Walk past any physics lab at 2am and you will hear it: someone alone at a whiteboard, muttering. Not to a colleague. Not into a phone. To themselves. The habit that gets children scolded in classrooms turns out to be one of the most reliable cognitive tools available to the adult brain — and the research explaining why has been piling up for two decades.
Cognitive psychologists have a name for it: private speech. The Russian developmental psychologist Lev Vygotsky described this phenomenon in the early 20th century, arguing that children who talk to themselves while solving puzzles were not exhibiting immaturity but rehearsing a skill that would eventually go silent and become adult reasoning. Modern experimental work has largely vindicated him — and then gone further. Adults who never stopped talking to themselves out loud appear to hold an edge on tasks involving working memory, visual search, and complex problem-solving.
The experiments that broke the taboo
The clearest case for talking out loud comes from a laboratory at Bangor University. In a 2017 experiment, neuropsychologist Paloma Mari-Beffa and Alexander Kirkham gave 28 participants a set of written instructions and asked them to read those instructions either silently or aloud. Concentration and performance both improved when the instructions were spoken. As the researchers reported the result, “talking out loud actually improves control over a task, above and beyond what is achieved by inner speech.” Much of the benefit appears to come from simply hearing oneself — auditory commands turn out to be better controllers of behaviour than written ones. Mari-Beffa went further still, suggesting that talking out loud, when the mind is not wandering, can be a marker of high cognitive functioning rather than a sign that anything is wrong.
A separate line of work explains part of the mechanism. When participants search for specific objects in cluttered displays, those who repeat the target word aloud find their targets faster — sometimes dramatically so. Psychologists Gary Lupyan and Daniel Swingley documented this in 2012, finding that speaking facilitated the search especially when the word and the object were strongly associated. Hearing yourself say the name of the thing you are looking for activates additional visual information about that object in the brain. Saying “keys” primes the shape, the metallic sheen, the size — thinking “keys” silently does not do it as efficiently.
The verdict was clear. Speech is not a symptom of the search. It is part of the machinery.
Why saying it beats thinking it
The reason external speech outperforms internal speech has to do with how loosely thoughts are held together when they never leave your head. A thought in silent form can be half-formed, associative, contradictory, or partially imagistic. The moment you commit to speaking it, syntax intervenes. Subject. Verb. Object. Your brain is forced to pick a version.
That forced clarification is the point. Speaking a problem aloud engages the prefrontal cortex more deeply than silent rumination, essentially routing the thought through the brain’s executive control system rather than letting it drift through default-mode wandering. The result is the sensation many people describe when they finally say something out loud and realise, mid-sentence, what they actually think.
This is also why writers talk to themselves. Why programmers explain code to rubber ducks. Why chess players mutter through positions. Why surgeons narrate steps during unfamiliar procedures. The externalisation is doing cognitive work.
The negative critic in your head is not the enemy
Self-talk gets complicated when the content turns critical. As Forbes contributor Mark Travers has explored, negative self-talk can sharpen focus and drive — but only when it prompts self-correction rather than self-blame.
Too much positive self-talk can tip into overconfidence and shut down self-awareness. Too much criticism drains motivation. The productive middle looks less like a cheerleader and more like an honest coach — one you happen to be having out loud, because the discipline of speaking forces the criticism into an actionable form. For example, thinking You’re rushing this section is more useful when spoken than the fog of unease it started as.
The stigma is cultural, not clinical
Most people who talk to themselves in public do so quietly, glancing around, embarrassed. The embarrassment is learned. Children between three and five talk to themselves constantly during play and problem-solving. The behaviour goes underground around age seven, not because it stops being useful but because it starts drawing social penalties.
The residual awkwardness follows people into adulthood. Someone caught muttering to themselves on a train assumes they have done something odd. In fact they have done something the brain has been doing since childhood — the same brain that, as Silicon Canals has explored in a piece on unrecognised signs of intelligence, often expresses itself through habits that get mistaken for eccentricity.
