So, far, many definitions of language forgetting have been provided. It could
not only refer to the loss of any language skills due to some physical or pathological
damage on the brain but also refer to the non-pathological, gradual loss through
lack of practice or exposure (Schmid, 2006). Many terms,
such as attrition, degradation, forgetting, loss and erosion, were used to refer
to this phenomenon. This study prefers forgetting in order to keep consistent
with studies in the field of psychology.
Four areas of language forgetting were discussed and explored. They could be
summarized in Table 1 based on what language (L1 or L2) is
forgotten and where (L1 or L2 context) the language is forgotten. The description
of language forgetting based on these criteria is referred to as the van
Els taxonomy (Van Els, 1986).
As illustrated in Table 1, L1 loss in an L1 environment can
be found among people with dementia or those with aphasia in the situation of
native language. L1 loss in an L2 setting can be found among immigrants who
lose their first language in the new environment. L2 loss in an L1 setting is
usually observed in individuals who have lost the ability to use an L2 that
was perhaps studied at school in their L1 setting. Finally, L2 loss in an L2
environment is most commonly observed among immigrant communities without formal
training in or immediate access to their L2 who lose that L2 as they age and
revert to their L1. This study mainly concentrates on the forgetting of L2 (English)
learners in an L1 (Chinese mandarin) language environment.
The most prominent theory to explain the retention of language proficiency
is the threshold hypothesis proposed by Neisser (1984),
stating that if learners can command a level of language ability up to a critical
threshold, then it will get resistant against forgetting, i.e., the forgetting
process will be hindered or slowed down.
|| Taxonomy of language forgetting
Studies on this hypothesis argue that threshold hypothesis should shed some
light on future forgetting studies (Grendel, 1993; Weltens
and Grendel, 1993; Weltens et al., 1989).
The inverse hypothesis claims that the higher the level of the language proficiency
is, the slower the forgetting process will be. This view is supported by many
researchers (Weltens et al., 1989; Edwards,
1977; Snow et al., 1984; Vechter
et al., 1990), though there are still some disagreements. In order
to study the inverse hypothesis to a further extent, Neisser
(1984) thought that participants in Bahrick and Phelps
(1988) might have attained to a critical threshold making them
less subject to forgetting. This suggested that there was a critical point
in overall language proficiency below which forgetting is rapid and extensive,
but above which, a large proportion of the initially acquired material is retained.
It is claimed that those who have passed the threshold level must have developed
a systematic ability to understand and internalize the knowledge structures,
which is called a schema by Neisser (1984). He has
argued that having this schema would enable the learner to be less vulnerable
to forgetting. On the contrary, forgetting tends to be faster when information
is made up of separate elements.
The distinguished linguistic expert Michel Paradis voiced his viewpoint clearly:
attrition is the result of long-term lack of simulation (Paradis,
2007). He has put forward a rule named the Activation Threshold Hypothesis
(ATH) which sounds reasonable and well accepted in the field of language forgetting.
The ATH is on the basis of the theory that, when on some level of knowledge
proficiency in human brain, humans feel easy to retrieve the information they
desire. However, if the level is not reached, it will be difficult to retrieve.
An example is that at times it is hard to remember even very familiar issues.
(This is referred to as the tip-of-the-tongue state). This is because accessing
something that is stored in memory needs a certain number of neural impulses.
The more frequently the item has been used before, the less effort it is needed
to activate it again. However, if something is not accessed for a long time,
the amount of energy that is necessary to access it again slowly goes up- that
is, the Activation Threshold increases (Schmid, 2011).
A bilingual who speaks his or her second language every day but has not used
the first for a long time,therefore has words and structures that belong to
the L2 which are highly active and easy to access but the corresponding bits
of the L1 may have a very high Activation Threshold. This is why the L2 can
often get in the way when a speaker attempts to use the L1 (Schmid,
This study focuses on the forgetting of English as an L2 among healthy English
language learners who have learned English for some time with the aid of computer
technology and later have not learned English for two years. The research question
raised in this study is: Does threshold hypothesis hold water?
