The influence of bilingualism on cognitive functioning is currently a topic of intense scientific debate. The strongest evidence for a cognitive benefit of bilingualism has been demonstrated in executive functions. However, the causal direction of the relationship remains unclear: does learning other languages improve executive functions or are people with better executive abilities more likely to become bilingual? To address this, we examined 90 male participants of the Lothian Birth Cohort 1936; 26 were bilingual, 64 monolingual. All participants underwent an intelligence test at age 11 years and were assessed on a wide range of executive and social cognition tasks at age 74. The only notable differences between both groups were found for the Simon Effect (which indexes stimulus-response conflict resolution; β=-.518, p=0.025) and a trend effect for the Faux Pas task (a measure of complex theory of mind; ToM, β=0.432, p=0.060). Controlling for the influence of childhood intelligence, parental and own social class significantly attenuated the bilingual advantage on the Faux Pas test (β=0.058, p=0.816), whereas the Simon task advantage remained (β=-.589, p=0.049). We find some weak evidence that the relationship between bilingualism and cognitive functions may be selective and bi-directional. Pre-existing cognitive and social class differences from childhood may influence both ToM ability in older age and the likelihood of learning another language; yet, bilingualism does not appear to independently contribute to Faux Pas score. Conversely, learning a second language is related to better conflict processing, irrespective of initial childhood ability or social class.