Psychologist logo
Illustrated dream, by Julia Lockheart

Does dreaming have a function?

Mark Blagrove considers the evidence, and a new theory.

15 March 2024

'I am driving on a motorway or American freeway, with lots of lanes. All is grey. My daughter is in the passenger seat on the left. There is a grey bridge ahead and under it the road is blocked by some cars. I stop before getting to them and get out of my car, going round to the passenger side. There is a group of young men, who are from the city where I live. One man sprays petrol on my car, and strikes a match to set fire to it. I try to stop him, saying he will kill my daughter. They say they are doing this as she wouldn't go out or be friends with one of them.'

That dream, reported in my recent book with Julia Lockheart, The Science and Art of Dreaming, is a good illustration of various theories of dreaming – that they are about simulating threat, or consolidating memories or processing emotions. Alternatively, perhaps dreams are ephemeral images, with no function at all.

This debate has run for many years, but now a new theory proposes that dreams have a group-bonding and empathy function when shared with others, and that this function developed during recent human evolution.

The standard view: information or emotion processing during sleep

Many proposals have been made for possible functions of dreams. For example, in the threat simulation theory (Revonsuo, 2000) people who practise overcoming threats in the dream environment will have an evolutionary fitness benefit in waking life. Alternatively, dreams are proposed to be the experience of the brain undertaking the consolidation of memories, so that dreaming reflects, or is part of, these functional neural processes during sleep (Wamsley & Stickgold, 2011). On this account dreams do not copy previous experiences but instead creatively link specific recent experiences with pre-existing knowledge structures, which allows for the extraction of generalisations across experiences.

The notion that Rapid Eye Movement (REM) sleep, and possibly dreaming, produce novel associations and combinations of recent and past memories has been proposed by many researchers over the years. An early version was by Ernest Hartmann, in his 1995 paper Making connections in a safe place: is dreaming psychotherapy? He proposed that dreaming and psychotherapy involve the freeing and making of associations in a safe place without actual behaviours occurring. For Hartmann, the clearest case for the associative function of dreams is in the associations made between recent memories of a trauma and previous memories that are already in long-term memory.

Such proposals hold that the function of dreams occurs during sleep and does not require the dream to be recalled after waking for the function to have its effect. But, as yet, there's no experimental (as opposed to correlational) evidence that dreaming alters the brain during sleep or causes a change in waking life behaviour. For example, in Wamsley et al. (2010) and Wamsley and Stickgold (2019), improvement in learning task performance across sleep was found to be associated with dreaming of the learning task. However, dreaming of the task in both these studies was also found to be related to poor performance on the task before sleep. Task-related dream content in these studies may thus be emotional residues of waking life concerns of having shown, before sleep, poor performance on a task set by the experimenters. The dream content might thus not be part of functional brain processes during sleep. To show such a causal effect of dreams what is needed is to assign participants randomly to dream content conditions during sleep, but such methods of experimentally altering dream content are not yet available.

The epiphenomenal view: why propose a function for dreams?

For all the hypotheses of dreaming having a function, there remains the alternative, the null hypothesis, that there is no purpose or function to our dreams – even when they are very meaningful and related to waking life fears, threats and memories.

The null hypothesis does not mean that dreams are a scrambled version of waking life memories, with dreaming being a type of delirium. Rather, the null hypothesis allows that dreams can still have some meaning, in that they refer to our individual waking life experiences, and therefore do differ between people, just as cognitions or memories or personalities differ between people. They thus depict, often metaphorically, the conceptualisations and pre-sleep emotional events and concerns of the dreamer, but might have no function when doing so (Domhoff, 2022).

Owen Flanagan's (2000) book Dreaming Souls supports the null hypothesis that dreams do not have a function, and he describes dreams as decorative 'spandrels'. The term spandrel is used in evolutionary theory to describe a functionless feature of an organism. In evolution such by-products might not in themselves be harmful, and so would not be selected out or removed by evolution. If our dreams are functionless, evolution might not have acted to stop us having them, although it may have acted to make them more difficult to remember.

There is neural replay during sleep, in which sequences of neurons that are activated in waking life then replay during sleep, such as has been found for rats after learning a path or maze when awake (Gillespie et al., 2021). However, neural replay has not been shown empirically to be related to dream content, and neural replay is mostly found in non-Rapid Eye Movement (NREM) sleep and wake rather than REM sleep (Findlay et al., 2020).

