a self constructed from borrowed expectations rather than authentic impulse (Winnicott, 1965). The therapeutic danger of rapid labeling is that it can produce an emotional false self: a person who names feelings in the language of their therapist, their workbook, their meditation app, without those names corresponding to anything still alive inside them. The labels are correct. The experience is gone.
This brings me to the concept I have been developing under the heading of Affective Sovereignty (Kim, 2026, Minds and Machines; Kim, 2026, Discover Artificial Intelligence, DOI: 10.1007/s44163-026-01000-0). Affective Sovereignty asks: who holds the final authority over the interpretation of an emotional experience? In clinical settings, interpretive authority often migrates from client to framework. In AI-mediated contexts, it migrates from user to system. In both cases, the transfer is invisible. The person believes they are generating their own understanding. They are operating within a borrowed grammar.
The therapeutic goal, I am increasingly convinced, should not be faster or more accurate labeling. It should be the preservation of the capacity to remain with an emotion without naming it. Not indefinitely. Not as a refusal of language. But long enough for the full architecture of the feeling to become legible on its own terms, before a word arrives and closes the door.
The emotions that last longest are the ones that were never named in time.
The ones that were named immediately are the ones we remember least.
References
Barrett, L. F. (2017). How Emotions Are Made: The Secret Life of the Brain. Houghton Mifflin Harcourt.
Gross, J. J. (2015). Emotion regulation: Current status and future prospects. Psychological Inquiry, 26(1), 1-26.
Kim, R. S. (2026). Affective Suppression Fatigue: Intensity-dependent bifurcation in emotion regulation. Under review, Cognition and Emotion.
Kim, R. S. (2026). Predictive Emotional Selfhood and Affective Modeling (PESAM). Under review, Acta Psychologica.
Kim, R. S. (2026). Linguistic precision and emotional trajectory in relational narratives. Under review, PLOS ONE.
Kim, R. S. (2026). Affective Sovereignty and the right to emotional interpretation. Minds and Machines.
Kim, R. S. (2026). Algorithmic affective blunting and emotional interpretive authority. Discover Artificial Intelligence. DOI: 10.1007/s44163-026-01000-0.
Kircanski, K., et al. (2021). Emotion naming impedes both cognitive reappraisal and mindful acceptance strategies of emotion regulation. Affective Science, 2, 389-398.
Krishnamurti, J. (1954). The First and Last Freedom. Harper and Brothers.
Lieberman, M. D., et al. (2007). Putting feelings into words: Affect labeling disrupts amygdala activity in response to affective stimuli. Psychological Science, 18(5), 421-428.
Pennebaker, J. W. (1997). Writing about emotional experiences as a therapeutic process. Psychological Science, 8(3), 162-166.
Winnicott, D. W. (1965). The Maturational Processes and the Facilitating Environment. Hogarth Press.