Talk by Merel Keijzer: March 4th

The causal mechanisms underlying first language reversion in elderly

bilingual migrants


Due to increased international mobility and longer life expectancies,

developed countries  are increasingly characterized by a growing

proportion of multicultural – and multilingual – elderly.

According to self-reports and reports of caretakers, aging migrants

tend to show a preoccupation with the culture of origin, often also

 exhibiting a return to the language they may have hardly used for

decades (Schmid & Keijzer, 2009). This development is known as

language reversion (de Bot & Clyne, 1989). In the past,  reversion has

 been ascribed to socially-induced changes: the use of the L2 becomes

less necessary due to retirement and in the absence of children

(Clyne, 1977), but this account leaves unexplained the fact that

reversion also occurs in elderly migrants who are still actively

taking part in the work force and cannot use the L1 at home due to

their spouses not speaking the language.


This paper revisits the topic of reversion on the basis of two age

 groups of elderly Dutch migrants in Australia: 60-65  and 75+ years old and

compares their performance against a baseline of 40-45 year-old Dutch

migrants. All participants were subjected to a test battery of 

language tasks measuring overall proficiency, grammaticality judgments

and vocabulary, but also to a number of working memory and executive

functioning tasks. Recent  cognitive aging research has revealed

age-related decrements in cognitive control and working memory (Burke

& Shafto, 2008), which in turn might explain shifts in language

dominance. Impaired inhibitory control, for instance, may result in

more L1 interferences in the L2 and vice versa. The scores on the

language tasks and cognitive measures were correlated and subjected to

a regression analysis.


The preliminary results from this investigation show that language reversion is

nowhere near as common as has been reported in previous literature and

self-reports.  Those subjects who did show a reversion to the

L1 produced similarly low scores on most of the working memory and

executive functioning measures, suggesting that reversion is very much

a cognitive control phenomenon.  Moreover, the results show that

reversion is most likely to occur in those subjects whose command of

the L2 was never strong to start with, heavily tying in with the

predictor variable of educational level. An active lifestyle was also

found to contribute significantly to the model.