This study attempts to integrate qualitative research method into the quantitative
one. The data were obtained from the test scores of 38 participants who had
undergone computer aided English instruction and had been beyond English formal
education for two years. All of them did not need to use English when working.
Among them, 19 successfully passed College English Test Band 6 in China (CET6),
while the remaining 19 participants failed to pass it. All of the participants
took CET6 (the test in June, 2010) at two time points, i.e., before and after
two years. The test settings and requirements were both the same. The degree
of contact with English made the only difference. The participants had sustained
computer aided English education for three years before June, 2010. Later, however,
they have been beyond formal English education for two years. The timeline of
the tests is shown in Fig. 1.
As shown in Fig. 1. participants took both pre and post tests
with two years break. The pretest was taken in June, 2010 in Room 204, Teaching
Building 2, Xianlin Campus, Nanjing University of Posts and Telecommunications
(NJUPT); the post test was taken in the same room in June, 2012. The test scores
were entered into SPSS 13.0 and analyzed through computing.
CET 6 has experienced convincible validation and a strong reliability has been
established to test English learners short conversation and passage listening
comprehension, speed and in-depth reading comprehension, cloze and writing skills.
Nearly 20 years has witnessed CET6s wide application in evaluating learners
English proficiency since it was born in the 20th century in China. The Ministry
of Education of China launched this testing project and made it popular in China
to measure learners comprehensive English proficiency in terms of listening,
speaking, reading and writing skills (Ministry of Education of the Peoples
Republic of China (MEPRC, 2001).
|| Timeline of pre and post tests
Yang and Weir (1998) conducted an empirical study whose
results supported the validity and reliability of CET6 measurements of listening,
speaking, reading and writing skills.
CET 6 is composed of writing, reading, listening and cloze test items:
Writing task: For the writing part, participants are allowed 30 min
to write a short essay on the topic of due attention should be given to spelling.
You should write at least 150 words following the outline given below:
||Nowadays, many students pay less attention to spelling when
||The reasons for this phenomenon are...
||In order to improve spelling, I think...
In addition, the translation task is one component of writing. Participants
are required to complete the sentences by translating into English the Chinese
given in brackets.
Reading comprehension: Two components were included in this part as
||Skimming and scanning: In this part, participants will
have 15 min to go over the passage quickly and answer the questions. For
questions 1-7, they should choose the best answer from the four choices
marked A), B), C) and D). For questions 8-10, they should complete the sentences
with the information given in the passage
||Reading in depth: This part includes two sections, sections A and
B. In section A, there is a short passage with 5 questions or incomplete
statements. Participants are required to read the passage carefully and
then answer the questions or complete the statements in the fewest possible
words. In Section B, there are 2 passages. Each passage is followed by some
questions or unfinished statements. For each of them there are four choices
marked A), B), C) and D). Participants should decide on the best choice
||Listening comprehension: This part involves 7 short conversations,
2 long conversations, 3 short passages and 1 compound dictation. Participants
should decide on the right choice according to the recording
||Cloze: In this part, there are 20 blanks in the following passage.
For each blank there are four choices marked A), B), C) and D) on the right
side of the paper. Participants should choose the one that best fits into
The writing scoring process is based on national standard scoring criteria
of CET6 set by Ministry of Education of China as follows.
The full score is 15 points. The scoring consists of five levels: 2 points,
5 points, 8 points, 11 points and 14 points. There are two standard samples
for each level.
Scorers have all gone through strict training and been evidenced qualified.
