Right now, as you read this sentence, there's a voice in your head. It's reading these words to you. You can hear it — not with your ears, but inside. It has a cadence, a rhythm, something close to a tone. It's narrating this experience to you as you have it.
You didn't start it. You didn't decide, this morning, to begin narrating your day. You didn't choose the first thought that appeared when you woke up, or the second, or the one that's running right now underneath the reading. The voice was already there — already talking — before you had a chance to consider whether you wanted it.
This isn't a disorder. It isn't a symptom. It's the most ordinary feature of human consciousness. Nearly everyone experiences some form of inner speech — a running stream of verbal thought that accompanies daily life, comments on it, evaluates it, replays it, and rehearses what hasn't happened yet. It's so constant and so familiar that most people never stop to ask the obvious question.
In 2020, a tweet by a user named Kyle Kidd went viral. "Hey so I just found out some people don't have an internal monologue," he wrote. "Like they think in concepts rather than words. This is blowing my mind right now." The tweet went viral. Many people were stunned to learn that some people don't hear a voice in their heads at all — that they think in concepts, images, or spatial relationships rather than words. Millions of people realised, for the first time, that the voice they'd assumed was universal might not be — and that they'd never once questioned its presence.
Who is talking? And who is listening? And if you're both, why does the conversation need to happen at all?
The voice nobody taught you
Children talk to themselves out loud. You can see it in any playground — a four-year-old narrating their own actions while they play, giving instructions to themselves, commenting on what's happening. "Now I'm going to put this one here. No, there. Okay." The psychologist Lev Vygotsky, working in the 1930s, proposed that this wasn't a quirk of childhood but a developmental stage. The child's outer speech, he argued, gradually moves inward — becoming quieter, more compressed, more abbreviated — until, around the age of seven, it disappears from public view entirely.
It doesn't disappear. It goes inside.
Vygotsky's theory has been broadly supported by decades of research since. Functional MRI studies confirm that inner speech activates the same brain regions as spoken language — Broca's area, the left frontal lobe, areas involved in speech production and auditory processing. When you talk to yourself silently, the brain is doing something remarkably similar to what it does when you talk to someone else out loud. The machinery is the same. The audience has changed.
What Vygotsky couldn't have known — because the technology didn't exist — is that the brain has a dedicated network for producing this internal commentary when no external task demands attention. In the late 1990s, the neuroscientist Marcus Raichle and his colleagues at Washington University discovered what they called the Default Mode Network — a set of brain regions that become active precisely when you stop focusing on the outside world.
The DMN is sometimes described as the brain's resting state. But it isn't restful. It's busy. It generates the inner monologue. It replays memories. It projects into the future. It runs simulations of social encounters — conversations you might have, things you should have said, how someone might react if you say this instead of that. When nothing external requires your attention, the brain doesn't go quiet. It starts talking to you.
Why it won't stop
You've noticed this. Everyone has. You finish a task — close the laptop, leave the meeting, get off the bus — and within seconds, the voice fills the gap. Not because you invited it. Not because you have something to think about. The gap appears, and the voice appears, and the two events are so close together that you rarely catch the gap itself. By the time you notice the silence, it's already been filled.
Try the opposite. Try to stop the voice. Sit still, close your eyes, and attempt to produce ten seconds of complete mental silence. No words. No narration. No commentary about whether you're succeeding.
You'll find it remarkably difficult. Not because you lack discipline. Because the instruction "stop talking to yourself" is itself a piece of inner speech, delivered by the voice you're trying to silence. The command and the problem are the same activity. You can't use the voice to turn the voice off any more than you can use a fire to extinguish itself.
This is the structural problem that meditation traditions have been pointing at for thousands of years, though they expressed it differently. The mind cannot transcend itself through its own activity. Every effort the mind makes to be silent is another sound.
Less constant than it feels
Despite the feeling of a voice that never stops, research suggests it may be less constant than most people assume.
Russell Hurlburt, a psychologist at the University of Nevada, has spent decades studying inner experience using a method designed to cut through the unreliability of introspection. He gives participants a beeper that sounds at random moments throughout the day. When it beeps, they note — immediately, before memory distorts it — what was happening in their mind at that exact moment. Over thousands of samples across many studies, Hurlburt found that inner speech was present in roughly a quarter of conscious moments. The rest of the time, thought occurred as images, feelings, sensory impressions, or what he calls "unsymbolised thinking" — knowing something without the knowledge taking any particular verbal form.
