People who seem to have a knack for calming down a heated moment are often described as having some innate gift for reading a room. The research behind one of their most reliable tools points somewhere more mundane and more useful than a gift: a specific, learnable habit of naming what’s actually happening emotionally, out loud, in plain words.
The clearest evidence for why this works comes from a 2007 study by Matthew Lieberman and colleagues at UCLA, published in Psychological Science. Participants underwent brain scans while looking at photographs of faces showing strong emotions, like anger or fear. In one condition, they simply looked at the images. In another, they were asked to label the emotion they saw, choosing a word like “angry” or “afraid.” The scans showed that labeling the emotion measurably reduced activity in the amygdala, the brain region most associated with detecting threat and generating a fear or anger response, compared with passively viewing the same image. At the same time, labeling increased activity in the right ventrolateral prefrontal cortex, a region involved in regulating emotional responses. The researchers described this as evidence that putting a feeling into words engages a kind of built-in brake on the brain’s alarm system, not just a way of describing distress after the fact, but something that appears to dampen the distress itself while it’s happening.
That finding, on its own, is about looking at a photograph in a scanner, not about defusing an actual argument between two people. But it maps onto something a more applied body of research has tested directly: whether training people in specific verbal techniques, including naming and reflecting emotions back to someone in distress, actually changes outcomes in high-stakes, tense real-world situations. A 2022 cluster randomized study led by Andreja Celofiga and colleagues, conducted across the acute psychiatric wards of hospitals throughout Slovenia, tested exactly this. Staff on some wards were trained in verbal and non-verbal de-escalation techniques, including naming a patient’s emotional state and responding to it directly, while other wards continued as usual. Compared with the untrained wards, the wards where staff had been trained saw a 73 percent lower rate of aggressive incidents and an 86 percent lower rate of severe incidents. In a setting where tense, potentially dangerous situations happen routinely, a fairly specific set of learnable verbal habits corresponded with a large, measurable drop in how often things escalated.
The Artful Age
A weekly letter on aging well, family across generations, and the creative life after the kids leave home.
It would be satisfying to conclude that these verbal techniques are close to a reliable fix wherever tension runs high, but the broader research doesn’t fully support that. A systematic review of de-escalation training in forensic psychiatric settings, a higher-security and often more institutionally rigid environment than a general acute ward, found no clear reduction in the overall rate of violent incidents after staff were trained, even though staff themselves reported feeling safer. Reviewers examining the evidence base for de-escalation training more broadly have described the underlying research as genuinely mixed in quality, with results that vary a good deal depending on the setting, the specific training used, and how rigorously the study measuring it was designed. The honest takeaway isn’t that these techniques don’t work. It’s that they seem to work better in some contexts than others, and the exact conditions that make the difference aren’t yet fully mapped out.
Put together, the amygdala research and the de-escalation training research are answering two different but related questions. Lieberman’s study explains a plausible mechanism, why naming an emotion might calm a nervous system down at all, at the level of individual brain activity. The Slovenia trial and the broader training literature test whether that mechanism translates into something that actually changes outcomes when it’s deployed deliberately, by trained people, in real tense situations. The mechanism looks solid. The real-world payoff looks real but uneven, strongest in some settings and unproven in others.
What that leaves is a genuinely useful, if less dramatic, practical takeaway. Naming what’s happening emotionally, saying “you seem really frustrated right now” or simply “I’m feeling angry” instead of letting a feeling sit unspoken and building, has a documented basis for calming a threat response down, both in a brain scanner and, in at least some tested settings, in the middle of an actual tense encounter. It isn’t a guaranteed fix for every kind of conflict. But it’s one of the few de-escalation habits with evidence behind it at more than one level, from what happens inside a single brain to what happens on an entire hospital ward.