Your Dreams Are Not a Search Engine
*The pop-psych revival sold you a conversation with your unconscious. The literature measured a rehearsal effect at sleep onset, and it's small.*
GPB ran a piece this week — "Can a dream help you solve a problem? Here's how to try tonight" — and it's part of a wave. Psychology Today, American Scientist, and a shelf of incubation blogs are all pushing the same pitch: pose a problem before sleep, dream on it, wake up with the answer. It's an old idea wearing new clothes, and the clothes are the problem. The revival has quietly relocated the effect from where every actual number in the literature lives — sleep onset, the hypnagogic drift-off — to REM dreaming, where the mysticism lives. Those are not the same address.
Start with the size, because the size is the honest part. By the cleanest denominator in the literature, dream incubation succeeds on something like 1.4% of attempts. That is a weak priming effect, not a technique. A practice that fails 98-plus times out of a hundred can still be real — priming effects are real — but "real" and "settled REM-dreaming practice you can try tonight" are different claims, and the gap between them is where the marketing lives.
Here's what I find genuinely interesting, though: the shape of the effect is more informative than its size. Two features stand out. There's a visual/associative asymmetry in what incubation actually moves. And there's the failure mode: incubation only works if you engage with the problem while awake first. Sit with that second one. A magic channel to unconscious computation wouldn't need you to do your homework before it picks up the phone. A memory-consolidation process would — because that's what consolidation is, rehearsal and reorganization of material you already fed it. Both features point the same direction: this is your brain filing what you studied, not your unconscious solving what you couldn't.
Which quietly kills the mystical version while leaving a real, testable, small effect standing. And it explains the anecdote stack. Kekulé and his dreamt benzene ring get trotted out in every one of these pieces, and the anecdotes are doing emotional work, not evidential work — a nineteenth-century chemist's retrospective story is not a denominator. The one figure in the revival that looks like evidence is the celebrated 42% success rate from the Dormio work, and I'll flag this honestly: I haven't pulled the primary paper, so this is a bet, not a claim. My bet is that 42% reduces to "cued rehearsal beats no rehearsal" — which we already knew, and which requires no dreams-as-oracle at all. One pass through the primary paper settles it. Until someone does that pass in public, the number is doing the same work Kekulé is.
What survives contact with the literature is modest and actually useful: a rehearsal protocol. Engage with the problem while awake, deliberately, at sleep onset, and let consolidation do what consolidation does. That's a technique. "Ask your dreams a question and wait for the answer" is a séance. The revival is selling the séance and shipping the rehearsal, and the difference between those two products is exactly the 1.4%.