2.9 — The Voice in Your Head

Ask a room full of people what their inner voice sounds like and you will get two reactions. Most will describe it immediately — their own voice, running commentary, sometimes an argument, sometimes a parent's phrasing. And a few will look confused, because they do not have one, and they are usually astonished to learn that everybody else does.

Both are normal. The variation in how people think is far larger than most people assume, and almost nobody discovers this, because the one thing you cannot compare directly is the inside of two heads.

Where the voice comes from

The best-supported account comes from Lev Vygotsky, working in the Soviet Union in the 1920s and 30s, and the developmental sequence he described is easy to verify by watching any small child.

Stage one: speech is between people. A parent says "careful, that's hot" and the child stops.

Stage two: the child says it out loud to themselves. Watch a four-year-old doing a puzzle and you will hear a running commentary — "no, not this one, turn it round". This is not showing off or confusion. The child is using speech as a tool to control their own behaviour, exactly as it was used on them.

Stage three: the speech goes quiet, then goes inside. By around seven it becomes muttering, then silence. The internal voice is an internalised conversation. It was outside first.

Two pieces of evidence support this rather than it being a nice story. Adults revert to stage two under difficulty — you talk out loud when the task is hard or when you are alone and stuck. And measurements during silent inner speech find faint activity in the speech muscles and in the brain regions used for producing speech, which is what you would expect from talking with the volume turned down rather than from some separate faculty.

This origin explains something important. The tone of your inner voice was learned from somewhere. If it is contemptuous, that contempt has a source, and the source is usually a person. This is not a metaphor and it is not blame — it is a developmental account, and it is the reason 15.6 asks you to write down what your inner voice actually says and then ask whose voice it is.

What it is for

Inner speech is not idle noise. It does specific jobs, and knowing them tells you when to encourage it and when to shut it down.

Self-control. Telling yourself to wait, to stop, to check. Children who spontaneously use self-talk do better on tasks requiring restraint, and teaching the strategy improves performance. The instruction in 5.10 to name what is happening out loud works partly through this channel.

Working memory. The phonological loop from 2.3 is inner speech — repeating a phone number to yourself is literally saying it silently. Block it by having someone repeat a syllable aloud continuously and your ability to hold verbal material collapses.

Planning and rehearsal. Running a conversation before you have it. Useful in moderation, and it becomes 5.5 when it will not stop.

Working out what you think. This one is under-appreciated. Ideas are often not fully formed until they are put into words, which is why explaining something to someone reveals holes you did not know were there, and why writing produces thinking rather than merely recording it.

Motivation and instruction. The commentary during difficult physical or mental effort. Sports research finds that trained self-talk measurably improves performance, and that the useful kind is instructional ("keep the elbow high") more reliably than the merely encouraging kind.

The one-word change that does the most work

Here is the most immediately usable finding in this chapter, and it is small enough to remember.

Talk to yourself in the second or third person rather than the first.

Not "I can do this" but "you can do this", or using your own name. In a series of experiments, people who used non-first-person self-talk before a stressful task — a speech, a difficult social situation — performed better as rated by observers, felt less anxious, and afterwards ruminated less about it.

The mechanism appears to be distance. "I am terrified" places you inside the feeling. "Indranil is nervous about this, and that is reasonable" places you slightly outside it, in the position from which you would advise a friend — the same distance effect that makes people give better advice to others than to themselves (1.7).

It sounds slightly ridiculous. It is one of the cheapest interventions in this volume, it takes no time, and the effect has held up across a decent number of studies.

The inner critic

Almost everyone has one, and the two things worth knowing about it are that it is not neutral and that it is not accurate.

It runs on a double standard. The reliable test: write down, word for word, what your inner voice said to you after your last mistake. Then ask whether you would say those words to a friend who had made the same mistake. Nearly everyone finds a gap so wide it is embarrassing to look at.

It believes it is helping. The usual defence of the inner critic is that it keeps standards high. This is checkable, and the check goes against it: self-criticism after failure predicts more avoidance and less subsequent effort, while self-compassion — treating yourself as you would treat a friend in the same position — predicts greater persistence and more willingness to try again. The critic is not maintaining your standards. It is making the next attempt more expensive. 9.6 covers this properly, along with the evidence that self-compassion is not self-indulgence.

And its content is frequently generic. "You always ruin things", "everyone can see you're a fraud". These are not observations about a specific event; they are templates. Noticing that the same seven sentences appear regardless of what actually happened is the beginning of not believing them.

Not everyone has a voice, and thinking comes in several formats

People vary enormously in the format of their inner life, and this only became widely known recently.

Some people have no inner voice at all and think in something more like wordless concepts or images. Some have no visual imagery either — a condition called aphantasia, in which asking someone to picture an apple produces nothing to look at, affecting a small percentage of people. Some think mainly in images, some in a felt sense that has to be translated into words to be examined.

None of these is better, and there is no evidence of a general advantage. What matters is knowing your own format, because advice written for one format is useless in another. If you have no visual imagery, every technique that begins "picture yourself on a beach" is going to fail and it is not your fault — and there are equivalents that work through other channels, which 7.4 provides.

And a related warning about hearing voices. Inner speech that you experience as your own is universal. Hearing a voice that seems to come from outside, or that you do not experience as yours, is a different phenomenon. It is more common in the general population than most people think, it is not automatically a sign of illness, and it becomes a clinical matter when it is distressing, commanding, or accompanied by other changes. 14.6 covers this with the care it deserves.

What to do with this

Find out what your voice actually says, in writing. For three days, whenever you notice self-talk during or after something difficult, write the sentence down verbatim. Not a summary — the actual words. Most people have never done this and are shocked at the result, because the voice is normally too fast and too familiar to be examined. This is also the raw material for the thought record in 15.3, and it is where the practical work of Part 15 begins.

Then apply the friend test to each sentence. Would you say this to someone you care about, in these words, in this tone? Where the answer is no, write what you would say to them instead. That second sentence is not a lie or a softening; it is usually the more accurate of the two, because it is not being generated by a system whose job is to punish.

Switch to the second person under pressure. Before the interview, during the argument, at the start of something you are dreading. Use your own name. It takes no preparation.

Use instruction rather than encouragement when you need to perform. "Slow down and start with the timeline" beats "you've got this" for anything with technique in it.

And give the voice a rest deliberately. Continuous inner narration is not a natural resting state, and one of the things attention practice does (Part 7) is show you what it is like without it, which is quieter than most people expect and considerably less frightening than they fear.

Next: Part 3 turns the machinery of Part 2 into a method. 3.1 starts with what learning physically is, so that every technique afterwards is something you understand rather than something you follow.