How does multilingualism slow down brain aging ?

Hi, hal­lo, hola : three lan­guages, one gree­ting and a slow­down of ageing of the brain. On 10^th Novem­ber 2025, the results of a large-scale epi­de­mio­lo­gi­cal stu­dy conduc­ted in 27 Euro­pean coun­tries were publi­shed in Nature Aging ¹. Its conclu­sions : mul­ti­lin­gua­lism halves the risk of brain decline, a bene­fit on a par with that of regu­lar phy­si­cal acti­vi­ty or res­to­ra­tive dai­ly sleep.

Being able to juggle lan­guages is not just about cros­sing lin­guis­tic boun­da­ries with ease or broa­de­ning one’s social and cultu­ral hori­zons : it is also about loo­se­ning the grip of time on our brains. The hypo­the­sis that the men­tal gym­nas­tics of mul­ti­lin­gua­lism pro­tects the brain is gai­ning cre­di­bi­li­ty. This is the conclu­sion of a large-scale stu­dy conduc­ted by Lucie Amo­ru­so, a neu­ros­cience spe­cia­list wor­king at the Basque Centre on Cog­ni­tion, Brain and Lan­guage (BCBL) in Spain, and her colleagues.

To test the sup­po­sed link bet­ween mul­ti­lin­gua­lism and brain health, the authors ana­ly­sed data from more than 86,000 elder­ly people living in 27 Euro­pean coun­tries. They desi­gned a pre­cise indi­ca­tor of ageing by mea­su­ring the gap bet­ween bio­lo­gi­cal age, esti­ma­ted from a set of life­style and health mar­kers, and chro­no­lo­gi­cal age, i.e. legal age. To objec­ti­ve­ly mea­sure this gap, they crea­ted the ABC index : bio-beha­viou­ral age. The ver­dict was clear, spea­king seve­ral lan­guages slows down the bio-beha­viou­ral clock, while mono­lin­gua­lism speeds it up. In other words, spea­king seve­ral lan­guages reduces a person’s risk of having a men­tal health pro­file that is “older” than their legal age, both now and in the future.

The risk of demen­tia is esti­ma­ted at around 5% for mono­lin­guals, 0.4% for bilin­guals, and remains very low for those who are fluent in three or more languages

For years, evi­dence has been moun­ting that mul­ti­lin­gua­lism seems to nou­rish cog­ni­tive reserve, the brain’s abi­li­ty to bend without brea­king under the effects of time or disease. Howe­ver, it is also true that meta-ana­lyses have so far offe­red an unclear pic­ture, ham­pe­red by over­ly small samples, over­ly indi­rect mar­kers of ageing, impre­cise or sub­jec­tive mea­sures of mul­ti­lin­gua­lism, and poor control of socio-envi­ron­men­tal contexts, some­times dis­tor­ted by stub­born ideo­lo­gi­cal assumptions.

Des­pite this uncer­tain­ty, the link bet­ween lan­guage lear­ning and pro­tec­tion against neu­ro­de­ge­ne­ra­tion has become increa­sin­gly clear. A com­mu­ni­ty stu­dy conduc­ted in 2024 with 1,234 par­ti­ci­pants in the Ban­ga­lore region of India, all over the age of 60 and spea­king Tamil, English or Hin­di² tes­ted the sub­jects’ gene­ral cog­ni­tive abi­li­ties, inclu­ding memo­ry, and used diag­nos­tic tools to assess demen­tia, Alzheimer’s disease and mild cog­ni­tive impair­ment. The degree of pro­fi­cien­cy in one or more lan­guages was mea­su­red in paral­lel. The ana­lyses revea­led a signi­fi­cant effect of bilin­gua­lism on the risk of demen­tia : esti­ma­ted at around 5% among mono­lin­guals, it falls to 0.4% among bilin­guals and remains very low among those who are pro­fi­cient in three or more languages.