They made their judgments on the basis of criteria and samples. If a writing
is similar to a sample (such as 8), then 8 points will be given; if it is slightly
better than or slightly inferior to the sample, the scorer can add (9 points)
or deduct one point (i.e., 7 points) but shall not add or deduct a half point.
||Two points: Sentences are incoherent; thoughts are
in disorder; there are many fragmental or wrong sentences, most of which
contain serious mistakes
||Five points: The writing basically keeps to the title;
thoughts are unclear and incoherent; there are many serious language errors
||Eight points: The writing fundamentally keeps to the title; thoughts
are not clear and coherent enough; there are many language errors, some
of which are serious
||Eleven points: The writing keeps to the title; thoughts are clear
and coherent; there are a few language errors
||Fourteen points: The writing keeps to the title; Thoughts are very
clear and coherent. There are few language errors
There are totally three independent scorers to evaluate writing task. The average
score is considered as the final one for each participant. In case there is
any wide gap between any score and the average, the score will be removed if
found carelessly decided.
Aiming to gain reliable data, the same scorer who received the same training
and scoring practice was selected to mark both pre and post writing tests. They,
willing to participate in the study, were all richly experienced in writing
scoring and acquainted with specific items in CET6.
After data of all tests were obtained, they were then entered into computer,
processed in the program paired-samples T tests in SPSS 16.0. The results were
summarized in Table 2.
|| Data of both pre and post tests
|Prelistening1, Prereading1, Prewriting1, Precloze1: Tests
taken by the participants who did not pass CET 6 before two years (2010-6),
Postlistening1, Postreading1, Postwriting1, Postcloze1: Tests taken by the
participants who did not pass CET 6 after two years (2012-6) Prelistening2,
Prereading2, Prewriting2, Precloze2: Tests taken by the participants who
passed CET 6 before two years (2010-6), Postlistening2, Postreading2, Postwriting2,
Postcloze2: Refers to the tests taken by the participants who passed CET
6 after two years (2012-6). SD: Standard Deviation
As shown in Table 2, the first column indicates the compared
pairs. The second one shows the means of different items. The third one presents
the differences of pairs. The fourth one reveals the standard deviations of
pairs. The fifth one shows the values of T and the last stands for the significance
levels. It could be found that participants failing to pass CET 6 showed significant
language attrition in terms of listening (t = 3.21, p = 0.02), reading (t =
5.14, p = 0.01), writing (t = 2.15, p = 0.03) and cloze (t = 7.25, p = 0.01).
On the contrary, participants who passed CET 6 failed to show any significant
language forgetting in terms of listening (t = 2.14, p = 0.06) and cloze (t
= 0.72, p = 0.42). They, however, showed significant language forgetting in
writing skills (t = 2.83, p = 0.01). Furthermore, it is uncommon to find that
they failed to show significant language forgetting in reading comprehension.
On the contrary, their reading comprehension significantly gained after two
years break (t = 2.22, p = 0.04).
The threshold hypothesis in language forgetting seemed selective rather than
overall. As could be seen, participants with higher initial proficiency did
not maintain language skills in all linguistic components. Reading comprehension
skills significantly gained although skills in listening and cloze decayed.
Receptive skills, such as reading comprehension might be more resistant against
forgetting than productive skills. Writing skills were considered as productive
skills and thus received significant decrease.
Reading skills having gained after certain time of nonuse were not in conformity
with initial consideration. However, a further consideration would bring reasonable
understanding to this phenomenon. Despite the fact that it was unnecessary for
participants to learn English any more after graduation, they could not keep
beyond any contact with English all the time in their daily life. Internet is
nowadays a convenient medium to get access to English, which consists of various
kinds of information in English. Participants have to learn English only if
they want to retrieve and understand the abundance of information displayed
on the screen. Even when they play computer games, they will easily get access
to English since many instructions are written in English and many words in
dialogues are also expressed in English. Furthermore, participants now frequently
communicate with each other in English through the Internet. Information of
various kinds is also often shown in English. Examples are international commercials,
job opportunities, admission instructions, computer-human dialogues, shopping
centers and vehicles manuals
As a productive skill, writing tends to be more subject to forgetting (Weltens
and van Els, 1986). Lack of practice might have contributed to this result.
