Sleep has a more powerful role in language-learning than was previously thought. What does this reveal about our night-time brain?
Just after I began work on this article, I had a very fitting dream. I was hosting a party in a hotel suite, with guests from the US, Pakistan, and other countries. Most of the guests were chatting away in English; one or two spoke German, my mother tongue. At one point I couldn’t find my son, and panicked. When I spotted him, I sighed a relieved “Ach, da bist du ja!” – “There you are!”, in German – and gave him a hug.
If you speak more than one language, you may have had similar experiences of them mingling in your sleep. My own dreams often feature English, which I speak in daily life here in London, as well as German, my childhood language. But how and why do our brains come up with these multilingual dreams – and could they have an impact on our real-life language skills?
Decoding our dream languages
At first glance, it may not seem surprising that many multilinguals who juggle different languages during the day, and even people who are only beginning to learn a foreign language, also use those languages in their dreams. After all, the language we speak during the day generally carries over into our nights. A study of deaf people and people with hearing loss found, for example, that they communicated in dreams as they did when awake, through sign language.
A closer look at multilingual dreams reveals a more complex picture, however. For a start, instead of randomly replaying linguistic snippets from our day, our brain appears to mash them up with all sorts of daytime worries, memories and problems. It may even create entire dialogues in an unknown, fantasy language, or in one the dreamers have come across in waking life, but don’t speak (in my dreams, I sometimes have lively conversations in Japanese, a language I’ve studied but failed to master in real life).
Our brains pick up new words through out our lives and they can become mingled together in different languages (Credit: Emmanuel Lafont/BBC)
Many of us seem to categorise our dream languages in certain ways, by person, location or life stage. For example, the people in the dream may speak the languages they would speak in real life, while dreams about one’s childhood home tend to be in one’s childhood language – though the idea of common patterns has to be approached cautiously since there have been only a few, small studies of multilingual dreams. In addition, the dream languages may be layered with questions of culture and identity, as in the case of a Thai-American woman who dreamed about shopping for a dress for her late sister, and debating the choice with her nieces in Thai and English.
There are also linguistic anxiety dreams, in which the speaker struggles to make themselves understood in a foreign language, has to catch a train or plane from one linguistic setting to another, or looks for words in a dream dictionary. A Polish study participant reported dreaming of an English word she couldn’t figure out – “haphazard” – then looking it up when awake. A Croatian participant dreamed of trying and failing to communicate with a stranger in Italian, German and English before realising they both spoke Polish, and laughing with relief.
Sleep researchers say that the exact mechanics and function of such dreams are quite hard to establish, partly because dreams are generally still quite a mysterious phenomenon. What is much better understood, however, is how and why our brains process languages and even learn new words in our sleep. This sheds at least some light on the puzzle of multilingual dreaming.
Crunching words in our sleep
To understand the link between sleep and language, let’s start with just one language: your own. You may think you mastered your native language long ago, but you are actually still constantly updating it. Even adults still learn about one new word every two days in their mother tongue.
“Obviously when we’re children there’s a lot of new word learning, particularly over the first 10 years. But we’re doing this all the time, we just don’t really notice,” says Gareth Gaskell, a psychology professor who leads the sleep, language and memory lab at the University of York.
Even in our own mother tongue, we still learn a new word every two days
When we learn a new word, we continuously update our knowledge around that word until we have a firm grasp of it, Gaskell says. He gives the example of “breakfast”, a word most of us use confidently. But when another, similar-sounding word comes along, it can renew our uncertainty around that existing word.
“At some point in the last five years or so, you would have learned the word ‘Brexit’ [referring to the UK’s vote to leave the European Union], and that’s a strong competitor to ‘breakfast’,” he says.
When the new word “Brexit” competed with the existing word “breakfast” in people’s minds, mix-ups happened. A plethora of newsreaders and politicians came up with phrases like “Brexit means breakfast” and “ploughing ahead with a hard breakfast“. To use the new word appropriately, and distinguish it from similar-sounding words, we need to link it to our existing knowledge, Gaskell says: “And in order to do that, you need to have some sleep.”
It’s during sleep that this integration of old and new knowledge happens. During the day, our hippocampus, which specialises in absorbing information quickly, soaks up new words. At night, it passes the new information on to other parts of the brain, where it can be stored and connected to other relevant information. This helps us choose the right word in any given situation, and suppress competing words.
