Dreams and creativity — and a myth check

The story that a discovery arrived in a dream is one of the most durable genres in the history of ideas. Some of those stories are well documented. Some are disputed. At least one is almost certainly false and is repeated constantly anyway. Sorting them is the point of this article, because a section about the science of dreams should be able to tell the difference.

The best-documented cases are self-reports, which is worth stating up front. Nobody was recording anyone's brain during these episodes. What we have is what the person said afterwards, sometimes years afterwards, and human memory of dreams is unreliable in the specific ways described in this section's article on forgetting dreams. That does not make the accounts worthless — it makes them testimony rather than data.

Several accounts are strong by that standard. Otto Loewi described dreaming the design of the experiment that demonstrated chemical transmission between nerves — the frog-heart experiment — waking, and, by his account, having to wait for a second night's dream after failing to read his own notes from the first. The work won him a Nobel Prize. Paul McCartney has consistently described waking with the melody of "Yesterday" fully formed, and being so convinced it must be someone else's that he played it to people asking where they had heard it. Mary Shelley attributed the origin of Frankenstein to a waking dream, a vivid vision of the scientist beside the thing he had made, in her own account of how the novel began. Salvador Dalí deliberately cultivated the state: he would doze in a chair holding a key over a plate, so that as he lost consciousness the key fell, the noise woke him, and he could seize the hypnagogic imagery from the threshold — a technique closely related to the material in this section's article on hypnagogia.

August Kekulé's benzene ring belongs in a different category. He described a reverie in which he saw a snake seizing its own tail, giving him the ring structure of benzene. The account is famous, it is his own, and it is disputed: he told it publicly many years after the fact, at an occasion honouring him, and historians have questioned both its accuracy and how much of the insight it actually explains. It is best presented as the famous but contested self-report it is, rather than as an established case.

And Mendeleev dreaming the periodic table is almost certainly apocryphal. It is the most repeated dream-discovery story of all, and the evidence for it is thin — it traces to a much later second-hand account, it sits awkwardly with the extensive documented work Mendeleev did on the arrangement, and it was not part of his own contemporaneous account of how he arrived at it. The story survives because it is a good story. That is not a reason to keep repeating it, and this myth-check is why the article exists.

On the research side, the effect is real but smaller and less romantic. Sleep favours insight on problems the sleeper has already engaged with. Ullrich Wagner and colleagues gave participants a number-reduction task with a hidden shortcut that could be discovered; those who slept between sessions were substantially more likely to find the shortcut than those who did not. That is a controlled demonstration that sleep can produce insight, rather than an anecdote about a Nobel laureate.

There is also active work on steering the process. Dream-incubation research at MIT has used devices that detect the moment of sleep onset and prompt the drifting sleeper with a chosen topic, then collect the imagery that follows — an attempt to make Dalí's key-and-plate technique repeatable and measurable, and to test whether incubated hypnagogic imagery feeds subsequent creative output.

The reasonable conclusion is neither the romantic one nor the debunking one. Sleep genuinely favours insight on material you have already worked on, hypnagogia genuinely produces unusual associations, and the famous stories are a mix of well-documented testimony, disputed recollection and outright myth. What sleep does not do is deliver a discovery to someone who has not done the work.

Frequently asked

Dream symbols on this topic

• ☽ Dreaming about singing — meaning and interpretation
• ☽ Solving a mystery dream meaning — interpretation

What this is based on

• Documented self-reports by Otto Loewi, Paul McCartney and Mary Shelley
• Historical assessments of Kekulé's benzene-ring account and the Mendeleev legend
• Ullrich Wagner's number-reduction task study on sleep and insight, and dream-incubation research at MIT