Spea­king seve­ral lan­guages the­re­fore helps to streng­then our brain’s abi­li­ty to func­tion smooth­ly, even when struc­tu­ral changes are taking place. This adap­tive facul­ty is known as cog­ni­tive reserve : a cere­bral resource that allows infor­ma­tion to forge new path­ways within neu­ral net­works to coun­ter­ba­lance age-rela­ted dys­func­tions. This pro­tec­tive wor­ka­round is due not only to the deve­lop­ment of cir­cuits used in mul­ti­lin­gual com­mu­ni­ca­tion, but above all to the men­tal flexi­bi­li­ty that this prac­tice requires. Swit­ching from one lan­guage to ano­ther mobi­lises a com­plex net­work of brain areas and requires tem­po­ra­ri­ly muting those that are not being used through a pro­cess of “cog­ni­tive inhi­bi­tion” (see the article “Why pau­sing intui­tive thin­king favours com­plex rea­so­ning”, Poly­tech­nique Insights) ³. This bypass mecha­nism pri­ma­ri­ly involves the regions asso­cia­ted with exe­cu­tive func­tions, which tend to decline with age.

Thus, prac­ti­sing seve­ral lan­guages may help to keep the mind sharp for lon­ger. By com­bi­ning cross-sec­tio­nal and lon­gi­tu­di­nal approaches within large repre­sen­ta­tive cohorts, and by inte­gra­ting the com­plexi­ty of the human expo­some – lan­guages, phy­si­cal acti­vi­ty, social fabric, varying degrees of inequa­li­ty, air qua­li­ty, diet, the role of ins­ti­tu­tions – Amo­ru­so and his col­leagues were able to esti­mate a bio-beha­viou­ral age based on a range of risk and pro­tec­tive fac­tors. The gap bet­ween pre­dic­ted age and legal age thus becomes a sen­si­tive indi­ca­tor, more refi­ned than a simple diag­no­sis of cog­ni­tive decline, for map­ping ageing tra­jec­to­ries at the popu­la­tion level.

The ques­tion of varia­tions bet­ween coun­tries remains. Even when natio­nal contexts are taken into account, the pro­tec­tive effect of mul­ti­lin­gua­lism per­sists. Bet­ter still, model­ling the ABC index allows us to direct­ly com­pare lin­guis­tic sta­tus with edu­ca­tion, phy­si­cal acti­vi­ty and chro­nic diseases, and to iden­ti­fy the spe­ci­fic contri­bu­tion of each. In various envi­ron­ments, albeit limi­ted to Europe, and under rigo­rous controls, mul­ti­lin­gua­lism has been shown to be a power­ful bul­wark against cere­bral ageing. Lear­ning an addi­tio­nal lan­guage the­re­fore appears to be a modi­fiable beha­viour, like other public health levers, and the ABC as an ope­ra­tio­nal tool for obser­ving, on a large scale, how dai­ly expe­rience modu­lates the pas­sage of time in our brains.

This large-scale epi­de­mio­lo­gi­cal stu­dy marks a major step for­ward in the deve­lop­ment of glo­bal public health stra­te­gies that inte­grate cog­ni­tive, social and cultu­ral fac­tors. It should help to shift the tone of a debate that has long been over­ly pola­ri­sed. The hypo­the­sis was cer­tain­ly plau­sible, fuel­led by beha­viou­ral obser­va­tions inhe­ri­ted from our ances­tors, but it was far from una­ni­mous. The most recent lite­ra­ture now invites us to think of mul­ti­lin­gua­lism not as a bina­ry label, but as a conti­nuum of enga­ge­ment : skills, fre­quen­cy of use, active alter­na­tion. Detai­led mea­su­re­ments alrea­dy show how this enga­ge­ment modu­lates cog­ni­tion and neu­ral dyna­mics throu­ghout life, exten­ding theo­ries that link mul­ti­lin­gua­lism, cog­ni­tive reserve and adap­tive control in sub­jects living in a world of constant change.

Demen­tia pre­ven­tion fra­me­works are pla­cing increa­sing empha­sis on the search for adjust­ment variables, and it is time to include the use of mul­tiple languages

Howe­ver, the very scale of the stu­dy by Amo­ru­so and his col­leagues is both its strength and its limi­ta­tion. By relying on pan-Euro­pean demo­gra­phic data, it smooths out the sin­gu­la­ri­ties of expe­rience : lear­ning a lan­guage at school is not the same as navi­ga­ting bet­ween seve­ral lan­guages on a dai­ly basis. This is also what makes the result more stri­king ; des­pite this approxi­ma­tion, the pro­tec­tive effect remains evident. The fin­dings argue for a tru­ly com­ple­men­ta­ry approach – detai­led stu­dies should focus on unders­tan­ding the mecha­nisms, while large-scale epi­de­mio­lo­gy should demons­trate that pro­tec­tion is wides­pread, even when the lin­guis­tic variable is only cap­tu­red in broad terms.

