The scene
You start a sentence in your head and feel, for a half-second, as if someone else is finishing it. The phrasing is a little too neat. Inner speech has always been a composite — caregivers, books, the last argument. What is new is the volume of a single non-human contributor.
What is, and is not, known
Alderson-Day and Fernyhough (2015) review inner speech as a scaffold for planning, self-control, and self-understanding, with a phenomenology that is varied and hard to measure. Johnson, Hashtroudi and Lindsay (1993) review source monitoring: the fallible process of knowing where a memory came from. Fluency makes misattribution easier.
There is not, at the time of writing, a robust empirical literature on “model residue in inner speech.” Reports from people who write with models all day are phenomenological. They are interesting because they match priming and source-monitoring accounts. They are not a syndrome, and this magazine will not invent one.
For living and designing
Write a messy paragraph before completion. Keep one channel — a notebook, a walk — where no model has spoken. Designers: do not autocomplete the first sentence by default. The first sentence is where a person finds out what they think.
The inner voice was never pure. It can still be yours in the sense that matters: formed under hesitation. Protect that hesitation. It is the sound of a mind arriving.
Evidence & provenance3 sources · production record
Source record
Sources stay inspectable so the boundary between reporting and interpretation remains visible.
- 01Alderson-Day B, Fernyhough C (2015). Inner speech: development, cognitive functions, phenomenology, and neurobiology — Psychological Bulletin, 141(5), 931–965
- 02Johnson MK, Hashtroudi S, Lindsay DS (1993). Source monitoring — Psychological Bulletin, 114(1), 3–28
- 03PSYTECH News editorial constitution — PSYTECH — House method: scene → claim → named evidence → implication. Editorial, not clinical.
Production
Editorial production
Human review
Not recorded for this release