Illiteracy is associated with poorer cognitive performance and increase risk of dementia among older adults, whereas multilingualism may be associated with a cognitive advantage and reduced dementia risk. It is unclear whether a multilingual cognitive advantage exists among illiterate older adults, such that multilingualism modifies the negative impact of illiteracy on cognition. Moreover, multilingualism is typically treated dichotomously, without acknowledging vast differences between multilingual individuals. This study examines the extent to which illiteracy and multilingualism are associated with cognition among older adults with low to no education in India. Additionally, we evaluated whether aspects of multilingualism (age of acquisition [AoA], language use, and self-reported proficiency) are associated with cognitive functioning after adjusting for relevant covariates.

Participants with <=4 years of education were included from the Longitudinal Aging Study in India  - Diagnostic Assessment of Dementia (LASI-DAD), and completed a comprehensive cognitive battery assessing executive function, language, memory, and visuospatial domains. The analytic sample included 2,553 older adults (Mage=70.31, SD=8.01, 65% female, 71% from rural areas), of which 2,137 were illiterate and 605 were multilingual (>=2 languages). A subsample of multilingual participants (n=121) completed a multilingual history questionnaire which characterized AoA (in childhood vs later), use (daily vs. less than daily), and proficiency (high vs low, averaging across speaking and understanding) across all languages. We evaluated the effects of illiteracy and multilingualism as well as their interaction on cognitive domain scores, adjusting for age, sex, years of education, and rurality.

Illiterate participants performed worse than literate participants on executive function (β=-0.448, [-0.522,  -0.374]), language (β=-0.568, [-0.648,  -.487]), memory (β=-0.312, [-0.405,  -0.219]), and visuospatial (β=-0.339, [-0.417,  -0.262]). Multilingual adults performed better than monolinguals in executive function (β=0.149, [0.093, 0.205]), language (β=0.091, [0.03, 0.153]), and visuospatial (β=0.094, [0.037, 0.152]), but not within memory (β=0.007, [-0.062, 0.076]). There was a significant interaction of multilingualism and illiteracy within executive function and memory. The association of multilingualism on executive function was stronger among literate (β=-.272, [.131, .412]) than illiterate participants (β=.114, [.055,  .173]), whereas – within memory – there was an association among literate (β=.256, [.085, .427]), but not among illiterate participants (β=-.054, [-.128,  .02]) Daily multilingual use was associated with higher language scores (β=0.442, [0.188, 0.697]), whereas neither AoA nor self-reported proficiency were associated with cognition (all p’s>.05).  

Among older adults in India with low to no education, illiteracy was associated with worse cognitive performance, whereas multilingualism was positively associated with cognition among both illiterate and literate adults. However, where literacy and multilingualism seem to be synergistic, multilingualism does not appear to modify the negative association of illiteracy on cognition. While multilingualism did not weaken the impact of illiteracy on cognition, results suggest that both early life educational experiences and cognitively stimulating activities like multilingualism can positively impact late-life cognition. Among multilingual adults, daily multilingual use was associated with better language performance, but neither age of acquisition nor proficiency were associated with cognition. Longitudinal studies are needed to evaluate the effect of illiteracy, multilingualism and frequency of language use on rate of cognitive decline and incident dementia.