As the evidence is not yet in, we don't yet know whether dreaming has a function. But then we don't know whether waking consciousness has a function either. Jeffrey Gray (2006) considered that waking consciousness is a way for us to scan for incongruities or unexpected events in the environment, and that consciousness has the function of drawing our attention to things that we did not expect to be happening. It could indeed be that dreaming has a similar role of alerting us to incongruities, or to things we have not quite fully noticed yet in waking life. But the epiphenomenal view, that dreams have no lasting effect on the brain during sleep, remains very plausible.

Social and adaptive effects of sharing dreams

Aside from this debate about functions or effects of dreams during sleep, there are known to be beneficial effects of sharing dreams. In Schredl and Bulkeley's (2019) diverse sociodemographic and ethnic background online survey, 23 per cent of the sample reported sharing dreams at least once per week. For Olsen et al. (2013), the main motivations for sharing dreams were, in order of importance, 'Entertainment', 'I want to understand what the dreams mean', and 'To let the other person know what is happening in my mind'. Dream-sharing can bring people closer together (Ijams & Miller, 2000; Vann & Alperstein, 2010), through providing a forum for self-awareness and self-disclosure (Duffey et al, 2004; Schredl & Schawinski, 2010).

As a result of our experiences in holding group dream discussions, Julia Lockheart and I hypothesised that dream-sharing might increase empathy between people. In Blagrove et al. (2019), we reported that trait empathy is significantly correlated with dream telling frequency, with frequency of listening to others' dreams, and with having a positive attitude towards dreams. Trait empathy was measured by the Toronto Empathy Questionnaire, which has 16 items, examples of which are: It upsets me to see someone being treated disrespectfully; and I become irritated when someone cries. This correlational study concerned empathy as a trait, which for a person is their long-standing level of empathy towards people in general. In the same paper we then addressed the relationship of dream-sharing to state empathy, which is the level of empathy that someone feels towards someone else at a particular time. In this we differentiated between empathy of a dream-sharer towards the person discussing the dream with them, and empathy of the discusser towards the dream-sharer.

We recruited 27 pairs of participants. Each pair applied to take part together, as friends or in a relationship, knowing that one would be sharing dreams and the other of the pair would discuss the dreams with them. The sharer was chosen as the member of the pair who recalls dreams the more often. At the start of the study each participant completed a state empathy scale, regarding their empathy towards the other member of the pair, this gave their baseline empathy score. The items included: My friend's / partner's emotions are genuine; I can feel my friend's / partner's emotions; I can see my friend's / partner's point of view; I can understand what my friend / partner goes through.

Upon having a dream, the dream-sharer arranged to meet the other member of the pair as soon as possible to discuss the dream with them. Due to the need for untrained participants to quickly learn and apply a dream exploration method, we taught them the technique devised by psychiatrist Montague Ullman (2006) for use in lay-person dream groups. The dream would be discussed for 15 to 30 minutes. Up to four dreams could be discussed, and after each discussion the participants again completed the state empathy questionnaire. The score on the last of these was used as their post-intervention measure, and compared to their baseline empathy score.

We found that dream discussers had a significant increase in empathy towards the dream-sharers as a result of discussing dreams, with a medium effect size. The dream-sharers had no significant change in empathy towards the discusser, no doubt because the dream-sharer is addressing their own dream and own life during the discussion process. An increase in empathy for both members of a pair would thus need them to take turns in sharing and discussing. We then replicated the dream-sharing and empathy effect in our paper Dream sharing and the enhancement of empathy: Theoretical and applied implications (Blagrove et al., 2021).

The main limitation to these findings is that there was no comparison condition in which some narrative material other than a dream report is used to elicit a meaningful discussion. Comparison conditions in future work could be the discussion of a recent significant event in the life of the dreamer, the dream-sharer telling someone else's dream, or telling a story based on an ambiguous photograph or drawing, or even talking about their favourite film. But whatever the effects on empathy for such comparison conditions, dreams are still valid and useful stimuli for discussion even if the outcomes and benefits from discussions are no higher than for other texts used to stimulate discussion. And it must be remembered that people often wake with a dream in mind that they want to tell.

