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.”