Language Attrition Research Archive

AAAL 2001.  St. Louis, Missouri.  February 26, 2001

Reactivating a ‘forgotten’ language: The Savings-paradigm applied

Organizers: Kees de Bot, University of Nijmegen and Lynne Hansen, BYU-Hawaii

l. to r.: Kees de Bot, Lynne Hansen, Tomoko Asao, Emiko Yukawa, Georgette Ioup, Machiko Tomoyama

While there is a general awareness that there is only partial retention of foreign language skills, there is hardly any research on the reacquisition of foreign language skills. In this symposium, recent research  language reactivation using the Savings paradigm, in which knowledge that seems to be forgotten is reactivated by short relearning activities, were discussed in relation to general theories on learning and forgetting. The research showed the potential of the Savings paradigm for a pedagogical approach to relearning. The following papers were presented: 


Emiko Yukawa                        Attrition, savings, and reactivation of L3 Swedish not used for 5 years


Georgette Ioup              Exploring age and loss using the savings paradigm


Lynne Hansen &            Beyond vocabulary: Applying the savings paradigm to the relearning of

Tomoko Asao                   Japanese complement structures


Machiko Tomiyama            Detecting a savings effect in longitudinal L2 attrition data


Kees de Bot &                      A longitudinal study on savings with French as foreign language

Vanessa Martens




Emiko Yukawa. Attrition, savings, and reactivation of L3 Swedish not used for 5 years


This study investigates how much of L3 Swedish learned in a natural milieu by an English-Japanese bilingual child is lost/maintained after 5 years of total non-use of the language.  The study also asks how much H's Swedish can be reactivated after some short relearning exercises, and whether the reactivation of a small part of the Swedish language system leads to that of the whole set.

The subject (H, 13;4 at the time of the investigation) spent 1.5 years in Sweden at the age 7;0-8;4.  During his stay in Sweden, he learned some Swedish mostly in a Swedish-medium daycare center he went to for two summers (3 months altogether).  Since he returned to Japan 5 years ago, H had no exposure to Swedish.

            H's Swedish data (conversation and narrative) had been collected in the last month of his stay in Sweden.  This was used as the point of reference.  The attrition/retention of H's Swedish was investigated by word and sentence recognition tests.  Then, his knowledge of the Swedish words, which were not recognized correctly and indicate no relationship with English (non-cognates) were further tapped via multiple savings tests.  The part-set effect and the degree of reactivation of H's Swedish after relearning were addressed by the results of the savings tests as well as the final sentence recognition test.

            H recognized only a few words correctly (9 out of 120) and only two sentences (out of 82) imperfectly.  He also identified/guessed 18 more words correctly in the sentences presented as the recognition test.  The results of the savings tests and the final sentence recognition test are still being analyzed, but it is expected that H will show some savings effect on the words he learned before, considering the results of the previous studies using the savings paradigm conducted by de Bot  & Stoessel (forthcoming) and Hansen, Umeda, & McKinney (2000).



Georgette Ioup.  Exploring Age and Loss Using the Savings Paradigm


            De Bot and Stoessel (1998) introduced a new technique for investigating language attrition - the savings paradigm.  It makes use of brief training in the attrited language as a means to determine whether residual knowledge of the attrited language remains.  Their initial study was successful in bringing to consciousness aspects of vocabulary of which subjects had no previous recollection.  A second researcher who has successfully used the technique to measure loss is Hansen (2000).  Both of these studies focus on the recall of select vocabulary items and address loss where age of attrition is not an issue.  The current study differs from the earlier research using the savings paradigm in two significant ways.  First, it investigates the retention of syntactic and morphological aspects of a language using savings, and second, it explores the relationship of recall to age of attrition onset.