In the age of ins­tant trans­la­tion tools, this abi­li­ty to learn a new lan­guage remains valuable, even as tech­no­lo­gies desi­gned to faci­li­tate glo­bal exchanges gain ground. Unlike many cost­ly inter­ven­tions, spea­king mul­tiple lan­guages is nei­ther a luxu­ry nor a pri­vi­lege ; ari­sing from neces­si­ty, is roo­ted in the com­mu­ni­ty, and seizes oppor­tu­ni­ty. It is part of eve­ry­day life, trans­cen­ding social, cultu­ral and eco­no­mic divides, and offers an inex­pen­sive and sca­lable lever for eve­ryone. If it increases resi­lience to ageing, then pro­mo­ting lan­guage lear­ning in schools, pro­tec­ting mino­ri­ty lan­guages and increa­sing oppor­tu­ni­ties for life­long prac­tice should be as impor­tant as cam­pai­gns to get us to exer­cise more or stop smo­king. Demen­tia pre­ven­tion fra­me­works are increa­sin­gly focu­sing on the search for adjust­ment variables ; it is time to include the use of mul­tiple languages.

One cru­cial ques­tion remains, with signi­fi­cant scien­ti­fic and, above all, prac­ti­cal impli­ca­tions. The gene­ric term “mul­ti­lin­gua­lism” actual­ly covers two very dif­ferent expe­riences – lear­ning new lan­guages requires a one-off, inten­sive effort, whe­reas remai­ning mul­ti­lin­gual requires prac­ti­sing dif­ferent lan­guages simul­ta­neous­ly, main­tai­ning them and kee­ping them alive. Both dimen­sions (lear­ning ver­sus main­te­nance) can shape cog­ni­tive reserve, but not neces­sa­ri­ly through the same mecha­nisms or at the same times. The for­mer is epi­so­dic and deman­ding ; the lat­ter is conti­nuous, invol­ving main­te­nance, adjust­ments and social inter­ac­tion, and requires tena­ci­ty and per­se­ve­rance on the part of the sub­ject. We can expect dif­ferent neu­ro­cog­ni­tive adap­ta­tions from each, both in terms of brain space and time. Research should the­re­fore be conduc­ted on two fronts. On the one hand, inter­ven­tion trials to test whe­ther tea­ching new lan­guages in later life improves cog­ni­tive func­tions, and with what effi­cien­cy, scope and dura­bi­li­ty ; on the other hand, lon­gi­tu­di­nal stu­dies, par­ti­cu­lar­ly in older people, to show how varia­tions in exis­ting mul­ti­lin­gua­lism modu­late the tra­jec­to­ries of ageing. It is not cer­tain that they all fol­low the same curve : reti­re­ment, the shrin­king of social net­works and changes in context can influence lin­guis­tic engagement.

Confir­ming these results with robust pro­to­cols will require more pre­cise tools and long-term lon­gi­tu­di­nal fol­low-ups. The poten­tial gain is consi­de­rable : unders­tan­ding not only whe­ther mul­ti­lin­gua­lism pro­tects, but also bet­ter defi­ning the notion of cog­ni­tive reserve is beco­ming increa­sin­gly neces­sa­ry to coun­te­ract the cog­ni­tive debt pro­du­ced due to use of gene­ra­tive arti­fi­cial intel­li­gence. While the idea is not new, the evi­dence is now gai­ning momen­tum and inter­dis­ci­pli­na­ry sup­port. The centre of gra­vi­ty of the debate is shif­ting, no lon­ger so much a ques­tion of confir­ming the now well-esta­bli­shed bene­fits of mul­ti­lin­gua­lism in slo­wing down brain ageing, but rather of map­ping its mecha­nisms, defi­ning its boun­da­ries and conver­ting its poten­tial into infor­med public poli­cies capable of coun­te­ring the deploy­ment of digi­tal trans­la­tors. As Hen­ry Ford remin­ded us : “Anyone who stops lear­ning a lan­guage is old, whe­ther they are twen­ty or eigh­ty. Anyone who conti­nues to learn remains fore­ver young.”