Participants might have rarely written English after graduation since there
was no immediate access to English in the working situation. Furthermore, participants
writing skills might be poor even when they were learning on campus which gave
rise to a relatively lower proficiency compared with reading and listening skills.
Lower proficiency might be more easily attrited than the higher proficiency
according to the threshold hypothesis. Reading is easily accessed due to the
fact that participants were immerged in English slogans, English ads, English
TV programs and English instructions nearly everywhere. It is not necessary
for participants to work hard if they want to read English. However, writing
needs enough patience and concentration to accomplish. Participants have to
focus on some topic and organize ideas by thinking over which needs painstaking
effort. This might have caused participants to be unwilling to write since it
was not so, easily handled.
Being unwilling to write might act as a factor which resulted in their writing
skills having not reached the threshold. The writing skills might be administered
in the brain by some interacted neural units. If the skills were not proficient
enough for them to unite together steadily, then the neural units would be ephemeral,
which led to the fact that the acquired writing skills were easily forgotten.
On the contrary, if the skills reached or over passed a threshold, then the
neural units might be able to unite together tightly and steadily which formed
a relatively consolidated block against forgetting. The leading reason why participants
writing skills decreased significantly might be that their skills had not attained
to the threshold level; the neural units managing the writing might have not
formed a steady block, while the reading might have.
The assessment criteria might have influenced the results of writing. Much
literature studied writing assessment. An example is ecological model of writing
assessment which seeks to provide students, teachers, departments and institutions
with fuller, richer accounts of the breadth of students literate experiences
and explores how those experiences impact their abilities to accomplish academic
tasks throughout the undergraduate years and beyond (Wardlea
and Roozenb, 2012). In this study, the writing performance was conducted
based on CET 6 scoring criteria ignoring ecological perspective of literate
development. Ecological factors such as scorers mood, physical condition,
surroundings and psychological state might have influenced the scoring results.
Listening, as a receptive skill, might be more resistant against forgetting
compared with writing. During the complex process of listening, a person should
be mentally active in order to differentiate the words, stress, intonation and
grammatical structure of listeners, to constitute meaningful combinations by
making classifications or combinations, to fill in the gaps logically by using
background knowledge, to keep in mind and evaluate what has been listened to
and to construct meaning (Long, 1989; Vandergrift,
The process of listening comprehension can only succeed with the participation
of nerves. As the stimuli increase, the response of nerves will be more intense
and the sensitivity to the stimuli will also be higher. With higher sensitivity,
the activation of nerves will be easier. After frequent English input in the
daily life, participants might have cultivated higher sensitivity to English
input, which might have been memorized by the brain. When hearing English next
time, the stored memory would be activated more easily than those not memorized
and stored. Participants might be able to access English listening frequently
since nowadays English is everywhere and participants can hear English without
any great effort, which might have cultivated the high sensitivity to English
listening and have made participants reach the threshold of listening comprehension.
Participants might have successfully reached the critical stage before two years
and then became more resistant against forgetting.
Research into the role of self-efficacy in listening seems to underline its
importance in terms of listening performance. Within a certain context, self-efficacy
would seem to have particular relevance. Nothing can fully prepare a second-language
student for the experience of listening to a full-length lecture or participating
in a rapid exchange of views in a seminar. However, it is important to ensure
that these experiences do not give rise to the kind of listening anxiety that
forces the student to rely almost exclusively upon visual input (PowerPoint
slides and handouts) in the first case and remain totally silent in the second.
The knowledge that he or she is capable of making sense of what is said to the
extent of picking out critical words and phrases and main ideas will give much-needed
confidence in the early days of an academic course. It will also mark the first
step in the gradual and usually imperceptible process that occurs as a listener
becomes attuned to a language by dint of extended exposure to it (Graham,
2011). Participants who must have known that this test would not mark their
proficiency of English, might have relaxed themselves when taking the examination.