It addresses the following questions:

1.  Is the language equally lost with both ages of onset?

2.  Can the language be equally accessed using the savings training with both ages of attrition onset?

3.  Are aspects of syntax and morphology equally accessed by the two age groups?

            Subjects in the study are two male siblings who began attrition at the ages of 6;9 and 13;1.  At the time of testing they were aged 53 and 59 respectively.  Neither subject had literacy in the attrited language, nor had they any significant contact with the language during the intervening 45 years.  At the onset of attrition the two were observed to have achieved equivalent levels of fluency. Results indicate that even  in the absence of literacy skills, the older leaner retained more of the language before training began, and recalled more after the training.  The younger sibling, who initially recalled almost nothing of the attrited language, was able to remember quite a bit after the savings training.  In terms of its ability to measure syntactic and morphological recall, the technique employed fell short. Problems encountered were discussed.



Lynne Hansen & Tomoko Asao.  Beyond Vocabulary: Applying the savings paradigm to the relearning of Japanese complement structures.


This paper extended the line of research which has recently applied the savings paradigm from cognitive psychology to the relearning of second language lexicon (de Bot & Stoessel, 1999, 2000; Hansen, Umeda & McKinney, 2000). In addition to further examining the savings advantage in the L2 Japanese vocabulary of American adults,  the study broke new ground in examining it in  the relearning of an aspect of Japanese syntax, case-marked complements. 

The English-speaking subjects, having  learned their second language in Japan, had been back in North America for times ranging from 1 to 45 years.  The data collected in two interviews, about a year apart, yielded findings of substantial and enduring savings effects in the relearning of L2 vocabulary. For both the lexical and grammatical knowledge, the ability to benefit from savings became significantly weaker only in the oldest subjects, those who had left Japan over 30 years previously. The current knowledge of words or complement structures related significantly to the ability to use the savings advantage, both presumably influenced by language aptitude. 


Machiko Tomiyama.  Detecting a savings effect in longitudinal L2 attrition data


This paper attempts to track down the extent and degree of productive vocabulary retention based on a four-year L2 attrition data of a Japaneseboy.  The subject, a returnee from the United States, was 8;0 at the time of his return to Japan.  He acquired English as his second language to thelevel of near-native proficiency during his seven years of residence. 

The subject's speech data, which were controlled to elicit targeted vocabulary items, formed the basis of analysis.  A comparison between "old items" and "new items" with respect to the effect of re-exposure orexposure to input were made.  An "old item" is defined here as a word which was once actively produced by the subject but was no longer recalled in a later data collecting session.  A "new item" is defined as a word which was confirmed not to have been acquired prior to the exposure of input.  A qualitative analysis of these active production data revealed that with a brief moment of "incidental instruction" from an interlocutor during a data collecting session, the subject was able to spontaneously produce the "old item" in the subsequent sessions.  He was further able to maintain it until the end of the four-year observation period.  On the other hand, for the "new items," the subject was not able to actively produce them in the subsequent sessions despite the fact that he was able to use them immediately after the exposure and throughout that particular data collecting session.

The results indicate a savings effect of productive vocabulary – the advantage of relearning over learning in terms of long-term retention.  For the "old items," they support the claim made by Bahrick and Phelps (1988) that "a brief exposure is sufficient to boost accessibility very significantly"  (p. 188).  If the claim is further confirmed by large scale experimental studies, it will be an encouraging finding for those foreign language maintenance programs where the attriters' exposure to input outside the classroom is severely limited.


Kees de Bot & Vanessa Martens.   A longitudinal study on Savings with French as a foreign language

On the basis of a number of pilot studies using the savings paradigm, an experiment was set up to test the long-term effects of small amounts of relearning. 60 subjects who had learned French as a foreign language in Dutch secondary education were tested 2 years after they had stopped learning French. After the first test and relearning session, the subjects were retested after 3, 6, 9 and 12 weeks. Words not known in a previous session were tested and relearned again to see what the cumulative effect of the learning sessions was. In the last session it was tested to what extent there was a transfer of receptive vocabulary to productive vocabulary. The findings from the various savings experiments were discussed in the light of current theories on forgetting and relearning.