The personal and social benefits of such dream-sharing thus do need to be investigated further, with comparison conditions that present or generate narratives other than dreams as the basis for discussion. But our conclusion is that increased dream-sharing across society might counteract current societal decreases in empathic concern and perspective taking, the main two components of empathy (Konrath et al., 2011).

Explaining the empathy effect: dreams, fiction and Human Self-Domestication

Dream-sharing may increase empathy because the dream acts as a piece of fiction, which is explored by the dreamer and others as part of the sharing process. Dreams would thus be like literary fiction, which has itself been shown to elicit empathy towards the characters that the fiction describes (Oatley, 2016). Dreams, like fiction, are a simulation of social experience, a term used by Mar and Oatley (2008) to describe fiction in their paper The function of fiction is the abstraction and simulation of social experience, where they state that 'engaging in the simulative experiences of fiction literature facilitates the understanding of others who are different from ourselves, and can augment our capacity for empathy and social inference.'

But why does our brain during sleep produce these imagined simulations of waking social life?  We speculate that dreams might have these characteristics because, across human evolution, there may have been a selection for fictional and story-like aspects of dream content that support a self-disclosure and empathy-eliciting function. This includes selection for the highly social and emotional characteristics of dream content. The functional and evolutionarily adaptive consequences of dream content would then occur after sleep, as a result of the sharing of dreams, and taking advantage of the long REM period that occurs for biological reasons near the end of the night, and which causes longer dreams to occur then. A rudimentary form of dreaming might have originated across many animal species, including birds and mammals, both of which have REM sleep, for memory consolidation or threat rehearsal or other functions, or indeed as no more than a spandrel, an epiphenomenon of sleep. But when dream-sharing started in humans, following the development of complex speech, new selective pressures would have begun for the components of dream content. As described by Barrett (2007), if spandrels become useful they can then become subject to natural selection.

Dream content that supports empathy and bonding when the dream is shared may thus have been selected for during early human evolution. This would have occurred on a timescale similar to that for the evolution of human language and story-telling, which support group cohesion and cooperation in humans (Smith et al., 2017), and which is estimated by Pagel (2017) as commencing from 150,000 to 200,000 years ago. These dream content characteristics that support empathy would also have been subject to sexual selection, where an inherited feature leads to the individual being more attractive to the opposite sex for mating (Verweij et al., 2014), as occurs for artistic virtuosity (Miller, 2000) and story-telling (Smith et al., 2017).

The wider theoretical context for this places dream-sharing as part of human self-domestication, which is a major theory of human social evolution that proposes that there has been a selective pressure for humans having reduced emotional reactivity, and, in particular, reduced aggression within their social group (Hare, 2017). This selection has resulted in humans exhibiting prosociality, self-control, tolerance, co-operation, empathy, and the ability to mentalise, that is, to recognise what others perceive, feel, intend and know.

We suggest that the function of dreams thus resides in their waking use, and that dreams are like blushing, a way for self-disclosure to be supported, for the benefit of the group, even if such self-disclosure can sometimes be uncomfortable for the individual. Remembering and telling dreams is thus essential to this function, in contrast to the theories of dream function reviewed above, where the functional effects occur during sleep, and occur also for unrecalled dreams.

This article is adapted from chapters 12, 13 and 17 of Blagrove, M. and Lockheart, J. (2023), The Science and Art of Dreaming. Routledge. The image above is taken from the book, by Julia Lockheart.

The British Psychological Society has a special bundle offer for World Sleep Day, saving 25% on two webinars. Find out more.


Barrett, D. (2007). An evolutionary theory of dreams and problem-solving In D. Barrett & P. McNamara (Eds.), The new science of dreaming, Vol. 3. Cultural and theoretical perspectives (pp. 133–153). Westport, CT: Praeger.

Blagrove, M., Hale, S., Lockheart, J., Carr, M., Jones, A., & Valli, K. (2019). Testing the empathy theory of dreaming: The relationships between dream sharing and trait and state empathy. Frontiers in Psychology, 10, 1351.

Blagrove, M., & Lockheart, J. (2023). The science and art of dreaming. Routledge. 

