How we know
The voice you hear when you think in words is called inner speech, and the standard review is Alderson-Day and Fernyhough 2015 in Psychological Bulletin. It pulls together decades of evidence that covert verbal thinking is produced by the same language system you use to speak aloud. When you talk out loud, your brain plans the words, sends them to the muscles of your lips, tongue, and larynx, and predicts how they should sound. Inner speech runs almost that whole sequence and then stops short of the final motor command. You get the words without the movement.
Imaging backs this up. Inner speech activates the left inferior frontal gyrus, which contains Broca’s area, along with the supplementary motor area and parts of the auditory cortex (Stephane et al. 2021 in Brain and Behavior, consistent with the wider literature). It is broadly the same map that lights up for speaking aloud, just dimmer, because the articulation is withheld. That is why the voice can feel like your own voice. It is built by your own speech machinery, for an audience of one.
It is also less constant than it feels. Using Descriptive Experience Sampling, where a beeper goes off at random moments and you report exactly what was in your awareness, Heavey and Hurlburt 2008 found inner speech present in only around a quarter of sampled moments on average, with huge variation between people. The sense of a non-stop narrator is partly an illusion of memory. You notice the voice when it is there and forget the silence in between.
What it means
The single most important caveat: not everyone has an inner voice. The experience of inner speech runs from near constant to essentially nonexistent across people. In 2024, Johanne Nedergaard and Gary Lupyan published four studies in Psychological Science and gave the low end a name, anendophasia. People who report little or no inner speech are not broken or unusual. They simply think in other ways, with images, abstractions, or wordless intuition.
Their work also found that the difference is not purely cosmetic. Adults with low inner speech did slightly worse on verbal working-memory and rhyme-judgment tasks, which makes sense if you cannot silently rehearse words. So if you have a constant inner monologue, that is common but not universal. And if you do not, you are not missing anything you needed.
The honest version
Inner speech is largely automatic and self-generated, which is why it is so hard to switch off on purpose. You did not choose to start it and you cannot simply will it silent. But the poetic framing of two selves, a narrator and a silent listener, is a way of describing experience, not a claim about two literal brain systems. The defensible science is narrower and stranger than the metaphor. The voice is yours, your brain made it, it is quieter than you think, and some people do not have it at all.