Tagging the ‘mental lexicon’
That process is essentially the same regardless of whether the word is in a first or second language, according to Gaskell. In the case of multilingual people, foreign words are also stored within that huge mental inventory, and are chosen or suppressed in a similar way.
“You can imagine that you’ve got some sort of tag in your memories,” says Gaskell. “So if you’ve got your mental lexicon for German and English, each of the words you know will be tagged for the language, and you suppress half of those words, and focus on the other half when you’re talking.”
Shortly after starting work on this article, the multilingual author had a dream where she spoke to her son in German at a party (Credit: Emmanuel Lafont/BBC)
Is that what I was doing with my dream of a hotel suite filled with people speaking English and German – sorting through my store of languages, and adding meaningful tags?
It would be a nice explanation, but unfortunately, the integration and consolidation process happens during a phase known as deep sleep, or slow-wave sleep. This phase is characterised by slow brain waves and higher-frequency spindles. Complex dreams like my hotel dream tend to happen during a different phase, known as the Rapid Eye Movement (REM) phase.
“Some people argue that REM sleep has a role to play in this whole consolidation process, and that its role is to tidy things up, and maybe smooth over the rough edges as it were,” Gaskell says. Referring to my dream, during which I slipped away from the party at one point to log on to a virtual BBC team meeting, he says: “That’s a really classic situation, where some of your recent memories are linked in with much longer-term knowledge. It fits really nicely with that story [of dreams helping to consolidate memories]. But it is at the moment pretty hypothetical.”
You can learn words in other languages during sleep, but you do it in a very different way than when you are awake – Matthieu Koroma
What we do know is that aside from processing daytime information, our brain can also learn new words while asleep.
Marc Züst is a research group leader at the University Hospital of Old Age Psychiatry and Psychotherapy in Bern, Switzerland, who specialises in the neuroscience of aging, sleep and memory. He and his collaborators created pseudo-words, like “tofer”, and paired each with a German word such as “Baum” (tree), changing the meaning across participants to ensure the pairing was random and free of any accidental sound associations. They then played the word pairs to participants while they were asleep.
The next morning, they asked them if “tofer” would fit into a shoebox. This roundabout question acknowledged a known limitation of learning any new information when asleep: we can’t generally use that information in a conscious, explicit way when awake.
“They couldn’t consciously reproduce that knowledge and say, ‘tofer clearly means tree’,” Züst says of the participants. “They had more of a gut feeling for whether it was a large or small object.” About 60% correctly answered that “tofer” would not fit into a shoebox.
Crucially, both words – “tofer”, and the German word – had to be played during slow-wave sleep, and specifically, during a peak of slow brain waves. When the researchers missed the peak, the pairing wasn’t learned.
Language, whether spoken, signed or written, marks us apart from other species in the animal kingdom. It has allowed us to communicate, pass on knowledge and learn. But it is also one of the main barriers that exist in human societies. Can we overcome those barriers and use languages to live better – more fully, more adventurously, more in touch with others and ourselves? Let’s Talk tackles some of the biggest discoveries, mysteries and taboos around languages.
Matthieu Koroma, a post-doctoral researcher at the University of Liège in Belgium, who specialises in sleep and cognition, co-authored a number of studies that add to this nuanced picture of how and when we engage with language in our sleep.
“Basically, the message is that you are able to learn [words in other] languages during sleep, even new languages you never heard before, but you do it in a very different way than when you are awake,” he says.
First, he and his team discovered that when we are asleep, we can still tell fake from real language. Sleeping participants were simultaneously played a recording of real speech in their native language into one ear, and meaningless, pseudo-speech into the other. Researchers recorded their brain activity, using electroencephalography (EEG), while this happened. The EEG results showed that the sleeping participants’ brains focused on the real speech, but not the fake one. However, during the dream-intense REM phase, the participants tended to shut out or suppress the incoming speech. Koroma suggest this might have been because the brain was focusing on inner processes: “When we are deeply immersed in dreams, we shut down from things that can perturb our dreams.”