Blagrove, M., Lockheart, J., Carr, M., Basra, S., Graham, H., Lewis, H., Murphy, E., Sakalauskaite, A., Trotman, C., & Valli, K. (2021). Dream sharing and the enhancement of empathy: Theoretical and applied implications. Dreaming, 31, 128–139.

Domhoff, G.W. (2022). The neurocognitive theory of dreaming: The where, how, when, what, and why of Dreams. MIT Press.

Duffey, T. H., Wooten, H. R., Lumadue, C. A., & Comstock, D. C. (2004). The effects of dream sharing on marital intimacy and satisfaction. Journal of Couple and Relationship Therapy, 3, 53–68.

Findlay, G., Tononi, G., & Cirelli, C. (2021). The evolving view of replay and its functions in wake and sleep. Sleep Advances : A journal of the Sleep Research Society1, zpab002.

Flanagan, O. (2000). Dreaming souls: Sleep, dreams and the evolution of the conscious mind. Oxford University Press.

Gray, J. (2006). Consciousness: Creeping up on the hard problem. Oxford University Press.

Hare, B. (2017). Survival of the friendliest: Homo sapiens evolved via selection for prosociality. Annual Review of Psychology, 68, 155-186.

Hartmann, E. (1995). Making connections in a safe place: Is dreaming psychotherapy? Dreaming, 5, 213–228.

Ijams, K., & Miller, L. D. (2000). Perceptions of dream‐disclosure: An exploratory study. Communication Studies, 51, 135–148.

Konrath, S. H., O'Brien, E. H., & Hsing, C. (2011). Changes in dispositional empathy in American college students over time: A meta-analysis. Personality and Social Psychology Review, 15, 180–198.

Mar, R. A., & Oatley, K. (2008). The function of fiction is the abstraction and simulation of social experience. Perspectives on Psychological Science, 3, 173–192.

Miller, G. (2000). The mating mind: How sexual choice shaped the evolution of human nature. New York: Doubleday and Co.

Oatley, K. (2016). Fiction: Simulation of social worlds. Trends in Cognitive Sciences, 20, 618–628.

Olsen, M. R., Schredl, M., & Carlsson, I. (2013). Sharing dreams: Frequency, motivations, and relationship intimacy. Dreaming, 23, 245–255.

Pagel, M. (2017). Q&A: What is human language, when did it evolve and why should we care? BMC Biology,15, 64.

Revonsuo, A. (2000). The reinterpretation of dreams: An evolutionary hypothesis of the function of dreaming. Behavioral and Brain Sciences, 23, 877–901.

Schredl, M., & Bulkeley, K. (2019). Dream sharing frequency: Associations with sociodemographic variables and attitudes toward dreams in an American sample. Dreaming, 29, 211–219.

Schredl, M., & Schawinski, J. A. (2010). Frequency of dream sharing: The effects of gender and personality. American Journal of Psychology, 123, 93–101.

Smith, D., Schlaepfer, P., Major, K., Dyble, M., Page, A. E., Thompson, J., Chaudhary, N., Salali, G. D., Mace, R., Astete, L., Ngales, M., Vinicius, L., & Migliano, A. B. (2017). Cooperation and the evolution of hunter-gatherer storytelling. Nature Communications, 8, 1853.

Ullman, M. (2006). Appreciating dreams: A group approach. (Originally published by Sage, 1996.) New York: Cosimo-on-Demand.

Vann, B., & Alperstein, N. (2010). Dream sharing as social interaction. Dreaming, 10, 111–119.

Verweij, K. J. H., Burri, A. V., & Zietsch, B. P. (2014). Testing the prediction from sexual selection of a positive genetic correlation between human mate preferences and corresponding traits. Evolution and Human Behavior, 35, 497–501.

Wamsley, E. J., Tucker, M., Payne, J. D., Benavides, J. A., & Stickgold, R. (2010). Dreaming of a learning task is associated with enhanced sleep-dependent memory consolidation. Current Biology, 20, 850–855.

Wamsley, E. J., & Stickgold, R. (2011). Memory, sleep and dreaming: Experiencing consolidation. Sleep Medicine Clinics, 6, 97–108.

Wamsley, E. J., & Stickgold, R. (2019). Dreaming of a learning task is associated with enhanced memory consolidation: Replication in an overnight sleep study. Journal of Sleep Research, 28, e12749.