The exact figure is debatable — introspection research is inherently difficult to validate, and other researchers have arrived at different numbers. But the direction of the finding is consistent: the voice is present far less often than the feeling of the voice would suggest. It dominates the gaps — the transitions, the idle moments, the spaces between tasks — and those gaps are so frequent in daily life that the voice feels continuous. But during absorption, during physical activity, during real conversation, during skilled work, the voice often drops away and thinking continues without it.
This raises a question that the voice itself would prefer not to consider: if thinking happens without narration much of the time, what is the narration for?
The spectrum of inner speech
Not everyone experiences the voice the same way.
In 2024, researchers Johanne Nedergaard and Gary Lupyan published a study in Psychological Science that gave a name to one end of the spectrum: anendophasia — the near-absence of inner speech. Their research suggested that between 5 and 10 percent of the population experiences very little verbal inner thought, relying instead on images, spatial reasoning, and non-verbal conceptual processing.
This exists on a spectrum, much like synesthesia — the neurological variation in which senses blend, so that numbers have colours or sounds have shapes. Synesthetes aren't disordered. They're wired differently, and the wiring comes with both advantages and limitations. The same appears to be true of people with reduced inner speech. Nedergaard and Lupyan's own data found measurable costs: participants with low inner speech performed worse on verbal working memory tasks and struggled more with tasks that require comparing the sounds of words — like judging whether two words rhyme without saying them aloud.
The finding doesn't prove that inner speech is optional. If anything, it suggests the opposite — that the voice serves genuine cognitive functions, and its reduction comes at a price. But it does reveal something interesting about the relationship between the voice and the self: people with very little inner narration still have a sense of identity. They know who they are. They make decisions. They hold values. They navigate the world. Their self doesn't depend on being constantly narrated into existence.
The voice, then, may be one way the self is maintained — perhaps the most common way — but not the only way. The self can apparently sustain itself through other means: images, embodied feeling, non-verbal knowing. The narration is a powerful tool for self-construction, but the self it constructs is not made of the narration. Something underneath the voice is holding the identity together, whether the voice is active or not.
What the voice is doing
The inner voice serves real functions. This is worth stating clearly, because the essay could be misread as suggesting the voice is a malfunction. It isn't.
Inner speech supports working memory. You repeat a phone number in your head while reaching for a pen. It aids planning — you rehearse a conversation before having it. It facilitates judgment — you weigh options by talking them through, testing one argument against another, asking yourself which is stronger. It assists in reading — most people vocalise the words on the page silently, and comprehension improves when they do.
These are genuine cognitive functions. Without the capacity for inner language, certain kinds of thinking — particularly evaluation, comparison, and self-correction — would be significantly harder. The voice isn't decorative. It's a tool, and for many of the most complex things humans do, it's the right tool.
But the functions don't explain the commentary.
Nobody needs an inner voice to rehearse a conversation that will never happen. Nobody needs a running evaluation of their performance while walking to the shop. Nobody needs, at 2am, a detailed replay of an argument from three years ago, complete with improved responses that will never be delivered.
The practical functions of inner speech — memory, planning, judgment — account for a fraction of what the voice actually does. The rest is narration. A continuous, unbidden, self-generating stream of commentary that has no practical purpose and cannot be stopped by the person it's talking to.
This excess is where the interesting question lives.
The narration that builds the narrator
Listen to the voice for a few minutes. Not to the content — to the activity. What is it doing, structurally?
It's producing a character.
"I need to send that email." "She probably thinks I was rude." "I should exercise more." "I wonder what he meant by that." Each thought, taken alone, is trivial. But the stream, taken together, is generating something: a me. A protagonist with preferences, opinions, a history, regrets, plans, anxieties, and a continuous sense of being someone specific.
This is what the voice does beneath its practical functions. It constructs the experiencer. The running commentary doesn't just describe your experience — it produces the one who is having it. The "I" that feels so solid, so continuous, so obviously real is, on close inspection, being assembled in real time by the narration.
