Résumé
Since attrition is generally defined as non-pathological loss of a language, comparisons with acquired language disorders, namely Alzheimer's Disease (AD), however fruitful they may be, are a largely neglected area in attrition research. One of the characteristics of neurogenerative diseases is the gradual continuous loss of cognitive skills raising theoret ical questions which are also highlighted by language attrition research. The vulnerability of languages acquired at different moments of life (L1, L2, L3…) has received most atten tion. Another question concerns the evolution of cognitive skills related to language con trol in demented patients as reflected in the specificities of code-mixing behaviour in bilinguals with AD. The hypothesis of protective effects of bilingualism in healthy and pathological cognitive ageing is then discussed. We suggest that further taking into ac count of the interaction between memory and language in cognition and language pro cessing, as in studies on AD, may be beneficial for attrition research.