Strong self-efficacy might have helped them perform better than the real CET
6 achievement test.
Computer anxiety and achievement in English were interrelated in a reverse
order, meaning that more computer anxiety caused poorer achievements and vice
versa. Students with higher English proficiency used computers (both online
and offline) more frequently than those with lower English proficiency. Computer
ownership was also significantly related to students success in English.
It was also found that except gender, achievement in English, PC time and computer
ownership were predictors of computer anxiety (Rahimi and
Yadollahi, 2011). All of the participants in this study received computer
aided English education. But two years later when they did not learn English
on campus, they did not learn English, let alone use computer as a tool to learn
English. This might account for the significant decrease in English skills.
Doubtlessly, computers are satisfactory tools for English learning. Language
acquisition can be facilitated by use of computers. Computers can realize various
modes of language acquisition, such as multi-media projecting, online listening,
speaking, reading, writing and testing. Computer programs can also be used to
analyze learning data and to design proper learning schedules. In the classroom,
a sea of information can be displayed for students to absorb. Clickers is also
considered a useful tool for students and instructors to provide effective and
timely response and thus interact with each other.
The understanding of human learning is increasingly informed by findings from
multiple fields-psychology, neuroscience, computer science, linguistic and education.
A convergence of insights is forging a new science of learning within
cognitive science (Heinz and Idsardi, 2011). To validate
the threshold hypothesis is a difficult task for linguists to accomplish alone.
To absorb cross-disciplinary knowledge and analyze data by using knowledge in
multi-faceted fields is necessary. Furthermore, even though threshold hypothesis
in language forgetting is nearly universally acknowledged, specific thresholds
for different languages in different countries and areas might still not be
easily determined. Scientists in many related fields, such as psycholinguistics,
neuro-linguistics, cognition, neurology, medicine and computer, should work
together. The integration between different disciplines is helpful for studies
on the threshold hypothesis in the field of language forgetting.
Computer technologies such as clickers might be able to facilitate the process
of reaching the threshold of English proficiency. Although not widely accepted
in China, use of clickers is a popular way in USA to produce an interactive
situation between students and teachers. Previous literature showed that students
thought clickers enhanced the learning experience (Prather
and Brissender, 2009). Several studies indicated that students believed
clickers aid in correcting misunderstandings of course concepts (Bode
et al., 2009). Clickers improved thelearning settings due to increase
in interaction between students and instructor; there was a natural participatory
quality to use of clickers that required student participation, inhibiting passive
learning (Hoekstra, 2008).
In general, the threshold hypothesis in computer aided English learning might
be well accepted in this study. Therefore, it is urgent for learners and instructors
to use computer technologies to improve learners
English learning so that their English proficiency can attain to the threshold
hypothesis as soon as possible. In this way, language forgetting may be avoided
to a certain extent.
The author wishes to thank the people who help this study and the projects
which financially support this study: 2011 Youth Fund of Humanities and Social
Sciences of Ministry of Education of China. The Regression and Threshold Hypotheses
of English Negative Sentences Attrition among English Learners in China, (Project
No.: 11YJC740138); 2011 Teaching Innovation Project of Tongda College of Nanjing
University of Posts and Telecommunications. The Regression and Threshold Hypotheses
of Foreign languages and Teaching Innovation of College English in Civil Colleges
(Project No.: TD02011JG02); The Second Batch of Post-doctoral Research Fund
of Jiangsu Province in 2012. The Regression and Threshold Hypotheses of English
Language Attrition among Students in China (Project No.: 1202112C), 2013 Philosophy
and Social Science Guidance Research Project of Education Bureau of Jiangsu
Province (Project No.: 2013SJD740005), and Special Fundamental Research Fund
for the Central Universities (Project No.: 2013B33914).