Consider what happens when the narration pauses. Not when you force it to pause — that's still the voice, wearing a silence costume. When it genuinely stops, briefly, on its own. In a moment of absorption — music, physical exertion, sudden beauty, danger. Something catches you fully, and for a few seconds, the commentary goes quiet.
What happens to the "I" in those moments?
It doesn't disappear. But it changes. It becomes less defined. Less located. Less like a thing and more like a space. The experience continues — you still see, hear, feel — but the narrator who was packaging the experience into a story about a specific someone is temporarily absent. And the experience, without the packaging, is often described the same way by everyone who notices it: lighter. Simpler. More direct. As though a filter has been removed and you're seeing without someone telling you what you're seeing.
Then the voice returns — "that was amazing" or "I should do this more often" — and the character is back. The narrator reclaims the experience, fits it into the ongoing story, and the sense of being a specific someone resumes.
The question behind the question
This is where the essay arrives at something that can't be answered by neuroscience and doesn't need to be.
Who is talking?
Not the brain — that's the hardware. Not you — because you can't control what the voice says, can't predict it, and can't stop it. The voice speaks, and you hear it, and you assume you're the speaker because it sounds like you and uses your vocabulary and talks about your life. But you didn't author it. You didn't choose the next sentence any more than you chose the next sentence of this essay before you read it.
And who is listening?
The voice assumes an audience. It wouldn't narrate if nobody were there to hear. But can you find the listener? Can you locate, in your experience, the one to whom the voice is speaking — as a separate entity, distinct from the voice itself?
When you look for the listener, you find — more voice. More commentary. "Am I the listener?" is just another line of inner speech. The listener and the speaker fold into each other, and what's left is the looking itself. The awareness that was there before the question, during the question, and after it. Not a listener. Not a speaker. Just the space in which the conversation has been happening.
Ramana Maharshi, the Indian sage whose sole method was the question "Who am I?", was pointing at exactly this. Follow the voice inward. Ask who is speaking. And when you look for the speaker, you won't find a person. You'll find awareness — silent, present, already here — in which the voice has been appearing this whole time.
The wallpaper and the wall
The inner monologue is the wallpaper of consciousness. It's been there so long you've stopped seeing it. It runs so continuously that its presence passes for the room itself — as though the wallpaper and the walls were the same thing.
They're not.
The walls are awareness. The wallpaper is the voice. The room — the space in which experience happens — doesn't require narration to exist. It was here before the voice started, in those first moments of infant consciousness before language arrived. It's here during the pauses — the seconds of absorption when the commentary drops and the world becomes suddenly vivid and immediate. And it will be here when the voice eventually stops, as it does every night in deep sleep, and the room holds up fine without a single word to furnish it.
The question for anyone reading this isn't how to tear the wallpaper down. The voice is part of you, it serves functions, it isn't going anywhere, and trying to silence it is just more wallpaper. The question is simpler, and it's available right now, in the time it takes to finish this sentence:
Can you notice the voice as something happening inside the room — rather than as the room itself?
Because the moment you hear the voice as a voice — as an activity occurring within a larger space — the relationship shifts. The voice keeps talking. But you stop being only its audience. You become the room.
And the room was always quiet.
Sources and further reading:
- Lev Vygotsky, "Thought and Language" (1934, trans. 1962) — on the developmental origins of inner speech from external speech
- Marcus Raichle et al., "A Default Mode of Brain Function," PNAS (2001) — on the discovery of the Default Mode Network
- Russell Hurlburt, "Investigating Pristine Inner Experience" — on sampling inner experience and finding inner speech present in roughly a quarter of conscious moments
- Nedergaard & Lupyan, "Not Everybody Has an Inner Voice: Behavioral Consequences of Anendophasia," Psychological Science (2024) — on the spectrum of inner speech and its cognitive consequences
- Fernyhough & Borghi, "Inner speech as language process and cognitive tool," Trends in Cognitive Sciences (2023) — on the functional and developmental dimensions of inner speech
- Ramana Maharshi, "Who Am I?" (Nan Yar?) — on tracing the inner voice to its source
- Alderson-Day et al., "The brain's conversation with itself: neural substrates of dialogic inner speech," Social Cognitive and Affective Neuroscience (2016) — on neuroimaging of inner speech