What the neuroimaging shows
Beyond behavioural experiments, imaging work has added mechanistic detail. Speaking aloud engages multiple brain regions including those responsible for speech production, motor control, and auditory processing. That auditory feedback loop is critical. When you hear yourself say something, your brain processes it partly as external input — which means self-talk effectively converts private thought into something the brain treats as new information worth attending to.
This is why people rehearse presentations by saying them out loud rather than reading them silently, and why students who explain material aloud to imaginary audiences retain it far better than those who reread notes. The retention gain is not motivational. It is structural.
Verbal rehearsal, whether whispered or fully vocalised, extends the duration items can be held in working memory. Speaking phone numbers aloud to remember them is not superstition. It is a compensatory technique your brain figured out before neuroscience did.
The overlap with other misread traits
Self-talk rarely travels alone. It tends to cluster with other habits that get pathologised in adulthood: fidgeting during meetings, zoning out mid-conversation, needing solitude to recover from social contact. An examination of behaviours that make highly intelligent people seem rude has highlighted how fidgeting can improve cognitive performance and help sustain attention in both children and adults — the same pattern as private speech. Motion and vocalisation are external supports that the brain uses to hold complex thought in place.
The pattern extends to how these minds experience downtime. As Silicon Canals reported in an examination of how highly intelligent people experience boredom, the same restlessness that shows up as muttering during a task can show up as extreme intolerance for under-stimulating environments. Both are the same nervous system asking for more to work with.
Where self-talk stops helping
The research is not a blanket endorsement. Self-talk becomes counterproductive under a few specific conditions. Rumination — the repetitive rehearsal of grievances or fears without any move toward resolution — tends to leave people feeling worse rather than better, and the longer it loops the deeper it can dig. The distinction between problem-solving self-talk and ruminative self-talk lies in direction of travel. Productive self-talk moves toward a decision. Rumination loops.
Second-person framing appears to help. People who address themselves by name or as “you” during stressful self-talk (such as “You’ve got this, Sarah”) may regulate emotion better than those who use first-person “I.” The linguistic distance seems to trigger the same neural circuits activated when giving advice to a friend — calmer, more strategic, less reactive.
And there is a difference between talking to yourself and being talked at by an intrusive voice you cannot control. The former is a tool. The latter is something to take seriously, and if it persists it is worth raising with a professional. Confusing the two is the source of most of the cultural taboo, and most of the muddle between an ordinary cognitive habit and something genuinely worth attention.
The professional world is already using it
Athletes have been coached in structured self-talk for decades. Golf, tennis, and gymnastics coaches routinely teach clients pre-shot and pre-routine verbal scripts, and the sports-psychology literature has consistently found performance gains from the practice. Military pilots use verbal callouts during high-load procedures. Emergency medicine training includes narrated checklists precisely because vocalisation catches errors that silent review misses.
In knowledge work, the technique has migrated into common practice under different names. Pair programming. Rubber-duck debugging. Think-aloud protocols in UX research. Each is a formal recognition that a thought spoken aloud is a thought that has been tested. The engineers doing it are not eccentric. They are using a well-documented augmentation of working memory.
The people who look strange doing it in public are, in most cases, doing exactly what the labs have been recommending — just without the professional cover that makes it socially legible. If the signs of intelligence you were never told about include habits you have quietly been apologising for, self-talk is near the top of the list.
A quiet rehabilitation
The rehabilitation of self-talk has been slow, partly because the behaviour still carries an image problem. Media depictions of the person muttering to themselves lean toward instability rather than cognition. But the labs have kept publishing, the sports psychologists have kept coaching, and the productivity culture has kept adopting the technique under sanitised names.
For anyone who has caught themselves narrating a stubborn spreadsheet, walking through a difficult email out loud, or arguing both sides of a decision on a solo drive: the science is on your side. The muttering is not the malfunction. It is the method. The brain thinks better when it can hear itself think.
Whether the person next to you on the train understands that yet is a separate question.
