Why we dream: the main scientific theories
There is no single accepted scientific explanation for dreaming. A map of the four main families of theory — and why they are not mutually exclusive.
- There is no single accepted scientific explanation for why we dream, and the possibility that dreaming is a by-product rather than an adaptation remains on the table.
- Four main families of theory: brainstem activity synthesised into narrative, threat rehearsal, memory and emotional processing, and continuity with waking thought.
- These families are not mutually exclusive — dreaming may serve several roles at once, or none.
- Content-analysis research consistently finds most dream content is mundane and tied to the dreamer's actual life.
Ask a sleep researcher why we dream and you will not get one answer. You will get a survey. This is not evasion, and it is not a field that has failed; it is an accurate description of where the science stands. Dreaming is universal, it happens every night to everyone, it has been studied with instruments for seventy years — and there is still no single accepted explanation for what it is for, or whether it is for anything at all.
That last possibility deserves to be on the table honestly. It is scientifically respectable to hold that dreaming is a by-product of how a sleeping brain operates rather than an adaptation with a function of its own — the way the heat from a computer is a real, reliable, entirely purposeless consequence of the work it is doing. Most researchers think dreaming does more than that, but the by-product position is not a fringe one, and any account that presents dreaming's purpose as settled is overstating.
The first family of explanations starts from the brain's own activity. Allan Hobson and Robert McCarley's activation–synthesis model proposed that activity originating in the brainstem during REM sleep is assembled by the forebrain into something narrative — the higher brain doing what it always does, making sense of the input it receives, only with unusual input. This is frequently and wrongly summarised as the claim that dreams are meaningless noise. Hobson's own later work moved considerably beyond that summary, and the article on activation–synthesis in this section covers what he actually went on to argue.
The second family says dreaming is rehearsal. Antti Revonsuo's threat simulation theory proposes that dreaming evolved as a safe space to practise recognising and escaping danger, and points to how over-represented threatening content is in dream reports relative to how often anything threatening happens in modern waking life. A related social simulation idea extends the same logic to rehearsing social interaction. The theory is genuinely interesting and has genuine problems, both covered in its own article here.
The third family ties dreaming to memory and emotion. Sleep demonstrably supports memory consolidation, and REM has a distinctive neurochemistry — noradrenaline at its lowest point of the 24-hour cycle — that has been proposed to let emotional memories be reprocessed with their charge reduced. On this view dreams are the experiential side of overnight processing: not messages, but what it feels like from the inside when the brain is sorting through what happened. This is probably the most actively researched family at present, and it is also the one with the most conflicting results.
The fourth family is the least dramatic and possibly the best supported by sheer volume of data. The continuity hypothesis holds that dreaming is simply thinking, continued into sleep under different conditions. Decades of dream-content analysis — the coding system developed by Calvin Hall and Robert Van de Castle, and G. William Domhoff's extensive work with it — find that most dream content is unremarkable and concerns the dreamer's actual life, relationships and preoccupations. Flying, teeth falling out and being chased are memorable precisely because they are exceptional; the bulk of the record is mundane.
These four families are not mutually exclusive, and treating them as rivals is the most common mistake in popular coverage. Dreaming could easily be a by-product of REM physiology that the brain also puts to use, that draws its raw material from memory consolidation, that is shaped by an evolved bias toward threatening content, and that mostly reflects waking concerns because those are what the brain has to work with. Different theories also explain different things well: the physiological accounts are strongest on why dreams are strange, the content-analysis tradition is strongest on what dreams are actually about.
For a reader arriving from the symbol dictionary, the useful conclusion is this. Science has a great deal to say about the machinery of dreaming and about what people dream, and comparatively little to say about what any individual dream means. Those are different questions, and the article on the science of dream meaning in this section takes the second one on directly.
Frequently asked
Dream symbols on this topic
Was your dream different?
Tell us exactly what you saw — our deeply trained AI will work through every detail and nuance of your specific dream.
Get a personal readingWhat this is based on
- Hobson and McCarley's activation–synthesis model of dreaming
- Antti Revonsuo's threat simulation theory
- The Hall and Van de Castle content-analysis tradition and G. William Domhoff's work
This article is for information only and is not a substitute for medical advice.