In a separate study by the team, participants were played Japanese words in their sleep, along with sounds that hinted at their meaning. For example, the word “inu” (dog) was played together with a barking sound, and the word “kane” (bell) played along with the sound of ringing bells. Different words were played during two different phases of sleep: light sleep, and the dream-intense REM phase. Again, researchers recorded the participants’ brain activity using EEG.
When awake, the participants were able to correctly associate the words heard during the light sleep phase with relevant pictures, with a better-than-chance outcome – pairing “inu” with the picture of a dog, for example. However, when it came to the words played during the REM phase, the outcome was no different from chance.
“Whenever we investigated REM sleep, so the phase where we have the most intense dreaming activity, we couldn’t find solid evidence that there was learning,” Koroma says. He adds that this doesn’t mean we can’t learn during that phase, just that more research is needed to understand if it’s possible.
Playing the sounds of barking dogs or ringing bells along with the words associated with them during the light phase of sleep increased learning (Credit: Emmanuel Lafont/BBC)
Boosting daytime learning
Does all this mean we can effortlessly learn Japanese in our sleep, as long as we play a language lesson all night long to make sure we catch the right sleep phase?
Not necessarily. It might actually backfire, by disturbing your rest, Koroma says. He also points out that in the study, participants learned the words much faster when they were awake than when they were asleep: “It’s way more efficient when you’re awake.” And they were able to use them more confidently, because they had learned them consciously.
“Wakefulness is good for learning, and sleep is more to replay, not for new language acquisition,” Koroma says. “It’s an interactive process, it’s complementary, meaning that you learn during the day, and during sleep you sort through this information, consolidate some of your memories, and try to put it in new contexts.”
Are there other ways we can use sleep to learn languages?
“The best way to do it is probably to learn a new language before going to sleep, and then play some of the words you just heard, while being asleep,” says Koroma. “Here the results are mostly that if you play them quietly enough, it will boost your learning abilities. But if you play it too loudly, it will actually lower you learning ability. So there is some fine-tuning.”
Züst at the University of Bern recommends studying new words during the day, but at night to “focus on getting enough sleep. Then the brain will do what it needs to do.”
Focus on getting enough sleep. Then the brain will do what it needs to do – Marc Züst
Problem-solving in our sleep
When it comes to the potential role of multilingual dreams in this night-time learning process, the researchers are cautious.
“It’s very, very hard to determine how multilingual dreams might fit into this,” says Züst.
That’s partly because the wider cognitive purpose of dreams is still unclear. One idea, according to Züst, is that they are more of a by-product “of the brain being active, and sorting through memory traces”. That does not mean dreams are completely unrelated to the language-learning process – just that they are perhaps a consequence, rather than the main event.
“It’s entirely possible that during multilingual dreams, the brain is trying to connect those two languages,” Züst says. But the chaotic, individual nature of dreams, and natural languages, makes it difficult to say anything more definitive.
Koroma points out that REM sleep is associated with problem-solving, and emotional regulation. In a similar vein, dreams may allow us to try out new words or phrases in different scenarios, he suggests, or explore emotions around the languages we speak.
Danuta Gabryś-Barker, a professor of psycholinguistics at the University of Silesia in Poland, comes to a similar conclusion in an analysis of multilingual people’s dreams, suggesting that such dreams can express “fears and desires” around learning a foreign language, including the yearning to be a native-like speaker.
That idea would nicely chime with studies showing that wrestling with words or tasks in our dreams may help with creative word-play and problem-solving when awake, as well as emotional processing. But as Koroma and the others emphasise, it is a possibility, not proven fact.
My multilingual dreams remain a bit of a mystery then, at least in terms of their practical function. But understanding my brain’s night-time acrobatics has certainly left me in awe of the hidden effort it takes to learn even a single word. And I did learn one new foreign word over the course of writing this article – though not while dreaming.
It’s “hypnopédie“, French for the act of learning in your sleep. I learned it from Koroma, the researcher in Belgium, who uses it in one of his articles. Several nights have passed since I first came across it. I wonder which tags and connections my night-time brain has attached to it – French, Belgium, sleep and deadline, perhaps? Now that could be the start of an interesting dream.
* Sophie Hardach is the author of Languages Are Good For Us, a book about strange and wonderful ways in which humans have used languages throughout history.
BBC News – February 16, 2023